53879 AGRICULTURAL PRICE DISTORTIONS, INEQUALITY, AND POVERTY Editors Kym Anderson · John Cockburn · Will Martin Agricultural Price Distortions, Inequality, and Poverty Agricultural Price Distortions, Inequality, and Poverty Kym Anderson, John Cockburn, and Will Martin, Editors Washington, D.C. © 2010 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 E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 13 12 11 10 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, denominations, 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 per- mission 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: 978-0-8213-8184-7 eISBN: 978-0-8213-8185-4 DOI: 10.1596/978-0-8213-8184-7 Library of Congress Cataloging-in-Publication Data Agricultural price distortions, inequality, and poverty / Kym Anderson, John Cockburn and Will Martin, editors. p. cm. ISBN 978-0-8213-8184-7 -- ISBN 978-0-8213-8185-4 (electronic) 1. Agricultural prices--Government policy. 2. Farm income--Developing countries. 3. Agricultural wages-- Developing countries. 4. Poverty. I. Anderson, Kym. II. Cockburn, John. III. Martin, Will, 1953 - HD1447.A37 2010 339.4'6--dc22 2009041807 Cover design: Tomoko Hirata/World Bank Cover photo: © Curt Carnemark/World Bank Photo Library OTHER TITLES IN THE SERIES Distortions to Agricultural Incentives: A Global Perspective, 1955­2007 edited by Kym Anderson, 2009 Distortions to Agricultural Incentives in Africa edited by Kym Anderson and William A. Masters, 2009 Distortions to Agricultural Incentives in Asia edited by Kym Anderson and Will Martin, 2009 Distortions to Agricultural Incentives in Europe's Transition Economies edited by Kym Anderson and Johan Swinnen, 2008 Distortions to Agricultural Incentives in Latin America edited by Kym Anderson and Alberto Valdés, 2008 CONTENTS Foreword xvii Acknowledgments xix Contributors xxi Abbreviations xxv PART I INTRODUCTION 1 1 Introduction and Summary 3 Kym Anderson, John Cockburn, and Will Martin PART II GLOBAL CGE APPROACHES 47 2 Global Welfare and Poverty Effects: Linkage Model Results 49 Kym Anderson, Ernesto Valenzuela, and Dominique van der Mensbrugghe 3 Global Poverty and Distributional Impacts: The GIDD Model 87 Maurizio Bussolo, Rafael De Hoyos, and Denis Medvedev 4 Poverty Impacts in 15 Countries: The GTAP Model 119 Thomas W. Hertel and Roman Keeney PART III NATIONAL CGE APPROACHES: ASIA 145 5 China 147 Fan Zhai and Thomas W. Hertel 6 Indonesia 179 Peter Warr 7 Pakistan 209 Caesar B. Cororaton and David Orden vii viii Contents 8 The Philippines 247 Caesar B. Cororaton, Erwin Corong, and John Cockburn 9 Thailand 283 Peter Warr PART IV NATIONAL CGE APPROACHES: AFRICA 301 10 Mozambique 303 Channing Arndt and James Thurlow 11 South Africa 331 Nicolas Hérault and James Thurlow PART V NATIONAL CGE APPROACHES: LATIN AMERICA 357 12 Argentina 359 Martín Cicowiez, Carolina Díaz-Bonilla, and Eugenio Díaz-Bonilla 13 Brazil 391 Joaquim Bento de Souza Ferreira Filho and Mark Horridge 14 Nicaragua 423 Marco V. Sánchez and Rob Vos Appendix: Border Price and Export Demand Shocks in Developing Countries from Rest-of-the-World Trade Liberalization: The Linkage Model 457 Dominique van der Mensbrugghe, Ernesto Valenzuela, and Kym Anderson Index 489 Contents ix Figures 3.1 Income Distributions among Agricultural and Nonagricultural Populations of the World, 2000 90 3.2 Relationship between Income Levels and Share of Employment in Agriculture, by Country, 2000 92 3.3 Share of the Population in Agriculture and of Agriculture in Total Income, Developing Countries, 2000 93 3.4 Inequality Variation in Agricultural, Nonagricultural, and All Households, Developing Countries, 2000 95 3.5 The GIDD Methodological Framework 97 3.6 The Effects on Real National Consumption of the Global Removal of Agricultural and All Merchandise Trade Distortions 100 3.7 The Effects on National Real Factor Rewards of the Global Removal of Only Agricultural Price and Trade Policies 102 3.8 National GICs: The Effects on the Per Capita Household Income Distribution of a Hypothetical 1 Percent Increase in Agricultural Versus Nonagricultural Incomes, Developing Countries 104 3.9 Regional and National GICs: The Effects on the Per Capita Household Income Distribution of Full Global Agricultural Policy Reform 108 3.10 Changes in Poverty as a Share of the Total Change among the Greatest Winners and Losers in Full Global Agricultural Policy Reform, Developing Countries 110 3.11 The Theil Index of Overall and Between-Group Income Distributional Changes after Full Global Agricultural Policy Reform 114 5.1 The Impacts of Prospective Liberalizations on Welfare of Five Types of Households, China 167 6.1 Initial and Simulated Levels of Poverty Incidence, Indonesia 201 7.1 Poverty Incidence, Pakistan, 1986­87 to 2004­05 224 8.1 Trends in Poverty Indexes, the Philippines, 1985­2000 253 12.1 Export Taxes, Argentina, 2001­07 362 12.2 Tax Revenue, Argentina, 1991­2007 363 12.3 Agricultural Producer Price Indexes, Argentina, 1992­2006 364 12.4 Indexes of International Agricultural Prices, Relative Domestic Agricultural Prices, and the Real Effective Exchange Rate, Argentina, 1996­2008 365 13.1 The Share of the Poor, by State, Brazil, 2001 399 14.1 Tariff Rates on Agricultural and Total Imports from the United States under DR-CAFTA, Nicaragua, 2003­20 428 14.2 The NRAs for Exportable, Importable, and All Covered Farm Prod- ucts, Nicaragua, 1991­2004 429 14.3 The NRAs for Agricultural and Nonagricultural Tradable Sectors and the RRAs for Farmers, Nicaragua, 1991­2004 430 14.4 The Fiscal Cost of Trade Liberalization, Nicaragua 446 x Contents Tables 1.1 Global Poverty and Inequality, by Region, 1981­2005 4 1.2 The NRAs for Tradable Agricultural and Nonagricultural Products and the RRA, Focus Regions, 1980­2004 7 1.3 The Linkage Model's Effects of Full Global Liberalization of Agricultural and All Merchandise Trade on Economic Welfare and Real GDP, by Country and Region 15 1.4 The Linkage Model's Effects of Full Global Merchandise Trade Liberalization on Real Factor Prices, by Country and Region 17 1.5 The Linkage Model's Effects of Full Global Merchandise Trade Liberalization on the Number of Extreme Poor, by Region 18 1.6 The GIDD Model's Effects of the Removal of Agricultural and All Merchandise Trade Distortions on the Number of Extreme Poor, by Region 20 1.7 The GIDD Model's Effects of the Removal of Agricultural and All Merchandise Trade Distortions on Global Poverty and Inequality, Farm and Nonfarm Households 21 1.8 The GIDD Model's Effects of the Removal of Agricultural and All Merchandise Trade Distortions on Global Poverty and Inequality, Farm and Nonfarm Households, by Region 22 1.9 The GTAP Model's Effects of Full Global Liberalization of Agricultural and All Merchandise Trade on the Number of Extreme Poor, by Country 24 1.10 Characteristics of the Models in the Global and National Country Studies 26 1.11 The Impact of Reform on the Incidence of Extreme Poverty, by Country 28 1.12 The Impact of Reform on the Incidence of Income Inequality, by Country 32 1.13 The Direction of the Effects of Global Reform on Extreme Poverty, by Country 36 2.1 Structure of Producer Price Distortions in Global Goods Markets, 2004 53 2.2 The Impact on Real Income of the Full Liberalization of Global Merchandise Trade, by Country and Region, 2004 58 2.3 Regional and Sectoral Sources of the Welfare Gains from the Full Liberalization of Global Merchandise Trade, 2004 61 2.4 The Impact of Full Global Liberalization on the Shares of Global Output Exported, by Product, 2004 62 2.5 The Impact of Full Global Trade Liberalization on Agricultural and Food Output and Trade, by Country and Region, 2004 63 2.6 The Impact of Global Liberalization on the Share of Agricultural and Food Production Exports, by Country and Region, 2004 65 Contents xi 2.7 The Impact of Global Liberalization on Self-sufficiency in Agricultural and Other Products, by Region, 2004 66 2.8 Shares of Production Exported and of Consumption Imported before and after Full Global Liberalization of all Merchandise Trade, by Product, 2004 67 2.9 The Impact of Full Global Liberalization on Real International Product Prices, 2004 68 2.10 The Impacts of Full Global Merchandise Trade Liberalization on Real Pretax Factor Prices, by Country and Region, 2004 70 2.11 The Effects of the Full Global Liberalization of Agricultural and Merchandise Trade on Sectoral Value Added, by Country and Region, 2004 73 2.12 The Poverty Effects of Full Global Liberalization of Merchandise Trade Reform, by Region, 2004 77 3.1 Characteristics of the Poor in Agricultural and Nonagricultural Households, Developing Countries, 2000 91 3.2 The Household Effects of the Removal of Agricultural and All Merchandise Trade Distortions on Global Poverty and Inequality 105 3.3 The GIDD Model: The Effects of the Removal of Agricultural and All Merchandise Trade Distortions on the Incidence of Poverty, by Region 107 3.4 The Effects of the Full Global Liberalization of Agricultural and All Merchandise Trade on the Incidence of Inequality and Poverty, Developing Countries 111 4.1 Estimates of Elasticities of the US$1-a-Day Poverty Headcount with Respect to Total Income, 15 Countries 122 4.2 Stratum Contributions to the US$1-a-Day Poverty Population, 15 Countries 124 4.3 Stratum- and Earnings-Specific Poverty Elasticities, Bangladesh 126 4.4 The Contributions of Earnings to the Total US$1-a-Day Poverty Response, 15 Countries 127 4.5 Summary of the Poverty Impacts of Global Reforms, 15 Countries 131 4.6 Earnings Differences after Agricultural and Nonagricultural Reforms, 15 Countries 136 4.7 Earnings-Specific Differences between Agricultural and Nonagricultural Reforms in the Changes in Poverty, 15 Countries 137 4.8 Stratum-Specific Differences in the Changes in Poverty after Agricultural and Nonagricultural Reforms, 15 Countries 138 4.9 Change in Poverty after Commodity-Specific Reforms, 15 Countries 140 4.10 Change in Poverty Caused by Foodgrain and Feedgrain Reforms, 15 Countries 141 5.1 The Sectoral Structure of GDP, Trade, Import Tariffs, and Export Subsidies, China, Around 2004 154 xii Contents 5.2 Modeled Liberalization Scenarios, China 157 5.3 Exogenous Demand and Price Shocks Caused by Liberalization in the Rest of the World, China 158 5.4 Aggregate Simulation Results of Prospective Liberalizations, China 161 5.5 The Effects of Prospective Liberalizations on Income Inequality and Poverty, China 164 5.6 The Effects of Prospective Liberalizations on Sectoral Outputs and Trade, China 171 5.7 Sensitivity Analysis of the Simulation Results, China 174 6.1 The Cost Shares of Major Factors of Production, Paddy and Other Industries, Indonesia, 2000 184 6.2 Sources of the Factor Incomes of 10 Broad Household Groups, Indonesia, 2000 187 6.3 Expenditure and Poverty Incidence by Household Category, Indonesia, 2000 187 6.4 Industry Assistance Rates Used in Modeling, Indonesia, 2004 191 6.5 Exogenous Border Price Shocks That Would Result from Rest-of-the-World Liberalization, Indonesia 193 6.6 Aggregate Simulation Results of the Prospective Liberalization of All Commodities, Indonesia 195 6.7 Poverty Effects of the Prospective Liberalization of All Commodities, Indonesia 198 6.8 Income Inequality Effects of the Prospective Liberalization of All Commodities, Indonesia 202 6.9 Aggregate Simulation Results of the Prospective Liberalization Only of Agricultural Commodities, Indonesia 203 6.10 Poverty Effects of the Prospective Liberalization of Only Agricultural Commodities, Indonesia 204 6.11 Income Inequality Effects of the Prospective Liberalization of Only Agricultural Commodities, Indonesia 206 7.1 Elasticity Parameters and Production Structure, Pakistan, 2001­02 215 7.2 Model Household Categories, Pakistan 218 7.3 Parameters and Exogenous Demand and Price Shocks Caused by Rest-of-the-World Liberalization, Pakistan 220 7.4 Household Income Sources and Income Taxes, Pakistan, 2001­02 222 7.5 Poverty Estimates, Pakistan, 1998­99, 2001­02, and 2004­05 224 7.6 The Poverty Effects of Prospective Liberalizations, Pakistan 227 7.7 Aggregate Simulation Results of Prospective Agricultural and Nonagricultural Liberalization, Pakistan 228 7.8 Household Welfare and Price Effects of the Liberalization in All Goods Trade by the Rest of the World, Pakistan 232 Contents xiii 7.9 The Sectoral Effects of the Unilateral Liberalization of All Goods Trade, Pakistan 236 7.10 Sensitivity Analysis of Household Welfare Effects, Two Tax Replacement Schemes, Pakistan 242 7.11 Sensitivity Analysis of Poverty Effects, Two Tax Replacement Schemes, Pakistan 243 8.1 The NRAs to Major Agricultural Commodities, the Philippines, 1960­2004 251 8.2 Poverty Incidence and Food Expenditure Shares, the Philippines, 1997 and 2000 252 8.3 Production Structure, the Philippines, 2000 255 8.4 Trade Structure and Elasticity Parameters, the Philippines, 2000 258 8.5 Structure of Household Expenditure, by Decile, the Philippines, 2000 261 8.6 Exogenous Demand and Price Shocks Caused by Rest-of-the-World Liberalization 263 8.7 Aggregate Simulation Results of Prospective Liberalizations in Agriculture and Nonagriculture, the Philippines 268 8.8 Effects of Prospective Liberalizations on Real Household Consumption, by Income Group, the Philippines 273 8.9 Income Inequality and the Poverty Effects of Prospective Liberalizations, by Location, Gender, and Skills, the Philippines 277 9.1 Cost Shares of the Major Factors of Production, Paddy and Other Industries, Thailand, 2000 286 9.2 Sources of the Factor Incomes of Broad Household Groups, Thailand, 2000 286 9.3 Characteristics of the Major Household Categories, Thailand, 2000 287 9.4 Industry Assistance Rates Used in Modeling, Thailand, 2004 289 9.5 Exogenous Border Price Shocks Caused by Rest-of-the-World Liberalization, Thailand 290 9.6 Aggregate Simulation Results of the Prospective Liberalization of All Commodities, Thailand 291 9.7 The Poverty and Inequality Effects of the Prospective Liberalization of All Commodities, Thailand 293 9.8 Aggregate Simulation Results of the Prospective Liberalization of Only Agricultural Commodities, Thailand 295 9.9 The Poverty and Inequality Effects of the Prospective Liberalization of Only Agricultural Commodities, Thailand 296 10.1 The Economic Structure in 2002 and Price Distortions in 2004, Mozambique 306 10.2 Factor Intensities of Production, Mozambique, 2002 309 xiv Contents 10.3 Household Income and Expenditure Shares, Mozambique, 2002 312 10.4 The Exogenous Demand and Price Shocks Caused by Rest-of-the-World Liberalization, Mozambique 316 10.5 The Macroeconomic Simulation Results of Prospective Liberalization Abroad and Nationally, Mozambique 318 10.6 The Effects of Prospective Liberalization Abroad and Nationally on GDP, by Sector, Mozambique 319 10.7 The Effects of Prospective Liberalization Abroad and Nationally on the Real Value of Exports, Mozambique 321 10.8 The Effects of Prospective Liberalization Abroad and Nationally on the Real Value of Imports, Mozambique 323 10.9 The Effects of Prospective Liberalization Abroad and Nationally on Employment, Welfare, and Poverty, Mozambique 325 11.1 Economic Structure and Price Distortions, South Africa, 2002 333 11.2 Direct Tax Rates on Households, South Africa, 2000 337 11.3 Household Income Shares, South Africa, 2000 338 11.4 Household Expenditure Shares, South Africa, 2000 339 11.5 Exogenous Demand and Border Price Shocks of Rest-of-the-World Liberalization, South Africa 341 11.6 Macroeconomic Simulation Results of Prospective Liberalization Abroad and Nationally, South Africa 342 11.7 The Effects of Prospective Liberalization Abroad and Nationally on GDP, by Sector at Factor Cost, South Africa 343 11.8 Factor Intensity Structure, South Africa, 2002 345 11.9 The Effects of Prospective Liberalization Abroad and Nationally on Factor Rewards, Employment, and Welfare, South Africa 346 11.10 The Effects of Prospective Liberalization Abroad and Nationally on Sectoral Employment, Income Inequality, and Poverty, South Africa 347 12.1 Components of the Social Accounting Matrix, Argentina, 2005 367 12.2 External Trade Structure, Argentina, 2005 369 12.3 Export Taxes and Import Tariffs, Argentina, 2005 370 12.4 Aggregate Simulation Results of Prospective Liberalization, Argentina 375 12.5 Sectoral Simulation Results of Prospective Liberalization, Argentina 377 12.6 Poverty Effects of Prospective Liberalization, Argentina 380 12.7 Income Inequality Effects of Prospective Liberalization, Argentina 381 12.8 Fiscal and Real Investment Effects of Prospective Liberalization, Argentina 384 13.1 Poverty Indicators, by Poverty Group, Brazil, 2001 398 13.2 Poverty and Income Inequality by State, Brazil, 2001 400 Contents xv 13.3 Share of Occupations in the Activity Labor Bill, by Wage Group, Brazil, 2001 402 13.4 The Structure of External Trade, Brazil, 2001 405 13.5 Exogenous Demand and Border Price Shocks Caused by Prospective Global Trade Liberalization, Brazil 407 13.6 The Macroeconomic Impacts of Prospective Trade Liberalizations, Brazil 410 13.7 The Effects of Prospective Global Trade Liberalization on Sectoral Outputs, Exports, and Imports, Brazil 411 13.8 The Effects of Prospective Global Trade Liberalization on Output, by State, Brazil 413 13.9 The Effects of Prospective Global Trade Liberalization on Poverty and Income Inequality, by Household Income Group, Brazil 414 13.10 Decomposition of the Effects of Liberalization on Real Incomes, by Household Income Group, Brazil 416 13.11 The Effects of Prospective Global Trade Liberalization on Poverty, by State, Brazil 417 14.1 Macroeconomic Indicators, Nicaragua, 1990­2005 425 14.2 Structure of Value Added by Sector at Factor Cost, Nicaragua, 1995, 2000, and 2004 426 14.3 The Poverty Headcount Ratio and the Gini Coefficient of Income Inequality, Nicaragua, 1993­2005 431 14.4 Trade Structure and World Price Shocks Imposed in the Global Trade Liberalization Simulations, Nicaragua, 2004 440 14.5 The Impact of Prospective Trade Liberalization on the Macroeconomy, Nicaragua, 2004 441 14.6 The Impact of Prospective Trade Liberalization on Real Sectoral GDP, Nicaragua 443 14.7 The Impact of Prospective Trade Liberalization on Employment and Real Wages, Nicaragua 447 14.8 The Impact of Prospective Trade Liberalization on Employment, by Sector, Nicaragua 448 14.9 Microsimulation Results for Poverty and Inequality, Nicaragua 451 A.1 Sectoral Concordance of the Linkage Model and the GTAP Version 7 Database 459 A.2 Regional Concordance of the Linkage Model and the GTAP Version 7 Database 460 A.3 Key Elasticities in the Linkage Model 464 A.4 Price Distortion Structure in GTAP Version 7p5 and the Distortion Rates Drawn from the World Bank Project, 2004 468 A.5 Export Price, Import Price, and Export Demand Shocks to National Models from Rest-of-the-World Agricultural and Trade Policy Reform, Selected Developing Countries 471 FOREWORD The vast majority of the world's poorest households depend to a considerable extent on farming for their incomes, while food represents a large component of the consumption of all poor households. The prices of farm products are thus crucial determinants of the extent of poverty and inequality in the world. Yet, for generations, these prices have been heavily distorted by government policies in high-income countries and in developing countries. True, many countries began to reform their agricultural price and trade policies in the 1980s, but Distortions to Agricultural Incentives: A Global Perspective, 1955­2007 (edited by Kym Anderson), a 2009 World Bank publication, shows that the extent of government policy inter- vention is still considerable and still favors farmers in high-income countries at the expense of farmers in developing countries. What would be the poverty and inequality consequences of the removal of the remaining distortions to agricultural incentives? This question is of great rele- vance to governments in evaluating ways to engage in multilateral and regional trade negotiations or to improve their own policies unilaterally. The answer is often not clear in any one country and is certainly an empirical matter for groups of countries because the positive and negative effects in different settings may be offsetting. Some analysts have sought answers in past events using ex post econo- metric analysis of historical data, but it is not easy to find natural experiments that are suitable for analysis and from which it is possible to generalize. An alternative approach--the one adopted in the present study--is to undertake ex ante analysis using economy-wide models, including global models, so as to be able to simulate the prospective effects of the multilateral removal of all price-distorting policies. This volume represents a first attempt to exploit new methodologies, models, and databases (developed partly as a consequence of recent World Bank­sponsored research) to assess the relative impacts on national, regional, and global poverty xvii xviii Foreword and inequality of agricultural and nonagricultural trade policies at home and abroad. While clear, definitive answers are not easy to determine, the volume extends our understanding of the contributions of various policies to inequality and poverty in selected developing countries and regions. Its broad finding is that the removal of all current price-distorting policies is likely to reduce global poverty and inequality, but there may be a few countries in which such a policy change on its own might worsen poverty nationally. In particular, it highlights the fact that the results are sensitive to assumptions about the changes in taxation needed to compensate the national treasury for losses in trade tax revenue, the degree of flexibility in each country's labor markets, the complementary measures taken to increase the opportunities for farm families and mitigate any adverse poverty impacts, and so on. As always, agricultural and trade policy reforms have better prospects of becoming politically and socially sustainable, the more govern- ments also provide optimal domestic safety net policies. Justin Yifu Lin Senior Vice President and Chief Economist The World Bank ACKNOWLEDGMENTS This book analyzes the effects of agricultural policies and farm and nonfarm trade policies around the world on national and regional economic welfare, on income inequality among and within countries, and, most importantly, on the level and incidence of poverty in developing countries. The studies in this volume include economy-wide analyses of the inequality and poverty effects of own-country poli- cies compared with rest-of-the-world policies for 10 individual developing coun- tries in three continents. In addition, the book includes three chapters that each use a separate global economic model to examine the effects of policies on aggre- gate poverty and the distribution of poverty across many identified developing countries. The book is motivated by two policy issues: first, the World Trade Organiza- tion's struggle to conclude the Doha Round of multilateral trade negotiations, in which agricultural policy reform is, once again, one of the most contentious top- ics in the talks, and, second, the struggle of the developing countries to achieve their Millennium Development Goals by 2015, notably the alleviation of hunger and poverty, which depends crucially on policies that affect agricultural incen- tives. The study's timing is facilitated by recent methodological developments involving microsimulation modeling based on household survey data, in conjunc- tion with economy-wide computable general equilibrium modeling. Recent surveys are now available at the World Bank on more than 100 countries. More- over, the World Bank has recently compiled a comprehensive new global database that updates and expands substantially our understanding of the distortions to agricultural incentives, particularly in developing countries. The authors of the chapters in this book are therefore indebted to the many contributors to these recent developments in methodology and data. In turn, we are grateful to the xix xx Acknowledgments authors for taking part in our project, which is an extension of the project that generated the new distortions database (details of which are available at http://go.worldbank.org/5XY7A7LH40). We extend our thanks to the senior advisory board of the overall project. The board's members have provided advice and encouragement throughout the plan- ning and implementation stages of the project. The board consists of Yujiro Hayami, Bernard Hoekman, Anne Krueger, John Nash, Johan Swinnen, Stefan Tangermann, Alberto Valdés, Alan Winters, and, until his untimely death in March 2008, Bruce Gardner. The editors, the other authors, and the advisory board are all grateful for the untiring administrative support provided throughout the entire project by Michelle Chester in the trade unit of the Development Research Group of the World Bank. They also appreciate the feedback and helpful comments of col- leagues at the 2008 and 2009 annual Global Economic Analysis Conference, in Helsinki and Santiago, respectively, where earlier versions of many of the chapters herein were presented. For financial assistance, grateful thanks go to the Development Research Group of the World Bank and the trust funds of the governments of the Netherlands and the United Kingdom. This combined support has made it possible for the study to include a wide range of developing countries. Kym Anderson, John Cockburn, and Will Martin August 2009 CONTRIBUTORS Kym Anderson is George Gollin Professor of Economics at the University of Adelaide, Australia and a fellow of the Centre for Economic Policy Research, London. During 2004­07, he was on an extended sabbatical as lead economist (trade policy) in the Development Research Group of the World Bank in Washington, DC. Channing Arndt, following a five-year period based in Maputo, Mozambique, is a professor at the Department of Economics at the University of Copenhagen. His publications cover poverty measurement, trade policy, the macroeconomic impli- cations of HIV/AIDS, agricultural productivity growth, and demand systems estimation. Maurizio Bussolo is a senior economist at the World Bank in Washington, DC. At the time of his contribution to this volume, he was with the Development Prospects Group; he is now in the economic policy unit of the Latin America and the Caribbean Region of the World Bank. Martin Cicowiez is a research economist and lecturer at the Center for Distribu- tional, Labor, and Social Studies at the Universidad Nacional de La Plata, in Buenos Aires, Argentina. John Cockburn is a professor of economics at the Université Laval in Quebec and co-director of the Poverty and Economic Policy Research Network. He conducts research on child welfare, poverty, growth, and trade policy. xxi xxii Contributors Erwin Corong is a PhD student at the Centre of Policy Studies at Monash Univer- sity, Melbourne. At the time of his contribution to this volume, he was with the Poverty and Economic Policy Research Network at the Université Laval in Quebec. He was subsequently with the International Food Policy Research Institute in Washington, DC. Caesar B. Cororaton is a research fellow at the Global Issues Initiative at the Institute for Society, Culture, and Environment of Virginia Polytechnic and State University in Alexandria, Virginia. At the time of his contribution to this volume, he was with the International Food Policy Research Institute in Washington, DC. Carolina Díaz-Bonilla is an economist in the Poverty Reduction and Economic Management Network of the Latin America and the Caribbean Region at the World Bank in Washington, DC. At the time of her contribution to this volume, she was in the Development Prospects Group of the World Bank. Eugenio Díaz-Bonilla is executive director for Argentina and Haiti at the Inter- American Development Bank, in Washington, DC. He was formerly a senior fellow in the Macro and Trade Division of the International Food Policy Research Institute in Washington, DC. Joaquim Bento de Souza Ferreira Filho is a professor in the Department of Economics in the Escola Superior de Agricultura "Luiz de Queiroz," Universidade de São Paulo at Piracicaba, São Paulo, Brazil. He has served as president of the Brazilian Society of Rural Economics, Management, and Sociology. Nicolas Hérault is a research fellow with the Melbourne Institute of Applied Economic and Social Research at the University of Melbourne, which he joined in March 2007 following doctoral studies at the University of Bordeaux IV in France, where he began his microsimulation analyses of South Africa. Thomas W. Hertel is a distinguished professor, Department of Agricultural Economics, and founding executive director, Center for Global Trade Analysis at Purdue University, West Lafayette, Indiana, where he teaches and conducts research on the economy-wide impacts of trade and environmental policies. Mark Horridge is a professor of economics at the Centre of Policy Studies at Monash University in Melbourne. He specializes in computable general equilib- rium modeling, including the ongoing development of widely used economic modeling software such as Gempack. Contributors xxiii Rafael De Hoyos, at the time of his contribution to this volume, was a senior economist in the Development Prospects Group of the World Bank in Washington, DC. He is now chief of advisers to the Mexican Under-Secretary of Education, Mexico City. Roman Keeney is an assistant professor in the Department of Agricultural Eco- nomics, Purdue University, West Lafayette, Indiana, where he specializes in teaching and research on the global economy-wide modeling of economic policies. Will Martin is the research manager of the Rural Development Unit in the Devel- opment Research Group at the World Bank in Washington, DC. He specializes in trade and agricultural policy issues globally, especially in Asia, and has written extensively on trade policies affecting developing countries. Denis Medvedev is a young professional in the World Bank in Washington, DC. At the time of his contribution to this volume, he was with the Development Prospects Group at the World Bank. He specializes in the global economy-wide modeling of economic policies, especially as these policies affect inequality and poverty. David Orden is a senior research fellow at the International Food Policy Research Institute in Washington, DC and director of the Global Issues Initiative at the Institute for Society, Culture, and Environment at Virginia Polytechnic Institute and State University in Alexandria, Virginia. His research focuses on the econom- ics and political economy of agricultural policies. Marco V. Sánchez is economic affairs officer at the Department of Economic and Social Affairs of the United Nations, New York. He was previously associated with the Sub-Regional Office in Mexico City of the United Nations Economic Commission for Latin America and the Caribbean. James Thurlow is a research fellow at the University of Copenhagen and at the International Food Policy Research Institute in Washington, DC. His research focuses on the role of agricultural transformation in the development process, particularly strategies for growth and poverty reduction in Sub-Saharan Africa. Ernesto Valenzuela is a lecturer and research fellow at the School of Economics and the Centre for International Economic Studies at the University of Adelaide, Australia. During 2005­07, he was a consultant at the Development Research Group of the World Bank in Washington, DC. xxiv Contributors Dominique van der Mensbrugghe is lead economist in the Development Prospects Group of the Development Economics Vice-Presidency of the World Bank in Washington, DC, where he specializes in global economy-wide modeling. He is the architect of the global Linkage model. Rob Vos is the director of the Development Policy and Analysis Division at the Department of Economic and Social Affairs of the United Nations, New York, and affiliated professor of finance and development at the Institute of Social Studies in The Hague. Peter Warr is the John Crawford Professor of Agricultural Economics and found- ing director of the Poverty Research Centre in the Arndt-Corden Division of Eco- nomics, Research School of Pacific and Asian Studies, at the Australian National University, Canberra. Fan Zhai was, at the time of his contribution to this volume, an economist in the Asian Development Bank in Manila on secondment to the Asian Development Bank Institute in Tokyo as a research fellow. He is now with the China Investment Bank in Beijing. His research has focused on modeling Asian economies. ABBREVIATIONS CGE computable general equilibrium (model) DR-CAFTA Dominican Republic­Central America Free Trade Agreement EU15 The 15 members of the European Union prior to 2004 GDP gross domestic product GIC growth incidence curve GIDD Global Income Distribution Dynamics (data set) GTAP Global Trade Analysis Project MS microsimulation NRA nominal rate of assistance OECD Organisation for Economic Co-operation and Development PPP purchasing power parity RRA relative rate of assistance SAM social accounting matrix WTO World Trade Organization Note: All dollar amounts are U.S. dollars (US$) unless otherwise indicated. xxv Part I Introduction 1 INTRODUCTION AND SUMMARY Kym Anderson, John Cockburn, and Will Martin For decades, the earnings from farming in many developing countries have been depressed because of a pro-urban, antiagricultural bias in own-country policies and because governments in more well off countries are favoring their farmers by imposing import barriers and providing subsidies. These policies have reduced national and global economic welfare, inhibited economic growth, and added to inequality and poverty because no less than three-quarters of the billion poorest people in the world have been dependent directly or indirectly on farming for their livelihoods (World Bank 2007). Over the past two or three decades, however, numerous developing-country governments have reduced sectoral and trade policy distortions, while governments in some high-income countries have also begun reforming protectionist farm policies. Partly as a consequence of these and associated domestic policy reforms and the consequent growth in incomes in many developing countries, the number of people living on less than US$1 a day fell by nearly half between 1981 and 2005, while the share of these people fell from 42 to 16 percent of the global population (table 1.1). Notwithstanding the dramatic achievement in poverty reduction, there were still almost 900 million extremely poor people in 2005, and the number may have risen following the eruption of the global financial crisis in 2008. Moreover, most of the improvement has been in Asia (especially China), while, in Sub-Saharan Africa, the incidence of poverty was not much different in 2005 than in 1981, at around 40 percent (amounting to 300 million people in 2005). Despite the success of China, more than 100 million people were still living on less than US$1 a day there in 2005, 90 percent of whom lived in rural areas. In India, the number of 3 4 Table 1.1. Global Poverty and Inequality, by Region, 1981­2005 (number and percent of people living on less than 2005 PPP US$1 a day) Share of Index of poor who income are rural, inequality, Region 1981 1987 1993 1999 2005 2002, % 2004a Number of people, millions World 1,528 1,228 1,237 1,146 879 74 -- Sub-Saharan Africa 157 202 247 299 299 69 -- East Asia and Pacific 948 598 600 425 180 85 0.37 China 730 412 444 302 106 90 0.36 South Asia 387 384 341 359 350 75 0.35 India 296 285 280 270 267 74 0.33 Latin America and the Caribbean 27 35 34 40 28 34 0.52 Rest of world 9 9 15 23 22 50 -- East and South Asia share in world, % 87 80 76 68 60 -- -- Share of population, % World 42 30 27 23 16 -- -- Sub-Saharan Africa 40 42 44 46 39 -- -- East Asia and Pacific 69 39 36 24 10 -- -- China 74 38 38 24 8 -- -- South Asia 42 37 29 27 24 -- -- India 42 36 31 27 24 -- -- Latin America and the Caribbean 7 8 7 8 5 -- -- Sources: Chen and Ravallion (2008); for the rural share, Ravallion, Chen, and Sangraula (2007); for the Gini coefficient, PovcalNet (2008). Note: The calculations are based on 2005 PPP (purchasing power parity) dollars. -- no data are available. a. The index is the Gini coefficient calculated as the population-weighted cross-country average of national Gini coefficients in the region for the available year nearest to 2004. Introduction and Summary 5 extreme poor remains stubbornly close to 300 million (74 percent in rural areas) despite large farm subsidies. Less pressing than the problem of extreme poverty, though nonetheless still important to the welfare of individuals, is the extent of income inequality.1 In the past, only inequality at the local level affected the utility of individuals, but the information and communication technology revolution has increased awareness of income differences nationally and even internationally. At national levels, there are concerns about rural-urban inequality, as well as inequality within other clas- sifications. Within rural areas, for example, differences in incomes may be vast among landless unskilled farm workers, subsistence farmers, the owners of larger commercial farms, and nonfarm workers in rural towns.2 At the global level, Milanovic (2005) points to three possible means of assess- ing the changes in income distribution in recent decades. One revolves around intercountry inequality, the comparison of average incomes across countries wherein each country has an equal weight in the world distribution regardless of population size; measured in this way, income distribution appears to have become more unequal. The second focuses on international inequality, the com- parison of average incomes across countries wherein each country is weighted by population; measured in this way, income inequality appears to have decreased, although the decrease has occurred mostly because of rapid population growth in China and India (see Bourguignon, Levin, and Rosenblatt 2004; Atkinson and Brandolini 2004). The third method of assessment focuses on global inequality, the comparison of individual incomes regardless of the country of citizenship. This method thus takes into account within-country inequality, which is ignored in the intercountry and international inequality approaches wherein individuals are all assumed to earn the average income in their countries. The rapid growth in the large emerging economies has tended to offset the increase in inequality within countries, and, so, by this last methodology, global inequality appears to have remained roughly constant since the late 1980s.3 Given the evidence currently available, our book focuses on one main question: how much scope is there to reduce poverty and inequality in the world and in spe- cific developing countries by unilaterally or globally eliminating the distortions in the incentives affecting the producers and consumers of tradable goods? This question is of great interest to the agricultural, trade, and development policy communities in many developing countries and in nongovernmental organiza- tions and international agencies. The answer is by no means obvious. While it is true that recent studies indicate that agricultural policies are responsible for the majority of the global welfare costs of the remaining distortions to goods markets, removing these policies could affect national poverty levels either negatively or positively. The answer in each country to our question depends on current food 6 Agricultural Price Distortions, Inequality, and Poverty and agricultural policies at home and abroad, as well as on earning and spending patterns and the taxes on the poor, among other factors. Account also needs to be taken of three other facts. First, the dependence of the extreme poor on agricul- ture for their livelihoods has been declining in numerous countries as alternative opportunities have emerged outside agriculture, especially in off-farm, part-time employment. Second, the lowering of trade barriers has improved the opportuni- ties for farmers to specialize in cash crops for export, increasing their potential benefits from any improvements in market access abroad for these crops or related products. Third, practical realities too important to ignore in some coun- tries are high levels of unemployment (as in South Africa) or policies that inhibit intersectoral labor mobility (as in China). Empirical studies undertaken as background for the World Trade Organiza- tion's ongoing Doha Round of multilateral trade negotiations suggest that, in 2001, when the Doha Round was launched, policy-driven distortions to agricultural incentives contributed around two-thirds of the global welfare cost of merchandise trade barriers and subsidies (for example, see Anderson, Martin, and van der Mensbrugghe 2006). While these studies did not have access to compre- hensive estimates of the distortions to farmer and food consumer incentives in developing countries other than the applied tariffs on imports, a more recent study that draws on a new database of distortions to agricultural incentives has confirmed the earlier result. Valenzuela, van der Mensbrugghe, and Anderson (2009) find that agricultural price and trade policies as of 2004 accounted for 60 percent of the global welfare cost of these and other merchandise trade policies. This is a striking outcome given that the shares of agriculture and food in global gross domestic product (GDP) and trade are less than 6 percent. The contribution of farm and food policies to the welfare cost of global trade-distorting policies in developing countries alone is estimated by these authors to be even greater, at 83 percent, of which a little more than one-third is generated by the policies of the developing countries themselves. Nonetheless, the price distortion estimates used in the modeling study (see Anderson and Valenzuela 2008) show that many devel- oping countries protect their less-competitive farmers from import competition; so, some of this subset of farmers might be hurt if all markets were opened. Thus, despite much reform over the past quarter century in policies leading to distortions in world trade, many of the relevant intervention measures, especially the agricultural ones, are still in place. Table 1.2 summarizes the average extent of these measures in developing and high-income countries. It shows that the rate of assistance to farmers relative to producers of nonfarm tradables has fallen by one- third in high-income countries since the late 1980s (from 51 to 32 percent), while, in developing countries, this relative rate of assistance has risen from 41 percent in the early 1980s to 1 percent in 2000­04. The latter trend in developing countries is caused partly by the phasing out of agricultural export taxes and partly by the Table 1.2. The NRAs for Tradable Agricultural and Nonagricultural Products and the RRA, Focus Regions, 1980­2004 (percent) Region, indicator 1980­84 1985­89 1990­94 1995­99 2000­04 Africa NRA agricultural exportables 35 37 36 26 25 NRA agricultural import- competing 13 58 5 10 2 NRA agricultural tradables 14 0 15 9 12 NRA nonagricultural tradables 2 9 3 2 7 RRA 13 8 17 10 18 South Asia NRA agricultural exportables 28 21 16 12 6 NRA agricultural import- competing 38 63 25 15 27 NRA agricultural tradables 2 47 0 2 13 NRA nonagricultural tradables 55 40 19 15 10 RRA 33 5 16 15 3 China and Southeast Asia NRA agricultural exportables 50 41 21 2 0 NRA agricultural import- competing 1 15 3 13 12 NRA agricultural tradables 35 28 12 5 7 NRA nonagricultural tradables 21 23 20 10 6 RRA 43 42 26 4 2 Latin America NRA agricultural exportables 27 25 11 4 5 NRA agricultural import- competing 14 5 19 13 21 NRA agricultural tradables 13 11 4 6 5 NRA nonagricultural tradables 19 17 7 7 5 RRA 27 24 3 1 1 All developing countries NRA agricultural exportables 41 36 19 6 3 NRA agricultural import- competing 17 38 23 22 23 NRA agricultural tradables 21 16 4 4 7 NRA nonagricultural tradables 35 27 17 10 6 RRA 41 34 18 5 1 High-income countries NRA agricultural exportables 12 22 16 8 7 NRA agricultural import- competing 58 71 62 54 51 NRA agricultural tradables 43 56 48 37 34 NRA nonagricultural tradables 3 3 3 2 1 RRA 38 51 45 34 32 Source: Anderson and Valenzuela (2008), based on estimates reported in the project country studies. Note: The relative rate of assistance (RRA) is defined as 100 * [(100 NRAagt) (100 NRAnonagt) 1], where NRAagt and NRAnonagt are the percentage nominal rates of assistance (NRAs) for the tradables parts of the agricultural and nonagricultural sectors, respectively, and NRAagt is the weighted average of the nominal rates of assistance for the exporting and import-competing subsectors of agriculture. 7 8 Agricultural Price Distortions, Inequality, and Poverty rise in assistance via food import restrictions over the period. Thus, both in high- income countries and in developing countries, there is now a large gap between the nominal rates of assistance for import-competing and export agriculture, as well as a continuing gap (albeit smaller than the corresponding gap in the 1980s) between the relative rates of assistance in the two groups of countries. In light of this evidence, our question above may be expressed more specifically for any developing country of interest as two additional questions: How important are a developing country's own agricultural and other trade policies compared with those of the rest of the world in determining the welfare of the poor in the coun- try? And what is the contribution of agricultural policies to these outcomes? Clear answers to these questions are crucial in guiding countries in national policy making and in negotiating bilateral and multilateral trade agreements. Now is an appropriate time to address these multifaceted questions for at least two policy reasons. One is that the World Trade Organization is struggling to con- clude the Doha Round of multilateral trade negotiations, and agricultural policy reform is once again one of the most contentious issues in these talks. The other is that poorer countries are striving to achieve their United Nations­encouraged Millennium Development Goals by 2015, and the prime goals are the reduction of hunger and poverty. Farm-subsidizing rich countries are not alone in resisting reform; some developing countries likewise do not wish to remove food import barriers and farmer subsidies. There are also several analytical reasons for focusing more thoroughly on this issue now. First, methodologies to address the issue have been improving rapidly, most notably through the combination of economy-wide computable general equilibrium (CGE) modeling and microsimulation modeling based on household survey data. Prominent examples include the studies in Hertel and Winters (2006) and in Bourguignon, Bussolo, and Pereira da Silva (2008). Household income information is increasingly important for poverty and inequality analysis because farm households and rural areas in developing countries are rapidly diversifying their sources of income beyond the income generated by agricultural land and farm labor, including part-time off-farm work and remittances (Otsuka and Yamano 2006; Otsuka, Estudillo, and Sawada 2009). Hence, the once close corre- spondence between net farm income or agricultural GDP and farm household welfare is fading even in low-income countries (Davis, Winters, and Carletto 2009). Second, the compilation of national household surveys that are comparable for the purpose of cross-country analysis has progressed rapidly. Recent surveys are now available at the World Bank for more than 100 countries. The Global Income Distribution Dynamics (GIDD) data set has already begun to be used in conjunc- tion with the World Bank's Linkage model of the global economy to assess global Introduction and Summary 9 income distribution issues (for example, see Bussolo, De Hoyos, and Medvedev 2008; see also the GIDD Database). Third, the World Bank has recently compiled a comprehensive global database that substantially updates and expands our information on the distortions to agri- cultural incentives in developing countries.4 The estimates in this database have since been expressed so as to make them usable in national and global economy- wide models (Valenzuela and Anderson 2008). They differ from the usual estimates relied on by trade modelers of developing-country policies in that they are based on direct domestic-to-border price comparisons rather than on applied rates of import tariffs and other key border measures (see Narayanan and Walmsley 2008 for the latest compilation of the Global Trade Analysis Project [GTAP] data set). The present volume is a first attempt to exploit these new methodologies and databases to assess the relative impacts on national, regional, and global poverty and inequality of agricultural and nonagricultural trade policies at home and abroad. Poverty is defined in purchasing power parity terms at a threshold of US$1 a day per capita (the extreme poverty line) and also sometimes at a thresh- old of US$2 a day (the moderate poverty line). If these indicators are not available, then the national poverty line is used. The incidence of poverty (the share of the population below the poverty line) and the headcount (the absolute number of poor people) are also used. The Gini coefficient of income distribution is the key measure of inequality adopted here. Where possible, the national indicators for both poverty and inequality are calculated for farm and nonfarm households separately, in addition to the national averages. In undertaking this set of studies, we are acutely aware that agricultural and trade or domestic price subsidies are by no means the first-best policy instru- ments for achieving national poverty or income distribution objectives; this is largely the prerogative of public finance policies such as the supply of public goods or tax and transfer measures, including the provision of social safety nets funded through general tax revenues. However, should studies reveal that national trade-related policies are worsening poverty or inequality in particular countries, they provide a reason, in addition to the usual national gains-from-trade reason, for unilateral policy reform in these countries. Should the inequality- and poverty-reducing effects of national trade-related policy reforms be contingent on the rest of the world also reforming, this would provide another reason for a country to participate actively in promoting multilateral trade negotiations at the World Trade Organization. Furthermore, if global modeling studies reveal that multilateral trade reform would reduce global inequality and poverty, this would underline the importance of bringing the Doha Development Agenda of the World Trade Organization expeditiously to a successful conclusion through com- mitments to ambitious agricultural reform. 9 10 Agricultural Price Distortions, Inequality, and Poverty A negative finding--for example, that trade liberalization or farm subsidy cuts would increase poverty in a particular developing country--need not be a reason to shun welfare-enhancing reform, but would be a reason to use the results to pro- vide guidance to determine where tax or social programs might become more well targeted so that all groups in society share in the economic benefits of reform (see Ravallion 2008). The results of global reform also provide bargaining power to developing countries that are seeking aid-for-trade side payments to reduce any increase in poverty projected to be generated by trade reform that has been multilaterally agreed upon. The purpose of the rest of this chapter is to outline the analytical framework and the common empirical methodology adopted in the global and national case studies reported in subsequent chapters, to summarize and compare the modeling results from the global and national models, and to draw some general policy implications. The findings are based on three chapters (part II) that each use a global model to examine the effects of farm and nonfarm price and trade policies on global poverty and the distribution of poverty within and across many of the countries identified, plus 10 individual developing-country studies (parts III­V) spanning the three key regions: Asia (where nearly two-thirds of the world's poor live), Sub-Saharan Africa, and Latin America. Analytical Framework To capture the poverty and inequality effects of price-distorting policies ade- quately, one must give careful consideration to the impacts of the policies on household incomes and expenditures. Many farm households in developing coun- tries rely on the farm enterprise for virtually all of their incomes, and, in the world's poorest countries, the share of national poverty concentrated in such households is large. That the poorest households in the poorest countries are concentrated in agriculture means these households are likely to benefit from farm producer price increases generated by global trade policy reform, all else being equal. However, this outcome is not certain for several reasons. First, if a country provides protec- tion from import competition for the farm commodities produced by the poor, the domestic prices of the commodities may decline after liberalization. Second, poor farm households also spend the majority of their incomes on staple foods (Cranfield et al. 2003); so, if food prices rise as a consequence of reform, then this adverse effect on household expenditure may more than offset any beneficial effect of higher earnings. Third, the rural nonfarm and urban poor, too, would be adversely affected by a rise in the consumer prices of staple foods. However, it is possible that a trade reform that induced a rise in food prices may also raise the demand for unskilled labor (according to the relative factor intensity of production Introduction and Summary 11 in the economy's expanding sectors), which--depending on the intersectoral mobility of labor--might raise the incomes of poor households more than it raises the price of the consumption bundle of these households. The outcome is therefore not always going to be clear for any particular country and is certainly an empirical matter for groups of countries because the positive and negative effects in different settings will be more or less offsetting. Some analysts have sought answers in past events by using ex post econometric or micro household data analysis (as in the set of studies in Harrison 2007), but it is not easy to find natural experiments of specific policy reforms that are appro- priate for analysis and from which it is possible to generalize. An alternative approach--the one adopted in the present study--is to undertake ex ante analysis using economy-wide models. While such models have well-recognized limita- tions, they are the only option available if one is seeking to simulate the prospec- tive effects of the removal of all remaining price-distorting policies (see, for exam- ple, Francois and Martin 2007). This is particularly the case if global reform is one of the scenarios of interest and, even more so, if insights into the effects on overall world poverty and inequality are being sought. The approach adopted in our study is a variant on the pathbreaking approach pioneered by Hertel and Winters (2006) in their study of the poverty consequences of a prospective Doha Round agreement under the World Trade Organization. Like Hertel and Winters (2006), our study uses global models to assess the impli- cations of global reform for poverty, plus a series of national models to focus more attention on specific aspects of importance to particular countries. However, the present study contrasts with the earlier one compiled by Hertel and Winters in three key respects. First, our study focuses on the impacts of agricultural policies and distinguishes these from the impacts of other merchandise trade policies; moreover, it relies on the new database on distortions to agricultural incentives in developing countries that has only recently become available in a format that is readily usable by CGE modelers (Valenzuela and Anderson 2008). These distor- tion estimates (for 2004) are used to represent agricultural and food policies in each of the 10 national CGE models employed in our country case studies in parts III­V, as well as in the three global models described in part II. A second dis- tinction is that our study examines inequality, in addition to poverty. The third difference is that the present study is able to draw on the massive data collection and modeling effort undertaken for the GIDD Database, which includes data on more than a million surveyed households in a set of countries representing more than 90 percent of the world's population. The national CGE models are able, on their own, to provide estimates of the effects of the unilateral reform of agricultural policies or of all merchandise trade- distorting policies in each relevant country. To estimate the effects of the policies 12 Agricultural Price Distortions, Inequality, and Poverty of other countries, however, a national modeler requires input from a global model. We have decided to use the World Bank Linkage model for this purpose. It is based on version 7 of the GTAP Database on global protection that is calibrated to 2004 except that we have replaced the applied agricultural tariffs for developing countries in the GTAP Database with the more comprehensive set of estimates of distortion rates derived during the World Bank's research project, Distortions to Agricultural Incentives, as collated by Valenzuela and Anderson (2008), which are also calibrated to 2004.5 The latter distortion estimates suggest that, despite reforms over the past 25 years, there was still a considerable range of price distor- tions across commodities and countries in 2004, including a strong antitrade bias in national agricultural policies in many developing countries, plus considerable nonagricultural protection in some developing countries (see table 1.2 above). There are various ways to transmit the results derived from a global CGE model such as Linkage to a single-country CGE model. Like Hertel and Winters (2006), we adopt the approach developed by Horridge and Zhai (2006). For imports, Horridge and Zhai propose the use of border price changes from the global model's simulation of rest-of-the-world liberalization (that is, without the developing country that is the focus). For the exports of the developing country of interest, the shift in the export demand curve following liberalization in the rest of the world is given as percentage changes, as follows: x (1 ).q, (1.1) where x is the percentage vertical shift in the export demand curve; is the elas- ticity of substitution between the exports of country i and the exports of other countries; and q is the percentage change in the quantity of exports under the sce- nario of liberalization in the rest of the world, excluding the country of focus. All the CGE models used in the present study are applied in the comparative static mode, and they assume constant returns to scale and perfectly competitive product markets and homogeneous firms. In all cases other than the exceptional case of South Africa and, to a much smaller extent, the cases of Argentina and Nicaragua, unemployment is assumed to be unaffected by changes in the trade policy regime. These assumptions are imposed simply because of insufficient empirical evidence for the use of alternative assumptions across all the countries modeled in our study. This application of a standard set of assumptions reduces the risk that differences in the results across countries are driven by differing assumptions about investment behavior, productivity growth, the degree of monopolistic competition or firm heterogeneity, economies of scale, or the aggre- gate employment response to changes in trade policy (see Helpman, Itskhoki, and Redding 2009). Our workhorse specifications almost certainly lead to underesti- mations of the welfare gains that would accrue from trade reform, however. In Introduction and Summary 13 particular, without dynamics, the models will not generate a growth dividend from the freeing up of markets or from the eventual productivity gains from trade. This div- idend may be substantial.6 Moreover, because economic growth is the predominant way poverty is reduced in developing countries (see the literature review in Ravallion 2006), the absence of dynamics implies that the results of this study will grossly underestimate the potential poverty-reducing consequences of liberalization and might, in some situations, indicate poverty increases when, in fact, they would be decreases had the growth consequences been incorporated. All the country case studies and two of the global modeling studies presented in this volume make use of household survey data, in addition to a social account- ing matrix. The matrix is the basis for the data in the CGE model, while the house- hold survey data are used in microsimulation modeling. Typically, the experiments are performed in two stages. The first stage involves the imposition on the national CGE model of the policy shock (either unilateral liberalization, or an exogenous shock to border prices and export demand pro- vided by the Linkage model). This generates changes in domestic product and fac- tor markets. The consequent changes in consumer and factor prices are then transmitted to the microsimulation model to determine how they alter the earn- ings of various household types (according to the shares of household income generated by the various factors) and the cost of living of these households (according to the shares of their expenditure on the various consumer products). In turn, this provides information on changes in the distribution of real house- hold incomes and, hence, in inequality, as well as in the number of people living below any poverty line, such as the US$1-a-day extreme poverty line. In all the country case studies, a common set of simulations has been run to compare the inequality and poverty effects of the own-country versus rest-of-the- world policies affecting the markets for agricultural goods (including lightly processed food) relative to the markets for other merchandise. The precise nature of the rest-of-the-world simulation, which employs the global Linkage model, is made clear in the next chapter and in the appendix. The other two global studies in part II use the same 2004 global protection data set, but rely on a different global model that includes national household survey data for microsimulations. In most cases, additional simulations have also been run, often to illustrate the sensitivity of the results to key assumptions pertinent to the particular case study. One assumption that the contributors to Hertel and Winters (2006) find impor- tant and that is confirmed in our study as well revolves around the nature of the change in raising tax revenue required to make up for the loss in tariff revenue associated with trade policy reform. Although the models used in this study are all standard, perfectly competitive, constant-returns-to-scale, comparative static, economy-wide CGE models, they 14 Agricultural Price Distortions, Inequality, and Poverty nonetheless differ somewhat so as to capture important realities, such as labor market characteristics or data limitations, in the particular national settings. How- ever, to ensure the comparability of the models within this volume, we have aimed at conforming the models to a common set of factor market assumptions and clo- sure rules, in addition to our use of 2004 as the base year for the data and under- taking a common set of simulations with the same global distortions data set. We know from trade theory that factor market assumptions are crucial deter- minants of the income distributional effects of trade policies; so, all modelers have assumed the following: (1) a fixed aggregate stock of factors (including no inter- national mobility in labor or capital or in international technology transfers), with the exception of labor in the studies on Argentina, Nicaragua, and South Africa, in which some aggregate employment responsiveness to trade policy is allowed because of the high level of unemployment in the baseline; (2) possibly some sector-specific capital and labor, but most capital and labor types are assumed to be intersectorally mobile and have a common, flexible rate of return or wage, except in the case of Argentina, in which the labor market is modeled with a switching regime between employment or wage adjustments; and (3) land is assumed to be specific to the agricultural sector, but mobile across crop and livestock activities within the sector. The key agreed macroeconomic closure rules that the authors have aimed to adopt in each case study are a fixed current account in foreign currency (so as to avoid foreign debt considerations) and fixed real government spending and fiscal balance (so as not to affect household utility other than through traceable changes in factor and product prices and taxes). Fiscal balance is achieved by using a uni- form (generally direct income) tax to replace the net losses in revenue caused by the elimination of sectoral trade taxes and subsidies. Technologies are also assumed to be unchanged by reform; so, no account is taken of any dynamic gains arising from the opening of trade and the prospective impacts on poverty and inequality generated by more rapid productivity growth. Synopsis of Empirical Findings: Global Model Results This section summarizes the results of the three global models (denoted Linkage, GIDD, and GTAP). The subsequent section then brings together the results from the 10 more-detailed national case studies. Finally, we draw together the lessons learned from both sets of analyses. It would be surprising if all the studies came to the same conclusions, but the strength of this blend of somewhat different global and national models is the fact that it is more likely to expose the various determi- nants of the measured effects in different settings than would be the case if only a single type of model were employed. Introduction and Summary 15 Linkage model results Chapter 2, by Anderson, Valenzuela, and van der Mensbrugghe, sets the scene for the rest of the book in that it uses the World Bank's global Linkage model (see van der Mensbrugghe 2005) to assess the market effects of the world's agricultural and trade policies as of 2004. It serves two purposes. One is to provide the basis for estimating, in each of the 10 country case studies in parts III­V, the effects of rest-of-the-world policies on the import and export prices and the demand for the various exports of any one developing country. The details of these results are reported in the appendix. The other purpose of chapter 2 is to provide estimates of various economic effects on individual countries and country groups so as to be able to say something about international inequality (in the Milanovic [2005] sense, that is, taking into account the economic size of countries) and poverty (using a simple elasticities approach). The Linkage model results reported in chapter 2 suggest that developing coun- tries would gain nearly two times more than high-income countries in welfare terms if 2004 agricultural and trade policies were removed globally (an average welfare increase of 0.9 percent in developing countries compared with 0.5 percent in high-income countries; see table 1.3, column 1). Thus, in this broad sense of a world of only two large country groups, completing the global trade reform Table 1.3. The Linkage Model's Effects of Full Global Liberalization of Agricultural and All Merchandise Trade on Economic Welfare and Real GDP, by Country and Region (percent change relative to benchmark data) Policies in Agricultural policies Policies in all sectors all sectors Economic Agric Nonag Agric Nonag Region welfare (EV) GDP GDP GDP GDP East and South Asia 0.9 0.3 0.7 0.5 2.9 China 0.2 2.8 0.2 5.7 3.0 India 0.2 6.1 1.4 8.3 0.3 Africa 0.2 0.1 0.8 0.9 0.0 Latin America 1.0 36.3 2.8 37.0 2.3 All developing countries 0.9 5.4 1.0 5.6 1.9 Eastern Europe and Central Asia 1.2 4.4 0.3 5.2 0.3 All high-income countries 0.5 13.8 0.2 14.7 0.1 World total 0.6 1.0 0.4 1.2 0.5 Source: Linkage model simulations of Anderson, Valenzuela, and van der Mensbrugghe (chapter 2). Note: EV equivalent variation in income. Agric agricultural. Nonag nonagricultural. 16 Agricultural Price Distortions, Inequality, and Poverty process would reduce international inequality.7 The results vary widely across developing countries, however, ranging from slight losses in the case of some South Asian and Sub-Saharan African countries that would suffer exceptionally large adverse changes in the terms of trade to an 8 percent increase in the case of Ecuador (of which the main export item, bananas, is currently facing heavy dis- crimination in the markets of the European Union, where former colonies and least developed countries enjoy preferential duty-free access).8 Because three-quarters of the world's poorest people depend directly or indi- rectly on agriculture for their main incomes and because farm sizes are far larger in high-income countries than in developing countries, chapter 2 also looks at the extent to which agricultural and trade policies in place as of 2004 reduced the rewards of farming in developing countries and thereby added to international inequality in farm incomes.9 It finds that net farm incomes in developing coun- tries would rise by 5.6 percent, compared with 1.9 percent in nonagricultural value added, if these policies were eliminated (see table 1.3, the final two columns). This suggests that the inequality between farm and nonfarm house- holds in developing countries would fall. In contrast, in high-income countries, net farm incomes would fall by 15 percent on average, compared with a slight rise in real nonfarm value added, that is, inequality between farm and nonfarm house- holds in high-income countries would probably increase.10 However, inequality between farm households in developing countries and those in high-income countries would decline substantially. These inequality results would not be so different if only agricultural policies were to be removed (see table 1.3, columns 2 and 3), underscoring the large magnitude of the distortions caused by agricultural trade policies relative to the effects of nonagricultural trade policies. Chapter 2 also reports that unskilled workers in developing countries--the majority of whom work on farms--would benefit most from reform (followed by skilled workers and then capital owners): the average change in the real unskilled wage over all developing countries would be a rise of 3.5 percent. How- ever, the most relevant consumer prices for the poor, including those many poor farm households and other rural households that earn most of their income from their labor and are net buyers of food, are the prices relating to food and clothing. Hence, if we deflate by a food and clothing price index rather than the aggregate consumer price index, we obtain a better indication of the welfare change among the poor. As shown in the final column of table 1.4, the real unskilled wage over all developing countries would show a rise of 5.9 percent if we use the food and clothing deflator, that is, the inequality in real incomes between unskilled wage earners and the much wealthier owners of (human or physical) capital within developing countries would likely be reduced if there were full trade reform. Introduction and Summary 17 Table 1.4. The Linkage Model's Effects of Full Global Merchandise Trade Liberalization on Real Factor Prices, by Country and Region (percent change relative to benchmark data) Nominal change deflated Real change in unskilled by the aggregate CPI wages, deflated Skilled Capital user Land user By aggregate By food By food and Region wages costa costa CPI CPI clothing CPI East and South Asia 3.4 3.0 1.8 3.2 4.6 4.8 Africa 4.7 4.3 0.1 4.4 5.8 6.9 Latin America 1.4 1.9 21.1 4.5 2.4 4.1 All developing countries 3.0 2.9 1.6 3.5 5.5 5.9 Eastern Europe and Central Asia 3.2 2.6 4.5 1.7 4.2 4.5 High-income countries 1.0 0.5 17.9 0.2 3.3 3.3 World total 1.3 1.2 3.1 0.9 3.6 3.8 Source: Anderson, Valenzuela, and van der Mensbrugghe (chapter 2). Note: CPI consumer price index. a. The user cost of capital and land represents the subsidy-inclusive rental cost. The results on real factor rewards and net farm incomes suggest that poverty, as well as international and intra-developing-country inequality, might be reduced globally by agricultural and trade policy liberalization. The authors of chapter 2 take a further step to assess the impacts of reform on poverty explicitly even though the Linkage model has only a single representative household per country. They do so using the elasticities approach, which involves taking the estimated impact on real household income and applying an estimated income to poverty elasticity to estimate the impacts on the poverty headcount index for each coun- try. They focus on the change in the average wage of unskilled workers, deflated by the food and clothing consumer price index and assume that these workers are exempt from the direct income tax imposed to replace the lost customs revenue following trade reform (a realistic assumption for many developing countries). Table 1.5 reports that, under the full merchandise trade reform scenario, extreme poverty--the number of people surviving on less than US$1 a day-- would drop by 26 million people in developing countries relative to the baseline level of slightly less than 1 billion, a reduction of 2.7 percent. The proportional reduction is much higher in China and Sub-Saharan Africa, falling in each by around 4 percent. It is even higher in Latin America (7 percent) and in South Asia outside India (10 percent). In contrast, the number of extreme poor in India is 18 Table 1.5. The Linkage Model's Effects of Full Global Merchandise Trade Liberalization on the Number of Extreme Poor, by Region Change in Change in number of number of poor relative poor relative Baseline to baseline to baseline Real change headcount, % New levels, US$1 a day New levels, US$2 a day levels, millions levels, % in average unskilled US$1 US$2 Headcount, Total poor, Headcount, Total poor, US$1 US$2 US$1 US$2 Region wage,a % a day a day % millions % millions a day a day a day a day East Asia 4.4 9 37 8 151 34 632 17 52 10.3 7.6 China 2.1 10 35 9 123 34 440 5 12 4.0 2.7 Other East Asia 8.1 9 50 6 29 42 192 12 40 30.1 17.1 South Asia 1.9 31 77 32 454 78 1,124 8 8 1.8 0.7 India 3.8 34 80 36 386 82 883 15 15 4.2 1.7 Other South Asia 4.0 29 94 26 68 92 241 8 7 9.9 2.7 Sub-Saharan Africa 5.3 41 72 39 287 70 508 11 14 3.8 2.7 Latin America 4.1 9 22 8 44 21 115 3 6 6.8 4.7 Middle East and North Africa 14.3 1 20 1 3 13 40 2 19 36.4 32.7 Developing-country total 5.9 18 48 18 944 46 2,462 26 87 2.7 3.4 Excluding China 6.5 21 52 20 820 50 2,022 21 74 2.5 4.7 Eastern Europe and Central Asia 4.5 1 10 1 4 9 43 0 4 6.8 8.0 Source: Anderson, Valenzuela, and van der Mensbrugghe (chapter 2). a. Nominal unskilled wage deflated by the food and clothing consumer price index. Introduction and Summary 19 estimated to rise, by 4 percent.11 This follows from the estimated decline in overall income in India following trade liberalization that is noted in table 1.3. Under the more moderate definition of poverty (people living on no more than US$2 per day), the number of poor in developing countries would fall by nearly 90 million compared to an aggregate baseline level of slightly less than 2.5 billion in 2004, or by 3.4 percent (although the number of people in India living on less than US$2 a day would still increase, but by only 1.7 percent). The GIDD model results Bussolo, De Hoyos, and Medvedev, in chapter 3, make direct use of the global CGE Linkage model, but combine it with the newly developed GIDD tool (Bussolo, De Hoyos, and Medvedev 2008). The GIDD is a framework for ex ante analysis of the effects on income distribution and poverty of changes in macroeco- nomic, trade, and sectoral policies or trends in global markets. It thus offers an alternative to the elasticity approach adopted in chapter 2. It complements a global CGE analysis by providing global microsimulations based on standardized house- hold surveys. The tool pools information from most of the currently available household surveys covering 1.2 million households in 73 developing countries. Information on households in developed countries and Eastern European transi- tion economies completes the data set. Overall, the GIDD sample countries cover more than 90 percent of the world's population.12 In contrast to the modeling approach used in chapter 2, the GIDD approach is able to distinguish between farm and nonfarm households by examining the employment of heads of household. However, because of differences across surveys in the coverage of sources of house- hold income, the database does not provide precise information on the sources of income for each household, and behind the data is the assumption that the propor- tional change in household incomes is driven only by changes in wages. The key inputs in the microsimulation model are the results on changes in incomes from labor that have been obtained through the use of a variation on the Linkage model that assumes full labor mobility.13 Two liberalization scenar- ios are examined: first, the full liberalization of the markets for agricultural prod- ucts and lightly processed food without the liberalization of nonfarm goods mar- kets and, second, the full liberalization of the markets for all goods. Neither scenario is accompanied by large effects on global poverty according to the GIDD. The results summarized in table 1.6 show the incidence of extreme poverty (US$1 a day) rising by 1.0 percent (0.5 percent each from full global trade reform in the farm sector and the nonfarm sector). This increase in poverty is largely caused by the increase in poverty in South Asia, where the number of poor people rises by 3.9 percent after complete global trade reform, a result 20 Agricultural Price Distortions, Inequality, and Poverty Table 1.6. The GIDD Model's Effects of the Removal of Agricultural and All Merchandise Trade Distortions on the Number of Extreme Poor, by Region (millions and percent) Change in number of poor after global trade reform Agricultural All merchandise Share of reform only trade reform global Region poverty, % millions % millions % Extreme poverty, US$1 a day East Asia 24 6.4 2.8 6.3 2.8 South Asia 50 15.4 3.3 18.2 3.9 Sub-Saharan Africa 21 1.0 0.5 0.5 0.3 Latin America 4 2.8 6.9 3.5 8.7 Globala 100 5.0 0.5 8.9 1.0 Moderate and extreme poverty, US$2 a day East Asia 33 12.8 1.6 13.2 1.7 South Asia 46 3.6 0.3 2.0 0.2 Sub-Saharan Africa 14 0.1 0.0 1.1 0.3 Latin America 4 4.8 4.6 5.7 5.4 Globala 100 22.1 0.9 19.8 0.8 Source: Bussolo, De Hoyos, and Medvedev (chapter 3). a. Includes Middle East and North Africa, Eastern Europe and Central Asia, and high-income countries, which, together, account for no more than 2 percent of the world's poor. similar to the result reported in chapter 2. Moderate poverty (US$2 a day), on the other hand, is projected to fall by a similar amount (0.9 percent because of agricultural reform alone and 0.8 percent if nonfarm reform is included). These small aggregate global changes are produced by a combination of offset- ting trends among farm and nonfarm households (table 1.7). At the US$1-a-day extreme poverty level, global liberalization would raise the share of agricultural households among the world's total poor households by one percentage point (from 76 to 77 percent). It would also increase the incidence of poverty among the world's agricultural households (from 32 to 33 percent), while the incidence of poverty among the world's nonfarm households would drop slightly, to 8 percent. However, at the moderate poverty line of US$2 a day, both agricultural and all merchandise trade liberalization would globally lower the incidence of poverty by nearly 1 percent, and it would reduce poverty among farm and nonfarm house- holds (compare table 1.7, last two columns). There are several possible explanations for the differences between the signs of the effects of reforms on extreme poverty in chapters 2 and 3. First is that the Introduction and Summary 21 Table 1.7. The GIDD Model's Effects of the Removal of Agricultural and All Merchandise Trade Distortions on Global Poverty and Inequality, Farm and Nonfarm Households US$1- US$1- US$2- US$2- Real average a-day a-day a-day a-day Gini monthly income, poverty poverty poverty poverty Indicator coefficient 2000 US$ PPP incidence share incidence share Initial levels Agricultural 0.45 65 31.5a 76a 73.8a 70a a Nonagricultural 0.63 320 8.3 24a 26.7a 30a All households 0.67 204 18.9a 100a 48.2a 100a Agricultural liberalization, change from baseline, percentage points Agricultural 0.7 1.1a 0.86 1.1 0.86 0.5 Nonagricultural 0.1 0.2a 0.29 1.1 0.90 0.5 All households 0.1 0.3a 0.23 0.0 0.88 0.0 All merchandise trade liberalization, change from baseline, percentage points Agricultural 0.8 0.8a 1.09 1.0 0.66 0.6 Nonagricultural 0.2 0.4a 0.19 1.0 0.95 0.6 All households 0.0 0.4a 0.39 0.0 0.82 0.0 Source: Bussolo, De Hoyos, and Medvedev (chapter 3). Note: PPP purchasing power parity. a. Expressed in percentages. GIDD poverty data refer to 2000, whereas the Linkage poverty numbers relate to 2004. A large share of the developing-country population was bunched around the extreme poverty line in 2000 (see chapter 3, figure 3.1), but, by 2004, poverty had shrunk quite a bit, at least in East Asia. Second, the GIDD results are based on changes in labor incomes only rather than on changes in incomes from all factors of production. In particular, by not including the effect on nonlabor (especially land) incomes, the study presented in chapter 3 understates the poverty-reducing impacts of trade reform on farm households, thereby contributing to its finding that extreme poverty among farm households would increase. Third, the assump- tion of full labor mobility implies that unskilled farm workers share in the gains from increased agricultural prices, along with their nonfarm counterparts, because less labor is required to migrate to nonfarm jobs. The GIDD results suggest that there might also be a considerable change in inequality following global trade reform. Indeed, table 1.7 shows that agricultural incomes would increase by twice as much as nonfarm household incomes in the all- goods reform scenario (0.8 compared with 0.4 percent) and by five times as much in the reform in agriculture only scenario (1.1 compared with 0.2 percent). While this reduction in the nonagricultural income premium on its own would reduce inequality, income dispersion within the agricultural sector would also increase 22 Agricultural Price Distortions, Inequality, and Poverty Table 1.8. The GIDD Model's Effects of the Removal of Agricultural and All Merchandise Trade Distortions on Global Poverty and Inequality, Farm and Nonfarm Households, by Region (percentage point change) Agricultural reform only All merchandise trade reform US$1 US$2 US$1 US$2 Gini a day, a day, Gini a day, a day, Region coefficient headcount headcount coefficient headcount headcount East Asia 0.72 0.38 0.76 0.62 0.37 0.78 South Asia 0.82 1.16 0.27 0.81 1.37 0.15 India 1.01 1.49 0.33 1.04 1.71 0.26 Other South Asia 0.22 0.06 0.09 0.02 0.21 0.17 Africa 0.04 0.23 0.02 0.06 0.11 0.25 Latin America 0.51 0.61 1.06 0.65 0.77 1.26 World 0.10 0.23 0.88 0.00 0.39 0.82 Source: Bussolo, De Hoyos, and Medvedev (chapter 3). a. Weighted averages across the selected countries in each region. given the differences in the impacts of reform on income distribution across developing-country regions, such that the final change in global inequality would be close to zero (table 1.7, column 1). Chapter 3 also provides results on poverty and inequality at the national and regional levels; these are summarized in table 1.8. Improvements (that is, reduc- tions) in these indicators are pervasive among the 19 countries of Latin America and the Caribbean and the 5 East Asian countries. There are far fewer African examples of improvements, regardless of whether the reform scenario is only in agriculture or also includes nonfarm goods, but most of the changes in the indica- tors on these countries are close to zero. It is mainly in India where extreme poverty--but not moderate poverty--would worsen according to the GIDD results; this was also the finding derived in the study in chapter 2 through the application of poverty elasticities directly to the Linkage model results. The impact of agricultural reform on poverty in India is important, but the existing evidence is quite mixed. In an econometric analysis of historical data, Topalova (2007) concludes that the reductions in agricultural protection associ- ated with India's tariff reforms of the 1990s increased national poverty. In con- trast, using detailed information on household incomes and expenditures, Cai, de Janvry, and Sadoulet (2008) conclude that 70 percent of the farmers in India (86 percent of those with less than 0.2 hectares, 73 percent of those with between 0.2 and 1 hectare, and 49 percent of those with more than 1 hectare) would have lost from increases in the prices of staple foods during 2007­08. According to their data, this is because even the smallest farmers receive only about half their Introduction and Summary 23 incomes from farming. Unfortunately, we have been unable to locate an appropri- ate model or commission a study for India to include in part III. One needs to bear in mind that the GIDD microsimulation model assumes that changes in total household incomes are proportional to the changes in the wage rates for the agricultural or nonagricultural labor of the households. While labor income is the most important source of income for households at or near the poverty line, it is not the only one. Thus, accounting for changes in other fac- tor returns may yield somewhat different results, especially in terms of inequality. In the remaining studies under review, all sources of income are taken into account so that we may provide more reliable national results, albeit for a smaller sample of developing countries. GTAP model results Hertel and Keeney, in chapter 4, draw on the widely used global economy-wide GTAP model. The model adopts the same price distortions as the other studies in this volume and runs the same scenarios, but generates its own world price changes from the GTAP model for scenarios of multilateral trade reform scenar- ios. These price changes alter the border prices in the various countries in the GTAP model, a subset of which has associated detailed household survey data. This allows the authors to make statements about poverty impacts across a range of diverse economies using an internally consistent framework that represents an alternative to the framework used by Bussolo, De Hoyos, and Medvedev in chap- ter 3. While the number of countries on which household survey data are available to Hertel and Keeney is much smaller, the income data are richer, allowing the authors to capture the distributive effects of all changes in factor incomes rather than restricting their examination only to labor income shocks, as in chapter 3. The multicountry study described in chapter 4 focuses on 15 developing countries: four African (Malawi, Mozambique, Uganda, and Zambia), five Asian (Bangladesh, Indonesia, the Philippines, Thailand, and Vietnam), and six Latin American countries (Brazil, Chile, Colombia, Mexico, Peru, and República Bolivariana de Venezuela). Overall, it concludes that the removal of current agricultural and trade policies globally would tend to reduce poverty, primarily via the agricultural reforms (table 1.9). The unweighted average for all 15 devel- oping countries is a decline in the headcount for extreme poverty (US$1 a day) of 1.7 percent. The average fall for the Asian subsample is two times higher, however, and nearly two-thirds of the world's extremely poor people live in Asia (although the sample does not include China and India). The results of Hertel and Keeney on specific countries indicate that, in the sample, the greatest reduc- tion in poverty would occur among the exporters of agricultural products, namely, Chile, Thailand, and Vietnam (table 1.9, column 3). The majority of the 24 Agricultural Price Distortions, Inequality, and Poverty Table 1.9. The GTAP Model's Effects of Full Global Liberalization of Agricultural and All Merchandise Trade on the Number of Extreme Poor, by Country (percentage point change in US$1-a-day poverty) Default tax replacement Alternative tax replacementa Agricultural Nonagricultural All merchandise All merchandise Country reform only reform only trade reform trade reform Asia Bangladesh 0.3 0.5 0.3 5.3 Indonesia 1.1 0.5 0.6 5.2 Philippines 1.4 0.4 1.0 6.4 Thailand 11.2 0.9 10.3 28.1 Vietnam 0.5 5.3 5.7 23.6 Africa Malawi 1.6 0.3 1.9 5.6 Mozambique 1.2 0.2 1.0 4.3 Uganda 0.0 0.1 0.1 6.0 Zambia 0.0 0.1 0.1 2.0 Latin America Brazil 2.5 0.4 2.2 10.0 Chile 4.8 0.1 4.6 12.3 Colombia 0.7 0.6 0.1 4.1 Mexico 0.8 0.4 1.1 0.5 Peru 0.6 0.2 0.8 5.2 Venezuela, R.B. de 0.2 0.7 0.9 2.1 Unweighted average Asia 2.9 0.6 3.5 13.7 Africa 0.7 0.1 0.7 4.5 Latin America 1.3 0.3 1.0 5.7 All 15 developing countries 1.7 0.1 1.7 8.0 Source: Hertel and Keeney (chapter 4), table 4.5. a. The poor are exempt under the alternative tax replacement. 15 countries would experience small increases in poverty after nonagricultural reforms, although the unweighted average across the 15 countries suggests that there would be a slight decrease, primarily because of a significant decline in Vietnam (table 1.9, column 2). The magnitude of the estimated reduction in extreme poverty in Asia and Latin America is somewhat larger according to Hertel and Keeney (chapter 4) than the average reductions estimated for the same countries by Bussolo, De Hoyos, and Medvedev (chapter 3, table 3.4) using the GIDD model. Hertel and Keeney also estimate a small reduction in poverty in Africa.14 These GTAP results are thus closer to the results of the Linkage model described in chapter 2. Introduction and Summary 25 Hertel and Keeney explore the relative poverty-friendliness of agricultural trade reforms in detail by examining the differential impacts on real after-tax fac- tor returns of agricultural versus nonagricultural trade policy reforms. They extend their analysis to the distribution of households by looking at changes in stratum-specific poverty. They find that the more favorable impacts of agricul- tural reforms are driven by increased returns to the labor of farm households, as well as higher returns to unskilled work off-farm. They also find that the liberal- ization of foodgrain markets accounts for the largest contribution to poverty reduction and that the removal of import tariffs in these commodity markets dominates among the poverty-increasing impacts of the removal of subsidies by high-income countries. The final column of table 1.9 reports the percentage change in the national poverty headcount if the poor are not subject to the rise in the income tax required to replace trade tax revenue following trade reform. This assumption represents a significant implicit income transfer from nonpoor to poor house- holds and thus generates a marked difference in the predicted reduction in poverty. Trade reforms are no longer marginally poverty reducing in most of the 15 cases, but, instead, are poverty reducing in all cases and by a considerable mag- nitude. This lowers the poverty rate by roughly one-quarter in Thailand and Vietnam, for example. Overall, the regional and total average extent of poverty reduction is around four times larger under this scenario relative to the scenario whereby the poor are also assumed to be subject to the income taxes levied to replace lost trade tax revenue. The reduction in the unweighted average poverty headcount in the three regions shown in the final column of table 1.9 is remarkably similar to the reduc- tion in the population-weighted averages found by Anderson, Valenzuela, and van der Mensbrugghe (chapter 2) and reported in table 1.5, above, under a similar tax- replacement assumption: the 17 percent for Asia, excluding China and India, and the 6.4 percent for Latin America found by the latter authors are only slightly above the 14 and 5.7 percent, respectively, of Hertel and Keeney, while the 3.7 per- cent for Sub-Saharan Africa in chapter 2 is only slightly below the 4.5 percent obtained by Hertel and Keeney. Synopsis of the Empirical Findings: National Model Results We turn now to a comparison of the results of the 10 detailed individual coun- try case studies with the results obtained from the global models described above. The features of the national models are summarized in table 1.10.15 Like the three global models, they focus on price-distorting policies as of 2004, even though the database for the CGE models and the household survey data on 26 Table 1.10. Characteristics of the Models in the Global and National Country Studies Protection Household Sectors, Products, Types of labor, Intersectoral Study Chapter data, year SAM,a year survey, year number number number labor mobility Global Linkage 2 2004 2004 None 23 23 2 Yes GIDD 3 2004 2004 circa 2000 23 23 4 Partial GTAP 4 2004 2001 circa 2000 31 6 6 Partial National China 5 2004 2002 2000 48 48 8 Partial Indonesia 6 2004 2000 1999 65 20 2 Yes Pakistan 7 2004 2000­01 2001­02 34 28 3 Farm, no Philippines 8 2004 2000 2000 41 34 2 Yes Thailand 9 2004 2000 2000 65 65 2 Yes Mozambique 10 2004 2003 2002 56 56 8 No South Africa 11 2004 2002 2000 110 110 3 Unemployment Argentina 12 2004 2005 2005 24 26 6 Unemployment Brazil 13 2004 2001 2001 42 52 10 Yes Nicaragua 14 2004 2000 2001 40 40 4 Unemployment Sources: Global and country case studies in parts II­V. a. Social accounting matrix of production and trade data. Introduction and Summary 27 which these models are based typically date back a little earlier in the decade. They all include more sectoral and product disaggregation than the global mod- els and cover multiple types of households and labor. All the national studies include microsimulations that draw on the CGE model results, as in the GIDD and GTAP global models. The results of the national studies on real GDP and household consumption suggest that GDP would increase after full global trade reform, but only by 1 or 2 percent, in all 10 countries (except in Argentina if export taxes are removed dur- ing full liberalization).16 Given falling consumer prices, real household consump- tion would increase by considerably more in most cases. (Argentina would again be the notable exception for the reasons discussed elsewhere below.) Generally, these numbers are a little larger than the numbers generated by the global Linkage model, but they are still usually much lower than would be the case had the authors used dynamic models. Like the global models, the studies therefore underestimate the poverty-reducing benefits of trade reform given the broad con- sensus in the literature that trade liberalization increases growth, which is, in turn, a major contributor to poverty reduction. Comparative tables 1.11 and 1.12 summarize the national results on the inci- dence of extreme poverty and income inequality, respectively, resulting from own- country and rest-of-the-world full global liberalization of either agricultural trade or the trade in all goods.17 Some authors ran only six of the nine simulations shown in these tables, but the authors who ran all nine found that the simulations sum almost exactly (to within one decimal place). We have therefore inferred the three missing results in the other country studies by assuming that the results in the agriculture only and nonagriculture only simulations sum to the results of the reform in the trade for all goods. The inferred numbers are shown in italics in tables 1.11 and 1.12. In each case, the total effects on poverty and inequality are subdivided into rural and urban effects. One should not necessarily expect the unweighted averages of the poverty results for each region to be similar to those generated by Hertel and Keeney (chapter 4) because only half of the 10 countries on which we have case studies are included among the 15 countries sampled by Hertel and Keeney. Nonetheless, the unweighted averages of the national poverty effects for each of the key developing country regions computed by Hertel and Keeney are reported in parentheses in the last 4 rows of table 1.11, panel c so that these may be easily compared with the unweighted regional averages derived from our national case studies. In all but 3 of the 12 comparisons on global liberalization (agriculture, nonagriculture, and all merchandise), the projected regional average reductions in poverty after global liberalization are larger in our sample of national case studies than in the sample of 15 countries examined by Hertel and Keeney. This may mean that the poverty 28 Table 1.11. The Impact of Reform on the Incidence of Extreme Poverty, by Country (percentage point change in the national or US$1-a-day poverty line) a. Rural poverty Agricultural reform only Nonagricultural reform only All merchandise trade reform Base Country study % Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China (US$2/day) 58 0.3 1.4 1.1 0.2 0.5 0.3 0.5 1.9 1.4 Indonesia 29 0.1 1.1 1.1 0.2 3.2 3.3 0.1 4.3 4.4 Pakistan 38 1.4 0.1 1.5 6.2 1.1 7.1 7.6 1.2 8.6 Philippines 49 0.0 0.6 0.3 0.6 0.3 0.2 0.6 0.9 0.1 Thailand 30 0.3 1.6 1.3 3.8 0.7 3.1 3.5 0.9 4.4 Mozambique 36 1.6 0.0 1.6 0.5 1.5 2.0 2.1 1.5 3.6 South Africa 17 0.3 0.3 0.7 0.8 0.0 0.8 1.1 0.4 1.4 Argentina -- -- -- -- -- -- -- -- -- -- Brazil -- -- -- -- -- -- -- -- -- -- Nicaragua 63 0.7 0.3 0.4 0.6 0.3 0.9 1.3 0.0 1.3 b. Urban poverty Agricultural reform only Nonagricultural reform only All merchandise trade reform Base Country study % Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China (US$2/day) 3 0.0 0.0 0.0 0.0 0.1 0.1 0.0 0.1 0.1 Indonesia 12 0.1 0.3 0.4 0.1 1.7 1.8 0.2 2.0 2.2 Pakistan 20 2.4 0.1 2.7 4.7 1.4 3.1 2.3 1.5 0.4 Philippines 19 0.8 0.9 0.2 1.2 0.7 0.3 2.0 1.6 0.1 Thailand 6 0.0 0.8 0.7 3.3 0.2 3.2 3.3 0.6 3.9 Mozambique 37 0.5 0.0 0.5 0.4 1.3 1.7 0.9 1.3 2.2 South Africa 4 0.1 0.2 0.3 0.4 0.0 0.4 0.5 0.2 0.7 Argentina 13 1.3 0.1 1.5 0.4 0.1 0.5 0.9 0.0 1.0 Brazil -- -- -- -- -- -- -- -- -- -- Nicaragua 27 0.3 0.5 0.2 1.0 1.4 0.4 0.7 0.9 0.2 (Table continues on the following page.) 29 30 Table 1.11. The Impact of Reform on the Incidence of Extreme Poverty, by Country (continued) (percentage point change in the national or US$1-a-day poverty line) c. Total poverty Agricultural reform only Nonagricultural reform only All merchandise trade reform Base Country study % Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China (US$2/day) 36 0.2 0.8 0.6 0.1 0.4 0.3 0.3 1.2 0.9 Indonesia 23 0.0 0.8 0.8 0.1 2.7 2.8 0.1 3.5 3.6 Pakistan 31 1.6 0.1 1.8 3.6 1.2 4.6 5.2 1.3 6.4 Philippines 34 0.4 0.6 0.1 0.7 0.3 0.2 1.1 0.9 0.1 Thailand 14 0.1 1.1 0.8 3.5 0.4 3.3 3.4 0.7 4.1 Mozambique 36 1.3 0.0 1.3 0.4 1.4 1.8 1.7 1.4 3.1 South Africa 10 0.2 0.3 0.5 0.6 0.1 0.6 0.8 0.3 1.1 Argentina -- -- -- -- -- -- -- -- -- -- Brazil 31 0.5 2.3 2.8 0.4 0.1 0.5 0.9 2.4 3.5 Nicaragua 41 0.1 0.2 0.3 0.9 0.3 0.6 1.0 0.1 0.9 Unweighted averages Asia 28 0.2 0.7 ( 2.9) 0.8 1.2 0.8 ( 0.6) 2.2 1.5 1.6 ( 3.5) 3.0 Africa 32 0.8 0.2 ( 0.7) 0.9 0.5 0.7 (0.1) 1.2 1.3 0.9 ( 0.7) 2.1 Latin America 36 0.3 1.3 ( 1.3) 1.6 0.7 0.1 (0.3) 0.6 1.0 1.2 ( 1.0) 2.2 All 9 developing countriesa 43 0.4 0.6 ( 1.7) 1.0 0.9 0.7 ( 0.1) 1.6 1.3 1.3 ( 2.6) 2.6 Sources: Studies in parts II­V. Hertel and Keeney (chapter 4), table 4.5, for the numbers in parentheses. Note: R of W rest of the world. The numbers in italics for individual countries have been inferred on the assumption that linearity holds (see the text). The numbers in parentheses are from the GTAP model results and are provided for comparison (see table 1.9). Numbers may not sum because of rounding or interaction effects. -- no estimates are available. a. Excludes Argentina. Introduction and Summary 31 elasticities used in the latter study (and also in the Linkage model, which gener- ated similar results) are too small given the greater possibilities for adaptation reflected in most of the national models.18 The individual country results show that poverty would be reduced in all countries by both global agricultural and, with the exception of the Philippines, nonagricultural liberalization (table 1.11, panel c). If all merchandise trade were liberalized, the extent of the reduction in poverty ranges from close to zero to about 3.5 percentage points, except in Pakistan, where it is more than 6.0 percent- age points.19 If we examine the unweighted averages, we find that a greater share of the reduction in poverty is generated by nonfarm trade reform, with the impor- tant exception of Brazil, where agricultural reform is the major contributor to the significant pro-poor outcome. However, if the average is weighted according to the number of people involved, agricultural reform would dominate, as it does in the results of the global modeling. The extreme Brazil result occurs despite the existence of tariff protection for the country's poor import-competing farmers; it is a consequence of the increase in the demand for unskilled labor following liber- alization, which evidently outweighs the poverty impact of the removal of farm tariffs. The contribution of own-country reforms to the decline in poverty appears to be as important as the rest-of-the-world reforms, on average, although there is considerable cross-country divergence in the extent of this parity in the farm reforms and in the nonfarm reforms. The reduction in poverty is divided into rural and urban sources of reduction in table 1.11, panels a and b. A glance at the final column in these panels reveals that rural poverty declines much more than urban poverty in every case. This is true after farm or nonfarm trade reform, as well as after own-country or rest-of- the-world reform. Since the rural poor are much poorer, on average, than the urban poor, this would lead one to expect trade reform to reduce inequality also (see Bussolo, De Hoyos, and Medvedev, chapter 3, figure 3.1). Indeed, the results shown at the bottom of table 1.12, panel c on this sample of countries indicate that inequality would decline in all three developing-country regions after full trade liberalization of all goods or of only agricultural products in the case of both own-country reform and rest-of-the-world reform. The effect of nonfarm trade reform alone is more mixed, providing another reason trade nego- tiators should not neglect agricultural reform in their discussions. Rest-of-the- world and global agricultural reform both lead to a reduction in inequality in every country in the sample except Thailand (plus Argentina and the Philippines slightly in the case of global reform). Meanwhile, unilateral agricultural reform reduces (or leaves constant) inequality in a small majority of countries; the exceptions are Argentina, China, the Philippines, and Thailand (although the effects are small). Nonfarm global reform increases inequality slightly in only three countries. 32 Table 1.12. The Impact of Reform on the Incidence of Income Inequality, by Country (percentage point change in the Gini coefficient) a. Rural Agricultural reform only Nonagricultural reform only All merchandise trade reform Base Country study % Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China 0.32 0.0 0.2 0.2 0.0 0.0 0.0 0.0 0.2 0.2 Indonesia 0.29 0.0 0.0 0.0 0.1 0.0 0.1 0.1 0.0 0.1 Pakistan 0.26 0.1 0.0 0.1 0.3 0.0 0.3 0.2 0.0 0.2 Philippines 0.43 0.2 0.1 0.1 0.3 0.0 0.1 0.5 0.1 0.2 Thailand 0.33 0.0 0.5 0.5 0.4 0.0 0.4 0.4 0.5 0.9 Mozambique -- -- -- -- -- -- -- -- -- -- South Africa 0.63 0.1 0.1 0.2 0.3 0.0 0.3 0.4 0.1 0.5 Argentina -- -- -- -- -- -- -- -- -- -- Brazil -- -- -- -- -- -- -- -- -- -- Nicaragua -- -- -- -- -- -- -- -- -- -- b. Urban Agricultural reform only Nonagricultural reform only All merchandise trade reform Base Country study % Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China 0.26 0.0 0.1 0.1 0.0 0.1 0.1 0.0 0.0 0.0 Indonesia 0.36 0.0 0.1 0.1 0.3 0.3 0.6 0.3 0.2 0.5 Pakistan 0.40 0.1 0.0 0.1 1.9 0.0 1.9 2.0 0.0 2.0 Philippines 0.48 0.3 0.2 0.1 0.1 0.0 0.1 0.4 0.2 0.2 Thailand 0.15 0.1 0.6 0.7 0.5 0.0 0.5 0.6 0.6 1.2 Mozambique -- -- -- -- -- -- -- -- -- -- South Africa 0.62 0.1 0.1 0.2 0.5 0.0 0.5 0.6 0.1 0.7 Argentina 0.50 0.3 0.1 0.2 0.2 0.1 0.3 0.1 0.2 0.0 Brazil -- -- -- -- -- -- -- -- -- -- Nicaragua -- -- -- -- -- -- -- -- -- -- (Table continues on the following page.) 33 34 Table 1.12. The Impact of Reform on the Incidence of Income Inequality, by Country (continued) (percentage point change in the Gini coefficient) c. Total Agricultural reform only Nonagricultural reform only All merchandise trade reform Base Country study % Unilateral R of W Global Unilateral R of W Global Unilateral R of W Global China 0.44 0.1 0.4 0.3 0.0 0.1 0.1 0.1 0.5 0.4 Indonesia 0.34 0.0 0.1 0.1 0.2 0.2 0.4 0.2 0.1 0.3 Pakistan 0.34 0.1 0.0 0.2 3.2 0.1 3.1 3.3 0.1 3.3 Philippines 0.51 0.3 0.2 0.1 0.1 0.0 0.1 0.4 0.2 0.2 Thailand 0.34 0.1 0.7 0.8 0.4 0.0 0.4 0.5 0.7 1.2 Mozambique 0.48 1.2 0.1 1.3 0.3 0.2 0.1 1.5 0.1 1.4 South Africa 0.67 0.1 0.1 0.2 0.4 0.0 0.4 0.5 0.1 0.6 Argentina -- -- -- -- -- -- -- -- -- -- Brazil 0.58 0.2 1.4 1.6 0.1 0.1 0.0 0.1 1.5 1.7 Nicaragua 0.53 0.1 0.1 0.0 0.1 0.2 0.3 0.2 0.1 0.3 Unweighted averages Asia 0.39 0.1 0.0 0.1 0.5 0.0 0.5 0.4 0.0 0.4 Africa 0.58 0.7 0.1 0.8 0.4 0.1 0.3 1.0 0.0 1.0 Latin America 0.56 0.2 0.7 0.8 0.0 0.2 0.1 0.2 0.8 1.0 All 9 developing countriesa 0.59 0.2 0.2 0.4 0.3 0.0 0.3 0.5 0.2 0.7 Sources: Country case studies in parts III­V. Note: R of W rest of the world. The numbers in italics for individual countries have been inferred on the assumption that linearity holds (see the text). Numbers may not sum because of rounding or interaction effects. -- no data are available. a. Excludes Argentina. Introduction and Summary 35 In Indonesia and Thailand, the inequality-increasing impact of nonfarm reform more than offsets the egalitarian effect of farm trade reform, whereas both types of reform increase inequality in the Philippines and Thailand. Inequality within the rural or urban household groups is not altered much by trade reform relative to overall national inequality (compare table 1.12, panels a and b with panel c). This underlines the point that trade reform would tend to reduce urban-rural inequality predominantly rather than inequality separately within either group. Several of the national studies investigate the impact of reforms that might complement trade reform, notably various approaches to address the elimination of trade tax revenue. If the revenues can be recouped through taxes that do not bear on the poor, then the impact of reform on poverty reduction is more favor- able. The China study focuses on lowering the barriers to migration out of agri- culture by improving the operation of land markets and diminishing the barriers to mobility created by the hukou system, the household registration system in China. Each of these two initiatives, along with international trade liberalization (which increases access to foreign markets), would reduce poverty, and a combi- nation of such measures would benefit all major household groups. Argentina is a special case in several respects. First, Cicowiez, Díaz-Bonilla, and Díaz-Bonilla, the authors of the country study (chapter 12), had access only to an urban household survey; they were thus unable to make any determinations about the effect of policy reform on rural poverty or urban-rural income inequality. Second, Argentina imposed export taxes on farm products in 2002 and has increased the taxes a number of times since then. Removing these taxes as part of a shift toward free trade would clearly benefit farmers and rural areas, but would also raise the price of food in urban areas, which, all else being equal, would tend to increase urban poverty (see the results in table 1.11, panel b). Third, by their assumptions, the authors allow reform to alter aggregate employment, unlike most of the other studies, making their results less comparable. Together, these features mean that the authors have found that global trade reform would reduce urban poverty and inequality in Argentina, but only if export taxes are not included in the reform. If export taxes are eliminated as well, the results in table 1.11, panel b and table 1.12, panel b show that urban inequality would change little, but urban poverty would rise. Although, on its own, nonfarm reform would reduce urban poverty in Argentina, urban poverty would rise because of the strong negative impact on the urban poor of the higher food prices resulting from the removal of export taxes. In a global reform scenario in which export taxes are left unchanged, the authors find that both poverty and inequality would fall in Argentina because there would be less unemployment with respect to the scenario involving the removal of export taxes. 36 Agricultural Price Distortions, Inequality, and Poverty What Have We Learned? As in previous studies, whether based on ex post econometrics (as in Harrison 2007) or on ex ante economy-wide simulation (as in Hertel and Winters 2006), this study also finds that the results are mixed and not easy to summarize, partic- ularly with regard to the poverty effects. There is nonetheless a high degree of sim- ilarity in the most important sign: the direction of the estimated national effect on extreme poverty that would result from freeing all merchandise trade globally. The greatest amount of overlap across the studies occurs in relation to this effect; the signs, summarized in table 1.13, agree in all but one-seventh of our available cases. Moreover, apart from India, there is no case in which the majority of the signs indicate that reform would increase poverty. This beneficial impact of the full liberalization of global merchandise trade on the world's poor would be more readily generated by agricultural reform than by Table 1.13. The Direction of the Effects of Global Reform on Extreme Poverty, by Country (sign of the change in the share of the population living on less than US$1 a day or the national poverty line) Agricultural reform only All merchandise trade reform GIDD GTAP National Linkage GIDD GTAP National Country model model model model model model model Brazil Chile China Colombia India Indonesia Mexico Mozambique Nicaragua Pakistan Peru Philippines South Africa Thailand Uganda Venezuela, R.B. de Vietnam Sources: Country case studies in parts III­V. Note: The table shows the only countries in our study on which results are available from at least two of the models reported in the subsequent chapters. Blank cells indicate that there are no estimates. Introduction and Summary 37 nonagricultural reform and, within agricultural reform, by the removal of the substantial supports provided to farmers in developed countries rather than by policy reform in developing countries. According to the economy-wide models used in our study, such reform would raise the real earnings of unskilled laborers in developing countries, most of whom are working in agriculture. The earnings of these laborers would rise relative to the earnings of unskilled workers in devel- oped countries and relative to the earnings of other income earners in developing countries. In addition to reducing poverty, such reform would thus lower the inequality within developing countries and between developing countries and developed countries. According to the estimates of the Linkage model, the number of extremely poor people in developing countries--people living on less than US$1 a day-- would fall by 2.7 percent after the global opening of all goods markets; the num- ber would decline by 4.0 percent in China and Sub-Saharan Africa, but rise by 4.0 percent in India (or by 1.7 percent if the more moderate US$2-a-day poverty line is used). The GIDD model suggests that the decline in moderate poverty would be less than the decline estimated in the Linkage model and that extreme poverty would actually rise by 1.0 percent globally after full global trade reform (almost all because of a rise in India). However, we should recall that the GIDD model only takes labor income effects into account. The results for the 15 coun- tries in the GTAP model are more in line with the Linkage results. They suggest that, in Asia and Latin America, the poverty-reducing effect of global reform would be twice as large as the estimates in the GIDD model and that, in Africa, there would be a small decline (rather than a small rise) in poverty. The 10 national case studies all find that global trade liberalization is poverty reducing (if the removal of export taxes is not part of the full liberalization in Argentina), regardless of whether the reform involves only agricultural goods or all goods; the benefit would arise roughly equally from reform at home and from reform abroad. The case studies also find that rural poverty would be cut much more than urban poverty in all cases, whether by reform at home or abroad and whether or not the reform includes nonfarm goods. Again according to the Linkage model, global trade liberalization would reduce international inequality between developing countries and high-income countries both in total and among only farm households. However, it cannot be guaranteed that every developing country would be made more well off unless there is a strong economic growth dividend associated with reform (which is not captured in the comparative static modeling used in our study). The message emerging from the GIDD analysis is less optimistic: the model finds that inequality would change little after full global reform (inequality would fall in Latin America, but rise in South Asia). This is mainly because of increased income dispersion within 38 Agricultural Price Distortions, Inequality, and Poverty the agricultural sector, despite a reduction in the gap in farm-nonfarm household incomes. The analysis based on the GTAP model, which reinforces the findings in the Linkage model with respect to poverty, does not cover the inequality effects. The full trade liberalization of all goods or only of agricultural products would also cause inequality to decline after both own-country reform and rest-of-the- world reform within each of the three developing country regions covered by our sample. Inequality within the rural or urban household groups would not alter much following full trade reform, suggesting that the predominant impact of trade reform would be a reduction in urban-rural inequality. The mechanism used by governments to adapt to the fall in tariff revenue is shown to be crucial. If one assumes that, rather than distributing it proportion- ately, governments do not require the poor to bear any of the tax burden in the effort to replace lost trade tax revenue, the estimated degree of poverty reduction is about four times greater in the 15 countries studied in the GTAP model. The results of the three global analyses all indicate that the removal of the remaining agricultural policies would have a much stronger impact on poverty and inequality than would nonagricultural trade reforms. A weighted average across the 10 country case studies would probably show similar results. This con- trasts with the outcomes of reforms over the past three decades according to Valenzuela, van der Mensbrugghe, and Anderson (2009): they estimate that global reforms in nonfarm trade policy between the early 1980s and 2004 boosted the value added in developing-country agriculture by more than twice as much as the global reforms in agricultural policy lowered it, and the former might therefore be expected to have had a dominant impact in any reduction in poverty and inequal- ity over the period. The 10 national case studies also shine some light on the relative importance of domestic versus rest-of-the-world reform in these countries. The contribution of own-country reforms to the fall in poverty appears to be equally as important as the contribution of rest-of-the-world reform, on average, although there is con- siderable cross-country divergence in the extent of the effects of both farm reform and nonfarm reform. Caveats The impacts of agricultural and other trade reforms are complex. Such reforms simultaneously affect product and factor markets, government budgets, and external trade. The studies in this volume provide a broad range of ex ante model- ing perspectives, including global and national models. They devote considerable attention to capturing poverty effects through microsimulation and poverty elas- ticity approaches and to using the same price distortion estimates, the same global Introduction and Summary 39 model for measuring rest-of-the-world border shocks in the 10 national models, and similar behavioral assumptions, tax replacement assumptions, and model closures. Nonetheless, there is ample scope for further exploration of this issue through additional comparisons, including by drilling down to examine the ori- gins of each modeling result. Our space limitations mean that such exploratory work needs to be left to future research efforts. The reforms considered here cover only the liberalization of the trade in goods. Freeing up the global trade in services would also likely produce gains in most national economies, including among farmers. Freeing up capital would add to the gains, as would freeing the international movement of low-skilled labor from developing countries to higher-income countries (Prasad et al. 2007, World Bank 2005). How those reforms would interact with farm and other goods trade reforms in terms of the impacts on global poverty and inequality awaits the devel- opment of more sophisticated global simulation models. Another key challenge that remains is to capture the growth effects of liberal- ization and, in particular, the general equilibrium distributive (poverty and inequality) consequences. This area of research has only recently been addressed in the empirical literature by building on the advances in the theoretical literature in the 1990s on endogenous growth (beginning with Grossman and Helpman 1991). Existing partial equilibrium analyses strongly suggest that the trade- growth-poverty nexus is extremely important, possibly much more important than the static reallocative impacts captured in the current set of studies. There is every reason to believe that, once dynamics are included, they will reinforce the basic finding of our study that agricultural and other merchandise trade policy reforms are poverty and inequality reducing. A further modeling change involves introducing a stochastic dimension so as to capture changes in the probability of falling into poverty. This is important if greater openness alters the risk of food price spikes given, for example, that an upward spike might cause a food-deficit household to starve. Such general equi- librium empirical modeling that contains sufficient sectoral and household detail to be useful for poverty analysis, even without a dynamic component, is still in infancy. However, this field may develop rapidly in response to the demand for climate change studies; an early prototype is Ahmed, Diffenbaugh, and Hertel (2009). There is significant scope also for exploring empirically the possible effects of complementary domestic reforms that might accompany agricultural price and trade policy reforms. This is well illustrated in the China case study, which shows that, if labor market reform were to accompany trade reform, the reduction in poverty would be several times greater. Even in the extreme case of India, trade reform would probably not increase poverty if more efficient transfer mechanisms 40 Agricultural Price Distortions, Inequality, and Poverty were in place and high-payoff infrastructure investments were undertaken. The politics behind implementing first-best domestic policies is not necessarily any less complex than the politics associated with trade policies, however; this under- scores the need for comprehensive political economy analysis that is not limited only to border policy measures.20 Policy Implications The empirical findings described above have a number of policy implications. First, the generally attractive results in terms of the poverty- and inequality- reducing effects of trade policy reforms, whether unilateral or multilateral, pro- vide yet another reason why it is in the interest of countries to seek the further liberalization of national and world markets. Second, a recurring theme in the national case studies is that the gains in terms of poverty and inequality reduction, in addition to the standard aggregate real income gains associated with trade liberalization, are generally much greater after global reform than after only own-country reform. In the Indonesia study, for example, unilateral trade liberalization is expected to reduce poverty only slightly, but liberalization by the rest of the world is expected to lower poverty substan- tially. In the Philippines, domestic reform of the current levels of protection alone may marginally increase poverty rates, whereas rest-of-the-world liberalization would almost fully offset this increase (and more than offset it in the case of only agricultural reform). Third, the results of this set of studies show that the winners in trade reform would overwhelmingly be found among the poorer countries and the poorest individuals within countries. However, it is also clear that, even among the extreme poor, some will lose out. Hence, there is merit in compensatory policies, ideally ones that focus not on private goods, but on public goods that reduce underinvestment in pro-growth factors such as rural human capital. At the national level, India appears to be an important example of a potential loser from global trade reform in terms of welfare, poverty, and inequality. The government of India might therefore consider replacing its current extensive agricultural input subsidies and import tariffs by targeted assistance aimed only at the poor- est farmers and rural areas (which may also help the urban poor, save govern- ment spending on fair price shops to offset the effects of tariffs on food, and reduce the adverse environmental effects of subsidies for irrigation and farm chemicals).21 Fourth, the most substantial benefits would be generated by agricultural reform. This underscores the economic and social importance of securing reforms in the agricultural sector, in addition to manufacturing, notwithstanding Introduction and Summary 41 the political sensitivities involved. Other domestic policy instruments would more directly address the Millennium Development Goals of governments in poverty and hunger reduction and, hence, be more efficient than trade policies in this effort. However, the former generally represent a greater net drain on the treasury, which may be a challenge in low-income countries that still rely heavily on trade tax revenue. One solution to this dilemma involves expanding aid-for-trade fund- ing as part of official development assistance programs. Finally, most of the national case studies find that unilateral policy reform alone may represent a way to reduce poverty and inequality. This suggests that developing countries should not hold back on national reforms while they are negotiating in the World Trade Organization Doha Round or other international forums. It also suggests that, from a poverty-reducing perspective, developing countries have little to gain and, potentially, much to lose by negotiating exemp- tions or delays in national reforms within the framework of World Trade Organi- zation multilateral agreements. Notes 1. For a review of the theoretical literature and the empirical evidence on individual and societal preferences for redistribution, see Alesina and Giuliano (2009). Prasad et al. (2007) make the point that, as the number of the extreme poor decline, concerns about poverty will diminish and be gradu- ally replaced by concerns about income inequality. 2. Political economists are also interested in the ways policies affect the incomes and asset values of various vested interests, but this is not a focus of our analysis. 3. A study by Sala-i-Martin (2006) finds that economies have converged in the sense that dispari- ties in GDP per capita across countries have shrunk in recent decades. Analyses based on household survey data rather than GDP per capita include the studies by Milanovic (2002, 2005, 2006). A recent review of the evidence on global poverty and inequality is available in Ferreira and Ravallion (2008). 4. The distortions database is documented fully in Anderson and Valenzuela (2008). It is based on the methodology summarized in Anderson et al. (2008) and detailed in appendix A of Anderson (2009). 5. We were fortunate to have early access to the P5 preliminary version of the GTAP Database ahead of the final release. Details on the GTAP Database are available in Narayanan and Walmsley (2008). 6. See Wacziarg and Welch (2008) and Krueger (2010), as well as the collection of seminal papers in Winters (2007). Rutherford and Tarr (2002) bring together these ideas using a numerical open- economy growth model. This model allows for product variety, imperfect competition, economies of scale, and international capital flows. It is also dynamic so that the model may be used to trace out an adjustment path to trade reform. It is stochastic in that it draws randomly from uniform probability distributions for eight key parameters of the model. The authors simulate a halving of the only policy intervention (a 20 percent tariff on imports) and thereby fully replace the government's lost tariff rev- enue by a lump-sum tax. This modest trade reform produces a welfare increase (in terms of a Hicksian equivalent variation) of 11 percent of the present value of consumption in the central model. System- atic sensitivity analysis with 34,000 simulations shows that there is virtually no chance of a welfare gain of less than 3 percent and a 7 percent chance of a welfare gain larger than 18 percent of consumption. See also the empirical study of four developing countries in Cockburn et al. (2008). 42 Agricultural Price Distortions, Inequality, and Poverty 7. This would continue a process that was initiated in the 1980s when many countries began to reform trade and exchange rate regimes. Using the same Linkage model and database as the present study, Valenzuela, van der Mensbrugghe, and Anderson (2009) find that the global reforms between 1980­84 and 2004 also boosted economic welfare proportionately more in developing countries than in high-income economies (by 1.0 percent compared with 0.7 percent, respectively). 8. Even so, if one were to treat each of the 60 countries or the groups of countries in the global study described in chapter 2 in terms of a single household (that is, ignoring intracountry inequality), then intercountry income inequality (not taking account of the differing economic size of countries) would be reduced at least slightly, as measured by the Gini coefficient, from 0.8513 to 0.8506. 9. According to data in the FAOSTAT Database, less than 15 million relatively wealthy farmers in developed countries, with an average of almost 80 hectares per worker, are currently being helped, at the expense of not only consumers and taxpayers in these rich countries, but also the majority of the 1.3 billion relatively impoverished farmers and their large families in developing countries, who, on average, must earn a living from only 2.5 hectares per worker. 10. In some high-income economies, however, farm households now have higher incomes than nonfarm households (Gardner and Sumner 2007, OECD 2009). 11. The rise in India is caused partly by the removal of the large subsidies and import tariffs that assist Indian farmers and partly by the greater imports of farm products that raise the border prices of these imports. 12. Information on the GIDD data set, methodology, and applications are available at http:// go.worldbank.org/YADEAFEJ30. 13. While changes in incomes from labor are the most important income change among house- holds at or near the poverty line, accounting for the changes in other sources of income may yield somewhat different results, particularly as they relate to inequality. The results for the Linkage model used in chapter 3 are not identical to those in chapter 2 because, to make their results com- patible with the GIDD, the authors of chapter 3 had to assume that labor is less than fully mobile across sectors. 14. An African comparison is not possible because there was only one African country common to the two sets of sample countries. 15. The 10 national studies cover Argentina (Cicowiez, Díaz-Bonilla, and Díaz-Bonilla; chap- ter 12), Brazil (Ferreira Filho and Horridge; chapter 13), China (Zhai and Hertel; chapter 5), Indonesia (Warr; chapter 6), Mozambique (Arndt and Thurlow; chapter 10), Nicaragua (Sanchez and Vos; chapter 14), Pakistan (Cororaton and Orden; chapter 7), the Philippines (Cororaton, Corong, and Cockburn; chapter 8), South Africa (Hérault and Thurlow; chapter 11), and Thailand (Warr; chapter 9). 16. The results on Argentina are included only in the urban parts of tables 11 and 12 because the relevant household survey does not include rural areas. However, it should be kept in mind that Argentina is the most urbanized developing country in the sample; only 8 percent of the population was living in rural areas in 2007. Hence, even if the poverty effects in rural areas had the opposite sign, they might not offset substantially the results in the urban sector. 17. The data are based on national or US$1-a-day poverty lines, except in the case of China; the results on China are available only at the US$2-a-day poverty line. 18. Hertel and Keeney (chapter 4) use stratum-specific poverty elasticities to map the impacts on poverty of the changes in average incomes from all sources. 19. The Pakistan results have been generated on the assumption that the loss in trade taxes would be offset by a rise in direct income taxes. Only nonpoor urban households pay direct taxes in Pakistan; so, the removal of tariffs would raise the direct taxes paid by the urban nonpoor such that the benefits of trade reform would go mainly to the poor. 20. A beginning has been made in such political econometric analysis in a set of studies reported in Anderson (2010) that use the World Bank agricultural distortions database compiled by Anderson and Valenzuela (2008). Introduction and Summary 43 21. Caution is needed, as always, in proposing such interventions in countries in which it is diffi- cult to implement even the most well intentioned policy. As Pritchett (2009) vividly points out, India especially suffers from this problem of underdevelopment. References Ahmed, S. A., N. S. Diffenbaugh, and T. W. Hertel. 2009. "Climate Volatility Deepens Poverty Vulnera- bility in Developing Countries." Purdue Climate Change Research Center Paper 0921, Purdue Uni- versity, West Lafayette, IN. Also published in Environmental Research Letters 4 (3): 034004 (1­8). Alesina, A. F., and P. Giuliano. 2009. "Preferences for Redistribution." NBER Working Paper 14825, National Bureau of Economic Research, Cambridge, MA. Anderson, K., ed. 2009. Distortions to Agricultural Incentives: A Global Perspective, 1955­2007. London: Palgrave Macmillan; Washington, DC: World Bank. ------, ed. 2010. The Political Economy of Agricultural Price Distortions. New York: Cambridge University Press. Anderson, K., M. Kurzweil, W. Martin, D. Sandri, and E. Valenzuela. 2008. "Measuring Distortions to Agricultural Incentives, Revisited." World Trade Review 7 (4): 1­30. Anderson, K., W. Martin, and D. van der Mensbrugghe. 2006. "Distortions to World Trade: Impacts on Agricultural Markets and Incomes." Review of Agricultural Economics 28 (2): 168­94. Anderson, K., and E. Valenzuela. 2008. "Estimates of Global Distortions to Agricultural Incentives, 1955­2007." Data spreadsheet, October, World Bank, Washington, DC. http://go.worldbank.org/ YAO39F35E0. Atkinson, A. B., and A. Brandolini. 2004. "Global World Inequality: Absolute, Relative or Intermedi- ate." Paper prepared for the 28th Conference of the International Association for Research in Income and Wealth, Cork, Ireland, August 22­28. Bourguignon, F., M. Bussolo, and L. A. Pereira da Silva, eds. 2008. The Impact of Macroeconomic Policies on Poverty and Income Distribution: Macro-Micro Evaluation Techniques and Tools. London: Palgrave Macmillan; Washington, DC: World Bank. Bourguignon, F., V. Levin, and D. Rosenblatt. 2004. "Declining International Inequality and Economic Divergence: Reviewing the Evidence through Different Lenses." Economie Internationale (4Q): 13­26. Bourguignon, F., and C. Morrison. 2002. "Inequality among World Citizens: 1820­1992." American Economic Review 92 (4): 727­44. Bussolo, M., R. De Hoyos, and D. Medvedev. 2008. "Is the Developing World Catching Up? Global Convergence and National Rising Dispersion." Policy Research Working Paper 4733, World Bank, Washington, DC. Cai, J., A. de Janvry, and E. Sadoulet. 2008. "Rural Households in India: Sources of Income and Inci- dence of Burden of Rising Prices." Unpublished report, July, University of California Berkeley, Berkeley, CA. Chen, S., and M. Ravallion. 2008. "The Developing World Is Poorer Than We Thought, but No Less Successful in the Fight against Poverty." Policy Research Working Paper 4703, World Bank, Washington, DC. Cockburn, J., B. Decaluwé, I. Fofana, and V. Robichaud. 2008. "Trade, Growth, and Gender in Develop- ing Countries: Comparison of Ghana, Honduras, Senegal, and Uganda." In Gender Aspects of the Trade and Poverty Nexus: A Macro-Micro Approach, ed. M. Bussolo and R. De Hoyos, 111­63. London: Palgrave Macmillan; Washington, DC: World Bank. Cranfield, J. A. L., P. V. Preckel, J. S. Eales, and T. W. Hertel. 2003. "Estimating Consumer Demands across the Development Spectrum: Maximum Likelihood Estimates of an Implicit Direct Additivity Model." Journal of Development Economics 68 (2): 289­307. Davis, B., P. Winters, and G. Carletto. 2009. "A Cross-Country Comparison of Rural Income-Generating Activities." Unpublished report, January, Agricultural Development Economics Division, Food and Agriculture Organization of the United Nations, Rome. 44 Agricultural Price Distortions, Inequality, and Poverty FAOSTAT Database. Food and Agriculture Organization of the United Nations. http://faostat.fao.org/ default.aspx (accessed June 2008). Ferreira, F. H.G., and M. Ravallion. 2008. "Global Poverty and Inequality: A Review of the Evidence," Policy Research Working Paper 4623, Poverty Team, Development Research Group, World Bank, Washington, DC. Francois, J. F., and W. Martin. 2007. "Great Expectations: Ex Ante Assessments of the Welfare Impacts of Trade Reforms." Unpublished report, January, Erasmus University, Rotterdam. Gardner, B. L., and D. A. Sumner. 2007. "U.S. Agricultural Policy Reform in 2007 and Beyond." Short publication, American Enterprise Institute, Washington, DC. GIDD Database (Global Income Distribution Dynamics Database). World Bank. http://go.worldbank. org/YADEAFEJ30. Grossman, G. M., and E. Helpman. 1991. Innovation and Growth in the Global Economy. Cambridge, MA: MIT Press. Harrison, A. E., ed. 2007. Globalization and Poverty. National Bureau of Economic Research Confer- ence Report Series. Chicago: University of Chicago Press; Cambridge, MA: National Bureau of Economic Research. Helpman, E., O. Itskhoki, and S. J. Redding. 2009. "Inequality and Unemployment in a Global Economy." CEPR Discussion Paper 7375, Centre for Economic Policy Research, London. Hertel, T. W., M. Ivanic, P. V. Preckel, and J. A. L. Cranfield. 2004. "The Earnings Effects of Multilateral Trade Liberalization: Implications for Poverty." World Bank Economic Review 18 (2): 205­36. Hertel, T. W., and L. A. Winters, eds. 2006. Poverty and the WTO: Impacts of the Doha Development Agenda. London: Palgrave Macmillan; Washington, DC: World Bank. Horridge, M., and F. Zhai. 2006. "Shocking a Single-Country CGE Model with Export Prices and Quantities from a Global Model." In Poverty and the WTO: Impacts of the Doha Development Agenda, ed. T. W. Hertel and L. A. Winters, 94­104. London: Palgrave Macmillan; Washington, DC: World Bank. Ivanic, M., and W. Martin. 2008. "Implications of Higher Global Food Prices for Poverty in Low- Income Countries." Policy Research Working Paper 4594, World Bank, Washington, DC. Krueger, A. O. 2010. "Trade Liberalization and Growth in Developing Countries." In Better Living through Economics, ed. J. J. Siegfried. Cambridge, MA: Harvard University Press. Milanovic, B. 2002. "True World Income Distribution, 1988 and 1993: First Calculation Based on Household Surveys Alone." Economic Journal 112 (476): 51­92. ------. 2005. Worlds Apart: Measuring International and Global Inequality. Princeton, NJ: Princeton University Press. ------. 2006. "Global Income Inequality: What It Is and Why It Matters." Policy Research Working Paper 3865, World Bank, Washington, DC. Narayanan, B. G., and T. L. Walmsley, eds. 2008. Global Trade, Assistance, and Production: The GTAP 7 Data Base. West Lafayette, IN: Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University. https://www.gtap.agecon.purdue.edu/databases/v7/v7_doco.asp. OECD (Organisation for Economic Co-operation and Development). 2009. The Role of Agriculture and Farm Household Diversification in the Rural Economy: Evidence and Initial Policy Implications. Report TAD/CA/APM/WP(2009)1/Final (April 21). Paris: OECD. Otsuka, K., J. P. Estudillo, and Y. Sawada, eds. 2009. Rural Poverty and Income Dynamics in Asia and Africa. New York: Routledge. Otsuka, K., and T. Yamano. 2006. "Introduction to the Special Issue on the Role of Nonfarm Income in Poverty Reduction: Evidence from Asia and East Africa." Agricultural Economics 35 (supplement): 373­97. PovcalNet. World Bank, Washington, DC. http://go.worldbank.org/NT2A1XUWP0 (accessed May 2008). Prasad, E. S., K. Rogoff, S.-J. Wei, and M. A. Kose. 2007. "Financial Globalization, Growth, and Volatil- ity in Developing Countries." In Globalization and Poverty, ed. A. E. Harrison, 457­516. National Introduction and Summary 45 Bureau of Economic Research Conference Report Series. Chicago: University of Chicago Press; Cambridge, MA: National Bureau of Economic Research. Pritchett, L. 2009. "A Review of Edward Luce's In Spite of the Gods: The Strange Rise of Modern India." Journal of Economic Literature 97 (3): 771­80. Ravallion, M. 2006. "Looking Beyond Averages in the Trade and Poverty Debate." World Development 34 (8): 1374­92. ------. 2008. "Bailing Out the World's Poorest." Policy Research Working Paper 4763, World Bank, Washington, DC. Ravallion, M., S. Chen, and P. Sangraula. 2007. "New Evidence on the Urbanization of Poverty." Policy Research Working Paper 4199, World Bank, Washington, DC. Rutherford, T. F., and D. G. Tarr. 2002. "Trade Liberalization, Product Variety, and Growth in a Small Open Economy: A Quantitative Assessment." Journal of International Economics 56 (2): 247­72. Sala-i-Martin, X. 2006. "The World Distribution of Income: Falling Poverty and . . . Convergence, Period." Quarterly Journal of Economics 121 (2): 351­97. Topalova, P. 2007. "Trade Liberalization, Poverty, and Inequality: Evidence from Indian Districts." In Globalization and Poverty, ed. A. E. Harrison, 291­336. National Bureau of Economic Research Conference Report Series. Chicago: University of Chicago Press; Cambridge, MA: National Bureau of Economic Research. Valenzuela, E., and K. Anderson. 2008. "Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980­84." Research Memorandum 13 (December), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925. Valenzuela, E., D. van der Mensbrugghe, and K. Anderson. 2009. "General Equilibrium Effects of Price Distortions on Global Markets, Farm Incomes, and Welfare." In Distortions to Agricultural Incen- tives: A Global Perspective, 1955­2007, ed. K. Anderson, chap. 13. London: Palgrave Macmillan; Washington, DC: World Bank. van der Mensbrugghe, D. 2005. "Linkage Technical Reference Document: Version 6.0." December, World Bank, Washington, DC. http://go.worldbank.org/7NP2KK1OH0. Wacziarg, R., and K. H. Welch. 2008. "Trade Liberalization and Growth: New Evidence." World Bank Economic Review 22 (2): 187­231. Winters, L. A., ed. 2007. The WTO and Poverty and Inequality. 2 vols. Critical Perspectives on the Global Trading System and the WTO 13. London: Edward Elgar. World Bank. 2005. Global Economic Prospects 2006: Economic Implications of Remittances and Migration. Washington, DC: World Bank. ------. 2007. World Development Report 2008: Agriculture for Development. Washington, DC: World Bank. Part II GLOBAL CGE APPROACHES 2 Global Welfare and Poverty Effects: Linkage Model Results Kym Anderson, Ernesto Valenzuela, and Dominique van der Mensbrugghe* Despite much reform over the past quarter century in the policy distortions to agricultural incentives, many relevant intervention measures remain active (Anderson 2009). How these policies affect economic welfare, inequality, and poverty levels is an issue of great interest to the agricultural, trade, and develop- ment policy communities in international agencies and in many developing coun- tries. More specifically, for any developing country of interest, how important are its own policies compared with those of the rest of the world in affecting the welfare of the poor in that country? And, given that three-quarters of the world's poor depend on agriculture directly or indirectly for their livelihoods and that, according to an earlier finding, farm policies in 2001 were responsible for more than three-fifths of the global welfare cost of trade distortions, what contributions do agricultural policies contribute to these outcomes (World Bank 2007; Anderson, Martin, and van der Mensbrugghe 2006a, 2006b)? Now is an appropriate time to address this set of questions for at least three reasons. One is that the World Trade Organization is struggling to conclude the Doha Round of multilateral trade negotiations; agricultural policy reform is once again one of the most contentious issues in these talks. Another reason is that *The authors are grateful for the distortion estimates provided by authors of the focus country case studies; for the assistance with spreadsheets by Johanna Croser, Marianne Kurzweil, and Signe Nelgen; and for the helpful comments of workshop participants. 49 50 Agricultural Price Distortions, Inequality, and Poverty poorer countries and their development partners are striving to achieve the United Nations­encouraged Millennium Development Goals by 2015, of which the prime goals are the reduction of hunger and poverty. The third reason is that a new set of estimates of distortions to agricultural incentives in many countries has been brought together recently by the World Bank (Anderson and Valenzuela 2008), and these estimates have since been provided as alternative measures of price distortions for use in computable general equilibrium models (Valenzuela and Anderson 2008). The estimates differ from the usual ones used by trade mod- elers in that they are based on direct domestic-to-border price comparisons rather than on merely applied import tariff rates (as with the Global Trade Analysis Project [GTAP] data set; see Narayanan and Walmsley 2008). This chapter serves two purposes. The first is to offer an analysis of the eco- nomic effects of agricultural price and merchandise trade policies around the world as of 2004 on global markets, net farm incomes, and national and regional economic welfare and poverty, that is, it assesses the effort still necessary to remove the disarray in world agriculture--to use the title of the seminal study by Johnson (1991)--and to provide at least a crude indication of the poverty that might be reduced by such a reform. In doing so, the chapter also serves a second purpose: to explain the origin of the exogenous shocks used in the global model- ing described in the following chapter and in the national modeling studies described in parts III­V that show the market effects on particular developing countries of rest-of-the-world agricultural and trade policies. To quantify the impacts of current policies, we first amend the distortions in version 7 of the GTAP global protection database (Narayanan and Walmsley 2008) by replacing the applied tariffs there with distortion rates that reproduce the distortions estimated by contributors to the World Bank's research project, Distortions to Agricultural Incentives, as collated by Valenzuela and Anderson (2008).1 These distortion estimates suggest that, despite the reforms of the past 25 years, there was still a considerable range of rates across commodities and countries in 2004, including a strong antitrade bias in national agricultural and trade policies in many developing countries. Furthermore, nonagricultural pro- tectionism is still rife in some developing countries, and agricultural price supports in some high-income countries remain high. The present analysis addresses the following questions: To what extent were policies as of 2004 still reducing the rewards of farming in developing countries and thereby adding to inequality in farm household incomes across countries? Are policies depressing value added more in primary agriculture than in the rest of the economies of developing countries? And are they depressing the earnings of unskilled workers more than the earnings of the owners of other factors of pro- duction, thereby potentially contributing to inequality and poverty within these Global Welfare and Poverty Effects: Linkage Model Results 51 developing countries (given that farm incomes are well below nonfarm incomes in most developing countries and that agriculture in these countries is intensive in the use of unskilled labor)? To provide answers to these and related questions, we use our amended GTAP distortion database in a global computable general equilibrium model (the Linkage model; see van der Mensbrugghe 2005). We use the model to assess how agricul- tural markets, factor prices, and the value added in agriculture versus in nonfarm sectors would differ if all price and trade policies that distort the markets for farm and nonfarm goods (as of 2004) were removed. It is important to include nonagri- cultural trade policies in the reform experiment because, as shown in the new study reported in Anderson (2009), these policies may have been more harmful than agricultural policies in depressing farmer incentives in many developing countries. We present the results for the key countries and regions of the world and for the world as a whole. While no one anticipates a complete freeing of global markets in the near future, this prospective analysis serves as a benchmark to appreciate the stakes in terms of further reforms through rounds of multilateral World Trade Organization negotiations. At the same time, by showing how different the trade patterns of various countries would be without distortions, we also provide indica- tions of agricultural comparative advantages in different parts of the world that are more accurate than the perspective offered through an examination of actual trade and self-sufficiency indicators derived in the current distortion-ridden situation. The chapter begins with an examination of the extent of price distortions in 2004 provided by various policy instruments as calibrated in Valenzuela and Anderson (2008). The emphasis is mainly on import tariffs in the case of nonfarm products, but, in the case of agriculture, it is on the full range of production, con- sumption, and trade taxes and subsidies and their equivalent in the case of quan- titative restrictions on markets. This is followed by a description of the Linkage model of the global economy that we use to analyze the consequences of the removal of the distortions. The key results of the two simulations are then pre- sented: the full global liberalization of markets for all merchandise and--to understand the relative contribution of farm policies to these outcomes--the full global reform only of agricultural policies.2 The chapter concludes by highlighting the main messages that emerge from the results. They are as follows: in a shift to free markets, income inequality across countries would be reduced at least slightly; all but one-sixth of the gains to developing countries would be generated by agricultural policy reform; unskilled workers in developing countries--the majority of whom work on farms--would benefit most from reform; net farm incomes in developing countries would rise by 6 percent compared with 2 percent in nonagricultural value added; and the number of people surviving on less than US$1 a day would drop by 3 percent globally. 52 Agricultural Price Distortions, Inequality, and Poverty Key Distortions in Global Markets Border measures have traditionally been the main means by which governments distort prices in their domestic markets for tradable products, given that the rela- tive prices of the various goods are affected by trade taxes or subsidies. Product- specific domestic output subsidies and farm input subsidies have played a more limited role in part because of their much greater overt cost to the treasury.3 To quantify the impacts of current policies, we use the Altertax procedure (Malcolm 1998) to amend the distortions in the prerelease of version 7 of the GTAP global protection database. The amendments relate mainly to developing countries, but, following Anderson and Valenzuela (2007b), we also alter cotton distortions in the United States to reflect the policies there more accurately. To simplify the discussion below, we treat European transition economies (among which we include Turkey) as one of the world's developing-country regions, alongside Africa, Asia, and Latin America.4 Version 7 of the GTAP database includes estimates of bilateral tariffs and export subsidies and of domestic supports as of 2004 for more than 100 countries and country groups spanning the world. As with version 6 of the GTAP data set (which relates to 2001; see Dimaranan 2006), the protection data come from MAcMaps, a joint project of the Centre d'Etudes Prospectives et d'Informations Internationale (Paris) and the International Trade Centre (Geneva). MAcMaps is a detailed database on bilateral import protection at the Harmonized System 6-digit tariff line classification level that integrates trade preferences, specific and compound tariffs, and a partial evaluation of nontariff barriers such as tariff rate quotas.5 The new 2004 version 7 of the GTAP database has lower tariffs than the previous, 2001 version 6 database because of major reforms such as the comple- tion of the implementation of the Uruguay Round agreements and unilateral reforms, including those resulting from the World Trade Organization accession negotiations of China and other recent acceding countries. As noted above, the agricultural price distortion rates in the GTAP version 7 database have been replaced here by an alternative data set for numerous develop- ing countries based on the estimates of the nominal rates of assistance for 2004 contained in Valenzuela and Anderson (2008). The sectoral averages of these amended values are shown in table 2.1. In the case of the amendments to the import tariffs on individual farm products for any particular developing country, the bilateral tariff structure in the GTAP version 7 database is preserved by simply lowering or raising the bilateral tariffs by the same proportion we use to amend the country's average import tariff on each product for 2004. According to this amended data set, the weighted average applied tariff for agri- culture and lightly processed food in 2004 was 21.8 percent for developing coun- tries and 22.3 percent for high-income countries, while, for nonfarm goods, it was Global Welfare and Poverty Effects: Linkage Model Results 53 Table 2.1. Structure of Producer Price Distortions in Global Goods Markets, 2004 (percent) Agriculture and Primary lightly processed Other agriculture food goods Domestic Export Region, country support subsidy Tariff Tariff Africa 0.8 0.1 20.4 11.2 Egypt, Arab Rep. 0.0 0.0 5.0 13.5 Madagascar 0.0 4.4 3.4 2.7 Mozambique 0.2 0.0 14.5 10.9 Nigeria 0.1 0.0 76.1 17.2 Senegal 0.0 1.1 6.2 8.9 South Africa 0.0 0.0 10.2 6.5 Tanzania 0.3 0.0 11.8 13.7 Uganda 0.0 2.6 9.2 5.5 Zambia 0.8 0.0 7.0 9.0 Zimbabwe 3.2 0.0 8.9 15.4 Rest of Africa 1.2 0.3 19.0 13.4 East and South Asia 2.4 0.6 29.6 8.1 China 0.0 0.2 6.5 7.1 Indonesia 0.0 1.6 7.3 4.9 Korea, Rep. 0.0 0.0 319.4 5.9 Malaysia 0.0 0.2 5.0 5.9 Philippines 4.7 0.0 7.1 3.4 Taiwan, China 0.4 0.0 84.2 3.9 Thailand 0.2 0.0 26.2 12.9 Vietnam 3.6 0.5 21.5 18.5 Bangladesh 1.0 0.0 9.9 22.5 India 10.1 2.5 2.9 20.8 Pakistan 0.0 0.2 19.4 18.5 Sri Lanka 0.6 0.3 23.8 5.8 Rest of East and South Asia 0.7 0.0 4.3 2.7 Latin America 0.2 1.4 7.2 6.7 Argentina 0.0 14.8 0.0 5.8 Brazil 0.0 0.0 4.8 8.9 Chile 0.0 0.0 2.4 1.8 Colombia 0.0 0.0 21.6 9.8 Ecuador 0.0 0.0 13.4 10.4 Mexico 1.2 0.0 6.2 3.4 Nicaragua 0.0 2.8 9.6 3.9 Rest of Latin America 1.7 0.3 9.9 9.9 (Table continues on the following page.) 54 Agricultural Price Distortions, Inequality, and Poverty Table 2.1. Structure of Producer Price Distortions in Global Goods Markets, 2004 (continued ) (percent) Agriculture and Primary lightly processed Other agriculture food goods Domestic Export Region, country support subsidy Tariff Tariff Eastern Europe and Central Asia 0.8 0.3 15.9 4.8 Baltic States 3.4 0.0 8.2 0.9 Bulgaria 0.6 0.0 14.8 11.5 Czech Republic 0.6 0.0 3.0 0.5 Hungary 3.1 0.0 6.2 0.5 Poland 0.4 0.0 6.2 0.8 Romania 1.3 0.0 18.0 9.8 Russian Federation 1.7 0.9 18.9 7.4 Slovak Republic 0.0 0.0 5.2 0.4 Slovenia 0.0 0.0 7.8 0.4 Kazakhstan 0.9 0.0 3.4 2.7 Turkey 0.8 0.0 33.3 3.1 Rest of Eastern Europe and Central Asia 1.1 0.9 9.7 5.7 High-income countries 2.6 7.2 22.3 1.2 Australia 0.0 0.0 0.5 3.3 Canada 1.6 3.6 18.9 1.4 EU15 1.2 12.8 6.9 0.7 Japan 2.0 0.0 151.7 1.7 New Zealand 0.0 0.2 0.7 3.3 Rest of Western Europe 2.6 13.4 53.9 2.2 United States 5.2 0.6 6.1 1.3 Developing countries 1.4 0.0 21.8 7.5 Africa 0.8 0.1 20.4 11.2 East Asia 0.3 0.0 41.6 6.7 South Asia 7.2 1.7 6.9 20.2 Latin America 0.2 1.4 7.2 6.7 Middle East 12.4 0.0 7.5 5.7 Eastern Europe and Central Asia 0.8 0.3 15.9 4.8 World total 1.9 3.5 22.1 3.3 Sources: Valenzuela and Anderson (2008) based on calculations compiled by Anderson and Valenzuela (2008). Note: Data are weighted by the value of production at undistorted prices. EU15 the 15 members of the European Union prior to 2004. Global Welfare and Poverty Effects: Linkage Model Results 55 7.5 percent for developing countries and only 1.2 percent for high-income coun- tries. Export subsidies for farm products in a few high-income regions and export taxes in a few developing countries were still in place in 2004, but they are generally small in impact compared with tariffs, as are production subsidies and taxes.6 The averages alone are not necessarily good indicators of the overall distortions to farmer incentives. Also of importance is the composition of each country's trade. Two examples serve to illustrate the point. First, if the tariffs of high-income countries are at a near-prohibitive level for temperate farm products, but are zero on tropical products such as coffee beans, the import-weighted average agricul- tural tariff of these countries may be quite low even if the agricultural value added has been enhanced substantially in these rich countries. Second, the nonagricul- tural primary sector and the farm sector may receive a similar level of import pro- tection (but less than the manufacturing sector), although the former is much more export-focused than agriculture. In this case, trade reform may cause the nonagricultural primary sector to expand at the expense not only of manufactur- ing, but also of farming. Although we have used production rather than trade weights to obtain sectoral average rates of distortion in table 2.1 and although the ratio of agricultural tariffs to the tariffs on other goods in 2004 shown in the table is well above unity in many of the regions, it is not possible to say from those distortion rates alone whether developing-country policies have an antiagricultural bias. Likewise, it is not possi- ble to know how the benefits of the removal of agricultural tariffs in the protective countries would be distributed among the various agricultural exporting coun- tries. To address such issues, a global general equilibrium model is needed to esti- mate the net effects of the distortions in all sectors in all countries in terms of the agricultural markets and net farm incomes of the various nations. We now turn to such a model. The Linkage Model of the Global Economy The model used for this analysis is the World Bank's global computable general equilibrium model known as Linkage (van der Mensbrugghe 2005). For most of this decade, this model has formed the basis for the standard World Bank long-term projections of the world economy, for much of world trade policy analysis, and, more recently, for migration policy analysis (see World Bank 2001, 2003, 2004, 2005, 2006). It is a relatively straightforward computable general equilibrium model, but has some characteristics that distinguish it from other comparative static models such as the GTAP model (described in Hertel 1997). Factor stocks are fixed, which means, in the case of labor, that the extent of unemployment (if any) in the baseline remains unchanged. Producers minimize costs, subject to constant returns to scale in production; consumers maximize utility; and all markets, 56 Agricultural Price Distortions, Inequality, and Poverty including for labor, are cleared with flexible prices. There are three types of produc- tion structures. The crop sectors reflect the substitution possibilities between extensive and intensive farming; the livestock sectors reflect the substitution possi- bilities between pasture and intensive feeding; and all other sectors reflect standard capital-labor substitution. There are two types of labor, skilled and unskilled, and the total employment of each is assumed to be fixed (so there is no change in unem- ployment levels), although both are assumed to be intersectorally mobile. There is a single representative household per modeled region, and this household allocates income to consumption using the extended linear expenditure system. Trade is modeled using a nested Armington structure whereby aggregate import demand is the outcome of the allocation of domestic absorption between domestic goods and aggregate imports, and then aggregate import demand is allocated across source countries to determine the bilateral trade flows.7 Government fiscal balances are fixed in U.S. dollar terms, and the fiscal objective is met by changing the level of lump sum taxes on households. This implies that the losses in tariff revenues are replaced by higher direct taxes on households. The cur- rent account balance is also fixed. Given that other external financial flows are fixed, this implies that ex ante changes to the trade balance are reflected in ex post changes to the real exchange rate. For example, if import tariffs are reduced, the propensity to import increases, and additional imports are financed by boosting export rev- enues. This last step is typically achieved through depreciation of the real exchange rate. Finally, investment is driven by savings. With fixed public and foreign savings, investment occurs through changes in the savings behavior of households and changes in the unit cost of investment. The model only solves for relative prices; the numéraire, or price anchor, is the export price index of manufactured exports from high-income countries. This price is fixed at unity in the base year. A virtue of beginning with the latest GTAP database is that the database includes bilateral tariffs that capture not only reciprocal, but also nonreciprocal preferential trade agreements; the latter provide low-income exporters duty-free access to protected markets in high-income countries. This allows us to take into account the fact that future reform may cause a decline in the international terms of trade in those developing countries that are enjoying preferential access to the agricultural markets and other markets in high-income countries (in addition to those that are net food importers because their comparative advantage lies in other sectors, such as labor-intensive manufacturing). The version of the Linkage model used in our study is based on an aggregation involving 23 sectors and 49 individual countries, plus 11 country groups spanning the world (see the appendix). There is an emphasis on agriculture and food, which account for 16 of the 23 sectors. Note that, consistent with the World Trade Organi- zation, we include the Republic of Korea and Taiwan, China in the developing- country category.8 Global Welfare and Poverty Effects: Linkage Model Results 57 The results below are comparative static results; so, they do not include the (often much larger) dynamic gains that result from an acceleration in investment arising from the reduction in tariffs on industrial goods that lowers the cost of investment. Also missing are therefore any costs of adjustment to reform. More- over, because this version of the Linkage model assumes perfect competition and constant returns to scale, it captures none of the benefits of freeing markets that might be generated by accelerated productivity growth, scale economies, and the creation of new markets (extensification versus intensification). There is also a dampening effect on the estimates of welfare gains because of product and regional aggregation, which hides many of the differences across products in rates of distortions. The results should thus be treated as much lower-bound estimates of the net economic welfare benefits of policy reform.9 The Prospective Effects of the Global Removal of Price-Distorting Policies In this section, to explore the possible outcome of the removal of the policies in force as of 2004, we examine the results of two modeling simulations. The main one involves the full global liberalization of agricultural policies and of trade poli- cies in nonagricultural goods. We also discuss an additional simulation, which involves the global liberalization of agricultural policies only, to give a sense of the relative contribution of farm policies to various outcomes. Global and national economic welfare Beginning with the baseline projection of the world economy in 2004, we remove globally all agricultural subsidies and taxes, plus import tariffs on other merchan- dise (as summarized in table 2.1).10 Our Linkage model suggests that this would lead to a global gain of US$168 billion per year (table 2.2). As a share of national income, developing countries would gain nearly twice as much as high-income countries by completing the reform process (an average increase of 0.9 compared with 0.5 percent, respectively). Thus, in this broad sense of a world of only two large country groups, completing the global reform process would reduce interna- tional inequality, to use the Milanovic (2005) term, taking in to account the eco- nomic size of each country.11 The results vary widely across developing countries, however, ranging from slight losses in the case of some South Asian and Sub- Saharan African countries that would suffer exceptionally large adverse terms of trade changes to 8 percent increases in the case of Ecuador (of which the main export item, bananas, is currently facing heavy discrimination in the markets of the European Union, where former colonies and least developed countries enjoy preferential duty-free access). 58 Agricultural Price Distortions, Inequality, and Poverty Table 2.2. The Impact on Real Income of the Full Liberalization of Global Merchandise Trade, by Country and Region, 2004 (2004 U.S. dollars and percent) Change in annual Change in real income, real income, US$, billions % of benchmark Change Change deriving deriving from from Total change Total terms of income in terms income trade Country and region gain of trade gain effects North Africa and Sub-Saharan Africa 0.9 6.0 0.2 1.1 Egypt, Arab Rep. 0.2 0.6 0.3 0.9 Madagascar 0.0 0.0 0.9 1.2 Mozambique 0.1 0.1 2.4 2.0 Nigeria 0.3 0.6 0.7 1.3 Senegal 0.0 0.1 2.3 4.0 South Africa 0.2 0.7 0.1 0.5 Tanzania 0.0 0.0 0.5 0.4 Uganda 0.0 0.0 0.6 0.1 Zambia 0.0 0.0 0.1 0.3 Zimbabwe 0.1 0.0 3.4 0.5 Rest of Africa 0.5 3.8 0.2 1.5 East and South Asia 29.7 4.9 0.9 0.1 China 3.3 0.5 0.2 0.0 Indonesia 0.5 0.0 0.2 0.0 Korea, Rep. 14.0 0.2 2.8 0.0 Malaysia 4.2 1.0 4.7 1.1 Philippines 0.0 0.5 0.1 0.7 Taiwan, China 1.0 0.0 0.4 0.0 Thailand 3.3 0.1 1.4 0.1 Vietnam 1.9 0.9 5.3 2.5 Bangladesh 0.2 0.8 0.4 1.7 India 0.8 2.9 0.2 0.6 Pakistan 0.1 0.6 0.2 0.8 Sri Lanka 0.8 0.5 5.1 3.1 Rest of East and South Asia 1.9 0.8 1.4 0.5 Latin America 15.8 2.5 1.0 0.2 Argentina 3.2 0.7 2.6 0.6 Brazil 6.8 5.6 1.6 1.3 Chile 0.3 0.2 0.4 0.3 Colombia 2.2 0.7 3.1 1.0 Ecuador 2.0 1.1 8.2 4.4 Mexico 0.7 3.4 0.1 0.6 Nicaragua 0.0 0.0 1.3 0.4 Rest of Latin America 2.0 1.0 0.5 0.3 Global Welfare and Poverty Effects: Linkage Model Results 59 Table 2.2. The Impact on Real Income of the Full Liberalization of Global Merchandise Trade, by Country and Region, 2004 (continued) (2004 U.S. dollars and percent) Change in annual Change in real income, real income, US$, billions % of benchmark Change Change deriving deriving from from Total change Total terms of income in terms income trade Country and region gain of trade gain effects Eastern Europe and Central Asia 14.2 3.6 1.2 0.3 Baltic States 0.5 0.1 1.8 0.3 Bulgaria 0.2 0.2 1.4 1.4 Czech Republic 1.0 0.1 1.4 0.2 Hungary 0.4 0.1 0.6 0.1 Poland 2.0 0.1 1.2 0.1 Romania 0.1 0.7 0.3 1.9 Russian Federation 5.4 3.1 1.2 0.7 Slovak Republic 0.7 0.1 2.3 0.4 Slovenia 0.3 0.1 1.5 0.3 Kazakhstan 0.4 0.2 1.1 0.6 Turkey 1.3 0.5 0.6 0.2 Rest of Eastern Europe and Central Asia 2.2 0.5 2.1 0.4 High-income countries 102.8 11.3 0.5 0.1 Australia 2.4 1.9 0.5 0.4 Canada 0.6 1.2 0.1 0.2 EU15 56.8 3.8 0.7 0.0 Japan 23.1 10.4 0.7 0.3 New Zealand 2.2 1.8 3.2 2.6 Rest of Western Europe 13.1 0.1 2.7 0.0 United States 2.8 0.9 0.0 0.0 Hong Kong, China; Singapore 1.7 1.4 1.4 1.1 Developing countries 64.9 12.2 0.9 0.2 North Africa 0.9 2.8 0.5 1.5 Sub-Saharan Africa 0.0 3.2 0.0 0.9 East Asia 30.1 1.0 1.1 0.0 South Asia 0.4 3.9 0.1 0.6 Latin America 15.8 2.5 1.0 0.2 Middle East 4.2 0.2 0.8 0.0 Eastern Europe and Central Asia 14.2 3.6 1.2 0.3 World total 167.7 1.0 0.6 0.0 Source: World Bank Linkage model simulations by the authors. Note: The table shows results relative to the 2004 benchmark data. EU15 the 15 members of the European Union prior to 2004. 60 Agricultural Price Distortions, Inequality, and Poverty If one were to represent each of the 60 countries and regions in table 2.2 by a single household (that is, ignoring intraregional inequality), the income inequal- ity across countries as measured by the Gini coefficient would be reduced at least slightly, from 0.8513 to 0.8506.12 The second and fourth columns in table 2.2 show the amount of the welfare gain deriving from changes in the international terms of trade for each country. For developing countries as a group, this terms of trade effect is slightly negative; the converse is true for high-income countries. The regional and sectoral distribution of welfare effects One way to decompose the real income gains from the full removal of price distor- tions globally so that we may understand more clearly the sources of the gains for each region is to assess the impacts in various economic sectors of liberalization in developing countries versus liberalization in high-income countries. These results are provided in table 2.3. They suggest that the global liberalization of agriculture and food markets would contribute 70 percent of the total global gains from mer- chandise reform. This is slightly greater than the 63 percent found for 2015 by Anderson, Martin, and van der Mensbrugghe (2006b) using the earlier version 6 of the GTAP database anchored on 2001 estimates of distortions. This result is remark- able given the low shares of agriculture and food in global GDP and global merchan- dise trade (3 and 6 percent, respectively). For developing countries, the importance of agricultural policies is even slightly greater, at 72 percent (see row 7 in table 2.3). Slightly more than two-thirds of the global gains that may arise because of the removal of agricultural policies are accounted for by the farm policies of high- income countries (column 3 in table 2.3; US$82 billion of the US$117 billion per year). These policies also account for nearly one-quarter of the overall gains to developing countries from all global agricultural and trade policy reforms (col- umn 1 in table 2.3; US$15 billion of the US$65 billion per year). The quantities produced and traded The results of full global liberalization suggest that there would be little change in the aggregate shares of developing countries in the global output and exports of nonfarm products other than textiles and apparel. The shares of these countries in agricultural and processed food markets change noticeably, however. The export share rises from 54 to 64 percent, and the output share rises from 46 to 50 percent. More significantly, the rises occur in nearly all agricultural and food industries. The share of the global production of farm products that is exported thus rises dramatically in many industries and, for the sector as a whole, increases from 8 to 13 percent, excluding intra­European Union trade (table 2.4). This thickening of international food markets would reduce substantially the fluctuations in food prices and in the quantities of food traded in these markets. Table 2.3. Regional and Sectoral Sources of the Welfare Gains from the Full Liberalization of Global Merchandise Trade, 2004 (2004 U.S. dollars and percent) Gains by region,a US$, billions Share of regional gain, % Sector Developing High-income World Developing High-income World Developing countries liberalize Agriculture and light processing 31.8 3.9 35.6 48.6 3.8 21.2 Manufacturing and services 5.6 36.7 42.3 8.6 35.9 25.2 Total 37.4 40.6 77.9 57.2 39.6 46.5 High-income countries liberalize Agriculture and light processing 15.1 66.4 81.6 23.2 64.9 48.6 Manufacturing and services 12.8 4.6 8.2 19.6 4.5 4.9 Total 28.0 61.8 89.8 42.8 60.4 53.5 All countries liberalize Agriculture and light processing 46.9 70.3 117.2 71.8 68.7 69.9 Manufacturing and services 18.4 32.1 50.5 28.2 31.3 30.1 Total 65.3 102.3 167.7 100.0 100.0 100.0 Source: World Bank Linkage model simulations by the authors. Note: The table shows results relative to the 2004 benchmark data. a. Small interaction effects are distributed proportionately, and the numbers are rounded to sum to 100 percent. 61 Table 2.4. The Impact of Full Global Liberalization on the Shares of Global Output Exported, by Product, 2004 (percent) 62 Share in global output Developing countries in Developing countries in exported global output global exports Full global Full global Full global Benchmark liberalization Benchmark liberalization Benchmark liberalization Paddy rice 1 2 81 82 56 42 Wheat 16 22 67 71 25 39 Other grains 11 15 55 57 35 56 Oilseeds 21 28 69 74 54 68 Plant-based fibers 25 25 74 83 50 79 Vegetables and fruits 9 15 72 77 69 80 Other crops 14 17 49 49 75 62 Cattle, sheep, and so on 2 2 43 48 56 59 Other livestock 4 4 65 67 43 46 Wool 13 14 82 81 16 18 Beef and sheep meat 7 21 27 41 31 68 Other meat products 7 12 32 34 42 45 Vegetable oils and fats 20 30 52 58 80 84 Dairy products 5 11 29 33 28 41 Processed rice 5 7 76 79 85 87 Refined sugar 8 42 52 85 78 90 Other food, beverages, tobacco 9 12 35 36 50 59 Other primary products 31 33 64 63 76 76 Textiles and wearing apparel 28 35 53 57 74 77 Other manufacturing 24 26 32 31 43 43 Services 3 3 20 20 31 30 Agriculture and food 8 13 46 50 54 64 Agriculture 8 11 62 65 55 64 Processed foods 8 14 37 40 52 63 Source: World Bank Linkage model simulations by the authors. Note: The developing-country shares of global output and exports exclude intra­European Union trade. Global Welfare and Poverty Effects: Linkage Model Results 63 The impact of full trade reform on agricultural and food output and trade for each country and region is shown in table 2.5. It is clear that global farm trade is enhanced by more than one-third (39 percent), whereas the global value of output is virtually unchanged (dropping only 2.6 percent). This suggests that, in aggregate, the pro-agricultural policies of high-income countries are not quite fully offset by the Table 2.5. The Impact of Full Global Trade Liberalization on Agricultural and Food Output and Trade, by Country and Region, 2004 (2004 U.S. dollars and percent) Change relative US$, billions to baseline, % Country, region Output Exports Imports Output Exports Imports North Africa and Sub-Saharan Africa 13.8 20.5 10.0 7.2 99.1 46.0 Egypt, Arab Rep. 0.4 0.5 0.1 2.2 39.2 4.2 Madagascar 0.0 0.0 0.0 0.4 2.7 4.3 Mozambique 0.9 1.0 0.1 52.3 597.1 33.3 Nigeria 0.5 0.4 0.7 2.9 92.8 43.1 Senegal 0.0 0.0 0.0 1.9 35.0 0.3 South Africa 0.7 0.9 0.8 2.4 26.7 42.9 Tanzania 0.0 0.2 0.1 0.7 28.5 31.2 Uganda 0.0 0.0 0.0 0.6 1.3 1.5 Zambia 0.1 0.1 0.0 5.2 22.3 35.9 Zimbabwe 0.4 0.3 0.1 25.7 38.0 39.2 Rest of Africa 12.0 17.0 8.3 10.5 133.1 64.3 East and South Asia 25.0 39.5 24.7 2.7 83.4 36.7 China 6.2 7.7 6.7 1.7 76.5 27.5 Indonesia 1.1 1.6 1.0 1.8 21.6 21.5 Korea, Rep. 1.0 1.0 6.2 1.7 194.1 75.0 Malaysia 1.6 1.3 0.7 8.9 17.0 17.8 Philippines 1.1 1.9 0.8 3.5 120.5 35.0 Taiwan, China 1.9 0.3 1.5 9.1 62.8 35.5 Thailand 9.5 8.3 1.9 17.4 133.0 78.1 Vietnam 0.5 1.1 0.6 3.3 54.0 55.6 Bangladesh 0.6 0.4 0.8 2.4 261.2 38.3 India 1.1 9.0 1.4 0.5 131.2 24.2 Pakistan 0.6 0.5 1.0 1.3 45.0 43.0 Sri Lanka 0.1 0.1 0.6 1.2 18.2 69.3 Rest of East and South Asia 8.0 6.4 1.4 41.5 266.1 29.5 (Table continues on the following page.) Table 2.5. The Impact of Full Global Trade Liberalization on Agricultural and Food Output and Trade, by Country and Region, 2004 (continued) (2004 U.S. dollars and percent) Change relative US$, billions to baseline, % Country, region Output Exports Imports Output Exports Imports Latin America 87.2 71.5 7.2 26.8 106.4 29.8 Argentina 12.2 15.1 0.3 37.8 95.6 81.8 Brazil 45.8 25.7 2.1 45.3 100.7 94.8 Chile 0.5 0.4 0.2 4.7 11.3 15.8 Colombia 3.1 4.9 1.1 14.6 161.4 81.7 Ecuador 4.2 4.6 0.3 46.1 198.7 71.8 Mexico 0.3 0.3 0.4 0.4 5.8 4.3 Nicaragua 0.0 0.1 0.0 2.9 21.6 19.4 Rest of Latin America 21.6 20.4 2.8 25.7 175.9 30.4 Eastern Europe and Central Asia 10.4 17.4 20.3 2.6 79.7 77.6 Baltic States 1.2 0.1 0.4 16.9 15.5 30.9 Bulgaria 4.2 2.6 0.6 6.6 366.5 118.1 Czech Republic 2.2 0.1 0.7 12.0 10.9 40.5 Hungary 0.9 0.4 0.8 6.0 17.1 66.6 Poland 1.7 2.5 2.5 3.9 80.7 88.8 Romania 0.2 1.3 1.1 1.0 190.5 78.3 Russian Federation 12.9 3.2 8.8 13.1 179.4 98.9 Slovak Republic 0.9 0.1 0.4 11.3 12.0 64.1 Slovenia 0.6 0.1 0.2 17.1 54.1 26.2 Kazakhstan 1.5 1.4 0.0 11.8 142.9 11.6 Turkey 2.0 2.3 2.9 3.1 61.5 92.1 Rest of Eastern Europe and Central Asia 3.0 4.1 2.0 7.7 71.3 53.4 High-income countries 233.2 9.2 89.8 13.1 4.0 38.3 Australia 12.0 7.0 0.2 19.8 41.2 11.1 Canada 1.6 3.6 2.7 2.4 24.1 32.8 EU15 190.9 38.8 50.9 21.2 29.2 31.9 Japan 39.1 0.4 16.8 22.9 87.7 69.1 New Zealand 10.6 6.4 0.2 46.6 74.3 27.1 Rest of Western Europe 11.6 11.7 9.8 19.4 312.0 132.7 United States 12.8 0.6 9.3 2.6 1.1 32.4 Hong Kong, China; Singapore 0.1 0.0 0.1 2.1 6.3 1.6 Developing countries 137.6 163.6 64.6 7.1 100.0 40.4 North Africa 11.4 13.3 6.1 17.3 377.2 62.5 Sub-Saharan Africa 2.5 7.2 3.8 1.9 41.9 32.3 East Asia 25.1 29.5 20.8 4.0 77.4 37.4 South Asia 0.1 10.0 3.9 0.0 108.3 33.2 Latin America 87.2 71.5 7.2 26.8 106.4 29.8 Middle East 22.0 14.8 2.5 21.5 222.7 12.1 Eastern Europe and Central Asia 10.4 17.4 20.3 2.6 79.7 77.6 World total 95.7 154.4 154.4 2.6 39.1 39.1 Source: World Bank Linkage model simulations by the authors. Note: The table shows results relative to the 2004 benchmark data. EU15 the 15 members of the European Union prior to 2004. 64 Global Welfare and Poverty Effects: Linkage Model Results 65 Table 2.6. The Impact of Global Liberalization on the Share of Agricultural and Food Production Exports, by Country and Region, 2004 (percent) 2004 benchmark Full global Country, region data liberalization Developing countries 9.5 16.9 North Africa 6.3 20.6 Sub-Saharan Africa 13.8 19.3 East Asia 8.4 15.1 South Asia 3.7 7.5 Latin America 18.1 28.2 Middle East 7.4 17.2 Eastern Europe and Central Asia 6.8 11.1 High-income countries 13.0 14.1 World total 11.4 15.4 Source: World Bank Linkage model simulations by the authors. antiagricultural policies of developing countries, whereas the antitrade bias in the policies of both groups of countries reinforce each other. The increase in the exports of these goods from developing countries would be a huge US$163 billion per year. Latin America accounts for nearly half this increase, but the exports of all developing regions expand. This means that the share of these countries in production for export would be much higher. It would increase in the case of almost all developing coun- tries, rising in aggregate for the group from 10 to 17 percent (table 2.6). Also of interest is the course of agricultural imports. Developing countries as a group would see them growing less than farm exports (table 2.5). This means that the food and agricultural self-sufficiency of these countries would rise, although, in aggregate, only slightly. For high-income countries, it would fall five percentage points (slightly less if Eastern Europe is included); for East Asia and Africa, it would rise two or three points; for South Asia, it would be unchanged; and, for Latin America, it would jump from 112 to 126 percent (table 2.7). Such reform also raises substantially the share of agricultural and food produc- tion that is exported globally, thereby thickening international markets. This would restrict international food price fluctuations and thereby reduce the concerns about vulnerability to import dependence. The extent of this global public good aspect of agricultural and trade reform may be sensed for various products from the results reported in table 2.8. The case of highly protected sugar and milk, as well as grains and oilseeds, is especially noteworthy. Also noteworthy is the extent to which the developing-country shares of output exported rise for certain products. The share of grain production that is exported by these countries would double, and Table 2.7. The Impact of Global Liberalization on Self-sufficiency in Agricultural and Other Products, by Region, 2004 66 (percent) North Africa and Eastern High-income Developing Sub-Saharan Europe and countries countries Africa Latin America East Asia South Asia Central Asia Product BK GL BK GL BK GL BK GL BK GL BK GL BK GL Paddy rice 101 105 100 99 97 96 93 72 100 101 101 101 95 92 Wheat 141 140 88 89 67 46 80 98 68 65 100 98 102 117 Other grains 108 102 94 98 94 91 98 119 88 81 103 105 103 113 Oilseeds 104 92 97 103 104 130 140 167 66 51 100 101 106 115 Plant-based fibers 161 112 88 97 177 265 94 107 54 58 93 95 104 118 Vegetables and fruits 90 78 105 109 108 103 153 221 102 104 99 98 99 92 Other crops 90 91 113 110 138 138 143 133 110 104 104 104 90 88 Cattle, sheep, and so on 100 100 100 100 101 99 102 102 98 97 100 100 102 102 Other livestock 101 101 100 100 101 100 101 100 99 99 100 100 99 98 Wool 161 180 92 91 103 104 103 102 78 75 96 93 96 99 Beef and sheep meat 101 85 97 134 96 102 108 183 83 77 126 652 95 85 Other meat products 100 99 100 103 92 85 121 143 101 103 96 95 96 93 Vegetable oils and fats 95 85 103 114 69 191 141 143 115 116 78 66 93 96 Dairy products 103 100 94 101 76 79 97 102 78 78 99 99 102 104 Processed rice 99 95 100 101 69 63 94 85 104 108 104 104 92 87 Refined sugar 98 41 102 133 95 100 131 227 98 196 96 91 98 70 Other food, beverages, tobacco 99 97 103 105 101 100 108 112 105 113 106 94 100 98 Other primary products 76 76 122 122 180 189 148 155 84 82 75 69 115 116 Textiles and wearing apparel 81 76 123 128 98 91 104 91 144 155 144 153 101 95 Other manufacturing 101 102 98 96 77 74 96 91 106 105 90 89 95 95 Services 101 101 101 101 101 102 100 100 101 100 100 101 101 101 Agriculture and food 100 95 101 105 100 103 112 126 100 102 100 100 99 98 Agriculture 99 96 100 102 104 103 115 126 96 95 100 100 100 101 Processed foods 100 95 101 108 94 103 110 126 104 111 100 101 99 96 Source: World Bank Linkage model simulations by the authors. Note: Self-sufficiency domestic production as a percentage of domestic consumption measured in value terms at free-on-board prices. BK benchmark. GL global liberalization. Table 2.8. Shares of Production Exported and of Consumption Imported before and after Full Global Liberalization of all Merchandise Trade, by Product, 2004 (percent) Share of production exported Share of consumption imported High-income countries Developing countries High-income countries Developing countries 2004 Global 2004 Global 2004 Global 2004 Global Product benchmark liberalization benchmark liberalization benchmark liberalization benchmark liberalization Paddy rice 3 7 1 1 2 3 1 2 Wheat 37 47 6 12 11 25 17 21 Other grains 15 16 7 15 9 14 11 15 Oilseeds 31 34 16 25 26 36 16 22 Plant-based fibers 50 31 17 24 18 22 26 25 Vegetables and fruits 10 13 9 15 18 30 4 7 Other crops 7 13 21 22 16 20 11 14 Cattle, sheep, and so on 1 2 2 2 2 2 2 2 Other livestock 6 7 3 3 6 6 3 3 Wool 60 62 2 3 35 31 10 12 Beef and sheep meat 6 11 7 35 5 24 10 13 Other meat products 6 10 9 16 6 12 8 14 Vegetable oils and fats 8 11 31 43 12 24 26 34 Dairy products 5 10 4 14 2 10 10 14 Processed rice 3 4 5 8 4 9 5 7 Refined sugar 4 30 12 44 5 66 10 25 Other food, beverages, tobacco 7 8 12 20 8 10 9 16 Other primary products 20 21 37 39 38 39 22 24 Textiles and wearing apparel 15 19 39 48 30 37 23 31 Other manufacturing 20 21 32 36 19 20 32 38 Services 3 3 5 4 2 2 5 5 Agriculture and food 7 9 9 17 8 13 8 12 Agriculture 9 11 7 11 10 15 7 9 Processed foods 6 9 12 23 7 13 10 16 67 Source: World Bank Linkage model simulations by the authors. Note: The data on the high-income countries exclude intra­European Union trade. 68 Agricultural Price Distortions, Inequality, and Poverty the corresponding share of meat would more than double, while the share of sugar would rise nearly fourfold. Global exports of cotton (plant-based fibers) would become more dominated by developing countries, while the share of cotton production exported by high-income countries would fall from 50 to 31 percent. The effects on product and factor prices The average real international prices of agricultural and lightly processed food prod- ucts would be only 1.3 percent higher in the absence of all merchandise trade distor- tions, or 2.0 percent if only agricultural policies were liberalized (see table 2.9). The Table 2.9. The Impact of Full Global Liberalization on Real International Product Prices, 2004 (percent relative to 2004 baseline) Policies in all Product Agricultural policies goods sectors Paddy rice 6.9 6.6 Wheat 1.8 1.4 Other grains 2.6 2.7 Oilseeds 2.2 2.4 Sugarcane and beets 1.1 2.0 Plant-based fibers 4.7 2.9 Vegetables and fruits 2.4 1.8 Other crops 1.7 1.0 Cattle, sheep, and so on 0.2 1.1 Other livestock 1.2 2.1 Raw milk 0.7 0.2 Wool 3.5 3.3 Beef and sheep meat 5.6 4.6 Other meat products 1.3 0.6 Vegetable oils and fats 1.4 1.9 Dairy products 4.6 3.8 Processed rice 2.8 2.9 Refined sugar 2.5 1.3 Other food, beverages, tobacco 1.7 1.3 Textiles and wearing apparel 0.3 1.2 Other manufacturing 0.2 0.2 Merchandise trade 0.3 0.2 Agriculture and food 0.8 0.3 Agriculture 1.5 0.9 Agriculture and light processing 2.0 1.3 Source: World Bank Linkage model simulations by the authors. Note: The model numéraire is the export price index of the manufactured exports of high-income countries. Global Welfare and Poverty Effects: Linkage Model Results 69 net effects of distortions (as of 2004) are especially dampening on the international prices of beef, milk, rice, and cotton. However, they prop up the international prices of some other products because export taxes are still in place in some developing countries, most notably Argentina. The size of the redistribution of welfare among groups within each country following trade reform may be much larger than the aggregate change partly because of the impacts on real pretax rewards to labor, capital, and land. These effects are reported in table 2.10, in which the data on factor rewards are deflated by the overall consumer price index and also, in the case of unskilled wages, by the food and the food plus clothing consumer price indexes (since these items are so prominent in the spending of unskilled workers). Consistent with trade theory, these results suggest that unskilled workers in developing countries--the majority of whom work on farms--would benefit most from reform, followed by skilled workers, then capital owners. The returns to immobile agricultural land would also rise in developing countries, but by less than the rise in more mobile factors. Land returns fall substantially in highly protected Japan and Western Europe, change little in the United States, rise considerably in Australia and Canada, and rise dramatically in dairy-intensive New Zealand. The effects on sectoral value added Also of crucial interest in terms of the impact of these policies on inequality and poverty is the effect the policies have on value added in agriculture, that is, on net farm incomes. The results of this issue in full global reform are reported in the first four columns of table 2.11. They show that, for developing countries as a group, the value added in agriculture rises by 5.6 percent following full global reform of all merchandise trade. This compares with only 1.9 percent for nonagriculture. Net farm income expands the most in Latin America, where the average rise is 37 per- cent, but exceeds 100 percent in Argentina and Ecuador and 40­50 percent in Brazil and Colombia. In East Asia, it also expands considerably, more than the expansion in nonagricultural value added, including in China. However, among the countries listed in Africa, net farm incomes would increase substantially only in Mozambique, Zambia, and Zimbabwe, and, in the continent as a whole, including North Africa, they would fall only slightly (by less than 1 percent). Partly this is because the nonagricultural primary sector--in which numerous African coun- tries have a strong comparative advantage--would expand (raising Africa's self- sufficiency in that sector from 180 to 189 percent; see table 2.7), and this would, in turn, boost the production and employment in nontradable goods and services. It is estimated that net farm incomes would also fall in South Asia (by 7 percent), although, there, textiles and clothing would expand (raising the self-sufficiency in 70 Table 2.10. The Impacts of Full Global Merchandise Trade Liberalization on Real Pretax Factor Prices, by Country and Region, 2004 (percent) Nominal change deflated by the Real change in unskilled wages, aggregate CPI deflated Skilled Capitala Landa By aggregate By food By food and Country, region wages user cost user cost CPI CPI clothing CPI North Africa and Sub-Saharan Africa 4.7 4.3 0.1 4.4 5.8 6.9 Egypt, Arab Rep. 3.2 1.7 3.4 2.3 2.2 2.9 Madagascar 2.0 0.2 0.3 0.8 0.8 0.7 Mozambique 0.2 4.3 5.8 10.7 12.0 13.0 Nigeria 10.1 10.5 1.3 3.2 7.7 8.6 Senegal 2.4 3.2 0.7 3.1 2.6 3.6 South Africa 1.8 2.4 0.1 1.6 1.7 3.6 Tanzania 2.1 2.4 1.6 2.8 2.6 3.0 Uganda 2.2 0.7 0.8 0.2 0.2 0.5 Zambia 2.6 3.2 0.7 3.0 3.1 3.5 Zimbabwe 6.7 11.8 23.1 13.6 15.9 16.8 Rest of Africa 6.2 5.5 1.1 6.4 7.2 8.5 East and South Asia 3.4 3.0 1.8 3.2 4.6 4.8 China 1.9 2.0 3.6 2.6 1.6 2.1 Indonesia 0.7 0.7 0.1 1.4 1.5 1.8 Korea, Rep. 7.1 6.5 14.5 5.8 26.6 22.7 Malaysia 10.7 10.2 3.6 11.1 32.3 29.2 Philippines 1.0 1.4 7.2 8.5 9.2 9.4 Taiwan, China 2.4 2.8 11.8 3.3 10.9 10.3 Thailand 2.6 3.5 7.5 5.6 8.5 7.8 Vietnam 17.7 16.0 9.1 19.6 23.9 26.7 Bangladesh 2.1 2.8 2.8 3.3 4.5 5.3 India 2.1 0.2 8.5 1.9 4.4 3.8 Pakistan 3.9 3.5 1.5 3.8 3.3 3.5 Sri Lanka 0.1 14.1 0.9 3.0 4.8 4.8 Rest of East and South Asia 2.0 2.9 6.5 5.1 8.3 9.1 Latin America 1.4 1.9 21.1 4.5 2.4 4.1 Argentina 3.1 4.1 43.6 8.8 4.9 7.2 Brazil 1.3 2.7 26.5 1.4 0.2 1.1 Chile 1.3 0.5 3.0 1.3 1.1 1.9 Colombia 0.8 0.3 30.2 6.0 4.4 5.6 Ecuador 2.1 1.2 61.7 15.1 12.1 13.9 Mexico 0.5 0.5 2.3 0.8 2.3 0.9 Nicaragua 1.9 2.5 2.1 3.8 3.9 4.7 Rest of Latin America 0.8 1.5 18.0 5.4 4.9 6.1 Eastern Europe and Central Asia 3.2 2.6 4.5 1.7 4.2 4.5 Baltic States 3.9 2.0 9.8 1.3 5.6 5.0 Bulgaria 0.3 1.6 5.9 2.5 1.8 2.3 Czech Republic 2.4 1.5 26.1 0.7 3.9 3.3 Hungary 2.2 1.1 19.9 1.0 1.6 1.4 Poland 3.9 2.6 24.6 0.5 5.2 4.6 Romania 4.5 3.5 3.4 3.4 4.5 5.7 Slovak Republic 2.9 2.2 15.9 1.0 4.9 4.0 Slovenia 2.3 1.8 17.2 1.3 4.9 4.2 Russian Federation 3.9 3.8 1.9 2.5 4.3 5.1 Kazakhstan 1.0 1.4 14.0 3.0 2.4 3.3 Turkey 2.1 1.4 3.4 0.4 1.8 1.8 71 Rest of Eastern Europe and Central Asia 3.7 4.2 12.7 6.3 9.7 10.3 (Table continues on the following page.) 72 Table 2.10. The Impacts of Full Global Merchandise Trade Liberalization on Real Pretax Factor Prices, by Country and Region, 2004 (continued) (percent) Nominal change deflated by the Real change in unskilled wages, aggregate CPI deflated Skilled Capitala Landa By aggregate By food By food and Country, region wages user cost user cost CPI CPI clothing CPI High-income countries 1.0 0.5 17.9 0.2 3.3 3.3 Australia 0.4 0.8 9.4 1.3 0.0 1.6 Canada 0.5 0.4 6.3 0.4 1.7 2.7 EU15 1.7 0.6 39.5 0.1 4.2 3.6 Japan 1.7 1.2 29.3 0.9 6.5 6.0 New Zealand 1.2 1.5 34.8 5.9 6.2 7.3 Rest of Western Europe 3.1 3.1 50.6 0.8 19.3 14.0 United States 0.2 0.1 2.9 0.1 2.0 0.0 Hong Kong, China; Singapore 0.3 0.8 0.4 2.2 1.8 2.4 Developing countries 3.0 2.9 1.6 3.5 5.5 5.9 North Africa 7.7 5.3 0.5 7.0 9.3 10.4 Sub-Saharan Africa 3.2 3.8 0.2 3.2 4.4 5.3 East Asia and Pacific 3.4 3.3 1.9 4.0 6.9 6.9 South Asia 2.3 1.2 6.2 0.6 2.5 1.9 Latin America 1.4 1.9 21.1 4.5 2.4 4.1 Middle East 2.9 4.7 43.8 8.3 17.0 16.5 Eastern Europe and Central Asia 3.2 2.6 4.5 1.7 4.2 4.5 World total 1.3 1.2 3.1 0.9 3.6 3.8 Source: World Bank Linkage model simulations by the authors. Note: The table shows results relative to the 2004 benchmark data. CPI consumer price index. EU15 the 15 members of the European Union prior to 2004. a. The user cost of capital and land represents the subsidy-inclusive rental cost. Table 2.11. The Effects of the Full Global Liberalization of Agricultural and Merchandise Trade on Sectoral Value Added, by Country and Region, 2004 (2004 U.S. dollars and percent) US$, billions Percent Agricultural Policies in all Agricultural Policies in all policies sectors policies sectors Country, region Ag Nonag Ag Nonag Ag Nonag Ag Nonag North and Sub Saharan Africa 0.1 5.1 0.9 0.2 0.1 0.8 0.9 0.0 Egypt, Arab Rep. 0.1 0.2 0.0 0.7 1.3 0.4 0.1 1.1 Madagascar 0.0 0.0 0.0 0.1 3.2 0.1 3.4 3.1 Mozambique 0.3 0.0 0.3 0.0 23.6 0.6 22.7 0.1 Nigeria 0.6 0.2 1.2 0.8 4.8 0.5 9.3 1.7 Senegal 0.0 0.0 0.0 0.0 1.5 0.8 1.1 0.8 South Africa 0.2 0.7 0.1 0.1 2.7 0.4 0.7 0.1 Tanzania 0.0 0.0 0.0 0.1 0.6 0.3 0.3 1.3 Uganda 0.1 0.0 0.1 0.1 1.6 0.4 2.9 1.6 Zambia 0.0 0.0 0.0 0.0 0.7 0.5 0.6 0.6 Zimbabwe 0.1 0.0 0.2 0.2 24.2 0.8 38.9 4.9 Rest of Africa 0.5 3.9 0.0 1.4 0.7 1.4 0.1 0.5 East and South Asia 1.4 24.4 2.0 100.7 0.3 0.7 0.5 2.9 China 4.6 2.5 9.4 37.5 2.8 0.2 5.7 3.0 Indonesia 0.3 1.1 0.2 2.7 1.1 0.5 0.8 1.2 Korea, Rep. 4.0 7.2 3.2 31.3 18.7 1.2 15.1 5.4 Malaysia 0.2 0.9 0.1 4.0 6.3 0.8 2.0 3.8 Philippines 1.7 0.3 1.9 1.0 13.8 0.5 15.6 1.7 Taiwan, China 0.5 0.8 0.5 10.1 11.3 0.3 9.9 3.7 73 (Table continues on the following pages.) 74 Table 2.11. The Effects of the Full Global Liberalization of Agricultural and Merchandise Trade on Sectoral Value Added, by Country and Region, 2004 (continued) (2004 U.S. dollars and percent) US$, billions Percent Agricultural Policies in all Agricultural Policies in all policies sectors policies sectors Country, region Ag Nonag Ag Nonag Ag Nonag Ag Nonag Thailand 2.9 2.7 3.0 7.3 14.0 1.0 14.3 2.8 Vietnam 1.4 0.0 1.2 4.5 22.8 0.0 18.8 15.6 Bangladesh 0.2 0.4 0.3 2.1 2.6 0.9 3.8 4.4 India 7.8 6.3 10.6 1.3 6.1 1.4 8.3 0.3 Pakistan 0.2 0.1 0.1 0.2 1.0 0.1 0.5 0.2 Sri Lanka 0.0 0.0 0.3 1.3 0.0 0.1 7.1 9.6 Rest of East and South Asia 0.6 2.3 0.7 4.3 9.6 1.4 11.2 2.7 Latin America 40.0 42.2 40.7 34.6 36.3 2.8 37.0 2.3 Argentina 12.4 8.1 10.9 15.1 116.8 7.4 103.5 13.8 Brazil 12.2 22.7 13.0 21.3 40.1 4.4 42.6 4.2 Chile 0.2 0.3 0.2 0.7 5.0 0.3 5.5 0.9 Colombia 5.0 2.1 5.0 1.2 53.5 2.7 53.5 1.5 Ecuador 2.6 2.9 2.9 1.7 113.1 11.4 126.0 6.7 Mexico 0.2 0.6 0.1 3.4 1.0 0.2 0.3 1.0 Nicaragua 0.0 0.0 0.0 0.1 3.0 1.4 2.4 2.3 Rest of Latin America 7.9 5.5 8.6 2.1 26.3 1.5 28.7 0.6 Eastern Europe and Central Asia 5.2 4.4 6.2 4.4 4.4 0.3 5.2 0.3 Baltic States 0.1 0.1 0.1 0.2 7.5 0.3 8.9 0.5 Bulgaria 0.3 0.1 0.4 0.1 5.1 0.4 5.6 0.3 Czech Republic 0.7 0.4 0.7 0.3 19.2 0.4 20.9 0.3 Hungary 0.7 0.3 0.7 0.1 16.8 0.4 17.9 0.1 Poland 2.4 2.1 2.5 1.7 21.8 1.1 22.6 0.9 Romania 0.3 0.2 0.5 0.3 3.7 0.4 5.8 0.5 Russian Federation 2.2 0.7 2.3 1.3 6.3 0.2 6.6 0.3 Slovak Republic 0.1 0.1 0.1 0.1 11.8 0.2 13.5 0.4 Slovenia 0.0 0.1 0.0 0.1 9.2 0.4 11.1 0.4 Kazakhstan 0.5 0.4 0.5 0.5 23.1 1.1 23.0 1.2 Turkey 1.0 0.9 1.5 0.9 3.2 0.4 4.7 0.4 Rest of Eastern Europe and Central Asia 1.5 0.5 1.5 2.1 11.1 0.4 11.1 1.8 High-income countries 55.1 61.9 58.5 28.6 13.8 0.2 14.7 0.1 Australia 2.2 8.4 2.7 11.7 10.9 1.5 13.7 2.1 Canada 0.4 2.5 0.7 4.6 3.4 0.3 5.3 0.5 EU15 42.9 16.7 47.4 45.9 23.0 0.2 25.4 0.4 Japan 7.6 4.5 7.6 93.2 16.7 0.1 16.8 2.3 New Zealand 2.7 4.1 2.7 4.4 57.7 5.0 57.2 5.4 Rest of Western Europe 3.6 6.5 3.6 8.4 25.8 1.0 25.8 1.3 United States 6.4 18.6 6.0 25.2 5.7 0.2 5.3 0.2 Hong Kong, China; Singapore 0.0 0.6 0.0 3.4 3.7 0.4 2.2 2.1 Developing countries 42.7 79.5 44.4 145.6 5.4 1.0 5.6 1.9 North Africa 0.1 3.9 0.3 1.8 0.4 1.8 1.1 0.8 Sub-Saharan Africa 0.2 1.2 0.6 2.0 0.3 0.3 0.8 0.5 East Asia 6.8 17.7 12.6 102.8 2.6 0.6 4.7 3.5 South Asia 8.2 6.7 10.7 2.1 5.1 1.1 6.7 0.3 Latin America 40.0 42.2 40.7 34.6 36.3 2.8 37.0 2.3 Eastern Europe and Central Asia 5.2 4.4 6.2 4.4 4.4 0.3 5.2 0.3 World total 12.4 141.4 14.2 174.2 1.0 0.4 1.2 0.5 75 Source: World Bank Linkage model simulations by the authors. Note: The table shows results relative to the 2004 benchmark data. Ag agriculture. Nonag nonagriculture. EU15 the 15 members of the European Union prior to 2004. 76 Agricultural Price Distortions, Inequality, and Poverty the sector from 144 to 153 percent). For India, the skilled-unskilled wage differen- tial rises, as does skill-intensive goods and services production. The effects on poverty measured using the elasticities approach The results for real factor rewards and net farm income suggest that inequality and poverty might be reduced globally through agricultural and trade policy lib- eralization. It is possible for us to take a step or two further in assessing the impacts of reform on poverty through a global model, even with only a single rep- resentative household per country. This involves using the elasticities approach, which is employed here in two ways. The first way focuses on the impact on real household income by applying an estimated income-to-poverty elasticity and then assessing the impacts on the poverty headcount index for each country. This simple approach assumes that there is distributional neutrality: the poor receive the same proportional increase in real income as the average household in the economy, and all are subject to the same higher rate of direct income taxation to replace the customs revenue forgone because of trade liberalization. A slightly more complex, but more reasonable approach involves linking key model variables to the possible change in the average per capita consumption of the poor, that is, to capture from the model's results some of the distributional aspects of the changes in real income, rather than simply the average gain. We have accomplished this by calculating the change in the average (pretax) wage of unskilled workers, deflated by the food and clothing consumer price index. (Food and clothing prices are presumably the most relevant consumer items among the poor, including people in the many poor farm households and other rural house- holds who earn most of their income from wages and are net buyers of food.) These workers are assumed to be exempt from the direct income tax imposed to replace the lost customs revenue following trade reform, a realistic assumption for many developing countries.13 Table 2.12 summarizes the key poverty results that emerge from the global reform scenario using these two approaches. As is clear from the comparison, the more naive first approach yields little change in poverty numbers. We therefore concentrate our attention here on the results generated using the more realistic second approach. Under the full merchandise trade reform scenario, extreme poverty--the num- ber of people surviving on less than US$1 a day--would drop in developing coun- tries by 26 million relative to the baseline level of slightly under 1 billion, a reduc- tion of 2.7 percent. The proportional reduction is much higher in China than in other developing countries, however: 3.7 compared with 2.6 percent. This would continue the trend of the recent past whereby China has been the region in which Table 2.12. The Poverty Effects of Full Global Liberalization of Merchandise Trade Reform, by Region, 2004 (number and percent) a. The benchmark Benchmark Poverty US$1 a day US$2 a day elasticities Number of poor, Number of poor, US$1 US$2 Region Headcount, % millions Headcount, % millions a day a day East Asia 9 169 37 684 n.a. n.a. China 10 128 35 452 1.9 1.3 Other East Asia 9 41 50 232 3.7 2.1 South Asia 31 446 77 1,116 n.a. n.a. India 34 371 80 868 1.1 0.5 Other South Asia 29 76 94 248 2.5 0.7 Eastern Europe and Central Asia 1 4 10 46 1.7 1.7 Middle East and North Africa 1 4 20 59 2.5 2.3 Sub-Saharan Africa 41 298 72 522 0.7 0.5 Latin America 9 47 22 121 1.7 1.1 Developing-country total 18 969 48 2,548 n.a. n.a. Excluding China 21 841 52 2,096 n.a. n.a. (Table continues on the following page.) 77 78 Table 2.12. The Poverty Effects of Full Global Liberalization of Merchandise Trade Reform, by Region, 2004 (continued) (number and percent) b. All are assumed to obtain the average income gain Change in US$1 a day US$2 a day number of poor Average income Number of Number of US$1 US$2 change, real, Headcount, poor, Headcount, poor, a day, a day, Region % % millions % millions millions millions East Asia 1.1 9 166 36 675 2.9 8.9 China 0.2 10 128 35 451 0.6 1.4 Other East Asia 1.9 8 38 49 224 2.3 7.5 South Asia 0.1 31 446 77 1,116 0.1 0.3 India 0.2 34 371 80 868 0.6 0.7 Other South Asia 0.3 29 75 94 248 0.5 0.4 Eastern Europe and Central Asia 1.2 1 4 10 45 0.1 1.0 Middle East and North Africa 0.7 1 4 19 58 0.1 1.0 Sub-Saharan Africa 0.0 41 298 72 522 0.0 0.0 Latin America 1.0 9 46 22 119 0.8 1.3 Developing-country total 0.9 18 966 47 2,536 3.7 12.0 Excluding China 1.1 21 838 51 2,085 3.1 10.6 c. The gain in the real earnings of unskilled workers Change in US$1 a day US$2 a day number of poor Average income Number of Number of US$1 US$2 change, real,a Headcount, poor, Headcount, poor, a day, a day, Region % % millions % millions millions millions East Asia 4.4 8 151 34 632 17 52 China 2.1 9 123 34 440 5 12 Other East Asia 8.1 6 29 42 192 12 40 South Asia 1.9 32 454 78 1,124 8 8 India 3.8 36 386 82 883 15 15 Other South Asia 4.0 26 68 92 241 8 7 Eastern Europe and Central Asia 4.5 1 4 9 43 0 4 Middle East and North Africa 14.3 1 3 13 40 2 19 Sub-Saharan Africa 5.3 39 287 70 508 11 14 Latin America 4.1 8 44 21 115 3 6 Developing-country total 5.9 18 944 46 2,462 26 87 Excluding China 6.5 20 820 50 2,022 21 74 Source: World Bank Linkage model simulations by the authors. Note: n.a. not applicable. a. Nominal unskilled wage deflated by the food and clothing consumer price index. 79 80 Agricultural Price Distortions, Inequality, and Poverty poverty reduction has been most numerous (see Chen and Ravallion 2007, 2008). Nonetheless, in this scenario, the number of the extreme poor in Sub-Saharan Africa would fall by 3.7 percent. In India (though not in the rest of South Asia), by contrast, we estimate that the number of the extreme poor would climb by 4.0 percent. Recall that this set of poverty calculations is based on the change in the real wage of unskilled workers, deflated by the food and clothing consumer price index. The average change in real unskilled wages across all developing countries is 5.9 percent, which is six times greater than the average net income increase in developing countries, after accounting for a rise in direct taxes to offset the loss in tariff revenues and assuming that the change in unskilled wages is fully passed through to households. This suggests that such a reform would deliver a marked average reduction in income inequality within developing countries. Under the broader definition of poverty (a subsistence threshold of no more than US$2 a day), the number of poor in developing countries would fall by 87 million under the full reform scenario relative to the aggregate baseline level of nearly 2.5 billion poor. This represents a somewhat larger proportion- ate reduction in the number of poor in developing countries, by 3.4 percent, or 3.7 percent if China is excluded. The proportionate decline in Sub-Saharan Africa is 2.7 percent, while, in India, there is an increase by 1.7 percent. Caveats As with all modeling, our results depend on our assumptions in structuring the model. Several assumptions place a downward bias on our estimates of the welfare gains arising from trade. They include the assumptions that returns to scale are constant (rather than increasing), that there are no productivity effects of reform (for example, of the sort stressed by Melitz 2003), and that there is no possibility for the creation of new markets following reform. There is also always the issue of product and regional aggregation: the less disaggregated the specification of the world economy, the smaller the estimated benefits of reform. This is because there is no accounting for welfare gains from adjustments within aggregated sectors or regions. As for the effects on poverty, the crude methodology used at the end of the pre- vious subsection is meant simply as a beginning in our examination of the poverty consequences of global trade reform. The results of global reform that we have identified based on the Linkage model and presented above are used in the next chapter, in association with microsimulation survey data on individual coun- tries, to assess the effects on the distribution of income across and within 101 countries. Bussolo, De Hoyos, and Medvedev, the authors of the next chapter, are Global Welfare and Poverty Effects: Linkage Model Results 81 thus able to identify much more precisely the inequality and poverty effects worldwide of the policies as of 2004. In parts III­V, in individual developing- country case studies, authors examine--also with the help of national microsim- ulation survey data--the effects of own-country policies, but also the effects of policies in the rest of the world. The border price and export demand shocks asso- ciated with rest-of-the-world liberalization are almost the same as those outlined above. The only difference is that, for each of the countries in the case studies, a separate global simulation has been run that excludes the reforms undertaken by the countries under examination. The border shocks in all the case studies are reported in the appendix. Conclusions The findings presented above are aimed at indicating the global economic effects of the agricultural and trade policies in place as of 2004. They may be summarized as follows: · As a share of national income, developing countries would gain nearly twice as much as high-income countries by removing the policies (an average increase of 0.9 percent compared with 0.5 percent, respectively), thereby reducing income inequality across countries. · Even intercountry inequality (not taking into account differences in the eco- nomic size of nations) as measured by the Gini coefficient would improve slightly, although some developing countries (notably in South Asia) would lose, and a few (for example, Ecuador) would gain many times more than the average. · Of the prospective welfare gains from global liberalization, 60 percent would be generated by agricultural and food policy reform, which is a striking result given that the shares of agriculture and food in global GDP and global merchandise trade are less than 9 percent. · The contribution of agricultural policy reform to the prospective welfare gain in developing countries is even greater, at 83 percent. · Through the full liberalization of the trade in goods, the share of exports in the global production of farm products would rise from 8 to 13 percent (excluding intra­European Union trade), thereby thickening international food markets and reducing the price fluctuations and the instability in the quantities traded on these markets. · Unskilled workers in developing countries--the majority of whom work on farms--would benefit most from reform, followed by skilled workers and then capital owners. The average change in the real unskilled wage across all 82 Agricultural Price Distortions, Inequality, and Poverty developing countries would rise 3.8 percent or nearly five times more than the average increase in net incomes in developing countries. · Net farm incomes in developing countries would rise by 5.6 percent, compared with 1.9 percent for nonagricultural value added, suggesting that inequality between farm and nonfarm households in developing countries would fall. · In contrast, in high-income countries, net farm incomes would fall by 15 per- cent on average (compared with a slight rise in real nonfarm value added). So, inequality between farm households in these countries and farm households in developing countries would decline substantially. · The number of people in extreme poverty--the number surviving on less than US$1 a day--in developing countries would drop by 26 million relative to the baseline level of slightly under 1 billion, a reduction of 3 percent (4 percent if the more moderate US$2-a-day poverty level is used). To obtain a more precise sense of the inequality and poverty effects within countries and to explore the extent to which own-country policies rather than rest-of world policies are causing the harm, we need country case studies using national economy-wide models that are enhanced through the inclusion of detailed earning and expenditure information on numerous types of urban and rural households. This is the purpose of parts III­V. Notes 1. This distortions database is documented fully in Anderson and Valenzuela (2008). It is based on the methodology summarized in Anderson et al. (2008a, 2008b). 2. Some of the questions raised here have been addressed by Anderson, Martin, and van der Mensbrugghe (2006a, 2006b), who use the same Linkage model we use in the present analysis, and by Anderson and Valenzuela (2007a), who use the GTAP-AGR model. However, in each case, these authors have relied on the GTAP version 6 protection database for 2001, which includes only the applied import tariffs in developing countries. 3. In principle, service trade and foreign investment distortions may also affect incentives in the agricultural and industrial sectors, but they are ignored here because much controversy still surrounds the measurement of these distortions and the way they should be modeled. This is reflected in the widely differing results emerging from attempts to include service distortions in trade reform model- ing. Compare, for example, Brown, Deardorff, and Stern (2003); Francois, van Meijl, and van Tongeren (2005); and Hertel and Keeney (2006). 4. We have no new distortion estimates for countries in the Middle East; so, in what follows, we give little attention to this small and relatively affluent part of the global agricultural economy. 5. More information on the MAcMaps database is available in Bouët et al. (2008) and at http://www. cepii.fr/anglaisgraph/bdd/macmap.htm. For details on the incorporation of the database into the GTAP version 7 data set, see Narayanan and Walmsley (2008). 6. Using the GTAP version 6 protection database for 2001, Anderson, Martin, and Valenzuela (2006) find that agricultural production and export subsidies together contributed only 7 percent of the global welfare cost of agricultural protection. 7. The size of the Armington elasticities is important; see Valenzuela, Anderson, and Hertel (2008) and Zhang and Osborne (2009). The Linkage model assumes larger values relative to some other Global Welfare and Poverty Effects: Linkage Model Results 83 models because it is seeking to estimate the long-run consequences of liberalization. An example of the differences this may generate in the results is detailed in Anderson and Martin (2006, table 12A.2). 8. The more affluent economies of Hong Kong, China and of Singapore are in our high-income category, but, since they have trade policies that approximate free trade policies and almost no farm production anyway, their influence on the results is not noticeable. 9. The model also does not include any divergences between the private and social marginal costs and the benefits that might arise from externalities, market failures, and other behind-the-border policies not represented in our amended GTAP protection database. These omissions may affect the welfare estimates in either direction. 10. The only other policy change is the removal of the export taxes on nonfarm products in Argentina. We do this because these taxes were introduced at the same time (at the end of 2001) and for the same reason (so the government might gain the support of the urban poor) as the country's export taxes on farm products. 11. This would continue a process that began in the 1980s when many countries began to reform their trade and exchange rate regimes. Using the same Linkage model and database as the present study, Valenzuela, van der Mensbrugghe, and Anderson (2009) find that the global reforms between 1980­84 and 2004 also boosted economic welfare proportionately more in developing countries than in high-income economies (by 1.0 percent compared with 0.7 percent, respectively). 12. The Gini coefficient is a measure of intercountry inequality in the Milanovic (2005) sense, whereby each country is treated as a single observation without regard to the economic size of the country. The measure is calculated using the 60 regions and Deaton's Gini coefficient calculation: N + 1 2 G = - (© n P X ) N - 1 N(N - 1)u i = 1 i i where N is the number of regions, u is the sample average GDP, Pi is the GDP sample rank (setting the highest at 1 and the lowest at N), and Xi is the GDP of country i. 13. Even if the fiscal closure affects a domestic sales or value added tax instead of direct taxes on households, food is exempt from taxation in many countries, or the tax is difficult to collect in practice because of the informal nature of many food markets. References Anderson, K., ed. 2009. Distortions to Agricultural Incentives: A Global Perspective, 1955­2007. London: Palgrave Macmillan; Washington, DC: World Bank. Anderson, K., and Y. Hayami, eds. 1986. The Political Economy of Agricultural Protection: East Asia in International Perspective. Sydney: Allen and Unwin. Anderson, K., M. Kurzweil, W. Martin, D. Sandri, and E. Valenzuela. 2008a. "Methodology for Measur- ing Distortions to Agricultural Incentives." In Distortions to Agricultural Incentives: A Global Perspective, 1955­2007, ed. K. Anderson, Appendix A. London: Palgrave Macmillan; Washington, DC: World Bank. ------. 2008b. "Measuring Distortions to Agricultural Incentives, Revisited." World Trade Review 7 (4): 1­30. Anderson, K., and W. Martin, eds. 2006. Agricultural Trade Reform and the Doha Development Agenda. London: Palgrave Macmillan; Washington, DC: World Bank. Anderson, K., W. Martin, and D. van der Mensbrugghe. 2006a. "Distortions to World Trade: Impacts on Agricultural Markets and Incomes." Review of Agricultural Economics 28 (2): 168­94. ------. 2006b. "Market and Welfare Implications of the Doha Reform Scenarios." In Agricultural Trade Reform and the Doha Development Agenda, ed. K. Anderson and W. Martin, chap. 12. London: Palgrave Macmillan; Washington, DC: World Bank. 84 Agricultural Price Distortions, Inequality, and Poverty Anderson, K., W. Martin, and E. Valenzuela. 2006. "The Relative Importance of Global Agricultural Subsidies and Market Access." World Trade Review 5 (3): 357­76. Anderson, K., and E. Valenzuela. 2007a. "Do Global Trade Distortions Still Harm Developing Country Farmers?" Review of World Economics 143 (1): 108­39. ------. 2007b. "The World Trade Organization's Doha Cotton Initiative: A Tale of Two Issues." World Economy 30 (8): 1281­1304. ------. 2008. "Estimates of Global Distortions to Agricultural Incentives, 1955­2007." Data spread- sheet, October, World Bank, Washington, DC. http://go.worldbank.org/YAO39F35E0. Bautista, R. M., S. Robinson, P. Wobst, and F. Tarp. 2001. "Policy Bias and Agriculture: Partial and General Equilibrium Measures." Review of Development Economics 5 (1): 89­104. Bouët, A., Y. Decreux, L. Fontagné, S. Jean, and D. Laborde. 2008. "Assessing Applied Protection across the World." Review of International Economics 16 (5): 850­63. Brown, D. K., A. V. Deardorff, and R. M. Stern. 2003. "Multilateral, Regional, and Bilateral Trade-Policy Options for the United States and Japan." World Economy 26 (6): 803­28. Chen, S., and M. Ravallion. 2007. "Absolute Poverty Measures for the Developing World, 1981­2004." Policy Research Working Paper 4211, World Bank, Washington, DC. ------. 2008. "The Developing World Is Poorer Than We Thought, but No Less Successful in the Fight against Poverty." Policy Research Working Paper 4703, World Bank, Washington, DC. Dimaranan, B. D., ed. 2006. Global Trade, Assistance, and Protection: The GTAP 6 Data Base. West Lafayette, IN: Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University. https://www.gtap.agecon.purdue.edu/databases/v6/v6_doco.asp. Francois, J. F., H. van Meijl, and F. van Tongeren. 2005. "Trade Liberalization in the Doha Development Round." Economic Policy 20 (42): 349­91. Hertel, T. W., ed. 1997. Global Trade Analysis: Modeling and Applications. New York: Cambridge University Press. Hertel, T. W., and R. Keeney. 2006. "What's at Stake: The Relative Importance of Import Barriers, Export Subsidies, and Domestic Support." In Agricultural Trade Reform and the Doha Development Agenda, ed. K. Anderson and W. Martin, chap. 2. London: Palgrave Macmillan; Washington, DC: World Bank. Jensen, H. T., S. Robinson, and F. Tarp. 2002. "General Equilibrium Measures of Agricultural Policy Bias in Fifteen Developing Countries." TMD Discussion Paper 105, Trade and Macroeconomics Division, International Food Policy Research Institute, Washington, DC. Johnson, D. G. 1991. World Agriculture in Disarray, rev. ed. London: St. Martin's Press. Krueger, A. O., M. Schiff, and A. Valdés. 1988. "Agricultural Incentives in Developing Countries: Mea- suring the Effect of Sectoral and Economy-wide Policies." World Bank Economic Review 2 (3): 255­72. Malcolm, G. 1998. "Adjusting Tax Rates in the GTAP Data Base." GTAP Technical Paper 12 (Septem- ber), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/download/580.pdf. Melitz, M. 2003. "The Impact of Trade on Intra-industry Reallocations and Aggregate Industry Productivity." Econometrica 71 (6): 1692­1725. Milanovic, B. 2005. Worlds Apart: Measuring International and Global Inequality. Princeton, NJ: Princeton University Press. Narayanan, B. G., and T. L. Walmsley, eds. 2008. Global Trade, Assistance, and Production: The GTAP 7 Data Base. West Lafayette, IN: Center for Global Trade Analysis, Department of Agricultural Eco- nomics, Purdue University. https://www.gtap.agecon.purdue.edu/databases/v7/v7_doco.asp. Valenzuela, E., and K. Anderson. 2008. "Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980­84." Research Memorandum 13 (December), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925. Valenzuela, E., K. Anderson, and T. W. Hertel. 2008. "Impacts of Trade Reform: Sensitivity of Model Results to Key Assumptions." International Economics and Economic Policy 4 (4): 395­420. Global Welfare and Poverty Effects: Linkage Model Results 85 Valenzuela, E., D. van der Mensbrugghe, and K. Anderson. 2009. "General Equilibrium Effects of Price Distortions on Global Markets, Farm Incomes, and Welfare." In Distortions to Agricultural Incen- tives: A Global Perspective, 1955­2007, ed. K. Anderson, chap. 13. London: Palgrave Macmillan; Washington, DC: World Bank. van der Mensbrugghe, D. 2005. "Linkage Technical Reference Document: Version 6.0." December, World Bank, Washington, DC. http://go.worldbank.org/7NP2KK1OH0. ------. 2006. "Estimating the Benefits: Why Numbers Change." In Trade, Doha, and Development: A Window into the Issues, ed. R. Newfarmer, 59­75. Washington, DC: World Bank. World Bank. 2001. Global Economic Prospects 2002: Making Trade Work for the World's Poor. Washington, DC: World Bank. ------. 2003. Global Economic Prospects 2004: Realizing the Development Promise of the Doha Agenda. Washington, DC: World Bank. ------. 2004. Global Economic Prospects 2005: Trade, Regionalism, and Development. Washington, DC: World Bank. ------. 2005. Global Economic Prospects 2006: Economic Implications of Remittances and Migration. Washington, DC: World Bank. ------. 2006. Global Economic Prospects 2007: Managing the Next Wave of Globalization. Washington, DC: World Bank. ------. 2007. World Development Report 2008: Agriculture for Development. Washington, DC: World Bank. Zhang, X. G., and M. Osborne. 2009. "Developing an Armington-Heckscher-Ohlin Database: Splitting Global Trade (GTAP) Data into Homogeneous and Differentiated Products." Internal Research Memorandum 09-02 (January), Productivity Commission, Melbourne. 3 Global Poverty and Distributional Impacts: The GIDD Model Maurizio Bussolo, Rafael De Hoyos, and Denis Medvedev* Trade liberalization is almost always welfare increasing nationally and globally, but it also brings about large redistributions of income. Simulation models calibrated on real world data show that the aggregate gains for a country that elimi- nates tariffs are, at best, within only a few percentage points of the initial GDP. Similarly, the gains from multilateral trade policy reforms for the whole world tend to be small. In contrast, the losses suffered by specific, initially protected sectors or factors may be much larger. As Rodrik (1998) puts it, the static efficiency conse- quences of trade reform pale in comparison with the redistributive effects. These effects often create complicated policy challenges at domestic and inter- national levels because, in most cases, losers tend to be a smaller and more vocal group than winners.1 Perhaps the most recent and glaring example of this trade- related distributional tension is the impasse in the Doha Round of the World Trade Organization. Disputes over the reduction of agricultural market distortions have stalled the entire multilateral trade negotiation process. The controversy is cen- tered on the demands of developing and agricultural exporting countries to phase out export subsidies and domestic farm supports that are mainly applied in devel- oped countries, in addition to the reduction of import barriers. This example illustrates that a distributional tension among countries may have important implications for international relations and global welfare. An *The authors are grateful to Rebecca Lessem for research assistance and to Will Martin, Hans Timmer, and Dominique van der Mensbrugghe for discussions on the issues. 87 88 Agricultural Price Distortions, Inequality, and Poverty additional question is this: would resolving trade disputes improve the distribu- tion of income not only between countries, but also within national economies? The answer depends, in part, on own-country distortions to agricultural and other producer incentives in individual developing countries. Frequently, the related policies privilege urban dwellers by protecting urban industries and main- taining low prices for food items, which works to the disadvantage of (often) poorer local farmers, although this is much less so now than it was in the 1960s and 1970s (Krueger, Schiff, and Valdés 1988; Anderson 2009). Given that poverty is greatest among farmers, the poverty reduction potential of agricultural trade liberalization is promising (Chen and Ravallion 2007). Using an ex ante simulation analysis, this chapter answers the following ques- tions: How much would global inequality and poverty be reduced if all distortions to trade in agricultural and other goods were removed? How much of this change would be generated by agricultural policy reform only? What share of the change in inequality would arise from changes across countries versus changes within countries (one should bear in mind that a lowering of the inequality between agricultural and nonagricultural groups might be offset by, for example, increased inequality within the agricultural sector)? What would happen to global poverty and to the incidence of poverty in specific countries and regions? Does it matter whether we use the US$1-a-day or US$2-a-day international poverty line (because, for example, more nonagricultural households than farm households may be clus- tered between these two poverty lines)? The empirical results of the study represented in this chapter have been pro- duced using the World Bank's Linkage global general equilibrium model and the newly developed Global Income Distribution Dynamics tool (GIDD) (van der Mensbrugghe 2005; Bussolo, De Hoyos, and Medvedev 2008; see also the GIDD Database). The GIDD is a framework for ex ante analysis of the income distribu- tional and poverty effects of changes in macroeconomic, trade, and sectoral poli- cies and trends in global markets. It complements a global computable general equilibrium (CGE) analysis by providing global microsimulations based on stan- dardized household surveys. The tool pools most of the currently available household surveys covering 1.2 million households in 73 developing countries. Information on households in developed countries completes the data set. Overall, the GIDD sample countries cover more than 90 percent of the world's population.2 The chapter is organized as follows. The next section presents the GIDD data set and the main features of global income distribution as a way of establishing the initial conditions or baseline. This descriptive analysis sets the stage for the subsequent sections, which illustrate the modeling methodology, lay out the reform scenarios, and summarize the main results. The two core simulations involve the case of the liberalization of all merchandise trade and agricultural Global Poverty and Distributional Impacts: The GIDD Model 89 domestic distortions and--to examine the contribution of farm and food policies--the case of the liberalization of only agricultural trade and domestic price-distorting measures. Final remarks are provided in the concluding section. What Is at Stake? The Initial Conditions Almost 45 percent of the world's people, most of them in developing countries, live in households in which agricultural activities represent the main occupation of the heads of household. A large share of this agriculture-dependent group, close to 32 percent, is poor. Agricultural households thus contribute dispropor- tionably to global poverty: three of every four poor people belong to this group. Improving the economic opportunities available in agriculture may therefore significantly affect global poverty and inequality. The specific opportunity consid- ered in this study is the removal of agricultural subsidies and taxes and all merchandise trade distortions. The direct effects of this global liberalization would include changes in the international prices of food and other agricultural products and changes in the returns to the factors used intensively in agriculture. These changes would determine the winners and losers of reform through the impacts on household earnings and spending. Before examining these effects in detail, this section describes the stakes involved in the reforms by considering the socioeconomic characteristics of the world's population, especially the people engaged in agriculture. This ini- tial descriptive analysis is based on the GIDD data set that has been recently developed at the World Bank. The data set consists of 73 detailed household surveys on low- and middle-income countries, complemented by more highly aggregated information on income distribution among 25 high-income coun- tries and 22 other developing countries.3 Together, the 120 countries in the data set cover more than 90 percent of the global population. The country cov- erage varies by region: while the countries in the GIDD data set represent more than 97 percent of the population in East Asia, South Asia, Latin America, and Eastern Europe and Central Asia, the coverage in Sub-Saharan Africa and in the Middle East and North Africa is limited to 76 and 58 percent of the popu- lation, respectively. Among the detailed surveys, the majority (54) rely on per capita consumption as the main indicator of welfare, while the remaining surveys--all but one are on countries in Latin America--rely on per capita income alone as a measure of household welfare. The income and consump- tion data are monthly; the data are standardized to the year 2000 and are expressed in 1993 purchasing power parity (PPP) prices for consistency with the US$1-a-day and US$2-a-day poverty lines, which are calculated at 1993 PPP exchange rates.4 90 Agricultural Price Distortions, Inequality, and Poverty Figure 3.1. Income Distributions among Agricultural and Nonagricultural Populations of the World, 2000 0.4 0.3 density 0.2 0.1 0 1 1 3 5 7 9 monthly household per capita income (1993 PPP, log) global agricultural population global income distribution, 2000 global nonagricultural population poverty line Source: Author compilation based on the GIDD Database. Note: See the text for an explanation of the figure. Three facts about the agricultural sector help identify the welfare effects of a global-scale removal of trade distortions: the proportion of the world's people whose real incomes depend on the agricultural sector, the initial position of the agricultural population in the global income distribution, and the dispersion of incomes among the agricultural population. Based on the GIDD data set, figure 3.1 shows a kernel density of the global income distribution of household per capita income and consumption and kernel densities of the income and consumption of the population inside and outside the agricultural sector.5 The area below the kernel density for the agricultural sector is equal to 0.45, indicating that 45 percent of the world population relies on agricul- ture for their livelihoods. The distribution of the agricultural population is located to the left of the nonagricultural distribution, implying that households in the agricultural sector earn, on average, lower incomes than their counterparts in other sectors. In U.S. PPP dollars, the average agricultural household's per capita monthly income is US$65, only 20 percent of the US$320 per capita income earned by the average household in the nonagricultural group. The differences in the shapes of the two distributions corroborate the hypothesis of Kuznets more than 50 years ago that incomes in the traditional sector are less well dispersed than incomes in the modern sectors. A more egalitarian traditional sector is indicated by the taller and thinner distribution for the agricultural population in figure 3.1. Global Poverty and Distributional Impacts: The GIDD Model 91 Income inequality is estimated based on the global income distribution data depicted in figure 3.1. The Gini coefficient for the world is equal to 0.67, which denotes a high level of inequality. In fact, this global Gini is about 0.28 points higher (worse) than the Gini of the United Sates and even higher than the level observed in extremely unequal countries such as Mexico. As Bourguignon, Levin, and Rosenblatt (2004: 15) note, "if the world [were] a country, it would be one of the most unequal countries in the world." How much of this inequality may be explained by the disparity in average incomes between the agricultural group and the rest? Inequality decomposition analysis shows that one-quarter of the global income disparities may be explained by the difference in average incomes between the two groups of households; the remaining three-quarters are generated by within-group income variation. Based on the preestablished poverty line of US$1 a day in PPP dollars, the GIDD global income data also provide information about the differences in poverty inci- dence among the two population subgroups. Despite the fact that incomes are more evenly distributed among the agricultural population (the Gini coefficient is 0.18 points lower in agriculture than in nonagriculture), lower average incomes in this sector result in a much higher poverty incidence: 32 percent of agricultural households are poor compared with 8 percent of nonagricultural households. In terms of the personal characteristics of the poor inside and outside the agri- cultural sector, no noticeable differences are observed in the average age of the heads of household or in the household size. However, poor people in agriculture tend to have lower educational attainment: only 32 percent have completed pri- mary education compared with 46 percent among nonfarm households. In agri- culture, poor woman-headed households are a small minority, less than 9 percent, which is significantly below the 14 percent observed among nonagricultural households (table 3.1). Up to this point, our welfare information on agricultural and nonagricultural populations has been derived by agglomerating all households within these two groups irrespective of nationality. In fact, the kernel densities in figure 3.1 exploit Table 3.1. Characteristics of the Poor in Agricultural and Nonagricultural Households, Developing Countries, 2000 Head of household Age of Household Woman- completed household members, headed Population primary head, average households, Household type share, % school, % average years number % Agricultural 44.8 32.3 44.8 7.11 8.7 Nonagricultural 55.2 46.0 44.4 7.06 14.0 Source: Author compilation based on the GIDD Database. 92 Agricultural Price Distortions, Inequality, and Poverty Figure 3.2. Relationship between Income Levels and Share of Employment in Agriculture, by Country, 2000 80 employment in agriculture (%) China 60 India 40 20 Brazil European Union United States Russian Federation 0 0 500 1,000 1,500 monthly household per capita income (1993 ppp) Source: Author compilation based on the GIDD Database. Note: Area of symbol proportional to the country's population. full income heterogeneity across households, including variations across and within countries. Countries display large differences in terms of population size, level of development, and the importance of the agricultural sector in the econ- omy. These three country-specific characteristics are important determinants in any prospective change in global poverty and global inequality. As we may see in figure 3.2, global poverty would be strongly reduced if China and India were to achieve higher income levels. Given the initial large share of these countries in the global population and their position in the global income distribution, the eco- nomic expansion of these two giants is a key factor shaping the evolution of the world economy.6 Figure 3.2 also depicts a negative relationship between income levels and the share of workers in agriculture, and, although it is imperfectly inferred from a cross-section of countries at a particular point in time, this rela- tionship suggests that structural shifts would likely affect income distribution within countries. Figure 3.3 shows this heterogeneity by displaying, for each county in our sample, the shares of agriculture in total population and in national income. Given the large variation in the proportion of the population for which the incomes depend on the agricultural sector, the income effects following a Figure 3.3. Share of the Population in Agriculture and of Agriculture in Total Income, Developing Countries, 2000 Burundi Burkina Faso Lao PDR Nepal Tanzania Ethiopia Kenya Madagascar Cambodia Vietnam Bangladesh Uganda Cameroon Yemen, Rep. Kyrgyz Republic Guyana Albania Gambia, The Nigeria Côte d'Ivoire Pakistan Moldova India Ghana Haiti Thailand Georgia Morocco Philippines Guatemala Uzbekistan Kazakhstan Honduras Armenia country Azerbaijan Sri Lanka Nicaragua Indonesia Turkey Paraguay Bolivia Tajikistan Peru Mali Macedonia, FYR Guinea Dominican Republic Jamaica Costa Rica Colombia South Africa El Salvador Ukraine Senegal Panama Benin Romania Lithuania Brazil Venezuela, R.B. de Chile Mauritania Poland Ecuador Estonia Belarus Hungary Mexico agriculture incomes (% of total) Jordan agriculture populations (% of total) Russian Federation Bulgaria 0 10 20 30 40 50 60 70 80 90 100 percent Source: Author compilation based on the GIDD Database. Note: The figure is based on the data set for developing countries and transition economies. 94 Agricultural Price Distortions, Inequality, and Poverty removal of agricultural distortions would differ greatly between countries and be especially important in countries in which more than half the population is in agriculture. The majority of these agriculture-based countries are located in the poorest region of the world, namely, Sub-Saharan Africa. Of the 25 countries in the agriculture-based group, 12 are in Sub-Saharan Africa, 4 are in South Asia, 3 are in East Asia, 3 are in Eastern Europe and Central Asia, 2 are in Latin America (Guyana and Haiti), and only 1 (the Republic of Yemen) is in the Middle East and North Africa. The pattern observed at the global level, namely, that agriculture-dependent households, on average, earn less than other households, is replicated in all devel- oping countries in the GIDD Database (figure 3.3); the share of the total popula- tion employed in agricultural activities is always larger than the share of this population in total income. The average income of nonagricultural households is 2.25 times greater than the average agricultural household income. This difference is unconditional in the sense it does not take into account the fact that, in agricul- ture, low-earning unskilled workers tend to be more abundant than skilled work- ers or that other factors may explain the observed income gap. However, we have undertaken a simple multivariate regression analysis on the GIDD data set, and it shows that, even controlling for education, age, gender, household size, geo- graphic region, and country fixed-effects, agriculture-related incomes are still 23 percent lower than incomes derived from other sectors. An important element hidden in figure 3.1 and only partially shown in figure 3.3 is the degree of cross-country variation in income inequality. Figure 3.4 shows that the differences in the Gini coefficient across countries are enormous. Former cen- trally planned economies such as Hungary, Romania, and Ukraine show a coeffi- cient below 0.3, whereas, in South Africa and much of Latin America, the coefficient reaches values well above 0.5. Once again, the tendency toward higher inequality within the agricultural group observed at the global level is corroborated by analysis of country-specific inequality. For more than three-quarters of the countries included in our data (56 of 73), Gini indicators of inequality within the agricultural group are higher than those within the nonagricultural group (figure 3.4). A global trade reform would be expected to reallocate resources intersectorally within national states and between them. Given the global variations in the impor- tance of the agricultural sector, the ratio of nonagricultural to agricultural incomes, and the within-sector income inequality, the resource reallocation following the trade reform would have significant distributional effects across and within coun- tries. Is economic theory able to provide guidance on the expected global welfare effects after the removal of agricultural distortions and other trade distortions? As shown by Winters (2002) and McCulloch, Winters, and Cirera (2001), trade liberalization and household welfare are linked via prices, factor markets, and Figure 3.4. Inequality Variation in Agricultural, Nonagricultural, and All Households, Developing Countries, 2000 Denmark Finland Sweden Netherlands Korea, Rep. Austria Luxembourg Germany Albania France Tajikistan Switzerland Canada Australia European Union Spain Greece Italy New Zealand Georgia Portugal Ghana Israel Nigeria United States Singapore China Hong Kong, China Gambia, The El Salvador Mexico Nicaragua Colombia Ecuador Bolivia Brazil Guatemala Hungary Ukraine Czech Republic Slovenia Pakistan Belarus Russian Federation Romania Kyrgyz Republic Lithuania Armenia country Kazakhstan Ethiopia Bulgaria Poland Mongolia Bangladesh Estonia India Indonesia Yemen, Rep. Ireland Uzbekistan Moldova Tanzania Vietnam Lao PDR Nepal Benin Azerbaijan Turkey Macedonia, FYR Sri Lanka United Kingdom Morocco Mali Senegal Guinea Thailand Mauritania Jamaica Guyana Cambodia Jordan Burundi Venezuela, R.B. de Uganda Cameroon Kenya Côte d'Ivoire Costa Rica Madagascar Philippines Burkina Faso Dominican Republic Gini in agriculture Chile Gini in nonagriculture Peru Honduras overall Gini Panama Paraguay South Africa Haiti 0.15 0.25 0.35 0.45 0.55 0.65 0.75 Gini coefficient Source: Author compilation based on the GIDD Database. 96 Agricultural Price Distortions, Inequality, and Poverty consumer preferences. The international prices of many agricultural products would increase after the removal of trade barriers (see chapter 2, by Anderson, Valenzuela, and van der Mensbrugghe; see also Anderson and Martin 2005). If we assume some degree of pass-through, the increase in international prices would be followed by a rise in domestic agricultural prices, which would encourage a redistribution of resources from nonagricultural activities to the agricultural sec- tor of the economy. Based on figure 3.1, this redistribution might help reduce global poverty and inequality. However, household consumption patterns would also change as a result of the shift in prices, making the link between trade liberal- ization and global household welfare more complex. As a consequence of the agri- cultural price changes, a redistribution of real income would take place between net sellers and net buyers of agricultural products; the welfare of the former would improve at the expense of the welfare of the latter.7 Finally, factor prices would also change after trade liberalization, thus changing the real incomes of house- holds that are not directly involved in agricultural production. The transition from trade theory to real world analysis presents serious chal- lenges. A sound empirical strategy must estimate the effects of the reform on prices, monetary incomes (via profits in the case of farm households and returns to factors of production in the case of nonfarm households), consumption, and transfers.8 The framework used in this chapter and described in more detail below accounts for the impact of trade liberalization through at least some of these channels. Methodology The empirical analysis in this chapter relies on the GIDD Database, a newly devel- oped tool for analyzing the dynamics of global income distribution.9 Compiled by the Development Prospects Group of the World Bank, the GIDD combines a con- sistent set of price and volume changes from a global CGE model with microdata at the household level to create a hypothetical or counterfactual income distribution capturing the welfare effects of policies under evaluation.10 The GIDD therefore has the ability to map CGE-consistent macroeconomic outcomes to disaggregated household survey data. The GIDD framework is based on microsimulation methodologies developed in the recent literature, including Bourguignon and Pereira da Silva (2003), Ferreira and Leite (2003, 2004), Chen and Ravallion (2003), and Bussolo, Lay, and van der Mensbrugghe (2006). The starting point is the global income distribution in 2000, which has been assembled using data from household surveys (see above).11 The hypothetical distribution is then obtained by applying three main exogenous changes to the initial distribution: changes in relative wages across skills and sectors for each country, changes in household purchasing power Global Poverty and Distributional Impacts: The GIDD Model 97 Figure 3.5. The GIDD Methodological Framework removal of global agricultural distortions CGE--linkage real wages (by growth effects skill and food prices sector) hypothetical income distribution Source: Construction by the authors. caused by shifts in food prices and other prices, and changes in the average level of welfare (real income) for each country. The methodological framework used here is depicted in figure 3.5. The starting point is the price and quantity effects after the removal of trade distortions; one computes these by using the global CGE Linkage model (top part of figure 3.5). The CGE model will compute the values of the three variables linking the macro- and microlevels of the model (middle part of figure 3.5): overall economic growth, real wage premiums among agricultural and nonagricultural and skilled and unskilled groups, and the consumption (or real income) effects brought about by the change in the relative prices of food. These CGE results are passed on to the household survey data, creating a new, hypothetical household income distribution (bottom part of figure 3.5). This is accomplished by differentiating four types of households: those in which the household head is an: (1) unskilled agricultural worker, (2) skilled agricultural worker, (3) unskilled nonagricultural worker, or (4) skilled nonagricultural worker. The initial income premium earned by groups (2)­(4) relative to group (1) is changed in accordance with changes in the wage premiums in the CGE model, which uses the GIDD information on a number of workers in each of the segments (1)­(4). For example, if, initially, a household headed by an unskilled worker in nonagriculture earns 50 percent more than a household headed by an unskilled worker in agriculture and if the CGE results show that this premium would decline by a tenth, the microsimula- tion part of the GIDD changes the incomes of all households headed by unskilled 98 Agricultural Price Distortions, Inequality, and Poverty workers in nonagriculture such that the new wage premium is 45 percent. In addi- tion to these wage shocks, the GIDD also accounts for changes in relative prices and changes in per capita incomes, thereby indirectly picking up the impact on the returns to factors other than labor. In the real world, the changes depicted in figure 3.5 take place simultaneously, but, in the GIDD simplified framework, they are accommodated in a sequential fashion. In the first step, changes in labor remuneration by skill level and sectoral location are applied to each household in the sample, depending on educational attainment and sector of employment. In the second step, consistent with an over- all growth rate of real income per capita, real household incomes are affected by the change in the prices of food items versus the prices of nonfood; households with a higher share of household income allocated to food consumption bear a larger proportional impact after a change in the prices of food. Comparisons between the initial and the counterfactual income distributions capture the welfare and inequality effects of the removal of global trade distor- tions. By taking into account labor market effects (returns to skills in the agricul- tural sector and in the rest of the economy) and consumption effects, the GIDD framework closely maps the theoretical links outlined in the previous section.12 However, the framework reshapes national income distributions under a set of strong assumptions. In particular, income inequality within population sub- groups formed by skills and by sector of employment is assumed to remain con- stant after the trade reform. Moreover, data limitations affect the estimates of the initial inequality and of the evolution of inequality. Although consumption expenditure is a more reliable welfare measure than income and although its dis- tribution is normally more equal than the distribution of income, consumption data are not available in all country surveys. To obtain a global picture, we have been obliged to include countries for which only income data are available, along with countries for which consumption information is available. Finally, measure- ment errors implicit in the PPP exchange rates that have been used to convert local currency units also affect the comparability across countries. The resulting hypothetical income distribution should thus not be considered a forecast of the future distribution; rather, it should be interpreted as the result of an exercise that captures the distributional effect of trade liberalization, all else being equal. What Happens to Poverty and Income Distribution If Trade Is Liberalized Globally? In this section, we link the macro-outcomes of global agricultural reform and other trade policy reforms to the changes in the distribution of income among and within countries. Our analysis is carried out in three stages. First, we briefly Global Poverty and Distributional Impacts: The GIDD Model 99 examine the macroeconomic results from the Linkage model simulations of global trade reform (which are similar, but not identical to those presented in chapter 2, by Anderson, Valenzuela, and van der Mensbrugghe) and focus on the variables that are passed on to the household survey data.13 Second, we consider the income distributional results from a global perspective by quantifying the likely changes in global poverty and inequality and identifying groups of coun- tries and individuals that are likely to benefit the most (least) from global trade reform. Third, we assess the potential trends in the distribution of income within countries by identifying countries in which inequality and poverty pressures may heighten and thus erode the support for additional reforms. Macroeconomic general equilibrium results We have carried out the Linkage simulation analysis using version 7 of the Global Trade Analysis Project Database, which has been amended by Valenzuela and Anderson (2008) to take account of new estimates of the distortions to agricultural incentives in developing countries compiled by these same authors. The Linkage model disaggregates global trade into bilateral flows between more than 100 coun- tries and regions in 57 commodity groups. The base year for the simulations is 2004, and the baseline data have taken into account changes in the global trade and tariff structure generated by the implementation of the Uruguay Round commitments, European Union enlargement, China's accession to the World Trade Organization, and most major preferential trade agreements in place at the time. The model is solved in a comparative static mode, which means that simulations are imple- mented as one-time shocks and do not take into account potential growth effects through changes in capital accumulation rates or variations in productivity. Our two simulations envision the full removal globally of trade taxes and sub- sidies on all agricultural goods and lightly processed food without and then with the trade reform of nonagricultural goods. With these two scenarios, we are able to see the relative contribution to changes in the global economy after the removal of agricultural distortions alone. The removal of distortions to the trade in agricultural products causes global consumption to rise by 0.29 percent, or two-thirds of the improvement expected under a trade liberalization scenario involving all goods. Developing countries gain more than the average; their consumption rises by 0.47 percent compared with 0.24 percent for high-income countries. No less than 50 of 60 Linkage coun- tries and regions--representing nearly 95 percent of the world's people--would experience positive changes in consumption following the removal of agricultural distortions compared with 47 countries and regions that would enjoy consump- tion gains from the liberalization of all goods trade (figure 3.6). 100 Agricultural Price Distortions, Inequality, and Poverty Figure 3.6. The Effects on Real National Consumption of the Global Removal of Agricultural and All Merchandise Trade Distortions Ecuador Lithuania New Zealand Bulgaria Nigeria Rest of North Africa Colombia Rest of Western Europe Argentina Mozambique Slovak Republic Estonia Czech Republic Brazil Morocco Slovenia Poland Russian Federation Rest of East Asia Vietnam Nicaragua Hungary Malaysia Thailand Zimbabwe Latvia Rest of Latin America and Caribbean Korea, Rep. Indonesia country Rest of Eastern Europe and Central Asia Kazakhstan Philippines Middle East EU15 Bangladesh Romania Turkey Australia Japan Zambia Canada Rest of Sub-Saharan Africa Tanzania Taiwan, China Pakistan India Hong Kong, China; Singapore United States Egypt, Arab Rep. South Africa Other South Asia Sri Lanka Mexico China Rest of West and Central Africa Chile global liberalization of trade in Madagascar agriculture and food products Uganda global liberalization of all trade Senegal Kyrgyz Republic 4 2 0 2 4 6 8 10 12 percent change in real consumption Source: Author compilation using Linkage model results similar to those in chapter 2, by Anderson, Valenzuela, and van der Mensbrugghe, except that factor mobility is more limited. Note: The darker bars show the percent increase in consumption (at prereform prices) generated by the removal of price and trade distortions in agriculture and food products (excluding beverages and tobacco). The lighter bars show the additional gains in consumption generated by the removal of all remaining barriers to merchandise trade. The combined length of the two bars shows the consumption gains from the global trade reform of all merchandise. EU15 the 15 members of the European Union prior to 2004. Global Poverty and Distributional Impacts: The GIDD Model 101 There are three main channels that transmit the trade reform shocks to house- hold consumption in the Linkage model and help explain the heterogeneity of the results in figure 3.6. The first channel is changes in the terms of trade, that is, the ratio of export prices to import prices, without taking into account domestic price distortions caused by own-country policies. Net exporters of agriculture and food, such as Brazil, Ecuador, and New Zealand, reap significant welfare gains if the export prices of their farm commodities rise by an average of 8, 19, and 11 percent, respec- tively.14 On the other hand, net importers of food, such as China, Mexico, and Senegal, experience real consumption losses due to higher import prices. The second channel is tightly linked to the first and involves the impact of own-country policies. Thus, countries with high prereform tariffs or export taxes, such as Lithuania, Nigeria, and the group in North Africa, tend to experience larger consumption gains relative to countries in which the initial distortions are low. If initial agricultural import barriers are sufficiently high, consumers may face lower postreform prices for food even if import prices are rising. This is the case in North Africa, which experiences an increase in real consumption although it is a net food importer. The third channel is the impact of trade reform on government budgets. Since the model does not include an explicit transversality condition, we maintain a fixed budget deficit closure, which means that any losses in public revenue (such as a reduction in tariff income) must be offset by a compensatory increase in the direct tax rate on households.15 Therefore, welfare gains are more limited in coun- tries, such as Tanzania and Zimbabwe, that rely heavily on international trade taxes as an important provider of public revenue.16 In addition to changes in the levels of per capita consumption across countries, the Linkage results hint at important distributional consequences of trade reform within countries, which come about through changes in returns to labor in differ- ent sectors and at varying skill levels. Figure 3.7 shows the contributions of the payments for different factors to the total change in real GDP at factor cost (in percentage points) after the removal of agricultural distortions alone. With the exception of China, all the countries experiencing an increase in the payments for unskilled labor in agriculture also register consumption gains after agricultural policy reform. However, the converse does not hold: real consumption rises in 32 of the 41 countries that show a decline in unskilled agricultural wages. Since unskilled workers in agriculture tend to be in the poorest part of the population, these results suggest that pressures toward increased inequality may intensify.17 Furthermore, losses and gains in agricultural wages exhibit strong regional pat- terns: the real wages of unskilled farmers rise in Latin America, the Middle East, and East Asia, but decline in other developing regions and decline much more in high-income countries. 102 Agricultural Price Distortions, Inequality, and Poverty Figure 3.7. The Effects on National Real Factor Rewards of the Global Removal of Only Agricultural Price and Trade Policies Argentina Ecuador Mozambique New Zealand Colombia Bulgaria Philippines Vietnam Zimbabwe Kazakhstan Middle East Brazil Rest of Latin America and Caribbean Thailand China Rest of North Africa Rest of East Asia Australia Rest of Eastern Europe and Central Asia Hong Kong, China; Singapore Canada unskilled agricultural labor Indonesia unskilled nonagricultural labor Chile Senegal skilled agricultural labor Egypt, Arab Rep. skilled nonagricultural labor United States capital Malaysia Mexico land Estonia country South Africa Pakistan Other South Asia Kyrgyz Republic Slovenia Nicaragua Taiwan, China Tanzania West and Central Africa Japan Rest of Western Europe Russian Federation Latvia Sri Lanka Zambia Bangladesh Romania EU15 Slovak Republic Rest of Sub-Saharan Africa Korea, Rep. Turkey Hungary Lithuania Czech Republic India Uganda Poland Nigeria Madagascar Morocco 4 3 2 1 0 1 2 3 4 5 percent change in real value added Source: Author compilation using Linkage model results similar to those in chapter 2, by Anderson, Valenzuela, and van der Mensbrugghe, except that factor mobility is more limited. Note: Each bar shows a contribution of the changes in the value added by a specific factor to the total change in value added, deflated by the price of GDP at factor cost. Countries are sorted (in descending order) by the increase in the payments to unskilled farm labor. Global Poverty and Distributional Impacts: The GIDD Model 103 The initial level of protection in agriculture, combined with the terms of trade shock, represents the main determinant of the trends in farm factor prices. Con- sider the example of India, where unskilled farm wages decline by 6.1 percent fol- lowing trade reform.18 Indian farmers must contend with the falling international prices of imported farm products (a decline of 1.7 percent), as well as a loss in tar- iff protection (2.0 percent), export subsidies (3.3 percent), and output subsidies (6.9 percent) if all farm policies are scrapped. The first two channels decrease the competitiveness of the farmers on the domestic market and lead to higher import penetration; the third channel erodes their competitiveness on international mar- kets; and the fourth channel increases production costs and makes Indian farmers less competitive overall. Together, these effects result in a strong negative shock to farm labor earnings. In Mexico, the income losses among unskilled farmers are lower than they are in India. This is partially attributable to the close trading relationship between Mexico and the United States. Mexico purchases 75 percent of its agricultural imports from the United States, the export prices of which rise by 5.7 percent because of the elimination of export and production subsidies. Thus, the removal of agricultural price supports in the United States exerts upward pressure on the import prices of farm products in Mexico, which hurts consumers, but increases the competitiveness of Mexican farmers on the domestic market. The trend is counteracted by the removal of Mexico's tariff protection on agriculture (1.2 per- cent) and farm output subsidies (0.8 percent), which leads to a decrease in the competitiveness of farmers in Mexico and market share losses among them in the domestic market and in export markets. In contrast, Brazil is an example of a country in which a number of positive developments combine to produce a nearly 34 percent gain in the wages of unskilled agricultural workers.19 The import prices of farm products in Brazil rise by 1.8 percent, bolstering the domestic competitiveness of Brazilian farmers, while export prices increase by more than 10 percent. Because the farmers do not receive export or production subsidies, they are well positioned to take advantage of these opportunities and gain market share both domestically and abroad. Although some of the gains among agricultural producers are offset by the loss in domestic protection (import tariffs of 2.4 percent), Brazilian agriculture is still able to increase production volumes by 18 percent after farm trade reform. Microsimulation results: A first look at global poverty and inequality effects In this subsection, we use the GIDD model and data to simulate the likely changes in global poverty and inequality arising from the elimination of trade distortions. 104 Agricultural Price Distortions, Inequality, and Poverty Given the richness of the data and the numerous factors affecting global poverty and inequality within the GIDD, this subsection starts with two simulations that illustrate, in a simple way, the expected effects of global trade policy reform. Focusing only on the developing countries in our data set, both simulations raise the average income in the developing world by 1 percent. In the first simulation, national income increases because of an increase in the incomes of agricultural households only, while, in the second simulation, the increase originates entirely in an expansion in nonagricultural incomes. The results of these examples are shown in two growth incidence curves (GICs) in figure 3.8.20 The broken GIC line captures the effects of assigning income gains only to agricultural households, while the solid GIC line raises incomes only among those households in which the heads work in nonagricultural activities. As expected, the increase in agricultural incomes is more pro-poor than the same income change taking place in other sectors in which households are rela- tively richer. This pro-poor bias in the growth in agricultural incomes is reflected in the downward slope of the broken GIC line, indicating that the poorest households Figure 3.8. National GICs: The Effects on the Per Capita Household Income Distribution of a Hypothetical 1 Percent Increase in Agricultural Versus Nonagricultural Incomes, Developing Countries 10 percent rise in household income 0.8 0.6 0.4 0.2 0 0 20 40 60 80 100 percentiles 1% increase in nonagricultural incomes 1% increase in agricultural incomes extreme poverty line Source: Author compilation based on the GIDD Database. Note: The figure is based on the data set for developing countries only. Global Poverty and Distributional Impacts: The GIDD Model 105 reap the largest income gains. A different way of interpreting the results shown in figure 3.8 is to observe that, if the agricultural sectors in all developing countries receive income gains above the gains in the nonagricultural sectors as an outcome of the elimination of distortions, global poverty and inequality would fall. As shown in the discussion below, the reality tends to be much more complex than these simple simulations, but the central message of figure 3.8 nonetheless captures the essence of the GIDD simulations. Impacts on global poverty and inequality Using the GIDD to translate the shocks from the Linkage model into poverty and inequality outcomes, we find that the effects of the full removal of agricultural price and trade distortions on extreme poverty globally are close to zero (table 3.2, col- umn 3). This limited impact is explained by several factors. First, the estimated impacts arise from a comparative static model rather than a dynamic model, and, so, they do not capture the growth effects of the reform. Thus, the changes in per capita consumption are small and arise only because of the income boost from more- efficient resource allocation. According to the GIDD, the world's average monthly Table 3.2. The Household Effects of the Removal of Agricultural and All Merchandise Trade Distortions on Global Poverty and Inequality (percentage point change and U.S. dollars) US$1-a-day US$2-a-day Real average poverty, % poverty, % Gini monthly income, Household type coefficient 2000 US$ PPP Incidence Share Incidence Share Baseline Agricultural 0.45 65 31.5 76 73.8 70 Nonagricultural 0.63 320 8.3 24 26.7 30 All households 0.67 204 18.9 100 48.2 100 Agricultural liberalization, percentage point change Agricultural 0.7 1.1a 0.86 1.13 0.86 0.47 Nonagricultural 0.1 0.2a 0.29 1.13 0.90 0.47 All households 0.1 0.3a 0.23 0.00 0.88 0.00 All merchandise trade liberalization, percentage point change Agricultural 0.8 0.8a 1.09 1.02 0.66 0.57 Nonagricultural 0.2 0.4a 0.19 1.02 0.95 0.57 All households 0.0 0.4a 0.39 0.00 0.82 0.00 Source: Author compilation based on the GIDD Database and using Linkage model results similar to those in chapter 2, by Anderson, Valenzuela, and van der Mensbrugghe, except that factor mobility is more limited. a. Changes in average income are expressed in percentages. 106 Agricultural Price Distortions, Inequality, and Poverty household income increases 0.3 percent after the removal of only agricultural dis- tortions, passing from an initial level of US$204 to a final value of US$210 at 1993 PPP. If nonagricultural trade is also liberalized, global income rises slightly more, but the total number of households living on less than US$1 a day remains the same. Second, there is an income redistribution from farm households to nonfarm households, so that the incidence of extreme poverty among farm households rises by around 1 percent in both scenarios, while the incidence among nonfarm households falls by 0.2 or 0.3 percent. As a result, the share of extremely poor households that are farm households rises from 76 to 77 percent (table 3.2, column 4). Third, the definition of poverty is important. The extreme level, at less than US$1 a day, suggests that only 8 percent of nonfarm households are poor and that they account for only one in four poor households, whereas, at the moderate poverty def- inition, at less than US$2 a day, the corresponding share is 27 percent, accounting for nearly one in three poor households. If we use the moderate definition of poverty, we find that both agricultural trade liberalization and all merchandise trade liberaliza- tion globally lower the total poverty incidence by nearly 1 percent by reducing it among farm and nonfarm households (compare table 3.2, columns 3 and 5). The policy reforms also have only a slight impact in reducing income inequal- ity at the global level. Incomes rise in both the agricultural and nonagricultural sectors, and agricultural incomes increase by twice as much in the all-goods reform scenario and by five times as much in the reform of agriculture only (1.1 percent compared with only 0.2 percent; see table 3.2, column 2). Yet, while the reduction in the nonagricultural income premium, on its own, reduces inequality, income dispersion within the agricultural sector also increases such that the final change in global income distribution is close to zero, judged accord- ing to the changes in the Gini coefficient shown in table 3.2, column 1. Because of these distributional changes taking place within the agricultural sector, the incidence of extreme poverty (under US$1 a day at PPP) rises in the agricultural sector. It increases by 0.9 percentage points as a consequence of the elimination of agricultural price and trade distortions and by 1.1 points if non- farm trade policies are also reformed. Meanwhile, poverty among nonagricul- tural households experiences a reduction equal to 0.3 or 0.2 percentage points, respectively. The combination of poverty changes occurring inside and outside the agricultural sector increases extreme poverty by 0.2 percentage points with respect to agricultural reform alone and by 0.4 points if all goods markets are freed, the latter pushing 9 million additional individuals below the extreme poverty line of US$1 a day. We must bear in mind, though, that the poverty effect of these reforms depends on where the poverty line is set. While the global number of people in poverty measured by the US$1-a-day poverty line shows Global Poverty and Distributional Impacts: The GIDD Model 107 Table 3.3. The GIDD Model: The Effects of the Removal of Agricultural and All Merchandise Trade Distortions on the Incidence of Poverty, by Region (number and percent) Change in number of poor, global trade reform Share of Agriculture only All merchandise global Region poverty, % Millions % Millions % Extremely poor, under US$1 a day East Asia 24 6.4 2.8 6.3 2.8 South Asia 50 15.4 3.3 18.2 3.9 Sub-Saharan Africa 21 1.0 0.5 0.5 0.3 Latin America 4 2.8 6.9 3.5 8.7 Globala 100 5.0 0.5 8.9 1.0 Moderately and extremely poor, under US$2 a day East Asia 33 12.8 1.6 13.2 1.7 South Asia 46 3.6 0.3 2.0 0.2 Sub-Saharan Africa 14 0.1 0.0 1.1 0.3 Latin America 4 4.8 4.6 5.7 5.4 Globala 100 22.1 0.9 19.8 0.8 Source: Author compilation based on the GIDD Database and using Linkage model results similar to those in chapter 2, by Anderson, Valenzuela, and van der Mensbrugghe, except that factor mobility is more limited. a. Includes Middle East and North Africa, Eastern Europe and Central Asia, and high-income countries. Together, these account for no more than 2 percent of the world's poor. an increase of 1.0 percent as a consequence of the reform in all goods, the num- ber measured using the US$2-a-day criterion actually falls by about 20 million or 0.8 percent (table 3.3, columns 4 and 5). These results treat the world as a single entity, making no distinction among regions or countries. Indeed, the lack of major changes at the global level may be the outcome of offsetting impacts across regions. Farmers in Latin America are big winners from agricultural price and trade reform, with an impressive increase of 16 percent in household incomes. In contrast, the incomes of farmers in South Asia would shrink more than 3 percent if agricultural distortions were dismantled globally. To show the incidence of these changes among the population across regions, figure 3.9 plots the GICs for Latin America, South Asia, and the rest of the world (see chart a). The GIC for Latin America indicates that agriculture-based growth in the region is quite pro-poor. In contrast, agriculture-based growth in South Asia is highly regressive: the poorest households are the losers in such reform. East Asia and, to a lesser extent, Sub-Saharan Africa would benefit from 108 Agricultural Price Distortions, Inequality, and Poverty Figure 3.9. Regional and National GICs: The Effects on the Per Capita Household Income Distribution of Full Global Agricultural Policy Reform a. Regional 12 % of poor in % of poor in South Asia Latin America 10 % change in per capita income 8 6 4 2 0 2 4 0 20 40 60 80 100 percentiles Latin America other developing countries South Asia b. National: Brazil and India 10 % of poor in Brazil % of poor in India % change in per capita income 5 0 5 0 20 40 60 80 100 percentiles India Brazil Source: Author compilation based on the GIDD Database. Note: The figure is based on the data set for developing countries only. Global Poverty and Distributional Impacts: The GIDD Model 109 global agricultural reform, and the effects of such reform would be progressive, albeit small, in the rest of the world. The differences in reform outcomes across regions help explain the lack of sig- nificant change in global poverty. With nearly a half billion of its people in extreme poverty and another 625 million moderately poor, South Asia alone accounts for about half of global poverty, while Latin America accounts for less than 5 percent (table 3.3, column 1). Hence, although removing agricultural dis- tortions reduces extreme poverty in most regions of the world, the increase in the headcount ratio in South Asia slightly more than offsets these gains. The results based on the US$2-a-day poverty line show a more positive picture. Poverty gauged by this standard is reduced in all key regions except (marginally) Sub-Saharan Africa (table 3.3, final column). The results at the moderate poverty line are particularly interesting in the case of South Asia, where agricultural reform becomes pro-poor, instead of antipoor as it was according to the US$1-a-day extreme poverty line. This result is explained by the large number of nonagricultural households between the two poverty lines in South Asia; these would experience an increase in purchasing power if global agricultural markets were liberalized. The reduction in moderate poverty is somewhat less if the trade in nonfarm goods is also liberalized, but is still 0.8 percent, or 20 million people globally. Poverty and inequality effects within regions and countries Global agricultural liberalization has distributional and poverty effects that vary not only across regions, but also between neighboring countries and within coun- tries. This subsection summarizes the poverty effects and distributional changes in each of the countries included in our sample. Table 3.3 shows that the 5 million individuals who would be pushed into poverty as a consequence of agricultural reform may be broken down into an increase of 15 million people living in poverty in South Asia and a decrease of 10 million people living in poverty in all other regions. Figure 3.10 shows the specific countries that contribute, respec- tively, the most to the reduction and the most to the increase in global poverty. On the one hand, among the new poor, 92 percent--about 15 million--are in India, while 2.2 percent are in Mexico, and 1.8 percent are in South Africa; both Mexico and South Africa currently protect their farmers and would face higher prices for imported farm products. On the other hand, the gross reduction in global poverty is distributed more evenly among many winners among countries; the great majority are located in East Asia and Latin America. In fact, no country in East Asia and only Chile and Mexico in Latin America experience an increase in the number of extreme poor as a result of global agricultural reform, and even the lat- ter two would experience poverty reduction if the reform were broadened to include nonfarm goods as well (table 3.4). 110 Agricultural Price Distortions, Inequality, and Poverty Figure 3.10. Changes in Poverty as a Share of the Total Change among the Greatest Winners and Losers in Full Global Agricultural Policy Reform, Developing Countries a. Share of the global increase in poverty India 92.0 Mexico 2.2 South Africa 1.8 Kenya 0.7 country Pakistan 0.7 Morocco 0.7 Bangladesh 0.5 Turkey 0.3 Sri Lanka 0.2 rest of the world 1.0 0 20 40 60 80 100 percent b. Share of the global reduction in poverty Philippines 38.6 Brazil 12.1 Nigeria 11.5 Indonesia 8.3 Vietnam 6.1 country Peru 5.5 Thailand 2.1 Yemen, Rep. 2.1 Ecuador 1.9 Colombia 1.6 rest of the world 10.1 0 10 20 30 40 percent Source: Author compilation based on the GIDD Database. Note: The figure is based on the data set for developing countries only. Global Poverty and Distributional Impacts: The GIDD Model 111 Table 3.4. The Effects of the Full Global Liberalization of Agricultural and All Merchandise Trade on the Incidence of Inequality and Poverty, Developing Countries (percentage point change) Agricultural reform only All merchandise trade reform US$1 US$2 US$1 US$2 Gini a day, a day, Gini a day, a day, Region, country coefficient headcount headcount coefficient headcount headcount East Asia 0.72 0.38 0.76 0.62 0.37 0.78 Cambodia 0.75 0.40 0.17 0.69 0.90 0.05 China 0.37 0.00 0.00 0.35 0.00 0.00 Indonesia 0.05 0.49 0.83 0.04 0.57 1.26 Philippines 6.88 6.31 9.91 6.65 6.15 9.80 Thailand 1.25 0.40 2.29 1.10 0.38 2.29 Vietnam 1.85 0.89 3.57 0.08 0.66 2.97 South Asia 0.82 1.16 0.27 0.81 1.37 0.15 Bangladesh 0.46 0.06 0.19 0.09 0.07 0.05 India 1.01 1.49 0.33 1.04 1.71 0.26 Nepal 0.05 0.12 0.04 0.05 0.11 0.18 Pakistan 0.04 0.09 0.03 0.10 0.49 0.44 Sri Lanka 0.03 0.14 0.01 0.17 0.20 0.69 Sub-Saharan Africa 0.04 0.23 0.02 0.06 0.11 0.25 Benin 0.00 0.05 0.09 0.02 0.79 0.74 Burkina Faso 0.02 0.00 0.05 0.02 0.78 0.43 Burundi 0.27 0.35 0.17 0.26 0.12 0.37 Cameroon 0.01 0.24 0.06 0.05 0.72 0.70 Côte d'Ivoire 0.02 0.14 0.61 0.06 0.43 0.89 Ethiopia 0.13 0.27 0.20 0.07 0.24 0.22 Gambia, The 0.12 0.05 0.00 0.11 0.61 0.40 Ghana 0.08 0.10 0.09 0.08 0.62 0.62 Guinea 0.04 0.10 0.04 0.03 0.62 0.42 Kenya 0.30 0.41 0.54 0.38 0.52 0.94 Madagascar 0.95 0.24 0.06 0.98 0.56 0.36 Mali 0.01 0.01 0.03 0.03 0.93 0.45 Mauritania 0.03 0.09 0.18 0.07 0.60 0.60 Nigeria 0.10 1.04 0.30 0.27 0.75 0.21 Senegal 0.28 0.12 0.24 0.01 0.75 0.31 South Africa 0.32 0.68 0.68 0.21 0.39 0.50 Tanzania 0.03 0.01 0.03 0.02 0.54 0.26 Uganda 0.30 0.05 0.01 0.45 0.12 0.03 Latin America 0.51 0.61 1.06 0.65 0.77 1.26 Bolivia 0.55 0.64 0.65 0.71 0.74 0.61 Brazil 0.58 0.82 1.16 0.70 0.92 1.31 Chile 0.03 0.02 0.02 0.09 0.02 0.10 Colombia 0.71 0.45 1.30 0.85 0.49 1.40 Costa Rica 0.56 0.13 0.50 0.72 0.22 0.62 (Table continues on the following page.) 112 Agricultural Price Distortions, Inequality, and Poverty Table 3.4. The Effects of the Full Global Liberalization of Agricultural and All Merchandise Trade on the Incidence of Inequality and Poverty, Developing Countries (continued) (percentage point change) Agricultural reform only All merchandise trade reform US$1 US$2 US$1 US$2 Gini a day, a day, Gini a day, a day, Region, country coefficient headcount headcount coefficient headcount headcount Dominican Republic 0.76 0.24 0.75 0.95 0.24 0.89 Ecuador 0.90 1.75 2.78 1.29 1.98 3.09 El Salvador 1.09 1.25 0.79 1.29 1.34 0.88 Guatemala 0.87 1.26 1.54 1.11 1.59 1.44 Guyana 1.49 0.27 0.39 1.66 0.46 0.54 Honduras 1.06 0.57 1.58 1.30 0.72 1.63 Haiti 0.80 0.61 0.55 1.04 0.65 0.53 Jamaica 0.75 0.10 1.44 0.91 0.05 1.49 Mexico 0.20 0.36 0.35 0.05 0.01 0.07 Nicaragua 2.42 3.03 1.42 2.38 3.07 1.42 Panama 0.40 0.44 0.74 0.53 0.51 0.85 Paraguay 0.77 1.30 1.41 0.99 1.38 1.92 Peru 1.21 2.92 6.41 1.51 3.10 6.77 Venezuela, R.B. de 0.29 0.44 0.71 0.42 0.55 0.82 Europe and Central Asia 0.07 0.02 0.26 0.22 0.00 0.12 Armenia 0.42 0.34 0.14 0.45 0.23 0.13 Azerbaijan 0.23 0.10 1.36 0.12 0.14 1.21 Georgia 0.48 0.49 0.57 0.37 0.45 0.61 Kazakhstan 0.97 0.39 2.07 1.00 0.43 2.22 Kyrgyz Republic 0.50 0.08 0.57 0.27 0.19 1.09 Moldova 0.20 0.22 0.75 0.06 0.09 0.24 Russian Federation 0.16 0.00 0.32 0.06 0.00 0.31 Tajikistan 0.12 0.77 0.75 0.32 0.67 0.74 Ukraine 0.11 0.03 0.33 0.02 0.00 0.04 Uzbekistan 0.07 0.00 0.06 0.29 0.00 0.03 World 0.10 0.23 0.88 0.00 0.39 0.82 Source: Author compilation based on the GIDD Database and using Linkage model results similar to those in chapter 2, by Anderson, Valenzuela, and van der Mensbrugghe, except that factor mobility is more limited. The contributions to the global entries into and exits out of poverty depicted in figure 3.10 are, to a certain extent, the outcomes of differences in population size. For instance, a populous country such as India may have a substantial contri- bution to global poverty without necessarily implying a large increase in the head- count ratio of the country. Another way of ranking countries in terms of poverty Global Poverty and Distributional Impacts: The GIDD Model 113 outcomes is to consider the postreform change in the headcount ratio following all merchandise trade liberalization globally. Undertaking this exercise, we show that, among countries in which poverty falls, Ecuador and Peru would enjoy the largest declines in Latin America, while, in Asia, the Philippines, Thailand, and Vietnam (and Kazakhstan) would be in this category. India is still the country showing the largest increase in the incidence of extreme poverty, but the incidence of moderate poverty (US$2 a day) falls by 0.3 percentage points (table 3.4). Note that these changes in India occur although average household income remains virtually constant; the changes are therefore entirely a result of a deterioration in income distribution. This is reflected in the increase by 1.0 percentage points in the Gini coefficient of India. Three-quarters of this increase in inequality is attrib- utable to a rise in the agricultural-nonagricultural income gap in India. In contrast, the poverty reduction in Brazil is the outcome of a combination of a 1 percent increase in average incomes and a reduction in the Gini coefficient of inequality of around 0.6­0.7 percentage points (table 3.4). The changes in overall growth and distribution that would take place in Brazil and India after global agricultural reform are summarized through the GICs for these two countries plotted in figure 3.9, chart b. Given the importance of Brazil and India in their respective regions, it is not surprising that the shapes of the GICs for these countries are similar to the GICs of their respective regions plotted in figure 3.9, chart a. Chart b in Figure 3.9 shows that the only beneficiaries of global agricultural liberalization in India are people in households in the top 22 percent of the distri- bution; given that 83 percent of the Indian population is living below the US$2-a- day poverty line, part of the top 22 percent is accounted for by households living in moderate poverty. Agricultural reforms may have important real income distributional effects in terms of the gap between agricultural and nonagricultural households (see elsewhere above). Our results show that, for many countries in our sample, the removal of distortions would have considerable distributional effects. In one- third of the countries listed in table 3.4, the Gini coefficient shows a decline of more than one-half of a percentage point. This pattern is also observed in the changes in the country-specific Theil index plotted in figure 3.11. There are distinguishable regional differences in the distributional effects of the reform; countries in East Asia and Latin America experience a significant reduction in income inequality, while the inequality in countries outside these regions remains constant or rises marginally. The advantage of using the Theil index as the inequality measure is that we are able to decompose the changes in the index into an effect attributable to shifts in the agricultural-to-nonagricultural wage gap (the between effect) and the effects arising from income changes 114 Agricultural Price Distortions, Inequality, and Poverty Figure 3.11. The Theil Index of Overall and Between-Group Income Distributional Changes after Full Global Agricultural Policy Reform Yemen, Rep. Haiti Vietnam change in the overall Honduras Theil index Guatemala Peru change in the Philippines between-group Paraguay (agri/non-agri) Thailand component of Ecuador the Theil index Jordan Cambodia Brazil Colombia Bolivia Guyana Jamaica Dominican Republic El Salvador Costa Rica Kazakhstan Panama Azerbaijan Venezuela, R.B. de China Kyrgyz Republic Georgia Russian Federation Moldova Ukraine Albania Tajikistan Bulgaria country Indonesia Chile Senegal Tanzania Uzbekistan Mongolia Mauritania Côte d'Ivoire Benin Guinea Pakistan Mali Romania Ghana Nicaragua Gambia, The Cameroon Mexico Burkina Faso Nepal Sri Lanka Estonia Armenia Nigeria Uganda Hungary Turkey Burundi South Africa Poland Bangladesh India Kenya Ethiopia Lithuania 0.035 0.025 0.015 0.005 0 0.005 0.015 0.025 Theil index Source: Author compilation based on the GIDD Database. Note: The Theil index is an inequality measure of the family of general entropy that has the property of yielding perfectly additive inequality decompositions by population subgroups. The figure is based on the data set for developing countries only. Global Poverty and Distributional Impacts: The GIDD Model 115 within these two groups. Figure 3.11 shows the total changes in the Theil index (depicted by an x) and the changes attributable to movements in the nonagri- cultural income premium (the small horizontal bar). Since the between effect is close to the total distributional effect for the majority of countries, we may conclude that the total change in income distribution in these economies is mainly the outcome of changes in the mean incomes in the agricultural and nonagricultural sectors. Conclusions Trade distortions in agriculture and food represent the last major bastion of merchandise protection and continue to be the main point of contention in multilateral and preferential trade negotiations. Using a newly developed data set and methodological approach for evaluating the poverty and inequality effects of policy reforms, the GIDD, this chapter evaluates the potential effects on global income distribution and poverty exerted by the removal of agricul- tural distortions and other trade distortions. There are three main messages emerging from our analysis. First, the liberaliza- tion of agricultural and food markets is unlikely to have large effects on global poverty. Our results show that the incidence of extreme poverty (US$1 a day, PPP) might rise by 0.5 percent each from farm and nonfarm full global trade reform, while moderate poverty (US$2 a day, PPP) is likely to fall by 0.9 percent from agri- cultural reform alone. The second message is that these small aggregate changes are produced by a combination of offsetting trends at the regional and national levels. Farmers in Latin America, the region that accounts for less than 5 percent of global poverty, experience significant income gains, while 15 million more people in South Asia, where half of the world's poor reside, would fall below the extreme poverty line after the liberalization of world agricultural markets, and another 2.8 million would fall below this line if the barriers to the trade in nonfarm goods were also removed. Some countries or regions would experience considerable distributional changes following global trade reform. Inequality is projected to fall in regions, such as Latin America, that are characterized by high initial inequality, but is projected to rise in South Asia, which is characterized by low initial inequality. As a result, the projected changes for developing countries as a group are small overall. There are several important caveats to our analysis. First, it should be empha- sized that, although poverty reduction is a most worthy goal, it should not be the only or even the first metric with which to measure trade policy. Trade reform should not be expected to benefit all constituents and may do so only if it is accompanied by other, complementary domestic policies. Second, our analysis is 116 Agricultural Price Distortions, Inequality, and Poverty confined to an examination of the effects of static efficiency gains and does not consider the potential growth effects of trade liberalization. Although our results show that the static gains from agricultural reform and other trade reform may not contribute to a reduction in extreme poverty and may do little to combat moderate poverty, they do not imply that this pattern of trade liberalization is not an effective tool for poverty reduction. Finally, our microsimulation model con- siders only changes in labor income; while this is the most important income source for households at or near the poverty line, accounting for changes in other factor returns may yield somewhat different inequality results. Notes 1. According to Anderson and Martin (2005), self-interested vocal groups lobbying aggressively in favor of excluding agricultural liberalization from multilateral negotiations encompass not only farm- ers in the highly protected countries and net food-importing developing countries, but also food exporters receiving preferential access to these markets, including holders of tariff rate quotas, mem- bers of regional trading agreements, and parties to nonreciprocal preference agreements, such as all least developed countries. 2. The GIDD data set, methodology, and applications are available at http://go.worldbank.org/ YADEAFEJ30. 3. This more highly aggregated information usually consists of 20 data points for each country, each data point representing the average per capita income (or consumption) of 5 percent of the coun- try's population. In the absence of full survey data, the use of these vingtile data provides a close approximation to most economy-wide measures of inequality. 4. The adjustment procedure for expressing welfare indicators in 1993 international dollars (PPP) is as follows. First, for countries for which the survey year is not 2000, the welfare indicator (household per capita income or consumption) is scaled to the year 2000 using the cumulative growth in real income per capita between the survey year and 2000. Then, the welfare indicator is converted to 1993 national prices by multiplying the welfare indicator by the ratio of the consumer price index in 1993 to the consumer price index in the survey year. Finally, the welfare indicator is converted to 1993 interna- tional prices by multiplying the outcome of the above calculations by the 1993 PPP exchange rate. 5. The distributions for the agricultural and nonagricultural populations are not, strictly speaking, density functions because the area below the curve does not add to 1. The densities of the agricultural and nonagricultural populations were rescaled so that the area under the curve represents the share of the world's population in these two groups. 6. For a specific analysis of the importance of China and India to global growth and income distri- bution, see Bussolo et al. (2007). 7. A household is defined as a net producer (consumer) of agricultural products if the monetary income it derives from merchandising these products is greater (smaller) than the amount spent on them. 8. For empirical applications of the effect of trade on household welfare that take into account these effects, see, for example, the case studies on Mexico by Nicita (2004) and De Hoyos (2007). 9. A detailed methodological description of the GIDD may be found in Bussolo, De Hoyos, and Medvedev (2008). 10. The GIDD uses the Linkage model as a global CGE framework. See van der Mensbrugghe (2005) for a detailed description of the Linkage model. 11. Throughout this chapter, global distribution means the GIDD sample of countries, which include 92 percent of the world's population. Global Poverty and Distributional Impacts: The GIDD Model 117 12. The GIDD does not take into account the welfare impacts via any changes in remittances or transfers between households that result from trade reform. 13. The Linkage results used here are not identical to those in chapter 2 because labor mobility in our chapter has been restricted to match the available microdata more closely. In the version of the Linkage model used in chapter 2, the assumption of full factor mobility leads to an equalization of fac- tor prices across sectors. However, the household survey data show large and persistent differences between the labor earnings in agriculture and the labor earnings in nonagriculture after we have con- trolled for other relevant characteristics. Imposing an equalization of wages in the GIDD would lead to large and implausible changes in the distribution of income; to maintain consistency between macro- and micromodels, we have therefore limited labor mobility in the macromodel. 14. The price increases are calculated using the Paasche price index, that is, using postreform exports as weights to aggregate the prices of individual commodities. Unless explicitly noted, all price indexes in this section are calculated using the Paasche formula. Price indexes differ by country because of differences in the composition of exports (that is, aggregation weights). 15. This closure choice thus gives rise to a consistent measurement of household utility because the utility function does not include the consumption of public goods. 16. In this situation, whether households gain or lose from trade reform, in addition to the impacts of the first two channels, depends on their ability to substitute out of more expensive goods into now- cheaper alternatives. 17. Note that trends in consumption per capita are unlikely to be representative of the welfare of agricultural households given that, because of limited incomes and high poverty incidence, the weight of these households in total consumption is low. 18. The 6.1 percent figure refers to the decline in nominal wages. The change in real wages depends on the choice of deflator: while the consumer price index increases by 2 percent relative to the base year, the GDP deflator falls by 1 percent. 19. This is a nominal increase, not a real increase. Consumer prices in Brazil rise by 4 percent after trade reform. 20. A GIC shows the changes in welfare along the entire income distribution. It therefore captures, in a single graph, the growth and distributional components of overall welfare changes. For a detailed description of the properties of GICs, see Ravallion and Chen (2003). References Anderson, K., ed. 2009. Distortions to Agricultural Incentives: A Global Perspective, 1955­2007. London: Palgrave Macmillan; Washington, DC: World Bank. Anderson, K., and W. Martin. 2005. "Agricultural Trade Reform and the Doha Development Agenda." World Economy 28 (9): 1301­27. Anderson, K., and E. Valenzuela. 2008. "Estimates of Global Distortions to Agricultural Incentives, 1955­2007." Data spreadsheet, October, World Bank, Washington, DC. http://go.worldbank.org/ YAO39F35E0. Bourguignon, F., V. Levin, and D. Rosenblatt. 2004. "Declining International Inequality and Economic Divergence: Reviewing the Evidence through Different Lenses." Economie Internationale 2004 (4Q): 13­26. Bourguignon, F., and L. A. Pereira da Silva, eds. 2003. The Impact of Economic Policies on Poverty and Income Distribution: Evaluation Techniques and Tools. London: Oxford University Press; Washington, DC: World Bank. Bussolo, M., R. De Hoyos, and D. Medvedev. 2008. "Economic Growth and Income Distribution: Linking Macroeconomic Models with Household Survey Data at the Global Level." Paper presented at the International Association for Research in Income and Wealth 30th general confer- ence, Portoroz, Slovenia, August 24­30. 118 Agricultural Price Distortions, Inequality, and Poverty Bussolo, M., R. De Hoyos, D. Medvedev, and D. van der Mensbrugghe. 2007. "Global Growth and Dis- tribution: Are China and India Reshaping the World?" Policy Research Working Paper 4392, World Bank, Washington, DC. Bussolo, M., J. Lay, and D. van der Mensbrugghe. 2006. "Structural Change and Poverty Reduction in Brazil: The Impact of the Doha Round." In Poverty and the WTO: Impacts of the Doha Development Agenda, ed. T. W. Hertel and L. A. Winters, 249­84. London: Palgrave Macmillan; Washington, DC: World Bank. Chen, S., and M. Ravallion. 2003. "Household Welfare Impacts of China's Accession to the World Trade Organization." World Bank Economic Review 18 (1): 29­57. ------. 2007. "Absolute Poverty Measures for the Developing World, 1981­2004." Policy Research Working Paper 4211, World Bank, Washington, DC. De Hoyos, R. E. 2007. "North-South Trade Agreements and Household Welfare: Mexico under NAFTA." mimeo, World Bank, Washington, DC. Ferreira, F. H. G., and P. G. Leite. 2003. "Meeting the Millennium Development Goals in Brazil: Can Microsimulations Help?" Economía 3 (2): 235­79. ------. 2004. "Educational Expansion and Income Distribution: A Microsimulation for Ceará." In Growth, Inequality, and Poverty: Prospects for Pro-Poor Economic Development, ed. A. Shorrocks and R. van der Hoeven, 222­50. London: Oxford University Press. GIDD Database (Global Income Distribution Dynamics Database). World Bank. http://go.worldbank.org/ YADEAFEJ30. Krueger, A. O., M. Schiff, and A. Valdés. 1988. "Agricultural Incentives in Developing Countries: Measuring the Effect of Sectoral and Economy-wide Policies." World Bank Economic Review 2 (3): 255­72. McCulloch, N., L. A. Winters, and X. Cirera. 2001. Trade Liberalization and Poverty: A Handbook. London: Centre for Economic Policy Research. Nicita, A. 2004. "Who Benefited from Trade Liberalization in Mexico? Measuring the Effects on House- hold Welfare." Policy Research Working Paper 3265, World Bank, Washington, DC. Ravallion, M., and S. Chen. 2003. "Measuring Pro-poor Growth." Economics Letters 78 (1): 93­99. Rodrik, D. 1998. "Why Is Trade Reform So Difficult in Africa?" Journal of African Economies 7 (0): 43­69. Valenzuela, E., and K. Anderson. 2008. "Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980­84." Research Memorandum 13 (December), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925. van der Mensbrugghe, D. 2005. "Linkage Technical Reference Document: Version 6.0." December, World Bank, Washington, DC. http://go.worldbank.org/7NP2KK1OH0. Winters, L. A. 2002. "Trade Liberalization and Poverty: What Are the Links?" World Economy 25 (9): 1339­67. 4 Poverty Impacts in 15 Countries: The GTAP Model Thomas W. Hertel and Roman Keeney Despite slow progress toward a multilateral trade reform agreement, the Doha Development Agenda negotiations of the World Trade Organization continue to generate interest because of their poverty reduction potential. A distinguishing feature of the Doha agenda has been the lack of commitment to trade policy reform by developing countries, particularly the poorest countries, which would not be required to commit to reduce tariffs following a Doha agreement (Anderson and Martin 2006). Recent research suggests that developing-country tariff cuts, particularly in agriculture, are among the most poverty-friendly elements of a broader multilat- eral trade policy agenda (Hertel et al. 2009). Such analyses hinge critically on the measured protection for agriculture in developing countries. Unlike the countries of the Organisation for Economic Co-operation and Development (OECD), in which the measurement of agricultural protection has received considerable attention over the past two decades through the regular publication of producer support estimates, there was considerable uncertainty about the current support and recent trends in agricultural protection in developing countries, particularly the smaller low-income economies, until the new World Bank agricultural distor- tions database was compiled (Anderson and Valenzuela 2008). In this chapter, we incorporate the new information on price distortions in an assessment of the impact of agricultural and other trade reforms on poverty. The new data on developing-country protection in agriculture afford us the opportunity to consider the relative importance of agricultural versus nonagricultural protection at home and abroad more accurately than has been previously possible.1 119 120 Agricultural Price Distortions, Inequality, and Poverty Tracing the impacts of developments in multilateral trade policy from interna- tional markets to the household level is a long and complex process (Winters, McCulloch, and McKay 2004). A natural method for accomplishing this is to divide the task into parts. This approach is taken in Hertel and Winters (2006), who use a global model to generate the changes in world prices following various multilateral trade reform scenarios and then national models rich in household- level detail to assess the ensuing national poverty impacts. This method allows the authors to benefit from the comparative advantage of separate modeling tools. It has become widely accepted as the standard method for reaching definitive answers to ex ante questions on the impacts of trade reforms on poverty in individual countries. An inherent limitation of the country-specific studies underlying the final estimates on the changes in poverty derived through this approach is the short- age of general conclusions on the broad pattern of impacts on poverty. Are rich- country reforms pro-poor on average? How do they compare with poor-country reforms? For this purpose, it is important to assemble a framework that is based on household surveys. The framework should allow us to make comparisons across countries and develop conclusions about poverty impacts in a wide range of diverse economies that are representative of the developing world. This approach has been adopted by Ivanic (2006) and Hertel et al. (2007) in their analyses of the Doha Development Agenda. These authors find that the set of measures envi- sioned under the Doha Development Agenda are less poverty-friendly than alter- native policy measures not being considered.2 Hertel et al. (2009) explore this question in detail and determine that an overemphasis on export and domestic production subsidies relative to the attention paid to market access in the rich countries, together with the absence of commitments to deep agricultural tariff cuts in developing countries, is the primary culprit in diminishing the prospects for poverty reduction through the Doha Development Agenda. The international cross-section approach--the framework approach mentioned above--has proven especially fruitful in providing insights that are complementary to those provided by more-detailed country-specific studies, despite the former's relatively undiffer- entiated treatment of developing-country household behavior. In addition to making use of the new developing-country protection estimates, we also examine the poverty impacts of trade reform through a different lens, namely, the lens of commodity-specific support. Given that many political economy and distributional causes of agricultural protection are inherently commodity specific, viewing the trade-poverty link through the commodity lens in a comparative fashion represents an area of relative inattention and one that a Poverty Impacts in 15 Countries: The GTAP Model 121 global model with a diverse developing-country sample is well suited to address. For our present purposes, we focus on 15 developing countries: four in Africa (Malawi, Mozambique, Uganda, and Zambia), five in Asia (Bangladesh, Indonesia, the Philippines, Thailand, and Vietnam), and six in Latin America (Brazil, Chile, Colombia, Mexico, Peru, and República Bolivariana de Venezuela). Our Analytical Approach to Poverty Modeling Our poverty analysis begins with the specification of a utility function and an associated consumer demand system for determining household consumption and the maximum utility attainable at a given level of prices and incomes. The utility of the household at the poverty line is defined as the poverty level of utility. Households with utility at or below this level are deemed to be in poverty. In this study, we follow Hertel et al. (2004, 2007) in using the AIDADS demand system of Rimmer and Powell (1992) to represent consumer preferences. AIDADS (an implicitly directly additive demand system) is particularly effective in poverty analysis because it lends itself to international cross-section estimation and devotes two-thirds of its parameters to consumption behavior in the neighbor- hood of the poverty line (Cranfield et al. 2003). The estimation of this demand system is undertaken using the 80-country per capita consumption data set offered in version 6.1 of the Global Trade Analysis Project (GTAP) data set; the resulting parameters are reported in Hertel et al. (2009). The demand system estimates are then calibrated to reproduce base year per capita demands in each country following the approach of Golub and Hertel (2008). A key finding in the work of Hertel et al. (2004) is the importance of stratifying households by primary source of income. Farm households in developing coun- tries often rely on the farm enterprise for virtually all income, and the share of national poverty concentrated in agriculture-specialized households is large in the poorest countries in our sample: between one-quarter and one-half of the US$1-a-day poverty line headcount in Chile, Colombia, Indonesia, Malawi, Mozambique, and Zambia. Not only are farm households in the poorest countries more likely to be spe- cialized in farm earnings, these specialized farm households also tend to be poorer than the rest of the population (Hertel et al. 2004). The implication is that the poorest households in the poorest countries are more concentrated in agriculture and therefore more likely to benefit from producer price increases engendered by multilateral trade reforms. We follow Hertel et al. (2004) in identifying five house- hold groups that rely almost exclusively (greater than 95 percent) on one source of income: agricultural self-employment, nonagricultural self-employment, rural 122 Agricultural Price Distortions, Inequality, and Poverty wage labor, urban wage labor, or transfer payments. The remaining households are grouped into rural or urban diversified strata, leading to seven total strata.3 Hertel et al. (2004) have simulated the impact of trade reform on the full dis- tribution of households within each of the seven strata using a global computable general equilibrium (CGE) model and a household microsimulation framework. Given our emphasis on poverty in this chapter, we follow Hertel et al. (2009) by focusing on the households in the neighborhood of the poverty line and making use of a highly disaggregated poverty elasticity approach. Specifically, we compute the stratum-specific elasticity of the poverty headcount with respect to a change in average income in the stratum. We denote this elasticity by ers and report the com- puted values for the 15 countries in our sample in table 4.1. The values range from a low of 0.0006 in the self-employed agriculture stratum in Zambia, where nearly Table 4.1. Estimates of Elasticities of the US$1-a-Day Poverty Headcount with Respect to Total Income, 15 Countries Strata Urban Rural Urban Rural National Country Ag Nonag labor labor Transfer diverse diverse elasticity Bangladesh 1.64 2.02 1.58 0.63 0.56 1.74 1.09 1.24 Brazil 0.75 1.28 1.94 2.19 0.34 3.63 2.69 1.35 Chile 1.90 2.24 2.06 1.55 2.45 2.29 2.60 2.18 Colombia 0.79 0.60 1.73 1.72 0.93 1.14 1.00 0.82 Indonesia 2.35 2.14 2.38 2.89 1.17 2.58 2.87 2.47 Malawi 0.49 0.30 2.26 1.97 0.43 1.04 0.76 0.58 Mexico 1.73 1.90 3.33 2.08 2.28 1.63 1.80 2.02 Mozambique 0.28 0.94 0.97 0.76 0.48 1.58 0.99 0.64 Peru 1.50 1.32 2.37 1.73 0.44 1.09 1.05 1.07 Philippines 2.25 1.96 2.98 2.44 1.69 2.42 1.98 2.15 Thailand 2.30 2.42 2.98 2.45 2.78 2.42 2.59 2.57 Uganda 0.28 0.40 1.71 0.34 0.01 0.36 0.21 0.24 Venezuela, R.B. de 0.69 1.16 2.57 2.17 0.01 1.72 1.53 1.20 Vietnam 0.48 1.12 2.81 8.98 0.84 0.86 1.01 0.98 Zambia 0.00 0.64 2.28 0.91 0.45 1.29 0.37 0.61 Source: Author compilation. Note: The values in the strata columns are the elasticities of the poverty headcount with respect to changes in earnings. National elasticity in the final column is the aggregate elasticity for each country weighted by the poverty share (see table 4.2). The authors have estimated the elasticities using country- specific household survey data. For the first five strata, more than 95 percent of household income is derived from only one income source. Ag agriculture. Nonag nonagriculture. Poverty Impacts in 15 Countries: The GTAP Model 123 all of the population is living well below the poverty line, to a high of 3.63 in the urban diversified stratum of Brazil, where the population density at the poverty line is quite high. The proportional change in the real incomes of households at the poverty line in stratum s of region r may be written as the sum of the real after-tax factor earn- ings of these households weighted by the income share, as follows: Np N Np y rs = a a p (W rj - C r ), rsj (4.1) j where a p is the share of income obtained from factor j by households at the rsj N poverty line in stratum s of region r; W rj is the proportional change in after-tax earnings of factor j in region r; and C N r is the proportional change in the cost of p living at the poverty line in region r, that is, the expenditure required to remain at the poverty level of utility. This last is obtained by solving the AIDADS demand system for the expenditure required to remain at the poverty level of utility given the new prices (that is, the post-trade reform prices). We may now express the proportional change in the poverty headcount in stratum s of region r as follows: N N N Hrs = ers # y p = ers # a # a p (W rj - C rp) N rs rsj (4.2) j p The earnings shares at the poverty line, a rsj, will play a critical role in our analysis. Regardless of the household type (with the exception of transfer-dependent households), the income from unskilled labor tends to dominate the stratum- specific earnings shares (Hertel et al. 2009). For example, in the case of agriculture- dependent households, most earnings show up in unskilled agricultural labor. In the case of rural diversified households, earnings usually show a mix of sources, including agricultural self-employed unskilled labor, unskilled wage labor, and unskilled nonfarm (self-employed) labor. The fact that these households are so poor means that they have little income from land or capital. Having established the determinants of the stratum poverty headcount, we may now progress to the national poverty headcount, Hr, which may be expressed as a function of the stratum headcounts and stratum populations (POPrs): Hr = c a POPrs * Hrs d n POPr , (4.3) s where POPr = © POPrs. So, the proportional change in the national poverty s N headcount is Hr = © b rs * Hrs . s 124 Agricultural Price Distortions, Inequality, and Poverty Here, b rs = C (POPrs * Hrs)>POPr D >Hr = (POPrs * Hrs)> a (POPrs * Hrs) (4.4) s is the share of stratum s poverty in nationwide poverty within region r. These shares are reported in table 4.2 for our 15 focus countries. Agriculture-specialized households and rural diversified households tend to dominate in the poverty headcount although exceptions are evident in Colombia, Peru, and República Bolivariana de Venezuela, where self-employed nonagricultural households represent a large share of the poor. Combining equations (4.2) and (4.4), we obtain a useful expression for evalu- ating the change in the national poverty headcount in response to a small change in factor and commodity prices: N N p Hr = a b rs # ers # a a p (W rj - C r ) rsj (4.5) s j Table 4.2. Stratum Contributions to the US$1-a-Day Poverty Population, 15 Countries (percentage shares) Strata Urban Rural Urban Rural Country Ag Nonag labor labor Transfer diverse diverse Total Bangladesh 0.15 0.13 0.04 0.22 0.03 0.07 0.37 1.00 Brazil 0.14 0.09 0.24 0.15 0.32 0.04 0.03 1.00 Chile 0.26 0.01 0.09 0.09 0.28 0.15 0.12 1.00 Colombia 0.28 0.43 0.03 0.04 0.12 0.05 0.04 1.00 Indonesia 0.42 0.12 0.02 0.07 0.04 0.06 0.28 1.00 Malawi 0.54 0.11 0.00 0.03 0.07 0.01 0.25 1.00 Mexico 0.05 0.06 0.05 0.12 0.28 0.14 0.29 1.00 Mozambique 0.41 0.13 0.01 0.05 0.14 0.06 0.19 1.00 Peru 0.07 0.35 0.01 0.02 0.22 0.11 0.23 1.00 Philippines 0.12 0.06 0.03 0.05 0.03 0.23 0.49 1.00 Thailand 0.06 0.02 0.00 0.06 0.11 0.07 0.68 1.00 Uganda 0.10 0.04 0.00 0.03 0.02 0.07 0.75 1.00 Venezuela, R.B. de 0.08 0.24 0.17 0.10 0.28 0.08 0.05 1.00 Vietnam 0.04 0.11 0.00 0.00 0.05 0.10 0.70 1.00 Zambia 0.34 0.23 0.10 0.07 0.07 0.09 0.11 1.00 Source: Author compilation. Note: The values are the shares of the impoverished population that are specialized in a particular stratum of earnings. The shares are derived from country-specific household surveys. The total column reflects the assumption that the entire population in poverty is allocated among the seven strata. Ag agriculture. Nonag nonagriculture. Poverty Impacts in 15 Countries: The GTAP Model 125 Because the expression in parentheses in equation (4.5) denotes the proportional change in the real after-tax income associated with each of the factors of produc- N rj tion (W R), we may rewrite this as follows: N N rj Hr = a a b rs # ers # a p # W R rsj (4.6) s j From equation (4.6), it is clear that, to obtain the change in national poverty, each real after-tax factor return must be premultiplied by a region-, stratum-, and factor-specific poverty elasticity. For example, table 4.3 reports the product b rs # ers # ap for Bangladesh. The rsj rows in the table correspond to strata, and the columns to earnings types. Because © j arsj = 1, the sums in the rows of the table simply give the elasticity of national poverty with respect to a 1 percent rise in stratum income, that is, b rs # ers. The sums are obviously heavily influenced by the national poverty shares reported in the Bangladesh row of table 4.2. The rural diversified stratum comprises nearly 37 percent of the poor; so it is not surprising that this row total is the largest in table 4.3. This stratum is followed in importance by the self-employed agricultural and nonagricultural strata, which show relatively high stratum-specific poverty elasticities (recall the first row of table 4.1), and then by the rural labor stratum and the urban diversified stratum. The column sums in table 4.3 allow us to identify the factors that are most important to poverty reduction in Bangladesh. In this case, it is clear that unskilled labor is the primary endowment of the poor in Bangladesh. The key issues are the strata of these workers and the effect of trade reform on their rela- tive wages. From the bottom row in table 4.3, we learn that unskilled wage labor is the most important stratum from the perspective of national poverty reduction, followed closely by self-employed nonagricultural labor and unskilled agricultural labor. Transfer payments, skilled wage labor, capital, and land play a much smaller role in poverty reduction at the margin, according to the final row in the table. The grand total in table 4.3 gives the national US$1-a-day poverty elasticity (bottom right corner, 1.24) with respect to a uniform 1 percent rise in real after- tax income from all sources. The first row of table 4.4 takes the final row of table 4.3 for Bangladesh and divides all the entries by 1.24 so that we may see the share of each earnings source in national poverty reduction in the context of an across-the-board rise in real after-tax incomes. Thus, unskilled wage labor (the fifth column entry in row 1) contributes 33 percent of the total and so on. The remaining rows in table 4.4 pro- vide the same calculation for the other 14 developing countries in our sample. If we treat each country as an observation, we find that unskilled wage labor shows the highest average share (table 4.4, bottom row), followed by agricultural unskilled labor and then nonagricultural unskilled labor and transfers. Transfer 126 Table 4.3. Stratum- and Earnings-Specific Poverty Elasticities, Bangladesh Ag Ag Unskilled Skilled Nonag Nonag unskilled skilled Ag wage wage unskilled skilled Nonag Strata Land labor labor capital labor labor labor labor capital Transfers Total Ag 0.01 0.23 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.24 Nonag 0.00 0.00 0.00 0.00 0.00 0.00 0.25 0.00 0.01 0.00 0.26 Urban labor 0.00 0.00 0.00 0.00 0.06 0.00 0.00 0.00 0.00 0.00 0.06 Rural labor 0.00 0.00 0.00 0.00 0.12 0.01 0.00 0.00 0.00 0.00 0.14 Transfer 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.02 Urban diverse 0.00 0.02 0.00 0.00 0.05 0.01 0.02 0.00 0.00 0.01 0.12 Rural diverse 0.00 0.07 0.00 0.00 0.17 0.02 0.08 0.00 0.01 0.04 0.40 Total 0.01 0.33 0.00 0.01 0.41 0.03 0.36 0.00 0.02 0.07 1.24 Source: Author compilation. Note: The table refers to households living on less than US$1 a day. The elasticities are calculated by multiplying earnings shares by stratum-specific elasticities for Bangladesh. The total column shows the change in stratum poverty resulting from a 1 percent increase in income for each household type. The total row shows the national change in poverty resulting from a 1 percent increase in the factor income of each column type. Ag agriculture. Nonag nonagriculture. Table 4.4. The Contributions of Earnings to the Total US$1-a-Day Poverty Response, 15 Countries (percent share of total) Ag Ag Unskilled Skilled Nonag Nonag unskilled skilled Ag wage wage unskilled skilled Nonag Country Land labor labor capital labor labor labor labor capital Transfers Total Bangladesh 0.01 0.27 0.00 0.01 0.33 0.02 0.29 0.00 0.02 0.06 1.00 (1.24) Brazil 0.00 0.07 0.03 0.01 0.60 0.04 0.10 0.01 0.00 0.15 1.00 (1.35) Chile 0.07 0.15 0.00 0.10 0.26 0.00 0.02 0.00 0.00 0.40 1.00 (2.18) Colombia 0.00 0.29 0.00 0.00 0.18 0.00 0.34 0.00 0.01 0.16 1.00 (0.82) Indonesia 0.03 0.50 0.00 0.02 0.19 0.01 0.17 0.00 0.04 0.03 1.00 (2.47) Malawi 0.03 0.52 0.00 0.07 0.12 0.02 0.07 0.00 0.05 0.14 1.00 (0.58) Mexico 0.00 0.09 0.00 0.00 0.38 0.00 0.08 0.00 0.00 0.43 1.00 (2.02) Mozambique 0.00 0.35 0.00 0.02 0.11 0.00 0.14 0.00 0.15 0.20 1.00 (0.64) Peru 0.01 0.16 0.00 0.01 0.09 0.00 0.45 0.07 0.06 0.14 1.00 (1.07) Philippines 0.23 0.00 0.01 0.12 0.31 0.01 0.14 0.00 0.07 0.10 1.00 (2.15) Thailand 0.03 0.20 0.02 0.02 0.24 0.05 0.04 0.01 0.01 0.38 1.00 (2.57) Uganda 0.13 0.13 0.00 0.25 0.08 0.08 0.08 0.00 0.13 0.08 1.00 (0.24) Venezuela, R.B. de 0.00 0.06 0.00 0.00 0.58 0.03 0.28 0.01 0.00 0.05 1.00 (1.20) Vietnam 0.01 0.12 0.00 0.00 0.00 0.00 0.18 0.00 0.47 0.20 1.00 (0.98) Zambia 0.00 0.02 0.00 0.02 0.43 0.16 0.20 0.00 0.13 0.08 1.00 (0.61) Average 0.04 0.19 0.00 0.04 0.26 0.03 0.17 0.01 0.08 0.17 n.a. Source: Author compilation. Note: The values are earnings-specific elasticities, divided by national poverty elasticities (in parentheses in the total column). The average row is the simple average across countries of the contribution of a specific earnings source to total poverty responsiveness. The total column reflects the fact that contributions to total 127 responsiveness add to 1. Ag agriculture. Nonag nonagriculture. n.a. not applicable. 128 Agricultural Price Distortions, Inequality, and Poverty payments are important in some of the richer developing countries (Brazil, Chile, Colombia, Mexico, Thailand), as well as in lower-income countries with a large share of migrant labor (for example, Mozambique). The entries in table 4.4 give a clear idea of the factor price increases that are most likely to lower the national poverty headcount in each country. The Global CGE Model Because the household-level poverty impacts hinge critically on factor rewards that depend not only on the type of endowment, but also on where it is employed, we need a global modeling framework with sufficient detail to separate out these differential returns. We have adopted the GTAP model (Hertel 1997) and the ver- sion 6.1 GTAP Database calibrated to 2001 (Dimaranan 2006). We have modified both to be consistent with our needs for differentiated factor returns, as well as the changes in price distortions identified in Anderson and Valenzuela (2008) and prepared for CGE modeling by Valenzuela and Anderson (2008). We must also make some updates to account for changes in market conditions and policy reforms since 2001, the year of the version 6.1 database. The model is imple- mented in Gempack software (Harrison and Pearson 1996). We now turn to a discussion of the modeling assumptions and data changes incorporated into our CGE modeling framework. Database adjustments Our starting point for the global CGE analysis of the impacts of trade policy is the GTAP version 6.1 database (Dimaranan 2006). For our present purposes, we have aggregated the database to 31 productive sectors and six groups of household con- sumption items, following Ivanic (2006). We have updated this database by solving an experiment that accounts for key policy reforms in border protection that were undertaken between 2001 and 2005, most notably the accession of China to the World Trade Organization, the enlargement of the European Union, and the com- pletion of the Uruguay Round commitments by some members of the World Trade Organization. We have also altered the database, now with a base year of 2005, to reflect the recent estimates of agricultural price distortions in developing countries by Anderson and Valenzuela (2008). We have likewise altered our database to reflect the most recent OECD commodity support data on agricultural output subsidies as of 2004. The aggregate changes in commodity support as measured by the producer support estimate are relatively small over this time frame (OECD 2007). However, in some instances, the support for particular commodities changed dramatically. Because of our desired emphasis on viewing poverty impacts through a commodity Poverty Impacts in 15 Countries: The GTAP Model 129 lens, our use of the most recent available information on initial protection is warranted. In total, this has meant that we have adjusted the output subsidy levels among 25 OECD members in the initial database.4 Modifications to the GTAP model Our modifications to the standard GTAP model focus on features that enhance the analysis of agricultural reforms and the simulation of impacts on poverty. We retain the simplistic, yet empirically robust assumptions of constant returns to scale and perfect competition typically featured in agricultural trade studies.5 The remaining modifications are aimed at permitting us to shed new light on the distributional consequences of trade reforms. They focus particularly on unraveling the puzzle involved in the question, why is such reform not more poverty-friendly? On the demand side of the model, we ensure consistency with the poverty analysis by modifying the GTAP model to incorporate the AIDADS demand system as discussed in Hertel et al. (2007). Thus, the aggregate preferences are consistent with the preferences used to evaluate the impact of price changes on households at the poverty line although expenditure patterns differ across income levels because of the nonhomotheticity of demands. The other modifications relate to the factor markets and follow from the model changes made by Keeney and Hertel (2005). Frictions in agricultural factor markets have featured prominently in the development economics literature, particularly as an explanation for low agricultural supply response (de Janvry, Fafchamps, and Sadoulet 1991). Modeling the complex processes leading to limited farm-nonfarm and rural-urban mobility for the full range of countries in our model is beyond the scope of this chapter and is better suited to the individual country case study approach (see elsewhere above). To maintain a general framework that reflects the imperfect factor mobility between rural and urban employment, we specify a constant elasticity of transfor- mation function that transforms farm-employed factors into nonfarm employ- ment and vice-versa.6 This allows factor rewards to diverge between the farm and nonfarm sectors and supplies us with the factor market segmentation we require for our distributional analysis. We use the factor supply parameters that have been adopted by Keeney and Hertel (2005) and that are drawn from the OECD (2001). We assume that aggregate endowment levels are fixed in our static analysis. This reflects a belief that the aggregate supply of factors is unaffected by trade policy. This is not the full employment assumption sometimes ridiculed by advocates of structural models of development. Rather, it holds that aggregate employment is primarily determined by factors, such as labor market norms and regulation, that are largely independent of trade policy in the long run. 130 Agricultural Price Distortions, Inequality, and Poverty Recalling equation (4.6), we must tie our results to the factor earnings shares for all household types in each region. We map the factor returns from the general equilibrium model to the various earnings types according to the following method. Agricultural labor and capital receive the corresponding farm factor returns from the general equilibrium model, as do nonagricultural labor and cap- ital. Wage labor in the household surveys is not distinguished by place of employ- ment; so, we use the economy-wide average change. Transfer payments represent an important source of earnings for many households and have no obvious corol- lary in the CGE model. We choose to index these to the growth rate in net national incomes, which allows us to maintain consistency with the representative house- hold approach of the global model (see Hertel et al. 2007). Finally, a few words about our macroclosure are in order. In this chapter, we fix the ratio of key macroeconomic aggregates relative to net national income. These include government spending, total tax revenues (net of subsidies), net national savings, and the trade balance. In this way, we also ensure that public transfer pay- ments (not explicitly modeled in our study) are implicitly fixed relative to net national income. This provides a convenient method for indexing the transfer payments accruing to households. Because tariff liberalization typically results in a reduction in tax revenues, a replacement tax is needed. In this chapter, we assume that income taxes on all earnings rise by an equal proportion so to ensure that tax revenues remain fixed relative to net national income. Of course, in those rich countries in which tariffs are low and agricultural subsidies are high, this tax rate may fall in the wake of trade liberalization. While we do not believe that the income tax will be the replacement tax of choice in many economies, particularly the poorest economies, it is a convenient tool, and we do not possess sufficient detail on the tax structure in many of these economies to improve greatly on this simple assumption. As we show below, the omission of the tax replacement effect has a dramatic impact on our poverty results. This impact is thereby highlighted as a key issue for consideration in the country-specific case studies in this volume and elsewhere. Model Results and Discussion In this chapter, we report the results of only one core simulation: the removal globally of all agricultural production and export subsidies and all agricultural and nonagricultural merchandise trade taxes.7 Table 4.5 reports our estimates of the percent change in the national poverty headcount for each of our focus coun- tries based on this global liberalization simulation. We use the decomposition method of Harrison, Horridge, and Pearson (2000) to identify the impacts of agricultural and nonagricultural policy reform separately from the total in Table 4.5. Summary of the Poverty Impacts of Global Reforms, 15 Countries (percent change in the headcount) Core results of this study Different Different tax replacement data, Country Ag reform Nonag reform Total Level change, 1,000s GTAP 6 Poor are exempt VAT Bangladesh 0.25 0.51 0.26 116 0.28 5.30 0.01 Brazil 2.53 0.38 2.15 50 1.41 10.00 1.42 Chile 4.76 0.12 4.64 14 4.99 12.25 4.79 Colombia 0.72 0.63 0.09 3 0.10 4.05 0.03 Indonesia 1.05 0.49 0.56 84 1.45 5.23 0.53 Malawi 1.64 0.26 1.91 81 1.84 5.62 1.31 Mexico 0.78 0.35 1.13 105 1.35 0.48 1.15 Mozambique 1.15 0.15 1.00 61 0.69 4.34 0.29 Peru 0.64 0.16 0.80 35 0.79 5.24 0.67 Philippines 1.37 0.42 0.95 108 0.75 6.39 1.92 Thailand 11.19 0.93 10.26 121 8.87 28.05 5.83 Uganda 0.01 0.09 0.09 15 0.06 5.96 0.09 Venezuela, R.B. de 0.15 0.71 0.86 28 0.86 2.12 1.14 Vietnam 0.48 5.26 5.74 88 5.85 23.58 6.96 Zambia 0.02 0.13 0.11 7 0.09 2.00 1.25 Average 1.66 0.05 1.71 n.a. 1.59 8.04 1.30 Absolute average 1.78 0.71 2.04 n.a. 1.96 8.04 1.83 Sign consistency 0.93 0.07 0.84 n.a. 0.81 1.00 0.71 Source: Model simulations by the authors. Note: The table shows the results of simulations by the authors using Gempack software (Harrison and Pearson 1996) and the GTAP Database (Dimaranan 2006). The impacts of agricultural (Ag) and nonagricultural (Nonag) reform add to the total percent change in poverty. We use the subtotal routine in Gempack that has been developed by Harrison, Horridge, and Pearson (2000) to isolate these portions of the total impact. Level change is the calculated number of persons moving out of poverty given the initial headcount and the predicted percent change. The different data simulation uses the GTAP version 6 (base year 2001) protection data with the modifications of Hertel et al. (2007, 2009). The different tax replacement, poor are exempt scenario assumes that the poor are not subject to the higher income tax needed to replace lost tariff revenue. The different tax replacement, value added tax (VAT) scenario replaces lost tariff revenue through a consumption tax that may be 131 viewed as a value added tax equivalent under which imports are taxed, while exports are exempted. Average is the simple average of the percent changes, while absolute average is the simple average of the absolute value of individual percent changes. Sign consistency measures the consistency between the direction of effects and the level of impact and is calculated as the ratio of the average to the absolute average, AV AAV (Ivanic 2006). n.a. not applicable. 132 Agricultural Price Distortions, Inequality, and Poverty percentage terms. We also provide level changes (in thousands) in the national poverty headcount (table 4.5, column 4). Impacts of the new data on price distortions Table 4.5 offers a comparison of the total poverty reduction based on previous work (Hertel et al. 2009) that did not feature the distortion estimates compiled by Anderson and Valenzuela (2008) or the updated OECD estimates of protection in OECD member countries. Comparing table 4.5, columns 3 and 5, we find mostly small differences in the percent changes in the poverty headcount of our study (column 3) compared with the changes calculated in the GTAP Database (column 5). However, some differences are worth noting. For Colombia, the previous results indicated that global liberalization would lead to a slight increase in poverty (a 0.1 percent rise in the national headcount), whereas our results anticipate that global trade reform will lead to a slight reduction in poverty. Colombia is one of the countries in which the information on agricultural protection changed signif- icantly because of the incorporation of the estimates of Anderson and Valenzuela (2008). More significantly, our predicted poverty reduction in Indonesia is some- what lower and, in Brazil, somewhat higher than the poverty reduction estimates reported previously. The crucial role of tax replacement The last two columns of table 4.5 present differing estimates of the predicted poverty changes that follow if we alter our assumption about tax adjustments. One variant is to assume that the poor are not subjected to the income tax replacement mechanism. Another is to use an alternative tax replacement instrument, namely, a value added tax equivalent. This allows us to compare our results with the results of studies that rely on different assumptions, and it highlights the importance of the tax replacement assumption on the predicted changes in poverty. Table 4.5, column 6 reports the percent changes in the national poverty head- counts if the poor are not subject to the replacement income tax. This is the assumption of Anderson, Martin, and van der Mensbrugghe (2006) in their analy- sis of the poverty impacts of the Doha Development Agenda. As one may see, this accounts for a marked difference in the predicted poverty reduction. Trade reforms that, without this assumption, were marginally poverty reducing in most cases are now universally poverty reducing by a considerable magnitude. The assumption reduces the poverty rate by roughly one-quarter in Thailand and Vietnam, for example. In this scenario, the poor are being given access to com- modities or are able to sell them at undistorted prices without having to bear any Poverty Impacts in 15 Countries: The GTAP Model 133 of the direct tax burden involved in replacing the lost tariff revenue. Indeed, this reform represents a significant implicit income transfer from nonpoor house- holds to poor households. We do not argue that such a fiscal transfer is not to be desired, of course, because the transfer would have tremendous poverty reduction benefits. However, we do not believe this is the measure that would most likely be used, given that most developing countries seek to make up for lost tariff revenue by resorting to a value added tax. The final column in table 4.5 reports the poverty results if we replace the lost tariff revenue by adjusting the value added tax. However, care must be taken in adjusting this tax because some sectors are exempt from the tax (public con- sumption, for example, and, often, basic foods). In this alternative scenario, we adjust the tax through an equiproportional adjustment in the power of the con- sumption tax (that is, one, plus the consumption tax rate) on taxable items in each focus country. This might be viewed as equivalent, in effect, to a value added tax replacement experiment if the value added tax applies to all imports and exempts all exports. However, if the existing consumption tax structure is already distorted (as in the case in our model, for example, because some sectors are exempt), this replacement consumption tax exacerbates the distortions. In particular, because it does not apply to public consumption, the tax distorts the allocation of resources between public consumption and private consumption. Therefore, in our framework, this value added tax scenario is expected to pro- duce less beneficial outcomes in terms of poverty reduction relative to the core scenario. Indeed, this is confirmed by a comparison of table 4.5, columns 3 and 7: if lost tariff revenue is replaced by a consumption tax instead of an income tax, the poverty gains are more modest, and the mix of countries that reduce poverty is altered slightly. Summary of the poverty impacts In the final three rows of table 4.5, columns 1­3, we report on the summary mea- sures introduced by Ivanic (2006). Specifically, we compute, first, the average value across countries (treating each country as an observation with equal weight); then, we report the average absolute value, which indicates the impor- tance of a given change regardless of sign; and, finally, we report the ratio of these two: AV AAV SC (average/absolute average sign consistency). Note that this ratio--which may be viewed as the tendency for trade reforms to reduce (or increase) poverty--is constructed such that 1 SC 1. If SC 1, then a given trade reform (or set of reforms) is poverty reducing for all countries in the sample because the average change and the absolute value of the changes are of the same magnitude, but opposite in sign. 134 Agricultural Price Distortions, Inequality, and Poverty From the average across countries, we see that our current results, based on updated protection data, predict larger poverty-reducing impacts in developing countries relative to the previous analysis (AV 1.71 versus AV 1.59) and a slightly greater prevalence of poverty-reducing impacts across the countries in our sample (AAV 0.84 versus AAV 0.81). In summary, the results predict that full global trade liberalization would reduce poverty in 10 of our 15 focus countries. Even if we ignore the growth-enhancing effects of trade reform (which are not included in our comparative static framework), this would mean that no less than 816,000 people would be lifted out of poverty. The decomposition of total impacts between agricultural and nonagricultural reforms (table 4.5, first two columns) shows that the impact of the global agricul- tural reforms on the poor is more than twice the impact of the nonagricultural reforms (AAVag 1.78 versus AAVnonag 0.71). Furthermore, the agricultural reforms are nearly always poverty reducing (SC 0.93), whereas the nonagricul- tural reforms tend to be only marginally poverty reducing among this sample of countries (SC 0.07). Of the five countries experiencing increases in poverty in the wake of combined agricultural and nonagricultural trade reforms, Bangladesh and Mexico are the most important in terms of absolute numbers (more than 100,000 in each case). These countries currently enjoy preferential market access in their most important markets (the European Union for Bangladesh and the United States for Mexico). Global trade liberalization would lead to substantial preference erosion for both countries in these preferred markets. Why are agricultural reforms more poverty-friendly than nonfarm trade reform? We begin our quest to understand the differential impact of trade reforms across products by focusing on the difference between the poverty impacts of agricul- tural and nonagricultural reforms; we then turn to an individual commodity decomposition of agricultural reforms. A natural method of investigating the difference between farm and nonfarm reforms is via the decomposition proposed in equation (4.6). Because the elastici- ties in this expression are the same (initially) in both experiments, the entire dif- ference in the change in the poverty headcount is accounted for by real after-tax wage changes. Specifically, we have the following decomposition of the difference in poverty headcount by region: N p N Hr,diff = a a b rs # ers # a rsj(W R NR rj,agr - W rj,nagr), (4.7) s j N rj,agr where W R is the real after-tax change in earnings for endowment j in region r N rj,nagr is the nonagricultural counterpart. owing to agricultural reforms, and W R Poverty Impacts in 15 Countries: The GTAP Model 135 Our first task is to explain why the earnings are differentially affected; then, we must consider the interaction of these changes with region, stratum, and factor elasticities to determine the differential impact on national poverty. N rj,agr - W R ) for all sources of earnings and Table 4.6 reports the values of (W R N rj,nagr all focus regions in our analysis. Relative to nonagricultural trade reforms, agricul- tural trade reforms raise the returns to farming in all our focus developing countries. Relative to nonagricultural reforms, they also raise the returns to unskilled wage labor in most countries as the unskilled labor-intensive agricultural sector expands and boosts wages among unskilled labor. (Indonesia, Mexico, and Vietnam are the excep- tions because the nonagricultural reforms exert a stronger influence on unskilled wages there.) Agricultural reforms are less favorable for skilled labor,but the reward of this factor is relatively unimportant for the poor (see table 4.4). Continuing across the columns in table 4.6, we see that agricultural reforms often, but not always lead to relatively lower real after-tax earnings relative to nonagricultural reforms for self-employed nonagricultural endowments; this is evident in the many negative entries in columns 7­9. Relative to nonagricultural reforms, agricultural reforms that increase food prices tend to hurt transfer- dependent households because transfers (by assumption) are indexed to net national income and not to the cost of living at the poverty line (column 10). This effect is attributable to the relatively large share of food expenditures in the consumption baskets of the poorest households. If we take into account the aggregate poverty elasticities with respect to factor earnings reported in table 4.4, it is not immediately clear from the pattern of earn- ings differences shown in table 4.6 why agricultural trade reforms dominate nona- gricultural reforms as a poverty reduction tool. However, this result depends on the responsiveness of the national poverty headcounts to each of the sources of earnings. We gain more insight into this issue by examining table 4.7, which reports the poverty-weighted elasticity counterparts to the earnings differences N rj,nfarm). N rj,farm - W R reported in table 4.6, that is, © b rs # ers # a p (W R rsj s If we sum across a row in table 4.7 (that is, if we sum across all endowments in a given country), we obtain the percent difference in the poverty headcount stem- ming from agricultural and nonagricultural reforms (subject to any rounding error arising from the differencing of percentages; this is the reason these entries differ from the simple difference between the first two columns of table 4.5). The final col- umn in table 4.7 indicates that, in all countries except Mexico and Vietnam, agricul- tural reforms are more poverty friendly relative to nonagricultural reforms. What is special about Mexico and Vietnam? We note that these are two of the three countries in which real after-tax unskilled wages rise more under nonagri- cultural trade reforms (the other country is Indonesia; see table 4.6). Unlike Indonesia, where the share of the poor in agriculture-specialized households is Table 4.6. Earnings Differences after Agricultural and Nonagricultural Reforms, 15 Countries 136 (percent change) Ag Ag Unskilled Skilled Nonag Nonag unskilled skilled Ag wage wage unskilled skilled Nonag Country Land labor labor capital labor labor labor labor capital Transfers Bangladesh 0.33 0.38 0.95 0.86 0.56 1.68 0.62 1.69 1.52 1.17 Brazil 40.01 16.35 15.10 14.81 1.34 0.75 0.72 0.91 1.58 0.13 Chile 15.35 7.57 6.93 6.74 0.29 0.86 1.03 0.88 1.33 0.85 Colombia 8.22 4.40 3.97 3.49 1.10 0.30 0.36 0.30 0.81 0.35 Indonesia 3.87 1.76 1.59 1.49 0.50 0.83 1.53 0.86 1.09 0.59 Malawi 4.76 3.20 2.68 2.82 1.75 0.77 0.75 0.77 0.68 1.21 Mexico 0.18 0.10 0.06 0.06 0.19 0.10 0.21 0.10 0.20 0.26 Mozambique 8.58 4.93 3.52 4.16 1.42 1.41 0.18 1.43 0.50 0.60 Peru 12.95 6.94 5.69 5.92 1.07 1.25 1.44 1.49 0.92 0.11 Philippines 3.07 1.68 1.53 1.21 0.35 0.07 0.56 0.03 0.67 0.36 Thailand 34.07 17.46 14.29 12.92 4.19 1.59 2.01 1.99 4.38 2.11 Uganda 0.62 0.50 0.47 0.44 0.40 0.34 0.25 0.34 0.20 0.37 Venezuela, R.B. de 0.87 0.68 0.70 0.44 0.49 0.55 0.46 0.55 0.01 0.11 Vietnam 15.75 3.87 3.89 5.83 9.07 8.93 12.81 8.93 4.87 6.40 Zambia 4.00 1.87 1.60 1.47 0.56 0.03 0.17 0.02 0.49 0.15 Average 10.15 4.76 4.19 4.17 0.25 0.80 1.19 0.86 0.96 0.49 Absolute average 10.17 4.78 4.20 4.18 1.55 1.30 1.54 1.35 1.28 0.98 Sign consistency 1.00 1.00 1.00 1.00 0.16 0.62 0.77 0.64 0.75 0.50 Source: Model simulations by the authors. Note: The table shows the differences in impacts on real after-tax earnings according to endowment following agricultural and nonagricultural reforms. See the note to table 4.5 for simulation information. Table 4.7. Earnings-Specific Differences between Agricultural and Nonagricultural Reforms in the Changes in Poverty, 15 Countries (percent change in the headcount) Ag Ag Unskilled Skilled Nonag Nonag unskilled skilled Ag wage wage unskilled skilled Nonag Country Land labor labor capital labor labor labor labor capital Transfers Total Bangladesh 0.00 0.12 0.00 0.01 0.23 0.05 0.22 0.00 0.04 0.08 0.76 Brazil 0.10 1.43 0.60 0.15 1.09 0.04 0.10 0.00 0.00 0.03 3.21 Chile 2.30 2.45 0.00 1.48 0.16 0.00 0.04 0.00 0.00 0.75 5.60 Colombia 0.00 1.07 0.00 0.00 0.17 0.00 0.10 0.00 0.01 0.04 1.38 Indonesia 0.30 2.16 0.02 0.08 0.23 0.02 0.66 0.01 0.10 0.05 1.49 Malawi 0.10 0.95 0.00 0.11 0.12 0.00 0.03 0.00 0.02 0.10 1.43 Mexico 0.00 0.02 0.00 0.00 0.15 0.00 0.03 0.00 0.00 0.23 0.43 Mozambique 0.04 1.12 0.00 0.04 0.11 0.00 0.02 0.00 0.05 0.08 1.31 Peru 0.12 1.17 0.00 0.07 0.11 0.01 0.70 0.11 0.06 0.02 0.59 Philippines 1.54 0.01 0.04 0.32 0.23 0.00 0.17 0.00 0.09 0.08 1.79 Thailand 2.95 9.08 0.89 0.54 2.55 0.21 0.21 0.03 0.13 2.06 13.38 Uganda 0.02 0.02 0.00 0.03 0.01 0.01 0.00 0.00 0.01 0.01 0.10 Venezuela, R.B. de 0.00 0.05 0.00 0.00 0.34 0.02 0.16 0.00 0.00 0.01 0.56 Vietnam 0.17 0.45 0.00 0.03 0.01 0.00 2.27 0.01 2.24 1.31 5.20 Zambia 0.01 0.01 0.00 0.01 0.14 0.00 0.02 0.00 0.04 0.01 0.15 Average 0.51 1.34 0.10 0.19 0.32 0.01 0.24 0.01 0.18 0.28 1.74 Absolute average 0.51 1.34 0.10 0.19 0.38 0.02 0.32 0.01 0.19 0.32 2.49 Sign consistency 1.00 1.00 1.00 1.00 0.86 0.53 0.77 0.93 0.95 0.87 0.70 Source: Model simulations by the authors. 137 Note: The results correspond with equation (4.7) and represent the differences in the impacts of agricultural and nonagricultural reforms on poverty according to endowment. See the note to table 4.5 for simulation information. 138 Agricultural Price Distortions, Inequality, and Poverty about 40 percent (see table 4.2), Mexico and Vietnam show relatively low shares of agricultural households in total poverty. Both these countries have relatively high shares of rural diversified households in poverty that rely heavily on unskilled wage labor, which is more favorably affected by nonagricultural trade reforms. While higher returns to unskilled family labor in agriculture represent the dominant driver of agricultural poverty reduction relative to nonagricultural poverty reduction in most countries, there are important exceptions. In Bangladesh and Zambia, agricultural reforms have a more favorable impact on poverty than do nonagricultural reforms largely because of the wage labor chan- nel. In the Philippines, farm households benefit relatively more from higher returns to agricultural land. We also examine the relative impact of agricultural and nonagricultural trade reforms on poverty by stratum. This is shown in table 4.8, which reports the Table 4.8. Stratum-Specific Differences in the Changes in Poverty after Agricultural and Nonagricultural Reforms, 15 Countries (contribution to percent change in the headcount) Urban Rural Urban Rural Country Ag Nonag labor labor Transfer diverse diverse Total Bangladesh 0.10 0.17 0.03 0.09 0.02 0.08 0.27 0.76 Brazil 1.67 0.08 0.55 0.40 0.01 0.46 0.23 3.21 Chile 4.47 0.02 0.05 0.04 0.58 0.97 0.67 5.60 Colombia 0.98 0.09 0.06 0.07 0.04 0.08 0.06 1.38 Indonesia 1.76 0.39 0.02 0.10 0.03 0.03 0.23 1.49 Malawi 0.86 0.02 0.00 0.09 0.03 0.02 0.40 1.43 Mexico 0.01 0.02 0.03 0.05 0.17 0.05 0.10 0.43 Mozambique 0.58 0.04 0.02 0.05 0.04 0.21 0.45 1.31 Peru 0.68 0.64 0.01 0.03 0.01 0.18 0.33 0.59 Philippines 0.62 0.06 0.03 0.04 0.02 0.41 0.77 1.79 Thailand 2.62 0.09 0.05 0.57 0.63 0.94 9.94 13.38 Uganda 0.01 0.00 0.00 0.00 0.00 0.01 0.06 0.10 Venezuela, R.B. de 0.04 0.13 0.21 0.11 0.00 0.04 0.03 0.56 Vietnam 0.08 1.16 0.00 0.00 0.27 0.11 3.74 5.20 Zambia 0.00 0.01 0.08 0.03 0.00 0.04 0.01 0.15 Average 0.96 0.14 0.07 0.09 0.11 0.22 0.64 1.74 Absolute average 0.97 0.20 0.08 0.11 0.12 0.24 1.15 2.49 Sign consistency 1.00 0.72 0.91 0.82 0.86 0.91 0.56 0.70 Source: Model simulations by the authors. Note: The results correspond with equation (4.7) and represent the differences in the impacts of agricul- tural and nonagricultural reforms on poverty according to household type. See the note to table 4.5 for simulation information. Poverty Impacts in 15 Countries: The GTAP Model 139 N rj,nfarm) for all 15 countries and all N rj,farm - W R elements of © j b rs # ers # a p (W R rsj seven strata. The columns now refer to strata, and each element represents the combined impact of all changes in earnings (adjusted for changes in taxes and the cost of living) on national poverty through the changes in poverty in the individ- ual household strata. Once again, the final column records the differences between agricultural and nonagricultural reforms in terms of the percent change in the national poverty headcount. Here, we see that agricultural trade reforms reduce poverty among agriculture-specialized households in nearly all countries (note that SC, the ratio of the average and absolute average headcounts, is 1.00 under this stratum). The contribution of the agricultural stratum to poverty reduction is negligible only in Mexico and Vietnam, and these are the only two countries in which poverty among diversified rural households would fall more after nonagricultural trade reforms than after agricultural reforms (leading to positive entries in the rural diversified column in table 4.8). In both Mexico and Vietnam, the contribu- tions of all nonagricultural household strata to reductions in poverty are more significant after multilateral nonagricultural reforms than after agricultural reforms. Understanding the impacts on poverty of particular farm commodity policies Because of the importance of agricultural reform, we now turn to the task of decomposing poverty impacts according to agricultural commodity. In table 4.9, we decompose the reduction in poverty after agricultural reform (see table 4.5) into component parts; in this case, our breakdown is according to the global commodity market in which the reform has occurred. Thus, the first set of columns reports the percent change in national poverty headcounts caused by global reforms in the foodgrains sector. As a result, we see that, with the exception of Vietnam (and Mexico and República Bolivariana de Venezuela, in which there is no effect), the liberalization in foodgrain markets is generally poverty reducing. The disaggregation of the foodgrains reforms according to instrument reveals that tariff cuts in foodgrains--as well as feed- grains, which are also shown in table 4.10--are universally poverty reducing (the sign consistency index is 1). If poverty increases in a country because of reforms in food or feedgrains, this effect is caused by adverse price impacts that arise from the elimination of export subsidies and domestic support (see the export subsidies and domestic subsidies columns in table 4.10). Apart from foodgrains, the reforms in other crops are the most poverty friendly. These 140 Agricultural Price Distortions, Inequality, and Poverty Table 4.9. Change in Poverty after Commodity-Specific Reforms, 15 Countries (percent change in the headcount) Other Country Foodgrains Feedgrains Sugar Cotton crops Dairy Meat Bangladesh 0.07 0.16 0.02 0.18 0.11 0.04 0.02 Brazil 0.29 0.20 0.20 0.01 0.27 0.01 1.56 Chile 0.18 0.04 0.03 0.03 3.11 0.40 1.03 Colombia 0.11 0.06 0.08 0.01 0.55 0.09 0.04 Indonesia 0.09 0.00 0.29 0.00 0.36 0.03 0.86 Malawi 0.14 0.02 0.05 0.04 1.51 0.02 0.00 Mexico 0.00 0.48 0.03 0.06 0.23 0.03 0.02 Mozambique 0.10 0.17 0.53 0.08 0.36 0.01 0.09 Peru 0.07 0.01 0.01 0.02 0.01 0.02 0.60 Philippines 0.20 0.23 0.01 0.07 0.79 0.11 0.19 Thailand 5.63 0.91 0.88 0.06 2.89 0.01 0.93 Uganda 0.01 0.02 0.00 0.02 0.03 0.01 0.00 Venezuela, R.B. de 0.00 0.02 0.00 0.00 0.04 0.03 0.04 Vietnam 0.03 0.40 0.07 0.00 0.18 0.13 0.01 Zambia 0.01 0.04 0.01 0.01 0.04 0.01 0.03 Average 0.46 0.11 0.10 0.01 0.64 0.01 0.33 Absolute average 0.46 0.18 0.15 0.04 0.70 0.06 0.36 Sign consistency 0.99 0.57 0.65 0.31 0.92 0.23 0.90 Source: Model simulations by the authors. Note: The results correspond with equation (4.7) and represent the impacts of reforms on poverty according to commodity sector. See the note to table 4.5 for simulation information. crops include many of the tropical products of which developing countries are net exporters. We report only foodgrains and feedgrains here, but an all- commodity ranking on poverty friendliness, as we have used the term, would appear as follows: foodgrains, other food and beverages, other crops, meats, sugar, feedgrains, cotton, and dairy. In terms of the absolute size of the impacts on poverty, the liberalization of other crops is the most significant, followed by liberalization in foodgrains and then meats. Summary and Conclusions This chapter offers a complementary perspective to the detailed country case studies of trade reform and poverty provided by others in this volume. By looking at a wide range of developing countries in Africa, Asia, and Latin America, we Poverty Impacts in 15 Countries: The GTAP Model 141 Table 4.10. Change in Poverty Caused by Foodgrain and Feedgrain Reforms, 15 Countries (percent change in the headcount) Foodgrains Feedgrains Export Domestic Export Domestic Country Tariffs subsidies subsidies Total Tariffs subsidies subsidies Total Bangladesh 0.02 0.02 0.03 0.07 0.17 0.00 0.01 0.16 Brazil 0.31 0.01 0.01 0.29 0.07 0.01 0.13 0.20 Chile 0.22 0.01 0.05 0.18 0.00 0.02 0.02 0.04 Colombia 0.14 0.00 0.02 0.11 0.12 0.03 0.14 0.06 Indonesia 0.18 0.01 0.08 0.09 0.05 0.01 0.04 0.00 Malawi 0.18 0.00 0.03 0.14 0.02 0.00 0.04 0.02 Mexico 0.16 0.00 0.15 0.00 0.44 0.01 0.92 0.48 Mozambique 0.13 0.01 0.02 0.10 0.17 0.00 0.01 0.17 Peru 0.11 0.01 0.04 0.07 0.05 0.02 0.04 0.01 Philippines 0.34 0.02 0.12 0.20 0.31 0.01 0.08 0.23 Thailand 5.45 0.03 0.15 5.63 1.09 0.00 0.18 0.91 Uganda 0.02 0.00 0.01 0.01 0.02 0.00 0.00 0.02 Venezuela, R.B. de 0.02 0.00 0.02 0.00 0.03 0.00 0.05 0.02 Vietnam 0.02 0.01 0.04 0.03 0.44 0.02 0.01 0.40 Zambia 0.01 0.00 0.00 0.01 0.05 0.00 0.01 0.04 Average 0.49 0.00 0.03 0.46 0.20 0.01 0.09 0.11 Absolute average 0.49 0.01 0.05 0.46 0.20 0.01 0.11 0.18 Sign consistency 1.00 0.17 0.54 0.99 1.00 0.60 0.82 0.57 Source: Model simulations by the authors. Note: The results correspond with equation (4.7) and represent the impacts of reforms on poverty according to type of reform in two commodity sectors. The totals are the sums across instruments and reflect the total contribution of sector-specific reforms to the percent change in poverty headcounts nationwide. See the note to table 4.5 for simulation information. provide more general conclusions about the poverty impacts of commodity trade reform. We find that, overall, trade reform tends to reduce poverty, and it does so because of the agricultural trade reform components in the total package of reforms. Indeed, nonagricultural trade reforms tend to increase poverty in most of our focus developing countries. We explore the relative poverty friendliness of agricultural trade reforms in more detail by examining the differential impacts (relative to nonagricultural reforms) on real after-tax factor returns and on poverty by stratum. Overall, the more favorable impacts of agricultural reforms are driven by the increased factor rewards for farm households, as well as the higher returns to unskilled wage labor, which we have evaluated relative to the real cost of living at the poverty line. 142 Agricultural Price Distortions, Inequality, and Poverty Finally, we have examined the poverty impacts of trade reform according to agri- cultural commodity groups across our sample of countries. Not surprisingly, foodgrain reforms are the most poverty-friendly group. Notes 1. Note that, throughout this chapter, as with the rest of the book, nonagriculture includes highly processed foods and beverages, while farm goods that require light processing before they may be traded easily, such as rice, sugar, dairy products, and meat, are included in agriculture. 2. Specifically, the set of measures envisioned under the Doha Development Agenda focus on the implementation of the July 2004 framework agreement of the Doha Development Agenda. 3. A clear limitation of this approach stems from the rigidity of the households' classification by earnings specialization. Obviously, households may be induced to change their specialization or diver- sify in response to changing relative factor returns. We believe that the relatively broad definition of the seven strata circumvents this problem for the majority of households in the face of modest earnings changes. However, this important qualification is considered in more detail in the subsequent results section. 4. We make adjustments to the output subsidy ad valorem rate in instances in which the differ- ence between the reported output subsidy rate in 2001 and the corresponding rate in 2004 is at least 1 percent. Most of the changes occur in the case of the United States, where the ad valorem output sub- sidy for rice falls by 77 points and the subsidy for oilseeds falls by 24 points. The world prices for oilseeds and rice were low in 2001 relative to 2004, and these commodities factor heavily in our updates of protection in the OECD countries. 5. Francois, van Meijl, and van Tongeren (2005) introduce monopolistic competition in the manu- facturing sector into their analysis of World Trade Organization reforms. The resulting variety and scale effects generally boost the gains to rich countries and dampen the gains to poor countries gener- ated by rich-country reforms. However, this makes their model less stable, and, given our focus on agricultural reforms, this feature seems less critical. 6. Land in the GTAP Database is specific to agriculture. Therefore, we model imperfect mobility across agricultural uses to represent the cost of converting land from one use to another. 7. More specifically, our global liberalization includes the removal of all border measures (export subsidies and trade taxes) in all regions of the model and the removal of all input and output subsidies in agriculture in the OECD and in those developing countries on which Anderson and Valenzuela (2008) provide new information (see the text for explanations). References Anderson, K., and W. Martin, eds. 2006. Agricultural Trade Reform and the Doha Development Agenda. London: Palgrave Macmillan; Washington, DC: World Bank. Anderson, K., W. Martin, and D. van der Mensbrugghe. 2006. "Global Impacts of the Doha Scenarios on Poverty." In Poverty and the WTO: Impacts of the Doha Development Agenda, ed. T. W. Hertel and L. A. Winters, 497­528. London: Palgrave Macmillan; Washington, DC: World Bank. Anderson, K., and E. Valenzuela. 2008. "Estimates of Global Distortions to Agricultural Incentives, 1955­2007." Data spreadsheet, October, World Bank, Washington, DC. http://go.worldbank.org/ YAO39F35E0. Cranfield, J. A. L., P. V. Preckel, J. S. Eales, and T. W. Hertel. 2003. "Estimating Consumer Demands across the Development Spectrum: Maximum Likelihood Estimates of an Implicit Direct Additiv- ity Model." Journal of Development Economics 68 (2): 289­307. Poverty Impacts in 15 Countries: The GTAP Model 143 de Janvry, A., M. Fafchamps, and E. Sadoulet. 1991. "Peasant Household Behavior with Missing Markets: Some Paradoxes Explained." Economic Journal 101 (409): 1400­17. Dimaranan, B. D., ed. 2006. Global Trade, Assistance, and Protection: The GTAP 6 Data Base. West Lafayette, IN: Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University. https://www.gtap.agecon.purdue.edu/databases/v6/v6_doco.asp. Francois, J. F., H. van Meijl, and F. van Tongeren. 2005. "Trade Liberalization in the Doha Development Round." Economic Policy 20 (42): 349­91. Golub, A., and T. W. Hertel. 2008. "Global Economic Integration and Land Use Change." Journal of Economic Integration 23 (3): 463­88. Harrison, W. J., M. Horridge, and K. R. Pearson. 2000. "Decomposing Simulation Results with Respect to Exogenous Shocks." Computational Economics 15 (3): 227­49. Harrison, W. J., and K. R. Pearson. 1996. "Computing Solutions for Large General Equilibrium Models using Gempack." Computational Economics 9 (2): 83­127. Hertel, T. W., ed. 1997. Global Trade Analysis: Modeling and Applications. New York: Cambridge University Press. Hertel, T. W., M. Ivanic, P. V. Preckel, and J. A. L. Cranfield. 2004. "The Earnings Effects of Multilateral Trade Liberalization: Implications for Poverty." World Bank Economic Review 18 (2): 205­36. Hertel, T. W., R. Keeney, M. Ivanic, and L. A. Winters. 2007. "Distributional Effects of WTO Agricul- tural Reforms in Rich and Poor Countries." Economic Policy 22 (50): 289­337. ------. 2009. "Why Isn't the Doha Development Agenda More Poverty Friendly?" Review of Devel- opment Economics. Wiley InterScience. http://www3.interscience.wiley.com/journal/122211891/ abstract (published online February 23). Hertel, T. W., and L. A. Winters, eds. 2006. Poverty and the WTO: Impacts of the Doha Development Agenda. London: Palgrave Macmillan; Washington, DC: World Bank. Ivanic, M. 2006. "The Effects of a Prospective Multilateral Trade Reform on Poverty in Developing Countries." In Poverty and the WTO: Impacts of the Doha Development Agenda, ed. T. W. Hertel and L. A. Winters, 405­26. London: Palgrave Macmillan; Washington, DC: World Bank. Keeney, R., and T. W. Hertel. 2005. "GTAP-AGR: A Framework for Assessing the Implications of Mul- tilateral Changes in Agricultural Policies." GTAP Technical Paper 24 (August), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=1869. OECD (Organisation for Economic Co-operation and Development). 2001. Market Effects of Crop Support Measures. Paris: OECD. ------. 2007. Agricultural Policies in OECD Countries: Monitoring and Evaluation 2007. Paris: OECD. Rimmer, M., and A. Powell. 1992. "An Implicitly Additive Demand System." Applied Economics 28 (12): 1613­22. Valenzuela, E., and K. Anderson. 2008. "Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980­84." Research Memorandum 13 (December), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925. Winters, L. A., N. McCulloch, and A. McKay. 2004. "Trade Liberalization and Poverty: The Evidence So Far." Journal of Economic Literature 42 (1): 72­115. Part III NATIONAL CGE APPROACHES: ASIA 5 China Fan Zhai and Thomas W. Hertel As the most populous nation in the world, China plays a critical role in the deter- mination of the global poverty headcount. Indeed, a considerable portion of the reduction in the headcount may be attributable to the remarkable reduction in the incidence of poverty in China over the past two decades. Chen and Ravallion (2004) find that, in 1981, 65 percent of the population in China was in extreme poverty (US$1 a day), whereas by 2001, this share had fallen to around 12 percent. They show that much of this poverty reduction was driven by reforms in the agri- cultural sector. These advances notwithstanding, rural poverty continues to dom- inate the national poverty headcount in China, and the headcount is highest among households that are specialized in farming. Furthermore, there is evidence that, despite rapid economic growth, the rural-urban wage gap is widening (Sicular et al. 2007). Within the rural sector, rapid nonagricultural income growth and slow agricultural income growth since the 1990s is contributing to increased rural inequality (Benjamin et al. 2007). The agricultural sector therefore contin- ues to play an important role in the determination of national poverty and inequality in China. In this chapter, we focus on the impact on poverty and inequality in China of agricultural, trade, and rural policy reforms both at home and abroad. The impact of trade reforms on poverty and inequality in China has been a topic of intense research over the past decade, culminating in a number of studies focusing on the impact on poverty of China's accession to the World Trade Organ- ization (WTO) (for example, see Bhattasali, Li, and Martin 2004). Chen and Ravallion (2004) examine these impacts at a highly disaggregate level using earn- ings and price estimates from another study; they estimate that WTO accession will benefit urban households, particularly poor urban households. However, they predict that, as a result of the WTO accession, falling rural wages and increases in 147 148 Agricultural Price Distortions, Inequality, and Poverty the consumer prices faced by rural households are likely to hurt the rural poor. In a companion study, Hertel, Zhai, and Wang (2004) aggregate households to a greater degree, but incorporate them directly into their computable general equi- librium (CGE) model of China. They, too, conclude that WTO accession will be relatively more favorable for urban households, but they argue that whether or not rural households will lose from these reforms depends critically on the degree of off-farm labor mobility. At low (or zero) mobility, as assumed by Chen and Ravallion, the poorest rural households lose from reform, but, as the off-farm labor supply elasticity rises, the potential for farm households to gain increases. In closely related work, Hertel and Zhai (2006) contrast the impacts of com- modity market reforms, such as those initiated through China's WTO accession, with factor market reforms aimed at facilitating an improved flow of labor out of agriculture and between the rural and urban markets. They find that the latter may result in significant gains for rural households. Specifically, these authors explore the implications of (1) reforming agricultural land markets to permit arm's-length land rental in all rural areas, thereby facilitating the permanent movement of labor out of farming; (2) enhancing off-farm labor mobility; and (3) abolishing the hukou system, the household registration system in China, thereby reducing the transaction costs imposed on rural-urban migrants. If combined, these reforms reduce the estimated 2007 urban-rural income ratio from 2.58 (in the absence of WTO accession) to 2.09. If WTO accession is added to this mix of policy reforms, the 2007 urban-rural income ratio is still reduced, though not quite as much, to 2.12. Given the importance of labor market distortions to poverty and inequality in China, we pay special attention in our study to their presence as well. In this chapter, we update the model used in earlier studies (to reflect the most recent social accounting matrix for China) and capitalize on the most recent esti- mates of agricultural price distortions, which have changed significantly since China's accession to the WTO. We also bring to bear new farm price distortion estimates for other developing countries to assess the impact of global trade reform on poverty and inequality in China.1 Unlike our earlier work (see below), which focused solely on China's own reforms associated with WTO accession, we explore here the impacts of reforms in the rest of the world, as well as in China. Additionally, we decompose these impacts in two ways: first, by region (China versus the rest of the world) and, second, by sector (agriculture versus nonagricul- ture). We also examine the interplay between these commodity market reforms and the factor market reforms aimed at improving the allocation of labor within the Chinese economy. This chapter is organized as follows. The next section describes the speci- fication of the CGE model used in this study. We then assess the effects on the macroeconomy, agricultural production, and poverty in China that are generated China 149 by the elimination of agricultural price distortions in the rest of the world, as well as by the reduction of China's own distortions in commodity and factor markets. The final section offers conclusions. The Model and the Data The Chinese model we use is an updated version of the household-disaggregated CGE model that we have relied on previously to study the economic and poverty effects in China of WTO accession and Doha Round trade liberalization (see Hertel, Zhai, and Wang 2004; Hertel and Zhai 2006; Zhai and Hertel 2006). The model has intellectual roots in the group of single-country applied general equi- librium models used over the past two decades to analyze the impact of trade policy reform. The updated version in this chapter has a more recent benchmark data set based on the 2002 input-output table (NBS 2006) and a detailed sectoral disaggregation for agriculture and food. In this section, we describe the main features of the model. Household behavior Following our previous work, we disaggregate rural and urban households into 40 rural and 60 urban representative households according to primary source of income and relative income level. In light of the fact that our focus here is on agricultural incentives, we seek to highlight those households that depend exclu- sively on farming for their incomes. Accordingly, we stratify the rural households by agriculture-specialized households (more than 95 percent of household income is obtained through farming) and diversified (all other sources of income). We are also interested in the impacts of restrictions on rural-urban labor mobility; so, we separately identify urban households and group them into three strata: transfer-specialized, labor-specialized, and diversified households. Within each stratum, we order households from poorest to richest based on per capita income and then group them into 20 vingtiles, each containing 5 percent of the stratum population. Household income is derived from labor income, the profits from family-owned agricultural and nonagricultural enterprises, property income, and transfers. House- holds consume goods and services according to a preference structure determined by the extended linear expenditure system. Through the specification of a subsistence quantity of each good or service, this expenditure function generates nonhomothetic demands, whereby the greater the relative importance of subsistence consumption to a household (for example, it would be high for rice, but low for automobiles), the more income inelastic the demand of the household for that good. 150 Agricultural Price Distortions, Inequality, and Poverty The other important dimension of household behavior is the supply of labor for off-farm activities. In China, the off-farm labor supply decision is complicated by institutional factors that have been built into the system to keep the agricul- tural population in place; among these, the rural land tenure system is one of the most widely discussed (Zhao 1999a). The absence of well-defined land tenure has served to raise the opportunity cost of leaving the farm (Yang 1997). Households that cease to farm the land may lose the rights to the land; so, they have a strong incentive to continue some level of agricultural activity even if profitability is quite low (Zhao 1999b). Because of only modest growth in rural nonfarm activi- ties, this seriously limits the ability of households to obtain off-farm work (Zhao 1999a).2 Although an active land rental market has emerged in some regions in recent years, the overall level of land rental transactions is still low; only around 10­15 percent of rural households rent land out or are land renters (Deininger and Jin 2005, 2007; Wang, Herzfeld, and Glauben 2007). Empirical studies have found that the transaction costs associated with land rental are significant, and the absence of an efficient land rental market remains a substantial barrier to the facilitation of off-farm labor participation by rural labor (Deininger and Jin 2005; Wang, Herzfeld, and Glauben 2007). In this chapter, we model rural households as if they maximize the total return to their labor supply, which is offered on both the on-farm and the off-farm labor markets. However, the ability of households to shift labor between these two labor markets is constrained by a number of factors, including education, experience, and simple geography, which may isolate farm households from the nonfarm labor market. We proxy the combined impact of these factors through a single, finite, constant elasticity of transformation. The labor allocation between farm and off-farm jobs is determined by the ratio of the shadow value of labor in agri- culture relative to the off-farm wage rate and this elasticity of transformation.3 The constant elasticity of transformation parameter governs the off-farm labor supply elasticity, for which we adopt the estimate of 2.67 of Sicular and Zhao (2004) as the overall farm­off-farm transformation elasticity for the total rural labor force. The empirical study by Zhang, Huang, and Rozelle (2002) suggests that this elasticity increases by 0.58 for an additional year of schooling. This is translated into the farm­off-farm transformation elasticity of 0.68 for unskilled labor and 4.01 for semiskilled labor.4 Owing to the absence of an effective land market, the shadow value of labor in agriculture in this function takes into account the potential impact that reducing agricultural employment would have on the household's claim to farmland. This incremental factor is calculated as the marginal value product of land, multiplied by the probability that the household will lose its land as a result of off-farm migra- tion. To make this amenable to use in a model of a representative farm household China 151 with continuously variable labor and land use, we translate the probability into a simple elasticity of land income with respect to on-farm labor. The higher this elasticity, the greater the probability that the farm household will lose its land if the farmer shifts to an off-farm job. The benchmark elasticity in our model is 0.5, that is, a 10 percent reduction in on-farm work results in a 5 percent loss of land income. However, for purposes of sensitivity analysis, we also report results from two extreme simulations. In the first, the elasticity of land income with respect to off-farm work is zero. This is the case of a perfectly functioning land rental market with no chance of land loss. In the second sensitivity analysis, the elasticity is set equal to 1, such that the farmer leaving the farm to work in the city is virtually guar- anteed to lose his land. By comparing these two extremes, we gain an appreciation of the importance of land market reform in addressing inequality. Rural-urban migration Migration is a key part of the rural economy in China. According to rural household survey data collected in 2003 and compiled by Liu, Park, and Zhao (2006), 19.4 per- cent of all rural workers participated in migratory work in 2003, and more than 40 percent of all households had at least one member who was a labor migrant in that year. More than half the migrants had left their provinces, and most of these had migrated to the coastal provinces, where manufacturing activity and exports have been booming. The 2000 census estimated that the total number of migrants in China was 131 million, of which nearly two-thirds were non-hukou migrants. (Households lacking the hukou urban registration face limited access to many urban amenities, including housing and education.) Rural-urban migration was the largest form of migration and accounted for more than 50 million people in the 2000 cen- sus (Cai, Park, and Zhao 2007). This massive migration is a rational response to the enormous rural-urban wage gap that exists in China, which Sicular et al. (2007) recently placed at 2.27 (the ratio of urban to rural per capita disposable income in 2002) after adjusting for housing subsidies and spatial price differences. Sicular et al. find no evidence of this gap declining. Indeed, the ratio of urban to rural incomes appears to have risen slightly between 1997 and 2002. This is hardly the outcome that a standard general equilibrium model with perfect labor mobility would predict. Clearly, there are important barriers to labor movement in China that need to be considered if one hopes to assess accurately the impact of commodity market reforms on rural and urban employment, wages, and household incomes. While the rural-urban per capita income gap is an indication of a potential labor market distortion, we are more interested in discovering the hourly wage differential among workers of comparable skill and ability. If there were no barri- ers to the movement of labor between rural and urban areas, we would expect real 152 Agricultural Price Distortions, Inequality, and Poverty wages to be equalized for an individual worker with given characteristics. In an exploration of the issue of rural-urban inequality in greater detail in nine provinces using the China Health and Nutrition Survey, Shi, Sicular, and Zhao (2002) con- clude that the apparent labor market distortion is about 42 percent of the rural- urban labor income differential and 48 percent of the hourly earnings differential.5 Applied to their estimated average wage differential, this amounts to an ad valorem apparent transactions tax rate on rural wages of 81 percent.6 We model these transaction costs as real costs that are assumed by the tempo- rary rural migrants who move to urban areas without possessing hukou urban registrations. Of course, these migrants are heterogeneous, and the extent of the burden varies widely. Individuals who are single and live close to the urban area in which they are working are likely to experience minor inconvenience as a result of the temporary migration. We expect them to be the first to migrate (all else being equal) in response to higher urban wages. Other migrants have large fami- lies and must travel a great distance. Their urban living conditions are often poor, and, among these migrants, it is not uncommon to be robbed on their return home by train after their jobs. For such individuals, the decision to migrate tem- porarily is likely to be difficult, and they may or may not choose to repeat the experience. With this heterogeneous population in mind, we postulate a transac- tion cost function in which the cost increases in proportion to the increase in the rural population engaged in temporary work. This transaction cost function has a simple constant elasticity functional form, which begins at the origin, reflecting those migrants for whom there is essentially no cost because of their proximity to urban areas, and reaches the observed wage gap (adjusted for transportation and living costs) at the current level of temporary migration (about 70 million work- ers). We assume that additional increases in temporary migration have only a modest impact on these transaction costs.7 Finally, only a portion of these observed transaction costs may be attributed to the government's formal policy of migration restriction, the hukou system. Indeed, Shi (2002) finds that only 28 percent of the rural-urban wage difference may be explained directly by the coefficient on the hukou registration variable. We take this into account elsewhere below in our investigation of prospective labor market reforms by the Chinese government. Production and trade Production in each sector of the economy is modeled using nested constant elas- ticity of substitution functions, and constant returns to scale are assumed. In the top level of the nest, value added and a composite of intermediate inputs produce outputs. Then, another constant elasticity of substitution function disaggregates China 153 the value added into a capital-labor composite and agricultural land. The capital- labor composite is split into the capital­skilled labor composite and aggregated less-skilled labor. The capital­skilled labor composite consists of capital and skilled labor, while aggregated less-skilled labor is composed of semiskilled labor and unskilled labor. A low substitution elasticity of 0.3 between capital and skilled labor is assumed to introduce the capital-skill complementarity. The elasticity of substitution between semiskilled labor and unskilled labor is set at 1.5 based on estimates for the United States by Katz and Murphy (1992) and Heckman, Lochner, and Taber (1998). Each sector employs a labor composite comprising both rural and urban labor that substitute imperfectly. This is an indirect means of building a geographic fla- vor into the model because some sectors will be located largely in urban areas, while others will be predominantly in rural areas. By limiting the substitutability of rural and urban labor in each sector, we are able to proxy the economic effect of geographically distributed activity. Ideally, we would model the geographic distri- bution of industrial activity, but the data do not exist to support this split. All commodity and factor markets are assumed to clear through prices. In the case of rural labor markets, there is a segmentation between agricultural and nonagricultural labor: these two markets are linked imperfectly through the con- stant elasticity of transformation parameter (see elsewhere above). Once the transaction costs associated with temporary migration are accounted for, rural wages are equated with urban wages. Capital is assumed to be fully mobile across sectors. Import demand is modeled using the Armington assumption, that is, domestic products are assumed to be differentiated from foreign products. On the export side, it is assumed that firms treat the domestic and export markets equally. Thus, the law of one price holds, that is, the export price is identical to the price of domestic supply. The small country assumption is assumed for imports; so, world import prices are exogenous in terms of foreign currency. Exports are demanded according to constant elasticity demand curves. The terms of trade for China are therefore endogenous in the simulations. The value of export demand and the Armington elasticities are based on the elasticities used in the global CGE Linkage model (van der Mensbrugghe 2005). The benchmark data A social accounting matrix is estimated for China for the year 2002 to serve as the benchmark data set for model calibration. The matrix contains 48 sectors of production and 100 representative households based on 2000 household survey data for three provinces (Guangdong, Liaoning, and Sichuan) and the 2002 input- output table (NBS 2006). Because the 2002 input-output table has only one crop 154 Agricultural Price Distortions, Inequality, and Poverty sector and one livestock sector, we disaggregate these two sectors into eight crop sectors and four livestock sectors according to the corresponding Global Trade Analysis Project sector classification (Hertel 1997). The information on the struc- ture of production, demand, inputs, and trade in the Global Trade Analysis Project Database, version 7 are used for the sectoral disaggregation, and we employ the cross-entropy method to balance the social accounting matrix (see Robinson, Cattaneo, and El-Said 2001). The base year tariffs and export subsidies and taxes are reported in table 5.1. Protection rates for lightly processed food and agricultural products have been obtained from Huang et al. (2009) and from the Global Trade Analysis Project Database, version 7. For other primary goods and manufacturing products, tariffs are estimated based on the collected revenue from import tariffs and base year imports by commodity. As shown in table 5.1, China's tariff structure provides more protection for food and agricultural products than for nonfood manufacturing Table 5.1. The Sectoral Structure of GDP, Trade, Import Tariffs, and Export Subsidies, China, Around 2004 (percent) Tariff Export GDP Export Import Product rate subsidy ratea share share share Agriculture 6.5 0.8 13.4 1.6 2.5 Paddy rice 0.0 1.0 0.8 0.0 0.0 Wheat 4.0 0.0 0.3 0.0 0.1 Other grains 3.4 13.0 0.3 0.1 0.0 Vegetables and fruits 14.8 0.0 5.5 0.7 0.1 Oilseeds 15.9 0.0 0.4 0.1 0.7 Sugarcane and beets 15.3 0.0 0.0 0.0 0.0 Plant-based fibers 5.3 0.0 0.2 0.0 0.3 Other crops 9.4 0.0 0.1 0.3 0.0 Cattle, sheep, and so on 3.9 0.0 0.3 0.0 0.0 Other livestock 0.0 0.0 2.7 0.1 0.2 Raw milk 0.0 0.0 0.1 0.0 0.0 Wool 7.0 0.0 0.1 0.0 0.1 Forestry 2.8 0.0 1.3 0.0 0.9 Fishing 5.2 0.0 1.3 0.1 0.0 Mining 0.7 0.0 4.9 1.5 6.2 Coal mining 3.1 0.0 1.9 0.5 0.1 Crude oil and natural gas 0.9 0.0 1.9 0.4 4.1 Ore mining 0.0 0.0 0.5 0.1 1.4 Other mining 0.5 0.0 0.6 0.5 0.7 (Table continues on the following page.) China 155 Table 5.1. The Sectoral Structure of GDP, Trade, Import Tariffs, and Export Subsidies, China, Around 2004 (continued) (percent) Tariff Export GDP Export Import Product rate subsidy ratea share share share Food manufacturing 5.0 0.0 3.8 3.0 2.0 Meat products 10.5 0.0 0.2 0.5 0.3 Vegetable oils 12.5 0.0 0.3 0.1 0.5 Grain, milled 0.0 1.0 0.2 0.0 0.0 Sugar, refined 17.3 0.0 0.1 0.0 0.1 Forage 11.5 0.0 0.2 0.0 0.0 Prepared fish products 0.9 0.0 0.2 1.0 0.7 Other processed food 9.4 0.0 0.8 1.1 0.3 Beverages 12.7 0.0 0.6 0.2 0.1 Tobacco products 8.9 0.0 1.1 0.1 0.1 Nonfood manufacturing 2.9 0.0 28.6 74.9 80.5 Textiles 0.2 0.0 2.0 9.1 4.5 Apparel and leather 0.2 0.0 1.5 9.3 1.6 Sawmills and furniture 1.9 0.0 0.9 2.2 0.7 Paper, printing, and so on 3.0 0.0 2.0 3.3 2.0 Petroleum refining 3.6 0.0 0.9 0.9 1.5 Chemicals 3.3 0.0 4.9 7.3 13.0 Building materials 2.7 0.0 1.6 1.4 0.7 Metals 1.7 0.0 3.1 1.5 5.9 Metal products 2.1 0.0 1.2 3.6 2.0 Machinery 3.3 0.0 3.1 4.4 11.6 Transport equipment 16.4 0.0 2.1 2.2 3.7 Electrical machinery 2.9 0.0 1.5 6.8 6.2 Electronics 1.4 0.0 2.3 16.6 20.7 Instruments 2.1 0.0 0.4 5.0 6.0 Other manufacturing goods 0.7 0.0 1.2 1.4 0.4 Utilities, construction, services 0.0 0.0 49.3 19.0 8.8 Sources: Huang et al. (2009), drawing on version 7 of the Global Trade Analysis Database and the social accounting matrix for China for 2002. a. Negative numbers indicate the existence of an export tax. goods. Moreover, the import tariff rates show considerable cross-sectoral variation within agriculture: vegetables and fruits, oilseeds, and sugarcane and beets have high tariff rates of around 15 percent, while imports of plant-based fibers appear to be effectively subsidized. Table 5.1, columns 3­5 present China's sectoral shares of GDP, exports, and imports. Despite the diminishing importance of the agricultural sector in the Chinese economy over the last two decades, agriculture still accounts for 13.4 percent 156 Agricultural Price Distortions, Inequality, and Poverty of GDP. Vegetables, fruits, and livestock are key sources of agricultural value added. Manufacturing value added represents 32 percent of the economy-wide total. Chem- icals, metals, and machinery lead the way, followed by sectors related to electronics, textiles, and apparel. Nonfood manufacturing is export intensive, accounting for 75 percent of Chinese exports. Electronics, textiles, apparel, chemicals, and machin- ery are the major exporting sectors. These sectors also represent a relatively large share of imports, reflecting the significant presence of the processing trade in China. Meanwhile, China's agricultural and food manufacturing sectors have limited trade exposure. Agriculture accounts for only 1.6 percent of exports and 2.5 percent of imports. Vegetables and fruits are major agricultural exports, while agricultural imports are concentrated in oilseeds, cotton, and forestry products. Simulation Design To explore the implications for the Chinese economy of agricultural distortions at home and abroad, we consider six policy reform scenarios that eliminate various distortions in global trade and in China's domestic commodity and factor mar- kets. These scenarios are summarized in table 5.2. The first two scenarios examine the effects of trade liberalization in the rest of the world. ROW-Ag considers the impact of agricultural liberalization in the form of the elimination of import tariffs and export subsidies, as well as subsidies for domestic production, in the agricultural and lightly processed food sectors in the rest of the world. The second scenario involving the rest of the world policies across all merchandise commodi- ties (ROW) looks at more broadly based trade liberalization. It combines the removal of policy distortions in the agricultural and lightly processed food sectors in scenario 1 and the elimination of tariffs on nonagricultural goods in the rest of the world. We incorporate the impacts of trade reforms in the rest of the world in the Chinese CGE model through exogenous changes in import prices and export demands. The sizes of these exogenous trade shocks are obtained from the global CGE Linkage model, omitting China's reforms in the process. Table 5.3 lists the external shocks imposed in the ROW and ROW-Ag scenarios.8 It shows that enormous percent increases in China's agricultural and food export volumes (export demand) are generated by the elimination of high rates of protection elsewhere in Asia. Rice, other grains, vegetables and fruits, and refined sugar all show large proportionate increases. Of course, the associated changes in output volume are often quite modest because China is not a large exporter of most of these products (see table 5.1). China's export volume declines in most livestock subsectors, reflecting the relatively smaller ROW barriers faced by China's exporters in these subsectors. China 157 Table 5.2. Modeled Liberalization Scenarios, China Scenario Description Agricultural liberalization in the rest of the world ROW-Ag ­ Elimination of production taxes and subsidies in the agricultural and lightly processed food sectors ­ Elimination of export taxes and subsidies in the agricultural and lightly processed food sectors ­ Elimination of import tariffs in the agricultural and lightly processed food sectors All merchandise trade liberalization in the rest of the world ROW ­ Elimination of production taxes and subsidies in the agricultural and lightly processed food sectors ­ Elimination of export taxes and subsidies in the agricultural and lightly processed food sectors ­ Elimination of import tariffs in all sectors Agricultural liberalization in China DOM-Ag ­ Elimination of export taxes and subsidies in the agricultural and lightly processed food sectors ­ Elimination of import tariffs in the agricultural and lightly processed food sectors All merchandise trade liberalization in China DOM ­ Elimination of export taxes and subsidies in the agricultural and lightly processed food sectors ­ Elimination of import tariffs in all sectors Relaxation of the hukou system LABOR ­ Cut the indirect transaction costs from 81 to 34 percent of the nonfarm rural wage Introduction of land reform LAND ­ Farm households do not include the returns to land in decisions to migrate temporarily Source: Specifications of the authors. In the case of broadbased trade reform, the average export price in China increases by 2.4 percent, while the average import price increases by only 0.6 per- cent (both relative to the price of the manufacturing exports of members of the Organisation for Economic Co-operation and Development), indicating a gain in the country's terms of trade if other countries liberalize and China does not. However, given the relatively greater importance to China of manufacturing exports, if liberalization is confined to the agricultural sector, the improvement in the terms of trade diminishes; the average increase in the export price is only 0.6 percent, and the average increase in the import price is only 0.4 percent. 158 Agricultural Price Distortions, Inequality, and Poverty Table 5.3. Exogenous Demand and Price Shocks Caused by Liberalization in the Rest of the World, China (percent change) Elimination of all trade Elimination of agricultural distortions in ROW distortions in ROW Export Export Import Export Export Import Sector, products demand prices prices demand prices prices Agriculture Paddy rice 94.9 4.2 n.a. 123.6 0.0 1.8 Wheat 15.5 3.5 2.8 45.8 1.4 3.6 Other grains 105.1 3.9 6.5 157.7 1.6 6.5 Vegetables and fruits 185.5 4.2 1.9 232.9 1.8 1.6 Oilseeds 10.3 4.0 2.8 42.9 1.7 2.3 Sugarcane and beets n.a. n.a. n.a. n.a. n.a. n.a. Plant-based fibers 30.0 3.3 10.0 51.4 1.3 11.5 Other crops 12.7 4.5 1.3 8.4 2.0 1.5 Cattle, sheep, and so on 18.6 4.4 6.5 3.1 1.9 6.6 Other livestock 20.8 3.8 0.7 0.2 1.6 1.6 Raw milk 48.3 4.1 1.8 31.7 1.7 0.7 Wool 13.1 3.8 4.9 10.1 1.6 4.9 Other primary products 7.8 2.7 0.5 2.0 0.6 1.1 Lightly processed food Meat products 29.2 3.5 4.9 56.3 1.3 5.6 Vegetable oils 6.4 1.8 0.2 5.7 0.3 0.9 Grain, milled 148.8 3.0 4.2 192.1 0.9 3.4 Sugar, refined 410.2 3.0 1.4 560.4 0.8 2.0 Highly processed food 67.3 2.9 0.8 14.1 0.8 0.2 Nonfood manufacturing Textiles; apparel and leather 13.7 2.6 0.2 2.1 0.8 0.4 Other manufacturing sectors 3.3 2.2 0.7 1.6 0.5 0.3 Services 10.5 2.5 0.1 0.9 0.5 0.2 Total 2.2 2.4 0.6 0.3 0.6 0.4 Source: Linkage model simulations by van der Mensbrugghe, Valenzuela, and Anderson (see the appendix). Note: n.a. not applicable. The increase in sectoral export prices range from 1.8 to 4.5 percent in the case of broadbased trade liberalization and from 0.3 to 2.0 percent in the case of agricultural liberalization only; food and agricultural prices rise relative to non- food prices in both cases. China 159 The changes in China's import prices show much greater sectoral variation. The import prices of most food and agricultural products rise more than the import prices of nonfood products, reflecting the elimination of agricultural sub- sidies in the countries of the Organisation for Economic Co-operation and Devel- opment. However, the world price of China's oilseed imports declines by 2.8 and 2.3 percent, respectively, in these two scenarios, largely because of the elimination of the high export taxes on soybean exports from Argentina; in the wake of the elimination of these export taxes, the soybean exports become a dominant source of oilseed imports into China. The next two scenarios focus on the impacts of China's own trade liberaliza- tion. Scenario 3 (DOM-Ag) eliminates the import tariffs and export taxes and subsidies on China's agricultural goods and lightly processed foods. In scenario 4, DOM, the tariff elimination is extended to nonagricultural sectors. These two scenarios are intended to show the effects of distortions in China's factor mar- kets. Scenario 5 (LABOR) examines the impact of a relaxation of the hukou sys- tem such that the ad valorem tax equivalent of the indirect transaction costs are reduced from 81 to 34 percent (if evaluated at current levels of migration). This is the portion of the observed differential in wages that has been directly attrib- uted to possession of a hukou certificate (see elsewhere above). In scenario 6 (LAND), we consider the impact of the relaxation of one of the important bar- riers to off-farm labor mobility, namely, the absence of well-defined property rights for agricultural land. The existence of this barrier leads to the retention of additional labor in the farm sector. The reason for this is that farm households presently tend to include the returns to communal land in their decision to work on- or off-farm because leaving the farm means potentially forgoing the rights to the farmland. Scenario 6 introduces a land reform whereby farm households migrating to cities may keep full land returns by renting their land out; they thereby only need to consider the ratio of the marginal value of the products of their labor in agriculture to their nonfarm rural wages in deciding where to work. In all six scenarios, government real spending and real savings (deflated by the GDP deflator) are fixed at the levels of the base year. Thus, the policy reforms are assumed to be revenue neutral and associated with a unified, endogenous factor income tax designed to replace lost government tariff rev- enue. The goal of this tax replacement closure is to avoid unrealistic macroeco- nomic effects of tariff removal, while exerting a relatively neutral impact on inequality. Foreign savings are also fixed in foreign currency terms, and the real exchange rate adjusts endogenously to maintain the current account balance. Total investment is endogenously adjusted according to changes in the savings levels of households and enterprises. 160 Agricultural Price Distortions, Inequality, and Poverty The Impacts of the Reforms on China's Economy In this section, we examine the impacts of the six scenarios on the macroeconomy, on poverty and inequality, and on households and sectors. Macroeconomic effects The macroeconomic outcomes of these simulations are reported in table 5.4. We begin by focusing on the two scenarios of broadbased commodity trade liberaliza- tion (ROW and DOM, reported in the first two columns). The elimination of trade distortions in all commodity sectors gives a substantial boost to trade in China; exports and imports rise by more than 5 percent in the unilateral liberal- ization scenario and by 2­4 percent in the scenario of trade liberalization in the rest of the world. Aggregate welfare effects, which we proxy using the sum of the equivalent vari- ation of individual households and a representative firm, are generally quite small, as one would expect in a model with fixed endowments, perfect competition, and constant returns to scale.9 The composite equivalent variation is projected to increase by 0.5 percent of GDP in the case of trade liberalization in the rest of the world because of improved terms of trade. In contrast, there is a small deteriora- tion in welfare under unilateral liberalization because of a deterioration in China's terms of trade. This reflects China's relatively limited import protection following the WTO accession, as well as the country's growing influence in world export markets, in which trade expansion tends to depress export prices. At fixed labor endowments and capital stocks and on the assumption of fixed unemployment and no productivity changes, real GDP changes little under both trade liberalization scenarios. The small decrease under ROW is driven by the ensu- ing labor reallocation from nonagriculture to agriculture. The stronger demand for China's agricultural exports following the elimination of trade barriers in the rest of the world diverts the labor force from high-productivity manufacturing sectors to lower-productivity agricultural subsectors.10 As a consequence, real GDP declines slightly. This contrasts with the case of China's unilateral trade liberalization, whereby the elimination of the relatively higher import protection in the agricul- tural sector encourages the movement of the labor force from the rural agricultural sector to urban nonagricultural activities, leading to an increase in GDP. As the bottom section of table 5.4 indicates, temporary migration from the rural to urban sectors slows down as a result of the trade liberalization in the rest of the world; this boosts the economic prospects in agriculture. Under ROW, there are about 5.9 million fewer rural-urban migrants in the new equilibrium relative to the base year. The larger rural labor force is mainly generated by the retention China 161 Table 5.4. Aggregate Simulation Results of Prospective Liberalizations, China Indicator ROW DOM ROW-Ag DOM-Ag LABOR LAND Macroeconomy, % change Welfare, equivalent variation 0.5 0.1 0.04 0.01 1.0 0.1 Real GDP 0.1 0.2 0.2 0.1 0.8 0.3 Exports 1.9 5.8 0.3 0.7 1.6 0.6 Imports 4.3 5.5 0.1 0.7 1.4 0.6 Terms of trade 1.8 0.8 0.3 0.1 0.3 0.1 Consumer price index 2.9 0.9 1.0 0.3 1.4 0.7 Factor prices, % Return to agricultural land 16.3 3.5 13.5 3.1 7.3 2.5 Return to capital 2.2 0.8 0.0 0.0 1.5 0.6 Unskilled wages Urban 3.7 1.1 1.2 0.3 17.7 3.1 Rural nonagricultural 3.9 1.3 1.3 0.4 6.9 3.9 Agricultural 4.4 1.8 1.3 0.4 23.7 8.8 Semiskilled wages Urban 3.9 1.2 1.3 0.3 5.4 3.1 Rural nonagricultural 4.9 1.1 2.2 0.4 25.5 4.5 Agricultural 2.7 1.1 0.0 0.0 20.1 11.7 Skilled wages Urban 1.9 0.9 0.0 0.0 0.9 0.3 Rural nonagricultural 1.9 1.0 0.1 0.0 0.9 0.2 Labor force, millions Farm labor 6.4 1.6 5.7 1.5 27.9 13.2 Unskilled 0.7 0.2 0.6 0.2 15.6 1.8 Semiskilled 5.7 1.4 5.1 1.3 12.3 11.3 Rural-urban temporary migration 5.9 1.5 5.3 1.4 35.7 12.1 Unskilled 0.6 0.1 0.5 0.1 18.2 1.5 Semiskilled 5.3 1.3 4.8 1.3 17.6 10.6 Skilled 0.0 0.0 0.0 0.0 0.0 0.0 Labor force, % Farm labor 1.7 0.4 1.6 0.4 7.6 3.6 Unskilled 0.4 0.1 0.4 0.1 9.8 1.2 Semiskilled 2.7 0.7 2.4 0.6 5.9 5.4 Rural-urban temporary migration 6.0 1.5 5.4 1.4 36.5 12.3 Unskilled 1.5 0.4 1.4 0.4 46.7 3.9 Semiskilled 10.4 2.6 9.3 2.5 34.3 20.6 Skilled 0.0 0.0 0.0 0.0 0.0 0.0 Source: CGE model simulations of the authors. 162 Agricultural Price Distortions, Inequality, and Poverty of additional on-farm labor (increased by 6.4 million) in the ROW scenario. In contrast, China's unilateral trade liberalization accelerates off-farm migration; about 1.5 million to 1.6 million workers would leave agriculture and migrate to urban areas (relative to the baseline). Table 5.4 also reports changes in factor prices. It is clear from these that trade liberalization in the rest of the world favors unskilled and semiskilled labor over skilled labor and rural labor over urban labor. This is caused by the relative increase in the demand for agricultural exports and the rise in the prices of com- peting agricultural imports. Agricultural profitability in China is also boosted by the trade reforms in the rest of the world; this is reflected in the rise in the returns to agricultural land under ROW. The returns to capital and skilled wages increase less than the consumer price index, which rises by 2.9 percent under ROW. This pattern of changes in factor prices contrasts sharply with the pattern emerging after China's unilateral liberalization, whereby the returns to capital and skilled wages increase the most relative to the consumer price index, while the returns to agricultural land decrease the most. In the macroeconomic results reported in table 5.4, columns 3 and 4, namely, the results stemming from the liberalization of the agricultural and lightly processed food sectors only, we see that agricultural liberalization has only modest impacts on aggregate exports and imports, reflecting the minor role of the agri- cultural and food sectors in China's total trade (see table 5.1). Consequently, China's welfare gains from agricultural liberalization are trivial, ranging from 0.01 to 0.04 percent of GDP. In contrast to unilateral trade liberalization in all sectors, China's unilateral agricultural liberalization leads to an aggregate gain in equiva- lent variation of 0.01 percent of GDP, mainly because of much smaller losses in the terms of trade. The changes in factor prices induced by agricultural liberaliza- tion show patterns that are similar to the patterns resulting from broadbased commodity trade liberalization, that is, agricultural liberalization in the rest of the world would favor unskilled and semiskilled labor, as well as agricultural land, in China, while China's unilateral agricultural reforms would favor the capital and skilled labor that are intensively employed in the relatively lightly protected man- ufacturing sectors. The changes in off-farm employment and rural-urban migra- tion under the two agricultural liberalization scenarios are comparable to the cor- responding changes in the broadbased trade liberalization scenarios, indicating the dominant role played by distortions in the agricultural subsectors in deter- mining the mobility of the rural labor force in China. Relative to the reduction in China's trade distortions, the labor market reforms investigated in scenarios LABOR and LAND generally have larger impacts on wel- fare, GDP, and other macroeconomic aggregates. This reflects the large and per- sistent rural-urban distortions in China's labor markets. It is evident from the last China 163 two columns of table 5.4 that both the factor market reforms serve to increase migration out of the relatively low-productivity agricultural sector into the higher-productivity nonagricultural sectors and from the rural to the urban econ- omy. In the case of land reform, 13.2 million additional workers leave agriculture after they have been assured that they will retain ownership of their land even if they migrate (see the farm labor row in table 5.4, final column). These individuals initially migrate to the off-farm rural labor market, which releases an additional 12.1 million temporary rural migrants to the urban sector to restore equality in rural and urban wages, net of transaction costs. The release of workers from agri- culture tends to depress wages in the rural nonfarm economy, where wages fall by 3.9 percent in the case of land reform. This wage drop plays a role in dampening the out-migration from agriculture. While the LAND reform scenario focuses on the barriers to the off-farm mobility of labor, the LABOR scenario focuses on rural-urban migration. If the transaction costs associated with temporary migration are reduced because of the elimination of the hukou system, rural-urban migration expands by 35.7 million workers. Because the transaction costs associated with temporary rural-urban migration operate, in effect, as a tax on rural labor, the first impact of their reduc- tion is to increase the supply of rural labor to the urban economy, thereby boost- ing rural wages and depressing urban wages. This represents a redistribution of the rents associated with the hukou system from urban to rural households. In addition, by raising rural wages, this hukou reform scenario also draws 27.9 million additional workers out of agriculture. Poverty and inequality impacts Because poverty and income distribution are central to our study, we provide sev- eral related measures of inequality and poverty in table 5.5. The first column in the table reports the initial level of each indicator in our database, while subse- quent columns report changes or percentage changes in the indicators. The initial urban-rural income ratio, at 3.5, is higher than the ratios in some of the studies cited elsewhere above that are based on household surveys. This is largely because of our inability to adjust for spatial price variations which, if taken fully into account, would reduce the ratio considerably. The initial Gini coefficient in our model, 0.442, is heavily influenced by rural-urban income disparity. This estimate is also consistent with the recent work of Benjamin et al. (2007), who identify limitations with many of the existing estimates of inequality and place the Gini in the 0.4­0.5 range. Using the US$2-a-day poverty line and the 1993 purchasing power parity exchange rate, the World Bank estimates that 58.1 percent of the rural population 164 Table 5.5. The Effects of Prospective Liberalizations on Income Inequality and Poverty, China Indicator Base ROW DOM ROW-Ag DOM-Ag LABOR LAND Inequality Urban/rural income ratio 3.538 0.052 0.009 0.042 0.010 0.303 0.167 Gini coefficient 0.442 0.005 0.001 0.004 0.001 0.021 0.008 Urban 0.259 0.000 0.000 0.001 0.000 0.006 0.003 Rural 0.315 0.002 0.000 0.002 0.000 0.008 0.003 Poverty headcount, US$2 a day Ratio, % Changes, percentage points Total 36.4 1.2 0.3 0.8 0.2 4.1 1.0 Urban 2.5 0.1 0.0 0.0 0.0 0.7 0.3 Transfer specialized 0.5 0.0 0.0 0.0 0.0 0.0 0.0 Labor specialized 4.0 0.1 0.0 0.0 0.0 0.9 0.4 Diversified 1.6 0.0 0.0 0.0 0.0 0.6 0.3 Rural 58.1 1.9 0.5 1.4 0.3 7.1 1.9 Agriculture specialized 63.6 1.8 0.4 1.4 0.3 6.8 3.5 Diversified 57.5 1.9 0.5 1.4 0.3 7.1 1.7 Poverty headcount, US$2 a day Millions Changes, millions Total 467.3 14.9 3.6 10.8 2.3 52.1 13.4 Urban 12.6 0.3 0.1 0.0 0.0 3.3 1.6 Transfer specialized 0.1 0.0 0.0 0.0 0.0 0.0 0.0 Labor specialized 8.1 0.2 0.1 0.0 0.0 1.7 0.9 Diversified 4.4 0.1 0.0 0.0 0.0 1.6 0.8 Rural 454.7 14.6 3.5 10.8 2.4 55.5 15.0 Agriculture specialized 52.2 1.5 0.4 1.1 0.3 5.6 2.9 Diversified 402.5 13.1 3.2 9.7 2.1 49.9 12.1 Source: CGE model simulations of the authors. China 165 and 2.5 percent of the urban population were living in poverty in China in 2004.11 We start with these target rates of poverty and compute the poverty line in our data set that reproduces this same poverty headcount. This yields a poverty line of Y 3,520 per person for urban areas and Y 2,591 per person for rural areas. By assuming a uniform distribution of the population within each of the income vingtiles in our source data from the National Bureau of Statistics of China (NBS 2006), we are able to estimate the poverty headcounts in each stratum. This infor- mation is also reported in table 5.5. As one may see there, the national poverty sit- uation in China is largely driven by rural poverty; 455 million poor reside in rural areas. The poverty headcount rate is highest in the agriculture-dependent house- hold stratum, in which nearly two-thirds (63.7 percent) of the population is poor. The two reform scenarios that do not reduce the rural-urban income disparity are DOM and DOM-Ag (China's unilateral liberalization) because rural house- holds generally lose if agricultural factor returns decline. Although the magnitude of the change in the rural-urban income ratio is small in the cases of trade liberal- ization, it is substantial in the factor market reform scenarios. In the case where the hukou registration system is abolished (LABOR), for example, the ratio declines from 3.54 to 3.23. The decline in the land reform scenario (LAND) is likewise large, at 0.17 points. Table 5.5 also reports the absolute changes in Gini coefficients. Because income inequality in China is dominated by urban-rural inequality, the narrower urban- rural income gaps in the scenarios of trade liberalization in the rest of the world and of reforms in factor markets are reflected in an improvement in overall inequality, as measured in the national Gini coefficient. There are no discernible changes in inequality within urban and rural areas in the unilateral liberalization scenarios. However, in the two factor market reform scenarios, the Gini coefficients show a slight increase in inequality within urban areas and a slight decline within rural areas. This is because low-income urban households dependent on the earnings from unskilled labor are hurt most by the increase in the rural-urban migration of unskilled workers after labor market reforms, whereas low-income, diversified rural households gain more from the resulting increase in rural unskilled wages than do households at high-income levels. In the scenario of broadbased trade liberalization in the rest of the world (ROW), the monetary poverty line increases by 2.9 percent after the change in the consumer price index (table 5.4). Nonetheless, higher factor earnings mean that the poverty headcount ratio declines among all household groups living in significant poverty. Urban poverty decreases by 0.3 million people (the bottom panel reports changes in poverty in millions), while rural households enjoy a 1.9 percentage point reduction in the poverty headcount (the middle panel reports percentage point changes in poverty). Because of the large population base in rural China, this translates into a rural poverty reduction of 14.6 million people. In the case of agricultural liberalization 166 Agricultural Price Distortions, Inequality, and Poverty in the rest of the world, the poverty reduction is smaller, but still significant; the rural poverty headcount declines by 10.8 million. Given the adverse impacts on agricultural subsectors after China's unilateral liber- alization, rural poverty increases slightly, by 3.5 million people, in the DOM scenario and by 2.4 million in the DOM-Ag scenario. Because of the predominance of rural poverty in China, these increases in rural poverty translate into comparable changes in total poverty. However, labor market reforms would significantly reduce rural poverty, but slightly increase urban poverty. The rural poverty headcount ratio declines from 58.1 percent in the base year to 51.0 percent in the LABOR scenario and 56.2 percent in the LAND scenario, while the urban headcount ratio rises slightly, from 2.5 to 2.8­3.2 percent. Overall, the share of the national population that is impoverished falls quite sharply in the case of hukou reforms, from 36.4 percent of the total population to 32.3 percent in the LABOR scenario (hukou reforms) and to 35.4 percent in the LAND scenario. If combined, these two scenarios, together, gener- ate a poverty reduction of 65.5 million. It is thus clear that, if poverty reduction and greater income equality are the objectives of the next round of reforms in China, then factor market reforms should be part of the package. Household impacts It is important to dig below the aggregate indicators of poverty and inequality and consider the disaggregated household incidence curves reported in figure 5.1, charts a-f. These show the percentage changes in welfare (the equivalent variation as a per- centage of initial income) by stratum across the income vingtile spectrum. The largest increases in welfare after both trade and agricultural liberalization in ROW (on the order of 2 percent) accrue to agriculture-specialized households (figure 5.1, charts a and b). These households benefit from the fact that the returns to agricul- tural land increase relative to other factor prices. Real incomes rise less among rural diversified households because of the dominance of nonfarm wage earnings in the income portfolios of these households. Among urban households, the largest wel- fare increases in chart a of figure 5.1 are associated with labor-specialized house- holds, followed by urban diversified households. This is consistent with the larger increases in wage rates relative to the returns to capital. Because the transfers are constant in real terms and because transfers make up most of the incomes of these households, the real incomes of the transfer group are little affected by the agricul- tural reform. However, because the share of food consumption in the total expendi- ture of households at low income levels tend to be larger, the relative increase in food prices leads to a higher household-specific consumer price index for these low- income households relative to the national average consumer price index, causing the modest welfare losses among the lowest income transfer groups. China 167 Figure 5.1. The Impacts of Prospective Liberalizations on Welfare of Five Types of Households, China a. All goods liberalization in the rest of the world (ROW) 3.5 equivalent variation as a % of income 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.5 1.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 vingtile urban transfer specialized rural diversified urban diversified urban labor specialized rural agriculture specialized b. Agricultural liberalization in the rest of the world (ROW-ag) 3.0 equivalent variation as a % of income 2.5 2.0 1.5 1.0 0.5 0.0 0.5 1.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 vingtile urban transfer specialized rural diversified urban diversified urban labor specialized rural agriculture specialized (Figure continues on the following page.) 168 Agricultural Price Distortions, Inequality, and Poverty Figure 5.1. (continued) c. Unilateral liberalization of all goods trade (TRA) 0.6 equivalent variation as a % of income 0.4 0.2 0.0 0.2 0.4 0.6 0.8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 vingtile urban transfer specialized rural diversified urban diversified urban labor specialized rural agriculture specialized d. Unilateral liberalization of agricultural trade (TRA-ag) 0.3 equivalent variation as a % of income 0.2 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 vingtile urban transfer specialized rural diversified urban diversified urban labor specialized rural agriculture specialized China 169 Figure 5.1. (continued) e. Hukou reform (LABOR) 15.0 equivalent variation as a % of income 10.0 5.0 0.0 5.0 10.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 vingtile urban transfer specialized rural diversified urban diversified urban labor specialized rural agriculture specialized f. Land reform (LAND) 8.0 equivalent variation as a % of income 6.0 4.0 2.0 0.0 2.0 4.0 6.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 vingtile urban transfer specialized rural diversified urban diversified urban labor specialized rural agriculture specialized Source: Author simulations. 170 Agricultural Price Distortions, Inequality, and Poverty China's unilateral trade liberalization hurts all households except urban transfer- specialized households (figure 5.1, chart c), although the magnitude of the welfare losses is small. Rural agriculture-specialized households experience the largest welfare losses, followed by rural diversified households, which suffer from the depressed returns to agricultural activity. The welfare losses of urban households are small, amounting to only around 0.1­0.2 percent of household income among diversified households and labor-specialized households. The increased income tax rate required to replace the loss in tariff revenue is the major factor contribut- ing to the welfare losses among urban households. In the case of China's unilateral agricultural liberalization, rural households are still the major losers, but all urban households gain slightly because of the smaller tax replacement effects associated with the less serious loss in tariff revenue (figure 5.1, chart d). Recall from the preceding discussion that the largest poverty and inequality impacts stem from the hukou reform. Chart e, figure 5.1 shows why this is true. The population stratum showing the highest poverty headcount, the agriculture- specialized households, is also the one reaping the largest proportionate gains in this labor market reform scenario. The stratum benefits from the significant increase in agricultural wages. Diversified rural households also benefit from the rise in rural wages, although their welfare gains are somewhat less. These households supply less of the temporary migrant labor to urban areas. These migrants bear the direct bur- den of the heightened transaction costs associated with the hukou system; so, if this burden is eliminated, the migrants benefit the most. While the benefits from hukou reform are spread relatively evenly across income levels within each of the rural strata, the higher-income households within the diversified strata, which have more capital earnings in their incomes, tend to experience smaller proportionate gains; this contributes to the decline in the Gini coefficient within the rural sector. Most urban households suffer from the influx of unskilled and semiskilled rural migrants, whose presence drags down the wage rates in urban areas. Almost all urban households experience welfare losses, with the minor exception of the richest transfer-specialized households. Overall, the urban index of income inequality worsens slightly. However, this is overwhelmed by the reduction in between-sector, rural-urban inequality; if coupled with the decline in rural inequality, this leads to a decline in the national Gini coefficient of inequality of 0.021 (from 0.442 to 0.421). This is a substantial movement in a coefficient that is generally quite robust in the face of policy reforms. Similar to the case of the hukou reform, the largest gains from land reform accrue to agriculture-specialized rural households (figure 5.1, chart f). These households are currently constrained to remain active on the farm if they wish to retain the rights to their land. By permitting some of these households to rent the land and migrate to the city if they wish, land market reform raises the shadow value of the labor remain- ing in agriculture substantially across all income levels. Diversified rural households China 171 also gain; some of the highest gains come at the lowest income levels, where house- holds are more heavily reliant on incomes from agriculture. Urban household welfare falls across the board in this experiment, and it falls most among the poorest house- holds. This is because of the large boost in rural-urban migration among unskilled and semiskilled labor, as well as the increase in food prices following the reduction in the agricultural labor force. As a consequence, the urban Gini coefficient rises. How- ever, from the point of view of overall inequality in China, the main consequence of this experiment is to redistribute income from urban to rural households, which lowers the Gini coefficient by 0.008. Sectoral impacts In table 5.6, we report the effects of the policy reforms in the six aggregated sce- narios we have investigated on output, exports, and imports. The first row of the table shows that highly processed food products are the major winners from the Table 5.6. The Effects of Prospective Liberalizations on Sectoral Outputs and Trade, China (percent) Subsector ROW DOM ROW-Ag DOM-Ag LABOR LAND Output Agriculture 1.7 0.3 1.6 0.4 2.9 1.3 Other primary goods 0.8 0.1 0.2 0.1 0.8 0.5 Lightly processed food 1.7 2.0 1.7 2.2 2.6 1.1 Highly processed food 5.4 0.03 1.6 0.44 2.3 1.1 Nonfood manufacturing 0.5 0.4 0.9 0.3 2.0 0.9 Services 0.4 0.2 0.2 0.1 1.3 0.5 Exports Agriculture 71.5 5.7 100.2 3.0 39.2 23.6 Other primary goods 6.0 5.6 1.7 0.6 3.7 1.3 Lightly processed food 31.2 11.0 58.3 7.7 25.0 14.4 Highly processed food 64.8 7.0 14.8 2.8 17.4 9.6 Nonfood manufacturing 1.6 6.4 2.1 0.7 3.0 1.4 Services 10.2 2.7 1.2 0.3 2.1 1.1 Imports Agriculture 11.6 21.0 1.2 22.1 18.6 9.8 Other primary goods 5.5 1.2 2.4 0.1 6.9 3.0 Lightly processed food 8.8 46.4 0.1 48.6 10.3 5.8 Highly processed food 8.8 16.3 3.5 0.9 8.0 4.3 Nonfood manufacturing 3.9 5.7 0.2 0.0 0.5 0.1 Services 4.8 1.2 0.3 0.1 0.4 0.1 Source: CGE model simulations of the authors. 172 Agricultural Price Distortions, Inequality, and Poverty elimination of market distortions in the rest of the world, with an average output expansion of 5.4 percent. Production in the agriculture and lightly processed food sectors expands by 1.7 percent. Substantial increases in exports are the key drivers of the expansion in China's food and agricultural sectors; this flows from the significant increase in international demand. Exports of agricultural products, lightly processed food, and highly processed food increase by 72, 31, and 65 per- cent, respectively, in the ROW scenario. Despite the absence of any decline in pro- tection for agriculture in the ROW scenario, China's agricultural and food imports increase by around 10 percent following agricultural liberalization in the rest of the world because of the decline in the world prices of some of China's major agri- cultural imports such as oilseeds and vegetable oils. In addition, there is a real appreciation in China's currency that tends to boost the demand for imports across the board. If ROW liberalization is confined to the agricultural and lightly processed food sectors, they become the only two aggregate sectors with expanding exports and output. All the other aggregate sectors experience declining output and exports. The impact of agricultural liberalization in the rest of the world on China's imports is modest in comparison with the impact of broadbased trade liberaliza- tion because the decline in total exports and a depreciation in the real exchange rate both serve to dampen the expansion of imports in the ROW-Ag scenario. The sectoral impacts of China's own reforms suggest that the current distor- tions arising from China's tariff protection and labor market barriers generally support the size of agriculture relative to the size of other industries. Under all four scenarios involving China's own reforms, agriculture experiences output losses, while the nonfood manufacturing sectors and services expand. The impact of reducing China's distortions in commodity and factor markets on highly processed food is mixed: this sector benefits from the elimination of import tariffs, but loses from the reforms in factor markets. In the two scenarios involving liberalization in the rest of the world, the disag- gregated changes in sectoral output (not shown) generally follow the changes in export demand reported in table 5.3: the sectors with larger increases (decreases) in export demand and higher export dependence, such as prepared fish products, sugar, textiles, and apparel and leather, experience relatively large increases (decreases) in output. However, imports also play a role in determining changes in sectoral output in oilseeds, in which output shrinks by 7.4 percent in the ROW scenario as a result of the 2.8 percent decline in import prices, which spurs the growth of oilseed imports. Under China's unilateral trade liberalization (DOM), instruments, electronics, textiles, apparel and leather are major manufacturing sectors with rapid output expansion. As the most highly export-oriented sectors, they benefit from the real China 173 depreciation in the Chinese currency in the wake of the country's unilateral trade liberalization. At the other end of the spectrum, the most heavily protected sec- tors, with sizable trade exposure, experience declining output, including oilseeds, sugar, transportation equipment, other grains, and vegetable oil. In the case of China's unilateral agricultural liberalization, the agricultural subsectors with high levels of protection experience significant contractions in output. In the scenarios of hukou reform (reduced transaction costs) and land reform, agricultural output falls sharply because the farm labor force is diverted to off- farm rural activities, as well as urban-based manufacturing. Within manufactur- ing, the consumption goods subsectors experience declining output, but most capital goods sectors expand because the changes in final demand favor invest- ment over consumption in these two scenarios. The sensitivity of the results to the land rental market assumption China's rural land markets have been undergoing reform, and a nascent market for land is emerging in many areas (see elsewhere above). In principle, this should facilitate off-farm migration because migrants may no longer risk losing control of their land if they leave their farms. However, to date, these reforms have been restricted to certain regions, and it is not clear how efficiently this market is functioning even in those special cases. Therefore, in our base case results, we have assumed that the transfer of rural labor from farm activities to off-farm activities would diminish the earnings from land rents by 50 percent, on average (that is, there is a 50 percent probability that migrants will lose control of their land). Since this parameter choice is somewhat arbitrary, we contrast the base case results with the results arising from the two extreme assumptions about the functioning of the rural land rental market: one in which there is a zero loss in land returns following the onset of off-farm employment and the other in which there is no land rental market (representing a 100 percent loss in land returns if farmers switch to off-farm jobs). We then repeat the two trade liberal- ization scenarios, ROW and DOM. The key simulation results are presented in table 5.7. Since the macro aggregate results are essentially unchanged from our base line results, only revised results on factor prices and labor migration are reported. The first pair of columns in table 5.7 report the results if the land market is entirely absent so that migration produces the loss of all land farmed by the migrants. The second pair of columns reports the results under the opposite assumption, namely, the existence of a fully functioning land market. Consider, first, the case of trade reforms in the rest of the world (ROW). In this case, the returns to land and the wage rates in agriculture rise. Furthermore, the rise in 174 Agricultural Price Distortions, Inequality, and Poverty Table 5.7. Sensitivity Analysis of the Simulation Results, China No land rental market Fully functioning land market Indicator ROW DOM ROW DOM Factor prices, % Return to agricultural land 16.8 3.6 15.8 3.4 Return to capital 2.1 0.7 2.2 0.8 Unskilled wages Urban 4.0 1.1 3.4 1.0 Rural nonagricultural 4.3 1.3 3.5 1.3 Agricultural 3.5 1.3 5.3 1.8 Semiskilled wages Urban 4.4 1.3 3.5 1.1 Rural nonagricultural 5.6 1.1 4.1 1.1 Agricultural 1.5 0.8 4.2 1.5 Skilled wages Urban 1.8 0.9 1.9 0.9 Rural nonagricultural 1.7 1.0 1.9 1.0 Labor force, millions Farm labor 8.0 1.9 4.7 1.3 Unskilled 0.8 0.2 0.5 0.1 Semiskilled 7.2 1.7 4.2 1.1 Rural-urban temporary migration 7.4 1.7 4.4 1.2 Unskilled 0.7 0.2 0.4 0.1 Semiskilled 6.7 1.6 4.0 1.1 Skilled 0.0 0.0 0.0 0.0 Source: CGE model simulations of the authors. the returns to land is greater than the rise in wages. Therefore, households that had hitherto been considering leaving agriculture because of depressed factor returns have an even stronger incentive to continue to devote their labor to agriculture than do those households that, at the margin, had been indifferent to the wage differential between the farm and nonfarm sectors (the fully func- tioning land market). Thus, the movement of labor into agriculture in the ROW scenario is greater if the land market is not functioning rather than functioning, provided there is no change in the underlying structure of the land market. The same situation applies, but in reverse, in the case of unilateral trade reforms if labor is leaving agriculture. Here, by including the returns to land in the off-farm migration decision (because these fall more than the fall in wages), the incentive to work off-farm is accentuated in the absence of a land market. As China 175 we see above, the shift from the lack of a land market to a fully functioning land market (the LAND experiment) generates a much larger flow of workers from agriculture to the rest of the economy (more than 13 million), as well as a signifi- cant reduction in poverty. So, we are not concluding that a poorly functioning land market is good for poverty reduction. Because the impact of land reforms dominates the impacts of trade reform on labor markets, the former rules the day if both are undertaken together. However, this sensitivity analysis does show that our predictions about the impact of trade reforms on intersectoral labor mobility depend importantly on the extent to which farmers are able to lease their land if they are migrating to the city for work. Conclusions and Policy Implications Absolute poverty in China is now largely a rural problem, and, within the rural sector, the intensity of poverty is greatest on the farm. Thus, policy reforms that either boost the returns to farming or enhance the off-farm opportunities for those individuals presently working in agriculture offer the best prospects for reducing poverty and inequality in China. Of the reforms considered, trade reforms in the rest of the world, land reform, and hukou reform all reduce poverty in China, while unilateral trade reforms result in a small increase in poverty. Domestic agricultural distortions are important factors in determining the distri- butional and poverty effects of trade reform packages, although the impacts of the reforms on aggregate trade and welfare are small. Furthermore, the ROW trade reforms, as well as the land and hukou reforms tend to favor rural households over urban households, while the opposite is true of the unilateral tariff reforms. So, it would seem desirable to bundle these reforms together in such a way that all these broad household groups stand to benefit from the reform package. For example, by combining the ROW and domestic trade reforms, a policy package would reduce poverty and inequality, while benefitting all the household groups in our study. The land reform and the hukou reform scenarios benefit rural areas much more significantly than urban ones. In the case of land reform, the changes may hurt the lower-income urban households that currently benefit from the artificial restriction of rural-urban labor mobility. This outcome may be avoided, however, if the reforms are phased in over time; in the context of continued rapid economic growth in the urban and coastal regions in China, the urban losses are likely to be more than offset by ongoing income growth. Indeed, this is what appears to be happening in many regions of China in which the restrictions on labor mobility are being eroded, and land markets are emerging. This study suggests that such labor and land market reforms are particularly impressive in their potential for 176 Agricultural Price Distortions, Inequality, and Poverty reducing inequality and rural poverty in China, as well as their scope for allowing China to realize more effectively the potential of the vast rural labor force. Notes 1. Estimates of agricultural protection and assistance for China based on Huang et al. (2009) are incorporated in the World Bank's Distortions to Agricultural Incentives Project Database (Anderson and Valenzuela 2008). The former cover five decades, but the representative values for developing- country agriculture as of 2004 that are used in the global CGE modeling for this study are summarized in Valenzuela and Anderson (2008). 2. However, as noted by Parish, Zhe, and Li (1995), the rural labor market is looking more and more like an open market. 3. See Hertel and Zhai (2006) for details on the behavior of off-farm labor supply in the model. 4. Given the small number of skilled farm workers in China, we ignore this group in our analysis. See Zhai and Hertel (2006) for details of the derivation of the off-farm labor supply elasticity. 5. There are likely other, unobserved factors inducing this rural-urban wage differential, in which case, the estimation of the labor market distortion via the subtraction of known factors is biased in the direction of overstating the hukou-related distortion. Therefore, it is useful also to estimate the direct impact of household registration status on the observed wage difference among households. Shi (2002) takes this approach to the problem. Using the same China Health and Nutrition Survey data set, he finds that only 28 percent of the rural-urban wage difference may be explained directly via the coefficient on the hukou registration variable. This is quite a bit less than the 48 percent left unexplained by the subtraction approach of Shi, Sicular, and Zhao (2002). 6. See Hertel and Zhai (2006) for a detailed description of the way this ad valorem distortion is obtained. 7. We assume that a doubling in the number of temporary migrants would only increase the marginal cost of migration by 10 percent. 8. Zhai and Hertel (2009), appendix table A.1 provides the mapping between 48 sectors in the Chinese CGE model and 23 sectors in the Linkage model. 9. The equivalent variation of the representative firm is calculated based on the firm's utility derived from the part of investment financed by the firm's retained earnings. 10. In reality, this is likely to show up in slower rates of out-migration from agriculture. 11. The World Bank's poverty estimates are available at http://go.worldbank.org/NT2A1XUWP0. References Anderson, K., and E. Valenzuela. 2008. "Estimates of Global Distortions to Agricultural Incentives, 1955­2007." Data spreadsheet, October, World Bank, Washington, DC. http://go.worldbank.org/ YAO39F35E0. Benjamin, D., L. Brandt, J. Giles, and S. Wang. 2007. "Inequality and Poverty in China during Reform." PEP-PMMA Working Paper 2007-07 (March), Poverty Monitoring, Measurement, and Analysis Program, Poverty and Economic Policy Research Network, Université Laval, Quebec. Bhattasali, D., S. Li, and W. Martin. 2004. China and the WTO: Accession, Policy Reform, and Poverty Reduction Strategies. London: Oxford University Press; Washington, DC: World Bank. Cai, F., A. Park, and Y. Zhao. 2007. "The Chinese Labor Market in the Reform Era." In China's Great Eco- nomic Transformation, ed. L. Brandt and T. G. Rawski, 167­214. New York: Cambridge University Press. Chen, S., and M. Ravallion. 2004. "Welfare Impacts of China's Accession to the WTO." In China and the WTO: Accession, Policy Reform, and Poverty Reduction Strategies, ed. D. Bhattasali, S. Li, and W. Martin, 261­82. London: Oxford University Press; Washington, DC: World Bank. China 177 Deininger, K., and S. Jin. 2005. "The Potential of Land Rental Markets in the Process of Economic Development: Evidence from China." Journal of Development Economics 78 (1): 241­70. ------. 2007. "Land Rental Markets in the Process of Rural Structural Transformation: Productivity and Equity Impacts in China." Policy Research Working Paper 4454, World Bank, Washington, DC. Heckman, J. J., L. Lochner, and C. Taber. 1998. "Explaining Rising Wage Inequality: Explanations with a Dynamic General Equilibrium Model of Labor Earnings with Heterogeneous Agents." Review of Economic Dynamics 1 (1): 1­58. Hertel, T. W., ed. 1997. Global Trade Analysis: Modeling and Applications. New York: Cambridge University Press. Hertel, T. W., P. V. Preckel, J. Cranfield, and M. Ivanic. 2004. "Poverty Impacts of Multilateral Trade Liberalization." World Bank Economic Review 18 (2): 205­36. Hertel, T. W., and F. Zhai. 2006. "Labor Market Distortions, Rural-Urban Inequality, and the Opening of China's Economy." Economic Modeling 23 (1): 76­109. Hertel, T. W., F. Zhai, and Z. Wang. 2004. "Implications of WTO Accession for Poverty in China." In China and the WTO: Accession, Policy Reform, and Poverty Reduction Strategies, ed. D. Bhattasali, S. Li, and W. Martin, 283­303. London: Oxford University Press; Washington, DC: World Bank. Huang, J., S. Rozelle, W. Martin, and Y. Liu. 2009. "China." In Distortions to Agricultural Incentives in Asia, ed. K. Anderson and W. Martin, 117­61. Washington, DC: World Bank. Katz, L. F., and K. M. Murphy. 1992. "Changes in Relative Wages, 1963­1987: Supply and Demand Factors." Quarterly Journal of Economics 107 (1): 35­78. Liu, X., A. Park, and Y. Zhao. 2006. "Determinants of Labor Migration: Recent Evidence from Rural China." Unpublished paper, Beijing University, Beijing. NBS (National Bureau of Statistics of China). 2006. Input-Output Table of China 2002. Beijing: China Statistics Press. Parish, W. L., X. Zhe, and F. Li. 1995. "Nonfarm Work and Marketization of the Chinese Countryside." China Quarterly 143 (September): 697­730. Robinson, S., A. Cattaneo, and M. El-Said. 2001. "Updating and Estimating a Social Accounting Matrix Using Cross Entropy Methods." Economic Systems Research 13 (1): 47­64. Shi, X. 2002. "Empirical Research on Urban-Rural Income Differentials: The Case of China." Unpub- lished paper, China Center for Economic Research, Beijing University, Beijing. Shi, X., T. Sicular, and Y. Zhao. 2002. "Analyzing Urban-Rural Income Inequality in China." Paper pre- sented at "International Symposium on Equity and Social Justice in Transitional China," Beijing, July 11­12. Sicular, T., and Y. Zhao. 2004. "Earnings and Labor Mobility in Rural China: Implications for China's Accession to the WTO." In China and the WTO: Accession, Policy Reform, and Poverty Reduction Strategies, ed. D. Bhattasali, S. Li, and W. Martin, 239­60. London: Oxford University Press; Washington, DC: World Bank. Sicular, T., Y. Ximing, B. Gustafsson, and S. Li. 2007. "The Urban-Rural Income Gap and Inequality in China." Review of Income and Wealth 53 (1): 93­126. Valenzuela, E., and K. Anderson. 2008. "Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980­84." Research Memorandum 13 (December), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925. van der Mensbrugghe, D. 2005. "Linkage Technical Reference Document: Version 6.0." December, World Bank, Washington, DC. http://go.worldbank.org/7NP2KK1OH0. Wang, X., T. Herzfeld, and T. Glauben. 2007. "Labor Allocation in Transition: Evidence from Chinese Rural Households." China Economic Review 18 (3): 287­308. Yang, D. T. 1997. "China's Land Arrangements and Rural Labor Mobility." China Economic Review 8 (2): 101­16. Zhai, F., and T. W. Hertel. 2006. "Impacts of the Doha Development Agenda on China: The Role of Labor Markets and Complementary Education Reforms." In Poverty and the WTO: Impacts of 178 Agricultural Price Distortions, Inequality, and Poverty the Doha Development Agenda, ed. T. W. Hertel and L. A. Winters, 285­318. London: Palgrave Macmillan; Washington, DC: World Bank. ------. 2009. "Economic and Poverty Impacts of Agricultural, Trade, and Factor Market Reforms in China." Agricultural Distortions Working Paper 98, World Bank, Washington, DC. Zhang, L., J. Huang, and S. Rozelle. 2002. "Employment, Emerging Labor Markets, and the Role of Education in Rural China." China Economic Review 13 (2­3): 313­28. Zhao, Y. 1999a. "Leaving the Countryside: Rural-to-Urban Migration Decisions in China." American Economic Review 89 (2): 281­86. ------. 1999b. "Labor Migration and Earnings Differences: The Case of Rural China." Economic Development and Cultural Change 47 (4): 767­82. 6 Indonesia Peter Warr* Since the independence of Indonesia, the country's trade policies have taxed agri- culture relative to manufacturing. However, since around 2000, the net impact of the trade policies has been roughly neutral between these broad sectors. The reversal occurred immediately following the Asian financial crisis of the late 1990s. It took the form of increases in the protection of the import-competing commodities sugar and rice, declines in the taxation of agricultural exports, espe- cially rubber and copra, and declines in manufacturing protection. The shift toward a more democratic form of government has weakened the influence of Indonesia's technocrats, who have generally favored liberal trade policies. The greater protection of some key agricultural commodities has been a consequence. The protection of agriculture has primarily involved import restrictions in the import-competing sugar and rice sectors. Other agricultural sectors receive virtu- ally no direct protection. Subsidies for fertilizer and other inputs have been an indirect source of assistance to agriculture, but the rates of these subsidies have declined. The political explanations for the protection of the sugar and rice industries are quite distinct to each case. The protection of the sugar industry is a conse- quence of the political power of the highly concentrated sugar refining industry, including the state-owned component of this industry, which is closely linked with large-scale sugar plantations.1 In contrast, Indonesia's paddy industry--the farm-level production of rice--is dominated by small-scale farm-level producers. The rice milling sector is much more concentrated and well organized, however, *The author has benefited greatly from the research assistance of Arief Anshory Yusuf in conducting the model simulations, from helpful discussions with Ernesto Valenzuela, and from the comments of John Cockburn. 179 180 Agricultural Price Distortions, Inequality, and Poverty and this is relevant because imports compete with milled rice rather than with the raw unmilled product (paddy) produced by farmers. The political power of the rice millers has been an important source of support for the protection of the rice industry. The enhanced political power of the Indonesian parliament since the upheavals induced by the Asian crisis, together with the economic nationalism that dominates the membership of the parliament, has strengthened this support for the protection of the rice industry. Since 2000, imports of rice have officially been banned. In part, this policy has reflected the dubious claim, advanced by supporters of rice industry protection, that restricting rice imports reduces poverty. A general equilibrium analysis presented in Warr (2005) indicates that the policy boosts poverty in rural and urban areas because the poverty-increasing effects of raising the consumer price of rice far exceed the poverty-reducing effects of raising the producer price. The purpose of our study is to analyze the effects of agricultural and other trade policies both in Indonesia and at a global level. We focus particularly on the effects on the incidence of poverty in Indonesia. We examine the effects of liberal- ization in the markets for all tradable goods and the effects of liberalization in agricultural markets alone. For this kind of analysis, a general equilibrium approach is essential. For illustrative purposes, consider the impact of reducing the protection of the rice industry, a highly controversial issue in Indonesia. An adequate analysis of the distributional impact of this policy needs to take account of the policy effects on the expenditures of households, disaggregated by household group, but also the policy effects on household incomes. This requires an examination of the impact on wages, which operate through effects on the labor market, as well as the impact on the returns to the agricultural land and capital owned by poor people. In undertaking this analysis, however, we should not consider the rice industry in isolation. A reduction in rice prices will induce some contraction in rice (paddy) production. The paddy industry is a large employer of unskilled labor in absolute terms. Depending on the labor intensity of this industry relative to the labor intensity of other industries, a contraction of output might induce an increase or a decrease in real unskilled wages. Any change in unskilled wages would affect the profitability in other industries and generate impacts on outputs and prices in those industries as well. These effects would have repercussions on household incomes. The repercussions on incomes would have to be balanced against the effects on consumers of a reduction in the price of rice. Moreover, the consump- tion of rice may not be considered in isolation either. A reduction in the price of rice will have implications for the demand for other staple foods, such as corn and wheat flour, another significant import. Finally, the reduced protection may lower government revenue if the instrument of protection is a tariff, or it Indonesia 181 may lower private rents if the instrument is an import quota. The way this revenue is spent by the government or the private quota holders will also influence the net distributional outcome. The debate over Indonesia's rice protection illustrates the necessity of a general equilibrium approach. The economic issues involved are complex and interrelated. A framework is required that accounts for these interactions and simultaneously satisfies all relevant market clearing conditions and macroeconomic constraints. To address the issues of poverty and inequality, such a framework must include a disaggregated household sector. Moreover, as the above discussion illustrates, the full impact of a reduction in the protection of the rice industry depends on the val- ues of key economic parameters. In the case of the rice example, these include the supply response of domestic producers and the elasticity of supply of rice imports in Indonesia. However, the true values of these parameters are uncertain. A frame- work is therefore needed in which the values of key parameters may be varied, where appropriate, to determine the sensitivity of the results to the assumed values of the parameters. The next section describes the Wayang general equilibrium model of the Indone- sian economy, the principal analytic tool we use in this study. The subsequent sec- tion describes the simulations performed with this model, in combination with the World Bank Linkage model of the world economy (van der Mensbrugghe 2005). The simulations involve unilateral agricultural and trade policy reform in Indonesia and reform in the rest of the world. Our purpose is to assess the relative importance of own-country versus rest-of-the-world policies in terms of the effects on Indone- sian households. The results are presented with a focus on the implications for the incidence of poverty in Indonesia. The final section concludes. The Wayang General Equilibrium Model of the Indonesian Economy This section briefly describes the major elements of the Wayang model. The household sector of the model is crucial to the analysis of the incidence of poverty. The most important features of the model are summarized in the overview below. Following the overview, we describe the theoretical structure of the model and the related database. We then discuss important features of the parameter assumptions. Overview The Wayang model identifies 10 household types, which represent 10 socioeco- nomic groups as defined in the social accounting matrix (SAM) published by 182 Agricultural Price Distortions, Inequality, and Poverty Statistics Indonesia (BPS 2005). For the purposes of our study, each of these 10 SAM household categories is divided into centile groups--100 subcategories, with an equal population segment in each--arranged by per capita consumption expenditure. Working with a general equilibrium model containing a highly dis- aggregated household sector allows us to conduct controlled experiments that focus on the consequences for household incomes, expenditures, poverty, and inequality that arise from selected economic shocks, taken one at a time. The Wayang model also has a disaggregated industry and commodity struc- ture. According to the assumptions adopted about microeconomic behavior, all firms seek competitive profit maximization, and all consumers seek competitive utility maximization. In the simulations reported in this chapter, the markets for final outputs, intermediate goods, and factors of production are all assumed to clear at prices that are determined endogenously within the model.2 The nominal exchange rate between the rupiah and the U.S. dollar may be thought of as exoge- nously fixed. The role in the model of the exogenous nominal exchange rate is to determine, along with international prices, the nominal domestic price level. Given that prices adjust flexibly to clear markets, a 1 percent increase in the rupiah-dollar exchange rate will result in a 1 percent increase in all nominal domestic prices, leaving all real variables unchanged. The Wayang model belongs to the class of general equilibrium models that are linear in proportional changes, meaning that relative effects are not altered by the size of a shock. These models are sometimes referred to as Johansen models, after the seminal work of Johansen (1964), who used this approach. The Wayang model shares many structural features with the highly influential ORANI general equi- librium model of the Australian economy (Dixon et al. 1982) and the general equilibrium model of the global economy established through the Global Trade Analysis Project (Hertel 1997). These two models also belong to the Johansen cat- egory. The specific structure of the Wayang model draws on an earlier version of the model (Warr et al. 1998) and on a revised version of the ORANI model, called ORANI-G (Horridge 2004). The features of the Wayang model have been adapted to reflect important aspects of the Indonesian economy and to facilitate the analysis of poverty and inequality in Indonesia.3 The principal components of the model are summarized below. Industries The national model contains 65 producer goods and services offered by 65 corre- sponding industries, of which 18 are agricultural industries, 6 are resource sectors (wood, other forestry, fishing, minerals mining, crude oil, and other mining), and 41 are other industries. Each industry produces a single output; so, the set of Indonesia 183 commodities coincides with the set of industries. The various industries in the model are classified as either export oriented or import competing. The level of exports of an export-oriented industry is treated as endogenous, while any exports of an import- competing industry are treated as exogenous.4 The criterion used to classify these industries is the ratio of an industry's imports to the industry's exports. If this ratio exceeds 1.5, then the industry is said to be producing an importable. If the import- export ratio is less than 0.5, then the industry is deemed export oriented. For ratios between 0.5 and 1.5, additional relevant information is used to classify the industry. Commodities The Wayang model contains two types of commodities: producer goods and con- sumer goods. Producer goods are either domestically produced or imported. All 65 producer goods are, in principle, capable of being imported, although some show zero import levels in the database, in which services and utilities account for most of the examples. The 20 consumer goods identified in the model are each transformed from the producer goods. The proportions of domestically produced and imported producer goods of each kind used in this transformation are sensitive to the (Armington) elasticities of substitution and the relative price changes of the goods. Factors of production The mobility of factors of production is a critical feature of any general equilib- rium system. Mobility here means mobility across economic activities (industries) rather than geographical mobility, although the two are clearly connected. The greater the factor mobility built into the model, the greater the economy's simu- lated capacity to respond to changes in the economic environment. It is essential that assumptions about the mobility of factors of production be consistent with the length of the run that the model is intended to represent. Two types of labor are identified: unskilled and skilled. They are distinguished by the educational characteristics of the workforce. Skilled labor is defined as those workers with at least a lower secondary education. Both types of labor are assumed to be fully mobile across all sectors. These assumptions imply that skilled wages must be equal in all sectors and move together. The same applies to unskilled wages, although the two need not be the same and need not move together. An alternative treatment, popular in general equilibrium modeling studies, is to assume labor mobility within the agricultural and nonagricultural sectors, but not between them. This approach is rejected here because it denies a central reality of the Indonesian economy (and of many other developing economies): the mobility of labor between rural and urban areas even in the short run. The approach would 184 Agricultural Price Distortions, Inequality, and Poverty rule out all resource mobility between the agricultural and nonagricultural indus- tries, greatly limiting the scope for economic adjustment to changes in patterns of incentives such as the changes that would be produced by trade liberalization. In Indonesia, unskilled and semiskilled workers move readily, often seasonally, between the agricultural and nonagricultural sectors of the economy. Indeed, this mobility is more important than direct mobility among the various agricultural regions of the country, although the latter sort of mobility also occurs. The mobility of capital is a somewhat different case. It is assumed that there are two kinds of mobile capital: one kind that is mobile among agricultural sectors and another kind that is mobile among nonagricultural industries. Mobile agricultural capital cannot be used outside agriculture, and mobile nonagricultural capital can- not be used in agriculture. Mobile agricultural capital includes machinery such as tractors of various types, which may be used in a range of agricultural activities. It is best to consider land (as we use the term) as an immobile form of agricultural capi- tal, which, in the short run, includes much true land. In nonagriculture, plant and buildings are classified as mobile because they may be used for many purposes. A factory building is a good example. Machinery is considered immobile because most such machinery is more industry specific than are tractors in agriculture. Table 6.1 summarizes some features of the cost structure of the paddy industry and compares this cost structure with the situation in the rest of the agricultural sec- tor and the rest of the economy. The paddy industry is intensive in its use of unskilled labor, which accounts for 18.5 percent of the total costs and 31 percent of the total variable costs excluding capital and land. Both shares are well above the Table 6.1. The Cost Shares of Major Factors of Production, Paddy and Other Industries, Indonesia, 2000 (percent of total costs) Other Non- All Cost components Paddy agriculture agriculture Industries Unskilled labor 18.5 9.0 6.3 8.1 Skilled labor 3.1 6.6 7.1 7.0 Mobile agricultural capital 20.6 21.3 0.0 1.7 Mobile nonagricultural capital 0.0 0.0 25.3 23.2 Land 18.1 20.2 0.0 1.6 Nonland fixed capital 0.0 0.0 25.3 21.2 Intermediate inputs 39.7 42.9 36.0 37.2 Total 100.0 100.0 100.0 100.0 Source: Wayang model database, based on data of Statistics Indonesia (BPS 2001) and unpublished agricultural cost survey data accessed at the Center for Agro-Social and Economic Research, Ministry of Agriculture, Bogor, Indonesia, in March 2008. Indonesia 185 shares in other agricultural industries and the rest of the economy, on average. This point is important in our discussion elsewhere below. It is assumed that, in every sector, the production technology is characterized by constant elasticity of substitution and diminishing returns to scale to variable factors alone. However, we introduce a sector-specific fixed factor in each sector to assure that there are constant returns to scale in production to all factors. We refer to the set of specific factors in the agricultural sectors as land and to the set of spe- cific factors in the nonagricultural sectors as fixed capital. The assumption of con- stant returns means that all factor demand functions are homogeneous of degree 1 in output. In each sector, there is a zero profit condition that equates the price of output to the minimum unit cost of production. This condition may be thought of as determining the price of the fixed factor in that sector. Households The model contains 10 major household categories--seven rural and three urban--that are differentiated by socioeconomic group. The sources of income of each household type depend on the ownership by the households of factors of production. These differ among the household categories and are estimated from the 2000 SAM compiled by Statistics Indonesia (BPS 2005). The SAM is based pri- marily on the National Socioeconomic Survey--the Susenas Survey--conducted by Statistics Indonesia. Drawing on the 1999 Susenas data (BPS 2000), we subdi- vide each of the 10 household categories into 100 subcategories, each of which has the same population size, and arrange them according to per capita real consump- tion expenditure, thus establishing a total of 1,000 subcategories.5 The consumer demand equations for the various household types are based on the linear expen- diture system. Within each of the 10 major categories, the 100 subcategories differ according to the budget shares in consumption. Because the focus of our study centers on income distribution, the sources of income of the various households are of particular interest. The source of the factor ownership matrix used in the model is the SAM for Indonesia in 2000. The households are described as follows: · Rural 1, agricultural employees: agricultural workers who do not own land · Rural 2, small farmers: agricultural workers with less than 0.5 hectare of land · Rural 3, medium farmers: agricultural workers with 0.5 to 1 hectare of land · Rural 4, large farmers: agricultural workers with more than 1 hectare of land · Rural 5, low-income nonagricultural rural households: small retail store own- ers, small entrepreneurs, small personal service providers, and clerical and manual workers in rural areas 186 Agricultural Price Distortions, Inequality, and Poverty · Rural 6, nonlabor rural households: nonlabor-force and unclassified house- holds in rural areas · Rural 7, high-income nonagricultural rural households: managers, technicians, professionals, military officers, teachers, large entrepreneurs, large retail store owners, large personal service providers, and skilled clerical workers in rural areas · Urban 1, low-income urban households: small retail store owners, small entre- preneurs, small personal service providers, and clerical and manual workers in urban areas · Urban 2, nonlabor urban households: nonlabor-force and unclassified house- holds in urban areas · Urban 3, high-income urban households: managers, technicians, professionals, military officers, teachers, large entrepreneurs, large retail store owners, large personal service providers, and skilled clerical workers in urban areas In the SAM, the income and expenditure items of each household are classified as follows: wages and salaries, rent from capital, incoming transfers, outgoing transfers, income taxes, final consumption, and savings. The wages and salaries and rent from capital categories are each subdivided into various subcategories. These categories do not correspond exactly to the cat- egories in the model. In agriculture, returns to land and capital are not separated in the SAM, but returns to owner-provided labor are separated. We have used a previous study on the cost structure of paddy production to allocate returns among the land and capital categories, and the various farming households receive the same proportionate breakdown from this total (see Warr 2005). The factor ownership characteristics of the 10 major household categories are summarized in table 6.2. These household categories vary considerably in the composition of factor incomes, but, for the purposes of our study, the limitations in the available data have obliged us to impose the assumption that the composi- tion of factor incomes is uniform across the 100 subcategories in each of the 10 major categories. These 100 subcategories thus obtain incomes from factors of production in the same proportions. Of course, because the incomes of these 100 subcategories vary greatly, we should think of them as owning different quan- tities of a uniform bundle of factors. The composition of the factor bundles varies across the 10 major household categories, but is uniform within each category. The composition of expenditures on final commodities does vary among the 100 subcategories, however, and also across the 10 major household categories. The characteristics of the 10 household categories described above are summa- rized in table 6.3. The table shows the importance of each category in the overall incidence of poverty in Indonesia using the government's official poverty line. As Indonesia 187 Table 6.2. Sources of the Factor Incomes of 10 Broad Household Groups, Indonesia, 2000 Mobile Mobile Fixed Total Skilled Unskilled agricultural nonagricultural capital factor Category Land labor labor capital capital and land income Rural 1 4.1 1.4 53.6 2.1 9.3 29.5 100.0 Rural 2 1.6 6.1 26.7 1.4 16.3 47.9 100.0 Rural 3 9.8 2.7 14.1 4.8 16.1 52.6 100.0 Rural 4 9.7 4.0 7.8 4.9 17.4 56.3 100.0 Rural 5 7.6 7.0 43.3 3.6 8.7 30.0 100.0 Rural 6 2.8 29.2 15.2 1.7 12.7 38.4 100.0 Rural 7 12.6 20.7 4.5 5.9 12.4 44.0 100.0 Urban 1 4.1 12.8 24.4 2.4 13.8 42.5 100.0 Urban 2 3.2 22.0 42.3 1.7 7.4 23.4 100.0 Urban 3 4.1 23.8 1.3 2.5 17.0 51.4 100.0 Source: Wayang model database (see Warr et al. 1998), based on data of Statistics Indonesia (BPS 2000). Table 6.3. Expenditure and Poverty Incidence by Household Category, Indonesia, 2000 Mean per Share of Share of Share of total capita monthly poor in category in Category population, % expenditure, Rp category, % all poor, % Rural 1 8.0 6,358 39.8 13.9 Rural 2 14.8 3,608 34.9 22.4 Rural 3 7.1 7,584 32.3 9.9 Rural 4 9.0 6,618 27.8 10.9 Rural 5 16.0 3,891 23.8 16.5 Rural 6 4.9 12,795 28.0 5.9 Rural 7 5.0 16,060 10.5 2.3 Urban 1 20.4 4,210 15.2 13.4 Urban 2 6.1 17,813 11.2 2.9 Urban 3 8.7 14,353 5.0 1.9 Indonesia 100.0 12,084 23.1 100.0 Memorandum items Headcount poverty incidence, national,% 23.10 Headcount poverty incidence, rural, % 29.09 Headcount poverty incidence, urban, % 11.97 Gini coefficient, national 0.335 Gini coefficient, rural 0.291 Gini coefficient, urban 0.356 Source: Wayang model database (see Warr et al. 1998), based on data of Statistics Indonesia (BPS 2000). 188 Agricultural Price Distortions, Inequality, and Poverty in many developing countries, the incidence of poverty in Indonesia is highest among the rural socioeconomic categories. Rural households account for 82 percent of all poor people in Indonesia, but only 65 percent of the total population. Theoretical structure of the model The analytical structure of the model includes the following major components: · The household consumption demands of each of the 10 broad household types for 20 categories of consumer goods are derived from the linear expendi- ture system. · The household supplies of skilled and unskilled labor are assumed to be exogenous. · The factor demand system is based on an assumption of constant elasticity of substitution in production technology that relates the demand for each pri- mary factor to industry outputs and to the prices of each of the primary fac- tors. This reflects the assumption that factors of production may be substituted for one another in ways that depend on factor prices and on the elasticities of substitution between the factors. · The distinction between skilled and unskilled labor, which are nested within the sectoral production functions, is addressed in each nonagricultural sec- tor by allowing skilled and unskilled labor to enter a constant elasticity of substitution production function to produce effective labor. Effective labor, variable capital, and fixed capital then enter the production functions for domestic output. · Leontief assumptions are used to represent the demand for intermediate goods. The demand for each intermediate good in each sector is assumed to be in fixed proportion to the gross output of the sector. · The demand for imported and domestically produced versions of each good incorporates Armington elasticities of substitution between the two versions. · The set of equations that determines the incomes of the 10 household types from the (exogenous) household ownership of factors of production reflects data derived from the official 2000 SAM (BPS 2005), the (endogenous) rates of return to these factors, and any net transfers from elsewhere into the system. · The import tariff and excise tax rates across commodities; the rates of business taxes, value added taxes, and corporate income taxes across industries; and the rates of personal income taxes across household types reflect the structure of the Indonesian tax system and rely on data of the Ministry of Finance.6 · A set of macroeconomic identities ensure that standard macroeconomic accounting conventions are observed. Indonesia 189 The Wayang model database This subsection provides a description of Indosam 2000, a disaggregated SAM for Indonesia with a base year of 2000. This SAM is intended to serve as the database for Wayang, but it has other, potential uses as well. At the time of our study, 2000 is the most recent year on which it is possible to assemble the information required for the construction of a SAM for Indonesia. Four data sources, all compiled by Statistics Indonesia, have been used to con- struct Indosam 2000: the input-output tables for 2000 (BPS 2001), the 2000 SAM (BPS 2005), the 1999 Susenas National Socioeconomic Survey (BPS 2000), and supplementary data sources used in the construction of specific tables. The prin- cipal data source is the 2000 SAM, produced by Statistics Indonesia (BPS 2005). It contains 22 production sectors, which is insufficient for the purposes of our study. Moreover, the SAM does not include adequate detail on tax payments and house- hold sources of income that is also required for the study. The 2000 input-output tables specify 66 production sectors. For the purposes of our study, modifications to the data contained in the input-output tables are needed for three reasons. First, the table specifies only the total transactions for intermediate goods and services for each pair of producing and purchasing industries at producer prices. Unlike the 1990 tables (BPS 1990), these transactions are not divided into goods and services provided through domestic and imported sources. Second, the tables include a sector--number 66, labeled as an unspecified sector--as a balancing item. Sector 66 does not describe a true sector of the economy, and, in any case, the data on this sector indicate neg- ative final demand, an economic impossibility. Third, the tables obtained from Statistics Indonesia are not fully balanced. The major imbalance is that, for most industries defined in the table, the industry-specific elements in row 210 (total input) are not equivalent to the corresponding elements in row 600 (total output), and the elements in row 200 (total imports), plus row 600 (total output) are not equivalent to the elements in row 700 (total supply). We have overcome these problems as follows. First, we have used the shares of imported intermediate goods and domestically produced intermediate goods for each cell of the table, as implied by the 1990 input-output tables (BPS 1990), to divide the transactions in intermediate goods into domestic and imported components, a distinction that is required by the Armington theoretical struc- ture of the demand for intermediate goods. Second, we have aggregated sector 66 with the much larger sector 65 (other services), which has eliminated the problem of negative final demand. Third, we have balanced the revised table using the RAS adjustment method to ensure that all required accounting identi- ties are observed. 190 Agricultural Price Distortions, Inequality, and Poverty The model's elasticities All export demand elasticities are set equal to 20. The elasticities of supply of imports in Indonesia are assumed to be infinite (import prices are set exoge- nously) except in the case of rice, for which the assumed elasticity is 10. All pro- duction functions are assumed to be constant elasticity of substitution functions in primary factors with elasticities of substitution of 0.5, except for the paddy pro- duction industry, for which this elasticity is set at 0.25, reflecting the empirical observation of low elasticities of supply response in this industry. The Armington elasticities of substitution in demand between imports and domestically pro- duced goods are set equal to 2, except for the case of rice, for which the assumed value is 6. The higher value for rice reflects the assumption that imported and domestically produced rice are close substitutes, unlike the situation in most other commodities. Simulations We first describe the policy shocks that are simulated and then summarize the model closure characteristics. The shocks We simulate the effects of policy reform using the Wayang model of the Indone- sian economy, combined with the Linkage model of the world economy (van der Mensbrugghe 2005). The simulations involve unilateral agricultural and trade policy reform in Indonesia and reform in the rest of the world. We do this to assess the relative importance of the impacts of own-country versus rest-of-the-world policies on Indonesian households. We also compare reform in agriculture alone with reform in all goods markets to gauge the relative contribution of agricultural policies to the measured impacts on Indonesian households. We use the word reform here to mean the complete elimination of all tariffs, the tariff equivalents of any nontariff barriers, export taxes and export subsidies, and domestic agricultural policies in so far as they alter the producer or consumer prices of farm products in various countries. We consider three sets of policy reforms below: unilateral reform in Indonesia (simulation A); global reform, excluding Indonesia (simulation B); and the combination, that is, global reform, including Indonesia (simulation C). In all three sets of simulations, the Wayang model treats as exogenous: (1) all rates of industry assistance in Indonesia (tariffs, the tariff equivalents of nontariff barriers, and export subsidies), (2) all import prices in Indonesia, and (3) all shifters in the inverse export demand functions in Indonesia Indonesia 191 Table 6.4. Industry Assistance Rates Used in Modeling, Indonesia, 2004 (percent) Commodity Tariff Export subsidy Output subsidy Paddy rice 15.0 0.0 0.0 Wheat 0.8 0.0 0.0 Other grains 15.0 0.0 0.0 Vegetables and fruits 4.8 0.0 0.0 Oilseeds 4.7 9.0 0.0 Sugarcane 0.0 0.0 0.0 Plant-based fibers 3.7 0.0 0.0 Other crops 4.5 8.0 0.0 Cattle, sheep, and so on 3.1 0.0 0.0 Other livestock 3.1 0.0 0.0 Raw milk 0.0 0.0 0.0 Wool 3.8 0.0 0.0 Other primary products 2.9 0.6 0.3 Beef and sheep meat 4.8 0.0 0.0 Other meat products 4.5 0.0 0.0 Vegetable oils and fats 2.9 0.0 0.0 Dairy products 3.9 0.0 0.0 Processed rice 15.0 0.0 0.0 Refined sugar 18.3 0.0 0.0 Other food, beverages, tobacco 15.3 0.0 0.0 Textiles and wearing apparel 8.0 1.4 0.0 Other manufacturing 5.1 1.2 0.1 Services 0.0 0.0 0.2 Source: Valenzuela and Anderson (2008), based on the estimates compiled by Anderson and Valenzuela (2008). (equivalent to shifts in the prices at which Indonesia may export a given vol- ume of exports). Simulation A depicts reform in Indonesia alone. This simulation relies only on the Wayang model and does not involve the Linkage model. The assumed rates of industry assistance in Indonesia are set out in table 6.4, and the database of the Wayang model has been amended to match these rates. Exogenous variables (1) are set to zero (changed exogenously by 100 percent) in this scenario, in which all Indonesian tariffs, the tariff equivalents of quantitative restrictions, and all export subsidies are eliminated. Exogenous variables (2) and (3), import prices and export demand shifters, do not change. Export prices are determined endoge- nously in the Wayang model by export demand equations for Indonesia that relate the export price of each commodity to the quantity of the commodity exported.7 192 Agricultural Price Distortions, Inequality, and Poverty The export quantities are endogenously determined within the Wayang model, and export prices are determined simultaneously by movements along the export demand equations (recalling that the export demand shifters are zero). Simulation B depicts reforms in the rest of the world (all countries except Indonesia). In this case, exogenous variables (1) do not change, but exogenous variables (2) and (3) do. The simulation uses the changes in import prices in Indonesia and the shifts in the export demand equations for Indonesia that are generated by simulations in the Linkage model. These Linkage model simulations, conducted by van der Mensbrugghe, Valenzuela, and Anderson (see the appen- dix), estimate the changes to these import prices and export demand shifters that result from liberalization in all countries except Indonesia; these results are then applied as shocks to the Wayang model. The changes to border prices derived from the Linkage model are shown in table 6.5. Simulation C combines simulations A and B to depict global reform, including Indonesia. In this case, exogenous variables (1), (2), and (3) all change. The non- linear approximation techniques used to solve the Wayang model mean that the results of simulation C are not exactly the arithmetic sum of the results of simula- tions A and B, but they are similar to the arithmetic sum. Each of the above simulations is conducted twice: once to show the case in which the reductions in the protection in Indonesia and in the rest of the world apply to all traded commodities (labeled simulations A1, B1, and C1) and once to show the case in which the reductions in protection apply only to agricultural and lightly processed food commodities (labeled simulations A2, B2, and C2). Model closure Because we have chosen the real consumption expenditure of each household as the basis for welfare measurement and the calculation of the incidence of poverty, we must make the macroeconomic closure compatible with this measure and with the single-period horizon of the model. We do this by ensuring that the full eco- nomic effects of the shocks to be introduced are channeled into current-period household consumption and do not leak in other directions so that we are unable to capture real world intertemporal welfare implications in our welfare measure. The choice of macroeconomic closure may thus be seen partly as a mechanism to minimize inconsistencies between the use of a single-period model to analyze welfare results and the multiperiod reality that the model imperfectly represents. To prevent intertemporal and other welfare leakages from occurring, we con- duct the simulations for the case of balanced trade (the exogenous balance on the current account). This ensures that the potential benefits of the liberalization do not flow to foreigners through a current account surplus or that increases in Indonesia 193 Table 6.5. Exogenous Border Price Shocks That Would Result from Rest-of-the-World Liberalization, Indonesia (percent deviation from base) Export price shocksa Import price shocksb Reform Only Reform Only of all agricultural of all agricultural Commodity goods reform goods reform Paddy rice 0.0 0.0 4.3 2.7 Wheat 0.0 0.0 6.8 7.1 Other grains 3.1 1.4 2.7 2.8 Oilseeds 3.0 1.3 1.8 1.3 Sugarcane and beets 0.0 0.0 0.0 0.0 Plant-based fibers 3.1 1.4 7.6 8.7 Vegetables and fruits 3.1 1.4 2.6 1.9 Other crops 3.0 1.3 1.5 1.8 Cattle, sheep, and so on 3.0 1.3 5.6 5.5 Other livestock 2.7 1.1 1.9 0.6 Raw milk 0.0 0.0 0.0 0.0 Wool 0.0 0.0 9.8 10.0 Beef and sheep meat 3.0 1.4 5.6 5.7 Other meat products 2.5 0.8 3.3 3.6 Vegetable oils and fats 2.5 0.9 0.4 1.1 Dairy products 2.7 1.1 8.6 8.8 Processed rice 3.0 1.3 3.7 2.8 Refined sugar 2.6 1.0 2.9 2.5 Other food, beverages, tobacco 2.4 0.9 0.8 0.7 Other primary products 2.1 0.6 1.3 0.8 Textiles and wearing apparel 2.1 1.0 0.3 0.4 Other manufacturing 1.9 0.6 0.4 0.4 Services 2.1 0.6 0.2 0.2 Agriculture and food 2.5 1.0 3.0 3.2 Agriculture 3.0 1.3 4.1 4.5 Processed foods 2.5 0.9 1.7 1.8 Other manufacturing 2.0 0.6 0.4 0.4 Nontradables 2.1 0.6 0.2 0.2 Total 2.1 0.7 0.5 0.6 Merchandise trade 2.1 0.7 0.7 0.8 Source: Linkage model simulations by van der Mensbrugghe, Valenzuela, and Anderson (see the appendix). a. Simulated as shocks to the inverse export demand equations for Indonesia. b. Simulated as shocks to the exogenous import prices for Indonesia. 194 Agricultural Price Distortions, Inequality, and Poverty domestic consumption are not achieved at the expense of borrowing from abroad in the case of a current account deficit. For the same reason, we exogenously fix real government spending and real investment demand for each good. We hold the government budget deficit fixed in nominal terms. We achieve this by endoge- nous across-the-board adjustments to the sales tax rate so as to restore the base level of the budgetary deficit. As the combined outcome of these features of the closure, the full effects of the changes in policy are channeled into household con- sumption and not into effects that are, in fact, relevant to economic welfare, but are ignored in the single-period focus of the model. The results of the liberalization of the markets for all goods While the emphasis in our study is on the effects on poverty and income inequality, an understanding of these effects requires that we first look at the macroeconomic impacts. Macroeconomic effects The macroeconomic effects of trade reform in the markets for all commodities are summarized in table 6.6. Real GDP rises in Indonesia under all three reform sce- narios. The increases are small in the case of unilateral liberalization (Sim A1) and moderate if other countries liberalize as well (Sim B1 and Sim C1). Under unilat- eral liberalization, the domestic price levels in Indonesia decline (measured as the GDP deflator and the consumer price index). Aggregate real household consump- tion rises marginally under unilateral reform (Sim A1) and rises more substan- tially if the rest of the world reforms (B1) and under global reform, including Indonesia (C1). As our description of model closure indicates, real investment, real inventories, and real government spending (each deflated by the relevant price deflator), as well as the real trade balance (measured in foreign exchange terms), are all held constant in these simulations. The nominal values of each of these categories thus change as the price levels of the components change. The nominal values of GDP and consumption change in line with the fact that the GDP deflator and the con- sumer price index both decline as tariffs are removed in Indonesia. The same is true of investment and inventories. However, nominal government spending increases. The reason is evident from the changes in factor prices. The real value of skilled labor increases by a propor- tion (5.7 percent) greater than the decline in the consumer price index (2.2 per- cent). Nominal skilled wages therefore rise (3.5 percent). Because government expenditure is heavily concentrated in the employment of skilled (educated) labor, nominal government spending must rise to maintain the real value of Indonesia 195 Table 6.6. Aggregate Simulation Results of the Prospective Liberalization of All Commodities, Indonesia Sim A1, Sim B1, Sim C1, unilateral rest-of-the-world global Indicator liberalization liberalization liberalization Macroeconomic aggregates, % change from base Real GDP, expenditure side, GDP deflator 0.54 1.12 1.37 Real household consumption, CPI deflator 0.54 5.21 5.78 Import volume index, duty- paid weights 11.95 12.54 27.09 Export volume index 9.84 4.59 15.56 GDP price index, expenditure side 1.75 9.06 7.59 Consumer price index 2.19 8.79 6.50 Nominal change, Rp, billions GDP 17,923 150,926 133,071 Consumption 14,854 129,693 113,574 Investment 4,561 12,890 9,135 Inventory 109 1,580 1,829 Government expenditure 1,600 9,922 12,191 Exports, net of imports 0 0 0 Real return to factors, % change from base, using the CPI deflator Unskilled labor 0.2 7.1 8.5 Skilled labor 5.7 1.8 8.1 Agricultural capital 5.6 11.9 9.0 Nonagricultural capital 3.1 4.6 1.5 Land 2.9 21.0 19.1 Real household expenditure, % change from base, using the CPI deflator Rural 1 0.8 3.9 3.0 Rural 2 1.1 5.9 4.7 Rural 3 0.1 6.8 6.8 Rural 4 0.7 4.4 5.1 Rural 5 1.0 5.5 6.7 Rural 6 0.3 5.6 6.0 Rural 7 1.1 5.2 6.3 Urban 1 0.1 4.6 4.5 Urban 2 1.6 5.1 7.0 Urban 3 1.9 5.5 7.4 Source: CGE model simulations for Indonesia by the author. Note: CPI consumer price index. 196 Agricultural Price Distortions, Inequality, and Poverty government spending. Consumption is the only component of expenditure on GDP the real value of which is not fixed exogenously. The increase in nominal government spending therefore limits the amount by which real household con- sumption may increase in response to the decline in protection. The real value of skilled wages rises because, according to the rates of industry assistance in Indonesia used in these simulations (see table 6.4), the country's pro- tective structure acts against the interests of industries that are intensive in the use of skilled labor. The protective structure is virtually neutral with regard to unskilled labor. The existence of this protection reduces the real value of skilled wages, and the removal of the protection accomplishes the reverse. The bottom panel of table 6.6 summarizes the changes in the real consumption expenditures of each of the 10 household groups. Under unilateral liberalization (Sim A1), the poorer rural household groups (Rural 1, 2, and 3) and the poorest urban household group (Urban 1) lose, and all other household groups gain. The changes in real factor prices and in the sources of household incomes (see table 6.2) provide the main explanation for these outcomes. Real unskilled wages remain virtually unchanged, and the real return to agricultural land and capital decline in this simulation, thus harming poorer rural households. At the same time, the real returns to skilled labor and nonagricultural capital increase, favor- ing the richer households, especially households in urban areas. In addition, the reduced protection directly lowers the consumer prices of the imported goods subject to the protection. These results necessarily reflect the structure of protection in Indonesia that is assumed in our study, as represented in the Distortions to Agricultural Incentives Project Database (Anderson and Valenzuela 2008). According to our model results, protection disfavors skilled labor-intensive industries overall, while favoring agricultural industries that are intensive in the use of land and agricultural capital. The removal of the protection harms agricultural land and capital and benefits skilled labor and nonagricultural capital. These factors are owned most intensively by the richest urban households, and these households are, in consequence, the largest beneficiaries of unilateral liberalization in Indonesia. Liberalization in the rest of the world (Sim B1) produces international price changes that raise real GDP in Indonesia by about twice the increase that would result from unilateral liberalization by Indonesia, but the effects within Indonesia are qualitatively different. Liberalization in the rest of the world raises the real value of both skilled and unskilled wages in Indonesia, especially the former, and also raises the returns to agricultural capital and (especially) agricultural land. This occurs because liberalization in the rest of the world raises agricultural prices relative to nonagricultural prices internationally, favoring the agricultural sectors in Indonesia. These results are important in our discussion of poverty (see below). Indonesia 197 The real returns to nonagricultural capital decline, but all other factor returns increase. A key point is that rest-of-the-world liberalization benefits unskilled labor proportionately more than skilled labor in Indonesia. The increase in real consumption in Indonesia in the scenario of rest-of-the-world liberalization is 10 times the corresponding increase in the scenario of unilateral liberalization. Real household expenditures increase in all 10 socioeconomic categories in the rest-of-the-world liberalization scenario. Global reform (Sim C1) is a combination of the previous two simulations and is dominated by rest-of-the-world liberalization. The results are qualitatively similar to the effects of simulation B1. Effects on inequality and the incidence of poverty The simulated effects on the incidence of poverty according to socioeconomic group broadly mimic the effects on average household consumption among these groups, which is discussed above. The level of the incidence of poverty depends on the poverty line used in the calculations, and this may also be true of the simu- lated changes in the incidence of poverty that result from particular economic shocks. The shifts in real expenditures generated by particular economic shocks are not uniform across income groups because of differences in expenditure pat- terns. Different poverty lines act on different sections of the cumulative distribu- tion of real expenditures and may thus produce different patterns of change in the incidence of poverty within the same simulation. We use three poverty lines to investigate the effects on the incidence of poverty: the government's national poverty line, the international US$1-a-day (extreme) poverty line at purchasing power parity, and the international US$2-a-day (mod- erate) poverty line at purchasing power parity. In the case of each of these poverty lines, we apply the following calibration method. First, we use the ex ante distribution of household expenditures revealed in the model database to calculate the particular level of poverty incidence associ- ated with the poverty line. Second, we use the published level of poverty incidence based on the particular poverty line to calculate the value of the poverty line in the domestic currency in Indonesia. The published levels of poverty incidence are contained in the Statisti- cal Yearbook of Indonesia in the case of the national poverty line (BPS 2007) and the World Bank's World Development Indicators in the case of the US$1-a-day and US$2-a-day poverty lines (PovcalNet 2008). These calculations yield the base levels of the poverty lines used in subsequent calculations. Third, we simulate the ex post levels of real expenditure for each household in the model to reflect the effects of the shocks. These calculations of real expendi- ture are performed using the individual household consumer price index as the deflator to reflect the particular household's consumption bundle. 198 Agricultural Price Distortions, Inequality, and Poverty Fourth, we compare these ex post real expenditures with the poverty lines described above to obtain ex post levels of poverty incidence. Fifth, the changes in poverty incidence reported in table 6.7 (see also table 6.10 elsewhere below) are the ex post levels of poverty incidence, minus the ex ante lev- els corresponding to each of the three poverty lines described above. A positive number thus indicates an increase in the simulated level of poverty incidence as a result of the relevant shocks. In addition to the effects on the incidence of poverty, we also report simulated effects on inequality in the distribution of household real expenditures using the Gini coefficient as the measure. The Gini coefficient has values between 0 and 1; the higher values reflect greater inequality. These coefficients are estimated by constructing Lorenz curves from the distributions of ex ante and ex post real expenditures and then calculating the Gini coefficients corresponding to these distributions. (These results are presented in tables 6.8 and 6.11; see elsewhere below.) Panel a in table 6.7 indicates that unilateral liberalization in Indonesia applied to all commodities (simulation A1) raises the incidence of poverty in the poorest Table 6.7. Poverty Effects of the Prospective Liberalization of All Commodities, Indonesia a. Measured at the national poverty line Changes in poverty incidence (%), Ex ante level ex post ex ante of poverty incidence, Sim A1, Sim B1, rest- Sim C1, Household % of group unilateral of-the-world global category population liberalization liberalization liberalization Rural 1 39.81 0.81 4.82 3.52 Rural 2 34.89 0.76 5.71 4.95 Rural 3 32.29 0.09 5.53 5.49 Rural 4 27.82 0.44 2.64 4.17 Rural 5 23.78 0.82 3.91 4.60 Rural 6 28.01 0.13 3.45 3.85 Rural 7 10.50 0.77 2.07 2.35 Urban 1 15.22 0.02 2.31 2.27 Urban 2 11.16 0.52 2.35 3.09 Urban 3 5.00 0.39 1.03 1.36 Urban households 11.98 0.16 1.99 2.18 Rural households 29.09 0.05 4.26 4.36 All households 23.10 0.12 3.49 3.60 Indonesia 199 b. Measured at the US$1-a-day poverty line Changes in poverty incidence (%), Ex ante level ex post ex ante of poverty incidence, Sim A1, Sim B1, Sim C1, Household % of group unilateral rest-of-the-world global category population liberalization liberalization liberalization Rural 1 2.40 0.11 0.47 0.39 Rural 2 13.09 0.43 3.46 2.72 Rural 3 8.74 0.06 2.07 2.06 Rural 4 18.55 0.67 3.13 3.48 Rural 5 8.67 0.54 2.20 2.73 Rural 6 1.80 0.03 0.47 0.51 Rural 7 0.00 0.00 0.00 0.00 Urban 1 7.08 0.00 1.43 1.40 Urban 2 2.66 0.18 0.51 0.68 Urban 3 0.00 0.00 0.00 0.00 Urban households 4.56 0.03 0.91 0.93 Rural households 9.09 0.11 2.09 2.09 All households 7.50 0.08 1.68 1.68 c. Measured at the US$2-a-day poverty line Changes in poverty incidence (%), Ex ante level ex post ex ante of poverty incidence, Sim A1, Sim B1, Sim C1, Household % of group unilateral rest-of-the-world global category population liberalization liberalization liberalization Rural 1 48.79 2.03 3.29 2.87 Rural 2 74.36 0.62 5.05 3.28 Rural 3 66.73 0.28 6.32 6.30 Rural 4 78.64 0.26 2.79 3.49 Rural 5 66.93 0.63 4.56 5.84 Rural 6 30.59 0.72 4.66 4.81 Rural 7 2.38 0.19 0.53 0.59 Urban 1 53.11 0.05 4.04 3.95 Urban 2 31.43 0.96 2.96 4.48 Urban 3 6.79 0.62 1.39 1.78 Urban households 37.87 0.29 3.19 3.50 Rural households 60.27 0.16 4.16 4.13 All households 52.40 0.00 3.82 3.91 Source: CGE model simulations for Indonesia by the author. 200 Agricultural Price Distortions, Inequality, and Poverty three rural household categories, but lowers it among the richer rural categories and in all but the poorest urban categories. The aggregate urban and rural inci- dence of poverty both decline, but the decline is larger among urban households. National poverty incidence necessarily declines. All these effects are quite small. Table 6.7, panels b and c show that these effects are not particularly sensitive to the poverty line used except that, at the US$2-a-day poverty line, the incidence of rural poverty rises. The effects of liberalization in the rest of the world (simulation B1) are quite different. The incidence of poverty in Indonesia declines significantly in this sce- nario. The decline occurs in all 10 socioeconomic groups, but is largest in rural areas. At the national level, poverty incidence declines by 3.5 percent (national poverty line). This pattern of results is not sensitive to the poverty line used. At the US$1-a-day and national poverty lines, the incidence of poverty declines signifi- cantly in both rural and urban areas. The explanation for these outcomes is evi- dent from the changes in real factor returns described above. Rest-of-the-world liberalization raises the international prices of agricultural commodities relative to nonagricultural prices, and this produces an increase in the real value of unskilled wages and of the returns to agricultural capital and land in Indonesia. These effects benefit poor households, especially those in rural areas. The inci- dence of poverty in both urban and, especially, rural areas declines. Finally, we may compare the combined effects of unilateral reform in Indonesia and liberalization elsewhere, summarized in the tables as the effects of global reform (simulation C1). As in the case of simulation B1, the real returns to unskilled and skilled labor rise significantly, along with the returns to agricultural land and capital. But the real return to nonagricultural capital declines marginally. Both rural poverty and urban poverty decline significantly, but, especially, rural poverty. The central result is that the effects of the rest-of-the-world reform dom- inate the effects of unilateral liberalization. The outcomes in this scenario are essentially the same as those of the rest-of-the-world scenario (B1) tempered by the counteracting domestic liberalization (A1) effects. The method we use to estimate the changes in the incidence of poverty is illus- trated in figure 6.1. The figure shows the ex ante (initial) distribution of expendi- tures for the socioeconomic household group Rural 3 (medium-size farmers with 0.5 to 1 hectare of land), along with the simulated ex post (new) distribution of expenditures that results from simulation C1. The two curves thus indicate the cumulative distribution of expenditures per person at constant (year 2000) prices before and after the shock, that is, the change in nominal expenditures for each household is deflated by a household-specific index of consumer prices, reflecting that household's base-period (initial) expenditure pattern. For any level of expenditure (horizontal axis), each curve shows (on the vertical axis) the share Indonesia 201 Figure 6.1. Initial and Simulated Levels of Poverty Incidence, Indonesia (as illustrated by socioeconomic household category Rural 3, simulation C1) 100 90 80 percent of population 70 60 50 40 30 32.3 28.8 20 10 0 0 5 10 15 20 25 30 35 expenditure per capita per month (Rp, thousands, at constant 1,000 prices) intial simulation C1 poverty line Source: CGE model simulations for Indonesia by the author. Note: The cumulative distributions of expenditures corresponding to simulation C1 are calculated in real terms using household-specific consumer price indexes at year 2000 prices as the deflator for each household. It is therefore possible to compare the initial (year 2000) distribution of expenditures and the distribution marked simulation C1 with the poverty line for the year 2000. The initial and simulated (postliberalization) levels of poverty incidence for this household category, using the national poverty line, resulting from simulation C1 are 32.3 percent and 26.8 percent, respectively. See table 6.7, panel a. of the population with expenditures less than or equal to that level of expenditure. For any poverty line, the incidence of poverty may thus be read as the vertical value of the intersection between the poverty line (horizontal axis) and the cumula- tive distribution. Using the national poverty line for 2000, we find that the initial level of poverty incidence among the Rural 3 group was 32.3 percent, the intersection of the poverty line with the cumulative initial distribution in figure 6.1. Simulation C1 (global reform in all commodities) shifts the entire distribution to the right, though not uniformly, producing the simulated cumulative distribution of expenditures at con- stant prices, which is labeled simulation C1 in the figure. It is important to note that this is a shift in real expenditures measured at 2000 prices. The new level of poverty incidence may therefore be read using the same poverty line that we use elsewhere above. Poverty incidence declines to 26.8 percent, a decrease of 5.5 percent. It is apparent from figure 6.1 that the incidence of poverty declined no matter which 202 Agricultural Price Distortions, Inequality, and Poverty Table 6.8. Income Inequality Effects of the Prospective Liberalization of All Commodities, Indonesia Changes in Gini coefficient, ex post ex ante Ex ante Sim A1, Sim B1, Sim C1, Household level of Gini unilateral rest-of-the-world global category coefficient liberalization liberalization liberalization Urban households 0.3559 0.0025 0.0019 0.0045 Rural households 0.2912 0.0011 0.0002 0.0013 All households 0.3351 0.0023 0.0006 0.0030 Source: CGE model simulations for Indonesia by the author. poverty line might be chosen. In the case of our simulation here, the conclusion that the incidence of poverty declined among socioeconomic group Rural 3 is therefore not dependent on the particular poverty line that we have selected, although the magnitude of the decline is affected by the choice of poverty line. The effects on inequality in Indonesia are summarized in table 6.8. Both uni- lateral liberalization and global liberalization in all commodities raise inequality. This effect is largest in the case of unilateral liberalization. The effect operates through the increased returns to skilled labor and nonagricultural capital, which are factors owned primarily by more affluent urban households. Results of the liberalization only of agricultural markets If liberalization is confined to agricultural and lightly processed food products only (simulations A2, B2, and C2), unilateral liberalization in Indonesia (simulation A2) lowers the returns to unskilled labor, agricultural capital, and land and raises the returns to skilled labor and nonagricultural capital (table 6.9). Rice and sugar dom- inate agricultural protection in Indonesia. These industries are intensive in unskilled labor and land relative to the rest of the economy. Reducing protection in Indonesia reduces the real incomes of the owners of unskilled labor and agricultural land (table 6.10, panels a and c). Real incomes fall in most rural household categories and rise in urban categories, but reduced agricultural protection also lowers food prices. The net effects on real expenditures are small. Most rural household categories are small net losers, and all urban categories are net gainers. It has been argued elsewhere that the reduced protection of the rice industry alone reduces the incidence of poverty among both urban and rural households in Indonesia (Warr 2005). Rice differs from the rest of agriculture in two impor- tant respects. First, it is more intensive in the use of unskilled labor than the rest of agriculture. Second, rice is a staple food for most Indonesians and forms a high share of the total expenditures among the poorest groups. Thus, reductions in the Indonesia 203 Table 6.9. Aggregate Simulation Results of the Prospective Liberalization Only of Agricultural Commodities, Indonesia Sim A2, Sim B2, Sim C2, unilateral rest-of-the-world global Indicator liberalization liberalization liberalization Macroeconomic aggregates, % change from base Real GDP, expenditure side, GDP deflator 0.01 0.25 0.26 Real household consumption, CPI deflator 0.00 1.07 1.08 Import volume index, duty-paid weights 0.51 0.57 1.17 Export volume index 0.42 0.43 0.05 GDP price index, expenditure side 0.17 1.46 1.25 Consumer price index 0.28 1.55 1.20 Nominal change, Rp, billions GDP 2,411 25,137 22,206 Consumption 2,556 23,709 20,658 Investment 1 2,319 2,322 Inventory 81 728 840 Government expenditure 228 162 65 Exports, net of imports 0 0 0 Real return to factors, % change from base, using the CPI deflator Unskilled labor 0.5 3.2 2.7 Skilled labor 0.7 2.5 1.8 Agricultural capital 2.0 15.7 13.4 Nonagricultural capital 0.4 1.6 1.2 Land 1.1 29.7 28.8 Real household expenditure, % change from base, using the CPI deflator Rural 1 0.0 1.1 1.2 Rural 2 0.4 2.1 1.7 Rural 3 0.3 2.3 2.0 Rural 4 0.1 1.4 1.3 Rural 5 0.1 0.6 0.8 Rural 6 0.2 1.8 1.6 Rural 7 0.0 1.1 1.1 Urban 1 0.0 1.1 1.1 Urban 2 0.2 0.8 1.0 Urban 3 0.2 0.2 0.4 Source: CGE model simulations for Indonesia by the author. Note: CPI consumer price index. 204 Agricultural Price Distortions, Inequality, and Poverty Table 6.10. Poverty Effects of the Prospective Liberalization of Only Agricultural Commodities, Indonesia a. Measured at the national poverty line Changes in poverty incidence (%), Ex ante level ex post ex ante of poverty incidence, Sim A2, Sim B2, Sim C2, Household % of group unilateral rest-of-the-world global category population liberalization liberalization liberalization Rural 1 39.81 0.00 1.93 2.00 Rural 2 34.89 0.25 1.73 1.37 Rural 3 32.29 0.17 1.40 1.25 Rural 4 27.82 0.06 0.78 0.74 Rural 5 23.78 0.15 0.39 0.58 Rural 6 28.01 0.11 0.76 0.70 Rural 7 10.50 0.02 0.73 0.72 Urban 1 15.22 0.11 0.47 0.52 Urban 2 11.16 0.05 0.16 0.25 Urban 3 5.00 0.05 0.00 0.06 Urban households 11.97 0.08 0.29 0.35 Rural households 29.09 0.05 1.11 1.06 All households 23.10 0.02 0.85 0.83 b. Measured at the US$1-a-day poverty line Changes in poverty incidence (%), Ex ante level ex post ex ante of poverty incidence, Sim A2, Sim B2, Sim C2, Household % of group unilateral rest-of-the-world global category population liberalization liberalization liberalization Rural 1 2.40 0.00 0.14 0.15 Rural 2 13.09 0.13 0.81 0.67 Rural 3 8.74 0.11 0.84 0.77 Rural 4 18.55 0.09 0.96 0.92 Rural 5 8.67 0.10 0.23 0.35 Rural 6 1.80 0.01 0.15 0.14 Rural 7 0.00 0.00 0.00 0.00 Urban 1 7.08 0.04 0.30 0.35 Urban 2 2.66 0.03 0.05 0.09 Urban 3 0.00 0.00 0.00 0.00 Urban households 4.56 0.03 0.18 0.22 Rural households 9.09 0.03 0.50 0.48 All households 7.50 0.01 0.38 0.39 Indonesia 205 c. Measured at the US$2-a-day poverty line Changes in poverty incidence (%), Ex ante level ex post ex ante of poverty incidence, Sim A2, Sim B2, Sim C2, Household % of group unilateral rest-of-the-world global category population liberalization liberalization liberalization Rural 1 48.79 0.01 0.89 0.92 Rural 2 74.36 0.20 1.27 1.03 Rural 3 66.73 0.41 2.42 2.21 Rural 4 78.64 0.05 0.56 0.51 Rural 5 66.93 0.09 0.39 0.51 Rural 6 30.59 0.46 1.70 1.58 Rural 7 2.38 0.00 0.17 0.17 Urban 1 53.11 0.04 1.04 1.12 Urban 2 31.43 0.38 0.60 0.69 Urban 3 6.79 0.08 0.01 0.10 Urban households 37.87 0.11 0.71 0.79 Rural households 60.27 0.11 0.98 0.92 All households 52.40 0.03 0.89 0.88 Source: CGE model simulations for Indonesia by the author. protection of rice reduce real unskilled wages, but they also reduce the consumer price of rice, the staple food of the poor. The latter effect dominates. However, if the reduction in protection applies to all agricultural products, including sugar, the effect on the incidence of poverty becomes more ambiguous. Reductions in protection still reduce unskilled wages, but the poverty-reducing effect of lowering consumer prices is less strong in protected agriculture in general than in rice alone. Unilateral agricultural liberalization causes the incidence of poverty to rise in rural areas and to fall in urban areas. The net effect on the national incidence of poverty is a small decline, measured by the national poverty line, but a small increase, measured by the US$1-a-day and US$2-a-day poverty lines. Agricultural liberalization in the rest of the world (simulation B2) raises the prices of agricultural commodities relative to the prices of nonagricultural com- modities on international markets. The effect is to raise the returns to unskilled labor in Indonesia and to lower the returns to skilled labor. The incidence of poverty declines in both rural and urban areas, and this qualitative result holds regardless of the poverty line. If the world liberalizes, including Indonesia (simu- lation C2), the quantitative effects of this liberalization also dominate the effects of domestic liberalization. The results of global agricultural liberalization are thus qualitatively similar to the results of rest-of-the-world liberalization. 206 Agricultural Price Distortions, Inequality, and Poverty Table 6.11. Income Inequality Effects of the Prospective Liberalization of Only Agricultural Commodities, Indonesia Changes in Gini coefficient, ex post ex ante Ex ante Sim A2, Sim B2, Sim C2, Household level of Gini unilateral rest-of-the-world global category coefficient liberalization liberalization liberalization Urban households 0.3559 0.0001 0.0006 0.0006 Rural households 0.2912 0.0000 0.0001 0.0000 All households 0.3351 0.0002 0.0007 0.0006 Source: CGE model simulations for Indonesia by the author. Inequality in Indonesia rises slightly under unilateral agricultural liberalization because unskilled wages decline relative to skilled wages. Within rural areas alone, there is no change in inequality because, although real wages decline, this effect is offset by the reduced returns to agricultural capital and land. Rest-of-the-world reform and global agricultural reform reduce inequality in Indonesia because the returns to unskilled labor, agricultural capital, and land (owned by the poor, especially the rural poor) rise relative to other factor returns. These effects are summarized in table 6.11. Comparing the results of the liberalization in all commodities (simulations A1 to C1) with the liberalization in agricultural products only (simulations A2 to C2), six key points emerge, as follows: · Trade liberalization reduces the incidence of poverty in Indonesia, but the effects are large only if they apply at a global level and to all commodities. · The effects of rest-of-the-world liberalization dominate the effects of unilateral liberalization, and rest-of-the-world liberalization is more strongly poverty reducing than liberalization in Indonesia alone. · Unilateral across-the-board liberalization in Indonesia reduces food prices and raises the real returns to skilled labor, while lowering the returns to agriculture- specific factors of production,which produces small net reductions in the incidence of poverty overall, though the effects on individual rural household categories are mixed. · Liberalization in all commodities is more strongly poverty reducing in Indone- sia than is liberalization in agricultural products alone, whether the liberaliza- tion occurs unilaterally in Indonesia, or in the rest of the world, or both. · Unilateral liberalization confined to agricultural products generates benefits mainly among urban households; these benefits operate through reduced food prices, plus increased returns to skilled labor and nonagricultural capital. Indonesia 207 · Liberalization in all commodities raises inequality in Indonesia, whether the liberalization is unilateral or global, but especially the former, whereas liberal- ization only in agricultural products raises inequality slightly if the liberaliza- tion is unilateral and reduces inequality if the liberalization is global. Conclusions The comparative static nature of our analysis in this chapter limits our capacity to capture the full economic gains available through liberalization. Dynamic effects are not captured, and these may be significant sources of additional welfare gains and poverty reduction from trade policy liberalization. Nonetheless, within this comparative static limitation, the analysis indicates that the global reform of trade policies in all commodities would be a significant potential source of poverty reduction in Indonesia. The poor in Indonesia--rural and urban--have a strong interest in global trade policy reform, whether Indonesia is part of the liberaliza- tion or not. If Indonesia liberalizes unilaterally, poverty incidence also declines, but the comparative static benefits are much smaller. If liberalization is confined to agricultural products, the pattern of the effects is similar, but the declines in the incidence of poverty in Indonesia are much more modest. Global reform in agricultural products generates significant reductions in both rural and urban poverty. Indonesia's rural and urban poor have a strong interest in the global reform of agricultural trade policy, but--according to the comparative static analysis of our study--the rural poor do not necessarily have an interest in unilateral agricultural liberalization because some rural socioeco- nomic groups would lose out from this sort of liberalization. Notes 1. For a fuller discussion of agricultural assistance in Indonesia, see Fane and Warr (2008, 2009). The estimates by these authors of agricultural assistance in Indonesia are incorporated in the World Bank database on global agricultural distortions (see Anderson and Valenzuela 2008). These estimates cover over four decades. The representative values for the computable general equilibrium modeling as of 2004 that are used here are available in Valenzuela and Anderson (2008). 2. Variations on this assumption are possible. For example, the possibility of unemployment may be introduced by varying the closure to make either real or nominal wages exogenous, thereby allowing the level of employment to be endogenously determined by demand. 3. For an application of an earlier version of this model to the relationship between economic growth and the incidence of poverty in Indonesia, see Fane and Warr (2003). 4. Given that the exported and domestically sold goods are considered identical, this assumption is required to separate the domestic price of the import-competing good from the price of the exported good. Otherwise, the Armington structure we have described above would be redundant. An alterna- tive treatment is to distinguish between the exported commodity and the commodity sold domesti- cally; between them would be a finite elasticity of transformation in production. 208 Agricultural Price Distortions, Inequality, and Poverty 5. The population size of each of the 10 major categories is not the same, but, within each of these 10 categories, the population size of each of the 100 subcategories is the same. 6. Unpublished tax rate data of the Ministry of Finance were accessed in Jakarta in March 2008. 7. The assumption that each export demand elasticity is 20 means that export prices are close to exogenous. References Anderson, K., and E. Valenzuela. 2008. "Estimates of Global Distortions to Agricultural Incentives, 1955­2007." Data spreadsheet, October, World Bank, Washington, DC. http://go.worldbank.org/ YAO39F35E0. BPS (Badan Pusat Statistik, Statistics Indonesia). 1990. Input-Output Tables of Indonesia. Jakarta: BPS. ------. 2000. The National Socioeconomic Survey (Susenas), 1999. Jakarta: BPS. ------. 2001. Input-Output Tables of Indonesia, 2000. Jakarta: BPS. ------. 2005. Social Accounting Matrix for Indonesia, 2000. Jakarta: BPS. ------. 2007. Statistik Indonesia: Statistical Yearbook of Indonesia 2007. Jakarta: BPS. Dixon, P. B., B. R. Parmenter, J. Sutton, and D. P. Vincent. 1982. ORANI: A Multisectoral Model of the Australian Economy. Amsterdam: North-Holland. Fane, G., and P. Warr. 2003. "How Economic Growth Reduces Poverty: A General Equilibrium Analy- sis for Indonesia." In Perspectives on Growth and Poverty, ed. R. van der Hoeven and A. Shorrocks, 217­34. New York: United Nations University Press. ------. 2008. "Agricultural Protection in Indonesia." Bulletin of Indonesian Economic Studies 44 (1): 133­50. ------. 2009. "Indonesia." In Distortions to Agricultural Incentives in Asia, ed. K. Anderson and W. Martin, 165­96. Washington, DC: World Bank. Hertel, T. W., ed. 1997. Global Trade Analysis: Modeling and Applications. New York: Cambridge University Press. Horridge, M. 2004. "ORANI-G: A Generic Single-Country Computable General Equilibrium Model." Centre of Policy Studies, Monash University, Melbourne. Johansen, L. 1964. A Multi-sectoral Study of Economic Growth. Amsterdam: North-Holland. PovcalNet. World Bank, Washington, DC. http://go.worldbank.org/NT2A1XUWP0 (accessed May 2008). Valenzuela, E., and K. Anderson. 2008. "Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980­84." Research Memorandum 13 (December), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925. van der Mensbrugghe, D. 2005. "Linkage Technical Reference Document: Version 6.0." December, World Bank, Washington, DC. http://go.worldbank.org/7NP2KK1OH0. Warr, P. 2005. "Food Policy and Poverty in Indonesia: A General Equilibrium Analysis." Australian Journal of Agricultural and Resource Economics 49 (4): 429­51. Warr, P., M. Aziz, H. da Costa, and P. J. Thapa. 1998. "Wayang: An Empirically Based Applied General Equilibrium Model of the Indonesian Economy." Unpublished report, Australian National University, Canberra. 7 Pakistan Caesar B. Cororaton and David Orden This chapter analyzes the macroeconomic, sectoral, and poverty implications of the removal of agricultural and nonagricultural price distortions in the domestic markets of Pakistan and in the rest of the world. Our analysis uses the simulated results of rest-of-the-world trade liberalization in the World Bank's global Linkage model (hereafter referred to as the global model; see van der Mensbrugghe 2005) and derives results for rest-of-the-world and own-country liberalization from the Pakistan computable general equilibrium (CGE) model of Cororaton and Orden (2008). The global model incorporates new estimates of assistance to farm indus- tries for various developed and developing countries, including Pakistan, from the World Bank's research project, Distortions to Agricultural Incentives.1 Using these new estimates, the global model simulates two separate scenarios: one involving full trade liberalization and one involving trade liberalization in agriculture alone. Both scenarios exclude Pakistan. The global model simulations generate changes in the import prices for Pakistan at the border, together with changes in world export prices and shifts in the export demand for Pakistan products. We utilize these results, along with the new estimates of industry assistance in Pakistan gen- erated by Dorosh and Salam (2009), to analyze various liberalization scenarios and measure their impacts on national welfare, income inequality, and poverty using the Pakistan CGE model. Trade reform entails a fiscal revenue loss to the government because trade taxes are an important source of revenue. We conduct experiments using two alterna- tive tax replacement schemes to retain a fixed fiscal balance: a direct tax on house- hold incomes and an indirect tax on consumption. We are thus able to show how the results differ according to the choice of the tax replacement method. The simulation analysis is conducted in stages. In the first stage, we run two separate experiments. One experiment involves using the changes in the border 209 210 Agricultural Price Distortions, Inequality, and Poverty prices and the computed shifts in the world export demand for Pakistan products from the global model (see the appendix by van der Mensbrugghe, Valenzuela, and Anderson) as an exogenous shock to the Pakistan model without altering the exist- ing structure of price-distorting policies in Pakistan. The other experiment involves simulating unilateral trade liberalization in Pakistan without incorporating the changes from the global model. In the second stage, we combine these two separate experiments to examine the total effects. We conduct separate experiments in each stage for trade liberalization in all tradable goods sectors and in agriculture only (including lightly processed food). The simulations generate vectors of household income and consumer prices, which we use in conjunction with data from the 2001­02 Pakistan Household Integration Economic Survey (Federal Bureau of Statistics 2003) to calculate the impact on national income inequality and poverty. The chapter is organized as follows. The next section discusses the structure of agricultural and trade distortions in Pakistan based on the new estimates of indus- try assistance. The Pakistan CGE model is then outlined. This section covers the model database (which reveals the structure of sectoral production, trade, and con- sumption), the sources of household income, and the tax structure based on the 2001­02 social accounting matrix (SAM). This is followed by a description of trends in rural and urban poverty in Pakistan. The policy experiments and the results generated by the various modeling scenarios are subsequently discussed in detail. The last section presents a summary of the findings and policy insights. The choice of tax replacement schemes plays an important role in the results we discuss. Agricultural Policies and Industry Assistance in Pakistan The period from the 1960s to the mid-1980s involved heavy government interven- tion in Pakistan (Dorosh and Salam 2009). The government's hand on agricul- tural markets, trade policies, and the market for foreign exchange depressed the real prices of tradable agricultural commodities. The fixed exchange rate policy during these years, together with high domestic inflation, eroded significantly the competitiveness of the export sectors. However, during these years, the green rev- olution took place in agriculture. The revolution involved a package of inputs such as seeds, fertilizer, and irrigation that boosted agricultural production through higher farm productivity. Then, from the mid-1980s to the early 1990s, the government started to liberalize the agricultural sector, though it maintained control over the domestic wheat market and imposed high tariffs on vegetable oils and milk products. Prior to the 1990s, Pakistan had been pursuing an import-substituting industri- alization strategy that involved high tariff rates and quantitative import restrictions Pakistan 211 to promote the manufacturing sector. Then, major reforms were implemented in 1991 and 1997 that involved a series of tariff rate cuts and the phasing out of the quantitative import restrictions. The maximum tariff rates were reduced from 65 to 45 percent, and the number of tariff categories was cut from 13 to 5. This led to a significant drop in government revenue from trade taxes because tariffs had been the major contributor to government funds. The key policy changes affecting agricultural prices are summarized in the rest of this section, while the policy changes affecting the manufacturing sector are described elsewhere below. Wheat is the staple food in Pakistan. The wheat market is still heavily con- trolled by the government through various instruments, including government procurement (to stabilize supply), support pricing (to assist farmers), and ceiling prices (to ensure affordability among consumers). However, Pakistan's trade and pricing policies on wheat effectively tax wheat producers, while providing sub- stantial fiscal subsidies to wheat millers through the government sale of wheat at below market prices (Dorosh 2005). Government involvement in the market for cotton, which is the largest cash crop in Pakistan, has changed substantially over time. In 1974, the government prevented the private sector from engaging in the international cotton trade, but this changed in 1989, when the private sector was allowed to buy cotton directly from ginners, export, and sell cotton domestically. Also, exports of cotton were subjected to an export tax. After the abolition of the export duty on cotton in 1994, domestic prices were more in line with international prices (Cororaton and Orden 2008). Since the mid-1990s, exports and imports of cotton have been prac- tically duty free, although seed cotton continues to enjoy indirect protection because of import tariffs on vegetable oils that increase the price of cotton seed oil. Otherwise, government intervention has recently been limited to an annual review of the support prices for seed cotton and some public sector procurement to maintain these prices. Rice is the third largest crop after wheat and cotton. There were heavy controls on rice in the early 1970s when the government instituted a monopoly procure- ment scheme to limit domestic consumption and expand exports. The two varieties of rice (basmati and the ordinary coarse rice called IRRI rice) are exported. The intervention system still exists but, since 2003­04, government procurement has been minimal. There were no export taxes on rice in the mid-2000s, but imports were subject to a 10 percent customs duty. The average domestic price of rice is below the export price (often about 20 percent) because of quality differences. The domestic marketing and processing of sugarcane were highly regulated until the mid-1980s. The zoning of sugar mills required farmers to sell sugarcane to mills inside their zone until 1987. There has been no government procurement of 212 Agricultural Price Distortions, Inequality, and Poverty sugarcane, but the government annually announces a support price that greatly assists sugarcane and refined sugar production, and it adjusts import tariffs and the related taxes to stabilize domestic prices. There are export bans on sugarcane and refined sugar, but they do little to reduce the high level of assistance to the industry. There was a minor tax on vegetable oils in the 1970s and 1980s, but, since the 1990s, vegetable oil imports have been taxed heavily. For example, in 2005­06, the tariff was 32 percent on imported soybean oil and 40 percent on palm oil. Like- wise, the domestic prices of sunflower oil are considerably higher than the border price. Nonetheless, two-thirds of the edible oil requirements of Pakistan are supplied by imports. Maize is mainly used as feed in the livestock and poultry sectors. Maize pro- duction has expanded rapidly in recent years because of the strong demand for poultry products. The government has not intervened in the production and mar- keting of maize. However, there are tariffs on imported maize; these range from 10 to 25 percent. Maize was a nontradable crop between 1990 and 2005; thus, import tariffs had only minor effects on domestic prices. The import tariffs on milk are high. In the 1970s and 1980s, the average pro- tection was estimated at 74 percent, but the extent of protection has diminished, and, in the first half of the present decade, protection averaged about 35 percent (Dorosh and Salem 2009). The Pakistan CGE Model This section summarizes the structure of the Pakistan CGE model; more details may be found in Cororaton and Orden (2008). It also discusses the method we use to introduce changes in the model to interface with the results generated by the global model. The model database representing the Pakistan economy is also summarized, along with the key parameters of the model. Structure of the national model The Pakistan CGE model of Cororaton and Orden (2008) is calibrated to the 2001­02 SAM constructed by Dorosh, Niazi, and Nazli (2004).2 The model has 34 production sectors in primary agriculture, lightly processed food, other manu- facturing, and services. There are five categories of productive factors: three types of labor (skilled labor, unskilled labor, and farm labor), as well as capital and land. There are also 19 household categories, a government sector, a firm sector, and the rest of the world. In the model, output (X) is a composite of value added (VA) and intermediate inputs. Output is sold to the domestic market (D) and may also be sold to the export market (E). Goods E and goods D are perfect substitutes. The sources of Pakistan 213 supply in the domestic market are domestic output and imports (M); substitution between D and M are dependent on the change in the relative prices of D and M and on the substitution parameter in a constant elasticity of substitution function. The primary factors of production in agriculture are unskilled labor (a com- posite of the labor of farmers and hired unskilled labor), land, and capital, while, in nonagriculture, they are skilled labor, unskilled labor, and capital. The on-farm labor of farmers is used only in primary agriculture. Other unskilled labor (including by farmers) is mobile across sectors and is employed in agricultural and nonagricultural sectors, while skilled labor is only mobile among nonagricul- tural sectors. Capital is fixed in each sector; so, the rates of return may differ across sectors.3 The use of land may shift among agricultural industries. The sources of household income are factors of production, transfers, foreign remittances, and dividends. Household savings are a fixed proportion of dispos- able income. According to the SAM, nonpoor urban households pay direct income taxes to the government, while other households do not. Household demand is specified as a linear expenditure system. The government obtains revenue from direct taxes on household and firm incomes, tariffs, other receipts, and indirect (consumption) taxes on domestic and imported goods. It spends on the consumption of goods and services, transfers, and other payments. We assume a fixed government fiscal balance in nominal terms. Tariff policy reforms result in changes in government income and expendi- ture, but the government balance is fixed through a tax replacement. We use a direct income tax replacement, but also compare the results under an adjustment via an indirect sales tax replacement on domestic consumption.4 In either case, the tax replacement is endogenously determined so as to maintain a fixed level of government balance. Foreign savings are also fixed. The numéraire is a weighted index of the price of value added in which the weights are the shares of the sectoral value added in the base calibration. The nominal exchange rate is flexible. Furthermore, we introduce a weighted price of investment and derive total investment in real prices. We hold total investment in real prices fixed by introducing an adjustment factor in the household savings function. Equilibrium in the model is achieved if the supply and demand of goods and services are equal and if investment is equal to savings. Linking the global model and the Pakistan model There are various methods for transmitting the results derived from a global CGE model to a single-country CGE model. Horridge and Zhai (2006) propose, for imports, the use of border price changes from the global model simulation of rest- of-the-world liberalization (that is, without Pakistan). For Pakistan's exports, their proposed scheme is as follows. 214 Agricultural Price Distortions, Inequality, and Poverty The export demand in the Pakistan model is PWE0 h E = E0 c d , (7.1) PWE where E is exports; PWE0 is international prices; PWE is the free on board (bor- der) prices of Pakistan's exports; h is the export supply elasticity, the value of which is equal to ESUBM (the Armington parameter in the global model); and E0 is the scale parameter in the demand function. Because exports and domestic goods are perfect substitutes, the export price in local currency is equal to the local price, which does not include indirect taxes. The change in the export demand shifter, E0, is derived as: E0 = 100 # (a - 1), (7.2) where a = (1 + 0.01p)([1 + 0.01q][1/ESUBM]), in which p is the change in the border export price, and q is the change in the export volume from the global model of rest-of-the-world liberalization, excluding Pakistan (Horridge and Zhai 2006). The purpose of the introduction of the export demand shift calculated in (7.2) is to allow the Pakistan model--not the simpler representation of Pakistan in the global model--to determine export supply behavior and the equilibrium prices and quantities of Pakistan's exports, taking into account the world demand shift in the global model. Economic structure in the SAM and key parameters in the Pakistan model Table 7.1 shows the sectoral structure of production and trade in the model based on the 2001­02 SAM. Of the 34 sectors, 12 are primary agricultural sectors (sectors 1 to 12), while sectors 14 to 18, which represent lightly processed food, are part of the broadly defined agricultural sector in this analysis. The nonagricultural sectors include the mining industry (sector 13), other food (sector 19), manufacturing industries (sectors 20 to 27), energy (sector 28), construction (sector 29), and five service sectors (sectors 30 to 34). Among these broad sectoral groupings, agriculture produces 27 percent of the sectoral value added and 28 percent of the gross value of sectoral output. In the SAM, it accounts for 13 percent of total employment.5 There are 19 household groups in the model. The agriculture-based groups are categorized according to household location (Punjab, Sindh, and other Pakistan) and the size of land holdings (large, medium, and small farms; landless small- farm renters, and agricultural workers without land). In addition, there are four nonfarm national aggregates: rural nonfarm poor, rural nonfarm nonpoor, urban poor, and urban nonpoor. Table 7.2 shows the 19 households in the SAM and the corresponding characteristics of the 19 household groups in the Household Integration Economic Survey. Table 7.1. Elasticity Parameters and Production Structure, Pakistan, 2001­02 Production Value added Value Employ- Land- Indirect Trade Share (%) ratio added Output Capital- ment output tax Elasticities Exports (%) Imports (%) va/x, share, share, labor share, Skilled Unskilled ratio, rate, Sectors % % % ratioa % labor labor Farmers % sig_va b % eta c sig_m d share intensity e share intensityf Agriculture 26.8 27.7 12.6 8.5 6.6 Primary agriculture 23.2 20.1 10.7 3.9 3.1 1. Wheat, irrigated 50.8 1.8 1.8 0.3 1.58 18.86 81.14 27.82 0.75 0.10 5.85 2.93 0.64 3.56 0.30 2.53 2. Wheat, nonirrigated 50.9 0.1 0.1 0.3 0.07 18.85 81.15 27.25 0.75 0.00 5.85 2.93 3. Paddy, IRRI 60.2 0.2 0.2 0.5 0.10 18.84 81.16 45.35 0.75 0.30 4.45 2.23 4. Paddy, basmati 60.2 0.5 0.4 0.5 0.12 18.86 81.14 51.27 0.75 0.00 4.45 2.23 5. Raw cotton 61.2 1.4 1.1 0.3 1.11 18.87 81.13 35.97 0.75 0.04 3.94 1.97 6. Sugarcane 60.0 1.0 0.8 0.7 0.32 18.87 81.13 46.68 0.75 0.07 5.91 2.96 7. Other major crops 71.0 2.8 2.0 0.3 2.42 18.87 81.13 38.88 0.75 0.05 3.94 1.97 0.52 2.65 0.60 4.53 8. Fruits and vegetables 64.2 3.6 2.8 0.6 1.75 18.87 81.13 44.37 0.75 0.34 3.94 1.97 1.05 3.78 1.31 6.94 9. Livestock, cattle and dairy 53.2 10.3 9.7 9.0 2.56 100.00 0.75 0.00 3.94 1.97 0.05 0.06 0.70 1.08 10. Poultry 51.6 0.7 0.7 9.0 0.18 100.00 0.75 0.00 3.94 1.97 0.01 0.11 11. Forestry 82.1 0.3 0.2 0.0 0.12 18.88 81.12 65.68 0.75 0.00 4.31 2.15 0.48 31.36 0.23 25.16 12. Fishing industry 57.1 0.6 0.5 2.3 0.41 100.00 0.75 0.00 4.31 2.15 1.14 23.79 0.00 0.08 Lightly processed food 3.6 7.6 1.8 4.6 3.4 14. Vegetable oil 7.9 0.2 1.4 6.7 0.07 60.28 39.72 1.50 0.02 3.94 1.97 0.00 0.02 2.33 19.99 15. Wheat, milled 21.8 1.2 2.8 4.4 0.56 64.94 35.06 1.50 0.02 4.45 2.22 0.51 1.82 0.82 4.31 16. Rice, milled IRRI 30.7 0.2 0.4 3.7 0.12 56.75 43.25 1.50 0.00 4.45 2.22 1.72 46.60 17. Rice, milled basmati 29.0 0.5 0.8 3.7 0.25 56.77 43.23 1.50 0.00 4.45 2.22 2.34 28.58 18. Sugar 32.2 1.4 2.2 3.3 0.82 69.96 30.04 1.50 6.75 5.91 2.96 0.03 0.11 0.28 1.89 215 (Table continues on the following pages.) 216 Table 7.1. Elasticity Parameters and Production Structure, Pakistan, 2001­02 (continued) Production Value added Value Employ- Land- Indirect Trade Share (%) ratio added Output Capital- ment output tax Elasticities Exports (%) Imports (%) va/x, share, share, labor share, Skilled Unskilled ratio, rate, Sectors % % % ratioa % labor labor Farmers % sig_va b % eta c sig_m d share intensity e share intensityf Nonagriculture 73.2 72.3 87.4 91.5 93.4 Mining and manufacturing 13.2 24.2 7.9 74.1 88.2 13. Mining 74.6 0.6 0.4 2.3 0.47 85.00 15.00 1.50 14.50 4.31 2.16 0.78 18.61 9.29 80.53 19. Other food 36.9 1.7 2.3 4.7 0.75 61.57 38.43 1.50 44.69 3.94 1.97 12.07 51.47 1.06 12.45 20. Cotton lint and yarn 21.6 1.4 3.3 3.3 0.82 85.52 14.48 1.50 12.05 4.11 2.06 8.97 27.10 0.71 4.27 21. Textiles 22.2 3.6 8.0 2.7 2.43 78.91 21.09 1.50 0.00 4.11 2.06 31.91 39.66 1.61 4.81 22. Leather 8.3 0.1 0.5 2.9 0.06 60.36 39.64 1.50 0.00 4.11 2.06 2.27 42.79 0.11 5.21 23. Wood products 36.3 0.7 0.9 1.8 0.57 67.98 32.02 1.50 0.18 4.09 2.04 0.03 0.34 0.56 8.61 24. Chemicals 28.2 0.5 0.9 3.8 0.25 55.16 44.84 1.50 2.58 4.09 2.05 1.38 15.90 11.16 69.89 25. Cement and bricks 55.0 1.4 1.3 7.4 0.42 68.98 31.02 1.50 24.15 4.09 2.05 0.03 0.21 26. Petroleum refining 19.4 0.6 1.5 2.9 0.36 71.95 28.05 1.50 28.96 4.09 2.05 9.71 50.11 27. Other manufacturing 25.4 2.6 5.0 2.6 1.75 67.99 32.01 1.50 4.39 4.09 2.05 16.61 33.17 54.00 71.03 Other industry 6.6 6.6 7.2 0.0 0.0 28. Energy 60.8 3.4 2.8 4.0 1.69 80.00 20.00 1.50 5.02 2.08 1.04 29. Construction 41.6 3.2 3.8 0.4 5.48 50.00 50.00 1.50 0.16 2.08 1.04 Services 53.5 41.5 72.4 17.5 5.2 30. Commerce 84.0 15.3 9.1 0.4 26.52 20.00 80.00 1.50 0.00 2.08 1.04 0.07 0.07 0.21 0.35 31. Transport 53.9 11.8 10.9 1.5 11.73 20.00 80.00 1.50 0.27 2.08 1.04 17.38 15.88 32. Housing 81.8 4.9 3.0 1.50 0.03 2.08 1.04 33. Private services 53.5 12.9 12.0 1.5 12.79 20.00 80.00 1.50 0.00 2.08 1.04 0.01 0.01 5.03 5.98 34. Public services 66.2 8.6 6.5 21.35 100.00 0.00 2.08 1.04 Total 49.8 100.0 100 100.0 100.0 10.0 100.0 14.45 Source: Dorosh, Niazi, and Nazli (2004). Note: va-value added; x-output. a. total labor total capital b. sig_va substitution parameter in CES production c. eta export demand elasticity d. sig_m substitution parameter in CES composite good e. export output f. imports composite good 217 218 Agricultural Price Distortions, Inequality, and Poverty Table 7.2. Model Household Categories, Pakistan 2001­02 social 2001­02 household accounting matrix integrated economic survey Large-farm households Landowners with more than (Punjab, Sindh, other Pakistan) 50 acres Medium-farm households Landowners with more than 12.5, (Punjab, Sindh, other Pakistan) but less than 50 acres Small-farm households Landowners with less than (Punjab, Sindh, other Pakistan) 12.5 acres Small-farm renters, landless No landholdings, but rent land for (Punjab, Sindh, other Pakistan) farm activities Rural agricultural workers, landless No landholdings, agricultural (Punjab, Sindh, other Pakistan) workers Rural nonfarm nonpoor Rural nonpoor, nonfarmers, nonagricultural workers Rural nonfarm poor Rural poor, nonfarmers, nonagricultural workers Urban nonpoor Urban nonpoor Urban poor Urban poor Sources: Dorosh, Niazi, and Nazli (2004); Federal Bureau of Statistics (2003). Note: The SAM covers three major provinces: Punjab, Sindh, and other Pakistan, which consists of Azad Kashmir, Balochistan, North-West Frontier Province, Northern Areas (Gilgit-Baltistan), Federally Administered, and Federally Administered Tribal Areas. The structure of consumption varies among household groups. A composite sector--livestock, cattle, and dairy--shows the highest share in the consumption basket, varying from 14 percent among large- and medium-farm households in other Pakistan to 25 percent among agricultural worker households in Punjab. The other major items in the consumption basket are private services (about 14 per- cent), transport (about 13 percent), milled wheat (from 4 percent among the urban nonpoor to 12 percent among agricultural workers in other Pakistan), textiles (from 5 percent among large- and medium-farm households in other Pakistan to 7 percent among agricultural worker households in Punjab and the urban poor), other manufacturing (from 1 percent among agricultural worker households in Sindh to 10 percent among large- and medium-farm households in other Pakistan), sugar (from 3 percent among urban nonpoor households to 10 percent among agricultural worker households in other Pakistan), and fruits and vegeta- bles (from 4 percent among large- and medium-farm households in Punjab to 7 percent among agricultural worker households in other Pakistan). Commodities with high foreign trade content are impacted significantly by changes in trade policies and world prices. This has varying effects across household groups because of differences in the household consumption bundles. Pakistan 219 The sectoral indirect tax structure is presented in table 7.1. The highest tax rate, 45 percent, applies to other food, the share of which in the consumption of house- holds is only about 1 percent. Indirect taxes are also relatively high on cement, bricks, and petroleum refining, which generally account for less than 1 percent of direct household consumption, but affect housing and transportation costs. The tax rate on cotton lint and yarn is 12 percent, while it is zero on textiles. However, because cotton lint and yarn are major inputs into textile production, an increase in the tax on the former raises the cost of production of textiles. This affects con- sumers because the share of textiles in the consumption basket is about 5 percent. Sugar shows the highest tariff rate, 59 percent (table 7.3). Another commodity showing high tariffs, averaging 55 percent, is livestock, cattle, and diary, which account for a large share in the consumption basket of households. Other agricul- tural commodities showing high tariffs and substantial consumption shares are milled wheat and vegetable oil. A few primary agricultural and light food process- ing sectors show low or even negative import tariffs. In contrast, tariffs are uni- formly relatively high across the manufacturing sectors. Overall, the foreign trade sector in Pakistan is not large relative to the domestic sector (table 7.1). Of the total domestic output, only 10 percent goes to the export market. Only 15 percent of the total goods and services available in the domestic market are imported. However, there are large differences across sectors. Within agriculture, the sectors with the highest share of production exported are milled IRRI rice, 47 percent; forestry, 31 percent; and fishing, 24 percent. The shares exported are small in the rest of the agricultural sectors. Within the nonagricultural sectors, other food, at 52 percent, has the highest share of production exported; leather is at 43 percent; textiles, 40 percent; and cotton lint and yarn, 27 percent. The textile sector dominates in exports. In the SAM, textiles account for 32 percent of total exports, cotton lint and yarn for 9 percent, and other food for 12 percent. Because of the large volume of crude oil imports, the mining sector has the high- est share of domestic consumption that is imported, at 81 percent. The share of other manufacturing is 71 percent; of chemicals, 70 percent; and of petroleum, 50 percent. Other manufacturing accounts for 54 percent of overall imports; chem- icals, 11 percent; and mining and petroleum refining, each about 9 percent. Except for forestry (25 percent) and vegetable oil (20 percent), the shares of domestic con- sumption that are imported in the agricultural sectors are well under 10 percent. Table 7.1 includes the values of key elasticity parameters in the model: the import substitution elasticity (sig_m) in the constant elasticity of substitution composite good function and the production substitution elasticity (sig_va) in the constant elasticity of substitution value added production function.6 The values of the export demand elasticity (eta) are the Armington parameters of the global model. The sources of household income in the model are labor income, capital income, income from land, and other income (table 7.4). Other income is Table 7.3. Parameters and Exogenous Demand and Price Shocks Caused by Rest-of-the-World Liberalization, Pakistan 220 Full trade Agricultural liberalization, excluding trade liberalization, Pakistan CGE model Linkage model Trade distortions Pakistan excluding Pakistan World import World price Export import Export Sector Sector Sector Tariff, Export change, demand price, % demand no. description description % tax, % % shifterb change shifterb Agriculture Primary agriculture 1. Wheat, irrigated Wheat 4.9 0.0 2.41 1.0 3.12 1.00 2. Wheat, nonirrigated Wheat 4.9 0.0 2.41 1.0 3.12 1.00 3. Paddy, IRRI Paddy rice 0.0 4.0 0.00 1.2 0.00 1.23 4. Paddy, basmati Paddy rice 0.0 4.0 0.00 1.2 0.00 1.23 5. Cotton Plant-based fibers 4.9 0.0 4.44 1.1 6.68 1.14 6. Sugarcane Sugarcane and beets 4.3 0.0 0.00 1.0 0.00 1.00 7. Other major crops Other crops 15.3 0.0 1.91 1.0 0.00 1.01 8. Fruits and vegetables Vegetables and fruits 16.5 0.0 2.93 1.0 2.62 0.97 9. Livestock, cattle, and Cattle, sheep, and so on 55.4 0.0 4.41 1.0 5.17 1.00 dairya 10. Poultry Other livestock 10.8 0.0 8.00 1.0 6.77 0.99 11. Forestry Other primary products 14.3 18.1 0.14 1.1 0.79 1.01 Lightly processed food 12. Fishing industry Other primary products 14.3 18.1 0.14 1.1 0.79 1.01 14. Vegetable oil Vegetable oils and fats 23.7 0.0 1.78 0.9 0.40 0.93 15. Wheat, milled Other food, beverages 24.9 0.0 0.00 1.0 1.68 0.97 and tobacco 16. Rice, milled IRRI Processed rice 0.0 4.0 8.21 1.1 10.18 1.08 17. Rice, milled basmati Processed rice 0.0 4.0 8.21 1.1 10.18 1.08 18. Sugar Refined sugar 59.0 0.0 1.62 1.0 3.44 1.00 Nonagriculture Mining and manufacturing 13. Mining Other primary products 14.3 18.1 0.14 1.1 0.79 1.01 19. Other food Other food, beverages 24.9 0.0 0.00 1.0 1.68 0.97 and tobacco 20. Cotton lint and yarn Textile and wearing 19.9 1.1 0.68 1.0 0.48 1.00 apparel 21. Textiles Textile and wearing 19.9 1.1 0.68 1.0 0.48 1.00 apparel 22. Leather Textile and wearing 19.9 1.1 0.68 1.0 0.48 1.00 apparel 23. Wood products Other manufacturing 16.6 5.4 0.38 1.0 0.38 1.00 24. Chemicals Other manufacturing 16.6 5.4 0.38 1.0 0.38 1.00 25. Cement and bricks Other manufacturing 16.6 5.4 0.38 1.0 0.38 1.00 26. Petroleum refining Other manufacturing 16.6 5.4 0.38 1.0 0.38 1.00 27. Other manufacturing Other manufacturing 16.6 5.4 0.38 1.0 0.38 1.00 Other industry 28. Energy Services 0.0 0.0 0.22 1.0 0.19 1.00 29. Construction Services 0.0 0.0 0.22 1.0 0.19 1.00 Services 30. Commerce Services 0.0 0.0 0.22 1.0 0.19 1.00 31. Transport Services 0.0 0.0 0.22 1.0 0.19 1.00 32. Housing Services 0.0 0.0 0.22 1.0 0.19 1.00 33. Private services Services 0.0 0.0 0.22 1.0 0.19 1.00 34. Public services Services 0.0 0.0 0.22 1.0 0.19 1.00 Source: Linkage model simulations by van der Mensbrugghe, Valenzuela, and Anderson (see the appendix). 221 a. This is the trade weighted average of cattle, sheep, other livestock, and dairy in the Linkage model. b. In equation 2, this is a (1 0.01 * p)(1 0.01 * q)^(1/ESUBM); where p is export price change, q export volume change; and ESBUM Arimington elasticity, all from the Linkage model. Table 7.4. Household Income Sources and Income Taxes, Pakistan, 2001­02 222 2001­02 income Income sources, % Per capita, 2001­02 population Labor Total PRs, Direct Households mil Rs 1,000s 1,000s % distribution Farm Unskilled Skilled Capital Land Other tax, % Large-farm households ­Sindh 19,079 112.9 169 0.12 13.2 0.3 -- 31.4 49.8 5.3 -- ­Punjab 64,116 173.7 369 0.25 8.6 0.5 -- 43.3 42.2 5.4 -- ­Other Pakistana 10,755 152.9 70 0.05 9.8 0.2 -- 52.4 32.3 5.4 -- Medium-farm households ­Sindh 44,625 30.4 1,466 1.00 14.5 3.1 -- 39.6 37.4 5.3 -- ­Punjab 145,995 48.4 3,014 2.07 10.8 4.3 -- 52.3 27.2 5.4 -- ­Other Pakistan 35,572 34.2 1,040 0.71 14.9 2.9 -- 38.2 38.4 5.5 -- Small-farm households ­Sindh 57,648 14.9 3,873 2.65 6.8 9.0 -- 57.9 20.4 5.8 -- ­Punjab 318,888 18.1 17,605 12.06 7.9 14.2 -- 51.8 20.3 5.8 -- ­Other Pakistan 124,985 11.9 10,493 7.19 6.0 11.6 -- 63.6 12.5 6.2 -- Small-farm renters, landless ­Sindh 43,672 7.7 5,682 3.89 11.6 18.7 -- 48.5 15.6 5.7 -- ­Punjab 45,963 10.7 4,307 2.95 9.0 20.5 -- 48.7 16.1 5.8 -- ­Other Pakistan 14,970 8.2 1,818 1.25 10.1 14.5 -- 55.0 14.7 5.7 -- Rural agricultural workers, landless ­Sindh 20,782 6.4 3,241 2.22 -- 51.0 -- 42.8 --- 6.2 -- ­Punjab 68,172 12.0 5,693 3.90 -- 49.3 -- 45.2 -- 5.5 -- ­Other Pakistan 9,513 14.6 653 0.45 -- 18.7 -- 76.0 -- 5.4 -- Rural nonfarm households ­nonpoor 400,771 19.8 20,233 13.86 -- 43.0 -- 49.9 -- 7.2 -- ­poor 134,398 5.5 24,525 16.80 -- 29.7 -- 63.4 -- 6.9 -- Urban households ­nonpoor 1,744,119 58.5 29,829 20.44 -- 11.9 33.3 10.6 -- 44.2 8.4 ­poor 181,413 15.3 11,880 8.14 -- 76.2 -- 18.0 -- 5.8 -- Sources: Dorosh, Niazi, and Nazli (2004). a. Other Pakistan--Azad Kashmir, Balochistan, North-West Frontier Province, Northern Areas (Gilgit-Baltistan), Federally Administered, and Federally Administered Tribal Areas. Pakistan 223 composed of foreign remittances, which is assumed in the SAM to be distributed proportionately among all households, and dividend income, which is earned only by urban nonpoor households. The sources of income vary across house- hold groups. Farm households are dependent on income from land, farm labor, and capital. Other rural households depend on income from unskilled labor and capital. About three-fourths of the income of urban poor households comes from unskilled labor. Urban nonpoor households derive 44 percent of their incomes from other income (composed largely of dividend income) and 33 per- cent from skilled labor income. According to the SAM, it is only the urban nonpoor household group that pays income taxes, amounting to 8.4 percent of their incomes. Poverty Indicators The overall poverty rate based on the official national poverty line declined in Pakistan from about 30 percent in the late 1980s to 26 percent in 1990­91. During these years, both urban and rural poverty decreased. However, in 1993­94, the incidence of rural poverty and the incidence of urban poverty started to move in dif- ferent directions: urban poverty continued to decline, while rural poverty began to rise, thereby widening the poverty gap between urban and rural areas (figure 7.1). The gap reached a peak in 2001­02, which was largely caused by the crippling drought that severely affected agricultural output that year, together with rela- tively low international agricultural commodity prices. Almost 70 percent of the people live in rural areas, and, because the majority of these people (40 percent of all households nationally) depend on agriculture for income, the incidence of rural poverty increased to 39 percent that year, while urban poverty was stable at 23 percent. There is some disagreement about the accuracy of recent estimates of poverty. For 2004­05, the estimates of the Planning Commission of Pakistan show the overall incidence of poverty declining from the peak of 34 percent in 2001­02 to 24 percent in 2004­05 (Dorosh and Salam 2009). The World Bank (2007) esti- mates a smaller decline, to 29 percent. Despite the disparity between these estimates (arising primarily because of the inflation factor used in computing the relevant poverty lines), the estimates show the incidence of poverty declining in urban and rural areas in the most recent years, but highlight the large gap between the rural and urban poverty rates. The depth of poverty indicated by the Foster, Greer, and Thorbecke (1984) poverty gap and the squared poverty gap also suggest that the problem of poverty is more severe in rural areas and that this was especially true during the 2001­02 drought (table 7.5). 224 Agricultural Price Distortions, Inequality, and Poverty Figure 7.1. Poverty Incidence, Pakistan, 1986­87 to 2004­05 45 40 35 poverty headcount, % 30 25 20 15 10 5 0 1986­87 1987­88 1990­91 1992­93 1993­94 1996­97 1998­99 2001­02 2004­05 Pakistan urban rural Sources: Ministry of Finance (2003); World Bank (2007) for the 2004­05 estimates. Note: Poverty is defined as the share of the population living below the national poverty line. Official data for 1993­94 indicate that the level of overall poverty was above the incidence of poverty in urban and rural areas; see Ministry of Finance (2003). Table 7.5. Poverty Estimates, Pakistan, 1998­99, 2001­02, and 2004­05 Poverty index 1998­99 2001­02 2004­05 Poverty incidence Pakistan 30.0 31.2 29.2 Urban 21.0 19.9 19.1 Rural 33.8 38.2 34.0 Poverty gap Pakistan 6.3 6.5 6.1 Urban 4.3 3.9 3.9 Rural 7.1 8.0 7.2 Squared poverty gap, poverty severity Pakistan 2.0 2.0 2.0 Urban 1.3 1.2 1.2 Rural 2.2 2.5 2.3 Sources: Ministry of Finance (2003); World Bank (2007) for the 2004­05 estimates. Note: The table reflects the poverty concepts of Foster, Greer, and Thorbecke (1984). Pakistan 225 Simulations The first part of this section describes our six policy experiments, while the second part discusses the results. The experiments use direct tax replacement to hold the government fiscal balance fixed. The idea is to replace distorting trade taxes with less-distorting income taxes. The fiscal burden falls on urban nonpoor house- holds because, according to the SAM, other household groups do not pay income taxes (table 7.4). We have also conducted an alternative indirect tax replacement experiment to check the sensitivity of the results to the specification, given that financing a trade reform is a nontrivial issue from the government's point of view (Ahmed, Abbas, and Ahmed 2009). In our analysis, we separate the effects on the economy of reducing distortions in the rest of the world and in domestic markets in Pakistan, and we evaluate the effects of both on income inequality and poverty. Design of the policy experiments Table 7.3 shows the sectoral correspondence between the Pakistan model and the global model. It also shows the sectoral tariff rates and export taxes, which, where possible, are based on the set of estimates on nominal rates of assistance for Pakistan by Dorosh and Salam (2009). We use these trade distortions in all our policy experiments. The table presents the changes in border import prices found through the global model under full trade liberalization and agricultural liberal- ization by the rest of the world, and it lists sectoral export demand shifters calcu- lated on the basis of equation (7.2). We have used these data as inputs in the six policy experiments we have conducted, as follows: · S1A, full world trade liberalization in all tradable goods sectors by all countries, excluding Pakistan: This experiment uses the results of the global model under full trade liberalization in table 7.3. It retains all existing trade distortions in Pakistan. · S1B, agricultural price and trade liberalization by all countries, excluding Pakistan: This scenario uses the results of the global model. As with S1A, all existing distortions in Pakistan are retained. · S2A, full goods trade liberalization carried out unilaterally in Pakistan: All Pakistan trade distortions are set to zero. There are no changes in the sectoral border export and import prices or in the export demand shifters because there is no rest-of-the-world trade liberalization. · S2B, agricultural trade liberalization carried out unilaterally in Pakistan: All Pakistan distortions in primary agriculture and in lightly processed food are set to zero. Similar to S2A, there are no changes in the sectoral border export and 226 Agricultural Price Distortions, Inequality, and Poverty import prices and in the export demand shifters because there is no rest-of- the-world trade liberalization. · S3A, full world trade liberalization of all tradable goods, including in Pakistan: This combines S1A and S2A. · S3B, agricultural world trade liberalization, including in Pakistan: This combines S1B and S2B. In analyzing the results under each of the scenarios, we indicate, first, the effects on poverty for the whole of Pakistan, for rural and urban areas, and for major household groups. The results on poverty include changes in the incidence of poverty and in the depth of poverty measured by the poverty gap and the squared poverty gap. These poverty effects are traced and analyzed through the various determining channels: macro, sectoral, commodity, and factor prices, as well as household incomes. In estimating the poverty effects, we apply the results on household incomes and consumer prices for each of the 19 household groups in the CGE model simulations to the household classifications in the Household Integration Economic Survey. Each of the CGE simulations generates a new vec- tor of household income and consumer price for each of the groups, which we use to compute new sets of poverty indexes to compare with the baseline indexes. Simulation results In this subsection, we present the modeling results of our six policy experiments. We continue the discussion with additional results that show the sensitivity of the core results to changes in the treatment of tax adjustments in the model. S1A: Trade liberalization by the rest of the world without Pakistan Full trade liberalization abroad, while retaining all existing trade distortions in Pakistan, causes the overall poverty incidence index to decline by 1.3 percent from its base value, from 31.2 to 30.8 (table 7.6). Households at the bottom of the income ladder benefit the most, as indicated by the greater reduction in the poverty gap (1.6 percent) and the squared poverty gap (1.9 percent). Among rural house- holds, the poorest households--those in the rural nonfarmer group--benefit the most. Thus, rural-urban income inequality is lowered in this scenario also. What are the forces that drive these reductions in poverty and income inequal- ity? The S1A simulation leads to a real exchange rate appreciation of 1.24 percent (table 7.7).7 The terms of trade (the ratio of domestic export to import prices) improve by 1.38 percent in agriculture and by 1.56 percent in nonagriculture. This is because of the lower world import prices of some of the agricultural products, as well as most of the nonfarm goods (table 7.3). Table 7.6. The Poverty Effects of Prospective Liberalizations, Pakistan 2001­02 % change from 2001­02 index Index in population Indicator 2001­02 distribution, % S1A S1B S2A S2B S3A S3B All Pakistan ­P0 31.23 1.3 0.1 5.2 1.6 6.4 1.8 ­P1 6.46 1.6 0.1 10.0 2.3 11.5 2.5 ­P2 1.97 1.9 0.1 12.1 2.8 13.8 3.0 Gini coefficient 0.34 0.08 0.02 3.3 0.1 3.3 0.2 All urban ­P0 19.86 29.7 1.5 0.1 2.3 2.4 0.4 2.7 ­P1 3.91 1.5 0.1 13.3 3.6 14.8 3.7 ­P2 1.16 1.7 0.1 16.9 4.2 18.3 4.3 Gini coefficient 0.40 0.03 0.01 2.0 0.1 2.0 0.1 All rural ­P0 38.16 70.3 1.2 0.1 7.6 1.4 8.6 1.5 ­P1 8.02 1.6 0.1 9.0 2.0 10.5 2.1 ­P2 2.47 2.0 0.1 10.8 2.4 12.5 2.6 Gini coefficient 0.26 0.01 0.02 0.2 0.1 0.2 0.1 Large- and medium-farm households ­P0 22.82 4.0 0.0 0.0 7.9 3.4 7.9 3.4 ­P1 4.06 0.7 0.4 9.9 4.9 10.0 5.3 ­P2 1.13 0.7 0.5 10.5 5.2 10.6 5.8 Small-farm households and agricultural workers ­P0 37.40 30.6 1.4 0.1 8.3 0.9 8.7 0.9 ­P1 7.47 1.3 0.0 9.3 1.3 10.3 1.3 ­P2 2.20 1.6 0.0 11.2 1.7 12.4 1.6 Rural nonfarm households ­P0 39.89 35.7 1.1 0.2 7.2 1.9 8.5 2.1 ­P1 8.71 1.9 0.2 8.9 2.6 10.6 2.8 ­P2 2.76 2.2 0.2 10.6 3.1 12.6 3.4 Source: National CGE model simulations by the authors. 227 Note: P0 poverty headcount. P1 poverty gap. P2 poverty severity. 228 Table 7.7. Aggregate Simulation Results of Prospective Agricultural and Nonagricultural Liberalization, Pakistan (% change from the base) S1A S1B S2A S2B S3A S3B Variables Ag Nonag Ag Nonag Ag Nonag Ag Nonag Ag Nonag Ag Nonag Real GDP 0.15 0.04 0.81 0.26 0.95 0.29 Real and relative prices Real exchange rate 1.24 0.13 6.09 1.87 4.79 1.73 Domestic terms of tradea 1.38 1.56 0.98 0.13 12.61 9.37 17.30 0.84 14.19 11.06 18.52 0.71 Prices Export price in local currency 0.26 0.82 0.33 0.13 1.92 0.09 0.63 1.70 2.16 0.66 0.30 1.59 Import price in local currency 1.70 2.35 0.73 0.24 12.26 8.34 14.49 2.57 13.74 10.41 15.13 2.33 Domestic price 0.29 0.55 0.07 0.08 2.47 1.23 1.53 1.11 2.74 1.69 1.56 1.06 Output price 0.29 0.58 0.06 0.08 2.45 1.07 1.50 1.18 2.73 1.57 1.52 1.12 Value added price 0.16 0.12 0.03 0.06 3.14 1.32 2.88 1.01 2.98 1.24 2.79 0.97 Consumer price index 0.51 0.11 2.81 0.27 3.24 0.34 Volume Imports 2.41 0.96 0.89 0.12 22.15 4.45 29.64 0.48 24.75 5.46 30.42 0.36 Exports 0.73 0.16 1.79 0.11 4.78 13.24 6.20 3.07 5.28 13.07 8.00 3.17 Domestic demand 0.09 0.07 0.01 0.02 0.87 1.12 1.05 0.03 0.76 1.19 1.06 0.01 Composite good 0.18 0.12 0.02 0.01 0.09 0.08 0.20 0.07 0.30 0.05 0.22 0.06 Output 0.11 0.08 0.05 0.01 0.69 0.39 0.82 0.39 0.57 0.50 0.77 0.38 Value added 0.07 0.03 0.03 0.01 0.51 0.19 0.39 0.13 0.43 0.16 0.35 0.12 Factor prices Farm wages 0.95 0.27 0.58 2.16 1.67 2.44 Wages of unskilled labor 0.14 0.05 1.49 0.47 1.38 0.45 Wages of skilled labor 0.04 0.00 2.06 0.73 2.04 0.74 Return to land 1.00 0.32 1.90 2.52 3.06 2.85 Return to capital 1.00 0.14 0.27 0.09 5.02 1.00 3.80 1.41 3.92 0.93 3.41 1.34 Factor prices less inflation Farm wages 0.44 0.16 2.23 1.88 1.56 2.10 Wages of unskilled labor 0.37 0.05 4.30 0.74 4.61 0.79 Wages of skilled labor 0.47 0.11 4.87 1.00 5.28 1.09 Return to land 0.49 0.21 0.91 2.25 0.18 2.51 Return to capital 1.51 0.37 0.37 0.02 2.21 3.81 3.53 1.68 0.69 4.17 3.07 1.68 Source: National CGE model simulations by the authors. Note: Ag agriculture. Nonag nonagriculture. a Change in the ratio of domestic export and import prices. 229 230 Agricultural Price Distortions, Inequality, and Poverty The import prices of agricultural goods drop by 1.7 percent (table 7.7), despite the increases in the import prices of livestock, wheat, vegetable oil, and sugar (table 7.3). This is caused by a number of factors, including the appre- ciation of the currency; the decline in the border import prices of fruits and vegetables and other major crops (table 7.3), all of which have relatively large import components (table 7.1); and the slight reduction in the border import prices of forestry products, which show high import intensity. The domestic prices of farm products decline by 0.3 percent, which is lower than the drop in import prices. This results in more imports of agricultural goods (a rise of 2.4 percent) and a marginal increase in the domestic demand for agriculture (a rise of 0.1 percent). Because the demand for imported and domestic agricul- tural products increases, the domestic consumption of farm products increases (by 0.2 percent). Table 7.3 shows that the border import prices of nonagricultural goods decline. This, together with the appreciation of the exchange rate, reduces the import prices of nonagricultural goods by 2.4 percent (table 7.7). The domestic prices of nonagricultural products also decline, by 0.6 percent, which is lower than the decline in the import prices. Thus, the imports of nonfarm products increase, by 1.0 percent. At the sectoral level, there is a relatively large rise in imports of cotton lint and yarn, textiles, and leather because of the relatively greater decline in the border prices of these products. More imports of nonfarm goods reduce margin- ally the domestic demand, but, despite this, overall domestic nonagricultural consumption increases by 0.1 percent. The export prices of farm products decline by 0.3 percent. Because the border prices of these products rise, the decline is caused by appreciation in the exchange rate. There is a slightly greater decrease in the domestic prices of agricultural products. Thus, exports of agriculture improve by 0.73 percent, and the overall output of agriculture increases by 0.11 percent. The effects on value added, value added prices, and factor prices in agriculture are explained by the changes in sectoral export prices, factor intensities, and import and export intensities. The overall output price of agriculture declines by 0.29, while the value added price increases by 0.16 percent. The difference in the sign is generated by a relatively higher increase in the value added price of milled rice (2 percent) and vegetable oil (1.7 percent).8 The increase by 1.18 percent in the border export price of milled rice has larger effects on the value added price of milled rice because rice has a high export-intensity ratio (table 7.1). Although the border import price of milled rice increases more (10.18 percent), this has no effects because there are zero imports. The increase by 1.78 percent in the import border price of vegetable oil increases the value added price of vegetable oil because vegetable oil has a high import-intensity ratio. Pakistan 231 Farm wages and the returns to land decline by about 1.0 percent. This is caused by the decline in the output and value added prices in primary agriculture, which employs farmers and uses land. The average rate of return to capital in agriculture improves by 1 percent. This is caused by the increase in the value added price of milled rice and vegetable oil. These sectors are relatively capital intensive, showing capital-labor ratios of 3.7 for rice and 6.7 for vegetable oil (table 7.1). Because wage rates increase less than value added prices, returns to capital rise. The returns to capital in these sectors increase by more than 2 percent for milled rice and 1.9 percent for vegetable oil. The change in the returns to capital in livestock and poultry is also positive, but smaller. The change in the returns to capital in the other primary agricultural commodities is negative. The decline in the value added prices in primary agriculture and in nonagri- culture lowers the wages of unskilled labor by 0.14 percent. However, because of the increase in the value added prices of milled rice and vegetable oil, the wages of skilled workers decrease by only 0.04 percent. The average returns to the capital used in nonagriculture decline by 0.14 percent. We have also included the results on factor prices net of the inflation effects. The overall consumer price index in this experiment decreases by 0.5 percent. Net of the inflation effects, there is a negative result for farm wages and the returns to land, but the other factors show positive net price effects. All these effects lead to changes in household incomes (table 7.8). The changes in the nominal incomes of households are negative across groups except rural nonfarmers and rural agricultural workers, the latter because of their heavy reliance on agricultural capital incomes (mostly informal capital), as shown in table 7.4, and the increase in the average returns to capital in agriculture (1 per- cent; see table 7.7). However, the consumer prices for each of the groups decline more rapidly than the drop in nominal incomes because of the greater reduction in import prices. Thus, all household groups realize improvements in real incomes. The largest increases in real incomes occur among rural nonfarmers (0.63 and 0.53 for the nonpoor and poor, respectively; see table 7.8) and agricul- tural workers in other Pakistan (0.58 percent). This explains the significant reduc- tion in the depth of poverty in rural areas, particularly among rural nonfarmers. In the scenario, full trade liberalization by the rest of the world thus reduces both poverty and income inequality. It reduces import prices, especially on com- modities that have relatively large shares in the consumption basket of consumers. This translates into declining consumer prices. It also enhances agricultural pro- duction relative to nonagricultural production because of improvements in the world price of farm commodities. The poorest nonfarm households in rural areas benefit the most from the improvement in the real wages of unskilled labor and in the returns to capital and from the reduction in consumer prices. 232 Agricultural Price Distortions, Inequality, and Poverty Table 7.8. Household Welfare and Price Effects of the Liberalization in All Goods Trade by the Rest of the World, Pakistan S1A S1B 2001­02 Change in Change in EV/ Change in Change in EV/ Change in population nominal consumer income, nominal consumer income, nominal Households distribution, % income, % price, % % income, % price, % % income, % Large-farm households ­Sindh 0.14 0.49 0.51 0.02 0.24 0.08 0.16 0.53 ­Punjab 0.17 0.35 0.58 0.23 0.20 0.09 0.11 0.48 ­Other Pakistan 0.01 0.24 0.67 0.43 0.17 0.13 0.04 0.39 Medium-farm households ­Sindh 0.99 0.46 0.51 0.05 0.20 0.08 0.11 0.61 ­Punjab 1.89 0.30 0.58 0.28 0.15 0.09 0.06 0.50 ­Other Pakistan 0.75 0.47 0.67 0.20 0.20 0.13 0.07 0.62 Small-farm households ­Sindh 2.79 0.19 0.44 0.25 0.12 0.06 0.06 0.33 ­Punjab 12.87 0.22 0.45 0.23 0.13 0.09 0.04 0.20 ­Other Pakistan 5.67 0.10 0.46 0.36 0.09 0.11 0.01 0.19 Small-farm renters, landless ­Sindh 0.16 0.22 0.38 0.16 0.12 0.07 0.06 0.03 ­Punjab 1.50 0.20 0.47 0.27 0.12 0.10 0.02 0.00 ­Other Pakistan 0.58 0.18 0.52 0.35 0.11 0.12 0.01 0.13 Rural agricultural workers, landless ­Sindh 4.12 0.01 0.38 0.37 0.06 0.09 0.03 0.87 ­Punjab 2.02 0.00 0.40 0.39 0.06 0.09 0.03 0.82 ­Other Pakistan 0.86 0.10 0.48 0.58 0.04 0.13 0.09 0.07 Rural nonfarm households ­nonpoor 17.60 0.11 0.52 0.63 0.06 0.11 0.04 0.94 ­poor 18.14 0.06 0.46 0.53 0.05 0.10 0.06 0.35 Urban households ­nonpoor 22.50 0.22 0.53 0.31 0.11 0.12 0.01 7.21 ­poor 7.23 0.09 0.46 0.37 0.08 0.11 0.03 1.49 Source: National CGE model simulations by the authors. Note: EV equivalent variation. S1B: Agricultural liberalization by the rest of the world This second experiment incorporates the results of the global model for agricul- tural liberalization by the rest of the world, while retaining all existing trade distor- tions in Pakistan. Compared with scenario S1A, the border import prices of some commodities increase more in this scenario. For example, there is a greater increase in the border import prices of wheat, livestock, cotton, milled rice, and sugar (table 7.3). Furthermore, the border import prices of nonagricultural prod- ucts increase in this scenario, while they decline in scenario S1A (table 7.7). Also, for commodities that show declining border import prices, the drop is relatively greater Pakistan 233 S2A S2B S3A S3B Change in EV/ Change in Change in EV/ Change in Change in EV/ Change in Change in EV/ consumer income, nominal consumer income, nominal consumer income, nominal consumer income, price, % % income, % price, % % income, % price, % % income, % price, % % 2.60 2.08 1.63 0.28 1.34 1.09 3.03 1.96 1.84 0.32 1.51 2.82 2.35 1.34 0.16 1.17 0.86 3.31 2.47 1.51 0.22 1.29 2.94 2.57 1.11 0.03 1.08 0.64 3.52 2.90 1.25 0.12 1.12 2.60 2.00 1.21 0.28 0.93 1.11 3.03 1.94 1.39 0.32 1.06 2.82 2.33 0.87 0.16 0.71 0.81 3.31 2.52 0.99 0.22 0.77 2.94 2.33 1.25 0.03 1.22 1.14 3.52 2.40 1.43 0.12 1.30 2.40 2.08 0.60 0.40 0.19 0.51 2.78 2.28 0.68 0.43 0.25 2.47 2.28 0.59 0.53 0.06 0.41 2.84 2.44 0.68 0.57 0.10 2.37 2.19 0.37 0.53 0.16 0.26 2.77 2.52 0.42 0.60 0.18 2.16 2.14 0.53 0.56 0.03 0.24 2.48 2.25 0.62 0.59 0.03 2.41 2.42 0.48 0.52 0.05 0.19 2.82 2.64 0.57 0.59 0.03 2.41 2.29 0.50 0.44 0.05 0.29 2.87 2.59 0.58 0.53 0.04 2.08 2.96 0.26 0.68 0.95 0.91 2.41 3.33 0.24 0.73 0.98 2.19 3.02 0.25 0.71 0.97 0.87 2.52 3.40 0.23 0.76 1.00 2.20 2.28 0.10 0.60 0.71 0.24 2.62 2.87 0.11 0.70 0.82 2.68 3.63 0.10 0.30 0.41 1.12 3.12 4.25 0.09 0.37 0.47 2.23 2.58 0.16 0.57 0.73 0.47 2.63 3.11 0.16 0.64 0.80 3.13 4.06 0.25 0.11 0.13 7.27 3.57 3.69 0.32 0.19 0.12 2.44 3.94 0.38 0.51 0.90 1.44 2.83 4.28 0.34 0.58 0.93 in this scenario compared with scenario S1A. Thus, the increase in the terms of trade for both agriculture and nonagriculture is lower in this experiment compared with scenario S1A. Also, the increase in the terms of trade is significantly lower in nonagriculture than in agriculture in this scenario relative to scenario S1A. Table 7.6 shows that, while Pakistan's overall poverty incidence index declines marginally, poverty is not reduced across the board. Poverty in urban areas declines, but not all rural households experience a drop in poverty. Rural non- farmers show the highest poverty reduction, but, among farmers and agricultural workers, there is a slight increase in poverty. 234 Agricultural Price Distortions, Inequality, and Poverty What are the factors that drive these poverty results? The import prices of agri- culture decline by 0.7 percent (table 7.7). This is caused by the real exchange rate appreciation of 0.13 percent and the reduction in the border prices of milled wheat, as well as fruits and vegetables, which are import intensive. There are a number of primary agricultural commodities that show relatively greater increases in import prices, but these commodities are not imported. The domestic prices of agricultural goods decrease, but by less than the decline in the import prices of these goods. Thus, the imports of agricultural goods increase by 0.9 percent. In nonagriculture, the smaller decline in domestic prices relative to import prices leads to a marginal increase in imports by 0.12 percent. This raises slightly the domestic consumption of nonagricultural products. The increase in the export prices in agriculture by 0.33 percent and the decline in the domestic prices in agriculture by 0.07 percent result in a rise in exports by 1.8 percent. This increases the overall output of agriculture slightly, despite the decline in the domestic demand because of more imports. However, the increase in the exports of nonagricultural goods is not quite sufficient to offset the decline in domestic demand; so, the overall output of nonagriculture declines by 0.01 percent. The difference in the results between the prices of value added and the output in agriculture is caused by variations in the results across industries within the agricultural sector. The greater increase in the border price of milled rice leads to a higher value added price, which offsets the decline in value added prices in the rest of agriculture. The decline in farm wages by 0.27 percent and the decline in the returns to land by 0.32 percent are caused by the decrease in the value added prices in primary agriculture. There is an increase in the returns to capital in agri- culture by 0.27 percent mainly because of the improvement in the value added price of milled rice, a sector which has a high capital-labor ratio. The decline in wages in unskilled labor is smaller than the decline in farm wages because of the increase in the value added price of milled rice, which neutralizes much of the effect of the falling value added prices in the rest of agriculture and in some non- agricultural sectors. Because milled rice requires more skilled labor than unskilled labor (table 7.1), the increase in the value added price of milled rice also offsets the negative effects arising in the rest of the economy; thus, the wages of skilled labor do not change. Net of the inflation effects, the changes in factor prices involve declining farm wages and declining returns to land, but otherwise rising factor prices. The nomi- nal income effects are negative among all household groups (table 7.8), but they are smaller than the effects generated in scenario S1A. Consumer prices decline, but the decline is not sufficient to offset the drop in the nominal incomes of farm- ers. However, rural nonfarmers and urban households enjoy marginal improve- ments in real incomes. Pakistan 235 In sum, agricultural liberalization by the rest of the world would generate a marginal change in the terms of trade that favors agriculture compared with the results of scenario S1A. Furthermore, although overall import prices decline, the drop is much smaller in the present case than in scenario S1A. This translates into a smaller decline in consumer prices across household groups that is not sufficient to offset the drop in nominal incomes in some groups. These groups--farmers and agricultural workers--experience a slight increase in poverty. Moreover, given the small share of agriculture in the overall trade of Pakistan (less than 10 percent; see table 7.1), a liberalization only in agriculture has much less impact on the Pakistan economy than a liberalization in all goods trade. Thus, the poverty impact is significantly less in the present case than in scenario S1A. S2A: The unilateral liberalization of all goods trade by Pakistan Our third experiment sets all sectoral import tariffs and export taxes in Pakistan to zero and assumes that there are no changes in policies abroad. Table 7.6 shows that this generates a significant drop in poverty, by 5.2 percent overall. There is also a significant reduction in the depth of poverty: the poverty gap drops by 10 percent, and the squared poverty gap drops by 12 percent. However, the inci- dence of poverty in urban areas increases by 2.3 percent. The detailed results show that the urban nonpoor suffer a decline in incomes because of the additional tax burden. This is the outcome of the tax replacement method, whereby we replace trade-distorting taxes in Pakistan with a less-distorting income tax that falls disproportionately on urban nonpoor households.9 The rest of the household groups enjoy higher incomes and therefore lower poverty. Overall, income inequality is also reduced. Most of the effects arise from the elimination of tariffs, although there are also effects because of the dismantling of the export taxes in a number of sectors (table 7.3). The elimination of tariffs leads to a substantial reduction in import prices. The greatest reduction is in sugar and in livestock, cattle, and dairy because these sectors are exposed to the highest tariff rates. The import prices of vegetable oil, milled wheat, other food, cotton lint and yarn, and textiles also decline notably (table 7.9). Overall, the import prices in agriculture decline by 12 percent, while, in nonagriculture, the import prices decline by 8.3 percent (table 7.7). Domestic prices also decline. However, the decline in domestic prices in most sectors is less than the decline in import prices. Thus, imports in these sectors surge. The imports of sugar increase by 215 percent; livestock, cattle, and dairy, by 99 percent; milled wheat, by 40 percent; other food, by 38 percent; cotton lint and yarn, by 38 percent; textiles, by 33 percent; and leather, by 28 percent (table 7.9). Other sectors show notable increases as well. Overall, imports in agriculture rise by 22 percent compared with only 4 percent in nonagriculture. 236 Table 7.9. The Sectoral Effects of the Unilateral Liberalization of All Goods Trade, Pakistan Domestic Composite Value Output demand good Exports Imports added Return to Sectors x px d pd q pq e pe m pm va pva capital Agriculture Primary agriculture 1. Wheat, irrigated 2.2 2.2 0.6 2.2 0.4 1.9 58.0 2.2 30.9 11.2 2.2 0.9 2.0 2. Wheat, nonirrigated 0.6 2.5 0.6 2.5 0.6 2.5 0.0 0.0 0.0 0.0 0.6 1.3 2.1 3. Paddy, IRRI 0.7 2.5 0.7 2.5 0.7 2.5 0.0 0.0 0.0 0.0 0.7 1.5 0.5 4. Paddy, basmati 0.9 2.6 0.9 2.6 0.9 2.6 0.0 0.0 0.0 0.0 0.9 1.7 0.5 5. Raw cotton 4.3 2.6 4.3 2.6 4.3 2.6 0.0 0.0 0.0 0.0 4.3 0.8 5.0 6. Sugarcane 1.3 2.7 1.3 2.7 1.3 2.7 0.0 0.0 0.0 0.0 1.3 1.8 3.5 7. Other major crops 0.2 1.6 1.0 1.6 0.4 1.9 32.6 1.6 13.9 8.3 0.2 1.3 1.5 8. Fruits and vegetables 0.8 1.5 2.0 1.5 0.4 2.3 32.2 1.5 15.2 9.3 0.8 1.7 2.7 9. Livestock, cattle, and dairy 0.5 3.3 0.6 3.3 0.1 3.6 41.8 3.3 99.0 32.0 0.5 5.0 5.7 10. Poultry 0.6 3.6 0.7 3.6 0.7 3.6 43.9 3.6 0.0 0.0 0.6 5.7 6.5 11. Forestry 21.2 1.9 3.3 1.9 0.3 3.2 41.7 1.9 9.9 7.5 21.2 1.6 0.0 12. Fishing industry 4.9 9.2 0.6 9.2 0.6 9.2 18.5 9.2 3.4 7.5 4.9 13.7 19.3 Lightly processed food 14. Vegetable oil 6.8 3.8 6.8 3.8 0.1 7.2 44.8 3.8 17.7 14.5 6.8 34.9 37.9 15. Wheat, milled 0.6 1.2 1.1 1.2 0.2 1.7 34.8 1.2 39.6 15.4 0.6 0.1 0.3 16. Rice, milled IRRI 1.5 0.7 0.1 0.7 0.1 0.7 3.3 0.7 0.0 0.0 1.5 5.5 6.5 17. Rice, milled basmati 1.3 0.3 0.3 0.3 0.3 0.3 5.2 0.3 0.0 0.0 1.3 5.0 5.9 18. Sugar 1.4 1.5 1.4 1.5 0.0 2.0 51.5 1.5 215.2 33.5 1.4 1.2 2.1 Nonagriculture Mining and manufacturing 13. Mining 4.0 8.6 1.1 8.6 1.3 7.5 21.0 8.6 1.9 7.3 4.0 10.4 15.1 19. Other food 2.6 2.9 6.0 2.9 1.1 0.2 11.3 2.9 38.1 15.4 2.6 10.3 12.2 20. Cotton lint and yarn 5.0 1.7 2.5 1.7 3.7 1.1 12.0 1.7 37.5 11.8 5.0 13.8 17.6 21. Textiles 4.1 1.6 0.9 1.6 0.8 0.8 12.3 1.6 32.8 11.8 4.1 9.6 12.6 22. Leather 8.9 0.5 0.2 0.5 1.1 1.2 22.7 0.5 27.8 11.8 8.9 20.0 27.0 23. Wood products 1.8 1.7 1.8 1.7 0.5 2.3 7.2 1.7 15.9 9.3 1.8 0.3 1.5 24. Chemicals 2.5 4.6 6.7 4.6 0.6 8.0 21.2 4.6 3.6 9.3 2.5 4.6 6.2 25. Cement and bricks 1.4 4.7 1.4 4.7 1.4 4.7 21.7 4.7 0.0 0.0 1.4 4.8 5.7 26. Petroleum refining 3.8 4.8 3.8 4.8 1.1 7.1 0.0 0.0 6.3 9.3 3.8 5.5 7.9 27. Other manufacturing 0.8 3.5 7.9 3.5 1.0 7.7 15.5 3.5 4.7 9.3 0.8 0.5 0.0 Other industry 28. Energy 0.3 0.5 0.3 0.5 0.3 0.5 0.0 0.0 0.0 0.0 0.3 2.7 2.8 29. Construction 1.7 2.9 1.7 2.9 1.7 2.9 0.0 0.0 0.0 0.0 1.7 1.3 0.1 Services 30. Commerce 0.4 0.5 0.4 0.5 0.4 0.5 11.1 0.5 5.5 5.7 0.4 1.5 1.2 31. Transport 1.7 1.3 0.6 1.3 0.6 1.3 15.2 1.3 0.0 0.0 1.7 3.4 4.6 32. Housing 0.0 11.7 0.0 11.7 0.0 11.7 0.0 0.0 0.0 0.0 0.0 13.6 13.6 33. Private services 0.4 0.7 0.4 0.7 0.8 0.3 13.9 0.7 6.7 5.7 0.4 1.2 0.9 34. Public services 0.7 0.5 0.7 0.5 0.7 0.5 0.0 0.0 0.0 0.0 0.7 2.1 0.0 Source: National CGE model simulations by the authors. Note: x output. px output price. d domestic demand. pd domestic price. q composite good. pq composite price. e exports. pe export price. m imports. pm import price. va value added. pva value added price. 237 238 Agricultural Price Distortions, Inequality, and Poverty Because world prices are fixed, the decline in prices in Pakistan because of the trade reform increases the country's competitiveness.10 There is a real depreci- ation of 6.1 percent in the exchange rate. The results, shown in table 7.9, indicate that, except for forestry and fishing, the exports of agriculture--primary agricul- ture and lightly processed food--improve. Overall, agricultural exports expand by 4.8 percent. However, this increase does not offset the displacement effects of the surge in imports by 22 percent. Thus, the overall output of agriculture declines by 0.7 percent. The biggest reductions are in forestry (21 percent), vegetable oil (7 percent), and fishing (5 percent). However, there is an improvement in raw cot- ton production because of the increase in the output of cotton lint and yarn, as well as textiles (see elsewhere below). In nonagriculture, almost all sectors realize positive growth in exports. Overall, exports in nonagriculture increase by 13 percent. The increase in manufacturing exports is also substantial, especially in major export items such as cotton lint and yarn, textiles, other food, and other manufacturing. There is also a notable increase in the exports of services such as commerce, transport, and private services. For other food, the increase in imports displaces domestic demand by 6 per- cent. However, this is offset by the increase in exports; thus, the output of other food improves by 2.6 percent, while the output price improves by 2.9 percent. The impact on textiles may be analyzed in relation to the effects on the cotton lint and yarn sector and the raw cotton sector. The increase in textile imports displaces domestic demand by 0.9 percent. However, this is offset by the increase in textile exports; thus, textile output improves by 4.1 percent, while the output price improves by 1.6 percent. Because the cotton lint and yarn sector supplies materials to the textile sector, the improvement in the output of textiles because of the greater exports leads to an improvement in the domestic demand for cotton lint and yarn by 2.5 percent. The increase in both the exports and the domestic demand for cotton lint and yarn leads to 5.0 percent more output and a rise in the output price by 1.7 percent, which, in turn, leads to 4.3 percent more output of raw cotton. The negative change in the value added prices in agriculture leads to lower prices for factors that are used heavily in agriculture. The wages of farmers decrease by 0.6 percent; the returns to land fall by 1.9 percent; and the average returns to agricultural capital fall by 5 percent. The average output price in nonagriculture decreases by 1.1 percent, but the value added price improves by 1.3 percent. The increase arises from the notable improvement in the value added prices of leather (20 percent), cotton lint and yarn (14 percent), other food (10 percent), textiles (10 percent), and transport (3 percent) (see table 7.9). Thus, the prices of the factors used in nonagriculture improve. The wages of unskilled workers increase by 1.5 percent and of skilled labor by 2.1 percent, while the average returns to nonagricultural capital rise by Pakistan 239 1 percent. Furthermore, there is a significant decline in the consumer price index. Thus, net of the inflation effects, factor prices improve except for the average returns to the capital used in agriculture. The nominal incomes of farmers drop (table 7.8). This occurs largely because of the declines in the wages of farmers, in the returns to land, and in the average returns to capital in agriculture. Because of the higher wages of workers, the nom- inal incomes of nonfarmers improve except for the nominal incomes of urban nonpoor households. The incomes of the urban nonpoor decline because of the income tax replacement imposed on this group. However, the decline in con- sumer prices is significant for all groups. This offsets the decline in nominal incomes except in the case of the urban nonpoor. In sum, all households, except urban nonpoor households, realize an increase in real incomes, which leads to a significant decline in poverty. The urban poor experience the largest increase in incomes and the largest drop in the depth of poverty. Income inequality is reduced. S2B: Unilateral agricultural liberalization by Pakistan Our fourth experiment sets only agricultural price distortions in Pakistan at zero, while retaining all nonagricultural trade taxes and assuming no changes in the global model.11 Overall, the poverty effects are significantly lower in this experiment compared with the S2A scenario. Furthermore, there are differences in the effects across households. Urban households experience a decline in poverty, and, although overall poverty in rural areas declines, large- and medium-farm households face increasing poverty. The results at the macro, sectoral, factor, and commodity price levels explain these poverty effects. At the sectoral level, the import prices of agriculture drop by 14 percent (table 7.7). The largest declines are in sugar (36 percent); livestock, cattle, and dairy (34 percent); milled wheat (18 percent); and vegetable oil (18 percent).12 There is also a reduction in domestic prices, but this is significantly smaller than the drop in import prices. Thus, the imports of agricultural goods surge by 30 percent. This agricultural liberalization results in a real exchange rate depreciation. Because tariffs and subsidies in nonagriculture are retained, the average import prices in nonagriculture increase by only 2.6 percent, while the domestic prices in nonagriculture increase by 1.11 percent. Thus, the imports of nonagricultural products decline by 0.5 percent. On the other hand, the exports of nonagricul- tural products improve by 3.1 percent. At the sectoral level, the increase is caused by the strong export effect on leather, wood products, cotton lint and yarn, and commerce. Because world prices are fixed and the domestic and output prices of nonagriculture increase, the increase in nonagricultural exports is caused by the 240 Agricultural Price Distortions, Inequality, and Poverty depreciation in the exchange rate. The increase in nonagricultural exports, together with the marginal increase in the domestic demand for nonagricultural products, leads to an improvement in nonagricultural output by 0.4 percent. The prices of the factors used in agriculture decline. The wages of farmers decrease by 2.2 percent; the returns to land, by 2.5 percent; and the average returns to capital, by 3.8 percent. However, the prices of factors used heavily in nonagri- culture improve. We observe a similar pattern in factor prices after we net out the marginal decline of 0.27 percent in the consumer price index. The nominal incomes of farmers decline, while the nominal incomes of non- farmers improve. The marginal decline in the consumer price index does not offset the decrease in the nominal incomes of farmers, especially large- and medium-farm households. Thus, the real income of these farm households is lower. However, non- farmers enjoy higher real incomes, except for the urban nonpoor, among whom real incomes fall slightly as a result of the tax burden they bear. Nonetheless, the addi- tional tax burden is not sufficiently large to push the urban nonpoor below the poverty line (unlike the case of scenario S2A); so, poverty declines in urban areas. Although the overall poverty in rural areas declines, large- and medium-farm households face increasing poverty because of declining real incomes. S3A: Full trade liberalization by Pakistan and the rest of the world Our fifth experiment combines the trade liberalization in the rest of the world with that in Pakistan in all sectors. The effects of the unilateral trade liberalization in Pakistan are larger than the effects of the rest-of-the-world trade liberalization. The combined impact on both exports and imports is strongly positive. There is also a large decline in the consumer price index. Factor prices in agriculture decline, but they improve in nonagriculture. However, net of the inflation effects, the only decline in factor returns occurs in the average returns to the capital used in agriculture. The nominal incomes of farmers decline, while the nominal incomes among nonfarmers improve. The large reduction in the consumer price index contributes to an increase in the real incomes of all households except urban nonpoor households. This scenario generates the largest reduction in poverty. Another point worth highlighting is that, while the incidence of poverty among the urban nonpoor increases, the increase is much smaller in this experiment than in scenario S2A. S3B: Agricultural liberalization by Pakistan and the rest of the world Our sixth experiment combines the agricultural liberalization of the rest of the world with that in Pakistan. It turns out that the effects of the reform in Pakistan dominate those of the agricultural liberalization in the rest of the world. There is also an upward response in imports and exports, but in agriculture only. The Pakistan 241 surge in agricultural imports displaces local production. This results in lower prices for the factors used in agriculture. Factor prices in nonagriculture increase because the sector remains protected. Therefore, farmers have lower incomes, while nonfarmers benefit. Sensitivity analysis: Indirect versus direct tax replacement The results discussed above are derived using a replacement income tax. Because the SAM we have used to calibrate the model has an income tax on urban non- poor households only (table 7.4), the direct tax replacement puts all the burden of financing the trade reform on this group. As an alternative, we consider, in this subsection, the assessment of indirect taxes to offset the losses in government tar- iff revenue that are generated by reform. We focus on the poverty effects of these two tax replacement schemes in scenarios S3A (full trade liberalization of all goods in the rest of the world and in Pakistan) and S3B (agricultural liberalization in the rest of the world and in Pakistan). The effects on real income across households are presented in table 7.10. In scenario S3A, in which all sectors are liberalized, changing the tax replacement from a direct tax to an indirect tax completely alters the results. Under the direct tax replacement, all households enjoy higher real incomes except urban nonpoor households. This tax replacement scheme therefore redistributes income from the urban nonpoor to the rest of the household groups. These latter household groups benefit from the reduction in consumer prices and from the redistribution of income from the urban nonpoor households. However, if an indirect tax replacement is used, consumer prices increase because of the taxes, and the bur- den is shared among all household groups depending upon household consump- tion structures. There is a reduction in household incomes among most of the groups except the three relatively wealthy groups (the large-farm households in other Pakistan, the rural nonpoor, and the urban nonpoor). Under this tax replacement scheme, there is a significant increase in domestic prices because of the higher indirect taxes. If the trade liberalization is focused on agriculture only under S3B, the income results are not sensitive to the tax replacement scheme used. This is because the net government budget implications of the elimination of the distortions are not as large in agriculture as in nonagriculture. Thus, the impact on domestic prices through higher indirect taxes is not as significant in the case of agricultural liber- alization as in the case of the trade liberalization of all goods. In both tax replace- ment schemes, farmers, particularly large- and medium-farm households, are negatively affected, while nonfarmers are favorably affected. However, in the direct tax replacement scheme, the urban nonpoor are still negatively affected, while they are favorably affected in the indirect tax replacement scheme. 242 Agricultural Price Distortions, Inequality, and Poverty Table 7.10. Sensitivity Analysis of Household Welfare Effects, Two Tax Replacement Schemes, Pakistan S3A S3B 2001­02 population Direct Indirect Direct Indirect Households distribution, % tax tax tax tax Large-farm households ­Sindh 0.1 1.96 0.49 1.51 1.65 ­Punjab 0.2 2.47 0.08 1.29 1.42 ­Other Pakistan 0.0 2.90 0.18 1.12 1.26 Medium-farm households ­Sindh 1.0 1.94 0.75 1.06 1.20 ­Punjab 1.9 2.52 0.31 0.77 0.92 ­Other Pakistan 0.8 2.40 0.40 1.30 1.45 Small-farm households ­Sindh 2.8 2.28 0.70 0.25 0.40 ­Punjab 12.9 2.44 0.55 0.10 0.25 ­Other Pakistan 5.7 2.52 0.56 0.18 0.03 Small-farm renters, landless ­Sindh 0.2 2.25 0.92 0.03 0.18 ­Punjab 1.5 2.64 0.53 0.03 0.13 ­Other Pakistan 0.6 2.59 0.51 0.04 0.20 Rural agricultural workers, landless ­Sindh 4.1 3.33 0.48 0.98 0.80 ­Punjab 2.0 3.40 0.36 1.00 0.82 ­Other Pakistan 0.9 2.87 0.47 0.82 0.66 Rural nonfarm households ­nonpoor 17.6 4.25 0.55 0.47 0.29 ­poor 18.1 3.11 0.41 0.80 0.63 Urban households ­nonpoor 22.5 3.69 0.63 0.12 0.10 ­poor 7.2 4.28 0.14 0.93 0.71 Source: National CGE model simulations by the authors. Table 7.11 presents the poverty results based on this sensitivity analysis. Trade liberalization in all goods globally under the indirect tax replacement scheme is poverty increasing in scenario S3A. This is because of the declining real incomes among most groups. This effect arises largely because of the higher consumer prices resulting from the indirect tax replacement scheme. The higher consumer prices wipe out the gains accruing because of the higher border export prices, the lower border import prices, and the lower tariffs. Agricultural liberalization alone entails a smaller fiscal burden. Therefore, both the direct income tax replacement scheme and the indirect tax replacement scheme Pakistan 243 Table 7.11. Sensitivity Analysis of Poverty Effects, Two Tax Replacement Schemes, Pakistan S3A S3B 2001­02 Direct Indirect Direct Indirect Poverty index index tax tax tax tax All Pakistan ­P0 31.2 6.4 0.5 1.8 1.2 ­P1 6.5 11.5 1.5 2.5 1.8 ­P2 2.0 13.8 1.8 3.0 2.2 All urban ­P0 19.9 0.4 0.0 2.7 2.3 ­P1 3.9 14.8 0.6 3.7 2.9 ­P2 1.2 18.3 0.7 4.3 3.3 All rural ­P0 38.2 8.6 0.6 1.5 0.9 ­P1 8.0 10.5 1.7 2.1 1.5 ­P2 2.5 12.5 2.1 2.6 1.9 Large- and medium-farm households ­P0 22.8 7.9 2.2 3.4 3.4 ­P1 4.1 10.0 2.4 5.3 6.0 ­P2 1.1 10.6 2.8 5.8 6.6 Small-farm households and agricultural workers ­P0 37.4 8.7 1.5 0.9 0.3 ­P1 7.5 10.3 2.2 1.3 0.6 ­P2 2.2 12.4 2.6 1.6 0.9 Rural nonfarm households ­P0 39.9 8.5 0.0 2.1 1.5 ­P1 8.7 10.6 1.5 2.8 2.2 ­P2 2.8 12.6 1.8 3.4 2.7 Source: National CGE model simulations by the authors. Note: PO = poverty headcount. P1 = poverty gap. P2 = poverty severity. generate favorable effects on poverty. In the case of the indirect income tax replace- ment scheme, although it increases consumer prices, it does not wipe out the gains achieved through the higher border export prices, the lower border import prices, and the lower trade taxes on agricultural commodities. Because of the negative effects of the agricultural liberalization on domestic agriculture in Pakistan, farm- ers will be hurt, especially large- and medium-farm households. However, this is a small segment in the total population and exhibits the lowest incidence of poverty (23 percent in 2001­02 compared with 37 percent among small-farm households and agricultural workers and 40 percent among rural nonfarmers). 244 Agricultural Price Distortions, Inequality, and Poverty Summary and Policy Implications In this chapter, we have linked the results of two economic models--the Linkage model of the World Bank and the Pakistan CGE model that we have developed-- to analyze and compare the poverty effects of trade liberalization abroad with the corresponding effects of unilateral reform by Pakistan. We have conducted six policy experiments: two rest-of-the-world trade liberalization experiments (full liberalization that covers all goods sectors and agriculture only), two unilateral trade liberalization cases (all goods and agriculture only), and two combined sce- narios. The results are evaluated under a direct tax replacement scheme among household incomes; the direct tax is paid only by the urban nonpoor. We also examine an alternative tax replacement scheme, an indirect tax on commodities. A number of policy insights may be drawn from our simulation results. The impacts on the Pakistan economy and on the extent of poverty in the country from own-country liberalization are significantly larger than the impacts from rest-of-the-world trade liberalization. The effects of agricultural liberalization (both in the rest-of-the-world market and in Pakistan) are considerably smaller than the effects of the liberalization of all goods trade. This is because of the smaller share of agricultural trade in overall exports and imports in Pakistan, in which trade is dominated by nonagricultural products. The income from trade taxes is a major source of revenue for the government. The trade tax revenue is considerably lower from agricultural commodities than from nonagricultural products. Thus, the elimination of the trade taxes on all tradable commodities creates a large dent in government income and in the fiscal balance. It therefore entails a significant government demand for tax revenue from other sources. The poverty and income effects of full trade liberalization greatly depend upon how the tax replacement is implemented. If an additional tax is imposed on household incomes to generate funds to finance the reduction in trade taxes in all sectors, there is a notable decline of consumer prices and a large income redistribution from urban nonpoor households to the rest of the house- hold groups. There is therefore a considerable decline in the incidence of poverty, in the depth of poverty, and in income inequality. This is because the burden of the additional tax falls entirely on the urban nonpoor, while the rest of the groups benefit from higher real factor prices and larger reductions in consumer prices. However, if the tax replacement is imposed as additional indirect taxes on com- modities, consumer prices increase and eliminate the benefits generated from the reduction in trade distortions. In this case, poverty rises. The trade tax revenue is considerably lower from agricultural commodities than from nonagricultural products. If trade liberalization is focused on agricul- tural commodities only, the fiscal refinancing requirement is substantially less urgent. The poverty reduction effects, although smaller, are robust to the change Pakistan 245 in tax policy, that is, poverty is reduced under both tax replacement schemes if only agricultural markets are liberalized. All these results are derived using a static model. The dynamic impact of trade reform on capital accumulation from the changes in prices has not been accounted for. For example, if the rates of return to capital are high in sectors in which the poor are heavily engaged, this will attract investment, thereby increas- ing capital accumulation in these sectors and output from these sectors. This would have favorable implications for poverty. (It is also possible that the results would be reversed and would therefore generate negative effects on the poor.) Furthermore, the dynamic effects would have an impact on technological progress, the movement of the labor of farm households into nonfarm employ- ment, factor and total productivity, and the flow of foreign direct investment. These are all empirical issues that are relevant topics for further research. Notes 1. Estimates of agricultural assistance in Pakistan, based on Dorosh and Salam (2009), are incorpo- rated in the World Bank's Distortions to Agricultural Incentives Project Database (Anderson and Valenzuela 2008). The estimates cover five decades. The representative values for CGE modeling as of 2004 that are used here are available in Valenzuela and Anderson (2008). 2. The specification of the Pakistan CGE model is based on EXTER (see Decaluwé, Dumont, and Robichaud 2000). 3. Cororaton and Orden (2008) include a dynamic analysis in which sectoral capital adjusts over time. 4. The direct tax replacement on household income is specified as: dyh yh(1 dtxrh[1 ndtxrh]), where dyh is disposable income; yh is income before income taxes; dtxrh is the income tax rate at the base; and ndtxrh is the income tax replacement. The indirect tax replacement on commodities is specified as: pd pl(1 itxr)(1 nitx), where pd is the domestic price; pl is the local price before indirect taxes; itxr is the indirect tax rate at the base; and nitx is the indirect tax replacement. 5. Sectoral informal capital is also in the SAM. The returns to informal capital may be considered primarily as payments to labor outside the formal labor market. However, instead of modeling infor- mal capital separately, we have aggregated it with formal capital. There is no significant underestima- tion of household incomes because informal capital is still being paid based on the returns to capital. However, this aggregation makes the labor share in agriculture appear relatively low. 6. We have set the sectoral values of the parameter eta in the export demand function equal to the Armington elasticities in the global model. The sectoral values of the parameter sig_e in the export supply function and the sectoral values of the parameter sig_m in the import demand function are half the values of eta. 7. There is no real exchange rate variable in the model. The real exchange rate is defined as the world price, multiplied by the nominal exchange rate, divided by the local price, where the world price is the trade-weighted world import and export prices and the local price is the sectoral output- weighted local prices. 8. Detailed sectoral results are shown only for scenario S2A (see table 7.9). Detailed comparable sectoral results for the other scenarios are available from the authors on request. 246 Agricultural Price Distortions, Inequality, and Poverty 9. In the model, the overall government revenue from tariffs is PRs 154 billion and from export taxes, PRs 15 billion. Total government revenue is PRs 446 billion. The total income of the urban nonpoor is PRs 1.73 trillion. 10. In our model, Pakistan is facing a downward-sloping world demand curve. Because the perfect substitution assumption is imposed between exports and domestic sales in Pakistan, the export supply curve is horizontal. The decrease in output prices raises export supply, which shifts the horizontal export supply curve downward. 11. The total tariff revenue from agricultural imports is PRs 14.2 billion, while farm export tax revenue is PRs 4.3 billion in the baseline. 12. Detailed sectoral results generated under this scenario are available from the authors upon request. References Ahmed, V., A. Abbas, and S. Ahmed. 2009. "Taxation Reforms: A CGE-Microsimulation Analysis for Pakistan." Unpublished working paper, Modeling and Policy Impact Analysis Program, Poverty and Economic Policy Research Network, Université Laval, Quebec. Anderson, K., and E. Valenzuela. 2008. "Estimates of Global Distortions to Agricultural Incentives, 1955­2007." Data spreadsheet, October, World Bank, Washington, DC. http://go.worldbank .org/YAO39F35E0. Cororaton, B. C., and D. Orden. 2008. Pakistan's Cotton and Textile Economy: Intersectoral Linkages and Effects on Rural and Urban Poverty. IFPRI Research Report 158. Washington, DC: International Food Policy Research Institute. Decaluwé, B., J.-C. Dumont, and V. Robichaud. 2000. Basic CGE Models.Vol. II of MIMAP Training Session on CGE Modeling. Micro Impacts of Macroeconomic and Adjustment Policies Program, Interna- tional Development Research Centre, Canadian Centre for International Studies and Cooperation, and Université Laval, Quebec. http://www.pep-net.org/fileadmin/medias/pdf/volume2.pdf. Dorosh, P. 2005. "Wheat Markets and Pricing in Pakistan: Political Economy and Policy Options." Wheat Policy Note, South Asia Rural Development Unit, World Bank, Washington, DC. Dorosh, P., M. K. Niazi, and H. Nazli. 2004. "A Social Accounting Matrix for Pakistan, 2001­02: Methodology and Results." PIDE Working Paper 2006­9, Pakistan Institute of Development Economics, Islamabad. http://econpapers.repec.org/paper/pidwpaper/2006_3a9.htm. Dorosh, P., and A. Salam. 2009. "Pakistan." In Distortions to Agricultural Incentives in Asia, ed. K. Anderson and W. Martin, 379­407. Washington, DC: World Bank. Federal Bureau of Statistics. 2003. "Household Integrated Economic Survey: Round 4, 2001­02." Federal Bureau of Statistics, Islamabad. http://www.statpak.gov.pk/depts/fbs/statistics/hies0102/ hies0102.html. Foster, J., J. Greer, and E. Thorbecke. 1984. "A Class of Decomposable Poverty Measures." Econometrica 52 (3): 761­66. Horridge, M., and F. Zhai. 2006. "Shocking a Single-Country CGE Model with Export Prices and Quan- tities from a Global Model." In Poverty and the WTO: Impacts of the Doha Development Agenda, ed. T. W. Hertel and L. A. Winters, 94­104. London: Palgrave Macmillan; Washington, DC: World Bank. Ministry of Finance. 2003. Economic Survey of Pakistan 2002­03. Islamabad: Economic Advisers Wing, Finance Division, Government of Pakistan. http://www.accountancy.com.pk/docs/Economic _Survey_2002-03.pdf. Valenzuela, E., and K. Anderson. 2008. "Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980­84." Research Memorandum 13 (December), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925. van der Mensbrugghe, D. 2005. "Linkage Technical Reference Document: Version 6.0." December, World Bank, Washington, DC. http://go.worldbank.org/7NP2KK1OH0. World Bank. 2007. Pakistan, Promoting Rural Growth and Poverty Reduction. Report 39303-PK. Washington, DC: Sustainable Development Unit, South Asia Region, World Bank. 8 The Philippines Caesar B. Cororaton, Erwin Corong, and John Cockburn* The agricultural sector in the Philippines employed 36 percent of the labor force and accounted for roughly 14 percent of the nation's GDP in 2004, or 26 percent of GDP if the agriculture-based food processing sector is included. From the 1950s to the 1970s, government policies were biased against agricul- ture. Until the 1980s, these policies included the import-substitution policy, which created a bias in favor of manufacturing and penalized the returns to agri- cultural investments and exports. Export taxes and exchange rate overvaluation greatly reduced earnings from agriculture, and government intervention through the creation of government corporations siphoned off much of the gains from trade (Intal and Power 1990, David 2003). Then, the trade reform program in the 1980s led to a shift from taxing to protecting agriculture relative to nonagricul- tural sectors, and these policies became more pronounced when the country became a member of the World Trade Organization in 1995. As a result, the cur- rent system of protection favors agriculture through both applied tariff rates and nominal rates of assistance (NRAs) in agriculture that are substantially higher than the NRAs in manufacturing (Aldaba 2005; David, Intal, and Balisacan 2009). However, two decades of protection have failed to induce competitiveness and productivity growth in agriculture. This chapter analyzes the poverty and inequality implications of the removal of agricultural and nonagricultural price distortions in the domestic markets of the *The authors are grateful to Kym Anderson, Thomas Hertel, Will Martin, Ernesto Valenzuela, Dominique van der Mensbrugghe, and workshop participants for helpful comments. 247 248 Agricultural Price Distortions, Inequality, and Poverty Philippines and in markets abroad. As exogenous shocks, our analysis relies on the results of the rest-of-the-world trade liberalization simulations of the Linkage model of the World Bank, here also called the global model (see the appendix by van der Mensbrugghe, Valenzuela, and Anderson; see also van der Mensbrugghe 2005). We use these exogenous shocks, along with national liberalization shocks, to derive effects based on the computable general equilibrium (CGE) model for the Philippines by Cororaton and Corong (2009). The global model incorporates new estimates of agricultural protection and assistance for various developing countries, including the Philippines, and simulates scenarios involving full world trade liberalization in all sectors and in agriculture alone.1 The global simulations generate changes in export and import prices at the border in the Philippines, as well as changes in world export demand for the products of the Philippines.2 We apply these results, together with the new estimates of protection and assistance for the Philippines by David, Intal, and Balisacan (2009), as shocks to the CGE model for the Philippines to analyze the distributional, welfare, and poverty impacts of various trade liberalization scenarios for the country. We conduct our simulation analysis in stages to assess the differing impacts that international market liberalization and domestic market liberalization may entail. In the first stage, we extract changes in the border export and import prices and changes in the world export demand for the products of the Philippines from the global model, and we apply these in the model for the Philippines without altering the existing trade protection system in the country. In the second stage, we simulate unilateral trade liberalization in the Philippines without incorporat- ing any changes from the global model. Finally, we combine the rest-of-the-world liberalization shocks and the unilateral liberalization shocks to assess the total effects. We conduct six policy experiments using separate scenarios for trade liberal- ization in all sectors and trade liberalization in the agricultural sector alone. We define the agricultural sector to include primary agriculture and lightly processed food.3 In each scenario, we generate results at the macrolevel and sectoral level, as well as vectors of changes in household incomes, consumer prices, and sectoral employment shares. We then use the vectors of changes as inputs in a microsimu- lation procedure to calculate the impact on poverty and inequality. This calcula- tion draws on data from a national household survey conducted in 2000. The chapter is organized as follows. The next section sheds light on the degree of price distortions and trade protection and on poverty trends in the Philippines. The subsequent section presents the structure of the CGE model for the Philippines, which is based on the national social accounting matrix as of 2000. We then discuss the policy experiments and the results. The final section provides a summary of our findings and examines some policy implications. The Philippines 249 Trade and Assistance Policies and Poverty Trends In 1949, the Philippines embarked on a development strategy geared toward industrial import substitution with a lesser emphasis on the agricultural and export sectors. The strategy provided protection to domestic producers of final goods through high tariff rates on nonessential consumer goods and low tariff rates on essential producer inputs. However, the policy was not effective: the growth of manufacturing value added and industrial employment increased min- imally. In 1970, the government shifted toward export promotion by offering tax exemptions and fiscal incentives to capital-intensive firms located in export pro- cessing zones. This strategy also achieved little: the highly skewed structure of the tariff protection in favor of import-substituting manufactured goods remained. Moreover, the imposition of export taxes, an overvalued exchange rate, and the presence of government corporations that not only regulated domestic prices, but also siphoned off much of the gains from domestic and international trade created a strong bias against agriculture and exports. The restrictive trade policies adopted between the 1950s and the late 1970s pre- vented efficient resource allocation and the smooth functioning of markets and penalized the domestic economy in three respects. First, import controls resulted in an overvalued exchange rate that favored import-substituting firms. Second, continued protection increased domestic output prices, which impeded forward links. Third, tariff escalations and import controls weakened backward links because the tariffs on capital and intermediate goods were maintained at low lev- els relative to the tariffs on finished products (Austria and Medalla 1996). This also promoted rent seeking and distorted the economic incentives for investment in agriculture. The agricultural sector, which served as the country's backbone in providing the necessary foreign exchange needed by the import-dependent man- ufacturing sector, stagnated. Meanwhile, the industrial sector ventured into import-dependent assembly operations with minimal value added content and little or no forward and backward links. Realizing the pitfalls of the import-substitution policy and the unsuccessful export-promotion strategy, the government implemented a series of trade reform programs starting in 1981.4 The Philippines has also been participating more actively in the multilateral trading system since it joined the World Trade Organi- zation in 1995. For example, between 1995 and 1999, the country complied with all its multilateral commitments within the prescribed time frames (WTO 1999). These commitments included tariff bindings at a maximum of 10 percentage points over the 1995 applied rate on roughly 65 percent of all tariff lines, tariff bindings on selected information technology products, and bindings on market access in selected service sectors (Austria 2001). Nonetheless, there continues to be a substantial tariff binding overhang, especially in agriculture.5 250 Agricultural Price Distortions, Inequality, and Poverty By around 2000, the country had slowed the pace of trade reform (WTO 2005). Although the average applied most favored nation tariff rate declined from 9.7 to 5.8 percent between 1999 and 2003, it then rose to 7.4 percent in 2004. This rever- sal of the tariff adjustment process was brought about by presidential discretion and was aimed at helping problematic domestic industries and responding to lobbying from domestic interest groups. Estimates of the NRA to agriculture David, Intal, and Balisacan (2009) have estimated the NRA to key industries in the agricultural sector. The NRA is the percentage difference between the domestic and the border price; it thus measures the direct effect of policy-induced distortions on producer incentives. The NRA for coconuts (copra or dried coconuts) is negative throughout the years shown in table 8.1, largely because of export taxes, a coconut levy, and a copra export ban. The currency devaluation in the 1970s and the world commod- ity boom in the middle of that decade did not translate into higher profits for coconut farmers; rather, it translated into higher revenue for the government and lower raw material costs for the coconut oil milling industry. Although these poli- cies began to be eliminated in 1986, coconut farmers remain penalized owing to the continued existence of a government corporation that controls 70­80 percent of coconut oil milling, thereby retaining a monopsonistic command over the domestic price of copra. The NRA for corn has always been positive and exhibits an increasing trend. There is not much political pressure on corn compared with rice because corn is a subsistence crop among upland farmers in the southern part of the country. Nonetheless, it is also a major animal feed ingredient. Among agricultural crops, sugar has had the highest NRA since the 1960s. In the 1960s and early 1970s, a major part of domestic production was exported to the high-priced U.S. market through a preferential access quota allocated to Philippine producers and sugar processors (known as the Laurel-Langley Agree- ment). This agreement ended in 1974, resulting in a dramatic drop in sugar exports from the Philippines to the United States.6 The NRA for chicken has been consistently high and well above the NRA for pork. However, the government imposed the same level of in-quota and out-of- quota tariffs for both commodities after the ratification of the World Trade Organization agreement in 1995. Cattle were not included on the sensitive products list for the Philippines at the World Trade Organization. The NRA for cattle has therefore been low relative to the NRAs for chicken and pork. In the early 1990s, the government attempted to Table 8.1. The NRAs to Major Agricultural Commodities, the Philippines, 1960­2004 (percent) Commodity 1960­64 1965­69 1970­74 1975­79 1980­84 1985­89 1990­94 1995­99 2000­04 Rice 6 1 10 18 16 14 21 53 51 Corn 19 38 14 24 20 60 63 79 55 Sugar 18 121 12 2 60 13 49 97 79 Domestic 4 78 39 29 14 112 45 99 75 Export 28 154 16 17 89 161 77 90 130 Coconut Copra 12 20 25 17 27 21 15 8 14 Coconut oil 3 18 21 8 17 4 7 1 6 Beef 60 16 47 18 2 8 26 15 17 Pork 30 14 3 6 36 51 25 21 8 Chicken -- 67 29 28 38 43 57 42 52 Other 10 10 32 32 16 17 10 5 5 Source: David, Intal, and Balisacan (2009). Note: -- no data are available. 251 252 Agricultural Price Distortions, Inequality, and Poverty promote the fattening of cattle and allowed duty-free imports of young cattle, while imposing more restrictive nontariff barriers on beef. Nonetheless, cattle fattening activities did not prosper because the tariffs on beef imports fell. Before the mid-1980s, the NRAs for agricultural inputs such as fertilizer, agri- cultural chemicals, and farm machinery were generally higher than the NRAs for agricultural crops, averaging well above 20 percent (David, Intal, and Balisacan 2009). This was largely caused by the government's industrial promotion policies, which raised the domestic prices of the manufactured inputs to agriculture. How- ever, after this period and during the trade liberalization process, there were sub- stantial reductions in agricultural input protection, down to around 10 percent in the late 1990s and to a uniform 3 percent in 2000­04. Poverty trends In rural and urban areas, over 60 percent of the expenditure of poor households goes for food, of which almost half is cereals, primarily rice and corn (table 8.2). Rural dwellers spend proportionately more than their urban counterparts on food. Food consumption among nonpoor households is relatively somewhat less (39 percent), and urban nonpoor households spend the least on food and cereals in relative terms (8 percent). Figure 8.1 presents the evolution of the poverty headcount index between 1985 and 2000. The national headcount index decreased from almost 50 to roughly 34 percent over the period. However, the decline was concentrated mainly in urban areas, especially in the National Capital Region, where poverty was already low. In contrast, the rural headcount index fell only modestly, from 56 to 49 per- cent, compared with the fall in urban areas from 44 to 23 percent. Table 8.2. Poverty Incidence and Food Expenditure Shares, the Philippines, 1997 and 2000 (percent) Rural Urban Indicator 1997 2000 1997 2000 Poverty incidence, % of population 50.7 48.8 21.6 18.6 Poor Nonpoor Poor Nonpoor Indicator 1997 2000 1997 2000 1997 2000 1997 2000 Expenditure shares, % of total All food 63.6 63.6 47.6 47.6 61.4 60.8 38.8 38.7 Cereals (mostly rice) 29.5 28.8 15.4 14.6 24.5 23.0 8.6 8.2 Source: NSO (1997, 2000). The Philippines 253 Figure 8.1. Trends in Poverty Indexes, the Philippines, 1985­2000 60 50 40 percent 30 20 10 0 1985 1988 1991 1994 1997 2000 year all national capital region rural urban Source: NSO (1985­). The CGE Model The national CGE model we use in our study is calibrated to the social accounting matrix of the Philippines for 2000.7 There are 41 production sectors and four fac- tors: two labor types (skilled workers with at least a college diploma and unskilled workers), plus capital and land. Institutions include the government, firms, households, and the rest of the world. Household categories are defined according to income deciles. Output (X) is a composite of value added (VA) and intermedi- ate inputs. Output is sold either to the domestic market (D), or to the export mar- ket (E), or both. The model assumes perfect substitutability between E and D. We assume a finite elasticity of export demand. There are two sources of domestic market supply: domestic output and imports (M). There is substitution between D and M depending on the changes in the relative prices of D and M and on a constant elasticity of substitution function. Sectoral output is a Leontief function of intermediate inputs and value added. Value added in agriculture is a constant elasticity of substitution function of skilled labor, unskilled labor, capital, and land. Nonagricultural value added is also a constant elasticity of substitution function of the same factors, except land. Capital and land are each sector specific. Skilled and unskilled labor are mobile 254 Agricultural Price Distortions, Inequality, and Poverty across sectors, but limited within skill category, and land use is mobile within the agricultural sector. Households earn incomes from factors of production, transfers, foreign remit- tances, and dividends, while paying direct income taxes to the government. Household savings are a fixed proportion of disposable incomes, and household demand is represented by a linear expenditure system. Government revenue is the sum of the direct taxes on the incomes of house- holds and firms, indirect taxes on domestic and imported goods, and other receipts. The government spends on the consumption of goods and services, transfers, and other payments. In the present version of the model, we assume a fixed government balance. Since shifts in policy will result in changes in govern- ment income and expenditure, the government balance is held fixed through a tax replacement variable. For the present analysis, we use an indirect tax replacement on domestic sales, but we also compare the results with the effects of a direct tax replacement on household incomes. In either case, the tax replacement is endoge- nously determined so as to maintain a fixed government balance. Foreign savings are held fixed. The nominal exchange rate is the model's numéraire. We introduce a weighted price of investment and derive total invest- ment in real prices; total investment is held fixed by introducing an adjustment factor in the household savings function. Equilibrium is achieved in the model if the supply of and the demand for goods and services are equal and if investment is equal to savings. Table 8.3 presents the production structure in the social accounting matrix. Generally, the agricultural and service sectors show greater shares of value added (as a percent of output) relative to the industrial sector. In agriculture, coconut and forestry have the largest shares of value added, almost 90 percent, while petro- leum refining has the lowest share among the industrial sectors, at 14 percent. The capital-output ratio is generally lower in agriculture than in industry and the ser- vice sectors. The largest employer of labor is the service sector. More than 90 per- cent of the labor input in agricultural production is represented by unskilled labor. The share of skilled labor employed in the industrial sector is substantially greater than the corresponding share in the agricultural sector. The structure of indirect taxes reveals that tobacco and alcohol (23 percent), followed by petro- leum (18 percent), are exposed to the highest indirect tax (table 8.3, last column). Table 8.4 shows that almost 50 percent of total exports are electrical prod- ucts. Almost 90 percent of the production of electrical products is exported. A major part of this sector is the semiconductor industry. A sizable share of exports are represented by machinery and transport equipment. This industry has a high export-intensity ratio, 73 percent.8 It is followed in the size of the ratio by other manufacturing, coconut oil, leather, fertilizer, other chemicals, Table 8.3. Production Structure, the Philippines, 2000 (percent) Value % of % of % of added/ value % of % of Capital/ skilled unskilled Land/ Indirect Sector output, % added output employment labor, % labor labor output, % tax rate Agriculture Primary agriculture Palay (rice) 77.5 2.0 1.4 3.1 41 6.2 93.8 7.3 3.3 Corn 78.5 0.6 0.4 1.0 25 6.2 93.8 5.3 3.5 Coconuts 88.9 0.6 0.4 0.8 59 6.2 93.8 10.3 0.9 Fruits and vegetables 79.7 2.2 1.5 2.4 88 6.2 93.8 11.3 3.4 Sugar 69.7 0.3 0.2 0.3 83 6.2 93.8 11.2 1.8 Other crops 77.3 0.6 0.4 0.5 105 6.2 93.8 13.7 1.3 Hogs 63.7 1.4 1.1 1.6 84 9.5 90.5 6.8 2.2 Cattle 71.9 0.4 0.3 0.4 111 9.5 90.5 11.0 1.2 Chickens 60.7 1.3 1.1 1.4 92 9.5 90.5 8.7 2.4 Lightly processed food Meat processing 20.5 1.1 2.8 0.8 196 25.0 75.0 n.a. 1.6 Milk and dairy 31.1 0.3 0.5 0.2 210 25.0 75.0 n.a. 1.0 Coconut and edible oil 28.7 0.5 0.9 0.2 574 25.0 75.0 n.a. 0.9 Milled rice and corn 34.8 1.4 2.1 1.2 126 25.0 75.0 n.a. 2.0 Milled sugar 22.0 0.2 0.4 0.1 191 25.0 75.0 n.a. 1.4 (Table continues on the following pages.) 255 256 Table 8.3. Production Structure, the Philippines, 2000 (continued) (percent) Value % of % of % of added/ value % of % of Capital/ skilled unskilled Land/ Indirect Sector output, % added output employment labor, % labor labor output, % tax rate Nonagriculture Other primary products and mining Agricultural services 84.7 0.4 0.2 0.5 61 6.2 93.8 10.0 2.8 Forestry 89.4 0.2 0.1 0.1 217 16.9 83.1 33.0 3.9 Fishing 77.4 2.8 1.9 2.1 216 2.4 97.6 3.8 1.7 Mining 63.0 0.6 0.5 0.4 253 30.5 69.5 n.a. 2.2 Crude oil and natural gas 34.6 0.0 0.0 0.0 n.a. n.a. n.a. n.a. 0.0 Highly processed food and tobacco Fruit processing 36.5 0.4 0.5 0.3 166 25.0 75.0 n.a. 2.2 Fish processing 28.5 0.3 0.6 0.2 355 25.0 75.0 n.a. 1.3 Other processed food 30.9 1.3 2.3 1.2 162 25.0 75.0 n.a. 1.6 Tobacco and alcohol 40.4 1.0 1.4 1.0 156 57.7 42.3 n.a. 22.9 Manufacturing Textiles 37.3 1.0 1.4 1.0 130 6.4 93.6 n.a. 0.7 Garments and footwear 46.1 2.1 2.4 1.9 162 4.5 95.5 n.a. 0.5 Leather and rubber-wear 42.9 0.7 0.9 0.7 143 9.8 90.2 n.a. 0.4 Paper and wood products 39.3 1.7 2.3 1.5 163 23.5 76.5 n.a. 0.7 Fertilizer 39.7 0.1 0.2 0.1 140 37.8 62.2 n.a. 0.5 Other chemicals 41.1 1.9 2.4 1.5 201 37.8 62.2 n.a. 1.0 Petroleum 14.2 0.7 2.6 0.8 114 42.4 57.6 n.a. 17.7 Cement and related products 41.7 0.7 0.9 0.6 165 29.8 70.2 n.a. 1.9 Metal and related products 36.9 1.9 2.7 1.4 210 8.4 91.6 n.a. 1.1 Machinery and transport equipment 40.0 3.6 4.8 1.8 368 30.4 69.6 n.a. 1.7 Electrical and related products 45.5 8.5 9.9 7.3 171 39.5 60.5 n.a. 1.2 Other manufacturing 48.1 1.4 1.6 1.4 135 6.7 93.3 n.a. 1.8 Other industry Construction 53.0 3.9 3.9 5.5 67 14.9 85.1 n.a. 1.4 Utilities 68.3 3.4 2.6 1.9 324 43.7 56.3 n.a. 3.2 Services Transportation and communications 53.6 7.0 6.9 5.3 210 18.2 81.8 n.a. 1.2 Wholesale trade 66.1 13.2 10.6 10.7 192 25.6 74.4 n.a. 1.1 Other services 63.5 20.2 16.8 17.5 171 31.5 68.5 n.a. 2.9 Public services 72.2 8.2 6.0 19.3 41 60.7 39.3 n.a. 0.0 Source: Based on the national model in Cororaton and Corong (2009). Note: n.a. not applicable. 257 258 Agricultural Price Distortions, Inequality, and Poverty Table 8.4. Trade Structure and Elasticity Parameters, the Philippines, 2000 Elasticitiesa Exports, % Imports, % b Sector sig_va sig_m eta sig_e Share Intensity Share Intensityb Agriculture Primary agriculture Palay (rice) 0.8 2.2 4.5 2.2 0.0 0.0 0.0 0.0 Corn 0.8 2.5 4.9 2.5 0.0 0.1 0.1 8.4 Coconuts 0.8 2.4 4.8 2.4 0.0 0.2 0.0 0.0 Fruits and vegetables 0.8 2.0 3.9 2.0 1.2 15.1 0.3 6.2 Sugar 0.8 3.0 5.9 3.0 0.0 0.0 0.0 0.0 Other crops 0.8 2.0 3.9 2.0 0.1 2.8 1.2 44.2 Hogs 0.8 2.0 3.9 2.0 0.0 0.0 0.0 0.0 Cattle 0.8 2.0 3.9 2.0 0.0 0.2 0.1 9.2 Chickens 0.8 2.0 3.9 2.0 0.0 0.0 0.0 0.4 Lightly processed food Meat processing 1.5 2.0 3.9 2.0 0.0 0.0 0.4 3.4 Milk and dairy 1.5 2.0 3.9 2.0 0.0 1.7 1.0 33.6 Coconut and edible oil 1.5 2.0 3.9 2.0 1.5 32.9 0.6 19.0 Milled rice and corn 1.5 2.2 4.5 2.2 0.0 0.0 0.8 8.8 Milled sugar 1.5 3.0 5.9 3.0 0.2 8.3 0.1 8.2 Nonagriculture Other primary products and mining Agricultural services 0.8 2.2 4.3 2.2 0.0 0.0 0.0 0.1 Forestry 0.8 2.2 4.3 2.2 0.1 10.3 0.0 0.6 Fishing 0.8 2.2 4.3 2.2 0.8 7.9 0.0 0.3 Mining 0.8 2.2 4.3 2.2 0.4 15.8 1.4 45.8 Crude oil and natural gas 0.8 2.2 4.3 2.2 0.0 0.0 7.5 99.6 Highly processed food and tobacco Fruit processing 1.5 2.0 3.9 2.0 0.7 24.1 0.3 13.9 Fish processing 1.5 2.0 3.9 2.0 0.7 22.0 0.2 7.4 Other processed food 1.5 2.0 3.9 2.0 0.6 4.8 0.9 9.3 Tobacco and alcohol 1.5 2.0 3.9 2.0 0.1 1.4 0.3 5.7 Manufacturing Textiles 1.5 2.1 4.1 2.1 1.2 16.9 2.7 36.7 Garments and footwear 1.5 2.1 4.1 2.1 0.2 1.8 0.1 1.3 Leather and rubber-wear 1.5 2.0 4.1 2.0 1.3 26.6 2.3 45.6 Paper and wood products 1.5 2.0 4.1 2.0 2.3 19.7 1.8 19.3 Fertilizer 1.5 2.0 4.1 2.0 0.1 16.8 0.5 49.4 Other chemicals 1.5 2.0 4.1 2.0 0.9 7.4 5.0 35.4 Petroleum 1.5 2.0 4.1 2.0 1.6 11.8 1.8 16.6 The Philippines 259 Table 8.4. Trade Structure and Elasticity Parameters, the Philippines, 2000 (continued ) Elasticitiesa Exports, % Imports, % b Sector sig_va sig_m eta sig_e Share Intensity Share Intensityb Cement and related products 1.5 2.0 4.1 2.0 0.4 9.5 0.5 13.8 Metal and related products 1.5 2.0 4.1 2.0 2.5 17.4 4.2 31.7 Machinery and transport equipment 1.5 2.0 4.1 2.0 18.3 73.2 12.5 70.6 Electrical and related products 1.5 2.0 4.1 2.0 45.9 89.0 35.2 88.9 Other manufacturing 1.5 2.0 4.1 2.0 3.7 44.3 2.0 36.1 Other industry Construction 1.5 1.0 2.1 1.0 0.3 1.5 0.3 1.9 Utilities 1.5 1.0 2.1 1.0 0.0 0.0 0.0 0.0 Services Transportation and communications 1.5 1.0 2.1 1.0 3.7 10.2 8.1 24.2 Wholesale trade 1.5 1.0 2.1 1.0 2.9 5.2 0.6 1.5 Other services 1.5 1.0 2.1 1.0 8.4 9.5 6.9 10.0 Public services n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. Source: Based on the national model in Cororaton and Corong (2009). Note: n.a. not applicable. a. sig_va substitution parameter in the constant elasticity of substitution value added function. sig_m substitution parameter in the constant elasticity of substitution composite good function. eta export demand elasticity. sig_e substitution parameter in the constant elasticity of transformation. b. Exports, divided by output; imports divided by the composite good's consumption. garments, fruit processing, and fish processing. On the import side, electrical products, as well as machinery and transport equipment, have high import- intensity ratios, accounting for 35 and 12 percent of total domestic sales, respectively. Other sectors in which imports are a major source of domestic supply include other crops, cattle, mining and crude oil, milk and dairy, fruit processing, fish processing, coconut oil, sugar milling, other food, textiles, leather, paper, fertilizer, other chemicals, petroleum, cement, and transporta- tion and communications. Table 8.4 also shows the values of key elasticity parameters used in the model: the import substitution elasticity (sig_m) in the constant elasticity of substitution composite good function, the production substitution elasticity (sig_va) in the 260 Agricultural Price Distortions, Inequality, and Poverty constant elasticity of substitution value added production function, and the export demand elasticity (eta), which is obtained from the Armington parameter of the global model. The consumption structure of households is presented in table 8.5. Rice is a significant staple among Filipinos, especially among poorer households. It accounts for 14.3 of total expenditure among the first household decile, but the share decreases substantially as households become richer. Fish and meat, fruits and vegetables, and other food are the other significant items in household con- sumption. Generally, the lower-income groups show substantial expenditure on food and food-related products. For instance, food items account for 42.4 percent of total expenditure among the first decile compared with 13.4 percent among the tenth decile. Richer households spend more on services relative to poorer house- holds. Products of special interest are corn, sugar, chicken, meat processing, milk and dairy, processed fruit, processed fish, milled rice and corn, and milled sugar. The shares of expenditure on these special products decline as households become richer: they account for 25 percent of consumption among the first decile, but only 8.6 percent among the tenth decile. Simulations All policy experiments reported in our study rely on an indirect tax replacement scheme to maintain fixed government balance. We generate results at the macrolevel and the sectoral level, as well as vectors of changes in household incomes, consumer prices, and sectoral employment shares. We then use the vectors of changes as inputs in a microsimulation procedure to calculate the impacts on poverty and inequality based on a household survey in 2000. We also undertake a sensitivity analysis with an alternative direct tax replacement scheme. Definition of policy experiments Table 8.6 shows the sectoral correspondence between the CGE model for the Philippines and the global model, as well as information on the sectoral tariff rates and export subsidies based on new estimates of nominal rates of protection and assistance for the Philippines. The CGE model for the Philippines is initially solved using these new estimates of protection and assistance to create the base with which all subsequent simulations may be compared. In certain policy exper- iments, we use the global simulation results from the global model as policy shocks for the model for the Philippines, following the method proposed by Horridge and Zhai (2006). Table 8.5. Structure of Household Expenditure, by Decile, the Philippines, 2000 (percent) Decile Sector 1 2 3 4 5 6 7 8 9 10 Agriculture Primary agriculture Corn 0.5 0.4 0.4 0.3 0.3 0.2 0.2 0.2 0.1 0.1 Coconuts 0.3 0.3 0.3 0.3 0.2 0.2 0.2 0.2 0.2 0.1 Fruits and vegetables 4.1 3.8 3.6 3.4 3.1 2.8 2.5 2.2 1.9 1.3 Sugar 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other crops 0.2 0.2 0.2 0.2 0.2 0.1 0.1 0.1 0.1 0.0 Chickens 0.8 0.9 0.9 1.0 1.1 1.1 1.1 1.1 1.0 0.7 Lightly processed food Meat processing 4.2 4.6 4.9 5.6 6.2 6.8 7.1 6.8 6.3 4.2 Milk and dairy 1.1 1.2 1.3 1.3 1.4 1.3 1.3 1.2 1.1 0.8 Coconut and edible oil 0.7 0.6 0.6 0.6 0.5 0.5 0.4 0.4 0.3 0.2 Milled rice and corn 14.3 12.9 11.7 10.0 8.4 6.9 5.7 4.5 3.4 1.8 Milled sugar 1.2 1.1 1.0 1.0 0.9 0.8 0.7 0.6 0.5 0.3 Nonagriculture Other primary products and mining Forestry 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Fishing 6.8 6.4 6.1 5.5 4.9 4.2 3.6 3.1 2.5 1.5 Mining 0.1 0.1 0.1 0.0 0.0 0.0 0.1 0.1 0.1 0.1 Highly processed food, and tobacco Fruit processing 1.2 1.1 1.0 0.9 0.9 0.8 0.7 0.6 0.5 0.4 Fish processing 2.0 1.9 1.8 1.6 1.4 1.2 1.1 0.9 0.7 0.4 261 (Table continues on the following page.) 262 Table 8.5. Structure of Household Expenditure, by Decile, the Philippines, 2000 (continued ) (percent) Decile Sector 1 2 3 4 5 6 7 8 9 10 Other processed food 5.1 4.8 4.7 4.3 4.0 3.7 3.3 2.9 2.5 1.6 Tobacco and alcohol 4.5 4.8 4.9 4.8 4.5 4.2 3.6 3.1 2.6 1.6 Mining and manufacturing Textiles 0.8 0.9 1.0 1.0 1.0 1.0 0.9 0.9 0.9 0.8 Garments and footwear 1.7 1.9 2.1 2.2 2.2 2.1 2.1 2.0 2.0 1.7 Leather and rubber-wear 0.3 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.3 Paper and wood products 0.8 0.7 0.7 0.7 0.6 0.6 0.6 0.6 0.7 0.9 Fertilizer 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other chemicals 2.7 2.4 2.2 2.1 1.9 1.8 1.8 1.9 2.2 3.1 Petroleum 1.9 1.6 1.6 1.6 1.6 1.5 1.5 1.4 1.3 0.9 Cement and related products 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 Machinery and transport equipment 0.1 0.3 0.3 0.5 0.7 0.9 1.0 1.1 1.1 1.3 Electrical and related products 0.3 0.7 0.8 1.1 1.5 1.8 1.9 2.1 2.2 2.4 Other manufacturing 0.6 0.8 0.9 0.9 1.0 1.1 1.1 1.1 1.1 1.0 Other industry Utilities 3.4 3.0 2.9 2.9 2.9 2.8 2.8 2.6 2.3 1.7 Services Transportation and communications 6.0 7.0 7.3 8.2 9.4 10.1 11.5 12.9 14.7 17.4 Wholesale trade 17.8 17.5 17.1 16.7 16.3 15.9 15.7 15.5 15.3 14.6 Other services 16.5 17.5 18.8 20.8 22.2 24.8 26.9 29.3 32.0 38.7 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: Based on the national model in Cororaton and Corong (2009). Note: There is no household consumption of agricultural services, crude oil and natural gas mining, and several other categories. Table 8.6. Exogenous Demand and Price Shocks Caused by Rest-of-the-World Liberalization (percent) Full trade liberalization, Agricultural trade liberalization, Trade distortions excluding Philippinesa excluding Philippinesb Export Export Import Export Export Import subsidy, price, price, demand price, price, Export Tariff, % % % shifter,c % % demand Philippine CGE model Linkage model % ( 0 tax) change change % change change shifter,c % Agriculture Primary agriculture Palay (rice) Paddy rice 19.6 0.0 0.0 0.0 1.0 0.0 0.0 1.0 Corn Other grains 29.6 0.0 0.0 6.1 1.0 0.0 5.7 1.0 Coconuts Oilseeds 4.8 10.0 0.0 0.8 1.0 0.0 0.5 1.0 Fruits and vegetables Vegetables and fruits 8.7 0.0 5.7 2.4 1.1 3.8 1.7 1.2 Sugar Sugarcane and beets 0.0 0.0 0.0 0.0 1.0 0.0 0.0 1.0 Other crops Other crops 3.9 0.0 5.9 1.3 1.0 3.9 1.4 1.0 Hogs Other livestock 18.7 0.0 5.6 1.0 1.0 3.6 0.1 1.0 Cattle 18.7 0.0 5.6 1.0 1.0 3.6 0.1 1.0 Chickens Cattle, sheep, other 10.0 0.0 0.0 5.6 1.0 0.0 5.5 1.0 Lightly processed food Meat processing Beef and sheep meat 9.0 0.0 3.7 2.8 0.5 2.0 4.5 0.5 Milk and dairy Dairy products 4.1 0.0 4.9 7.0 1.1 4.2 7.4 1.1 Coconut and edible oil Vegetable oils and fats 4.4 0.0 2.6 1.1 1.0 0.9 1.7 1.0 Milled rice and corn Processed rice 29.0 0.0 5.3 4.3 0.8 3.3 1.6 0.8 Milled sugar Refined sugar 73.2 0.0 3.9 2.1 1.5 2.0 0.8 1.6 263 (Table continues on the following pages.) 264 Table 8.6. Exogenous Demand and Price Shocks Caused by Rest-of-the-World Liberalization (continued ) (percent) Full trade liberalization, Agricultural trade liberalization, Trade distortions excluding Philippinesa excluding Philippinesb Export Export Import Export Export Import subsidy, price, price, demand price, price, Export Tariff, % % % shifter,c % % demand Philippine CGE model Linkage model % ( 0 tax) change change % change change shifter,c % Nonagriculture Other primary products and mining Agricultural services 2.7 1.0 2.8 0.6 1.1 1.0 0.9 1.0 Forestry 2.7 1.0 2.8 0.6 1.1 1.0 0.9 1.0 Fishing Other primary products 2.7 1.0 2.8 0.6 1.1 1.0 0.9 1.0 Mining 2.7 1.0 2.8 0.6 1.1 1.0 0.9 1.0 Crude oil and natural gas 2.7 1.0 2.8 0.6 1.1 1.0 0.9 1.0 Highly processed food and tobacco Fruit processing 6.0 0.0 3.6 1.6 1.2 2.0 0.4 1.0 Fish processing Other food, beverages, 6.0 0.0 3.6 1.6 1.2 2.0 0.4 1.0 Other processed food and tobacco 6.0 0.0 3.6 1.6 1.2 2.0 0.4 1.0 Tobacco and alcohol 6.0 0.0 3.6 1.6 1.2 2.0 0.4 1.0 Manufacturing Textiles Textiles and 8.0 1.7 2.0 0.2 1.0 1.1 0.4 1.0 Garments and footwear wearing apparel 8.0 1.7 2.0 0.2 1.0 1.1 0.4 1.0 Leather and rubber-wear 8.0 1.7 2.0 0.2 1.0 1.1 0.4 1.0 Paper and wood products 3.5 0.0 2.1 1.5 1.0 0.7 0.3 1.0 Fertilizer 3.5 0.0 2.1 1.5 1.0 0.7 0.3 1.0 Other chemicals 3.5 0.0 2.1 1.5 1.0 0.7 0.3 1.0 Petroleum 3.5 0.0 2.1 1.5 1.0 0.7 0.3 1.0 Cement and related Other manufacturing products 3.5 0.0 2.1 1.5 1.0 0.7 0.3 1.0 Metal and related products 3.5 0.0 2.1 1.5 1.0 0.7 0.3 1.0 Machinery and transport equipment 3.5 0.0 2.1 1.5 1.0 0.7 0.3 1.0 Electrical and related products 3.5 0.0 2.1 1.5 1.0 0.7 0.3 1.0 Other manufacturing 3.5 0.0 2.1 1.5 1.0 0.7 0.3 1.0 Other industry Construction Services 0.0 0.0 2.9 0.1 1.0 1.1 0.2 1.0 Utilities 0.0 0.0 2.9 0.1 1.0 1.1 0.2 1.0 Services Transport and communications 0.0 0.0 2.9 0.1 1.0 1.1 0.2 1.0 Wholesale trade Services 0.0 0.0 2.9 0.1 1.0 1.1 0.2 1.0 Other services 0.0 0.0 2.9 0.1 1.0 1.1 0.2 1.0 Source: Linkage model simulations by van der Mensbrugghe, Valenzuela, and Anderson (see the appendix). a. Rest-of-the-world liberalization in all sectors, excluding the Philippines. b. Rest-of-the-world liberalization in agriculture only, excluding the Philippines. c. Derived using (1 0.01 * p) (1 0.01 * q)^(1/ESUBM), where p is the export price change; q is the export volume change; and ESUBM is the Armington import elasticity taken from the global (Linkage) model. 265 266 Agricultural Price Distortions, Inequality, and Poverty The six policy experiments are as follows: · ROW-ALL, rest-of-the-world trade liberalization in all sectors, excluding the Philippines: This experiment uses the results of the global model under full rest-of-the-world liberalization in table 8.6 and retains all existing trade distortions (tariff rates and export subsidies) in the Philippines. · ROW-AGR, rest-of-the-world trade liberalization in agriculture and lightly processed food only: As with ROW-ALL, this experiment uses the results of the global model, but only under rest-of-the-world liberalization in agriculture and lightly processed food, and retains all existing trade distortions in the Philippines. · PHIL-ALL, unilateral trade liberalization in all sectors: All trade distortions in the Philippines are eliminated. No changes in the sectoral border export prices and import prices or in export demand are included in this unilateral liberalization. · PHIL-AGR, unilateral agriculture trade liberalization: All trade distortions in primary agriculture and lightly processed foods in the Philippines are elimi- nated. Similar to PHIL-ALL, there are no changes in the sectoral border export prices and import prices or in export demand in this unilateral liberalization. · COMB-ALL, full trade liberalization in the rest of the world and in the Philippines: This scenario combines ROW-ALL and PHIL-ALL in a global liberalization. · COMB-AGR, liberalization in agriculture and lightly processed foods in the rest of the world and in the Philippines: This scenario combines ROW-AGR and PHIL-AGR. Results In this subsection, we present the modeling results of our six policy experiments and then report the impacts on household income and welfare and on poverty and inequality. The subsection concludes with additional results that show the sensitivity of the core results to changes in the treatment of tax adjustments in the model. Rest-of-the-world trade liberalization in all sectors (ROW-ALL) The results of the Linkage model in table 8.6 indicate that this first policy experi- ment leads to higher export prices and greater export demand for the products of the Philippines. Within agriculture, a significant shift in export demand is observed in milled sugar, as well as raw fruits and vegetables (at 1.5 and 1.1 per- cent, respectively). This is also true in fruit and fish processing (1.2 percent), The Philippines 267 whereas slightly more modest shifts in export demand are observed in other industrial and service sectors. Meanwhile, full rest-of-the-world liberalization leads to higher world import prices for most Philippine goods. Table 8.7 shows that export prices overall rise more in the agricultural sectors than in the nonagricultural sectors (by an average of 3.6 versus 2.4 percent in local currency). So, too, do export volumes: they expand 9.8 percent in agriculture compared with a modest 0.3 percent rise in nonagriculture. Local import prices also increase more in agriculture than in nonagriculture (3.0 versus 1.1 percent). Substitution toward imported goods is observed owing to a larger rise in the prices of domestically produced goods relative to the prices of the corresponding imported products. Because of this, agricultural and nonagricultural import volumes increase by 0.9 and 1.1 percent, respectively (table 8.7, second column). The entire agricultural sector benefits from the improved international market conditions. Agricultural output and value added prices increase by 3.5 and 3.9 percent, respectively. Thus, the returns to agriculture-specific factors, particularly land and agricultural capital (which increase by 5.1 and 4.7 percent, respectively), rise relative to wage rates (3 percent) and to the returns to nonagricultural capital (2.9 percent). Unskilled wages increase slightly more than skilled wages because unskilled workers are used more intensively in the expanding agricultural sector. In contrast, in the nonagricultural sectors, the fall in domestic sales offsets the export expansion. Thus, the volume of output contracts by 0.1 percent. Essen- tially, this is traceable to the import­domestic price substitution effects (see else- where above), the fall in the world import prices for essential consumer goods such as garments (table 8.6), and the appreciation in the real exchange rate. Despite falling output volumes, nonagricultural output prices still increase by 2.5 percent owing to the higher export prices. Hence, the returns to factors such as nonagricultural capital and skilled workers, which are used intensively in nonagriculture, increase as well. Rest-of-the-world trade liberalization in agriculture only (ROW-AGR) The results of the ROW-AGR scenario are similar, but smaller in magnitude rela- tive to the results of the ROW-ALL scenario. We only focus on the results that are different in this scenario because the mechanisms driving the model results are essentially the same in the ROW-ALL and ROW-AGR scenarios. Agricultural exports increase by 11 percent mainly because of a significant export demand shift in sugar, as well as raw fruits and vegetables (1.6 and 1.2 percent, respectively; see tables 8.6 and 8.7). A distinct feature of this scenario is that domestic agricultural prices increase relatively less than agricultural import prices (1.7 versus 2.7 per- cent). In the face of more expensive agricultural imports, domestic demand expands, while imports fall (0.1 and 1.2 percent, respectively). Agricultural output 268 Table 8.7. Aggregate Simulation Results of Prospective Liberalizations in Agriculture and Nonagriculture, the Philippines (percent change from the baseline) ROW-ALL ROW-AGR PHIL-ALL PHIL-AGR COMB-ALL COMB-AGR Indicator Ag Nonag Ag Nonag Ag Nonag Ag Nonag Ag Nonag Ag Nonag Real GDP 0.3 0.1 0.8 0.2 1.1 0.3 Prices Real exchange rate 1.0 0.4 1.6 0.8 0.6 0.4 Export price in local currency 3.6 2.4 2.0 0.8 1.8 1.0 1.3 0.3 1.8 1.4 0.7 0.4 Import price in local currency 3.0 1.1 2.7 0.3 7.2 2.1 7.9 0.1 4.5 1.1 5.4 0.5 Domestic price 3.5 2.6 1.7 0.9 2.3 1.7 2.2 0.7 1.2 0.8 0.5 0.2 Output price 3.5 2.5 1.7 0.9 3.1 1.6 2.3 0.6 0.4 0.9 0.6 0.3 Value added price 3.9 2.9 2.0 1.0 4.0 1.6 2.9 0.7 0.1 1.3 0.8 0.2 Consumer price index 2.6 1.0 1.3 1.1 1.2 0.1 Volume Imports 0.9 1.1 1.2 0.3 15.1 2.1 17.0 0.7 16.2 3.2 15.8 1.0 Exports 9.8 0.3 11.0 0.1 8.5 3.5 6.2 1.1 19.5 3.8 18.2 0.9 Domestic demand 0.2 0.2 0.1 0.1 3.1 0.6 2.8 0.2 2.9 0.8 2.6 0.0 Composite good 0.2 0.2 0.0 0.0 0.8 0.0 0.4 0.1 0.6 0.1 0.4 0.1 Output 0.6 0.1 0.6 0.1 2.5 0.3 2.4 0.3 1.9 0.3 1.7 0.2 Value added 0.7 0.1 0.7 0.1 2.3 0.3 2.2 0.3 1.6 0.2 1.5 0.2 Factor prices Nominal wages of skilled workers 2.9 1.0 1.6 0.9 1.3 0.1 Nominal wages of unskilled workers 3.1 1.1 2.2 1.2 0.8 0.1 Nominal returns to land 5.1 n.a. 3.0 n.a. 4.7 n.a. 3.5 n.a. 0.3 n.a. 0.5 n.a. Nominal returns to capital 4.7 2.9 2.9 0.9 5.7 1.3 4.5 0.5 1.2 1.6 1.7 0.4 Source: National CGE model simulations of the authors. Note: Ag agriculture, which includes primary agriculture and lightly processed food. Nonag nonagriculture, which includes other primary products, highly processed foods, manufacturing, and services. n.a. not applicable. 269 270 Agricultural Price Distortions, Inequality, and Poverty therefore expands because domestic demand in agriculture accounts for a larger share of domestic agricultural output. The absence of nonagricultural liberalization results in a 0.1 percent decline in nonagricultural exports because most nonagricultural goods show little or no change in world export demand (table 8.6). Nonagricultural imports rise, while domestic demand declines (0.3 versus 0.1 percent) because domestic prices increase relatively more than import prices (0.9 versus 0.3 percent). The contrac- tion in domestic demand, together with the 0.1 percent decline in exports, leads to a 0.1 percent drop in nonagricultural output. Nonetheless, nonagricultural output and value added prices still increase, owing to higher export and domestic prices. Full unilateral trade liberalization in the Philippines (PHIL-ALL) This experiment eliminates all sectoral tariff rates and export subsidies in the Philippines and assumes that there are no changes in the global model. PHIL-ALL leads to a 7.2 and 2.1 percent decline in the local prices of imported agricultural and nonagricultural products, respectively (table 8.7). Import prices fall more, and import volumes correspondingly increase more in agriculture than in non- agriculture because the initial distortions are greater in agriculture. In the face of cheaper imports relative to domestic prices, the domestic demand declines for local agricultural and nonagricultural producers. Meanwhile, these producers benefit from cost savings on imported inputs, resulting in falls of 2.3 and 1.7 per- cent in the domestic cost of production in the agricultural and nonagricultural sectors, respectively. The real exchange rate depreciates by 1.6 percent, meaning that products made in the Philippines are relatively cheaper in international mar- kets. Coupled with falling domestic prices in the face of cheaper imports and input cost savings, this encourages producers to reallocate resources toward the export market (table 8.7). While exports rise in both agriculture and nonagriculture, domestic demand falls relatively more in agriculture. Because domestic demand represents a larger share in agricultural output, agricultural output contracts, while nonagricultural output expands. Output and value added prices in both agriculture and nonagri- culture fall, but the fall in the former is higher than the fall in the latter. The result of all these adjustments is a fall in all factor returns. The factors used intensively in agriculture experience a much greater reduction: the returns to agricultural capi- tal and land decline by 5.7 and 4.6 percent, respectively, whereas nonagricultural capital returns fall by only 1.3 percent. The nominal wages for unskilled workers fall relatively more than the wages for skilled workers because unskilled workers are used more intensively in agriculture. The Philippines 271 Unilateral trade liberalization only in agriculture in the Philippines (PHIL-AGR) Unilateral liberalization in agriculture and lightly processed food results in sub- stantial expansion in agricultural imports (17 percent) owing to the significant decline in local agricultural import prices (7.9 percent). At the same time, the removal of domestic agricultural distortions also generates cost savings in the export-oriented lightly processed food sector because of this sector's reliance on primary agricultural inputs. Thus, agricultural industries, which, in the context of our study, include lightly processed food, reorient production toward the export market, resulting in a 6.2 percent expansion in exports and a 2.8 percent reduc- tion in domestic sales. The net result is a contraction in agricultural output because domestic sales comprise a larger share of total agricultural output. As a result, agricultural value added and value added prices fall, along with the returns to all agricultural factors. The returns to land drop 3.5 percent, and the returns to agricultural capital decline by 4.5 percent, while the wages of unskilled workers fall by 1.2 percent. The results in the nonagricultural sectors are the opposite. Import prices increase marginally by 0.1 percent, while domestic prices fall by 0.7 percent, resulting in a 0.2 percent expansion in domestic sales. Domestic prices decrease relatively more than world prices (0.7 versus 0.3 percent), leading to a 1.0 percent export expansion. Together with greater domestic demand, this allows overall nonagricultural output to expand by 0.3 percent. A comparison of the sectoral results of PHIL-ALL and PHIL-AGR in table 8.7 confirms the heavier price burden of agricultural protection in the Philippine economy. Indeed, the removal of agricultural distortions accounts for at least two- thirds of the price reduction in exports, imports, domestic production, value added, and the consumer price index.9 This is also true of factor prices: 40 to 80 percent of the fall in factor returns is traceable to the removal of agricultural protection. Rest-of-the-world and Philippine trade liberalization in all sectors (COMB-ALL) This experiment combines rest-of-the-world trade liberalization and domestic trade liberalization. The impact of the rest-of-the-world trade liberalization dom- inates the effects of the unilateral trade liberalization in both agricultural and the nonagricultural sectors. Local import prices decline in the agricultural sector despite the increase in world commodity prices, indicating the substantial level of domestic distortions in the Philippines. The cheaper imports crowd out the corre- sponding domestic products, leading to a contraction in the domestic sales of domestic producers; this affects the agricultural sector more negatively than the 272 Agricultural Price Distortions, Inequality, and Poverty nonagricultural sector. At the same time, rising world export prices, real exchange rate depreciation, and cost savings on imported inputs allow domestic producers to reorient a large share of their production successfully toward the more prof- itable export market. Given the greater reliance of the agricultural sector on domestic sales, the net impact is a contraction in agriculture and an expansion in the nonagricultural sectors. The output prices for both agricultural and the nonagricultural sectors increase (0.4 and 0.9 percent, respectively). Agriculture experiences a smaller increase owing to the substantial level of domestic agricultural distortion in the base. With the exception of the returns to agricultural capital, all factor prices increase. Factor prices increase especially in the nonagricultural sectors, although they increase by less in this scenario than in the ROW-ALL scenario. The impacts in output volumes are the opposite: the effects of unilateral liberalization domi- nate the effects of rest-of-the-world trade liberalization. Thus, overall agricultural output declines, while overall output in the nonagricultural sectors improves ( 1.9 and 0.3 percent, respectively). Rest-of-the-world and Philippine trade liberalization in agriculture only (COMB-AGR) The unilateral agricultural trade liberalization scenario (PHIL-AGR) dominates the global agricultural trade liberalization scenario (ROW-AGR) within the COMB-AGR scenario. In the COMB-AGR scenario, output, prices, and volumes in agriculture decline, as do agricultural factor prices. Local import prices in agri- culture decline despite rising world commodity prices. This indicates that the level of domestic distortions is substantial in the Philippines. The positive impact of the higher world commodity prices is dominated by the negative impact of the domestic agricultural distortions imposed by the government. Thus, the returns to the factors used intensively in agriculture (land, agricultural capital, and unskilled wages) fall in response to the declining agricultural output prices. In contrast, local import prices in the nonagricultural sectors increase. This is expected because the country's nonagricultural trade distortions are already low relative to international standards.10 Thus, the overall output prices of nonagri- cultural products increase by 0.2 percent, resulting in output expansion and, consequently, higher returns to nonagricultural factors (table 8.7). Household income, the consumer price index, and welfare The changes in nominal household incomes, nominal consumer price indexes (based on household-specific consumer baskets), and real incomes and welfare are presented in table 8.8. In interpreting the changes in household-specific con- sumer prices, recall from above that primary and processed food account for a Table 8.8. Effects of Prospective Liberalizations on Real Household Consumption, by Income Group, the Philippines (percent) ROW-ALL ROW-AGR PHIL-ALL Change in Change in Change in Change in Change in Change in Household nominal income consumer price EV nominal income consumer price EV nominal income consumer price EV Decile 1 3.3 2.9 0.4 1.3 1.2 0.1 1.9 2.2 0.3 Decile 2 3.3 2.9 0.4 1.3 1.1 0.1 1.9 2.0 0.2 Decile 3 3.3 2.8 0.4 1.3 1.1 0.1 1.9 2.0 0.1 Decile 4 3.2 2.8 0.4 1.2 1.1 0.1 1.8 1.8 0.0 Decile 5 3.2 2.8 0.4 1.2 1.1 0.1 1.8 1.7 0.1 Decile 6 3.1 2.7 0.4 1.2 1.1 0.1 1.7 1.6 0.2 Decile 7 3.0 2.6 0.4 1.1 1.1 0.1 1.7 1.4 0.2 Decile 8 2.9 2.6 0.4 1.1 1.0 0.1 1.6 1.3 0.3 Decile 9 2.8 2.5 0.3 1.1 1.0 0.0 1.5 1.2 0.3 Decile 10 2.9 2.4 0.5 1.1 1.0 0.1 1.4 1.0 0.4 Overall 3.0 2.6 0.4 1.1 1.0 0.1 1.6 1.3 0.3 (Table continues on the following page.) 273 274 Table 8.8. Effects of Prospective Liberalizations on Real Household Consumption, by Income Group, the Philippines (continued ) (percent) ROW-ALL ROW-AGR PHIL-ALL Change in Change in Change in Change in Change in Change in Household nominal income consumer price EV nominal income consumer price EV nominal income consumer price EV Decile 1 1.2 2.0 0.8 1.4 0.7 0.7 0.1 0.8 0.9 Decile 2 1.2 1.8 0.7 1.4 0.8 0.6 0.1 0.7 0.8 Decile 3 1.2 1.8 0.6 1.4 0.8 0.5 0.1 0.6 0.7 Decile 4 1.2 1.6 0.4 1.4 0.9 0.4 0.1 0.5 0.5 Decile 5 1.1 1.5 0.3 1.4 1.0 0.3 0.1 0.4 0.4 Decile 6 1.1 1.3 0.2 1.3 1.1 0.2 0.1 0.2 0.3 Decile 7 1.1 1.2 0.1 1.3 1.2 0.1 0.0 0.1 0.2 Decile 8 1.1 1.1 0.0 1.3 1.2 0.1 0.0 0.0 0.1 Decile 9 1.1 1.0 0.1 1.3 1.3 0.0 0.0 0.0 0.0 Decile 10 1.0 0.8 0.3 1.4 1.4 0.0 0.0 0.2 0.1 Overall 1.1 1.1 0.0 1.4 1.2 0.1 0.0 0.1 0.1 Source: National CGE model simulations of the authors. Note: EV equivalent variation in income. The Philippines 275 significant share of household expenditure, especially among the lower-income groups, and that the initial tariff rates are higher on primary and processed food items than on other goods (tables 8.4 and 8.5). ROW-ALL, the scenario of global trade liberalization in all sectors, excluding the Philippines, registers the greatest increase in nominal household incomes because rising world prices translate into higher factor returns. For the same rea- son, consumer prices also increase the most in this scenario. Nonetheless, the greater nominal income growth among all households outweighs the detrimental effects of rising consumer prices, with the result that welfare improves among all household groups. Variations in incomes and consumer prices tend to be greater in the poorest deciles, which are more tightly linked to the agricultural sector and which post generally better welfare results in the scenario. The results under ROW-AGR are similar, but less than half as large. The results are qualitatively similar and also display a generally pro-poor effect (table 8.8). The two domestic liberalization scenarios (PHIL-ALL and PHIL-AGR) result in falling consumer prices, which are driven by the reduction in local import and export prices following the elimination of trade-related distortions in the Philippines. This price reduction is greater if the domestic agricultural and nona- gricultural liberalizations are combined. We see that the removal of domestic agricultural distortions reduces consumer prices by more relative to the corre- sponding effect of the removal of nonagricultural distortions given the high share of agriculture in household consumption and the higher initial levels of protec- tion and, hence, consumer prices. A comparison of the changes in consumer prices in scenarios PHIL-ALL and PHIL-AGR show that roughly 1.8 of the 2.2 per- cent reduction in the consumer price index for the first decile is generated by the elimination of domestic agricultural distortions alone (table 8.8). One may also observe that nominal incomes fall under the two unilateral liber- alization scenarios. However, the consumer price effects dominate such that, despite falling nominal incomes, welfare and real incomes increase more under agricultural trade liberalization. Furthermore, these welfare gains accrue propor- tionately more to the poorer deciles owing to the greater relative agricultural consumption among these deciles. These welfare gains are bolstered in rest-of-the-world liberalization and unilat- eral trade liberalization combined. Nominal incomes rise under the full rest-of- the-world and domestic trade liberalization scenario, but this is somewhat offset by soaring consumer prices. Overall, the combined global and domestic agricul- tural trade liberalization scenario (COMB-AGR) provides the greatest increase in welfare. This is because the nominal income gains from the rest-of-the-world trade liberalization are largely conserved, and to these gains are added the con- sumer price reductions from the domestic trade liberalization. In this case, the 276 Agricultural Price Distortions, Inequality, and Poverty poorest deciles emerge as the winners because of both the domestic agricultural trade liberalization and the pro-agricultural nature of the rest-of-the-world trade liberalization. The effects on poverty and inequality The microsimulation process we use in our study relies on the Family Income and Expenditure Survey of the Philippines in 2000 (NSO 2000).11 To estimate the likely poverty and inequality impacts of the labor market conditions arising from trade liberalization, we apply certain information from the CGE model as an input into the microsimulation procedure in a sequential manner. In particular, we use the vectors of changes in the total incomes of households; wage income, capital income, and other income; household-specific consumer price indexes (to update the nominal value of the poverty line); and sectoral employment shares. Through a random process, we incorporate changes in the employment status of households after the simulation. In this way, it is possible to capture household laborers moving in and out of employment (at the microlevel) by taking into account the changes in sectoral employment arising from a policy shift (at the macrolevel). For instance, households that have no labor income because of unemployment may become employed and, consequently, earn labor income. Similarly, employed households may become unemployed and earn no labor income at all after the policy change. Household labor income is affected by changes in wages, as well as by the chance of becoming unemployed after the policy shock. The microsimulation process is repeated 30 times, allowing us to derive confidence intervals on our Foster, Greer, and Thorbecke (1984) indexes and Gini coefficient estimates.12 To take advantage of the richness of the microsimulation procedure, we calcu- late poverty indexes and Gini coefficients based on the demographic characteris- tics of the household heads, especially gender, skill level, and location (urban or rural). In total, the final Foster, Greer, and Thorbecke indexes are derived for eight categories of household heads. Using demographic characteristics instead of income deciles to evaluate changes in poverty and income distribution is prefer- able because it allows for a more accurate policy evaluation and identification of the gainers and losers from trade liberalization. The poverty and inequality results in all experiments are summarized in table 8.9. Inequality marginally worsens in all unilateral liberalization scenarios, but slightly improves in the rest-of-the-world liberalization scenarios. The latter outcome arises because of the increase in nominal incomes among poorer house- holds, while the former outcome results from the greater decrease in the nominal incomes of poorer households relative to richer households (table 8.8). Table 8.9. Income Inequality and the Poverty Effects of Prospective Liberalizations, by Location, Gender, and Skills, the Philippines (percent change relative to the 2000 index) Direct tax Indirect tax replacement replacement 2000 ROW- ROW- PHIL- PHIL- COMB- COMB- COMB- COMB- Group Variable index ALL AGR ALL AGR ALL AGR ALL AGR All Philippines Gini 0.51 0.2 0.2 0.4 0.3 0.2 0.1 0.1 0.1 P0 33.5 0.9 0.6 1.1 0.4 0.1 0.1 0.0 0.1 P1 10.3 1.9 1.2 1.6 0.4 0.4 0.9 0.4 0.7 P2 4.3 2.8 2.0 2.0 0.4 0.8 1.5 0.9 1.3 All urban Gini 0.48 0.2 0.2 0.4 0.3 0.2 0.1 0.2 0.1 P0 18.6 1.6 0.9 2.0 0.8 0.1 0.2 0.1 0.0 P1 5.0 1.9 1.0 3.2 1.7 0.4 0.1 0.4 0.1 P2 2.0 2.0 1.1 4.7 2.7 0.8 0.0 0.9 0.5 Urban men, skilled P0 3.2 0.1 0.0 1.5 1.4 0.7 0.7 1.2 1.6 P1 0.7 1.4 0.2 2.0 3.4 2.2 2.6 3.4 3.4 P2 0.2 1.7 0.4 2.6 4.4 2.7 3.3 4.4 4.3 Urban men, unskilled P0 23.3 1.6 0.9 2.1 0.9 0.1 0.2 0.1 0.1 P1 6.4 1.8 1.0 3.2 1.7 0.5 0.1 0.5 0.2 P2 2.5 2.0 1.1 4.8 2.8 0.8 0.1 1.0 0.5 Urban women, skilled P0 0.9 0.0 0.0 2.3 1.9 1.0 1.5 1.2 1.4 P1 0.1 2.5 0.6 0.8 3.1 2.7 3.4 2.8 3.7 P2 0.0 4.1 1.0 0.2 4.4 4.1 5.2 4.4 5.5 Urban women, unskilled P0 15.2 2.1 1.2 1.8 0.6 0.1 0.4 0.1 0.2 P1 3.9 2.3 1.1 3.7 2.3 0.5 0.0 0.5 0.1 277 P2 1.6 2.4 1.3 5.1 3.4 1.0 0.1 1.0 0.6 (Table continues on the following page.) 278 Table 8.9. Income Inequality and the Poverty Effects of Prospective Liberalizations, by Location, Gender, and Skills, the Philippines (continued ) (percent change relative to the 2000 index) Direct tax Indirect tax replacement replacement 2000 ROW- ROW- PHIL- PHIL- COMB- COMB- COMB- COMB- Group Variable index ALL AGR ALL AGR ALL AGR ALL AGR All rural Gini 0.43 0.1 0.1 0.5 0.2 0.2 0.1 0.1 0.1 P0 48.7 0.9 0.6 0.6 0.0 0.1 0.3 0.2 0.3 P1 15.9 1.4 0.8 1.6 0.4 0.1 0.6 0.2 0.5 P2 6.8 1.7 0.9 2.7 1.1 0.1 0.6 0.0 0.4 Rural men, skilled P0 12.0 0.0 0.0 1.4 2.1 1.6 3.1 1.7 2.8 P1 3.5 1.0 0.2 1.4 2.3 1.8 2.2 1.7 1.9 P2 1.4 1.2 0.3 1.5 2.7 2.2 2.6 2.0 2.3 Rural men, unskilled P0 52.4 0.9 0.6 0.6 0.0 0.1 0.3 0.2 0.3 P1 17.2 1.3 0.7 1.6 0.5 0.1 0.6 0.2 0.5 P2 7.4 1.6 0.9 2.7 1.2 0.1 0.6 0.0 0.4 Rural women, skilled P0 14.7 0.0 0.0 0.7 2.1 0.0 0.8 0.2 1.0 P1 4.1 1.0 0.2 1.1 4.4 1.1 2.5 1.2 2.5 P2 1.4 1.6 0.3 1.6 6.3 1.7 3.7 1.9 3.8 Rural women, unskilled P0 34.9 1.3 0.9 0.8 0.2 0.2 0.4 0.2 0.1 P1 10.8 1.8 1.1 2.0 0.6 0.1 0.6 0.1 0.4 P2 4.4 2.3 1.3 3.7 1.6 0.3 0.4 0.2 0.1 Source: National CGE model simulations of the authors. Note: Gini Gini coefficient. P0 poverty headcount. P1 poverty gap. P2 poverty severity. The Philippines 279 Rest-of-the-world liberalization reduces poverty at the national level and favors unskilled households because rising world demand and the rising export prices of products made in the Philippines bring about higher agricultural factor returns (table 8.7). In contrast, unilateral liberalization favors skilled households such that national poverty indexes worsen. This is because the removal of distor- tions within the Philippines results in resource reallocations toward outward- oriented and externally competitive nonagricultural sectors that employ skilled workers intensively. Poverty generally falls under the combined global and domestic liberalization scenario, wherein the poverty-reducing impact of rest-of-the-world liberalization dominates the poverty-increasing effect of unilateral liberalization. In contrast, inequality marginally rises because the inequality-increasing effect of unilateral liberalization dominates the inequality-reducing effect of rest-of-the-world liber- alization. The combined global and domestic agricultural reform is the most poverty-friendly scenario. Although the national poverty headcount decreases marginally, all household groups, with the slight exception of urban households headed by unskilled workers, share in the benefits of the poverty friendliness of trade liberalization. Indeed, the poorest of the poor, particularly those poor resid- ing in rural areas, emerge as winners because they rely on agricultural production and the wages of unskilled labor. These results are consistent with those obtained by Cororaton, Cockburn, and Corong (2006). However, the results of these authors suggest a worsening in both the poverty gap and the severity of poverty, while our results show the opposite, especially under the combined rest-of-the-world and Philippine agricultural lib- eralization. This difference is traceable to the use of more-recent estimates of trade protection on key food items (such as rice, corn, sugar, and processed meat), which, if eliminated from the calculation, result in a significant fall in the con- sumer prices faced by lower-income groups (table 8.7). Sensitivity analysis: Indirect and direct tax replacement schemes The results discussed above are derived using an indirect tax replacement scheme. Are the results sensitive to the tax replacement scheme used? This subsection compares the above results with the results obtained if a direct tax replacement closure is adopted. We focus on an analysis of the poverty and inequality results of COMB-ALL (full rest-of-the-world and domestic trade liberalization) and COMB-AGR (rest-of-the-world and domestic agricultural trade liberalization). The sensitivity analysis is summarized in table 8.9. The directions of the changes in the poverty indexes and inequality are generally the same regardless of the tax replacement scheme. However, the magnitudes are marginally greater under the direct tax scenario owing to a smaller increase in consumer prices because direct taxes are used to compensate for foregone tariff revenues. 280 Agricultural Price Distortions, Inequality, and Poverty Summary and Policy Implications Starting in the 1980s, the government shifted from taxing to protecting agricul- ture relative to the nonagricultural sectors. However, two decades of protection failed to induce competitiveness and productivity growth because agriculture became inward looking and inefficient. This study analyzes the poverty and inequality implications of the removal of the agricultural and nonagricultural price distortions in the domestic markets of the Philippines as of 2004 and com- pares these effects with the results if policies abroad are liberalized. Rest-of-the-world liberalization reduces poverty at the national level and favors unskilled households in the Philippines because the higher world export prices and rising export demand for products made in the Philippines allow pro- ducers in the Philippines to benefit from favorable international market condi- tions. Nominal incomes improve significantly, outweighing the impact of higher consumer prices. Rest-of-the-world trade liberalization in all sectors generates almost uniform increases in real incomes across household types, while rest-of- the-world trade liberalization in agriculture brings about progressive changes in real incomes that benefit lower-income groups more. Unilateral trade liberalization leads to a drop in consumer prices, which is driven by the reduction in local import and export prices following the elimina- tion of trade-related distortions induced by policies within the Philippines. Import prices fall more, and import volumes correspondingly increase more in agriculture relative to nonagriculture because the initial distortions are greater in agriculture. However, unilateral liberalization favors skilled households such that national poverty indexes and inequality worsen. This is because the removal of the distortions leads to resource reallocation in the Philippines toward outward- oriented and externally competitive nonagricultural sectors that employ skilled workers intensively. The combined global and domestic agricultural reform appears to be the most poverty-friendly scenario for the Philippines. Although the national poverty headcount decreases only marginally, all household groups, with the slight excep- tion of urban households headed by unskilled workers, share in the benefits of the poverty friendliness of the trade liberalization. The poorest of the poor, particu- larly those poor living in rural areas, emerge as winners because of their reliance on agricultural production and the wages from unskilled labor. Thus, it appears to be in the best interest of the Philippines to take a proactive trade liberalization stance by fully participating with the rest of the world in the effort at trade liberal- ization through the inclusion of its own domestic liberalization. Our sensitivity analysis confirms that the results are not affected by differing tax replacement assumptions because a similar pattern of effects emerge regardless of whether indirect or direct tax replacement schemes are used. The Philippines 281 Notes 1. Estimates of agricultural protection and assistance in the Philippines, based on David, Intal, and Balisacan (2009), are incorporated in the World Bank's Distortions to Agricultural Incentives Project Database (Anderson and Valenzuela 2008). The estimates cover five decades. The representative values for CGE modeling as of 2004 that are used here are available in Valenzuela and Anderson (2008). 2. These vectors of changes are generated by simulating the global model with no Philippine trade liberalization. 3. This definition is maintained throughout the study. Agriculture is defined as primary agricul- ture (excluding fishing, forestry, and agricultural services) and lightly processed food, while nonagri- culture refers to all other sectors, including highly processed foods, tobacco, and beverages. 4. The trade reform programs were major components of the structural programs prescribed by multilateral organizations in the 1980s, including the International Monetary Fund and the World Bank. The Philippines is currently in the fourth phase of the trade reform programs. See Cororaton, Cockburn, and Corong (2006) for a detailed discussion. 5. Tariff binding overhang refers to the difference between a country's bound tariffs and its applied tariffs. Bound tariffs are tariff rates that, according to agreements with the World Trade Organ- ization, the country will not exceed. 6. Sugar exports accounted for only 10 percent of domestic production during this period. 7. The specification of the CGE model is based on EXTER (see Decaluwé, Dumont, and Robichaud 2000). For a complete discussion and specification of the model, see Cororaton and Corong (2009), who also provide details on the construction of the social accounting matrix. 8. The export (import) intensity ratio is the ratio of a sector's exports (imports) to output (domestic supply). 9. See the ratio of prices in PHIL-AGR relative to PHIL-ALL in table 8.7: exports ( 1.3 1.8 0.7), imports ( 7.9 7.2 1.1), domestic prices ( 2.2 2.3 1.0), output ( 2.3 3.1 0.8), value added ( 2.9 4.0 0.7), the consumer price index ( 1.1 1.3 0.8), skilled wages ( 0.9 1.6 0.4), unskilled wages ( 1.2 2.5 0.5), the returns to land ( 3.5 4.7 0.7), the returns to agricul- tural capital ( 4.5 5.7 0.8), and the returns to nonagricultural capital ( 0.5 1.3 0.4). 10. This is because previous rounds of trade reform in the country focused primarily on reducing nonagricultural distortions. 11. Our microsimulation is a modified version of the original approach proposed by Vos (2005). 12. Vos (2005) observes that 30 iterations are sufficient because repeating the process additionally does not significantly alter the results. The results on confidence intervals are available from the con- tact author, Caesar Cororaton, upon request. References Aldaba, R. M. 2005. "Policy Reversals, Lobby Groups, and Economic Distortions." PIDS Discussion Paper 2005­04, Philippine Institute for Development Studies, Makati, the Philippines. Anderson, K., and E. Valenzuela. 2008. "Estimates of Global Distortions to Agricultural Incentives, 1955­2007." Data spreadsheet, October, World Bank, Washington, DC. http://go.worldbank.org/ YAO39F35E0. Austria, M. S. 2001. "Liberalization and Regional Integration: The Philippines' Strategy for Rest of the World Competitiveness." Philippine Journal of Development 37 (2): 55­86. Austria, M. S., and E. M. Medalla. 1996. A Study on the Trade and Investment Policies of Developing Countries: The Case of the Philippines. PIDS Discussion Paper 96­03. Makati, the Philippines: Philippine Institute for Development Studies. Cororaton, C. B., J. Cockburn, and E. Corong. 2006. "Doha Scenarios, Trade Reforms, and Poverty in the Philippines: A CGE Analysis." In Poverty and the WTO: Impacts of the Doha Development Agenda, ed. T. W. Hertel and L. A. Winters, 375­402. London: Palgrave Macmillan; Washington, DC: World Bank. 282 Agricultural Price Distortions, Inequality, and Poverty Cororaton, C. B., and E. Corong. 2009. Philippine Agricultural and Food Policies: Implications for Poverty and Income Distribution. IFPRI Research Report 161. Washington, DC: International Food Policy Research Institute. David, C. C. 2003. "Agriculture." In The Philippine Economy: Development, Policies, and Challenges, ed. A. M. Balisacan and H. Hill, 175­218. Quezon City, the Philippines: Ateneo de Manila Press. David, C. C., P. S. Intal Jr., and A. M. Balisacan. 2009. "The Philippines." In Distortions to Agricultural Incentives in Asia, ed. K. Anderson and W. Martin, 223­54. Washington, DC: World Bank. Decaluwé, B., J.-C. Dumont, and V. Robichaud. 2000. Basic CGE Models. Vol. II of MIMAP Training Session on CGE Modeling. Micro Impacts of Macroeconomic and Adjustment Policies Program, International Development Research Centre, Canadian Centre for International Studies and Cooperation, and Université Laval, Quebec. http://www.pep-net.org/fileadmin/medias/pdf/ volume2.pdf. Foster, J., J. Greer, and E. Thorbecke. 1984. "A Class of Decomposable Poverty Measures." Econometrica 52 (3): 761­66. Horridge, M., and F. Zhai. 2006. "Shocking a Single-Country CGE Model with Export Prices and Quantities from a Global Model." In Poverty and the WTO: Impacts of the Doha Development Agenda, ed. T. W. Hertel and L. A. Winters, 94­104. London: Palgrave Macmillan; Washington, DC: World Bank. Intal, P. S., Jr., and J. H. Power. 1990. Trade, Exchange Rate, and Agricultural Pricing Policy in the Philippines. World Bank Comparative Studies. Washington, DC: World Bank. NSO (National Statistics Office). 1985­. Family Income and Expenditure Survey, the Philippines. Manila: NSO. http://www.census.gov.ph/data/publications/pubfies.html. Valenzuela, E., and K. Anderson. 2008. "Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980­84." Research Memorandum 13 (December), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925. van der Mensbrugghe, D. 2005. "Linkage Technical Reference Document: Version 6.0." December, World Bank, Washington, DC. http://go.worldbank.org/7NP2KK1OH0. Vos, R. 2005. "Microsimulation Methodology: Technical Note." Unpublished paper, Department of Economic and Social Affairs, United Nations, New York. WTO (World Trade Organization). 1999. Trade Policy Review: The Philippines. Geneva: WTO. ------. 2005. Trade Policy Review: The Philippines. Geneva: WTO. 9 Thailand Peter Warr* As Thailand has industrialized, successive Thai governments have become increasingly interested in intervening on behalf of producers in the key declining sector, namely, agriculture.1 Agricultural producers and processors are the intended beneficiaries of these interventions. However, the fact that Thailand is a major agricultural exporter has limited the scope of protection policies as a means of influencing domestic commodity prices. Over time, the direct taxation of agri- cultural exports has been gradually eliminated. This has been important in the case of rice, on which the high rates of export taxation prior to the mid-1980s were abolished in 1986. Similarly, rubber exports, taxed prior to 1990, have not been taxed since then. Cassava exports have continued to be taxed to a minor extent through the system of export quotas. Taxes on imports of fertilizer, a major input into agricultural production, have been steadily phased out since the early 1990s. Maize exports have been consistently untaxed, as have chicken exports, a commodity not covered by the present analysis because of the lack of suitable price data. Most of this history is a tale of the elimination of the price distortions that once acted against agricultural export industries. Four commodities depart from this general trend of the liberalization of agricul- tural markets. Soybeans were an export item prior to 1992 and have been a net import item since then; the imports have been subject to quota restrictions. The change from a net export item to a net import item coincided with a switch from *The author has benefited greatly from the research assistance of Krisada Bamrungwong and Arief Anshory Yusuf, from helpful discussions with Ernesto Valenzuela, and from the comments of John Cockburn. This chapter relies on a modeling approach that is similar to the approach in the author's chapter on Indonesia in this volume (chapter 6) and is meant to be read in conjunction with that chapter. Methodological points that are common to the two studies are not repeated here. The discussion here focuses on methodological points that are different and on the results of the analysis on Thailand. 283 284 Agricultural Price Distortions, Inequality, and Poverty negative to positive nominal rates of protection. Since the early 1990s, the domestic soybean industry has received a nominal rate of protection of between 30 and 40 percent. Sugar is an export commodity in Thailand, but the domestic sugar industry is protected by a home price system that taxes domestic consumers and transfers the revenue to producers. Nominal rates of protection have averaged over 60 percent. The political power of the highly capital-intensive sugar milling industry explains this pattern of protection. The case of palm oil is qualitatively similar to the case of sugar, but the rates of protection are somewhat lower. Finally, Thailand's small dairy industry is protected from competition by imported milk powder. It has not been possible to obtain sufficient data to quantify dairy protection for the purposes of our study, but informed sources report that the rate of protection is comparable with the rate for sugar. The prospects for additional trade liberalization in Thailand are not encouraging unless this occurs through bilateral preferential trading arrangements such as the scheme proposed with the United States.2 Almost all of Thailand's poor people reside in rural areas, and most are directly involved in agricultural production (Warr 2005). The Thai public is well disposed to finding ways to reduce rural poverty, and Thai governments have responded to this sentiment. Interventions on behalf of rural people have been important, but Thailand is unusual in that, except for the cases discussed above, these interven- tions have seldom taken the form of initiatives in agricultural commodity mar- kets. The strong export-orientation of Thai agriculture is an important reason for this because assistance through commodity markets would require a large pro- duction of export subsidy outlays. Instead, cash transfers to village organizations, subsidized loan schemes not linked to agricultural production, and a generally good system of public infrastructure have been the main instruments of interven- tion in support of rural areas. Unfortunately, with the exception of the investment in rural infrastructure, these transfers have seldom been directed in any systematic way at raising the productivity of rural people or at assisting them in finding bet- ter economic opportunities outside agriculture. The long-term contribution of these transfers to any reduction in rural poverty will probably be small. We use the JamlongThai general equilibrium model of the Thai economy (Warr 2008b), in conjunction with the Linkage global economic model (van der Mensbrugghe 2005), to analyze the effects on the incidence of poverty in Thailand of agricultural and other policies at home and abroad. JamlongThai is structurally similar in most respects to the Wayang model of the Indonesian economy, described in detail in chapter 6 by Warr, and the simulations we perform using this model are also similar to those described in the Indonesian case study. The next section describes the JamlongThai general equilibrium model of the Thai economy, the principal analytic tool we use in this study. The subsequent sec- tion describes the simulations we perform with this model, in combination with Thailand 285 the global Linkage model. The simulations involve both unilateral agricultural and trade policy reform in Thailand and reform by the rest of the world. We do this to assess the relative importance of own-country versus rest-of-the-world policies to Thai households. We present the results with a focus on the implica- tions for poverty incidence within Thailand. The final section concludes. A General Equilibrium Model of the Thai Economy: JamlongThai JamlongThai is a general equilibrium model of the Thai economy based on 65 sectors and 200 households. It has been constructed for the analysis of the effect of trade policy and other policy shocks on poverty incidence in Thailand. Unless otherwise stated, the database of the model refers to the year 2000. JamlongThai shares many structural features with the highly influential ORANI general equilibrium model of the Australian economy (Dixon et al. 1982) and the general equilibrium model of the global economy established through the Global Trade Analysis Project (Hertel 1997), which also belong to the Johansen (1964) category of economy-wide models. The specific structure of JamlongThai draws on a revised version of the ORANI model, called ORANI-G (Horridge 2004), and the PARA and Wayang general equilibrium models of the Thai and Indonesian economies, respectively, described in detail in Warr (2001, 2005). We have adapted this general structure of the model to reflect the specific objectives of our study and important features of the Thai economy. Industries The JamlongThai model contains 65 industries, of which 24 are in agriculture. Each industry produces a single output, and the set of commodities therefore coincides with the set of industries. Exports are not identical with domestically sold commodities. In each industry, the two are produced by a transformation process with a constant elasticity of transformation.3 The core of the production side of the model is an input-output table of 65 sectors. This table has been aggre- gated from an input-output table of 180 sectors that has been produced by the National Economic and Social Development Board (NESDB 2004). At the time of our study, the latest input-output table available for Thailand related to the year 2000. We have used this data set for our study. The cost structure of the Thai economy, with an emphasis on the agricultural sector, is summarized in table 9.1. The agricultural sector relies on almost no skilled (educated) labor, but it is intensive in the use of unskilled labor. Unskilled and skilled labor are fully mobile across all industries. 286 Agricultural Price Distortions, Inequality, and Poverty Table 9.1. Cost Shares of the Major Factors of Production, Paddy and Other Industries, Thailand, 2000 (percent of total costs) Other All Cost components Paddy agriculture Nonagriculture industries Skilled 0.0 0.0 10.5 9.9 Unskilled 24.9 15.0 4.6 5.2 Mobile capital 2.1 5.1 29.3 28.0 Land 45.7 44.1 0.0 2.3 Intermediate inputs 27.3 35.8 55.7 54.6 Total 100.0 100.0 100.0 100.0 Source: Database of the JamlongThai model, based on NESDB (2004) and agricultural cost survey data of the Ministry of Agriculture and Cooperatives. Table 9.2. Sources of the Factor Incomes of Broad Household Groups, Thailand, 2000 (percent) Total Mobile factor Household category Skilled Unskilled capital Land income 1. Farm, northeast 15.5 28.0 42.8 13.6 100.0 2. Farm, north 6.3 15.5 44.0 34.2 100.0 3. Farm, other 10.1 12.9 47.5 29.6 100.0 4. Farm worker, all regions 9.6 78.3 6.0 6.1 100.0 5. Entrepreneur, paid employees 2.2 0.7 97.0 0.1 100.0 6. Entrepreneur, no paid employees 1.6 1.6 96.6 0.2 100.0 7. Professional 71.2 4.8 20.1 3.9 100.0 8. Clerical employee 47.5 27.7 21.3 3.5 100.0 9. Production and construction 31.6 51.8 13.5 3.1 100.0 10. Economically inactive 7.8 4.6 38.2 50.1 100.0 Source: Database of the JamlongThai model, based on NSO (2003). Households The JamlongThai model contains 10 major household categories, which are based on the socioeconomic survey classifications used by the National Statistical Office (NSO 2003). Table 9.2 summarizes these 10 household categories, which are classi- fied according to the occupation of the heads of household. Four are farm house- holds: farmer, northeast; farmer, north; farmer, other region; and farm worker, all regions. Six are nonfarm: entrepreneur with paid employees; entrepreneur without paid employees; professional; clerical employee; production, construction, or general laborer; and economically inactive. Each of these categories is divided into Thailand 287 20 groups of equal population size that are sorted by household income per person.4 There are thus a total of 200 household subcategories. The incomes of each of these 200 household subcategories depend on the household ownership of factors of production, the returns to these factors, and nonfactor incomes, mainly transfers from others. Since our focus is on income distribution, the sources of the incomes of the various households are of particu- lar interest. These differ significantly among the 10 household categories. Our household data are extracted from the 2000 household socioeconomic sur- vey of the National Statistical Office (NSO 2003). A social accounting matrix has been constructed for Thailand based on data of the household socioeconomic sur- vey, the input-output tables (NESDB 2004), the national accounts, and trade data, all for 2000. The database for the JamlongThai model draws upon this matrix. The consumer demand equations for the various household types are based on a Cobb-Douglas demand system, using data on expenditure shares extracted from the 2000 household socioeconomic survey. Within each of the 10 major cate- gories, the 20 subcategories thus differ according to per capita expenditures, budget shares in consumption, and sources of factor and nonfactor incomes. Table 9.2 summarizes the sources of the factor incomes of the 10 major house- hold categories, and table 9.3 summarizes the share of these categories in the over- all population and in the overall incidence of poverty in Thailand based on the official poverty line. Table 9.3. Characteristics of the Major Household Categories, Thailand, 2000 (number and percent) Total population, Share of total Poverty Household category millions population, % incidence,a % 1. Farm, northeast 8,460 13.9 40.7 2. Farm, north 3,287 5.4 27.3 3. Farm, other 4,920 8.1 14.5 4. Farm worker, all regions 4,565 7.5 29.3 5. Entrepreneur, paid employees 1,953 3.2 0.7 6. Entrepreneur, no paid employees 8,354 13.7 5.9 7. Professional 4,866 8.0 1.6 8. Clerical employee 7,934 13.0 2.5 9. Production and construction 8,967 14.7 7.7 10. Economically inactive 7,613 12.5 11.6 1­10. Total population 60,916 100.0 14.4 Source: Author calculations based on NSO (2003). a. Headcount measure of poverty incidence using the national poverty line. 288 Agricultural Price Distortions, Inequality, and Poverty Simulations The effects of agricultural trade policy reform are simulated using the JamlongThai model of the Thai economy, combined with the Linkage model of the world economy (van der Mensbrugghe 2005). The simulations involve both unilateral agricultural and trade policy reform in Thailand and reform by the rest of the world. We do this to assess the relative importance of own-country versus rest-of-the-world policies for Thai households. We also compare the reform of agriculture alone with the reform of all goods markets to gauge the relative contribution of agricultural policies to the measured impacts on households. We consider three sets of policy reforms below: unilateral reform in Thailand (simulation A); global reform, excluding Thailand (simulation B); and the combi- nation of these two sets of reforms, that is, global reform, including Thailand (simulation C). The structure of these simulations is identical to the structure described for the simulations on Indonesia by Warr in chapter 6; the reader should therefore refer to chapter 6 for details. By reform, we mean the complete elimination of all tariffs, the tariff equivalents of any nontariff barriers, export taxes and export subsidies, and domestic agricultural policies in so far as they alter the producer or consumer prices of farm products in various countries. Each of our three sim- ulations is conducted twice: one time each in which the reductions to protection in Thailand and in the rest of the world apply to all commodities (labeled simu- lations A1, B1, and C1) and one time each in which the reductions apply only to agricultural commodities, including lightly processed food (labeled simulations A2, B2, and C2). The initial rates of industry assistance in Thailand used in the simulations are derived from the database of the Linkage model (see table 9.4). We have reduced these rates to zero in simulations A and C. In simulations B and C, we use the Linkage model simulations conducted by van der Mensbrugghe, Valenzuela, and Anderson (see the appendix) to estimate the changes in the import prices and export demand shifters that result from the liberalization in all countries except Thailand. We then apply these results as shocks in the JamlongThai model. The changes to border prices, derived from the Linkage model, are shown in table 9.5. The results of the liberalization of markets for all goods While our emphasis in this study is on the effects on poverty and income inequality, an understanding of these effects requires that we look first at the macroeconomic effects. Thailand 289 Table 9.4. Industry Assistance Rates Used in Modeling, Thailand, 2004 (percent) Commodity Tariff Export subsidy Output subsidy Paddy rice 12.5 0.0 0.3 Wheat 26.8 0.0 0.2 Other grains 26.5 0.0 0.2 Vegetables and fruits 44.6 0.0 0.1 Oilseeds 32.9 0.0 0.1 Sugarcane 0.0 0.0 0.1 Plant-based fibers 5.0 0.0 0.2 Other crops 38.0 0.0 0.1 Cattle, sheep, and so on 9.6 0.0 0.1 Other livestock 13.9 0.0 0.1 Raw milk 0.0 0.0 0.1 Wool 7.2 0.0 0.1 Other primary products 2.4 0.9 2.4 Beef and sheep meat 49.3 0.0 1.2 Other meat products 40.9 0.0 1.2 Vegetable oils and fats 39.2 0.0 1.2 Dairy products 18.1 0.0 1.2 Processed rice 19.6 0.0 0.5 Refined sugar 0.0 0.0 5.0 Other food, beverages, tobacco 39.7 0.0 19.9 Textiles and wearing apparel 23.5 0.6 1.1 Other manufacturing 13.9 0.0 3.3 Services 0.0 0.0 3.7 Source: Valenzuela and Anderson (2008), based on the estimates compiled by Anderson and Valenzuela (2008). Macroeconomic effects The simulated macroeconomic effects of trade reform in all goods markets are summarized in table 9.6 (simulations A1, B1, and C1). Real GDP rises in Thailand under all three reform scenarios. Aggregate real household consumption also rises. The increases in real GDP and aggregate real household consumption in simulation A1 far exceed the corresponding increases in simulation B1. Unilateral liberalization is more beneficial to Thailand in macroeconomic terms than the liberalization of all other countries combined. By turning to the effects of these reforms on real factor returns, we may discern clues to the income distributional impacts of the reforms. The real factor returns, deflated by the consumer price index, rise in the case of unskilled labor, skilled 290 Agricultural Price Distortions, Inequality, and Poverty Table 9.5. Exogenous Border Price Shocks Caused by Rest-of-the-World Liberalization, Thailand (percent deviation from base) Export price shocksa Import price shocksb Reform of Agricultural Reform of Agricultural Commodity all goods reform only all goods reform only Paddy rice 7.3 5.7 0.0 0.0 Wheat 0.0 0.0 3.3 4.0 Other grains 6.4 4.8 4.9 4.5 Oilseeds 6.0 4.7 6.0 5.6 Sugarcane and beets 0.0 0.0 0.0 0.0 Plant-based fibers 6.8 5.2 6.4 7.9 Vegetables and fruits 7.2 5.6 1.8 1.1 Other crops 7.3 5.8 1.7 1.6 Cattle, sheep, and so on 6.2 4.6 3.6 2.8 Other livestock 5.4 3.7 0.6 1.1 Raw milk 0.0 0.0 0.0 0.0 Wool 0.0 0.0 5.5 5.3 Beef and sheep meat 4.1 2.2 10.0 10.1 Other meat products 4.5 2.7 1.6 2.5 Vegetable oils and fats 2.6 0.9 0.7 1.2 Dairy products 4.4 2.6 12.2 12.5 Processed rice 6.6 4.9 1.4 0.5 Refined sugar 4.6 2.7 0.0 0.0 Other food, beverages, tobacco 3.5 1.7 0.9 1.4 Other primary products 3.3 1.2 0.5 0.8 Textiles and wearing apparel 3.2 1.4 0.4 0.5 Other manufacturing 2.9 1.1 1.3 0.3 Services 3.3 1.2 0.2 0.2 Agriculture and food 4.6 3.0 1.7 0.9 Agriculture 7.0 5.5 1.1 1.6 Processed foods 4.3 2.7 2.1 0.5 Other manufacturing 3.0 1.2 1.1 0.3 Nontradables 3.3 1.2 0.2 0.2 Total 3.3 1.4 1.0 0.4 Merchandise trade 3.3 1.5 1.2 0.4 Source: Linkage model simulations by van der Mensbrugghe, Valenzuela, and Anderson (see the appendix). a. Simulated as shocks to the inverse export demand equations for Thailand. b. Simulated as shocks to the exogenous import prices for Thailand. Thailand 291 Table 9.6. Aggregate Simulation Results of the Prospective Liberalization of All Commodities, Thailand Sim A1: Sim B1: rest- Sim C1: unilateral of-the-world global Indicator liberalization liberalization liberalization Macroeconomic aggregates, % change from base Real GDP, expenditure side, GDP deflator 0.27 0.04 0.30 Real household consumption, CPI deflator 4.32 0.16 4.43 Import volume index, duty-paid weights 4.68 0.12 4.79 Export volume index 0.04 0.01 0.03 GDP price index, expenditure side 8.24 2.11 10.49 Consumer price index 7.69 2.64 10.45 Nominal change, baht, millions GDP 482,697 121,566 611,829 Consumption 397,015 90,448 493,626 Investment 7,775 19,905 28,127 Inventory 2,516 38 2,610 Government expenditure 75,390 11,175 87,466 Real returns to factors, % change from base, using CPI deflator Unskilled labor 3.3 0.5 2.8 Skilled labor 10.0 0.8 9.1 Agricultural capital 3.3 0.3 3.5 Nonagricultural capital 9.9 0.3 9.5 Land 3.8 0.5 4.2 Real household expenditure, % change from base, using CPI deflator 1. Farm, northeast 2.6 0.8 3.3 2. Farm, north 2.9 0.7 3.5 3. Farm, other 4.4 0.6 4.9 4. Farm worker, all regions 1.1 0.5 1.6 5. Entrepreneur, paid employees 5.7 0.1 5.8 6. Entrepreneur, no paid employees 5.7 0.3 6.0 7. Professional 5.7 0.4 5.3 8. Clerical employee 4.3 0.1 4.1 9. Production and construction 2.7 0.1 2.7 10. Economically inactive 3.8 0.4 4.1 Source: Thailand computable general equilibrium model simulations by the author. Note: CPI consumer price index. 292 Agricultural Price Distortions, Inequality, and Poverty labor, and nonagricultural capital in simulation A1 (unilateral liberalization). The returns to agricultural capital and land decline following liberalization, indicating that the structure of protection favors industries that are intensive in the use of agricultural capital and land. Another way of viewing this result is that the struc- ture of protection in Thailand is such that the real returns to skilled and unskilled labor, as well as nonagricultural capital, are lowered, while the real returns to agricultural capital and land are raised. The removal of the protection has the opposite effect. Simulation C1 is approximately the sum of simulations A1 and B1. It is dominated by the effects of unilateral liberalization. In all three reform scenarios, real household expenditures increase among all household categories except household group 7, the households headed by profes- sionals. The negative effect under simulation B1 is small relative to the gains achieved by most other household categories. The effects on poverty incidence and inequality The poverty effects of the reforms are summarized in table 9.7. The measured level of the incidence of poverty depends on the poverty line used in the calcula- tions, and this may also be true of the simulated changes in poverty incidence that result from particular economic shocks. We present the effects on the incidence of poverty using two poverty lines: the national poverty line and the international US$2-a-day poverty line at purchasing power parity.5 For our calculations with each of the poverty lines, we use the same calibration method, as follows. First, we assemble the data on the ex ante distribution of household expenditures according to information in the model database. Second, we use the published level of poverty incidence based on one and then the other of the two poverty lines to find the value of the poverty line, measured in the domestic currency, that generates the particular level of poverty incidence we have found in the data on household expenditures in the model database. We take these published levels of poverty incidence from data of the National Economic and Social Development Board in the case of the national poverty line and from PovcalNet (2008) in the case of the US$2-a-day poverty line. This then becomes the base level of the poverty line we use in our subsequent calculations. Third, the ex post levels of real expenditure for each household group are sim- ulated within the model so as to reflect the effects of the shocks applied to the model. We perform these calculations of real household expenditures using the individual household consumer price index as the deflator so as to reflect the par- ticular household's consumption bundle. Fourth, we then compare these ex post real expenditures with the poverty line described above to obtain ex post levels of poverty incidence. Thailand 293 Table 9.7. The Poverty and Inequality Effects of the Prospective Liberalization of All Commodities, Thailand Change, ex post ­ ex ante Ex Sim A1: Sim B1: rest- Sim C1: ante unilateral of-the-world global Group level liberalization liberalization liberalization Poverty incidence at the national poverty line, % 1. Farm, northeast 40.7 6.2 2.9 9.1 2. Farm, north 27.3 2.6 1.3 4.0 3. Farm, other 14.5 12.2 5.8 18.0 4. Farm worker, all regions 29.3 3.4 0.2 3.6 5. Entrepreneur, paid employees 0.7 0.7 0.0 0.7 6. Entrepreneur, no paid employees 5.9 11.5 5.7 17.2 7. Professional 1.6 2.3 1.0 3.3 8. Clerical employee 2.5 3.0 1.0 4.1 9. Production and construction 7.7 9.0 6.7 15.7 10. Economically inactive 11.6 4.9 0.6 7.5 Farm households 30.1 3.5 0.9 4.4 Nonfarm households 5.9 3.3 0.6 3.9 All households 14.4 3.4 0.7 4.1 Poverty incidence at the US$2-a-day poverty line, % Farm households 21.03 1.55 0.31 1.86 Nonfarm households 15.19 1.10 1.18 2.28 All households 25.20 1.84 1.00 2.84 Inequality, Gini coefficient Urban households 0.152 0.006 0.006 0.012 Rural households 0.334 0.004 0.005 0.009 All households 0.339 0.005 0.007 0.012 Source: Thailand computable general equilibrium model simulations by the author. Finally, the changes in poverty incidence reported in table 9.7 and table 9.9 (see below) are the ex post levels of poverty incidence, minus the ex ante levels corre- sponding to each of the two poverty lines described above. A positive number thus indicates an increase in the simulated level of poverty incidence as a result of the shocks identified. In addition to the effects on poverty incidence, we also report simulated effects on inequality in the distribution of household real expenditures. We accomplish this using the Gini coefficient as the measure. The Gini coefficient takes values 294 Agricultural Price Distortions, Inequality, and Poverty between zero and 1; higher values reflecting greater inequality. These coefficients are estimated by constructing Lorenz curves from the distributions of ex ante and ex post real expenditures and then calculating the Gini coefficients corresponding to these distributions. These results are also presented in table 9.7. Unilateral liberalization (simulation A1) delivers reduced poverty to all house- hold categories and to farm and nonfarm households in aggregate. The benefits in terms of poverty reduction in Thailand that derive from global across-the-board liberalization (simulation C1) arise primarily from Thailand's own liberalization. Only one-fifth of the total reduction in the incidence of poverty may be attributed to the effects of liberalization elsewhere (simulation B1). Liberalization in Thailand and in the rest of the world increases inequality within Thailand some- what. Skilled labor gains proportionately more than unskilled labor. However, in absolute terms, poor households have a strong stake in trade liberalization, espe- cially liberalization within their own country regardless of whether or not other countries participate. The results of the liberalization of only agricultural markets The macroeconomic effects of liberalization that is confined to agricultural prod- ucts alone are summarized in table 9.8. The simulated effects are quite different from those arising from liberalization in the markets for all goods. Unilateral agri- cultural liberalization in Thailand raises real household consumption and mar- ginally raises the consumer price index. It reduces the real value of unskilled labor marginally and increases skilled wages significantly. It also reduces the returns to agricultural capital, along with the returns to land. Relative to the rest of the econ- omy, the agricultural industries in Thailand use much smaller shares of skilled labor and much larger shares of unskilled labor. Reduced agricultural protection reduces the size of the agricultural sector and raises the size of the nonagricultural sector. In the process, it raises real skilled wages and lowers real unskilled wages, along with the real returns to agricultural capital and land. Unilateral agricultural liberalization in Thailand raises the incidence of poverty among farm households (table 9.9) because it reduces the real returns to unskilled labor (slightly)--deflated by the consumer price index--and also the real returns to agricultural capital and land (more significantly). Agricultural lib- eralization in the rest of the world raises the international prices of Thailand's agricultural exports (table 9.4). This increases the real returns to skilled and unskilled labor and the returns to land in Thailand. Farm-level poverty in Thailand declines, along with poverty among nonfarm households. Measured solely by the incidence of poverty in Thailand, these results suggest that Thailand has a stake in agricultural liberalization, but only if other countries participate as well. Thailand 295 Table 9.8. Aggregate Simulation Results of the Prospective Liberalization of Only Agricultural Commodities, Thailand Sim A2: Sim B2: rest- Sim C2: unilateral of-the-world global Indicator liberalization liberalization liberalization Macroeconomic aggregates, % change from base Real GDP, expenditure side, GDP deflator 0.01 0.03 0.04 Real household consumption, CPI deflator 0.12 0.14 0.25 Import volume index, duty-paid weights 0.12 0.10 0.21 Export volume index 0.00 0.01 0.01 GDP price index, expenditure side 0.21 0.99 1.20 Consumer price index 0.12 1.25 1.37 Nominal change, baht, millions GDP 12,633 57,660 69,886 Consumption 7,733 44,919 52,352 Investment 2,587 7,767 10,304 Inventory 80 51 131 Government expenditure 2,232 4,923 7,099 Real returns to factors, % change from base, using CPI deflator Unskilled labor 0.07 0.25 0.18 Skilled labor 0.51 0.36 0.87 Agricultural capital 1.34 0.20 1.14 Nonagricultural capital 0.47 0.08 0.52 Land 1.46 0.42 1.04 Real household expenditure, % change from base, using CPI deflator 1. Farm, northeast 0.16 0.47 0.32 2. Farm, north 0.10 0.42 0.32 3. Farm, other 0.07 0.37 0.44 4. Farm worker, all regions 0.11 0.33 0.23 5. Entrepreneur, paid employees 0.20 0.14 0.33 6. Entrepreneur, no paid employees 0.24 0.26 0.48 7. Professional 0.26 0.15 0.10 8. Clerical employee 0.15 0.00 0.15 9. Production and construction 0.03 0.10 0.12 10. Economically inactive 0.05 0.26 0.30 Source: Thailand computable general equilibrium model simulations by the author. Note: CPI consumer price index. 296 Agricultural Price Distortions, Inequality, and Poverty Table 9.9. The Poverty and Inequality Effects of the Prospective Liberalization of Only Agricultural Commodities, Thailand Change, ex post ­ ex ante Ex Sim A2: Sim B2: rest- Sim C2: ante unilateral of-the-world global Group level liberalization liberalization liberalization Poverty incidence at the national poverty line, % 1. Farm, northeast 40.7 0.3 0.8 0.5 2. Farm, north 27.3 0.1 0.3 0.2 3. Farm, other 14.5 0.1 1.1 1.1 4. Farm worker, all regions 29.3 1.0 4.7 3.7 5. Entrepreneur, paid employees 0.7 0.4 0.7 1.1 6. Entrepreneur, no paid employees 5.9 0.1 0.2 0.3 7. Professional 1.6 0.0 0.6 0.6 8. Clerical employee 2.5 0.1 1.5 1.4 9. Production and construction 7.7 0.2 1.0 0.8 10. Economically inactive 11.6 0.1 0.5 0.4 Farm households 30.1 0.3 1.6 1.3 Nonfarm households 5.9 0.0 0.8 0.7 All households 14.4 0.1 1.1 0.8 Poverty incidence at the US$2-a-day poverty line, % Farm households 21.03 0.09 0.31 0.24 Nonfarm households 15.19 0.03 0.14 0.17 All households 25.20 0.04 0.24 0.28 Inequality, Gini coefficient Urban households 0.152 0.001 0.006 0.007 Rural households 0.334 0.000 0.005 0.005 All households 0.339 0.001 0.007 0.008 Source: Thailand computable general equilibrium model simulations by the author. Conclusions and Comparison with the Study on Indonesia The comparative static analysis in our study indicates that across-the-board trade liberalization is poverty reducing in Thailand, whether other countries participate in the liberalization or not. This poverty reduction occurs among farm and non- farm households, and the qualitative outcome is not dependent on the particular poverty line used in the analysis. The reduction in farm and nonfarm poverty Thailand 297 incidence occurs despite an increase in inequality in Thailand. Liberalization raises real skilled wages relative to real unskilled wages, and this effect increases inequality. Both Thailand's own liberalization and the liberalization of the rest of the world reduce poverty among farm and nonfarm households, but the largest benefits from across-the-board liberalization, measured in terms of the effects on poverty, arise from Thailand's own liberalization. If the trade liberalization is confined to agricultural products alone, the results are somewhat different. A similar increase in inequality occurs, but unilateral agri- cultural liberalization in Thailand raises the incidence of poverty among farm households, while reducing it slightly among nonfarm households. This negative effect on rural households arises from the reduction in real unskilled wages. If the rest of the world also liberalizes agricultural trade, the increase in farm-level poverty in Thailand disappears. Thailand's farm poor thus have an interest in agricultural liberalization, but only if the rest of the world also liberalizes. Our analyses of trade liberalization in Thailand in this chapter and the study by Warr on Indonesia in chapter 6 use similar modeling frameworks, but the struc- ture of agricultural trade and agricultural protection in the two countries is quite different. A direct comparison of the simulated effects of liberalized trade policies is therefore possible and offers insights. The unilateral liberalization of all commodities reduces the incidence of poverty among farm and nonfarm households in both countries. In both coun- tries, skilled and unskilled real wages rise as a result of this liberalization. How- ever, the magnitude of the rise in unskilled wages is greater in Thailand. This reveals a key difference in the effects of the overall structure of the distortions to incentives in the two countries. In Thailand, industry assistance policies act more strongly against the interests of unskilled workers by favoring industries that are less intensive in the use of this factor. The result is that across-the-board liberal- ization has a much larger poverty-reducing effect in Thailand than in Indonesia, although it reduces the incidence of poverty among all socioeconomic groups in both countries. Rest-of-the-world liberalization in all commodities is poverty reducing among all socioeconomic groups in both countries, but the magnitude of this effect is larger in Indonesia. Rest-of-the-world liberalization raises the prices of agricul- tural products relative to the prices of manufactured goods. Indonesia's poor are more dependent on agricultural production, and the magnitude of the poverty reduction that occurs following rest-of-the-world liberalization is therefore greater in Indonesia. If liberalization occurs only in agricultural products, the returns to agricultural capital and land decline in both countries. In Indonesia, unskilled wages also decline significantly, whereas, in Thailand, this effect is negligible. This divergence in 298 Agricultural Price Distortions, Inequality, and Poverty outcomes reflects a difference in the structure of agricultural assistance in the two countries. Indonesia's agricultural distortions are more biased toward unskilled labor­intensive agricultural industries, of which rice--highly protected in Indonesia, but not in Thailand--is the most important example. Meanwhile, because rice is a staple consumer good among Indonesia's poor, a reduction in agricultural assistance benefits many poor households, rural as well as urban. The outcome of unilateral agricultural liberalization is, in both countries, a reduction in the incidence of urban poverty and an increase in the incidence of rural poverty. The net effect is a small overall reduction in national poverty incidence in Indonesia and a small overall increase in Thailand. Rest-of-the-world liberalization in agricultural products alone is poverty reducing among all socioeconomic groups in both countries. Overall, the key results common to both countries are as follows: · The poor have a strong interest in across-the-board liberalization both in their own country and in the rest of the world; this applies to the rural poor and the urban poor. · The urban poor have an interest in unilateral agricultural liberalization, while the rural poor do not. · Agricultural liberalization in the rest of the world is poverty reducing among all socioeconomic groups in rural and urban areas. Notes 1. Structural change in the Thai economy is reviewed by Warr (2007), and a fuller discussion of Thailand's agricultural trade policies is provided by Warr (2008a) and Warr and Kohpaiboon (2009). The estimates of agricultural assistance for Thailand supplied by Warr and Kohpaiboon (2009) are incorporated in the World Bank's Distortions to Agricultural Incentives Project Database (Anderson and Valenzuela 2008). The estimates cover four or more decades. The representative values for com- putable general equilibrium modeling as of 2004 that are used here are available in Valenzuela and Anderson (2008). 2. A bilateral trading arrangement with the United States was under negotiation prior to February 2006, but the negotiations are currently suspended. The protection of Thailand's soybean industry would be an important issue in the negotiations. 3. This treatment differs from that used by us in the Indonesian case study in chapter 6, in which the commodities exported and sold domestically are identical, but the quantity of exports in import- competing industries is fixed exogenously. 4. The population sizes of the 10 major categories are not the same, but, within each category, the population sizes of the 20 subcategories are the same. 5. Poverty incidence at the US$1-a-day level is extremely low in Thailand. We have been unable to measure it with any hope of accuracy using our methods in this study. References Anderson, K., and E. Valenzuela. 2008. "Estimates of Global Distortions to Agricultural Incentives, 1955­2007." Data spreadsheet, October, World Bank, Washington, DC. http://go.worldbank.org/ YAO39F35E0. Thailand 299 Dixon, P. B., B. R. Parmenter, J. Sutton, and D. P. Vincent. 1982. ORANI: A Multisectoral Model of the Australian Economy. Amsterdam: North-Holland. Hertel, T. W., ed. 1997. Global Trade Analysis: Modeling and Applications. New York: Cambridge University Press. Horridge, M. 2004. "ORANI-G: A Generic Single-Country Computable General Equilibrium Model." Centre of Policy Studies, Monash University, Melbourne. Johansen, L. 1964. A Multi-sectoral Study of Economic Growth. Amsterdam: North-Holland. NESDB (Office of the National Economic and Social Development Board). 2004. Input-Output Tables of Thailand, 2000. Bangkok: National Accounts Office, NESDB. NSO (National Statistical Office). 2003. Report of the Household Socio-economic Survey, 2000. Bangkok: Economic Statistics Division, NSO. PovcalNet. World Bank, Washington, DC. http://go.worldbank.org/NT2A1XUWP0 (accessed May 2008). Valenzuela, E., and K. Anderson. 2008. "Alternative Agricultural Price Distortions for CGE Analysis of Developing Countries, 2004 and 1980­84." Research Memorandum 13 (December), Center for Global Trade Analysis, Department of Agricultural Economics, Purdue University, West Lafayette, IN. https://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID=2925. van der Mensbrugghe, D. 2005. "Linkage Technical Reference Document: Version 6.0." December, World Bank, Washington, DC. http://go.worldbank.org/7NP2KK1OH0. Warr, P. 2001. "Welfare Effects of an Export Tax: Thailand's Rice Premium." American Journal of Agricultural Economics 83 (4): 903­20. ------. 2005. "Food Policy and Poverty in Indonesia: A General Equilibrium Analysis." Australian Journal of Agricultural and Resource Economics 49 (4): 429­51. ------. 2007. "Long-Term Economic Performance in Thailand." ASEAN Economic Bulletin 24 (1): 1­26. ------. 2008a. "Trade Policy and the Structure of Incentives in Thai Agriculture." ASEAN Economic Bulletin 25 (3): 249­70. ------. 2008b. "World Food Prices and Poverty Incidence in a Food Exporting Country: A Multi- household General Equilibrium Analysis for Thailand." Agricultural Economics 39 (3): 525­37. Warr, P., and A. Kohpaiboon. 2009. "Thailand." In Distortions to Agricultural Incentives in Asia, ed. K. Anderson and W. Martin, 255­79. Washington, DC: World Bank. PART IV NATIONAL CGE APPROACHES: AFRICA 10 Mozambique Channing Arndt and James Thurlow Mozambique has considerable agricultural potential. Only about 20 percent of its vast tracts of decent quality land and an even smaller share of its water resources are currently being exploited. The country's long coastline contains multiple har- bors, which face eastward toward the dynamic markets of Asia. Regional markets also offer promise in the short and long term. Despite this potential, Mozambique earned the unwanted label of poorest country in the world in the early 1990s. The country's severe poverty and poor economic performance were the result of a number of factors, including the char- acter of Portuguese colonization, a failed socialist experiment, and more than a decade of vicious civil war that lasted until 1992 (Tarp et al. 2002). Since the end of the war, Mozambique has performed much better, and most development indi- cators have shown substantial improvement (Arndt, Jones, and Tarp 2006). Nonetheless, its low starting point underlines the country's need for substantial growth over an extended period if it is to reach even the average conditions among developing countries. So, while much improved over the past two decades, the economic situation in Mozambique remains sobering, particularly in rural areas, where 70 percent of the population resides. About half of the rural population is considered absolutely poor, meaning that these households have difficulty acquiring even the most basic necessities such as sufficient food to meet calorie requirements (Arndt and Simler 2004). Rural dwellers, particularly the poor, depend heavily on crop agriculture for their incomes. Technology is generally rudimentary, and agricultural value added remains concentrated in cassava, cereals (particularly maize), and beans. Only a small minority of rural households report using improved seeds, fertilizers, and 303 304 Agricultural Price Distortions, Inequality, and Poverty pesticides (Uaiene 2008). Rural households tend to consume most of their produc- tion directly; they market only a relatively small share. Overall, approximately three-quarters of the population (rural and urban) depend on agriculture for the majority, often a large majority, of their incomes. Urban households also rely on domestically produced agricultural goods, particularly urban households outside Maputo, the capital city. Given the weight of agriculture in the economic life of most Mozambicans, growth in agriculture is widely regarded as a potentially powerful lever for reducing poverty (Thurlow 2008, Tarp et al. 2002). Agriculture is also a potential driver of exports and economic growth. Mozam- bique already exports a range of traditional crops, such as tobacco, sugarcane, and cotton. These crops contribute to upstream processing in the manufacturing sector, which has grown rapidly over the last decade. Moreover, the country's underutilized natural resources have attracted considerable foreign investment interest.1 However, most options for rapid growth in agriculture depend on foreign demand through exports of primary or processed products. The critical role of international markets in agricultural growth stems from two facts. First, the import penetration in agriculture is relatively limited: imports represent only about 4 percent of the total demand for agricultural products.2 While there is room to displace processed food imports, the overall scope for growth via import substitution is narrow. Second, because of geographical factors, even the relatively small volume of existing imports would be difficult to displace. The large majority of agricultural imports are consumed in the south of the country, particularly in Maputo. However, the agricultural potential in the south is limited. The more favorable growing areas are located 1,000 to 2,000 kilometers away in the northern and central regions of the country. Because of the large dis- tances, inadequate infrastructure, the inefficiencies in storage, and, especially, the high cost of capital, exporting surplus from the north and importing the same prod- ucts in the south is generally more efficient than attempting to transport surpluses across space (north to south) and time (from the postharvest period to the hungry season; for example, see Arndt, Schiller, and Tarp 2001; Cruz 2006). Hence, for agri- culture to grow substantially, an export orientation is crucial. This may be achieved through exports of primary agriculture or processed farm products. The links among international agricultural markets, agricultural growth, and the prospects for overall growth and poverty reduction provide the motivation for our study. Similar to other chapters in this volume, this chapter examines whether distortions to agricultural markets hamper the prospects for reducing poverty and income inequality. Two sources of distortions are considered: the distortions to domestic agricultural markets imposed by the government of Mozambique and the distortions imposed by other countries that influence Mozambique's export prospects and import prices. Mozambique 305 The chapter is structured as follows. The next section provides basic informa- tion on the structure of the Mozambican economy and summarizes the results of a detailed study of domestic agricultural distortions. The subsequent section presents the modeling framework we use for our analysis. The following section discusses our simulations and results. A final section concludes. Economic Structure and Agricultural Distortions in Mozambique The structure of the Mozambican economy is summarized in table 10.1. Agricul- ture represents about 26 percent of total GDP and 19 percent of exports. Exports are dominated by fisheries, which generate two-thirds of raw agricultural export earnings. The low import penetration in agriculture is reflected in the small share of agriculture in total imports and by the small share of imports in the total domestic demand for agricultural goods. Agriculture has strong links to agriculture- related processing in the manufacturing sector. Sugar processing, for instance, relies on domestically sourced sugarcane, as do cotton and tobacco processing and grain milling. Together, these processing sectors represent about 0.6 percent of total GDP and 3 percent of exports. Agriculture's upstream links are therefore an important part of the sector's overall contribution to the economy. The relatively large share of manufacturing in total exports is driven by aluminum smelting, which accounts for around half of all export earnings and 5 percent of total GDP. As documented by Andersson (2001), aluminum smelting is largely an enclave sector characterized by high capital intensity, specific capital requirements, foreign ownership, limited taxation, expatriate labor, and imported intermediates. All aluminum production is exported, and the links between the aluminum sector and the rest of the economy are weak. If aluminum were removed, agriculture would represent 37 percent of exports and 28 percent of GDP. The agricultural sector has undergone progressive liberalization and the grad- ual elimination of government intervention since 1987. In particular, the country shifted from central planning to a market economy. The reform program has sub- stantially reduced government involvement in agriculture. A detailed study of the agricultural distortions in Mozambique finds that, since the period of reform from a centrally planned to a market economy, there has been hardly any substan- tial government intervention in the sector, and few distortions are observed (Alfieri, Arndt, and Cirera 2009). The government intervention that does exist relies primarily on the use of import tariffs. There are a few exceptions, such as the cotton, cashew, and sugar sectors, in which more complex policies have been implemented. Nonetheless, the overall picture is one of distinctly limited government involvement. 306 Table 10.1. The Economic Structure in 2002 and Price Distortions in 2004, Mozambique (percent) GDP Output Imports Exports Sectors share subsidy Share Intensity Tariff Share Intensity Tax rate Total 100.0 0.1 100.0 24.0 9.3 100.0 11.4 0.1 Agriculture 26.1 0.7 2.9 4.8 5.2 19.0 13.0 -- Maize 3.5 0.1 0.3 4.4 10.0 0.2 1.0 -- Sorghum 1.1 0.1 -- -- -- -- -- -- Unshelled rice 0.7 0.2 0.0 0.0 25.9 -- -- -- Wheat -- -- 1.7 100.0 2.5 -- -- -- Cassava 7.2 -- -- -- -- -- -- -- Roots and tubers 0.2 -- 0.0 2.2 10.9 0.0 0.0 -- Beans 1.0 -- 0.0 2.2 10.9 0.0 0.3 -- Groundnuts 0.9 24.6 0.1 2.1 0.8 0.0 0.0 -- Cashews 0.4 -- 0.0 0.2 10.9 1.6 43.6 -- Vegetables 1.7 0.9 0.0 1.5 19.0 0.0 0.0 -- Fruits 1.6 0.9 0.0 1.3 19.0 1.7 19.3 -- Leaf tea 0.0 -- -- -- -- -- -- -- Tobacco 0.3 -- 0.3 40.4 10.9 0.8 47.7 -- Sugarcane 0.1 8.4 -- -- -- -- -- -- Cotton 0.3 3.1 -- -- -- -- -- -- Other crops 0.8 0.8 0.2 6.5 8.2 0.0 0.1 -- Livestock 0.5 1.4 0.0 2.3 -- 0.0 0.4 -- Forestry 2.8 0.7 0.1 0.6 3.0 1.6 10.7 -- Fisheries 2.3 0.7 0.0 0.0 3.0 12.7 63.6 -- Industry 23.2 0.0 76.0 44.6 12.1 67.1 23.6 0.1 Mining 0.3 0.7 0.2 6.6 3.0 0.3 4.9 -- Manufacturing 13.7 0.0 70.3 57.9 12.4 55.5 35.7 0.1 Meat processing 1.5 -- 0.3 3.1 21.9 0.0 0.0 -- Other food products 1.4 -- 4.2 34.4 18.7 1.6 8.5 -- Grain milling 1.6 -- 8.5 45.0 18.7 0.2 0.7 -- Sugar processing 0.1 -- 0.7 49.9 98.9 0.5 15.1 -- Beverages 0.8 -- 1.3 18.6 18.7 -- -- -- Tobacco processing 0.1 -- 0.4 36.4 18.7 -- -- -- Cotton processing 0.4 0.2 2.5 66.2 18.8 2.7 64.5 2.0 Textiles and clothing 0.6 0.2 2.0 40.3 18.8 0.7 10.3 -- Wood products 0.8 0.2 3.1 39.0 18.8 0.5 4.5 -- Chemicals 0.4 -- 20.2 86.1 8.2 0.4 6.4 -- Nonmetals 0.7 -- 2.3 37.8 8.2 0.1 0.7 -- Metal products 5.2 -- 5.3 71.6 8.2 48.6 94.8 -- Machinery 0.0 -- 17.3 99.4 8.2 -- -- -- Other manufacturing 0.0 -- 2.4 96.8 8.2 0.2 46.7 -- Electricity 1.9 -- 5.5 81.0 8.2 11.3 79.9 -- Water 0.3 -- -- -- -- -- -- -- Construction 7.1 -- -- -- -- -- -- -- Services 50.7 -- 21.1 9.2 -- 13.9 3.2 -- Sources: Social accounting matrix for Mozambique in 2002 (Thurlow 2008) adjusted to reflect the developing-country distortions compiled by Anderson and Valenzuela (2008) and harmonized with the Global Trade Analysis Project model and the Linkage model in Valenzuela and Anderson (2008). Note: Import intensity is the share of imports in each sector's total domestic demand. Export intensity is the share of exports in each sector's total domestic output. -- data are not available or the indicator is not applicable. 307 308 Agricultural Price Distortions, Inequality, and Poverty The social accounting matrix we use for our analysis reflects this situation. The only meaningful distortions in the matrix are import tariffs. These are also shown in Table 10.1. To maintain consistency with other studies in this volume, we impose the tariffs in the Linkage model on the Mozambican computable general equilibrium (CGE) model.3 Similar to many other least developed countries, Mozambique views tariffs as a tool for both raising revenue and influencing prices and incentives within the economy (Arndt and Tarp 2008). Consistent with the revenue-raising goal, most rates are positive, but relatively low. The exceptions are the rates imposed on processed sugar (nearly 100 percent), rice (26 percent), and many processed food commodities (around 20 percent). More recently, the sugar tariff, which is implemented as a variable levy, has declined to near zero because of increases in the world price of sugar. The tariffs on rice have also declined. For our study, we retain the applied rates from 2004 in our estimates of the impact of the removal of domestic price distortions. The Modeling Framework The CGE model used in our study contains 56 activities and commodities, including 24 agricultural and 7 food processing sectors.4 We identify five factors of production: three types of labor (unskilled, semiskilled, and skilled), agricultural land, and capi- tal. Factor intensities for each sector are shown in table 10.2. Rural and urban labor markets are segmented, such that rural nonfarm and urban nonagriculture are dis- tinguished. We assume that the factors in the model are fully employed at flexible real wages. The only exceptions are rural and urban unskilled laborers, who are unem- ployed at fixed nominal wages, and capital in the metals and electricity sectors, which is immobile and earning sector-specific returns.5 The former captures the underem- ployment of lower-skilled workers in Mozambique, while the latter reflects a depend- ence on foreign direct investment. Using these factors, producers in the model maximize their profits under technologies exhibiting constant returns to scale, and the choice among factors is governed by a constant elasticity of substitution function. We then combine factors with fixed-share intermediates using a Leontief specifica- tion. Under profit maximization, the factors receive income such that marginal revenue equals marginal cost based on endogenous relative prices. The possibility of substitution exists between production for domestic and production for foreign markets. The decision of producers is governed by a con- stant elasticity of transformation function that distinguishes between exported and domestic goods to capture any time or quality differences between the two types of products. Profit maximization drives producers to sell in the markets in which they are able to achieve the highest returns. These returns are based on domestic and export prices (whereby the latter is determined by the world price, multiplied by the exchange rate and adjusted for any taxes). Under the small Table 10.2. Factor Intensities of Production, Mozambique, 2002 Labor Sector Skilled Semiskilled Unskilled All Capital Land Total Total 10.7 13.8 39.7 64.1 6.1 29.8 100.0 Agriculture 0.6 1.9 59.6 62.1 23.3 14.6 100.0 Maize 0.6 1.9 58.1 60.6 29.7 9.7 100.0 Sorghum 0.6 2.0 61.6 64.2 26.6 9.2 100.0 Unshelled rice 0.5 1.8 54.5 56.8 30.7 12.5 100.0 Wheat -- -- -- -- -- -- -- Cassava 0.5 1.8 56.0 58.3 30.3 11.4 100.0 Roots and tubers 0.4 1.3 39.6 41.3 42.1 16.6 100.0 Beans 0.4 1.4 42.7 44.6 38.7 16.8 100.0 Groundnuts 0.6 2.2 65.1 67.9 20.4 11.7 100.0 Cashews 0.3 1.2 35.8 37.3 42.7 20.0 100.0 Vegetables 0.6 2.0 60.2 62.8 26.8 10.5 100.0 Fruits 0.4 1.2 36.5 38.1 43.8 18.1 100.0 Leaf tea 0.4 1.3 38.5 40.1 41.2 18.6 100.0 Tobacco 0.6 1.9 56.2 58.6 29.7 11.6 100.0 Sugarcane 0.1 0.2 7.1 7.4 73.3 19.3 100.0 Cotton 0.4 1.3 38.6 40.2 41.4 18.4 100.0 Other crops 0.5 1.8 56.6 59.0 29.5 11.5 100.0 Livestock 0.8 2.5 78.0 81.3 18.7 -- 100.0 Forestry 0.6 2.0 63.2 65.8 34.2 -- 100.0 Fisheries 0.8 2.8 87.2 90.9 9.1 -- 100.0 (Table continues on the following page.) 309 310 Table 10.2. Factor Intensities of Production, Mozambique, 2002 (continued) Labor Sector Skilled Semiskilled Unskilled All Capital Land Total Industry 10.7 9.7 20.8 41.2 58.8 -- 100.0 Mining 1.4 2.8 11.6 15.7 84.3 -- 100.0 Manufacturing 10.9 8.3 17.3 36.6 63.4 -- 100.0 Meat processing 18.4 14.1 29.3 61.7 38.3 -- 100.0 Other food products 18.4 14.1 29.3 61.7 38.3 -- 100.0 Grain milling 24.5 18.8 39.1 82.3 17.7 -- 100.0 Sugar processing 15.8 12.1 25.2 53.2 46.8 -- 100.0 Beverages 3.3 2.5 5.3 11.1 88.9 -- 100.0 Tobacco processing 3.3 2.5 5.3 11.1 88.9 -- 100.0 Cotton processing 23.4 17.9 37.3 78.7 21.3 -- 100.0 Textiles and clothing 24.0 18.3 38.1 80.4 19.6 -- 100.0 Wood products 24.4 18.7 38.8 81.8 18.2 -- 100.0 Chemicals 8.9 6.8 14.1 29.8 70.2 -- 100.0 Nonmetals 3.0 2.3 4.8 10.0 90.0 -- 100.0 Metal products 0.2 0.1 0.3 0.6 99.4 -- 100.0 Machinery 8.4 6.4 13.3 28.1 71.9 -- 100.0 Other manufacturing 8.2 6.3 13.0 27.4 72.6 -- 100.0 Electricity 9.0 6.9 14.3 30.2 69.8 -- 100.0 Water 6.5 7.7 17.0 31.3 68.7 -- 100.0 Construction 11.4 13.5 29.9 54.8 45.2 -- 100.0 Services 15.9 21.7 38.0 75.7 24.3 -- 100.0 Source: CGE model for Mozambique by the authors (see Thurlow 2008). Note: -- data are not available. Mozambique 311 country assumption, Mozambique faces perfectly elastic world demand curves at fixed world prices (global liberalization is discussed below). The final ratio of exports to domestic goods is determined by the endogenous interaction of the relative prices for these two commodity types. Additional substitution possibilities exist between imported and domestic goods under a constant elasticity of substitution Armington specification. In this case, substitution may take place in both final and intermediate usages. The elasticities vary across sectors; the lower elasticities reflect greater differences between domestic and imported goods. Under the small country assumption, Mozambique faces an infinitely elastic world supply at fixed world prices. The final ratio of imports to domestic goods is determined by the cost-minimizing decision making of domestic demanders based on the relative prices of imports and domestic goods (both of which include the relevant taxes). In global liberalization, the results from the World Bank's model of global trade (the Linkage model; see van der Mensbrugghe 2005) are transmitted to the Mozambique model via changes in the import prices, export prices, and export quantities faced by Mozambique. We simply apply the import price changes to the exogenous import prices in the Mozambique model. We apply export price and quantity changes derived from the Linkage model in the manner developed by Horridge and Zhai (2006). Specifically, export demand functions of the form Q (FP P)ES (where Q is the quantity exported; P is the export price; ES is the elasticity of demand for exports; and FP is a shift parameter) have been added to the Mozambique model to mimic the global Linkage model. Horridge and Zhai (2006) show that the export price and quantity changes generated by Linkage may be mimicked in a country model through shocks to the shifter parameter FP. Using lowercase to indicate percentage change, one may derive the percentage change in FP applied to the Mozambique model as follows: fp p q ES. The model distinguishes among various institutions, including enterprises, the government, and 10 representative household groups. Households are disaggre- gated across rural and urban areas and national income quintiles. Households and enterprises receive incomes in payment for the use of their factors of production by producers. Households and enterprises pay direct taxes to the government (based on fixed tax rates), save (based on marginal propensities to save), and make transfers to the rest of the world. Enterprises pay their remaining incomes to households in the form of dividends. Households use their incomes to consume commodities under a linear expenditure system of demand. Home consumption is important among rural households. It represents about half of the commodity consumption among all but the top quintile of rural house- holds; among the top quintile, the share of home consumption in total commodity consumption is about one-quarter (table 10.3). Home consumption is driven in large 312 Table 10.3. Household Income and Expenditure Shares, Mozambique, 2002 (percent) Household income sources Labor Households Skilled Semiskilled Unskilled Capital Land Other income Total All households 11.7 15.0 43.4 22.4 6.7 0.8 100.0 Rural 0.8 5.2 68.6 7.7 17.5 0.2 100.0 Quintile 1 0.1 1.9 84.2 2.5 11.3 0.1 100.0 Quintile 2 0.2 2.4 74.5 3.2 19.6 0.2 100.0 Quintile 3 0.4 3.7 65.8 5.2 24.6 0.1 100.0 Quintile 4 0.2 4.3 69.5 5.8 20.0 0.2 100.0 Quintile 5 2.4 10.2 59.9 16.4 10.8 0.2 100.0 Urban 18.4 21.1 27.9 31.4 0.0 1.2 100.0 Quintile 1 0.1 12.4 77.9 9.5 0.0 0.1 100.0 Quintile 2 2.8 18.9 61.6 16.6 0.0 0.1 100.0 Quintile 3 1.4 15.5 69.8 13.0 0.0 0.3 100.0 Quintile 4 5.1 25.0 46.2 23.3 0.0 0.3 100.0 Quintile 5 23.5 21.3 18.0 35.8 0.0 1.5 100.0 Household expenditures Own goods Purchased goods Households Food Nonfood Food Nonfood Taxes Savings Total All households 8.6 6.7 19.0 62.7 1.7 1.3 100.0 Rural 22.6 17.7 12.3 46.5 0.3 0.7 100.0 Quintile 1 36.3 15.6 14.0 34.0 0.0 0.1 100.0 Quintile 2 30.1 16.4 13.3 39.7 0.0 0.5 100.0 Quintile 3 28.2 18.9 12.4 39.8 0.3 0.4 100.0 Quintile 4 20.3 21.5 12.7 45.2 0.1 0.3 100.0 Quintile 5 10.2 14.9 10.4 62.2 0.7 1.6 100.0 Urban 0.0 0.0 23.1 72.7 2.6 1.6 100.0 Quintile 1 0.0 0.0 56.8 42.5 0.3 0.3 100.0 Quintile 2 0.0 0.0 43.1 55.5 0.6 0.9 100.0 Quintile 3 0.0 0.0 44.9 53.1 0.8 1.2 100.0 Quintile 4 0.0 0.0 30.2 66.8 1.7 1.3 100.0 Quintile 5 0.0 0.0 18.1 77.1 3.0 1.8 100.0 Source: CGE model for Mozambique by the authors (see Thurlow 2008). Note: Other income refers to government transfers (for example, pensions) and foreign remittances received. Own goods are goods produced and consumed by a household. 313 314 Agricultural Price Distortions, Inequality, and Poverty measure by substantial divergences between farmgate and consumer prices because of high transaction costs. These margins are captured in the model, wherein transac- tions costs potentially differ among domestic, imported, and exported goods. The modeling of home consumption and margins follows Arndt et al. (2000). The government receives income through imposing activity, sales and direct taxes, and import tariffs and then makes transfers to households, enterprises, and the rest of the world. The government also purchases commodities in the form of government consumption expenditures. The remaining income of government is (dis)saved. All savings by households, enterprises, government, and the rest of the world (foreign savings) are collected in a savings pool from which investment is financed. The model includes three broad macroeconomic accounts: the government bal- ance, the current account, and the savings and investment account. To bring about balance among the various macroaccounts, we must specify a set of macroclosure rules. These provide a mechanism through which macroeconomic balance may be achieved. Consistent with other analyses in this volume, we assume a savings-driven closure to balance the savings and investment account. Under this closure, the mar- ginal propensities to save of households and enterprises are fixed, while investment adjusts to changes in incomes to ensure that the level of investment and savings are equal. For the current account, we assume that a flexible exchange rate adjusts to maintain a fixed level of foreign savings. Thus, the external balance is held fixed in foreign currency terms. Finally, in the government account, we assume that the fiscal deficit remains unchanged and that government revenues and expenditures are balanced through changes in the direct tax rates on households and enterprises. The CGE model is calibrated to a 2003 social accounting matrix (Thurlow 2008) that was constructed using information from national accounts, trade and tax data, and household income and expenditure data in the 2002 national house- hold survey (INE 2004). Trade elasticities are taken from the Global Trade Analysis Project (Dimaranan 2006). The model is calibrated so that the initial equilibrium reproduces the values of the base year in the social accounting matrix. The results of the CGE model are passed back down to the household survey on which the model is based and in which the poverty measures are calcu- lated. More specifically, the changes in the real commodity expenditures of each representative household in the CGE model are applied to the expenditures of the corresponding household in the survey. Total expenditures are compared to real expenditure poverty lines, and standard poverty measures are recalculated. The Simulations and the Results In this section, we present the results of four simulations. The first two simulations assess the impact of the removal of distortions in the rest of the world without any changes in the tariffs and subsidies in Mozambique. We do this for all commodities Mozambique 315 (simulation 1) and then only for agriculture and agriculture-related processing (simulation 2). As with other case studies in this volume, the impacts of global liberalization are taken from the Linkage model (see the appendix by van der Mensbrugghe, Valenzuela, and Anderson). The impacts include changes in world import and export prices, as well as changes in the demand for Mozambican exports (table 10.4). The implications of global liberalization as derived from the Linkage model are imposed as in Horridge and Zhai (2006). These two simulations therefore model the impact on Mozambique if the rest of the world removes agricultural and nonagricultural distortions, while Mozambique does not. The remaining two simulations (3 and 4) assess the impact if Mozambique removes its distortions. These include import tariffs and, to a much lesser extent, export and output taxes and subsidies. In both of these national simulations, there are no changes in the rest-of-the-world distortions. World prices therefore remain unchanged because we retain the small country assumption for Mozambican exports. Simulation 1: The global liberalization of all commodities The results from the Linkage model indicate that Mozambique's terms of trade improve by 1.3 percent once the distortions on all the commodities in the rest of the world are removed (table 10.5). World demand for Mozambique's main exports increases, which is a strong driver in the results that follow. World import prices decline for Mozambique's main imported goods, such as clothing and other heavier manufactures; together, these account for more than half of total imports. Import prices rise for some commodities, such as processed foods, but these are less significant import commodities and show relatively low import penetration ratios (table 10.1). Rising export prices encourage producers to increase their production for for- eign markets, thereby causing an appreciation of the real exchange rate of 2.8 per- cent (table 10.5). Import demand, which is already rising because of falling world prices, increases additionally as a result of the appreciation, while the expansion of exports is only partially offset. The appreciation of the exchange rate also reduces the value of foreign inflows (mostly the value of foreign assistance measured in local currency) and, hence, lowers investment demand. However, the real appreci- ation and cheaper imported goods drive down consumer prices, causing a substantial rise in private consumption. Overall, the increase in exports and consumer spending outweighs any additional import penetration, and there is an increase in total GDP of 0.9 percent. The increase in GDP is driven by agriculture and agriculture-related process- ing (table 10.6). Greater agricultural production occurs mainly because of increased fisheries exports, which already dominate exports (table 10.7) and are 316 Table 10.4. The Exogenous Demand and Price Shocks Caused by Rest-of-the-World Liberalization, Mozambique (percent change from the baseline) Simulation 1: All commodities Simulation 2: Agriculture only Export Export Sector Import price Price Quantity Import price Price Quantity Agriculture Maize 2.6 1.1 3.4 4.1 1.4 19.7 Sorghum 2.6 1.1 3.4 4.1 1.4 19.7 Unshelled rice 8.5 -- -- 9.7 -- -- Wheat 0.9 -- -- 0.8 -- -- Cassava 0.7 1.1 8.7 0.3 1.4 16.0 Roots and tubers 0.7 1.1 8.7 0.3 1.4 16.0 Beans 0.7 1.1 8.7 0.3 1.4 16.0 Groundnuts 1.2 1.0 9.7 0.1 1.3 2.8 Cashews 1.2 1.0 9.7 0.1 1.3 2.8 Vegetables 2.7 0.6 496.8 1.5 0.9 567.8 Fruits 2.7 0.6 496.8 1.5 0.9 567.8 Leaf tea 0.7 1.1 8.7 0.3 1.4 16.0 Tobacco 0.7 1.1 8.7 0.3 1.4 16.0 Sugarcane -- -- -- -- -- -- Cotton -- 0.7 48.4 -- 1.2 56.4 Other crops 0.7 1.1 8.7 0.3 1.4 16.0 Cattle -- -- -- -- -- -- Poultry 1.7 -- -- 0.1 -- -- Other livestock 1.7 -- -- 0.1 -- -- Forestry 0.6 0.8 38.4 0.4 1.1 3.1 Fisheries 0.6 0.8 38.4 0.4 1.1 3.1 Industry Mining 0.6 0.8 38.4 0.4 1.1 3.1 Meat processing 1.3 -- -- 0.1 -- -- Other food products 1.9 0.4 5.8 0.3 0.7 15.7 Grain milling 1.9 0.4 5.8 0.3 0.7 15.7 Sugar processing 1.4 0.4 21.9 0.1 0.9 11.9 Beverages 1.9 0.4 5.8 0.3 0.7 15.7 Tobacco processing 0.6 0.8 38.4 0.4 1.1 3.1 Cotton processing 0.6 0.8 38.4 0.4 1.1 3.1 Textiles and clothing 1.3 0.3 22.4 0.6 0.8 3.2 Wood products 0.8 0.2 7.6 0.2 0.7 5.1 Chemicals 0.8 0.2 7.6 0.2 0.7 5.1 Nonmetals 0.8 0.2 7.6 0.2 0.7 5.1 Metal products 0.8 0.2 7.6 0.2 0.7 5.1 Machinery 0.8 0.2 7.6 0.2 0.7 5.1 Other manufacturing 0.8 0.2 7.6 0.2 0.7 5.1 Electricity 0.3 0.3 1.7 0.3 0.7 1.0 Water 0.3 0.3 1.7 0.3 0.7 1.0 Construction 0.3 0.3 1.7 0.3 0.7 1.0 Services 0.3 0.3 1.7 0.3 0.7 1.0 Source: Results from the World Bank Linkage model (see the appendix by van der Mensbrugghe, Valenzuela, and Anderson). Note: -- data are not available or the indicator is not applicable. 317 318 Agricultural Price Distortions, Inequality, and Poverty Table 10.5. The Macroeconomic Simulation Results of Prospective Liberalization Abroad and Nationally, Mozambique (percent) Change from the base Rest-of-the-world Unilateral liberalization liberalization Indicator Base share Full Agriculture Full Agriculture Real GDP at market prices 73.5 0.9 0.1 1.2 0.7 Consumption 59.1 1.4 0.2 1.0 1.1 Investment 18.8 0.5 0.2 1.6 0.9 Government 10.4 0.0 0.0 0.0 0.0 Exports 14.1 3.8 0.1 9.8 6.4 Agriculture 19.0 19.5 0.1 28.4 19.5 Industry 67.1 0.6 0.2 2.8 1.4 Services 13.9 2.4 1.5 18.0 12.5 Imports 28.8 2.6 0.4 4.8 3.1 Agriculture 2.9 3.5 1.2 0.7 0.1 Industry 76.0 2.6 0.3 7.2 4.9 Services 21.1 2.2 0.4 3.5 3.0 Consumer price index -- 0.5 0.0 2.4 2.3 Real exchange rate -- 2.8 0.1 5.2 4.0 World export prices -- 1.0 0.4 0.0 0.0 World import prices -- 0.3 0.2 0.0 0.0 Terms of trade -- 1.3 0.6 0.0 0.0 Source: Simulations by the authors using the CGE model for Mozambique. Note: The domestic price index is the numéraire in the model. -- data are not available or the indicator is not applicable. generated by price rises and demand increases for Mozambican fish following global liberalization (table 10.4). There is also an expansion in cotton processing because the rice import and export prices and export demand increase. Some other traditional exports also benefit and contribute to the greater upstream pro- duction of processed goods. However, sectoral links work against sugarcane pro- duction because sugar processing declines as a result of falling world prices and greater import competition (tables 10.7 and 10.8). Finally, the overall increase in Mozambique's international trade generates additional demand for the domestic trade and transport sectors, which, together, drive most of the rise in service sector GDP under this simulation. The improved terms of trade and the stimulus to export demand for primary and processed agricultural products resulting from global liberalization boost the demand for unskilled workers (greater employment), as well as the returns to the Mozambique 319 Table 10.6. The Effects of Prospective Liberalization Abroad and Nationally on GDP, by Sector, Mozambique (percent) Change from the base Rest-of-the-world Unilateral liberalization liberalization Sector Base share Full Agriculture Full Agriculture Real GDP at factor cost 100.0 0.9 0.1 0.9 0.3 Agriculture 26.1 2.6 0.1 2.9 1.9 Maize 3.5 1.0 0.0 1.4 1.6 Sorghum 1.1 2.4 0.2 2.3 1.4 Unshelled rice 0.7 0.9 0.4 3.3 3.3 Wheat 0.0 0.0 0.0 0.0 0.0 Cassava 7.2 0.8 0.1 0.2 0.0 Roots and tubers 0.2 0.3 0.2 0.9 0.5 Beans 1.0 0.8 0.1 0.2 0.2 Groundnuts 0.9 0.9 0.0 21.9 22.0 Cashews 0.4 4.3 0.2 10.1 8.3 Vegetables 1.7 1.0 0.2 1.0 0.9 Fruits 1.6 0.6 0.2 0.9 0.9 Leaf tea 0.0 0.7 0.2 6.8 6.8 Tobacco 0.3 3.3 8.9 15.4 9.9 Sugarcane 0.1 0.8 0.4 2.3 2.2 Cotton 0.3 5.3 0.9 2.6 2.2 Other crops 0.8 0.5 0.4 3.5 3.6 Livestock 0.5 0.3 0.3 2.9 2.5 Forestry 2.8 2.8 0.1 0.8 0.4 Fisheries 2.3 20.2 1.0 21.5 13.8 Industry 23.2 0.1 0.1 0.7 1.3 Mining 0.3 0.1 0.0 0.1 0.1 Manufacturing 13.7 0.0 0.1 2.0 1.8 Meat processing 1.5 1.2 0.2 1.2 0.8 Other food products 1.4 0.8 0.8 1.9 2.1 Grain milling 1.6 1.2 0.3 3.5 3.8 Sugar processing 0.1 6.2 1.4 40.1 40.3 Beverages 0.8 0.5 0.2 2.3 1.8 Tobacco processing 0.1 0.6 0.1 1.8 1.6 Cotton processing 0.4 15.5 1.1 18.8 9.0 Textiles and clothing 0.6 3.0 0.6 10.5 11.1 Wood products 0.8 3.9 0.2 11.4 13.0 Chemicals 0.4 7.1 0.4 3.8 7.3 Nonmetals 0.7 3.6 0.2 1.7 2.6 (Table continues on the following page.) 320 Agricultural Price Distortions, Inequality, and Poverty Table 10.6. The Effects of Prospective Liberalization Abroad and Nationally on GDP, by Sector, Mozambique (continued) (percent) Change from the base Rest-of-the-world Unilateral liberalization liberalization Sector Base share Full Agriculture Full Agriculture Metal products 5.2 0.0 0.0 0.0 0.0 Machinery 0.0 5.8 0.5 4.8 6.4 Other manufacturing 0.0 10.7 0.9 64.6 62.3 Electricity 1.9 0.1 0.1 0.3 0.2 Water 0.3 0.5 0.3 4.4 3.2 Construction 7.1 0.5 0.1 1.5 0.8 Services 50.7 0.3 0.0 0.6 0.3 Source: Simulations by the authors using the CGE model for Mozambique. other factors (fully employed workers). Because agricultural production relies on unskilled rural labor and land intensively (the factor intensities are listed in table 10.2), the employment of unskilled rural labor rises by 3.2 percent, while the real returns to land rise by 3.0 percent (table 10.9). The increases in the demand for urban labor and more highly skilled labor are less pronounced, thereby reduc- ing the pressure on factor prices. In addition, skilled workers are often employed by the government, and we assume that recurrent and wage bill expenditures are unaffected by rest-of-the-world liberalization (see the unchanged government spending in table 10.5). Capital returns rise slightly in nominal terms, alongside mild increases in production in the industry and service sectors. All households in the model benefit from full global liberalization. However, the main beneficiaries are lower-income and rural households (measured by the equivalent variation). This is because these households derive a larger share of their incomes from agricultural production and processing, and these sectors are stimulated by the rest-of-the-world liberalization. In contrast, higher-income urban households receive a larger share of their incomes from capital earnings and skilled labor. The improved household welfare is also reflected in changes in the poverty headcount. The share of the population living below the US$1-a-day poverty line falls by 1.4 percent. The decline in poverty is only slightly larger among rural households, and there is no significant change in the inequality measures. Thus, the removal of global distortions on all commodities increases GDP and household welfare and reduces poverty in Mozambique, but has little impact on national inequality. Mozambique 321 Table 10.7. The Effects of Prospective Liberalization Abroad and Nationally on the Real Value of Exports, Mozambique (percent) Change from the base Rest-of-the-world Unilateral liberalization liberalization Sector Base share Full Agriculture Full Agriculture Total exports 100.0 3.8 0.1 9.8 6.4 Agriculture 19.0 19.5 0.1 28.4 19.5 Maize 0.2 0.2 12.2 4.8 3.6 Sorghum 0.0 0.0 0.0 0.0 0.0 Unshelled rice 0.0 0.0 0.0 0.0 0.0 Wheat 0.0 0.0 0.0 0.0 0.0 Cassava 0.0 0.0 0.0 0.0 0.0 Roots and tubers 0.0 1.4 11.4 10.0 9.2 Beans 0.0 0.4 13.8 23.8 20.6 Groundnuts 0.0 402.7 504.8 169.3 162.5 Cashews 1.6 10.6 0.3 24.3 20.6 Vegetables 0.0 8.0 0.2 5.0 4.2 Fruits 1.7 10.2 0.2 13.9 12.0 Leaf tea 0.0 0.0 0.0 0.0 0.0 Tobacco 0.8 2.1 15.6 33.0 23.8 Sugarcane 0.0 0.0 0.0 0.0 0.0 Cotton 0.0 0.0 0.0 0.0 0.0 Other crops 0.0 2.1 0.1 2.0 0.9 Livestock 0.0 2.4 0.0 4.2 3.0 Forestry 1.6 20.4 0.0 17.1 13.5 Fisheries 12.7 29.5 1.3 32.6 21.2 Industry 67.1 0.6 0.2 2.8 1.4 Mining 0.3 15.9 0.2 13.7 11.6 Manufacturing 55.5 0.6 0.4 2.8 1.5 Meat processing 0.0 6.1 0.2 19.3 17.4 Other food products 1.6 6.2 8.0 6.4 5.2 Grain milling 0.2 6.7 8.7 7.3 5.1 Sugar processing 0.5 18.4 5.9 27.7 29.7 Beverages 0.0 0.0 0.0 0.0 0.0 Tobacco processing 0.0 0.0 0.0 0.0 0.0 Cotton processing 2.7 23.7 1.2 35.7 22.1 Textiles and clothing 0.7 4.7 0.3 12.2 9.2 Wood products 0.5 10.1 1.3 8.1 1.6 Chemicals 0.4 10.6 1.2 15.3 18.1 Nonmetals 0.1 9.9 1.4 15.7 14.4 (Table continues on the following page.) 322 Agricultural Price Distortions, Inequality, and Poverty Table 10.7. The Effects of Prospective Liberalization Abroad and Nationally on the Real Value of Exports, Mozambique (continued) (percent) Change from the base Rest-of-the-world Unilateral liberalization liberalization Sector Base share Full Agriculture Full Agriculture Metal products 48.6 0.0 0.1 0.5 0.0 Machinery 0.0 0.0 0.0 0.0 0.0 Other manufacturing 0.2 12.6 1.7 107.3 92.5 Electricity 11.3 0.1 0.3 2.5 0.8 Water 0.0 0.0 0.0 0.0 0.0 Construction 0.0 0.0 0.0 0.0 0.0 Services 13.9 2.4 1.5 18.0 12.5 Source: Simulations by the authors using the CGE model for Mozambique. Simulation 2: The global liberalization of agricultural commodities only In the global agricultural simulation, we model the impact of the removal by the rest of the world of only agricultural distortions, while the distortions in Mozambique are unchanged. Unlike the full global liberalization simulation, Mozambique now experiences a deterioration in the terms of trade of 0.6 percent. However, the demand rises for certain agricultural exports, while the demand declines for all processed commodities, such as processed tobacco, cotton, and sugar. The principal macroeconomic effects of global agricultural liberalization are small and typically opposite in sign relative to the results of the full global simula- tion (compare table 10.5, columns 2 and 3). The changes at the sectoral level are correspondingly small (table 10.4), reflecting the relatively mild price and quan- tity changes that occur in the global markets for products of importance to Mozambique as a result of the rest-of-the-world agricultural trade liberalization (table 10.4). The impacts on factor rewards, welfare, and poverty are also relatively small (table 10.9). Simulation 3: The domestic liberalization of all commodities In the third simulation, we assess the impact of the removal by Mozambique of all its own distortions, while the distortions of the rest of the world remain unchanged. Mozambique's largest distortions revolve around its import tariffs (table 10.1). There is also a small export subsidy for cotton processing, as well as output taxes on Mozambique 323 Table 10.8. The Effects of Prospective Liberalization Abroad and Nationally on the Real Value of Imports, Mozambique (percent) Change from the base Rest-of-the-world Unilateral liberalization liberalization Sector Base share Full Agriculture Full Agriculture Total imports 100.0 2.6 0.4 4.8 3.1 Agriculture 2.9 3.5 1.2 0.7 0.1 Maize 0.3 2.1 4.4 3.4 4.1 Sorghum 0.0 0.0 0.0 0.0 0.0 Unshelled rice 0.0 17.1 36.6 146.1 152.5 Wheat 1.7 0.9 0.4 3.3 3.3 Cassava 0.0 0.0 0.0 0.0 0.0 Roots and tubers 0.0 6.8 0.7 8.1 9.9 Beans 0.0 11.2 0.9 14.4 16.3 Groundnuts 0.1 11.1 2.2 26.9 25.4 Cashews 0.0 10.3 0.4 0.3 1.6 Vegetables 0.0 8.0 0.5 21.3 23.4 Fruits 0.0 7.6 0.4 16.8 18.5 Leaf tea 0.0 0.0 0.0 0.0 0.0 Tobacco 0.3 9.2 3.6 4.0 6.6 Sugarcane 0.0 0.0 0.0 0.0 0.0 Cotton 0.0 0.0 0.0 0.0 0.0 Other crops 0.2 7.3 0.4 1.1 2.0 Livestock 0.0 6.2 0.3 6.7 5.4 Forestry 0.1 4.8 1.2 5.6 3.3 Fisheries 0.0 0.2 0.5 6.6 4.2 Industry 76.0 2.6 0.3 7.2 4.9 Mining 0.2 11.0 1.2 0.3 1.3 Manufacturing 70.3 2.6 0.3 7.8 5.8 Meat processing 0.3 16.1 1.1 59.0 64.1 Other food products 4.2 2.4 0.2 15.1 17.1 Grain milling 8.5 2.1 0.1 14.8 16.4 Sugar processing 0.7 8.4 0.0 137.1 142.0 Beverages 1.3 1.5 0.0 9.0 10.9 Tobacco processing 0.4 2.9 0.6 9.6 11.0 Cotton processing 2.5 3.0 0.7 12.8 14.2 Textiles and clothing 2.0 10.1 1.6 25.7 28.0 Wood products 3.1 6.9 0.7 22.0 23.8 (Table continues on the following page.) 324 Agricultural Price Distortions, Inequality, and Poverty Table 10.8. The Effects of Prospective Liberalization Abroad and Nationally on the Real Value of Imports, Mozambique (continued) (percent) Change from the base Rest-of-the-world Unilateral liberalization liberalization Sector Base share Full Agriculture Full Agriculture Chemicals 20.2 2.2 0.2 1.7 1.8 Nonmetals 2.3 6.9 0.7 6.2 6.0 Metal products 5.3 1.4 0.6 3.9 2.0 Machinery 17.3 1.0 0.2 1.2 1.6 Other manufacturing 2.4 2.3 0.4 1.3 4.4 Electricity 5.5 3.0 0.5 0.7 5.3 Water 0.0 0.0 0.0 0.0 0.0 Construction 0.0 0.0 0.0 0.0 0.0 Services 21.1 2.2 0.4 3.5 3.0 Source: Simulations by the authors using the CGE model for Mozambique. groundnuts and raw sugarcane. Given the relatively few nontariff distortions, the reduction in import tariffs dominates in the results of this simulation. The removal of import tariffs leads to a 4.8 percent increase in import demand. Highly protected sectors experience greater import penetration. Imports more than double in the most highly protected industry, sugar processing (table 10.8). However, highly protected commodities represent only a small share of total imports (see elsewhere above). Thus, the import tariffs on manufactures are low, and it is the increased domestic demand for these manufactured goods that drives the overall increase in imports. Furthermore, rapidly rising import demand places pressure on the current account, thereby inducing a 5.2 percent depreciation in the real exchange rate. The depreciation improves export competitiveness. Recall that the major export, aluminum, is highly capital intensive and tied to existing manufacturing facilities. As a result, aluminum exports effectively do not respond to exchange rate signals, transferring the onus of the export supply response onto primary and processed agriculture, the second-largest exporting sector. Agricultural exports grow by 28 percent. Large increases in exports cause agricultural GDP to rise by 2.9 percent. In contrast, the manufacturing sector experiences a contraction almost across the board, with particularly marked declines in sugar processing; the overall decline is 2 percent. This is, however, insufficient to offset the expansion in agricul- ture, and real GDP rises by 0.9 percent under full domestic liberalization. Mozambique 325 Table 10.9. The Effects of Prospective Liberalization Abroad and Nationally on Employment, Welfare, and Poverty, Mozambique Rest-of-the-world Unilateral liberalization liberalization Indicator Base value Full Agriculture Full Agriculture a Real factor returns Index Change from the base, % Rural skilled labor 1.0 4.5 0.2 6.2 4.2 Rural semiskilled labor 1.0 3.2 0.2 5.2 3.0 Rural unskilled 1.0 0.5 0.0 2.4 2.3 Urban skilled labor 1.0 1.5 0.1 4.5 2.2 Urban semiskilled labor 1.0 1.9 0.2 3.9 2.0 Urban unskilled 1.0 0.5 0.0 2.4 2.3 Capital 1.0 0.6 0.5 5.7 3.2 Agricultural land 1.0 3.0 0.1 5.4 3.8 Unskilled employment Index Rural 1.0 3.2 0.1 3.0 1.5 Urban 1.0 1.0 0.1 1.0 0.3 Equivalent variation, welfare n.a. 1.4 0.2 0.9 1.0 Rural Quintile 1 n.a. 2.8 0.1 3.9 2.9 Quintile 2 n.a. 2.8 0.1 4.3 3.2 Quintile 3 n.a. 2.7 0.1 4.1 3.2 Quintile 4 n.a. 2.8 0.1 4.3 3.2 Quintile 5 n.a. 2.6 0.2 3.1 2.4 Urban Quintile 1 n.a. 0.9 0.2 2.0 1.4 Quintile 2 n.a. 1.0 0.2 2.5 2.0 Quintile 3 n.a. 0.9 0.2 1.8 1.5 Quintile 4 n.a. 0.8 0.3 0.9 1.1 Quintile 5 n.a. 0.3 0.3 1.9 0.9 Poverty headcount ratio % Change from the base, percentage points US$1-a-day poverty line 36.2 1.4 0.0 1.7 1.3 Rural 36.0 1.5 0.0 2.1 1.6 Urban 36.5 1.3 0.0 0.9 0.5 Inequality measures Value Change from the base, percentage points Gini coefficient 0.477 0.1 0.1 1.5 1.2 Theil entropy 0.532 0.3 0.2 3.7 3.2 Source: Simulations by the authors using the CGE model for Mozambique. Note: n.a. not applicable. a. Real factor returns are adjusted to reflect changes in the consumer price index (hence, the change in unskilled wages that are fixed in nominal terms). 326 Agricultural Price Distortions, Inequality, and Poverty Cheaper imported capital goods reduce the cost of investment, while the depreciating exchange rate raises the domestic value of foreign inflows. This causes an increase in investment demand and explains the 1.5 percent rise in con- struction value added. Cheaper imports also lower consumer prices. This would typically boost consumer spending. However, to replace lost tariff revenues, the government must raise personal and corporate tax revenues. Thus, even with the increase in GDP, the share of direct taxes in GDP more than doubles, rising from 2.1 to 5.3 percent. We assume that tax rates adjust proportionally such that most of the additional tax burden falls on enterprises and higher-income urban households. This is real- istic. Obtaining direct tax revenue from rural and urban poor households is effec- tively impossible. The new tax burden causes the welfare outcomes to be negative among urban households in the top quintile (table 10.9), despite the relatively large increases in the real returns to skilled labor and capital. The growth in agri- cultural exports and output drives a 3.0 percent gain in employment in unskilled rural labor and increases the factor prices for rural skilled labor and land by even larger shares. Welfare improves among rural households and all but the richest quintile among urban households. Our assumptions about tax incidence greatly influence the distributional impacts of the removal of domestic distortions. If each household in Mozambique experiences the same percentage point increase in tax rates, then the gains are dis- tributed much more evenly across households. However, if tax rates increase pro- portionally, as we have assumed, then most of the tax burden falls on high-income urban households. The latter assumption represents the only feasible direct tax policy. Sensitivity analysis reveals that other policies, such as increasing activity taxes or sales taxes, also tend to concentrate the tax incidence on urban households because of the significant dependence of rural households on home consumption, which evades taxation. As a result, under all feasible revenue replacement options, rural and lower-income urban households tend to gain more than proportionately from own-country trade liberalization. Moreover, a larger reduction in rural house- hold incomes and poverty relative to urban household incomes and poverty causes national inequality to decline slightly, as reflected in the Gini coefficient. Simulation 4: The domestic liberalization of primary and processed agricultural commodities In our final simulation, we consider the impact of the removal by Mozambique of distortions only in the country's agricultural sector, including agricultural process- ing and textiles. As in simulation 3, we assume that other countries do not alter their Mozambique 327 own distortions. In many ways, the results of this simulation are similar to the results of simulation 3, though they are somewhat more limited (table 10.5). Reduc- ing tariffs causes imports to rise, the real exchange rate to depreciate, and exports to expand. Total GDP increases by 0.7 percent, in part because of an expansion in unskilled employment. However, there is a notable difference in the components of GDP compared with simulation 3. The combination of the less pronounced nomi- nal depreciation (with implications for the pool of foreign savings), the increases in taxes on high-income urban households and enterprises, and the price rises in the construction sector imply that real investment actually declines. This allows the consumption aggregate to grow more relative to the result in simulation 3. Because the largest distortions (the ones generated by import tariffs) occur in agricultural processing (textiles also enjoyed fairly significant protection, at 19 percent), the primary impact of the liberalization is to expand the imports of processed products, which are treated as part of industry in table 10.5. Primary agricultural imports actually decline slightly in the aggregate, despite the reduc- tion in protection. Because agriculture is the major source of exports at the mar- gin, the liberalization of agriculture, processed agriculture, and textiles actually stimulates agricultural production; the strongest production gains are registered in fisheries and groundnuts, which had low initial protection. Commensurate with the consumption aggregate, aggregate household welfare improves slightly more relative to the results under full liberalization. The distrib- utional outcomes are driven by the expansion in agriculture, which generally favors rural labor and land, and the mechanism used to replace lost tariff rev- enues, which affects the highest urban household quintile the most. Because the aggregate stimulus to agriculture is less in this simulation compared with the result in the full liberalization of all markets, the consumption gains are more evenly distributed across factors and, hence, across households. This leads to reductions in poverty that are slightly less pronounced than the ones in simula- tion 3 (table 10.9). There is, however, a similar reduction in national inequality because of the larger income gains and more significant poverty reductions among rural households relative to urban households. Conclusions Agriculture is adversely affected by current global distortions, which are biased against some of Mozambique's key export sectors, such as fisheries. Removing all global distortions would reduce the prices for key imported commodities. It would, however, raise the prices of imported food items, dampening the gains in the terms of trade. Overall, Mozambique's terms of trade would improve. The production responses to a new global environment favor agriculture because of 328 Agricultural Price Distortions, Inequality, and Poverty the stimulation of agricultural exports and import-competing food sectors. Agri- cultural GDP rises once all trade distortions in the rest of the world are elimi- nated, which reduces poverty in Mozambique. In contrast, the removal of only agricultural price distortions abroad has little effect on Mozambique's agricultural sector. The gains in traditional export crops, such as cotton and tobacco, are offset by heightened import competition in processed foods, which has adverse effects on downstream food-crop farmers. Thus, the net impact is a small decline in agricultural GDP and a small increase in national poverty if only agricultural distortions are removed in the rest of the world. Mozambique's own distortions are also biased against agriculture. The produc- ers of processed agricultural products enjoy high levels of protection. The removal of these barriers causes a significant expansion in agricultural GDP despite the associated elimination of tariffs on primary products. The primary agricultural sectors lose little from these tariff reductions because of the relatively low initial protection rates and low import penetration. The rise in agricultural GDP is driven by increased agricultural exports, especially fisheries. Full liberalization nonetheless provides a bigger stimulus to the agricultural sector. Poverty reduc- tion is greater in the full liberalization case because of the relatively greater pro- poor distribution o