60810 Managing Openness TRADE AND OUTWARD-ORIENTED GROWTH AFTER THE CRISIS H Managing Openness Managing Openness Trade and Outward- Oriented Growth after the Crisis Mona Haddad and Ben Shepherd, Editors © 2011 The International Bank for Reconstruction and Development/The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org All rights reserved 1 2 3 4 14 13 12 11 This volume is a product of the staff of the International Bank for Reconstruction and Development/The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denomina- tions, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permis- sion may be a violation of applicable law. The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750- 4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN: 978-0-8213-8631-6 eISBN: 978-0-8213-8632-3 DOI: 10.1596/978-0-8213-8631-6 Library of Congress Cataloging-in-Publication Data Managing openness : trade and outward-oriented growth after the crisis / edited by Mona Haddad and Ben Shepherd. p. cm. Includes bibliographical references and index. ISBN 978-0-8213-8631-6--ISBN 978-0-8213-8632-3 (electronic) 1. International trade. 2. Balance of trade. 3. International economic relations. 4. Global economy. I. Haddad, Mona. II. Shepherd, Ben. III. World Bank. HF1379.M3555 2011 382'.3--dc22 2011001903 Cover photo: Corbis Cover design: Tomoko Hirata, World Bank contents About the Editors and Contributors xv Acknowledgments xvii Abbreviations xix 1 Managing Openness: From Crisis to Export-Led Growth, Version 2.0 1 Mona Haddad and Ben Shepherd Roots of the Crisis and Global Imbalances 1 The Impact of the Global Financial Crisis on Trade and Protectionism 3 Should Countries Reassess Export Orientation? 5 Changing Dynamics in Global Trade 7 The New Trade Policy Agenda 8 Conclusion 9 Bibliography 10 Part I THE CRISIS, GLOBAL IMBALANCES, AND REBALANCING 11 2 Managing Openness: Lessons from the Crisis for Emerging Markets 13 Barry Eichengreen Who Was Hit, and Why? 13 Why Was the Collapse of Trade So Dramatic? 17 The Role of Global Imbalances 20 How Should Emerging Markets Respond? 22 In Sum 23 Notes 24 Bibliography 24 3 Global Imbalances: Past and Future 27 Luis Servén and Ha Nguyen The Nature of Global Imbalances 29 The Future of Global Imbalances and Implications for Developing Countries 33 Conclusion 37 Notes 38 Bibliography 38 4 Rebalancing Trade after the Crisis 41 Caroline Freund The Decline in Global Imbalances: Rebalancing versus the Trade Collapse 42 Rebalancing across Countries and Future Trade Growth 48 v vi Contents Conclusion 53 Notes 54 Bibliography 54 5 An Anatomy of Trade in the 2008­09 Crisis 55 Mona Haddad, Ann Harrison, and Catherine Hausman Changes in the Intensive and the Extensive Margins 55 Demand and Supply Shocks 56 Variation by Product Type 58 Variation by Income Group 59 Were Any of These Trends Present before the Crisis? 60 Conclusion 60 Notes 62 Bibliography 62 6 Developing Countries, New Trade Barriers, and the Global Economic Crisis 63 Chad P. Bown and Hiau Looi Kee Protectionism from the Perspective of Domestic Industries and Importing Economies 64 Developing-Country Exporters and the Incidence of Crisis-Era Protectionism 73 Policy Implications and Conclusions 81 Notes 82 Bibliography 82 7 Turning toward China? The 2008 Crisis and Its Influence on Brazil's Development Model 85 Gustavo H. B. Franco and Fausto J. A. Vieira Trends in Brazilian Trade Orientation before the 2008 Crisis 85 The 2008 Crisis: Impacts and Responses 89 The "Exit" and New Directions in Development 93 Bibliography 96 Part II OPEN BUT NOT DEPENDENT: SOUTH-SOUTH TRADE AND EXPORT DIVERSIFICATION 97 8 Changing Dynamics in Global Trade 99 Gordon Hanson The New Drivers of Global Trade 100 The Effects of Changes in Demand on Supply Patterns 102 Conclusion 105 Notes 105 Bibliography 106 9 Sources of Export Growth in Developing Countries 107 Gordon Hanson Recent Trends in Global Trade and GDP 107 Are Global Trade Patterns Sustainable in the Future? 111 Conclusion 116 Notes 116 Bibliography 117 Contents vii 10 China's Trade and Investment with the South Pre- and Postcrisis 119 Jing Wang and John Whalley China's Trade with the South, 1995­2007 120 China's Future Trade with the South 122 China's FDI Flows with the South 123 The Financial Crisis and China's Southern Trade 126 China's Trade Performance and That of Other Asian Countries during the Financial Crisis 127 Institutional Dimensions of China's Growing Southern Links 132 Concluding Remarks 134 Note 134 Bibliography 134 11 Volatility, Export Diversification, and Policy 135 Mona Haddad, Jamus Lim, Laura Munro, Christian Saborowski, and Ben Shepherd The Effect of Increased Trade Openness on Vulnerability to Global Shocks 136 Most Countries Benefit from Increased Openness 136 Facilitating Export Diversification 139 Conclusion 142 Notes 143 Bibliography 143 12 The Effects of Exports on Productivity and Volatility: Evidence from Malaysian Firm-Level Data 145 Mona Haddad, Deborah Winkler, and Albert Zeufack The Effect of Exports on Productivity 145 The Effect of Exports on Output Growth Volatility 149 Conclusion 151 Bibliography 153 Part III LESSONS IN MANAGING OPENNESS FROM COUNTRY AND REGIONAL EXPERIENCES 155 13 The International Crisis and Latin America: Growth Effects and Development Strategies 157 Vittorio Corbo and Klaus Schmidt-Hebbel Latin America's Growth Performance 158 Explaining the Amplitude of the 1998­99 and 2008­09 Recessions 158 Implications for Policies and Growth Strategies 162 Final Remarks 166 Notes 167 Bibliography 167 14 The Economic Crisis of 2008­09 and Development Strategy: The Mexican Case 169 Jaime Ros Dimensions of the Crisis 169 The Channels of Transmission 171 The Policy Response 173 Why Was the Crisis So Severe? 174 Looking Ahead: The Crisis, Macroeconomic Policy, and Development Strategy 176 The Reform of Monetary and Exchange-Rate Policy 179 viii Contents Conclusions 181 Notes 181 Bibliography 181 15 The International Crisis and Development Strategies: The Case of Chile 183 Roberto Zahler Chilean Development Strategy 183 World Economic Cycle and Chile's Growth 187 Transmission Mechanisms of the 2008­09 Economic Crisis 190 Overall Impact of the 2008­09 Economic Crisis on Chile 199 Chilean Policy Reaction to the 2008­09 Crisis 200 Conclusions 205 16 The International Crisis and Development Strategies: The Case of Malaysia 209 Mahani Zainal Abidin State of the Malaysian Economy before the Global Crisis 209 Impact of the 2008 Global Crisis 210 Recovery 212 Malaysia's Development Strategies Going Forward 213 Proposals for Bringing Sustained High Growth to Malaysia 215 17 Should Indonesia Say Goodbye to Its Strategy of Facilitating Exports? 217 Muhammad Chatib Basri and Sjamsu Rahardja The Impact of the Global Financial Crisis on Indonesia 217 Should We Say Goodbye to the Strategy of Facilitating Exports? 218 Export Diversification: The Indonesian Experience 221 How to Promote Export Diversification: The Role of Policy 227 Conclusion 229 Notes 230 Bibliography 230 18 India: Managing Openness for a Rapidly Developing Domestic Market 233 Ulrich Bartsch and Abhijit Sen Gupta India's Recent Growth Experience 234 India's Growth Model: Greater Openness but Also a Rapidly Developing Domestic Market 236 India's Approach to Capital Account Liberalization 239 Macroeconomic Management and Capital Flows 240 Annex: Framework for Quantifying Policy Choices under the Impossible Trinity 244 Notes 245 Bibliography 245 19 Exports and Export Diversification in Sub-Saharan Africa: A Strategy for Postcrisis Growth 247 Vera Songwe and Deborah Winkler The Importance of Export Structure 248 Trends in Exports and Export Diversification 248 The Effect of Exports and Export Diversification on Growth 250 The Economic Crisis in Sub-Saharan Africa 254 Policy Implications for Postcrisis Export Strategies 256 Bibliography 257 Contents ix Part IV EMERGING TRADE POLICY ISSUES IN THE POSTCRISIS ENVIRONMENT 259 20 Structural Changes in Commodity Markets: New Opportunities and Policy Challenges for Commodity Exporters 261 Donald Mitchell and Enrique Aldaz-Carroll Structural Changes in Commodity Markets 262 Prospects for Commodity Prices 265 A Historical Account of the Commodities Literature 267 Policy Recommendations to Foster a Sectorally Balanced, Resource-Based Economy 269 Conclusions 273 Notes 273 Bibliography 274 21 Global Production Networks in the Postcrisis Era 275 Alyson Ma and Ari Van Assche Drivers in the Spread of GPNs 275 Mapping Global Production Networks: China's Processing Trade Regime 276 Peak Oil and Intra-GPN Trade 279 Intra-GPN Trade and the Global Recession 280 Conclusion 283 Notes 283 Bibliography 284 22 The 2008­09 Recession: Implications for International Labor Migration 287 Philip Martin Recent Trends in Population and Migration 288 Future Supply of Migrants 290 Economic Implications of Migration 295 Conclusions and Recommendations 296 Notes 297 Bibliography 298 23 Trade and Climate Policies after the Crisis 301 Jaime de Melo and Nicole A. Mathys The Contours of the Next Multilateral Climate Agreement 302 The Role of Trade in GHG Mitigation 305 Climate Change and the WTO 308 Final Reflections: Lessons from World Trade for Governing Climate Change 313 Notes 314 Bibliography 315 Index 317 Figures 1.1 Current Account Imbalances for Selected Countries as a Percentage of World GDP, 1980­2009 2 1.2 Gross Capital Inflows to the United States from Emerging Markets, 2000­09 3 1.3 Monthly Import Growth by Region, January 2008­November 2009 4 1.4 Diversification Trends among Selected Countries, 1981­2005 6 1.5 Combined G-20 Use of Selected Temporary Trade Barriers by Import Source, 1997­2009 8 1.6 Real Commodity Price Indexes, History, and Projection, 1960­2020 9 x Contents 2.1 Ratio of Current Account to GDP, 2007, versus Change in Growth, 2007­09, All Emerging/Developing Countries 14 2.2 Exports as a Percentage of GDP, 2007, versus Change in Growth, 2008­09, All Emerging/Developing Countries 15 2.3 Government Budget as a Percentage of GDP, 2007, versus Change in Growth, 2008­09, All Emerging/Developing Countries 16 2.4 International Reserves as a Percentage of Short-term External Debt, 2007, versus Change in Growth, 2008­09, All Emerging/Developing Countries 17 2.5 Short-term External Debt as a Percentage of GDP, 2007, versus Change in Growth, 2008­09, All Emerging/Developing Countries 18 2.6 Average Financial Credit Restrictions, 2005, versus Change in Growth, 2008­09, All Emerging/Developing Countries 19 3.1 Current Account Imbalances for Selected Countries as a Percentage of World GDP, 1980­2009 28 3.2 U.S. Bilateral Current Accounts with Other Countries, 1999­2009 28 3.3 Saving and Investment Rates as a Percentage of GDP in Selected Countries, 1990­2008 29 3.4 U.S. Multilateral Real Exchange Rate Index, 1980­2009 30 3.5 U.S. Valuation Effects and Current Account as a Percentage of GDP, 1994­2007 31 3.6 Foreign Reserve Stock of Industrial and Emerging Countries, 1991­2008 33 3.7 Gross Capital Inflows to the United States from Emerging Markets, 2000­09 34 3.8 U.S. Current Account as a Percentage of GDP, Quarters 2 and 4, 1990­2010 34 3.9 U.S. Gross Capital Inflows in Long-term Securities from Residents and Nonresidents, 1989­2010 35 3.10 Net Foreign Purchases of U.S. Long-term Securities and China's Net Foreign Asset Position, Quarters 1 and 3, 2000­10 35 3.11 Saving and Investment Forecasts for China and the United States as a Percentage of GDP, 1991­2013 36 4.1 Global Imbalances and Global Trade, 1970­2007 42 4.2 Growth of Imports in Brazil, China, India, and Indonesia and in the Rest of the World, 2006­10 51 4.3 The Changing Nature of Imports in Brazil, China, India, and Indonesia, 1998 and 2008 51 4.4 Share of Exports by Region and Income Group to Brazil, China, India, and Indonesia, 1997­98 and 2007­08 52 5.1 Index of Total Number of Products Traded in Brazil, the European Union, Indonesia, and the United States, January 2007­September 2009 56 5.2 Index of Total Value of Imports for Brazil, the European Union, Indonesia, and the United States, January 2007­September 2009 57 5.3 Change in Total Import Value of All Products in Brazil-Indonesia and United States­EU Trade, by Margin, 2008­09 57 5.4 Variations in Price and Quantity in Commodities and Manufactures, 2008 58 5.5 Changes in Import Values across Product Types and Shipment Methods for U.S. Imports as a Percentage of Value by Margin, 2008 59 5.6 Percentage Change in Imports to Brazil and Indonesia and to the United States and the EU from Countries in Different Income Groups, 2008­09 60 5.7 Percentage Changes in Imports of Manufactures and Commodities to the United States and Indonesia, Second Quarter, 2007­Third Quarter, 2009 61 6.1 Monthly Import Growth by Region, January 2008­November 2009 64 6.2 Combined G-20 Use of Temporary Trade Barriers, 1997­2009 65 6.3 Combined G-20 Use of Selected Temporary Trade Barriers by Import Source, 1997­2009 66 6.4 G-20 Developing Economies' Use of Selected Temporary Trade Barriers by Import Source, 1990­2009 69 6.5 G-20 High-Income Economies' Use of Selected Temporary Trade Barriers by Import Source, 1990­2009 70 6.6 Exports of High-Income Economies Subject to Selected Temporary Trade Barriers, 1990­2009 74 6.7 Exports of Developing Economies Subject to Selected Temporary Trade Barriers, 1990­2009 75 7.1 Trade to GDP for Selected Economies, 1995­2000 86 7.2 Openness Indicators, Brazil, 1960­2009 87 7.3 Investment and Growth of GDP in China and Brazil, 1967­2009 88 7.4 Ratio of Outstanding Credit to Total Assets in Brazil's Medium-Size and State-Owned Banks, 2008­09 92 7.5 Industrial Production in Brazil, 2000­10 93 7.6 GDP Growth in Brazil, 2008­10 94 Contents xi 8.1 Share of World Imports by Importer Income Group, 1996­2008 100 8.2 Share of World Exports by Exporter Income Group, 1996­2008 103 8.3 Average Export HHIs of Non-oil-Exporting Countries Weighted by Exports, 1997 and 2007 105 9.1 Log GDP with Trend and Cyclical Components in Selected Countries, 1978­2008 109 9.2 Change in Relative Export Competitiveness of Selected Countries by Industry, 2000­07 114 10.1 China's Exports and Imports with the South, 1994­2008 121 10.2 Projections of China's Southern Export Share, 2008­30 123 10.3 Projections of China's Southern Import Share, 2008­30 123 10.4 Projections of Southern FDI Inflow Share for China, 2008­30 126 10.5 Growth Rates of China's Southern Trade Pre­ and Post­financial Crisis Compared to Same Period of Previous Year 128 10.6 China's Regional Export Shares Pre­ and Post­financial Crisis, 2007­09 128 10.7 China's Regional Import Shares Pre­ and Post­financial Crisis, 2007­09 129 10.8 Export Growth Rates of Key Asian Economies during the Financial Crisis Compared with Same Month of Previous Year, 2008­10 129 10.9 Import Growth Rates of Key Asian Economies during the Financial Crisis Compared with Same Month of Previous Year, 2008­10 130 10.10 Growth Rates of Asian Economies' Exports to the United States during the Financial Crisis, 2008­10 131 10.11 Growth Rates of Asian Economies' Imports from the United States during the Financial Crisis, 2008­10 131 10.12 Growth Rates of Asian Economies' Exports to China in the Financial Crisis, 2008­10 132 10.13 Growth Rates of Asian Economies' Imports from China in the Financial Crisis, 2008­10 133 11.1 Total Effect of Trade Openness on Growth Volatility 137 11.2 Distribution of Countries by Product Diversification, Trade Openness, and Income per Capita 138 11.3 Diversification Trends among Selected Countries, 1981­2005 139 11.4 Histogram of the Number of 8-Digit Product Lines Exported to the EU by Each Partner Country 140 11.5 Histogram of the Number of Export Destinations Served by Each Country 140 11.6 Number of Products Exported versus Doing Business Cost of Exporting 141 11.7 Number of Export Markets versus Doing Business Cost of Exporting 142 12.1 Malaysian Exports, 1971­2008, and Their Contribution to Real GDP Growth, 1988­2009 146 12.2 Firm-Level Exports and Total Factor Productivity for the Malaysian Manufacturing Sector, 2006 148 12.3 Firm-Level Exports and Labor Productivity in the Malaysian Manufacturing Sector, 2006 148 12.4 Average Annual Exports and Volatility Growth by One-Digit SITC Sector, 1992­2005 151 12.5 Product and Market Herfindahl-Hirschman Index by SITC1, 1989­2007 152 13.1 Average GDP Growth in Latin America, 1990­2010 158 13.2 GDP Growth in Selected Latin American Countries, 1990­2010 159 13.3 Average Output Gap in Latin America, 1990­2010 160 13.4 Average GDP Growth around the Asian Crisis, 1998­99, and the Global Financial Crisis, 2008­09 161 14.1 Growth Rates of Components of Mexico's GDP, 2008­09 171 14.2 Value of Mexico's Exports of Goods, 2008­09 172 14.3 Mexico's Exports to the World, 2008 175 14.4 Infrastructure Investment by Selected Latin American Countries, 1981­86 and 2001­06 178 15.1 Effective Average Tariff on Chilean Imports, 1998­2009 185 15.2 Growth in Exports of Chilean Goods and Services Relative to GDP, 1987­2009 186 15.3 Nominal and Real Price of Copper in Chile, 1960­2010 187 15.4 Growth of GDP in Chile and Selected Country Groups, 1980­2009 188 15.5 World, Latin American, and Chilean Growth in GDP, 1961­2009 189 15.6 Growth and Volatility of Chilean GDP, 1965­2009 189 15.7 Evolution of Chile's Real GDP in the Past Two Crises, 1998­2000 and 2008­09 190 15.8 Ratio of Exports to GDP in 2008 and GDP Growth in Selected Countries, 2009 192 15.9 Chile's Commercial Openness, 1986­2009 192 15.10 Chile's External and Internal Demand for Goods and Services, 1970­2009 193 15.11 Changes in Prices and Quantities of Chilean Exports, 1997­2009 194 15.12 Diversification of Chilean Exports by Destination, 1980­2009 194 15.13 Chile's Exports to the United States as a Percentage of Total Exports, 2003­09 194 xii Contents 15.14 Changes in Price and Quantity of Chilean Mining Exports, 1997­2009, as of Third Quarter 2009 196 15.15 Chilean Industrial Exports, 2009 196 15.16 Change in Price and Quantity of Chilean Industrial Exports, 1997­2009 197 15.17 Chile's Terms of Trade, 2003­09 197 15.18 Value of Chilean Exports of Services and Percentage Change, 2003­09 198 15.19 Foreign Direct Investment and Portfolio Investment in Chile, 1996­2009 198 15.20 NAFTA Foreign Direct Investment in Chile, 1974­2008 199 15.21 Three-Month LIBOR and Five-Year CDS for Chile, 2008­09 200 15.22 Forecasts of Chile's GDP Growth for 2009, June 2008­December 2009 201 15.23 Contribution of the Components of Aggregate Demand to Chile's GDP Growth, First Quarter 2008­Fourth Quarter 2009 201 15.24 Index of Consumer Confidence in Chile, October 2008­December 2009 202 15.25 Monetary Policy Interest Rates in Selected Latin American Countries, January 2007­March 2010 204 15.26 Chile's Real Exchange Rate Index, 1976­2008 204 15.27 Market Value of Chile's Sovereign Wealth Fund, 2006­09 206 15.28 Chile's Fiscal Balance, 2003­09 206 16.1 Real GDP Growth in Malaysia, 2008­10 211 16.2 Merchandise Export Growth in Malaysia, Quarter 1, 2008­Quarter 1, 2010 211 17.1 Export Values in Selected Markets, Quarterly Growth, Seasonally Adjusted, 2008­09 220 17.2 Impulse Response Function of GDP due to Export and Domestic Demand Shocks in Indonesia 220 17.3 Variance Decomposition of GDP due to Exports and Domestic Demand Shocks in Indonesia 220 17.4 Expansion outside Java because of the Commodity Boom, 2007­10 222 17.5 Indonesian Oil and Non-oil Exports, 1970­96 223 17.6 Composition of Indonesia's Non-oil Exports, 1970­96 223 17.7 Concentration of Indonesia's Export Products, 1980­95 224 17.8 Product and Country Concentration of Indonesia's Exports in the 1980s and 1990s 224 17.9 Pairs of Markets and Products in Indonesia's Exports, 2004­08 225 17.10 Index of Export Product Concentration in Indonesia, 2001­08 225 17.11 Relationship of Indonesia's Commodity Exports, Manufacturing Exports, and Real Exchange Rate, 2000­08 226 17.12 Contribution of Intensive and Extensive Margins to Indonesian Export Growth, 1990­2008 227 18.1 GDP Per Capita Growth in India, 1970­2009 234 18.2 Composition of GDP in India, 1965­2009 235 18.3 Direction of India's Trade, 1987­88 through 2007­08 237 18.4 Income Strata in India, 1985­2025 240 18.5 Cross-Country Comparison of Capital Account Openness, Selected Countries, 2000­07 241 18.6 Configurations of the Impossible Trinity and Reserve Accumulation in India, 1993­2007 243 19.1 Exports and Exports of Goods and Services in Sub-Saharan Africa 249 19.2 Annual Average Growth of Export Goods by Destination, Sub-Saharan Africa, 2000­08 250 19.3 Annual Commodity Export versus Value-Added Growth in 30 Sub-Saharan African Countries, 1995­2008 251 19.4 Annual Manufacturing Export versus Value-Added Growth in 30 Sub-Saharan African Countries, 1995­2008 251 19.5 Export Concentration of Products versus Value Added in 30 Sub-Saharan African Countries, 1995­2008 252 19.6 Export Concentration of Markets versus Value Added in 30 Sub-Saharan Countries, 1995­2008 253 19.7 Trade Volume Decline from Peak to Trough, by Region, 2008­09 255 19.8 Goods Exports for Six World Regions, Constant and Seasonally Adjusted, 2007­10 255 19.9 Shares of Sub-Saharan African Goods Exports, by Destination, 2000, 2008, and 2009 256 20.1 Real Commodity Prices, 1960­2009 263 20.2 China's Shares of Global Metals Consumption, 1996­2008 264 20.3 Share of Growth of World Oil Consumption of OECD Countries, Non-OECD Countries, and China, 1988­2008 264 20.4 Per Capita Food Production of World and Selected Regions, 1961­2007 265 20.5 Real Commodity Price Indexes, Historical and Projected, 1960­2020 266 20.6 China's Metals Consumption versus GDP as Compared to Selected Other Economies, Various Years 268 Contents xiii 20.7 Energy Subsidies in Indonesia, 2008 271 20.8 Benefits of Fuel Subsidies for the Rich and the Poor, 2007 271 21.1 Processing Trade in China as a Share of Total Trade, 1988­2008 276 21.2 Domestic and Foreign Content as a Share of China's Processing and Nonprocessing Exports, 2006 277 21.3 Share of China's Processing Imports by Region of Origin, 1988­2008 277 21.4 Share of China's Processing Exports by Destination, 1988­2008 277 21.5 Processing Imports as a Share of China's Total Imports by Economy of Origin, 2007 278 21.6 Processing Exports as a Share of China's Total Exports by Destination Economy, 2007 279 21.7 Average Prices for Crude Oil, 1980­2009 279 21.8 Processing Exports as a Share of China's Total Exports by Technology Level, 1992­2007 280 21.9 China's Processing Imports, 2008, versus China's Import Growth by Economy of Origin, Quarter 1, 2008­Quarter 1, 2009 283 23.1 Total and Per Capita Carbon Emissions for the World and Various Regions, 1960­2002 303 23.2 Embodied CO2 Emissions in Imports and Exports for Selected Countries, 2001 306 23.3 Justifying Border Measures under WTO Law in a Nutshell 310 Tables 4.1 Change in Trade Imbalances due to Declines in Trade and Rebalancing Selected Economies, 2009 43 4.2 Change in Global Trade Imbalance, 2007­09 46 4.3 Global Trade Balance Adjustment in Asian Crisis Economies, 1996­98 47 4.4 Import and Export Growth in Deficit and Surplus Economies, 2007­09 49 4.5 Intensive and Extensive Margin of Total Imports for Brazil, China, India, and Indonesia by Income Group, 2007­09 52 6.1 Stocks of Temporary Trade Barriers Imposed by Individual G-20 Economies, 2009 67 6.2 G-20 Imports Subject to Newly Imposed Temporary Trade Barriers, 2008­09 71 6.3 Overall Trade Restrictiveness of G-20 Economies, 2008­09 72 6.4 Major Exporters Subject to Stock of Selected G-20 Temporary Trade Barriers, 2009 77 6.5 The Value of Exports Subject to G-20 Temporary Trade Barriers Newly Imposed in 2008­09 78 6.6 Examples of Major Changes in MA-OTRI due to G-20 Trade Policy Changes, 2008­09 80 7.1 GDP Growth and Various Estimates of Labor and Total Factor Productivity Growth in Brazil, 1985­2007 88 7.2 Estimation Results, Model of Recession, and GDP Growth in Brazil 90 8.1 Share of Exports by Country Income Groups and Destination Market, 2000 and 2008 101 9.1 Export Shares by Sector and Developing-Country Income Group, 2000 and 2007 108 9.2 Export Shares by Destination and Developing-Country Income Group, 2000 and 2007 109 9.3 Mean Change in Export Competitiveness by Exporter Income Group, 2000­07 112 9.4 Mean Change in Export Competitiveness by Exporter Region, 2000­07 113 10.1 China's Total Trade (Imports and Exports) with the South, 1995, 2001, and 2007 120 10.2 Average Annual Growth Rates in China's Trade with the South, Pre- and Post-2001 122 10.3 China's Southern Trade Shares in Major Product Groups, 2007 122 10.4 China's Average FDI Inflows and Outflows to the South and the World, 1997­2007 124 10.5 China's FDI Inflows from and Outflows to the South, 2003 and 2007 124 10.6 Average Annual Growth of China's Southern and Northern FDI Flows, 1998­2007 125 10.7 Growth of China's Southern FDI and Total FDI, 1998­2007 125 10.8 China's Southern Trade Pre­ and Post­financial Crisis, 2008 and 2009 126 10.9 Commodity Composition of China's Imports in the Financial Crisis, 2008 and 2009 127 11.1 Thresholds and Shares Based on Error Components and System GMM Regressions for Growth Volatility on Openness and Product Diversification 137 13.1 Growth of GDP in Latin American Recessions, 1998­99 and 2008­09 161 13.2 Decomposition of Latin America's Recessions, 1998­99 and 2008­09 162 13.3 Public and Publicly Guaranteed External Debt in Latin America as a Percentage of GDP, 1990­2009 163 13.4 Inflation in Latin America, 1990­2009 164 13.5 Trade Openness in Latin America, 1990­2009 165 13.6 Financial Openness in Latin America, 1990­2009 166 xiv Contents 14.1 GDP Growth in Mexico by Sector and Employment Indicators, 2008 and 2009 170 14.2 Mexico's Balance of Payments, 2008 and 2009 172 14.3 Size of the Fiscal Stimulus as Percentage of GDP, 2008­10 175 14.4 A Comparison of Selected Economic Data on Canada and Mexico 176 15.1 Basic Macroeconomic Data for Chile, 1986­2009 184 15.2 World GDP Growth in Recessions, 1982­83, 1998­99, and 2008­09 188 15.3 Effect of the International Crisis on the Exports of Selected Countries, as of December 2009 191 15.4 Chilean Export Diversification by Destination, 1982­2009 195 15.5 Chile's Export Products, 1982­2009 195 15.6 Percentage Change in Chilean GDP, Year on Year, First Quarter 2008­Fourth Quarter 2009 200 16.1 Malaysia's Economic Conditions before the Global Crisis, 2007 210 16.2 Malaysia's GDP by Sector, 2007­Quarter 1, 2010 212 17.1 Exports at Current and Constant Prices, by Major Commodities, 1975 to Quarter 1, 2009 219 17.2 Co-movements of Innovations in Private Consumption and Innovations of GDP Components, Quarter 1, 2000­Quarter 4, 2008 221 18.1 Factor Intensities of Indian Exports and Domestic Production, Selected Years, 1962­2008 238 18.2 Growth in Exports, Different Factor Intensities, 1990­2008 239 18.3 Household Consumption and Exports, Level and Increment, 2002­08 240 19.1 Actual and Projected GDP Growth Rates, Selected Regions, 2007­11 254 20.1 Percentage Growth of Merchandise, Manufacturing, and Commodities Export Values and Shares of Merchandise Trade, 2005­08 262 20.2 Share of Commodities in Total Merchandise Exports for 57 Commodity-Dependent Countries, 2007 262 20.3 Summary of Policy Options for Mitigating the Impact of Price Volatility 270 21.1 Breakdown of China's Exports by Level of Technology, 2008­09 281 21.2 China's Processing Exports as a Share of Total Exports, HS Eight-Digit Level, Quarter 1, 2008, and Quarter 1, 2009 281 21.3 Changes in China's Processing Trade by Industry, Quarter 1, 2008­Quarter 1, 2009 282 22.1 Percentage of World Population by Continent, 1800­2050 288 22.2 Population and Migration, 2010 289 22.3 The World's Economically Active Population, 1980­2050 290 22.4 Economically Active Population in More- and Less-Developed Countries, 1980­2020 291 22.5 Replacement Migration in Europe and the United States, 1995 293 About the Editors and Contributors Mona Haddad is Sector Manager of the World Bank's International Trade Department. In this capacity, she manages the group responsible for supporting the development and implementation of trade-related activities at both country and regional levels. These include trade policy analysis, competitiveness, and trade facilitation. Prior to joining the Trade Department, she was the Regional Trade Coordinator for the East Asia Region, where she worked on trade issues in various countries, including China, Indonesia, the Lao People's Democratic Republic, and Vietnam. Ben Shepherd is the Principal of Developing Trade Consultants Ltd. He is a trade economist and international develop- ment consultant, having worked on a wide range of trade and development issues with organizations such as the World Bank, the Organisation for Economic Co-operation and Development, the Asian Development Bank, the United Nations, and the Food and Agriculture Organization. He specializes in providing policy-relevant research, as well as capacity- building seminars for researchers working in trade and development. About the Contributors Mahani Zainal Abidin, Chief Executive, Institute of Strategic and International Studies, Malaysia Enrique Aldaz-Carroll, Senior Economist, World Bank, Indonesia Ulrich Bartsch, Senior Economist, World Bank, India Muhammad Chatib Basri, Senior Lecturer, Department of Economics, University of Indonesia Chad P. Bown, Senior Economist, Development Research Group, World Bank Vittorio Corbo, Centro de Estudios Públicos, Chile Jaime de Melo, Professor, Department of Economics, University of Geneva; Center for Economic Policy Research; Founda- tion for International Development Study and Research Barry Eichengreen, George C. Pardee and Helen N. Pardee Professor of Economics and Political Science, University of California, Berkeley Gustavo H. B. Franco, Chairman, Rio Bravo Investimentos Caroline Freund, Lead Economist, Development Research Group, World Bank Abhijit Sen Gupta, Associate Professor, Jawaharlal Nehru University, India, and Consultant, World Bank, India Mona Haddad, Sector Manager, International Trade Department, World Bank Gordon Hanson, Professor of Economics and Director, Center on Pacific Economies, School of International Relations and Pacific Studies, and Department of Economics, University of California, San Diego Ann Harrison, Director of Development Policy, Development Research Group, World Bank xv xvi About the Editors and Contributors Catherine Hausman, Ph.D. Student, Department of Agricultural and Resource Economics, University of California, Berkeley Hiau Looi Kee, Senior Economist, Development Research Group, World Bank Jamus Lim, Economist, Development Prospects Group, World Bank Alyson Ma, Associate Professor of Economics, School of Business Administration, University of San Diego Philip Martin, Professor and Chair, Comparative Immigration and Integration Program, Department of Agricultural and Resource Economics, University of California, Davis Nicole A. Mathys, Swiss Federal Office of Energy and University of Neuchatel, Switzerland Donald Mitchell, Consultant, World Bank, Tanzania Laura Munro, Graduate Student, Columbia University and the London School of Economics Ha Nguyen, Economist, Development Research Group, World Bank Sjamsu Rahardja, Senior Economist, East Asia Financial and Private Sector Unit, World Bank Jaime Ros, Professor of Economics, Universidad Nacional Autónoma de México and University of Notre Dame Christian Saborowski, Consultant, World Bank Klaus Schmidt-Hebbel, Professor of Economics, Catholic University of Chile Luis Servén, Senior Adviser, Development Research Group, World Bank Ben Shepherd, Principal, Developing Trade Consultants Ltd. Vera Songwe, Lead Economist and Adviser, Office of the Managing Director Africa, South Asia, Eastern and Central Europe, and Human Resources, World Bank Ari Van Assche, Associate Professor, Department of International Business, HEC Montréal Fausto J. A. Vieira, Chief Economist, Rio Bravo Investimentos Jing Wang, Postdoctoral Fellow, University of Western Ontario; Associate Researcher, Chinese Academy of Social Science John Whalley, Professor of Economics, University of Western Ontario, and Distinguished Fellow, Center for International Governance Innovation, Waterloo, Ontario Deborah Winkler, Consultant, International Trade Department, World Bank Roberto Zahler, President, Zahler & Co. Albert Zeufack, Director, Khazanah Research and Investment Strategy, Malaysia Acknowledgments First versions of a number of chapters in this volume were presented at the "Managing Openness: Outward-Oriented Growth Strategies after the Crisis" conference held at the World Bank in Washington, D.C., on May 10, 2010. The editors are grateful to Ann Harrison, who co-organized the conference, as well as to Chad Bown, Barry Eichengreen, Gustavo Franco, Klaus Schmidt-Hebbel, and John Whalley, who presented their work. Otaviano Canuto, Luis Ernesto Derbez, Homi Kha- ras, and Justin Lin also participated as panelists, and we are grateful for their many questions and insights. The conference brought together academics, researchers, and policy makers from around the world, and participants' contributions were very helpful in shaping the final product. Many people provided assistance during the manuscript's preparation. We are grateful to Homi Kharas for providing overall guidance. Laura Munro contributed outstanding editorial assistance. Sebastian Saez also reviewed key chapters and provided important inputs. Stacey Chow played a key role in organizing the May 2010 conference, and she provided con- stant program support during the book's preparation, as did Cynthia Abidin. The World Bank's Office of the Publisher managed the publication process very efficiently, and we are grateful in particular to Mary Fisk and Stephen McGroarty for their help and guidance. xvii Abbreviations ASEAN Association of Southeast Asian Nations ASI Annual Survey of Industry BRICs Brazil, the Russian Federation, India, and China BTA border tax adjustment CDS credit default swaps CEPAL United Nations Economic Commission for Latin America CPO crude palm oil EAP economically active population ETS emission-trading system EU European Union FTA free trade agreement GATS General Agreement on Trade in Services GATT General Agreement on Tariffs and Trade GDP gross domestic product GEM growing emerging market GMM generalized method of moments GHG greenhouse gas GNI gross national income GPN global production network GTI global trade imbalance H1N1 a type of influenza virus HHI Herfindahl-Hirschman Index HS harmonized system IMF International Monetary Fund KP1 first Kyoto Protocol LDC least-developed country LNG liquefied natural gas LP labor productivity M2 the total amount of money available in a given economy at a particular time MA-OTRI market access overall trade restrictiveness index MIER Malaysian Institute of Economic Research MERCOSUR Southern Common Market MFN most-favored nation MP Montreal Protocol MSS market stabilization scheme NAFTA North American Free Trade Agreement NBER National Bureau of Economic Research n.e.s. not elsewhere specified OECD Organisation for Economic Co-operation and Development OPEC Organization of the Petroleum Exporting Countries OTRI overall trade restrictiveness index xix xx Abbreviations PDBC bonds Pagarés Descontables del Banco Central PEMEX Mexico's state oil monopoly R&D research and development RBI Reserve Bank of India Re Indian rupee repo repurchase SACU Southern Africa Customs Union SAR special administrative region SITC Standard International Trade Classification SRR statutory reserve requirement TFP total factor productivity ToT terms of trade TTB temporary trade barrier TWP temporary worker program UNCTAD United Nations Conference on Trade and Development UNFCC United Nations Framework Convention on Climate Change WTO World Trade Organization yoy year on year 1 Managing Openness: From Crisis to Export-led Growth, Version 2.0 Mona Haddad and Ben Shepherd The global financial crisis is stimulating a broad reassess- thanks to a solid multilateral regime as well as to a new sys- ment of economic integration policies in developed and tem of production sharing across countries, which does developing countries alike. The crisis was associated with a not lend itself naturally to broad-based protectionism. great trade collapse--the sharpest in recorded history and Moreover, the role of South-South trade is growing, giving the deepest since World War II (Baldwin 2009). The trade developing countries new opportunities to export and new collapse affected all countries and products, although to opportunities to import cheaper capital goods--now pro- different extents. While signs of recovery are starting to duced in countries like China or India--that allow them to solidify, deeper questioning of the causes of the crisis and industrialize faster. Thus, while outward-oriented growth the merits of globalization has surfaced. The emergence of is here to stay, it needs to be put in a different perspective China and the imbalances of its trade with the United and packaged with additional policies. As the world States are shaking the stability of the global system. Are emerges from the crisis, we expect to see the development these imbalances sustainable, or do they need to be of an "export-led growth version 2.0" model that reflects adjusted to avoid another global crisis? What impact would these new dynamics. these adjustments have on the trade of developing coun- tries if they mean that China consumes more and the Roots of the Crisis and Global Imbalances United States saves more? Openness has helped support growth in many countries--to unprecedented levels in The crisis started in the industrial world but spread rap- Brazil, China, Indonesia, Malaysia, and others. Yet today idly to developing countries through financial and trade many are concerned that openness is creating vulnerability, channels. Its financial origins are embedded in the and vulnerability can hurt growth. No one believes that flawed regulation and perverse incentives that governed inward orientation is the solution or that domestic con- financial markets at least from the 1999 repeal of provi- sumption alone can boost growth, even in large coun- sions in U.S. law that controlled speculation and sepa- tries. The longer-term benefits of openness more than rated investment banks from depository banks. The crisis compensate for the short-term negative impacts of trade manifested itself in developing countries through shocks shocks. The question is not whether to remain open but related to the balance of payments--a sudden stop in rather what kind of safety and insurance systems--at the capital flows; the collapse of export demand and prices; micro- and macrolevels--to put in place to better hedge and the disappearance of trade credit lines, repatriations, against shocks from globalization. and dividend remittances. As developing countries try to find answers to these Despite its financial origins, the crisis had enormous questions, they also face a drastically changed trade envi- effects on the real economy. As credit conditions tightened, ronment. The crisis proved that protectionism is no longer employment and consumer confidence fell, and U.S. eco- the name of the game; it remained largely under control, nomic activity slumped, while the voracious appetite of the 1 2 Managing Openness United States for imports started to flag. Global trade val- and financial innovations. As the world's largest debtor-- ues plummeted in the second half of 2008, before starting at 25 percent of gross domestic product (GDP) in to recover in 2009. Besides a sudden stop in capital flows, 2009--the United States must necessarily undergo a the crisis was "the first modern sudden stop in trade flows" depreciation of the dollar larger than the 30 percent drop (Eichengreen, chap. 2, this volume). since its 2002 peak. This depreciation will lead to an During the crisis and recovery period, there have been improved trade balance and a slowdown of capital renewed calls, particularly in the United States but also inflows into the United States. among the Group of 20 (G-20), for a comprehensive Others believe that imbalances are the result of struc- process of global rebalancing. Although global imbalances tural factors in other countries, which if unchanged will may not have caused the crisis, they certainly played a sup- cause the imbalances to persist mainly because of interna- porting role in its genesis. The prospect that some degree of tional asymmetries in the supply and demand for financial global rebalancing could take place in the aftermath of the assets (equilibrium approach). Underdeveloped financial crisis has important implications for developing countries, markets in emerging economies cause capital to flow to since it would imply a major realignment of global trade rich countries--driven by volatility, risk perception, and a flows. The question of whether global imbalances are sus- savings glut in developing countries. This process can con- tainable in the medium to long term is thus not only an tinue as long as financial markets in emerging economies important financial question, but also an emerging issue remain underdeveloped and U.S. assets remain readily for trade policy. available. On the demand side, the absence of social safety Only in the past few years have global imbalances nets encourages emerging markets to have high levels of become a bilateral story of the U.S. deficit and China's sur- private savings, thus constantly putting them in a net cred- plus (Servén and Nguyen, chap. 3, this volume). China's itor position. Moreover, imbalances can be maintained at surplus only recently displaced the surplus of countries like equilibrium if emerging markets hoard foreign assets as a Japan and the rest of Asia, and the surplus of the European policy choice in pursuit of export-led growth by depressing Union (EU) disappeared only around 2005 (figure 1.1). consumption or accumulating foreign assets as a precau- Some believe that global imbalances are unsustainable tion against external shocks such as a sudden stop of for- (disequilibrium approach). They argue that the United eign capital inflows. States has been recklessly increasing public and private Clearly, the sudden stop of capital inflows into the expenditure, prompted respectively by fiscal expansions United States implied by more extreme versions of the Figure 1.1. Current Account Imbalances for Selected Countries as a Percentage of World GDP, 1980­2009 3.0 2.5 2.0 1.5 1.0 percent 0.5 0.0 ­0.5 ­1.0 ­1.5 ­2.0 ­2.5 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 year China Asia (excluding China) European Union United States oil exporters other countries Source: Servén and Nguyen, chap. 3, this volume. Note: GDP = gross domestic product. Managing Openness: From Crisis to Export-led Growth, Version 2.0 3 disequilibrium approach did not actually happen. In fact, their accompanying positive externalities are both positive quite the reverse occurred. Capital fled to the safety of low- signs. Further changes that would facilitate adjustment could risk U.S. Treasuries during the crisis, with U.S. residents include the development of social safety nets, the liberalizing repatriating capital and with inflows from foreign investors of financial markets, and the reform of corporate governance picking up after 2008 in the large-scale acquisition of to encourage payment of dividends. These changes, however, Treasuries (figure 1.2). With the continued absence of would lead only to a gradual appreciation of the exchange mature financial markets in emerging economies and the rate and not a sudden adjustment of global imbalances. cushion provided by reserves in the immediate aftermath of the crisis, capital could plausibly continue to flow to The Impact of the Global Financial Crisis developed countries, and global imbalances could persist on Trade and Protectionism for some time yet in the postcrisis economy. Indeed, the cri- sis has highlighted the intrinsic instability of financial mar- The financial crisis led to the largest drop in global trade kets and may even encourage some emerging economies to volumes since World War II, with a simultaneous drop in go slow on financial development. global trade flows in the fourth quarter of 2008 for all One path toward reducing global imbalances is an major regions of the world. Real trade fell 12 percent in increased emphasis on domestic markets by developing 2009, with World Trade Organization (WTO) estimates of economies. This process is evident in China's stated inten- a year-on-year decline of 30 percent in the first quarter of tion to ease dependence on exports, for example. Amid the 2009 (figure 1.3). For some countries, the drop was enor- clamor for renminbi appreciation, Eichengreen (chap. 2, this mous: Japanese exports fell by 50 percent between February volume) argues that the current level of the renminbi 2008 and February 2009. However, while imports collapsed reflects, rather than determines, the current structural policy in most countries, four growing emerging markets recorded followed by China: depressing consumption, holding down positive import growth from 2007 to 2009: Brazil, China, the real exchange rate, and plowing savings into investment India, and Indonesia (Freund, chap. 4, this volume). This of tradable manufactures, with a view to encouraging growth outcome could herald the beginnings of a "new world trade of the manufacturing sector. It is therefore the surrounding order," with these countries showing resilient growth in policies that would have to change for the renminbi to appre- global import demand in the face of the crisis, despite a ciate. Nascent financial development in China--allowing depreciation of the exchange rate in the case of Brazil firms to issue bonds and borrow from banks rather than and Indonesia. rely solely on retained earnings--and the increasing devel- The extent of the drop in trade has puzzled econo- opment of supply chains and production networks with mists. They have attributed the magnitude of the collapse Figure 1.2. Gross Capital Inflows to the United States from Emerging Markets, 2000­09 1,200 1,000 800 US$, billions 600 400 200 0 00 01 02 03 04 05 06 07 08 09 20 20 20 20 20 20 20 20 20 20 year total inflows inflows from official sources Source: Servén and Nguyen, chap. 3, this volume. 4 Managing Openness Figure 1.3. Monthly Import Growth by Region, January 2008­November 2009 80 60 import growth month over same month, 40 previous year (%) 20 0 ­20 ­40 ­60 09 08 08 08 8 09 09 9 00 00 20 20 20 20 20 20 2 2 y ril ly er ry ly er ril ar Ju Ju Ap ua ob ob Ap nu n ct ct Ja Ja O O month East Asia and Pacific Europe and Central Asia Latin America and the Caribbean Middle East and North Africa North America South Asia Sub-Saharan Africa Source: Bown and Kee, chap. 6, this volume. to protectionist measures, disruptions to the supply of trade finance in foreign exchange (Eichengreen, chap. 2, trade credit, and the development of global supply chains, this volume). among other possible causes. Trade finance was certainly Global supply chain production has become increasingly disrupted when the crisis first started and credit markets prevalent in recent years. Households put off the purchase froze, but the joint response of official export credit agen- of "big-ticket" items such as consumer durables--vehicles cies, intergovernmental organizations, and governments and electronics, for example--when the economy is down. quickly limited its effect. Because trade credit is collateral- These are exactly the types of goods that now tend to be ized, it was possible to keep credit flowing. For longer-term produced through global supply chains. A significant drop policy, however, central banks and national export credit in demand for these goods thus contributed to the magni- agencies need to devise stronger crisis responses and move tude of the trade collapse. Furthermore, in practical terms, away from reliance on multilaterals. In addition, the need this form of production is more sensitive to disruptions in for trade finance highlights the importance of holding trade credit: if component exporters cannot get credit, then reserves: where parts, components, and other inputs going assemblers cannot get parts, and even a limited financial into the manufacture of exports are themselves imported, disruption can break all the links in the chain. Interna- central banks and export credit agencies have to provide tional production networks can be particularly sensitive to Managing Openness: From Crisis to Export-led Growth, Version 2.0 5 these types of changes in trade costs (Ma and Van Assche, protectionism gives some cause for concern in light of the chap. 21, this volume). increasing importance of that type of trade. As the financial crisis unfolded over 2008­09, some feared that countries would resort to "beggar-thy-neighbor" Should Countries Reassess Export trade policies similar to those in place during the Great Orientation? Depression. However, despite a significant drop in inter- national trade, "the world did not witness a `disintegra- Such a huge shock to world trade, combined with severe tion' of globalization in the sense of a policy-reinforced macroeconomic instability, made it natural for policy rush to disconnect from the global economy as seen in the makers and commentators to call into question some of 1930s" (Franco and Vieira, chap. 7, this volume). Protec- the basic assumptions underlying the system. Trade and tionism, while "murky"--Evenett (2010) identifies 300 economic integration policies are no exception. Outward- trade-restricting measures of some sort between the oriented or export-led growth strategies are being fourth quarter of 2008 and the fourth quarter of 2009-- reassessed for a variety of reasons. First, the trading system was neither as intense nor as broadly based as the tariffs quickly transmitted simultaneous negative shocks in major and quantitative restrictions seen in the 1930s. markets to the rest of the world. Second, global rebalancing The good news is that the protectionist measures that over the medium term might limit the capacity of newly come up are now mainly WTO-consistent temporary trade developing countries to rely on overseas demand in the barriers (TTBs): antidumping, countervailing duties, and same way that China has or that the Asian Tigers and Japan safeguard measures (Bown and Kee, chap. 6, this volume). did before it. Some claim that "the era of export-led growth The trade impacts of these measures have been extremely is over in its current form" (Klein and Cukier 2009, 11). limited, affecting on average 0.5 percent to 1.0 percent of A broad policy implication that has grown from a total imports, with overall increases in protectionism of reassessment of export-led growth is that developing coun- roughly the same order of magnitude. tries should focus more on building internal markets and This outcome is a positive affirmation of the global supplying domestic demand. They will thus be relatively multilateral trading system under the WTO, owing perhaps less exposed to negative demand shocks from abroad. to the design of WTO rules that allow for a relatively small Recent statements by China suggest that it may indeed be incidence of new protectionism through permissible TTBs. moving toward greater reliance on domestic consumption. Countries have committed to the rules of the global trad- In addition, Rodrik (2009) suggests that a particular ing system and the WTO, and they now have an interest in type of industrial policy should be used to promote devel- upholding them. The muted protectionism was also possi- opment of the domestic tradables sector, but without the bly a result of the increasing globalization of supply chains. currency devaluations that many countries have relied on Although the overall resort to protectionism has been in the past. Developing countries could then reap many of limited, the pattern of trade protection that has emerged the gains associated with export-led growth and not have has some important implications for South-South trade. to rely on large external surpluses. However, the efficacy of The biggest imposers of TTBs were not the large, devel- industrial policy is hotly debated, and the quantitative evi- oped markets, but Argentina, India, Indonesia, and Turkey. dence in favor of its effects on trade and growth is scarce China's exporters bore the brunt of the new protectionist and inconclusive (Harrison and Rodríguez-Clare 2009). measures. Eighty percent of the total value of trade on Moreover, the type of industrial policy envisaged by Rodrik which the G-20 imposed new TTBs in 2008­09 was on (2009) relies heavily on relatively efficient and transparent imports from China. The new TTBs in 2008­09 affected governance institutions, the lack of which is a major devel- more than US$20 billion worth of China's exports, or opment constraint in many countries. about 2 percent of its precrisis, 2007 volume. The emerging An important lesson to emerge from Eichengreen and pattern of protectionist trade policies thus has a strong from Haddad et al. in this volume (chaps. 2 and 11, respec- South-South flavor: South-North trade was largely unaf- tively) is that outward-oriented growth is indeed not a risk- fected in this way, with developing countries continuing to free strategy. In fact, openness brings a number of economic have access to rich-country markets. risks with it, most notably the potential for increased volatil- All in all, the multilateral trading system, with the WTO ity of the national economy due to its openness to shocks at its center, has weathered the storm relatively well. The abroad. This factor has important implications for the pol- global trade environment is likely to remain conducive to icy instruments needed to manage the risks that come with outward-oriented growth strategies over the medium outward orientation--policy instruments that should be an term, although the emerging pattern of South-South integral part of the export-led growth version 2.0 model. 6 Managing Openness Export diversification should remain a key policy objec- of barriers to market entry stand out. Moreover, they have tive for developing countries, because it can help manage the important advantage of keeping incentives neutral the risks that come with the benefits of openness. Coun- across firms within sectors and across sectors in the econ- tries with more diversified export bundles exhibit a weaker omy. Trade facilitation and lower entry barriers intensify link between openness and increased volatility of output. competitive pressures and speed export innovation. They To some extent, the process of market diversification is help create a situation in which winners can pick themselves. already well under way with the rise of South-South trade. The ongoing emergence of an export-led growth ver- Developing countries still have considerable scope, how- sion 2.0 model suggests that most countries have not lost ever, for intensifying the process, in particular for further sight of the important dynamic benefits that openness diversifying the range of products they export. Despite can bring. Examples include technology transfer, learning- recent gains, low-income countries remain relatively by-exporting, increased competitive pressure that reduces dependent on a concentrated bundle of export goods, markups and improves efficiency, and economies of many of which are resource dependent; that dependence, scale--particularly for small countries. Although the in turn, makes them more susceptible to international macrolevel "openness and growth" literature of the 1990s shocks in those markets (Hanson, chap. 8, this volume). ground to a stalemate of result and counterresult in the A threshold effect influences the relationship between early 2000s (Rodriguez and Rodrik 2001), the new firm- diversification and the trade-off between openness and level evidence provides consistent and highly convincing volatility (Haddad et al., chap. 11, this volume). However, evidence that more open sectors experience faster produc- because most countries are already at a sufficiently high tivity growth (Bernard et al. 2007). level of export diversification, an increase in openness is Outward-oriented growth strategies have brought not likely to be associated with increased economic volatil- considerable economic benefits to a range of countries ity. Many other countries are heading in the right direction over recent decades. The examples of Malaysia (Haddad, (figure 1.4). Of course, countries with highly concentrated Winkler, and Zeufack, chap. 12, this volume) and even export bundles--such as some commodity exporters--are Africa (Songwe and Winkler, chap. 19, this volume) con- not yet in that situation, and measures to promote diversi- firm that those benefits remain important for developing fication could be beneficial. countries, because exports tend to increase firm productiv- Among the available policy instruments for promot- ity, which is the engine of long-term economic growth. ing export diversification, trade facilitation and lowering These mechanisms are more important than ever following the trade and financial crises. As economies around the world emerge from the crisis, they will naturally look to engagement with world markets as a source of growth and Figure 1.4. Diversification Trends among Selected Countries, 1981­2005 development. This impetus behind outward-oriented growth is therefore likely to remain firmly in place. 0.9 Besides the role of export diversification in helping a country mitigate the impact of external and internal 0.8 export concentration shocks, sound macroeconomic fundamentals are vital 0.7 too--especially for smaller and more open economies that 0.6 are subject to greater income volatility risk. Such measures 0.5 include holding high reserves, especially to fill the trade- threshold credit gap, and establishing and funding specialized 0.4 credit agencies for trade finance. Reducing the budget 0.3 deficit and maintaining policies for greater regulation of 0.2 financial markets, particularly for dealing with the capi- tal inflows of the carry trade, are also important. Those 5 0 5 0 5 ­8 ­9 ­9 00 ­0 81 86 91 00 economies that rely heavily on durable manufactures ­2 19 19 19 96 20 19 produced through global supply chains should focus year especially on export diversification. Colombia South Africa Kenya A number of countries were able to move proactively to support their economies through the crisis, thanks to fore- Nicaragua Jordan Mexico sight and prudent macroeconomic management. Chile is Source: Haddad et al., chap. 11, this volume. an important example (Zahler, chap. 15, this volume). Managing Openness: From Crisis to Export-led Growth, Version 2.0 7 Despite having a more open and outward-oriented econ- specialized in resource-intensive exports, but they are tak- omy than many other countries in Latin America, Chile ing small steps up the value chain through expansion recovered quickly from the crisis, helped along by healthy in apparel production. Although the export baskets of international copper prices. Its "sound fundamentals and middle-income countries are more diversified than those strong macroeconomic management provided a buffer of low-income countries, low-income countries are steadily against the global economic recession" (Zahler, chap. 15, closing the gap. this volume). Prudent fiscal policy during times of strong How sustainable is this new trade pattern? The recent growth--with a fiscal surplus reaching nearly 9 percent of diversification is based primarily on strong fundamen- GDP in 2007 compared with an average fiscal deficit of 0.2 tals, namely, economic growth in importing countries percent of GDP in other Latin American countries--gave and greater competitiveness in exporting countries. If the government room for significant macroeconomic middle-income countries continue along the recent stimulus in the face of the recent crisis. upward trend, import demand will continue to grow in Another important feature of the crisis is that although the next decade. Developing countries experienced strong the major developing countries were hit hard by the trade improvements in relative competitiveness in most sectors. effects of the crisis, they have rebounded strongly. Haddad And this phenomenon is not limited to "acronym" coun- et al. (chap. 11) highlight the fact that the medium-term tries such as the BRICs and GEMs (growing emerging benefits of outward-oriented policies outweigh the short- markets, including Brazil, China, India, and Indonesia): term costs associated with the negative impact of the shocks other middle-income nations have enjoyed improvements from the crisis. This point is true not only for major devel- in export competitiveness similar to those the BRICs oping countries, but also for many small, open, developing experienced. The role of South-South trade is therefore countries, thanks in part to the growing strength of trade likely to be greatly enhanced in the export-led growth version links between China and the South (Wang and Whalley, 2.0 model. chap. 10, this volume). On the one hand, this dynamic Among potential developing-country import markets, underscores the importance of promoting South-South China is obviously a major player. Wang and Whalley trade in the future. On the other hand, it also shows the (chap. 10, this volume) examine China's South-South stabilizing effect of the greater integration of a number of trade relations before and after the crisis. By 2007, China's developing countries with large internal markets-- exports to and imports from the South accounted for primarily Brazil, the Russian Federation, India, and China 21 percent and 35 percent of total trade, respectively. India (BRICs)--into the world economy. As development contin- and Brazil are the most rapidly growing trade partners, hav- ues in those countries and purchasing power increases, their ing experienced thirty-two- and fifteenfold increases, role as demand stabilizers will also become more important. respectively, in their bilateral trade with China since 1995. Latin America and Africa are the largest regional sources of trade growth, with more dispersion across countries in Asia. Changing Dynamics in Global Trade In contrast to its total trade, China runs a significant trade In the future, the likely drivers of export growth and deficit with the South because of its imports of resource diversification in developing countries are the develop- products. Assuming unchanged growth rates, by 2015 ing countries themselves; these countries will play a vital Southern trade will account for over 50 percent of China's role through South-South trade (Hanson, chaps. 8 and 9, total, and by 2025, India will account for over 50 percent of this volume). GDP growth in low-to-middle-income China's trade. Finally, China's trade with the South was rel- developing countries has led to increased import demand, atively less affected than average by the great trade collapse, which has played an important role in boosting exports in part because of a successful stimulus package. and promoting export diversification in low-income The overall improvements in South-South trade notwith- countries. Low- and middle-income countries accounted standing, the temporary trade barriers imposed during the for 31 percent of world imports in 2008. crisis were South-South in nature (figure 1.5). Of the stock Concerns that weak consumption growth in the United of TTBs that developing-country users had in place in 2009, States might undermine the viability of export-led devel- 68 percent were imposed on imports from other developing opment may be exaggerated. Dependence on U.S. demand economies. Because TTBs fell less heavily on developing- for imports has generally lessened, although the United country exports trying to reach high-income markets, it States is still a key source of demand for exports from the is clear that improving the security of access for developing Western Hemisphere, for oil products, and for apparel countries to Southern markets should remain a priority. from low-income countries. Low-income countries remain Enhanced South-South regional engagement--through 8 Managing Openness Figure 1.5. Combined G-20 Use of Selected Temporary Trade Barriers by Import Source, 1997­2009 3,500 unique HS06 product-exporter 3,000 2,500 combinations 2,000 1,500 1,000 500 0 97 98 99 00 01 02 03 04 05 06 07 08 09 19 19 19 20 20 20 20 20 20 20 20 20 20 year China as exporter developing economy as exporter (excluding China) developed economy as exporter Source: Bown and Kee, chap. 6, this volume. Note: HS06 refers to the six-digit level of the Harmonized System. WTO-consistent agreements--would be one step in this (Mitchell and Aldaz-Carroll, chap. 20, this volume). First, direction. Another possible line of action would be to commodities have had a relatively less severe crisis: thanks ensure better access to the WTO dispute settlement system to strong demand from developing countries--especially for Southern countries, so that they can continue to open China--the drop in commodities trade was less than in desired markets in other developing economies. other sectors, and the rebound has been stronger. Second, commodity markets are now more closely than ever linked to energy markets because of the increasingly important The New Trade Policy Agenda role played by biofuels. As a result, price prospects are The emerging postcrisis environment poses a number of strong in oil and agriculture markets alike (figure 1.6). Of challenges and presents important opportunities for devel- course, all is not unmitigated good news for commodity oping countries. Commodities trade is booming, largely to exporters: stronger governance and efficient regulation will the benefit of developing countries. Demand for tempo- be necessary to manage commodity wealth and to avoid rary workers is also rising because of shortages in labor some of the negative impacts that have accompanied previ- markets in developed countries and increasingly in middle- ous instances of high prices, such as the "Dutch disease." income countries. The nature of production and trade is A much newer issue on the trade policy agenda relates changing, with much of it taking place through produc- to the increasing importance of global production net- tion-sharing networks. This evolution opens new opportu- works (GPNs). Their prevalence may have been one vector nities for countries that can latch onto these networks, but by which the global trade collapse was propagated, as final it will leave out less competitive countries. Finally, we can demand shocks were transferred to suppliers of intermediate no longer ignore the impact of production and trade on inputs. Evidence based on Chinese processing trade shows the environment. that intranetwork trade suffered a disproportionately large While trade policy specialists are no strangers to com- fall during the crisis: it appears to be more sensitive to cycli- modity markets, a number of new dynamics will prove cal fluctuations than regular trade (Ma and Van Assche, particularly important in the postcrisis environment chap. 21, this volume). Evidence of a "bullwhip" effect is Managing Openness: From Crisis to Export-led Growth, Version 2.0 9 Figure 1.6. Real Commodity Price Indexes, History, and Projection, 1960­2020 300 forecast 250 constant 2000 US$ 200 150 100 50 0 19 0 19 2 19 4 19 6 68 19 0 19 2 19 4 76 19 8 80 19 2 19 4 19 6 19 8 19 0 19 2 19 4 19 6 20 8 20 0 20 2 20 4 20 6 20 8 20 0 20 2 20 4 20 6 20 8 20 6 6 6 6 7 7 7 7 8 8 8 8 9 9 9 9 9 0 0 0 0 0 1 1 1 1 1 19 19 19 19 agriculture crude oil metals and minerals Source: Mitchell and Aldaz-Carroll, chap. 20, this volume. also clear: the drop in Northern demand for Chinese goods Finally, perhaps the most crucial part of the "new" trade caused local manufacturers to run down inventories of agenda is the emerging links between trade policy and the imported intermediates, which magnified the overall trade environment, particularly climate change. The December drop. Following the crisis, GPNs are being geographically 2009 Copenhagen meeting marked a fundamental change consolidated (Milberg and Winkler 2010). With the crisis in direction, which potentially puts the global climate over, the challenges are rising for developing countries change and trade regimes on a collision course (de Melo seeking to enter or upgrade within a GPN. The consolida- and Mathys, chap. 23, this volume). Alternative architec- tion has contributed to fiercer competition and allowed tures for the next climate change agreement lend them- only those with specific capabilities to remain in the game. selves to four areas in which trade will play a role: as a pur- Business as usual will not work: developing countries need veyor of technological transfer; as a mechanism for to design new strategies to benefit from opportunities from separating where abatement takes place from who bears the growing presence, and importance, of GPNs (Cattaneo, the costs of abatement; as a participation mechanism; and Gereffi, and Staritz 2010). as a way to address the pressures for border adjustments. Shifting from trade in goods to movements of people Perhaps an umbrella agreement with leeway--where much through international migration, we see more opportuni- initial mitigation would take place unilaterally, as under ties for developing countries (Martin, chap. 22, this vol- the early days of the General Agreement on Tariffs and ume). The economic gains from freeing up migration Trade--might be the most promising way ahead for pre- could potentially be larger than those stemming from a fea- serving an open world trading system. sible Doha Round of trade liberalization. The demand for migrant labor is expected to continue to expand in the Conclusion coming decades: the decline in labor migration associated with the global financial crisis is likely a temporary hiccup According to the work brought together in this volume, in an otherwise upward trend. Most of this demand for we argue that reports of the death of export-led growth migrants will be in the industrial countries, but some of are greatly exaggerated. Although the nature of outward- the sharpest increases in the demand for migrant labor are oriented growth strategies is likely to change over the expected to be in middle-income developing countries medium term to give greater prominence to South-South such as Brazil, Costa Rica, Libya, Malaysia, South Africa, trade, the fundamental mechanisms are likely to remain and Thailand. This trend parallels the role of South-South in place. Similarly, there is a good chance that some relations in the trade area and raises the question of the degree of global imbalance will persist over the medium capacity of nontraditional destinations to absorb new term, which means that any realignment of trade flows migrant flows. that takes place is unlikely to be so sharp that it becomes 10 Managing Openness impossible for developing countries to continue leverag- broad-based liberalization commitments. Indeed, given the ing international markets as part of their growth strat- rise of production networks, particularly in Asia, regional egy. It is significant that from the broad range of country agreements are likely to move more and more in this direc- experiences discussed in part V of this volume, the over- tion, thereby facilitating the emergence of cross-border all impression that emerges is one of a continuing policy- supply chains. level commitment to outward-oriented growth. There is If the global economic crises highlight anything about very little evidence that major economies are seriously policies of outward-oriented growth, it is that they are obvi- reconsidering their degree of engagement with the global ously not a panacea for development. No set of policies is. trading system. Nevertheless, we should continue to Rather, outward-oriented growth brings both opportunities monitor trade policies for signs of retreat from this posi- and risks for developing countries. The key is in finding the tion should economic hard times continue longer than appropriate set of policies to manage openness, one that expected. The weight of evidence suggests, however, that maximizes its benefits and minimizes its costs. The evidence the trade and trade policy implications of the global presented in this volume suggests that many developing crises should not be overstated. countries are now engaged in that process and will continue Going forward, policy makers can do much to ensure along that course for the foreseeable future. that the global trading environment remains as receptive as possible to outward-oriented growth strategies. On the one Bibliography hand, concluding the WTO's Doha Round would provide Baldwin, R. E., ed. 2009. The Great Trade Collapse: Causes, Consequences, business with additional certainty about the future shape and Prospects. London: Centre for Economic Policy Research. of trade policy and would encourage investment in export- Bernard, A. B., J. B. Jensen, S. J. Redding, and P. K. Schott. 2007. "Firms in oriented industries. One important benefit the Round International Trade." Journal of Economic Perspectives 21 (3): 105­30. could deliver would be to establish binding trade policy Cattaneo, O., G. Gereffi, and C. Staritz. 2010. Global Value Chains in a Postcrisis World: A Development Perspective. Washington, DC: World ceilings that are closer to the current levels of applied tariffs Bank. (Messerlin, forthcoming). The crisis experience shows that Evenett, S. J., ed. 2010. Tensions Contained . . . for Now: The 8th GTA countries are reluctant to increase tariffs other than at the Report. London: Centre for Economic Policy Research. Harrison, A., and A. Rodríguez-Clare. 2009. "Trade, Foreign Investment, margin, for good economic reasons. They can enhance cer- and Industrial Policy for Developing Countries." Working Paper tainty and promote trade over the longer term by locking 15261, National Bureau of Economic Research, Cambridge, MA. in that position through tariff bindings. Klein, B. P., and K. N. Cukier. 2009. "Tamed Tigers, Distressed Dragon." Foreign Affairs 88 (4): 8­16. Policy makers will also need to address the increasing Messerlin, P. A. Forthcoming. "The Doha Round." In Handbook of Trade importance of South-South trade. They should pay greater Policy for Development, ed. A. Lukauskas, R. M. Stern, and G. Zanini. attention to the reduction of South-South trade barriers, Oxford: Oxford University Press. particularly nontariff barriers that can hold back the Milberg, H., and Winkler, D. 2010. "Trade, Crisis, and Recovery: Restruc- turing Global Value Chains." In Global Value Chains in a Postcrisis emergence of the "export-led growth version 2.0" model World: A Development Perspective, ed. O. Cattaneo, G. Gereffi, and we referred to earlier. Enhancing the engagement of some C. Staritz, 23­72. Washington, DC: World Bank. developing countries in the Doha Round could be one Rodriguez, F., and D. Rodrik. 2001. "Trade Policy and Economic Growth: A Skeptic's Guide to the Cross-National Evidence." NBER Macroeco- useful way of moving forward in this area. Another possi- nomics Annual 15: 261­338. bility is for the emerging body of South-South regional Rodrik, D. 2009. "Growth after the Crisis." Working Paper, Commission agreements to focus on the implementation of deep and on Growth and Development, Washington, DC. Part I II The Crisis, Global Imbalances, and Rebalancing 2 Managing Openness: Lessons from the Crisis for Emerging Markets Barry Eichengreen Another discussion of the global financial crisis requires positive and negative, already points to hypotheses. More some justification.1 The justification for this one is that the open economies were hit harder. Countries with large cur- lessons of the crisis for emerging markets and their man- rent account deficits were hit harder (see figures 2.1 and agement of openness are still not adequately understood. 2.2.) Countries that had restrained the rate of growth of Important questions remain unanswered. This chapter credit and had more flexible exchange rates did better. In focuses on three. contrast, the role of the government budget deficit is not First, who was hit, and why? And, in a related question, clear; it is not obvious that countries with larger fiscal sur- what policies should emerging markets follow to minimize pluses did better, in other words (see figure 2.3).2 the effects of global volatility? While more than a little has The question is whether these and other regularities stand been written on this subject, it is not clear that a consensus up to scrutiny when analyzed using data for a larger sample of has yet formed. emerging markets. Rose and Spiegel (2009) are skeptical. Second, what explains the outsize response of trade that They link the severity of the growth decline, along with some was one of the principal transmission belts for the crisis? ancillary measures of financial distress, to a set of indicator This crisis may have been just another "sudden stop" of cap- variables in 2006, the eve of the crisis, and find few robust ital flows, not unlike the sudden stops of the past, but it was regularities. One interpretation of this finding is as confirm- the first modern sudden stop in trade flows--something ing the weak predictive power of so-called early-warning that deserves further analysis. indicators, something to which some have pointed previ- Third, and finally, what was the role of global imbal- ously.3 Crises differ. Market behavior and policy responses ances in the crisis? The answer to this last question again change, not least in response to the development of early- has implications for what kind of policy adjustments warning indicators themselves. In this view, there is no emerging markets should make going forward. telling when a country will be hit or how hard. The appro- priate policy response is therefore to invest in insurance. This view, for example, has some appeal to those who live on Who Was Hit, and Why? active earthquake faults and have learned to keep flashlights The impact of the crisis varied enormously. Comparing and bottled water on hand. demeaned real gross domestic product (GDP) growth in Another interpretation is that it is not so much crises the third quarter of 2008 to the first quarter of 2009 at sea- that differ as countries. The impact of the same shocks and sonally adjusted annual rates, growth fell by an astounding policies may be different in low-income, low-to-middle- 35 percentage points in Latvia, 30 in Lithuania, and 25 in income, and middle-income countries, given differences in Estonia, compared to less than 5 percentage points in market structure and development. Mody (2010) finds, for Argentina, India, and Poland. This handful of outliers, both example, that the positive correlation that one might 13 14 Managing Openness Figure 2.1. Ratio of Current Account to GDP, 2007, versus Change in Growth, 2007­09, All Emerging/Developing Countries 10 COG MWI MAR GNB IRQ CIV LBN NPL COM TGO YEM MLI BDI LAO KSV ZMB SYR HTI BFA 0 MOZ KIRBGD CMR LSO NGA DZA MRT GMB BEN CAFUGA GHA TZA SLE TCDGIN ETHRWA ECU UZB IDN BLZ EGYPNG SUR BOL SEN CPV PAKTUN SWZ VNM ALB GUY LKA COD IND SAU JAM NIC TJK MUS NER OMN CHN DOM URY KEN POL JOR SDN GTM PHL NAM IRN CHL change in growth (%) MKD KGZ BRA GAB HUN COL MMR ZAF KAZ ARG THA TKM HND SLV BLR PER MYS MDA CRI TUR SRB BIH PAN MEX ­10 MDG PRY BWA HRV MNGVEN KHM ROM BTN SVK GEO RUS GNQ ­20 AGO EST UKR LTU LVA ARM ­30 ­20 ­10 0 10 20 ratio of current account to GDP Source: World Economic Outlook (database), International Monetary Fund (IMF), http://www.imf.org/external/pubs/ft/weo/2010/01/weodata/index.aspx. expect between large current account deficits and the fall in find no evidence that countries with more reserves had output (as countries with large deficits found them better crises. Blanchard, Faruqee, and Das (2010) report increasingly difficult to finance) is evident only in lower- the same negative conclusion: when they include both middle-income countries (the middle tier of developing reserves and short-term liabilities as shares of GDP, the lat- countries), not in upper-middle or low-income economies. ter matters but the former does not. Others like Obstfeld, Berkmen et al. (2009) find that the financial channel was Shambaugh, and Taylor (2009), in contrast, find a link more important than the trade channel for emerging mar- between reserves and financial stability. Policy makers like kets (defined as developing countries with reasonably open Brazilian central bank governor Henrique Meirelles have capital markets) but that the trade channel was more similarly argued that they played an important stabilizing important for a broader sample of developing countries role in the crisis (see MercoPress 2010). The postcrisis (trade mattered more for the financially less connected behavior of emerging markets, which has been to accumu- low-income countries). late more, is certainly consistent with this view. Alternatively, the difficulty of identifying sources of vul- The obvious reconciliation is that of Moghadam (2010). nerability may reflect neither that crises differ nor that Reserves play a stabilizing role but only to a point. In a liq- countries differ but rather that the link between a country's uidity crisis in which investors are deleveraging, foreign characteristics and its susceptibility to disturbances is non- borrowings must be repaid, and the scarcity of foreign linear. An example is the role of reserve accumulation in exchange puts severe downward pressure on the local cur- providing insulation from shocks. Berkmen et al. (2009) rency. Having the reserves needed to repay most or all of Managing Openness: Lessons from the Crisis for Emerging Markets 15 Figure 2.2. Exports as a Percentage of GDP, 2007, versus Change in Growth, 2008­09, All Emerging/Developing Countries 5 HTI COG CIVBTN GUY KEN TGO ZMB COM LBN PAK 0 CAF MLI LAO NGA GNB N CHN DZA NPL BGDSEN IR MARMOZ CMR UZB TCD VNM ETH SLE IDN SYR KSV INDMWI GMB UGACPV DOM TUN SWZ SDN LKA BEN GTM COL KAZ PNG EGY BOL PHL POL LSO MUS GHA CRI TJK NIC NAM GAB SAU MRT THA BRA TUR ZAFALB GIN CHL MKD BLZ JOR GNQ change in growth (%) ­5 ARGURY HND SLVGEO KGZ TKM MYS RWA ECU HUN MEX VEN HRV PAN SRB PER BIH BWA OMN KHM BLR ­10 PRY MNG EST SVK MDG RUS LVA AGO ROM MDA ­15 UKR LTU ­20 ARM 0 50 100 ratio of exports to GDP Source: World Economic Outlook (database), IMF, http://www.imf.org/external/pubs/ft/weo/2010/01/weodata/index.aspx; World Development Indicators (database), World Bank, http://data.worldbank.org. those short-term foreign obligations, provide banks and continue to yield significant stability benefits beyond that firms with scarce foreign exchange, and support the exchange point. It can be the stock rather than simply the maturing rate is of considerable value--as anyone who was in the portion of the foreign debt that matters if investors, in a Republic of Korea in November 2008 would know. At some panic, scramble to sell it off. It may be M2 that matters if point, however, perhaps when reserves match the value of the liabilities of the banking system are in foreign currency short-term obligations coming due, the marginal benefit of or the country is committed to pegging the exchange rate. having more reserves begins to diminish. Whether, beyond The exchange rate is another variable that appears to that point, they do anything to enhance stability further is bear a nonlinear relationship to the impact of the crisis. questionable. In other words, the relationship between The evidence is quite strong that countries pegging their reserves and stability is nonlinear. The least squares parabola currencies had worse crises, other things equal. Flexibil- in figure 2.4 is consistent with this view.4 ity helps when confronted by an unprecedented shock. This much is intuitive. The problem is that there is less Berkmen et al. (2009) and Blanchard, Faruqee, and Das than full agreement on the point at which diminishing (2010) both report that countries with pegged rates suf- returns set in. Moghadam's data suggest that this happens fered deeper output collapses even after controlling for a around the point at which reserves match a country's range of other economic and financial variables.5 But external financing requirement (the sum of the current both also suggest that more flexibility was not always account deficit, short-term debt, and medium- and long- better. What significantly enhanced stability was moving term amortizations of the public and private sectors) (see from a peg to a managed float, not moving from man- figure 2.5.) Others like Wyplosz (2007) argue that reserves aged flexibility to a free float. 16 Managing Openness Figure 2.3. Government Budget as a Percentage of GDP, 2007, versus Change in Growth, 2008­09, All Emerging/ Developing Countries 5 HTI COG GUY MMR CIV KEN TGO ZMB LBN YEM COM PAKMOZ NGA LAO MAR N IRMLI CAF DZA 0 VNM SENBGD CHN BDI ETH SLE GMB UZB CMR SYR IDN TCD MWI UGA JAM IND DOM PNG BFA TZA TUN CPV SWZ KAZ EGY SDN LKA MUS GTM BENCOL PHL BOL LSO POL GHA CRI NAM GAB TJK NIC SAU JOR ALBTHA GIN BLZ MRT BRA TUR URY MKDZAF CHL IRQ GNQ ­5 change in growth (%) HND SLVARG KGZ GEOMYS TKM HUN RWA ECU HRV MEX VEN PAN SRB BIH KHM OMN PER BWA NER BLR ­10 PRY EST SVK MDG LVA RUS AGO ROM MDA ­15 UKR LTU ­20 ARM ­10 0 10 20 government budget as percentage of GDP Source: World Economic Outlook (database), IMF, http://www.imf.org/external/pubs/ft/weo/2010/01/weodata/index.aspx; Economist Intelligence Unit (database), Economist, http://www.eiu.com. Perhaps the most striking correlation is that countries than countries with some restrictions, shown in the mid- with larger, better-developed, and more open financial sys- dle. (Note that the index of restrictions on inflows and tems did worse in the crisis.6 This observation again makes outflows is from Schindler 2009, where a higher value sense intuitively. The shock originated in financial markets. means more restrictive.) Financial links were an important transmission belt. Hence, The question is what to do about it. If the crisis just past is countries with relatively large financial systems and whose a once in a 100-year firestorm, then the correct answer, pre- markets were open to foreign investors felt the crisis first sumably, is nothing. But if it was a salutary if expensive and most acutely. The Republic of Korea, for example, suf- reminder of the intrinsic instability of financial markets, then fered because half its stock market capitalization was in the the lesson must be "go slow on financial liberalization and hands of foreign investors, who held a fire sale in response opening." The Indian approach of going slow on domestic to their own financial distress. Countries with better- deregulation and opening is the right one. The Brazilian developed financial systems had tended to have more approach of using taxes to discourage short-term foreign short-term external debt, which made for a more serious capital is the right one. The implication, like it or not (and crisis (figure 2.5).7 Figure 2.6 shows that countries with some in the World Bank will not like it), is that financial devel- repressed financial systems had their own problems in the opment in emerging markets will be slower than otherwise. crisis, but it confirms that countries with highly open The problem, again, is that the information needed to financial systems, shown at the right of graph, did worse know how far to go in this direction is not available. There Managing Openness: Lessons from the Crisis for Emerging Markets 17 Figure 2.4. International Reserves as a Percentage of Short-term External Debt, 2007, versus Change in Growth, 2008­09, All Emerging/Developing Countries 5 COG CIVGUY TGO ZMB KEN YEM LBN PAK COMMAR CAF VNM 0 GNB BGD CHN NGA BDI MOZSEN CMR ETH TCD DOM ID TUNN JAM COL BFA GMB IND SLE SDN TZA KAZ LKA PHL GTM SW EGY Z PNG BOL POL CRI change in growth (%) NI C GAB MRT THA BLZ ­5 MKD URY JOR ALB CHL BRA TUR ZAF ARG HND SLV GEO KGZ ECU MYS RWA VENMEX SRB BIH PER KHM BLR PRY NER ­10 MDG LVA RUS AGO MDA ROM ­15 UKR LTU ­20 ARM 0 10 20 30 40 international reserves/short-term external debt Source: World Economic Outlook (database), IMF, http://www.imf.org/external/pubs/ft/weo/2010/01/weodata/index.aspx; World Development Indicators (database), World Bank, http://data.worldbank.org. is ample and convincing evidence that financial develop- Asian crisis; Tong and Wei (2009) and World Bank (2010a) ment and openness have a positive impact on growth and show that it continues to hold. Markets that were permis- inclusiveness.8 The evidence is ample and convincing, that sively regulated, resulting in the strongly procyclical behav- is, when one draws it from good times. But it is equally ior of credit, and not simply markets were the problem.10 clear that financial development and openness expose coun- tries to additional problems in bad times, when financial Why Was the Collapse of Trade So Dramatic? markets fail. Two studies establishing that point are Vlachos and Waldenstrom (2005) and Eichengreen, Gullapalli, and The magnitude of the collapse of trade is a second important Panizza (2009). The problem is that there are no good esti- mystery to be unraveled before moving to policy recommen- mates with which to balance the marginal benefits of the dations. The suspects are clear: protectionist measures, dis- first effect against the marginal costs of the second. Again, ruptions to the supply of trade credit, and the development the relationship may be nonlinear: the early stages of finan- of global supply chains. It is unclear, though, how much cial development and integration may have significant net weight to attach to them. benefits, but those benefits diminish subsequently. Protectionism was likely averted largely as a result of It is, of course, possible to give more nuanced advice. lessons learned from historical experience. Comparisons of That is, rather than slowing financial development, perhaps the Great Recession with the Great Depression, which were developing countries should slow only certain forms of rife in 2008­09, pointed to the importance of avoiding the financial development. Countries whose banks funded kind of protectionism that compounded the earlier slump. themselves on the wholesale market, especially abroad, were World Trade Organization disciplines helped, as did coop- vulnerable when liquidity evaporated. Those in which the eration among the Group of 20 and monitoring of coun- ratio of deposits to domestic private sector loans was high tries' compliance by organizations like the World Bank and did relatively well.9 Highly leveraged as opposed to highly Global Trade Alert. But there was still a good deal of murky developed financial markets can be especially dangerous, in protectionism. Evenett (2010) identifies more than 300 other words. Countries in which a relatively high share of trade-restricting measures of one sort or another between foreign capital inflows were in the form of portfolio capital the fourth quarter of 2008 and the fourth quarter of 2009. (short-term portfolio flows in particular) did poorly. This is Eichengreen and Irwin (2009) exploited the Great an old finding from statistical post mortems on the 1997 Depression parallel to suggest where the danger was greatest. 18 Managing Openness Figure 2.5. Short-term External Debt as a Percentage of GDP, 2007, versus Change in Growth, 2008­09, All Emerging/ Developing Countries 5 HTI COG BTN CIV GUY ZMB KEN TGO YEM COM PAK LBN LAONGA MOZ DZACAF MAR MLISENIR N GNB 0 NPL CHN BGD CMR VNM UZB BDI TCD ETH MWIGMBIDN TUN SLE UGA Z DOM JAM CPV INDTZA KAZ SW BFA PNG SDN BEN LKA COL EGY GTM MUS PHL BOL LSO POL GAB GHA CRI TJK NIC BLZ GIN CHL MRT THA BRAJOR ALB ­5 URYTUR ZAF MKD change in growth (%) KGZSLVHND ARG TKMGEO MYS RW A ECU MEX VEN PAN SRB BWA PER BIH KHM NER BLR ­10 MNG PRY MDG AGO RUS LVA ROM MDA ­15 UKR LTU ­20 ARM 0 20 40 60 short-term external debt/GDP Source: World Economic Outlook (database), IMF, http://www.imf.org/external/pubs/ft/weo/2010/01/weodata/index.aspx; World Development Indicators (database), World Bank, http://data.worldbank.org. In the 1930s, recovery policy meant monetary policy. To transmitted the contraction internationally. Protectionism promote recovery, countries abandoned defense of their was a byproduct of their failure to act and, more generally, exchange-rate pegs, cut interest rates, and allowed their of the inadequate coordination of stimulus policies. currencies to decline. Unlike during the recent crisis, This time, recovery policy meant not only sharp reduc- reductions in interest rates in the 1930s were not accompa- tions in interest rates, often to zero, but also aggressive nied by aggressive quantitative easing.11 Other countries quantitative easing and fiscal stimulus. With quantitative felt the effects through two channels. To the extent that easing, the locomotive effect as opposed to the beggar-thy- they saw their currencies appreciate as a result of their neighbor effect of expansionary monetary policy was neighbors' policies, their competitiveness worsened and stronger. The cross-border spillovers of expansionary fis- their problems deepened. They lost reserves, and, to main- cal policy were positive as well.13 Whereas in the Depres- tain their pegs to gold, their central banks were forced to sion, it had been the passive countries--those that did not tighten. But to the extent that their neighbors began to take a policy response to the crisis--that had the strongest recover and, as a result, consumed more foreign as well as incentive to protect, this time it was the active countries domestic goods, they also felt a positive locomotive effect. that saw other countries as free riding on their efforts. This The evidence for the 1930s is that the first channel domi- explanation, in a nutshell, is the genesis of "buy America" nated: depreciation was beggar thy neighbor.12 Countries policies: some U.S. policy makers saw an expensive but that felt themselves beggared responded with restrictive necessary US$787 billion fiscal stimulus as also benefiting trade policies that distorted their economies and further other countries insofar as the associated spending fell on Managing Openness: Lessons from the Crisis for Emerging Markets 19 Figure 2.6. Average Financial Credit Restrictions, 2005, versus Change in Growth, 2008­09, All Emerging/Developing Countries 5 CIV ZMB KEN TGO YEM LBN PAK 0 MAR BGD CHN UZB UGA IDN IND TUN DOM JAM BFA KAZ SWZ TZA EGY GTM MUS LKA BOL PHL CRI GHA NIC SAU THA CHL BRA TUR ZAF change in growth (%) ­5 URY ARG SLV KGZ GEO MYS ECU HUN VEN PAN MEX OMN PER ­10 PRY LVA RUS AGO ROM MDA ­15 ­20 0 0.2 0.4 0.6 0.8 1 average financial credit restrictions Source: World Economic Outlook (database), IMF, http://www.imf.org/external/pubs/ft/weo/2010/01/weodata/index.aspx; Schindler 2009. imports as well as U.S. goods, and unfairly so insofar as It is plausible that disruptions to the supply of trade other countries did not respond with stimulus programs finance might have been important in the collapse of trade. of their own. The evidence of protectionism in the past Trade, by virtue of its time-intensive nature, depends on three years is broadly consistent with this pattern. finance, and this was, after all, a financial crisis. Iacovone To be sure, countries were much more successful now and Zavacka (2009) show that the exports of firms more than 80 years ago in coordinating their policy responses dependent on external finance fall by more in banking to the crisis, which appears to be more evidence of their crises than those of firms that self-finance and that have having learned from history. That increased coordination more tangible assets and hence better collateral. limited complaints about free riding and contained the On the other hand, Mora and Powers (2009) argue that protectionist impulse. Kee, Neagu, and Nicita (2010) con- this effect was quantitatively small because the disruption to clude that only 2 percent of the decline in world trade in flows of trade credit was limited in duration and extent. 2008 was attributable to increased protectionism. That Although other credit markets froze up, trade finance number may be an underestimate, however; unlike other declined to only a limited extent, a few exceptional cases authors (Evenett 2010, for example) who consider trade notwithstanding. Because trade credit is collateralized, it was restrictions broadly defined, Kee, Neagu, and Nicita (2010) possible to keep credit flowing. Official export credit agen- look only at tariffs. A more encompassing measure would cies, for their part, stepped in to help. Developing countries, yield a somewhat higher number. Still, the conclusion that then, need to put their central banks and national export trade policy was not a major factor in the collapse of trade credit agencies in a position where they can also help, rather would probably still stand. than relying just on the multilaterals. To the extent that 20 Managing Openness parts, components, and other inputs going into the manu- One wonders also about the interaction of production facture of exports are themselves imported, the central fragmentation with the two earlier explanations for the banks and export credit agencies in question have to pro- collapse of trade. It could be that the articulation of supply vide trade finance in foreign exchange--another reason, chains renders trade more sensitive to disruptions in the above and beyond those discussed earlier, to hold reserves. provision of trade credit. If component exporters cannot These last observations lead to a word on the role of trade get credit, then assemblers cannot get parts, and even a in parts and components. This trend is relatively new in limited financial disruption can break all the links in the developing countries but one in which Asian countries in chain.16 This explanation is the trade equivalent of the particular have become deeply implicated. It is widely cited as O-ring theory of economic development.17 a factor in the strong reaction of trade in 2008­09. The expla- In this case, it is of course in the interest of the assem- nation appears to be especially popular among Japanese blers to provide the component exporters the credit they economists (for example, Tanaka 2009), who must account need. But it is not obvious that the assemblers will be able for the fact that Japanese trade fell so dramatically in the cri- to obtain credit in a truly global credit crisis or that they sis (export volumes fell by an astounding 50 percent between will have the earnings with which to fund such credit February 2008 and February 2009). Japan's extensive involve- themselves, absent an ability to get the parts they need to ment in trade in parts and components is an alluring expla- assemble and export. Perhaps, then, disruptions to the sup- nation. ply of trade credit and production fragmentation interact. The evidence is not convincing, however. If the differ- Similarly, protectionism and supply chains may inter- ence now is that the parts and components in laptops are act. Freund (2009) observes that firms using global supply produced in Taiwan, China, but the machine is assembled chains tend to alter the location of production in a slump. in China, causing the components to cross national bor- She gives the example of Porsche, which decided to cut the ders and be counted twice in the trade statistics, it is true assembly of its cars in Finland in 2009 while maintaining that the same decline in the demand for laptops can result its operations in Germany, one presumes for political econ- in a larger recorded drop in trade, since it causes the vol- omy reasons, given that Porsche is a German-owned com- ume of global trade to fall by approximately the value of pany. In this case, it is precisely the exports assembled two laptops (ignoring the value added in assembly). But through international supply chains that disappear, despite while this factor can explain why the absolute value of the the fact that those products are identical down to the finest fall in trade was large, it cannot by itself explain why the detail to those that the German plant assembles for export. percentage fall in trade was so large or why the elasticity of trade with respect to income has been rising.14 With assem- The Role of Global Imbalances bly through global supply chains, there is twice as much trade in laptop parts and components. A fall in demand by As for the role of global imbalances in the crisis, some one laptop causes recorded trade to fall by twice as much. might expect them to have a place of prominence in this But with both the numerator and denominator multiplied discussion, since I have written in the past of the dangers of by two, elasticities are unchanged.15 their disorderly correction.18 Of course, the crisis about For production fragmentation to be implicated in the which I worried then was not exactly the same as the crisis collapse of trade, it is necessary to argue two things: first, the world went on to experience. With benefit of hindsight, that only some goods are produced using global supply most of the blame for the crisis appears to belong else- chains and, second, that goods so produced were affected where, although global imbalances probably played a role. most strongly by the negative demand shock. It is possible Fundamentally, I see the crisis as the result of flawed regu- to defend both arguments. In periods of high uncertainty, lation and perverse incentives in financial markets. Regulators firms and households will put off spending on big-ticket bought into the arguments of those regulated that financial items. (Baldwin 2009 refers to these items as "postpon- institutions could safely operate with a thinner capital cush- ables.") These postponements occur especially in the ion. They accepted the premise that capital adequacy could be uncertain conditions associated with financial disruptions, gauged on the basis of banks' internal models and, where these since big-ticket purchases have to be financed. Romer were absent, ratings of securities provided by commercial (1990) showed that it was demand for consumer durables credit rating agencies--notwithstanding the incentives for the that fell off so significantly in the early stages of the Great proprietors of the banks to tweak their models to minimize Depression. And a number of those products--motor estimated risks and capital requirements and the tendency for vehicles, consumer electronics--are now heavily involved the credit rating agencies, as investment advisers as well as in global supply chains. issuers of ratings, to fall prey to conflicts of interest. The Managing Openness: Lessons from the Crisis for Emerging Markets 21 regime that resulted was capital poor and dangerously pro- the center of the housing boom in the United States, they cyclical. Regulators neglected liquidity, assuming away prob- played a supporting role. lems in wholesale money markets. Banks were allowed to Beyond the housing market, the downward pressure hide risks in conduits and structured investment vehicles and on U.S. interest rates resulting from foreign official and window dress their balance sheets. Agency problems flour- private purchases of U.S. Treasury and agency securities ished at each stage of the originate-and-distribute process. could have contributed to the crisis through a number of Mortgage brokers had no fiduciary responsibility to home- channels. First, lower nominal interest rates encouraged owners. Banks not keeping a participation in the complex institutions to take on more risk to match previous nom- derivative securities they originated felt no responsibility to inal returns.20 Investors use nominal returns as a gauge investors. The structure of compensation encouraged bank of manager performance. If nominal returns go down, executives to roll the dice, disregarding the implications of they may take this decline as the manager's fault and their actions for the survival of the firm. And the regulators withdraw their funds. To retain their clients, managers averted their eyes. That is my summary of the crisis, in one are then forced to move into riskier assets and employ paragraph. more leverage. Of course, this summary goes only an inch below the Second, some investors, such as pension funds and surface. The deeper question is how these extraordinary cir- insurance companies, have fixed contractual liabilities. cumstances were allowed to arise. Here, I would cite a pow- They are required to pay out fixed nominal amounts to erful ideology of deregulation stretching back to at least the their investors. If market interest rates go down more than Reagan-Thatcher years. I would cite excessive confidence in the company expected when signing the contract, the yield quantitative methods of risk management, Value at Risk, on safe securities may not be enough for it to meet its obli- and of asset pricing. The academy, too, fell prey to a power- gations. Again, survival will require portfolio managers to ful collective psychology.19 In addition, the intensification move into riskier investments or take on more leverage. of competition, with the Glass-Steagall restrictions starting Banks that have issued certificates of deposit to their cus- to crumble even before passage of the Gramm-Bliley-Leach tomers and whose other liabilities bear fixed interest rates Act in 1999, encouraged banks to take on additional lever- may likewise find themselves squeezed. age in their desperation to maintain normal returns. Third, lower interest rates cheapen wholesale funding. Finally, a conscious policy in the United States of starving Lower wholesale money market rates encourage financial the regulators of human and financial resources had a role intermediaries to expand their balance sheets. The impact in the crisis. It is hard to understand the precrisis behavior will be most visible among broker-dealers relying on the of the Securities and Exchange Commission any other way. wholesale money market for much of their funding and That is my summary of the deeper causes of the crisis, among conduits and special-purpose vehicles that issue again in one paragraph. commercial paper to fund their investments in specula- But if the match that ignited the fire lay elsewhere, in lax tive assets.21 regulation and perverse incentives in financial markets, Finally, if lower interest rates and more ample liquidity global imbalances poured fuel on the flames. With signifi- boost equity prices, including the equity prices of financial cant amounts of foreign capital, official capital in particular, institutions themselves, those institutions will want to flowing toward the United States, long-term interest rates increase their lending to restore previous levels of leverage. were lower than they otherwise would have been. This Higher share prices for banks mean that they have more influx of capital fed the housing boom. Reinhart and Rogoff capital but also that they are not fully "loaned up." Some of (2009) show that the connection between capital inflows their capital is effectively sitting idle. If the firm's lending and housing booms is a historical regularity. My own capacity is not being fully used, it will seek to correct that ongoing work with Kevin O'Rourke and Augustin Benetrix situation. Low interest rates that translate into higher on housing booms and busts, using data for a panel of equity prices will thus trigger a lending boom. countries in the Organisation for Economic Co-operation The question is how much difference capital inflows and Development (OECD) in recent years, again suggests made for U.S. rates. Craine and Martin (2009) estimate that that developments in house prices are strongly correlated 10-year bond yields were at least 50 basis points lower in with capital flows. Foreign capital inflows into U.S. housing 2005 than they would have been had there been no addi- markets made it easier for financial institutions to finance tional foreign purchases from the beginning of 2004. the teaser rates on option-ARMs (adjustable-rate mort- Bandholz, Clostermann, and Seitz (2009) suggest that gages) that sucked more households into the market. 10-year bond yields were 70 basis points lower as a result Again, although global imbalances might not have been at of foreign capital inflows. Warnock and Warnock (2009) 22 Managing Openness suggest that the increase in U.S. Treasuries held by foreign- or a relative price that results from the elements constitut- ers depressed Treasury yields by 90 basis points. These views ing the development strategy, not a policy variable in and show a reasonably high degree of consensus on magnitudes, of itself.22 In China, to pick an example not entirely at ran- at least by the standards of the economics profession. dom, the strategy has been to restrain domestic consump- In the end, one must ask how different the course of the tion to mobilize large amounts of domestic savings for crisis would have been had 10-year bond yields been 50, 70, investment in capacity to produce tradable manufactures. or even 90 basis points higher. One answer is, not very dif- Limited financial development, a limited social safety net, ferent. The problems of lax regulation and skewed incen- and limited pressure on enterprise managers to pay out tives in financial markets would still have been there. The dividends are all mechanisms that help maintain this bal- problems implicit in the originate-and-distribute model ance of consumption and investment. With domestic would still have been there. Problems in the mortgage- consumption low, the relative price of nontraded goods is broking industry would still have been there. The conflicts low. The prices of exportables are relatively high. To of interest of the rating agencies would still have been there. observers ignorant of the policy mix, the renminbi looks The incentives for risk taking created by the structure of undervalued. But given the policy mix, the prevailing real executive compensation and too-big-to-fail would still have rate is the market equilibrium. Were it not, China would been there. With wholesale funding modestly more expen- experience faster inflation, and the real exchange rate sive, leverage modestly less, and investors stretching less for would adjust through this mechanism. yield, outcomes would have been less extreme. When the Should China now change its policy mix (more rap- boom unwound, it would have unwound less violently. But, idly)? The answer, logically, should flow from an analysis of qualitatively, outcomes would have been the same. the conditions that made the original policy mix desirable. Another answer to the "how different" question is, very The policy mix has been beneficial for some years now as a different. Economic dynamics are nonlinear. Crises are way of promoting the flow of resources into a manufactur- nonlinear. It is just conceivable that a difference of 70 basis ing sector that would have been suboptimally small, owing points would have meant an entirely different outcome. We to other distortions, in its absence. A policy mix that will never know. depresses the real exchange rate may be a second-best way of overcoming distortions (financial market underdevel- opment that limits the availability of start-up capital, for How Should Emerging Markets Respond? example) that would otherwise discourage the growth of The financial crisis was born and bred in the United States. high-value-added manufacturing.23 Or it may be a way of To the extent that global imbalances played a role, low U.S. encouraging activities that throw off positive externalities saving rates were, in turn, central to the development of (learning effects external to the manufacturing firm, for those imbalances. But it takes two to tango. The story example) that would otherwise be undersupplied by even a would be incomplete without also acknowledging the con- well-developed market. Both kinds of distortions have tribution of the surplus countries: China, emerging East probably been present in China, which is why this particu- Asia, the Middle East oil exporters, and surplus OECD lar development strategy has been so successful. countries like Germany and Japan. The question is whether those distortions have now The roles of these different countries and regions of become less pronounced, so that the authorities can begin course varied over time. Early on, surpluses were relatively modifying the policy mix. This is properly a question for evenly balanced, while more recently China in particular specialists on Chinese capital markets and Chinese manu- has dominated the surplus side of the equation. Emerging facturing, not for me. For what it is worth, I think China markets, as exemplified by China, now face a dilemma. has made good progress in terms of financial development. Should they stick with their tried-and-true development Enterprises are increasingly able to float bonds and borrow strategy, which has entailed restraining domestic consump- from banks, permitting them to rely less on the retained tion, keeping the real exchange rate low, and plowing sav- earnings they amass as a result of the prevailing policy mix. ings into investment in tradable manufactures, and thereby (To be clear, increasingly able does not mean freely able.) risk the reemergence of global imbalances and the associ- Through integration, collaboration, and the development ated crisis risks as demand again picks up in the United of supply chains and production networks, manufacturing States? Or should they abandon that strategy for another? firms are better able to appropriate some of the positive In thinking about this problem, economists should not externalities thrown off by their activity. (In this case, better become fixated on the nominal exchange rate (the politi- appropriate is different from fully appropriate.) If this cians can argue over that). The exchange rate is an outcome assumption is correct, then the policy response should be Managing Openness: Lessons from the Crisis for Emerging Markets 23 to begin to move gradually away from the prevailing policy to shield relatively fragile developing economies. For devel- mix. Policy makers can encourage consumption (by devel- oping countries where trade remains the principal channel oping the social safety net and liberalizing financial mar- through which shocks are transmitted, recent events under- kets). They can encourage enterprises to pay out dividends score the importance of making contingency plans for the (by reforming corporate governance). As consumption of possibility that trade credit might dry up and exports may nontraded goods, among other things, rises in response, collapse. Central banks should hold reserves to fill the the real exchange rate will adjust. China can take the trade-credit gap. They should establish and fund specialized adjustment either through inflation (which will raise the export credit agencies. They should prearrange support relative price of nontraded goods) or renminbi apprecia- with multilaterals and other extranational agencies in a tion (which has the same effect). My own preference would position to help. Given the special sensitivity to such dis- be for the latter. ruptions of durable manufactures produced through global How quickly should it move? The answer depends on supply chains, countries heavily dependent on these prod- how quickly the distortions just described diminish; that ucts should redouble their efforts at export diversification. much is obvious. But it also follows that, since the diminu- For emerging markets where financial links are now the tion of financial market constraints, the development of principal channel through which foreign shocks are trans- collaborative relationships among firms are gradual rather mitted, the regulatory framework for domestic financial mar- than discontinuous processes, the change in the policy mix, kets needs to be strengthened. Such strengthening means and therefore the level of the real exchange rate, should also tightening supervision and regulation along the obvious adjust gradually rather than discontinuously. A sudden dimensions and, given the crisis, worrying more about lever- sharp appreciation of the renminbi, in other words, does age, liquidity. and transparency. It means using a portfolio of not seem the best approach. This logic calls for gradual policies to deal with capital inflows associated with the carry appreciation over time. trade: first, fiscal tightening; second, tightening limits on But if one believes that global imbalances contributed to lending by domestic banks; third, additional exchange-rate the crisis and that China's large surpluses, emanating from flexibility to introduce two-way bets into financial markets; its policy mix, contributed to those imbalances, then this fourth, sterilized intervention; and fifth (and finally, if the argues for rather faster appreciation. If one believes that preceding measures do not work), taxes on capital inflows. China's policies (of course, not only China's), operating It means holding reserves adequate for dealing with the through the channel of global imbalances, have implica- consequences of sudden stops and, indeed, with the whole- tions for global financial stability (and, thus, implications sale liquidation of foreign holdings. Achieving this last goal external to the country), then it should optimally step up means two things: one, identifying more precisely what the pace of renminbi appreciation. And what is logically constitutes an adequate level of reserves under these cir- true of China is true, to a greater or lesser extent, of other cumstances; and, two, negotiating reserve-pooling and emerging economies in East Asia and other parts of the emergency-swap facilities to minimize the cost of reserves, world. whether at the regional level--such as the Chiang Mai Ini- tiative Multilateralization, CMIM, or the Latin American Reserve Fund, FLAR--or through bilateral swaps with the In Sum Federal Reserve and the European Central Bank, or at the Openness has been tremendously beneficial for emerging International Monetary Fund. markets. Over the past 30 years, the shift toward a more Finally, emerging markets must think about gradually market-led system, stable macroeconomic and financial transitioning away from a tried-and-true growth model policies, and greater openness in international transactions that has emphasized saving at the expense of consumption, have yielded enormous benefits in economic development slowed financial development, and successfully promoted and growth. Notably, greater openness is only one of a con- export-led growth but at the same time contributed to stellation of related policies: it is not openness per se that global imbalances. China and others are already committed matters but its combination with other policies. While this to this transition. But to complete it successfully, they need makes it difficult to isolate the contribution of openness a clearer understanding of the underlying distortions that among other factors to the improvement in economic per- made for the success of the earlier strategy. Without this formance, most would share a strong intuitive sense that insight, it is hard to know how quickly to move away from openness has played an important role. it. And they need to bear in mind that policies that had But openness also has a downside in that it exposes coun- unquestionable benefits domestically also added fuel to the tries to external shocks. It heightens the need for policies fire that resulted in the financial crisis. 24 Managing Openness Notes Alexander, L., et al. 2008. "Global Recession and Response." Citigroup Eco- nomic and Market Analysis. December 3. http://www.realclear 1. This chapter was first presented as a paper at the World Bank Con- markets.com/articles/econ1.pdf. ference on Managing Openness, May 10, 2010. I thank Gisela Rua for Baldwin, R. 2009. "The Great Trade Collapse: What Caused It and What helpful research assistance. Does it Mean?" VoxEU. November 27. http://voxeu.org. 2. Berkmen et al. (2009) find some support for the hypothesis that Bandholz, H., J. Clostermann, and F. Seitz. 2009. "Explaining the US Bond countries with stronger fiscal positions were hit less severely but caution Yield Conundrum." Applied Financial Economics 19: 539­50. that this evidence is weak. Budget data here are from the Economist Intel- Berkmen, P., G. Gelos, R. Rennhack, and J. Walsh. 2009. "The Global ligence Unit. Financial Crisis: Explaining Cross-country Differences in the Output 3. As in Eichengreen, Rose, and Wyplosz (1995). Impact." IMF Working Paper WP/09/280, International Monetary 4. Data for five outliers--Algeria, Botswana, Benin, Cape Verde, and Fund, Washington, DC. Uganda--were dropped for clarity. Including them deforms the parabola Blanchard, O., H. Faruqee, and M. Das. 2010. "The Initial Impact of the a bit but does not change the story. Crisis on Emerging Market Countries." Brookings Papers on Eco- 5. IMF (2010) dissents from this emerging consensus, concluding nomic Activity. Brookings Institution, Washington, DC. that there was no difference in the depth of the recession between coun- Craine, R., and V. Martin. 2009. "Interest Rate Conundrum." B. E. Journal tries with pegs and floats. of Macroeconomics 9: 1­27. 6. Alexander et al. (2008) is an example of a study showing that the Eichengreen, B. 2007. Global Imbalances and the Lessons of Bretton Woods. severity of the crisis was increasing in the size of the financial sector. Cambridge, MA: MIT Press. 7. Note that I show the least squares regression line both with and ------. 2008. "The Real Exchange Rate and Economic Growth." Growth without the outlier, Latvia. Commission Working Paper 4, World Bank, Washington, DC. 8. See, for example, World Bank (2010b), chapter 2, for a summary of ------. 2009. "The Last Temptation of Risk." National Interest. May­June. the evidentiary base. http://www.nationalinterest.org. 9. Latvia and Korea were among the countries with the lowest ratios of Eichengreen, B., R. Gullapalli, and U. Panizza. 2009. "Capital Account Lib- deposits to private sector loans; neither did well during the crisis. More eralization, Financial Development and Industry Growth: A Synthetic generally, Berkmen et al. (2009) and World Bank (2010a) show that coun- View." UNCTAD Polis Working Paper 144, United Nations Confer- tries with more leveraged domestic financial systems (higher ratio of ence on Trade and Development, Geneva, Switzerland. domestic credit to domestic deposits) did poorly in the crisis. Eichengreen, B., and D. Irwin. 2009. "The Slide to Protectionism in the 10. Thus, Mody (2010) finds that economies that had overheated in Great Depression: Who Succumbed and Why." NBER Working Paper 2008 saw larger decelerations in 2009. Berkmen et al. (2009) and World 15142, National Bureau of Economic Research, Cambridge, MA. Bank (2010a) similarly find that countries with more rapid credit growth Eichengreen, B., A. Rose, and C. Wyplosz. 1995. "Exchange Market May- tended to suffer larger growth decelerations. Of course, any Polish policy hem: The Antecedents and Aftermath of Speculative Attacks." Eco- maker could have told you this. nomic Policy 21: 249­313. 11. A few dramatic counterexamples like Japan notwithstanding. Eichengreen, B., and J. Sachs. 1985. "Exchange Rates and Economic 12. This was the influential view of Nurkse (1944). Evidence for it is in Recovery in the 1930s." Journal of Economic History 49: 341­59. Eichengreen and Sachs (1985). Evenett, S., ed. 2010. Will Stabilization Limit Protectionism? The Fourth 13. Normally one would think them ambiguous: the direct spending Global Trade Alert. London: Centre for Economic Policy Research. effect on other countries is positive, but the positive impact on interest Freund, C. 2009. "Demystifying the Collapse in Trade." VoxEU. July 3. rates of fiscal expansion is negative, since it crowds out investment in http://voxeu.org. neighboring countries. In a little trap, of course, the second channel is Gambacorta, L. 2009. "Monetary Policy and the Risk-Taking Channel." rendered inoperative. BIS Quarterly Review (December): 43­53. 14. As documented by Freund (2009). Iacovone, L., and V. Zavacka. 2009. "Banking Crises and Exports: Lessons 15. A nice exposition of this is O'Rourke (2009). from the Past." Policy Research Paper 5016, World Bank, Washington, DC. 16. One is reminded of some of the incipient disruptions to trade IMF (International Monetary Fund). 2010. World Economic Outlook. and production in Europe with the Icelandic volcano eruptions of April Washington, D.C.: IMF. 2010. Kee, H. L., C. Neagu, and A. Nicita. 2010. "Is Protectionism on the Rise? 17. See Kremer (1993). Assessing National Trade Policies during the Crisis of 2008." Policy 18. In Eichengreen (2007). Research Working Paper 5274, World Bank, Washington, DC. 19. A longer reflection on the role of economists in the crisis is Kremer, M. 1993. "The O-Ring Theory of Economic Development." Eichengreen (2009). Quarterly Journal of Economics 108: 551­75. 20. This is the mechanism discussed by Gambacorta (2009). MercoPress. 2010. "Crisis Lesson: Better to Rely on International Reserves 21. The effect will be less, though by no means absent, among com- than IMF Credit says Brazil." http://www.en.mercopress.com. April 24. mercial banks relying on retail deposits for most of their funding. That Mody, A. 2010. "Who Fell in 2009." VoxEU. January 21. http://voxeu.org. the expansion of balance sheets should be proportionately greater among Moghadam, R. 2010. "Emerging Market Countries and the Crisis: How broker-dealers than commercial banks is emphasized by Adrian and Have They Coped?" April 19. http://www.imf.org. Shin (2009). Mora, J., and W. Powers. 2009. "Decline and Gradual Recovery of Global 22. As I argue at more length in Eichengreen (2008). A similar argu- Trade Financing: U.S. and Global Perspectives." VoxEU. November 27. ment is Song, Storensletten, and Zilibotti (2010). http://voxeu.org. 23. High-value-added relative, specifically, to agriculture and tradi- Nurkse, R. 1944. International Currency Experience. Geneva: League of tional manufacturing. Nations. Obstfeld, M., J. Shambaugh, and A. Taylor. 2009. "Financial Instability, Reserves and Central Bank Swap Lines in the Panic of 2008." NBER Bibliography Working Paper 14826, National Bureau of Economic Research, Cam- bridge, MA. Adrian, T., and H. Shin. 2009. "Financial Intermediaries and Monetary O'Rourke, K. 2009. "Collapsing Trade in a Barbie World." http://www Economics." Staff Report 398, Federal Reserve Bank of New York. .irisheconomy.ie/index.php/2009/06/18/collapsing-trade-in-a-barbie- New York. world/. Managing Openness: Lessons from the Crisis for Emerging Markets 25 Reinhart, C., and K. Rogoff. 2009. This Time is Different: Eight Centuries of Tong, H., and S.-J. Wei. 2009. "Composition Matters: Capital Inflows and Financial Folly. Princeton, NJ: Princeton University Press. Liquidity Crunch during a Global Economic Crisis." NBER Working Romer, C. 1990. "The Great Crash and the Onset of the Great Depres- Paper 15207, National Bureau of Economic Research, Cambridge, MA. sion." Quarterly Journal of Economics 98: 85­106. Vlachos, J., and D. Waldenstrom. 2005. "International Financial Liberal- Rose, A., and M. Spiegel. 2009. "Cross-country Causes and Consequences ization and Industry Growth." Journal of International Finance and of the 2008 Crisis: International Linkages and American Exposure." Economics 10: 264­84. NBER Working Paper 15358, National Bureau of Economic Research, Warnock, F., and V. Warnock. 2009. "International Capital Flows and U.S. Cambridge, MA. Interest Rates." Journal of International Money and Finance 28: 903­19. Schindler, M. 2009. "Managing Financial Integration: A New Data Set." World Bank. 2010a. "From Global Collapse to Recovery." Office of the IMF Staff Papers 56, International Monetary Fund, Washington, DC. Chief Economist for Latin America and the Caribbean, World Bank, Song, Z., K. Storesletten, and F. Zilibotti. 2010. "The `Real' Causes of Washington, DC. China's Trade Surplus." VoxEU. May 3. http://voxeu.org. ------. 2010b. Global Economic Prospects. Washington, DC: World Bank. Tanaka, K. 2009. "Trade Collapse and Vertical Foreign Direct Investment." Wyplosz, C. 2007. "The Fuss over Foreign Exchange Reserve VoxEU. May 7. http://voxeu.org. Accumulation." VoxEU. May 28. http://voxeu.org. 3 Global Imbalances: Past and Future Luis Servén and Ha Nguyen Global imbalances--that is, large current account deficits in these two variables in different countries and regions and surpluses in the global economy--have taken center provide some information about the sources of changes in stage in the debate on the international economic outlook. their respective external imbalances. Figure 3.3 shows that Academic and policy scholars have offered contrasting rising saving has led to widening current account sur- views about their role in precipitating the global crisis, the pluses in emerging economies (namely, oil-exporting potential threat of the imbalances to economic stability countries and Asia). In oil-exporting countries, this was around the world, and the policy measures that should be driven by the persistent rise in world oil prices, while in taken to "rebalance" the global economy. This chapter dis- Asia this reflected rising saving in the region's emerging cusses the nature of global imbalances and their implica- markets. In particular, extremely high levels of saving are tions for economic stability and for developing-country observed in China (over half of GDP). Saving and invest- growth strategies. ment rates have both risen in China since 2000, but sav- The U.S. current account deficit grew virtually without ings grew at a faster rate. A disaggregate analysis reveals interruption from 1995 to 2005 (figure 3.1). It peaked in that the increase in total saving reflects primarily the 2005 and 2006 at over 1.5 percent of world gross domestic rise of corporate saving, which in recent years has hov- product (GDP). Thereafter, it declined to about 1.2 percent ered around 20 percent of GDP, although household of world GDP in 2008, and preliminary estimates suggest saving has also been on the rise (Kujis 2005, 2006; that it fell under 1 percent in 2009. During the late 1990s, Prasad 2009). Many observers attribute China's high lev- the counterparts to the U.S. deficits were large surpluses els of corporate saving to the weak corporate gover- from Japan and emerging Asian countries, excluding nance of large enterprises.1 In any event, the result has China, as well as the surplus of the European Union (EU) been a major increase in China's current account sur- during the years of the Asian crisis (1996­97). After 2001, plus, which peaked at 10 percent of GDP in 2007. however, the situation changed radically. As the U.S. deficit China's high level of saving stands in sharp contrast with grew, the surpluses of China and the oil-exporting coun- low saving in the United States. Indeed, falling saving was tries accounted for an increasing share of the U.S. overall the principal source of the increasing external deficits of the current account deficit (figure 3.2). Moreover, since 2005, United States from the end of the 1990s to 2003. After 2003, the surplus of the EU has almost completely disappeared, the trends in the current account were driven primarily by and China's surplus has grown to exceed the combined changes in investment--increasing until 2006 and then surpluses of Japan and the rest of emerging Asia. During decreasing. In particular, the reduction of the external 2007­08, the bilateral deficit with China accounted for imbalance of the United States in 2007­08 reflects the fact around 40 percent of the overall U.S. deficit. Hence, the that, although saving rates remained on a declining trend, popular (but inaccurate) view of global imbalances as a investment rates fell even more quickly. Aside from the problem of the United States versus China has started United States, other industrial countries have undergone making some sense only in the past two or three years. relatively modest changes in saving and investment rates. Since the current account surplus is identically equal to In the EU, the slightly decreasing trend in saving over the the difference between saving and investment, the trends past decade led to a gradual reduction, and eventually a 27 28 Managing Openness Figure 3.1. Current Account Imbalances for Selected Countries as a Percentage of World GDP, 1980­2009 3.0 2.5 2.0 1.5 1.0 percent 0.5 0.0 ­0.5 ­1.0 ­1.5 ­2.0 ­2.5 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 year China Asia (excluding China) European Union United States oil exporters other countries Source: World Economic Outlook (database). International Monetary Fund (IMF), http://www.imf.org/external/ns/cs.aspx?id=28. Note: GDP = gross domestic product. The 2009 data are the International Monetary Fund's estimates. Figure 3.2. U.S. Bilateral Current Accounts with Other Countries, 1999­2009 100 0 ­100 ­200 US$, billions ­300 ­400 ­500 ­600 ­700 ­800 ­900 99 00 01 02 03 04 05 06 07 08 09 19 20 20 20 20 20 20 20 20 20 20 year China Japan European Union Organization of the Petroleum Exporting Countries other countries Source: International Economic Accounts (database), Bureau of Economic Analysis, http://www.bea.gov/international. reversal, of the area's current account surplus. In Japan, In spite of their recent rise to prominence in the debate saving and investment rates both followed a downward over the roots of the crisis, global imbalances are not a new trend over the 1990s. However, in the past decade, both phenomenon. In fact, the 1980s witnessed a situation qual- rates remained fairly stable, and the current account main- itatively similar to that observed in recent years, character- tained a modest surplus. ized by large U.S. current account deficits funded by other Global Imbalances: Past and Future 29 Figure 3.3. Saving and Investment Rates as a Percentage of GDP in Selected Countries, 1990­2008 a. United States b. China 60 60 50 50 saving 40 40 percent percent 30 30 investment investment 20 20 10 saving 10 0 0 90 92 94 96 98 00 02 04 06 08 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 20 20 20 20 20 19 19 19 19 19 20 20 20 20 20 year year c. European Union d. Japan 60 60 50 50 40 percent saving percent 40 30 saving 30 20 20 investment investment 10 10 0 0 90 92 94 96 98 00 02 04 06 08 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 20 20 20 20 20 19 19 19 19 19 20 20 20 20 20 year year e. Asia (excluding China, including Japan) f. Organization of the Petroleum-Exporting Countries 60 60 50 50 saving percent percent 40 40 saving 30 30 20 20 10 investment 10 investment 0 0 90 92 94 96 98 00 02 04 06 08 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 20 20 20 20 20 19 19 19 19 19 20 20 20 20 20 year year Source: World Development Indicators (database), World Bank, http://data.worldbank.org. World Economic Outlook (database), IMF, http://www.imf.org/external/ns/cs.aspx?id=28. countries (figure 3.1). Two important differences, however, different from those at play in the 1980s. Understanding distinguish these episodes. First, both the duration and the such factors is important to assessing how global imbal- magnitude of the U.S. deficits of the 1980s were substan- ances may evolve after the world crisis and to gauging the tially lower than those in recent years. In particular, in the potential risks that their continued presence could pose to 1980s U.S. deficits exceeded 1 percent of world GDP in the world economy. only 3 years, as opposed to 10 years in the present circum- The rest of this chapter is organized as follows. The next stance. Second, in recent years emerging economies have section summarizes various views in the academic literature been primarily responsible for funding the U.S. external on the nature of global imbalances. The following section deficit (as well as the deficits of other advanced countries), reviews recent trends in global imbalances and discusses unlike in the 1980s when such funding came mostly from alternative scenarios for their future evolution, their impact other rich countries, primarily Japan. In other words, the on developing countries, and the appropriate policy recent global imbalances involve a flow of capital from responses. The final section concludes. poor countries to rich countries, against the prediction of conventional economic theory that developing countries The Nature of Global Imbalances should be net capital importers.2 These distinguishing features of the global imbalances At the risk of over-simplifying, we can distinguish two basic of recent years suggest that the factors behind them may be views among academics and policy analysts on the nature of 30 Managing Openness global imbalances. What we shall label for want of a better Figure 3.4. U.S. Multilateral Real Exchange Rate Index, term as the disequilibrium approach considers global imbal- 1980­2009 quarter 1, 2002 = 100 ances an unsustainable phenomenon, requiring adjustment of the U.S. current account and a major depreciation of the 150 dollar. This correction could come in the form of a sudden 140 stoppage of capital flows into the United States and collapse 130 120 of the exchange rate (Roubini 2009, for example). The sec- 110 index ond view, or equilibrium approach, asserts that global 100 imbalances represent a situation that, absent changes in its 90 deep determinants, can be self-sustaining. We next review 80 70 the main lines of both approaches. 60 50 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 The Disequilibrium Approach 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 year The view that current global imbalances are unsustainable Source: International Financial Statistics (database), IMF, Washington, DC. http://www.imfstatistics.org. starts from the intertemporal budget constraint dictating that a country's net liability position against the rest of the world at any given time cannot exceed the present value of its future current account surpluses.3 This requirement the trends in the real effective exchange rate of the dollar makes it entirely possible for a country to run current during that episode (1981­92) with those observed in the account deficits for a long time, as long as it is capable of past decade (1998­2009). From its peak at the beginning of running sufficiently large surpluses in the future. Such 1985, the dollar had depreciated over 40 percent by the end could be the case of developing countries that borrow from of 1991. The virtual elimination of the current account developed countries to invest and accumulate capital and deficit accompanied the depreciation. In contrast, over the repay their debts once they reach a higher stage of develop- past decade the dollar has followed a pattern of deprecia- ment (Kraay et al. 2005). Alternatively, a developed country tion similar to that observed in the 1980s, although the could also run current account deficits if in the future it is magnitude of the depreciation since the peak in early 2002 expected to grow faster than the rest of the world. In effect, to date has been smaller--around 30 percent. Moreover, the country finances its consumption by borrowing against external deficits have remained quite large, at least until its future income. In this vein, Engel and Rogers (2006) 2008. Under the disequilibrium view, these deficits suggest argue that the U.S. current account deficit might be an out- that further real depreciation of the dollar is still to come. come of such intertemporally optimizing behavior. In fact, numerous studies have subjected the magnitude of Many observers are concerned, however, that the large the trade balance correction, and of the depreciation neces- U.S. external deficits are the result of an unsustainable sary to achieve it, to detailed analysis (Obstfeld and Rogoff increase in public or private expenditure, prompted 2005, 2009, for example). respectively by fiscal expansions and financial innovations. Correcting these external imbalances, according to the Whether the deficits reflect intertemporally optimizing disequilibrium view, demands a real adjustment--a reversal behavior or excessive spending, the net foreign asset posi- of the trade balance. But recent literature has underscored tion of the United States has clearly undergone a steep the potentially important role that financial adjustment decline. From a creditor position amounting to 10 percent can also play. Changes in the prices of a country's foreign of its GDP at the beginning of the 1980s, the United States assets and liabilities also affect its net foreign asset position. had shifted into a debtor position approaching 25 percent In particular, changes in asset prices--that is, capital gains of GDP in 2009. In absolute terms, this debtor position is and losses on foreign assets and liabilities--imply that the the biggest in the world. Thus, according to the disequilib- current account balance no longer determines the change rium approach, this trend is unsustainable, and the coun- in net foreign assets. try needs to change the sign of its trade balance at some Although financial adjustment has received little atten- point. Such a shift, in turn, would entail a depreciation of tion, it is especially important in the U.S. case owing to the dollar to increase U.S. net exports. return differentials: assets held by U.S. investors abroad According to this view, the adjustment process might yield higher returns than U.S. assets held by foreign not be very different from the one that led to elimination of investors (Hausmann and Sturzenegger 2004; Gourinchas the U.S. external deficits of the 1980s. Figure 3.4 compares and Rey 2007a; Forbes 2010). Hausmann and Sturzenegger Global Imbalances: Past and Future 31 (2004) note that, despite holding a negative net foreign ciently detailed information on the composition of U.S. asset position of nearly 20 percent of GDP in 2004, the external assets and liabilities (Curcuru, Thomas, and United States still earned a US$30 billion positive net Warnock 2008). Observers, however, generally agree that return that year.4 financial adjustment does play a significant role and that In turn, depreciation of the dollar also favors the United ensuring sustainability of the U.S. external position will States because its external liabilities are denominated require a much more modest depreciation of the dollar mostly in dollars, while its assets are denominated in other than what would be necessary if external adjustment had currencies. Therefore, a depreciation of the dollar gener- to take place only from the trade balance. ates a wealth transfer in favor of the United States: the As an illustration, figure 3.5 (from Nguyen 2010) offers value of its liabilities falls relative to the value of its assets. an assessment of the financial adjustment that occurred This process is exactly the reverse of what usually happens during the period 1994­2007 through changes in stock and in emerging markets when their exchange rate undergoes a bond prices. After 2002, it generated large gains for the real depreciation because they are typically indebted in for- United States, which peaked at 4 percent of U.S. GDP in eign currency. 2006 and 2007--an order of magnitude similar to that of Hence, depreciation of the dollar has a dual effect on the the current account deficits of those years. Two main fac- external asset position of the United States. On the one tors were at play. First was the relative decline in U.S. stock hand, it generates a real adjustment through an improving market prices since 2000, which generated capital losses for trade balance. On the other hand, it generates a financial foreign investors (see Kraay and Ventura 2005 for a detailed adjustment through capital gains for the United States discussion). Second was the depreciation of the dollar after (losses for the rest of the world). Simplistic assessments of 2002. the depreciation of the dollar--and the trade surplus-- required to put the external position of the United States The Equilibrium Approach on a sustainable trajectory can lead to exaggerated conclu- sions unless they take into consideration this second effect, In contrast to the disequilibrium view, the equilibrium which is becoming increasingly important given the sharp approach explains global imbalances as the result of struc- rise in cross-country asset holdings in the past two decades. tural factors or policies in other countries that have led to a How big is the financial adjustment effect? The topic steady accumulation of assets in the United States by the rest has been debated extensively (for example, Blanchard, of world. Absent changes in such structural factors and pol- Giavazzi, and Sa 2005; Gourinchas and Rey 2007b). It is icy choices, global imbalances could persist. Although details hard to give a precise answer because of the lack of suffi- vary in different versions of the equilibrium approach, one Figure 3.5. U.S. Valuation Effects and Current Account as a Percentage of GDP, 1994­2007 6 4 2 percent 0 ­2 ­4 ­6 ­8 94 95 96 97 98 99 00 01 02 03 04 05 06 07 19 19 19 19 19 19 20 20 20 20 20 20 20 20 year U.S. valuation effects U.S. current account Source: Nguyen 2010. 32 Managing Openness common feature is an emphasis on the capital account. This Under appropriate circumstances, the mercantilist emphasis contrasts with that of the disequilibrium approach strategy may succeed in accelerating economic growth. on the current account. However, its merits from the welfare standpoint are ques- We can further distinguish two main versions of the tionable, because large-scale accumulation of external (and equilibrium approach. The first one underscores interna- typically low-yield) assets involves major opportunity costs tional asymmetries in the supply of or demand for financial in forgone consumption and investment (Korinek and assets. Caballero, Fahri, and Gourinchas (2008a, 2008b) Servén 2010). Nevertheless, as long as Asian emerging mar- suggest that international savers prefer assets of coun- kets stick to the export-led strategy, global imbalances can tries with more advanced financial markets--the United remain in operation indefinitely. States in particular--because local assets in emerging Alternatively, emerging-market authorities may choose countries are plagued by volatile yields or risk of expropri- to accumulate foreign assets for precautionary reasons.6 In ation. A growth acceleration in emerging countries (or an the absence of mechanisms for international diversification oil price boom) that increases their wealth and saving-- of aggregate risk, emerging countries integrated into the the ultimate causes of the so-called global saving glut global financial system need to self-insure against external (Bernanke 2005)--leads them to expand their holdings of shocks such as disruptions of international capital flows U.S. assets. Such an expansion, however, can occur only by (for instance, like what occurred in the crises of Asia and raising the volume of U.S. assets available to international the Russian Federation in the 1990s). Emerging countries investors and thus increasing the U.S. current account accumulate external assets, preferably short-term instru- deficit. This process can persist as long as its driving force-- ments, from which they can draw in the event of a "sudden the underdevelopment of financial markets in emerging stop." Unless the global financial system generates new countries--remains unchanged. As a result, capital flows international diversification mechanisms, this precaution- "uphill," from poor to rich countries. ary accumulation of foreign assets is unlikely to end. On the demand side, international asymmetries may What does the empirical evidence say about the validity arise from the limited appropriability of returns on emerg- of these arguments? To begin with, the massive accumula- ing-market assets (Mendoza, Quadrini, and Rios-Rull tion of international reserves by emerging economies dur- 2009). Asymmetries may also arise from the shortcomings ing the past decade seems to confirm that deliberate of the social protection system in emerging economies hoarding has played an important role. Between 1998 and (Carroll and Jeanne 2009), which force individuals to save 2008, international reserve holdings of emerging countries more for retirement or to protect themselves from the risk (figure 3.6, panel a) increased fourfold, while those of indus- of unemployment. In either case, the result is that savers in trial countries rose only 50 percent (figure 3.6, panel b). As a emerging countries tend to save more than those in indus- result, the volume of international reserves held by emerg- trial countries. In a context of international financial inte- ing markets at present greatly exceeds that of industrial gration, this tendency leads to a global equilibrium in countries. For example, at the end of 2008, China's foreign which emerging countries acquire a creditor position, reserve stock was almost as large as that of all industrial whereas advanced countries are net debtors. If the ultimate countries combined. Reserve holdings in the rest of determinants of this equilibrium remain unaltered, global emerging Asia have also increased dramatically. But the imbalances and uphill capital flows can persist indefinitely. phenomenon is not confined to Asia; Latin American The second version of the equilibrium approach economies (with Chile at the top) and oil-exporting coun- emphasizes policy makers' choices as the main cause of tries have also accumulated large volumes of international the accumulation of external assets by emerging markets. reserves over the past decade. Again, the argument has two variants. The first variant Even if the accumulation of international reserves, or asserts that hoarding foreign assets results from the so- external assets more broadly, was a deliberate policy choice called new mercantilist development strategy: the of emerging economies, the question remains whether cau- attempt of a number of emerging markets--particularly tion in the face of volatile international capital flows or the in East Asia--to pursue export-led growth. The mercan- pursuit of competitive exchange rates was the driver. tilist strategy calls for an undervalued real exchange rate Aizenman and Lee (2007) examine the question empiri- to preserve export competitiveness.5 The best way to cally using data from 49 countries in the 1980­2000 achieve this outcome is by compressing domestic spend- period; they conclude that both motives were at work but ing, particularly consumption, which inevitably leads to that the precautionary saving motive was more important. persistent current account surpluses and accumulation Jeanne and Ranciere (2009), in turn, conclude that the of foreign reserves. accumulation of external assets since 2000 has been too Global Imbalances: Past and Future 33 Figure 3.6. Foreign Reserve Stock of Industrial and Emerging Countries, 1991­2008 constant 2000 US$, billions a. Emerging countries b. Industrial countries 4,500 4,500 3,500 3,500 2,500 2,500 1,500 1,500 500 500 0 0 92 94 96 98 00 02 04 06 08 92 94 96 98 00 02 04 06 08 19 19 19 19 20 20 20 20 20 19 19 19 19 20 20 20 20 20 China United States Japan EU Emerging Asia (excluding China) Latin America and the Caribbean oil exporters Source: World Development Indicators (database), World Bank, http://data.worldbank.org. large, particularly in Asia, to be justified by the precaution deficit of the United States, measured as a proportion of motive alone. GDP, has fallen to less than half of its peak--from 6.1 per- If deliberate policy choices were the main force behind cent of GDP in the second quarter of 2006 to 2.8 percent in the accumulation of U.S. assets, however, official capital the second quarter of 2009 (figure 3.8). Moreover, the inflows to the United States should predominate over pri- decline accelerated during 2009, undoubtedly helped by vate flows. Figure 3.7, though, shows that the picture is the recession.7 The contraction of the external imbalance more mixed. Net purchases of U.S. assets by central banks of the United States reflects, on the one hand, the decline of and other government bodies from emerging markets in private investment, especially in construction, and, on the Latin America, Asia, and the Middle East have grown other hand, an increase in private saving. increasingly large in the 2000s. After the onset of the crisis Thus, are global imbalances going to disappear? It is dif- in 2007, they became the sole source of inflows from ficult to give a conclusive answer about their future. It these countries. But in the years up to 2007, non-official depends on a constellation of real and financial forces investors accounted for the majority of emerging-market whose evolution is difficult to predict. Nevertheless, it may inflows to the U.S. economy. be useful to think in terms of two broad scenarios--the The predominance of private capital in the total flows return of global imbalances versus the narrowing of global from emerging markets in the run-up to the crisis lends some imbalances--and their consequences for developing coun- support to the first version of the equilibrium approach: that tries' growth strategies. global imbalances are primarily caused by asymmetries in the supply of or demand for international assets. Forbes (2010) The Return of Global Imbalances offers some corroborating evidence, based on an analysis of the geographical origin of private capital flows to the United To the extent that the deep determinants of the imbalances States. She finds that investors from countries with less- remain largely unchanged, global imbalances may well be developed financial markets tend to hold greater shares restored after the crisis. Several indicators seem to point to of their investment portfolios in the United States, which their likely return. is in accordance with the equilibrium approach espoused First of all, the crisis did not lead to a "sudden stop" of by Caballero, Farhi, and Gourinchas (2008a, 2008b) and capital flows to the United States, which could have given Mendoza, Quadrini, and Rios-Bull (2009) discussed above. rise to a disorderly adjustment of global imbalances and an abrupt depreciation of the dollar. Instead of the deprecia- tion that many had predicted, the dollar experienced an The Future of Global Imbalances and initial appreciation (see again figure 3.4). International Implications for Developing Countries investors "flew to safety" in low-risk U.S. Treasury debt, at The world crisis has led to a reduction, or at least a tempo- the expense of all risky assets--from corporate debt to rary one, in global imbalances. Indeed, the current account emerging-market assets. The dollar became the reserve 34 Managing Openness Figure 3.7. Gross Capital Inflows to the United States from Emerging Markets, 2000­09 1,200 1,000 800 US$, billions 600 400 200 0 00 01 02 03 04 05 06 07 08 09 20 20 20 20 20 20 20 20 20 20 year total inflows inflows from official sources Source: International Economics Accounts (database), Bureau of Economic Analysis, http://www.bea.gov/international. Figure 3.8. U.S. Current Account as a Percentage of GDP, Quarters 2 and 4, 1990­2010 2 1 0 ­1 percent ­2 ­3 ­4 ­5 ­6 ­7 90 91 92 93 2 4 2 5 2 6 2 7 2 8 2 9 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 10 Q 99 Q 9 Q 99 Q 99 Q 99 Q 99 Q 00 Q 00 Q 00 Q 00 Q 00 Q 00 Q 00 Q 00 Q 0 Q 00 19 19 19 19 19 20 20 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Q Q Q Q Q quarter Source: International Economics Accounts (database), Bureau of Economic Analysis, http://www.bea.gov/international. currency of last resort, and the government of the United prime turmoil in mid-2007, these patterns changed States, the borrower of last resort. Paradoxically, the United abruptly: capital inflows from nonresidents collapsed, States, undeniably the source of the crisis, also turned out and outflows of residents reversed, reflecting capital repa- to be the last refuge of international investors. triation by residents to stem losses in domestic markets or Throughout the global turmoil, the United States has to seek safe haven from the global turbulence. In 2009, had no difficulty financing its (narrowed) external imbal- however, the data reveal a return to precrisis trends: capi- ance. However, capital flows to the United States have tal inflows of foreign investors and outflows of resident undergone important changes (see figure 3.9). Capital investors. flows from nonresident investors were positive and fol- Figure 3.10 offers an additional look at the capital flow lowed an upward trend from the late 1970s until late pattern. Foreign investors steadily accumulated all types of 2006. Capital flows from resident investors in recent U.S. assets until 2008. As the crisis hit in 2008, investors decades were mostly negative, indicating capital outflows withdrew from U.S. assets for a short time, selling off from the United States. However, at the onset of the sub- agency bonds in particular. Thereafter, investors returned Global Imbalances: Past and Future 35 Figure 3.9. U.S. Gross Capital Inflows in Long-term Securities from Residents and Nonresidents, 1989­2010 500 400 300 US$, billions 200 gross inflows by foreigners 100 0 ­100 gross inflows by U.S. residents ­200 ­300 ­400 ­500 10 89 90 91 92 93 94 95 96 97 98 99 00 20 1 02 03 04 05 06 07 08 2009 0 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 year Source: U.S. Department of the Treasury, updated from De la Torre, Schmukler, and Servén 2009. Figure 3.10. Net Foreign Purchases of U.S. Long-term Securities and China's Net Foreign Asset Position, Quarters 1 and 3, 2000­10 500,000 40,000 35,000 400,000 30,000 300,000 25,000 yuan, millions US$, millions 20,000 200,000 15,000 100,000 10,000 0 5,000 0 ­100,000 ­5,000 ­200,000 ­10,000 0 1 02 03 04 05 06 07 08 09 10 0 0 20 20 20 20 20 20 20 20 20 20 20 1 1 1 1 1 1 1 1 1 1 1 Q Q Q Q Q Q Q Q Q Q Q year treasury agency corporate equities change in net foreign assets Source: Treasury International Capital System (database), U.S. Department of the Treasury, http://www.ustreas.gov/tic/ticsec.shtml; Statistics Database, People's Bank of China, http://www.pbc.gov.cn. to large-scale acquisition of U.S. Treasuries and, to a more large volumes of financial assets from more developed modest extent, equities. Interestingly, the capital inflow markets, particularly in the absence of deep--and, to date, cycle of 2008­09 was closely matched by the timing of unforeseen--reforms to speed up the development of China's accumulation of foreign assets. emerging countries' financial markets. This pattern of capital inflows suggests that savers from In addition, from a global perspective, the crisis has developing countries will very likely continue to demand underscored the effectiveness of the self-insurance strategy 36 Managing Openness pursued by emerging economies, as countries that had Vanishing Global Imbalances amassed large volumes of external assets managed to Although less likely, a sustained narrowing of global weather the global storm better than the rest. The crisis imbalances could result from several causes. First, new experience may encourage these and other countries to mechanisms for diversifying international risk (such as the hold even bigger stocks of liquid foreign assets in the contingent credit facility recently established by the IMF) future, especially because even at the height of the turmoil might begin to reduce the incentives for self-insurance. some emerging economies feared weakening the confi- Governments in emerging economies could then diversify dence of international investors and were thus reluctant to their portfolios away from low-yield short-term foreign use up their vast reserves (Aizenman and Sun 2009). assets and into more profitable investment opportunities. These global factors imply that the uphill pattern of In turn, this decline in emerging countries' reserve hold- capital flows is likely to persist. Indeed, recent forecasts of ings would help reduce systemic risk in the global financial the International Monetary Fund (IMF) of China's gap system. between saving and investment suggest little decline in the Second, a shift in portfolio diversification by interna- external surplus from its high precrisis levels (figure 3.11, tional savers away from U.S. assets would also contribute to panel a). In turn, from the U.S. perspective, a quick a narrowing of global imbalances. While savers have been rebound of the economy from the crisis could lead the way diversifying since the late 1990s, the rate of diversification to the recovery of world trade and commodity prices and could accelerate in the face of renewed doubts about the firm up the comeback of capital flows from emerging future performance of the U.S. economy and the dollar, or countries. Furthermore, the record U.S. public deficits if the recent turmoil in the U.S. financial system were to could well prevail over the rise in private saving prompted weaken the perceived appeal of U.S. assets. by the fall in asset prices and household net worth, halting Third, an early withdrawal of fiscal stimuli in the United (although perhaps not reversing) the decline of the current States and other advanced countries would also reduce global account deficit (figure 3.11, panel b). imbalances. Such a course of action, however, could also For small developing countries, this scenario would delay the world recovery and put in jeopardy the export-led come close to business as usual, at least for some time. In a growth strategy pursued by a number of emerging markets. context of rapid recovery, Asian emerging countries could Although China and other major developing countries have continue to pursue their export-led growth strategy based weathered the crisis reasonably well thus far, they have done on currency undervaluation, further fueling the return of so in the context of aggressive fiscal stimuli in advanced global imbalances. One potentially important difference, countries. A reversal of the fiscal expansions in rich countries, however, is that in the postcrisis world improved financial and a longer-lasting global slump, would imply a further regulation, as well as enhanced investor awareness, will slowdown in global demand for developing-country exports. likely bring to an end the underpricing of risk that charac- From the perspective of low-income countries, the terized the run-up to the global financial crisis. As a result, increase in import demand from middle-income economies the cost of capital, especially for developing countries, is could pick up some of the slack (see the chapters by Hanson likely to be higher than it was in the precrisis world, so that in this volume). Low-income countries might then be able the efficiency of investment will become a more pressing to offset, at least in part, the slowdown in demand from the concern from the perspective of growth. Figure 3.11. Saving and Investment Forecasts for China and the United States as a Percentage of GDP, 1991­2013 a. Forecast for China b. Forecast for the United States 70 25 60 investment saving 20 50 % of GDP % of GDP 40 15 saving 30 investment 10 20 5 10 0 0 91 93 95 97 99 01 03 05 07 09 11 13 91 93 95 97 99 01 03 05 07 09 11 13 19 19 19 19 19 20 20 20 20 20 20 20 19 19 19 19 19 20 20 20 20 20 20 20 year year Source: World Economic Outlook (database), IMF, http://www.imf.org/external/ns/cs.aspx?id=28. Note: The forecasts are shown in the shaded areas. Global Imbalances: Past and Future 37 advanced economies. However, the export-led growth strat- Another factor behind high corporate saving is the lack egy pursued by a number of middle-income countries, of financial options, such as a deep corporate bond market, notably in East Asia, would come under stress if global which encourages firms to retain their earnings to finance demand were to fall. future investments. In addition, repressed interest rates on alternative assets (bank deposits, for example) give firms an incentive to recycle their retained earnings into further Promoting Growth in a Postcrisis World investments, including those in marginally productive Developing countries could encourage growth through projects. Reforms to speed up the development of bond several policy measures that take global imbalances into and equity markets would reduce firms' incentives to retain account. First, surplus emerging markets could rebalance earnings and undertake possibly inefficient investments. their economies through an increased reliance on inward- A different approach to speeding growth in developing looking growth. Such a shift may have already started, as countries in the postcrisis world is to actively promote the suggested by China's massive fiscal stimulus implemented expansion of modern, high-value-added sectors through in 2009 with the stated objective of easing the economy's tax-cum-subsidy industrial policies (Rodrik 2009). In the- dependence on growth of world export markets.8 In the ory, these can be designed to replicate the effects of under- short run, the stimulus helped China maintain its high valued real exchange rates, that is, by encouraging the growth rate, but concerns have arisen recently about the growth of tradable goods industries, without the dampen- efficiency of the expenditures involved, as well as their pos- ing down of domestic demand and the accumulation of sible contribution to asset bubbles. foreign assets that so far have characterized neomercan- In general, an orderly rebalancing of emerging economies tilist policies (and further increased global imbalances). with strong external positions and sound macroeconomic The risk of such a policy, however, is that the govern- frameworks would also likely require an adjustment of ment, rather than the discipline of world markets, would exchange rates to reduce the accumulation of foreign assets dictate the sector composition of growth. Moreover, such a and encourage domestic demand (Blanchard and Milesi- policy strategy would run counter to existing World Trade Ferreti 2009). Policy measures would also be needed to Organization agreements. More generally, it is subject to attack entrenched distortions, such as those causing exces- the same caveats and controversies that have long sur- sively high saving rates (both in household and in corporate rounded the use of industrial policy. While its theoretical sectors) and those hampering development of the financial justification is clear, its practical implementation often system. Indeed, while high saving rates in China and other runs into insurmountable difficulties because governments emerging Asian countries reflect, in part, frugality and might not have sufficient information to identify the deep-rooted cultural values,9 they are also partly attributa- "right" products or industries and because institutional ble to weak social protection systems (Carroll and Jeanne weaknesses common in developing countries make such 2009). As such, the strengthening of social safety nets interventions prone to rent seeking and corruption13--and underway in China and other emerging markets will likely possibly even detrimental to the efficient allocation of reduce exorbitant household saving rates and promote con- resources, which is key to growth in developing economies. sumption.10 The process may take considerable time, but it should eventually allow a substantial increase in domestic Conclusion demand.11 In China, and to a lesser extent in other surplus emerg- The nature of global imbalances, their role in the world cri- ing markets, the distortions behind high corporate saving sis, and their likely path in the postcrisis world have taken should also be addressed. Large state subsidies, monopolis- center stage in the debate over the international economic tic conditions, and fast growth have all generated massive outlook. Although the crisis has temporarily reduced the profits for some Chinese sectors. Since the fiscal reform in magnitude of global imbalances, their future is far from 1994, however, most state-owned enterprises have not dis- certain. Absent radical national and global policy action, the tributed dividends to shareholders or the state, choosing structural distortions at the root of the imbalances are likely instead to retain their profits to finance potentially ineffi- to remain in place, suggesting that global imbalances will cient investment in the Chinese corporate sector (Kuijs return. In such a case, the global environment for develop- 2005, 2006). A requirement for state-owned enterprises to ing economies could show little change relative to the pre- distribute dividends could help channel saving into more crisis situation, although stricter prudential regulation of productive uses through financial markets or would per- the financial system postcrisis and a renewed perception of haps boost consumption.12 risk might lead to a persistent increase in the cost of capital 38 Managing Openness and make its efficient use an even higher priority than when such policy would take place ("China Sets Dividend Scale for Its State-Owned Firms" 2007). before, particularly for deficit developing countries. 13. In fact, undervaluation through reserve hoarding can be viewed It is also possible, however, although less likely, that other as a way to achieve the same objective, overcoming (at a cost) these obsta- forces triggered by the crisis would lead to a significant nar- cles; see Korinek and Servén (2010). rowing of global imbalances in the future. In that case, advanced countries' import demand growth could slow, Bibliography and thereby pose a challenge for developing countries-- particularly for surplus emerging markets that pursue Acharya, V., and P. Schnabl. 2009. "Do Global Banks Spread Global Imbal- ances? The Case of Asset-Backed Commercial Paper during the Finan- export-led growth. Rising South-South trade might fill part cial Crisis of 2007­09." NBER Working Paper 16079, National Bureau of the global demand gap, but not all. The remedy would be of Economic Research, Cambridge, MA. for surplus developing countries to increase their reliance Adrian, T., and M. Shin. 2009. "Money, Liquidity and Monetary Policy." American Economic Review 99: 600­605. on inward-looking growth. Policy responses should empha- Aizenman, J., and J. Lee. 2007. "International Reserves: Precautionary vs size reducing excessive savings and increasing domestic Mercantilist Views." Open Economies Review 18: 191­214. consumption and investment. Aizenman, J., and Y. Sun. 2009. "The Financial Crisis and Sizable Interna- tional Reserves Depletion: From `Fear of Floating' to the `Fear of Los- ing International Reserves'?" NBER Working Paper 15308, National Notes Bureau of Economic Research, Cambridge, MA. Bernanke, B. 2005. "The Global Saving Glut and the U.S. Current Account 1. For instance, since the fiscal reform in 1994, large enterprises have Deficit." Federal Reserve Board of Governors. http://www.federal the right to retain their profits. reserve.gov/boarddocs/speeches/2005/200503102/default.htm. 2. This is the so-called "Lucas paradox"; see Lucas (1990). Blanchard, O., F. Giavazzi, and F. Sa. 2005. "The U.S. Current Account and 3. Excluding the possibility of default and abstracting from capital the Dollar." Brookings Papers on Economic Activity, Brookings Insti- gains and losses on external assets and liabilities. tution, Washington, DC. 4. They argue that if assets are valued according to the income they Blanchard, O., and G. Milesi-Ferretti. 2009. "Global Imbalances: In Mid- generate, the value of U.S. assets held abroad is much larger than officially stream." IMF Staff Position Note, International Monetary Fund, measured. They term this discrepancy dark matter and attribute it to Washington, DC. intangible assets such as superior technology and organizational knowl- Caballero, R. 2010. "The Other Imbalance and the Financial Crisis." NBER edge embedded in foreign direct investment. Working Paper 15636, National Bureau of Economic Research, Cam- 5. For example, for the case of China, Cline and Williamson (2008) bridge, MA. survey existing estimates of the equilibrium value of the renminbi. Caballero, R., E. Farhi, and P. Gourinchas. 2008a. "An Equilibrium Model Only 1 of the 18 studies in their survey concludes that the renminbi is of Global Imbalances and Low Interest Rates." American Economic overvalued. On average, the estimates indicate substantial renminbi Review 98 (1): 358­393. undervaluation--on the order of 20 percent for the real effective ------. 2008b. "Financial Crash, Commodity Prices, and Global Imbal- exchange rate and 40 percent for the nominal bilateral renminbi-dollar ances." Brookings Papers on Economic Activity 1­55. exchange rate. Carroll, C., and O. Jeanne. 2009. "A Tractable Model of Precautionary 6. This of course is closely related to the explanations based on Reserves, Net Foreign Assets, or Sovereign Wealth Funds." NBER asset demand of Mendoza et al. (2009) and Carroll and Jeanne (2009) Working Paper 15228, National Bureau of Economic Research, Cam- summarized earlier. In both cases, the underlying force is the lack of ade- bridge, MA. quate insurance mechanisms. The main difference is that now the focus is "China Auto Sale Surge." 2009. Wall Street Journal. September 9. on country-level insurance against external shocks rather than on indi- "China Sets Dividend Scale for Its State-Owned Firms." 2007. Wall Street vidual insurance against idiosyncratic shocks. Journal. December 12. 7. Recent International Monetary Fund estimates, reflected in figure "China Wows to Improve Social Safety Net." 2010. Associated Press. 3.1 above, suggest that the U.S. current account deficit fell to less than 1 March 8. percent of world GDP in 2009, as opposed to 1.2 percent in 2008. Cline, W., and J. Williamson. 2008. "Estimates of the Equilibrium 8. However, recent signals from Chinese authorities do not offer Exchange Rate of the Renminbi: Is There a Consensus and If Not, Why much indication of an impending change in growth strategy, in spite of Not?" In Debating China's Exchange Rate Policy, ed. M. Goldstein and pressures from the IMF, the United States, and other advanced countries N. Lardy, 131­54. Washington, DC: Peterson Institute of International ("Wen Dismisses Currency Pressure" 2009). Economics. 9. For example, Wei and Zhang (2009) attribute the high household Curcuru, S., C. Thomas, and F. Warnock. 2008. "Current Account Sustain- saving rates in China to a rising sex ratio imbalance and an increasingly ability and Relative Reliability." NBER Working Paper 14295, National competitive marriage market in the country. Bureau of Economic Research, Cambridge, MA. 10. The top stated priorities of the 2010 Chinese People's Congress De la Torre, A., S. Schmukler, and L. Servén. 2009. "Back to Global Imbal- included establishing a safety net of pension, health care, and unemploy- ances?" http://www.roubini.com/globalmacro-monitor/257276/back_ ment benefits and providing free primary and secondary education to_global_imbalances. ("China Wows to Improve Social Safety Net" 2010). Engel, C., and J. Rogers. 2006. "The U.S. Current Account Deficit and the 11. In China, this process may be already underway. For example, car Expected Value of World Output." Journal of Monetary Economics 53: sales recently exceeded those in the United States for the first time, partly 1063­93. as a result of tax incentives to buy cars with small engines ("China Auto Forbes, K. 2010. "Why Do Foreigners Invest in the United States?" Journal Sale Surge" 2009). of International Economics 80: 3­21. 12. The Chinese government announced some time ago its intention Gourinchas, P., and H. Rey. 2007a. "From World Banker to World Venture to require state-owned enterprises to pay dividends but did not specify Capitalist: The US External Adjustment and the Exorbitant Privilege." Global Imbalances: Past and Future 39 In G7 Current Account Imbalances: Sustainability and Adjustment, ed. Laibson, D., and J. Mollerstrom. 2010. "Capital Flows, Consumption R. Clarida, 11-66. Chicago, IL: University of Chicago Press. Booms and Asset Bubbles: A Behavioural Alternative to the Savings ------. 2007b. "International Financial Adjustment." Journal of Political Glut Hypothesis." NBER Working Paper 15759. National Bureau of Economy 115 (4): 665­703. Economic Research, Cambridge, MA. Hausmann R., and F. Sturzenegger. 2004. "Global Imbalances or Bad Lucas, R. 1990. "Why Doesn't Capital Flow from Rich to Poor Countries?" Accounting? The Missing Dark Matter in the Wealth of Nations." Cen- American Economic Review 80: 92­96. ter for International Development Working Paper 124, Harvard Uni- Mendoza, E., V. Quadrini and J. Rios-Rull. 2009. "Financial Integration, versity, Cambridge, MA. Financial Deepness and Global Imbalances." Journal of Political Econ- Iossifov, P., M. Cihák, and A. ar Shanghavi. 2008. "Interest Rate Elasticity omy 117: 371­416. of Residential Housing Prices." IMF Working Paper WP/08/247, Inter- Nguyen, H. 2010. "Valuation Effects with Transitory and Trend Productivity national Monetary Fund, Washington, DC. Shocks." Research Working Paper 5174, World Bank, Washington, DC. Jeanne, O., and R. Ranciere. 2009. "The Optimal Level of International Obstfeld. M., and K. Rogoff. 2005. "The Unsustainable U.S. Current Account Reserves for Emerging Market Countries: A New Formula and Some Position, Revisited." http://elsa.berkeley.edu/~obstfeld/NBER_ final.pdf. Applications." http://www.econ.jhu.edu/People/Jeanne/JeanneRanciere_ ------. 2009. "Global Imbalances and the Financial Crisis: Products of a Feb09.pdf. Common Cause." Mimeo. Kashyap, A. Forthcoming. "Comment on `The Financial Crisis and Global Portes, R. 2009. "Global Imbalances." In Macroeconomic Stability and Policy Reforms' by Barry Eichengreen." In Asia and the Global Finan- Financial Regulation: Key Issues for the G20, ed. Mathias Dewatripont, cial Crisis. San Francisco: Federal Reserve Bank of San Francisco. Xavier Freixas, and Richard Portes, 19­26. London: Centre for Eco- Korinek, A., and L. Servén. 2010 "Undervaluation through Foreign nomic Policy Research. Reserve Accumulation: Static Losses, Dynamic Gains." Policy Research Prasad, E. 2009. "Rebalancing Growth in Asia." Discussion Paper 4298, Working Paper 5250, Washington, DC, World Bank. Institute for the Study of Labor (IZA), Bonn. Kraay, A., N. Loayza, L. Servén, and J. Ventura. 2005. "Country Portfolios." Rodrik, D. 2009. "Growth after the Crisis." Discussion Paper 7480, Centre Journal of the European Economic Association 3 (4): 914­45. for Economic Policy Research, London. Kraay, A., and J. Ventura. 2005. "The Dot-Com Bubble, the Bush Deficits, Roubini, N. 2009. "Will the Bretton Woods 2 (BW2) Regime Collapse Like and the US Current Account." NBER Working Paper 11543, National the Original Bretton Woods Regime Did? The Coming End Game of Bureau of Economic Research, Cambridge, MA. BW2." http://www.rgemonitor.com/. Kuijs, L. 2005. "Investment and Savings in China." Policy Research Work- Wei, S.-J., and X. Zhang. 2009. "The Competitive Saving Motive: Evidence ing Paper 3633, World Bank, Washington, DC. from Rising Sex Ratios and Savings Rates in China." NBER Working ------. L. 2006. "China in the Future: A Large Net Saver or Net Bor- Paper 15093, National Bureau of Economic Research, Cambridge, MA. rower?" Mimeo, World Bank. "Wen Dismisses Currency Pressure." 2009. Financial Times. December 27. 4 Rebalancing Trade after the Crisis Caroline Freund Global trade imbalances have surged since the early 1990s. levels. If capital flows do become more balanced, what will Figure 4.1 shows an index of global trade imbalances drive future trade growth is not clear. (GTIs)--the sum of the absolute values of trade balances The financial crisis has already shocked trade patterns, across countries--in log levels. The average annual growth leading to a reduction in real trade of 12.2 percent in 2009,1 of global imbalances was 11 percent from 1990 to 2007, rel- a magnitude not seen since the Great Depression. The ative to only 1 percent in the previous 20 years. In contrast collapse of trade can be explained in part by the global to global imbalances, global trade grew at a strong and imbalance view of the crisis. This view maintains that steady pace of about 6 percent a year over the whole period. because of a global savings glut, global interest rates were The growth of global imbalances has been a cause for con- severely depressed. The crisis caused interest rates to shoot cern since the new millennium, when the GTI approached a up. Rising interest rates, in turn, led to a sharp drop in level nearly twice the previous peak. As discussed in earlier investment and consumption in borrower countries. The chapters, some economists believe the surge in global imbal- effects on industrial production and consumer goods, and ances contributed to the onset and severity of the financial therefore on trade, were especially strong. As both exports crisis. For instance, Bernanke (2009) argues that imbalances and imports fell dramatically, so too did the gap between were an important cause of the crisis because they depressed them, immediately cutting imbalances to respectable levels. global interest rates, leading investors to search for higher The important question is whether the fall in trade and yields and to underprice risk. global imbalances is a short-run phenomenon or a struc- Obstfeld and Rogoff (2009) suggest that large global tural change brought about by the crisis.2 If it is a short- imbalances could also be a symptom of financial distress, run change, many of the same issues that plagued the without being a cause. They contend that policies pur- financial system in recent years are likely to reemerge. If it sued in Asian countries and the United States led to is a shift to more balanced flows, the world may be moving unstable financial conditions and global imbalances. As to a more stable financial system. In that case, however, time passed, ever-growing imbalances magnified finan- trade patterns will look very different in the future. cial instabilities. Whether global imbalances are a cause or In this chapter, we examine the extent to which the crisis a symptom, however, they are malignant and are associ- is responsible for rebalancing trade and whether that shift ated with an unstable financial system. A more stable is sustainable. Unlike other papers on current account financial system must therefore involve more balanced adjustment, we approach the question from the real side.3 capital flows. First, we investigate how trade balances have adjusted fol- A world with more balanced capital flows has impor- lowing the financial crisis. Specifically, we calculate how tant implications for bilateral trade and global trade much of the adjustment is a result of the drop in trade and growth. As capital has migrated to the United States, so how much is a result of rebalancing between export and have imports. With more balanced capital flows, the import growth. We argue that a shift due to the drop in United States could no longer be a rapidly growing market trade is likely to be reversed as global income expands but for the world's exports. And, with more balanced capital that rebalancing likely reflects shifts in attitudes toward flows, China could not maintain export growth at precrisis saving and investment. 41 42 Managing Openness Figure 4.1. Global Imbalances and Global Trade, 1970­2007 logarithms normalized to 1970 = 0 2.50 2.00 1.50 1970 = 0 1.00 log(Trade) 0.50 log(GTI) 0.00 ­0.50 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 year Source: Author's calculations and World Development Indicators (database), World Bank, http://data.worldbank.org. Note: Data are in constant U.S. dollars, for a balanced sample of 73 countries that make up about 85 percent of global trade. Data are in logs and normalized to start from zero. Second, we examine which countries are well posi- The Decline in Global Imbalances: tioned to drive trade growth and which countries are Rebalancing versus the Trade Collapse likely to benefit from changing demand patterns. Four Is an adjustment of global imbalances underway? And large emerging-market countries have fared remarkably what would that imply for the pattern of trade and world well in the crisis: Brazil, China, India, and Indonesia. In trade growth? Using aggregate trade data, we find strong these countries, imports in 2009 increased above precrisis evidence that imbalances are adjusting: they fell by 26 per- levels. Standard economic models imply that these rap- cent from 2007 to 2009.4 In this section, we analyze how idly growing emerging markets should be net importers much of the change in global imbalances during 2009 is of capital and goods; yet all but India have been running due to rebalancing and how much is a result of the collapse sizable trade surpluses in recent years. Conditions, how- in trade. ever, may have now changed so that more typical patterns may emerge: that is, imports may grow rapidly in these large emerging markets as they become the future of Average Change in Trade trade growth. The rest of this chapter is organized as follows. The next Indeed, 69 percent of countries experienced a smaller sur- section examines how much of the decline in global imbal- plus or smaller deficit following the crisis (table 4.1). In the ances can be explained by rebalancing and how much is typical country, more than two-thirds of the reduction was due to the fall in trade. The following section discusses the the result of rebalancing, whether the finding is based on a changing patterns of global demand and the sustainability country's average change in trade or on the global average of those trends. The final section concludes with a discus- change in trade. Moreover, significant rebalancing in coun- sion of the main risks that lie ahead. The presentation here tries such as the United States and China appears to be focuses on results and policy implications and is nontech- underway. This change is a positive sign, because as trade nical. For full details of data and methodology, readers are expands after the crisis, major shifts will be required to referred to the working paper version of this chapter. maintain global balances. Table 4.1. Change in Trade Imbalances due to Declines in Trade and Rebalancing Selected Economies, 2009 Economy average Global trade trade growth growth Change in net Change in Net exports/ Economy Trade drop Rebalancing Trade drop Rebalancing exports (US$, millions) net exports (%) GDP (2007) Algeria ­10.5 110.5 14.3 85.7 ­26,600 ­82 23.21 Australia ­7.2 107.2 16.6 83.4 11,600 ­70 ­2.01 Austria 1.4 98.6 1.1 98.9 ­5,564 ­1,044 0.14 Bolivia ­56.4 156.4 37.5 62.5 ­403 ­31 9.93 Bosnia and Herzegovina 58.1 41.9 90.0 10.0 724 ­13 ­37.12 Brazil ­0.7 100.7 30.6 69.4 ­15,300 ­38 3.08 Bulgaria 38.8 61.2 26.8 73.2 4,392 ­44 ­25.00 Canada 17.7 82.3 10.4 89.6 ­44,600 ­113 2.83 Cape Verde ­334.9 434.9 102.3 ­2.3 82 ­11 ­51.29 Chile 36.8 63.2 28.1 71.9 ­9,960 ­42 14.94 China ­7.5 107.5 47.5 52.5 ­64,600 ­25 7.71 Colombia ­4.4 104.4 12.1 87.9 2,804 ­97 ­1.38 Croatia 81.8 18.2 57.4 42.6 2,746 ­20 ­22.88 Cyprus 98.1 1.9 119.4 ­19.4 709 ­10 ­34.45 Dominican Republic 164.9 ­64.9 129.0 ­29.0 735 ­9 ­19.75 Ecuador ­0.6 100.6 2.9 97.1 ­1,724 ­403 0.93 El Salvador 39.8 60.2 43.6 56.4 1,270 ­27 ­23.64 Estonia 34.8 65.2 15.2 84.8 3,580 ­77 ­22.19 Faeroe Islands 11.3 88.7 12.7 87.3 248 ­92 ­ Finland 38.4 61.6 16.0 84.0 ­6,105 ­73 3.33 Germany 47.5 52.5 40.8 59.2 ­76,800 ­29 8.09 Honduras ­2.7 102.7 59.5 40.5 814 ­20 ­34.49 Hungary 0.4 99.6 0.3 99.7 5,758 ­4,089 ­0.10 Iceland 25.5 74.5 9.6 90.4 2,378 ­122 ­9.73 Indonesia ­31.9 131.9 23.3 76.7 ­19,900 ­50 9.21 Israel 26.3 73.7 24.1 75.9 4,951 ­49 ­6.00 Italy 37.5 62.5 22.5 77.5 6,024 ­52 ­0.55 Japan 21.3 78.7 17.0 83.0 ­63,200 ­69 2.09 Jordan ­18,149.9 18,249.9 26,152.3 ­26,052.3 4 0 ­53.21 Kazakhstan 807.7 ­707.7 828.5 ­728.5 ­212 ­1 15.00 Latvia 34.9 65.1 16.4 83.6 5,208 ­71 ­25.23 Lithuania 19.4 80.6 15.5 84.5 5,498 ­76 ­18.65 Luxembourg 379.4 ­279.4 231.1 ­131.1 310 ­5 ­12.24 (continued) 43 44 Table 4.1. (continued) Economy average Global trade trade growth growth Change in net Change in Net exports/ Economy Trade drop Rebalancing Trade drop Rebalancing exports (US$, millions) net exports (%) GDP (2007) Malta 289.1 ­189.1 143.1 ­43.1 145 ­8 ­23.90 Mexico 30.2 69.8 21.9 78.1 5,396 ­54 ­1.01 Moldova 77.6 22.4 95.5 4.5 348 ­12 ­64.60 Netherlands 62.2 37.8 79.6 20.4 ­8,130.3 ­15 7.09 New Zealand 14.7 85.3 14.0 86.0 3,274.1 ­83 ­3.02 Norway 190.3 ­90.3 166.7 ­66.7 ­3,925.9 ­7 14.33 Pakistan 8.4 91.6 151.1 ­51.1 1,189 ­8 ­11.00 Peru ­6.3 106.3 40.2 59.8 ­2,413 ­29 7.53 Philippines 297.4 ­197.4 150.7 ­50.7 392 ­8 ­3.61 Poland 14.1 85.9 22.2 77.8 13,500 ­53 ­5.98 Portugal 1,269.0 ­1,169.0 1,115.6 ­1,015.6 281 ­1 ­12.18 Romania 20.7 79.3 21.5 78.5 16,300 ­54 ­17.59 Russian Federation 96.7 3.3 79.3 20.7 ­19,300 ­15 10.08 Senegal ­71.9 171.9 61.8 38.2 630 ­19 ­30.23 Serbia 45.5 54.5 56.2 43.8 2,027 ­21 ­ Singapore 24.3 75.7 35.0 65.0 ­12,100 ­33 21.24 Slovak Republic 6.3 93.7 4.4 95.6 2,766 ­263 ­1.25 Slovenia 34.4 65.6 22.8 77.2 1,494 ­51 ­6.19 South Africa 18.4 81.6 14.6 85.4 7,796 ­80 ­3.47 Spain 36.8 63.2 24.1 75.9 66,000 ­49 ­9.71 Sweden 80.2 19.8 42.4 57.6 ­4,280 ­28 3.44 Tanzania ­302.7 402.7 114.1 ­14.1 406 ­10 ­23.26 Turkey 28.7 71.3 30.6 69.4 24,000 ­38 ­9.66 United Kingdom 71.6 28.4 39.9 60.1 53,200 ­29 ­6.46 United States 40.0 60.0 31.9 68.1 290,000 ­37 ­5.65 Zambia 40.1 59.9 78.5 21.5 ­91 ­15 5.55 Median 28.7 71.3 31.9 68.1 629.6 ­36.7 ­3.6 Economy average Global trade trade growth growth Imbalance-widening Change in net Change in Net exports/ economy Trade drop Diverging growth Trade drop Diverging growth exports (US$, millions) net exports (%) GDP (2007) Albania 42.6 57.4 ­106.9 206.9 ­341 11 ­28.35 Argentina ­13.9 113.9 ­23.5 123.5 5,614 50 4.35 Armenia ­83.0 183.0 ­78.5 178.5 ­333 15 ­24.30 Azerbaijan 3.1 96.9 ­0.5 100.5 8,224 2,376 1.05 Belarus ­9.9 109.9 ­18.1 118.1 ­2,863 65 ­9.82 Belgium ­352.7 452.7 ­284.4 384.4 718 4 3.80 Czech Republic ­11.4 111.4 ­13.5 113.5 3,716 86 2.53 Denmark ­9.8 109.8 ­9.2 109.2 5,685 128 1.44 Egypt, Arab Rep. 53.4 46.6 ­11.5 111.5 ­11,000 101 ­8.38 Ethiopia 83.3 16.7 ­12.5 112.5 ­4,238 94 ­23.85 France ­350.4 450.4 ­322.8 422.8 ­2,116 4 ­2.25 French Polynesia ­19.7 119.7 ­121.6 221.6 ­138 10 ­ Hong Kong SAR, China ­28.3 128.3 ­50.6 150.6 ­5,426 23 ­11.19 India 41.9 58.1 ­81.5 181.5 ­10,000 14 ­5.81 Ireland ­33.0 133.0 ­23.2 123.2 18,000 50 13.73 Korea, Rep. ­3.3 103.3 ­6.6 106.6 25,800 176 1.46 Macao SAR, China ­299.3 399.3 ­76.3 176.3 ­565 15 ­19.39 Macedonia, FYR ­45.5 145.5 ­45.6 145.6 ­480 26 ­23.69 Malaysia ­100.0 200.0 ­91.1 191.1 3,825 13 15.68 Mauritius ­154.8 254.8 ­203.1 303.1 ­96 6 ­22.30 Mozambique 4.1 95.9 ­7.6 107.6 ­979 154 ­7.97 Paraguay 78.2 21.8 ­53.5 153.5 ­677 22 ­25.80 Switzerland ­2.3 102.3 ­19.9 119.9 6,348 59 2.51 Taiwan, China ­246.5 346.5 ­153.0 253.0 2,044 8 6.79 Thailand ­7.7 107.7 ­33.9 133.9 4,796 34 5.56 Uruguay 44.7 55.3 ­24.6 124.6 ­369 48 ­3.23 Zimbabwe ­3.4 103.4 ­2.8 102.8 ­1,092 425 ­ Median ­9.9 109.9 ­33.9 133.9 ­333.3 34.5 ­3.2 Median full sample 18.0 82.0 16.8 83.2 370.2 ­14.8 ­3.5 Source: Author's calculations based on the following databases: Datastream, Thomson Reuters, http://www.datastream.com; World Integrated Trade Solution (WITS), World Bank, http://wits.worldbank.org; World Development Indicators, World Bank, http://data.worldbank.org; National Statistics, Taiwan, China, http://eng.stat.gov.tw. 45 46 Managing Openness For the remaining 31 percent of countries, imbalances that imbalance-augmenting trade growth contributed to swelled. In the typical country, the trade drop caused expanding the GTI by 20 percent.6 imbalances to shrink by 10­34 percent, but this decline was Similar effects from rebalancing are also found using more than offset by imbalance-expanding trade growth. growth in country trade. In this case, however, the decline The trade drop appears to have affected imbalances to a due to the trade drop is smaller (39.6 percent) because it takes lesser extent when we employ country-specific trade into account that trade grew or fell by less than the global growth rates. This is logical, however, because the method- average in some countries. Expansion due to imbalance- ology takes into account the fact that trade did not fall very augmenting growth is also smaller for this reason. much and even expanded in some of the countries with This phenomenon does not apply only to small coun- larger imbalances. Since the trade drop is smaller or absent, tries. Three quarters of the decline in the global trade it has a smaller effect on imbalances. imbalance, which puts more weight on larger imbalances, What does this breakdown imply for the sustainability stems from rebalancing. In sum, the trade drop was not the of the adjustment in global imbalances? The contraction in main reason behind the improvement in trade balances in imbalances due to rebalancing is likely to be sustainable 2009; instead, countries tended to rebalance trade flows.7 because it reflects changes in the underlying fundamentals, savings and investment. Rebalancing occurs when coun- What Happened in the Asian Financial Crisis? tries with large trade deficits have reduced their imports to a far greater extent than their exports and vice versa. For a better understanding of whether the adjustment of In contrast, the reduction of imbalances due to the global imbalances is sustainable over the medium to long sharp drop in trade is unlikely to be sustainable because a term, we review the pattern of adjustment in the Asian cri- large drop in trade (such as the one that occurred in 2009) sis countries. The Asian financial crisis revealed a pattern causes trade imbalances to retreat only if it affects exports similar to the recent crisis, with over 70 percent of coun- and imports proportionately. If both imports and exports tries showing greatly reduced imbalances immediately fol- decline by a given percentage, the larger flow must decline lowing the crisis (table 4.3). Four countries (the Republic by absolutely more and the difference will therefore also of Korea, Malaysia, Singapore, and Thailand) show com- shrink. Thus, it follows that a reversal of the trade collapse plete reversals, moving from a deficit to a large surplus. The would lead the proportion of the imbalance driven by adjustment in net exports was almost entirely due to rebal- trade to reemerge. ancing (92 percent), partly because the drop in trade was more moderate in the Asian crisis (9 percent on average) relative to the current event (15 percent). The Global Trade Imbalance What is especially striking is that the improvements in This section examines how the GTI was affected. Looking net exports tended to be sustained in the years following at the global trade balance, as opposed to the average the crisis. Indeed, the Asian case shows that rebalancing as change, puts more weight on large countries. It also better a result of major crises can be sustainable. In five out of reflects the magnitude of the "global savings glut." Based on the six economies with sizable deficits before the crisis our sample, the GTI fell by 26 percent over the period (Hong Kong SAR, China; Korea; Singapore; Sri Lanka; and 2007­09. Given that aggregate trade fell by about 11 per- Thailand), the trade balance remained positive or contin- cent, the trade drop contributed to about 42 percent of the ued improving over the next five years. Only the Philippines decline in GTI. Moreover, rebalancing contributed 78 per- experienced a widening of the deficit after 1998, although it cent of the reduction in GTI (table 4.2).5 Finally, we find did not approach precrisis levels. Table 4.2. Change in Global Trade Imbalance, 2007­09 Imbalance-augmenting Rebalancing % Change in GTI Trade drop Rebalancing growth Global average % change in trade 26 39.6 78.1 ­17.7 Country average % change in trade 26 42.5 78.2 ­20.7 Source: Author's calculations based on the following databases: Datastream, Thomson Reuters, http://www.datastream.com; WITS, World Bank, http://wits.worldbank.org; World Development Indicators, World Bank, http://data.worldbank.org; National Statistics, Taiwan, China, http://eng.stat.gov.tw. Table 4.3. Global Trade Balance Adjustment in Asian Crisis Economies, 1996­98 Economy average Global trade trade growth growth Change in net Change in exports net exports (%) Net exports/ Net exports/ Net exports/ Economy Trade drop Rebalancing Trade drop Rebalancing (US$, millions) 1996­98 GDP (1996) GDP (1998) GDP (2003) Balancing Malaysia 0.2 99.8 0.0 100.0 14,800 ­13,566 ­0.11 20.42 18.18 Korea, Rep. 7.5 92.5 0.2 99.8 63,900 ­282 ­4.05 11.77 2.19 Singapore 10.8 89.2 0.3 99.7 12,400 ­178 ­7.51 6.63 17.1 Thailand 14.6 85.4 0.3 99.7 30,600 ­171 ­9.84 11.64 3.34 Philippines ­14.5 114.5 0.6 99.4 12,400 ­95 ­15.69 ­1.03 ­4.96 Taiwan, China 9.2 90.8 1.1 98.9 ­8,250 ­53 5.42 2.69 6.69 Hong Kong SAR, China 20.9 79.1 1.3 98.7 8,360 ­43 ­12.33 ­6.59 ­4.99 Sri Lanka ­50.5 150.5 3.8 96.2 211 ­15 ­10.36 ­7.69 ­7.58 Median balancing 8.3 91.7 0.5 99.5 12,400.0 ­133.1 ­8.7 4.7 2.8 Imbalance worsening India 4.4 95.6 ­0.4 100.4 ­6,040 152 ­1.02 ­2.38 ­2.12 Indonesia ­11.8 111.8 ­0.3 100.3 14,800 187 3.49 24.0 11.61 China 2.9 97.1 ­0.2 100.2 32,500 244 1.55 4.58 1.49 Median imbalancing ­1.5 101.5 ­0.3 100.3 13,753.3 194.3 1.3 8.7 3.7 Median full sample 4.4 95.6 0.3 99.7 12,400.0 ­52.7 ­4.1 4.6 2.2 Source: Author's calculations based on the following databases: Datastream, Thomson Reuters, http://www.datastream.com; WITS, World Bank, http://wits.worldbank.org; World Development Indicators, World Bank, http://data.worldbank.org; National Statistics, Taiwan, China, http://eng.stat.gov.tw. 47 48 Managing Openness Rebalancing across Countries correlation between fiscal stimulus and import growth is and Future Trade Growth small and insignificant, perhaps because most packages are geared toward domestic job expansion or because packages Which countries are well positioned to drive trade growth in are slow to have an impact on import growth. Together with the future? While imports collapsed in most countries in the rebalancing of imports and exports that has occurred, 2009, four growing emerging markets (GEMs) bucked the the continuation of import growth in these countries is a trend. Indeed, Brazil, China, India, and Indonesia recorded very positive sign. Indeed, these large emerging markets positive import growth from 2007 to 2009. This section may be the future source of growth in trade. examines the recent emergence of the GEMs as drivers of trade growth and how this trend will affect exports from the rest of the world. First, the section discusses how rebalancing Micro Effects can occur and the role of these four countries in future trade Import growth over the past 10 years has been robust, par- growth. Then the focus shifts to a micro perspective and iden- ticularly for China and India. These two countries report tifies the sectors and countries that will benefit from demand nominal dollar export growth of over 500 percent. Figure 4.3 growth in the GEMs. And finally, the section discusses how presents the sectors that are primarily responsible for the much of import growth in these countries is at the intensive recent growth of imports in the GEMs. One of the most (old products) versus the extensive (new products) margin. notable trends emerging in this figure is the sharp increase in imported fuels in all four countries. Rebalancing and Import Growth in the GEMs Different patterns emerge among industrial goods, how- Rebalancing can occur in three different ways: (a) both ever. Overall, increases are recorded in machinery and trans- imports and exports decline, and the larger flow decreases port and in miscellaneous manufactures, which include by relatively more than the smaller flow; (b) the larger flow items such as prefabricated buildings, furniture, travel declines and the smaller flow increases; or (c) both flows goods, clothing, footwear, and professional and scientific expand, and the larger flow expands by relatively less than equipment. In particular, Brazil shows a jump in the imports the smaller flow. These three modes have different impacts of chemical products, China and Indonesia in miscellaneous on global trade growth. The first way suggests that trade manufactures, and India in machinery and transport. growth may stagnate in the near term as trade flows adjust, When we examine exports to the GEMs as a share of while the third way suggests positive trade growth and total exports, interesting trends emerge among exporters declining imbalances. at the regional and income group level. For regional trends, From 2007 to 2009, rebalancing occurred through the the share of exports to the GEMs has increased most in first mode in most countries: flows decreased and trade Sub-Saharan African countries, followed by South Asia imbalances declined by relatively more because the bigger (figure 4.4). With respect to income groupings, low-income flow fell by a larger amount (table 4.4). However, in some countries, which have many of the natural resources countries, an increase in exports, imports, or both occurred needed for growth,8 have seen an increasing share of their following the crisis. For instance, imports increased in exports going to the GEMs, as have the high-income coun- Brazil and China. Moreover, exports and imports both tries, which have capital goods. Middle-income countries, increased in India and Indonesia. These are four of the five especially in Europe and Central Asia, have seen the smallest largest countries in the world by population, making up gains. over 40 percent of the total. Collectively, they accounted for 15 percent of the GTI in 2009. In theory, such fast-growing Intensive and Extensive Margins countries should be running external deficits and import- ing heavily for future growth. Finally, consideration should be given to whether import To illustrate how these countries diverged from the rest growth in the GEMs is occurring at the intensive margin or of the world, figure 4.2 shows imports on a log scale and at the extensive margin. Recent trade literature emphasizes adjusted to begin at the same point in 2006. These four the gains from trade as importing countries access new countries have seen a much stronger bounce back in product varieties (that is, import growth at the extensive imports than that observed in the rest of the world. margin). For example, Broda and Weinstein (2006) find Moreover, while the surge in imports could be a tempo- that 30 percent of U.S. import growth between 1972 and rary feature due to fiscal stimulus, this does not appear to be 2001 was in new varieties (the extensive margin). In addi- the case. China recorded a large fiscal stimulus in 2009, but tion, growth at the extensive margin would also suggest an Brazil, India, and Indonesia did not. Furthermore, a simple additional welfare gain among the GEMs. Table 4.4. Import and Export Growth in Deficit and Surplus Economies, 2007­09 percent Net exports/ Change in Change in Surplus Net exports/ Change in Change in Deficit economy GDP (2007) exports (%) imports (%) economy GDP (2007) exports (%) imports (%) Iceland ­9.59 ­15.78 ­46.60 Ireland 13.67 ­5.18 ­28.11 Latvia ­25.43 ­10.20 ­39.49 Finland 3.39 ­30.35 ­25.96 Estonia ­21.73 ­17.98 ­35.49 Faeroe Islands 0.00 2.08 ­22.88 Lithuania ­18.71 ­4.01 ­25.34 Sweden 3.42 ­22.42 ­21.89 Spain ­9.44 ­11.32 ­24.49 Taiwan, China 6.79 ­17.34 ­20.39 Romania ­17.69 0.40 ­22.92 Denmark 1.44 ­9.36 ­15.64 Bulgaria ­25.40 ­11.24 ­22.65 Canada 2.77 ­24.52 ­15.31 United Kingdom ­6.47 ­19.58 ­22.43 Malaysia 15.94 ­10.47 ­15.23 Philippines ­3.50 ­23.84 ­22.38 Norway 14.39 ­11.73 ­15.01 United States ­5.76 ­9.10 ­20.26 Belgium 3.81 ­14.13 ­14.90 Malta ­23.74 ­27.05 ­20.25 Russian Federation 10.11 ­14.39 ­14.18 Italy ­0.55 ­19.17 ­19.92 Serbia 0.00 ­5.44 ­13.51 Slovenia ­6.17 ­15.91 ­19.41 Kazakhstan 14.30 ­9.55 ­13.27 South Africa ­3.42 ­10.53 ­19.04 Argentina 4.30 ­0.56 ­13.26 Slovak Republic ­1.25 ­14.61 ­18.57 Austria 0.14 ­16.09 ­12.58 Hungary ­0.10 ­12.31 ­18.36 Germany 8.07 ­15.34 ­11.94 Croatia ­23.00 ­15.35 ­17.98 Czech Republic 2.47 ­8.16 ­11.60 Turkey ­9.69 ­4.79 ­17.13 Japan 2.10 ­18.42 ­10.96 New Zealand ­2.92 ­7.42 ­17.08 Chile 14.61 ­20.94 ­9.71 Luxembourg ­12.28 ­21.47 ­16.97 Korea, Rep. 1.40 ­2.14 ­9.46 Mexico ­0.99 ­15.51 ­16.87 Netherlands 7.12 ­9.49 ­8.81 El Salvador ­23.21 ­4.69 ­16.72 Singapore 21.62 ­9.75 ­6.51 Israel ­6.10 ­9.23 ­16.38 Zambia 5.35 ­6.61 ­5.35 Macao SAR, China ­19.80 ­77.88 ­13.89 Thailand 5.63 ­0.89 ­4.40 Dominican Republic ­19.59 ­17.41 ­12.51 Switzerland 2.53 0.59 ­3.32 (continued) 49 50 Table 4.4. (continued) Net exports/ Change in Change in Surplus Net exports/ Change in Change in Deficit economy GDP (2007) exports (%) imports (%) economy GDP (2007) exports (%) imports(%) France ­2.25 ­13.52 ­11.89 Mayotte 0.00 ­13.22 0.51 Moldova ­64.56 ­7.78 ­11.23 Zimbabwe 0.00 ­31.60 2.47 Poland ­6.04 ­3.61 ­11.20 China 7.76 ­1.34 5.05 Portugal ­11.98 ­15.85 ­10.78 Brazil 3.00 ­5.22 5.75 Cyprus ­33.73 ­9.48 ­9.75 Azerbaijan 1.05 142.46 7.12 Bosnia and Herzegovina ­36.57 ­5.36 ­9.74 Peru 7.71 ­3.58 7.23 Cape Verde ­52.87 85.62 ­8.95 French Polynesia 0.00 ­11.24 7.44 Honduras ­34.53 9.90 ­8.83 Ecuador 0.93 ­3.66 8.63 Hong Kong SAR, China ­11.34 ­7.52 ­5.57 Bolivia 9.84 10.06 25.19 Mauritius ­22.21 ­12.93 ­4.92 Indonesia 9.17 2.11 30.02 Macedonia, FYR ­23.61 ­19.81 ­3.53 Algeria 23.91 ­24.88 42.08 Senegal ­29.37 30.47 ­3.25 Median 4.06 ­9.52 ­10.34 Pakistan ­10.72 1.53 ­2.84 Belarus ­9.76 ­12.33 ­0.45 Colombia ­1.40 8.94 ­0.38 Australia ­2.01 9.26 0.93 Armenia ­24.29 ­28.15 3.41 Jordan ­53.20 12.23 4.02 India ­5.92 4.42 7.62 Albania ­28.79 0.95 8.38 Tanzania ­23.50 51.78 10.34 Paraguay ­25.33 15.37 18.81 Uruguay ­3.19 19.21 23.37 Mozambique ­7.93 ­10.98 23.43 Egypt, Arab Rep. ­8.33 42.35 66.07 Ethiopia ­23.64 67.99 87.91 Median ­11.98 ­9.48 ­11.89 Source: Author's calculations based on the following databases: Datastream, Thomson Reuters, http://www.datastream.com; WITS, World Bank, http://wits.worldbank.org; World Development Indicators, World Bank, http://data.worldbank.org; National Statistics, Taiwan, China, http://eng.stat.gov.tw. Note: Countries are sorted by import growth. fo cr od m ud in e an percent er al m be d at ve liv fu e 0 5 10 15 20 25 30 35 40 er ra el s, ia ge an an lu ls, sa im im in al al br ic ed nd s an an ib to d ts le ba m ve ,a ,e cc xc o logarithms, 2006 = 0 an ch ge nd ep uf em ta re tf ac bl la tu ic e te ue re al oi d ls 2006 = 0 d sa nd ls, m go fa at 1.2 1 0.8 0.6 0.4 0.2 0 ­0.2 re ts er od s la ,a i 20 m te nd als 06 ac clas d w hi s pr ax ne ifie od ry d uc es m an ch ts isc d ie n. el e. la tra fly b s. ns y co neo po m m a. All sectors us rt at m eq er od ma ia iti nu ui l 20 es fa pm 07 an ct en d ur t Brazil tra ed Source: Datastream, Thomson Reuters, http://www.datastream.com. ns ar ac tic le Indonesia tio s ns ch n. em e. m ic s Source: World Integrated Trade Solution, World Bank, http://wits.worldbank.org. an al 20 uf sa 1998 ac nd 08 tu year re re d la te percent go d od pr China 0 10 20 30 40 50 60 sc od 2008 la uc ss ts rest of world ifi n. ed m ch e. ac s. hi ie 20 ne fly 09 ry by an m d at m tra er ia l India isc ns el po la rt ne eq ou Figure 4.3. The Changing Nature of Imports in Brazil, China, India, and Indonesia, 1998 and 2008 ui sm pm co an en 20 m uf t m ac 10 od tu Figure 4.2. Growth of Imports in Brazil, China, India, and Indonesia and in the Rest of the World, 2006­10 iti re es d an ar d tic b. Industrial sectors le tra s ns ac tio ns n. Rebalancing Trade after the Crisis e. s. 51 52 Managing Openness Other authors highlight that export growth at the extensive Overall, import growth in the GEMs has been largely at the margin may be an important indicator of successful intensive margin (table 4.5). However, some growth has growth. Indeed, a strong positive correlation has been taken place at the extensive margin, especially in Indonesia. found between the number of export varieties a country Thirty percent of Indonesia's imports in 2007­08 were produces and its living standard (Funke and Ruhwedel from new exporters in that product. 2001; Hummels and Klenow 2005). These findings suggest For all the GEMs, low- and middle-income countries that future growth of exporting countries could look very have the highest extensive share. High-income countries, different if import growth in GEMs is largely at the exten- especially countries in the Organisation for Economic Co- sive margin. operation and Development, tend to have lower extensive Figure 4.4. Share of Exports by Region and Income Group to Brazil, China, India, and Indonesia, 1997­98 and 2007­08 a. Share of exports by region b. Share of exports by income 16.00 14.00 14.00 12.00 12.00 10.00 10.00 percent percent 8.00 8.00 6.00 6.00 4.00 4.00 2.00 2.00 0.00 0.00 e m e m e EC e, EC e, ci d As ral be nd Af and ia Af aran m co dl co dl Pa an O m O m As e e D D nt in id in id rib a a fic ia an a a co co n- co ric ric h ia th st -m -m Ce h in Ca ic Sa -in no in ut or a As e er N eE er er w d So b- gh gh st th Am an w p lo Su dl up Ea hi lo hi id pe tin M ro La Eu region income group 1997­98 2007­08 Source: WITS, World Bank, http://wits.worldbank.org; World Development Indicators, World Bank, http://data.worldbank.org. Note: OECD = Organisation for Economic Co-operation and Development. Table 4.5. Intensive and Extensive Margin of Total Imports for Brazil, China, India, and Indonesia by Income Group, 2007­09 Brazil Import Import Income growth Extensive Intensive growth Extensive Intensive level (US$, millions) (%) (%) Region (US$, millions) (%) (%) All 141.03 17.91 81.91 East Asia and Pacific 709.61 12.35 87.86 High income, Europe and Central non-OECD 248.82 33.63 66.20 Asia 586.50 44.13 55.70 High income, Latin America and OECD 70.27 6.25 93.65 the Caribbean 100.00 21.27 78.70 Low income Middle East and 601.83 24.02 75.79 North Africa 905.39 83.76 15.98 Lower-middle income 873.21 30.68 69.32 South Asia 904.98 61.48 38.52 Upper-middle income 107.76 21.81 78.19 Sub-Saharan Africa 656.63 10.38 89.27 (continued) Rebalancing Trade after the Crisis 53 Table 4.5. (continued) China Import Import Income growth Extensive Intensive growth Extensive Intensive level (US$, millions) (%) (%) Region (US$, millions) (%) (%) All 592.31 8.41 91.53 East Asia and Pacific 1,058.54 9.47 90.53 High income, non-OECD 504.55 8.57 91.73 Europe and Central Asia 694.30 29.42 70.58 High income, Latin America and OECD 477.69 4.19 95.87 the Caribbean 1,786.41 18.76 81.40 Low income Middle East and 1,189.67 28.21 71.66 North Africa 1,289.81 11.19 88.84 Lower-middle income 1,410.49 13.10 86.57 South Asia 1,316.14 10.41 89.70 Upper-middle income 1,147.75 15.09 84.84 Sub-Saharan Africa 2,440.96 22.48 77.52 India Import Import Income growth Extensive Intensive growth Extensive Intensive level (US$, millions) (%) (%) Region (US$, millions) (%) (%) All 561.95 15.67 84.27 East Asia and Pacific 1,065.02 16.21 83.79 High income, non-OECD 803.70 7.22 92.62 Europe and Central Asia 705.04 46.17 53.51 High income, OECD Latin America and 404.59 4.78 95.45 the Caribbean 1,154.62 56.58 43.16 Low income Middle East and 655.15 25.29 74.82 North Africa 1,490.56 19.13 81.05 Lower-middle income 1,476.42 17.12 83.20 South Asia 470.20 47.21 52.79 Upper-middle income 612.71 38.62 61.19 Sub-Saharan Africa 607.96 22.18 77.78 Indonesia Import Import Income growth Extensive Intensive growth Extensive Intensive level (US$, millions) (%) (%) Region (US$, millions) (%) (%) All 210.45 30.30 69.85 East Asia and Pacific 721.80 26.17 74.02 High income, non-OECD 944.30 51.35 48.84 Europe and Central Asia 701.56 53.55 46.45 High income, Latin America and OECD 77.27 11.33 88.64 the Caribbean 191.37 34.94 65.06 Low income Middle East and 247.78 43.51 56.49 North Africa 160.71 36.55 63.45 Lower-middle income 637.74 16.76 83.24 South Asia 273.15 43.81 56.19 Upper-middle income 537.38 48.79 51.21 Sub-Saharan Africa 193.37 46.98 53.02 Source: Author's calculations based on data from WITS, World Bank, http://wits.worldbank.org; World Development Indicators, World Bank, http://data.worldbank.org. Note: OECD = Organisation for Economic Co-operation and Development. shares. In China, where the extensive margin was less than 10 Conclusion percent of imports in 2007­08, nearly 30 percent of imports In the coming years, global imbalances must be limited to from low-income countries were goods that were not previ- ensure financial stability, with important implications for ously imported from those countries. Thus, for developing trade patterns and trade growth. More balanced capital countries to take advantage of demand in the GEMs, they flows imply that trade flows must also be more balanced will need to continue finding new products to export. 54 Managing Openness and that the primary source of global demand can no 5. Change due to rebalancing is calculated as the change in the GTI due to rebalancing in countries where the imbalance shrank. longer be the U.S. market. 6. Change due to imbalance-augmenting trade growth is calculated as A new world trade order is already emerging, with large the change in GTI not due to the trade drop in the countries where imbal- and growing emerging markets absorbing capital and ances expanded. goods from the rest of the world. In particular, Brazil, 7. This is consistent with Blanchard and Milessi-Ferretti (2009). They examine global imbalances during the crisis from a macro perspective and China, India, and Indonesia have demonstrated resilient find that significant adjustment in savings and investment patterns have global demand growth throughout the financial crisis, occurred. despite sizable exchange-rate depreciation in some coun- 8. See the chapter by Mitchell and Aldaz-Carroll for further discussion of the future of commodity markets. tries when the crisis began (for instance, in Brazil and Indonesia), and without extensive fiscal support. Financial stability and continued trade growth rely, in Bibliography part, on the sustainability of this diversification of global Baldwin, R., and D. Taglioni. 2009. "The Illusion of Improving Global demand. Fast-growing economies such as the GEMs need Imbalances." VoxEU. November. http://www.voxeu.org/index.php?q= to continue pulling capital to higher yields and importing node/4209. Bernanke, B. 2009. "Financial Reform to Address Systemic Risk." Speech at raw materials and machinery to ensure future growth. the Council on Foreign Relations, Washington, DC, March 10. http:// These requirements will tend to support the external sector www.federalreserve.gov/newsevents/speech/bernanke20090310a .htm. in low-income countries, especially those rich in natural Blanchard, O., and G. M. Milesi-Ferretti. 2009. "Global Imbalances: In resources, and high-income countries with strong machin- Midstream?" IMF Staff Position Note SPN/09/29, International Mon- etary Fund, Washington, DC. ery sectors or natural resources. Broda, C., and D. Weinstein. 2006. "Globalization and the Gains from While the adjustment process from a trade perspective Variety." Quarterly Journal of Economics 121 (2): 541­85. appears to be moving in a positive direction, serious risks Colacelli, M. 2006. "Export Responses to Real Exchange Rate Fluctuations: An Empirical Analysis." Unpublished paper, Department of Econom- remain. Among these are a return to low savings in the United ics, Columbia University, New York. States and a reemergence of large imbalances. In addition, Evenett, S., and J. Francois. 2010. "Will Chinese Revaluation Create China's import strength may be temporary, as relatively cheap American Jobs." VoxEU. http://www.voxeu.org/index.php?q=node/4931. Fang, W., Y. Lai, and S. Miller. 2006. "Export Promotion through Exchange natural resources are purchased for future use and domestic Rate Changes: Exchange Rate Depreciation or Stabilization?" Southern consumption has tended not to expand in a sustainable way. Economic Journal 72 (3): 611­26. Without the participation of these two countries, a more sta- Freund, C. 2009. "The Trade Response to Global Downturns: Historical ble order cannot take hold. In Europe, continued financial Evidence." Working Paper 5015, World Bank, Washington, DC. Funke, M., and R. Ruhwedel. 2001. "Product Variety and Economic problems, coupled with localized trade patterns, could lead to Growth: Empirical Evidence for the OECD Countries." IMF Staff a much slower trade recovery. These countries will do well to Papers 48 (2): 225­42. look toward Asia over the coming years. Hummels, D., and P. Klenow. 2005. "The Variety and Quality of a Nation's Exports." American Economic Review 95: 704­23. Obstfeld, M. 2002. "Exchange Rates and Adjustment: Perspectives from Notes the New Open Economy Macroeconomics." NBER Working Paper 9118, National Bureau of Economic Research, Cambridge, MA. 1. See World Trade Organization (2010). Obstfeld, M., and K. Rogoff. 2009. "Global Imbalances and the Financial 2. Baldwin and Taglioni (2009) argue that a strong recovery has fol- Crisis: Products of Common Causes." Paper prepared for the Federal lowed the sharp drop in trade and that it is likely to be accompanied by Reserve Bank of San Francisco "Asia Economic Policy Conference," the same worrisome global imbalances that defined previous years. Santa Barbara, CA, October 18­20. http://elsa.berkeley.edu/~obstfeld/ 3. For example, Blanchard and Milesi-Ferretti (2009) examine how santabarbara.pdf. imbalances have adjusted based on changes in savings and investment Prasad, E., and I. Sorkin. 2009. "Understanding the G-20 Economic Stim- patterns across countries. ulus Plans." Brookings Institution, Washington, DC. http://www 4. The sample in this chapter includes 86 countries through 2009, .brookings.edu/articles/2009/03_g20_stimulus_prasad.aspx. which together account for over 85 percent of world trade. Data are World Trade Organization. 2010. "Trade to Expand by 9.5% in 2010 gathered from various sources and are measured in nominal U.S. after a Dismal 2009, WTO Reports." March 26. News release. dollars. http://www.wto.org/english/news_e/pres10_e/pr598_e.htm. 5 An Anatomy of Trade in the 2008­09 Crisis Mona Haddad, Ann Harrison, and Catherine Hausman As discussed in the previous chapters, the global economic constraints in explaining some--but obviously not the crisis of 2008­09 was accompanied by a severe fall in inter- majority--of the collapse in world trade. national trade. Baldwin (2009) characterizes the collapse as The rest of this chapter is organized as follows. The next "sudden, severe, and synchronized . . . the sharpest in section discusses recent changes in the extensive and inten- recorded history and deepest since WWII." sive margin. The following section presents evidence of the A number of hypotheses have attempted to explain role of demand and supply shocks in the onset of the trade what caused the collapse and why it became so wide- collapse. The subsequent sections analyze how these trends spread and deep.1 While many highlight the fall in aggre- vary by product type and income group. The following sec- gate demand, some suggest that several supply-side factors tion discusses whether any of these trends were present may have played a role. Supporters of the demand shock before the crisis. The final section concludes. For full hypothesis argue that the collapse in trade was the result details of our methodology and data sources, readers are of a synchronized postponement of purchases, especially referred to the working paper version of this chapter. of durable consumer and investment products. Eaton et al. (2010) and Levchenko, Lewis, and Tesar (2010) are among Changes in the Intensive and those arguing that the collapse in trade was primarily the the Extensive Margins result of demand-side shocks. By contrast, supporters of the supply shock hypothesis suggest that the collapse in The speed and sustainability of recovery from the crisis trade was a consequence of the sudden financial arrest, depend partly on which margin, extensive or intensive, has which froze global credit markets and spilled over onto the been more affected by the crisis and by how fast it responds specialized financial instruments that finance interna- to fiscal stimulus. The intensive margin refers to changes in tional trade. Others have noted that with the globalization the value of exports due to changes in the quantities or of supply chains, a fall in manufactures could lead to an prices of already exported goods. The extensive margin outsized fall in total trade, particularly if supply chains refers to changes in the value of exports due to changes in are disrupted. Finally, some highlight the role of rising the number of goods exported or changes in the number of protectionism. destinations to which a country exports old or new goods. In this chapter, we identify a new set of stylized facts on If significant fixed costs are associated with exporting or the 2008­09 trade collapse that can shed light on the importing new products (that is, the extensive margin), importance of demand- and supply-side factors in explain- then identifying the relative size of changes along the ing the fall in trade. In particular, we decompose the fall in extensive and intensive margins is useful in predicting the international trade into product entry and exit, price speed of the recovery. For instance, we would expect a changes, and quantity changes for imports to Brazil, the faster recovery if most of the changes in trade were on the European Union (EU), Indonesia, and the United States. intensive margin. Evidence from U.S. and French firms Our ability to separate price and quantity changes allows (Schott 2009; Bricongne et al. 2009, respectively) suggests us to identify the upward trajectory of prices for manufac- that the intensive margin was more affected during this cri- turing, which could be consistent with a role for credit sis than the extensive margin. These results are consistent 55 56 Managing Openness with Bernard et al. (2009), who analyze past recessionary dropped from US$4 trillion in 2008 to US$3 trillion in periods. Yet none of these studies has examined price 2009, a total fall of 25.2 percent from the first half of 2008 changes. In this section, we discuss changes in the extensive to the first half of 2009. The quantity effect accounted for and intensive margins in recent years. over 15.9 percent out of the total 25.2 percent value drop, In October 2008, the total number of bilaterally traded and the decline in prices accounted for only 5.5 percent. products began to fall in Brazil, Indonesia, and the United Net entry, the sum of exit and entry, was negative, but States (figure 5.1). The percentage fall in the total number accounted for only a small portion of the change in total of products was greatest for Brazil and Indonesia, reaching value. Imports by Brazil and Indonesia dropped from about 10 percent from peak to trough for both countries. US$135 billion to US$109 billion, a fall of 18.9 percent By contrast, the percentage of products traded for the from the first half of 2008 to the first half of 2009. For European Union did not decline, and the decline for the Brazil and Indonesia, changes in quantity accounted for United States was half of that for Brazil and Indonesia, 18.5 percent out of the total 18.9 percent drop in the around 5 percent. value of trade, with product net exit accounting for Brazil, the European Union,2 Indonesia, and the United 1 percent and a small price increase partially offsetting States all experienced a sharp drop in the total value of these effects. The results for the intensive and extensive imports beginning in October 2008, with recovery begin- margins match evidence by Schott (2009) and Bricongne ning in early 2009 (figure 5.2). None managed to re-attain et al. (2009), who find that for U.S. and French firms, precrisis levels by the end of the period shown (September changes in the intensive margin outweighed changes in 2009). Developing countries experienced much greater the extensive margin. volatility than the developed countries. Indeed, before beginning to recover, the total value of imports fell by Demand and Supply Shocks nearly half for Brazil and Indonesia. At the end of the sam- ple period, however, the total loss in trade was about the Determining what happened to traded prices and quanti- same for the developed and developing countries (around ties between 2008 and 2009 offers a new approach to deter- one-third of aggregate trade).3 mining the role of demand versus supply factors in the Effects along the intensive margin dramatically out- trade collapse. If the decline in trade was driven mostly by a weighed the effects along the extensive margin, both in negative demand shock, we would expect both prices and the U.S.-EU markets and in the Brazil-Indonesia markets quantities to have been negatively affected. However, if (figure 5.3). Imports by the United States and the EU supply-side shocks were important, we would expect an Figure 5.1. Index of Total Number of Products Traded in Brazil, the European Union, Indonesia, and the United States, January 2007­September 2009 1.15 1.1 January 2007 = 1 1.05 1 0.95 0.9 0.85 0.8 7 8 09 0 0 20 20 20 year United States average across EU Brazil Indonesia Source: Authors' calculations based on data from national statistics agencies. Note: Products are tallied by partner country. An Anatomy of Trade in the 2008­09 Crisis 57 Figure 5.2. Index of Total Value of Imports for Brazil, the European Union, Indonesia, and the United States, January 2007­September 2009 2 1.8 1.6 January 2007 = 1 1.4 1.2 1 0.8 0.6 0.4 0.2 0 07 08 09 20 20 20 year United States average across EU Brazil Indonesia Source: Authors' calculations based on data from national statistics agencies. Figure 5.3. Change in Total Import Value of All Products level,4 the evidence is consistent with a pure demand shock in Brazil-Indonesia and United States­EU Trade, by Margin, story: a fall in demand generated declines in prices and 2008­09 percent quantities. For the United States, 36 percent of the reduc- tion in trade across all trading partners can be explained by a decline in prices, while 57 percent can be explained by a quantity effect decline in quantity. The contribution of entry and exit accounted for less than 1 percent of the observed changes. price effect In both the EU and Indonesia, the quantity declines domi- exiting products nated the price effect, while for Brazil the opposite is true. For all countries, the intensive margin again is substantially new products larger than the extensive margin. Exit again outweighs entry, leading to negative net entry in all regions. ­0.20 ­0.15 ­0.10 ­0.05 0 0.05 However, if we restrict the sample to manufacturing, a dif- percent ferent story emerges. We see that for manufactures, the fall in Brazil-Indonesia United States­EU quantity continues to account for the major share of the Source: Authors' calculations based on data from national statistics observed trade collapse. However, for Brazil, Indonesia, and agencies. the United States there is also evidence of a supply shock: price increases offset the contribution of declining prices to the fall in trade. This finding is particularly striking for upward pressure on prices because a reduction in trade Indonesia. All in all, the evidence suggests that supply-side credit would lead to a reduction in the supply of traded disruptions play a more important role in manufacturing. goods independently of the negative demand shock. Several outliers in these trends should be noted. For While the largest contributor to the crisis seems to be instance, members of the Organization of the Petroleum- the demand shock, there is also systematic evidence of fac- Exporting Countries experienced larger price effects on tors generating negative supply shocks in manufacturing. their exports to the United States than did other countries. By extending work by Bernard et al. (2009) to decompose Oil-exporting countries, including Algeria, Angola, Iraq, the intensive margin into price and quantity effects in an the Russian Federation, Saudia Arabia, and República Boli- ordinary least squares linear regression, we find evidence of variana de Venezuela, show price effects that are larger both demand and supply shocks. Indeed, at the aggregate (more negative) than the average. Their quantity effects 58 Managing Openness are, correspondingly, smaller than the average. Other large commodities; it is, however, noteworthy that the price trading partners--such as France; Germany; Italy; Japan; effect was generally limited to commodities; the price effect the Republic of Korea; Malaysia; Taiwan, China; and the for manufactures was positive. United Kingdom--show larger quantity effects than the Since we know that demand for manufactures fell dur- average across countries. ing the crisis, the effect on prices can tell us something In an alternative decomposition of the total trade value5 about what happened to supply. Where prices rose or by trading partner, product, and average value,6 we find where they fell only slightly, it is plausible that supply that while the extensive margin experienced negative shifted in. Thus, the evidence on manufactures, contrasted changes, the intensive margin mattered more. No country with commodities and particularly in the case of Brazil and saw an increase in the number of trading partners from Indonesia, points to a negative supply shock in manufac- 2008 to 2009. In fact, Brazil lost almost 10 trading partners. tures in addition to the negative demand shock. This nega- All four regions also saw a fall in the number of products tive supply shock could be from fragmentation of the traded, although this was a very small percentage of the global supply chain or from reductions in trade finance. total loss. The largest changes were in average value, which The prevalence of supply shocks in the manufacturing fell by 10 to 28 percent across the various regions. sector is also confirmed when we analyze differences across product types. The largest total value change is in minerals, which fell by US$134 billion. Almost 90 percent of that Variation by Product Type decline was due to price falls. Machinery and electrical Our research shows that these average effects mask enor- equipment and transportation equipment also experienced mous differences across different products. Figure 5.4 large falls in total volume, but almost all of the changes shows the difference between commodities and manufac- were due to a decline in quantity for these products. tures, where each margin is shown as a percentage of the Prices increased in the following product categories: change in total import value in 2008. The differences across chemicals, footwear, leather, miscellaneous, and trans- product classes are striking. The negative price effect, portation equipment (note that because the total value apparent for the United States and EU when aggregating change is negative, a positive total price effect is repre- across all goods, is still evident for commodities but not for sented by a negative percentage of the value change). Most manufactures. With commodity prices falling during the of the large negative price effects were in product cate- crisis, it is not surprising that the price effect was large for gories made up largely of commodities (animal products, Figure 5.4. Variations in Price and Quantity in Commodities and Manufactures, 2008 a. Commodities b. Manufactures price effect (Brazil) price effect (Brazil) price effect (Indonesia) price effect (Indonesia) price effect (EU) price effect (EU) price effect (United States) price effect (United States) quantity effect (Brazil) quantity effect (Brazil) quantity effect (Indonesia) quantity effect (Indonesia) quantity effect (EU) quantity effect (EU) quantity effect (United States) quantity effect (United States) ­0.5 ­0.4 ­0.3 ­0.2 ­0.1 0 ­0.4 ­0.2 0 0.2 0.4 percent percent Source: Authors' calculations based on data from national statistics agencies. An Anatomy of Trade in the 2008­09 Crisis 59 minerals, and vegetable products). All products saw falls in We also examine evidence on credit constraints by adopt- quantity, with the quantity effect generally contributing to ing the classification scheme of Bricongne et al. (2009) to over half of the total effect. The quantity effect was rela- separate products according to sectoral dependence on tively small for minerals, where the price effect was largest. external finance. We restrict our analysis to manufactures. As figure 5.5 illustrates, differences in trade responses For the United States, price increases were most significant across modes of transport are not significant. However, a in sectors that are typically credit constrained, partially distinction can be made between manufactures and com- counteracting the large quantity effect. The EU does not modities. Commodities exhibited both price and quantity show this effect. For Brazil, the overall import value for the declines.7 In contrast, manufactures showed quantity finance-dependent sectors dropped (by 3 percent) but rose declines but mild price increases. The overall trends in for the low-dependence sectors (by 9 percent). The overall figure 5.5 are consistent with an overall pattern of import value also fell more for finance-dependent sectors demand contraction, with some evidence of supply con- (19 percent as opposed to 6 percent) in Indonesian imports. straints in manufacturing. The largest price increases were in the leather and footwear sector, where credit con- Variation by Income Group straints and trade frictions may have restricted supply. Finally, it is interesting to note that for many products, While the financial crisis originated in high-income coun- the quantity effect is larger for transport by air than for tries, its effects on trade were rapidly transmitted to low- transport by sea.8 income countries. It has been hypothesized that the effects Figure 5.5. Changes in Import Values across Product Types and Shipment Methods for U.S. Imports as a Percentage of Value by Margin, 2008 animal: air, sea chemical: air, sea foodstuffs: air, sea footwear, headgear: air, sea leather: air, sea machinery: air, sea metals: air, sea minerals: air, sea miscellaneous: air, sea plastics: air, sea stone, glass: air, sea textiles: air, sea transportation: air, sea vegetables: air, sea wood products: air, sea ­0.60 ­0.50 ­0.40 ­0.30 ­0.20 ­0.10 0 0.10 0.20 0.30 percent price effect quantity effect Source: Authors' calculations based on data from national statistics agencies. Note: Within each product grouping (for example, animal products), the upper bar shows the effect on air shipments and the lower bar shows the effect on vessel shipments. The effects are percentage changes of the total effect within that shipment type. For instance, the price effect for animal products shipped by sea is divided by the total value in 2008 of animal products shipped by sea. Note that typically sea shipments were much larger in 2008, so the gross value changes for sea shipments are larger compared to those for air shipments than what is shown. Entry and exit are not shown and represent less than 5 percent of the change in any given category. Continuing products with unobserved price or quantity are also not shown. 60 Managing Openness of constrained trade finance could vary by exporter very small. The most striking result is the large increase in income (Malouche 2009; Berman and Martin 2010) and exports of low-income countries to Brazil and Indonesia by geographic region (Berman and Martin 2010). On the of nearly 30 percent between 2008 and 2009. This finding one hand, high-income countries with well-developed confirms that South-South trade is becoming increas- markets were most affected in the financial crisis; on the ingly important and was reinforced during the crisis. other hand, low-income exporters with less-developed Sub-Saharan Africa, however, was not able to take advan- financial markets may be more reliant on trade finance tage of these South-South trade opportunities as its originating in their trading partners. Countries with dif- exports to Brazil and Indonesia dropped 27 percent. ferent levels of income export different baskets of goods, China's exports to the United States and the EU dropped which embody different levels of quality and variety. by only 8 percent, in line with other lower-middle- To analyze how the response in trade volumes changes income countries; but China's exports to Brazil and with the income level of the exporting country, we clas- Indonesia increased by 5 percent. sify trading partners in four categories: high, upper- middle, lower-middle, and low income, according to the Were Any of These Trends Present World Bank's country classification. We also add China before the Crisis? and Sub-Saharan Africa as separate categories, two regions where researchers have hypothesized that the We examine whether these findings are unique to the crisis trade fall was unique. As mentioned above, it has been or whether they represent the continuation of historical theorized that the trade finance constraint could have trends. Figure 5.7 shows changes in each margin (in billions been either much more severe or much less severe in of U.S. dollars) for U.S. and Indonesian imports across Sub-Saharan Africa. China has been unique both because quarters for 2007­09. Price and quantity changes are now it recovered from the crisis more rapidly and because it defined relative to the previous quarter rather than to the has been a target for protectionism (Bown 2009). The same quarter of the previous year. Hence, the magnitudes of story can be further elaborated by returning to the het- these changes do not match the magnitudes given in the erogeneity across trading partner income but restricting other figures, but they do show the specific timing of the the sample to manufactures. The results are shown in collapse in trade. figure 5.6. As expected, entry and exit do not play a large role in Overall, upper- and upper-middle-income exports to U.S. imports (before and after the crisis) but are more developed countries were most affected by the crisis, with important in a developing country like Indonesia, where falls in the value of their exports reaching 25 percent to the trade relations are thinner and less established. The quan- United States and EU. Low-income countries were able to tity and price effects are not part of broader historical increase their exports to the United States and EU by trends; rather, they match the timing of the global eco- 7 percent. The impact of the crisis on the exports of coun- nomic crisis. Manufacturing imports to the United States tries of various income groups to Brazil and Indonesia was level off in the third quarter of 2008, as the crisis is begin- ning, and then plummet in the following two quarters. For this whole period, the negative quantity effect dominates, and there is a smaller positive price effect for the fourth Figure 5.6. Percentage Change in Imports to Brazil and quarter of 2008. Manufacturing imports to Indonesia fol- Indonesia and to the United States and the EU from Countries low a similar pattern, with a negative quantity effect begin- in Different Income Groups, 2008­09 ning in the fourth quarter of 2008 and an initial positive price effect. Commodity imports to the United States also Sub-Saharan Africa plummet in the fourth quarter of 2008, but here the nega- China tive price effect dominates. For commodity imports to Low-income trading partner Indonesia, both quantity and price effects are negative and Lower-middle income trading partner begin around the fourth quarter of 2008. Upper-middle income trading partner High-income trading partner ­0.4 ­0.2 0 0.2 0.4 Conclusion percent Brazil-Indonesia United States­EU In summary, the great trade collapse occurred along both Source: Authors' calculations based on data from national statistics intensive and extensive margins. The intensive margin had agencies. much greater impacts, with negative effects for both prices An Anatomy of Trade in the 2008­09 Crisis 61 Figure 5.7. Percentage Changes in Imports of Manufactures and Commodities to the United States and Indonesia, Second Quarter, 2007­Third Quarter, 2009 a. U.S. imports of manufactures b. U.S. imports of commodities 30 30 20 20 10 10 0 0 percent percent ­10 ­10 ­20 ­20 ­30 ­40 ­30 ­50 ­40 ­60 ­50 07 Q 07 Q 07 Q 08 Q 08 Q 08 Q 08 Q 09 Q 09 09 07 Q 07 Q 07 Q 08 Q 08 Q 08 Q 08 Q 09 Q 09 09 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 2 3 4 1 2 3 4 1 2 3 2 3 4 1 2 3 4 1 2 3 Q Q Q Q year year c. Indonesian imports of manufactures d. Indonesian imports of commodities 3.0 1.5 2.0 1.0 1.0 0.5 percent percent 0.0 0.0 ­1.0 ­0.5 ­2.0 ­1.0 ­3.0 ­1.5 ­4.0 ­2.0 07 07 Q 07 Q 08 Q 08 Q 08 Q 08 Q 09 Q 09 09 07 07 Q 07 Q 08 Q 08 Q 08 Q 08 Q 09 Q 09 09 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 2 3 4 1 2 3 4 1 2 3 2 3 4 1 2 3 4 1 2 3 Q Q Q Q Q Q year year quantity effect price effect entry exit change in value Source: Authors' calculations based on data from national statistics agencies. and quantities. Across all products, most of the countries conclusions reached by Eaton et al. (2010) and Levchenko, analyzed experienced a decline in new products, a rise in Lewis, and Tesar (2010), who argue that the collapse in product exit, and reductions in quantity for product lines trade was caused primarily by a synchronized demand- that continued to be traded. These effects are similar for side shock. high-income and middle-income partner countries, but However, these average effects mask enormous differ- trade with low-income partner countries was much less ences across different product types. Disaggregating the affected. data into manufactures and nonmanufactures, we find that The evidence suggests that the intensive rather than the price declines are driven primarily by commodities. the extensive margin mattered the most, consistent with Within manufacturing, however, while most quantity studies of other countries and previous recessionary peri- changes were negative, in many cases price changes moved ods. On average, quantities declined and prices fell, which in the opposite direction, particularly for products is consistent with a story in which the demand shock imported by developing-country trading partners. Conse- played a dominant role. Aggregating across all product quently, within manufacturing, some evidence indicates categories, we find that the evidence is consistent with the that supply-side frictions did play a role. 62 Managing Openness Considerable differences emerge across product types Bibliography and exporter incomes. Some have argued that South- Baldwin, R, ed. 2009. The Great Trade Collapse: Causes, Consequences and South trade was less disrupted during the crisis. While Prospects, VoxEU.org Ebook, available at http://www.voxeu.org/index this appears to be the case if one examines aggregate trade .php?q=node/4297. Berman, N., and P. Martin. 2010. "The Vulnerability of Sub-Saharan patterns across all goods, a large decline occurred in man- Africa to the Financial Crisis: The Case of Trade." CEPR Discussion ufacturing trade with Sub-Saharan Africa. The most strik- Paper 7765, Centre for Economic Policy Research, London, United ing result is the large increase in exports of low-income Kingdom. Bernard, A. B., J. B. Jensen, S. J. Redding, and P. K. Schott. 2009. "The Mar- countries to Brazil and Indonesia--nearly 30 percent gins of U.S. Trade." American Economic Review: Papers & Proceedings between 2008 and 2009. We also find evidence consistent 99 (2): 487­93. with the view that credit constraints could account for the Bown, C. 2009. "The Global Resort to Antidumping, Safeguards, and price increases occurring in the manufacturing sector. For Other Trade Remedies amidst the Economic Crisis." Policy Research Working Paper 5051, World Bank, Washington, DC. the United States, for example, price increases were most Bricongne, J.-C., L. Fontagne, G. Gauler, D. Taglioni, and V. Vicard. 2009. significant in sectors that are typically credit constrained. "Firms and the Global Crisis: French Exports in the Turmoil." Docu- ment de Travail 265, Banque de France, Paris. Council of Economic Advisers. 2010. Economic Report of the President. Notes U.S. Government Printing Office, Washington, DC. Eaton, J., S. Kortum, B. Neiman, and J. Romalis. 2010. "Trade and the 1. For summaries, see Baldwin (2009) and Council of Economic Global Recession." Working Paper 16, National Bank of Belgium, Advisers (2010). Brussels. 2. This is an average across the European Union. Levchenko, A. A., L. T. Lewis, and L. L. Tesar. 2010. "The Collapse of 3. Since the sample ends in September 2009, we do not capture the International Trade during the 2008-2009 Crisis: In Search of the continued slow recovery since that period. Smoking Gun." Discussion Paper 592, Ford School of Public Policy, 4. This level includes commodities. University of Michigan, Ann Arbor, MI. 5. The change in total trade value is included. Malouche, M. 2009. "Trade and Trade Finance Developments in 14 Devel- 6. This follows Bernard et al. (2009). oping Countries Post September 2008: A World Bank Survey." Policy 7. Declines are indicated by both the dark and the light bars located Research Working Paper 5138, World Bank, Washington, DC. on the left-hand side of the graph for minerals, stone and glass, animal Rauch, J. E. 1999. "Networks versus Markets in International Trade." Jour- products, and vegetables. nal of International Economics 48 (1): 7­35. 8. Products do not include chemicals, foodstuffs, and transportation Schott, P. 2009. "US Trade Margins during the 2008 Crisis." VoxEU. equipment. November 27. http://voxeu.org. 6 Developing Countries, New Trade Barriers, and the Global Economic Crisis Chad P. Bown and Hiau Looi Kee The financial crisis that began in mid-2008 quickly spread developing economies--both as imposers of new trade globally and resulted in a major shock to the international barriers and as exporters whose trade is most likely to be economy. The initial concern was whether the global econ- adversely affected by such barriers.2 omy would suffer as deep an economic decline as in the To begin, figure 6.1 presents monthly data on the Great Depression of the 1930s. The onset of a coordinated growth of imports from Freund and Horenstein (2010) for recession across the world's major economies and a severe the period January 2008 through November 2009. The collapse in international trade flows further stoked fears of fourth quarter of 2008 saw a sudden and almost simulta- an impending protectionist backlash. Just as the 2008­09 neous drop in global trade flows for virtually all the major recession led to injured industries and massive unemploy- regions of the world. The trend continued through the ment and drew comparisons to the Great Depression, the first three quarters of 2009, which continued to register specter of the 1929 U.S. Smoot-Hawley tariffs and the subse- negative rates of import growth. The World Trade Organi- quent international protectionist response of the 1930s cre- zation (WTO) (2009, 3) estimates a year-on-year decline ated the worry that the recent crisis would result in a similar in world trade in the fourth quarter of 2008 alone as over global imposition of new trade barriers and severely curtail 10 percent, with another year-on-year decline of 30 per- the timely resumption of international trade (Eichengreen cent in the first quarter 2009. In our focus on protection- and Irwin 2009a, forthcoming; Irwin, forthcoming). ism during the crisis, we first clarify the limited role that This chapter presents a set of stylized facts on the new trade barriers contributed to the trade collapse of protectionism that has emerged during the global eco- 2008­09. The evidence we present on the timing and the nomic crisis of 2008­09. We assess the magnitude and scale of the new import protection under these barriers is distribution of the policy changes that occurred by plac- consistent with the existing evidence that such protection ing them into recent historical context. In line with the likely contributed only slightly to the sharp decline in types of explicit trade barriers that countries have global trade flows.3 imposed during the crisis, we place special emphasis on We examine the evolving changes in protectionism the role of import restrictions resulting from national across countries resulting from the crisis to infer how such use of antidumping, countervailing duty, and safeguard barriers may affect future patterns of trade, including the policies.1 Examination of the detailed evidence from the potential for a V-shaped postcrisis trade recovery. While World Bank's Temporary Trade Barriers Database (Bown the newly documented protectionism imposed during 2010c) on the evolving use of these policies also man- 2008­09 may have played a minuscule role in causing the dates that we pay particular attention to the interests of global trade collapse, the emerging pattern of new trade Thanks to Aksel Erbahar for outstanding research assistance. 63 64 Managing Openness Figure 6.1. Monthly Import Growth by Region, January 2008­November 2009 80.0 import growth month over same month, previous year (%) 60.0 40.0 20.0 0 ­20.0 ­40.0 ­60.0 08 08 08 8 09 09 09 9 00 00 20 20 20 20 20 20 2 2 y ril ly er y ril y er ar ar l Ju Ju Ap Ap ob ob nu nu ct ct Ja Ja O O East Asia and Pacific Europe and Central Asia Latin America and the Caribbean Middle East and North Africa North America South Asia Sub-Saharan Africa Source: Authors, based on data in Freund and Horenstein 2010. barriers imposed across countries and exporters during Protectionism from the Perspective the crisis is likely to have longer-term implications for the of Domestic Industries and potential trade recovery in 2010 and beyond. Much of the Importing Economies evidence of new crisis-era protectionism is in the form of We begin our analysis of protectionism during the crisis by South-South trade barriers--policies such as antidump- examining it from the perspective of the policy-imposing ing that one developing economy imposes on the imports economies and their imports that are likely affected by the of other developing economies, including, but not limited imposition of new barriers to trade. We focus primarily on to, imports from China. While this phenomenon is not the Group of 20 (G-20) members as the policy-imposing new--that is, it had been trending in this direction long economies. before 2008­09--it was certainly accentuated during the crisis period. Temporary Trade Barriers: Antidumping, Countervailing The chapter proceeds as follows. In the next section, we Duties, and Safeguards describe newly compiled, detailed data on the imposition of trade barriers from the perspective of the policy-imposing Long before the onset of the 2008­09 crisis, most of the economies. The following section turns its attention to G-20 economies had a significant history of using at least exporting countries and the impacts they have seen. We one of the four policies that we refer to collectively as tem- conclude the chapter with a final discussion of lingering porary trade barriers (TTBs): antidumping, countervailing issues and policy implications. duties, global safeguards, and the China-specific safeguard.4 Developing Countries, New Trade Barriers, and the Global Economic Crisis 65 For the purpose of tracking trends in levels of import protec- TTB protection over time. The underlying stock of prod- tion over time, we group together these four TTBs because ucts that could be subject to a TTB is relatively fixed over they are relatively substitutable forms of import protection. time (at least at the six-digit HS level), and thus, our stock The WTO establishes a minimal set of conditions for each measure is a simple indicator of the scope of imported of the four different TTBs that national policy makers must products affected by an economy's use of TTBs over time. follow in order to implement new import protection. The first stylized fact is that the combined G-20 use of such temporary trade policies has resulted in a signifi- cantly higher stock of products covered by import protec- The Products Affected by TTBs and the Global tion in 2009 relative to precrisis levels.6 The dark black line Economic Crisis of figure 6.2 provides a summary from the World Bank's In this section, we follow the methodology described in Temporary Trade Barriers Database of the combined major detail in Bown (2010b) and begin by assessing the "stock" G-20 users of these four policies over the 1997­2009 of six-digit Harmonized Commodity Description and period. While the database contains information for many Coding System (HS) products on which an economy has of these economies that dates back further and even into imposed at least one of the four import-restricting TTB the 1980s, 1997 is a useful starting point for the current policies.5 The basic argument for focusing first on the analysis, because that is the first year in this sample in stock of products covered is that changes in the flow of which each of the G-20 economies that we analyze started new barriers imposed and removed can affect the level of using these temporary trade barriers. Figure 6.2. Combined G-20 Use of Temporary Trade Barriers, 1997­2009 1,600 1,400 1,200 number of products 1,000 800 600 400 200 0 97 98 99 00 01 02 03 04 05 6 07 08 9 0 0 19 19 19 20 20 20 20 20 20 20 20 20 20 stock: products under trade barrier (AD, CVD, SG, CSG) stock: products under trade barrier (AD only) flow: products subject to newly initiated trade barrier investigation (AD, CVD, SG, CSG) flow: products subject to newly initiated trade barrier investigation (AD only) Source: Bown 2010b. Note: Data on the stock of policies imposed and removed over 1988­2009 were compiled from the Temporary Trade Barriers Database. A "product" is defined at the six-digit HS level. Figure 6.2 illustrates the number of importing country­product combinations affected by policies such as antidumping, countervailing duties, global safeguards, and China-specific transitional safeguards. The data are aggregated over the following 12 G-20 economies: Argentina, Australia, Brazil, Canada, China, the European Union, India, Indonesia, the Republic of Korea, South Africa, Turkey, and the United States. The only major G-20 user of such policies not included in the figure is Mexico, for reasons described in the text. The "stock" includes both imposition and removal of import restrictions after terminations or sunset reviews. With roughly 5,000 six-digit HS product categories per importing economy and 12 policy-imposing countries, the maximum value that the vertical axis could possibly take is 60,000. AD = antidumping, CSG = China-specific transitional safeguards, CVD = countervailing duties, SG = global safeguards. 66 Managing Openness According to figure 6.2, by the end of 2009 the major G- Figure 6.3. Combined G-20 Use of Selected Temporary 20 users of TTBs together subjected 25 percent more Trade Barriers by Import Source, 1997­2009 import product lines to these trade barriers than they did 3,500 in 2007. While the detailed information in the Temporary Trade Barriers Database does indicate that each of the four 3,000 unique HS06 product-exporter TTB policies was used during the crisis, the gray line in 2,500 figure 6.2 shows that most of the products subject to the combinations TTBs in place during 2007 through 2009 are covered by 2,000 antidumping policies, which is still the primary TTB of 1,500 choice for many governments. The dashed and dotted lines in figure 6.2 provide information on the potential "flow" of 1,000 products that may be subject to new TTBs by illustrating 500 the number of six-digit HS lines subject to newly initiated investigations each year. 0 97 98 99 00 01 02 03 04 05 06 07 08 09 As is also clear from figure 6.2, the increase in imported 19 19 19 20 20 20 20 20 20 20 20 20 20 products subject to TTBs between 2007 and 2009 is part of China as exporter a broader, longer-term upward trend in the use of these developing economy as exporter (excluding China) sorts of trade barriers. A number of these G-20 member developed economy as exporter economies undertook extensive tariff-cutting, trade liber- Source: Data on the stock of policies imposed and removed over alization episodes in the 1990s and early 2000s and subse- 1988­2009 compiled from the Temporary Trade Barriers Database. quently promised to keep their applied tariffs low--either Note: A "product" is defined at the six-digit HS level. Figure 6.3 illustrates through WTO tariff "binding" commitments or through the number of importing country-product­exporting country target combinations affected by policies such as antidumping, countervailing preferential trade agreements. Thus, a general increase in duties, and China-specific transitional safeguards. The policies are the stock of products subject to these other, more flexible aggregated over the following 12 G-20 economies: Argentina, Australia, Brazil, Canada, China, the European Union, India, Indonesia, the Republic TTB policies, which may be thought of as imperfect substi- of Korea, South Africa, Turkey, and the United States. Mexico is the only tutes for tariffs, is perhaps not surprising. The broad sense major G-20 user of such policies not included, and the reason for its exclusion in this figure is described in the text. The "stock" includes both from figure 6.2 is that this trend has been ongoing at least imposition and removal of import restrictions after terminations or sunset since 1997 and perhaps would have continued at some level reviews. Unlike figure 6.2, this figure does not reflect the economy's potential use of the global safeguards policy, which is not exporting-country specific. irrespective of the crisis. Figure 6.3 presents a second way to measure and exam- ine the extent of imports covered by the TTBs imposed by than the 2007 stock of product-exporter combinations, the major G-20 economies in effect over time. The unit of and starting from a lower baseline leads to higher growth measurement here goes beyond the product to include the rate. Second, as we describe in more detail below, the inci- number of distinct exporting-country targets that a new dence of new barriers imposed between 2007 and 2009 was TTB over a given product affects. It therefore measures the increasingly on a single foreign supplier (that is, products combination of products and exporting countries affected the G-20 imported from China). The stock of products the by (only) antidumping, countervailing duty, and the G-20 had covered with TTBs that was in effect by 2007 China-specific safeguard policies: that is, omitting prod- affected many more foreign suppliers on average than the ucts subject only to global safeguard policies since such new TTBs that added to that stock in 2008­09.7 policies are applied on a most-favored nation basis against Figure 6.3 illustrates this second point: how the incidence all foreign sources. The figure also breaks out the incidence of the new TTBs added to the stock during 2008­09 was not of the exporters affected by these policies into three groups: uniform across export sources. China had 40 percent more developed-economy exporters, China, and other (non- exported product lines subject to these G-20 TTBs by the China) developing economies. end of 2009 relative to the precrisis level of 2007. The com- Compared to 2007, figure 6.3 illustrates that the major bined increase for all other developing-economy exporters users of these country-specific policies had a 12 percent was 4 percent, while developed-economy exporters faced higher stock of product-exporter combinations subject to roughly the same number of products affected in 2009 as in TTBs by the end of 2009. Two main factors explain why the 2007. We describe in substantially more detail these and 12 percent increase from 2007 to 2009 of product-exporter other insights from the exporter's perspective below. combinations is lower than the 25 percent increase for Table 6.1 summarizes the economy-by-economy differ- products alone. First, the 2007 stock of products was less ences across the major G-20 users of these TTBs. Column (1) Developing Countries, New Trade Barriers, and the Global Economic Crisis 67 Table 6.1. Stocks of Temporary Trade Barriers Imposed by Individual G-20 Economies, 2009 Stock of % Share of (3) % change in product­exporter change in imposed Share of (3) Stock of products (1) relative to combinations (3) relative against all imposed subject to such precrisis 2007 subject to such to precrisis developing against barriers in 2009 level barriers in 2009 2007 level economies China only G-20 economy imposer (1) (2) (3) (4) (5) (6) Developing economies India 287 0.61 527 0.41 0.58 0.39 Turkey 256 0.46 276 0.19 0.88 0.55 Argentina 139 0.48 342 0.26 0.77 0.22 Brazil 82 0.22 113 0.06 0.70 0.42 Mexico 61 ­2.84 80 ­2.59 0.56 0.19 China 46 ­0.10 132 ­0.26 0.16 NA South Africa 40 ­0.18 59 ­0.17 0.80 0.34 Indonesia 24 0.69 66 1.05 0.74 0.18 Total developing economies 935 0.41a 1,595 0.22a 0.68b 0.35b High-income economies United States 260 0.10 781 0.05 0.63 0.22 European Union 137 ­0.05 249 ­0.07 0.86 0.41 Canada 69 0.16 238 0.08 0.70 0.26 Korea, Rep. 38 ­0.44 72 ­0.25 0.60 0.31 Australia 31 0.39 43 0.26 0.65 0.44 Total high-income economies 535 0.03 1,383 0.02 0.68 0.27 Source: Authors, based on data from the Temporary Trade Barriers Database (Bown 2010c). a. Data on the stock of policies in place in 2009 compiled by the authors from the Temporary Trade Barriers Database (Bown 2010c). Column (1) includes products affected by the use of one or more of the following four policies: antidumping, countervailing duties, global safeguards, and China-specific transi- tional safeguards. Column (3) documents the number of exporting country­product combinations affected by the use of only three policies: antidumping, countervailing duties, and China-specific transitional safeguards. The "stock" includes both imposition and removal of import restrictions after terminations or sunset reviews. The maximum value that column (1) could take on for any one economy is roughly 5,000 six-digit HS products. In columns (2) and (4), the percent change from 2007 to 2009 aggregated for the developing economies does not include Mexico for reasons described in the text. b. In columns (5) and (6), the developing-economy aggregated total does not include China's use. lists the stock of six-digit HS products covered by at least close to the ordering based on (1), it is not identical for two one TTB in effect in 2009, defined consistently with the basic reasons. Most important, a larger economy may apply aggregated figure 6.2. Column (2) illustrates the percent- the same product-level TTB to more foreign sources simply age change in this stock for each economy when compared because it imports a given product from more trading part- to the stock of products covered in 2007. Eight of the 13 ners on average owing to its larger market. Furthermore, G-20 economies listed in the table increased the number of countries like India and Turkey were more likely to have products subject to these import restrictions in 2009 when global safeguard policies in effect in 2009. While any prod- compared to 2007--including Argentina (48 percent more uct covered by the global safeguard policy is captured in (1), products covered), Australia (39 percent), Brazil (22 per- we have chosen not to include product-exporter combina- cent), Canada (16 percent), India (61 percent), Indonesia tions subject to global safeguards in (3). Next, column (4) (69 percent), Turkey (46 percent), and the United States reports growth rates of TTB coverage of product­foreign (10 percent). Only China, the European Union, Mexico, the source combinations in 2009 compared to 2007. Similar to Republic of Korea, and South Africa reduced the number the aggregated statistics described in reference to figure 6.3, of products subject to such import barriers between 2007 on a country-by-country basis this statistic is generally and 2009. lower than the percentage change for product coverage Columns (3) and (4) of table 6.1 present economy-by- alone reported in column (2). economy differences of the second approach of measuring Columns (5) and (6) of table 6.1 report the incidence of the stock of imports covered by TTBs in 2009 (consistent the stock of temporary trade barriers in place in 2009 that with figure 6.3); that is, the product-exporter combinations were imposed on imports from developing-economy affected by country-specific TTBs such as antidumping, exporters overall and then specifically against China. Over- countervailing duties and the China-specific safeguard. all, 68 percent of the TTBs that G-20 developing economies First, while the rank ordering of countries based on (3) is had imposed by the end of 2009 were on the imports from 68 Managing Openness other developing countries. They ranged from a low of 56 use of these TTBs. Although there is some heterogeneity, percent (Mexico) to a high of 88 percent (Turkey). Of this incidence of South-South protectionism through this number, a large share of the TTBs were being TTBs has been increasing over time. Some of the protec- imposed on imports from China--an average of 35 per- tionism is certainly to be expected given the evolution of cent of all product-exporter combinations involved China, trade patterns and emerging-economy exports. Because a ranging from a low of 19 percent (Mexico) to a high of number of developing-economy exporters have become 55 percent (Turkey). The users of TTBs from high-income more successful at penetrating new markets for their prod- economies listed in table 6.1 were imposing roughly the ucts, including many developing-economy import markets, same share of barriers on imports from developing adjustment in national import-competing industries needs economies overall (68 percent). However, and perhaps sur- to occur. In some instances, domestic industries resist such prisingly, in 2009 high-income economies imposed a adjustments and will request (and sometimes be granted) smaller share of their TTBs against China (27 percent) than protection from the new import competition through did developing economies (35 percent). additional TTBs. More broadly, figures 6.4 and 6.5 illustrate the evolution Figure 6.5, which illustrates the time-varying stock of of each policy-using economy's stock of imposed TTBs TTBs imposed by high-income economies broken out by over time and across the three different trading partner the same three categories of exporters, shows several categories; that is, the figures decompose figure 6.3 on the broad trends. First, the stock of TTBs in place by 2009 had basis of policy-imposing countries.8 Before turning to declined relative to the post-1995 peak for all five of these interesting differences in time trends across policy-using economies. For Australia and the EU, the peak was around countries, we begin by noting substantial differences in the 1997; for Canada and the United States, it was during the underlying levels of product coverage across the policy- recession of 2001­02.10 Second, for Australia, Canada, the using countries. Specifically, the cumulative levels in 2009 EU, and the United States, the stock of TTBs imposed in each figure correspond to column (3) in table 6.1; for on products from developed-economy exporters either example, in figure 6.4, India had 527 product-exporter remained unchanged or even decreased between 2005 and combinations covered by the TTBs imposed by 2009, 2009. In contrast, for the products subject to the continu- which was almost 10 times as many as the 2009 number of ing TTBs, there is a compositional change toward using only 66 for Indonesia. such barriers to confront imports from China and other Figure 6.4 illustrates the major developing-economy developing economies. members of the G-20 that are users of the TTBs. As described above and summarized in table 6.1 column (4), The Potential Trade Impact of the New TTBs the stock of product-exporter combinations targeted in Imposed during the Crisis 2009 increased in comparison to the precrisis 2007 levels for the following economies: Argentina (26 percent), Brazil The previous section provided one way to assess the chang- (6 percent), India (41 percent), Indonesia (105 percent), ing nature of TTB protectionism: through examination of and Turkey (19 percent). The only decrease in the stock of the stocks of products and product-exporter combinations product-exporter combinations covered by TTBs during affected by TTBs and of changes in those stocks during the this period took place in China, Mexico, and South Africa. 2008­09 crisis. Next, we provide estimates of the potential The illustration for Mexico in figure 6.4 underscores why trade impact of the flow of TTBs newly imposed during we have chosen to exclude it from the aggregated analysis 2008 and 2009.11 We begin by matching the tariff line TTB pictured in figures 6.2 and 6.3.9 In 1993, Mexico imposed policy data to 2007 (precrisis) bilateral, tariff line import new antidumping duties on China covering more than a data for the G-20 economies of interest. We hope to come thousand six-digit HS product lines (roughly 20 percent of up with a basic estimate of how much trade might be elim- all of its six-digit HS codes), and these duties stayed in place inated if the newly imposed TTBs were high enough to be until October 2008 when an agreement was finally reached prohibitive and with an estimate of the importance of the to remove them. Because this removal coincides with the lost trade to total trade. timing of the 2008­09 global economic crisis (but is unre- As an example, the value of 2007 Indian imports of lated to the crisis) and is so extreme in scale, we have chosen products on which India would subsequently impose TTBs to leave Mexico out of figures 6.2 and 6.3 to better capture during 2008­09 was US$3.1 billion, which was 1.4 percent the impact of the crisis on aggregate protectionist trends. of its total 2007 imports (table 6.2). Of the US$3.1 billion Figure 6.4 also reveals the prevalence of China and other of Indian imports subject to new TTBs in 2008­09, roughly developing economies as targets for developing-country 74 percent (US$2.3 billion) were new Indian TTBs 69 unique HS06 product-exporter unique HS06 product-exporter combinations combinations 0 100 200 300 400 500 600 0 50 100 150 200 250 300 350 400 19 19 9 9 19 0 19 0 9 9 19 1 19 1 92 9 19 19 2 9 93 19 3 19 9 9 19 4 19 4 9 9 19 5 19 5 96 9 19 19 6 9 9 19 7 19 7 9 9 19 8 19 8 unique HS06 product-exporter 9 9 20 9 20 9 combinations 0 0 d. India 20 0 20 0 0 0 a. Argentina 0 20 40 60 80 100 120 19 20 1 20 1 9 0 0 19 0 20 2 20 2 9 0 0 19 1 20 3 20 3 9 0 0 19 2 20 4 20 4 9 0 0 19 3 20 5 20 5 9 0 0 19 4 20 6 20 6 9 0 0 19 5 20 7 20 7 9 0 0 19 6 20 8 20 8 9 09 09 19 7 9 19 8 9 unique HS06 product-exporter unique HS06 product-exporter 20 9 0 combinations combinations China as exporter 20 0 0 0 10 20 30 40 50 60 70 80 0 20 40 60 80 100 120 140 19 19 20 1 9 9 g. South Africa 0 19 0 19 0 20 2 0 91 9 19 19 1 20 3 04 9 9 19 2 19 2 20 0 9 9 19 3 19 3 20 5 0 94 9 19 19 4 20 6 0 9 9 19 5 19 5 20 7 0 9 9 19 6 19 6 20 8 09 9 9 19 7 19 7 9 9 19 8 19 8 9 9 unique HS06 product-exporter 20 9 20 9 combinations 0 0 20 0 20 0 b. Brazil 0 0 e. Indonesia 0 50 100 150 200 250 300 19 20 1 20 1 9 0 02 19 0 20 2 20 91 0 0 19 20 3 20 3 9 0 0 19 2 20 4 20 4 9 0 developing economy exporter (excluding China) 05 19 3 20 20 5 9 0 0 19 4 20 6 20 6 Source: Data on the stock of policies imposed and removed over 1988­2009 compiled from the Temporary Trade Barriers Database. 9 0 0 19 5 20 7 20 7 9 0 0 19 6 20 8 20 8 9 09 09 Figure 6.4. G-20 Developing Economies' Use of Selected Temporary Trade Barriers by Import Source, 1990­2009 19 7 9 19 8 9 20 9 unique HS06 product-exporter unique HS06 product-exporter 0 combinations combinations 20 0 h. Turkey 0 20 1 0 200 400 600 800 1,000 1,200 0 20 40 60 80 100 120 140 160 180 200 19 19 20 02 9 9 19 0 19 0 0 9 9 20 3 19 1 19 1 0 9 9 developed economy exporter 20 4 19 2 19 2 0 93 9 20 5 19 19 3 0 9 9 20 6 19 4 19 4 0 9 9 20 7 19 5 19 5 0 9 9 20 8 19 6 19 6 09 9 97 19 7 19 9 9 19 8 19 8 9 9 20 9 20 9 0 0 20 0 20 0 c. China 0 f. Mexico 01 20 20 1 0 0 20 2 20 2 0 0 20 3 20 3 0 0 20 4 20 4 0 05 20 5 20 0 0 20 6 20 6 0 0 20 7 20 7 antidumping, countervailing duties, and China-specific transitional safeguards. The figures do not reflect the economy's potential use of the global safeguards policy which is not exporting-country specific. Note: A "product" is defined at the six-digit HS level. Each figure illustrates the number of importing country-product combinations affected because of the use of exporting-country-specific policies such as 0 0 20 8 20 8 09 09 70 Managing Openness Figure 6.5. G-20 High-Income Economies' Use of Selected Temporary Trade Barriers by Import Source, 1990­2009 a. Australia b. Canada 100 700 unique HS06 product-exporter unique HS06 product-exporter 90 600 80 70 500 combinations combinations 60 400 50 40 300 30 200 20 100 10 0 0 19 0 19 1 92 19 3 19 4 19 5 96 19 7 19 8 99 20 0 20 1 20 2 03 20 4 20 5 06 20 7 20 8 09 19 0 19 1 19 2 19 3 94 19 5 96 19 7 19 8 20 9 00 20 1 02 20 3 04 20 5 20 6 20 7 20 8 09 9 9 9 9 9 9 9 0 0 0 0 0 0 0 9 9 9 9 9 9 9 9 0 0 0 0 0 0 19 19 19 20 20 20 19 19 19 20 20 20 c. European Union d. Korea, Rep. 500 100 unique HS06 product-exporter unique HS06 product-exporter 450 90 400 80 350 70 combinations combinations 300 60 250 50 200 40 150 30 100 20 50 10 0 0 19 0 19 1 19 2 93 19 4 19 5 19 6 19 7 98 20 9 20 0 20 1 20 2 20 3 20 4 20 5 20 6 07 20 8 09 19 0 19 1 19 2 93 19 4 95 19 6 19 7 98 20 9 20 0 20 1 20 2 20 3 20 4 20 5 20 6 07 20 8 09 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 19 19 19 20 19 19 19 19 20 e. United States 1,400 unique HS06 product-exporter 1,200 1,000 combinations 800 600 400 200 0 19 0 91 19 2 19 3 94 19 5 19 6 19 7 98 20 9 20 0 01 20 2 20 3 04 20 5 20 6 20 7 20 8 09 9 9 9 9 9 9 9 0 0 0 0 0 0 0 19 19 19 19 20 20 China as exporter developing economy exporter (excluding China) developed economy exporter Source: Data on the stock of policies imposed and removed over 1988­2009 compiled from the Temporary Trade Barriers Database. against imports deriving from China alone. Finally, of each of these economies' overall 2007 imports (0.5 percent US$2.3 billion in imports from China subject to new and 0.2 percent, respectively). Indian import restrictions covered products that made up One common theme from the table is the extent to 9.5 percent of all Indian imports from China in 2007. which China's exports have been subject to new TTBs Country-by-country information provided in table 6.2 imposed during the crisis. As we have described in the con- also indicates that the magnitude of trade potentially text of figure 6.3, the stock of Chinese exported product affected by India's new TTB policies was not typical of each lines facing TTBs in 2009 was 40 percent higher than the of the major G-20 economies. Only Turkey (1.1 percent of measured stock before the crisis in 2007. Table 6.2 trans- 2007 total imports) imposed new TTBs in 2008­09 over a lates the new TTBs in 2008­09 imposed on China's exports comparable share of its total imports. And while the United as covering an estimated US$20.5 billion in trade in 2007. States (US$10 billion) and European Union (US$7.8 bil- This number is roughly 80 percent of the entire amount of lion) imposed TTBs covering a greater value of imports these economies' total 2007 imports covered by new TTBs than India, these new TTBs covered a much smaller fraction imposed during 2008­09. Furthermore, the estimated Developing Countries, New Trade Barriers, and the Global Economic Crisis 71 Table 6.2. G-20 Imports Subject to Newly Imposed Temporary Trade Barriers, 2008­09 2007 imports for 2007 imports from products subject to China subject to (3) as a share of new TTBs in 2008­09 Share of economy's new TTBs in 2008­09 (3) as a share all 2007 imports G-20 (US$, thousands) total 2007 imports (US$, thousands) of (1) from China economy imposer (1) (2) (3) (4) (5) Developing economies India 3,140,000 0.014 2,332,000 0.743 0.095 China 2,447,700 0.003 n.a. n.a. n.a. Turkey 1,940,000 0.011 640,000 0.323 0.048 Brazil 1,218,000 0.005 653,000 0.536 0.026 Argentina 303,800 0.007 167,400 0.551 0.033 Indonesia 289,026 0.004 123,533 0.427 0.014 Mexico 76,400 0.000 39,600 0.518 0.001 South Africa 7,803 0.000 5,631 0.722 0.001 High-income economies United States 9,990,000 0.005 9,080,000 0.909 0.027 European Union 7,750,000 0.002 6,540,000 0.844 0.010 Canada 673,000 0.002 622,000 0.924 0.017 Australia 281,600 0.002 272,200 0.967 0.012 Total 28,117,330 0.003 20,475,365 0.798a 0.018 Source: Temporary trade barriers imposed during 2008 or 2009 and 2007 import data on the economy's national tariff line level from Comtrade matched to tariff line policy data from the Temporary Trade Barriers Database (Bown 2010c). Note: n.a. = not applicable. a. The aggregated total subtracts out from the denominator the value of China's imports subject to its own new TTBs, since China does not impose TTBs on its own exports. US$20.5 billion was 1.8 percent of the value of total Chinese results for the G-20 policy-imposing economies broken exports to these particular G-20 markets in 2007. into three categories: all sectors, manufacturing, and agri- Finally, table 6.2 presents data for the potential trade culture. The middle three columns present the estimates of affected by the aggregated G-20 use of TTBs during the OTRI for 2008, the estimates of the OTRI for 2009, and 2008­09. While over US$28 billion in G-20 member econ- the change from 2008 to 2009, respectively. The right-most omy imports from 2007 was subsequently subjected to new column is the estimated trade impact in dollars of the TTBs during 2008­09, that number was only 0.3 percent of changes in these economies' OTRI between these two years. these G-20 member economies' total imports. Among the G-20 economies, Argentina, the Russian Federation, and Turkey have the largest increases in their OTRIs, at 0.9, 1.2, and 0.8 percentage points, respectively. Tariffs, Antidumping, and the Overall Trade Since we do not have detailed data on antidumping use Restrictiveness Index for Russia, its increase in OTRI is driven entirely by its Another way to study the potential crisis-era trade impact increase in tariffs, mainly in the manufacturing sector. of policies such as tariffs and antidumping duties is to con- Similarly for Turkey, the increase in its agricultural tariff struct an overall trade restrictiveness index (OTRI) for was the main culprit for its OTRI change, even outweigh- each policy-imposing economy and compare their changes ing the magnitude of its substantial resort to new over time. The OTRI is a more sophisticated way to meas- antidumping duties during the crisis period.13 For ure average tariffs, which takes into account the composi- Argentina, roughly two-thirds of its OTRI increase and the tion of import volume and import demand elasticities. A resulting trade loss can be attributed to antidumping country's OTRI is higher than its import-weighted average duties, while the remaining share is due to the tariff tariff when it levies higher tariffs on the more elastic increases in its manufacturing sector. imported products: that is, when tariffs and import For the EU and the United States, changes in antidump- demand elasticities are positively correlated. Kee, Neagu, ing duties had relatively more impact than tariff changes, and Nicita (2010) constructed the OTRI for a wide range of particularly for manufactured products from China and countries over the crisis period using data on product-line other exporters.14 These duties have resulted in a large loss tariff changes (International Trade Centre) and imposed in trade that overshadows the EU's tariff liberalization in antidumping duties (Bown 2010a).12 Table 6.3 presents the agricultural sector. 72 Managing Openness Table 6.3. Overall Trade Restrictiveness of G-20 Economies, 2008­09 Aggregate trade effects Imposing economy Sector OTRI 2008 OTRI 2009 Change (US$, thousands) Argentina All 0.039 0.048 0.009 ­914,534 Australia All 0.044 0.031 ­0.014 4,575,676 Brazil All 0.080 0.083 0.003 ­991,122 Canada All 0.013 0.016 0.003 ­1,857,762 China All 0.064 0.068 0.003 ­5,263,381 European Union All 0.017 0.018 0.001 ­1,935,871 Indonesia All 0.033 0.027 ­0.006 990,744 India All 0.064 0.067 0.004 ­1,833,246 Japan All 0.044 0.044 0.000 ­75,556 Korea, Rep. All 0.093 0.093 0.000 ­78,149 Mexico All 0.017 0.015 ­0.002 707,280 Russian Federation All 0.096 0.108 0.012 ­4,834,623 Saudi Arabia All 0.040 0.040 0.000 ­1,186 South Africa All 0.033 0.031 ­0.002 199,176 Turkey All 0.020 0.028 0.008 ­2,218,696 United States All 0.011 0.017 0.005 ­24,100,000 Argentina MF 0.045 0.056 0.011 ­926,261 Australia MF 0.047 0.033 ­0.014 4,447,388 Brazil MF 0.089 0.093 0.004 -986,717 Canada MF 0.009 0.010 0.001 ­401,627 China MF 0.055 0.055 0.000 646,881 European Union MF 0.012 0.014 0.001 ­3,648,412 Indonesia MF 0.032 0.026 ­0.006 902,396 India MF 0.057 0.057 0.000 25,154 Japan MF 0.011 0.011 0.000 130,473 Korea, Rep. MF 0.042 0.042 0.000 ­15,776 Mexico MF 0.016 0.014 ­0.002 745,403 Russian Federation MF 0.076 0.088 0.012 ­4,381,372 Saudi Arabia MF 0.039 0.039 0.000 ­304 South Africa MF 0.030 0.029 ­0.001 110,732 Turkey MF 0.008 0.011 0.002 ­628,322 United States MF 0.011 0.017 0.006 ­24,100,000 Argentina AG 0.011 0.011 ­0.001 11,728 Australia AG 0.012 0.006 ­0.005 128,288 Brazil AG 0.033 0.033 0.000 ­4,404 Canada AG 0.058 0.087 0.029 ­1,456,455 China AG 0.161 0.203 0.042 ­5,910,257 European Union AG 0.081 0.073 ­0.008 1,712,236 Indonesia AG 0.044 0.038 ­0.005 88,347 India AG 0.248 0.352 0.104 ­1,858,403 Japan AG 0.316 0.317 0.001 ­206,023 Korea, Rep. AG 0.604 0.605 0.001 ­62,371 Mexico AG 0.021 0.022 0.001 ­38,123 Russian Federation AG 0.204 0.211 0.007 ­453,253 Saudi Arabia AG 0.043 0.043 0.000 ­883 South Africa AG 0.066 0.058 ­0.008 88,444 Turkey AG 0.212 0.314 0.102 ­1,590,374 United States AG 0.019 0.019 0.000 3,707 Source: Kee, Neagu, and Nicita 2010. Note: All calculations are based on bilateral tariffs and antidumping duties, with the exception of India, Japan, and Korea when most-favored nation tariffs are used because of missing data. Changes reported in parentheses are estimated decreases in OTRI. AG = agriculture; MF = manufacturing. Developing Countries, New Trade Barriers, and the Global Economic Crisis 73 Developing-Country Exporters and the years, even during the crisis, a longer time frame indicates Incidence of Crisis-Era Protectionism some change in the composition of TTBs by policy- imposing economy. In each of the panels of figure 6.6, an In this section, we consider the impact of new, crisis-era increasing share of the products covered by TTBs imposed G-20 protectionism from the perspective of the exporting over time derives from policies imposed by developing economies subject to the trade barriers. We follow the same economies. basic presentation approach as the previous section. Figure 6.7 presents the same breakdown of information Because so much of crisis-induced protection has come in for the TTBs affecting a number of developing economies the form of TTBs, we begin by examining changes in the whose exporters are subject to TTBs. As observed in the pattern of the stock of exported products subject to TTBs context of the aggregated figure 6.3, the stock of G-20 TTBs over time. We then use the tariff-line import data matched confronted 884 different Chinese exported product-import to these TTBs imposed during the crisis to project the market combinations in 2009, a 40 percent increase over the amount and distribution of exporters' trade possibly affected stock of product-market combinations subject to G-20 by the imposition of new TTBs in 2008­09. Finally, we TTBs in 2007. Brazil, India, Indonesia, Russia, South turn to model-based estimates of the market access­OTRI. Africa, Thailand, and Ukraine are examples of other devel- oping-economy exporters that confronted a relatively high Temporary Trade Barriers and the Incidence number of their exported products subject to foreign- on Developing-Country Exports imposed TTBs in 2009. Nevertheless, as the economy- While the stock of products subject to new temporary specific panels in figure 6.7 illustrate, there are differences trade barriers in 2009 has increased 25 percent over the across these economies in (a) the frequency with which they precrisis levels of 2007, the data also reveal a nonuniform are subject to such TTBs, and (b) whether the importing- impact of such barriers across exporters: both in the level economy trading partner imposing the new TTBs is a devel- and in the growth rate. As can be surmised from table 6.2, oped or a developing economy. Finally, with the exception many of the new trade barriers during this period have of Vietnam, it is also important to note that few countries affected China's exports. classified as low income by the World Bank have a high Figures 6.6 and 6.7 provide evidence on how the stock of exported products subject to TTBs. imposed exporter-specific trade barriers are affecting a Table 6.4 summarizes key elements of information con- number of exporting economies over time, including dur- tained in figures 6.6 and 6.7. The first column provides data ing the crisis era. Both figures show data on the number on which exporting countries face the greatest total quan- of an economy's exported products subject to a G-20- tity of imposed import restrictions on their traded prod- imposed TTB each year, further broken down into two cat- ucts. By the end of 2009, China's exported products faced egories based on whether the policy-imposing country was roughly four times as many G-20 TTBs as the second-most- a developed or a developing G-20 member economy.15 In affected exporting economy (European Union). Table 6.4 addition to the interesting trends in the data, there are also also provides estimates of the change in the 2009 stock rela- substantial differences in the level of TTB stocks imposed tive to the precrisis year of 2007. The middle column of data across exporting economies. For example in figure 6.6, again shows that, in the aggregate, exporters in developed European Union exporters had 223 (six-digit HS) products economies face roughly the same number of such barriers subject to foreign-imposed TTBs by the end of 2009, in 2009 as they did in 2007. And while the stock of barriers whereas exporters from the United States had only 91 facing all developing economies in the aggregate products subject to foreign-imposed TTBs. increased by 18 percent, those facing all developing coun- Figure 6.6 identifies and examines the main high- tries aside from China increased by only 4 percent. Never- income exporting economies that have been subject to theless, the increase in total combined product-import TTBs over time. As we have already mentioned (with market coverage between 2007 and 2009 was particularly respect to figure 6.3), in the aggregate exporters from pronounced for developing-economy exporters such high-income economies were subject to roughly the same as India (17 percent), Indonesia (25 percent), Thailand stock of imposed TTBs in 2009 as before the onset of the (23 percent), and Vietnam (73 percent). crisis in 2007. The economy-specific panels in figure 6.6 In addition to the exporter incidence associated with also illustrate little heterogeneity during the crisis period the crisis-era growth of such trade barriers, the last col- across these different exporting economies. While the umn of table 6.4 provides summary statistics on the extent overall stock of TTBs imposed on these high-income to which the particular stock of exported products exporting economies has remained relatively flat in recent affected by TTBs in place by 2009 is South-South in 74 Managing Openness Figure 6.6. Exports of High-Income Economies Subject to Selected Temporary Trade Barriers, 1990­2009 a. European Union b. Japan 350 300 number of unique import market­ number of unique import market­ HS06 product combinations HS06 product combinations 300 250 250 200 200 150 150 100 100 50 50 0 0 19 0 19 1 92 19 3 19 4 19 5 96 19 7 19 8 20 9 00 20 1 20 2 20 3 04 20 5 06 20 7 08 09 19 0 19 1 19 2 93 19 4 19 5 96 19 7 19 8 20 9 00 20 1 02 20 3 20 4 05 20 6 20 7 08 09 9 9 9 9 9 9 9 9 0 0 0 0 0 9 9 9 9 9 9 9 9 0 0 0 0 0 19 19 19 20 20 20 20 19 19 19 20 20 20 20 c. Korea, Rep. d. Taiwan, China number of unique import market­ number of unique import market­ 250 200 HS06 product combinations HS06 product combinations 180 200 160 140 150 120 100 100 80 60 50 40 20 0 0 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 98 20 9 20 0 01 20 2 20 3 20 4 20 5 06 20 7 20 8 09 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 19 0 19 1 19 2 93 19 4 19 5 19 6 19 7 98 20 9 00 20 1 20 2 03 20 4 20 5 20 6 07 20 8 09 19 19 20 20 9 9 9 9 9 9 9 9 0 0 0 0 0 0 19 19 19 20 20 20 e. United States f. Other high-income economies number of unique import market­ number of unique import market­ 120 300 HS06 product combinations HS06 product combinations 100 250 80 200 60 150 40 100 20 50 0 0 19 0 91 19 2 19 3 94 19 5 96 19 7 98 20 9 00 20 1 20 2 20 3 04 20 5 06 20 7 08 09 19 0 19 1 92 19 3 19 4 19 5 19 6 97 19 8 20 9 20 0 20 1 02 20 3 20 4 20 5 06 20 7 20 8 09 9 9 9 9 9 9 0 0 0 0 0 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 19 19 19 19 19 20 20 20 20 19 19 19 20 20 stock: developing economy imposition of discriminatory trade barriers stock: developed economy imposition of discriminatory trade barriers Source: Data on the stock of policies imposed and removed over 1988­2009 compiled by the authors from the Temporary Trade Barriers Database. Note: The figures illustrate the number of importing country­product combinations affected by the use of policies such as antidumping, countervailing duties, and China-specific transitional safeguards aggregated over the following G-20 economies: seven developing (Argentina, Brazil, China, India, Indonesia, South Africa, and Turkey) and five developed (Australia, Canada, the European Union, the Republic of Korea, and the United States). The only major G-20 user of such policies not included in the figures is Mexico. The "stock" includes both imposition and removal of import restrictions after termina- tions or sunset reviews. With roughly 5,000 six-digit HS product categories per importing economy and 12 policy-imposing countries, the maximum value that the vertical axis could possibly take is 60,000. "Other" includes Australia; Canada; Croatia; Denmark; the Faeroe Islands; Hong Kong SAR, China; Israel; Kuwait; Liechtenstein; Macau SAR, China; Norway; Oman; Qatar; Saudi Arabia; Singapore; Trinidad and Tobago; and the United Arab Emirates. nature. Overall, other developing economies imposed (85 percent), South Africa (57 percent), and Thailand 52 percent of the TTBs facing developing-economy (60 percent). Only for major emerging-market exporters exporters in 2009. With a (somewhat arbitrary) threshold like Brazil (38 percent), India (37 percent), Russia (40 per- of 90 product-import market combinations as the cutoff cent), and Ukraine (25 percent) did developing economies defining which exporting economies were most severely impose a minority of the TTBs they faced in 2009. And as affected by TTBs as of 2009, developing economies were figure 6.7 illustrates, even for these emerging economies, responsible for more than 50 percent of the TTB-affected over time other developing economies are imposing products for the following exporters above the cutoff: an increasing share of the TTBs that confront their China (58 percent), Indonesia (53 percent), Malaysia exporters. number of unique import market­ number of unique import market­ number of unique import market­ number of unique import market­ HS06 product combinations HS06 product combinations HS06 product combinations HS06 product combinations 0 10 20 30 40 50 60 70 19 0 10 20 30 40 50 60 0 20 40 60 80 100 120 140 160 180 19 19 90 0 100 200 300 400 500 600 700 800 900 1,000 90 90 19 19 19 9 19 19 0 91 91 91 9 19 19 19 19 1 92 92 92 92 19 19 19 19 93 93 93 9 19 19 19 19 3 94 94 94 94 19 19 19 19 95 95 95 9 19 19 19 19 5 96 96 96 9 19 19 19 19 6 97 97 97 97 19 19 19 19 98 98 98 9 19 19 19 19 8 99 99 99 9 20 20 20 20 9 00 0 00 00 e. India 20 c. China 20 20 20 0 a. Argentina 01 g. Malaysia 01 01 01 20 20 20 20 02 02 0 02 20 20 2 20 20 03 03 03 03 20 20 20 20 04 04 0 04 20 20 4 20 20 05 05 0 05 20 20 5 20 20 06 0 06 06 20 20 6 20 20 07 0 07 07 20 20 7 20 20 08 08 08 08 20 20 20 20 09 09 09 09 number of unique import market­ number of unique import market­ number of unique import market­ number of unique import market­ stock: developing economy imposition of discriminatory trade barriers HS06 product combinations HS06 product combinations HS06 product combinations HS06 product combinations 0 10 20 30 40 50 60 70 80 0 10 20 30 40 50 60 70 80 19 19 0 20 40 60 80 100 120 140 160 180 200 0 20 40 60 80 100 120 140 19 19 90 90 9 90 19 19 19 0 19 91 91 91 9 19 19 19 19 1 92 92 9 92 19 19 19 2 19 93 93 93 9 19 19 19 19 3 94 94 9 94 19 19 19 4 19 95 95 95 95 19 19 19 19 96 96 9 96 19 19 19 6 19 97 97 97 97 19 19 19 19 98 98 9 9 19 19 19 8 19 8 99 99 9 99 20 20 20 9 20 00 00 0 00 b. Brazil 20 20 20 0 20 h. Mexico 0 Figure 6.7. Exports of Developing Economies Subject to Selected Temporary Trade Barriers, 1990­2009 f. Kazakhstan 01 d. Indonesia 01 01 20 20 20 1 20 02 02 02 02 20 20 20 20 03 03 0 03 20 20 20 3 20 04 04 0 04 20 20 20 4 20 05 05 05 05 20 20 20 20 06 06 0 06 20 20 6 20 stock: developed economy imposition of discriminatory trade barriers 20 07 0 0 Developing Countries, New Trade Barriers, and the Global Economic Crisis 07 20 20 7 20 7 20 08 08 08 0 20 20 20 8 20 09 09 09 09 (continued) 75 number of unique import market­ number of unique import market­ number of unique import market­ number of unique import market­ 76 HS06 product combinations HS06 product combinations HS06 product combinations HS06 product combinations 0 20 40 60 80 100 120 0 5 10 15 20 25 30 19 19 0 10 20 30 40 50 60 0 20 40 60 80 100 120 140 160 180 19 90 19 90 90 19 90 19 19 9 19 91 91 91 19 1 19 19 92 19 92 92 19 92 19 19 9 19 93 93 19 3 93 19 94 19 19 94 94 94 19 19 19 95 19 95 95 95 19 Figure 6.7. (continued) 19 96 19 19 96 19 96 96 19 19 Managing Openness 19 97 97 19 97 97 19 98 19 19 Source: See source notes to figure 6.6. 98 19 98 98 19 99 19 19 99 20 99 99 20 00 20 20 00 20 00 00 m. Turkey 20 20 i. Pakistan 20 01 o. Vietnam 20 01 01 k. South Africa 01 20 02 20 20 02 20 02 02 20 03 20 20 03 20 03 03 04 20 20 20 20 04 04 04 05 20 20 20 20 05 05 05 06 20 20 20 20 06 06 06 07 20 20 20 20 07 07 07 08 20 20 08 20 20 08 08 09 20 20 09 20 09 09 stock: developing economy imposition of discriminatory trade barriers number of unique import market­ number of unique import market­ number of unique import market­ number of unique import market­ HS06 product combinations HS06 product combinations HS06 product combinations HS06 product combinations 0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 160 19 0 20 40 60 80 100 120 140 160 180 200 19 19 19 90 90 90 90 19 19 19 19 91 91 91 91 19 19 19 19 92 92 92 92 19 19 19 19 93 93 93 93 19 19 19 19 94 94 94 94 19 19 19 19 95 95 95 95 19 19 19 19 96 96 96 96 19 19 19 19 97 97 97 97 19 19 19 19 98 98 98 98 19 19 19 19 99 99 99 99 20 20 20 20 00 00 00 00 20 20 20 20 n. Ukraine l. Thailand 01 01 01 01 20 20 20 20 02 02 02 j. Russian Federation 02 20 20 20 20 03 03 03 03 20 20 20 20 04 04 04 04 20 20 20 20 05 05 05 05 p. Other low- and middle-income economies 20 20 20 20 06 06 06 06 20 stock: developed economy imposition of discriminatory trade barriers 20 20 20 07 07 07 07 20 20 20 Cuba, Dominican Republic, Ecuador, the Arab Republic of Egypt, Georgia, the Islamic Republic of Iran, Kyrgyz Republic, Lebanon, the Former Yugoslav Republic of Macedonia, Malawi, Moldova, Nepal, Nigeria, Paraguay, Peru, the Philippines, Sri Lanka, Uruguay, and República Bolivariana de Venezuela. Note: "Other" includes Albania, Algeria, Armenia, Azerbaijan, Bangladesh, Belarus, Bosnia and Herzegovina, Chile, Colombia, Costa Rica, Côte d'Ivoire, 08 20 08 08 08 20 20 20 20 09 09 09 09 Developing Countries, New Trade Barriers, and the Global Economic Crisis 77 Table 6.4. Major Exporters Subject to Stock of Selected G-20 Temporary Trade Barriers, 2009 Stock of product ­ import % change in (1) relative market combinations subject to precrisis 2007 level Share of (1) to such barriers in 2009 developing G-20 economies imposed on exporter by Exporting economy (1) (2) (3) Developing-country exporters China 884 0.40 0.58 India 159 0.17 0.37 Indonesia 129 0.25 0.53 Thailand 119 0.23 0.60 Brazil 107 ­0.01 0.38 Ukraine 107 ­0.13 0.25 Russian Federation 97 ­0.10 0.40 South Africa 90 ­0.01 0.57 Malaysia 52 0.06 0.85 Vietnam 50 0.73 0.38 Kazakhstan 38 ­0.25 0.84 Mexico 33 ­0.17 0.15 Turkey 17 ­0.46 0.59 Argentina 7 ­0.76 0.14 Pakistan 5 0.00 0.00 Other developing countries 59 ­0.10 0.53 Total developing-economy exporters 1,953 0.18 0.52 High-income exporting economies European Union 223 0.01 0.48 Korea, Rep. 199 0.02 0.60 Taiwan, China 185 0.07 0.52 Japan 138 ­0.07 0.37 United States 91 ­0.04 0.73 Other high-income countries 89 ­0.04 0.75 Total high-income economy exporters 925 0.00 0.55 Source: Data on the stock of policies imposed and removed over 1988­2009 compiled by the authors from the Temporary Trade Barriers Database (Bown 2010a). Note: The table illustrates the number of importing country­product combinations affected by the use of policies such as antidumping, countervailing duties, and China-specific transitional safeguards aggregated over the following twelve G-20 economies: Argentina, Australia, Brazil, Canada, China, the European Union, India, Indonesia, Korea, South Africa, Turkey, and the United States. Mexico is the only major G-20 user of such policies not included, and the reason for its exclusion in this table is described in the text. The "stock" includes both imposition and removal of import restrictions after terminations or sunset reviews. With roughly 5,000 six-digit HS product categories per importing economy and 12 policy-imposing countries, the maximum value that column (1) for any one exporting economy could possibly take is 60,000. The Potential Trade Impact of New TTBs on Exports exported to developed economies, which was only 1.6 per- During the Crisis cent of China's total exports to those economies. As such, a Analogous to the approach adopted above, we move much larger fraction of China's exports to other develop- beyond examining stocks of six-digit HS exported prod- ing economies was subject to new TTBs in 2008­2009 than ucts subject to TTBs and consider the value of exports its exports to high-income economies. potentially affected by the flow of the main G-20 users' A similar phenomenon holds for seven of the next most imposition of new TTBs during 2008­09. frequently targeted countries on the list of developing- As an example, the value of China's 2007 exports that economy exporters in table 6.5. India, Indonesia, Pakistan, would subsequently be subject to new TTBs imposed by Russia, Thailand, Turkmenistan, and Uzbekistan each had the other G-20 members during 2008­09 was US$20.5 bil- a larger share of its 2007 exports to other developing lion, which was 1.8 percent of its total exports to those economies become subject to new TTBs in 2008­09 than economies in 2007 (table 6.5). Of the US$20.5 billion in the share of the 2007 exports sent to high-income exports, US$4 billion was exported to developing economies. In the aggregate, 1.3 percent of the 2007 value economies, which was 3.5 percent of China's total exports of all developing-economy exports to G-20 developing to those economies. In contrast, US$16.5 billion was economies would be subject to new TTBs in 2008­09, 78 Table 6.5. The Value of Exports Subject to G-20 Temporary Trade Barriers Newly Imposed in 2008­09 2007 exports to 2007 exports to G-20 2007 exports to G-20 G-20 subject to % of total developing economies % of total exports developing economies % of total exports new TTBs in exports to to subject to new TTBs to G-20 developing subject to new TTBs to G-20 developing 2008­09 G-20 subject in 2008­09 economies subject in 2008­09 economies subject Exporting economy (US$, thousands) new TTB (US$, thousands) to new TTB (US$, thousands) to new TTB Developing-economy exporters total 23,401,616 0.008 6,140,224 0.013 17,261,392 0.007 China 20,470,000 0.018 3,970,000 0.035 16,500,000 0.016 India 583,729 0.005 575,000 0.022 8,729 0.000 Indonesia 416,200 0.005 358,000 0.016 58,200 0.001 Russian Federation 389,500 0.001 343,000 0.007 46,500 0.000 Moldova 345,000 0.166 -- 0.000 345,000 0.171 Thailand 278,000 0.002 278,000 0.008 -- 0.000 Uzbekistan 186,000 0.038 186,000 0.182 -- 0.000 Vietnam 128,861 0.003 7,861 0.001 121,000 0.003 Pakistan 128,000 0.008 128,000 0.054 -- 0.000 Malaysia 99,000 0.001 99,000 0.002 -- 0.000 Mexico 85,800 0.000 -- 0.000 85,800 0.000 Turkmenistan 69,000 0.042 69,000 0.151 -- 0.000 Brazil 62,400 0.000 45,200 0.001 17,200 0.000 Other developing countries 160,125 0.000 81,161 0.001 78,963 0.000 Total high-income exporting economies 4,720,216 0.001 3,287,316 0.002 1,432,900 0.000 European Union 1,308,000 0.002 1,130,000 0.003 178,000 0.000 United States 1,279,000 0.001 486,000 0.002 793,000 0.001 Taiwan, China 1,094,000 0.005 993,000 0.008 101,000 0.001 Korea, Rep. 552,000 0.002 408,000 0.003 144,000 0.001 Canada 136,032 0.000 32 0.000 136,000 0.000 Other high-income countries 351,184 0.000 270,284 0.001 80,900 0.000 Source: Temporary trade barriers imposed during 2008 or 2009, 2007 import data at the tariff line level from Comtrade matched to tariff line policy data from the Temporary Trade Barriers Database (Bown 2010a). Note: Exports aggregated over the following twelve G-20 economies: Argentina, Australia, Brazil, Canada, China, the European Union, India, Indonesia, Mexico, South Africa, Turkey, and the United States. Korea is the only major G-20 user of such policies not included. The aggregated figures in this table may not match exactly those in table 6.2 due to rounding. -- = not available. Developing Countries, New Trade Barriers, and the Global Economic Crisis 79 whereas only 0.7 percent of their exports to G-20 high- faced an increasingly adverse impact of crisis-era restric- income economies would be affected. These data provide tions on exports. In particular, China increased its tariff on additional evidence of the increasing prevalence of South- salmon from Norway from 10 to 40 percent, which severely South protectionism in the crisis era. restricted Norway's access to the Chinese market, given the very elastic import demand for that product. Other important insights arise from analysis of the MA-OTRI MA-OTRI, given that we can also calculate the indexes at A final approach used to identify which exporting coun- the sectoral level. Table 6.6 decomposes the overall tries are hardest hit by the G-20's changing trade policies changes into compositional changes for the agriculture during the crisis period is to use the market access overall and manufacturing sector MA-OTRIs during the crisis trade restrictiveness index (MA-OTRI) (see table 6.6). period. For agricultural products, increased tariff protec- Analogous to the OTRI described earlier, the MA-OTRI tion in Turkey on meslin and wheat spelt, as well as the measures the average foreign tariff faced by a given coun- European Union's on bananas, severely restricted the mar- try's exporters, taking into account export composition ket access of Kazakhstan, Panama, and Russia. The agri- and the import demand elasticities of the importing culture MA-OTRI of these countries increased from 2 to economies. Kee, Neagu, and Nicita (2010) calculate the 3.7 percentage points, resulting in a trade loss of US$738 changes in the MA-OTRI for a wide range of countries million. The above-mentioned tariff reversal on palm oil over the crisis period due to trade policy changes of G-20 by India explains a large part of Indonesia's MA-OTRI and non-G-20 countries. increase of 7.4 percentage points and the corresponding Consistent with the impact described above that trade loss of $1.5 billion. focused exclusively on TTBs and disregarded tariffs, the Argentina, Bosnia, Botswana, Brazil, Indonesia, Sri exporter most affected by G-20 changes in trade policy Lanka, and Ukraine are other countries aside from China during the crisis was China. China's MA-OTRI increased whose manufacturing exports were adversely affected by by 1.5 percentage points, which translates to an estimated increased barriers. Each of these countries saw its manu- US$28 billion reduction in exports (if the model allows facturing MA-OTRI increase by 0.1 to 0.4 percentage trade to fall more than the 2008 level) or US$5 billion (if points, and jointly their exports were decreased by the model restricts the value of the fall in trade to an US$550 million. For most countries, the main new G-20 amount no larger than the 2008 level). The biggest trade trade barriers came in the form of increased tariffs, while the value impact facing China comes in the form of new G-20 impact on Brazil and Indonesia was the result of additional antidumping duties imposed on manufacturing exports, antidumping duties imposed on manufactured goods. particularly those imposed by developed economies such Finally, while the intensity of G-20 imposed TTBs such as the United States. This finding too is consistent with the as antidumping did increase during the crisis period, for a results of table 6.5. number of exporting economies the adverse impact on Nevertheless, the MA-OTRI results indicate that a num- market access may be offset by simultaneous G-20 tariff ber of other developing countries witnessed severe erosion reductions for other imported product lines. As such, we of access to G-20 markets. Bosnia, Indonesia, Kazakhstan, can reconcile some of the different magnitudes to the esti- and Panama each experienced an MA-OTRI increase in the mates presented in tables 6.5 (TTBs only) and 6.6 (tariffs range of 0.12 to 0.4 percentage points. Most of Indonesia's and antidumping only). For example, table 6.5 indicates trade loss came from India's removal of a temporary tariff that India's exporters were subject to substantially more reduction on palm oil. Kazakhstan and Panama suffered new TTBs and even that 2.2 percent (or US$575 million) of mainly from additional tariff restrictions in the agricul- its 2007 exports to other G-20 developing economies tural sector in China, India, and Turkey. Bosnia lost access would become subject to newly imposed TTBs during the to export markets in manufactured products to the EU. crisis. Nevertheless, table 6.6 indicates that this adverse Bhutan, Botswana, Brazil, the Kyrgyz Republic, Lebanon, impact on market access due to TTBs was at least partially Russia, Sri Lanka, Tanzania, and Ukraine also experienced offset by new market access opportunities for Indian export reductions due to G-20 changes in trade policy, exporters of other products; one specific example turns out with a joint loss in access to export markets of roughly to be due to China's reducing its applied tariff on cotton US$621 million. imports from 40 percent to 6.4 percent. Furthermore, Besides the developing countries, high-income while Vietnam was also adversely affected by new TTBs, its economies such as the European Union; Hong Kong SAR, MA-OTRI is positively affected by China's reducing its tar- China; Israel; Korea; Norway; and the United States also iff from 378 percent to 25 percent for television imports 80 Managing Openness Table 6.6. Examples of Major Changes in MA-OTRI due to G-20 Trade Policy Changes, 2008­09 Aggregate trade Exporting economy Sector 2008 2009 Change effect (US$, thousands) Norway All 0.018 0.049 0.032 ­8,307,226 China All 0.045 0.060 0.015 ­27,500,000 Indonesia All 0.037 0.049 0.012 ­1,676,246 Panama All 0.024 0.031 0.007 ­30,388 Kazakhstan All 0.005 0.011 0.005 ­338,276 Hong Kong SAR, China All 0.026 0.030 0.004 ­388,736 Bosnia And Herzegowina All 0.005 0.009 0.004 ­11,076 Botswana All 0.002 0.005 0.003 ­3,992 Tanzania All 0.023 0.026 0.003 ­6,707 United States All 0.043 0.046 0.002 ­3,688,027 Kyrgyz Republic All 0.027 0.029 0.002 ­1,605 Sri Lanka All 0.052 0.053 0.002 ­20,263 Israel All 0.012 0.013 0.001 ­101,276 Ukraine All 0.030 0.031 0.001 ­111,810 Korea, Rep. All 0.045 0.047 0.001 ­525,618 Brazil All 0.041 0.041 0.001 ­176,619 Russian Federation All 0.009 0.010 0.001 ­298,196 Lebanon All 0.008 0.009 0.001 ­1,185 Bhutan All 0.000 0.001 0.001 ­170 European Union All 0.012 0.012 0.001 ­4,134,099 China MF 0.041 0.056 0.016 ­27,500,000 Norway MF 0.008 0.017 0.010 ­2,092,733 Hong Kong SAR, China MF 0.026 0.031 0.004 ­390,103 Bosnia And Herzegovina MF 0.003 0.007 0.004 ­11,950 Botswana MF 0.002 0.005 0.003 ­3,992 Ukraine MF 0.011 0.013 0.002 ­131,633 Sri Lanka MF 0.051 0.053 0.002 ­19,061 Israel MF 0.009 0.011 0.002 ­107,292 United States MF 0.017 0.019 0.001 ­1,982,447 Brazil MF 0.007 0.008 0.001 ­217,877 Argentina MF 0.007 0.008 0.001 ­43,196 Korea, Rep. MF 0.042 0.043 0.001 ­532,827 Indonesia MF 0.033 0.034 0.001 ­122,079 Maldives MF 0.007 0.007 0.001 ­8 European Union MF 0.010 0.011 0.001 ­3,407,266 Norway AG 0.070 0.218 0.148 ­6,214,493 Indonesia AG 0.062 0.136 0.074 ­1,554,168 Russian Federation AG 0.094 0.131 0.037 ­380,338 Kazakhstan AG 0.019 0.056 0.036 ­325,425 Panama AG 0.029 0.049 0.020 ­32,022 United States AG 0.239 0.248 0.009 ­1,697,471 Malaysia AG 0.059 0.068 0.008 ­175,319 Tanzania AG 0.030 0.037 0.007 ­8,690 Bhutan AG 0.002 0.008 0.005 ­145 Bahrain AG 0.087 0.092 0.004 ­349 Kyrgyz Republic AG 0.046 0.050 0.004 ­1,620 Lebanon AG 0.022 0.025 0.003 ­1,038 Switzerland AG 0.100 0.102 0.002 ­14,096 European Union AG 0.029 0.030 0.001 ­726,839 Canada AG 0.026 0.027 0.001 ­221,993 China AG 0.187 0.188 0.001 ­39,327 Sri Lanka AG 0.054 0.054 0.001 ­1,202 Source: Kee, Neagu, and Nicita 2010. Developing Countries, New Trade Barriers, and the Global Economic Crisis 81 and Russia's lowering of its tariff on imported rice. These new TTBs being imposed by developing economies such as are examples of countries whose overall market access Argentina, India, Indonesia, and Turkey. In many opportunities did not deteriorate over the crisis period by instances, the new TTBs continue a precrisis trend of as much as one might infer by relying solely on estimates of affecting South-South trade: in particular, 68 percent of the their realized loss of market access stemming from newly stock of 2009 TTBs that developing-economy users had in imposed TTBs. The overall impact of G-20 trade policy place were imposed on imports from other developing changes during the crisis era will ultimately be judged economies. In the aggregate, 1.3 percent of the 2007 value by the larger impact of two competing forces: whether of developing-economy exports to G-20 developing exporters receiving new opportunities for market access economies would be subject to new TTBs in 2008­09. This were ultimately able to capitalize on them, or whether percentage is almost twice as much as the only 0.7 percent exporters that faced the imposition of new trade barriers of developing-economy exports to G-20 high-income were unable to identify ways to overcome them and had to economies that would subsequently be subjected to newly reduce sales or exit the market. imposed TTBs during the crisis. One interpretation of these data from the crisis era is that the world trading system somehow better shielded Policy Implications and Conclusions exports from developing economies to the higher-income This chapter has identified a number of stylized facts markets from protectionism. The access of developing- regarding the evolving pattern of import protection associ- country exporters to markets in high-income economies ated with the global economic crisis. Overall, the major turned out to be more "secure" than their access to markets economies of the international trading system--in partic- in other developing economies. Improving the security of ular the G-20 members--largely refrained from using pro- market access associated with South-South trade is an tectionist instruments that had been used during earlier important agenda item for the trading system. One policy crisis eras, such as across-the-board increases in applied implication is the possibility of providing developing- tariffs and the imposition of new quantitative restrictions. country exporters better access to the WTO dispute settle- Instead, most of the new protectionism came in the form ment system to continue to open up desired markets in of potentially WTO-consistent use of temporary trade other developing economies.16 barriers such as antidumping, countervailing duties, and Furthermore, China's exporters present a special case safeguards. for consideration. Overall, imports sourced from China Developing economies can take away important accounted for 80 percent of the total value of trade on insights from the crisis-era implementation of new TTBs which the G-20 imposed new TTBs in 2008­09. The stock in particular. On one hand, the global economy could be of products exported from China subject to G-20 use of heartened by the resilience of the world trading system these TTBs in 2009 increased 40 percent relative to precri- under the WTO. Perhaps the resilience was due to the sis level of 2007. The new TTBs in 2008­09 affected more design of WTO rules allowing for a relatively small inci- than US$20 billion worth of China's exports, or almost dence of new protectionism through permissible TTBs. 2 percent of its (precrisis) 2007 level of exports. The fact Perhaps other factors, such as the globalization of supply that China's exports are subject to TTBs is not a new crisis- chains, have effectively reduced the threat of protectionism era phenomenon (Bown 2010c), but it is one that may have stemming from traditional political-economic forces. been heightened by the crisis. In part, the endogeneity of Whatever the cause, the new protectionism that emerged the G-20 policy response is likely affected by China's con- in 2008­09 was certainly not as bad as it might have been. tinued export successes even during the crisis. Overall, by 2009 the stock of products affected by G-20 use Finally, it is worth noting that the final cross-country of such TTBs had increased by 25 percent over those in pattern as well as the depth of the new crisis-era protec- place in 2007. Even this 25 percent increase in affected tionism is not yet completely known. Notwithstanding the products in the aggregate is estimated to affect less than 0.3 possibility of a further deepening of the global economic percent of total trade. recession begun in 2008 that may lead to a substantial On the other hand, it is also clear that the limited inci- increase in the flow of new government-conducted TTB dence of protectionism that did take place during the cri- investigations, the amount and distribution of G-20 sis was developing economy­centric in nature: it was import protection after the crisis will be the result of two disproportionately imposed by developing economies on yet-to-be-resolved policy questions. First, the postcrisis developing economy exporters. Policy-imposing countries stock of TTBs will partly reflect dozens of forthcoming show substantial differences, with the largest increases in government policy-making decisions over whether to 82 Managing Openness impose new TTBs that have yet to be concluded. Second, 8. Just as in figure 6.3, figures 6.4 and 6.5 are again limited to captur- ing only these economies' use of the three exporting-country specific poli- the postcrisis stock of TTBs that we have described cies (antidumping, countervailing duties, and China-specific safeguards) throughout is only "temporary" if they are someday and thus the figures omit any use of global safeguards. removed. Policy makers will ultimately be responsible for 9. Niels and Francois (2006) provide an empirical analysis of Mexico's how the postcrisis trading system responds to the TTBs earlier antidumping use. See also Finger and Nogués (2005). 10. Knetter and Prusa (2003) link macroeconomic determinants such now imposed--whether they are removed through WTO as recessions and exchange rate fluctuations to the earlier use of dispute settlement or countries' adherence to sunset antidumping by Australia, Canada, the EU, and the United States. reviews and safeguards expirations--and this too will also 11. While comparable to the first estimates of this issue contained in table 7.4 of Bown (2009a), these figures are "improved" estimates of substantially affect the legacy of crisis-era protectionism.17 potential impacts for the following reasons. First, whereas the results in Bown (2009a) covered all newly initiated investigations between first quarter (1Q) 2008 through 1Q 2009, this table reports all temporary trade Notes barriers imposed (preliminary and final) between 1Q 2008 and 4Q 2009. 1. In our discussion of separate estimates using the overall trade Second, the results in Bown (2009a) were estimated from 6-digit Har- restrictiveness index (OTRI) and market access overall trade restrictive- monised System (HS) level data, whereas the results above are computed ness index (MA-OTRI) described later, we also examine the impact of the from actual tariff line import data (at the 8,9,10, or 12 digit level, limited incidence of increases in applied tariffs. Nevertheless, aside from depending on the reporting convention to UN Comtrade of the importing antidumping, countervailing duties, and safeguards, we do not examine economy). the potential trade impact of other examples of "murky" nontariff barri- 12. When comparing the magnitude and distribution of the esti- ers to trade taking place during the crisis (Baldwin and Evenett, 2009; mated trade impacts of crisis-era policies, we should note that the exer- Evenett, Hoekman, and Cattaneo, 2009). For example, we leave to future cises reported here drawing from Kee, Neagu, and Nicita (2010) and work the more difficult task of assessing the trade impact of subsidies or Bown (2010b) are not strictly comparable for a number of subtle, data- government bailouts during the crisis, many of which are captured in the driven reasons. First, in constructing estimates for potential trade impact, Global Trade Alert (Evenett 2009) database. Bown uses 2007 import data while Kee, Neagu and Nicita use 2008 import 2. This chapter draws heavily from two separate pieces of research to data. Second, Bown does not examine tariff changes but does examine which the reader should refer for more detailed discussion, especially antidumping, countervailing duties, global safeguards, and China-specific regarding methodology. See Bown (2010b) and Kee, Neagu, and Nicita safeguards, while Kee, Neagu, and Nicita examine tariff changes and (2010). antidumping (but not countervailing duties, global safeguards, and 3. Baldwin (2009) presents a set of early research examining likely China-specific safeguards). Third, Bown relies on policies imposed in all culprits behind the trade collapse of 2008­09, most of the evidence point- of 2008 and 2009, whereas Kee, Neagu, and Nicita examine tariff and ing toward an adverse demand shock. See also Freund (2009a, 2009b). antidumping duties imposed between June 2008 and September 2009 4. Antidumping has historically been the most prevalent of these four only. Despite these slight differences in approach, the results are broadly policy instruments (Prusa 2001; Zanardi 2004). As such, there is well- consistent in magnitude of effects across policy-imposing economies and established theoretical and empirical literature examining determinants the distribution of the incidence across exporting countries. and impacts of the use of antidumping across countries. Recent empirical 13. Turkey is one of the few economies for which newly imposed contributions, including examinations of developing economies, using global safeguards may have also had a large trade impact during the crisis, detailed data provided in early versions of the Temporary Trade Barriers and these are not captured in the OTRI estimates of table 6.3. While such Database include Egger and Nelson (forthcoming), Moore and Zanardi policies are captured in table 6.2, the table 6.2 estimates would not capture (2009), Vandenbussche and Zanardi (2010), Bown (2008), Reynolds the impact of Turkey's new tariff impositions in the agricultural sector (2009), and Bown and Tovar (forthcoming). For a survey of the literature that are captured in table 6.3. on antidumping, see Blonigen and Prusa (2003). 14. These results would be reinforced by inclusion of estimates of the 5. The yearly stock is computed through examination of all initiated September 2009 U.S. imposition of the China-specific safeguard on investigations, the dates of imposition of the first trade barriers during imports of Chinese tires. (preliminary) or at the end of a (completed) investigation, and the date of 15. To be consistent with the analysis presented above, again we removal for investigations. There are roughly 5,000 six-digit HS product exclude Mexico and thus focus on the other 12 major TTB-using G-20 categories per importing economy. economies. 6. The twelve G-20 members included in figures 6.1 and 6.2 are 16. Bown (2009b) presents a set of proposals describing how devel- Argentina, Australia, Brazil, Canada, China, European Union, India, oping-country exporters might use WTO dispute settlement to better Indonesia, the Republic of Korea, South Africa, Turkey, and the United secure their access to other developing-country markets. States. G-20 economy member and TTB user Mexico is not included in 17. While safeguards have typically been removed as scheduled, the the aggregated figures 6.2 and 6.3 for reasons described below. Japan, the evidence on timely removals for antidumping is much less convincing Russian Federation, and Saudi Arabia are the only G-20 economies not (Moore 2006; Cadot, de Melo, and Tumurchudur 2007). represented in the analysis of the use of these temporary trade barriers since they did not actively use such policies during this time period. G-20 member countries France, Germany, Italy, and the United Kingdom are Bibliography not included separately because their trade policy is determined by the European Union, the 20th member of the G-20. Even though many of Baldwin, R., ed. 2009. The Great Trade Collapse: Causes, Consequences and these economies' use of TTBs started much earlier, we begin in 1997 Prospects. VoxEU.org. E-book. because that is the time period in which each of the 12 G-20 economies in Baldwin, R., and S. J. Evenett, eds. 2009. The Collapse of Global Trade, the sample were using at least one of their TTB policies, China being the Murky Protectionism, and the Crisis: Recommendations for the G20. last of the 12 as it adopted use of antidumping in 1997. VoxEU.org. E-book. 7. A third factor is that the set of underlying policies is also not iden- Blonigen, B. A., and T. J. Prusa. 2003. "Antidumping." In Handbook of tical, since figure 6.2 includes the products affected by the use of the International Trade, ed. E. K. Choi and J. Harrigan, 251­84. Oxford: global safeguards policy whereas figure 6.3 does not. Blackwell Publishers. Developing Countries, New Trade Barriers, and the Global Economic Crisis 83 Bown, C. P. 2008. "The WTO and Antidumping in Developing Countries." Freund, C. 2009a. "The Trade Response to Global Crises: Historical Economics and Politics 20 (2): 255­88. Evidence." Policy Research Working Paper 5015, World Bank, ------. 2009a. "The Global Resort to Antidumping, Safeguards, and Washington, DC. other Trade Remedies amidst the Economic Crisis." In Effective Crisis ------. 2009b. "The Trade Response to Global Downturns." In The Great Response and Openness: Implications for the Trading System, ed. Simon Trade Collapse: Causes, Consequences and Prospects. VoxEU.org. J. Evenett, Bernard M. Hoekman, and Olivier Cattaneo, 91­118. E-book. London: Centre for Economic Policy Research and World Bank. Freund, C., and M. D. Horenstein. 2010. "Trade Watch Data." World Bank, ------. 2009b. Self-enforcing Trade: Developing Countries and WTO Dis- Washington, DC. http://go.worldbank.org/EWEDUHSI50. pute Settlement. Washington, DC: Brookings Institution Press. Irwin, D. A. Forthcoming. Peddling Protectionism: Smoot-Hawley and the ------.2010a. "China's WTO Entry: Antidumping, Safeguards, and Dispute Great Depression. Princeton, NJ: Princeton University Press. Settlement." In Handbook of International Trade, ed. Robert C. Feenstra Kee, H. L., I. C. Neagu, and A. Nicita. 2010. "Is Protectionism on the and Shang-Jin Wei. Chicago: University of Chicago Press for the NBER. Rise? Assessing National Trade Policies during the Crisis of 2008." ------. 2010b. "Taking Stock of Antidumping, Safeguards, and Counter- Policy Research Working Paper 5274, World Bank, Washington, vailing Duties, 1990­2009." Policy Research Working Paper 5436, DC. World Bank, Washington, DC. Knetter, M. M., and T. J. Prusa. 2003. "Macroeconomic Factors and ------. 2010c. "Temporary Trade Barriers Database." World Bank, Antidumping Filings: Evidence from Four Countries." Journal of Inter- Washington, DC. http://econ.worldbank.org/ttbd. national Economics 61 (1): 1­17 Bown, C. P., and P. Tovar. Forthcoming. "Trade Liberalization, Antidump- Moore, M. O. 2006. "An Econometric Analysis of U.S. Antidumping ing and Safeguards: Evidence from India's Tariff Reform." Journal of Sunset Review Decisions." Weltwirtschaftliches Archiv 142 (1): Development Economics. 122­50. Cadot, O., J. de Melo, and B. Tumurchudur. 2007. "Anti-dumping Sunset Moore, M. O., and M. Zanardi. 2009. "Does Antidumping Use Contribute Reviews: The Uneven Reach of WTO Disciplines." Working Paper to Trade Liberalization in Developing Countries?" Canadian Journal 6502, Centre for Economic Policy Research, London. of Economics 42 (2): 469­95. Egger, P., and D. Nelson. Forthcoming. "How Bad Is Antidumping? Evi- Niels, G., and J. Francois. 2006. "Business Cycles, the Exchange Rate, and dence from Panel Data." Review of Economics and Statistics. Demand for Antidumping Protection in Mexico." Review of Develop- Eichengreen, B., and D. A. Irwin. Forthcoming. "The Great Depression ment Economics 10 (3): 388­99. and the Protectionist Temptation: Who Succumbed and Why?" Jour- Prusa, T. J. 2001. "On the Spread and Impact of Antidumping." Canadian nal of Economic History. Journal of Economics 34 (3): 591­611. ------. 2009a. "The Protectionist Temptation: Lessons from the Great Reynolds, K. M. 2009. "From Agreement to Application: An Analysis of Depression for Today." VoxEU. March 17. http://voxeu.org. Determinations under the WTO Antidumping Agreement." Review of Evenett, S. J. 2009. "Global Trade Alert: Motivation and Launch." World International Economics 17 (5): 969­85. Trade Review 8 (4): 607­09. Vandenbussche, H., and M. Zanardi. 2010. "The Chilling Trade Effects of Evenett, S. J., B. M. Hoekman, and O. Cattaneo, eds. 2009. Effective Crisis Antidumping Proliferation." European Economic Review 54 (6): Response and Openness: Implications for the Trading System. London: 760­77. World Bank and Centre for Economic Policy Research. WTO (World Trade Organization). 2009. International Trade Statistics Finger, J. M., and J. J. Nogués, eds. 2005. Safeguards and Antidumping in 2009. Geneva: WTO. Latin American Trade Liberalization: Fighting Fire with Fire. New York: Zanardi, M. 2004. "Antidumping: What Are the Numbers to Discuss at World Bank and Palgrave. Doha?" World Economy 27 (3): 403­33. 7 Turning toward China? The 2008 Crisis and Its Influence on Brazil's Development Model Gustavo H. B. Franco and Fausto J. A. Vieira This chapter addresses issues pertaining to the advancement Trends in Brazilian Trade Orientation and style of Brazilian economic development in the years to before the 2008 Crisis come in the aftermath of the 2008 crises and the lessons it An outward-oriented policy stance could have been a natu- entails. Globalization, the distinguishably superior perform- ral casualty of a worldwide financial crisis originating in ance of export promotion over import substitution policies, the world's core economy, and with a variety of aggravat- and especially the rise of China are factors that overwhelmed ing circumstances coming mostly from the developed the debates on openness and on trade and industrial policies world. This possibility was all the stronger for a country to support economic development. Against the background like Brazil, which has a long tradition of inward-oriented of the consequences of the first systemic crisis of the global- economic growth. "Turning inward" would not be in disso- ized economy, policy makers in Brazil are facing hard choices nance with what happened from 1929 onward in a similar on the next steps in economic development. situation of global crisis. Carlos Diaz Alejandro has said of The analysis proceeds as follows. The next section that period that developing nations will never forgive devel- briefly reviews the reform record and some of the singular oped economies for the massive betrayal of the Ricardian features affecting Brazilian openness during the crucial theories they always preached. Yet, despite the precipitous years of stabilization and redefinition of the Brazilian fall in international trade in 2008, the world did not wit- development model. The next section chronicles events ness a "disintegration" of globalization in the sense of a and policy responses after the global panic produced by the policy-reinforced rush to disconnect from the global econ- demise of Lehman Brothers. We start from a discussion of omy as was seen in the 1930s. Much to the surprise of some the reasons why, for Brazil, some external shocks turn into observers, that kind of de-globalization has gained little or severe crises while others go almost unnoticed. The section no traction in Brazil. then provides a detailed analysis of the transmission mech- It is true that there are isolated signs of protectionist anisms functioning in this crisis, some quite novel, on practices and mercantilistic leanings and some complaints which we will focus. In the last section, we discuss the way about "asymmetries" or anticompetitive practices, with ahead, beyond "exit strategies" from the exceptional meas- regard to China in particular. But the magnitude of ures deployed during the more acute moments of the crisis. exchange-rate fluctuations overshadows the importance of In fact, in some cases atypical measures were reversed the discussion on competitiveness differentials and, for during the crisis, while other policy measures advanced that reason as well, of the discussion of more structural and even accelerated. Finally, we draw conclusions about ways to develop innovation and comparative advantage. It the future of Brazilian development. 85 86 Managing Openness is as if the capital account, or short-term financial factors, Although there is no sign of a crisis-related backlash in at least in times of turmoil, have become a dominant favor of import substitution, protectionism, and other influence over the "factory level" elements traditionally practices of the past, that is not to say that there have been associated with trade and industrial policies. As turmoil no changes in development thinking. In fact, the crisis appears to be all the more frequent, one wonders what is could be a decisive element in swinging the balance of ideas left to discuss in the field of trade and industrial policies in toward more "Chinese-inspired" policies that are in sym- this brave new world. pathy with past practices and institutions in Brazil and For all the talk of globalization, Brazil remains one of whose suitability is very much contested. the least-open economies on the planet, at least as far as Before extending the argument, let us take a step back ratios of trade to gross domestic product (GDP) are con- and understand first how firmly Brazil had come to cerned (Figure 7.1). It has a contentious track record on embrace globalization in the past two decades. Several trade liberalization and an extended history of inward- good arguments explain these new attitudes, nearly all oriented industrialization strategies. Yet, the crisis was rela- related to the hyperinflation experience ending in 1994, tively mild in Brazil, thanks to its domestic market. which many saw as a clear indication of the exhaustion of The crisis brought the usual plethora of shocks related the old development model. Starting from the late 1980s, to the balance of payments--collapsing export demand when defenders of an import substitution strategy could and prices, disappearance of trade credit lines, sudden stop still be found, hyperinflation and stabilization policies had of capital flows, repatriations, and dividend remittances. The an overwhelming influence in all spheres of thought. Ulti- response was a major exchange-rate depreciation, some mately, this attitude helped establish the concept that deployment of international reserves to sustain trade flows, extreme inflation was but a clear expression of the collapse and some action in the field of derivatives to prevent of the old inward-oriented development model based on excesses that could further destabilize the currency. An addi- "heterodox" (so-called Keynesian) notions of public tional impact on Brazilian banks, to be examined in more finance and heavy protectionism. This approach proved detail below--although not directly related to the banking to be so powerful that the design of policies to end hyper- stress going on in the Northern Hemisphere--was perhaps inflation and stabilize the economy turned into a broad the key element in explaining the unusually sharp drop in program of reforms affecting every aspect of the old industrial production and GDP in the last quarter of 2008. If model. it were not for this accident, the picture of economic growth The strength of this revisionist thinking was propor- in Brazil in 2009, where it was near zero, could look much tional to the size and resilience of inflation. According to more like the stronger outcomes in India and China. the Fischer, Sahay, and Végh (2002) redefinition of "high Figure 7.1. Trade to GDP for Selected Economies, 1995­2000 180 trade to GDP (openness in constant prices, average) 160 140 120 100 80 60 40 20 0 Ar Jap zil nt n ite In na St ia Co P es lo eru st ia ia U Ch ia Fe ug a ut rat y Af n G rica Fr ece Bo ce In It ia ne y Sp ia Tu ain m y ew Po any al d Ve d C nd zu g ile a m re ma . a, rk rt p. e l Fi xico or d n y e a ra n Pa Isr e Ph rag ael pp y Ta T us s an ila a , C nd Cz et t a ec he nam p ds l ic lg ia Ir ium al d sia M uga Ko en . B A ne de ua do al er e Ca wa ili ua n r in Sw ad iw ha tri N Vie hin in ge a h io Uk de Ze lan N lan M elan Bu ubl d d Au mb l liv s Be gar Po Re el do at a ra ne Kin h Re n an G rk i ay a Br D ,R i re h rla n So Un ite N Un ss Ru Source: Penn World Tables 6.1 (database), Center for International Comparisons, University of Pennsylvania, http://pwt.econ.upenn.edu/. Turning toward China? The 2008 Crisis and Its Influence on Brazil's Development Model 87 inflation" as 100 percent or more on a 12-month rolling No doubt crises in 1994, 1997, 1998, 1999, 2001, and basis, Brazil started its high inflation episode in April 1980 2002--some of which were made more serious by inci- and finished it 182 months later in June 1995. In this dents linked to the stabilization process--may have pre- period, accumulated inflation reached the extraordinary vented more substantial gains from reforms and from the number of 20,759,903,275,651 percent--equivalent to an new policy stance. Promises related to growth and compet- average monthly inflation rate during these 15 years of itiveness, all crucial to opening the political gates to large- nearly 16 percent. It is hard to argue that intelligent eco- scale reforms, seemed unfulfilled. It appeared that the nomic life can exist in such conditions, let alone healthy reform program had stopped short of full implementation economic growth. The 1994 monetary reform leading to a or, alternatively, that its allegiance to the canons of the so- successful stabilization--the Real Plan--marked a major called Washington Consensus was misdirected. Ultimately, turn toward market-driven mechanisms in trade and good results in growth and productivity are the elements industrial policies as opposed to anything Brazil had expe- around which one builds the constituencies supportive of rienced in the past. reforms. Otherwise, reformers are bound to political weak- A major wave of liberalizing measures followed stabi- ness, as was indeed the case in Brazil. lization, along with ambitious deregulation and privatiza- Extensive soul-searching inquiries were conducted tion programs. To many, this process was a kind of "fall of while the leftist administration of President Luis Inacio the Wall": reforms seemed as profound and far reaching Lula da Silva came to power in 2003. Its policies, however, as those seen in Eastern Europe. Yet, macroeconomic insta- did little to settle the issue. There was no sign of a move bility, most notably affecting exchange rates, prevented away from conventional macroeconomic policies. But nei- price signals produced by the milder structure of protec- ther was there any sign of a backlash against the most con- tion from having their full impact on productivity and tentious reform processes associated with the Washington competitiveness. Openness improved, though not in a very Consensus, foremost among which were privatization and striking way, as seen in figure 7.2. Despite this progress, trade liberalization. No further advancements, however, Brazil remained stuck with dismal levels of openness as were considered. Reforms came to a halt altogether, as if compared to other countries, as seen in figure 7.1. the country needed to heal from the political wounds of The magnitude of real exchange-rate variations may the aggressive reforms conducted during the Cardoso have been one explanation for the good yet unimpressive years. The maintenance of the policy stance, most notably a record of productivity and GDP growth, particularly the primary surplus amounting to 4 percent of GDP, especially indications on the correlation between total factor produc- in the absence of external shocks, led to an extended period tivity (TFP) and openness suggested in table 7.1. of good macroeconomic performance produced mostly by increased credit and personal consumption. Many wel- come signs of a "crowding-in" phenomenon were visible and pointed to promising new directions for the composi- Figure 7.2. Openness Indicators, Brazil, 1960­2009 tion of growth. 30 Growth was significantly higher than observed during the Cardoso years: 4.4 percent on average in 2002­09, 25 against 1.8 percent on average for 1995­2001. This differ- 20 ence obviously reflects the fact that Cardoso carried the load of fighting hyperinflation and was hit by three exter- ratio (%) 15 nal crises--worse results, nevertheless. Little change could 10 actually be seen in the rates of aggregate investment in the two presidencies: gross fixed capital formation remained 5 under 20 percent of GDP, as seen in figure 7.3, throughout 0 the 1990s and 2000s. Insufficient aggregate investment has been constantly 60 00 05 65 90 95 70 75 85 80 19 20 20 19 19 19 19 19 19 19 highlighted as the most important obstacle to sustained export share of GDP high growth, and the contrast to China highlighted import share of internal absoprtion openness (import + export as % of GDP) in figure 7.3 is anything but accidental. During the years of military rule, public savings were at a historical high, Source: Penn World Tables 6.1 (database), Center for International Comparisons, University of Pennsylvania, http://pwt.econ.upenn.edu/; adding to private savings to sustain investment rates over Bloomberg, www.bloomberg.com. 25 percent of GDP. With additional help from "external 88 Managing Openness Table 7.1. GDP Growth and Various Estimates of Labor and Total Factor Productivity Growth in Brazil, 1985­2007 average annual growth rates, % Source 1985­92 1993­97 1998­2002 2003­07 Measurement GDP growth 2.3 4.0 1.7 3.6 Rossi and Ferreira (1999) ­2.49a 2.15a n.a. n.a. TFP 1.1b 6.21/7.97c n.a. LP Ferreira (2001) ­0.48 7.45 n.a. n.a. LP ­1.03 2.0/4.3 n.a. n.a. TFP Pinheiro et al. (2001) ­0.7­0.65d 2.1­2.6d n.a. n.a. TFP Bonelli (2002) ­0.68e 7.19/8.31e n.a. n.a. LP Authors' estimate ­2.3 1.6 ­0.7 0.9 TFP Gomes, Pessoa, and Veloso (2003) ­2.0/­2.9f 1.0/0.1f n.a. n.a. TFPf Souza (2007) -- 1.35/1.69g ­0.45/­0.16g 0.94/1.67g TFP Sources: Rossi and Ferreira 1999; Ferreira 2001; Pinheiro, I. Gill, L. Severn, and Thomas 2001; Bonelli 2002; Gomes, Pessoa, and Veloso 2003; Souza 2007. Note: LP = labor productivity; TFP = total factor productivity; n.a. = not applicable; -- = not available. a. 1985­90 and 1991­97. b. 1985­89, c. 1990­93 and 1994­97. d. Extreme values on sectoral estimates, 1981­93 and 1994­2000. e. 1985­90, 1990­95, and 1995­2000. f. 1976­1992 and 1992­2000, estimates for TFP and "discounted" TPF, capturing shifts at the technological frontier. g. Estimates with reference to the national accounts of 1985 and 2000 for quarter 1, 1992­quarter 4, 1997; quarter 1, 1998­quarter 3, 2003; and quarter 4, 2003­quarter 4, 2006. Figure 7.3. Investment and Growth of GDP in China and Brazil, 1967­2009 50 14 45 12 Growth (annual change, moving 40 10 investment (% of GDP) average of four years) 35 8 30 6 25 4 20 2 15 10 0 5 ­2 19 7 69 71 19 3 19 5 19 7 79 19 1 19 3 19 5 19 7 89 19 1 19 3 95 19 7 99 20 1 20 3 20 5 20 7 09 6 7 7 7 8 8 8 8 9 9 9 0 0 0 0 19 19 19 19 19 19 20 GDP growth, Brazil investment, Brazil GDP growth, China investment, China Sources: Penn World Tables 6.1 (database), Center for International Comparisons, University of Pennsylvania, http://pwt.econ.upenn.edu/; Bloomberg, www.bloomberg.com. savings"--that is, current account deficits--Brazil lived its not last long, though. The transition to democracy started in "economic miracle" during these years. The similarities of the first half of the 1980s and was expedited by the deterio- the growth model, mechanisms, and discourse with con- rating performance of the economy, especially after the hit temporary China are remarkable and at the same time puz- taken in 1982 with the Mexican and Brazilian moratoria. zling. The average annual GDP growth rate for Brazil in The first civilian government was inaugurated in 1985, 1967­78 was 9.5 percent, the very same rate observed in with inflation running at 215 percent for the calendar year China for 1990­2000. The Brazilian economic miracle did 1984. When President Sarney's administration left office in Turning toward China? The 2008 Crisis and Its Influence on Brazil's Development Model 89 March 1990, inflation reached an all-time high of 83 per- address so that private investment can approach levels cent in that month. During these incredible five years of observed in emerging Asia. Much progress was made, the so-called New Republic, an explosion of social however, in reducing the size of the problem through the demands all converged on the budget. They found expres- years, and many would explain the flamboyant behavior of sion in the 1988 Constitution, which led to a fiscal crisis Brazilian capital markets up to late 2008 as an indication and hyperinflation. This episode was a rare case of peace- of the beginning of a crowding-in process. Yet, resistance time hyperinflation, of which the most visible cause was to further improvements in fiscal policies drove the econ- the inconsistency between social demands expressed in omy toward overheating and toward renewed inflationary sharply increased government consumption (mostly pressures. The central bank was indeed initiating a tight- related to social security, health care, education, and social ening cycle when the Lehman Brothers event changed overheads, all summarized by the expression social debt), everything. ambitious public investment programs, and society's stiff resistance to raising the tax burden. These mechanisms The 2008 Crisis: Impacts and Responses have been extensively discussed elsewhere (see, for instance, Franco 1999) and the issue to note for our pur- At first glance, the 2008 crisis was an external shock and, in poses is that, after the Real Plan, a new "growth equation" this regard, was very similar to others that had taken place would have naturally evolved if private savings and invest- on several occasions in the past. Some of these external ment made up for the eroded savings and investment capa- shocks turned into crises, understood as financial distress bility of the public sector. Budgetary, fiscal, social security, always accompanied by significant adverse changes in GDP and tax reforms, along with extensive internal debt and growth. Other shocks did not, and an interesting issue is to restructuring of public banks, addressed the fiscal crisis investigate factors that leverage the impacts of such events. and set the stage for a crowding-in process. Privatization It light of the myriad external shocks experienced was crucial in transferring investment responsibilities to through the years--1974, 1979, 1982, 1987, 1989, 1991, the private sector in important sectors: petrochemicals, 1994, 1997, 1998, 1999, 2001, 2002 and 2008--it is inter- steel, mining, telecommunications, and electricity, among esting to ask when and why they "cause" a recession. As a others. To judge from the huge magnitude of investment prelude to the discussion of the 2008 shock and ensuing programs of privatized enterprises, this was by far the most recession, we ran two equations to offer a summary of past important contribution privatization has made to the long- experience, as shown in table 7.2. They are probit and logit term development of the country, much beyond the debt equations assessing the factors affecting the probability of cancelations derived from direct privatization revenues. a recession for the reference period. The years following the success in stabilization saw The coefficients reveal a curious "counter-Keynesian" developments in all these fields, but they were not broad result, whereby higher fiscal deficits increase the likelihood enough to produce any significant increase in private of a recession, although we cannot rule out reverse causal- savings and investment, as seen in figure 7.3. In fact, the ity here. An important point is that an increase in the continuous difficulty in improving the fiscal situation deficit of one percentage point does not make a huge dif- maintained the economy in a chronic crowding-out situa- ference in the range between a fiscal surplus and a deficit of tion, whereby the continuous growth of the private econ- 3 percent. But a one percentage point increase above the omy was not matched by budgetary improvements to allow threshold of a 3 percent deficit increases the probability of a changed aggregate demand composition. Unusually high a new recession by 22 percent. More important, the interest rates--as a matter of fact, the highest in the strongest element reducing the probability of a recession is world--offered a clear expression of the problem and the ratio of international reserves to external debt. It works represented, unquestionably, a leftover from the hyperin- like an insurance policy: when the international reserves flation years. It was as if Brazil had succeeded in ending the ratio is low (0­20 percent), increasing it by one percentage infection producing hyperinflation, but the antibiotics did point reduces the probability of a recession by 2 percent. As not completely eradicate fiscal malpractices that remained in the 1980s, the mix of low reserves and a high current in the system. Although the remaining malpractices were account deficit is conducive to frequent recessions (at the on a much smaller scale, they were large enough to produce time designated as "stop and go" policies). Nevertheless, a classic crowding-out situation, maintaining interest rates when the international reserves ratio is higher than 20 per- at levels that were and remain exceedingly high. No wonder cent, increasing it by one percentage point reduces by investment ratios are so small in the country with the high- 0.5 percent the probability of a recession, instead of 2 per- est real interest rate in the world; this is the challenge to cent when the ratio is lower. Another point is that when the 90 Managing Openness Table 7.2. Estimation Results, Model of Recession, and and sharply increased country risk premium. The direct GDP Growth in Brazil impact on the exchange rate was brutal: from September Recession dummy GDP growth onward, in 45 days the real lost approximately 45 percent Probit Logit OLS of its value. Some very unusual influences were behind such a big Global growth ­0.82 ­1.49 0.48 depreciation of the currency, and the problem was, unfor- (0.02) (0.05) (0.04) tunately, connected to foreign-exchange derivatives. The Reserves to external debt ­0.05 ­0.09 0.04 surprisingly wide dissemination of "toxic" derivatives (0.09) (0.10) (0.03) products affecting balance sheets of nonfinancial corpora- Fiscal deficit 0.38 0.64 0.00 tions of all sizes multiplied the impacts of the real depreci- (0.03) (0.05) (0.98) ation on Brazilian companies in novel and dangerous d (saving ratio) ­0.25 ­0.44 0.40 ways. Some interesting parallels are found here with what (0.03) (0.05) (0.03) happened in the United States related to asymmetry of d (terms of trade) 0.00 ­0.01 0.22 (0.98) (0.91) (0.00) information: some banks sold highly dangerous products Constant 1.63 3.03 ­0.25 to their clients with little explanation of implicit risks, or (0.24) (0.22) (0.89) with outright misrepresentation, and buyers lacked the Number of skills to discuss the products or the capability to manage observations 29 29 29 the risk, or they simply acted on trust. The template is very R-square 0.49 0.49 0.52 much the same for the structured products based on sub- Source: Authors' calculations. prime mortgages in the United States and for the deriva- Note: t-statistics are in parentheses. Recession dummy is a variable for Brazilian recessions as indicated by Fundacao Getulio Vargas. tives based on exchange rates in Brazil. OLS = ordinary least squares. The scant available evidence indicates that a few hun- dred companies were offered to launch far out-of-the- world is in a recession or growing below trend, an increase money put options, which they sold to the banks from of one percentage point in global growth reduces by 18 per- which they drew their working capital financing. Thanks to cent the probability of a new Brazilian recession. When the the premium of these options, companies could reduce the world is growing above trend (higher than 2.5 percent), the costs of bank financing in exchange for the puts they sold, impact is much lower: an increase of one percentage point albeit with a very clear underestimation of the risks in global growth reduces the probability of a Brazilian involved. Notional amounts were huge and potential losses recession by just 1.3 percent instead of 18 percent. unlimited, if the exchange rate were to depreciate enough The third equation is an ordinary least squares version to reach the strike prices. If not, it appeared to be only a of the first two but with economic growth as the dependent sophisticated mechanism to reduce the cost of capital. variable. The importance of international reserves in this With the unusually large devaluation taking place from equation can be explained only if one accepts that instabil- September 2008 onward, and at a very rapid pace, the dis- ity produced by external shocks has had such an over- aster materialized, and these options reached their strike whelming effect on growth that its antidote, reserves, points. Many prime companies entered acute distress, seems to be more important at first glance than any other especially in the case of listed companies, where the size of long-run structural elements. In a less shock-prone econ- the exposure had to be made public, with truly devastating omy, this result would appear surprising, but in a shattered effects for those involved: Aracruz, Sadia, and Votarantim emerging economy with a long history of macroeconomic were the outstanding examples of massive value destruc- instability, it makes perfect sense. tion, while a few hundred other nonpublic companies in External shocks are usually thematic, often related to the same situation were able to negotiate quietly with their geographies (Asia, the Russian Federation, Mexico, and so creditors. There was no precise account of how many com- forth) or to specific issues (oil, commodities, or hedge panies and how much was involved in such operations. funds in trouble). This time, the theme was broad and Some of the largest banks reported some features of their far reaching: banks. The worldwide deleveraging and exposures, with numbers varying from US$500 million to risk-reduction movements produced sudden stops in US$1.5 billion in the three banks reporting on these deals. almost all types of capital inflows, sharp terms-of-trade These disclosures unquestionably represented only a frac- losses, falling external demand, repatriations of portfolio tion of the size of the problem. The Aracruz situation investments, and even some disruption in foreign trade alone (the equity value of the company was around mechanisms because of the disappearance of trade lines US$9.6 billion in June 2008) involved losses of approximately Turning toward China? The 2008 Crisis and Its Influence on Brazil's Development Model 91 US$8 billion and presumed notional amounts between 5 and although for a short time only. No other casualties 10 times this. How exposures of this order of magnitude occurred in this episode. could be entered without appearing on any regulator's radar During the second half of 2008, and very clearly after screen is a very relevant and difficult question to address, the Lehman failure, Brazilian banks found themselves in a but it is outside the scope of this chapter. Yet, this was one situation that resembled the aftermath of the Banco Santos of the key features of the crisis transmission into Brazil. intervention. It appeared more serious than the direct The rush to stop or to hedge these exposures in futures impact of the crisis on banks through reductions in trade markets for the real at the derivatives exchange in Brazil lines and in all other forms of external funding. Indications (BM&F Bovespa) put enormous pressure against the real are that medium-sized banks as a group lost nearly half and forced the central bank to act very swiftly on the sale of their deposits in the two months following the Lehman foreign exchange swaps to feed those willing to endure the Brothers event. The central bank deployed offsetting meas- losses and close these exposures. In parallel, the Brazilian ures on a massive scale, and again no casualties resulted Development Bank (BNDES) provided direct support to from what was a genuine bank run. One of these measures some of the companies in distress. The central bank was the creation of specially insured deposits up to R$20 reported some US$33 billion in sales of such swaps, which, million for every individual investor, although with a limit given their nondeliverable characteristic, did not affect for the receiving bank of up to either twice the value of its international reserves. A comparable amount in hard capital or the level of deposits in June of 2008, whichever reserves was deployed in a number of ways (outright sales, was larger, but limited to R$5 billion per institution. This funding of trade-related credit lines, purchase of Brazilian very exceptional guarantee has not been removed so far, paper, and the like) less to lean against the wind than to raising real concerns about moral hazard. reinstate markets for trade financing. BNDES resources One important factor that helps explain the system's deployed in bailouts cannot be precisely ascertained; yet resilience is that Basel ratios have been significantly higher figures for overall credit provided by BNDES in January in Brazil than in most other countries: Basel ratios are 13 2009 increased to some 9 percent of GDP from less than 6 percent of risk-weighted capital, and the average for the percent in August 2008! system is 18 percent. These higher ratios are due not just It is quite remarkable that all such actions and direct to regulatory requirements, but also to the absence of lim- bailouts were directed to help companies remedy the losses ited liability protection for controlling shareholders and produced by an undetected and yet incredibly dangerous directors in the event of intervention and liquidation. This dissemination of toxic products with a huge potential for latter factor results in overly conservative banks, with destruction. Nothing like this had been seen in other crises. chronic excess capital, little maturity mismatching (trans- Yet, by far the strongest and most worrisome aspect of the formation), and asset concentration of very short-term crisis was felt in some Brazilian banks that were in no way operations. This feature of Brazilian banking legislation involved with toxic derivatives, except perhaps indirectly may surely be a new twist in the regulatory issues under through the heightened public concern about banks and discussion in the United States. (This point is extensively the presumption that they could be affected by toxic discussed in Franco and Rosman 2010.) In addition, 11 derivatives. Brazilian banks at large had little or no direct middle-sized banks among the ones most pressured listed connection to events going on in the United States and their shares in the two years before the crises, each one Europe. Nevertheless, the transmission of a "risk aversion raising its capital by R$1.5 billion on average. shock" to Brazil was very concrete, and in many respects its The tradition of heavy regulation--combined with the impact seemed like what had happened a few years previ- weight of state-owned banks, notably Banco do Brasil, ous when Banco Santos failed. This episode involved a BNDES and Caixa Economica Federal, and the novel func- small-to-midsize bank specializing in wholesale funding tion of the FGC (Fundo Garantidor de Créditos, the local from institutional investors on the liability side and in the equivalent of the U.S. Federal Deposit Insurance Corpora- corporate middle market on the asset side. When the tion)--increased the "arms-length" intervention capacity central bank intervened in Banco Santos in 2004, many of the authorities very considerably. In this particular concerns spread through the wholesale time deposit episode, both federal banks were extremely active in the market, producing a major increase in standards and risk secondary market of credit portfolios. They decisively aversion on the part of the major players in this market, helped banks in distress sell their less liquid assets in order mostly pension funds. Other middle-market banks with to honor redemptions. funding needs similar to Banco Santos suffered significant Another important prudential resource was the fact that losses in their funding, producing considerable stress, reserve requirements in Brazil are very high compared to 92 Managing Openness other countries. This requirement is not justifiable from a The return to normality in the financial system was monetary policy viewpoint, since these reserves serve to quick, despite the huge depreciation of the real and its provide funding for government credit programs, mostly impact on the corporate sector. It is true that there was related to agriculture and housing. These excess reserve little foreign-denominated public debt and thus no bal- requirements should rightly be considered a distortion. ance sheet losses to the public sector. Yet, the most impor- However, it is no less true that they may have an important tant of all differences with respect to previous external "precautionary" function as the requirements can be shock-induced crises was related to the inflationary conse- adjusted to effectuate sharp changes in bank liquidity. Sig- quences of external adjustment. Previously, big devalua- nificant reductions in such requirements undertaken tions were needed to restore external balance, and working during the critical days of the crisis represented some addi- out the adjustments for the inflationary consequences tional R$100 billion in cash to banks (total time deposits was always painful. This time, however, the terms of trade were R$1.25 trillion in June 2008). Figure 7.4 shows how loss was so severe that it more than offset the impact of state-owned banks maintained the level of new credit, the real's depreciation on prices. Wholesale prices fell, avoiding a worse deterioration in consumption, when pri- bringing down consumer price indexes just when the cen- vate banks were decreasing their new loans. tral bank was at the beginning of a tightening cycle. In The monetary authorities, state banks included, played this respect, the crisis provided a free ride for monetary their part and avoided any accidents in the banking system, policy, because the central bank could reverse course and which would have been disastrous to the economy. Yet, the pursue aggressive reductions in interest rates where just acute bank distress had very considerable repercussions. the opposite was expected before the Lehman Brothers The credit contraction produced by the banks in distress event. This uncommon circumstance allowed monetary was particularly strong in the sales cycle of durable goods, policy to help banks and the real sector without departing most notably vehicles (30 percent year on year). The paral- from inflation targeting and without the central bank ysis in credit was sharp: credit came to a standstill, surely openly admitting any other influence in fixing the interest causing the rarely seen contraction of industrial produc- rate other than the normal course of business in pursuit tion observed in the fourth quarter. As seen in figure 7.5, of the inflation target. industrial production indexes displayed volatility typical of With the rapid normalization of the credit channels financial data, and the most impressive decreases were affecting automakers and durable manufacturers, the effects associated with durable goods, thus less affected by con- of historically low interest rates started to be strongly felt traction of external demand than by paralysis of domestic in the ensuing recovery. In this particular issue, the crisis credit stemming from financial distress. may have advanced the clock, as the economy reacted Figure 7.4. Ratio of Outstanding Credit to Total Assets in Brazil's Medium-Size and State-Owned Banks, 2008­09 0.30 0.12 0.11 medium-size banks state-owned banks 0.25 0.1 0.20 0.09 0.15 0.08 08 ly 8 08 08 ob 008 8 8 8 ry 9 ch 9 09 ay 9 ne 9 09 0 00 00 00 00 0 0 0 20 20 20 20 20 20 20 20 20 r2 ov er 2 ec er 2 r2 Fe ry 2 ay ne st ril be be Ju gu Ap b a ua M M ar Ju Ju nu em em em M Au br ct Ja O pt Se D N medium-size banks state-owned banks Source: Brazilian Central Bank databases, http://www.bcb.gov.br. Turning toward China? The 2008 Crisis and Its Influence on Brazil's Development Model 93 Figure 7.5. Industrial Production in Brazil, 2000­10 200 150 index number 100 50 00 01 02 03 04 05 06 07 08 09 10 20 20 20 20 20 20 20 20 20 20 20 total business equipment materials durable goods nondurable goods Sources: IBGE (Brazilian Institute of Geography and Statistics) databases, http://www.ibge.gov.br. beautifully to interest rates closer to civilized levels. Yet, as a capital inflows of certain types through taxation at the byproduct of the crisis, the government increased fiscal moment of entry. These restrictions were instantly spending and openly alluded to "countercyclical policies." removed as the crisis set in. Soon Brazil was again facing crowding-out symptoms and overheating just as before the Lehman event, and possibly The "Exit" and New Directions in a stronger way. in Development For completion of the picture, a few words are necessary on what happened to foreign trade. The fall in export vol- The crisis and recovery in Brazil, to judge from the descrip- umes helped deepen the contraction in activity, but as tion in the last section, had a somewhat more complex shown by the growth decomposition methodology (figure relation with the global financial crisis than one would 7.6), the fall in import volumes was bigger, and, conse- infer from the usual transmission channels. Moreover, and quently, net exports were positive for three quarters after in contrast with other countries, it is not exactly right to the beginning of the crisis. Indeed, the domestic market argue that Brazil deployed very exceptional policy meas- was much more relevant both to the contraction in the ures unlikely to remain or to be sustainable in normal cir- fourth quarter of 2008 and most especially to the recovery cumstances. Whereas interest rates in developed countries in the second quarter of 2009. In the first quarter of 2009, a were set close to zero in quantitative easing schemes to be huge fiscal stimulus reduced the decrease in GDP. After reversed at some point in the future, in Brazil, they that, the restoration of new loans by state-owned banks remained at a solid 8.75 percent throughout the crisis. This (figure 7.4) encouraged personal consumption and private nominal interest rate is, unfortunately, the lowest at least investment. The massive depreciation of the currency had a since the 1950s. Reserve requirements, even after being strong impact on the trade balance and the current substantially reduced, still represented a remarkably high account, and for this particular reason the authorities, percentage of all deposits (31 percent), hardly justifiable on especially in the Finance Ministry, seemed not entirely monetary policy grounds. Fiscal policy was already on an unhappy with developments in the foreign exchange mar- expansionary footing before the crisis, but the chance was ket. Complaints about continued currency overvaluation not missed to ratchet up a few more notches, given the had been frequent in the years before the crisis and had alleged need for "countercyclical" stimuli. The impact of prompted policy makers to implement restrictions on this display of naive Keynesianism is very much a matter 94 Managing Openness Figure 7.6. GDP Growth in Brazil, 2008­10 5.0 4.0 1.71 3.0 0.71 0.43 1.33 2.0 0.75 0.33 1.21 0.10 1.33 0.52 0.15 GDP growth (%) 1.0 0.41 0.17 1.85 1.60 1.07 0.92 1.27 0.93 0.54 0 0.02 ­0.05 ­0.62 ­0.38 -1.04 ­1.09 ­0.48 ­1.09 ­1.38 ­0.06 ­1.0 ­1.95 ­0.72 ­2.0 ­0.91 ­1.74 ­3.0 ­0.35 ­4.0 ­5.0 3Q-2008 4Q-2008 1Q-2009 2Q-2009 3Q-2009 4Q-2009 1Q-2010 change in inventories net exports fixed investment government expenditures personal consumption Source: IBGE National Accounts (database), http://www.ibge.gov.br. for discussion; it was not very prudent, especially in light of dealing with China as a trading and investment partner has Brazil's delicate track record of broken promises in the fis- been a rewarding experience, given the complementarities cal domain, and one can legitimately question whether it between the two economies, although with the usual com- was really necessary. mercial tensions. Emulating China is an entirely different In the first quarter of 2010, thanks to sharply increased proposition. It is an idea that seduces the emerging world private consumption, the economy was overheating again; partly in view of the legacy of the Bush administration and inflation was displaying worrying signs, capacity utiliza- of the banking crisis, which could do little else but elimi- tion and the labor market were stretched, and the central nate what was left of the Washington Consensus. But it also bank started a new tightening cycle. The return of a has something to do with the Chinese policy mix. Perhaps crowding-out configuration on the back of a heightened the first thing to catch the attention of a Brazilian observer fiscal and credit impulse loaded with rhetorical overtones is the importance of the accumulation of reserves. There somewhat deviant from orthodox thinking could be was a time when one would diminish the importance of explained, as a first approximation, by the political cycle. reserves, evoking the New Zealand example of a country in Election-motivated spending sprees can be seen all over which a pure float made it unnecessary to hold expensive the planet, and Brazil is no exception. Yet, there seemed to international reserves (and the more expensive, the higher be more to it, as the nature and extent of government the interest differential). Emerging markets hardly ever activism during the crisis and recovery may well be indica- adopted this wisdom; "fear of floating" is by far the rule tive of new directions in economic policy in the coming with very few exceptions, as is well known. The crisis has years. The challenge of increasing investment rates and only reinforced in a dramatic way the mercantilistic view thus attaining sustainable high growth looms large and that current account surpluses and reserve accumulation open, and in emerging markets in general, and in Brazil in are the shortest way to prevent external shocks from hurt- particular, one can say that policy paradigms have been ing domestic growth, as indeed suggested by the empirical significantly affected. In the decade of reforms, emerging exercise in table 7.2. Yet, the lure of the Chinese examples markets looked to Washington for recipes, most notably goes well beyond that. from the multilaterals from which emerged the much vili- In Brazil, the debate over the next steps in economic fied "consensus." Now, all curiosity is directed toward China. development can be said to refer particularly to China, or China's continued flamboyant success in economic more specifically to some key features of the Chinese policy growth invites speculation on how replicable the "Chinese mix that seem somewhat similar to the realities of the model" might be in other emerging economies. For Brazil, Brazilian economy. It was noted above, in connection to Turning toward China? The 2008 Crisis and Its Influence on Brazil's Development Model 95 figure 7.3, that there are similarities between China today appreciation trend. In the early years of stabilization, given and Brazil during the "economic miracle" years of the its exchange-rate­based character, the problem appeared military regime. It is interesting to push this analogy a bit to be temporary. But the currency experience years later, further. On the demand side, one can argue that not only under Lula, was similar and all the more revealing: the was government investment the key driver of growth, but policy mix in Brazil biases the economy toward a stronger also that "financial repression" was essential for mobiliz- currency, which is by no means "Chinese" and, for this ing "forced savings" out of the private sector, which, in particular reason, detrimental to growth. Again, it should itself, biases demand toward investment to the detriment be argued that the mix might be a low-growth one, having of consumption and thus biases relative prices toward appreciation as a consequence, which is entirely different tradables or toward a "competitive" exchange rate. Up to from arguing that the fixing of the exchange rate at over- this point, Brazil in the 1970s and China today look very valued levels was in itself a drag on growth. much alike. If one looks at Brazil having in mind the demand side of On the supply side, however, the distinguishing feature the Chinese model, many common features can be found. of China is the demography, or the excess supply of labor, Mechanisms of "financial repression" producing "forced combined with the deficit of democracy. This latter feature savings," many of them created in the 1970s, are still in is also important for the demand side of the model to the place, though with much less weight than they once had. extent that it reduced pressures toward public spending on Several important mechanisms are worth mentioning, social security and overheads, which further biases relative given the present administration's declared or revealed prices toward tradables. This was less clearly the case of intentions of using them more intensively: public commer- Brazil in the 1970s. cial and development banks (sometimes public commercial The effect of these demand and supply elements is the banks behaving as development banks) growing in impor- unusual combination of low interest rates and high savings tance, thanks to acquisitions, heavy capitalizations, and (and public investment) and a very "competitive," or expansionary credit policies; public enterprise­sponsored undervalued, exchange rate, which is, as put by Eichen- pension funds engaged in government programs; tax-fed green in this volume, "an outcome, or a relative price that budget funds with "organic" connections with develop- results from the elements comprising the development ment banks (like FAT/BNDES and FGTS/CEF); and reserve strategy, not a policy variable in and of itself." requirements and credit "directions" imposed on private In fact, the Chinese "model" outlined above is not at all banks. All these mechanisms are contentious, because they distant from Arthur Lewis's (1954) classic analysis of involve conflicts of interest between private shareholders growth under "unlimited supply of labor," an analogy that or participants and public endeavors and also because they was common in Brazil in the 1970s. In this model, wages are are ultimately "para-fiscal" activity somehow affecting the permanently at a "subsistence level," and growth results in a public deficit. To that extent, these mechanisms only number of paradoxes: capital accumulation does not aggravate a crowding-out situation that is typical of Brazil increase wages but profits, which means increasing inequal- and not at all Chinese. The root of Brazil's difficult fiscal ity (share of profits in GDP) and a rising investment rate situation, as mentioned in the first section of this chapter, (gross fixed capital formation), because capital accumula- is spending related to social security, health, and education tion does not affect wages so long as surplus labor prevails. to an extent unthinkable in China, yet normal for any Since savings result exclusively from profits, the rising share industrialized Western democracy. Hyperinflation was of profits in GDP implies a rising savings rate, or increased indeed a demonstration that Brazil could not have public "forced savings," consistent with a current account surplus. investment at levels comparable to the ones seen in The Lewis analysis as applied to China is intriguing and China and have government consumption at the levels of not at all inconsistent with Easterly's (2005) finding of a southern Europe. negative correlation between real exchange-rate overvalu- On the supply side of the Chinese model, Brazil no ation and per capita growth rates. It may be incompatible longer possesses the demography and the (lack of) democ- with Rodrik's (2008) contention that undervaluation pro- racy that maintain China under the regime of "unlimited motes economic growth to the extent that it provides a supply of labor," with all its implications for wages and second-best policy that offsets shortcomings in competi- the exchange rate. Absent these conditions, the movement tiveness and infrastructure deficiencies, precluding export of public investment toward a "Chinese model" is bound performance. The issue is at the very core of the current to produce a marked deterioration in the fiscal accounts Brazilian debates, because the currency has been, since the and the necessity of crowding out private spending, monetary reform in 1994 and absent crises, under a constant most likely investment demand, to accommodate 96 Managing Openness government programs and credit on an increased scale. nothing simple about it; in fact, it is as complex as saying Following the recovery from the 2008 crisis, the "Chinese" that inflation can be stopped if only the central bank would trends in Brazil are visible in the authorities' articulations stop printing money. Fiscal restraint is counterintuitive in of economic policy. The proximity of elections seems to an economy with substandard growth, especially in politi- produce spending sprees and rhetoric hostile to U.S.-style cians' minds. In fact, politicians tend to behave in just the neoliberalism, a combination that seems to point to the opposite way, expanding public spending along Keynesian East as paradigm. Yet, the social and economic realities of lines and ultimately making things worse. Brazil do not seem to allow much replication of the Chi- nese policy mix. In fact, the fiscal limitations would seem to point quite otherwise. Bibliography The question of what the competing model should be is Bonelli. R. 2002. "Labor Productivity in Brazil during the 1990s." Texto not that difficult. Its basic assumption is simple enough: para Discussão 906, IPEA, Rio de Janeiro. Easterly, W. 2005. "National Policies and Economic Growth." In Handbook further fiscal restraint would succeed in lowering the cost of Economic Growth, ed. P. Aghion and S. Durlauf. Amsterdam: Elsevier. of capital, whether because it would soften the "loose fiscal Ferreira, P.C. 2001. "Grupos de interesse, determinantes da política com- and tight money" deadlock or because it would inevitably ercial e produtividade industrial." Artigo preparado para o Instituto Futuro Brasin, São Paolo. improve sovereign risk ratings a couple of notches above Fischer, S., R. Sahay, and C. Végh. 2002. "Modern Hyper- and High Infla- the "investment grade" status. These are, in essence, the tions." Journal of Economic Literature 40 (3): 837­80. fundamental explanations for the one economic pathology Franco, G. H. B. 1999. O Desafio Brasileiro: ensaios sobre desenvolvimento, that singles out Brazil as a low-investment country: the globalização e moeda. São Paulo: Editora 34. ------. 2000. "The Real Plan and the Exchange Rate."Essays in Interna- interest rate. The brief period of time in which Brazil had tional Finance. Department of Economics, International Finance Sec- nominal interest rates below 10 percent, for the first time in tion. Princeton University, Princeton, NJ. the past 40 years, offered a glimpse of a new reality regard- Franco, G. H. B. and L. A. C. Rosman. 2010. "A crise bancária norte-ameri- cana: algumas lições da experiência brasileira." In Risco e Regulação, ing borrowing, leveraging, capital markets, and equity val- orgs. M. Garcia and F. Giambiagi. Rio de Janeiro: Campus & Elsevier. uations that captured people's imagination. In many Gomes, V., S. Pessoa and F. Veloso. 2003. "Evolução da produtividade total respects, the sensation was as far reaching as the one caused de fatores na economia brasileira: uma análise comparativa." Pesquisa e Planejamento Econômico 33 (3): 389­434. by the end of high inflation, and there is no coincidence in Lewis, W.A. 1954. "Economic Growth under Unlimited Supplies of this analogy, given that these pathologies shared the same Labor." The Manchester School 22 (2): 139­91. origin. Historically, Brazilian companies are extremely Moreira, M. M. 2004. "Brazil's Trade Liberalization and Growth: Has It averse to any form of indebtedness--exceptions are prime Failed?" Occasional Paper 24, INTAL, Buenos Aires. Pinheiro, A. C., I. Gill, L. Severn, and M. Thomas. 2001. "Brazilian Eco- corporations with access to international capital markets-- nomic Growth in 1900­2000: Lessons and Policy Implications." Paper which emphasizes as one forceful implication the fact that presented at the Third Annual Conference on Global Development capital expenditures are made mostly out of retained earn- Network, Rio de Janeiro, December 9­12. Rodrik, D. 2008. "The Real Exchange Rate and Economic Growth." Brook- ings. This is actually a microexpression of a low rate of ings Papers on Economic Activity 2: 365­412. gross fixed capital formation in the aggregate and a con- Rossi, J. L. Jr., and P. C. Ferreira. 1999. "Evolução da produtividade indus- venient way to look at the reason why interest rates are trial brasileira e a abertura comercial." Pesquisa e Planejamento indeed the one important obstacle preventing private Econômico 29 (1): 365­412. Souza, J. R. C. 2007. "Estimativa do produto potencial para a economia investment and saving from reaching levels consistent with brasileira: atualização utilizando o sistema de contas nacionais refer- high economic growth on a sustainable basis. There is ência 2000." Nota Técnica, IPEA, Brasilia. Part II II Open but not dependent: South-South Trade and Export Diversification 8 Changing Dynamics in Global Trade Gordon Hanson The economic crisis of 2007­09 has led to a severe decline demand side, how important is the United States for the in world trade. Between the fourth quarter of 2007 and the growth of developing-country exports? On the product side, second quarter of 2009, world merchandise imports has trade expanded more for countries that export manufac- dropped by a staggering 36 percent. Trade levels began a tures than for those that produce primary products? modest recovery in the third quarter of 2009, but they are One of the main findings of this chapter is that develop- still far below precrisis highs. ing countries are the new drivers of global trade. Middle- Will the Great Recession encourage developing coun- income countries (including, but not limited to the BRICs, tries to adopt more inward-oriented development strate- or Brazil, the Russian Federation, India, and China) have gies? Although U.S. import demand helped sustain export become an important source of demand for exports from growth in developing countries over the past two decades, low- and middle-income countries. Gravity model esti- it may not fully recover from the crisis of 2007­09. Coun- mates suggest that growth of gross domestic product tries following an export-led development strategy may (GDP) in developing countries has led to increased find themselves producing goods for which there are insuf- demand for imports, which has played an important role in ficient foreign buyers to sustain steady increases in per boosting exports of low-income countries. Concerns that capita income (Rodrik 2009). Conceivably, the response of weak consumption growth in the United States might some countries could be a shift away from trade and invest- undermine the viability of export-led development would ment openness. Developing countries may start to reverse appear to be exaggerated. Dependence on U.S. demand for reforms that have reduced barriers to foreign trade and imports has generally lessened, although the United States investment and eased regulations on industries and labor is still a key source of demand for exports from the Western markets in recent decades. Hemisphere, for oil products, and for apparel from low- It is easy to overstate the potential for de-globalization. income countries. No prominent international economists are advocating a This chapter also focuses on the new supply patterns wholesale rejection of openness as a foundation for eco- that have emerged since the mid-1990s. Developing coun- nomic development. The countries that have adopted tries have captured a larger share of world exports in the more inward-oriented development strategies recently, past decade. Low-income countries remain specialized in such as Argentina, Bolivia, and República Bolivariana de resource-intensive exports, but they are making small steps Venezuela, began doing so before the crisis hit. However, up the value chain through expansion in apparel produc- many developing-country policy makers have a sense that tion. Middle-income countries' export baskets are more emerging markets are too reliant on the world economy diversified than those of low-income countries, but low- and on the United States, in particular.1 income countries are steadily closing the gap. In considering the future of globalization in the devel- The next section discusses the relative role of developing oping world, we must first assess the nature of global trade countries and the United States in driving trade demand. today. How dependent are developing countries on partic- The following section examines changing dynamics on the ular markets or particular goods for their exports? On the supply side. The final section concludes. 99 100 Managing Openness The New Drivers of Global Trade nations (table 8.1). Significantly, the shift in demand away from high-income markets is not driven only by the BRICs. In the past decade, the destination markets for developing- Other low- and middle-income countries are also driving country exports have shifted. Although these exports used to new demand: of the 13 percent increase in market share for be directed largely at high-income countries, they are now low- and middle-income nations, the BRICs account for moving increasingly toward other developing economies 5 percent and other countries account for 8 percent. Demand instead. As a result, the share of low- and middle-income from the BRICs is generally growing at a faster rate, but other countries in total world imports has nearly tripled, from low- and middle-income countries still import a bigger share 12 percent in 1996 to 31 percent in 2008 (figure 8.1). of exports from all income groups, apart from upper-middle- Within this group, the BRICs' share of world imports more income countries. than tripled, from 4 percent to 12 percent. Consequently, In general, the rising share of exports to developing coun- high-income countries now account for less than 70 percent tries has meant a natural fall in the share of exports to high- of world imports, compared with nearly 90 percent a little income nations. The share of imports to the European over a decade ago. Union (EU) has shrunk since 2000: its import share from the This shift in demand has had varying impacts on specific BRICs and other high-income countries has increased, but exporter groups. Imports of low- and middle-income coun- the share has fallen significantly for imports from low- and tries from all other income groups (that is, low-, middle-, and lower-middle-income countries. The share of exports of high-income nations) increased since 2000. In particular, developing countries to the United States fell for all income low- and middle-income countries absorbed half of the groups except for low-income countries. However, low- export growth of low-income and lower-middle-income income country exports to the United States still account for only 18 percent of total exports from these countries. Figure 8.1. Share of World Imports by Importer The varied changes in bilateral absorption patterns sug- Income Group, 1996­2008 gest that the story is not just about differential rates of income growth across high-income countries. Were this 100 the case, then the logic of the gravity model of trade (Feen- 90 stra 2004) would suggest that countries increase or decrease their consumption of goods across exporters in 80 constant proportions. Other factors--for instance, bilateral 70 trade or investment agreements, the strategies of multina- tional firms, and the effect of income levels and income 60 distribution on patterns of import demand--clearly percent 50 account for why the share of goods going to the United States or the EU rises for some groups but falls for others. 40 30 Effects of a Drop in U.S. Consumption 20 Despite these changing trends, the United States continues 10 to remain a major export destination for Latin America 0 and the Caribbean, for oil exporters worldwide, and for low-income countries in East and South Asia. With respect 96 98 00 02 04 06 08 19 19 20 20 20 20 20 year to Latin America and the Caribbean, Mexico and Central America are particularly dependent on U.S. demand. The low- and middle-income countries United States also absorbs large amounts of oil from the BRICs Middle East and other petroleum exporters, but the oil other low- and middle-income countries import demand of middle-income countries such as the high-income countries BRICs is steadily increasing. For low-income countries in European Union 25 Asia, the importance of the U.S. market is predominantly United States in apparel exports. The United States and the EU absorbed other high-income countries over three quarters of the growth in recent (2000­08) Source: Author's calculations, based on data from UN Comtrade (database), apparel exports from low-income nations, with East and United Nations, http://comtrade.un.org. South Asian nations being the main producers. Changing Dynamics in Global Trade 101 Table 8.1. Share of Exports by Country Income Groups and Destination Market, 2000 and 2008 percent Income group of exporting country Low (%) Lower middle (%) Upper middle (%) BRICs (%) High (%) Importing region 2000 2008 2000 2008 2000 2008 2000 2008 2000 2008 Low- and middle- income countries 28 41 · 23 41 · 18 31 · 17 30 · 21 30 · BRICs 11 16 · 9 23 · 6 13 · 3 8 · 5 12 · Others 17 25 · 14 18 · 12 19 · 14 22 · 15 18 · High-income countries 72 59 , 77 59 , 82 69 , 83 70 , 79 70 , EU25 36 25 , 26 21 , 26 31 · 25 28 · 43 43 · United States 15 18 · 25 19 , 40 27 , 24 20 , 17 12 , Others 21 16 , 27 19 , 15 11 , 35 22 , 20 15 , Source: Author's calculations, based on data from UN Comtrade (database), United Nations, http://comtrade.un.org. But these findings suggest that concerns that a con- Nigeria's oil exports), the upper-middle-income countries sumption bust in the United States might threaten global in the Middle East and North Africa (primarily because of economic cooperation are exaggerated. The most adverse oil exports from the Persian Gulf), and China. impacts of a slowdown in growth of U.S. import demand would likely be felt by Latin America, the Caribbean area, Continued U.S. Demand for Oil and Apparel Exports and China. With the potential for growth in domestic consumption in China, however, the displacement of From 2000 to 2008, the United States absorbed 20 percent of Chinese goods from the U.S. market may not be a grave export growth from low-income countries, but 14 percent concern. came from just two sectors: petroleum and apparel. Simi- Distinct changes in import demand have also occurred larly, the United States absorbed 15 percent of export growth at the regional level. In Europe and Central Asia, export in lower-middle-income countries, with 13 percent of that destinations have diversified, moving away from the EU growth in petroleum. U.S. absorption of export growth is and the United States and toward low- and middle-income spread across more than two sectors only for upper-middle- countries, including the BRICs (at least for Bulgaria and income countries and the BRICs, reflecting the diversifica- Poland). In countries from the Middle East and North tion in the exports from these countries. Africa, South Asia, and Sub-Saharan Africa, there is a shift Absorption by the EU, though also concentrated, is not as in exports toward the BRICs and low- and middle-income extreme as the United States. The EU absorbs 20 percent of countries in general. East Asian exports are shifting exports from low-income countries, with about half coming strongly toward the BRICs and other low- and middle- from apparel and one-quarter from food products. The EU income economies in all countries except Cambodia. Latin and the United States together absorbed 79 percent of low- American exports too are shifting toward other low- and income countries' export growth in apparel and footwear middle-income countries but more modestly than in the over the 2000­08 period, making these destination markets other regions. The United States remains the dominant overwhelmingly important for exporters of low-end manu- destination for nearby Guatemala and Mexico. factures. Low-income countries in Latin America and South Although the United States captures a significant share Asia are the most dependent on import demand from of oil and apparel exports, producers of these goods are not high-income nations. The United States absorbed more necessarily at risk from U.S. belt-tightening. Given the rela- than 15 percent of recent (2000­08) export growth in just tively inelastic demand for oil and semidurable apparel, a 7 of the 22 regional income-group exporters: 3 are in Latin major reduction in U.S. consumption growth in these America and the Caribbean, and the others are low- products seems unlikely. More likely candidates for import income countries in East Asia and the Pacific (because of reductions would be durable goods in the machinery, elec- apparel exporters in the region), lower-middle-income tronics, and transportation equipment sectors, which are countries in Sub-Saharan Africa (primarily because of major exports for upper-middle-income countries. 102 Managing Openness The United States clearly remains an important mar- makes a larger contribution to export growth in lower- ket for developing countries, but developing countries, as middle-income countries in Sub-Saharan Africa (in part a group, have greatly increased in importance. Low- and because of strong trade ties between the United States and middle-income countries, including the BRICs, absorb Nigeria), low-income South Asia, and upper-middle- 63 percent of low-income-country exports of petroleum, income East Asia and the Pacific, with contributions above 70 percent of their exports of iron and steel, and 52 percent 5 percent in each case. Not surprisingly, the most important of their exports of food products. The only sectors with U.S. contribution is in Latin America and the Caribbean, minimal absorption are electronics and other products where it accounts for 13 percent of export growth. (which include capital-intensive instruments and arms). Relative to the United States, GDP growth in the EU Low- and middle-income countries thus appear to have makes a larger contribution to export growth in Europe and done more to help low-income exporters diversify their Central Asia, the Middle East and North Africa, and South exports than has the United States or the EU. Asia. It makes a smaller contribution to export growth in East Asia and the Pacific and in Latin America and the Caribbean. These results are as expected, given that Europe The Role of GDP Growth in Developing Countries' has relatively stronger trade ties with nearby North Africa, Demand for Imports Central and Eastern Europe, and Central and South Asia. The expanding role of low- and middle-income countries The gravity model captures only partial equilibrium as a source of import demand can be partially explained by relationships between growth in importer GDP and growth the rapid growth in their GDP. As these economies grow rel- in developing-country exports. GDP growth may have indi- ative to rich countries, their expanding import demand rect effects on trade that the model does not pick up. That absorbs an ever-larger share of global exports. Over the is, the model analyzes the direct impact of GDP growth in period 2000 to 2008, average annual growth in real GDP an importing country on growth in bilateral imports from was 9.7 percent in China, 7.3 percent in India, 6.5 percent in one of its trading partners, disregarding the potential indi- Russia, and 3.5 percent in Brazil. In other low- and middle- rect effects of off-shoring or global production networks. income countries, it was 4.3 percent. By contrast, GDP Income growth in the United States, for instance, may growth averaged approximately 2 percent in United States increase demand for apparel imports from Vietnam, which and the EU. The much stronger growth performance of may in turn increase Vietnam's demand for textile imports low- and middle-income nations in the past decade is from Malaysia.3 associated with export growth in their developing-country The conceptually correct way to measure exports would trading partners. A gravity model provides quantitative be as in-country value added, excluding the value of confirmation of this relationship by decomposing sources imported intermediates. However, no nation follows this of growth in developing-country exports into portions convention in measuring trade flows. As a result, observed associated with their own GDP growth and the GDP trade overstates the shipment of value added between growth of their trade partners.2 nations. Although it is a valid concern, the amount of dou- Higher GDP growth rates in low- and middle-income ble counting would have to be immense to explain away the economies explain a significant portion of the growth of observed increase in absorption of exports by low- and exports from low-income countries: 51 percent of export middle-income countries. growth in low-income Middle East and North Africa, In summary, middle-income countries are emerging as 42 percent of export growth in low-income Europe and powerful drivers of growth in global import demand. Central Asia, and 21 percent of export growth in low-income However, a drop in U.S. import demand could still have an Sub-Saharan Africa. All of these figures exceed the contribu- adverse impact on the Western Hemisphere, oil producers, tion of the United States and EU combined. In middle- and apparel-exporting nations. Meanwhile, GDP growth in income countries, the contribution of GDP growth in low developing countries will continue to play a stronger role and middle-income countries to export growth is smaller. in growing global import demand in the future. Despite similar growth performances, the gravity model suggests that the United States and the EU make very dif- The Effects of Changes in Demand ferent contributions to export growth in particular devel- on Supply Patterns oping regions, owing to differences in the strength of their trade ties. In Europe, Central Asia, and the Middle East From 1996 to 2008, low-income countries' share of world and North Africa, the contribution of U.S. GDP growth to exports rose from 0.7 percent to 1.1 percent (figure 8.2). regional export growth is less than 3 percent. U.S. GDP Low-income countries are expanding trade, but overall Changing Dynamics in Global Trade 103 Figure 8.2. Share of World Exports by Exporter countries, export growth appears to be almost entirely Income Group, 1996­2008 resource dependent, relying on raw labor (apparel), mineral resources (petroleum, iron ore), or agricultural land (food 90 products). Low-income countries export little in the way of 80 chemicals, electronics, machinery, or transportation equip- ment, all of which are intensive in human or physical capital. 70 One way in which low-income countries are moving up the value chain is by expanding apparel production. In 60 1996, 86 percent of exports from low-income countries was concentrated in two major product categories (food percent 50 products, apparel and footwear) and three minor products 40 (textiles, petroleum, iron and steel). In 2008, these five products continued to account for nearly 80 percent of 30 low-income country exports, but their relative importance 20 had changed. Apparel, petroleum, and iron and steel became even more important, while textiles and food 10 products became less important. The growing export share of apparel products relative to textiles marks a step up the 0 value chain for low-income countries. The changing prod- 96 98 00 02 04 06 08 uct mix of exports from low-income countries can also be 19 19 20 20 20 20 20 year seen as a response to the declining relative demand from low-income countries the EU (a main importer of food products from low- lower-middle-income countries (excluding BRICs) income countries) and increasing relative demand from upper-middle-income-countries (excluding BRICs) middle-income countries such as the BRICs (key importers BRICs of oil, iron, and steel). high-income countries As low-income countries have increased apparel exports, middle-income countries' share of apparel exports Source: Author's calculations, based on data from UN Comtrade (database), United Nations, http://comtrade.un.org. has declined. In 1996, for example, apparel accounted for 23 percent of BRIC exports, due almost entirely to the con- tribution of China. Yet, by 2008 the share of apparel exports had fallen to 11 percent, as China experienced they still remain small players in global commerce. Lower- more rapid export growth in other sectors. China, as middle- and upper-middle-income countries (excluding well, is moving away from apparel and footwear (at least the BRICs) saw their combined export share rise from in a relative sense), opening the way for lower-income 13 percent to 18 percent. The BRICs' global share of nations to fill global demand for these goods. One possi- exports shows an impressive rise from 9 percent to 17 per- ble explanation for this shift is that productivity growth cent. They now export nearly as much as all other low- and in middle-income countries is putting upward pressure middle-income countries combined.4 As the export share on wages, thereby pricing them out of the labor-intensive of low- and middle-income groups grew, the global export apparel sector. This phenomenon has been occurring in share for high-income countries correspondingly fell from high-income countries for several decades, although import 77 percent to 65 percent. For the first time since the early barriers in these countries have prevented them from losing 20th century, high-income countries accounted for less their apparel edge entirely. This factor could also account than two-thirds of world exports, with their share likely to for the entry of low-wage nations into markets that were continue shrinking in the years to come. earlier the bastions of middle- or high-income nations. Specialization in Low-Income Countries The Increasing Diversity of Exports from Middle-Income Countries From 1998 to 2006, just two products,5 apparel and petro- leum, accounted for 51 percent of the growth in exports of Reliance on a dominant export product is lower for richer low-income countries. The next two, food products and iron countries in all regions except South Asia (because middle- and steel, bring the total to 76 percent. For low-income income Pakistan is more specialized than low-income 104 Managing Openness Bangladesh) and Sub-Saharan Africa (because of the The BRIC countries have little in common in their export presence of oil-exporting Nigeria among middle-income patterns except for scale. Export growth in Brazil is concen- countries in the region). 6 The average export share of trated in food products and iron and steel. Export growth the dominant export exceeds 20 percent in all lower- in Russia is concentrated in oil. India's export growth is middle-income regions but exceeds only 13 percent in all concentrated in chemicals, nonmetallic minerals, and iron upper-middle-income regions. The export share of the dom- and steel.8 Export growth in China is concentrated in inant export for low-income countries equals or exceeds machinery and electronics. Thus, while the BRICs as a 25 percent in all regions except South Asia. Thus, not until group resemble other middle-income countries, individu- countries exceed a gross national income per capita of ally they are heterogeneous. around US$4,000 does their most important export account for less than a fifth of total shipments abroad. Changing Patterns of Export Diversification Of course, diversification at the income group level may hide specialization among individual countries. However, As seen in the HHI in figure 8.3, specialization levels have these results are roughly consistent with recent literature that notably declined for low-income countries, from an HHI finds that from the 1970s to the mid-1990s, countries value of 0.25 to 0.16 over the 10-year period.9 For these became more diversified in the allocation of employment countries, this change is equivalent to going from an equal across sectors as their incomes rose from low levels to higher export spread across just four products to an equal export levels. Imbs and Wacziarg (2003) document that the rela- spread across about seven products (out of the over 1,200 tionship between the sector concentration of employment products they could export). While either case represents a and per capita GDP is U-shaped. Countries become more high degree of specialization, the movement toward greater diversified during initial phases of development and less diversification among low-income countries is pronounced. diversified in later stages. An estimation of export diversifi- This increasing diversification among low-income coun- cation levels with the Herfindahl-Hirschman Index (HHI) tries is significant, given that all other country-income also confirms that low-income countries are the most spe- groups have become somewhat more specialized in their cialized, while high-income countries are the least. Interest- exports over the period. In 2007, the average HHI for lower- ingly, the difference between low-income countries and all middle-income countries was 0.12 and 0.07 for upper- other countries is much larger than the difference between middle-income countries. It should be emphasized that the middle- and high-income countries. Distinct from Imbs and changes in the HHI distributions for middle- and high- Wacziarg (2003), much of the diversification in exports income countries are slight and therefore unlikely to be sta- appears to occur in the early stages of economic develop- tistically significant. Still, the pronounced increase in export ment. Little apparent increase in specialization occurs as a diversification of low-income countries stands in stark con- country moves from the middle-income to the high-income trast to changes for middle- and high-income countries. group. The HHI results can be partially explained by the Although middle-income countries have diversified decline in food exports, which fell from nearly one-third of exports more than low-income countries, some specializa- low-income country exports to 17 percent in 2008. The tion still exists. Given the prevalence of middle-income movement out of food products and into other goods, countries in the Middle East and North Africa, petroleum especially petroleum, apparel, and iron and steel, appears is the most common product specialization among mid- to have helped make low-income countries more diversi- dle-income countries. In 2007, the dominant export of fied, although they are still more specialized than higher- middle-income countries in the Middle East and North income countries. Africa accounted for 60 percent of shipments abroad, ver- One might imagine that the distribution of HHIs would sus 92 percent for low-income countries in the region.7 depend on whether oil exporters are included in the data. In contrast, the BRICs are more diversified across prod- These countries tend to be highly specialized in crude petro- ucts, exhibiting a pattern of sector growth that looks simi- leum and petroleum refining. Yet, even when the 20 major lar to high-income countries. That said, the BRICs do have oil exporters are dropped from the sample, the plots look relatively high shares in exports of apparel and electronics highly similar.10 The phenomenon of increasing diversifica- and relatively low shares in chemicals and transportation tion among low-income countries and modestly decreasing equipment. The first two sectors are intensive in lower- diversification among middle- and high-income countries skilled labor (China specializes in electronics product thus holds for non-oil exporters. assembly, which is not skill intensive), whereas the second In summary, middle-income countries are much more two sectors are relatively intensive in physical capital. diversified in their export mix than low-income countries Changing Dynamics in Global Trade 105 Figure 8.3. Average Export HHIs of Non-oil- In response to changing dynamics on the demand side, Exporting Countries Weighted by Exports, 1997 and new supply patterns have emerged. Developing countries 2007 have captured a larger share of world exports in the past decade. Low-income countries remain specialized in 0.30 resource-intensive exports, but they have diversified to a average export HHI (scale of 0 to 1) 0.25 certain extent, both in the regions they supply and in the types of goods they produce. Low-income countries are 0.20 also making small steps up the value chain through expan- sion into apparel production. 0.15 The degree of export diversification in middle-income countries is quite similar to that in high-income countries, 0.10 with richer nations being only somewhat more diversified. Yet, despite the greater export diversification in middle- and 0.05 high-income countries, it is low-income countries that have experienced the greatest increase in diversification, closing a 0 significant part of the gap that exists between them and tri e- s tri e tri e tri e- IC un om un om un dl un dl es es es es BR richer nations. The recent phase of globalization therefore co id co id co inc co inc e -m e -m - - gh w m er m er appears to have helped low-income nations expand the lo co w co p hi in up in lo range of goods they produce for foreign markets. 1997 2007 Notes Source: Author's calculations, based on data from UN Comtrade (database), United Nations, http://comtrade.un.org. 1. Interestingly, dependence on the United States does not appear to have been a predictor for which countries suffered most from the eco- nomic crisis (Rose and Spiegel 2009). and are approaching diversification levels comparable to 2. For full details of data, methodology, and estimation results, see the working paper version of this chapter. some high-income countries. However, low-income coun- 3. It should be noted that the gravity model does not focus on where tries are now closing part of the export diversification gap. GDP growth in importers and exporters comes from. Nothing in the analy- Furthermore, middle- and high-income countries have sis is informative about whether rapid GDP growth in low- and middle- income countries is sustainable. We can only interpret from the model that actually become slightly less diversified in their exports in if growth continues, it is likely to generate further increases in demand for recent years. The proximate cause of rising absolute and rel- exports from other developing countries. ative export diversification in low-income countries is a 4. Among the BRICs, China and India are classified as lower-middle- movement away from heavy reliance on food products income countries and Brazil and Russia are upper-middle-income countries. 5. Products are grouped according to factor intensity. Using four-digit toward other sectors. 1996 Harmonized System (HS) product classification codes, we construct 12 aggregates: raw and processed food products (HS 1­24); petroleum products (HS 27); chemicals (HS 28­40); textiles (HS 41, 50­60); wood and paper Conclusion products (HS 44­49); apparel and footwear (HS 42­43, 61­67); nonmetallic minerals (HS 25, 68­71); iron and steel (HS 26, 72­83); machinery (HS 84); Trade levels have fallen since the crisis, but the growing electronics (HS 85); transportation equipment (HS 86­89); and other man- demand for exports in developing countries suggests that ufactures (HS 90­97). Other manufactures include optical instruments, clocks, musical instruments, arms, furniture, toys and games, miscellaneous falling consumption in the United States is not as harmful goods, and works of art. These product aggregates are similar to those in to global trade as some have claimed. Middle-income Leamer (1984), who examines industry factor intensity more formally. countries (including, but not limited to, the BRICs) have 6. The dominant export is defined as the product with the highest become an important source of demand for exports from export levels according to four-digit 1996 HS product classification codes. The dominant export varies across countries, regions, and low-income and lower-middle-income countries. A simple income levels. The most common dominant export good is crude petro- gravity model suggests that GDP growth in developing leum (HS 2709), which meets the criterion for 6 low-income countries, countries plays a role in their growing demand for imports. 18 middle-income countries, and 8 high-income countries, followed by refined petroleum (HS 2710), for 3 low-income, 9 middle-income, and And as demand from middle-income countries has risen, 13 high-income countries. Other (less) common dominant exports dependence on the United States as a destination market include frozen fish, petroleum gases, passenger and transport ships, and for developing-country products has lessened. However, rough lumber. Most of these goods are intensive in the use of natural resources of some kind. Export specialization frequently follows from the United States is still a key driver of demand for exports resource abundance, especially for low-income countries. from the Western Hemisphere, for oil products, and for 7. This figure is calculated as a centered moving average for 2006­08. apparel from low-income countries. We average the maximum shares across years rather than show shares for 106 Managing Openness individual years to account for the impact of temporary global price Bibliography shocks on export values. The data are weighted by country shares of global exports. Feenstra, R. C. 2004. Advanced International Trade. Princeton, NJ: Princeton 8. Service exports associated with the off-shoring of call centers and University Press. other backbone services are important in India's total exports but are not Imbs, J., and R. Wacziarg. 2003. "Stages of Diversification." American Eco- included in the data we use. nomic Review 93 (1): 63­86. 9. To examine export specialization patterns more systematically, we Leamer, E. E. 1984. Sources of Comparative Advantage. Cambridge, MA: calculate the export Herfindahl-Hirschman Index for each country, which MIT Press. measures the concentration of exports in individual products. This pro- Rodrik, D. 2009. "Growth after the Crisis." Discussion Paper 7480, Centre vides a scalar measure for each country's export specialization. for Economic Policy Research, London. 10. Major oil exporters are defined as countries for which petroleum Rose, A. K., and M. M. Spiegel. 2009. "Cross-Country Causes and Conse- or petroleum-related products were the dominant export in both 1997 quences of the 2008 Crisis: International Linkages and American and 2007 and for which the share of the dominant export in a country's Exposure." Working Paper 15358, National Bureau of Economic total export exceeded 25 percent. Research, Cambridge, MA. 9 Sources of Export Growth in Developing Countries Gordon Hanson The world economy is beginning to recover from the Great factor productivity, capital investment, and skill upgrad- Recession of 2007­09. Industrial production in emerging ing, or is it simply a temporary deviation from a more economies, after falling sharply in the second half of 2008, modest trend? Second, is export growth in developing began to grow in early 2009. World trade also began to nations being matched by improvements in their export rebound at this time. The financial crisis, though crippling capacity, allowing the expansion of exports to be sustained? in its effects on rich countries, did not infect developing To address these questions, we examine the contribution economies outside of Eastern Europe (de la Torre 2010). of changes in export competitiveness and importer growth Indeed, the exposure of emerging economies to the reces- in gross domestic product (GDP) to growth in exports by sion has been limited largely to a contraction in the global developing countries over the period 2000­2007. The rest of demand for their exports. With that contraction easing, the chapter is organized as follows. The next section dis- world trade is likely to continue to expand in 2010. cusses changes in global trade patterns and GDP growth As discussed in the previous chapter, trade patterns have over the past decade. The following section explores the sus- changed dramatically over the past decade. Most notably, tainability of these trends based on the findings of a gravity developing countries have become a prime source of model. The final section concludes. growth in demand for exports from other developing countries. And while developing countries have increased Recent Trends in Global Trade and GDP their share of global imports, high-income countries have become relatively less important destinations for develop- To understand the relationship between economic growth ing-country goods. So what has driven this change in trade and global trade patterns, we must examine basic trends in flows? The primary reason is that the locus of economic trade flows and GDP growth. growth has shifted from high-income to middle-income nations. Middle-income countries account for an ever- Changing Trade Patterns larger source of global demand because they are now grow- ing much more rapidly than high-income economies. A growing trend of export diversification has emerged in In this chapter, we explore whether recent changes in the past decade. In particular, exports from low-income global trade patterns are likely to be sustained as the global countries have become more diversified (table 9.1). While economy recovers. Naturally, a prolonged collapse in eco- low-income countries are still specialized in agriculture; nomic growth in the BRICs (Brazil, the Russian Federa- extractive industries; and apparel, footwear, and textiles, tion, India, and China) or other middle-income economies they have also begun shifting into sectors in which they had would profoundly affect global trade. But without such an relatively low export participation in the past, including interruption, are trends in global trade expected to con- iron and steel (metals) and machinery; electronics; and tinue? The answer depends on two factors. First, has recent transport equipment. growth in middle-income economies been robust; that is, Relative to low-income countries, lower-middle-income does it represent long-run growth based on growth in total countries are more diversified. In 2007, the most important 107 108 Managing Openness Table 9.1. Export Shares by Sector and Developing-Country Income Group, 2000 and 2007 percent Income group Low Lower middle Upper middle Sector 2000 2007 2000 2007 2000 2007 Agriculture 16.4 12.8 5.0 2.8 6.2 6.0 Food, beverage, tobacco, wood 6.7 6.3 5.4 4.1 7.5 7.1 Extractive 26.3 24.0 16.8 17.9 27.3 36.2 Chemicals 2.6 3.9 6.3 7.1 5.9 6.3 Apparel, textiles, footwear, leather 38.1 34.2 22.0 14.4 7.8 4.3 Metals 3.1 7.7 4.6 7.6 8.9 10.8 Machinery, electronics, transport 4.8 7.2 28.5 37.0 32.6 25.9 Other 2.0 3.9 11.4 9.1 3.7 3.4 Source: Author's calculations based on UN Comtrade (database), United Nations, http://comtrade.un.org. sector was machinery, electronics, and transportation East Asia and the Pacific and for Latin America and the equipment (37 percent of exports); and extractive indus- Caribbean, while extractive industries are the major source tries was the second most important (18 percent of exports of exports for the Middle East and North Africa, Europe in 2007) for lower-middle-income countries. One notable and Central Asia, and Sub-Saharan Africa. shift for lower-middle-income countries is the decreasing New trends have also emerged with respect to destina- importance of apparel and textiles, whose share of exports tion markets. Developing countries are an increasingly fell from 22 percent in 2000 to 14 percent in 2007. Of important source of export growth for low- and middle- course, this change can be attributed largely to the decreas- income economies (table 9.2). The share of exports des- ing relative importance of apparel in China's exports. tined for low- and middle-income countries has risen Upper-middle-income countries exhibit similar diversi- significantly, with the most pronounced growth occurring fication patterns as lower-middle-income countries, with in lower-middle-income exporters.1 The BRICs have the primary difference being that extractive industries attracted much attention owing to their large size and rapid (36 percent of exports in 2007) are more important than rates of recent economic growth, but they are by no means machinery and electronics (26 percent of exports in 2007). the only drivers of export growth in developing countries. This finding is not surprising, given the preponderance of oil exporters in the Middle East and North Africa in this Trends in GDP Growth income group. Over the past decade, export patterns have also varied To understand whether the above-mentioned patterns in within income groups across regions. While apparel is global trade are sustainable, we first take a closer look at the important for low-income countries overall, much of this decomposition of GDP. Figure 9.1 presents the trend and significance is due to East Asia and the Pacific and South cyclical components of GDP for several groups of coun- Asia, where in 2007 apparel accounted for 37 percent and tries. Each plot shows log GDP (left-hand scale), the trend 89 percent of exports, respectively. In the Middle East and component of log GDP (left-hand scale), and the cyclical North Africa and Sub-Saharan Africa, in contrast, the pri- deviation from trend (right-hand scale). mary export sectors were agriculture, extractive industries, Among the BRICs, GDP growth seems to be generally and metals. associated with trend GDP rather than cyclical variations Among lower-middle-income countries, the shift into (figure 9.1a). China and India have clearly experienced machinery and electronics was strongest in East Asia and trend growth over the past three decades. Moreover, after a the Pacific, where the sector accounted for 46 percent of recession in the early 1990s, China's GDP has shown little exports in 2007. Extractive industries are important for cyclical variation. The cyclical component of India's GDP is Sub-Saharan Africa, the Middle East and North Africa, similarly modest. In Brazil, trend growth has accelerated Europe and Central Asia, and Latin America and the and cyclical variation has dampened over time. Russia Caribbean. Finally, for upper-middle-income countries, shows a sharp break in its trend growth in the late 1990s, machinery and electronics are a major export industry for after which point cyclical variation declined. Sources of Export Growth in Developing Countries 109 For the most part, other middle-income countries expe- Mexico, Poland, and Turkey accelerated after 2000. Trend rienced more cyclical variation in GDP and sharper differ- growth was modest in Argentina and South Africa, with the ences in trend GDP growth than the BRICs (figure 9.1b). former showing large cyclical variation and the latter The Republic of Korea has had a strong positive GDP showing relatively little. Indonesia and Thailand experi- trend that weakened somewhat after 2000. Trend growth in enced large cyclical variation in GDP around the Asian Table 9.2. Export Shares by Destination and Developing-Country Income Group, 2000 and 2007 percent Income level Low Lower middle Upper middle Destination 2000 2007 2000 2007 2000 2007 Low and middle income East Asia and Pacific 12.4 15.5 6.7 10.1 4.7 9.8 Europe and Central Asia 4.3 5.3 2.1 4.4 3.2 5.9 Latin America and the Caribbean 1.0 1.1 2.1 3.8 6.8 6.5 Middle East and North Africa 0.1 0.1 0.1 0.2 0.2 0.2 South Asia 2.6 3.8 0.8 2.5 0.7 1.3 Sub-Saharan Africa 0.9 1.9 0.5 1.2 0.3 0.6 Subtotal 21.3 27.7 12.3 22.2 15.9 24.3 High income Americas 17.2 22.3 28.6 24.7 40.0 28.4 Asia 17.8 13.9 35.0 26.7 12.3 9.6 Europe 40.0 32.5 22.3 24.4 31.2 37.1 Pacific 3.8 3.6 1.7 2.0 0.7 0.7 Subtotal 78.8 72.3 87.6 77.8 84.2 75.8 Source: Author's calculations based on Comtrade (database), United Nations, http://comtrade.un.org. Figure 9.1. Log GDP with Trend and Cyclical Components in Selected Countries, 1978­2008 a. BRICs Brazil China 29 0.05 GDP and noncyclical component of GDP 28 0 27 26 ­0.05 India Russian Federation 29 0.05 28 0 27 26 ­0.05 1978 1983 1988 1993 1998 2003 2008 1978 1983 1988 1993 1998 2003 2008 GDP without cyclical component cyclical component (continued) 110 Managing Openness Figure 9.1. (continued) b. Selected other middle-income countries Korea, Rep. Mexico 27.5 0.05 27 0 26.5 26 ­0.05 GDP and noncyclical component of GDP 25.5 ­0.1 Poland Turkey 27.5 0.05 27 0 26.5 26 ­0.05 25.5 ­0.1 1978 1983 1988 1993 1998 2003 2008 1978 1983 1988 1993 1998 2003 2008 Argentina Indonesia 27 0.1 0.05 26 0 25 ­0.05 GDP and noncyclical component of GDP 24 ­0.1 South Africa Thailand 27 0.1 26 0.05 0 25 ­0.05 24 ­0.1 1978 1983 1988 1993 1998 2003 2008 1978 1983 1988 1993 1998 2003 2008 Egpyt, Arab Rep. Malaysia 26 0.05 25.5 0 25 ­0.05 GDP and noncyclical component of GDP 24.5 ­0.1 24 ­0.15 Pakistan Venezuela, R. B. 26 0.05 25.5 0 25 ­0.05 24.5 ­0.1 24 ­0.15 1978 1983 1988 1993 1998 2003 2008 1978 1983 1988 1993 1998 2003 2008 GDP without cyclical component cyclical component (continued) Sources of Export Growth in Developing Countries 111 Figure 9.1. (continued) c. Selected high-income countries Australia Canada 0.04 GDP and noncyclical component of GDP 28 27.5 0.02 27 0 26.5 ­0.02 26 ­0.04 France Spain 28 0.04 27.5 0.02 27 0 26.5 ­0.02 26 ­0.04 1978 1983 1988 1993 1998 2003 2008 1978 1983 1988 1993 1998 2003 2008 Germany Japan 30 0.02 GDP and noncyclical component of GDP 29 0 28 ­0.02 27 ­0.04 United Kingdom United States 30 0.02 29 0 28 ­0.02 27 ­0.04 1978 1983 1988 1993 1998 2003 2008 1978 1983 1988 1993 1998 2003 2008 GDP without cyclical component cyclical component Source: Author's calculations. financial crisis in 1997­98. The crisis was accompanied by trend growth in the early 1990s, consistent with the begin- a dampening of trend growth, which later accelerated in ning of the country's economic stagnation. the early 2000s. Malaysia resembles other Asian countries All in all, middle- and high-income countries show con- with volatility in the late 1990s and an uptick in trend growth siderable variation in the pace and stability of trend growth in the 2000s. The Arab Republic of Egypt and Pakistan in GDP and in the contribution of the cyclical component exhibit stable trend growth, with only modest cyclical vari- to GDP changes. From these figures, it is not immediately ation. The República Bolivariana de Venezuela follows the apparent which component of GDP growth, the trend or Latin American pattern of strong cyclical variability and the cycle, is most important for growth in import demand. several breaks in trend growth. The magnitude of variations in the business cycle is rela- Are Global Trade Patterns Sustainable tively similar across high-income countries (figure 9.1c). in the Future? Australia, Canada, and Spain exhibit more rapid trend growth in GDP than France, Germany, the United Kingdom, To shed light on the question of the sustainability of trade and the United States. Japan shows a marked slowing in its flows, we employ a gravity model of trade. There are two 112 Managing Openness steps to this analysis. First, we apply the Eaton and Kortum foundations as a basis for production.3 Based on this model, (2002) gravity model to determine the contribution of the following findings emerge. changes in supply conditions to export growth and how First, developing countries experienced pronounced these conditions vary across exporting countries. Second, improvements in export competitiveness relative to devel- we use the Hodrick-Prescott (1997) filter to examine how oped economies in the 2000s (table 9.3). Low-income coun- trend growth and cyclical growth are related to changes in tries increased competitiveness in agriculture, apparel and import demand. The rest of this section discusses the find- textiles, metals, and other industries and lost competitive- ings of these models. Full technical details are available in ness in food processing and extractive industries.4 Lower- the working paper version of this chapter. middle-income countries improved competitiveness in all sectors except extractive industries. Lower-middle-income countries also show the strongest relative improvement Changes in Export Competitiveness and among all income groups in five of the eight sectors (food Implications for Export Growth processing; apparel and textiles; chemicals; metals; and Is export growth in developing nations being matched by machinery and electronics). Upper-middle-income coun- improvements in their export capacity, allowing the expan- tries increased competitiveness in agriculture, food process- sion to be sustained? To address this question, we specify a ing, extractive industries, apparel and textiles, and other gravity model based on the Eaton and Kortum (2002) industries. Because developing nations saw improvements model of trade. Our sample is composed of 49 countries, (or no change) in their competitiveness in most sectors, it is which collectively accounted for 98 percent of global GDP not surprising that high-income countries saw a decline: in 2000.2 The sample covers more than 36,000 changes in high-income countries of the Organisation for Economic bilateral trade flows at the sector level from 2000 to 2007. Co-operation and Development (OECD) lost competitive- Product-level data from the Comtrade database are ness in all sectors, and high-income non-OECD countries aggregated into eight sectors: (1) agriculture, meat and lost competitiveness in all sectors but one (chemicals). dairy, seafood; (2) food, beverages, tobacco, wood, paper; Moreover, the results are largely unchanged when the (3) extractive industries; (4) chemicals, plastics, rubber; BRICs are excluded from their respective income categories. (5) textiles, apparel, leather, footwear; (6) iron, steel, and This finding indicates that other middle-income nations other metals; (7) machinery, electronics, transportation have enjoyed improvements in export competitiveness sim- equipment; and (8) other industries. Each sector is a collec- ilar to those experienced by the BRICs. tion of industries that share similar factor intensities and Second, changes in export competitiveness, whether are likely to rely on similar technological or institutional associated with changes in technology or with changes in Table 9.3. Mean Change in Export Competitiveness by Exporter Income Group, 2000­07 Income group (1) (2) (3) (4) (5) (6) (7) (8) Including BRICs Low income 0.023 ­0.030 ­0.013 0.001 0.055 0.046 ­0.003 0.064 Lower-middle income 0.014 0.032 ­0.014 0.062 0.090 0.081 0.093 0.034 Upper-middle income 0.033 0.008 0.041 ­0.003 0.055 ­0.018 0.007 0.038 High income OECD ­0.022 ­0.051 ­0.014 ­0.029 ­0.019 ­0.032 ­0.061 ­0.058 Non-OECD ­0.059 ­0.014 ­0.010 0.014 ­0.036 ­0.011 ­0.077 ­0.063 Excluding BRICs Low income 0.023 ­0.021 ­0.010 ­0.011 0.069 0.024 -- 0.064 Lower-middle income 0.012 0.035 ­0.012 0.064 0.091 0.088 0.095 0.034 Upper-middle income 0.028 0.007 0.040 ­0.003 0.056 ­0.016 0.008 0.038 High income OECD ­0.021 ­0.050 ­0.014 ­0.028 ­0.015 ­0.031 ­0.061 ­0.058 Non-OECD ­0.058 ­0.011 ­0.015 0.016 ­0.034 ­0.006 ­0.076 ­0.059 Source: Author's calculations. Note: The mean change in exporter competitiveness is the mean exporter fixed effect from gravity regressions reported in the working paper version of this chapter. Means are weighted by 2000 country export levels. Columns indicate broad sectors: (1) agriculture; (2) food, beverages, tobacco, wood products; (3) extractive industries; (4) chemicals; (5) apparel, textiles, footwear, leather; (6) iron, steel, other metals; (7) machinery, electronics, transportation equipment; (8) other industries. -- = not available. Sources of Export Growth in Developing Countries 113 production costs, account for approximately 11 percent of competitiveness in all eight sectors, India is in the top five the variation in bilateral export growth from 2000 to 2007. in seven sectors, Brazil is in the top five in two sectors, and GDP growth in exporting countries explains a small frac- Russia is in the top five in one sector. However, other mid- tion of changes in export competitiveness, suggesting that dle-income nations have also shown strong growth in changes in export capabilities are not simply a byproduct export competitiveness, with Turkey in the top five in six of overall economic expansion. sectors, Poland in the top five in six sectors, and Argentina Third, trends in changes in export competitiveness in the top five in four sectors.5 High-income nations, in emerge at the regional level (table 9.4). East Asia and the contrast, show weak improvements in export competitive- Pacific showed improvements in export competitiveness ness. Japan and the United Kingdom are among the bot- across all sectors (with or without China included). Europe tom five performers in six of the eight sectors, and the and Central Asia, Latin America and the Caribbean, and United States is among the bottom five performers in four South Asia increased competitiveness in all sectors but one sectors. The República Bolivariana de Venezuela is the sole (metals, chemicals, and agriculture, respectively). The middle-income nation with consistent poor performance, Middle East and North Africa increased export competi- showing up in the bottom five countries in seven of the tiveness in three sectors: agriculture, apparel and textiles, eight sectors. and machinery and electronics. Sub-Saharan Africa, in Finally, several secondary trends can be seen in the contrast to other developing regions, did not improve results of the gravity model. First, for apparel and other export competitiveness in any sector. Overall, East Asia industries (primarily instruments and specialized equip- showed the largest or second-largest improvement in ment), distance became a less important impediment to export competitiveness (competitiveness increased in six trade in the 2000s. Second, in agriculture, metals, and of the eight sectors) and Sub-Saharan Africa showed the machinery and electronics, lack of a common language largest decline (competitiveness decreased in six of the became a more important impediment to trade over the eight sectors). period. Third, colonial heritage became a more important While BRICS are not the only drivers of the rising com- factor for trade in the agriculture, chemicals, apparel, and petitiveness, their performance is still remarkably impres- metal sectors over the period. Finally, having an existing sive relative to other large economies (figure 9.2). China is regional trade agreement or a new agreement post-2000 is one of the top five countries in improvement of export uncorrelated with trade growth. Table 9.4. Mean Change in Export Competitiveness by Exporter Region, 2000­07 Region (1) (2) (3) (4) (5) (6) (7) (8) a Low- and middle-income countries East Asia and Pacific 0.037 0.035 0.027 0.052 0.096 0.107 0.054 0.033 Europe and Central Asia 0.038 0.029 0.075 0.041 0.087 ­0.041 0.092 0.161 Latin America and the Caribbean 0.028 0.011 0.029 ­0.025 0.042 0.007 0.034 0.011 Middle East and North Africa 0.025 ­0.006 ­0.006 0.000 0.073 ­0.014 0.031 ­0.021 South Asia ­0.010 0.036 0.103 0.065 0.060 0.086 0.078 0.018 Sub-Saharan Africa ­0.036 ­0.076 ­0.061 ­0.037 ­0.062 ­0.018 ­0.037 ­0.038 High income all regions ­0.022 ­0.050 ­0.013 ­0.027 ­0.021 ­0.031 ­0.062 ­0.059 Low- and middle-income countriesb East Asia and Pacific 0.037 0.042 0.027 0.055 0.097 0.109 0.056 0.033 Europe and Central Asia 0.026 0.027 0.070 0.044 0.096 ­0.041 0.096 0.166 Latin America and the Caribbean 0.022 0.009 0.028 ­0.026 0.042 0.009 0.034 0.011 Middle East and North Africa 0.026 ­0.006 ­0.003 0.001 0.074 ­0.013 0.032 ­0.020 South Asia ­0.010 0.038 0.102 0.069 0.060 0.088 0.078 0.018 Sub-Saharan Africa ­0.034 ­0.070 ­0.059 ­0.043 ­0.054 ­0.019 ­0.039 ­0.038 High income all regions ­0.022 ­0.050 ­0.014 ­0.027 ­0.017 ­0.031 ­0.062 ­0.058 Source: Author's calculations. Note: The mean change in exporter competitiveness is the mean exporter fixed effect from gravity regressions reported in the working paper version of this chapter. Means are weighted by 2000 country export levels. Columns indicate broad sectors: (1) agriculture; (2) food, beverages, tobacco, wood products; (3) extractive industries; (4) chemicals; (5) apparel, textiles, footwear, leather; (6) iron, steel, other metals; (7) machinery, electronics, transportation equipment; (8) other industries. a. Including BRICs. b. Excluding BRICs. 114 Managing Openness Figure 9.2. Change in Relative Export Competitiveness of Selected Countries by Industry, 2000­07 a. Agriculture b. Food, beverages, tobacco, and wood products Brazil China Egypt, Arab Rep. Poland Poland Argentina Argentina Egypt, Arab Rep. China India Turkey Turkey South Africa Malaysia Malaysia Thailand Pakistan Pakistan Spain Brazil Japan Spain Thailand Australia Canada France Germany Russian Federation Korea, Rep. Germany India Indonesia France Mexico Russian Federation Korea, Rep. Indonesia United Kingdom United States South Africa Mexico Japan United Kingdom United States Venezuela, R. B. Canada Australia Venezuela, R. B. ­0.1 ­0.05 0 0.05 0.1 0.15 ­0.1 ­0.05 0 0.50 0.1 0.15 change in export competitiveness change in export competitiveness c. Extractive industries d. Chemicals Egypt, Arab Rep. Egypt, Arab Rep. Turkey China China India India Turkey Argentina Poland Brazil Thailand Pakistan Pakistan Russian Federation Brazil Poland Canada Mexico Russian Federation Australia Spain United States Indonesia Korea, Rep. Malaysia Canada Argentina Venezuela, R. B. Australia Japan Germany Thailand France South Africa Korea, Rep. Malaysia United States Germany South Africa Indonesia Mexico France Japan Spain United Kingdom United Kingdom Venezuela, R. B. ­0.1 ­0.05 0 0.05 0.1 0.15 ­0.15 ­0.1 ­0.05 0 0.05 0.1 change in export competitiveness change in export competitiveness (continued) Sources of Export Growth in Developing Countries 115 Figure 9.2. (continued) e. Apparel, textiles, footwear, and leather f. Iron, steel, and other metals Turkey China China India Brazil Turkey Poland Mexico India Thailand Egypt, Arab Rep. Malaysia Spain Pakistan Pakistan Poland Argentina Indonesia France Korea, Rep. Thailand Brazil Germany Spain Malaysia South Africa Mexico Canada Indonesia Germany United Kingdom United States South Africa Egypt, Arab Rep. United States France Canada Australia Japan Venezuela, R. B. Russian Federation United Kingdom Venezuela, R. B. Japan Australia Russian Federation Korea, Rep. Argentina ­0.1 ­0.05 0 0.05 0.1 0.15 ­0.1 ­0.05 0 0.05 0.1 0.15 change in export competitiveness change in export competitiveness g. Machinery, electronics, and transport equipment h. Other industries China Poland Turkey Turkey Poland Russian Federation India India Argentina China Egypt, Arab Rep. Mexico Pakistan Argentina Mexico Egypt, Arab Rep. Thailand Malaysia South Africa Germany Brazil Brazil Russian Federation Indonesia Indonesia South Africa Korea, Rep. Australia Spain Korea, Rep. Germany Pakistan France France Malaysia Spain Australia Thailand Venezuela, R. B. Canada Japan United Kingdom United Kingdom United States Canada Japan United States Venezuela, R. B. ­0.1 ­0.05 0 0.05 0.1 0.15 ­0.2 ­0.1 0 0.1 0.2 change in export competitiveness change in export competitiveness What Is Driving the Demand for Imports--Trend contribution of different components of GDP growth to Growth or Cyclical Growth? growth in demand for imports. We replace importer fixed effects in the gravity model with importer GDP broken Are middle-income economies experiencing long-run down into two components: change in the trend compo- growth based on growth in total factor productivity, capital nent captures long-run economic growth, and change in the investment, and skill upgrading, or are they experiencing cyclical component captures changes in business cycle con- simply cyclical growth? In a second exercise, we estimate a ditions. We use the Hodrick-Prescott filter to decompose modified version of the gravity model to evaluate the 116 Managing Openness GDP based on annual data from the World Development Africa shows improvements in export competitiveness in Indicators database over the period 1960­2008.6 no sectors. Overall, the primary source of variation in import We also find that changes in relative exporter capabil- demand seems to be variation in trend GDP, rather than ities can account for approximately 11 percent of the variation in business cycle fluctuations. While growth in variation in bilateral export growth over 2000­07. Thus, bilateral trade is strongly positively correlated with changes important sources of trade growth are improvements in in trend GDP, it is weakly correlated with changes in the exporter technology and reductions in exporter produc- cyclical component of GDP, suggesting that the recent tion costs (part of which may be related to global pro- growth in trade derives from structural economic growth duction networks). in importing nations. Moreover, it seems that GDP growth plays only a mod- At the country level, more specific trends emerge in est role in changes in countries' export competitiveness. the 2000s. Income growth was associated almost entirely That countries vary in their sectoral rankings of improve- with the GDP trend in the BRICs and many Asian ments in export competitiveness is initial evidence this is nations7 (for example, Indonesia, Malaysia, and Thailand), not the case. Thus, other factors must also be at work, pre- but other middle-income countries had a more varied sumably related to technological advancement in exporter experience: in Latin American nations (for example, technology and reductions in exporter production costs. Argentina, Mexico, and the República Bolivariana de Eaton and Kortum (2002) demonstrate that countries that Venezuela) the cyclical variation seemed to be a more are more integrated into global production networks may important component of growth. Moreover, the GDP have lower production costs, which would show up in a trend in large high-income nations is considerably smaller higher exporter fixed effect (or in time differences in larger in comparison to middle-income countries (and to increases in exporter fixed effects). However, given the con- smaller high-income nations, such as Australia, Canada, centration of these networks in just two of the eight sectors and Spain). (apparel and textiles; machinery, electronics, and trans- Finally, the gravity estimation reveals that changes in portation), cross-border trade in inputs is unlikely to be an trend GDP are strongly positively associated with trade important factor in changes in export competitiveness in growth, while cyclical variation in GDP positively affects the other sectors (agriculture, food processing, extractive trade in just one industry, metals. These results are silent industries, chemicals, and metals). on whether the flexible trend GDP growth estimated over Finally, we find that the recent growth in trade was the 1960­2008 period will continue. What they do show is driven by trend growth, as opposed to cyclical growth, in that the trend component of GDP is what matters for trade importing nations. What does this mean for future growth growth. To the extent that GDP trends in middle-income in trade among developing countries? If GDP trends in nations continue, there is every reason to believe they will middle-income countries persist, we have a strong basis for continue to be an important source of demand for growth predicting that developing nations will be a significant in imports. source of growth in demand for imports in the coming decade. This possibility is especially important for growth in trade, given the likelihood that the European Union, Conclusion Japan, and the United States may have anemic growth in In this chapter, we examine the contribution of changes the medium run. in export competitiveness and importer GDP growth to growth in exports by developing countries over the Notes 2000­07 period. One of the main findings is that develop- 1. It should be emphasized that these figures understate the impor- ing countries have enjoyed significant improvements in tance of low- and middle-income destinations, owing to the sample their export competitiveness since 2000. East Asia and restrictions placed on the data. See the working paper version of this the Pacific exhibited the strongest performance, with chapter for further details. 2. The list of importers is presented in the working paper version of improvement in export competitiveness in all sectors this chapter. To remove zero trade associated with small samples, the sam- (with or without China); Europe and Central Asia, Latin ple excludes very small exporters (countries with a population in 2000 of America and the Caribbean, and South Asia show less than 1 million inhabitants) and small importers (countries with less improvements in export competitiveness in most sectors; than 0.1 percent share of world GDP in 2000). Due to missing GDP data in the World Development Indicators our sample also excludes Taiwan, the Middle East and North Africa shows improvements in China; several Gulf States (Bahrain, Kuwait, Oman, Qatar, and the United export competitiveness in a few sectors; and Sub-Saharan Arab Emirates); war-torn countries (Afghanistan, Iraq, and Somalia, and Sources of Export Growth in Developing Countries 117 the Palestinian territories); and tightly controlled economies (Cuba, the Bibliography Democratic People's Republic of Korea, Myanmar, and Zimbabwe). 3. This aggregation scheme is similar to that used by Harrigan (1997) de la Torre, A. 2010. "Update on LAC and the Global Crisis: The Worst Is and Romalis (2004) for Standard Industrial Classification industries. See Over but What Lies Behind?" Mimeo, World Bank, Washington, DC. the working paper version of this chapter for further details. Eaton, J., and S. Kortum. 2002. "Technology, Geography, and Trade." 4. There was little relative change in the other industries. Econometrica 70: 1741­79. 5. Changes in export competitiveness could in part reflect unmea- Harrigan, J. 1997. "Technology, Factor Supplies, and International Spe- sured changes in trade barriers in destination countries (for example, for cialization: Estimating the Neoclassical Model." American Economic Poland and Turkey) or the effects of currency devaluation during the 2000 Review 87 (4): 475­94. to 2007 period (for example, in Argentina). Hodrick, R. J., and E. C. Prescott. 1997. "Postwar U.S. Business Cycles: An 6. We use the same sample of 49 countries as in the first exercise. See Empirical Investigation." Journal of Money, Credit and Banking 29 (1): the working paper version of this chapter for further details on the con- 1­16. struction of this model. Romalis, J. 2004. "Factor Proportions and the Structure of Commodity 7. This occurred outside of the Asian financial crisis period in 1997­98. Trade." American Economic Review 94 (1): 67­97. 10 China's Trade and Investment with the South Pre- and Postcrisis Jing Wang and John Whalley China has seen a remarkable eightfold increase in real per rates always mean the rate compared with the same period capita gross domestic product (GDP) since the mid-1970s. of the previous year). From small levels of trade in 1995, That growth has been accompanied by a large increase in China had become India's largest export market and India trade and, since the early 1990s, inward foreign investment. had become China's fourth-largest market by 2007. Trade More recently, it is China's expanding engagement with the with Brazil also showed rapid growth, at a factor of 18 South--Africa, developing Asia, and Latin America-- between 1995 and 2007. which has been changing rapidly and growing at higher Given this pace of change, we make some forward pro- rates than total trade and investment, albeit from a small jections assuming that average 2005­07 growth rates of base. China's growing engagement with other developing trade remain unchanged (which in the current crisis is a countries is the focus of this chapter, and we summarize strong assumption). Our findings show an even greater data both pre- and postcrisis, as well as making forward acceleration in these trends as the levels of fast-growing projections. To our knowledge, no others have attempted Southern trade rise. Our projections indicate that China's this analysis. We draw heavily on Chinese source material. Southern trade will grow to 50 percent of its total trade by The picture that emerges is one of rapid change. The 2020 and to 60 percent of its total trade by 2027. The com- share of Southern trade in China's total trade shows little pounding effects of the extremely high growth rates of change in the 1990s but then begins to increase at an accel- China's trade with India and Brazil are such that in a sur- erating rate. Southern trade is now over 25 percent of prisingly short period these bilateral trade flows will come China's total trade--35 percent of imports by value and in to dominate China's trade. China, in turn, will even more the 20s for exports. Trade with developing countries--both quickly become the world's largest trading entity. imports and exports--relative to trade with developed We then turn to China's foreign direct investment (FDI) countries, after some relative slippage in the 1990s, began to involvement with the South. Data show China's FDI pick up speed following China's entry into the World Trade inflows from the South as small and only slowly changing, Organization (WTO), and its share increased consistently with a small dip in the early 2000s and a rise in 2006­07. though the 2000s until 2007. On the import side, imports of Shares are steady in the 10 percent range of total FDI oil and materials, especially from Africa, are key. On the inflows. On the FDI outflow side, the picture is different, export side, exports to countries elsewhere in Asia show the with sharp growth in recent years. This increase is reflected largest growth. in even higher growth rates of FDI from China to Africa, By country, some notable changes occur in China's Brazil, India, and elsewhere than occurs with trade. FDI trade. Especially prominent is trade between India and flows to India increased by over ninetyfold during the China, which increased by a factor of 33 between 1995 and 1995­2007 period. 2007, or at a little less than four times faster than the We also discuss how the financial crisis has seemingly growth rate of China's total trade (in this chapter, growth affected the rapid growth of China's involvement with the 119 120 Managing Openness South. We present data on trade for 2008 and early 2009 Hong Kong SAR, China; Japan; the Republic of Korea; that show a fall in China's Southern trade but at a much Singapore; and Taiwan, China. Trade with Hong Kong smaller rate than the drop in China's total trade. The share SAR, China, and Taiwan, China, grows at about the same of Southern in total trade thus continues to rise. According rate as China's total trade, and we emphasize that including to data on FDI flows in 2008, China's inflows and outflows those trade flows in Southern trade produces considerably both kept growing, but Southern growth rates were a little lower growth rates in China-South trade flows than we below those of total growth. FDI outflows doubled in 2008, report later. This grouping of Asian countries that we and inflows increased by over 20 percent. denote as Asia* in this chapter includes India, Indonesia, Finally, we place this discussion in a broader policy con- Malaysia, Thailand, and Vietnam. The South here is thus text, noting China's growing ties with India and China's Africa, Asia*, and Latin America. Because we exclude other regional trade arrangements, which are heavily Southern in countries, a small downward bias may therefore exist in focus. Combined with the emergence of an assertive South our calculations of Southern trade shares for China. on climate change policy, given the data we present, South- We begin with the precrisis period (1995­2007). The ern countries appear likely to increase their focus on China size of China's trade (imports and exports) with major both diplomatically and strategically. Southern countries and regions for 1995, 2001, and 2007 is set out in table 10.1. The absolute value of exports to Southern regions and major countries almost doubled China's Trade with the South, 1995­2007 (except Mexico with a ninefold increase) between 1995 and Although, as we show, China's trade shares with the South 2001, and exports in 2007 increased by a factor over 6 to have been increasing, we need to give a careful definition of every region compared to 2001. China's exports to India South in this context. The issue is particularly important in increased by a factor of 13 between 2001 and 2007. China's the cases of Hong Kong SAR, China, and Taiwan, China, 2007 exports to and imports from the South are 21 percent which are not members of the Organisation for Economic and 35 percent of total trade, respectively. Over 70 percent Co-operation and Development (OECD), have substantial of China's Southern trade is with developing countries in trade with China, and have high GDP per capita. Here we Asia. Thus, trade has increased steadily over time, with a define the South as Africa, Latin America, and Asia, excluding dramatic change between 2001 and 2007 (figure 10.1). Table 10.1. China's Total Trade (Imports and Exports) with the South, 1995, 2001, and 2007 US$, billions Exports Imports Region or country 1995 2001 2007 1995 2001 2007 Africa 2.5 6.0 37.3 1.4 4.8 36.4 Asia*a 14.3 26.1 172.3 12.0 39.1 250.9 Latin America 3.1 8.2 51.5 3.0 6.7 51.1 Total South 19.9 40.4 261.1 16.4 50.6 338.3 East and South Asia India 0.8 1.9 24.0 0.4 1.7 14.6 Indonesia 1.4 2.8 12.6 2.1 3.9 12.4 Malaysia 1.3 3.2 17.7 2.1 6.2 28.7 Thailand 1.8 2.3 12.0 1.6 4.7 22.7 Vietnam 0.7 1.8 11.9 0.3 1.0 3.2 Latin America Argentina 0.3 0.6 3.6 0.4 1.3 6.3 Brazil 0.8 1.4 11.4 1.2 2.3 18.3 Mexico 0.2 1.8 11.7 0.2 0.8 3.3 Other countries Egypt, Arab Rep. 0.4 0.8 4.4 0.03 0.08 0.2 South Africa 0.6 1.0 7.4 0.7 1.2 6.6 Source: China Statistical Yearbook 1996, 2002, 2008, National Bureau of Statistics, http://www.stats.gov.cn/english/ statisticaldata/yearlydata/. a. Asia* excludes Japan; Hong Kong SAR, China; Korea; Singapore; and Taiwan, China. China's Trade and Investment with the South Pre- and Postcrisis 121 Figure 10.1. China's Exports and Imports with the South, 1994­2008 500 450 400 350 US$, billions 300 250 200 150 100 50 0 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 year export import Source: Authors' calculations based on data from China Statistical Yearbook, 1996­2009, National Bureau of Statistics, http://www.stats.gov.cn/ english/ statisticaldata/yearlydata/. The growth of China's trade with Southern regions and Export growth rates after 2001 were at least 40 percent major countries is remarkably high, with the highest, beyond those before 2001 (Indonesia) and increased by Indian-Chinese trade, running at 35 percent per year over three times in the highest case (India). Import growth rates the whole period and accelerating to 64 percent on the increased by at least 10 percent (Africa before 2001 had export side in 2007. The growth rates of China's Southern high growth of 39 percent) and at most 1.8 times (South trade in all time periods are much higher than those of Africa). For total Southern trade, growth rates of exports Northern trade, as would be expected in light of figure and imports both doubled after 2001. These differences in 10.1. On the export side, China's exports to India are the the growth rates of China's Southern and Northern trade highest; growth rates of exports to Brazil are also high. On are large, reflecting the more rapid growth trend of China's the import side, China's imports from Africa increased Southern trade after 2001. most during 1995 to 2007, but these growth rates have We can also compare the growth rates of China's Southern been slowing in recent years. The same is true for Asia* and trade to those of worldwide South-South trade. These data for total trade with the South. Growth rates of imports show China's share of South-South trade growing by a little from Latin America and South Africa, however, continue to over 3 percent to 9 percent on the export side and 12 per- increase. cent on the import side by 2006. China's Southern trade China's entry into the WTO has also apparently had growth was more rapid than world South-South trade significant impacts on China's Southern trade. Kowalski growth, and this activity is reflected in China's share of and Shepherd (2006) observe that trade barriers affect- world South-South trade, which increased consistently ing South-South trade are much higher than those affect- between 1995 and 2002. For 2007, the ratio is likely higher ing North-North or North-South trade, according to again, given the growth of China's trade. 1985­2002 data. Their econometric analysis suggests that Table 10.3 reports the shares of China's trade with South- South-South trade is more sensitive to tariff reductions, ern countries in major product groups in 2007. On the and their simulations also suggest that China would gain export side, China exported 22 percent of its total exports to more than twice as much from liberalization of trade with the South. Among major product groups, 36.4 percent of Latin America, the Middle East and North Africa, and Sub- China's exports of chemicals and 29.3 percent of its exports Saharan African countries as from liberalization with other of transport equipment went to the South. Also, the South Asian countries. bought about 30 percent of textiles, clothing, and base metal China's Southern trade performance after entry into the and related products. The rapid growth of China's exports WTO supports this position. Table 10.2 reports changes in to the South, especially after its entry into the WTO, average growth rates for subperiods before and after 2001. implies growing Southern demand for Chinese products. 122 Managing Openness Table 10.2. Average Annual Growth Rates in China's Trade with the South, Pre- and Post-2001 percent Export Import Region or country 1995­2001 2002­07 Changea 1995­2001 2002­07 Changea Africa 20.0 35.9 1.8 38.8 42.0 1.1 Asia* 15.6 37.0 2.4 25.3 36.8 1.5 Latin America 20.1 36.5 1.8 20.3 41.6 2.0 South (above) 16.8 36.6 2.2 25.2 37.8 1.5 North (rest of world) 11.7 27.3 2.3 9.6 21.2 2.2 South Brazil 25.9 44.1 1.7 15.8 42.9 2.7 India 19.7 53.7 2.7 30.1 46.0 1.5 Indonesia 20.7 28.5 1.4 15.2 21.5 1.4 South Africa 18.5 39.1 2.1 12.7 35.2 2.8 Source: Authors' calculations based on data from China Statistical Yearbook 1996­2008, National Bureau of Statistics, http://www .stats.gov.cn/english/statisticaldata/yearlydata/. a. Change = growth rate after 2001/growth rate before 2001. Table 10.3. China's Southern Trade Shares in Major Product applied in 2005­07. These projections inevitably imply Groups, 2007 growing Southern trade for China, but the speed and percent potential implications for the trading system are striking. Exports to Imports from The simulation considers what could happen if China's the South Share the South Share Southern trade were to keep growing at these rates out to Nonmanufactured 24.5 Nonmanufactured 70.5 different dates. We also assume China's trade with the Manufactures 21.3 Manufactures 26.4 world grows at rates applying in 2005­07. For these projec- Chemicals 36.4 Oil 78.9 tions, the sum of China's Southern and Northern trade in Textiles, yarn, and Minerals the year yields China's total trade, and we project trade levels clothing 28.8 (excluding oil) 63.8 with the South in the future. Base metals and Jewelry, precious related products 28.7 metals and products 36.9 According to these projections, China's trade with the Transport equipment 29.3 Electronics 32.5 South will grow rapidly. Although at present China's Source: Authors' calculations based on Development Research Center of the Southern trade is substantially less than half of its total State Council of China database, www.drcnet.com.cn. trade, its more rapid growth implies that China's Southern export and import shares will keep rising relentlessly. Also, as China's exports grow more rapidly than imports, China's On the import side, the South dominated China's trade surplus will keep growing. If China's trade with India imports of nonmanufactured products in 2007. And the were to continue on the same track as during 2005­07, South is an extremely important supplier of strategic mate- India would become China's largest market in about rials for China, accounting for 78.9 percent of China's 10 years, and China's trade with the South would be about imports of oil and 63.8 percent of its imports of other min- twice that with the North. Even though China's productiv- erals. These data reinforce the impression of increasing ity growth, the absorptive capacities of Brazilian and trade links between China and the South. Indian markets, and countries' economic growth in the All these data thus seem to point in one direction: medium and long term are uncertain, these projections China's trade involvement with the South is growing rap- indicate what could happen if China's Southern trade were idly and is accelerating. The fastest-growing patterns of this to keep growing at 2005­07 rates. trade appear to be with other large and rapidly growing Using the same assumptions, we can also project when Southern economies: Brazil and India. China's Southern trade will dominate China's foreign trade. These projections are shown in figures 10.2 and 10.3. China's exports to the South will exceed 50 percent and 60 China's Future Trade with the South percent of its total exports by 2023 and 2030, respectively, Next we turn to some projections of China's future South- and imports from the South will exceed 50 percent and ern trade based on the annual rates of trade growth that 60 percent of China's total imports by 2013 and 2017, China's Trade and Investment with the South Pre- and Postcrisis 123 Figure 10.2. Projections of China's Southern Export Share, 2008­30 70 export share (%) 60 50 40 30 20 10 0 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Source: Authors' calculations. Figure 10.3. Projections of China's Southern Import Share, 2008­30 90 import share (%) 80 70 60 50 40 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Source: Authors' calculations. respectively. South-South trade will exceed 50 percent of reports the levels involved for various subregions, and table China's total trade by 2020 and will exceed 60 percent by 10.5 gives the country breakdown for 2007. Total Chinese 2027. The South will be China's major trade partner sur- outward foreign investment over the whole period prisingly soon if these growth rates are maintained. 1978­2001 was only US$8.4 billion, in contrast to annual flows today of over US$2 billion. Although statistics before 2003 are lacking, using the limited information available China's FDI Flows with the South we can still calculate growth rates for China's foreign In addition to its trade relationships, China has also experi- investment. enced growing interactions with Southern economies China's annual FDI inflows increased from an average through FDI flows, especially through outflows from of US$45 billion between 1997 and 2002 to US$74 billion China. FDI inflows and outflows both began from very in 2007, and the Southern share increased from 3.2 percent small bases in the 1990s, but on the outflow side they have to 6.2 percent. Outflows grew from US$2.8 billion in 2003 experienced rapid growth over the past few years. As with to US$26 billion in 2007, and the Southern share doubled trade, we can evaluate whether China's Southern invest- from 8 percent to 16.5 percent. Much more rapid growth ment (both inflows and outflows) grew more quickly than therefore occurs for the Southern share than for the North- China's total foreign investment flows. In passing, we note ern share, and the change (from a small base) is larger than that the South here is the South as above, and in the process for trade. In addition, by region and country the geograph- we exclude the Cayman Islands and the British Virgin ical composition of FDI flows are more balanced than Islands as tax haven­based conduits. Both of them rank in those for trade. the top five country destinations for Chinese outward FDI. Table 10.6 reports the annual average growth rates of Before China's entry into the WTO, its outward invest- China's FDI flows by region and country. The contrast ment was very small. Data on China's Southern investment between Northern and Southern flows is striking. Inflows activity is reported in tables 10.4 and 10.5. Table 10.4 from the South grew substantially but declined in 2007, 124 Managing Openness Table 10.4. China's Average FDI Inflows and Outflows to the South and the World, 1997­2007 Inflows from Outflows to South World South share South World South share Time period (US$, millions) (US$, millions) (%) (US$, millions) (US$, millions) (%) 1997­2002 (average) 1,690.3 45,229.0 3.7 -- -- -- 2003­06 (average) 3,064.5 59,369.9 5.2 760.9 9561.9 8.0 2007 4,669.1 74,767.9 6.2 4,366.4 26,506.1 16.5 Source: Authors' calculations. Inflow data based on China Statistical Yearbook 2008, National Bureau of Statistics, China, http://www.stats.gov.cn/english/ statisticaldata/yearlydata/. Outflows data are from 2007 Statistical Bulletin of China's Outward Foreign Direct Investment, Ministry of Commerce of the People's Republic of China, http://hzs.mofcom.gov.cn/accessory/200909/1253868856016.pdf. Note: The inflow data use China's foreign direct investment flows and are smaller than data from the UNCTAD database. The outflow data are nonfinancial outward FDI flows. -- = not available. Table 10.5. China's FDI Inflows from and Outflows to the South, 2003 and 2007 US$, millions Inflows Outflows Region, country, 2003 2007 2003 2007 or group (US$, millions) (US$, millions) 2007/2003 (US$, millions) (US$, millions) 2007/2003 Africa 617.8 1,486.8 2.4 74.8 1,574.3 21.0 Asia*a 1,423.2 2,187.5 1.5 198.0 2,367.4 12.0 Latin America*b 263.6 994.8 3.8 21.9 424.7 19.4 Total 2,304.6 4,669.1 2.0 294.6 4,366.4 14.8 Southern regions, countries, or groups ASEAN*c 867.0 1,206.7 1.4 122.5 570.4 4.7 ASEAN 2,925.4 4,391.2 1.5 119.3 968.1 8.1 Brazil 16.7 31.6 1.9 6.7 51.1 7.7 India 15.9 34.0 2.1 0.2 22.0 146.8 South Africa 32.5 69.2 2.1 8.9 454.4 51.3 Source: Authors' calculations based on China Statistical Yearbook 2008, 2007, National Bureau of Statistics, http://www.stats.gov.cn/english/statisticaldata/ yearlydata/; Statistical Bulletin of China's Outward Foreign Direct Investment, Ministry of Commerce of the People's Republic of China, http://hzs.mofcom .gov.cn/accessory/200909/1253868856016.pdf. Note: ASEAN = Association of Southeast Asian Nations. a. Asia* excludes Japan; Hong Kong SAR, China; Korea; Singapore; and Taiwan, China. b. Latin America* denotes Latin America except the Cayman Islands and the British Virgin Islands. c. ASEAN* denotes the ASEAN except Singapore (Brunei Darussalam, Cambodia, Indonesia, Lao People's Democratic Republic, Malaysia, Myanmar, the Philippines, Thailand, and Vietnam). with sharp declines for Brazil, India, and South Africa. The are volatile: both grew, but in 2007 the outflow to the South growth rates on outflows from China are dramatically dif- was sharply higher, which produces a higher average growth ferent, with a growth rate in 2007 of many hundreds of rate over the period 2004­07. As with trade, the picture con- percent. According to the World Investment Report 2008 veyed by FDI data is one of relatively accelerating integra- (UNCTAD 2008), FDI inflows into developing countries tion with the South through FDI flows, especially on the increased 21 percent over 2006, which is almost the same outflow side. as the growth rates of African and Latin American FDI Similar to the earlier trade projections, we can project into China but a slower growth rate for FDI of Asian China's FDI with regions and major countries assuming that developing countries into China. China's Southern FDI growth rates between 2005 and 2007 remain unchanged into inflow increased only 16.7 percent. In 2007, FDI outflows the future. Again, the sum of Southern and Northern pro- from South, East, and Southeast Asia reached a new high, jected FDI yields China's total FDI. with growth of 46 percent. According to these calculations, China's FDI inflows and Table 10.7 reports growth rates of China's Southern FDI outflows will grow very quickly if growth rates between and total FDI. Here, the striking feature on the inflow side 2005 and 2007 are sustained. More striking growth would is the persistently higher growth rate of Southern-origin occur on the outflow side. China's outward investment FDI. On the outflow side, the relative North-South outflows boomed in recent years, and its growth rates between 2005 China's Trade and Investment with the South Pre- and Postcrisis 125 Table 10.6. Average Annual Growth of China's Southern and Northern FDI Flows, 1998­2007 percent Inflows Outflows Region or country 1998­2007 2005­07 2007 2004­07 2005­07 2007 Africa 35.8 24.6 22.1 145.8 86.3 202.8 Asia* 6.5 7.1 11.4 115.9 140.3 343.9 Latin America* 32.4 64.7 21.4 165.5 115.1 333.7 South (above) 13.2 18.2 16.7 118.0 112.5 279.3 North (rest of the world) 5.5 7.6 19.7 75.9 120.1 157.7 Southern countries, regions, or groups ASEAN* 5.0 4.9 10.6 60.5 73.8 180.1 ASEAN 3.6 13.2 31.0 86.4 94.0 181.1 Brazil 202.4 21.0 43.1 126.2 169.4 406.7 India 356.9 39.9 35.0 866.2 1110.5 292.5 South Africa 100.4 13.6 27.1 368.2 457.3 1015.4 Source: Authors' calculations based on national sources. China Statistical Yearbook 2008, 2007, National Bureau of Statistics, http://www.stats.gov.cn/english/ statisticaldata/yearlydata/; Statistical Bulletin of China's Outward Foreign Direct Investment, Ministry of Commerce of the People's Republic of China, http://hzs .mofcom.gov.cn/accessory/200909/1253868856016.pdf. Note: ASEAN = Association of Southeast Asian Nations. a. Asia* excludes Japan; Hong Kong SAR, China; Korea; Singapore; and Taiwan, China. b. Latin America* denotes Latin America except the Cayman Islands and the British Virgin Islands. c. ASEAN* denotes the ASEAN except Singapore (Brunei Darussalam, Cambodia, Indonesia, Lao People's Democratic Republic, Malaysia, Myanmar, the Philippines, Thailand, and Vietnam). Table 10.7. Growth of China's Southern FDI and Total FDI, 1998­2007 percent Inflow from Outflow to Year South World Difference South World Difference 1998 13.5 0.5 13.0 -- -- -- 1999 19.3 11.3 8.0 -- -- -- 2000 15.7 1.0 14.8 -- -- -- 2001 15.8 15.1 0.7 -- -- -- 2002 28.1 12.5 15.6 -- -- -- 2003 0.1 1.4 1.6 -- -- -- 2004 23.4 13.3 10.0 134.4 92.6 41.8 2005 9.4 0.5 9.9 31.4 123.0 91.7 2006 28.6 4.5 24.1 26.9 43.8 16.9 2007 16.7 18.6 1.9 279.3 50.3 229.0 Average 13.2 5.5 7.7 118.0 77.4 40.6 Source: Authors' calculations based on national sources: China Statistical Yearbook 2008, 2007, National Bureau of Statistics, http://www.stats.gov.cn/english/statisticaldata/yearlydata/; Statistical Bulletin of China's Outward Foreign Direct Investment, Ministry of Commerce of the People's Republic of China, http://hzs.mofcom.gov.cn/accessory/200909/ 1253868856016.pdf. Note: -- = not available. and 2007 were very large. China's investments in India, for tions suggest that the Southern share of China's total out- instance, increased by 11 times in these years. These projec- ward investments will decline in the future. The picture tions thus paint a dramatic picture, but China would on the inflow side, however, is the opposite. China's FDI remain a large source of FDI even if these growth rates inflows from the South grew more rapidly in recent years were only half as high. than those from the North, with the result that China's Since developed-country markets are the main destina- FDI inflows from the South will exceed 50 percent and 60 tion for China's outward investments and have had higher percent of China's total FDI inflow by 2022 and 2024, growth rates in recent years than the South, these projec- respectively. These projections are displayed in figure 10.4. 126 Managing Openness Figure 10.4. Projections of Southern FDI Inflow Share for China, 2008­30 90 80 70 60 percent 50 40 30 20 10 0 29 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 30 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Source: Authors' calculations. Table 10.8. China's Southern Trade Pre­ and Post­financial Crisis, 2008 and 2009 US$, billions Exports Imports 2008 2009 2008 2009 Region or country Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Region Africa 10.1 12.9 14.6 13.2 10.1 11.7 12.8 13.9 16.2 15.6 10.2 5.7 9.5 13.3 Asia* 46.4 57.4 64.7 55.8 39.5 47.8 54.3 71.3 79.6 88.5 64.3 46.8 60.3 71.7 Latin America 13.9 18.1 22.5 17.0 10.5 12.6 16.2 14.6 19.3 22.8 14.8 9.4 16.5 20.6 South 70 89 102 86 60 72 83 100 115 127 89 62 86 106 Country India 6.9 8.7 8.8 7.1 6.0 7.0 8.1 6.3 7.1 4.3 2.7 3.3 3.3 3.2 Indonesia 3.6 4.4 5.2 4.0 2.6 3.5 3.9 3.7 4.1 3.9 2.6 2.3 3.3 3.7 Malaysia 4.7 5.3 5.8 5.5 3.6 4.7 5.3 7.4 8.5 9.5 6.7 5.5 7.0 9.2 Thailand 3.5 4.2 4.4 3.4 2.6 3.0 3.7 6.2 6.4 7.0 5.9 4.4 6.0 7.0 Vietnam 4.1 4.4 3.3 3.3 2.8 3.6 4.5 1.1 1.0 1.3 1.0 0.9 1.2 1.2 Argentina 1.1 1.5 1.6 0.9 0.6 0.8 1.1 1.7 1.8 3.4 2.5 0.9 1.2 1.3 Brazil 3.6 4.8 6.4 3.9 2.3 2.7 4.3 4.9 8.1 10.5 6.1 3.4 8.1 9.7 Mexico 2.8 3.6 4.0 3.4 2.3 2.8 3.4 0.9 1.0 1.0 0.8 0.8 0.9 1.0 Egypt, Arab Rep. 1.3 1.5 1.7 1.4 1.1 1.3 1.3 0.1 0.2 0.1 0.1 0.0 0.1 0.3 South Africa 1.9 2.2 2.4 2.1 1.6 1.6 2.0 2.0 2.6 2.4 2.2 1.5 2.2 2.5 Source: Authors' calculations based on monthly data of China Customs Statistics, China Customs Information Center, http://www.haiguan.info. The Financial Crisis and China's China's trade with major Southern countries and regions Southern Trade presents the same trend between the first quarter of 2008 and the third quarter of 2009. Trading volumes increased China's export-oriented economy has been severely steadily before the third quarter of 2008; then they decreased affected by the financial crisis, and so we report only pre- for two quarters with a trough in the first quarter of 2009, liminary data for 2008 and 2009 on Southern trade and rebounding slowly from the second quarter of 2009. shares (table 10.8). For trade data, the absolute values of Table 10.9 reports the commodity composition of China's Southern trade are considerably higher in 2008 China's total imports between 2008 and 2009 (bilateral than 2007 but decline substantially in the spring of 2009 data by commodity are unavailable). These largely remain and increase again in the summer of 2009. Trading levels in as they were precrisis, with declines in mineral fuels and the second and third quarters of 2009 are still far less than growth in industrial products. These data suggest that the levels in 2008. On both the export and the import side, China's trade structure has not changed during the crisis China's Trade and Investment with the South Pre- and Postcrisis 127 Table 10.9. Commodity Composition of China's Imports in the Financial Crisis, 2008 and 2009 percent 2008 2009 Item Jan. Mar. May Jul. Sept. Nov. Jan. Mar. May Jul. Sept. Primary products 30.7 31.9 34.0 33.5 32.4 29.1 27.0 27.0 28.6 30.4 28.5 Food and live animals 1.3 1.2 1.3 1.1 1.2 1.2 1.7 1.7 1.6 1.5 1.4 Beverages and tobacco category 0.2 0.1 0.2 0.1 0.1 0.4 0.2 0.1 0.2 0.1 0.2 Non-food raw materials (excluding fuel) 15.3 14.3 15.0 14.8 15.4 14.8 14.2 14.5 14.2 14.2 14.1 Oil 12.9 14.9 16.3 16.2 14.3 11.1 9.9 9.1 10.9 12.5 10.8 Mineral fuels, lubricants and related materials 13.2 15.2 16.6 16.5 14.8 11.5 10.4 9.9 11.9 13.8 11.9 Animal and vegetable oils, grease, wax 0.7 1.0 1.0 1.0 0.8 1.2 0.4 0.8 0.8 0.9 0.9 Industrial products 69.3 68.1 66.0 66.5 67.6 70.9 73.0 73.0 71.4 69.6 71.5 Chemical products and related products 11.1 10.7 10.4 11.4 10.3 9.7 11.4 11.7 11.5 11.1 11.0 Manufactured goods classified by raw materials 9.9 9.8 9.6 9.3 8.9 9.7 9.9 10.9 12.1 11.1 10.6 Steel 2.4 2.5 2.2 2.3 2.4 2.6 3.0 3.1 3.0 2.6 2.6 Base Metals and Related Products 6.6 6.6 6.2 6.0 5.9 6.1 6.7 7.5 8.5 7.7 7.5 Machinery and transport equipment 38.7 38.4 36.8 37.3 39.8 42.2 43.1 41.7 39.3 38.7 40.9 Miscellaneous manufactured articles 9.3 8.9 8.8 8.2 8.2 8.6 8.1 8.4 8.2 8.3 8.7 Unclassified goods 0.2 0.3 0.4 0.2 0.4 0.7 0.5 0.3 0.3 0.3 0.3 Source: Authors' calculations based on China Customs Statistics, China Customs Information Center, http://www.haiguan.info. but that China's manufacturing industries may have On the import side, China's Northern trade performed declined because of the financial crisis. better than Southern trade. Imports from India were only As table 10.3 shows in part, the South is the major source half those of the spring and summer of 2008. of China's oil and mineral imports, with the oil share increas- Figures 10.6 and 10.7 suggest that China's imports from ing from 54 percent to 79 percent during 2004 and 2007 and the South in summer 2009 increased over the same period the mineral shares constant around 60 percent between 1999 for the previous year at a slower rate. They grew more rapidly and 2007. Also China's import shares of other product compared to spring 2009, raising the share of Southern groups from the South have increased in recent years. Table trade in the summer of 2009. China's exports to and 10.9 shows that China's import share of oil deceased signifi- imports from the South in the second quarter of 2009 were cantly after November 2008 and that it began to rebound 26 percent and 36 percent of total trade, respectively, slowly after May 2009, while the level of September 2009 is increasing by 1.6 points and 3 points, respectively, com- still lower than that of 2008. The level of September 2009 pared to shares in the first quarter of 2009. This share of imports partly reflects the fall in oil prices and partly the exports is higher than that of 2008, but the share of imports compression of China's domestic industries, findings sup- is still lower by 2.3 percentage points than that of 2008. ported by data on industrial consumption of power. These numbers suggest that China's Southern trade has Figure 10.5 reports growth rates of China's Southern played a more prominent role in its trade recovery on the trade postcrisis, which plays a more important role during export side. The relative fall on the import side reflects lower this time. It grew by more than 30 percent in 2008; in con- oil prices and large oil imports from Africa. trast, the rate of China's Northern trade growth was only 13 percent. (China's total trade grew 18 percent in 2008.) China's Trade Performance and That of Other Compared with the same period in 2008, China's trade Asian Countries during the Financial Crisis with almost every market contracted during the spring and summer of 2009. The exceptions are exports to Africa, We next draw a comparison between the trade performance which increased by 0.7 percent in the first quarter of 2009, of other Asian countries and China during the financial cri- and imports from Vietnam, which increased by 22 percent sis. Monthly growth rates of major Asian economies' total in the second quarter of 2009. On the export side, China's exports and imports (compared with the same month of Southern trade fell less than Northern trade during the the previous year) are reported in figures 10.8 and 10.9. first half of 2009, but Latin American trade fell the most. Asian trade fell rapidly amid the global financial crisis. 128 Managing Openness Figure 10.5. Growth Rates of China's Southern Trade Pre­ and Post­financial Crisis Compared to Same Period of Previous Year 50 40 30 20 10 percent 0 ­10 ­20 ­30 ­40 ­50 1 2 3 4 1 2 3 Q Q Q Q Q Q Q 08 08 08 08 09 09 09 20 20 20 20 20 20 20 China's exports south's exports north's exports China's imports south's imports north's imports Source: Authors' calculations based on data from China Customs Statistics, China Customs Information Center, http://www.haiguan.info. Figure 10.6. China's Regional Export Shares Pre­ and Post­financial Crisis, 2007­09 30 25 20 percent 15 10 5 0 1 2 3 4 1 2 3 4 1 2 3 Q Q Q Q Q Q Q Q Q Q Q 07 07 07 07 08 08 08 8 9 9 09 0 0 0 20 20 20 20 20 20 20 20 20 20 20 quarter Asia* Africa Latin America South ASEAN* European Union Source: Authors' calculations based on the data from China Customs Statistics, China Customs Information Center, http://www.haiguan.info. Note: ASEAN = Association of Southeast Asian Nations; ASEAN* = ASEAN excluding Singapore. China's Trade and Investment with the South Pre- and Postcrisis 129 Figure 10.7. China's Regional Import Shares Pre­ and Post­financial Crisis, 2007­09 45 40 35 30 percent 25 20 15 10 5 0 1 2 3 4 1 2 3 4 1 2 3 Q Q Q Q Q Q Q Q Q Q Q 07 07 07 07 08 08 08 08 09 09 09 20 20 20 20 20 20 20 20 20 20 20 quarter Asia* Africa Latin America South ASEAN* European Union Source: Authors' calculations based on the data from China Customs Statistics, China Customs Information Center, http://www.haiguan.info. Note: ASEAN = Association of Southeast Asian Nations; ASEAN* = ASEAN excluding Singapore. Figure 10.8. Export Growth Rates of Key Asian Economies during the Financial Crisis Compared with Same Month of Previous Year, 2008­10 80 60 40 % change 20 0 ­20 ­40 ­60 08 8 08 08 08 8 09 9 09 09 09 9 10 0 00 00 00 00 01 20 20 20 20 20 20 20 20 20 .2 .2 .2 .2 .2 n. ay l. p. n. ay l. p. n. ar ov ar ov ar Ju Ju Ja Ja Ja Se Se M M M M M N N month China India Japan Korea, Rep. Source: Authors' calculations based on national sources: China Customs Statistics, China Customs Information Center, http://www.haiguan.info; Indian DGCIS (Kolkata) Database, Ministry of Commerce and Industry of India, http://www.dgciskol.nic.in/; Trade Statistics of Japan, Ministry of Finance of Japan, http://www.customs.go.jp/toukei/info/tsdl_e.htm; Korean Customs Statistics, Korea Customs Service, http://english.customs.go.kr. 130 Managing Openness Figure 10.9. Import Growth Rates of Key Asian Economies during the Financial Crisis Compared with Same Month of Previous Year, 2008­10 100 80 60 40 % change 20 0 ­20 ­40 ­60 08 8 08 08 08 8 09 9 09 09 09 9 10 0 00 00 00 00 01 20 20 20 20 20 20 20 20 20 .2 .2 .2 .2 .2 n. ay l. p. n. ay l. p. n. ar ov ar ov ar Ju Ju Ja Ja Ja Se Se M M M M M N N month China India Japan Korea, Rep. Source: Authors' calculations based on national sources: China Customs Statistics, China Customs Information Center, http://www.haiguan.info; Indian DGCIS (Kolkata) Database, Ministry of Commerce and Industry of India, http://www.dgciskol.nic.in/; Trade Statistics of Japan, Ministry of Finance of Japan, http://www.customs.go.jp/toukei/info/tsdl_e.htm; Korean Customs Statistics, Korea Customs Service, http://english.customs.go.kr. Among these countries, Korea has performed much bet- related products, Korean imports grew the most rapidly ter than the others, with strong positive growth in the in 2008 and fell sharply in 2009, while Japan's grew the whole of 2008 on both the export and the import sides, and most slowly precrisis and dropped most postcrisis. has experienced a smaller decline than the others in 2009. China's demand for base metals dropped considerably On the export side, the worst performer was Japan. With during the crisis and started to rebound in May 2009, fol- sluggish growth precrisis, Japan's exports shrank over 30 lowing China's stimulus package, almost two quarters percent for 10 months, and they were almost half those of ahead of the rebounds in other countries' imports of base the previous year in the first quarter of 2009. On the metals. import side, in addition to the strong performance by Next we compare two major partners and assess their Korea, data on the other countries' imports seemed to fol- trade performance with individual Asian countries. One is low a pattern that the faster the growth rate had been pre- the United States, the largest developed country and the crisis, the larger the decline was during the crisis. The trade source of the global financial crisis, and the other is China, of all countries started to rebound at the end of 2009, with the largest developing country and neighbor of these Asian a solid rebound in 2010. countries. Growth rates of trade between the United States Next we compare the situation of Asian countries' and Asian countries are reported in figures 10.10 and imports during the crisis. Demand for these materials 10.11. Japan's exports to the United States started to decline reflected the situation of the real economies. For imports in February 2008 and fell the most in the first quarter of of oil, these countries' demand grew rapidly before Octo- 2009, with less than half the exports of the same period in ber 2008 and then fell sharply (except for Korea). Because 2008. Asian countries' exports to the United States all fell the fluctuations in global oil prices influenced every sharply from January 2009 to August 2009, with a monthly country equally, the less significant decline in Korean decline of 23.9 percent for China, 58.4 percent for Japan, demand for oil indicates that the crisis had a weaker 28.2 percent for Korea, and over 30 percent for other coun- impact on the Korean economy. For other minerals, India tries. Thailand's and China's exports to the United States experienced the most rapid growth precrisis and the rebounded at the end of 2009 and followed other countries sharpest decline during the crisis. For base metals and with a significant rebound in 2010. China's Trade and Investment with the South Pre- and Postcrisis 131 Figure 10.10. Growth Rates of Asian Economies' Exports to the United States during the Financial Crisis, 2008­10 80 60 40 20 percent 0 ­20 ­40 ­60 ­80 8 8 8 8 9 09 9 9 0 00 00 00 00 00 00 00 01 20 .2 2 .2 .2 .2 .2 .2 .2 r. r. n l ct n l ct n Ju Ju Ap Ap Ja Ja Ja O O month China India Japan Korea, Rep. Malaysia Taiwan, China Thailand Source: Authors' calculations. Figure 10.11. Growth Rates of Asian Economies' Imports from the United States during the Financial Crisis, 2008­10 150 100 50 percent 0 ­50 ­100 08 8 08 08 08 8 09 9 09 09 09 9 10 0 00 00 00 00 01 20 20 20 20 20 20 20 20 20 .2 .2 .2 .2 .2 n. ay l. p. n. ay l. p. n. ar ov ar ov ar Ju Ju Ja Ja Ja Se Se M M M M M N N month China India Japan Korea, Rep. Malaysia Taiwan, China Thailand Source: Authors' calculations. 132 Managing Openness On the import side, the demand of India; Korea; Institutional Dimensions of China's Taiwan, China; and Thailand for U.S. products declined Growing Southern Links more than their exports to the United States, with monthly China's growing trade and investment links are matched declines of 46.3 percent, 44.4 percent, 65.2 percent, and by growing institutional connections with the South, 41.7 percent, respectively. China's exports to and imports based on the recognition by China and other developing from the United States are similar in trend, but imports countries of common interests. As a result, China is likely show a stronger rebound than exports in 2010, which con- to emphasize cooperation with other developing countries tributed to China's trade deficit in March 2010. Similar to as the starting point for its Southern diplomacy. As Wen exports, the most severe decline in imports also occurred in Jiabao indicated in 2004, China has good political relation- first quarter of 2009. ships with other developing countries, and economic The growth rates of other Asian countries' trade with cooperation should mutually empower them. China thus China are reported in figures 10.12 and 10.13. Exports by seeks to combine trade, investment, and foreign aid to India; Korea; Taiwan, China; and Thailand fell sharply deepen multilateral and regional cooperation with devel- from the last quarter of 2008 and declined more than oping countries and safeguard the interests of developing exports to the United States in percentage terms. Japan's countries in international trade organizations and multi- exports to China declined less than those to the United lateral mechanisms. States during the crisis, and this result is slightly better than China's diplomatic initiatives toward the South thus Japan's total export performance. All countries' exports to reflect the growing economic interest in those countries. China rebounded strongly from the end of 2009. These After its entry into the WTO, China focused on two major countries' imports from China fell sharply from the first directions in foreign trade policy. One was to support the quarter of 2009, lagging a couple of months behind exports WTO system, and the other was to try to play a more impor- to China, and started to rebound from the end of 2009, tant role in negotiations. China first tried to strengthen almost at the same time as exports. The rebound speed of cooperation, especially with developing countries, and imports is significantly lower than that of exports, which emphasize links to least-developed countries and small suggests that China's large market for foreign goods is con- economies. Another approach was to deepen bilateral and tributing to the recovery of its neighbors' trade. Figure 10.12. Growth Rates of Asian Economies' Exports to China in the Financial Crisis, 2008­10 200 150 100 percent 50 0 ­50 ­100 08 8 08 08 08 8 09 9 09 09 09 9 10 0 00 00 00 00 01 20 20 20 20 20 20 20 20 20 .2 .2 .2 .2 .2 n. ay l. p. n. ay l. p. n. ar ov ar ov ar Ju Ju Ja Ja Ja Se Se M M M M M N N month India Japan Korea, Rep. Malaysia Taiwan, China Thailand Source: Authors' calculations. China's Trade and Investment with the South Pre- and Postcrisis 133 Figure 10.13. Growth Rates of Asian Economies' Imports from China in the Financial Crisis, 2008­10 120 100 80 60 percent 40 20 0 ­20 ­40 ­60 08 8 08 08 08 8 9 9 09 09 09 9 10 0 00 00 00 00 00 01 20 20 20 20 20 20 20 20 .2 .2 2 .2 .2 .2 n. ay l. p. n. ay l. p. n. ar ov ar ov ar Ju Ju Ja Ja Ja Se Se M M M M M N N month India Japan Korea, Rep. Malaysia Thailand Taiwan, China Source: Authors' calculations. regional trade cooperation, and especially building free- Formal negotiations on a China-India FTA are still not trade agreements (FTAs). This effort began with adjacent underway, and the progress of bilateral FTA negotiations countries and then extended to a wider area, especially between China and India is slower than that between India emerging economies in Africa and Latin America. At and Japan and Korea. the same time, China sought closer cooperation with energy China has also reached agreements with Africa. Based and strategic material suppliers, especially Arab countries. on good political relations since 1949, and the simultaneous China's regional free-trade arrangements with developing liberalization of China and Africa since the 1970s, economic countries began when it joined the Bangkok Agreement in cooperation between China and Africa is growing rapidly 2001. The Bangkok Agreement settled the origin of products (Zhang Qingmin 2007). Evidence of these closer ties is the eligible for preferential concessions among members. China Forum on China-Africa Cooperation jointly set up by then concluded a framework agreement on comprehensive China and Africa in October 2000. China made major economic cooperation with the Association of Southeast commitments including debt relief of 10 billion yuan for Asian Nations (ASEAN) in 2002, a starting point for China's African countries. China also gave some of Africa's least- FTAs. China and ASEAN then agreed on tariff concessions developed countries tariff-free treatment for 190 tariff from 2005 and will now build a China and ASEAN Free- headings of goods. This tariff preference was later extended Trade Area (CAFTA) by 2015. to nearly 500 tariff headings. Negotiations on an FTA With FTAs in place with ASEAN and Pakistan, China between China and the Southern Africa Customs Union already has closer trade relationships with most major (SACU), including Botswana, Lesotho, Namibia, South developing countries in Asia except India, although here Africa, and Swaziland, started in 2004; and South Africa there is the prospect of an agreement. In April 2005, during recognized China's market economy status at the same Wen Jiabao's visit to India, the two governments announced time. Both sides are now pushing the negotiations forward a joint feasibility study on a China-India regional trade to a substantive stage. arrangement. The study was completed in October 2007 China's relationship with Latin America has also devel- and reached consensus on trade in goods, trade in services, oped rapidly following the improved openness of develop- investment, trade facilitation, and economic cooperation. ing countries in Latin America from the end of the 134 Managing Openness 1980s. A China-Chile FTA was signed in 2005 and covers links. These ties may well be poised to accelerate further in trade issues such as market access, rules of origin, sanitary the next decade. and phytosanitary measures, technical trade barriers, trade remedies, and dispute settlement mechanisms. A China- Chile investment agreement is now under negotiation. In Concluding Remarks addition, a China­Costa Rica FTA is now under negotia- This chapter assesses China's growing interactions through tion. A feasibility study was completed in 2008, and three trade and FDI with the South and also provides an initial negotiations have been carried out thus far, with some assessment of the impact of the financial crisis on China's progress on trade in goods and services. trade. Based on Chinese source data, the picture that A China-Peru FTA was signed in 2009. On trade in emerges shows rapidly growing trade and FDI links, which goods, over 90 percent of Chinese and Peruvian goods in recent years (precrisis) have been accelerating. At 2005­07 will receive zero tariff treatment in stages. On services, growth rates, we project that China's Southern trade will be both countries will open further their service markets to around 50 percent of its total by 2015 and that trade with each other, based on WTO commitments. On the invest- India will account for around 50 percent of China's trade by ment side, both countries will grant investors national 2023 (such trade increased by 33 percent between 1997 and treatment and most-favored nation treatment. China 2007). Initial data show this rapid relative growth in South- and Peru also reached consensus in other fields such as ern engagement, if anything, has been accelerated further by intellectual property rights, trade remedies, rules of ori- the financial crisis, especially on the export side. gin, customs procedures, technical barriers to trade, and China is rapidly become the manufacturing center of sanitary and phytosanitary measures. the global economy, and with more rapid growth in Brazil, The China-Peru FTA has become a symbol for new India, Indonesia, and other Southern countries than in the bilateral relations between China and other developing OECD, the prospect is for even more significant Southern countries in Latin America. For instance, Brazil is a found- investment by China. In FDI flows, this trend is seemingly ing member of the Rio Group, a member of MERCOSUR accelerating even more rapidly, if from a lower base. The (Southern Cone Common Market), a leading member of South may be poised to become the epicenter of trade in the Group of 77 and the Group of 15, and an observer of the global economy by 2030, with China at its hub. the Non-Aligned Movement and is rich in minerals and oil. Chinese and Brazilian governments have proposed in-depth consultations on issues of common interest such Note as an FTA.1 Economic cooperation between China and 1. Ministry of Foreign Affairs, People's Republic of China. "China and Brazil has become closer in recent years, and China became Brazil Signed Joint Communique." http://www.chineseembassy.org/chn/ gxh/zlb/smgg/t118446.htm. Brazil's largest trading partner in 2009. President Luiz Inácio Lula da Silva sought to strengthen cooperation between these two countries and to play a greater role in dealing with the global financial crisis when he visited Bibliography China in April 2009. During this visit, Brazil and China Kowalski, P., and B. Shepherd. 2006. "South-South Trade in Goods." Trade signed cooperation protocols involving fields such as agri- Policy Working Paper 40, Organisation for Economic Co-operation and Development, Paris. culture, science and technology, space, energy, ports coop- UNCTAD (United Nations Conference on Trade and Development). eration, mining, and renewable energy. With these, a 2007. World Investment Report 2007: Transnational Corporations, "loans-for-oil" deal over 10 years involving US$10 billion Extractive Industries and Development. Geneva: United Nations Con- ference on Trade and Development. http://www.unctad.org/en/ between the China Development Bank and the Brazilian docs/wir2007_en.pdf. national oil company (Petrobras) was signed. In addition, ------. 2008. World Investment Report 2008: Transnational Corporations, the two governments plan a Joint Action Plan 2010­14. and the Infrastructure Challenge. Geneva: United Nations Conference The picture therefore shows rapidly deepening institu- on Trade and Development. http://www.unctad.org/en/docs/wir 2008_en.pdf. tional involvement of China with the South, complementing Zhang, Q. 2007. "China's Foreign Relations." Foreign Affair Review and supporting growing Southern trade and investment 94: 22­28. 11 Volatility, Export Diversification, and Policy Mona Haddad, Jamus Lim, Laura Munro, Christian Saborowski, and Ben Shepherd Although the global economic crisis erupted in the finan- shocks are indisputably crucial in accounting for external cial markets of the industrial world, developing countries sources of variation. However, they can explain only a small have not escaped its effects. Many, including those without fraction of the long-run variation in real per capita gross close financial ties to the developed world, were driven into domestic product (GDP) (Ahmed 2003; Becker and Mauro recession, as global demand plummeted. The recent crisis 2006). The underlying institutional and policy environment has resulted in the largest drop in global trade volumes cannot be ignored (Easterly, Islam, and Stiglitz 2001). since the Second World War. Naturally, open economies In studying the mechanisms by which the trade channel heavily reliant on export revenues were among those hard- affects growth volatility, this chapter considers three ques- est hit by the crisis. This observation has led to renewed tions that have been neglected thus far. First, does the effect of questioning of the merits of export-led growth strategies trade openness on growth volatility vary with the degree of for developing countries (Harrison and Rodriguez-Clare diversification of a country's export basket? Second, if a con- 2009; Rodrik 2009). ditional relationship exists, is there a threshold--in a given While the relationship between openness and growth measure of export concentration--above which the total has been investigated thoroughly, the link between open- effect of trade openness on growth volatility changes from ness and growth volatility is less well understood. Various negative to positive?3 Third, what policies are available to studies have argued that trade openness increases macro- governments to promote export diversification efficiently? economic volatility (Rodrik 1997); yet there is no clear This chapter is based on the empirical work reported consensus in the literature to date. Di Giovanni and in technical detail in our working papers (Haddad, Lim, Levchenko (2009) find that trade openness leads countries and Saborowski 2010; Shepherd 2010; and Dennis and to become more specialized in their exports. This finding Shepherd, 2011). We show that the effect of openness on implies that a higher degree of openness not only exposes volatility indeed depends on the degree to which a coun- countries to a larger number of external shocks but that try's export basket is diversified. We then go on to demon- greater openness also makes them more vulnerable.1 At a strate that policies such as trade facilitation--reducing the more aggregate level, Easterly and Kraay (2000) find that fixed and variable costs facing exporters and importers-- terms-of-trade volatility is an important driver of growth can be effective in promoting diversification. volatility, especially for smaller states. Yet, they argue that The rest of this chapter is organized as follows. The next the high income volatility typically experienced by small section discusses the relationship among trade openness, economies is due mainly to their openness and that export growth volatility, and export diversification. The following concentration plays only a minor role.2 section shows how most countries' export baskets meet Raddatz (2007) shows that external shocks--such as the threshold of diversification and indeed benefit from those transmitted by prices, foreign growth, and real inter- trade openness. We then discuss the policies available to est rates--have a substantial and significant impact on the governments to promote export diversification and help volatility of real activity in low-income economies. External manage openness, focusing on trade facilitation. The final 135 136 Managing Openness section concludes with policy implications flowing from We find that the degree of diversification matters in the our findings. determination of this sign. The evidence suggests that the vulnerability of countries to some types of external shocks is reduced when these countries are better diversified in The Effect of Increased Trade Openness their exports. on Vulnerability to Global Shocks Our empirical analysis also suggests that while trade Due to the large number of variables included in the openness, at least initially, may induce production special- dataset, we limit our discussion here to the key dependent ization and concentration through comparative advantage, and independent variables. Full technical details are avail- financial openness may result in production diversification, able in the working paper version of this chapter.4 Our which reduces growth volatility. This argument is similar in main dependent variable is volatility of output growth, spirit to the central message of our chapter, namely, that measured as the standard deviation of GDP per capita export diversification reduces growth volatility through a across each five-year period. country's improved ability to weather economic storms. The two main independent variables of interest are Better integration of a country into a broader range of export diversification and trade openness. We include a global value chains and implicit or explicit insurance variety of export diversification measures that capture the schemes also have this effect. dimensions of product and geographical market concen- These findings are based on a panel of 77 developing tration for any given country. These are fairly standard and and developed economies over the period 1976­2005. The include the top 5 and top 10 shares of products and mar- main findings do not change much when high-income kets as well as Herfindahl-Hirschman Indexes for products economies are excluded from the analysis.7 By contrast, the and markets. We supplement these direct diversification relationship does not always hold when we exclude low- measures with summary ones that we construct using income economies from the analysis. This finding suggests principal-components analysis. We extract the first princi- that much of the action driving our results indeed lies with pal component of the three product and three market low- and middle-income economies, for which export diversification measures.5 diversification matters more in shielding their economies Consistent with much of the literature, we compute from the adverse effects of external shocks. A likely expla- trade openness as the ratio of the sum of exports and nation is that developed economies have better access to imports to GDP, while financial openness is measured with other forms of insurance schemes, whereas developing an index of restrictions on cross-border transactions taken countries depend more strongly on implicit insurance as from Chinn and Ito (2008).6 Both these indicators provide represented by a more diversified export structure. measures of the actual exposure of a country to interna- tional markets. They reflect structural as well as policy- Does Product or Market Diversification Matter Most? related characteristics of a country. Product diversification clearly moderates the effect of trade openness on growth volatility. Market diversification The Effect of Openness and Diversification on Volatility measures yield more mixed results. Evidence in favor of a One of the chief arguments against export-led growth role for market alongside product diversification in shield- strategies for developing countries is that economically ing an economy from shocks is limited at best. Further- open countries are more prone to external shocks. But are more, thresholds (at which the total effect of openness on they necessarily more strongly affected by external shocks growth volatility changes sign) of the market diversifica- by way of higher volatility? The total effect of openness on tion indicators cannot be established with confidence. In volatility could in fact be positive, due to the enhanced pos- the product diversification regression, this is not the case, sibility of international risk sharing through explicit and as the following section demonstrates. implicit insurance schemes. Examples include joint ven- tures, international lending, production diversification, Most Countries Benefit from and formal insurance contracts. Moreover, the disciplinary Increased Openness effect of international competition and the prevalence of formal international contracts could potentially limit the The second research question of this chapter concerns the risk of domestic policy mistakes. existence and extent of a threshold at which export diversi- It is therefore unclear, ex ante, whether the effect of open- fication is sufficient to ensure that openness does not have ness on growth volatility should be positive or negative. an overly large effect on volatility. Based on the econometric Volatility, Export Diversification, and Policy 137 Figure 11.1. Total Effect of Trade Openness on Growth Volatility 25 total effect of openness on volatility 20 15 10 5 0 ­5 ­10 ­15 ­20 ­25 0 10 20 30 40 50 60 70 80 90 100 export concentration (%) total effect of trade openness on growth volatility 90% confidence interval Source: Authors' calculations. results reported in the working paper, we can identify posi- Table 11.1. Thresholds and Shares Based on Error tive thresholds in our product diversification indicators at Components and System GMM Regressions for Growth Volatility on Openness and Product Diversification which the effect of openness on volatility changes sign. On the basis of our preferred model,8 this threshold occurs in Indicator Threshold Share the upper part of the distribution of the respective diversifi- Herfindahl 0.154 0.797 cation indicators. The current degrees of diversification in 5-product 0.481 0.563 the export baskets of the countries in our sample suggest 10-product 0.575 0.469 that most countries benefit from increased openness. Principal component (product) ­0.516 0.672 This relationship is illustrated by the plot in figure 11.1, Source: Authors' calculations. which uses the share of the five most important products Note: GMM = generalized method of moments. in total exports as a measure of diversification. We can see that the impact of trade openness on volatility is signifi- The value of the five-product measure lies below the cantly lower than zero, with 90 percent confidence, as long threshold of 0.48 for 56 percent of all countries (table 11.1). as a country scores lower than about 0.24 on the diversifi- In the case of the 10-product indicator, we see a similar pic- cation variable. The effect gradually increases and changes ture emerge. The total effect of trade openness on growth sign at the threshold, which is at about 0.48. In contrast, volatility is again highly significant, and the system general- above a value of about 0.71, the impact of trade openness ized method of moments (GMM) estimator points to a on growth volatility is significantly positive. With indica- threshold that lies at 0.58, which covers about 47 percent of tors of geographic diversification, however, the results are countries. For the Herfindahl indicator, this share of coun- not as telling. tries is even higher, at 80 percent. It is important to put the threshold value of 0.48 in con- The threshold of diversification can also be understood text. Based on figure 11.1, determining the share of countries relative to degrees of trade openness (figure 11.2, panel a) in the sample whose value on the five-product indicator and income per capita (figure 11.2, panel b). When com- lies below the threshold and the share of those whose value pared to trade openness, the distribution of countries lies above it is straightforward. Performing this exercise for a below the diversification threshold appears to be relatively variety of measures of export diversification, we consistently even. In contrast, countries above the threshold appear to find that the threshold value is located in the upper part of be largely clustered around moderate levels of openness the respective indicator's distribution. A large share of the (with the exception of outliers, such as Zambia). This find- sample of countries benefits from trade openness in the ing suggests that the countries that currently experience sense that it reduces the volatility of income growth. reduced volatility as a consequence of diversification are 138 Managing Openness Figure 11.2. Distribution of Countries by Product Diversification, Trade Openness, and Income per Capita a. Distribution by 10-product diversification and trade openness 100 Algeria Nigeria Botswana Iran, Islamic Rep. Malawi Zambia Syrian Arab Republic Ghana Ecuador Burkina Faso Paraguay 80 Madagascar Norway Trinidad and Tobago Bangladesh Gambia, The Bolivia Togo Philippines export concentration (%) Senegal Honduras Costa Rica Peru Chile Zimbabwe Nicaragua Kenya Jordan Ireland 60 Pakistan Colombia Israel El Salvador Dominican Republic Guatemala Morocco Tunisia Sri Lanka Malaysia Argentina Australia Panama Uruguay South Africa New Zealand Mexico Canada 40 Japan Indonesia India United Kingdom Portugal Belgium Brazil Spain Sweden France Thailand Turkey Denmark China United States Netherlands Italy 20 0.6 0.7 0.8 0.9 1.0 1.1 trade openness b. Distribution by 10-product diversification and income per capita Nigeria 100 Algeria Botswana Zambia Iran, Islamic Rep. Malawi Ghana Syrian Arab Republic Burkina Faso Ecuador Paraguay 80 Madagascar Gambia, The Bangladesh Trinidad and Tobago Bolivia Norway Togo Senegal 10-product concentration (%) Philippines Honduras Costa Rica Zimbabwe Peru Chile Jordan Nicaragua El Salvador Ireland 60 Kenya Israel Dominican Republic Pakistan Colombia Guatemala Tunisia Sri Lanka Malaysia Panama Argentina Australia Morocco Uruguay New Zealand South Africa Mexico Canada 40 Indonesia Japan United Kingdom India Portugal Belgium Brazil Spain Sweden Thailand Turkey France China Denmark United States Netherlands 20 Italy 0 10,000 20,000 30,000 40,000 GDP per capita in initial period (2000 prices, US$) Source: Authors' calculations. Volatility, Export Diversification, and Policy 139 not limited to the most open economies. Conversely, the countries that choose to pursue a diversification path do so economies that do not benefit from the diversification expediently, so that they quickly fall below the diversifica- effects of reduced volatility are not necessarily closed. tion threshold? Naturally, countries in the lower-right quadrant of the The answer is yes, as figure 11.3 clearly illustrates; it fol- plot are in a beneficial situation, given their very open and lows the path of six developing countries toward greater well-diversified economies. They should be well shielded diversification over the sample period, using the five-product against foreign shocks and benefit from less growth volatil- measure. Although countries such as Colombia, Kenya, and ity. Countries in the upper-left quadrant of the plot, in Nicaragua had very concentrated export baskets at the begin- contrast, are in a problematic situation. Botswana and ning of our sample period, they have successfully diversified Malawi, for instance, lie well above the threshold but have to levels close to or below the threshold value we have identi- relatively open economies. Their relatively open economies fied. This example illustrates that diversification--as a means may expose them to a large number of global shocks, and of deriving larger benefits from trade openness and at the their highly concentrated export baskets may make them same time shielding the economy against global shocks--is very vulnerable to those shocks. indeed a feasible and realistic policy goal. The threshold of diversification can also be understood relative to income per capita. Figure 11.2, panel b, illustrates Facilitating Export Diversification that, as expected, all high-income economies, with the exception of Norway and Ireland, have attained levels of Thus far, this chapter has highlighted the importance of diversification that lie substantially below the threshold export diversification in mediating the links between eco- value. Thus, they are likely to enjoy the benefits of trade nomic openness and volatility. In the remainder of the text, openness while being well shielded from foreign shocks we look at the ways in which developing countries can pro- through the participation in a large number of global value mote export diversification. Indeed, they have enormous chains. The vast majority of countries above the diversifica- scope to do so: although diversification has been taking tion threshold are low-income countries, although a large place in recent years, most countries still export only a rel- number of low-income economies also fall below the atively small number of products (figure 11.4) to a rela- threshold. Whereas countries such as Botswana and Nigeria tively narrow range of foreign markets (figure 11.5). are troubled by extremely high export concentration, China Dennis and Shepherd (2011) examine the policy deter- has reached a level of diversification that falls well below minants of export diversification using an econometric the threshold. The question then arises: Can developing model. They focus in particular on trade facilitation, that is, Figure 11.3. Diversification Trends among Selected Countries, 1981­2005 90 80 export concentration (%) 70 60 50 threshold 40 30 20 1981­85 1986­90 1991­95 1996­2000 2000­05 year Colombia South Africa Kenya Nicaragua Jordan Mexico Source: Authors' calculations. 140 Managing Openness Figure 11.4. Histogram of the Number of 8-Digit Product Lines Exported to the EU by Each Partner Country 8.0e-04 6.0e-04 density 4.0e-04 2.0e-04 0 0 2,000 4,000 6,000 8,000 number of products exported Source: COMEXT and authors' calculations. Figure 11.5. Histogram of the Number of Export Destinations Served by Each Country 0.06 0.04 density 0.02 0.00 0 50 100 150 number of export destinations Source: COMTRADE and authors' calculations. the range of policies that reduce the cost of exporting goods required for trading, such as a letter of credit, bill of lading, and the costs of market entry. Data on export costs are and the like; the costs related to the transportation of goods sourced from the World Bank's Doing Business project to the relevant seaport; the administrative costs related to (Doing Business Indicators database, http://www.doing customs clearance, technical controls, and inspections; and business.org). They are defined as the total official costs for ports and terminal handling charges. The indicator thus exporting a standardized cargo of goods, excluding ocean provides a useful cross-section of information on a country's transit and trade policy measures such as tariffs. The four approach to trade facilitation. It covers elements of variable main components of the costs of exporting captured here costs (transportation and handling charges) and fixed costs are the costs related to the preparation of documents (standardized document preparation). These Doing Business Volatility, Export Diversification, and Policy 141 data are collected from local freight forwarders, shipping countries. Although it is true that export diversification is lines, customs brokers, and port officials, based on a stan- thereby measured vis-à-vis the EU and not the world as a dard set of assumptions, including the traded cargo travels whole, these measures remain highly relevant, given that the in a 20-foot full container load; the cargo is valued at EU is one of the most important outlets for developing- US$20,000; and the goods do not require any special phy- country exports. tosanitary, environmental, or safety standards beyond what These results are broadly consistent with previous work is required internationally. These export operations cost as examining the diversification impacts of GDP, per capita little as US$300­US$400 in China, Israel, Singapore, income, and trade costs, such as that of Hummels and Tonga, and the United Arab Emirates, whereas they run Klenow (2005), Debaere and Mostashari (2005), and nearly 10 times that amount in Gabon and Tajikistan. On Feenstra and Kee (2008). Moreover, the finding that lower average, the cost is around US$1,278 per container, excluding market-entry costs are associated with greater export countries in the Organisation for Economic Co-operation diversification is consistent with the results of Helpman, and Development and the European Union (EU). Melitz, and Rubinstein (2008). Market-entry costs are also sourced from Doing Business. In practice, trade facilitation can have a significant The data provide indicators on the costs, time, and number impact on the range of products that countries export (fig- of procedures required for an entrepreneur to start up and ure 11.6). Concretely, a 10 percent improvement in trade formally operate a local limited liability company with gen- facilitation--as measured by the Doing Business data set's eral industrial or commercial activities. This process cost of exporting--is associated with product diversity gains includes legal preregistration, registration, and postregistra- on the order of 3­4 percent. Moreover, differentiated goods tion requirements. Only official costs are considered, based (such as manufactures) appear to have stronger diversifica- on information gathered from the company law, commer- tion responses to trade facilitation than do homogeneous cial code, and specific regulation and fee schedules. goods (such as agricultural products). When export costs are Dennis and Shepherd (2011) find that market-entry disaggregated into four components--customs costs, docu- costs, export costs, distance, and foreign tariffs negatively mentation costs, inland transport costs, and port costs--it affect product diversification. Market size, GDP per capita, appears that customs costs are the primary driver of the and exporter country tariffs, however, are positively related finding on trade facilitation. to the degree of market diversification. These findings are Recent findings also suggest that geographical diversifica- based on new measures of export diversification con- tion--an expansion in the number of foreign markets served structed from EU data on imports from 118 developing by a country's exports--can be an important mechanism Figure 11.6. Number of Products Exported versus Doing Business Cost of Exporting 8,000 6,000 number of products 4,000 2,000 0 0 1,000 2,000 3,000 4,000 export cost ($) number of products exported fitted values Source: Authors' calculations. 142 Managing Openness Figure 11.7. Number of Export Markets versus Doing Business Cost of Exporting 150 number of products 100 50 0 0 1,000 2,000 3,000 4,000 export cost (US$) number of export markets fitted values Source: Authors' calculations. through which developing countries can become more tariffs increase the number of export markets by 4 percent integrated in the world trading system. Shepherd (2010) and 3.5 percent, respectively. A similar reduction in export finds that increases in export costs (figure 11.7), distance as a costs relative to per capita GDP increases geographical proxy for transport costs, and tariffs are consistently asso- diversification by more than 12 percent. These results line ciated with a decrease in geographical diversification of up well with the trade facilitation literature, in which meas- exports. However, as the size and development of the home ures that reduce nontariff trade costs are usually found to economy increase, the number of export markets tends to have bigger trade impacts than tariff cuts (see, for example, grow. Moreover, improved trade facilitation--that is, lower Hertel and Keeney 2006). export costs at home--has stronger potential for increasing Another way of looking at the impact of lower export geographical diversification than do comparable changes in costs is in terms of their absolute U.S. dollar level. With market access abroad or international transport costs. per capita income constant, reducing the U.S. dollar cost Concretely, 10 percent reductions in international of exporting in Tajikistan (the highest-cost market, transport costs and importer tariffs are associated with US$4,300) to the level of the median country (St. Lucia, increases of 2 percent and 5 percent, respectively, in the US$1,053) would be associated with an increase of nearly number of export destinations. A 10 percent increase in the 40 percent in the number of foreign markets entered. size of the domestic market is associated with increased geographical diversification of 3 percent. The elasticity of Conclusion export destinations with respect to export costs is weaker than for distance or tariffs: a reduction of 10 percentage It is widely believed that trade openness is, under suitable points in the ratio of export costs to per capita income is conditions, positively associated with growth. But does that associated with a nearly 1.5 percent increase in the number growth come at the cost of more volatility due to greater of foreign markets served. vulnerability to global shocks? After all, an open economy How important are trade costs for geographical diversi- may face a larger number of adverse shocks than a more fication of exports in a quantitative sense? To examine this closed economy. question, we consider one-standard-deviation decreases in This study explores the relationship among diversifica- each of the three factors of trade cost independently, that tion, trade openness, volatility, and policy. We find strong is, changing one variable but keeping all others constant. evidence for the role of export diversification in reducing the Evaluated at the sample mean, one-standard-deviation vulnerability of countries to global shocks. Indeed, openness reductions in transport costs (distance) and in overseas appears to reduce volatility in diversified economies and to Volatility, Export Diversification, and Policy 143 have the opposite effect in poorly diversified economies. variable costs of moving goods across borders) can be Our empirical findings suggest that product diversification highly effective in promoting export diversification. Focus- in particular plays an important role in shielding an econ- ing on removal of red tape affecting exports and imports omy from the detrimental impact of foreign shocks. The and promoting the development of trade-related infra- evidence for market diversification is more mixed. There structure and services sectors can, therefore, make a major appears to be a threshold for product diversification at contribution to diversifying exports and helping manage which the effect of openness on growth volatility switches outward orientation. sign. Most countries appear to experience less growth volatility as they pursue increased openness to trade. These findings are of major relevance for policy makers Notes in developing countries. One case against export-led 1. Finally, the authors illustrate that more open sectors also become increasingly detached from the overall economy in their growth processes, growth strategies for developing countries is that econom- thus leading to less volatility in growth economy-wide. However, they ically open countries are more likely to be buffeted by do find that this latter effect is smaller in magnitude relative to the ear- external shocks. This observation could well be true, but lier two. the relevant question is whether the combined impact of 2. A number of papers--including Bevan, Collier and Gunning (1993), Dehn (2000), and Kose and Riezman (2001)--have documented these shocks is large and whether the effect of trade open- important effects of commodity price shocks on growth volatility. ness on volatility is indeed positive. For reasons discussed 3. To our knowledge, these questions have not been addressed in the before, the theoretical case is indeterminate. This chapter empirical literature. The closest study in spirit to ours is Jansen (2004). She uses a cross-section of countries to show, first, that export concentra- has shown that the effect of openness on growth volatility tion determines terms of trade (ToT) volatility and, second, that ToT is likely to be negative when a country possesses a suffi- volatility drives income volatility. However, the paper does not directly ciently diversified export basket. These findings amount to test how the link between openness and income volatility is affected by different levels of diversification. It does not establish confidence-bound a powerful argument in favor of making export diversifica- thresholds at which the total effect of openness on growth volatility tion a first-order policy concern for developing countries changes sign. Last but not least, the author does not use market and prod- as they consider exit strategies from the global financial uct conceptions of diversification, as we do in this study. crisis. At the same time, this research serves as an impor- 4. In our preferred specification, we include the following control variables: inflation volatility, exchange-rate volatility, the volatility of cap- tant counterargument against the recent rise in protection- ital flows to the region, and an indicator for the frequency of systemic ist sentiment worldwide (Baldwin and Evenett 2009). banking crises, as well as the volatility of foreign shocks, such as foreign How can policy be used to enhance diversification? We growth volatility and terms-of-trade volatility. The robustness checks in the working paper expand this set to include several additional controls show that reducing the costs facing exporters and poten- that may also potentially affect volatility. tial exporters--particularly export costs and market-entry 5. Both first-principal components capture more than 85 percent of barriers--can be highly effective in promoting export the variation in the underlying variables, which makes it acceptable to rely on the first-principal component alone. diversification, both in the product and in the geographi- 6. In addition to these measures, we have explored alternative meas- cal dimensions. Of course, these policies not only affect ures of trade and financial openness, such as the import share of GDP and export diversification but also have a range of other bene- the ratio of foreign direct investment and portfolio liabilities to GDP, ficial impacts on the economy. The overall cost-benefit respectively. Our central results are not altered, although some of the con- trol variables lost statistical significance (while maintaining their signs). balance of reform is likely to be strongly in their favor. 7. See the working paper version of this chapter (Haddad et al. 2010) Furthermore, our findings suggest a more phased for further information on robustness checks. approach toward introducing trade reform. Countries 8. The preferred model is the system GMM estimator. See the work- ing paper for full details. should pursue an expanded production base and export diversification strategies before broad tariff removal. This sequencing of liberalization efforts--especially for coun- Bibliography tries that currently have a very concentrated export base-- Ahmed, S. 2003. "Sources of Economic Fluctuations in Latin America and may be important for minimizing the disruptive effects Implications for Choice of Exchange Rate Regimes." Journal of Devel- that expanded trade could have on growth volatility. opment Economics 72 (1): 181­202. This approach does not mean protecting domestic pro- Baldwin, R. E., and S. J. Evenett, eds. 2009. The Collapse of Global Trade, Murky Protectionism, and the Crisis: Recommendations for the G20. ducers with "infant industry" tariffs--a classic inward- London: Centre for Economic Policy Research/VoxEU.org. oriented strategy--but rather an outward-oriented one in Becker, T. I., and P. Mauro. 2006. "Output Drops and the Shocks which barriers to domestic market entry are removed to that Matter." Working Paper 06/172, International Monetary Fund, Washington, DC. encourage innovation and development of new markets by Bevan, D. L., P. Collier, and J. W. Gunning. 1993. "Trade Shocks in Devel- companies at home. Strong evidence also suggests that bet- oping Countries: Consequences and Policy Responses." European ter trade facilitation (through the reduction of fixed and Economic Review 37 (2-3): 557­65. 144 Managing Openness Chinn, M. D., and H. Ito. 2008. "A New Measure of Financial Openness." Paper 15261, National Bureau of Economic Research, Cambridge, Journal of Comparative Policy Analysis 10 (3): 309­22. MA. Debaere, P., and S. Mostashari. 2005. "Do Tariffs Matter for the Extensive Helpman, E., M. Melitz, and Y. Rubinstein. 2008. "Estimating Trade Flows: Margin of International Trade? An Empirical Analysis." Discussion Trading Partners and Trading Volumes." Quarterly Journal of Econom- Paper 5260, Centre for Economic Policy Research, London. ics 123 (2): 441­87. Dehn, J. 2000. "The Effects on Growth of Commodity Price Uncertainty Hertel, T., and R. Keeney. 2006. "What Is at Stake? The Relative Impor- and Shocks." Policy Research Working Paper 2455, World Bank, tance of Import Barriers, Export Subsidies, and Domestic Sup- Washington, DC. port." In Agricultural Trade Reform and the Doha Development Dennis, A., and B. Shepherd. 2011. "Trade Facilitation and Export Diversi- Agenda, ed. K. Anderson and W. Martin, 37­62. Washington, DC: fication." World Economy 34 (1): 101­22. World Bank. Di Giovanni, J., and A. A. Levchenko. 2009. "Trade Openness and Volatil- Hummels, D., and P. Klenow. 2005. "The Variety and Quality of a Nation's ity." Review of Economics and Statistics 91 (3): 558­85. Exports." American Economic Review 95 (3): 704­23. Easterly, W. R., R. Islam, and J. E. Stiglitz. 2001. "Shaken and Stirred: Jansen, M. 2004. "Income Volatility in Small and Developing Economies: Explaining Growth Volatility." In Annual World Bank Conference on Export Concentration Matters." Discussion Paper 3, World Trade Development Economics, ed. B. Pleskovic and N. H. Stern, 191­211. Organization, Geneva. Washington, DC: World Bank. Kose, M.A., and R. Riezman. 2001. "Trade Shocks and Macroeconomic Easterly, W. R., and A. C. Kraay. 2000. "Small States, Small Problems? Fluctuations in Africa." Journal of Development Economics 65 (1): Income, Growth, and Volatility in Small States." World Development 55­80. 28 (11): 2013­27. Raddatz, C. E. 2007. "Are External Shocks Responsible for the Instabil- Feenstra, R. C., and H. L. Kee. 2008. "Export Variety and Country Pro- ity of Output in Low-Income Countries?" Journal of Development ductivity: Estimating the Monopolistic Competition Model with Economics 84 (1): 155­87. Endogenous Productivity." Journal of International Economics 74 (2): Rodrik, D. 1997. "Has Globalization Gone Too Far?" Washington, DC: 500­518. Institute for International Economics. http://bookstore.piie.com/ Haddad, M., J. J. Lim, and C. Saborowski. 2010. "Trade Openness Reduces book-store/57.html. Growth Volatility When Countries Are Well Diversified." Policy ------. 2009. "Growth after the Crisis." Discussion Paper 7480, Centre for Research Working Paper 5222, World Bank, Washington, DC. Economic Policy Research, London. Harrison, A. E., and A. Rodriguez-Clare. 2009. "Trade, Foreign Invest- Shepherd, B. 2010. "Geographical Diversification of Developing Country ment, and Industrial Policy for Developing Countries." Working Exports." World Development 38 (9): 1217­28. 12 The Effects of Exports on Productivity and Volatility: Evidence from Malaysian Firm-level Data Mona Haddad, Deborah Winkler, and Albert Zeufack The large and rapid slowdown in economic activity since (see figure 12.1, panel a). In 2007, Malaysia was ranked the 2007 has resulted in even larger and more rapid declines in world's 18th largest goods exporter, compared to 23rd in international trade, with potentially devastating effects on 1990 and 30th 1980. It exceeded the export volumes of much economic growth, value added, and employment in export- larger countries, such as Brazil, India, and South Africa. oriented developing countries. The situation is particularly While exports grew strongly, imports expanded even more serious, because since the 1980s developing countries have quickly, resulting in a mostly negative contribution of net experienced large increases in the ratio of exports to gross exports--that is, exports minus imports--to real GDP domestic product (GDP) (Milberg and Winkler 2010). That growth (see figure 12.1, panel b). This result reflects the change has led to a renewed interest in the merits and costs country's assembly-type export sector, which is character- of export-led growth strategies, particularly for developing ized by a large import content (World Bank 2010). countries (Rodrik 2009). Using the example of Malaysia, This chapter is structured as follows. In section two, we this chapter addresses the gains from exports in the form of estimate the impact of exports on productivity at the plant higher productivity and the risk of outward orientation in level in Malaysia and also take plant-level characteristics the form of volatility in output growth. into account. Section three focuses on the effect of sectoral Malaysia's experience is particularly useful for our study exports on the volatility of output growth at the firm level because government policy gradually shifted from import and on the role of product and market diversification of substitution in the 1960s to export-oriented industrializa- exports. Section four concludes and makes some policy tion from the 1970s onward. Export promotion was "the recommendations for Malaysia. vehicle for achieving the twin objectives--alleviation of poverty and restructuring of employment and the owner- The Effect of Exports on Productivity ship of assets--of the New Economic Policy" (Khalafalla Motivation and Webb 2001). Moreover, since the establishment of a free-trade zone in Penang in 1971, Malaysia has success- Advocates of the export-led growth hypothesis generally fully attracted multinational firms, initially from Japan and refer to the success stories of the Asian Tigers--Hong Kong the United Sates and later from Europe. Export-led indus- SAR, China; the Republic of Korea; Singapore; and Taiwan, trialization has transformed Malaysia into Asia's third- China--that were characterized by high and continuous most-open economy in trade terms (World Bank 2010). growth and rapid industrialization between the early 1960s As a result, Malaysia's combined exports of goods and and the 1990s. That success is attributed to their export- services as a percentage of GDP more than doubled oriented free-market economies. Most developing coun- from 46 percent in 1974 to its peak of 121 percent in 1999 tries that followed an inward-oriented import substitution 145 146 Managing Openness Figure 12.1. Malaysian Exports, 1971­2008, and Their Contribution to Real GDP Growth, 1988­2009 a. Exports 110 100 90 80 share of GDP (%) 70 60 50 40 30 20 10 0 19 1 19 2 19 3 19 4 75 19 6 77 19 8 79 19 0 19 1 19 2 83 19 4 19 5 19 6 87 19 8 89 19 0 19 1 19 2 93 19 4 19 5 19 6 97 19 8 20 9 20 0 20 1 02 20 3 20 4 05 20 6 20 7 08 7 7 7 7 7 7 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 0 0 0 0 0 0 19 19 19 19 19 19 19 19 19 20 20 year goods exports services exports b. Contribution of net exports to real GDP growth 20 15 10 5 percent 0 ­5 ­10 ­15 ­20 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 year net exports, contribution to real GDP growth (points) real GDP growth (annual) Sources: World Development Indicators (panel a) and EIU Country Data (panel b). strategy (mostly in Latin America) experienced relatively The channels through which export expansion poor economic achievements, which were taken as further enhances aggregate productivity and growth are well evidence of the validity of the export-led growth model known. Exports allow for specialization according to a (Medina-Smith 2001; Furuoka 2007). country's comparative advantage, which thereby raises The Effects of Exports on Productivity and Volatility: Evidence from Malaysian Firm-Level Data 147 growth. Export orientation also includes dynamic effi- characteristics--namely, export duration, the share of ciency gains in the more productive export sector due to skilled labor, the share of private domestic ownership, higher competition, greater economies of scale, better and the share of private foreign ownership--to test capacity utilization, the dissemination of knowledge and whether the impact of exports on productivity is different technological progress, and improved allocation of scarce taking these characteristics into account. Fourth, we apply resources throughout the economy. Moreover, higher modern estimation techniques including the dynamic sys- exports are linked to productivity gains, which should tem generalized method of moments (GMM) estimator. lead to real wage increases and potentially to employment And fifth, our data allow us to test the export-led growth creation, thereby fostering domestic demand, growth, hypothesis for more recent years. and consequently the economic well-being of a country. Finally, increased export earnings relax current account Key Findings pressures by making it easier to import necessary inter- mediate and capital goods and to attract foreign invest- We estimate the impact of exports on both TFP and LP. At ment. Such activity stimulates growth through capital the firm level, several determinants of TFP have been iden- accumulation (World Bank 1993; Medina-Smith 2001; tified. We capture knowledge by the share of high-skilled Mahadevan 2007). labor in a firm's total labor force. In general, technology is At the firm level, several determinants of total factor pro- measured by a firm's R&D intensity. However, our dataset ductivity (TFP) have been identified, of which knowledge reports R&D expenses only for 2001 and 2006 and only for (human capital), technology, and technology spillovers a few firms. We therefore use a firm's telecommunication appear to be the most important. Besides a firm's own costs as a proxy for telecommunication technology. Finally, research and development (R&D) intensity, technology can we include several variables as proxies for technology be acquired by technology spillovers through economic spillovers; most important are exports, but we also include integration into world markets, including exports and for- the share of imported material inputs in total material eign direct investment but also through imported interme- inputs (off-shoring) and sectoral trade liberalization. We diates and trade liberalization (see, for example, Damijan include tariffs in the equation to account for the literature et al. 2003; Fernandes 2008). Labor productivity (LP) at the around the Melitz (2003) model that focuses on the effect firm level is affected mainly by capital intensity and TFP. of trade liberalization on productivity. The hypothesis of the export-led growth model is far We hypothesize that a higher share of high-skilled labor from being universally accepted, because of ambiguous and a larger use of telecommunication technology both results and econometric weaknesses of earlier studies. We increase TFP. Analogously, we expect exports to have a pos- put this hypothesis to the test by estimating the impact of itive effect on TFP. The impact of materials off-shoring is exports on TFP and LP in the Malaysian manufacturing not straightforward and depends on whether Malaysian sector using plant-level data for 1999­2006. To our knowl- manufacturing firms source high- or low-quality material edge, no study at the firm level has yet focused on the inputs from abroad. We expect trade liberalization to impact of exports on productivity. Our analysis attempts increase TFP because of more competition among firms. to fill this gap by using plant-level data of Malaysian We estimate the impact of exports on TFP for the manufacturing firms. The data are retrieved from the Malaysian manufacturing sector using plant-level data for "Productivity and Investment Climate Survey," which is a the period 1999­2006. Our data cover two subperiods, collaborative effort of the Malaysian government and the 1999­2001 and 2004­06. In our baseline specification, we World Bank. Especially relevant to our study is the avail- find that the share of high-skilled labor in total labor and ability of plant-level data on exports. The presentation spending on telecommunication technology both signifi- here is nontechnical and focuses on key results and inter- cantly increase TFP. Exports also have a significantly posi- pretations. Full details of data, estimation techniques, and tive impact on TFP and show the biggest positive elasticity regression results are available in the working paper ver- (figure 12.2). Materials off-shoring significantly lowers sion of this chapter. TFP, indicating that Malaysian manufacturing firms source This section contributes to the vast export-led growth low-quality inputs from abroad. As expected, higher tariffs literature in several ways. First, no previous studies to our at the sectoral level lower TFP. knowledge have directly measured the effect of export The effect of exports on TFP might be influenced by indi- levels on productivity. Second, we use plant-level data for vidual firm-level characteristics such as export duration, the Malaysia, while most of the export-led growth literature skill intensity of labor, the share of private domestic owner- focuses on either time-series data or cross-country ship, and the share of foreign ownership. In a second step, we regressions. Third, we interact exports with four plant interact exports with such characteristics. We expect that 148 Managing Openness Figure 12.2. Firm-Level Exports and Total Factor Figure 12.3. Firm-Level Exports and Labor Productivity in the Productivity for the Malaysian Manufacturing Sector, 2006 Malaysian Manufacturing Sector, 2006 10 16 14 8 log(TFP) log(LP) 12 6 10 4 8 15 20 25 30 15 20 25 30 log(exports) log(exports) Source: Authors' calculations. Source: Authors' calculations. firms with a longer export duration might have knowledge A larger share of private foreign ownership, in contrast, and network advantages over younger firms; that is, the contributes positively to the overall effect of exports on TFP. effect of exports on productivity should be positively corre- We also estimate the effect of exports on LP. As hypoth- lated with export duration. Likewise, a higher share of esized, capital intensity and TFP have a significantly posi- skilled labor might translate into higher productivity gains tive effect on LP in the baseline specification. Machinery from exports due to faster learning. However, this effect and equipment intensity has a larger positive effect on LP depends on the skill intensity of labor used in manufactur- than building intensity when capital intensity is subdivided ing operations in Malaysia. Finally, we expect that a higher into its subcomponents. Exports also show a significantly share of private foreign ownership will increase the positive positive effect on LP (figure 12.3). Export duration has no productivity effect of exports, while a higher share of pri- significant impact on the productivity effects from exports. vate domestic ownership will reduce it, based on the Surprisingly, the interaction of exports with the share of assumption that foreign firms are more productive than high-skilled labor significantly reduces the positive pro- domestic firms, a finding supported by several firm-level ductivity effects from exports, which is not in line with our studies (see, for example, Temouri, Driffield, and Higón conjecture. We interpret this as a result of the fact that [2008] for a literature review focusing on industrial coun- Malaysia's export-oriented manufacturing sector is domi- tries or Fernandes [2008] for Bangladesh). nated by low-skilled assembly-type operations, which cre- When we interact exports with four plant characteristics-- ate low value added. Although Malaysia has been successful export duration, the share of skilled labor, the share of pri- in integrating its manufacturing sector into regional and vate domestic ownership, and the share of private foreign global production networks, it nevertheless remains highly ownership--we find as follows. The interaction of exports reliant on low- and semi-skill intensive assembly-type with export duration and the share of high-skilled labor manufacturing (World Bank 2010). A larger share of pri- has no significant effect on TFP. As hypothesized, a larger vate domestic ownership significantly reduces the positive share of private domestic ownership reduces the positive productivity effects from exports, while a larger share of productivity effects from exports, whereas a larger share of private foreign ownership significantly increases it, which private foreign ownership contributes positively to the is in line with our conjecture. overall effect of exports on productivity. In quantitative terms, our results suggest that a 1 percent What do our results mean in quantitative terms? Using increase in real exports raised LP by on average 0.09 to our baseline specification, we find that a 1 percent increase 0.21 percent between 1999 and 2006. When we control for in real exports raised TFP in Malaysian manufacturing firm characteristics, we find that this positive effect of firms by on average 0.31 to 0.44 percent between 1999 and exports is lowered or even reversed, the larger the share of 2006. When we control for firm characteristics, we find that high-skilled labor and private domestic ownership is, while a larger share of private domestic ownership can cause the a larger share of foreign domestic ownership significantly positive effect of exports to be lowered or even reversed. increases the gains from exports. The Effects of Exports on Productivity and Volatility: Evidence from Malaysian Firm-Level Data 149 The Effect of Exports on Output refers to imperfect correlations between domestic and for- Growth Volatility eign shocks, which can have a stabilizing effect on volatility Motivation (Buch, Döpke, and Strotmann 2009). Fourth, exports can lead to changes in specialization patterns. If exports lead to a The "great trade collapse," in particular the scale and more diversified production structure, volatility is expected velocity of the export decline during the recent economic to be reduced and vice versa (Giovanni and Levchenko crisis, has revealed an increased growth volatility in export- 2009). Because of these potentially conflicting effects, the net dependent countries and has led some to question the effect of exports on volatility is ambiguous. We explore the export-led strategy of development (Rodrik 2009). One of fourth channel in more detail when testing for the effect of the main arguments against trade openness is the vulnera- exports on the volatility of output growth. bility of outward-oriented economies to external shocks Firm-level volatility could be a concern for policy such as changes in the terms of trade, the interest rate, cap- makers for the following reasons (see Buch, Döpke, and ital inflows, and foreign demand. Strotmann 2009). First, changes in the volatility of output What are the determinants of the volatility of output growth can affect economic welfare, which matters at the growth? Ahmed and Suardi (2009) provide a number of firm level rather than at the aggregate level for households economic, noneconomic, and institutional factors that and consumers, as they must find ways to protect their affect a country's aggregate volatility, including country income and consumption against fluctuations of output. size and the level of economic development, trade open- While access to financial markets could enable economic ness and export structure, development of a country's actors to insure against such volatility, many households financial market, the role of institutions, natural disas- hold few financial assets and would rely on public insur- ters, and distortionary macroeconomic policies. Loayza ance mechanisms. Second, a higher exposure of house- et al. (2007) explain why developing countries face more holds to risk could lead to resistance to open markets if this aggregate volatility than developed countries and iden- risk is associated with globalization. Third, changes in tify three sources: developing countries experience big- firm-level volatility can influence growth, but the empirical ger exogenous shocks, suffer more domestic shocks, and literature has remained ambiguous about this link. have weaker "shock absorbers." In this second part of the chapter, we focus on the costs At the firm level, there has been little consensus on the of trade openness by assessing the impact of sectoral causes of the volatility of output growth. The following exports on firm-level volatility of output growth in the factors, among others, have been shown to influence Malaysian manufacturing sector for the period 1989­2007. volatility at the firm level: firm size, R&D intensity, owner- Again, the presentation is nontechnical and focuses on key ship status (publicly traded firms versus small private results and interpretation. For full details on data, estima- firms), inventory-sales ratio, access to external financing, tion results, and methodology, readers are referred to the deregulation of product markets, and greater international working paper version of this chapter. competition, exports, and export diversification (see, for This section thus adds to the small amount of empirical example, Comin and Philippon 2005; Davis and Kahn evidence on this subject, especially of studies using firm- 2008; Buch, Döpke, and Strotmann 2009; Buch, Doepke, level data. Second, we identify several variables that have a and Stahn 2009; the chapter by Haddad et al. in this vol- significant impact on volatility. Third, we interact sectoral ume). Our focus is on how exports and export structure in exports with sectoral product and market diversification of the form of product and market diversification affect the exports to test whether the impact of exports on the volatility of output growth at the firm level. volatility of output growth is different with diversification Exports have an influence on this volatility through at taken into account. And fourth, our data cover recent years, least four channels. The first channel is the exposure of while other studies focus on older time series. output to shocks in the form of volatility of aggregate shocks and the correlation of shocks across countries Key Findings (Buch, Döpke, and Strotmann 2009). The second channel is the reaction of output to shocks in the form of the elas- We define volatility as the standard deviation of output ticity of factor demand and supply (Buch, Döpke, and growth on a five-year moving average, an approach used in Strotmann 2009). Both channels are likely affected by a other studies measuring volatility (see, for example, Buch, firm's degree of openness (Rodrik 1997; Slaughter 2001). Döpke, and Strotmann 2009 or Davis and Kahn 2008). Exports are thus expected to increase the volatility of out- While the Productivity and Investment Climate Survey has put growth through these two channels. The third channel the advantage of reporting information on a plant's 150 Managing Openness exports and imports, its disadvantage lies in the short export directly to C, its output will be indirectly affected by period from 1999 to 2006, especially for measuring volatil- demand changes of C for assembled components from A. ity. We therefore use firm-level data from Compustat, which In a first step, we estimate the impact of exports on covers the period 1989­2007. the volatility of output growth, based on our baseline At the firm level, there has been little consensus on the specification. Labor shows a negative coefficient sign, while causes of the volatility of output growth. The following fac- capital shows a positive one, both being significant. What is tors among others have been identified as increasing the explanation for the differential impact of firm size on volatility at the firm level, at least for developed countries volatility? Labor has a stabilizing effect on volatility, (see, for example, Comin and Philippon [2005] and Davis because firms might not immediately adjust their required and Kahn [2008] for the United States and countries in number of employees to demand changes. This delay the Organisation for Economic Co-operation and Devel- might be because firms are not willing to lay off redundant opment, and Buch, Döpke, and Strotmann [2009] and workers immediately during an economic downturn, as Buch, Doepke, and Stahn [2009] for Germany): R&D search costs might be involved in identifying new workers intensity, ownership status (publicly traded firms versus once the business cycle picks up again. In addition, firms small private firms), access to external financing, deregu- might be restricted in adjusting their amount of labor lation of product markets and greater international com- because of labor market rigidities. The required amount of petition, and exports. Firm size, however, has a stabilizing capital, however, can be more easily adjusted to market effect. needs and thus might favor volatility. The impact of the inventory-sales ratio is ambiguous, Firm-level R&D intensity shows no significant impact depending on whether productivity or demand shocks on the volatility of output growth. The firm's share of dominate. While a higher inventory-to-sales ratio can inventory stock significantly raises volatility, suggesting smooth the impact of volatile production on sales (pro- that demand shocks dominate productivity shocks in ductivity shock), a lower inventory-to-sales ratio, reflecting Malaysian firms, while the opposite has been confirmed for improved inventory management, can help firms reduce Germany (Buch, Doepke, and Stahn 2009). Thus, a lower their volatility of sales (demand shock). Buch, Doepke, and ratio of inventory to sales, reflecting improved inventory Stahn (2009) find for Germany that the first effect domi- management, helps firms reduce their volatility of sales. nates. In addition, our chapter includes sectoral market and Moreover, we find that a higher share of long-term debt product diversification of exports, which have been identi- in total sales significantly lowers volatility in Malaysia. fied as reducing the volatility-augmenting effects from Why would a higher share of long-term debt in total sales exports but only at the national level (see Haddad et al. in at the firm level lower volatility in Malaysia but increase it this volume). in Germany (Buch, Doepke, and Stahn 2009)? One expla- In our regression analysis, we include the number of nation would be that a higher share of long-term debt is a employees and capital stock to capture firm size. R&D good sign in developing countries, as it reflects a firm's intensity is measured by calculating total R&D expenses creditworthiness (firms that are not creditworthy would over total sales. We obtained the ratio of inventory stock to simply not be given any credit). In developed countries, sales analogously. The share of long-term debts over total however, almost all firms are creditworthy, and a higher sales as a measure for access to external financing is calcu- share might thus reflect unsustainability. As hypothesized, lated following Comin and Philippon (2005). Since Com- higher sectoral import tariffs (more regulation) reduce pustat covers only publicly listed firms, we cannot control volatility, while larger sectoral exports significantly for the type of ownership. As in the previous section, we increase it (figure 12.4). use sectoral import tariffs as an inverse measure for trade In the second step of our analysis, we evaluate whether liberalization. this export-induced volatility can be reduced by higher Since the Compustat database does not include firm- product and market diversification of exports. We interact level trade data, we match these data with sectoral export sectoral exports with measures of sectoral diversification. data. This strategy follows the assumption that export We measure product diversification using a modified ver- activity at the sectoral level can influence firm-level output sion of the Hirschman-Herfindahl Index (HHI). The HHI through spillover effects. Assume, for instance, that a is a measure of concentration and is defined as the sum of domestic supplier B of automotive components purchases the squared world market shares of each exported product automotive parts from an upstream domestic supplier A by sector j. The HHI can range between 1/n and 1, where n before exporting the assembled components to a foreign designates the total number of products being exported. automobile manufacturer C. Although supplier A does not The lower bound represents total diversification, that is, all The Effects of Exports on Productivity and Volatility: Evidence from Malaysian Firm-Level Data 151 Figure 12.4. Average Annual Exports and Volatility Growth 0.252 percentage points in Malaysian manufacturing over by One-Digit SITC Sector, 1992­2005 the period 1989­2007. Controlling for export diversifica- tion, we find that market diversification significantly 7 reduced this effect, while such a stabilizing effect could not SITC4 SITC1 6 be confirmed with product diversification of exports. volatility growth by one-digit SITC (%) 5 SITC6 4 Conclusion SITC7 3 SITC5 This chapter evaluated the gains from exports in the form 2 SITC8 SITC0 of higher productivity and the risk of outward orientation 1 in the form of volatility of output growth using recent data 0 of Malaysian manufacturing firms. The hypothesis of the 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 export growth by one-digit SITC (%) export-led growth model is far from being universally Source: Authors' calculations. accepted, because of ambiguous results and econometric Note: Calculating volatility using five-year moving averages yields no weaknesses of earlier studies. We put this hypothesis to the volatility measures for 1989­91 and 2006­07. The x-axis shows the growth test by estimating the impact of exports on productivity rate of exports in logarithms by SITC. In the y-axis, sectoral volatility is calculated using averages over available firms by SITC. The base year is in the Malaysian manufacturing sector using plant-level 1997 for SITC4. SITC = Standard International Trade Classification. data for the period 1999­2006. Our estimation method- ologies also addressed the endogeneity problem between products have an equal market share, while the upper exports and productivity. Our results confirm the positive bound reflects total concentration, that is, one product has effect of exports on productivity. Thus, exports signifi- the market share of 1. Analogously, we also focus on the cantly increased TFP and LP in Malaysian manufacturing number of export markets instead of products by each sec- firms over the period 1999­2006. Sectoral trade liberal- tor to measure market diversification. Since our focus in ization also has a positive impact on productivity. this study is on diversification and not concentration, we Controlling for plant-level characteristics, we show that define DIV = 1/HHI. the positive impact of exports is lowered or even reversed Findings from this analysis are as follows. Market diver- the bigger its share of private domestic ownership. On the sification of exports significantly reduces the positive contrary, a bigger share of foreign domestic ownership impact of exports on the volatility of output growth. Such contributes positively to the effect from exports. Surpris- a stabilizing effect cannot be confirmed with product ingly, we find that a higher skill intensity of labor reduces diversification of exports, however. the effect of exports on labor productivity. We interpret What could explain the insignificant effect of product this outcome as a result of the dominance of the export- diversification? Empirical observation shows a nonlinear oriented manufacturing sector by low-skilled assembly- relationship between product diversification and income. type manufacturing in Malaysia, which creates low value Thus, middle-income countries appear to be much more added (World Bank 2010). diversified than low-income countries, whereas rich coun- In a second step, we focused on the costs of trade open- tries have a much less diversified economic structure ness by assessing the impact of sectoral exports on firm- (World Bank 2010). Figure 12.5 shows the sectoral devel- level volatility of output growth in the Malaysian manufac- opment of product and market diversification for the turing sector for the period 1989­2007. Our results show period 1989­2007 measured by the HHI. While this that sectoral exports raise firm-level volatility, which we measure shows a clear trend of increasing market diversi- relate to a higher dependency on world markets. We also fication (panel b), the same does not apply to product tested whether the export-induced volatility can be diversification (panel a). Many sectors show a trend reduced by sectoral product and market diversification of toward more product diversification until the mid-1990s, exports. Market diversification of exports significantly followed by growing specialization. This finding could be reduces the volatility-augmenting effect of exports. Such a an indication that Malaysia is in the process of growing stabilizing effect cannot be confirmed for sectoral product from a middle- into a high-income country, and that diversification of exports. Finally, we find that a firm's would explain the insignificant effects of product diversi- number of employees, its share of long-term debt in total fication in the regressions. sales, and sectoral import tariffs lower volatility, while a To sum up, sectoral exports significantly raised firm- firm's capital stock and its share of inventory stock in total level volatility of output growth by between 0.065 and sales increase it. Since our data source restricted us to 152 Managing Openness Figure 12.5. Product and Market Herfindahl-Hirschman Index by SITC1, 1989­2007 a. Herfindahl-Hirschman Index, product 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 b. Herfindahl-Hirschman Index, market 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 SITC0 SITC1 SITC4 SITC5 SITC6 SITC7 SITC8 Source: Authors' calculations. Note: A lower HHI reflects more diversification. SITC = Standard International Trade Classification. sectoral export and diversification data, future research TFP and LP in the Malaysian manufacturing firms between could estimate the effect of firm-level exports and diversifi- 1999 and 2006, they seem to have prevented a more serious cation on firm-level volatility. productivity slowdown in Malaysia. Our results imply that exports have counteracted the Finally, countries need to assess which type of export- productivity slowdown Malaysia has experienced during led growth model is appropriate given their stage of devel- the past decade. Average Malaysian LP grew at a rate of opment. The growth process from middle- to high-income only 2.9 percent following the Asian crisis in 1997, com- countries seems to be associated with greater specialization pared to 5.5 percent in the decade preceding the crisis. The of production and a greater focus on higher-value-added reasons for the productivity slowdown include a reduction products. Malaysia has been characterized by a greater in private investment since the Asian crisis, skills shortages, sophistication of export mix over the past four decades-- and lack of significant innovation. Average Malaysian TFP shifting from a reliance on the resource-based sector to a growth, however, remained stagnant over the two periods growing reliance on the electrical and electronics sectors-- (World Bank 2010). Since exports had a positive effect on now showing one of the largest shares of high-tech exports The Effects of Exports on Productivity and Volatility: Evidence from Malaysian Firm-Level Data 153 in total exports in the world. However, the share of domestic di Giovanni, J., and A. Levchenko. 2009. "Trade Openness and Volatility." Review of Economics and Statistics 91 (3): 558­85. value added in total output value stagnated or declined for Fernandes, A. 2008. "Firm Productivity in Bangladesh Manufacturing most manufacturing sectors between 1981 and 2002, sug- Industries." World Development 36 (10): 1725­44. gesting that Malaysia's exports still remain highly reliant on Furuoka, F. 2007. "Do Exports Act as `Engine' of Growth? Evidence from low- and semi-skill-intensive assembly-type manufacturing, Malaysia." Economics Bulletin 6 (37): 1­14. Khalafalla, K., and A. Webb. 2001. "Export-Led Growth and Structural which creates low value added (World Bank 2010). Specializ- Change: Evidence from Malaysia." Applied Economics 33 (13): 1703­15. ing into higher-value-added activities seems to be the right Loayza, N., R. Rancière, L. Servén, and J. Ventura. 2007. "Macroeconomic way forward for Malaysia. Although this approach might Volatility and Welfare in Developing Countries: An Introduction." World Bank Economic Review 21 (3): 343­57. increase the volatility of output growth due to a reduction in Mahadevan, R. 2007. "New Evidence on the Export-Led Growth Nexus: the product spectrum, a higher market diversification of A Case Study of Malaysia." World Economy 30 (7): 1069­83. exports could significantly counterbalance such effects. Medina-Smith, E. 2001. "Is the Export-Led Growth Hypothesis Valid for Developing Countries? A Case Study of Costa Rica." Policy Issues in International Trade and Commodities, Study Series 7, United Nations Bibliography Conference on Trade and Development, Geneva, Switzerland. Melitz, M. 2003. "The Impact of Trade on Intra-Industry Reallocations and Ahmed, A. and S. Suardi. 2009. "Macroeconomic Volatility, Trade and Finan- Aggregate Industry Productivity." Econometrica 71 (6): 1695­1725. cial Liberalization in Africa." World Development 37 (10): 1623­36. Milberg, W., and D. Winkler. 2010. "Trade Crisis and Recovery: Restruc- Buch, C., J. Doepke, and K. Stahn. 2009. "Great Moderation at the Firm turing of Global Value Chains." Policy Research Working Paper 5294, Level? Unconditional vs. Conditional Output Volatility." B. E. Journal World Bank, Washington, DC. of Economic Analysis and Policy 9 (1): Contributions, Article 20. Rodrik, D. 1997. Has Globalization Gone Too Far? Washington, DC: Insti- Buch, C., J. Döpke, and H. Strotmann. 2009. "Does Export Openness tute for International Economics. Increase Firm-Level Output Volatility?" World Economy 32 (4): 531­51. Rodrik, D. 2009. "Growth after the Crisis." Paper prepared for the Com- Comin, D., and T. Philippon. 2005. "The Rise in Firm-Level Volatility: mission on Growth and Development, Kennedy School, Harvard Causes and Consequences." Working Paper 11388, National Bureau University. of Economic Research, Cambridge, MA. Slaughter, M. 2001. "International Trade and Labor-Demand Elasticities." Damijan, J., M. Knell, B. Majcen, and M. Rojec. 2003. "The Role of FDI, Journal of International Economics 54 (1): 27­56. R&D Accumulation and Trade in Transferring Technology to Transi- Temouri, Y., N. Driffield, and D. Higón. 2008. "Analysis of Productivity tion Countries: Evidence from Firm Panel Data for Eight Transition Differences among Foreign and Domestic Firms: Evidence from Countries." Economic Systems 27 (2): 189­204. Germany." Review of World Economics 144 (1): 32-54. Davis, S., and J. Kahn. 2008. "Interpreting the Great Moderation: Changes World Bank. 1993. The East Asian Miracle: Economic Growth and Public in the Volatility of Economic Activity at the Macro and Micro Levels." Policy. Oxford: Oxford University Press. Working Paper 14048, National Bureau of Economic Research, Cam- ------. 2010. "Growth through Innovation." Malaysia Economic Monitor. bridge, MA. April. Part III II lessons in managing openness from country and regional experiences 13 The International Crisis and Latin America: Growth Effects and Development Strategies Vittorio Corbo and Klaus Schmidt-Hebbel The international economy is recovering from the worst mechanisms, and how have the region's structural condi- financial crisis since the 1930s. While the origin of the cri- tions affected its sensitivity to foreign shocks? sis was at the heart of the world's financial centers, it spread This chapter addresses the latter issues by assessing the to other regions and countries through different transmis- performance of growth in Latin America's seven major sion mechanisms. The financial crisis in major industrial economies during 1990­2009. Results from an economet- economies was halted only by massive financial support ric model are used to decompose growth into long-term and rescue programs, while the free fall of demand, output, and cyclical determinants to explain the amplitude of the and employment was reversed only by the combination of decline in gross domestic product (GDP) during the large-scale financial intervention and the most aggressive 1998­99 Asian crisis and the 2008­09 global crisis. This macroeconomic expansion recorded in history. All other approach allows us to quantify and identify several factors: economies where financial systems were not in crisis-- (a) the differences in unconditional and conditional effects industrial and developing alike--suffered from interna- of the global crisis on Latin America for both crises; (b) the tional contagion from the financial centers' crisis and the role of structural and policy variables that have improved industrial world's recession through conventional financial the region's resilience in the face of foreign shocks and and trade transmission channels. crises; and (c) the main implications of the dominant devel- This global financial crisis has raised concerns in devel- opment strategy adopted by the region since the 1990s. The oping economies about their macroeconomic policy presentation here is nontechnical and focuses on policy frameworks and their development strategies. Among the implications. For full details of the model and estimation questions raised by the crisis are, Which policies can best results, readers are referred to the working paper version of protect developing countries from world crises and shocks? this chapter (Corbo and Schmidt-Hebbel 2010). What role does domestic demand play in shielding them The next section of the chapter describes the growth from crises? And to what extent should they rely on a strat- performance of Latin America during 1990­2009 and justi- egy of close trade and financial integration into a world fies the focus on the two regional recessions: the 1998­99 economy punctuated by shocks and crises? recession associated with the Asian crisis and the 2008­09 Latin America has been strongly affected by the interna- recession caused by the global financial crisis. The follow- tional crisis and recession since late 2008. In comparison to ing section uses results from a growth regression model to previous crises, how has Latin America coped with the decompose the amplitude of both recessions, comparing the global crisis, what has been the role of different transmission very different roles of external and domestic growth factors 157 158 Managing Openness in the downturns. The next section draws implications factors (a deep and generalized crisis in Argentina and a from the previous results for the choice of policy regimes temporary collapse of oil production in the República Boli- and development strategies in support of the region's variana de Venezuela associated with a strike in the sector), growth and resilience in the face of foreign shocks and with almost no consequences for other countries in the crises. Final remarks close the chapter. region. In contrast to those two country-specific episodes, five of the seven countries suffered a recession during the 1998­99 regional contraction, and all seven countries suf- Latin America's Growth Performance fered a recession during the 2008­09 contraction. Hence, This study focuses on Latin America's seven largest we focus in this study on the two latter recessions only. economies--Argentina, Brazil, Chile, Colombia, Mexico, We now turn to dating the precise extent of the reces- Peru, and the República Bolivariana de Venezuela--that sion. One possibility is to stick to the two windows of account jointly for 91 percent of Latin America's 2008 consecutive negative growth, depicted in figure 13.1. GDP. The time sample spans the quarters from quarter one However, this aggregate regional growth behavior may of 1990 through quarter four of 2009. The main variable of mask significant country heterogeneity. Therefore, we interest is the countries' annualized quarterly growth rate exploit the full panel-data sample to test for recessions of seasonally adjusted real GDP. combining alternative recession windows for the 1998­99 Figures 13.1 and 13.2, panels a, b, and c, depict quarterly recession with different windows for the 2008­09 reces- GDP growth rates for the region and the seven individual sion, using panel-data estimations.2 We find that the best countries, respectively.1 Figure 13.1 reflects four periods of results are those for the four-quarter window spanning at least two consecutive quarters of negative average quarter three of 1998­quarter two of 1999 (Asian crisis) growth in the seven countries that represent Latin America and the two-quarter window from quarter four of 2008 to in our study: quarter three 1998­quarter two of 1992, quarter one of 2009 (global financial crisis). The latter quarter three of 2001­quarter one of 2002, quarter four of results are identical to the recession periods for aggregate 2002­quarter one of 2003, and quarter four of 2008­ Latin American GDP, depicted in figure 13.1. quarter one of 2009. The first episode is linked to the However, for the purpose of the final choice of contrac- 1997­98 Asian crisis and the last to the 2008­09 global tion periods relevant for our growth decomposition financial crisis and world recession. The second and third analysis performed below, we also consider the behavior of episodes reflect two very deep but idiosyncratic recessions output gaps around recessions (figure 13.3).3 The average in Argentina and the República Bolivariana de Venezuela, output gap in Latin America during the first recession more clearly visible in figure 13.2, panels a and c. The two period declines precisely during the four-quarter window episodes in Argentina and the República Bolivariana de that was selected above, that is, from quarter three of 1998 Venezuela were not caused by international but by domestic through quarter two of 1999. The output gap starts to close in quarter three of 1999; that is, actual GDP growth exceeds estimated trend growth since the latter quarter. Figure 13.1. Average GDP Growth in Latin America, After the second recession period, however, the output gap 1990­2010 continues to widen in quarters two and three of 2009, 25 reflecting a weak growth recovery in the aftermath of the global financial crisis. This factor leads us to extend the 20 contraction period relevant for our 1998­99 growth 15 decomposition by one quarter to obtain a three-quarter GDP growth (%) recession period. Accordingly, we have identified quarter 10 three of 1998--quarter two of 1999 (four quarters) and 5 quarter four of 2008­quarter two of 2009 (three quarters) as the recession periods in this study. 0 ­5 Explaining the Amplitude of the 1998­99 ­10 and 2008­09 Recessions 1990-Q1 1995-Q1 2000-Q1 2005-Q1 2010-Q1 year The literature on long-term growth is very broad on both Source: Authors' calculations. the theoretical and the empirical sides. While theoretical Note: Q = quarter. studies usually analyze the role of a key growth determinant The International Crisis and Latin America: Growth Effects and Development Strategies 159 Figure 13.2. GDP Growth in Selected Latin American Countries, 1990­2010 a. Argentina, Brazil, and Mexico 30 20 10 GDP growth (%) 0 ­10 ­20 ­30 1990-Q1 1995-Q1 2000-Q1 2005-Q1 2010-Q1 year Argentina Brazil Mexico b. Chile, Colombia, and Peru 25 20 15 GDP growth (%) 10 5 0 ­5 ­10 ­15 1990-Q1 1995-Q1 2000-Q1 2005-Q1 2010-Q1 year Chile Colombia Peru (continued) 160 Managing Openness Figure 13.2. (continued) c. Venezuela, R. B. 120 100 80 GDP growth (%) 60 40 20 0 ­20 ­40 1990-Q1 1995-Q1 2000-Q1 2005-Q1 2010-Q1 year Source: Authors' calculations. Note: Q = quarter. Figure 13.3. Average Output Gap in Latin America, We put our regression results to work by using them to 1990­2010 explain the amplitude of Latin America's decline in GDP 4 growth in the aftermath of both crises. To start, we compute the amplitude of the growth reduction in the seven sample 3 countries during both recessions--that is, the cumulative average output gap (% of GDP) 2 GDP reduction (expressed in annualized terms) observed 1 between the peak quarter before the recession (labeled in 0 figure 13.4 as quarter 0) and the trough quarter of our selected recession periods (labeled in figure 13.4 as quarter ­1 4 or quarter two of 1999 for the first recession and quarter 3 ­2 or quarter two of 2009 for the second recession). Table 13.1 ­3 reports the annualized recession amplitude for the seven individual countries and the region at large. The peak-to- ­4 1990-Q1 1995-Q1 2000-Q1 2005-Q1 2010-Q1 trough cumulative GDP change ranges from a GDP loss of year 8.5 percent in the República Bolivariana de Venezuela to a Source: Authors' calculations. GDP gain of 3.4 percent in Mexico during the four-quarter Note: Q = quarter. 1998­99 recession. In contrast to the latter, the full country range is in negative terrain during the three-quarter in isolation, the empirical literature takes a wider view, 2008­09 recession, with cumulative GDP losses that range considering several structural and policy growth factors. from 0.9 percent in Colombia to 11.1 percent in Mexico. Our approach is to estimate a growth model encompassing Simple (weighted) country averages of recession ampli- the largest possible set of structural, institutional, policy, tudes for the region stand at ­3.0 percent (­1.2 percent) for and cyclical determinants of short- and long-term growth, the first recession and ­4.2 percent (­5.2 percent) for the anchored in theory and international evidence. Our second recession. By any of the weighted averages, it is clear regression models, data sources, and estimation results are that the second recession was much deeper than the first presented in full detail in the working paper version of this one. Our next task is to explain a significant part of the chapter. observed simple-average recession amplitude, making use The International Crisis and Latin America: Growth Effects and Development Strategies 161 Figure 13.4. Average GDP Growth around the Asian Crisis, with an amplitude of ­4.2 percent, we explain some 95 per- 1998­99, and the Global Financial Crisis, 2008­09 cent, that is, an annualized output decline of 4.1 percent. What are the factors driving these results? 10 Asian crisis We start with foreign cyclical variables, which reflect 8 the transmission mechanisms from international crises 6 global financial and recessions to the region. A striking difference emerges GDP growth (%) 4 crisis between Latin America's first and second recessions. On 2 average (across countries and across the five foreign cycli- 0 cal variables), international conditions improved during ­2 the first recession, contributing 0.5 percent to higher ­4 cumulative growth.4 The opposite is observed during the ­6 recent recession, when international conditions deterio- ­8 ­6 ­5 ­4 ­3 ­2 ­1 0 1 2 3 4 5 6 rated on average massively for Latin America, contribut- quarter ing ­2.7 percent to (or more than half of) the recession's Asian crisis global financial crisis amplitude. In 1998­99, three out of five foreign variables improved for Latin America. However, in 2008­09 all five Source: Authors' calculations. cyclical variables deteriorated, and the largest single external driver of the recession was the massive decline in Table 13.1. Growth of GDP in Latin American Recessions, trading partners' economic growth. Hence, the 1998­99 1998­99 and 2008­09 recession was largely homemade, while the global finan- percent cial crisis and world recession were the main drivers of Asian crisis, Global financial crisis, the 2008­09 downturn. Countries Q3 1998­Q2 1999 Q4 2008­Q2 2009 We now turn to long-term growth variables. They dete- Argentina ­5.20 ­1.55 riorated significantly on average during the first recession, Brazil ­1.03 ­3.99 explaining a sizable ­1.7 percent, which is more than half Chile ­3.88 ­4.40 the 1998­99 recession's amplitude. In contrast, long-term Colombia ­6.82 ­0.87 Mexico 3.37 ­11.09 variables improved on average during the second recession, Peru 1.15 ­3.64 contributing 0.8 percent to higher cumulative growth in Venezuela, R. B. ­8.51 ­3.59 2008­09. Higher private credit flows (relative to GDP) and Simple average ­2.99 ­4.16 lower inflation contributed most to positive growth, while Weighted average ­1.15 ­5.24 the deterioration in fiscal balances (relative to GDP) weak- Source: Authors' calculations. ened growth. When we consider the reduced inflation coeffi- Note: Cumulative GDP growth rates within the reference period. Series cient observed since 2002, we find that the growth gain from de-seasonalized using ARIMA X­11. Q = quarter. lower inflation is much smaller in 2008­09. Therefore, when changes both in variables and in coefficients are combined, of our coefficient estimates and the changes in independent the contribution of long-term variables to the amplitude of variables (and in coefficient estimates, when applicable), the second recession is close to nil. according to our decomposition method, summarized in We come to similar conclusions on the very different the working paper version of this chapter. role of changes in structural variables during both reces- The results are reported in table 13.2, based on our most sions: they deepen the recession in 1998­99 (by ­0.6 per- comprehensive regression results. There we report the cent), while they dampen the recession in 2008­09 (by recession amplitude decomposition for the Asian crisis 0.6 percent). While our ex post measures of financial and (column 1) and for the global financial crisis (column 2). trade openness decline significantly during the most recent Column (2) is divided into three parts: the first is based on recession, the buildup of international reserves more than changes in explanatory variables only, the second is based offsets the latter. Once we consider the large changes in on changes in estimated parameters only, and the third is coefficients after 2000 (smaller for financial openness, the total contribution, the sum of the two previous parts. larger for trade openness, and smaller for international The amplitude of the first recession is ­3.0 percent reserves), however, the overall contribution of structural (reported in the bottom line of table 13.2), of which we variables to the amplitude of the 2008­09 recession-- explain some 90 percent, that is, an annualized output combining changes in their values and their estimated decline of 2.7 percent. Of the much deeper second recession, parameters--is very negative, equaling ­1.7 percent. 162 Managing Openness Table 13.2. Decomposition of Latin America's Recessions, 1998­99 and 2008­09 percent Asian crisis, Global financial crisis, Q3 1998­Q2 1999 Q4 2008­Q2 2009 Amplitude of GDP growth decline ­2.99 ­4.16 Structural changes No Changes Yes Long-term variables ­1.68 0.77 0.05 Private credit 0.24 0.44 0.44 Inflation 0.65 0.97 ­0.73 0.24 Secondary school enrollment ­0.14 0.15 0.15 Fiscal balance ­1.17 ­0.73 ­0.73 Political certainty ­1.26 ­0.06 0.01 ­0.05 Structural variables ­0.57 0.59 ­1.70 Financial openness 0.73 ­0.60 0.14 ­0.46 Trade openness ­0.53 ­1.32 ­0.79 ­2.11 Net external assets ­0.08 0.08 0.08 International reserves ­0.68 2.43 ­1.64 0.79 Exchange rate regime ­0.01 0.00 0.00 0.00 Foreign cyclical variables 0.54 ­2.60 ­2.74 Terms of trade growth 0.02 ­0.32 ­0.32 Growth of trading partners 0.26 ­1.36 ­1.36 Growth of world exports 0.53 ­0.05 ­0.05 Capital inflows to Latin America ­0.05 ­0.68 ­0.68 Sovereign spreads ­0.22 ­0.19 ­0.14 ­0.33 Domestic policy variables ­0.99 ­0.14 0.99 Government consumption 0.69 1.12 1.12 Real interest rate ­1.68 ­1.26 1.13 ­0.13 Interactions ­0.02 ­0.67 ­0.67 Growth of trading partners * trade openness 0.00 ­0.19 ­0.19 Growth of trading partners * financial openness 0.10 ­0.35 ­0.35 Capital inflows to Latin America * financial openness ­0.09 ­0.10 ­0.10 Sovereign spreads * net external assets ­0.02 ­0.03 ­0.03 Structural changes post-2000 ­2.02 Explained variation ­2.72 ­4.07 ­4.07 Unexplained variation ­0.26 ­0.09 ­0.09 Total variation ­2.99 ­4.16 ­4.16 Source: Authors' calculations. Note: Q = quarter. Domestic macroeconomic policy played on average a Finally, the growth effects of interactions between struc- contractionary role in 1998­99 and an expansionary role tural conditions and foreign shocks were neutral for the in 2008­09. Fiscal policy was expansionary in both reces- first recession but deepened significantly the second reces- sions but much more so in the second experience, when it sion, by 0.7 percent. This result is not surprising because made a positive contribution of 1.1 percent to cumulative the interaction terms largely reflect the amplifying effects growth. As opposed to the latter, monetary policy was of the deterioration in foreign conditions observed in highly contractionary in both recessions (due to higher 2008­09 but not in 1998­99. nominal interest rates in 1998­99 and negative inflation expectations in 2008­09), but much less so in the recent Implications for Policies and experience. Higher real interest rates deepened the 1998­99 Growth Strategies recession by 1.0 percent, while higher real rates (combined with the decline in the absolute coefficient of the real interest The evidence presented in this chapter on Latin America's rate) deepened the 2008­09 recession by just 0.1 percent. performance during its two last crises, 1998­99 and 2008­09, The International Crisis and Latin America: Growth Effects and Development Strategies 163 shows striking differences between the very different roles reduced to low and sustainable levels. Average public and played by foreign and domestic growth factors. The first publicly guaranteed debt fell from 30.1 percent of GDP in (less intense) recession was largely homemade, while the the early 1990s to 14.3 percent of GDP in the late 2000s second (more intense) recession was due largely to a deteri- (table 13.3). A final step toward further strengthening of orating world economy. The combined effect of foreign fiscal frameworks in the region--adopting formal fiscal cyclical factors was positive for Latin America's growth rules and fiscal councils--is still pending. Chile is the only during the first recession, while all foreign cyclical variables country that has in put in place a fiscal rule since 2001. deteriorated sharply during the world financial crisis, Our results provide strong evidence on the growth explaining more than half the last recession. In contrast to impact of the shift in the region's fiscal policy. First, the fis- foreign variables, all domestic variables explain more than cal balance makes a robust and economically large contri- 100 percent of the first recession and less than half the bution to growth. Second, government consumption has a 2008­09 downturn. significant stabilizing effect on short-term growth. Our The latter result is due to the large changes in develop- growth decomposition shows that the stabilizing role of ment strategies and policy regimes that Latin America government consumption was more heavily used during started in the 1990s and deepened in the 2000s. While pop- the 2008­09 contraction, when countries had more room ulist policies have reemerged in some countries, the for countercyclical fiscal policy. region's dominant development approach relies on market The second regime change in macroeconomic policies and private sector development, strong commitment to was the shift from inflexible toward flexible exchange-rate global integration, adoption of sustainable macroeco- regimes, largely implemented after the Asian crisis. Either nomic and financial regimes, and some progress on forced by markets or prompted by policy makers' convic- reforms to make governments more effective in their pro- tions, many countries replaced their crawling pegs or vision of public goods. Next, we derive the implications of exchange-rate bands with floats, which occasionally are of our empirical findings for evaluating the region's develop- the "clean" type (as in Mexico) and more frequently of the ment strategy in three key areas: macroeconomic regimes "dirty" type, that is, with high-frequency nonannounced and policies, domestic financial development, and interna- interventions (as in Brazil or Peru) or low-frequency pre- tional integration of goods and financial markets. announced intervention periods (as in Chile). Latin America Latin America began a major revamping of its macro- has reaped three benefits from flexible exchange rates: avoid- economic policy frameworks in the 1990s, a drive that was ance of recurring currency crises (that often lead to reces- consolidated in the 2000s. Fiscal policy had been unsustain- sions), use of nominal (and, hence, real) exchange-rate able in many countries since the 1970s and through the adjustment as a buffer against adverse foreign shocks (there- early 1990s, leading to fiscal crises and hyperinflation. Fiscal fore avoiding costly unemployment and output losses), and orthodoxy replaced profligacy in the 1990s, a trend that allowance of full conduct of an independent monetary intensified in the 2000s, when a significant part of commod- policy. ity windfalls was saved. In turn, fiscal policy was used as a Flexible exchange rates have not prevented countries countercyclical stabilizing tool during the 2008­09 recession. from engaging in trend accumulation of international Trends in fiscal deficits were dramatically curtailed or reserves to strengthen their foreign liquidity positions. turned into surpluses, and public debts were generally Drawing lessons from recurring past experience with Table 13.3. Public and Publicly Guaranteed External Debt in Latin America as a Percentage of GDP, 1990­2009 percent 1990­94 1995­99 2000­04 2005­09 Argentina 23.59 23.92 56.35 25.84 Brazil 20.31 12.35 16.91 7.26 Chile 23.42 7.16 9.15 6.27 Colombia 28.04 17.05 22.71 14.10 Mexico 22.03 24.06 14.80 10.93 Peru 45.23 35.13 36.18 21.43 Venezuela, R. B. 48.10 34.11 24.51 14.41 Simple average 30.10 21.97 25.80 14.32 Weighted average 23.56 18.42 22.51 11.62 Source: World Development Indicators (database), World Bank, http://data.worldbank.org. 164 Managing Openness inflexible exchange-rate regimes and currency crises, Latin (and sometimes without it), central banks have made sig- America has adopted an eclectic framework that combines nificant progress in adopting a framework of careful and exchange-rate flexibility with self-insurance in the form of responsible exercise of monetary policy. The success of holding significant levels of international reserves. Our monetary policy is reflected in lower inflation, declining in empirical evidence shows that both a flexible exchange-rate Latin America from an annual average of 34 percent in the regime and foreign exchange holdings contribute to early 1990s to 7 percent in the past five years (table 13.4). growth in Latin America. Most revealing is our finding that Our findings support the conclusion that lower inflation while reserve holdings had a very large effect and the also contributes significantly to higher growth. exchange-rate regime a nonsignificant effect on growth in The gains in the credibility of monetary policy reaped the 1990s, the relative importance of both variables was from low inflation gradually allow central banks to adopt reversed after the shift toward floats. Since 2000­2001, the countercyclical monetary policies. While central banks flexible exchange-rate regime has had a significant and were busy defending their inflexible exchange rates during large effect on growth, while the effect of reserve holdings the 1998­99 recession, they allowed their local currencies has declined in size albeit not in statistical significance. to depreciate in 2008­09 and exercised countercyclical Moreover, during the 1998­99 recession, central banks sold monetary policy. Our evidence shows that central banks reserves and therefore contributed to deepening the reces- raised nominal (and, hence, real) interest rates in 1998­99, sion, while in 2008­09 they did the opposite, contributing while they cut nominal interest rates in 2008­09. Although to higher growth. the latter cuts were not sufficient to compensate for a sig- The third component of macroeconomic policy is the nificant decline in inflation expectations, they helped avoid monetary regime. As noted above, a flexible exchange rate is excessively high real interest rates. Our evidence shows a necessary condition for exercising an independent mone- that growth was significantly curtailed by contractionary tary policy. Fiscal sustainability and responsibility preclude monetary policy in 1998­99, as opposed to the 2008­09 fiscal dominance over monetary policy, which is a second experience. condition of a macroeconomic regime for the exercise of an The macroeconomic regime shifts that Latin America independent and credible monetary policy. Finally, de jure has implemented in the past decade have contributed to (or, at least, de facto) central bank independence strength- holding aggregate demand growth in check during the past ens the conduct of a monetary policy that is independent of decade, leading to healthy current account balances and direct interference by government or private sector inter- significant reductions in public and private net external ests. Adoption of inflation targeting, today's monetary liabilities. Our findings confirm that the buildup of net regime of choice among many central banks in the world, external assets has had a significant positive effect on the requires the satisfaction of the three latter conditions. region's growth performance, either directly or interacting Therefore, it is no coincidence that several central banks with sovereign debt premiums. Moreover, when the global adopted inflation targeting in Latin America after obtaining financial crisis and world recession of 2008­09 hit, Latin legal or de facto independence, after severing their links America's fiscal and external position was healthy, and policy with government budgets, and during or after their transi- regimes were strong, enabling the region to face very well the tion toward floating exchange rates. With inflation targeting severe deterioration in international conditions (compared Table 13.4. Inflation in Latin America, 1990­2009 percent 1990­94 1995­99 2000­04 2005­09 Argentina 30.46 0.21 6.73 8.26 Brazil 85.91 8.56 7.79 4.54 Chile 13.66 5.26 2.68 3.69 Colombia 20.02 14.32 6.55 4.69 Mexico 12.32 19.01 5.40 4.04 Peru 47.09 7.08 2.19 2.54 Venezuela, R.B. 30.12 30.74 16.75 18.06 Simple average 34.23 12.17 6.87 6.55 Weighted average 51.68 11.16 7.11 5.45 Source: Authors' calculations. The International Crisis and Latin America: Growth Effects and Development Strategies 165 to 1998­99 or 1981­82) and to adopt effective countercycli- both domestically and internationally. The latter systems cal policies for the first time in its recorded history. contribute to financial deepening (and financial opening), The second area of significant progress in the region has improve domestic corporate governance, and raise aggre- been in the development of domestic financial and capital gate efficiency. Hence, structural pension reform can con- markets. During the past decade, Latin America's banking tribute significantly to economic growth, as shown for the sector has developed both in size and in diversity of finan- Chilean case (Corbo and Schmidt-Hebbel 2003). cial services, while improving its health and resilience in the The third key area of the region's development strategy face of domestic and external shocks. Domestic financial is globalization. Latin America in general has deepened its deepening (and financial integration) has been facilitated trade and financial integration with the world economy. by macroeconomic stability, deregulation of domestic During the past two decades, the region has largely dis- financial activities, privatization of banks, opening up to mantled its massive historical barriers to trade in goods, foreign ownership of banks, privatization of nonfinancial services, and capital flows. firms, and reduction of controls on foreign capital flows. Latin American countries have made much progress in Restrained from excessive risk taking by reformed financial reducing import tariffs, eliminating most nontariff barri- regulation and supervision (showing that the right lessons ers, and putting in place a large number of multilateral and were derived from previous financial crises), the region's bilateral preferential trade agreements with major world banks have avoided exposure to U.S. toxic assets and have trading partners. An open trade regime contributes to generally resisted well the recession of 2008­09. In fact, no higher long-term growth by reaping the well-known bene- financial crises were observed during 2008­09 in a region fits of improved resource allocation and helps to cushion that had suffered recurring banking crises in the past when the negative growth effects of adverse regional shocks hit by severe foreign shocks and domestic recessions. In our (such as the 2008­09 recession in industrial countries) findings, the ratio of private credit from commercial banks through a regionally more diversified trade pattern. The to GDP contributes significantly to the region's growth. region's large progress in trade integration is reflected by Moreover, the increase in that ratio had a mild stabilizing an increase in its average total trade ratio to GDP from effect during the 1998­99 recession and a larger expansion- 32 percent in the early 1990s to 49 percent in the late 2000s ary influence during the 2008­09 recession. (table 13.5). The countries that have progressed most in Beyond banking, the region adopted reforms in capital trade integration are Chile and Mexico--a result of their markets that boosted the development of private debt and low general trade barriers and the fact that a dominant equity markets, insurance markets, and pension funds. share of their foreign trade is conducted under preferential Financial and capital market development is a major and trade agreements. According to our findings, higher trade robust growth determinant acting through several channels openness has a very significant and large effect on the of transmission on saving and investment and, fundamen- region's growth performance. The drawback of this posi- tally, on productivity growth, as shown by a long literature tive impact on long-term growth is that during recessions, (for example, Levine 2005). Deep pension reforms in many when trade declines more than domestic output, shrinking Latin American countries have replaced state-run pay-as- trade ratios deepen domestic recessions; this pattern was you-go pension systems with defined-contribution systems observed moderately in 1998­99 and massively in 2008­09, managed by private companies that invest pension funds according to our results. Table 13.5. Trade Openness in Latin America, 1990­2009 Exports and imports as a percent of GDP 1990­94 1995­99 2000­04 2005­09 Argentina 17.20 22.12 22.60 25.98 Brazil 15.45 20.44 22.36 27.40 Chile 49.72 60.85 68.41 83.56 Colombia 29.96 37.50 36.76 44.27 Mexico 27.26 40.47 53.32 60.89 Peru 26.00 32.74 35.43 40.56 Venezuela, R.B. 61.37 56.22 52.46 61.29 Simple average 32.42 38.62 41.62 49.14 Weighted average 22.64 29.53 33.74 39.77 Source: Authors' calculations. 166 Managing Openness Regarding financial integration, Latin America has equate public policies are very high is in education, which complemented domestic financial liberalization with is of very low quality. Although much progress has been external financial opening: reducing restrictions on hold- made in school enrollment and educational attainment, ings, on inflows and outflows of short- and long-term for- Latin American countries still rank very low in interna- eign direct investment, on loans, and on portfolio and tional education achievement tests, even when controlling equity flows. Restrictions on short-term capital inflows-- for per capita income. Public education suffers from low prevalent in some countries during the 1990s--have been budgets, poor incentives, lack of accountability, and barri- abolished or not restarted in most countries. International ers to educational reforms aimed at improving teaching financial integration leads to larger gross external asset methods and raising teachers' productivity. Finally, regional and liability holdings, which contribute to more efficient growth is hampered by widespread government corruption resource allocation and better insurance against idiosyn- and inefficient public administration. Government bureau- cratic national shocks and hence to higher growth and crats are selected largely on the basis of party affiliation lower volatility in income and output. The region's instead of professional merit, resulting not only in the low progress in financial integration is reflected by a rise in quality of government bureaucracies but also in their short the average total external asset and liability ratio to GDP tenure, linked to government mandates. Notable excep- from 88 percent in the early 1990s to 114 percent in the tions are Brazil and Chile, which have introduced, at least late 2000s (table 13.6). We have also found that higher partly, meritocratic hiring of government managers and financial openness has a very significant and large effect staff. Hence, government reform at all levels--from munici- on the region's growth. However, while during the palities to public enterprises and central governments--is 1998­99 recession the GDP ratio of external asset and also a major development challenge in the region's quest to liability holdings increased, hence lessening the recession, attain higher growth and more equity. the opposite occurred during 2008­09, when the signifi- cant decline of the latter ratio (reflecting in part the Final Remarks decline in capital inflows to the region) contributed to deepening the recession. We conclude that Latin America changed significantly Despite considerable progress in applying a coherent and between the late 1990s and the 2000s. In this chapter's sustainable development strategy, Latin America still faces a empirical results, we show that putting in place a better and large pending agenda of raising growth further and making stronger development strategy since the late 1990s has faster progress in reducing poverty and improving income raised the region's growth rate. While there is still signifi- distribution. On growth, the region's main shortcoming is cant intraregional heterogeneity in economic regimes and low productivity and slow productivity growth. The region policies, the predominant development strategy is based on has much room for improving the efficiency and competi- the adoption of prudent and rule-based macroeconomic tiveness of domestic markets and for facilitating the policies, deeper and healthier financial systems and capital process of creative destruction of firms. Labor markets are markets, and strong integration into world goods and cap- excessively regulated in the formal sector, leading to high ital markets. Our results show that improvements in many structural unemployment and informal employment. specific variables associated with these three areas have led Another area where the equity and efficiency costs of inad- to higher average growth. Table 13.6. Financial Openness in Latin America, 1990­2009 Total external assets and liabilities as a percent of GDP 1990­94 1995­99 2000­04 2005­09 Argentina 78.47 103.80 176.51 147.57 Brazil 45.84 53.18 86.77 82.94 Chile 119.02 126.87 192.10 184.57 Colombia 51.70 61.62 87.07 78.97 Mexico 62.99 81.79 70.28 79.52 Peru 97.99 100.91 103.79 102.45 Venezuela, R.B. 156.85 131.10 145.50 122.00 Simple average 87.55 94.18 123.14 114.00 Weighted average 63.19 74.23 100.77 95.70 Source: Authors' calculations. The International Crisis and Latin America: Growth Effects and Development Strategies 167 Moreover, adopting that development strategy has greatly region's business environment, labor market regulations, improved Latin America's resilience in the face of adverse quality of education, and government efficiency has to be foreign shocks. According to our results, the last recessions tackled to raise Latin America's efficiency and equity. Lack suffered by the region were very different--in magnitude, in of progress in those areas could result in frustration with the role of foreign shocks, and in the contribution of domes- macroeconomic responsibility and structural achieve- tic conditions and policies. The 1998­99 recession--of a ments, creating conditions for further spreading of pop- smaller magnitude--was largely homemade, related to the ulist policies that have inflicted so much damage on the weak macroeconomic and structural policy framework that region in the past 50 years. To make significant progress Latin America had in place in the 1990s. In contrast, the sec- requires significantly improving the quality and independ- ond recession--much deeper and affecting all major Latin ence of the public sector, learning from the successful expe- American economies--was largely due to deteriorating con- rience of countries like Australia, Canada, Finland, New ditions in the world economy. The improved resilience of Zealand, and Sweden. Latin America in the face of foreign shocks and world reces- sions is reflected in our results in four ways. First is the suc- cess in adopting macroeconomic policy regimes that better Notes protect domestic economies against external shocks (like 1. Seasonally adjusted GDP data are from official national sources. The full database used in this chapter is available upon request. exchange-rate floats, lower foreign net liabilities, and higher 2. Results are not reported here but are available on request. gross international reserves) and strengthen adoption of 3. Output gap series are built for each country using 2010­14 GDP countercyclical policies (like inflation targeting, contributing projections from Consensus Forecast. Then we use the 1990­2014 quar- to lower inflation, and improved fiscal policy frameworks, terly country time series for past and projected future GDP levels to esti- mate trend GDP series based on the Baxter-King filtering method. The reflected in lower public debts and deficits). Second is the output gap is defined as the percentage deviation of actual (or projected success in building up deeper and healthier financial systems future) GDP from trend GDP. and capital markets. Third is the attainment of larger trade 4. For simplicity, we use the term percent change instead of the more precise percentage-point change throughout this section. and financial integration. Finally are the indirect benefits of these improvements in reducing the sensitivity of growth to adverse conditions, reflected for example by the post-2000 Bibliography reduction in the sensitivity of growth (that is, in growth coefficients) to inflation and political uncertainty and the Corbo, V., and K. Schmidt-Hebbel. 2003. "Macroeconomic Effects of Pension Reform in Chile." In Pension Reforms: Results and Challenges. increase in the sensitivity of growth to trade openness and Central Bank of Chile, Santiago. exchange-rate floats. ------. 2010. "Is the Global Financial Crisis Different for Latin America?" Although much has changed in Latin America in the Catholic University of Chile and Centro de Estudios Publicos, Santiago. past two decades, many impediments to achieving higher Levine, R. 2005. "Finance and Growth: Theory, Evidence and Mecha- and sustained growth and better opportunities for the poor nisms." In Handbook of Economic Growth, ed. P. Aghion and S. Durlauf. still exist. An ambitious reform agenda for improving the Amsterdam: North-Holland. 14 The Economic Crisis of 2008­09 and Development Strategy: The Mexican Case Jaime Ros The international economic and financial crisis of could have done more to moderate the recession. The fiscal 2008­09 has dramatically affected the Mexican economy. stimulus and the management of the interest rate were too The contraction in economic activity and employment cautious, given the size of the shock. Then the chapter was considerable in late 2008 and early 2009. The fall in brings together the conclusions of previous sections to economic activity in 2009 was worse than the decline in explore why the crisis was so severe in the historical and gross domestic product (GDP) in 1983 in the wake of the international contexts, highlighting the role of the external 1982 debt crisis and was even greater than the drop in shock and the domestic policy response. The next section 1995 during the so-called Tequila crisis, which was the looks ahead: the sharp output contraction was followed by worst recession since the Great Depression of the 1930s. a rather vigorous recovery in the third and fourth quarters Mexico's contraction in output was by far the most severe of 2009. The section continues with a discussion of the in Latin America. In addition, the economic crisis of changes in macroeconomic policy and development strat- 2008­09 also caused a sharp increase in unemployment, egy that are necessary to sustain a faster pace of economic underemployment, and poverty rates. growth than that in the recent past. The final section offers This chapter examines the economic crisis of the past some conclusions. two years in Mexico. The first section looks at the evolu- tion of the main sectors of economic activity and the Dimensions of the Crisis behavior of the components of aggregate demand, show- ing that the recession was largely an industrial recession Mexico's economic expansion from 2002 onward came to led by the contraction in exports of manufactures. Com- an end in the first quarter of 2008, when the economy paring the economic performance of Mexico with that of entered a recession that lasted over the following two quar- the rest of Latin America puts in perspective the size of the ters (table 14.1). Then, led by a precipitous decline in man- external demand shock to Mexico's exports. The next sec- ufacturing production, overall economic activity contracted tion looks at the channels of transmission of the interna- sharply in the fourth quarter of 2008 and the first quarter of tional crisis. The trade channel is certainly the most 2009 (at annualized rates of 7.5 percent and 24.9 percent, important but not the only one; the section also looks at respectively). Construction was also severely affected by the the capital account shock caused by the "flight to quality" contraction of activity, while commerce and services largely and the impact of the crisis on migration flows and remit- mirrored the fall of total GDP. The turnaround began in the tances. The following section examines the fiscal, mone- second quarter of 2009 and gained momentum in the third tary, and exchange-rate policy response. It shows that and fourth quarters, when total GDP grew at 2.5 percent while not quite procyclical (largely because the deprecia- and 2.0 percent over the previous quarter (at annualized tion of the real exchange rate helped moderate the adverse rates of 10.4 percent and 8.2 percent, respectively). The effects of the external demand shock), the policy response recession thus appears to have been relatively brief but very 169 170 Managing Openness Table 14.1. GDP Growth in Mexico by Sector and Employment Indicators, 2008 and 2009 percent 2008 2009 Sector and employment indicator Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Sectors Agriculturea ­0.5 2.8 0.1 0.4 ­1.7 4.1 ­1.3 0.9 Commerce and services 0.7 0.4 1.2 ­2.5 ­7.0 ­1.5 4.9 1.0 Construction 2.9 ­4.0 ­0.1 ­1.9 ­4.0 ­0.1 ­1.2 ­1.8 Electricityb ­2.5 ­1.9 ­2.5 0.6 1.3 0.5 1.5 ­0.1 Manufacturing 1.0 ­0.4 ­1.3 ­2.4 ­11.9 1.0 4.4 5.9 Mining 0.8 ­2.0 0.6 0.9 ­0.5 0.2 1.8 ­0.1 Whole economy 1.2 ­0.4 ­0.1 ­1.9 ­6.9 0.3 2.5 2.0 Employment indicatorsc Unemploymentd 3.80 3.72 3.86 4.56 4.87 5.49 5.75 5.72 Underemploymente 6.75 6.55 6.86 7.25 8.38 10.16 8.99 8.93 Informal employmente 27.38 27.40 27.27 27.02 28.13 28.02 28.31 28.36 Source: Based on Instituto Nacional de Estadística y Geografía (INEGI), Banco de Información Económica, Indicadores de coyuntura (database), http://www.inegi.org.mx. Note: GDP growth is seasonally adjusted. The percentages indicate changes over previous quarter. GDP levels are at 2003 prices. Q = quarter. a. Includes livestock, forestry, and fishing, b. Includes gas and water. c. Quarterly average rates (seasonally adjusted). d. Percentage of economically active population. e. Percentages of employed population. Underemployment refers to the employed population that would work more hours than the current job allows. deep, driven by the sharp contraction of industry and man- and in public consumption in the second quarter of 2009. ufacturing in particular. Taken together, the picture that emerges from table 14.1 The recession led to a significant deterioration in and figure 14.1 on GDP by sector and the components of employment indicators. The unemployment rate rose by aggregate demand is one of a predominantly industrial about two percentage points between the second quarter recession driven by a collapse in exports. of 2008 and the third quarter of 2009 (peaking at a rate of Mexico's contraction in GDP has been by far the worst 5.75 percent) and started falling only in the fourth quarter in Latin America (see table 14.2). The decline of GDP in of 2009.1 The underemployment rate recorded a similar Latin America and the Caribbean has been on the order of evolution, rising almost four percentage points and peak- 1.8 percent, compared to a contraction of 6.7 percent in ing at 10.16 percent in the second quarter of 2009. The Mexico.2 Mexico's output contraction was much greater share of unemployed and underemployed in the labor than the contraction in world output (­0.8 percent) and force thus reached a level of 15.65 percent in the second greater than those in most regions in the world, comparable quarter of 2009. The share of the informal sector in the only, in fact, to that of the Commonwealth of Independent total labor force also rose and continued to climb in the States (­7.5 percent) (see IMF 2010c). This severity is fourth quarter of 2009. explained partly by the fact that, as noted, the recession in Figure 14.1 shows the evolution of the main compo- Mexico--driven by the collapse of exports and the shock to nents of GDP. Clearly, exports were the component of the volume of exports--was one of the worst in Latin aggregate demand that fell most sharply during the America: a 14.3 percent decline compared to a fall of 9.6 recession, collapsing at annualized rates of 50 percent percent for the region as a whole. and 21.9 percent during the fourth quarter of 2008 and the However, the size of the external shock may not explain first quarter of 2009, respectively. From their peak in the why the recession in Mexico was the worst in Latin America. second quarter of 2008 to the trough in the second quarter The difference in the size of the negative shock to export of 2009, exports fell 24.4 percent. Private investment was volumes in Latin America (an average of ­9.6 percent) and also severely affected and has yet to start a recovery, Mexico (­14.3 percent) does not appear to justify the differ- unlike exports, which began growing vigorously in the ence in the size of the decline of GDP (1.8 percent versus third quarter of 2009. Overall, public consumption and 6.7 percent), given that the degree of openness to trade is investment rose throughout the recession, despite small not much higher in Mexico than in Latin America as a declines in public investment in the first quarter of 2009 whole (48.2 percent versus 58.9 percent). In addition, a few The Economic Crisis of 2008­09 and Development Strategy: The Mexican Case 171 Figure 14.1. Growth Rates of Components of Mexico's This puzzle will be addressed later, after we review the chan- GDP, 2008­09 nels of transmission of the international crisis. 15 The Channels of Transmission 10 This section discusses the ways in which the crisis was transmitted to the Mexican economy. Consideration is 5 given to both real and financial sector mechanisms. % growth 0 The Trade Channel and the Collapse of Exports Among the various mechanisms through which the inter- ­5 national economic crisis has affected Mexico, the trade channel stands out, given that the United States, the epi- ­10 center of the world crisis, represents around 80 percent of Mexico's exports at a time when exports account for ­15 around 30 percent of Mexico's GDP (in 2008 at 2003 prices, INEGI database). The contraction in the value of ­20 total exports, coming largely from the fall in oil prices and manufacturing export volumes, proceeded at quarterly 1 2 3 4 1 2 3 4 -Q -Q -Q -Q -Q -Q -Q -Q rates of nearly 20 percent in the fourth quarter of 2008 08 08 08 08 09 09 09 09 20 20 20 20 20 20 20 20 and nearly 14 percent in the first quarter of 2009. Indeed, quarter the international recession brought about a sharp decline private consumption in the price of oil, one of Mexico's main export products, public consumption shown in a steep decline in the value of oil exports at private investment quarterly rates of 50.3 percent in the fourth quarter of 2008 and 25.7 percent in the first quarter of 2009 (figure public investment 14.2).3 This fall affected not only the economy's terms of exports of goods and services trade, but also government revenues and the fiscal deficit, imports of goods and services given that Petróleos Mexicanos, (PEMEX), the state oil monopoly, is a major source of fiscal revenues (approxi- Source: Based on INEGI, Banco de Información Económica, Indicadores de coyuntura (database), http://www.inegi.org.mx. mately 37 percent of the total in 2008). Note: Data are seasonally adjusted. Percentage changes are over previous Most important, the U.S. recession sharply affected quarter. GDP levels are at 2003 prices. Q = quarter. export volumes, particularly manufacturing exports. The value of manufacturing exports (around 95 percent of non- countries that experienced greater shocks to their export oil exports and around 78 percent of total exports) fell at volumes than Mexico had more open economies but did quarterly rates of over 13 percent in the fourth quarter of not record the kind of output contraction that character- 2008 and the first quarter of 2009. Exports of the automo- ized the Mexican recession. The Dominican Republic (with bile industry, which make up around 22 percent of total an 18.1 percent decline in exports and an increase of exports (and 26 percent of non-oil exports), were particu- 2.5 percent in GDP), for example, and Paraguay (with larly affected, with a contraction of nearly 38 percent in declines of 19.1 percent and 3.5 percent in exports and the first quarter of 2009 (compared to the previous quar- GDP, respectively) are two such cases. Nor did Bolivia and ter). The UN Economic Commission for Latin America and the República Bolivariana de Venezuela, with sharp contrac- the Caribbean (CEPAL) has estimated that as many as tions in export volumes, experience the size of the Mexican 1.1 percentage points of the contraction of GDP in 2009 recession. Moreover, the shock to the value of Mexican may be attributed solely to the direct effects (before consid- exports is of the same order of magnitude as that for Latin ering multiplier effects) of the contraction in the automo- America and the Caribbean as a whole and similar, or even bile industry (around 35 percent), given the weight of this smaller, than the shocks in several countries that recorded industry in total GDP (3.3 percent) (CEPAL 2009). Another sharp declines in export prices (Brazil, Chile, Colombia, important export-oriented sector that was severely affected Ecuador, Peru, and the República Bolivariana de Venezuela). was electrical and electronic products and equipment. 172 Managing Openness Table 14.2. Mexico's Balance of Payments, 2008 and 2009 US$, billions 2008 2009 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Current account deficit 2.9 2.0 4.3 6.7 1.2 ­0.4 3.7 0.7 Capital account surplus 9.1 1.4 2.1 12.0 ­2.7 ­3.0 6.5 13.8 Foreign direct investment 6.4 7.2 3.5 6.1 4.0 5.1 1.6 0.7 Portfolio foreign investment 4.2 5.8 0.6 ­5.8 ­1.1 1.0 7.8 7.6 Assets in foreign banks ­4.2 ­7.5 0.9 8.1 1.7 ­7.1 ­2.2 2.9 Other 2.7 ­4.1 ­2.9 3.6 ­7.3 ­2.0 ­0.7 2.6 Errors and omissions ­0.2 2.3 ­0.1 ­3.2 ­2.7 ­2.0 ­0.8 1.6 Change in reserves 6.1 1.6 ­2.4 2.1 ­6.6 ­4.7 1.9 14.7 Source: INEGI, Banco de Información Económica, Balanza de pagos (database), INEGI, http://www.inegi.org.mx. Note: Q = quarter. Figure 14.2. Value of Mexico's Exports of Goods, declined starting in the third quarter of 2008 and only 2008­09 began to recover significantly in the third quarter of 2009. The fall in portfolio flows was especially deep in the fourth 90 quarter of 2008, when the capital outflow reached almost 80 US$6 billion. By contrast, foreign direct investment was much less volatile during the period of recession, and, 70 adjusting with a relatively long lag to the output contrac- 60 tion, it declined significantly during the second half of 2009. US$, billions The flight to quality led to an increase in risk spreads 50 that was wider than for many other emerging markets and 40 to a significant depreciation of the Mexican peso. The risk 30 premium reached a peak of nearly 600 points in October 2008 but fell afterward as a result of lesser risk aversion as 20 well as a flexible credit line by the International Monetary 10 Fund (IMF) (US$47 billion, March 2009), a swap line with the U.S. Federal Reserve (US$30 billion), and resources 0 from multilateral development banks. Today, the risk 2 3 4 1 2 3 4 -Q -Q -Q -Q -Q -Q -Q spread has returned to precrisis levels. 08 08 08 09 09 09 09 20 20 20 20 20 20 20 In turn, the depreciation of the peso led to large unex- quarter pected losses on corporate derivative exposures, which caused financing difficulties for some large Mexican corpo- other non oil rations from November 2008 to March 2009. These financ- manufactures (including automobiles) ing difficulties are likely to have contributed to the sharp oil decline in private investment, especially in the fourth quar- Source: Based on INEGI, Banco de Información Económica, Indicadores ter of 2008. Given the widespread ownership of the domes- de coyuntura (database), INEGI, http://www.inegi.org.mx. tic banking system by global banks,4 the negative direction Note: Figures are seasonally adjusted. Current prices are provided. Q = quarter. of private investment may also have been affected by con- straints on credit supply in the case of some global banks, derived from pressures on parent balance sheets. The Capital Account Shock and the Fall in Private Investment Migration and Remittances In addition to the initial deterioration of the current account balance, the balance of payments was affected by The crisis and recession in the United States have brought the "flight to quality," the increased demand for safe for- with them a sharp reduction in the net flow of migration eign assets typical in situations of high uncertainty. As out of Mexico, as the demand for migrant workers has shown in table 14.2, portfolio foreign investment sharply fallen. Family remittances have also declined, to the extent The Economic Crisis of 2008­09 and Development Strategy: The Mexican Case 173 that unemployment among migrants has increased with and commerce). CEPAL estimates the negative impact of the U.S. recession. Indeed, the net migratory flow has the epidemic at 0.7 percent of GDP (CEPAL 2009) and the sharply diminished. Mexico's National Survey of Employ- IMF at 0.5 percent (IMF 2010a). ment and Occupation, which began in 2006, estimates that the net flow of migrants fell from 547,000 in the period The Policy Response February 2006 to February 2007 to 203,000 in the period February 2008 to February 2009, a 63 percent decline. This As discussed in the previous section, the crisis had both drop has been the result of a very sharp contraction in financial and real sector effects. This section examines the migration from Mexico in the face of a relatively stable government response to them, focusing on fiscal policy, migration to Mexico. It is worth noting that in the period monetary policy, and the exchange rate. November 2008 to February 2009 net migration even turned slightly negative (­1,000). Fiscal Policy Since the reduction in migration flows took place dur- ing a severe recession in Mexico, the most likely factors The impact of the global recession on Mexico's public behind the contraction come from the demand side finances was felt through a fall in oil revenues (2.1 percent- (together with tighter border controls) rather than the age points of GDP), aggravated by a decline in domestic oil supply side. The role of tighter border controls must also production and tax revenues (0.7 percentage points of have been important, because the contraction in migra- GDP). According to the Ministry of Finance, the fall in tion began before the recession. government revenues was the worst in at least 30 years More than 10 million Mexican immigrants are esti- (SHCP 2010). Despite the balanced-budget rule, which mated to reside in the United States. They sent about severely constrained the use of a countercyclical fiscal pol- US$26 billion back to Mexico in family remittances in icy in the face of this reduction in government revenues, 2007, the highest in the world after India. Although the government managed to take a number of fiscal meas- compared to other countries of outmigration (Central ures to respond to the impact of the global crisis. The American countries, Portugal, and Turkey, for example), program to boost growth and employment, initiated in remittances as a percentage of GDP are relatively small at October 2008, included a fiscal stimulus package aimed at less than 3 percent of GDP, at their peak in 2007 they nev- increasing transportation and infrastructure investments.5 ertheless amounted to around two-thirds of fuel exports The removal of long-term PEMEX investment schemes and more than 100 percent of the net inflows of foreign from the balanced-budget rule freed up some Mex$78.3 direct investment (see World Development Indicators, billion in the 2009 budget (or about US$5.5 billion), two- http://data.worldbank.org). The contraction of migration thirds of which was allocated to additional infrastructure flows and the recession in the United States have been investment6 (see OECD 2009c; SHCP 2010). Overall, the accompanied by a decline in family remittances. They fiscal stimulus for 2009 has been estimated at 1.5 percent of started falling in 2008 and sharply declined in 2009: a GDP by the IMF (see SHCP 2010) and at 1.4 percent of reduction of nearly US$4 billion, a 15.7 percent decline or GDP by the Organisation for Economic Co-operation and around 0.4 percentage points of GDP. Development (OECD) (OECD 2009c).7 All this means that the escape valve that migration has The maintenance of a small budget deficit--or even a represented in the Mexican labor market--and its posi- balanced budget when excluding PEMEX investment--in tive effects (along with those of remittances) on poverty the face of a moderate increase in public spending as a alleviation--no longer has the important role that it has percentage of GDP and a sharp fall in oil prices and tax had in the recent past. As expected--given these condi- revenues was made possible by a number of factors. First, tions and the employment and income losses generated the hedging of oil prices for 2009 at US$70 per barrel by the recession--unemployment, underemployment, (twice the level at the end of 2008) generated savings of and poverty have risen substantially in Mexico. about Mex$118.4 billion (about US$8.5 billion). Second, Other factors, of a domestic nature, further compli- as already noted, in 2008 the balanced-budget rule was cated the economic difficulties Mexico faced in 2009. One amended to exclude the capital investments of PEMEX, example was the case of the outbreak of influenza an action that reduced budget expenditures and created a A (H1N1) in April and May of 2009, which paralyzed 13 percent increase within the budget for investment. Mexico City (given the measures taken to address the epi- Third, the depreciation of the peso increased the domes- demic) and had a significant adverse effect on tourism and tic currency value of revenues from oil exports. Finally, the economic activities linked to it (hotels, restaurants, the government made use of nonrecurrent revenues in 174 Managing Openness the oil stabilization fund and the operating surplus of the April 2009 by the IMF of a flexible credit line arrangement, central bank for fiscal 2008 (SHCP 2010). with no conditionalities, of approximately US$47 billion. The real exchange rate followed closely, but with lesser amplitude, the evolution of the nominal exchange rate. Monetary Policy Overall, the real depreciation of the peso was on the order In industrial countries, a loosening of monetary policy of 20 percent between August 2008 and December 2009 generally began in the third quarter of 2008, and the same and on the order of 11 percent between January 2008 and happened during the fourth quarter in some emerging- December 2009. market economies. By contrast, in Mexico monetary policy was tightened at the beginning of the crisis in response to Why Was the Crisis So Severe? the capital account shock and the resulting exchange-rate pressures. The target interest rate of the Banco de México To answer the question of why the crisis was so severe, one was increased by 25 basis points twice in June 2008 and has to start by looking at the characteristics of Mexico's again in July and August, remaining at 8.25 percent foreign trade. Since the early 1980s and later with the through January 2009. Then, as price pressures receded advent of the North American Free Trade Agreement and the recession deepened, the policy rate was reduced in (NAFTA), the processes of trade liberalization and inter- a series of steps starting in January 2009 (50 basis points) national trade integration have gone hand in hand with an and continuing in February (25 basis points), March increasing importance of foreign trade in GDP, a rising (75 basis points), April (75 basis points), May (75 basis share of manufactures in total exports, and a high concen- points), June (50 basis points), and July (25 basis points), tration of exports in the U.S. market. The ratio of foreign remaining thereafter constant at 4.50 percent. It is also trade to GDP rose from 27 percent in 1982­84 to 62.6 per- worth noting the role of development banks in relaxing the cent in 2004­06 (World Development Indicators, http:// credit constraints that resulted from the financial crisis by data.worldbank.org) as a result of a boom in imports as providing financing to the most affected sectors (Banco de well as exports. This ratio is well above that of the other Mexico 2010). two largest Latin American economies (and in fact more than twice Brazil's ratio). The increase in exports was fundamentally a result of the boom in manufacturing Exchange-Rate Policy exports. As a consequence, the share of manufactures in After a gradual appreciation of the peso that took it to total merchandise exports rose to around 80 percent in around 10 pesos per U.S. dollar in July 2008, the foreign 2008, compared to about 12 percent in 1980 (World exchange pressures of the third quarter of 2008, derived Development Indicators, http://data.worldbank.org), when from capital outflows and the shock to external demand, the export structure was dominated by oil (which then produced a rapid depreciation of the peso through Febru- represented over two-thirds of total exports). Export con- ary 2009, when the exchange rate reached a peak of around centration in the U.S. market increased significantly after 15 pesos per U.S. dollar. The central bank then took a num- NAFTA came into effect in 1994, jumping from 82.7 percent ber of measures to reverse the depreciation and eventually in 1993 to 88.7 percent in 2000. Since then, however, the stabilized the exchange rate at around 13 pesos per U.S. export structure has diversified as all other regions increased dollar. In March 2009, through an auction mechanism the their importance in Mexican exports, especially the rest of bank started daily sales for US$100 million. These daily the Americas (which increased its share from 6 percent to sales fell to US$50 million from the beginning of June to 10 percent between 2000 and 2008). But dependence on the early September, when the announcement was made U.S. market continues to be very high, with the United States that the sales would continue until the end of September as the destination for more than 80 percent of Mexico's total and then stop on October 1. In addition, daily auctions exports of goods (figure 14.3). of US$250 million­US$400 million were undertaken at a minimum exchange rate 2 percent higher than the New Vulnerabilities exchange rate the day before. Moreover, two measures to help bolster confidence during the global crisis were These three features--a high degree of trade openness, a adopted: in October 2008 the establishment by the U.S. large concentration in manufacturing exports, and a high Federal Reserve of temporary liquidity swap facilities with dependence on the U.S. market--have solved old problems: the Banco de México of up to US$30 billion8 (extended in the excessively inward orientation of the economy before February and again in June 2009); and the approval in trade liberalization, the high dependence of foreign trade The Economic Crisis of 2008­09 and Development Strategy: The Mexican Case 175 Figure 14.3. Mexico's Exports to the World, 2008 Table 14.3. Size of the Fiscal Stimulus as Percentage of GDP, % of total exports 2008­10 Country % Country % 0.6% 3% United States 5.6 Netherlands 1.5 Korea, Rep. 4.9 United Kingdom 1.4 6.2% Australia 4.6 Mexicoa 1.3 New Zealand 4.3 Austria 1.1 Canada 4.1 Slovak Republic 1.1 10% Luxemburg 3.6 Poland 1.0 Spain 3.5 Norwaya 0.8 Finland 3.1 Portugal 0.8 Czech Republic 3.0 France 0.6 Germany 3.0 Switzerland 0.5 Sweden 2.8 Italy 0.0 Denmark 2.5 Hungary ­4.4 Japan 2.0 Ireland ­4.4 80.2% Belgium 1.6 Iceland ­9.4 Source: OECD 2009a. a. Unavailable information for 2010. The role of these new vulnerabilities should not, how- United States rest of America ever, be exaggerated. First, these "vulnerabilities" become "strengths" in times of expansion of the U.S. economy. In Europe Asia other particular, the composition of exports, strongly biased Sources: INEGI, Estadísticas Históricas de México, 2009; Banco de México, toward manufactures with a high income elasticity of Informe Anual, 2008b. demand, gives exports a very high dynamism when exter- nal demand grows. Second, because, as we show in the next on oil exports and high vulnerability of the economy to section, Canada, where foreign trade has a similar impor- fluctuations in oil exports, and the failure to exploit the tance and structure, suffered much less from the U.S. reces- opportunities presented by proximity to the largest market sion than Mexico, by reason of its different policy response in the world. But these features have at the same time cre- to the crisis. ated new vulnerabilities, which have exacerbated the nega- tive effects of the external shock on the domestic economy. The Role of the Policy Response The high trade ratio enhances the impact on GDP of a fall in exports of a given product. Moreover, Mexico's exports The policy response also needs to be considered in are predominantly manufactures, in particular consumer explaining the severity of the crisis. Given the size of the durable and capital goods, and international trade in these external shock, Mexico's domestic policy response could products has been affected the most severely by the global have been more aggressive. The fiscal policy response, in crisis (SHCP 2010). In contrast, one consequence of the particular, was timid. First, note that the size of the fiscal increased economic integration with the U.S. economy has stimulus is one of the smallest among the OECD countries been the synchronization of the economic cycles in the (table 14.3), while the size of the external demand shock United States and Mexico (see Sosa 2008). The correlation was probably one of the largest. Second, the fiscal space between the GDP cycles in Mexico and the United States available for a greater stimulus seemed to have been there. was above 0.8 in 1996­2008, while between 1985 and 1995 The central government debt as a percentage of GDP is on this correlation was practically nonexistent (SHCP 2010). the order of 25 percent (2007), a very low value compared The correlation between the cycles of industrial activity to other OECD countries. In a group of 20 OECD coun- since 1997 is even greater (close to 0.99) (Esquivel 2009). tries, only three (Australia, New Zealand, and Norway) had Moreover, the industrial sector is more important in total a ratio lower than Mexico, which in turn had less debt as a GDP in Mexico than in the United States (30.3 percent percentage of GDP than, for example, France, Germany, versus 19.1 percent), so that for a similar fall in industrial the United Kingdom, or the United States (OECD 2009c).9 output, the impact on Mexico's GDP is greater than that on The balanced-budget rule and, probably too, fears that a U.S. GDP (SHCP 2010). larger fiscal deficit would have led to a deterioration in the 176 Managing Openness country's credit rating prevented a greater mobilization of GDP (2.7 percent) was one of the mildest in the OECD this fiscal space. area (where the overall decline in GDP was 3.5 percent). Monetary policy could also have been loosened earlier The difference probably has most to do with the more aggres- and more aggressively. The central bank's target interest rate sive domestic policy response in Canada, where the fiscal was increased by 75 basis points in the second and third stimulus package was far larger than in Mexico (4.1 versus quarters of 2008, at a time when developed economies were 1.3 percent of GDP) and the reduction in the central bank's well into the process of aggressively cutting interest rates target interest rate more pronounced (table 14.4). (see, in particular, the case of Australia, Canada, and the Republic of Korea, countries with inflation-targeting Looking Ahead: The Crisis, Macroeconomic regimes similar to Mexico's). Inflationary pressures, result- Policy, and Development Strategy ing largely from the increase in commodity prices in the first half of 2008, and fears that these pressures would lead to an A vigorous recovery in the third and fourth quarters of upward revision of inflation expectations by economic 2009 followed the recession of 2008­09. The economy of agents were the main reason that the Banco de México Mexico, largely as a result of what has already happened, is increased the target interest rate (see Banco de México expected to grow by about 4 percent in 2010. However, 2008a). Later, when in the first quarter of 2009 the central beyond the short-term recovery, Mexico faces the danger bank began reducing its target interest rate in the face of that without a number of policy reforms the economy the sharp contraction of economic activity, it did so less could return to the lackluster growth performance that aggressively than other developing countries with inflation- characterized the period 2000­2007, when GDP grew at targeting regimes (such as Brazil, Colombia, and Peru). 2.4 percent per year. The successful attempt to reverse the depreciation of the In the current strategy to put the Mexican economy on a peso, after the exchange rate reached a peak of about 15 path of high and sustained growth, the emphasis has been pesos per dollar in early 2009, is also questionable. With on microeconomic reforms in the labor market, the energy fiscal policy subject to a balanced-budget rule and mone- sector, competition policy, and other areas (see SHCP tary policy focused exclusively on price stability, the 2010, 6 and 7, for the government's reform agenda). In this exchange rate is practically the only automatic stabilizer strategy, macroeconomic policy reforms are hardly men- that the economy has when confronted by external shocks tioned (with the exception of tax reform). It would appear to aggregate demand. that, according to this approach, all that macroeconomic A comparison between Canada and Mexico illustrates the policy can do to foster growth is to control inflation and overall role of the policy response in explaining the severity thus that, even when inflation is low, inflation control must of the recession. As shown in table 14.4, both countries have be its main, if not its only, objective. In effect, with macro- nearly identical degrees of openness and concentration of economic stability (narrowly defined as low and stable their exports in the U.S. market. Yet, the recession was far inflation) having been achieved, macroeconomic policy more severe in Mexico than in Canada, where the fall of can do little, if anything, to accelerate growth. While microeconomic reforms in a number of areas are desirable in their own right, the disappointing growth Table 14.4. A Comparison of Selected Economic Data on Canada and Mexico performance in recent decades is probably not due to the lack of these reforms as often asserted (for a discussion of Canada Mexico this point, see Ros 2008 and Moreno-Brid and Ros 2009, Trade openness (%) 67.8 66.9 ch. 10). Thus, the currently dominant view fails to recog- U.S. share of total exports (%) 77 80 nize that the lackluster growth performance of Mexico in 2009 change in GDP (%) ­2.7 ­6.5 recent decades has macroeconomic roots and that greater Fiscal stimulus (% of GDP) 4.1 1.3 emphasis should be put on the need to reform macroeco- Reduction in interest rate in 2008 and 2009 (basis points) 450 300 nomic policies to provide a more growth-oriented macro- economic framework. Sources: Trade openness: Penn World table 6.3 (database), University of Pennsylvania, http://pwt.econ.upenn.edu; U.S. share of exports: Statistics Canada (database), Statistics Canada, http:// www.statcan.gc.ca; INEGI database, INEGI, http://www.inegi.org.mx; GDP growth: INEGI database, Growth, Public Investment, Procyclical Fiscal Policy, INEGI, http://www.inegi.org.mx; OECD Economic Outlook 86 database, and Exchange-Rate Appreciation OECD, http://www.oecd.org/document/18/0,3343,en_2649_34109 _20347538_1_1_1_1,00.html; fiscal stimulus: table 14.3; Interest rate reduction: Banco de México, http://www.banxico.org.mx, and Bank of Canada, A solid diagnostic of the lack of growth in the Mexican http://www.bankofcanada.ca. economy must start by recognizing that the slowdown in The Economic Crisis of 2008­09 and Development Strategy: The Mexican Case 177 productivity growth that "explains," in a growth account- 2001­06, Mexico was last, among the large Latin American ing sense, the growth slowdown of the Mexican economy economies for which information is available, in infra- in recent decades is largely an endogenous phenomenon, structure investment as a fraction of GDP, in both public determined by the lack of growth itself and the resulting and private investment (figure 14.4). During that period, expansion of employment in the low-productivity sectors Mexico invested in infrastructure four percentage points of of the economy (see Ros 2008 and Moreno-Brid and Ros GDP less than Chile, the only country in figure 14.4 with 2009, ch. 10). Rather than a deceleration in exogenous an increase in infrastructure investment between 1981­86 productivity growth, the proximate determinant of slow and 2001­06. With the exception of Brazil, Mexico also growth is the low rate of physical capital formation, which stands out for having the largest contraction in infrastruc- fell from 7.3 percent per year in 1960­81 to 4.0 percent in ture investment (50 percent). The fall in infrastructure 1990­2008 and was only 3.4 percent for the period investment took place in road construction, water provi- 1981­2008 as a whole.10 The inability of capital formation sion, and electricity. Only in the case of telecommunica- to grow at a fast pace--after the years of decline during the tions was there a recovery of investment in the 1990s. Even debt crisis--has prevented the expansion of employment in this case, however, Mexico is today lagging behind other in high-productivity sectors and the modernization of pro- Latin American countries such as Brazil and Chile, which ductive capacity while at the same time constraining the were behind Mexico in 1980 (Calderón and Servén 2004). growth of aggregate demand. A second macroeconomic factor that constraints invest- In turn, the proximate determinant of the reduction in ment and growth is the tendency of fiscal policy to operate the investment rate has been the contraction of public procyclically. This tendency has been amply illustrated by investment. While total fixed investment fell by almost the available empirical studies. Pastor and Villagómez 5 percentage points of GDP between 1979­81 and (2007) discuss the procyclical behavior of fiscal policy 2004­07, public investment fell even more (collapsing by during the period 1990­2003, showing how the struc- 6.6 percentage points). Whether there are crowding-out or tural balance of the public sector increases in times of crowding-in effects of public investment on private invest- recession (1995, for example) and falls during expan- ment is subject to controversy (see, for opposite views, sions (1992­94 and 1999­2000). Other studies finding Lachler and Aschauer 1998, who find a partial crowding- procyclical behavior in Mexico's fiscal policy are Talvi out effect, and Ramirez 2004, who finds an important and Végh (2000), World Bank (2001), and Kaminsky, crowding-in effect). There is, however, consensus on the Reinhart, and Végh (2004). view that, even if crowding-out effects exist, they are at The tendency of fiscal policy to be procyclical has its worst partial; that is, an increase in public investment origin in the procyclical behavior of capital markets and in increases total investment rather than displacing fully an a very low tax burden, which implies that public finances equal amount of private investment. It follows that the continue to be very vulnerable to changes in volatile oil decline in public investment is partly responsible for the revenues. Moreover, the balanced-budget rule has further fall in the overall investment rate and may even have had accentuated the procyclical character of fiscal policy. As is an adverse effect on private investment (if crowding-in well known, maintaining a constant fiscal deficit through effects predominate). the business cycle exacerbates the economic cycle, because The fall in public investment has partly to do with pri- it ensures that the structural balance increases in recession vatizations and partly also to do with the type of fiscal years and falls during periods of expansion. Yet, this princi- adjustment followed after the debt crisis. Mexico's macro- ple has inspired fiscal policy management during the past economic adjustment was successful in correcting fiscal several administrations and was institutionalized in the imbalances, in the sense of eliminating high and unsustain- legislation on fiscal responsibility approved in 2006. able public deficits. However, it relied excessively on the The procyclical behavior of fiscal policy exacerbates the contraction of public investment and failed dramatically in negative effects of shocks to economic activity and has the task of strengthening non-oil tax revenues. As shown adverse consequences on long-run growth. Indeed, a by Giugale, Lafourcade, and Nguyen (2001), since 1980 fis- greater volatility in economic activity, with the correspon- cal deficit reductions have correlated closely with the fall in ding increase in uncertainty, reduces the rate of invest- public investment (the correlation coefficient between the ment and modifies the composition of investment in favor two turning out to be 0.82 between 1980 and 1997). of short-term investments (with negative effects on pro- Infrastructure investment, which has the largest poten- ductivity growth), while deeper recessions can have irre- tial for affecting productivity growth and private invest- versible adverse effects on "learning by doing" and on ment, has suffered in this contraction. In the period workers' skills. 178 Managing Openness Figure 14.4. Infrastructure Investment by Selected Latin American Countries, 1981­86 and 2001­06 6 5 4 % of GDP 3 2 1 0 ile a il ru a o ile a il ru a o az az bi in bi in ic ic Pe Pe Ch Ch ex ex nt nt m m Br Br lo lo ge ge M M Co Co Ar Ar 1981­86 2001­06 private public Source: Calderón and Servén (forthcoming). The appreciation of the real exchange rate in 1988­94 overvalued real exchange rates and the extraordinarily high and later in the period 2000­2007 further conspired growth rates in countries that have deliberately underval- against investment in manufacturing and more generally ued their real exchange rate (as in China and, in the Latin in tradable goods sectors. While real exchange-rate appre- American context and to a lesser extent, Argentina). As a ciation can encourage fixed investment in developing result, today a large literature documents a positive rela- countries by lowering the relative prices of imported tionship between a high real exchange rate and growth. machinery and equipment, it also shifts relative prices in In this recent literature, the effect of the real exchange favor of nontradable goods sectors, reducing the profitabil- rate on investment profitability, already mentioned, is the ity of the tradable goods sectors and inhibiting capital main but not the only mechanism through which the real accumulation in these activities. Blecker (2009) estimates exchange rate affects growth. The impact of investment on that the real appreciation of the peso between 1996­2000 growth is amplified by the greater productivity of invested and 2003­07 (a 9 percent appreciation) reduced Mexico's capital, mainly in sectors subject to international competi- GDP growth rate by as much as 1.3 percentage points.11 He tion (Polterovich and Popov 2002). Moreover, a high real also finds that an appreciated peso has had a more negative exchange rate reallocates resources toward the tradable effect on growth since trade liberalization and especially goods sectors, which present increasing returns to scale and since NAFTA took effect. Ibarra (2010) also finds a highly whose expansion generates economies of specialization and significant effect of the real exchange rate on investment technological externalities that are captured by the less even after controlling for industrial production and dynamic sectors (Frenkel and Ros 2006; Rodrik 2008). Levy exports, suggesting, therefore, a very significant "profitabil- Yeyati and Sturzenegger (2007) highlight the increase in ity effect" of the real exchange rate. domestic saving that results from the redistribution of The empirical evidence for other countries is also quite income in favor of profits derived from a high real exchange conclusive. Indeed, the relationship between the real rate. In small and medium firms with limited access to exchange rate and growth has received much attention in external finance, the resulting increase in internal funds has recent years after the slow growth of many countries with a positive effect on investment. The increase in overall The Economic Crisis of 2008­09 and Development Strategy: The Mexican Case 179 savings also tends to reduce the cost of capital for large rates or intervening in the foreign exchange market at firms that finance their investments in the capital market. times when the exchange rate hits the floor but allowing Although macroeconomic factors are not the only the exchange rate to float freely otherwise. Thus, under this determinants of slow growth--broader diagnostics should alternative the central bank does not target a particular real also refer to the role of industrial policy and the lack of exchange rate but only establishes a floor on it is value. bank financing for productive activities (see, on the sub- An objection to such a proposal is that by defending the ject, Moreno-Brid and Ros 2009)--this analysis of the floor the central bank loses control of the money supply causes of slow growth in the Mexican economy suggests and thus could fail to maintain the inflation target (for a that a good deal of the economic policy agenda for growth fuller discussion, see Frenkel and Rapetti 2004). The prob- should focus on the reform of macroeconomic policy. lem arises at times of excess supply of foreign currency as a result, in particular, of massive capital inflows. It is worth noting, however, that speculative capital inflows will tend The Reform of Monetary and to be deterred to the extent that the central bank clearly Exchange-Rate Policy signals that it will prevent the appreciation of the domestic Consider first monetary and exchange-rate policy. The currency, thus stabilizing exchange-rate expectations.14 empirical evidence seems to suggest that Mexico's central bank has adopted a strict inflation-targeting regime, Tax Reform and Public Investment focused exclusively on inflation targets, rather than a flexi- ble framework that takes into account the output gap in Relaxing the constraints on public investment leads the design of monetary policy (for a review of the evidence, directly to the issue of tax reform and the mobilization of see Esquivel 2009).12 Some related evidence also shows that the fiscal space given by low public debt. Despite recent monetary policy reacts asymmetrically to exchange-rate attempts at tax reform by the Fox and Calderón administra- shocks (tightening in the face of depreciations while not tions, non-oil tax revenues, at about 11­12 percent by the loosening enough in the face of appreciations) (Galindo mid- to late-2000s, continue to be extremely low by inter- and Ros 2008). national standards, well below those prevailing in OECD If one accepts that price stability is consistent with mul- countries and even below those of Latin American coun- tiple configurations of real wages, interest rates, and tries with similar per capita incomes (see OECD 2007). exchange rate and that some of them are more favorable A priority area for investing the increased resources than others to economic growth, it follows that, without from a major tax reform is the development of the poor violating the constitutional mandate that requires the cen- southern part of Mexico. A "new deal" that builds on the tral bank to pursue price stability, monetary policy should productive potential of the south and allows it to reach the seek such stability within the set of configurations favor- level of development of the rest of the country would both able to growth.13 Moreover, given that the configurations generate further growth and reduce regional inequalities. of relative prices that inhibit growth have often proved to This effort requires eliminating the regional bias that most be unsustainable, prudence dictates seeking price stability public capital policies or development incentives have only within a context favorable to growth. There is thus a shown so far. Infrastructure investment in these regions case for increasing the flexibility of the currently very strict and the introduction of positive discrimination in their inflation-targeting framework and its combination with favor in other policies would open up new areas of invest- some form of real exchange-rate targeting. Such a change ment and new markets, thus liberating a growth potential would strengthen resistance to the temptation to use the that would contribute, for a considerable period, to a exchange rate as a counterinflationary policy instrument, higher rate of growth of the economy as a whole (see, on which is important for curbing the tendency toward real the subject, Dávila, Kessel, and Levy 2002). Although there currency appreciation, and establish a competitive real are no estimates for Mexico, the available estimates for exchange rate. This approach requires systematically avoid- Latin America suggest that the effects of such policies ing exchange-rate overvaluation especially at times of could be considerable. Indeed, using a cross-country recession: to avoid, that is, the implementation of a pro- regression framework, Calderón and Servén (forthcom- cyclical monetary policy such as the one followed in the ing) estimate that growth in Latin America would rise by early 2000s. More precisely, the central bank could pro- approximately two percentage points per year if the infra- mote a competitive exchange rate by establishing a sliding structure development of each Latin American country floor to prevent excessive appreciation (see Galindo and were to rise to match the average among non­Latin Ros 2008). This measure would imply managing interest American middle-income countries. 180 Managing Openness Toward a Countercyclical Fiscal Policy explanation for Mexico's slow rate of economic growth in the postreform period suggests that other interpretations Another aspect of the reform of fiscal policy refers thus to of Mexico's growth slowdown may be missing the point. its procyclical behavior. The most frequently given justifi- The alleged incompleteness of microeconomic reforms is cation for the adoption of a procyclical fiscal policy holds sometimes put forward as an explanation of Mexico's that governments characterized by their "fiscal excesses" in slow growth. This topic has been addressed elsewhere the past have credibility problems. As discussed in Casar (see Ros 2008; Moreno-Brid and Ros 2009, ch. 10), and Ros (2004), in the case of Mexico this argument seems where the argument is that the slow productivity growth an exaggeration after four six-year administrations in rate, which is supposedly determined by the lack of which the principal banner of economic policy has been microeconomic reforms, is in fact a consequence of the the achievement of "healthy public finances." Moreover, low rate of investment. even if the diagnosis of a lack of credibility were true, the Mexico has integrated into the international economy as response should be to establish credibility rather than to an exporter not only of oil but also of manufactures, both continue with a practice that deepens recessions. One can labor-intensive manufacturing (garment and assembly of think, for example, of institutional arrangements--such electronic products) and medium- and high-technology as a multiyear fiscal rule with a ceiling on public indebted- intensity manufacturing (automobiles, metal, and mechani- ness as a proportion of GDP--that guarantee the stability cal industries). Thus, the share of manufactures in total of public finances over the medium term without elimi- exports climbed from below 20 percent in the early 1980s to nating the room for fiscal policy to maneuver in the short nearly 80 percent today, at the expense of the share of oil run (for a proposal in this direction based on the adoption exports, which fell from over two-thirds of the total in the of a structural balance, see Pastor and Villagómez 2007). early 1980s to less than 20 percent today. This transformation The growth benefits of a countercyclical fiscal policy is remarkable in the international context. Today, the share of should not be underestimated, as suggested by recent manufacturing exports in total exports is higher in Mexico research (Aghion, Barro, and Marinescu 2006; Aghion and than in several Latin American and East Asian countries. Marinescu 2006) and its successful application in, for Moreover, manufacturing exports have a high share of example, Chile. consumer durable and capital goods with a high income It is worth noting that both the IMF and the OECD are elasticity of demand. The technology intensity of Mexican advocating a reform of the balanced-budget rule. For exports has been increasing over time with the share of example, in the most recent Article IV consultation the high-technology exports in total manufacturing exports IMF staff report states: reaching close to 20 percent by 2006. Although lower than The balanced budget rule has been a key anchor, but there that of China, Korea, Malaysia, and Thailand, this share is is a case for moving towards a structural rule over time. well above that of Argentina, Brazil, and Chile. A significant The current rule has built up the credibility of the fiscal share of medium- and high-technology-intensive exports framework. But moving to a structural rule would lower has also been found in other studies for the Mexican case procyclicality and reduce the asymmetry inherent in the (Moreno-Brid, Santamaría, and Rivas 2005). current framework. (IMF 2010b, 32) It is also important to observe that among the major Similarly, the OECD in the 2009 Economic Survey competitors in the U.S. market, Mexico is the only one of Mexico states: that has not suffered a major decline in its share of U.S. imports in the face of the rapidly increasing participation Fiscal policy is highly dependent on volatile oil income. The balanced budget rule can create a bias for spending oil rev- of China. In fact, despite a slight decline throughout the enues as they are earned, especially as transfers to the stabi- 2000s, Mexico's share in U.S. imports was higher in 2001­08 lization funds are limited by caps at low levels. This can than in 1994­2000. This position is in stark contrast with potentially lead to a pro-cyclical bias in fiscal policy . . . To the decline of almost six percentage points in Japan's share better manage budget cycles and oil wealth, Mexico should or the fall on the order of two percentage points in Canada's establish a structural deficit fiscal rule. (OECD 2009b) share, the other major trading partners of the United States besides China and Mexico. All this suggests that the pattern of integration with Is There Anything Wrong with Mexico's Insertion the world economy has not been a drag on Mexico's eco- in the World Economy? nomic growth. In fact, the nature of Mexican exports The emphasis on the low rate of investment in physical should, if anything, have generated faster economic capital and the macroeconomic factors behind it as an growth. Hausmann, Hwang, and Rodrik (2005) show that The Economic Crisis of 2008­09 and Development Strategy: The Mexican Case 181 the level of technological sophistication of a country's 2. According to CEPAL estimates. The most recent Instituto Nacional de Estadística y Geografía figure for GDP in 2009 shows a decline of exports relative to its per capita income is a good predictor of 6.5 percent. a country's subsequent growth, and it is clear from their find- 3. The price of the Mexican oil export mix rose from an annual average ings that the level of sophistication of Mexico's export basket of US$53.04 per barrel in 2006 to a peak of US$132.71 (July 14, 2008), falling is relatively high, higher certainly than that of other Latin thereafter to a minimum of US$26.23 on December 24, 2008 (SHCP 2010). 4. Foreign banks represent about 80 percent of the Mexican banking American economies with similar or even higher per capita system (Sosa 2008). income (Argentina and Chile, for example). In other words, 5. The program also aims to ease credit to small and medium enter- Mexico has had poor growth performance in recent decades prises and a gradual reduction in tariffs applied to a wide range of goods between 2009 and 2013 (OECD 2009c). despite a rather favorable pattern of trade specialization. 6. Also worth noting is the national agreement in support of house- holds and employment, announced in January 2009, although the main measures of this program, besides investments in infrastructure, are not Conclusions fiscal policy actions but rather aim at improving the competitiveness of small and medium enterprises (OECD 2009c). The deep recession of 2008­09 in the Mexican economy 7. IMF (2010b) estimates the fiscal stimulus at 2.5 percent of GDP, confirms that shocks to U.S. industrial production consti- but it is unclear if this is net of the budget cuts undertaken in the second tute the largest foreign source of macroeconomic fluctua- half of 2009. 8. These new facilities also included the central banks of Brazil, the tions in Mexico in the post-NAFTA period. The external Republic of Korea, and Singapore. demand shock caused a severe industrial recession led by 9. In the nontraditional measure of the public debt--which includes manufacturing exports. It was further complicated by the the net debt of the federal government, nonfinancial public enterprises, development banks, official trust funds, liabilities related to banking sec- collapse of private investment arising from the capital tor restructuring, and PIDIREGAS--the ratio to GDP was 40 percent at account shock and, less dramatically, by the fall in family the end of 2008 (OECD 2009c) and also quite low by OECD standards. remittances in 2008 and 2009. The recession was the most 10. These amounts are based on estimates of the nonresidential net severe since the 1930s and the deepest in Latin America, a capital stock provided by André Hofman. 11. Other factors contributing to the fall in the growth rate (of about feature explained by the very close ties of the Mexican 2.1 percentage points between the two periods) were a slower growth of economy with the U.S. economy and also by the limited the U.S. economy and lower net financial inflows while a higher real oil countercyclical response of economic policy, especially fis- price had a positive effect on the growth rate. 12. See, however, Villagómez and Orellana (2009) for an opposite view. cal policy whose room for maneuver was constrained by 13. The mandate could of course also be changed to consider eco- the balanced-budget rule. By contrast, the depreciation of nomic growth as an additional objective. The Mexican Senate is currently the exchange rate served as a cushion that made the col- considering such reform. 14. When the increase in foreign currency is associated with massive lapse of exports less severe than otherwise and helps inflows of family remittances or sharp improvements in the terms of explain the early and vigorous recovery of exports and trade, the challenge of avoiding a persistent real exchange-rate apprecia- GDP since the second half of 2009. tion may be more complex but still manageable with appropriate interest A growth diagnostics exercise suggests that past devel- rate management and intervention in the foreign exchange market. opment strategy has neglected the macroeconomic fac- tors--such as low public investment, a procyclical fiscal Bibliography policy, and a tendency toward the real appreciation of the Aghion, P., R. Barro, and I. Marinescu. 2006. "Cyclical Budgetary Policies: peso--that are behind the lackluster growth of the econ- Their Determinants and Effects on Growth." Mimeo, Harvard Uni- omy in the 2000s. A redirection of development strategy versity, Cambridge, MA. should thus consider the reform of macroeconomic poli- Aghion, P., and I. Marinescu. 2006. "Cyclical Budgetary Policy and Eco- cies as a central ingredient of the reforms necessary to put nomic Growth: What Do We Learn from OECD Panel Data?" Mimeo, Harvard University, Cambridge, MA. Mexico on a path of fast and self-sustained growth. That Banco de Mexico. 2008a. "Informe sobre la inflación." Julio­Septiembre. goal implies the mobilization of the fiscal space, including http://www.banxico.org.mx. tax reform, necessary to recover high rates of public invest- ------ . 2008b. Informe Anual. Mexico: Banco de Mexico. ------ . 2010. Informe anual 2009. www.banxico.org.mx. ment, the establishment of a competitive and stable real Blecker, R. 2009. "External Shocks, Structural Change, and Economic exchange rate, and the move toward a structural budget Growth in Mexico, 1979­2000." World Development 37 (7): 1274­84. rule that allows fiscal policy to act countercyclically. Calderón, C., and L. Servén. 2004. "Trends in Infrastructure in Latin America, 1980­2001."Policy Research Working Paper 3401, World Bank, Washington, DC. Notes ------. Forthcoming. "Infrastructure in Latin America." In Oxford Hand- book of Latin American Economics, ed. J. A. Ocampo and J. Ros. 1. The increase in unemployment is considerable, given that the Casar, J., and J. Ros. 2004., "¿Porqué no crecemos?" Nexos. Octubre. unemployment rate is usually very low due to the definition of unemploy- CEPAL (United Nations Economic Comission for Latin America and the ment and the role of the informal sector as an employment shock Caribbean). 2009. Balance preliminar de las economías de América absorber (see Frenkel and Ros 2004, for further discussion). Latina y el Caribe, 2009. Santiago: CEPAL. 182 Managing Openness Dávila, E., G. Kessel, and S. Levy. 2002. "El Sur también existe: Un ensayo Levy Yeyati, E., and F. Sturzenegger. 2007. "Fear of Appreciation." Policy sobre el desarrollo regional de México." Economía Mexicana 11 (2): Research Working Paper 4387, World Bank, Washington, DC. 205­60. Moreno-Brid, J. C., and J. Ros. 2009. Development and Growth in the Mexican Esquivel, G. 2009. "De la inestabilidad macroeconómica al estancamiento Economy: A Historical Perspective. New York: Oxford University Press. estabilizador: El papel del disenño y conducción de la política Moreno-Brid, J. C., J. Santamaría, and J. C. Rivas. 2005. "Industrialization económica en México." Mimeo. and Economic Growth in Mexico after NAFTA: The Road Traveled." Frenkel, R., and M. Rapetti. 2004. "Políticas macroeconómicas para el Development and Change 36 (6): 1095­119. crecimiento y el empleo." Paper prepared for the International Labour OECD (Organisation for Economic Co-operation and Development). Organization. 2007. Getting It Right: OECD Perspectives on Policy Challenges in Frenkel, R., and J. Ros. 2004. "Desempleo, políticas macroeconómicas y Mexico. Paris: OECD. flexibilidad del mercado laboral. Argentina y México en los noventa." ------. 2009a. Economic Outlook: Interim Report March 2009. Paris: Desarrollo Económico 173 (44): 33­56. OECD. ------. 2006. "Unemployment and the Real Exchange Rate in Latin ------. 2009b. Economic Survey of Mexico 2009. Paris: OECD. America." World Development 34 (4): 631­46. ------. 2009c. OECD Review of Budgeting in Mexico 2009. Supplement 1. Galindo, L. M., and J. Ros. 2008. "Alternatives to Inflation Targeting." Paris: OECD. International Review of Applied Economics 22 (2): 201­14. Pastor, J., and A. Villagómez. 2007. "The Structural Budget Balance: A Pre- Giugale, M., O. Lafourcade, and V. H. Nguyen, eds. 2001. Mexico: liminary Estimation for Mexico." Applied Economics 39 (12): 1599­607. A Comprehensive Development Agenda for the New Era. Washington, Polterovich, V., and V. Popov. 2002. "Accumulation of Foreign Exchange DC: World Bank. Reserves and Long Term Growth." New Economic School, Moscow. Hausmann, R., J. Hwang, and D. Rodrik. 2005. "What you Export Mat- http://mpra.ub.uni-muenchen.de/20069/1/ExchangeGrowth.pdf. ters". Working Paper 11905, National Bureau of Economic Research, Ramirez, M. 2004. "Is Public Infrastructure Investment Productive in the Cambridge, MA. Mexican case? A Vector Error Correction Analysis." Journal of Interna- Ibarra, C. 2010. "Capital Flows, Real Exchange Rate, and Growth Con- tional Trade and Economic Development 13 (2): 159­78. straints in Mexico." Mimeo, Department of Economics, Universidad Rodrik, D. 2008. "The Real Exchange Rate and Economic Growth." de las Américas. Brookings Papers on Economic Activity (Fall). IMF (International Monetary Fund). 2010a. "Mexico: Arrangement Ros, J. 2008. "La desaceleración del crecimiento económico en México under the Flexible Credit Line and Cancellation of the Current desde 1982." El Trimestre Económico 299: 537­60. Arrangement." Country Report 10/81. International Monetary Fund, SHCP (Secretaría de Hacienda y Crédito Público). 2010. Weekly Report. Washington, DC. March 22­ 26. Mexico City: SHCP. ------. 2010b. "Staff Report for the 2010 Article IV Consultation." IMF, Sosa, S. 2008. "External Shocks and Business Cycle Fluctuations in Washington, DC. Mexico: How Important Are U.S. Factors?" Working Paper 08/100, ------. 2010c. "World Economic Outlook Update: January 2010." IMF, International Monetary Fund, Washington, DC. Washington, DC. Talvi, E., and C. Végh. 2000. "Tax Base Variability and Procyclical Fiscal INEGI (Instituto Nacional de Estadística y Geografía). 2009. Estadísticas Policy." Working Paper 7499, National Bureau of Economic Research, Históricas de Mexico: INEGI. Cambridge, MA. Kaminsky, G., C. Reinhart, and C. Végh. 2004. "When It Rains, It Pours: Villagómez, A., and J. Orellana. 2009. "Monetary Policy Rules in a Small Procyclical Capital Flows and Macroeconomic Policies." Working Open Economy: An Application to Mexico." Working Paper, Tecnológico Paper 10780, National Bureau of Economic Research, Cambridge, MA. de Monterrey, Monterrey. Lachler U., and D. Aschauer. 1998. "Public Investment and Economic World Bank. 2001. "Fiscal Policy, Business Cycles and Growth in Mexico: Growth in Mexico." Policy Research Working Paper 1964, World Bank, Fiscal Sustainability." Report 20236-ME, World Bank, Washington, Washington, DC. DC. 15 The International Crisis and Development Strategies: The Case of Chile Roberto Zahler During the past 25 years, the economic performance of very sporadic interventions by the Central Bank of Chile in Chile has been quite satisfactory. Although it is embedded the foreign exchange market, the existence of a maximum in a relatively unstable region, far away from the great trade conventional interest rate, and a minimum wage. centers, this small country functions well, is open to goods The state has contributed to increasing private sector and financial markets, and has a vibrant private sector. participation in the economy through the privatization of public companies and the franchising of infrastructure Chilean Development Strategy works. This effort has enabled the private sector to build and operate highways, airports, prisons, and hospitals. The allocation of resources in Chile takes place through The most important results for the past 25 years may be decentralized markets. Most of the production of goods classified in the following categories: (a) macroeconomic and services, as well as investment, functions within the strength; (b) commercial and financial integration into the private sector. As a productive agent, the state focuses on world economy; (c) a sound financial sector; and (d) global some specific lines of business, such as copper, petroleum, competitiveness. In contrast, Chile's main constraints dur- and railways, as well as other, less significant areas. Apart ing this period have been the limited internal demand from its traditional roles, the state's action centers on regu- faced by domestic companies because of the size of the lating noncompetitive markets and on social areas for country, its per capita income, and its total factor produc- redistributive purposes. tivity (TFP), which after having grown at very high rates The Chilean economy is very open to international trade until the mid-1990s, has experienced less dynamism and and financial markets: the effective average tariff rate on has stagnated during the past decade. imports is less than 2 percent, while goods exports and for- eign direct investment (FDI) account for 34 percent and Macroeconomic Strength 9.9 percent of gross domestic product (GDP), respectively. Markets are able to operate with price freedom. However, the During the past 25 years, the Chilean economy grew, in real state intervenes in infrastructure industries that are subject to terms, relatively quickly, with some variation in subperi- price regulations, such as electricity, telecommunications, ods (table 15.1). In 1986­90 (the recovery period follow- and water, and also intervenes in oil prices through subsidies ing the profound crisis suffered by the Chilean economy in and indirect tax changes. Exchange rates, interest rates, and the early 1980s), the annual average growth of GDP was salaries are freely determined, with the sole exceptions of 6.8 percent. Then, from 1991 to 1995 that rate accelerated I would like to thank Hermann Gonzalez for his contribution to and comments on this chapter. 183 184 Managing Openness Table 15.1. Basic Macroeconomic Data for Chile, 1986­2009 1986­90 1991­95 1996­2000 2001­05 2006­09 GDP growth (year-on-year, %) 6.8 8.7 4.2 4.2 2.7 Inflation rate (average, %) 19.4 14.0 5.2 2.6 4.5 Fiscal balance (% of GDP) 1.3 2.0 0.4 0.9 4.4 Investment rate (% of GDP) 15.4 19.2 21.5 21.1 26.3 International reserves (% of imports) 83.3 100.6 92.5 79.7 49.2 External debt (% of GDP) 77.2 39.1 40.1 51.8 38.3 Unemployment rate (%) 9.7 7.3 7.7 9.7 8.1 Source: Zahler & Co., based on data from the Central Bank of Chile and the Instituto Nacional de Estadísticas: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl; and online database, Instituto Nacional de Estadísticas, Santiago, http://www.ine.cl. to 8.7 percent. Subsequently, as a consequence of the Asian sovereign wealth funds and establishing the basic institu- crisis and the domestic policy response to it, during the fol- tional framework necessary for their management. The lowing two five-year periods the economy's economic Pension Reserve Fund, created at the end of 2006, was growth fell to an annual average of 4.2 percent, and from established to finance minimum pensions; and the Eco- 2006 to 2009 GDP growth fell further, to only 2.7 percent, a nomic and Social Stabilization Fund, launched in early figure determined by the international financial crisis. 2007, was set up to finance other spending and public GDP growth of 4.5­5.0 percent is expected for 2010, as a investment in the future. result of an improved external scenario, expansionary fiscal Fiscal income, particularly that resulting from the high and monetary conditions, and the reestablishment of con- international price of copper, led to the accumulation of fidence among entrepreneurs and consumers. significant surpluses between 2004 and 2008--5.7 percent Chilean inflation was, until the late 1980s, historically and of average GDP--a trend that came to an abrupt stop in chronically high and volatile. In 1990, the central bank 2009. That year recorded a fiscal deficit of 4.5 percent of became autonomous, and since then the 12-month inflation GDP as a consequence of the international crisis, which rate has fallen continuously from levels higher than 20 per- implied a significant drop in fiscal income and required an cent in the early 1990s to an annual average of 2.6 percent in important increase in public spending to partially com- 2001­05, well within the range the central bank established pensate for the fall of both external and internal private as the inflation target (2­4 percent). The years 2007 and demand. 2008 were characterized by higher annual inflation--close The economy's investment rate, measured in real terms, to 10 percent in 2008--as a consequence of the various has increased significantly from the late 1980s, reaching supply shocks affecting international food and fuel prices, over 20 percent of GDP during the past 25 years. Before the with strong repercussions for domestic prices. But since late 2008­09 international crisis, the investment rate reached 2008 and throughout 2009, 12-month inflation dropped an important benchmark, coming close to 30 percent of quite strongly, becoming negative for many months. It is GDP in 2008. expected that during the inflation target period set by the International reserves fluctuated in the range of US$14 central bank (24 months), it will return to the target. billion­US$16 billion from 1994 to 2004, equivalent to just Management of public finances has been prudent, and, under a year of the country's goods imports. Subsequently, on average, during the past 25 years all five-year subperiods they vastly increased, surpassing US$25 billion in 2009. have recorded a surplus. In addition, in the past nine years However, the strong rise in imports during the last five- the public sector has been able to play a countercyclical role year period implied a reduction in the ratio of reserves to without jeopardizing medium-term fiscal indicators. In imports, to half the average of the previous 20 years. 2001, Chile adopted a rule of targeting a structural budget Chilean external debt dropped considerably from surplus as a percentage of GDP. That surplus reflects the the early 1990s, when it was equivalent to 90 percent of level of fiscal revenues and spending if GDP were at its GDP. By the end of 2009, it amounted to US$74 billion potential level and the international price of copper were at (47 percent of GDP). Only 3 percent corresponds to the its long-term trend. Chile took another important step for- public sector and is mainly a long-term obligation. The ward in this direction in 2006 by approving the Fiscal rest originated in the private sector, including 21 percent Responsibility Law. This law allowed for setting up two owed by banks. Seventy-six percent of Chile's total external The International Crisis and Development Strategies: The Case of Chile 185 debt corresponds to foreign commitments by nonbank- surprises have little impact on real wages. In addition, wage ing private enterprises. contracts tend to have a two-year duration, reducing their On average, the unemployment rate has remained rel- ability to react quickly to changes in economic conditions, atively high. Only during the most dynamic growth thus further contributing to wage rigidity. Chile also has a period, in the first half of the 1990s, did it drop to relatively high minimum wage. around 6 percent. However, the low GDP growth in the past three five-year periods contributed to the increase in Integration into the World Economy the unemployment rate once more. In fact, immediately An important part of Chile's economic success is due to after the Asian crisis, Chile's unemployment rose to growth and diversification of its exports, especially with approximately 10 percent. Only from 2006 to 2008 did it respect to their destinations. This result came about ini- fall again, averaging 7.6 percent. In 2009, as a consequence tially when Chile dismantled most quantitative import of the international crisis it increased again, reaching restrictions and began a process of continuously reducing approximately 11 percent in some months. import tariffs (figure 15.1). That process was reinforced Most of the literature that has analyzed this increase in later by an increasing number of free trade agreements, Chile's unemployment highlights two considerations. The which include China, the European Union (EU), Mexico, first is international and consists of successive external and the United States, among others. shocks that affected the Chilean economy after the Asian In fact, exports have been one of the main sources of crisis, such as a fall in the country's terms of trade, the growth in the Chilean economy. Figure 15.2 shows the impact of the 9/11 attack on the United States, and the tur- ratio of the growth of exports of goods and services, in bulence affecting Argentina and Brazil, among others. The real terms, to the growth of GDP, also in real terms. In the second is a local consideration, associated with rigidities in period 1987­2009, excluding the crisis years of 1999 and the labor market that impede adjustment through a lower- 2009, on average the real rate of growth of exports ing of real wages. exceeded by 1.4 times the average rate of growth of GDP. The empirical evidence suggests that labor market rigidi- Too little evidence is available to analyze contraction ties in Chile are stronger than in Argentina and Mexico as periods, because the only year in which exports fell, in well as in most East Asian countries. In Chile, wages respond real terms, was 2009. During the recession of 1999, little to an increase in unemployment and to negative exter- Chilean GDP fell by 0.7 percent, while exports increased nal shocks. In the years following the Asian crisis, many by 7.3 percent. In 2009, both GDP and exports fell, by workers, especially those with low education and qualifica- 1.5 percent and 5.6 percent, respectively. tions and little labor experience, were affected by strong Financial opening up to the international economy was increases in the minimum wage. Furthermore, nominal initially executed gradually, selectively, and pragmatically. wages in Chile are generally indexed to changes in the con- This approach partly explains why the major crises that sumer price index so that inflationary (or deflationary) took place in the 1990s--such as the Mexican ("tequila") crisis and the Russian crisis--had a minimal effect on the Figure 15.1. Effective Average Tariff on Chilean Imports, Chilean economy. Financial openness is reflected in the 1998­2009 growing flows of capital to and from the country, bank percent financing, and other sources of funding and foreign 10 investment. FDI is not discriminatory with respect to resi- 9 dents' investments. In the past 10 years, FDI averaged 8 6.6 percent of Chile's GDP, reaching a maximum of 7 9.9 percent in 2008. FDI is present in many sectors of eco- 6 nomic activity, including mining, industry, finance (banks tariff (%) 5 and insurance companies), energy, telecommunications, 4 and highways. 3 Chilean companies began to invest directly abroad in 2 the early 1990s, the annual peak coming in 2000, when it 1 represented 5.3 percent of GDP; the average of the past 0 10 years was 2.7 percent of GDP. Most of Chilean FDI 98 99 00 01 02 03 04 05 06 07 08 09 abroad is in Latin America--mainly Argentina, Brazil, 19 19 20 20 20 20 20 20 20 20 20 20 year Colombia, and Peru--and is concentrated in retail, energy, Source: Cámara de Comercio de Santiago, Santiago, http://www.ccs.cl/. financing, and cellulose. 186 Managing Openness Figure 15.2. Growth in Exports of Chilean Goods and Services Relative to GDP, 1987­2009 6 3.7 4 2.3 2.0 2.2 2.2 1.6 1.5 1.6 1.6 1.7 1.6 1.6 1.6 2 1.0 1.1 1.0 1.1 1.1 0.5 0.7 0.8 0.8 0 percent ­2 ­4 ­6 ­8 ­10 ­10.0 ­12 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 year exports growth/GDP growth average growth excluding 1999 and 2009 Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. In addition, the country's financial investment abroad is agreements (inflation-indexed financial contracts), the quite significant, through both the sovereign wealth funds presence of institutional investors (mainly private pen- and the private pension funds. Foreign investments in pri- sion funds and life insurance companies), and a housing vate pension funds are concentrated in stocks and bonds. By finance circuit. All this has meant that much of the the end of 2009, they amounted to around US$52 billion, domestic financial savings invested in Chile becomes which represents 44 percent of total pension fund assets and productive investment in the real sector of the Chilean 33 percent of Chile's GDP. economy. Sound Financial Sector Obstacles to Growth Following the deep economic crisis of 1982­83, regulation Three main factors have limited Chile's economic growth. and supervision have been great strengths in the Chilean First, Latin America has been unstable economically and financial system. As the system has developed and become politically; that instability affects capital flows to the region more professional, legislation and regulations have been and the demand for Chilean exports, of which 16 percent adjusted and modernized. are shipped to the region. One significant asset of the Chilean economy in the past Second, the size of the country (GDP of US$157 billion 25 years is the dynamism and strength of its banking sys- in 2009), together with its per capita income, limits inter- tem. From 1995 to 2009, total bank loans grew at an annual nal markets and forces companies to sell abroad to exploit real average rate of 9.3 percent. By the end of 2009, the pro- economies of scale and create efficiency gains. As a result, portion of total past due loans amounted to only 1.35 per- the local economy depends heavily on the health of the cent, and the system's average capitalization index (Basel global economy. Index) was 14.3 percent. A third element that has restricted economic growth, In the early 1980s, social security was reformed from a particularly in the past decade, is low growth in TFP. pay-as-you-go system to one of individual capitalization. Although macroeconomic reforms and a policy framework This reform created institutional investors, which have based on sound and countercyclical fiscal policy, inflation since been a major source of demand for long-term targeting, and a free exchange rate have contributed to financial instruments and have thus helped develop the overall stability and to a reduction in the country's vulner- domestic capital market. That market also benefited ability to shocks, the slower rate of microeconomic reforms from the generalized use of price indexation in financial has limited growth. The International Crisis and Development Strategies: The Case of Chile 187 World Economic Cycle and Chile's Growth growth rate in 2008­10 was lower than that registered in past episodes of high international copper prices (when the Its high degree of financial and commercial openness fiscal rule was nonexistent). The same stabilizing effect implies that Chile is very dependent on the world economy allowed the Chilean economy to grow, on average, more and especially on the country's main commercial partners. than 3 percent at the beginning of the current decade, Therefore, a strong and widespread international recession, when the real price of copper was the lowest since the Great such as the one that took place in 2008­09, is almost sure to Depression. In short, although Chile is very much affected have a negative impact on Chile's economic growth. by the international economy, it now depends much less on In addition to being very open to world trade, Chile has fluctuations in the international copper price. been the biggest producer of copper in the world since Table 15.2 shows GDP growth rates in three recession 1982. Nearly 98 percent of Chilean copper output is episodes: 1982­83, the 1998­99 Asian crisis, and the most exported, and copper has represented, depending on its recent crisis of 2008­09. During 1982­83, world GDP grew price, from a third to a half of total Chilean exports. In the 1.9 percent in purchasing power parity terms, and in 1982 four major international crises of the past 40 years, the real world GDP growth was only 1.2 percent. In that crisis, only copper price has dropped significantly (figure 15.3). In Latin American economies (where the crisis originated) suf- magnitude and speed of adjustment, the 2009 shock resem- fered a drop in GDP. The 1998­99 crisis appears, by far, to bles that of the mid-1970s. In fact, in the mid-1970s, the real have had the least severe effect on world GDP, which grew at copper price fell 64 percent in nine months. Subsequently, an annual average of 3.1 percent. The 2008­09 recession is in the early 1980s, the copper price fell 62 percent, but the the most severe of these three episodes, resulting in an annual drop was much smoother, lasting 28 months, from March average world GDP growth of only 1.2 percent and a fall in 1980 to June 1982. During the Asian crisis, the real copper GDP of 0.6 percent in 2009. In this recession, the unprece- price fall was lower (47 percent) and took place over 21 dented decline in GDP of developed countries and the months, from July 1997 to March 1999. In the latest crisis, 1.8 percent drop suffered by Latin American countries in 2009 the real price of copper fell 63 percent in only eight months. also stand out, comparable only to the occurrence in 1983. Recent fiscal policy has reduced the impact of fluctua- Figure 15.4 shows the annual rate of change of Chilean tions in international copper prices on the Chilean eco- GDP, compared with that of the world and with selected nomic cycle. In fact, Chilean fiscal policy--accompanied economic regions. During the 1980s and 1990s, Chilean by coherent monetary and exchange-rate policies--has growth significantly differs--both in its level and its played a key stabilizing role in the Chilean economy since volatility--from the rest. It is clear that in the early 1980s 2002. Therefore, it is not surprising that Chile's GDP Figure 15.3. Nominal and Real Price of Copper in Chile, 1960­2010 4.0 3.5 copper price (US$ per pound) 3.0 2.5 2.0 1.5 1.0 0.5 0.0 60 63 66 69 72 75 78 81 84 87 90 93 96 99 02 05 08 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 year real nominal Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. 188 Managing Openness Table 15.2. World GDP Growth in Recessions, 1982­83, 1998­99, and 2008­09 1982­83 1998­99 2008­09 1982 1983 average 1998 1999 average 2008 2009 average World 0.9 2.9 1.9 2.6 3.6 3.1 3.0 ­0.6 1.2 Major advanced economies (G-7) ­0.1 3.2 1.6 2.7 3.3 3.0 0.2 ­3.4 ­1.6 Emerging and developing economies 2.2 2.7 2.4 2.5 3.5 3.0 6.1 2.4 4.3 Western Hemisphere ­0.6 ­2.5 ­1.6 2.3 0.4 1.4 4.3 ­1.8 1.3 Source: Zahler & Co., based on purchasing-power parity data from the IMF: International Financial Statistics (database), IMF, Washington, DC, http://www .imfstatistics.org. Note: G-7 = Group of Seven (Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States). Figure 15.4. Growth of GDP in Chile and Selected Country Groups, 1980­2009 15 10 GDP growth (%) 5 0 ­5 ­10 ­15 80 85 90 95 00 05 19 19 19 19 20 20 year world major advanced economies (G-7) emerging and developing economies Chile Source: Zahler & Co., based on International Monetary Fund (IMF) data: International Financial Statistics (database), IMF, http://www.imfstatistics.org. Note: G-7 = Group of Seven (Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States). and late 1990s, the contraction of Chile's growth was recessions (figure 15.5). The 1.5 percent drop in Chile's greater than in all the regions considered. Particularly rele- GDP in 2009 compares with drops in its GDP of around 13 vant is what happened in 1982, when Chile's GDP fell close percent to 14 percent in both 1975 and 1982. Part of this to 14 percent, while world GDP grew by 0.9 percent and improved performance is explained by the different macro- the Latin American region as a whole suffered a contraction economic policy frameworks prevailing in the three of only 0.6 percent. However, Chile's recovery from that cri- episodes and, especially, the capacity of the Chilean econ- sis was quite rapid, and during the rest of the 1980s it grew at omy to apply countercyclical monetary and fiscal policies a much higher rate than the rest of Latin America. in the most recent crisis. In the past 10 years, Chile's growth rate has been very Despite the increasing commercial and financial integra- similar to that of the world as a whole. In 2009, Chile's tion of the Chilean economy, the volatility of its GDP has 1.5 percent fall compares to a decrease of 0.6 percent of been falling since the mid-1980s (figure 15.6). During the world GDP and is clearly lower than the contraction suffered 1998­99 crisis, GDP volatility temporarily increased, but its by developed countries, where the crisis originated. Even level was still less than half that experienced in the 1970s and with the data from the 1960s, it is possible to observe that, 1980s. And volatility continued to fall in the recent crisis. with the greatest external crisis of the past 60 years and with Figure 15.7 shows Chile's real seasonally adjusted GDP the country's increasing financial and commercial integra- for 1998­2000 and 2008­09. With the quarter in which tion into the international economy, Chile's performance in GDP reached its maximum precrisis level as a starting 2009 was much less negative than in the previous world point (the second quarter of 1998 and the second quarter The International Crisis and Development Strategies: The Case of Chile 189 Figure 15.5. World, Latin American, and Chilean Growth in GDP, 1961­2009 15 10 5 GDP growth (%) 0 ­5 ­10 ­15 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05 07 09 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 year Latin America Chile world Source: Zahler & Co., based on data from the IMF, the Central Bank of Chile, and the OECD: International Financial Statistics (database), IMF, http://www.imfstatistics.org; Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl; and OECD Stat (database), OECD, http://stats.oecd.org. Figure 15.6. Growth and Volatility of Chilean GDP, 1965­2009 10 15 9 10 8 7 5 GDP growth (%) volatility (%) 6 5 0 4 ­5 3 2 ­10 1 0 ­15 65 68 71 74 77 80 83 86 89 92 95 98 01 04 07 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 year volatility (5-year window, left) GDP growth (right) Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. of 2008), the impact of both crises on Chile's GDP is very ery speed. In the current crisis, recovery is taking place at a similar as well as the time elapsed between peak and slower rhythm than during the Asian crisis and, after six trough: four quarters. However, with the information up semesters since activity began to diminish, GDP still has to the fourth quarter of 2009, the episodes differ in recov- not recovered to its level before the beginning of the 190 Managing Openness Figure 15.7. Evolution of Chile's Real GDP in the Past Two Crises, 1998­2000 and 2008­09 103 index (June 1998 and June 2008 = 100) 102 101 100 99 98 97 96 95 T­1 T T+1 T+2 T+3 T+4 T+5 T+6 T+7 time period 1998­2000 2008­09 Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Note: GDP is seasonally adjusted. T represents the period of maximum precrisis GDP. contraction. During the Asian crisis, GDP had recovered specifically with regard to Chile. The final part of this section its precrisis level in the fourth quarter of 1999, two quarters discusses the shock that came from global capital markets. after having reached the minimum during the contrac- tionary phase. One possible explanation for the slower speed External Demand Shock: An International Comparison of recovery is that, on average, in 1998­99 world GDP grew 3.1 percent, while in 2008­09 it grew by only 1.2 percent. The external demand shock experienced in 2008 and 2009 From a longer-term perspective, it is interesting to ana- affected all countries in Latin America through two chan- lyze Chile's GDP growth trend before and after the Asian nels: lower demand for their exports and a fall in the price crisis. During the 1990s, 1999 aside, the average growth of exported goods and services. Table 15.3 shows an inter- rate of GDP was 7.1 percent, while from 2000 to 2008 it national comparison of the external demand shock expe- was 4.3 percent. In short, after the Asian crisis, Chile's rienced by several Latin American countries and other growth was much lower than before that crisis. One of the high-income exporters of raw materials. main reasons for the high GDP growth from 1990 until the The upper part of table 15.3 shows Latin American Asian crisis was TFP growth of around 4 percent, a situa- economies. Despite differences in the degree of openness tion that changed dramatically after 1997. and in the composition of exports, the fall in value of Evidence on the determinants of TFP suggests that in Chilean exports does not differ substantially from the aver- addition to cyclical factors, TFP growth reflects the effects age drop in the region, around 50 percent in six months. of macroeconomic stability as well as of structural reforms. The greatest decline was in Ecuador--whose exports Given that Chile's macroeconomic environment has been account for a high proportion of GDP--followed by Brazil, stable for many years, the empirical evidence suggests that the economy with the lowest ratio of exports to GDP. higher GDP growth depends, temporally, on improvements The recovery of exports shows greater homogeneity than in the terms of trade, such as what Chile experienced in the contractionary phase, essentially accounting for the 2004, and, more permanently, on the implementation of speed of recovery in their prices. In Chile, export value additional structural reforms. increased by 60 percent between March and December 2009, after Peru, which had an 80 percent rise in the same period. Both economies are within the group of countries Transmission Mechanisms of with greater commercial openness in the region, but they the 2008­09 Economic Crisis also have another element in common: the relevance of The 2008­09 crisis was transmitted to Chile through two copper in their exports. Given the speedy recovery shown main mechanisms. The first was an external demand by the price of this metal, it is the main candidate for shock, which is discussed in comparative perspective and explaining the increased value of exports in both countries. The International Crisis and Development Strategies: The Case of Chile 191 Table 15.3. Effect of the International Crisis on the Exports of Selected Countries, as of December 2009 Fall (maximum-minimum) Recovery to Dec. 2009 % months % months exports/GDP Argentina ­49 5 30 11 22 Brazil ­53 7 51 10 13 Chile ­51 11 60 10 39 Colombia ­40 4 40 13 16 Ecuador ­59 9 50 9 34 Mexico ­45 6 51 11 27 Peru ­47 6 80 11 25 Uruguay ­45 6 33 11 19 Average excluding Chile ­48 6 48 11 22 Australia ­33 7 15 11 23 Canada ­45 9 24 8 23 Chile ­51 11 60 10 39 New Zealand ­46 9 43 11 24 Norway ­46 13 24 7 38 South Africa ­57 6 41 11 29 Average excluding Chile ­46 9 29 10 27 Source: Zahler & Co., based on data from Bloomberg (database): New York, http://www.bloomberg.com. In contrast, Argentina, Colombia, and Uruguay--countries External Demand Shock: Effects on Chile that did not see such a significant improvement in the price Considering constant 2003 prices (the base year of the lat- of their exported products during the second half of 2009-- est available national accounts), Chile's ratio of exports of also experienced lower growth in the value of exports. goods and services to GDP practically doubled in 20 years, Figure 15.8 shows the relation in this wide sample of from 23 percent in the mid-1980s to nearly 40 percent in countries between the ratio of exports to GDP in 2008 and the mid-2000s (figure 15.9). Imports experienced an even the GDP rate of change in 2009. There is a negative rela- larger jump, supported by the growth in national income tion: that is, the higher the ratio of exports to GDP, the and the continuous reduction in tariffs due to Chile's free greater tends to be the fall in GDP. Given the linear adjust- trade agreements. In real terms, imports of goods and serv- ment, the drop in Chile's GDP in 2009 could have been ices went from representing 15 percent of GDP in the mid- around 2 percent because of its degree of commercial 1980s to over 50 percent in 2008. In 2009, real exports and openness. imports both decreased as a percentage of GDP. With the exception of Mexico, all other Latin American From 2005 on, in real terms, imports were higher than countries show a better relative performance than Chile, exports, a finding explained mostly by the differences in the with GDP variations close to zero or even positive. Mexico, evolution of export and import prices. Indeed, from 2005 to most likely because of its heavy ties to the U.S. economy, 2007, the increase in prices of imported goods was less than presents a significant GDP contraction, which goes further the increase in prices of exported goods. As from the third than what its degree of openness would suggest. quarter of 2008, the price of imported goods decreased to The higher-income commodity-exporting countries are levels similar to those that prevailed in 2005 and stabilized generally positioned closer to the linear relation. It stands at that level, while the price of exports had a temporary out that Norway, with an openness similar to Chile's, experi- drop but, by late 2009, had recovered considerably. In spite enced a GDP contraction also similar to that of Chile. GDP of everything, in current dollars, not only has the Chilean changes in New Zealand and South Africa were consistent balance of trade maintained a surplus during the past five with their degree of commercial openness, while Australia years, but also that surplus continued to increase until 2007. did relatively better, and Canada, worse. It is possible that the The relevance of external demand to Chile's economic close commercial and financial ties of Canada and Mexico performance has grown significantly throughout the to the United States may have had an influence, given that course of the past decades. The largest increase took place both economies performed more poorly than what their from 1970 to 1990, when exports of goods and services as a overall ratio of exports to GDP would explain. 192 Managing Openness Figure 15.8. Ratio of Exports to GDP in 2008 and GDP Growth in Selected Countries, 2009 4 Uruguay 2 Australia Argentina GDP growth in 2009 (%) Peru Ecuador 0 Brazil Colombia Norway New Zealand South Africa Chile ­2 Canada ­4 ­6 Mexico ­8 10 15 20 25 30 35 40 exports/GDP in 2008 (%) Source: Zahler & Co., based on data from Bloomberg and the IMF: Bloomberg (database), New York, http://www.bloomberg.com; and International Financial Statistics (database), IMF, http://www.imfstatistics.org. Figure 15.9. Chile's Commercial Openness, 1986­2009 55 100 50 90 45 80 40 70 % of GDP % of GDP 35 30 60 25 50 20 40 15 10 30 86 87 88 89 90 91 92 93 94 19 5 19 6 97 98 99 00 01 02 03 20 4 05 06 07 08 09 9 9 0 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 year exports (% GDP) imports (% GDP) exports and imports (% GDP) (right axis) Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. The International Crisis and Development Strategies: The Case of Chile 193 proportion of GDP tripled, from 11 percent in 1970 to the United States were greatly affected (figure 15.13) by 33 percent in 1990. Subsequently, the increase in external the subprime crisis and the lower U.S. demand for hous- demand became more gradual and at present represents ing construction materials. slightly less than 40 percent of GDP (figure 15.10). Table 15.4 shows that in the early 1980s, U.S. participa- Chilean exports grew substantially from 2003 to 2007, tion increased from 22 percent to 28 percent, while Latin with important increases in prices and quantities (figure America's dropped from 19 percent to 12 percent, evidence 15.11). In 2008, prices and quantities stagnated, and in that Latin America was one of the most affected regions 2009, the Chilean economy suffered a negative shock. Dur- during that crisis. During the Asian crisis, no major ing the four quarters preceding the third quarter of 2009, changes occurred in the relevant participation of the vari- export prices fell by 25 percent (year-on-year average), ous economic zones. Also noticeable is the relatively low while quantities exported fell 4.6 percent. These events dif- participation of China, which in 1999 represented only fer considerably from what happened during the 1998 3 percent of total Chilean exports. In the most recent crisis Asian crisis, when prices fell by 15 percent but export (2008­09), the principal change is the extreme increase in quantities increased by 8 percent. China's participation, reaching 23 percent of the total, This difference in the evolution of export quantities more than double that of the United States. during these two episodes may be explained, on the one However, most Chilean exports to China are mining hand, by the much greater magnitude of the 2008­09 products. In 2009, 85 percent of Chilean exports to China shock relative to that of the late 1990s. On the other hand, came from that sector, 14 percent were industrial ship- Japan, Europe, and the United States, whose markets were ments, and only 1 percent was agricultural exports. India most affected by the 2008­09 crisis, played an important is not a major destination for Chilean exports. In 2009, part in Chile's decline in exports, notwithstanding the sig- only 2 percent of Chilean exports (US$1.1 billion) were nificant diversification of Chilean export destinations. shipped to India. Of this amount, 93 percent were mining Diversification of Chilean exports has changed signif- exports, 6 percent were industrial products, and 1 percent icantly since the beginning of the 1980s. Figure 15.12 was agricultural products. shows that the relevance of Europe and Latin America as The global nature of the 2008­09 crisis reduced the destinations of Chilean exports has decreased during the effectiveness of the export diversification strategy as a past 30 years, mainly in favor of Asia. U.S. participation mechanism to "shield" the Chilean economy from external in total exports, which increased from 12.6 percent in shocks. According to estimates of the International Mone- 1980 to around 17 percent in 1990 and 2000, had fallen tary Fund (IMF), 49 percent of the countries for which to 11.3 percent in 2009. In 2007­09, Chilean exports to information was available experienced a fall in GDP during 2009. In 1999, the percentage of countries whose GDP fell Figure 15.10. Chile's External and Internal Demand was 19 percent, and in 1983 it reached 33 percent. for Goods and Services, 1970­2009 Furthermore, even with the high degree of export 100 market diversification, more than 40 percent of Chilean 90 11% shipments go to Europe, Japan, and the United States--the 24% 33% 35% 38% countries most affected in the recent crisis. Nearly 60 per- 80 70 cent of Chilean exports go to the world's four largest eco- nomic zones or countries: China, Europe, Japan, and the % of GDP 60 50 89% United States. Until 2006, shipments to Europe, Japan, and 76% 67% 65% 62% the United States represented over 50 percent of the total, 40 30 and China represented around 10 percent. The 2008­09 cri- 20 sis, however, generated a significant change in the destina- 10 tion of Chilean exports. In fact, 40 percent of the shipments 0 in 2009 went to Europe, Japan, and the United States, while 1970 1980 1990 2000 2009 23 percent went to China, which has become the principal year country destination for Chilean exports. domestic demand (C + I + G ­ M) Chilean exports are much less diversified by product exports (goods and services) than by destination. They are still highly dependent on the Source: Zahler & Co., based on data from the Central Bank of Chile: situation of copper, the main export product. In 2009, Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Note: C = consumption; G = government spending; I = investment; mining exports represented close to 60 percent of total M = imports. Chilean exports, and copper accounted for 90 percent of 194 Managing Openness Figure 15.11. Changes in Prices and Quantities of Chilean Exports, 1997­2009 50 40 30 change (yoy, %) 20 10 0 ­10 ­20 97 98 99 00 01 02 03 04 05 06 07 08 09 19 19 19 20 20 20 20 20 20 20 20 20 20 year quantity price Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Note: yoy = year over year. Figure 15.12. Diversification of Chilean Exports by Destination, 1980­2009 Figure 15.13. Chile's Exports to the United States as a Percentage of Total Exports, 2003­09 100 3.6 4.4 3.7 7.0 20 17.7 18.5 18 80 27.0 30.8 16.1 15.9 16 15.2 45.7 14 12.9 60 12.0 % of total 40.9 12 11.3 percent 26.5 39.1 10 40 19.6 8 12.6 17.3 6 20 17.0 11.3 4 24.4 21.7 12.5 16.4 2 0 0 1980 1990 2000 2009 03 04 05 06 07 08 09 year 20 20 20 20 20 20 20 year Latin America United States Europe Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Asia rest of world Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Table 15.5 also shows that, at an aggregate level, the mining sector exports (table 15.5). Asia remains the main composition of exported goods has not significantly buyer of copper, with 58 percent of the total, followed by changed in the past 25 years. Although Chile's policies on Europe with 22 percent and the Americas (16 percent). exports were neutral during those years, on occasion gov- Asia's demand is dominated by three countries: China (31 ernment officials stated that their objective was to change percent of the total), Japan (11 percent), and the Republic Chile's export structure to one less dependent on natural of Korea (8 percent). The United States is the fourth-largest resources; however, this shift has not occurred. buyer with 7 percent of the total and Brazil the fifth largest, In a comparison of Chile's GDP per capita performance with 6 percent. to that of Australia and New Zealand, Chile had a higher The International Crisis and Development Strategies: The Case of Chile 195 Table 15.4. Chilean Export Diversification by Destination, caused by a decrease in prices, because volumes did not 1982­2009 experience significant changes (figure 15.14). percent Industrial exports, which in 2008 and 2009 represented Asia total one-third of total shipments, are more diversified by prod- 1982 1983 1998 1999 2008 2009 uct than mining exports, as may be seen in figure 15.15. The two most important products are salmon and trout, with United States 22 28 17 19 12 11 Europe 37 39 29 27 26 20 12 percent, and cellulose, with 11 percent. Despite this Asia total 18 17 27 29 36 45 greater diversification, however, industrial exports also suf- Japan 12 9 15 14 10 9 fered a major fall during 2008­09. Figure 15.16 shows that, China 2 2 4 3 14 23 contrary to what happened with mining shipments, prices Latin America 19 12 23 21 16 15 and quantities of industrial exports both suffered a signifi- Other 3 4 3 3 9 8 cant drop as a result of the latest crisis. Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Overall international goods prices faced by Chile Database, Central Bank of Chile. Santiago, Chile, http://www.bcentral.cl. (figure 15.17) indicate that its terms of trade in the recent world crisis began falling in the first quarter of 2008 and, Table 15.5. Chile's Export Products, 1982­2009 after having dropped by close to 30 percent, they reached a percent minimum in the first quarter of 2009. Since then, the Product 1982 1983 1998 1999 2008 2009 increase in the terms of trade was quite fast, by the end of Agriculture, fruit, 2009 recovering practically the entire fall experienced livestock, forestry, and fish 10.1 8.5 11.4 10.8 6.2 6.9 during the previous quarters. Minerals 58.1 59.9 41.3 42.9 60.4 59.5 Chile's exports of services and transport, including air Copper 46.7 47.9 33.7 36.4 51.8 53.5 and sea as well as freight and passengers, accounted for Industry 31.8 31.6 46.8 45.6 33.5 33.6 58 percent of the total in 2003­09. Other important services Other 0.0 0.0 0.5 0.6 0.0 0.0 exported by Chile were travel and tourism and business Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics services; the former were 17 percent and the latter 16 per- Database, Central Bank of Chile, Santiago, Chile, http://www.bcentral.cl. cent of total export services in 2003­09. Chilean exports of services (in value terms) grew at a high rate (16.4 percent) GDP growth rate than either Australia or New Zealand, in 2004­08, a trend that was abruptly interrupted in 2009, after Latin America's lost decade of the 1980s. Although when they fell 21 percent as a consequence of the crisis this outcome is the result of many factors, it may indicate (figure 15.18). that the Chilean development strategy of the past 25 years In recent years, growth in services trade in China, India, has been quite successful. Despite that success, however, the and the Russian Federation was at more than double the same strategy should not necessarily be pursued in the world average rate. Latin America's growth in services has same manner in the future. In fact, Chile's GDP per capita been below the average, and therefore the region's participa- in 2009 was similar to that of Australia in 1985 and of New tion in world trade in services has fallen in recent years. Zealand in the mid-1990s; therefore, Chile should seriously China and India and, to a certain extent, Brazil and Mexico consider development strategies followed by Australia and can rely on an abundant and low-wage labor force, a New Zealand, countries that have changed their export resource not available in Chile, a country that faces other structure quite significantly. If implemented, such policies difficulties in substantially increasing its exports of services: may lead to a change in Chile's productive structure, allow- the lack of knowledge of English and the scarce, although ing more rapid growth of industrial goods production, increasing, technically skilled labor force. Chile's macroeco- services, and technology. nomic and institutional stability, however, is an important The recent international crisis caused a tremendous asset that could favor the development of service exports. drop in the price of copper, as well as in the value of total exports. In effect, in April 2008 the price of copper reached International Capital Markets Shock: Effect on Chile a maximum of US$3.90 per pound, and in December of the same year, it fell to US$1.40 per pound, a deterioration Chile's positioning as an attractive country for FDI is due of 65 percent. Sixty percent of this fall took place after to its relative openness to international capital flows, its September 2008, the month when Lehman Brothers went political and economical stability, its high connectivity, its bankrupt. The greatest part of the fall in the value of min- wide network of international agreements, and the legal ing exports during the first three quarters of 2009 was certainty and stability it offers. These factors have given rise 196 Managing Openness Figure 15.14. Changes in Price and Quantity of Chilean Mining Exports, 1997­2009, as of Third Quarter 2009 80 60 40 change (%, yoy) 20 0 ­20 ­40 97 98 99 00 01 02 03 04 05 06 07 08 09 19 19 19 20 20 20 20 20 20 20 20 20 20 year quantity price Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Figure 15.15. Chilean Industrial Exports, 2009 uncertainty and volatility (see figure 15.19). In 2009, both FDI and portfolio investment fell; however, the magnitude 8% 2% of this reduction was not significant and appears secondary in comparison with the shock to export prices and quanti- 6% ties exported described above. In 2009, FDI into Chile amounted to around US$12 37% billion, equivalent to 7.6 percent of GDP, lower than the previous year's figure of US$17 billion but much higher 14% than the average of 1996­2007 (around US$6 billion annu- ally). Portfolio investment exceeded US$2 billion, a lower figure than that for 2008 but much higher than the flows of 1996­2007 (annual average of US$1.2 billion). Although FDI increased significantly in 2007­09, U.S. participation in those flows was relatively low. In fact, U.S. FDI in Chile has fallen systematically over time. In 15% 1974­84, it reached 50 percent of total FDI in Chile. From 9% 1985 to 1995, U.S. FDI fell to 36 percent, falling further to 9% around 21 percent in 1996­2003. From 2004 to 2008, FDI food alcohol and tobacco originating in the United States represented, on average, around 9 percent of total FDI in Chile. forestry and wooden furniture cellulose, paper, and others chemical products basic metal industries The major part (in absolute terms) of U.S. FDI took metal products and machinery others place from the mid-1990s to the early 2000s. Due to the high economic integration among the signatories to the Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics North American Free Trade Agreement (NAFTA), any eval- Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. uation of a shock in the United States must also take into account the potential impact on investment flows originat- to foreign companies' interest in investing in Chile, partic- ing in Canada and Mexico. FDI from Mexico into Chile is ularly during times of economic turbulence. Thus, during still minor; however, that from Canada has been quite 1999 and 2008, FDI increased, perhaps the result of Chile's steady since the early 1990s, totaling around US$2 billion attractiveness as a shelter during times of greater regional in 2006 and 2008 (figure 15.20). The International Crisis and Development Strategies: The Case of Chile 197 Figure 15.16. Change in Price and Quantity of Chilean Industrial Exports, 1997­2009 20 15 10 change (yoy, %) 5 0 ­5 ­10 ­15 ­20 97 98 99 00 01 02 03 04 05 06 07 08 09 19 19 19 20 20 20 20 20 20 20 20 20 20 year quantity price Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Note: As of third quarter 2009. yoy = year over year. Figure 15.17. Chile's Terms of Trade, 2003­09 200 180 export price/import price (%) 160 140 120 100 80 60 40 20 0 3 4 5 6 7 8 9 00 00 00 00 00 00 00 .2 .2 .2 .2 .2 .2 .2 ar ar ar ar ar ar ar M M M M M M M year Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. China and India are not major sources of FDI into As for portfolio investment, most of Chile's sovereign Chile. In 1974­2008, FDI from India amounted to only wealth fund investments are registered, for balance-of- US$27 million, equivalent to 0.04 percent of Chile's total. payments purposes, as portfolio investments. During 2009, Similarly, FDI from China from 1974 to 2008 totaled the balance-of-payments financial account recorded an US$84 million, or 0.12 percent of the total. income of US$7.95 billion from the central government, 198 Managing Openness Figure 15.18. Value of Chilean Exports of Services and Percentage Change, 2003­09 US$, millions 12,000 25 20 10,000 15 10 8,000 US$, millions 5 % change 6,000 0 ­5 4,000 ­10 ­15 2,000 ­20 0 ­25 03 04 05 06 07 08 09 20 20 20 20 20 20 20 year transportation travel communication insurance other business services other change yoy (%), right axis Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Figure 15.19. Foreign Direct Investment and Portfolio Investment in Chile, 1996­2009 18,000 16,000 14,000 12,000 US$, millions 10,000 8,000 6,000 4,000 2,000 0 ­2,000 96 97 8 99 0 01 02 03 04 05 06 07 8 09 9 0 0 19 19 19 19 20 20 20 20 20 20 20 20 20 20 year FDI portfolio investment Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. as a result of the repatriation of foreign exchange to that was reduced to less than half in 2009. In their turn, finance the countercyclical fiscal spending of that year. As pension funds remitted abroad US$16.4 billion in 2009, a for the other institutional players, in 2008 commercial figure that is far above the average of the past six years banks repatriated US$6.3 billion from abroad, a figure (US$2.6 billion). The International Crisis and Development Strategies: The Case of Chile 199 Figure 15.20. NAFTA Foreign Direct Investment in Chile, 1974­2008 3,000 2,500 2,000 US$, millions 1,500 1,000 500 0 4 86 88 90 92 94 96 98 00 02 04 06 08 ­8 19 19 19 19 19 19 19 20 20 20 20 20 74 19 year United States Mexico Canada Source: Zahler & Co., based on data from Foreign Investment Committee: Government of Chile, Santiago, http://www.foreigninvestment.cl. In brief, neither FDI nor portfolio flows into Chile were crisis on the terms of trade and the decrease in expected seriously affected during the 2008­09 crisis; in particular, the GDP growth of the country's principal commercial part- Chilean economy did not experience a reversal of capital ners became clear. In April 2009, consensus was already flows. It did, however, face an increase in the cost of foreign considering the possibility of a fall in GDP, and by mid- financing (figure 15.21). The three-month LIBOR (London year projections stabilized around the figure by which GDP interbank offered rate) in U.S. dollars rose over 100 basis finally fell: 1.5 percent. points from August to October 2008; in addition, there was a The deterioration of credit terms, the loss of confi- significant adjustment suffered by credit default swaps dence felt by consumers and entrepreneurs, and the (CDS) for Chile's sovereign debt. If the two effects are added, increase in unemployment had to be added to the external the cost of a company's borrowing in international markets demand shock faced by the Chilean economy--all leading increased over 200 basis points in just a few weeks. to a contraction in domestic demand. Firms lowered their inventories quickly and significantly, and families reduced their purchases of consumer goods, particularly durables Overall Impact of the 2008­09 (figure 15.23). Economic Crisis on Chile Banks began to restrict their loan offers as far back as From 2003 to 2007, Chile's GDP grew at an average rate of 2007. According to a quarterly survey prepared by the Cen- 5 percent. This trend was interrupted by the 2008­09 inter- tral Bank of Chile, credit supply conditions started to national crisis. Chile grew 3.7 percent in 2008, but during become more restrictive for consumer loans in the second the first three quarters GDP growth was 4.7 percent. In the quarter of 2007, while for companies they began to become fourth quarter of 2008, GDP grew only 0.7 percent and, in more restrictive in the third quarter of 2007. The fall in terms of quarter-on-quarter annualized change, GDP fell credit demand came later, beginning in the first quarter of by 9.5 percent, the worst figure in 2008 and 2009 (table 2008 with respect to consumer credit and in the second 15.6). GDP also contracted in 2009, but, as noted, the mag- quarter of 2008 in the case of companies. The global liq- nitude of the fall was considerably less than contractions uidity crisis, however, raised the cost of external financing in earlier episodes of world recession. significantly, although relatively briefly, as noted. In mid-2008, figure 15.22 shows that market consensus Consumer confidence began to deteriorate in early expected the Chilean economy to grow by 4.5 percent in 2008, a pessimistic outlook that lasted until September 2009, a projection that was corrected downward during the 2009 (figure 15.24). Likewise, entrepreneurial confidence fourth quarter of 2008, when the impact of the international also fell in June 2008 and remained low until August 200 Managing Openness Figure 15.21. Three-Month LIBOR and Five-Year CDS for Chile, 2008­09 8 7 6 5 percentage 4 3 2 1 0 08 8 09 10 00 20 20 20 .2 n. ay n. pt Ja Ja M Se year LIBOR (3 months) CDS (5 years) Source: Bloomberg (database): New York, http://www.bloomberg.com. Note: CDS = credit default swaps, LIBOR = London interbank offered rate. Table 15.6. Percentage Change in Chilean GDP, Year on Year, First Quarter 2008­Fourth Quarter 2009 1Q 2008 2Q 2008 3Q 2008 4Q 2008 1Q 2009 2Q 2009 3Q 2009 4Q 2009 GDP (yoy) 3.7 5.1 5.2 0.7 ­2.1 ­4.5 ­1.4 2.1 qoq SAAR 8.0 8.5 ­3.2 ­9.5 ­3.0 ­0.5 6.6 5.9 Internal demand 8.2 11.7 10.8 0.2 ­6.6 ­9.9 ­8.1 1.4 Private consumption 5.5 6.2 5.7 1.3 ­1.0 ­2.0 0.8 5.5 Investment 13.9 22.7 29.0 9.8 ­9.9 ­19.4 ­19.1 ­11.9 Government consumption 2.1 ­0.1 ­0.3 0.5 7.2 7.8 5.8 6.4 Exports 2.3 0.0 7.2 3.4 ­4.9 ­7.2 ­6.7 ­3.7 Imports 12.8 16.1 19.4 1.5 ­14.5 ­18.9 ­19.2 ­4.0 Investment/GDP 26.4 29.1 31.1 30.1 24.3 24.6 25.5 26.0 Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Note: Q = quarter, qoq SAAR = quarter-on-quarter seasonally adjusted annualized rate, yoy = year on year. 2009. The unemployment rate rose from 7.4 percent in foreign currency (mainly U.S. dollars) liquidity and then September 2008 to 10.1 percent in August 2009; in the counteracting the fall in foreign demand by means of same period, the number of unemployed increased by greater public spending, a significant reduction in interest 228,000. rates, state guarantees of access to credit, tax incentives, and The terms of trade fell 28 percent in the fourth quarter incentives to hire workers, among others. These policy of 2008, and the drop in the price of exports between the measures helped avoid a greater fall in GDP, which finally maximum and minimum in 2009 was one of the highest in contracted 1.5 percent in 2009. Latin America. In 2009, exports, in real terms, fell 5.6 per- cent and gross capital formation, which in 2008 had Chilean Policy Reaction to the 2008­09 Crisis increased by 18.8 percent, fell by 15.3 percent in 2009. Eco- nomic authorities--monetary and fiscal--acted decisively Chilean policy makers responded to the 2008­09 crisis on and in coordination, implementing a set of measures that a number of levels. This section discusses the implications were among the largest in the world. In the first instance, of the crisis for monetary policy, the exchange rate, and these measures aimed at improving the availability of fiscal policy. The International Crisis and Development Strategies: The Case of Chile 201 Figure 15.22. Forecasts of Chile's GDP Growth for 2009, June 2008­December 2009 5 4.5 4.2 4.2 4.2 4 3.5 3 2.5 GDP growth (%) 2.0 2 1.5 1.2 1 0.2 0 ­0.5 ­1 ­0.7 ­1.0 ­1.5 ­1.5 ­1.5 ­1.5 ­1.5 ­2 ­1.8 8 8 8 9 9 9 9 00 00 00 00 00 00 00 -2 r-2 r-2 -2 -2 r-2 r-2 ne ch ne be be be be ar Ju Ju em em em em M pt ec pt ec Se Se D D month Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Figure 15.23. Contribution of the Components of Aggregate Demand to Chile's GDP Growth, First Quarter 2008­Fourth Quarter 2009 20 5.2 15 5.1 3.7 10 0.7 contribution (%) 5 2.1 0 ­1.4 ­5 ­10 ­2.1 ­15 ­4.5 ­20 08 08 08 08 09 09 09 09 20 20 20 20 20 20 20 20 1- 2- 3­ 4- 1- 2- 3- 4- Q Q Q Q Q Q Q Q quarter private consumption government consumption fixed investment inventories exports imports GDP Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Note: Q = quarter. 202 Managing Openness Figure 15.24. Index of Consumer Confidence in Chile, October 2008­December 2009 60 56 55 54 52 53 50 49 index value 45 45 40 41 40 39 37 36 36 36 35 35 35 34 30 8 8 09 09 9 9 9 9 00 00 00 00 00 00 20 20 .2 .2 2 .2 .2 .2 b. r. n. ct ec g ct ec Ap Ju Fe Au O O D D month index threshold Source: Zahler & Co., based on data from Adimark, Santiago, http://www.adimark.cl. Monetary Policy Chile. This development put an end to the policy, just when the central bank had accumulated nearly US$6 billion, put- Internal demand grew quite rapidly between 2004 and ting it in a comfortable position for mitigating a possible 2007. This growth was not entirely reflected in a higher sudden stop in net capital inflows into Chile. GDP growth rate, because part of this demand went to the In the fourth quarter of 2008, the impact of the interna- rest of the world through higher imports. During this tional crisis was added to the decelerating phase of the period, internal demand grew an average of 8.1 percent domestic cycle. The Central Bank of Chile then reversed annually, 2.9 percentage points higher than GDP growth. the monetary policy contraction. During the most critical Private consumption grew by 7.2 percent annually, and stage of the crisis, the first policy reaction by the central investment grew 12.1 percent, more than doubling the GDP bank was to establish extraordinary measures for liquidity growth rate. This trend had to be slowed, a process initiated provision, both in national currency and in U.S. dollars. This in 2007, by increasing the Central Bank of Chile's monetary action ensured the normal operation of the financial system, policy interest rate. This policy continued during 2008. notwithstanding the great increase in risk perception. In In addition, during the first part of 2008 the Chilean addition, motivated by local dollar liquidity restrictions and peso experienced an unexpected and clearly excessive the shortage of foreign credit lines for banking institutions, appreciation, at a moment when there were clear signs of a the Treasury began to place resources in dollars in the local deteriorating international situation. In this context, the financial system--resources originating from one of the sov- central bank decided to strengthen its international ereign wealth funds--in the amount of US$1.05 billion-- reserves by accumulating US$8 billion during an eight- while the central bank started to provide liquidity in local month period. To preserve monetary policy independence currency and in dollars through repurchase (repo) and swap from the exchange-rate policy, the central bank announced operations with banking institutions. Furthermore, the a predetermined and mechanical purchase of US$50 mil- Treasury made a one-off increase in U.S. dollar deposits for lion per day. By mid-September 2008, liquidity tensions US$700 million in the domestic banking system. and constraints became severe worldwide, including in The International Crisis and Development Strategies: The Case of Chile 203 The first central bank policy measures to mitigate the These monetary policy actions were decisive in reducing impact of the crisis were taken in October 2008: the interest rates faced by the financial system's clients. In addition, the government capitalized Banco Estado, a · Auctioned US$500 million for a term of 28 days, under commercial 100 percent government-owned bank in the the modality of foreign exchange swaps amount of US$500 million. Banco Estado played a very · Extended the swap purchase of dollars from one to six important de facto and nontraditional monetary policy months for a maximum amount of US$5 billion role (the Chilean version of quantitative easing) by expand- · Opened the offer of cashier repo transactions for a term ing its credit in very significant amounts while putting of 28 days downward pressure on interest rates charged by the private · Offered repo transactions at 60 and 90 days' term and banking system. In fact, Banco Estado's participation in broadened eligible collateral total bank credit increased from 13.4 percent at the end of · Temporarily relaxed the norms for requirements of 2008 to 15.8 percent at the end of 2009. bank reserves. In late 2009, the Central Bank of Chile announced the scheduled and gradual withdrawal of the complementary In January 2009, there was a clear assessment of reduced measures introduced earlier to strengthen monetary policy. inflationary pressures: consumer price index inflation fell Specifically, the liquidity facility would be reduced from from 9.9 percent in October 2008 to 6.3 percent in January 180 days to 150 days from December 2009. Subsequently, 2009, and inflation expectations for December 2009 fell in the term would continue to be reduced monthly by 30 January 2009 to 3.5 percent from 4.9 percent in September days, bringing access to this facility to an end in May 2010. 2008. Consequently, the central bank lowered the mone- tary policy interest rate by 100 basis points. Later, with evi- Exchange-Rate Policy dence of further alleviated inflationary pressures and drastic increases in downward risks to growth and infla- Chile has a floating exchange regime, although the central tion, the central bank decided to cut the rate by 250 basis bank reserves the right to intervene under exceptional cir- points at both its meetings in February and March. Thus, cumstances, when it sees a significant misalignment in the in the first quarter of 2008 the rate was cut by 600 basis real exchange rate. This exchange-rate policy--in combi- points, from 8.25 percent to 2.25 percent. The monetary nation with inflation targeting and a prudent and counter- easing continued, bringing the rate down by 775 basis cyclical fiscal policy--contributes to stability, allowing for points in seven monthly meetings and lowering the mone- changes in the real exchange rate, which, in the face of tary policy interest rate to 0.50 percent, the minimum unexpected shocks, helps minimize effects on production deemed adequate for a normal functioning of money and employment. All three crises in the past 35 years markets and the lowest of any Latin American economy (1981­82, 1998­99, and 2008­09) have coincided with a (figure 15.25). The central bank stated that it would keep previous appreciation of the Chilean peso until it reached a the rate constant for a prolonged period, at least until the critical level (figure 15.26). second quarter of 2010. The crisis with the largest impact on Chile's GDP and To reinforce this decision and align financial asset prices unemployment rate (1981­82) coincided with the most with the path of monetary policy, the central bank imple- significant peso appreciation, when the real exchange rate mented complementary monetary policy measures: fell to an all-time low of 50 (base 1986 = 100; average real exchange rate from 1986 to 2010 = 96). Before the 1998­99 · It established a term liquidity facility (Facilidad de Liq- crisis, the real exchange rate also fell, although from a peak uidez a Plazo, or FLAP) for banking institutions, whereby of 115 in the early 1990s to less than 80. In April 2008, as it granted 90- and 180-day liquidity at the prevailing noted, the central bank began to accumulate international monetary policy interest rate. reserves, in part because of the appreciation of the Chilean · It adjusted its program of note issuance (Pagarés peso (real exchange rate of 85 in March 2008). Descontables del Banco Central, or PDBC bonds) at maturities below one year. Fiscal Policy · It suspended the issuance of debt instruments maturing in or after one year, corresponding to two-year central Chile's institutions of fiscal policy have constantly bank peso-denominated bonds (BCP-2) and one-year improved. Already noted, for example, were the adoption of central bank notes (PDBC-360). a rule targeting a structural budget surplus as a percentage 204 Managing Openness Figure 15.25. Monetary Policy Interest Rates in Selected Latin American Countries, January 2007­March 2010 16 14 12 interest rate (%) 10 8 6 4 2 0 07 07 07 07 07 08 08 08 08 08 09 09 09 09 09 10 10 7 8 9 00 00 00 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 /2 /2 /2 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ 2/ /2 /2 /2 1/ 3/ 5/ 7/ 9/ 1/ 3/ 5/ 7/ 9/ 1/ 3/ 5/ 7/ 9/ 1/ 3/ 11 11 11 month Brazil Chile Colombia Mexico Peru Source: Zahler & Co., based on Bloomberg data: Bloomberg (database), New York, http://www.bloomberg.com. Figure 15.26. Chile's Real Exchange Rate Index, 1976­2008 1986 = 100 Apr. 82 Jul. 98 Apr. 08 120 110 100 90 index 80 70 60 50 40 Dec. 76 Dec. 80 Dec. 84 Dec. 88 Dec. 92 Dec. 96 Dec. 00 Dec. 04 Dec. 08 year RER index average Jan. 1986­Feb. 2010 Source: Zahler & Co., based on data from the Central Bank of Chile: Statistics Database, Central Bank of Chile, Santiago, http://www.bcentral.cl. Note: RER = real exchange rate. The International Crisis and Development Strategies: The Case of Chile 205 of GDP (2002) and the mandate that the surplus be equiv- contributions to Corporación Nacional del Cobre (Codelco) alent to 1 percent of GDP (which was reduced to 0.5 per- and Corporación de Fomento de la Producción (Corfo). cent in the fiscal budget for 2008 and subsequently the ex The plan's financing originated from the resources of ante target was reduced to 0 percent of GDP in 2009 the Economic and Social Stabilization Fund and the because of the crisis). Such a budget surplus allows fiscal issuance of bonds authorized by the 2009 Budget Law. Of policy to be carried out in a countercyclical manner. This the US$4 billion, it was announced that US$3 billion rule was formalized in the second semester of 2006, when would be assigned to finance spending and investments in the Fiscal Responsibility Law was enacted together with the Chilean pesos, while the remaining US$1 billion would creation of the Pension Reserve Fund and the Economic be used to finance spending and investments in dollars. In and Social Stabilization Fund. the case of the US$3 billion, the Treasury requested that the The structural budget surplus rule was introduced to central bank, acting as fiscal agent, sell foreign exchange specify an annual level of fiscal spending in accordance through a system of competitive bidding, consisting of with the structural (permanent) income of the central gov- daily sales of US$50 million. ernment. Therefore, the level of public spending rules out As a result of the lower tax revenue triggered by the 2009 the cyclical fluctuations of economic activity, the copper recession and the greater public spending to counteract the price, and other factors that determine actual fiscal contraction in private demand, government spending in income. This policy implies saving during the up phase of 2009 increased by 17.8 percent in real terms--three times the economic cycle and allows avoiding drastic adjust- the average increase registered from 2000 to 2008--and the ments of fiscal spending in the downturn of the economic fiscal deficit increased to 4.5 percent of GDP. This increase cycle, such as the one that took place in late 2008 and part in spending implied a larger fiscal stimulus in Chile than in of 2009. other Latin American countries (figure 15.28). As a result of the very high prices of copper in the three- year period 2006­08, these funds accumulated more than Conclusions US$22 billion. During 2009, the Economic and Social Sta- bilization Fund contributed to financing the incentive Chile's sound fundamentals and strong macroeconomic plans and the fiscal deficit, initiatives that allowed Chile to management provided a buffer against the global eco- face the financial and economic world turbulence. The nomic recession, which nevertheless hit the country value of the Pension Reserve Fund as of December 31, through a sharp deterioration in its terms of trade and the 2009, was US$3.4 billion, while the value of the Economic fall in external demand for Chilean exports. There was and Social Stabilization Fund reached US$11.3 billion room for significant macroeconomic stimulus, thanks to (figure 15.27). prudent fiscal policy during the boom years. This provi- To mitigate the huge fall in external demand and in sion, together with the rebound in copper prices and the internal demand for capital goods, as well as the expected revival of global trade, contributed to a turnaround in - fall in private consumption, the government launched a activity in 2010. Although the economy is coming out of fiscal incentive plan in 2009 committing more than recession, unemployment is projected to remain initially US$4 billion, equivalent to 2.8 percent of GDP. The fiscal high, and inflation is likely to stay low. incentive package contemplated providing support to indi- From an intertemporal perspective, Chile was much viduals and families, boosting public investment, reducing more integrated into the world economy, both in trade and taxes, giving other incentives to private investment, and in finance, in 2008­09 than in the mid-1970s and in the strengthening the access to financing for small and medium early 1980s. However, when faced with the 2008­09 inter- enterprises, as well as measures to protect employment, national crisis, Chile performed much better than in those among other initiatives. previous crises. Indeed, in comparison with the very signif- The plan would increase public spending in 2009 by icant international crises of the mid-1970s and early 1980s close to 1 percent of GDP (US$1.485 billion), and con- (both much smaller than the 2008­09 crisis), the decline in templated the temporary reduction of the structural Chilean GDP was significantly (10 times) lower; similarly, surplus target of 0.5 percent to 0 percent of GDP in 2009. the highest unemployment rate registered in those two The plan also considered temporarily reducing fiscal previous crises was 1.5 and 2 times bigger, respectively, than income by US$1.455 billion, or the equivalent of 1 percent the 10 percent registered in 2009. And although in 2009 the of GDP, which does not affect the fiscal structural income. It decline in Chilean GDP (­1.5 percent) was bigger than the also included disbursements that do not constitute spending 0.7 percent drop in 1999 (Asian crisis), the size of the 2009 but rather the acquisition of financial assets, such as capital international shock was also much greater than in 1999. 206 Managing Openness Figure 15.27. Market Value of Chile's Sovereign Wealth Fund, 2006­09 25,000 20,211 20,000 US$, millions 15,000 14,033 11,285 10,000 5,000 3,421 2,507 1,466 605 0 2006 2007 2008 2009 year pension reserve fund economic and social stabilization fund Source: Statistical Database, Ministry of Finance, Santiago, http://www.hacienda.cl/english/estadisticas/. Figure 15.28. Chile's Fiscal Balance, 2003­09 10 8.8 8 7.7 6 5.3 4.6 4 % of GDP 2.1 2 0 ­0.5 ­0.6 ­2 ­1.1 ­1.3 ­1.2 ­2.0 ­2.9 ­2.6 ­4 ­4.5 ­6 2003 2004 2005 2006 2007 2008 2009 year Latin America Chile Source: Zahler & Co., based on data from the United Nations Economic Commission for Latin America and the Caribbean (ECLAC) and the Ministry of Finance: CEPALSTAT (database), ECLAC, Santiago, http://websie.eclac.cl/infest/ajax/cepalstat.asp?carpeta=estadisticas&idioma=i; and Statistical Database, Ministry of Finance, Santiago, http://www.hacienda.cl/english/estadisticas/. A macroeconomic framework characterized by a prudent mid-1970s and the early 1980s. More specifically, the accu- fiscal policy, a monetary policy of inflation targeting and mulation of international reserves, both by the central a floating exchange rate, and a solvent and well-regulated bank and by the government, proved to be a good insur- financial system allowed the implementation of strong ance policy against a major foreign shock. In fact, Chile's countercyclical fiscal and monetary policies in 2009. These accumulation of significant fiscal surpluses during the policies contributed to substantially improving the per- boom years of the world economy and high copper prices formance of the Chilean economy in comparison with the (2003­07) allowed the country to engage in self-financed The International Crisis and Development Strategies: The Case of Chile 207 stimulus without jeopardizing the policies or the authori- Although the medium-term macroeconomic policy ties' credibility. In short, the most recent world trade framework has worked well and the Chilean economy ben- contraction hit Chilean external demand, but, thanks to efited from its implementation by successive governments, countercyclical monetary and fiscal policies in 2009, inter- there is room for further improvement. Chile could con- nal demand turned out to be quite resilient, although not sider complementing its fiscal rule with a ceiling on growth enough to lift global demand. This macroeconomic frame- in government spending. During booms, such ceilings help work was not in place during the two prior crises, which accumulate additional funds, which can then be used required strong domestic adjustment through major fiscal countercyclically in sharp downturns. The financial system contractions and increases in interest rates, together with is generally well regulated, but a consolidated approach to abrupt and significant devaluations, all of which amplified financial conglomerates and stronger supervision of non- the effects of those two previous world crises on the bank financial institutions would be useful. Additional Chilean economy. strengthening of the insurance element of the unemploy- In contrast, from a cross-section perspective, the perform- ment benefit system, in combination with lowering sever- ance of the Chilean economy in 2008­09 was not better ance pay, would provide more effective protection for the than that of several other Latin American countries, despite unemployed and would contribute to greater flexibility in the magnitude of Chile's expansionary policies. In fact, the the labor market. Although the valued-added content of fall in Chile's GDP in 2009 was surpassed only by Mexico-- exports and product diversification are still lagging, this lag a country with a high dependence on the U.S. economy-- may be explained by Chile's development stage, in that it and the Républica Bolivariana de Venezuela. Other countries, can still benefit from comparative advantages based on such as Argentina, Brazil, Colombia, Ecuador, and Peru saw natural resource abundance and their "light" industrializa- their economies growing marginally in 2009 or experienc- tion, while the country deals with--and perhaps solves-- ing a smaller drop in GDP than the Chilean one. It is possi- its major challenge: increasing the growth of total factor ble that, in this comparison, the Chilean economy was more productivity. affected because of its very high degree of trade and finan- Although Chile's per capita income on the basis of pur- cial openness to the international economy. chasing power parity has increased markedly over the past However, when the crisis became acute, the conditions two decades, TFP has stagnated in the past decade. In addi- faced by Chile were quite different from the other countries tion, although poverty has been reduced quite significantly, considered. In 2008 and 2009, Chilean industrial activity income distribution remains extremely skewed, and many suffered two major shocks not related to the crisis sectors of the population do not feel integrated into the (methanol production and exports fell significantly country's development and modernization process. because of lack of gas in the south of Chile); and salmon Income inequality, as measured by the Gini coefficient, has output and exports also fell drastically because of inade- not declined much over the past 20 years and remains very quate sanitary conditions and lack of appropriate sector high by standards of the Organisation for Economic Co- regulation. Before the crisis erupted, Chile's GDP growth operation and Development (OECD). Sustained growth rate had been trending downward from 2004 (6 percent) to will need to be accompanied by the right social policies to 2007 (3.7 percent), and the 2009 figure of 1.5 percent reduce poverty further and improve income distribution. implies that, with the single exception of Colombia, Chile The OECD, Chilean think tanks, and academics generally had, among the countries considered, the smallest drop in agree that TFP growth should be dealt with by microeco- GDP growth rate between 2009 and 2008. In the same vein, nomic policies and additional reforms, aimed at fostering Chile's average rate of inflation in 2008 (8.7 percent) was competition, entrepreneurship, and innovation; improving the highest, with the exception of the Républica Bolivari- the quality of education; and modernizing labor market ana de Venezuela, of all Latin American countries consid- practices and regulation. ered, requiring a stronger price stabilization policy (that Facilitating entrepreneurship could have beneficial effects is, higher interest rates), which inevitably had a negative on productivity and economic dynamism. For this purpose, effect on Chile's GDP growth rate in 2009. In fact, in 2009 the regulatory red tape that burdens start-ups should be Chile's average rate of inflation fell eight percentage points reduced, and bankruptcy procedures could be further sim- from 2008, by far the largest drop in the sample considered. plified. Until recently, the innovation policy framework Furthermore, Chile's current account improved by 3.7 per- focused on basic research in public institutes and universi- cent of GDP in 2009 over 2008, the second-highest ties. As a consequence, the private business sector's propen- improvement in the sample considered and very similar to sity to engage in technological (product or process) and the 3.9 percent registered by Peru. nontechnological (marketing or organizational) innovation 208 Managing Openness is still low. Several measures have been taken to improve of total world consumption, in the past 10 years, from less industry-science relationships to make scientific research than 40 percent in 1999 to nearly 55 percent in 2008, and more market relevant. Efforts to strengthen links between this trend is likely to continue. The relative scarcity of nat- companies and universities should be continued. ural resources, made more acute by rapid urbanization, has The quality and equity of primary and secondary edu- made Asia an important importer of commodities and cop- cation have to be improved. Notwithstanding impressive per. Furthermore, most of Asia and Chile have complemen- progress in school enrollment, much remains to be done if tary roles in endowments, which are increasingly reflected Chilean children are to reach OECD standards in learning in growing trade relations. Therefore, and in spite of high outcomes. Better-qualified teachers and improved initial transport and communications costs, long delivery time, teacher education and training are key to this purpose. very different time zones, and no direct air links between As mentioned, rigidities in the labor market--relatively Chile and Asia--given the high complementarity between high minimum wages and wage indexation--tend to impede China's and Chile's export structure and the number of the response of wages to an increase in unemployment and free trade agreements signed with Asian countries--Chile's to negative external shocks. In addition, severance pay thus far rather neutral outward-oriented development rights are often several times higher than unemployment strategy appears to be validated. However, given that Chile's benefits, contributing to slower adjustment of the economy 2009 per capita GDP was similar to that of Australia in after adverse shocks. Unemployment benefits rest largely 1985 and of New Zealand in the mid-1990s, Chile should on individual savings accounts with small supplements seriously consider development strategies followed by those from an insurance fund, the Fondo Solidario, which had two countries, which changed their export structure quite very restrictive access until recently. Increasing replace- significantly. Such policies could lead to a change in Chile's ment rates for all workers and extending duration, while productive structure, allowing more rapid growth in pro- simultaneously restricting severance pay, would provide duction of industrial goods, services, and technology. more effective workers' protection and would allow for Chile will go on benefiting from Asia, especially from more efficient job search, thus maximizing workers' pro- China's increasing demand for copper and cellulose and ductivity. If these reforms could be put in place, Chile not eventually for other Chilean exportable goods based on only would likely be better prepared to face international raw materials. This likely outcome should act as an incen- shocks but also could enter another development stage, tive for further inward FDI in sectors related to raw materi- where exports have a higher degree of value added as well als. In this regard, China, which wants to ensure its access as greater diversification of goods and services. to raw materials and guarantee future supplies, could Regarding Chile's development strategy, expected trends extend to Chile its recent policy of funding investment in the international economy tend to support the overall projects in other regions, because Chile is a country with a approach of the past 25 years. The most likely scenario long-standing policy of openness and nondiscrimination after the 2008­09 crisis is that Asia will be the most against FDI. Chinese and Indian FDI into Chile is extremely dynamic region in the world economy. Asia has been small, and there is huge opportunity to increase it and help steadily increasing its consumption of copper, as a share boost trade with Asia. 16 The International Crisis and Development Strategies: The Case of Malaysia Mahani Zainal Abidin Because it is a small and very open economy, where total This chapter will discuss the policy options available to trade is twice the size of its gross domestic product (GDP), Malaysia for producing sustained high growth based on its Malaysia is very vulnerable to international economic extensive global links and strong dependency on interna- events. All three crises that Malaysia has experienced in the tional trade and on financial and investment flows. It will past 25 years were triggered by external factors. The 1985 focus on two issues in Malaysia's outward-oriented devel- crisis, in which GDP contracted by 1 percent, was due to opment strategy: first, its potential for producing sustained the collapse of commodity prices. The 1998 crisis, with a high growth and, second, its ability to cope with global fall in GDP of 7.4 percent, was caused by regional conta- volatility. The chapter begins by analyzing the state of the gion from massive short-term capital outflows. Finally, the Malaysian economy before the 2008­09 crisis and the main 2009 crisis, in which GDP fell by 1.7 percent, was the result differences between the present crisis and the other crises of an export slump. that Malaysia has faced in the past. The impact of the pres- Malaysia's economic growth has long been dependent ent crisis will then be discussed in the second part, followed on exports, starting with primary commodities (rubber by Malaysia's economic recovery. The chapter concludes and tin) in the 1950s and now including manufacturing with an analysis of options for Malaysia's outward-oriented exports (in particular electrical goods and electronics), development strategy. palm oil, petroleum, and tourism. It is not surprising that international developments and global volatility are major State of the Malaysian Economy concerns, because these factors can have a significant impact before the Global Crisis on Malaysia's short-term economic performance as well as on its medium- and long-term development strategies. The The 1998 Asian financial crisis was a seminal experience for 2008­09 global crisis raises a fundamental question about Malaysia. It provided the incentive for building up addi- Malaysia's development strategy, in particular, its depend- tional strength to mitigate the effects of future crises. One ence on manufacturing exports. This issue has led some of the key reform measures undertaken was to strengthen to suggest a more balanced approach that would involve the financial sector. The overcrowded banking sector (with strengthening domestic demand as a counterbalance in 58 financial institutions) was consolidated into 10 banking case exports drop sharply in the future. This concern is based groups with larger capital requirements. Improvements in on the faltering economic recovery in the United States and the banking sector are reflected in the low level of non- the prospects of weak growth in the Euro Area, which may performing loans: the Malaysian banking sector's non- dampen global trade and economic development for some performing loans as a share of financial assets were one of time to come. the lowest among Asian economies for 2008. Exposure to 209 210 Managing Openness foreign loans was also kept to a minimum. Additional pru- in particular the manufacturing sector, escaped deep dential measures were put in place, and corporate gover- restructuring. The exception was the financial sector, as nance was strengthened. Malaysia took a more cautious mentioned earlier. approach to financial liberalization and set out guidelines and plans in the Financial Sector Masterplan and Capital Impact of the 2008 Global Crisis Market Masterplan, both introduced in 2001. As a result, Malaysia was very cautious when approving the use of The global financial crisis of 2008­09 with its epicenter in complex, innovative, yet risky financial instruments such the United States has had enormous ramifications for the as collateralized debt obligations. world economy. It started as an asset bubble caused by an Since 1998, Malaysia has produced continuous trade array of financial derivatives that, among other things, surpluses (for 151 months), which enabled the accumula- caused the subprime mortgage boom and bust. It then sent tion of large international reserves and a healthy balance- shock waves through banking and financial institutions of-payments position. Manufactured exports continue to with a subsequent effect on consumer demand. As con- be the major component of exports, while palm oil and sumer demand in the United States shrank, export-oriented petroleum have increased their share, mainly thanks to Asian economies began to face an economic contraction in high commodity prices. Table 16.1 gives the summary indi- the real sector. Hence, although the Malaysian economy was cators of the health of the Malaysian economy in 2007, that insulated from the direct effects of financial exposure is immediately before the current crisis. (partly because derivatives were not widely permitted) the Another good lesson that Malaysia learned from the crisis has affected the economy through a collapse in 1998 Asian crisis was the importance of having sufficient exports, a large capital outflow, and a slowdown in foreign international reserves. The possibility of large and volatile direct investment (FDI). The Malaysian financial sector capital outflows through the loss of market confidence or escaped unscathed from the global crisis because it was in a activities of currency speculators made Malaysia more cau- strong position as a result of the measures taken in the tious in liberalizing its financial sector and capital markets. 1998 crisis and higher capital requirements. It began accumulating reserves to meet such eventualities. As a consequence of the crisis, the Malaysian economy Notwithstanding these sound macroeconomic funda- began to slow down in the fourth quarter of 2008, when mentals, Malaysia has a number of vulnerabilities. Since GDP grew by only 0.1 percent (figure 16.1). The full 1998, Malaysia has had a persistent fiscal deficit and has impact of the global crisis was felt in the first three quarters proved unable to revert to a surplus. This situation has of 2009, when GDP declined by 6.2 percent, 3.9 percent, constrained Malaysia's ability to stimulate the economy for and 1.2 percent, respectively. The indirect sources of a sustained period of time in the event of an export slow- contraction came from all sectors dependent on GDP down. Because of the rapid recovery from the 1998 crisis and growth--consumer and investment demand. Falling demand Malaysia's unwillingness to undergo the structural adjust- has already driven firms to retrench and lay off workers. ment programs imposed by the International Monetary Although retrenchments reported by government sources are Fund on other crisis-hit countries (Indonesia, the Republic of few, it is believed that both underemployment (and with it Korea, and Thailand, for example), the Malaysian economy, the consequent contraction in incomes) and unemployment are really much higher. During this global crisis, transmission of the impact to Table 16.1. Malaysia's Economic Conditions before the Global Crisis, 2007 the Malaysian economy was different from that during the Asian crisis in 1998. This time, transmission was through Indicator Malaysia the real sector, namely, exports, investment flows, and Current account balance (% of GDP) 15.5 prices of exports, unlike the developed countries, where the Foreign exchange reserves 101.2 primary channel was through the financial sector. While the External debt (% of GDP) 32.2 impact on the real economy for the developed countries Fiscal balance (% of GDP) ­3.2 Government debt (% of GDP) 38.5 came later, for Malaysia the real sector was the first point of Domestic credit (% change) 8.6 impact--the financial sector was almost unaffected. Nonperforming loans (% of total loans) 3.8 Given Malaysia's high ratio of exports to GDP, contrac- Interest rates (%) 3.54 tion in external demand was the most serious factor bur- Consumer price index (% change year on year) 2.0 dening the economy. The direct sources of problems came Source: World Development Indicators database, World Bank, http://data from a contraction in export demand--particularly in .worldbank.org. manufacturing--as the developed markets (the European The International Crisis and Development Strategies: The Case of Malaysia 211 Figure 16.1. Real GDP Growth in Malaysia, 2008­10 Figure 16.2. Merchandise Export Growth in Malaysia, Quarter 1, 2008­Quarter 1, 2010 12 40 10 30 8 real GDP growth (%) 6 20 growth (yoy, %) 4 10 2 0 0 ­10 ­2 ­4 ­20 ­6 ­30 ­8 ­40 1 2 3 4 1 2 3 4 1 1 2 3 4 1 2 3 4 1 -Q -Q -Q -Q -Q -Q -Q -Q -Q -Q -Q -Q -Q -Q -Q -Q -Q -Q 08 08 08 08 09 09 09 09 10 08 08 08 08 09 09 09 09 10 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 quarter 20 quarter Source: East Asia Update database, World Bank, http://www.worldbank.org/ Source: External Trade Database, Department of Statistics, Malaysia, eapupdate. http://www.statistics.gov.my. Note: yoy = year on year. Union [EU] and the United States) were gripped by a reces- 2008), or nine months of retained imports. These reserves sion. The slowdown through exports was also caused by were able to meet the demand of the large capital outflows East Asian trade and production links through intricate of US$8.7 billion in the second quarter and US$11 billion production networks and vertical specialization. As a result, in the third quarter of 2008. Capital outflows gave rise to a demand for intermediate goods from Malaysia was affected small ringgit depreciation against the U.S. dollar. when demand for final goods manufactured in regional The crisis also triggered a slump in commodity prices production networks fell. from their high in the first quarter of 2008. The collapse of Malaysia has never before seen such steep drops in its commodity prices has affected the income of key sectors in manufactured exports, as shown by figure 16.2. In 2008, the Malaysian economy, palm oil and petroleum. At the manufactured exports expanded by 13.4 percent, but in same time, though, it has removed the inflationary pres- 2009, they fell by 21.1 percent. However, the drop in exports sure experienced in the first half of 2008, when oil prices did not cause a deficit in the trade account, because imports reached US$147 per barrel. fell even more sharply. The structure of its manufacturing The deteriorating trade and production conditions and sector provides a balancing mechanism that helps the bal- the large outflow of short-term capital did not significantly ance of trade remain positive when exports decline. Imports erode consumer confidence and business conditions. The are used mostly for producing export goods, while exports consumer sentiment index survey conducted by the of resource-based industries use relatively few imported Malaysian Institute of Economic Research shows that con- inputs. Consequently, the manufacturing sector was the sumer confidence is still relatively intact. Although con- worst hit among the economic sectors, with a contraction of sumer sentiment dipped slightly in the second quarter of 9.4 percent in 2009 as shown in table 16.2. 2008, it then rebounded in the third quarter. The institute's The second direct source is a contraction in inflows of business conditions index showed that the private sector FDI, particularly from the developed countries. Overall, expected deteriorating prospects for business conditions in FDI inflows into Malaysia started falling from the third the middle of 2008. quarter of 2008 until the second quarter of 2009. Diminishing industrial production led to layoffs and Massive global risk aversion and greater demand for liq- higher unemployment. The number of laid-off workers due uidity, coupled with pessimistic growth prospects, have to the crisis was not very serious, however. According to the resulted in sharp declines in equity prices. The decline was Ministry of Human Resources, total layoffs from October 1, most pronounced in September 2008. Continuous and 2008, to January 22, 2009, were 14,158, comprising largely healthy trade surpluses had enabled Malaysia to accumu- local workers (10,636). The ministry tends to view the situa- late a comparatively large stock of international reserves, tion as manageable, because there were also 16,600 vacancies, which at its peak stood at about US$125 billion (as of June that is, a number greater than the total number of layoffs. 212 Managing Openness Table 16.2. Malaysia's GDP by Sector, 2007­Quarter 1, 2010 annual % change Q1 Q2 Q3 Q4 Q1 2007 2008 2009 2009 2009 2009 2009 2010 Agriculture 1.4 4.3 0.4 ­4.4 0.4 ­0.4 5.9 6.8 Construction 4.7 4.2 5.8 1.2 4.5 7.9 9.3 8.7 Manufacturing 3.1 1.3 ­9.4 ­17.9 ­14.5 ­8.6 5.0 16.9 Mining 2.0 ­2.4 ­3.8 ­5.2 ­3.5 ­3.6 ­2.8 2.1 Services 9.6 7.4 2.6 ­0.2 1.7 3.4 5.2 8.5 Source: Quarterly Malaysian Economy, various issues, Ministry of Finance, Malaysia, http://www.treasury.gov.my. Recovery ESP 2 covers two years, 2009 and 2010. The main purpose of the packages was to boost domestic aggregate demand. Malaysia decided that domestic demand should play a bigger Nearly 43 percent of the first package was intended for role in mitigating the economic recession and introduced infrastructure, to upgrade, repair, and maintain public fiscal stimulus programs to overcome the export contraction amenities, including schools, hospitals, roads, quarters for and maintain stability in the economy. The government police and armed forces, and police stations, for example. It implemented a monetary policy and fiscal injections to miti- also covered the building of more low-cost houses, more gate the negative impact of the global crisis. However, the fis- public transportation, and the implementation of high- cal injections were the more pronounced because they were speed broadband. The package contained three main parts: quite big and covered many economic areas. The benefits of ensuring citizens' well-being, developing quality human fiscal stimulus include, among others, raising output and capital, and strengthening national resilience. While proj- incomes in the short run when the economy is operating ects relating to these three main thrusts vary in nature, they below its potential. This approach is considerably different focus primarily on construction and infrastructure, trans- from the 1998 capital controls because the approach now port, the banking and finance sectors, and education. retains a flexible exchange rate with minimum regulation Malaysia's second stimulus package was 8.5 times larger from the central bank compared to the fixing of the ringgit than the first and equivalent to 9 percent of its GDP. against the dollar and the banning of ringgit transactions Nearly 48 percent went to assist the private sector through outside the country in September 1998. The stock market was bank guarantees for small- and medium-size companies. also shielded in 1998, whereas it remains fairly liberal now. Another 32 percent went to infrastructure. But of this sum, The initial reaction of the government to the global cri- a substantial portion went to maintenance rather than to sis was to relax monetary policy. Expansionary monetary new spending on public facilities. Seventeen percent of the policy was implemented as early as December 2008. Bank spending from the second stimulus was targeted to the vul- Negara Malaysia (the Malaysian central bank) had main- nerable through food, toll, and fuel subsidies and for tained reasonably low interest rates before the global crisis, support of low-cost housing and laid-off workers. The despite high inflation in early 2008 caused by the increase remaining 3 percent went toward reducing unemployment in world crude oil prices and the removal of petrol and and increasing job and training opportunities. Although a diesel subsidies in the country. The liquidity ratio for the total sum of RM 60 billion was announced for this second economy was reported to be satisfactory despite gloomy stimulus, the actual spending in 2009 and 2010 was only reports from export industries and increasing retrench- RM 10 billion. Tax incentives amounted to RM 3 billion, ments in the manufacturing sector. Accordingly, the revi- and RM 10 billion is for strategic investment by the sion of the overnight policy rate and the statutory reserve national sovereign wealth fund. With this large stimulus requirement (SRR) led to a reduction in the base lending package relative to GDP, Malaysia's fiscal deficit rose from rate; in February, the overnight policy rate was reduced to 4.8 percent in 2008 to 7.6 percent in 2009. 2 percent from 3.5 percent in 2007. ESP 2 covers the period 2009 and 2010 and is focused The government unveiled two economic stimulus pack- on four main areas: reducing unemployment and increas- ages (ESP) allocating RM 7 billion (US$1.9 billion) or 1.04 ing employment opportunities, easing the burden on citi- percent of GDP (ESP1) in November 2008, and RM 60 billion zens, assisting the private sector in facing the crisis, and (US$16.2 billion) or 9 percent of GDP (ESP2) in March 2009. building capacity for the future. Like other countries in the The ESP 1 was targeted to cover the first quarter of 2009. The The International Crisis and Development Strategies: The Case of Malaysia 213 region, Malaysia recovered relatively quickly. After negative Malaysia (the developed markets) has an important and GDP growth in the third quarter of 2009, the fourth quar- immediate impact on Malaysian exports. In this case, the ter registered growth of 4.5 percent. Growth in the first issue of competitiveness and the state of Malaysian export quarter of 2010 was very encouraging at 10.1 percent: industries are secondary, because the collapse in global Malaysia's growth was among the highest in East Asia. This demand automatically results in a fall in demand for commendable performance is attributed to the strong Malaysian exports and leads to an economic slowdown or regional growth and revival of exports due to restocking by even a recession. industries, although fiscal stimulus programs could have Second, because of Malaysia's heavy dependence on kept the economy buoyant during the export slowdown. exports, the revival from the economic slowdown is very The high growth rate in the first quarter 2010 is perhaps much dependent on external demand, be it regional or assisted by the low base (year-on-year). global. Recovery by expanding domestic demand may not be as quick because measures such as public sector infra- Malaysia's Development Strategies structure investment or tax deductions will take time to Going Forward produce the required results. Recovery very much depended on exports in both the 1998 and the 2009 crises. The 2008­09 global crisis has sharpened the debate on the Third, for the financial sector, the state of the domestic future of Malaysia's outward-oriented development strategy sector is critical in determining its ability to withstand that began during the Asian crisis in 1998. This call for review global shocks. When that sector is strong and well gov- of Malaysia's growth path is not unexpected, given the erned, the impact of external volatility will be manageable. importance of the contribution of its external sector and its Fourth, building safeguards is important to assuring integration into the global economy. The contribution of domestic and foreign constituencies that economic funda- the external sector comes in three ways: exports, financial mentals are strong. In this regard, sufficiently large interna- flows, and FDI. Export industries have created employment, tional reserves, low short-term external debt, and a realistic upgraded the skills of workers, acquired technology, and put exchange rate are some of the fundamentals that maintain Malaysia in regional production networks. Export surpluses public and investor confidence. have provided the liquidity to expand economic activities Fifth, it is important to differentiate between external and buy imports. The financial flows and FDI are important shocks that affect demand and those that affect the com- sources of investment. Increasing global volatility has height- petitiveness and strength of the external sector--export ened concerns that the Malaysian growth strategy might be industries, the financial sector, or the investment climate. It too dependent on external factors. is difficult to be insulated from demand shocks, but recov- Although the two crises--the 1998 Asian and 2008­09 ery will be quick once demand is restored. However, if the global--have had different impacts on the Malaysian econ- external sector is uncompetitive, medium- and long-term omy, they both point to some very useful and pertinent les- growth will be jeopardized. sons. On the one hand, in the 1998 crisis, the external shock had severe consequences for the Malaysian financial sector, whereas the export sector was only marginally affected and Is Malaysia's Outward-Oriented Strategy in fact led the recovery process. On the other hand, in the the Right Approach? 2008­09 global crisis, the Malaysian financial sector escaped Why is Malaysia's external dependency and outward- the spillover effects of the collapse of the developed coun- oriented development strategy being questioned now? The tries' financial sector, especially major financial institutions. seeds of doubt about its strategy were sown during the The unexpected outcome was the deep contraction of real 1998 Asian crisis. Before that crisis, Malaysia was very pos- sector exports. Likewise, the movement of the ringgit was itive about globalization and its close integration with the very different in the two crises: in 1998, the ringgit depreci- international economy. Moreover, the rapid recovery from ated by 45 percent causing massive damage to the economy, crisis, partly based on the newly introduced measures, gave which prompted the imposition of capital controls and the impression that the crisis was a one-off event and that pegging of the ringgit. In contrast, the movement of the global volatility would be confined to short-term capital ringgit was mild in the 2008­09 crisis. flows and currency speculators. In addition, expanding international trade was seen as a key reason for the recovery. Lessons from the Two Crises Unlike the 1998 crisis, this 2008­09 global crisis has called A number of lessons can be drawn from these two crises. into question Malaysia's outward development strategies First, the health of the real sector in the major markets for because of the gloomy prospects for trade growth as a result 214 Managing Openness of the economic difficulties in its two major markets--the global demand, as well as Malaysia's own difficulties in EU and the United States. This time the fundamental issue is sharpening its export competitiveness and attracting larger weak demand for exports, and there is relatively little FDI volumes. But in the short and medium term, Malaysia Malaysia can do about that. The massive export decline in is likely to continue depending on the international econ- 2009 has highlighted Malaysia's heavy reliance on global omy for its growth. Switching to domestic demand can be trade: cutting of costs, improvement of competitiveness, or only a long-term strategy, because Malaysia has to increase currency depreciation will not revive demand for exports. its domestic market and purchasing power first before that The more cautious approach to financial liberalization and strategy can be effective. Notwithstanding the importance integration taken after the 1998 crisis was seen as having of outward orientation, it is imperative for Malaysia to saved Malaysia from the potentially devastating damage that have a balanced growth strategy that expands domestic was suffered by the financial sectors in developed countries. demand while at the same time improving its global com- But the larger concern is the prospect of growth in petitiveness and links. Malaysia's largest export markets, the EU the United States. Malaysia realizes that it needs to continue the high- EU fiscal adjustments may see stringent government cut- global-dependency approach; yet it is also aware that that backs that can reduce growth significantly, and the United approach will expose its economy to unpredictable swings States has to tackle its high unemployment if it is to have a with serious negative consequences. Continuing the pres- sustainable recovery. With the developed countries likely to ent outward-oriented strategy, however, will also require grow slower for some time, what are the prospects for frequent fine-tuning and refocusing. In the meantime, demand for Malaysian exports? Although East Asia, China Malaysia still needs to mitigate possible negative impacts in particular, has stepped in to assist the global recovery, coming from future external shocks and volatility, since there are doubts about whether the region can create sus- these are part and parcel of globalization. For this purpose, tainable demand to replace that lost from the EU and the it is useful to give examples of how Malaysia has managed United States. some past vulnerabilities. The call to review Malaysia's outward-oriented strategy Primary commodities are important export products was reinforced further with declining growth of manufac- for Malaysia in revenue contribution and participation in turing exports and FDI inflows. Before 1996, Malaysian the global economy. For example, in 2008, exports of pri- manufacturing exports were growing at over 20 percent mary commodities--oil and gas, palm oil, sawn timber-- per year (for example, in 1994 the growth rate was 34 per- contributed about 25 percent of total exports. Changes in cent). However, the annual growth rate for 1996­99 was commodity prices can negatively affect smallholders in the 13.2 percent, and for the period 2000­09, it was only palm oil and rubber industries, whose incomes are already 5 percent. Besides this performance, an added concern is at low levels. Long cycles of commodity prices have pro- the inability of manufacturing exports to move up the duced surpluses that formed a substantial source of gov- value chain: some industries continue making existing ernment revenues. These revenues were used to finance products by keeping costs low through the use of low- Malaysia's development program by building basic infra- skilled, low-cost foreign workers. Smaller FDI flows are structure such as schools and hospitals. In the case of oil, another concern because Malaysia's growth strategy has a significant proportion of these revenues were invested been premised on a high investment ratio. With the abroad to find new oil and gas reserves as well as in other public sector cutting its investment to reduce the fiscal business activities to diversify income sources. For the deficit, more efforts have been made to attract FDI, but palm oil industry, a levy collected from large producers is with few results to show so far. Many incentives were used to fund development of the industry and to provide a offered for this purpose, which resulted in some parties' buffer in case of a fall in prices. questioning the net benefit of FDI. The most obvious The increasing global competition for FDI has prompted benefit is employment creation, but the net value added Malaysia to improve its investment climate. A number of of multinational companies' exports, links to the domestic measures have been introduced for this purpose, including economy, and transfer of technology are generally small. a wide range of new incentives, reducing red tape for faster Some have recommended promotion of domestic invest- approval processes, liberalizing equity ownership rules (for ment and technology development to deepen industrial example in 1998, full foreign ownership was allowed in the integration. manufacturing sector), and establishing a more flexible In short, the calls for Malaysia to review its outward- policy on recruitment of skilled foreign workers. oriented development strategies are prompted by the vul- Stronger governance, adherence to international stan- nerability caused by global shocks and potentially lower dards, higher capital requirements, and consolidation of The International Crisis and Development Strategies: The Case of Malaysia 215 financial institutions are among the measures introduced Technology acquisition and availability of skilled labor for the financial sector. These enable such institutions to are essential for this transformation. Another important cope better with external shocks and deeper liberalization. measure is to develop more exports that originate in Malaysia. Currently, a sizable portion of Malaysian manufacturing exports are produced as part of regional Proposals for Bringing Sustained production networks (primarily in the electrical and High Growth to Malaysia electronics industry). Although this form of export is very As noted earlier, Malaysia should continue with its out- important and has produced large benefits, Malaysia ward-oriented development strategy, but it should refocus should also develop its own export capability separate and improve that strategy. What sorts of policies can help from these production chains. For primary commodities, Malaysia mitigate the impact of future global financial it is imperative to increase the proportion of downstream crises on trade? Can Malaysia minimize its exposure to the exports, especially with the emergence of more low-cost volatility that comes with globalization without moving producers of upstream products. away from an export-oriented growth strategy? The fol- · Diversify markets. East Asia has now become Malaysia's lowing are proposals for improving the outward-oriented largest trading partner, although markets outside the strategy and bringing sustained high growth for Malaysia: region are the final destination of a large part of those exports. Notably, China's share in manufactured exports · Have a balanced export structure. At present, Malaysia's has increased from 1.7 percent in 1996 to 12 percent in exports come from three major groups--manufacturing, 2009. The share of the Association of Southeast Asian mining, and primary commodities. In the past, this Nations remains at about 27 percent. The share of tradi- structure has produced a good counterbalance when tional markets has declined, however: the United States one component suffered a slowdown; the other groups to 13.2 percent in 2009 from 21.7 in 1996, the EU to grew quite well because they had different consumers 12.2 percent from 14.5 percent, and Japan to 7 percent and demand cycles. Consequently, total exports did not from 11.1 percent for the same period. Diversification is experience a large decline. For example, when manu- important especially in view of the moderate growth facturing exports fell, exports of primary commodities prospects and consequent lower demand for imports increased (sometimes because of either demand or from these traditional markets. If China and other Asian price increases or both). Therefore, Malaysia should countries are able to generate their own domestic keep this balanced export structure. The structure of demand, then the prospects for Malaysia's exports will manufacturing exports, however, is too skewed toward be bright. the electrical and electronics industry. Expansion of · Prepare the financial sector for more global competition. other manufacturing exports is needed if Malaysia is to Malaysia has already undertaken measures to prepare avoid any sharp drop in manufacturing exports as a result the financial sector for global competition, as dis- of a fall in global demand for electrical and electronics cussed earlier. It should implement measures proposed products. in the Financial Sector Masterplan and Capital Market · Accelerate the expansion of services exports. Malaysia has Masterplan as well as introducing additional steps to made good progress in promoting services exports: improve governance, enhance standards, and strengthen tourism is a major export earner, and the Multimedia industry players. Super Corridor is focusing on information technologies · Improve the investment climate to attract FDI. Recently, and business off-shoring services. Other services exports Malaysia has restructured its main investment promo- that have been identified are Islamic finance, health and tion agency (Malaysia Industrial Development Authority) education services, and air passenger transport services. as part of a larger initiative to attract both domestic and Malaysia has strategic advantages in these areas, includ- foreign investment. Another focus area is reducing ing its diverse cultural and ethnic society, good health unnecessary barriers in the investment approval process. and education services, and first-mover advantage in Since Malaysia is already host to a large amount of FDI, low-cost air transport services. more efforts should be made to encourage reinvestment · Upgrade existing exports. Manufacturing and primary because these investors are already operating locally and commodity exports should be upgraded to incorporate relationships have been formed. higher value added and innovation. This move will reduce the possibility of hollowing out the production Besides these specific proposals, the ability to withstand of manufacturing exports to other low-cost producers. global volatility and to continue to be a destination of 216 Managing Openness choice for businesses will depend on Malaysia's own initial In its present trade agreements, Malaysia has not made conditions. Strong macroeconomic fundamentals such as deep liberalization commitments because of political econ- low national and foreign debts, a sound fiscal position, low omy considerations. In the future, it should strive for inflation and unemployment, a current account surplus, higher-quality free trade agreements or economic integra- high international competitiveness, capacity for research tion pacts as part of its outward-oriented development and development and technological development, and strategy. Many beyond-the-border issues such as govern- availability of skilled workers are factors that provide ment procurement, opening up of sensitive sectors, own- confidence to the market and investors. In a way, the ership rules, and competition policy are liberalized in above recommendations are the insurance policy that high-quality free trade agreements. Taking into account Malaysia can establish to protect itself from global volatil- political economy sensitivities, Malaysia can still achieve ity, because markets and investors will be confident about high-quality agreements by selectively liberalizing the con- the health of the domestic economy and of its ability to tentious areas and doing so progressively. Recently, the ride out the volatility. government has taken the bold step of removing equity The currency exchange rate also has a similar role in an restrictions in 27 services subsectors. outward-oriented development strategy as an important In summary, although Malaysia wants to minimize its component of macroeconomic fundamentals and as an exposure to global volatility, it would be difficult to find a instrument for determining export competitiveness. An substitute for the outward-oriented strategy to drive its unsustainable exchange rate that does not reflect eco- growth. Turning to domestic demand as a source of growth nomic fundamentals can lead to speculative attacks and is an option that must be considered seriously because it undermine market confidence. At the same time, if the counterbalances sharp global swings. This effort will take exchange rate is kept at an unrealistically low level to time, however, because the government's fiscal position ensure export competitiveness, there will be an upward limits its ability to stimulate the economy. Also, household pressure to reflect the true economic fundamentals. purchasing power needs to be increased through rising Thus, management of the exchange rate to balance export wages matched by productivity improvements and a better competitiveness against economic fundamentals is criti- income distribution. A policy of ramping up demand when cal for stability in the outward-oriented development households have low savings or limited disposable income strategy. will only increase household debt. Preferential trade agreements, economic integration, Going forward, Malaysia should improve its outward- and regionalism are initiatives for ensuring a stable and oriented development strategy as suggested above. This predictable environment for external economic relation- process should strengthen its competitiveness and make it ships. Liberalization is the major part of these initiatives more able to cope with both global volatility and structural wherein rules and regulations are set to ease market changes in the international economy. Although the East access and promote investment. For this purpose, Malaysia Asia region and other emerging markets will be important has embarked on a number of bilateral, regional, and multi- to Malaysia, Malaysia does not seek to decouple itself from lateral trade agreements. They provide certainty in conduct- the EU and the United States. Finally, sharp global swings ing trade, which is essential for increasing trade volume; trade are likely to be increasingly common. Malaysia should get agreements are, however, powerless to prevent global volatil- ready now, by buttressing its fundamentals. That approach ity. Nonetheless, such agreements are important for keeping will reassure domestic and international markets and give trade channels open and preventing protectionism when it the ability to take whatever measures are necessary to economies face a slowdown. The 2008­09 crisis showed that. deal with short-term impacts. 17 Should Indonesia Say Goodbye to Its Strategy of Facilitating Exports? Muhammad Chatib Basri and Sjamsu Rahardja Millions of Indonesians have witnessed their economy's or protectionist views for a new reason: many have argued relative resilience during the global financial crisis. The that Indonesia should limit openness and integration with Indonesian economy continues to grow even though its the global economy to prevent itself from being dragged export engine has been hit as badly as that in other Asian into the global economic slump. This view adds to the exist- economies. Although Indonesia's share of exports in gross ing notion that openness exposes Indonesian firms to domestic product (GDP) is lower than that of its Asian unfair competition. Both views are popular. Political pres- neighbors, the collapse of global trade was powerful sures against more openness sometimes influence policy enough to lower GDP growth from 1.5 percent in the third makers to implement more inward-looking policies and quarter of 2008 to 0.2 percent in the fourth quarter of 2008 rely less on exports. What, then, is the relevance of a strategy (quarter-to-quarter, seasonally adjusted). But domestic that facilitates exports and openness to economic growth? demand--that is, private and government consumption-- The chapter examines the future of Indonesia's export- expanded by 4.9 percent in 2009 and provided a cushion led growth strategy, focusing on questions such as, What for Indonesian economic growth during the early period of reforms can help Indonesia benefit from continuing to the global crisis. It has also supported a quicker economic open up to trade and investment? What kinds of policies recovery than in other countries. Many believe that its rela- best serve both domestic and global orientation? What was tively insulated economy is the reason that Indonesia has the impact of the global economic crisis on the Indonesian performed relatively better than other economies during economy? Is the choice of an inward-looking strategy the the global financial crisis (Basri and Rahardja 2010). Given right one? What role do composition and diversification of that what saved Indonesia from the dreadful effects of the exports play in reducing vulnerability? What policies are global financial crisis was the domestic economy, the ques- needed to support export diversification? tion arises whether an export-led growth strategy is still relevant. This question is not specific to Indonesia, because The Impact of the Global Financial many other countries that are more supported by their Crisis on Indonesia domestic economies have performed better than export- dependent countries during the global financial crisis On the financial side, the effects of the global crisis on the (Basri and Rahardja 2010). Indonesian economy appeared through several indicators, The challenge for Indonesia is to manage its economic such as a depreciation of the exchange rate and a decline in policy so that it can navigate through the global economic the stock market.1 The rupiah exchange rate had fallen by crisis. The crisis provides important lessons for Indonesia, 30 percent by the end of 2008. The Indonesia Stock Mar- including the need to strengthen domestic demand. How- ket Index experienced a drop of 50 percent in 2008. ever, this experience leads commentators, politicians, and Growth of banking credit also experienced a significant some policy makers to embrace somewhat more nationalist drop, from 32 percent to 10 percent (Basri and Siregar 217 218 Managing Openness 2009). In addition, confidence among banks declined, as the Philippines, Singapore, and Thailand experienced a seen in the shrinking size of interbank borrowing and contraction in export growth of around 30 percent in the lending, which was down by 59.3 percent to Rp 83.8 trillion fourth quarter of 2008 and the third quarter of 2009. in December 2008 from Rp 206.0 trillion in December Why did this sharp drop in exports have a limited effect 2007 (Gunawan, Arman, and Hendranata 2009). The need on the Indonesian economy? We argue that this was to carry out the expansion of banks' funding bases, added because the contribution of exports within the Indonesian to again by increases in interbank rates, had already cre- economy was relatively small compared with countries like ated sharp competition between banks, which in turn Malaysia, Singapore, and Thailand. The total share of had supported higher interest rates. Indonesian exports in GDP was only as high as 29 percent. Weak global economic growth began reducing demand This percentage is far smaller than the share in other for Indonesian exports in the fourth quarter of 2008. In countries such as Singapore (234 percent); Taiwan, China addition, the drop in global demand also resulted in weak (74); and the Republic of Korea (45).3 demand for primary and mining exports, which led to a This phenomenon has led to lively discussions among drop in the price of commodities and mining goods. The commentators, politicians, and some policy makers about fall in global economic growth weakened demand for the importance of relying less on exports and focusing energy so much that the price of oil also declined. Papanek, more on the domestic market. Often the conclusion of Basri, and Schydlowsky (2010) point to an extremely sharp such discussions is advice to policy makers in Indonesia to decline in exports in the first quarter of 2009 compared to pay less attention to "openness" to trade and investment the same period in the previous year (table 17.1), a decline and to concentrate more on protecting the domestic econ- driven more by price than volume. In fact, the demand for omy against external volatility. Given the close link primary commodity exports, especially agriculture and between exports and openness to trade and investment, mining, was relatively stable in quantity terms, as indicated this issue leads to the question of whether Indonesia by the relative stability of the export volume of agriculture should leave behind its export-led growth strategy and and mining commodities, such as crude palm oil (CPO), adopt an inward-looking, more protectionist strategy. coal, and copper (Basri and Rahardja 2010). With natural Amid that debate, a study by Basri and Rahardja (2010) resources accounting for more than half of Indonesia's indicates that exports are in fact an important source of exports, they provided some life support for the Indone- Indonesia's economic growth. Exports have a large support- sian economy (figure 17.1). In addition, the depreciation of ing effect on economic growth, albeit a less stable effect the rupiah that took place after September 2008 also com- than domestic demand (figures 17.2 and 17.3). Therefore, a pensated for the effects of the collapse in the demand for strategy that safeguards a balance between the domestic exports. However, the data show that the increase in economy and global orientation--such as joining produc- demand due to the depreciation of the rupiah (substitution tion networks and promoting export-oriented growth-- effect) was smaller than the fall in demand due to the must be part of the development strategy of the national decline in income (income effect). As a result, exports economy. Strengthening domestic demand can be done experienced an overall reduction. without resorting to protectionist policies. With the export weakness that started in the fourth quar- The study by Basri and Rahardja (2010) also shows the ter of 2008, economic growth slowed to 5.2 percent year on strong link between exports and the strength of the domes- year in the fourth quarter of 2009. Even so, Indonesia's tic economy. Using quarterly data, they found that the rela- overall economic growth was still 6.1 percent, the highest tively strong growth in consumption during the crisis growth in Asia, after China and India. period was a lagged effect from the relatively strong exports in the previous two to three quarters (table 17.2). Certainly, we are cautious in not interpreting this finding as a causal Should We Say Goodbye to the Strategy relationship. Given the scope of data covering the first of Facilitating Exports? quarter of 2001 until the last quarter of 2008, we are also Indonesia was not the only country to experience a sharp taking the results as valid only for those time periods. Nev- decline in exports.2 A similar decline was experienced by ertheless, the exercise gives some idea of how much private many countries, including China, Malaysia, Singapore, and consumption moves together with exports and government Thailand. In fact, the large export contractions that consumption. The comovement between private consump- occurred suggest that the force of the global economic tion and government consumption is somewhat expected. crisis hitting the Indonesian economy was in fact very sim- As a response to the global economic downturn, Indonesia ilar. Figure 17.1 shows how China, Indonesia, Malaysia, implemented a fiscal stimulus targeted at infrastructure Table 17.1. Exports at Current and Constant Prices, by Major Commodities, 1975 to Quarter 1, 2009 Constant prices (annual % change) Current prices (annual % change) At 1980­2008 average prices At 2008­09 prices 1975­85 1985­90 1990­96 1996­08 2004­08 Q1 2008­Q1 2009 1975­85 1985­90 1990­96 1996­08 2004­08 Q1 2008­Q1 2009 Q1 2008­Q1 2009 Total exports 10.1 6.4 11.8 8.9 17.8 ­31.8 0.7 7.3 9.1 5.1 5.4 ­19.5 ­16.2 Non-oil, gas total 12.6 19.4 17.6 9.2 18.0 ­25.3 0.8 11.7 13.9 7.6 7.2 ­20.5 ­17.3 Labor-intensive manufacturesa 61.8 44.6 17.1 3.2 7.1 ­13.8 49.5 31.9 17.1 3.9 3.1 ­13.6 ­13.2 Commodity-based exports 9.6 2.6 7.5 9.9 21.7 ­38.3 0.2 4.5 5.0 3.4 5.5 ­22.8 ­16.0 Increase in US$ billions at constant prices = change in quantity weighted by price and unit value Increase in US$ billions in current prices At 1980­2008 average prices At 2008­09 prices 1975­85 1985­90 1990­96 1996­08 2004­08 Q1 2008 ­Q1 2009 1975­85 1985­90 1990­96 1996­08 2004­08 Q1 2008­Q1 2009 Q1 2008­Q1 2009 Total exports 11.5 6.8 24.1 87.6 65.8 ­10.7 1.6 9.7 22.4 45.4 19.1 ­5.0 ­5.0 Non-oil, gas total 4.1 8.4 23.4 70.2 52.3 ­6.7 .7 7.3 20.5 52.7 21.8 ­4.7 ­4.4 Labor-intensive manufacturesa 0.6 3.2 6.0 4.5 3.4 ­0.5 .8 2.4 4.9 4.7 1.5 ­.4 ­.5 Commodity-based exports 10.6 2.4 10.9 64.8 52.2 ­9.2 .5 5.3 9.3 9.3 10.4 ­3.4 ­3.4 Source: Papanek, Basri, and Schydlowsky 2010. Note: Q = quarter. a. Includes textiles, garments, footwear, and furniture. 219 220 Managing Openness Figure 17.1. Export Values in Selected Markets, Quarterly Growth, Seasonally Adjusted, 2008­09 40 30 20 10 percent 0 ­10 ­20 ­30 ­40 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 2008 2009 year Singapore China Indonesia Malaysia Thailand Philippines Source: Basri and Rahardja 2010. Note: Q = quarter. Figure 17.2. Impulse Response Function of GDP due to Figure 17.3. Variance Decomposition of GDP due to Exports Export and Domestic Demand Shocks in Indonesia and Domestic Demand Shocks in Indonesia 3.00E-03 45 impulse response function 2.50E-03 40 2.00E-03 35 1.50E-03 30 percent 1.00E-03 25 5.00E-04 20 0.00E+00 15 ­5.00E-04 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 periods after shock 5 exports domestic demand shocks 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Source: Basri and Rahardja 2010. periods after shock exports domestic demand shocks spending. The fiscal stimulus has a rather poor disburse- Source: Basri and Rahardja 2010. ment record, however, and therefore one could argue that it was less effective than it should have been in stimulating the economy. economic activities is explained as follows. Economic Because of the data period used, it is also likely that activities outside Java increased as a result of the commod- commodity exports played an important role in driving ity boom that occurred several years before. This increase consumption. The effect of the commodity boom on was reflected in relatively high credit growth outside Java Should Indonesia Say Goodbye to Its Strategy of Facilitating Exports? 221 Table 17.2. Co-movements of Innovations in Private Consumption and Innovations of GDP Components, Quarter 1, 2000­Quarter 4, 2008 Components of GDP Lags (quarters) Government consumption Gross fixed-capital formation Exports 0 0.12 0.06 0.24 ­1 ­0.16 ­0.04 ­0.27 ­2 ­0.22 ­0.01 ­0.41 ­3 0.26 ­0.07 0.29 ­4 0.20 ­0.13 0.49 Source: Basri and Rahardja 2010. Note: Comovements between innovations of each component derived from original data that span the first quarter of 2000 to the fourth quarter of 2008. Here growth is expressed as annual (year-to-year) growth. several years ago (figure 17.4). The growth of third-party the 1970s to support a nationalist economic agenda through funds in commodity-producing regions also experienced a expansion of state-owned businesses (see Hill 1996). slow increase. These data strengthen the argument that the The high crude price had a profound impact on dis- economy outside Java improved as a result of the commod- couraging diversification away from the oil sector. Aside ity boom, and during the crisis period, residents outside from the inflationary pressure from swelling foreign Java were capable of making use of their accumulated exchange reserves, the high crude price provided incentives savings to smooth their consumption during the global for state and private businesses to expand their activities financial crisis. In addition, services exports played an around the business of Pertamina, the state-owned oil important role because of surprisingly strong exports in company. As inflationary pressure built up, the demand for tourism and creative designs, while workers' remittances protection also increased, and the government regulated are likely to have a direct link to private consumption. business activities through multiple permits and prohibi- tions. Trade facilitation was poorly managed; for example, corruption in customs was costly for the private sector. Tariffs Export Diversification: The Indonesian were raised, and nontariff barriers, such as import restric- Experience tions or trade through state trading enterprises, were put in Chapter 11 of this volume shows a positive connection place. With the inflationary pressure and incentive to between export concentration and the total effect of eco- expand oil-related businesses, Indonesia finally found that nomic openness on volatility. The more concentrated its industrialization program did not go far and that the oil exports are, the higher the total effect of openness on sector still dominated the Indonesian economy.5 As time volatility. By implication, export diversification would passed, Indonesia found itself entrenched in a high-cost diminish the effect of growth volatility on an economy fol- economy with low diversification. lowing an export-led growth strategy. This result provides However, the game totally changed when the crude a way for Indonesia to safeguard its export-led growth price started falling. The government had to confront the strategy so long as it diversifies its exports. Therefore, it is reality that its main engine for economic growth, oil rev- very important to look at the experience of export diversi- enue, had fallen away and that it would have to jump- fication in Indonesia. start a new engine. The fall of oil prices to US$14.30 per Diversifying exports was not part of Indonesia's earlier barrel in 1987 from an average of US$36.90 per barrel in development agenda. In addition to foreign aid and rev- 1980 was used by policy makers as the spur for a series of enue from non-oil exports and gas, Indonesia had been structural reforms to ensure the survival of Indonesia's relying on revenues from exporting rubber and crude oil in development. Following a devaluation of the rupiah in the late 1960s and early 1970s to finance its development. 1983 and fiscal austerity measures, Indonesia finally The quadrupling of the crude price in 1973 caused windfall started a sequence of deregulation to bring down the gains for Indonesia, which, unfortunately, resulted in infla- "high-cost" economy, that is, bureaucracy and red tape, to tionary pressure.4 After forceful anti-Japanese protests and free up its non-oil sectors to private businesses. The gov- riots, Indonesia embarked on a state-sponsored industrial- ernment introduced often heavy-handed measures to ization program to substitute imports. Trade and foreign slash licenses and procedures for opening businesses, investment policy also became relatively more restrictive in impose market discipline on state-owned enterprises 222 Managing Openness Figure 17.4. Expansion outside Java because of the Commodity Boom, 2007­10 60 50 loan growth (% change) 40 30 20 10 0 ­10 8 08 8 9 9 09 9 10 00 00 00 00 00 20 20 20 2 .2 .2 2 .2 r. l. r. l. n. ct n ct Ju Ju Ap Ap Ja Ja O O month Gorontalo Kalimantan, East Papua Riau North Sulawesi Banten Java, West Java, Central Java, East Source: Indonesia Financial Statistics, Bank Indonesia, http://www.bi.go.id/web/en/Statistik/Statistik+Ekonomi+dan+Keuangan+Indonesia/. through privatization, increase competition from private value added, such as garments.6 Indonesia also took the businesses, and reform the tax system. unprecedented policy step of stripping customs of its man- Aside from reducing bureaucracy and red tape, policy date to perform inspections. Customs inspections were makers also introduced sweeping trade reforms. Starting delegated to SGS, a Swiss-based company, which brought in the mid-1980s, the government implemented a series down corruption and sped up import clearances. Indonesia of serious trade liberalizations. With respect to tariffs, also replaced priority sectors for investment with a more the study by Fane and Condon (1996) showed that the relaxed negative list approach to attract foreign direct weighted average rate of effective protection for manufac- investment (FDI). turing (excluding the special case of oil and gas processing) That deregulation in the 1980s was the main impetus declined from 59 percent to 16 percent over the period for Indonesia's export diversification up to the Asian eco- 1987­95, while the dispersion (standard deviation) fell nomic crisis. As a result of those policies, Indonesia's private from 102 to 39. They also showed a sharp reduction in the sector had incentives to exploit the country's comparative coverage of nontariff barriers: the percentage of non-oil advantages in non-oil industries. In 1985, Indonesia's non- manufacturing value added affected by these barriers oil exports took off, growing at an average rate of 17.8 per- declined from 77 percent to 17 percent. The government cent between 1985 and 1996, higher than the average also introduced a duty drawback system for exporters and growth of total exports (9.3 percent). By 1996, Indonesia reformed customs procedures. Starting from a point of had successfully diversified its exports away from oil prod- high import tariffs, the duty drawback system on imports ucts, with non-oil exports accounting for 76.4 percent of for export provided a fiscal incentive for businesses to enter total exports compared to only 28 percent in 1980 (figures export sectors, particularly activities with low domestic 17.5 and 17.6). Should Indonesia Say Goodbye to Its Strategy of Facilitating Exports? 223 Figure 17.5. Indonesian Oil and Non-oil Exports, 1970­96 due to the powerful state enterprises that often influenced policy in favor of rent-seeking activities (Soesastro 1989). 40 The deregulation packages helped transform the Indone- 35 sian manufacturing industry from inward looking to increasingly outward looking and internationally competi- 30 tive (Hill 1996). Indonesian manufacturers were quick to 25 exploit investment opportunities in footloose industries US$, billions catering to the export market. Lower import tariffs, the 20 duty drawback system, and improved trade facilitation 15 enabled exporters to access materials and technology and combine them with low-cost labor to produce clothing, 10 footwear, and furniture. By 1996, Indonesia managed to 5 book US$20 billion worth of exports in textiles, clothing, footwear, and furniture, from less than US$2.5 billion 0 in 1986. 70 73 76 79 82 85 88 91 94 96 Figure 17.7 indicates a dramatic increase in the diversifi- 19 19 19 19 19 19 19 19 19 19 year cation of export products following the reform package, as non-oil and gas oil and gas measured by a Herfindahl-Hirschman Index of Indonesia's export products from 1980 to 1995. It is apparent that prod- Source: Authors' calculations based on UN Comtrade (database), United Nations, http://comtrade.un.org. uct concentration fell dramatically after the government launched trade and investment reforms in 1985. By 1994, Figure 17.6. Composition of Indonesia's Non-oil Exports, the concentration of Indonesia's export products was only 1970­96 41 percent of its 1984 level. However, other important driv- ers were also behind the increase in export diversification in 25 the late 1980s and early 1990s. First was the slump in com- modity prices. Following the collapse of oil prices in 1986, 20 exports of resource commodities were pushed back by the collapsing price of crude oil, which dropped to an average US$, billions 15 of US$14.35 per barrel in 1986 from US$36.80 per barrel in 1980. Second, the government imposed measures prevent- 10 ing exports of raw agricultural materials, such as logs and rattan, which made exporting raw materials less attractive 5 and helped wood-based manufacturing activities. However, the increase in market diversification was not as fast as the increase in product diversification. To see how 0 concentrated Indonesia's exports were in certain markets 70 72 74 19 6 78 80 82 19 4 86 88 90 92 94 96 7 8 19 19 19 19 19 19 19 19 19 19 19 19 for certain products, we construct the following bar charts year with product and partner country pairs on the horizontal furniture, textiles, clothing, footwear, etc. axis and share in total exports (percent) on the vertical electronics and machinery axis. Between 1980 and 1984, there were only 4,569 prod- agriculture mining uct-country pairs, while in 1990 and 1994 the number expanded to 21,161 pairs. But the more interesting fact was Source: Authors' calculations based on UN Comtrade (database), United Nations, New York, http://comtrade.un.org. Indonesia's complete dependence on one product and a few markets. Between 1980 and 1984, about 26 percent of The reform package of the mid-1980s quickly reduced Indonesia's exports were due to one product, petroleum Indonesia's dependence on oil exports and significantly exported to Japan (figure 17.8, panel a). But reform packages increased the share of manufactures in total exports. Before in the mid-1980s and the drop in the price of crude brought deregulation, private investment in the manufacturing sec- down the importance of oil in Indonesia's total exports. The tor was constrained by bureaucracy that created high costs. importance of oil exports continued to drop as export Those costs included business licenses, a complicated tax growth in footloose and wood-based industries rose dra- regime, and difficulties in obtaining inputs at world prices matically. The share of a single market and a single product 224 Managing Openness Figure 17.7. Concentration of Indonesia's Export Products, 1980­95 3,500 Herfindahl-Hirschman index 3,000 2,500 2,000 1,500 1,000 500 0 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 year Source: Authors' calculations based on UN Comtrade (database), United Nations, http://comtrade.un.org. Figure 17.8. Product and Country Concentration of Indonesia's Exports in the 1980s and 1990s a. Export concentration, 1980­84 b. Export concentration, 1990­94 25 10 20 8 % in total exports % in total exports 15 6 10 4 5 2 0 0 1 4,569 1 21,161 number of country and product pairs number of country and product pairs Source: Authors' calculations using UN Comtrade (database), United Nations, http://comtrade.un.org. Note: Calculations are based on trade data up to the five-digit level of the Standard International Trade Classification (SITC). in Indonesia's exports in 1990­94 dropped significantly After the Asian crisis, has Indonesia managed to pro- compared to the share in 1980­84 (figure 17.8, panel b). duce more new products or to sell more to markets that it Ten years later, however, Indonesia's exports were still has not traditionally served? Unfortunately, casual obser- largely dependent on the Japanese market. One of the vation using the same method as above does not produce a reasons was the emergence of liquefied natural gas (LNG), strong conclusion. Indeed, on the one hand, figure 17.9 which quickly became the most important export com- suggests that Indonesia managed to double the number of modity. About 9.4 percent of Indonesia's total exports in country-product pairs, from 14,551 between 1990 and that period were accounted for by LNG exported to Japan. 1994 to 51,631 between 2004 and 2008. We could interpret As for petroleum, 8.7 percent of Indonesia's total exports this as a significant achievement in export diversification were due to exports of that product to Japan. Meanwhile, because more products are sold to more countries. China, plywood products were already Indonesia's third-largest India, and Malaysia are increasingly important markets for export commodity, and most of it went to the Japanese Indonesia's exports, taking over from the European and market. U.S. markets. The figure also indicates, however, that the Should Indonesia Say Goodbye to Its Strategy of Facilitating Exports? 225 Figure 17.9. Pairs of Markets and Products in Indonesia's Figure 17.10. Index of Export Product Concentration Exports, 2004­08 in Indonesia, 2001­08 6 1,200 Herfindahl-Hirschman index 1,000 % in total exports 4 800 600 2 400 200 0 0 01 02 03 04 05 06 07 08 1 51,631 20 20 20 20 20 20 20 20 number of country and product pairs Source: Authors' calculations based on data from UN Comtrade, Source: Authors' calculations using UN Comtrade (database), United United Nations, http://comtrade.un.org. Nations, http://comtrade.un.org. Note: Calculations are based on trade data up to the five-digit level of the Standard International Trade Classification (SITC). role of a few markets and a handful of products remain firms can no longer use the provisions of the Multi-Fiber important. About 14.2 percent of Indonesia's total exports Arrangement to secure the value of their textile and in 2004 and 2008 were due to LNG, petroleum, and CPO clothing exports and have to compete head to head with sold to China, India, Japan, and Korea. In fact, 5.8 percent Chinese products in the U.S. market. Meanwhile, local and 2.8 percent of Indonesian exports in 2004 through manufacturing firms have been concerned about the 2008 were due to LNG and petroleum exports, respectively, potential adverse impact of bilateral and regional free trade for the Japanese market. The period also marked the emer- agreements, especially with large developing economies gence of CPO as an important export commodity. CPO such as China. exports to India made up about 1.8 percent of Indonesia's Another possible explanation is that supply-side prob- total exports in that period. lems are affecting the competitiveness of Indonesia's manu- The process of diversifying export products has been facturing firms. Indonesian manufacturing firms may showing slow progress in recent years. Figure 17.10 pres- currently face greater global competition while equipped ents a Herfindahl-Hirschman Index of product concentra- with worse physical and regulatory infrastructure compared tion of Indonesia's exports between 2000 and 2008. The to other middle-income economies. Indonesia's relatively figure suggests that concentration has not significantly rigid labor law could affect the ability of private businesses to declined since 2003. One possible explanation for the set up larger manufacturing operations. Uncertainties in relatively flat concentration index is a lack of "discovery"-- regulation could also reduce the expected return from that is, new export products--in Indonesia's export port- investing in sectors with low economic rent. Indonesian folio. As we will show later, the data suggest that either not manufacturing firms also contend with worse transporta- many new export products are coming out of Indonesia or tion infrastructure relative to those in other emerging that exports of new products have not been able to fully economies.7 Because discovering a new exportable product is take off. The other possible explanation is that the rising a costly process, firms will not undertake such activity unless exports of agriculture and resource commodities are off- they are rewarded with adequate economic rents (Lederman setting the declining performance of Indonesia's "tradi- and Klinger 2004; Rodrik 2004). In that context, Indonesia tional" exports of labor-intensive manufactures, such as also lacks effective public institutions, such as an export textiles, clothing, and footwear. credit agency, private equity or venture capital, and a public Nevertheless, the cause behind the declining per- research and development (R&D) institute, all of which can formance of manufacturing exports in recent years is an provide manufacturing firms with the incentive to pursue interesting phenomenon that researchers have not fully discovery of new exportable products. Finally, coordinating explained. Popular reasons typically center on market coherent policy making is often difficult in a fully decentral- access issues, as Indonesian firms are facing much tougher ized government system and in a noisy democracy. competition from products made in China and Vietnam, Interestingly, the declining performance of Indonesia's both at home and in third markets. For example, Indonesian manufacturing exports coincides with an appreciation in 226 Managing Openness the real exchange rate and rising exports of resource-based where those products have not been sold before; (b) commodities. Figure 17.11 indicates that the (one period exports of old products to new markets; (c) exports of new lagged) real exchange rate is positively correlated with an products (never produced before) to existing markets; and increase in commodity exports and a decline in exports of (d) exports of new products to new markets. manufactures. It is possible that all these coincidences Decomposing export growth between 1990 and 2008 are symptoms of a mild case of "Dutch disease." Because yields interesting results. About 71 percent of the increase Indonesia is one of the largest exporters of minerals, rubber, in Indonesia's exports in the past 18 years was due to and crude palm oil, it has some market power to influence growth in the same set of products sold to the same mar- the world price. Increases in the prices of commodities raise kets. Given this finding along with the result of decompos- export revenues and increase pressure for a real apprecia- ing export growth in East Asia and the Pacific (Brenton and tion of the rupiah. Meanwhile, an increase in the price of Newfarmer 2007), Indonesia may well be relatively less commodities could also increase factors of production dependent on the intensive margin. Nevertheless, this find- intensively used in commodity sectors, either labor or cap- ing stresses the importance of maintaining the export per- ital, and squeeze the profitability of traditional manufac- formance of existing products to existing markets for turing sectors that are facing competitive world prices and Indonesia. the strengthening of the rupiah. The other interesting findings relate to the extensive Given the success in diversifying export products in the margin. First is the significant role of "switching" pairs 1990s and the increase in new export markets in recent between existing markets and products in driving new years, has Indonesia managed to rely on selling new prod- exports. As Indonesia successfully exported certain prod- ucts to new markets? Or is it that exports have largely ucts to certain markets, other markets that Indonesia depended on the same "old products" sold to "old mar- already served started demanding those products. A simple kets"? To answer this question, we turn to an analysis of the example is the demand for footwear from China in 2008, intensive and extensive margins of Indonesia's exports as Indonesia already exported footwear to the United (figure 17.12). The intensive margin is defined as exports States and petroleum to China in 1990. These switching going to the same set of products and markets. We use activities accounted for 26.5 percent of growth in Indonesia's the approach introduced by Zahler (2007) to decompose exports. Meanwhile, the rapid increase in Indonesia's the extensive margin into the following four compo- exports to new markets is apparently still relatively small nents: (a) exports of old products to existing markets compared to the intensive margin or switching between Figure 17.11. Relationship of Indonesia's Commodity Exports, Manufacturing Exports, and Real Exchange Rate, 2000­08 70 160 share of products in total exports (%) 60 140 1 period lag in RER index 120 50 100 40 80 30 60 20 40 10 20 0 0 00 01 02 03 04 5 06 07 08 0 20 20 20 20 20 20 20 20 20 year primary and resource-based products TCF and low technology medium and high technology 1 period lag in RER index (right axis) Source: Authors' calculations based on data from: Indonesia Financial Statistics, Bank Indonesia, http://www.bi.go.id/web/en/Statistik/Statistik+Ekonomi +dan+Keuangan+Indonesia/; and UN Comtrade, United Nations, http://comtrade.un.org. Note: Classification is based on Lall (2000). RER = real exchange rate, TCF = textiles, clothing, and footwear. Should Indonesia Say Goodbye to Its Strategy of Facilitating Exports? 227 Figure 17.12. Contribution of Intensive and Extensive Margins to Indonesian Export Growth, 1990­2008 80 70 60 50 contribution (%) extensive margin: `'new exports'' 40 30 20 `'discovery'' 10 0 ­10 ar ive ke o ke o ke o ke o uc d ar t ar t ar t ar t od he n ts ts ts ts ts m cts m cts m cts m cts m ns pr anis gi te g u w du g u w du in od in od in v ne pro ne pro ist pr ist pr ex ld d ex w w ol ne ne o Source: Authors' calculations using UN Comtrade (database), United Nations, http://comtrade.un.org. pairs of existing products and markets. Exports of existing This argument is extremely relevant for Indonesia, because products to new markets comprised only 2.9 percent of the the choice of an inward-looking strategy is not the increase in total exports between 1990 and 2008. Finally, choice that will provide the best results for the economy. "discovery" plays little part in Indonesia's exports. Exports We therefore argue that the trade reforms that Indonesia of new products, either to existing or to new markets, con- has carried out up to now and the choice of an export-led tributed only 2 percent of the increase in total exports in growth strategy have to be accompanied by policies to that period. support export diversification. The government needs to implement several specific policies to minimize export volatility due to shocks such as the global financial crisis. How to Promote Export Diversification: The Role of Policy Development of the Financial Sector Experiences in Indonesia and other countries suggest the important role that policy can play in supporting export Agosin, Alvarez, and Bravo-Ortega (2009) show a positive diversification. Agosin, Alvarez, and Bravo-Ortega (2009) statistical relationship between financial sector develop- use a large data set for a 40-year period to show that export ment and export diversification. This finding is in fact con- diversification is influenced by several factors such as trade sistent with the argument of Rodrik (2007) that there is an openness, financial development, the real exchange rate, uncertainty problem in producing new products. Some human capital accumulation, and trade costs determined by upfront investments are needed to solve this problem, but remoteness and terms of trade. In addition, Agosin, Alvarez, those investments must be financed. Unfortunately, com- and Bravo-Ortega (2009) and di Giovanni and Levchenko, mercial banks cannot fully deal with this problem, and the (2009) argue that openness will induce specialization and corporate debt market, equity markets, and private venture not necessarily a higher diversification of exports. As a capital funds are not likely to finance innovative invest- result, as indicated by Haddad, Lim, Munro, Saborowski, and ments (Rodrik 2007). Other sources of longer-term financ- Shepherd in chapter 11 of this volume, openness must be ing, such as development banks or venture funds, must be accompanied by policies to support export diversification. found. In Indonesia itself, financing still largely relies on 228 Managing Openness banking sources. The Indonesian government must reform regions in Indonesia. "Trade logistics"--the capacity for the financial sector so that insurance, the capital market, integrating domestic economies and connecting domestic pension funds, venture funds, and development banks can economies with international markets through goods dis- finance new products to promote export diversification. patch--is an extremely important factor in the potential for economic growth of a country. We therefore argue that Indonesia will obtain support for self-discovery and export Better Connectivity to Reduce Trade Costs diversification if logistics and infrastructure problems are Agosin, Alvarez, and Bravo-Ortega (2009) show that a dealt with. Although progress certainly has already begun reduction in trade costs will support new export production in the logistics and infrastructure development system in activities. Hausmann and Rodrik (2003) argue that entre- Indonesia, it is still moving extremely slowly. Such delay is preneurs have to continue experimenting with new products connected with land clearance problems, complicated bid- and adapt foreign technology to local needs, a process ding processes, and lack of coordination between govern- referred to as self-discovery. However, there is a problem: if ment institutions in solving infrastructure problems. entrepreneurs fail in these experiments, they will absorb all losses, while if they succeed, other entrepreneurs will start Primary Products and Real Exchange-Rate Appreciation to imitate and enter into these same activities. As a result, in practice no one is interested in self-discovery. Self-discovery Carrère, Strauss-Kahn, and Cadot (2007) show a relatively means that there are rents to be provided to entrepreneurs, high export concentration in countries that export pri- leaving insufficient incentive to carry out innovation in mary products. High concentration in the export of new products. Trade costs that arise because of logistics primary goods carries with it some risk of Dutch disease. costs or high transaction costs will also be a disincentive for We showed above that in the past Indonesia was extremely entrepreneurs to carry out diversification and develop new dependent on oil products and particular destination products, including those for export. countries. At the end of the 1970s, for example, oil price This problem is relevant for Indonesia, because in an increases and the dependence of the Indonesian economy archipelago nation transaction costs and especially logistics on oil brought with them the Dutch disease (Pangestu costs are relatively higher than in continental countries. 1986; Warr 1992). Transport costs, inventory costs, and the need for interisland It is interesting to see the changes that happened in shipping are larger. A study by LPEM (2005) shows that Indonesia after the increase in commodity prices in 2000. logistics costs in Indonesia amount to 14 percent of total The commodity boom that happened after that year had production costs, far higher than in Japan, for example, already raised credit growth in many provinces in Sumatra, where they reach only 4.9 percent. Patunru, Nurridzki, and Kalimatan, and Sulawesi (see figure 17.4). Symptoms of a Rivayani (2007) show that inefficient harbors make trans- mild case of Dutch disease were present: the boom in the port more expensive, especially in export-oriented and mining and plantation sectors had already supported import-based industries. Meanwhile, a study by LPEM- expansion, and then there was a spending effect outside of FEUI and Asia Foundation (2009) shows that the combina- Java. In parallel was growth in the labor-intensive textile tion of complicated regulations and high domestic transport and leather products sectors, whose growth had declined costs has already disrupted Indonesia's competitiveness. over the previous several years. Consistent with these other For several export commodities such as cocoa, rubber, and developments, the real exchange-rate value continued to coffee, more than 40 percent of total logistics and transport appreciate (Soesastro and Basri 2005; Athukorala 2006; costs are incurred before dispatch and in land transport, not Basri and Patunru 2006). As a result, competiveness in yet including international dispatch costs (Carana 2004). labor-intensive sectors seemed to decline. Consistent with With high logistics costs, only products with a high profit that fall, the index of Indonesia's export concentration margin can be exported, and this factor is especially pro- remained relatively flat, suggesting slow progress in diversi- nounced among primary products, in particular in a com- fying export products. We therefore argue that to support modity boom. High logistics costs lower the profit margin export diversification, Indonesia must maintain its compet- and provide insufficient incentives for creating new products itiveness in exchange rates. Of course, under a flexible in the manufacturing sector or other sector where the profit exchange rate the nominal exchange rate cannot be depreci- margin is small because of extreme competition. As a result, ated; however, Indonesia can maintain its real exchange- Indonesia's effort to diversify its exports has been hampered. rate competitiveness by curbing inflation. In addition, labor In this sort of situation, logistics may be the key that will market reform is also extremely important for avoiding unlock the door to more export diversification for many Dutch disease and for supporting export diversification. Should Indonesia Say Goodbye to Its Strategy of Facilitating Exports? 229 The Role of R&D and Improvement of Product Quality in the Asia-Pacific region will be experiencing aging prob- lems, while Indonesia will have a demographic dividend As discussed earlier, the diversification that occurred in with a high proportion of residents of productive age. In Indonesian exports was driven more by the intensive mar- this situation, the demand for working-age professionals gin than the extensive margin. The ability to create new will increase. If Indonesia increases the quality of its human products for new markets as well as for old markets was resources, it will have the opportunity to become a supplier relatively small. Supporting new product lines requires of skilled labor to the region. For this result, investment in R&D support. Rodrik (2007) states that technology cannot human capital, such as education and health, becomes very be obtained from advanced countries only. Therefore, the important. Unfortunately, up to now, Indonesia has lagged need for R&D financed by the public has to be identified behind in health and education (Woo and Hong 2010). The and adapted with transfer of technology from developed development of these services sectors will make Indonesian countries. Woo and Hong (2010) believe that Indonesia exports become more diversified. Up to now, attention paid should move toward a science-based economy. Indonesia to these services sectors in Indonesia has been small. needs R&D in agriculture, for example, new crop varieties (including agro-biotechnology), a new approach to the management of water and the environment, mechaniza- Promotion and Marketing tion, improvements in superior seeds for agricultural prod- Rose (2007) shows a positive relationship between foreign ucts, infrastructure to support agriculture, and so forth. In missions (consulates and trade promotion agencies) and the case of agriculture, the World Bank (2009) has shown exports. Furthermore, the first foreign mission has a larger that expenditure for agriculture has already increased sig- effect on the increase in exports than subsequent missions. nificantly (in real terms) but that production is still low. Therefore, we see that promotion, marketing, and trade Production in major crops has been stagnant since 1990 promotion organizations are important for supporting and is relatively low compared with other Asian countries. export diversification. LPEM-FEUI (2008) shows that the World Bank (2009) also points out that the ratio of majority of world exports come from products manufac- Indonesia's agricultural research intensity is relatively low.8 tured for markets with the characteristic of monopolistic In the future, to support diversification in exports and competition. In a monopolistically competitive market, the especially to increase the extensive margin, improvements roles of nonprice variables such as advertising, promotion, in R&D are required. This R&D is also needed to improve and information are important. Until now, Indonesian product quality, so that better-quality products can satisfy exports have been supported by the results of the primary the demands of the new export markets. sectors (agriculture and mining). If Indonesia wants to pen- etrate monopolistically competitive markets, then exporters must be capable of competing not only on price but also on The Development of the Services Sector nonprice factors. One of the problems facing Indonesia is As pointed out by Brenton, Newfarmer, and Walkenhorst the asymmetric information between domestic producers (2007), the services sector is important for achieving and international consumers because of the lack of infor- export diversification. The services sector constitutes mation about global demand, as well as the relative lack almost half the Indonesian economy. Links from the serv- of promotion of Indonesian products. LPEM-FEUI (2008) ices sector are also very strong compared to other sectors showed that marketing and promotion problems were one (van Diermen, Basri, and Shan, forthcoming). Indonesia of the main factors that delayed exports. Respondents has several sectors that have great potential for the future, supported the introduction of trade policies to overcome such as tourism and the movement of service providers this problem. Some 70 percent of respondents stated that (natural persons). Tourism is a sector with potential for promotional agencies were important. LPEM-FEUI 2008 growth and absorption of manpower. The cultural and also indicated that trade, tourism, and FDI were sources of geographical diversity of the Indonesian archipelago offers foreign reserve and should be connected naturally. Thus, to a range of trade in tourism opportunities. Particularly support export diversification, a promotion agency needs to with increased stability, the potential exists for substantial be integrated with these three agencies. growth in tourist numbers (van Diermen, Basri, and Shan, forthcoming). Conclusion In addition, the movement of natural persons (mode 4 in the General Agreement on Trade in Services, [GATS]) has The above discussion shows that the effects of the global great potential for Indonesia. By 2025, many of the countries crisis in 2008 on the Indonesian economy were relatively 230 Managing Openness small, when compared with several other countries in reducing dependency on primary exports, safeguarding a Southeast Asia. The main reason for the relatively small competitive exchange rate (preventing Dutch disease), rais- effects on Indonesia's economic growth was the limited ing R&D and product quality, and increasing the role of the openness in the Indonesian economy. This result gives services sector as well as improvements in promotion and rise to the question whether Indonesia should give up its marketing. If these policies are implemented, then Indonesia strategy of facilitating trade. Our answer to this is that can maintain an export-led growth strategy while still sup- the strategy of facilitating exports is still the right choice porting domestic consumption. for Indonesia as long as it is accompanied with export diversification. Notes Our study shows that Indonesia can do much more to 1. For the details of the impact of the global financial crisis on the diversify its exports. It is true that the reforms in the 1980s Indonesian economy, see Basri and Siregar (2009); Gunawan and Siregar dramatically increased Indonesia's export product diversifi- (2009); Gunawan et al. (2009); Basri and Rahardja (2010); Papanek, Basri cation: that set of policies lowered barriers to entry, improved and Schydlowsky (2010). 2. This section is summarized and drawn from Basri and Rahardja trade facilitation, and reduced bureaucratic inefficiencies, (2010). thus unlocking business opportunities in Indonesia's non-oil 3. Total export of goods and services in national account as a percent- sectors. As a result, Indonesia became one of the platforms age of GDP. 4. The Arab oil embargo caused the average price of crude oil to jump for footloose manufacturing industries. The process was from US$2.80 per barrel in 1973 to US$11.00 per barrel in 1974 and also accompanied by a global quota arrangement on tex- finally to hit US$36.90 per barrel in 1980. tiles and clothing, measures discouraging exports of raw 5. The government had to devalue the rupiah in 1978 to avoid a total agricultural commodities, and relatively low global com- loss of competitiveness. 6. However, there are reasons to be cautious in imposing the duty modity prices that made exporting manufactured products drawback system on exporters. Cadot, de Melo, and Olarreaga (2003) more attractive than exporting commodities. Our findings demonstrated that if the industry actively lobbies for the policy, the duty also suggest that Indonesia has increased exports of exist- drawback can result in slower trade liberalization because exporters would have lower incentives to prevent high tariffs on their input. ing products to some new markets, part of the extensive Another popular objection to the duty drawback system is the cost of margin. However, we also find that the discovery of new administering the system, particularly the potential for redundancy if products and exports to new markets are still quite low. import tariffs are further lowered. 7. See Basri and Rahardja (2010) for comparisons. Recent efforts to diversify exports, both markets and 8. This ratio is defined as the ratio of public agricultural R&D expen- products, have become increasingly more challenging. Com- ditures to agricultural output or agricultural gross domestic product. petitiveness issues are constraining Indonesian manufactur- ers and limiting their returns as they face an increasingly Bibliography competitive global market. We also believe that the boom in commodity prices and the appreciation of the rupiah real Agosin, M. R., R. Alvarez, and C. Bravo-Ortega. 2009. "Determinants of Export Diversification around the World 1962­2000." Department of exchange rate have increased the opportunity costs of invest- Economics Working Paper, University of Chile, Santiago. ing outside commodity-related businesses and have lowered Athukorala, P. 2006. "Post-Crisis Export Performance: The Indonesian the margins for manufactures to compete globally. Mean- Experience in Regional Perspective. Bulletin of Indonesian Economic Studies 42 (2): 177­211. while, insufficient capacity and lack of effective public insti- Basri, M. C., and A. A. Patunru. 2006. "Survey of Recent Developments." tutions for facilitating R&D and new exports could limit the Bulletin of Indonesian Economic Studies 42 (3): 295­319. incentive for manufacturers to engage in new product dis- Basri, M. C., and S. Rahardja. 2010. "The Indonesian Economy amidst the covery. As a consequence, Indonesia has depended more on Global Crisis: Good Policy and Good Luck." ASEAN Economic Bulletin 27 (1): 77­98. exports of old products to existing markets as the main Basri, M. C., and R. Y. Siregar. 2009. "Navigating Policy Responses at the driver of export growth. National Level in the Midst of the Global Financial Crisis: The Experi- For these reasons, the relatively large decline in Indone- ence of Indonesia." Asian Economic Papers 8 (3): 1­35. Brenton, P., and R. Newfarmer. 2007. "Watching More than the Discovery sian exports--comparable with other countries in the Channel: Export Cycles and Diversification in Development." Policy Association of Southeast Asian Nations--did not translate Research Working Paper 4302, World Bank, Washington, DC. into a major set of economic effects during the global Brenton, P., R. Newfarmer, and P. Walkenhorst. 2007. "Export Diversifi- cation: A Portfolio Approach." Paper Presented at the Growth Com- financial crisis. Indonesia's experience in the past, when mission Conference on Development, Yale University, September. dependence on oil commodities was large, also highlights Cadot, O., J. de Melo, and M. Olarreaga. 2003. "The Protectionist Bias of the risks for the Indonesian economy. In the future, there- Duty Drawbacks: Evidence from MERCOSUR." Journal of Interna- fore, Indonesia has to come up with policies to diversify tional Economics 59 (1): 161­82. Carana Corporation. 2004. "Impact of Transport and Logistics on Indone- exports. These policies include improving the financial sia's Trade Competitiveness." Paper for review by U.S. Agency for Inter- sector, strengthening the logistics system or connectivity, national Development, Jakarta. Should Indonesia Say Goodbye to Its Strategy of Facilitating Exports? 231 Carrère, C., V. Strauss-Kahn, and O. Cadot. 2007. "Export Diversification: Papanek, G., M. C. Basri, and D. Schydlowsky. 2010. "The Impact of the What's behind the Hump." Mimeo. http://r0.unctad.org/ditc/tab/ World Recession on Indonesia and an Appropriate Policy Response: events/emstrade/Speakers/Cadot_doc.pdf. Some Lesson for Asia." In Poverty and Sustainable Development in di Giovanni, J., and A. A. Levchenko. 2009. "Trade Openness and Volatil- Asia: Impacts and Responses to the Global Financial Crisis, ed. Armin ity." Review of Economics and Statistics 91 (3): 558­85. Bauer and Myo Thant, 13­50. Manila: Asian Development Bank. Fane, G., and T. Condon. 1996. "Trade Reform in Indonesia, 1987­1995." Patunru, A. A., N. Nurridzki, and Rivayani. 2007. "Port Competition in Bulletin of Indonesian Economic Studies 32 (3): 33­54. Indonesia." Paper prepared for Asian Development Bank Institute. Gunawan, A. H., H. Arman, and A. Hendranata. 2009. "Indonesia 2009 Forthcoming as "Port Competitiveness: A Case Study of Semarang Economic Outlook: Slowing, Not Falling." PT Bank Danamon Indone- and Surabaya, Indonesia." In Infrastructure's Role in Lowering Asia's sia Tbk. Jakarta, January 7. Trade Costs: Building for Trade, ed. D. Brooks and D. Hummels. Gunawan, A. H., and R. Y. Siregar. 2009. "Survey of Recent Develop- Cheltenham, U.K.: Edward Elgar. ments." Bulletin of Indonesian Economic Studies 45 (1): 7­36. Rodrik, D. 2004. "Industrial Policy for the Twenty First Century." Dis- Hausmann, R., and D. Rodrik. 2003. "Economic Development as Self-dis- cussion Paper 4767, Centre for Economic Policy Research, covery." Journal of Development Economics 72: 603­33. London. Hill, H. 1996. The Indonesian Economy since 1966. Cambridge, U.K.: ------. 2007. One Economics, Many Recipes: Globalization, Institutions, Cambridge University Press. and Economic Growth. Princeton, NJ: Princeton University Press. Lall, S. 2000. "The Technological Structure and Performance of Develop- Rose, A. 2007. "The Foreign Service and Foreign Trade: Embassies as ing Country Manufactured Exports, 1985­98." Oxford Development Export Promotion." World Economy 30 (1): 22­38. Studies 28 (3): 337­69. Soesastro, H. M. 1989. "The Political Economy of Deregulation in Lederman, D., and B. Klinger. 2004. "Discovery and Development: An India." Asian Survey 29 (9): 853­69. Empirical Exploration of New Products." Policy Research Working Soesastro, H., and M. C. Basri. 2005. "The Political Economy Paper 3450, World Bank, Washington, DC. of Trade Policy in Indonesia." ASEAN Economic Bulletin 22 (1): LPEM (Institute for Economic and Social Research). 2005. "Ineffi- 161­78. ciency in the Logistics of Export Industries: The Case of Indonesia." van Diermen, P., M. C. Basri, and E. Shan. Forthcoming. "Trade, Aid and Report in collaboration with Japan Bank for International Cooper- Development: An Analytical Paper." Report prepared for Aus Aid, ation, Jakarta. Indonesia Branch. LPEM-FEUI (Institute for Economic and Social Research­Faculty of Eco- Warr, P. G. 1992. "Exchange Rate Policy, Petroleum Prices, and the Balance nomics, University of Indonesia). 2008. "Studi Lanjutan Komisi Perda- of Payments." In The Oil Boom and After: Indonesian Economic Policy gangan Luar Negeri & Analisis Kelayakan Lembaga Promosi Indonesia" and Performance in the Soeharto Era, ed. A. Booth, 132­56. Kuala [Study on the Foreign Trade Commissioner and Indonesian Promotion Lumpur: Oxford University Press. Council]. Unpublished report to the Ministry of Trade, Jakarta. World Bank. 2009. "Revitalizing Agriculture in Indonesia." Policy Notes, LPEM-FEUI (Institute for Economic and Social Research­Faculty of Eco- World Bank, Washngton, DC. nomics, University of Indonesia) and Asia Foundation. 2009. The Cost Woo, W. T., and C. Hong. 2010. "Indonesia's Economic Performance in of Moving Goods, Transportation, Regulations and Charges in Indone- Comparative Perspective and a New Policy Framework 2049." Bulletin sia. Jakarta: Asia Foundation. of Indonesian Economic Studies 46 (1): 33­64. Pangestu, M. E. 1986. "The Effects of an Oil Boom on a Small Oil Export- Zahler, A. 2007. "Decomposing World Export Growth and the Relevance ing Country: The Case of Indonesia." Ph.D. diss. University of Califor- of New Destinations." Working Paper 20, Center for International nia Davis. Development, Harvard University, Cambridge, MA. 18 India: Managing Openness for a Rapidly Developing Domestic Market Ulrich Bartsch and Abhijit Sen Gupta India joined the elite group of fast-growing economies macroeconomic management with an ever more open cap- in the 2000s, when its gross domestic product (GDP) ital account. The first section gives some background infor- expanded by 9 percent per year during 2002­07 (figure 18.1). mation on the Indian growth experience, which differed Since the early 1990s, India has been rapidly opening up from other countries in that the services sector played a to trade and capital flows, but because the opening much larger role than in countries with comparable per started from a very low level, India's economy remains capita incomes. relatively closed, especially compared to East Asian coun- The second part is devoted to an analysis of the role of tries. Indicators of economic openness increased along exports in India's economic growth by looking at factor with those of overall economic growth, as promarket intensities of exports overall and to specific country reforms, chipping away at the elaborate edifice of the groups. This approach highlights the importance of the "license raj," unshackled entrepreneurs after the 1991 "old" countries of the Organisation for Economic Co- balance-of-payments crisis. operation and Development (OECD)--Europe, Japan, and The recent growth spurt seems to have taken only a the United States--for exports of products that embody short pause during the global financial crisis, when India's skills and technology and thereby provide impetus to fac- growth slowed down significantly but the economy tor productivity growth in India. Slow growth in the old avoided a recession. In the first half of 2010, signs emerged OECD with the lingering effects of the global crisis could that India is returning to precrisis growth rates. India was dampen the export demand crucial for economic growth. affected by the shocks of 2008--first the commodity price However, with India not easily fitting into the world of boom and then the global financial crisis--along with small open economies most often described by theorists, most other countries in the world, but showed remark- the domestic market seems to be no less demanding of able resilience. In fact, with Indian banks unaffected by sophisticated goods than the export market. Exports, the toxic asset crisis, India faced some limited capital therefore, can be expected to play a less prominent role in outflows and a big decline in demand for its exports, but India than in some smaller emerging-market countries. the main impact of the crisis was psychological: con- The third section analyzes challenges to macroeco- sumers and businesses took fright and postponed some nomic management posed by volatile capital inflows. The decisions. section shows the loss of monetary policy independence in After the crisis, policy makers turned attention to the the face of massive capital inflows in 2006­07 and during requirements of fast growth for the medium term. In earlier such episodes. It highlights the importance of a the context of this book, we discuss two aspects: first, the careful approach to capital account liberalization and poli- prospects for fast growth in exports and their impact on cies that can limit the economy's vulnerability to the most factor productivity growth; and, second, the challenges of volatile types of capital flows. 233 234 Managing Openness Figure 18.1. GDP Per Capita Growth in India, 1970­2009 9.0 8.0 5-year moving average change (%) 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 ­1.0 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 year Source: Central Statistics Office database, http://www.mospi.gov.in/cso_test1.htm. India's Recent Growth Experience fiscal 2004­05. Growth accounting, therefore, shows that accumulation of capital increased significantly starting The recent growth spurt of the Indian economy is remark- in the late 1990s, but total factor productivity (TFP) also able. Over three decades, India's anemic economic growth shifted into higher gear (see Bosworth, Collins, and rate had gained notoriety as the "Hindu rate of growth." Virmani 2007). Changes began in the 1980s and accelerated after a bal- In contrast to fast-growing emerging economies in East ance-of-payments crisis in the early 1990s. Broad reforms Asia, India's manufacturing sector did not spearhead eco- aimed at dismantling the so-called license raj freed up the nomic growth. As in other developing countries, the agri- dynamism of India's entrepreneurs, and economic growth culture sector's share in production declined. The share of accelerated significantly. At the same time, the rise in per industry remained relatively stable, however, while the capita incomes was also helped by declining population services sector showed the most dynamism (figure 18.2). growth. Per capita growth averaged only 0.7 percent From a share of 38 percent of GDP in 1965, the services between the early 1950s and the end of the 1970s. It accel- sector produced 55 percent of GDP in fiscal 2009­10. erated to 3.3 percent in the 1980s and 1990s. However, in During the same period, the share of agriculture declined the five years 2002­07 average real GDP grew at almost from 41 percent to 17 percent of GDP. The industrial sec- 9 percent, and per capita growth at more than 7 percent. At tor increased its share only from 21 percent to 28 percent this rate, India's economic growth was the second highest of GDP. The high share of the services sector in GDP of any major economy in the world, behind only China's. A means that the structure of India's GDP resembles that of sustained performance at this rate would double real high-income countries more than that of countries at incomes every 10 years. similar levels of per capita income. High growth was driven by a significant shift in domes- tic saving and investment, but the trade balance remained negative in almost all years. Private saving as a share of India's Resilience during the Global Crisis GDP languished below 20 percent until the beginning of the 1990s. From 24 percent in fiscal 1996­97, however, it India was hit by the commodity price boom before the increased steadily and reached 32 percent in fiscal global financial crisis, which peaked in the summer of 2009­10.1 Private investment oscillated around 18 percent 2008. The terms-of-trade shock--mainly from rising oil of GDP until fiscal 2004­05, when it jumped to 24 percent prices--is estimated at 2.5 percent of GDP. Rising com- and continued climbing to nearly 28 percent in fiscal modity prices led to higher inflation in India, although 2007­08. The share of exports and imports in GDP crossed the government did not fully pass on higher oil prices into double digits in the mid-1990s and nearly doubled by to Indian consumers and increased fuel and fertilizer India: Managing Openness for a Rapidly Developing Domestic Market 235 Figure 18.2. Composition of GDP in India, 1965­2009 100 90 80 70 sectoral shares (%) 60 50 40 30 20 10 0 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05 07 09 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 year agriculture industry services, etc. Source: Central Statistics Office database, http://www.mospi.gov.in/cso_test1.htm. subsidies instead. Wholesale price inflation nevertheless In the immediate aftermath of the Lehman Brothers peaked at nearly 13 percent in August 2008, an almost collapse, financial sector liquidity dried up and caused a 10-percentage-point increase from the beginning of the sharp spike in interbank lending rates. As in other coun- year. Monetary policy reacted to higher inflation with tries, the Reserve Bank of India (RBI) scrambled to restore higher interest rates and prudential requirements: the cash liquidity to the system. In a step-wise manner over a period reserve requirement was raised by 150 basis points and the of six months, it lowered prudential ratios and policy inter- repurchase rate by 125 basis points between January and est rates and directly injected liquidity through open mar- September 2008. With these actions, policy rates reached ket operations. According to its own estimates, the different the highest level since the beginning of the decade, liquidity operations injected rupees worth about 9 percent although the tightening obviously trailed inflation and of GDP into the financial system. Interbank rates fell sig- was relatively modest. nificantly and stabilized by November 2009. Expansionary At the time of the global financial crisis, India's econ- fiscal policy, however, set in at the same time as the Lehman omy suffered a significant slowdown in growth. However, collapse, was eroding confidence because of India's elec- India did not suffer a recession. In the fourth quarter of toral cycle: populist policies of farm debt waivers and 2008, GDP growth was more than three percentage points higher civil servant salaries greatly supported private con- lower than the five-year average before the crisis. But at its sumption demand in the last quarter of 2008, when the lowest point during the crisis, growth was still higher than economy was hit by a significant slump in investment in any major developed country during the boom period demand and destocking.2 Three explicit stimulus packages of the early 2000s. From a high of 10.1 percent in the sec- added to government demand by lowering taxes and ond quarter of fiscal 2006­07, GDP growth (at factor cost) increasing spending, for example, on the National Rural slumped to 5.8 percent in the fourth quarter of fiscal Employment Guarantee Schemes. Overall, the general 2008­09, and the latest available data show a recovery to government fiscal deficit reached 8.8 percent of GDP in 8.6 percent in the last quarter of fiscal 2009­10. The down- fiscal 2008­09, a widening by 3.8 percent of GDP over turn and recovery are much stronger in the growth rate of fiscal 2007­08. the monthly index of industrial production, which acceler- At first sight, the global financial crisis should have had ated to a peak of 13 percent in January 2007, dropped to only a mild impact on India's economy because of the near zero in December 2009, and recovered to 17 percent in resilience of the country's banking system and relatively May 2010. closed current and capital accounts. First, the banking 236 Managing Openness system faced no exposure to the toxic assets that led to crisis, as imports grew faster than exports. India has diver- widespread financial distress in developed countries. Sec- sified its trading partners, with China becoming increas- ond, India's development model is not that of export-led ingly more important (figure 18.3). While the shares of the growth, as in some other emerging-market countries, most European Union (EU) and the United States in total trade notably in East Asia, although the importance of external have declined over the past 20 years, China captured about transactions, both current and capital, has greatly expanded 11 percent of India's exports and imports in fiscal 2009­10, over the past two decades (see "India's Growth Model" as compared with negligible shares in fiscal 1987­88. below). The percentage of exports in GDP reached the The global financial crisis and its lingering effects have mid-20s before the global crisis, compared with around led to concerns that increasing openness--that is, the 50 percent in many East Asian countries. Moreover, in the share of trade and capital flows in the economy--cannot boom years before the global crisis and the severe recession drive growth the way that it did in many countries before in international trade, imports were growing faster than the global crisis (see, for example, World Bank 2010). exports, and the net contribution of trade to GDP growth According to this line of thought, emerging markets was therefore negative. Finally, India faced a sudden stop in would derive smaller benefits than before from an export- capital flows but avoided large-scale capital flight. Portfolio led industrialization strategy because of the lingering flows rushed into India's equity market before the crisis recession in the OECD countries and higher risk aversion and led to concerns about the competitiveness of domestic by foreign investors. tradables. Nevertheless, some capital fled to safe havens, Economists have argued that trade can be a powerful which the RBI managed by letting the exchange rate depre- driver of factor productivity growth. They point out that ciate and intervening only marginally in the foreign competing in the global marketplace through trade--and exchange markets. The low level of short-term external in particular exporting manufactured products--generally borrowing, the relative resilience of foreign direct invest- requires higher technology and skills than production for ment (FDI), and the confidence-inspiring level of RBI for- the domestic market. Because growth requires the accumu- eign reserves prevented a stronger upset. lation of capital and skills and adoption of better technol- The crisis nevertheless had a strong impact on business ogy to increase factor productivity, trade drives growth and consumer confidence. The main channel of transmis- because it pushes domestic workers to acquire skills, entre- sion of the global financial crisis to India's real economy preneurs to innovate, and international investors to trans- seems to have been psychological: uncertainty over the fer technology through FDI. The corollary to this view is future caused consumers to postpone purchases of durable that production for the domestic market would employ goods and businesses to postpone investment projects. less technology and human-capital-intensive processes This reaction led to the transmission of the global financial because domestic consumers demand less sophisticated crisis to the real economy. However, it did not lead India products than overseas markets. However, we point out into recession. Growth of industrial production slumped below that the production of India's exports does not seem to zero in December 2009 but did not enter negative terri- to employ higher technology and more human capital than tory. Even fixed investment in the national accounts the production for the domestic market. Exports, there- showed a near-zero growth rate, which means capacity fore, may not be as important for growth as they are in continued to expand in the midst of the global crisis at the some other, smaller emerging markets. same level as before the crisis. This view of trade as an engine of growth and structural change stands in marked contrast to the earlier view that countries engage in trade in accordance with their compar- India's Growth Model: Greater Openness but ative advantage. In the case of India, that process would Also a Rapidly Developing Domestic Market mean a specialization in exports of goods that incorporate During the past decade, India's economy has greatly the country's relatively abundant natural resources and opened up to the world. Since 2001, its exports and unskilled labor. While the type of trade based on compara- imports have grown by around 20 percent per year on aver- tive advantage increases demand for goods, it cannot age and amounted to 24 percent and 30 percent of GDP, explain the kind of total factor productivity growth and respectively, in fiscal 2009­10. India has also emerged as structural change observed in the fast-growing countries the world's biggest recipient of worker remittances, which such as the Asian Tigers (Hong Kong SAR, China; the amounted to US$55 billion in fiscal 2009­10.3 However, the Republic of Korea; Singapore; and Taiwan, China). Their net impact of trade on growth--in a national accounting growth relied to a large extent on shifting production into sense--was negative over the five years before the global more sophisticated products. Exports of natural resources India: Managing Openness for a Rapidly Developing Domestic Market 237 Figure 18.3. Direction of India's Trade, 1987­88 through 2007­08 a. Exports 100 80 shares of total (%) 60 40 20 0 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 year b. Imports 100 80 shares of total (%) 60 40 20 0 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 year European Union United States China others Source: Reserve Bank of India. and unskilled labor­intensive goods would not have pro- of four factors: natural resources, unskilled labor, tech- vided incentives for investment in technology and skills, nology, and human capital (skills) using the three-digit while the domestic markets for sophisticated products in Standard International Trade Classification (SITC) (see these countries were too small for production to benefit Pitigala 2010).4 from economies of scale and scope. However, much of the While we can safely assume that the TFP spillover literature discusses the experience of small, open economies, from "sophisticated" exports had some impact on domes- where domestic market size is limited and with it the tic production, exports on the whole do not seem to opportunities for economies of scale and scope and where, employ more technology and skills than home market therefore, international investors have little reason to set up production. In fact, at first glance the data show the oppo- manufacturing capacity if it is not meant to provide for site. The share of such sophisticated processes was actually exports. The question is whether there is evidence that this higher in home production than in exports, with 41 percent analysis also holds true for India. of domestic output coming from technology-and-skills- Exports of human-capital-and-technology-intensive intensive industries compared with 37 percent of exports. manufactures have grown fast in India since 1990. Until However, this comparison rests on somewhat shaky then, more than 80 percent of exports were based on ground. It compares data on trade with data from India's natural resources and unskilled labor, arguably in line Annual Survey of Industry (ASI), although the two use dif- with India's endowments. However, by 2008 this share ferent classifications.5 More important, the total value of had declined to 63 percent (see table 18.1). The data in industrial production shown in the ASI is only about table 18.1 are derived from categorizing India's exports 40 percent of that in the national accounts (ASI data for according to the relative intensity in their production 2008 show value added as Re 4.7 trillion as compared with 238 Managing Openness Table 18.1. Factor Intensities of Indian Exports and Domestic Production, Selected Years, 1962­2008 percentage of total Year Natural resources Unskilled labor Technology intensive Human capital intensive All countries 2008 46.5 16.8 20.3 16.4 2000 39.7 24.7 15.4 14.2 1990 52.1 31.5 7.4 8.9 1970 59.9 27.0 4.5 8.6 1962 62.9 33.3 1.6 2.2 Domestic production 2008 49.6 9.0 25.3 16.0 2004 47.1 9.8 25.9 17.2 Old OECD 2008 36.6 14.2 26.7 22.5 High-income East Asia 2008 64.1 8.7 17.7 9.4 China and Hong Kong SAR, China 2008 83.4 2.9 8.9 4.8 Sources: Authors' calculations based on UN Comtrade (database), United Nations, http://comtrade.un.org. Re 12.7 trillion in the national accounts). At least part of labor­intensive production, respectively. Tobacco consti- the discrepancy would most likely be explained by prob- tutes an impressive 43 percent of what high-income Asian lems in coverage: the ASI probably underrepresents small- countries buy from India, but technology and human scale industry in the informal sector. Because this sector capital­intensive goods make up 27 percent of the total. In accounts for 80 percent of nonfarm employment in India, contrast, nearly 50 percent of Indian exports to old OECD the true share of unskilled labor­intensive production is countries consist of products using relatively more tech- probably much higher than shown in the ASI data. If the nology and human capital.6 "missing" value added in the survey is added to unskilled Growth in exports in the 1990s was predominantly in labor­intensive industries, their share rises to 20 percent, exports of technology and human capital­intensive pro- and sophisticated processes now produce 38 percent of duction. They grew by around 17 percent per year as domestic output, equal to their share in exports. against 9 percent growth for all exports (see table 18.2). In Thus, from the limited data available, we cannot con- contrast, export growth in the 2000s was much more clude that exports are more sophisticated goods than those balanced between sophisticated goods and goods more in produced for the home market. Possibly a finer classifica- line with India's static comparative advantage, natural tion of goods distinguishing between those produced for resources and unskilled labor. In the 2000s, however, serv- exports and those for the home market within categories ices exports with arguably much higher human-capital under the three-digit SITC or International Standard intensity took off with growth of 18 percent per year. Industrial Classification (ISIC) could have shown that tele- The sources of global trade growth provide no strong visions for sale in India, for example, are less sophisticated reasons for export pessimism. In recent years before the than televisions exported by domestic producers. global crisis, high-income country imports have grown It has to be noted, though, that the composition of faster than GDP, driven by differentiation of goods and India's exports differs strongly between regions. India's outsourcing of some elements of production. Developing- exports to Europe, Japan, and the United States (the old country exports, in contrast, have risen faster than global OECD) are fairly diversified, but exports to China and GDP because of continuing economic integration, frag- Hong Kong SAR, China, are narrowly based on just two mentation of production, and specialization in globalized product categories. In its exports to higher-income East production networks. The forces that drove these develop- Asia (Indonesia; Korea; Singapore; Taiwan, China; and ments before the global crisis continue to operate. India Thailand), India provides natural resource­based inputs stands to benefit from these developments, because its to the manufacturing powerhouses, but it is also increas- integration into East Asian production networks is still ingly integrated into production networks for electrical in its infancy and new opportunities arise as China goods, electronics, car parts, and machinery. A large share moves up the value chain. The most hopeful recent of India's exports to China and Hong Kong SAR, China development is the emergence of India as a hub for the (more than 60 percent), consists of mineral ores and production of cars and car components for export. woven fabrics using natural resources and unskilled Hyundai and Suzuki already export an important share India: Managing Openness for a Rapidly Developing Domestic Market 239 Table 18.2. Growth in Exports, Different Factor Intensities, 1990­2008 US$, millions Annual growth rates (%) 1990 2000 2008 1990­2008 1990­2000 2000­08 Export of goods 18,477 45,452 175,184 12.5 9.0 13.5 NR and UL 15,447 29,271 110,891 11.0 6.4 13.3 TN and HC 3,030 16,181 64,293 17.0 16.8 13.8 Export of services 4,551 16,268 101,224 17.2 12.7 19.3 Sources: CSO Balance of Payments, and UN Comtrade. Note: NR = natural resources, UL = unskilled labor, TN = technology, HC = skills. of their India-produced vehicles today, and Ford and the cost of savings--by encouraging higher consumer Nissan will commence shortly. spending--would be risky because it would require India Domestic market opportunities are growing fast. In fact, to rely to a greater degree on uncertain external financing the domestic market was about three times the size of the of the massive investment needs of the country. Conse- export market during the high-growth years before the global quently, a balanced approach that builds on both domestic crisis on average (US$458 billion as compared with US$150 and external demand is needed to provide a solid founda- billion for exports, table 18.3). And even though exports were tion for medium-term growth. growing fast, the additional domestic market demand was higher than additional demand from international markets India's Approach to Capital Account in every year during the 2002­07 period. Over the five years Liberalization on average, domestic demand increased by US$64 billion every year, as compared with US$35 billion for exports. Alongside the opening and expansion of the current The Indian middle class--defined as people living in account, capital account flows also increased massively households with consumption patterns comparable to the after the 1991 balance-of-payments crisis. The higher middle class in the United States based on purchasing- volatility of capital account flows relative to current power-parity exchange rates--was estimated as 5 percent account flows has posed challenges to macroeconomic of the population, or 50 million people in 2005 (see figure management. India witnessed its first surge in capital 18.4).7 At the high rates of growth in per capita income inflows during 1993 and 1994, when net capital inflows observed before the global financial crisis, the middle class increased to more than US$6.7 billion in 1993 and is projected to expand to 43 percent of the population or US$10.6 billion in 1994, from an average of US$3.7 billion 500 million people by 2025. With 50 million consumers, during the previous two years. Capital flows declined in the the market size for many products is currently not big second half of the 1990s for a variety of domestic and enough to allow for economies of scale and scope.8 The external reasons, including the contagion effects of differ- outlook for the next 20 years, however, makes India an ent crises in Latin America and East Asia and sanctions attractive market to invest in, and many international con- after India's nuclear tests in May 1998. Net capital flows sumer goods companies do so. For now, however, many of dropped from US$10.5 billion in 1994 to US$3.8 billion in them content themselves with sales outlets rather than 1995. There was some recovery in 1996 with net capital production bases, because economies of scale allow them flows reaching US$12 billion, but subsequent years saw an to produce cheaper elsewhere. This situation is expected to average of US$9 billion between 1997 and the early 2000s. change, however, in particular because production costs in Net capital inflows picked up after 2002: they first touched East Asia are rising with rising wages. US$12 billion with an improvement in global conditions, It would be risky for India to adopt policies that aim at doubled by 2004, and rose quickly to US$37.7 billion in replacing export growth with faster growth of the domestic 2006 and US$95 billion in 2007. As a result of the onset of market. Such a rebalancing is proposed for countries with the subprime crisis in the United States and the associated current account surpluses such as China to reduce global "flight to safety," portfolio capital started flowing out of imbalances, which are believed to be one of the factors that India in early 2008 and continued until early 2009, contributed to the global crisis. In India, however, with a although FDI flows remained quite robust. Capital flows small deficit in the current account over recent years, a resumed again in early 2009, with India receiving net capi- strategy aimed at increasing domestic demand growth at tal inflows of US$53 billion during fiscal 2009. 240 Managing Openness Table 18.3. Household Consumption and Exports, Level and Increment, 2002­08 US$, billions 2002­03 2003­04 2004­05 2005­06 2006­07 2007­08 Average Level Household consumption 324.1 379.2 416.7 466.1 516.4 643.7 457.7 Exports 73.5 88.7 126.6 160.8 203.1 249.0 150.3 Increment Household consumption 16.6 55.1 37.5 49.4 50.3 127.3 63.9 Exports 12.5 15.3 37.9 34.2 42.3 45.9 35.1 Source: Central Statistics Office National Account Statistics, Balance of Payments Statistics, RBI exchange rates, http://www.mospi.gov.in/cso_test1.htm. Figure 18.4. Income Strata in India, 1985­2025 1 1 2 1 1 2 100 6 4 9 90 18 19 80 share of total population (%) 41 32 70 60 43 50 93 40 80 36 30 54 20 35 10 22 0 1985 1995 2005 2015 2025 year globals strivers seekers aspirers deprived persons middle class Source: Beinhocker, Farrell, and Zainulbhai 2007. Despite the significant changes in the importance of cross-country comparison India is placed at the restrictive external flows, India is still a relatively closed economy end of the spectrum, for both de jure and de facto openness. according to several benchmarks (figure 18.5). In fact, when It has been overtaken by several other countries like Brazil, evaluated on the basis of legal restrictions on cross-border Chile, China, and others, which lagged behind India in cap- capital flows, the openness of India's capital account has ital account openness during the 1980s. remained stagnant since the 1970s (Chinn and Ito 2008).9 In contrast, a de facto measure of openness shows that Macroeconomic Management India's international financial integration progressed signif- and Capital Flows icantly: the ratio of the sum of foreign assets and foreign liabilities to GDP increased from 17.8 percent in 1970 to The increased volatility of international capital flows has 85.4 percent in 2007 (Lane and Milessi-Ferreti 2007). How- rekindled the debate over capital account liberalization. ever, India has not kept pace with the liberalization of the Sharp changes in the direction of capital flows during the past capital account in other emerging markets. As a result, in a two years have presented difficulties for macroeconomic India: Managing Openness for a Rapidly Developing Domestic Market 241 Figure 18.5. Cross-Country Comparison of Capital Account Openness, Selected Countries, 2000­07 a. Lane and Milessi-Ferreti de facto openness average, 2000­07 300 percent 200 100 0 M an ge a a ile n ili ca Th nes nd re ru do . sia lo il a M a o ic a p. In Rep Co az Ar ysi in bi in am di ic ut io Re Ko Pe Ph Afri Ch rd ne la ex nt Ch m In So erat Br i a pp ai Jo al a, h d Fe Isl n n, ia ss Ira country Ru b. Chinn and Ito de jure openness average, 2000­07 3 2 1 index 0 ­1 ­2 ru an do ile M ia o p il a, s M p. Th sia ss Arg nd de a lo n a a a Tu a ey e az Fe tin bi in ut di ric ic Co tio s Re Ko pin Pe Ch rk rd ne ay la ex Ch m So In Br Af en ra ai Jo al h ili re In Ph n ia Ru country Sources: Lane and Milessi-Ferreti 2007; Chinn and Ito 2008. management. Proponents of capital account liberalization extremely volatile for reasons outside the control of individ- have argued that it creates opportunities for diversifying ual countries and that opening an economy to such capital portfolios, smoothing consumption, sharing risk, and results in financial crises. They argue, therefore, for the addressing contracts and payments in trade. Furthermore, imposition of frictions to limit the short-term cross-border the threat of capital outflows in the face of opportunistic trade in financial assets (see, for example, Bhagwati 1998, policies could have a disciplining effect on policy makers Rodrik 1998, and Stiglitz 2000). In a recent paper, the Inter- (see, for example, Fischer 1998 and Summers 2000). Con- national Monetary Fund--long-time champion of capital versely, critics point toward the East Asian crises at the end account opening--seems to concur with this view (Ostry of the 1990s and maintain that speculative capital flows are et al. 2010). Furthermore, current thinking emphasizes the 242 Managing Openness need for appropriate sequencing of capital account liberal- (defined as the ratio of foreign assets and liabilities to ization in the overall reform process and prioritizing liber- GDP) over the period 1993­2007 (see Aizenman, Chin, alization of certain selected capital flows and taxing others and Ito 2010).11 All indexes have been normalized between (Rogoff 2002; Feldstein 2003). In particular, some are con- 0 and 1. The origin in the charts indicates a completely cerned that the low interest rates found in the high-income floating exchange rate, zero monetary independence, a countries today could lead to a surge in capital flows to completely closed capital account, and zero reserve accu- emerging markets through the carry trade, which could mulation. We distinguish six periods: choke off the recovery in the receiving countries by affect- ing the competitiveness of domestic tradables production. · 1993­96, exchange-rate peg. During the first surge in Consequently, a number of countries have taken measures capital inflows during 1993 and 1994, the RBI inter- to deter inflows of short-term capital. vened heavily in the foreign exchange market to main- India has adopted a gradual, calibrated approach toward tain the rupee peg to the U.S. dollar. As a result, India's liberalizing its capital account, where certain flows and spe- reserve holdings more than doubled from US$12 billion cific agents have been accorded priority. On one side of the in July 1993 to US$24 billion in October 1994. This spectrum, FDI has been significantly liberalized during the intervention could not be sterilized because of a lack of past few years. Currently, barring a few sectors, FDI is uni- instruments and an illiquid bond market. The interven- versally allowed.10 Net FDI flows into India increased from tion therefore contributed to reserve money growth of just US$97 million in fiscal 1990­91 to nearly US$35 bil- 23 percent in 1994 compared to 16.5 percent in 1992 lion in fiscal 2008­09. India has been more cautious in lib- and is reflected in low values on the monetary inde- eralizing portfolio investment and has maintained separate pendence index. investment caps on subaccounts of foreign institutional · 1997­2000, low capital flows, moderate exchange-rate investors. Despite these caps, recent years have seen a sharp depreciation. In the second half of the 1990s, capital increase and reversals in net portfolio flows, which flows dropped, and the rupee depreciated by around increased from less than US$2.0 billion in fiscal 2001­02 to 8.5 percent against the U.S. dollar. The weaker peg is US$29.5 billion in fiscal 2007­08. reflected in the drop in the index of exchange-rate sta- Treatment of debt-creating flows has remained much bility in 1997 and 1998. In contrast, the drop in capital more restrictive: both borrowers and lenders have to satisfy flows and increased flexibility of the exchange rate eligibility criteria with a limit of US$500 million per bor- allowed a greater degree of monetary independence. rower per financial year and the Reserve Bank of India · 2001­05, increasing costs of sterilization, weakening imposes minimum maturity and interest rate ceilings. exchange-rate peg. Net capital inflows picked up after However, attractive global interest rates have led to a sharp 2002, and the RBI intervened heavily. It purchased more rise in the inflow of external commercial borrowing from than US$45 billion of foreign assets between January US$2.7 billion in fiscal 2001­02 to over US$30.0 billion in 2001 and December 2003. As a result, the RBI's stock of fiscal 2007­08. Recent years have also witnessed a significant foreign assets increased by more than Re 2.6 trillion liberalization of outward FDI and portfolio investment, but over this period, which is reflected in a jump in the upper limits still exist. reserve accumulation index and high values of Macroeconomic management faces increasing chal- exchange-rate stability. To limit the increase in the mon- lenges with rising integration in the global capital market. etary base resulting from this intervention, the RBI In fact, with free capital flows, policy makers cannot simul- reduced its holdings of government bonds by about half taneously aim to achieve a stable rate of capacity utiliza- the increase in foreign assets. The policy trade-offs were tion, a low and stable rate of inflation, and a stable becoming increasingly visible after 2003. Initially, the exchange rate, a problem known in the literature as the RBI continued to intervene heavily in the foreign "impossible trinity." Capital account opening thus reduces exchange market. However, by late 2003, it started to policy makers' ability to run a monetary policy independ- run out of government bonds for sterilization, and in ently from policies of their trading partners. The panels in January 2004, a new instrument for sterilization--the figure 18.6 illustrate India's macroeconomic management market stabilization scheme (MSS) bond--was intro- in a period of increasing integration with the global finan- duced. By August 2005, the amount of outstanding MSS cial markets. They compare monetary independence bonds had increased to Re 0.71 trillion. However, with a (defined on the basis of the correlation between Indian and rising number of outstanding MSS bonds, the quasi- U.S. interest rates), exchange-rate stability, reserve accu- fiscal costs of sterilization were also increasing. The RBI mulation as a share of GDP, and capital account openness therefore gave up sterilization and reduced interven- India: Managing Openness for a Rapidly Developing Domestic Market 243 Figure 18.6. Configurations of the Impossible Trinity and Reserve Accumulation in India, 1993­2007 a. 1993­96 b. 1997­2000 MI MI 1.0 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 cap open 0.0 ERS cap open 0 ERS RA RA 1993 1994 1997 1998 1995 1996 1999 2000 c. 2001­05 d. 2006­07 MI MI 1 1 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 cap open 0 ERS cap open 0 ERS RA RA 2001 2002 2006 2007 2003 2004 2005 Source: Authors' calculations. Note: MI = monetary independence; ERS = exchange-rate stability; RA = reserve accumulation; cap open = capital account openness. tions.12 Instead, the rupee appreciated from Re 46.3 to purchasing more than US$95 billion between January Re 43.5 per U.S. dollar between August 2004 and July 2006 and December 2007, resulting in the sharp increase 2005, reflected in a drop in the exchange-rate stabiliza- in the reserve accumulation index in these years. To ster- tion index in 2004 and 2005. ilize this intervention, the RBI issued additional MSS · 2006­07, surge in capital flows, appreciation of the bonds, and the outstanding stock of MSS bonds exchange rate. In 2006 and 2007, the RBI developed its increased by Re 1.5 trillion between April 2006 and "intermediate regime" in the face of a sharp increase in November 2007. Liquidity was also absorbed with an capital flows. The rupee was allowed to appreciate by increase in the cash reserve ratio by 200 basis points in 17 percent against the U.S. dollar between June 2006 2007. However, this measure was still not enough to and December 2007, as reflected in figure 18.6, panel d fully sterilize the market interventions, and the growth in a drop in the exchange-rate index. The RBI also in reserve money jumped to 18 percent in 2006 and resumed intervention in the foreign exchange market by 26 percent in 2007. In addition, a number of measures 244 Managing Openness were introduced in 2007 to curb capital inflows. These have resulted in higher interest rates, which the govern- included restrictions on external commercial borrow- ment wanted to avoid. At the same time, with inflation ing, curbs on the use of participatory notes, and restric- inching up and crossing 10 percent in March 2010, the RBI tions on foreign institutional investors. The RBI also may have been more wary of increasing the money supply introduced measures to limit loans to both foreign and and exacerbating inflationary pressure. In addition, it may domestically held mutual funds operating in India. have welcomed an appreciation of the rupee to help to However, the effectiveness of these measures was lim- keep inflation in check by reducing the cost of imports. ited, and capital continued to pour into the country. However, the RBI has left open the option for sterilization in · 2008­March 2009, portfolio capital outflow, loss of the future by replenishing the MSS bonds to the extent of Re reserves, and weakening of the rupee. As a result of the 500 billion (US$12 billion). This amount seems a paltry onset of the subprime crisis in the United States and the sum compared with the inflows of US$95 billion in the associated flight to safety, portfolio capital flowed out of year before the global crisis. India in early 2008, continuing until early 2009. In addi- In conclusion, India has resorted to multiple instruments tion, in the worst crisis quarters between September to deal with capital flows with the aim of restricting those 2008 and March 2009, banking assets were transferred that are highly volatile and those that create debt. It had a abroad, and suppliers' credit was withdrawn. The capital moderately flexible exchange-rate regime with interventions outflow exacerbated a rise in the current account deficit to prevent excessive volatility, sterilization of these interven- because of high commodity prices and dwindling tions through multiple instruments like MSS bonds and export receipts. The RBI reacted by selling foreign cur- cash reserve requirements, and building up of reserves. This rency assets worth US$40 billion between June 2008 and approach has suited India well, given that its economy has March 2009.13 In addition, the rupee was allowed to been able to maintain healthy growth, targeted monetary depreciate by more than 23 percent against the U.S. dol- and credit growth, moderate inflation, and a sustainable cur- lar in the same period. A number of monetary measures rent account deficit through most of the period. were also introduced to bolster liquidity, such as cuts in policy rates, opening up of refinance windows to allow Annex: Framework for Quantifying Policy easy access to credit for some of the troubled sectors, Choices under the Impossible Trinity and reduction of prudential norms relating to provi- sioning and risk weights. We measure the extent of monetary independence as the · The present situation. Capital flows resumed again in inverse of the annual correlation of the monthly interest early 2009, with India receiving net capital inflows rates between India and the United States. The United of US$53 billion during fiscal 2009. This resurgence States is taken as the base country following Aizenman, prompted the RBI to re-introduce restrictions on capital Chinn, and Ito (2010) and Shambaugh (2004), who argue inflows. The all-in-cost ceilings, which were withdrawn that Indian monetary policy has been most closely linked in January 2009 to encourage external commercial bor- to the United States. We use the money market rates for the rowing, were re-imposed in December 2009, with the interest rates. In India, the Reserve Bank of India uses sev- re-imposed ceilings being higher than the precrisis lev- eral policy tools like the repurchase rate, reverse repurchase els. Foreign currency convertible buybacks were discon- rate, and reserve ratio, among others, to conduct monetary tinued. The rupee was allowed to appreciate by nearly policy. Any change in these policy tools is likely to have an 17.5 percent between March 2009 and April 2010. Even impact on the money market rate. Following Aizenman, the 36-currency nominal effective exchange rate appreci- Chinn, and Ito (2010), the index for extent of monetary ated by more than 9 percent. independence is given by corr (i ,i * ) ­(­1) Interestingly, the RBI has lately refrained from interven- MI = 1 ­ , ing in the foreign exchange market. While it actually sold 1 ­ (­1) more than US$6 billion of foreign currency between where i and i* are the money market rates in India and March and November 2009, it did not intervene at all in the United States. This index is bound between 0 and 1. By foreign exchange market in the following five months. A construction, a higher value implies greater monetary number of reasons could be behind RBI's reluctance to independence. Data on money market rates are taken from intervene, including its preoccupation with managing the International Financial Statistics database. record borrowing requirements of the government in fiscal The index for exchange-rate stability (ERS) is calculated 2009­10 and fiscal 2010­11. Sterilization of inflows could using the annual standard deviations of the monthly log India: Managing Openness for a Rapidly Developing Domestic Market 245 change in the exchange rate between India and the United 3. Biggest in absolute value. Of course, many countries received higher remittances when measured against their GDPs. States, 4. The methodology was first proposed by Krause (1987). 0.01 5. The United Nations' SITC classification for exports is not easily ERS = 0.01 + ( ) mapped into the classification used in the ASI by the Central Statistical Organization, which is based on the UN's International Standard Indus- trial Classification (ISIC). where s is the standard deviation, is the first difference 6. Data are from the UN Comtrade database, factor intensity classifi- operator, and e is log of the exchange rate between the cation from Krause (1987). The three country groups account for more Indian rupee and the U.S. dollar. The ERS index also lies than 50 percent of India's exports, and the four product categories account for 92 percent of the total. See also Pitigala (2010). between 0 and 1 with a higher number indicating greater 7. The study by Beinhocker, Farrell, and Zainulbhai (2007) from the exchange-rate rigidity. Aizenman, Chinn, and Ito (2010) McKinsey Global Institute is based on household surveys by the National argue that a simple application of this formula exaggerates Council for Applied Economic Research. The study defines annual income brackets as follows: Globals = > Re 1,000,000, Strivers = the flexibility in cases where the exchange rate has followed Re 500,000­1,000,000, Seekers = Re 200,000­500,000, Aspirers = a narrow band but has been devalued or revalued infre- Re 90,000­200,000, Deprived = < Re 90,000. The middle class encom- quently. To overcome this problem, they apply a thresh- passes Globals, Strivers, and Seekers. They use Re 45.7/US$1 in real 2000 old--and the regime is labeled fixed if the rate of monthly U.S. dollars, and Re 8.5/US$1 adjusted for purchasing power. Thus, the middle class is defined as people earning US$23,500­US$117,650 per year change in the exchange rate has stayed within ±0.33 per- in 2000 international dollars. cent bands--and assign a value of 1 for the ERS index. The 8. For example, about 1.3 million cars were sold in India in fiscal data on exchange rate are also taken from the International 2009­10, a market hotly contested by three major players (Tata, Maruti- Suzuki, and Hyundai) and many others. By contrast, the top 15 global car Financial Statistics. companies sold more than that number of cars each during the year, and We deviate from Aizenman, Chinn, and Ito (2010) in the top four sold more than 6 million each. construction of the index of capital account openness, 9. This measure is based on the International Monetary Fund's Annual Report on Exchange Arrangements and Exchange Restrictions, in who use the de jure measure of openness developed in which a score of 0 indicates presence of some restrictions on a particular Chinn and Ito (2008), for two main reasons. First, as dis- transaction, while 1 indicates free movement of capital. Hence, even if cussed above, according to the Chinn-Ito measure, India's capital transactions have been progressively liberalized, they would attract extent of openness remained virtually stagnant since the a score of 0 so long as there are some minimal restrictions. In India; while controls on capital flows have been eased over time, the presence of some 1970s. Second, we think that it is the actual quantum of minimal restrictions explains the low score. flows that create a conflict between retaining monetary 10. Sectors where FDI is prohibited include atomic energy, certain independence and maintaining exchange-rate stability agriculture and plantation activities etc. Some sectors which require an industrial license (alcoholic beverages, defense equipments etc.) or are and not the regulations. A country with high de jure reserved for the small units are constrained in getting FDI. Some sensitive openness can have low inflow of foreign capital and not sectors such as banking, insurance etc. are subject to caps. be worried about simultaneously stabilizing the exchange 11. The methodology is described in detail in the annex. The period coverage is limited by the availability of the measure for capital account rate and exerting monetary autonomy. A country with openness. Presentation in calendar years. low de jure openness, however, can experience a large 12. In India, sterilization has an associated cost as the interest paid volume of flows because of low enforcement of capital out on the sterilization bonds are higher than the interest earned on for- controls and face problems in ensuring monetary inde- eign currency assets. 13. Foreign reserves dropped by US$62 billion because of valuation pendence and exchange-rate stability. As a result, we use effects. the de facto measure outlined in Lane and Milessi-Ferreti (2007), where capital account openness is described as the sum of foreign assets (FA) and liabilities (FL) as a Bibliography share of GDP: Aizenman, J., M. D. Chinn, and H. Ito. 2010. "The Emerging Global FA + FL Financial Architecture: Tracing and Evaluating the New Patterns of CapOpen = . the Trilemma Configurations." Journal of International Money and GDP Finance 29 (4): 615­41. Beinhocker, E., D. Farrell, and A. S. Zainulbhai. 2007. "Tracking the To make this measure comparable with monetary inde- Growth of India's Middle Class." McKinsey Quarterly 3: 51­61. pendence and exchange-rate stability, we also normalize it Bhagwati, J. 1998. "The Capital Myth: The Difference between Trade in between 0 and 1. Widgets and Dollars." Foreign Affairs 7 (3): 7­12. Bosworth, B., S. Collins, and A. Virmani. 2007. "Sources of Growth in the Indian Economy. Working Paper 12901, National Bureau of Economic Notes Research, Cambridge, MA. Chinn, M. D., and H. Ito. 2008. "A New Measure of Financial Openness." 1. India's fiscal year runs April 1 to March 31. Journal of Comparative Policy Analysis 10 (3): 307­20. 2. General elections were held in May 2010, which returned to power Feldstein, M. 2003. "Economic and Financial Crises in Emerging the incumbent United Progressive Alliance led by the Congress Party. Market Economies: Overview of Prevention and Management." In 246 Managing Openness Economic and Financial Crises in Emerging Market Economies, ed. M. Pitigala, N. 2010. "Background Paper for South Asia Regional Economic Feldstein, 1­20. Chicago: University of Chicago Press. Outlook." World Bank, Washington, DC. Fischer, S. 1998. "Capital Account Liberalization and the Role of the Rodrik, D. 1998. "Who Needs Capital Account Convertibility?" In Should IMF." In Should the IMF Pursue Capital-Account Convertibility? ed. the IMF Pursue Capital-Account Convertibility? ed. P. Kenen, Vol 207, P. Kenen, 1­10. Vol. 207, Essays in International Finance. Department 55­65. Essays in International Finance, Department of Economics, of Economics, Princeton University. Princeton University, Princeton, NJ. Krause, L. 1987. "The Structure of Trade in Manufactured Goods in the Rogoff, K. 2002. "Rethinking Capital Controls: When Should We Keep East and Southeast Asian Region." In Trade and Structural Change an Open Mind?" Finance and Development 39 (4): 55­56. in Pacific Asia ed. C. Bradford and W. H. Branson. Chicago: Univer- Shambaugh, J. C. 2004. "The Effects of Fixed Exchange Rates on Monetary sity of Chicago Press. Policy." Quarterly Journal of Economics 119 (1): 301­52. Lane, P., and G.-M. Milessi-Ferreti. 2007. "The External Wealth of Stiglitz, J. 2000. "Capital Market Liberalization, Economic Growth, and Nations Mark II: Revised and Extended Estimates of Foreign Assets Instability." World Development 28 (6): 1075­86. and Liabilities: 1970­2004." Journal of International Economics 73 Summers, L. H. 2000. "International Financial Crises: Causes, Prevention, (2): 223­50. and Cures." American Economic Review 90 (2): 1­16. Ostry, J., A. R. Ghosh, K. Habermeier, M. Chamon, M. S. Qureshi, and UN Comtrade (database), United Nations, http://comtrade.un.org. D. B. S. Reinhardt. 2010. "Capital Inflows: The Role of Controls." World Bank. 2010. Global Economic Prospects 2010. Washington, DC: Staff Position Note SPN/10/4, International Monetary Fund, World Bank. Washington, DC. 19 Exports and Export Diversification in Sub-Saharan Africa: A Strategy for Postcrisis Growth Vera Songwe and Deborah Winkler The past decade has been one of great volatility for Africa per year, as the continent became increasingly more open but also of substantial progress. Africa has become fully and globally connected. integrated into the world economy. At the turn of the The channels through which export expansion enhances decade, many in the developing world wondered if Africa aggregate productivity and growth are well known. Exports would become "the doomed continent" (Quenum 2000), allow for specialization in a country's comparative advan- crippled by political and ethnic tensions (Easterly and tage and thereby raise growth. Ricardo, in the famed theory Levine 1997), or if in fact Africa could claim the 21st cen- of comparative advantage, showed that countries benefit by tury (Gelb 2000). In this environment, predictions that specializing in the production of those goods with the low- Sub-Saharan Africa as a continent was about to enter the est opportunity cost and trading the surplus of production fastest growth period of its young 50-year history would over domestic demand, taking as given appropriate have seemed impossible. However, between 2002 and exchange-rate regimes. Under this model, a country should 2008 gross domestic product (GDP) grew by 6.5 percent quickly specialize in sectors in which it has a comparative annually in Sub-Saharan Africa. Commodity-exporting advantage. The new trade theory associated with Helpman countries as well as non-commodity-exporting countries and Krugman (1985) and generalized by Grossman and experienced high growth rates. In fact, some of the non- Helpman (1991), however, has shifted the focus of the trade commodity countries such as Burkina Faso, Mali, and debate from the static gains from trade to one in which Rwanda grew faster. increased investment, knowledge, and technology associ- The hitherto poor macroeconomic indicators that had ated with increased productivity growth can transform become synonymous with Africa also changed. Inflation in trade patterns and accelerate overall economic growth. most countries was brought down to single digits for the first Under the new theory, specialization is a result of scale and time, debt ratios fell to sustainable levels, and deficits were concomitant efficiencies. reduced as countries moved to consolidate the size of gov- However, even as gross output increased in both com- ernment, rationalize spending, and obtain debt write-offs. In modity- and non-commodity-exporting countries, the an overall favorable external economic environment, these debate over the quality of growth in Africa continued. reforms quickly began to produce results. Foreign exchange This debate shifted from the need to support export- reserves, including gold, increased more than 300 percent based growth to the quality of exports and its impact on from US$37 billion in 2001 to US$154 billion in 2008. growth--that is, what a country exports matters. The Net flows of foreign direct investment (FDI) more than dou- argument was as follows: a reliance on a less sophisticated bled from US$14 billion in 2001 to US$34 billion in 2008. export base is not sufficient to guarantee sustained Goods exports over the period 2000­08 grew by 18 percent long-term growth. Hausman, Hwang, and Rodrik (2006) 247 248 Managing Openness developed an indicator that measures the productivity capital stock. Second, we find that a 1 percent increase in associated with a country's export basket. This research exports has a higher impact on labor productivity than concluded that Africa needs to diversify its export base increases in capital intensity by the same percentage. These away from less sophisticated primary commodities into results are robust even if we exclude resource-intensive high-productivity sectors such as manufacturing to enjoy exporters from the country sample. Third, we find that faster growth. product and market diversification increase value added and labor productivity. In addition and contrary to expec- tations, we find that the positive value-added and labor- The Importance of Export Structure productivity effects from exports are larger the more This chapter argues that the export structure of Africa is one concentrated the countries in our sample are in export of the main reasons that Africa has been able to get a head products and markets. start out of the recession. Africa has not missed the boat, as These results indicate that for Sub-Saharan African many predicted, because of a reliance on commodity exports; countries to develop product value chains as suggested by on the contrary, it has benefited from its export structure, the new trade theory, they must continue to focus on which enabled it to rebound quickly after the crisis. Unlike exports of goods in which they have a revealed comparative the countries in East Asia, which followed a different export advantage to help them move toward more value-added trajectory of diversification before specialization, Africa exports based on these commodities. For example, Nigeria increased its export specialization first, and in some cases this and other oil-exporting countries could move into the strategy has served it well. In sum, while Africa's quick recov- pharmaceuticals industry, while Côte d'Ivoire and other ery from the crisis was mainly a result of sound precrisis cocoa-producing countries could move into the manufac- macroeconomic fundamentals, the structural composition of turing of chocolate-based products such as biscuits and its exports was the main reason for its quick rebound in eco- drinks. These results also indicate that while product and nomic activity. market diversification are important for growth, the effects We show that Africa's high export concentration in a few may be nonlinear; that is, product specialization even at sectors helped improve productivity and build resilience. In the commodity stage might be more beneficial for short- addition, the chapter shows that the direction of exports term growth than is product diversification over the matters just as much for resilience as it does for productivity. medium term. Exports to more sophisticated markets lead to higher The remainder of the chapter is organized as follows. In productivity, while exports to less sophisticated markets do the next section, we provide a brief overview of trends in not. Finally, we argue that African countries, especially the exports and export diversification in Sub-Saharan Africa. resource-based economies, need to concentrate on improv- In the following section, we estimate the effects of exports ing productivity in areas where they have a comparative and export diversification on value added and labor pro- advantage and on moving up the value chain in those com- ductivity using regression analysis for 30 Sub-Saharan modities. Oil-producing countries might, for example, move African countries over the period 1995­2008. We also into pharmaceuticals, and agriculture-based economies examine whether product and market diversification, as might expand into agroprocessing and marketing. measured by the Herfindahl-Hirschman Index, and the In this chapter, we examine the effects of exports and export share to China and the United States increase or export diversification on value added and labor produc- lower the positive effects from exports on labor productiv- tivity in 30 Sub-Saharan African countries. We find that ity and value added. We then analyze the impact of the eco- exports are critical for increasing value added and nomic crisis on exports from Sub-Saharan Africa. The final improving labor productivity. These results confirm section points to the policy implications for postcrisis many studies, which have shown that increased trade may export strategies. lead to growth and that a strong, outward-oriented trade regime facilitates the exchange and adoption of new tech- Trends in Exports and Export Diversification nology and improves productivity. An extensive literature review on the relationship between trade openness and We first examine trends in exports and export diversifica- growth since the 1970s can be found in Harrison and tion for Sub-Saharan Africa. Figure 19.1, panel a plots Sub- Rodríguez-Clare (2009). Saharan African exports by broad category for the period For 30 selected African countries, we find, first, that an 1976­2008. While exports have increased in all categories increase in exports of 1 percent has a higher and more sig- over this period, exports of fuels, manufactures, ores nificant impact on value added than a 1 percent increase in and metals, and, to some extent, commercial services have Exports and Export Diversification in Sub-Saharan Africa: A Strategy for Postcrisis Growth 249 Figure 19.1. Exports and Exports of Goods and Services in Sub-Saharan Africa a. Exports by broad category, 1976­2008 120 110 100 90 80 US$, billions 70 60 50 40 30 20 10 0 agricultural food fuel manufactures ores and commercial raw materials metals services category 1976 1996 2006 2007 2008 b. Exports of goods and services, 1970­2008 50 45 40 35 % of GDP 30 25 20 15 10 5 0 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 year East Asia and Pacific Middle East and North Africa Sub-Saharan Africa Europe and Central Asia Latin America and the Caribbean South Asia Source: Authors' calculations based on World Development Indicators (database), World Bank, http://data.worldbank.org. grown more strongly since 1996. Moreover, while African The share of goods and services exports in GDP in Sub- exports are still dominated by fuel, exports of manufactures Saharan Africa increased by more than eight percentage have expanded significantly over the past 20 years, growing points, from 27.7 percent in 1995 to 36.2 percent in 2008 at an average annualized rate of 14 percent between 1996 (see figure 19.1, panel b). This expansion was driven mainly and 2008 compared to an average 12 percent in all merchan- by goods exports, whose share in GDP rose from 23.4 per- dise exports. Only exports of ores and metals showed a higher cent to 34.4 percent over the same period. An international annual growth rate of 19 percent, while exports of fuel (12 comparison reveals that in 2008 Sub-Saharan Africa percent), commercial services (9 percent), food (8 percent), showed the third-largest share of exports in GDP after East and agricultural raw materials (4 percent) all grew at lower Asia and Pacific and the Middle East and North Africa. rates over this period. Moreover, while the export share over the 1980s through 250 Managing Openness 2000--the so-called lost years--remained more or less region's exports, as we will show in the section on the constant in Sub-Saharan Africa, a clear upward trend took effects of the economic crisis. hold at the beginning of the 1990s. No other continent has Next, we analyze the growth of Sub-Saharan African experienced such a long period of export stagnation. The exports by product category for 2000­08. The region's region that comes closest to Sub-Saharan Africa in this annual average export growth was highest in electricity respect is Latin America and the Caribbean, which saw (30 percent), mining of metal ores (28 percent), tobacco exports as a share of GDP plateau for a decade between (27 percent), and petroleum and natural gas (24 percent). 1984 and 1994 before growing again. Among manufactured products, export growth was highest Which are the most important destination countries for in motor vehicles, trailers, and semitrailers (20 percent) export growth in Sub-Saharan Africa? Figure 19.2 plots and in machinery and equipment (17 percent). Export average annual growth of Sub-Saharan African goods shares in 2008 reveal the strong dependency of Sub- exports over the period 2000­08 by destination. Total Saharan African exports on commodities, especially of goods exported from the region showed an impressive petroleum and natural gas (55 percent), or on manufactur- average growth of 18 percent per year over the period. ing sectors that make use of commodities, especially basic Exports to low- and middle-income countries experienced metals (11 percent). The structure of export shares also higher growth than exports to high-income countries. indicates that the composition of Sub-Saharan African Regional exports to high-income countries showed an exports is shifting into second-generation exports, as wit- annual average growth of 15.5 percent, while exports to nessed by the increase in exports of motor vehicles and low- and middle-income countries increased by 23 percent machinery. per year over the same period. Among high-income coun- tries, Sub-Saharan African exports to the United States The Effect of Exports and Export increased above average at 17 percent per year. Among low- Diversification on Growth and middle-income countries, annual export growth was highest for East Asia and Pacific and for South Asia, the In this section, we examine the impact of exports and result of strong demand from China and India. Sub-Saharan export diversification on growth. We first analyze the effect African exports to China alone grew at an annualized rate of exports on value added and labor productivity. Second, of 33 percent over the period. While developing countries we focus on the effect of export diversification of products played a bigger role in Sub-Saharan African export growth, and markets on value added and labor productivity for high-income countries still absorb the major share of the these countries. This chapter focuses only on the direct Figure 19.2. Annual Average Growth of Export Goods by Destination, Sub-Saharan Africa, 2000­08 40 32.8 31.1 30 29.3 27.7 22.6 23.5 21.8 % 20 17.3 15.5 14.8 15.7 14.0 10 0 IC 5 es IC fic ia id be d Af and ia a IC na ric -2 M rib an As As at H LM BR ci hi N eE n So ica EU Af Pa St a C a a st al h r C ic ut or a n tr d d e er ra en te an th Am ha ni dl C ia U Sa d As tin an b- th La st Su pe Ea ro Eu destination Source: Authors' calculations based on UN Comtrade (database), United Nations, http://comtrade.un.org. Note: BRIC = Brazil, the Russian Federation, India, and China; HIC = high-income countries; LMIC = low- and middle-income countries. Exports and Export Diversification in Sub-Saharan Africa: A Strategy for Postcrisis Growth 251 effects of exports and export diversification on higher Figure 19.3. Annual Commodity Export versus Value-Added value added and labor productivity. Indirect effects of Growth in 30 Sub-Saharan African Countries, 1995­2008 exports take a variety of forms and are basically induced by 0.4 changes in prices and income. Indirect effects include a reduction of the foreign exchange gap--which makes imports of capital goods cheaper and, thus, facilitates tech- 0.2 value-added growth nology spillovers--and also the effects of exports on employment and skills, among others. 0 Our empirical analysis covers 30 Sub-Saharan African countries for the period 1995­2008. The choice of coun- tries was based on data availability only. Our sample of 30 ­0.2 Sub-Saharan African countries includes Botswana, Burkina Faso, Cameroon, the Central African Republic, Chad, ­0.4 Comoros, Côte d'Ivoire, Eritrea, Ethiopia, Ghana, Guinea, ­2 ­1 0 1 Guinea-Bissau, Kenya, Lesotho, Madagascar, Malawi, Mali, commodity export growth Mauritania, Mauritius, Mozambique, Namibia, Senegal, Source: Authors' calculations based on UN Comtrade (database), United South Africa, Sudan, Swaziland, Tanzania, Togo, Uganda, Nations, http://comtrade.un.org. Zambia, and Zimbabwe. For full details of our data set, see Note: Annual export and value-added growth rates are based on values in the working paper version of this chapter. logarithms. Figure 19.4. Annual Manufacturing Export versus The Effect of Exports on Value Added and Value-Added Growth in 30 Sub-Saharan African Countries, Labor Productivity 1995­2008 We estimate an aggregate value-added production function 0.4 in log-linear form in which capital, labor, and technology are the input factors (see the working paper version of this chapter for the empirical model). Value added is the dif- 0.2 value-added growth ference between output and intermediate inputs. The technology shifter is a function of exports and tariffs. We 0 also control for fixed-country effects and fixed-year effects. We hypothesize that capital, labor, and exports ­0.2 have a positive impact on value added, while tariffs--as an inverse measure of trade liberalization--should have a negative influence. ­0.4 As a first indicator of the relationship between exports ­2 ­1 0 1 2 3 and growth, we plot annual export and value-added manufacturing export growth growth rates for our country sample for 1995­2008. Since Source: Authors' calculations based on UN Comtrade (database), United there might be a differential effect on value-added growth Nations, http://comtrade.un.org. depending on the type of exports, we split the product Note: Annual export and value-added growth rates are based on values in sample into commodities (figure 19.3) and manufactured logarithms. goods (figure 19.4). The bivariate regression lines indicate a stronger positive relationship between annual export achieve rapid economic development in emerging markets. growth of manufactured goods and annual value-added The East Asian successes, for instance, relied heavily on the growth than for commodity export growth. This finding manufacturing sector to achieve rapid economic growth. confirms the result that growth is more responsive to The East Asian Tigers (Hong Kong SAR, China; the Republic exports of manufactured goods than it is to exports of of Korea; Singapore; and Taiwan, China), Japan, and the commodities. Countries exporting goods with higher value second-generation successes (Indonesia, Malaysia, Thailand, added grow faster. While commodity exports interact posi- and, recently, China) all have shares of manufacturing tively with value added, the effects are less pronounced. value added that exceed the global average (UNIDO 2009). This result confirms the "law of development," accord- Full details of our regression results are contained in the ing to which industrialization is the well-proven way to working paper version of this chapter. Here, we focus on 252 Managing Openness an intuitive presentation of the main results. As hypothe- The first involves increasing the number of products sized, capital, labor, and exports have a significantly posi- exported to international markets. This effort requires dis- tive effect on value added. Remarkably, the elasticity of covering new products and moving up the value chain to exports is larger than that of capital, while the labor elas- produce products of higher value and sophistication. Much ticity is the highest. This finding reflects the labor-intensive of the focus in the past has been on the process of discover- production structure in these Sub-Saharan African coun- ing new exports. A second component of diversification tries. Tariffs have no influence on value added. Since we relates to breaking into new geographical markets, that is, are concerned that the results above could be driven by expanding market reach in products that have already natural resource­intensive countries, we drop four natural proven competitive. resource­intensive economies as identified by the Inter- One of the primary challenges facing Africa's resource- national Monetary Fund (IMF), namely, Cameroon, Chad, rich economies is how to diversify production beyond the Côte d'Ivoire, and Zambia, in the next regressions. The natural resource sector. Natural resource­based products results above continue to hold. have dominated exports for the past 50 years, but reliance In a second step, we formulate the value-added equation on such products has not made African countries richer. as a function of labor productivity. Labor productivity, Some economists refer to the "resource curse" as a reason defined as value added per worker, depends on capital that some African countries have not been able to use their intensity and the technology shifter, that is, exports and wealth to drive economic growth. However, others believe tariffs. Our study differs from studies that measure the that it is export concentration per se and not natural effect on per capita GDP in a new growth theory model, resources in particular that is negatively associated with which includes exports among other control variables such as growth (Lederman and Maloney 2007). In spite of the cir- the initial per capita GDP, human capital, population growth, cumstances, resource-rich countries like Norway, Indone- terms of trade, and investment ratio (see, for example, sia, and Malaysia have demonstrated that it is possible to Greenaway, Morgan, and Wright 1999). We hypothesize the use natural resource wealth to diversify and support eco- same coefficient signs as for the value-added regressions. nomic growth. Capital intensity and exports show a significantly posi- As a first indication, we show the relationship between tive effect on labor productivity, while tariffs have a signifi- export diversification of both products (figure 19.5) and cantly negative impact, which is line with our conjecture. markets (figure 19.6) and value added for our 30 Sub- Interestingly, the export elasticity is higher than that of Saharan African countries over 1995­2008. We use the capital intensity; that is, a 1 percent increase in exports Herfindahl-Hirschman Index (HHI) of market and prod- results in larger productivity gains than a 1 percent uct concentration as an inverse measure of export diversifi- increase in capital intensity. This finding reveals the poten- cation. The HHI of market concentration is defined as the tial of exports for increasing the region's competitiveness sum over a country's squared market shares of export and growth. The results become more significant when we drop the four natural resource­intensive countries. Figure 19.5. Export Concentration of Products versus Value Added in 30 Sub-Saharan African Countries, 1995­2008 The Effect of Export Diversification on Value Added and Labor Productivity 26 In addition to export growth, export diversification has been shown to be strongly associated with economic 24 growth (Lederman and Maloney 2007; Hesse 2009), partic- value added ularly for developing countries. Diversity in exports can 22 reduce income volatility for countries with large popula- tions living in poverty and reduce vulnerability to sharp declines in the terms of trade. Diversification also increases 20 the potential for generating spillovers, whereas reliance on only a few exports generally has greater negative conse- 18 quences for growth (Lederman and Maloney 2007). This 0 0.2 0.4 0.6 0.8 1.0 factor holds true especially for the countries in transition Herfindahl-Hirschman Index product from low-income status to middle-income status. Source: Authors' calculations based on UN Comtrade (database), United Export diversification can take place in two ways, by Nations, http://comtrade.un.org. exporting new products and by exporting to new markets. Note: Value added is in logarithms. Exports and Export Diversification in Sub-Saharan Africa: A Strategy for Postcrisis Growth 253 Figure 19.6. Export Concentration of Markets versus Value the more concentrated exports are in different products Added in 30 Sub-Saharan Countries, 1995­2008 and markets. Finally, we interact exports with a country's export 26 share to China, the EU-25 and the United States. In the full-country sample, a larger export share to China signifi- 24 cantly increases value added, but the interaction with exports is negative. That is, the gains from exports are low- value added ered the more dependent a country is on China. When we 22 drop natural resource­intensive countries, a higher export share to the United States now significantly increases value 20 added, but the positive value-added effects from exports are lowered for countries that depend more on the United States as an export destination. A higher market share to 18 China, however, no longer shows significant effects on value 0 0.2 0.4 0.6 0.8 1.0 added. The reason for these results might be that less natural Herfindahl-Hirschman Index market resource­intensive countries have a greater dependence on Source: Authors' calculations based on UN Comtrade (database), United Nations, http://comtrade.un.org. the United States, while more natural resource­intensive Note: Value added is in logarithms. countries have a greater dependence on China as a major export destination. A larger export share to the EU-25 has no influence on value added in both the full- and the reduced- destinations. If a country exports to only one destination, country sample. the HHI would be 1, while a lower HHI reflects a higher Next, we show the effects of export diversification and regional export diversification. Accordingly, the HHI of its interaction with exports on labor productivity. A higher product concentration is defined as the sum over a country's HHI of market concentration of exports markedly lowers squared market shares of different export products. Export labor productivity, which is significant only when natural concentration and value added are negatively correlated, or, resource­intensive countries are excluded from the sample. analogously, export diversification and value added have a The HHI of product concentration has no impact on labor positive relationship. The steeper regression line in figure productivity. When we interact the HHI with exports, a 19.6 suggests that the effect is stronger for export diversifica- higher product concentration of exports significantly lowers tion of markets. labor productivity but significantly increases the positive We then estimate the effects of export diversification productivity effects from exports. Analogously, a higher mar- and its interaction with exports on value added and labor ket concentration of exports has a strong negative effect on productivity. We hypothesize that a higher HHI--that is, labor productivity but significantly increases the positive less export diversification--has a negative effect on value productivity effects from exports. This finding holds for added. While these hypotheses are straightforward in both the full- and the reduced-country sample. Finally, value-added equations, we cannot unambiguously predict higher market shares to China and the United States and the effect of the interaction term with exports. The coeffi- their interaction with exports have no influence on labor cient sign of the interaction variable depends on whether productivity in the full-country sample. However, when we export growth is higher in countries with a higher export drop natural resource­intensive countries, a higher export concentration or a higher export diversification. share to the United States significantly increases labor pro- The regression results using the full-country sample show ductivity, but the positive productivity gains from exports are that a higher HHI of product concentration of exports sig- lowered for countries that depend more on the United States nificantly lowers value added, which is in line with our con- as an export destination. Again, a higher export share to the jecture, while a higher HHI of market concentration of EU-25 has no effects on labor productivity in the full- and the exports has no significant effect. The negative effect of the reduced-country samples. HHI of market concentration on value added becomes sig- In sum, the results show that exports significantly nificant when we drop the natural resource­intensive coun- increase value added and labor productivity. The export tries. We then interact exports with the two measures of elasticity is larger than the elasticity of capital in the value- export concentration. Interestingly, the interaction term of added regressions, while it is larger than the elasticity of the HHI with exports yields a positive coefficient, which is capital intensity in the labor-productivity regressions. This significant in the full- and reduced-country samples. That finding reveals the potential of exports for increasing the is, the positive impact of exports on value added is higher region's competitiveness and growth. Moreover, the results 254 Managing Openness imply that greater diversification of export products and and a growing world population. The rising commodity markets has a positive impact on value added and labor prices had a mixed impact on the continent: commodity- productivity. exporting countries such as Angola, the Democratic Repub- Surprisingly, the positive value-added and labor lic of Congo, and Nigeria benefited from the steep increase productivity effects from exports are larger the more con- in prices, while the non-commodity-exporting countries centrated the export markets are in our sample countries. In suffered from increases in food prices and a deterioration in effect, export growth seems to be greater in countries with their balance of payments. Third, the collapse of Lehman a higher concentration of exports. Finally, countries that Brothers in September 2008 marked the beginning of the are less natural resource­intensive depend more strongly financial crisis. Asset values contracted, as stock markets hit on the United States, while natural resource­intensive historic lows. According to some estimates (Bollard and countries depend more strongly on China as an export des- Gaitanos 2010), the amount of money lost in global stock tination. Future research could include indirect effects of markets totaled US$30 trillion dollars--a sum that could exports such as imports of capital goods and technology eradicate poverty in the developing world in 10 years. spillovers or the effects of exports on employment and The crisis reached Sub-Saharan Africa through three skills in Sub-Saharan Africa. main channels of transmission: labor markets, capital mar- kets, and export markets. The impact of the crisis on trade was devastating for Sub-Saharan Africa. Global trade The Economic Crisis in Sub-Saharan Africa growth dropped from 7.2 percent in 2007 to ­12.3 percent In 2008, the world economy was plunged into a deep and in 2009, while Africa's exports of goods and services in prolonged crisis--the worst recession since the Great 2009 dropped by about 16 percent compared to the 22.8 Depression. Global GDP contracted for the first time on percent increase in 2008. Sub-Saharan Africa was the second- record. Africa's hard-won gains came under threat from the hardest-hit region on the trade front after the Middle East crisis. Between 2007 and 2008, Sub-Saharan Africa was hit and North Africa (see figure 19.7). The trade collapse in by triple shocks--the food, fuel, and financial crises. Sub-Saharan Africa is also reflected in exports alone, which Growth plunged from a 6.5 percent average between 2000 fell by more than 40 percent in January 2009 on a year-to- and 2007 to 1.6 percent in 2009 (see table 19.1). year basis (figure 19.8). It is no surprise that the collapse in The first hint of an impending crisis was the rapid rise commodity trade following the real sector collapse in the in food prices caused by the high price volatility of oil mar- West affected Sub-Saharan Africa so severely, given that kets, which was transferred to corn markets as oil prices one-third of the world's resource-dependent countries are rose above US$50 per barrel and use of corn-based ethanol in the region. increased. Second, many financial institutions, seeking During the second half of 2008, nonenergy commod- safety and an exit from the more risky and less transparent ity prices plunged 38 percent. In December, nonenergy derivatives market, diversified into commodity markets, prices fell 6.8 percent, down for the fifth consecutive putting further pressure on commodity prices (Songwe month. Oil prices fell 69 percent between July and forthcoming). The standard literature on the food price December 2008, reversing the oil price increases of the crisis has attributed the rise in prices to increasing use of previous three and a half years. Oil exporters suffered a food crops for biofuels, and rapidly rising oil prices, as loss, but many of them such as Nigeria had built up a sav- mentioned here, but also to increased food and meat con- ings cushion during the boom years and were better able sumption in emerging markets because of rising incomes to withstand the crisis. By end 2009, however, oil prices Table 19.1. Actual and Projected GDP Growth Rates, Selected Regions, 2007­11 percent Actual Projected Growth 2007 2008 2009 2010 2011 World 5.2 3 ­0.8 3.9 4.3 Advanced economies 2.7 0.5 ­3.2 2.1 2.4 Sub-Saharan Africa 6.9 5.6 1.6 4.3 5.5 Central and Eastern Europe 5.4 3.1 ­4.3 2 3.7 Middle East 6.4 5.3 2.2 4.5 4.8 Source: IMF 2009, 2010. Exports and Export Diversification in Sub-Saharan Africa: A Strategy for Postcrisis Growth 255 Figure 19.7. Trade Volume Decline from Peak to Trough, by Region, 2008­09 0 ­10 percent ­20 ­30 ­40 ­50 Af and a l A nd ia ie e ci d be nd ric om m Pa an As ra a rib a a a si a s fic an on co Af ric th st nt pe ia h ec -in Ca ic ut or a As n Ce uro e er N le E ra So gh st th Am ha E Ea d Hi Sa id n tin er M b- st La Su Ea region Source: Didier, Hevia, and Schmukler (2010). Note: Percentage decline in trade from peak to trough is shown for each region. Figure 19.8. Goods Exports for Six World Regions, Constant and Seasonally Adjusted, 2007­10 120 110 index (January 2008 = 100) 100 90 80 70 60 50 ar 007 ay 7 l. 7 p. 07 . 7 n. 7 ar 08 ay 08 l. 8 p. 08 . 8 n. 8 ar 09 ay 9 l. 9 p. 09 . 9 n. 9 ar 10 ay 10 Ju 010 10 M 00 Ju 00 ov 0 Ja 200 Ju 00 ov 0 Ja 200 M 200 Ju 00 ov 0 Ja 200 Se 20 N 20 M 20 M 20 Se 20 N 20 M 20 Se 20 N 20 M 20 M 20 20 2 .2 2 2 2 2 n. . . . l. Ja M month East Asia and Pacific South Asia Latin America and the Caribbean Middle East and North Africa Europe and Central Asia Sub-Saharan Africa Source: Authors' calculations, based on Global Economic Monitor (database), World Bank, http://www.worldbank.org/gem. had recovered from their low point of US$40 in December share of exports from Sub-Saharan Africa to high-income 2008 to about US$70. Some poorer countries suffered partic- countries declined sharply from almost 70 percent of the ularly large shocks. Many poor African countries experienced total in 2000 to 60 percent in 2008 and fell further to terms-of-trade losses of over 3 percent of 2008 GDP, with 55 percent in 2009. This sharp drop was caused by a big losses exceeding 5 percent in Chad, Guinea, Mozambique, decline in exports to the EU-25 countries, which fell from and Zambia. 35 percent in 2000 to 27 percent in 2008 and 22 percent in Has the crisis changed the pattern of export destinations 2009. Exports to the United States, which absorbs almost a for Sub-Saharan Africa? Figure 19.9 shows the share of quarter of Sub-Saharan African exports, showed a weak goods exports by destination as a percentage of total goods decline during and after the crisis. In contrast, Sub-Saharan exports from Sub-Saharan Africa pre- and postcrisis. The African exports to low- and middle-income countries as a 256 Managing Openness Figure 19.9. Shares of Sub-Saharan African Goods Exports, by Destination, 2000, 2008, and 2009 70 % of total goods exports 60 50 40 30 20 10 0 C 5 es IC c sia be nd Af and ia a IC a fi ric in -2 HI As at LM BR ci lA rib a a an a Ch EU Af ric Pa St th st h ra C a ic ut or a n d d nt e er N eE ra ite So an Ce th A m ha dl Un ia Sa id d As tin an M b- La st Su e Ea rop Eu 2000 2008 2009 Source: Authors' calculations based on UN Comtrade (database), United Nations, http://comtrade.un.org. Note: BRIC = Brazil, the Russian Federation, India, and China; HIC = high-income countries; LMIC = low- and middle-income countries. percentage of total exports increased strongly from 27 per- from Sub-Saharan Africa to high-income countries as a per- cent in 2000 to 38 percent in 2008 and 43 percent in 2009. centage of total intermediate goods exports dropped sharply Among low- and middle-income countries, exports from from 71 percent to 60 percent between 2000 and 2008, a fall Sub-Saharan Africa expanded in all regions except for the driven by declining demand in the EU-25. This drop was off- Middle East and North Africa and within Sub-Saharan set by increasing demand for intermediates from the BRICs. Africa itself. The expansion of export share from Sub- Intermediate goods exports from Sub-Saharan Africa to Saharan Africa was strongest in East Asia and Pacific and China alone rose from 6 percent to 15 percent over the period grew from 7 percent in 2000 to 15 percent in 2008 and 19 2000­08. The crisis further accentuated the trends for capital percent in 2009. China was the main driver, whose export and intermediate goods in 2009. shares increased from 5 percent in 2000 to 14 percent in Intraregional trade continued to drop during the crisis. 2008 and 17 percent in 2009. Africa trades more with itself than it does with Latin Amer- The share of exports to the EU-25 in all export categories ica and the Caribbean, Europe and Central Asia, and the dropped over the period 2000­08; however, this drop was Middle East and North Africa combined. And before the compensated for by growth in intraregional trade and crisis, trade within Africa was higher than trade with South emerging markets' exports for some categories. The share of Asia, including India. However, the share of intraregional capital goods exports from Sub-Saharan Africa to high- trade has been dropping, falling from 11.2 percent before income countries as a percentage of total capital goods the crisis in 2000 to 8.8 percent postcrisis. For Africa to exports fell from 49 percent to 44 percent between 2000 and realize its economic potential and achieve fast and sus- 2008, with exports of capital goods falling especially sharply tained growth, it will have to increase intraregional trade. to the United States, while intraregional trade increased strongly from 37 percent to 42 percent over the same period. Policy Implications for Postcrisis Sub-Saharan Africa's export share of consumption goods to Export Strategies high-income countries as a percentage of total consumption goods exports declined from 71 percent to 68 percent over Over the past decade, many developing countries embraced the period 2000­08, a drop driven mainly by a drastic decline export-led strategies as an engine for growth and have in EU-25 demand, which fell from 50 percent to 42 percent. increasingly diversified both export markets and products. Growing demand by the United States and the BRICs (Brazil, The food, fuel, and financial crises were the first test of the the Russian Federation, India, and China), however, counter- resilience of this strategy to shocks. Our analysis shows balanced this effect. The share of intermediate goods exports that this strategy has served Africa well, helping facilitate Exports and Export Diversification in Sub-Saharan Africa: A Strategy for Postcrisis Growth 257 its early recovery from the global crisis. Africa's GDP is Despite the strong rebound, the crisis has interrupted projected to come out of the crisis faster than that of most the growth acceleration experienced on the continent; other regions (see table 19.1). Africa's growth is pro- growth averages have dropped. African countries will need jected to rebound to 5.5 percent in 2011, up from to do more, better, and faster to catch up to pre-2008 1.6 percent in 2009. While this increase is still signifi- growth levels. The first order of business should be to pro- cantly below precrisis levels compared to other regions, tect past gains in macroeconomic stability and continue Africa has bounced back faster than both Europe and structural reforms. Second, countries should put in place Central Asia and the Middle East and North Africa. policies that exploit the increasing benefits of South-South Africa has performed well but has also exhibited some trade, including improved intraregional trade. Growth in structural weaknesses that must be addressed if it is to exports to low- and middle-income countries offers Africa accelerate growth and reduce poverty. a chance to diversify export markets further while it works The recent global economic downturn has highlighted to diversify its products. the critical importance of trade as an engine for Africa's Most important, Sub-Saharan African countries must growth and, more important, as a way of smoothing out pursue a dual strategy of diversifying both export markets the crisis impact. In view of our regression results, we con- and product markets to accelerate growth. African countries clude that the export structure of Sub-Saharan Africa is need to work to protect market share in existing commodity one of the main reasons why it has been able to get a head markets by retaining and increasing their competitiveness in start out of the recession. Africa has not missed the boat, as the areas where they have a comparative advantage. The cri- many predicted, because of its reliance on commodity sis has shown that this strategy allows countries to manage exports; on the contrary, it has benefited from its export economic downturns better. The above analysis also shows structure, particularly because of high demand from that exports of manufactures contribute more to growth China. In sum, while Africa's rally was mainly the result of than exports of commodities. sound macroeconomic fundamentals in place before the Therefore, while acquiring new markets for commodi- crisis, the structural composition of Africa's exports also ties, countries also need to expand and diversify exports to supported the region's economic resurgence. high-value markets such as the EU and the United States. While precrisis trade discussion on Africa focused on Our analysis shows that while export market share to the the need for increased product diversification, the crisis developed countries is dropping, a higher export share to has demonstrated that increased market diversification is the United States has positive growth effects in non- equally important. For resource-dependent exporting resource-intensive countries. Part of the strategy is to put countries, especially those whose exports are highly con- in place policies that attract more foreign direct investment centrated in a few commodities, we conclude that market to facilitate more technology transfer. Countries therefore diversification is equally important for growth as prod- need to improve their business regulation, property rights uct diversification. This important finding suggests that legislation, and, most important, the governance environ- export concentration in a few products in which coun- ment for business. tries have a high comparative advantage yields more ben- efits than product diversification in goods in which they Bibliography have less comparative advantage. In cases where countries have a comparative advantage in a few commodities, spe- Bollard, A., and S. Gaitanos. 2010. Crisis: One Central Bank Governor and the Global Financial Collapse. Auckland: Auckland University Press. cialization should precede diversification. Product diver- Didier, T., C. Hevia, and S. Schmukler. 2010. "How Resilient Have Devel- sification must therefore be managed to safeguard market oping Countries Been during the Global Crisis" Background Paper, share of exports. World Bank, Washington, DC. Easterly, W., and R. Levine. 1997. "Africa's Growth Tragedy: Policies and The analysis also suggests that the export landscape in Ethnic Division." Quarterly Journal of Economics 112 (4): 1203­50. the region is changing. Primary-commodity export coun- Gelb, Alan H. 2000. Can Africa Claim the 21st Century? Washington, DC: tries are more dependent on emerging markets for exports, World Bank. Greenaway, D., W. Morgan, and P. Wright. 1999. "Exports, Export Compo- while as countries move up the value chain, they turn sition and Growth." Journal of International Trade and Economic increasingly to the high-income countries as export desti- Development 8 (1): 41­51. nations. If managed well, this pattern of diversification of Grossman, G., and E. Helpman. 1991. Innovation and Growth in the Global export markets could help underpin the transition from Economy. Cambridge, MA: MIT Press. Harrison, A., and A. Rodríguez-Clare. 2009. "Trade, Foreign Investment, market to product diversification on the continent, as more and Industrial Policy for Developing Countries." Working Paper countries enter second-generation export phases. 15261, National Bureau of Economic Research, Cambridge, MA. 258 Managing Openness Hausman, R., J. Hwang, and D. Rodrik. 2006. "What You Export Matters." Lederman, D., and W. F. Maloney. 2007. "Trade Structure and Growth." In Working Paper, Center for International Development, Harvard Natural Resources: Neither Curse nor Destiny, ed. D. Lederman and University, Cambridge, MA. W. F. Maloney, 15­40. Palo Alto, CA: Stanford University Press. Helpman, E., and P. Krugman. 1985. Market Structure and Foreign Trade: Quenum, B. M. 2000. "Is Africa Doomed? The African Challenge." Africa Increasing Returns, Imperfect Competition, and the International Econ- Business Magazine. December. omy. Cambridge, MA: MIT Press. Songwe, V. Forthcoming. "A Tale of Two Crises: Innovation and Diversifi- Hesse, H. 2009. "Export Diversification and Economic Growth." In Break- cation in Financial Markets and Impact on Africa." In Food and Finan- ing into New Markets: Emerging Lessons for Export Diversification, ed. cial Crises Impacts on Sub-Saharan Africa, ed. D. R. Lee and M. Ndulo. R. Newfarmer, W. Shaw, and P. Walkenhorst, 55­80. Washington, DC: Wallingford, U.K.: CABI. World Bank. UNIDO (United Nations Industrial Development Organization). 2009. IMF (International Monetary Fund). 2009. World Economic Outlook: Cri- Industrial Development Report 2009. Breaking In and Moving Up: New sis and Recovery. Washington, DC: IMF. Industrial Challenges for the Bottom Billion and the Middle-Income ------. 2010. World Economic Outlook: Rebalancing Growth. Washington, Countries. Vienna: UNIDO. DC: IMF. Part IV II Emerging Trade Policy ISSUES in the Postcrisis Environment 20 Structural Changes in Commodity Markets: New Opportunities and Policy Challenges for Commodity Exporters Donald Mitchell and Enrique Aldaz-Carroll The global financial crisis led to a sharp decline in merchan- opportunities for export growth. Commodity price booms dise trade. World trade volumes declined by an estimated have often led to increased growth in commodity-dependent 17.6 percent in 2009 from 2008 (World Bank 2010c). The countries, but over the longer term these economies have decline in trade from peak to trough was even greater at had slower economic growth than those with more diversi- 24­25 percent for both developed and developing coun- fied export bundles, mainly because of the high price tries. This sharp decline in merchandise trade exposed the volatility that results from commodity booms and busts. vulnerability of countries heavily dependent on exports Prices fluctuate because of the relatively low short-run and has led to a reevaluation of the export-led growth supply response and low price elasticity of demand. More- model for developing countries. over, demand growth has been slower because of low Commodities trade, however, declined less than other income elasticities for most commodities and slowing pop- merchandise trade and has rebounded more sharply since ulation growth, particularly in developed countries. the global recession ended, largely because of strong However, the rapid economic growth of developing growth in developing-country demand for imports (and countries over the past two decades has begun to change demand for commodities, in particular). This relatively these historical relationships. Growth in demand for com- better performance of commodities suggests new opportu- modities has increasingly shifted from developed to devel- nities as well as policy challenges for commodity exporters. oping countries, where income elasticities and population Commodities account for roughly one quarter of the dollar and income growth rates are higher. Indeed, the share of value of global merchandise exports (table 20.1).1 Although global commodity imports going to low- and middle- commodities have historically had slower growth than income countries increased from 17 percent in 1990 to manufactures and declining relative prices, their share has 27 percent in 2007 (World Bank 2009b). In China, for been increasing in recent years. example, gross domestic product (GDP) grew more than Commodity exports accounted for only one quarter of 10 percent annually from 1990 to 2007, and the country global merchandise exports in 2007, but they accounted now accounts for nearly 40 percent of global consump- for more than half of export earnings for 57 countries tion of most metals (compared to 10­15 percent a decade (table 20.2). Twenty-two countries received at least 90 per- ago). India, which had almost 7 percent annual GDP cent of their merchandise export earnings from commodi- growth over this same period, has increased oil imports at ties, and 37 countries received at least 80 percent of their 5 percent per year--twice the global average--albeit from merchandise export earnings from commodities. a low base. Many other countries--including Chile, Historically, countries that have relied on commodity Malaysia, Mozambique, Uganda, and Vietnam--have had exports have faced declining terms of trade and stagnant more than 5 percent annual GDP growth over the same 261 262 Managing Openness Table 20.1. Percentage Growth of Merchandise, Manufacturing, and Commodities Export Values and Shares of Merchandise Trade, 2005­08 2005 2006 2007 2008 Merchandise 12.8 16.8 13.7 12.2 Manufacturing 10.6 13.5 14.4 8.8 Commodities 22.9 25.2 11.6 20.2 Share of merchandise trade Manufacturing 71.3 69.3 69.3 65.6 Commodities 23.7 25.4 24.9 26.7 Source: World Bank 2009b. Table 20.2. Share of Commodities in Total Merchandise Exports for 57 Commodity-Dependent Countries, 2007 Low-income Lower-middle- Upper-middle High-income Share of commodities countries income countries income countries countries Total countries 90­100 7 8 4 3 22 80­89 7 3 3 2 15 70­79 2 1 2 2 7 60­69 4 4 1 0 9 50­59 1 2 1 0 4 50­100 21 18 11 7 57 Source: World Bank 2009b. period and are becoming larger consumers and importers A combination of factors drove the rise in commodity of commodities. prices. These factors include years of low investment in While the prospects for commodity-dependent coun- response to the low prices of the 1980s and 1990s, a surge tries have brightened, vulnerability to price volatility and in demand for commodities in developing countries, the other risks still persists. Given that the majority of these declining value of the dollar relative to other major cur- countries are classified as low- and middle-income countries, rencies, supply disruptions in major producing countries, the urgency of crafting policies to turn the current boom in policy changes, and temporary bans on exports of some commodity prices into sustainable economic growth is clear. commodities. Some of these factors were of short duration, This chapter discusses the structural changes in commodities and their impacts have already dissipated. However, others markets over the past decade and the policy implications of reflected structural changes that will have longer-lasting these trends. impacts and important consequences for the global econ- The rest of the chapter is structured as follows. The next omy and developing countries. section discusses recent structural changes in commodity One of the main reasons that commodity prices rose so markets. The following section considers the prospects for sharply during 2000­08 was underinvestment. During the commodity markets in light of these changes. The subse- 1980s and 1990s, investment in commodity production quent section presents a historical account of the literature capacity fell (especially for crude oil, metals, and minerals) on commodity dependence. The next section discusses in response to a severe depression in commodity prices.2 policy recommendations for resource-based economies. In particular, real spending by major U.S. multinational The final section concludes. oil companies on exploration and development declined by half from 1981 to 1985 and stagnated at that depressed level for 15 years before rising in response to higher oil Structural Changes in Commodity Markets prices in the past decade (World Bank 2009a). Commodity prices have increased sharply in real terms However, slow demand growth in the 1980s and 1990s since their cyclical lows at the turn of the 21st century. In largely offset underinvestment and allowed existing capac- fact, the increases from 2000­01 to 2008 were larger than ity to suffice. Global oil consumption increased by only any commodity boom of the 20th century: prices for crude 0.7 percent per year during the 1980s and 1.3 percent per oil, metals, and minerals nearly tripled, while agricultural year during the 1990s (BP 2009). Investment in agriculture prices nearly doubled (figure 20.1). also declined, but demand growth fell so rapidly that there Structural Changes in Commodity Markets: New Opportunities and Policy Challenges for Commodity Exporters 263 Figure 20.1. Real Commodity Prices, 1960­2009 300 250 200 index 150 100 50 0 60 62 64 66 19 8 70 72 19 4 76 78 80 82 84 19 6 88 90 92 94 19 6 98 00 02 04 06 08 6 7 8 9 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 year agriculture crude oil metals and minerals Source: World Bank 2010a. Note: Prices are measured relative to the manufacturing unit value of the world exports index in constant 2000 U.S. dollars. was little upward pressure on real prices of agricultural Nearly all the growth in world oil consumption in the products. As a result of slower population growth and past decade has come from countries outside the Organisa- falling income elasticities as incomes rose, growth of tion for Economic Co-operation and Development (OECD), global grain consumption fell from about 2.8 percent per and about one-third of that growth has come from China year during the 1970s to 1.7 percent during the 1980s and (figure 20.3). China's share of world oil consumption 1.0 percent during the 1990s (Production, Supply, and Distri- increased from 3.5 percent in 1990 to 9.5 percent in 2008, bution database, U.S. Department of Agriculture, http:// while the share of other Asian developing countries increased www.fas.usda.gov/psdonline). By the early 2000s, surplus from 17 percent to 21 percent over the same period. China capacity had waned, and demand had begun to increase, thus and India are expected to account for just over half the setting the stage for the rise in commodity prices. increase in global demand for primary energy between 2006 A second reason for the rise in commodity prices was the and 2030, based on projections of continuing strong eco- rapid economic growth of developing countries. Eighteen nomic growth (IEA 2008). low- and middle-income countries had real average GDP growth of at least 5 percent per year during 1990­2007.3 Trends in the Agriculture Sector These 18 countries have a combined population of nearly 3.0 billion people--nearly half the global population--and Agricultural prices have also risen in the past decade but account for a large share of the increase in commodity much less than the prices of oil and metal products. One of demand. the main reasons for this slower rate of growth in prices was the rapid increase in per capita food production dur- ing the 1970s, 1980s, and 1990s (figure 20.4).4 For instance, Metals, Minerals, Crude Oil, and the Role of China per capita calorie supplies in China rose by more than Demand growth in China was the dominant force behind 50 percent from 1970 to 2000 and reached levels that the dramatic increase in the prices of metals, minerals, and rivaled consumption in high-income Asian countries such crude oil from 2001 to 2008, during which period they as Japan and the Republic of Korea (FAO 2010).5 India, nearly tripled. China's consumption of metals has risen with its largely vegetarian diets, had lower per capita calo- significantly in the past decade (figure 20.2), and that rie supplies than China but still witnessed a 12 percent country consumes between 30 and 45 percent of most increase in per capita calorie supplies from 1970 to 2000 metals produced. and a 29 percent increase in per capita food production. 264 Managing Openness Figure 20.2. China's Shares of Global Metals Consumption, 1996­2008 50 40 share (%) 30 20 10 0 96 98 00 02 04 06 08 19 19 20 20 20 20 20 year aluminum copper nickel steel zinc Source: World Bureau of Metals Statistics 2010. Figure 20.3. Share of Growth of World Oil Consumption of OECD Countries, Non-OECD Countries, and China, 1988­2008 140 120 100 80 share (%) 60 40 20 0 ­20 ­40 1988­93 1993­98 1998­2003 2003­08 OECD countries non-OECD countries China Source: BP 2009. India also became an exporter of wheat and rice, its staple Although its role was much less significant than in fuels food grains, in the 1990s. The increase in per capita food or metals, China has still contributed to the rising demand production in South America was even greater than in for agricultural imports, particularly nonfood imports. For Asia. South America emerged as a main exporter of grains instance, China has increased imports of natural rubber and oilseeds during this period. In contrast, per capita food and cotton to support its tire and textile industries. China production in Sub-Saharan Africa remained relatively has also become the largest soybean importer in order to unchanged from 1970 to 2000, as modern inputs were not support its rapidly growing pork and poultry industries widely used and population growth rates remained high. (World Bank 2010b). Sub-Saharan Africa has become a large importer of grains While increased production has placed downward and other food crops. pressure on agricultural prices, the expansion of biofuel Structural Changes in Commodity Markets: New Opportunities and Policy Challenges for Commodity Exporters 265 Figure 20.4. Per Capita Food Production of World and Selected Regions, 1961­2007 200 175 150 index (1970 = 100) 125 100 75 50 25 0 60 70 80 90 00 19 19 19 19 20 year world Africa Asia South America Source: FAO 2010. production in recent years has encouraged a rise in agricul- price indexes are projected to decline from their 2008 tural prices. The price of food crops has steadily increased peaks but remain well above 1990­2000 levels (World in recent years, culminating in a price spike in 2008 because Bank 2010c). the demand for food crops used to produce biofuels grew faster than the increase in supplies. The Future of Oil Brazil, the European Union (EU), and the United States account for 90 percent of current global biofuel According to the World Bank (2010b), crude oil prices are production. The United States uses almost one-third of expected to remain high (at more than double 2000 prices) its maize production (13.2 percent of global production) until 2020. This projection is based on rapid demand for ethanol. Brazil currently uses approximately half its growth in developing countries, declining production from sugar cane crop (18 percent of global production) for mature fields, and higher costs for new production in ethanol. The European Union produces both ethanol remote areas and unstable regions. from grains (wheat and maize) and biodiesel from Crude oil is expected to be more costly in the future vegetable oils (rapeseed and soybean and sunflower oils). because many of the most easily accessible and richest In 2009, the EU-27 used an estimated 7.4 million tons fields and deposits have already been developed. Countries of vegetable oils for biodiesel (5.4 percent of global that restrict access and plan to develop their resources production). slowly are where a significant proportion of remaining resources are located, a factor that will limit supplies and create upward pressure on prices. Still other resources are Prospects for Commodity Prices located in politically unstable regions such as the Niger The World Bank and many other organizations-- the Inter- Delta and are costly and dangerous to develop. national Energy Agency, International Monetary Fund, Restricted access to potentially rich oil fields has encour- Food and Agriculture Organization, OECD, U.S. Depart- aged the development of offshore production. In fact, nearly ment of Agriculture, U.S. Department of Energy, Food all the increase in global oil production since 1970 has come and Agriculture Policy Research Institute, and Interna- from offshore fields (World Bank 2009a). However, as the oil tional Food Policy Research Institute--forecast high spill in the Gulf of Mexico in April 2010 illustrates, such commodity prices relative to historical levels for at least fields can pose severe environmental risks. Thus, future the next decade. As shown in figure 20.5, real commodity development of offshore fields will likely require more 266 Managing Openness Figure 20.5. Real Commodity Price Indexes, Historical and Projected, 1960­2020 300 forecast 250 indexes (2000 = 100) 200 150 100 50 0 60 70 80 90 00 10 20 19 19 19 19 20 20 20 year agriculture crude oil metals and minerals Source: World Bank 2010c. Note: Prices are measured relative to the manufacturing unit value of the world exports index in constant 2000 U.S. dollars. costly safety measures and location restrictions, which will greenhouse gases could push energy prices lower and car- limit supplies and increase costs. bon prices higher, to over US$100 per ton by 2030, resulting Most experts seem to agree that while natural resources in greater demand for low-carbon biofuels (IEA 2008). remain abundant, many are of lower quality than resources The U.S. Department of Energy also predicts a sharp rise already developed. However, as prices rise, oil fields that in energy prices as a result of growing energy demand-- were previously too small or too costly to develop will particularly in China, India, and other developing coun- become profitable. For instance, the Bakken Formation oil tries--as well as efforts by many countries to limit access to field in the northern United States and southern Canada their oil resources (2009). In the Department of Energy has pockets of oil dispersed throughout the formation that reference case, world oil prices are predicted to rise to can now be tapped with newer, more expensive technology US$130 per barrel in real dollars in 2030. However, they such as horizontal drilling and fracturing techniques note significant uncertainty in their projections and fore- (USGS 2008). By some estimates, this formation may have cast a range of oil prices from US$50 to US$200 per barrel more recoverable oil than all other known U.S. reserves in alternative scenarios. In the low-price scenario, many of combined, but extraction will be more expensive than has the major oil-producing countries expand output more been the case with discrete pools of oil in the past. rapidly than in the reference case and increase their share Indeed, the International Energy Agency recently declared of world production beyond current levels. In contrast, in that "the era of cheap oil is over" (IEA 2008). The agency the high-price scenario, major oil producers maintain tight projects that the price of crude oil will average US$100 per control over access to their resources and develop them barrel over the 2008­20 period and will then rise to over slowly. Other organizations generally share the view that US$120 per barrel in 2030 in constant 2007 dollars. This dra- energy prices will be higher in the future because of strong matic increase is based on the expectation that the output demand growth and dwindling supplies of easily accessible from mature fields will suffer a severe decline, developing crude oil. countries will experience rapid demand growth, and investment in new capacity will fail to keep up with growth The Future of Agriculture in demand (despite adequate global oil reserves). Assuming continued strong economic growth, China and India are Rising energy prices are expected, in turn, to put upward expected to account for just over half the increase in global pressure on agricultural prices. Energy is an important demand for primary energy between 2006 and 2030. input to agricultural production as fuel and in the produc- However, the International Energy Agency also suggests tion of fertilizer. Moreover, as discussed in the previous that climate change and policies to mitigate emissions of section, growth of the biofuels industry has strengthened Structural Changes in Commodity Markets: New Opportunities and Policy Challenges for Commodity Exporters 267 the connection between energy and agriculture by linking capita consumption of metals is following historical demand for food crops to energy prices. The energy mar- trends of other rapidly growing developing countries at ket is so large relative to the volumes of biofuels that can similar income levels (figure 20.6). However, China is be produced from food crops that energy prices will such a large country, and its economic growth has been so drive up food crop prices as long as policies encourage rapid that following historical trends in per capita con- biofuel consumption and the use of food crops for biofu- sumption has led to very large increases in metals consump- els is unrestricted.6 tion and a growing share of global imports. If historical Efforts to mitigate greenhouse gas emissions may also increases in China's economic growth and exports con- contribute to the increase in biofuel production. Many tinue, per capita metals consumption is expected to con- countries appear committed to reducing dependence on tinue increasing for at least another decade. Indeed, China's fossil fuels and increasing reliance on renewable energy share of global metals consumption is forecast to increase sources in the long term. Brazil, Canada, the EU, the United from 40 percent to 50 percent or more for many metals States, and other countries have mandated large increases during this period (Mitchell, Tan, and Timmer 2007). in biofuel production, and other countries have set non- Once China moves through its period of rapid eco- binding targets (China, India, Japan, and others).7 Con- nomic growth, its demand for metals will fall, and no other sumption mandates in many countries, but especially in countries are expected to replace China's demand for met- the EU and the United States, will require large increases in als and minerals. As a result, metals prices could be sharply biofuels production through at least 2020.8 lower in a decade or more. Although supplies of global In the long term, second-generation biofuel technology metals and minerals deposits are abundant (USGS 2008),9 may limit the upward pressure on prices by allowing agri- many of these deposits will probably be costly to develop cultural waste to be used to produce biofuels. However, this and under the best of conditions will take many years to technology has been slow to develop and may not be eco- begin producing. nomically viable for a decade or more. Even when it In conclusion, while the boom in commodity prices becomes profitable, it will take many years to scale up pro- may be long lasting, it is unlikely to be permanent because duction to significant levels. And until second-generation commodity demand will slow in rapidly growing develop- technology develops, the increased use of food crops for ing countries as their economies mature and new resource biofuels will divert resources from other food, feed, and supplies are developed. However, the period of high prices fiber uses and raise all agricultural commodity prices and strong demand growth may last for at least a decade (Mitchell forthcoming). and provide an opportunity for resource-rich countries to If crop prices remain high, agricultural production may enjoy faster growth than in the past. How they manage increase. Argentina, Brazil, and other parts of Latin America such growth will largely determine whether it will be sus- (as well as Eastern Europe and Sub-Saharan Africa) have tainable. Past efforts at managing commodity booms have land that is suitable for crops and would likely come into not been very successful. For example, countries with production. Much of this land is already in pasture and abundant natural resources have experienced slowed eco- could be used for crops if the necessary infrastructure were nomic growth because of overvalued exchange rates and made available. Brazil, for example, has about 200 million volatile export earnings. However, countries that manage hectares of pasture, and a portion of this land could be their policies well can avoid these problems. used for crop production without undue environmental stress. This area potentially rivals the 140 million hectares A Historical Account of the of cropland currently used for production of major field Commodities Literature crops in the United States. Sub-Saharan Africa has more than 1 billion hectares of land with potential for rain-fed Recent trends in commodity prices have revived discussion crop production, with less than one quarter of this land of the differing impact of commodity and manufactures now being cropped, according to the Food and Agriculture production on economic growth. This debate has been Organization (FAO 2008). Marginal and abandoned lands emerging cyclically since the 1950s, starting with Raul Pre- may be even more abundant. bisch (1959) and Hans Singer (1950). Prebisch and Singer argued that natural-resource dependency leads to low growth because of the long-term deterioration of com- The Future of Metal and Minerals modity prices relative to manufactures and the belief that The future of metal and mineral prices will depend opportunities for technical progress were limited relative mostly on demand from China. Currently, China's per to those in manufactures. Their analyses were further 268 Managing Openness Figure 20.6. China's Metals Consumption versus GDP as Compared to Selected Other Economies, Various Years a. Aluminum consumption b. Copper consumption 25 20 20 metal consumption 16 metal consumption (kg per capita) (kg per capita) 15 12 10 8 4 5 0 0 0.0 4.0 8.0 12.0 0.0 4.0 8.0 12.0 real GDP per capita (PPP adjusted) real GDP per capita (PPP adjusted) c. Steel consumption d. Zinc consumption 1,000 10.0 metal consumption 800 metal consumption 7.5 (kg per capita) (kg per capita) 600 5.0 400 2.5 200 0 0.0 0.0 4.0 8.0 12.0 0.0 4.0 8.0 12.0 real GDP per capita (PPP adjusted) real GDP per capita (PPP adjusted) India Brazil Taiwan, China Korea, Rep. China Source: Mitchell, Tan, and Timmer 2007. Note: Data for all countries ends in 2005, and the beginning year for each country is shown in parenthesis: Brazil (1964); China (1962); EU-12 (1960); India (1975); Taiwan, China (1975); and the United States (1930). kg = kilogram, PPP = purchasing power parity. supported by Kaldor's (1967) empirical observation of a link a country's volatile real exchange rate--resulting from the between industrial growth, productivity growth, and GDP. inherent price volatility of natural resource exports--fuels In the 1970s, the so-called Dutch disease phenomenon uncertainty about future economic conditions and thus may linked the exploitation of natural resources and declines in reduce investment (Gylfason, Herbertsson, and Zoega the manufacturing sector. While a boom in production of 1999). Third, the greater income inequality associated with natural resources increases wealth in the economy, it may Dutch disease can make governance issues and internal also drive up the real exchange rate and draw away factors conflicts more prevalent. Finally, Dutch disease can lead to of production from the nonresource tradable sector to corruption and protectionist policies for lagging industrial the resource tradable sector and the nontradable sector. sectors, which, in turn, may increase inefficiencies. These outcomes cause the nonresource tradable sector to However, recent research has shown that many of the lose competitiveness and shrink. shortcomings of commodities identified in the past are no Dutch disease can cripple a country's long-term pros- longer relevant. For instance, a more recent analysis of the perity in four ways. First, if natural resources begin to run Prebisch-Singer hypothesis concludes that from the begin- out or if there is a downturn in commodity prices, the ning of the 20th century to 1973 there was no trend in (previously) competitive manufacturing industries will relative commodity prices (Cuddington, Ludema, and have difficulty recovering because they will have become Jayasuriya 2007). The recent price increase and apparent technologically backward (Van Wijnbergen 1984). Second, structural break reinforce this view. Moreover, Martin and Structural Changes in Commodity Markets: New Opportunities and Policy Challenges for Commodity Exporters 269 Mitra (2001) show that total factor productivity (TFP) (De Ferranti et al. 2002). This phenomenon can be seen in growth in agriculture from 1967 to 1992 was 50 percent the Scandinavian countries that built on their strengths faster than in manufactures, although the industrial coun- in natural resources to produce aircraft, luxury cars, tries experienced rates substantially above those of less designer furniture, and, most recently, advanced telecom- developed countries. Lederman and Maloney (2007) draw munications products. Sweden's Volvo and Saab emerged on this work and find econometric evidence that natural- partially from backward linkages with the forestry indus- resource-based activities can spur long-term growth. Sim- try. Similarly, Finland's Nokia, originally a wood-pulp ilarly, De Ferranti et al. (2002) argue that several of the producer, became a major player in the global cellular natural resource success stories--Denmark and Sweden-- phone industry. continue to show the highest TFP growth rates in the agri- In light of these findings, the gap between the manufac- culture sector. tures and commodity camps can be bridged by distin- De Ferranti et al. (2002) and Lederman and Maloney guishing between high-productivity and low-productivity (2007) also question the view that manufacturing has activities, within both manufactures and commodities. something special in terms of backward and forward Rodrik (2009) argues that the engine of growth since the linkages, technological innovation, and other potential end of the Second World War has been rapid structural externalities. The most convincing evidence is offered by change in the developing nations from low-productivity history: Australia, Canada, Finland, Sweden, and the United ("traditional") to high-productivity ("modern") activities. States developed and achieved technological progress Although Rodrik is seen as a proponent of manufactures, based on their natural resources. The main driver of he acknowledges that high-productivity activities are not growth and industrialization was mining in Australia and exclusively in the manufacturing sector: they are also pres- the United States and forestry in Finland and Sweden. And ent in commodities (for example, horticulture) and serv- even today, these countries remain exporters of natural- ices (for example, call centers). As De Ferranti et al. (2002) resource-based products, along with high-tech products and Lederman and Maloney (2007) note, what is impor- (De Ferranti et al. 2002). The success of Chile over the past tant is not what is produced but how it is produced--the two decades has been almost fully led by exports of natural- ability to produce and commercialize knowledge. A key resource-based products. policy objective of countries with abundant natural Moreover, the remaining shortcomings of commodities resources should therefore be to embed knowledge into can be overcome with sound policy choices. The historical natural-resource production. record is clear: when managed well and coupled with a strong institutional framework, natural resources can be vital for development. For instance, strong evidence now Policy Recommendations to Foster a Sectorally Balanced, suggests that the U.S. industrial success resulted from a Resource-Based Economy gradual transition to resource-intensive manufacturing industries and most recently in its development to more High commodity prices offer countries with natural knowledge-intensive industries. Along these lines, the case resources an opportunity to generate strong, broad-based of Canada inspired "staples theory"--where exports of pri- growth. Indeed, commodities can generate needed revenue mary goods drive development over an extended period of to improve competitiveness in the overall economy and pro- time through either demand or supply links. And although mote development. However, the increase in export earnings wool is Australia's most famous staple, success in mining can also result in Dutch disease. If the government does not (and the derivative industries of both) made that country implement policies to counter this possibility, the economy one of the richest economies in the world by the early 20th may become concentrated in the resource and nontradable century. Australia's discovery of new deposits and the gener- sectors. Because the resource sector is not labor intensive, ation and export of mining-related knowledge--in mineral higher unemployment, increased exposure to costly boom- detection, environmentally sound mining practices, and and-bust cycles, and unstable growth could result. processing, all based on a massive educational and research Appropriate policies can mitigate the inherent risks of infrastructure--may put it near the top of the list again. a commodity-dependent economy. Interventions can In particular, natural resources seem to have furthered mitigate the impact of price volatility, limit exchange- development in countries with strong organizational rate appreciation and increase competitiveness of the structures, knowledge networks, and aggressive human tradables sector, for example. The following section dis- capital policies (which were developed to pursue resource cusses how policy measures can foster sustainable growth processing but were transferable to high-tech industries) in commodity-dependent countries. 270 Managing Openness Dealing with High and Volatile Commodity Prices employment-based safety net program, with 8 million ben- eficiaries in 2008; and Indonesia's unconditional cash trans- In addition to remaining high, commodity prices are also fer for the poor, the largest cash transfer in the world, which likely to experience considerable volatility over the next covered 15.5 million households in 2008 (28 percent of the decade. High commodity prices will continue to create population) (FAO 2008). income opportunities for producers in resource-rich coun- Historical experience also highlights certain policies to tries. However, price volatility will have negative macroeco- be avoided. First, policy makers should steer clear of inter- nomic and microeconomic effects. It can lead to higher ventions that are expected to generate huge efficiency, equity, inflation and unemployment and could cause variability in or fiscal costs over the medium term. Second, policies GDP, exchange rates, and public expenditures. The effect of that are very costly to introduce and maintain should be such volatility on these variables will be greater in avoided, because they may create disincentives on the sup- resource-rich economies, where the link between com- ply side or promote black markets and inefficiencies. modity prices and revenues is stronger. Third, it is important to consider whether policies are not As shown in table 20.3, historical experience suggests well targeted or even biased against the poor. For instance, that some policy options are better than others at mitigating universal subsidies are poorly targeted and difficult to the impact of price volatility.10 For instance, several short- remove once introduced. Price controls are also biased run measures have proven effective at reducing the impact against the poor, difficult to remove, and conducive to of volatility on the poor. These include transfers to target black market promotion. groups (for instance, cash transfers, feeding programs, and Fourth, interventions should be avoided if they are food for work), responses to macroeconomic problems pro- politically difficult to scale back or remove down the line. voked by rising food prices (balance of payment support, For example, in 2007 and 2008, Indonesia's government for example), and quick measures to expand supply and spent almost two and a half times as much on energy sub- reduce prices (lifting of import restrictions, drawing down sidies as on social assistance. Although subsidy spending of food stocks, and reduction of tariffs and other taxes on was finally scaled back in May 2008, about two-thirds of key staples, for example). Four examples of targeted safety fuel subsidies still benefit the top 20 percent of the popula- net measures that were successfully expanded during the tion, and the bottom 10 percent receive only 1 percent of past food crisis are Mexico's conditional cash transfer (Pro- the direct benefits (see figures 20.7 and 20.8). gresa, now called Oportunidades), which covered 5 million Fifth and finally, food exporters should avoid applying households in 2008 (one out of four Mexican families); "beggar-thy-neighbor" policies, which consist of restric- Bangladesh's Public Food Distribution System, which cov- tions or taxes on food exports. These policies push world ered 30.5 million people during fiscal 2007­08; Ethiopia's Table 20.3. Summary of Policy Options for Mitigating the Impact of Price Volatility Transition toward market Transfers to poor households Public price stabilization stabilization measures Immediate responses · Feeding programs · Lifting of import restrictions on · Reduction of red tape in food and quotas transporting goods across regions · Food for work programs · Draw down of food stocks · Limiting of intervention using variable tariff · Expansion of existing cash-transfer programs · Limited subsidies Long-term responses · Development of cash-transfer · Improvement of farm productivity · Encouragement of investments in programs (where previously · Improvement of village private storage and nonexistent) or improvement infrastructure warehouse receipt of targeting (where existent) · Improvement of food logistics · Forward contracts network · Domestic market efficiency · Future Market, Index-based weather insurance Policies to avoid · Universal subsidies · Export bans · Import quota or import bans · In-kind transfers · Price controls · Price controls Source: Based on a World Bank note prepared by Cord et al. 2008. Structural Changes in Commodity Markets: New Opportunities and Policy Challenges for Commodity Exporters 271 Figure 20.7. Energy Subsidies in Indonesia, 2008 place. It is generally better to set a "tolerable" level of price variability than try to stabilize prices too much. Further- 21 more, the burden of proof for the need for stabilization electri- should rest on the private sector (rather than on the gov- city* ernment) to ensure that complaints are backed by data and US$, billions 14 decisions are based on cost-benefit analyses. It is also advisable to use variable tariffs and targeted 7 fuel* subsidies, because they are less costly and more effective than public marketing agencies. If the government chooses 0 to intervene, it is preferable to rely more on the use of a s t s en ie m small financial reserve for imports or domestic purchases id ra m bs og st of privately held stocks rather than relying mainly on gov- su e pr nv al li ci ernmental strategic commodity reserves. It is less costly for ta so pi ca the government to tender public procurement, imports, central government spending category and storage to the private sector. This approach will Source: World Bank staff calculations based on data from the Ministry of Finance of Indonesia. increase efficiency and develop storage capacity in the pri- * assumes ICP oil at US$95. vate sector. The government can be more effective at help- ing the poorest consumers deal with commodity price Figure 20.8. Benefits of Fuel Subsidies for the Rich and the shocks by strengthening its social safety net program. Poor, 2007 Strengthening market price stabilization (private stor- age and warehouse receipts, forward contracts, futures 50 markets, and index-based weather insurance) would allow market-based instruments to help stabilize short-term 40 price fluctuations. Moreover, strong intermediary institu- tions are needed to pool and repackage the risks facing small-scale producers, traders, and processors and then 30 hedge the pooled risks using global futures, options, and percent insurance markets. Improved statistical capacity, technical support, and education are also needed to facilitate use of 20 global futures and options markets by large domestic firms. Indeed, market stabilization mechanisms can help stabilize prices, thus minimizing the need for government 10 stabilization and saving public funds. At the same time, it is important to develop an environ- ment that promotes a quick supply response that benefits 0 1 2 3 4 5 6 7 8 9 10 producers and helps reduce a sudden price spike. A key step poor rich toward this objective is to expand the generally neglected household consumption decile agricultural extension services. It is also important to Source: World Bank staff calculations based on 2007 Susenas data. strengthen institutions that support rural finance markets and expand the availability of credit. Finally, regulations prices even higher, as the rice price bubble in the first half should be streamlined to ensure competitive behavior and of 2008 clearly illustrated (Aldaz-Carroll, Slayton, and reduce transport costs. As the national market becomes Timmer 2010). more integrated and subregional trade flows increase, price In the medium to long term, countries need to make differences across regions and regional shocks will diminish. their price stabilization policies more predictable, better targeted, less costly, and more effective. Above all, policy Complementary Policies for Promoting Growth makers need to develop and publish well-defined stabiliza- in Commodity-Dependent Countries tion strategies, moving from discretionary to rule-based interventions. For policy formulation to be transparent, Humphreys, Sachs, and Stiglitz (2007, 323­26) identify consistent, and consultative, countries must ensure that several other policies that can help countries maximize standard operating procedures for price stabilization are in the economic benefits of a natural-resource boom while 272 Managing Openness minimizing governance issues and appreciation pressure logistics costs can be caused by underdeveloped or poorly on the exchange rate: maintained infrastructure and a constraining business environment for logistics services. · Contracts on future prices, output, and other economic Second, governments should also foster foreign direct circumstances should be designed to avoid situations in investment (FDI) flows and maintain an open economy to which corporations receive an unconscionably high attract the skills and knowledge needed. Greater opening to return and the government a marginal amount. FDI is crucial for attracting needed technological and mar- · Auctions should foster increased competition between ket knowledge that can lead to greater export diversification corporations to reduce the risk of cronyism and increase (within and outside natural-resource-based activities), returns to the country. higher intra-industry trade, innovation and skills promo- · Contracts should be made transparent, and domestic tion (De Ferranti et al. 2002). Establishing an enabling envi- constituents should be involved in contract evaluation ronment for FDI by improving the business regulatory to reduce corruption and increase the bargaining power environment, human capital, public infrastructure, and of domestic negotiators vis-à-vis corporations. knowledge clusters is therefore crucial. Usual regulatory · Corporations should be required to post bonds in antic- issues inhibiting FDI in commodity-abundant countries ipation of future cleanups to protect the environment. tend to be those surrounding land entitlement, labor laws, · National wealth should be calculated by taking into and conflicting regulations. account not only earnings but also depletion of stocks and Third, government policies should foster the develop- degradation of the environment. Doing so will avoid the ment of knowledge industries in natural-resource-based temptation to spend too much of the depletable resource. activities where the country has a comparative advantage. · Action should be taken to prevent inequities and poten- Natural-resource-based activities can become knowledge tial social disturbance by ensuring a fair distribution of industries. That was the case, for instance, with mining in benefits within the country.11 the United States, which led to the development of a strong · Expenditures should be stabilized to avoid costly technological system; forestry and forest products in Finland boom-and-bust patterns. A sovereign wealth fund can and Sweden, which are highly knowledge intensive; and be established to sterilize the boom revenues and spread fresh-fruit production and marketing in Chile, which have expenditures across time, transforming a nonrenewable a high technology content. Developing a dynamic natural- resource into a financial asset that will last forever. resource-based sector is not incompatible with building new comparative advantages in footloose and high-tech In addition, governments should pursue a range of policies manufacturing. Both sectors coexist not only in natural to increase the technological sophistication and dynamism resource­rich developed economies but also in the already of tradable sectors. First, they should improve logistics. Poor highly diversified export structures of Brazil and Mexico logistics reduce the competitiveness of products and can (De Ferranti et al. 2002). For success, natural resource wealth inhibit greater processing of commodities and the export needs to be complemented with human capital, knowl- of higher-value-added products. The high cost of transport edge, and good institutions. For instance, a country can leads to higher prices for products and inputs in remote move from just extracting minerals to promoting the areas and to producers' having poor access to domestic development of engineering services for mineral extraction and international markets. Furthermore, in countries and from just producing raw commodities to downstream where the commodity-producing regions are poorly con- higher value and knowledge derivatives. nected to industrial areas, commodities are more likely to Fourth, governments need to build new endowments in be exported in raw form to be processed abroad rather human capital and knowledge. As discussed in De Ferranti than domestically. For instance, as a result of the high et al. (2002), a key lesson from the experience of other nat- domestic logistics costs to Java-based processing facilities, ural-resource-rich countries is that investing in knowledge some high-quality commodities with great potential, such and skills to build "new endowments" is crucial to promot- as shrimp from eastern Indonesia, cannot be exported, and ing sustained and inclusive growth. Public policy has a others, like pineapples, are canned abroad. A weak distribu- large role to play by providing quality education in general tion and marketing network can also discourage a country and lifelong training to support product upgrading and from venturing into higher-value-added downstream activ- innovation and by promoting research and development ities. High cost and uncertainty in distribution channels can incentives and innovation systems to lead to knowledge and also prevent countries from integration into just-in-time technological progress that raises productivity growth and production networks of higher-value-added products. High promotes new comparative advantages. In addition, public Structural Changes in Commodity Markets: New Opportunities and Policy Challenges for Commodity Exporters 273 policy should promote the establishment of knowledge clus- support services, market research, and policy advocacy. ters and networks, encompassing private firms, independent Export promotion services have been shown to be poten- research institutions and universities and the public sector. tially beneficial in terms of returns to invested public funds This step is critically important not just for high-tech manu- (see Lederman, Olarreaga, and Payton 2006) and could factures but also for many natural resource activities. possibly help lower the costs of entering export markets Fifth, governments should develop better institutions. To and promote diversification (Iacovone and Javorcik 2008). achieve productivity growth and develop comparative But export promotion needs to be well targeted and appro- advantages in technologically sophisticated industries, gov- priately designed. ernments need good institutions (rule of law, security, effective property rights, transparency, removal of excessive Conclusions regulatory burdens, and efficiency of public service deliv- ery). The experience of various Latin American countries The main conclusion of this chapter is that countries with with rich natural resources shows that the "curse" of natu- abundant natural resources have new opportunities. To ral resources appears to be driven not simply by the exis- capitalize on these opportunities, the challenge for such tence of rich endowments of natural resources per se but economies is to improve the investment climate, build bet- rather by multiple barriers to technology adoption usually ter institutions, and improve infrastructure, human capi- associated with weak institutions, burdensome regulations, tal, and knowledge while maintaining price stability. and artificially created monopolies (Lederman and Mal- Windfalls from commodity revenues should be invested oney 2007). to increase the technological sophistication of production Finally, governments should promote the development of across sectors. The competitiveness of the domestic and the services sector. Services have become a significant export sectors can be increased by relieving infrastructure engine of growth in East Asia since the East Asian crisis, bottlenecks, reducing regulatory uncertainty, and promot- and they continue to have great potential for further devel- ing the development of a knowledge economy. If these goals opment. Developing a competitive services sector will can be achieved, greater technology absorption and the increase the competitiveness of other sectors and facilitate development of knowledge industries in natural resources diversification into higher-value-added products, spurring and other sectors will spur sustainable growth. dynamism in the economy. For instance, in Indonesia one- The government should also redistribute the resource third of inputs used in the mining sector are services and windfall in a way that promotes social and political stability around 15 percent in the case of agriculture, forestry, and and mitigates the negative effects of volatile commodity fishing (Atje and Rahardja 2010). More specifically, greater prices on vulnerable households. Safety net approaches like efficiency of services would result in reduced service costs, cash transfers and food aid are more effective than price greater variety of services, increased investments in related controls in ensuring food security. To lower domestic and supported sectors, enhanced goods and services prices, the more effective options are to reduce tariffs and exports, and higher economic growth. Improving access to other taxes on key staples. Other policy actions that aim to competitive cyber-communication, design and marketing lower domestic prices may be administratively easy to services, financial instruments, and tailored supply chain implement but often have unintended consequences. How- management, for example, would raise the efficiency of ever, the government should not attempt to protect all local producers in upgrading their products and finding consumers from price changes since that effort can lead to new market niches, making them more competitive and burdensome social programs that limit investment in areas increasing global market shares. that will foster sustained growth. The development of information and communications technology can reduce coordination and marketing costs and enable firms to hook into international industrial clus- Notes ters, expand resource-based tourism, and facilitate mar- 1. Commodities in this chapter refer to agricultural commodities, ket access for nontraditional agricultural products and fuels, ores, and metals. 2. The depression of commodity prices in the 1980s and 1990s was processed goods. For instance, setting up a website has caused by two main factors: overinvestment (in response to sharp price allowed companies to serve the global market with little in increases in crude oil in the 1970s) and surplus capacity (as a result of the the way of advertising costs and achieve significant growth collapse of demand for metals and minerals in the former Soviet Union). 3. These countries, ranked from highest to lowest based on GDP growth in sales. Some services like export promotion are "public rates (percent), were China (10.5), Cambodia (8.5), Vietnam (7.9), Myanmar goods" that can be supported by the public sector. These (8.1), Uganda (7.1), Mozambique (7.1), India (6.9), Lao People's Democratic export promotion services include image building, export Republic (6.6), Sudan (6.3), Malaysia (6.2), Botswana (5.7), Burkina Faso 274 Managing Openness (5.7), Jordan (5.7), Chile (5.6), the Dominican Republic (5.4), Costa Rica Iacovone, L., and B. S. Javorcik. 2008. "Using Export Promotion to Stim- (5.4), Sri Lanka (5.3), and Bangladesh (5.3) (World Bank 2009b). ulate Export Diversification in Indonesia." Mimeo, World Bank, 4 Production levels increased as a result of increased fertilizer use, Washington, DC. expanded irrigation, improved seeds, and slowing population growth. IEA (International Energy Agency). 2008. World Energy Outlook. Paris: 5. Calorie supplies measure calories available for consumption and Organisation for Economic Co-operation and Development. are used as a proxy for calorie consumption since actual consumption is Kaldor, N. 1967. Strategic Factors in Economic Development. Ithaca, NY: difficult to estimate and not available for most countries. New York State School of Industrial and Labor Relations, Cornell 6. Current biofuels production accounts for only 1.5 percent of global University. transport fuels. Lederman, D., M. Olarreaga, and L. Payton. 2006. "Export Promotion 7. These mandates and targets are not tied directly to energy prices Agencies: What Works and What Doesn't." Discussion Paper 5810, and should continue regardless of energy prices unless new legislation Centre for Economic Policy Research, London. repeals existing laws. Lederman, D., and W. F. Maloney. 2007. "Trade Structure and Growth." 8. The EU has mandated that 10 percent of transport fuels come from In Natural Resources: Neither Curse nor Destiny, ed. D. Lederman and renewable sources by 2020, which will require almost tripling the approx- W. F. Maloney, 15­40. Palo Alto, CA: Stanford University Press. imately 15 billion liters of biofuels consumed in 2009. The United States Martin, W., and D. Mitra. 2001. "Productivity, Growth and Convergence has mandated that 36.0 billion gallons (136 billion liters) of biofuels be in Agriculture and Manufacturing." Economic Development and Cul- consumed by 2022, which requires more than tripling the 11.1 billion gal- tural Change 49 (2): 403­22. lons (42 billion liters) of biofiuels consumed in 2009. Mitchell, D. 2008. "A Note of Rising Food Prices." Policy Research Work- 9. The U.S. Geological Survey is a government body assigned to ing Paper 4682, World Bank, Washington, DC. quantify and monitor such deposits. Mitchell, D. Forthcoming. Biofuels in Africa: Prospects for Sustainable 10. Some of these recommendations come from a World Bank note Development. Washington, DC: World Bank. prepared in 2008 by Louise Cord, Eduardo Ley, Hassan Zaman, Elena Mitchell, D., S. Tan, and H. Timmer. 2007. "China's Demand for Primary Ianchovichina, C. Hull, Emmanuel Skoufias, Mark Thomas, Brian Pinto, Commodities and the Impacts on Global Markets." Unpublished and Tahrat Shahid. manuscript, World Bank, Washington DC. 11. Humphreys, Sachs, and Stiglitz (2007) suggest that given the Mulat Demeke, M., G. Pangrazio, and M. Maetz. 2009. Country Responses weaker capacity of subnational governments to manage extreme to the Food Security Crisis: Nature and Preliminary Implications of the volatility of revenues and ensure oversight, it is better in most country Policies Pursued. Rome: Food and Agriculture Organization. cases to centralize revenue collection while allowing for decentralized Prebisch, R. 1959. "Commercial Policy in the Underdeveloped Countries." expenditure to subnational governments. A redistribution of revenues American Economic Review: Papers and Proceedings 49 (2): 251­73. between regions and the set-up of social protection to assist the poor Rodrik, D. 2009. "Growth after the Crisis." Paper prepared for the Com- will help make growth more inclusive and reduce potential political mission on Growth and Development. http://www.growthcom instability. mission.org/storage/cgdev/documents/financial_crisis/rodrikafterthe crisis.pdf. Sachs, J. D., and A. Warner. 1995. "Natural Resource Abundance and Eco- Bibliography nomic Growth." Working Paper 5398, National Bureau of Economic Research, Cambridge, MA. http://www.nber.org/papers/w5398.pdf Aldaz-Carroll, E., T. Slayton, and P. Timmer. 2010. "Pricking the Price ------. 2001. "The Curse of Natural Resources." European Economic Bubble to Avert a World Rice Crisis." Background paper, World Bank, Review 45 (4­6): 827­38. Washington, DC. Singer, H. W. 1950. "The Distribution of Gains between Investing and Atje, R., and S. Rahardja. 2010. "Indonesia's Services Industries: Questions Borrowing Countries." American Economic Review 40 (2): 473­85. Moving Forward." Presentation at workshop "Dialogue among Stake- USDA (U.S. Department of Agriculture). 2010. PSD Online (database). holders: Growth and Efficiency in the Indonesian Services Sector," USDA, Washington, DC. http://www.fas.usda.gov/psdonline. organized by the World Bank, Kadin, and the Center for International USDOE (U.S. Department of Energy). 2009. Annual Energy Outlook 2009 Studies, Jakarta, January 14. with Projections to 2030. Washington, DC: USDOE. BP (British Petroleum). 2009. Database on Energy Demand, Supply and USGS (U.S. Geological Survey). 2008. "Assessment of Undiscovered Oil Prices. British Petroleum, London. Resources in Devonian-Mississippian Bakken Formation Williston Cuddington, J., R. Ludema, and S. Jayasuriya. 2007. "Prebisch-Singer Basin Province, Montana and North Dakota, 2008." USGS, Reston, Redux." In Natural Resources: Neither Curse nor Destiny, ed. Daniel Virginia. Lederman and William Maloney, 103­40. Washington, DC: World Bank Van Wijnbergen, S. 1984. "The `Dutch Disease': A Disease after All?" Eco- and Stanford University Press. nomic Journal 94 (373): 41­55. De Ferranti, D., G. Perry, D. Lederman, and W. Maloney. 2002. From Nat- World Bank 2009a. Global Economic Prospects: Commodities at the Cross- ural Resources to the Knowledge Economy. Washington, DC: World roads 2009. Washington, DC: World Bank. Bank. ------. 2009b. World Development Indicators 2009. Washington, DC: FAO (Food and Agriculture Organization). 2008. The State of Food and World Bank. Agriculture 2008: Biofuels: Prospects, Risks and Opportunities. Rome: ------. 2010a. Commodity Price Data. Washington, DC: World Bank. Food and Agriculture Organization. http://econ.worldbank.org/WBSITE/EXTERNAL/EXTDEC/EXTDEC ------. 2010. FAOSTAT (database). Food and Agriculture Organization, PROSPECTS/0,,contentMDK:21148472~menuPK:556802~pagePK: Rome. http://faostat.fao.org. 64165401~piPK:64165026~theSitePK:476883,00.html. Gylfason, T., T. T. Herbertsson, and G. Zoega. 1999. "A Mixed Blessing." ------. 2010b. Global Commodity Markets: Review and Price Forecasts. Macroeconomic Dynamics 3 (2): 204­25. Washington, DC: World Bank. Humphreys, M., J. D. Sachs, and J. E. Stiglitz. 2007. "Future Directions for ------. 2010c. Global Economic Prospects: Crisis, Finance, and Growth. the Management of Natural Resources." In Escaping the Resource Washington, DC: World Bank. Curse, ed. M. Humphreys, J. D. Sachs, and J. E. Stiglitz, 233­36. New World Bureau of Metals Statistics. 2010. London. http://www.world- York: Columbia University Press. bureau.com. 21 Global Production Networks in the PostCrisis Era Alyson Ma and Ari Van Assche Vertical specialization has been one of the most notable In this chapter, we take advantage of a unique data set trends in the international organization of production on China's processing trade regime from 1988 to 2008 to during the past few decades (Spencer 2005; Helpman analyze the impact of trade costs on intra-GPN trade. 2006; Desai 2009). Thanks to reductions in communica- Under this customs regime, firms are granted duty tion costs, transportation costs, and other trade barriers, exemptions on imported raw materials and other inputs multinational firms have sliced up their supply chains and as long as they are used solely for export purposes. As a dispersed their production activities across multiple coun- result, the data set provides information on trade among tries. As a result, each production step in the supply chain three sequential nodes of a vertically specialized global of a single final good is performed in the most advanta- supply chain: the location of input production, the loca- geous location for that particular process. tion of processing (in China), and the location of further A recent series of macroeconomic events, however, is consumption. This makes it possible to examine the role of threatening to reverse this trend of vertical specialization. both trade costs related to the import of inputs (upstream First, the era of cheap and plentiful oil seems to be draw- trade costs) and trade costs related to the export of final ing to a close, thus leading to increases in transportation goods (downstream trade costs) on intra-GPN trade. costs (Rubin and Tal 2008; Rubin 2009). Second, the The rest of this chapter is organized as follows. In the global recession of 2008­09 has put pressure on many next section, we discuss the key forces driving the vertical governments to raise trade and investment barriers that specialization process. In the following section, we analyze privilege domestic over foreign production. Because ver- the role that both upstream and downstream trade costs tical specialization is considered especially sensitive to play in China's processing trade. In the subsequent section, such changes in trade costs (Yi 2003), these trend reversals we investigate whether rising oil prices have rendered could cause firms to rethink their global production net- intra-GPN trade more sensitive to distance. The next sec- works (GPNs) and abandon distant production sites for tion analyzes the role that GPNs played in the great export closer locations. collapse that occurred during the recent global recession. What will GPNs look like in the postcrisis era? To what The final section concludes. extent will they become smaller if these macroeconomic events continue? To evaluate these questions, observers Drivers in the Spread of GPNs must understand and quantify the sensitivity of intra-GPN trade to trade costs. Yet, because of data limitations, empir- Reductions in trade costs are generally considered the key ical research on this topic is scant. A notable exception is driving force behind the vertical specialization process. Hanson, Mataloni, and Slaughter (2005), who have exam- Three developments in the past few decades have been ined the role of trade costs in the decisions of U.S. multi- particularly important in decreasing trade costs within nationals to export intermediate goods to their foreign GPNs. First, the modularization of production in many affiliates for processing. industries has made it easier for firms to slice up their value 275 276 Managing Openness chain into separable production stages, thus reducing trans- us to overcome some of the shortcomings in the existing action costs. Second, technological innovations in commu- literature. Specifically, we use a data set collected by the nication and transportation have improved the efficiency General Administration of Customs of the People's Republic of coordinating geographically dispersed production of China on China's processing trade regime for the period stages. Third, liberalization policy reforms in home and 1988­2008. Under this regime, firms are granted duty host countries have considerably reduced barriers to exemptions on imported raw materials and other inputs as trade and investment. long as they are used solely for export purposes. As a result, A minimum prerequisite for vertical specialization is the the data set provides a unique mapping of the source coun- distinct separability of production stages. And in the past tries of processing imports and the destination countries of few decades, the modularization of products has signifi- processed exports for each Chinese processing location. This cantly increased the separability of production (Sturgeon information makes it possible to examine the role of both 2002; Gereffi, Humphrey, and Sturgeon 2005; Van Assche trade costs related to the import of inputs (upstream trade 2008). Modularity refers to a technological property of a costs) and trade costs related to the export of final goods product that describes how different components of a final (downstream trade costs) on trade within GPNs. Such good interact with one another. When a product is non- mapping of GPNs cannot be conducted with regular trade modular, components need to be specifically adjusted to data, since imports are not necessarily used solely for export one another, and it is thus difficult to locate their produc- purposes but can also be consumed domestically. tion in different places. In contrast, modular products con- China's processing trade regime was set up in the mid- sist of loosely coupled components that interact with one 1980s to attract foreign direct investment and promote another through well-defined and codified architectural exports. Largely ignored by many scholars, the regime standards. Advances in information technology in recent reached farther than similar systems introduced in other decades have enabled firms to standardize the interfaces East Asian countries. China's concessionary provisions between components more easily, thus allowing for pro- applied to its entire territory, unlike in neighboring coun- duction separability. tries where concessions were geographically limited to At the same time, technological advances have also export processing zones (Naughton 2006). As a result, driven vertical specialization through reduced communi- China's processing trade regime has played an important cation and transportation costs. In the past few decades, role in its overall trade performance. Indeed, between significantly cheaper communication technologies have 1988 and 2008, the share of processing exports (that is, become available (for example, international telephone exports conducted under the processing regime) in China's service, fax machines, and the Internet). These reductions total exports rose from 30 percent to 51 percent, while the in communication costs have made it easier for firms to share of processing imports in total imports increased coordinate, manage, and monitor production activities in from 27 percent to 38 percent (figure 21.1). The rest of this different locations. Furthermore, global transportation section discusses three stylized facts that suggest that both costs for distance and time-sensitive goods have declined upstream and downstream trade costs play an important (Hummels 2007, 2009; Behar and Venables 2010), benefit- role in China's processing trade. ing the international trade of goods produced in GPNs. A final driver of vertical specialization has been the con- tinued political efforts by governments to eliminate tariffs Figure 21.1. Processing Trade in China as a Share of Total and nontariff barriers. Multilateral trade negotiations under Trade, 1988­2008 the auspices of the General Agreement on Tariffs and Trade 60 and the World Trade Organization have helped reduce aver- 50 age tariffs for manufacturing products to under 5 percent 40 percent and have phased out quantitative restrictions. 30 20 Mapping Global Production Networks: 10 China's Processing Trade Regime 0 88 90 92 94 96 98 00 02 04 06 19 19 19 19 19 19 20 20 20 20 While trade costs reductions are generally accepted as the year main driver of vertical specialization, empirical evidence exports imports on the sensitivity of intra-GPN trade to trade costs remains remarkably scant.1 We exploit a unique data set that allows Source: Authors' calculations based on China's Customs Statistics. Global Production Networks in the Postcrisis Era 277 Figure 21.2. Domestic and Foreign Content as a Share of China's Processing and Nonprocessing Exports, 2006 a. Nonprocessing exports b. Processing exports foreign Chinese content content 11% 18% Chinese content foreign content 89% 82% Source: Koopman, Wang, and Wei 2008. First, China's processing exports rely heavily on foreign Figure 21.3. Share of China's Processing Imports by Region of Origin, 1988­2008 inputs, with a relatively low share of the value made in China. According to a recent estimate by Koopman, Wang, 90 and Wei (2008), only 18 percent of the value of China's processing imports (%) 80 processing exports is produced in China, while the remain- 70 ing 82 percent consists of the value of imported processing 60 50 inputs (figure 21.2). In comparison, the domestic content 40 share of China's nonprocessing exports is 88.7 percent, 30 meaning that imported inputs represent only 11.3 percent 20 10 of the export value. 0 Second, the average distance traveled by processing 88 90 92 94 96 98 00 02 04 06 08 19 19 19 19 19 19 20 20 20 20 20 imports (import distance) is shorter than the average dis- year tance traveled by processing exports (export distance). As East Asia non-Asian OECD countries shown in figure 21.3, 75.1 percent of China's processing rest of the world imports originated from within East Asia in 2008.2 In con- trast, Canada, the EU-19,3 and the United States together Source: Authors' calculations based on China's Customs Statistics. Note: OECD = Organisation for Economic Co-operation and Development. accounted for less than 19 percent of processing imports in 2008. This asymmetric sourcing pattern of processing Figure 21.4. Share of China's Processing Exports inputs has become more pronounced over time. Between by Destination, 1988­2008 1988 and 2008, the share of processing imports originating from China's most important East Asian trading partners 70 processing exports (%) rose from 59.6 percent to 75.1 percent, while the share of 60 processing imports originating from non-Asian coun- 50 tries in the Organisation for Economic Co-operation and 40 30 Development (OECD) decreased from 37.7 percent to 20 18.7 percent. 10 Conversely, most processing exports were destined for 0 non-Asian OECD countries from 1988 to 2008 (except 8 0 2 4 6 98 00 02 4 06 08 8 9 9 9 9 0 19 19 19 19 19 19 20 20 20 20 20 for an interlude between 1992 and 1997). As shown in year figure 21.4, the share of processing exports destined for East Asia non-Asian OECD countries non-Asian OECD countries rose from 54.7 percent in rest of the world 1997 to 59.4 percent in 2008. In contrast, the share of processing exports destined for the East Asian region Source: Authors' calculations based on China's Customs Statistics. 278 Managing Openness declined from 36.0 percent to 28.3 percent during the the exports that China sends to the EU-19, Japan, and the same period. United States are processing exports (figure 21.6). The This unbalanced processing trade pattern is generally share of processing exports is significantly lower for most attributed to the reorganization of GPNs in East Asia developing East Asian countries. (Yoshida and Ito 2006; Gaulier, Lemoine, and Ünal-Kesenci Third, this spatial pattern is not consistent across 2007; Haddad 2007). With rising costs in Japan and the processing locations. In a cross-section of 29 Chinese newly industrialized economies--Hong Kong SAR, China; provinces, the weighted average distance traveled by pro- the Republic of Korea; Singapore; and Taiwan, China--East cessing imports (import distance) has been negatively Asian firms are increasingly using China as a lower-cost correlated to the weighted average distance traveled by pro- export platform. Instead of directly exporting their final cessing exports (export distance) for most years between goods to the Western markets, these firms now export high- 1995 to 2008 (Ma, Van Assche, and Hong 2009). In other value intermediate goods to their processing plants in words, locations in China that import their processing China and then export them on to the West after assembly. inputs from nearby tend to export their processed goods As a result, a triangular trade pattern seems to have emerged far away and vice versa. in the GPNs in which China relies heavily on processing Our econometric analysis provides further empirical inputs from East Asia and sends processed goods predomi- support that China's processing trade depends heavily on nantly to the West. both upstream and downstream trade costs.4 Specifically, Moreover, the share of processing imports in China's we find that both import and export distance affect pro- total imports is greater for imports from East Asian coun- cessing exports negatively. Furthermore, we show that tries than for imports from outside East Asia. Except for processing exports to East Asian countries are more sensi- Indonesia and Vietnam, more than 35 percent of China's tive to export distance and less sensitive to import distance imports from its major East Asian trading partners were than processing exports to non-Asian OECD countries. As processing imports in 2007 (figure 21.5). Almost 40 percent a result, for firms in advanced East Asian countries, the key of its imports from Japan and between 40 and 60 percent distance factor that determines China's attractiveness as a of its imports from the newly industrialized economies processing location is its vicinity to Eastern input suppli- were aimed at supplying inputs for processing industries. ers, that is, import distance. As import distance increases, This share is significantly higher than for Western coun- China becomes less attractive as a location for processing tries. The share of processing imports in China's total activities, and therefore the volume of processed exports imports from the EU-19, Canada, and the United States from China decreases. Conversely, for firms in Europe and amounted to 15.4 percent, 17.6 percent, and 25.0 percent, North America, the critical determinant of China's attrac- respectively. tiveness as a processing location is its proximity to the East At the same time, processing exports represent a greater Asian market, that is, export distance. As export distance proportion of China's total exports to developed countries increases, China becomes less attractive as a location for than to its East Asian neighbors. More than 50 percent of processing activities. Figure 21.5. Processing Imports as a Share of China's Total Imports by Economy of Origin, 2007 70 processing imports (%) 60 50 40 30 20 10 0 na es lia m sia 9 da es sia Th n ng d e p. or -1 pa n in Re na at ra hi na ne ay la ap EU Ta ipp Ja St ,C st ai et al Ca do a, Au Vi M d re an il In ite Ph Si Ko iw Un economy Source: Authors' calculations based on China's Customs Statistics. Global Production Networks in the Postcrisis Era 279 Figure 21.6. Processing Exports as a Share of China's Total Exports by Destination Economy, 2007 70 60 processing exports (%) 50 40 30 20 10 0 m Th ia d Ca s Au a lia p. na 9 sia ng n ite ore es e d -1 pa Ph lan s Re na in at ra hi na ne ay Un ap EU Ja pp St st ,C ai et al do a, Vi M d ili re an In Si Ko iw Ta economy Source: Authors' calculations based on China's Customs Statistics. Peak Oil and Intra-GPN Trade Figure 21.7. Average Prices for Crude Oil, 1980­2009 Recent macroeconomic events may have reversed the 120 declining trend in trade costs, thus threatening to undo average price (US$ per barrel) some vertical specialization and making GPNs more local. 100 Among them, the era of cheap and plentiful oil seems to be 80 drawing to a close, thus leading to increases in transporta- 60 tion costs (Rubin and Tal 2008; Rubin 2009). Just how vulnerable, then, are GPNs to such changes in 40 trade costs? Will GPNs get smaller if oil prices continue to 20 rise in the postcrisis era? In this section, we provide new insights into these questions by investigating whether ris- 0 80 83 86 89 92 95 98 01 04 07 ing oil prices have rendered China's processing trade more 19 19 19 19 19 19 19 20 20 20 sensitive to distance. year As discussed in chapter 20 of this volume, oil prices have Source: International Financial Statistics (database), International Monetary risen dramatically in the past decade. While crude oil prices Fund, http://www.imfstatistics.org. were relatively stable and had even declined over the period 1980­99, they grew at an annualized rate of 20.6 percent with the oil price hikes, the cost of shipping a standard over the period 1999­2008 (figure 21.7). Oil prices retreated 40-foot container from Shanghai to the U.S. eastern during the global recession of 2008­09, but this decline is seaboard has risen from US$3,000 in 2000 to US$8,000 in likely a temporary phenomenon. As the global economy 2008. Other studies, however, have estimated that the sen- comes out of the recession, oil prices are expected to return sitivity of shipping freight rates to oil prices remains rela- to and even exceed precrisis levels. Oil experts generally tively low, thus limiting the threat that rising oil prices agree that peak oil will be reached before 2015 (De Almeida will make GPNs significantly smaller. Hummels (2007) and Silva 2009).5 When this moment occurs, the gap and UNCTAD (2010) estimate an elasticity of maritime between oil production and demand is expected to increase. cargo costs with respect to fuel prices of between 0.20 and As a consequence, the price of oil is expected to rise signifi- 0.40. Mirza and Zitouna (2009) and OECD (forthcoming) cantly and become more volatile. estimate an even lower elasticity of freight rates to oil Rubin and Tal (2008) argue that rising oil prices are prices ranging from 0.02 to 0.15. likely to lead to significant hikes in international trans- To determine if rising oil prices have made intra-GPN portation charges and thus reverse the trend of vertical trade more sensitive to distance, we have taken advantage specialization. They highlight the fact that, hand in hand of the panel data structure of the Chinese processing trade 280 Managing Openness data.6 According to our results for 1988­2008, China's pro- sectors,7 trade and especially intra-GPN trade have become cessing exports have become more sensitive to both import more intensive in durable goods than overall GDP. This and export distance in times of rising oil prices but not compositional change has made trade more sensitive to more sensitive to internal distance (between a processing fluctuations in business cycles.8 In times of recession, location and the nearest large seaport). We also find that households and companies tend to delay purchases of processing exports are more sensitive to oil price move- durable goods--not only because tightening budget con- ments than nonprocessing exports. Specifically, an increase straints render high-ticket goods unaffordable for some in oil prices tends to reduce the share of processing exports but also because consumers and firms want to postpone in total exports, especially when destined for far away their purchases until they know with more certainty countries. These results are in line with Yi's (2003) predic- whether and when the economic climate will improve. tion that intra-GPN trade is especially sensitive to changes Just as an economic crisis leads to a disproportionate in trade costs. drop in the demand for durable goods, so the composi- tional effect can explain the lopsided collapse in trade Intra-GPN Trade and the Global Recession compared to GDP. We find evidence of a compositional effect in China's What has been the impact of the global recession on exports during the global recession. Specifically, the sectors intra-GPN trade flows? This question has become partic- that contributed most to the export collapse in China are ularly relevant in light of the collapse of trade during the those in which vertical specialization is more prevalent.9 crisis, which was significantly larger than the drop in Figure 21.8 shows that vertical specialization in China is world gross domestic product (GDP). A number of schol- more common in the higher-technology categories than in ars have attributed the disproportionate trade collapse to the lower-technology categories.10 In 2007, processing vertical specialization. Barry Eichengreen, for example, exports accounted for 84.9 percent of high-technology has stated that "the most important factor is probably the exports; 45.6 percent of medium-high-technology exports; growth of global supply chains, which has magnified the 26.6 percent of medium-low-technology exports; and impact of declining final demand on trade" (Interna- 29.8 percent of low-technology exports. tional Economy 2009). Bems, Johnson, and Yi (2009) The higher-technology categories are also the indus- argue that "international supply chains are a leading con- tries that contributed most to the drop in China's exports tender for explaining why the great collapse was so great." during the global recession. High-technology exports The channels through which vertical specialization contributed to 37.6 percent of the collapse in exports; exacerbated the trade collapse, however, have become the medium-high-technology exports contributed to 25.4 per- source of a heated debate. Using a simple Barbie doll exam- cent; medium-low-technology exports contributed to ple, O'Rourke (2009) demonstrated that vertical specializa- 16.0 percent; low-technology exports contributed to tion does not automatically explain why the fall in world 10.9 percent; and other nonmanufacturing exports con- trade overshot the drop in world GDP. The fact that com- tributed to 10.1 percent (table 21.1). ponents of the Barbie doll cross borders multiple times to produce a final doll does not necessarily imply that a drop Figure 21.8. Processing Exports as a Share of China's Total in Barbie sales should lead to a disproportionate drop in Exports by Technology Level, 1992­2007 trade. To explain the role of vertical specialization in the trade collapse, scholars have therefore focused on three 95 processing exports (%) additional effects: a compositional effect, a trade cost 85 effect, and a bullwhip effect. The rest of this section dis- 75 cusses evidence for the existence of these three channels in 65 55 China's trade during the global recession. 45 35 The Compositional Effect 25 92 94 96 98 00 02 04 06 A number of studies have argued that vertical specializa- 19 19 19 19 20 20 20 20 year tion worsened the trade collapse through a compositional high technology medium-high technology effect (Francois and Woerz 2009; Levchenko, Lewis, and medium-low technology low technology Tesar 2009; Eaton et al. 2010). Because vertical special- ization has taken place primarily in the durable goods Source: Authors' calculations, based on China's Customs Statistics. Global Production Networks in the Postcrisis Era 281 Table 21.1. Breakdown of China's Exports by Level of Technology, 2008­09 Exports value Contribution to (US$, billions) Growth rate, Share of exports (%) total export growth, Q1 2008­Q1 2009 Q1 2008­Q1 2009 Category Q1 2008 Q1 2009 (%) Q1 2008 Q1 2009 (%) High technology 94.6 71.8 ­24.1 31.1 29.5 37.6 Medium-high technology 70.0 54.6 ­22.0 23.0 22.5 25.4 Medium-low technology 44.9 35.2 ­21.6 14.8 14.5 16.0 Low technology 73.8 67.2 ­8.9 24.3 27.6 10.9 Other 20.5 14.4 ­29.8 6.7 5.9 10.1 Total 303.8 243.2 ­19.9 100.0 100.0 100.0 Source: Authors' calculations based on China's Customs Statistics Data. Note: Q = quarter. The Trade Cost Effect Table 21.2. China's Processing Exports as a Share of Total Exports, HS Eight-Digit Level, Quarter 1, 2008, and A second set of studies argues that vertical specialization Quarter 1, 2009 intensified the trade collapse through rising trade costs Number of Standard associated with evaporating credit, increasing nontariff Variables observations Mean error barriers, and home bias in government stimulus plans Share of processing (Jacks, Meissner, and Novy 2009; Yi 2009). Jacks, Meissner, exports in total and Novy (2009) estimate that trade costs increased by exports, Q1 2008 4,760 0.31 0.004 Share of processing 11 percent on average between the second quarter of exports in total 2008 and the first quarter of 2009. As the previous sec- exports, Q1 2009 4,760 0.29 0.004 tion showed, intra-GPN trade is especially sensitive to Difference 9,520 0.020*** 0.003 changes in trade costs (Yi 2003), which can explain the Source: Authors' calculations based on China's Customs Statistics Data. greater fall in trade than in GDP. Note: Q = quarter. * = significant at 10 percent; ** = significant at 5 percent; Our analysis finds evidence of a trade cost effect in *** = significant at 1 percent. China's exports during the global recession. Controlling for compositional effects, we find that the share of China's processing exports in total exports at the eight-digit level downstream activities translates into a larger reduction in of the Harmonized System (HS) significantly declined the demand for upstream inputs.11 Because fluctuations in from the first quarter of 2008 to the first quarter of 2009 final demand become amplified as one moves upstream (table 21.2). This finding confirms that processing exports along the supply chain, this effect provides an alternative consistently dropped more than nonprocessing exports explanation of the magnified fall in trade compared to GDP. during the global recession. This result is in line with the Our analysis finds evidence of a bullwhip effect in trade cost effect and the findings of the previous section, intra-GPN trade during the global recession. The percent- given that intra-GPN trade should be more sensitive to age drop in China's processing imports was larger than trade costs than regular trade. that of processing exports in the first quarter of 2009 compared to a year earlier.12 China's processing exports dropped 23.7 percent, while processing imports declined The Bullwhip Effect 36.2 percent. Moreover, disaggregating the analysis at the A final explanation of how global supply chains intensified industry level provides further evidence of a bullwhip the trade collapse is the bullwhip effect (Escaith and effect: the percentage change in processing imports was Gonguet 2009; Escaith, Lindenberg, and Miroudot 2010; more pronounced than the percentage change in process- Ma and Van Assche 2010). The bullwhip effect is one of the ing exports in 15 out of 20 industries (table 21.3). most researched and documented symptoms in the field The existence of a bullwhip effect in China's processing of supply chain management (Lee, Padmanabhan, and trade helps at least partially to explain the resilience of its Wang 1997; Cachon, Randall, and Schmidt 2007). In the economy when faced with the great trade collapse (Ma and bullwhip effect, when a downstream firm is confronted Van Assche 2009). When the crisis hit China in the second with a drop in demand for its final products, its first reac- half of 2008, its economy was able to rapidly pass on the tion is to run down its inventories. Thus, a slump in negative export demand shock to its input suppliers 282 Managing Openness Table 21.3. Changes in China's Processing Trade by Industry, Quarter 1, 2008­Quarter 1, 2009 Processing Processing Growth in Growth of exports imports processing exports, processing imports, (US$, billions) (US$, billions) Q1 2008­Q1 2009 Q1 2008­Q1 2009 Q1 2008 Q1 2009 (%) Q1 2008 Q1 2009 (%) Bullwhip effect High technology 78.32 57.23 ­26.9 45.88 29.03 ­36.6 Yes Aircraft 0.21 0.23 8.6 0.11 0.09 ­17.8 No Pharmaceuticals 0.37 0.36 ­3.3 0.08 0.09 24.0 No Office and computing machinery 30.70 23.80 ­22.4 3.59 2.43 ­32.5 Yes Radio, TV and communications equipment 38.50 27.80 ­27.7 29.70 20.20 ­31.9 Yes Medical, precision, and optical instruments 8.54 5.05 ­40.9 12.40 6.22 ­49.8 Yes Medium-high technology 34.01 24.90 ­26.9 13.59 9.79 ­27.9 Yes Electrical machinery 10.50 7.43 ­29.4 6.27 4.34 ­30.7 Yes Motor vehicles 3.37 1.37 ­59.4 0.26 0.17 ­35.0 No Chemicals 2.47 2.01 ­18.4 2.10 1.41 ­32.9 Yes Other transport equipment 0.87 0.79 ­9.5 0.07 0.06 ­13.1 Yes Machinery and equipment 16.80 13.30 ­21.0 4.89 3.81 ­22.0 Yes Medium-low technology 12.67 12.04 ­5.1 10.12 5.69 ­43.8 Yes Shipbuilding and repairing 3.36 5.30 57.5 0.01 0.04 377.6 Yes Rubber and plastic products 4.57 3.59 ­21.5 1.59 1.09 ­31.5 Yes Petroleum products 0.05 0.04 ­14.7 0.04 0.18 395.9 No Nonmetallic mineral products 0.71 0.49 ­31.1 0.66 0.37 ­44.5 Yes Metal products 3.98 2.62 ­34.3 7.82 4.01 ­48.7 Yes Low technology 22.69 19.38 ­14.6 6.57 4.69 ­28.7 Yes Manufacturing 8.17 6.79 ­16.9 0.73 0.52 ­27.7 Yes Paper and paper products 1.38 0.98 ­29.0 1.28 0.72 ­44.0 Yes Printing and publishing 0.32 0.32 ­1.3 0.06 0.05 ­13.3 Yes Food, beverages, and tobacco 1.32 1.35 2.5 0.71 0.50 ­29.5 No Textiles, apparel, and leather 11.50 9.94 ­13.5 3.80 2.90 ­23.8 Yes Other 4.55 2.63 ­42.3 11.80 6.85 ­41.8 No Total 152.24 116.18 ­23.7 87.95 56.06 ­36.2 Yes Source: Authors' calculations based on China's Customs Statistics. Note: Q = quarter. through a reduction in demand for processing inputs. figure 21.9, the global recession has affected most severely Indeed, since the drop in imports was larger than the drop China's imports from countries that more intensively supply in exports in the first quarter of 2009 compared to a year processing inputs to China (that is, its East Asian neighbors). earlier, China's net exports actually increased. With the exception of Indonesia and Vietnam, more than Those worst hit by the bullwhip effect in China's pro- 40 percent of China's imports from its major East Asian cessing trade were its East Asian neighbors. As shown in trading partners in 2006 were processing imports, which is a Global Production Networks in the Postcrisis Era 283 Figure 21.9. China's Processing Imports, 2008, versus China's Import Growth by Economy of Origin, Quarter 1, 2008­Quarter 1, 2009 0 Australia ­10 EU-19 Vietnam United States ­20 Canada Singapore import growth (%) Malaysia ­30 Thailand Japan Korea, Rep. Indonesia ­40 Taiwan, China ­50 ­60 Philippines ­70 0 10 20 30 40 50 60 processing imports (as a % of China's total imports) Source: Authors' calculations based on China's Customs Statistics. significantly higher share than for countries outside of East gather strength. Full recovery of intra-GPN trade, however, Asia. During the recent global economic crisis, these East will depend on the ability of policy makers to keep protec- Asian countries witnessed a larger decline in imports than tionism (and the related trade cost effect) in check. other countries. Compared to the previous year, China's imports from its major East Asian trading partners all Conclusion declined between 25 percent and 61 percent in the first quarter of 2009. In contrast, China's imports from its major In this chapter, we have relied on data from China's pro- non-Asian trading partners fell less than 20 percent. cessing trade regime to analyze the impact of trade costs on In sum, we find supporting evidence that intra-GPN trade intra-GPN trade. We found that intra-GPN trade differs has been especially vulnerable to the global recession. It has from regular trade in that it depends not only on down- exacerbated the great trade collapse through three channels: a stream trade costs but also on upstream trade costs and the compositional effect, a trade cost effect, and a bullwhip effect. interaction of both. Moreover, we showed that intra-GPN First, in line with the compositional effect, we find that the trade is more sensitive to changes in trade costs than regu- sectors that contributed most to the collapse of Chinese lar trade. Specifically, intra-GPN trade is more sensitive to exports are those in which processing trade is more prevalent. oil price movements, trade policy changes, and business Second, we show that, within industries, processing exports cycle fluctuations than regular trade. These results suggest consistently dropped more than nonprocessing exports dur- that the resilience of vertical specialization and GPNs in ing the global recession. This finding is in line with the trade the postcrisis era will critically depend on our ability to cost effect since intra-GPN trade should be more sensitive to keep trade costs in check. trade costs than regular trade. Third, in line with the bullwhip effect, we show that in virtually all industries, the drop in demand for China's processing exports led to a bigger drop in Notes processing imports. 1. See the background paper by Ma and Van Assche (2010) for further details. The larger sensitivity of intra-GPN trade to business cycle 2. The source and destination countries of Chinese processing trade movements of course also suggests that intra-GPN trade is are adjusted for Hong Kong re-exports. See Ma and Van Assche (2010) for likely to rebound rapidly as the global economy continues to further details on how we address this in the empirical analysis. 284 Managing Openness 3. The EU-19 include all European Union countries prior to the Escaith, H., N. Lindenberg, and S. Miroudot. 2010. "International Supply accession of the 10 candidate countries on May 1, 2004, plus the four Chains and Trade Elasticity in Times of Global Crisis." Staff Working Eastern European member countries of the Organisation for Economic Paper ERSD-2010-08, World Trade Organization, Geneva. Co-operation and Development, namely, the Czech Republic, Hungary, Evans, C. and J. Harrigan. 2005. "Distance, Time, and Specialization: Lean Poland, and the Slovak Republic. retailing in General Equilibrium." American Economic Review 95 (1): 4. See Ma and Van Assche (2010) for full details. 292­313. 5. Peak oil refers to the attainment of the maximum conventional Francois, J., and J. Woerz. 2009. "The Big Drop: Trade and the Great Reces- oil output (that is, excluding heavy oil from tar sands, oil shale, and the sion." VoxEU.org. May 2. http://www.voxeu.org/index.php?q=node/ like), expressed as millions of barrels of crude oil extracted per day 3527. (Deffeyes 2001). Gaulier, G., F. Lemoine, and D. Ünal-Kesenci. 2007. "China's Integration 6. See Ma and Van Assche (2010) for further details on hypotheses in East Asia: Production Sharing, FDI and High-technology Trade." and construction of the model. Economic Change and Restructuring 40: 27­63. 7. For example, consumer electronics, automobile and transport Gereffi, G., J. Humphrey, and T. Sturgeon. 2005. "The Governance of the equipment, office equipment, and computers. Global Value Chains." Review of International Political Economy 12 (1): 8. Engel and Wang (2008) found that U.S. durable goods imports are 78­104. more sensitive to business cycles than nondurable goods imports. Grossman, G., E. Helpman, and A. Szeidl. 2006. "Optimal Integration 9. In line with this result, Aziz and Li (2008) demonstrate that China's Strategies for the Multinational Firm." Journal of International Eco- increasing specialization in electronics exports has led to an overall rise in nomics 70: 216­38. the income elasticity of China's exports. Haddad, M. 2007. "Trade Integration in East Asia: The Role of China and 10. We use the OECD's technology classification (Hatzichronoglou Production Networks." Policy Research Working Paper 4160, World 1997) to disaggregate China's exports into four categories: high-technology Bank, Washington, DC. exports, medium-high-technology exports, medium-low-technology exports, Hanson, G., R. Mataloni, and M. Slaughter. 2005. "Vertical Production and low-technology exports. Networks in Multinational Firms." Review of Economics and Statistics 11. Mason-Jones and Towill (2000) provide the following explana- 87 (4): 664­78. tion of a bullwhip effect: "If demand for products is transmitted along a Harrigan, J., and A. Venables. 2006. "Timeliness and Agglomeration." series of inventories using stock control ordering, then the demand varia- Journal of Urban Economics 59: 300­316. tion will increase with each transfer." Hatzichronoglou, T. 1997. "Revision of the High-Technology Sector and 12. In the first and second quarter of 2009, China's GDP expanded at Product Classification." Science, Technology and Industry Working an annualized rate of 6.1 percent and 7.9 percent, respectively. Papers 1997/2, Organisation for Economic Co-operation and Devel- opment, Paris. Helpman, E. 2006. "Trade, FDI, and the Organization of Firms." Journal of Bibliography Economic Literature 44 (3): 589­630. Hummels, D. 2007. "Transportation Costs and International Trade in the Aziz, J., and X. Li. 2008. "China's Changing Trade Elasticities." China and Second Era of Globalization." Journal of Economic Perspectives 21 (3): the World Economy 16 (3): 1­21. 131­54. Behar, A., and A. Venables. 2010. "Transport Costs and International Trade." ------. 2009. Globalization and Freight Transport Costs in Maritime Ship- In Handbook of Transport Economics, ed. A. de Palma, R. Lindsey, ping and Aviation. Paper 2009-3, International Transport Forum, Paris. E. Quinet, and R. Vickerman. Cheltenham, U.K.: Edward Elgar. International Economy. 2009. "Collapse in World Trade: A Symposium of Bems, R., R. Johnson, and K.-M. Yi. 2009. "The Role of Vertical Linkages Views." International Economy. Spring: 28­38. in the Propagation of the Global Downturn of 2008." Unpublished. Jacks, D., C. Meissner, and D. Novy. 2009. "The Role of Trade Costs in the http://www.imf.org/external/np/res/seminars/2010/paris/pdf/keymu Great Trade Collapse." VoxEU.org. November 27. http://www.voxeu .pdf. .org/index.php?q=node/4300. Cachon, G., T. Randall, and G. Schmidt. 2007. "In Search of the Bullwhip Koopman, R., Z. Wang, and S.-J. Wei. 2008. "How Much of Chinese Effect." Manufacturing and Service Operations Management 9 (4): Exports Is Really Made in China? Assessing Domestic Value-added 457­79. When Processing Trade Is Pervasive." Working Paper 14109, National De Almeida, P., and P. Silva. 2009. "The Peak Oil Production: Timings Bureau of Economic Research, Cambridge, MA. and Market Recognition." Energy Policy 37: 1276­76. Lee, H., S. Padmanabhan, and S. Wang. 1997. "Information Distortion in a Deffeyes, K. 2001. Hubbert's Peak: The Impending World Oil Shortage. Supply Chain: The Bullwhip Effect." Management Science 43 (4): Princeton, NJ: Princeton University Press. 546­58. Desai, M. 2009. "The Decentering of the Global Firm." World Economy Levchenko, A., L. Lewis, and L. Tesar. 2009. "The Collapse of U.S. Trade: In 32 (9): 1271­90. Search of the Smoking Gun." VoxEU.org. November 27. http://www Eaton, J., S. Kortum, B. Neiman, and J. Romalis. 2010. "Trade and the .voxeu.org/index.php?q=node/4280. Global Recession." Research Series 201010­196, National Bank of Ma, A., and A. Van Assche. 2009. "China's Trade in Crisis." AIB Insights Belgium, Brussels. 9 (2): 11­15. Egger, P. 2008. "On the Role of Distance for Bilateral Trade." World ------. 2010. "The Role of Trade Costs in Global Production Networks." Economy 31 (5): 653­62. Policy Research Working Paper 5490, World Bank, Washington, DC. Ekholm, K., R. Forslid, and J. Markusen. 2007. "Export-platform Foreign Ma, A., A. Van Assche, and C. Hong. 2009. "Global Production Networks Direct Investment." Journal of the European Economic Association 5 (4): and China's Processing Trade." Journal of Asian Economics 20 (6): 776­95. 640­54. Engel, C., and J. Wang. 2008. "International Trade in Durable Goods: Mason-Jones, R., and D. Towill. 2000. "Coping with Uncertainty: Reduc- Understanding Volatility, Cyclicality and Elasticities." Working Paper ing `Bullwhip' Behavior in Global Supply Chains." Supply Chain 13814, National Bureau of Economic Research, Cambridge, MA. Forum 1: 40­45. Escaith, H., and F. Gonguet. 2009. "International Trade and Real Transmis- Mirza, D., and H. Zitouna. 2009. "Oil Prices, Geography and Endogenous sion Channels of Financial Shocks in Globalized Production Net- Regionalism: Too Much Ado about (Almost) Nothing." Working works." Staff Working Paper ERSD-2009-06, World Trade Organiza- Paper 2009-26, Centre d'Etudes Prospectives et d'Informations Inter- tion, Geneva. nationales (CEPII), Paris. Global Production Networks in the Postcrisis Era 285 Naughton, B. 2006. "China's Emergence and Prospects as a Trading UNCTAD (United Nations Conference on Trade and Development). 2010. Nation." Brookings Papers on Economic Activity 2: 273­313. "Oil Prices and Maritime Freight Rates: An Empirical Investigation." OECD (Organisation for Economic Co-operation and Development). Geneva: United Nations Conference on Trade and Development. Forthcoming. "Determinants of Maritime Transport Costs." Mimeo. Van Assche, A. 2008. "Modularity and the Organization of International O'Rourke, K. 2009. "Collapsing Trade in a Barbie World." The Irish Econ- Production." Japan and the World Economy 20: 353­68. omy (blog). June 18. http://www.irisheconomy.ie/index.php/2009/06/ Yeaple, S. 2003. "The Complex Integration Strategies of Multinationals 18/collapsing-trade-in-a-barbie-world/. and Cross Country Dependencies in the Structure of Foreign Direct Rubin, J. 2009. Why Your World Is Going to Get a Whole Lot Smaller: Oil Investment." Journal of International Economics 60: 293­314. and the End of Globalization. New York: Random House. Yi, K.-M. 2003. "Can Vertical Specialization Explain the Growth of World Rubin, J., and B. Tal. 2008. "Will Soaring Transport Costs Reverse Global- Trade?" Journal of Political Economy 111: 52­102. ization?" CIBC World Markets. http://research.cibcwm.com/economic ------. 2009. "The Collapse of Global Trade: The Role of Vertical Special- _public/download/smay08.pdf. ization." In The Collapse of Global Trade, Murky Protectionism, and the Spencer, B. 2005. "International Outsourcing and Incomplete Contracts." Crisis: Recommendations for the G20, ed. R. Baldwin and S. Evenett. Canadian Journal of Economics 38: 1107­35. London: VoxEU. Sturgeon, T. 2002. "Modular Production Networks: A New American Yoshida, Y., and H. Ito. 2006. "How Do the Asian Economies Compete Model of Industrial Organization." Industrial and Corporate Change with Japan in the U.S. Market? Is China Exceptional? A triangular 11 (3): 451­96. Trade Approach." Asia Pacific Business Review 12 (3): 285­307. 22 The 2008­09 Recession: Implications for International Labor Migration Philip Martin Demographic and economic inequalities between countries, crisis was followed by a relatively quick resumption of combined with revolutions in communications, trans- economic and job growth, with more migrant workers portation, and rights, encourage and enable people to cross moving within Asia than before the crisis. national borders in search of economic opportunity. In What, then, are the implications of increased migra- recent decades, these factors have contributed to a dramatic tion from lower- to higher-wage countries for the global increase in international migration flows. Between 1985 economy? The Global Economic Prospects report (World and 2010, the global stock of migrants almost doubled, Bank 2005) estimated that increasing the number of from 111 million to 214 million. The increase was concen- developing-country migrant workers in industrial coun- trated in industrial countries, where the stock of migrants tries by 50 percent could increase global economic output rose from 50 million to 128 million. Thus, the share of the by US$356 billion or 0.6 percent of global gross domestic world's migrants in industrial countries rose from 48 per- product (GDP). This number is more than the estimated cent to 60 percent.1 gains from reducing trade restrictions in the Doha round The demand for migrant labor is expected to continue of negotiations. to expand in the coming decades. Most of this demand While increased migration may benefit the economy as for migrants will come from the industrial countries, but a whole, it is less clear to what extent foreign jobs and some of the sharpest increases in the demand are remittances can speed development in migrant countries expected to be in middle-income developing countries of origin. Remittances to developing countries quadrupled such as Brazil, Costa Rica, Libya, Malaysia, South Africa, in recent decades, from less than US$60 billion in 1990 to and Thailand. US$338 billion in 2008 (Ratha, Chami, and Fullenkamp While migration flows are expected to increase, the 2009). Both remittances and migration increased, raising 2008­09 global recession raises questions about future questions about the relationship between remittances and migrant patterns. Historical precedent suggests that the what might be called stay-at-home development, the type recent recession may alter international labor migration of development that ensures that the children of migrants patterns. The recession that followed oil price hikes in do not have to follow their parents abroad. 1973­74 stopped guest worker recruitment in Europe but The rest of this chapter discusses recent trends in inter- opened new opportunities for migrant workers in the national labor migration and the factors that will affect oil-exporting nations, especially in the member coun- migrant flows in the coming decades. The next section tries of the Gulf Cooperation Council. The 1981­82 reces- discusses recent trends in population growth and migration. sion accelerated labor-saving changes in industrial country The following section analyzes projections of future migrant manufacturing sectors, prompting European countries patterns and the factors that influence these flows. The sub- such as France and Germany to offer departure bonuses to sequent section addresses the economic implications of jobless migrants. Moreover, the 1997­98 Asian financial international migration. The final section concludes. 287 288 Managing Openness Recent Trends in Population and Migration less-developed countries. The share of global population of more-developed countries is expected to fall from 18 per- This section provides an overview of recent dynamics in cent in 2009 to 13 percent by 2050. While the populations world population and migration patterns. Against that of Europe and Africa are currently similar in size, there are background, it then discusses the factors that influence projected to be three Africans for every European in 2050. migration flows in the global economy. Recent Trends in International Migration Population Trends In 2010, the stock of international migrants was 214 mil- Between 1950 and 2000, the world's population rose by lion, meaning that 3 percent of the world's 6.9 billion almost 150 percent, from 2.5 billion to 6.1 billion (table people were living outside their country of birth.5 One 22.1).2 As of 2009, the world's population had reached 6.8 hundred twenty-eight million of these migrants were liv- billion, including 1.2 billion residents in more-developed ing in more-developed countries, making up 10 percent of countries and 5.6 billion (or 82 percent) in less-developed their 1.2 billion population. The other 86 million were living countries (UN 2009a). in less-developed countries, making up less than 2 percent However, the rate of population growth varied by of their 5.7 billion population. The ratio of the share of region and by country. Indeed, while Africa's share of the migrants to the share of population in industrial countries world's population rose from 9 percent to 13 percent, was 3.3, reflecting the fact that industrial countries had 60 Europe's share declined from 22 percent to 12 percent. percent of the world's migrants and less than 20 percent of Moreover, the populations of major migrant-sending the world's people (table 22.2). countries such as Mexico and the Philippines doubled or By region, immigrants were concentrated in Europe, tripled during this period. Asia, and North America in 2010. Europe had the most The global population is expected to continue growing migrants in absolute terms: Europe's 70 million migrants until 2050, but at a slower rate and with significant vari- constituted almost 10 percent of the continent's popula- ability. Fertility rates are projected to decline from an aver- tion.6 Asia followed with 61 million migrants, who made age 2.6 children per woman in 2005 to 2.05 by 2050.3 Based up 1.5 percent of the Asian population. Almost half of on these estimations, the world's population is expected to Asia's migrants were in western Asia, that is, the Persian reach 9.2 billion by 2050 (UN 2009a).4 If fertility does Gulf countries and the Middle East. Last, North America not decline to below the replacement level of 2.1 as pro- had an immigrant population of 50 million in 2010, jected, the global population could approach 11 billion including 43 million in the United States and 7.2 million in by 2050. Canada. One-seventh of North American residents were Moreover, significant variability is predicted in popula- foreign born. tion growth across regions and income groups. A quarter of In contrast, Africa and Latin America had relatively few the world's countries are expected to have shrinking popu- migrants in 2010 (about 19 million and 7 million, respec- lations over the next decade, while another quarter are tively). However, in both continents, a few countries had expected to grow at rates that will double their populations high shares of migrants in their population (for example, within a half-century (table 22.1). Almost all population Costa Rica, Côte d'Ivoire, and Libya each had more than growth is projected to occur in what are now considered 10 percent migrants). Oceania also had a relatively lower number of immigrants in absolute terms (6 million), but it Table 22.1. Percentage of World Population by Continent, was the region with the highest share of migrants: 17 per- 1800­2050 cent of the residents of Oceania were foreign born. 1800 1950 2000 2050a Six countries had 40 percent of the total migrant popu- Africa 11 9 13 20 lation: 43 million in the United States, 12 million in the Asia 65 56 61 59 Russian Federation, 11 million in Germany, and about Europe 21 22 12 7 7 million each in Canada, France, and Saudi Arabia. Latin America Countries with the highest share of migrants in their and the Caribbean 3 7 9 9 North America 1 7 5 4 populations were mostly Gulf oil exporters such as Qatar, Oceania 1 1 1 where over 85 percent of residents were migrants, and World (millions) 978 2,535 6,124 9,191 Kuwait and the United Arab Emirates, where 70 percent of Source: UN 1999, 2009a. residents were migrants. The countries with the lowest a. Projected. shares of migrants in their population include China, The 2008­09 Recession: Implications for International Labor Migration 289 Table 22.2. Population and Migration, 2010 Ratio of % Population Migrants Migrants as % % of total % of of migrants to % (millions) (millions) of population population migrants of total population World 6,909 214 3 100 100 1.0 More-developed countries 1,237 128 10 18 60 3.3 Less-developed countries 5,671 86 2 82 40 0.5 Africa 1,033 19 2 15 9 0.6 Asia 4,167 61 1 60 29 0.5 Europe 733 70 10 11 33 3.1 Latin America 589 7 1 9 3 0.4 North America 352 50 14 5 23 4.6 Oceania 36 6 17 1 3 5.4 Source: UN 2009b. Cuba, Indonesia, Peru, and Vietnam, where less than private recruiters. Recently, the role of direct employer 0.1 percent of residents were migrants. recruitment and public employment services has been Shifting patterns can also be seen in migrant destina- declining, as the role of networks and private agents has tions by income level. In particular, migration to middle- been increasing. income developing countries has been on the rise in recent The transportation revolution has facilitated migration years. The cases of Malaysia and Thailand illustrate several through the declining cost of travel. British migrants common features of migration trends in fast-growing unable to pay one-way passage to North American colonies developing countries. These two countries attract signifi- in the 18th century often indentured themselves, signing cant numbers of migrants from poorer neighbors and have contracts that obliged them to work for three to six years become increasingly dependent on migrant workers over for whoever met the ship and paid the captain. In contrast, time. While flows have increased because of the attraction transportation costs today are typically less than US$2,500 of higher-wage jobs, the governments of both countries to travel anywhere in the world legally and US$1,000 to acknowledge significant gaps between the goals and the US$20,000 for unauthorized migration. Most studies sug- outcomes of their migrant labor policies. Malaysia and gest faster payback times for migrants today, so that even Thailand have been taking an ad hoc approach to employer migrants who pay high recruitment or smuggling fees can requests for migrant workers and migrant worker policy, usually repay them within two or three years. wavering between periods of openness and periods of The rights revolution has also facilitated migration by restrictions. However, they are now making efforts to increasing the ability of migrants to stay abroad. After reduce irregular migration and employment in ways that World War II, most industrial countries strengthened the avoid the settlement of immigrants. Indeed, in contrast to constitutional and political rights of residents and granted many European nations, most Asian nations do not have social or economic rights to individuals in their evolving the goal of not becoming destinations for immigrants. welfare states without distinguishing between citizens and migrants. As migration increased in the 1990s, policy mak- ers began to roll back socioeconomic rights for migrants in What Drives Migration? an effort to manage migration.7 Adjusting rights is widely Two key factors have contributed to the rise in migrant acknowledged to be a blunt and often inefficient instru- flows in recent decades. First, the revolution in communi- ment for managing economically motivated migration. cations, transportation, and rights, has made international A second factor that can explain the rise in international migration easier. Cheaper communication has facilitated migration is the persistence of demographic and economic the growth of "migration infrastructure," a substantive net- inequalities. Demographic differences between areas have work of private and public intermediaries that helps historically driven large-scale migration in some parts of migrants cross national borders. Intermediaries include the world. For instance, in 1800, when 21 percent of the employers seeking migrants, public employment services world's population was in Europe and 4 percent was in the that match local workers with foreign jobs, social networks Americas, millions of Europeans emigrated to the Americas such as friends and family who are or were abroad, and in search of economic opportunity as well as religious and 290 Managing Openness political freedom. Will history repeat itself? Africa is pro- labor force will increase from about 3 billion in 2005 to jected to have three times more residents than Europe by 4.4 billion in 2050 (table 22.3). However, the rate of labor 2050. If Africa remains poorer than Europe, the diverging force growth is projected to be only 4 percent in the decade demographic trajectories of these two continents may pro- of 2040­50, a dramatic decrease relative to labor force pel increased migration flows. growth of over 20 percent in the 1980s and 1990s.10 Related to demographic differences, economic inequal- In 2005, about 20 percent of the world's workers were in ity has also been driving international migration. The 30 more-developed countries, and 80 percent were in less- high-income countries account for a seventh of the world's developed countries. If the economically active population people but over 70 percent of the world's economic output remains at 48 percent, the labor force in more-developed in 2008. Average gross national income per capita was countries is expected to stay constant, largely because of US$40,000 a year in high-income countries, 14 times the the stable population size of industrial countries. The labor US$2,800 average in the poorer 170 countries (World Bank force of more-developed countries is projected to stop 2009).8 Such income differences encourage (especially, growing after 2020, but the labor force of less-developed young) people to migrate over national borders. countries is expected to continue rising and to converge with the 48 percent of more-developed countries after 2020 (table 22.4). Future Supply of Migrants By continent, Africa is expected to have the fastest- This section looks at the factors influencing the future sup- growing labor force. The 357 million African workers in ply of migrants. After examining baseline projections of 2005 amounted to 15 percent of the total workforce in the labor force and migration flows, it turns attention to developing countries. By 2050, Africa's labor force is issues that can have a major impact on migration, such as expected to be 930 million people. Thus, 40 percent of the climate change. 1.4 billion expected growth in developing-country labor forces will be in Africa. As the most rural continent,11 Africa is likely to experience large-scale rural-urban migra- Labor Force Projections tion, some of which may spill over national borders. The world's economically active population (EAP)9 is In contrast to the projections for Africa, the labor force expected to gradually rise from 43 percent in 1980 and in Europe is expected to shrink. For example, Bijak et al. 47 percent in 2005 to 48 percent in 2020 (ILO 2008). And if (2007) have projected that the labor force of the EU-27 will the economically active share of the global population fall from 233 million in 2002 to 210 million in 2052. Three remains at 48 percent between 2020 and 2050, the world's policy levers could be used to offset projected declines in Table 22.3. The World's Economically Active Population, 1980­2050 millions 1980 1985 1990 1995 2000 Total population 4,451 4,855 5,295 5,719 6,124 World EAP 1,930 2,160 2,406 2,605 2,818 % EAP 43 44 45 46 46 2005 2010 2015 2020 2025 Total population 6,514 6,910 7,295 7,667 8,010 World EAP 3,050 3,279 3,481 3,651 3,845 % EAP 47 47 48 48 48 2030 2035 2040 2045 2050 Total population 8,317 8,587 8,824 9,025 9,191 World EAP 3,992 4,122 4,236 4,332 4,412 % EAP 48 48 48 48 48 1980­90 1990­2000 2000­10 2010­20 2020­30 2030­40 2040­50 % change in world EAP 25 21 17 17 9 6 4 Source: LABORSTA (database), International Labour Organization (ILO), http://laborsta.ilo.org. Population data from UN 2009a, medium variant. The 2008­09 Recession: Implications for International Labor Migration 291 Table 22.4. Economically Active Population in More- and Less-Developed Countries, 1980­2020 millions 1980 1985 1990 1995 2000 More developed Population 1,083 1,115 1,149 1,175 1,194 EAP 523 545 569 574 589 % EAP 48 49 50 49 49 Less developed Population 3,368 3,740 4,146 4,544 4,930 EAP 1,407 1,616 1,837 2,031 2,229 % EAP 42 43 44 45 45 2005 2010 2015 2020 2025 More developed Population 1,216 1,232 1,245 1,254 1,259 EAP 604 613 611 603 604 % EAP 50 50 49 48 48 Less developed Population 5,299 5,674 6,050 6,413 6,752 EAP 2,446 2,666 2,870 3,048 3,241 % EAP 46 47 47 48 48 2030 2035 2040 2045 2050 More developed Population 1,261 1,260 1,257 1,252 1,245 EAP 605 605 603 601 598 % EAP 48 48 48 48 48 Less developed Population 7,057 7,327 7,567 7,774 7,946 EAP 3,387 3,517 3,632 3,732 3,814 % EAP 48 48 48 48 48 % Change in EAP 1980­90 1990­2000 2000­10 2010­20 2020­30 2030­40 2040­50 More-developed EAP 9 5 4 5 0 0 ­1 Less-developed EAP 31 26 21 20 11 7 5 Source: LABORSTA (database), ILO, http://laborsta.ilo.org. Population data from UN 2009a, medium variant. the number of workers: raising fertility, encouraging more United States have sharply increased both guest worker residents to work longer, and admitting immigrants. Most and foreign student admissions. Governments navigating European studies put more emphasis on the fertility and between employers requesting foreign workers and restric- employment policy levers than migration. tionist public opinion are likely to continue to offer many of these so-called side door entrants a path to immigrant sta- tus or permanent residence if a local employer offers a job. Migrant Flow Projections Moreover, the decline in labor migration in the 2008­09 The Global Forum on Migration and Development and recession is likely a temporary hiccup in the upward trajec- most academic literature expect international labor migra- tory of international labor migration. While the recession tion levels to continue to rise as a result of demographic prompted a reduction of migrant worker inflows into and economic inequalities in a period of globalization countries that had been receiving large numbers,12 rela- (Pritchett 2006; Martin and Zuercher 2008; UNDP 2009). tively few migrant workers were pushed out of host coun- Indeed, migrant-receiving countries appear poised to accept tries overall. What is more, the recruitment of migrant more migrant workers over the next decade. Traditional workers resumed with economic growth in many countries immigration countries such as Australia, Canada, and the in 2010. 292 Managing Openness Factors Affecting Future Migrant Flows However, if governments provide the elderly with generous pensions and health care benefits that are financed from In addition to the driving effect of inequalities and the facil- taxes paid by current workers, making it unnecessary for itating effect of falling transportation and communication workers to save during their working lives for retirement, costs, three other factors are likely to affect migration flows economic growth may slow because of rising taxes on a in the coming decades. First, population aging is expected declining workforce. to increase demand for migrant workers. Second, climate Across industrial countries, the share of government change is predicted to drive a significant increase in rural- resources devoted to the elderly is about the same as their urban internal and international migration. And third, share of the population (Papadimitriou 2007). With little changes in international regulation are expected to affect prospect for an upsurge in fertility, many observers expect future migrant patterns. The rest of this section discusses governments in countries experiencing both population the implications of these factors for future migration. aging and decline to reduce benefits for the elderly and to raise income taxes to support retirees. Alternatively, the The Age of Aging United Nations (UN) estimated that raising the retirement Global population aging is unprecedented. While the age to 75 could maintain current support ratios16 in most world's population is growing by about 1.2 percent a year, industrial countries (UN 2000, 5). One justification for the population of persons 60 and older is growing at least more years of work is that older people are healthier than a twice as fast. Moreover, population aging appears to be an generation ago; a French woman of 77 in 2000, in health and enduring trend (UN 2002, 28).13 The median age of all fitness, had the health equivalent to a 63-year old in 1900 people on Earth, which was 24 in 1950 and 27 in 2000, is (Bourdelais 1999). projected to be 38 in 2050.14 About 20 percent of the popu- Migration is also touted as at least a partial "solution" to lation in developed countries is 60 and older, and that the related issues of population aging and financing benefits proportion is expected to rise to 33 percent by 2050. By of retirees. Working-age migrants can pay taxes that help contrast, the share of elderly among the residents of less- sustain pay-as-you-go social security systems. However, developed countries is expected to double from 10 percent unless immigration increases over time, migrants have a today to 20 percent by 2050. one-time effect on social security systems, since they also Many industrial countries will experience population age and become eligible for government-provided benefits. aging amid population decline. These countries, including Moreover, immigration rates would have to increase Italy, Japan, and the Russian Federation, have not yet devel- quite substantially to compensate for population aging and oped migration responses that could include, for instance, decline. For instance, the UN estimated that for France, admitting migrants to maintain their populations or the Germany, Italy, and the United Kingdom to maintain the ratios between dependent and working-age populations. 1995 ratio of workers to retirees,17 immigration would have While the aging trend poses challenges for growth of the to increase 37-fold over 1995 levels, to almost 9 million a global economy, economists highlight two "demographic year (table 22.5). By contrast, the United States would have dividends" that can speed economic and job growth to a to increase immigration to almost 12 million to maintain certain extent. First, declining fertility will progressively the 1995 ratio of workers to retirees (UN 2000). As opinion reduce the number of youth to educate and thus could lead polls suggest that neither Europeans nor Americans want to to an increase in the number of women who elect to work increase migrant flows, the feasibility of this solution to for wages (Bloom, Canning, and Sevilla 2003). Moreover, population aging is limited, at least in the short term.18 as families tend to invest more in each of their fewer children, However, aging is likely to prompt some migration of the growing availability of well-educated young workers the elderly to lower-cost areas and to increase migration could fuel economic and job growth, as with the East Asian of caregivers from lower- to higher-wage countries. The "economic miracle" economies (World Bank 1993).15 exact mix of responses is likely to vary across countries Second, as longevity prompts workers to increase their and depend on individual and policy decisions. Because savings, the cost of capital will decline, thus promoting aging is projected to increase the fastest in less-developed investment and raising productivity (Lee and Mason countries, the demand for caregivers for the elderly is 2006). The life-cycle theory of consumption and sav- likely to rise fastest there. It could also encourage some ings assumes that households accumulate wealth during elderly to move to places with warmer weather and lower working years to maintain consumption in retirement. This living costs for retirement, an idea that has prompted pattern can result in a virtuous circle of low interest rates, proposals for "retirement villages" in countries with low- more investment, and rising productivity and earnings. cost care workers such as the Philippines.19 Moreover, The 2008­09 Recession: Implications for International Labor Migration 293 Table 22.5. Replacement Migration in Europe and the United States, 1995 thousands Average annual number of migrants required, 2000­50 To maintain Actual To maintain Multiple To maintain Multiple population Multiple immigration 1995 of 1995 1995 working- of 1995 support of 1995 in 1995 population immigration age population Immigration ratioa immigration EU-15b 270,000 949,000 4 1,588,000 6 13,480,000 50 EU-4c 237,000 677,000 3 1,093,000 5 8,884,000 37 France 7,000 29,000 4 109,000 16 1,792,000 256 Germany 204,000 344,000 2 487,000 2 3,630,000 18 Italy 6,000 251,000 42 372,000 62 2,268,000 378 United Kingdom 20,000 53,000 3 125,000 6 1,194,000 60 Other EU countries 33,000 272,000 8 495,000 15 4,596,000 139 United States 760,000 128,000 0 359,000 0 11,851,000 16 Source: UN 2000. a. Migrants necessary to maintain 1995 population ratio of persons ages 15­64 to those age 65 or older. b. EU-15: Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal, Spain, Sweden, and the United Kingdom. c. EU-4: France, Germany, Italy, and the United Kingdom. aging is likely to encourage the development of technologies farming opportunities in areas now too cold for crop pro- that enable the elderly to care for themselves where they duction. Agricultural productivity is expected to decrease live. Japanese firms, for example, have pioneered a vari- in Sub-Saharan Africa, South Asia, and parts of South ety of technologies aimed at helping the elderly in their America and may perhaps increase in colder areas such as homes, such as robots that assist with daily chores and Canada and the Russian Federation (Darwin et al. 1995; summoning help.20 World Bank 2008, 16­17). However, new farming opera- tions in sparsely populated regions that are now too cold Climate Change and Migration for farming are not likely to be labor intensive. Thus, the A second factor that may affect international migration in net effect of these changes is likely to be more migration the coming decades is climate change.21 There are three out of agriculture. major channels through which climate change could trig- Observers generally agree that climate change will affect ger new migrant patterns. First, climate change is expected developing countries more adversely than industrial coun- to generate more severe storms, such as hurricanes, that tries because less-developed countries have fewer resources destroy housing and erode land, thus encouraging migra- to invest in prevention and adaptation. And within devel- tion, at least until recovery. Second, competition for scarce oping countries, the poorest residents are likely to suffer land and water could increase, especially in arid areas with most. Because three-fourths of the poor in developing rapidly growing populations, as in Sub-Saharan Africa. countries live in rural areas, climate changes that adversely Competition for land and water can lead to conflict and affect agriculture are expected to increase out-migration migration, as when herders come into conflict with crop from rural areas (World Bank 2008, 1). farmers. Third, rising and more variable temperatures are The question is, What is the optimal policy response? likely to make agriculture less productive in densely popu- Current government policies include helping rural-urban lated areas in developing countries. Since 40 percent of the migrants integrate in urban areas by removing barriers to world's workers obtain their income from farming,22 cli- government services, as in China23 and India, and creating mate change is expected to increase rural-urban internal nonfarm jobs in urban centers for migrants. Other policies and international migration. This increased migration will to ease the transition of workers from rural to urban labor likely follow well-established migration networks, making markets include extending labor exchanges to rural areas to it hard to isolate the extra migration due to climate change expedite worker-employer job matching, providing accom- (World Bank 2009, 110). modation for the often single youth who migrate, and Of course, the impacts of climate change are not uni- ensuring that migrants who settle in urban areas have form across regions, as global warming may open up access to government services. 294 Managing Openness Alternatively, the World Bank (2008) made three policy The most likely scenario lies between these extremes, recommendations to mitigate the pressure for developing- meaning that a variety of national and regional migration country farmers to migrate: systems will probably prevail. Rentier societies such as Gulf oil exporters seem poised to continue relying on migrant · First, additional investments should be made in agricul- workers in private labor markets. And more industrial tural research and development, including research on countries are joining the four traditional immigration how to produce food and fiber despite climate change. countries--Australia, Canada, New Zealand, and the · Second, governments should implement policies to United States--to select a certain number of immigrants help small farmers and landless workers increase their each year and plan for their integration. incomes through reform of land, commodity, and water The great unknown is policy development in middle- markets and create opportunities for rural people to income migrant-receiving countries such as Brazil, supplement their incomes in rural nonfarm industries. Malaysia, Mexico, Morocco, South Africa, and Thailand. · Third, governments should change policies so that farm- Most migration literature predicts and endorses regional ers and other rural residents have incentives to man- and global liberalization of labor migration (Ghosh 2000; age limited supplies of land, water, and other natural Bhagwati 2003). But will migrant-receiving countries, resources in sustainable ways. some of which also send workers abroad, follow the lead of However, assessments of these policy options should industrial countries and plan for immigration based on also consider that nonfarm industries often need subsidies family, economic, or other criteria? Will they favor the to induce them to locate in rural areas that may have higher return of descendants of previous emigrants, or will they transportation costs and lack workers with needed skills. follow the model of Singapore, welcoming professionals to Mass migrations due to climate changes that reduce the settle with their families while aiming to rotate less skilled viability of farming will be much easier to identify but per- workers in and out of the labor force and country? haps harder to deal with effectively. For example, storms The answer to many of these questions depends on the that destroy shrimp farms, cover land with salt water, or nature of migration management at the global level. Cur- erode top soil can displace large numbers of people quickly. rently, the global migration policy regime is primarily nor- The initial policy response is likely to be temporary sup- mative, with International Labour Organization (ILO) and port until the agricultural system can be repaired and UN conventions to ensure that migrant workers have the those who were displaced can resume farming. Similarly, same wages and rights as national workers. International the response to flooding that covers low-lying land could migration enables many people to earn higher wages and be temporary shelter for the displaced followed by a return improve their standard of living, but it can also be associ- to farming. What is not clear is when or whether a mass ated with violations of fundamental rights and freedoms, movement as a result of a climate-related event will or unequal treatment and abuse, and sometimes trafficking should lead to permanent resettlement. and forced labor. Thus, the ILO's mission is to provide a framework to maximize the positive outcomes of migra- International Management of Labor Migration tion and minimize negative outcomes. In particular, the Global politics will also affect future trends in interna- ILO's Multilateral Framework on Labor Migration aims to tional migration. At the extremes, two scenarios for the protect migrant worker rights, but it does not spell out how management of international labor migration are possi- many migrants should be admitted. ble. In the first scenario, additional bilateral and regional This normative or rights-based policy regime may be free-movement regimes could emerge, such as that between altered by Mode 4 of the General Agreement on Trade in Australia and New Zealand or among member states of Services (GATS) negotiations aimed at facilitating the the European Union. Eventually, this development could movement of service providers over national borders.25 lead to the creation of a global World Migration Organi- While GATS is directly concerned with trade in services, zation analogous to the World Trade Organization (WTO) not labor migration, GATS does apply to foreigners who to foster the migration of workers over national borders provide services as self-employed independent contractors to, among other things, increase global economic effi- or as foreign employees of foreign firms. Foreign employee ciency.24 The second extreme scenario is that current levels service providers covered by GATS are considered migrant of migration will persist as national governments limit workers under ILO Conventions. migrant admissions for reasons ranging from fears of Liberalizing trade in services to promote economic crime and terrorism to the adverse effects of migrants on growth is one of the main goals of the current Doha develop- native workers and productivity. ment round of WTO negotiations.26 The value of services The 2008­09 Recession: Implications for International Labor Migration 295 is about 75 percent of the economy in high-income coun- labor-demand curve, increasing employment and lowering tries and half the economy in low-income countries. Trade wages (Borjas 1994). The major economic beneficiaries of liberalization occurs through requests and offers, and most immigration are the immigrants who earn higher wages in developing countries want to liberalize Mode 4 movements the destination area and employers who pay lower wages of natural persons, reflecting their comparative advantage because migrants are available; the major losers are workers in lower wages (Mattoo and Olarreaga 2004). India has led employed before the migrants arrived, whose wages are other developing countries in the request for four major lowered by immigration.29 changes that would make it easier for service providers to However, immigration also expands economic output. cross borders:27 eliminating the economic needs tests, The presence of more workers can attract capital, which, in expediting the issuance of visas and work permits, facilitat- turn, expands employment. For instance, the availability of ing the recognition of credentials and the obtaining of migrant farm workers can preserve jobs for local workers needed licenses, and exempting foreign service providers in packing and transportation industries. In this way, the from participating in work-related benefit programs and unemployment rate can fall or remain unchanged in an the payroll taxes that finance them. Ideally, liberalizers area despite in-migration. would like a "GATS visa" that would be uniform across WTO member countries (Chanda 2001, 648) and allow The Role of Temporary Worker Programs multiple visits within a given period of time, say, one to In recent years, growth in the number of temporary three years.28 migrant workers has been faster than growth in the number However, developed countries are more concerned with of permanent or long-term immigrants. This trend holds liberalization of Mode 3 (commercial presence) and have true both in countries that are newcomers to immigration, given few signs of intentions to liberalize Mode 4. Except such as Ireland, Spain, and the United Kingdom, and for some liberalization of intracompany transfers, few traditional immigration countries such as Australia, migrant-receiving countries have expanded existing chan- Canada, and the United States. Rapid growth in temporary nels or opened new ones for low-skilled migrant service foreign workers raises several questions, including how to providers. If developed countries make more commit- determine if foreign-born workers are "needed" and their ments under GATS to facilitate the entry of foreigners to effects on local workers in the short and long term. "provide services," developing-country governments could Almost all countries that are richer than their neighbors file trade complaints against the immigration and labor have some type of temporary worker programs (TWPs), departments of WTO member countries that slow or block often called guest worker programs, to allow local employ- the entry and work of their service providers. The Doha ers to employ foreign workers. These programs aim to add round is thus unlikely to lead to significantly more move- foreign workers to the labor force but not necessarily settled ments of service providers, which has disappointed many residents to the population. TWPs can often be a solution developing-country negotiators. for governments caught between competing interests, including employers requesting migrants, migrants eager for higher wages, and public constituencies that want Economic Implications of Migration migration reduced. However, the administrative elements This section examines the economic impacts of migration. of TWPs are often debated in migrant-receiving countries. It considers the perspectives of both receiving and sending For instance, What criteria must employers satisfy to countries. receive permission to recruit and employ migrants? How is migrant employment monitored? How do governments ensure that migrant workers depart as scheduled? These Economic Impact on Migrant-Receiving Countries issues may obscure an even more important issue--how The World Bank estimated that adding 14 million migrant are employer preferences, local workers' attitudes, and the workers from developing countries to the 28 million operation of labor markets affected by the presence of already in industrial countries would generate global migrants over time? benefits of US$355 billion. This number is more than the An evaluation of the economic impact of TWPs should estimated gains from reducing trade restrictions in the also consider the potential for path dependence to emerge Doha round of negotiations (World Bank 2006). from short-term hiring of foreign workers. Employers hire The standard analysis of the economic impact of immi- migrants because they cannot find workers willing to fill gration assumes that the entry of foreign workers shifts the jobs for the wages and benefits they offer. However, eco- aggregate labor supply curve to the right along a fixed nomic distortion or path dependence can increase the 296 Managing Openness reliance of employers on migrant workers over time as they children of migrants do not have to follow their parents make investments based on the assumption that doors to abroad. migrants will remain open. The lower labor costs made However, international labor migration does have the possible by the availability of migrants are sometimes capi- potential to speed development in migrant-sending coun- talized into asset values, giving asset owners an incentive to tries. Governments that have most successfully used migra- keep hiring migrants. tion to foster development often have more than just workers A similar path dependence can develop among migrants, abroad. For instance, the government of the Republic of as families, regions, and countries become dependent on Korea supported Korean construction firms that employed earnings and remittances from a foreign labor market. If Korean migrant workers in Gulf countries in the 1980s. the money remitted by migrants is not invested to create China has had a similar strategy of providing both engineer- jobs and generate stay-at-home development, incentives to ing services and migrant workers for foreign infrastructure migrate for employment may persist. In both Europe and projects. In contrast, countries that send mostly low-skilled the United States, decades of guest worker recruitment workers abroad, from Bangladesh to Mexico to the were followed, with lags, by asylum seeking from Turkey Philippines, generate remittances that reduce poverty but and unauthorized migration from Mexico (Martin 2009, do not necessarily set stay-at-home development in motion. chap. 2; Martin 2004). Given these opportunities and challenges, migrant- Path dependence helps explain why temporary worker sending countries must decide whether to encourage, programs often get larger and last longer than expected. tolerate, or try to discourage the out-migration of their Many migrant worker programs were begun in wartime or workers. What interventions are appropriate for migrant- at other "extraordinary" times, allowing employers to per- sending governments? That is, should they aim to regulate suade governments to admit migrant workers to cope with private sector recruiters or develop government alterna- "emergency" labor shortages. The anticipation that TWPs tives to them? If they choose the regulation approach, what are short-lived minimizes serious discussion of alternatives is the optimal amount of regulation and predeparture to migrants and the trade-offs between migrant numbers orientation and training? and migrant rights (Ruhs and Martin 2008).30 Conclusions and Recommendations Economic Impact on Migrant-Sending Countries In spring 2010, unemployment rates remained high and Migrant-sending nations also face difficult policy deci- international labor migration flows were smaller than in sions. Recognizing the lack of decent work at home, many 2006­07, but developing and industrial countries alike migrant-sending nations look to remittances to alleviate appeared to be recovering from the worst recession in a poverty and provide the investment needed to speed up half-century. Structural demographic and economic fac- development. In many developing countries, migrants gen- tors suggest that international labor migration will increase. erate more foreign exchange than foreign direct investment The major destinations are likely to be the industrial does. Moreover, migrants often send more money home countries that already have 60 percent of the world's after disasters and during recessions. In the wake of the migrants. Industrial countries with aging and shrinking recent financial crisis, remittances fell less than deploy- populations, employers complaining of labor shortages, ments of new migrant workers, reflecting the tendency of and large numbers of foreign students and others who remittances to act as a countercyclical stabilizer as migrants can work at least part time are likely to expand current resist returning and some remit more to family and friends programs and experiment with new programs that admit in need.31 In some countries, remittances rose even though migrant workers and allow some foreigners within their migrant deployments fell. For example, the number of borders to become guest workers or immigrants. Migration Bangladeshis deployed to foreign jobs fell from 875,000 in to middle-income developing nations is also likely to 2008 to 475,000 in 2009, but remittances rose 20 percent to increase, because the same demographic and economic US$10.7 billion in 2009. inequalities that encourage migration to industrial coun- Most studies show that remittances reduce poverty tries also move migrants from poorer to richer developing and increase spending on health and education in the countries. families that receive them, but the effects of migration The migration and development agenda is lengthy. As and remittances on development are more ambiguous migrants tend to be concentrated in particular occupations (Adams 2004). Indeed, in many cases, remittances do not and industries, migration policy has significant implica- foster the type of development that ensures that the tions for labor demand, supply, and labor market operation The 2008­09 Recession: Implications for International Labor Migration 297 in particular sectors of migrant-receiving countries. Migra- more migration are based in part on assumptions that may tion policy may become one of the most important factors not square with the politicians' experience, such as the shaping the structure and functioning of labor markets assumption that migrants get jobs without displacing (Martin, Abella, and Kuptsch 2006). local workers or depressing wages. What would happen to Richer countries need to deal with issues ranging from the estimated economic benefits of migrant workers if employers' requests for foreign workers to the options for researchers allowed for some displacement and depression? sustaining the financing of pay-as-you-go social security Second, we need to examine the longer-term impacts of systems in countries with shrinking and aging populations. more migration. Trade theory recognizes the possibility of Migrant workers are likely to be a key ingredient of the different short- and long-term effects of freer trade, but policies adopted to cope with demographic and economic migration theory, with some exceptions, rarely draws dis- change, but the fact that changes to immigration, social tinctions between short- and long-run effects. For exam- security, and health care policies are controversial makes it ple, what effects does easy access to migrants have on hard to debate policy options openly in some countries. productivity trends, an issue recently being debated in Instead, governments often tolerate the irregular migration Singapore?32 If easy access to migrants leads to "underinvest- that fills jobs and provides some short-term help for social ment" in capital, what is the best corrective tool: quotas on security systems. migrants, taxes or levies on migrant employment that can be Middle-income developing countries similarly avoid used to invest in labor-saving changes, or other tools? explicit discussions of the hard choices posed by migrant And finally, further study of the link between migration, workers. Because most are avowedly not countries of immi- remittances, and development is needed. Many govern- gration, governments periodically announce their intentions ments have embraced more labor migration in an effort to to "eventually" reduce or eliminate migrant workers. How- maximize remittances and accelerate development. As more ever, many governments are trying to achieve migrant governments adopt policies to promote foreign employ- worker reductions with rules that contradict the incentives ment, a careful evaluation of what types of migration lead of employers and migrants who want to prolong their to the most rapid development is needed. In particular, are employment relationship, resulting in irregular migration remittances a useful shorthand measure of the benefits of and thus uncertainties for both employers and migrants. migration for development? Remittances clearly reduce Some countries also actively seek to send more workers poverty for families receiving them, and their spending has abroad to increase remittances, speed up development, and spillover effects that benefit nonmigrants, but it is less clear relieve some of the pressures resulting from displacements that remittances set stay-at-home development in motion related to climate change. These countries have been (Ratha, Mohapatra, and Silwal 2009). among the most active in international forums, seeking to use the GATS Mode 4 negotiations, the Global Forum on Notes Migration and Development, and climate-change forums 1. The stock of migrants in developing countries rose from 57 million to argue that richer countries should open their doors to 86 million or 50 percent. The share of migrants in developing countries wider to workers from poorer countries. Thus far, however, fell from 52 percent to 40 percent between 1985 and 2010. 2. This increase was driven by baby booms in both industrial and these governmental efforts have not been successful. developing countries (Cohen 2003). Achieving international consensus to increase migrant 3. Fertility in the more-developed regions in 2005 was 1.6, versus 4.6 numbers appears to be a very distant goal. The major goal in less-developed regions. Fertility is expected to rise to 1.8 by 2050 in more-developed regions. of researchers and development institutions has been to 4. Keilman (1998, 15) finds that United Nations projections of popu- highlight the economic benefits of more labor migration, lation are most reliable for 10­15 years into the future. emphasizing the cost of restrictions on the entry of more 5. See International Migration (database), United Nations, http://www migrants. However, these economic arguments have so .unmigration.org. 6. Europe is defined as 48 countries, including the Russian Federation. far not been persuasive in international forums. Many 7. For instance, European nations preserved the right to asylum but migrant-receiving countries feel free to raise barriers to made it more difficult for asylum seekers to apply by designating some migrants, especially in recession. countries as generally safe, so that asylum seekers could be returned while their applications were pending. In the United States, Congress elected in The experience with migrant workers during the 1996 to maintain the number of immigrants admitted but to reduce their 2008­09 recession prompts three major recommendations. access to welfare benefits. Some of these benefit reductions were reversed First, the economic benefits of more migration need to be in the late 1990s (Migration News 1996). 8. The amount was not adjusted for purchasing power parity. reevaluated. Politicians may be reluctant to embrace more 9. The International Labour Organization defines the share of per- migration for many reasons, including the fact that many sons who are in the "economically active population" as those who are of the studies estimating significant economic benefits to employed or looking for work. 298 Managing Openness 10. Growth in the 1980s and 1990s can be explained primarily by the abroad; Mode 3 is commercial presence; and Mode 4 is temporary baby boom of the 1960s and 1970s. movement of natural persons. Thus, while "temporary" is not defined in 11. In 2008, 38 percent of Africa's population was urban, against the GATS, and GATS explicitly does not apply to permanent migration, a global average of 49 percent. See Population Reference Bureau, http:// Mode 4 refers to migration in broad terms. www.prb.org. 26. GATS does not include all services: it excludes most air transport 12. Those countries include, for instance, Ireland, Spain, the United services as well as "services supplied in the exercise of governmental Kingdom, and the United States. There were also a few well-publicized authority." cases: (a) of migrants being blocked from entry, for example, 55,000 27. Chanda (2004, 634) calls these four categories restrictions on Bangladeshi migrants were prevented from entering Malaysia early in entry and stay, recognition of credentials, differential treatment, and regu- 2009; (b) some Filipinos lost jobs and were forced to depart before the end lations on commercial presence, a taxonomy that groups economic needs of their contracts when the Taiwanese factories that employed them test and visa-work permit issuance. closed; and (c) some migrant construction and service workers in Gulf 28. The Coalition of Service Industries says that a "GATS visa" allow- Cooperation Council countries lost jobs and returned to their countries ing multiple short-term visits would be limited to professionals and of origin. highly skilled individuals and proposed a model of how countries could 13. In 2007, there were more people 65 and older than under age five, implement a GATS visa regime. GATS visas would be given to employees and this is unlikely to change in the future. of established foreign firms, which would post bonds on each GATS visa 14. For more-developed countries, median ages were 29 in 1950, 37 in holder that would be forfeited if the visa holder did not obey the terms of 2000, and 46 projected for 2050. the visa. 15. Real per capita GDP in East Asia rose almost 5 percent a year 29. In a simple two-factor production function with homogeneous between 1960 and 2005. Credit is attributed to sound macroeconomic labor and constant returns to scale, adding immigrants to the workforce policies, infrastructure investment, and policies that favored exporters. reduces wages in the short run (assuming full employment) but does not The ensuing discussion focused on whether productivity growth or large change the long-run return to capital and labor. If immigrant and native increases in physical capital and labor were largely responsible for the labor is not homogeneous, adding immigrants to the workforce has distri- rapid economic growth--that is, was the miracle due to innovation or butional consequences, helping complementary workers and hurting perspiration. workers who are substitutes for the immigrants. 16. The definition is the ratio of retired persons to those of work- 30. Migrant rights can cost employers money. With a negatively ing age. sloped demand for labor, ensuring full or equal rights can reduce the 17. Measurement is at prevailing fertility rates in 1995. number of temporary foreign workers employed. 18. See Pew Research Center, "World Publics Welcome Global Trade-- 31. The World Bank in November 2009 projected that remittances but Not Immigration," http://pewglobal.org/reports/display.php?ReportID= to developing countries would fall from US$338 billion in 2008 to 258; Polling Report, http://www.pollingreport.com/immigration.htm. US$317 billion in 2009, a drop of 6 percent. World Bank, "Workers' 19. One proposal to attract elderly Japanese to the Philippines is at Remittances Fall Less Than Expected, But 2010 Recovery Likely To Japanese External Trade Organization, http://www3.jetro.go.jp/ttppoas/ Be Shallow," http://web.worldbank.org/WBSITE/EXTERNAL/NEWS/ anken/0001051000/1051449_e.html. 0,,contenMDK:22394180~pagePK:64257043~piPK:437376~theSitePK: 20. Economist (2005) says that "although they are at ease with robots, 4607,00.html. many Japanese are not as comfortable around other people." 32. Between 2005 and 2009, Singapore's population rose by an 21. Climate change is defined as a significant change in temperature average 150,000 a year, mostly due to immigration. In 2010, a third of or precipitation that persists for several decades. Climate change can Singapore's 5 million residents were foreigners; in 1990, a seventh of occur because of natural factors such as changes in the sun's intensity, nat- 3 million residents were foreigners. Singapore has no unemployment ural processes such as changes in ocean currents, and human activities insurance and a very thin social safety net. The open-door policy for that change the atmosphere, such as burning fossil fuels or deforestation. low-skilled guest workers may have slowed productivity growth, which 22. Climate change is not the only factor affecting agriculture. Other averaged only 1 percent a year between 1988 and 2008. The government- trends helping to transform agriculture include (a) a rapidly rising demand appointed Economic Strategies Committee (http://www.esc.gov.sg) in for meat in middle-income developing countries, which can accelerate 2010 recommended an increase in the monthly employer levy to encour- deforestation and speed up global warming; (b) increased demand for bio- age employers to make the investments needed to raise productivity fuels that can push up food prices as crops are shifted from food to fuel pro- growth and train local workers. duction; and (c) a rising demand for seafood that encourages production in coastal areas in ways that can increase storm-related damage. 23. China is debating how to modify and perhaps eliminate the hukou Bibliography or household registration system that limits access to public housing, edu- cation, medical, and other benefits to the place where a person is registered. Adams, R. 2004. "Remittances and Poverty in Guatemala." Policy Research The Organisation for Economic Co-ordination and Development's Working Paper 3418, World Bank, Washington, DC. Economic Survey of China, released in February 2010, recommended that Bhagwati, J. 2003. "Borders beyond Control." Foreign Affairs (January­ the hukou system be phased out and that the Chinese government develop February): 98­104. national pension and health insurance systems to promote internal Bijak, J., D. Kupiszewska, M. Kupiszewski, K. Saczuk, and A. Kicinger. mobility. About 45 percent of China's 1.3 billion residents live in urban 2007. "Population and Labor Force Projections for 27 European areas, but 104 million of the 612 million urban residents do not have urban Countries, 2002­2052: Impact of International Migration on Popula- hukous. In 2006, the government began allowing migrant children without tion Ageing." European Journal of Population 23: 1­31. local hukous to attend local public schools, and city governments started Bloom, D., D. Canning, and J. Sevilla. 2003. A New Perspective on taking control of private schools for migrant children. the Economic Consequences of Population Change. Santa Monica, 24. A World Trade Organization or World Migration Organization CA: Rand. could also give national governments facing antimigrant publics "cover" Borjas, G. 1994. "The Economics of Immigration." Journal of Economic to liberalize labor flows over national borders, as when they say they do Literature 32 (4): 1667­717. not want to allow the entry of migrants but their international obligations Bourdelais, P. 1999. "Demographic Aging: A Notion to Revisit." History require them to do so. of the Family 4 (1): 31­50. 25. There are four major modes or ways to provide services over Chanda, R. 2001. "Movement of Natural Persons and the GATS." World national borders: Mode 1 is cross-border supply; Mode 2 is consumption Economy 24 (5): 631­54. The 2008­09 Recession: Implications for International Labor Migration 299 ------. 2004. "Movement and Presence of Natural Persons and Develop- Migration News. 1996. "Welfare Overhaul and Minimum Wage Changes." ing Countries: Issues and Proposals for the GATS Negotiations." 3 (9). http://migration.ucdavis.edu/mn/more.php?id=1022_0_2_0. Working Paper 19, South Centre, Geneva. Papadimitriou, D., ed. 2007. Government Spending on the Elderly. Cohen, J. 2003. "Human Population: The Next Half Century." Science 302 New York: Palgrave Macmillan. (5648): 1172­75. Pritchett, L. 2006. Let Their People Come: Breaking the Gridlock on Global Darwin, R., M. Tsigas, J. Lewandrowski, and A. Raneses. 1995. "World Labor Mobility. Washington, DC: Center for Global Development. Agriculture and Climate Change: Economic Adaptations." Agricul- Ratha, D., R. Chami, and C. Fullenkamp. 2009. "Remittances in Develop- tural Economic Report 703, Natural Resources and Environment ment." Finance and Development 46 (4): 30. Division, Economic Research Service, U.S. Department of Agricul- Ratha, D., S. Mohapatra, and A. Silwal. 2009. "Migration and ture, Washington, DC. Remittance Trends 2009." Migration and Development Brief 11, Economist. 2005. "Better than People," December 20. http://www World Bank, November 3. http://www.worldbank.org/prospects/ .economist.com/world/asia/displaystory.cfm?story_id=5323427. migrationandremittances. Ghosh, B., ed. 2000. Managing Migration: Time for a New International Ruhs, M., and P. Martin. 2008. "Numbers vs. Rights: Trade-Offs and Regime? New York: Oxford University Press. Guest Worker Programs." International Migration Review 42 (1): ILO (International Labour Organization). 2008. "Estimates and Projections 249­65. of the Economically Active Population, 1980­2050." August. http:// UN (United Nations). 1999. The World at Six Billion. New York: UN. laborsta.ilo.org/applv8/data/EAPEP/EAPEP_methodology.pdf ------. 2000. Replacement Migration: Is It a Solution to Declining and Keilman, N. 1998. "How Accurate Are the United Nations' World Popula- Aging Populations? New York: UN. tion Projections?" Frontiers of Population Forecasting: Supplement to ------. 2002. World Population Aging: 1950­2050. New York: UN. Data Population and Development Review 24: 15­41. available online at http://esa.un.org/unpp/index.asp?panel=1. Lee, R., and A. Mason. 2006. "What Is the Demographic Dividend?" ------. 2009a. World Population Prospects: The 2008 Revision. New York: Finance and Development 43 (3): 16­17. UN. http://esa.un.org/unpp/. Martin, P. 2004. "Germany: Managing Migration in the 21st Century." ------. 2009b. "International Migration 2009 Wallchart." http://www.un In Controlling Immigration: A Global Perspective, ed. W. A. Cornelius, .org/esa/population/. T. Tsuda, P. L. Martin, and J. F. Hollifield, 221­52. Palo Alto, CA: Stan- UNDP (United Nations Development Programme). 2009. Overcoming ford University Press. Barriers: Human Mobility and Development. New York: UNDP. Martin, P. 2009. Importing Poverty? Immigration and the Changing Face of World Bank. 1993. The East Asian Miracle: Economic Growth and Public Rural America. New Haven, CT: Yale University Press. Policy. New York: Oxford University Press. Martin, P., M. Abella, and C. Kuptsch. 2006. Managing Labor Migration in ------. 2006. "The Economic Implications of Remittances and Migra- the Twenty-First Century. New Haven, CT: Yale University Press. tion." Global Economic Prospects. http://www.worldbank.org/prospects/ Martin, P., and G. Zuercher. 2008. "Managing Migration: The Global gep2006. Challenge." Population Bulletin 63 (1): 3­20. ------. 2008. World Development Report 2008: Agriculture for Develop- Mattoo, A., and M. Olarreaga. 2004. "Reciprocity across Modes of Supply ment. Washington, DC: World Bank. in the WTO: A Negotiating Formula." International Trade Journal ------. 2009. World Development Report 2010: Development and Climate 18 (1): 1­22. Change. Washington, DC: World Bank. 23 Trade and Climate Policies after the Crisis Jaime de Melo and Nicole A. Mathys As the world begins to leave the financial crisis behind, represented by the WTO are on a collision course. Businesses sustained growth is expected to continue largely because fear that mitigation policies will affect their competitiveness, the globalization of recent decades has unleashed techni- and environmentalists fear that international trade will cal progress that has spread increasingly rapidly across undercut policies to reduce greenhouse gas (GHG) emis- countries. That globalization is also expected to continue, sions. Both groups are appealing to the WTO for rescue.1 because it is taken for granted that today's open world trad- For a long time, trade and the environment have been ing system will continue to function adequately. At the odd bedfellows, as environmentalists have claimed that the same time, the continuing difficulties at the climate change interests of the trade community have trumped their con- and Doha trade negotiations echo concerns that the equity cerns. In acknowledgment of this concern, the launch of the challenges of growth recovery are not being met. The stale- Doha round explicitly recognized that environmental con- mates in these negotiations have the same losers: the least- cerns would be fully taken into account to address fears that developed countries and other vulnerable nations with the gains from growth and globalization could be undermined weak institutional infrastructures. These countries have by their environmental side effects. First, globalization-induced difficulties carrying out the policies needed for a sustained increases in trade can magnify cross-border pollution. Second, recovery, policies that include mitigation and adaptation improvements in technology make it increasingly easier to measures to deal with climate change. intensify the exploitation of natural resources, potentially exac- Several observers view the shift at the Copenhagen meet- erbating the depletion of natural capital. ing in 2009 from a top-down approach with a collective tar- As a consequence of globalization, national decisions get under the current Kyoto Protocol (KP1) to a bottom-up have a growing impact on other countries. This changing accord as a shift from an approach favoring environmental landscape motivated the creation of the Global Task Force objectives to one favoring political feasibility. Under this on Public Goods in 2003, and many policy publications have approach, the world lacks an agreed standard for determining addressed the need to expand the provision of the global whether a country is doing its fair share to limit global commons. Frequently mentioned areas include biodiversity warming. Because the United Nations Framework Conven- and ecosystems; water resources; fisheries; and, now, miti- tion on Climate Change works by consensus, a meaningful gating climate change, the "ultimate" global public good. But binding agreement with a mechanism for dispute settlement the sheer magnitude of the climate problem justifies fears akin to the one in the World Trade Organization (WTO) that trade and climate change objectives may be at odds. does not appear likely at the next meetings. Some say that This chapter takes stock of the challenges ahead, start- the global climate regime and the global trade policy regime ing from the observation that trade and environmental This is a shortened version of de Melo and Mathys (2010). We thank Jean-Marie Grether, Patrick Messerlin, Richard Newfarmer, Gisèle Schmid, and Ludivine Tamiotti for helpful comments on an earlier draft and the World Bank and the Swiss National Science Foundation (grant 100012-109926) for financial support. De Melo also thanks FERDI (Fondation pour les etudes et recherches sur le développement international) for support. All remaining errors are our sole responsibility. 301 302 Managing Openness negotiations both face the same problem: providing a ex post leakage in the case of SO2, that is, sulfur dioxide, a "public good" whose supply requires widespread partici- pollutant prevalent in the same sectors as CO2. We then pation and compliance. Drawing on this analogy, we point report on the ex ante estimates for CO2, questioning the out that the environmental negotiations can learn from likely accuracy of these estimates. We also discuss the con- the past evolution of the world trading system, first under sequences for developing countries of alternative border the General Agreement on Tariffs and Trade (GATT) and adjustment measures. The next section tackles the uncer- then under the WTO. We recognize, though, that the com- tain legality of border adjustment measures at the WTO pliance necessary to bring about collective action in the case and the recent stalled initiatives to reduce protection on of climate change policies is fundamentally different. As environmental goods. The final section closes with some noted by Barrett (2010), trade is bilateral, so that trade agree- thoughts about lessons that the multilateral trading system ments can more easily be enforced by a strategy of reciproc- could offer to the climate change negotiators. ity. In contrast, climate change mitigation is a global public good in which reciprocity is a weak enforcement tool, The Contours of the Next Multilateral because those who impose sanctions are hurt in the process. Climate Agreement It is not surprising, then, that KP1 did not include trade sanctions as an enforcement mechanism. The challenge ahead is clear from figure 23.1, which shows Four aspects that complicate the design of appropriate the evolution of CO2 emissions (outside of agriculture) trade policies for dealing with mitigation and its conse- over the past 40 years for a very large sample of countries. quences are addressed. First, any serious attempt at mitiga- Even though it does not include all sources and sinks, the tion will require a higher price, perhaps around US$100 per figure shows the emissions that are likely to be amenable to metric ton of CO2e, that is, carbon dioxide equivalent. the trade policy levers, which are the concern of this chap- (Currently, the price in the over-the-counter market in the ter. For the world as a whole, emissions growth has not emission-trading system, or ETS, is around US$20 per met- slowed down since concerns about climate change surfaced ric ton.) The higher price will result in huge rents, whose around 1990. Since GHGs stay in the atmosphere for over a contestability will lead to much resource waste that should century, the stock is increasing. Much of this growth is be addressed in the architecture of the next climate agree- catch-up from Brazil, China, and India, shown in panels c ments. Second, all forms of energy generation require and d. As with almost everything else on climate in which irreversible investments in capital equipment. Very forward- the only "certainty is uncertainty," orders of magnitude of looking decisions under uncertainty require a relatively pre- the marginal damage are largely unknown. There is now dictable price of carbon, the predictability being enhanced by broad consensus, however, that damage from emissions a multilateral framework outlining the contours of the path will increase at least until stabilization in the stock is ahead, a requirement that cannot be achieved by unilateral achieved. Figure 23.1 also shows that per capita emissions action. (The life of an investment project is more than 50 have stabilized for the three high-income heavy per capita years for electricity generators, for example, and 60­100 years emitters (Australia, Canada, and the United States), in for residential buildings.) Third, measuring the emission of spite of growth in gross domestic product (GDP) and GHGs in manufacturing and in agriculture is difficult, com- population, although still at about four times above safe plicating the measurement of leakage. Furthermore, the levels. It is also evident from panels c and d that differ- carbon content of energy is not visible in a product, compli- ences in per capita emissions are huge across country cating the application of border measures to counter any groupings and that Australia, Canada, and the United leakage toward nonsignatories. Fourth, since the damages States have a long way to go to reduce their per capita from GHGs are truly global, it is extremely difficult to achieve emissions to that of the OECD-26 group. Finally, CO2 the necessary collective action, given that everyone shares the emissions fell sharply during the two oil-price spikes in benefits while only the participants bear the costs. As a result, the 1970s, suggesting that a tax on the price of fossil fuels starting with unilateral actions to build confidence and would lead to a fall in emissions. reduce the gaps might be necessary. Three arguments favor acting now on climate change. The next section delineates desirable features for a First, the uncertainty involved is not marginal, and the pos- multilateral agreement and the role of trade and trade sibility of a tipping point lowers the discount rate and calls policy in an evolving world trading system where climate for early action. Intergenerational equity is another issue change occupies a progressively larger role. The following that has been debated in connection with the Stern report section then examines the evidence of leakage effects (Stern 2006). Second, because the hardest hit will be the from a carbon tax by examining first the evidence of poor, action needs to be taken now to reduce the adverse Trade and Climate Policies after the Crisis 303 Figure 23.1. Total and Per Capita Carbon Emissions for the World and Various Regions, 1960­2002 a. CDIAC166 world sample, total b. CDIAC166 world sample, PCE 7,000 1.2 tons per capita 6,000 millions (mt) 1.1 5,000 1.0 4,000 3,000 0.9 2,000 0.8 60 70 80 90 00 60 70 80 90 00 19 19 19 19 20 19 19 19 19 20 year year c. US-CA-AUS, OECD-26, BRIC-3, d. US-CA-AUS, OECD-26, BRIC-3, RoW-134, total RoW-134, PCE 2,000 6 tons per capita 5 millions (mt) 1,500 4 1,000 3 2 500 1 0 0 60 70 80 90 00 60 70 80 90 00 19 19 19 19 20 19 19 19 19 20 year year US-CA-AUS BRIC-3 OECD-26 RoW-134 e. World regions, total f. World regions, PCE 2,500 6 tons per capita millions (mt) 2,000 5 1,500 4 3 1,000 2 500 1 0 0 60 70 80 90 00 60 70 80 90 00 19 19 19 19 20 19 19 19 19 20 year year ECA NA EAP and SA MENA LAC SSA Source: Adapted from Ordás Criado and Grether forthcoming. Note: BRIC-3 = Brazil, China, and India; USA­CA­AUS = United States, Canada, and Australia; OECD-26 = Organisation for Economic Co-operation and Development without Australia, Canada, and the United States; RoW-134 = the remaining countries, hence a total of 3 + 3 + 26 + 134 = 166 countries; ECA = Europe and Central Asia; EAP = East Asia and Pacific; SA = South Asia; MENA = Middle East and North Africa; LAC = Latin America and the Caribbean; SSA = Sub-Saharan Africa; PCE = per capita carbon emissions. Initial data are from the Carbon Dioxide Information Analsysis Center (CDIAC) and reflect anthropogenic emissions from fossil fuel consumption, cement manufacturing, and gas flaring, ignoring fuels supplied to ships and crafts. To convert carbon emissions into CO2 emissions, apply a multiplicative factor of 3.67. effects to which they will be subject. Third, the investments · Increasing spending on research and development (R&D), in green renewable energy called for are long term, and both public and private to develop green energies investment decisions are inherently dynamic. The longer · Combating deforestation, because it is an important action is delayed, the more investment will take place in source of GHG emissions emission-intensive activities because they will remain rela- · Providing aid to support adaptation in the often-hostile tively more profitable than clean energies. However, the environments in less-developed countries slow pace at which collective action is building indicates · Establishing an effective emission-trading regime when that acting rapidly will not happen.2 countries choose to reduce emissions by issuing permits, Because of the magnitude of the problem, mitigating a choice that could be adopted along with a tax and climate change will require action on several fronts: technical regulations in certain sectors.3 304 Managing Openness Addressing Equity Concerns on several commentators observed that its architecture-- which was "deep and shallow" rather than "broad, then The United Nations Framework Convention on Climate deep"--had received insufficient attention (see, for exam- Change calls for a common but differentiated responsibil- ple, Schmalensee 1998). The lack of compliance mecha- ity in addressing climate change as well as long-term flexi- nisms is particularly noteworthy, and a comparison with bility in implementation. In particular, any successful cli- the hugely successful Montreal Protocol (MP) that mate change regime will have to address burden sharing to addresses ozone depletion, another global public good, is take into account that since around 1990 (when global instructive, even though ozone depletion and climate warming became a widespread public concern), high- change are very different problems. income countries have been responsible for a larger share Barrett (2010) argues that trade policy played a signifi- of the current stock of emissions. Taking into account cant compliance role in the MP. First, all countries complied intra- and intergenerational equity will also require that with the MP through strong incentives against deviation countries converge on some per capita measure. The two (trade sanctions) and in favor of participation (financial broad alternatives are "per capita comparability" and "car- payments from rich to poor nations). Also the MP mini- bon price equivalency." Because countries are reluctant to mized trade leakage by putting a cap on consumption acquiesce to subsidiarity on tax matters, a top-down (which is derived as production ­ exports + imports), approach with an agreement calling on countries to con- whereas KP1 put a cap only on production. Finally, the MP's verge on a carbon price subject to sanctions for deviation caps are permanent, thereby stimulating innovation. Inter- is unlikely ever to be agreed, despite its superiority on eco- estingly, the MP is estimated to have been indirectly far more nomic grounds. effective at mitigating hydrofluorocarbons (HFCs), a by- As a result, any broadly supported agreement will have to product of hydrochlorofluorocarbons (HCFCs), an ozone- rely on some convergence toward equal entitlements on a depleting substance that is a GHG not covered by KP1. Had per capita basis, some arguing that it should be in terms of HFCs been addressed in a separate (that is, a sectoral) agree- access to energy needs rather than emissions. Taking con- ment, they could have been cut dramatically, and perhaps vergence in per capita emissions as an equitable target, phased out. Interestingly, the MP was also strengthened over Spence (2009) computes such emissions paths over a time with seven amendments (see Barrett 2010, appendix 50-year period. Per capita emissions would have to decline B); that is, its architecture was "broad, then deep," rather by 4 percent per year in Australia, Canada, and the United than "deep and shallow" as under KP1. States and by 2.6 percent per year in the other high-income If participation and compliance were the missing ele- countries, mostly the OECD (Organisation for Economic ments in KP1 and trade was integral in achieving both Co-operation and Development)-26 in figure 23.1. All but compliance and participation under the MP, what poten- the slow-growing developing countries, therefore, will have tial role is there for trade policy under KP2 and its succes- to accept an agreement that will call on emissions caps in sors? Trade could come into play through four channels. the near future. First, an open trading system with high trade volumes is These trajectories suggest some building blocks for the essential, including intellectual property rights that recog- next climate agreement. First, there should be broad partic- nize that technology transfers will have to be encouraged at ipation in any future agreement and recognition that fast- the same time that incentives for R&D are maintained. The growing developing countries should "graduate" soon from reason is that much technology development and transfer their current status under KP1. This graduation means take place through trade. being part of a cap on emissions. Second, because the Second, it will be necessary to separate where abatement uncertainty is great--a factor of 3 according to Schelling takes place from who bears the costs of abatement. Marginal (2007)--regular negotiations are needed along the way (the costs of abatement differ widely with many "no-regrets" specificities of each successive agreement taking place every energy-saving opportunities in developing countries. five years). KP1, which covers the period 2008­12, has set Under those circumstances, a global carbon credit trading up targets for renegotiation after a five-year period. system building on an improved Clean Development Mechanism introduced under KP1 will be necessary. Implementing the mechanism involves not only trade in Trade and Trade Policy to Address the Kyoto Failure credits, but also trade through technology transfer. KP1 is generally seen as largely ineffective because it lacks Third, trade sanctions could enter as a participation enforcement and participation has been low. This failure mechanism, although how the role of sanctions would be calls for a new architecture, and it is interesting that early rendered effective is difficult to assess. That role is likely to Trade and Climate Policies after the Crisis 305 be more promising if the new architecture leans toward a transport, which accounts for 13 percent of total emissions, system of treaties. Countries participate in an interna- much is national transport that does not involve trade. In tional treaty mostly to influence the participation of oth- sum, about 15 percent of emissions involve international ers and of nonsignatories rather than to tie their hands. trade. The issue then is how much trade really matters and Members of an international agreement can establish what the potential is for leakage, a major source of concern trade sanctions for nonparticipants who do not comply, as in the climate negotiations. in the case of the MP. Although to be successful, this Almost all the evidence on the environmental effects of approach requires a high degree of participation, thereby trade has focused on local pollutants. The early evidence obviating the need to use trade sanctions. This approach (for example, Grossman and Krueger 1993), identifies has the advantage that new principles could be drawn up three effects through which trade liberalization may affect for a climate change treaty rather than relying on the cur- the environment: rent complicated rules governing the use of border adjust- ment measures to address competitiveness effects (see the · A scale effect. Increased economic activity from trade section on climate change and the WTO). For example, liberalization leads, ceteris paribus, to increased emis- only a handful of high emitters of CO2 would be covered sions. under the new set of rules. Building such an agreement · A composition effect. Trade liberalization may lead to might be easiest with few participants but would come at a changed specialization patterns across countries and cost, because a substantial amount of trade might then sectors with different emission intensities, which can take place among nonparticipants, making it costly (and, trigger changes in emissions. hence, not credible) for participants to apply the agreed- · A technique effect. Through increased income and tech- upon sanctions. It is necessary to be cautious, however, in nology transfer, trade can lead to cleaner production transposing trade measures that worked under the MP to technologies. GHG agreements. The dimension of the problem is much larger in the latter case, and hence, any trade sanctions It is important, however, to know the origin of comparative would have larger effects and might potentially be much advantage in polluting industries (is it energy subsidies, for more destructive. example, or a lax environmental policy?) and whether those Fourth is the issue of leakage and the pressure for bor- industries might migrate to "pollution havens" with less der measures discussed in the section on climate change stringent environmental policies. In the case of the Kyoto and the WTO. Non-CO2 evidence suggests that leakage Protocol, would caps among signatories lead to a migration may not be as important as some might believe. Nonethe- of industries toward nonsignatories (or under KP1 to signa- less, the pressure for countervailing action against coun- tories that are not under a cap)? In an open-economy tries (like the United States) that in effect subsidize their model incorporating trade and environmental policies, industries by not correcting the externality due to CO2 comparative advantage in polluting industries is shaped by emissions will be great. It has already surfaced under two forces working in opposite directions: the mild cuts of KP1. It is hard to imagine that this pres- sure will not be strengthened in the successors to KP1, and · A pollution haven effect. Lax environmental regulation trade wars are a real possibility as the parties involved are in low-income countries would give them a compara- major players in international trade (China, the European tive advantage in emission-intensive industries. Union, and the United States, for example). Dealing with · A factor endowment effect. Capital abundance in high- the pressure for border adjustments will be an integral part income countries would give them a comparative advan- of the current negotiations on climate change. As discussed tage in emission-intensive sectors because they also in the section on climate change and the WTO, the WTO- happen to be capital intensive. legality of border measures is uncertain, and current rules appear inadequate to deal with the leakage issues raised by The first issue, then, is whether the possibility of trade CO2 emissions. results in more CO2 emissions than if there were no trade. For example, is it the case that exporters of chemical prod- ucts, an energy-intensive sector, emit more than countries The Role of Trade in GHG Mitigation that import chemical products? Next, is it likely that Many sources of CO2 emissions such as power (mainly energy-intensive industries would migrate if the price of electricity production) and buildings are nontradable energy were to change differentially across countries as cli- activities, hence not directly subject to leakage. As for mate change policies are put in place? Finally, what level of 306 Managing Openness emissions is embodied in international transport? We embodied emissions and energy content increased between examine the evidence on all three issues below. 2001 and 2006 (Pan, Philips, and Cheng 2008). Although balances of emissions embodied in trade are not a measure of leakage (see later text), these patterns, especially those Emissions Embodied in Trade for China, suggest the possibility of carbon leakage. If one takes into account indirect emissions through interindustry links, international trade and transport have Estimates of Carbon Leakage made up roughly 23 percent of production-based emissions in recent years (see figure 23.2 and Davis and Caldeira 2010). Leakage4 effects (known as the "pollution heaven" effect in Figure 23.2 also shows the emission content of imports and the trade and environment literature) during reduction in exports from which one can compute a country's balance GHG emissions come about through two channels. First, of emissions embodied in trade, expressed as a percentage of emission-intensive industries could simply relocate to production-based emissions. A positive (negative) balance countries where policies are less stringent (see, for exam- means that the country is a net importer (exporter) of CO2 ple, Copeland and Taylor 2003). This effect can take place emissions in its trade bundle. Developed countries are net either through a reallocation of market shares or through importers (for example, ­7 percent for the United States increased investment in unconstrained industries. The and ­15 percent for Japan). Developing countries, in con- second effect is indirect. By increasing the cost of emission- trast, are net exporters of emissions, with China, Malaysia, intensive goods through climate policy, the derived South Africa, and Venezuela having positive balances demand for carbon-intensive inputs (oil, for example) is above 15 percent. Other estimates indicate that China's net reduced, potentially lowering the price of oil on world Figure 23.2. Embodied CO2 Emissions in Imports and Exports for Selected Countries, 2001 Australia Brazil Canada China France Germany India Italy Japan Korea, Rep. Mexico Russian Federation South Africa United Kingdom United States world 0 10 20 30 40 50 percentage of total domestic emissions (production) imports exports Source: Peters and Hertwich 2008. Note: Originally based on data from the Global Trade Analysis Project (GTAP) version 6 database, GTAP v6 data for 87 countries in 2001 using domestic input-output data for bilateral trade flows. Trade and Climate Policies after the Crisis 307 markets. This indirect effect would also shift comparative and between-country effects contributed to a reduction in advantage toward countries that are not implementing a emissions. Particularly noteworthy is the absence of leak- climate policy. In both cases, the effectiveness of GHG mit- age effects for the United States (the mandated reductions igation would be reduced. in SO2 and the emission-trading system entered into effect Instructive predictions about the likely effects of a rise in in 1996). In sum, insofar as SO2 is comparable to CO2, SO2 the price of carbon can be gleaned by looking at the case of estimates suggest that leakage effects would be small. In SO2. SO2 is responsible for acid rain, a regional phenome- addition, reduction in emissions came through the tech- non. SO2 and CO2 (emission) intensities are in fact highly nique effect, pointing toward the (obvious) necessity of correlated across industries. Not surprisingly, the same six developing clean sources of energy in the struggle to miti- industries are the main emitters of both gases: petroleum gate CO2 emissions.7 products, pulp and paper, nonferrous metals, iron and steel, Most estimates, however, come from ex ante simulation chemicals, and building material (cement). Therefore, it is models (multiregional general-equilibrium, or MR-GE, worthwhile to review studies based on SO2 and other models). These models also provide estimates of the effects energy-intensive industries to get a feeling for the potential of border tax adjustments (BTA) to prevent carbon leakage leakage under likely policies mitigating climate change. (the legal aspects of the introduction of border measures First, the more tradable among these sectors are largely are discussed later). MR-GE estimates suggest that a BTA weight-reducing industries. Smelting nonferrous metals will reduce leakage by half. The reason for the relative inef- (and the processing of paper from wood) usually takes ficiency of a BTA is that a tax on the CO2 content of place close to extraction sites to avoid transport costs. imports has a strong terms-of-trade effect in favor of the Grether and de Melo (2004) estimate a bilateral trade country that imposes the BTA, thereby leading it to model for each of these dirty industries and an aggregate of increase its volume of imports. "clean" industries and find a consistently higher absolute These models also give estimates of the different BTAs value of the distance coefficient for dirty industries (in the that have been proposed in the political debate. One of the range ­1.10 to ­1.40 except for nonferrous metals estimated proposals circulated in the United States was to adjust the at ­.95, while the average for clean industries is ­0.82).5 price of imports by applying the CO2 tax in the United This finding, and the fact that extraction in natural States to the total (direct and indirect) carbon content of resource­based industries cannot migrate, suggest that imports, perhaps along with relief from paying the tax for transport costs would deter relocation of much processing exporters. Another proposal was to tax imports on the to countries with lower regulation standards. basis of their carbon content (U.S. legislation makes it pos- Other work has also tried to detect pollution-haven sible to require importers to buy emission allowances effects in the case of SO2, where national attempts at con- equivalent to the carbon content of imports). Mattoo et al. trolling acid rain could have led to leakage. Using concen- (2009a, 2009b) estimate that if industrial countries reduce tration data, Antweiler, Copeland, and Taylor (2001), esti- emissions by 17 percent without applying a BTA, manufac- mate that if an increase in trade openness generates a turing exports by developing countries remain unchanged 1 percent increase in income and output, then pollution but fall by about 2 percent under the first proposal and by will fall by approximately 1 percent.6 Drawing on an exten- 15 percent under the second proposal. Should developed sive database covering the same six sectors and a "clean" countries try to impose across-the-board taxation on sector for 62 countries with emission coefficients for 1990 imports based on their carbon content, there would be a and 2000, Grether, Mathys, and de Melo (2009) decompose collision between developed and developing countries at changes in SO2 emissions. The scale effect accounted for a the WTO. 10 percent increase in emissions, while the shift toward While these model-based estimates are transparent and cleaner industries within (across) countries accounted for are arguably the best game in town so far, they differ a decrease in emissions of 2 percent (3 percent) and the sharply from the few econometric and ex post studies. technique effect accounted for a decrease in emissions of Drawing on data for 38 countries (26 of which have ratified 14 percent. As a result, while manufacturing growth was Kyoto) and 12 sectors over the period 1995­2005, Aichele 8 percent over the decade, SO2 emissions fell by 10 percent and Felbermayr (2010) estimate the impact of different or 1 percent per year, about half the needed reduction in GHG policies on trade flows and emissions. They estimate emissions for the OECD countries for a convergence in that carbon imports are on average 12 percent higher if CO2 emissions over a 50-year period. the importer has ratified Kyoto but the exporter has not. On a sectoral level, only nonferrous metals had an Confirming previous work, the effect is most important in increase in emissions per unit of output. Between-sector energy-intensive industries, where robust evidence for 308 Managing Openness carbon leakage was found for seven sectors. Their findings 1990 and by only 3.5 percent in 2000, largely a reflection of suggest that, even though the volume of trade "caused" by the all-around decline in emission intensity. Using a rough Kyoto is rather small, on average about 40 percent of car- decomposition by mode of transport, they estimate that bon savings due to the ratification of the Kyoto Protocol emissions related to international trade accounted for about have been offset by increasing emissions in noncommitting 5­9 percent of total manufacturing emissions. Adding emis- countries. sions related to trade and transport costs emissions, their We noted earlier that an open world economy is impor- estimates suggest that transport-related emissions rose from tant in facilitating technological transfers. Di Maria and accounting for roughly one-third to accounting for three- van den Werf (2008) and Acemoglu et al. (2009) argue that fourths of total trade-related emissions over 1990­2000. To effective long-run carbon leakage might be smaller than put it differently, if one adds emissions coming from trade- the results from the previously mentioned studies, because related composition effects to those coming from trade- induced technological change might have a positive effect. related transport activities, the sum shows an increase in Using panel data in a gravity model, Costantini and Maz- global manufacturing emissions through trade of 16 percent zanti (2010) analyze the effect of energy and environmen- in 1990 and 13 percent in 2000. tal policy in the European Union (EU) on export dynamics Both sets of estimates are only rough orders of magni- of four sectors over 1996­2007. Including patent data, they tude. Estimates of carbon footprints should factor in that find evidence that the overall effect of energy policy is not favorable climatic conditions and the use of low energy- in conflict with export competitiveness. In general, these intensive production technologies in developing countries results support the Porter hypothesis (Porter and van der may lead to carbon efficiencies that outweigh additional Linde 1995), which states that stricter environmental regu- emissions due to transport for products shipped to and lation can have a "halo" rather than a "pollution haven" consumed in developed countries.9 Reducing trade-related effect by inducing efficiency and encouraging innovation. GHG emissions would be best achieved by an agreement on transport that would impose low-emitting technologies (for example, hydrogen). Short of such an approach, rais- The Environmental Consequences of Transport ing transport costs by taxes on transport will reduce trade, Transport costs (national and international) account for which will in turn reduce GDP growth. If externalities were approximately 13 percent of global CO2 emissions (about corrected where they occur (that is, in the GHG-emitting two-thirds of the emissions related to manufacturing activ- sectors rather than by raising taxes on trade), however, then ity), with shipping-related emissions raising particularly world welfare would increase. strong concerns. How much, then, does international trade contribute to CO2 emissions?8 Climate Change and the WTO Two recent estimates provide orders of magnitude for a large number of countries. Hummels (2009) uses the Some fear that the gains from trade--which are huge-- Global Trade Analysis Project model to estimate that full could easily be eroded by trade policies seeking to counter- trade liberalization would increase trade volumes by act spillover effects from mitigation measures. Many also 5.8 percent. He then converts this growth into weight- agree that subsidies to energy and to substitutes for fossil distance profiles by modal usage (truck, air, and mar- fuels (biofuels, for example) should be removed. Here we itime) to which he applies GHG emission estimates. His focus on the interface of climate change and the WTO. First, results show a shift in trade away from close to distant are WTO rules and previous dispute settlement processes at partners and estimates that CO2 emissions could increase the WTO adequate to deal with the pressures to apply bor- by as much as 10 percent. Thus, insofar as trade liberal- der measures to counter leakage toward nonparticipants? ization is likely to involve growth at the extensive margin Second, how much progress has taken place on liberaliza- (new partners), it is not friendly to the environment. tion of tariffs on environmental goods and services?10 Hummels's estimates would thus suggest an elasticity of emission to trade growth of around 1.5. Border Measures to Address Competitiveness Grether, Mathys, and de Melo (2010) carry out a similar and Leakage Concerns exercise for worldwide SO2 emissions over 1990­2000 dur- ing which manufacturing output grew by 8 percent. They By changing carbon prices across countries, climate poli- first define an anti-monde in which countries produce what cies will lead to demand for border adjustments to they actually consume in the trade equilibrium and find address competitiveness effects. Indeed, proposed U.S. that international trade increases emissions by 10 percent in legislation (the Warner-Lieberman and Waxman-Markey Trade and Climate Policies after the Crisis 309 bills) included the possibility of border adjustments for Hence, the thorniest issue for the functioning of the energy-intensive industries that would lose competitive- world trade system is how to resolve the conflicts from dif- ness to countries without a similar policy. Such pressures ferent carbon policies. Would trade intervention justified have already appeared under KP1 against nonsignatories by differences in climate policy be allowed under WTO (for semifinished steel, clinker, aluminum, cement, and law?15 So far, this territory is uncharted in spite of several some basic chemicals).11 environmental cases that have been settled by the dispute Ideally, a worldwide price on GHG emissions (through settlement process. The answer depends crucially on the a tax or an emission-trading system) should be applied at specific design and implementation of the intervention. the marginal social cost of the emissions. This approach Generally, two types of trade measures could be imposed would allow internalizing external costs, such as negative on imports to complement mitigation policies. Restric- effects of emissions that are not taken into account by indi- tions could be with respect to "locally emitted" GHGs or viduals, since property rights are badly defined when it with respect to "foreign emitted" gases. In the first category, comes to the world's climate.12 However, as long as emit- emissions take place when the imported goods are "con- ting activities (such as burning coal, oil, and gas) are not sumed." Emission standards on cars, for instance, would subject to a cost internalizing these negative effects in some fall in this group. Pauwelyn (2007) states that if such meas- countries (that is, the nonsignatories of KP1), energy- ures do not discriminate between imports and local prod- intensive industries are implicitly subsidized. Except in ucts, they should generally be accepted under WTO law. limited circumstances like agriculture, the WTO does not If, however, trade restrictions address GHGs of imports allow subsidies. Thus, nonsignatories of KP1 like the that are emitted in the exporting country (that is, embod- United States are in effect enjoying an unfair trade advan- ied emissions in imported products), compliance with tage that KP1 signatories could bring to the WTO. Given WTO law is more controversial.16 An example would be that a single worldwide price on GHG emissions is not imported aluminum produced with high CO2-emitting politically achievable, many will call for border adjust- Söderberg technology. Consider then the application of a ments for the climate-induced change in competitiveness BTA on imported aluminum.17 Could it be defended under among WTO members. international trade law? Yes, if it were possible to convince Thus, much of the debate on the relation between trade the WTO panel that the imposed competitiveness provi- and the environment revolves around a border adjustment sions are only an extension of domestic climate policy to bring about a level playing field across countries in applied on an equal footing to imports. We illustrate the energy-intensive industries, following differential efforts at difficulties in the case where one country (party B in figure reducing emissions. Another motivation would be lever- 23.3) enacts a GHG tax and its trading partner (party A in age, in the sense that the threat of border adjustment could figure 23.3) does not have an equivalent climate policy be used to force countries to participate in climate negotia- under the assumption that leakage has occurred, since tions. While the first motivation calls for trade controls, the losses of competitiveness by domestic firms will not be second calls for trade sanctions. The focus here is on trade considered as a relevant argument. The idea is to apply at controls, applied to environmentally relevant sectors, as the border a charge equivalent to an internal tax, and con- opposed to trade sanctions, where targeted products are versely, one could rebate the same tax on domestic prod- unrelated to environmental discrepancies (Charnovitz ucts that are exported (this was the practice under the 2003). In the case of KP1, no trade measures were included application of the value added tax). Figure 23.3 shows the in the treaty, while in the MP, trade bans were included two main possibilities of how such a trade intervention from the beginning as an instrument for achieving full par- could be justified under WTO law. ticipation in the treaty.13 To be applied, a BTA must first be confirmed as cov- From an economic point of view, if the externality in ered by GATT articles II (tariff obligation, which production were corrected through a tax on carbon, any requires members to fix tariff levels at bound rates) and tax imposed at the border would lead to inefficiencies. It III (national treatment, that is, not to apply internal taxes would lead to trade diversion because countries without to protect domestic producers). Such BTA restrictions stringent climate policy would shift their exports to coun- are not covered by GATT article II on tariffs but by tries that do not apply any kind of adjustment at the bor- GATT article III:2 on domestic taxes (imports can be der. For example, in the EU, where the tax on gasoline is subject to adjustable product taxes but not producer about four times as high as in the United States, the exter- taxes). Article III.2 states that the instrument against nality is corrected. Then, a border tax on imports from, say, which adjustment is requested must be an indirect prod- the United States would create an inefficiency.14 uct tax (value added tax or excise tax, for example) and 310 Managing Openness Figure 23.3. Justifying Border Measures under WTO Law in a Nutshell Party A complains about discrimination of its imports in Party B. Is coverage provided by GATT Articles II and III? yes no Is there consistency with GATT Articles Party B continues II and III? Border tax adjustment is allowed with its practice. in respect of taxes "applied directly or indirectly, to like domestic products." yes no There is no discrimination Exception is possible under GATT Article XX-- with respect to "like products," so Chapeau: has not to be "unjustifiable or arbitrary"; and Party B continues with its practice. paragraph (g): "relating to the conservation of exhaustible natural resources ...." yes no Party B loses and has to Party B wins and continues change its practice to conform to with its practice. WTO standards; Party A can take counter measures. Source: Authors. not a direct producer tax (payroll tax or income tax, for If the previously mentioned justification for imposing example) and that it must be nondiscriminatory (see trade restrictions were not allowed (left-arm in figure 23.3), later text). This distinction is based on the destination violation of GATT articles II and III is still possible if envi- principle in which each product should be taxed only ronmental exceptions under article XX can be invoked. once, namely, in the country of consumption. As dis- Paragraph (g) is the key article related to the environ- cussed in Demaret and Stewardson (1994), under GATT ment,19 and it asks for three cumulative conditions likely to article III.2 border adjustment is allowed under taxes be satisfied. First, it has to be shown that the planet's "applied directly or indirectly, to like domestic prod- atmosphere is an exhaustible natural resource. Second, it ucts." The crucial question then is whether this rule also has to be shown that the given climate change policy as a applies to inputs that are physically not incorporated whole has a "substantial relationship" with the conserva- directly (such as GHG emissions released during produc- tion of the atmosphere. And third, broadly similar condi- tion) in the final product. It has been argued that the tions must be imposed on domestic producers. Accord- GATT Superfund case18 confirms that BTAs are possible ing to Pauwelyn (2007) and the latest scientific evidence, for inputs that are not physically incorporated in the the first two conditions should be easily met, and a stan- traded product. Under the presumption that the BTA can dard carbon tax should also meet the third. be justified under articles II and III, then the adjustment However, to be WTO-law compatible, an adjustment at must meet the WTO principles of national treatment the border must also satisfy the conditions under the intro- (article III) and most-favored-nation (article I). ductory (chapeau) phrase of article XX GATT. So far, Trade and Climate Policies after the Crisis 311 whenever the Appellate Body has found that the article XX and multiplying these quantities by emission coefficients. exception was not met, it has done so under the introduc- A special case is electricity, where the mix varies consid- tory phrase. Under this phrase, discrimination between erably between countries and, therefore, contains very different foreign countries (most-favored-nation type) and different levels of embodied emissions. Moreover, since between foreign countries and the home country (national exporting countries would be unlikely to truthfully treatment type) are not allowed. Three specific elements cooperate in the computation of the carbon content of are important. First, the legislation has to take into account their products, Pauwelyn (2007) has suggested computing local conditions in foreign countries. On the one hand, an implicit carbon content based on production techniques there are other ways to address climate change, and, on the dominant in importing countries in this case. Ismer and other hand, developing countries may not be asked to carry Neuhoff (2007) propose a processed-material approach with the same burden of adjustment as industrial countries. the best available technology. Second, before imposing the unilateral trade restriction, Still, implementing such a scheme would be a night- the country must have previously engaged in serious mare, even more so for developing countries, as ample across-the-board negotiations with the objective of con- experience with the less controversial tax rebate for tariffs cluding a bilateral or multilateral agreement. Finally, the on imported inputs in free trade zones has shown. Further, implementation and administration of the imposed trade BTAs may open the door to capture by powerful domestic policy have to respect "basic fairness and due process." For political interest groups and become the source of dis- example, concrete problems, like delayed imports related to guised protection. Finally, imposing carbon tariffs could be border controls, should be avoided. Jurisprudence has seen as leverage for inducing countries to participate in cli- highlighted the importance of flexibility in the design of mate negotiations, but, more likely, it could undermine the measure to take into account different situations in dif- trust between trading partners and poison the hard-to- ferent countries. acquire trust in multilateral negotiations and endanger In sum, the conditions related to the introductory future agreements. phrase of article XX GATT are still highly controversial for In sum, even if carefully constructed and implemented leakage effects arising from cost differentials due to differ- border measures could in theory pass WTO law, considera- ent energy and environment policies. Notwithstanding the tions of political economy suggest that it might not be a recent environmental disputes that the WTO has adjudi- good instrument for leveling the playing field for energy- cated, it is easy to imagine that leakage concerns will be intensive industries. Pauwelyn (2007) puts forward several subject to extensive litigation. other instruments that might be used instead. As discussed In the case where an ETS is enacted instead of a carbon earlier, a country that adopts an ETS can alleviate competi- tax, concerns discussed earlier are also raised. In addition, tion effects by adopting flexibility mechanisms (such as the it must be shown that the obligation to hold permits is Clean Development Mechanism under the Kyoto Protocol) equivalent to an internal tax since the ETS is also impos- that allow companies to trade emission permits and, hence, ing a cost on firms if they have to be purchased. More- reduce abatement costs. Other possibilities are free distribu- over, if some permits are handed out for free, instead of tion of all (or at least part of) emission permits or exclusion being auctioned off, some kind of average market price of industries that are most exposed to international compe- would have to be used to determine the level of the effec- tition from stringent emission-reduction programs. The tive cost and, hence, the tax. Ismer and Neuhoff (2007) problem with this measure is that it is in effect a subsidy that investigate conformity and feasibility of a BTA for an could be actionable under GATT law (that is, it does not fol- ETS. They propose an adjustment level that does not low the "polluter-pays" principle, and it might create wind- exceed the upper bound of the amount payable for pro- fall profits for highly polluting industries). Also, safety valves duction in the country enacting the ETS using the best could be imposed, thereby assuring companies that permit available technology. prices would not exceed a maximum level. Several other problems lurk on the horizon. To begin with, as others have pointed out (for example, Bhagwati Liberalization of Environmental Goods and Services and Mavroidis 2007), the implementation of such border adjustment with respect to embodied carbon would be Under the 2001 Doha Ministerial Declaration, WTO mem- very costly.20 Indeed, Ismer and Neuhoff (2007) say that bers have been asked to negotiate on the reduction or, as only an indirect scheme, using a measure highly correlated appropriate, elimination of tariff and nontariff barriers with effective embodied carbon emissions, would be on environmental goods and services.21 The aim is to cre- implementable. They suggest using specific input mixes ate a triple-win situation--for trade, the environment, and 312 Managing Openness development. If negotiations are successful, trade will be 40 percent of trade in environmental goods and services, facilitated through reduced or eliminated tariffs and nontar- with 40 percent of imports and 42 percent of exports. Con- iff barriers on environmental goods and services. Agreement cerning tariffs, however, there is a clear difference between on this measure would decrease the cost of environmental countries of various income levels. Developed countries technologies, increase their use, and stimulate innovation have MFN tariffs below 2 percent on average, whereas and technology transfer. low-income developing countries and other developing Developing countries would benefit in two ways. First, countries have average MFN tariffs of 11 percent and producers of environmental goods and services would have 6 percent, respectively. Overall, the trade and protection better access to large markets in Europe, the United States, pattern in the tariff lines submitted for negotiations over and high-income Asia, and, second, it would be easier for environmental goods and services seems to be very similar developing countries as a whole to obtain high-quality to the overall pattern. environmental goods on world markets. Such access should The triple-win expectation from a liberalization of trade increase energy efficiency and improve the water and sani- in environmental goods and services has been challenged. tation situation in developing countries. Increased trade in Consider a vertical industry in which the downstream these products would lead to overall welfare gains and good's production is polluting and the upstream industry would help developing countries reduce emissions through that faces international competition is engaged in R&D to technology transfer. develop abatement technology that it can sell to the down- Even though liberalization in these particular product stream firm. In this setting, Dijkstra and Mathew (2009) categories should be win-win, WTO members have been show that it might be best for developing countries to first unable to agree on which goods should be liberalized. liberalize trade in environmental goods with countries Three approaches have been proposed to identify these whose environmental technologies are not too much better goods. In the list approach, WTO members would have to than their own. This approach would stimulate R&D in the agree on a list of products (at the six-digit level of the Har- domestic eco-industry and put it in a better position to face monized System, or HS) for which trade barriers should be competition from more advanced eco-firms later. They reduced. So far, WTO members have been unable to agree also show that when cleaner technologies become avail- on such a list. Developing countries argue that a number of able, governments have an incentive to reduce the effective products on the list are of export interest primarily to tax on polluting output to take the opportunity to increase industrial countries. Also problematic is the issue of dual welfare through the higher profits of the downstream firm use, because many product categories proposed also and increased consumer surplus. Hence, they show that include products that have nonenvironmental uses. It liberalization in environmental goods and services might would be possible to select "ex-outs" (goods that are not lead to a situation in which consumer benefits come at the separately identified at the 6-digit level of the HS and have expense of the environment. Next, consider an environ- to be identified in national tariff schedules at the 8- or 10- ment in which the North innovates and its technologies digit level), but this approach would be particularly costly diffuse to the South. In a stylized model, Di Maria and and difficult for developing countries. In response to this Smulders (2005) investigate the role of endogenous tech- issue, Argentina and India proposed a define-by-doing nology and technology spillovers. They conclude that approach, in which national authorities would select proj- endogeneous technological change is potentially but not ects. Environmental goods and services necessary for the necessarily a blessing. The lack of intellectual property selected projects would then temporarily benefit from rights in the South and innovation costs born asymmetri- enhanced market access. Categories of goods could be iden- cally by consumers in the North create distortions. tified multilaterally in advance. Finally, Brazil suggested a There is little evidence on the effects of liberalization of request-offer approach. Under this approach, countries environmental goods and services. Using a partial equilib- would request specific commitments on tariff cuts on prod- rium simulation model applying import elasticities to ucts of interest and then extend those tariff cuts to all WTO trade data for the 18 most important developing-country members on the basis of the most-favored nation (MFN) emitters of GHGs, the World Bank (2008) assessed the effect clause. Such an approach might better take into account of trade liberalization (elimination of tariffs and nontariff developing countries' specific interests.22 barriers) for the following four technologies: clean coal Including all HS-6 tariff lines that have so far been pro- technologies, wind energy, solar photovoltaic systems, and posed in the lists officially submitted to the WTO,23 envi- energy-efficient lighting. Eliminating trade barriers would ronmental goods and services make up about 20 percent of increase import volumes by up to 13 percent. More gener- total trade value. Developing countries account for about ally, Dechezleprêtre, Glachant, and Ménière (2009) used Trade and Climate Policies after the Crisis 313 patent data from 66 countries for the period 1990­2003 sovereignty. As proposed by Antholis (2009), taking their and investigated the forces that drive transfer of technologies inspiration from the GATT, climate negotiators could set up mitigating climate change. Their regression results show that a General Agreement to Reduce Emissions, avoiding moving restrictions to international trade (treated as an exogenous too quickly into a full-blown organization like a world envi- variable) negatively influence the diffusion of patented ronment organization or a treaty under the UN process, knowledge. However, Managi and Kumar (2009) estimated which also requires unanimity. exogeneous and trade-induced technological change for It is noteworthy that under the GATT, sustained 76 countries over the period 1963­2000 and found that national tariff cuts took place without the appearance of trade-induced technological change, which makes up the loss of sovereignty. Much tariff reduction occurred uni- roughly one-third of overall technological change, had a laterally. Adjudication and dispute settlement could evolve negative effect on GDP and increased pollution. progressively as they did under the GATT and the WTO so The lack of progress on negotiations on environmental that fears of loss of sovereignty would be alleviated (under goods and services could be attributed to traditional fears the GATT, a panel decision against a member could be invoked by developing countries to protect sectors from blocked by the member). international markets dominated by industrial countries. What should the principles for an umbrella climate code Infant industries would then deserve temporary protection. be? Commentators broadly agree:24 carbon taxes should be Also, if a first-mover advantage or lock-in effects exist, then encouraged, though not required; auctioning of permits developing countries may have incentives not to liberalize should likewise be encouraged; trade-related GHG meas- these markets too quickly. Finally, negotiations over envi- ures should be limited; like products should be defined at a ronmental goods and services are only a secondary arena in broad enough level of aggregation (four-digit HS for the Doha negotiation round. Developing countries might Hufbauer, Charnovitz, and Kim 2009); and the modalities well be using the possibility to block negotiations in envi- for border adjustments and the management of a Carbon ronmental goods and services to better defend their inter- Added Tax should be flexible. Countries that subscribe to ests in the agricultural negotiations, in which developing this "green code" would benefit from a "peace clause" and, countries have greater interests at stake. thus, avoid being subject to WTO disputes. It is likely, how- ever, that this sensible approach would be difficult to estab- lish, because all activities would want to qualify for "green Final Reflections: Lessons from World space status" and the request for flexibility could easily lead Trade for Governing Climate Change to a made-to-measure rather than a transparent code. Reflecting on the outcome of the Earth Summit in Rio in This umbrella approach would give countries maxi- 1992, as KP1 was under negotiation, Norwegian Prime Min- mum leeway and create a more enabling environment for ister Brundtland said: "We knew the basic principles on international negotiations, as countries would be more which to build: cost effectiveness, equity, joint implementa- likely to commit. With well over two-thirds of CO2 emis- tion and comprehensiveness. But not how to make them sions coming from nontradable activities, and, hence, not operational" (Schmalensee 1998). Fifteen years later, with subject to direct carbon leakage, countries with large KP1 nearing its end and negotiations for KP2 underway, energy subsidies would first enact energy legislation. For what has been learned? Certainly, one would add that cost example, the United States could first pass an energy bill effectiveness has to go beyond flexibility mechanisms: it that would raise the price of carbon.25 requires participation and compliance as shown by the suc- Under the umbrella of a broad multilateral agreement, cess of the MP. Are there any lessons for the design of climate- which would set the basic rules like MFN and national treat- related trade policies from the evolution of the world trading ment under the GATT, much progress could still take place system, first under the GATT and then under the WTO? either in a small group, perhaps two or three countries, or even unilaterally, as it has so far in many instances. This route could be much easier than a treaty, at least for a country like An Umbrella Agreement with Leeway the United States.26 As mentioned earlier, unilateral reduction In the early days of the GATT, participation was among a in tariffs was the way most progress was made in the early small group of countries where negotiation was easier rounds of trade negotiations. Of course, unilateral action than under the now unwieldy WTO where unanimity is is certainly easier to envisage in the case of tariff reductions, required for all major decisions. The GATT thus made where most gains are internalized, than under GHG emis- progress toward free trade with agreements that bound sions, where all gains are equally shared so that the need nations in ways that did not impinge on their national for collective action is much greater. Under this simpler 314 Managing Openness architecture, in the initial steps forward, the UN process, come to the same conclusion that these same polluting industries are not footloose. which requires unanimity, would be bypassed. In the words 6. Applied to CO2 emissions, the Antweiler, Copeland, and Taylor of Antholis (2009), the new climate architecture would "gear (2001) framework suggests that trade is likely to increase emissions in the up" domestic steps that nations are willing to take. case of carbon dioxide (Cole and Elliott 2003). Managi and Kumar (2009) find that trade has a beneficial effect on emissions for OECD countries, but a detrimental effect for non-OECD countries. A Polycentric Approach 7. Levinson (2009) also carries out retrospective calculations for the United States over 1972­2001 using U.S. emission data for 1997 and Drawing on the extensive experience with the provision of input-output data to capture the indirect SO2-content for imports. Although his data do not allow him to capture the technical change in national and environmental public goods, Ostrom (2009) emissions or changes in emissions due to a within-sector shift toward pro- argues that, without denying the global nature of the prob- ducers with higher per unit emissions, he also finds a strong composi- lem, much progress on climate change can be achieved by tional effect away from the production of SO2-intensive manufacturing activities coupled with a shift away from SO2-intensive imports over this actions at multiple scales (the household, the region, and longer period of time. the country). Ostrom argues that this approach is invalu- 8. International transport is only a fraction of transport. Cadarso able in building the trust necessary to achieve the collective et al. (2010) estimate that international trade accounts for 4 percent of action that is still so elusive for climate change. In fact, all transport embodied in Spanish consumption. 9. Brenton, Edwards-Jones, and Jensen (2009) survey results from progress is currently taking place at the local and national carbon-accounting and carbon-labeling studies involving developing levels rather than at the multilateral level.27 While this countries. decentralized approach has the advantage of building con- 10. WTO and UNEP (2009) covers these issues in greater depth. 11. President Sarkozy of France proposed a BTA on U.S. exports to fidence, the very different approaches across countries and the EU because the United States did not participate in KP1, and it was a states will require much effort later on at the national level great cost in terms of concessions to industry (quasi-automatic access to to bring those measures into conformity with any future free allowances) to obtain that BTAs would be left out of the EU proposal for KP2. rules in the multilateral trading system. But the multilat- 12. Allocative efficiency would be achieved when marginal abatement eral trading system has withstood several shocks success- costs are equal to the marginal social costs of abatement. And if the same fully, including the rise of regionalism. Ideally, the confi- price applies everywhere, worldwide efficiency would be achieved. But it is dence built by decentralized measures will pay off and will very difficult to estimate with any confidence the social cost of carbon because it requires estimating the marginal damage from an extra unit of help build the needed commitment to achieve a binding carbon, which depends on assumptions about the future path of the econ- General Agreement to Reduce Emissions. omy, the discount rate, science, and so forth. 13. See, for example, Frankel (2009). In fact, the variable of concern under the MP was consumption, defined as "production ­ exports + Notes imports." Hence, trade flows were directly included. 14. As an example, since 2007 China has implemented export tariffs 1. Waters are further muddied by the disagreement in the scientific ranging from 5 to 25 percent on carbon-intensive products, including iron, community and its impact on public opinion as well as differences about steel, coke, and cement, so the United States would not seek to impose a what needs to be done as reflected in the very different prognoses in BTA. See the comment by Hu Tao in Brainard and Sorkin (2009). Nordhaus (2008), Stern (2009), and McKibbin (2010). Frankel (2008), 15. Relevant environmental WTO cases are US-Gasoline, US-Shrimp, Hufbauer and Kim (2010), and others see a train wreck between climate and Brazil-Retreaded Tyres. In a first decision by the Appellate Body in the change objectives and trade. One source of relative consensus is that dam- Tuna-dolphin case involving Mexico and the United States (under the ages will fall mostly on poor countries with little capacity to adapt. Recent GATT 1991), the United States lost the case when it argued that tuna estimates put up to 80 percent of the damage from climate change in low- could be imported only if it was not caught in purse-seine nets, jurispru- latitude countries (Mendelsohn, Dinar, and Williams 2006). dence that was overturned later in the shrimp-turtle decision, which in 2. Against acting rapidly is the concern that the marginal cost func- effect ruled that process and production methods could be invoked at the tion for mitigation may be very steep, and for some because marginal WTO for contingent protection. Technically, Shrimp-turtle did not explic- damages rise as GHGs accumulate, the optimal policy is dynamic, grow- itly approve that so-called nonproduct-related production and process ing stricter over time (see, for example, Nordhaus 2008). methods, but it provided that they could in certain circumstances be jus- 3. All three approaches require measurement of carbon emitted. tified under the GATT's article XX exception clauses. Trading schemes beyond hydrocarbon products will be severely limited by 16. Following Pauwelyn (2007), three measures in that category are the costs and difficulty of measuring emissions--hence, the three levers most likely to violate international trade law. Import bans or punitive tar- for reducing emissions (Schmalensee 1998). Geo-engineering is also iffs would violate article XI of the GATT, which imposes the elimination likely to be on the menu, but for most, it is considered a last-resort of quantitative restrictions, but also article I, which states that discrimina- option with carbon capture and storage probably a more promising tion between imports from different countries must be avoided. The same avenue (see Barrett 2008). holds for contingent measures that would also likely violate WTO law. 4. The carbon leakage rate = (increase in emissions due to climate Thus, antidumping duties against "environmental dumping" could be policy outside enacting country/decrease in emissions due to climate pol- levied only if import prices are below the prices in the country of origin. icy in enacting country). This is not a measure adapted to address competitiveness concerns ema- 5. A distance coefficient estimate of ­1.40 (­0.95) implies that if trade nating from different climate policies. Countervailing duties to offset the flows are normalized to 1 for a distance of 1,000 kilometers, a doubling de facto "subsidy" of imports in their country of origin are the third meas- of distance to 2,000 would reduce bilateral trade volume to 0.38 (0.52). ure likely to violate WTO law. Several conditions are necessary to show On the basis of U.S, data, Ederington, Levison, and Minier (2005) also that the government "fails to act"--that is, to impose a carbon policy Trade and Climate Policies after the Crisis 315 instrument--to be allowed to impose countervailing duties. Most impor- Antholis, W. 2009. "Five Gs: Lessons from World Trade for Governing Global tant, not imposing a climate policy is not likely to be considered as either Climate." Mimeo, Brookings Institution, Washington, DC. http://www a specific subsidy (specific to a group of enterprises or industries) or an .brookings .edu/articles/2009/0403_trade_climate_antholis.aspx. export subsidy. Antweiler, W., B. Copeland, and S. Taylor. 2001. "Is Free Trade Good for 17. Border measures could also take the form of mandatory allowance the Environment?" American Economic Review 91 (4): 877­908. purchases by importers or embedded carbon product standards. Barrett, S. 2008. "A Portfolio System of Climate Treaties." Discussion 18. In the case of the GATT Superfund case, the dispute panel found Paper 08-13, Harvard Project on International Climate Agreements, that U.S. taxes on "chemicals used as materials in the manufacture or pro- Harvard Kennedy School, Cambridge, MA. duction of the imported substances" might be taken into account when ------. 2010. "Climate Change and International Trade: Lessons on Their imposing BTAs on imported like products. Linkage from International Environment Agreements." Paper pre- 19. The defendant party could also invoke paragraph (b) in article sented at TAIT (Thinking Ahead on International Trade) conference XX, which is more difficult to fulfill because it asks for a closer link sponsored by the Graduate Institute Centre for Trade and Economic between the policy and the environmental goal. It reads: "necessary to Integration, the World Bank, and the World Trade Organization, protect human, animal or plant life or health." Geneva, June 16­18. 20. In Messerlin's (2010) words: one would need a gigantic database, Bhagwati, J., and P. C. Mavroidis. 2007. "Is Action against U.S. Exports for generating astronomical transaction costs, and the risk of corruption Failure to Sign Kyoto Protocol WTO-legal?" World Trade Review 6 (2): would be huge. 299­310. 21. Negotiations take place in the Special Session of the Committee Brainard, L., and I. Sorkin, eds. 2009. "Climate Change, Trade and Com- on Trade and the Environment. The Doha negotiations also address the petitiveness: Is a Collision Inevitable?" Brookings Trade Forum 2008/ question of the relationship between the WTO and multilateral environ- 2009. Washington, DC: Brookings Press. mental agreements. For more on the link between the WTO and these Brenton P., G. Edwards-Jones, and M. F. Jensen. 2009. "Carbon Labelling environmental agreements, see WTO and UNEP (2009, 82­83). and Low-income Country Exports: A Review of the Development 22. One could also classify environmental goods and services along a Issues." Development Policy Review 27 (3): 243­67. different dimension. "Traditional" environmental goods are goods whose Cadarso, M.-A., L.-A. López, N. Gómez, and M.-A. Tobarra. 2010. "CO2 main purpose is to treat a specific environmental problem (for example, Emissions of International Freight Transport and Offshoring: Meas- wastewater treatment equipment). A second, broader type of goods urement and Allocation." Ecological Economics 69 (8): 1682­94. relates to the so-called environmentally preferable goods. For each of Charnovitz, S. 2003. "Trade and Climate: Potential Conflicts and Synergies." those, there exists a close substitute with a similar use but which is less In Beyond Kyoto: Advancing the International Effort against Climate environmentally friendly. An example of an environmentally preferable Change, 141­70. Arlington, VA: Pew Center on Global Climate Change. good is an energy-efficient washing machine. Another that is more doubt- Cole, M., and R. J. R. Elliott. 2003. "Determining the Trade­Environment ful is the use of biofuel to save on energy and reduce CO2 emissions. Composition Effect: The Role of Capital, Labor and Environmental 23. Four submissions have been made: a group of nine members Regulations." Journal of Environmental Economics and Management (Canada; the EU; Japan; the Republic of Korea; New Zealand; Norway; 46 (3): 363­83. Switzerland; Taiwan, China; and the United States), Saudi Arabia, Japan, Copeland, B., and S. Taylor. 2003. Trade and the Environment: Theory and and the Philippines. Evidence. Princeton, N.J.: Princeton University Press. 24. Frankel (2008), Hufbauer et al. (2009), and Messerlin (2010) are Costantini, V., and M. Mazzanti. 2010. "On the Green Side of Trade the main recent commentators on a trade and climate code. Competitiveness?" Sustainable Development Papers, University of 25. Following the demise of the Kerry-Lieberman bill, a new biparti- Minnesota, St. Paul. san bill (CLEAR [s.2877]) has been proposed by Senator Maria Cantwell Davis, S. J., and K. Caldeira. 2010. "Consumption-Based Accounting for (D-WA) and Senator Susan Collins (R-ME), in which no offsets would be CO2 Emissions." Proceedings of the National Academy of Sciences (early allowed, all permits would be auctioned, and revenues from the tax would edition). http://www.pnas.org/content/early/2010/02/23/0906974107 be distributed to U.S. residents. This unilateral approach would certainly .full.pdf+html. put the United States in a better position to enter multilateral negotiations Dechezleprêtre, A., M. Glachant, and Y. Ménière. 2009. "What Drives the and may get support if American households believe that the political International Transfer of Climate Change Mitigation Technologies? economy process will return to them a fair share of the revenues. Empirical Evidence from Patent Data." Working Paper 14, Grantham 26. Such agreements would require a simple majority in the U.S. Research Institute on Climate Change and the Environment, London. House rather the supermajority needed in the U.S. Senate for a treaty. Demaret, P., and R. Stewardson. 1994. "Border Tax Adjustments under 27. Wheeler and Shome (2010) estimate that India, which is seriously GATT and EC Law: General Implications for Environmental Taxes." considering a goal of 15 percent renewable energy in its power mix by 2020, Journal of World Trade 28 (4): 5­66. could effect the shift from coal-fired plants to renewables at a cost of de Melo, J., and N. A. Mathys. 2010. "Trade and Climate Change: The US$50 billion. They note that India is contemplating this option despite the Challenges Ahead." Discussion Paper 8032. Centre for Economic Pol- absence of any meaningful international pressure to cut emissions and no icy Research, London. guarantees of compensatory financing and argue that the government has Di Maria, C., and S. A. Smulders. 2005. "Trade Pessimists vs. Technology concluded that it should promote clean power anyway to develop an inter- Optimists: Induced Technological Change and the Pollution Havens." nationally competitive supplier industry, to bolster energy independence, Advances in Economic Analysis and Policy 4 (2): article 7. and to help limit climate change because it poses dire threats to India itself. Di Maria, C., and E. van den Werf. 2008. "Carbon Leakage Revisited: Uni- lateral Climate Policy with Directed Technical Change." Environmen- tal and Resource Economics 39 (2): 9­15. Bibliography Dijkstra, B. R., and A. J. Mathew. 2009. "Liberalizing Trade in Environ- mental Goods." Working Paper, University of Nottingham, Notting- Acemoglu D., P. Aghion, L. Bursztyn, and D. Hemous. 2009. "The Envi- ham, United Kingdom. ronment and Directed Technological Change." Working Paper 15451, Ederington, J., A. Levison, and J. Minier. 2005. "Footloose and Pollution National Bureau of Economic Research, Cambridge, MA. Free." Review of Economics and Statistics 87 (1): 92­99. Aichele, R., and G. Felbermayr. 2010. "Kyoto and the Carbon Content of Frankel, J. 2008. "Global Environmental Policy and Global Trade Policy." Trade." Discussion Paper 10-2010, Center for Research on Innovation Discussion Paper 08-14, Harvard Project on International Climate and Services, University of Hohenheim. Agreements, Harvard Kennedy School, Cambridge, MA. 316 Managing Openness ------. 2009. "Addressing the Leakage/Competitiveness Issue in Climate Nordhaus, W. 2008. A Question of Balance. New Haven, CT: Yale Univer- Change Policy Proposals." Trade Forum Working Paper, Brookings sity Press. Institution, Washington, DC. Ordás Criado, C., and J.-M. Grether. Forthcoming. "Convergence in per Grether J.-M., N. A. Mathys, and J. de Melo. 2009. "Scale, Technique and Capita CO2 Emissions: A Robust Distributional Approach." Resource Composition Effects in Manufacturing SO2 Emissions." Environmen- and Energy Economics. tal and Resource Economics 43 (2): 257­74. Ostrom, E. 2009. "A Polycentric Approach for Coping with ------. 2010. "Is Trade Bad for the Environment? Decomposing World- Climate Change." Policy Research Working Paper 5095, World Bank, wide SO2 Emissions." Review of World Economics 145 (4): 713­29. Washington, DC. Grether, J.-M., and J. de Melo. 2004. "Globalization and Dirty Industries: Pan, J., J. Philips, and Y. Cheng. 2008. "China's Balance of Emissions Do Pollution Havens Matter?" In Challenges to Globalization: Analyzing Embodied in Trade: Approaches to Measurement and Allocating the Economics, ed. R. E. Baldwin and A. Winters, 167­208. Chicago: International Responsibility." Oxford Review of Economic Policy 24 University of Chicago Press. (2): 354­76. Grossman, G. M., and A. B. Krueger. 1993. "Environmental Impacts of a Pauwelyn, J. 2007. "U.S. Federal Climate Policy and Competitiveness Con- North American Free Trade Agreement." In The Mexico-US Free Trade cerns: The Limits and Options of International Trade Law." Working Agreement, ed. Peter M. Garber, 13­56. Cambridge, MA: MIT Press. Paper 07-02, Duke University, Durham, NC. Hufbauer, G., S. Charnovitz, and J. Kim. 2009. Global Warming and the Peters, G. P., and E. G. Hertwich. 2008. "CO2 Embodied in International World Trading System. Washington, DC: Peterson Institute for Inter- Trade with Implications for Global Climate Policy." Environmental national Economics. Science and Technology 42 (5): 1401­07. Hufbauer, G., and J. Kim. 2010. "Climate Change and Trade: Searching for Porter, M. E., and C. vander Linde. 1995. "Toward a New Conception of Ways to Avoid a Train Wreck." Paper presented at TAIT (Thinking the Environment-Competitiveness Relationship." Journal of Economic Ahead on International Trade) conference sponsored by the Graduate Perspectives 9 (4): 97­118. Institute Centre for Trade and Economic Integration, the World Bank, Reinaud, J. 2008. "Ex Post Evaluation of the EU ETS: Impacts on the Alu- and the World Trade Organization, Geneva, June 16­18. minium Sector." Presentation at a Conference sponsored by Massa- Hummels, D. 2009. "How Further Trade Liberalization Would Change chusetts Institute of Technology Center for Energy and Environmen- Greenhouse-Gas Emissions from International Freight Transport." tal Policy Research, Washington, DC, January 28. Mimeo, Purdue University, West Lafayette, IN. Schelling, T. 2007. "Climate Change: The Uncertainties, the Certain- Ismer, R., and K. Neuhoff. 2007. "Border Tax Adjustment: A Feasible Way ties and What They Imply about Action." Economist's Voice 4 (3), to Support Stringent Emission Trading." European Journal of Law and article 3. Economics 24 (2): 137­64. Schmalensee, R. 1998. "Greenhouse Policy Architectures and Institutions." Levinson, A. 2009. "Comment." In Brookings Trade Forum 2008/2009, ed. In Economics and Policy Issues in Climate Change, ed. W. D. Nordhaus, Brainard and Sorkin, 28­32. Washington, DC. Brookings Press. 137­58. Washington, DC: Resources for the Future. Managi, S., and S. Kumar. 2009. "Trade-induced Technological Change: Spence, M. 2009. "Climate Change, Mitigation, and Developing Country Analyzing Economic and Environmental Outcomes." Economic Mod- Growth." Working Paper 64, Commission on Growth and Develop- elling 26 (3): 721­32. ment, Washington, DC. Mattoo, A., A. Subramanian, D. van der Mensbrugghe, and Jianwu He. Stern, N. 2006. Stern Review on the Economics of Climate Change. London: 2009a. "Can Climate Change Mitigation De-Industrialize Developing Her Majesty's Treasury. Countries?" Mimeo, World Bank, Washington, DC. ------. 2009. The Global Deal: Climate Change and the Creation of a New ------. 2009b. "Reconciling Climate Change and Trade Policy." Policy Area of Progress and Prosperity. New York: Public Affairs. Research Working Paper 5123, World Bank, Washington, DC. Wheeler, D., and S. Shome. 2010. "Less Smoke, More Mirrors: Where India McKibben, Bill. 2010. Earth: Making a Life on a Tough New Planet. New Really Stands on Solar Power and Other Renewables." Working Paper York: Times Books. 204, Center for Global Development, Washington, DC. Mendelsohn, R., A. Dinar, and L. Williams. 2006. "The Distributional World Bank. 2008. International Trade and Climate Change: Impact of Climate Change on Rich and Poor Countries." Environment Economic, Legal, and Institutional Experiences. Washington, DC: and Development 11: 1­20. World Bank. Messerlin, P. 2010. "Climate Change and Trade: From Mutual Destruction WTO (World Trade Organization) and UNEP (United Nations Environ- to Mutual Support." Policy Research Working Paper 5378, World ment Programme). 2009. Trade and Climate Change: A Report by Bank, Washington, DC. UNEP and the WTO. Geneva: World Trade Organization. Index Figures, notes, and tables are indicated with f, n, and t following the page number. A Association of Southeast Asian Nations Abidin, Mahani Zainal, 209 (ASEAN), 133 Africa. See Middle East and North Africa; Sub-Saharan Africa; Australia specific countries GDP growth trends, 111, 111f aging populations, 292­93, 293t temporary trade barriers imposed Agosin, M. R., 227, 228 by, 67, 68, 70f agriculture automobile industry, 171, 238­39, 245n8, 309 climate change and, 293 Aziz, J., 284n9 export competitiveness and, 113, 114f exports from low-income countries, 102, 103, 107, 108t B price projections, 8, 266­67, 266f balanced-budget rule, 175­76, 180 structural market changes, 263­65, 263f, 265f balance of payments temporary trade barriers and, 71, 79 Chile, 197­98, 197f Ahmed, A., 149 financial crisis impact on, 1 Aizenman, J., 32, 244, 245 India, 234 Aldaz-Carroll, Enrique, 261 Mexico, 172, 172t Alejandro, Carlos Diaz, 85 Baldwin, R., 54n2, 55, 82n3 Algeria, oil exports, 57 Banco de México, 174, 176 Alvarez, R., 227, 228 Banco do Brasil, 91 Angola Banco Estado (Chile), 203 commodity prices and, 254 Banco Santos (Brazil), 91 oil exports, 57 Bandholz, H., 21 Annual Survey of Industry (India), 237 Bangkok Agreement, 133 Antholis, W., 314 Bangladesh, export diversification, 103­4 antidumping duties, 5, 64­65, 68, 71, 81, 82n4 banking sector. See financial sector Antweiler, W., 307, 314n6 Bank Negara Malaysia, 212 apparel Barrett, S., 302, 304 export competitiveness and, 113, 115f Bartsch, Ulrich, 233 exports from low-income countries, 107, 108t Basri, Muhammad Chatib, 217, 218 South-South trade, 121 Beinhocker, E., 245n7 U.S. demand for, 7, 100, 101­2 Bems, R., 280 Argentina Benetrix, Augustin, 21 development strategies, 99 Berkmen, P., 14, 15, 24nn9­10 exports Bernanke, B., 41 competitiveness of, 113, 114­15f Bernard, A. B., 56, 57 subject to TTBs, 75f Bhutan, market access in, 79 FDI in, 185 Bijak, J., 290 financial crisis impact, 13, 191, 191t biofuels, 264­65, 267, 274n6, 274n8 GDP growth trends, 109, 110f Blanchard, O., 14, 15, 54n3, 54n7 growth performance, 158, 159f Blecker, R., 178 market access index for, 79 BNDES (Brazilian Development Bank), 91 temporary trade barriers imposed by, 5, 67, 68, 69f, 81 Bolivia trade restrictiveness index for, 71 development strategies, 99 Asian crisis (1996­98) financial crisis impact, 171 copper prices during, 187 border tax adjustments (BTAs), 307, Latin America and, 158­62, 161f, 162t 308­11, 310f trade impact of, 46, 47t Bosnia, market access in, 79 317 318 Index Botswana Carroll, C., 38n6 export diversification, 138f, 139 Casare, J., 180 market access index for, 79 cellulose, 195 regional trade agreements and, 133 Central Asia. See Europe and Central Asia Bown, Chad P., 63, 65, 82nn11­12, 82n16 Central Bank of Chile, 199, 202, 203 Bravo-Ortega, C., 227, 228 CEPAL (Economic Commission for Latin America and Brazil, 85­96 the Caribbean), 171, 173 biofuel production, 265 Chad, terms-of-trade volatility in, 255 bureaucracy in, 166 Chanda, R., 298n27 Chile's exports to, 194 chemicals, 121 China's trade with, 119, 134 Chiang Mai Initiative Multilateralization (CMIM), 23 demand shocks and, 57 Chile, 183­208 development policy, 93­96 bureaucracy, 166 economic policy, 85­89, 87­88f, 88t commodity markets, 261 exchange-rate policy, 95 development strategy, 183­86, 208 exports economic policy, 183­86, 184t, 207 competitiveness of, 113, 114­15f exchange-rate policy, 203, 204f diversification of, 104 export diversification, 187­95, 187f, 191t, 192­94f, 196­98f subject to TTBs, 73, 74, 75f financial crisis impact, 6­7, 190­200, 191t, 192­94f FDI in, 119, 124, 185 financial sector, 186 financial crisis impact, 3, 89­93, 90t, 92­94f, 190, 191t fiscal policy, 187, 203­5, 206f financial market development, 16 foreign direct investment in, 196­99, 198­99f fiscal policy, 93­94 free-trade agreement with China, 134 GDP growth trends, 108, 109f growth performance, 158, 159f, 186 growth performance, 158, 159f monetary policy, 202­3, 204f, 206f import growth in, 42, 48, 52t, 60, 102 policy responses, 200­205 intensive vs. extensive trade, 56 trade openness, 191, 192f market access index for, 79 trade policy, 165, 185­86, 185­86f migrant labor demand in, 9 China rebalancing trade, 48 agricultural commodities in, 263 temporary trade barriers imposed by, 67, 68, 69f Brazil's trade with, 119 Brazilian Development Bank (BNDES), 91 Chile's exports to, 193, 194, 208 Brenton, P., 229, 314n9 commodities market changes and, 263, 268f Bricongne, J.-C., 56, 59 domestic market development, 22 BRICs. See also specific countries exchange-rate policy, 22­23, 38n5 export exports competitiveness of, 113, 113t competitiveness, 113, 114­15f exports diversification of, 104, 138f, 139 diversification of, 104, 105f growth in, 218 growth in, 103, 103f subject to TTBs, 74, 75f, 81 GDP growth trends, 108­9, 109f financial crisis impact, 3, 127­32, 128­33f import growth, 100 foreign reserves, 32 Broda, C., 48 free-trade agreement with Chile, 185 BTAs. See border tax adjustments GDP growth trends, 108, 109f Buch, C., 150 global imbalances and, 2, 2f, 36 Bulgaria, exports from, 101 global production networks and, 8­9, 276­78, 276­79f, bullwhip effect, 8­9, 281­83 281­82t, 283f bureaucracy household registration system, 298n23 Latin America and the Caribbean, 166 import growth, 7, 42, 48, 53t, 102 Malaysia, 223 India's trade with, 119, 120­21 Burkina Faso, financial crisis impact on, 247 labor migration and, 296 market access index for, 79 C as model for Brazil, 94­95 Caballero, R., 32, 33 rebalancing trade, 48 Cadarso, M.-A., 314n8 savings and investment rate in, 29f, 36, 36f, 38n9 Cadot, O., 228, 230n6 South-South trade by, 119­34 Caixa Economica Federal (Brazil), 91 FDI flows, 123­25, 124­25t, 126f Calderón, C., 179 future projections for, 122­23, 123f Canada institutional dimensions, 132­34 GDP growth trends, 111, 111f post-financial crisis, 126­27, 126­27t migration trends, 288 prior to financial crisis, 120­22, 120t, 121f, 122t temporary trade barriers imposed by, 67, 68, 70f Sub-Saharan Africa exports to, 253 trade openness, 176, 176t temporary trade barriers and, 5, 66­67, 67t Cantwell, Maria, 315n25 U.S. trade imbalance with, 1, 2, 27 capital accounts China Development Bank, 134 Chile, 195­99 Chinn, M. D., 136, 244, 245 flight to safety and, 3, 3f, 33­34, 35f civil service, 166 housing boom and, 21 Clean Development Mechanism, 304, 311 India, 239­44, 243f climate change Mexico, 172, 172t agricultural impact of, 298n22 Capital Market Masterplan (Malaysia), 210 defined, 298n21 carbon leakage, 306­11 equity concerns, 304 Carrère, C., 228 greenhouse gas mitigation, 267, 305­8 Index 319 labor migration and, 293­94 FDI flows into, 124 trade policy and, 9, 302­5, 303f financial crisis impact on, 1 UN Framework Convention on Climate Change, 301, 304 GDP growth trends, 102, 108­11, 109­11f WTO and, 308­13 global imbalances and, 33­37, 34­35f Cline, W., 38n5 greenhouse gas emissions and, 306 Clostermann, J., 21 protectionism and, 73­81, 75­76f, 77­78t CMIM (Chiang Mai Initiative Multilateralization), 23 savings and investment rate, 35 Coalition of Service Industries, 298n28 trade barriers of, 63­83 Collins, Susan, 315n25 development strategies Colombia Argentina, 99 export diversification, 138­39f, 139 Bolivia, 99 FDI in, 185 Brazil, 93­96 financial crisis impact, 191, 191t Chile, 183­86, 208 growth performance, 158, 159f, 160 India, 16 Comin, D., 150 Malaysia, 213­15 commercial banks, 24n21 Mexico, 176­79, 178f commodities, 261­74. See also specific commodities Di Giovanni, J., 135, 227 financial crisis impact on, 58­59, 58f, 61f Dijkstra, B. R., 312 India exports of, 218 Di Maria, C., 308, 312 Indonesia, 221­29, 222­26f disequilibrium approach, 30­31 literature review, 267­69 documentation costs, 141 Malaysian exports of, 214 Doepke, J., 150 policy recommendations, 8­9, 269­73, 270t domestic markets price projections, 265­67 China, 101 South-South trade, 126­27 India, 236­39 structural market changes, 262­65 industrial policy and, 5 trade policy agenda for, 8­9, 9f Dominican Republic, financial crisis impact on, 171 volatility of prices, 143n2, 253, 262­65, 270­71 downstream trade costs, 275, 278 compositional effect, 280­81, 305 dual-use technologies, 312 Compustat, 150 Dutch disease, 8, 228, 268 Condon, T., 222 Constantini, V., 308 E construction industry, 169 EAP (economically active population), 290­91, 290­91t consumer durables, 4, 20, 92, 175 early-warning indicators, 13 Copeland, B., 307, 314n6 East Asia and the Pacific. See also specific countries copper, 187, 187f, 193­94, 208 apparel exports, 108 Corbo, Vittorio, 157 exports Corporación de Fomento de la Producción (Corfo, Chile), 205 competitiveness of, 113, 113t Corporación Nacional del Cobre (Codelco, Chile), 205 growth in, 102 Costa Rica Easterly, W. R., 95, 135 free-trade agreement with China, 134 Eaton, J., 55, 112, 116 migrant labor demand in, 9 economically active population (EAP), 290­91, 290­91t migration trends, 288 Economic and Social Stabilization Fund (Chile), 184, 205 costs of exporting, 140­42, 141­42f Economic Commission for Latin America and the Caribbean Côte d'Ivoire, migration trends, 288 (CEPAL), 171, 173 countercyclical policies, 93 economic policy. See also fiscal policy; monetary policy countervailing duties, 5, 64­65, 81 Brazil, 85­89, 87­88f, 88t Craine, R., 21 Chile, 183­86, 184t, 207 credit default swaps, 199 India, 240­44 current account imbalances, 27, 28f Ecuador, financial crisis impact on, 190, 191t customs costs, 141 Ederington, J., 314n5 Edwards-Jones, G., 314n9 D Egypt, GDP growth trends, 110f, 111 dark matter, 38n4 Eichengreen, Barry, 3, 5, 13, 17, 280 Das, M., 14, 15 electronics, 108, 108t, 113, 115f, 171 Debaere, P., 141 emerging market economies, 13­25. See also growing Dechezleprêtre, A., 312­13 emerging markets (GEMs) De Ferranti, D., 269, 272 financial crisis impact on, 13­17, 14­17f defined-contribution pension systems, 165 global imbalances, role of, 20­22 demand shocks, 55, 56­58 import growth in, 48, 51f, 52f Demaret, P., 310 policy response, 22­23 De Melo, Jaime, 230n6, 301, 307, 308 trade collapse, analysis of, 17­20, 18­19f demographic trends, 288, 288t, 289t, 292­93, 293t employment. See also unemployment Dennis, A., 139, 141 financial crisis impact on, 1 deregulation informal employment, 166, 181n1 Brazil, 87 Energy Department (U.S.), 266 Latin America and the Caribbean, 165 Engel, C., 30, 284n8 Malaysia, 149, 221, 223 environmental goods and services, 311­13. developing countries See also climate change exports equilibrium approach, 2, 31­33 competitiveness, 112­13t, 112­15, 114­15f ERS (exchange-rate stability), 244­45 diversification of, 6 Estonia, financial crisis impact on, 13 growth in, 5, 107­17 European Central Bank, 23 320 Index Europe and Central Asia. See also specific countries financial crisis (2008­09). See also specific countries for impact exports protectionism and, 3­5, 63­83 competitiveness of, 113, 113t roots of, 1­3 growth in, 102 trade and, 3­5, 55­62 European Union financial sector. See also specific institutions biofuel production, 265 Chile, 186 Chile's exports to, 193 in emerging markets, 16 demand shocks and, 57 Indonesia, 227­28 exports subject to TTBs, 73, 74f Latin America and the Caribbean, 165, 166 free-trade agreement with Chile, 185 openness of, 136, 166, 166t global imbalances and, 2, 2f Financial Sector Masterplan (Malaysia), 210 imports, 60, 100, 101 fiscal policy intensive vs. extensive trade, 56 Brazil, 93­94 migration trends, 288 Chile, 187, 203­5, 206f renewable fuels mandate, 274n8 Latin America and the Caribbean, 162 savings and investment rate in, 27, 29f Mexico, 173­74, 175t, 176­79, 180 Sub-Saharan Africa exports to, 253 Fiscal Responsibility Law of 2006 (Chile), 184, 205 temporary trade barriers imposed by, 68, 70f, 79 Fischer, S., 86­87 trade restrictiveness index for, 71 Fondo Solidario (Chile), 208 U.S. trade deficit with, 27 Food and Agriculture Organization (FAO), 267 Evenett, S. J., 5, 17 Forbes, K., 33 exchange-rate policy foreign direct investment (FDI) Brazil, 95 in Chile, 183, 185, 196­97, 198f Chile, 202, 203, 204f from China, 119, 123­25, 124­25t, 126f China, 22­23, 38n5 commodities markets and, 272 financial crisis impact on, 15 to developing countries, 124 India, 217, 242 housing boom and, 21 Latin America and the Caribbean, 163­64 in India, 236 Malaysia, 216 in Malaysia, 214, 222 Mexico, 174, 176­81 in Mexico, 172 reserves and, 38n13 in Sub-Saharan Africa, 247 U.S., 2, 30­31 foreign-exchange derivatives, 90­91 exchange-rate stability (ERS), 244­45 Forum on China-Africa Cooperation, 133 export competitiveness, 112­13t, 112­15, 114­15f France export credit agencies, 19 aging population in, 292 export diversification demand shocks and, 58 Chile, 193 GDP growth trends, 111, 111f Indonesia, 221­29, 222­26f intensive margin trade, 55 in middle-income countries, 103­4, 105f migration trends, 288 pattern changes in, 104­5, 105f Franco, Gustavo H. B., 85 policy creation and, 6, 6f, 139­42, 141­42f Francois, J., 82n9 productivity impact of, 252­53f, 252­54 free-trade agreements (FTAs), 133­34, 185 Sub-Saharan Africa, 248­50, 249­50f NAFTA, 174, 196 volatility and, 135­44, 137t, 138­40f Freund, Caroline, 20, 41, 63 exports fuel subsidies, 270, 271, 271f demand and supply pattern Fundo Garantido de Créditos (FGC, Brazil), 91 changes, 102­5, 103f developing countries, 107­17 G India, 236­39, 239t G-20 Mexico, 171­72, 172f, 175f global imbalances and, 2 productivity impact of, 145­48, 146f, 148f temporary trade barriers and, 17, 64­67, 65­66f, 67t, 69­70f, 71t specialization in low-income countries, 103 GDP growth temporary trade barriers and, 73­79 Brazil, 87­89, 88f, 88t volatility impact of, 149­51, 151f Chile, 187­90, 188­90f, 188t, 200t extensive margins financial crisis impact on, 13­17 Indonesia exports, 227f import demand and, 7 Malaysian exports, 226, 227f trends in, 108­11, 109­11f trade, 48­53, 52­53t GEMs. See growing emerging markets external demand shock, 135, 190­95, 191t, 192­94f General Agreement on Trade in Services (GATS), 229, 294­95 extractive industries, 108, 108t. See also minerals; oil Germany aging population in, 292 F demand shocks and, 58 factor endowment effect, 305 export diversification, 150 Fane, G., 222 GDP growth trends, 111, 111f Farhi, E., 32, 33 migration trends, 288 Farrell, D., 245n7 Giugale, M., 177 Faruqee, H., 14, 15 Glachant, M., 312­13 FDI. See foreign direct investment Global Economic Prospects (World Bank), 287 Federal Reserve (U.S.) global financial crisis. See financial crisis (2008­09) bilateral currency swaps, 23 Global Forum on Migration and Development, 291, 297 swap facility with Mexico, 172, 174 global imbalances, 27­39 Feenstra, R. C., 141 developing countries and, 33­37, 34­35f FGC (Fundo Garantido de Créditos, Brazil), 91 disequilibrium approach, 30­31 Index 321 emerging market economies and, 20­22 I equilibrium approach, 31­33 Iacovone, L., 19 financial crisis and, 1­3, 2­3f Ibarra, C., 178 future of, 2, 33­37 ILO (International Labour Organization), 294, 297n9 growth promotion and, 37 Imbs, Jean, 104 nature of, 28­30f, 29­33 IMF. See International Monetary Fund sustainability of global trade patterns, 111­16 imports in trade, 42­47, 42f demand and supply pattern changes, 102­5, 103f, 115­16 global production networks (GPNs), 275­85 developing countries' demand for, 102 China's processing regime, 276­78, 276­79f, 281­82t, 283f growth in, 3­4, 4f, 7, 63­64, 64f expansion of, 275­76 import substitution policy, 145 intra-GPN trade, 280­83 impossible trinity, 242, 243f oil prices and, 279­80, 279f income inequality, 207 policy recommendations, 8­9 India, 233­46 protectionism and, 1 agricultural commodities in, 263­64 global savings glut, 32 capital account liberalization, 239­40, 241f Global Task Force on Public Goods, 301 capital flows, 240­44, 243f Global Trade Alert, 17 China's trade with, 119, 120­21 Global Trade Analysis Project, 308 domestic market, 236­39, 237f global trade imbalances (GTIs), 41, 42f, 46, 46t. See also economic growth in, 234­35f, 234­36 global imbalances economic policy, 240­44 Gourinchas, P., 32, 33 exchange-rate policy, 217, 245n9 governance, 5, 214­15 exports, 236­39, 237f, 239t GPNs. See global production networks competitiveness of, 113, 114­15f Gramm-Bliley-Leach Act of 1999, 21 diversification of, 104 gravity model of trade, 99, 100, 102, 105n3, 111 subject to TTBs, 73, 74, 75f, 77 Great Depression, 17­18 FDI from China, 119, 124 greenhouse gases, 267, 305­8 financial crisis impact, 3, 13 Grether, J.-M., 307, 308 financial market development, 16 Grossman, G., 247 GDP growth trends, 108, 109f Group of 20 growth model of, 236­39 global imbalances and, 2 import growth in, 42, 48, 53t, 102, 130, 132 temporary trade barriers and, 17, 64­67, 65­66f, 67t, 69­70f, 71t rebalancing trade, 48 growing emerging markets (GEMs). See also emerging regional trade agreements and, 133 market economies temporary trade barriers imposed by, 5, 67, 68, 69f, 70, 81 import growth in, 48, 51f, 52­53, 52f Indonesia, 217­31 GTIs. See global trade imbalances commodity imports, 60­61, 61f Guatemala, exports to U.S., 101 demand shocks and, 57 Guinea, terms-of-trade volatility in, 255 exchange-rate policy, 228 Gullapalli, R., 17 exports Gupta, Abhijit Sen, 233 diversification of, 221­29, 222­26f facilitation policies, 218­21, 219t, 220f H growth in, 218 Haddad, Mona, 1, 7, 55, 135, 145 marketing of, 229 Hanson, Gordon, 99, 107, 275 subject to TTBs, 73, 74, 75f, 77 Harrison, Ann, 55, 248 financial crisis impact, 3, 217­18 Hausman, Catherine, 55 financial sector, 227­28 Hausmann, R., 30­31, 180­81, 228, 247­48 GDP growth trends, 109, 110f Helpman, E., 141, 247 import growth in, 42, 48, 52, 53t, 60 Herfindahl-Hirschman Index (HHI), 104, 106n9, intensive vs. extensive trade, 56 150­51, 223, 252­53 manufactured imports, 60­61, 61f HFCs (hydrofluorocarbons), 304 market access index for, 79 high-income countries rebalancing trade, 48 exports research and development, 229 competitiveness of, 112, 112t services sector, 229 growth of, 103, 103f temporary trade barriers imposed by, 5, 67, 68, 69f, 81 subject to TTBs, 73­74, 74f trade policy, 228 extensive margin import growth, 52­53 industrial policy, 5, 207, 236 financial crisis impact, 60 "infant industry" tariffs, 143 labor migration and, 297 inflation Hodrick, R. J., 112, 115­16 Brazil, 86­87, 89, 95 Hong, C., 229 Chile, 184, 184t, 207 Hong Kong SAR, China India, 235 Asian financial crisis and, 46, 47t Latin America and the Caribbean, 164, 164t growth model of, 145 Malaysia, 221 Horenstein, M. D., 63 Mexico, 176 housing boom, 21 inflation-indexed financial contracts, 186 human capital, 147, 238, 272 informal employment, 166, 181n1 Hummels, D., 141, 279, 308 infrastructure investments, 173, 177, 178f, 179, 181n6, 214, 228 Humphreys, M., 271­72, 274n11 intensive margins Hwang, J., 180­81, 247­48 Indonesia exports, 227f hydrofluorocarbons (HFCs), 304 Malaysian exports, 226, 227f hyperinflation, 86, 89, 95 trade, 48­53, 52­53t 322 Index intergenerational equity, 302, 304 reserve ratios, 24n9 International Energy Agency, 266 temporary trade barriers imposed by, 68, 70f International Labour Organization Kortum, S., 112, 116 (ILO), 294, 297n9 Kraay, A. C., 135 international migration. See labor migration Krugman, P., 247 International Monetary Fund (IMF) Kumar, S., 313 bilateral currency swaps, 23 Kuwait, migration trends, 288 on capital account liberalization, 241 Kyoto Protocol, 301, 304­5, 311 on China's savings and investment rate, 36 Kyrgyz Republic, market access in, 79 credit line for Mexico, 172 on exchange-rate policy, 24n5, 245n9 L on Mexico's balanced-budget rule, 180 labor markets on stimulus packages, 181n7 Chile, 185, 208 intra-GPN trade, 280­83 labor force projections, 290­91, 290­91t inventory-sales ratio, 148, 150 Latin America and the Caribbean, 166 investment climate, 147, 215 labor migration, 287­99 investment rates, 27­28 aging populations and, 292­93, 293t Iraq, oil exports, 57 climate change and, 293­94 Ireland, export diversification, 138f, 139 drivers of, 289­90 iron exports, 102, 103 economic implications of, 295­96 Irwin, D., 17 future supply of, 290­91t, 290­95 Ismer, R., 311 international management of, 294­95 Italy from Mexico, 172­73 aging population in, 292 policy recommendations, 9 demand shocks and, 58 recommendations, 296­97 Ito, H., 136, 244, 245 to Singapore, 298n32 trends in, 288­89, 289t, 291 J labor productivity Jacks, D., 281 export diversification's impact on, 147, 148, 252­53f, 252­54 Jansen, M., 143n3 exports' effect on, 251­52, 251­52f Japan Malaysia, 152 aging population in, 293 Lafourcade, O., 177 agricultural commodities in, 263 Lane, P., 245 Chile's exports to, 193, 194 Latin America and the Caribbean, 157­67. See also specific countries demand shocks and, 58 Asian crisis impact on, 158­62, 161f, 162t exports China's trade with, 7, 121, 133­34 competitiveness of, 113, 114­15f exchange-rate policy, 163­64 growth in, 130 exports subject to TTBs, 74f competitiveness of, 113, 113t financial crisis impact, 3, 20 growth in, 102 GDP growth trends, 111, 111f market shifts in, 101 import growth in, 130 financial crisis impact on, 158­62, 161f, 162t savings and investment rate in, 28, 29f financial markets, 166 U.S. trade deficit with, 27 fiscal policy, 162­63 Jeanne, O., 32, 38n6 growth performance, 158­60f Jensen, M. F., 314n9 migration trends, 288 Johnson, R., 280 monetary policy, 162, 164 joint ventures, 136 policy implications for, 162­66, 163­64t trade openness, 165­66, 165­66t K U.S. imports from, 100 Kaldor, N., 268 Latin American Reserve Fund (FLAR), 23 Kaminsky, G., 177 Latvia Kazakhstan financial crisis impact, 13 exports subject to TTBs, 75f, 79 reserve ratios, 24n9 market access index for, 79 learning-by-exporting, 6 Kee, Hiau Looi, 19, 63, 71, 79, 82n12, 141 Lebanon, market access in, 79 Keilman, N., 297n4 Lederman, D., 269 Kenya, export diversification, 138­39f, 139 Lee, J., 32 Klenow, P., 141 Lesotho, regional trade agreements by, 133 Knetter, M. M., 82n10 Levchenko, A. A., 55, 135, 227 Korea, Republic of leverage ratios, 17 agricultural commodities in, 263 Levinson, A., 314n5, 314n7 Asian financial crisis and, 46, 47t Lewis, Arthur, 95 Chile's exports to, 194 Lewis, L. T., 55 demand shocks and, 58 Li, X., 284n9 exports Libya, migration trends in, 9, 288 growth in, 130 life-cycle theory of consumption, 292 subject to TTBs, 73, 74f Lim, Jamus, 135 foreign exchange reserves, 15, 16 liquefied natural gas (LNG), 224 GDP growth trends, 109, 110f liquidity, 14, 212 growth model of, 145 Lithuania, financial crisis impact on, 13 import growth, 130, 132 Loayza, N., 149 labor migration and, 296 logistics costs, 228, 272 Index 323 lower-middle income countries Melitz, M., 141 current account deficits, 14, 14f Mendoza, E., 33, 38n6 exports Ménière, Y., 312­13 competitiveness of, 112, 112t Messerlin, P., 315n20 diversification of, 107­8, 108t metals. See also specific metals low-income countries export competitiveness and, 113, 115f current account deficits, 14, 14f price projections, 266f, 267 exports South-South trade, 121 competitiveness of, 112, 112t structural market changes, 263­64f, 268f diversification of, 6, 100, 104­5, 105f, 107, 108t Mexico, 169­82 to emerging markets, 48 balance of payments, 172, 172t growth in, 103, 103f capital account shock, 172, 172t specialization in, 7, 99, 103 development strategies, 176­79, 178f subject to TTBs, 76f exchange-rate policy, 174, 176­81 extensive margin import growth, 52­53 exports, 171­72, 172f, 175f financial crisis impact, 60 subject to TTBs, 74, 75f GDP growth and import demand in, 7 to U.S., 101 global imbalances and, 36 financial crisis impact, 169­71, 170t, 171f, 174­76, 191, 191t imports from low-income countries, 102 fiscal policy, 173­74, 175t, 176­79, 180 Sub-Saharan Africa exports to, 250 free-trade agreement with Chile, 185 "Lucas paradox," 38n2 GDP growth trends, 109, 110f Lula da Silva, Luis Inacio, 87, 95, 134 growth performance, 158, 159f, 160 migration and remittances, 172­73 M monetary policy, 174, 176, 179­81 Ma, Alyson, 275 policy implications, 176­79 machinery, 108, 108t, 113, 115f population trends, 288 Malawi, export diversification, 138f, 139 private investment, 170, 172 Malaysia, 209­16 public investment, 176­79 Asian financial crisis and, 46, 47t tax reforms, 179 commodity markets, 261 temporary trade barriers imposed by, 68, 69f demand shocks and, 58 Tequila crisis (1995), 169 development strategies, 213­15 trade openness, 176, 176t economy prior to financial crisis, 209­10, 210t trade policy, 165 exchange-rate policy, 216 unemployment, 170 exports Middle East and North Africa. See also specific countries growth in, 218 exports productivity impact of, 145­48, 146f, 148f competitiveness of, 113, 113t subject to TTBs, 74, 75f growth in, 102 volatility impact of, 149­51, 151f market shifts in, 101 financial crisis impact, 210­12, 211­12f migration trends, 288 GDP growth trends, 110f, 111 oil exports, 108 growth strategies, 6, 215­16 middle-income countries migrant labor demand in, 9 apparel exports from, 103 migration trends, 289 exports recovery strategies, 212­13 competitiveness of, 112, 112t textile exports from, 102 diversification of, 48, 99, 100, 103­5, 105f, 108, 108t Malaysian Institute of Economic Research, 211 growth in, 103, 103f Mali, financial crisis impact on, 247 GDP growth and import demand in, 7 Maloney, W. F., 269 import growth, 52­53, 99, 102 Managi, S., 313 labor migration and, 297 manufactures migrant labor demand in, 9 financial crisis impact on, 58­59, 58f, 61f Sub-Saharan Africa exports to, 250 Indonesia exports of, 226f migration. See labor migration Malaysia, 147, 211, 214 Milesi-Ferretti, G. M., 54n3, 54n7, 245 Mexico, 169, 175 minerals supply shocks in, 57 exports from low-income countries, 103 temporary trade barriers and, 71 price projections, 266f, 267 market access overall trade restrictiveness index South-South trade, 122, 127 (MA-OTRI), 79­81, 80t, 82n1 structural market changes, 263, 263f market diversification, 136, 151, 215, 257. See also Minier, J., 314n5 export diversification Mirza, D., 279 market-entry costs, 141 Mitchell, Donald, 261 marketing of exports, 229 Mitra, D., 268­69 Martin, Philip, 287 Mody, A., 13, 24n10 Martin, V., 21 Moghadam, R., 14, 15 Martin, W., 268­69 monetary policy Mason-Jones, R., 284n11 Chile, 202­3, 204f, 206f Mataloni, R., 275 in Great Depression, 18 Mathew, A. J., 312 India, 242­44 Mathys, Nicole A., 301, 307, 308 Latin America and the Caribbean, 162, 164 Mazzanti, M., 308 Mexico, 174, 176, 179­81 Meirelles, Henrique, 14 Montreal Protocol, 304 Meissner, C., 281 Mora, J., 19 324 Index Mostashari, S., 141 Peru Mozambique FDI in, 185 commodity markets, 261 financial crisis impact, 190, 191t terms-of-trade volatility, 255 free-trade agreement with China, 134 multiregional general-equilibrium (MR-GE), 307 growth performance, 158, 159f Munro, Laura, 135 Petrobras (Brazil), 134 Petróleos Mexicanos (PEMEX), 171, 173 N Philippines NAFTA (North American Free Trade Agreement), 174, 196 Asian financial crisis and, 46, 47t Namibia, regional trade agreements by, 133 export growth, 218 National Rural Employment Guarantee Schemes (India), 235 population trends, 288 National Survey of Employment and Occupation Philippon, T., 150 (Mexico), 173 Poland Neagu, C., 19, 71, 79, 82n12 exports, 101 Neuhoff, K., 311 competitiveness of, 113, 114­15f Newfarmer, R., 229 financial crisis impact, 13 New Zealand, financial crisis impact on, 191 GDP growth trends, 109, 110f Nguyen, Ha, 27 pollution heaven effect, 305, 306 Nguyen, V. H., 177 population trends, 288­89t Nicaragua, export diversification, 138­39f, 139 port costs, 141 Nicita, A., 19, 71, 79, 82n12 Powers, W., 19 Niels, G., 82n9 Prebisch, Raul, 267 Nigeria Prescott, E. C., 112, 115­16 commodity prices and, 254 price indexation, 186 export diversification, 138f, 139 private investment North American Free Trade Agreement (NAFTA), 174, 196 Chile, 183, 202, 207­8 Norway Malaysia, 152 exports Mexico, 170, 172 diversification of, 138f, 139 productivity and, 148 subject to TTBs, 79 privatization, 87, 89, 165 Novy, D., 281 product diversification. See export diversification Nurridzki, N., 228 productivity export diversification's impact on, 252­53f, 252­54 O exports' effect on, 145­48, 146f, 148f, 251­52, 251­52f Obstfeld, M., 14, 41 Productivity and Investment Climate Survey oil (Malaysia & World Bank), 148, 149­50 exports from low-income countries, 102 protectionism. See also specific protectionist measures global production networks and, 279­80, 279f developing-country exporters and, 73­81, 75­76f, 77­78t Indonesia exports of, 223f domestic industries and, 64­72 Mexico's exports of, 175 financial crisis and, 3­5, 8f price projections, 8, 265­66, 266f global production networks and, 1 South-South trade, 122, 127 Prusa, T. J., 82n10 structural market changes, 263­64f public investment Sub-Saharan Africa exports of, 254 Brazil, 87­88 U.S. demand for, 7, 100, 101­2 Mexico, 170, 176­79 Olarreaga, M, 230n6 O'Rourke, Kevin, 21, 280 Q Ostrom, E., 314 Qatar, migration trends, 288 overall trade restrictiveness index (OTRI), 71, 72t, 82n1. See also Quadrini, V., 33, 38n6 market access overall trade restrictiveness index (MA-OTRI) R Raddatz, C. E., 135 P Rahardja, Sjamsu, 217, 218 Pakistan Ranciere, R., 32 exports R&D. See research and development diversification of, 103­4 RBI. See Reserve Bank of India subject to TTBs, 76f, 77 regional trade agreements, 133­34, 174, 196 free-trade agreement with China, 133 Reinhart, C., 21, 177 GDP growth trends, 110f, 111 remittances, 172­73, 296, 298n31 Panama renewable fuels, 274n8 exports subject to TTBs, 79 replacement migration, 292­93, 293t market access index for, 79 research and development (R&D), 147, 150, 229, 303 Panizza, U., 17 Reserve Bank of India (RBI), 235, 242, 244 Papanek, G., 218 reserves Paraguay, financial crisis impact on, 171 Brazil, 89­90, 91­92 Pastor, J., 177 Chile, 184 Patunru, A. A., 228 China, 32 Pauwelyn, J., 309, 310, 311, 314n16 exchange-rate policy and, 38n13 peak oil, 279­80, 279f, 284n5 liquidity crisis and, 14­15 PEMEX (Petróleos Mexicanos), 171, 173 Malaysia, 210 pension reforms, 165 Sub-Saharan Africa, 247 Pension Reserve Fund (Chile), 184, 205 trade finance and, 4 Pertamina (Malaysia), 221 Rios-Rull, J., 33, 38n6 Index 325 risk aversion shock, 91 regional trade agreements and, 133 Rodríguez-Clare, A., 248 temporary trade barriers imposed by, 68, 69f Rodrik, D., 5, 95, 180­81, 227, 228, 229, 247­48, 269 South Asia. See also specific countries Rogers, J., 30 apparel exports, 108 Rogoff, K., 21, 41 climate change and, 293 Ros, Jaime, 169, 180 exports Rose, A., 13, 229 competitiveness of, 113, 113t Rubin, J., 279 growth in, 102 Rubinstein, Y., 141 market shifts in, 101 Russian Federation Southern Africa Customs Union (SACU), 133 exports South-South trade competitiveness of, 113, 114­15f by China, 119­34 diversification of, 104 FDI flows, 123­25, 124­25t, 126f oil exports, 57 future projections for, 122­23, 123f subject to TTBs, 73, 74, 76f, 77, 79 institutional dimensions, 132­34 GDP growth trends, 108, 109f post-financial crisis, 126­27, 126­27t import growth in, 102 prior to financial crisis, 120­22, 120t, 121f, 122t market access index for, 79 export diversification and, 6 migration trends, 288 growth of, 1 trade restrictiveness index for, 71 low-income countries and, 60 Rwanda, financial crisis impact on, 247 migrant labor demand and, 9 temporary trade barriers and, 5, 7, 8f, 64, 68, 73­74, 81, 121 Spain, GDP growth trends, 111, 111f S Spiegel, M., 13 Saborowski, Christian, 135 Sri Lanka Sachs, J. D., 271­72, 274n11 Asian financial crisis and, 46, 47t SACU (Southern Africa Customs Union), 133 market access index for, 79 safeguards for domestic industries, 5, 64­65, 81 Stahn, K., 150 Sahay, R., 86­87 "staples theory," 269 Saudi Arabia state-owned enterprises (SOEs) migration trends, 288 Brazil, 91 oil exports, 57 dividend distributions from, 37, 38n12 savings and investment rate export diversification and, 221­22 Brazil, 87­88 steel exports, 102, 103 China, 36, 36f, 38n9 Stewardson, R., 310 developing countries, 35 Stiglitz, J. E., 271­72, 274n11 equilibrium approach and, 32 stimulus packages global imbalances and, 27­28 global imbalances and, 37 U.S., 36, 36f India, 235 scale effect, 305 Malaysia, 212, 220 Schmidt-Hebbel, Klaus, 157 Mexico, 173, 175t, 181n7 Schott, P., 56 protectionism and, 18 Schydlowsky, D., 218 Strauss-Kahn, V., 228 Securities and Exchange Commission, 21 Sturzenegger, F., 30­31 Seitz, F., 21 Suardi, S., 149 Servén, Luis, 27, 179 Sub-Saharan Africa, 247­58. See also specific countries services sector agricultural commodities in, 264 Chile, 195 China's trade with, 7 commodities markets and, 273 climate change and, 293 exports, 106n8 exports Indonesia, 229 competitiveness of, 113, 113t Malaysia, 215 diversification of, 104, 248­50, 249­50f Mexico, 169 growth impact of, 250­54, 251­52f Shambaugh, J. C., 14, 244 to emerging markets, 48 Shepherd, Ben, 1, 121, 135, 139, 141, 142 growth in, 102 Shome, S., 315n27 market shifts in, 101 Singapore FDI from China, 119 Asian financial crisis and, 46, 47t financial crisis impact, 254­56, 254t, 255­56f export growth, 218 growth strategies, 6 growth model of, 145 migration trends, 288 labor migration to, 298n32 policy implications, 256­57 Singer, Hans, 267 subsidies, 270, 271, 271f Slaughter, M., 275 supply shocks, 55, 56­58 Smulders, S. A., 312 Swaziland, regional trade agreements by, 133 social safety nets, 37, 38n10, 186, 208 SOEs. See state-owned enterprises Songwe, Vera, 247 T South Africa Taglioni, D., 54n2 China's trade with, 121 Taiwan, China exports subject to TTBs, 73, 74, 76f demand shocks and, 58 FDI from China, 124 exports subject to TTBs, 74f financial crisis impact, 191 growth model of, 145 GDP growth trends, 109, 110f import growth in, 132 migrant labor demand in, 9 Tajikistan, export costs in, 142 326 Index Tal, B., 279 Turkey Talvi, E., 177 exports Tanzania, market access in, 79 competitiveness of, 113, 114­15f tariffs subject to TTBs, 76f on Chile's imports, 185 GDP growth trends, 109, 110f "infant industry" tariffs, 143 temporary trade barriers imposed by, 5, 67, 68, 69f, Latin America and the Caribbean, 165 70, 79, 81, 82n13 trade impact of, 71 trade restrictiveness index for, 71 trade liberalization and, 66 Turkmenistan, exports subject to TTBs, 77 Taylor, A., 14 TWPs (temporary worker programs), 295­96 Taylor, S., 307, 314n6 technique effect, 305 U technology transfer, 6, 9, 147, 308 Uganda, commodity markets in, 261 telecommunications technology, 147 Ukraine temporary trade barriers (TTBs). See also specific barriers exports subject to TTBs, 73, 74, 76f developing countries and, 73­79, 75­76f, 77­78t market access index for, 79 domestic industries and, 64­65, 65f unemployment financial crisis impact on, 65­71 Chile, 185, 200, 208 high-income economies and, 74f labor migration and, 295 policy implications, 81­82 Latin America and the Caribbean, 166 products affected by, 65­68, 66f, 67t Malaysia, 211 South-South trade and, 5, 7, 8f Mexico, 170, 173, 181n1 temporary worker programs (TWPs), 295­96 United Arab Emirates, migration trends, 288 Tequila crisis (1995), 169 United Kingdom term liquidity facility, 203 aging population in, 292 terms-of-trade volatility, 135, 143n3, 200, 234, 255 demand shocks and, 58 Tesar, L. L., 55 export competitiveness, 113, 114­15f textiles, 107, 108t, 121 GDP growth trends, 111, 111f TFP. See total factor productivity temporary trade barriers imposed by, 70 Thailand United Nations Asian financial crisis and, 46, 47t on aging populations, 292 exports Economic Commission for Latin America and growth in, 218 the Caribbean (CEPAL), 171, 173 subject to TTBs, 73, 74, 76f, 77 Framework Convention on Climate Change, 301, 304 GDP growth trends, 109, 110f United States import growth in, 132 biofuel production, 265 migration trends, 9, 289 capital flows to, 3, 33­34, 35f Tong, H., 17 Chile's exports to, 193, 194 total factor productivity (TFP) China trade imbalance with, 1, 2, 130 Chile, 190 consumption drop, impact of, 100­101 commodity markets and, 269 current account deficit, 27 export-oriented growth and, 147 demand shocks and, 57 India, 234, 236­37 exports openness and, 87, 87f competitiveness of, 113, 114­15f Towill, D., 284n11 growth in, 132 trade. See also temporary trade barriers (TTBs) subject to TTBs, 73, 74f Asian crisis (1996­98) impact on, 46, 47t extensive margin import growth, 48 barriers, 63­83 free-trade agreement with Chile, 185 changing dynamics of, 7­8, 42­46, 43­45t, 99­108, 108­9t GDP growth trends, 111, 111f demand shocks, 56­58 global imbalances and, 2, 2f, 36 financial crisis impact on, 3­5, 4f, 55­62, 56­57f imports, 7, 100­102 by income group, 59­60, 60f commodities vs. manufactures, 60­61, 61f by product types, 58­59, 58­59f, 61f extensive margin growth, 48 global imbalances in, 42­47, 42f from low-income countries, 60, 100 greenhouse gas mitigation and, 305­8, 306f intensive margin trade, 55 import growth, 63, 64f migration trends, 288 intensive vs. extensive margins, 48­53, 52­53t, 55­56 savings and investment rate in, 27, 29f, 36, 36f policy agenda, 8­9 Sub-Saharan Africa exports to, 250, 253 rebalancing, 41­54, 49­50t temporary trade barriers imposed by, 67, 70, 70f supply shocks, 56­58 trade restrictiveness index for, 71 sustainability of global trade patterns, 111­16 upper-middle income countries trade cost effect, 281 current account deficits, 14, 14f trade credit, 4, 19, 60 exports transparency, 5 competitiveness of, 112, 112t transportation sector diversification of, 107­8, 108t Chile, 195 upstream trade costs, 275, 278 costs, 141, 142 Uruguay, financial crisis impact on, 191, 191t environmental impact of, 308 Uzbekistan, exports subject to TTBs, 77 equipment exports, 108, 108t infrastructure investments, 173 V South-South trade, 121 value-added productivity travel and tourism industry, 195, 229 export diversification's impact on, 252­53f, 252­54 TTBs. See temporary trade barriers exports' effect on, 251­52, 251­52f Index 327 Van Assche, Ari, 275 Wheeler, D., 315n27 Van den Werf, E., 308 Williamson, J., 38n5 Végh, C., 86­87, 177 Winkler, Deborah, 145, 247 Venezuela, República Bolivariana de Woo, W. T., 229 development strategies, 99 World Bank export competitiveness, 113, 114­15f Doing Business Indicators database, 140­41 financial crisis impact, 171 on environmental goods and services, 312 GDP growth trends, 110f, 111 on foreign capital inflows, 17 growth performance, 158, 160, 160f Global Economic Prospects, 287 oil exports, 57 on labor migration, 294 vertical specialization, 276 on Malaysian investment climate, 147 Vieira, Fausto J. A., 85 on Mexico's fiscal policy, 177 Vietnam on oil prices, 265 apparel exports from, 102 on remittances, 298n31 commodity markets, 261 temporary trade barriers and, 17 exports subject to TTBs, 73, 76f Temporary Trade Barriers Database, 63, 65 Villagómez, A., 177 World Investment Report 2008 (UNCTAD), 124 Vlachos, J., 17 World Trade Organization (WTO) volatility China's entry into, 119, 121, 132 export diversification and, 135­44 climate change and, 301, 308­13 exports' effect on, 149­51, 151f temporary trade barriers and, 5, 17 openness of markets and, 136­39, 137f, 137t on trade reduction in wake of financial crisis, 3, 63 terms-of-trade volatility, 135, 143n3 Wyplosz, C., 15 W Y Wacziarg, Romain, 104 Yi, K.-M., 280 Waldenstrom, D., 17 Walkenhorst, P., 229 Z Wang, Jing, 7, 119, 284n8 Zahler, Roberto, 183, 226 Warnock, F., 21­22 Zainulbhai, A. S., 245n7 Warnock, V., 21­22 Zambia, terms-of-trade volatility in, 255 Wei, S.-J., 17, 38n9 Zavacka, V., 19 Weinstein, D., 48 Zeufack, Albert, 145 Wen Jiabao, 132, 133 Zhang, X., 38n9 Whalley, John, 7, 119 Zitouna, H., 279 ECO-AUDIT Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. The · 16 trees Office of the Publisher has chosen to print · 5 million Btu of Managing Openness: Trade and Outward- total energy Oriented Growth after the Crisis on recycled · 1,537 lb. of net paper with 50 percent postconsumer fiber in greenhouse gases accordance with the recommended standards · 7,404 gal. of waste for paper usage set by the Green Press Initiative, water a nonprofit program supporting publishers in · 450 lb. of solid using fiber that is not sourced from endangered waste forests. For more information, visit www.green pressinitiative.org. T he global financial crisis and the attendant collapse in trade--which affected all countries and all product categories--have triggered a broad reassessment of economic integration policies in developed and developing countries worldwide. In the aftermath of such macroeconomic instability, it is not surprising that some commentators are calling into question the underlying assumptions of trade liberalization and openness. In particular, the association between openness and volatility, as well as the prospect of global economic rebalancing, have led some commentators to reconsider outward-oriented or export-led growth strategies. If the global economic crisis highlights anything about policies of outward-oriented growth, it is that they are not a panacea for development. No set of policies is. Rather, outward-oriented growth offers both opportunities and risks for developing countries. The key is in finding the appropriate set of policies to manage openness, one that maximizes its benefits and minimizes its costs. The economic benefits that openness can offer are well known, and most policy makers in developing and developed countries alike remain committed to the progressive opening of international markets. Although there is a need for continued policy monitoring, most countries' policy-level commitment to openness has not been undermined by the crisis--witness the relatively limited recourse to protectionist policies since the crisis began, despite its magnitude. The evidence presented in this volume suggests that many developing countries will continue along that course for the foreseeable future. Managing Openness: Trade and Outward-Oriented Growth after the Crisis contributes to this important policy debate by presenting a comprehensive analysis of recent empirical work on the trade-related aspects of the crisis and providing relevant insights on how countries can manage outward-oriented growth in the postcrisis environment. The authors address critical policy issues revolving around the topic of outward- oriented growth, including the following: · The continued relevance of the export-led growth model in the postcrisis environment, and the increasing importance of South-South trade · Policy instruments to help manage the risks that come with increased openness · Lessons learned from the crisis for particular countries and regions · Effects of emerging trade policy issues--such as climate change, commodities, global production networking, and migration--on the prospects for recovery and outward-oriented growth. This book will be of interest to policy makers and civil society organizations in developing countries, as well as to students of international economics. About our Program The World Bank's International Trade Department produces and disseminates policy-oriented knowledge products and forges partnerships on trade to advance an inclusive trade agenda for developing countries and to enhance developing countries' trade competitiveness in global markets. Learn more about the World Bank's trade portfolio at: www.worldbank.org/trade. ISBN 978-0-8213-8631-6 SKU 18631