china and the wto - isbn: 0821356674 - world bank

C H I N A A N D T H E

WTO

2

T R A D E A N D D E V E L O P M E N T S E R I E S

CH

INA

AN

D T

HE

WT

O

EditorsDeepak Bhattasali • Shantong Li • Will Martin

ACCESSION, POLICY REFORM, AND POVERTY REDUCTION STRATEGIES

T his book fills a large gap in the analysis of China’s WTOaccession. Evidence-based investigations of the likely effect offurther opening of China’s economy, following the steady

globalization that had occurred since 1979, have been largely unavailable.Moreover, the strategic dimensions of the policy changes implemented byChina’s far-seeing leaders have not been addressed adequately until now, andnot as thoroughly as in this book. This collection of first-rate technicalstudies, drawn from good understanding of Chinese and internationaleconomic institutions, and tracing the likely effects of China’s globalizationon different segments of its economy — especially its poor — is essentialreading for students of public policy.”

Long Yongtu, Secretary General, BOAO Forum for Asia

“This book provides the most comprehensive and up-to-date evaluation ofChina’s entry into the WTO. It provides a succinct summary of China’scommitments to the WTO and a detailed analysis of the economicconsequences of the fulfillment of these obligations. A particular strength ofthe volume is its treatment of the relationship between trade liberalizationand the distribution of income in China.”

Nicholas R. Lardy, Senior Fellow, Institute for International Economics

™xHSKIMBy356678zv,:&:<:':(0-8213-5667-4

THE WORLD BANK

3

“

Administrator

29897

China and the WTO Accession, Policy Reform,

and Poverty ReductionStrategies

China andthe WTO

Accession, PolicyReform, and Poverty

ReductionStrategies

Editors

Deepak Bhattasali, Shantong Li,and Will Martin

A copublication of the World Bank and Oxford University Press

© 2004 The International Bank for Reconstruction and Development / The World Bank1818 H Street, NWWashington, DC 20433Telephone 202-473-1000Internet www.worldbank.orgE-mail [email protected]

All rights reserved.A copublication of the World Bank and Oxford University Press.

1 2 3 4 07 06 05 04

The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarilyreflect the views of the Board of 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 judgment on the part of theWorld Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries.

Rights and PermissionsThe material in this work is copyrighted. Copying and/or transmitting portions or all of this work without

permission may be a violation of applicable law. The World Bank encourages dissemination of its work and willnormally grant permission promptly.

For permission to photocopy or reprint any part of this work, please send a request with complete information tothe Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400,fax 978-750-4470, www.copyright.com.

All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422,e-mail [email protected].

Cover collage created by Tomoko Hirata using the following photographs:© Michael S. Yamashita/CORBIS (farmer); © Enzo & Paolo Ragazzini/CORBIS (woman);© Yang Liu/CORBIS (automobile factory); © How-Man Wong/CORBIS (man).

Library of Congress Cataloging-in-Publication Data

China and the WTO: accession, policy reform, and poverty reduction strategies / edited byDeepak Bhattasali, Shantong Li, William J. Martin.

p. cm. – (Trade and development series)Includes bibliographical references and index.ISBN 0-8213-5667-4

1. China—Commercial policy. 2. World Trade Organization—China. 3. Foreign tradeRegulation—China. 4. Agriculture and state—China. 5. Poverty—Government policy—China.6. China—Economic policy—2000-. I. Bhattasali, Deepak. II. Li, Shantong. III. Martin, Will, 1953- IV. Series.

HF1604.C4426 2004382'.3'0951—dc22 2004044097

Contents

Foreword xiAcknowledgments xiiAbout the Editors and Authors xiiiAcronyms and Abbreviations xiv

1. IMPACTS AND POLICY IMPLICATIONS OF WTO ACCESSION FOR CHINA 1Deepak Bhattasali, Shantong Li, and Will Martin

Part I. Policy Reforms Associated with Accession 19

2. WHAT CHINA’S WTO ACCESSION IS ALL ABOUT 21Jeffrey L. Gertler

3. CHINA IN THE WTO: ANTIDUMPING AND SAFEGUARDS 29Patrick A. Messerlin

4. INTELLECTUAL PROPERTY RIGHTS IN THE WTO ACCESSION PACKAGE:ASSESSING CHINA’S REFORMS 49Keith E. Maskus

5. CHINA’S AGRICULTURAL DEVELOPMENT AND POLICY READJUSTMENT AFTERITS WTO ACCESSION 69Chen Xiwen

6. THE NATURE OF DISTORTIONS TO AGRICULTURAL INCENTIVES IN CHINAAND IMPLICATIONS OF WTO ACCESSION 81Jikun Huang, Scott Rozelle, and Min Chang

Part II. Economic Impacts of Accession 99

7. THE IMPACTS OF WTO ACCESSION ON CHINESE AGRICULTUREAND RURAL POVERTY 101Kym Anderson, Jikun Huang, and Elena Ianchovichina

8. THE SERVICES DIMENSION OF CHINA’S ACCESSION TO THE WTO 117Aaditya Mattoo

9. LOGISTICS IN CHINA: IMPLICATIONS OF ACCESSION TO THE WTO 141Wenping Luo and Christopher Findlay

v

10. TELECOMMUNICATIONS SERVICES IN CHINA: FACING THE CHALLENGESOF WTO ACCESSION 157Mari Pangestu and Debbie Mrongowius

11. ACCELERATING FINANCIAL MARKET RESTRUCTURING IN CHINA 181Deepak Bhattasali

12. WTO ACCESSION AND THE STRUCTURE OF CHINA’S MOTOR VEHICLE SECTOR 191Joseph F. Francois and Dean Spinanger

13. ECONOMIC IMPACTS OF CHINA’S ACCESSION TO THE WTO 211Elena Ianchovichina and Will Martin

Part III. Impacts on Households and on Poverty 237

14. EARNINGS AND LABOR MOBILITY IN RURAL CHINA: IMPLICATIONSFOR CHINA’S ACCESSION TO THE WTO 239Terry Sicular and Yaohui Zhao

15. WELFARE IMPACTS OF CHINA’S ACCESSION TO THE WTO 261Shaohua Chen and Martin Ravallion

16. IMPLICATIONS OF WTO ACCESSION FOR POVERTY IN CHINA 283Thomas W. Hertel, Fan Zhai, and Zhi Wang

17. COPING WITH AND ADAPTING TO JOB LOSSES AND DECLINES IN FARM EARNINGS IN CHINA 305Athar Hussain

Index 325

Boxes

9.1 Local Policies and Their Implications for Logistics Operations 14810.1 Summary of China’s WTO Commitments 16610.2 Coverage and Definition of China’s Telecommunications Sector 16710.3 Experience of Chile, Peru, South Africa, and Hong Kong (China) with Universal Access 17511.1 Modeling State Bank Performance: Can the State Banks

Survive Liberalization? 18812.1 Summary of Developments in China’s Automotive Sector 19317.1 Advantages and Disadvantages of Decentralization 312

Figures

8.1 Country Services Commitments by Mode 1198.2 Sequence of Telecommunications Reform in Nine Asian Countries, 1989–99 1359.1 Choice of Transport Modes: China, 1970–2000 143

vi Contents

10.1 Mobile and Fixed-Line Use in China, 1996–2003 16310.2 Internet Users in China, 1996–2002 16410.3 Restructuring of China’s Telecommunications Industry, 1994–2003 17010.4 Share of Revenue by Operator, 2001 and 2002 17212.1 China’s Production of Motor Vehicles since Adoption of the Open-Door Policy 19412.2 Scale and Average Cost (with a CDR of 0.125) 20112.3 Motor Vehicle Production in China and United States 20112.4 Import Share of Total Auto Parts 20312.5 Location of Automobile Production in China 20815.1 Poverty Incidence Curves, Rural 26815.2 Poverty Incidence Curves, Urban 26915.3 Mean Gains by Provinces: Absolute Gain in Yuan Per Capita 27015.4 Mean Gains by Provinces: Proportionate Gains in Percent 27015.5 Mean Gains by Provinces: Percentage of Gainers by Provinces 27115.6 Mean Gains in Yuan by Income Percentile 27115.7 Mean Percentile Gain by Income Percentile 27215.8 Percentage of Gainers by Income Percentile 27216.1 Change in Sector Output due to WTO Accession, China 29816.2 Impact of WTO Accession on Rural Households 30016.3 Impact of WTO Accession on Urban Households 300

Tables

1.1 Some Measures of Import Protection in China’s Agricultural Sector 63.1 Top 10 Antidumping Users, Measures in Force, 1995–2001 303.2 Major Antidumping Targets, Measures in Force, 1995–2001 313.3 Shares of Antidumping Measures in Force against Imports from China, 1995–2002 323.4 Antidumping Measures in Force in the World by Sector, 1995–2000 343.5 EC and U.S. Echoing Antidumping Cases against Chinese Producers, 1980–99 373.6 China’s Antidumping Measures in Force and Investigations, as of 30 June 2003 403.7 Do Antidumping Investigations Really Look at Dumping? 424.1 Substantive Requirements of the TRIPS Agreement in the WTO 514.2 An Assessment of China’s IPRs System for Development Purposes 584.3 Patent Applications by Type and Nationality, 1994–2000 624.4 Patent Grants by Type and Nationality, 1994–2000 634.5 Bilateral Invention Patent Applications, 1994–98 634.6 Bilateral Invention Patent Grants, 1994–98 644.7 Patenting Indicators for Top Patenting Regions, 1985–96 and 2000 645.1 Changes in Farm Output Value Shares 705.2 Sources of Farmers’ Incomes, 1997–2000 715.3 Declines in Per Capita Farm Incomes in Six Key Provinces 725.4 Changes in Farm Household Spending, 1997–2000 725.5 Share of Farmers in Consumer Spending at County and Lower Levels 735.6 Changes in Urban and Rural Resident Income and Spending, 1996–2001 756.1 Changes in Nominal Rates of Protection Over Time of China’s Major

Agricultural Commodities, 1978–2000 836.2 China’s Agricultural Trade (US$ millions) by Factor Intensity, 1985–97 836.3 Subsidies and Tax Rebates for Exports of Selected Agricultural

Commodities in China, 2001 85

Contents vii

6.4 Disaggregated Nominal Rates of Protection for Selected Grains in China,October 2001 87

6.5 Average Nominal Protection Rates for Major Imports and Exports in China,October 2001 88

6.6 Percentage of Market Pairs that Test Positive for Being Integrated Based on Dickey-Fuller Test in Rural China, 1988 to 2000 89

6.7 Soybean, Corn, and Wheat Village Price Regression, 2000 936.8 Transmission Coefficients for Rice and Maize Measuring the Percentage of

Price Shock at Port that Is Transmitted to Inland Markets in China, 1996–2000 937.1 Growth Rates of China’s Economy, 1970–2000 1037.2 Farm and Rural Enterprise (RE) Developments in China, 1980–99 1037.3 Changing Structure of China’s Economy, 1970–2000 1047.4 China’s Grain Production, Trade, and Consumption, 1980–2000 1047.5 Structure of China’s Food and Feed Trade, 1980–99 1057.6 Nominal Rates of Protection (NRP, Applied Tariff or Tariff Equivalent), Agricultural

Products: China, 1995–2007 1067.7 Rural Income, Employment, and Food Output by Region: China, 2000 1087.8 Income and Its Distribution: Rural and Urban China, 1980–99 1087.9 Sectoral Volume Effects of China’s WTO Accession Reforms (Core Case), 2002–07 1107.10 Change in China’s Real Factor Prices and National Economic Welfare

Attributable to WTO Accession, 2001–07 1117.11 Changes in China’s Farm Output, Employment, and Trade Volumes: Without WTO

Accession, 1995–2007, and Resulting from China’s WTO Accession Reforms,2002–07 113

7.12 Effect on Agricultural Self-Sufficiency of China’s WTO Accession without and with Cutsin Agricultural Protection in OECD Countries 114

8.1 Coverage of Specific Commitments by China to Services in Trade 1208.2 China’s Commitments: Past, Present, and Future 1229.1 Starting Position of China’s Logistics Sector 1459.2 Position of China’s Logistics Sector One Year after WTO Accession 1469.3 Position of China’s Logistics Sector Three Years after WTO Accession 1469.4 Summary of Indirect Effects of WTO Accession on China’s Logistics Sector 1479.5 Logistics Costs as a Share of Total Revenue, Europe and Asia, 2001 149

10.1 Network Operators and Licensing, China 15910.2 Network Operators in China: Description of Network and Coverage 16010.3 Impacts of Reforms on China’s Telecommunications Industry, 1998–2003 16511.1 China’s Financial Services Industry, 2001 18311.2 China’s WTO Commitments in the Financial Services Sector 18612.1 GNP, Population, and Stocks of Cars: Selected Countries, 2000 19212.2 Auto Production by Plant in China, 1995–2002 19512.3 Import and Domestic Shipment Values of China’s Motor Vehicle Industry, 1997 19612.4 Tariffs on Chinese Motor Vehicles 19612.5 Regional and Sectoral Breakdown of the Model 19812.6 Mainland China’s Pre- and Post-WTO Accession Tariff Rates 19912.7 Impact of WTO Accession of Greater China on Output 20212.8 Mainland China Motor Vehicle Market 20312.9 Impact of Greater China Accession on Value Added 20412.10 Impact of Greater China Accession on Value of Exports 20512.11 Impact of Greater China Accession on Value of Exports 206

viii Contents

12.12 China Export Shares: Baseline and Scenario 20713.1 Import Coverage of Nontariff Barriers 21413.2 China’s Average Statutory Rates 21513.3 China’s Import Protection before and after WTO Accession 21613.4 Projected Growth in Factor Inputs and Total Factor Productivity 1997–2007 21813.5 Changes in China’s Key Economic Indicators after 2001 as a Result of WTO Accession 22013.6 Change in Real Factor Prices in China as a Result of Accession, 2001–07 22113.7 Welfare Change and Sources of Welfare Change as a Result of China’s WTO Accession 22213.8 Change in Output and Employment in China as a Result of WTO Accession and

Other Reforms 22513.A.1 Tariff Lines Subject to Import NTBs, China, 2001 22713.A.2 The Import Coverage of Nontariff Barriers in China, 2001 22813.A.3 Commodity Import Shares by NTB Measure 22913.A.4 Elasticity of Substitution and Changes in Real Consumer Prices due to China’s

WTO Accession 23013.A.5 Changes in China’s Key Economic Indicators in the Baseline, without WTO

Accession, 1995–2007 23113.A.6 Changes in China’s Key Economic Indicators due to WTO Accession for the Period

before 2001 23213.A.7 Sensitivity Analysis with Respect to the Elasticity of Transformation � (2.67) 23314.1 Descriptive Statistics, CHNS, 1997 24414.2 Income Generation Functions 24514.3 Descriptive Statistics for Variables Used in the Income Generation Functions,

CHNS, 1997 24614.4 Decomposition of Household Net Earnings Differentials between Households

Belonging to the Top and Bottom 20 Percent 24814.5 Agricultural Production Function 25114.6 Nonagricultural Production Function 25214.7 Shadow Wage Equations for Household Agricultural and Nonagricultural Production 25314.8 Employed Wage Equation 25414.9 Summary Statistics for Wage Rates and Shadow Wages 25514.10 Household Labor Supply Equations 25715.1 Predicted Price Changes from GTAP Model and Per Capita Net Gain or Loss

for Rural and Urban Households, 1995–2001 26615.2 Predicted Price Changes from GTAP Model and Per Capita Net Gain or Loss

for Rural and Urban Households, 2001–07 26715.3 Summary Statistics on Aggregate Welfare Impacts, 1995–2001 and 2001–07 26815.4 Regression Results for Level of Gain (Yuan) in Rural Areas of Three Provinces, 2001–07 27415.5 Regression Results for Percentage Gains in Rural Areas of Three Provinces, 2001–07 27515.6 Regression Results for Level of Gain (Yuan) in Urban Areas of Three Provinces, 2001–07 27615.7 Regression Results for Percentage Gains in Urban Areas of Three Provinces, 2001–07 27715.8 Average Impacts for Agriculture Households in Selected Countries, 2001–07 27815.A.1 Rural Gains and Losses by Province, 2001–07 28015.A.2 Urban Gains and Losses by Province, 2001–07 28116.1 Per Capita Income by Location, Stratum, and Vingtile 28516.2 Economic Structure and Market Openness in China, 1997 28716.3 Summary of Baseline Calibration 29316.4 Summary of Experiments 29416.5 Implications of China’s WTO Accession, 2007 296

Contents ix

x Contents

16.6 Sector Impacts of China’s WTO Accession in 2007: Percentage Deviation from Baseline 29716.7 Household Impacts of China’s WTO Accession, 2007 29917.1 Classification of Schemes and Measures 30817.2 Differences between Social Insurance and Social Assistance 31017.3 Income Maintenance Schemes in China 311

China’s era of reform has seen enormous progressin reducing poverty at the same time as increasingintegration with the world economy. The process ofaccession to the WTO contributed to a decisivedeepening of the process of integration with theworld and a move to a much stronger emphasis onrules-based reform. However, in recent years, therehas been increasing concern about an increase ininequality within China and a rising gap betweenthe fortunes of urban and rural families.

The aim of the project on which this book isbased was to bring together information on theimplications of China’s WTO accession agreement—an enormous and complex watershed agreement—and to assess the overall impacts for householdsthroughout China. This required the collection ofinformation and the analysis of the policy reformsinvolved in the accession agreement, followed by anassessment of the effects of these policies on themajor elements of the economy. Finally, drawing onthis analysis, the impact of the reforms on individualhouseholds—and especially those of the poor—wasevaluated. The results of the analysis provide a basisfor recommending policies to capitalize on the bene-fits resulting from accession and to mitigate theadverse impacts.

The work on which this book is based has beendisseminated in China through numerous presen-tations and seminars and through inputs into a

number of training courses undertaken by theWorld Bank Institute and partner organizations indifferent parts of China. This book will provide alasting reference to this major step in the evolutionof the Chinese economy.

This book is the result of a number of years ofwork by a team of Chinese and foreign scholars.Just as China and her trading partners have bene-fited from the win-win nature of internationaltrade, members of this team have benefited fromone another’s knowledge and understanding ofdifferent facets of this enormously complex anddynamic economy. The work is testimony to thesuccess of the partnership between the World Bankand the Development Research Centre of StateCouncil—a partnership that has contributed agreat deal to both of our organizations. We lookforward to even stronger cooperation between ourtwo organizations in the future, as China enters thenext phase of development.

François BourguignonSenior Vice President and Chief EconomistThe World Bank

Wang MengkuiDirectorDevelopment Research Centre of State Council

Foreword

xi

xii

We owe a debt of gratitude to the United KingdomDepartment for International Development (DFID)for its generous support to this project. Particularthanks are due to the DFID staff who provided sup-port and encouragement at critical times in the lifeof the project—in particular, Paul Whittingham forhis encouragement in the original design; CatherineMartin for her continuing advice and encourage-ment from Beijing; and Fiona McConnon, ClareRoberts, and Kebur Azbaha for their assistance withthe implementation of the project.

In addition to our authors—from the Develop-ment Research Center (DRC) of the State Councilof China, the World Trade Organization, the WorldBank, and other institutions around the world—wewould like to thank numerous staff within theDRC, especially Sun Zhiyan and Liu Peilin, whoassisted this project at every step, including the finaldissemination workshop in Wuhan.

At the World Bank, we would like to thankJianqing Chen, Leona Luo, and Rebecca Martin forthe superb support and assistance they have pro-vided throughout the life of the project. The assis-tance of the Beijing Office was invaluable also in

organizing the many workshops and interim publi-cations associated with this project.

Many others have assisted by commenting ondrafts of this book, through participation in theworkshops, or by providing advice. Among others,we would like to thank Long Yongtu, FrançoisBourguignon, Christopher Hum, T. N. Srinivasan,Kong-Yam Tan, L. Alan Winters, Justin Yifu Lin,Wang Xinkui, Nicholas Lardy, Ke Binsheng, WangJun, Bernard Hoekman, Liu Xiaofan, Xu Lin, WangYan, Lu Feng, Yu Jianhua, Sarah Cook, YangGuohua, Yu An, Sheng Laiyun, Wang Pingping,Wang Yongjuan, Huang Langhui, Li Yan, ChenXiaodong, Zhang Xiangchen, Wang Kan, NicholasBridge, Gillian Cull, Yukon Huang, Charles Piggott,Jeffrey Liang, Li Shi, Edwin A.L.M. Vermulst, andJianguang Shen for this help.

We are very grateful to Mary Fisk for her excel-lent editorial assistance and to Santiago Pombo-Bejarano for his enthusiastic support in convertingthe manuscript into a finished volume. We wouldalso like to thank the anonymous reviewers, whoprovided excellent comments on the manuscript;may their tribe increase!

Deepak Bhattasali, Shantong Li, Will Martin

Acknowledgments

xiii

EditorsDeepak Bhattasali, Lead Economist, Poverty

Reduction and Economic Management SectorDevelopment, World Bank, Beijing

Shantong Li, Director, Department of Develop-ment Strategy and Regional Economy, Develop-ment Research Center of the State Council, P.R.China, Beijing

Will Martin, Lead Economist, DevelopmentResearch Group, World Bank, Washington, D.C.

AuthorsKym Anderson, Centre for Economic Studies, Uni-

versity of Adelaide, Adelaide, AustraliaMin Chang, Graduate Student, Department of

Agricultural and Resource Economics, Univer-sity of California, Davis

Shaohua Chen, Senior Information Officer,Development Research Group, World Bank,Washington, D.C.

Christopher Findlay, Professor, Asia-Pacific Schoolof Economics and Government, The AustralianNational University, Canberra

Joseph F. Francois, Professor, Tinbergen Institute,Rotterdam, and Fellow, Centre for EconomicPolicy Research, London

Jeffrey L. Gertler, Legal Affairs Division and for-mer Secretary, Working Party on the Accessionof China, World Trade Organization, Geneva

Thomas W. Hertel, Distinguished Professor,Department of Agricultural Economics, andExecutive Director, Center for Global TradeAnalysis, Purdue University, West Lafayette,Indiana

Jikun Huang, Director and Professor, Center forChinese Agricultural Policy Institute forGeographical Sciences and Natural ResourceResearch, Chinese Academy of Sciences, Beijing

Athar Hussain, Department of Economics, LondonSchool of Economics, London

Elena Ianchovichina, Economist, Poverty Reduc-tion and Economic Management, World Bank,Washington, D.C.

Wenping Luo, Professor, Shanghai Maritime Uni-versity, Shanghai

Keith E. Maskus, Professor, Department of Eco-nomics, University of Colorado, Boulder

Aaditya Mattoo, Lead Economist, DevelopmentResearch Group, World Bank, Washington,D.C.

Patrick A. Messerlin, Professor of Economics, Insti-tutes d’Etudes Politiques, Groupe d’EconomieMondiale de Sciences Po, Paris

Debbie Mrongowius, Independent Researcher,Shanghai

Mari Pangestu, Director, Centre for Strategic andInternational Studies, Jakarta

Martin Ravallion, Research Manager, Develop-ment Research Group, World Bank, Washington,D.C.

Scott Rozelle, Professor, Department of Agricul-tural and Resource Economics, University ofCalifornia, Davis

Terry Sicular, Associate Professor of Economics,University of Western Ontario, London, Ontario

Dean Spinanger, Professor, Institute for WorldEconomics, Kiel University, Kiel

Zhi Wang, Economic Research Service, UnitedStates Department of Agriculture, Washington,D.C.

Chen Xiwen, Vice President (Vice Minister), Devel-opment Research Center of the State Council,Beijing

Fan Zhai, Department of Policy and Fiscal Affairs,Ministry of Finance, Beijing

Yaohui Zhao, Professor of Economics, Beijing Uni-versity, Beijing

About the Editors and authors

xiv

ADSL asymmetric digital subscriber lineAQSIQ Administration for Quality

Supervision, Inspection andQuarantine

CAAC Civil Aviation Administration ofChina

CGA Customs General AdministrationCGE Computable general equilibrium CHNS China Health and Nutrition Survey CIRC China Insurance Regulatory

Commission CJV contractual joint venture CPI consumer price index CSRC Chinese Security Regulatory

CommissionDDA Doha Development Agenda EC European CommunityEJV equity joint venture FDI foreign direct investment GATS General Agreement on Trade in

Services GATT General Agreement on Tariffs and

TradeGDP gross domestic product GSM global system for mobile

communicationsGTAP Global Trade Analysis ProjectHTS Harmonized Tariff System IP Internet protocol IPR intellectual property rightsISDN integrated services digital network MEI Ministry of Electronics Industry MFA Multi-Fiber Arrangement MFN most favored nation MII Ministry of Information Industry MOFTEC Ministry of Foreign Trade and

Economic CooperationMPT Ministry of Posts and

Telecommunications

NME non-market economy NRP nominal rate of protectionOECD Organisation for Economic

Co-operation and Development PBC People’s Bank of China PCT Patent Cooperation Treaty PTA provincial telecommunications

authorityPTNR permanent normal trade relationsSARFT State Administration of Radio, Film

and TelevisionSCIO State Council Information Office SDPC State Development Planning

Commission SETC State Economic and Trade

Commission SEZ special economic zone SILG State Information Leading Group TCSC Tariff Commission under the State

Council TDSCDMA time division synchronous code

division multiple access3PL third-party logisticsTPS transitional product-specific

safeguard TRIPS Trade-Related Aspects of

Intellectual Property Rights TRQ tariff-rate quota TVE township and village enterprises URAA Uruguay Round Agreement on

AgricultureVAT value-added tax VoIP Voice over Internet Protocol WCDMA wideband code division multiple

access WIPO World Intellectual Property

Organization WTO World Trade Organization

Acronyms andAbbreviations

There is every indication that China’s policy-makers see China’s WTO accession agreement asa means to fulfilling broader goals. One broaderstrategic goal is to facilitate the peaceful emergenceof China as a great trading nation—and to avoidthe trade tensions associated with the emergence ofmajor new traders in the past. Another goal is toaccelerate the process of domestic reform—withthe WTO acting, in the colorful words of Jin(2002), as a wrecking ball for what remains of theearlier closed economy (Jin 2002). The process ofaccession has also unleashed a massive process ofreforming the laws covering trade issues (WTO2002), with the result that many reforms go beyondthose narrowly required by the WTO accessionagreement. In addition, China’s policymakers havealso clearly recognized the need to deal with theyawning gap between urban and rural incomes—agap that has become more evident since China’saccession to the WTO (Wen 2004).

The importance of China’s accession to theWTO has been widely recognized and studied. Boththe China Quarterly (see Fewsmith 2001) andChina Economic Review (see Chun, Fleisher, andParker 2001) have published special issues on thetopic, and it has been the subject of books, includ-ing those by Panitchpakdi and Clifford (2002) andLardy (2002). Other investigators have undertaken

After 15 arduous years as a candidate, China’sapplication for membership in the World TradeOrganization (WTO) was accepted at the WTO’sDoha ministerial meeting in November 2001. Asexpected by observers such as Panitchpakdi andClifford (2002), China has become an active andprominent member of the WTO, and has sharplyincreased its share of world trade since joining.Domestically, however, China faces many chal-lenges, particularly in dealing with the widening gapbetween urban and rural incomes. The objective ofthis study is to analyze the impacts of China’s acces-sion to the WTO and the policy reforms needed tohelp facilitate China’s economic development.

Of the many approaches that might be used toanalyze China’s accession to the WTO, an importantone deals with the legal rights and responsibilities—that is, the challenges involved in meeting China’slegal commitments and in ensuring that China’srights are maintained. Another focuses on the tradeand policy changes required to open China’s econ-omy and integrate it into the global economy. Yeteven though these are important, they are not acomplete characterization of the implications ofaccession. For us, perhaps the most compellingaspect is one identified by Woo (2001): WTO acces-sion as a key component in the restructuring of theChinese economy.

1

1

impacts and policyimplications of wtoaccession for china

Deepak Bhattasali, Shantong Li, and Will Martin

quantitative studies of the effects of trade liberal-ization on China (McKibbin and Tang 2000;Ianchovichina and Martin 2001; Gilbert and Wahl2002). Some studies, including those by Martin andIanchovichina (2001) and Shafaeddin (2004) havelooked into the implications of China’s WTOaccession for other countries and for the WTOsystem itself. The studies presented in this volumeattempt to go beyond previous studies in examin-ing, within an integrated framework, China’s policyreforms and their implications for the Chineseeconomy and for individual households.

This volume is divided into three parts. The firstpart contains studies that cover the most importantpolicy reforms associated with accession to theWTO. The second part consists of studies thatexamine the consequences of reform for the econ-omy and for particular sectors. The studies in thethird part build on the earlier studies and examinethe implications of reform for households, andparticularly for those households vulnerable topoverty. Only with the insights available from thesethree approaches is it possible to provide soundpolicy recommendations. In the remainder of thischapter, we review and summarize some of the keyfindings emerging from the entire set of studies.

China’s Policies and the ReformsAssociated with Accession

In chapter 2 of this volume, Jeffrey L. Gertler de-scribes the long process that culminated in China’srecent accession to the WTO. As in other acces-sions, the process involved bilateral negotiations onissues of market access and a multilateral process inwhich the accession agreement was formulated andthe best market access given to any one memberwas extended to all other members under the most-favored-nation (MFN) rule.

The five basic principles of the General Agree-ment on Tariffs and Trade (GATT) and the WTOoutlined by Gertler are useful in organizing ourassessment of the implications of China’s accession.The five principles are: (1) nondiscrimination,(2) market opening, (3) transparency and pre-dictability, (4) undistorted trade, and (5) prefer-ential treatment for developing countries. Theagreement on Trade-Related Aspects of IntellectualProperty Rights (TRIPS) incorporates some ofthese principles as well as others, such as striking a

balance between creating incentives for innovationand restricting competition by giving monopolypower to the inventor. The TRIPS agreement is dis-cussed at the end of this section.

The general principle of nondiscriminationrequires that WTO members give equal treatmentto competing suppliers (the most-favored-nationprinciple), and that they not discriminate betweendomestically produced and imported goods orservices in their internal markets. The applicationof this general principle to China involved someadditional commitments, including the eliminationof dual pricing systems, the phasing out of restric-tions on trading, and the introduction of more uni-form administrative arrangements and of judicialreview. These commitments are important not justfor the central authorities, but also for the lowertiers of government that deal with internal tradeand regulation.

In China’s accession, the market opening principlehad three major dimensions: (1) commitments byChina to abolish nontariff barriers, reduce tariffs,and open service sectors; (2) commitments by theimporting countries to abolish the quotas on textilesand clothing originally imposed under the Multi-Fiber Arrangement (MFA); and (3) agreement bythe United States and other countries to imposeMFN tariffs on China. Analysis in chapter 13 byElena Ianchovichina and Will Martin suggests thatthe tariff cuts offered by China were very substantial,and will result in a reduction in the weighted averagetariff from 12 percent in 2001 to 6.8 percent at theend of the implementation period.1 However, thesereductions are quite small relative to the reductionof 29 percentage points between 1992 and 2001. Theanalysis reported by Aaditya Mattoo in chapter 8suggests that China’s commitments to liberalizetrade in services are extremely large relative to thoseof almost all other countries. Indeed, Mattoo goes sofar as to say that China’s accession commitments onservices constitute the most radical reform of tradein services ever undertaken under the WTO. Henotes, however, that China’s commitments weremore frequently subject to qualifications or reserva-tions than those of other countries. Details of theliberalization in specific sectors are considered inthe next section, as a prelude to assessing the likelyimpact of accession on the Chinese economy.

Although most of the market opening agreed inChina’s accession, like in other WTO accession

2 China and the WTO

agreements, is to be carried out by the accedingcountry, China’s trading partners made an impor-tant “concession” in agreeing to abolish the quotason textiles and clothing originally imposed underthe MFA. This agreement appears likely to beextremely important for China given its strongcomparative advantage in textiles and clothing,although it is qualified by the ability of the im-porters to impose special textile and clothing safe-guards for one year at a time during a transitionperiod up to 2008. Another important “concession”by almost all existing members of the WTO was torefrain from invoking nonapplication provisions of the type widely invoked against Japan when itjoined the GATT. As a result, China has receivedmost-favored-nation treatment in virtually all mar-kets. Clearly, this status is important because it freesChina from onerous, one-sided review procedures,2

such as the former annual review of MFN in theUnited States, and increases the confidence ofinvestors in China’s export industries that foreignmarkets will be available to them on a continuingbasis.

The transparency and predictability of trade pol-icy are enhanced through both the general WTOpolicy rules, such as the need to publish trade rulesand regulations, and some of the specific commit-ments made by China, such as those on uniformapplication of the trade regime and independentjudicial review, as well as the provision of a mecha-nism through which concerned parties can bringproblems of local protectionism to the attention ofthe central government.3 Another important con-tributing factor is China’s binding of its entire tariffschedule for goods, almost always at tariff levelsbelow prior applied rates. The transitional reviewheld each year for eight years after China’s accessionprovides additional information about China’strade regime and its reforms during that period.

China’s commitments to phase out restrictionson trading rights for all products except a short listof commodities that may remain subject to statetrading are important for increasing transparency.Commitments to allow entry into distribution andwholesale services provide further transparency asa side effect of liberalization of these importantservice activities.

The WTO principle of undistorted trade involvesdisciplines on issues such as subsidies and counter-vailing measures, antidumping, and safeguards.

China has made more stringent commitments,including one to refrain from using export subsi-dies on agricultural goods.

The issue of antidumping and safeguard meas-ures is troubling, particularly because it concernsChina’s market access opportunities. As Patrick A.Messerlin notes in chapter 3, the WTO rules onantidumping are biased toward finding dumping,even where no economically meaningful dump-ing exists. The situation is worse for China thanfor other WTO members, because 70 percent ofChina’s exports are products most vulnerable toantidumping measures. Furthermore, China willpotentially remain vulnerable, for up to 15 years, tohighly discriminatory non–market economy provi-sions. These provisions dramatically increase theprobability that dumping will be found. And whenantidumping duties are applied, they are generallymuch higher than when market economy provi-sions are used—in the United States, the average40 percent duty applied using nonmarket economyprovisions was more than 10 times as high whennormal approaches were used.

A particularly worrying feature of the accessionagreement is the product-specific transitional safe-guard provisions that may be applied by any WTOMember, and may then trigger actions againstdiversion of Chinese exports to other markets(Panitchpakdi and Clifford 2002). These provisionsare, in a sense, worse than the provisions onnon–market economy treatment in that they intro-duce an entirely new form of protection, targetedspecifically against China, more readily triggeredthan regular safeguards and available to China’strading partners for up to 12 years from the date ofaccession (Anderson and Lau 2001). The tradediversion measures allowed under this agreementare especially troublesome, because they provideeven less procedural protection than is availableunder regular safeguards.

Messerlin believes that China has two majorchoices in responding to these measures: to retaliateby, for example, increasing use of antidumping andcontesting safeguard actions in dispute settlement;or to mount a concerted campaign for reform ofthe rules in the Doha negotiations, most particu-larly in the area of antidumping. Clearly, China willhave to contest unjustified actions through disputesettlement. However, a retaliatory approach oflaunching antidumping actions could be extremely

Impacts and Policy Implications of WTO Accession for China 3

costly to China’s economy both by reversing liberal-ization and by increasing the degree of uncertaintyabout trade policy. Even though this approach islikely to be politically attractive, as is confirmedby the recent upsurge in antidumping actions inChina—and the dramatic upsurge in these actionsin developing countries such as Argentina, India,Mexico, and South Africa—the economic costs toChina in particular suggest it needs be avoided asmuch as possible.

If China elects instead to lead a push for reformof the antidumping and safeguard rules that wouldreduce the abuses of these protectionist measures, itcould greatly improve the performance of its owneconomy in the short run and the global tradingsystem in the longer term. Here Messerlin providestwo specific suggestions. On non–market economytreatment, he argues that China could press for newrules on the automatic granting of market economystatus in a particular commodity as long as thecountry meets basic conditions such as low rates ofprotection, an absence of core gray-area measures,and an absence of state monopoly in the distribu-tion of that good. On antidumping measures moregenerally, China could put forward, or stronglysupport, proposals to narrow the use of antidump-ing measures and to reduce their severity. Chinamight also seek similar relief on the product-specific safeguards.

Preferential treatment for developing countrieswas a particularly vexing issue throughout theWTO accession negotiations. Although China isclearly a low-income country, because of its sizeand growth performance WTO members werereluctant to give it full developing-country treat-ment. In many areas of the agreement, it is likely tohave full access to the developing-country provi-sions, although in some others—such as agricul-ture where it had to accept a limit of 8.5 percent onde minimis domestic support versus the usual10 percent limit for developing countries—it facestighter restrictions than other developing countries.At the same time, China obtained specific transi-tional arrangements in areas, such as the phasingout of quotas and licenses and phased entry offoreign enterprises, that are not generally avail-able to developing-country Members. Special anddifferential treatment in the form of preferentialaccess to industrial-country markets is not impor-tant for China, which increases the importance of

multilateral trade reform as a means of creatingmarket access opportunities.

A central aspect of the WTO agreement forChina is the TRIPS agreement. In response to itsrecognized need to stimulate innovation in Chinaand access foreign technology and in responseto pressure from its trading partners, China hasstrengthened its intellectual property rights regime.Since 1990, China has updated its laws on copy-rights, trademarks, patents, and trade secrets, andadopted protection for new plant varieties andintegrated circuits.

Intellectual property rights involve a fine balancebetween creating the incentive to innovate andrestricting competition in the market for the goodsand services that result from innovation. The bal-ance between these two objectives differs betweendeveloped and developing countries, so that im-plementing an appropriate intellectual propertyregime in a developing country is not simply a mat-ter of emulating the state-of-the-art regime of adeveloped country. Such a regime may inhibitgrowth by limiting innovation and diffusion, andmay result in excessive transfers to foreign producersof intellectual property. Appropriate regulations areneeded to ensure that markets remain competitivewithout excessive reductions in the incentive toinnovate. The TRIPS agreement generally appears toallow the needed flexibility, but taking advantage ofsuch flexibility is a nontrivial task.

In chapter 4, Keith E. Maskus assesses China’sintellectual property regime—that is, its provisionsfor patents, trademarks, trade secrets, and copy-rights. He bases his assessment in part on a com-parison of China’s, intellectual property rights(IPR) with benchmark standards for middle-income developing countries (World Bank 2001,and on interviews with market participants). Heconcludes that China’s TRIPS regime is broadlyappropriate to China’s situation. He believes, inparticular, that China’s policy of public procure-ment of pharmaceuticals at negotiated prices isappropriate for providing public health services. Healso concluded that, with the reforms underway ataccession, the regime would be fully consistent withTRIPS requirements.

Maskus does, however, raise some importantconcerns about China’s policies related to TRIPS.For example, he believes proposals to extendpatent protection to computer software—a level of

4 China and the WTO

protection currently provided only in the UnitedStates, Japan, and Australia—may be excessive in ayoung industry such as China’s. Maskus also raisesthe serious problems of IPR enforcement, particu-larly of trademarks, patents, and trade secrets. Hebelieves these enforcement problems will inhibitthe transfer of technology in China and the devel-opment of innovative domestic businesses. Abroader problem relates to China’s current lowallocation of resources to research and develop-ment and the therefore limited benefits of protect-ing Chinese innovations. Important issues for thefuture include enhancing pricing regulations onpharmaceuticals as patent protection becomesstronger and developing a broader competitionpolicy regime to deal with abuses of IPRs suchas monopoly pricing and restrictive licensingarrangements.

In chapter 5, Chen Xiwen explores the situationof the agricultural sector in China from the per-spective of a senior policymaker. For basic foods,he emphasizes the shift from a supply-constrainedenvironment to a demand-constrained environ-ment. The demand for food has grown very slowly,while the growth of supplies has accelerated. Mean-while, farmers’ incomes grew very slowly in the sec-ond half of the 1990s. In this environment, Chensuggests that Chinese agricultural policymakersgive priority to, among other things, improvingthe quality of food produced and moving towardhigher-value labor-intensive commodities moresuited to China’s comparative advantage and toraising farmers’ incomes. He sees very clearly thatit is important to allow many farmers to leavethe agricultural sector, and believes the coordi-nated expansion of cities and small towns is central to achieving this relocation of labor. Giventhat the shift from supply-constrained to demand-constrained has been, to a degree, fortuitous andunexpected, a key question is whether the tradi-tional emphasis on food self-sufficiency willreemerge as a policy driver when imports of keyagricultural commodities increase.

Economic Impacts of Accession

A comprehensive evaluation of the effects of majorchanges in trade policy must be based on an assess-ment of the stance of policy prior to accession—an assessment that was particularly difficult in

agriculture and services in China, and nontrivial inmanufacturing. The evaluation must then take intoaccount the implications of the policy measuresbeing introduced. Only then can it take into accountthe impact of these measures on economic variablessuch as output and trade levels by sector, and on theincome levels of people, and particularly poor peo-ple. We first examine the impacts on agriculture,then on manufacturing, then on services.

Agriculture

Many authors have raised concerns about theimpact of WTO accession on China’s agriculturalsector and the many poor people engaged in thissector. Much of this concern has been based oncomparisons of China’s statutory tariffs on agricul-ture in the 1990s with the rates agreed in theaccession process (e.g., Schmidhuber 2001). Otherauthors, such as Johnson (2000) and Lin (2000),have recognized that these statutory tariff ratesbore little relationship to the protection (or taxa-tion) actually experienced by China’s agriculturalsector. However, the evidence about the actual ratesof protection applying to agriculture has remainedextremely limited and frequently contradictory.

Agricultural trade in China has been influencedby a bewildering array of policies on imports andexports, including state trading, designated trading,quotas, licenses, tariffs, and tariff-rate quotas. Manystudies have attempted to deal with this problem bysummarizing the protective impact of agriculturaltrade policies through the price distortions createdby these measures. The more restrictive the trademeasure, in general, the larger will be the gapbetween the domestic price and the internationalprice.

A number of studies have estimated the magni-tude of the agricultural price distortions using theavailable series on domestic and internationalprices. Unfortunately, the results obtained have var-ied wildly. Huang, Chen, and Rozelle (1999), forexample, estimated the protection applying to rice,wheat, and maize in the mid-1990s at 4 percent,20 percent, and 25 percent, respectively. By contrast,Tuan and Cheng (1999) estimated these protectionrates to be �29 percent, 62 percent, and 15 percent,respectively. Carter (2000, p. 80) relied on producerprice data and found generally negative pricedistortions. More generally, Carter (2000) and


6 China and the WTO

Martin (2001) believed that WTO accession wouldrequire relatively little agricultural liberalization inChina, while Schmidhuber (2001) and many othersbelieved that the agricultural liberalization requiredby accession would be dramatic. Clearly, whenscholars reach such different conclusions from thesame facts, a new approach is required, especiallygiven that the policy implications are so large.

For this project, Jikun Huang, Scott Rozelle, andMin Chang adopted a new approach by basingtheir analysis of policy impacts on detailed inter-views with participants in China’s agricultural mar-kets rather than on readily available price series (seechapter 6). This approach provides a much clearerindication of the implications of agricultural tradepolicies for product prices, and of the real-worldimpacts of policies, than would otherwise be possi-ble. Their research reveals, for example, that amajor source of the discrepancies in earlierresearch is differences in the quality of domesticproducts and those traded internationally. Theyalso identify important features of the traderegime, such as export subsidies for maize and cot-ton, that have important impacts on product mar-kets. Finally, they raise an important questionabout the implications of the manner in which thevalue-added tax (VAT) is collected, at a 13 percentrate, for imported agricultural commodities. Itappears that, for administrative reasons, this tax isnot collected on domestic agricultural output at the

farm level. Rather, it is collected only on intermedi-ate inputs into production.

Once the implications of agricultural policieshave been assessed, it is possible to begin to assessthe implications of China’s accession commitmentsfor agricultural markets. For those products pro-tected by ad valorem tariffs, this evaluation isstraightforward. The reduction in the tariff corre-sponds quite directly to the reduction in the domes-tic price of the good, and this change, together withinformation on the slope of the import demandcurve, can be used to estimate the cost of protec-tion. For those products protected by both a tariffand an export subsidy, it may be necessary toconsider changes in both variables. The situationbecomes considerably more complex, however,when products are being protected, or are to beprotected, using tariff-rate quotas. In this situation,the impact depends greatly on whether the in-quota or out-of-quota tariff determines the price ofthe good. When the quota will be filled in someyears but not in others, the average rate of protec-tion may be a combination of the two tariff rates.

Some key assessments that have been made ofthe implications of the level of protection and thechanges associated with WTO accession are givenin table 1.1. The statutory tariff rates for 1998 usedby Schmidhuber and others are given in the firstcolumn of the table; Huang, Rozelle, and Min’s esti-mates of the actual protection provided in 2001 are

TABLE 1.1 Some Measures of Import Protection in China’s Agricultural Sector (percent)

Postaccession 1998 Statutory Tariffs 2001 Protection Protection

Rice 127 �3.3 �3.3Wheat 133 12.0 12.0Maize 130 32.0 32.0Vegetables and fruits 15 �4.0 �4.0Oilseeds 132 20.0 3.0Sugar 30 40.0 20.0Cotton 3 17.0 20.0Livestock and meat 35 �15.0 �15.0Dairy 46 30.0 11.0

Note: The third column of the table shows anticipated average rates of import protection after accessiontaking into account the reforms required by accession and likely market outcomes.Sources: Average statutory rates taken from Schmidhuber (2001) and www.chinavista.com. Estimates ofprotection in 2001 provided by the authors.

given in the second column. The third columnshows the anticipated average rates of importprotection after accession taking into account thereforms required by accession and the likely marketoutcomes.

Huang, Rozelle, and Min estimated the averagerate of protection for rice to be slightly negative,implying that China’s system of state trading for riceoperated to tax rice exports slightly in 2001. Afteraccession this rate is expected to remain the same,because accession to WTO does not require reduc-tions in such protection for a state-traded commod-ity. For wheat, protection averaged an estimated12 percent. This rate of protection need not begreatly reduced, on average, because it seems likelythat wheat imports will exceed the tariff-rate quotareasonably frequently (see Martin 2001), allowingimposition of a tariff of up to 65 percent. For maize,the rate of import protection utilized was higher, at32 percent, because of an export subsidy. Although,on average, the level of import protection need notchange greatly because there is a significant proba-bility that the tariff-rate quota will bind by the endof the decade (Martin 2001), the export subsidymust be abolished, implying a potentially substan-tial reduction in the price support given to maize.Labor-intensive vegetables and fruits, like livestockproducts, experienced negative rates of protectionin 2001, and the WTO accession commitments areunlikely to require changes in protection.

Oilseeds present an entirely different picture inwhich the principal form of protection has been atariff, and the tariff is being reduced substantially.The protection provided to sugar must be reducedto meet China’s commitments to a bound tariff of20 percent. As for cotton, import protection willnot change greatly, but export subsidies, such as the10-percent export subsidy observed in 2001, areruled out in the future. Protection of livestock andmeat could remain negative as a consequence ofexport restrictions to markets such as Hong Kong.Finally, protection of dairy products can beexpected to decline to meet China’s tariff-bindingcommitments.

The reductions in protection between secondand third columns of table 1.1 are just one possibleoutcome in a situation in which rates of agricul-tural protection can vary substantially, particularlyif import levels exceed the tariff-rate quotas. How-ever, the agricultural protection measures illustrate

the importance of the discretion remaining toChina’s policymakers even after implementation ofsuch comprehensive commitments on agriculture.Removal of the negative protection applying tolabor-intensive products would also be consistentwith WTO rules, and is likely to be particularlyhelpful for employment in rural areas, as well aseconomic efficiency.

Participation by China in the WTO agriculturalnegotiations being conducted under the DohaDevelopment Agenda could potentially reinforcethese benefits by opening large, and currentlyhighly protected, markets for China’s labor-intensive agricultural exports. Unfortunately, thehigh rates of agricultural protection that arosewhen GATT rules on agriculture were extremelyweak mean that China faces barriers to its agricul-tural exports that are four times higher than those itfaces on its other merchandise exports (Martin2001).

In chapter 7, Kym Anderson, Jikun Huang, andElena Ianchovichina examine the implications ofChina’s agricultural commitments for agriculturalprotection and for the agricultural sector. They findthat the basic WTO accession commitments toreduce agricultural import protection and elimi-nate agricultural subsidies will make those farmhouseholds dependent on agriculture worse offrelative to urban households. If China elects as wellto remove the negative protection of importantcommodities such as rice, vegetables, and meats,then returns to unskilled rural labor, and to farm-land would rise slightly, and the impact of accessionand the elimination of negative protection on ruralwages would be �0.5 percent instead of the–0.7 percent observed in the accession case. Theyalso highlight the importance of China pursuingimprovements in agricultural market access,citing results from Yu and Frandsen (2002) thatsuggest that agricultural liberalization by membercountries of the Organisation for EconomicCo-operation and Development (OECD) wouldbenefit China and improve its agricultural tradebalance.

Industrial Products

In chapter 13, Ianchovichina and Martin examinethe reductions in protection under way in manufac-turing and services as well as in agriculture. Within


industry, they find that accession builds on the sub-stantial reductions in tariffs undertaken during the1990s, when weighted average tariffs on manufac-tures fell from 46.5 percent in 1992 to 25 percent in1995. By the time of accession in 2001, weightedaverage tariffs on manufactures had fallen to around13 percent. With full implementation of China’saccession commitments, they will fall to 6.9 percent.The six-percentage-point reduction in average tar-iffs to be implemented after accession is important,but very small relative to the 33-percentage-pointreduction undertaken since 1992.

The largest reductions in industrial tariffs arerequired for beverages and tobacco—an almost28-percentage-point reduction from 2001 levels. Inautomobiles, the reduction was just over 15 per-centage points. Although large, these reductionswere very much smaller than those undertakensince 1995. The reduction in protection to the auto-mobile sector is particularly important because ofthe high profile of this industry and its pervasivelinkages throughout the economy. Other industriesin which substantial reductions in tariffs wererequired included textiles, clothing, electronics, andlight manufactures. Many of these industries arerelatively labor-intensive ones in which China has acomparative advantage and in which liberalizationwill help maintain efficiency and competitiveness.

The liberalization considered in this projectis, apart from the export quotas on textiles andclothing, focused entirely on tariffs. This emphasisunderstates the degree of liberalization resultingfrom China’s accession, because it ignores the abo-lition of nontariff barriers such as designated trad-ing, quotas, and licenses. The seriousness of thisexclusion is much less than it would have been inearlier years because of the dramatic reduction inthe coverage of nontariff barriers in China—thefrequency of import licenses, in particular, fell fromalmost half 4 in the late 1980s (Lardy 2002) to under5 percent in 2001. Given the great uncertaintyabout the protective impacts of nontariff barriersin China and their limited remaining coverage,Ianchovichina and Martin concluded that it wasperhaps better to use tariffs only when consideringliberalization, and therefore to present somethingof a lower-bound estimate of the benefits of the lib-eralization associated with accession.

It is also important to take into account thereduction in the barriers facing China’s exports of

textiles and clothing. Because these barriers requirean exporter to purchase an export quota—or toforego the opportunity to sell quotas alreadyheld—they impose a cost on exports that is similarto an export tax. Based on information on quotaprices, this tax was estimated to be about 15 percentfor clothing and 10 percent for textiles.5

The estimates of the extent of merchandise tradeliberalization used in this study omit some impor-tant and potential elements of trade policy. Oneis the possibility of antidumping and safeguardsmeasures being applied against China. Another isthe increasing use of measures of this type inChina. The potential use of the product-specificsafeguards against China would be particularlyimportant in this respect, because no such measuretargeted specifically at China existed prior toChina’s accession. The risk that China will increaseits use of antidumping and safeguard measuresbeyond the current levels is also a concern fordevelopment policy. Such an action would be atriumph of a rules-focused approach to WTOimplementation—it is legal; therefore, we shoulddo it!!!—over the sharp focus on promoting devel-opment that has characterized China’s trade policyagenda since the beginning of the reform era.

Services

Trade in services was a key area in China’s WTOaccession negotiations. To obtain a clear under-standing of the implications of China’s commit-ments in this area, and the potential usefulness ofthe WTO process for China’s economic develop-ment, we consider first Mattoo’s overview ofChina’s commitments in services (chapter 8), andthen detailed case studies of logistics (chapter 9),telecommunications (chapter 10), and the financialsector (chapter 11).

Chapter 8 by Aaditya Mattoo compares the lib-eralization undertaken by China in services withthat undertaken by other groups of countries, andit concludes that China’s services reform is the mostradical ever negotiated in the WTO. In cross-bordertrade (mode 1), consumption abroad (mode 2),and establishment trade (mode 3), China has mademore commitments in more service sectors thanthe industrial countries, other developing coun-tries, or the group of countries that recentlyacceded to the WTO. Of course, any assessment of

8 China and the WTO


the importance of these commitments depends ontheir implementation, which has drawn reserva-tions from Whalley (2003). However, the first tworeviews of China’s services commitments at theWTO appear to indicate that considerable progresshas been made (see WTO 2003a, 2003b).

Mattoo notes, however, that China’s commit-ments to service market liberalization were notindiscriminate. The number of sectors with guar-anteed unrestricted access is lower for the first twomodes than that in most other countries, andessentially zero for mode 3. An important feature ofChina’s commitments is that they focus on marketaccess and provide national treatment in notdiscriminating between domestic and foreignsuppliers.

Frequently applied restrictions on the establish-ment of services enterprises include:

• restrictions on the form of establishment • restrictions on geographic scope• regulatory requirements.

Restrictions on the form of establishment, suchas requirements for joint ventures, have a longhistory in China and have frequently been justifiedas a means of achieving technology transfer or asmeans of obtaining a share of monopoly rents.Mattoo, however, points out that requirements toform joint ventures might inhibit the transfer oftechnology. He also points out that a more thor-ough approach to the problem of monopoly rentswould be to ensure that competition betweenfirms, whether domestic or foreign, eliminatesthese rents.

Restrictions on geographic scope also have along history—from an era when experimentationwith market-oriented approaches had to be isolatedbecause of the inconsistencies between, for exam-ple, planned and market prices. Confining foreignventures, such as insurance, to five cities for fiveyears might encourage agglomeration of theseactivities in these cities that will not be reversedwhen the geographic restrictions are later lifted.Moreover, such an approach may reduce theopportunities for other parts of China, such as inte-rior cities with a potential comparative advantagein these activities, to get started in these activities.Yet China’s WTO commitments represent only alower bound, and so they have not prevented

Chinese authorities from moving ahead faster thanthe minimum to which they have committed insome cases. Given the risks of exacerbating thealready substantial inequalities between coastal andinterior provinces, there would appear to be gooddevelopment reasons to phase out these geographicrestrictions more quickly than is required by theWTO commitments.

Imposing regulatory requirements is an impor-tant role of government.Some of the most importantobjectives of regulations are to make competitionwork, improve the availability of information to con-sumers, and ensure universal service. Making com-petition work is particularly important in networkindustries such as telecommunications, where indi-vidual firms frequently do not find it in their interestto allow interconnection by new firms. Improvingthe availability of information is especially impor-tant in financial services. Finally, developing efficientprovisions on universal service is important forensuring that all parts of China have access to tele-communications services.

In chapter 9, Wenping Luo and ChristopherFindlay provide a detailed assessment of the re-forms required by WTO accession in the range ofservice activities that make up the logistics chain.They note that although China has made substan-tial progress in many of the component activities,logistics costs are disproportionately high in China,and service quality is lower than is desirable, in partas an enduring legacy of the planned economy. Luoand Findlay conclude that logistics accounts for30–40 percent of the wholesale cost of manufac-tured goods, as contrasted with 5–20 percent in theUnited States, implying that the scope for gain fromliberalization is particularly large. High logisticscosts are a particularly important problem for peo-ple in the poorer areas of China, whose ability totrade, and consequently their real incomes, are sig-nificantly reduced by these excessive costs.

China’s WTO commitments on logistics applyto a range of specific General Agreement on Tradein Services (GATS) service sectors, including pack-aging and courier services; maritime and rail trans-portation; freight forwarding; and storage andwarehousing services. The commitments made inthese sectors promise to increase competition insome key areas, including road transport, rail trans-port, warehousing, and freight forwarding. Thebreadth of these commitments also provides a

much stronger basis for development of integratedthird-party logistics firms able to reduce the costsand increase the quality of logistics services inChina. If the envisaged reductions in logistics costsare brought to fruition, Luo and Findlay concludethat the costs of a wide range of goods and servicesmight be reduced by about 10 percent.

Yet the commitments outlined in the GATS agree-ment are not sufficient to achieve the full potential oflogistics in China. Regulatory reforms are needed toremove discrimination against particular enterprisetypes, to separate local administrations from enter-prises, and to eliminate local protectionism. In addi-tion, substantial investments in infrastructure areneeded to improve the timeliness and reduce the costof providing logistics services.

In chapter 10, Mari Pangestu and DebbieMrongowius explore China’s commitments intelecommunications—commitments that take onparticular importance given that China is expectedto be the largest market for telecommunications inthe world by 2010. These commitments are pro-found in that they allow foreign entry to a widerange of activities that are currently closed to foreigninvestment. Furthermore, this entry takes place in asector that was monopolized by China Telecom until1994 and is currently dominated by a small numberof state-owned firms (see DeWoskin 2001).

In basic telecommunications, China has com-mitted to the disciplines of the WTO referencepaper on the regulatory framework for telecommu-nications. The objective is to create a competitiveenvironment in which interconnection betweensystems is allowed under reasonable and nondis-criminatory conditions, and it allows for universalservice provisions. It also requires the existence of aregulator independent of the telecom provider andsets criteria for licensing entry and for allocatingscarce commodities such as the mobile telephonespectrum.

In the context of a basic telecommunicationssystem governed in line with the regulatory paper,China’s other GATS commitments cover value-added services such as voice mail and online infor-mation services; mobile voice and data services;and domestic and international services such asprivate leased circuit services. Most of these ser-vices are initially subjected to a combination ofownership restrictions and geographic restrictionswithin China. Although the geographic restrictions

will be phased out over several years, China has notcommitted to allowing more than 49 percent for-eign ownership for important services such asmobile telephony. Because allowing higher levelsof foreign ownership would be consistent withChina’s GATS obligations, China may do so in thefuture if such a move appears to promise worth-while gains in a particular activity.

Clearly, the challenge of establishing a telecom-munications sector that makes the maximumcontribution to China’s overall growth and devel-opment will require further expansion of the regu-latory framework. Pangestu and Mrongowiusbelieve the central issues here will include ensuringthe independence of the regulator, ensuring thatinterconnection works adequately, and makingpricing regulations more flexible.

In chapter 11, Deepak Bhattasali examines theimplications of China’s commitments in the finan-cial sector, including banking, security trading, fundmanagement, and insurance. According toBhattasali, reforms up to 1997 focused on institu-tional diversification and strengthened administra-tive oversight. Since 1997, reforms have addressedthe portfolio problems of the banks and governanceof the financial sector in preparation for WTOaccession. In 2001, however, the four large statebanks still accounted for 67 percent of bank depositsand 56 percent of total financial assets. The shareof loans going to small and medium-size non–state enterprises—the most dynamic part of theeconomy—remains small, a finding evident inenterprise survey work reported by Dollar (2002).Another key concern is the dearth of the informa-tion needed to assess the performance of the finan-cial sector.

Although a large number of foreign banks wasactive in China on the eve of accession, theyoperated almost exclusively on an offshore orenclave basis and accounted for less than 3 percentof bank assets. However, two years after accessionthey became able to provide local currency servicesto Chinese enterprises, and they will becomeeligible to offer these services to individuals withinfive years. Other areas of financial services, such asstockbroking, fund management, and insurance arealso being opened up on quite short timetables.

Like other scholars such as Langlois (2001),Bhattasali believes that the reforms agreed upon inthis vitally important sector were a clear attempt to

10 China and the WTO

increase competition, performance, and the rangeof products available in financial services. The keychallenges will lie in managing the transition to amarket-based system without serious problems or afinancial crises. He believes that the state banks willcome under serious pressures from their nonper-forming loans, weak management systems, lowoperating margins, and the strong competitionthey will face. His assessment is that financial re-engineering and fairly radical actions to reduceoperating costs will be required, but that, givensuitable reforms, the rehabilitation of the statebanks is unlikely to present major problems.

Impacts on the Economy

Because China’s reforms are so broad-ranging, andtheir economy-wide interactions are so extensive,they must be evaluated on an economy-wide basis.In chapter 13, Ianchovichina and Martin assess theimpacts of liberalization on agriculture, manufac-tures, and services. Previous research, however,revealed that undertaking any such assessment in asatisfactory manner required special attention tothe motor vehicle sector. In an analysis in whichthey did not account for the restructuring of thisindustry, Ianchovichina and Martin (2001) foundthat output in this industry would decline abso-lutely over the period to 2007, despite the strongincreases in the Chinese demand for automobilesand the shift in China’s comparative advantage tomore capital- and skill-intensive products such asmotor vehicles.

In their contribution to this study, Joseph F.Francois and Dean Spinanger examine China’smotor vehicle industry (chapter 12). Like previousstudies such those by Harwit (2001) and theChinese Academy of Engineering and NationalAcademy of Engineering/National Research Coun-cil (2003), they conclude that it has been shaped bypolicies that have encouraged market segmentationand suboptimal plant size—a structure that is seenfrequently in countries with highly protected auto-mobile industries. As a result, the industry is veryinefficient, with most plants operating well belowglobal standards for efficient production. Unlessthese problems can be overcome, the industry isexpected to respond to the fall in protection, andthe increased competition for labor from expand-ing sectors such as textiles and clothing, with a

sharp decline in output achieved through wide-spread plant closures.

Francois and Spinanger conclude that restruc-turing in the industry to achieve scale economies infinal assembly would reduce costs by about 20 per-cent. This reduction would more than reverse thenegative impact on output of the reduction in pro-tection from 1997 levels, allowing the industry toexpand relative to the no-accession case—and toexpand dramatically as China’s growth and shiftingcomparative advantage shift resources into sectorssuch as motor vehicles. There also would be impor-tant changes within the motor vehicle sector.According to the analysis by Francois andSpinanger, the increase in efficiency of the finalassembly industry relative to the production ofintermediate parts is likely to increase the demandfor imported parts substantially, with their sharerising from 39 percent to 52 percent. If the restruc-turing of the industry is achieved successfully,exports of finished motor vehicles are projected toincrease very rapidly, resulting in an increase intotal exports of vehicles and parts of over US$4 bil-lion6 a year. Clearly, the reforms associated withaccession will require major changes through-out the industry and considerable restructuring,although this process has perhaps been eased by thevery rapid growth in demand during recent years.

In chapter 13, Ianchovichina and Martin ana-lyze the impacts of liberalization associated withaccession in agriculture, manufactures, and serv-ices,7 and the opportunities arising from the elimi-nation of the quotas against China’s (and othercountries’) exports of textiles and clothing. Theiranalysis takes into account China’s importantexport processing arrangements, and it builds onthe labor market study by Sicular and Zhao (chap-ter 14), a study by Shi Xinzheng (2002) on labormarkets, and the Francois and Spinanger analysis ofautomobile industry restructuring (chapter 12).The resulting changes in the specification of theirmodel greatly increase the realism of their analysisand have very important implications for theirresults.

Ianchovichina and Martin divide the effects ofWTO accession into a component of liberalizationundertaken between 1995 and 2001 in preparationfor accession and the remaining component to beundertaken after 2001 in order to meet China’saccession commitments. The choice of 1995 as a



starting period is somewhat arbitrary in view of thefact that China reformed its trade regime through-out the 1990s. However, 1995 was an importantturning point, when China had to forgo its hopeof resuming its seat in the GATT and apply as anewcomer to the WTO under a process much morefocused on the commercial implications of theaccession package. Ianchovichina and Martin rec-ognize that China’s product mix and the world’sdemand for China’s exports are changing rapidly.They therefore superimpose the impact of liberal-ization on a situation in which China’s industrialstructure and output and trade patterns are veeringsharply toward more capital- and skill-intensivegoods in response to high rates of investment andrapid growth in educational levels.

The authors find that the liberalization associ-ated with WTO accession results in substantialgrowth in trade relative to output, with the totalvolume of exports rising by 17 percent as a conse-quence of the liberalization after 2001. The mostrapid growth in exports is in apparel; exportsexpand by over 100 percent in response to the abo-lition of the export quotas on clothing. Exports ofmost agricultural products rise because of a declinein agricultural input costs and constraints on theout-migration of labor from the agricultural sector.However, exports of plant-based fibers (predomi-nantly cotton) are projected to fall in response tothe increase in demand for use in export produc-tion. Feed grain exports also fall because of the abo-lition of the export subsidy on exports of maize.Exports of automobiles rise substantially becauseof the increase in the efficiency of the sector as itexploits economies of scale and the greater tradeexposure of the industry in the more liberalizedpostaccession economy. None of the increases inexports takes into account the possible benefitsto China of being able to expand its marketaccess through participation in WTO negotiations.The expansion in textile and clothing exports is,in fact, a delayed benefit from the Uruguay Round,previously denied to China as a nonmember ofthe WTO.

Imports rise in a range of sectors in which thereare substantial reductions in trade barriers, includ-ing beverages and tobacco, processed food, textiles,clothing, oilseeds, dairy products, and sugar.Because Ianchovichina and Martin have repre-sented trade liberalization in services as reducing

trade barriers, there are also substantial increases inimports of services. The increase in imports of bev-erages and tobacco is the largest because of the sharpreduction in the tariffs for these commodities.

The biggest change in employment after acces-sion is an increase in employment in the apparelsector of more than 50 percent. Employment in thetextile sector and in plant fibers also increase tomeet the demand from the apparel sector. Tradereform leads to small reductions in employment inmost agricultural sectors and in manufacturingsectors such as petrochemicals, metals, and auto-mobiles. Overall, however, the movements of laborbetween sectors are generally small relative to thechanges in trade patterns.

The overall welfare gains to China are substan-tial, particularly from the liberalization undertakenbetween 1995 and 2001 to prove China’s bona fidesand to prepare for accession. The benefit from theliberalization undertaken during this period isestimated to be a continuing gain of about $30 bil-lion per year. The smaller reduction in protectionbetween 2001 and the end of the implementationperiod will generate incremental gains to China of$10 billion a year. The measured gains in exportand income growth are very much lower-boundestimates in that they ignore the benefits from abo-lition of nontariff barriers and they involve seriousaggregation biases. Moreover, models of this typeappear to greatly understate the implications ofmajor trade liberalizations, particularly those asso-ciated with the rapid emergence of new products(Kehoe 2002), which Martin and Manole (2003)have found to be a major feature of China’s reformexperience.

The gains from reform do not accrue evenlywithin China, with wages of skilled and unskilledurban workers rising modestly and wages forunskilled farm workers declining by 0.7 percent inreal terms. Ianchovichina and Martin also examinethe implications of some potential complementarypolicies that might be able to deal with this prob-lem. Reform of the hukou system (a system ofresidence permits regulating movement betweenurban and rural residence) and other labor marketbarriers is highly desirable because of the ineffi-ciencies and inequities associated with this segmen-tation of the labor market. However, such reformincreases in importance after WTO accessionbecause of the partial liberalization of the

agricultural sector, which increases the pressuresfor workers to leave agriculture. Adding abolitionof the hukou system to the policy package is foundto result in an increase in rural wages of almost17 percent. Hukou removal completely overwhelmsthe reduction in rural wages associated with tradeliberalization, as an estimated 28 million agricul-tural workers leave the agricultural sector. Com-plete abolition of the hukou system would putdownward pressure on unskilled wages in urbanareas, although this reduction of an estimated3.8 percent would be much smaller than the gain torural workers.

Another complementary reform thatIanchovichina and Martin consider is an expansionin access to education—a change that would helpto raise unskilled wages in both urban and ruralareas. Improving agricultural technology is anotherpolicy option with potentially large benefits forpoor rural households that are able to adopt thenew production techniques.

The results of this quantitative analysis arehighly stylized, because they assume that enter-prises and households are able to adjust success-fully to the changes in incentives created by WTOaccession. As the OECD (2002) has pointed out,successfully making these changes is likely torequire considerable strengthening of the economicsystem in areas such as enterprise governance andreform of the banking system.

Impacts on Households andon Poverty

The simplest possible approach to capturing theeffects of WTO accession at the household levelrequires an assessment of its impacts on the pricesconsumers pay; on the prices that owners of labor,capital, and other factors receive for their resources;and on the government’s ability to provide transfersor public goods. In addition, it is useful to be ableto assess the ability of households to adjust to thechanges resulting from accession, perhaps bychanging occupations or activities.

Because ongoing work on trade and poverty hasfound that impacts through factor markets are con-sistently more important than impacts throughconsumer prices, it is important to examine factormarket effects carefully. This is especially true inChina because of the large income differences

between urban and rural workers and because ofthe explicit policy barriers resisting the movementof labor between urban and rural sectors.

In chapter 14, Terry Sicular and Yaohui Zhaoexamine the options that households in China facein selling their labor—the most important resourceof the poor. They find substantial differencesbetween the earnings of rural and urban house-holds. Although some of these differences stemfrom differences in the quality of labor and otherresources, much is attributable to differences inthe earnings received for resources of the samequality—a finding that is consistent with the exis-tence of substantial barriers between urban andrural labor markets. These barriers appear to beparticularly strong for poorer households, con-tributing both to depressed income levels and todifficulty in adapting to changes in economicopportunities.

Another important feature of Chinese labormarkets captured by Sicular and Zhao is the reluc-tance of China’s farm households to transfer laborfrom agriculture to other sectors. Although suchreluctance is frequently observed in countries thatare attempting to transfer labor out of agriculture,it is exacerbated in China by restrictions on the saleof farmland usage rights. Permanent movement islikely to require families to relinquish their landrights without compensation. A main objective ofthe study by Sicular and Zhao is to provide esti-mates of the responsiveness of rural workers tochanges in relative returns. They find rural house-holds quite responsive to movements in agricul-tural returns relative to market wage rates, withan estimated response elasticity of 2.67, but theresponse to changes in the ratio of agricultural andnonfarm business returns is much lower. Althoughlabor is not perfectly mobile between urban andrural labor markets, these results make clear that itis also far from immobile.

Two studies in this volume integrate the findingsof the different studies and analyze their implicationsfor poverty. They make these assessments by usinggeneral equilibrium models of China’s economy toevaluate the implications of reform for the prices andfactor returns faced by China’s households and thenexamining the effects on particular households.

The study by Shaohua Chen and MartinRavallion presented in chapter 15 considers theimpact of WTO accession on income distribution


and poverty drawing on the simulation modelresults provided by Ianchovichina and Martin. Thisstudy uses a sample of 84,000 households—17,000urban and 67,000 rural—from surveys by China’sNational Bureau of Statistics. The price impactsimplied by Ianchovichina and Martin’s GlobalTrade Analysis Project (GTAP) model analysis areapplied to the households, taking into account theimpacts on the prices households must pay fortheir consumption goods and purchases of inputs,and the prices they receive for their sales of goodsand of labor and other factors. The loss of govern-ment revenue from falling tariffs is restored verysimply a tax that raises the price of all consumptiongoods to maintain government revenues.

Chen and Ravallion’s analysis is focused on theshort run, in which households are limited in theirresponses to changes in prices, and on the periodafter accession in 2001, when the impact on theaverage household and on poverty rates was verysmall. They consider the impacts on individualhouseholds and find sharp differences, particularlybetween urban and rural households. Most urbanhouseholds, and especially the relatively poorurban households, gain from WTO accession, withthe poorest urban group gaining by about 1.5 per-cent of initial incomes. This is not true for ruralhouseholds. The very poorest rural householdsexperience a sharp reduction in their livingstandards—about 6 percent for the poorest per-centile. This reduction reflects a combination offalling rural wages and increases in the prices ofconsumption goods—items consumed in substan-tial quantities by members of this group. Overall,almost 90 percent of urban households gain fromWTO accession, while over three-quarters of ruralhouseholds lose, although the losses are generallyquite small. The estimated losses to rural house-holds are larger in the Northeast than in otherareas, with average losses to rural households ofover 2 percent in Heilongjiang and Jilin.

The study by Thomas W. Hertel, Fan Zhai, andZhi Wang in chapter 16 uses a model of theChinese economy that takes into account impor-tant features such as the duty exemptions forintermediate goods used in the production ofexports. Because of limitations on the availabilityof household data for analytical purposes, itfocuses on the three relatively diverse provinces ofLiaoning, Sichuan, and Guangdong. This study

takes a longer-run perspective than the Chen andRavallion study, and it assumes that households aremore able to move labor between agriculture andother activities, with an elasticity of transformationof 2.67 for the movement of labor between agricul-tural and nonagricultural activities. Like Chen andRavallion, the authors find that urban householdsbenefit substantially more than rural householdsfrom WTO accession. However, in this longer-runanalysis even agriculture-specialized householdsgain on average. Only in a shorter-run experiment,where households are much less able to movebetween sectors, do some agricultural householdssuffer small losses.

Hertel, Zhai, and Wang also consider a range ofcomplementary policy reforms that might help todeal with adverse impacts on rural households.Abolition of the hukou system is found to bestrongly beneficial for rural workers, as are expan-sions in education. Both of these reforms are foundto be substantially favorable for rural households.

In chapter 17, Athar Hussain analyzes theChinese system of social protection and its value asa safety net for shocks of the type involved in WTOaccession. In his view, China’s social protection sys-tems should not attempt to compensate losers fromthe reforms; it is simply too difficult to determinethe magnitude of compensation that is appropriate.Rather, Hussain believes the focus should be onpreventing poverty. He notes that the wide distri-bution of land ownership by rural households hasplayed an important role in maintaining livingstandards and providing insurance against adverseshocks, but views this insurance as less valuableagainst shocks to grain prices resulting from liber-alization than against shocks such as the loss of anurban job.

Hussain observes that China’s systems forincome maintenance and poverty relief have threebroad characteristics: a stark dichotomy betweenurban and rural systems; a focus on the reductionof absolute poverty; and a high degree of decentral-ization in financing. He finds the system in urbanareas relatively comprehensive, but views the sys-tem in rural areas as seriously deficient. The locallyfunded system in place in many areas is unable todeal with a large-scale shock of the type likely tooccur with accession to WTO—“the social safetynet is full of holes.” Hussain goes on to argue forextending basic safety net protection into rural


areas. One first step might be to extend a mecha-nism like the urban unemployment insuranceschemes to wage employees in township and villageenterprises. However, Hussain believes the centralelement is the development of a national scheme,such as the Minimum Living Standard Scheme,that targets poor rural households.

Conclusions

China’s WTO accession agreement calls for sub-stantial reductions in protection, for a strengthen-ing of intellectual property rights protection, andfor adoption of a framework of trade rules at homeand abroad to facilitate trade growth. While requir-ing many policy changes, it leaves open a widerange of policy choices and, in these, China shouldcontinue to focus strongly on its developmentneeds—the central perspective that has guided thesteady transformation of its trade regime, and itseconomy generally as it moves from a planned to amarket economy.

One key concern in the accession agreement isthe provisions on antidumping and safeguards. Thenonmarket provisions that countries are permittedto invoke against China for up to 15 years are likelyto result in antidumping duties substantially higherthan those in other countries—in a situation inwhich China faces seven times as many antidump-ing actions per dollar of exports as the UnitedStates. The special product safeguards applicableagainst China for the next 12 years are a new formof protection, applicable only against China. Theprovisions on trade diversion lack even basic proce-dural restraints, and pose a potentially seriousthreat to China’s export development. The tempta-tion to retaliate, particularly with antidumpingactions of its own, will be strong but would damageChina more than other countries. A better optionfor China would be to seek reform of WTO rules inthese areas.

China’s TRIPS arrangements generally appear tohave been implemented in a manner consistentwith both the legal requirements of the agreementand China’s development needs. They generallyprovide a reasonable balance between incentives toinnovate and access to those innovations. Someconcerns have emerged about proposals that mightstrengthen protection too much to allow access toinnovations such as those for patent protection on

software and about problems with inadequateenforcement of IPRs that reduces the incentive toprovide innovations.

Agriculture is being liberalized by less than wassuggested by some earlier studies that began fromthe assumption that agricultural protection wouldbe reduced from initial statutory tariff levels. Signifi-cant liberalizations have been achieved in areas suchas maize, cotton, and sugar, and it appears that theadjustment pressures in these industries will be sig-nificant. There will, however, be opportunities toexpand exports of some labor-intensive exportsas part of a broader policy reform and by seekingincreases in agricultural market access in the Dohanegotiations.

The industrial sector will face substantial adjust-ment pressures in key sectors such as automobiles,and beverages and tobacco, where external protec-tion is being substantially reduced. Restructuring ofscale-intensive sectors such as the automobile sec-tor will be essential and can generate substantialproductivity gains. Overall, however, most of theadjustment in this industry has already occurred,and what remains is an expansion of both importsand exports.

China’s GATS commitments represent perhapsthe most thorough-going liberalization of the serv-ices trade ever undertaken in the GATT. The rangeof commitments is extremely broad, although somecommitments involve restrictions on ownership,business scope, or region. Critical sectors such astelecommunications, logistics, and financial serv-ices are to be confronted with renewed competition,and they are likely to see a burst of innovation andproductivity growth as restructuring proceeds.

The impacts of China’s WTO commitments—and China’s successful economic development—depend heavily on the ability of China’s labormarkets to reallocate labor from agriculture toother activities. Analysis reported in this volumeconcludes that these markets are adversely affectedby a range of regulations such as the hukou system.Other features of the labor market, such as the “tie”to the land where households have use rights toland but cannot sell it because property rights arenot sufficiently well defined, restrict the mobility oflabor out of agriculture. These features of the econ-omy inhibit the adjustment needed after accessionand increase the vulnerability of poor people todownturns in agricultural prices.


We report in this volume two studies that evalu-ate the impacts of trade reform on poverty. Bothstudies find that the rural sector is more vulnerableto this reform than is the urban population, andboth conclude that urban households benefit morethan rural households from the package of reformsassociated with accession. The study that looks atthe shorter term finds rural households are moresubject to negative shocks. Although these impactsare in general relatively small, they affect most ruralhouseholds. The effects are largest in proportionalterms among the poorest households, and tend tobe geographically concentrated, particularly in thenortheastern provinces such as Heilongjiang andJilin. The second study provides a relatively opti-mistic longer-run picture, where virtually allhouseholds gain from the reform.

These models and the overall modeling studiessuggest that the answer to mitigating these prob-lems is to include reforms such as reductions in thebarriers against movement of labor out of agricul-ture. Removal of the hukou system would increaserural wages substantially, with a relatively modestnegative effect on higher-income households. Arange of other policy options directed towardimproving the welfare of rural people, such as ruralresearch and development, improvements in ruralinfrastructure, expansion of educational opportu-nities in rural areas, and improvements in ruralhealth services, are all fully WTO consistent andcould have powerful beneficial effects in both theshort and the long run. As an important WTOmember, China also has an opportunity to press forgreater market access opportunities for labor-intensive agricultural exports that could help gen-erate employment in rural areas.

The network of social protection measures inChina is quite underdeveloped and a constraint onthe country’s ability to grow while dealing withwidely held concerns about the need to compensatethe potential “losers” from the policy reforms. Par-ticular attention should be given to strengtheningthe social welfare systems available to ruralresidents.

China’s accession to the WTO was a definingmoment for both China and the WTO. It necessar-ily moves the process of trade reform in China awayfrom an incremental approach to one incorporat-ing quite detailed rules for trade policy. It is impor-tant to combine the spirit of China’s reform

process—with its emphasis on continuing reformdesigned to improve economic performance—withthe rules-oriented approach of the WTO. The ben-efits of accession to China will be substantial, butreaping them will require continuing reforms in theadventurous spirit of China’s economic reforms—reforms that focus on economic development andgo beyond trade policy into areas such as labormarket reform and rural development.

Notes

1. See chapter 13 for a discussion of the methodology usedfor these computations.

2. Although China’s transitional review at the WTO allowsother WTO members to seek information and raise concernsabout China’s trade policies, the process is multilateral, andChina can raise concerns about implementation of other Mem-bers’ commitments in the protocol.

3. Although internal protectionism appears to have abatedin recent years, survey results suggest that problems remain, par-ticularly local government interventions in labor markets andadministrative discrimination (DRC 2003). These interventionsappear to hinder foreign investment (Amiti and Javorcik 2003).

4. Two-thirds of exports were also subject to licensing.5. See www.chinaquota.com.6. All dollar amounts are current U.S. dollars.7. The estimated liberalization of services is very crudely

approximated by halving the barriers to trade in these activitiesestimated by Francois.

References

The word processed describes informally reproduced works thatmay not be commonly available through libraries.

Amiti, M., and Javorcik, B. 2003. “Trade Costs and Location ofForeign Firms in China.” Paper presented at Workshop onNational Market Integration. Development Research Centreof State Council and World Bank. Beijing, September 6.

Anderson, S., and C. Lau. 2001. “Hedging Hopes with Fears inChina’s Accession to the World Trade Organization: TheTransitional Special Product Safeguard for Chinese Exports.”Powell, Goldstein, Frazer, and Murphy LLP, Geneva.Processed.

Carter, C. 2000. “China’s Trade Integration and Impacts on Fac-tor Markets.” In Organisation for Economic Co-operationand Development, China’s Agriculture in the InternationalTrading System, OECD Proceedings. April.

Chinese Academy of Engineering and National Academy ofEngineering/National Research Council. 2003. Personal Carsand China. Washington, D.C.: National Academies Press.

Chun, C., B. Fleisher, and E. Parker. 2001. “The Impact ofChina’s Entry into the WTO: Overview.” China EconomicReview 11(4): 319–22.

DeWoskin, K. 2001. “The WTO and the TelecommunicationsSector in China.” China Quarterly 167: 630–54.

Dollar, D. 2002. “Improving the investment Climate in China.”Presentation to the Conference on China’s InvestmentClimate, Beijing, December 3.

DRC (Development Research Centre). 2003. “Measures, Objectsand Degrees of Local Protection in the Chinese Market—An


Analysis Based on a Sample Survey.” Paper presented atWorkshop on National Market Integration. DevelopmentResearch Centre of State Council and World Bank. Beijing,September 6.

Fewsmith, J. 2001. “The Political and Social Implicationsof China’s Accession to the WTO.” China Quarterly 167:573–91.

Gilbert, J., and M. Wahl. 2002. “Applied General EquilibriumAssessments of Trade Liberalization in China.” WorldEconomy 25(5): 697–731.

Harwit, E. 2001. “The Impact of WTO Membership on theAutomobile Industry in China.” China Quarterly 167:655–70.

Huang, J., M. Chen, and S. Rozelle. 1999. “Reform, Trade Liber-alization and Their Impacts on China’s Agriculture.” Paperpresented at International Agricultural Trade Research Con-sortium Symposium on “China’s Agricultural Trade Policy:Issues, Analysis and Global Consequences.” San Francisco,June 25–26.

Ianchovichina, E., and W. Martin. 2001. “Trade Liberalization inChina’s Accession to WTO.” Journal of Economic Integration16(4): 421–45.

Jin, Liqun. 2002. “China: One Year into the WTO Process.”Paper presented at World Bank, Washington, D.C., October22. Available at www.worldbank.org/wbi/B-SPAN/docs/IMF-WB_address_final.pdf.

Johnson, D. G. 2000. “The WTO and Agriculture in China.”China Economic Review 11: 402–4.

Kehoe, T. 2002. “How Well Did Computable General Equilib-rium Models Predict the Effects of NAFTA?” Paper pre-sented at Conference on General Equilibrium Modeling.Cowles Foundation, Yale University. April.

Lardy, N. 2002. Integrating China into the Global Economy.Washington, D.C.: Brookings Institution Press.

Langlois, J. 2001. “The WTO and China’s Financial System.”China Quarterly 167: 610–29.

Lin, J. Y. 2000. “WTO Accession and China’s Agriculture.” ChinaEconomic Review 11: 405–8.

Martin, W. 2001. “Implications of Reform and WTO Accessionfor China’s Agricultural Policies.” Economics of Transition9(3): 717–42.

Martin, W., and E. Ianchovichina. 2001. “Implications of China’sAccession to the WTO for China and the WTO.” World Econ-omy 24(9).

Martin, W., and V. Manole. 2003. “China’s Emergence as theWorkshop of the World.” Paper presented at Stanford Centerfor International Development Conference on ChineseEconomic Policy Reform. Stanford University, September18–20.

McKibbin, W., and K. Tang. 2000. “Trade and Financial Reformin China: Impacts on the World Economy.” World Economy23(8).

OECD (Organisation for Economic Co-operation and Develop-ment). 2002. China in the World Economy: The DomesticPolicy Challenges. Paris.

Panitchpakdi, S., and M. Clifford. 2002. China and the WTO:Changing China, Changing World Trade. Singapore: JohnWiley and Sons.

Schmidhuber, J. 2001. “Changes in China’s Agricultural TradePolicy Regime: Impacts on Agricultural Production, Con-sumption, Prices, and Trade.” In Organisation for EconomicCo-operation and Development, China’s Agriculture in theInternational Trading System, OECD Proceedings. April.

Shafaeddin, S. 2004. “Is China’s Accession to WTO ThreateningExports of Developing Countries?” China Economic Review,forthcoming.

Shi Xinzheng. 2002. “Empirical Research on Urban-RuralIncome Differentials: A Case of China.” Peking University.Processed.

Tuan, F., and G. Cheng. 1999. “A Review of China’s AgriculturalTrade Policy.” Paper presented at International AgriculturalTrade Research Consortium Symposium on China’s Agricul-tural Trade Policy: Issues, Analysis and Global Conse-quences. San Francisco, June 25–26.

Wen, J. 2004. “Report on the Work of the Government.” Reportdelivered at the second session of the National People’s Con-gress. Beijing, March 5.

Whalley, J. 2003. “Liberalization in China’s Key Service SectorsFollowing WTO Accession: Some Scenarios and Issues ofMeasurement” NBER Working Paper 10143, NationalBureau of Economic Research, Cambridge, Mass.

Woo, W. T. 2001. “Recent Claims of China’s Economic Excep-tionalism: Reflections Inspired by WTO Accession.” ChinaEconomic Review 12: 107–36.

World Bank. 2001. Global Economic Prospects 2002: Making TradeWork for the World’s Poor. Washington D.C: World Bank.

WTO (World Trade Organization). 2002. “Transitional Reviewunder Article 18 of the Protocol of Accession of the Peoples’Republic of China.” G/C/W/438. Geneva.

———. 2003a. “Report of the Meeting Held on 5 December2003.” Council for Trade in Services, S/C/M/69. Geneva.

———. 2003b. “Report of the Meeting held on 1 December2003. Council for Trade in Financial Services, S/FIN/M/43.Geneva.

Yu, W., and S. E. Frandsen. 2002. “China’s WTO Commitmentsin Agriculture: Does the Impact Depend on OECD Agricul-tural Policies?” Paper presented at Fifth Conference onGlobal Economic Analysis. Taipei, June 5–7.


Part I

Policy ReformsAssociated with

Accession

China would leave the GATT. Although the govern-ment in Beijing never recognized this withdrawaldecision, nearly 40 years later, in 1986, China noti-fied the GATT of its wish to resume its status as acontracting party and its willingness to renegotiatethe terms of its membership.

A working party to examine China’s statusestablished in March 1987 met for the first time inOctober of that year. The GATT Working Party onChina’s Status met on over 20 occasions but with-out conclusion. With the formation of the WTO in1995, the GATT Working Party was converted intoa WTO Working Party on the Accession of China.The Working Party, chaired by Ambassador Pierre-Louis Girard of Switzerland, met 18 times.

Article XII of the Marrakesh Agreement Estab-lishing the WTO (the “WTO Agreement”), whichgoverns accessions, is striking in its brevity. Theoperative provision reads: “Any State or separatecustoms territory . . . may accede to this Agree-ment, on terms to be agreed between it and theWTO.” By early April 2004, more than 20 newmembers had joined the WTO since its establish-ment, bringing membership to 146, includingthree separate customs territories (Chinese Taipei,1

Hong Kong [China], and Macau). Cambodia andNepal had been accepted for membership but

With the gavelling of the accession package at the conclusion of the Working Party meeting onSeptember 17, 2001, the negotiation on China’saccession to the World Trade Organization (WTO)finally was concluded. The WTO Ministerial Con-ference subsequently approved the terms of China’saccession in Doha, Qatar, on November 10, 2001,and the Chinese Government notified its accept-ance on November 11. In line with customary prac-tice and as set out in China’s Protocol of Accession,China formally became a Member of the WTO 30 days later on December 11, 2001.

Each accession to the WTO is a unique event,but China’s accession has been particularly note-worthy. China had been one of the 23 original con-tracting parties to the General Agreement on Tariffsand Trade (GATT) in 1948. However, the applica-tion process for readmission to the multilateraltrading system took 15 years from its submission inJuly 1986, making it easily the longest and mostarduous accession negotiation in the history of theGATT/WTO. A review of the application processcontributes to an understanding of the overallprocess and highlights some key features of China’sAccession agreement.

After China’s revolution in 1949, the govern-ment in Taiwan (China) announced in 1950 that

21

2

What China’s WTOAccession is

All About

Jeffrey L. Gertler

were still in the process of completing ratificationprocedures.

The final stages of the China accession processcan be classified rather naturally under threeheadings:

• conclusion of bilateral market-access negotia-tions

• conclusion of multilateral negotiations in theWorking Party and the corresponding documentsthat together stipulate the terms of China’saccession, including the draft Protocol and itsAnnexes, and the Working Party Report

• approval and acceptance of these terms ofaccession by WTO members and by China,respectively.

Before delving into a description of these threesteps, it is worth reviewing briefly the many ups anddowns China experienced in the course of theaccession process. I refer in particular to significantprogress made just prior to the Tiananmen “inci-dent,” followed by almost two-and-a-half years withvirtually no accession activity, China’s subsequentparticipation in the Uruguay Round negotiations,and its failure to conclude negotiations on its statusas a GATT-contracting party in time to become anoriginal member of the WTO. Next were the con-version of the GATT Working Party into a WTOaccession working party in December 1995, theconsiderable optimism about an accelerated pro-cess in early 1997, the near-conclusion of a bilateralagreement with the United States in April 1999, theU.S. bombing of the Chinese Embassy in Belgradethe following month, and the conclusion of a bilat-eral agreement with the United States in November1999. These developments were followed by a spateof additional bilateral agreements in the first half of2000, including that with the European Union inMay. Further hiccups occurred subsequently, in-cluding the downing of a U.S. spy plane over theStraits of Taiwan, the Bush administration’s aggres-sive enthusiasm for a new form of the StrategicDefense Initiative called the Missile Defense System,and finally, the terrorist attack in the United Statesand the U.S.-led retaliatory actions since that time.Clearly, China and WTO members have been on aroller-coaster ride of major proportions over muchof the accession period.

BILATERAL NEGOTIATIONS

China’s importance as a trading nation generatedconsiderable enthusiasm for its participation in theaccession negotiations. Some 44 WTO members,including the 15 member States of the EuropeanUnion as one entity, expressed interest in conclud-ing bilateral market-access negotiations with China.

As these bilateral deals were struck and notifiedto the WTO, China’s consolidated Schedule of Con-cessions and Commitments on Goods and its con-solidated Schedule of Specific Commitments onServices were prepared with assistance from theWTO Secretariat. Thereafter, they were reviewedin the Working Party and multilateralized, thatis, extended on a most-favoured-nation (MFN)basis to all WTO members, as China’s Goods andServices Schedules, annexed to the Protocol ofAccession. In this way, these bilateral commitmentsbecame part of the multilateral treaty terms ofChina’s membership in the WTO.

China was only able to make rapid progress inconcluding its bilateral negotiations with mostother WTO member governments once it reachedbilateral agreement with the United States in Novem-ber 1999 and then with the European Communitiesin May 2000. Thereafter the negotiations proceededapace.

Only in the final days of the Working Party wasChina able to conclude negotiations with Mexico,the last of the 44 WTO members seeking bilateralmarket-access commitments. The sticking pointhad been the hundreds of antidumping orders thatMexico continued to maintain against products ofChinese origin. Mexico eventually agreed to termi-nate these allegedly WTO-inconsistent measuressix years after China’s accession. In addition,El Salvador, which recognizes Taiwan (China) andnot China (as is true of other WTO members),did not request bilateral negotiations with Chinaand invoked the non-application provision ofArticle XIII of the WTO Agreement against China.

One element contributing to pressure on Chinaand Members to conclude bilateral accords was theU.S. administration’s agreement with China. As aquid pro quo for China’s market-access concessions,the United States would provide China with per-manent MFN status, thereby eliminating the annu-ally renewed conditional MFN provided under theJackson-Vanik Amendment to the U.S. trade act of


1974. After much debate, the U.S. Congress finallypassed unconditional MFN status (what it callspermanent normal trade relations or PNTR) forChina in September 2000.

MULTILATERAL STEPS

With the bilateral market-access negotiations near-ing completion, members and China showedrenewed interest in wrapping up the many out-standing multilateral elements of the accessionpackage. In order to finalize the negotiated packageand in recognition of the fact that much of theinformation China had submitted to the WorkingParty was incomplete or out of date, the WorkingParty requested that China submit updatedinformation—including notifications of laws, reg-ulations, and other policy measures—on all keyaspects of China’s trade regime.

Clearly, such information was indispensable toassessing the congruency of China’s trade regimewith WTO rules. It was also indispensable to final-izing negotiations on key provisions of the Protocoland the Report. Identifying the trouble spots andagreeing on the timing—including possible tran-sition periods—for China to bring any WTO-inconsistent policy measures into compliance withWTO obligations, presented major challenges.

The final meeting of the Working Party inSeptember 2001, with informal sessions the weekof September 10 and the formal meeting onSeptember 17, was devoted to completing the tech-nical cleanup and verification of the Goods andServices Schedules, followed by an overall review ofthe documents to ensure consistency among thevarious elements of the accession package.

China and Working Party members finallyreached agreement on all outstanding issues:

• preambular and general provisions • the administration of the trade regime, includ-

ing uniform administration, special economicareas, transparency, and judicial review

• nondiscrimination, special trade arrangements,state trading, nontariff measures, tariff-ratequota administration, import and export licens-ing, price controls, taxes and charges levied on imports and exports, export subsidies anddomestic support in agriculture, and sanitaryand phytosanitary measures

• trading rights • standards and technical regulations.

Additional agreements were reached on thefollowing:

• a special transitional provision lasting 15 yearson price comparability in determining subsidiesand dumping

• the establishment of both a transitionalproduct-specific safeguard mechanism and aseparate transitional textile safeguard

• immediate implementation of the Agreementon Trade-Related Aspects of Intellectual Prop-erty Rights (TRIPS)

• a transitional review mechanism to overseecompliance with the terms of the Protocol

• a host of technical sectoral issues in trade inservices.

While China has reserved the right to exclusivestate trading for products such as cereals, tobacco,fuels, and minerals, and to maintain some restric-tions on transportation and distribution of goodsinside China, many of the restrictions that foreigncompanies currently face in China will be phasedout over a three-year transition period. During a12-year period after accession, WTO members willhave access to a transitional safeguard mechanismin cases where imports of products of Chinese ori-gin cause or threaten to cause market disruption tomember’s domestic producers.

Upon accession China became a party to theAgreement on Textiles and Clothing; accordingly,as for all WTO members, quotas on textiles will endon December 31, 2004. However, a special, negoti-ated, safeguard mechanism will remain in placeuntil the end of 2008, permitting WTO members to take action to curb imports in case of marketdisruption caused by Chinese exports of textileproducts.

Several areas of negotiation were problematicuntil very late in the process. First, there was lack ofagreement on the availability to China of WTOprovisions for developing countries in areas such asdomestic support to agriculture and industrial sub-sidies. The United States, in particular, objected toproviding China with the full benefit of developingcountry provisions in these areas. The matter even-tually was resolved; China agreed not to resort to

What China’s WTO Accession Is All About 23

certain of these WTO provisions and accepted a capof 8.5 percent on domestic support in agriculture,below the 10 percent available generally to develop-ing country members under the WTO Agreement.Because this 8.5 percent cap is still well above exist-ing budgetary outlays by the Chinese government,it is doubtful whether the cap will have a majorimpact on China’s ability to help its farming sec-tor adapt to the new and evolving conditions ofcompetition.

A second area that was among the last resolvedwas the regime relating to trading rights. The basicregime eventually agreed to calls for a fully liberal-ized right for foreign companies to gain tradingrights in China after a three-year transition period;state-traded products remain an exception.

A third area of considerable difficulty was agree-ment on China’s handling of its regime dealing withtechnical regulations and standards; the key issuewas how to ensure nondiscrimination (nationaltreatment) in the application of this regime. Afterrepeated urging spanning several years, Chinafinally committed to unifying its administrativestructure responsible for the inspection and con-formity assessment procedures (under the Admin-istration for Quality Supervision, Inspection andQuarantine [AQSIQ]) for both domestic andimported goods, thereby allaying many of the con-cerns raised by the dual and separate systems ofinspection that previously existed.

As part of the concluding phase, the WorkingParty also reviewed and obtained amendments andclarifications to many of the transitional Annexesof the Protocol. This was done on the basis ofupdated and revised drafts of these Annexesprovided by China. They included the followingannexes:

• Products subject to state trading • Products subject to designated trading • Nontariff measures subject to phased

elimination • Products and services subject to price controls • Notification and phase-out of subsidies • Export taxes and charges • Restrictions maintained against China • Issues to be addressed in the transitional review• the Schedule of Concessions and Commitments

on Goods, as well as the Schedule of SpecificCommitments on Services.

The Protocol and Working Party Report essen-tially contain a one-way set of commitments (fromChina’s side only), although these documents alsocontain some “soft” commitments by Members, forexample, regarding the non-abuse of domestic pro-cedures in anti-dumping actions and restraint inthe use of the special safeguard. Additionally, thereis an unusual Annex to the Protocol, containingcommitments by certain Members to phase outinconsistent measures maintained against Chinaover a transition period lasting up to five years.

APPROVAL AND ACCEPTANCE

Once consensus was achieved in the Working Partyon the final accession package, this was forwardedto the General Council for decision. Given thetiming of the Fourth Ministerial Conference inmid-November 2001, it was decided that these doc-uments should be forwarded to Doha for approvalby Ministers, rather than be decided upon at thelevel of the General Council in Geneva.

In accordance with established procedures, theMinisterial Conference approved the Decision onAccession and the Protocol on the terms of China’saccession on November 10, 2001. China acceptedthe WTO’s Protocol of Accession on November 11,2001, and became a Member of the WTO 30 dayslater. Notifying the Director-General that the Stand-ing Committee of the People’s Congress had rati-fied the terms of accession, China became the 143rd

member of the WTO on December 11, 2001.2

Article XII:2 of the WTO Agreement providesthat “the Ministerial Conference shall approve theagreement on the terms of accession by a two-thirds majority of the Members of the WTO.”However, pursuant to Article IX:1 of the same agree-ment and a 1995 decision of the General Council,all accession decisions are to be approved by con-sensus (with possible recourse to voting only whereconsensus is not achievable). All 16 WTO accessiondecisions to date, including China’s, have beentaken by consensus.

CHINA’S ACCESSION IN THE CONTEXT OF THE WTO’S BASIC PRINCIPLES

It could be interesting to consider briefly howChina’s accession fits within the context of the fivefundamental principles of the GATT and the WTO.


Simply put, these principles cover the following:

• Nondiscrimination • Market opening• Transparency and predictability • Undistorted trade• Preferential treatment for developing countries.

Non-Discrimination

Two types of nondiscrimination are of interest:the MFN principle and the national treatment prin-ciple. Under the MFN principle, a member may notdiscriminate between its trading partners: goodsand services and service providers are to be accordedMFN, that is, equal, treatment. At the same time, amember must provide national treatment: it maynot discriminate on its internal market between itsown and foreign products, services, and nationals.

Where do things stand in terms of China’s acces-sion vis-à-vis the principle of nondiscrimination?China, like other members, has committed itselfto abide by all WTO agreements, including thoseprovisions requiring application of MFN andnational treatment. In its Protocol of Accession,China has agreed to undertake additional commit-ments to ensure the smooth phasing in of these non-discrimination principles. Of particular note arecommitments to eliminate dual pricing practicesand to phase out within three years most of therestrictions on importing, exporting, and tradingcurrently faced by foreign enterprises. All foreignenterprises, including those not invested or regis-tered in China, are to be accorded treatment no lessfavorable than that accorded to enterprises in China.

Market Opening

The principle of market opening is promoted inthe WTO through successive rounds of multilateraltrade negotiations aimed at the progressive lower-ing of trade barriers. New members are pressed toliberalize their trade regimes during accession nego-tiations. Trade ministers also initiated the latestround of multilateral negotiations at the DohaMinisterial Meeting in November 2001.

With respect to market opening, China has sig-nificantly reduced its tariff and nontariff barriers aspart of its bid to join the WTO. The breadth anddepth of the cuts are evident. China’s willingness to

progressively and substantially open up its servicessectors to foreign competition also is undeniable.Moreover, China has already demonstrated its inten-tion to play a significant role in the Doha Develop-ment Agenda.

Transparency and Predictability

These are key elements of the multilateral tradingsystem. The basic transparency principle, containedin GATT Article X, calls on member governments topromptly publish all trade-related laws, regulations,judicial decisions, and administrative rulings ofgeneral application; to administer all such measuresin a uniform, impartial, and reasonable manner;and to provide for independent judicial review pro-cedures for the prompt review and correction ofadministrative actions. The predictability principleis ensured through a legal hierarchy giving prefer-ence to tariffs over less transparent and less securenon-tariff measures such as quotas and licences,and through encouraging members to “bind” theirmarket opening commitments in goods and serv-ices. In the goods area, this binding amounts tosetting ceilings on customs tariff rates.

China has committed to abide by the WTO’stransparency obligations across the board—including with respect to uniform application of itstrade regime and independent judicial review—and has made additional commitments in each ofthese areas. While difficulties may exist with respectto variations in treatment in different parts ofChina’s customs territory, as well as with the per-ceived lack of independence of the judiciary, therecan be little doubt that the Chinese government iscommitted to carrying through the necessaryreforms to implement these obligations in a uni-form and impartial manner. Also, as noted, China’saccession commitments will be the subject of aspecial transitional review mechanism for the first10 years of membership.

With China as a member, her producers andexporters will more confidently be able to makelong-term business decisions on the expansion oftheir activities.The more open the Chinese economybecomes, the more China will benefit from the legalsecurity of the rules-based trading system. Not justforeign investors, exporters, and importers, but alsoall Chinese citizens, will benefit from the more open,nondiscriminatory reforms China is undertaking.


In terms of producing a more predictable andsecure trading environment, China has bound allits import tariffs in the goods area. China also hascommitted to the phased reduction and removal oftariff barriers, mostly by 2004, but no later than2010. China’s average bound tariff level willdecrease to 15 percent for agricultural products,ranging from 0 to 65 percent, with the higher ratesapplied to cereals. For industrial goods, the averagebound tariff level will go down to 8.9 percent, witha range from 0 to 47 percent, with the highest ratesapplied to photographic film and automobiles andrelated products. In services, China has made a morecomprehensive set of initial commitments thanthose offered by most developed countries duringthe Uruguay Round. Of particular note are China’scommitments in services sectors covering telecom-munications, banking, and insurance.

Undistorted Trade

The WTO system also promotes undistortedtrade through the establishment of disciplines onsubsidies and dumping, allowing members torespond to unfair trade through the imposition ofcountervailing or antidumping duties. The treatyallows individual members to impose temporarysafeguard measures, under strict rules, when facedwith a sudden surge in imports causing seriousinjury to a domestic industry.

As in other areas, China has agreed to abide byall WTO disciplines relating to subsidies and coun-tervailing measures, protection provided underrules for antidumping, and safeguard measures.As noted above, it has also committed not to useexport subsidies on either industrial or agriculturalgoods and has accepted special provisions soughtby other members in relation to determinations ofdumping or subsidies, as well as a special product-specific safeguard mechanism and a separate textilesafeguard. China has indicated its intention to jointhe plurilateral Agreement on Government Pro-curement, which is aimed at ensuring fair competi-tion rules in purchases by government procurementagencies.

Preferential Treatment for Developing Countries

This principle permeates the entire WTO Agree-ment, providing transition periods to developing

countries and countries in transition to marketeconomies to adjust their systems to many ofthe new obligations resulting from the UruguayRound. A Ministerial Decision gives additionalflexibility to the least-developed countries in imple-menting the various Uruguay Round agreementsand calls on developed country Members to accel-erate their implementation of market accesscommitments on goods exported by the least-developed countries.

Although China has not been granted across-the-board preferential treatment as a developingcountry, it has negotiated specific transitionalarrangements in certain areas of its trade regime.Examples include the phasing out of quotas andimport licenses, and the phased liberalization ofthe right for foreign entities to trade in China. Incontrast, despite the availability of more preferen-tial treatment under the WTO agreements, Chinahas accepted a special cap on its ability to providedomestic production subsidies in agriculture, hasagreed not to use export subsidies, and has com-mitted to immediate implementation of the TRIPSAgreement.

INSTITUTIONAL IMPLICATIONS

No one can contest that China’s participation in theWTO will affect the operations of this organizationin substantial ways and over the long term. China isjoining as the seventh largest exporter and eighthlargest importer of merchandise trade and as thetwelfth leading exporter and tenth leading importerof commercial services. It has the largest popula-tion and largest potential market of any WTOmember.

Undoubtedly, China’s membership will haveimplications for the regular work of the WTO’smany committees administering the many agree-ments of this institution. China will surely be activein the newly launched and future rounds of multi-lateral trade negotiations, in agriculture, and inservices, but also in other areas of mutual concern.It will participate in fashioning the improved insti-tutional operations of the WTO. Clearly, China’smembership is likely to result in expanded recourseto the dispute settlement procedures of the WTO,both by China and by other members in relation toChina’s implementation of its WTO commitments.The first such case was that recently brought by


China, similar to that brought by many other WTOmembers, against the U.S. steel safeguard measures.We should, of course, expect to see some new facesin the Secretariat.

CONCLUDING REMARKS

Since the mid-1980s, the process of reform inChina has matured considerably, and China’s tradeperformance has reflected this maturation. Chinahas become a very important player in interna-tional trade, on both the import and the exportsides. Moving from an earlier phase of import plan-ning to one of import licensing and then, more sig-nificantly, to one of import tariffs, has broughtprices and the market mechanism into play as keydeterminants of China’s future trade relations withthe rest of the world.

Accession should allow China to lock-in theaccumulated benefits of the trade reform processthat the Chinese government has undertaken todate, and it should provide a platform from whichChina can sustain its reform process into the future.

By placing China’s reforms within the broadercontext of trade liberalization by all WTO mem-bers, Chinese producers and exporters can increasethe returns from trade reform in China throughreciprocal market access abroad and help theChinese government resist pressure domestically toreverse the process of reform.

For China, WTO accession will provide the 1.3 billion Chinese people with secure, predictable,and nondiscriminatory access to the markets of143 trading partners. It will also give this sameenormous population secure and nondiscrimina-tory access to the goods and services of these otherWTO members.

However, from China’s perspective, membershipalso for the first time commits this new WTOplayer at the international level to implement legaland domestic policy reform, ensuring much greatertransparency and security on a uniform basis.China has made impressive strides at reform overthe past 20 years. However, committing itself toabide by international treaty rules and the rule oflaw in the conduct of trade and in domestic policyreform is likely to take this process forward at aneven more impressive pace.

Accession will also mean that China can replacethe many risky and uncertain bilateral relationships

it has had to use until now, to shape its tradewith its major trading partners, with a single,multilateral trade relationship with the rest ofthe world.

Of course, what lies beyond China’s accession isthe major, continuing, and, in many ways, impon-derable task of implementation by China of itsWTO accession commitments. A question upper-most in the minds of many Chinese and foreignersalike is whether and how China will be able toensure uniform and impartial implementation oftrade commitments.

At this stage, it is difficult to predict the speedwith which WTO members and China will resort tothe WTO’s dispute settlement procedures. There islittle doubt, however, that China and its tradingpartners will eventually take full advantage of WTOmechanisms to resolve trade disputes. While con-siderable emphasis has been placed on the 10-yeartransitional review mechanism provided in China’sProtocol, this mechanism does not contain anyenforcement provisions.

It is hard to overstate the difficulties many sec-tors of Chinese society will face in the months andyears following accession. The impact on loss-making state industries, less developed agriculturalcommunities, and myriad government-financedprojects throughout the country will be dramatic.Moreover, the adjustment to new, more competi-tive market conditions, will, for many millions ofindividuals and families, mean unemployment andsignificant displacement. It will take many years forlarge segments of China to establish a new equilib-rium, during which time many citizens may wellface considerable hardships.

However, we should bear in mind that thisadjustment process has already begun. The Chinesepeople are hardly strangers to this process. Alreadyin the early 1990s China had introduced abankruptcy law and other legislation making stateindustries in principle responsible for their ownprofits and losses. For at least the past decade,China has radically reduced state subsidies andencouraged development of private enterprise inmany sectors. Since the mid-1990s in particular,and as a member of the International MonetaryFund, China has rationalized and liberalizedhandling of its foreign exchange market. In addi-tion, China has progressively and dramaticallyreduced its import tariffs and other nontariff


restrictions on foreign participation in the Chinesemarket.

The difficult adjustment is far from complete. Atthis stage, therefore, we can do little more than wishChina and its people “bon courage” as they venturedown the extremely challenging path that stretchesbefore them.

Notes

1. The nation of Taiwan (China) became a member of theWTO as Chinese Taipei.

2. Chinese Taipei, whose terms of accession were approvedby Ministers on 11 November 2001, became a WTO member on1 January 2002.


These include continued use of non-market econ-omy (NME) status in antidumping investigationsand the use of a special transitional product-specificsafeguard (TPS) provision. These two provisionsare scheduled to last 15 and 12 years, respectively.1

Second, China, by far the main target of existingantidumping measures, to date is one of the smallestusers of such measures. However, the past decadehas shown how quickly large developing countrieswilling to use antidumping rules can become inten-sive users of this instrument, and the evolution ofChina’s antidumping enforcement in 2002 and early2003 raises legitimate concerns in this respect.

The chapter is organized as follows. The first sec-tion describes the current situation. Antidumping isused massively by only 10 countries (four industrialand six developing), and a strong asymmetry, bestillustrated by China, exists between countriesenforcing antidumping measures and those targetedby antidumping measures. The second sectionexamines how China could minimize exposure toforeign antidumping cases—an option that wouldbe a recipe for both trade success and China’s lead-ing role in reforming WTO antidumping rules.The third section analyzes China’s antidumping

A few weeks prior to China’s acceptance as a fullmember of the World Trade Organization (WTO),chief negotiator Long Yongtu had ranked stricterantidumping rules second among China’s prioritiesin the WTO. Antidumping rules define the condi-tions under which a WTO member can, if it wishes,counter-balance dumping, that is, the fact that for-eign firms exporting to the markets of this WTOmember price their products more cheaply in thesemarkets than in their own domestic markets. At thattime, the United States was still fighting to excludeantidumping from the topics to be discussed atthe WTO Doha Ministerial, and the EuropeanCommunity (EC) was adopting an ambiguousposition.

In the early stages of the negotiations under theDoha Development Agenda (DDA), China findsitself in an unique situation on the antidumping andsafeguard issues. (Safeguard rules define the condi-tions under which a WTO member can, if it wishesso, give a transitional relief to a domestic industryfacing an unforeseen surge in competing foreignimports.) First, the WTO accession protocol ofChina includes very special provisions that China’strading partners may use against Chinese exports.

29

3

China in the WTO:Antidumping and

Safeguards

Patrick A. Messerlin

I would like to thank Mike Finger, Edwin Vermulst, Will Martin, and two anonymous referees for very usefulcomments on earlier drafts of this paper.

regulations and first cases, including their crucialrelations with the existing web of the U.S. and ECantidumping cases. The fourth section examines theopportunities that the DDA offers to China for nego-tiating stricter disciplines both on WTO contingentprotection, and on the use of the NME and TPS pro-visions by China’s trading partners. The conclusionsummarizes the crucial choices to be made by Chinawith respect to antidumping and safeguard policies.

The Current Situation

During the November 2001 Doha WTO MinisterialConference, antidumping was perceived as an issuepitting developing countries anxious to disciplinethe use of this instrument against the United States,which was (and still is) very reluctant to change itsown antidumping regulations. However, a closeexamination of the current situation suggests amuch more complex picture. This examination isbased on the antidumping measures in force at theend of each year of the period 1995–2002 that arenotified to the WTO Secretariat by members.2

Antidumping Users vs. Targeted Countries: A Key Asymmetry

Table 3.1, which presents the stock of antidumpingmeasures in force by antidumping users, illustrates

two main results. First, worldwide antidumpingenforcement is thus highly concentrated in lessthan a dozen countries. The top 10 antidumpingusers enforce 90 percent of the antidumping meas-ures notified in the WTO; they represent 70 per-cent of the world GDP and 50 percent of worldtrade.

Second, the situation prevailing during theUruguay Round—antidumping users were almostexclusively industrial countries—is no longer thecase. Six “new” antidumping-intensive users (all of them developing countries: Argentina, Brazil,India, Mexico, South Africa, and Turkey) havealmost caught up with the four major “old” users.These new users implement more than one-third ofthe total number of antidumping measures in forcein 2002, compared with less than one-fourth in1995. Meanwhile, the share of measures of the fourold users has declined from more than two-thirdsto half of the total number of antidumping meas-ures in force during the period. A last worrisomesign is that the rest of the developing countries,while still small users individually, have doubledtheir global share of measures in force during theobservation period.3

Table 3.2 presents the stock of antidumpingmeasures in force by targeted country for theperiod 1995–2002. It shows a marked asymmetry


TABLE 3.1 Top 10 Antidumping Users, Measures in Force, 1995–2001

Average AverageNumber of Measures in Forcea

Numberper AppliedUser Country 1995 1996 1997 1998 1999 2000 2001 2002 Importb Tariffc

U.S. 265 271 271 281 282 202 227 239 0.29 4.3EC 140 138 138 139 159 175 175 183 0.19 4.6Australia 78 46 40 49 39 44 59 38 0.77 5.8Canada 79 78 78 65 72 71 85 83 0.38 4.8

Mexico 93 92 81 86 80 77 61 55 0.72 12.6South Africa 12 29 42 56 87 96 94 80 1.81 15.0India 13 15 24 44 58 94 115 181 1.28 39.6Argentina 15 30 33 39 45 42 45 60 1.17 13.7Turkey 37 37 34 34 35 14 16 30 0.61 12.7Brazil 21 28 23 28 38 43 49 54 0.51 12.5All Others 50 59 84 102 122 117 97 — 0.04 —of which China — — 0 0 4 8 11 17 0.03 15.8

All countries 803 823 848 923 1017 975 1023 — 0.21 —

a. Antidumping duties and undertakings in force as of 31 December of the year.b. Per thousand of US$ of the 1997 imports of the user country.c. Average applied tariffs (WTO 2001).Sources: WTO Reports on Antidumping (G/ADP/N series), WTO trade data 2001; Author’s computations.

between antidumping users and targets. The top 10users are the targets of less than one-third of all themeasures in force, with a strong amplification ofthis gap in 2001 and 2002. In sum, antidumping isan instrument enforced by a few large countriesagainst the smaller economies of the rest of theworld—hence, the absence of incentives comingfrom the rest of the world that would induce theexisting antidumping-intensive users to restraintheir use of antidumping actions.

Combining tables 3.1 and 3.2 shows that for alltop 10 antidumping users, with the exception ofBrazil, the country’s interests hurt by foreign anti-dumping measures are smaller than those benefit-ing from domestic antidumping protection. This isbased on the well-known economic propositionthat analyzes protection more as a conflict betweendomestic forces (namely export interests versusimport-competing interests) than as a conflictbetween countries. To capture this aspect, onecan calculate the number of foreign antidumpingmeasures in force imposed on exports from a topuser adjusted by the size of the country’s exports(in thousands of U.S. dollars). Such a trade-adjusted number of measures (shown in table 3.2)mirrors the intensity of the foreign pressuresimposed on the export interests of a country, giving

an indication of the incentives of these exportinterests to contribute to the opening of the mar-kets of their country. These numbers can be use-fully compared to the symmetrical numbers ofantidumping measures in force by thousands ofU.S. dollars imported by the country in question(shown in table 3.1), which can be interpreted as anindication of the strength of the incentives of these import-competing interests to induce their own government to use antidumping. Theobserved imbalance between export interests andimport-competing antidumping beneficiaries inthe top 10 antidumping users suggests that it also isunlikely that domestic coalitions are strong enoughto support antidumping reforms in the WTO inusers that are key WTO players.

This situation raises a question that should becarefully examined in the future. Table 3.1 suggeststhat one could reasonably argue that antidumpingmeasures enforced by the six major developingcountry antidumping users impose welfare costson their own domestic economies higher than thecosts imposed on industrial economies by thedeveloped country–imposed antidumping meas-ures for two reasons. First is the marked differencebetween the number of measures imposed bydeveloping countries and industrial countries,

China in the WTO: Antidumping and Safeguards 31

TABLE 3.2 Major Antidumping Targets, Measures in Force, 1995–2001

Targeted Number of Measures Number perCountries 1995 1996 1997 1998 1999 2000 2001 2002a Exportb

U.S. 60 66 66 68 66 62 57 67 0.09EC 77 88 89 102 132 149 99 (98) 0.13Australia 5 6 5 6 5 5 5 5 0.08Canada 19 19 19 19 20 18 8 8 0.08

Mexico 11 15 17 17 19 21 17 17 0.15South Africa 7 10 11 11 12 15 16 24 0.39India 15 15 15 21 29 35 42 44 0.72Argentina 9 8 7 7 7 7 9 8 0.30Turkey 9 9 6 8 10 13 12 18 0.37Brazil 48 51 52 45 42 43 34 51 0.85All others 543 536 561 619 675 611 724 — 0.29of which China 143 148 180 193 202 207 199 212 0.99

All countries 803 823 848 923 1017 979 1023 — 0.22

a. Incomplete estimates for the EC.b. Per thousand of US$ of the 1997 exports of the targeted country.Sources: WTO Reports on Antidumping (G/ADP/N series); WTO trade data; Author’s computations.

once adjusted by trade size. The average numberof measures in force per thousands of U.S. dollarsof goods imported in 1997 by an antidumpinguser is a better indicator of the potential harmdone by antidumping on the domestic economythan the mere absolute number of measures. Thisindicator is much higher for developing countriesthan for industrial countries: it ranges from 0.5(Brazil) to 1.8 (South Africa), whereas it rangesfrom 0.2 to 0.4 for industrial countries (with theexception of Australia). These differences would bemuch larger if the number of antidumping meas-ures to be adjusted were to take into account thenumber of tariff lines concerned, since developingcountries tend to cover many more tariff itemswith antidumping cases than industrial countries.The second reason for arguably higher welfarecosts from antidumping measures in major devel-oping country users is that available (though notsystematic) information suggests that antidump-ing duties enforced by developing countries are,on average, more severe than those imposed byindustrial countries—and economic analysisshows that welfare costs increase more rapidlythan tariffs.

China’s Special Situation

Table 3.2 shows that China has been the main targetof all the antidumping measures enforced in theworld—18 percent of the antidumping cases in1995, almost 20 percent in 2001–2002. Table 3.3refines this information by presenting the unad-justed number of antidumping measures imposedon Chinese exports by the top 10 antidumpingusers, and this number once adjusted by the tradevalue between each trade partner and China (inother words, the average number of cases imposedby foreign antidumping users per thousand U.S.dollars of exports from China to these users). Bothsets of numbers show that China is much moretargeted by the developing countries than by theindustrial countries—especially when one looks atthe figures adjusted for trade size. In fact, China isalmost exclusively targeted by the top users, all ofthem relatively large economies.

The above information on antidumping usersversus targets raises a key question about China’srole in future world antidumping activities. WillChina follow the same evolution as the other largedeveloping countries, that is, will it increase rapidly


TABLE 3.3 Shares of Antidumping Measures in Force against Imports from China,1995–2002

Average

Shares of Antidumping Measures Against Chinaa Average Number perShare Export from

User Countries 1995 1996 1997 1998 1999 2000 2001 2002 (excl. 2002) Chinaa

U.S. 12.8 13.7 15.1 14.6 14.5 16.5 18.5 18.0 15.1 1.72EC 20.7 21.7 23.2 23.0 20.8 19.4 19.4 15.8 21.2 1.04Australia 9.0 4.3 10.0 8.2 7.7 4.5 5.1 7.9 7.0 0.71Canada 7.6 7.7 10.3 10.8 8.3 8.5 10.6 10.8 9.1 1.83

Mexico 33.3 28.3 40.7 38.4 36.3 35.1 44.3 41.8 36.6 60.27South Africa 8.3 27.6 28.6 23.2 19.5 18.8 19.1 22.5 20.7 11.98India 38.5 46.7 33.3 27.3 32.8 22.3 25.2 27.1 32.3 10.66Argentina 33.3 20.0 30.3 35.9 31.1 21.4 15.6 20.0 26.8 9.98Turkey 13.5 13.5 14.7 17.6 17.1 14.3 25.0 40.0 16.5 4.37Brazil 14.3 14.3 21.7 28.6 28.9 25.6 22.4 20.4 22.3 4.67All Others 34.0 28.8 26.2 22.5 18.9 12.8 15.5 — 22.7 —

Total (number)b 143 148 180 193 202 179 199 — 174 —in % of total 17.8 18.0 21.2 20.9 19.9 18.3 19.5 — 19.4 —

a. in percent of the total number of antidumping measures in force imposed by the user country.b. Per thousands of US dollars of the 1997 imports from China of the user country.c. provisionally based on Lindsey and Ikenson 2001.Sources: WTO Reports on Antidumping (G/ADP/N series). WTO trade data; Author’s computations.

the number of antidumping cases? Or will Chinaadopt a different approach, that is, will Chinaminimize its own antidumping use and investits negotiating strength in the WTO to obtain stricter antidumping rules, as claimed by its chieftrade negotiator? Clearly, China’s choice betweenthese alternatives will have a decisive impact on theevolution of world antidumping enforcement andon WTO trade disciplines.

Minimizing China’s Exposure to Foreign Antidumping

China’s second alternative will be a difficult exer-cise, particularly in the coming years. The usualslowness of WTO negotiations means that Chinacould not get reforms of WTO antidumping rulesbefore two (and more likely four) years. In theinterim, it will be hard for Chinese authorities to resist pressures from import-competing firmsbased in China that demand intensive use of China’santidumping procedures. This difficulty raises thefollowing question: could China adopt policiesminimizing, as quickly as possible, its exposure toforeign antidumping measures—thereby alleviat-ing the political costs of playing a reforming role inWTO antidumping rules?

Trade problems are fundamentally domesticconflicts between export-oriented and import-competing industries. Country-based data, as thoseprovided above, do not permit examination of thisdeeper aspect of protection. They can even be mis-leading. For instance, they can suggest that thestrong asymmetry between antidumping users andtargets observed in tables 3.1 and 3.2 wouldprogressively disappear because each WTO mem-ber would be induced to introduce antidumpingregulations and to enforce them—with such aworsening of the situation ultimately inducingWTO members to adopt collectively stricter disci-plines. (Arguments have already been presentedsuggesting that the top antidumping users will be unlikely to follow this path.) Taking into accountfirms’ behavior suggests that an even darkerscenario is quite plausible. Firms that are thepetitioners—hence the driving forces in antidump-ing enforcement—may well want to lodge similarantidumping complaints in several key countries tosegment the world markets of their productsthrough antidumping measures; evidence support-

ing this hypothesis is shown below. In such a case,the worldwide spread of antidumping regulationsand the increase of antidumping measures areseen as a positive development by the firms in ques-tion, rather than as incentives to discipline anti-dumping use. Examining such deeper aspects ofantidumping protection requires looking at thedistribution of antidumping measures in force bysector or product rather than by country.

A Few “Antidumping-Intensive” Sectors

Table 3.4 provides a broad picture based on thenumber and average shares of antidumping meas-ures in force by the most aggregated section of theHarmonized Tariff System (instead of specific prod-ucts, as below). It shows that antidumping measuresare concentrated in a handful of Harmonized TariffSystem (HTS) sections. Antidumping measures inmetals, chemicals, machinery and electrical equip-ment, textiles and clothing, and plastics and rubbersamount to 75 percent of the total number of meas-ures, whereas trade in these sectors amounts to lessthan half of total world trade. These few HTS sec-tions happen to be key sources of exports fordynamic developing countries in the first stages oftheir industrial development. Second, these HTSsections are characterized by a high proportion ofrelatively standard products and by oligopolisticmarket structures. The metals and chemicals sec-tions clearly fit these features. The other HTS sectorsrequire a closer look at subsectors. Few antidumpingactions in the machinery and clothing subsectors arecharacterized by many firms and highly differenti-ated products, and most of the antidumping actionsin the electrical equipment and textiles subsectorsare characterized by relatively standard productsand by oligopolistic firms. These features explain thewide use of antidumping measures in the textilesubsector since the late 1980s, before any effectiveliberalization of the Multi-Fibre Arrangement quo-tas, as decided by the Uruguay Round (GATT, 1994).

Such a pattern strongly suggests that complain-ing firms use antidumping as an additional—cheapand powerful—instrument for segmenting themarkets that ongoing or scheduled trade liberaliza-tions aim at making more competitive. It also sug-gests that the spread of antidumping cases tocertain sectors, such as clothing, should be expectedto the extent that these markets will be increasingly



TAB

LE 3

.4A

nti

du

mp

ing

Mea

sure

s in

Fo

rce

in t

he

Wo

rld

by

Sect

or,

1995–2

000

Num

ber

of A

nti-D

umpi

ng M

easu

res

AD

Ms

Chi

na’s

Patt

ern

AD

Ms

Ave

rag

eTa

riff

Lin

esb

y H

SG

row

thEx

po

rtIm

po

rtA

pp

lied

wit

h T

arif

fsH

arm

on

ized

Sys

tem

Sect

ion

s19

9519

9619

9719

9819

9920

00Se

ctio

nIn

dex

esa

Patt

ern

Patt

ern

Tari

ffb

<10%

c

15M

etal

s, m

etal

art

icle

s27

629

030

531

637

240

133

.511

7.0

6.7

9.2

9.6

68.6

6C

hem

ical

, alli

ed in

dust

ry p

rodu

cts

201

203

204

210

217

228

21.6

104.

14.

78.

09.

478

.216

Mac

hine

ry, e

lect

rical

eq

uip

.81

8594

9510

081

9.2

116.

129

.237

.814

.430

.011

Text

iles,

clo

thin

g48

6061

5470

796.

411

4.2

19.8

7.4

21.0

12.6

7Pl

aste

rs, r

ubbe

rs53

5455

5366

746.

110

8.4

3.2

7.3

15.9

18.0

17Ve

hicl

es, t

rans

por

t eq

uip

men

t42

4040

4243

304.

010

1.6

3.7

2.8

23.8

43.4

13A

rtic

les

of s

tone

, pla

ster

, cem

ent

3437

3742

4440

4.0

115.

51.

60.

817

.712

.110

Woo

dpul

p, p

aper

, pap

erbo

ard

1817

2144

4439

3.1

207.

60.

73.

113

.426

.14

Prep

ared

food

stuf

fs35

3129

3128

273.

188

.92.

10.

828

.415

.72

Vege

tabl

e p

rodu

cts

1519

1924

2322

2.1

129.

42.

11.

921

.329

.212

Foot

wea

r, he

adw

ear

2116

917

1316

1.6

70.3

4.8

0.2

23.2

0.0

18O

ptic

al, c

inem

a in

stru

men

ts14

1110

920

231.

510

4.0

3.4

3.6

14.3

30.1

20M

isce

llane

ous

man

ufac

ture

d ar

ticle

s14

1514

1312

141.

489

.77.

10.

420

.21.

29

Woo

d, a

rtic

les

of w

ood

66

65

1111

0.8

122.

21.

11.

79.

861

.91

Live

ani

mal

s, a

nim

al p

rodu

cts

43

46

77

0.5

161.

91.

71.

018

.921

.53

Ani

mal

, veg

etab

le fa

ts, o

ils9

99

21

10.

544

.40.

10.

540

.65.

95

Min

eral

pro

duct

s0

00

99

90.

5—

3.7

10.9

4.1

96.3

19A

rms,

am

mun

ition

s3

33

30

00.

266

.70.

00.

013

.00.

08

Raw

hid

es, s

kins

, lea

ther

pro

duct

s0

01

11

10.

1—

3.0

1.4

16.4

25.0

14G

ems,

jew

elle

ry0

00

00

00.

0—

1.0

0.5

12.6

67.2

21A

rt o

bjec

ts0

00

00

00.

0—

0.2

0.8

10.9

25.0

Tota

l87

489

992

197

610

8111

0310

0.0

112.

010

0.0

100.

015

.838

.7

a. G

row

th b

etw

een

1998

–200

0 an

d 19

97–1

995.

b. i

n 20

01, c

. in

per

cent

of t

otal

tar

iff li

nes,

by

sect

or.

Sour

ce:L

inds

ey a

nd Ik

enso

n 20

01. I

n co

ntra

st w

ith t

able

s 3.

1 to

3.3

, AD

Ms

imp

osed

on

imp

orts

of t

he s

ame

pro

duct

by

vario

us E

C m

embe

r st

ates

are

not

syst

emat

ical

ly a

ggre

gate

d in

one

EC

cas

e.

subjected to product differentiation and imperfectcompetition (based on such factors as trademarks,goodwill, and distribution channels). In sum, theobserved sectoral pattern of antidumping reflectsthe increasing privatization of trade policy by firmsthat enjoy enough initial oligopolistic power tofully use the pro-collusive bias embedded inantidumping regulations—a key lesson that shouldbe kept in mind when implementing these regula-tions, in China as elsewhere.

China’s Sensitivity to Antidumping-IntensiveSectors

Does China tend to import products pertaining toantidumping-intensive HTS sections (that is, toindustries subjected to many antidumping meas-ures in the world)? If that is the case, one shouldexpect Chinese authorities to be under strong pres-sures for taking and enforcing antidumping meas-ures on imports of such goods—that is, forparticipating to the worldwide segmentation gamein this sector that constitutes the ultimate goal ofantidumping enforcement. Table 3.4 shows that thefive most antidumping-intensive HTS sections rep-resent almost 70 percent of total Chinese imports,opening the possibility that Chinese firms or foreignfirms producing in China could table antidumpingcomplaints to segment world markets.

China’s sensitivity to antidumping-sensitivesectors could also be assessed from an export per-spective. Table 3.4 underlines how sensitive Chineseexports are to the worldwide antidumping activity,particularly in machinery-electrical equipment andin textiles and clothing. The above observations onthe key role of economic characteristics (type ofproduct, market structure) as the dominant factorsdriving world antidumping activity suggest Chinaas a target of foreign antidumping measures muchmore because of the intrinsic economic features ofher exported products (differentiation level andoligopolistic markets) than because it was not aWTO member.

Chinese Policies as the Best Ways to MinimizeExposure to Foreign Antidumping

A key corollary of this conclusion is that Chinashould not expect to face less antidumping meas-ures in the coming years because of her WTO acces-

sion but rather should expect to continue to face alarge number of antidumping cases. Of course,WTO membership gives China access to the WTOdispute settlement mechanism, which could pro-vide some relief to Chinese exporters harassed byforeign antidumping. However, this relief probablywill be only marginal and short-lived. For instance,the 2001 WTO dispute settlement ruling banningthe averaging method is already in the process ofbeing circumvented by alternative procedures(hastily developed by creative petitioners) that mayhurt the defendants differently, but not necessarilyless than the condemned averaging method.

A more promising route for minimizing China’sexposure to foreign antidumping measures may beChinese policies—and pre-eminently China’s owntrade policy. If China still imposes high tariffs inantidumping-intensive sectors at the end of theaccession period, it may not adopt many or severeantidumping measures on imports from the rest ofthe world in these products (though section 1shows that the top developing country antidump-ing users have not hesitated to add high antidump-ing measures on the top of unfinished tradeliberalization). However, such a costly tariff policyfor China will not ensure less foreign antidumpingcases against Chinese exports. On the contrary,high Chinese tariffs will facilitate the introductionof foreign antidumping measures against Chineseexports in antidumping-intensive sectors. This isbecause high Chinese tariffs will allow the existenceof high prices in China, making the existence ofdumping by Chinese exporters easier to prove, allthe more because the Chinese exports in questionconsist mostly of basic products, for which Chineseexporters are likely to align their prices to thoseprevailing in foreign markets since there is minimalor no premium for differentiation.

Accordingly, a first way to minimize exposure toforeign antidumping measures would be the adop-tion by China of a relatively uniform and moderatetariff policy to reduce distortions in the domesticproduction pattern, foreign antidumping investi-gators interpret such distortions as sources ofdumping as much as possible. Table 3.4 provides theaverage applied tariffs by HTS section that Chinaimposed in 2001. Though averaging tariffs for suchwidely defined sectors gives very imperfect infor-mation on the tariff structure (peak tariffs withineach sector are eroded by low tariffs), it suggests that


tariff peaks in China are concentrated in HTS sec-tions that are not antidumping-intensive activities,with the exception of textiles and clothing.

Implementing a uniform tariff policy may help China to shift export products away fromantidumping-intensive sectors for another reasonalready observed in certain countries, such as HongKong (China), Japan, the Republic of Korea,Singapore, and Taipei. These five industrial coun-tries have reduced their exposure to foreignantidumping measures by upgrading theirexported products—to the point that one couldargue that, in fact, foreign antidumping measuresmay have accelerated the economic development ofthese industrial countries by inducing them to shiftfaster to anticipated comparative advantages inhighly differentiated products.

EC and U.S. Antidumping “Echoing” against ChinaExports

Table 3.4 relies on sectors too broad for capturingantidumping protection as embedded in the productstrategies of a few large firms. This aspect requiresinformation at the product level. Table 3.5 providessuch detailed information for the EC and U.S.antidumping actions against China. More precisely,Table 3.5 provides the list of the EC and U.S. “echo-ing” cases, that is, those cases that have targeted thesame products exported by China between 1980 and1999. However, the European Community and theUnited States are the top antidumping users in termsof the absolute number of measures in force, andthey represent the two largest markets for Chineseexports (15 and 21 percent, respectively).

Table 3.5 provides two general observations:

• First, the number of echoing cases is large: 58cases, that is, 75 and 68 percent of the totalnumber of antidumping cases initiated againstChinese exports by the United States and theEuropean Community, respectively. All of thesecases echo each other generally within a year orless, though of course they subsequently mayfollow different developments in these markets.All but three (cycles, hammers, and pocketlighters) resulted in antidumping measures ofsome kind. Such a large proportion of echoingcases and the similarity of their outcomes con-stitute signs that antidumping is a protectionist

instrument that petitioners are successfullyusing in a strategic way to segment the twolargest world markets. This observation leads toan important corollary. Assessing the welfarecosts of antidumping measures should takeinto account not only the severity of the meas-ures per se (the high level of antidumpingduties, or the restrictiveness of quantitativerestrictions), but also the fact that the levelof competition prevailing in the importingmarkets is dramatically reduced by suchmeasures. “Indirect” welfare costs generated byantidumping-caused collusion compound thedirect welfare costs generated by the antidump-ing duties.

• Second, U.S. antidumping duties are, on average,higher than EC duties: 104 percent and 38 per-cent, respectively. However, comparing in detailthe measures in the EC and U.S. echoing cases isdifficult because of the differences in regulationsand the lack of sufficiently detailed information.For instance, some EC cases are terminated bythe withdrawal of the complaint by the petition-ers. The effective impact of such withdrawals ishard to assess. It may be limited to a “chilling”effect on Chinese exporters who may be forcedto export less, or at higher prices, or both,to limit the risks of facing new antidumpingcomplaints. However, withdrawals may merelyreflect a lack of cooperation of the domesticindustry or a failure to get the minimum num-ber of EC member states to support an EC Com-mission proposal to impose measures. They alsomay reflect the fact that petitioners have beenable to impose quantity or price restraints onChinese exports on a “private” basis—with thecorresponding full-fledged impact to beexpected from these hidden restraints.

China’s Antidumping Enforcement:At the Crossroads

The following first assessment of China’s anti-dumping regulations and enforcement proceduresshould be taken as provisional because it is con-strained by the limited number of ongoing cases todate; a substantial number of cases is often neededfor a robust assessment of the effective antidump-ing enforcement by a country.



TABLE 3.5 EC and U.S. Echoing Antidumping Cases against Chinese Producers,1980–99

DumpingDecisions Antidumping

Initiating Margin Positive Negative Duties

Years Country Products (EC) EC U.S. EC U.S. EC U.S.

1992 EC antimony trioxide 43.2 noi

1991 U.S. antimony trioxide N

1988 EC barium chloride 50.1 D 25.8

1982 EC barium chloride 75.0 Ds

1983 U.S. barium chloride N

1983 U.S. barium chloride A 14.5

1999 EC brushes, hair

1992 U.S. brushes, hair T

1986 EC brushes, paint 100.0 Und

1992 EC brushes, paint noi

1985 U.S. brushes, paint A 127.1

1999 U.S. brushes, paint

1994 EC coumarin 50.0 Ds

1994 U.S. coumarin A 160.8

1991 EC cycles 30.6 D 30.6

1995 U.S. cycles N

1984 EC cycles, chains 45.0 Und 45.0

1999 EC cycles, forks

1999 EC cycles, frames

1996 EC cycles, parts 30.6 D 30.6

1999 EC cycles, wheels

1992 EC ferrosilicon 49.7 D 49.7

1992 U.S. ferrosilicon A 137.7

1995 EC furfuryl alcohol

1994 U.S. furfuryl alcohol A 45.3

1999 EC glycine

1994 U.S. glycine A 155.9

1985 EC handtools: hammers

1990 U.S. handtools: hammers A 45.4

1994 U.S. lighters, disposable N

1990 EC lighters, pocket 16.9 D 16.9

1991 EC magnesium oxide 27.7 Und

1994 U.S. magnesium, alloy A 79.4

1994 U.S. magnesium, pure A 108.3

1997 EC magnesium, unwrought 31.7 D�Und 31.7

1994 U.S. manganese, metal A 143.3

1992 EC manganese, unwrought

1996 U.S. persulfates

1994 EC persulfates, peroxodisulfates 110.1 D 83.3

1982 U.S. polyester, cotton cloth A 36.2

1990 EC polyester, yarn 23.5 D 23.5

1986 EC potassium permanganate 94.5 D�Und 28.0

1983 U.S. potassium permanganate A 39.6

1984 EC silicon carbide 31.5 Und

1993 U.S. silicon carbide N

1989 EC silicon, metal 38.7 Ds 18.7

1991 EC silicon, metal 178.0 Ds

1990 U.S. silicon, metal A 139.5

China’s Antidumping Regulations

China adopted its first antidumping regulations onMarch 25, 1997, and the antidumping guidelinesnecessary for implementing the law later the sameyear. Following accession to the WTO, this “old”regulation was replaced by the “new” regulation inJanuary 2002 with another set of guidelines (see,on the WTO website, G/ADP/N/1/CHN1 andG/ADP/ N/1/CHN2; see also Wang 2003 for a legalanalysis).

Chinese regulations follow the usual structure ofantidumping legislation, including proof of theexistence and estimate of the magnitude of dump-ing and of material injury, and proof of the causalrelation between dumping and injury. However,they have a few striking features. First, many detailsare left to the detailed guidelines or to case-by-casepractices. This is often the case among countriesthat have recently adopted antidumping regula-tions. However, it has the great inconvenience ofgenerating a high level of legal uncertainty in theprocess as a whole.

The second feature is that all the well-knownprotectionist biases of the WTO antidumping pro-visions are included in China’s regulations, asshown by the following non-exhaustive list:

• use of the concept of “a major proportion” ofthe industry as the threshold level for accepting

complaints (a condition that domestic monopo-lies, oligopolies, or cartels fit much more easilythan do competitive industries)

• possibility of ex officio initiation of cases by theChinese authorities

• screening of the complaints by the antidumpingoffice, exposing this office to strong and hiddenpressures by vested interests

• possibility of withdrawal by the petitioners,facilitating private collusion between them andthe defendants

• cumulation of imports, facilitating the demon-stration of injury and widening the geographicalscope of protective measures

• recourse to constructed normal values, enablingmanipulation of costs and reasonable profit incase of the absence of comparable prices in theexporting country

• broad definition of the confidentiality of theinformation limiting the rights of the defen-dants (and arguably having equal impacts ondomestic and export firms)

• possibility of imposing undertakings as anti-dumping measures and the mandatory provi-sion that antidumping duties shall be borne bythe importers (the no-absorption provision)

• possibility of imposing retroactive antidumpingduties in the case of a history of dumping, thatis, recurrent antidumping complaints


TABLE 3.5 (Continued)

DumpingDecisions Antidumping

Initiating Margin Positive Negative Duties

Years Country Products (EC) EC U.S. EC U.S. EC U.S.

1994 EC steel, pipe or tubes fittings 58.6 D 58.6

1999 EC steel, pipe or tubes fittings 49.4 D 49.4

1991 U.S. steel, pipes A 182.9

1985 U.S. steel, pipes N

1988 EC tungstate, ammon. para- 75.7 noi

1988 EC tungsten, carbide 73.1 D�Und

1988 EC tungsten, metal powder noi

1991 U.S. tungsten, ore A 151.0

1989 EC tungsten, ores 50.3 D�Und

1988 EC tungstic, oxide & acid 85.8 D�Und

Total 58 Average margins and duties 59.2 29 15 5 7 37.8 104.5

Sources: Bloningen 2001 for the U.S.; Official Journal of the EC.

Notes: EC decisions: D � ad valorem duty; Ds � specific duty; Und � undertaking; with � withdrawal; noi � no injury.ep � expired deadline.

U.S. decisions: A � affirmative; N � negative.

• possibility of taking “appropriate” measures incase of circumvention of the antidumpingmeasures by the foreign firms.

The third feature of the Chinese antidumpingregulations prior to the 2003 government reformwas a highly complex coordination of the variousadministrative agencies involved in antidumpinginvestigations. The Fair Trade Administration for Import and Export (FTA) of the Ministry of Foreign Trade and Economic Cooperation(MOFTEC) received complaints, decided those itwould accept or reject, and was involved in thewhole investigation and process. The State Eco-nomic and Trade Commission (SETC) joinedMOFTEC for determining the existence of injury atthe preliminary stage and for the final investiga-tions. The Customs General Administration (CGA)joined MOFTEC for some parts of the investiga-tions. For the imposition of antidumping duties,MOFTEC made proposals to the Tariff Commis-sion under the State Council (TCSC), which tookthe decisions. The merger of SETC and MOFTECin early 2003 brought these two functions into oneministry.

Lastly, the Old Regulation included the follow-ing Article 40: “In the event that any country orregion applies discriminatory antidumping orcountervailing measures against the exports fromthe People’s Republic of China, the People’sRepublic of China may, as the case may be, takecounter-measures against the country or region inquestion.” It is not known whether this provisionhas led to cases, but clearly it opened the possibil-ity for China to use her antidumping rules as aretaliatory instrument. Interestingly, Article 56 ofthe New Regulation is only slightly more diplo-matic: “Where a country (region) discriminatorilyimposes antidumping measures on the exportsfrom the People’s Republic of China, China may, onthe basis of the actual situations, take correspon-ding measures against that country (region).”

Antidumping Enforcement by China

Table 3.6 provides the list of the antidumping casesinitiated by China between 1997 and May 2003,based on information provided by MOFTEC forthe cases under the old regulation and on China’slatest notification (G/ADP/N/105/CHN) to the

WTO for the cases under the new regulation.Sixty-nine cases have been initiated, with a veryuneven time pattern: after a slow start in 1997–98,the number of cases has increased rapidly (somecases under the Old Regulation are left withunknown status), reaching 24 cases in 2002 and 11cases for the first five months of 2003. This increasecan only be a source of serious concern. It remainsto be seen whether it simply mirrors cases thatwere “in the pipeline” for a long time, or whetherChina has begun to follow the drift observed in thesix top developing country antidumping users andwill soon become another antidumping intensiveuser, putting in danger her so-far successfulliberalization.

Although it is too early to know whetherChina’s first antidumping measures will be repre-sentative of future antidumping enforcement,these measures deserve a few observations. First isthat antidumping measures have been taken inalmost all the cases. This is a very high percentage,compared to generally observed levels of 60–70 per-cent in industrial countries. The second observa-tion is the relatively high level of the measuresadopted by the Chinese authorities, although itseems that the most recent measures are less severethan the ones taken under the Old Regulation.The main countries targeted are industrial andadvanced developing countries—not quite the pat-tern observed for the other developing countryantidumping users.

The cases initiated since 2001 deserve an addi-tional remark. Their product pattern is closer towhat is observed for the other antidumping users,as best illustrated by the steel cases (echoing the ECand U.S. safeguards) and by the ethanolamine cases(observed in several other antidumping users). Thisincreasingly similar product pattern suggests thatChina’s antidumping enforcement may be startingto be part of the ongoing process of segmentingworld markets through worldwide antidumpingactivity. It also raises the issue of a progressive cap-ture of China’s trade policy by firms, similar towhat is observed in the 10 major antidumpingusers. In this context, it would be important toknow whether petitioners in China are Chinesefirms (private or state-owned) or whether they arefirms having strong links (such as joint ventures,technical relations, vertical integration) with for-eign firms well experienced in “antidumping art.”



TABLE 3.6 China’s Antidumping Measures in Force and Investigations, as of30 June 2003

Provisional DefinitiveAntidumping Antidumping

Initiation Duties (percent) Duties (percent)

Year Country Products Minimum Maximum Minimum Maximum

Cases initiated under the old regulation1997 Canada newsprint 17.1 78.9 9.0 78.01997 Korea, Rep. of newsprint 17.1 78.9 9.0 78.01997 U.S. newsprint 17.1 78.9 9.0 78.01999 Russia steel, cold-rolled silicon 11.0 73.0 0.6 62.01999 Korea, Rep. of polyester film 21.0 72.0 13.0 46.01999 Japan steel, cold-rolled stainless 4.0 75.0 17.0 58.01999 Korea, Rep. of steel, cold-rolled stainless 4.0 75.0 17.0 58.01999 EC Germany acrylates 24.0 74.0 0.0 0.01999 Japan acrylates 24.0 74.0 31.0 69.01999 U.S. acrylates 24.0 74.0 31.0 69.02000 EC Britain methylene chloride 7.0 39.0 6.0 39.02000 EC France methylene chloride 28.0 75.0 28.0 75.02000 EC Germany methylene chloride 67.0 67.0 66.0 66.02000 EC Netherlands methylene chloride 10.0 58.0 9.0 57.02000 Korea, Rep. of methylene chloride 7.0 28.0 4.0 28.02000 U.S. methylene chloride 49.0 58.0 49.0 58.02001 Japan polystyrene a a a a2001 Korea, Rep. of polystyrene a a a a2001 Thailande polystyrene a a a a2001 Indonesia lysine a a a a2001 Korea, Rep. of lysine a a a a2001 U.S. lysine a a a a2001 Korea, Rep. of polyester, chips a a a a

Cases initiated under the new regulation2001 Korea, Rep. of polyester, staple fiber 4.0 48.0 2.0 48.02001 Korea, Rep. of pet chips 6.0 52.0 5.0 52.02001 Indonesia acrylates 11.0 24.0 11.0 24.02001 Korea, Rep. of acrylates 11.0 20.0 2.0 20.02001 Malaysia acrylates 13.0 38.0 4.0 38.02001 Singapore acrylates 46.0 49.0 30.0 49.02001 EC Belgium caprolactam 6.0 16.0 6.0 16.02001 EC Germany caprolactam 28.0 38.0 28.0 38.02001 EC Netherlands caprolactam 9.0 18.0 6.0 18.02001 Japan caprolactam 5.0 21.0 5.0 18.02001 Russia caprolactam 6.0 29.0 7.0 16.02002 Korea, Rep. of polyester, film — — 0.0 0.02002 India anhydride, purified 33.0 33.0 b b2002 Japan anhydride, purified 66.0 66.0 b b2002 Korea, Rep. of anhydride, purified 14.0 33.0 b b2002 Japan styrene butadiene 0.0 33.0 b b2002 Korea, Rep. of styrene butadiene 10.0 27.0 b b2002 Russia styrene butadiene 16.0 46.0 b b2002 Kazakstan steel, cold-rolled products 21.0 48.0 b b2002 Korea, Rep. of steel, cold-rolled products 9.0 40.0 b b

China’s Options in the WTONegotiations on ContingentProtection

China’s antidumping enforcement needs to takeinto account an aspect that the six top developingcountry antidumping users have not had toaddress: China’s WTO accession protocol incorpo-rates specific provisions on antidumping and safe-guards (for a legal analysis, see Vermulst 2000). Thesection argues that this special feature that is ini-tially a handicap can be turned into a positiveinstrument. More precisely, it could induce Chinato negotiate, in the DDA, a more economically

sound “interpretation” of the specific provisions onantidumping. Such an interpretation could createstrong incentives in China for both restraining itsown antidumping use and fighting for stricterWTO rules on antidumping. It also could givestrong incentives to China’s trading partners to easeChina’s transitional period of accession. It is fair tosay that dealing with the provision on safeguardsseems more difficult. This provision is so much incontradiction with the general WTO rules andspirit that its use will raise a large systemic risk forthe whole WTO. However, various options areexamined at the end of the section.



Provisional DefinitiveAntidumping Antidumping

Initiation Duties (percent) Duties (percent)

Year Country Products Minimum Maximum Minimum Maximum

2002 Russia steel, cold-rolled products 9.0 29.0 b b2002 Taiwan (China) steel, cold-rolled products 8.0 55.0 b b2002 Ukraine steel, cold-rolled products 12.0 22.0 b b2002 Japan polyvinyl chloride 32.0 115.0 b b2002 Korea, Rep. of polyvinyl chloride 10.0 76.0 b b2002 Russia polyvinyl chloride 34.0 82.0 b b2002 Taiwan (China) polyvinyl chloride 10.0 27.0 b b2002 U.S. polyvinyl chloride 25.0 83.0 b b2002 Japan toluene 19.0 49.0 b b2002 Korea, Rep. of toluene 6.0 22.0 b b2002 U.S. toluene 23.0 28.0 b b2002 Japan phenol 7.0 144.0 b b2002 Korea, Rep. of phenol 10.0 10.0 b b2002 Taiwan (China) phenol 7.0 20.0 b b2002 U.S. phenol 29.0 29.0 b b2003 EC Germany ethanolamine b b b b2003 Iran ethanolamine b b b b2003 Japan ethanolamine b b b b2003 Malaysia ethanolamine b b b b2003 Mexico ethanolamine b b b b2003 Taiwan (China) ethanolamine b b b b2003 U.S. ethanolamine b b b b2003 EC chloroform b b b b2003 India chloroform b b b b2003 Korea, Rep. of chloroform b b b b2003 U.S. chloroform b b b b

69 All cases average antidumping 17.2 43.4 9.8 23.7duty

a. Status unknown; b. Ongoing investigations.Sources: MOFTEC, WTO Semi-annual report G/ADP/N/105/CHN, 22 August 2003.

Linking China’s Effective Liberalization to a BetterTreatment of Exports by Antidumping Users

China’s protocol of accession allows China’s tradingpartners to use the non-market economy (NME)status in their investigations against allegedlydumped Chinese exports. This NME status, whichis scheduled to last 15 years until 2017, allows for-eign antidumping investigators to use proxies forestimating the home market prices or costs ofChinese exporters. Such proxies make the proofof the existence of antidumping much easier than under the antidumping rules for marketeconomies, and they inflate the magnitude of theestimated antidumping margins compared tothose, already high, imposed on market economies.

Table 3.7 gives a sense of the intrinsic biases ofthe NME procedure by summarizing the informa-tion available from 208 EC and U.S. antidumpingcases initiated between 1995 and 1998 (Lindsey1999; Messerlin 2000a). It shows that the furtherfrom pure price comparisons the methodologyused for estimating dumping margins stays, thehigher these margins have been: from 3 percent(U.S.) and 22 percent (EC) under pure price com-parisons, to 14 percent (U.S.) and 24 percent (EC)under a mix of price comparisons and constructedvalues, and to 25 percent (U.S. and EC) under thevarious constructed value methods. Moreover, useof NME status is clearly linked to the highestdumping margins found (40 percent in the U.S.and 46 percent in the EC).4

It is almost impossible to eliminate a provisionincluded in a country’s protocol of accession.However, during the DDA, it may be possible to

negotiate an economically sound interpretation ofthe use of the NME status by WTO members. Thebenefits from such an interpretation rely on the factthat China and her trading partners have a com-mon interest in establishing the strongest possiblelink between Chinese effective liberalization andthe elimination of the NME use by foreignantidumping authorities against Chinese exporters.It is important to underline that what is at stake isnot the elimination of the NME provision itself(once again impossible to obtain) but its effectiveuse in the future.5

The argument aims at mobilizing the exportinterests in both China and the rest of the worldduring the implementation period of China’saccession to the WTO. On the one hand, if Chineseexporters know that they will face a less unfairtreatment (i.e., that they will not be subject to theNME status) in foreign antidumping cases, theywill be induced to monitor China’s liberalizationmore closely—to check whether China is effectivelyopening her markets in accordance with her acces-sion protocol—and to provide stronger support toit, including stricter use of China’s antidumpingregulations. On the other hand, if foreign exportersare convinced that they will get more effective andstable access to Chinese markets, they will beinduced to fight for restraining the antidumpingenforcement by their own authorities, in particularfor limiting the use of the NME status.

Doing so could be achieved by implementingthe following simple rule. Foreign antidumpinginvestigators will automatically grant the market-economy status to Chinese exporters of any given


TABLE 3.7 Do Antidumping Investigations Really Look at Dumping?

U.S. Cases (1995–98) EC cases (1995–97)

Number Average Number AverageBases of Estimated Normal Values of Cases Dumping of Cases Dumpingof Exporters Examined Margins Examined Margins

Price comparisons only 5 3.2 8 22.7Price comparisons and constructed values 33 14.2 33 24.4Constructed values 20 25.1 12 25.1Non-market economies 47 40.0 12 45.6“Best available facts” 36 95.6 2 74.5

All cases examined 141 44.7 67 29.6

Sources: Lindsey 1999 for the U.S. cases; Messerlin 2000a for the EC cases.

product that would meet the three followingconditions:

• The Chinese MFN tariff on the product involvedis moderate (e.g., 10 percent or less). This thresh-old tariff will be one of the core components of amore economically sound “interpretation” ofChina’s NME status to be agreed upon duringthe DDA. It could be stable over time, or it couldincrease as time passes—showing an increasingconfidence in the ongoing Chinese liberalizationprocess among China’s trading partners.

• No “core grey measures” shall be imposed on theproduct in question by the Chinese authorities.The list of the core gray-area measures to beintroduced in the interpretation agreement alsoshould be negotiated during the DDA. It shouldbe short (e.g., specific tariffs, quantitativerestrictions, and minimum prices) and only themeasures listed should be considered as parts ofthe conditions.

• No state-owned monopolies shall distribute inChina the competing foreign and Chinese vari-eties of the product in question. Chinese state-owned sole producers are acceptable because,as shown by economic analysis, a protectiongranted exclusively by a moderate tariff elimi-nates the risk of monopoly power of the domes-tic sole producer.

Table 3.4 shows that Chinese ad valorem tariffsapplied in 2001 lower than 10 percent amount to 38 percent of all the tariff lines. The fact that theChinese tariff schedule has only roughly 7,000 tarifflines suggests that it does not offer many opportu-nities to create narrow niches of protection forcarefully defined tariff items. Applying the aboveconditions thus would substantially reinforce therights of Chinese exporters in the antidumpingcases lodged in these two HTS sections. However,China should take some initiatives for improvingthe situation in the other antidumping-intensiveHTS sections, in particular in textiles and clothing.

The rationale for the three conditions is thatthey make unlikely the fact that Chinese exporterswould dump for other reasons than those econom-ically sound, such as differences in demand pattern,and need to meet foreign demand and to makeChinese products known in foreign markets.Hence, the minimum that WTO China’s trading

partners should grant to China is the unconditionalbenefit of market economy status in the antidump-ing investigations faced by the Chinese exportsmeeting these conditions.6

Three final observations are necessary. First, thenegotiations on the improved implementation ofthe NME status for China should be as swift as pos-sible. Second, the conditions can be easily definedon a tariff line (HTS) basis. For instance, Chinacould notify the WTO, on a regular basis, of thetariff (HTS) lines for which these conditions aremet; this procedure could readily be included in thegeneral monitoring procedures of China accessionprocess. Cross-notifications by China’s tradingpartners could be added to the process, under thecondition that they will not slow it. Finally, weakervariants of the above suggestions could be exam-ined, if necessary. For instance, the NME statuscould only be eliminated for the goods notifiedafter a specified period, for example, one yearinstead of immediately. Suggesting weaker variantsis beyond the scope of this paper and is left to tradenegotiators. However, it is worth noting that anyweakening of the suggested approach may havehuge costs in terms of decreasing incentives forexport interests in both China and the rest of theworld to support the transition process of theChinese accession to the WTO.

Stricter Rules on Antidumping

Ambassador Long Yongtu’s desire to introducestricter rules on antidumping in the WTO is a“natural” extension of the negotiation on the NMEuse proposed above because it focuses on theantidumping rules faced by allegedly dumpedexports from market economies. China’s efforts tointroduce stricter rules on WTO antidumping couldfollow two very different approaches.

A cautious approach would be to merely table aseries of proposals or to support those alreadytabled in Geneva by a few WTO members forimproving WTO-based antidumping regulations atthe margin. For instance, the following suggestions,derived from proposals tabled by the SwedishKommerzkollegium 1999, could receive China’ssupport:

• Dumping should be the principal cause of mate-rial injury.


• Double protection (for instance, antidumpingmeasures imposed on the top of quantitativerestrictions) should not be allowed.

• Measures should last a maximum of five years(implying stronger limits to review).

• Repeated initiations in a short period of timeshould not be allowed.

• Cumulation of imports from different countriesshould be banned or severely restricted, unlessthey come from the same firms or from the sub-sidiaries of the same firms.

• Aggregating products under the “one single prod-uct” procedure should be severely restricted.

• All zeroing practices (only export transactionsthat have been found to be dumped are used tocalculate dumping margins) should be banned;all export transactions should be included in theinvestigation.

• The antidumping authorities should producedisclosure documents.

• The use of the de minimis rule should beexpanded in an economically sound way.

Alternatively, China could adopt a bolderapproach. For instance, reforming antidumpingcould take, as often as possible, the form of negoti-ated “quantitative thresholds” (Messerlin 2000b).WTO members could agree that no antidumpingmeasure should be imposed in cases where thelevel of injury losses is less than an agreed thresh-old of the complainants’ revenues of the year(s)used as the reference (pre-dumping) period. Anapproach based on quantitative thresholds is con-ceptually equivalent to a tariffication process. Thisapproach tends to give a sense of the magnitudeof the concessions granted by both sides, bringingit more in line with usual WTO negotiating tech-niques. It is also flexible enough to permit incre-mental reforms and to deliver the progressiveliberalization that WTO members seek throughprogressive increases of the thresholds, therebyavoiding the current deadlock of binary choicesbetween fully enforcing antidumping regulationsor rejecting them totally.

The TPS Provision and the General Issueof Contingent Protection in the WTO

Section 16 of the China Protocol of Accession cre-ates the TPS mechanism. The TPS makes it legally

much easier for WTO members to impose safeguardmeasures against Chinese exports during the next12 years, until 2014 (for a detailed description froma legal perspective, see Andersen and Lau 2001).All the terms defining the use of a safeguard actionin the traditional GATT-WTO context (underArticle XIX) have been systematically weakened: norequirement of “unforeseen circumstances,” noMFN requirement, the need for “material” ratherthan “serious” injury, fewer factors related to thecondition of the domestic industry, a weaker causallink between increased imports and injury, theabsence of a non-attribution causation analysis.

The most important—and potentially the mostdevastating for the WTO—provision of Section 16of the WTO China Protocol states that WTO mem-bers have the unprecedented possibility to use a“trade-diversion” clause. This means that, as soonas one WTO Member implements a TPS measureagainst Chinese exports, all the others couldenforce a similar measure at almost no cost interms of legal procedures (e.g., no investigation, noprior notification, no input from Chinese parties).In short, the trade-diversion clause makes almostunchallengeable the hypothesis that Chineseexports will be diverted from the first closed marketto the rest of the world.

All these features put the TPS largely in contra-diction with the WTO’s usual deep concerns abouta fair balance of rights and obligations. It is a provi-sion so unbalanced that one can fear that its usewill trigger feelings in China close to those associ-ated to the “Unequal Treaties” of the nineteenthcentury. As a result, it represents a serious systemicrisk to the WTO regime.

One may argue that the TPS is such a politicallyaggressive instrument that one may wonderwhether it will be ever used. This argument is farfrom convincing. Its politically explosive contentmakes the TPS unlikely to be initiated by a WTOmember other than a very large industrial country(the United States, the European Community, orJapan). However, all the other WTO members thatwould have waited for such a move would benefitfrom the “trade diversion” clause as soon as a largeindustrial country takes the lead.

What are the possible actions left to China if theTPS were to be invoked? A first possibility would beto negotiate the same “preemptive” approach as forthe NME provision by negotiating during the Doha


Round an interpretation agreement on the effectiveuse of this procedure. WTO members would agreenot to use the TPS when Chinese products wouldmeet the three conditions listed above. Rather, theywould use the normal WTO safeguard provisionunder Article XIX.

Putting antidumping and safeguard regulationson a par would make a lot of sense from the per-spective of the global WTO architecture. MostWTO members use antidumping as a substitute forsafeguards, with many antidumping measures infact being safeguard actions dealing with industriesin difficulties. In fact, the TPS provision expandsChina’s strong stake in substantial improvements ofantidumping rules to the whole WTO contingentprotection regime—that is, antidumping andsafeguard regulations. As a result, during the DDA,China may not only be very active in the antidump-ing negotiations but could also try to expand thesenegotiations to safeguards—so far not explicitlyincluded in the Doha negotiating program—tomake the whole WTO contingent protection regimemore consistent.

A promising way to improve the current WTOSafeguard Agreements would require tying togetherthe concept of temporary protection embedded insafeguards and the basic concept of renegotiationunder GATT Article XXVIII (Messerlin 2000b).For instance, at the end of the second period ofenforcing a safeguard measure allowed by the cur-rent Safeguards Agreement (based in turn onGATT Article XIX), the country shall be requestedto choose between two alternatives: either to rene-gotiate the tariff of the product subjected to thesafeguard measure in question, or to eliminate thesafeguard measure (using then antidumping or anyother trade remedy should be prohibited since allthe instruments of contingent protection are sub-stitutable). The logic of the mandatory aspect ofsuch a choice is that current safeguard andantidumping procedures transform what should betransitory protection into permanent protection.

The possibility remains that such a globalapproach to antidumping, safeguards, and TPSmeasures may face entrenched hostility from WTOmembers, in particular from the top 10 antidump-ing users. In such a case, the remaining options forChina would rely on some kind of threats of retali-ation. The least aggressive approach would be forChina to announce its intention to systematically

use the dispute settlement mechanism as soon as aWTO member notifies TPS use to the WTO Secre-tariat. Lawyers tend to overstate the benefits of suchan approach, by ignoring the full development ofthe trade conflicts. It is almost certain that WTOdispute settlement cases dealing with the TPSwould leave the two parties in a difficult politicalsituation. A more aggressive approach by Chinawould be based on Article 56 of the New Regula-tion. However, such an approach should not ignorethe basic principle of deterrence: trade deterrence,as nuclear deterrence, is good as long as it remains athreat—as long as it stops short of being an effec-tive action.

Conclusion

The chapter provides two results of prime impor-tance for China. First, the countries that would gainmost from better disciplines on antidumping rulesare the few developing countries which, since the1995 Uruguay Round, are intensive antidumpingusers. Because antidumping measures imposed bythese developing countries tend to be more fre-quent and severe than those imposed by industrialcountries, they hurt developing countries’ domesticeconomies much more than industrial country-imposed antidumping measures harm theirs. IfChina wants to continue to enjoy successful liberal-ization, it should not become an intensive anti-dumping user.

Second, few economic and political forces wouldintroduce some automatic counterweights andrestraints to the situation currently existing in termsof antidumping enforcement. Major current usershave few incentives to reform their very discrimina-tory use of the antidumping instrument, mostlytargeting other countries than the top 10 users ifone accepts, of course, the economically soundwelfare-based incentive that protection is harmfulto the economy of the country imposing it. Mean-while, smaller countries have few incentives to usethe antidumping instrument in a retaliatory move.

As a result, China is at crossroads—and withChina, so is the rest of the world, and in particular,the United States and the European Community,the world’s largest economies. On the one hand,China may be increasingly tempted to use anti-dumping rules more intensively for several reasons:as a back door of old-time protection, at the risk of


unraveling her scheduled trade liberalization; as aprogressive integration in the worldwide collusivedimension of antidumping, used as an instrumentfor segmenting world markets for the benefit oflarge firms; and as a retaliatory instrument againstforeign antidumping measures. This last reason isspecific to China, and it deserves a comment. Retal-iatory antidumping measures can have some appealonly for countries with markets large enough togenerate some deterrence power. Chinese marketsare not yet large enough to meet this condition.However, strong Chinese growth will change thissituation relatively rapidly, and China’s increasinglycredible threat of antidumping retaliation impliesthat WTO disciplines on antidumping shouldbecome increasingly attractive to major tradingpartners (if the United States and the EuropeanCommunity are not myopic). Since WTO Roundstend to last for extended periods, this evolutionshould be taken into account by the United Statesand the European Community as early as the cur-rent Doha Round—thereby triggering a potentially“virtuous” dynamic game.

On the other hand, being still a small antidump-ing user and a key target of foreign antidumpingmeasures, China has expressed the desire to imposestricter WTO antidumping rules. It will be one ofthe main beneficiaries of such a move because itwould allow it to preserve its successful trade liber-alization. Such a move may have another key posi-tive outcome: it will help China to negotiate aneconomically sound interpretation of the specialprovisions on antidumping and safeguard includedin its WTO accession protocol. This new interpreta-tion of these special provisions should be based ona few key and economically sound conditions tobe met by China: low tariffs, no “core gray” meas-ures, and no distribution monopolies. Such aninterpretation relies on the following strong eco-nomic and political argument. China and its trad-ing partners have a common interest in establishingthe strongest possible link between Chinese effec-tive liberalization and the elimination of the useof these special provisions by foreign authoritiesagainst Chinese exporters. In other words, the sug-gested interpretation of the specific antidumpingand safeguard provisions included in China’s acces-sion protocol is an effort to mobilize the exportinterests in both China and the rest of the worldduring the difficult implementation period of theChinese accession to the WTO.

Notes

1. In Spring 2002, the EC and the United States declared thatthey will consider Russia as a market economy for purposes ofantidumping investigations. However, the case for the introduc-tion of a TPS provision in Russia’s accession protocol seems stillopen. It would be interesting to make a parallel between the con-ditions imposed on Japan’s accession to GATT, and thoseimposed on China in its accession to the WTO.

2. Tables 3.1 to 3.3 treat measures taken against differentindividual member-states of the EC as one aggregated measureif adopted at the same time and for the same product (datafor 2002 are still not totally complete). By contrast, table 3.4follows the notifications of EC trading partners that vary intheir treatment of the EC (as one entity or as a set of distinctmember-states).

3. Are antidumping measures used by developing countriesas a “safety valve” for softening the impact of their ongoing tradeliberalizations? Rather, comparing the intensity of antidumpinguse with three indicators capturing key aspects (simplicity, irre-versibility, and openness, as estimated by Laird and Messerlin(2002) of the trade policy of the six developing countries, whichare intensive antidumping users, suggests a correlation betweenantidumping intensity and a flagging trade liberalization.

4. For instance, under the NME status, it is possible to useindustrial countries (such as the United States or Sweden) as“reference” countries for China. That introduces systemic errorsabout the product and/or the production process. For instance,it makes no sense to consider, without deep economic analysis,the calcium metal produced in small quantities by a U.S.monopolist for its own use as similar to the calcium metalproduced by China and Russia in large quantities for sales oninternational markets. The U.S. product is likely to have charac-teristics in terms of quality and/or availability making it verydifferent from the Russian or Chinese calcium metal, and it issold and bought in a market structure very different from themarkets of its Russian and Chinese counterparts. In the samevein, trying to estimate production costs by combining inputprices in industrial countries and input quantities used in adeveloping country makes little economic sense.

5. Antidumping authorities in the EC and the Untied States insome cases already adopt more liberal interpretations vis-a-visChinese exporters than the Protocol allows them. In the EuropeanCommunity, for example, this would be the case where theauthorities accept individual treatment or market-economystatus with respect to individual Chinese exporters.

6. It could be argued that market forces in China for theserelatively unprotected products could be distorted by Chineseregulations on inputs for such goods (for instance, subsidies).However, there are WTO instruments for addressing suchan argument (which could be applied to most China tradingpartners).

References

Andersen, Scott, and Christian Lau. 2001. “Hedging Hopes withFears in China’ Accession to the World Trade Organization:The Transitional Special Product Safeguard for ChineseExports.” Mimeo. Geneva: Powell, Goldstein, Frazer andMurphy.

Bloningen, Bruce. 2001. Database on U.S. Antidumping Cases.(www.uoregon.edu).

GATT Secretariat. 1994. The Results of the Uruguay RoundMultilateral Trade Negotiations. Geneva: World TradeOrganization.


Kommerzkollegium. 1999. Further Actions within the EU inthe Field of Antidumping. Stockholm: Kommerzkollegium.Photocopy (17 June).

Laird, Sam, and Patrick A. Messerlin. 2002.“Trade Policy Regimesand Development Strategies: A Comparative Study.” Mimeo.Washington, D.C.: InterAmerican Development Bank.

Lindsey, Brink. 1999. The US Antidumping Law: Rhetoric versusReality. Washington, D.C.: Cato Institute.

Lindsey, Brink, and Dan Ikenson. 2001. Coming Home to Roost:Proliferating Antidumping Laws and the Growing Threat toUS Exports. Washington, D.C.: Cato Institute.

Messerlin, Patrick A. 2000a. “An Economic Perspective onAntidumping and Safeguard.” Seminar on Antidumping and Safeguards. Department of International Trade andEconomic Affairs. Beijing: Ministry of Foreign trade andEconomic Cooperation.

Messerlin, Patrick A. 2000b. “Antidumping and Safeguards.” InJeff Schott, ed. The WTO After Seattle. Washington, D.C.:Institute for International Economics.

Vermulst, Edwin. 2000. “Contingent Protection Provisions inChina’s Draft Protocol on Accession to the WTO.” Seminaron Antidumping and Safeguards. Department of Interna-tional Trade and Economic Affairs. Beijing: Ministry ofForeign Trade and Economic Cooperation.

Wang, Lei. 2003. “China’s New Antidumping Regulations:Improvement to Comply with the WTO Rules.” Mimeo([email protected]). Beijing: East Associates.

World Trade Organization (WTO). 2001. Market Access: Unfin-ished Business, Post-Uruguay Round Inventory and Issues.Special Studies 6. Geneva.


property laws. Since 1990 the government hasupdated its laws covering copyrights, trademarks,patents, and trade secrets (or “anti-unfair competi-tion”) and has adopted protection for new plantvarieties and integrated circuits. The country hasjoined nearly all major international IPRs conven-tions and is a member of international agreementson classification of patents and trademarks and thedeposit of microorganisms.

Most recently China made further revisions toconform to the requirements of the Agreement onTrade-Related Aspects of Intellectual PropertyRights (TRIPS) in the World Trade Organization(WTO). In its accession agreement with WTOmembers, China agreed to implement intellectualproperty laws that are fully consistent with TRIPS.When legislative reforms are fully implemented,China will have a modern structure for IPRs on apar with many developed economies.

China also has established education and train-ing programs in IPRs, upgraded its administrativeand legal systems for enforcing these rights, andundertaken numerous anticounterfeiting pro-grams. Nevertheless, problems remain in the areasof administration and enforcement. Victims ofinfringement complain about weak monetaryand civil penalties, delays in administrative and

After a long period of rapid growth and significantstructural change, the Chinese economy increas-ingly makes use of advanced production technolo-gies, while demand shifts toward higher-qualitygoods and services. Chinese enterprises place grow-ing emphasis on developing brand-name recogni-tion, reputation for quality, and product innova-tion. In such an environment, the provision andenforcement of intellectual property rights (IPRs)help promote further economic development. Withsubstantial structural reform ongoing in Chineseenterprises, it is important to establish incentivesfor the development and expansion of businesses inhigh-growth sectors and to support innovation inconsumer products. Properly structured, IPRs canhelp achieve these goals.

The Chinese central government recognizes theneed for a workable IPRs system. Support is alsogrowing among innovative Chinese enterprises,which are likely to suffer the largest losses fromtrademark and copyright infringement in the econ-omy. Chinese enterprises understand that theiraccess to new foreign technologies depends partiallyon IPRs.

In response both to changes in internal prefer-ences and to considerable external pressure, Chinahas undertaken a dramatic reform of its intellectual

49

4

IntellectualProperty Rights inthe WTO AccessionPack age: Assessing

China’s Reforms

Keith E. Maskus

court procedures, and “local protectionism”that makes enforcement difficult in regionaljurisdictions.

The evolving system presents both opportuni-ties and challenges for the Chinese economy. Theopportunities arise from the improved environ-ment for technical innovation, product develop-ment, and inward technology and investmentflows. The challenges include moving resources outof infringing activities into legitimate businesses,coping with higher costs of imitating products andtechnologies, and absorbing the costs of adminis-tering a stronger system.

Over time, stronger IPRs will shift incentivesaway from encouraging static competition throughcopying and imitation toward promoting dynamiccompetition through innovation, technologyabsorption, and product design. The latter incen-tives are increasingly appropriate for China, whichplans to become a leader in technology develop-ment. However, stronger intellectual property pro-tection will place competitive pressures on laggingenterprises and will raise concerns about the distri-bution of costs and benefits among individuals,enterprises, and regions.

The ultimate objectives of an IPRs system are toincrease competition through innovation and tech-nology acquisition and to encourage innovatorsto make their products available to consumers.Strengthening IPRs improves such incentives but isnot sufficient on its own. Rather, the system needsto be developed within a broader set of policies,including further enterprise reform, developmentof financial and innovation systems, expansion ofeducational opportunities, and means for sustain-ing competition in Chinese markets.

This chapter contains the following sections:

• a description of the progress in China’s WTOcommitments in the IPRs area

• an overview of the economic rationale for intel-lectual property rights

• a discussion of the intricate relationshipsbetween IPRs and economic development,including a review of available evidence on thatsubject and assessment of the potential of theChina’s regime to encourage growth

• an analysis of recent trends in the use of IPRsin China, considering both data and informa-tion learned from a series of interviews in 1998and 2001, and some simple indications of how

Chinese economic development could be affectedby stronger IPRs

• conclusions and recommendations.

WTO Commitments

Since the mid-1980s China has implemented anumber of laws and administrative regulations cov-ering intellectual property protection (see Lacroixand Konan 2002; Maskus, Dougherty, and Mertha1998; Potter 2001). External pressure has been animportant impetus for legal change. Many of thesechanges were made as a result of three agreements,each a Memorandum of Understanding, with theUnited States.

This regulatory process has culminated in theintroduction of numerous changes in China’s IPRsregime, both in anticipation of joining the WTOand in becoming fully compliant after the country’saccession in late 2001. For example, a substantialsecond revision of the patent law was achieved in2000, becoming effective in July 2001. This revisionestablishes full TRIPS compliance in patent regula-tions and clarifies and strengthens certain adminis-trative and judicial procedures, including the use ofpreliminary injunctions. In June 2003 the govern-ment issued procedures for issuing compulsorylicenses for patented inventions, consistent withTRIPS Article 31, while the patent law itself wasrevised in August 2003 (WTO 2003).

Similarly, in 1997 the government promulgatednew rules on the protection of new plant varieties,establishing sui generis, TRIPS-consistent protec-tion along the lines of the 1978 Union for theProtection of New Varieties of Plants (UPOV)treaty.1 A new set of regulations for the protectionof layout designs of integrated circuits came intoforce October 1, 2001.

Just prior to China’s accession to the WTO at theDoha Ministerial in November 2001, some differ-ences with required TRIPS standards remained.2

Table 4.1 provides a list of areas in which standardsare mandated, the norms (minimum standards) inTRIPS, the status of China’s laws and regulationsjust before Doha, and China’s actions in caseswhere a divergence exists. Most of these discrepan-cies arose in the trademark and copyright areas. Forthis reason, on October 27, 2001, the People’s Con-gress enacted a substantial revision to its trademarkand copyright laws to make both consistent withTRIPS obligations.3


TABLE 4.1 Substantive Requirements of the TRIPS Agreement in the WTO

Intellectual Property Rights in the WTO Accession Package: Assessing China’s Reforms 51

Pre-WTO Status of TRIPS Norm Chinese Law Actions

General Obligations

National Treatment Applied for persons Discrimination in copyright Remove enforcement, trademark discriminationagents, and trade secrets protection

Most Favored Nation MFN with reciprocity Member of Berne Convention; exemptions for copyright TRIPS-compliant

Copyrights andNeighboring Rights

Term of protection Life � 50 years; 50 years TRIPS-compliantcorporate

Computer software CR Copyright TRIPS-compliant Discussing patents

Data compilations Copyright Not protected Protect withcopyright

Phonogram producer Right to prevent fixation, Inconsistent with TRIPS Clarify compensation and performer rights reproduction, or system; strengthen

broadcasting for rights50 years

Broadcast rights Right to prevent fixation, Inconsistent with TRIPS Provide right of reproduction, or communication to broadcasting for 20 years, publicor copyright

Rental rights Right to prohibit rental of Not protected Provide rental rightscomputer programs and movies

Discrimination in National Treatment Foreigners could not use Remove discriminationenforcement local copyright bureausprocedures

Trademarks

Well-known marks Protected without requiring No criteria for defining Protect well-known registration “well-known”; none granted marks; establish

to foreigners criteria

Use restrictions Use not required for registration; Law is unclear on prior use Comply with TRIPSimport restraints cannot be used to invalidate use

Symbols protected Rights extend to distinguishing Certain signs are ineligible Comply with TRIPSnames, letters, numerals, colors

Geographical Indications

Basic protection Prevent misleading claims Not protected Comply with TRIPSof origin

Wines and spirits Prevent use of such words as Not protected Comply with TRIPS“style” or “like”

Patents

Eligibility Basic exemptions Probably TRIPS-compliant Clarify compatibilitywith TRIPS

Pharmaceutical Covered; interim marketing TRIPS-compliantproducts rights

Living organisms Micro-organisms and biological TRIPS-compliant; plant and Considering patentsproduction processes covered; animal varieties excluded“higher-order” life optional

Term of protection 20 years from filing TRIPS-compliant



Pre-WTO Status of TRIPS Norm Chinese Law Actions

Patents

Rights Exclude others from production, TRIPS-compliantuse, or distribution

Compulsory licenses Wide scope for use with TRIPS-compliantcompensation and limiting conditions

Burden of proof in Falls on defendant TRIPS-compliantprocess patent infringement

Industrial Designs


Textile designs Covered Protected by copyright Considering designpatents

Plant Varieties

Basic protection Plant breeders’ rights to Plant breeders’ rights with prevent commercial farmer’s privilege and production or marketing research exemption of propagating material (UPOV 1978)

Stronger protection Patents optional No patents

Integrated Circuits


Rights Prevent distribution of IC’s TRIPS-compliantor IC-using products

Exceptions Non-voluntary license TRIPS-compliant

Trade Secrets

Protection from Defines boundaries of unfair TRIPS-compliantunfair disclosure practices

Test data for Protection from disclosure Unfair use not prohibited Protection for six pharmaceuticals for unspecified period and years from date and agricultural unfair use of undisclosed of marketing chemicals data approval

Control of Anti-Competitive Practices

Compulsory licenses Wide latitude for use subject to Weak consultation provisions Comply with TRIPSconditions and consultations

Exhaustion No standard Depends on form of IPR

Enforcement

Sanctions Civil and criminal sanctions In existence but weak Enhance and border measures enforcement action enforcement

Provisional measures Preliminary injunctions and Not fully available Comply with TRIPSseizures

Damages Adequate to compensate Generally low or no Comply with TRIPSvictim of infringement compensation

Administrative actions Enforcement may be through Available but costly and tends Enhance administrative actions to result in small fines enforcement

Judicial review Must be available Not widely available Enhance reviewprocedures

The trademark law establishes the right of indi-vidual Chinese persons to apply to register trade-marks. It clarifies the definition of collective marks,which are marks that several firms in a collective

association (e.g., for producing fruits and vegeta-bles of a given quality standard) may use jointly. Italso establishes joint ownership and protects collec-tive marks and certificate marks for the first time. It


further broadens the range of symbols that may beused as a distinctive mark, extends protection ofwell-known trademarks to their unauthorized useon different products, and sets out criteria forascertaining well-known marks. These regulationson well-known marks were further updated andadopted by the National Industrial and Commer-cial Administration in June 2003 (WTO 2003). Thetrade mark law also sets out protection for geo-graphical indications in accordance with TRIPS.

The copyright law establishes a communicationright over the Internet, sets out fair use limitationsfor electronic content consistent with two newtreaties reached under the aegis of the World Intel-lectual Property Organization (WIPO), clarifiesbroadcast and rental rights, and recognizes thatdatabases are copyrightable. Further changes weremade in August 2002 with respect to rental rights,performance rights, rights of communication andbroadcast, and database protection.

Both the trademark and copyright laws clarifythe amounts of compensation available to plaintiffsand the methods for their calculation. In copy-rights, foreign firms are now permitted to pleadtheir cases with local copyright bureaus in additionto the National Copyright Administration, remov-ing a prior form of discrimination. A number ofclarifications regarding enforcement, administra-tion, and penalties were made in 2002 and 2003(WTO 2003).

This recent and ongoing activity should makeChina’s IPRs legal regime fully consistent withTRIPS in the near term. It is evident that the prom-ise of WTO membership significantly acceleratedChina’s legal reforms in IPRs, because consistencywith at least minimum TRIPS standards is a neces-sary condition for accession (Maskus 2000a).Moreover, China’s commercial interests in innova-tion, development of consumer products, andaccess to high-level foreign technologies providedsignificant domestic motivation for such reforms.Nevertheless, because China remains a countrywith relatively low per-capita income and techno-logical development has not spread widely throughthe economy, it is likely that significant problemswith enforcement will remain for some time. Chinamay become one focus of dispute resolution casesat the WTO as intellectual property interests fromdeveloped countries demand stronger enforcementand protection.

The Nature of Intellectual Property Rights

To analyze how IPRs influence economic develop-ment, it is important to understand their economicunderpinnings. The need for IPRs arises from thesocial objective of promoting the creation of newtypes of information adding to the economy’sknowledge base. These types of informationinclude new products and technologies; new liter-ary, musical, and artistic expressions; and indica-tors of product quality. An intellectual propertyright defines the extent to which its owners maylegally prevent others from taking actions thatinfringe or damage the property. It also may bedefined as the legal ability to set terms on which itmay be used, subject to public-interest limitationson the scope of that ability. For example, patentsprovide exclusive use of a technology, but they aregranted only for a fixed period of time and inreturn for disclosing sufficient technical informa-tion to allow competitors to understand it and tryto improve on it.

Public intervention is needed because the out-come of some intellectual effort may be potentiallyvaluable but also easily copied and used by others,leaving little incentive to incur the original invest-ment costs or to improve it. Without public sup-port for innovation, the economy suffers frominsufficient incentives to develop new products,technologies, and cultural works, making citizensworse off in the long run. An additional problem isthat the social value of information is often greaterthan its private value because there are externalbenefits from new inventions. Examples includespillover cost reductions from new technologies toinput users and network efficiencies from softwaresystems.

At the same time, IPRs generate costs. Legalexcludability imposes a static cost on users, reflectedby the excess of price over marginal productioncost. In the case of intellectual creations, this distor-tion can be significant because the marginal cost ofsupplying additional blueprints, digital video discs(DVDs), and computer programs is small. Addi-tional costs are incurred because IPRs encourageduplication of investment in research and develop-ment (R&D) through patent races, and generatewasteful efforts to assert and defend ownershiprights and to extend those rights beyond the scopegranted. Enforcement costs may be substantial.


Thus, there are complex tradeoffs facing thedesign of IPRs. China needs to strike an appropriatebalance among the needs of creators, developers,and users, in accordance with minimum standardsrequired by TRIPS.

One alternative to IPRs is direct governmentsupport for invention, including public monopoliesin technology development, research institutes inagriculture and industry, and government sub-sidies to university research. Many such programsare important complements to IPRs in nationalinnovation systems. However, the performanceof China’s public organizations and state-ownedenterprises in commercializing products and serv-ices has been poor. This failure arose largely frominadequate incentives within public research insti-tutions to focus on new and marketable inventions.

Intellectual Property Rights andEconomic Development

Relationships between IPRs and economic develop-ment are extremely complex and available evidenceis difficult to interpret (see Evenson and Westphal1997). However there is a growing consensus thatstronger IPRs can improve development prospectsif they are structured properly.

Types of Intellectual Property Rights

It is useful to provide a brief description of themain forms of IPRs. Rights to exploit inventionswith commercial uses are granted through patents,utility models or petty patents; these terms refer toshort-term exclusive rights for small and incremen-tal innovations, and industrial designs. Patents pro-vide the right to prevent for a fixed time (20 yearsunder TRIPS) the unauthorized making, selling, orusing of the product or process described. Utilitymodels provide exclusive rights for a shorter periodto inventions that embody only a small inventivestep. The scope of patent coverage is limited to usesof the novelty claimed by the inventor and recog-nized by patent examiners. The technology mustmeet technical criteria for novelty (or non-obviousness), inventiveness, and industrial utilityand must survive procedures challenging validity.Patent applications are published for inspection byinterested persons. The essential tradeoff in patentsis to create a protected market position in returnfor disclosure of technical knowledge.

A related form of industrial property is plantbreeders’ rights (PBRs), which permit developers tocontrol the marketing and use of new plant vari-eties. These rights operate much like patents, pro-viding for fixed terms, requirements of distinctive-ness in new plant strains, and disclosure rules. Theyare intended to encourage development of newseed strains for agricultural use.

Rights to market goods and services underexclusive names and symbols are protected bytrademarks and service marks. Registration may berenewed indefinitely, subject to use requirements.An important related device is the geographicalindication, which permits the use of a particularplace name where a good was produced to ensurethat the product embodies quality characteristicsof that region. Trademarks provide incentives forfirms to invest in brand-name reputation andproduct quality and for their licensees to produceand sell high-quality goods. If marks were not pro-tected, rival firms would ruin their value by sellingcheaper items under those marks. Thus, the socialbenefits from trademarks include greater productvariety and lower consumer search costs due to theabsence of confusion.

Firms develop technological know-how that isimportant for production but may not be patentableor may have greater economic value if it remainsundisclosed. Such trade secrets are protected bylegal rules against unfair misappropriation. There isno exclusive right to the process if it is discoveredby fair means, such as reverse engineering. Tradesecrets protection is important because it supportsthe introduction of sub-patentable technologicalprocesses into commerce and also promotes compe-tition through reverse engineering.

Artistic and literary creations are protected bycopyrights, which are exclusive privileges to copyand distribute particular creative expressions for afixed term. Related IPRs include neighboring rightsof performers and broadcasters and moral rights oforiginal creators to prevent future alterations oftheir works. Copyrights are subject to limitationsfor social purposes, with the most prominentexception being the fair-use doctrine, under whichothers may use copies for scientific and educationalpurposes. Countries vary widely in the scope oftheir fair-use exceptions to copyrights, particularlywith regard to reverse engineering of computerprograms.

Some new technologies do not fit easily withinthis framework. Computer software contains ele-ments of both literary expression and industrialutility, raising questions about whether it should becovered by copyrights or patents. The TRIPS stan-dard is for copyright protection; the United States,Japan, and other countries also provide patent pro-tection. Similar questions arise regarding semicon-ductor chip designs, which achieve a special form ofprotection under TRIPS. Electronic transmissionsof Internet materials, databases, and broadcastsalso raise concerns about the adequacy of copyrightprotection to encourage their development.

IPRs are enforced both to deter and punishinfringement and to discipline rights holders thatabuse their market power. Enforcement againstbasic infringement involves seizing unauthorizedcopies, destroying the associated facilities if neces-sary, and imposing fines and criminal sanctions.However, as claims about infringement or abuse,particularly in the patent realm, become more com-plicated, courts must decide on their legality, whichrequires considerable legal and scientific expertise.

How IPRs Stimulate Economic Development

IPRs protection can spur economic developmentand growth in several ways. A weak regime can stifleboth invention and innovation even at low levels ofeconomic development. Most inventions are spe-cific to local market circumstances and can benefitfrom patent or utility model protection. Innovationusually involves minor adaptations of existing tech-nologies, management systems, and quality controlmechanisms, which can stimulate growth. Suchinvestments tend to have high economic and socialreturns by raising productivity toward internationallevels. Evidence from Brazil and the Philippinessuggests that effective systems of utility models canpromote innovation (see Maskus and McDaniel1999). Another study demonstrated econometri-cally that Japan’s system of utility models con-tributed positively and significantly to its postwarrise in productivity (see Scotchmer et al. 1991).

Trademarks provide incentives for the entry ofnew firms and the development of new products,even in poor nations. Firms find it easier to innovatecumulatively as they grow larger and their trade-marks are better recognized. This process has twopositive effects on industrial development. First, it

stimulates the entry of small and medium-sizedenterprises into specific markets. Second, it encour-ages more successful enterprises to grow and takeadvantage of scale economies through interregionalproduction and marketing. Some may even becomesignificant exporters as they improve quality.

Similar comments apply to copyrights. Sectorsthat are dependent on copyrights, such as publish-ing, recorded entertainment, and software, willexperience limited entry by local firms in theirabsence. Creation of new films, music, and softwareis expensive and little worth the investment by localentrepreneurs if their products will be copied.Accordingly, society’s long-run cultural and eco-nomic development is impaired.

Innovation goes beyond developing new prod-ucts to establishing marketing and distribution net-works. It is difficult to do this in an environment ofweak IPRs because rights holders cannot readilyprevent infringement. IPRs improve the certaintyof contracts, and permitting better monitoringand enforcement of rights at all levels of the supplynetwork. In turn, both innovative firms and theirdistributors are more willing to invest in marketingand brand-name reputation.

As firms build reputations through trademarks,incentives grow to deter false use of those marks.Fake products sold under a misappropriated trade-mark can ruin reputations, particularly for newfirms, and overcoming such damage can be costly.Thus, effective trademark enforcement shouldincrease the average quality of products over timeand permit consumers to be less wary of counter-feit goods. This is particularly important in cases ofbeverages, foodstuffs, and medicines.

Finally, IPRs can help disseminate knowledge.Patent claims are published, and competitors mayuse the disclosed technical knowledge to developfurther inventions. This cumulative process of in-vention, which depends on the narrowness ofpatent claims, can be an important source of tech-nical change (see Maskus and Penubarti 1995;Smith 1999). Moreover, patents provide a legalbasis for trading and licensing technologies. Trade-marks and trade secrets facilitate informationexchange through ensuring that licensees do notabandon their contracts.

Considerable evidence suggests that interna-tional flows of technology depend on the strengthof IPRs, among many other factors. For example,



international trade in manufactures is positivelyaffected by the strength of patent regimes in largedeveloping countries (see Coe, Helpman, andWhoffmaister 1997). This trade often embodiestechnical knowledge that may be learned in recipientcountries and adapted to local technological capa-bilities (see Maskus 1998a; Lee and Mansfield 1996).

Foreign direct investment, joint ventures, andtechnology-licensing contracts also transfer produc-tion knowledge. It is clear that the strength of IPRsinfluences choices by multinational firms on whereto invest and whether to transfer advanced tech-nologies. Studies of U.S.-sourced foreign directinvestment (FDI) find evidence that firms limit theirinvestments in countries with weak patents (seeMansfield 1995; Contractor 1980). Survey evidenceindicates that the level of technology transferreddepends on the ability to maintain control overthe technology through defense of intellectual prop-erty (see Yang and Maskus 2001). Licensing alsotends to rise with stronger IPRs because of reducedcontracting costs and greater legal certainty.4

How IPRs Limit Economic Development

Tightened IPRs also impose economic costs. Poorcountries may experience net losses in the short-run because the dynamic gains tend to take a longtime to emerge.

The costs of administering and enforcing amodern IPRs system are high. For China they easilywill amount to annual sums in excess of $10 mil-lion (see Watal 1999; Lanjouw 1998). These costsinclude training of examiners, judges, lawyers, andenforcement officers, along with the costs of run-ning various offices. Many of these costs may becovered by administrative fees charged to apply forand register patents and trademarks, while othersmay be reduced by using international registrationagencies such as the Patent Cooperation Treaty(PCT), as does China.

The most visible aspect of IPRs infringement inChina is unauthorized copying of recorded enter-tainment and software and selling of productsbearing counterfeit trademarks. Undoubtedly, sig-nificant amounts of labor are employed in copyingand retailing illegitimate products in China, and animportant short-run cost of stronger IPRs will belabor displacement. The adjustment costs tend tobe smaller in economies with flexible labor markets

and rapid growth, making it easier to shift workersinto legitimate activities. China has mixed prospectson this score, in part because of restrictions oninternal labor mobility. It is conceivable that copy-right and trademark enforcement will contributesignificantly to looming unemployment problemsassociated with economic reform.

Because IPRs raise the costs of copying and imi-tating products and technologies, acquiring tech-nological knowledge through simple imitationcould become more expensive in China. Consider-able anecdotal evidence suggests that firms operat-ing in China lose technologies to potential rivalsthrough such avenues as defection of technical per-sonnel, misappropriation by input suppliers, andcopying of blueprints. Without effective tradesecret protection, these activities are common andhelp establish competition. However, some impor-tant practical effects are that foreign firms tend totransfer older technologies, engage in less techni-cal training, hide key aspects of know-how fromsubcontractors and suppliers, and not establishfirst-rate R&D facilities.

Accordingly, stronger IPRs entail a balancingact. Stronger IPRs make uncompensated imitationmore difficult but improve the quality of technol-ogy flows. Countries wishing to become significanttechnology developers should favor IPRs for thatreason.

Stronger IPRs create market power, from whichfirms may be able to raise prices to monopolisticlevels. This concern is particularly relevant in devel-oping countries for two reasons. First, applicationsfor protection come overwhelmingly from foreignfirms, meaning that the associated profits are trans-ferred abroad. Second, market competition may beweak, supporting monopolization. As will be dis-cussed in the next section, China is in an intermedi-ate position in this regard. Patent applications bydomestic firms are rising rapidly but lag behindforeign applications. Broader economic reformshave improved competitive processes, but the econ-omy remains far from a situation of free entry andvibrant competition in technology and productmarkets.

Chinese authorities may be especially concernedabout the implications of patent rights for theprices of medicines. There is evidence that patentscan support markedly higher prices for pro-tected drugs than for copied and generic drugs.5


However, the extent of these price increasesdepends on the competitive aspects of markets. Themore competitive the local drugs market is beforepatents are awarded, the larger is the share of drugproduction that consists of copies of patentabledrugs; the more inelastic the demand for medicinesis, the higher will be the price increases caused bypatents. These conditions suggest that countrieswith extensive drug imitation and high demandcould experience substantial price increases forprotected drugs. In this regard, China’s policy of public procurement at negotiated prices isappropriate for ensuring the continued provisionof public health services.

Another area of concern is computer software. Itis often claimed that software would be much moreexpensive with enforced copyrights because thecurrent prices of legitimate copies in developingnations are very high compared to prices of unau-thorized copies. However, in countries with highpiracy rates software producers or their distributorstend to sell at low volumes and high markups,reflecting small markets with inelastic demand. InChina such markets are largely limited to foreign-owned enterprises and government agencies. Asmarkets broaden under copyright enforcement,foreign and domestic firms should supply morelegitimate copies at lower prices, suggesting thatultimate price increases could be modest (seePotter 2001).

Thus, there are legitimate concerns aboutmarket power supported by IPRs. However, com-petitive markets and appropriate regulation canmitigate these impacts without unduly reducinginnovative incentives. IPRs need to be introducedinto markets in which other competitive processes,such as firm entry, labor flexibility, distributionsystems, and international trade, are operatingeffectively.

Competition is important because IPRs may beabused, as shown by litigation problems in theUnited States, the European Union, and elsewhere.Such abuses include bad-faith lawsuits, hiddenownership of intellectual property, restrictivepatent pooling agreements, refusals to license tech-nologies, tie-in sales in related markets, and insis-tence on exclusive rights to competing technolo-gies. Thus, China must develop mechanisms forensuring competition maintenance in marketsaffected by IPRs.

A Scorecard for China’s IPRs Regime

With this background it is useful to assess infor-mally the legal standards China has adopted interms of how these standards may affect prospectsfor technical change and the provision of certainpublic goods.6 For this purpose, China’s regula-tions in several critical areas are compared againstbenchmark standards consistent with TRIPSrequirements for middle-income countries as setout by the World Bank (2001). China has character-istics of the dynamic middle-income countries,including a growing base of human capital andsophisticated capabilities in science and technol-ogy, suggesting that this comparison may be appro-priate. However, it also suffers from substantialrural and urban poverty, and many of its enter-prises use technologies that lag significantly behindthose in the modern sectors. Accordingly, it is diffi-cult to provide a definitive analysis of the nation’slaws in this context, and the following analysisis offered largely for purposes of illustration anddiscussion.

An overview is provided in table 4.2. In thecopyrights area, China provides TRIPS-standardterms of protection for creative goods, such as aonce-renewable 25-year term for software copy-rights (50-year maximum period). The essentialquestion is whether the particular conditions oflegal protection are appropriate for China’s devel-opment prospects. For example, it is recommendedthat middle-income countries provide fairly wideexceptions for fair use of copyrighted materials ineducation and scientific research. China’s Copy-right Law embraces this concept, permitting freeuse of copyrighted material in journals, periodicals,and broadcast media for purposes of disseminatingnews. It also allows the uncompensated making ofcopies for classroom use and scientific research andallows a free right to translate works from Han intominority languages. In software, users are permitteda limited right of decompilation for purposes ofdeveloping new programs, which should help theindustry remain fairly open to incremental innova-tion and competition. However, the governmentis considering extending patents to computerprograms, a standard that would exceed TRIPSrequirements and is found only in the United States,Japan, and Australia (see Caves 2000). For a soft-ware sector that remains young and subject to


considerable cross-fertilization through learningand reverse engineering, such a choice seems ques-tionable for the medium term. Neither shouldChina extend patents to methods of instruction orof doing business.

Other important issues exist in the copyrightarea. China has relatively weak complementary

institutions for realizing economic returns tocreative activity, including collection societies forlicensing recorded music and gathering royalties.Indeed, only under the Amended Copyright Lawof 2001 were copyright owners given the right toauthorize collection associations to administer theirrights. Further, because contract enforcement can be

TABLE 4.2 An Assessment of China’s IPRs System for Development Purposes

China’s Post-WTO Area of IPRs Middle-Income Standards Standards Commentary

Copyrights

Fair use Liberal exceptions for Same Important for research and exceptions education and research technology access

Computer Copyrights with decompilation Permit limited decompilation Patents may be overly software and fair use and fair use; considering protective

patents

Market Improve collection societies Weak institutions Improvement would be institutions and rights contracts beneficial

WIPO treaties Adopt minimum standards Considering ratification Could improve internetcontent and access

Data Copyrights with creativity Same Avoid EU-type protectioncompilations requirement

Trademarks

Well-known Narrow definition of sectoral Unclear Clarify scope and recognize marks coverage well-known marks

Domain names Protect against fraudulent Same Important for promoting registration internet use

Confidential Information

Test data Short period of protection Six years of protection Stronger than U.S.from disclosure standard

Disclosure Limited restraints on what Unclear limitation on Ambiguity may be restraints employees may reveal to “promise of gain” detrimental

new employers

Patents

Exemptions Exemptions for discoveries, Same Appropriate to retain from eligibility algorithms, medical exemptions

treatment methods, plants and animals

Novelty Consider oral prior art Oral prior art in China Scope could be widened

Inventiveness High step Notable progress Unclear

Scope of claims Narrow claims and narrow Single claim; scope Depends on examinersdoctrine of equivalents unspecified

Experimental use Permit experimental use Not permitted May be overly strong

Exhaustion International exhaustion National exhaustion Strong protection

Government use For clear public interest Same Appropriate

Compulsory Permitted under TRIPS Same Appropriatelicenses conditions

Utility models Short duration, low 10-year duration, same Appropriateand designs inventiveness

Plant Varieties

Farmers’ privilege Recognize farmers’ privilege Recognized Appropriate

Breeders’ Use for breeding and Permitted Appropriateexemption scientific research permitted

weak and uncertain, the ability of enterprises to allo-cate rights in creative works is limited. Such institu-tions are important for providing a full frameworkwithin which the development of artistic and literarywork can grow beyond intermediate stages (see alsoDahlman and Aubert 2001). To date China protectsdatabases solely with copyright protection, as man-dated by TRIPS, and clarifies that such protectioncannot interfere with the independent rights ofthose who develop components of compilations.This minimum standard is appropriate for a countrywith a strong interest in access to information data-bases, and China should be wary of moving towardthe much stronger standards of the EuropeanUnion. Finally, ratification of the WIPO CopyrightTreaty and the Treaty on Performances and Phono-grams could be beneficial in sorting out copyrightprotection for Internet transmissions, so long asappropriate fair-use limitations are provided.

The Amended Trademark Law of 2001 clarifiesthe definition of well-known trademarks by settingout five criteria for achieving protection withoutregistration. It remains to be seen whether thesecriteria will limit the unauthorized use of suchmarks by others, and the regulation is unclearabout how widely the restrictions on use will applyacross sectors of business. This law also recognizesthat well-known marks should be applicable todomain names on the Internet, a provision thatshould promote wider content available to ChineseInternet users.

Protection of confidential test data from disclo-sure provides applicants for patents in pharmaceu-tical and chemical products a period of exclusivityin the use of results from clinical trial. It seemsadvisable for middle-income economies withdomestic research capabilities in medicines andbiotechnology to provide such exclusivity for someperiod, but TRIPS is silent on the length of anysuch requirement. Surprisingly, China has opted toprotect it for six years from application date, incomparison to the U.S standard of five years.China’s law may be overly protective from thestandpoint of the encouragement of domesticcompetition. Next, China’s law on unfair competi-tion has no explicit language on the legality ofrestraints on the ability of employees to reveal tech-nical secrets to rivals that may hire them, other thanto declare void unfair “promises of gain.” It isunclear what this statute covers, and the ambiguity

may be costly in terms of sorting out the scope ofunfair competition in this area.

China does not patent higher-order life forms orbiological research tools. In its 2001 patent law,China retains appropriate exemptions from cover-age for discoveries of nature, mental methods ofarriving at results (such as computer algorithmsand mathematical formulas), diagnostic and surgi-cal treatments, and plant and animal varieties.These limitations are widely advocated for develop-ing countries with emerging biotechnology sectorsin order to avoid locking up critical technologiesthat support additional research and learning.China’s patent law does not permit experimentaluse of patented materials, however, which may beoverly strong in the context of its developmentstrategies. The country’s standards covering gov-ernment use, compulsory licenses, and utility mod-els and designs are typical of middle-incomeeconomies.

Finally, China’s plant variety law also seemsappropriate for its needs. The patent law excludessuch inventions from coverage, leaving them tovariety protection. China’s regulation permits thefarmers’ privilege to use propagating materialsfor re-planting and also permits experimental usefor science and for rival breeders to develop newvarieties.

The Issue of Enforcement

The largest remaining obstacle to effective use ofIPRs in China is weakness in enforcement proce-dures. Recent legal changes increased the scope ofenforcement considerably. As noted in table 4.1,criminal sanctions are now available for cases ofwillful infringement, while the maximum permis-sible monetary sanctions were increased. Prelimi-nary injunctions and orders for seizure of sus-pected infringing goods may be issued, which areimportant components of timely relief for IPRsowners. Courts may order compensatory damages,though this standard is weaker than what might berequired to deter infringement a priori. Finally,enhanced access to judicial review is provided,which is consistent with prior efforts by theChinese government to strengthen its enforcementmechanisms.

Looking forward, perhaps the most significantissue for China with respect to IPRs and the WTO



is that country’s commitment to enforce againstinfringing activities and prevent exports of piratedand counterfeited goods. Because China remains acountry with relatively low per-capita income andtechnological development has not spread widelythrough the economy, it is likely that significantproblems with enforcement will remain for sometime. Because the scale of activity in China is solarge, weak enforcement against domestic infringe-ment would attract the concerns of patent, trade-mark, and copyright holders. Thus, China maybecome one focus of dispute resolution cases at theWTO as intellectual property interests from devel-oped countries demand stronger enforcement andprotection.

Whether improvements in enforcement areliable to be in China’s favor as a development issuedepends on how economic agents respond to thechanged incentives. As noted above, it is likely thatthis factor will generate higher equilibrium inflowsof technology transfer. However, the essential ques-tion is whether it will expand incentives for localbusiness development. The following section pro-vides evidence that this expansion is likely to occur.

The IPRs Situation in China

Two sources of information about the current situ-ation regarding the use and adequacy of IPRs pro-vide more concrete perspective on the potentialimpacts of IPRs reform. First, the results of inter-views of public officials, university scholars, andenterprise managers conducted in 1998 and 2001paint a consistent picture that has not changed sig-nificantly over the three-year period. In fact, anumber of interviewees described the situationwith respect to enforcement of IPRs to have deteri-orated in that period. Second, analysis of recentpatent and trademark statistics in China suggestthat the use of formal IPRs is growing rapidlybut significant regional disparities exist. Overallthe analysis suggests that the IPRs regime for inven-tion and innovation is improving in China butsignificant problems remain.7

Discussion of Interview Findings

Management officials and intellectual propertymanagers of enterprises from several IPR-sensitiveindustries were interviewed. The enterprises

represented a mix of state-owned enterprises(SOEs), private Chinese enterprises, joint ventureswith international firms, and majority-owned sub-sidiaries of multinational enterprises. Most firms inthe last category are in high-technology sectors andhave significant R&D programs in their homecountries, although some undertake R&D in Chinaas well. Thus, the sample is not representative of thebulk of Chinese industry at this time but is morefocused on product and technology development.

Overwhelmingly, enterprise managers believethat the legal structure for IPRs in China hasimproved markedly and is adequate. However, themajority think the enforcement environmentremains quite weak, while the rest find it to be weakbut improving. Interestingly, Chinese enterprisestend to view the system as improving more rapidlythan do foreign-owned enterprises and joint ven-tures. Many high-technology Chinese enterprisesapplaud the new legal climate, which allows themscope for defending their intellectual property.

Unquestionably, inadequate enforcement is themain problem facing firms wishing to exploit intel-lectual property in China. There can be long delaysin enforcement actions and court rulings. Prior tothe legislative changes in 2001, monetary penaltieswere small even in cases of significant infringe-ment, and there was little scope for criminal prose-cution of willful and ongoing violations. Thenew laws increase maximum fines and clarify thenature of criminal activity, but some intervieweesthought these changes would be insufficient todeter infringement. Enforcement actions may betaken arbitrarily and may lack transparency. Thecentral government and certain regional andmunicipal governments are taking steps to reducethese problems, and several enterprise managers in2001 positively commented on this change.

There are several structural sources of weakenforcement:

• Trademark infringement and illegal copyingremain profitable and face little opposition,especially in rural and inland regions.

• Enterprises engaged in infringement often areimportant employers and sources of revenue forlocal governments.

• Low salaries for public officials may reducetheir effectiveness as enforcement agents, whileadministrative programs may be underfunded.


• Legal and technical expertise for administrativeand judicial operations is limited despite theexistence of special training programs in IPRs.

Among these problems, “regional protection-ism” in IPRs is regarded as the most difficult toconfront by enterprises suffering infringement.There is little coordination among regional bureausof the Administration for Industry and Commerce(AIC). Moreover, the regional AICs have weakadministrative powers and actions of municipalgovernments may supercede them. Municipal gov-ernment officials may well have priorities that takeprecedence over IPRs enforcement.

Managers of both Chinese and foreign-affiliatedenterprises expressed the view that weak enforce-ment of IPRs results in widespread copyright andtrademark infringement. A major problem is thattrademark violations often target innovativeChinese enterprises and thereby deter local businessdevelopment. Examples were given of problemsfacing Chinese-brand producers of such consumergoods as medicines, soft drinks, processed foods,tobacco products, and clothing. Enterprises sellingelectronics products seemed particularly vulnera-ble. Once brand recognition is achieved, domesticenterprises find their trademarks applied to unau-thorized products of lower quality, damaging theoriginal enterprise’s reputation. In some cases,Chinese enterprises either had to give up on theirtrademarks and become licensees of better-knownenterprises or undertake extensive private and pub-lic enforcement actions. It is impossible to know theextent to which this problem hampers industrialdevelopment, but the impact could be significant.

Weak enforcement also impedes efficient use ofpatents and trade secrets. Patent infringementseems to be most common in utility models, whichare easy to copy but are overwhelmingly owned byChinese enterprises. Several foreign enterprises alsoclaimed to have lost patented technologies throughunfair means, such as the selling of design specifi-cations and technical manuals by former employ-ees. According to one industry association, suchcases are becoming more common and increasinglytargeted on sophisticated technologies. Interview-ers claimed that many foreign companies areconsidering more carefully whether they wish totransfer advanced technologies into the Chineseeconomy.

Defection of technical and managerial employ-ees remains a basic problem for both foreign-owned and Chinese enterprises. In economic termsa balance is needed between promoting mobility ofskilled labor, which raises diffusion and competi-tion, and discouraging uncompensated losses oftechnical knowledge, which can reduce competi-tion over time. Both foreign and domestic enter-prises attempt to manage the problem with tempo-rary anti-disclosure clauses, but such contractshave been difficult to enforce in China and recentchanges in the law have not clarified this issue.

Most respondents agreed that the environmentfor selling copyrighted materials is improving inChina, although pirating of software, games, DVDs,and music remains common. While large foreignfirms claim significant harm from such copying, itis likely that relatively larger losses are suffered byChinese entertainment and publishing interests.For its part, the Chinese software industry is grow-ing rapidly, largely because of a substantial base ofskilled software engineers and managers. However,according to many interviewees, such firms concen-trate on developing small-scale programs thatattract less copying, such as business applicationsor limited-run games. This problem could delay theestablishment of Chinese-developed software stan-dards and networking software.

Significant differences exist between Chinese andforeign-owned businesses in their ability to dealwith trademark and other IPRs-related violations.Foreign companies have more resources to combatinfringement than domestic enterprises. Insofar asan enforcement action is a significant expense for asmall or medium-sized Chinese operation, enforce-ment difficulties are biased against Chinese businessdevelopment. Furthermore, foreign companies,particularly Western ones, are more inclined to seeklegal solutions to IPRs problems.

Foreign companies may undertake more defen-sive actions in the presence of weak IPRs. Managersof most foreign enterprises indicated a reluctanceto locate R&D facilities in China, although thisis changing rapidly as the legal environmentimproves. Nearly all indicated that in the past theytransferred technologies that are at least five yearsbehind global standards in the expectation thatthose technologies would be lost to local competi-tion, or they brought in technologies that would beobsolete quickly. Foreign enterprise managers are


TABLE 4.3 Patent Applications by Type and Nationality, 1994–2000

Invention Patents Utility Models Design Patents Total Patents

Domestic Foreign Domestic Foreign Domestic Foreign Domestic Foreign

1994 11,191 7,876 45,188 323 11,428 1,729 67,807 9,9281995 10,018 11,618 43,429 312 15,433 2,235 68,880 14,1651996 11,471 17,046 49,341 263 21,395 3,219 82,207 20,5281997 12,713 20,953 49,902 227 27,456 2,957 90,071 24,1371998 13,726 22,234 51,220 177 31,287 3,345 96,233 25,7561999 15,596 21,098 57,214 278 37,148 2,905 109,958 24,2812000 25,346 26,401 68,461 354 46,532 3,588 140,339 30,343

Growth 126.50% 235.20% 51.50% 9.60% 307.20% 107.50% 107.00% 205.60%Total 100,061 127,226 364,755 1,934 190,679 19,978 655,495 149,138

often reluctant to license technologies, preferringjoint ventures and majority-owned subsidiaries inwhich they can exercise greater control of propri-etary secrets. Enterprises are unlikely to integratefully their Chinese operations, splitting variousproduction processes among facilities in order notto reveal fully the underlying know-how.

Other defensive measures are used by bothChinese and foreign companies. One is to sell onlyto established customers that need assured quality,such as hospitals, large enterprises, and publicagencies. This acts as a barrier to the entry of smallfirms needing the associated products or inputs. Asecond is to establish strict vertical supply anddistribution chains to permit monitoring of qual-ity. A third is to employ technical safeguards, suchas software locks and encrypted source codes thatmust be decoded to operate software upgrades.

It is impossible to know how these distortionsassociated with weak IPRs contribute to economicinefficiency in China, although the effects presum-ably are significant. If so, stronger IPRs over time willgenerate important static and dynamic efficiencies.

Public officials often raise concerns about thepotential impacts of stronger IPRs on prices andcompetition. Some officials also recognize thatstronger IPRs need to be accompanied by otherpolicy measures to build technological capacitiesand maintain competition. However, this recogni-tion only recently has been translated into suchpolicy initiatives as the Standards Office within theState Industry Commission, which will establishuniform national standards for information net-works, including copyright provisions.

Patent and Trademark Activity in China

Despite these problems, data on patent and trade-mark use indicate that both foreign and domesticenterprises are applying for more protection.Table 4.3 presents figures on applications to theState Intellectual Property Organization for allthree types of patents from 1994–2000. Domesticenterprises more than doubled their applicationsfor invention patents, while foreign applicationsrose by 235 percent. From 1996 through 1999 for-eign applications considerably exceeded domesticapplications. A significant rise in Chinese domesticapplications in 2000 virtually equalized the numberfor that year, however. Chinese enterprises nowapply for nearly as many invention patents, withtheir higher inventive content, as do foreignenterprises.8

In contrast, applications for utility models anddesign patents overwhelmingly are filed by Chineseorganizations. In both categories domestic applica-tions rose far faster than foreign applications. Thus,these rewards aimed at encouraging small-scaleinvention seem to be having their desired effect ondomestic innovation.

Table 4.4 shows data for patent grants and the ratio of cumulative grants to cumulativeapplications over the same period. Grants of inven-tion patents to both domestic and foreign appli-cants increased rapidly, with the former nearlycatching up to the latter by 2000. The aggregategrants ratios for invention patents are surprisinglylow, perhaps reflecting long examination delays.Grant rates are much higher in utility models and


TABLE 4.4 Patent Grants by Type and Nationality, 1994–2000

Invention Patents Utility Models Design Patents Total Patents

Domestic Foreign Domestic Foreign Domestic Foreign Domestic Foreign

1994 1,659 2,224 32,611 208 5,507 1,088 39,877 3,5201995 1,530 1,863 30,195 276 9,523 1,677 41,248 3,8161996 1,383 1,593 26,961 210 11,381 2,252 39,725 4,0551997 1,532 1,962 27,185 153 17,672 2,488 46,389 4,6031998 1,655 3,078 33,717 185 26,006 3,248 61,378 6,5111999 3,097 4,540 56,094 274 32,910 3,241 92,101 8,0552000 6,177 6,506 54,407 336 34,652 3,267 95,236 10,109

Growth 272.30% 192.50% 66.80% 61.50% 529.20% 200.20% 138.80% 187.20%Total 17,033 21,766 261,170 1,642 137,651 17,261 415,954 40,669Grants Ratio 17.00% 17.10% 71.60% 84.90% 72.20% 86.40% 63.50% 27.30%

Source: State Intellectual Property Office 2000.

TABLE 4.5 Bilateral Invention Patent Applications, 1994–98

1994 1996 1998

by China in China by China in China by China in ChinaCountry in Foreign by Foreign in Foreign by Foreign in Foreign by Foreign

U.S. 190 8,105 246 14,892 436 25,634Japan 137 3,742 145 7,212 373 11,301Germany 172 2,094 210 3,631 624 6,599U.K. 188 2,028 224 2,656 627 4,216Australia 87 692 101 746 295 1,067Korea, Rep. of 84 569 97 1,645 309 2,076Brazil 85 20 100 39 290 104India 21 8 54 6 15 21

design patents, which are easier to examine andcarry shorter protection periods.

Tables 4.5 and 4.6 indicate trends in bilateralinvention patenting activity between China and keytrading partners.9 Chinese patent applicationsabroad rose sharply between 1996 and 1998 in allcountries listed except India, suggesting an increas-ing international orientation of Chinese innovation.However, China remains well behind most devel-oped countries in terms of bilateral applicationsflows, with the United States and Japan togetherapplying for some 37,000 patents in 1998 in thatcountry. Despite the increase in Chinese applica-tions abroad, only Japan actually granted risingnumbers of invention patents to Chinese inventors.

Table 4.7 provides a breakdown of total domesticpatent applications for the top 11 patenting regions

in China during 1985–96 and 2000. Residents ofGuangdong applied for over 21,000 patents in2000, while people in Hebei applied for less than4,000. Better measures of inventive capacity aregiven in the final two columns as applications permillion people and applications per million yuan ofregional GDP. In these rankings Beijing is at the topof the list, with far more applications per capita andper unit of output than any other province. Thisreflects both Beijing’s status as a technology devel-oper and the fact that many patent registrationscome through legal offices in the capital. Shanghaihas the second highest applications per person butranks sixth in applications per yuan of GDP. Fujianand Hebei rank low in both categories.

The middle column ranks these regions interms of average income per capita. There are large


regional disparities in income levels, ranging fromSichuan at the bottom to Shanghai at the top. Thedifference between them is a factor of 5.6, which isextraordinarily high for regions within a country.It is interesting to correlate per-capita GDP withthe relative patent application figures; there is astrong positive correlation (0.54) between GDP percapita and patent applications per million people.Thus, richer provinces apply for more patents(develop more products) per person than poorprovinces. The correlation between GDP per capita

and applications per million yuan of GDP is higher(0.78). Accordingly, higher incomes are associatedwith greater innovation propensities, which in turnraise regional economic growth.

Data for trademark registrations not shown heretell a similar story. In particular, trademark applica-tions have risen rapidly since 1994, especially thosethrough the Madrid Protocol. Far more domesticmarks are registered than foreign ones, but foreignapplications have increased at a faster pace. Interms of regional performance, Guangdong had

TABLE 4.7 Patenting Indicators for Top Patenting Regions, 1985–96 and 2000

2000 Applications 2000 Applications Applications 2000 GDP per per Million of per Million Yuan of Region 1985–96 2000 Capita, Yuan Population GDP

Guangdong 42,159 21,123 11,180.55 487.84 436.33Shanghai 21,758 11,337 27,187.57 1,299.76 478.07Beijing 54,348 10,344 17,936.32 3,932.56 2,192.51Zhejiang 29,197 10,316 12,906.56 624.27 483.68Shandong 37,082 10,019 9,408.97 408.44 434.09Jiangsu 34,983 8,211 11,538.99 470.33 407.60Liaoning 38,768 7,151 11,017.23 914.77 830.31Sichuan 27,046 4,496 4,814.86 324.72 674.41Fujian 11,027 4,211 11,293.86 317.69 281.29Hunan 26,400 4,117 5,732.76 409.94 715.08Hebei 20,584 3,848 7,545.91 305.22 404.48

Correlation with GDP per capita 0.54 0.78

Sources: The Patent Office, Annual Report, 2000; State Statistical Bureau, China Statistical Yearbook 2000;Author’s calculations.

TABLE 4.6 Bilateral Invention Patent Grants, 1994–98

1994 1996 1998

to China by China to China by China to China by ChinaCountry by Foreign to Foreign by Foreign to Foreign by Foreign to Foreign

U.S. 48 701 46 449 72 785Japan 5 579 29 445 7 927Germany 16 213 15 148 7 318U.K. 21 94 29 66 12 133Australia 4 24 5 20 14 36Korea, Rep. of 0 48 0 49 2 149Brazil 2 4 0 1 5 2India 5 2 1 0 0 3

Note: Data include applications under Patent Cooperation Treaty.Source: WIPO, various years.


the largest absolute number of applications in2000, followed by Zhejiang and Jiangsu. Scaled bypopulation, however, Shanghai ranked first by alarge margin, followed by Beijing, Zhejiang, andGuangdong. There is a very high correlation (0.81)between per-capita GDP and per-capita applica-tions, reflecting again that trademark applicationsrise with income levels.

Thus, the use of patents and trademarks is risingrapidly in China for several reasons:

• Laws have been strengthened and fees reduced,encouraging more applications.

• As trademark and patent infringement haveincreased, both domestic and foreign enterprisesrecognized the importance of establishing intel-lectual property protection, even in an environ-ment of weak but improving IPRs.

• Chinese markets are getting deeper as incomegrows, despite the substantial barriers to inter-regional integration, And registration of IPRs isimportant for exploiting deeper markets.

• Chinese research organizations and enterprisesare engaged in more inventions, and Chinesefirms are undertaking more innovative activity.

Conclusions and Recommendations

In recent years China has made significant progresson the legislative end of intellectual property rights,especially in preparation for its entry into theWTO. The specific standards it has adopted acrossthe range of intellectual property regimes arelargely consistent with what might be recom-mended for middle-income developing countrieswith strong innovation potential. However, in somedimensions the new Chinese standards may beoverly protective for an economy that remainslargely a net importer of new technology and infor-mation. At the same time, China continues to expe-rience severe enforcement problems. Enforcementis likely to be problematic for the intermediate termbecause of structural difficulties with the system andbecause costs of copying and counterfeiting remainsmall relative to prices of legitimate products.

China is undergoing a long process of increasingsophistication in technology use and development.Three important problems arise with IPRs thatmay influence the pace and characteristics of thistransformation. First, inadequate enforcement of

IPRs limits incentives to develop products andbrand names, especially on the part of small andmedium-sized domestic enterprises. This structuraldifficulty likely limits entry of new firms and thedevelopment of entrepreneurial skills. It alsorestricts the ability of enterprises to market nation-ally and to take advantage of economies of scale, andtends to reduce investment in quality improve-ments. Over time, this situation could make itincreasingly difficult to break into export marketsfor high-quality and high-technology goods.

Second, Chinese enterprises and research organ-izations are engaging in more innovation, assuggested by the patent and trademark statistics.However, the country remains behind global stan-dards in allocating resources to R&D and science(Maskus, Dougherty, and Mertha 1998). Moreover,interviewees noted that SOEs and state researchinstitutions face structural difficulties in commer-cializing the results of invention. This points outthe importance of continuing to develop a technol-ogy innovation system that encourages innovativeactivity.10 The state has important roles to play inpromoting pre-competitive research and removingdisincentives to commercialization. China has madeprogress toward these goals, with support programsin information technology, biotechnology, andother important areas, along with efforts to raise theflow of knowledge from institutes and universitiesto producing enterprises. Nonetheless, ambiguitiesremain about effective ownership of intellectualproperty rights. This is another reason that the newsystem of IPRs should be an important componentof the evolving innovation system.

Third, stronger IPRs alone are not sufficient toestablish effective conditions for further technologydevelopment and growth. Rather, they must beembedded in a broader set of complementaryinitiatives that maximize the potential for IPRsto be pro-competitive over the long term. Animportant complement is development of humancapital through education in science, technology,and law, and acquisition of skills through trainingin enterprises. Enterprises should be more willingto undertake such training under an improved IPRs regime. Both directly and indirectly, then,effective IPRs can help Chinese enterprises raisetheir technological capabilities, which is critical foradaptation of foreign technologies and innovationof new products.


Another supporting factor is to ensure that com-petition on domestic markets is sufficient to preventstronger IPRs from becoming a damaging source ofmarket power. Further enterprise reform and dereg-ulation are important in this context. Over time theliberalization commitments made in the WTO willprovide important competition as well.

Finally, China like other countries, has the rightto safeguard its interests in competition and socialobjectives through effective regulation of IPRs asthose rights become stronger. Thus, the govern-ment should further analyze the appropriate formof pricing regulations and compulsory licensing in its drug procurement programs as medicinesreceive stronger protection.

Moreover, an opportunity arises for China toconsider what form of competition regime it willimplement as it shifts further toward the market.Currently China tries to maintain competitionthrough centralized regulation of market structure,ownership, and innovation, a system that willbecome increasingly incompatible with needs fortechnological change. A shift toward antimonopolyregulation of such IPRs abuses as monopoly pric-ing, restrictive licensing arrangements, and refusalsto deal is important. China also may need to employcompulsory licenses for this purpose. To be effec-tive, such regulation needs to be well defined, non-discriminatory, and professionally applied by thecompetition authorities and courts. This pointsagain to the need for building legal expertise in IPRsover the long term.

The authorities in China expect that the strongerIPRs regime set in place by new legislation andincreased efforts at enforcement will supportdynamic gains in technology acquisition and inno-vation. This outcome seems achievable in thoseregions and sectors that are technologicallydynamic. At the same time, however, substantialdifferences in incomes, education, and entrepre-neurship persist among regions, while much of thecountry remains poor. Developing a comprehensiveapproach to regulating the use of IPRs that helpsbring these poorer regions and groups more fullyinto the modern commercial system will pose asignificant challenge for China. In the long run,however, the technological dynamism that shouldbe facilitated by the new IPRs regime will greatlyassist the country in its efforts to spread widely thegains from competition.

Notes

1. See “Draft Protocol on the Accession of China,” July 10,2001.

2. For policies taken before 2003, the description in theseparagraphs relies on information from an interview with a sen-ior judicial official in Beijing and on Lehman, Lee and Xu, ChinaIntellectual Property Newsletter: Special Issue, 2001.

3. For extensive reviews see Evenson and Westphal (1997),Maskus (2000a), and Primo Braga, and others (1998).

4. UNCTAD (1996) presents estimates of such costs inseveral developing countries.

5. Prices of copyrighted goods have fallen sharply in Taiwansince the aggressive crackdown on counterfeiting in the mid-1990s, in part because of additional competition from legitimatelocal developers.

6. The European Union has moved recently toward theprovision of patent protection.

7. It is conceivable that the increase in Chinese applicationscome primarily from joint venture partners of foreign enter-prises but unfortunately the available data do not make thisdistinction.

8. Because these data were taken from WIPO, they includeapplications made under the Patent Cooperation Treaty.

9. Dahlman and Aubert (2001) discuss this in detail.10. See Maskus (2000a) for more detailed discussion.

References

Caves, Richard E. 2000. Creative Industries. Cambridge: HarvardUniversity Press.

Coe, David T., Elhanan Helpman, and Alexander Whoffmaister.1997. “North-South R&D Spillovers.” The Economic Journal107: 134–49.

Dahlman, Carl J., and Jean-Eric Aubert. 2001. China and theKnowledge Economy: Seizing the 21st Century. Washington,D.C.: World Bank.

Dougherty, Sean M. 1997. “The Role of Foreign Technology inImproving Chinese Productivity.” MIT Science and Technol-ogy Initiative, Beijing.

Evenson, Robert E., and Larry E. Westphal. 1997. “TechnologicalChange and Technology Strategy.” In Handbook of Develop-ment Economics: Volume 3. Amsterdam: North-Holland.

Jiang, L. 1996. “Technological Innovation in Business Strategy.”Science and Technology International 1: 60–3.

Lacroix, Sumner, and Denise Eby Konan. 2002. “IntellectualProperty Rights in China: The Changing Political Economyof Chinese-American Interests.” The World Economy 25:759–88.

Lanjouw, Jean O. 1997. “The Introduction of PharmaceuticalProduct Patents in India: ‘Heartless Exploitation of the Poorand Suffering’?” Discussion paper no. 775. New HavenConnecticut, Economic Growth Center: Yale University.

Lee, Jeong-Yeon, and Edwin Mansfield. 1996. “Intellectual Prop-erty Protection and U.S. Foreign Direct Investment.” Reviewof Economics and Statistics 28: 181–6.

Ma, Chi, and Gao Chang Lin. 1998. “Technological Innovationin China’s Manufacturing,” State Science and TechnologyCommission, Beijing, manuscript.

Mansfield, Edwin. 1995. “Intellectual Property Protection,Direct Investment, and Technology Transfer.” Discussionpaper 27. Washington: International Finance Corporation.

Maskus, Keith E. 1998. “The International Regulation of Intel-lectual Property.” Weltwirtschaftliches Archiv 123: 186–208.

———. 2000a. Intellectual Property Rights in the Global Economy.Washington D.C.: Institute for International Economics.

———. 2000. “Intellectual Property Rights in Lebanon.” In B.Hoekman and J. Zarrouk, eds. Catching Up with the Compe-tition. Ann Arbor: University of Michigan Press.

Maskus, Keith E., and Christine McDaniel. 1999. “The Impactsof the Japanese Patent System on Post-War ProductivityGrowth.” Japan and the World Economy 11: 557–74.

Maskus, Keith E., Sean Dougherty, and Andrew Mertha. 1998.“Intellectual Property Rights and Economic Development inChina,” manuscript.

Maskus, Keith E., and Mohan Penubarti. 1995. “How Trade-Related Are Intellectual Property Rights?” Journal of Interna-tional Economics 39: 227–48.

Potter, Pitman B. 2001. The Chinese Legal System: Globalizationand Local Legal Culture. London: Routledge.

Primo Braga, Carlos A., Carsten Fink, and Claudia PazSepulveda. 1998.“Intellectual Property Rights and EconomicDevelopment.” Manuscript. Washington, D.C.: World Bank.

Scotchmer, Suzanne. 1991. “Standing on the Shoulders ofGiants: Cumulative Research and the Patent Law.” Journal ofEconomic Perspectives 5, Winter, 29–42.

Smith, Pamela J. 1999. “Are Weak Patent Rights a Barrier to U.S.Exports?” Journal of International Economics 48: 151–77.

UNCTAD (United Nations Conference on Trade and Develop-ment). 1996. The TRIPS Agreement and Developing Coun-tries, Geneva: UNCTAD.

Watal, Jayashree. 1999. “Pharmaceutical Patents, Prices andWelfare Losses: A Simulation Study of Policy Options forIndia under the WTO TRIPS Agreement.” Washington, D.C.:Institute for International Economics. Photocopy.

World Bank. 2001. Global Economic Prospects 2002: Making TradeWork for the World’s Poor. Washington, D.C.: World Bank.

WTO (World Trade Organization). 2003. “Transitional ReviewMechanism of China: Communication to the Council forTrade-Related Aspects of Intellectual Property Rights,”IP/C/W/415.

Yang, Guifang, and Keith E. Maskus. 2001. “Intellectual PropertyRights and Licensing: An Econometric Investigation.”Weltwirtschaftliches Archiv 137, 1, 58–79.


to 15.9 percent, 4.6 percent lower than in 1995. Thereduced share of agriculture in the national econ-omy, which is a normal development in a period ofaccelerated industrialization, does not alter theimportant position of agriculture in the nationaleconomy. The agricultural growth was remarkablyhigher than the population growth; annual popula-tion growth on average was 0.91 percent, havingprovided a reliable guarantee for improvements inthe standard of living, especially diet and nutrition.

The agricultural and rural economic structureshave improved gradually. In agricultural production,the share of forestry, animal husbandry, and fisherycontinued to increase compared to crop farming.Promoted by technical progress, the capacity ofcrop farming to ensure provision of farm produceand support agricultural restructuring is increasing(table 5.1).

At the same time, the percentage of farmers’incomes derived from secondary and tertiary indus-tries has been increasing. In 1995, the average netincome from secondary and tertiary industriesaccounted for 32.65 percent of the net productiveincomes; this percentage increased to 46.65 percentin 2000. The value-added of township and villageenterprises (TVEs) was equivalent to 121.7 percentof the gross value of agricultural production in1995 and increased to 191.1 percent in 2000.

At the time of its accession to the World TradeOrganization (WTO), China was following theninth Five-Year Plan for the Development of theNational Economy and Society. The Plan requiredChina to accomplish two major goals in its nationaleconomic and social development: to establish thesystem of the socialist market economy at the pri-mary stage and to attain a higher standard of living.Accomplishing these two goals would provide animportant foundation for reform and developmentand for the accomplishment of the third step, thestrategic goal of modernization. While remarkableachievements had been made in agriculture andrural economy during the ninth Five-Year Plan, sig-nificant problems had developed that could not beignored.

Agriculture and Rural Developmentduring the Ninth Five-Year Plan

During this period, the gross value of agriculturalproduction in China grew from RMB1199.3 billionin 1995 to RMB1421.2 billion in 2000, an averageannual increase of 3.5 percent in constant prices,0.5 percent lower than annual growth during theeighth Five-Year Plan. In 2000, the share of agricul-ture in Gross Domestic Product (GDP) decreased

69

5

China’s AgriculturalDevelopment and

Policy Readjustmentafter Its WTO

Accession

Chen Xiwen

The average standard of living of farmers con-tinued to improve. In the ninth Five-Year Plan, eventhough the growth of farmers’ income sloweddown year by year, the per capita net incomeincreased from RMB1,577.7 in 1995 to RMB2,253.4in 2000, an average annual increase of 4.7 percentafter adjustment for price changes. This growth ex-ceeded the target of 4 percent set in the outline ofthe ninth Five-Year Plan. Farmers’ standard of liv-ing also continued to improve. In 1995, spendingon food made up 58.62 percent of per capita livingexpenses of farmers; this percentage was reduced to49.13 percent in 2000. Compared with 1995, elec-tric fans owned by every 100 farmers’ householdsincreased by 34 sets, black-and-white televisions by38 sets, color televisions by 32 sets, refrigerators by7 sets, washing machines by 12 sets, and motorcycles by 17 sets. The per capita housing spaceincreased from 21 to 24.82 square meters.

The most remarkable achievements were thesignificant expansion of the production capacity ofgrain and other major farm produce and the histor-ical transition from the long-standing short supplyof major agricultural products to balance in aggre-gate and surplus in years of good harvests.

The achievements in grain production madeduring the ninth Five-Year Plan period are of his-torical significance to the development process ofthe Chinese economy and society. The pressure ofpopulation growth on demand for food has been a significant economic and social problem con-fronting China for close to 400 years. In the lateMing Dynasty and the early Qing Dynasty, China’spopulation was less than 100 million and had 800million mu of farmland. By 2000, the total popula-tion had grown to 1.266 billion, while the farmlandarea only expanded to 1.924 billion mu. The popu-lation grew over 11-fold, while the farmland areaincreased by a factor of little more than two. Afterthe country’s founding in 1949, “taking grain as the

key link” was stressed for years to alter the situationof food shortage, but the problem was not resolved.The grain rationing system was terminated as lateas 1992, producing improvements in the foodsupply situation. Beginning in 1995, China hadenjoyed good harvests in grain production for fiveconsecutive years. In the first four years of the Planperiod, the annual grain production surpassed500 million tons, leading to a continuing situationof supply exceeding the demand.

Four causes contributed to the significant growthin grain production during the ninth Five-YearPlan period:

• The basic rural policies were improved continu-ously. The household contract responsibility sys-tem featuring a combination of centralizationand decentralization played an important role in expanding grain production. In November1993, the central government explicitly extendedthe land contract for another 30 years after the expiration of the original 15-year contractperiod. This policy gave farmers a sense of sta-bility in terms of land contract rights and stimu-lated their initiative to increase investment in farmland. Many farmers have been diggingmotor-pumped wells and irrigation ditches onthe contracted land and purchasing sprinklerirrigation devices to develop water-saving irriga-tion. The area of irrigated farmland increased by68.55 million mu or 9.3 percent.

• Increases in the contract purchase price for grainprovided an incentive for farmers to increase pro-duction. In 1994, consumer prices rose. To makeup for the effect of these rises on farmers, thegovernment raised the contract purchase pricefor grain by 40 percent; the price was furtherincreased to 42 percent in 1996.

• Scientific and technological progress has beenresponsible for nearly 40 percent of the agricul-tural growth in China. The extension of severalkey technologies, in particular, played an impor-tant role in increasing grain production. The“seed project” introduced a large number ofimproved varieties and eliminated poor varietiesthat did not suit market needs. The “WhiteRevolution” extended mulching technology inthe cold and dry areas in the northern part ofChina to advance sowing time and prolong theperiod of crop production while improving soil


TABLE 5.1 Changes in Farm OutputValue Shares (percent)

Crop AnimalYear Farming Forestry Husbandry Fishery

1995 58.43 3.49 29.72 8.362000 55.68 3.76 29.67 10.89

Source: China Statistical Yearbook 2001.

moisture conservation and repressing weedgrowth. Water-saving irrigation technologieshave been developed, and demonstration coun-ties where these technologies are being appliedhave been designated. The flood irrigation prac-tice has been replaced by wide application ofspraying, trickle, and micro-irrigation methodsto meet the needs of crop growth with less water.

• Favorable climatic conditions produced adequaterainfall that alleviated long-standing droughtconditions and facilitated increases in grain pro-duction. The most serious threat to agriculturalproduction in China traditionally has beendrought.

Generally, the agricultural policies during thisperiod were stable and explicit, and many of thescientific achievements facilitated growth in grainproduction. However, factors such as major rise inprices and favorable climatic conditions are incon-sistent, and reliance on a confluence of such favor-able factors in the future would be imprudent.Accordingly, the long-term development of grainproduction remains an issue meriting ongoing seri-ous attention.

The Biggest Problem in China’s RuralEconomy: Raising Farmers’ Incomes

The decreasing growth in farmers’ income was themost outstanding problem during the ninth Five-Year Plan period. Changes in the patterns of supplyand demand produced an excess of supply for mostfarm produce, leading to declines in prices and newdifficulties in raising farmers’ incomes. In 1996,farmers’ net income increased by 9 percent, but therate of growth declined to 4.6 percent in 1997,

4.3 percent in 1998, 3.8 percent in 1999, and2.1 percent in 2000.

The decreasing growth in farmers’ net income inrecent years is only a symptomatic manifestation ofthe problem. The severity of the problem lies pri-marily in the sustained decrease of income fromagricultural production. Table 5.2 shows the com-position of and changes in per capita net income offarmers from 1997 to 2000.

It can be clearly seen from Table 5.2 that, whilethe productive net income of farmers in 2000increased by RMB142 over 1997, the net incomefrom agricultural production decreased by 132, areduction of 10.4 percent. Between 1998 and 2000,the average farmer’s income from agriculturalproduction declined progressively, decreasing byRMB30 between 1998 and 1997, RMB57 between1999 and 1998, and further by RMB44 between1999 and 2000. According to the statistics in theagricultural survey, 59 percent of the rural house-holds in China are “pure farmer households” (i.e.,the family members are mainly engaged in farm-ing), and another 18 percent are “farmer house-holds with combined occupations” (i.e., thenumber of family members engaged in farming ismore than those engaged in nonfarm occupations).These two types of households with farming astheir main source of income account for 78 percentof households in rural China. Accordingly, due tothe progressively reduced income from agriculturein the last three years, the per capita net income ofthe households with farming as the main source ofincome actually declined.

It was under just such circumstances that partof the agricultural provinces and the westernprovinces with underdeveloped nonfarm industriesfaced a severe reduction in the per capita income of

China’s Agricultural Development and Policy Readjustment after Its WTO Accession 71

TABLE 5.2 Sources of Farmers’ Incomes, 1997–2000

Productive Net Income from Net Income from Net Income fromNet Income Primary Industry Secondary Industry Tertiary Industry

Year (Yuan/Person) (Yuan/Person) (Yuan/Person) (Yuan/Person)

1997 1987 1,268 438 2811998 2040 1,237 499 3031999 2079 1,180 564 3342000 2130 1,136 598 3952000 vs. 97 +142 −132 +160 +113

farmers. In 1999 and 2000, farmers in six provincesand regions experienced decreases in their percapita net income over the previous year. Table 5.3shows the provinces and amounts of the reductionin income.

Per capita income decreased for two consecutiveyears in the provinces of Liaoning, Jilin, andHeilongjiang. The reduction in Jilin was as much asRMB361 in the same period, and per capita incomein the province was 15 percent lower in 2000 thanin 1998.

The net income from farming declined for yearsrunning, while the tax burden on farmers’ house-holds mainly included taxes on agricultural andanimal products, taxes on special products, anda contract land deduction. These factors havedeepened the feeling of households in which farm-ing is the main source of income that “income cannot go up and burdens can not come down.” Thissituation has forced some farmers to seek workfrom other sources to offset payment of agricul-tural taxes and fees with income; more farmershave had to reduce their spending on productionand living expenses. Since 1998, the per capitaspending on operation and living expenses hasdecreased or stopped increasing. Table 5.4 showsthe changes in this regard.

There were successive declines in 1998 and 1999in per capita spending on farm household opera-tions, with a cumulative 15 percent reductionbetween 1997 and 1999 (a reduction of RMB106).The increase in per capita spending on farmerhousehold operations in 2000 constituted some-thing of a recovery, but it still left spending over7 percent lower than in 1997. On the other hand,this increase was indirect, the result of increases indrought relief expenses and the rise in prices forfuels.

The changes in the per capita consumer spend-ing of farmers have similar characteristics: succes-sive declines in 1998 and 1999, with spending in1999 at 2.5 percent (RMB40) below the level in1997. Although the per capita living expendituresincreased by RMB93 in 2000 compared with theprevious year, the change was to a large extentpassive. First, the per capita spending on food con-tinued to decrease after the successive decrease fortwo years, while per capita spending on garmentsincreased slightly but was still lower than the levelsin 1997 and 1998. Second, much of the spendingincrease resulted from the rise in service prices,such as medical expenses that increased by RMB18or 25 percent, transport by RMB24 or 35.5 percent,and education and entertainment devices and


TABLE 5.3 Declines in Per Capita Farm Incomes in Six Key Provinces

Yuan Yuan Yuan Yuan Yuan Yuan

Reduction in Shanxi Liaoning Jilin Heilongjiang Gansu Xinjiangincome in 1999 86 79 123 87 36 127Reduction in Guangxi Liaoning Jilin Heilongjiang Shaanxi Ningxiaincome in 2000 184 145 238 18 12 30

TABLE 5.4 Changes in Farm Household Spending, 1997–2000

HouseholdTotal Operation Living Spending Spending on

Spending Spending Expenses on Food GarmentsYear (Yuan/Person) (Yuan/Person) (Yuan/Person) (Yuan/Person) (Yuan/Person)

1997 2,537 706 1617 890 1091998 2,457 652 1590 850 981999 2,390 600 1577 829 922000 2,652 654 1670 821 96

services by RMB18 or 11 percent. The above expen-diture increases, which were basically passive,increased by RMB60 or 65 percent of the totalincrease in living expenses. Therefore, even thoughfarmers’ per capita spending on living hadincreased, the amount of the spending that couldbe used to improve their quality of life was verylimited.

The data on per capita expenditures on produc-tion and living of farmers are only national aver-ages, and, very often, average figures conceal manydisparities. The situation was more serious in termsof the reduction in spending for most of the farmerhouseholds that mainly depend on farming forincomes. This is one of the major reasons for thedecline in farmers’ share of rural consumption evi-dent in table 5.5.

Thus it can be seen that the decrease and stagna-tion in farmers’ spending has actually producednegative impacts on the expansion of the domesticmarket and healthy development of the nationaleconomy. As the central government pointed out atthe Central Working Conference on Rural Issues,the issue of increasing tangible benefits to farmersis central to the continuation of the progressachieved in agriculture productivity. If the purchas-ing power of farmers does not continue to improve,the policy for expanding domestic demand will notachieve the desired results. If farmers’ living stan-dards cannot be improved, the risk of increasedsocial instability in the rural areas will increase. Ifthe ability of agriculture to save and accumulatecannot be constantly strengthened, it will be diffi-cult for China’s agricultural products to competein the global marketplace. If Chinese agricultureexperiences major setbacks, the development ofthe whole national economy and society will beaffected adversely.

WTO Accession and the NewStage of Development of ChineseAgriculture and Rural Economy

The Initiation of the New Stage and its Central Tasks

In October 1998, the Third Plenary Session of the15th Central Committee of the Communist Partyof China (CPC) adopted the Decisions of the CPCCentral Committee on Major Issues of Agricultureand Rural Work. The judgment of balance in aggre-gate and surplus in harvest years for grain andother major agricultural products was made in theDecisions. Based on this judgment, the central gov-ernment stated at the National Conference onRural Work at the end of 1998 that the develop-ment of Chinese agriculture and rural economyhad entered a new stage. The main basis for the newstage thinking is the major change in the pattern ofagricultural supply and demand. One of the keyfeatures of the situation is that the quantitativecontradictions in the supply of and demand foragricultural products have been relieved. Anotheris that past agricultural production had beenrestricted by natural resources; presently, however,restrictions of market demand on production havebecome more acute. What farmers are worryingabout is not whether enough products can be pro-duced, but whether the products can be sold atgood prices. Precisely because of such changes, thecentral government explicitly directed at the end of1999 that the central task for agriculture and ruraleconomy in the new stage should be implementa-tion of strategic restructuring.

The Crux of the Strategic Restructuring of Agriculture and Rural Economy

Since China adopted reform and open economicpolicies, Chinese agriculture has experiencedrestructuring. The starting point of the strategicrestructuring lies in the adjustment of variety andquality instead of quantity. The adjustment in thearea of crop sowing is no more than which cropshould have more sowing area and which shouldhave less. This kind of adjustment was effective inthe past, is effective at the present, and will be effec-tive in the future. However, adjustment in quantityhas major limitations. For example, judging fromthe existing production level and consumer demandfor grain in China, the total sowing area for grain


TABLE 5.5 Share of Farmers in ConsumerSpending at County and LowerLevels (percent)

Year Percent

1996 401997 391998 391999 392000 38

crops should be maintained at 1.65 billion mu to1.7 billion mu, with only about 50 million mu as amargin for adjustment.

Strategic restructuring differs substantially.Adjustment can be made on every piece of landwithin more than 2.3 billion mu of sowing area,such as replacing the unmarketable poor qualityvarieties with high quality marketable ones. Adjust-ment can be made not only on cultivated land, butalso in forest products, fruits, animal products, andaquatic products. Accordingly, adjustments empha-sizing optimizing varieties, improving quality, andefficiency can be made everywhere. Only by suchstrategic restructuring focused on quality and effi-ciency can the overall quality of Chinese agriculturebe improved to meet the requirements and chal-lenges posed by WTO accession and economic glob-alization, to ensure improvement in the people’sliving standards, and to accelerate the process ofagricultural modernization in China.

The Fundamental Objective of the StrategicRestructuring

The new stage of strategic restructuring was initi-ated early in 2001 when the central governmentspecified that the fundamental objective of thestrategic restructuring is to ensure increases infarmers’ income. After entering the new stage andbasically solving the quantitative contradictionin agricultural supply and demand, increasingfarmers’ income has increasingly become a priorityof the Chinese central government. In November2000, Secretary General Jiang Zemin and PremierZhu Rongji stressed at the National EconomicConference that increasing farmers’ income shouldbe a key goal in the entire economic work program.

The government has paid close attention to issuesconcerning agriculture, rural areas, and farmers,with the focus on increasing farmers’ income. Thefundamental objective set at the Central WorkingConference on Rural Affairs at the beginning of2001 was to achieve growth in farmers’ income.In mid January 2001, the central government con-vened a National Conference on AgriculturalScience and Technology, at which a group of agri-cultural scientists was commended and the Outlinefor the Development of Agricultural Science andTechnology in the Next 10 Years was adopted. InFebruary 2001, the State Council held a working

conference in Anhui Province on the reform exper-iment in rural taxes and fees to explore fundamen-tal solutions to the burdens on farmers. At the turnof the century, the central government held a seriesof major conferences on rural restructuring,increasing farmers’ income, and alleviating burdenson farmers.

Securing income growth for farmers is by nomeans a simple matter. The most importantimmediate cause of the slow increase in farmers’income is the current difficulty in selling agricul-tural products and the declines in prices. However,problems of product sales and price declines arenot only evident in agriculture, but in most otherindustries as well. Therefore, the current difficul-ties facing farmers’ income are to a great extentrelated directly to the economic cycles at bothhome and abroad. Achievement of sustainedgrowth in farmers’ income is a complicated projectof system engineering. One or two specific policiescannot be expected to provide fundamental solu-tions. The key is to make great efforts in all aspectsof agricultural and rural work. This also dependsto a large degree on the improvement in thenational economy as well as the world economy soas to create a broader market for the agriculturaldevelopment.

The Basic Direction of the Strategic Restructuringof Agriculture and the Rural Economy

As directed by the Chinese government, the centraltask of the new stage of agricultural development isto implement strategic restructuring of agriculturalproduction. This restructuring is strategic for tworeasons. The first is to put the long-standing prac-tice of pursuing quantitative growth in agriculturalproducts on the track of stressing optimized varietyand improved quality and efficiency. The second isthat the ongoing development of Chinese agricul-ture needs to continue to consider the constraintson the resource side and to pay more attention toconstraints from demand side. To achieve this, theregulation of agricultural production by marketforces should be expanded.

It can clearly be seen that WTO accession willaccelerate agricultural restructuring and adjust-ment in agricultural policies. The agriculturalrestructuring consistent with the above goals willrequire changes in the relevant policies. Although


these policy adjustments primarily resulted frommajor changes in the patterns of agricultural supplyand demand in China, they are fundamentallyconsistent with the direction of China’s WTO com-mitments. Even if no WTO accession issues are pre-sented, such adjustments in agricultural policieswould be required.

The restructuring of agriculture and rural econ-omy, revealed at the end of the ninth Five-YearPlan, has five major orientations:

1. Optimize varieties of farm produce andimprove the overall product quality. Optimizationof varieties and improvement in quality and effi-ciency should be the main orientation of adjust-ment for all agricultural production, including cropfarming, animal husbandry, fishery, forestry, andfruit production. With the issue of quantity in agri-cultural supply and demand resolved, optimizingvarieties and improving quality will constitute anobjective requirement for the current stage of agri-cultural development facing supply and demandchanges. It is preferable in agricultural productionto have smaller quantities of higher quality prod-ucts to avoid wasting the human, financial, andmaterial resources of farmers. To this end, the cen-tral government announced the termination ofpurchasing at protective prices for long-grainednonglutinous rice, northeastern spring wheat,southern winter wheat, and other unmarketablevarieties, starting in 2000. The purpose of thisreform was to guide farmers to replace inferiorvarieties with superior ones.

2. Expand processing of agricultural products.Agricultural processing will become more and moreimportant in agricultural development in China.With continuous increases in people’s incomes andliving standards, the Engel coefficients and incomeelasticity of demand can be expected to fall, whichcould pose a grave challenge to the continued devel-opment of agriculture.

Currently, the main factor restricting the growthof farmers’ income is the decrease of income fromagricultural production. This is related to the stageof social and economic development in China.Profound changes have taken place in the currentconsumption structure of residents. One is thedecrease of the Engel coefficient and the associateddecline in the income elasticity of demand for food.In 1991, the Engel coefficients of China’s urban andrural residents were 54 percent and 58 percent,respectively. These declined to 38 percent and48 percent in 2001.

The decline in the propensity of households tospend their income on food is evident in Table 5.6.Between 1996 and 2001, per capita disposableincome increased by RMB2 021 and consumerexpenditure increased by RMB1,390, while foodexpenditure only increased by RMB109, eventhough expenditures on food eaten outside thehome increased by RMB128. Thus, the urban resi-dent expenditures on grain, oil, meat, and vegeta-bles decreased by RMB131.

Only by expanding farm produce processingand encouraging the consumption of higherquality, more nutritious, and safer food can the


TABLE 5.6 Changes in Urban and Rural Resident Income and Spending, 1996–2001

2001 (Yuan) 1996 (Yuan)

Per capita disposable income 6,860 4839Per capita consumer expenditure (total) 5,309 3919

Food 2,014 1905Grain 188 272Oil 59 69Meat, poultry, and products 413 439Aquatic products 152 132Vegetables 194 207Tobacco 104 84Liquor and beverage 104 85Fresh and dried fruits 131 118Dining out 314 186

continued expansion of agricultural markets beachieved. Food processing can not only increase thevalue of agricultural products, but it also can guideconsumption and develop new markets to make itpossible to continuously increase farmers’ income.Two concepts should be altered in terms of devel-oping agricultural processing industry. First, foodprocessing should not be regarded as an industrywith low-technology content and without the needfor substantial investment. On the internationalmarket, brand-name and marketable processedfoods all have high-technology content. Theseproducts simply can not be produced without theapplication of advanced technology. Second, itshould not be believed that unsaleable fresh prod-ucts can be processed into high-quality processedproducts. The raw materials for food processingshould be special-purpose farm produce, such asspecial wheat and special grapes, which are distinc-tive from those directly used for food. Therefore,adjustments should be made in the varieties ofagricultural products to achieve better results fromfood processing.

3. Bring into play the local comparative advan-tages in agriculture. In the past, widespread short-ages of agricultural products left localities with no choice but to pursue higher degrees of self-sufficiency. Now that the supply and demand situa-tion has changed, local authorities should beliberalized from such traditional thinking andbring into play local comparative advantage inagriculture. Particularly, the coastal regions andsuburbs of large and medium cities should makeuse of capital and technology-intensive techniques.These localities should properly reduce their grainproduction and develop higher value-added animalhusbandry, fishery, and horticulture so as to givemore market space to the main grain-producingregions in a pattern of mutual benefit and relativecomplementarity. At the same time, the variouslocalities should be encouraged to develop theirown agriculture with distinctive features.

To promote restructuring of the regional distri-bution of agriculture, the central governmentdecided in 2001 to liberalize grain markets in themain coastal grain-consuming regions, includingBeijing, Tianjin, Shanghai, Jiangsu, Zhejiang,Fujian, Guangdong, and Hainan. Particularly, in theface of the challenges and opportunities after WTOaccession, much more attention should be paid to

giving play to local comparative advantage in agri-culture. China has a large population and littleland. Most of the land-intensive agricultural prod-ucts, such as grain, cotton, and edible oil are rela-tively high-cost in China. Most labor-intensiveagricultural products, such as animal and poultryproducts, aquatic products, and horticultural prod-ucts (fruits, vegetables, flowers, and bonsai), have acomparative advantage in international markets.Currently, there is an urgent need for Chinese agri-culture to foster strengths and circumvent weak-nesses and to give full play to China’s comparativeadvantage in international market competition. Totap the comparative advantage of Chinese agricul-ture on the international market, it is first necessaryto bring into full play the agricultural comparativeadvantages of different domestic regions. In thisway, it will be possible to create a pattern of mutualbenefit and complementarity between regions toimprove the competitiveness of Chinese agricul-tural products on international markets and topromote the development of China’s agriculture as a whole.

4. While continuing to support the develop-ment of TVEs, make positive and steady efforts toadvance urbanization. In the second half of theninth Five-Year Plan period, the per capita netincome of farmers increased, although incomefrom farming declined. This increase dependedalmost entirely on income growth from nonfarmsectors. This shows that the fundamental way toenrich farmers is to reduce the number of farmers.Agricultural resources are limited and effortsshould be made to transfer rural labor and popula-tion to allow the farmers remaining on the land toexpand their scale of production. The townshipand village enterprises (TVEs) in the countrysidehave provided 130 million job opportunities tofarmers, and their role in alleviating the unem-ployment problem can hardly be replaced. TVEsshould be guided to conduct restructuring andinstitutional innovation to promote their furtherdevelopment and to make a greater contributionto increasing the employment and income offarmers.

Another major source of income growth offarmers during the ninth Five-Year Plan was tran-sient employment of farmers in towns and cities.This part of income has provided an increasingshare in the net income of farmers. According to an


estimation by Sichuan, 6 million farmers in theprovince seek transient employment and remitnearly RMB30 billion each year. The transientemployment of farmers, which should be encour-aged, should be managed to keep it orderly. How-ever, many large and medium cities have set upvarious restrictions and rules on employing farm-ers, and many of these rules and regulations areboth unreasonable and ineffective. On one hand,these transient farmer workers should be guided toflow in an orderly manner and to abide by relevantlaws and regulations. On the other hand, the vari-ous rules and regulations restricting farmers seek-ing jobs in urban areas should be reviewed so as toalleviate the burden on the transient farmers. In theprocess of economic development in China, thetransient employment of farmers will certainly be along-term process. A large number of farmers willbe transferred to nonfarm employment, but thetransition to settlements in towns will occur overtime and will require an extended period of tran-sient employment. Conscientious research shouldbe conducted, and effective policies should beadopted to guide and manage such a large scale andlong-term population flow.

Developing small towns is an important way totransfer rural population and accelerate the processof urbanization in China. In the decisions adoptedat the Third Plenary Session of the 15th CentralCommittee of the Chinese Party Congress, it wasspecified that developing small towns would be amajor strategy for the development of the ruraleconomy and society.

There are two schools of thought on the specificways in which urbanization should be advanced.The first places the emphasis on large and mediumcities. Through a host of analyses and researches,many scholars think that large cities can generatehigher economic benefits and provide a large num-ber of job opportunities. The second school ofthought stresses the effect of developing smalltowns. Its advocates believe that the basic goal ofurbanization at the current stage is to speed up thetransfer of rural surplus labor and rural popula-tion, and these advocates view developing smalltowns as a more effective means of achieving thisgoal.

The path to urbanization should be one withChinese characteristics, featuring coordinateddevelopment of large and medium cities and small

towns. Medium and small cities should be devel-oped, the regional center cities should have distinc-tive features, and large cities should bring into playtheir leading role and growth pole functions.However, measured with the goal of transferringthe rural population, developing small towns ismore realistic. Presently, large and medium citiesfrequently provide job opportunities for transientfarmers and opportunities for them to accumulatecapital, but it is very difficult for these farmers tosettle in these cities. After a certain period of capitalaccumulation, rural residents typically leave thelarge cities, but not all of them return to rural areas.Some of them settle in small towns. This process ofresettlement will provide practical opportunitiesfor the development of small towns. However thedevelopment of small towns should not rush head-long into mass action and “blossom” everywhere,or the results will be counterproductive. There areabout 45,000 towns and townships in China, inwhich 19,000 are governmentally nominated town-ships. Considerations of how many small townsshould be established should proceed from realisticassessments of China’s actual conditions andnational strength. Towns with better basic condi-tions and development potential should be givenpriority in construction. The most important pre-condition for the development of small towns is aprosperous local economy.

Construction of small towns differs from that of residential quarters. A residential district canaccommodate several thousand residents and hassound living environments and complete livingfacilities; however, most of the residents are notemployed within the district, that is, residential dis-tricts have no function to create job opportunities.The basic function of small towns is to attract thetransferred farmers and provide job opportunities.Accordingly, emphasis should be laid on giving playto functions of towns; the basic function is to createa prosperous local economy and to provideemployment. The most basic judgment on whetherconditions exist for constructing a town is whethera planned town can become a source of economicgrowth. Only by making this judgment can weachieve the goal of accelerated transfer of rural sur-plus labor and population, provide better condi-tions for agricultural and rural development, andprovide more opportunities for income growth offarmers.


5. Intensify the construction of ecologicalenvironment and implement sustainable devel-opment. In ecologically vulnerable regions such asthose along the upper reaches of major riversand arid areas, the government should providefinancial aid for farmers to return grain plots toforestry to conduct ecological construction includ-ing re-vegetation and soil and water conservation.Starting in 1999, experiments on returning grainplots to forestry were conducted in some provincesand regions. By the end of 2001, 11 million mu ofgrain plots had been returned to forestry nation-wide. In 2002, the central government officiallyadopted policies for returning grain plots toforestry. The government will subsidize 300 jin ofgrain to farmers in the southern part of China and200 jin to farmers in the north for every 1 mu ofgrain plots returned to forestry, plus RMB50 foreach mu incurred in sapling and grass seedexpenses and RMB20 in living expenses. Under thispolicy, implementation of returning grain plots toforestry not only can recover balance of ecologyeffectively, but it also can offer opportunities forfarmers to receive direct benefits. By the end of2002, the total area of returning grain plots toforestry reached 34 million mu and is expected toreach 100 million mu during the whole tenth Five-Year Plan.

Policy Orientations for PromotingRestructuring of Agriculture andthe Rural Economy

To promote the strategic restructuring of agricul-ture and rural economy and to improve the over-all quality and international competitiveness ofChinese agriculture, the government is acceleratinga series of policy readjustments to meet the needsof current agricultural development. The mainreadjustments include the following:

• accelerate the establishment of market informa-tion systems for agricultural products andprovide farmers with timely, comprehensive,accurate, and authoritative market informationon agricultural products

• accelerate the establishment of quality, security,and standards systems, and improve the meansof inspection and testing for agricultural prod-ucts, particularly food products, guiding the

farmers to produce high-quality, low-residue,and safe farm produce

• adjust the existing supportive and protectivepolicies for agriculture and establish a domesticprotective system for agriculture that conformsto WTO rules

• accelerate the innovation of agricultural man-agement system; develop an integrated agri-cultural management system with “companyplus farmer households” as the main form toimprove farm organization and farmers’ accessto markets

• further reform and perfect the circulation sys-tem for grain and other major agricultural prod-ucts to reduce circulation expenses and improvecompetitiveness

• further restructure the rural financial systemand explore the establishment of an agriculturalinsurance system that suits the national condi-tions in China

• restructure the agricultural science and tech-nology research and extension systems andintensify international cooperation and technol-ogy import to accelerate technical progress inagriculture

• reform the household registration system in smalltowns and encourage the rural population thathas met necessary conditions to settle in smalltowns.

Remaining Policy Issues

A number of policy issues needs to be examined inmore detail in the future. Some of the most impor-tant of these issues are as follows:

• Correct analysis of price issues for staple agricul-tural products, particularly grain, soybeans, andcotton. The relationship among producer prices,wholesale prices, and retail prices should be dis-tinguished, and the actual gap between pricesfor China’s staple agricultural products andinternational market prices should be analyzedobjectively. Monopoly still exists in the circula-tion of staple agricultural products. Therefore, itis absolutely possible to reduce the circulationexpenses of China’s staple agricultural productsby further reforming the agricultural circulationsystem to reduce the gap between the Chineseand international market prices.


• The comparative advantages of Chinese agricul-tural products; labor costs give China a clear com-parative advantage in many animal and aquaticproducts. However, given the aggravated “greenbarriers” in international agricultural trade, lowlabor costs are not enough to constitute interna-tional competitiveness. The key is to find ways to produce agricultural products that meetinternational quality and safety standards. At thesame time, the question remains whether otherWTO members will open their markets toChinese agricultural products.

• Reforming the foreign trade system for agricul-tural products. The key is to properly resolvethe problems in giving nonstate firms thathave met the necessary conditions of foreigntrade rights for importing and exporting agri-cultural products.

• Both tariff concessions and tariff-rate quotas forimports of agricultural products are hard to bearin the transitional period. However, it is crucialto study how to deal with the pressures from

the international agricultural markets. Bothresearch and preparations for the new round ofWTO agricultural negotiations are inadequate.The central government should set up specialbodies to intensify research on the relevantpolicy recommendations.

References

“The Outline of the Ninth Five-Year Plan (1996–2000) forNational Economic and Social Development and the Long-range Objectives to the Year 2010.” Delivered at the FifthPlenary Session of the 14th Central Committee of the Com-munist Party of China. Xinhua News Agency, September 28,1995.

National Bureau of Statistics. Statistical Yearbook of China 2002.Beijing. China Statistics Press.

National Bureau of Statistics. Statistical Yearbook of China 2001.Beijing. China Statistics Press.

“The Decisions of the CPC Central Committee on Major Issuesof Agriculture and Rural Work.” Delivered at the Third Ple-nary Session of the 15th Central Committee of the Commu-nist Party of China. People’s Daily. October 15, 1998.

The Outline for the Development of Agricultural Science andTechnology in the Next 10 Years: 2001–2010. People’s Daily.May 24, 2001.


efficiency gains from decollectivization, the countrystill enjoyed agricultural growth rates that have out-paced the rise in population. New opportunities inthe off-farm sectors have allowed farm families toshift part of their household labor out of the agri-cultural sector into higher paying off-farm jobs.

Despite the healthy expansion of agriculture, thesector still faces serious challenges. According tothe World Bank (2000), more than 100 million peo-ple fell below the poverty line in the late 1990s,earning less than one dollar (U.S.) per day in pur-chasing power parity (PPP) terms. Some regionsof the nation remain highly dependent on cropproduction, including farmers in some Northeastprovinces (maize and soybeans) and the NorthChina Plain (wheat and maize) (Ministry of Agri-culture 2001). In recent years, expanding suppliesand increased liberalization have pushed real agri-cultural prices to their lowest levels in history (Parkand others 2002). With the retreat of the state

The initiation of economic liberalization and struc-tural change in 1978 has produced substantialgrowth in China’s economy. The annual growth rateof gross domestic product (GDP) was 8.5 percentfrom 1979–84 and 9.7 percent from 1985–95.Despite the Asian financial crisis, the GDP growthcontinued at 8.2 percent annually from 1996 to 2000.

Although reform has penetrated the entire econ-omy since the early 1980s, most of the successivetransformations began with, and to some extentwere dependent on, growth in the agricultural sec-tor (Nyberg and Rozelle 1999). After 1978, decollec-tivization, price increases, and the relaxation oflocal trade restrictions on most agricultural prod-ucts accompanied the substantial growth of theagricultural economy from 1978 to 1984. Grainproduction increased by 4.7 percent per year. Evenhigher growth was enjoyed in horticulture, live-stock, and aquatic products. Although agriculturalgrowth decelerated after 1985 after the one-time

81

6The Nature of

Distortions toAgricultur al

Incentives in Chinaand Implications of

WTO Accession

Jikun Huang, Scott Rozelle, and Min Chang

The authors would like to thank the research assistance of Yuping Xie. This work could not have been donewithout him. We also acknowledge the helpful comments and suggestions of Kym Anderson, Fred Crook,Tom Hertel, Elena Ianchovichina, Will Martin, Francis Tuan, and participants in the World Bank-sponsoredmeeting on the impact of WTO accession on China. We acknowledge the financial support of the WorldBank’s Trade and Rural Development units.

occurring in many sectors of the economy, such asrural health care and provision of welfare services(Nyberg and Rozelle 1999), large numbers of peo-ple remain poor and vulnerable to even relativelyminor income shocks.

Agriculture has been at the center of discussionof China’s entry into the World Trade Organiza-tion (WTO), due in part to the vulnerability ofparts of the rural economy and in part to theimportance of agriculture in the political economyof a number of developed nations with whomChina negotiated its accession to the WTO. How-ever, the likely shifts in China’s future agriculturalpolicy and their impacts are not well understood,and debates on the future of China’s agriculturecontinue. Some argue that the impact of WTOaccession on China’s agriculture will be substan-tial, adversely affecting hundreds of millions offarmers (Carter and Estrin 2001; Li and others1999). Others believe that, although some impactswill be negative and even severe in specific areas,the overall effect of accession on agriculture will bemodest (Anderson and Peng 1998). In part, theconfusion about the ultimate impact of WTOaccession on agriculture can be traced to a generallack of understanding of the policy changes thataccession will engender (Martin 2002). However,in perhaps an even greater way, the lack of clarityof the debate can be traced to a lack of under-standing of the fundamental facts about the natureof the distortions to China’s economy on the eve ofits WTO entry.

In this chapter, we provide an approach to helpresearchers and trade officials better understandhow trade liberalization will affect agriculturalprices and how the price changes will be experi-enced in different parts of the nation. First, wedescribe a new way of collecting data that can beused to design more accurate, disaggregated meas-ures of protection (specifically, nominal rates ofprotection NRPs). Then, we use price determina-tion and market integration analyses to study thenature of domestic markets to begin to understandhow price shifts at the borders that arise from tradeliberalization measures affect producers in differentparts of the country.

We report the results of a case study of theimpact of China’s WTO accession on agriculturalprices. Since many of China’s poor rural house-holds live in remote regions and tend to rely more

than any other group on the income earned fromcropping, they are the ones most likely to beaffected by the WTO’s liberalizing measures (Chenand Ravallion 2004). In conducting our case study,we focus on the effects of the WTO accession onagricultural prices even though the other effects onthe rural population are likely to be at least asimportant as cropping (Zhao and Sicular 2002). Wedo not want to quantify the total welfare effect;instead, we focus primarily on the qualitativeeffects on China’s farmers to illustrate how toapproach assessing the effects of trade liberaliza-tion on agricultural prices.

The rest of the chapter is organized as follows.First, after providing a brief context for our analysisand discussing our data, we present measures ofNRPs for a set of commodities for China. Next wediscuss how these distortions to agricultural pricesshould be expected to change as China implementsits WTO obligations and gains access (or not) to thepromises that were made to China. The next sec-tion analyzes the transmission of prices throughthe economy, and the final section discusses theimplication of the findings.

Gradual Opening and RemainingDistortions

Although agriculture has been at the center ofChina’s negotiations over its WTO entry, the likelyshifts in future agricultural policy and its impactsare not well understood. Debates on future foodsecurity are growing. The fundamental confusionabout the ultimate impact of WTO accession inpart can be traced to a widespread lack of under-standing of the policy changes that may result fromChina’s WTO accession (Martin 2002). Tradition-ally, analysts have focused on four sets of trade poli-cies, measures that are most frequently used byother countries to protect their agricultural sectors.In examining the previous work (e.g., CARD 2001;Tuan and Cheng 1999; OECD 2001), we find thatalmost all of the discussion is directed at tariffs,quotas and licensing, state trading, and traditionalnon-tariff barriers (NTBs). It is implicitly assumedthat the WTO agreement is focused solely on thesepolicies, that these policies are responsible for mostif not all of the protection that China enjoyed priorto accession, and that accession represents China’sinitial assault on protection at the border. In fact,


while at one time these policies were the source ofhigh distortions, some of the worst distortionscaused by these policies have already disappearedafter nearly two decades of reform in the externaleconomy. The experience of China’s agriculturaltariff policy illustrates the gradual but dramaticchanges that China has experienced in the past twodecades. In the late 1970s and early 1980s, thedomestic wholesale price of China’s four majorcommodities, converted at the official exchangerate far exceeded the world price measured atChina’s border (table 6.1). For example, in 1978,China’s rice price was 10 percent above the worldmarket price. The nation’s wheat and maize pricesexceeded the world price by approximately 90 per-cent. However, over the next two decades, the NRPfor rice fell to �6 percent and to about 30 percentfor wheat and maize. Although the NRP for soy-beans fell similarly between the late 1980s and themid 1990s, the rate rose in the late 1990s beforefalling to less than 20 percent in 2001 (table 6.1; Xie2002). During this period, the intervention by statetraders and the use of NTBs also fell gradually(Martin 2002).

Falling protection and changes in internationaltrade and domestic marketing policies have resultedin dramatically shifting trade patterns. Disaggre-gated, crop-specific trade trends show how exportsand imports increasingly are moving in a directionthat is more consistent with China’s comparativeadvantage. For example, the proportion of grain

exports in the 1990s, approximately 20 percent oftotal agricultural exports, is less than half of what itwas in the early 1980s (Huang and Chen 1999). Bythe late 1990s horticultural products and animal andaquatic products accounted for about 80 percent ofagricultural exports. These trends are even moreevident when reorganizing the trade data groupingthem on the basis of factor intensity (table 6.2). The

The Nature of Distortions to Agricultural Incentives in China and Implications of WTO Accession 83

TABLE 6.1 Changes in Nominal Ratesof Protection Over Time of China’s Major AgriculturalCommodities, 1978–2000a

Nominal Rates of Protection (Percent)

Rice Wheat Maize Soybean

1978–79 10 89 92 40

1980–84 9 58 46 44

1985–89 �4 52 37 39

1990–94 �7 30 12 26

1995–97 �1 19 20 19

1998–00 �6 26 32 49

1998 �6 22 40 37

1999 �9 30 33 67

2000 �2 26 23 44

aNominal rates of protection (NRPs) measured asdifference (percentage) between average borderprices and average domestic wholesale (market)prices.

Source: Huang 2001.

TABLE 6.2 China’s Agricultural Trade (US$ millions) by Factor Intensity, 1985–97

Land Intensive Labor Intensive Labor/Capital Products Products Intensive Products

Value Share Value Share Value ShareYear (US$ Million) (Percent) (US$ Million) (Percent) (US$ Million) (Percent)

Agricultural exports1985 2,119 36.4 2,199 37.8 1,497 25.71990 1,689 17.7 4,971 52.1 2,881 30.21995 875 6.0 7,095 48.4 6,704 45.71997 2,158 14.1 6,538 42.6 6,642 43.3

Agricultural imports1985 1,072 43.8 680 27.8 695 28.41990 4,032 71.9 642 11.5 935 16.71995 6,575 54.5 3,278 27.2 2,216 18.41997 4,644 47.3 2,179 22.2 2,987 30.5

Source: Huang and Chen 1999.

imports of land-intensive bulk commodities, such asgrains, oilseeds, and sugar crops, have risen (or netexports have fallen); while exports of higher-valued,more labor-intensive products, such as horticulturaland animal (including aquaculture) products, haverisen (table 6.2).

Nontraditional Sources of Trade Liberalization or Protection

Based on the preceding discussion, two factsbecome clear. First, distortions have declined sig-nificantly in the past 20 years. Considering this fact,the current episode of policy reform that accompa-nied China’s accession to the WTO should be con-sidered an extension of past efforts. Second, muchof the falling protection has come from decentraliz-ing authority and relaxing licensing procedures forsome crops (e.g., moving oil and oil seed importsaway from state trading firms), reducing the scopeof NTBs, relaxing real tariff rates at the border,and changing quotas (Huang and Chen 1999). It isperhaps for these reasons that much research onChina’s entry into the WTO focuses on the policiesresponsible for much of the earlier progress andstudying these policy tools might be beneficial.Undoubtedly, changes in China’s tariff regimes,state trading system, and matrix of NTBs will play acontinuing role in creating or eliminating distor-tions in China’s agriculture. However, partlybecause many of the gains from traditional tradereforms have already been experienced, China maybe able to affect further trade liberalization viaother, less discussed policies. Even if traditionalpolicies are still important, the gains from theseother policy reforms may be as important as thosefrom traditional trade reform.

For example, China also has used its taxationpolicy to protect its agriculture, especially in certainsectors, such as soybeans, that have been most liber-alized. In the early 1990s, leaders radically revisedChina’s fiscal system, making it much more reliantfor revenue generation on a value-added tax system(Nyberg and Rozelle 1999). The theory of the tax isthat it is assessed on value added in all goods duringtheir manufacture and sales process from the timethe raw material comes out of the ground until itreaches the consumer. National regulations statethat imported goods that are not for immediate

re-export also are to be assessed the value-addedtax. Although rates vary, the typical value-added taxrate ranges from 13 to 17 percent.

For a variety of political and tax collection rea-sons, in the early stages of the implementation ofthe tax, authorities decided to exempt farmerswhen they sold their products into the market.Traders that purchased grain, for example, from afarmer in his home or in a local market would nothave to pay the value-added tax. When the goodwas resold in a downstream wholesale market, thevalue-added tax was assessed, but the trader onlyowed the tax on the amount of the marketingmargin, or the difference between the procurementprice and sale price. Xie (2002) and Rozelle (2003)found that in China’s competitive marketing regime,the marketing margins (the difference between thebuying and selling prices) are extremely low. Forshipments from markets that are 500 to 1,500 kilo-meters from the port, traders report margins thatrange from 10 to 20 percent. Assuming an averageof 15 percent, the real tax rates on domestic agri-cultural goods are only 15 percent of those onimported goods.

Such a tax system can provide some of China’sfarmers with significant protection. For example, inthe case of soybeans, there is only a 3-percent tariffon imports. In recent years, however, many tradershave been given the right to import soybeans. Theo-retically, then, the international price of soybeansupon arrival at China’s borders should differ by only3 percent. However, as soybeans enter the country,soybean importers also must pay a 13-percent tax tomeet their value-added tax obligation. Domesticsoybeans, in contrast, are taxed, on average, at lessthan 3 percent. Through this means, then, China’ssoybean producers have an added 10 percent ofprice protection.1

China also has used export subsidies in recentyears to increase exports of some commodities,thereby increasing protection by raising the priceof domestic commodities (table 6.3). Maize andcotton are the two crops that have received the mostsubstantial export subsidies. In general, the subsi-dies are paid out of the national budget on approvalof the State Council. Only approved state tradingorganizations are authorized to export with thesubsidies. During interviews in the field during2001, we found that maize exporters, especially



those in Northeast China, received subsidies thataveraged 34 percent of export price (row 3). Forexample, one trader said that for each ton of maizethat his company exported in 2001, it received back378 yuan per ton (45.7 US$ per ton) after produc-ing an export bill of sale with the export sales price.With a sale price of 104 US$ per ton, the tradingcompany received a subsidy of 44 percent, 10 per-centage points above the average. The total pay-ment received (export earnings plus subsidy) was1,240 yuan per ton, which was about 90 yuanhigher than what could have been earned in thedomestic market (1,150 yuan). We also discoveredthat cotton exporters received fairly large subsidieswhen they exported raw cotton, up to 10 percent ormore (table 6.3). Finally, in several isolated cases,rice exporters reported that they received smallsubsidies (though no more than 5 percent for anysingle trade) from municipal and prefectural gov-ernments, a subsidy that we only documented insouth China (although it should be noted that wedid not have the opportunity to interview manyrice traders in north China—that is, ones thatmight be exporting japonica varieties into thenortheast Asia market). Most of the rice traders wespoke with received no subsidy for their exports,meaning the average subsidy almost certainly wasless than 1 percent (table 6.3).

Because of the higher numbers of meat exportersand the private or commercialized public natureof the companies, subsidies are not provided;

however, tax policies help exporters compete on anequal basis in international markets. Based on thetrade ministry’s estimate of the average value-added tax paid on the products exported by meattraders, when a meat exporter executes a contract,the company can receive a rebate equal to the esti-mated value of the value-added tax (table 6.4).For example, pork and beef exporters receive arebate equal to 5.2 percent of the value of theirtransaction (rows 4 and 5). Poultry exportersreceived 13 percent rebates (row 6). When interna-tional prices for meat commodities are higher thanthose inside China, rebates encourage domesticproducers and traders to export, since the perceivedprice gap between the domestic and internationalprice will be greater (which, of course, will also helpraise China’s domestic price as exports rise).

In summary, then, as China enters the WTO,those officials interested in liberalizing China’s tradewill face a number of challenges. Alternatively,China also has a number of instruments that it hasbeen using and may continue to use (legally or not)in managing its domestic economy. In addition totraditional trade policies, tariffs, quotas and licens-ing, state trading, and NTBs, China has protectedand/or has the potential to protect its agriculturewith a number of other policy measures. In partic-ular, our analysis has shown that taxation policymay still be a tool that China could try to use toprotect or further open its agricultural sector. Itmay also use other policies, such as sanitary andphytosanitary (SPS) regulations. In the past, it alsohas used export subsidies and rebates (when usedfor commodities on which the domestic value-added tax is not assessed) to create wedges betweenthe domestic and international prices of importablecommodities and to decrease the domestic price rel-ative to the world price of exportable goods. Whileexport rebates can still be used legally (assuming thetaxes that are being rebated have been paid domesti-cally), some instruments will not be able to be usedin the future (such as, export subsidies).

New Estimates of China’s NominalRates of Protection

In this section, we estimate a new set of NRPs onthe eve of China’s accession to the WTO. Theseestimates will attempt to overcome some of the

TABLE 6.3 Subsidies and Tax Rebates forExports of Selected AgriculturalCommodities in China, 2001

Rebate of Value-Added Tax

for ExportedExport Agricultural

Subsidies CommoditiesCommodity (Percent) (Percent)

Rice �1 0Cotton 10 0Maize 34 0Pork 0 5.2Beef 0 5.2Chicken 0 13

Source: Authors’ survey.


previous problems faced by researchers. In particu-lar, we try to understand in a more disaggregatedway the part of certain markets (in terms of vari-eties or commodity type) that China is protecting.Such an analysis should help us more accuratelyassess what the impacts will be after China imple-ments it WTO obligations. To do so, we first explainhow we collected our data. Next, we look at the dis-aggregated results. Finally, to make the informationmore useful to policy makers and other researchers,we create a series of more aggregate NRPs. Theaggregation of our disaggregated NRPs into a singlecrop-specific figure allows us to assess how ourmethods would have changed had we used tradi-tional methods of estimating NRPs. Appendix Asummarizes some of the difficulties that practition-ers face when trying to measure NRP for China’sagriculture.

To overcome the shortcomings of previous NRPstudies, we conducted a set of interviews and surveyswith the stated goal of precisely identifying thedifferences in prices at a precise point of time and aparticular location between an imported good onone side of the border (outside China) and a domes-tic good of identical quality on the other side (insideChina). Similarly, we also wanted to identify thesame price gap between exportable domestic goodsas they leave the country and the same goods fromother countries that are being traded in interna-tional markets. Conducted in 2001, the enumerationteam was in the field more than three months, fromAugust to November. The team visited seven coastalcities—Guangzhou, Shenzhen, Ningbo, Shanghai,Lianyungang, Qinghuangdao, and Dalian—and thetwo more inland cities of Beijing and Changchun. Ineach port, a number of “sampling frames” was usedto select a sample of domestic traders, importersand exporters, wholesalers, grain and oilseed users,trade regulators, agents, and other grain and fiberofficials. More than 100 people were interviewed;2

less than 10 percent of those contacted refused to beinterviewed.3

Dissagregated NRPs for Selected AgriculturalCommodities in China

The results of our analysis clearly illustrate theproblems with a strategy of NRP estimation thatattempts to come up with a single rate of protection

for a commodity. For example, it is difficult toprovide one single NRP of wheat in China, one ofthe world’s largest importers of wheat (table 6.4).Traders reported that the price of very high qualitywheat from North America was 20 to 50 percenthigher in the domestic markets of China’s majorports than when it was on a ship in China’s portready to be brought into the country (table 6.4).More precisely, the average trader told us that if aton of Canadian Number 3 hard white wheat werebrought in and auctioned off in October 2001, thecompetitive bid price would have been 20.5 percenthigher than the international price on a cost, insur-ance, and freight (CIF) basis. The same traders toldus that the competitive bid price of high qualityU.S. wheat would have been 50 percent higher.Hence, based on this price gap, one would have toassume that China’s protection price is high, and ifit were to open its markets completely, wheat priceswould plummet and import volume soar.

However, traders were quick to point out thatthey did not think that even with open marketsChina’s wheat price would fall anywhere near50 percent even if there was no effect on the worldprice; they were not considering the impact ofChina’s imports on the world price.4 According toour interviews, the market for baking-qualitywheat, the main use for hard white wheat fromNorth America, is actually relatively small in China,at most only several million metric tons (MMTs).We also were told that few users in China outsidethose who demanded flour for making cakes, pas-tries, and high-quality breads would use this type ofwheat and that only a small group of farmers andprocessors inside China could supply this type ofwheat. If this is in fact the case, this would meanthat even in a world free of any trade restrictions,imports would come into China until demand wasfilled and the domestic price for that variety fell tointernational levels.

Alternatively, it could be the case that all pro-duction of that particular variety would shift out ofChina if all of China’s farmers abandoned thembecause they could not make a profit. In such anextreme case, with few domestic supplies and withlittle or no substitution of the baking-quality wheatfor other domestic uses, the price impact on mostdomestic producers would be small. Growers of thehigh-quality wheat would lose; they would have to

keep growing at a lower price, switch to anotherwheat variety, or change cropping patterns. Sincethe quantities of such grain are so small, however,the overall impact would be minimal.

While not as extreme as the case for NorthAmerican baking-quality wheat, traders reportedthat arbitrage possibilities in other markets(table 6.4). With a remarkable degree of consis-tency, the CIF price of medium-quality wheatimports from Australia, England, and the PacificNorthwest of the United States (hard red) wasreported to be 10 percent lower than the price thatthey believed the same wheat would command inChina’s domestic market. Used for more commonbread, cheaper pastries, industrial uses, and high-quality noodles, interviewees believed that thismarket accounted for approximately 10 to 15 per-cent of China’s wheat demand. However, unlike thecase of the highest-quality baking wheat, produc-tion was greater in China. In fact, in 2001 domesticproducers supplied most of the wheat of this qualityinto this segment of China’s wheat market. In

China’s domestic market, however, this wheat wasconsidered to be high-quality Chinese wheat. Inter-estingly, evidence that medium-quality wheat oninternational markets is the same as high-qualitywheat supplied by China’s farmers is found in theanswer to the question that we asked our intervie-wees: if China’s higher-quality wheat were soldon international markets, how much loss woulda trader incur? Our survey found that this rate,10 percent, was almost exactly the same as thepremium importers would make from bringingin medium-quality grain from the internationalmarket.

Finally, although there have been no imports oflow- or lower-medium quality wheat from inter-national markets, it appears that China’s medium-quality wheat, by far the biggest part of China’sproduction (estimated to be more than 60 per-cent) is only marginally protected (table 6.4). Oursurvey found that traders believed if China’smedium-quality wheat was sold on the interna-tional market in late 2001, it would sell at a discount


TABLE 6.4 Disaggregated Nominal Rates of Protection for Selected Grains inChina, October 2001

Comparable Domestic Border PricesPrice (US$/Ton)

Variety or Quality Yuan/Ton US$/Ton C.I.F F.O.B NPR (%)

Estimated at official exchange ratea

Rice Weighted average �3Thai jasmine rice 3,690 446 380 17High-quality japonica 2,930 354 398 �11Medium-quality indica 1,519 184 185 �0.5

Wheat Weighted average 12US DNS 2,350 284 190 49Canadian #3 1,800 218 181 20Australian soft 1,625 196 175 12US hard red 1,550 187 169 11UK 1,350 163 145 12China high quality 1,350 163 145 12China medium quality 1,250 151 140 8China low quality 1,100 133 133 �0.1

Soybean Common variety 1,950 236 205 15Maize Common variety 1,150 139 105 32

C.I.F. Cost, insurance, and freightF.O.B. Free on boardaThe estimated official exchange rate is 8.28.Source: Authors’ survey.


of about 8 percent. Another way to interpret thisresult is that if international traders can ship thisquality of wheat to China, it would command apremium of 8 percent. It is likely that imports ofwheat will increase after WTO accession becauseof the persistent price gap. The effect, however,appears to be less than 10 percent. China’s lowestquality of wheat (about 10 to 15 percent of itsharvest) is at the world’s feed wheat price (table6.5). China did export some feed wheat intointernational markets in 2001 (mostly to Asia,according to an interviewee). Similar differencesin the size of the price gap among varieties of asingle grain are found for rice, though not forwheat and maize, which are more homogeneousproducts.

New NRPs for China

Although differences exist among major typesof any individual agricultural commodity, byweighting them by their sown area (for crops) andproduction (for meats) shares, a set of by cropaggregate NPRs can be created (table 6.5). Wheat,for example, has an NRP of 12 percent (row 1)when the individual NRPs from table 6.4 are

weighted by their area shares. On average, the pricesof all varieties of domestically produced wheat soldin the domestic markets of China’s major port (andinland) cities are 12 percent above the average CIFprice of all types of imported wheat varieties. Rice,on the other hand, is implicitly taxed at 3 percent.The aggregate figures, although helpful (and per-haps needed for analysis that is only disaggregatedto the crop level, are less interesting and providemuch less insight into which groups of farmers inwhich areas are producing which varieties will behurt or helped if trade liberalization reduces thedistortions.

However, to the extent that certain commoditieshave less intracrop quality differences, the aggregatemeasures have more inherent interest. For example,maize, soybeans, cotton, and sugar have far lessquality differences among varieties than rice andwheat. This is partly due to the fact that maize andsoybeans are rarely consumed directly (as are riceand wheat, making them more sensitive to humantastes and preferences). Instead, maize and soybeansare used mostly as feed or are processed. As aconsequence, in our analysis we only examineaggregate crop NRPs for maize, soybeans, cotton,and sugar.5

Our findings show not only that significantlypositive rates of protection exist for a number ofChina’s major field crops, but also that they varyover the nation and according to the position inwhich China finds itself as a net importer or as anet exporter. Maize prices, according to exporters,were more than 30 percent, on average, aboveworld prices. In other words, they would have lostmore than 30 percent of the value of their ship-ment if the government had not paid a subsidy.Protection rates for maize as an import differed.For example, traders in the northeast told our sur-vey team that if they were not exporting and for-eign maize was to come into China, the importercould gain 21 percent since the price of importedmaize is 21 percent lower than the domestic maizeprice. Our interviews in south China, however,found that the price gap between imported maize(CIF) and maize being traded in the domesticmarket in and around Guangzhou was more than30 to 40 percent. Aggregated across areas on thebasis of meat consumption shares, we estimatethat China’s maize NRP was 32 percent in 2001(table 6.5).

TABLE 6.5 Average Nominal ProtectionRates for Major Imports andExports in China, October 2001

NominalMajor Imports Domestic Price Rate of Protectionand Exports (Yuan per Ton) (Percent)

Imports

Wheata 1,250 12

Maize 1,150 32

Soybeans 1,950 15

Cotton 9,500 17

Sugar 2,612 40

Exports

Ricea 1,954 �3

Porka 11,442 �30

Beefa 13,743 �10

Poultrya 9,904 �17

Fresh fruits 5,472 �4

aAverage Nominal Rates of Protection are created bysumming the NRP rates of individual varieties weightingwith the sown area (production) share.

Source: Authors’ survey.


Interviewees also reported that despite the largevolume of increase of soybean imports in recentyears, a difference remains between the CIF anddomestic price in the port (table 6.6). The averagedifference between the domestic price and theinternational price was 15 percent. In one sense, thefact that there is a remaining price gap is remarkablegiven that China imported almost 15 MMT ofsoybeans in 2001, the official tariff is only 3 percent,and the commodity is freely traded without secur-ing a license or quota allocation. On the other hand,the remaining price gap reminds us that there maybe other reasons for distortions beyond tariffs andstate trading.

Our results also demonstrate that cotton andsugar were fairly highly protected in October 2001(table 6.5). The case of cotton, however, is an exam-ple of how fast the NRP can change across time. TheNRP was measured at 17 percent in October 2001.When our team returned for follow-up work at theend of November, the domestic price of cotton hadfallen from 9,500 yuan per ton in October to lessthan 8,000 yuan per ton. With this fall, the NRPwent to less than zero. However, later in the year, theinternational price of cotton also dipped, a fact thatwould lead to a higher NRP. Being less variable in2001 in both China’s and international markets, the

NRP of sugar remained about 40 percent through-out the year.

Assessing the New Methodology

Since one of our objectives was to use a new datasource and method for aggregating NRPs data togenerate crop specific NRPs, it would be interestingto analyze what would have happened had we notused this time- and data-intensive survey method-ology. To conduct such an experiment we used thesame methodology, data sources, and assumptionsthat many people use for calculating NRPs tocalculate an NRP for China in 2001.6 Although thetwo approaches give almost the same answers forsome commodities, such as soybeans and maize(though soybeans were still overstated, in partbecause of difference in prices over the entire year—China’s domestic prices fell sharply over the year,suggesting that the NRP in late 2001 was lower thanit was in early 2001), the answers vary considerablyfor other commodities. For example, the nationalaverage price for wheat in 2001 reported from theMinistry of Agriculture reporting system was1,113 yuan per ton. The average price of importscalculated by dividing total import value by totalimport quantity was 1,393 yuan per ton. In otherwords, the domestic price of wheat using thesesources of data about prices is 21 percent below theCIF price of imports. From this standard methodol-ogy, one would come to the conclusion that wheat,rather than being protected (by 12 percent—seetables 6.4 and 6.5), was actually being taxed by trad-ing policies. Yet, as we have seen the main reason forgenerating a negative rate of protection is thatChina is importing almost exclusively very high-grade, baking-quality wheat, while its domesticconsumers use mostly medium- and lower-qualitywheat. The wrong conclusion is reached when oneuses the specialty prices for imports as an interna-tional reference price for types of wheat that aremuch lower quality and are lower priced.

The same problem is found for rice. BecauseChina imports only high-quality jasmine rice fromThailand, the international price of rice (3,908 yuanper ton—that is calculated by total import valuedivided by total import quantity) appears to be morethan 150 percent higher than the average domesticprice (1464 yuan per ton). In fact, as shown intable 6.5, China’s average price protection (tax) rate,

TABLE 6.6 Percentage of Market Pairs that Test Positive for BeingIntegrated based on Dickey-Fuller Test in Rural China, 1988 to 2000

Commodity 1989–95 1996–2000

(Percent of Market Pairs)Maize 28 89Soybeans 28 68Japonica Rice 25 60

(Yellow River Valley)

Indica Rice 25 47(Yangtse Valley and South China)

Note: Results for two periods from same dataset. For results from 1989–95 for maize andrice, see Rozelle and others (2000). Rice resultsare for the whole country in 1989–95. Resultsfrom soybeans for 1989 to 1995 and all resultsfrom 1996 to 2000 are by authors.

calculated on a variety-by-variety basis, is almostzero (�3).

Thus, according to this illustrative example, wecan see the necessity of approaching the estimationof NRPs in a more careful way for some commodi-ties. Using the traditional approaches work fairlywell for commodities that are relatively homoge-nous in their quality characteristics, such as maizeand soybeans. We have seen for the case of wheatand rice for China in 2001, however, that compar-ing average prices inside and outside of the nationcan lead to misleading results. Based on this exam-ple, one might conjecture that traditional estimatesof NRP for some products, such as sugar and edibleoils, may be fairly reliable. Those for meat products,cotton, and horticulture crops, however, could bemisleading.

WTO Effects Away from the Border

While important in determining the size of theshock at the border, the broader magnitude of theeffect of the WTO agreement on China’s farmersdepends not only on the size of the distortion, butalso on the size of the area across which it willbe felt. This second factor, in turn, is a function ofthe size and nature of China’s market. In fact, thereare at least three factors:

• policy safeguards that limit market forces fromfully equilibrating domestic and internationalprices

• household responses by which households areable to move into the production of higherprofitability commodities and away from thosethat experience price falls

• high transaction costs that possibly can serve tobuffer the effects of liberalization policies onthose who live in rural areas in China.

The policy safeguards are discussed elsewhere inthis volume. The effects of household responses arediscussed in Taylor (1998) and Huang and Rozelle(2002). In this section, we focus on the nature ofmarkets.

Ultimately, the distributional impacts of WTOwill depend on the nature of China’s markets. Iflarge areas of the country are isolated from coastalmarkets where imports land, then the effects ofWTO may be circumscribed to restricted parts of

the country and should not be expected to havehighly adverse impacts on the poor who are largelylocated in inland areas far from major urban cen-ters. While being isolated from negative externalshocks is a plus, there is also a cost. Those livingin poor, isolated areas also would not benefitfrom price rises and opportunities to export, andthey are potentially vulnerable to price shocks asregional production and consumption demandchange. However, if markets exist that link togetherdistant regions with the coast and price changes inone part of the economy quickly ripple through theeconomy, even though imports are infused into(and exports flow out of) areas concentratedaround a few large coastal cities, they could haveramifications for poor households thousands ofkilometers away.

To the extent that there are high transactioncosts inside China and to the extent that certaindomestic markets are isolated from others in thecountry—especially those inland areas isolatedfrom port regions where imports land—the impactof WTO policies may not be evenly distributed. Inprevious work done on China’s agricultural mar-kets (e.g., Park and others 2002), it was found that,in general, China’s markets were becoming fairlyintegrated by the mid-1990s. However, this con-clusion should be qualified. First, although mar-kets improved greatly during the early 1990s, theanalysis still found large parts of the country,especially poorer areas, were not completely inte-grated into national markets in all years. More-over, the study’s dataset is dated. Since the finalyear of the available data, more than seven yearshave passed. It is unclear whether markets havematured since that time or whether the actionstaken by leaders have led to greater fragmentation(Nyberg and Rozelle 1999). Surprisingly, given thefragile nature of reforming China’s agriculturalmarkets, almost no recent work addresses thesequestions.

Assessing the Determination of Price and MarketIntegration in China

To assess the state of the integrated and developedmarkets in rural China in the late 1990s and 2000,we first describe the data. Second, we test for inte-gration and conduct direct tests of how well pricesin different markets move together and if prices are



integrated between the market town and China’svillages. Finally, we measure the degree of pricetransmission.

Data The data come from a unique price datasetcollected by China’s State Market AdministrationBureau (SMAB). Nearly 50 sample sites from 15 ofChina’s provinces report prices of agriculturalcommodities every 10 days. The prices are the aver-age price of transactions that day in the local ruralperiodic market. The Ministry of Agricultureassembles the data in Beijing, making them avail-able to researchers and policymakers.

We examine rice, maize, and soybean pricesfrom 1996 to 2000 (except for maize that was onlyavailable only through 1998). The three crops areproduced and consumed in nearly every provincein China. Rice price data are available for 31 mar-kets. Because of quality differences among rice vari-eties in different regions of China, we look at priceintegration between markets within four regions,South China, the Yangtse Valley, the North ChinaPlain and Northeast China. For the provincesincluded in the sample, rice prices are available forover 90 percent of the time periods. Prices formaize and soybean data are available for 13 and 20markets, respectively.7 Product homogeneity makesit possible to include a broader geographic range ofbuyers and sellers in a single analysis, and we areable to assess the integration of markets spread outover thousands of kilometers. We compare theseresults to results from 1988 to 1995 that were pro-duced with the same data and published in Parkand others (2002).

Integration Tests In this section we use more for-mal tests of market integration. To do so, we applythe Engle-Granger cointegration approach to testfor the integration of China’s grain markets. Two ormore price series are cointegrated (even if each isindividually nonstationary) if a linear combinationof the variables (e.g., the differences of the prices) isstationary. Following Engle and Granger (1987), weapply a two-step residual-based test. The first stepuses the ordinary least square (OLS) regression ofone price series on another:

P it = α + λt + β P j

t + et , (6.1)

where t is the common trend of the two price-series and where et is the error term. The main rea-son for running the first step is that it provides the

residuals, et , for the second step. The second stepthen tests for the stationarity of the residuals fromequation (6.1) using the augmented Dickey-Fullertest:

�et = δet−1 +N∑

j=2

γ�et− j + ξt (6.2)

If the test statistic on the δ coefficient is less (i.e.,more negative) than the relevant critical value fromthe Dickey-Fuller (D-F) table, the null hypothesismay be rejected and the two series are said to becointegrated of order (1,1). When the series arecointegrated this way, it implies that the two mar-kets from which the price series come are inte-grated. The absolute value of the test statisticsshould be greater than 3 at 10% significant level. Inour paper, we are conducting only the unit roottest, where j equals zero, since the error term fromequation (6.1) is an AR(1) process.

The results of the cointegration analysis supportboth our descriptive findings and the conclusionsof the determinants of commodity price analysis inthe previous page, especially when they are com-pared to the findings of research on market inte-gration in the late 1980s and early 1990s (table 6.6).In middle part of the reform era (1988–1995), atime when markets were starting to emerge,between 20 to 25 percent of markets showed signsthe prices were moving together during the studyperiods and sub-periods. According to their find-ings, although there were many market pairs inwhich prices did not move together, between thelate 1980s and mid-1990s, there was evidence ofrising integration.

Using the results from the early 1990s as a baseline, our current analysis shows that during the late1990s, China’s markets continued along their pre-vious path of maturation; markets in China, espe-cially those for maize, are remarkably integrated.In the late 1990s, examining the co-movement ofprices between pairs of markets in our sample, wesee a large increase in the number of integratedmarkets. In the case of maize, for example, in89 percent of the cases, prices in one market moveat the same time as in another (table 6.6). This is upfrom only 28 percent of the time in the early 1990s.The number of pairs of markets for soybeans,japonica, rice, and indica rice shows similarincreases (table 6.6). The integration of thesemarkets is notable because, in many cases, the pairs

of markets are separated by more than a 1000 kilo-meters. For example, we find prices in almost allyears to be integrated between markets in Shaanxiand Guangdong provinces and between those inSichuan province and southern Jiangsu.

Despite the significant progress in terms of inte-grations, our results also show pairs of marketsduring different years that are not integrated. Forexample, in one third of the cases, japonica riceprices moved in one market but did not in another.The case of indica rice trade is even more notable.In the case of more than half the time (and places),prices do not move together in China’s indicarice producing and consuming regions. Oneexplanation for such a result is that there issome kind of institution (policy or infrastructure/communication) breakdown that is creatingChina’s fragmentation, as shown in Park and others(2002). It is also the case that since every provincein China has rice production and consumption, ifduring a certain year in a certain area, supply inthat region is just equal to demand and price differ-entials between regions stay within the bandbetween regional “export” and “import” prices,moderate price movements in another area maynot necessarily induce a flow into or out of theregion that is in equilibrium.

Even with the nontrivial number of cases inthe late 1990s in which market prices in pairs ofmarkets do not move together, based on each ofthe market performance analyses, one must con-clude that the impacts of WTO on China’s agricul-ture will be experienced across wide regions of thenation from coastal to inland areas. However, this isonly half of the story. The discussion of table 6.6demonstrates a remarkable degree of integrationbetween markets on the coast and those inland,such an analysis is still not sufficient to ensurethat many of China’s villages will be affected bythe shocks that hit the coast and are transmittedinland.

To do so, in this part of this section we examinethe extent to which villages are integrated intoregional markets. Our test of integration will essen-tially identify if farmers in China’s villages are pricetakers or are villages isolated, making prices deter-mined by local supply and demand. In briefest terms,if variables that affect local supply significantlyaffect prices, we will assume villages are isolated and

markets are not integrated to the village level; in con-trast, if the local supply shock does not affect theprice, villagers are price takers and markets will bethought to be integrated.

The data for this study were collected in a ran-domly selected, nearly nationally representativesample of 60 villages in six provinces of ruralChina the China National Rural Survey (CNRS).To accurately reflect varying income distributionswithin each province, one county was randomlyselected from within each income quintile for theprovince, as measured by the gross value of indus-trial output. Two villages were randomly selectedwithin each county. The survey teams used villagerosters and our own counts to randomly choosetwenty households, both those with their residencypermits (hukou) in the village and those without.A total of 1,199 households was surveyed. TheCNRS project team gathered detailed informationon both the production and marketing behavior ofall of the farmers in the sample and the character-istics of each village and its relationship to thenearest regional market. From each respondent inthe survey in each village, we identified the priceand timing of the sale for each commodity. Fromthese data, we construct an average village pricefor each month in yuan per kilogram. In a com-munity questionnaire, we know how far the vil-lage’s center is from the nearest paved road and thedistance to the county market in kilometers.Finally, for each crop that the farmer cultivated, weknow whether the farmer’s crop suffered a shock,recording both the incidence and the percentageby which the yield fell. We do not include any vari-able that controls for the presence of a communitybuffer stock system, primarily because such aninstitution is almost never observed in modernChina.8 In addition, sales among farmers within avillage are rare (according to our data, less than5 percent of sales). County dummies hold qualityconstant.

Our regression analysis clearly shows marketsin China are integrated down to the village level(table 6.7). The signs and level of significance of thecoefficients on variables, such as the distance that avillage is from the market, demonstrate that thefarther a village is from a market, the lower theprice the farmer receives, which is the expectedresult. More importantly for our purposes, the



t-ratios of the coefficients of the village supplyshock variables are all less in absolute value termsthan 1.35, signifying that the output of the localvillage’s crops do not affect the local price. Oneimplication of this result is that factors outside thevillage affect the price that farmers receive, makingthem price takers. In other words, farmers, even inChina’s remote villages, are linked to the markets ofits main commodities.

Price Transmission Coefficients

While integrated markets mean that inland marketswill experience price changes in the direction of theprice movements at the ports (for those crops thatare imported and exported), frictions in the mar-keting system may shield inland producers fromsome of the effects.9 Moreover, despite the rise ofprice integration in China’s domestic markets dur-ing the reform period, there are still a significantnumber of market pairs during certain years thatdo not move together. Hence, when assessing theimpact of WTO-induced price shocks at the borderon farmers inland, we need to examine the degreeof transmission of these effects.

To examine the proportion of the price changesthat would be experienced in regions away from theport, we conducted a series of analyses to try tomeasure the extent of the change of prices inlandfor a percentage change at the port. We do theanalysis for the two major crops for which we havecomplete data series: rice and maize. In the first

analysis, we stack the price data from various mar-kets for the last three years (1998–2000) and regressthe price of the inland market (in logs in timeperiod t) on the price at the port market (in logs intime period t) and three lags (t-1; t-2; and t-3—table 6.8). The sum of the coefficients on the portvariables provide an intuitive measure for the totaltransmission of price shocks in percentage terms.The price transmission coefficients range from 42to 51 percent for rice (the lower range being ameasure of the price impact that does not includecoefficients with t-ratios under 1.580 (table 6.8).According to these measures, about half of the pricechange at the port is transmitted to the inland mar-ket. The transmission coefficients for maize rangefrom 51 to 57 percent (table 6.8). This means that if

TABLE 6.8 Transmission Coefficients forRice and Maize Measuring thePercentage of Price Shock atPort that Is Transmitted toInland Markets in China,1996–2000a

Maize Rice

Standard Vector 51–57 42–51Autogression Model(VAR)b

VAR with corrections 49 10–13for autocorrelation.

Vector Autogression 20–35 12–25Model with Impulse Response Simulationd

aThe transmission coefficient is interpreted atthe average proportion of a price shock in theport market that is experienced by the marketsinland.bStandard Vector Autoregression Analysis stacksthe price data from various markets for the lastthree years (1998 to 2000) and regresses theprice of the inland market (in logs in timeperiod t) on the price at the port market (inlogs in time period t) and three lags (t-1; t-2;and t-3). The sum of the coefficients on theport variables is a measure of the total transmis-sion of price shocks in terms of proportions.cThis uses the same model as in row 1, but alsoincludes lags of the dependent variable for t-1,t-2 and t-3. Simulation analysis proceeds byshocking the price at the border and followingthe price from port to inland, holding the ownmarket’s price generation constant.

TABLE 6.7 Soybean, Corn, and WheatVillage Price Regression, 2000

Explanatory Soybean Corn WheatVariable Price Price Price

Distance to the �0.029 �0.00064 �0.0095nearest countrymarket

(2.37)** (�1.63)* (3.24)**

Village-level �0.04 0.12 0.081shock to production

(�0.17) (�1.34) (�1.02)

Other variables timing of sales/net purchase not shown or seller


maize prices fell 10 percent at the border due tothe import of an amount up to China’s quotalimits, the price of maize inland would fall around5 percent.

The analysis in the above paragraph, however,does not account for the fact that the error term inthe equation could be subject to autocorrelation.The corrected transmission coefficients fall onlymodestly for maize, declining to 49 percent(table 6.8). In other words, even after accountingfor autocorrelation, if the price at the border (orport) changes by 10 percent, the inland price fallsby nearly 5 percent. After the same treatment in therice equations, however, the price transmissioncoefficient falls sharply (to around only 10 percent(table 6.8)). According to this result, we find thatrice markets are subject to much more frictionthan those for maize, a result that may result fromthe inherent differences between rice (a commod-ity with a wide range of qualities) and maize (amore homogeneous commodity). In other words,when we are observing a price shock in the portmarket, it may be caused by the new inflow orshortage of a particular type of rice. For example,when a certain type of rice in Guangzhou suffers ashock (e.g., the harvest in one of the productionbases is greatly reduced, although the price inanother production base that produces that varietymay move, the price in an area that does not pro-duce that variety may not change. In the case ofmaize, however, its homogeneous nature meansthat most of the price shock in one market ispassed through by traders.

A third analysis confirms the finding of the moretraditional price transmission model (table 6.8).Using a impulse response analysis (which basicallyuses the price transmission model and adds a set oflagged own prices the price in the inland market inperiod t-1, t-2 and t-3), we find that a 10 percentprice change in the port price of maize changes theinland price by up to 35 percent and up to 28 per-cent in the case of rice.

Conclusions

The purpose of our study is to examine the effectsof China’s accession to the WTO on the agriculturalsector. Although there are many possible effects, wehave focused on the impact of trade policy changes

on agricultural prices. Moreover while trade willalso affect other subsectors of the rural economy,this study’s focus on the agricultural sector showedthat there will be an impact. However, like othereffects, those in the agricultural sector may not allbe negative. Our findings, based on new methods tocollect data and create NRPs, show that indeed forsome crops WTO accession will likely lead to a fallin prices and a rise in imports. Maize and cottonmay be most affected. It is possible that soybeansand sugar could be significantly affected in thelonger run. There are also commodities in whichChina has considerable comparative advantage—for example, rice, meats, and horticultural productsand, hence, WTO accession could provide benefitsto those engaged in these activities. The prospect ofincreased imports of feed grains (e.g., maize andsoybeans) at lower prices means that livestock pro-ducers could become even more competitive.

The extent to which prices fall from risingimports or rise from rising exports in part dependson how China executes its WTO obligations.Although there may be room for footdragging(which could delay that negative effects), the natureof the agreement also provides many means to limitthe downside effects. Likewise, China’s benefits aregoing to depend on how well trading partnershonor their commitment and provide better accessto global markets. We suggest that rather than foot-dragging on its own reforms, China would be betterto use some of its capital and goodwill to fightmeasures in its trading partners, such as Japan’ssafeguards against mushrooms. Here, China hasalready had a huge, unheralded win by gettingJapan to move from its original way of conductingbilateral trade with China (based on blatantlyWTO–illegal measures to safeguards that are trans-parent and temporary. China also should realizethat, as a WTO member, it now has a means to file aprotest if it believes it is being treated unfairly.When countries use the WTO enforcement organi-zations to fight unauthorized trade barriers, com-plainants tend to have a high success rate. In gen-eral, to gain the most in the long run from thisagricultural agreement, both China and its partnersneed to endeavor to live up to their agreements.

We also found that unlike the case of Mexico, itappears as if most of China’s markets may be well-integrated into the economy. This is good news and


bad news for poor farmers. The good news is thatthey can benefit from falling input prices and risingexport opportunities. The bad news is that unlike alarge number of maize farmers in Mexico who werenot affected by the North American Free TradeAgreement reduction in maize import restriction, ifour results are correct for large parts of China, itsfarmers will be affected. The problem, although it isa short-run one, may be that it is this group of ruralhouseholds that are most dependent on agricultureand least able to be flexible. As a consequence, ourfindings should be taken as a warning to govern-ment leaders that they need to be concerned aboutthe welfare of these susceptible groups.

Our results also generate findings that show theclose relationship that exists among the degree ofintegration of an economy, the size of the priceeffect, and the amount that fraction of the TRQthat will be imported. If China’s markets are reallyso integrated, and leaders do not artificially delaythe ability of traders to execute TRQs, our findingssuggest that the price effects may not be too largebecause they will be spread out across a large areaof the country. However, if the price effects for agiven quantity of imports are not large, the vol-ume of imports may be larger than predicted bysome and the bindings may be more likely to takeeffect. We do not expect in any circumstance thatimports will ever exceed the limits put on by theTRQs.

Appendix: Challenges and Issues in Measuring Nominal ProtectionRates

Although measuring the difference in pricebetween an economy’s domestic price and theinternational price, the wide range of estimates ofNPRs that exist for China demonstrate it is not astraightforward process. In fact, a number of issuescomplicate NPR measurement. First, confusion maystem from the way analysts have asked their ques-tion about NPRs. Policy makers and researchershave sought to summarize the impact of variouscommodities with a single number. Trade modelersneed a single number to make their analyticalframeworks tractable. People want to know what isthe price of wheat in China and compare that to theworld price of wheat. With this information, the

NPR of a commodity is simply the difference ofthese two numbers.

However, more careful observation shows thesearch for a single number may be one of the mainreasons different analysts can come to so many dif-ferent conclusions. In fact, there are many prices forwheat in China. Prices vary across time within ayear. Prices vary across regions within a timeperiod. When calculating the NPR, does one look atthe price of corn in a Guangzhou feedlot or theprice of corn sitting in storage in a farmer’s home-made silo in Northeast China? Moreover, rice is notrice is not rice. There are many different varietiesand types, all of which command different prices atdifferent places at different times during the year. Infact, for some commodities, such as rice, China isexporting one type at the same time it is importinganother. The same sets of issues face analysts whenthey attempt to choose a price series (or more diffi-cult yet, the single price) to represent the interna-tional price. Which price should an analyst choose?Should it be FOB or CIF? Should it be the averageannual price or a price during one particularperiod? If there are many different types ofimported varieties, which type should be chosen?

In part because previous studies have not dealtwith these issues (at least explicitly), it is unsurpris-ing that different research efforts have generateddifferent estimates of NPRs. For example, Tuan andCheng (1999) estimated quite high and variablenominal rates of protection for agricultural com-modities. Their estimates for wheat, maize, andsoybeans in 1997 were 62, 15, and 140 percentrespectively. On the other hand, Carter and Estrin(2001) find generally negative price distortions.Huang (2001) provides sets of estimates that showsome products are highly protected and in othercases there are negative rates of protection.

Notes

1. Some scholars in China have also pointed out that sincepart of the value of agricultural commodity production usesinputs on which the value added tax has been assessed, the “real”tax rate on agricultural commodities is actually higher. Althoughcertainly this is the case, the maximum that could be addedwould only be an additional 2 to 4 percentage points (15 percenttimes the share of the inputs that were taxed—about 10 to30 percent—depending on the commodity, the technology, andregion of production).


2. Because of the absence of a single central authority thatmanages grain flows, the enumeration team chose its sample in anumber of ways. In each location, we first visited the local grainbureau and obtained access to a list of all grain bureaus, the firmsthat they were running on a commercial basis, and their sub-sidiaries. We interviewed an official in the grain marketing divi-sion and transportation division. We also chose three firms thatwere owned directly by the grain bureau and three that wereaffiliated with the grain bureau. In several cities, the grain bureauhad a list of large grain trading and grain using firms (e.g., millsand feed lots). In others, this was obtained from the marketadministration bureau. Five firms were chosen on the basis thatthey were private and had yearly sales that exceeded one millionyuan. We interviewed at least two flour or rice mills and feedmills in each location. Finally, we visited the wholesale marketand randomly chose five stalls to interview. The team also visiteda number of other entities, such as the grain reserve, the localCOFCO agency, and supermarket chains. In some cases, themanagers of these entities knew the grain trade business wellenough to answer our questions; in other cases they did not.

3. During the interviews, a survey instrument was filled outdocumenting the scope of the interviewee’s participation inChina’s domestic and international food and fiber trade. Wewere particularly concerned with understanding the transac-tions that the interviewees were involved with or knew aboutthat concerned imported or exported grains, fiber, meat andother goods. The survey recorded the characteristics of the com-modities that were involved in trade in the immediate marketingarea during the fall of 2001. Enumerators then asked the inter-viewee a series of questions about commodities about which thetraders were most familiar. For imported commodities, intervie-wees first told the enumerators the international CIF price of thegood. Second, the interviewee then told enumerators what thegood would sell for if auctioned of in a competitive auction. Inother words, we elicited a series of price gaps for a carefullydefined set of goods. Since, on average, each interviewee hadinformation about a number of commodities, we had severalhundred observations. A similar set of questions was askedabout exportable goods, including maize, rice, cotton, and meatproducts.

4. The assertion that consumption will be limited is based onthe assumption that tastes will not change. It could be, however,that when the changes in consumption due to rising incomes areadded to those due to changing tastes as higher quality productsbecome cheaper, the scope for much greater penetration of thedomestic market clearly seems possible. After years or centuriesof consuming certain foods, the Chinese might find as othershave that better quality bread or noodles are desirable options,in some cases offering nutritional advantages as well.

5. We should stress, however, our survey was conducted thesame way. In most cases, interviewees told us that there were nota lot of quality differences among maize varieties. Moreover,there was only a slight (2 to 3 percent) of price differencebetween imported and domestic soybeans from quality.

6. These are computed by comparing the domestic wholesaleprice with the average implicit price of trade. For the importable(exportable) it is total value of import (export) divided by totalvolume of import (export).

7. Since we use data over time, we need to convert prices to areal basis. Nominal prices from our data set are deflated usingmonthly consumer price indices calculated and reported bythe China National Statistical Bureau. Deflation facilitatestransaction cost comparisons across time and allows us todisregard transaction cost increases within periods associatedwith inflation.

8. We do, however, in an alternative specification (not shownhere) include a variable that controls for the average amount ofprivate grain storage in the village. It is not significant and itsinclusion does not affect the results.

9. It is also possible that there could be transmission elastici-ties greater than unity. Producers in an exporting region mayfind their price moves proportionately more than in the port.For example, consider the case if the CIF price was 100 and theinterior price was 50. If the marketing margin were additive, therise in the world price of 10 percent.

References

Anderson, K., and C. Y. Peng. 1998. “Feeding and Fueling Chinain the 21st Century.” World Development 26(8): 1413–29.

CARD. 2001.“Effect of WTO on Agriculture.” Paper Presented atthe 2001 annual meetings of the American Agricultural Eco-nomics Association, August 5–8. Chicago, IL.

Carter, C. A., and A. Estrin. 2001. “China’s Trade Integration andImpacts on Factor Markets.” Mimeo. Department of Agri-cultural and Resource Economics, University of California,Davis.

Chen, Shaohua, and Martin Ravallion. 2004. “Welfare Impacts ofChina’s Accession to the WTO.” In D. Bhattasali, S. Li, and W.Martin, eds., China and the WTO: Accession, Policy Reform,and Poverty Reduction Strategies. Washington, D.C.: WorldBank and Oxford University Press.

Huang, Jikun. 2001. “Agricultural Policy and Food Security.”Working paper, Center for Chinese Agricultural Policy,Chinese Academy of Sciences. Beijing.

Huang, Jikun, and C. Chen. 1999. “Effects of Trade Liberaliza-tion on Agriculture in China: Institutional and StructuralAspects.” Working Paper 42, United Nations ESCAP CGPRTCentre, Bogor, Indonesia.

Huang, Jikun, and Scott Rozelle. 2002. “Market Distortions andthe Impact of WTO on China’s Rural Economy.” WorkingPaper, University of California, Davis, CA.

Li, S., F. Zhai, and Z. Wang. 1999. “The Global and DomesticImpact of China Joining the World Trade Organization, AProject Report.” Development Research Center, the StateCouncil, China.

Martin, W. 2002. “Implication of Reform and WTO Accessionfor China’s Agricultural Policies.” Economies in Transition(forthcoming).

Ministry of Agriculture. 2001. China Statistical Yearbook.Beijing: China Agricultural Press.

Nyberg, A., and S. Rozelle. 1999. Accelerating China’s RuralTransformation. World Bank, Washington, D.C.

OECD. 2001. “Implications of Trade and Investment Liberaliza-tion for China’s Agriculture.” Working Paper, Organizationfor Economic Cooperation and Development, Paris.

Park, A., H. Jin, Scott Rozelle, and Jikun Huang. 2002. “MarketEmergence and Transition: Transition Costs, Arbitrage, andAutarky in China’s Grain Market,” American Journal of Agri-cultural Economics. Forthcoming.

Rozelle, Scott. 2003. “After Accession to the WTO: Corn TradeWithin China and Between China and the Rest of theWorld.” A Report written for the U.S. Grains Council,Beijing, China.

Rozelle, Scott, Albert Park, Jikun Huang, and Hehui Jin. 2000.“Bureaucrat to Entrepreneur: The Changing Role of theState in China’s Transitional Commodity Economy,” Eco-nomic Development and Cultural Change 48, 2: 227–52.

Taylor, J. Edward. 1998. “Trade Liberalization and the Impact onSmall Holders in Rural Mexico,” Working Paper, Depart-ment of Agricultural and Resource Economics, University ofCalifornia, Davis.

Tuan F.C., and Cheng, G. 1999. “A Review of China’s Agricul-tural Trade Policy,” paper prepared for the IATRC SummerMeeting. June 25–26. San Francisco.

Xie, Yuping. 2002. “WTO and China’s Agricultural Trade andImpact on Domestic Markets,” Unpublished Masters Thesis.

Center for Chinese Agricultural Policy, Chinese Academy ofSciences, Beijing.

World Bank. 2000. “An Assessment of China’s Poverty Policy.”Working Paper, World Bank, Washington, D.C.

Zhao, Yaohui, and Terry Sicular. 2002. “Affects of WTO Acces-sion on China: Labor Market Responses.” Working Paper,Center for Chinese Economic Reform, Peking University,Beijing.


Part II

Economicimpacts ofAccession

in rural poverty (reported in Kanbur and Zhang2001). So high is the level of concern that at its firstWTO ministerial meeting in Cancún in September2003, China joined a new group of more than 20developing countries that demanded that membercountries of the Organisation for Economic Co-operation and Development (OECD) reform theiragriculture sectors without requiring developingcountries to lower their farm trade barriers.

The policy changes that China still must make tofulfill its WTO obligations include major changesin its farm trade policies by 2005—that is, cuts inprotection that appear to be far greater, and faster,than those imposed on any other developing coun-try in the Uruguay Round Agreement on Agricul-ture. With one-quarter of rural households inChina living on less than $1 a day in 1999 (versus1 percent of urban households) and more thanthree-quarters of all poor Chinese people living inrural areas, concerns about the impact of WTOaccession on rural poverty are understandable. Yetreduced protectionism also may boost output andexports of some labor-intensive farm products inwhich China still has a comparative advantage.1 In

After 15 years of negotiations, China acceded to theWorld Trade Organization (WTO) at the end of2001. During those negotiations, China was contin-ually opening up and reforming its economy, andfurther reforms will be introduced over the nextfew years to fulfill its legal obligations under itsWTO Protocol of Accession. Such reform necessar-ily involves structural adjustments by households,firms, and bureaucracies, and although the econ-omy as a whole can gain substantially from thoseadjustments, losses and even hardship can result forsome households unless complementary domesticpolicies are in place to facilitate adjustment orcompensate losers. This situation underscores theimportance of first analyzing the likely distribu-tional consequences of the reforms themselves andthen considering what complementary policies areneeded to provide adequate safety nets for potentiallosers.

Of particular concern in China’s case is that thereforms may cause farm incomes to fall, exacerbat-ing the rise since the mid-1980s in farm–nonfarmand inland–coastal aggregate income inequality andpossibly reversing the decline since the late 1970s

101

7

The Impacts of WTOAccession on Chinese

Agriculture andRur al Poverty

Kym Anderson, Jikun Huang, and Elena Ianchovichina

Our thanks are extended to the World Bank and Australian Research Council for their financial support andto Will Martin, Tom Hertel, Scott Rozelle, and three referees for helpful comments on earlier versions of thischapter. The usual disclaimer applies.

addition, farm households will be affected indi-rectly by many of the other commitments Chinahas made in its WTO Accession Protocol. Especiallyimportant will be the arrangements for phasing outthe “voluntary” export restraints on China’s textileand clothing trade and the reductions in protectionof the motor vehicle and parts industry. Thosechanges, together with the promised increase ina wide range of agricultural imports, will allowChina to exploit more fully its strong comparativeadvantage in unskilled labor-intensive products—both farm and nonfarm.

An assessment of the impacts on farmer incomesand on rural areas of the remaining reformsrequired to meet China’s commitments to WTOmembership is best conducted by seeing thoseimpacts in the context of ongoing economic growthand structural change. This chapter thereforebegins with a brief summary of rural developmentssince the initial reforms began in the late 1970s, ofrecent policies affecting rural households, andof the pertinent reforms still to be delivered as partof China’s WTO commitments. With that back-ground, the chapter then provides some indicationof the likely effects of the reforms to be imple-mented between 2002 and 2007 on the welfare ofdifferent factor markets and thus different types offarm households. Because even the direction, letalone the magnitude, of some of the effects cannotbe discerned from theory (McCulloch, Winters, andCirera 2001), we use the numerical simulationmodel known as GTAP (named after Purdue Uni-versity’s Global Trade Analysis Project) to addressthese issues. Our experimental design, which isdetailed in the companion paper by Ianchovichinaand Martin (2004), goes beyond previous studies(see the survey by Gilbert and Wahl 2002) in severalrespects: (a) it focuses just on the liberalizationrequired to meet China’s WTO accession commit-ments from 2002 (by projecting the GTAP databaseto 2001 with China’s prior reforms and all coun-tries’ Uruguay Round commitments implemented);(b) it incorporates new estimates of nominal ratesof agricultural protection as of 2001 and at the endof accession; (c) it takes account of existing dutydrawbacks so that it avoids overstating the extent ofreform; (d) it binds the government budget byadding a uniform consumption tax to offset the lossof tariff revenue; (d) it takes account of differencesbetween farm and nonfarm unskilled labor; and(e) it examines the real income effects on different

types of farm households. The chapter concludes bydrawing out implications for Chinese policymakerswishing to preempt any increases in food insecurityor rural poverty.

The Setting

To put prospective changes in perspective, it ishelpful to review rural developments over the past25 years and then the policies that have affectedthose changes.

Rural Developments since the Late 1970s

China’s decision in December 1978 to open up itseconomy acted as a major stimulus to economicgrowth: the pre-reform rate of growth in the percapita gross domestic product (GDP) of 3.1 percenta year more than doubled, and it has remainedabove 7 percent for the past two decades (see lastrow of table 7.1). Rapid economic growth is nor-mally accompanied by a relative decline in the farmsector, but in China that decline was initially tem-pered by introducing the farm household responsi-bility system (which led to replacing collectivefarms with individually managed holdings) and byraising the prices received by farmers. So began theprocess of moving away from taxing agriculturerelative to other sectors—a process followed bymost of the advanced economies in the early stagesof their industrialization (Anderson and others1986; Lindert 1991).

Agriculture grew nearly as rapidly as industryfrom 1979 to 1984. However, the one-off efficiencyeffects of moving to the household responsibil-ity system and raising relative prices for farm prod-ucts were mostly reaped by the mid-1980s, afterwhich agriculture grew at only one-third the paceof industry and less than half that of the servicessector as industrialization boomed on the easternseaboard (table 7.1). Meanwhile, the employment,output, and exports of rural township and villageenterprises also boomed (see the first threecolumns of table 7.2). Despite that migration offarm workers to rural industrial and service activi-ties (not to mention the temporary unauthorizedmigration to urban jobs such as those in construc-tion), the average area per farm worker has fallensteadily since the late 1970s, while the share of farmhousehold income from nonfarm sources has risen(see last two columns of table 7.2). Table 7.3 shows


the slowdown in the decline in agriculture’s sharesof GDP and employment in the 1980s and its sub-sequent acceleration in the 1990s.

Income growth has boosted the demand forfoods that are high in protein and nutrients relativeto those high in carbohydrates, which in turn hasstimulated major structural changes within agri-culture as farmers respond to the changes indomestic demand. For example, livestock and fishincreased their share of agricultural output from

less than 20 percent in the late 1970s to 30 percentby the late 1990s (table 7.3). Within the crop sub-sector, fruit and vegetable production grew two tothree times as fast as grain output (table 7.1). Theprices and marketing of grain and oilseed productshave continued to be highly regulated, whereas themarkets for horticultural, livestock, and fish prod-ucts have been greatly liberalized. This liberaliza-tion has accentuated the growth in the output ofthe latter group relative to grain and oilseed output

The Impacts of WTO Accession on Chinese Agriculture and Rural Poverty 103

TABLE 7.1 Growth Rates of China’s Economy, 1970–2000 (percent per year)

Pre-Reform Reform Period

1970–78 1979–84 1985–95 1996–2000

Gross domestic product a 4.9 8.8 9.7 7.9Agriculture 2.7 7.1 4.0 3.4Industry 6.8 8.2 12.8 9.6Services n.a. 11.6 9.7 8.3

Food production volumeGrain 2.8 4.7 1.7 0.03Oilseed crops 2.1 14.9 4.4 5.6Fruit and vegetables 6.6 7.2 12.7 8.6Red meat 4.4 9.1 8.8 6.5Fish 5.0 7.9 13.7 10.2

Value of output of nonfarm n.a. 12.3 24.1 14.0rural enterprises

Population 1.80 1.15 1.43 0.9

Per capita GDP 3.1 7.7 8.3 6.9

n.a. Not applicable.aFigures for GDP in 1970–78 are the growth rate of national income in real terms. Growth rates arecomputed using the regression method. Growth rates of individual and groups of commodities arebased on volume of production data; sectoral growth rates refer to value added in real terms.Sources: National Bureau of Statistics of China, various issues; Ministry of Agriculture, Agricultural Yearbookof China, various issues.

TABLE 7.2 Farm and Rural Enterprise (RE) Developments in China, 1980–99

RE’s Share Share of Farmof Total RE’s Share RE’s Share of Household Income

Rural Labor of Total GDP Total Exports Farmland Size from Nonfarm(%) (%) (%) (Hectares/Farm) Sources (%)

1980 9 4 0 0.56 171985 19 9 15 0.51 251990 23 14 43 0.43 261995 34 25 48 0.41 371999 35 30 48 0.40 47

Sources: National Bureau of Statistics of China, a, b; various issues.

since the mid-1980s (table 7.1). Meanwhile, thedirect consumption of grain by both rural andurban households has virtually ceased growing(table 7.4)—a consequence not only of incomesrising but also of population growth slowing to lessthan 1 percent a year and of cuts in the implicitconsumption subsidy for food grains.

The use of grain for animal feed continuesto grow. To date, that increase has been supplied

almost completely by rising domestic production, sothat the trend level of grain self-sufficiency hasremained close to 100 percent. Table 7.5 shows thatthere are nevertheless considerable changes fromyear to year in grain exports and imports. It alsoshows that, overall, China has remained a netexporter of food and feed, with meat, fish, and fruitsand vegetables providing most of the growth in netexport earnings.2


TABLE 7.3 Changing Structure of China’s Economy, 1970–2000 (percent, based oncurrent prices)

1970 1980 1985 1990 1995 2000

Share of GDPAgriculture 40 30 28 27 20 16Industry 46 49 43 42 49 51Services 13 21 29 31 31 33

Share of employmentAgriculture 81 69 62 60 52 49Industry 10 18 21 21 23 23Services 9 13 17 19 25 28

Share of agricultural outputCrops 82 76 69 65 58 56Livestock 14 18 22 26 30 30Fish 2 2 3 5 8 10Forestry 2 4 5 4 3 4

Share of population 83 81 76 72 71 64that is rural

Sources: National Bureau of Statistics of China, a, b; various issues.

TABLE 7.4 China’s Grain Production, Trade, and Consumption, 1980–2000(million metric tons)

1980–89 1990–94 1996–2000

Production 332 396 389Net imports 8 �1 �3Change in stocks 1 11 12Consumption 339 384 374

Food—urban 35 42 41Food—rural 177 190 177Feed 64 86 91Other (seed, industrial 63 66 65use, waste)

Source: Huang 2001, based on China National Statistics Bureau publications and theCenter for Chinese Agricultural Policy’s database.


TABLE 7.5 Structure of China’s Food and Feed Trade, 1980–99 (millions of US$)

SITC 1980 1985 1990 1995 1999

ExportsLive animals 00 384 304 430 473 374Meat 01 361 448 791 1,349 1,054Dairy products 02 71 57 55 61 71Fish 03 380 283 1,370 2,875 2,969Grains 04 423 1,065 614 281 1,273Fruits and 05 746 825 1,759 3,399 3,150

vegetablesSugar 06 221 79 317 321 214Coffee and tea 07 328 435 534 523 561Animal feeds 08 58 241 623 351 239Other foods 09 49 66 107 290 541Oilseeds 22 — — — 522 373Vegetable oils 4 — — — 454 132

Total food 3,021 3,803 6,600 10,899 10,951

ImportsLive animals 00 5 18 14 18 22Meat 01 1 6 54 97 503Dairy products 02 5 31 81 60 160Fish 03 13 44 102 609 890Grains 04 2,458 982 2,353 3,631 574Fruits and 05 48 52 83 185 384

vegetablesSugar 06 316 274 390 935 183Coffee and tea 07 56 40 30 74 72Animal feeds 08 14 83 182 423 620Other foods 09 2 23 46 92 182Oilseeds 22 — — — 110 1,531Vegetable oils 4 — — — 2,596 1,352

Total food 2,918 1,553 3,335 8,828 6,474

Net exportsLive animals 00 379 286 416 455 352Meat 01 360 442 737 1,252 551Dairy products 02 66 26 �26 1 �89Fish 03 367 239 1,268 2,266 2,079Grains 04 �2,035 83 �1,939 �3,350 663Fruits and 05 698 773 1,676 3,214 2,766

vegetablesSugar 06 �95 �195 �73 �614 31Coffee and tea 07 272 395 504 449 489Animal feeds 08 44 158 441 �72 �381Other foods 09 47 43 61 198 359Oilseeds 22 — — — 412 �1,158Vegetable oils 4 — — — �2,142 �1,220

Total food 103 2,250 3,265 2,071 4,477

— Not available.SITC = Standard International Trade Classification.Source: Mathews 2001, based on UN COMTRADE statistics.

Recent and Prospective Policies AffectingRural Areas

As in most developing countries,3 agriculture inChina was squeezed at the early stages of industrial-ization when gross fiscal contributions to the sectorwere more than outweighed by implicit taxation inthe form of depressed prices for farm products, neg-lect of public infrastructure in rural areas relative tothat in urban areas, and capital outflows via thefinancial system (Huang and Ma 1998). Then, agri-culture experienced the introduction of price andother market reforms associated with China’s policyshift from a socialist to a market-oriented economy,starting with nonstrategic commodities such as veg-etables, fruits, fish, livestock, and oil and sugar crops.The early reforms were aimed at raising farm-levelprices and gradually deregulating the market. As theright to private trading was extended to include sur-plus output of all categories of agricultural productsafter contractual obligations to the state were ful-filled, the foundations of the state marketing systembegan to be undermined (Rozelle and others 1997).Despite periodic stop-go cycles in the reformprocess, the proportion of retail commodities soldat market prices has continued to rise. Accordingto Lardy (2002), the share for agriculture was just6 percent in 1978, but had risen to 40 percent by1985, 79 percent by 1995, and 83 percent by 1999.

What have these policies meant for the nominalrates of agricultural protection in China (the per-centage by which domestic prices exceed prices at the

country’s border)? Table 7.6 shows new estimates byHuang, Rozelle, and Chang (2004) of nominal pro-tection rates for key agricultural commodities forsome recent years. It suggests that rice, meat, fish, andfruits and vegetables have been priced at less thanborder prices, while other grains, oilseeds, sugar,cotton, and milk have been priced at 20–40 per-cent above border prices. Maize and cotton alsoenjoyed export subsidies in 2001, amounting to one-third and one-tenth of f.o.b. prices, respectively.

Table 7.6 also shows China’s commitments in itsWTO Protocol of Accession: the tariff-rate quotawill apply to grains, sugar, and cotton whose out-of-quota tariffs are quite high, but, otherwise, after thephase-in period the tariffs range between just 1 and15 percent—representing substantial liberalizationsover 2001 levels. As well, producers of major cropsmay continue to be affected by commodity-specificpolicies of government procurement of a portion ofthe crop at lower-than-market prices as in the past(see Sicular 1988) or at higher-than-market prices,as in 1998 (see Huang 1998).

What will those reforms mean for agriculturaltrade? Many analysts expected China to becomeever more dependent on agricultural imports in thecourse of the economy’s rapid industrializationover the past two-plus decades. Some extremists(e.g., Brown 1995) even suggested that China couldseriously deprive other developing countries offood. Yet, as reported earlier in this chapter, growthin net food imports has not yet happened, at least


TABLE 7.6 Nominal Rates of Protection (NRP, Applied Tariff or Tariff Equivalent),Agricultural Products: China, 1995–2007 (percent)

2007 Assumed NRP2007 Out-of-Quota in 2007

1995 1997 2001 In-Quota Tariff Tariff (Core Scenario)

Rice �5 �5 �3 1 65 �3Wheat 25 17 12 1 65 12Coarse grains 20 28 32 1 65 32Fruits and vegetables �10 �8 �4 11 11 �4Oilseeds 30 28 20 3 3 3Sugar 44 42 40 15 50 20Cotton 20 17 17 1 40 20Meats �20 �19 �15 12 12 �15Milk 30 30 30 11 11 11

Source: Based on research subsequently reported in summary form in Huang, Rozelle, and Chang 2004.

not in a sustained way, and China has continued tobe a net exporter of meat, fish, and fruits and veg-etables. Indeed, on occasion in the latter 1990sChina also was a net exporter of grain and cotton.How much of that situation was attributable togovernment policies that constrained domesticdemand, including occasional export subsidies, is amoot point.

In its WTO Protocol of Accession (WTO 2001),China agreed to have no agricultural export subsi-dies and to limit its domestic support for farmers to8.5 percent of the value of production (comparedwith 10 percent for other developing countries).The import market access commitments China hasmade to WTO members look substantial on paper.Tariff-rate quotas (TRQs) will be retained only onwheat, rice, maize, edible oils, sugar, cotton, andwool—the domestic production of which in aggre-gate makes up only about one-sixth of China’s agri-cultural GDP. The quota volumes are to grow over2003–05 at annual rates ranging from 5 to 19 per-cent. China has also agreed to weaken the monopo-lies previously held by state trading enterprises,except for tobacco. Even though some state tradingenterprises will continue to operate, there will besome competition from private firms in theimporting and exporting of farm products, at leastwithin the tariff-rate quotas.

Farmers and the rural sector more broadlywill also be affected by China’s commitment to pro-vide improved and WTO-bound market access forindustrial products. Mineral and manufacturingtariffs will be bound and generally reduced on abroad basis, with many tariffs falling to 10 percent orless. Tariffs were to be cut on accession, and furthercuts will be phased in by 2005 (with just a few excep-tions). Furthermore, for industrial products Chinawill reduce significantly its nontariff measures andwill eliminate all quotas, tendering, and importlicensing on nonfarm merchandise by no later than2005. Quotas on Chinese imports of automobilesand parts will grow by 15 percent annually, froma level of about US$6 billion4 in 2000, and thesequotas will be eliminated by 2005. For textiles andclothing, however, the current “voluntary” exportrestraints will be phased out by the end of 2004,although importing countries will be able to reim-pose quotas under a special textile safeguard mecha-nism until the end of 2008. China also has madecommitments to open up its services markets.

Over the 1990s the average scheduled tariff ratesfor manufacturing initially exceeded those for agri-culture, but later fell more than those for agricul-ture. By 2010 the manufacturing average will bewell below that for agriculture (a simple average of9 percent versus 16 percent for agriculture—seeWTO 2003: Appendix Table IIB). However, thesenumbers do not give a true indication of the extentof change in protection that is taking place, becausein the 1990s many manufactures were enteringChina at reduced or zero tariffs via duty drawbacksto encourage foreign investment in the process-ing of imported intermediate goods for subse-quent export. Some agricultural products also haveentered at less than the scheduled rate, includingthrough smuggling.

What all this information means for incentivesfor each industry is difficult to discern precisely, butit is better information than has been available todate for analyzing empirically the economic effectsof the reforms associated with China’s WTO acces-sion, including the impacts on factor rewards andprices from which inferences about income distrib-utional effects can be made. We make these infer-ences bearing in mind the marked differences inper capita incomes between eastern, central, andwestern provinces, and between urban and ruralareas (as shown in tables 7.7 and 7.8).

Applying the GTAP Model

Version 5 of the computable general equilibriummodel of the global economy known as GTAP isused here.5 As an economy-wide model, GTAPdescribes both the vertical and horizontal linkagesbetween all product markets within the model’sindividual countries and regions as well as betweencountries and regions via their bilateral trade flows.For present purposes, the 1997 database is aggre-gated to 25 sectors and 20 regions and projectedforward first to 2001 and then to 2007, usingWorld Bank projections of population, income,and endowments of productive factors (agricul-tural land, other natural resources, unskilled labor,skilled labor, and other capital). The initial basecase assumes that China retains its protection poli-cies as of 1995 and that Taiwan (China) retains itsprotection as of 1997, but that all other countriesfully implement their Uruguay Round obligationson schedule before 2005. China’s trade policy


changes between 1995 and 2001 are assumed tohave been in anticipation of the requirements of,and thus part of, China’s WTO accession. Thesechanges are analyzed in detail in Ianchovichina andMartin (2004), together with the effects of imple-menting over the next few years the remainder ofChina’s commitments as recorded in its WTO Pro-tocol of Accession.6 In this chapter we focus just on

the additional reform commitments to be imple-mented after 2001, relative to the revised base casein which China’s reforms only up to 2001 are inplace and there are no further reforms to 2007. Forkey agricultural import policies, these remainingreform commitments are assumed to shift thenominal rates of protection (NRPs) from column3 to column 6 of table 7.6. As well, the export


TABLE 7.7 Rural Income, Employment, and Food Output by Region: China, 2000

Western Central EasternProvinces Provinces Provinces

Rural income per capita (yuan) 1,557 2,030 2,994Share of population that is rural (%) 81 75 71Share of rural labor force in agriculture (%) 75 73 61Share of farm household income from off-farm (%):

Wages 20 22 36Total 38 35 58

Share of national agricultural GDP (%) 17 34 49Share of regional food output (%) from:

Rice 8 11 6Wheat 5 5 4Maize 5 5 3Fruits and vegetables 11 10 14Oilseeds 4 6 3Sugar 0 0 1Cotton 3 2 1Dairy products 2 1 1Other livestock products 20 19 17Fish products 2 6 20Other food, agricultural, and forest products 40 35 30Total 100 100 100

Sources: Center for Chinese Agricultural Policy’s CAPSIM Model; China National Statistics Bureau, ChinaRural Statistical Yearbook 2001.

TABLE 7.8 Income and Its Distribution: Rural and Urban China, 1980–99

Real Per Capita Income IndexIncome Disparity

Urban/RuralYear Rural Average Poorest 20 Percent Gini Coefficient Income Ratio

1980 100 100 0.24 3.41985 175 157 0.23 2.21990 183 154 0.31 2.31995 239 179 0.34 2.81999 330 224 0.32 2.9

Sources: China National Statistics Bureau (CNSB), China Statistical Yearbook 1989–2000; CNSB’s ruralhousehold income and expenditure surveys.

subsidies in place in 2001 (34 percent for maize,10 percent for cotton) are eliminated, and weassume no new farm production subsidies areintroduced.7 The choices of new agricultural NRPsfall into three categories:

• no change if they were negative in 2001 (rice,meats, fruits and vegetables

• a move to part way between the in-quota andout-of-quota tariffs if the TRQs become bindingon imports (wheat, coarse grains, sugar, cotton)

• a move down to the new in-quota tariffs(oilseeds, milk).

The sensitivity of the results for the first two cate-gories is explored in the results section later in thischapter.

If this reform were to require a movement ofunskilled labor out of farm activities, three impedi-ments have to be kept in mind. The first is thatthose farm workers would be less than perfect sub-stitutes for those already in nonfarm pursuits.Econometric work by Sicular and Zhao (2004) sug-gests that restraints on mobility could be approxi-mated via a constant elasticity of transformation(CET) function with an elasticity of transformationof 1.32.8 Accordingly, we have incorporated thatfunction in the GTAP model for China. The secondimpediment to off-farm migration is that urbansocial welfare benefits such as subsidies for housing,food, education, and health care are not available tononurban people, except by purchasing a residencepermit or hukou (Zhao 1999). And the third imped-iment is that farm workers who permanently cuttheir ties with agriculture may lose entitlement toreturns from their family’s land, and even the directsupport and assistance of family members (Hussain2004). These latter two impediments have con-tributed to the persistence of a large gap in farmversus nonfarm returns to unskilled labor. Wemodel that gap as a “tax” wedge that raises the costof labor to urban employers, with urban unskilledworkers receiving the tax-inclusive wage.

The closure adopted is a long-run one in which,in addition to the earlier assumptions aboutunskilled labor, we assume agricultural land ismobile between industries within the agriculturesector, and skilled labor and capital are mobilewithin and between sectors. This closure contrastswith the short-run closure used by Chen and

Ravallion (2004). It keeps the aggregate trade bal-ance and government tax revenue as a fixed share ofGDP (with little change in net international capitalflows), and holds employment constant so thatwages adjust endogenously.

What Do the Results Show?

We begin with the bottom line of the main scenariobefore revealing the details.

The Core WTO Accession Scenario

The core empirical results suggest that WTO acces-sion will increase farm–nonfarm income inequal-ity. The main reason for this greater inequality isthat the relative producer prices of farm productsare projected by the GTAP model to fall more thanthe prices of labor-intensive nonfarm productsafter the completion of WTO accession reforms(table 7.9).

The removal of the “voluntary” export restrainton sales of textile products to the United States andEuropean Union is not the only boost to unskilledlabor off farms. There is, in addition, a lower coststructure in unskilled, labor-intensive manufactur-ing activities for three reasons:

• The lower demand for labor on farms lowers thecost of unskilled labor in manufacturing.

• Import taxes on the intermediate inputs used inthose manufacturing activities are lower becauseof the accession process.

• The real exchange rate effect of the tariff reduc-tions lowers the cost of nontraded goods andother factors used as inputs into manufacturingproduction.

As a result, the quantity of unskilled nonfarm labordemanded is greater (by 0.8 percent). However,lower farm product prices mean the quantity ofunskilled farm labor demanded is less (by 1.7 per-cent), resulting in a 0.7 percent fall in the real wagefor unskilled farm labor and a rise in real wagesfor unskilled nonfarm labor of 1.2 percent, afteradjusting for the change in the aggregate cost ofliving. Farmers are also made worse off by the lowerdemand for farmland, the return from which is5.5 percent lower in 2007 after WTO accessionreforms. Meanwhile, the real wages of skilled labor


increase by 0.8 percent, and the rewards to nonfarmcapital are 1.3 percent higher.

Together, these results suggest that the owners ofnonfarm capital gain almost the same in propor-tional terms as unskilled laborers in nonfarmemployment, but the latter do better than skilledworkers. Therefore, on balance, income inequalitymay improve slightly among nonfarm householdsdependent mainly on labor income.9

However, income distribution can be expectedto worsen slightly, just as between farm and non-farm households, although the degree depends onthe proportion of farm household income earnedoff the farm. With only 1.7 percent (or about 6 mil-lion) of unskilled farm workers leaving agriculturefor nonfarm work and with land returns depressedby 5.5 percent in addition to farm labor returnsbeing 0.7 percent lower, the gap between farm andnonfarm incomes even within rural areas—and

certainly between rural and urban areas—appearspoised to increase slightly unless remedial policyaction is forthcoming. For farm householdsentirely dependent on earnings from agriculture(type A in table 7.10), income would fall 1.6 percenton average. This decrease would differ little acrossthe three regions identified in table 7.7, becauseproduct shares for farm output—when fish prod-ucts are ignored—are reasonably similar in thewestern, central, and eastern provinces.10 For farmhouseholds earning 30 percent of their incomefrom nonfarm unskilled work, however, thatincome fall is only half as large (0.8 percent), andfor farm households earning 60 percent of theirincome from nonfarm unskilled work, theirincomes would not decline at all (types B and C,respectively, in table 7.10, rows 7 and 8).

The sectoral details of the GTAP results aresummarized in table 7.9. The real consumer prices,


TABLE 7.9 Sectoral Volume Effects of China’s WTO Accession Reforms(Core Case), 2002–07

Household Trade Balance Producer ConsumerOutput Employment Consumption Exports Imports (Millions of Prices Prices

(%) (%) (%) (%) (%) U.S. Dollars) (%) (%)

Rice �2.1 �2.3 �0.1 6.1 �7.1 64 �0.9 0.9

Wheat �2.0 �2.3 0.0 18.9 �10.1 174 �1.7 0.4

Feed grains �2.3 �2.6 �0.1 �77.8 �2.4 �596 �1.9 1.9

Fruits and vegetables �3.4 �3.7 0.1 14.6 �6.3 214 �1.9 �0.1

Oilseeds �7.9 �8.4 0.9 29.8 20.9 �789 �2.8 �4.7

Sugar �6.5 �7.4 0.6 13.9 24.1 �73 �1.9 �3.1

Plant-based fibers 15.8 16.4 �0.6 �51.8 7.7 �189 0.1 3.1

Livestock and meat 1.3 1.1 0.0 15.5 �8.9 837 �1.6 0.2

Dairy �2.0 �2.4 0.0 13.5 23.8 �143 �1.5 0.2

Other food �5.9 �6.4 0.4 11.4 62.6 �3460 �1.7 �1.8

Beverages and tobacco �33.0 �33.1 1.5 9.7 112.4 �14,222 �1.8 �6.9

Extractive industries �1.0 �1.3 �0.2 7.5 �4.4 2,088 �0.7 1.2

Textiles 15.6 15.5 0.7 32.7 38.5 �10,366 �1.7 �3.2

Apparel 57.3 56.1 0.5 105.8 30.9 49,690 �0.5 �1.9

Light manufacturing 3.7 3.7 0.0 5.9 6.8 1,786 �0.9 0.0

Petrochemical industry �2.3 �2.3 �0.2 3.1 11.8 �8,810 �0.7 0.8

Metals �2.1 �2.1 �0.3 3.7 6.8 �1,893 �0.4 1.3

Automobiles 1.4 �2.2 1.0 27.7 24.0 516 �3.9 �4.2

Electronics 0.6 0.4 0.5 6.7 6.8 453 �1.3 �1.7

Other manufactures �2.1 �2.2 �0.2 4.1 18.9 �11,291 �0.5 0.8

Trade and transport 0.0 0.0 �0.3 0.8 �0.4 493 �0.2 1.6

Construction 0.9 0.9 �0.4 2.7 17.5 �436 �0.2 1.7

Communication �0.5 �0.5 �0.4 �0.5 10.9 �56 0.1 1.9

Commercial services �2.0 �2.0 �0.5 �0.4 35.4 �1,749 0.2 1.9

Other services �1.7 �1.8 �0.3 1.4 33.6 �1,525 �0.1 1.6

Source: Authors’ GTAP results.

relative to the consumer price index (CPI) drivenlower the most by WTO accession are those formotor vehicles, oilseeds, and sugar (and for bever-ages and tobacco, although if China was usingimport taxes on those items as a form of consump-tion tax and their decline were to be matched by anincrease in domestic sales taxation, those pricedeclines may not materialize). The prices for textileproducts, and to a lesser extent clothing, are drivenlower as well. Among the farm products, consumerprices rise slightly for livestock products, somewhatmore for grains, and significantly for cotton (plant-based fibers).

Producer prices (also shown relative to the CPI intable 7.9) fall more than consumer prices becauseof a uniform consumption tax of 1.9 percent that islevied to compensate for the loss of import tariff rev-enue. The prices of farm products fall more thanthose of most other products except autos, and farmoutput is down for all but cotton and meat.Moreover,feed grain exports shrink by three-quarters and cot-ton exports by half with the abolishment of exportsubsidies. The difference in the effects on productionand consumption shown in table 7.9 reveal that

China’s food, feed, and fiber self-sufficiency will bereduced at least slightly by these reforms. But theextent is really quite minor: the trade balance columnin table 7.9 suggests that for all agricultural and foodproducts net imports would be greater because of theremaining accession reforms by only $3.96 billion ayear by 2007 (in 1997 U.S. dollars), which representsonly 1 percent of total imports.

The results described here depend as always onthe assumptions in the model. To check the sensi-tivity of some of those assumptions, two alternativescenarios—one with greater agricultural protectioncuts and the other removing negative agriculturalprotection—were run to compare their results withthose in the base accession scenario.

Alternative Scenario 1: Greater AgriculturalProtection Cuts

In this scenario the grain, sugar, and cottonNRPs drop to the in-quota tariff levels shown intable 7.6. In the additional GTAP simulation run,the differences for factor rewards are not huge inthe aggregate, but they would be in the direction of


TABLE 7.10 Change in China’s Real Factor Prices and National Economic WelfareAttributable to WTO Accession, 2001–07 (percent)

Alternative 2:Alternative 1: Core Case Plus Also

Core Accession Greater Agricultural Removing NegativeCase Protection Cuts Agricultural Protection

Farm unskilled wages �0.7 �0.9 �0.5Rental price of land �5.5 �6.4 �4.7Nonfarm unskilled wages 1.2 1.4 1.1Skilled labor wages 0.8 1.0 1.5Rental price of capital 1.3 1.5 1.5

Farm householda income–A �1.6 �1.9 �1.3Farm householda income–B �0.8 �0.9 �0.6Farm householda income–C 0.1 0.1 0.1

National economic welfare(increase in billions of1997 US$) 9.56 9.57 9.87

aAccording to the GTAP database, 57 percent of farm income from agriculture comes from unskilled labor,26 percent from land, and 17 percent from capital. In 1999 on average 51 percent of rural householdincome in China was earned outside agriculture, mostly from unskilled labor. Therefore, to illustrate theimportance of those off-farm earnings, three types of farm households are shown in this table: type A,where it is assumed that nonfarm unskilled labor contributes 0 percent of farm household income; type B,30 percent; and type C, 60 percent. It is assumed that all households have the same expenditure pattern,and so all changes are net of the consumer price index change.Source: Calculated from the authors’ GTAP results.

worsening income inequality: unskilled farm wageswould fall 0.9 instead of 0.7 percent, and, on theone hand, rewards to farmland would fall 6.4instead of 5.5 percent, while, on the other, nonfarmwages would rise 1.4 instead of 1.2 percent forunskilled workers and 1.0 instead of 0.8 percent forskilled workers (table 7.10). These changes wouldattract only another million workers from farms.Even though agricultural incomes would be lower,farm household income would not fall if at least60 percent of income comes from wages of non-farm unskilled labor (see row 8 of table 7.10).Domestic production of grains, sugar, and cottonwould, however, be less, and domestic consump-tion would be greater, so that self-sufficiency inthose products would be slightly lower. Even so, thenet imports of all food and agricultural productswould be greater by only $1.5 billion a year by 2007($5.43 billion instead of $3.96 billion). Such animport increase would be within the tariff-ratequotas for those items, with the possible exceptionof maize (depending on the extent to which otherfeed grains that are not TRQ-restricted, such asbarley, are substitutable for maize). National eco-nomic welfare would be only very slightly greater inthis case as compared with the core scenario (seebottom row of table 7.10).

Alternative Scenario 2: Removing NegativeAgricultural Protection

If the negative NRPs for rice, meats, and fruits andvegetables were to be raised to zero, the income dis-tributional effects would go in the direction oppositeto that in the first alternative scenario—that is,toward less inequality between farm and nonfarmhouseholds. The changes are not great though, eventhough these products account for nearly 40 per-cent of the value of food and agricultural outputin China. A comparison of columns 1, 2, and 3 oftable 7.10 reveal they would involve about as muchimprovement in income distribution as the previousalternative scenario would worsen it. This caseinvolves a 3 percent larger national economic welfaregain than the core case (bottom row of table 7.10).

Conclusions and Policy Implications

Our initial analysis suggests that rural nonfarmincomes will rise on average absolutely and possiblyeven faster than urban incomes for households

depending just on labor income (assuming urbanlaborers are more skilled). However, some farmhouseholds facing increased import competitionmay be worse off in this case, other things beingequal, if they are

• unable to send household members to jobs inexpanding industrial and service industries

• too poorly served with infrastructure to attractsuch activities to their own region

• unable to diversify into producing farm goodswhose relative prices have risen

• lacking relatives who are able to repatriate non-farm earnings to them.

Thus in the core scenario the incidence of ruralnonfarm poverty will fall mainly because of thegrowth in wages for unskilled workers in rural non-farm activities, while poverty may well increase inagriculture-based hinterland provinces a long wayfrom markets and in regions poorly served bythe infrastructure needed to attract investment inexpanding activities such as textiles and clothing.

The first alternative scenario shows that this sit-uation would be exacerbated slightly if the TRQ-protected items (grains, sugar, and cotton) wereto become even less protected than we initiallyassumed. By way of contrast, the second alternativescenario suggests that the situation could be ame-liorated slightly by removing the negative protec-tion affecting rice, meats, and fruits and vegetables.But both of these alternatives involve only smallchanges to the magnitudes of effects rather thanaltering the sign of those effects, and both add onlya small amount to the aggregate gains from tradeliberalization.

National self-sufficiency in farm products woulddecrease slightly, particularly for feed grains andcotton as demand for livestock products grows,driven by the income gains from trade reform, andas the production of natural fiber-based textilesand clothing expands. But, overall, most of thedeclines in domestic agricultural production asa consequence of the remaining postaccessionreforms are relatively tiny in magnitude, especiallywhen compared with the growth in farm outputthat would occur as a result of normal economicgrowth (compare columns 1 and 5 of table 7.11).

What should be done if some farmers’ incomesworsen relative to those of nonfarm households,


and if there is concern about the fall in agriculturalself-sufficiency? Rather than argue against tradereform, policymakers should seek the best ways todeal with those concerns (and with any transitoryunemployment that might follow reform). Themost efficient policy responses are likely to involveinvestments in rural human capital, rural infra-structure, and agricultural research and develop-ment (Fan, Zhang, and Zhang 2002); improve-ments in the land tenure system and rural financialmarkets; reductions in the informal taxes/levies onfarmers by local governments; and changes in grainmarketing.

First, the government might consider furtherinvestments in basic rural education and healthservices to reduce the adverse effects of tradereform on the incidence of poverty and perceivedfood security. Better education and health for farm-ers’ children not only boosts their farm productiv-ity should they choose to stay on the farm after fin-ishing school, but also increases their capacity tofind more lucrative off-farm work and to adjust tononagricultural employment and living (Schultz1975; Zhang, Huang, and Rozelle 2002). In additionto those longer-term benefits, there could also bean immediate poverty-alleviating effect if the gov-ernment were to cut basic school fees and make upthe shortfall with a bigger direct grant to rural pri-mary and middle schools.

Second, improvements in rural infrastructuresuch as roads and rail mean that a larger share ofthe price eventually received at the end of the mar-keting chain for farm products can be passed backto farmers. Such improvements also lower the bar-rier for off-farm work by members of farm house-holds, making it easier for them to take advantageof expanding employment opportunities in ruraltownships.

Third, agricultural research and development canease both urban and rural poverty (see Fan, Fang,and Zhang 2001; Hazell and Haddad 2001). A boostin agricultural productivity could significantly offsetthe 2–8 percent drop in agricultural production thatis estimated in the core scenario to result from WTOaccession. An important policy issue here is whetherChina should deny itself the use of genetically mod-ified organisms in food production.11

Fourth, improvements in the land tenure systemwould not only increase the incentive to investmore in land but would also enhance the collateralof farm households. If accompanied by improve-ments in rural financial markets, investments byfarmers back into agriculture would rise. Theywould rise even further if returns were increasedvia reductions in the informal taxes/levies imposedby local governments on farmers.12

Finally, the government might reduce its regula-tion of grain marketing and, in particular, cease


TABLE 7.11 Changes in China’s Farm Output, Employment, and Trade Volumes:Without WTO Accession, 1995–2007, and Resulting from China’s WTOAccession Reforms, 2002–07 (percent)

Changes Without WTO Changes Due to WTOAccession, 1995–2007 Accession, 2002–07

Output Employment Exports Imports Output Employment Exports Imports

Rice 63.8 �11.5 134.7 �8.8 �2.1 �2.3 6.1 �7.1Wheat 81.4 6.4 �15.2 126.3 �2.0 �2.3 18.9 �10.1Feed grains 109.5 23.8 �0.6 95.9 �2.3 �2.6 �77.8 �2.4Fruits and 98.2 16.8 �10.8 122.1 �3.4 �3.7 14.6 �6.3

vegetablesOilseeds 100.9 18.4 �36.0 151.7 �7.9 �8.4 29.8 20.9Sugar 112.5 14.5 109.4 88.7 �6.5 �7.4 13.9 24.1Plant-based fibers 137.2 41.1 �8.5 146.1 15.8 16.4 �51.8 7.7Livestock and meat 121.9 25.6 12.8 135.3 1.3 1.1 15.5 �8.9Dairy products 122.5 18.8 60.5 100.3 �2.0 �2.4 13.5 23.8Other food 110.8 �1.5 76.8 58.5 �5.9 �6.4 11.4 62.6

Source: Authors’ GTAP results.

compulsory procurement from farmers at less thanmarket prices and reduce the provision of grain tourban consumers at less than market prices. De-emphasizing the governors’ grain responsibilitysystem (provincial self-sufficiency) would allowmore exploitation of comparative advantage withinChina as well.

If all these measures proved insufficient to sup-port the incomes of the poorest farm households,short-term adjustment assistance via inframarginal(and therefore not output-inducing) producerprice subsidies could be provided as an efficientway to boost those households’ incomes withoutboosting farm output in an equal but opposite wayto that used to tax farmers in earlier decades (seeShea 2003). Such an intervention could well bedeemed WTO-consistent because of its decouplednature. In any case, if it just targets poor farmers itis unlikely to ever exceed 8.5 percent of the value ofChina’s output of the product concerned (China’sde minimis exemption limit for product-specificsupport under Article 6.4 of the WTO’s Agreementon Agriculture).

Finally, now that China is in the WTO it can takepart in the new rounds of multilateral trade negoti-ations and thereby seek increased market access forits exports of farm (and other) products abroad.Although this factor is not taken into account inthe analysis described here, if WTO membership

enhances China’s chances of expanding its access toagricultural more than other markets abroad in thefuture, that would be a positive benefit of WTOaccession for China’s farmers and rural areas.Martin (forthcoming) points out that Chinese farmexports face particularly high barriers abroad, sothis potential benefit is nontrivial in principle(although in practice it may be difficult to secure,especially if the main barriers are sanitary andphytosanitary measures). That proposition is testedby Yu and Frandsen (2002), also using the GTAPmodel. They find that reductions in barriersto agricultural imports and in domestic support tofarmers in OECD countries reduce the extentto which China’s farm output would fall with WTOaccession, and that in some cases they lead toexpansion rather than contraction of outputs. As aconsequence, China’s agricultural imports wouldfall slightly and its agricultural exports would begreater. These changes are reflected in table 7.12.It shows that not only would China’s food self-sufficiency be higher with than without agricul-tural protection in the European Union, UnitedStates, and Japan, but that the difference is in mostcases more than enough to offset the fall in self-sufficiency that is projected to result from China’sWTO accession. Such reform in the OECD wouldclearly benefit farm households in China, providinga further pro-poor consequence of trade reform.


TABLE 7.12 Effect on Agricultural Self-Sufficiency of China’s WTO Accessionwithout and with Cuts in Agricultural Protection in OECD Countries(domestic production as a percentage of domestic consumption)

China’s WTO Accessionwith EU, United States, and

Base Case (No China’s WTO Accession Japan Removing AgriculturalWTO Accession with No Change in Tariffs, Exports Subsidies,

for China) OECD Policies and Domestic Support

Rice 99.9 99.4 99.6Wheat 93.0 88.2 89.1Feed grains 98.1 98.1 99.6Fruits and vegetables 99.7 99.7 99.9Oilseeds 81.0 81.3 82.3Sugar 77.2 77.4 82.9Plant-based fibers 83.8 83.7 83.9Ruminant meat 88.5 88.2 89.4Nonruminant meat 97.7 97.8 98.9Dairy products 81.8 80.0 84.9

Source: Yu and Frandsen 2002, table 4.

Notes

1. The difficulties China has had in exporting food productsto, for example, Japan, the Republic of Korea, and the UnitedKingdom in recent years because of those countries’ quarantineand sanitary and phytosanitary measures should ease after WTOaccession, or at least be challengeable under WTO dispute settle-ment provisions. Because those measures are notoriously diffi-cult to predict, they are ignored in our empirical analysis (as areprospective changes to other nontariff measures that may limitChina’s access to foreign markets for farm or nonfarm prod-ucts). Also ignored is the possibility that China itself may usequarantine measures to limit its imports of farm products.

2. However, soybean imports grew substantially when thegovernment (in anticipation of WTO accession) lowered thetariff on soybeans from 114 percent to 3 percent in 2000.

3. See Sah and Stiglitz (1992) and Anderson (1995).4. All dollar amounts are current U.S. dollars.5. The GTAP model is a multiregional, static, applied general

equilibrium model based on neoclassical microeconomic the-ory, including full employment of all factors of production, con-stant returns to scale, and perfect competition. For a discussionof the prospect of greater unemployment during the accessionperiod, see Zhai and Wang (2002). The numeraire is the worldprice of exports. See Hertel (1997) for comprehensive documen-tation. The Version 5 database is described at www.gtap.org.

6. A particularly important feature of their analysis is theinclusion of China’s duty exemptions in the base scenario,because, otherwise, the model would overstate the gains fromtariff reductions. Tariff cuts are from the 2001 applied rates tothe postaccession bound rates (or zero if the latter exceed theformer). In this application the aggregate trade balance and gov-ernment tax revenue are both assumed to remain a fixed share ofGDP. The 2001 trade data are from the UN’s COMTRADE data-base, and the 2001 applied tariffs for China are from CDS Con-sulting Co. (2002).

7. Three nonfarm reforms are worthy of mention. The “vol-untary” export restraint on China’s textile and clothing exportsto the United States and European Union, expressed in the basescenario as taxes on those exports, are removed; restructuring ofthe motor vehicle and parts industry after WTO accession ismodeled as a 20 percent productivity boost to vehicle assembly,following François and Spinanger (2004); and liberalization ofChina’s services trade also follows François and Spinanger(2004).

8. In a subsequent analysis, Sicular and Zhao (2004) provideestimates of 2.67 for the “push” elasticity of transformation fromchanges in agricultural returns to changes in the supply of laborout of agriculture, but only 0.6 from changes in nonagriculturalreturns. Ianchovichina and Martin (2004) conducted sensitivityanalysis on the implications of changes in these parameters, andthey found the results are not altered greatly.

9. Wages of skilled workers might increase more than sug-gested here because we do not capture the endogenous produc-tivity growth resulting from the substantial liberalization ofthe service sectors. For a recent study that does incorporatethat effect using a dynamic version of the GTAP model, seeIanchovichina and Walmsley (2003).

10. However, in the northern and southern provinces ofChina the differences could be large because wheat, maize, soy-bean, and cotton are planted primarily in northern China, andrice, horticultural products, livestock, and fish are raised mostlyin southern China.

11. For an empirical analysis using GTAP of the impact ofthe adoption or nonadoption of genetically modified varieties of

first maize and oilseeds and then rice and wheat on China, India,and other countries, see Anderson and Jackson (2003).

12. According to Lin and others (2002), such levies amountto as much as one-quarter of the net incomes of poorer farmersin the hinterlands.

References


Anderson, K. 1995. “Lobbying Incentives and the Pattern of Pro-tection in Rich and Poor Countries.” Economic Developmentand Cultural Change 43 (2): 401–23.

Anderson, K., Y. Hayami, and [others]. 1986. The Political Econ-omy of Agricultural Protection: East Asia in International Per-spective. Boston, London, and Sydney: Allen and Unwin (alsoavailable in an expanded Chinese language edition translatedby F. Cai. Tianjin: People’s Publishing House, 1996).

Anderson, K., and L. A. Jackson. 2003. “Standards, Trade andProtection: The Case of GMOs.” Paper presented at a semi-nar at the World Bank, October 2. Washington, D.C.

Brown, L. R. 1995. Who Will Feed China? Wake-up Call for aSmall Planet. New York: Norton.

CDS Consulting Co. 2002. “Applied Tariffs in China, 2001.”Washington, D.C.

Chen, S., and M. Ravallion. 2004. “Household Welfare Impactsof China’s Accession to the WTO.” In D. Bhattasali, S. Li, andW. Martin, eds., China and the WTO: Accession, PolicyReform, and Poverty Reduction Strategies. Washington, D.C.:World Bank and Oxford University Press.

Fan, S., C. Fang, and X. Zhang. 2001. “How AgriculturalResearch Affects Urban Poverty in Developing Countries:The Case of China.” Environment and Production Technol-ogy Discussion Paper 83. International Food Policy ResearchInstitute, Washington, D.C.

Fan, S., L. Zhang, and X. Zhang. 2002. Growth, Inequality, andPoverty in China: The Role of Public Investments. IFPRIResearch Report 125. Washington, D.C.: International FoodPolicy Research Institute.

Francois, J., and D. Spinanger. 2004. “Regulated Efficiency, WTOAccession and the Motor Vehicle Sector in China.” In D.Bhattasali, S. Li, and W. Martin, eds. China and the WTO:Accession, Policy Reform, and Poverty Reduction Strategies.Washington, D.C.: World Bank and Oxford University Press.

Gilbert, J., and T. Wahl. 2002. “Applied General EquilibriumAssessments of Trade Liberalization in China.” World Econ-omy 25 (5): 697–731.

Hazell, P., and L. Haddad. 2001. “Agricultural Research andPoverty Reduction.” Food, Agriculture, and the Environ-ment. Discussion Paper 34. International Food PolicyResearch Institute, Washington, D.C.

Hertel, T. W., ed. 1997. Global Trade Analysis: Modeling andApplications. Cambridge and New York: Cambridge Univer-sity Press.

Huang, J. 1998. “Evaluation of Current Three-Grain-Policy.”CCAP’s Working Paper WP-00-C01. Center for ChineseAgricultural Policy, Chinese Academy of Agricultural Sci-ences, Beijing.

———. 2001. “Agricultural Policy and Food Security in China.”Center for Chinese Agricultural Policy, Chinese Academy ofSciences, Beijing. Processed.

Huang, J., and H. Ma. 1998. “The 20-Year Reform and the Roleof Agriculture in China: Capital Flow from Rural to Urbanand from Agriculture to Industry.” Reform 5: 56–63.


Huang, J., S. Rozelle, and M. Chang. 2004. “The Nature of Dis-tortions to Agricultural Incentives in China and Implica-tions of WTO Accession.” In D. Bhattasali, S. Li, and W.Martin, eds., China and the WTO: Accession, Policy Reform,and Poverty Reduction Strategies. Washington, D.C.: WorldBank and Oxford University Press.

Hussain, A. 2004. “Coping and Adapting to Job Losses andDeclines in Farm Earnings.” In D. Bhattasali, S. Li, and W.Martin, eds., China and the WTO: Accession, Policy Reform,and Poverty Reduction Strategies. Washington, D.C.: WorldBank and Oxford University Press.

Ianchovichina, E., and W. Martin. 2004. “Economic Impacts ofChina’s Accession to the WTO.” In D. Bhattasali, S. Li, andW. Martin, eds., China and the WTO: Accession, PolicyReform, and Poverty Reduction Strategies. Washington, D.C.:World Bank and Oxford University Press.

Ianchovichina, E., and T. Walmsley. 2003. “Impact of China’sWTO Accession on East Asia.” Policy Research WorkingPaper 3109. World Bank, Washington, D.C.

Kanbur, R., and X. Zhang. 2001.“Fifty Years of Regional Inequal-ity in China: A Journey through Evolution, Reform andOpenness.” CEPR Discussion Paper No. 2887. Centre forEconomic Policy Research, London.

Lardy, N. 2002. Integrating China in the Global Economy.Washington, D.C.: Brookings Institution Press.

Lin, Y. F., R. Tao, M. Liu and Q. Zhang. 2002. “Rural Direct Taxa-tion and Government Regulation in China: EconomicAnalysis and Policy Implications.” Working Paper. ChinaCenter for Economic Research, Beijing University.

Lindert, P. 1991. “Historical Patterns of Agricultural Protection.”In P. Timmer, ed. Agriculture and the State. Ithaca: CornellUniversity Press.

Martin, W. Forthcoming. “Implication of Reform and WTOAccession for China’s Agricultural Policies.” Economies inTransition.

Mathews, A. 2001. “The Possible Impact of China’s WTO Acces-sion on the WTO Agricultural Negotiations.” Trinity Eco-nomic Papers No. 15. Department of Economics, TrinityCollege, Dublin.

McCulloch, N., L. A. Winters, and X. Cirera. 2001. Trade Liberal-ization and Poverty: A Handbook. London: Centre for Eco-nomic Policy Research.

Ministry of Agriculture. China’s Township and Village EnterpriseStatistical Yearbook. Beijing: China Statistics Press.

National Bureau of Statistics of China a. China Statistical Year-book (various issues). Beijing: China Statistics Press.

National Bureau of Statistics of China b. China Rural StatisticalYearbook. Beijing: China Statistics Press.

Rozelle, S., A. Park, J. Huang, and H. Jin. 1997. “Liberalizationand Rural Market Integration in China.” American Journal ofAgricultural Economics 79 (2): 635–42.

Sah, R., and J. E. Stiglitz. 1992. Peasants versus City-Dwellers:Taxation and the Burden of Development. Oxford: ClarendonPress.

Schultz, T. W. 1975. “The Value of the Ability to Deal with Dise-quilibria.” Journal of Economic Literature 13 (3): 827–46.

Shea, Y. P. 2003. The Political Economy of China’s Grain PolicyReform. Ph.D. diss., University of Adelaide.

Sicular, T. 1988. “Plan and Market in China’s Agricultural Com-merce.” Journal of Political Economy 96 (2): 383–87.

Sicular, T., and Y. Zhao. 2004. “Earnings and Labor Mobility inRural China: Implications for China’s WTO Entry.” In D.Bhattasali, S. Li, and W. Martin, eds., China and the WTO:Accession, Policy Reform, and Poverty Reduction Strategies.Washington, D.C.: World Bank and Oxford University Press.

WTO (World Trade Organization). 2001. Report of the WorkingParty on the Accession of China., WT/MIN(01)/3. Geneva.

———. 2003. World Trade Report 2003. Geneva.Yu, W., and S. E. Frandsen. 2002. “China’s WTO Commitments

in Agriculture: Does the Impact Depend on OECD Agricul-tural Policies?” Paper presented at the Fifth Conference onGlobal Economic Analysis, Taipei, June 5–7.

Zhai, F., and Z. Wang. 2002. “WTO Accession, Rural LaborMigration, and Urban Unemployment in China.” UrbanStudies 39 (12): 2199–217.

Zhang, L., J. Huang, and S. Rozelle. 2002. “Employment, Reces-sions, and the Role of Education in Rural China.” China Eco-nomic Review 114: 1–16.

Zhao, Y. 1999.“Leaving the Countryside: Rural-to-Urban Migra-tion Decision in China.” American Economic Review 89 (2):281–86.


cross-subsidization ends, and in the factor markets,as the scale of activity expands but surplus labor isshed. It is, unfortunately, not easy to quantify theseeffects.

It is possible, however, to analyze a key determi-nant of these changes: how the transition to moreopen markets is managed. Already known are thepolicy status quo and the destination as specified inthe commitments, but how much discretion doesChina retain in its choice of policy? China’s sched-ule of commitments specifies not only the terminalsituation, but also how many of the modificationsof policy will be phased in. Even though changesin the degree of competition and the pattern ofownership are largely prespecified, they can beaccelerated—commitments in the General Agree-ment on Trade in Services (GATS) do not prevent acountry from liberalizing faster than it has said itwould. More important, China retains much free-dom in designing domestic regulation, and reformsin this area will crucially affect the magnitude anddistribution of gains from liberalization.

The fulfillment of China’s accession commitmentswill lead to one of the most dramatic episodes ofservices sector liberalization seen in any country.Over the space of some six years, one of the mostclosed services markets has promised to becomeone of the most open. This chapter begins bydescribing China’s policy commitments, and it thenasks whether the disciplines implied by these com-mitments are desirable and, more important, howthe process of implementing these commitments isbest managed.

Successful reforms will lead to significantimprovements in the services markets themselvesin terms of prices, quality, product variety, and theavailability of new products.1 More efficient provi-sion of telecommunications, transport, and otherservices will also advance the integration of theChinese economy with that of the rest of the worldand within the country itself—provided access toimproved services is not limited to the coastalenclaves. The poor are also likely to be affected—inthe product market, as efficiency increases but

117

8

The ServicesDimension of China’s

Accession to the WTO

Aaditya Mattoo

The views contained in this paper are those of the author and should not be attributed to the World Bank.This paper has benefited greatly from the comments of Carsten Fink and two anonymous referees. AntoniaCarzeniga, Stijn Claessens, Caroline Freund, Christina Lund, Will Martin, Cristina Neagu, RandeepRathindran, and Beata Smarzynska also made valuable contributions.

The chapter begins by describing China’s acces-sion commitments. It then examines whether theimplied loss of policy discretion is desirable andwhether China’s commitments promote good policy.Finally, the chapter identifies the priorities for com-plementary reform and offers some conclusions.

China’s Accession Commitments

The GATS Framework and Services Liberalization

The GATS covers all measures taken by membersof the World Trade Organization (WTO) affectingtrade in services and all service sectors.2 The agree-ment is unusual in taking a wide view of what con-stitutes trade, and it defines trade in services as thesupply of a service through any of four modes.Mode 1, cross-border supply, is analogous to trade ingoods and arises when a service crosses a nationalfrontier, for example, the purchase of software orinsurance by a consumer from a supplier locatedabroad. Mode 2, consumption abroad, arises whenthe consumer travels to the territory of a servicesupplier to purchase, for example, tourism, educa-tion, or health services. Mode 3, commercial pres-ence, involves foreign direct investment—for exam-ple, when a foreign bank or telecommunicationsfirm establishes a branch or subsidiary in the terri-tory of a country. And mode 4, movement of indi-viduals, occurs when independent service providersor employees of a multinational firm temporarilymove to another country.

Certain GATS obligations apply across theboard, while others depend on the sector-specificcommitments assumed by individual Members.The most important of the general obligations aretransparency and the most-favored-nation (MFN)principle. The transparency obligation requires,among other things, that each member publishpromptly “all relevant measures of general applica-tion” (that is, measures other than those involvingonly individual service suppliers) to trade in ser-vices. The MFN obligation prevents members fromdiscriminating among their trading partners. Theagreement, however, permits members to list tem-porary exemptions to MFN status.

The liberalizing content of the GATS dependson the extent and nature of sector-specific commit-ments assumed by individual members. The coreprovisions of the GATS in this context relate tomarket access (Article XVI) and national treatment

(Article XVII). These provisions apply only to sec-tors explicitly included by a member in its scheduleof commitments—a “positive list” approach—andthese sectoral commitments, too, are subject to thelimitations that the member has scheduled. GATScommitments are guarantees, and the absence ofsuch guarantees need not mean that access to a par-ticular market is denied in practice.

The market access provision prohibits six typesof limitations, unless they have been inscribed by aMember in its schedule. These are as follows:

• limitations on the number of suppliers• limitations on the total value of service transac-

tions or assets• limitations on the total number of service

operations or on the total quantity of serviceoutput

• limitations on the total number of naturalpersons (i.e. individuals rather than juridicalpersons) that may be employed

• measures that restrict or require specific types oflegal entity or joint venture

• limitations on the participation of foreigncapital.

In scheduled sectors, the existence of any of theselimitations has to be indicated with respect to eachof the four modes of supply described earlier.

In the traditional manner followed by the Gen-eral Agreement on Tariffs and Trade (GATT),national treatment is defined under Article XVII astreatment no less favorable than that accorded tolike domestic services or service suppliers. In con-trast to the GATT approach, however, membersmay inscribe limitations on national treatment intheir schedules with respect to each of the fourmodes of supply, as in the market access provision.

Numerical Overview of Commitments: HowDifferent Is China?

This section compares China’s services commit-ments with those of other WTO members. It beginswith a brief look at the state of access today, beforeChina’s liberalization commitments have beenphased in, for a representative sample of sectors.Figure 8.1 compares China’s commitments foreach mode with those of industrial, developing,and acceding countries. On consumption abroad


(mode 2), commercial presence (mode 3), and thepresence of natural persons (mode 4), China hasmade at least partial commitments in all the sec-tors, and on cross-border supply (mode 1) for over80 percent of the sectors. This situation comparesfavorably with the commitments of all other coun-try groups. However, the number of sectors with aguarantee of full access in China’s schedule today isless than those for the other country groups, with avery significant difference for mode 3.

The picture changes quite drastically whenlooking at the period after all the liberalizationhas been phased in. Three indicators are computedby Ianchovichina, Martin, and Wood (2001) forboth market access and national treatmentcommitments:

• “unweighted average count,” indicating thenumber of sector/mode-of-supply combina-tions where a commitment was made about themaximum number possible

• “average coverage,” providing an arithmeticweighted average of the weights (0 for no com-mitment, 1 for a full commitment, and 0.5 for allpartial commitments) allocated to each cell

• the “no-restriction commitments share” in acountry’s total commitments as well as in themaximum possible commitments.

Overall for China, the coverage of market accesscommitments (i.e., the unweighted average count)was 57.4 percent (table 8.1). This coverage is muchmore extensive than the commitments offered inthe Uruguay Round by any other group of countries(including high-income ones). The “average cover-age,” a measure of coverage that better reflects theextent of liberalization of services, was 38.1 percentfor China—again showing more openness than thatof even the high-income countries. China’s share ofcompletely liberal commitments (i.e., no restric-tions) of the maximum possible was 23.1 percent,which is much higher than that for any other groupof developing countries but somewhat lower thanthat for high-income countries. China’s commit-ments on national treatment are deeper and widerthan those of all other country groups.

A Closer Look, Sector by Sector

Typical Restrictions In some respects, China’scommitments resemble those made by other coun-tries. For most sectors, modes 1 and 2 are eitherfully open or unbound and are not subject tospecific restrictions. Commitments on mode 4,specified horizontally rather than sector by sector,are also standard: entry is guaranteed only for man-agers, executives, and specialists—who must be

The Services Dimension of China’s Accession to the WTO 119

0

10

20

30

40

50

60

70

80

90

100

Percent

IC DC AC CHNMode 1 Mode 2 Mode 3 Mode 4

IC DC AC CHN IC DC AC CHN IC DC AC CHN

FIGURE 8.1 Country Services Commitments by Mode

IC � industrial countries, DC � developing countries, AC � acceding countries, CHN � China.Note: Calculated on the basis of a sample of 37 sectors deemed representative for various service sectors. Seechapter appendix table 8.3 for details on China and WTO (1999) for details on other countries. The upper partof each bar represents partial commitments, the lower part full commitments.Source: WTO.

either intracorporate transferees or employees offoreign invested enterprises—and for services sales-persons on exploratory business visits. No commit-ments are made on other categories of natural per-sons such as unskilled personnel or movement notlinked to a commercial presence.

It is for the commitments on commercial pres-ence that a range of restrictive measures emerge.They are related to the following:

• Form of establishment. The typical restrictionis the requirement to form a joint venture thatis either an equity joint venture (EJV) or acontractual joint venture (CJV). Foreign owner-ship in EJVs is frequently restricted to speci-fied levels, ranging from minority ownership

(49 percent or less), to 50 percent ownership, tomajority ownership, to full ownership. There isno commitment to allowing establishment ofbranches by foreign enterprises, except in spe-cific sectors.

• Geographic scope. Business activity may beallowed only in specified cities (e.g., Shanghai)or in special economic zones (SEZs).

• Business scope. Transactions may be permittedonly with a subset of consumers or restricted insome other way.

• Regulatory requirements. Foreign firms may berequired to possess a certain minimum amountof assets and be established as a representativeoffice for a certain period of time before com-mencing full business operations.


TABLE 8.1 Coverage of Specific Commitments by China to Services in Trade (percent)

Low- and LargeHigh-Income Middle-Income Developing

Countries Countries Nations China

Market accessUnweighted average count (sectors/modes 47.3 16.2 38.6 57.4

listed as a share of maximum possible)Average coverage (sectors/modes listed as a 35.9 10.3 22.9 38.1

share of maximum possible, weighted by openness or binding factors)

Coverage/count (average coverage as a share 75.9 63.6 59.3 66.4of the average count)

No restrictions as a share of total offer 57.3 45.5 38.7 40.2(unweighted count)

No restrictions as a share of maximum possible 27.1 7.3 14.9 23.1

National treatmentUnweighted average count (sectors/modes 47.3 16.2 38.8 57.4

listed as a share of maximum possible)Average coverage (sectors/modes listed as a 37.2 11.2 25.5 45.0

share of maximum possible, weighted by openness or binding factors)

Coverage/count (average coverage as a share 78.6 69.1 66.1 78.4of average count)

No restrictions as a share of total offer 65.1 58.0 52.3 63.5(unweighted count)

No restrictions as a share of maximum possible 30.8 9.4 20.2 36.5

Memo itemNo restrictions on market access and national 24.8 6.9 14.3 29.8

treatment as a share of maximum possibleNumber of sectors committed 293.0 100.0 239.0 356.0

Source: Ianchovichina, Martin, and Wood 2001.Note: The breadth and depth of commitments by other countries are understated because their more recentcommitments in telecommunications and financial services have not been taken fully into account.

Interestingly, most restrictions pertain to marketaccess; relatively few limitations affect nationaltreatment.3 In fact, one of the striking aspects ofChina’s schedules is the country’s willingness tocommit across modes and sectors to full nationaltreatment for foreign providers.

Commitments: Past, Present, and Future Howmuch has China’s trade policy already changed, andhow much will it change over the next few years?China’s schedules of commitments provide a firstsource of information. China participated in theUruguay Round services negotiations and submit-ted a schedule under GATS in April 1994. BecauseChina did not become a member of the WTO atthis time, this schedule did not have legal status.More important, it is not clear how much thisschedule reflected the actual openness of Chineseservices markets at that time. Nevertheless, it doesprovide the basis for an interesting comparisonwith the final schedule.

China’s schedule of commitments under GATShas had legal status ever since China became aWTO member, and it is the outcome of the tough-est services negotiations undertaken in the WTO.Whereas most WTO members have merely boundthe policy status quo or even less, China agreed toallow significant liberalization to be implementedeither immediately or in the near future. One con-sequence is that the schedule describes the state ofactual policy at the time of accession and over thenext few years.

Relying on these two schedules, table 8.2 revealsa rough picture of the state of “policy” at threepoints of time: 1994; 2001, the date of accession;and 2008, the date by which all liberalization com-mitments will have been phased in (i.e., seven yearsafter accession). The earlier date shows how farChina has already come, and the latter date showshow far it is committed to go.

Professional Services Today, commitments onprofessional services are far more liberal than in1994, and the sector will be highly liberalized overthe next few years. Unlike in 1994, China is nowcommitted to allowing the cross-border supplyof professional services; quantitative limitationsno longer apply to accountancy firms; taxationservices can be offered outside of “economically

developed areas”; and there are meaningful com-mitments in urban planning and legal services.

By 2007 geographic and quantitative limitationswill be eliminated in legal services, and fully ownedforeign subsidiaries can operate in accounting, tax-ation, architecture, engineering, and urban plan-ning services. But some restrictions will persist,especially in legal and medical services. Foreignfirms are not allowed directly to participate in legalactivity in China and are only entitled to work onlegal affairs related to their home country or toentrust work to Chinese firms on behalf of theirclients. In medical services, hospitals cannot befully foreign-owned and are still subject to quanti-tative limitations in line with China’s assessedneeds.

Computer and Related Services This is the onesector where some commitments have actuallybecome less liberal. In 1994 there were no restric-tions on establishment in software implementa-tion, systems and software consulting, and systemsanalysis. Now establishment can only take placethrough joint ventures, although foreign majorityownership is permitted. Cross-border delivery inthese services remains unbound, but it is fully openin all other computer and related services. Nofuture liberalization has been specified.

Telecommunications Much change is anticipatedin this sector over the next few years. Today foreignproviders can provide value added and mobile ser-vices in and between Shanghai, Guangzhou, andBeijing through joint ventures, subject to stringentrestrictions on foreign ownership. By 2004 fixed-line services can be provided on similar terms inand between the same cities. By 2007 all geographicrestrictions will be eliminated and equity restric-tions will be relaxed. But even at the end of theperiod majority foreign ownership will not beallowed in any area. Furthermore, there is nocommitment to allow cross-border any of theseservices.4

Construction and Engineering Services Whereascommercial presence was unbound in 1994, today itis allowed through joint ventures with foreignmajority ownership permitted, but only in foreign-invested construction projects. By 2004 full foreign



TABLE 8.2 China’s Commitments: Past, Present, and Future

Sector 1994 2001 2008

Professional services

Legal services

Accounting,auditing, andbookkeepingservices

Taxation

Architecture andengineering

Urban planning(excludinggeneral urbanplanning)

Medical anddental services

CONTINUED RESTRICTIONSON BUSINESS SCOPE.

Mode 3: Geographic andquantitative limitationswere to be eliminated by 2002.

FULLY LIBERALIZED Mode 3: None; wholly

foreign-owned subsidiariespermitted by 2007.

FULLY LIBERALIZED EXCEPT FOR MODE 1RESTRICTIONS.

Mode 3: Wholly foreign-owned subsidiariespermitted by 2006.

FULLY LIBERALIZED EXCEPT FOR MODE 1RESTRICTIONS.

Mode 3: Wholly foreign-owned subsidiariespermitted by 2006.

FULL FOREIGN OWNERSHIPNOT ALLOWED ANDNEEDS-BASED QUOTAS.

No commitments

Modes 1 and 2: UnboundMode 3: Through branch

offices and CJVs subject tolimitations on minimumsize, aggregate number(15), and geographic scope(SEZs). Auditing reports areonly valid if a Chinese CPAtitle is obtained.

Mode 1: NoneMode 2: UnboundMode 3: Through branch

offices subject to limitationson minimum size andgeographic scope (SEZs).

Mode 1: UnboundMode 2: NoneMode 3: Only through an EJV

or CJV. Registered in owncountry.

Mode 1: UnboundMode 2: NoneMode 3: Unbound

Mode 1: UnboundMode 2: UnboundMode 3: Only through an EJV

or CJV with a quantitativelimitation based on a needstest and approval by theMinistry of Public Healthand Ministry of ForeignTrade and EconomicCooperation. CJV or EJVsolely responsible forforeign exchange balanceand profits and losses.Majority of personnel must be Chinese.

Mode 4: Foreign doctors canprovide services for sixmonths (may extend to ayear) provided a license isobtained at provincial leveland they are contracted by Chinese medicalinstitutions.

Modes 1 and 2: NoneMode 3: Only through one rep.

office which is allowed toengage in profit-makingactivities, but only in specifiedcities. Business scope restrictedto home country legal affairsfor Chinese and China-basedclients and to entrusting, onbehalf of foreign clients,Chinese law firms to deal withChinese legal affairs.

FULLY LIBERALIZED except thatpartnerships and incorporatedaccounting firms are limited to CPAs licensed by Chineseauthorities.

Mode 2: NoneMode 3: Only through CJVs, with

majority foreign ownershippermitted.

Mode 1: None for schemedesign; otherwise, cooperationwith Chinese professionalorganizations is required.

Mode 3: Only through an EJV orCJV. Registered in own countryand engaged in architecture/engineering services in homecountry.

Mode 1: None for schemedesign; otherwise, cooperationwith Chinese professionalorganizations is required.

Mode 3: Only through an EJV orCJV.

Mode 1: NoneMode 2: NoneMode 3: Foreign majority

ownership explicitly permitted.Not required to accept soleresponsibility for foreignexchange balance and profitsand losses, but still subject toquantitative limitations basedon a needs test.

Mode 4: Licenses can beobtained from the Ministry ofPublic Health, and a contract is not required.



Sector 1994 2001 2008

Computer andrelated services

Consultancyservices relatedto theinstallation ofcomputerhardware

Data processingand tabulation

Time-sharingSoftwareimplementation

Systems andsoftwareconsulting

Systems analysis

Systems designProgrammingSystems

maintenanceData processingInput preparation

Telecommunica-tions

Value added

Basic telecom-munications:mobile voiceand data

Basic telecom-munications:fixed-lineservices

Mode 1: UnboundMode 2: NoneMode 3: NoneMode 4: Qualifications:

B.A. and five years ofexperience.

Mode 1: UnboundMode 2: NoneMode 3: NoneMode 4: Qualifications:


Mode 1: UnboundMode 2: NoneMode 3: Through EJV only.Mode 4: Qualifications:


No commitments

No commitments

No commitments

Modes 1–3: FULLY LIBERALIZEDMode 4: Qualifications: B.A. and

three years of experience.

Mode 1: UnboundMode 2: NoneMode 3: Only through JVs with

foreign majority ownershippermitted.

Mode 4: Qualifications: B.A. andthree years of experience.

Mode 1: NoneMode 2: NoneMode 3: Only through JVs with


Mode 4: Qualifications: B.A. andthree years of experience.

Mode 1: UnclearMode 2: NoneMode 3: Through JVs with a

foreign investment limit of 30 percent only in Shanghai,Guangzhou, and Beijing.

Mode 1: UnclearMode 2: NoneMode 3: Through JVs with a

foreign investment limit of 25 percent only in andbetween Shanghai,Guangzhou, and Beijing.

Mode 1: UnclearMode 2: NoneMode 3: Unbound

By 2002 expansion ingeographic area andforeign investment limitto 49 percent.

By 2003 no geographicrestriction and FOREIGNINVESTMENT LIMIT TO50 PERCENT.

By 2002 expansion ingeographic area andforeign investment limitto 35 percent.

By 2004 FOREIGNINVESTMENT LIMIT TO49 PERCENT.

By 2006 no geographicrestriction.

By 2004 through JVs with aforeign investment limitof 25 percent only in and between Shanghai,Guangzhou, and Beijing.

By 2006 expansion ingeographic area, andforeign investment limitto 35 percent.

By 2007 no geographicrestriction and FOREIGNINVESTMENT LIMIT TO49 PERCENT.



Sector 1994 2001 2008

Construction

Construction and relatedengineering

Distribution

Commissionagents andwholesaletrade, and afull range ofsubordinatedservices,including after-salesservices.

Retailing and afull range ofsubordinatedservices,including after-salesservices.

Franchising

Educational andenvironmentalservicesEducational

services,excludingspecialeducation(e.g., militaryand political)and nationalcompulsoryeducation

Environmentalservices

RESTRICTIONS ONBUSINESS SCOPE OFFULLY FOREIGN-OWNEDENTERPRISES.

Mode 3: By 2004 fullyforeign-owned enterprisespermitted but only inprojects financed byforeign investment and/orgrants, or by loans frominternational financialinstitutions, or thoseprojects that aretechnically difficult forChinese enterprises.

LIBERALIZED EXCEPTCROSS-BORDER DELIVERYAND TWO PRODUCTS.

Mode 3: By 2002 throughJVs subject to restrictionson products, to bephased out by 2006(except salt and tobacco).By 2003 foreign majorityownership allowedand no geographic orquantitative restrictions.

CONTINUEDRESTRICTIONS ON LARGECHAIN STORES.

Mode 3: By 2003 allprovincial capitals open.By 2004 no moregeographic restrictions.By 2006 no restrictionson products. Foreignmajority control allowedexcept in chain storeswith more than 30 outletsselling a range ofproducts.

FULLY LIBERALIZED BY 2004.

Mode 3: By 2004 none.

Mode 1: UnboundMode 2: NoneMode 3: Unbound

No commitments

No commitments

No commitments

Mode 1: UnboundMode 2: NoneMode 3: UnboundMode 4: Subject to licensing

from the State Bureau ofForeign Experts (SBFE) and State EducationCommission (SEC), andpossession of M.A. andprofessional title.

No commitments

Mode 1: UnboundMode 2: NoneMode 3: Through JVs with foreign

majority ownership permittedand only foreign-investedconstruction projects.

Mode 1: UnboundMode 2: NoneMode 3: Foreign-invested

enterprises are permittedto distribute their productsmanufactured in China.

Mode 1: Unbound except formail order.

Mode 2: NoneMode 3: Through JVs (not foreign

majority controlled) in five SEZsand eight cities subject toquotas (e.g., four in Beijing and Shanghai), restrictions on products (not books,newspapers, pharmaceuticals,pesticides, chemical fertilizers, etc.).

Mode 1: NoneMode 2: NoneMode 3: Unbound

Mode 1: NoneMode 2: NoneMode 3: Only through JVs with

foreign majority ownershippermitted (national treatment:unbound).

Mode 4: Subject to invitation oremployment by Chineseinstitution and possession ofB.A., two years of experience,and professional title.

Mode 1: Unbound except forconsultation services.

Mode 2: NoneMode 3: Through JVs with foreign

majority ownership permitted.


Sector 1994 2001 2008

Financial services

Insurance(exceptstatutoryinsurance)

Banking

Securities

Tourism andtravel-relatedservices

Hotels

Travel agencyand touroperator

Mode 1: UnboundMode 2: UnboundMode 3: Through a branch or

JV only in Shanghai, subjectto minimum local andglobal asset and localpresence (as rep. office)requirements.

Mode 1: UnboundMode 2: UnboundMode 3: Through a branch,

subsidiary JV only inspecified regions, subject tominimum asset and localpresence (as rep. office)requirements; acceptance ofdeposits only from nonresi-dents in foreign currencies(with some exceptions) andno loans to Chinese citizens.

No commitments

Mode 1: UnboundMode 2: NoneMode 3: Through JVs subject

to needs test at central andlocal levels.

No commitments

Mode 1: Unbound except forinternational maritime, aviation,and transport insurance andreinsurance, and certain types ofbrokerage.

Mode 2: None, but unbound forbrokerage.

Mode 3: Form of establishment:non-life: through a branch orJVs with 51 percent foreignownership; life: through JVs with50 percent foreign ownership.

Geographic limitation: only infive cities.

Business scope: only selected forms of non–life insurance. Lifeonly to individuals, not groups.

Licenses: no quotas but subject tominimum asset and duration ofestablishment requirements.

Upon accession, a 20 percentcession required of all lines of theprimary risks for non-life,personal accident, and healthinsurance business with anappointed Chinese reinsurancecompany.

Mode 1: Unbound except forprovision of data, advice, etc.

Mode 2: None.Mode 3: Geographic limitation:

none for foreign currencybusiness, but local currency onlyin four cities. Interregionalsupply of services permitted.

Clients: only foreign currencybusiness.

Licenses: only prudential criteria.

Mode 1: Unbound except B sharebusiness.

Mode 2: NoneMode 3: Unbound, except rep.

offices may become specialmembers of the Chinese StockExchange (CSE)s, and throughJVs with up to 33 percent foreignownership to conduct domesticsecurities investment fundmanagement business.

Mode 1: UnboundMode 2: NoneMode 3: Through JVs with foreign

majority ownership permitted.

Mode 1: UnboundMode 2: NoneMode 3: Through JVs subject to

geographic and business scoperestrictions.


By 2004, FULLYLIBERALIZED EXCEPT 50 percent FOREIGNOWNERSHIP LIMIT IN LIFEINSURANCE.

Mode 3: By 2003 noestablishment restrictionsin non-life.

By 2004 no geographicrestrictions.

By 2004 no restrictions onbusiness scope

By 2005 no cessionrequirement.


Mode 3: Geographiclimitations phased outgradually by 2006.

Clients: local currencybusiness with Chineseenterprises by 2003 andall clients by 2006.

Mode 3: By 2004 49 percentforeign ownership in JVs to conduct domesticsecurities investment fundmanagement business,and through JVs with upto 33 percent foreignownership to underwrite Ashares and underwrite andtrade B and H shares, aswell as government andcorporate debts,launching of funds.




Sector 1994 2001 2008

Transportservices

MARITIME

TRANSPORT

Internationaltransport

Auxiliaryservices

INTERNAL

wATERWAYS

AIR TRANSPORT

Aircraft repairandmaintenance

Computerreservations

SPACE

TRANSPORT

RAIL TRANSPORT

ROAD TRANSPORT

(FREIGHT)

SERVICES

AUXILIARY TO

ALL MODES OF

TRANSPORT

Storage andwarehousing

Freightforwardingagencyservices

FULLY LIBERALIZED BY 2007.Mode 3: Majority

ownership by 2004.


ownership by 2002.


ownership by 2002.


ownership by 2002.

Mode 1: NoneMode 2: NoneMode 3: Unbound

Mode 1: UnboundMode 2: NoneMode 3: Through JVs only.

Mode 1: Only internationalshipping in ports open toforeign vessels permitted.

Mode 2: NoneMode 3: Unbound

No commitments

No commitments

No commitments

No commitments

Mode 1: UnboundMode 2: NoneMode 3: Through JVs subject

to an economic needs test.

Commitments only formaritime transport, asabove.

Commitments only formaritime transport, asabove.

Mode 1: NoneMode 2: NoneMode 3: Through JVs subject to

49 percent foreign ownershiplimits to operate only a Chineseflag fleet

Mode 1: UnboundMode 2: NoneMode 3: Through JVs only with


As in 1994.

Mode 1: UnboundMode 2: NoneMode 3: Through Chinese

controlled JVs and subject to aneconomic needs test.

Mode 1: By connection withChinese computer reservationsystem, etc.

Mode 2: NoneMode 3: Unbound

No commitments

Mode 1: NoneMode 2: NoneMode 3: Through JVs with a

foreign ownership limit of 49 percent.



Mode 1: UnboundMode 2: NoneMode 3: Through JVs with a



foreign ownership limit of 50 percent and subject tominimum capital requirements.

JV � joint venture; CJV � contractual joint venture; EJV � equity joint venture; SEZ � special economic zone; CPA � certifiedpublic accountant.Source: Compiled by the author from China’s schedules of commitments under the GATS.



ownership will be permitted but subject to certainrestrictions on business scope.

Distribution Services No commitments at allwere shown in the 1994 schedule. Substantial liber-alization has already taken place, but restrictionspersist on establishment (only through jointventures), geographic scope (retailing in only fivespecial economic zones) and products sold (e.g.,not books, newspapers, pharmaceuticals, and pesti-cides). By 2006 the sector will be largely open. Per-haps most strikingly, China has agreed to open upthe whole logistical chain of related services,including inventory management; assembly, sort-ing, and grading of bulk lots; breaking bulk lots andredistributing into smaller lots; delivery services;refrigeration, storage, warehousing, and garageservices; sales promotion, marketing and advertis-ing; and installation and after-sales services, includ-ing maintenance and repair and training services.No WTO member has made such deep commit-ments in this sector.

Some restrictions will remain. Salt and tobaccoare excluded from the scope of commitments oncommission agents and wholesalers. Furthermore,majority foreign ownership will not be allowed inretail chain stores that sell multiple products anddifferent brands of products such as books, news-papers, pharmaceuticals, and chemical fertilizersand have more than 30 outlets. There are no com-mitments on the cross-border supply of commis-sion agents’ services and delivery of wholesale tradeservices. More interestingly, cross-border supply ofretail services is allowed only through mail order,which presumably covers electronic commerce.

Educational Services Cross-border delivery wasunbound in 1994, but it is now fully open.Commercial presence was also unbound, but it nowcan be established through joint ventures with for-eign majority ownership permitted. National treat-ment is not, however, guaranteed for foreign educa-tional institutions.

Financial Services The 1994 commitments speci-fied that insurance services could be suppliedthrough a branch or joint venture only in Shanghai,subject to conditions pertaining to minimum capi-tal and prior presence, globally (30 years as aninsurance company) and locally (three years as a

representative office). On accession, non–life insur-ers are permitted to open a branch or joint venturewith 51 percent foreign ownership, whereas lifeinsurers are permitted 50 percent ownership of ajoint venture with a partner of their choice.Non–life insurers can provide “master policy”insurance and insurance of large-scale commercialrisks without geographic restrictions, and insur-ance of enterprises abroad as well as propertyinsurance, related liability insurance, and creditinsurance of foreign-invested enterprises in fivecities: Shanghai, Guangzhou, Dalian, Shenzhen,and Foshan. Life insurers are permitted to provideindividual (not group) insurance to foreigners andChinese citizens in the same five cities. By 2004 allrestrictions will disappear except the foreign own-ership limit on life insurers. Licenses are to beawarded solely on the basis of prudential criteria,with no application of quantitative limitations oreconomic needs tests.

Under the 1994 commitments, foreign bankscould only operate in specified regions, acceptdeposits only from nonresidents and only in for-eign currencies (with some exceptions), and makeno loans to Chinese citizens. On accession, geo-graphic and client limitations will be eliminated forthe foreign currency business. Even though theschedule states that on accession local currencybusiness will be allowed in four cities (Shanghai,Shenzhen, Tianjin, and Dalian), there seems to be abinding restriction on clients that will only berelaxed in two years. The entire banking sector willbe fully liberalized by 2006. As for insurance,licenses are to be awarded solely on the basis ofprudential criteria, with no quantitative limitationsor economic needs test applied.

There were no commitments on securities in1994. On accession, joint ventures with up to 33 per-cent foreign ownership will be allowed to conductbusiness in domestic securities investment fundmanagement. By 2004 foreign ownership of suchventures will be allowed to increase to 49 percent.Furthermore, joint ventures with up to 33 percentforeign ownership will be allowed to underwritedomestic equity issues and underwrite and trade ininternational equity and all corporate and govern-ment debt issues.

Grandfather Provisions One of the key difficul-ties in the accession negotiations arose because of

the presence in the Chinese insurance market offirms that enjoyed better conditions than Chinawas prepared to guarantee to new entrants. In par-ticular, the U.S. life insurance firm AIG was estab-lished in Shanghai with full foreign ownership;Allianz (German), Axa (French), and Manulife(Canadian) were 51 percent foreign-owned. Asnoted earlier, the best China was prepared to offernew entrants was entry through joint ventures thatwere 50 percent foreign owned. Apparently thisasymmetry in itself was generally acceptable andled to the following grandfather provision inChina’s schedule:

The conditions of ownership, operation andscope of activities, as set out in the respectivecontractual or shareholder agreement or in alicense establishing or authorizing the opera-tion or supply of services by an existing foreignservice supplier, will not be made more restric-tive than they exist as of the date of China’saccession to the WTO.

The problem arose because of the implicationsfor branching rights. The key element of China’scommitment was the guarantee that “internalbranching for an insurance firm will be permittedconsistent with the phase out of geographic restric-tions.” In effect, AIG would be able to expandoperations through branches of its fully ownedsubsidiary in China, whereas other new firms couldonly do so through branches of their 50 percent–owned joint ventures. This situation was apparentlynot acceptable to the European Union. The even-tual compromise was a rather convoluted footnotein the schedule, which states:

Any further authorization provided to foreigninsurers after accession under more favorableconditions than those contained in this schedule(including the extension of grandfathered invest-ments through branching, sub-branching or anyother legal form), will be made available to otherforeign service suppliers which so requested.

This solution is messy and could lead to adispute. If AIG is denied the right to branch from itsfully owned subsidiary, the United States can claimthat the assurance contained in the grandfatherand internal branching commitments is not beingrespected. If AIG is allowed the right to establishbranches and the European Union is denied the

right to establish fully owned subsidiaries, then thelatter can claim that the assurance contained inthe footnote is not being respected.5

Transport Services Liberal commitments hadalready been made in international maritime trans-port and certain supporting services, even in 1994.The major change is that the greater depth of com-mitments across the whole range of transport andauxiliary services will significantly facilitate theprovision of multimodal transport services. In road,rail, and key auxiliary services, notably storage andwarehousing and freight forwarding agency ser-vices, restrictive or no commitments will bereplaced on accession by the requirement to enteras joint ventures, and by 2007 there will be full lib-eralization. The exception will be hard rights in airtransport, which are excluded from the scope of theGATS.

Do China’s Commitments PromoteGood Policy?

The commitments just described imply a dramaticloss of discretion in policymaking. China has prom-ised to give up over the next few years much of thefreedom it had to restrict new entry and foreignownership, to discriminate between trading part-ners and in favor of its own firms, and, more broadly,to change its mind. The central argument of this sec-tion is that, for the most part, the discipline impliedby the commitments is desirable. However, the per-sistence of restrictions on foreign ownership maydeprive China of the full benefits of foreign invest-ment, and the scheduled sequence of liberalizationacross regions may accentuate regional inequalities.Moreover, as discussed in the next major section, thesize and distribution of the gains from liberalizationwill depend on how China uses the freedom it stillhas to strengthen its regulatory framework andchoose an appropriate sequence of reform. Giventhe limited empirical evidence and the uniquenessof China’s experience, these issues are at this stagenot much more than plausible hypotheses thatdefine an agenda for future research.

Eliminating Discretion

In many areas the Chinese government has beenreluctant to liberalize immediately. Some of the


protection is probably a consequence of politicaleconomic pressures from vested interests. In somecases, the government evidently has felt the need toprotect the incumbent suppliers from competitionbecause of infant industry–type arguments, tofacilitate an “orderly exit,” or to allow complemen-tary reforms sufficient time—all of which are argu-ments for temporary protection. But protectiononce granted can be difficult to remove. The failureof infant industry policies in the past and the innu-merable examples of perpetual infancy (or senility)are attributable in part to the inability of theChinese government to commit itself to liberalizeat some future date and thus to confront incum-bents with a credible deadline. The binding com-mitments under the GATS to provide market accessby a precise future date should play a valuable rolein overcoming the credibility problem. Failure tohonor these commitments creates an obligation tocompensate those who are deprived of benefits,making the commitment more credible than amere announcement of liberalizing intent in thenational context. The price in terms of the loss offlexibility would seem to be worth paying.

It is also worth emphasizing that GATS commit-ments represent only a ceiling on protection andnot a floor. China has given up the freedom to liber-alize more slowly than specified in its GATS com-mitments, but it still has the freedom to move faster.

Eliminating Discrimination

Discrimination either between trading partners orin favor of domestic firms results in economicallycostly distortions, and so its prohibition is generallydesirable. The two pillars of nondiscriminationunder the GATS are the MFN and national treat-ment obligations, but both allow exceptions to belisted. China’s MFN exemptions are relatively nar-row. The only one listed applies to internationalmaritime transport: to permit cargo-sharing agree-ments only with certain countries and to allowjoint ventures and wholly owned shipping sub-sidiaries to be formed on the basis of bilateralagreements. In any case, because China has com-mitted to full market access and national treatmenton cross-border supply, there seems to be littlescope for discriminatory cargo sharing.

There are also surprisingly few limitations onnational treatment, and so, for the most part, China

has given up the right to offer preferential treat-ment to domestic enterprises through any measureaffecting the supply of services. Virtually no scopeexists for discriminatory taxation, and discrimina-tory subsidies can be awarded only in aviation,audiovisual, and medical services—and there, too,only to the extent that such subsidies already exist.The other instances of discriminatory provisionsare the following: all legal representatives arerequired to be resident in China for no less than sixmonths each year; a majority of doctors in jointventure hospitals must be Chinese; the existing reg-istered capital requirements for joint venture con-struction enterprises are slightly different fromthose of domestic enterprises; joint venture travelagencies are not allowed to provide services toChinese traveling abroad; and foreign insurers aresubject to a 20-percent cession requirement with aChinese reinsurance company, to be phased out infour years.

Although the scope for explicit discriminationhas shrunk, the nature of measures affecting tradein services does offer considerable scope forimplicit discrimination. First, there is the possibil-ity of variable treatment through domestic regula-tions such as licensing, recognition of qualifica-tions, and various other technical regulations. Suchdiscretion is likely to be constrained, but probablynot eliminated, by the assurance that licenses infinancial services will be issued solely on the basisof prudential criteria and by the requirement fortransparency of licensing criteria and decisionprocesses in other areas. Furthermore, Article VII ofthe GATS, which deals specifically with recognitionagreements pertaining to educational qualifica-tions, licenses, and such, strikes a delicate balanceby allowing such agreements, provided they are notused as a means of discrimination, and third coun-tries have an opportunity to accede or demonstrateequivalence.

Second, China has retained the freedom toimpose explicit quantitative restrictions on thenumber of providers in legal, medical, and retailingservices. De facto quotas also may exist in someareas, such as those imposed by the scarcity of radiospectrum needed for the provision of mobiletelecommunications services and the scarcity ofspace needed for department stores or airportsin a city. Apart from the general MFN obligation,no specific rules are in place to ensure the


nondiscriminatory allocation of quotas in services(WTO rules are, however, in place for goods). In the past, this lack of rules was not a major issuebecause commitments reflected the status quo, andquotas, particularly those for service suppliers, weredescriptions of the existing market structure. Butwhen genuine liberalizing commitments are made,like those by China, the nondiscriminatory alloca-tion of quotas is bound to be an important issue.The instinctive candidate for a nondiscriminatoryrule is an auction, which also has the virtue oftransferring quota rents to the government. Itremains to be seen whether the Chinese regulatorscan resist the temptation to resort to other, morediscretionary methods of allocating quotas.

Adhering to the principle of nondiscriminationis important, even though it may seem at first sightthat there is no real cost to granting preferentialtreatment in services. Because the protective instru-ment is often a restrictive regulation that does notgenerate revenue (or rents), there is no cost togranting preferential access because there is no rev-enue (or rents) to lose. When the protective instru-ment is a quota, the implications of preferencesdepend on who appropriates the rents. Where rentsare appropriated by exporters or dissipated, prefer-ential liberalization is again necessarily welfare-enhancing for the importing country.

Preferences in services may, however, imposeother, more subtle costs. The cost of trade diversionmay well be the establishment of poorer-qualityproviders, not just in terms of poorer-qualityservices but also in terms of smaller social benefitsin the form of knowledge and technologicalspillovers. Furthermore, the greater importance ofsunk costs in some service sectors, ranging frombasic telecommunications to financial services, sug-gests that preferential liberalization may have moredurable consequences than in the case of goods. Forexample, allowing the second-best provider toestablish a business may mean that a country isstuck with such a provider even when it subse-quently liberalizes on an MFN basis.6

These considerations suggest that China woulddo well to resist the reported pressure to grant pref-erential access to the service suppliers of certainWTO members. Such pressure resulted in, for exam-ple, the grandfather provisions in China’s commit-ments on insurance services, described earlier in thischapter. These provisions reflect an emphasis on

guaranteeing the rights of incumbents. Insofar asthey provide the benefits of security to investors whoare already present in the market rather than to newinvestors, they may not do enough to make marketsmorecontestable.Newentrantsmayevenbeplacedata competitive disadvantage insofar as differences inownership and legal form affect firm performance.Forexample,largerequitysharesmaymakeiteasiertoexercise effective control over the operations of a firmand ensure efficient production, and the marginalcost of providing a service through a branch may belower than through a subsidiary. Fortunately, the factthat several foreign and domestic firms are alreadycompeting in the market may limit the impact ofthese grandfather provisions.

Eliminating Barriers to Entry

Apart from the few instances just noted, China has chosen not to limit the number of serviceproviders. The only barrier to new entry is therequirement to enter as a joint venture, which isdiscussed in the next section. But two questionsarise: Is it good policy to allow unrestricted entry inall service sectors? Will China’s commitmentsdeprive the government of an important policyinstrument? One reason for entry restrictionsmight be the existence of significant economies ofscale. For example, networks incur substantial fixedcosts, and competitive entry could lead to ineffi-cient network duplication.7

For several reasons, however, entry restrictionsare increasingly difficult to defend in principle inthe face of technological change and in the face ofmounting evidence that competition works. First,entry restrictions change the nature of interactionbetween incumbents and may well make collusionmore likely. Second, such restrictions dampen theimpact of competition on productive efficiency.Third, the regulator is usually not better placedthan the competitive process to determine the opti-mal number of firms in the market, especially giventhe difficulty of obtaining information about thecost structure of firms and other sources of regula-tory failure. Fourth, technological advances havesignificantly lowered network costs, even in a sectorlike fixed-line telecommunications, and verticalseparation (e.g., through network unbundling) haswidened the scope for competitive entry (Smith1995). Inefficiencies introduced by the duplication


of networks may be small compared with the oper-ational inefficiencies that can result from a lack ofcompetitive pressure.8 On this basis, it would seemthat there is little reason to worry about excessiveentry; rather, the priority should be eliminating thebarriers to entry where they remain (e.g., in med-ical services and retailing).

Eliminating Restrictions on Foreign Ownership

The most important restriction on foreign presencein China is the requirement to enter as a joint ven-ture, often with limits on the extent of foreign own-ership. In most areas these limits on ownership arebeing gradually phased out, but in some areas suchas telecommunications and life insurance they willremain even after all the liberalization commit-ments have been phased out. What is the rationalefor such restrictions, and what are their implica-tions? Furthermore, what are the consequences ofthe manner in which they are being phased out?

Joint ventures may, of course, be the preferredchoice of the foreign investor if, for example, localfirms have specific assets that can be accessed onlythrough collaboration. However, binding owner-ship restrictions may adversely affect firm perform-ance, because the incentive to undertake costlytransfers of technology and improvements in man-agement is related to the expected gains, which inturn are related to an owner’s share of the profitsof. Moreover, changes in the permissible share ofownership in the vicinity of 50 percent (e.g., from49 percent to 51 percent) may have a particularlylarge impact on performance as a firm obtainsfuller control of its operations and has greater free-dom to make the changes it deems necessary.

At this stage the evidence available is not suffi-cient to evaluate the magnitude of these effects, butthe plausibility of their existence raises the questionof why the Chinese government is willing to bearsuch costs. Four reasons are possible. First, limi-tations on ownership may seek to balance theefficiency enhancing and the rent-appropriatingaspects of foreign investment. However, rent appro-priation could to some extent be prevented by exante auctions of equity or ex post taxation of prof-its.9 A more basic question is why rent-generatingrestrictions on competition continue to exist. Sec-ond is the “infant entrepreneur” argument: foreign-ers are induced to form EJVs so that local investors

can learn by collaborating. As with all such argu-ments, it is difficult to judge whether the currentcosts are likely to be offset by the eventual benefits.Third is the adjustment cost argument: an immedi-ate transfer of control could lead to drastic cuts insurplus labor, which gradual reductions in owner-ship help prevent. A central issue is whether it ispossible to address these adjustment costs throughdirect support to the affected factors rather than bystaggered liberalization. Finally, the most impor-tant reason is probably a purely political reluctanceto allow foreign control of an essential service.Again, these political concerns should diminish ifmultiple competing foreign firms, not just one for-eign monopolist, are providing the service.

Eliminating Restrictions Gradually Across Regions

A remarkable feature of China’s dramatic expan-sion in international trade over the past twodecades has been the concentration of export-oriented industries in coastal regions. Three coastalprovinces (Guangdong, Jiangsu, and Fujian) andShanghai have been the main recipients ofoutward-oriented foreign investment, with theremaining portion going either to other coastalprovinces or to regions adjoining coastal areas.Thus, although China’s economic reforms havebeen successful in raising living standards for aconsiderable share of the population, a large num-ber of Chinese people in inland provinces havebenefited substantially less.

A factor responsible for coastal agglomerationhas been the inefficiencies in China’s internal serv-ice systems, most obviously transport and telecom-munications, but also financial, business, and otherservices. Transport infrastructure disparitiesbetween the coastal and inland provinces narrowedconsiderably after policies were adopted in 1990aimed at promoting more regionally balanced eco-nomic development. However, there is evidence tosuggest that it is not the availability of transportinfrastructure per se that has precluded inlandprovinces from actively participating in foreigntrade. Rather, the inadequacies associated withtransport services have been the more serious bind-ing constraints on better integrating China’s hinter-land economy.10

The geographic limitations on liberalizationcommitments in certain sectors could well


accentuate these interregional inequalities, eventhough the limitations do not apply to transportservices and most are to be phased out over time.The existing enclaves of development are likely tosee faster improvements in service quality fromearly liberalization, and these improvements willcause even more economic activity to gravitate tothese areas. Eventually, though, liberalization ofservices in the hinterland may lead to greater diffu-sion of activity. Yet insofar as sunk costs are impor-tant, the earlier agglomeration is unlikely to becompletely reversible and inequalities may persist.The importance of these considerations needs to beestablished through further empirical research, butat least intuitively would seem to strengthen thecase for eliminating the geographic restrictionssimultaneously rather than sequentially.

What Are the Priorities forComplementary Domestic Reform?

If China is to make the most of the liberalization itis now committed to undertake, improving the reg-ulatory framework is critical. Regulation in services,as in goods, arises essentially from the need to rem-edy market failure—attributable to the problems ofnatural monopoly and inadequate consumerinformation—and to ensure equitable access.

Efficient Regulation: Making Competition Work

The first regulatory priority arises in the so-calledlocational services. These markets are concentratedbecause of the large fixed costs required to createspecialized distribution networks: roads and railsfor land transport, cables and satellites for commu-nications, and pipes for sewage and energy distri-bution (UNCTAD and World Bank 1994). Unlessthe appropriate regulatory mechanisms are put inplace, the incumbent can frustrate competition bydenying rivals access to essential facilities such asdistribution networks and terminals.

China has accepted the regulatory principlesspecified in the Telecommunications ReferencePaper. Thus it has committed to instituting anindependent regulator for basic telecommunica-tions services to ensure that the incumbent supplierdoes not undermine market access by charging pro-hibitive rates for interconnection to its establishednetworks.11 Since 1998 the Ministry of Information

Industry (MII) has been the regulator of telecom-munications and other information technology-related products and services, including datacommunications, wireless communications, elec-tronics, computers, the Internet, and software.In August 2001 the high-profile State CouncilInformation Office (SCIO) was formed with a man-date to develop and implement information policyin China (see chapter 10). Similar steps may needto be taken in a variety of other network services,including transport (terminals and infrastructure)and energy services (distribution networks).

The creation of a regulator is only a first step.Persuading the dominant interest groups to con-cede control is fraught with difficulty. For example,in India a conflict between the Department ofTelecommunications (DOT) and the Telecommu-nications Regulatory Authority of India (TRAI), asit was initially constituted, hampered progresstoward an efficient telecom infrastructure.12 Absenta truly independent regulator empowered to rebal-ance tariffs, enforce fair interconnection agreements,and ensure rapid, equitable issuance of radio spec-trum, the benefits of a sector opened to allow pri-vate participation and foreign investment could besignificantly limited.

In certain market segments it may not be possi-ble to create conditions for effective competition inthe supply of certain telecommunications, trans-port, and financial services, even if all barriers toentry are eliminated. Two related factors are likely tobe involved. First, unlike for goods, national marketsare often segmented from the international marketby the infeasibility of cross-border delivery. Second,changing technologies may have reduced the opti-mal scale of operation as well as the sunk costs inthese sectors, but not enough for small markets tosustain competitive market structures. Some formof final price regulation may therefore be needed.

Regulation to Remedy Inadequate ConsumerInformation

In China’s increasingly open markets, priority mustbe given to strengthening the quality of prudentialregulation in intermediation and knowledge-basedservices, where consumers have difficulty securingfull information about the quality of service theyare buying (UNCTAD and World Bank 1994).For example, consumers cannot easily assess the


competence of professionals such as doctors andlawyers, the safety of transport services, or thesoundness of banks and insurance companies.When such information is costly to obtain and dis-seminate and consumers have similar preferencesabout the relevant attributes of the service supplier,the regulation of entry and operations in a sectorcould increase social welfare. However, the estab-lishment of institutions competent to regulate wellis a serious challenge, as revealed by the difficultiesin the financial sectors elsewhere—not only in var-ious developing countries but also in the UnitedStates, Sweden, and Finland in the 1980s and 1990s.

A separate problem is that domestic regulationsto deal with market failure may themselves becomeimpediments to competition and trade as a resultof differences across jurisdictions in technical stan-dards, prudential regulations, and qualificationrequirements in professional, financial, and numer-ous other services. For example, China requires for-eign doctors to obtain a license from the Ministryof Public Health and foreign accountants to passthe Chinese national certified public accountantexamination. In many instances the impact ontrade is an incidental consequence of the pursuit ofa legitimate objective, but in some instances reg-ulation can be a particularly attractive means ofprotecting domestic suppliers from foreign compe-tition. Multilateral trade rules on domestic regula-tions may contribute to domestic reform by help-ing to sift the legitimate from the protectionist. Tothis end, negotiations are under way to developGATS rules for domestic regulations. The core ofthese disciplines may well be the so-called necessitytest, which seeks to establish whether a particularregulation is more burdensome than necessary toachieve a legitimate objective.

Regulation to Ensure Universal Service

Opening up essential services to foreign or domesticcompetition could have an adverse effect on thepoor—a point that is often cited as a reason for thepersistence of public monopolies. A more efficientsolution is to have regulations with a social purpose.

Where China is a relatively inefficient producerof a service, liberalization and the resulting foreigncompetition are likely to lead to a decline in domes-tic prices and improvements in quality. However,there is a twist. Frequently, the prices before liberal-

ization were not determined by the market, butwere set administratively and then kept artificiallylow for certain categories of end users or types ofservice products. Liberalization threatens thesearrangements. The elimination of restrictions onentry implies an end to cross-subsidization, becauseit is no longer possible for firms to make extranor-mal profits in certain market segments. And privati-zation could mean the end of government support.

Reform programs can accommodate universalservice obligations by imposing this requirementon new entrants in a nondiscriminatory way. Thusin several countries such obligations have been partof the license conditions for new entrants into fixednetwork telephony and transport. Often, however,subsidies have proved more successful than directregulation in ensuring universal access (Estache,Foster, and Wodon 2001). The Chilean governmentadopted a scheme that permitted it to leverage over$2 million in public funds into $40 million in privateinvestment. As a result, telephones were installed in1,000 localities at about 10 percent of the cost ofdirect public provision. Public subsidies can also bedirected to the consumer rather than the provider(Cowhey and Klimenko 1999). The choice ofappropriate instrument will have to be made on asector-by-sector basis.

Addressing Adjustment Costs

Different modes of supply have different effects onfactor markets. Cross-border trade and consump-tion abroad resemble the goods trade in theirimplications. The impact of the movement of fac-tors depends critically on whether they are substi-tutes or complements for domestic factor services.Given the structure of factor prices in China, liber-alization would typically lead to an inflow of capitaland skilled workers. Such inflows would tend to beto the advantage of the unskilled poor—increasingemployment opportunities and wages.13 Interest-ingly, it has been shown that even when foreignerscompete with local skilled workers in a service sec-tor, the productivity boost to the sector from allow-ing foreigners access could lead to an increase in thedemand for domestic skilled workers—that is, thescale effect could outweigh the substitution effect(Markusen, Rutherford, and Tarr 2000). Giventhese predictions, why are workers in China some-times skeptical about the benefits of liberalization?


One concern is the possible reduction in employ-ment in the formerly public monopolies that havefrequently employed surplus labor.

Evidence suggests that such pessimism maynot always be justified. For example, some develop-ing countries have managed to maintain or evenincrease employment in their liberalized telecom-munications sectors. Because China still has a lowteledensity outside the big cities (about 10 lines per100 people), much of telecom investment is beingdirected toward building wire line and mobile net-works that are labor-intensive and that are probablyhelping to maintain or raise employment levels. Theintroduction of competition should help. For exam-ple, Petrazzini and Lovelock (1996) found in a studyof 26 Latin American and Asian economies that tele-com markets with competition were the only onesthat consistently increased employment levels, whiletwo-thirds of the countries with monopolies sawconsiderable declines in their telecom work force.Despite these optimistic projections, it is likely thatreform programs will require complementary poli-cies to mitigate any social and economic costs ofadjustments in factor markets.

Sequencing Regulatory Reform and Trade andInvestment Liberalization

Regulatory improvements take time. Changes in thepatterns of competition and ownership can beimplemented instantaneously in principle, butChina, like many other countries, has chosen tointroduce changes gradually. One question thatarises in this context is: How does the impact of dif-ferent elements of reform depend on the extent towhich other reforms have been implemented—thatis, how does the interaction of the different elementsof policy reform affect performance? The other,more subtle question is: Does the sequence of reformshave transitory and permanent effects on perform-ance? Some preliminary observations on these ques-tions, which are relevant to the design of transitionstrategies, are presented in the sections that follow.

Telecommunications Figure 8.2 depicts thesequence of telecommunications liberalizationin nine Asian countries. China is among the few thathave allowed some degree of competition (in long-distance services) before allowing a change of owner-ship in the incumbent supplier and creating an inde-pendent regulator. Fink, Mattoo, and Rathindran

(2001), in a study of telecommunications reform inAsia, Africa, and Latin America, found that althougheach element of reform has a positive impact on per-formance, the effect of each is magnified when othersare also implemented. Moreover, privatizing theincumbent after introducing competition (theChinese route) is likely to lead to a higher level ofmainline penetration than the opposite sequence.

Although China has chosen the preferredsequence in telecommunications, it must still makeimportant choices in other areas, such as postal andair transport services. Introducing privatizationbefore introducing competition may create a privi-leged incumbent that has a first-mover advantagein the market. Among other things, the incumbentmay be able to make certain strategic decisions orindulge in lobbying to affect the eventual form ofcompetition.14 For example, in South Africa foreignand domestic shareholders in the privatizedtelecommunications monopoly managed to per-suade the government to allow only one newentrant rather than the planned two. Regulating theterms of access to essential facilities for newentrants may also be more difficult in an arm’slength relationship with a private provider whosecosts are difficult to observe than with a publicprovider whose cost information may be easierto access. Entry may be easier on symmetricterms with an inefficient public incumbent thanon asymmetric terms with an efficient privateincumbent.

Financial Services In financial services, interna-tionalization raises several concerns: the threat tothe survival of local banks and financial companies;the loss of monetary autonomy; and the increasedvolatility of capital flows. Many of these concernsdo not relate just to the internationalization offinancial services, but also to the processes of finan-cial deregulation and capital account liberalization.However, the extent of the benefits and costs ofinternationalization depend to a great extent onhow it is phased in with these other two types offinancial reform, and, in particular, the strengthen-ing of prudential regulation and supervision.

Many countries that have had successful experi-ences opening up to foreign financial firms (Brazil,Chile, Hungary, Ireland, Poland, Portugal, andSpain, among others) also engaged in a process ofdomestic deregulation and consequently reapedsubstantial gains (World Bank 2001). The experience


of the countries acceding to the European Unionsuggests that internationalization and domesticderegulation can be mutually reinforcing.Increased foreign entry bolstered the financialsector framework by creating a constituency forimproved regulation and supervision, better disclo-sure rules, and improvements in the legal and

regulatory framework for the provision of financialservices. It also added to the credibility of rules.

Most of these considerations are relevant toChina. As discussed by Bhattasali in chapter 11,on-performing loans account for a large fraction ofthe assets of the big four banks, which account formore than two-thirds of the assets of the banking


Sri Lanka

1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

Singapore

Philippines

Pakistan

Malaysia

Korea,Republic of

Indonesia

India

China

LDILD

ILD

ILD

Local

Local

Local, LD

Local, LD

Local, LD

Local

LD

LDILD

ILD

ILD

Local

Local

Local, LD

Local, LD

Local, LD

2 3

1

2

1

1 2 3 4

2

4

2 5

7 8

2 5

Local201919148

10%

12%

11% 17%

34%

20% 29%

25%

100%

7431 6

19% 23%

LD21

Privatization Mobile

RegulationFixed competition

FIGURE 8.2 Sequence of Telecommunications Reform in Nine Asian Countries, 1989–99

Notes: The percentages indicate the share of private equity ownership in the incumbent operator. Local, LD, andILD refer to the local, long distance, and international fixed-line service segments, respectively. The number(s) inthe mobile row corresponds to the number of cellular operators in the country. “Regulation” only captures theexistence of a separate regulatory agency.Source: World Bank/ITU telecommunications policy database.

sector. The big four are already thought to be insol-vent (Fitch and Moody’s). One of the main reasonsfor the many bad loans is that interest rates are stillcontrolled by the People’s Bank of China (PBC).They are unusually low to make it easier for state-owned enterprises to borrow funds at well belowthe market-clearing equilibrium rate of interest.15

This situation has been responsible for a hugetransfer of wealth from individual savers to state-owned enterprises. Two-thirds of credit resourceswent to state-owned enterprises that generate onlyone-third of industrial output. The official target of2003 for interest rate liberalization has already beenpushed back to 2005 unofficially for fear that theensuing competition for deposits will drive busi-ness away from the four large banks, which wouldthen collapse, leading to a serious banking crisis(Business China, May 7, 2001).

Although the two reform processes (interna-tionalization and domestic financial deregulation)are mutually reinforcing, they are not sufficient inthemselves. More than in other sectors, the gainsand costs of financial reform depend on the regula-tory and supervisory framework (Barth, Caprio,and Levine 2001). Experience shows that it is vitalto strengthen the supporting institutional frame-work in parallel with domestic deregulation andinternationalization. In the absence of suchstrengthening, foreign entry may entail risks. For-eign bank entry can destabilize local banks by tak-ing away the lowest-risk business—including largeexporting firms—leaving local banks to venturefurther out on the risk frontier.

The benefits of a supportive institutional frame-work are even more obvious when it comes to cap-ital account liberalization.16 Experiences in recentyears, most recently in Asia, have shown thatachieving the potential gains, and avoiding therisks, of capital account liberalization depend to agreat extent on whether domestic institutions andprudential authorities have developed sufficientlyto ensure that foreign finance will be channeled inproductive directions (Eichengreen 2001).

Conclusions

China’s GATS commitments represent the mostradical services reform program negotiated in theWTO. China has promised to eliminate over thenext few years most restrictions on foreign entryand ownership, as well as most forms of discrimi-nation against foreign firms. Trading partners arenaturally interested in the market access dimensionof these commitments and are keen to see themfully implemented.17 However, realizing the gainsfrom, and perhaps even the sustainability of, liber-alization will require the implementation of com-plementary regulatory reform and the appropriatesequencing of reforms.

Although China’s commitments, for the mostpart, promote good policy, two elements raise someconcern. First, restrictions on foreign ownership—temporary in most sectors, but more durable intelecommunications and life insurance—maydampen the incentives for foreign investors toimprove firm performance. The rationale for theserestrictions merits greater scrutiny. Second, initialrestrictions on the geographic scope of servicesliberalization could encourage the further agglom-eration of economic activity in certain regions—indeed, to an extent that is unlikely to be reversedcompletely by subsequent country-wide liberaliza-tion. It may therefore be worth examining whetherthese restrictions could be phased out more quickly.

This chapter has identified three main prioritiesfor complementary reform. First, improved pru-dential regulation and measures to deal with thelarge burden of nonperforming loans by statebanks are necessary to realize the benefits of liber-alization in financial services. Second, in basictelecommunications and other network-basedservices meaningful liberalization will be difficultto achieve without strengthened pro-competitiveregulation. Finally, in virtually every importantservice sector, an effective universal service policymust be put in place because liberalization alonewill not deliver improved access for the poor.



TABLE 8.A.1

Mode 1 Mode 2 Mode 3 Mode 4

Sector F P N F P N F P N F P N

Business servicesLegal services 0 1 0 0 1 0 0 1 0 0 1 0Accounting/auditing/ 1 0 0 1 0 0 0 1 0 0 1 0

bookkeepingArchitectural services 0 1 0 0 1 0 0 1 0 0 1 0Medical and dental services 1 0 0 1 0 0 0 1 0 0 1 0Data processing services 0 1 0 0 1 0 0 1 0 0 1 0RandD services (natural 0 1 0 0 1 0 0 1 0 0 1 0

sciences)Advertising services 1 0 0 1 0 0 0 1 0 0 1 0Management consulting 0 1 0 0 1 0 0 1 0 0 1 0

servicesCommunication services

Courier services 0 1 0 0 1 0 0 1 0 0 1 0Voice telephone services 0 1 0 0 1 0 0 1 0 0 1 0Private leased circuit services 0 1 0 0 1 0 0 1 0 0 1 0Electronic mail 0 1 0 1 0 0 0 1 0 0 1 0Online info and database 0 1 0 1 0 0 0 1 0 0 1 0

retrievalAudiovisual services 0 1 0 0 1 0 0 1 0 0 1 0

Construction, engineeringConstruction work (building) 0 0 1 0 1 0 0 1 0 0 1 0Construction work (civil engin.) 0 0 1 0 1 0 0 1 0 0 1 0

DistributionWholesale trade 0 1 0 0 1 0 0 1 0 0 1 0Retailing services 0 1 0 0 1 0 0 1 0 0 1 0

Educational servicesSecondary education 0 0 1 0 1 0 0 1 0 0 1 0Adult education 0 0 1 0 1 0 0 1 0 0 1 0

Environmental servicesSewage services 0 1 0 0 1 0 0 1 0 0 1 0Refuse disposal 0 1 0 0 1 0 0 1 0 0 1 0

Financial servicesNon–life insurance 0 1 0 0 1 0 0 1 0 0 1 0Acceptance of deposits 0 1 0 0 1 0 0 1 0 0 1 0Lending of all types 0 1 0 0 1 0 0 1 0 0 1 0Trading in securities 0 1 0 1 0 0 0 1 0 0 1 0

Health-related, social servicesHospital servicesSocial services

Tourism servicesHotels and restaurants 1 0 0 1 0 0 0 1 0 0 1 0Travel agencies 1 0 0 1 0 0 0 1 0 0 1 0

Recreational servicesEntertainment servicesNews agency services

Transport servicesMaritime (freight) 0 1 0 0 1 0 0 1 0 0 1 0Rail (passenger)Rail (freight) 1 0 0 1 0 0 0 1 0 0 1 0Road (passenger)Road (freight) 1 0 0 1 0 0 0 1 0 0 1 0

Total number 7 20 4 10 21 0 0 31 0 0 31 0(31 subsectors)

Percent 23 64 13 32 68 0 0 100 0 0 100 0

F � full commitment; P � partial commitment; N � no commitment.Sources: WTO services database.

Appendix: Structure of China’s Market Access Commitments

Notes

1. See, for example, chapter 9 by Luo and Findlay, chapter 10by Pangestu and Mrongowius, and chapter 11 by Bhattasali.

2. The only explicit sectoral exclusion from the GATS is cer-tain “hard” rights in the aviation sector.

3. Note that a GATS scheduling rule requires that discrimi-natory quantitative restrictions be listed as market access restric-tions even though they are also inconsistent with nationaltreatment.

4. The commitment on mode 1 indicates a cross reference tomode 3, the meaning of which is not clear.

5. It was reported in the Financial Times of December 6,2001, that a compromise had been reached on this issue. Appar-ently, AIG would be given permission to open two more 100 percent–owned branches, but would thereafter have to abideby the same 50 percent rule as all other foreign insurers inthe Chinese market. The deal was seen as consistent with WTOrules because the two extra licenses were to be awarded to AIGbefore China entered the WTO on December 11, 2001. Thelicenses were to be granted for the cities of Suzhou, nearShanghai, and Beijing, the capital.

6. Where this happens will depend on the importance ofsunk costs relative to differences in costs and quality.

7. One such possibility is in the “nonsustainability” of a nat-ural monopoly. This situation could arise, for example, undersome natural monopoly cost conditions, when there exist noprices that will not attract entry, even though supply by a singlefirm is efficient. Armstrong, Cowan, and Vickers (1994, p. 106)conclude: “Notwithstanding the logical possibility of this hap-pening, we are doubtful whether it provides a good case forentry restrictions in the utility industries, which are not for themost part remotely contestable and where there is little evidencethat cost conditions give rise to non-sustainability.”

8. Interesting evidence in this context is available from theIndian telecommunications sector. Das (2000) estimates afrontier multiproduct cost function for the incumbent fixed-line operator, covering 25 years from 1969 to 1994. The studyfinds the existence of very high economies of both scale andscope in the technology used—the parameter estimates evensuggest that telecommunications in India is a natural monop-oly. However, the incumbent operator displays great ineffi-ciency, leading to a 26 percent increase in the operator’s cost ofproduction. Based on these findings, Das concludes that India’smarket liberalization program, started in the mid-1990s, is jus-tified, but he argues that there may be a need to regulate entryin order to reduce the unnecessary duplication of commoncosts. Moreover, with continued improvements in technology,the fixed costs of entrants are likely to fall, reducing lossesof scale economies and thus increasing the costs of entryrestrictions.

9. The fear of creating a disincentive for investors might be areason to refrain from taxation.

10. For example, although there has been a significantincrease in the volume of container traffic in China since 1990,the increase is largely confined to coastal regions and associatedwith the ocean-going leg of travel. Container traffic in inlandareas is much less, with no significant change in the percentageof seaborne containers traveling beyond port cities and coastalprovinces. Truck rates for moving a container 500 kilometersinland are estimated to be about three times more, and the tripduration five times longer, than they would be in Europe orUnited States. The intermodal transport system was found to bepoorly integrated, with no streamlined procedures to supportthe continual movement of containers between the coast andinlands.

11. Several countries have found it difficult to create open,competitive telecommunications sectors because of a weak regu-latory environment. Poland opened up its telecommunicationssector to private competition as early as 1990. There was a rushto invest, and about 200 licenses were awarded in the first sixyears of the newly liberalized regime. However, the dominantstate operator, operating in a weak regulatory system, limitedaccess to its network and benefited from unequal terms for rev-enue sharing. By 1996 only 12 of the 200 licenses were still beingused by the few competitive operators to survive.

12. The Indian government announced a new telecommuni-cations policy on March 26, 1999, that addressed several of thesekey outstanding issues.

13. Because the poor are likely to be unskilled, the questionarises as to which service sectors are likely to employ them? Dataon the skill composition of the work force in service sectors areavailable only at a rather aggregate level. Still, a certain patterncan be inferred. Construction, distribution, and personalservices tend to be unskilled labor–intensive, whereas commu-nications, financial, and business services tend to be skilledlabor–intensive.

14. A public sector provider could also behave in this way,but presumably if the government’s objective is liberalization, itis somewhat easier to draw along a public provider.

15. In 1995 savings deposits received a paltry 3.15 percent ininterest, while loans for working capital were made at only11 percent—they should have been made at about 21 percentafter factoring in the rate of inflation. The current one-year rateon savings deposits is about 2.25 percent.

16. Because China has made only limited commitments oncross-border trade in financial services, its GATS commitmentsdo not require it to allow a high degree of capital mobility,except insofar as capital inflows are required to establish com-mercial presence.

17. In fact, in the WTO’s Services Council the United Stateshas already raised the issue of whether China’s current rules forexpress delivery services and branching by non–life insurancecompanies conform to its GATS commitments (Inside US Trade,March 29, 2002).

References


Armstrong, M., S. Cowan, and J. Vickers. 1994. RegulatoryReform: Economic Analysis and British Experience. Cambridge,Mass.: MIT Press.

Barth, J. R., G. Caprio Jr., and R. Levine. 2001.“Bank Regulationsand Supervision: What Works Best?” Paper presented atThirteenth Annual World Bank Conference on DevelopmentEconomics, April.

Cowhey, P., and M. M. Klimenko. 1999. “The WTO Agreementand Telecommunication Policy Reforms.” Draft report forthe World Bank. Graduate School of International Relationsand Pacific Studies, University of California at San Diego,March.

Das, N. 2000. “Technology, Efficiency and Sustainability ofCompetition in the Indian Telecommunications Sector.”Information Economics and Policy 12: 133–54.

Eichengreen, B. 2001. “Capital Account Liberalization: What Dothe Cross-Country Studies Tell Us?” World Bank EconomicReview 15 (3): 341–65.

Estache, A., Vivian Foster, and Quentin Wodon. 2001. “Account-ing for Poverty in Infrastructure Reform: Learning from



Latin America’s Experience.” WBI Development Studies No.23950. World Bank Institute, Washington, D.C.

Fink, Carsten, Aaditya Mattoo, and Randeep Rathindran. 2001.“Liberalizing Basic Telecommunications: The Asian Experi-ence.” Policy Research Working Paper 2718. World Bank,Washington, D.C., November.

Ianchovichina, E., W. Martin, and C. Wood. 2001. “EconomicEffects of the Vietnam–US Bilateral Trade Agreement.”World Bank, Washington, D.C. Processed.

Markusen, James, Thomas F. Rutherford, and David Tarr. 2000.“Foreign Direct Investment in Services and the DomesticMarket for Expertise.” Policy Research Working Paper 2413.World Bank, Washington, D.C., August.

Petrazzini, B. A., and P. Lovelock. 1996. “Telecommunications inthe Region: Comparative Case Studies.” Paper presented at

International Institute for Communication Telecommunica-tions Forum, April 22–23. Sydney, Australia.

Smith, P. 1995. “End of the Line for the Local Loop Monopoly.”Public Policy for the Private Sector, Note No. 63. WorldBank, Washington, D.C.

United Nations Conference on Trade and Development(UNCTAD) and World Bank. 1994. Liberalizing Interna-tional Transactions in Services: A Handbook. New York andGeneva: United Nations.

World Bank. 2001. “Finance for Growth: Policy Choices in aVolatile World.” World Bank Policy Research Report. WorldBank and Oxford University Press.

WTO (World Trade Organization). 1999. “Structure of Com-mitments for Modes 1, 2 and 3.” Background Note by theSecretariat, Council for Trade in Services, S/C/W/99, March.

evolve, and recently it became a “hot” topic inChina, with logistics services eliciting stronginterest from both businesses and policymakers.Moreover, a significant part of seminar and man-agement training time in China is now devoted todeveloping what might be called the “logistics wayof thinking.” There is an appreciation that systems,processes, and perspectives—that is, the software ofthe business—are just as important as the physicalassets, if not more so, in the growth of enterprisesthat are internationally competitive in logistics.

We argue in this chapter that WTO accession hassignificant implications for the logistics sector.Although there is no easy match between the vari-ous categories of the General Agreement on Tradein Services (GATS) and the modern scope of logis-tics, it is possible to make a concordance. We com-plete that work in order to comment on theimplications of China’s accession commitments.The requirement to make these associations illus-trates an important feature of scheduling commit-ments under GATS in emerging business areas.

The next section of this chapter reviews the def-inition of logistics services. It is followed by anoverview of their evolution in China.1

Study of the geography of poverty has revealed howchanges in transport costs, even when they make upa small share of the value of a product, can lead tolarge changes in value added and in growth rates(see Radelet and Sachs 1998 and Redding andVenables 2000). The rate at which transport costsaccumulate over distance is usually taken as givenin this research, and the questions asked are aboutthe links between location and rural producerincomes. Policy prescriptions then usually refer toprocesses under direct government control such asport operations or customs procedures. However,in this chapter we stress that policy reform, espe-cially through commitments to the World TradeOrganization (WTO), can have significant effectsnot just on transport prices, but also on the othercosts of getting goods to market in the form thatconsumers prefer. These effects are important bothin international trade and in trade within an econ-omy. Reductions in this set of costs can add to valueadded, especially in remote, poor areas.

We illustrate these points in this study of thelogistics sector in China and of the impact ofChina’s accession to the WTO on the performanceof that sector. The concept of logistics continues to

141

9

Logistics in China:Implications of

Accession to the WTO

Wenping Luo and Christopher Findlay

The authors thank the editor of this volume and seminar participants for comments. Errors remain their own.

We then outline the WTO commitments madeby China related to this sector, and reach a set ofconclusions on their direct effects and their indirectimplications. The direct effects, we emphasize,include substantial savings in real resources and notjust transfer differences that represent relativelysmall gains in efficiency.

The indirect effects include reference to policy-making processes and administrative structures.The analysis here highlights the value of comple-mentary reforms in domestic institutions and regu-latory agencies and their processes in order tomaximize the gains from liberalization.

The chapter then continues with a discussion ofthe likely orders of magnitude of the effects of amore open logistics sector on total logistics costs,and discusses the significance of these changes forpoor areas in China. We pay special attention to therole of the gains in increasing the value added in pro-duction processes in poor areas. Some of the mainpoints are revisited in the final section of the chapter.

Logistics Defined

Logistics is the process of planning, implementing,and controlling the efficient flow and storage ofgoods, services, and related information from thepoint of origin to the point of consumption to meetcustomers’ requirements. The provision of logis-tics services requires inputs from various serviceproviders, including the providers of transport andwarehousing as well as other value-adding activities.

The concept of logistics has undergone manysignificant changes. Coyle, Bardi, and Langley(1996) divided the development of logistics intothree stages:

• Stage 1: physical distribution or outbound logis-tics system (during 1960s and 1970s). Businessesattempted to manage systematically a set ofinterrelated activities, including transportation,distribution, warehousing, finished goods,inventory levels, packaging, and materials han-dling, to ensure the efficient delivery of finishedgoods to customers.

• Stage 2: integrated logistics management (dur-ing the 1970s and 1980s). Companies began torecognize the additional opportunities for sav-ings by combining the inbound side (materialsmanagement) with the outbound side (physicaldistribution). Initially, savings was achieved by

having a single transportation manager whocould coordinate inbound and outbound trans-portation. Then companies became aware of theopportunities to view the whole process, fromraw materials to work-in-process inventory tofinished goods, as a continuum that, managedfrom a systems perspective, could lead to moreefficient operation.

• Stage 3: supply chain management (from 1980sto 1990s). Companies expanded their perspectiveon the logistics processes to include all the firmsinvolved, making use of partnerships/alliancesbetween manufacturing companies and theirsuppliers/vendors, customers (channels of dis-tribution), and other logistics-related partiessuch as transportation and public warehousingcompanies.

The concentration on fewer carriers operatingunder long-term contracts and on outsourcing thewhole package of services stimulated the develop-ment of third-party logistics (3PL) firms. As noted,several subsectors are involved in logistics—thetransport sector (including different modes oftransportation), the warehousing sector, andrelated providers of value-added services. The 3PLbusiness is complex, because it operates across allthe subsectors involved.

Luo and Findlay (2001) reviewed developmentsin each subsector in China, including the basic factsof capacity, volume, and growth; evolution andreform of the subsectors; subsector infrastructureand organization; and administrative, regulatory,and policy issues. Some of the results relevant tothis chapter are outlined in the next section.

The Logistics Sector in China

Growth of the logistics sector in China has beendramatic. Freight turnover in China increased from76 billion ton-kilometers in 1952 to 4,381 billionton-kilometers in 2000. The volume increased by3.8 times between 1980 and 2000.

Evolution of Transport Services

In the political and economic environment of theearly 1950s, China gave priority to heavy industries.A precondition for their development was a supplyof cheap raw materials. This development in turndemanded a supply of cheap transport, and rail and


water were two modes of transportation that couldprovide transport services for large volumes at arelatively low price.

Figure 9.1 shows the composition of freightcarried by mode (measured in weight) from 1970to 2000. After 1978 the transport system in Chinawas characterized by the rapid development of roadand air transportation. The market shares of waterand rail transport decreased sharply, reflecting theshift from the traditional heavy industries to lightindustries that demanded speedy, flexible trans-portation services and that could bear a highertransport service price. Figure 9.1 shows the risingshare of road transport and the fall in the shares ofrail and water in the total tonnage. Although it isnot evident, data for air transport were included inthe construction of the chart. Those numbers arevery small compared with those for other modes,and therefore are not evident in the columns of thechart, but air freight volumes are growing fastestamong all modes. As for trip lengths, they varysignificantly by mode. The length of road trips isrelatively short on average. Rail and water showlonger trip lengths and therefore account for muchhigher shares of the transport task measured interms of ton-kilometers. But even in these terms,the road share continues to grow, mainly at theexpense of rail.

Constraints on Development

China’s program of economic reform has graduallychanged the ownership of its enterprises, especially

in the manufacturing sector. But in the transporta-tion sector the change has been relatively small;large, state-owned enterprises still dominate thesector. In recent years, the use of joint ventures hasincreased, especially in the container berth, con-tainer freight station, and motor carrier activities.However, these firms still account for only a smallportion of the total number of enterprises. Roadtransport and shipping were deregulated in 1986,and rail and air transportation were also decentral-ized to some extent. The ownership structure haschanged to different degrees in different modes oftransportation. The road and inland shipping com-panies are dominated by collective and privateenterprises, and coastal and international shipping,rail, and air transportation enterprises are stillmostly state-owned.

The transport sector is still a weak link in China’seconomy:

• The number of either rail or road trunk lineslinking regions is insufficient. The major railwaylines are overloaded. The trunk lines of the high-way system linking provinces are not well net-worked and cannot meet the demand. Theseimpediments also contribute to differences ingrowth rates between regions.

• The coordination between different modesrequires planning and development. Withineach mode of transport, there are imbalances. Inthe rail sector, the mixing together of passengertrains with cargo trains on the busy lines leadsto slower speeds and low efficiency. In the air

Logistics in China: Implications of Accession to the WTO 143

FIGURE 9.1 Choice of Transport Modes: China, 1970–2000

0

20

40

60

80

100Percent

1970 1980 1990 2000

Freight traffic by mode (metric tons)

Rail Water AirRoad

Source: Luo and Findlay 2001.

sector, there is an imbalance between the trunkand feeder lines and in the number of big andsmall aircraft. In the road sector, the construc-tion of the national-level trunk road system hasbeen accelerated, but in the rural areas the roadnetwork is far less developed.

• Service quality is generally low, as is the techni-cal level of the equipment. For example, in therail sector double lines and electrified lines rep-resent a relatively small share of the total lengthof lines. Less than 20 percent of the freighttrucks can carry containers.

The warehousing and storage system in theplanned economy before reform was character-ized by department or sector administration andoperations—that is, sectors built and owned theirown warehousing and storage facilities and servedtheir own demands only. Many sectors formed theirown closed warehousing systems.2 Meanwhile, eachlocal government at the provincial, city, and countylevels established their own warehousing system toserve their local economy. Overall, therefore, govern-ment administration and enterprise functions arenot separated, which constrains enterprise perform-ance; the administration is fragmented in both verti-cal and horizontal dimensions, which inhibits theintegration of services; and in some activities thequality of the physical infrastructure is insufficient.

Demand for 3PL Services

The demand for specialized logistics or 3PL serv-ices has begun to grow faster in recent years,especially that originating from the multinationalcompanies, the leading Chinese state-owned com-panies, and some of the private companies, as wellas that arising from the e-commerce sector. In therecent years, quite a few 3PL companies haveappeared, including (a) international companiessuch as Maersk, UPS, and TNT; (b) traditionalChinese transport, warehousing, and forwardingcompanies such as COSCO and Sinotrans; and(c) emerging companies such as the Bao GongLogistics Group.

Relevant departments in the central governmenthave paid greater attention from different perspec-tives to the logistics sector and its development, andthey have studied policies that they expect willpromote it. Local governments such as Shenzhen,

Beijing, Tianjing, Shanghai, Guandong, andShandong have also paid attention to the develop-ment of their local logistics sectors.

Most of logistics activities are still operated in-house, and demand for 3PL services is low amongtraditional Chinese firms. State-owned companiesin China have inherited an operating model basedon the provision of services in-house, includinglogistics. According to investigations by the ChinaWarehousing and Logistics Council, a third partyundertakes only 18 percent of the inbound logistics.As for outbound logistics, 59.8 percent is operatedtogether with a 3PL firm, and only 16.1 percent isoperated totally by a 3PL firm. Most Chinese firmskeep their own logistics assets and personnel (Luoand Findlay 2001).

Although many 3PL firms have appeared in themarketplace, most of them offer a low standard ofservice and efficiency. For one thing, most of thefirms provide only simple delivery and warehous-ing services: value-added services such informationservices, inventory management, logistics cost con-trol, or logistics system design are not offered. Foranother, the scale of the firms is usually small, andmost of them come from among traditional firmswith weak network and organizational capabilities.

Moreover, modern logistics facilities such astransshipment facilities for different modes oftransport, public warehousing and transport facili-ties, and logistics centers are not available. Atlow levels are the technology used for logisticsequipment and the extent to which facilities arestandardized and information technology isapplied. Finally, logistics facilities are distributedunevenly—for example, transport facilities are lessdeveloped in the middle and western parts of thecountry.

An important constraint on the capacity of ser-vice providers to offer integrated services lies in theadministration system—in particular, in the verti-cal and horizontal separations in the administrativesystems that have carried over from the planningsystem, the self-protection that remains, the mixingof the enterprise with administration functions,and the lack of transparency in policy measures. Allthese issues, and the extent to which they impedethe scope to gain from reform in the logistics sector,are examined in more detail after a review ofthe implications of China’s WTO accession for thesector.


Accession to the WTO and Implicationsfor the Logistics Sector

To provide integrated logistics services, a logisticssector must cover the following service elements:international and domestic transport (all modes);warehousing and storage; international and domes-tic freight forwarding; other value-added servicessuch as logistics information (tracing and track-ing), packaging and labeling, and quality control/inspection; and third-party logistics services.

Various GATS categories must be examined inorder to incorporate the scope of logistics servicesas just defined. Although these categories are listedin the appendix to this chapter, this list is not com-prehensive, and in some cases the impact of logis-tics activities must be inferred from packages ofmore specific commitments—for example, in rela-tion to 3PL services.

Another issue related to the definition of thesector is coverage of the GATS category called “dis-tribution.” Distribution includes some activitiesthat might normally be associated with logistics,such as warehousing and inventory management.3

It appears, therefore, that some parts of a logisticssector, as might be defined in ideal terms, arealready included in the category of distribution.However, distribution refers to these activitieswhen undertaken in the context of wholesaling andretailing. Thus their coverage is much narrowerthan that which is relevant here.

Our assessments of the effects of the likely policychanges associated with WTO accession are dividedinto two types: direct and indirect. Direct effects arethose on the costs and quality of services provided.Indirect effects are those on the regulatory oradministrative environment.

Direct Effects

The direct effects of WTO accession on the logisticssector are presented in the appendix to this chapterand are summarized in tables 9.1–9.3. The tablesshow the starting position (table 9.1) and theimpacts of the commitments one year after acces-sion (table 9.2) and in the period three years afteraccession (table 9.3).

In general, China’s accession to the WTO willhave a positive affect on performance in the logisticsservices sector as a whole. Of the elements listed inthe tables, only the air transport and water transportsectors are isolated from a significant change in pol-icy. The commitments in many areas extend existingpolicy (e.g., customs clearance, container depotservices, maritime agency services, road transport,rail transport, warehousing and forwarding, courierservices). In others, the commitments bind currentpolicy (international shipping). In some cases, geo-graphic limits apply—in internal waterway trans-port, foreign firms can operate only on routes toports open to foreign shipping.


TABLE 9.1 Starting Position of China’s Logistics Sector

Area Starting Position

Air transport Relatively closed in both domestic and international routesWater transport Inland shipping open to domestic companies; coastal and international

shipping relatively closedRoad transport No restriction for domestic entry; foreign entry only by joint venturesRail transport State monopolyForwarding International forwarding: licensing system applies; foreign participation

only via a joint ventureDomestic forwarding: relatively open for domestic companies; not open

to foreign companiesStorage and warehousing No restriction for domestic entry; foreign entry only by joint venturesIntegrated services (3PL) Determined by the aboveOther (packaging) No restriction for domestic entry; foreign entry only by joint venturesCourier service No restriction for domestic entry; foreign entry only by joint ventures

Source: The authors.

Discrimination against foreign suppliers willcontinue (mainly in terms of rules on the forms ofestablishment) for varying periods, which delaysthe benefits of reform. A schedule might bedesigned to capture some of the rents created in thetransition to full openness. It also might bedesigned to slow down the transition and therefore(it is hoped) reduce the costs of adjustment. Theconsequence, however, may only be one of delayingthose costs and adding to the risks of backsliding.Another goal may be to limit for a longer time for-eign ownership in sensitive sectors. However, costswould be incurred, including the imposition on

foreign suppliers of second-best forms of participa-tion in the market (joint ventures instead of whollyowned operations).

As discussed earlier, the demand for integratedservices (3PL) is increasing, and so, more and more,logistics service firms are trying to provide or con-trol the whole chain of the services. Therefore, a3PL firm needs to have access to all the relevantlicenses, which could include those for interna-tional freight forwarding, different modes of trans-portation (air, international shipping, domesticshipping, road, rail), storage and warehousing,container station and depot services, and courier


TABLE 9.2 Position of China’s Logistics Sector One Year after WTO Accession

Evaluation of Direct Impacts

Area High Medium Low

Air transport �

Water transport �

Road transport �

Rail transport �

Forwarding �

Storage and warehousing �

Integrated services (3PL) �

Courier service �

Other (packaging) �


TABLE 9.3 Position of China’s Logistics Sector Three Years after WTO Accession

Evaluation of Direct Impacts

Area High Medium Low

Air transport �

Water transport �

Road transport (3) �

Rail transport (6) �

Forwarding (4) �

Storage and warehousing (3) �

Integrated services (3PL) �

Courier service �

Other (packaging) �

Note: The numbers in parentheses show the number of years afteraccession when wholly owned foreign establishments will be permitted.Source: The authors.

services. The opening up of the different sectors inlogistics, as discussed earlier, would have a signifi-cant impact on the ability of logistics firms to pro-vide the whole chain of services.

Indirect Effects

This section assesses the indirect effects on thelogistics sector of the reforms associated withChina’s accession to the WTO. These effects aresummarized in table 9.4.

Local Protectionism Vertical and horizontaldivisions within administrative systems constraindevelopment of the logistics sector. For example,the administrative system that applies to trans-portation is divided according to the mode oftransport, and the same systems operate at both thelocal and higher levels of government. The result isoverlapping responsibility and power between thedepartmental and district administrations. Theseadministrative structures are no longer suitable forthe development of logistics activities for severalreasons:

• Coordination of the different modes of trans-port services is difficult because the separationof different modes has resulted in differences inorganization, service standards, and technicaland equipment standards. This lack of consis-tency leads to the difficulties in developing mul-timodal services.

• Coordination of the construction of infrastruc-ture is poor, which leads both to duplication ofinvestments and to neglect of the developmentof transshipment centers.

• Self-protection is evident at the departmentaland district levels. Policies and regulations areoften designed according to the self-interest ofthe department, sector, or district involved.A local logistics firm will often seek protectionfrom competition through the application of pol-icy by departmental or district agencies, whichhampers the development of cross-departmentor cross-district logistics networks (see box 9.1for an example of local protectionism).

Reform of the administrative system will have tobe accelerated to meet WTO rules. The more rapidreform will break down the scope to apply protec-tion at the departmental and district levels, and it will facilitate competition in the market at a national level. Bosworth (2002) points out,however, that the obligations under the GATS cov-ering subnational governments are weaker thanthose applying to national governments. Therequirement is that WTO members take only“reasonable measures” to ensure that subnationalgovernments meet their obligations. However,China’s commitments under its Protocol of Acces-sion specifically require that it maintain a uniformsystem of administration. Furthermore, the centralgovernment is required to establish a mechanismwhereby those concerned about problems ofregional protection can bring their concerns to itsattention. These specific provisions seem likely tomake the disciplines on China stronger than thoseunder general WTO rules.

Separation of Government Administration fromEnterprises Some government departments arestill directly or indirectly involved in enterprise


TABLE 9.4 Summary of Indirect Effects of WTO Accession on China’s Logistics Sector

Area of Existing Problems Evaluation of Indirect Impact

Local protectionism GATS obligations also apply at subnational government level.Mixing of government administration State-owned enterprises are under pressure to transform

and enterprise management to a market-oriented operation, and government adminis-Inadequate infrastructure and tration can be separated from enterprise management.

facilities/equipment Sector likely to attract more foreign investment.Lack of research and professional skills Foreign-invested enterprises will bring know-how, innovation,

technology, and management.


activities. As already noted, in the railway and airsectors state ownership is still dominant. Portadministration is mixed with port enterprise oper-ations. These situations not only affect governmentfunctioning, but also weaken the competitivenessof the enterprises.

Logistics Infrastructure and Facilities/EquipmentThe scale of the infrastructure and the quality offacilities and equipment available to logistics ser-vices suppliers are relatively poor. The assessmentby the Hong Kong Trade Development Council(2002) is that the airport infrastructure is a “majorimpediment to aviation industry growth.” Further-more, there is an imbalance in developmentbetween the eastern and western parts of the coun-try. Road and rail transport, storage and warehous-ing, container station and depot services, andairports are some of the areas that could attractmore foreign investment after accession.

Human Resources Research on logistics, espe-cially among business firms, is rare, and the profes-sional education system has so far included littlelogistics training. Foreign investment and establish-ments in the logistics services sector will bring withthem technology and management know-how. The

relevant know-how includes not only the operationof particular segments of the logistics operation,but also the manner in which the elements fittogether, as well as the management of the relation-ships between all the suppliers and the customers.Opening other service sectors, such as education,will also raise the quality of options available foreducation and training in this sector.

Cost Savings in the Third-PartyLogistics Sector

Little empirical work has been conducted on theefficiency of the logistics sector in China or on theimpacts of reform. Some ideas about methodologyand some indicators of the scale of impacts areavailable, however, from work done in othermarkets—for example, on the effects of the integra-tion of transport markets in Europe (EuropeanCommission 1997).4

Those studying the European logistics servicesmarkets identified eight input-oriented measuresof logistics performance: (1) customer orientation;(2) integrated long-range planning; (3) supplierpartnerships; (4) cross-functional operations;(5) continuous improvement process; (6) emplo-yee empowerment; (7) integrated information


BOX 9.1 Local Policies and Their Implications for Logistics Operations

The Jia Yu Freight Group Company is a new pri-vate transport and logistics services company. Itsmore than 20 branches cover most of the coun-try to provide road express service in a fast-growing business. The company manager,Mr. Zhai, complains about the business environ-ment, especially the local protection:

We face many obstacles in operating ourbusiness and providing our services to cus-tomers. Transport needs both registration inthe Administration Bureau of Industry andCommerce of the local government and alicense from the Road Administration Bureau[RAB]. Each local RAB has a different prac-tice. Some cities like Changzhou in JiangsuProvince and Wenzhou in Zhejian Provincedo not issue licenses to companies from otherregions. Our trucks are not allowed to enterthese cities. If we carry cargoes from theircities, we will be fined. Some cities are not soextreme in their treatment of the outside

transport companies, but we have to fol-low their special practice. For example, inGuangzhou we have to rent their designatedoffice room with a higher rental in a specifiedplace. Otherwise, we are not allowed to enterthe market. In Guangzhou, we have todeposit a larger amount of money andadministration fee. But whenever we meetdifficulties, there is no assistance from them.

In general, local governments do not have iden-tical rules and regulations. The road sector hasbeen deregulated, but the local governmentalways wants to protect its local companies byvarious means, which limits the developmentof cross-regional logistics services. “The strangething is the more developed the region, themore restrictions there are on the outside firms.For example, in Chongqing we haven’t metmuch difficulty in establishing our presence andoperating our business,” Mr. Zhai added.

Source: Author interviews.

technology (IT) systems; and (8) measurement,comparison, and action.

Different stages of development of the logisticssector can be identified through the use of theseindicators. Higher levels of performance, as meas-ured by these indicators, lead to better performancein terms of the costs of each element of the serviceand in terms of other indicators such as on-timedelivery, order completeness, invoice accuracy, anddamage-free delivery. All these features can thereforebe summarized in the total cost of logistics services,which refers to the sum of inventory, administrative,warehousing, and transportation costs. Not onlydoes market opening lower the prices of the directinputs (e.g., transport), but it also raises servicesquality and thereby lowers other costs (e.g., ware-housing or inventory, or allowances for losses).

Transport is usually the single major componentof logistics costs. For example, in 1992 in Europelogistics costs were on average 10.1 percent of rev-enue, down from 14.3 percent in 1987. The 1992share comprised 3 points for transport, 2.3 pointsfor warehousing, 1.9 points for administration, and2.9 points for carrying inventory. The logistics costshare of revenue was expected to fall to 9.1 percentin 1997. In the firms regarded as leaders in the oper-ations of logistics, the logistics share of revenue wasestimated to be as low as 6.5 percent (1992), 36 per-cent lower than the overall average. The largestdecreases over this period in Europe were observedin the component associated with inventory costs,followed by those linked with administration, ware-housing, and transportation. Transport, even inthose leading companies, remains the largest singlecomponent of total logistics costs.

The significance of logistics costs varies betweensectors of the economy as illustrated in table 9.5.

According to Gibson (2001), electronics is associ-ated with a significantly higher share of logisticscosts in total revenue than textiles. Gibson alsonotes that in some sectors, such as fresh food, logis-tics costs can be 50–60 percent of total revenue.

The table also illustrates the variation in the sig-nificance of logistics costs in Asia and those inEurope—the latter are lower by 24 percent. Gibson(2001) suggests that in China logistics could accountfor 30–40 percent of the total cost of manufacturedgoods. Another source suggests that “supply-chain-related costs can be 30% to 40% of wholesale pricesin China, compared to 5% to 20% in the U.S.”5 TheHong Kong Trade Development Council (2002)refers to an assessment that logistics account for40 percent of general production costs and that“logistics . . . takes 90% of the whole productioncycle time.”

A key factor in the performance of the logisticssector is the damage rate. One source reports thatfor fast-moving consumer goods, the damage rateis 5 percent in China compared with “well below”1 percent in developed economies.6 High rates ofproduct loss lead to requirements for higher inven-tory levels. The costs of inventory and product losscan be significant, which reduces the relativeimportance of the transport component of the totalcost of distribution. Switching to a higher-qualitylogistics service, even one that charges higher fees,for example, a foreign provider can then loweroverall logistics costs if the costs associated withproduct loss are reduced sufficiently.

Boillot and Michelon (2000) provide a break-down of the total logistics costs in China. Theyreport that losses and damage account for 48 per-cent of the total, compared with 21 percent fortransport activities. They also quote industry


TABLE 9.5 Logistics Costs as a Share of Total Revenue,Europe and Asia, 2001 (percent)

Sector Europe Asia

Textile 8.7 11.5Automotive 8.9 13.5Chemicals 10.2 13.8Fast-moving consumer goods 10.4 14.0Electronics 12.6 14.0Average 10.4 13.7

Source: Gibson 2001.

estimates that total logistics costs could fall by37 percent if the current arrangements for manag-ing logistics were replaced by a subcontractingapproach. The bulk of this reduction appears, intheir assessment, to be attributable to the reductionin losses and damage rather than a reduction inprofit margins. Management fees may increase toreflect the higher costs of the higher levels of man-agement inputs in time and experience.

In summary, a more open market for logisticsservices that transfers the technology for theredesign and management of integrated services islikely to have significant effects on total logisticscosts in China. The composition of those costswill change—for example, management fees willincrease, but this change will be offset by reduc-tions in other items, not necessarily in the transportcomponent but more likely in the costs associatedwith damage, losses, and inventory.

To summarize this discussion, if total logisticscosts are 30–40 percent of wholesale prices inChina, as suggested earlier, and if those costs fall by35 percent, then logistics costs as a share of whole-sale prices would fall by 10–14 percentage points, to20–26 percent of wholesale prices. But even thisassessment may be conservative, because logisticscosts as a share of the wholesale price in China isstill above the average for Asia in table 9.5. It alsolies above that observed in developed marketeconomies, although there are limitations on theextent to which comparisons of these ratios can bemade across countries.7

Implications for Less-developedAreas in China

As noted in the previous section, reforms associ-ated with China’s accession to the WTO couldresult in a fall in logistics costs equal to 10–14 per-cent of wholesale prices in China. In an exampledeveloped by Radelet and Sachs (1998), supposethat a region of China faces a perfectly elasticdemand for its output and a perfectly elastic supplyof purchased intermediate inputs that come fromoutside the region. Suppose that the share of pur-chased inputs in gross value is 30 percent. If thenlogistics costs are 35 percent of the value of outputand inputs, the value added is 24.5 percent of thevalue of output. If logistics costs fall to 23 percentof the value of output, the value added rises by64 percent, to reach 40 percent of the value of

output. Changes in logistics costs even in the rangesuggested here can make a significant difference inthe value added available for distribution to the fac-tors of production. If the payment to capital isfixed, then the amount of valued added available tolabor changes by an even greater proportion.8 Thisexample illustrates the significance of logistics effi-ciency for inland areas of China.

A more efficient logistics sector will benefit peo-ple living in poor areas in several ways. For exam-ple, as implicit in the example just given, theirterms of trade will improve; the costs of items theybuy from the rest of China will be less, and theprices they receive for the items they export to therest of China will be higher. The impact of reform,which reduces logistics costs, also generates exten-sive real gains: the impacts on real incomes aregreater than those of the removal of a tax, for exam-ple. The impact of tax removal includes transfereffects, but when logistics costs are reduced,resources are saved. In other words, the rectangleeffects are gains, not just transfers (see Deardorff2001).

Reductions in transport costs can also lead tosignificant increases in trade orientation. Some ofthe orders of magnitude involved in these effectsare illustrated by the work on Africa by Limao andVenables (2001) who estimate the response of tradeto reductions in transport costs. They find anelasticity of 2.5, which could be applied to thereduction of 35 percent in logistics costs, accordingto the assessment reported earlier.

The opening up of markets stemming fromimprovements in transport and logistics servicesalso can have effects on competition in the localmarkets. And competition itself has additionaleffects. One is the impact on the rents available.The suppliers who previously had monopolypower are constrained by the options now avail-able to their customers to seek supplies from out-side the region. Thus not only does regulatoryreform lower costs, but also its competitive effectsforce the passing on of those cost reductions toconsumers in business and in households.9 Firmsthat previously earned rents from their protectedposition are clearly worse off, but overall theregion will realize a welfare gain from the intro-duction of competition. In terms of the welfare ofthose people living in the poor area, this effect iseven larger when those who captured the rentswere not local firms.


Competition, and a greater foreign (both fromout of the country and out of the region) presencein all markets within a poor area, also could have adynamic effect on, for example, the transfer oftechnology and productivity growth. This effect islikely to be greater in those services markets wherean establishment is required to deliver the activity,and where there is therefore more interactionbetween the new supplier and local firms.

After markets become more competitive, somelocalities may continue to lack services of a qualitytypical of those in more developed areas. Govern-ments often respond to this situation by subsidiz-ing the provision of services. Subsidies might bearranged within the sector (supported by a regula-tory structure that facilitates these transfersbetween consumers), or through explicit paymentsfrom the budget. The design of these policiesincludes important efficiency considerations, bothon the funding side and in the manner in which thesubsidies are applied.

Gruen (2001) has questioned and proposedoptions for the management of service obligations.The popular approach to a service obligation policy isto use cross-subsidies between consumer groups toprovide specific services (e.g., a rail link) to all resi-dents in a particular area. Gruen suggests insteadstrategies in which services are provided to well-defined target households and where the recipientshave options for the manner in which they receivethat support (“cashing out,” for example, or at leastsubstitutability between modes of supply—in thiscase, of transport services).Analysis of costs and ben-efits of alternatives is important—for example, thoseof building a road to an existing community or imple-menting an option for relocating that community.10

An important consideration in the design ofpolicies to support provision of services to poorareas is their consistency with obligations underthe WTO on subsidies. Models for the design ofthese policies are available, however—for example,in the WTO’s April 1996 reference paper on basictelecommunications, which notes:

Any Member has the right to define the kindof universal service obligation it wishes to main-tain. Such obligations will not be regarded asanti-competitive per se, provided they are admin-istered in a transparent, non-discriminatory andcompetitively neutral manner and are not moreburdensome than necessary for the kind of uni-versal service defined by the Member.11

China can learn much from the considerableexperience in the rest of the world with communityservice obligation policies, and the scope for Chinato innovate in policy development in this field isgreat.

Conclusions

The China experience illustrates that the design ofcommitments in the GATS is complicated in newbusiness areas. For example, the GATS categoriesdo not match exactly the business structures usedto provide logistics services. However, it is possibleto make a concordance: China’s commitments,which are significant and which in most parts gobeyond a binding of current policy, provide a tem-plate that other developing economies may con-sider for this sector.

China’s experience also illustrates how WTOcommitments can open up the component activi-ties of the logistics sector, including areas that inthe past might have been regarded as nontrade-ables. The commitments thereby also facilitate thegrowth of third-party logistics providers. Thedesign and management of the provision of ser-vices in this field is important; not just the extentand quality of physical capacity matters. The pres-ence of foreign providers of these integrating ser-vices will be facilitated by the WTO commitments.

A feature of the set of commitments reviewedhere was the application of timetables for theirimplementation. Some motivations for thatapproach can be identified, but it comes at somecost, including the higher risks of backsliding andthe use of inefficient forms of business organ-ization by firms that enter in anticipation offurther market opening—for example, where jointventures are permitted but full foreign ownershipis not.

One of the challenges, however, is to promotecomplementary administrative and regulatoryreform. The China case highlights the risks associ-ated with administrative structures, even at the cen-tral level, that add to the complexity of establishingthe most efficient business structures (e.g., forintermodal transfers of freight). Perhaps an evenmore important issue is to translate national com-mitments to the WTO into local government pol-icy, especially in regions where there has been atradition of using regulatory instruments to protectlocal businesses. Implementation of the principles


of transparency and nondiscrimination in theapplication of business regulation are importantelements of the extent to which markets are actuallyopened to foreign competition.

A more open logistics sector, especially throughthe growth of 3PL providers, can lead to significantcost reductions. One estimate, based on assess-ments of the state of logistics services in China byindustry commentators, is that the more wide-spread presence and use of efficient and interna-tionally competitive 3PL providers could lowercosts by an amount equivalent to at least 10 percentof the wholesale price of manufactured goods onaverage. The impact on these relative terms may beeven higher for fresh food produce. These amountsare not simply transfers, but real gains associatedwith the saving of resources. For example, animportant source of gains is the lower wastage ratesin transport and warehousing.

Changes of these magnitudes in the quality oftransport services could have important effects forpoor areas. Their lack of access to markets in therest of the country and in the rest of the world is animportant constraint on their development. Amore open and competitive logistics sector willremove some of the impediments to their partici-pation in trade within the economy and with therest of the world. The effects of lower logistics costson value added and on the total value of wages paidand available for distribution to workers are signif-icant. In an example reported here, a 12 percentagepoint reduction in the logistics costs in the whole-sale value of goods leads to an increase in valueadded of more than 60 percent.

However, some impediments remain to the sup-ply of services in remote areas. Policies to support theprovision of services where these impediments applyare not inconsistent with the GATS, but they shouldbe evaluated against alternative strategies for pro-moting the income growth of those communities.

Appendix: China’s Commitmentson Logistics Services–RelatedSectors and Their Implications

Air Transport

Air transport remains quarantined from GATS cov-erage except for a limited number of complementaryservices.12 Even in these areas, China’s commitmentsare relatively limited.

1. Aircraft Repair and Maintenance Services

COMMITMENT

Foreign service suppliers are permitted to estab-lish joint venture aircraft repair and maintenanceenterprises in China. The Chinese side should holda controlling share or be in a dominant position inthe joint ventures. Licenses for the establishment ofjoint ventures are subject to an economic needs test.

IMPLICATIONS

The commitment effectively binds current pol-icy. Two joint venture aircraft repair and mainte-nance service firms are already located in China:AMECO (founded in July 1989) with a 40 percentforeign share and GAMECO (founded in August1989) with 50 percent share. Binding policy mayreduce uncertainty from the perspective of poten-tial entrants. Further entry would add to competi-tion and may facilitate the introduction of newtechnology and management expertise in aircraftrepair and maintenance, so that a higher volume ofqualified services are made available within China.

2. Computer Reservation System (CRS) Services

COMMITMENT

For cross-border supply:

• Foreign computer reservation systems, when hav-ing agreements with Chinese aviation enterprisesand the Chinese computer reservation system,may provide services to Chinese aviation enter-prises and Chinese aviation agents by connectingwith the Chinese computer reservation system.

• Foreign computer reservation systems may pro-vide services to representative offices and salesoffices established in the destination cities inChina by foreign aviation enterprises that havethe right to engage in business under the bilat-eral aviation agreements.

• Direct access to and use of foreign computerreservation systems by Chinese aviation enter-prises and agents of foreign aviation enterprisesare subject to the approval of the general admin-istration of the Civil Aviation Administration ofChina (CAAC).

IMPLICATIONS

The commitment refers only to foreign connec-tion with the Chinese system. The commitment


confirms policy as it is now. Chinese carriers havecode-sharing arrangements with carriers in morethan 10 foreign countries, but direct foreign partic-ipation in the provision of CRS services is notallowed. Commercial presence by foreign providersis still not permitted.

Water Transport (International and Domestic Shipping)

1. Maritime Transport Services: InternationalTransport (Freight and Passengers, (CPC 7211and 7212 Less Cabotage Transport Services)

COMMITMENT

Registered companies may be added for the pur-pose of operating a fleet under the national flag ofChina. Foreign service suppliers are permitted toestablish joint venture shipping companies, but theforeign investment shall not exceed 49 percent ofthe total registered capital of the joint venture. Thechairman of the board of directors and the generalmanager of the joint venture shall be appointed bythe Chinese side.

IMPLICATIONS

The commitment is the same as current policy.

2. Auxiliary Services: Maritime Services (CPC 741),Customs Clearance Services for MaritimeTransport

COMMITMENT

Foreign participation is possible only in theform of joint ventures, but foreign majority owner-ship is permitted.

IMPLICATIONS

Joint ventures are already allowed, but currentlythe maximum foreign share is 49 percent. Thecommitment relaxes that constraint.

3. Container Station and Depot Services

COMMITMENT

Foreign participation is possible only in theform of joint ventures, but foreign majority owner-ship is permitted.

IMPLICATIONS

Joint ventures are already allowed, but currentlythe maximum foreign share is 49 percent. Thecommitment relaxes that constraint.

4. Maritime Agency Services

COMMITMENT

Foreign participation is permitted in the form ofjoint ventures, but the foreign equity share may notexceed 49 percent.

IMPLICATIONS

Maritime agency services are not presently open,so this commitment is significant. The level of for-eign ownership remains constrained, however.

5. Internal Waterways Transport: Freight Transport(CPC 7222)

COMMITMENT

For cross-border supply: international shippingis possible on routes to ports that are open to for-eign vessels.

IMPLICATIONS

This policy will apply only in ports open to for-eign shipping. In 1999 there were 130 such ports.

Road Transport

Road Transport Services: Freight Transportation byRoad in Trucks or Cars (CPC 7123), Storage andWarehousing Service (CPC 742)

COMMITMENT

Foreign participation is permitted in the form ofjoint ventures, but the foreign equity share may notexceed 49 percent. Within one year of China’saccession, foreign majority ownership is permitted.Within three years of China’s accession, whollyforeign-owned subsidiaries are permitted.

IMPLICATIONS

In the road transport sector joint ventures areallowed at present, but foreign partners are allowedonly a minor share (less than 49 percent). Thissector is now very competitive with low marginsbecause it was subject to domestic deregulation in1986. Road transport is an essential link of thelogistics chain and, of all the transport sectors,plays the most important role. Moreover, of thefour modes of transport, the road sector is thefastest growing. Foreign logistics firms might havean interest in participating in the road transportsector by either providing new capacity (perhapsin special road transport services) or buying anexisting fleet. Some of the effects of deregulation


include greater use of subcontracting and larger-scale and lower-cost operations.

Rail Transport

Rail Transport Services: Freight Transportation byRail (CPC 7112)

COMMITMENT

Foreign participation is permitted in the form ofjoint ventures, but the foreign equity share may notexceed 49 percent. Within three years of China’saccession, foreign majority ownership will be per-mitted and within six years of accession whollyforeign-owned subsidiaries will be permitted.

IMPLICATIONS

Railway transport continues to be monopolizedby the government. China’s WTO commitment inthis sector reveals its intention to open this trans-port up to the outside world. The process is, how-ever, slower than that in the road sector. Only sixyears after China’s accession to the WTO willwholly foreign-owned subsidiaries be permitted.Some parts of the rail sector are profitable, and oth-ers are not (e.g., in rural areas). Foreign firms willtake an interest in the profitable lines or regions.

Evidence from other markets indicates that pro-ductivity in rail is related to government policy. Keyfactors include degree of managerial autonomy andthe extent of subsidies made available. Higher levelsof private (and foreign) ownership would beexpected to add to productivity in this sector. Oneimportant issue is how the unprofitable lines can bemaintained if that is required to meet service obliga-tions. The effects of reform in this area could be sig-nificant. According to one source, “Most logisticscompanies and distribution managers continue toargue that there is never a good reason to use rail fortransport.” Factors in this view include damagerates, delays, slow speed, lack of information andbooking requirements, and lack of route flexibility.13

Warehousing and Storage

Storage and Warehousing Services (CPC 742)

COMMITMENT

Foreign participation is permitted in the formof joint ventures, but the foreign equity share maynot exceed 49 percent. Within one year of China’saccession, foreign majority ownership is permitted,

and within three years of accession wholly foreign-owned subsidiaries are permitted.

IMPLICATIONS

Storage and warehousing are an important linkin logistics services. Joint ventures are allowed atpresent, but foreign partners are allowed only aminor share. The margins in this sector are low.There also appears to be excess capacity, but theservice standard is low, and the mix of servicesavailable does not match demand. The opening upof the market could help foreign-owned logisticssuppliers establish their own presence in this sectorby either building new warehouses or buying theexisting warehouses that are appropriate.

Forwarding

Freight Forwarding Agency Services (CPC 748 and749, Excluding Freight Inspection)

COMMITMENT

Upon China’s accession, foreign forwarding agen-cies that have at least three consecutive years of expe-rience are permitted to set up freight forwardingagency joint ventures in China, with foreign invest-ment not to exceed 50 percent. Within one year ofChina’s accession, foreign majority ownership ispermitted. Within four years of accession, whollyforeign-owned subsidiaries will be permitted. Theminimum registered capital for a joint venture is tobe no less than US$1 million. With four years ofaccession, national treatment will be accorded in thisrespect. The term of the joint ventures shall notexceed 20 years. After one year of operating in China,the joint venture can set up branches, but anotherUS$120,000 should be added to the original regis-tered capital of the joint venture for the setup of eachbranch. Within two years of China’s accession to theWTO, this rule on an additional registered capitalrequirement will be implemented on the nationaltreatment basis. A foreign freight forwarding agencymay set up a second joint venture after its first jointventure has been in operation for five years. Withintwo years of China’s accession to WTO, this require-ment will be reduced to two years.

IMPLICATIONS

Joint ventures are allowed at present, but theforeign partner is allowed only a minor share.Within four years of China’s accession, wholly


foreign-owned subsidiaries will be permitted.Foreign-owned firms could enter the market easilyafter the restriction is lifted. Many firms havealready presented themselves in the market in theform of a joint venture. Although, according to theregulations, only a minor share is allowed, the for-eign partner actually may control the business.Some firms enter the market by using the “shell” ofa domestic company. Complete opening of themarket will reduce for these firms the costs of pro-viding forwarding services and also enhance theirability to provide “whole of chain” services.

Related Areas

Courier Services (CPC 75121, Except Those CurrentlySpecifically Reserved to Chinese Postal Authorities byLaw)

COMMITMENT

Upon China’s accession, foreign service suppli-ers are permitted to establish a joint venture, withforeign investment not exceeding 49 percent.Within one year of accession, foreign majorityownership is permitted. Within three years ofaccession, foreign service suppliers are permitted toestablish wholly foreign-owned subsidiaries.

IMPLICATIONS

Courier services are an important part of logis-tics services. A joint venture is allowed at present,but only with a minor share for foreigners. Theopening up of this sector could bring up morecompetition in this market, and there is already evi-dence of the interest of foreign firms in this sector.

Packaging Services (CPC 876)

COMMITMENT

Foreign service suppliers will be allowed toestablish a joint venture in China. Within one yearof China’s accession, foreign majority ownership ispermitted. Within three years of China’s accession,foreign service suppliers are permitted to establishwholly foreign-owned subsidiaries.

Notes

1. These sections are summaries of the material in Luo andFindlay (2001). For a review of developments in the logistics sec-tors in three East Asian economies, including China, see Heaver(2001).

2. These systems operated in these sectors: raw material andcomponents (wu zi), manufactured consumer products (shangye), rural area suppliers and distributors (gong xiao), foreign

trade goods (wai mao), railway, road, and water transportation(jiaotong), military, and basic food (liang shi).

3. Distribution (trade) is different from logistics. But therelated subordinated services are increasingly likely to beincluded in the logistics sector, because the sector has taken onmore and more of the value-added services demanded by ship-pers. As defined, distribution trade services are composed offour main subsectors: commission agents’ services, wholesaling,retailing, and franchising. The principal services rendered ineach subsector can be characterized as reselling merchandise,accompanied by a variety of related subordinated services,including inventory management; assembly, sorting, and grad-ing of bulk lots; breaking bulk lots and redistributing intosmaller lots; delivery services; refrigeration, storage, warehous-ing, and garage services; sales promotion, marketing, and adver-tising; and installation and after-sales services, includingmaintenance and repair and training services (covered by CPC[Central Product Classification] 61, 62, 63, and 8929).

4. The experience of regulatory reform in the United Statescan be used as another comparator. A longer time-series of dataon the European experience may also now be available.

5. Andrew Tanzer quoting McKinsey & Co. in “ChineseWalls,” Forbes Global, November 12, 2001. Available atwww.forbes.com/global/2001/1112/091.html.

6. Economic Intelligence Unit, China Hand (database),December 1999, Chap. 12, Sec. 1.

7. The ratio of logistics costs to wholesale prices in differentcountries will depend not only on the characteristics of theproduct, but also on the factor intensities of the different pro-duction processes of both the logistics services and the manufac-tured products and the relative factor prices in the economyconcerned.

8. For example, if the cost of payments for capital is 20 per-cent of the value of output, the funds available for the wage billrise by a factor of 3.5 as logistics costs fall from 35 to 23 percentof wholesale prices.

9.According to anecdotal evidence, this effect has appeared inmarkets for light industrial products made in the past by town-ship and village enterprises. Brand-name products from suppli-ers based in eastern China are now much more widespread.

10. Gruen also discusses the value of broadening the fundingbase for service obligation policies.

11. Available at www.wto.org/english/tratop_e/serv_e/telecom_e/tel23_e.htm.

12. See Hong Kong Trade Development Council (2002) foranother review of China’s WTO commitments in transportationand logistics.

13. Economic Intelligence Unit, China Hand (database),December 1999, chap. 12, sec. 3.

References

Boillot, Jean-Joseph, and Nicholas Michelon. 2000. “The NewEconomic Geography of Greater China.” China Perspectives30 (July–August): 18–30.

Bosworth, M. 2002. “Most-Favoured-Nation Treatment andNational Treatment in the GATS.” In Sherry Stephenson andChristopher Findlay, with Soonhwa Yi, eds., Services TradeLiberalisation and Facilitation. Canberra: Asia Pacific Press,ANU.

Coyle, J. J., E. J. Bardi, and J. C. Langley. 1996. The Management ofBusiness Logistics. 6th ed. St. Paul, Minn.: West Publishing Co.

Deardorff, A. 2001. “International Provision of Trade Services,Trade, and Fragmentation.” Review of International Economics9 (2): 233–48.


European Commission. 1997. Impact on Services: TransportNetworks. The Single Market Review, subseries II, vol. 11.Luxembourg: Office for Official Publications of theEuropean Communities.

Gibson, Ken. 2001. “Analysing the Trends: Predicting the Futureof Logistics in Asia.” Presentation to the conference “Towardthe Final Frontier: Logistics and the Efficient Supply Chain.”September 27. Available at www.dnmstrategies.com/presentation/tnt/sep_2001/KenGibsonV1.0-0945.pdf.

Gruen, N. 2001.“Beyond the Safety Net—A View from Outside.”Presentation to the APEC Telecommunications WorkingGroup, Canberra, March.

Heaver, T. 2001. “Logistics in East Asia.” Development EconomicResearch Group, World Bank, Washington, D.C.

Hong Kong Trade Development Council. 2002. “China’s WTO Accession: Enhancing Supply Chain Efficiency—Transportation and Logistics.” Hong Kong.

Limao, N., and T. Venables. 2001. “Infrastructure, GeographicalDisadvantage and Transport Costs.” World Bank EconomicReview 15: 315–43.

Luo W., and C. Findlay. 2001. “Analysis of Logistics Service Sec-tor in China: Accession to the WTO and Its Implications.”Paper prepared for Authors’ Workshop on WTO Accessionand Poverty Reduction, Beijing, October 26–27.

Radelet, S., and J. Sachs. 1998. “Shipping Costs, ManufacturedExports and Economic Growth.”Paper presented at AmericanEconomic Association Meetings, Harvard University.Processed. Available at www2.cid.harvard.edu/hiidpapers/shipcost.pdf.

Redding, S., and T. Venables. 2000. “Economic Geography andInternational Inequality.” CEPR Discussion Paper SeriesNo. 2568. Centre for Economic Policy Research, London,September.

WTO (World Trade Organization). 1996. Telecommunica-tions services: Reference Paper, Negotiating Group onbasic telecommunications. WTO, Geneva. Available athttp://www.wto.org/english/tratop_e/serv_e/telecom_e/tel23_e.htm.


the five-year phase-in period under its WTO acces-sion agreement. The second section sets out toanalyze the effects of current reforms on greater effi-ciency, reduced costs, improved quality, and greateraccess in the sector. The third and fourth sectionsevaluate the post–WTO accession environmentfaced by China and analyze the possible direct andindirect implications. The final section draws gen-eral conclusions and policy implications about thetelecommunications sector in China.

The Landscape of theTelecommunications Industrybefore WTO Accession4

In anticipation of the liberalization commitments itwould have to make upon its accession to the WTO,China began in 1994 to undertake reforms toincrease competition in its telecommunications sec-tor. It took tentative first steps in 1994 to change themonopoly structure, but only in the last five years,beginning in 1998, has it undertaken major reforms.

Tentative First Steps: Constrained Competition

Up to 1994 China Telecom was the sole basictelecommunications services provider in China.New market entry was prohibited, and the Ministry

Despite some significant reforms, China’s telecom-munications sector since the country’s accession tothe World Trade Organization (WTO) remains oneof the most restricted and regulated of those in themajor developing countries in the region. Duringthe WTO negotiations, liberalization of this sectorwas a critical issue both because of its growthpotential and because it was considered one ofChina’s “key national industries.”1 By the end of thedecade, China is projected to become the largestmarket for telecommunications services in theworld, creating a strong incentive for foreign play-ers to enter the market.2

Reforms were sought before and in associationwith China’s WTO accession, as well as in responseto technological advances and the gradual com-moditization of services.3 However, competitionremains constrained, and obstacles include lack ofclear and concise legislation, regulatory implemen-tation, and regulatory independence. The key chal-lenge for China is to ensure that regulatory reformsand new institutions will create greater competitionthat will maximize the benefits for the Chineseeconomy as well as meet equity objectives.

This chapter has several goals. The first majorsection seeks to examine the pace and progress ofchange in China’s telecommunications services sec-tor and to assess China’s position at the beginning of

157

10

TelecommunicationsServices in China:

Facing theChallenges of WTO

Accession

Mari Pangestu and Debbie Mrongowius

of Posts and Telecommunications (MPT) not onlyregulated the sector but also operated telecommuni-cations services. In 1994 the China United Telecom-munications Corporation (Unicom) was created toimprove service quality by introducing managedcompetition (see chapter appendix). However, pric-ing policies, such as service fees, remained under theMPT, and individual operators were not able tocompete through the price mechanism. Unicomfaced an uphill battle in its early years.5

The biggest problem faced by Unicom was get-ting access to the backbone infrastructure con-trolled by China Telecom. When Unicom was firstgranted fixed-line services in some provinces, thelocal China Telecom operators refused to allocatephone numbers to Unicom, which held up inter-connection for over a year. Another problem wassecuring access to capital, which Unicom didlargely by using “gray area” measures, the most wellknown of which was the Chinese-Chinese-Foreign(Zhong-Zhong-Wai or CCF) joint venture enter-prise structure, to circumvent the ban on foreigndirect investment (FDI) in the sector. The govern-ment subsequently initiated an investigation intothe legality of CCF ventures in light of the 1993directive forbidding foreign investment in the sec-tor. The investigation resulted in restrictions onCCF arrangements: a mandatory limit on agree-ments of 15 years, a five-year limit on profit shar-ing, and transfer of 90 percent of all network assetsto the Chinese party.

Another company, Jitong, was created in 1994 aspart of the “Golden Bridge Network”—a govern-ment initiative to develop a series of informationnetworks. Jitong’s original objective was to providedata services via satellite connections, and it wasintended to compete with China Telecom. Most ofJitong’s services have now been shifted to an Internetprotocol (IP) network infrastructure. These servicesinclude Internet access and gateway services; value-added services such as Web site hosting, server man-agement, and e-commerce solutions; and VoIP(Voice over Internet Protocol). In early 1999 Jitongwas the first Chinese operator to offer VoIP, and itbecame at once a dominant player in the market.

The reforms undertaken up to 1998 were signif-icant first steps, but the competition remained con-strained, the service quality poor, and the pricesuncompetitive by global standards. Prices were stillset by the MPT, the State Pricing Board, and the

State Development Planning Commission (SDPC).In response to the inefficiency in the telecommuni-cations system and its inability to provide connec-tion, rural communities began to finance andestablish their own systems to provide basic servicein their areas. These town and village enterprises(TVEs) then merged into larger organizations pro-viding wider coverage and eventually set up admin-istrative relationships with the local office ofthe MPT. In addition, various networks had alsoevolved within separate industries and institutionsin order to compensate for the supply shortage inthe public network maintained by the MPT. Thesenetworks included those for the academic andresearch communities, the military, the railways,and the oil industry.

To separate the regulatory and operational func-tion of the MPT, as well as to respond to changes intechnology and administrative reforms, the Chinesegovernment created the Ministry of InformationIndustry (MII) in March 1998.6 This “super agency”was created by merging the MPT; the Ministry of Electronics Industry (MEI); and parts of theState Administration of Radio, Film and Television(SARFT), China Aerospace Industry Corporation,and China Aviation Corporation. The MII workedclosely with the State Economic and Trade Com-mission (SETC) in setting policies for the informa-tion industry about the introduction of foreigninvestment and technologies, as well as in establish-ing industry strategic development plans. However,the independence of the MPT and China Telecomwas not convincingly demonstrated, in part becausemany of the senior staff at the MII came from theMPT.

Breaking Up the China Telecom Monopoly

The MII’s fear that the monopoly it enjoyed wouldbe sacrificed as a “pawn” for China’s accession tothe WTO, and its conviction that national interestswould be compromised if China’s telecom industrydid not survive the resulting competition, galva-nized the ministry into devising a survival plan,beginning in 1998. This strategy involved creating afavorable policy environment for China Unicomand breaking up the China Telecom monopoly.

In 1999 the MII launched major restructuringefforts by splitting China Telecom into three state-owned companies and reinvigorating Unicom. The


three companies emerging from China Telecomwere: China Telecom for fixed-line business; ChinaMobile Communications Corporation (CMCC orChina Mobile) for mobile phone business; and ChinaSatellite (ChinaSat) for satellite communications. Inaddition, a new company, China Netcom Corpora-tion (CNC) was established to build a broadbandInternet protocol network. The division of enter-prises by subsector reflected MII’s plan to speed upthe development of individual service sectors.

In 2000 a license was also given to Railcom, whichwas part of the Ministry of Railways, to provide allbasic telecommunications services except mobile.Thus, together with Jitong, at the end of 2001 therewere seven major operators in the telecommunica-tions sector. More than 3,000 different enterprisesare engaged in Internet-related and other value-added businesses, which enjoy greater operationalfreedom and are market-oriented. Current opera-tors are only granted licenses that specify the typeof telecom service they may offer (table 10.1).

Only Unicom has a license to undertake all serv-ices except for satellite. Although mobile telephonyis the most rapidly growing sector in the industry,currently there are only two mobile phone licenseesand no clear indications about whether an addi-tional license will be issued despite much specula-tion that China Telecom will be awarded the covetedlicense. Other advantages of Unicom include being

allowed to charge 10 percent lower rates than ChinaMobile, the right to operate a code division multipleaccess (CDMA) mobile phone system7 and to list itsshares on the Hong Kong (China) stock exchange.Unicom also has a modern management structure,because it did not inherit an existing state-ownedenterprise structure.

China Mobile has listed shares from its mostprofitable subsidiaries through China Mobile Group(Hong Kong) Co. Ltd. The latter company listed24.4 percent of its shares on the Hong Kong and NewYork stock exchanges. Its stated development strat-egy is to shift from traditional circuit-switched net-works to broadband packet-switched IP networksand to diversify from pure wireless voice service tomobile image and data services. Its competitive edgeover Unicom in the mobile sector is its financialresources, geographic coverage, and brand name.

China Telecom owned most of the backboneinfrastructure and leased it to the others. Only ChinaTelecom and China Mobile, which originally waspart of China Telecom, operated in all provinces. Asfor mobile phone market, Unicom has been given aboost since its first uphill struggle due to some of theadvantages mentioned above, and it has reached amarket share of about one-third. Even though otheroperators are covering some areas, Unicom’s cover-age appears to be quite extensive, with expansionplans in place.

Telecommunications Services in China: Facing the Challenges of WTO Accession 159

TABLE 10.1 Network Operators and Licensing, China (end of 2001)

License

InternationalCompany Fixed-Line Cellular Paging VoIP Dataa Gateway Satellite

China Telecom � � � � �

Unicom � � � � � �

China Mobile � � �b �

China Netcom � � �

Jitongc � � � �d

Railcome � � � �

ChinaSat �

aData includes fax and Internet transmission. bOnly mobile data transmission. cJitong serves only corporate clients and has since merged with China Netcom and part of China Telecom.dOnly VSAT.eAlthough Railcom is licensed to provide services other than fixed-line, it has yet to develop the capabilityto do so.Source: Compiled by authors.

As for universal service obligations (USOs), alloperators are subject to these obligations, but thereare no clear implementing regulations. BecauseChina Telecom is the primary owner of the func-tional local end loops, some believe it eventually willbe the major deliverer of universal service. Alreadyrequired to provide telecommunications services togovernment organizations and emergency services,it has implicitly been the one given the task ofproviding universal service.

Regulatory Framework

Since 1998 the Ministry of Information Industryhas been the regulator of telecommunications andother information technology–related productsand services, including data communications, wire-less communications, electronics, computers, theInternet, and software. The MII reports to the StateCouncil and is a member of the State InformationLeading Group (SILG), which is the governmentagency that links the central government (State

Council) and the MII and other government bodiesin charge of information security and enforcement.The SILG was created in the mid-1990s in responseto the growth of Internet service and computercrimes. The group wields much control and influ-ence over the development of the sector, andapproves and modifies the framework for industryregulations and the future direction for the indus-try, and measures to implement policy. The StateCouncil Informatization Office (SCIO) was formedunder the SILG in August 2001 as the executive armwith a mandate to explain and carry out govern-ment policy. SCIO is a high-profile organization,and it plays a more direct role than SILG in develop-ing and implementing information policy in China.

Because the Internet involves information flowsto the public, which raises issues of censorship andsecurity, other relevant authorities are the StateSecrets Bureau, Ministry of Public Security(network security), and the State Administrationfor Industry and Commerce, which is responsiblefor registering Internet service providers (ISPs) and


TABLE 10.2 Network Operators in China: Description of Network and Coverage

Networks Coverage

Company Own Leasea Current Planned

China Telecom � All provinces Add capacity for 18.5 millionnew telephone subscribers and 7.1 million datacommunication users.

Unicom � � 180 cities; mobile: CDMA coverage for 200 cities322 cities

China Mobile � All provinces; GPRS GPRS for all provincesservice in 25 citiesand 16 provinces

China Netcom � � 18 major cities Add 15,000 km of fiber-opticcable

Jitong � � 120 cities; 30 cities 180 cities by 2002covered by city-areanetworks; 1,000 VSATstations

Railcom � 28 cities, 14 provinces 100 cities, all provincesChinaSat � All provincesb n.a.

n.a. Not available. Note: GPRS � general packet radio service.aAlthough all operators have built their own backbone networks, China Telecom still maintains amonopoly over end local loops throughout the country. Currently, interconnection fees must benegotiated individually at every China Telecom province administration.bThis includes 120 duplex remote stations providing access to remote regions such as Guangxi, Guizhou,Inner Mongolia, Qinghai, Tibet, and Yunnan.Source: Compiled by the authors.

Internet content providers (ICPs). The MII, StateDevelopment Planning Commission, and Ministryof Finance (MOF), with approval from the StateCouncil, determine prices and rates.

Prior to the reforms that began in 2001, variousissues arose about the regulatory framework. Onewas the independence of MII from the incumbenttelecom operators, since the latter still comprised ofofficials from the Ministry of Post and Telecommu-nications (MPT) structure and still influential in thenew MII structure. Other issues are that althoughthe MII supervises the telecom network, SARFT pre-sides over cable markets, and other departments reg-ulate the Internet. Given the number of authoritiesin charge of the overall sector, it is often not clearwho is in charge of policies and guidelines for thesector. Another general problem—not just for regu-lation of the Internet but more generally—is the gapbetween national and local policies. Shanghai,for example, has taken the lead in encouraginge-commerce, and local authorities have forged aheadwithout approval from the central government.

To address the problems between national andregional policies, provincial telecommunicationsauthority (PTA) offices were set up in 2001 to con-duct the regulatory function within each province.

Performance of the Sector: Impactof Reforms Prior to WTO Accession

The effects of reforms and the introduction ofcompetition, albeit controlled by the government,are expected to include lower prices, better qualityof service, high growth and telecom penetrationrates, and major changes in the structure of theindustry.

Decline in Prices

Prices for basic telecom services have come downconsiderably in the last few years, but more becauseof mandated price reductions than pure competi-tion. After all, the MII retains a considerable degreeof control over the sector through price controls aswell as through licensing. The MII justifies its con-trol of prices by pointing to its desire to maintainstability in the market and avoid disruptive pricewars. The procedure for telecom rate changes beginswhen an operator submits a proposal, which is fol-lowed by State Council department and ministry

discussions. Public hearings on the proposed pricesmust then be held before the prices are resubmittedto the State Council for final approval and imple-mentation. In practice, basic telecommunicationsservice fees have come down since 1999.

In 1999 the MII slashed service fees, in part toaddress consumer complaints, but also to increaseconsumer access to advanced technology and torationalize the fee structure in anticipation ofChina’s WTO entry. The move was also part of astrategy to provide universal service. The drastic cutsaccelerated telephone popularization in rural areas,doubled the number of Chinese accessing the Inter-net, and enabled Chinese Internet service providersto expand their business and telecom facilities, whichin turn benefited Chinese Internet content providers.Reducing long-distance rates was intended to cutdown on “information smuggling”—that is, chan-neling calls to China through Hong Kong (China),where rates were lower.

In early 2001, the MII, the SDPC, and the MOFjointly announced massive cuts in telecommunica-tions and Internet usage fees. Those cuts were fol-lowed by a second round of cuts in July 2001. Localphone charges per unit dropped from being chargedat three minute to one minute intervals, and thecharge was set at 0.10 yuan ($0.018) per minute.Monthly service fees for fixed-line telephones alsodropped from 24 yuan ($2.90) to 18 yuan ($2.17),and fixed-line phone installation fees were elimi-nated completely. Domestic long-distance chargeshave been standardized at 0.70 yuan ($0.08) perminute and additional charges on long-distancecalls will no longer apply. Long-distance charges forcalls to Hong Kong, Macao, and Taiwan were stan-dardized at 2 yuan ($0.24) per minute, while the ratefor international calls was set at about 8 yuan($0.97) per minute. Initial installation fees havebeen eliminated as well, and the price to lease linesfrom China Telecom was lowered by 72.8 percent.

On top of the mandated price cuts, several opera-tors continued to announce further rate cuts of up to40 percent for domestic long-distance calls, whiledoubling the rates for local calls to maintain theiroperational profits. At the local level, individualoperators have underpriced themselves slightlywhen they have felt local conditions required it.Most competition in pricing takes place throughpackaging mechanisms, which are not subject to theconstraints of the established price ranges.


An example of price competition and its effecton reducing prices is evident in those services, suchas IP telephony, where the government does not haveprice controls. There, intense competition has ledto reductions in the price of long-distance calls. Aprice war on IP telephony was introduced in 2001by China Netcom. The IP phone card, a prepaiddevice for long-distance calling, became very popu-lar because it can save callers up to 70 percent ontheir long-distance call charges. CNC then furtherreduced their charges by another 50 percent, pric-ing IP domestic long-distance calls at 0.3 yuan($0.036) per minute and overseas calls at 2.4 yuan($0.28) per minute. When the other operators—China Telecom, Unicom, Jitong, and ChinaMobile—followed suit and reduced their IP pre-paid cards, the government had to lower long-distance rates in the face of such competition.

Most of these rates appear to be artificially lowand to require significant rebalancing before thesector can become fully competitive. IP card ratesmay be the only ones that closely reflect the truecosts of these services.

Profitability of the Industry

The resulting decline in prices for long-distancecalls had a serious effect on China Telecom’s over-seas listing plans, because China Telecom holds afranchise on fixed-line phone service and makesmost of its money from long-distance calls. ChinaMobile (listed in Hong Kong) and Unicom shouldbe able to benefit significantly from the changesbecause both companies lease lines from ChinaTelecom for their voice and Internet businesses. Bycontrast, China Mobile, Unicom, China Netcom,and Jitong could be adversely affected by thechanges in Internet protocol telephony. With ChinaTelecom’s rates falling, the discount margin forusing IP telephony has been narrowed significantly,removing much of the attraction for using the serv-ice in the first place. In 2001 China Mobile’s averagerevenues per user fell by 35 percent, leading to a fallin its share prices and those of Unicom’s. Theadditional users of mobile are not subscribers; theyare using prepaid cards for which the cost ofadministration and use is much lower.

It remains unclear how profitable each operatoris under current conditions. Those operators whohave partially listed on the stock exchange must

regularly reveal the financial conditions of thelisted subentities, but a significant portion of theirbusinesses remains undisclosed. Corporatizationhas not severed the state connection, and it is notpossible to discern exactly how much support theystill receive from their shareholders and how manyare actually sustained purely by generated revenues.Furthermore, many board members and top-levelmanagement continue to be closely connected withthe state authorities. The support that these rela-tionships could continue to generate, in any terms,is utterly unquantifiable. China Telecom, ChinaMobile, Unicom, and China Netcom claim thatthey are no longer subsidized by the state and areable to run a stand-alone operation profitably.

Licensing and the forced divestiture of ChinaTelecom’s mobile and paging components havereduced cross subsidization between products orinterproduct subsidization, although this possibil-ity has not been eliminated entirely.

Growth and Competition in the Different Subsectors

Fixed-Line The number of fixed-line subscribershas increased substantially in recent years, espe-cially in the post-1998 reform era. In 1996 therewere 54.9 million fixed-line subscribers—a numberthat had doubled to 108.8 million by 1999. By theend of 2002 this number had almost doubled again,with the number of users reaching 180.4 million.By the end of 2003 the number of fixed-line usersreached 210 million (figure 10.1) or slightly lowerthan the number of mobile users. This increase rep-resents a growth rate of 27 percent p.a. or per yearover the 1996–2003 period. The penetration ratesof the population for fixed-line telephony havegone up from 10 percent in 1998 to 14 percent in2003, still leaving considerable scope for growth.Even though Unicom and Railcom had licenses to provide fixed-line services, China Telecomremained the dominant provider with 99 percent ofthe market.

Mobile Phones Mobile telephony began toincrease significantly with the introduction of theGSM (global system for mobile communications)networks in 1995. Domestic and internationalroaming capability was introduced in 1996. The


number of mobile subscribers in China grew from7.3 to 44.3 million over the 1996 to 1998 period,and surpassed the number of users in the UnitedStates in 2001, reaching 144.8 million (comparedwith 120 million in the United States) (fig-ure 10.1).8 Growth has continued, and the numberof subscribers reached 268.7 million by the end of2003, and for the first time was slightly higher thanthe number of fixed-line users at 263.3 million.This increase represents a dramatic growth rate of86 percent over the 2001–2003 period. As a result,the penetration rates have also increased. Animportant contributing factor to the increase in thenumber of mobile phone users has been the intro-duction of prepaid cards, which do not require sub-scriptions or monthly user fees.

Of the only two licensed mobile phone opera-tors in China, both are state-owned companies.Although the market share of UNICOM hasincreased, China Mobile’s share of the market isstill dominant at about two-thirds. In this mobiletelephony market, the two participants in theduopoly have been reluctant to work together. Forexample, lack of number portability eliminatesmuch of the incentive for consumers to switchservice operators. Newly attained market sharerarely comes at the expense of the incumbent;instead, most of the uptake comes from new sub-

scribers. In an effort to eliminate this type of“arbitrage” opportunity between operators, the MIIhas begun to standardize.

The issuance of additional mobile licenses is thesubject of great debate. It remains unclear whetherthe government will issue new licenses, despite indi-cations that such plans are under way to increasecompetition in the sector and despite the need tohave new licenses issued to operators during thetransition to third-generation (3G) technology.However, it is unclear which technology standardsand which 3G technology should be used. A timedivision synchronous code division multiple access(TDSCDMA) was recently created within China,with standards parallel to those of the widebandcode division multiple access (WCDMA) standardfavored by European and Japanese companies.TDSCDMA was also approved by the InternationalTelecommunication Union’s telecom division asone of the three international 3G standards.

Internet The number of Internet users9 in Chinahas also increased rapidly, rising 20-fold—fromabout 1.2 million in 1997 to reach 22.5 million in2000 (figure 10.2). By the end of 2002 the numberhad more than doubled again, to 59.1 million.About two-thirds are dial-up users, and theremainder are connected through leased-line


0

50

100

150

200

250

300

1996 1997 1998 1999 2000 2001 2002 2003(est.)

Millions

Mobile Fixed-line

FIGURE 10.1 Mobile and Fixed-Line Use inChina, 1996–2003 (millions)

Source: Ministry of Information Industry.

connections or use both dial-up and leased lines.Despite the high growth rates, China is a long wayfrom being considered a “wired” nation.

The chief bottlenecks for further expansion ofInternet use include high costs and slow speeds inmost areas. The maximum transmission rate of thetelephone lines used by most people to access theInternet is only 56 kilobits per second, makingdownloading and uploading very time-consuming.Although an integrated services digital network(ISDN) was introduced in 1998, it remains costlyand is still of insufficient bandwidth to allow formore sophisticated online usage and programs.

Shanghai has been able to benefit by using themore advanced asynchronous digital subscriberline (ADSL) services since the beginning of 2000.10

ADSL is not the only broadband technology avail-able. More mature technologies include hybridfiber/coax (HFC), digital subscriber line access(DSLx), local multipoint distribution service(LMDS), and fiber-optic cable. HFC has to datebeen applied in only a few selected pilot projectlocations. Much of the difficulty in getting this tech-nology installed in a wider area arises from issues ofconvergence. The HFC broadband access technol-ogy is the cable modem, which allows transmissionof data to computers and television sets at the sametime. Regulatory reforms are required for this tech-nology to be approved on a nationwide basis.

Perhaps one of the more interesting develop-ments in the industry—and a step on the way

to convergence—is the evolution of Voice overInternet Protocol (VoIP) services. Despite signifi-cant resistance from existing operators and the MII,the possibility of providing reduced long-distancecosts, up to 70 percent less than conventional long-distance rates over the fixed line provided by ChinaTelecom, led to intense competition for the provi-sion of this service. In recognition that it would bedifficult to control this development, MII legalizedVoIP, and incumbents such as China Telecom haveseen their revenues drop by an estimated $130 mil-lion because of the increase in VoIP usage. Thisincrease in competition could come about veryquickly because IP networks can be built anddeployed more rapidly than other networks. TheMII was forced to create a framework to allow theincumbents to provide these services as well. Cur-rently, five of the six telecommunications carriersare licensed to provide VoIP services. Foreign net-work providers and local paging companies havecome together in alliances to seek a share of thelucrative market.11

Competition and Improvements in Quality of Service

Five regulations governing the quality of serviceshave been promulgated by the MII in the last fewyears related to periodic reports about the qualityof service, frequency of publication, and servicequality maintenance, including how to manage


0

10

20

30

40

50

60

70

0

20

40

60

80

100

120

140

160

180

1996 1997 1998 1999 2000 2001 2002

Millions Percent per year

Millions Growth

FIGURE 10.2 Internet Users in China, 1996–2002

Source: Ministry of Information Industry 2003.

complaints from customers. It is difficult to meas-ure whether the limited increase in competition todate has led to improvements in quality. However,interviews do reveal that there has been a change inattitude at China Telecom, which is servicing itscustomers better.

Since 2000, the Telecom Consumer ComplaintsHandling Center under MII has made publicall complaints made against telecom companies.

Notices on complaint handling will be publishedevery three months. Service quality reports submit-ted regularly by the telecom agencies to the MII willbe published every six months. The results of sur-veys and evaluations of customer satisfaction will bepublished annually. All other information relatedto service quality will be published irregularly andaccording to circumstance. Recently, telecom com-panies were warned that if they do not provide good


TABLE 10.3 Impacts of Reforms on China’s Telecommunications Industry, 1998–2003

1998 1999 2001 2003

Backbone optical cable line 180.8 194.1 340(yuan per thousand km)

Fixed asset investment (billions of yuan) 168.1 151.8Telecom services business volume 226.49 313.24 350

(10 million of yuan) (est.)

Price (yuan) June 1Fixed (LAN) line installation (analog) 1,010 725 Free FreeDomestic leased lines (digital, per month)

For telecommunications services 9,000 3,000 1,500For Internet services 4,500 2,000 1,000

International leased lines (per month)Hong Kong, Macao, and Taiwan (China) 27,400 2,800Other Asia 28,900 14,000Other 30,800 15,000

Domestic long-distance fee (per minute) 0.5–1.00 0.7 (depending (flat fee)on range)

Internet connection fee (per minute) 0.02Mobile network access fee 800 500 Free

Users (millions)Fixed-line 87.4 108.8 180.4 263.3

Urban 62.6 74.6 171.3Rural 24.8 34.1 92.0

Mobile phonesa 23.9 43.2 144.8 268.7Fixed-line penetration (per 100 people) 10.5 13 18.9 21.2Computers with Internet access 2 8.9 22.5

CompetitionNumber of players 2 5 7 6b

Mobile 2 2 2 2Fixed-line 2 2 3 3Data 1 2 2 2VoIP 0 1 5 5

Market share of China Mobile (percent) 95 89 75 68Market share of Unicom mobile (percent) 5 11 25 32

aVarious estimates say mobile phone use will hit 300 million by 2005.bAfter the breakup of China Telecom into north and south, with north and Jitong being absorbed byChinaNetCom.Source: Various issues of China Daily and MII.


BOX 10.1 Summary of China’s WTO Commitments

The main commitments are to open up for for-eign entry, which is currently prohibited. Theopening up is undertaken in phases for geo-graphic area and percentage of foreign owner-ship, and agreement to give national treatmentto foreign firms. China also has to abide with theWTO’s April 1996 reference paper on basictelecommunications.

Value-added services (electronic mail, voicemail, online information and database retrieval,electronic data interchange, enhanced facsimileservices, code and protocol conversion, online infor-mation and data); paging services: no limitationson cross-border supply except for right of estab-lishment, no limitations on consumption abroad,and no limitations on national treatment. Onright of establishment, upon accession jointventures with up to 30 percent foreign ownershipwill be allowed to provide services in Beijing,Guangzhou, and Shanghai. Within one year, thisarrangement will be expanded to Chengdu,Chongqing, Dalian, Fuzhou, Hangzhou, Nanjing,Ningbo, Qingdao, Shenyang, Shenzhen, Xiamen,Xian, Taiyuan, and Wuhan. The amount of foreigninvestment is not to exceed 49 percent. Withintwo years, there will be no geographic restriction,and foreign ownership can be up to 50 percent.

Mobile voice and data services (analog/digital/cellular services, personal communicationsservices): no limitation on cross-border supplyexcept under right of establishment, no limita-tions on consumption abroad, and no limitation

on national treatment. On right of establish-ment, upon accession joint ventures will beallowed in Beijing, Guangzhou, and Shanghai,with foreign ownership of up to 25 percent.Within one year, areas will be expanded toinclude services in Chengdu, Chongqing,Dalian, Fuzhou, Hangzhou, Nanjing, Ningbo,Qingdao, Shenyang, Shenzhen, Xiamen, Xian,Taiyuan, and Wuhan, with foreign investmentallowed up to 35 percent. Within three years ofaccession, foreign investment will be no morethan 49 percent, and within five years the geo-graphic restriction will be removed.

Domestic and international services (voice,packet-switched, circuit-switched, facsimile, domes-tic private leased circuit services, internationalclosed user group voice and data services): no limi-tation on cross-border supply except under rightof establishment, no limitations on consumptionabroad, and no limitation on national treatment.On right of establishment, upon accession jointventures will be allowed in Beijing Guangzhou,and Shanghai, with foreign ownership of up to25 percent. Within five years, areas will beexpanded to include services in Chengdu,Chongqing, Dalian, Fuzhou, Hangzhou, Nan-jing, Ningbo, Qingdao, Shenyang, Shenzhen,Xiamen, Xian, Taiyuan, and Wuhan, with foreigninvestment allowed up to 35 percent. Within sixyears of accession, foreign investment will be nomore than 49 percent, and the geographicrestriction will be removed.

service, they will be given a warning by the MII orwill face losing their license altogether. However, theeffectiveness of this mechanism on the quality ofservice remains to be seen.

China’s WTO Commitmentsand Changes in the RegulatoryFramework

A series of reforms and regulatory changes pre-ceded China’s accession to theWTO on December 11,2001, and reforms continue in line with meetingChina’s WTO commitments, which are to be com-pleted over the next five years. However, delays incompletion of the regulatory framework promptedsome of China’s fellow WTO Members to expressconcerns during China’s transitional review (WTO2003).

WTO Commitments

China’s WTO commitments in the telecommunica-tions sector open up the telecom market to foreignparticipation, and bring the domestic regulatoryand business environment in line with interna-tional standards (box 10.1). The commitmentshave two main components. The first is removal of limitations on market access by allowing right of establishment and removal of limitations onnational treatment. Foreign investment is allowedto enter, but initially at a lower ownership level andwith geographic restrictions. Over a period—twoyears for value-added services, five years for mobiletelephony; and six years for domestic and interna-tional services—the geographic restrictions areremoved. However, foreign ownership is capped at50 percent for value-added services and 49 percent

for mobile telephony and domestic and interna-tional services, indicating that a foreign majority isnot desired. The second component requires Chinato adhere to the WTO’s April 1996 reference paperon basic telecommunications, which sets out prin-ciples on the regulatory framework for basic tele-communications services.

September 2001 Regulations

Prior to China’s formal accession to the WTO inDecember 2001, some observers had been antici-pating the introduction of a telecommunicationslaw, because work on one had been under way since1986, with earlier attempts dating back to the early1980s. However, the law was not announced priorto China’s accession and provisional telecommuni-cations regulations were issued in September 2001in line with China’s WTO commitments in the tele-com sector.12 These regulations and some follow-upregulations are a very important first step towarddeveloping a comprehensive and pro-competitiveregulatory framework, but many issues of interpre-tation, clarity, and implementation remain, andsome areas have not been adequately addressed.Moreover, as regulations, they do not have the fulleffectiveness of a true telecommunications law. The

major elements of the regulations are described inthe following sections.

Defining Coverage The working definition usedin the WTO schedule is that basic telecom serviceencompasses local, long-distance, and internationalservices for public and non-public use; may be pro-vided on a facilities basis or by resale; and may beprovided through any means of technology (e.g.,cable, wireless, satellite). In anticipation of thisworking definition, the 2001 regulations definedtelecom services as services provided by means ofcarrying, sending, or receiving sound, data, images,or any other information through a hard-wired orwireless system. This definition is intended to coverbroadcast networks, the Internet, and relatedservices in order to provide the legal basis for the“convergence” of information technologies (Horsley2000). A summary of definition of coverage is pro-vided in box 10.2.

Foreign Investment Foreign investment issuesare not explicitly addressed in the regulations,although a schedule for phasing in foreign owner-ship was set out in China’s protocol of accession tothe WTO. In addition, Regulations for the Adminis-tration of Foreign Invested Telecom Enterprises


BOX 10.2 Coverage and Definition of China’s Telecommunications Sector

Basic telecommunications services are fixed-linelocal and long-distance telephony; mobile tele-communications including analog and digitaltrunk systems, analogue cellular services, second(2G) and third (3G) generation mobile telecom-munication services; satellite communications;Internet, and information transmission; sale orlease of bandwidth, wavelength, fiber optics,cable, or any other network elements; networkaccess and out-sourcing; international telecom-munications infrastructure and internationaltelephony business; wireless paging; and resaleof basic telecom services.

Value-added services refers to services inwhich the supplier adds value to the informationby enhancing its content or providing storageand retrieval. These include nine categories: e-mail, voice mail, online information databaseand retrieval, electronic data exchange, online

data processing and transaction handling,value-added fax, Internet access service, Internetinformation service, and videoconferencing.

An additional five categories were introducedin 2001: fixed-line-based, value-added telecomservices, including telephone information, pag-ing, message recording, and restoring and videoconferencing services; mobile phone value-added telecom services; satellite value-addedtelecom services; Internet value-added telecomservices, including Internet access, data center,information, analog private network, videocon-ferencing imaging, and other Internet-basedservices; and other data transmission andnetwork telecom services, including computerinformation, electronic data exchange, messag-ing, e-mail, fax restoring and forwarding, andanalog private network services.

Source: Telecom Service Classification Catalogue (annex to PRC Telecommunication Regulations, October 1,2000) and Notice Regarding Adjustment of the Telecom Service Classification Catalogue, June 21, 2001.

(FITE Regulations) came into effect on January 1,2002. An additional MII directive was issued in earlyJanuary 2002. The Administrative Measures forTelecommunications Business Operating Permitsdefined changes in registered capital requirements,permit applications, and approvals processes, as wellas operating and annual auditing requirements.

Responding to the Requirements of the Refer-ence Paper The September 2001 regulations alsoincorporated many of the specific requirements oncreating pro-competitive regulations and adminis-tration that are contained in the WTO Agreementon Basic Telecommunications Services. The mainelements of pro-competitive regulations and proce-dures contained in the regulations are as follows.

Competitive safeguards are incorporated to pro-hibit unreasonable cross-subsidization of otherbusinesses, actions that limit subscribers fromusing services of other operators, and the provisionof below-cost services. In addition, an operator isnot allowed to restrict subscribers to using onlyservices or equipment specified by the operator,refuse or delay service without justification,increase rates or items for which it charges feeswithout the approval of the regulator or customer,provide misleading information, or refuse to carryout its commitment to subscribers.

As for licensing, all telecom service providersneed a license and have to meet the transparencyrequirements and the criteria of eligibility for alicense. A time limit has been set for making deci-sions on licensing applications and for making avail-able the reasons for denial. In issuing the license, theissuing agency must consider its impact on nationalsecurity, network safety, the sustainability of tele-com resources, environmental protection, andcompetition. The pending issues here are the imple-menting regulations and details on how the MII andthe provincial regulators will evaluate these factorsand handle issues such as auctions.

The provisions related to domestic interconnec-tion are in line with the reference paper in requiringmajor operators not to refuse other operators’requests for interconnection and the use of non-discriminatory and transparent interconnectionprocedures, including how disputes should beresolved. The operators must also “unbundle” thenetwork elements they sell so that the supplier onlypays for the network components or facilitiesrequired.

Universal service is to be provided in a transpar-ent, nondiscriminatory, and competitively neutralway, and operators should carry out any relevantnational provisions for universal service. The regu-lations refer to some form of subsidy or cost recov-ery without being specific and linking it to thecompetitive safeguards. However the regulationsdid not make clear how universal service will beimplemented.

Resource allocation of scarce resources such asfrequencies are to be done in an objective, transpar-ent, timely, and nondiscriminatory way. The regu-lations in fact go beyond the reference paper byreferring to the establishment of an auction-basedsystem of allocating telecom resources includingfrequencies and network numbers. Leading opera-tors are required to ensure that such users can uti-lize their network resources. However, it is not clearwhat form the implementing regulations will take.

Restructuring of the regulation and administra-tion of the telecom sector in line with the referencepaper requirement to have an independent regula-tor has begun, but it is still unclear what the finaloutcome will be and how long the process will take.The reference paper calls for the regulator to beseparate and not accountable to operators and sup-pliers, and be impartial and nondiscriminatory indealing with all market participants.

In the September 2001 regulations the MIIremained the institution designated “to oversee and control the telecommunications industry ofChina.” Surveillance and control of the sector wereto abide by the following principles: “the separationof administrative departments and enterprises; thebreakdown of monopolies and the encouragementof competition, transparency, fairness and justpractice.” The official position is that the opera-tional and regulatory functions were separatedthrough the creation of the MII in 1998 (WTO2003: para. 78). How objective and impartial MIIcan be in dealing with China Telecom and ChinaMobile remains a source of concern, given closeassociations between the incumbents and MII.

In September 2001 the government set up a newcommission to decide on major policies for thetelecommunications sector and prepare the newlaw. The MII will be in charge of implementingthose rules and administering standards (DowJones Newswire, September 20, 2001). The settingup of the commission is intended to lead to therequired objective changes in policies and laws and


will pave the way for the independent regulatorybody. However, since at present MII is still chargedwith implementing the rules and administrativestandards, one has to wait for the implementingregulations and mechanisms to be in place to assesshow objective the outcomes will be.

Other provisions in the telecom regulationsaddress the provision and quality of telecom serv-ices, network construction, and security. Require-ments of greater transparency imply that there couldbe a greater role for the nongovernment sector—operators, users, and other relevant parties—inproviding input to rates, as well as implementationof regulations, setting of standards, and so on.

Regulations Regarding Internet and E-Commerce13

Given the rapid development of this segment of theindustry and of technology, regulations are stillcatching up with current conditions. Currently,China has over 300 Internet service providers andover 600 Internet content providers. Of the ISPs,only 53 have MII approval to provide nationwideservices, and the remainder are approved by thelocal provincial telecommunications authorities, orPTAs, to provide restricted services. The latest regu-lations require ISPs to obtain permits to do busi-ness and require ICPs to register and record withinthe MII/PTA system currently in place. ISPs mustno longer require foreign investors to find aChinese partner in order to gain approval.

Regulations issued in 2001 on Internet informa-tion services, electronic bulletin board services, andonline news publishing specify registration proce-dures for providers. Even though the rules exist,local authorities have reiterated the rules, and insome cases extended the deadlines, for the neces-sary approvals. In conjunction with the Ministry ofPublic Security, the Ministry of Culture, and theState Administration for Industry and Commerce(SAIC), the MII also issued in 2001 new measuresfor the “Administration of Places that ProvideInternet Access Services”—that is, they governInternet cafes. Those cafes that have already beenapproved by the relevant authorities are nowrequired to apply for reexamination and approval.For example, Internet cafes may not set up within a200-meter radius of primary and middle schoolsand are forbidden to set up in a residential building.

Whereas for E-Commerce the issue of whetherto tax e-commerce has not been resolved, and

China’s State Administration of Taxation has set upa strategy and research group to discuss the issue ofe-commerce taxation and conduct research on thelogistics of an e-commerce tax policy. The group’sfindings will be the basis for new e-commerce regu-lations to meet the requirements of China’s Inter-net economic development, differentiate and stan-dardize business activities on the Internet, enhancethe government’s role in the sector, protect the law-ful rights and interests of enterprises and con-sumers, crack down on illegal business activities,maintain the socioeconomic order, and establish ane-commerce operations registration system.14

Recent Changes in Institutional Structure andPreparation of a New Telecommunications Law

The MII minister was appointed director ofthe State Council Informatization Office in June2003—the first time one person was placed incharge of both MII and SCIO. Some say it is a stepcloser to merging the two agencies to oversee thecountry’s telecom, electronics, and informationtechnology industry and become the new inde-pendent regulatory agency. However, at this stage itis not clear how and when the regulatory agencywill be set up.

Meanwhile, the telecommunications law wasstill being drafted in late 2003 (WTO 2003: para.71); policy reforms and rapid changes in technol-ogy had forced revisions of the drafts before theywere even issued. There have been various updatesand clarifications since the September 2001 regula-tions and they are going in the right direction. Thisincludes greater transparency such as specificationof deadlines for approval of licenses in basic tele-com (180 days) and value added services (60 days)and requiring the authorities to provide an expla-nation for decisions. In August 2003 the StateCouncil announced its “opinion” on the broadframework for telecommunication market admin-istration as part of the results of the commission setup to look into telcom regulations.15 This includedstrengthening and facilitation interconnection pol-icy with fees to be based on costs incurred by theoperator, and sanctions and penalties on violationof interconnection. This issue is important in lightof the disagreements on interconnection pricingbetween mobile and fixed line networks. The scopeof the telecommunication law confirms that theauthorities have convergence in mind, that so, that



China Telecom

China Telecom

Unicom

Jitong

China Mobile

China Satellite

Guoxin Paging

China Netcom

China Railcom

CT (all fixed-line, south)

CSAT (satellite comm.)

Unicom (all, CDMA)

CNC (Internet, fixed-line, north)

CRC (fixed-line, long-dist.,Internet)

CMCC (GSM, long-dist. IP)

MII

Today200219981994 2000

Source: Compiled by authors.

FIGURE 10.3 Restructuring of China’s Telecommunications Industry,1994–2003

the framework law is intended to include opera-tions of telephony, internet, and cable TV indicat-ing that all of the subsectors will come under oneframework law. However, even though the organi-zation and administration of the new regulatoryauthority was announced in broad terms, it was notclarified how the responsibilities between MII andSARFT will be reallocated. Other issues addressedwith greater transparency and market based princi-ples in mind, were implementing regulations formarket entry, distribution and administration ofscare resources (e.g. frequencies), service fees, qual-ity standards, and information security.

Restructuring: Government-Driven Mergers and Consolidation

Other than regulatory changes, the authorities havealso restructured the sector through governmentrather than market driven mergers and consolida-tions. In 2002 China Telecom was split alongregional lines. Operations covering the northern 10provinces were merged with China Netcom and

Jitong to form the China Netcom CommunicationsGroup Corporation. The 21 southern operationsretained the China Telecom name and operate as aseparate entity. Figure 10.3 shows the restructuringefforts since 1994 up to the last change.

Corporatization efforts to improve the perform-ance of the sector and raise capital continue. InNovember 2002 China Telecom listed the fourprofitable networks in Guandong, Jiangsu,Shanghai, and Zhejiang in the Hong Kong and New York capital markets, raising $1.5 billion.Meanwhile, China Netcom is still undertakinginternal reorganization in merging the three sepa-rate entities (China Telecom north, Jitong, and theoriginal China Netcom), and, as a result, its plans tolist shares on the capital markets have been delayed,with a public offering now planned for 2004.

The Potential Direct Impact ofReforms and Outstanding Issues

What is the expected impact of the current set ofreforms and WTO accession commitments on the


telecommunications sector? First, the sector will bechallenged to be competitive and efficient beforethe likely entry of foreign investment stemmingfrom China’s WTO commitments, and it will faceincreased pressure to meet customer satisfaction, aswell cope with the restructuring that is under waydomestically. Second, the industry has to respondto rapid changes in technology. Meeting these chal-lenges will require further restructuring of the sec-tor, new investments from both the governmentand private sector, and the transfer and develop-ment of the appropriate technology and infrastruc-ture (hard and soft). The evolution of the sectorshould be managed carefully, because there is thedanger of over investment and over capacity, whichwill be influenced by the regulatory environmentfacing the sector. The sections below address thepotential competition and impact on the differentsubsectors, and issues related to developing theappropriate regulatory environment.

The Potential for New Entrants and Managing Competition

It is unlikely that China would have opened up itstelecommunications sector to foreign entry with-out the pressure it faced in bargaining for WTOaccession (Mueller and Lovelock 2000; Zhang2001). Now, even though the schedule allows newentry and permits foreign investors to come in overthe next few years under the WTO commitments, itis unlikely that a flood of new foreign entrants willmaterialize. Because of the large size of the invest-ment and the higher risks of an uncertain regula-tory environment, foreign interest will be limitedunless there is a credible commitment from theregulators to implement the planned regulations.In any event, foreign majority ownership will notbe allowed in key areas, and foreign ownership iscapped at 49–50 percent, with entry still based onobtaining a license to operate.

Interviews with foreign telecom operators andindustry people conducted from mid-2001 to mid-2002 indicate that capping foreign ownership wasnot a major disincentive, but the lack of specificityon the regulations for foreign investors was. Animportant barrier to entry to fixed-line operators isthe dominance of the incumbent, China Telecom;the lack of certainty surrounding interconnectionto its backbone infrastructure; and the drastic

decline in prices and the fall in monopoly rentswith the commoditization16 of fixed-line use. Theonly area in which prices could decline further isinternational rates, which will face increased pres-sures to fall given the advent of VoIP. Perhaps amore important concern is that most foreign tele-com operators are currently not in a position toundertake huge investments in China.

Thus it is unlikely that new entrants will enterthe fixed-line subsector, except possibly in segmentsrequiring low capital requirements. Increasedcompetition in fixed-line is unlikely to come fromliberalization or privatization, because the majorityownership is expected to remain with the state.Rather, the priority will be on creating the regula-tory framework, including the institution, to createthe pro-competitive operating environment andcomplete the corporatization process so that ChinaTelecom and China Netcom will operate more effi-ciently. The government-driven process of mergersand consolidations is likely to continue to be themodus operandi.

As for mobile phone services, the rapid progressof technology in this subsector, combined withdeclining costs and an infrastructure cost lowerthan that for fixed-line telephony, make it attractiveto investors. Yet toppling the major barrier to entry,obtaining a license, remains elusive. Licenses tooperate mobile services, as with any other basicservice, are issued at the discretion of the centralauthorities, and although the procedures forgreater transparency in the issuance of licenseswere laid out in the September 2001 regulations,there is no indication when new licenses will beopened. Additional licensing to operate mobileservices has so far been linked to technology such as3G and standards such as those represented byUNICOM’s CDMA, although a decision on thetechnology standard has yet to be made.

The mobile telephone market is still a duopolywith very little competition in this subsector.Backdoor attempts at entering the market by, forexample, mobiles that operate from the fixed-linenetwork within limited areas have been haltedindefinitely.17 Moreover, because the applicationsfor value-added services that use the latest technol-ogy do not yet exist in China, technology alone willnot provide enough of a catalyst for users to switch.Another problem for competition is that as yetthere is no number portability—that is, users cannot


switch services without switching their numbers.Therefore, the technology alone does not provideenough of an incentive to switch services. Theattractiveness of mobile telephony is also reducedbecause roaming charges are based on the caller payparty principle.

The main challenge facing the mobile industry inChina—as everywhere—is the commoditization ofcall fees as the average revenue per user continues todecline. To find new revenue streams, mobile opera-tors need to seek out new partnerships with value-added service providers. One model is the coopera-tion between Unicom and ICPs (e.g., Montro).Another value-added service that is booming is theshort messaging service (SMS). Foreign and newentrants are likely to take part in the value-addedand specialized services segments of this subsector.

Another issue is what will happen to the existinginternal networks that developed in response to theinefficiencies and high cost of using China Tele-com? Many ministries such as the State PowerCommission and the Ministry of Railways builttheir own internal networks. One of the importantgovernment-supported networks is the one devel-oped by the agency that administers the state radio,SARFT. The network already has 90 million usersand is planning to connect all existing networks.

Finally, the government also plans to upgradethe CATV network so that it can also transmit

Internet data. The first broadband services systemusing television cable was recently introduced inDaqing City in Heilongjiang Province. The openingof the system on CATV represents a major break-through in the country’s broadband progress. Thebenefits to the industry could be substantial, if thegovernment can come to a final decision on howfast it wants convergence to take place.

Development of Internet and the New Economy

Although certain barriers such as low credit cardpenetration and an underdeveloped distributioninfrastructure will exist for some time in China,Internet usage and everything associated with it,including e-commerce, are expected to continue togrow rapidly. Although foreign companies will beable to enter this market to provide value-addedservices, it seems likely that partnerships betweenthe two will become the order of the day. The foreigncompanies will provide the technology, brand-namereputation, and skills, and the Chinese companieswill provide the local connections, language abili-ties, and legal foundations for the new venture’sexistence. Again, however, as the recent changes inregulations have demonstrated, the lack of regula-tory certainty, including on cross-border taxation,security, and content, have to be resolved before thesubsector can really take off.

FIGURE 10.4 Share of Revenue by Operator,2001 and 2002

0

10

20

30

40

50

60

Percent

2001 2002

CT CNC Unicom CMCC Railcom CSAT

Source: Ministry of Information Industry.

Overall, then, China does not yet have the infra-structure needed to take advantage of the new econ-omy, and the chief bottlenecks to such developmentare less in the technology than in finance and logis-tics. Much needs to be done in terms of developingthese sectors of China’s old economy before it canleapfrog its way into the new global economy. Thisconstraint also points to the need for a comprehen-sive approach in developing the services sector. Cur-rently, the only real possible use of new economysystems in China rests with the foreign multination-als that have already implemented various supplychain management systems, customer relationsmanagement systems, and other e-solutions systemsthroughout their global operations. In many cases,China continues to be the one country in which theiroperations cannot be fully implemented.

Managing Policy Reforms and IncreasedCompetition

The discussion in this chapter thus far clearlypoints to the importance of creating an appropriateregulatory environment to attain the emergence ofhealthy competition and an efficient telecommuni-cations sector. Some of the key regulatory issuesrelated to policies on convergence, government-mandated restructuring, consolidation of the state-owned enterprises, and the creation ofpro-competitive regulations are described in thissection.

Cable and Convergence Issues China’s cable tel-evision network now reaches about 90 million sub-scribers throughout the country, and it is expectedto eventually be the largest in the world. Majorrecent regulatory reforms have enabled fixed-lineoperators such as China Telecom, Unicom, CNCC,and Railcom to offer cable television over theirbroadband networks. Cable TV companies werealso looking forward to being able to deliver Inter-net and voice services over their networks. How-ever, until there is an implementation schedule onconvergence, explicitly indicating what is or is notpermissible for cable operators, the cable industryremains reluctant to see convergence because itfears broadband would significantly erode its cur-rent subscriber base.

The MII has remained generally unsupportiveof convergence. Regulations introduced in 2002

required joint ventures between existing telecomoperators and new partners to seek new telecomlicenses and to adhere strictly to the business cate-gory and geographic area of operation specified bythe telecom license already in hand.

Toward a Pro-Competitive Regulatory Frame-work and Institutions Given the nature of anddevelopments in the telecommunications sector, itis unlikely that increased competition will comefrom liberalization or privatization. In any case,studies have shown that privatization will flowfrom the introduction of competition, but that thereverse is not necessarily true (Petrazinni 1996;Fink, Mattooo, and Rathindran 2001). In China,inefficiency has stemmed not from governmentownership per se, but from a lack of competitionand ineffective government regulation (Stiglitz1998). Given that the major operators in basic tele-com are state-owned and that the situation isunlikely to change even in the short term despitethe liberalization of the sector through WTO acces-sion as already outlined, the key challenge forChina’s telecom sector is to ensure that the few andpredominantly state-owned players will be compet-ing in a pro-competitive environment.

The September 2001 regulations go some waytoward responding to some of the obligations forpro-competitive regulations and institutions as setout in the WTO agreement on basic telecommuni-cations, but there are still areas of uncertainty,issues of implementation, and lack of a timetablefor implementation. Given the current politicalinstitutions and structure in China, it is unrealisticto expect full regulatory independence in the nearterm. Not only does the Communist Party haveoverall responsibility for policy, but the MII still hasa strong and complex relationship with the majoroperators. This situation gives rise to specific andmajor challenges to the creation of a truly inde-pendent regulator (Zhang 2001: 47). If regulatoryreform continues to proceed slowly, the impactof WTO accession on liberalization of China’stelecommunications market could end up beinglimited.

The problems and the issues surrounding thecreation of an independent regulatory agency arenot just confined to China, but also are beingreplayed in most developing countries attemptingto open up their telecom sector. Therefore, careful



thought should be given to the sequence of stepstaken in developing such an institution. The firststeps should be adopting transparency and traininghuman resources adequately, as well as settingadministrative guidelines as to just what constitutesan “independent” agency. These measures could befollowed by a step-by-step phase-in of a more inde-pendent regulatory agency.

Provision of Universal Service to Poor and Isolated Areas18

As in other countries, before reforms China Tele-com was the only operator in China and had anobligation to provide universal service. The modeladopted for China Telecom was the traditionalcross-subsidization model in which subsidizationflowed from the eastern and more developedcoastal areas to the western provinces of China andfrom wireless and long-distance services to localservices. The need to provide all households andsmall firms with telecommunications was used as areason to defend monopolistic telecom companies.Furthermore, it was believed that profit-maximizingprivate firms would not go to poor and remoteareas. However, experience with protected statemonopolies has shown that the outcome is likely tobe one of inefficiency and underinvestment, withlittle or no service actually provided to the poorareas. Indeed, many developing countries have paidcapital costs of $4,000 per line, or three to four timesthe achievable cost (Stiglitz 1998). Evidence alsosuggests that liberalization in the form of greatercompetition or privatization will generally increasenetwork availability (Petrazinni 1996, p. 37).

With the restructuring of the China’s telecomsector, the authorities are shifting from the tradi-tional cross-subsidization model to the provision ofuniversal service by popularizing access throughsignificant price reductions. By means of suchactions, it is hoped that telecom services will gradu-ally become more accessible, despite the lack of anexplicit universality mandate. The growth rates offixed-line and mobile phone users have indeedbeen dramatic, leading to penetration rates ofabout 20 percent and 21 percent of the population,respectively, based on 2003 estimates. This averageis, however, misleading, because a much lower rateof teledensity—only about 6 percent—prevails inrural areas and western provinces (MII statistics).Furthermore, in large areas of the country the basic

infrastructure for telecommunications connectivitystill does not exist, and mobile stations are limitedto serving only the large cities.

The objective of the universal service obligationis to ensure that the public has access to affordabletelecommunications services, including people liv-ing in high-cost service areas such as rural andremote regions as well as low-income groups. Theauthorities have not yet assigned the delivery ofuniversal service to any of the current operators,although the expectation is that China Telecom willremain the major provider for the less developedareas and the poor regions.

The WTO reference paper on basic telecommu-nications services does not provide guidelines onuniversal service, although it does make clear that itshould be implemented in an anticompetitive fash-ion, that it should be administered in a transparent,non-discriminatory, and competitively neutralmanner, and that it should be no more burdensomethan necessary for the kind of universal servicedefined by the WTO Member.

The experience of and lessons learned by othercountries in implementing universal service pro-grams have yielded some caveats and models thatChina could consider (World Bank 2000, Box 10.3).First, it is important that the definition of the uni-versality target is clear and that it is linked to realis-tic implementation measures, including how itshould be funded. The universality target differsfrom country to country, depending on local eco-nomic and sector conditions, the distribution ofthe population, geographic considerations, andsecurity issues. Universal service refers to individualor private access, whereas universal access refers tocommunity or public (shared) access (e.g., a phonebooth in every village), and in developing countriesthe latter is more important.

There are several common models for achievinguniversal access: applying market-based reformsthrough privatization, competition, and cost-basedpricing; requiring operators, including new entrants,to assume mandatory service obligations; imple-menting cross-subsidies between and within servicesprovided by incumbent operators; using accessdeficit charges paid by telecommunications opera-tors to subsidize the access deficit of incumbents; andcreation of a fund to provide for universal service(World Bank 2000). Other models and combina-tions of these approaches are available, but the choicedepends on country-specific issues and conditions.


Cross-subsidization has traditionally been usedwhen a state has one dominant provider. The disad-vantages, however, are by now well known, such asinefficiency and not reaching the target group. Chinahas used this model in the past, and so now mustrethink how to deliver universal access.

The MII plans to set up a universal service fundthat will be used for developing the western and

rural regions, which currently lag far behind theireastern coastal counterparts. The fund is intendedto separate government administration from enter-prise management in the sector, shifting responsibil-ity for infrastructure, operations, and fund-raisingto the telecom companies themselves. Because thecosts of these activities are high and have lowreturns, the MII is proposing tax incentives and

BOX 10.3 Experiences of Chile, Peru, South Africa, and Hong Kong (China) with Universal Access

Chile

After market opening and privatization of itstelecommunications sector, Chile introducedpolicy intended especially to provide universalaccess to people living in low-income and ruralareas. The supporting funds, which came fromthe central government budget, took the formof targeted subsidies to private operators for theinstallation of public telephones in the unservedareas. The process included competitive biddingand the funds were not linked to the rollout obli-gations of operators to provide services in cer-tain areas Over the period 1993–99, a total of183 projects to serve almost 6,000 localitieswere approved. Competition between biddersreduced the actual subsidies paid, althoughthere were some delays in installation. The tar-gets are now to finance community Telecentersthat have access to the Internet and other newinformation and communications technology.

Peru

Peru imposed rollout obligations as part of priva-tization and established universal accesstelecommunications funds, whose revenueswere taken from the 1 percent of the gross rev-enues of the telecommunications sector. Thefunds were used to meet the target of extendingservices to 5,000 unserved localities by the year2003. Universal access was defined as servicesprovided by public operators and available to amajority of users, including voice telephony,low-speed fax and data, and free emergencycalls. Criteria for the selection of localities wereset as rural towns (300–400,000 population),district capitals, and towns in high social interestareas. Several innovations were introduced inthe bidding process such as, for economies ofscale, allowing a single operator to bid simulta-

neously on three projects and permitting bid-ders to bid on any combination of three proj-ects. In 2000 Peru modified its target to fundaccess to the Internet and other advanced serv-ices, and to allow for funding the operation andmaintenance of designated services, not justinstallation.

South Africa

In South Africa, rollout obligations are imposedon operators in underserved areas and com-bined with a universal service fund. The fund isgenerated by annual contributions from alltelecommunications licensees, and is used toprovide direct subsidies to a targeted prioritypopulation to subsidize the higher costs oftelecommunications and to subsidize the cost ofnetwork rollout to underserved areas by opera-tors. The network rollout, service quality targets,and pricing are closely monitored by the respon-sible agency. The funds have been used to assignpriority to the setting up of telecenters in part-nership with communities and donor agencies.The telecenters typically contain telephones, faxand photocopy machines, personal computers,and access to the Internet.

Hong Kong (China)

In Hong Kong, the cost-based universal serviceregime is funded through charges on external(i.e., international) traffic. The designated serv-ice provider has the obligation to provide PSTN(public switched telephone network) accessservice in Hong Kong and receives a fair contri-bution from other licensees toward the net costof serving customers and providing public tele-phones. In 1999 the funds accounted for about1 percent of total sector revenues.

Source: Kerf, Schiffler, and Torres 2001; Wellenius 1997.


investment, financing, and human resource policiesfor the industry, in tandem with the continuedopening of pricing issues toward full market forcemanagement. However, the plan is still under discus-sion. In February 2003 the government announcedthat it would reinstate the universal service fundwith would require contribution from all domesticoperators to subsidize the construction of telecominfrastructure in China’s rural areas. However thedetails on the size of contributions, how they will bedisbursed and managed remains unclear.19

The approach of using universal funds has theadvantage of being most effective in providing tar-geted subsidies to expand or support uneconomicservices, such as the basic telecommunicationsservices needed in the underdeveloped westernprovinces and rural areas. Indeed, it is potentiallythe most efficient and transparent approach, butmuch depends on how it is structured and admin-istered. The scheme can be administratively com-plex, so that the transaction cost are higher than the subsidy, and there is the potential for badgovernance if accountability and monitoringmechanisms are not built in and implemented.Furthermore, it is often difficult to predict theassociated costs and revenues.

Experience has produced important lessons foranyone using the universal funds approach to pro-vide universal access. First, the target must beclearly specified (e.g., high-cost regions such as thewestern provinces of China, rural areas without any access, a low-income population). Second, thefinancing, be it from direct government funding,contributions from operators, or proceeds fromprivatization, must be transparent. Third, the bodythat administers and disburses the funds must beindependent from the operators. It also should bemarket-neutral in that it does not favor the incum-bent vis-à-vis new entrants, and it must seek com-petitive bidding. Fourth, the subsidies given (i.e., inthe form of a fiscal or investment incentive) shouldcover only the uneconomic portion of the cost ofproviding the services, and the operators, whetherprivate or government, should finance the rest ofthe cost from their revenue.

Conclusions

In recent years the view of the telecommunicationssector has changed—from that of a natural

monopoly with large fixed costs to that of a set ofmore complex and competitive activities. Newtechnologies have lowered unit costs dramatically.Since the mid-1990s, there has been a tendencytoward a convergence of telecommunications,computing, and broadcasting. New services, such asthe Internet, mobile, and wireless telecom, usedifferent platforms and transmit through variousmedia, and they provide alternatives to traditionalwire line telephony at a much lower cost. Theseservices have developed beyond regulators’ controland can bypass the public switched network andcontrol of the incumbent. These changes have alsomeant moves away from traditional forms of own-ership and the market structure of state monopolyto one of joint ventures, strategic alliances, and newand foreign entrants (Petrazzini 1996, p. 24; Intven2000). In addition the justification of maintainingstate monopoly in telecom and undertaking crosssubsidization has lost ground.

China’s reforms in the last few years have beenquite dramatic compared with the situation inother countries, and currently the telecom servicessector in China is moving in this direction also.Increased competition has occurred, although ithas been “managed” by the government throughintroducing new state owned players, breaking upthe state monopoly and restructuring as well asconsolidation efforts, and through governmentdriven price reductions. Increased competition hasin turn led to greater efficiency, price reductionsbeyond the government mandates, competitionthrough better service and product differentiation,and greater penetration rates. Even though manychallenges lie ahead, considering the dramaticchanges that the sector has witnessed since 1998and the unstoppable pace of global technologicalchange, it seems reasonable to expect that the sectorwill continue to liberalize.

Whilst a number of foreign players have enteredand will continue to enter value added services, asalready discussed above, even after basic telecomsector is opened up under China’s commitments,competition will remain between domestic (whichwill remain largely state owned) companies. There-fore to reach the desired level of fair competitionand efficiency, and attract new investments, muchdepends, however, on the ability of the authorities tocome up with a solid legislative framework that fos-ters a stable business environment by ensuring and

enforcing a level industry playing field and privateproperty rights, and how the state owned companieswill be “corporatized” to be managed in a commer-cially viable manner. In 1999 the MII issued a state-ment that a telecom law would be finished prior toaccession, but in December 2003 this law was stillbeing drafted. Given that even the United States,considered to be the most wired nation, continuesto have considerable difficulty with creating andupdating regulations on various subsectors of thetelecom industry, it is perhaps not surprising thatthe process of drafting this legislation has been sotime-consuming. A number of issues related to thenew telecommunications law such as interconnec-tion policy, the creation of an independent regula-tory authority and convergence appear to be movingin the right direction. However, a number of otherissues still need to be clarified:

• Regulate and enforce a level playing field for pri-vate and foreign investment. Although hedgingagainst risk is an inherent part of good businesspractice, the current environment creates a majordisincentive for investment of the foreign andprivate funds needed to develop the industry.

• Consider implementation of less interventionistand government driven price mechanism, such asintroducing price caps, that incorporate qualityand other self-regulating factors. Current pricesreflect an effort to broaden the use of telecom-munications services, and significant sector andgeographic rebalancing will be required. But theywill affect cross-subsidization and will not reflecttrue costs.

• Decide, like all other countries, which 3G tech-nology standard to adopt. The debate overwhich 3G technology to use contributes to theenvironment of uncertainty and has the poten-tial to hold back investments.

• Adopt a clear and concise, but flexible, frame-work for the regulation of value-added services,including e- and m-commerce platforms.

• Devise anticollusion mechanisms given the like-lihood of mergers between the current majormarket players. Such mechanisms become par-ticularly important once the MII is relieved of itscontrol over the industry and pricing issues areleft to market forces.

• Establish an independent regulatory agency.Given past experience, it is likely that this will

take some time and the issue of how “independ-ent” the regulator will be from the domesticoperators, if the regulator emerges out of MII,will continue. However, it is important thatChina addresses these issues by beginning withthe introduction of greater transparency. Forinstance, by establishing a clear and transparentscheme within which licenses can be allocated.The deadlines provided for consideration oflicense approval are welcome, but more detailsare needed with regard to the criteria for eligibil-ity of licenses.

Other than the regulatory framework andimplementation thereof to ensure greater trans-parency and a level playing field, another majorchallenges facing the sector of providing universalservice. China is already moving in the right direc-tion by recognizing that separate policies areneeded to address this social objective and notbeing burdened on the objective of developing anefficient and competitive telcom sector. Chinashould take note of the experience of other coun-tries, and choose what will be appropriate for itsconditions. Given China’s size and vast differencesbetween provinces, more than one model is likelyto work depending on the existing conditions andgeographical location in question. However, theprinciples and lessons from other countries shouldbe borne in mind—especially with respect to itbeing market based, transparent and with the rightincentive structure—when designing the policy.

Appendix: Summary of WTOReference Paper on Basic Communications (Annex to FourthProtocol of General Agreement onTrade in Services [GATS], Agreementon Basic Telecommunications,Effective January 1, 1998)

China’s WTO accession commitments in telecom-munications oblige it to also adhere to the princi-ples of the WTO reference paper. The paper pro-vides for a pro-competitive environment of policyand regulation.

Competitive Safeguards

Prevention of Anticompetitive Practices in Tele-communications Appropriate measures shall bemaintained for the purpose of preventing suppliers


who, alone or together, are a major supplier fromengaging in or continuing anticompetitive practices.

Safeguards Anticompetitive practices include, inparticular, engaging in anticompetitive cross-subsidization, using information obtained fromcompetitors with anticompetitive results, and notmaking available to other services suppliers on atimely basis technical information about essentialfacilities and commercially relevant informationthat are necessary for them to provide services.

Interconnection

Interconnection is to be provided under nondis-criminatory terms, conditions, and rates, and in atimely fashion, and, upon request, at points inaddition to the network termination points offeredto the majority of uses. The procedures applicablefor interconnection to a major supplier will bemade publicly available, and a major supplier willmake publicly available either its interconnectionagreements or a reference interconnection offer. Anindependent domestic body will resolve intercon-nection disputes.

Universal Service

Any Member has a right to define the kind ofuniversal service obligation it wishes to maintain.Such obligations would not be considered anti-competitive as long as they are administered in atransparent, nondiscriminatory, and competitivelyneutral manner.

Licensing Criteria

The following should be made publicly available:all licensing criteria and the period of time requiredto reach a decision on a license application, theterms and conditions of individual licenses, and thereasons for denial of a license.

Independent Regulators

The regulatory body should be separate from andnot accountable to any supplier of basic telecom-munications services. The decisions of and proce-dures used by regulators shall be impartial to allmarket participants.

Allocation and Use of Scarce Resources

Allocation of scarce resources such as frequencies,numbers, and rights of way will be carried outin an objective, timely, transparent, and non-discriminatory way. The current state of allocatedfrequency bands will be made publicly available.

Notes

1. For the United States, European Union, and Japan, open-ing up telecommunications services was a major point of con-tention during their bilateral negotiations with China. China’srejection of foreign direct investment (FDI) in telecommunica-tions services was unusual when compared with its stance inother economic sectors or with that taken by other countries at asimilar level of development.

2. About 40 foreign operators lost up to US$1.4 billion whenthe government cracked down on investments that circum-vented the FDI ban in the sector. Despite this, the U.S. companyAT&T recently entered into a joint venture to form ShanghaiSymphony Telecommunications, and all other major foreignoperators remain in China, awaiting further entry opportunities.

3. A decline in average revenue per unit without increasingusage.

4. See Gao, Pin, and Kalle (2000); Perkins (2001), PriceWaterhouse Coopers (2001), World Bank (1992, 2000) andZhang (2000) for greater detail and background.

5. The lack of cooperation is not surprising, because Unicomwas created by the rival Ministry of Electronics and an official inthe State Council—not by MPT. Unicom’s original objective wasto share 10 percent of the fixed-line market and 30 percent of themobile phone market with China Telecom by 2000 (Asian Com-munications, September 1998).

6. Creation of the agency was part of the institutional reor-ganization undertaken at the Ninth National People’s Congress.

7. In 1998 the government forced the People’s LiberationArmy (PLA) to divest itself of all of its business interests as partof a government crackdown on corruption and forced a han-dover of Great Wall Telecom Corporation, which had the rightto operate a CDMA system, to Unicom.

8. As reported in China Daily, 5 August 2001.9. The Chinese Internet Network Information Center

(CNNIC) semiannual survey defines Internet users in China asChinese citizens who use the Internet at least one hour a week.

10. ADSL employs the existing double twisted copper tele-phone lines to provide broadband network access, enabling adownstream rate of 8 megabits per second and an upstream rateof 760 kilobits per second. However, transmission quality can bemaintained only within a distance of 5 kilometers, thus requir-ing one base site every 5 kilometers, and with heavy initialinvestment by telecom operators in network construction. SinceADSL was first introduced, charges have dropped from about450–500 yuan to about 130 yuan per month.

11. Companies such as AT&T and Edge2net Inc. have alreadyentered the market. The latter has teamed up with China MotionTelecom to create Shenzhen China Motion Telecom.

12. The regulations have since been updated with a revisedcatalogue of value-added services in June 2001.

13. See Baker and McKenzie (2000); Liu He, Qin Hai, and Lu Yanrong (2001) for more details.

14. New regulations were announced in early 2003 that cov-ered the legal definition of E-Commerce as trading, production,


and services provided over the electronic network (e.g. informa-tion search, ordering, and payment) and authentication ofe-signatures. They did not as yet discuss the issue of taxation.

15. State Council, Opinions Regarding Further Strengthen-ing the Supervision and Administration of Telecom Market,August 2003.

16. A decline in average revenue per unit without anadequate increase in minutes of usage.

17. After essentially being cut out of the mobile market withits divestiture of CMCC, China Telecom tried to enter themobile business by introducing a mobile phone that functionedessentially as an extension of its fixed-line networks. Therefore,the mobiles only functioned within a limited geographic areaaround the line. Nevertheless, 2.5 million lines in about 50 citieswere already installed when the MII, in response to the lobbyingby two licensed mobile operators, ordered China Telecom tocease such operations. The attraction for users was that handsetscost about half the price of normal handsets.

18. See Petrazinni (1996); Smith (1997a, 1997b); Stiglitz(1999); and Wallenius (1997) for more information.

19. It was recently announced under the Rural Communica-tion and Universal Services Initiative that, by 2005, 40,000 vil-lages will be reached and each of the six major telecom operatorswill be asked to cover certain regions: China Telecom (Shanxi);China Netcom (Inner Mongolia autonomous region); ChinaMobile (Sichuan); Unicom (Guanxi autonomous region); ChinaSatcom (Sichuan); and China Railcom (Henan).

References

Baker and McKenzie. 2000. Asia Information Associates,E-Commerce in China.

McKinsey Quarterly. China Statistics Yearbooks 1998–2000.Beijing, China Statistics Press.

Fink, Carsten, Aaditya Mattoo, and Randeep Rathindran. 2001.“Liberalizing Basic Telecommunications: the Asian Experi-ence.”Paper presented at the Conference on Trade, Investmentand Competition Policy in the Global Economy; the case ofthe International Telecom Regime. Germany, January.

Gao, Pin, and Lyytinen, Kalle. 2000. “Transformation of China’sTelecommunication Sector: A Macro-perspective.” Telecom-munications Policy 24: 719–30.

Intven, Hank, ed., “Telecommunications Regulation Hand-book.” 2000. Infodev, World Bank.

Klein, Michael, and Philip Gray. “Competition in NetworkIndustries—Where and How to Introduce It.” Public Policyfor the Private Sector, World Bank.

Liu He, Qin Hai, and Lu Yanrong. 2001.“E-Commerce in China”(First Draft).

Mueller, Milton, and Peter Lovelock. 2000. “The WTO andChina’s Ban on Foreign Investment in the Telecommunica-tions Sector: A Game-theoretic Analysis.” Telecommunica-tions Policy 24: 731–59.

Perkins, Tony, and Steven Shaw. 2000. “China and the WTO:What Will Really Change?” McKinsey Quarterly no. 2.

Petrazinni, Ben A. 1996. “Competition in Telecoms—Implications for Universal Service and Employment.”Public Policy for the Private Sector. World Bank Group,Washington, D.C., October.

Price Waterhouse Coopers. 2001. “Briefing Overview of ChinaTelecommunications Sector,” February.

Qiang, C.Z.W., and Xu, L.C. 2000. “Reforming China Telecom,”World Bank DECRG.

Smith, Peter. 1997. “What the Transformation of TelecomMarkets Means for Regulation.” Public Policy for the PrivateSector. World Bank Group, Washington, D.C., July.

Smith, Warrick. 1997. “Utility Regulators—The IndependenceDebate.” Public Policy for the Private Sector. World BankGroup, Washington, D.C. October.

Stiglitz, Joseph. 1998. “Creating Competition in Telecommuni-cations.” Paper presented at World Bank Conferenceon Managing the Telecommunications Sector Post-Privatization, Washington, D.C., April.

Stiglitz, Joseph. 1999. “Promoting Competition and RegulatoryPolicy: With Examples from Network Industries.”Washington, D.C., World Bank, Beijing, July.

Wellenius, Bjorn. 1997. “Extending Telecommunications Ser-vice to Rural Areas—The Chilean Experience.” PublicPolicy for the Private Sector. World Bank, Washington,D.C., February.

World Bank. 1992. “People’s Republic of China Telecommunica-tions Sector Study: Survey, Assessment and Strategy Recom-mendations.” Report No. 9413-CHA, February.

World Bank. 2000. “China: Services Sector Development andCompetitiveness.” Washington, D.C., December.

World Bank. 2001. “Seizing the 21st Century: Using Knowledgefor China’s Development.” Washington, D.C., June.

WTO (World Trade Organization). 2003. “Report of the Meet-ing held on 5 December 2003: Note by the Secretariat.”Council for Trade in Services. WTO, Geneva.

Zhang, Bing. 2001. “Assessing WTO Agreements on China’sTelecommunications Regulatory Reform and IndustrialLiberalization.” TPRC.


a joint venture with one of the largest foreign lifeinsurance firms, which led to a further round of newassociations between Chinese financial enterprisegroups and foreign insurance companies. A similarpattern of mergers, acquisitions, and partnershipshas emerged in investment banking, funds manage-ment, securities houses, and housing finance. Inanother development, a leading foreign manufac-turing firm entered into an innovative agreementwith a wholly owned foreign bank for loans againstits receivables and promptly used the funds to payoff outstanding credits to two large local banks. Tocap it all, data available during the first quarter of2002 suggested that a credit crunch had occurred inthe second half of 2001, as banks tightened lendingin order to improve their portfolios to meet thepost–WTO accession challenge.

By the standards of most financial markets, theseevents were unremarkable, either singly or together.However, to observers of China’s financial market itwas clear that something momentous was happen-ing; the genie was out of the bottle, competition wasabout to break out! Members of China’s moribundfinancial community, especially bankers, saw theseevents as the first signs of the imminent destructionof the domestic financial system, justifying the fearsof protectionists who had argued against many ofthe liberal measures included in China’s accession

Close to the date of China’s entry into the WorldTrade Organization (WTO) in December 2001,several events appeared to perk up the domesticfinancial market. A leading international bankannounced that it would charge a small monthlyfee if the amount in an individual bank deposit fell below US$5,000.1 A leading local bank thenpointed out that about 20 percent of its savingsdeposits were in the amount of 100 yuan or less andhinted that this situation was uneconomical. TheChina Bank Association then wrote a detailedproposal that reportedly was discussed with thePeople’s Bank of China (PBC)—the central bank—before submission for approval to the State Devel-opment Planning Commission. The proposal rec-ommended that banks be allowed to levy fees forabout 40 different services that traditionally theyhad provided free of charge.

A little earlier, fierce competition had broken outamong Shanghai banks when they slashed buy/sell margins on foreign currency transactions to10 basis points (from an average of 40 basispoints in 2001), but order was quickly restoredby the Shanghai Banking Association. In paralleldevelopments, a foreign-invested bank and a fullyowned foreign bank received their first Chinese cus-tomers. Meanwhile, a major local financial con-glomerate opened a new life insurance company in

181

11

AcceleratingFinancial MarketRestructuring in

China

Deepak Bhattasali

agreement (see table 11.2). Against this backdrop itis interesting to assess the following issues: How didthe accession agreement threaten the status quo inthe domestic banking market?2 If the policymakerswere aware of the threat, why did they agree to takesuch radical measures to permit the entry of foreignbanks? What are some likely scenarios for the evo-lution of the banking sector in China during theperiod of accession and beyond?

The next section of this assessment describes thestructure and behavior of the Chinese bankingindustry at the time that the final WTO-related nego-tiations were under way.It is followed by a descriptionof China’s WTO accession commitments. The finalsection examines the likely impact of foreign entryon China’s banking sector and related policy issues.

Initial Conditions

The post-1978 economic reforms in China—whenagriculture and industry began to grow rapidlyin response to improved incentive systems—ledto a large increase in incomes, savings, and bankdeposits. For the most part, deposits were channeledto state-owned enterprises (SOEs) through the state-owned banking system, with insufficient regard forcredit quality. Although investment decisions werecontrolled by the central government in theory, inpractice subnational governments enjoyed remark-able degrees of freedom in choosing investmentsand appropriating finance from the local branchesof the state banks.

The banking system, and later the securities mar-ket, played an even more important role in support-ing the level of economic activity in the early 1990s,as the government budget became increasingly con-strained by a drop in revenues. Revenues fell fromabout 34 percent of the gross domestic product(GDP) in the early to mid-1980s to below 12 per-cent by the mid-1990s. However, the legacy of poorloans began to affect the profitability of banksacutely by the mid-1990s, although the propensityof Chinese households to save a large part of theirincome (averaging 28 percent of GDP), the relativeabsence of alternative savings vehicles, and theimplicit deposit guarantees at the state banks pro-vided high levels of liquidity to the banking system,helping to mask its problems. The rapid growth ofhousehold savings in the domestic banking systemhas also been possible because of capital controls

and the successful nationalization of capital flightthrough the creation of an onshore foreign currencydeposit and loan market. Uncertainties about thesolvency of pension plans, the absence of a mort-gage market (until 1999), and funds for educationneeds and purchases of other services are alsobelieved to have stimulated savings in the form ofbank deposits. However, this proposition has not yetbeen tested rigorously.

Financial Sector Reform

China’s attempts to modernize its financial markethave been steady, but, beyond an occasional flurryof precautionary measures introduced after themini-crises of the 1990s, they have focused mainlyon institutional diversification and strengthenedadministrative oversight. It is possible to distin-guish four distinct phases of institutional reform inChina in the 22 years leading up to WTO accession.

The first phase (1979–86) was characterized bythe breakup of the mono-bank system. PBCbecame the central bank, and the Industrial andCommercial Bank of China (ICBC) was created tohandle urban commercial banking, thus joiningother specialized institutions for large-scale con-struction (China Construction Bank [CCB]), for-eign exchange transactions and international trade(Bank of China [BOC]), and rural lending (Agri-cultural Bank of China [ABC]). Foreign banks wereinitially allowed to open representative offices, butby the end of this phase they also were permitted toestablish commercial branches, albeit with tightgeographic, product, and customer restrictions.

The second phase (1987–91) was marked by areduction in administratively governed specializa-tion among institutions, the rapid growth of non-bank financial intermediaries, the establishment ofthe first joint stock universal bank (Bank of Com-munications [BOCOM]), the establishment of twostate-owned insurance companies (in addition tothe People’s Insurance Company of China [PICC]established in 1949), and the beginnings of a capitalmarket by the introduction of secondary markettrading in government securities.

The third phase (1992–96) saw even greater diver-sification in the financial market. Stock exchangeswere established in Shanghai and Shenzhen, theinterbank market was developed, some interest ratesbecame flexible, and life and non–life insurance


licenses were extended to foreign firms, but withtighter product and geographic restrictions than fordomestic providers.

In the final pre-WTO accession phase of reform(1997–2001), the emphasis shifted to addressing theportfolio problems of the commercial banks andgovernance of the financial market as a whole. In asense, the preparations for WTO accession effectivelybegan during this period. They entailed restructur-ing the PBC, attempting to improve bank supervi-sion and the recognition of bad loans in the portfo-lios of state banks, creating four asset managementcompanies (to segment a portion of such loans),clar-ifying the roles of the China Securities RegulatoryCommission (CSRC) and the China Insurance Regu-latory Commission (CIRC), and diversifying finan-cial instruments to meet the emerging needs of savers

(e.g., mortgage loans and automobile finance toaccompany the rapid growth of newly created hous-ing markets and the anticipated increase in car buy-ing as their prices dropped after China’s accession tothe WTO).Moreover, the pressure to improve corpo-rate governance within financial institutions andincrease the transparency of financial informationprovided by listed companies and banks was notchedup, using a combination of sanctions, personnelmovements, moral suasion, and new regulations.

Preaccession Situation

Despite these measures, China’s financial servicesindustry seemed far from ready for increased com-petition from abroad at the time of WTO accession(table 11.1).3 The size of the Chinese financial

Accelerating Financial Market Restructuring in China 183

TABLE 11.1 China’s Financial Services Industry, 2001

Size (Trillions of Yuan) Share (%)

Type of Institution Assets Loans Deposits Assets Loans Deposits

Deposit money banks 15.0 10.7 13.1 100.0 100.0 100.0State banks 10.1 7.2 8.7 67.3 67.3 66.4Other banks 1.9 1.2 1.6 12.7 11.2 12.2Foreign funded banks 0.4 0.2 0.01 2.7 1.9 0.1Urban credit cooperatives 0.8 0.6 0.8 5.3 5.6 6.1Rural credit cooperatives 1.6 1.3 1.7 10.7 12.1 13.0Finance companies 0.2 0.2 0.3 1.3 1.9 2.2

Other deposit institutions 1.1 1.4 0.2

Size (Trillions of Yuan) Prevalence

Assets Premium Income Depth (%) Penetration (Yuan)

Insurance companies 0.46 0.21 2.2 169

Trading Volume (Trillions of Yuan)

Overnight 7 Days 30 Days 90 Days Total

Interbank market 0.10 0.56 0.04 0.005 0.81

Size

Market Capitalization Turnover Trading Volume (Trillions of Yuan) (Trillions of Yuan) (Trillion Shares)

Securities market 4.4 3.8 31.5

Size (Trillions of Yuan) Share (%)

Treasury Financial Corporate Treasury Financial Corporate

Bond market 1.95 0.69 0.03 73.2 25.7 1.1

Sources: People’s Bank of China 2002–03; Almanac of China’s Finance and Banking 2002, Yearbook ofChina’s Insurance 2002.

market was remarkable. Total deposits in banksreached the exceptionally high level of 150 percentof GDP in 2001. China had the second-largest secu-rities market in Asia, with 1,189 listed companies,about 60 million brokerage accounts, and a marketcapitalization approaching 45 percent of GDP. Thefour large state banks—which had been trans-formed into universal banks and constituted thecore of the financial system—accounted for 67 per-cent of deposit bank assets and 56 percent of totalfinancial assets.

Nearly 77 percent of the annual financial savingsof households was deposited in the banking systemduring 1998–2001. Consequently, banks handledmost of the finance available to enterprises and thegovernment, of which the share held by small andmedium nonstate enterprises—the most dynamicsegment of the Chinese economy—was tiny. More-over, the rural finance system was malfunctioning,with the Agricultural Development Bank of China(policy bank), the Agricultural Bank of China, andthe rural credit cooperatives facing the severestrepayment problems in the banking sector. During2001, total bank loans expanded by 1.3 trillion yuan,while both new equity issues and corporate bondissues were a negligible fraction of this amount.Corporate (i.e., state enterprise) sector leverage wasvery high; equally, a high proportion of loans takenby state enterprises was nonperforming. Govern-ment estimates placed the nonperforming loans(NPLs) of the four large state banks at about 27 per-cent of their loans outstanding, which excluded theNPLs transferred in 2000 to four asset managementcompanies (nearly 20 percent of loans outstandingat the four banks at that time). Unofficial estimatesof the remaining NPLs in the state banks have run ashigh as 40–50 percent of loans outstanding.4 Inaddition, it is likely that the other domestic bankshave a larger proportion of NPLs.5 However, neitherthe application of existing loan classification sys-tems nor the current state of management informa-tion systems within the banks gives analysts any rea-son to believe that the various estimates present anaccurate picture of the condition of bank portfolios.

Current accounting, auditing, and reportingstandards and practices do not provide sufficientinformation for making judgments about the prof-itability of China’s financial institutions, the ade-quacy of their capital bases, and the efficiency oftheir operations.6 The information available on the

condition of their portfolios or in balance sheetsand income statements can, at best, help only inunderstanding broad changes in the direction ofthese parameters over time. However, the informa-tion available on the operating practices of Chinesefinancial institutions is sufficient to develop a set ofcharacterizations to describe the state of the majorindustry segments on the eve of WTO accession.7

Each of China’s four state banks is big—they havean average asset size of more than $400 billion, aver-age employment of over 415,000 people, and from15,000 to 58,000 branches (see Almanac of China’sFinance and Banking 2002). These numbers suggestthat, at least in terms of resources, they are largeenough to compete against most foreign banks, evenafter factoring in the potential entry of major inter-national players that have grown rapidly in recentyears through aggressive mergers and acquisitions.

The core institutional problem, however, is therelative lack of profitability and insolvency of thefour state banks, which undermines the many bene-fits that could result from size, incumbency, andreach.8 Furthermore,although improving, the bankshave limited capacity to make lending decisionsbased purely on the commercial appraisal of repay-ment ability and credit risk. Until recently, the PBC-regulated spread between deposit and lending rates(about 400 basis points, which are high by interna-tional standards but not unusual by developingcountry standards) was sufficient to sustain thereported profitability of banks, but it removedincentives to extend credit to underserved but rela-tively dynamic segments of the economy, such assmall enterprises. The competitive prospects of thestate banks are further undermined by their relativeinability so far to implement efficiency reforms—for example, by cutting staff as much as requiredor closing unprofitable branches. Since 1997, thestate banks have reportedly closed 29 percent oftheir offices and cut staff by 15 percent; not only arethese measures considered insufficient, the feasiblelimits under current ownership conditions seemto have been reached. Thus, as of the end of 2001,the pace of innovation in retail banking—wheresuch banks would normally be expected to havean advantage—was unsatisfactory, despite thefact that such services are relatively underdevel-oped in China, where they have massive potential.9

Finally, these banks’ lack of experience with inter-national banking—and their unfamiliarity with


internationally benchmarked banking practices,technology, and management—is a handicap.10

If information about the state banks is scanty,that on the other parts of the banking industry, withthe exception of the four listed commercial banks, iseven more deficient. The published data do notengender much confidence, and little systematicresearch has been undertaken into their operatingconditions. In general, however, it is believed thatthe urban and rural credit cooperatives (as well assome city commercial banks) are insolvent andmostly illiquid, with periodic reports of failuresamong them. In addition to poor management andbanking practices, the rural credit cooperatives havebeen affected adversely by the stagnation of therural economy since the mid-1990s. Although theABC generally withdrew from subcounty-levelbanking in 1996, leaving the field clear for the coop-eratives, the franchise value of rural banking didnot rise, and may have declined in the face of large-scale rural distress.

The other development of note in the periodimmediately preceding accession to the WTO wasthe transfer of a significant portion of directedlending to the three policy banks.11 It is not imme-diately apparent that this transfer reduced the levelof credit risk in the banking industry, because thereis little evidence that these banks are better placedthan the state banks to evaluate such risks. More-over, directed lending has played an important rolein the support state banks have given to the fiscalstimulus program introduced in 1998 and currentlyunder implementation. Finally, because the twomain policy banks (China Development Bank andChina Import and Export Bank, or Eximbank) fundthemselves mainly through the issuance of long-term local currency bonds, the majority of whichare sold to the state banks, covariant risk existsamong the major players in the banking system.

Despite these weaknesses, other segments of thebanking market show potential for growth andinnovation. Although deposit growth at the statebanks continues to be strong—rising nearly 26 per-cent during 1999–2001—joint stock and othercommercial banks, with a significantly smallerreach in terms of branch networks, have made rapidgains. Thus, despite the implicit guarantee thatdeposits at state banks are thought to enjoy, and thewidespread belief among banking analysts that con-cerns about financial sector solvency would result

in a flight of deposits to the big banks, the nonstatebanks saw an increase in deposits of 77 percent overthe same period. The same is true of loans, withthese banks accounting for 42 percent of theincrease over the period, compared with 21 percentfor the state banks. Agility in trawling for savings inurban areas, combined with deft exploitation of thenon-SOE segment of the corporate market, pro-duced such outcomes. Both areas of concentrationrepresent the fastest-growing parts of the Chineseeconomy. Therefore, it is likely that, unless theyprove to be equally dexterous, the state banks willhave a smaller market share in 2006, by which timeall of China’s existing WTO commitments on bank-ing liberalization are expected to come into force.

On the eve of WTO accession, the 191 foreignbanks in China operated almost as off-shore orenclave banks.12 The foreign bank penetration rate, ifmeasured by the number of banks or branches,seemed quite high—52 percent and 14 percent ofthe total Chinese banking industry, respectively.However, they accounted for less than 3 percent ofassets, 2 percent of loans, and about 0.1 percent ofdeposits. Despite these low numbers, several foreignbanks were able to become profitable within12–18 months of entering the Chinese market,although market reports also suggest that they hadnonperforming assets equal to 20 percent of theirtotal assets (and some as high as 90 percent),and theirbusiness was almost exclusively with enterprises withforeign investment.Because foreignbankswerebeinglimited mostly to foreign currency transactions—commanding a 24 percent market share in foreigncurrency loans and 45 percent share in internationalsettlements—such banks had co-located with theirclients, the overwhelming majority of whom wereforeign-domestic joint ventures in the coastal belt ofChina. Only a handful of banks—mainly with strongHong Kong (China) or other Asian institutionalpresence—had developed longer-term strategiesaimed at other local market segments. The banksoperated in an environment that was radically differ-ent fromthat for thedomesticbanks.Theyhadtodealwith onerous location, product, and investmentrestrictions,and a vastly different supervisory and taxregime (foreign banks are taxed at a 33 percent rate,while local banks are taxed at a 50 percent rate).Over-all, few linkages exist between foreign and domesticbanks—which limited technological spillover andcompetition in the domestic banking industry.


China’s WTO Commitmentsin Financial Services

China’s commitments to liberalizing its financialservices industry, and indeed its General Agree-ment on Trade in Services (GATS) commitments asa whole, are possibly the most radical offerings evernegotiated in the history of the WTO and in theUruguay Round (see table 11.2). Although haz-ardous, it is interesting to speculate briefly aboutthe possible motivation for such a fundamentalattempt to liberalize the financial services industry.

At the time the Chinese government was negoti-ating its specific commitments, a significant body ofliterature had been building up on the effect of for-eign entry into developed and developing countryfinancial markets. Issues related to the survival ofdomestic financial institutions and the possibility ofmore complex macroeconomic risks and manage-ment demands were known and debated. At thesame time, the evidence of a favorable impact onfinancial market efficiency was equally impressive,especiallybecauseitwasbasedonverycarefulanalysis


TABLE 11.2 China’s WTO Commitments in the Financial Services Sector

Subsector Commitment

Banking

Securities

Fundmanagement

Insurance

Location: Upon accession, foreign currency business is allowed without geographicrestriction. Geographic restrictions on the local currency business of foreign bankswill be phased out over five years; four cities will be opened upon accession, fouradditional cities thereafter.

Products: Within two years of accession, China will permit foreign banks to providelocal currency services to Chinese enterprises; within five years, to all Chinese.

Investment: Within five years of accession, all current nonprudential measuresregarding the ownership, operation, and establishment of foreign banks, as wellas those concerning their branches and restrictions on issuing licenses, will beeliminated (national treatment).

Product: Foreign securities companies may engage directly in B share business.Investment: Within three years of accession, foreign investment banks will be

permitted to establish joint ventures, with foreign ownership not exceeding33 percent, to engage (without Chinese intermediary) in underwriting domesticshares (A shares) and underwriting and trading in foreign currency denominatedsecurities (B and H shares, government and corporate debts).

Representative offices of foreign securities companies may become special membersof Chinese stock exchanges.

Investment: Upon accession, the establishment of joint venture fund managementcompanies will be permitted, with foreign ownership not exceeding 33 percent,to conduct domestic fund management business.

Foreign investment shall be increased to 49 percent after three years.

Location: All geographic restrictions will be lifted in three years after the entry.Product: Upon accession, foreign life insurers will be permitted to provide individual

(nongroup) life insurance services. Two years after entry, they will be permitted toprovide health insurance, group insurance, pension insurance, and annuities toChinese and foreign customers.

Reinsurance is completely open upon accession, with no restrictions.Investment: Upon accession, foreign life insurers will be allowed to hold 50 percent

ownership in joint ventures. They may choose their own joint venture partners.For nonlife, China will allow branching or 51 percent foreign ownership upon

accession and wholly owned subsidiaries in two years after the entry (i.e., norestriction on the form of enterprise establishment).

Licenses will be granted solely on the basis of prudential criteria with no economicneeds test or quantitative limits on the number of licenses granted.

Source: Ministry of Commerce 2002.

of country experience and microdata.13 China’sdecision to join the WTO was a clear attempt tobenchmark its economic activities to internationalstandards and to ensure that the future growth ofthe economy was along the lines of its comparativeadvantage. Within this broader context, and givencertain institutional and political considerations rel-evant to the financial sector, the evidence on foreignentry and efficiency seems to have been compelling.

The basic impulse for financial sector liberaliza-tion seems to have been based less on the effects itwas likely to have on the mobilization of a greatervolume of financial savings and the promotion ofa higher level of investment—which in standardmodels appears to lead to a temporary increase inthe aggregate growth rate of the economy—thanon the need for technical innovation that wouldlower the cost of financial services and permanentlyboost growth. Moreover, in recent years Chinesepolicymakers have expressed greater concern aboutthe efficient allocation of resources in the economy.This concern has manifested itself in pricing, com-petition, and fiscal measures, as well as financial sec-tor reform. The rapid liberalization of many realsectors—for example, housing and other services—provided the opportunity to apply a similar compe-tition policy model in banking reform, although lessso in other financial market segments which,because of their infancy, attracted stronger protec-tionist sentiment.

Equally important, given the size of the domi-nant players in the banking market, the informa-tion gaps and span of control problems thataccounted for their ineffective management, andthe difficulties of severing or weakening the linksbetween SOEs and banks, fairly strong pro-competitive measures had to be employed. Foreignentry, suitably phased to provide a manageableadjustment process, addressed many of the per-ceived problems with the financial services indus-try in China. The expected diversification of finan-cial instruments for saving, liquidity management,and financing of economic activity would alsosupport the transition to market finance. Guid-ance from the recent Sixteenth National Congressof the Communist Party on active promotion ofthe private sector is bound to be translated intodemands for a significant redirection of debt andequity flows. More significant, it sets China inex-orably on the path toward market-based finance

and highlights the need to manage the transitionactively.

Major Issues and Implications

This concluding section examines briefly some ofthe major banking issues related to WTO accessionthat will have to be addressed in the transition tomarket-based finance. It is useful to start with thelikely effect on general economic management,because accession to the WTO is likely to result in asignificant remapping of the flow of funds in thefinancial market.

Interestingly, many of the factors traditionallythought to have permitted relatively uncomplicatedfinancial and macroeconomic management, despiteinadequate economic information, weak accounta-bility of institutions, and distortions in market sig-nals, are likely to become high-pressure points inthe transition period. They include control overmajor interest rates, capital flows, and the level ofthe exchange rate. The ability to capture and retainwithin the financial system a large proportion ofhousehold savings, then to channel it to enterprisesand government directly or indirectly to supportfiscal policy with quasi-fiscal stimulus through thefinancial system, is likely to be affected the most. Asforeign banks and other financial institutions areallowed to offer a broader range of domestic andforeign currency savings instruments, the excessliquidity of local banks will shrink. If expectationsabout the selective credit reach of foreign institu-tions are realized, the liquidity position for meetingthe burgeoning needs of the domestic banks’ largecustomer base is likely to tighten. Finally, the needto correct disparities that have resulted from usingan unbalanced growth model to promote coastalprovinces represents a special challenge. The effectsof geography and favorable government policiestoward such provinces have been reinforced by lop-sided financial sector development, which discrimi-nates against the growth of the interior and westernprovinces of China.14 The geographic liberalizationbuilt into China’s WTO accession commitmentswill tend to foster more rapid development of thefinancial services industry—and the benefits tosavers and investors that stem from it—in thesesame regions.

A central issue related to financial stability andthe future of the banking system is the fate of the


state banks. Clearly, rapid development of the non-bank financial market will erode the role of allbanks—a particular concern with respect to theirability to capture household savings and maintainlending in the face of declining levels of liquidity.This problem could, of course, be mitigated partlyby removing gradually and according to satisfactoryprudential control the current investment restric-tions on banks—for example, by permitting pru-dent participation in equity markets and investmentfunds. However, the recent portfolio problems of banks will also limit their future risk-bearingcapacity, and possibly their ability to compete withforeign banks on the development of new financialinstruments. State banks are particularly vulnerable

because their high operating costs and inability toreduce NPLs stem as much from weak managementas from domestic political economy pressures.15

Despite recent attempts to cut the number ofbranches and attract new sources of equity (as, forexample, in the recent initial public offering by theBOC), it is unlikely that small changes in macro-economic parameters, banking behavior, or thelevel of nonperforming loans will restore the healthof their balance sheets.

Extrapolations from published data on the statebank income statements and balance sheets and avariety of assumptions about macroeconomicparameters are shown in box 11.1. They suggestthat, without fairly radical actions to reduce


BOX 11.1 Modeling State Bank Performance: Can the State BanksSurvive Liberalization?

Within the limitations of the data, it is useful toassess the prospects of China’s main banks afterthe post-WTO accession period. The official andindustry response to banking system liberaliza-tion hinges critically on whether the banks areable to earn their way out of their current diffi-culties while facing foreign entry and associatedchanges. The results of a simple financial projec-tions exercise, which draws on publicly availableinformation on the condition of the portfolios ofthe four state-owned commercial banks, theiraccounts, and assumptions about key macro-economic parameters, suggest the followingbroad trends:

• By the end of the accession period in 2007, thebanks will not, either individually or as a group,be able to cope with their current and prospec-tive nonperforming loans (a level of 30 percentis used for the base case, ranging from 25 to 40percent for the individual banks). It is likely thatall but one bank will have negative capital by2007, and that, as a group, the four banks willhave negative capital.

• The operating margins are too low through-out the post-accession period, and not onlydo the banks not make profits during thetransition period, but their capital turns evenmore negative over time.

• Changes in assumptions about the currentstock of nonperforming loans make a signifi-cant difference to the results of the base case.However, within a zone of 20 percent aroundthe base case assumption (about 5 percent-age points above and below the 30 percentnonperforming loan level assumed in the base

case), the banks as a group are likely to stillhave negative capital by 2007.

• Competitive pressures, either in the form ofrising interest rates to protect deposit basesor shrinking loan margins, would worsen thebase case results.

• The group results are not affected significantlyby small changes in the assumptions aboutmacroeconomic growth, inflation, or theoperating costs of banks. However, the effecton individual banks could differ substantially.Also, even large changes in macroeconomicvariables—for example, higher credit growthresulting from 50 percent higher outputgrowth in the economy—will be inadequatefor stopping the erosion of the capital base ofthe banks.

The conclusion to be drawn from this simplefinancial modeling exercise is that the statebanks will come under severe pressure duringpost–WTO accession liberalization. Fairly sub-stantial changes in operating costs would berequired to improve the condition of the banks.Failing this, because they are unlikely to be ableto simply “grow out of their problems”—that is,they will be unable to recapitalize themselvesfrom provisions and retained earnings—a recap-italization of the state banks from outsidesources is likely to be required. A combination ofinstruments most probably will be used torestore the health of the state banks’ balancesheets. These results support broadly the find-ings of other, similar modeling exercises relatedto China’s state banks (see, for example, Li andMa 2001).

operating costs, the banks will be unable to earntheir way to improved health. In such an event, toavoid a loss of confidence in the state banks or asharp erosion of their competitive position withrespect to foreign and smaller domestic banks,financial reengineering will have to accompany theneeded changes in banking practices. Toward thisend, it is possible that bolder actions to inject newcapital from private and public sources will have tobe taken. Private capital would be preferable, espe-cially if it is bundled with access to new bankingtechnologies and financial expertise, but manymarket analysts believe that the use of public capi-tal is unavoidable. If so, the internal reform of thebanks should be speeded up and strengthened, sothey can move toward self-sufficency. Reforms willreduce the size of the public capital infusion thatwill be needed which, given many other claims ongovernment finances, would help smooth post-WTO accession and structural adjustment else-where in the Chinese economy.

Notes

1. All dollar amounts are current U.S. dollars.2. This chapter focuses on banking, although it refers to

other financial market segments where relevant.3. Domestic banks compete vigorously in extending credit to

the larger corporate customers. One indicator of price competi-tion is the tendency to make full, but one-sided, use of the lee-way allowed by the PBC on loan rates (from 10 percent below to30 percent above the base rate). Because banks have little experi-ence with the estimation of risks, when they charge more thanthe base rate it is usually for the full amount allowed, which is30 percent, rather than for some amount in-between that wouldsuggest finer differences among their customers.

4. See, for example, Lardy (2000).5. These banks comprise 10 national commercial banks

(accounting for 12 percent of banking assets), 99 city commercialbanks, 1,695 urban credit cooperatives (5 percent of bankingassets), 39,255 rural credit cooperatives (10 percent of bankingassets), and 3 policy banks (10 percent of banking assets). Theother financial institutions were 12 financial leasing companies,100 trust and investment companies and regional policy banks,98 securities firms, 13 insurance companies, 191 foreign-fundedfinancial institutions, and 71 enterprise group finance companies.

6. China’s WTO accession agreement for the financial servicesindustry does not seem to have relied significantly on technicalcalculations of profitability or solvency. Instead, broad product/client/geographic criteria were used to give domestic institutionsand labor markets time to adjust. In this sense, the financial serv-ices sector is radically different from other sectors, such as agricul-ture, where the information base was much stronger and permit-ted disaggregated calculations of impacts, to which a host of finelysequenced market opening measures were tied.

7. The literature on which to draw is vast. See, for example,Lardy (1998), Cheng (1999), Bonin and Huang (2001a, 2001b),Fang (2001), OECD (2002).

8. Given their role and their dominance of the financial mar-ket, the current health and future performance of the state bankscannot help but be a major source of uncertainty about theprospects for development of the financial services industry andChina’s economy.

9. Most indicators—number of checking accounts, share ofpersonal loans, mortgages or automobile finance in total creditor normalized by GDP, automated teller machines per capita—show that Chinese banks lag substantially when compared withthose in neighboring East Asian countries. By contrast, China’sretail banking through the postal system has expanded rapidly,managing over $46 billion in 110 million accounts at the end of1999 through a network of 66,649 postal branches nationwide.The Postal Savings Bank, reconstituted in 2000, has establishedtransactions links with more than 17 countries and boasts a bet-ter electronic network than most state banks. See Leong (2000).

10. Almost all of the expertise was located within the BOC,which accounted for 80 percent of China’s financial sectorabroad, 40 percent of international trade settlement, and 85 per-cent of foreign exchange clearing.

11. They are the Agricultural Development Bank, ChinaDevelopment Bank, and China Export and Import Bank, whichtogether accounted for 11 percent of assets, 13 percent of loans,and less than 1 percent of deposits. They have mostly fundedtheir operations through long-term local currency bonds,mainly purchased by the state banks.

12. Of these, 92 percent were branches of foreign banks, andthe remainder were locally registered banks with full foreignownership or joint venture arrangements. In addition, 263 rep-resentative offices of foreign banks were situated in China.

13. See, for example, the excellent collection of papers pre-sented at the Third Annual International Financial Markets andDevelopment Conference, entitled“Open Doors: Foreign Partici-pation in Financial Systems in Developing Countries.”Sponsoredby the World Bank, International Monetary Fund, and BrookingsInstitution, it was held April 19–21, 2001, in New York City. Thepapers by Mathieson and Roldos (2001) and Pomerleano andVojta (2001) contain useful overviews of the issues. Also seeClaessens, Demigurc-Krunt, and Huizinga (1998).

14. See Boyreau-Debray (2002) and Hu and Zhou (2001) foranalyses of regional disparities in financial development.

15. In addition, the risks that could arise from a change inmarket conditions should be considered. For example, given thevery generous debt-service-to-household-income criteria in therapidly growing home mortgage market where state bank expo-sure is rising, an increase in market interest rates could sparkdefaults among households.

References


Almanac Editorial Board. 2002. Almanac of China’s Finance andBanking. Beijing.

Bonin, John P., and Yiping Huang. 2001a. “Dealing with the BadLoans of the Chinese Banks.” Journal of Asian Economics(summer).

———. 2001b. “Foreign Entry into Chinese Banking: DoesWTO Membership Threaten Domestic Banks?” Paper pre-sented at an international conference on Greater China andthe WTO, March 22–24, Hong Kong.

Boyreau-Debray, Genevieve. 2002. “Financial Intermediationand Growth: Chinese Style.” World Bank, Washington, D.C.,March. Processed.



Cheng Hang-Sheng. 1999. “Can China Achieve Bank Marketiza-tion without Bank Privatization? Paper presented at WorldBank, March. Washington, D.C.

Claessens, Stijn, Asli Demigurc-Kunt, and Harry Huizinga. 1998.“How Does Foreign Entry Affect the Domestic BankingMarket?” Policy Research Working Paper 1918. World Bank,Washington, D.C., May.

Fang Xinghai. 2001. “Reconstructing the Micro-Foundation ofChina’s Financial Sector.” Paper presented at Conference onFinancial Sector Reform in China, Harvard University,Cambridge, Mass., September.

Hu Angang, and Li Zhou. 2001. “Market Openness and GoodGovernance: The Changing Regional Disparity of FinancialDevelopment in China (1978–1999).” Paper presented atConference on Financial Sector Reform in China, HarvardUniversity, Cambridge, Mass.

Lardy, Nicholas R. 1998. China’s Unfinished EconomicRevolution. Washington, D.C.: Brookings Institution andInstitute for International Economics.

———. 2000. “When Will China’s Financial System MeetChina’s Needs?” Revised version of paper presented atConference on Policy Reform in China, February. StanfordUniversity.

Leong, Elaine. 2000. “China’s Postal Savings Service ChallengesState Banks.” FinanceAsia (September).

Li Kui-Wai, and Jun Ma. 2001. “China’s Accession to the WorldTrade Organization and Policy Options for BankingReform.” Paper presented at Greater China and the WTOConference, March. City University of Hong Kong, HongKong.

Mathieson, Donald J., and Jorge Roldos. 2001. “The Role ofForeign Banks in Emerging Markets.” Paper presented atThird Annual International Financial Markets and Develop-ment Conference. April 19–21. Sponsored by the WorldBank, International Monetary Fund, and Brookings Institu-tion, New York.

Organisation for Economic Co-operation and Development(OECD). 2002. China in the World Economy: The DomesticPolicy Challenges. OECD: Paris.

People’s Bank of China. People’s Bank of China QuarterlyBulletin. 2002–03. Vol. 27.

Pomerleano, Michael, and George J. Vojta. 2001. “What DoForeign Banks Do in Emerging Markets? An InstitutionalStudy.” Paper presented at Third Annual InternationalFinancial Markets and Development Conference, April19–21. Sponsored by the World Bank, International Mone-tary Fund, and Brookings Institution, New York.

Yearbook of China’s Insurance. 2002. Beijing: Insurance YearbookEditorial Board.

trade restrictions in Canada). The Chinese govern-ment has also set prices and limited competitionthrough a raft of import restrictions, which haveincluded quotas, high tariffs, and differential taxesthat favor local suppliers (see, for example, Li2003). The limitation of trade has encouragedinefficient production and has allowed market seg-mentation.1

The integration of Greater China (MainlandChina and Chinese Tapei) into the World TradeOrganization (WTO), and thus into most-favored-nation (MFN) principles, has important implica-tions for the Chinese economy, not least of all forthe motor vehicle sector. For example, the accessionagreements specify major changes in tariffs, theelimination of quotas as well as local contentrequirements, and changes in the rules governingforeign investment. The perception of the Chinesemarket by outside investors has already changed,because the application of WTO rules covering thetreatment of foreign firms has reduced uncertainty

An automobile industry has often been a symbol ofeconomic prestige in the developing world. Brazil,China, India, Indonesia, and Malaysia, among othercountries, have all promoted, and sometimes evenshowcased, the development of a domestic motorvehicle industry. In China, which has a huge popu-lation together with a surface area roughly as largeas that of the United States and almost 15 percentlarger than that of Brazil (see table 12.1), almostevery province has its own motor vehicle factoryand satellite factories. Yet, among the world’s majoreconomies, China has the largest number of peopleper vehicle. Even Indonesia, with a 30 percent lowerper capita income than China, has over 50 percentfewer people per automobile.

China’s situation is the result of a series of policymeasures dating from 1949 (see box 12.1). Furtherdistorting an efficient structuring of the automo-bile industry have been the internal measures thathave limited and even prohibited trade throughlocal protectionism (analogous to inter-provincial

191

12

WTO Accession andthe Structure of

China’s MotorVehicle Sector

Joseph F. Francois and Dean Spinanger

This paper builds on Francois and Spinanger (forthcoming 2004). Special thanks are extended to Zhang Wenkuifor help with data, to L. Alan Winters and Will Martin for detailed comments on an earlier draft, and to fiveanonymous referees, who offered valuable suggestions. Thanks also go to the participants in a World Bank-sponsored conference in Beijing for their helpful discussion. Finally, the authors would like to thank each other fortheir individual efforts to improve the quality of the output. Unfortunately, no agreement has been reached onwho accepts final responsibility for errors or misinterpretations, and so we blame each other for any shortfalls.

about the general economic climate, thereby induc-ing notable increases in investment and promptingnew decisions to enter the market.

This chapter assesses the effect of these broadchanges on the Chinese motor vehicle sector. Itemphasizes the role of administratively imposedinefficiencies within the sector and the role of suchinefficiencies in structural adjustment. We callthese regulated efficiencies. A 1993 World Bankreport pointed out that “it cannot be too muchemphasized that in the long run the Chinese auto-motive sector will have a most efficient industrialorganization when its development, dictated byinternational competition and state participationin the sector, is eventually replaced by entrepre-neurial ownership” (World Bank 1993, xix).

The industry itself anticipates significant change.In recent years, growth in the motor vehicle sectorhas been very rapid, with output expanding at anannualized rate of 13 percent in the four yearsending in 1999, at a 26 percent rate in the three yearsto 2002, and at more than 52 percent in 2003. Inview of the new, modern plants that came on line in2001 and 2002 and the additional facilities expectedto increase capacity by over 150 percent from 2002to 2005, a large, discrete change in production levelsis expected.

At the same time, WTO membership implieslower prices and steeper foreign competition inthe sector. The response to this shift in the compet-itive landscape will be shaped by continuing prob-lems with local government protection, lack of


TABLE 12.1 GNP, Population, and Stocks of Cars: Selected Countries, 2000

PPP GNP Stocks ofper Capita Population Cars/Trucks(2002 U.S. (Millions, (Millions, People Surface Area

Dollars) 2001) 2001) per Car (1,000 sq. km)

India 2,570 1,032.4 6.3/5.9 163.2 3,287Indonesia 2,990 209.0 3.0/2.4 68.8 1,905China 4,390 1,271.8 8.5/15.4 149.0 9,598Colombia 5,870 43.0 1.8/0.8 23.4 1,139Turkey 6,120 66.2 4.5/1.6 14.6 775Thailand 6,680 61.2 2.9/4.1 21.4 513Brazil 7,250 172.4 15.8/4.0 10.9 8,547Russia 7,820 144.8 21.2/5.1 6.8 17,075Malaysia 8,280 23.8 4.2/1.0 5.6 330Mexico 8,540 99.4 12.2/5.6 8.2 1,958South Africa 9,870 43.2 41.0/2.5 1.1 1,221Argentina 9,930 37.5 5.4/1.6 7.0 2,780Rep. of Korea 16,480 47.3 8.9/4.0 5.3 99Chinese Taipei 17,730 22.4 4.8/0.9 4.6 36Spain 20,460 41.1 18.2/4.2 2.3 506Italy 25,320 57.9 33.2/3.8 1.7 301United Kingdom 25,870 58.8 27.8/3.4 2.1 243Japan 26,070 127.0 53.5/19.9 2.4 378France 26,180 59.2 28.7/5.9 2.1 552Germany 26,220 82.3 44.4/3.6 1.9 357Canada 28,070 31.1 17.1/0.7 1.8 9,971United States 35,060 285.3 128.7/88.0 2.2 9,629

Total 9,224 4,017.1 492.2/184.3 8.2 71,200Low and middle income 4,682 3,274.4 140.6/54.9 23.3 49,263High income 29,248 742.7 351.6/129.4 2.1 21,937

PPP GNP = GNP adjusted for purchasing power parity exchange rate differences.Sources: World Bank, World Development Indicators, various issues; VDA, various issues.

automobile infrastructure such as roads, parking,and service facilities, and related factors that act asconstraints on growth of the sector. Even so, theindustry itself expects continued strong growth.2

Notwithstanding industry expectations aboutits prospects, what is really likely to happen oncethe competitive landscape has changed in criticalways? The primary approach employed here toexplore this question is general equilibrium, involv-ing the application of a computational model.The next section discusses some basic issues aboutthe structure of China’s motor vehicle industry,in particular the impact of government interven-tion on its efficiency. A brief description of themodel framework and then the experiments them-selves follow. Conclusions are offered in the finalsection.

China’s Auto Industry

Because of China’s own national and regional poli-cies, the country’s motor vehicle industry is highlyfragmented and inefficient by global standards.This situation is not just the result of the introduc-tion of Soviet-style industrialization in the 1950s,when firms were merely production units andquestions about efficiency were irrelevant becausethe dictated output was all that mattered. As notedby Zhang and Taylor (2001), the First AutomobileWorks (FAW) provides ample evidence of theimpact evolving from the various policies in placein China over the past 50 years. Between 1959 and1981, FAW produced “a mere 1,542 units, an aver-age 67 units per annum.” As Zhang and Taylor alsonote, in 1970 the production cost of a particular

WTO Accession and the Structure of China’s Motor Vehicle Sector 193

BOX 12.1 Summary of Developments in China’s Automotive Sector

The following summary is based on the authors’survey of the literature:

1953–65: Self-Reliance Policy

• About 60,000 vehicles were produced eachyear.

• Production relied on Soviet technologies.• Sector had no other international contacts.• Provincial governments set up production

units.• By 1960, 16 auto producers and 28 assem-

bly companies were in place.

1966–80: Security-Oriented Policy

• Government invested heavily in westernregions (Sichuan, Shanxi, and Hubei).

• Remote locations caused severe problemsand overcapacity.

• Focus was on heavy vehicles for militarypurposes.

• Car demand increased rapidly and capacitieswere expanded to 160,000 units per year.

• By 1980, 58 carmakers, 192 assemblycompanies, and 2,000 spare parts producerswere operating in China.

1981–98: Initial Fruits of the Open Door Policy

• The Open Door policy adopted in 1978 kick-started the industry.

• From 1983 to 1985, the number of automo-tive companies almost doubled, from 65 to114 units.

• By 1998, roughly 2,500 production unitswere in place.

• Provincial governments further regionalizedproduction.

• Major international firms began to invest,and then toward 1998 began to invest quiterapidly.

• Volkswagen had already established opera-tions in China—in 1978.

• These joint ventures accounted for about60 percent of production during the period1981–98.

1999– Opening Up and Beyond

• Major investments have been initiated byforeign companies.

• All major Japanese companies now haveoperations in China.

• All major German producers now haveoperations in China.

• French and Italian producers are nominallypresent.

• U.S. producers also are nominally present.• Currently, the expansion is rapid; capacity is

now close to 2.5 million units.• Capacity is especially growing in coastal

areas.

model (the CA72) was 220,000 yuan, but “the salesprice was only 40,000 yuan. . . . In the absence ofcompetition, all production units ran at low levelsof productivity and efficiency. . . . The legacy left byMao’s regime loomed well into the 1980s. By 1980the number of automotive enterprises had risen to2,379, consisting of 56 vehicle manufacturers . . .[producing, among other things] 5,418 cars.”

China’s fragmented and inefficient automotiveindustry is also the result of import substitutionpolicies, together with cooperation agreementsmade with foreign companies beginning in the1980s that were meant to fill the increasing gapbetween production and the rapidly expandingdemand for automobiles. After all, the major thrustof policies until then had been to build trucks.2

These trends are depicted in figure 12.1.Thus China’s automotive companies are operat-

ing with cost structures well within the global fron-tier, and the plants are producing considerablybelow global standards for efficient scale. Thisobservation is illustrated in table 12.2 for the sedanindustry, assuming that plants of minimum effi-cient scale (MES) for final assembly of cars arethose that produce at least 200,000 units a year ofone model (see Huang 2002, p. 543). China’s entiresedan production in 1998 was 507,000 vehicles,produced by 13 factories. The fact that only twofactories produced more than 100,000 sedans eachimplies that fewer than 40,000 sedans were pro-duced by each of the other factories.

Actually, a great deal of variation in plant scaleexists around this average (see table 12.2). The lead-ing producer, Shanghai-Volkswagen, made 248,000sedans in 2002 (and therefore operated at MES lev-els if producing one model). Several other plantshad production runs of less than 20,000. In thissense, there are strong parallels to the situationin Mexico prior to implementation of the NAFTA inwhich protected, inefficient factories operated wellwithin the global technology frontier (see Lopez deSilanes, Markusen, and Rutherford 1994). Overall,China’s motor vehicle industry has about 2,400industrial enterprises. In 1998, these included122 motor vehicle manufacturing plants, 520 auto-refitting factories, 130 motorcycle factories, 62 carengine factories, and 1,589 auto and motorcyclespare parts factories. Annual production capacityexceeds 2.3 million motor vehicles and 10 millionmotorcycles. Since 1995, the general pattern hasbeen closure of the smaller plants (generally rele-gated to the “other” category in table 12.2) andexpansion of production runs in the larger plants.With foreign investment and the rapid growth inthe industry, the number of plants producing atleast 25,000 vehicles rose from 3 in 1995 to 11 in2002 (see U.S. and Chinese National Academies ofEngineering 2003).

Import and domestic shipment data in valueterms are summarized in table 12.3 (the data arefor 1997, which serves as the “pre-accession” refer-ence point). Import tariff protection is summarized


0

1,000

500

1,500

2,000

2,500

3,000

3,500

1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Thousands

Total motor vehicles

Trucks and buses

Passenger cars

FIGURE 12.1 China’s Production of Motor Vehicles sinceAdoption of the Open-Door Policy

Sources: Bessum 2002; VDA, various issues.


TAB

LE 1

2.2

Au

to P

rod

uct

ion

by

Pla

nt

in C

hin

a, 1

99

5–2

00

2

Ran

k20

02/1

995

Plan

t19

9519

9619

9719

9819

9920

0020

0120

02a

1/1

Shan

ghai

-VW

160,

070

200,

222

230,

443

235,

000

230,

946

221,

524

230,

378

248,

000

2/4

FAW

-VW

24,5

5344

,825

46,4

0566

,000

81,4

6494

,147

101,

622

131,

000

3/n.

a.Sh

angh

ai-G

M—

——

——

30,0

2458

,548

106,

000

4/2

Tian

jin X

iali

(Dai

hats

u)65

,258

88,2

3295

,155

100,

021

101,

828

81,9

5141

,703

93,0

005/

5FA

W-A

udi-H

ongq

i19

,350

15,0

0015

,731

31,2

2552

,667

78,0

006/

9Sh

enlo

ng (

Citr

oen)

3,79

79,

228

30,0

3536

,240

40,2

0053

,900

52,8

5068

,000

7/6

Cha

ng’a

n (S

uzuk

i)17

,770

16,4

2035

,160

36,2

3944

,583

48,2

3550

,573

64,0

008/

n.a.

Gua

ngzh

ou-H

onda

——

—2,

246

10,0

0832

,228

51,1

5360

,000

9/n.

a.Sh

angh

ai-Q

irui

——

——

—2,

767

30,0

8547

,000

10/n

.a.

Gee

ly G

roup

——

——

—14

,594

21,7

0238

,000

11/n

.a.

Don

gfen

g Fe

ngsh

en—

——

——

3,15

98,

000

32,0

0012

/n.a

.H

aim

a (N

aina

n-M

azda

)—

——

——

3,05

97,

800

20,0

0013

/n.a

.Yu

edo-

KIA

——

——

—2,

423

6,21

016

,000

14/n

.a.

Qin

chua

n—

——

——

5,38

05,

686

16,0

0015

/n.a

.N

anya

——

——

—1,

000

8,00

013

,500

16/3

Beiji

ng (

Jeep

)25

,127

26,0

5119

,377

8,34

49,

294

4,86

74,

663

4,40

017

/7G

uizh

ou Y

unq

ue (

Suba

ru)

7,10

579

81,

000

——

859

1,25

32,

100

18/n

.a.

Tian

jin-T

oyot

a—

——

——

——

2,00

0n.

a./8

Gua

ngzh

ou-P

euge

ot6,

698

2,41

61,

557

——

——

—O

ther

22,5

70—

22,4

798,

013

31,3

1217

,930

—1,

900

Tota

l35

2,29

838

8,19

248

1,61

150

7,10

356

5,36

664

9,27

273

2,88

31,

040,

900

Num

ber

of p

lant

s3

45

55

89

11p

rodu

cing

>25

,000

Num

ber

of p

lant

s2

22

33

47

8p

rodu

cing

>50

,000

Num

ber

of p

lant

s1

11

22

12

3p

rodu

cing

>10

0,00

0

— N

ot a

vaila

ble

beca

use

pla

nts

have

clo

sed

or w

ere

not

yet

open

ed.

n.a.

Not

ap

plic

able

.a Th

e 20

02 v

alue

s ar

e ba

sed

on c

omp

any

pro

ject

ions

.So

urce

s:Be

ssum

200

2; C

hine

se M

otor

Veh

icle

Doc

umen

tatio

n C

ente

r 20

02.

in table 12.4. China’s pre-accession average tariff onauto products (vehicles and parts and electronics)was 35 percent. The rate for vehicles averaged 70percent, with sedans subject to tariffs of between 80percent and 100 percent. Motor vehicle parts weresubject to an average tariff of only 23 percent.Import shares were relatively low, averaging per-haps 3 percent in the years 1995–2002. Officially,only 20,000 sedans were imported.3 Governmentpolicy encouraged the use of domestic parts, and,better yet, locally produced parts. For new invest-ments, domestic content rules applied, stipulating80 percent domestic content by the third year. Allthis is reflected as well in the low share of automo-tive parts imports in total production. As for theownership of companies, even after full implemen-tation of the WTO, foreign ownership will belimited to 50 percent.4

China’s tariff rates are scheduled to come downsubstantially (table 12.4)—25 percent for vehiclesand 10 percent for parts on an MFN basis as part of

WTO accession. In addition, quotas on automo-biles and on key components will be phased out by2006, after being expanded by 15 percent a yearuntil then. Likewise, domestic content require-ments have already been removed. (Both of thesenontariff measures violate basic WTO rules.)Clearly, these changes in the structure of protectionwill have significant implications for the structureof the automotive sector. Critically, other WTOobligations imply free movement of importedautos (free of import quotas) within the Chinamarket. This situation implies tremendous pressurefor a breakdown of internal barriers for domesticproduction and for a rationalization of the domes-tic industry. The internal barriers to trade simplycannot be sustained if China’s new WTO obliga-tions are to be taken seriously.

The government has itself recognized this situa-tion. Official and industry sources have signaledtheir intention to support only a small number ofdomestic production groups, perhaps including the


TABLE 12.3 Import and Domestic Shipment Values of China’sMotor Vehicle Industry, 1997 (millions of US$)

Value

Imported motor vehicles and parts, world prices 3,607.71Imported motor vehicles and parts, internal prices 4,849.31

Imported parts, internal prices 3,239.45Imported motor vehicles, internal prices 1,609.86

Domestic intermediates and parts 32,812.46Domestic intermediate parts 10,896.15Industry consumption of motor vehicles 21,625.50Final consumption of motor vehicles 290.81

Source: Global Trade Analysis Project (GTAP) version 5 database.

TABLE 12.4 Tariffs on Chinese Motor Vehicles (percent)

Base Rate Final Rate

Finished motor vehicles 70.5 25Motor vehicle parts 23.4 10Electronic vehicle parts 12.0 10Average, vehicles and parts 34.7 15

Sources: China’s WTO accession schedule; GTAP data; Office of the U.S. TradeRepresentative.

Shanghai group (Volkswagen), China’s First AutoWorks (Volkswagen), Shanghai-GM (Buick), andthe Dongfeng Group (Citroën). These groups withtheir foreign partners already account for over70 percent of production in China. Such a sharprationalization would undoubtedly be painful, butit could allow the industry to consolidate produc-tion and work its way down the average cost curvefor vehicle production.

The Modeling Framework

The next section sketches a quantitative assessmentof the possible impact of WTO accession on China’smotor vehicle industry—an assessment made byapplying the computable general equilibrium(CGE) model. This section provides a briefoverview of the model. More technical details andreferences are provided in Francois and Spinanger(2001) and in the technical annex available fordownloading with the model files.5 For multi-sectorpolicy initiatives (such as WTO accession) in whichinteractions between sectors have a major effect onresults, the use of CGE models has become arelatively standard approach (see François 2000).We recognize that the results of these exercises arehampered both by the assumptions made and bythe quality of the data available. Even so, we believetheir utility in estimating the possible overall pat-tern of the effect of broad policy changes—bothdirect and indirect—is well proven.

The Model Data

The model data, taken from a variety of sources,were organized into 23 sectors and 25 regions.(Note that we included some detail on the value-added chain, linking fibers with textile and clothingproduction to better capture the initial impact ofthe Agreement on Textiles and Clothing [ATC] onour base scenario.) The sectors and regions for this23 × 25 aggregation of the data are detailed intable 12.5.

Data on production and trade are based onnational accounting data linked through tradeflows and drawn directly from the Global TradeAnalysis Project (GTAP) version 5 data set(McDougall 2001). The GTAP version 5 data set isbenchmarked to 1997, and it includes detailed

national input-output, trade, and final demandstructures. We modified the basic database. In par-ticular, we updated the data set to better reflectactual import protection for goods and services.

The basic data on current tariff rates are drawnfrom United Nations Conference on Trade andDevelopment (UNCTAD) and WTO data on theschedules of applied and bound tariff rates, andthey are integrated into the core GTAP database.They are supplemented with data from the Officeof the U.S. Trade Representative and the U.S.International Trade Commission on regional pref-erence schemes in the Western Hemisphere. Foragriculture, protection is based on Organisationfor Economic Co-operation and Development(OECD) and U.S. Department of Agriculture(USDA) estimates of agricultural protection, asintegrated into the GTAP core database. Tariff andnontariff barrier estimates are further adjusted toreflect remaining Uruguay Round commitments,including the phaseout of remaining textile andclothing quotas under the ATC. Data on post–Uruguay Round tariffs are taken from recent esti-mates reported by Francois and Strutt (1999).These estimates are taken primarily from theWTO’s integrated database, with supplementalinformation from the World Bank’s recent assess-ment of detailed pre– and post–Uruguay Roundtariff schedules. All of this tariff information hasbeen matched to the current model sectors.Services trade barriers are based on the estimatesdescribed in the technical annex and are shown intable 12.6. (The basic GTAP database includes noinformation at all on trade barriers for services, forexample.)

Although the basic GTAP data set is bench-marked to 1997 and reflects applied tariffs actu-ally in place in 1997, we want to work with arepresentation of a post–Uruguay Round world.To accomplish this, before conducting any policyexperiments we run a “pre-experiment” in whichwe implement the remaining Uruguay Round tar-iff cuts. Most of these cuts are already in place inthe 1997 benchmark data set. At the same time,the data are adjusted to reflect regional preferenceschemes in Latin America (not represented in thecore GTAP database). The data set we work withfor actual experiments is therefore a representa-tion of a notional world economy (values are in



TABLE 12.5 Regional and Sectoral Breakdown of the Model

Model Regions/Economies Model Sectors

Economy Description Sector Description

Hong Kong Hong Kong (China) Wool Wool

China (PRC) People’s Republic of China Other natural fibers Natural fibers (cotton, etc.)

Chinese Taipei Chinese Taipei Primary food Primary food production

Japan Japan Other primary production Other primary production

Korea, Rep. of Korea, Rep. of Sugar Sugar

ASEAN5 ASEAN5 member statesa Processed foods Processed foods, tobacco, and beverages

Vietnam Vietnam

India India Textiles Textiles

Bangladesh Bangladesh Clothing Wearing apparel

Other South Asia Other South Asian economiesb Leather goods Leather products

Australia Australia Chemicals, rubber, and Chemicals, refinery products, refinery products rubber, plastics

New Zealand New Zealand Primary steel Steel refinery products

Canada Canada Primary nonferrous metals Nonferrous metal products

United States United States of America Motor vehicles and parts Motor vehicles and parts

Mexico Mexico Electronics Electronic machinery and equipment

Brazil Brazil Other machinery and Other machinery and equipmentequipment

MERCOSUR, other MERCOSURc Other manufactures Other manufactured goods

CBI Caribbean Basin Initiativeeconomiesd

ATP Andean Trade Pact Wholesale and retail trade Wholesale and retail trade serviceseconomiesd

Chile Chiled Transport services Transportation services(land, water, air)

Other Latin America Other Latin Americad Communications Communications services

European Union(15) European Union, Construction Construction15 economies

Turkey Turkey Finance, insurance, and Finance, insurance, and real real estate estate services

Africa, Mideast Africa and the Middle East Commercial services Other commercial services

Rest of world Rest of world Other services Other services (public,health, etc.)

aASEAN5 � Philippines, Thailand, Indonesia, Singapore, and Malaysia.bPakistan, Sri Lanka, and Nepal.cMERCOSUR � Argentina, Paraguay, Uruguay. Brazil is represented separately.dNot treated in tables and figures.


1997 U.S. dollars), with full implementation ofUruguay Round tariff cuts. We then examine boththe ATC phaseout and Greater China accessionwith reference to this post–Uruguay Round tariffbenchmark.

Model Structure

We turn next to the basic analytical features of themodel. Except for automobiles, we use a very stan-dard CGE model structure. On the production side,


TABLE 12.6 Mainland China’s Pre- and Post-WTO Accession Tariff Rates (as modeled)

Sector Model Base Rate Accession Rate New Bound Rate

MerchandiseWool 14.76 42.00 38.00Natural fibers (cotton, etc.) 3.14 17.38 13.58Primary food production 58.80 58.13 46.83Other primary production 0.48 6.94 5.08Sugar 29.49 30.00 20.00Processed food, tobacco, and beverages 37.65 40.66 23.18Textiles 25.09 25.43 10.21Wearing apparel 31.75 32.80 16.05Leather products 12.10 20.94 17.02Chemicals, refinery products, rubber, plastics 12.62 14.85 7.17Steel refinery products 9.68 8.92 5.10Nonferrous metal products 7.83 8.20 5.52Motor vehicles and parts 34.42 38.65 15.41

Motor vehicles 70.50 70.50 25.00Parts 23.40 23.40 10.00

Electronic machinery and equipment 11.93 16.90 9.62Other machinery and equipment 12.83 15.37 10.14Other manufactured goods 14.51 21.99 16.29

ServicesWholesale and retail trade services 0.00 n.a. 0.00Transportation services (land, water, air) 3.97 n.a. 1.99Communications services 9.18 n.a. 4.59Construction 13.68 n.a. 6.84Finance, insurance, and real estate services 8.08 n.a. 4.04Other commercial services 47.92 n.a. 23.96Other services (public, health, etc.) 25.74 n.a. 12.87

n.a. Not applicable.Note: Services barriers are based on gravity equation estimates. Accession rates reflect an assumed 50percent drop in cross-border trading cost estimates.Sources: China WTO accession schedule; GTAP data; Office of the U.S. Trade Representative. Gravityestimates are based on trade and macroeconomic data and cross-country regressions. See François andSpinanger 2001.

all firms are assumed to minimize costs, employingdomestic production factors (capital, labor, andland) and intermediate inputs from domestic andforeign sources to produce goods and services.These technologies are modeled as constant elastic-ity of substitution (CES) technologies defined overprimary inputs and Leontief processes definedover intermediate inputs. Products from differentregions are assumed to be imperfect substitutes inaccordance with the so-called Armington assump-tion. Prices on goods and factors adjust until allmarkets are simultaneously in (general) equilib-rium, which means that we solve for equilibria

in which all markets clear. Although we modelchanges in gross trade flows, we do not modelchanges in net international capital flows. (Thisapproach does not by any means preclude intro-ducing changes in the level of gross capital flows.)Trade liberalization in the goods sectors involves areduction of tariffs. This reduction involves in turna shift from the model base rates in table 12.6 tothe new bound rates. The new bound rates are gen-erally quite close to our calculations of averageaccession rates, also shown in table 12.6. Servicesector liberalization is modeled as a reduction intrading costs, reflecting the barrier reductions


reported in table 12.6. These are modeled asSamuelson iceberg costs.

For the motor vehicle sector, we want to modelthe status quo in a stylized though representativeway. One option is to implement imperfect compe-tition in the model, but such an approach does notreally deal with the primary issue at hand. Govern-ment policy has led to market segmentation, as wellas to price setting and regulation. Although the bestapproach to use in this situation is ultimately some-thing of a judgment call, we have chosen to focus onrealized cost efficiency for the sector. The currentcost structure of the motor vehicle industry reflectsthe net effect of a basket of policies. Like clothing inIndia and automobiles in Mexico before the NorthAmerican Free Trade Agreement (NAFTA) thestructure of the auto sector in China is a product ofregulated efficiency—that is, the general regulatoryand administrative environment. The critical issueis actually these collective inefficiencies, which fol-low from the full set of industrial policies. Atthe same time, an implication of the reforms thatare being implemented alongside WTO accessionseems to be restructuring and consolidation, lead-ing to an improvement in regulated efficiency.

What shape will changes in regulated efficiencygains take? The industry, through rationalization,may move collectively down the relevant costcurves. A comparison of current average plant scale(table 12.2) with a global norm closer to 350,000units per plant implies that average costs are roughly20 percent higher simply because of inefficientscale. The 20 percent figure is based on the distri-bution of current plants in table 12.2. In particular,if we apply the formula � ln (Average Cost) =−CDR · � ln(Quantity), where CDR, the cost-disadvantage ratio, is the inverse elasticity of scale,defined as

CDR = −Average Cost − Marginal Cost

Average Cost

and where CDR is between 0.125 and 0.135 (therange of values found in engineering studies), wecan calculate an average cost index for the industry.If such an index is 100 at 350,000 units per plant(the approximate norm for European and NorthAmerican plants), current plant structure yields acost index of roughly 120.

Data from interviews with industry representa-tives (Feenstra and others 2001) point to similar

cost savings, with expectations even higher—in therange of 25–30 percent cost reductions. The WorldBank (1993: 57) has described quite succinctly theexpected gains from reaching MES: “If this cost-volume relationship is applied to the Chinese auto-motive industry, the passenger car segment has acost disadvantage of 20 to 30 percent compared withthe international producers having MES. This costdisadvantage could be an understatement, however,as there are already eight producers in the market.”

These efficiencies are illustrated in figure 12.2,where we take sedan assembly as representative. Infigure 12.2, individual plants are plotted relative tothe cost:output mapping for an efficient plant. Thefigure depicts an expectation that significant costsavings may be realized if the government takesseriously its commitments to eliminate internalbarriers, and if it responds by encouraging arestructuring of the domestic industry around alimited number of motor vehicle groups.

This net cost effect is stressed here and sets thetreatment of motor vehicles apart from that ofother sectors in the model. We work with the lowerbound of these cost effect estimates. In particular,we focus on potential cost savings in the finalassembly of autos (stemming from consolidationand rationalization at the plant level), which yieldsa higher regulated efficiency level for the industry.In addition, the differential treatment of parts andfinished vehicles in the tariff schedule is tracked.

To set these issues in context, it may help to drawon developments in the automobile industry at anearlier point in time. That large gains can beachieved in rationalizing production and accord-ingly reducing costs in the automobile sector wasmost clearly demonstrated toward the beginning ofthe twentieth century. Henry Ford changed theparameters of producing automobiles radically byintroducing the assembly line in 1910. By 1914,“13,000 workers at Ford were producing 260,720cars. By comparison, in the rest of the industry, ittook 66,350 workers to make 286,770 [cars].”6 Butit was not just the shift in cost parameters that wassignificant in transforming the sector. Demand fac-tors were also important. Cars in the United States,as a result of Ford’s new production methods,moved from being a scarce good to one affordableby large segments of the U.S. population. China isalready in the process of moving into this phase, asis evident from figure 12.3.

Experiments and Results

The experiments are directed at full accession to theWTO for Greater China (Mainland China andChinese Taipei). The basic accession package coversthe changes in tariffs detailed in table 12.3. Forautomobiles, we model the following effects:7

• Tariffs on motor vehicles will decrease to 25 percent.

• Tariffs on auto parts will decrease from an aver-age of 23.4 percent to an average of 10 percent.

• Industry rationalization will implicitly involvethe elimination of internal regional barriers andwill allow for consolidation and rationalizationwithin the domestic market. Small, inefficientfactories will close.

To quantify the last effect, we take sedan produc-tion as representative. Given the typical scale ofdomestic production, we estimate that auto plantswill realize a 20 percent cost savings in assembly ifwe move plants to efficient scale (see note 2 and the


1.4

1.2

1.0

1.6

1.8

2.0

2.2

2.4

0 50,000 100,000 150,000 200,000 250,000 300,000 350,000

Average cost index

Production per plant

ShanghaiVolkswagen

FAWVolkswagen

Beijing Jeep

Tianjin XialiDaihatsu

Geely Group

FIGURE 12.2 Scale and Average Cost (with a CDR of 0.125)

Note: Moving from the current range of output to plants with an average of 300,000units implies a 20 percent cost savings.Source: Author’s calculations and data from table 12.2.

0

500

1,500

1,000

2,000

3,000

2,500

4,000

3,500

4,500

1901 1905 1909 1913 1917 1921 1925

United States 1901–1925

0

500

1,500

1,000

2,000

3,000

2,500

3,500

1984 1988 1992 1996 2000

China 1984–2002

FIGURE 12.3 Motor Vehicle Production in China and United States

Source: Bessum 2002; VDA, various issues.


TABLE 12.7 Impact of WTO Accession of Greater China on Output (percentchange)

A B C D � A � B E � A � C

Impact of Impact of Impact of Eliminating Mainland Mainland ATC Quotas China and China and

for WTO Chinese Chinese Members, Taipei TaipeiMainland Accession, Accession, Total Impact, Total Impact,

China, and without with Auto without with Auto Chinese Auto Sector Sector Auto Sector Sector Taipei Restructuring Restructuring Restructuring Restructuring

Wool 12.80 18.26 16.84 33.40 31.79Other natural fibers 12.11 17.86 16.41 32.13 30.51Primary food �0.43 �1.03 �0.92 �1.46 �1.34Other primary production �2.60 �3.57 �3.33 �6.07 �5.84Sugar �2.26 �7.93 �8.48 �10.01 �10.55Processed foods �1.02 �4.66 �4.74 �5.63 �5.71Textiles 13.93 32.00 30.57 50.39 48.75Clothing 50.26 75.46 73.03 163.65 159.98Leather goods �7.18 5.36 3.51 �2.20 �3.92Chemicals, rubber, and �2.03 �4.53 �4.27 �6.46 �6.21

refinery productsPrimary steel �3.99 �9.13 �7.86 �12.76 �11.54Primary nonferrous metals �5.42 �9.24 �8.94 �14.16 –13.87Motor vehicles and parts �4.11 �36.68 7.99 �39.28 3.54Electronics �5.06 �3.91 �4.43 �8.77 �9.26Other machinery �3.80 �5.39 �4.84 �8.98 �8.46

and equipmentOther manufactures �2.16 �0.34 0.14 �2.49 �2.02Wholesale and retail trade �0.25 1.39 1.93 1.14 1.68Transport services �1.94 �1.95 �1.39 �3.85 �3.31Communications �0.51 0.06 0.99 �0.45 0.47Construction 0.75 2.81 4.17 3.58 4.95Finance, insurance, �0.65 �0.40 0.22 �1.05 �0.44

and real estateCommercial services �0.78 �5.85 �5.41 �6.58 �6.15Other services 0.00 0.46 1.23 0.46 1.23

Source: Model estimates.

discussion in this overall section). This savings ismodeled at the assembly level.

The overall sectoral effects of the experimentsare presented in table 12.7. This table reportschanges in the quantity of output under our alter-native scenarios. Thus, as expected, the extension ofthe ATC phaseout to China and Chinese Taipeiimplies a rather dramatic expansion of the textileand clothing sectors—13.9 and 50.3 percent,respectively. There are important general equilib-rium effects, because the resources needed for this

experiment are drawn from other parts of the econ-omy, including the motor vehicle sector.

The results in columns B and C of table 12.7 arevery important for the motor vehicle sector. Theyreflect the incremental impact of China’s marketaccess commitments made as part of accession.Column B is a “business as usual” scenario, withoutthe restructuring discussed elsewhere in this chap-ter. It reflects a domestic motor vehicle industrythat continues to be fragmented, with favored pro-ducers in each region, small production runs, and

high costs. Such an industry is simply unable tocompete with imports, and indeed is very hard-hitby imports, with domestic production falling by36.7 percent. When that result is combined with theinitial impact of the ATC phaseout, a rather dra-matic retrenchment of the uncompetitive domesticindustry occurs in the face of imports (column D).

The contrast is offered in column C and the cor-responding total in column E. Column C representsthe elimination of internal barriers, rationalizationof plants (with smaller plants being closed), and a

realized efficiency gain of roughly 20 percent asscale economies are realized. This industry is muchdifferent from the one in column B. Productionactually goes up slightly (3 percent) in total, and theindustry emerges as a relatively competitive one,despite the loss of protection.

More information is provided on the differencesbetween the two scenarios in table 12.8 and fig-ure 12.4. The table expands on the informationoriginally provided in table 12.3, with a compara-ble breakdown corresponding to columns D and E


TABLE 12.8 Mainland China Motor Vehicle Market (millions of 1997 US$)

Impact of Impact ofMainland China Mainland China

and Chinese and ChineseTaipei Accession, Taipei Accession,

without Auto with Auto1997 Sector Sector

Benchmark Restructuring Restructuring

ValuesImported motor vehicles and parts, world prices 3,607.71 10,595.68 6,967.97Imported motor vehicles and parts, internal prices 4,806.39 12,080.71 7,995.72

Imported parts, internal prices 1,609.86 2,827.93 5,535.24Imported motor vehicles, internal prices 3,196.53 9,252.78 2,460.48

Domestic autos, intermediates and parts 32,812.46 19,401.89 24,249.56Domestic intermediate parts 10,896.15 4,493.95 5,189.12Industry consumption of motor vehicles 21,625.50 14,698.79 18,785.03Final consumption of motor vehicles 290.81 209.15 275.41

Indexes and sharesImport share of total auto parts (percent of value) 12.87 38.62 51.61Index of vehicle production 100.00 67.98 102.78Index of parts production 100.00 41.22 56.28

Sources: Author’s calculations and GTAP base data.

1997 benchmark Accession withoutrestructuring

Accession withrestructuring

Percent of value

51.61

100

20304050

38.62

12.87

FIGURE 12.4 Import Share of Total Auto Parts

in table 12.7; table 12.9 portrays the results just forvalue added. The most striking difference betweenthe two scenarios is the different impacts on inter-mediate parts production and final auto produc-tion, as illustrated in figure 12.4. Under the firstscenario, characterized by a domestic policy statusquo, imports of parts rise slightly and their share ofthe domestic parts market rises substantially. At thesame time, there is a dramatic surge in imports ofmotor vehicles, which displace more than one-thirdof existing domestic production. The overall mar-ket for parts falls because of the decline in domestic

vehicle production. Under the second scenario, thefinal assembly sector is rationalized, allowingthe sector to compete more directly with imports.The result is a shift to imported intermediates(rising to a market share of over 50 percent), a fallin domestic parts production (as they are displacedby imports), but a steady overall demand for parts.For the industry overall, although ground is stilllost to parts imports, sales of domestic vehiclesremain relatively steady in the face of imports.

One last view of the effect of accession ispresented in tables 12.10, 12.11, and 12.12, which


TABLE 12.9 Impact of Greater China Accession on Value Added (percent change)

A B C D � A � B E � A � C

Impact of Impact of Impact ofEliminating Mainland MainlandATC Quotas China and China and

for WTO Chinese ChineseMembers, Taipei Taipei Total TotalMainland Accession, Accession, Impact, Impact,

China, and without with Auto without with AutoChinese Auto Sector Sector Auto Sector SectorTaipei Restructuring Restructuring Restructuring Restructuring

Wool 11.55 12.97 10.62 26.02 23.40Other natural fibers 10.87 12.58 10.21 24.82 22.20Primary food −1.53 −5.46 −6.19 −6.91 −7.63Other primary production −3.68 −7.88 −8.47 −11.27 −11.84Sugar −3.34 −12.05 −13.36 −14.99 −16.25Processed foods −2.11 −8.92 −9.81 −10.85 −11.72Textiles 12.67 26.09 23.62 42.07 39.28Clothing 48.60 67.61 63.81 149.06 143.42Leather goods −8.20 0.64 −2.00 −7.61 −10.04Chemicals, rubber, and –3.11 −8.80 −9.37 −11.64 −12.19

refinery productsPrimary steel −5.06 −13.20 −12.76 −17.59 −17.17Primary nonferrous metals −6.46 −13.31 −13.79 −18.91 −19.36Motor vehicles and parts −5.17 −39.51 2.24 −42.64 −3.05Electronics –6.11 −8.21 −9.52 −13.82 −15.04Other machinery and −4.86 −9.62 −9.91 −14.02 −14.29

equipmentOther manufactures −3.24 −4.80 −5.19 −7.88 −8.26Wholesale and retail trade −1.35 −3.15 −3.50 −4.46 −4.80Transport services −3.03 −6.34 −6.64 −9.17 −9.47Communications −1.61 −4.42 −4.39 −5.96 −5.93Construction −0.36 −1.79 −1.38 −2.15 −1.73Finance, insurance, and −1.75 −4.86 −5.12 −6.53 −6.78

real estateCommercial services −1.87 −10.07 −10.45 −11.75 −12.13Other services −1.10 −4.04 −4.16 −5.10 −5.22



TABLE 12.10 Impact of Greater China Accession on Value of Exports (percentchange)

A B C D � A � B E � A � C




Wool –9.52 −18.73 −20.14 −26.47 −27.75Other natural fibers −4.67 −16.12 −16.67 −20.04 −20.56Primary food −4.16 −5.05 −7.29 −9.00 −11.14Other primary production −3.33 −2.98 −4.23 −6.21 −7.42Sugar −3.29 −5.19 −7.05 −8.31 −10.11Processed foods −4.99 −4.15 −5.96 −8.93 −10.65Textiles 6.18 32.74 32.01 40.93 40.16Clothing 87.81 80.03 77.48 238.11 233.33Leather goods −7.66 3.85 1.76 −4.11 −6.04Chemicals, rubber, and −3.29 −0.30 −1.01 −3.59 −4.27

refinery productsPrimary steel −5.56 −3.00 −3.99 −8.39 −9.33Primary nonferrous metals −5.87 −4.05 −5.36 −9.69 −10.92Motor vehicles and parts �12.33 16.17 392.33 1.85 331.64Electronics −4.56 −0.86 −1.75 −5.39 −6.24Other machinery and −6.69 −3.69 −4.52 −10.13 −10.91

equipmentOther manufactures −5.74 −2.57 −3.69 −8.16 −9.22Wholesale and retail trade −4.79 −5.19 −5.82 −9.73 −10.34Transport services −3.31 −2.28 −2.44 −5.52 −5.67Communications −5.05 −4.38 −4.90 −9.20 −9.70Construction −4.92 −4.34 −5.62 −9.05 −10.26Finance, insurance, and −5.33 −6.01 −7.10 −11.01 −12.05



reveal the export impacts observed under the vari-ous modeling scenarios. It is logical to expect someexport response, both because of the general liber-alization in trade and because pressure fromimports may force firms to seek other markets.China exports less than 4 percent of its productionin the motor vehicle sector, based on 1997 values.Of US$32 billion in production,8 only $1.3 billionin products is exported. To put this figure inperspective, Australia has a comparable level ofexports for an industry only one-third the size of

the Chinese industry. The export share for theRepublic of Korea is 10 times as large. China’s tradeis therefore well below global integration standardsas measured by exports.

Tables 12.10 and 12.11 compare export volumeeffects, as a percentage of base exports and in dollarterms. Clearly, restructuring accelerates the exportorientation of the industry, with a rapid growth inexports. Exports rise by about 300 percent, or$3.8 billion, reaching roughly 10 percent of pro-duction by value. Although this expansion seems


TABLE 12.11 Impact of Greater China Accession on Value of Exports (millions ofU.S. dollars)

A B C D � A � B E � A � C




Wool −4.95 −9.74 −10.47 −13.76 −14.42Other natural fibers −0.18 −0.63 −0.65 −0.78 −0.80Primary food −237.46 −288.39 −416.09 −513.86 −636.25Other primary production −153.94 −138.02 −195.70 −287.36 −343.13Sugar −4.58 −7.22 −9.82 −11.57 −14.08Processed foods −360.07 −299.33 −430.38 −644.47 −768.99Textiles 1,356.02 7,186.55 7,026.81 8,986.47 8,816.86Clothing 4,3997.88 40,096.19 38,820.21 119,303.97 116,907.51Leather goods −1,503.44 755.88 345.26 −805.49 −1,184.64Chemicals, rubber, and −560.97 −51.79 −171.59 −611.06 −726.91

refinery productsPrimary steel −225.93 −121.72 −162.30 −340.88 −379.21Primary nonferrous metals −130.28 −89.89 −118.97 −214.90 −242.26Motor vehicles and parts �141.58 185.76 4,505.86 21.28 3,808.80Electronics −1,365.27 −257.74 −524.98 −1,611.25 −1,866.29Other machinery and −2,145.66 −1,183.94 −1,450.56 −3,250.42 −3,499.21

equipmentOther manufactures −2,183.35 −977.14 −1,403.76 −3,104.39 −3,506.52Wholesale and retail trade −338.09 −365.87 −410.81 −686.43 −729.21Transport services −278.13 −191.84 −205.06 −463.62 −476.40Communications −19.84 −17.22 −19.27 −36.20 −38.14Construction −25.88 −22.84 −29.53 −47.59 −53.96Finance, insurance, and −26.63 −30.02 −35.49 −55.05 −60.23



dramatic, it should be kept in perspective. Exportshares are shown in table 12.12. Currently, auto-mobiles and parts are a small share of exports(0.6 percent in 1997), and they remain small (up to2 percent) even with the growth in automobileexports. In addition, most of the restructuringremains focused on the domestic market.

Summary and Conclusions

In this chapter we examine the interaction betweenregulated efficiency and China’s accession to the

WTO, and its impact on China’s motor vehicle sec-tor. The approach we take is general equilibrium,requiring application of a global general equilib-rium model.

We argue that regulatory reform and internalrestructuring are critical to any realized effect onthe auto sector. Such restructuring is representedhere by a cost reduction that follows from consoli-dation and rationalization. This representation issupported by a comparison of scale in a typicalauto plant in China with that in typical plants inNorth America and Europe, and by firm survey


TABLE 12.12 China Export Shares: Baseline and Scenario

Total Impact, Total Impact,without Auto with Auto

1997 Sector SectorExport Shares Baseline Restructuring Restructuring

Primary foods 0.046 0.033 0.033Textiles 0.084 0.098 0.097Clothing 0.102 0.303 0.298Motor vehicles and parts 0.006 0.004 0.019Electronics 0.133 0.100 0.099Other machinery and equipment 0.146 0.104 0.103Other manufactures 0.397 0.294 0.290Services 0.087 0.062 0.062

Sources: Author’s calculations; GTAP base data.

responses. We also draw on earlier similar estimatesof the benefits to be had by achieving minimumefficient scale and by radically structuring produc-tion more efficiently.

The net result is movement of costs towardglobal norms. Without such restructuring, thedomestic industry remains uncompetitive, andWTO accession will lead to a surge in imports offinal vehicles, though imports of parts will fall asproduction moves offshore. With restructuring, thefinal assembly industry will become competitive byworld standards, while the parts industry furtherintegrates with the global industry through exports(and through higher imports of parts). As high-lighted in figure 12.5, most automobile firms arelocated along the coast and thus are well placed totake advantage of global markets.

Two additional issues also require attention.First, as seen in table12.1, China’s ratio of popula-tion to motor vehicles is far higher than that foundin many other countries with similar income levels.Because this ratio reflects the impact of the existingpolicies, significant changes in these policies willshift demand back to what could be viewed asa normal pattern of consumption of cars, givenChina’s geographic attributes. Second, improvedaccess to car financing could further strengthen thedemand for cars. Whereas about 75 percent of U.S.and European automobile purchases are financedthrough loans, only 15 percent of car purchasesin China are financed this way (KPMG 2003, p. 7).Although China’s protocol of accession to the WTOstipulates that automobile finance will be liberal-ized, only draft legislation has been presented to

date.9 To the extent that this potential can betapped, the pressure on automotive firms to bemore productive and thus more competitive willbe all the greater. This, then, would be anotherfactor helping to ensure that the welfare gains cal-culated will come about.

The shortcomings of the analysis also need to behighlighted. We have worked with a very stylizedmodel, even though we feel it captures importantelements of the real world. Although restructuringhas positive overall implications for the motorvehicle industry, significant adjustment costs notpointed to explicitly in the model are bound toemerge. Even if value added is preserved within thesector, there will most likely be a dramatic reloca-tion of jobs toward a limited number of plants,with job losses in the other, smaller plants. The cur-rent regional scattering of final auto production(figure 12.5) will be replaced by a more geographi-cally concentrated pattern. Parts production willalso tend to concentrate. To the extent that partssuppliers are able to supply regional markets, thisconcentration is likely to mean an intensification ofthe clustering in the coastal regions, with partsshipments to Japan, Korea, the United States, andother regional centers of production.10 This resultis broadly consistent with the findings and recom-mendations of Takayasu and Mori (2004), whostress that the future of the automobile industry inthe entire region, and not just China, hinges on theability of local governments to adopt policies thathelp to better integrate national manufacturers intoregional and global supply chains. Overall, a rela-tively large share of value added is kept intact with


FIGURE 12.5 Location of Automobile Production in China

Location of foreign production Production capacities in provinces

Producer Foreign producer Capacity(cars/year)

Production,2002

123456789

1011121314151617181920212223242526272829303132

SAIC VWSAIC GMFAW VWFAW ToyotaDongfeng PSADongfeng HondaDongfeng YulongTianjing ToyotaJIangsu NanyaSAIC CheryZehjiang JiliChongqing Chang'an SuzukiChang'an FordDengfeng Yueda KiaFAW HainanBeijing HyundaiChina Guizhou Aviation Ind.Shenyang Brilliance JunbeiHarbin HafeiShangdong YanteiSoutheastBeijing JeepJinbei GMHunan ChangfengZhengzhou NissanRongcheng HuataiJiangxi FuqiTianjing HualiSAIC GM WulingSanjiang RenaultChengdu FAWYizhong

VWGMVWToyota/MazdaPSA/CitroënHondaNissan/YulongToyotaFiatDaewooDaewoo (geplant)Suzuki/YanjinFordKiaMazdaHyundaiWanhong/ChenchangBMW (mid-2003)MitsubishiGMZhonghuaDaimlerChryslerGMMitsubishiNissanHyundaiGolden LionGolden LionGMRenaultToyotaSAIC/RDS

450,000100,000270,00070,000

150,00060,00060,00030,000

100,00060,000

150,000150,00050,00050,00050,00030,00010,000

200,00030,00050,00060,00085,00030,00030,00030,00020,00020,00020,000

150,00030,0005,000

10,000

278,890111,623158,65430,16584,37859,02438,8972,147

23,39349,39747,44367,846

n.a.20,08011,9891,3561,831

n.a.14,577

n.a.16,9359,0523,751

15,067n.a.n.a.n.a.n.a.n.a.n.a.n.a.n.a.

Capacity(cars/year)

Production,2002

AnhuiBejingFujianGuandongGuangxi ZhuangGuizhouHainanHeilongjangHenanHubeiJiangsuJilinLiaomingShandongShanghaiShanxiSichuanTianjingZehjiang

60,000115,00080,000

120,000150,00010,00050,00030,00030,000

180,000130,000340,000230,00080,000

550,00050,000

205,00050,000

150,000

49,39710,40816,93597,921

n.a.1,831

11,98914,577

n.a.84,37838,460

188,8193,751

n.a.390,51320,08067,8462,147

47,443

Total 2,380,000 146,495

Other foreignfirms

Employees No. ofplants

BoschKolbenschmidtMichelinZF/Sachs

3,6001,5004,0002,100

6222

Total 11,200 12

n.a. Not applicable or available.Sources: VDA, various issues; Bessum 2002.

restructuring. From an employment perspective,value added results closely track the impact ofrestructuring on employment. Table 12.9, there-fore, identifies a range of effects on auto sectoremployment, from �43 percent (without restruc-turing) to �3 percent (with restructuring). Need-less to say, it is important that the structure of theindustry be rationalized.11

Notes

1. Smith and Venables (1991) dealt with the benefits arisingfrom such a situation, where access across the domestic marketis improved, when examining the implications of the formationof the European Union’s common internal market for domesticand foreign suppliers.

2. See, for example, China Online (2001). As WTO member-ship approached in late 2001, the opinions of the industry andrelated ministries, as reflected in the Chinese press, hinged criti-cally on whether restructuring of the domestic industry wouldbe allowed to proceed. Thus a report in Touzi Yu Hezuo (summa-rized in China Online 2001) stressed the expected injury to theindustry, while the industry itself was indicating optimism thatit could realize significant cost reductions, allowing it to remaincompetitive with imports (Feenstra and others 2001). In themeantime, price cuts by foreign producers in China havebecome commonplace, with some of them induced by increasedimport competition and others by more intense domestic com-petition. Buick, for example, reduced its prices on its domesti-cally produced models by 12 percent, and VW lowered Passatprices by 6.5 percent (indiacar.net, May 3, 2002). But, even moreimportant, nearly all major foreign producers have announcedplans to sizably establish or increase production capacities. Arecent major manufacturer to do so was DaimlerChrysler; inSeptember 2003 it finally ratified plans to establish facilities toproduce Mercedes C and E models in China (InternationalHerald Tribune, September 9, 2003).

3. According to unnamed sources in Hong Kong, 100,000 ormore sedans may have been imported into the country. Many ofthe smuggled cars tend to be luxury models.

4. In the past, foreign investors seeking partners frequentlyhad to go to provinces other than those on the coast (seeoverview of policies). These provinces often tried to ensure that“buy local” conditions prevailed for local authorities. InShanghai, taxi regulations “by chance” stipulated specificationsthat could be filled only by a Shanghai-VW model.

5. The model files themselves, along with the technicalannex describing the model, can be downloaded from www.intereconomics.com/francois. The model is implemented inGEMPACK.

6. See http://inventors.about.com/gi/dynamic/offsite.htm?site=http://www.wiley.com/products/subject/business/forbes/ford.html.

7. We did not model the impact of the elimination of quotasfor imported cars. Data on the value of quota rents could not beproduced. Assuming that positive quota rents did exist thewelfare gains would be all the greater. However, the qualitativemessage would not change.

8. All dollar amounts are 1998 U.S. dollars.9. Nonetheless, some major car companies (VW and

Ford) have reached agreements with Chinese banks (KPMG,2003, p. 7). According to the International Herald Tribune

(October 6, 2003), China has opened up this sector in line withits WTO commitments.

10. European manufacturers have already established12 plants in China, and one large American company (Delphi) isshifting from Mexico.

11. Again, this is a message not unique to China. As empha-sized by Doner, Noble, and Ravenhill (2004), the industries inChina, Malaysia, the Republic of Korea, Taiwan (China), andThailand all face significant pressure to restructure. In China’scase, like that of the Republic of Korea, the relatively large localmarket offers the scope for increased presence of foreign pro-ducers through joint ventures and foreign direct investment. Ifmanaged properly (that is, if realized in a context of integrationof China’s internal markets), net efficiency boosts as modeledshould be realized, with results as modeled in this chapter.

References

The word processed describes informally reproduced works thatmay not be commonly available through libraries.Bessum, F. 2002. “Global Car Production Statistics.” March.

Available at www.geocities.com/MotorCity/Speedway/4939/carprod.html.

China Online. 2001. “How WTO Membership Could AffectChina’s Auto Industry.” Available at www.chinaonline.com/issues/wto/NewsArchive/secure/2000/january/b200010319-3-SS.asp.

Chinese Motor Vehicle Documentation Center. 2002. Catalogueof the Present Chinese Motor Car Production. 2d ed.Aldeboarn: Netherlands.

Doner, R., G. Noble, and J. Ravenhill. 2004. “ProductionNetworks in East Asia’s Automobile Industry.” World Bank,Washington, D.C. Processed.

Feenstra, R., D. Sperling, L. Branstetter, E. Harwitt, and W. Hai.2001. “China’s Entry to the WTO: A View from the AutoIndustry.” Mimeo, University of California Davis. Processed.

François, J. F. 2000. “Assessing the Results of General Equilib-rium Studies of Multilateral Trade Negotiations.” UNCTAD/ITCD/TAB/4, UNCTAD Policy Issues in International Tradeand Commodities Study Series. UN Conference on Tradeand Development, Geneva, October.

François, J. F., and D. Spinanger. 2001. “Greater China’s Acces-sion to the WTO: Implications for International Trade/Production and for Hong Kong.” Paper prepared for HongKong Trade Development Council, December.

———. 2004. “Regulated Efficiency, WTO Accession, and theMotor Vehicle Sector in China.” World Bank EconomicReview; forthcoming.

François, J. F., and A. Strutt. 1999. “Post Uruguay RoundTariff Vectors for GTAP Version 4.” Erasmus University,Rotterdam.

Huang, Y. 2002. “Between Two Coordination Series: AutomotiveIndustrial Policy in China with a Comparison to Korea.”Review of International Political Economy 9(3): 538–73.

KPMG. 2003. China Automotive and Component Parts Market.Hong Kong, August.

Li, Shantong. 2003. “Survey Report on Regional Protection.” InWorkshop on National Market Integration, proceedings ofDevelopment Research Centre Workshop, September 6.

Lopez de Silanes, F., J. Markusen, and T. F. Rutherford. 1994.“Complementarity and Increasing Returns in IntermediateInputs.” Journal of Development Economics 45: 133–51.

McDougall, R., ed. 2001. The GTAP Database—Version 5. GlobalTrade Analysis Center, Purdue University.


Smith, A., and A. Venables. 1991. “Economic Integration andMarket Access.” European Economic Review 35: 388–95.

Takayasu, K., and M. Mori. 2004. “The Global Strategies ofJapanese Vehicle Assemblers and the Implications for theThai automobile Industry.” World Bank, Washington, D.C.Processed.

U.S. and Chinese Academy of Engineering. 2003. Personal Carsand China. National Academies Press. Washington, D.C.

VDA (Verband der Automobilindustrie). Various issues.Tatsachen und Zahlen. Frankfurt.

World Bank. 1993. China Industrial Organization and EfficiencyCase Study: The Automotive Sector. Report no. 12134-CHA.Washington, D.C.

———. Various issues. World Development Indicators.Washington, D.C.

Zhang, W., and R. Taylor. 2001. “EU Technology Transfer toChina: The Automobile Industry as a Case Study.” Journal ofAsia Pacific Economy 6(2): 261–74.


of basic information such as trade flows and patternsof production and consumption. What standardmodels such as GTAP (Hertel 1997; www.gtap.org)do not do, however, is incorporate the nonstandardfeatures of China’s partially reformed economy,where many imports enter duty-free if used inthe production of exports and labor market poli-cies result in serious barriers between urban andrural areas.

Like Ianchovichina and Martin (2001) andWang (2003), the analysis here explicitly allows forthe duty exemption arrangements that result inclose to half of China’s imports entering as duty-free inputs into the production of exports. Recentwork by Sicular and Zhao (2004) is drawn on to represent imperfect labor mobility and labormarket distortions. This chapter extended our ear-lier work reported in Ianchovichina and Martin(2001) by moving to the GTAP version 5 database(Dimaranan and McDougall 2002) based on 1997data rather than 1995 data; by incorporating

Trade policy reforms such as those flowing fromaccession to the World Trade Organization (WTO)lead directly to changes in policy instruments suchas tariffs, nontariff barriers, and coverage of traderules. The main policy concerns, however, are withthe impacts on such economic variables as prices;output, employment, and trade volumes; factorreturns; and household incomes. In this chapter weestimate the impacts on these key economic vari-ables of China’s accession to the WTO in 2001 as aguide to policy and as a basis for subsequent analy-sis at the household level.1

The obvious instrument for performing thistype of analysis is the computable general equilib-rium (CGE) model. Many such models now exist,and a cottage industry has emerged in estimatingthe impacts of trade reform in China (Gilbert andWahl 2001).2 The availability of the internationallystandard database of the Global Trade AnalysisProject (GTAP) has facilitated such modeling andreduced the burden involved in obtaining estimates

211

13

Economic Impacts ofChina’s Accession to

the WTO

Elena Ianchovichina and Will Martin

We thank Kym Anderson, Hana Polackova Brixi, Louise Fox, Thomas Hertel, T. N. Srinivasan, Alan Winters,and three anonymous reviewers for helpful comments, Prashant Dave and Zhi Wang for their generosity inproviding data. Elena Ianchovichina is Economist with the Economic Policy Unit of the Poverty Reductionand Economic Management Network at the World Bank; her email address is [email protected]. Will Martin is Lead Economist at the World Bank; his email is [email protected].

improved estimates of protection in agriculture(Huang, Rozelle, and Min 2004)3 and services(Francois and Spinanger 2004); by using measuresof liberalization based on the final, multilateralagreement; by taking into account the restructuringof the automobile sector (Francois and Spinanger2004); and by simulating the consequences ofmajor labor market reform. This last issue is a par-ticularly critical area for China, and there have beenfew simulation studies.4

This chapter first discusses the methodologyand then the policy changes associated with China’sWTO accession and the results of the simulationanalysis. It then considers some possible comple-mentary policy actions, such as reducing barriersto rural outmigration and expanding access toeducation.

Methodology

The standard GTAP model5 was adjusted to incor-porate China’s important export processingarrangements. Ianchovichina (2003) documentsthe approach used and shows that failing toaccount for China’s duty exemptions in analyzingWTO accession overstates the increase in China’strade flows by 40 percent and the increase inexports of selected sectors by 90 percent. Theadjusted model (GTAP-DD) also incorporatessome of China’s key labor market mechanisms andinstitutions that related research has shown mayhave a major influence on the impacts of WTOaccession (Sicular and Zhao 2002, 2004).

Export Processing Arrangements

Export processing arrangements in China takemany forms. Most arrangements allow firms pro-ducing goods for export to import intermediateinputs at world prices. These arrangements wereincorporated in the GTAP-DD model used in thisstudy by creating two activities for each sector. Forsectors covered—or potentially covered—by exportprocessing arrangements, one activity is specializedin production for export, and one is specialized inproduction for the domestic market.6 This separa-tion is preferable to representations based on asingle sector producing differentiated products fordomestic and export markets because it allows thetwo sectors to use different input mixes, and it

allows export-oriented activities to use much moreimport-intensive means of production. It providesa reasonably realistic depiction of China’s traderegime in the 1990s, when duty exemptions wereused to facilitate exports while protection in therest of the economy remained fairly high.7

The tax arrangements for export processing(duty and value-added tax [VAT] exemptions onimported intermediate inputs and VAT refunds onexports) discouraged export-oriented firms fromusing domestic intermediate materials and sellingin the local market.8 Furthermore, the vast majorityof exports were produced using imported interme-diates that were either exempt from duties or eligi-ble for refunds on taxes paid.9 China Customs datafor 2000 show that 60 percent of imports enteredduty free, 41 percentage points of which wereimports used for export processing, 13 percentagepoints were capital goods, and 6 percentage pointswere goods in special categories, such as materialsused by research institutes. Rough calculations basedon input-output information from the GTAP ver-sion 4 database (McDougall, Elbehri, and Truong1998) and data from China Customs suggest that26 percent of imports were used to produce for thedomestic market and only about 3 percent wereused to produce ordinary exports.10

Intermediate inputs for domestic and export-oriented activities were initially estimated by allo-cating them to each sector in proportion to sales inexport and domestic markets. However, this yieldedunsatisfactory results, with the database showingmuch less use of imported inputs in the export sec-tor than the reported imports of duty-free interme-diate inputs for export production obtained fromChina Customs (Li Yan, China Customs, personalcommunication). To deal with this, increased use ofimported intermediates was allowed for in exportactivities in accordance with the price changesinvolved in providing duty exemptions, and theelasticities of substitution between domestic andintermediate goods in the model.11 This modifica-tion more than doubled the import-intensity of theexporting activities and reduced that of the domes-tically oriented activities.

China’s Labor Market Policies

Perhaps the central labor market issue for theanalysis is the barriers to mobility between rural


and urban activities. Taking up employment in anurban area is inhibited by the need to obtain anurban residence permit (hukou). In addition, work-ers tend to be reluctant to permanently cut theirties with the rural sector because it is not possibleto sell the land to which a rural family has use rights(Hussain 2004). Many workers move temporarilyfrom rural to urban areas, although restrictions arefrequently imposed on such movements, and socialwelfare benefits such as health care and schoolingfor children enjoyed by urban residents are typi-cally not available to temporary migrants. While it is sometimes possible to overcome these prob-lems by purchasing an urban residence permit, thisimposes an additional cost on migrants from ruralto urban areas, a group with particularly limitedaccess to capital. Finally, as in all countries, rural–urban labor mobility is also inhibited by the sector-specific nature of farmers’ human capital and areluctance to cut family ties by migration to urbanareas.

The per capita income of agricultural workers isonly about one-third that of urban workers (WorldBank 2002). Not all of this difference can be attrib-uted to barriers to mobility between rural andurban areas, however. Urban workers typically havehigher skills, work more intensively, and face highercosts of living than rural workers (Sicular and Zhao2004).

To capture the effects of barriers to mobility,the model allows for both imperfect transforma-tion between unskilled workers in agriculturaland unskilled nonagricultural employment, andan implicit tax on nonagricultural employment.The imperfect transformation representation is de-signed to reflect the substantial differences in thecharacteristics of unskilled workers in rural andurban areas, and the ability, at a cost, to transformagricultural workers into nonagricultural workersthrough training, experience, and the creation ofnonagricultural jobs in rural areas. The tax is de-signed to reflect the pure policy-induced barriersbetween rural and urban workers, such as therequirement for a residence permit in urban areasand the barriers to mobility created by the inabilityto sell farm land. It is specified as a barrier thatraises the cost of labor to urban employers, withurban workers receiving the tax-inclusive wage.

The imperfect transformation between agricul-tural and nonagricultural workers is represented

using a constant elasticity of transformationbetween workers in agriculture and workers inother sectors in the following simple manner:

LNF/LF = α(WNF/WF)σ, (13.1)

where LNF is the number of nonfarm workers, LF isthe number of farm workers, α is a constant term,WNF and WF are nonfarm and farm wages, respec-tively, and σ is the elasticity of transformation. Thevalue of σ was set at 1.32, based on Sicular andZhao’s (2002) estimates of this parameter.12 The taxreflecting pure policy-induced barriers betweenrural and urban wages was estimated at 34 percentby Shi (2002). While there is some level of unem-ployment in China, particularly associated withstate enterprise reform and rural–urban migration,much of this unemployment is transitional, becauseof the weakness of the social safety net (Hussain2004). Given the long-run focus of the analysis here,total employment is treated as exogenous.

Trade Policy Reforms

This section examines the implications of tradepolicy reforms in China and in its trade partners inthe years leading up to accession.

Changes in China’s Trade Policies

During the 1990s China made substantial progressin reducing the coverage of nontariff barriers, low-ering tariffs, and abolishing the trade distortionscreated by the exchange rate regime. Lardy (2002)estimates that the number of tariff lines subject toquotas and licenses fell from 1,247 in 1992 to 261 in1999. By 2001, 257 tariff lines were covered by a combination of licenses and quotas and 47 bylicenses only, while 245 were subject to designatedtrading and 84 to state trading. Tendering andother registration requirements, primarily for ma-chinery and electrical products, covered an addi-tional 120 tariff lines. By 2001, nontariff barrierscovered 664 tariff lines, or less than 10 percent oftariff lines (see Appendix table 13.A.1), with over athird of these being subject to designated trading,one of the less intrusive forms of quantitativerestriction employed in China.

Data on nontariff barriers frequency alone maybe misleading because of the enormous varia-tions in the importance of tariff lines. To give some

Economic Impacts of China’s Accession to the WTO 213

indication of the potential importance of nontariffbarriers, the import coverage of the key nontariffbarriers was calculated using data on nontariff bar-rier coverage of tariff lines and on import data bytariff line.

The import coverage of all nontariff barriers inChina fell from 32.5 percent in 1996 (World Bank1997, p. 15) to 21.6 percent in 2001 (see table 13.1).Coverage of import licensing and quotas fell from18.5 percent in 1996 to 12.8 percent in 2001, andcoverage of state trading from 11 to 9.5 percent.The import coverage of tendering requirementsfell particularly rapidly, from 7.4 percent in 1996 to2.7 percent in 2001. For more details on coverageby type and commodity, see Appendix table 13.A.2.

Oil was by far the most important import sub-ject to nontariff barriers, accounting for almost halfthe value of imports subject to any nontariff barri-ers (see Appendix table 13.A.3). Ferrous metals,subject to designated trading arrangements, werethe second most important category. Imports of oiland oil products accounted for 84 percent of totalimports subject to state trading.

The average protective impact of the completeset of nontariff barriers in China was estimated(very crudely) to be 9.3 percent in the mid-1990s(World Bank 1997), with most of the protectiveeffect arising from licensing and quota-constrainedgoods. The protective effect of these nontariffbarriers has clearly declined since then because of a

number of factors such as the progressive phase-out of nontariff barriers, a standstill on new non-tariff barriers during the accession process, the gen-eral WTO prohibition against nontariff barriersafter accession, and a likely reduction in the severitywith which many of these measures have beenadministered. A simple rule of thumb that protec-tion provided by nontariff barriers declines withtheir import coverage would suggest that theirprotective impact has fallen to about 5 percent.However, given the large margin of uncertaintyassociated with this measure, the analysis herefocuses only on tariff liberalization, implying thatthe results should be taken as a lower bound to theoverall impact of liberalization.

The pace of tariff reform in China was also rapidduring the 1990s. While average tariffs were veryhigh in the early 1990s, they fell sharply after 1994(table 13.2). A significant tariff reform in October1997 reduced average tariffs well below 20 percent.Three subsequent reductions at the beginning of1999, 2000, and 2001 further reduced tariffs on awide range of items. The progressive reductions intariffs between 1992 and 2001 lowered average tar-iffs by two-thirds, with larger than average cuts in the manufacturing sector, thereby ensuring thatthe future reductions in tariffs required under theWTO accession agreement are much smaller thanthe reductions occurring before accession. Anotherimportant feature of the reforms has been a sub-stantial reduction in the dispersion of tariff rates—with the standard deviation falling from 32.1 per-cent in 1992 to 10 percent in 2001.

Examination of weighted average applied tariffsfor 1995 and 2001 and after implementation of thefinal tariff bindings agreed in the accession sched-ule suggests13 that substantial merchandise tradeliberalization occurred in China over the period1995–2001 (table 13.3). Weighted average tariffsdropped substantially for wheat, beverages andtobacco, textiles, apparel, light manufactures,petrochemicals, metals, automobiles, and electron-ics. Analysis by Huang, Rozelle, and Min (2004)suggests that some agricultural commodities suchas vegetables and fruits, livestock and meat, and ricefaced negative protection in 1995, generally as aresult of restrictions on exports. Protection onthese commodities rose (or negative protectionfell) over the period 1995–2001. Accession is notexpected to lead to a significant fall in protection on


TABLE 13.1 Import Coverage of NontariffBarriers (percent)

Barrier 1996 2001

Licenses and quotas 18.5 12.8Tendering 7.4 2.7Licensing only 2.2 0.5State trading 11.0 9.5Designated trading 7.3 6.2Any nontariff barriers 32.5 21.6No nontariff barriers 67.5 78.4Total 100 100

Note: For 1996 nontariff barrier coverage, thetrade weights used were for 1992, while for2001 the trade weights used were for 2000.Source: For 1996, Lardy 2002; for 2001, MeiZhen of the Ministry of Foreign Trade andEconomic Cooperation during an internship atthe World Bank, using data from WTO.


TABLE 13.2. China’s Average Statutory Tariff Rates (percent)

All Products Primary Products Manufactures

Year Simple Weighted Simple Weighted Simple Weighted

1992 42.9 40.6 36.2 22.3 44.9 46.51993 39.9 38.4 33.3 20.9 41.8 44.01994 36.3 35.5 32.1 19.6 37.6 40.61996 23.6 22.6 25.4 20.0 23.1 23.21997 17.6 18.2 17.9 20.0 17.5 17.81998 17.5 18.7 17.9 20.0 17.4 18.51999 17.2 14.2 21.8 21.8 16.8 13.42000 17.0 14.1 22.4 19.5 16.6 13.32001 16.6 12.0 21.6 17.7 16.2 13.0After accession 9.8 6.8 13.2 3.6 9.5 6.9

Source: World Bank 1999, p. 340 to 1998. Authors’ calculations for tariff lines with imports in 1999. Tradedata come from COMTRADE. Protection data for 1999 to 2001 come from CDS Consulting Co. and afteraccession from China’s WTO final offer.

most agricultural commodities after 2001. Importprotection is expected to remain unchanged formost commodities except oilseeds, sugar, and dairyproducts.

Protection will continue to fall for all other mer-chandise commodities, with especially big cuts forprocessed food, beverages and tobacco, automo-biles, electronics, and other manufactures. Francoisand Spinanger (2004) conclude that the automo-bile sector liberalization will be accompanied bymassive restructuring to realize economies of scaleand improve structural efficiency, perhaps increas-ing productivity by 20 percent during the accessionperiod (2001–07).14

A key element of China’s accession agreement isthe abolition of agricultural export subsidies.Huang, Rozelle, and Min (2004) estimate that therewas a 32-percent export subsidy on feedgrains anda 10 percent export subsidy on plant-based fibers in2001 (particularly cotton).

In addition to China’s barriers on merchandisetrade, border measures and domestic regulationson domestic service sectors and trade in these serv-ices have reduced the efficiency of services sectors.China has made substantial commitments to openits services sectors,15 and critical services such astelecommunications, logistics, and finance arelikely to benefit from inward foreign direct invest-ment and rising productivity as they are restruc-tured. Based on work by Francois and Spinanger(2001) reported in Francois and Spinanger (2004),

these measures are represented here as barriers totrade in services expressed in ad valorem terms.Following Francois and Spinanger (2004), theimpact of accession is represented as halving thenontariff barriers to services trade. Efficiencyimprovements in the services sectors are not mod-eled since there were no reliable estimates of thelikely productivity gains at the time this researchwas conducted.16

Changes in China’s Trade Partners’ Policies

The arrangements for textiles and clothing are aparticularly important element of China’s acces-sion. Unlike most other developing economy ex-porters, China was excluded from the liberalizingelements of the Uruguay Round Agreement on Tex-tiles and Clothing. This means that prior to acces-sion China did not benefit from the integration oftextile and clothing products into the GeneralAgreement on Tariffs and Trade (GATT) or fromthe increases in quota growth rates provided forunder this agreement. That placed upward pressureon the prices of these quotas in China, raising thecosts of exporters just as an equivalent export taxwould.17 Under its accession agreement China ben-efited immediately from the integration of textilesand clothing into the GATT, leading to the aboli-tion of quotas and increases in quota growth ratesthat have occurred since 1994 (WTO 1994). Allquotas are to be phased out by 2005. Importing

economies will be allowed to introduce special tex-tile safeguards during the period 2005–07, but thesewill be effective for only one year at a time.

The accession agreement includes a TransitionalProduct-specific Safeguard mechanism that allows

China’s trading partners to take safeguard actionsunder rules that are more permissive of protectionthan the usual WTO safeguard rules (Messerlin2004). These provisions have the potential forintroducing a new form of protection against


TABLE 13.3 China’s Import Protection before and after WTO Accession (tariff ortariff equivalent, percent)

China Taiwan, China

Product 1995 2001 Post-Accessionb 1997 2001 Post-Accessionb

AgricultureRice −5.0 −3.3 −3.3 2.2 0.0 0.0Wheat 25.0 12.0 12.0 6.5 6.5 6.5Feedgrains 20.0 32.0 32.0 1.0 1.0 0.0Vegetables and fruits −10.0 −4.0 −4.0 35.7 36.9 16.0Oilseeds 30.0 20.0 3.0 1.8 0.8 0.2Sugar 44.0 40.0 20.0 21.9 25.8 22.7Plant-based fibers 20.0 17.0 20.0 0.0 0.0 0.0Livestock and meat −20.0 −15.0 −15.0 7.5 6.5 4.0Dairy 30.0 30.0 11.0 16.6 9.3 5.9Processed food 20.1 26.2 9.9 14.9 14.2 9.9Beverages and tobacco 137.2 43.2 15.6 48.1 22.0 13.0

Total 4.8 7.6 3.6 9.1 6.9 4.6

ManufacturingExtractive industries 3.4 1.0 0.6 5.5 5.5 4.1Textiles 56.0 21.6 8.9 6.1 6.3 5.6Apparel 76.1 23.7 14.9 12.8 13.4 11.2Light manufactures 32.3 12.3 8.4 4.0 4.1 3.4Petrochemicals 20.2 12.8 7.1 4.2 4.2 2.9Metals 17.4 8.9 5.7 4.0 3.8 1.5Automobiles 123.1 28.9 13.8 23.9 21.5 13.3Electronics 24.4 10.3 2.3 2.9 0.5 0.3Other manufactures 22.0 12.9 6.6 4.4 3.3 2.1

Total 25.3 13.5 6.9 6.3 5.2 3.5Total merchandise tradea 24.3 13.3 6.8 6.5 5.2 3.6

ServicesTrade and transport 1.9 1.9 0.9 1.3 1.3 0.7Construction 13.7 13.7 6.8 5.9 5.9 2.9Communications 9.2 9.2 4.6 9.2 9.2 4.6Commercial services 29.4 29.4 14.7 3.7 3.7 1.9Other services 24.5 24.5 12.7 7.1 7.1 3.5

Total 10.3 10.3 5.2 3.2 3.2 1.6

aThose estimates are based on trade weights for the respective years. If trade weights for 2000 at the six-digit level of the harmonized system are used, the total weighted average tariffs in 2001 and 2007are 12.2 percent in 2001 and 6.3 percent in 2007 for China and 4.5 percent and 3.1 percent forTaiwan, China. bApplied rates at the end of the implementation period were estimated as the lesser of the bindingsand 2001 applied rates. In virtually all cases the bindings were lower than the applied rates.Source: Authors’ calculations based on agricultural protection data from Huang, Rozelle, and Min 2002;manufacturing protection data from GTAP in 1995, from CDS Consulting Co. in 2001 and fromChina’s WTO final offer for protection after accession; services protection data from Francois andSpinanger 2004.

China. This potential danger needs to be weighedagainst the substantial gains to China from hertrading partners being required to follow WTOrules in implementing contingent protection meas-ures against China. For simplicity, these gains andlosses are assumed to cancel each other out.

China’s accession also triggered a liberalizationof its partners’ trade policies. The average tariffs ofTaiwan (China) were estimated to fall by almost1.5 percentage points, from 4.5 percent in 1997 to3.1 percent after accession. Taiwan (China) commit-ted to tariff reductions on thousands of industrialand agricultural product lines, a phase-out of tariffson a number of products as part of the Zero-for-Zero program of the Uruguay Round, and reduc-tions in tariffs on chemical products as part of theChemical Harmonization program. Tariffs on thevast majority of products related to informationtechnology were also reduced in 2000 and, onceWTO accession commitments are implemented,the tariff on electronic products will fall to 0.3 per-cent (see table 13.3).

Taiwan (China) made horizontal and sector-specific commitments in the following servicesectors: business, communication, construction,engineering, distribution, education, environmen-tal, financial, health, social, transport services, andtourism and recreation. Like Francois and Spinanger(2002), we treat the WTO accession commitmentsof Taiwan (China) as halving nontariff barriers totrade in services.

Experimental Design

The impact of accession is evaluated here in thedynamic context of the growth and structuralchange expected in China and its trading partnersduring the period up to 2007, when almost all ofthe changes associated with accession will havecome into effect. A baseline scenario is constructedunder which the economies of the world growand experience the manifold structural changesassociated with economic growth up to 2007(table 13.4 and Appendix table 13.A.5). The GTAPmodel includes key elements such as changes indemand patterns as incomes rise, changes in theindustrial structure associated with changes in thestock of capital per worker, and changes in worldprices resulting from changes in world supply anddemand that allow the model to capture keychanges in the world economy over this period. The

baseline broadly replicates World Bank projectionsfor overall growth in each region and uses projec-tions of factor input growth and a residually deter-mined level of total factor productivity growth toensure consistency between the two.

The model considers the effects of the WTOaccession agreement signed at the Doha ministerialmeeting in November 2001 on protection pre-vailing in 2001. In addition, it separately takes into account the liberalization in China during1995–2001 since much of the liberalization duringthe 1990s was influenced by China’s desire to pre-pare for the type of trade regime needed for WTOaccession and to establish the credibility of its com-mitment to an open economy. These sharp reduc-tions in protection are unlikely to have occurredwithout the prospect of accession to WTO, andthey have been locked in by China’s WTO commit-ments. Thus results are presented for both theaccession period (2001–07) and the entire liberal-ization period (1995–2007).

The analysis starts with 1995 tariff levels be-cause 1995 was a major turning point in thenegotiations—the closing of the door on China’sattempt to enter the world trading system byresuming its status as a Contracting Party to theGATT. As Long (2000, p. 43) emphasized, Chinafocused more strongly on commercial considera-tions after 1995 than it had previously done—andits trading partners also strongly emphasized thecommercial aspects of the negotiations. To capturethe implications of WTO accession, 1997 protec-tion data for China in the benchmark data (GTAPversion 5) are adjusted to 1995 levels to obtain theinitial base.18 For Taiwan (China) liberalization isconsidered to have begun in 1997, the year forwhich tariff data are available in GTAP version 5.

Two experiments are conducted to evaluate theimpact of WTO accession. The first assesses theimpact of the fall in tariffs from 1995 to 2001 levelsand the restructuring of the automobile sectoraccompanying the reductions in tariffs on automo-biles and automobile parts during this period. Thesecond assesses the impact of the fall in tariffsfrom 1995 to post-accession (2007) tariff levels,liberalization of the services sectors, continuedrestructuring of the automobile sector, removalof quotas on China’s clothing and textiles exports,and removal of China’s agricultural export subsi-dies. The productivity shock designed to capturethe restructuring of the automobile sector is



TABLE 13.4 Projected Growth in Factor Inputs and Total Factor Productivity 1997–2007 (percent)

ManufacturingUnskilled Skilled Total Factor

Trading Partner Population Labor Labor Capital Productivitya

North America 11 11 12 49 High(1.05) (1.08) (1.11) (4.07)

Western Europe 0 −1 1 30 High(0.03) (−0.08) (0.07) (2.69)

Australia and New Zealand 10 12 10 55 High (0.98) (1.14) (0.99) (4.45)

Japan 1 −2 −7 35 Medium(0.06) (−0.19) (−0.71) (3.02)

China 8 13 50 174 High(0.81) (1.26) (4.15) (10.62)

Taiwan (China) 9 11 14 96 High(0.86) (1.05) (1.36) (6.97)

Other newly industrialized 10 −1 55 88 Mediumcountries in Asia (0.93) (−0.10) (4.47) (6.53)

Indonesia 16 17 123 25 Low (1.50) (1.59) (8.36) (2.27)

Vietnam 15 32 36 111 Medium(1.40) (2.79) (3.10) (7.78)

Other Southeast Asia 18 22 134 60 Low (1.70) (2.04) (8.87) (4.83)

India 18 23 78 88 Medium(1.67) (2.10) (5.92) (6.54)

Other South Asia 25 30 80 72 Medium(2.22) (2.69) (6.06) (5.55)

Brazil 14 19 72 31 Medium(1.31) (1.77) (5.60) (2.75)

Other Latin America 18 6 90 54 Low(1.68) (0.57) (6.65) (4.42)

Turkey 16 19 107 55 Low(1.47) (1.75) (7.55) (4.46)

Other Middle East and North Africa 24 37 67 28 Low (2.16) (3.23) (5.24) (2.50)

Economies in transition −1 6 9 33 High (−0.11) (0.56) (0.90) (2.88)

South African Customs Union 15 31 47 34 Low (1.39) (2.76) (3.92) (2.94)

Other Sub-Saharan Africa 30 40 54 38 Medium(2.65) (3.42) (4.42) (3.26)

Rest of world 18 23 35 68 Low(1.63) (2.10) (3.05) (5.32)

aLow = average annual growth rates of 0.1 percent, medium = 1.0 percent, high = between 2 percentand 4 percent.Note: Numbers in parentheses are annual growth rates.Source: World Bank; GTAP data.

proportionate to the fall in tariffs on automobilesin each simulation.

The same macroeconomic closure is used for allexperiments—full employment,19 perfect mobilityof skilled and unskilled workers between nonagri-cultural sectors, and perfect mobility of unskilledworkers within agriculture. Based on the workingassumption that trade balances are not changedsignificantly by WTO accession, trade balances asshares of GDP were fixed for China and Taiwan(China). While trade balances can be expected tovary during the transition, particularly if there aresubstantial changes in foreign investment levels,foreign investment levels are not determinedwithin the model.20

Taxes lost because of trade liberalization areassumed to be replaced by a uniform, nondistor-tionary consumption tax affecting both private andgovernment final consumption of all goods andservices. This hypothetical tax is included to ensurethat any adverse impacts of trade reform on gov-ernment revenues, and hence on the ability toprovide income transfers or public services, areallowed for in the analysis of impacts of the reformon households. Because the GTAP version 5 data-base appears to represent the VAT on domestic pro-duction as an output tax, the model generates taxlosses from the contraction of industries such astobacco and alcohol. These inward-oriented indus-tries have higher VAT rates than export-orientedsectors such as apparel because exports are exemptfrom the VAT. When the export-oriented sectorsexpand, the net impact of WTO accession is a sharpcontraction in tax revenues. In reality, such a con-traction will not be observed because VATs of thesame magnitude are levied on imports. To offsetthis impact, particularly in the poverty analysis, theconsumption tax had to be adjusted downward.This was done by first computing the consumptiontax that compensates for the loss in output taxes(this tax as a share of the total replacement tax isequal to the share of the output tax loss in the totaltax losses) and then adjusting the consumption taxrate to eliminate the component due to the changein output taxes.

To reflect the long-run change in the stance oftrade policy—phased in over many years—involvedin WTO accession, most of the analysis uses astandard long-run specification, with capital andlabor freely mobile between industrial sectors and

within agriculture. However, barriers to labormobility between rural and urban employment areincorporated.

Assessment of China’s Accession

This section assesses the impacts of China’s acces-sion to the WTO on China and its trading partners.

Impacts on China

The focus here is on the impacts of the trade policychanges remaining after 2001. (Detailed results forthe period before 2001 are presented in Appendixtable 13.A.6). One important development in theperiod after 2001 is the removal of the quotas onapparel and textiles imposed against China andother developing country exporters by majorindustrial country importers. These quotas arescheduled for abolition in January 2005 for all WTOmembers. Abolition gives a significant boost to thetextile and apparel sectors in China, which had beenone of the countries most tightly restricted bythe quotas.21 Output in these sectors rises substan-tially (table 13.5). That in turn stimulates the pro-duction of plant-based fibers (mainly cotton),which increases by 16 percent.

Output and employment in the other agriculturalsectors, with the exception of livestock and meat, areexpected to fall as unskilled agricultural labor movesinto the textile and apparel sectors and unskillednonfarm real wages rise (table 13.6). Oilseeds andsugar contract more than other agricultural sectorsas a result of falling protection. Tariffs on oilseeds fallfrom 20 percent to 3 percent, while tariffs on sugarfall from 40 percent to 20 percent. Protection inother agricultural sectors is assumed to remainalmost unchanged. The automobile and electronicssectors also expand slightly, creating employmentopportunities, particularly for skilled labor.22 Resultssuggest that approximately 6 million farm workersin China will leave their farm jobs as a result of WTOaccession reform after 2001 in pursuit of employ-ment in the non-agricultural sectors.23

For most merchandise goods real wholesale pricesfall as a result of trade liberalization after accession.Retail prices reflect a uniform consumption taxincrease of about 1.9 percent levied to compensatefor the loss of tariff revenue.24 For some products,such as beverages and tobacco, automobiles, and



TABLE 13.5 Changes in China’s Key Economic Indicators after 2001 as a Result of WTO Accession (percent unless otherwise indicated)

Trade Balance(Millions of Wholesale Consumer

Product Output Employment Exports Imports US$) Prices Prices

Agriculture

Rice −2.1 −2.3 6.1 −7.1 64 −0.9 0.9

Wheat −2.0 −2.3 18.9 −10.1 174 −1.7 0.4

Feedgrains −2.3 −2.6 −77.8 −2.4 −596 −1.9 1.9

Vegetables and fruits −3.4 −3.7 14.6 −6.3 214 −1.9 −0.1

Oilseeds −7.9 −8.4 29.8 20.9 −789 −2.8 −4.7

Sugar −6.5 −7.4 13.9 24.1 −73 −1.9 −3.1

Plant-based fibers 15.8 16.4 −51.8 7.7 −189 0.1 3.1

Livestock and meat 1.3 1.1 15.5 −8.9 837 −1.6 0.2

Dairy −2.0 −2.4 13.5 23.8 −143 −1.5 0.2

Other food −5.9 −6.4 11.4 62.6 −3,460 −1.7 −1.8

Beverages and tobacco −33.0 −33.1 9.7 112.4 −14,222 −1.8 −6.9

Manufacturing

Extractive industries −1.0 −1.3 7.5 −4.4 2,088 −0.7 1.2

Textiles 15.6 15.5 32.7 38.5 −10,366 −1.7 −3.2

Apparel 57.3 56.1 105.8 30.9 49,690 −0.5 −1.9

Light manufacturing 3.7 3.7 5.9 6.8 1,786 −0.9 0.0

Petrochemicals −2.3 −2.3 3.1 11.8 −8,810 −0.7 0.8

Metals −2.1 −2.1 3.7 6.8 −1,893 −0.4 1.3

Automobiles 1.4 −2.2 27.7 24.0 516 −3.9 −4.2

Electronics 0.6 0.4 6.7 6.8 453 −1.3 −1.7

Other manufactures −2.1 −2.2 4.1 18.9 −11,291 −0.5 0.8

Services

Trade and transport 0.0 0.0 0.8 −0.4 493 −0.2 1.6

Construction 0.9 0.9 2.7 17.5 −436 −0.2 1.7

Communications −0.5 −0.5 −0.5 10.9 −56 0.1 1.9

Commercial services −2.0 −2.0 −0.4 35.4 −1,749 0.2 1.9

Other services −1.7 −1.8 1.4 33.6 −1,525 −0.1 1.6

Total 1.0 0.0a 16.8 17.3 717 −0.7 −0.2

aReflects the fixed labor supply assumption.

Source: Authors’ simulations with modified GTAP model; see details in text.

sugar, the fall in real retail prices reflects a larger thanproportionate drop in protection.

Increased demand for nonagricultural labormeans higher real nonfarm wages and higherreturns to nonagricultural labor relative to agricul-tural labor. Removal of protection on some agricul-tural sectors additionally lowers the attractivenessof farming and implies falling returns to farm laborand land. Real farm wages fall 0.7 percent and thereal rental price of land falls 5.5 percent. Thedecline in farm incomes and the rise in the realretail price of many nonfarm products mean that

some farmers may be hurt by WTO accession. Non-farm wages rise 1.2 percent and skilled labor wagesrise 0.8 percent, implying that workers in urbancenters—and farmers who are able to engage innonfarm employment—are more likely to be betteroff as a result of WTO accession.25

Accession will make China a much bigger playerin world markets through three channels—therapid growth and structural change of its economy;the liberalization undertaken in preparation forWTO accession; and the liberalization under-taken after accession in 2001. The liberalization

undertaken after 2001 contributes to an increase inChina’s share in world exports from 4.4 percent to7.8 percent upon completion of accession. Simi-larly, China’s share in world import markets risesfrom 5.8 percent in 2001 to 6.4 percent in 2007.With the removal of textile and apparel quotas,apparel exports lead export expansion with anincrease in export volume of about 106 percent,followed by textiles and automobiles. The dramaticfall in protection of beverages and tobacco resultsin imports more than doubling, followed byincreases in imports of food products, textiles, agri-cultural products, automobile parts, and commer-cial services.

China’s total welfare gain from WTO accession isestimated at $40.6 billion (in 1997 dollars), or2.2 percent of per capita real income (table 13.7).26

Most of the gain ($31 billion) was realized follow-ing the massive liberalization between 1995 and2001 and the ongoing restructuring of the automo-bile industry. The remaining reforms will lead to anadditional welfare gain of $9.6 billion. The largestpart of this gain in welfare will come from fur-ther merchandise trade liberalization ($4.7 billion,nearly half the $9.6 billion), followed by $2.4 billion(25 percent) from the removal of quotas on textilesand apparel and $1.2 billion (12 percent) from serv-ices liberalization. Continuing automobile sectorrestructuring will generate $1.1 billion (11 per-cent), while the removal of agricultural export sub-sidies will provide only $275 million (3 percent) inadditional benefits.

Impacts on China’s Trading Partners

Among China’s trading partners the largestabsolute gains accrue to North America andWestern Europe with close to half of the gains com-ing from elimination of the quotas they impose onChina’s exports of textiles and clothing—and thuselimination of their efficiency losses and rent trans-fers to China. North America, Western Europe, andJapan also gain from China’s cuts in protection,which increase China’s efficiency as an exportsupplier and its demand for their exports.

The welfare gain by Taiwan (China) from its andChina’s accession to the WTO is estimated at $3 bil-lion per year—the second largest gain relative to thesize of the economy after China’s (see table 13.7).About half of the gain ($1.6 billion) was realizedas a result of the liberalization in China and inTaiwan (China) during 1997–2001. Remainingreforms will lead to an estimated real income gainof $1.4 billion a year after 2001. Other newly indus-trialized economies also benefit from China’s acces-sion. Most of these benefits are associated withtrade liberalization and removal of quotas on tex-tile and apparel, which translate into gains fromterms of trade improvements after 2001.

The world as a whole and key developing eco-nomies that trade directly with China benefit fromChina’s accession, but developing economies inSoutheast Asia, South Asia, and Latin America that compete with China in third markets maylose from the removal of textile and apparel quotas


TABLE 13.6 Change in Real Factor Prices in China as a Result of Accession, 2001–07 (percent)

Accession with Accession with Accession with Labor Accession Labor Market Increase in Skill Market Reform and

Item Alone Reform only Level only Increase in Skill Level

Farm unskilled wages −0.7 16.8 1.6 19.4Nonfarm unskilled wages 1.2 −3.8 2.7 −2.5Skilled labor wages 0.8 −1.7 −6.3 −8.7Rental price of land −5.5 −9.7 −6.4 −10.5Rental price of capital 1.3 −1.4 0.9 −1.8Price of capital goods −0.9 −3.6 −1.1 −3.9Migration from rural to 6 28 10 32.0

urban jobs (millions)National welfare 10.0 11.0 10.0 11.0

(billions of 1997 US$)



TABLE 13.7 Welfare Change and Sources of Welfare Change as a Result of China’sWTO Accession (millions of 1997 US$)

Total, Export1995– Tariff Quota Subsidy Liberalization Auto Sector Impact

Country or Group 2007 Cuts Reductions Reductions of Services Restructuring 2001–07

North America 6,072 3,207 2,713 24 172 −44 5,259(0.0)*

Western Europe 18,189 9,724 8,285 −51 338 −107 14,200(0.2)

Australia/New Zealand 136 175 −47 2 18 −12 152(0.0)

Japan 5,694 5,522 291 −22 5 −102 2,553(0.1)

Chinaa 40,552 29,452 2,389 275 1,160 7,276 9,563(2.2)

Taiwan (China) 2,985 2,300 338 −4 265 85 1,376(0.6)

Other Newly Industri- 6,831 6,539 −82 −185 49 511 1,456alized Countries (0.7)

Indonesia −408 −167 −216 −10 1 −16 −310(−0.2)

Vietnam −453 −63 −395 0 6 0 −405(−1.4)

Other Southeast Asia −585 −109 −464 −46 16 18 −268(−0.1)

India −3,357 −1,087 −2,338 −5 −23 96 −2,999(−0.4)

Other South Asia −1,622 −176 −1,427 −7 1 −12 −1,619(−0.8)

Brazil −76 −76 3 4 5 −12 359(−0.0)

Other Latin America −32 59 −171 20 32 29 −36(−0.0)

Turkey −338 −50 −295 −2 7 2 −327(−0.1)

Other Middle East 368 675 −467 −13 57 116 −365and North Africa (0.0)

Economies in 19 318 −321 4 15 3 −185Transition (0.0)

South African 78 89 −18 0 5 2 13Customs Union (0.0)

Other Sub-Saharan −45 71 −159 4 15 24 −78Africa (−0.0)

Rest of World 155 330 −210 −15 27 23 −78(0.0)

World 74,166 56,733 7,409 −27 2,171 7,880 28,261

aImpacts exclude output tax losses because of a compensating value-added tax levied uniformly on both imported anddomestic goods.

Numbers in parentheses are percentage changes in per capita utility. The impact for 1995–2001 is the differencebetween the impact for 1995–2007 and the impact for 2001–07.


after 2001. The losses will be largest for Vietnam—an economy that is following in China’s foot-steps and has a similar pattern of comparativeadvantage in labor-intensive products. The wel-fare loss for Vietnam is estimated as a 1.4 per-cent drop in per capita income (see table 13.7).The loss to India is estimated to be considerablysmaller as a share of per capita income, at 0.4 per-cent, while the percentage losses to other countriesare very small.

Complementary Policy Reforms

While the overall effects of WTO accession onChina’s economy are generally positive, there aresome concerns that declines in real returns tofarm labor may exacerbate poverty in rural areas.Approaches that deal directly with these prob-lems are more likely to succeed than approachesthat attempt to water down China’s trade policyreforms. Two policy tools that lend themselves toanalysis within the model framework used here arerelaxation of the barriers to labor migration fromrural to urban areas and skills upgrading for work-ers in rural areas.

Impact of Reducing the Policy Barriers to Labor Mobility

Abolishing policy barriers to labor mobility fromrural to urban areas—such as residence permits,differences in social insurance, and the inability tosell agricultural land—in conjunction with acces-sion leads to a nearly 17 percent increase in realreturns to rural workers (see table 13.6).27 Thiscontrasts sharply with the 0.7 percent reduction inreal farm wages for accession without labor marketreform. Rents to farmland would decline, withhigher farm wages leaving a smaller residual returnto farmland. Real urban unskilled wages woulddecline by an estimated 3.8 percent. Clearly, therewould be scope for partial reform of these arrange-ments that could leave both farm and nonfarmunskilled workers better off than in the absence oflabor market reform.

These results suggest that this reform wouldhave significant impacts on the number of peopleleaving their farm jobs for jobs in the nonfarm sec-tors and on the industry composition of China’seconomy. Some 28 million people would leave their

farm jobs if the government removed the policybarriers to labor movement from rural to urbanareas28—several times the estimated 6 million peo-ple who would move as a result of WTO accessionreforms alone between 2001 and 2007. The impacton the composition of Chinese industrial outputwould also be substantial (table 13.8). This wouldallow not only apparel production to expand morebut also metals, automobiles, electronics, machin-ery, other manufactures, and construction, all at theexpense of reductions in some agricultural sectors.

Impact of an Increase in Skill Levels

One of the central problems facing most rural work-ers is low levels of education. One way to get a senseof the likely impacts of improving access to educa-tion is to consider the impact of resultant increasesin the skill levels of rural workers on the perform-ance of the Chinese economy. This experiment looksonly at the impact of improvements in education onthe skills of rural workers. It ignores any potentialbenefits to rural households from improvementsin access to education for their children—such asreductions in school fees—and any changes in thegovernment budget associated with increases ingovernment spending on education.29

An increase in the provision of education thatwould boost the annual growth rate for skilledlabor from 4.15 percent to 5 percent and wouldlead to a decline in the annual growth rates forunskilled labor from 1.26 percent to 1.1 percentwas considered. This was found to have importantimpacts on the structure of the Chinese economy.An increase in skilled labor leads to a strongerexpansion, or a smaller contraction, in the manu-facturing sectors that are skilled labor–intensivethan does accession with labor market reform butno change in education spending (table 13.8,columns 2 and 4). Metals, automobiles, electronics,and other manufactures all expand.

While output in some sectors expands, the realwages of skilled workers fall as the supply of skilledworkers increases (see table 13.6) and world pricesof the outputs they produce decline. This contrastswith the case of accession alone, which results in anincrease in the real wages of skilled workers. How-ever, the real wages of generally much poorerunskilled workers rise with increased education,with the wages of unskilled nonfarm workers rising



more than those of unskilled farm workers (seetable 13.6). Of course, those who are able to trans-fer from agricultural to nonagricultural employ-ment as a result of increased educational opportu-nities are likely to be substantially better off.

Overall, it is clear that increasing access to train-ing and skills could substantially offset the adverseimpacts on rural labor of the trade reforms associ-ated with accession. Finally, increased educationboosts the need for migration as demand forunskilled workers increases in large urban areas.An estimated 10 million farm workers are expectedto exchange farm jobs for nonfarm ones (seetable 13.6). The impact on consumer prices issmall—with falling prices for farm products andrising prices for manufactured commodities.

Impact of Labor Market Reform and an Increase in Skill Levels

The combination of removing labor market barri-ers and increasing education spending creates themost favorable scenario for unskilled farm labor,leading to the largest increase in real farm wages(19.4 percent; see table 13.6). Farm output con-tracts more than in the case of labor marketreform alone, while skilled labor–intensive indus-tries such as metals, automobiles, electronics, othermanufactures, and services expand more than inthe case of labor market reform alone or increasededucation spending alone (see table 13.8). Underthis scenario an estimated 32 million farm workerswould leave their farm jobs for jobs in urban areas(table 13.6).

These results suggest that to generate pro-poorgrowth over the next decade, the governmentshould consider both removing policy barriers tolabor movement and changing the composition ofspending to favor education. Not only would thesepolicies facilitate the transformation of China’seconomy toward services and high-tech manufac-turing sectors, but they also have the potential tomore than offset any negative impacts of accessionon rural wages and incomes.

Conclusion

The analysis suggests that the reforming economiesand their close trading partners will be the biggestbeneficiaries of accession to the WTO. China is

undertaking the greatest reform and will gain themost. The North American and Western Europeaneconomies that abolish their export quotas on tex-tiles and clothing and increase their direct tradewith China will gain the most in absolute terms.Taiwan (China) will benefit substantially, both as aconsequence of its own liberalization and throughstrengthened trade links with China. Japan willgain substantially because of increased exportopportunities in China and China’s increasedcompetitiveness as a supplier. Other industrializ-ing and industrialized economies that are China’slargest trading partners will also be substantialgainers.

China’s WTO accession will have a noticeableimpact on the level and pattern of global trade.With accession, China is becoming a much biggerplayer in world markets. Apparel exports will leadChina’s export expansion, followed by textiles andautomobiles. In addition to being an importantsource of traded goods, China will become animportant destination for other economies’ prod-ucts. Imports of beverages and tobacco will morethan double, followed by imports of food prod-ucts, textiles, agricultural products, automobileparts, and commercial services. The expansion oftextiles, light manufactures, petrochemicals, andequipment exports from Taiwan (China) will bedriven almost entirely by demand for these prod-ucts in China.

Accession will have important distributionalconsequences for China. The wages of skilled work-ers and unskilled nonfarm workers will rise in realterms and relative to the wages of farm workers. Anestimated 6 million people will leave their farm jobsin pursuit of employment in industry and services.Real farm wages and land rental rates will decline.The decline in farm incomes and the rise in the realretail prices of many nonfarm products suggest thatsome farmers may be hurt by WTO accession after2001—an issue explored by Chen and Ravallion inchapter 15.

To help offset these adverse impacts on farmers,the Chinese government might make changes in itslabor market policies. Abolition of the hukou sys-tem and reform of the labor market more generallywould raise farm wages and allow 28 million peopleto migrate to nonfarm jobs in search of a better life.It would lead to an even bigger expansion of thelabor-intensive manufacturing sector. An increase

in education spending, to improve skills, wouldhave a positive impact on the structure of theChinese economy. The real wages of skilled work-ers would fall, while the real wages of unskilledworkers would rise with increased educationspending. Thus, on the income side, improved skillswould induce pro-poor growth and decreasepoverty and inequality.

A number of caveats are important. The gains toChina and Taiwan (China) are probably understatedbecause tariff aggregation in the GTAP model hides

much of the variation in tariffs and the welfare gainsfrom reducing this variation within the productaggregates used in the analyses (Bach and Martin2001; Martin, van der Mensbrugghe, and Manole2003). When Bach, Martin, and Stevens (1996)adjusted for this in a partial equilibrium context,gains to China almost doubled. The analysis hereassumes flexible wages and full employment. How-ever, trade liberalization and foreign competitionmay worsen unemployment and put downwardpressure on the wages of unskilled workers in the


TABLE 13.8 Change in Output and Employment in China as a Result of WTOAccession and Other Reforms (percent change over the period2001–07)

Output Employment

With labor With LaborMarket Without With With Market

Without With Labor With Reform and Labor Labor Increase Reform andHukou Market Increase in Increase in Market Market in skill Increase in

Product Removal Reform Skill Level Skill Level Reform Reform Level Skill Level

Agriculture

Rice −2.1 −4.3 −2.4 −4.6 −2.3 −7.4 −3.1 −8.2

Wheat −2.0 −11.5 −3.3 −12.9 −2.3 −13.3 −3.9 −14.9

Feedgrains −2.3 −7.8 −3.1 −8.6 −2.6 −9.7 −3.7 −10.6

Vegetables and fruits −3.4 −7.1 −3.9 −7.7 −3.7 −8.9 −4.6 −9.7

Oilseeds −7.9 −18.4 −9.4 −19.8 −8.4 −20.4 −10.2 −22.0

Sugar −6.5 −17.1 −8.0 −18.4 −7.4 −22.4 −9.6 −24.2

Plant-based fibers 15.8 12.8 15.1 12.1 16.4 11.6 15.5 10.6

Livestock and meat 1.3 −3.3 0.6 −4.0 1.1 −7.0 −0.3 −8.2

Dairy −2.0 −9.4 −3.1 −10.5 −2.4 −14.4 −4.3 −16.0

Other food −5.9 −13.4 −7.0 −14.5 −6.4 −13.2 −8.9 −15.5

Beverages and tobacco −33.0 −38.7 −33.7 −39.5 −33.1 −37.6 −35.0 −39.5

Maufacturing

Extractive industries −1.0 0.1 −1.2 −0.1 −1.3 0.2 −1.7 −0.2

Textiles 15.6 14.7 15.3 14.3 15.5 16.8 12.7 14.0

Apparel 57.3 61.4 56.7 60.7 56.1 62.6 52.7 59.1

Light manufacturing 3.7 −6.8 2.1 −8.5 3.7 −5.4 0.1 −8.9

Petrochemicals −2.3 −1.3 −2.3 −1.2 −2.3 0.7 −4.4 −1.4

Metals −2.1 0.8 −1.8 1.2 −2.1 2.4 −3.9 0.7

Automobiles 1.4 4.1 1.8 4.4 −2.2 2.3 −4.0 0.5

Electronics 0.6 4.5 1.1 5.1 0.4 6.3 −1.3 4.6

Other manufactures −2.1 0.3 −1.9 0.6 −2.2 2.2 −4.0 0.3

Services

Trade and transport 0.0 0.8 0.1 1.0 0.0 3.4 −3.1 0.4

Construction 0.9 2.0 0.9 1.9 0.9 3.4 −1.4 1.0

Communications −0.5 0.6 −0.3 0.9 −0.5 3.4 −3.0 0.8

Commercial services −2.0 −1.4 −1.8 −1.2 −2.0 1.0 −4.7 −1.8

Other services −1.7 −0.5 −0.9 0.3 −1.8 1.5 −6.2 −2.9

Source: Authors’ simulations with modified GTAP model.

short run. Furthermore, while the analysis hereimproves on Ianchovichina and Martin (2001),with better treatment of the extent of liberalizationin agriculture and services and the changes in theautomobile sector, there are still areas that havebeen ignored. One is nontariff barriers in themanufacturing sectors other than the quotas onapparel and textiles. Another is the impact of acces-sion on foreign direct investment (FDI) and thehard-to-measure efficiency gains in services that areassociated with this increased investment.

FDI has contributed significantly to China’s eco-nomic growth and will play an important role asChina continues to reform its economy. WTOaccession is likely to increase FDI in China, as trade

liberalization improves returns to investment andthe liberalization of rules on investment easesfinancial flows into previously restricted sectorssuch as services and automobile production. Thesubstantial productivity gaps between local andforeign firms imply that new FDI will raise produc-tivity.30 Walmsley, Hertel, and Ianchovichina(2004) take into account both the impact of FDIand increased productivity growth in services.They find that the impacts of accession are farlarger than those predicted by earlier studies,including this one, which ignore the potentialproductivity gains in the services sector and areabstracted from capital accumulation and foreigninvestment.



TABLE 13.A.1 Tariff Lines Subject to Import NTBs, China, 2001

Licenses & Licensing State Designated Any Total Quotas Tendering only Trading Trading Unrestricted NTB Tarifflines

Paddy rice 0 0 3 3 0 0 3 3

Wheat 0 0 3 3 0 0 3 3

Cereal grains nec 0 0 1 2 0 9 2 11

Vegetables, fruit, nuts 0 0 0 0 0 109 0 109

Oil seeds 0 0 0 0 0 25 0 25

Sugar cane, sugar beet 0 0 0 0 0 1 0 1

Plant-based fibers 1 0 0 1 0 6 1 7

Crops nec 5 0 0 5 0 96 5 101

Cattle, sheep, goat, horses 0 0 0 0 0 6 0 6

Animal products, nec 0 0 0 0 0 62 0 62

Wool, silk-worm cocoons 3 0 0 1 2 9 3 12

Forestry 0 0 0 0 12 23 12 35

Fishing 0 0 0 0 0 57 0 57

Coal 0 0 0 0 0 6 0 6

Oil 0 0 2 1 0 1 3 4

Gas 1 0 0 1 0 1 1 2

Minerals nec 0 0 0 0 0 106 0 106

Meat: cattle, sheep, goats 0 0 0 0 0 26 0 26

Meat products nec 0 0 0 0 0 47 0 47

Vegetable oils and fats 0 0 12 7 0 32 12 44

Dairy products 0 0 0 0 0 24 0 24

Processed rice 0 0 2 2 0 0 2 2

Sugar 7 0 0 9 0 3 9 12

Food products nec 0 0 0 8 0 311 8 319

Beverages and tobacco 1 0 9 5 1 17 15 32

Textiles 39 0 0 3 25 711 46 757

Wearing apparel 0 0 0 0 0 289 0 289

Leather products 11 16 0 0 0 73 27 100

Wood products 0 0 0 0 18 106 18 124

Paper products, publishing 0 0 0 3 0 160 3 163

Petroleum, coal products 8 0 0 7 0 25 8 33

Chemical, rubber, plastic 35 0 15 21 5 1248 51 1299

Mineral products nec 0 0 0 0 0 198 0 198

Ferrous metals 0 0 0 0 181 49 181 230

Metals nec 0 0 0 0 0 190 0 190

Metal products 0 0 0 0 1 264 1 265

Motor vehicles and parts 64 0 0 0 0 93 64 157

Transport equipment nec 7 10 0 0 0 72 17 89

Electronic equipment 36 17 0 0 0 205 53 258

Machinery and equipment 39 77 0 2 0 1199 116 1315

Manufactures nec 0 0 0 0 0 219 0 219

Electricity 0 0 0 0 0 1 0 1

Gas manuf, distribution 0 0 0 0 0 1 0 1

Total 257 120 47 84 245 6080 664 6744

Source: WTO 2001. Commodity definitions are for GTAP-5, see www.gtap.org for concordances.


TABLE 13.A.2 The Import Coverage of Nontariff Barriers in China, 2001

Licenses & Licensing State Designated Any Unrestricted Quotas Tendering only Trading Trading NTB Trading Total

% % % % % % % %

Paddy rice 100 0 100 100 0 100 0 100

Wheat 100 0 100 100 0 100 0 100

Cereal grains nec 0 0 0 0 0 0 100 100

Vegetables, fruit, nuts 0 0 0 0 0 0 100 100

Oil seeds 0 0 0 0 0 0 100 100

Sugar cane, sugar beet 0 0 0 0 0 0 100 100

Plant-based fibers 93 0 0 93 0 93 7 100

Crops nec 48 0 0 48 0 48 52 100

Cattle, sheep and 0 0 0 0 0 0 100 100goats, horses

Animal products, nec 0 0 0 0 0 0 100 100

Wool, silk-worm cocoons 0 0 0 0 95 95 5 100

Forestry 0 0 0 0 94 94 6 100

Fishing 0 0 0 0 0 0 100 100

Coal 0 0 0 0 0 0 100 100

Oil 100 0 0 100 0 100 0 100

Gas 0 0 0 0 0 0 100 100

Minerals nec 0 0 0 0 0 0 100 100

Meat: cattle, sheep, 0 0 0 0 0 0 100 100goats, horse

Meat products nec 0 0 0 0 0 0 100 100

Vegetable oils and fats 59 0 60 59 0 60 40 100

Dairy Products 0 0 0 0 0 0 100 100

Processed rice 100 0 100 100 0 100 0 100

Sugar 85 0 0 85 0 85 15 100

Food products nec 1 0 0 1 0 1 99 100

Beverages and tobacco 20 0 16 20 0 36 64 100products

Textiles 9 0 0 0 8 14 86 100

Wearing apparel 0 0 0 0 0 0 100 100

Leather products 0 0 0 0 0 0 100 100

Wood products 0 0 0 0 55 55 45 100

Paper products, publishing 0 0 0 0 0 0 100 100

Petroleum, coal products 58 0 0 58 0 58 42 100

Chemical, rubber, 5 0 1 5 2 7 93 100plastic prods

Mineral products nec 0 0 0 0 0 0 100 100

Ferrous metals 0 0 0 0 85 85 16 100

Metals nec 0 0 0 0 0 0 100 100

Metal products 0 0 0 0 1 1 99 100

Motor vehicles and parts 32 0 0 0 0 32 68 100

Transport equipment nec 1 3 0 0 0 4 96 100

Electronic equipment 9 5 0 0 0 14 86 100

Machinery and 1 8 0 0 0 10 90 100equipment nec

Manufactures nec 0 0 0 0 0 0 100 100

Total Import Coverage 12.8 2.7 0.5 9.5 6.2 21.6 78.4 100

Note: Based on WTO 2001 and import data from China Customs for 2000.


TABLE 13.A.3 Commodity Import Shares by NTB Measure

Licenses Licensing State Designated Any Unrestricted& Quotas Tendering Only Trading Trading NTB Trading

% % % % % % % Total

Paddy rice 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Wheat 0.0 0.0 13.0 0.7 0.0 0.6 0.0 0.1

Cereal grains nec 0.0 0.0 0.0 0.0 0.0 0.0 0.2 0.2

Vegetables, fruit, nuts 0.0 0.0 0.0 0.0 0.0 0.0 0.2 0.2

Oil seeds 0.0 0.0 0.0 0.0 0.0 0.0 1.7 1.5

Sugar cane, sugar beet 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Plant-based fibers 0.5 0.0 0.0 0.3 0.0 0.0 0.0 0.0

Crops nec 1.1 0.0 0.0 0.8 0.0 0.0 0.1 0.1

Cattle, sheep and goats 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Animal products, nec 0.0 0.0 0.0 0.0 0.0 0.0 0.5 0.4

Wool, silk 4.3 0.0 0.0 0.0 4.7 0.7 0.0 0.1

Forestry 0.0 0.0 0.0 0.0 11.7 2.4 0.1 0.3

Fishing 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1

Coal 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Oil 0.0 0.0 0.0 69.4 0.0 48.9 0.0 5.6

Gas 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Minerals nec 0.0 0.0 0.0 0.0 0.0 0.0 2.5 2.2

Meat: cattle, sheep, goat 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1

Meat products nec 0.0 0.0 0.0 0.0 0.0 0.0 0.4 0.3

Vegetable oils and fats 0.0 0.0 55.0 2.8 0.0 0.7 0.2 0.3

Diary products 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1

Processed rice 0.0 0.0 9.9 0.5 0.0 0.5 0.0 0.1

Sugar 0.8 0.0 0.0 0.5 0.0 0.4 0.0 0.1

Food products nec 0.0 0.0 0.0 0.1 0.0 0.0 1.4 1.2

Beverages and tobacco 0.0 0.0 2.9 0.2 0.0 0.2 0.1 0.1

Textiles 16.2 0.0 0.0 1.3 8.2 0.7 7.1 6.3

Wearing apparel 0.0 0.0 0.0 0.0 0.0 0.0 0.6 0.5

Leather products 0.0 0.0 0.0 0.0 0.0 0.0 1.6 1.4

Wood products 0.0 0.0 0.0 0.0 8.8 0.8 0.6 0.6

Paper products 0.0 0.0 0.0 0.1 0.0 0.0 4.0 3.5

Petroleum, coal prods 20.8 0.0 0.0 14.8 0.0 3.4 1.3 1.6

Chemical, rubber, plastic 17.0 0.0 19.1 8.1 4.2 0.5 17.9 15.9

Mineral products nec 0.0 0.0 0.0 0.0 0.0 0.0 1.1 1.0

Ferrous metals 0.0 0.0 0.0 0.0 62.1 29.9 0.9 4.3

Metals nec 0.0 0.0 0.0 0.0 0.0 0.0 5.2 4.6

Metal products 0.0 0.0 0.0 0.0 0.3 0.0 2.0 1.7

Motor vehicles and parts 8.7 0.0 0.0 0.0 0.0 2.4 1.6 1.7

Transport equipment nec 0.2 1.6 0.0 0.0 0.0 0.1 1.8 1.6

Electronic equipment 26.8 41.1 0.0 0.0 0.0 6.0 23.8 21.7

Machinery and equip 3.8 57.3 0.0 0.3 0.0 1.6 22.2 19.8

Manufactures nec 0.0 0.0 0.0 0.0 0.0 0.0 0.7 0.6

Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0


TABLE 13.A.4 Elasticity of Substitution and Changes in Real Consumer Prices dueto China’s WTO Accession

Elasticity of Substitution

between Domestic Products Short Run Closure Long Run Closure

and Imports* 1995–2007 2001–2007 1995–2007 2001–2007

Rice 4.4 2.2 0.7 2.5 1.0Wheat 4.4 −0.8 0.7 0.4 0.4Feed grains 4.4 13 2.1 13.5 1.9Vegetables and fruits 4.4 0.9 −0.6 1.4 −0.1Oilseeds 4.4 −6.7 −5.9 −4.8 −4.6Sugar 4.4 −2.1 −3.5 −1.8 −3.1Plant based fibers 4.4 2.1 4.1 3.1 3.1Livestock & meat 5.0 3.8 0.7 2.4 0.2Dairy 4.4 2 −0.5 2 0.2Other food 4.4 0.3 −2.7 0.4 −1.8Beverages & tobacco 6.2 −14.3 −7.7 −10.8 −6.9Extractive industries 5.6 2.5 1.7 2.3 1.2Textiles 4.4 −10.3 −1.5 −10.5 −3.1Apparel 8.8 −6.7 0.8 −8.6 −1.9Light manufacturing 8.8 −2 0.5 −2.6 0.0Petrochemical industry 4.1 0.7 0.8 0.9 0.9Metals 5.6 1.2 1.3 1.3 1.3Automobiles 10.4 −23.6 −4.0 −22.2 −4.2Electronics 5.6 −5.3 −1.4 −5.8 −1.7Other manufactures 5.5 0.5 0.8 0.4 0.9Trade and transport 3.8 3 1.7 2.6 1.7Construction 3.8 2.8 1.7 2.5 1.7Communication 3.8 3.6 1.7 2.8 1.9Commercial services 3.8 2.7 0.9 3.2 1.9Other services 4.0 2.4 1.3 2.5 1.6


TABLE 13.A.5 Changes in China’s Key Economic Indicators in the Baseline, without WTO Accession, 1995–2007

Output Employment Exports Imports(Percent) (Percent) (Percent) (Percent)

Rice 63.8 −11.5 134.7 −8.8Wheat 81.4 6.4 −15.2 126.3Feed grains 109.5 23.8 −0.6 95.9Vegetables and fruits 98.2 16.8 −10.8 122.1Oilseeds 100.9 18.4 −36 151.7Sugar 112.5 14.5 109.4 88.7Plant based fibers 137.2 41.1 −8.5 146.1Livestock & meat 121.9 25.6 12.8 135.3Dairy 122.5 18.8 60.5 100.3Other food 110.8 −1.5 76.8 58.5Beverages & tobacco 114.6 −9 166.7 65.9Extractive industries 77.8 67.2 −88 554.2Textiles 142.2 −1.7 95 70.4Apparel 110.5 −1.7 100.8 47.2Light manufacturing 135.8 11 117.3 60.1Petrochemical industry 126.7 −5.6 80.2 101.5Metals 144.4 11.5 114.4 99.6Automobiles 157.5 2.2 419.7 105.3Electronics 186.5 13.1 168.1 112.1Other manufactures 161.1 6.4 195.1 51.3Trade and transport 129.4 −3.3 129.4 69.3Construction 113.3 26.5 77.4 85.6Communication 133.2 −12.4 452.9 11.9Commercial services 133.4 1.8 211.7 55.1Other services 113.6 21 101.1 75.1Total 120.2 N/A 131.7 97


TABLE 13.A.6 Changes in China’s Key Economic Indicators due to WTO Accessionfor the Period before 2001 (1995–2001)

Trade Wholesale Consumer Output Employment Exports Imports Balance Prices Prices

% % % % US$ m. % %

Rice −0.7 −0.7 9.5 −11.3 101 0.5 1.5Wheat −2.8 −2.9 14.9 39.1 −484 0.1 0.0Feed grains 2.3 2.4 2.5 −20.9 244 0.8 11.4Vegetables and fruits −0.6 −0.7 8.6 −26.5 486 0.5 1.5Oilseeds −1.8 −1.9 10.9 16.6 −549 0.5 −0.2Sugar −0.1 −0.2 9.1 3.3 8 0.6 1.3Plant based fibers −11.6 −12.2 29.6 −9.8 264 −1.2 0.0Livestock & meat 1.3 1.4 8.7 −27.3 1888 1.1 2.2Dairy 2.0 2.2 10.3 2.6 −10 0.8 1.8Other food 4.6 4.8 10.9 −26.3 3537 0.3 2.2Beverages & tobacco −23.8 −23.8 16.8 614.9 −10656 −0.5 −4.2Extractive industries −0.4 −0.5 16.6 1.9 −781 0.1 1.1Textiles −12.1 −12.2 16.7 35.5 −9556 −3.0 −7.6Apparel −3.4 −3.9 23.7 344.6 −344 −1.9 −6.8Light manufacturing 3.7 3.6 14.6 48.9 2008 −0.5 −2.6Petrochemical industry −0.7 −0.7 12.9 11.3 −4433 −0.6 0.0Metals −0.5 −0.5 20.9 23.5 −3051 −0.6 0.0Automobiles 2.8 −13.2 326.8 195.5 2935 −15.9 −18.8Electronics 3.8 3.6 18.4 16.2 2657 −1.8 −4.2Other manufactures 1.9 1.9 17.6 22.2 8511 −0.8 −0.5Trade and transport 1.3 1.3 10.1 −7.3 5284 −0.2 0.9Construction 2.2 2.2 14.8 −6.3 301 −0.2 0.8Communication 1.0 1.0 11.1 −7.4 229 −0.1 0.9Commercial services 0.8 0.8 10.5 −6.3 798 0.3 1.3Other services 0.3 0.4 12.7 −8.1 749 −0.1 0.9Total 2.0 N/A 20.3 22.0 138 −1.4 −0.8


TABLE 13.A.7 Sensitivity Analysis with Respect to the Elasticity of Transformation� (2.67) (percentage changes in China after 2001)

Wholesale Consumer Output Employment Exports Imports Prices Prices

(Percent) (Percent) (Percent) (Percent) (Percent) (Percent)

Rice −2.2 −2.5 4.2 −6.3 −0.8 1.1Wheat −2.4 −2.8 15.6 −9.2 −1.5 0.6Feed grains −2.5 −2.9 −78.2 −1.9 −1.7 1.8Vegetables and fruits −3.5 −3.9 12.3 −5.4 −1.7 0.1Oilseeds −8.3 −8.9 26.8 21.3 −2.6 −4.5Sugar −6.9 −8.1 11.8 24.4 −1.7 −3.0Plant based fibers 15.7 16.3 −52.3 8.0 0.1 3.1Livestock & meat 1.2 0.8 12.6 −7.6 −1.4 0.4Dairy −2.3 −2.9 11.5 24.4 −1.3 0.3Other food −6.2 −6.7 10.2 63.3 −1.7 −1.8Beverages & tobacco −33.2 −33.2 8.8 112.9 −1.9 −7.0Extractive industries −0.9 −1.2 7.6 −4.4 −0.8 1.1Textiles 15.6 15.6 32.6 38.6 −1.8 −3.2Apparel 57.4 56.4 106.1 30.7 −0.7 −2.0Light manufacturing 3.3 3.3 5.5 6.9 −0.9 0.0Petrochemical industry −2.3 −2.2 3.2 11.7 −0.8 0.7Metals −2.0 −1.9 4.0 6.8 −0.5 1.2Automobiles 1.5 −2.0 27.9 24.1 −4.0 −4.3Electronics 0.8 0.7 6.8 6.9 −1.4 −1.8Other manufactures −2.0 −2.0 4.3 18.8 −0.7 0.7Trade and transport 0.0 0.1 1.0 −0.4 −0.3 1.5Construction 0.9 1.0 3.0 17.4 −0.3 1.6Communication −0.5 −0.3 −0.3 10.9 0.0 1.8Commercial services −2.0 −1.9 −0.2 35.4 0.1 1.8Other services −1.6 −1.7 1.7 33.5 −0.2 1.5Total 1.0 0 17.5 17.2 −0.8 −0.1Nonfarm unskilled wages 1.08 Migration 7* Unskilled wage 0.7Farm unskilled wage −0.02 Land rent −5.7 Skilled wage 0.8Price of capital goods −0.95 Capital rent 1.2 Welfare 9728**

*1997 US millions.**Millions of workers leaving their farm jobs for nonfarm jobs.


Notes

1. Because China’s accession was a necessary condition forthat of Taiwan, China, and because of the strong trade linkagesbetween the two economies, the impact of Taiwan, China’saccession to the WTO is also considered.

2. Among the studies using a general equilibrium approachto quantify the impact of China’s WTO accession are Lejour(2000); Zhai and Li (2000); Li and others (2000); McKibbin andTang (2000); Ianchovichina and Martin (2001); Walmsleyand Hertel (2001); Deutsche Bank (2001); Wang (2002);Zhai and Wang (2002); Walmsley, Hertel, and Ianchovichina(2004); and Ianchovichina and Walmsley (2003).

3. Huang, Rozelle, and Min (2004) find that nominal rates ofprotection on important agricultural commodities (rice, vegeta-bles and fruits, livestock and meat) were negative in 2001 and arelikely to remain unchanged in the post-accession period. Conse-quently, the reduction in agricultural protection is likely to befar less than presented in earlier studies. Nonetheless, greaterscope for imports is likely for a range of agricultural products(wheat, oilseeds, sugar, and dairy products), which are protectedby tariffs that are scheduled to be reduced substantially and forproducts (cotton and feedgrains) for which export subsidies areruled out. These important findings were not incorporated inearlier studies.

4. The current treatment differs from that in Zhai and Wang(2002), who represent imperfect labor mobility in a single coun-try model with endogenous urban unemployment but do notdifferentiate between skilled and unskilled labor and employ alow level of elasticity of labor mobility between rural and urbanareas (0.25). The results here are similar to those in Zhai andWang (2002) in the case of WTO accession with high labormobility (elasticity of labor mobility is doubled) and fixedurban unemployment (flexible urban wage).

5. GTAP is a standard global applied general equilibriummodel with perfectly competitive markets and constant returnsto scale technology. The model represents consumer demandsthrough a constant difference of elasticities functional form andon the supply side emphasizes the role of intersectoral factormobility in the determination of sectoral output. Product differ-entiation between imported and domestic goods and amongimports by region of origin allows for two-way trade in eachproduct category, depending on the ease of substitution betweenproducts from different regions. Land, capital, skilled andunskilled labor, and, in some sectors, a natural resource factorare used in production and are fully employed.

6. Some sectors, particularly the service sectors, do not par-ticipate in export processing arrangements and so are notexempt from duties on intermediate inputs used in the produc-tion of exports.

7. In a deterministic world, a producer of exports will alwaystake advantage of duty exemptions or rebates unless the admin-istrative costs are excessive, which does not appear to be the casein China. Many studies have either ignored the problem or, as inLejour (2000), treated duty exemptions as simple reductions ininitial tariffs, instead of exemptions on imports used specificallyin the production of exports.

8. Export processors can sell locally, but they have to payduty/VAT on the imported inputs used in the production of out-put sold on the domestic market.

9. The export processing arrangements did not prevent firmsproducing mainly for the domestic market from exporting.These firms produced exports, known as “ordinary” exports,using mainly domestic inputs and only a small portion of duty-or VAT-paid imported materials (Ianchovichina 2003).

10. According to GTAP version 4, 14 percent of imports werefor final consumption and according to China’s Customs 40 per-cent of imports were ordinary imports that were not dutyexempt. This means that approximately 26 percent were ordi-nary imports used as intermediates. Also according to GTAPversion 4, China’s firms exported an average of 10 percent oftheir output, implying that only about 3 percent of imports wereused for production of ordinary exports.

11. The GTAP version 5 database (Dimaranan andMcDougall 2002) is the source for the elasticities of substitutionbetween domestic and composite imported commodities in theArmington production structure of a sector. The values for theseelasticities are shown in column 1 of Appendix table 13.A.4.

12. In a more recent work Sicular and Zhao (2004) estimatethe responsiveness of rural labor supply to changes in agricul-tural returns. They present two “push” elasticities—2.67 fornonagricultural wage employment and 0.24 for nonagriculturalnonwage employment. Focusing on the push elasticity for nona-gricultural wage employment and testing the sensitivity of theresults by replacing the elasticity of 1.32 used in our analysiswith 2.67 leaves the aggregate results largely unchanged (Appen-dix table 13.A.7). The greater responsiveness of labor movementimplied by the larger elasticity of transformation (2.67) trans-lates into better poverty and inequality outcomes since farmwages remain nearly unchanged and an additional 1 millionfarm workers leave farming.

13. Applied rates at the end of the implementation periodwere estimated as the lesser of the bindings and 2001 appliedrates. In virtually all cases, the bindings were lower than theapplied rates.

14. Francois and Spinanger base their estimate of the 20 per-cent productivity increase on the distribution of current plantsin China and apply the formula �ln(AC) = CDR · �ln(Q),where AC is average cost, MC is marginal cost, Q is the quantityproduced, and CDR is the inverse elasticity of scale, defined asCDR = −(AC − MC)/AC, and varies between .125 and .135(the range of values found in engineering studies). Then theycalculate an average cost index for the industry. Assuming thatthe index is 100 at 350,000 units per plant, current plant struc-ture yields a cost index of roughly 120.

15. Mattoo (2004) argues that China’s commitments onservices were the most comprehensive ever made in the WTO.

16. In a recent paper Mai, Horridge, and Perkins (2003)estimate productivity increases over 10 years at 1.8 percent ayear for the strategic manufacturing industries and 2.7 percent ayear for the services sectors as reforms take place under WTOaccession.

17. These quotas are represented in the analysis as an exporttax. In some cases the proceeds of this implicit export tax areredistributed to quota holders, who may be quite different fromthe producers and exporters of the goods. In other cases the quo-tas are auctioned, with the quota rents accruing to the govern-ment. In either case, the marginal return from additional outputof textiles and apparel is net of the quota rent/export tax.

18. This adjustment was made with ALTERTAX (Malcolm1998), so that the consistency and the shares in the GTAP data-base would be preserved.

19. The fixed employment assumption may understate thecosts of accession to some degree. Zhai and Wang, who explorethe impact of WTO accession on migration and unemployment(2002), conclude that structural unemployment may rise follow-ing China’s WTO accession as farmers move to urban areas.

20. The assumption of fixed trade balance as a share of GDPis required when evaluating welfare impacts using a static trademodel such as GTAP.


21. This is a consensus finding supported by Ianchovichinaand Martin (2001), Deutsche Bank (2001), Wang (2002), andIanchovichina and Walmsley (2003).

22. The model underestimates the potential expansion andefficiency increase in the service sectors.With its promise to elim-inate over the next few years most restrictions on foreign entryand ownership, as well as most forms of discrimination againstforeign firms (Mattoo 2004), China has set the stage for increasesin foreign investment and productivity in these sectors. This inturn could lead to much larger income gains from WTO acces-sion and larger increases in wages of skilled workers than shownhere (see Walmsley, Hertel, and Ianchovichina 2004).

23. This estimate represents the number of “effective” farmworkers likely to migrate from rural to urban areas based onemployment data for 2000 from China Statistical Yearbook (NBS2001, pp. 111–12).

24. The consumption tax is close to nondistortionary,because it applies at the same rate to all components of privateand government consumption but not investment. Because theGTAP Version 5 database appears to represent the VAT ondomestic production as an output tax, the model generates taxlosses from the contraction of some industries (e.g., the tobaccoand alcohol industries). These inward-oriented industries havehigher average VAT rates than export-oriented sectors such asclothing because the VAT is exempted on exports. When theexport-oriented sectors expand, the net impact of WTO acces-sion is a sharp contraction in tax revenues. In reality, such a con-traction will not be observed, because VATs of the same magni-tude are levied on imports. To offset this impact, particularly inour poverty analyses, we had to adjust the consumption tax in adownward direction. We first computed the consumption taxthat compensates for the loss in output taxes. This tax as a shareof the total replacement tax is equal to the share of the outputtax loss in the total tax losses. Second, we adjusted the consump-tion tax rate to eliminate the component stemming from thechange in output taxes.

25. High unemployment due to the restructuring of state-owned-enterprises, privatization, and fierce competition inChina imply that WTO accession may dampen the effect on thewages of unskilled workers. By assuming full employment,the model overestimates the increase in wages of workers in thenonfarm sectors, underestimates the fall in the wages of farmworkers, and overestimates the increase in total welfare.

26. These estimates are in agreement with findings inIanchovichina and Martin (2001) and Wang (2002). These areconservative estimates since they do not reflect income increasesresulting from trade- and foreign direct investment–inducedproductivity gains, especially gains associated with liberalizationof China’s service sectors (Walmsley, Hertel, and Ianchovichina2004). Transaction cost savings from developing institutionscompatible with an open and modern market could be verylarge as well, but were not factored into the analysis.

27. Zhai and Wang (2002) obtained similar results for a com-bination of WTO accession and full labor market reform.

28. Since the tax on nonfarm employment of 34 percent rep-resents a bundle of policies that act as a barrier to rural-urbanmigration, this estimate is representative of the likely impact andcould change depending on the policy mix the governmentadopts.

29. The model does not track education spending as a com-ponent of the government budget constraint.

30. Claro (2001) estimates that in the apparel and footwearindustries, the adoption of foreign technology raises productivityby 30–62 percent in collective enterprises and 20–59 percent instate enterprises.

References

Bach, C., and W. Martin. 2001. “Would the Right Tariff Aggrega-tor for Policy Analysis Please Stand Up?” Journal of PolicyModeling 23: 621–35.

Bach, C., W. Martin, and J. Stevens. 1996. “China and the WTO:Tariff Offers, Exemptions and Welfare Implications.”Weltwirtschaftliches Archiv 132(3): 409–31.

Bhattasali, D., Li, Shantong, and Martin, W., eds., 2004. Chinaand the WTO: Accession, Policy Reform, and Poverty Reduc-tion Strategies. Washington, D.C.: World Bank and OxfordUniversity Press.

Chen, S., and M. Ravallion. 2004. “Household Welfare Impactsof China’s Accession to the WTO.” In Bhattasali, D., Li,Shantong and Martin, W., eds., China and the WTO: Acces-sion, Policy Reform, and Poverty Reduction Strategies. Chap-ter 15. Washington, D.C.: World Bank and Oxford UniversityPress.

Claro, S. 2001. “Tariff and FDI Liberalization: What to Expectfrom China’s Entry into WTO?” Paper presented at theEighth Annual Conference on Empirical Investigations inInternational Trade, Purdue University, West Lafayette, Ind.,November 9–11.

Deutsche Bank. 2001. “Quantifying the Impact of China’s WTOEntry.” DB Global Market Research. [www.deutsche-bank.de/index_e.htm.]

Dimaranan, B., and R. McDougall. 2002. Global Trade, Assis-tance, and Production: The GTAP 5 Data Base. West Lafayette,Ind.: Purdue University, Center for Global Trade Analysis.

Francois, J., and D. Spinanger. 2001. “Greater China’s Acces-sion to the WTO: Implications for International Trade/Production and for Hong Kong.” Paper prepared for theHong Kong Trade Development Council, Hong Kong,December 2001.

———. 2004. “Regulated Efficiency, WTO Accession and theMotor Vehicle Sector in China.” This issue.

Gilbert, J., and T. Wahl. 2001. “Applied General EquilibriumAssessments of Trade Liberalization in China.” WorldEconomy 25(5): 697–731.

Hertel, T., ed. 1997. Global Trade Analysis: Modeling and Applica-tions. Cambridge: Cambridge University Press.

Huang, J., S. Rozelle, and C. Min. 2004. “The Nature of Distor-tions to Agricultural Incentives in China and Implicationsof WTO Accession.” In D. Bhattasali, Shantong Li, andW. Martin, eds., China and the WTO: Accession, Policy Reform,and Poverty Reduction Strategies. Chapter 6. Washington,D.C.: World Bank and Oxford University Press.

Hussain, A. 2004. “Impact of WTO Membership on Inequalityand Poverty.” In D. Bhattasali, Shantong Li, and W. Martin,eds., China and the WTO: Accession, Policy Reform, andPoverty Reduction Strategies. Chapter 17. Washington, D.C.:World Bank and Oxford University Press.

Ianchovichina, E. 2003. “GTAP-DD: A Model for AnalyzingTrade Reforms in the Presence of Duty Drawbacks.” GTAPTechnical Paper 21. Purdue University, West Lafayette, Ind.[www.gtap.org/resources/tech_papers.asp] and Journal ofPolicy Modelling (forthcoming).

Ianchovichina, E., and W. Martin. 2001. “Trade Liberalization inChina’s Accession to WTO.” Journal of Economic Integration16(4): 421–45.

Ianchovichina, E., and T. Walmsley. 2003. “Impact of China’sWTO Accession on East Asia.” Policy Research WorkingPaper 3109. World Bank, Washington, D.C.

Lardy, N., ed. 2002. Integrating China in the Global Economy.Washington, D.C.: Brookings Institution Press.


Lejour, A. 2000. “China and the WTO: The Impact on China andthe World Economy.” The Netherlands Bureau for EconomicPolicy Analysis. The Hague, Netherlands.

Li, S., Z. Wang, F. Zhai, and L. Xu. 2000. WTO: China and theWorld. Beijing: China Development Press.

Long, Yongtu. 2000. “On the Question of Our Joining the WorldTrade Organization.” The Chinese Economy 33(1): 5–52.

Mai,Y., M. Horridge, and F. Perkins. 2003.“Estimating the Effectsof China’s Accession to the World Trade Organization.”CoPS General Paper, no. G-137. Centre of Policy Studies andImpact Project, Clayton, Australia.

Malcolm, G. 1998. “Adjusting Taxes in the GTAP Data Base.”GTAP Technical Paper 12. Center for Global Trade Analysis,Purdue University, West Lafayette, Ind. [www.gtap.org/resources/tech_papers.asp.]

Martin, W. 2001. “Implications of Reform and WTO Accessionfor China’s Agricultural Trade Policies.” Economics of Transi-tion 9(3): 717–42.

Martin, W., D. van der Mensbrugghe, and V. Manole. 2003.“Keeping the Devil in the Details: Aggregation and the Bene-fits of Trade Liberalization.” Presented at the Sixth AnnualConference on Global Economic Analysis, The Hague, June12–14. Available at http://www.gtap.agecon.purdue .edu.

Mattoo, A. 2004. “China’s Accession to the WTO: The ServicesDimension.” In D. Bhattasali, Shantong Li, and W. Martin,eds., China and the WTO: Accession, Policy Reform, andPoverty Reduction Strategies. Washington, D.C.: World Bankand Oxford University Press.

McDougall, R., A. Elbehri, and T. Truong. 1998. Global Trade,Assistance, and Production: The GTAP 4 Data Base. PurdueUniversity, Center for Global Trade Analysis, West Lafayette,Ind.

McKibbin, W., and K. Tang. 2000. “Trade and Financial Reformin China: Impacts on the World Economy.” World Economy23(8): 979–1003.

Messerlin, P. 2004. “China in the WTO: Antidumping andSafeguards.” In D. Bhattasali, Shantong Li, and W. Martin,eds., China and the WTO: Accession, Policy Reform, andPoverty Reduction Strategies. Washington, D.C.: World Bankand Oxford University Press.

NBS (National Bureau of Statistics of China). 2001. China Sta-tistical Yearbook. People’s Republic of China. China StatisticsPress.

Shi, Xinzheng. 2002.“Empirical Research on Urban-Rural IncomeDifferentials: A Case of China.” Peking University, Beijing.

Sicular, T., and Y. Zhao. 2002. “Employment, Earnings andLabor Market Responses in Rural China.” Mimeo, PekingUniversity and the University of Western Ontario.

———. 2004. “Labor Mobility and China’s Entry to the WTO.”In D. Bhattasali, Shantong Li, and W. Martin, eds., China andthe WTO: Accession, Policy Reform, and Poverty ReductionStrategies. Washington, D.C: World Bank and OxfordUniversity Press.

Walmsley, T. L., and T. W. Hertel. 2001. “China’s Accession to the WTO: Timing is Everything.” World Economy 24(8):1019–49.

Walmsley, T. L., T. Hertel, and E. Ianchovichina. 2004. “Assessingthe Impact of China’s WTO Accession on Foreign Owner-ship.” Pacific Economic Review (forthcoming).

Wang, Z. 2002.“WTO Accession, ‘Greater China’ Free Trade Areaand Economic Relations across the Taiwan Strait.” Paper pre-sented at the Fifth Conference on Global Economic Analysis,June 5–7, Taipei.

———. 2003.“The Impact of China’s WTO Accession on Patternsof World Trade.” Journal of Policy Modeling 25(1): 1–41.

World Bank. 1997. China Engaged: Integration with the WorldEconomy. Washington, D.C.

———. 1999. World Development Indicators. Washington, D.C.———. 2002. China: Country Economic Memorandum.

Washington, D.C.WTO (World Trade Organization). 1994. The Results of the

Uruguay Round of Multilateral Trade Negotiations. Geneva.———. 2001. “Accession of the People’s Republic of China.”

WT/L/432 (www.wto.org).Zhai, Fan, and S. Li. 2000. “The Implications of Accession to

WTO on China’s Economy.” Paper presented at the ThirdAnnual Conference on Global Economic Analysis, June28–30, Monash University, Melbourne, Australia.

Zhai, Fan, and Z. Wang. 2002. “WTO Accession, Rural LabourMigration and Urban Unemployment in China.” UrbanStudies 39(12): 2199–217.

Part III

Impacts onHouseholds

and on Poverty

benefits agriculture and if land holdings are highlyconcentrated, then inequality could increase. Third,the distributional impact of trade liberalizationdepends on the flexibility of domestic markets,especially for labor. Gains from trade liberalizationare less likely to be shared equally where labor mar-kets are segmented and barriers hinder the move-ment of labor among sectors.

These considerations are relevant to the impactof China’s accession to the World Trade Organiza-tion (WTO) and its concomitant trade liberaliza-tion on employment and incomes in China. Indeed,the Chinese case has some interesting features.China’s labor markets have historically (undersocialism) been inflexible and highly segmented.Domestic economic reforms have allowed greaterlabor mobility, but many observers believe thatsubstantial institutional barriers to the movementof labor persist. Also, in China certain assets such asland and education, while not equally distributed,are nevertheless relatively equally distributed bydeveloping country standards. These two featureswould have counterbalancing effects in that theformer would tend to cause the gains from WTOaccession to be concentrated, while the latter wouldtend to cause the gains to be shared more broadlycompared with the gains from similar liberalizationin other countries.

The literature on trade liberalization in developingcountries contains divergent views on the impactof liberalization on employment, incomes, andpoverty. Although most studies find aggregate wel-fare gains, they disagree over the distribution ofthese gains among households. One view is thattrade liberalization generates broad-based employ-ment gains across regions for skilled and unskilledlabor, and that consequently income gains areshared widely (Dollar and Kraay 2001). Underthese conditions, trade liberalization contributesto reductions in income inequality and poverty.The alternative view is that employment andincome gains go disproportionately to the alreadybetter-off groups, with negative implications forinequality and perhaps also for the poor (Rodrik2000).

The reality probably lies somewhere in between,with the distributional outcome depending on spe-cific conditions in the country in question (Winters2000). First, it depends on the level and structure ofpre-liberalization trade barriers, which determinethe sectors that gain and lose. A differential impacton sectors holds implications for the distribution of gains among regions, skill levels, and incomegroups. Second, it depends on the preexisting dis-tribution of assets—that is, of land, capital, andhuman capital. So, for example, if trade liberalization

239

14

Earnings and Labor Mobility in Rural

China: Implicationsfor China’s Accession

to the WTO

Terry Sicular and Yaohui Zhao

This chapter examines the microeconomic deter-minants of rural employment and incomes in China.More specifically, using survey data we estimateincome, wage, and labor supply functions for ruralhouseholds. Because the households derive incomefrom agriculture and sideline family businesses andsince few households hire labor to undertake suchactivities (Bowlus and Sicular 2003), we must imputeshadow wages derived from self-employment.Together, the income, wage, shadow wage, and laborsupply functions empirically describe householdincome generation from employment.

Our analysis fills a gap in the literature. Eventhough the literature examining employment andearnings in rural China is now quite substantial,most studies of China’s rural employment analyzethe determinants of occupational status—that is,whether or not individuals participate in differenttypes of work such as wage jobs or nonagriculturalsidelines (examples are Hare 1994, 1999a, 1999b;Knight and Song 1997, 1999; Michelson and Parish2000, Parish, Zhe, and Li 1995; Rozelle and others1999; Zhao 1999a, 1999b). Relatively few studiesestimate the rural labor supply per se—that is,hours or days worked (examples are Knight andSong 1997, and Yao 1999). Even fewer studies esti-mate labor supply as a function of wages, the rela-tionship of greatest interest here.

To our knowledge, the only study that estimateslabor supply for rural China and includes a measureof the wage as an explanatory variable is by Meng(2000). Like Meng, we estimate time worked as afunction of wages and other variables. Unlike Meng,and indeed unlike the relevant literature for otherdeveloping countries (e.g., Jacoby 1993; Skoufias1994), our labor supply functions allow for the pos-sibility that wages or shadow wages, and labor’sresponse to these wages, can differ depending on thetype of wage. Thus it is possible for labor supply tobe more responsive to the market wage than to theagricultural shadow wage, or vice versa. Finally,although other studies estimate total labor supply,we estimate total labor supply and its components—in this case, labor supply to household agriculturalproduction, to household nonagricultural produc-tion, and to market or wage employment. By esti-mating the components of labor supply, we obtaininformation about how wages and other variablesinfluence the composition of employment.

Our empirical results provide some of theunderlying parameters needed to understand the

effects of trade liberalization on levels of employ-ment and earnings. The income generation func-tions give estimates of the impact of agriculturalversus nonagricultural employment on incomefrom labor. We find that nonagricultural employ-ment generates substantially more income per hourworked than does agricultural employment. Thisresult is consistent with the findings of other studies(e.g., Knight and Song 1997; Meng 2000; Michelsonand Parish 2000), which generally conclude thatincome inequality among households reflects dif-ferences in access to higher paying, off-farm jobs.

We take this analysis one step further anddecompose the income gap between households inricher and poorer subsamples. Our decompositionreveals that the income gap is accounted for byboth differences in hours worked in different occu-pations and differences in the estimated returnsto labor in different occupations—that is, poorerhouseholds are poorer both because they supplyless labor to higher-paying wage employment andbecause the returns to the labor they supply in eachoccupation, and their returns to education, arelower. These findings suggest that the impact ofChina’s accession to the WTO on income distribu-tion will depend on how it affects the distributionof employment, wages, and the returns to educa-tion among households.

Our labor supply estimates reveal how workhours, and the composition of work hours, wouldrespond to changes in wages for agricultural andnonagricultural work. We find that wages do nothave a large effect on the total number of hoursworked. They do, however, significantly influencethe composition of hours worked. Our estimates oflabor supply by type of employment—in house-hold agriculture, household nonagriculture, andoff-farm wage jobs—indicate that most own- andcross-elasticities with respect to the different wagesare well below one. This finding suggests thatalthough labor moves among these types ofemployment, mobility is limited.

A notable exception to this pattern is the cross-elasticity with respect to the agricultural shadowwage of labor supplied to the market. This elasticityis significant and large, indicating that differencesin labor supply for wage employment amonghouseholds is driven by differences in the returns tolabor in household farming. Put differently, highmarket wages do not “pull” labor out of agriculture;rather, low marginal returns to work in agriculture


“push” labor into wage employment. We discussthese results more fully later in this chapter.

The next section of this chapter presents anoverview of aggregate trends in China’s ruralemployment, earnings, and labor markets, payingspecial attention to the institutions and policies thataffect labor mobility and the distribution of earn-ings. A description of the dataset follows. The nextsection examines household income generationand analyzes the income gap between richer andpoorer households. We then turn to our economet-ric estimates of wage and labor supply functions.

Throughout this chapter our focus is on ruralhouseholds. China’s accession to the WTO is affect-ing both urban and rural households, but the insti-tutional setting and economic behavior of thesetwo types of households are substantially differentand require separate analysis. Analysis of ruralhouseholds is important because the rural sectorcontains most of China’s population and also mostof its poor. Indeed, a disproportionate share ofChina’s poor population is located in rural areas.1

Recent estimates for 1999 by Chen and Wang(2001) report a poverty rate (using the $1/daypoverty line) of 24.9 percent in rural areas versusonly 0.5 percent in urban areas. Despite increasedurbanization, 64 percent of the population and74 percent of employed persons are still classifiedas rural (National Bureau of Statistics 2001, pp. 37,39). Average per capita income in rural areas is onlyabout one-third that in urban areas, and evidencesuggests a widening of the urban-rural income gapin recent years (Yang and Zhou 1999).

Rural Employment and Earnings inChina in the 1990s: The AggregatePicture

A review of aggregate trends in rural employmentand earnings in China provides a broad context inwhich to interpret our microeconomic results.Aggregate trends reflect the impact of major policyreforms, including domestic market liberalizationas well as trade liberalization in advance of WTOaccession, and so they also provide some cluesabout the potential impact of WTO accession onemployment and earnings.

Developments in the 1990s on both the supplyand demand side suggest movement toward fulleremployment and higher earnings in rural areas. Onthe supply side, in the 1990s China’s labor force

grew slowly both overall and in rural areas. Becauseof population planning policies and a markeddecline in fertility in the 1970s, cohorts entering thelabor force in the 1990s were smaller than those inthe 1970s and 1980s. New cohorts entering thelabor force should remain small or even declinefurther in the coming years because of the strictpopulation control policies adopted in the 1980s.

On the demand side, macroeconomic growthcombined with policy liberalization generated newjob opportunities for rural workers. Rural nonagri-cultural employment grew substantially. The offi-cial data on such employment is problematic, but itdoes provide a rough indication of trends. Ruralnonagricultural employment includes employmentin township and village enterprises (TVEs) and inprivate and individual enterprises.2 During the1990s, employment in these enterprises grew, onaverage, 5.4 percent annually. In absolute terms,during the 1990s the increase in rural enterpriseemployment exceeded 66 million jobs, and by 1999this employment was equivalent to one-third of therural labor force.

A notable aspect of growth in rural enterpriseemployment is that nearly half of it stemmed fromthe expansion of private and individual enterprises.By 1999 these enterprises were employing over 45million people, equivalent to about 10 percent ofthe rural labor force. Although some of this growthcould be the result of reclassification of collectiveTVEs as private businesses, it also reflects the grow-ing importance of household-based nonagricul-tural activities.

Also notable during the 1990s was a rise inrural–urban migration. After a relaxation of restric-tions on labor movement, rural–urban migrationappears to have grown substantially. Data onmigration are spotty, and definitions of what con-stitutes migration differ (see Wu and Zhou 1996;Rozelle and others 1999), but most estimates sug-gest at least a doubling of the number of migrantsbetween the late 1980s and the mid- or late 1990s.Sources suggest that by the mid- or late 1990s thenumber of migrants (excluding commuters) prob-ably exceeded 50 million, or about 10 percent of thenumber of rural employed persons (Wu and Zhou1996; Zhao 1999b).

Altogether, then, growth in employment byrural TVEs, by private and individual enterprises,and through migration increased from perhaps 130million in the early 1990s to roughly 230 million in

Earnings and Labor Mobility in Rural China: Implications for China’s Accession to the WTO 241

the late 1990s. By then, nonagricultural employmentof rural residents had risen from less than 30 per-cent to nearly 50 percent of the number of ruralemployed persons.

This substantial expansion of nonagriculturalemployment has spurred some debate about thenature of labor markets in rural China. Someauthors argue that labor markets are now fairly openand competitive, with considerable labor mobility(Rawski and Mead 1998). Others, however, arguethat while open and competitive labor markets areemerging in some regional pockets, and while gen-erally the direction of change has been toward moreopen, competitive conditions, rural labor marketsnevertheless continue to be imperfect, and institu-tional and administrative barriers persist (Parish,Zhe, and Li 1995; Knight and Song 1999). Theseissues are relevant to our analysis and are discussedfurther later in this chapter.

Discussion of employment cannot be completewithout mention of the largest employer, agricul-ture. Official statistics reveal cycles in agriculturalemployment. The number of rural employed per-sons rose to more than 340 million in 1991–92, fellto less than 325 million in 1995–97, and then roseagain to almost 330 million in 1999–2000 (NationalBureau of Statistics, various years). Interestingly,the upswings in these cycles occurred at the sametime as downswings in rural enterprise employ-ment, and vice versa, which suggests an inverse rela-tionship between the two kinds of employment.

The official data on rural employment are, how-ever, problematic. They simply count the numberof people by primary occupation. They do not cap-ture the fact that many rural workers engage inmultiple occupations and that hours worked in anyparticular occupation can fluctuate over time.Some studies indicate that changes in hours workedin agriculture have been significant (Rawski andMead 1998; World Bank 2001). Estimates of agri-cultural employment based on hours worked indi-cate that in the mid-1990s, when rural enterpriseemployment was expanding rapidly, agriculturallabor days worked actually rose (World Bank2001). These estimates indicate that growth innonagricultural employment and agriculturalemployment can occur concurrently. The relation-ship between these two kinds of employment is important to our analysis, because the willing-ness and ability of rural residents to supply labor

hours to different sectors can affect the impact oftrade liberalization.

What have been the effects of the just-describedtrends in employment on rural incomes andinequality? On average, the real per capita incomeof rural households rose during the 1990s. Between1990 and 1999, net income per capita rose about70 percent. This growth was derived from multiplesources. Wages contributed the largest share ofthe increase (38 percent), followed by agriculture(29 percent), and household nonagriculturalsidelines and businesses (22 percent).3 Althoughnonagricultural sources dominated, the fact thatagriculture contributed nearly one-third of theincrease in income is notable in view of the factthat, according to the official data, agriculturalemployment supposedly declined.

The average trends just outlined mask changesin distribution among poorer and richer house-holds. The general consensus is that rural incomeinequality increased during the 1990s, whilepoverty decreased. These apparently contradictorydevelopments reflect income growth for the poor,but faster income growth for richer groups.

A 2001 World Bank study outlines key aggregatefactors underlying the recent decline in ruralpoverty, some of which are relevant from the per-spective of China’s accession to the WTO (WorldBank 2001). First, aggregate growth in the grossdomestic product (GDP) appears to be important.Nationally, the reduction in poverty coincided witha period of rapid GDP growth, and poverty reduc-tion occurred more quickly in those regions thatexperienced the most rapid aggregate growth. Sec-ond, the composition of growth matters. The rateof poverty reduction has been faster in regionswhere agricultural growth has more or less keptpace with growth in other sectors. This situationreflects the fact that agriculture is the primarysource of employment and income for the poor.These points suggest that the effect of WTO acces-sion on poverty will be influenced by its impact onboth aggregate growth and the sectoral composi-tion of that growth.

The Survey Sample

For our empirical analysis, we use data from theChina Health and Nutrition Survey (CHNS).CHNS data were collected through an independent


survey conducted by an international team ofresearchers collaboratively sponsored by theCarolina Population Center at the University ofNorth Carolina at Chapel Hill, China’s Institute ofNutrition and Food Hygiene, and the ChineseAcademy of Preventive Medicine. Data are availablefor four years: 1989, 1991, 1993, and 1997. The sur-vey, which used a multistage, random cluster sam-pling method, covers about 3,800 households with14,000 members in nine provinces with differentgeographic and economic characteristics.4 TheCHNS also offers information on community-levelvariables such as market prices, health facilities, andsocial services.5

The CHNS survey data are useful here becausethey provide detailed information on incomes andhours worked in different occupations as well as on awide range of relevant individual, household, andcommunity characteristics. The sample includes ashousehold members migrant workers who work andlive out of town,but whose earnings and expenses areconsidered part of the household’s. The data do not,however, allow us to distinguish between work andincome from migrant versus local employment.Thus although we are able to investigate mobilityamong types of work (in household agriculturalproduction, household nonagricultural production,and wage employment), we cannot investigategeographic mobility.

For our estimations we use the most recent(1997) data, and we drop urban households—thatis, we use a subsample of households that includesonly those that reside in rural and suburban vil-lages and in county towns. We include suburbanvillages and county towns because they have closeties with rural areas, and a significant portion of thepopulation in these areas holds a rural residenceregistration or hukou. This subsample covers 3,239households with 8,590 working-age adults. Becausewe are interested in labor supply, we further restrictour sample to households that contain at least onehousehold member who works. This reduces thesubsample to 2,998 households with 8,326 working-age adults. The number of observations actuallyused in our analysis varies among regressions,depending on the extent of participation in theactivity being analyzed and on the prevalence ofmissing values in relevant variables.

Table 14.1 contains descriptive statistics for thenonurban subsample of the CHNS survey. It also

gives some comparable statistics from the 1997National Bureau of Statistics (NBS) official ruralhousehold survey, where available. Household sizeand structure are similar for the NBS data and forCHNS subsample. Income levels for the CHNSsubsample are somewhat lower than for the NBSsurvey, but this disparity could reflect differences inhow income is calculated. The structure of incomealso differs. In particular, agricultural income isnoticeably higher in the NBS survey than in theCHNS samples. This difference could reflect theinclusion of households in suburban villages andcounty towns in the CHNS subsample but not inthe NBS rural survey. For both the NBS and CHNSsamples, the major sources of income are agricul-ture and wage employment.

The CHNS data provide some informationabout relative earnings in different sectors. Dividingaverage earnings by hours worked for each sectorsuggests that the returns to labor in agriculture arelower than in other occupations. Average net earn-ings per hour worked in agriculture are roughly1.1 yuan as compared with 2.2 yuan in nonagricul-tural sidelines and 2.8 yuan in wage employment.When we use regression analysis to estimate thereturns to labor in the next section, these earningsdifferentials become larger.

Income Generation Functions

A common approach to analyzing the determinantsof income is to estimate an income generationfunction, where net income is a function of laborinputs, land and capital assets, and other householdor regional characteristics that contribute to thegeneration of earnings. This approach takes incomegeneration as a simple, linear accounting relation-ship, where income equals the sum of householdlabor and other assets times the returns to thoseassets. The regression coefficients provide estimatesof the marginal returns to each asset. We use thisapproach both to examine income generation onaverage and to explore differences between poorerand richer subgroups.

The first columns (Model I) of table 14.2 con-tain results from income generation regressions forall households with at least one working adult inthe CHNS nonurban subsample. The dependentvariable is the sum of net earnings from the threeoccupations—that is, net earnings (revenue minus



TABLE 14.1 Descriptive Statistics, China Health and Nutrition Survey, 1997

NBS RuralHousehold CHNS Nonurban Households

Survey, Standard Number ofVariable 1997 Mean Deviation Observations

Household size 4.35 3.83 1.42 3,239Number of adults 2.79 2.67 1.29 3,212Number of dependents per capita 0.30 0.26 3,212Number of male adults 1.36 0.80 3,212Number of female adults 1.31 0.76 3,212Number of male laborers (hours > 0) 1.18 0.73 3,239Number of female laborers (hours > 0) 1.10 0.72 3,239Mean age of working-age adults 35.39 14.30 3,238Mean schooling of working-age adults 6.94 2.81 3,033Ratio of male working-age adults to family 0.51 0.18 3,041

working-age adultsCultivated land (mua) 6.84 12.14 1,807Household total income (yuan) 9,094 9,871 1,966Household net earnings (yuan) 9,092 7,774 9,784 2,160Household net earnings from agricultural 5,081 2,069 5,095 2,380

sectors (yuan)Household net earnings from nonagricultural 1,009 1,267 4,627 3,080

sectors (yuan)Household earnings from employed 2,240 3,830 7,835 3,069

work (yuan)Household nonlabor income (yuan) 1,397 2,868 2,808Household total labor hours 3,786 2,874 2,904Household total labor hours in household 1,803 2,203 3,054

agricultural productionHousehold total labor hours in household 572 1,504 3,170

nonagricultural productionHousehold total labor hours in wage 1,359 2,115 3,122

employmentHeilongjiang 0.12 3,239Jiangsu 0.12 3,239Shandong 0.12 3,239Henan 0.13 3,239Hubei 0.12 3,239Hunan 0.12 3,239Guangxi 0.13 3,239Guizhou 0.13 3,239

aThe mu is a Chinese measure of area. Fifteen mu equal 1 hectare. Note: The National Bureau of Statistics (NBS) data are taken from the China Statistical Yearbook. The NBSrural survey sample includes only households in rural villages but in all provinces, whereas the subsampleof the China Health and Nutrition Survey (CHNS) data includes households in suburban villages andcounty towns but only in eight provinces. Note that income is calculated differently for the NBS andCHNS samples. Also, the NBS household income data are in per capita terms, and we calculate NBShousehold total income as mean household size times mean income. This method could create somebias, because household size and income per capita are typically inversely correlated. Nevertheless,comparison of the CHNS means with the NBS means gives a rough indication of how the CHNS samplecompares with the official national rural survey sample.Sources: National Bureau of Statistics, various years; authors’ calculations using CHNS data.


TAB

LE 1

4.2

Inco

me

Gen

erat

ion

Fu

nct

ion

s (d

epen

den

t va

riab

le: h

ou

seh

old

net

ear

nin

gs)

Mod

el I:

All

Hou

seho

lds

Mod

el II

: Bot

tom

20

Perc

ent

Mod

el II

I: To

p 20

Per

cent

Stan

dar

d

Stan

dar

dSt

and

ard

Co

effi

cien

tEr

ror

Co

effi

cien

tEr

ror

Co

effi

cien

tEr

ror

Agr

icul

tura

l p

rodu

ctio

n la

bor

hour

s0.

226*

*0.

101

�0.

043

0.13

20.

510*

0.27

4N

onag

ricul

tura

l p

rodu

ctio

n la

bor

hour

s1.

715*

**0.

137

1.26

6***

0.21

21.

958*

**0.

304

Wag

e em

plo

ymen

t la

bor

hour

s2.

372*

**0.

111

2.11

3***

0.15

33.

004*

**0.

230

Mea

n sc

hool

ing

of w

orki

ng-a

ge a

dults

241.

438*

**84

.215

167.

649

116.

527

395.

775*

*18

8.01

0Ra

tio o

f m

ale

wor

king

-age

adu

lts t

o �

206.

177

1,18

5.77

63,

709.

768*

*1,

597.

452

�44

7.15

72,

679.

677

fam

ily w

orki

ng-a

ge a

dults

La

nd75

.514

***

17.6

3620

0.76

4***

48.8

7317

3.03

0***

37.7

20To

tal

equi

pm

ent

(1,0

00 y

uan)

150.

298*

**16

.869

231.

095*

**58

.782

109.

221*

**26

.446

Mea

n ag

e of

wor

king

-age

adu

lts33

.799

26.7

2757

.601

38.5

7746

.073

57.4

21H

eilo

ngjia

ng77

0.72

172

8.72

3�

3,92

2.14

9*2,

331.

068

�4,

998.

626

3,41

6.96

6Jia

ngsu

1,98

2.58

9***

717.

698

�3,

473.

656

2,36

9.04

5�

3,59

4.46

83,

386.

499

Shan

dong

686.

911

851.

888

�4,

332.

018

3,37

4.84

4�

3,30

4.31

13,

352.

075

Hen

an�

1,01

5.65

174

8.15

8�

4,46

6.83

3**

2,28

8.17

1�

8,19

1.83

3**

3,65

8.16

5H

ubei

98.7

3969

8.34

6�

3,14

5.44

62,

242.

68�

5,58

6.84

7*3,

375.

095

Hun

an2,

167.

564*

*90

5.10

1�

1,26

2.59

42,

279.

672

1,27

0.76

74,

252.

05G

uang

xi�

480.

419

750.

139

�4,

669.

35*

2,50

2.90

3�

2,78

1.75

73,

617.

45G

uizh

oun.

a.n.

a.�

2,52

0.56

42,

185.

661

�5,

023.

888

4,36

8.55

4C

onst

ant

�1,

083.

532

1,51

7.58

n.a.

n.a.

n.a.

n.a.

Adj

uste

d R-

squa

re0.

352

0.72

20.

757

Num

ber

of o

bser

vatio

ns1,

671

333

349

n.a.

Not

ap

plic

able

. **

*p<

.01,

**p

< .0

5,*p

< .1

0.N

ote:

Gui

zhou

is t

he o

mitt

ed r

egio

nal d

umm

y va

riabl

e in

Mod

el I.

In M

odel

s II

and

III a

ll re

gion

al d

umm

y va

riabl

es a

re in

clud

ed, a

nd t

he c

onst

ant

term

is s

upp

ress

edto

allo

w a

naly

sis

of t

he r

egio

nal c

ontr

ibut

ion

in t

he O

axac

a de

com

pos

ition

. To

min

imiz

e sa

mp

le s

elec

tion

bias

, we

grou

p h

ouse

hold

s on

the

bas

is o

f ave

rage

vill

age

per

cap

ita h

ouse

hold

inco

me

as r

epor

ted

by v

illag

e le

ader

s in

a c

omm

unity

que

stio

nnai

re. T

hus

the

bott

om (

top

) 20

per

cent

ref

ers

to a

ll ho

useh

olds

tha

t liv

e in

villa

ges

that

hav

e av

erag

e p

er c

apita

hou

seho

ld n

et in

com

e in

the

bot

tom

(to

p)

20 p

erce

nt o

f the

inco

me

dist

ribut

ion

for

villa

ges

in t

he s

urve

y.So

urce

: Aut

hors

’ est

imat

es u

sing

CH

NS

data

.

nonlabor variable costs) from household agricul-tural production, net earnings from householdnonagricultural production, and wage income.Explanatory variables include actual labor hours ineach of household agricultural production, house-hold nonagricultural production, and wageemployment, as well as education, ratio of males tofemales, age, land and equipment assets to capturenonlabor inputs that contribute to income fromhousehold production, and provincial dummyvariables. The mean values of these variablesappear in table 14.3.

Of central interest are the returns to labor indifferent occupations. The estimated coefficient foragricultural hours of work is small, 0.23, and signifi-cant at the 5 percent level. The coefficient for laborhours in nonagricultural household production is1.72, and that for hours in wage employment is 2.37,both statistically significant at the 1 percent level.

These results indicate that income differencesamong households are generated more by nonagri-cultural work than by agricultural work; further-more, wage employment brings the highest returns.6

These results are consistent with the findings of otherstudies for China (Knight and Song 1997; Meng2000; Michelson and Parish 2000).

Education has a significant effect on householdincome. An additional year of average schooling forworking-age adults in the household increaseshousehold net income by 241 yuan, roughly 3 per-cent of average net income. Again, this result isconsistent with findings in other studies.

The average age of household workers and thesex composition of the work force do not have signif-icant effects on household earnings. As expected,land and equipment assets have significant, posi-tive effects. An additional mu of land (a Chinesemeasure in which 15 mu equal 1 hectare) increases


TABLE 14.3 Descriptive Statistics for Variables Used in the Income GenerationFunctions, CHNS, 1997

All Households Used in the Income

Generation Function Bottom 20 Percent Top 20 Percent

Standard Standard StandardVariable Mean Deviation Mean Deviation Mean Deviation

Household net earnings (yuan) 8,113.19 9,746.09 5,955.50 6,265.58 10,966.54 9,907.75

Household agricultural production labor 1,888.35 2,261.85 2,760.06 2,562.277 778.59 1,642.21hours

Household nonagricultural production 573.11 1,543.39 476.76 1307.40 529.67 1,484.76labor hours

Wage employment labor hours 1,578.73 2,093.83 1,127.63 1,883.53 2,525.19 2,217.55

Household total labor hours 4,040.18 2493.40 4,364.45 2,599.83 3,833.45 2,337.51

Mean schooling of working-age adults 7.20 2.83 6.77 2.91 8.17 2.55

Ratio of male working-age adults to 0.51 0.16 0.52 0.16 0.50 0.15family working-age adults

Cultivated land (mu) 4.80 12.16 4.85 6.75 5.72 13.75

Equipment (1,000 yuan) 2.24 11.64 1.06 4.36 2.92 15.18

Mean age of working-age adults 36.84 7.85 36.20 7.73 37.42 7.54

Heilongjiang 0.16 0.14 0.23

Jiangsu 0.15 0.08 0.23

Shandong 0.09 0.01 0.20

Henan 0.11 0.17 0.04

Hubei 0.14 0.24 0.14

Hunan 0.07 0.10 0.03

Guangxi 0.11 0.06 0.10

Guizhou 0.17 0.22 0.03

Number of observations 1,671 333 349

Source: Authors’ estimates using CHNS data.

net income by 76 yuan or 1 percent of mean income,and an additional 1,000 yuan of equipment assetsincreases net income by 150 yuan or 2 percent ofmean income. The presence of significant, large coef-ficients on some provincial dummy variables indi-cates that location of residence is also important.

We explore differences between poorer and richerhouseholds using a modified Oaxaca-type decom-position. This method involves first estimatingincome generation functions for richer and poorersubgroups. To minimize the problem of endogenoussample selection, we divide the sample using averagevillage incomes reported by village leaders in a sepa-rate community questionnaire. Models II and III intable 14.2 contain the results of income generationequations for households in villages in the top 20 per-cent and bottom 20 percent of the village average percapita net income distribution.

The two groups have significantly different coef-ficients. The returns to labor and education aresubstantially higher in the top 20 percent than thebottom 20 percent. For the bottom 20 percent,the returns to agricultural labor are not signifi-cantly different from zero. This finding would beconsistent with a standard surplus labor story. Bycontrast, for the top 20 percent the coefficient foragricultural labor is positive and significant, albeitsmall. The nonagricultural and wage labor groupshave positive and significant returns, but thereturns are lower for the poorer subsample (bottom20 percent).

Schooling has positive coefficients everywhere,but the return to education is smaller and insignifi-cant for the poorer subsample. The ratio of malesamong working-age adults is insignificant exceptfor the poorer group, for which it is large, positive,and significant. The presence of male workers thusappears to be most important in poorer areas. Thereturns to land and equipment are all positive andsignificant, with the magnitudes higher for the bot-tom 20 percent than the top 20 percent. This differ-ence could stem from the greater scarcity of landand capital in poorer villages than in richer villages.

The Oaxaca decomposition combines these esti-mated coefficients with the mean values of the vari-ables for the two groups. The mean values of house-hold characteristics reveal substantial differencesbetween households in richer and poorer villages(table 14.3). Average net earnings of the rich groupare 1.8 times those of the poor group, and average

education is 1.4 years longer. Land assets are 0.9 mularger (although not corrected for land quality),and equipment assets are substantially larger, too.

Employment also differs. Although both groupswork in all three sectors, household agriculturedominates for the poorer group and wage employ-ment dominates for the richer group. Overall,households in the poorer group work more hours.Total work time for households in the poorer groupis 4,364 hours, about 14 percent higher than the3,833 hours total work time for households in thericher group.7

Table 14.4 gives the results of our Oaxaca-typedecomposition. Although conventional decomposi-tion combines the contribution of regional dummyvariables together with that of other characteristics,we separate out the effects of regional differences onthe income gap. The overall difference of 5,011 yuanin mean net earnings is thus decomposed into threecomponents: (1) the income gap due to differencesin levels of household productive characteristicsother than regional location, (2) the income gapdue to differences in returns to these householdproductive characteristics, and (3) the income gapdue to regional factors.

The results show that all three components con-tribute to the income gap, but the differences inproductive characteristics are most important. Thiscomponent explains 82 percent of the income gapbetween the two groups. The largest contributorhere is the difference in wage employment hours,which accounts for 71 percent of the income gap.The richer households have fewer labor hours inagriculture (see table 14.4), which serves to reducethe income gap by 9 percent. Differences in educa-tion levels explain 8 percent of the gap. Differencesin mean endowments of land and productiveequipment, and in household structure and age, arerelatively unimportant.

Differences in the returns to productive charac-teristics—that is, differences in the estimated coef-ficients of the two groups—account for 34 percentof the income gap. The largest contributor here isthe difference in the returns to education, which byitself explains 34 percent of the income gap. Nextmost important is the difference in returns to wageemployment, which explains 32 percent of theincome gap. Differences in the returns to agricul-tural labor and sex composition of working-ageadults are also important.


Regional location, like other household charac-teristics, contributes to the income gap in two ways.First, the returns to region of residence or esti-mated coefficients for the provincial dummy vari-ables differ between the rich and poor groups. And,second, the regional distribution of the poor differsfrom the regional distribution of the rich. Togetherthese regional factors explain about 17 percent ofthe overall income gap. Virtually all of this is due todifferences in the returns to location.

In summary, our decomposition analysis showsthat households in poor villages are relatively poorboth because they have different characteristics than

richer households and because the returns to thosecharacteristics are lower. Here the key contributorsto the income gap are a lower supply of wage laborby households in poorer villages, as well as the lowerreturns received by poor households for education,wage employment, and agricultural labor.

New Estimates of Labor Supplyand Allocation

The estimates and decomposition just describedindicate that returns to labor and pattern ofemployment are important factors underlying


TABLE 14.4 Decomposition of Household Net Earnings Differentials betweenHouseholds Belonging to the Top and Bottom 20 Percent

Contributions to Income Gap between Households in the Top 20 Percent and

Bottom 20 Percent of Villages

Yuan Percent

Household net earnings, total gap 5,011.04 100.00

Amount due to(1) Differences in productive characteristics 4,116.19 82.14(2) Differences in returns to productive characteristics 1,723.47 34.40(3) Differences in regional location and returns to regional �828.61 �16.54

location(a) Agricultural production labor hours �462.40 �9.23(b) Nonagricultural production labor hours 85.28 1.70(c) Wage employment labor hours 3,575.93 71.36(d) Mean schooling of working-age adults 394.56 7.87(e) Ratio of male working-age adults to family �19.45 �0.39

working-age adults(f) Land 162.50 3.24(g) Total equipment (1,000 yuan) 316.56 6.32(h) Mean age of working-age adults 63.22 1.26

(2) Differences in returns to productive characteristics (total) 1,723.48 34.4(a) Agricultural production labor hours 978.92 19.54(b) Nonagricultural production labor hours 348.15 6.95(c) Wage employment labor hours 1,627.69 32.48(d) Mean schooling of working-age adults 1,703.70 34.00(e) Ratio of male working-age adults to family �2,121.71 �42.34

working-age adults(f) Land �146.48 �2.92(g) Total equipment (1,000 yuan) �242.43 �4.84(h) Mean age of working-age adults �424.36 �8.47

(3) Differences in regional location and returns to �828.62 �16.54regional location (total)(a) Differences in regional location 93.21 1.86(b) Differences in returns to regional location �921.83 �18.40


income differences. This finding raises questionsabout what explains the level and pattern of laborsupply and why the returns to labor vary amongsectors and households. Clearly, these two ques-tions are interrelated, because household labor sup-ply may depend on the returns to labor, and thereturns to labor (especially in household produc-tion) may depend on household production deci-sions, including decisions on the allocation oflabor. Here we pursue these questions by empiri-cally analyzing the determinants of the returns tolabor and of labor supply, and we employ instru-mental variable methods to address the endogene-ity of key explanatory variables.

We begin with a model of time allocation byrural households. The family is assumed to haveendowments of workers, land, and other assets.The household may allocate its resources amongthree possible income-generating activities: agricul-tural production, nonagricultural production, andwage or market employment. In the first two activi-ties, the household organizes production usingcertain quantities of family assets and labor andperhaps also purchased inputs, hired assets, andhired labor. In market employment the householdfaces an exogenous wage rate set by the market.

Labor supply decisions are the outcome ofhousehold utility maximization, where utility is afunction of the consumption of leisure and goodsand the budget constraint depends on full income,which includes profits from household productionas well as the value of the household’s time endow-ment. Where markets function well and factors areperfectly mobile, households maximize utility byfirst maximizing profits at market prices and thendeciding on optimal levels of leisure and goods con-sumption given market prices and their endow-ments. Labor allocation to the different activities—hours worked in agricultural production (HA),nonagricultural production (HN ) and wage or mar-ket employment (HM )8—are the outcome of thisdecision process. In theory, the allocation of timeamong these activities should equalize the returnsto labor in them. Optimal leisure consumption isthe difference between the household’s time endow-ment and total time worked in all the activities.

Where markets are imperfect, where transactioncosts or barriers to mobility are present, or wherehouseholds have preferences for certain types ofwork, the returns to labor in different activities may

no longer be equal. Moreover, in such situations thereturns to labor in household production may notbe observable. These returns or shadow wages willbe endogenous and a function of both production-and consumption-side variables. Households thusmay simultaneously face three different prices oflabor: a shadow wage for labor in agricultural pro-duction (WA), a shadow wage for labor in nonagri-cultural production (WN ), and an observed wagefor market employment (WM).

Factor markets in China are likely to be imper-fect, so we adopt an empirical strategy that allowsfor this possibility. Specifically, we use the approachof Jacoby (1993), who notes that at the householdoptimum, the household’s shadow price of laborwould equal the value of the marginal product of labor in household production. Labor supply,then, can be modeled as a function of the shadowwages. We thus specify labor supply as a function ofshadow wages, as well as of other relevant variables.Also, because we are interested in labor supply tothree activities, we have three different labor supplyfunctions. These take the form

Hi = f i (WA, WN , WM , I , X),

i = A, N, M(14.1)

where WA, WN , WM are the shadow wage in agri-cultural production, the shadow wage in nonagri-cultural production, and the wage rate in marketemployment, respectively; I is household nonlaborincome; and X is a vector of household characteris-tics. Total labor supply is the sum of these threelabor supplies.

Empirical Specification

The specification of our labor supply functionsbasically follows that of Jacoby and others,9 exceptthat we include multiple wages. The functions forlabor supply to household agricultural production,household nonagricultural production, and wageor market employment are, respectively,

log HA = α0 + αA ln WA + αN ln WN

+ αM ln WM + αI I

+ αX X + εA

(14.2)

log HN = β0 + βA ln WA + βN ln WN

+ βM ln WM + βI I

+ βX X + εN

(14.3)


log HM = γ0 + γA ln WA

+ γN ln WN + γM ln WM

+ γI I + γX X + εM

(14.4)

Overall labor supply (H) is the sum of laborsupply into three sectors. Alternatively, it can beestimated directly as

log H = δ0 + δA ln WA + δN ln WN

+ δM ln WM + δI I

+ δX X + ε

(14.5)

The signs of the effects of the three wage rates ontotal labor supply depend on the relative magni-tude of substitution and income effects.

The estimation of equations 14.2–14.5 requiresknowledge of wage or shadow wage rates in allactivities. For those engaged in agricultural ornonagricultural self-employment, shadow wagescan be calculated using estimated parameters fromhousehold production functions (Jacoby 1993). Wetherefore begin with estimates of production func-tions in the next section. We then present our esti-mates of the wage and shadow wage functions. Weuse these estimates to project (shadow) wages forhouseholds that do not participate in one or moreactivities. These projected (shadow) wages areemployed in our analysis of labor supply.

Estimation of Production Functions

We use the standard Cobb-Douglas functionalform to estimate the production functions for agri-cultural and nonagricultural activities. For each ofthese activities the production function takes theform

ln Q =n∑

j=1

αj ln Mj +m∑

k=1

γk Zk + ε (14.6)

where Q is the total value of output produced bythe household; M is a vector of production inputs,which includes family labor hours, the value ofvariable costs (including materials and hiredlabor),10 the value of fixed capital, and land used inagricultural production; and Z is a vector of othercontrol variables, including average education ofhousehold workers, average age of householdworkers, and provincial dummy variables.

Each production function is estimated using threeestimation methods—OLS, IV, and selectivity-bias

corrected estimation. In the IV or 2SLS estimations,labor and variable cost inputs are treated as endoge-nous. Instruments used in both the agriculturaland nonagricultural production functions areexogenous variables for the household and the com-munity, including household composition variables(number of working-age adults, children, and eld-erly) and local market prices of vegetables, pork,chicken, and gasoline. In the nonagricultural pro-duction function we also include as instruments thelocal market prices of honey-combed coal bri-quette, coal lumps, coal powder, and liquefied natu-ral gas, and the market wage for unskilled labor. Theselectivity-bias corrected estimation treats whetherthe household engages in the activity as an endoge-nous decision. The identification variable is thenumber of dependents in the household, and we usethe Heckman method (Heckman 1979). As noted intables 14.5 and 14.6, the IV estimation does notpass the Hausman joint exogeneity test for eitheragricultural or nonagricultural production. Selec-tivity terms in the selectivity-bias correction estima-tions (lambda statistics reported for Model III) arealso statistically insignificant for both equations.

The first column of table 14.5 reports OLS esti-mates of the agricultural production function. Theresults indicate that variable inputs are the domi-nant contributor to agricultural output, with anelasticity of 0.46. Land is the next most importantcontributor to agricultural production, with anelasticity of 0.32, and labor hours have an elasticityof 0.19. The coefficient for agricultural equipmentis also significant. These estimates are similar inmagnitude to those in other studies (e.g., Yang1997; Li and Zhang 1998). The coefficients of theseinputs add up to 0.999, which indicates that agri-culture displays constant returns to scale. Educa-tion is insignificant. A comparison of the first threecolumns of the table reveals that the instrumentalvariable and selectivity-bias corrected estimates arevery similar to the OLS estimates.

The first column of table 14.6 reports OLS esti-mates of the nonagricultural production function.The elasticity of labor inputs is 0.62, more thantriple the labor coefficient for agriculture. Outputvalue is less responsive to variable costs in nonagri-cultural production than in agricultural produc-tion. The coefficient of equipment is the same as inagriculture, 0.033. Education is statistically insignif-icant. Interestingly, the nonagricultural production


function displays decreasing returns to scale. Thisfinding suggests that rural households mayencounter difficulties in expanding the scale ofhousehold nonagricultural production (or perhapshigher-income households underreport suchincome). Like in table 14.5, a comparison of the firstthree columns reveals that the selectivity-bias cor-rected estimates are similar to the OLS estimates. Inthe IV regression, the coefficients on labor and vari-able costs differ from those in the other specifica-tions, but the Hausman test indicates that the overallthe results are not significantly different from thoseof OLS.

Although the Hausman tests do not providestrong support for preferring the IV results to thoseof the OLS, we nevertheless use the results from

the IV specifications to calculate shadow wages. Weprefer the IV results because, in theory, output,labor hours, and variable costs are jointly andendogenously determined, and so in principle IVmethods are needed to obtain unbiased results.

Estimation of Wages and Shadow Wages

We are interested in learning how rural householdsrespond to differentials between agricultural andnonagricultural wages. Although we observe thehourly wage rates received by employed workers,the same is not true for workers engaged in house-hold agricultural or nonagricultural production.We follow the approach of Jacoby (1993) and deriveshadow wages from the production function


TABLE 14.5 Agricultural Production Function (dependent variable: log of output value)

Model III: Corrected for Model I: OLS Model II: IV Sample Selection Bias

Standard Standard StandardCoefficient Error Coefficient Error Coefficient Error

Log of labor hoursa 0.192*** 0.035 0.260 0.197 0.179*** 0.035Log of variable costsa 0.458*** 0.031 0.422** 0.177 0.449*** 0.031Log of equipment 0.035*** 0.009 0.034*** 0.011 0.034*** 0.009Log of land 0.317*** 0.052 0.322*** 0.107 0.314*** 0.052Mean schooling of working- 0.010 0.012 0.014 0.016 0.009 0.011

age adultsMean age of working-age adults �0.0001 0.004 0.0002 0.004 �0.0001 0.004Dummy: telephone present in 0.095 0.062 0.089 0.064 0.104* 0.062

local areasHeilongjiang 0.092 0.116 0.135 0.176 0.088 0.115Jiangsu �0.069 0.097 �0.026 0.154 �0.081 0.097Shandong �0.163 0.139 �0.116 0.198 �0.167 0.139Henan �0.623*** 0.102 �0.617*** 0.131 �0.614*** 0.104Hubei 0.023 0.090 0.005 0.108 0.046 0.090Hunan 0.143 0.156 0.195 0.215 0.170 0.157Guangxi �0.312*** 0.096 �0.310*** 0.099 �0.265*** 0.096Constant 2.647*** 0.350 2.323** 1.146 2.811*** 0.353

Lambda n.a. n.a. �0.091 0.096Adjusted R-square 0.455 0.453 n.a.Log-likelihood n.a. n.a. �1,393.699Number of observations 1,066 1,066 1,095

n.a. Not applicable.*p < .01, **p < .05, ***p < .10.aIndicates endogenous variables. Instruments used are family composition variables (number of working-age adults,number of children, and number of elderly) and market prices for vegetables, pork, chicken, and gasoline. TheHausman statistic for the joint endogeneity test is 0.13, and the P-value is 0.937. Here and in later tables Guizhou is theomitted regional dummy variable.Source: Authors’ estimates using CHNS data.

estimates. Specifically, we calculate the shadowwage rates, or marginal products, of family laborhours from the IV estimates of the Cobb-Douglasproduction functions in tables 14.5 and 14.6 asfollows:

Wj = αj Qj

L j,

j = agriculture, nonagriculture

(14.7)

where Q is the value of output and L is family laborhours.

Equation 14.7 allows us to calculate averageshadow wages for households that engage in agri-cultural or nonagricultural production. Theseshadow wages are derived from household produc-tion functions and thus are at the household, ratherthan individual, level.

Households compare expected wage differen-tials among sectors when deciding their sectoral

labor allocation. After a selection is made, onlywages in the actually selected sectors are observed.To estimate the expected wage differential, we needthe counterfactual wage rates for households notparticipating in a sector. To obtain these rates, weregress the shadow wages computed as just out-lined against a range of household characteristicsthat potentially affect the productivity of workers.These include all exogenous variables in the pro-duction functions, the variables used as instru-ments for labor, and variable costs. We then use theresults from these regressions to predict shadowwages for all households.

Note that the projected shadow wages are usedlater in our labor supply functions, and so some ofthe explanatory variables here will serve as instru-ment variables in our labor supply regressions.Therefore, because all second-stage variablesshould be included in the first-stage regressions, we


TABLE 14.6 Nonagricultural Production Function (dependent variable: log of output value)

Model III: Corrected for Model I: OLS Model II: IV Sample Selection Bias

Standard Standard StandardCoefficient Error Coefficient Error Coefficient Error

Log of labor hoursa 0.618*** 0.039 0.486*** 0.183 0.656*** 0.048Log of variable costsa 0.167*** 0.012 0.253*** 0.069 0.154*** 0.015Log of equipment 0.033*** 0.010 0.024* 0.014 �0.030 0.032Mean schooling of working-age adults 0.012 0.016 0.007 0.019 �0.001 0.027Mean age of working-age adults �0.005 0.006 �0.007 0.006 0.012 0.008Dummy: telephone present in local areas 0.075 0.098 �0.041 0.149 0.100 0.114Heilongjiang 0.031 0.206 0.125 0.229 0.376 0.350Jiangsu 0.159 0.144 0.201 0.155 0.123 0.205Shandong �0.271 0.188 �0.354* 0.209 �0.181 0.265Henan �0.008 0.125 0.016 0.133 �0.126 0.163Hubei �0.028 0.143 �0.092 0.159 �0.143 0.214Hunan 0.066 0.135 0.059 0.142 0.223 0.241Guangxi �0.137 0.114 �0.128 0.120 �0.432*** 0.169Constant 2.968*** 0.362 3.670*** 1.136 2.985*** 0.661

Lambda n.a. n.a. �0.450 0.239Adjusted R-square 0.603 0.561 n.a.Log-likelihood n.a. n.a. �946.014Number of observations 507 507 1,611

n.a. Not applicable. ***p < .01, **p < .05, *p < .10.aIndicates endogenous variables. Instruments used are family composition variables (number of working-age adults,number of children, and number of elderly); market prices of honey-combed coal briquette, coal lumps, coal powder,and liquefied natural gas; and local market wage rate for the unskilled. The Hausman statistic for the joint endogeneitytest is 1.77, and the P-value is 0.413.Source: Authors’ estimates using CHNS data.

also include all exogenous variables from the laborsupply equations in the shadow wage equations.These variables include nonlabor income, numberof dependents, education and age of workers,health status of household members, and a fewcommunity-level variables that measure the preva-lence of nonfarm activities.

Table 14.7 reports estimates of the shadow wageequations for household agricultural productionand household nonagricultural production. We usethe OLS regression because, after including all the

variables mentioned above, it is difficult to findadditional identification variables for use in an IVor 2SLS regression. The agricultural shadow wagefunction has good explanatory power, and theresults are reasonable. As expected, equipment andland are significant in enhancing the marginal pro-ductivity of agricultural labor. Average schooling ofworking adults is also positive and significant.

The explanatory power of the nonagriculturalshadow wage function is low, perhaps reflecting thesmaller number of observations and heterogeneity


TABLE 14.7 Shadow Wage Equations for Household Agricultural and Nonagricultural Production (dependent variable: log of shadow wage)

Household Agricultural Household Nonagricultural Production (OLS) Production (OLS)

Coefficient Standard Error Coefficient Standard Error

Agricultural equipment (1,000 yuan) 0.020*** 0.006 0.037 0.025Nonagricultural equipment (1,000 yuan) 0.014*** 0.005 0.018*** 0.005Land 0.012*** 0.003 �0.016 0.012Number of family working-age adults �0.017 0.033 0.056 0.058Number of dependents 0.045 0.036 0.104 0.064Ratio of male working-age adults to �0.281 0.207 0.656 0.399

family working-age adults Family nonlabor income (1,000 yuan) �0.002 0.019 �0.052 0.033Mean schooling of working-age adults 0.036** 0.015 0.008 0.031Mean age of working-age adults 0.004 0.005 0.013 0.011Dummy: all family members are healthy 0.220*** 0.071 0.191 0.134Dummy: telephone present in local �0.008 0.084 0.330** 0.148

areasMarket price of kerosene �1.093*** 0.229 �0.732* 0.402Market price of rice most commonly �0.049 0.030 0.052 0.132

usedDummy: TVEs present in local areas 0.138* 0.076 �0.084 0.136Number of self-employed household 0.003 0.002 �0.003 0.003

enterprises in local areasHeilongjiang 0.604*** 0.153 �0.058 0.443Jiangsu 0.063 0.153 �0.266 0.296Shandong �0.057 0.236 �0.347 0.406Henan �0.653*** 0.132 �0.186 0.217Hubei 0.0002 0.114 �0.120 0.247Hunan 0.949*** 0.275 1.035*** 0.360Guangxi �0.458*** 0.135 �0.265 0.211Constant 0.169 0.509 0.206 0.990

Adjusted R-square 0.274 0.076Number of observations 998 274

***p < .01, **p < .05, *p < .10.Source: Authors’ estimates using CHNS data.

of labor inputs in nonagricultural activities. Herethe most significant variable is nonagriculturalequipment.

Not all households or individuals participate inwage employment. We therefore also need to pre-dict market wages for nonparticipants. In addition,market wages are reported at the individual level,but to analyze household labor supply we need ameasure of the expected wage at the householdlevel. To address these issues, we first estimate awage equation for all workers participating in wageemployment. We then use these estimates to predictwage rates for all workers in the sample, after whichwe aggregate for each household the predicted indi-vidual wage rates of family members to obtain ahousehold-level wage.

Table 14.8 gives the estimates of wage functionsfor the employed workers. Our specification fol-lows the standard approach used in the labor litera-

ture, but adds some variables relevant to the Chinacase. The dependent variable is the log of hourlywages. Independent variables include variables cap-turing individual human capital or labor quality(education, age, age squared, and health status),individual characteristics (marital status and sex),and regional dummy variables. We allow for possi-ble sample selection bias by applying the Heckmanmethod, but the results show that selectivity is notimportant.11

The employed wage function indicates that edu-cation receives a rate of return of 3 percent. Thiseffect is statistically significant at the 1 percent level.The age-earning profile has a concave shape, withthe maximum wage reached at age 48. Marriage hasa positive but statistically insignificant effect onwage. Once given marital status, female workers earn20.3 percent less than their male counterparts. Thosewho have urban hukou status earn 11.0 percent more


TABLE 14.8 Employed Wage Equation (dependent variable: log of hourly wages)

Model II: Selectivity-Bias Model I: OLS Correcteda

Independent Variable Coefficient Standard Error Coefficient Standard Error

Female �0.203*** 0.029 �0.203*** 0.033Marriage status 0.012 0.046 0.012 0.058Education level (years) 0.030*** 0.005 0.030*** 0.009Age 0.029*** 0.011 0.029** 0.012Age squared �0.0003** 0.0001 �0.0003** 0.0001Health status 0.066* 0.040 0.066* 0.039Urban hukou 0.110*** 0.029 0.110*** 0.029Heilongjiang �0.084 0.065 �0.084 0.064Jiangsu 0.181*** 0.058 0.181*** 0.069Shandong �0.057 0.061 �0.057 0.070Henan �0.236*** 0.072 �0.236*** 0.073Hubei �0.043 0.064 �0.043 0.063Hunan 0.229*** 0.063 0.229*** 0.065Guangxi �0.002 0.063 �0.002 0.062Constant �0.018 0.189 �0.018 n.a.

Lambda n.a. 0.00005 0.084Adjusted R-square 0.116 n.a.Log-likelihood n.a. �5,493.72Number of observations 1,910 8,123

n.a. Not applicable.***p < .01, **p < .05, *p < .10.aThe identification variable is the number of dependents.Source: Authors’ estimates using CHNS data.

than those who do not. Using the coefficients fromthe selectivity-bias corrected regression, we predictwages for all workers. To obtain a household-levelwage, we calculate the weighted average of pre-dicted wages for the workers in each household,where the weights are based on each worker’s hoursof work.

Table 14.9 shows the mean predicted shadowwages and market wages of all households and, forcomparison, the computed shadow wages andobserved market wages of participating house-holds. Not surprisingly, the mean wages of partici-pating households are higher than those predictedfor all households. This difference reflects that par-ticipation is more likely at higher wage rates. Thereturns to labor are lowest for agricultural produc-tion, followed by nonagricultural production andwage employment. These numbers are broadlyconsistent with the estimated returns to work timefrom the income generation regressions reportedearlier.

Estimation of Labor Supply Functions

Our household labor supply functions follow theform of equations 14.2–14.5. For the dependentvariable, we use hours per adult rather than totalhours to avoid possible correlation between house-hold labor supply and number of working-ageadults in the household. All specifications are esti-mated using IV (2SLS), with the wage and shadowwage equations reported earlier serving as the first-stage regressions. A large number of householdssupply zero hours to nonagricultural productionand wage employment, and so for these functions

we also run Tobit regressions and report the mar-ginal effects for all households and the standarderrors of the marginal effects.

Our results appear in table 14.10. The firstcolumns give estimated coefficients for total laborsupply per adult. These results indicate that totallabor supply is not sensitive to the marginal returnsto nonfarm labor. The returns to agricultural laborhave a coefficient that is negative, significant (at the10 percent level), and small. This coefficient reflectsthe net impact of the agricultural shadow wage onthe different components of the labor supply. Thisimpact is positive for the agricultural labor supplybut negative for the market labor supply (this is dis-cussed later in this chapter). The effect of nonlaborincome is negative and significant at the 10 percentlevel, indicating that leisure is a normal good. Inter-estingly, total labor supply per adult is higher themore dependents (elderly and children) in a house-hold. This finding suggests that the need to earnincome to support dependents outweighs the needto spend non-earning time caring for dependents.The effect of dependents also indicates that depend-ents cause households to reallocate labor away fromhigher-paying wage jobs to lower-income house-hold production activities.

Although the nonagricultural wage variables donot have a significant effect on total labor supply,they affect the allocation of labor among differentactivities. The shadow wage for work in householdagricultural production is significant for marketemployment, where it has a fairly large, negativecoefficient. A 1 percent increase in the shadow agri-cultural wage decreases labor supply to marketemployment by 2.6 percent.


TABLE 14.9 Summary Statistics for Wage Rates and Shadow Wages (yuan/hour)

Number of Standard Variable Observations Mean Deviation

Shadow wage for agricultural production (WA) 1,066 0.742 1.322Shadow wage for agricultural production, predicted (WA) 1,500 0.517 2.197Shadow wage for nonagricultural production (WN) 507 2.383 3.243Shadow wage for nonagricultural production, predicted (WN) 1,500 1.760 6.871Employed wage rate (individual level) (WM) 1,910 4.013 37.789Employed wage rate, predicted (individual level) (WMi) 8,270 2.239 0.554Employed wage rate, predicted (household level) (WM) 2,755 2.271 0.526



TAB

LE 1

4.1

0H

ou

seh

old

Lab

or

Sup

ply

Eq

uat

ion

s (d

epen

den

t va

riab

le: lo

g o

f w

ork

ing

ho

urs

per

ad

ult

)

Log

of A

gric

ultu

ral

Log

of N

onag

ricu

ltur

al

Log

of T

otal

Hou

rsH

ours

Self-

Empl

oym

ent

Hou

rsLo

g of

Wag

e Em

ploy

men

t H

ours

IVIV

IVTo

bit

Unc

ondi

tion

alIV

Tobi

t U

ncon

diti

onal

Stan

dar

d

Stan

dar

d

Stan

dar

dM

arg

inal

St

and

ard

Stan

dar

dM

arg

inal

Stan

dar

d

Co

effi

cien

tEr

ror

Co

effi

cien

tEr

ror

Co

effi

cien

tEr

ror

Effe

ctEr

ror

Co

effi

cien

tEr

ror

Effe

ctEr

ror

lnW

A�

0.14

4*0.

084

0.02

50.

098

0.14

10.

304

0.00

80.

376

�1.

137*

**0.

335

�2.

137*

**0.

485

lnW

N0.

051

0.05

4�

0.12

3**

0.06

20.

332*

0.19

40.

389

0.24

70.

617*

**0.

214

1.18

3***

0.29

6

lnW

M0.

067

0.11

3�

0.12

00.

129

0.57

70.

400

0.79

1*0.

476

0.55

50.

448

0.48

50.

414

Log

(non

labo

r in

com

e)�

0.00

9*0.

005

�0.

032*

**0.

006

�0.

052*

**0.

019

�0.

047*

**0.

018

0.11

8***

0.02

20.

111*

**0.

021

Num

ber

of d

epen

dent

s0.

117*

**0.

020

0.14

0***

0.02

30.

199*

**0.

070

0.14

1**

0.06

1�

0.31

8***

0.07

7�

0.31

0***

0.07

7

Mea

n sc

hool

ing

of w

orki

ng-a

ge a

dults

0.00

90.

010

�0.

049*

**0.

012

0.08

8**

0.03

60.

077*

*0.

035

0.23

6***

0.04

00.

249*

**0.

040

Ratio

of m

ale

wor

king

-age

adu

lts t

o �

0.12

80.

125

0.26

1*0.

144

�0.

292

0.44

5�

0.31

60.

455

�1.

598*

**0.

494

�2.

078*

**0.

527

fam

ily w

orki

ng-a

ge a

dults

Mea

n ag

e of

fam

ily w

orki

ng-a

ge a

dults

0.00

5*0.

003

0.00

7**

0.00

3�

0.01

30.

010

�0.

018*

0.01

0�

0.01

80.

011

�0.

026*

*0.

011

Dum

my:

goo

d he

alth

for

fam

ily�

0.01

20.

045

�0.

101*

0.05

20.

145

0.16

00.

075

0.14

50.

180

0.17

70.

249

0.17

1

Dum

my:

TVE

s p

rese

nt in

loca

l are

as0.

005

0.04

4�

0.20

6***

0.05

00.

193

0.15

40.

226

0.15

11.

054*

**0.

171

1.27

9***

0.20

4

Num

ber

of s

elf-

emp

loye

d ho

useh

old

�0.

001

0.00

1�

0.00

6***

0.00

10.

001

0.00

40.

0003

0.00

30.

010*

*0.

004

0.01

1***

0.00

4en

terp

rises

in lo

cal a

reas

Mar

ket

pric

e of

ker

osen

e0.

030

0.15

10.

021

0.17

40.

018

0.53

7�

0.27

20.

501

�1.

376*

*0.

595

�2.

190*

**0.

627

Mar

ket

pric

e of

com

mon

ric

e�

0.03

1**

0.01

5�

0.00

10.

018

�0.

095*

0.05

6�

0.13

9**

0.07

0�

0.15

8***

0.06

1�

0.20

8***

0.06

3

Hei

long

jiang

�0.

054

0.11

1�

0.15

00.

129

�1.

204*

**0.

400

�0.

944*

**0.

176

0.65

10.

440

2.05

1**

0.90

3

Jiang

su0.

137

0.08

5�

0.25

8***

0.09

8�

0.38

70.

302

�0.

429*

*0.

199

1.68

4***

0.33

51.

916*

**0.

529

Shan

dong

0.22

0*0.

113

�0.

443*

**0.

130

�0.

761*

0.38

9�

0.59

6***

0.17

61.

770*

**0.

437

1.76

7**

0.69

6

Hen

an�

0.17

2**

0.08

5�

0.31

2***

0.09

80.

475

0.30

40.

304

0.36

9�

0.65

3*0.

336

�0.

907*

**0.

247

Hub

ei0.

216*

**0.

064

0.34

4***

0.07

4�

0.60

7*0.

228

�0.

529*

**0.

145

�0.

111

0.25

2�

0.00

10.

253

Hun

an�

0.02

70.

161

0.19

10.

183

�0.

385

0.55

4�

0.36

20.

322

�0.

220

0.62

10.

521

0.79

2

Gua

ngxi

0.32

9***

0.08

60.

414*

**0.

099

1.23

4***

0.30

70.

893*

*0.

410

�0.

542

0.34

1�

0.62

1**

0.24

4

Con

stan

t6.

525*

**0.

268

6.88

1***

0.30

70.

931

0.94

2n.

a.n.

a.1.

466

1.04

8n.

a.n.

a.

Adj

uste

d R-

squa

re0.

098

0.19

80.

105

n.a.

0.17

5n.

a.

Pseu

do R

-sq

uare

n.a.

n.a.

n.a.

0.05

8n.

a.0.

073

Obs

erva

tions

1,36

01,

404

1,43

01,

430

1,42

11,

421

n.a.

Not

ap

plic

able

.

***p

< .0

1,**

p<

.05,

*p<

.10.

Sour

ce:A

utho

rs’ e

stim

ates

usi

ng C

HN

S da

ta.

The effect of the shadow wage for work inhousehold nonagricultural production is negativeand significant for agricultural labor and positiveand mostly significant for nonagricultural andmarket labor supply. In other words, householdswith higher returns to labor in nonagriculturalsidelines work fewer hours in agriculture and morehours in both nonagricultural and wage employ-ment. Here the elasticities are small for agricultureand nonagriculture, and close to one for wageemployment. The signs on the coefficient for themarket wage are the same as those for the shadownonagricultural wage, but this variable is signifi-cant in only one equation.

Overall, the coefficients on the wage variablesreveal complementarity between employment inhousehold nonagricultural production and off-farmwage jobs, but substitution between agriculture andboth types of nonagricultural employment. Further-more, the estimated coefficients on wage variablesare larger in magnitude and more significant in theoff-farm wage labor supply equations than in thehousehold agricultural or nonagricultural labor sup-ply equations. In other words, labor supply to house-hold production activities does not respond much towage levels; the supply of labor to off-farm wageemployment does.

From the wage elasticities we can calculatethe elasticity of labor transfers in response torelative wage differentials among sectors.12 Sub-tracting equation 14.2 from 14.3 gives the followingexpression:

log(HN/HA)

= (β0 − α0) + (βA − αA) ln WA

+ (βN − αN) ln WN

+ (βM − αM) ln WM + · · ·= (β0 − α0)

+ (βA − αA + βN − αN) ln WA

+ (βN − αN) ln(WN/WA)

+ (βM − αM) ln WM + · · ·

(14.8)

Similarly, subtracting (14.2) from (14.4) gives

log(HM/HA)

= (γ0 − α0)

+ (γA − αA + γM − αM) ln WA

+ (βN − αN) ln WN

+ (βM − αM) ln(WM/WA) + · · ·

(14.9)

Using these expressions and the results intable 14.10, we calculate the marginal response oflog(HN/HA) to log(WN/WA) as 0.512 (5 0.389 10.123), and the marginal response of log(HM/HA)to log(WM/WA) as 0.605 (5 0.485 1 0.120). Thesenumbers imply that a 1 percent increase in the ratioof either nonagricultural wage to the agriculturalwage raises the ratio of that nonagricultural workto agricultural work by about 0.5–0.6 percent.

Note that these numbers assume that the changein the wage ratio stems from a rise in the nonagri-cultural shadow or market wage, with the agricul-tural shadow wage remaining constant. Due to thepresence of cross-wage effects, shifts caused bychanges in the agricultural shadow wage (holdingthe nonagricultural and market wages constant)would be different. Here the relevant calculationsare as follows: the marginal response oflog(HA/HN) to log(WA/WN) is 0.017 (� 0.025 �

0.008), and the marginal response of log(HA/HM)to log(WA/WM) is 2.165 (� 0.028 � 2.137). Inother words, a 1 percent increase in the ratio of theagricultural shadow wage to the nonagriculturalshadow wage would increase the ratio of labor inagricultural to nonagricultural production by0.02 percent. For an increase in the ratio of theagricultural shadow wage to the market wage, theresponse would be much larger, 2.2 percent.

These numbers indicate that, for the most part,the effects of relative wages on sectoral labor alloca-tion are not overly large or significant (relative tothe standard errors). The one exception is the casein which a change in the agricultural shadow wagecauses a change in the ratio of the agriculturalshadow wage to the market wage. In this case, shiftsin labor allocation between agriculture and off-farmemployment may be relatively large.

These findings suggest that labor movement outof agriculture into other sectors is not driven by the“pull” of rising nonagricultural wages, but would bedriven by the “push” of lower agricultural wages.Why would these pull-and-push effects differ? Oneexplanation is that for rural families income fromagriculture provides the base income that ensuresthat the household achieves some acceptable orminimum standard of living. If agricultural incomefalls short—for example, in the event of drought,which would reduce overall agricultural incomeas well as the returns to labor in agriculture—households may feel compelled to seek income


from other sources. They do not typically seek thatadditional income from nonagricultural self-employment, because such activities usually requirestart-up time and some initial investment. Casualwage jobs entail lower initial costs and generate cashincome more quickly. Field interviews with ruralhouseholds that have experienced shortfalls in agri-cultural income indeed indicate that these house-holds often send family members out to find casualwage jobs, sometimes seasonal, in construction,services, and other unskilled occupations.

Several nonwage variables have significant andinteresting effects on labor allocation. Nonlaborincome is significant and negative for labor supplyto both agricultural and nonagricultural householdproduction, but is significant and positive for wageemployment. This result could indicate that higherincome increases access to wage employment by,for example, providing the resources needed tocover initial and search costs associated withobtaining wage jobs. It could also indicate thathouseholds gain some status or utility from off-farm work. The magnitudes of these income effects,however, are fairly small. Not surprisingly, thenumber of dependents significantly increases laborsupply to both types of household production, butreduces labor supply to off-farm jobs. One moredependent in the household increases workinghours that an adult devotes to agricultural produc-tion by 14 percent, increases hours devoted tononagricultural production by 14 percent, andreduces working hours in wage employment by31 percent. Age has a marginally significant andpositive effect on agricultural labor supply; fornonagricultural production and market employ-ment, its effects are negative, and for the Tobitregressions, its effects are significant. This result isconsistent with the observation that in rural Chinayounger adults are more mobile than older adults.

Interestingly, a few household characteristicsthat do not significantly influence total labor supplyhave significant effects on the allocation of laboramong activities. Education reduces hours workedin agricultural production by 5 percent, increaseshours in nonagricultural production by 8 per-cent, and increases hours in wage employment by asubstantial 25 percent. The proportion of malesamong working-age adults increases labor supplyto agriculture, but reduces labor supply to wageemployment.

Community-level variables also affect the com-position of the labor supply. We use a dummyvariable for the availability of TVEs and the num-ber of self-employed household businesses as indi-cators of the prevalence of nonfarm activities inlocal communities. As expected, both variables aresignificant in reducing labor allocation to agricul-tural production and raising participation innonagricultural production and wage employ-ment. Also, the coefficients on the market pricesfor fuel and rice are in some cases negative andsignificant.

In general, our estimates indicate that labor sup-ply to agriculture and labor supply to wage employ-ment have opposite responses to most explanatoryvariables. In other words, factors that increase wageemployment would tend to decrease agriculturalemployment, and vice versa. This conclusion isconsistent with aggregate statistics that show suchopposing trends in aggregate employment in thetwo sectors. Labor supply to nonagricultural house-hold businesses, however, moves with agriculturefor some variables and with wage employment forothers. Thus, depending on which determinants oflabor supply are driving aggregate trends, we mightexpect to observe concurrent growth in agriculturaland nonagricultural household employment, orgrowth in one and decline in the other.

Conclusions

The impact of trade liberalization in China dependsnot only on the aggregate sectoral and price shiftsthat follow liberalization, but also on the microeco-nomic response of households and individuals tothose shifts. In this chapter we have analyzed themicroeconomics of household earnings and employ-ment in rural China. Using household survey datawith broad regional coverage, we have examinedhousehold incomes, wage determination, and laborsupply in total and among sectors.

Our analysis of income generation providesinformation about earnings from labor. We findthat the returns to nonagricultural labor hours arehigher than for agricultural labor hours, a resultthat is consistent with other studies in the litera-ture. We then decompose the income gap betweenpoorer and richer groups. Key contributors to theincome gap are the lower returns received by poorhouseholds from education, agricultural labor, and


wage employment. Also important is the fact thatpoorer households work fewer hours in higher-paid wage employment. These results suggest thatanalysis of the impact of trade liberalization shouldconsider not only the fact that the level and compo-sition of employment differ among income groups,but also that the returns to employment in eachoccupation can differ among groups.

Our analysis of household labor supply providessome insights into how changes in wages afterChina’s accession to the WTO might affect patternsof employment and thus earnings. Total labor sup-ply is not sensitive to changes in wages, which sug-gests that any changes in wages resulting fromWTO accession would not have much impacton overall employment. We also find that, for themost part, the allocation of labor among self-employment in agriculture, self-employment innonagriculture, and off-farm wage jobs is notoverly sensitive to changes in wages. In particular,labor supply to both agricultural and nonagricul-tural household production is inelastic with respectto most wages. Put differently, changes in wages donot cause large changes in the amount of laborsupplied to household production.

The off-farm labor supply is more responsive,at least to the agricultural shadow wage. A loweragricultural shadow wage is associated with sub-stantially increased employment in off-farm jobs.Thus if WTO accession raises wages for off-farmjobs, it would not “pull” labor into wage jobs;however, if WTO accession reduces the returns toagricultural labor, it might “push” labor into wagejobs.

Although in general labor supply is not overlysensitive to wages, it is sensitive to other variables.Household characteristics such as education,household composition, and regional location havesignificant effects on the level and sectoral composi-tion of labor supply. Higher education is associatedwith less work time in agriculture and more worktime in nonagricultural types of work. Moredependents are associated with more work timeoverall, as well as with more work time in householdproduction and less in off-farm wage employment.These findings suggest that labor mobility amongsectors is significantly influenced by variables suchas education, health, and demographic structure.Such variables are not directly affected by WTOaccession, at least in the short term.

Notes

1. Official Chinese estimates show the number of rural poordeclining from 65 million in 1995 to 42 million in 1998 (WorldBank 2001), and show the number of urban poor ranging from10 to 15 million in the mid-1990s (UNDP 1998; NBS 1998).These numbers may not be entirely comparable, but they indi-cate that roughly 75–85 percent of the poor are rural.

2. Official Chinese statistics distinguish between private andindividual enterprises. Private enterprises are larger and rely onhired employees as the main source of labor. Individual enter-prises are smaller and, although they may hire employees, relylargely on family workers for labor. Both types of enterprises areprivately owned.

3. Income data in the text are deflated using the rural con-sumer price index and are taken from the National Bureau ofStatistics (various years) and the National Bureau of StatisticsRural Social and Economic Survey Team (2000).

4. The CHNS survey covers the provinces of Guangxi,Guizhou, Heilongjiang, Henan, Hubei, Hunan, Jiangsu,Liaoning, and Shandong. The 1997 survey, which we use for ouranalysis, does not include Liaoning and so covers only eight ofthe nine provinces.

5. Detailed information about the CHNS is available atwww.cpc.unc.edu/china/home.html.

6. Note that for agricultural and nonagricultural householdproduction, these estimated returns are substantially lower thanaverage earnings per hour worked given earlier (1.2 and2.2 yuan, respectively). This finding could reflect the fact that inhousehold production the average and marginal returns to labordiffer and that the income generation function controls for non-labor inputs. The estimated return to wage labor hours is alsosomewhat lower than average earnings (2.8 yuan).

7. Note that the difference in time worked per adult is similarto that for total hours, because the number of working-ageadults in the two groups is similar—on average, the poor house-holds had 2.74 adults and the rich households 2.71 adults.

8. Note that wage employment includes wage labor in agri-culture, but in our sample hired farm labor is a small proportionof total wage employment.

9. Other studies taking this approach include Skoufias(1994) and Abdulai and Regmi (2000).

10. The CHNS data do not distinguish hired labor fromother variable inputs.

11. Estimated coefficients for Models I and II are nearlyidentical despite the significance of the number of dependentsin sector choice. Note that lambda is also highly insignificant.This indicates that there is no selectivity bias.

12. We thank Martin Ravallion for suggesting this approach.

References


Abdulai, Awudu, and Punya Prasad Regmi. 2000. “EstimatingLabor Supply of Farm Households under Nonseparability:Empirical Evidence from Nepal.” Agricultural Economics 22:309–20.

Bowlus, Audra, and Terry Sicular. 2003. “Moving towardMarkets? Labor Allocation in Rural China.” Journal ofDevelopment Economics 71 (2): 561–83.

Chen Shaohua and Wang Yan. 2001. “China’s Growth andPoverty Reduction: Recent Trends between 1990 and 1999.”World Bank, July. Processed.


Dollar, David, and Aart Kraay. 2001. “Trade, Growth andPoverty.” Policy Research Working Paper 2615. World Bank,Washington, D.C.

Hare, Denise. 1994. “Rural Nonagricultural Activities and TheirImpact on the Distribution of Income: Evidence from FarmHouseholds in Southern China.” China Economic Review4 (1): 59–82.

———. 1999a. “ ‘Push’ versus ‘Pull’ Factors in Migration Out-flows and Returns: Determinants of Migration Status andSpell Duration among China’s Rural Population.” Journal ofDevelopment Studies 35 (3): 45–72.

———. 1999b. “Women’s Economic Status in Rural China:Household Contributions to Male-Female Disparities in theWage Labor Market.” World Development 27 (6): 1011–29.

Heckman, James. 1979. “Sample Selection Bias as a SpecificationError.” Econometrica 47 (1): 153–62.

Jacoby, Hanan. 1993. “Shadow Wages and Peasant FamilyLabour Supply: An Econometric Application to the PeruvianSierra.” Review of Economic Studies 60 (4): 903–21.

Knight, John, and Lina Song. 1997. “Chinese Peasant Choices:Migration, Rural Industry or Farming.” Oxford AppliedEconomics Discussion Paper No. 188. October.

———. 1999. The Rural–Urban Divide: Economic Disparitiesand Interactions in China. Studies on Contemporary China.Oxford and New York: Oxford University Press.

Li Tianyou and Junsen Zhang. 1998. “Returns to Educationunder Collective and Household Farming in China.” Journalof Development Economics 56 (2): 307–35.

Meng Xin. 2000. Labour Market Reform in China. Cambridge:Cambridge University Press.

Michelson, Ethan, and William Parish. 2000. “Gender Differen-tials in Economic Success: Rural China in 1991.” In BarbaraEntwisle and Gail E. Henderson, eds., Redrawing Boundaries:Work, Households and Gender in China. Berkeley: Universityof California Press.

National Bureau of Statistics. 1998. Zhongguo Shehui yu KejiFazhan Baogao [China Social and Science and TechnologyDevelopment Report]. Beijing: China Statistics Press.

———. 2001. China Statistical Abstract 2001. Beijing: ChinaStatistics Press.

———. Various years. China Statistical Yearbook. Beijing: ChinaStatistics Press.

National Bureau of Statistics Rural Social and Economic SurveyTeam. 2000. Zhongguo Nongcun Zhuhu Diaocha Nianjian

2000 [China Rural Household Survey Yearbook 2000].Beijing: China Statistics Press.

Parish, W. L., X. Zhe, and F. Li. 1995. “Nonfarm Work andMarketization of the Chinese Countryside.” China Quarterly143: 697–730.

Rawski, T. G., and R. W. Mead. 1998. “On the Trail of China’sPhantom Farmers.” World Development 26 (5): 767–81.

Rodrik, Dani. 2000. “Comments on ‘Trade, Growth, and Poverty,’by David Dollar and Aart Kraay.”Available at http://ksghome.harvard.edu/~.drodrik.academic.ksg/papers.html.

Rozelle, Scott, Guo Li, Minggao Shen, Amelia Hughart, and JohnGiles. 1999. “Leaving China’s Farms: Survey Results of NewPaths and Remaining Hurdles to Rural Migration.” ChinaQuarterly 158 (June): 367–93.

Skoufias, Emmanuel. 1994. “Using Shadow Wages to EstimateLabor Supply of Agricultural Households.” American Journalof Agricultural Economics 76 (May): 215–27.

UNDP (United Nations Development Programme). 1998. ChinaHuman Development Report: Human Development andPoverty Alleviation, 1997. Beijing: UNDP.

Winters, L. Alan. 2000. “Trade, Trade Policy, and Poverty: Whatare the Links?” Centre for Economic Policy Research PaperNo. 2382. Centre for Economic Policy Research, London.

World Bank. 2001. China: Overcoming Rural Poverty.Washington, D.C.

Wu, Harry X., and LI Zhou. 1996 “Research on Rural to UrbanLabour Migration in the Post-Reform China: A Survey.”Chinese Economy Research Unit Working Paper 96/4.University of Adelaide, February.

Yang, Dennis Tao. 1997. “Education and Off–Farm Work.”Economic Development and Cultural Change 45 (3): 613–32.

Yang, Dennis Tao, and Hao Zhou. 1999. “Rural-Urban Disparityand Sectoral Labour Allocation in China.” Journal ofDevelopment Studies 35 (3): 105–33.

Yao Yang. 1999. “Rural Industry and Labor Market Integrationin Eastern China.” Journal of Development Economics 59 (2):463–96.

Zhao Yaohui. 1999a. “Labor Migration and Earnings Differ-ences: The Case of Rural China.” Economic Development andCultural Change 47 (4): 767–82.

———. 1999b. “Leaving the Countryside: Rural-to-UrbanMigration Decisions in China.” American Economic Review89 (2): 281–86.


within countries can add considerable noise tothe measured levels and changes in poverty andinequality. It is unclear how much power cross-country data sets have for detecting any underlyingeffects of greater openness or other covariates.There is also the issue of whether trade volume canbe treated as exogenous in these cross-countryregressions; it is clearly not a policy variable andmay well be highly correlated with other (latent)attributes of country performance independently oftrade policy. The attribution of inequality impactsto trade policy reforms themselves is clearlyproblematic. The correlations (or their absence)found in cross-country studies can also be decep-tive because starting conditions can vary so muchbetween reforming countries. Averaging across thisdiversity in initial conditions can readily hide sys-tematic effects of relevance to policy (Ravallion2001).

In principle, such problems in cross-countrycomparative work can be dealt with by better dataand methods. However, the concerns go deeper.Aggregate inequality or poverty may not changewith trade reform even though there are both gain-ers and losers at all levels of living. Survey datatracking the same families over time commonlyshow considerable churning under the surface.1

The data show that many people have escaped

There has been much debate about the welfareimpacts of greater trade openness. Some argue thatexternal trade liberalization is beneficial to the poorin developing countries, while others argue that thebenefits will be captured by people who are notpoor. Expected impacts on relative wages (notablybetween skilled and unskilled labor) and relativeprices (such as between food staples and luxury im-ports) have figured prominently in debates aboutthe welfare impacts.

What does the evidence suggest? One mighthope to provide a conclusive answer by comparingchanges over time in measures of inequality orpoverty between countries that are open to externaltrade and countries that are not. A number ofattempts to throw empirical light on the welfareeffects of trade liberalization have been made usingaggregate cross-country data sets that combinesurvey-based measures of inequality or povertywith data on trade openness and other control vari-ables (see Bourguignon and Morisson 1990;Edwards 1997; Barro 2000; Dollar and Kraay 2002;Lundberg and Squire 2003).

However, caution in drawing implications fromsuch cross-country comparisons is appropriate.Concerns have been raised about data and econo-metric specifications. Differences in survey designand processing between countries and over time

261

15

Welfare Impacts ofChina’s Accession to

the WTO

Shaohua Chen and Martin Ravallion

poverty while others have fallen into poverty, eventhough the overall poverty rate is unchanged.

Numerous sources of such diverse impacts canbe found in developing country settings. For exam-ple, geographic disparities in access to human andphysical infrastructure affect prospects for partici-pating in economic growth.2 For China the eco-nomic geography of poverty and how this interactswith geographic diversity in the impacts of policyreforms are high on the domestic policy agenda. Apolicy analysis that simply averaged over such dif-ferences would miss a great deal of what matters tothe debate on policy.

This chapter follows a different approach inwhich the attribution to trade policy changes isunambiguous and the diversity of welfare impactsis not lost. The article examines the welfare impactsat the household level of the changes in commodityand factor prices attributed to a specific tradepolicy reform, namely China’s accession in 2001 tothe World Trade Organization (WTO). For China,this meant a sharp reduction in tariffs, quantitativerestrictions, and export subsidies, with implica-tions for the domestic structure of prices and wagesand thus for household welfare. Drawing on esti-mates by Ianchovichina and Martin (2002) of theimpacts of reform on prices (for both commoditiesand factors of production), the following analysisapplies standard methods of first-order welfareanalysis to measure the gains and losses at thehousehold level using large sample surveys col-lected by China’s National Bureau of Statistics.

Measuring the Welfare Impacts ofTrade Reform

Past approaches to studying the welfare impacts of specific trade reforms have tended to be eitherpartial equilibrium analyses, which measurehousehold-level welfare impacts of the direct pricechanges due to tariff changes using survey data(typically) covering many thousands of randomlychosen households, or general equilibrium analy-ses, which use a computable general equilibrium(CGE) model to capture second-round responses.3

While partial equilibrium analysis requires little orno aggregation of the primary household data, itmisses potentially important indirect effects onprices and wages. General equilibrium analysis hasthe power to capture these effects by simulating

economywide impacts on markets. However, stan-dard CGE models entail considerable aggregationacross household types, with rarely more than sixor so “representative households.” Such models arecrude tools for welfare-distributional analysis.

The challenge for applied work is to find anapproach that respects the richness of detail avail-able from modern integrated household surveyswhile ensuring that the price changes attributed toreform are internally consistent with economywideequilibrium conditions. In principle, the CGEmodel could be built onto the household survey, sothat the number of households in the model is thenumber sampled in the survey.4 For this study, thatdegree of integration would require an extraordi-narily high dimensional CGE model, with 85,000households. This is currently not a feasible route.

The intermediate approach used here carries the reform-induced commodity and factor pricechanges simulated from a general equilibriummodel to the level of all the sampled households inthe survey.5 The welfare impacts are measuredusing standard tools of analysis familiar from pastwork on the welfare effects of price changes associ-ated with tax and trade policy reform. This ap-proach imposes minimal aggregation conditionson the survey data, within unavoidable data limita-tions. In addition to calculating the trade reform’soverall effects on poverty and inequality, thisapproach provides a detailed socioeconomic mapof impact, showing how it varies with other nonin-come characteristics, such as location. This gener-ates better insights to the questions policymakersask about who gains and who loses from reform.

The general equilibrium analysis generates a setof price and wage changes. These embody both thedirect price effects of the trade policy change andsecond-round, indirect effects on the prices of non-traded goods and on factor returns, including ef-fects operating through the government’s budgetconstraint. Ianchovichina and Martin (2002) use acompetitive market-clearing model from the GlobalTrade Analysis Project.6 The revenue implicationsof the trade policy change are reflected in changes inindirect tax rates.7 Since the price changes are basedon an explicit model, their attribution to the tradepolicy reform is unambiguous, thus avoiding theidentification problems common to past attemptsto estimate distributional effects of trade policyreform using cross-country comparisons.


The approach can be outlined as follows. Eachhousehold has preferences over consumption andwork effort (under the standard assumption thatgoods have positive marginal utilities while laborsupply has negative marginal utility) represented bythe utility function ui (q d

i , L i ), where q di is an m-

dimension vector of the quantities of commoditiesconsumed by household i and L i is a vector of laborsupplies by activity, including supply to the house-hold’s own production activities. The household isassumed to be free to choose q d

i and L i subject to itsbudget constraint. Consistently with the generalequilibrium model that generated the price andwage changes, there is no rationing at the householdlevel; for example, involuntary unemployment isruled out.

The indirect utility function of household i isgiven by:

vi

[pd

i , wi , πi

] = max(q d

i , L i )

[ui

(q d

i , L i

) ∣∣pdi q d

i

= wi L i + πi

](15.1)

where pdi is the price vector (of dimension m) for

consumption, wi is the vector of wage rates, and πi

is the profit obtained from all household enter-prises as given by:

πi

(ps

i , pdi , wi

) = max(zi , L o

i )

[ps

i q si − pd

i zi − wi L oi

∣∣q sij

≤ fij

(zij, L o

ij

), j = 1, . . . , m;∑

j

zij ≤ zi ,∑

j

L oij ≤ L o

i

](15.2)

where psi is the m-vector of supply prices, q s

i is thecorresponding vector of quantities supplied, L o

i isthe labor input to own-production activities, ofwhich L o

ij is used in producing good j, fij is thehousehold-specific production function for good j(embodying fixed factors), and the zi terms are thecommodities used as production inputs, of whichzij is used in producing good j.

Measurement of the welfare impacts is of courseconstrained by the data, which do not include ini-tial price and wage levels.8 However, this data limi-tation does not matter in calculating a first-orderapproximation to the welfare impact in a neighbor-hood of the household’s optimum. Taking the dif-ferentials of equations 15.1 and 15.2 and using theenvelope property (whereby the welfare impacts ina neighborhood of an optimum can be evaluatedby treating the quantity choices as given), the

monetary value of the change in utility for house-hold i is given by:

gi ≡ dui

vπ i=

m∑j=1

[ps

ijqsij

dpsij

psij

− pdij

(q d

ij + zij

) dpdij

pdij

]

+n∑

k=1

(wk L s

ik

dwk

wk

)(15.3)

where vπ i is the marginal utility of income forhousehold i (the multiplier on the budget con-straint in equation 1) and L s

ik = L ik − L oik is the

household’s “external” labor supply to activity k.(Notice that gains in earnings from labor used inown-production are exactly matched by the highercost of this input to own-production.)

Equation 15.3 is the key formula used in calcu-lating the household-level welfare impacts of theprice changes implied by the general equilibriumanalysis of the trade policy reform. The propor-tionate changes in all prices and wages are weightedby their corresponding expenditure and incomeshares. The weight for the proportionate change inthe j’th selling price is ps

ijqsij , the revenue (selling

value) from household production activities insector j. Similarly, −pd

ij (q dij + zij) is the (negative)

weight for demand price changes, and wk L sik is the

weight for changes in the wage rate for activity k.The term ps

ijqsij − pd

ij (q dij + zij) is referred to as net

revenue, which (to a first-order approximation)gives the welfare impact of an equiproportionateincrease in the price of commodity j.

With the gain (or loss) to each householdcalculated based on equation 15.3, the covariatesof those gains can now be examined. One covariateof obvious interest is income, needed to assessimpacts on aggregate poverty and inequality. Ide-ally, one would use a money metric of utility basedon equation 15.1. However, that would require anexplicit model of the demand and supply system(that can be integrated back to obtain the indirectutility function). Again, feasibility becomes an issuebecause of the absence of complete data on priceand wage levels. Thus there is little choice but to useincome as the money metric of utility, in effectignoring all geographic differences in the pricesfaced or in the extent to which border price changesare passed on locally. However, we make a seem-ingly plausible allowance for urban–rural cost ofliving differences in this setting.

Welfare Impacts of China’s Accession to the WTO 263

Two further limitations of this approach shouldbe noted. First, applying the calculus in deriving(15.3) implicitly assumes small changes in prices.Relaxing this requires more information on pricelevels and the structure of the demand and supplysystem.9 This would entail considerable furthereffort, and the reliability of the results will be ques-tionable given the problem of incomplete price andwage data.

Second, as already noted and consistent with thegeneral equilibrium analysis, this approach alsorules out rationing in commodity or factor marketsor nonconvexities in consumption or production.In principle, these problems can also be handledthrough a completely specified demand model,which can be used to estimate the virtual prices atwhich the rationed demand or supply would bechosen. This is not feasible without data on priceand wage level.

Setting and Data

While the official date of China’s WTO accession is2001, it is clear that the Chinese economy hadalready started to adapt to this expected changewell before 2001. The trade reform can thus bethought of as having two stages, a lead-up period inwhich tariffs started to fall in anticipation of WTOaccession and the period from 2001 onward.Ianchovichina and Martin (2002) argue for 1995 asa plausible beginning of the lead-up period. Theirestimates of the price changes induced by WTOaccession for the periods 1995–2001 and 2001–07are used in this analysis. While the primary focus ison the second period, welfare impacts are also esti-mated for the lead-up period.

The measure of welfare impacts given by equa-tion 15.3 is calibrated to survey data for 1999, twoyears before official WTO accession and a few yearsafter the likely beginning of the lead-up period. Thechoice of 1999 was partly made for data reasons,since it was the most recent year for which themicro data were available. Choosing a year near themiddle of the lead-up period (rather than a surveyat the beginning or end) should also diminishbiases due to any nonlinearity in the welfareimpacts of price and wage changes.

Survey Data

The survey data used in this study are from the1999 Urban Household Survey and the 1999 Rural

Household Survey by China’s National Bureau ofStatistics (NBS). The sample size is 67,900 house-holds for the rural survey and 16,900 households(out of the survey total of 40,000 households) forthe urban survey.10 Over the past 15 years, NBS hasworked to improve both surveys, focusing on sam-ple coverage, questionnaire design, methodology,and data processing.11 The number of variables inthe surveys has increased dramatically, with addi-tional details on income, expenditure, savings,housing, and productivity, among others. NBS alsoprovided micro data for three provinces (Liaoning,Guangdong, and Sichuan—the “test provinces”).The computer program to implement the estima-tion method was written for these data, after whichthe program was run by NBS staff on the entirenational data set.

A number of problems remain in the 1999 sur-veys. For a sample frame, the rural survey relies onits sampled counties from 1985, which may nolonger be representative. The urban survey excludesrural migrants, since the base of the sample frame isthe legal registration system (hukou). As in othercountries the rural survey gives data on the remit-tances of migrants workers, but it does not provideinformation about the migrant workers themselves,who (unlike in other countries) are not sampled inthe urban survey either. This makes it difficult tomeasure impacts through labor mobility and rural-urban transfers.

Comparisons between the rural and urban sur-veys also present problems. For example, income inthe rural survey includes in-kind income (such asfrom own-farm production and other householdenterprises), but income in the urban survey ig-nores some in-kind components, notably subsidiesfrom the government.

Sampling Weights

The population census puts the 1999 urban popu-lation share at 34 percent while the sample-basedurban population share is 20 percent. To correctthe rural and urban sampling weights, the urbanpopulation share from the China Statistical Year-book (NBS 2000) was used to replace the surveysample weights to form the national figures.

Matching the Global Trade Analysis Project Modeland the Surveys

There are 57 sectors in the Global Trade AnalysisProject (GTAP) model. The China GTAP model


used in this study regroups these 57 sectors into 25:rice, wheat, feed grains, vegetables and fruits,oilseeds, sugar, plant-based fibers, livestock andmeat, dairy, processed food, beverages and tobacco,extractive industries, textiles, apparel, light manu-factures, petrochemicals, metals, automobiles, elec-tronics, other manufactures, trade and transport,construction, communications, commercial serv-ices, and other services. To these are added land,capital, and three types of labor (see below).

China’s rural and urban surveys have about 2,000categories for consumption and production. Thevariables from the household surveys are matchedto the closest category in the GTAP model. Forexample, corn, millet, and potatoes are placed in thecategory“feed grains”and cotton and fiber crops areplaced in the category of “plant-based fibers.” Theworking paper version, Chen and Ravallion 2003gives details on how the variables from the surveysare matched to the GTAP model sectors.

Definitions of Labor and Labor Earnings

The China GTAP model defines three types oflabor: unskilled farm labor, unskilled nonfarmlabor, and skilled nonfarm labor.12 Since the ruraland urban surveys have different questionnaires,rural and urban labor earnings are treated differ-ently. In the urban survey three variables—sector,occupation, and education—are used to determinelabor types. However, “sector” or “occupation”alone cannot indicate whether a person should beclassified as skilled labor. For example, the financialsector may hire some unskilled labor while theservices sector may hire some skilled labor. Simi-larly, a train driver in the occupation category“workers and staff-members in production andtransportation” counts as skilled labor. Accordingly,education is also taken into account. Workers whohave received education at the senior high schoollevel or above are considered skilled labor. Othersare classified as unskilled labor.

It is more difficult to determine the type of laborincome for rural areas. There is no information onhow much each person earns and from what work.Consequently, labor earnings can be classified onlyroughly by income source. For instance, all laborremuneration from agriculture is considered in-come from unskilled farm labor; earnings fromindustry or construction, grain processing, andthe like are considered income from unskillednonfarm labor; earnings from the services sector,

transportation and trade, and the like are consid-ered income from skilled nonfarm labor.

Land

Under China’s economic reforms, which began in1978, all farmers have land use rights but not theright to sell, although they can subcontract the allo-cated land to other farmers. Therefore, the changein land prices from the GTAP model affects onlythe value of land rentals paid and received.

Household Income

For assessing the overall impacts on poverty andinequality, rural and urban households are com-bined. There is no cost of living index betweenurban and rural areas of China; urban and ruralconsumer price indexes are both indexed to 100 atthe base date. The urban price level is assumed tobe 15 percent higher than the rural price level. Thisdifferential is less than for other developing coun-tries because subsidies to urban households inChina help to compensate for higher housing andfood costs than in rural areas.

Income per person is used as the welfare indica-tor (so that all households are ranked by per capitaincome, from the poorest to the richest). This istermed “net income” in the rural survey and “dis-posable income” in the urban survey. Post-reformincome is then income plus the estimated gain de-fined by equation 15.3.

Measured Welfare Impacts of WTOAccession

Based on the predicted relative price and wagechanges from the GTAP model for 1995–2000(table 15.1) and 2001–07 (table 15.2) and produc-tion and consumption shares from the 1999 ruraland urban household survey data, equation 15.3can be used to compute the net gain for eachhousehold. The first panel in table 15.3 gives themean gains for 1995–2001 and 2001–07, split byurban and rural areas. The second panel gives theGini indices, both actual (for the baseline year,1999) and simulated. The two simulated incomedistributions are obtained by subtracting theestimated gains over 1995–2001 from the 1999incomes at the household level and by adding thehousehold-specific gains from 2001–07 to the 1999incomes. Thus the first simulation shows thedistributional impact of the price changes during


the first stage of the trade reform (what the baselinedistribution would have looked like without thereforms) while the second shows the impact ofthe post-2001 price changes (how the changes areexpected to affect the baseline distribution, lookingforward). The third panel gives the headcountindex of poverty for the official poverty linebased on the poverty lines used by China’s

National Bureau of Statistics and for the $1 a dayand $2 a day poverty lines from Chen and Ravallion(2001).

There is an overall gain of about 1.5 percent ofmean income.All of this gain is in the period leadingup to WTO accession. There is almost no impact oninequality, either in the period leading up to WTOaccession or predicting forward. The aggregate Gini


TABLE 15.1 Predicted Price Changes from GTAP Model and Per Capita Net Gain or Loss for Rural and Urban Households, 1995–2001

Rural Urban

Wholesale Consumer Net Mean Welfare Net Mean WelfareExpenditures and Prices Prices Revenue Change Revenue ChangeIncome Sources (Percent) (Percent) (Yuan) (Yuan) (Yuan) (Yuan)

ExpendituresRice 0.5 1.5 73.66 0.15 −109.33 −1.64Wheat −1.7 −1.5 40.86 −0.74 0.00 0.00Feedgrains 2.6 10.7 117.04 2.15 0.00 0.00Vegetables and fruits 0.5 1.5 123.41 0.13 −378.69 −5.68Oilseeds −0.6 −0.8 37.05 −0.24 −1.04 0.01Sugar 0.7 1.4 13.74 0.05 −174.06 −2.44Plant-based fibers −3.6 −1.9 36.84 −1.34 0.00 0.00Livestock and meat 2.0 3.1 194.62 2.59 −500.65 −15.52Dairy 1.5 2.5 2.50 0.02 0.00 0.00Other food 1.2 3.1 −81.60 −3.39 −343.13 −10.64Beverages and tobacco −4.6 −7.2 −72.98 5.25 −197.20 14.20Extractive industries −0.2 0.8 17.99 −0.44 −173.03 −1.38Textiles −5.0 −8.9 −11.08 0.99 −53.50 4.76Apparel −2.7 −7.4 −64.13 4.75 −394.30 29.18Light manufacturing −0.3 −2.5 −16.15 0.40 −82.96 2.07Petrochemical industry −0.7 −0.1 −325.39 0.33 −398.23 0.40Metals −0.7 −0.1 −15.30 0.02 −24.02 0.02Automobiles −17.7 −20.4 −52.27 10.66 −37.76 7.70Electronics −1.5 −4.0 −24.27 0.97 −162.69 6.51Other manufactures −0.6 −0.3 −264.61 0.79 −431.16 1.29Trade and transport 0.2 1.3 −18.70 −0.24 −110.53 −1.44Construction 0.1 1.1 0.00 0.00 −31.11 −0.34Communication 0.9 1.9 −16.72 −0.32 −152.04 −2.89Commercial services 0.8 1.8 −61.37 −1.10 −533.33 −9.60Other services 0.1 1.1 −414.45 −4.56 −680.99 −7.49

Income sourcesFarm unskilled labor 1.7 1.7 313.58 5.22 0.00 0.00Nonfarm unskilled labor 1.7 1.7 287.19 4.78 1227.51 20.44Skilled labor 2.0 2.0 360.87 7.09 3391.11 66.64Land 1.3 1.3 17.08 0.22 0.00 0.00Capital 1.3 1.3 21.14 0.27 126.01 0.77

Sources: Ianchovichina and Martin 2002; author’s calculations from China National Bureau of Statistics1999 Rural Household Survey and 1999 Urban Household Survey.

index increased slightly, from 39.3 percent withoutWTO accession to 39.5 percent after accession.

The incidence of poverty would have beenslightly higher in 1999 if not for the trade policychanges in the lead-up to WTO accession, whilepoverty is predicted to increase slightly during2001–07 due to the expected price changes induced

by the remaining tariff changes during that period.The impacts on rural and urban poverty for a widerange of poverty lines can be seen in figures 15.1and 15.2, which give the cumulative distributions ofincome for both the baseline and the two simulateddistributions for the poorest 60 percent in ruralareas and 40 percent in urban areas, respectively.


TABLE 15.2 Predicted Price Changes from GTAP Model and Per Capita Net Gain orLoss for Rural and Urban Households, 2001–07

Rural Urban

Wholesale Consumer Net Mean Welfare Net Mean WelfareExpenditures and Prices Prices Revenue Change Revenue ChangeIncome Sources (Percent) (Percent) (Yuan) (Yuan) (Yuan) (Yuan)

ExpendituresRice −1.4 0.7 73.66 −1.39 −109.33 −0.75Wheat −1.5 0.7 40.86 −0.92 0.00 0.00Feedgrains −3.7 2.1 117.04 −4.90 0.00 0.00Vegetables and fruits −2.6 −0.6 123.41 −4.02 −378.69 2.24Oilseeds −5.7 −5.9 37.05 −2.10 −1.04 0.06Sugar −2.8 −3.5 13.74 −0.34 −174.06 6.01Plant-based fibers 1.6 4.1 36.84 0.56 0.00 0.00Livestock and meat −1.5 0.7 194.62 −5.21 −500.65 −3.40Dairy −2.4 −0.5 2.50 −0.09 0.00 0.00Other food −3.1 −2.7 −81.60 2.04 −343.13 9.32Beverages and tobacco −5.6 −7.7 −72.98 5.62 −197.20 15.09Extractive industries −0.4 1.7 17.99 −0.86 −173.03 −2.92Textiles −0.2 −1.5 −11.08 0.17 −53.50 0.82Apparel 2.6 0.8 −64.13 −0.51 −394.30 −2.98Light manufacturing −0.6 0.5 −16.15 −0.08 −82.96 −0.43Petrochemical industry −1.1 0.8 −325.39 −2.60 −398.23 −3.19Metals −0.6 1.3 −15.30 −0.20 −24.02 −0.31Automobiles −3.8 −4.0 −52.27 2.09 −37.76 1.52Electronics −1.2 −1.4 −24.27 0.34 −162.69 2.20Other manufactures −0.8 0.8 −264.61 −2.12 −431.16 −3.46Trade and transport −0.4 1.7 −18.70 −0.32 −110.53 −1.85Construction −0.4 1.7 0.00 0.00 −31.11 −0.52Communication −0.4 1.7 −16.72 −0.28 −152.04 −2.54Commercial services −1.1 0.9 −61.37 −0.55 −533.33 −4.72Other services −0.7 1.3 −414.45 −5.39 −680.99 −8.76

Income sourcesFarm unskilled labor −0.3 −0.3 313.58 −0.85Nonfarm unskilled labor 1.0 1.0 287.19 2.96 1227.51 12.64Skilled labor 0.4 0.4 360.87 1.55 3391.11 14.58Land −4.7 −4.7 17.08 −0.80Capital 0.6 0.6 21.14 0.13 126.01 0.80

Sources: Ianchovichina and Martin 2002; authors’ calculations from China National Bureau of Statistics1999 Rural Household Survey and 1999 Urban Household Survey.


TABLE 15.3 Summary Statistics on Aggregate Welfare Impacts, 1995–2001and 2001–07

Item Rural Urban National

Mean gains (yuan per capita)1995−2001 34.47 94.94 55.49

(1.54%)a

2001−07 −18.07 29.45 –1.54(–0.04%)a

Inequality impacts (Gini index as percentage)Baseline, 1999 33.95 29.72 39.31Simulated: Less gains 1995–2001 33.90 29.68 39.27Simulated: Plus gains 2001–07 34.06 29.65 39.53

Poverty impacts (headcount index, percentage)b

Official poverty lineBaseline, 1999 4.38 0.08 2.92Simulated: Less gains 1995–2001 4.56 0.08 3.04Simulated: Plus gains 2001–07 4.57 0.07 3.04

$1/day (1993 purchasing power parity)Baseline, 1999 10.51 0.29 7.04Simulated: Less gains 1995–2001 10.88 0.28 7.28Simulated: Plus gains 2001–07 10.81 0.28 7.23

$2/day (1993 purchasing power parity)Baseline, 1999 45.18 4.07 31.20Simulated: Less gains 1995–2001 46.10 4.27 31.88Simulated: Plus gains 2001–07 45.83 3.97 31.60

a. Percentage of mean income.b. Official poverty line is from China National Statistics Bureau; poverty lines of $1 a day and $2 a day arefrom Chen and Ravallion 2001.Sources: Authors’ computations from China National Bureau of Statistics 1999 Rural Household Surveyand 1999 Urban Household Survey.

60

50

40

30

20

10

0400 600

Annual per capita income (yuan)800 1,000 1,200 1,400 1,600 1,800 2,000

Percentage of population below poverty line

Pre-WTO Post-WTO Baseline distribution

Sources: China National Bureau of Statistics; Chen and Ravallion 2001.

FIGURE 15.1 Poverty Incidence Curves, Rural

While there is virtually zero aggregate impactwhen predicting forward from WTO accession, thedisaggregated results show a more nuanced picture.The analysis focuses on three measures of impact at the household level: the absolute gain or loss, gi ;the proportionate gain or loss, gi/yi ; and whetherthere is a gain or not, I (gi ), where I is the indicatorfunction. This third measure helps to determinewhere there might be high concentrations of losers,in specific areas or socioeconomic groups.

The results by provinces ranked by mean incomeper person are plotted in figure 15.3 for meanabsolute gains (gi in yuan per capita), in figure 15.4for proportionate gains (gi/yi , as a percentage),and in figure 15.5 for the proportion of householdsthat registered positive gains. The average gain orloss by province for urban and rural areas and thenumber of gainers in each case are shown in appen-dix tables 15.A.1 and 15.A.2, respectively; Chen andRavallion 2003 give the province rankings.

The same results are also plotted in figures 15.6–15.8 against percentiles of the income distribution.So, for example, to see the mean impact in yuan percapita at the median income one looks at the 50thpercentile of figure 15.6. (Notice that figure 15.6gives the horizontal differences in figures 15.1 and15.2 plotted against the point on the vertical axis.)

In the aggregate, about three-quarters of ruralhouseholds and one-tenth of urban households

will experience a real income loss. Farm income ispredicted to drop by 18 yuan per person whileurban income rises by 29 yuan per person. Thebreakdown by sectors in table 15.2 shows that the decline in rural income is due to the drop inwholesale prices for most farm products, plushigher prices for education and health care. Farm-ers will also benefit from the drop in some con-sumer prices and from the increase in nonfarmlabor wages. In urban areas residents will enjoylower prices for most farm products and higherwages, but they will also be hit by increases in serv-ice fees for education and health care.

Impacts differ considerably across regions (seefigures 15.3–15.5 and appendix tables 15.A.1 and15.A.2).The mean absolute gains tend to be highest inthe richest provinces in both urban and rural areas(figure 15.3), though there is no correlation betweenthe proportionate gains and mean income ofprovinces (figure 15.4). One spatially contiguousregion—the northeast provinces of Inner Mongolia,Liaoning, Jilin, and Heilongjiang—stands out as hav-ing the largest loss from the reform.Both absolute andproportionate impacts are highest in this region—more than 90 percent of farmers in Heilongjiang andJilin are predicted to experience a net loss.

Notice that these geographic differences inwelfare impacts arise entirely from differences inconsumption and production behavior. In reality,


35

40

30

25

20

15

5

10

0500 1,000

Annual per capita income (yuan)1,500 2,000 2,500 3,000 3,500 4,000 4,500

Percentage of population below poverty line

Pre-WTO income Baseline distributionin 1999

Post-WTO income

Source: Authors’ computations based on data from China NationalBureau of Statistics 1999 Rural Household Survey and 1999 UrbanHousehold Survey.

FIGURE 15.2 Poverty Incidence Curves, Urban


60

40

20

0

�20

�40

�60

�801

Provinces ranked by provincial per capita income6 11 16 21 26 31

Net gain/loss per capita (yuan)

Urban Total Rural


FIGURE 15.3 Mean Gains by Provinces: AbsoluteGain in Yuan Per Capita

2

1

0

�1

�2

�31


Percentage net gain/loss per capita

Urban Total Rural


FIGURE 15.4 Mean Gains by Provinces: ProportionateGains in Percent


100

90

70

80

60

40

50

30

20

10

01


Percentage of gainers

Urban Total Rural

Note: Urban prices are assumed to be 15 percent higher than ruralprices.Source: Authors’ computations based on data from China NationalBureau of Statistics 1999 Rural Household Survey and 1999 UrbanHousehold Survey.

FIGURE 15.5 Mean Gains by Provinces: Percentage of Gainers by Provinces

30

40

20

10

0

�20

�10

�30

�40

�50

�600

Percentage of population ranked by per capita income2010 30 40 6050 70 80 90 100

Net gain/loss per capita (yuan)

Urban National Rural

Note: Urban prices are assumed to be 15 percent higher than ruralprices.Source: Authors’ compilations based on data from China NationalBureau of Statistics 1999 Rural Household Survey and 1999 UrbanHousehold Survey.

FIGURE 15.6 Mean Gains in Yuan by Income Percentile


2.0

1.0

0.0

�2.0

�1.0

�3.0

�4.0

�5.0

�6.00


Percentage net gain or loss per capita

Urban National Rural


FIGURE 15.7 Mean Percentage Gain by IncomePercentile

100.0

90.0

80.0

60.0

70.0

50.0

40.0

30.0

20.0

10.00


Percentage of gainers

Urban Total Rural


FIGURE 15.8 Percentage of Gainers by IncomePercentile

differential impacts on local prices are likely, due totransport or other impediments to internal trade.The analysis here does not incorporate such differ-ences, and doing so would pose a number of dataand analytic problems. This might, however, be afruitful direction for future work where the neces-sary data on prices and wage levels are available bygeographic area.

When households are ranked by initial income,there is a notable difference between urban andrural households, with absolute gains tending to behigher for higher income households in urbanareas, but lower for higher income households inrural areas (see figure 15.6). Nationally (combiningurban and rural areas with the corrected weightsdiscussed above), there is the hint of a U-shapedrelationship, though still with the highest absolutegains for the rich.

This picture is reversed for proportionate gains,which tend to fall as income rises in urban areas,but to rise with income in rural areas and nation-ally (see figure 15.7). In the aggregate the propor-tion of gainers rises with income, a result that isdriven by the rise in the number of gainers asincome increases in rural areas (see figure 15.8).

Explaining the Incidence of Gainsand Losses

The way the problem of measuring welfare impactswas formulated in section II allows utility andprofit functions to vary between households atgiven prices. To explain the heterogeneity in meas-ured welfare impacts, these functions can insteadbe supposed to vary with observed household char-acteristics. The indirect utility function becomesvi ( pd

i , wi , πi ) = v( pdi , wi , πi , x1i ) where πi =

π( psi , pd

i , wi , x2i ) for vectors of characteristics x1i

and x2i that shift the utility functions in equation15.1 and the profit functions in 15.2. Note thatthe characteristics that influence preferences overconsumption (x1i ) are allowed to differ from thosethat influence the outputs from own-productionactivities (x2i ).

The gain from the price changes induced bytrade reform, as given by equation 15.3, depends onthe household’s consumption, labor supply, andproduction choices, which in turn depend on pricesand characteristics, x1i and x2i . For example, house-holds with a higher proportion of children will

naturally spend more on food, so if the relativeprice of food changes, the welfare impacts will becorrelated with this aspect of household demo-graphics. Similarly, there may be differences intastes associated with stage of the life cycle and edu-cation. There are also likely to be systematic covari-ates of the composition of income.

Generically, the gain can be written as:

gi = g(pd

i , psi , wi , x1i , x2i

)=

m∑j=1

[ps

ijqs(

pdi , ps

i , wi , x2i

)dpsij

psij

− pdij

[q d

(pd

i , wi , πi , x1i

)+ zij

(pd

i , psi , wi , x2i

)]dpdij

pdij

]

+n∑

k=1

wk

[L ik

(pd

i , wi , πi , x1i

)

− L oik

(pd

i , psi , wi , x2i

)]dwk

wk(15.4)

Notice that the gain from reform is inherently non-separable, in that it cannot be written as a functionsolely of pd

i , x1i and πi because the gain alsodepends on production choices.

However, as noted above, household-specificwages and prices are not observed, so furtherassumptions are required. In explaining variationsacross households in the predicted gains from tradereform, wage rates are assumed to be a function of prices and characteristics as wi = w( pd

i , psi ,

x1i , x2i ), and differences in prices faced areassumed to be adequately captured by a completeset of county-level dummy variables.

Under these assumptions, and the linearizationof equation 16.4 with an additive innovation errorterm, the following regression model applies for thegains:

gi = β1x1i + β2x2i +∑

k

γk Dki + εi (15.5)

where Dki = 1 if household i lives in county k andDki = 0 otherwise and εi is the error term.

The characteristics considered include age andage-squared of the household head, education anddemographic characteristics, and land (interpretedas a fixed factor of production, since it is allocatedlargely by administrative means in rural China).Also included are dummy variables describingsome key aspects of the occupation and princi-ple sector of employment, such as whether the


household is a registered agricultural household,whether there is wage employment, whether thereis state-sector employment, and whether there isparticipation in a township and village enterprise.There are endogeneity concerns about these vari-ables, but they appear to be minor in this context,especially when weighed against the concerns aboutomitted variable bias in estimates that exclude thesecharacteristics. Under the usual assumption thatthe error term is orthogonal to these regressors,equation 5 is estimated by ordinary least squares.The model is estimated separately for urban and

rural areas in each of the three test provinces(Liaoning, Guangdong, and Sichuan) for whichcomplete micro data are available.

There are some differences in the explanatoryvariables between rural (tables 15.4 and 15.5) andurban areas (tables 15.6 and 15.7). Results are pre-sented for both absolute gains (gi ) and proportion-ate gains (gi/yi ). Recall that these are averagesacross the impacts of these characteristics on theconsumption and production choices that deter-mine the welfare impact of given price and wagechanges. This makes interpretation difficult. These


TABLE 15.4 Regression Results for Level of Gain (Yuan) in Rural Areas of ThreeProvinces, 2001–07

Variable Liaoning Guangdong Sichuan

Log of household size 37.642 (6.42) 28.822 (2.64) 4.958 (2.16)Age of household head −2.425 (−3.11) −1.783 (−2.60) −0.548 (−1.51)Age of household head squared 0.026 (3.36) 0.017 (2.66) 0.005 (1.30)Agriculture household −10.942 (−3.31) −42.850 (−6.45) −37.723 (−6.54)Number of employees/ 12.665 (4.10) −6.932 (−0.29) 12.652 (3.02)

household sizeNumber of township and 10.768 (3.13) 29.466 (3.06) 15.327 (4.26)

village enterprise workers/household size

Number of migrant workers/ 5.399 (1.73) 7.798 (2.35) 7.067 (3.79)household size

Area of cultivated land −0.027 (−5.73) −0.002 (−1.00) −0.001 (−0.28)Area of hilly land 0.000 (−0.05) −0.001 (−0.87) 0.002 (1.94)Area of fishpond land −0.001 (−0.94) −0.070 (−2.85) 0.000 (0.04)

Highest education levelIlliterate or semi illiterate 7.926 (1.04) 19.016 (1.25) 8.387 (0.92)Primary school 0.071 (0.01) −2.148 (−0.13) 9.694 (1.06)Middle school −0.755 (−0.11) −4.261 (−0.26) 7.669 (0.84)High school 2.125 (0.31) 2.806 (0.18) 9.675 (1.03)Technical school −3.096 (−0.44) −36.482 (−1.09) 4.270 (0.38)College (default)

Labor force/household size 0.576 (0.08) 2.877 (0.15) −4.995 (−1.16)Children under 6/household 46.999 (2.71) 8.109 (0.35) −2.291 (−0.45)

sizeChildren ages 6−11/household 1.414 (0.11) 2.247 (0.10) −9.011 (−1.50)

sizeChildren ages 12−14/household −0.155 (−0.01) −24.489 (−1.20) −9.606 (−1.51)

sizeChildren ages 15−17/household −2.592 (−0.22) −23.390 (−1.02) −5.485 (−0.73)

sizeConstant −17.851 (−0.82) −17.742 (−0.65) −17.220 (−1.43)

R-square 0.278 0.116 0.116

Note: Numbers in parentheses are t-statistics.Source: Authors’ computations based on data from China National Bureau of Statistics 1999 RuralHousehold Survey and 1999 Urban Household Survey.

regressions are mainly of descriptive interest, tohelp isolate covariates of potential relevance inthinking about compensatory policy responses.

For rural areas the results show that the predictedgain from trade reform tends to be larger for largerhouseholds in all three provinces. There is also a U-shaped relationship with age of the household head:the gains reach a minimum around 50 years of age (47in Liaoning, 52 in Guangdong, and 55 in Sichuan).The gains are lower for agricultural households andhigher for households with more employees and

more township and village enterprise workers, withmore migrant workers, and with less cultivatedland (though significant only in Liaoning). The onlystrongdemographiceffect is thatyoungerhouseholds(with a higher proportion of children under six) tendto be gainers in Liaoning. While the results forthe county dummy variables are not shown (to savespace), losses were significantly higher than averagein six counties in Liaoning, seven in Guangdong, andsix in Sichuan. Table 16.8 gives the mean losses inthese counties for agricultural households.


TABLE 15.5 Regression Results for Percentage Gains in Rural Areas of ThreeProvinces, 2001–07


Log of household size 0.768 (2.46) 0.022 (0.20) 0.030 (0.40)Age of household head −0.108 (−2.17) −0.007 (−0.34) −0.004 (−0.31)Age of household head squared 0.001 (2.19) 0.000 (0.40) 0.000 (−0.02)Agriculture household −0.896 (−2.98) −1.365 (−14.85) −1.420 (−7.58)Number of employees/ 0.630 (2.76) 0.271 (2.57) 0.444 (3.61)

household sizeNumber of township and 0.669 (4.27) 0.585 (4.47) 0.548 (6.11)village enterprise workers/household sizeNumber of migrant workers/ 0.655 (3.59) 0.187 (3.59) 0.346 (7.08)

household sizeArea of cultivated land 0.000 (−1.77) 0.000 (−0.73) 0.000 (−1.61)Area of hilly land 0.000 (−0.48) 0.000 (−0.35) 0.000 (2.20)Area of fishpond land 0.000 (−0.17) −0.001 (−2.23) 0.000 (0.55)

Highest education levelIlliterate or semi illiterate 1.393 (2.18) 0.507 (1.26) −0.013 (−0.05)Primary school −0.634 (−2.01) −0.154 (−0.90) 0.069 (0.30)Middle school −0.891 (−3.08) −0.023 (−0.14) −0.011 (−0.05)High school −0.660 (−2.42) 0.010 (0.06) 0.006 (0.02)Technical school −0.573 (−1.87) −0.229 (−1.18) 0.038 (0.14)College (default)

Labor force/household size 0.456 (0.85) 0.323 (1.81) −0.099 (−0.71)Children under age 6/household 3.730 (3.61) 0.461 (1.49) −0.169 (−0.78)

sizeChildren ages 6−11/household 1.557 (1.41) 0.173 (0.72) −0.275 (−1.48)

sizeChildren ages 12−14/household 1.625 (1.54) −0.477 (−1.60) −0.343 (−1.85)

sizeChildren ages 15−17/household 1.325 (1.80) −0.289 (−0.91) −0.192 (−0.88)

size

Constant 0.788 (0.69) −0.709 (−1.39) −0.584 (−1.68)R-square 0.108 0.217 0.171

Note: Numbers in parentheses are t-statistics.Source: Authors’ computations based on data from China National Bureau of Statistics 1999 RuralHousehold Survey and 1999 Urban Household Survey.


TABLE 15.6 Regression Results for Level of Gain (Yuan) in Urban Areas of ThreeProvinces, 2001–07


Log of household size −5.627 (−1.81) 5.289 (0.27) −19.441 (−4.09)Single household head −1.366 (−0.4) −37.216 (−2.06) −17.369 (−3.61)Age of household head 0.531 (0.92) 5.266 (2.43) 1.542 (2.34)Age of household head squared −0.001 (−0.24) −0.040 (−1.8) −0.015 (−2.22)

Highest education level (default is college)

Primary school or lower 13.240 (2.95) 50.434 (2.4) 23.079 (3.11)Middle school 19.104 (5.99) 56.659 (3.58) 26.096 (4.34)High school 5.123 (1.62) 12.053 (0.95) 12.717 (2.39)Technical school 11.086 (3.23) 11.075 (0.88) 9.552 (1.62)College 3.974 (1.26) 3.447 (0.3) 11.013 (2.12)

Sector (default is government)Agriculture −16.310 (−1.22) −25.590 (−2.23) 17.293 (1.76)Mining −14.586 (−3.24) 19.351 (1.13) −3.851 (−0.53)Manufacturing −9.231 (−2.59) 17.773 (1.28) −4.634 (−1.2)Utility −9.387 (−1.63) −10.816 (−0.42) 1.516 (0.13)Construction −6.394 (−1.18) 8.622 (0.63) −4.409 (−0.92)Geological prospecting and −27.422 (−2.62) 20.089 (0.92) −16.585 (−0.83)

water conservancyTransportation and 6.368 (1.52) 16.525 (1.24) 1.644 (0.25)

telecommunicationsWholesale and retail −3.184 (−0.61) 5.664 (0.45) −1.983 (−0.4)Banking and finance −5.278 (−0.55) 3.888 (0.3) 9.491 (0.85)Real estate −11.708 (−1.71) 46.192 (1.35) 7.670 (0.37)Social services −5.542 (−1.02) −4.186 (−0.33) 0.504 (0.1)Health care −9.260 (−1.93) 0.683 (0.04) −1.049 (−0.17)Education −7.279 (−1.64) 7.649 (0.46) −5.219 (−0.87)Scientific research −20.982 (−4.06) 17.882 (1.14) −7.929 (−0.59)Other −7.784 (−1.42) −24.851 (−0.75) −7.012 (−0.73)

Type of employer (default is state owned)

Collective owned −1.927 (−0.76) 11.882 (0.54) −5.946 (−2.09)Foreign company −3.138 (−0.72) −10.988 (−1.22) 2.038 (0.31)Self-employed 4.278 (0.6) 9.448 (0.64) 10.582 (2.08)Privately owned business −9.587 (−1.41) −14.823 (−0.99) −4.601 (−0.57)Retirees reemployed −13.333 (−2.45) −35.591 (−1.82) −6.752 (−0.99)Retirees −15.569 (−3.66) −49.442 (−1.91) −12.218 (−1.95)Other −10.350 (−1.36) −6.568 (−0.34) −16.796 (−2.06)

Occupation (default is retiree)Engineer and technician 10.244 (1.66) 3.479 (0.12) 10.179 (1.49)Officers 12.747 (2.07) 17.701 (0.64) 10.564 (1.53)Staff in commerce 11.742 (2.08) 18.553 (0.65) 12.734 (1.92)Staff in services 19.940 (2.54) 3.380 (0.11) 4.057 (0.5)Worker in manufacturing 17.484 (2.02) 13.151 (0.47) 13.810 (1.86)Worker in transportation 21.469 (3.59) 9.637 (0.34) 16.117 (2.35)

and telecommunicationsOther 15.318 (2.05) 9.810 (0.27) −6.141 (−0.77)

Constant −10.744 (−0.77) −164.442 (−2.43) −17.611 (−1.1)R-square 0.265 0.131 0.181

Note: Numbers in parentheses are t-statistics.Source: Authors’ computations based on data from China National Bureau of Statistics 1999 Rural Household Survey and1999 Urban Household Survey.


TABLE 15.7 Regression Results for Percentage Gains in Urban Areas of ThreeProvinces, 2001–07


Log of household size 0.175 (3.54) −0.038 (−0.4) 0.036 (0.46)Single household head −0.022 (−0.36) −0.221 (−2.21) −0.259 (−3.07)Age of household head 0.000 (−0.01) 0.033 (2.55) 0.017 (1.53)Age of household head squared 0.000 (0.1) 0.000 (−2.12) 0.000 (−1.46)

Highest education level (default is university)

Primary school or lower 0.524 (6.43) 0.389 (3.7) 0.509 (5.15)Middle school 0.539 (10.41) 0.583 (7.25) 0.591 (8.27)High school 0.180 (3.56) 0.095 (1.46) 0.262 (3.83)Technical school 0.214 (4.04) 0.076 (1.22) 0.120 (1.79)College 0.054 (1.04) 0.015 (0.25) 0.125 (2.24)

Sector (default is government)Agriculture −0.079 (−0.32) 0.166 (2.2) 0.338 (2.64)Mining 0.183 (1.11) 0.346 (3.38) −0.129 (−1.01)Manufacturing −0.015 (−0.27) 0.114 (1.41) −0.021 (−0.34)Utility −0.040 (−0.36) −0.144 (−1.18) −0.134 (−0.84)Construction 0.095 (0.91) 0.109 (1.19) 0.036 (0.51)Geological prospecting −0.407 (−3.06) 0.178 (1.03) −0.228 (−0.53)

and water conservancyTransport and 0.206 (2.93) 0.060 (0.79) −0.036 (−0.4)

telecommunicationsWholesale and retail 0.060 (0.78) 0.081 (0.99) −0.015 (−0.18)Banking and finance −0.088 (−0.47) 0.049 (0.53) 0.013 (0.12)Real estate −0.108 (−0.91) 0.222 (1.16) 0.106 (0.29)Social services −0.090 (−1.09) 0.065 (0.69) 0.148 (1.37)Health care −0.088 (−1.1) 0.007 (0.06) −0.124 (−1.49)Education −0.057 (−0.75) 0.044 (0.44) −0.031 (−0.39)Scientific research −0.454 (−4.09) 0.126 (1.11) −0.082 (−0.73)Other 0.012 (0.14) 0.034 (0.25) −0.121 (−0.55)

Type of employer (default is state owned)

Collective owned 0.053 (1.16) 0.008 (0.08) 0.137 (1.73)Foreign company −0.046 (−0.54) −0.122 (−2.3) −0.193 (−2.08)Self-employed −0.069 (−0.59) −0.051 (−0.39) 0.317 (2.46)Privately owned business −0.182 (−1.65) −0.231 (−1.96) −0.037 (−0.22)Retirees reemployed −0.302 (−3.39) −0.242 (−1.41) −0.177 (−1.32)Retirees −0.341 (−4.2) −0.452 (−2.37) −0.359 (−3.42)Other −0.124 (−1.13) −0.187 (−1.24) −0.338 (−1.2)

Occupation (default is retiree)Engineer and technician −0.015 (−0.14) −0.141 (−0.69) −0.036 (−0.29)Officers −0.044 (−0.43) −0.063 (−0.31) −0.045 (−0.36)Staff in commerce 0.012 (0.12) −0.036 (−0.17) 0.029 (0.24)Staff in services 0.437 (3.08) 0.019 (0.09) −0.011 (−0.08)Worker in manufacturing 0.118 (0.82) 0.025 (0.12) 0.091 (0.56)Worker in transport and 0.209 (2.02) −0.018 (−0.09) 0.130 (1.03)

telecommunicationsOther 0.171 (1.33) −0.069 (−0.27) −0.636 (−4.2)

Constant 0.172 (0.7) −0.623 (−1.68) −0.197 (−0.71)R-square 0.401 0.290 0.359

Note: Numbers in the parentheses are t-statistics.Source: Authors’ computations based on data from China National Bureau of Statistics 1999 Rural Household Survey and1999 Urban Household Survey.

In urban areas the gains tend to be higher forsmaller households (except in Guangdong). As inrural areas there is a U-shaped pattern (except forLiaoning), with lowest gains at 66 years of age inGuangdong and 51 in Sichuan. While there is nopattern in the relationship between education andwelfare gains in rural areas, the gains in urban areastend to be larger for less educated households.However, this may be biased by the fact that educa-tion was used in identifying skilled labor (notingthat unskilled nonfarm wages are predicted toincrease relative to skilled labor; see table 15.2).There are some signs of sectoral effects, thoughonly significantly so in Liaoning, with higher gainsfor those with government jobs. Retirees tend tohave lower gains than others.

Conclusion

In the aggregate, the analysis finds that China’strade reforms have had only a small impact on

mean household income, inequality, and povertyincidence. There is, however, a sizable (and at leastpartly explicable) variance in impacts across house-hold characteristics. Rural families tend to lose;urban households tend to gain. Some provincesexperience larger impacts than others; highestimpacts occur in the northeast region of InnerMongolia, Liaoning, Jilin, and Heilongjiang, whererural households are more dependent on feed grainproduction (for which falling prices are expectedfrom WTO accession) than elsewhere in China.

Within rural or urban areas of a given province,the gains from trade reform vary with observablehousehold characteristics. The most vulnerablehouseholds tend to be in rural areas, dependent on agriculture, with relatively fewer workers andwith weak economic links to the outside economythrough migration. There are also some strong geo-graphic concentrations of adverse impacts. Forexample, agricultural households in some countiesincur welfare losses of 3–5 percent of their incomes.


TABLE 15.8 Average Impacts for Agricultural Households in Selected Counties, 2001–07

CountyGain Provincial Mean

Province Identifier Yuan Percent Yuan Percent

Liaoning 210181 −73.72 −3.07 −32.34 −1.29210212 −145.40 −2.99210381 −172.01 −5.57210921 −57.70 −5.21211321 −45.58 −3.78211322 −53.60 −3.23

Guangdong 440111 −107.31 −2.74 −29.34 −0.81440126 −183.63 −2.64440223 −102.33 −3.53440523 −148.90 −2.55440620 −227.23 −3.11440621 −109.59 −2.64441425 −316.49 −5.34

Sichuan 510121 −130.46 −2.86 −12.31 −0.67510125 −63.19 −3.81512425 −138.34 −5.71512610 −52.23 −3.11512825 −40.44 −2.80513021 −93.02 −4.07

Note: A negative sign means a net loss. Agricultural household means that more than 75 percent of incomeis from agriculture.Source: Authors’ computations based on data from China National Bureau of Statistics 1999 Rural HouseholdSurvey and 1999 Urban Household Survey.

Naturally, the approach taken here has limita-tions. For example, there may well be dynamicgains from greater trade openness that are notbeing captured by the model used to generate therelative price impacts. Trade may facilitate learningabout new technologies and innovation, bringinglonger term gains in productivity. Trade reformmay also come with (and possibly help induce)other policy reforms, such as in factor markets. Theapproach here has attempted to capture only thestatic welfare effects of WTO accession.

A further limitation was the need to make lin-ear approximations in the neighborhood of an ini-tial optimum for each household. In other appli-cations this could be deceptive if price or wagechanges are large or if the household were initiallyout of equilibrium, due to rationing (including

involuntary unemployment), for example. In prin-ciple, there are ways of dealing with these prob-lems by estimating complete demand and supplysystems that allow for rationing. This may prove afruitful avenue for future research, though itshould be noted that these methods generate theirown problems, such as those arising from incom-plete data on price and wage levels at householdlevel.

Despite these limitations, the type of approachfollowed here can usefully illuminate the range ofwelfare impacts to be expected from economywidereforms. By avoiding unnecessary aggregation ofthe primary household-level data, these relativelysimple tools can also offer insights into the sorts ofpolicy responses that might be called for to com-pensate losers from reform.


Appendix


TABLE 15.A.1 Rural Gains and Losses by Province, 2001–07

Number of Original Post-WTO Share ofSampled Number of Income Income Gain or Change Losers

Province Households Gainers (Yuan) (Yuan) Loss (Yuan) (Percent) (Percent)

Beijing 750 381 4,221.05 4,210.08 −10.96 −0.26 49.20Tianjin 595 219 3,401.71 3,380.48 −21.22 −0.62 63.19Hebei 4,200 1,310 2,441.50 2,426.82 −14.68 −0.60 68.81Shanxi 2,100 926 1,772.62 1,765.13 −7.49 −0.42 55.90Inner Mongolia 2,198 206 2,055.49 2,011.26 −44.22 −2.15 90.63Liaoning 1,886 353 2,501.98 2,469.64 −32.34 −1.29 81.28Jilin 1,598 132 2,260.12 2,210.46 −49.66 −2.20 91.74Heilongjiang 1,997 115 2,166.59 2,114.18 −52.41 −2.42 94.24Shanghai 600 416 5,409.11 5,428.79 19.68 0.36 30.67Jiangsu 3,400 1,209 3,495.20 3,486.78 −8.42 −0.24 64.44Zhejiang 2,693 1,148 3,946.44 3,934.92 −11.52 −0.29 57.37Anhui 3,095 676 1,900.76 1,885.79 −14.97 −0.79 78.16Fujian 1,750 469 3,091.39 3,071.40 −19.99 −0.65 73.20Jiangxi 2,450 553 2,129.45 2,117.26 −12.19 −0.57 77.43Shandong 4,200 822 2,520.76 2,494.89 −25.87 −1.03 80.43Henan 4,200 828 1,948.36 1,931.70 −16.66 −0.86 80.29Hubei 3,188 755 2,212.71 2,200.04 −12.68 −0.57 76.32Hunan 3,700 1,181 2,102.98 2,095.39 −7.60 −0.36 68.08Guangdong 2,560 514 3,628.95 3,599.61 −29.34 −0.81 79.92Guangxi 2,310 309 2,048.33 2,025.75 −22.58 −1.10 86.62Hainan 718 28 2,086.40 2,057.85 −28.55 −1.37 96.10Chongqing 1,500 404 1,736.63 1,730.20 −6.43 −0.37 73.07Sichuan 3,998 879 1,843.23 1,830.92 −12.31 −0.67 78.01Guizhou 2,240 417 1,363.07 1,354.03 −9.04 −0.66 81.38Yunnan 2,397 399 1,438.34 1,421.34 −17.00 −1.18 83.35Tibet 480 143 1,309.46 1,307.41 −2.05 −0.16 70.21Shaanxi 2,217 446 1,456.48 1,442.09 −14.39 −0.99 79.88Gansu 1,800 479 1,357.28 1,350.34 −6.95 −0.51 73.39Qinghai 600 135 1,466.67 1,452.61 −14.06 −0.96 77.50Ningxia 600 108 1,754.15 1,729.05 −25.11 −1.43 82.00Xinjiang 1,495 312 1,471.11 1,447.57 −23.55 −1.60 79.13Rural total 67,515 16,272 2,257.15 2,239.08 −18.07 −0.80 75.90

Source: Authors’ computations based on data from Ianchovichina and Martin 2002 and China NationalBureau of Statistics 1999 Rural Household Survey and 1999 Urban Household Survey.The ordering of provinces is the traditional administrative ordering as used in (for example) NBS 2000.


TABLE 15.A.2 Urban Gains and Losses by Province, 2001−07

Number of Original Post-WTO Share ofSampled Number of Income Income Gain or Change Losers

Province Households Gainers (Yuan) (Yuan) Loss (Yuan) (Percent) (Percent)

Beijing 500 430 9,388.88 9,431.72 42.84 0.46 14.00Tianjin 500 451 7,323.57 7,358.47 34.91 0.48 9.80Hebei 650 591 5,673.46 5,702.35 28.89 0.51 9.08Shanxi 650 598 4,519.20 4,549.94 30.74 0.68 8.00Inner Mongolia 550 495 4,491.87 4,516.19 24.32 0.54 10.00Liaoning 1000 916 5,257.42 5,285.65 28.23 0.54 8.40Jilin 700 610 4,630.13 4,650.46 20.33 0.44 12.86Heilongjiang 1000 887 4,798.92 4,820.50 21.58 0.45 11.30Shanghai 500 458 10,927.18 10,984.16 56.98 0.52 8.40Jiangsu 800 723 6,933.07 6,968.78 35.71 0.51 9.63Zhejiang 550 498 9,044.40 9,098.28 53.87 0.60 9.45Anhui 500 458 5,159.46 5,190.37 30.91 0.60 8.40Fujian 550 516 7,521.52 7,569.70 48.18 0.64 6.18Jiangxi 550 498 4,762.78 4,783.38 20.60 0.43 9.45Shandong 650 602 5,689.90 5,720.69 30.78 0.54 7.38Henan 600 565 4,689.43 4,717.89 28.46 0.61 5.83Hubei 750 619 5,743.18 5,765.29 22.11 0.38 17.47Hunan 700 612 5,727.42 5,750.43 23.00 0.40 12.57Guangdong 600 490 10,871.06 10,903.85 32.79 0.30 18.33Guangxi 600 496 6,011.10 6,033.40 22.30 0.37 17.33Hainan 200 172 5,766.33 5,787.64 21.31 0.37 14.00Chongqing 300 239 5,910.18 5,931.90 21.72 0.37 20.33Sichuan 800 691 5,610.29 5,634.60 24.30 0.43 13.63Guizhou 450 383 5,324.43 5,347.71 23.27 0.44 14.89Yunnan 650 566 5,939.69 5,973.23 33.54 0.56 12.92Tibet n.a.Shaanxi 500 427 4,768.99 4,788.25 19.26 0.40 14.60Gansu 400 372 4,610.86 4,641.27 30.41 0.66 7.00Qinghai 250 240 3,759.53 3,788.65 29.12 0.77 4.00Ningxia 200 177 4,472.43 4,493.27 20.84 0.47 11.50Xinjiang 250 214 5,277.25 5,295.94 18.69 0.35 14.40Urban total 16,900 14,994 6,046.13 6,075.60 29.45 0.49 11.28

Source: Authors’ computations based on data from Ianchovichina and Martin 2002 and China NationalBureau of Statistics 1999 Rural Household Survey and 1999 Urban Household Survey.The ordering of provinces is the traditional administrative ordering as used in (for example) NBS 2000.

Notes

1. Jalan and Ravallion (1998) report evidence of such churn-ing using panel data for rural China. Baulch and Hoddinott(2000) review evidence for a number of countries.

2. For China’s lagging poor areas see Jalan and Ravallion(2002).

3. Examples of partial equilibrium analysis of the welfaredistributional effects of price changes include King (1983),Deaton (1989), Ravallion and van de Walle (1991), andFriedman and Levinsohn (2002). On applications to tax policyreform, see Newbery and Stern (1987). On CGE models see

Decaluwe and Martens (1988) and Hertel (1997). For a usefuloverview of alternative approaches to assessing the welfareimpacts of trade policies and examples from the literature, seeMcCulloch, Winters, and Cirera (2001).

4. The only known example of this full integration isCockburn (2002), who built a classic trade-focused CGE modelonto the Nepal Living Standards Survey covering about 3,000households.

5. In an antecedent to the approach taken here, Bourguignon,Robilliard, and Robinson (2003) also take price changes gener-ated by a CGE model to survey data (for Indonesia). Method-ologically, the main difference is that Bourguignon, Robilliard,

and Robinson generate income impacts at the household levelfrom a microeconometric model of income determination,whereas this study derives first-order welfare impacts analyticallyfrom a standard competitive farm-household model.

6. Hertel (1997) contains descriptions of the standard GTAPmodel with applications.

7. A full discussion of the assumptions of the general equilib-rium model and the results of its application to China’s accessionto the WTO can be found in Ianchovichina and Martin (2002).

8. For food items, unit values can be calculated (expendituredivided by quantity) from the survey data, but there is no suchoption for food inputs to production, nonfood commoditiesconsumed or used in production, or wages (the survey data donot include labor supplies or quantities consumed of nonfoodgoods, including production inputs).

9. Examples of this approach can be found in King (1983)and Ravallion and van de Walle (1991).

10. The full sample of the urban survey was about 40,000households, but until 2002 the central NBS office kept individ-ual record data for only 16,900 households.

11. For further discussion in the context of the Rural House-hold Survey, see Chen and Ravallion (1996).

12. By the International Labor Organization’s definitions,skilled labor consists of managers and administrators, profes-sionals, and para-professionals, while unskilled labor consists oftradespeople, clerks, salespeople and personal service workers,plant and machine operators and drivers, laborers, and relatedworkers and farm workers.

References


Barro, Robert. 2000.“Inequality and Growth in a Panel of Coun-tries.” Journal of Economic Growth 5: 5–32.

Baulch, Bob, and John Hoddinott. 2000. “Economic Mobilityand Poverty Dynamics in Developing Countries.” Journal ofDevelopment Studies 36(6): 1–24.

Bourguignon, Francois, and C. Morisson. 1990. “IncomeDistribution, Development and Foreign Trade.” EuropeanEconomic Review 34: 1113–32.

Bourguignon, Francois, Anne-Sophie Robilliard, and ShermanRobinson. 2003. “Representative versus Real Households inthe Macro-Economic Modeling of Inequality.” WorkingPaper 2003-05. DELTA, Paris.

Chen, Shaohua, and Martin Ravallion. 1996.“Data in Transition:Assessing Rural Living Standards in Southern China.” ChinaEconomic Review 7: 23–56.

———. 2001. “How Did the World’s Poor Fare in the 1990s?”Review of Income and Wealth 47(3): 283–300.

———. 2003. “Household Welfare Impacts of China’s Accessionto the World Trade Organization,” Policy Research WorkingPaper 3040, World Bank, Washington, D.C.

Cockburn, John. 2002. “Trade Liberalization and Poverty inNepal: A Computable General Equilibrium Micro Simula-tion Analysis.” Processed. Quebec: University of Laval.

Deaton, Angus. 1989. “Rice Prices and Income Distribution inThailand: A Non-Parametric Analysis.” Economic Journal 99:1–37.

Decaluwe, B., and A. Martens. 1988. “CGE Modeling and Devel-oping Economies: A Concise Empirical Survey of 73 Appli-cations to 26 Countries.” Journal of Policy Modeling 10(4):529–68.

Dollar, David, and Aart Kraay. 2002. “Growth is Good for thePoor.” Journal of Economic Growth 7(3): 195–225.

Edwards, Sebastian. 1997. “Trade Policy, Growth and IncomeDistribution.” American Economic Review 87(2): 205–10.

Friedman, Jed, and James Levinsohn. 2002. “The DistributionalImpacts of Indonesia’s Financial Crisis on HouseholdWelfare: A ‘Rapid Response’ Methodology.” World BankEconomic Review 16(3): 397–424.

Hertel, T., ed. 1997. Global Trade Analysis: Modeling andApplications. Cambridge: Cambridge University Press.[www.gtap. org.]

Ianchovichina, Elena, and Will Martin. 2002.“Economic Impactsof China’s Accession to the WTO.” Development ResearchGroup, World Bank, Washington, D.C.

Jalan, Jyotsna, and Martin Ravallion. 1998. “Transient Poverty inPost-Reform Rural China.” Journal of Comparative Econo-mics 26: 338–57.

———. 2002. “Geographic Poverty Traps? A Micro Model ofConsumption Growth in Rural China.” Journal of AppliedEconometrics 17(4): 329–46.

King, Mervyn A. 1983. “Welfare Analysis of Tax Reforms UsingHousehold Level Data.” Journal of Public Economics 21:183–214.

Lundberg, Mattias, and Lyn Squire. 2003. “The SimultaneousEvolution of Growth and Inequality.” Economic Journal 113:326–44.

McCulloch, Neil, L. Alan Winters, and Xavier Cirera. 2001. TradeLiberalization and Poverty: A Handbook. London: Center forEconomic Policy Research and Department for Interna-tional Development.

NBS (National Bureau of Statistics). 2000. China Statistical Year-book. Beijing: China Statistics Press.

Newbery, David, and Nicholas Stern, eds., 1987. The Theory of Taxation for Developing Countries. Oxford: OxfordUniversity Press.

Ravallion, Martin. 1990. “Rural Welfare Effects of Food PriceChanges with Induced Wage Responses: Theory andEvidence for Bangladesh.” Oxford Economic Papers 42:574–85.

———. 2001. “Growth, Inequality and Poverty: LookingBeyond Averages.” World Development 29(11): 1803–15.

Ravallion, Martin, and Dominique van de Walle. 1991. “TheImpact of Food Pricing Reforms on Poverty: A DemandConsistent Welfare Analysis for Indonesia.” Journal of PolicyModeling 13: 281–300.


likely to face substantial adjustments, becausetariffs remain high and the scale of production isfar below international standards to date (François,2002). In addition, the imports of land-intensiveagricultural products, including corn and soybeans,are expected to surge under the accession agree-ment. The services sector is also expected to facesubstantial expansion and restructuring. Such abroad structural adjustment is likely to have aneffect on the distribution of income in China,potentially depressing rural incomes in some areasrelative to urban incomes. Similarly, it is anticipatedthat coastal regions might benefit at the expense ofthe interior. Yet the rural–urban, interior–coastalincome disparities in China are already at recordlevels (Yang and Zhou 1999).

Kanbur and Zhang (2001) track regional inequal-ity over the past 50 years and find that there arethree peaks over this five-decade period: the GreatFamine of the 1950s, the Cultural Revolution of thelate 1960s and early 1970s, and the period of open-ness and global integration of the 1990s. They alsofind that the interior–coastal disparities have espe-cially contributed to the recent growth in inequal-ity as the Chinese economy has opened up. Chenand Wang (2001) attempt to discern the separateimpacts of growth and inequality on poverty inChina since 1990. They find that the increase in

After a decade and half of negotiations, the WorldTrade Organization (WTO) successfully concludednegotiations on China’s terms of entry to the WTO atthe end of 2001. China’s accession is expected to lenda further boost to trade and foreign investment.China’s foreign trade quadrupled from 1990 to2000—growing much faster than world trade, whichonly increased 87 percent during the same period.By the year 2001, the volume of China’s merchandiseexports and imports had reached US$262 billion1

and US$244 billion, respectively, ranking it sixth inworld trade. China is also the largest recipient offoreign direct investment (FDI) among developingcountries; over US$320 billion flowed into its econ-omy during the last decade. Most studies predict thatWTO accession will bring with it added benefits,both for China as well as for most of its trading part-ners (Ianchovichina, McDougall, and Hertel 1999;McKibbin and Tang, 2000; Wang 2001).

At the sectoral level, however, the effects ofaccession are likely to be quite different acrosseconomies. China’s trade surplus in apparel prod-ucts is expected to increase even more with elimi-nation of the Multi-Fiber Arrangement (MFA)quotas under the Uruguay Round’s Agreementon Textiles and Clothing, or ATC (DevelopmentResearch Center 1998; Wang 1999; Walmsley andHertel 2000). By contrast, the automobile sector is

283

16

Implications of WTOAccession for

Poverty in China

Thomas W. Hertel, Fan Zhai, and Zhi Wang

inequality has contributed to greater poverty overthe past decade. However, the beneficial impact ofoverall economic growth has dominated, and theabsolute number of people in poverty (by severaldifferent measures) has been reduced.

For China’s WTO accession, the same two forcesare at work. On the one hand, further trade liberal-ization will accelerate the country’s integration intothe world economy and enhance economic effi-ciency (Yang and Huang 1997; Wang and Zhai1998). On the other hand, the experience of the lastdecade suggests that such liberalization might alsoincrease inequality (Kanbur and Zhang 2001). Thegoal of this chapter is to shed more light on thisempirical question. Our approach is model-based.We use an applied general equilibrium modeldesigned to capture the differential effects of acces-sion on each of China’s economic sectors. Specialattention is paid to the disaggregation of house-holds in order to permit us to say something aboutthe likely impacts of accession on income distribu-tion. In addition, we pay quite a bit of attention tothe modeling of China’s rural labor markets, whichwe see as the primary vehicle for poverty reductionin the wake of WTO accession.

This chapter is organized as follows. The nextsection provides an empirical overview of theChinese economy—with a special emphasis ontrade and poverty. This overview sets the stage forthe computable general equilibrium (CGE) modeldescribed in the section that follows. We then turnto a discussion of the baseline scenario, againstwhich China’s accession will be evaluated, alongwith a description of the accession scenario itself.The next section uses simulation results to assessthe impact of China’s WTO accession on poverty,and the final section offers some conclusions andsuggestions for future research.

Overview of the Chinese Economy,Poverty, and Trade

This section provides an overview of the Chineseeconomy. We start by looking at some stylized factsof household income and poverty and examiningpatterns of production and trade.

Households and Poverty

To facilitate analysis of the impacts of WTO acces-sion on poverty, we obtained from China’s National

Bureau of Statistics aggregations for 2000 of the ruraland urban household surveys for three provinces:Guangdong, Sichuan, and Liaoning. Together, theseprovinces are fairly representative of the diversitywithin China as a whole. Guangdong represents therelatively wealthy coastal region, which is heavilyexport-oriented; Sichuan represents the populous,relatively poor inland region in which agricultureplays a more important role in the economy; andLiaoning has a typical “old industrial base,” which isheavily urban and highly dependent on state-ownedenterprises. The latter are mainly concentrated inhighly protected, heavy industrial sectors.

Within these surveys, we have aggregated house-holds into several groups or “strata,” according totheir primary source of income. Recent analysis of trade and poverty by Hertel and others (2004)suggests the merit of distinguishing those house-holds that are specialized—that is, they receive 95 percent or more of their income from one source:transfer payments, labor wages and salaries, orself-employment income. Because of the limitednumber of specialized, nonfarm households in thesurvey, as well as the absence of transfer- and labor-specialized households in the rural surveys, we endup with two strata in the rural survey—agriculture-specialized and diversified (all other)—and threestrata in the urban survey—transfer-specialized,labor-specialized, and diversified. Within eachstratum, we order households from poorest to rich-est, based on per capita income, and then groupthem into 20 vingtiles, each containing 5 percent of the stratum population. We then aggregate thethree rural surveys, and the three urban surveys,before incorporating them into the national socialaccounting matrix. In this way, we obtain a dis-aggregated representation of households in thenational model, wherein the earnings and spendingprofiles reflect the diversity obtained from the threeprovincial surveys.

Table 16.1 reports per capita income by stratum,location, and vingtile for the 100 representativehouseholds in the model. Using the $1/day level of poverty, Chen and Wang (2001) estimate that24.9 percent of the rural population in China and 0.5 percent of the urban population are inpoverty. Using these figures as a guide, in table 16.1we place the rural poverty line in the fifth vingtileof the rural population and the urban poverty line in first vingtile of the urban population. The


average income in each vingtile is much higher forthe urban population, and even after adjusting for differences in cost of living, a far greater per-centage of the rural population falls below thepoverty line.

This preliminary analysis reveals that any policythat depresses incomes in agriculture is very likelyto lead to a short-run increase in poverty. If oppor-tunities arise elsewhere in the rural economy, how-ever, it is possible that, after suitable adjustment,many of these rural households could be madebetter off. Such an adjustment will likely be mostdifficult for those households that currently do notparticipate in the nonfarm economy, either asworkers or in a self-employed capacity. A criticaldeterminant of farm households’ ability to partici-pate in the nonfarm economy will be their level of educational attainment. In general, the higher-income vingtiles in both rural and urban areas havea higher level of educational attainment. Within the

rural population, the average educational attainmentof agriculture-specified households is lower thanthat of other households. There is also a significantgap in educational attainment between rural andurban households. Thus the most vulnerable farmhouseholds are likely to be those that are trappedin agricultural activity by virtue of their lack ofeducation.

The impact of WTO accession on rural non-farm households and especially urban householdsdepends importantly on what happens to the man-ufacturing and services sectors. Here, we expect amixed outcome, with profitability and employmentin light manufactures rising, but the heavy industryand automobile sectors will be adversely affected.Households dependent on transfer payments arealso potentially vulnerable after WTO accession,because reduced tariffs may bring additional pres-sure to bear on state-run enterprises and on gov-ernment spending—and thus transfer payments.

Implications of WTO Accession for Poverty in China 285

TABLE 16.1 Per Capita Income by Location, Stratum, and Vingtile (1997 yuan)

Vingtile Rural Households Urban Households

(Poorest � 1) Agric. Diverse Total Transfer Labor Diverse Total

1 845 889 874 2,903 2,135 2,454 2,3512 1,049 998 1,006 3,995 3,151 3,054 3,2123 1,156 1,162 1,161 4,674 3,790 3,703 3,8274 1,303 1,301 1,301 5,273 3,987 4,250 4,2165 1,433 1,432 1,432 5,595 4,513 4,452 4,6086 1,755 1,551 1,568 6,280 4,763 4,528 4,7697 1,675 1,679 1,678 6,594 5,237 4,884 5,1558 1,822 1,811 1,812 7,794 5,692 5,370 5,5889 1,947 1,944 1,944 8,643 6,096 5,786 6,045

10 2,099 2,095 2,096 8,142 6,694 6,334 6,56411 2,240 2,252 2,251 8,220 6,866 6,482 6,71812 2,415 2,411 2,411 8,946 7,420 6,901 7,19213 2,602 2,595 2,595 10,807 7,686 7,532 7,67114 2,835 2,818 2,819 12,973 8,432 7,974 8,30315 3,031 3,069 3,066 10,601 9,120 8,526 8,79916 3,344 3,353 3,352 12,925 9,709 8,727 9,13017 3,708 3,717 3,717 — 11,152 9,659 10,24018 4,306 4,258 4,261 18,821 12,749 10,985 11,79619 5,171 5,162 5,163 15,190 15,134 13,403 14,12520 9,712 8,345 8,485 — 21,997 19,659 20,522

Share of 7.35 62.73 70.08 1.44 12.01 16.47 29.92population (%)

Not available.Source: National Bureau of Statistics Rural and Urban Household Surveys 2000.

Finally, by virtue of the way they have been defined,diversified households are likely to be little affectedby trade liberalization. Decreases in one source ofincome will tend to be offset by increases elsewhere.

The fate of the labor-specialized households willdepend on what happens to wages after accession.If educational attainment is used as a proxy for skilllevel, then the poor are most heavily reliant on un-skilled labor. Therefore, from a poverty perspective,the key factor for labor-specialized householdsis what happens to wages for the unskilled afterWTO accession. What are the likely factor marketeffects of China’s WTO accession? These effects willdepend on the pattern of production, trade, andprotection in China, or the subject to which wenext turn.

China’s Production, Trade, and Protection

Having provided an overview of the Chinese econ-omy from the household perspective, we now offera perspective from the production point of view.table 16.2 has been assembled using the 1997Chinese social accounting matrix developed by theDevelopment Research Center (DRC) at the StateCouncil. We used the most recent input–outputtable. One of the most striking points abouttable 16.2 is the very high share of employment inagriculture. This sector accounted for more than55 percent of total employment in the Chineseeconomy in 1997, yet its contribution to economy-wide value-added (first column) is only about22 percent. This finding is consistent with thelow income levels observed in table 16.1 foragriculture-specialized households. Because thissector also relies almost entirely on unskilled labor,a reduction in the size of this sector will likelyrelease additional unskilled labor into the rural—and possibly urban—labor markets, therebydepressing wages.

Table 16.2 also reveals a relatively underdevel-oped services sector. For example, when comparedwith the 65 other regions in the Global Trade Analy-sis Project (GTAP) Version 5 database (Dimarananand McDougall 2002), China ranks last in the shareof economy-wide value-added generated in theservices sector, and this share is scarcely half theworld average: 37 percent versus 72 percent. Theshare of services in total employment in China is even smaller, amounting to less than 27 percentin 1997. This finding indicates the potential that

may be realized by means of WTO accession as the rules governing foreign investment in theservices sector are liberalized. In contrast to serv-ices, manufacturing’s share of value-added is quitesubstantial—33 percent versus a global average of19 percent based on GTAP data—placing Chinasecond, behind Thailand. Textiles, building materi-als, and chemicals lead the way, followed by sectorsrelated to metals, machinery, electronics, and asso-ciated products.

An important factor in the growth of manufac-turing activity in the past decade has been the rapidexpansion of exports. Textiles, apparel, and leatherproducts now account for more than one-quarter oftotal Chinese exports, and electronics account foranother 11.5 percent. The export dependency ofthese sectors, as well as for crafts and toys andinstruments, is high, with more than 30 percent oftheir output directed to foreign markets. (Textilesare important intermediate inputs, and so theexport dependency ratio is somewhat less—about18 percent.) Much of this export growth has beenfueled by the export processing or export promo-tion regime introduced in 1986 (Naughton 1996).The enterprises operated under this regime haveattracted a great deal of foreign investment as aresult of special preferences—most important, theexemption from the tariffs on imported intermedi-ate inputs to be used in the production of exports.

By 1997 the export processing regime hadgrown to account for 50 percent of China’s totalexports and 46 percent of its total imports. Nearly80 percent of China’s primary iron and steel,electric machinery, electronics, and instrumentexports, and more than two-thirds of its leather,toys, and chemical fiber exports were “processingexports”—that is, assembling or transformationof imported intermediate goods and re-export.Export processing also constitutes a significantportion of China’s exports of apparel, paper, andprinting products. The high shares of processingexports in these sectors require a large volume ofraw materials, components, and semi-processedproducts imported from abroad. Column 8 intable 16.2 shows that in textiles, apparel, andleather, ordinary imports constituted only 1–2 per-cent of total imports in 1997, and so almost allimports were used for the production of process-ing exports. Most imports of paper products, build-ing materials, chemicals, basic metal and metalproducts, machinery, and electronics were also



TAB

LE 1

6.2

Eco

no

mic

Str

uct

ure

an

d M

arket

Op

enn

ess

in C

hin

a, 1

997

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

Ord

inar

yO

rdin

ary

Co

llect

edO

rdin

ary

Tota

lEx

po

rts/

Imp

ort

s/N

etA

sPo

st-

Val

ue-

Imp

ort

s/Ex

po

rts/

Tota

lTo

tal

Exp

ort

sN

om

inal

Co

llect

edN

om

inal

acce

ssio

nA

dd

edEm

plo

ymen

tIm

po

rts

Exp

ort

sD

om

esti

cO

utp

uts

Exp

ort

sIm

po

rts

(Bill

ion

s o

fTa

riff

Tari

ffPe

rcen

tTa

riff

(%)

(%)

(%)

(%)

Use

(%

)(%

)(%

)(%

)Yu

an)

Rat

e (%

)R

ate

(%)

(%)

Rat

e (%

)

Rice

2.32

7.5

0.2

0.1

0.88

0.9

100

920.

01.

00.

440

.00.

5W

heat

1.27

4.1

0.9

0.0

6.93

0.0

—78

�11

.31.

00.

220

.00.

2C

orn

0.82

2.6

0.1

0.4

0.32

7.9

100

216.

31.

00.

00.

00.

0C

otto

n0.

622.

00.

60.

02.

960.

010

017

�7.

73.

50.

617

.11.

0O

ther

non

grai

n cr

ops

7.35

23.8

0.5

1.1

0.22

2.3

100

2512

.85.

75.

710

0.0

4.1

Fore

stry

0.90

2.0

0.5

0.2

3.81

3.6

9633

�2.

628

.68.

329

.011

.5W

ool

0.03

0.0

0.3

0.0

29.6

58.

910

06

�3.

815

.00.

74.

71.

0O

ther

live

stoc

k 5.

608.

20.

10.

40.

070.

997

476.

85.

02.

142

.03.

4Fi

shin

g2.

073.

20.

00.

20.

191.

292

772.

50.

80.

562

.50.

2O

ther

agr

icul

ture

0.95

2.2

0.0

0.1

0.00

1.2

8812

1.5

16.0

1.7

10.6

9.6

Coa

l min

ing

1.68

1.2

0.1

0.6

0.26

3.1

100

638.

66.

03.

355

.05.

3C

rude

oil

and

natu

ral g

as1.

670.

23.

61.

519

.07

14.4

100

71�

20.6

1.5

1.0

66.7

0.0

Ferr

ous

ore

min

ing

0.14

0.1

1.1

0.0

16.3

10.

010

043

�14

.40.

00.

0—

0.0

Non

ferr

ous

ore

min

ing

0.42

0.2

0.5

0.1

3.06

1.1

9237

�4.

90.

00.

0—

0.0

Oth

er m

inin

g1.

000.

80.

80.

62.

944.

379

36�

0.5

2.9

0.9

31.0

1.2

Vege

tabl

e oi

l0.

280.

11.

20.

38.

414.

319

46�

10.3

17.0

5.3

31.2

9.0

Gra

in m

ill a

nd fo

rage

0.78

0.2

1.1

0.3

4.08

1.6

7694

�7.

94.

74.

493

.61.

9Su

gar

0.05

0.1

0.2

0.1

3.07

3.3

1236

�0.

830

.09.

531

.715

.0Pr

oces

sed

food

1.72

0.6

1.0

3.3

0.74

9.7

7218

42.4

23.2

3.7

15.9

10.4

Beve

rage

s0.

760.

30.

10.

50.

323.

184

457.

560

.224

.039

.913

.7To

bacc

o0.

280.

10.

20.

31.

433.

296

243.

349

.110

.621

.611

.0Te

xtile

s3.

251.

56.

811

.40.

4118

.460

110

4.5

27.5

0.2

0.7

9.9

Ap

par

el1.

880.

60.

79.

90.

1737

.045

215

6.2

41.8

0.7

1.7

14.3

Leat

her

0.64

0.3

1.8

5.0

0.35

32.6

291

61.4

35.5

0.3

0.8

9.1

Saw

mill

s an

d fu

rnitu

re0.

800.

40.

92.

12.

4713

.161

2022

.814

.42.

517

.43.

7Pa

per

and

prin

ting

1.45

0.6

2.9

0.5

4.88

2.0

3533

�29

.211

.03.

128

.22.

5C

raft

s an

d to

ys0.

880.

31.

06.

03.

8640

.032

2887

.63.

11.

032

.30.

8Pe

trol

eum

ref

inin

g0.

610.

23.

11.

211

.08

5.7

6962

�19

.68.

74.

855

.25.

1C

hem

ical

s2.

521.

011

.84.

28.

488.

380

31�

77.8

10.8

3.0

27.8

4.7

Phar

mac

eutic

als

0.74

0.2

0.2

0.7

1.53

5.9

8475

8.5

10.9

7.2

66.1

4.0

Che

mic

al fi

bers

0.33

0.1

2.1

0.5

5.78

6.6

317

�17

.415

.51.

06.

53.

8


TAB

LE 1

6.2

(Co

nti

nu

ed)

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

Ord

inar

yO

rdin

ary

Co

llect

edO

rdin

ary

Tota

lEx

po

rts/

Imp

ort

s/N

etas

Post

-V

alue

-Im

po

rts/

Exp

ort

s/To

tal

Tota

lEx

po

rts

No

min

alC

olle

cted

No

min

alac

cess

ion

Ad

ded

Emp

loym

ent

Imp

ort

sEx

po

rts

Do

mes

tic

Out

put

sEx

po

rts

Imp

ort

s(B

illio

ns

of

Tari

ffTa

riff

Perc

ent

Tari

ff(%

)(%

)(%

)(%

)U

se (

%)

(%)

(%)

(%)

Yuan

)R

ate

(%)

Rat

e (%

)(%

)R

ate

(%)

Rubb

er a

nd p

last

ics

1.24

0.6

2.1

4.2

1.49

15.7

2812

42.8

19.8

2.0

10.1

7.2

Build

ing

mat

eria

ls3.

542.

10.

82.

10.

603.

482

2023

.920

.83.

617

.39.

3Pr

imar

y iro

n an

d st

eel

1.36

0.7

3.8

1.8

4.12

5.5

2128

�17

.58.

12.

024

.74.

2N

onfe

rrou

s m

etal

s0.

490.

32.

61.

26.

638.

056

14�

13.1

7.1

0.9

12.7

2.9

Met

al p

rodu

cts

1.48

0.7

2.6

4.1

4.36

13.1

5720

36.4

13.1

2.4

18.3

5.6

Mac

hine

ry2.

361.

05.

61.

811

.16

5.9

6135

�41

.013

.74.

230

.76.

2Sp

ecia

l eq

uip

men

t1.

310.

68.

01.

123

.65

5.6

6321

�82

.914

.12.

618

.45.

0A

utom

obile

s1.

090.

41.

10.

44.

111.

963

73�

7.6

50.7

32.6

64.3

11.2

Oth

er t

rans

por

t eq

uip

men

t0.

960.

42.

61.

69.

439.

724

28�

5.8

5.6

1.3

23.2

3.3

Elec

tric

mac

hine

ry1.

560.

64.

05.

54.

6915

.921

2042

.217

.93.

117

.36.

7El

ectr

onic

s1.

630.

413

.211

.513

.27

36.3

920

25.2

11.8

2.1

17.8

1.6

Inst

rum

ents

0.36

0.2

2.7

2.7

24.5

149

.515

2012

.312

.52.

318

.44.

0O

ther

man

ufac

turin

g1.

330.

60.

40.

90.

168.

129

39.

638

.90.

92.

312

.0U

tiliti

es2.

450.

50.

00.

20.

000.

910

0—

3.8

0.0

0.0

—0.

0C

onst

ruct

ion

7.03

5.8

0.4

0.1

0.28

0.1

100

100

�2.

60.

00.

0—

0.0

Tran

spor

tatio

n3.

913.

10.

72.

81.

839.

410

010

038

.90.

00.

0—

0.0

Post

and

com

mun

icat

ion

1.61

0.3

0.2

0.7

1.27

5.7

100

100

8.7

0.0

0.1

—0.

1C

omm

erci

al s

ervi

ces

7.47

9.0

1.2

0.7

1.31

0.9

100

100

�4.

20.

00.

0—

0.0

Fina

nce

2.09

0.5

0.4

0.1

1.22

0.5

100

100

�2.

70.

00.

0—

0.0

Soci

al s

ervi

ces

5.17

1.6

2.9

4.5

5.17

10.1

100

100

38.8

0.0

0.0

—0.

0Ed

ucat

ion

and

heal

th4.

694.

00.

20.

30.

440.

710

010

01.

70.

01.

7—

1.7

Publ

ic a

dmin

istr

atio

n3.

061.

80.

20.

00.

450.

110

010

0�

1.5

0.0

0.0

—0.

0To

tal/

aver

age

100

100

100

100

3.59

7.7

5132

409.

511

.22.

722

.34.

0

— N

ot a

vaila

ble.

Not

es:O

rdin

ary

imp

orts

/dom

estic

use

and

tot

al e

xpor

ts/o

utp

ut a

re a

t do

mes

tic p

rice.

The

sec

tora

l sha

re o

f im

por

ts a

nd e

xpor

ts a

re a

t w

orld

pric

e. T

he im

por

ts o

f ric

e, w

heat

, cor

n,co

tton

, gra

in m

ill a

nd fo

rage

, and

veg

etab

le o

il ar

e th

e av

erag

e of

199

3–97

.Th

e ta

riff r

ates

are

cal

cula

ted

base

d on

the

wei

ght

of o

rdin

ary

imp

orts

for

1997

at

the

eigh

t-di

git

leve

l of t

he h

arm

oniz

ed s

yste

m. T

he a

vera

ge t

ariff

rat

e is

for

mer

chan

dise

onl

y.

Sour

ce:C

hine

se s

ocia

l acc

ount

ing

mat

rix, 1

997,

Dev

elop

men

t Re

sear

ch C

ente

r, St

ate

Cou

ncil.

used by foreign or joint venture companies as inter-mediate inputs to produce processing exports.

The dominance of duty-exempt imports is alsoevident when one compares the nominal tariffrate with the collected rate reported in table 16.2(smuggling also plays a role here). The nominaltariff for textile products is 27.5 percent, but thecollected tariff averages 0.2 percent. Similarly, dra-matic differences appear in the apparel and leathersectors as well as many of the other export-orientedactivities. The lowest nominal/actual ratios appearin the pharmaceutical and automobile industrieswhere the nominal rate is only about twice as largeas the actual rate. These sectors have relatively lowexport/output ratios.

Table 16.2 also reports the ratio of ordinaryimports to domestic use (column 5) as an indica-tion of overall import dependency. The sectors withthe largest dependency on ordinary imports areenergy and mining, machinery, special equipment,and electronics, all of which show import shares inexcess of 10 percent of China’s domestic use ofthese commodities.

An overall picture of China’s trade is offeredby the sector-specific net exports also reported intable 16.2 (value of exports minus value of importsin billions of yuan). By and large, these exportsreflect China’s comparative advantage. China is anet exporter of unskilled labor-intensive manu-factures and a net importer of skilled labor andcapital-intensive manufactures. The largest share ofthe trade surplus in China is attributable to appareland textiles. Within the agricultural sector, China isa net importer of grains, but it has a trade surplusin other agricultural products.

China’s nominal tariff structure is typical of thatof many developing countries in that it provideshigh protection for the manufacturing sector, espe-cially capital-intensive manufactures and final con-sumption goods. Automobile imports face one ofthe highest nominal tariff rates (50 percent), andthe actual collection rate is also quite high (33 per-cent). The tariff rates in other manufactures and inthe textile and apparel sector are also relativelyhigh, but their effects are limited because the shareof duty-paying imports (ordinary imports) is verysmall.

The final column in table 16.2 reports China’spostaccession tariff rates, by sector. The trade-weighted tariff rate will drop by more than 60 percent

in the six years after China’s WTO accession. Thesectoral distribution of tariff cuts is relatively even.For most sectors, the tariff cuts range from 50 per-cent to 80 percent. Some light manufacturingsectors and capital goods sectors that historicallyhave been highly protected—such as beverages,tobacco, automobiles, and electronics—will experi-ence substantial tariff cuts. The dominant roleplayed by the export processing sector will be veryimportant in determining the impact of these tariffcuts. In sectors such as textiles, where virtually allimports come in under the export processing pro-visions, the impact of further tariff cuts on the costof intermediate inputs will be negligible. Theimportation of consumer goods, as well as importedintermediates for domestic use, also will be affectedby the WTO accession tariff cuts.

To the extent that accession leads to a decline inagricultural, relative to nonagricultural, profitabil-ity, we expect to see a rise in rural poverty andincreased pressure on the rural population to mi-grate to urban areas. However, a definitive answer tothe question posed by this chapter—Will China’saccession to the WTO increase poverty—and if so,among which groups?—requires use of a formalmodel. We therefore turn next to a summary of thesingle-country general equilibrium model used inthis paper.

Description of the Model

The CGE model of China used in this study is thelatest in a long line of models developed and main-tained by the Development Research Center of theState Council in Beijing. Earlier versions have beenused to analyze the economy-wide implications of China’s accession to the WTO (DevelopmentResearch Center 1998), the consequences thataccession might have for urban unemployment(Zhai and Wang 2002), and the income distributionconsequences of trade and tax reform (Wang andZhai 1998). The CGE model has its intellectualroots in the group of single-country, applied gen-eral equilibrium models used over the past twodecades to analyze the impacts of trade policyreforms (Dervis, de Melo, and Robinson 1982; deMelo 1988; Shoven and Whalley 1992; de Melo andTarr 1992). The China model began as a prototypeCGE model developed for the Trade and Environ-ment Program of the Organisation for Economic


Co-operation and Development (OECD) Develop-ment Center (Beghin and others 1994). Since thattime, however, significant modifications have beenmade to capture the major features of the tradingand tax system of the Chinese economy (Wang andZhai 1998; Zhai and Li 2000), to add a full-fledgeddemographic module (Zhai and Wang 2002), andto introduce disaggregated households and animproved treatment of labor markets (this study).Here, we focus on the main features of the model,especially those relevant for assessing the conse-quences of China’s WTO accession for poverty.

Modeling Household Behavior

To come to grips with the poverty question, it iscritical that we disaggregate households to themaximum extent possible, subject to the limita-tions posed by survey sampling, computationalconstraints, and human capacity for analysis. It isparticularly important to disaggregate householdsalong those dimensions most likely to be affectedby WTO accession. Thus, for example, one wouldnot want to group together rural and urban house-holds,or agriculture-and-nonagriculture specializedhouseholds. Otherwise, one would be introducingartificial diversification, thereby blurring the possi-ble consequences of WTO accession. Accordingly,we have used the grouping of households in table16.1: 20 vingtiles � 2 rural and 3 urban strata � 40rural and 60 urban representative households. Atotal of 100 household groups are therefore used inthe model.

Each household is endowed with three types oflabor: unskilled, semiskilled, and skilled. These threetypes are distinguished by educational attainment:2

unskilled workers are illiterate or semi-literate,semiskilled workers have a middle or high schooleducation, and skilled workers have schooling be-yond high school.3 Households are also endowedwith profits from family-owned agricultural andnonagricultural enterprises, property income, andtransfers. Agricultural profits represent returns tofamily labor, land, and capital. However, becausethe land market is not well developed in China, thereturns to land are effectively part of the return tofamily labor (Yang 1997). If a household migratesfrom the rural area and ceases to farm its land, theland cannot be sold or rented; Rather, it reverts to the state.

Households consume goods and servicesaccording to a preference structure determined bythe extended linear expenditure system. Throughspecification of a subsistence quantity of each goodor service, this expenditure function generates non-homothetic demands, whereby the larger the rela-tive importance of subsistence consumption (e.g.,importance would be high for rice and low forautomobiles), the more income-inelastic is thehousehold’s demand for that good.

The other important dimension of householdbehavior is the supply of labor to off-farm activi-ties. Whereas in developed economies this supplyis typically modeled as a function of the relativewages in the farm and nonfarm sectors, in Chinathe off-farm labor supply decision is complicatedby institutional factors that have been built into thesystem to keep the agricultural population in place(Zhao 1999b). For example, the Chinese governmenthas sought to make it costly for people to leaverural areas by tying incomes to daily participationin collective work. More recently, the absence ofwell-defined land tenure has raised the opportunitycost of leaving the farm. Because households thatcease to farm their land may lose their rights to it,they have a strong incentive to continue some levelof agricultural activity, even when profitability isquite low (Zhao 1999a). Low-skill farm householdsare prevented from moving to urban areas throughdenial of hukou (household registration), a govern-ment measure discussed in more detail later in thischapter. Because the growth in rural, nonfarmactivities has been only modest, the factors justdescribed seriously limit the ability of householdsto obtain off-farm work (Zhao 1999b).4

Our approach to modeling the off-farm laborsupply is to capitalize on the econometric work ofSicular and Zhao (2002). They report results from ahousehold labor supply model estimated using laborsurvey data from the 1997 Chinese Health andNutrition Survey (CHNS) of nine central provinces.This survey measures the labor supplied by mem-bers of each household to farm and nonfarmactivities. Sicular and Zhao estimate the implicit(shadow) wage for each individual in the sample ifhe or she were to work in agriculture or nonagri-cultural self-employment, and they also estimatethe nonagriculture wage that this person couldobtain. They then estimate labor supply equationsfor self-employed agricultural labor, self-employed


nonagricultural labor, and wage labor. From theseequations, it is possible to calculate elasticities of labor transfer between the farm and nonfarmsectors.

Because of the variety of labor supply elastici-ties produced in response to the three differentwages, the authors obtain a variety of labor trans-fer elasticities, depending on the “thought experi-ment” being conducted. These labor transferelasticities are asymmetric in their framework. Forexample, the responsiveness of labor transfersfrom agriculture to market wage employmentbased on a decline in the shadow wage of labor inagriculture is 2.67, whereas the elasticity of labortransfers in response to an increase in the marketwage is only 0.6. Yet this response is symmetric inour model, for which only relative prices matter,and so it is difficult to choose the correct parame-ter for our analysis. For our base case, we adoptthe “push” elasticity of labor transfer (2.67) (calledhere the high labor transfer elasticity). However,for purposes of sensitivity analysis we also reportresults from a simulation in which the “pull” elas-ticity of 0.6 is used (called here the low labortransfer elasticity).

Modeling Household Migration

Despite the large income and poverty differentialbetween rural and urban households, China haslimited permanent migration through a combina-tion of direct and indirect measures. First andforemost, households must have an appropriateregistration (hukou) to reside legally in an urbanarea. Without this registration, access to many ofthe urban amenities, including housing and educa-tion, is limited and quite expensive. Althoughhighly skilled persons and investors can purchase a“blue stamp hukou” (Chan and Zhang 1999), thisavenue is not available to the vast majority of ruralresidents. Because of these barriers to moving anentire household to an urban area, rural–urbanmigration is largely a transitory phenomenon—and one that is occurring on a massive scale. Bymost estimates, the number of “floating workers”(excluding commuters) exceeds 50 million, orabout 10 percent of China’s rural labor force. Ifcommuters and mobile self-employed individualsare included, the figure doubles to about 100million (Rozelle and others 1999).

For our modeling exercise, it is important toobtain an estimate of the wage gap motivating thetemporary migration of workers between the ruraland urban sectors in China. Zhao (1999a) docu-ments an average annual wage gap between ruraland urban work of 2,387.6 yuan for unskilled ruralworkers of comparable background and ability inSichuan Province in 1995. Much of the wage gapstems from the social costs associated with migra-tion, including the disutility of being away fromfamily, poor quality of housing, limited social serv-ices for migrants, the danger of being robbed enroute to and from the work location, and the gen-eral uncertainty associated with being a nonregis-tered worker in an urban area (Zhao 1999a, 1999b).Although these transaction costs are not observ-able, they clearly represent a very significant burdenon the migrants and their families.

If there were no barriers to the movement oflabor between rural and urban areas, we wouldexpect real wages to be equalized for an individualworker with given characteristics. Shi, Sicular, andZhao (2002) explore the question of rural–urbaninequality in greater detail for nine differentprovinces using the CHNS. The authors concludethat the apparent labor market distortion is about 42percent of the rural–urban labor income differentialand 48 percent of the hourly earnings differential.5

When applied to the average wage differential, thisdistortion amounts to an ad valorem “tax” on ruralwages of 81 percent.6

We model these transaction costs as real coststhat are assumed by the temporary migrants, recog-nizing that these migrants are heterogeneous andthat the extent of the burden varies widely. Thosemigrants who are single and live close to the urbanarea in which they are working are likely to experi-ence minor inconvenience as a result of this tempo-rary migration. We expect them to be the first tomigrate, other things being equal, in response tohigher urban wages. As for migrants who have largefamilies and come from a great distance, theirurban living conditions are often very poor, and itis not uncommon for them to be robbed on thetrain when they are returning home from work. Forsuch people, the decision to migrate temporarily islikely to be a marginal one—and one that they maynot choose to repeat. With this heterogeneous pop-ulation in mind, we postulate a transaction costfunction that is increasing in the proportion of the


rural population engaged in temporary work. Thistransaction cost function has a simple, constantelasticity functional form, which begins at the ori-gin and reaches the observed wage gap (adjustedfor transport and living costs) at the current level oftemporary migration (about 70 million workers).We assume that further increases in temporarymigration have only a modest impact on thesetransaction costs.7

In our subsequent analysis, these transactioncosts associated with temporary migration will playan important role in the unskilled and semiskilledlabor market. To the extent that WTO accessionincreases the demand for unskilled labor in urbanareas, urban wages will rise and more rural laborwill be drawn in. However, this supply of rurallabor will come at some cost—both in terms ofhigher transport and living costs for the worker,as well as indirect transaction costs—because theadditional workers are presumably being drawnfrom a greater distance or from less favorable fam-ily or social circumstances.

Modeling Production, Exports and Imports

An important characteristic of our CGE model isthe explicit treatment of two separate foreign trad-ing regimes. One is the export processing regime,discussed earlier in the context of table 16.2. Thissector receives duty-free imports and is thereforeextremely open, with considerable foreign invest-ment. The other sector is the ordinary trade regime.Since the 1990s, processing exports have grownrapidly as a result of their preferential treatment.They now account for more than half of China’stotal exports. Obviously, any analysis of externaltrade and the impact of changes in trade policymust have an explicit treatment of this dualisticforeign trading regime in the model.

Trade is modeled using the Armington assump-tion for import demand and a constant elasticity of transformation for export supply. Thus Chineseproducts are assumed to be differentiated fromforeign products, and exports from China aretreated as products different from those sold on thedomestic market. The small-country assumption isassumed for imports, and so world import pricesare exogenous in terms of foreign currency. Exportsare demanded according to constant-elasticitydemand curves, the price elasticities of which are

high but less than infinite. Therefore, the terms oftrade for China are endogenous in our simulation.

Production in each of the sectors of the econ-omy is modeled using nested constant elasticity ofsubstitution functions, and constant returns toscale are assumed. Sectors differentiate betweenrural and urban labor that substitute imperfectlybetween them. This is an indirect means of build-ing into the model a geographic flavor, becausesome sectors will be located largely in urban areas,while others will be predominantly in rural areas.By limiting the substitutability of rural and urbanlabor in each sector, we are able to proxy the eco-nomic effect of geographically distributed activity.Thus if WTO accession boosts the demand forgoods that are predominantly produced in urbanareas, then urban wages will rise relative to ruralwages, and migration will be encouraged. Ideally,we would model the geographic distribution ofindustrial activity, but unfortunately the data donot exist to support this type of split.

All commodity and nonlabor factor markets areassumed to clear through prices. With the excep-tion of the farm–nonfarm labor supply decision,labor is assumed to be perfectly mobile across sec-tors. Capital is assumed to be partially mobile,reflecting differences in the marketability of capitalgoods across sectors.

The current version of the DRC’s CGE modelhas a simple recursive dynamic structure. Dynam-ics in the model originate from the accumulation of productive factors and productivity changes.The model is benchmarked on China’s 1997 socialaccounting matrix and is solved for subsequentyears from 1998 to 2007. We turn now to the detailsof the baseline scenario.

Base Case Projections and WTOAccession Simulation

China’s WTO accession includes a complex packageof trade and investment liberalization. In this chap-ter, however, only four aspects are considered:

(1) tariff reduction on industrial products (2) agricultural trade liberalization (i.e., tariff

reduction for agricultural products), intro-duction of the tariff-rate quota system foragricultural goods, and elimination of exportsubsidies for corn and cotton

(3) liberalization of direct trade in services



TABLE 16.3 Summary of Baseline Calibration

1997 2000 2001 2002 2003 2004 2005 2006 2007

Exogenous specified variablesGDP growth rate (%) — 8.1 7.3 7.6 7.8 7.8 7.8 7.8 7.8Life expectancy (years) 70.3 71.8 72.0 72.2 73.2 73.4 73.6 73.8 74.0Total fertility rate

Urban 1.40 1.45 1.48 1.5 1.52 1.55 1.57 1.59 1.62Rural 2.01 1.95 1.97 1.99 2.01 2.02 2.04 2.06 2.07

Calibrated results:Macroeconomic trendsGrowth rate (%)Total absorption — 11.02 8.53 8.29 8.45 8.39 8.34 8.29 8.25Labor force — 1.23 1.21 1.17 1.15 1.17 1.16 1.15 1.15Capital stock — 10.08 10.51 10.45 10.28 10.18 10.08 9.98 9.88Total Factor Productivity — 2.87 1.74 2.16 2.43 2.44 2.49 2.53 2.56

Ratio to GDP ( %)Private consumption 44.3 47.0 47.3 47.4 47.4 47.4 47.5 47.6 47.6Investment 32.7 34.1 35.2 35.9 36.7 37.4 38.1 38.8 39.4Export 21.7 25.5 25.7 25.7 25.7 25.7 25.7 25.6 25.6Import 16.3 22.9 23.5 23.7 23.8 24.0 24.1 24.2 24.2

Urban/rural per capita income 3.03 2.77 2.74 2.71 2.68 2.65 2.63 2.60 2.58ratio

Population and labor forcePopulation (millions) 1,236.3 1,269.9 1,280.5 1,290.8 1,301.0 1,311.5 1,321.9 1,332.4 1,343.0

Urban 369.9 397.5 407.2 416.9 426.6 436.5 446.4 456.4 466.5Rural 866.4 872.4 873.3 873.9 874.3 875.0 875.5 876.0 876.4

Labor force (millions) 640.7 665.6 673.6 681.5 689.4 697.4 705.5 713.6 721.8Urban 181.1 196.9 202.5 208.2 213.8 219.5 225.3 231.2 237.2Rural 459.6 468.7 471.1 473.4 475.6 477.9 480.2 482.4 484.6

Grain self-sufficiency rate (%) 99.5 98.5 98.4 98.4 98.3 98.2 98.1 98.0 98.0Share of rural labor force (%) 71.7 70.4 69.9 69.5 69.0 68.5 68.1 67.6 67.2Share of agricultural employment 72.0 71.9 71.6 71.6 71.2 71.0 70.9 70.7 70.7

in total rural labor force (%)

Note: — Not available.

Source: Authors’ calculations.

(4) phaseout of MFA quotas on textile andclothing.

Now that China is a member of the WTO, itsexports of textiles and apparel products to NorthAmerican and European markets are being sub-jected to accelerated MFA quota growth (from 2002to 2006), similar to that affecting other developingcountries that are WTO members. The remainingquota restriction will be eliminated in 2007,according to the ATC.8 Therefore, the analysis atbest captures only parts of the issue. It does not takeinto account other major aspects of WTO member-ship, such as reducing barriers to foreign invest-ment, protecting intellectual property rights,securing market access, enforcing commitment,

and cooperating in dispute settlement. BecauseChina’s market accession commitments to WTOentry will be phased in over a seven-year transitionperiod, a baseline from 1998 to 2007 is establishedusing our recursive dynamic model under a set ofassumptions: China will continue its grain self-sufficiency policy, and the import quota of agricul-tural goods will grow at 3 percent annually from2000 to 2007. This calibrated “benchmark” willserve as a basis of comparison for counterfactualsimulation conducted in the WTO accession exper-iments. Major results from the baseline calibrationare summarized in table 16.3, and the key assump-tions are outlined in table 16.4. The baseline usesthe economy-wide total factor productivity (TFP)

variable as a residual and adjustment mechanism tomatch the pre-specified real gross domestic prod-uct (GDP) growth rate under an exogenously spec-ified life expectation rate and total fertility rate forboth the rural and urban population.

The WTO accession scenario is described indetail in table 16.4. Import tariffs are reduced gradu-ally, and the sectoral reduction rates are aggregatedfrom the Harmonized Commodity Description andCoding System tariff schedules for the period2002–2007 and weighted by 1997 ordinary tradedata. The quota control system is introduced for rice,wheat, corn, cotton, wool, vegetable oil, and sugar.Moreover, the tariffs for other agricultural goods will

also be reduced. Following François (2002), wemodel the impact of service sector liberalization ashalving the barriers to services trade. We also intro-duce a 20 percent productivity boost for the auto-mobile sector to reflect the efficiency gain fromindustrial restructuring and realization of economicscale in this sector after China WTO accession.

Simulation Results

We now turn to the results of our analysis of WTOaccession, beginning with the economy-wideeffects, then turning to the disaggregated house-hold impacts.


TABLE 16.4 Summary of Experiments

Experiment Description

Base Base case:— Real GDP exogenous— TFP growth rate

— Agricultural sector exogenous — Other sectors endogenous

— Three percent growth rate of import quota for goods subjected to quantitativerestriction (rice, wheat, corn, cotton, wool, vegetable oil, sugar, petroleumrefining, automobiles)

— Exogenous export quota growth for textiles (5.7 percent) and apparel(6.0 percent)—annual average

— All tax rates fixed at their base year level— Balance of payments gradually declines to 40 percent of its base year level in 2007

WTO WTO accessionTariff reduction — An average 60 percent cut in 2000 tariff level from 2000 to 2007Agricultural trade liberalization— Quota restriction (10 billion yuan, 1997)

Initial quota in 2000 Annual growth rate of quota (%)

Rice 1.073 26.4Wheat 1.252 10.5Corn 0.374 18.0 Cotton 1.153 4.5Wool 0.665 6.5Vegetable oil 3.706 27.0Sugar 0.408 7.6

— Tariff cut for other agricultural goods — Elimination of export subsidies for corn and cotton

Phaseout of MFA— Acceleration of MFA quota growth rate from 2001 to 2004— Zero export tax of textiles and apparel in 2005

Source: Protocol on the Accession of WTO; Authors’ assumptions.

Economy-Wide Impacts of WTO Accession

Our results, which are directed at the outcome in2007 when China’s accession to the WTO is com-plete, are reported in terms of deviations from thebaseline in 2007. The first column of results intable 16.5 reports the macroeconomic impacts ofaccession on China’s economy under the hightransfer elasticity assumption. Real GDP and wel-fare both increase by 0.65 percent and 0.73 percent,respectively, while consumption increases by over1 percent and investment by only half as much. Thereduction in trade barriers gives a substantial boostto trade in China, with both exports and importsrising by about 15 percent.

As for changes in factor prices, skilled wages risemore than semiskilled wages, which, in turn, risemore than unskilled wages. The relatively greaterincrease in skilled wages is fueled by the tendencyfor the manufacturing and service sectors to expandat the expense of agriculture. The manufacturingand service sectors are relatively intensive in theiruse of skilled and semiskilled labor, thereby boost-ing wages for these factors relative to unskilledwages. And then wages in the different markets arelinked. In particular, urban and rural nonfarmwages are linked through the temporary migrationof workers to the urban areas. Specifically, 1.46 mil-lion additional temporary migrants move to thecity in response to the higher urban wages. As aconsequence, the rural wage is once again equatedwith the urban wage, less the direct and indirectcosts associated with migration. With transactioncosts changing only modestly (–0.15 percent to 0.63 percent), the boost in wages is relatively moreimportant for rural households, whose wagesincrease by a greater proportion because of the(nearly fixed) transaction costs.

Returns to labor in agriculture are also linked tothe rural nonfarm wages via off-farm migration. Asdiscussed earlier, this decision is based on the combi-nation of agricultural wages and returns to land,because migrants risk losing the right to the land theyleave. Because the returns to land fall, the returns toremaining in farming rise very little, thereby leadingto additional off-farm migration (1.89 million).

Sector Impacts

Table 16.6 reports the impact of accession on out-put, imports, exports, and relative prices in the

Chinese economy. To facilitate analysis, figure 16.1reports the output changes in descending order,omitting the changes that are less than 2 percent inabsolute value. Clearly, the largest increases in out-put stem from elimination of the quotas on textilesand apparel exports to North America and Europe;these products, as well as the production of syn-thetic fibers and cotton, increase by substantialamounts. On the other end of the spectrum, themost heavily protected sectors, which have sizabletrade exposure, experience declining output. Thesesectors include wool, automobiles, vegetable oils,sugar, machinery, nonferrous metal products, spe-cial equipment, and wheat.

With the exception of a few mining productsand transport services for which there is no cut inprotection, import volumes increase for all sectorsin the economy. The largest increases are for foodand agricultural products (where there are deepcuts in protection), textiles and apparel (where the demand for intermediate inputs increasesstrongly), as well as automobiles (where there arestrong productivity gains from the rationalizationof production). Export volumes for most productsalso increase, fueled by cheaper inputs and realdepreciation, as well as the elimination of textileand apparel quotas. Only cotton and corn, forwhich export subsidies are removed, and wool,which is used heavily in the expanding textile sec-tor, experience reductions in export volumes.

Table 16.6 also reports the change in net exportsfor each of the product categories. It is not surpris-ing that the largest increase in net exports is in theapparel and textile sectors because of elimination ofthe quotas under the Agreement on Textiles andClothing. The net exports of apparel and textilesincrease by 286 billion yuan and 61 billion yuan,respectively. Because chemical fibers and cotton aremajor intermediate inputs in the textile sector, theirnet imports also increase sharply. In addition,China’s net imports of automobiles, chemicals,electronics, machinery, and commercial servicesalso increase strongly, offsetting the increasingtrade surplus from the textile and apparel sectors.

Finally, table 16.6 reports the percentage changein commodity prices relative to the Consumer PriceIndex in China. Wool, vegetable oil, and automo-biles experience substantial relative price declines,because imports have a high market penetrationratio and their rates of import protection are cut



TABLE 16.5 Implications of China’s WTO Accession, 2007 (percentage change)

High Labor Low LaborTransfer Elasticity Transfer Elasticity

Macroeconomic variablesWelfare (Equivalent Variation) 0.73 0.62GDP 0.65 0.54Consumption 1.02 0.94Investment 0.46 0.39Exports 14.98 14.35Imports 14.23 13.64

Factor pricesReturns to agricultural land �3.54 �1.65Capital returns �0.56 �0.35Unskilled wages

Urban �0.30 0.24Rural nonagricultural �0.34 0.23Agricultural �0.76 �0.61

without land return 0.13 �0.08Semiskilled wages

Urban 0.37 0.69Rural nonagricultural 0.26 0.64Agricultural �0.28 �0.54

without land return 0.66 0.00Skilled wages

Urban 1.11 0.89Rural nonagricultural 1.11 0.89Agricultural 0.42 �0.21

without land return 1.10 0.21

Inequality measurementa

Urban/rural income ratio 0.0107 0.0167Gini 0.0013 0.0012

Urban 0.0004 0.0002Rural 0.0001 0.0003

Transaction costsUnskilled �0.15 0.31Semiskilled 0.63 0.84

Labor migration (millions)Agricultural-nonagricultural 1.89 0.54Rural-urban 1.46 0.45

Unskilled 0.35 0.13Semiskilled 1.03 0.29Skilled 0.07 0.03

Labor migration (%)Agricultural-nonagricultural 2.42 1.55Rural-urban 2.05 1.07

Unskilled 1.20 0.62Semiskilled 2.86 1.75Skilled 1.28 0.59

aChange of original value, not percentage change.Source: Authors’ model simulation results.


TABLE 16.6 Sector Impacts of China’s WTO Accession in 2007: Percentage Deviation from Baseline

Net Trade Prices (Relative(Billions of to Consumer

Output Import Export Yuan) Price Index)

Rice �0.8 151.7 2.3 �4.5 �0.4Wheat �4.2 50.9 — �8.9 �1.8Corn �3.7 181.1 �66.5 �6.3 �0.2Cotton 3.2 78.6 �23.8 �8.2 �0.5Other nongrain crops �1.0 53.9 2.4 �7.8 �0.3Forestry �2.9 16.9 �2.2 �2.9 0.3Wool �32.9 24.8 �19.2 �1.8 �14.6Other livestock 1.3 1.6 3.5 0.3 0.2Fishing 0.5 3.3 2.1 0.1 0.6Other agriculture 0.1 5.7 6.2 0.1 �0.2Coal mining �0.8 �1.8 1.3 0.1 0.6Crude oil and natural gas 0.3 �0.3 5.7 1.9 �0.4Ferrous ore mining �2.3 �3.6 3.6 1.3 0.3Nonferrous ore mining �4.1 �4.1 �0.5 0.7 0.6Other mining �0.4 1.3 3.5 0.1 0.1Vegetable oil �14.6 167.2 170.3 �20.1 �7.9Grain mill and forage 0.8 5.1 27.9 �0.9 �0.6Sugar �10.9 180.1 295.5 �2.9 �4.5Processed food 1.3 17.0 10.2 5.9 �0.3Beverages �0.6 130.0 11.2 �5.0 �0.7Tobacco �2.1 81.4 6.6 �6.1 �1.3Textiles 17.9 62.0 45.3 60.6 �1.3Apparel 38.6 54.8 102.2 286.2 �2.7Leather �1.3 19.7 1.4 �6.4 �0.6Sawmills and furniture �1.0 14.9 3.3 �3.5 �0.4Paper and printing �3.3 13.6 5.5 �16.3 �0.5Crafts and toys 1.1 6.5 2.7 3.4 �0.2Petroleum refining 0.0 1.0 4.8 4.8 �0.6Chemicals �1.8 7.9 3.0 �26.1 �0.7Pharmaceuticals 0.3 14.1 5.2 0.3 �0.3Chemical fibers 8.0 48.0 6.8 �28.7 �1.1Rubber and plastics �2.1 10.7 1.3 �5.3 �0.4Building materials �0.4 13.2 2.7 �2.5 0.1Primary iron and steel �3.8 2.4 1.3 �3.8 �0.1Nonferrous metals �4.8 1.7 �0.1 �2.1 �0.1Metal products �2.4 10.1 1.5 �11.4 �0.4Machinery �5.1 9.1 0.1 �21.8 �0.7Special equipment �4.6 7.0 1.1 �22.1 �1.6Automobiles �19.7 83.5 37.9 �50.1 �11.3Other transport equipment 0.9 �0.1 2.2 1.3 �0.1Electric machinery �2.6 13.4 1.0 �21.0 �0.8Electronics �3.8 6.3 0.3 �29.2 �2.1Instruments �3.3 6.9 �0.3 �7.1 �2.0Other manufacturing �1.7 21.6 0.2 �2.4 0.5Utilities �0.6 16.5 2.9 0.4 0.1Construction 0.3 10.3 3.1 �2.6 0.2Transportation 0.3 �1.5 3.2 4.7 �0.1

sharply. For most other sectors the price changesare quite modest.

Household Impacts

Table 16.7 reports the household impacts in 2007 ofChina’s accession to the WTO. The first point tonote in table 16.7 is that WTO accession benefits all households, except those reliant on transfers. Thelatter group, which receives much of its transferincome from state-run enterprises, is adverselyaffected by the decline in rents accruing to these

enterprises. Government transfers are assumed tobe constant in real terms and are indexed by theConsumer Price Index.

The smallest welfare increases for the other(nontransfer) households in table 16.7 are associ-ated with agriculture-specialized households. Theyare adversely affected by the decline in returns toagricultural land, but rising returns to labor stillboost the average returns to farming overall assome individuals leave the sector. Thus, even theagriculture-specialized households benefit fromWTO accession. Figure 16.2 contrasts the gains for



Net Trade Prices (Relative(Billions of to Consumer

Output Import Export Yuan) Price Index)

Post and communication 0.5 4.0 4.3 2.1 �0.4Commercial sources 1.7 21.7 2.4 �15.8 �0.2Finance �0.5 17.4 1.7 �2.7 0.1Social services �1.0 14.9 2.1 �13.2 �0.8Education and health �0.3 18.6 0.9 �2.9 0.5Public administration �0.4 18.0 1.4 �1.9 0.3

— Not available.Source: Authors’ model simulation results.

�40.0

�30.0

�20.0

�10.0

0.0

10.0

20.0

30.0

40.0

50.0

Output (% change)

Appar

el

Textile

s

Chem

ical fi

bers

Cotto

n

Tobac

co

Rubbe

r and

plas

tics

Ferro

us o

re m

ining

Met

al pr

oduc

ts

Electri

c mac

hiner

y

Fores

try

Paper

and

prin

ting

Instr

umen

tsCor

n

Electro

nics

Iron

and

steel

Nonfe

rrous

ore

mini

ng

Specia

l equ

ipmen

t

Nonfe

rrous

met

als

Mac

hiner

y

Sugar

Veget

able

oils

Autom

obile

s

Woo

l

Sector

Whe

at

FIGURE 16.1 Change in Sector Output due to WTO Accession, China

Source: Authors’ model simulation results.


TABLE 16.7 Household Impacts of China’s WTO Accession, 2007 (EquivalentVariation as percentage of household income)

Urban Rural

Vingtile Transfer- Labor- Agriculture-(Poorest � 1) Specialized Specialized Diversified Specialized Diversified

High labor transfer elasticity1 �0.274 1.217 1.044 0.446 0.5342 �0.548 1.395 1.247 0.426 0.5713 �0.450 1.533 1.279 0.343 0.6064 �0.534 1.477 1.103 0.295 0.5915 �0.584 1.451 1.252 0.366 0.6096 �0.464 1.571 1.048 0.355 0.6367 �0.397 1.631 1.315 0.320 0.6588 �0.287 1.607 1.395 0.522 0.6369 �0.381 1.706 1.295 0.370 0.623

10 �0.550 1.844 1.308 0.409 0.67111 �0.457 1.709 1.269 0.316 0.65012 �0.397 1.731 1.271 0.367 0.64813 �0.356 1.766 1.267 0.345 0.63514 0.263 1.719 1.224 0.421 0.68315 �0.270 1.826 1.339 0.329 0.67816 �0.190 1.768 1.204 0.654 0.66417 — 1.631 1.102 0.782 0.74918 0.052 1.682 1.230 0.629 0.64219 �0.354 1.774 1.006 0.391 0.61820 — 1.743 0.506 0.469 0.619

Low labor transfer elasticity1 �0.270 1.260 1.103 0.129 0.3732 �0.566 1.418 1.272 0.090 0.4113 �0.463 1.453 1.279 0.020 0.4404 �0.553 1.471 1.128 �0.029 0.4355 �0.596 1.393 1.212 0.037 0.4646 �0.466 1.470 1.065 0.025 0.5017 �0.437 1.520 1.264 �0.011 0.5318 �0.291 1.526 1.337 0.188 0.5109 �0.393 1.570 1.239 0.042 0.519

10 �0.566 1.689 1.217 0.064 0.57311 �0.470 1.566 1.183 �0.014 0.55912 �0.398 1.570 1.179 0.021 0.56413 �0.369 1.570 1.186 0.021 0.55014 0.326 1.562 1.123 0.068 0.58815 �0.294 1.647 1.233 0.023 0.60116 �0.185 1.547 1.126 0.253 0.59617 — 1.488 0.996 0.410 0.67218 0.073 1.490 1.082 0.309 0.59819 �0.356 1.552 0.921 0.075 0.56120 — 1.484 0.477 0.146 0.519

Note: — Not available.Source: Authors’ model simulation results.

rural households specialized in farming with thosethat have diversified income sources. For the lattergroup, income rises more because of the increasedrole of nonfarm wage earnings. The largestincreases in welfare after WTO accession accrue tothe urban households—especially the wealthier,labor-specialized households. They benefit fromthe fact that, of all factor prices, skilled wagesincrease the most after WTO accession. Figure 16.3contrasts these changes with the changes in EV(relative to the initial expenditure) for the urban

diversified households who also have significantincome from capital. Because the latter rise moremodestly, the highest-income, diversified house-holds benefit proportionately less than the othernontransfer-specialized, urban household groups.

Sensitivity to the Elasticity of Labor Transfer out of Agriculture

As noted earlier, Sicular and Zhao (2002) provideseveral estimates of the labor transfer elasticity out


0.0

0.2

0.3

0.1

0.4

0.5

0.6

0.7

0.8

1 2 3 4 5 6 7 8 9 10Vingtile

11 12 13 14 15 16 17 18 19 20

EV as percent of income

Ag-specialized Diversified

FIGURE 16.2 Impact of WTO Accession on Rural Households

0.0

0.4

0.8

1.2

1.6

2.0

1 2 3 4 5 6 7 8 9 10Vingtile

11 12 13 14 15 16 17 18 19 20

EV as percent of income

Labor-specialized Diversified

FIGURE 16.3 Impact of WTO Access on Urban Households



of agriculture. In our base case results, we use theelasticity that relates the relative supply of labor toagriculture and nonagriculture (market wage)activities to a change in the relative wage rates—owing to a perturbation in agricultural returns.Our rationale for selecting this elasticity is that thelargest relative factor price effect of WTO accessionis the downward pressure on returns to agriculture.This “push” elasticity is 2.67. It is considerablyhigher than the “pull” elasticity engendered by anincrease in the market wage for labor, which is esti-mated to be 0.6. For purposes of sensitivity analy-sis, we turn now to the second set of results intables 16.5 and 16.7, which utilize the low labortransfer elasticity between agriculture and nonagri-culture. The second column of table 16.5 revealsthat the aggregate gains from WTO accession aresomewhat smaller in the low transfer elasticitycase—although not dramatically so.

The most important difference between thehigh and low labor transfer elasticity results is theimpact of accession on the agriculture-specializedhouseholds (table 16.7). In the case of reduced off-farm mobility of labor, the gains of agriculture-specialized households are much smaller than thosearising with a high off-farm labor transfer elasticity.In fact, a few of these farm household vingtiles nowlose from WTO accession, because they are not ableto take advantage of the improved opportunitiesoutside of the farm sector and fall prey to depressedfactor returns in agriculture. This finding confirmsthose of Chen and Ravallion (2003)—see Chap-ter 15 in this volume. By not modeling factor marketadjustment at all, Chen and Ravallion have effec-tively assumed an even smaller labor transfer elas-ticity out of agriculture—zero. In their chapter, theyconclude that the poor agricultural households willlose from WTO accession. This qualitative differ-ence in findings on rural incidence highlights theimportance of explicitly modeling labor markets inorder to obtain an accurate assessment of the impactof WTO accession on farm households in China.

Conclusions and Policy Implications

The aim of this study has been to quantify theimpact of China’s accession to the WTO onincome distribution and poverty. Toward this end,we have drawn on household surveys from three

diverse provinces in China: Guangdong, Liaoning,and Sichuan. Using these rural and urban surveys,we estimated earnings and spending patterns for 20income vingtiles across five representative house-hold groups in China. These groups were dividedbased on their location (rural versus urban) andtheir primary source of earnings, yielding five cate-gories: agriculture-specialized, rural-diversified,urban labor-specialized, urban transfer-specialized,and urban-diversified. By embedding these disag-gregated households in a national CGE model ofChina, we are able to identify the disaggregatedhousehold impacts of WTO accession.

We find that WTO accession benefits all house-hold groups—with the exception of the urbantransfer-specialized households, which experiencemodest losses because of the decline in rents accru-ing to state-run enterprises. The biggest gains go tourban labor-specialized households because of thestrong rise in wages for skilled and semiskilledlabor in urban areas. The smallest gains go to therural agriculture-specialized households, whichface declining returns to agriculture-specific factorsand which have limited labor mobility out of farm-ing. Indeed, some of these households could actu-ally lose from accession under low values of thefarm–nonfarm labor transfer elasticity. Becausethese households are also the poorest in China,WTO accession runs the risk of widening incomedisparities. Other studies that have abstracted fromintersector labor mobility altogether have come tothe same conclusion (Chen and Ravallion 2003).

The critical importance of labor markets to therural impact of China’s accession to the WTO sug-gests the potential for complementary labor marketreforms to offset some of these adverse impacts. Ina companion paper to this chapter (Zhai, Hertel,and Wang 2003), we find that reforms aimed atfacilitating an improved flow of labor out of agri-culture and between the rural and urban markets,can result in significant gains for rural households.In that study, we explore the following implications:

• reforming agricultural land markets to permitarms-length land rental in all rural areas,thereby facilitating the permanent movement oflabor out of farming

• enhancing off-farm labor mobility• abolishing the hukou system, thereby reducing

the transaction costs imposed on rural–urbanmigrants.


When combined, these reforms reduce the esti-mated 2007 urban–rural income ratio from 2.58 (inthe absence of WTO accession) to 2.09. When com-bined with WTO accession, the 2007 urban–ruralincome ratio is still reduced to 2.12 as a conse-quence of these three factor market reforms.Although these reform scenarios are perhaps moredramatic than those likely to be undertaken in thenear term, even a small movement in this directionwould clearly go a long way toward offsetting thepotentially adverse consequences of WTO accessionon rural–urban inequality.

Notes

1. All dollar amounts are current U.S. dollars.2. We would have preferred to base this split on occupa-

tion—that is, what they actually do—rather than potential asdetermined by education. However, the rural household surveydoes not support this split.

3. Because the rural survey reports only the highest educa-tional attainment of the household, attainment by worker isunknown. This factor biases the skill level of rural householdsupward, but, because the vast majority of rural households areunskilled, this bias is less of a problem in practice.

4. However, as noted by Parish, Zhe, and Li (1995), the rurallabor market is looking more like a market all the time.

5. Other, unobserved factors inducing this rural–urban wagedifferential are likely. If so, estimation of the labor market distor-tion via subtraction of known factors is biased in the directionof overstating the hukou-related distortion. Therefore, it is use-ful to estimate as well the direct impact of household registra-tion status on the observed wage difference among households.Shi (2002) takes this approach to the problem, using the sameCHNS data set. He finds that only 28 percent of the rural–urbanwage difference can be explained directly via the coefficient onthe hukou registration variable. This figure is quite a bit less thanthe 48 percent left unexplained via the subtraction approach ofShi, Sicular, and Zhao (2002).

6. See Zhai, Hertel, and Wang (2003) for a detailed descrip-tion of how this ad valorem distortion is obtained.

7. We assume that a doubling of temporary migration wouldonly increase the marginal cost of migration by 10 percent.

8. On January 1, 1995, the ATC entered into force andreplaced the old Multi-Fiber Arrangement. The ATC providesfor elimination of the quotas and the complete integration oftextiles and apparel into the WTO regime over a 10-year transi-tion period, ending on January 1, 2005. All WTO countries aresubject to ATC disciplines, and only WTO members are eligiblefor ATC benefits.

References

The word processed describes informally reproduced works thatmay not be commonly available through libraries.Beghin, John, Sébastien Dessus, David Roland-Holst, and

Dominique van der Mensbrugghe. 1994. “Prototype CGEModel for the Trade and the Environment Programme-Technical Specification.” OECD Development Centre, Paris.

Chan, Kam Wing, and Li Zhang. 1999. “The Hukou System andRural-Urban Migration in China: Processes and Changes.”China Quarterly 160 (December): 818–55.

Chen, Shaohua, and Martin Ravallion. 2003.“Welfare Impacts ofChina’s Accession to the WTO.” World Bank, Washington,D.C. Processed.

Chen, Shaohua, and Yan Wang. 2001. “China’s Growth andPoverty Reduction: Recent Trends between 1990 and 1999.”World Bank, Washington, D.C. Processed.

de Melo, J. 1988. “Computable General Equilibrium Models forTrade Policy Analysis in Developing Countries: A Survey.”Journal of Policy Modeling 10: 469–504.

de Melo, J., and David Tarr. 1992. A General EquilibriumAnalysis of US Foreign Trade Policy. Cambridge, Mass.: MITPress.

Dervis, K., J. de Melo, and Sherman Robinson. 1982. GeneralEquilibrium Models for Development Policy. Cambridge:Cambridge University Press.

Development Research Center. 1998. The Global and DomesticImpact of China Joining the World Trade Organization,Research Paper, Development Research Center of the StateCouncil, PRC, Beijing.

Dimaranan, Betina V., and Robert A. McDougall. 2002. GlobalTrade, Assistance, and Production: The GTAP 5 Data Base.West Lafayette, Indiana: Center for Global Trade Analysis,Purdue University.

Francois, Joseph F., and Spinanger, Dean. 2004. “Regulated Effi-ciency, WTO Accession, and the Motor Vehicle Sector inChina.” In D. Bhattasali, S. Li, and W. Martin, eds., China andthe WTO: Accession, Policy Reforms, and Poverty ReductionStrategies. Washington, D.C.: World Bank and OxfordUniversity Press.

Ianchovichina, Elena, R. McDougall, and T. Hertel. 1999. “China2005 Revisited: The Implications of International CapitalMobility.” Global Trade Analysis Project, Purdue University,West Lafayette, Indiana. Processed.

Hertel, Thomas W., Maros Ivanic, Paul V. Preckel, and JohnCranfield. 2004. “The Earnings Effects of MultilateralTrade Liberalization: Implications for Poverty.” World BankEconomic Review. Forthcoming.

Kanbur, R., and Zhang, Xiaobo. 2001. Fifty years of regionalinequality in China: A journey through revolution, reformand openness. Working Paper 2001-04, Department ofApplied Economics and Management, Cornell University,Ithaca, New York.

McKibbin, Warwick, and K. K. Tang. 2000. “Trade and FinancialReform in China: Impacts on the World Economy.” WorldEconomy 23(8): 979–1003.

Naughton, Barry. 1996. “China’s Emergence and Prospects as aTrading Nation.” Brookings Paper on Economic Activity 2,Washington, D.C.

Parish, W. L., X. Zhe, and F. Li. 1995. “Nonfarm Work andMarketization of the Chinese Countryside.” China Quarterly143: 697–730.

Rozelle, Scott, Guo Li, Minggao Shen, Amelia Hughart, and JohnGiles. 1999. “Leaving China’s Farms: Survey Results of NewPaths and Remaining Hurdles to Rural Migration.” ChinaQuarterly 158 (June): 367–93.

Shi, Xinzheng. 2002. Empirical research on urban-rural incomedifferentials: The case of China. unpublished manuscript,CCER. Beijing: Beijing University.

Shi, Xinzheng, Terry Sicular, and Yaohui Zhao. 2002. “AnalyzingUrban-Rural Income Inequality in China.” Paper presentedat International Symposium on Equity and Social Justice inTransitional China,” Beijing, July 11–12.


Shoven J. B., and J. Whalley. 1992. Applied General EquilibriumAnalysis. Cambridge: Cambridge University Press.

Sicular, Terry, and Yaohui Zhao. 2004. “Earnings and LaborMobility in Rural China: Implications for China’s Accessionto the WTO.” In D. Bhattasali, S. Li, and W. Martin, eds.,China and the WTO: Accession Policy Reform, and PovertyReduction Strategies. Washington, D.C.: World Bank andOxford University Press.

Walmsley, Terrie, and Thomas Hertel. 2000. “China’s Accessionto the WTO: Timing Is Everything.” Paper presented atInternational Conference on Global Economic Transforma-tion after the Asian Economic Crisis, Hong Kong.

Wang, Zhi. 1999. “Impact of China’s WTO Entry on LaborIntensive Export Market—A Recursive Dynamic CGEAnalysis.” World Economy 22 (3): 379–405.

———. 2001. “The Impact of China’s WTO Accession on Tradeand Economic Relations Cross the Taiwan Strait.” Economicsof Transition 9 (3).

Wang, Zhi, and Fan Zhai. 1998. “Tariff Reduction, Tax Replace-ment and Implication for Income Distribution in China.”Journal of Comparative Economics 26: 358–87.

WTO (World Trade Organization). Protocol on the Accession ofthe People’s Republic of China. Geneva: WTO.

Yang, Dennis T. 1997. “China’s Land Arrangements and RuralLabor Mobility.” China Economic Review 8 (2): 101–16.

Yang, Dennis Tao, and Hao Zhou. 1999. “Rural-Urban Disparityand Sectoral Labour Allocation in China.” Journal of Devel-opment Studies 35 (3): 105–33.

Yang, Y., and Y. Huang. 1997. “The Impact of Trade Liberaliza-tion on Income Distribution in China.” China EconomyWorking Paper 97/1. Research School of Pacific and AsiaStudies, Australia National University, Canberra.

Zhai, Fan, and Shantong Li. 2000.“The Implications of Accessionto WTO on China’s Economy.” Paper presented at ThirdAnnual Conference on Global Economic Analysis,Melbourne, Australia, June 27–30.

Zhai, Fan, and Zhi Wang. 2002. “WTO Accession, Rural LabourMigration and Urban Unemployment in China.” UrbanStudies 39, no.12.

Zhai, Fan, Thomas W. Hertel, and Zhi Wang. 2003. “Labor Mar-ket Distortions, Rural-Urban Inequality and the Opening ofChina’s Economy.” GTAP working paper 27. Center forGlobal Trade Analysis, Purdue University, West Lafayette,Indiana. Available at http://www.gtap.agecon.purdue.edu/resources/res_display.asp?RecordID �1323.

Zhao, Yaohui. 1999a. “Labor Migration and Earnings Differ-ences: The Case of Rural China.” Economic Development andCultural Change, 47(4): 767–82.

———. 1999b. “Leaving the Countryside: Rural-to-UrbanMigration Decisions in China.” American Economic Review(May 1999): 281–86.


losses, there is a social dimension to be consideredwhenever the interpersonal distribution of lossesand gains do not coincide. In the context of China’sWTO accession, this might happen in the followingfour ways:

1. Because of a reduction in the prices of importedcommodities, a large number of consumers gain,while a comparatively small number of work-ers lose jobs or farmers experience a decline inearnings.

2. Loss of jobs or decline in farm earnings precedesjob creation or increases in farm earnings due totrade liberalization.

3. Those losing jobs or experiencing an incomedecline differ from those recruited into new jobsor experiencing an income rise.

4. Regions reaping gains from WTO accession dif-fer from regions affected by the adverse effects ofWTO accession.

All four scenarios involve noncoincidencebetween the distribution of gains and losses andthus create a potential issue for social protection.The first concerns the distribution of gains betweenconsumers and producers (employees and farm-ers), and it arises from the simple fact that, formany consumer goods, the number of consumers

The knock-on effects of the conditions of China’saccession to the World Trade Organization (WTO)on the labor market (including on farmers) are amajor issue in analysis of the social impacts ofWTO membership, in particular its impact onpoverty and income equality. The nature of thelabor market effects can be expected to differ in theshort and the long run. In the short run, some ofthe effects will be adverse and will take two forms:(1) job shedding in particular industries, and (2) areduction in the earnings of some farmers, depend-ing on the region and the commodities produced.

Correlatively, the beneficial labor market effectswill accrue in the form of expanded employment inthe sectors opened wider to foreign direct invest-ment (FDI). Additional beneficial effects wouldstem from job creation and the rise in earningsfrom the lower barriers facing Chinese exports.Broadly, the consensus is that product and labormarkets effects all included, gains from WTOmembership will far outweigh losses. Moreover,given the increased efficiency flowing from WTOaccession in the long run the positive labor marketimpact (changes in jobs and earnings) would ex-ceed the negative impact.

As for the social impact of WTO accession, thedistribution of gains and losses is as important as thebalance between the two. Even when gains exceed

305

17

Coping with andAdapting to Job

Losses and Declinesin Farm Earnings

in China

Athar Hussain

far exceeds the number of producers, includingworkers employed directly and indirectly. Theimplication is that adverse effects in the form of joblosses or decline in earnings produced by WTOaccession are likely to be far more concentratedthan the favorable effects associated with a fall inprices dispersed over a large number of consumers.Such patterns of distribution of losses and gains arelikely to arise for consumer goods such as passengervehicles, making the adverse effects more visiblethan the favorable ones.

The second is particularly important becauseChina had to start lowering barriers to foreignimports from the date of accession, December 11,2001, but its exports will benefit from reduced bar-riers in other WTO member countries only after atime lag, in some cases a long and uncertain one. Inshort, whereas China’s WTO accession involves anobligation to lower barriers to imports, it providesno more than a distant prospect of lower barrierson exports. The abolition of the Multi-FiberArrangement, from which China is expected to be amajor beneficiary, is scheduled to take place in2005, and Chinese textiles may continue to face spe-cial textile safeguards until 2008. Similarly, barrierson foreign farm imports began to fall immediatelyupon China’s accession to the WTO, but it is uncer-tain when China’s farm exports will face lower bar-riers; it is contingent on the outcomes of bilateraltrade negotiations and of the Doha Round of tradetalks, which will take some years to conclude.

In the third and fourth scenarios in the list, amismatch occurs between the distributions oflosses and gains over individuals and regions. In thefirst of two examples of likely outcomes of WTOaccession, employment expands in the textile andclothing industry and the number of jobs in thevehicle and motor parts industries falls. Those get-ting jobs in the textile and clothing industries willnaturally not all be just those losing jobs in thevehicle industry, even when both sets of industriesare located in the same place. The second exampleconcerns agriculture, where grain farmers, espe-cially those heavily dependent on corn, wheat, andcoarse grain, will likely be losers. These losers willbe concentrated in the Northeast and Northwest.The gainers will consist of producers of exportcrops and animal products. Most of the gainers willbe in the coastal provinces of the East and theSoutheast. The Chinese economy is made up of

numerous subeconomies that are loosely connectedrather than fully integrated. Regional differences infactor endowments, labor productivity, or scaleeconomies within China are arguably as pro-nounced as those between other countries. Oneimplication is that the impact of WTO accession is likely to vary regionally. Moreover, the impact on the Chinese economy of WTO accession willdepend crucially on what happens to internal tradebarriers in the years after accession. A lowering ofsuch barriers induced by China’s membership inthe WTO may have the effect of a geographic redis-tribution of production within the Chinese eco-nomy as well as between China and the rest ofthe world. It is conceivable that the impact of theremoval of internal trade barriers may dominatethe impact of lower international trade barriers.

Empirical evidence on the seriousness of lossesin the Chinese context is not available. However, theevidence from the United States provides a bench-mark for estimating losses suffered by adverselyaffected workers. The re-employment rates of work-ers displaced by competition from imports are typ-ically low. The reason is not the nature of importcompetition. Rather, the explanation lies in themismatch between the characteristics of laid-offworkers and those of workers hired for new jobs.Workers to be laid off are not chosen randomly butselectively. In the United States, workers losing jobsare likely to be older and to have lower educationalattainment. In contrast, new jobs tend to demandyounger and better-educated workers. The patternsof selection of employees for redundancy and ofnew recruits are aimed at raising the efficiency ofthe U.S. labor force, and these patterns are likely toapply in China as well. Furthermore, the new jobsthat laid-off workers find tend to be lower payingthan the previous jobs.

In the United States, workers losing jobs becauseof competition from imports may experience along spell of unemployment and a significant dropin wages when they find a new job (Kletzer 2001).Arguably, in terms of spells of unemployment, thelosses suffered by Chinese workers losing jobsbecause of WTO accession are likely to be no lessthan those experienced by U.S. workers unem-ployed because of competition from imports.Indeed, they may be higher for three reasons. First,China, as a developing economy with a large poolof surplus labor, would have more job seekers


per vacancy than the United States. Added to this,workers losing jobs because of WTO accessionwould be competing with the large numbers ofworkers laid off from the state sector. Second, bar-riers to labor mobility in China serve to prolongunemployment between jobs. Third, the problemof a lower re-employment rate among older work-ers, who are likely to account for a substantial shareof employees selected to be laid off, is in theChinese context compounded further by the factthey that they have much lower educational attain-ment than that of younger workers. Because ofthe adverse impacts of the Cultural Revolution(1966–78) on secondary and higher education, alarge proportion of birth cohorts from 1955 to1965 (ages 37–47 in 2002) missed out entirely onpost-basic education and occupational training. Asa result, the difference in educational attainmentbetween younger and older workers in China isparticularly wide.

A preliminary hypothesis is that China’s WTOmembership is highly unlikely to have an identifi-able impact on the wages of skilled and unskilledemployees—an issue that dominates the discus-sion of the impact of imports from developingeconomies on U.S. workers (Collins 1998). Like-wise, a rise in the demand for unskilled laborbecause of a lowering of impediments to Chineseexports, such as textiles and clothing, is highlyunlikely to translate into a rise in the wages ofunskilled workers, as predicted by the standardinternational trade theory that proceeds on theassumption of full employment. In China, a devel-oping economy, the supply of unskilled workers farexceeds demand, and a rise in demand induced byWTO membership is likely to be small relative tothe excess supply. In fact, the wages of professionaland skilled employees have been rising more thanthose of unskilled workers in China, and the trendis likely to continue. This pattern runs contrary tothe expected change in relative wages, but it haslittle to do with international trade. Rather, it stemslargely to a correction of the highly compressedwage structure inherited from the planning period.This chapter proceeds on the assumption that theimpact of WTO entry on industry and services willprincipally be on jobs and that on farming will beon earnings.

For anyone studying the labor market impact ofWTO membership, the principal problem is that

this impact is intertwined with those of the twotransitions under way in the Chinese economy: thetransition from a planned to a market economyand the process of economic development that isseeing a transfer of labor out of farming. Both ofthese transitions involve job displacement and aredeployment of labor on a large scale. The laborrecruitment practices of the planned economy leftbehind a legacy of surplus labor in the state sector,which, forced by competition, has been sheddingthis excess baggage. In just six years, from 1996 to2002, the state sector (including enterprises, publicinstitutions, and the government) has lost about40 million jobs, or 37 percent of the total in 1996,and the trend is likely to continue for some years tocome. The argument is that job losses or job dis-placement arising from WTO membership wouldsimply be a part of the same phenomenon on alarger scale. Similarly, a reduction in income amongsections of the farming population, which is likelyto be one of the effects of WTO membership, hasfor other reasons been a common occurrence, espe-cially since the late 1990s.

China’s accession to the WTO in December2001, although a historical landmark, was just oneevent in a series of step-by-step openings of theChinese economy to international trade and for-eign direct investment since 1979, referred to as theOpen Door policy. From a historical perspective,the effects directly attributable to WTO accessionmay be seen as one chapter in the extensive story(i.e., ramifications) of the Open Door policy. Giventhe long time it has taken China to gain entry to theWTO, the event has been “expected” and “preparedfor,” at least to a degree. Some of the effects of WTOaccession have preceded the cause itself. This is par-ticularly true about the restructuring of industriesover the last few years and perhaps also about thedecision to convert marginal farming land intoforests. This situation is akin to the observationthat part of the relative decline in the wages ofunskilled works observed in the United States maybe attributable to employers securing restraint inwage demands by using the threat of develop-ing country imports—that is, the expectation ofheightened competition may have the same effect asthe actual event. Most of the likely impact of WTOmembership on the Chinese economy is attributa-ble to sharper competition in product and servicesmarkets. In this respect, increased competition

Coping with and Adapting to Job Losses and Declines in Farm Earnings in China 307

between domestic producers, which has risensharply since the mid-1990s, has the same effect asthat produced by heightened competition betweendomestic and foreign producers after WTO entry.

All this discussion leads to one implication andraises a question. The implication is that the adverselabor market impacts of WTO membership wouldbe neither novel nor out of scale with what theChinese economy has already experienced in recentyears, or more specifically since the second half ofthe 1990s. This implication suggests that observerskeep an open mind about the magnitude of theeffects, at least at the national level. One importantquestion is whether separate or special policies areneeded to deal with the adverse impacts of WTOentry. Arguably, the policy responses to job losses orfalls in farm incomes should be the same regardlessof their proximate causes. A justification for specialmeasures is that the adverse impact of the pre-scribed reduction in trade barriers after WTO entrymay be concentrated over a short time period andfall heavily on particular regions, such as the ruralareas of the already impoverished Northwest.

Social Protection and Adaptation

Schematically, the current social security schemesand policy measures that may be used to respond tothe adverse impacts of WTO accession on the labormarket (i.e., job losses and lower farm earnings inChina) fall into four categories, depending on thepolicy aim and the target (table 17.1).

Excluded from table 17.1 are the social securityschemes that are not directly related to a job loss or

fall in labor earnings, such as old-age pensions andhealth insurance. The category “social protection”covers various schemes differentiated by the targetof assistance and the conditions attached to theprovision of assistance, such as the requisite contri-bution record or insufficient means. These schemesdo not prevent job losses or declines in earnings;they operate after the event and provide no morethan partial compensation. Full compensation isruled out in order to maintain economic incentivesto remain employed and because of budgetary con-straints. The category “adaptation” covers a diverserange of schemes with the common feature of facil-itating individuals, localities, or industries to adaptto a new trade environment through labor rede-ployment and modification of the composition ofeconomic activities. Most of the schemes fallingunder this category have not been introduced todeal with the effects of WTO entry, but they lendthemselves to serving that purpose. A notableexception is the restructuring of various industriesover the last few years in the expectation of WTOaccession, such as those in the textile, steel, andvehicle industries. Some of the schemes, such aspublic works programs, contain elements of bothsocial protection and adaptation.

Two gaps in the schemes and measures(table 17.1) that would provide protection andfacilitate adaptation to the adverse impacts of WTOentry stand out. First, the schemes for households/individuals that fall under “social protection” varygreatly between rural and urban inhabitants, whichleaves some groups such as migrants unprotected.Second, among the schemes for localities/industries


TABLE 17.1 Classification of Schemes and Measures

Aim

Target Social Protection Adaptation

Households/individuals • Unemployment insurance • Job search facilities• Poverty relief • Retraining

• Assisted migration

Localities/industries • Schemes to assist poor • Development schemes for rural localities lagging regions

• Public works programs • Assisted industrial restructuring• Assisted conversion of marginal

forest land into forests


that fall under “adaptation,” there are no assistancemeasures targeted at towns and cities adverselyaffected by industrial restructuring. These meas-ures assume a special importance in the Chinesecontext because of impediments to labor migra-tion. The problem of towns and cities caught in anunemployment-and-stagnation trap has come tothe fore with the massive labor redundancies andlayoffs in the state sector since 1996. The problem isparticularly serious in medium and small townsheavily dependent on a few state enterprises thathave been forced to shed labor because of financialdifficulties. In such towns, job losses in a few enter-prises accounting for a substantial percentage oflocal employment can generate significant negativepecuniary externalities. Such externalities arisewhenever there is a combination of imperfect com-petition and increasing returns to scale. In theChinese context, the negative regional impact ofjob losses is further exacerbated by the arrange-ment whereby urban income maintenance schemesare largely financed at the city level. Cities with ahigher poverty or unemployment rate also tend tohave strained public finances and therefore mustreduce coverage or benefit levels. As described laterin this chapter, there is a rural assistance schemetargeted at poor localities, but that scheme is oflittle or no use as a cushion against the adverseimpacts of increased agricultural imports on farmincomes.

The framework employed in this chapter toanalyze social protection measures is structuredaround the following questions.

• Do the beneficiaries (individuals/industries/localities) of a social protection measure overlapwith those adversely affected by WTO entry? Ifyes, how significant is the overlap, and howeffective is the targeting?

• How does the measure address the adverseimpacts (e.g., assistance or facilitating adapta-tion)?

• How is the measure financed, and is thefinancing adequate to meet the measure’sobjectives?

• What effects does the measure have on incentiveand efficiency? In terms of efficiency, would analternative measure achieve the desired effect ata lower cost?

Instruments for Social Protection

Social protection schemes are intended to reducethe incidence of poverty and economic inequality.They provide a partial compensation for a job lossor fall in income, but on their own they do notfacilitate a redeployment of labor, which is the ulti-mate aim. Following the usual convention, suchschemes fall into two categories: social insuranceand social assistance (social safety nets).

China, which has both types of schemes, has amore extensive social insurance system than thosefound in most developing economies. This system,then, provides the Chinese government with twotypes of measures to compensate, partially, thelosers from trade liberalization. The differencesbetween the two sets of schemes arise from tworelated features: first, the conditions attached to theprovision of benefits (entitlement qualification)and, second, the method of financing. Under socialinsurance schemes, benefit payment is jointly con-ditional on the contribution record and on theoccurrence of specific contingencies, such as unem-ployment. The first condition makes social insur-ance similar to commercial insurance and limits itscoverage of the population. In China, as in manyeconomies, social insurance is confined to the laborforce in formal wage employment, and, unlike inother economies, is restricted to registered urbanresidents.1 Under formal wage employment, it iseasy to collect contributions in the form of a pay-roll tax and define and verify the state of unem-ployment, both of which raise difficulties in the case of informal wage employment or self-employment. The protective cover provided bysocial insurance in China is subject to two limita-tions. First, coverage is in practice limited to thelabor force in urban areas in formal wage employ-ment, which represents 14 percent (105.6 million)out of the total labor force of 737.4 million(National Bureau of Statistics 2003a).2 Second, thecontribution condition imposes a limit on unem-ployment compensation. In most schemes, theduration of compensation is linked to the durationof previous contributions and is subject to a maxi-mum, which in China is two years. The implicationis that unemployment insurance cannot be the solemainstay of a comprehensive social protectionsystem; it has to include a needs-based incomemaintenance scheme. Financed by contributions,


social insurance resembles commercial insurance,but it differs in two respects, which help to extendits role in mitigating risks:

1. Participation is invariably mandatory for a des-ignated section of the population, such as urbanwage employees, which avoids adverse selectionwhereby low-risk groups opt out, thus raisingthe percentage of higher risk groups in the cov-ered population.

2. The relationship between benefit and contribu-tion in social insurance is typically looser thanin commercial insurance and involves redistri-bution in favor of vulnerable groups.

In contrast to social insurance, social assistanceis noncontributory and thus far removed fromcommercial insurance. A key distinguishing featureis that it can, in principle, provide protection toeveryone in need in the relevant population. Nor-mally, social assistance is available to all peoplewhose resources fall under a specified minimum—for example, an income that falls under the povertyline. Furthermore, the benefit is available as long as

the means test and some supplementary conditionare satisfied.3 China has several social assistanceschemes, but each covers only a particular sectionof the population, which leaves some glaring blindspots in coverage (this is discussed later in thischapter). Table 17.2 presents the salient differencesbetween social insurance, with reference to unem-ployment insurance, and social assistance.

Though overlapping in aims, social assistanceand social insurance schemes are as much com-plements as they are substitutes. Developed andcomprehensive social security systems invariablyinclude both types of schemes. Does a low-incomedeveloping economy such as China need both typesof schemes, especially social insurance that coversonly a minority of the labor force, and usually thatin regular wage employment? A low income doesnot constitute a barrier against the institution andfunctioning of social insurance, because benefitsare tied to contributions that are usually a percent-age of payroll. The principal advantage of socialinsurance is that, when well designed, it generatesmost if not all of the funds needed for its opera-tion, which matters when possibilities for raising


TABLE 17.2 Differences between Social Insurance and Social Assistance

Social Insurance Social Assistance (Safety Nets)

Financing

Benefits

Coverage

Impact on poverty andeconomic inequality

Source: The author.

• Self-financing, at least partially ifnot totally

• Easy to combine with commercialinsurance to cover additional risks

• Can be tailored to serve a range ofobjectives

• Typically, benefits more generousthan those under social assistance,but only available for a limitedduration

• Limited to those who can makethe requisite contributions. Bydesign, such schemes cannotprovide universal coverage

• Provides protection againstpoverty to participants and thosewith an adequate contributionrecord

• Reduction in inequality limited bycoverage

• Must be financed from generalgovernment revenue, which posesproblems in low-income localitieswith a tight budget

• Usually set a level just sufficient toavert hardship, in some cases onlysevere hardship

• Restricted to those with insufficientincome and savings to cover basicneeds

• In principle, can cover the wholepopulation, depending on thebudgetary constraint

• Can serve as social safety net toensure protection against poverty

• Limited reduction in economicinequality constrained by targeting


revenue through alternative taxes are limited. Thecontribution condition that limits the coverage ofsocial insurance also makes it attractive from afiscal point of view. In contrast, social assistanceschemes are, by design, not self-financing. Theprincipal argument in favor of introducing andextending unemployment insurance is that it pro-vides a first line of defense against poverty andreduces the task of poverty alleviation left for socialassistance schemes. The latter is of particularimportance when the coverage of social assistanceis limited by budgetary stringency, as it is in Chinaand other developing economies.

Overview of Chinese Income Maintenance Schemes

Table 17.3 presents an overview of various incomemaintenance schemes in China relevant to copingwith job losses or declines in income stemmingfrom WTO accession, excluding old-age pensionsand disability compensation.

The land plots allocated to households areincluded as an income maintenance instrument intable 17.3 because of the importance accorded to

them in providing income security for the ruralpopulation. Three features characterize Chineseincome maintenance schemes. The first is the starkdichotomy between the urban and rural popula-tion. At present, no scheme covers both the popula-tions, which are defined differently in China thanelsewhere.4 Second, the schemes aim to do no morethan alleviate severe poverty. In the two urbanschemes, the benefits are set at a flat rate either justabove or at the local poverty line. The two ruralassistance schemes are narrowly targeted, thoughin many cases inaccurately. The third feature isa high degree of decentralization. With the excep-tion of the “Assistance to Poor Rural Localities”program, all schemes are in the first instancefinanced by low tiers of government: cities forthe two urban schemes and villages for the “Assis-tance to Poor Rural Households” program. Highergovernment tiers contribute toward the cost of theschemes. But there is as yet no settled formula forcost sharing, and lower tiers of government end upbearing a share of costs that is too high to be equi-table and consistent with the sustainability of theschemes.

TABLE 17.3 Income Maintenance Schemes in China

Urban Schemes Rural Schemes

Unemployment Minimum Living Assistance to Poor Assistance to Poor Allocation of a Insurance Standard Localities Households Land Plot

Coverage

Benefits

Financing

Adminis-tering agency

Source: The author.

Urban, formalwage employees

• Cash allowance

• Health insurance

• Free orsubsidizedservices

Contributions and residualgovernmentfinancing

City labor andsocial securitybureaus

Registered urbanresidents

• Cash allowance

• Health insuranceto cover largemedical costs

Central andterritorialgovernments

City civil affairsbureaus

Designated rurallocalities

• Investment insocial services,infrastructure,and economicactivities

• Subsidiestargeted athouseholds

Central andprovincialgovernments

Civil affairs bureaus

Usually restricted to very poorhouseholds andhighly variable across localities

Benefits in cash andkind

Locally financed

Villages

Long-standinginhabitants ofthe locality

A plot of land

Villages


The rural–urban segmentation that casts a longshadow on the Chinese social security system hasseveral facets. The urban and rural schemes notonly are mutually exclusive, but also are structureddifferently. For the rural population, there are noequivalents of unemployment insurance (UI) andthe urban Minimum Living Standard Insurance(MLSI), although the latter has begun to change.The explanation for confining UI to urban em-ployees is not that the rural labor force, being self-employed, does not face the risk of unemployment.Although a majority of the rural labor force isengaged in the cultivation of its own land plots, asubstantial percentage is employed in farming andnonfarming activities, and most of these are wage-employed. Over a third (35 percent) of the 490 mil-lion strong rural labor force is engaged full time innonfarming activities, which says that in Chinarural is far from synonymous with agricultural.Leaving aside household units, town and villageenterprises (TVEs) employ 133 million workers, or27 percent of the rural labor force and far morethan the total employment in the state sector(National Bureau of Statistics 2003a). The risk ofunemployment facing TVE employees is just as

palpable as that facing their counterparts in urbanenterprises; the labor retrenchment in TVEs inrecent years is evidence. Additionally, many of theTVE employees are just as exposed to the impact ofinternational trade as their urban counterparts.This being the case, the direct impact of WTO entryon rural China would not be confined to farming—it also would extend to rural industry. The impactof WTO entry on industry would straddle bothurban and rural areas.

The MLSI, a social assistance scheme based onmeans, marks another major difference betweensocial protection in urban and in rural areas. Thescheme, which was introduced in 1997 and was bythe middle of 2000 established and functioning inall towns and cities, aims to provide assistance to allregistered urban residents with a household percapita income below the local poverty line. Thenumber of beneficiaries under the scheme has risensharply since inception, indicating improved cover-age. The scheme marks a major change in the his-tory of social security in China, in that until itsintroduction all social assistance was confined to asmall section of the population characterized bythe “three nos”: no ability to work, no savings, and

BOX 17.1 Advantages and Disadvantages of Decentralization

Decentralization in the Chinese social securitysystem offers both advantages and disadvan-tages.

Advantages:• Decentralization allows flexibility to account

for local variations and gives local govern-ments the freedom to take initiatives. In fact,some local government initiatives have laterbeen adopted as national policy. For exam-ple, the Minimum Living Standard Insurance(MLSI) scheme for urban areas, now a nation-wide scheme, was first implemented in a fewcoastal cities. A uniform system for all ofChina runs the risk of being determined bythe lowest common denominators.

Disadvantages:• Municipalities and the tiers for the organiza-

tion of rural social security are, in manyinstances, too small to provide sufficient riskpooling for maintaining the financial integrityof schemes. Decentralization is associatedwith variations in standards of provision thatmay go beyond the socially acceptable

differences. Furthermore, such variations mayimpede the transferability of social securityentitlements (such as unemployment bene-fits) and thus impede labor mobility. Non-transferability is the principal barrier inbringing migrants under the social insurancecover.

• Decentralization accentuates disparities insocial security provisions. Localities with ahigher poverty level or a higher percentage ofthe needy in the population also tend to havemore strained public finances and thus areforced to reduce coverage or the benefitlevel.

Planned ReformThe pilot program for improving the urban

social security system, which is currently beingtested in Liaoning Province, recognizes theproblems created by decentralization. One of itsaims is to upgrade the level of budgeting forsocial insurance from the municipal to theprovincial level in selected localities, startingwith the old-age pension scheme.

no relatives to depend on. Under the planned eco-nomy system, those physically able to work wereexpected to work and were provided with a job,which the government was able to do by virtue ofits control of the economy. The introduction of theMLSI is underpinned by the government’s recogni-tion that it cannot provide a job guarantee to all ina market economy. With the combination of MLSIand UI, the urban population has all the ingredi-ents of a comprehensive insurance against poverty,including that caused by the adverse impacts ofWTO accession. As outlined later in this section,both schemes suffer from deficiencies, but thosestem largely from problems in the design andimplementation of the existing schemes. The lackof appropriate schemes does not create gaps insocial protection for urban households, but it cer-tainly does for the rural population.

Although the rural population benefits fromnumerous antipoverty schemes targeted at poorlocalities and at poor households (table 17.3), itdoes not enjoy the same degree of protection as theurban population. The program offering assistanceto poor localities is a big one that dates back to1994, when the government launched the so-calledthe “8-7” Plan.5 The plan targeted the 592 poorestcounties (designated as national poor counties)selected out of 2,000-plus rural counties. Thesecounties, many of which were in the westernregion, contained, it was estimated, more than70 percent of the 80 million people who remainedpoor. These counties were generally remote, char-acterized by backward socioeconomic conditions6

and poor natural endowments—that is, they wereareas in which geographic targeting is appropriate.The plan was financed entirely by the central gov-ernment and backed by a multitiered administra-tive structure from the central government down tothe counties. The principal instruments of assis-tance were tax reduction, financial grants, and asocioeconomic development program. The princi-pal thrust was toward building infrastructure anddeveloping the local economy rather than towardproviding poor households with assistance. Theprogram came to an end in 2000 and was replacedwith a new one with similar aims and mode ofoperation, but targeted differently. Putting aside theissue of the appropriate territorial units for target-ing,7 these programs, though valuable in reducingsevere rural poverty, are of little consequence, if

any, in cushioning the adverse impacts of WTOmembership on income from farming. In general,the poor rural localities are remote and close toautarky. Market transactions with the economyoutside the locality are sporadic and of low value.China’s accession to the WTO would have littledirect impact on such localities and their inhabi-tants. However, the program could be adapted tocope with the adverse impacts of WTO accessionon rural areas by redefining its scope.

Also in rural areas is a wide range of assistanceschemes that are targeting poor households. Theoldest of these is the so-called Five GuaranteesScheme, covering food, clothing, shelter, health care,and funerals. The recipients of assistance under thescheme have been mostly elderly persons withoutmeans and relatives.8 Therefore, the scheme is oflittle or no relevance in providing assistance torural inhabitants adversely affected by WTO entry.Other rural assistance schemes cover a wider cate-gory of the poor than just the indigent elderly.These schemes have two features in common. First,they are highly decentralized, mostly organized andfinanced at the grass-roots level of villages. Second,associated with this high degree of decentralization,the structure of the scheme, funding, eligibility forassistance, and details of assistance vary widely. Incontrast to the urban schemes (UI and MLSI), therural schemes lack the underpinning of a nationalframework and are, as a result, far more variable incoverage and benefits. In localities, the spectrumruns from well-functioning schemes similar to theurban MLSI to no social protection scheme at all.

Decentralization has two related consequences,both of which erode the degree of protection pro-vided by the rural schemes targeting households.First, localities with a higher incidence of povertytend to have poorly financed and organizedschemes, exactly the opposite of what is needed.Second, decentralized schemes are not able to copewell with correlated risks that adversely affect asubstantial percentage of households in the locality.Competition from imports is an example of such arisk when households in a locality have similar oridentical cropping pattern. The second feature ofrural assistance schemes is that they are generallydesigned to provide assistance to those in extremepoverty only. The rural population is expected torely on kin and community in case of all but direcontingencies. In sum, the rural areas lack schemes



that can directly address the adverse impacts ofWTO accession in the form of unemployment anda drop in farm earnings.

Land Plot as Insurance

One of the two rationales for the sparse and highlyselective social protection cover in rural areas isthat each rural household has been assigned a plotof agricultural land that serves as a floor to house-hold income, a feature that is particular to Chinaand a few other economies. The other is the highcost of introducing a social security regime compa-rable to the one in urban areas. In rural areas, thereis a limited capacity to collect taxes and social secu-rity contributions, which is a common problem indeveloping economies. Agricultural land in China,compared with most developing economies, is veryequally distributed. The removal of the adverseeconomic incentives that marred the collectiveeconomy and the provision of each rural house-hold with a piece of land have together contributedto the dramatic drop in the (official) headcount ofthe poor, from 250 million in 1978 to 70 million in1994, on the eve of the 8-7 Plan.9 Landlessness andrural poverty do not go together in China, inmarked contrast to the prevalent pattern in devel-oping economies. However, in the present contextthe pertinent question is how much of a cushion doland plots provide against job losses in rural areasand a decline in farm income in general and thosearising from China’s accession to the WTO.

Controlling for the impact of input and outputprices on farm earnings, income per capita from aland plot depends crucially on the land area perrural inhabitant, land quality and location, andtechnology. The constraint on income per capitaimposed by the land–rural population ratio can bepushed back through investment in transport andother infrastructure and through technologicalupgrading, but it cannot be eliminated. The land–rural population ratio acts as a limit on the incomesecurity provided by land. Because of China’s hugepopulation, the cultivated area per rural inhabitantis slightly less than a fifth of hectare (0.19).10 Thecultivated land per rural person has been fallingsteadily since the distribution of land to householdsin the early 1980s because of the combination ofrising population and various impediments to

rural–urban migration. This trend may continuefor some time.

After a sharp drop in the 1980s, the percentagegap between the rural and urban income per capitahas been widening particularly rapidly in recentyears. The implication is that the level of economicsecurity provided by land plots, as measured byaverage farm income per rural inhabitant, has beenfalling relative to per capita income. Added to this,the level of income security land plots provide torural households has been highly variable acrosslocalities. The distribution of land to rural house-holds takes place at the grass-roots level of villages.As a result, the average area of land plot per house-hold varies across villages even in close vicinity.Land plots have served to reduce the incidence ofrural poverty but they have not been sufficient toprevent poverty, not even extreme poverty. Most ofthe rural population with a per capita income ofless than 635 yuan per year—the official povertyline for rural inhabitants—has land.

It is also important to point out that not allrural households have a land plot, and the percent-age without land has been rising. Two examples aredescribed here—one from a prosperous region andanother from a poor region. In rural areas ofsouthern Jiangsu, one of the richest rural areas inChina which covers 12 counties11 known as Sunnanand has highly developed TVEs, a survey con-ducted in 1998 revealed that about 10.3 percent ofhouseholds no longer possessed land. Reasonsincluded the conversion of agricultural land tononfarm uses and the relinquishment of land byhouseholds in order to opt out of liability for theagricultural tax. Furthermore, among the 89.7 per-cent of households with a land plot, another 3 per-cent planned at the time of the survey to give uptheir land plot in the near future. However, even inpoorer rural regions not all rural households have aland plot. A survey of 1,500 households conductedin 1999 and drawn from 15 nationally designatedpoor counties in Gansu and Inner Mongolia foundthat about 3 percent of rural households for vari-ous reasons had no land. Most of the rural house-holds without land had in the past given up landvoluntarily and for good economic reasons. Theimportant point is that on giving up land thehouseholds are not included in another incomemaintenance scheme.


How much of a cushion do land plots provideagainst a decline in household income from, forexample, a job loss from a TVE or a fall in grainprices because of competition from imports? Landplots are not transferred to rural households inperpetuity but are leased for a definite period ona renewable basis.12 In principle, the land lease canbe sold or transferred, freely, but, in practice, it issubject to severe restrictions. One serious conse-quence of the limited tradability of land plots isthat they cannot be used as collateral for loans.Land plots provide rural households with theoption of an income generated by deploying house-hold labor on their own land plot. It is this incomethat determines the value of insurance provided byland plots. In the event of the loss of a job in thenonfarming sector, households have the option ofcompensating for the loss of income by redeployingthe now unemployed labor to their own land plot.13

The important feature of this scenario is thatincome obtainable from the land plot remainsunaffected by the adverse shock. However, whenthe decline in household income is the result ofcompetition from cheaper agricultural imports,household land plots provide no cushion againstthe fall in income. Such competition reduces theinsurance value of land. Because the competitivethreat posed by agricultural imports varies acrosscommodities, the extent of reduction would de-pend on the cropping pattern and the possibilityof changing it. More specifically, in terms of theimpact of WTO accession on farming, the North-east and the Northwest would be losers and the Eastand Southeast would be gainers.

Anomalies in the Classification of the Population

As noted earlier, under the current social securityarrangements any assistance that a person mayreceive after a job loss or reduction in earnings, forexample, stemming from WTO entry would de-pend crucially on whether the person is living in anurban or rural area. The categories urban and ruralare defined differently in China than elsewhere.This section will outline the anomalies created bythe combination of the control on populationmigration and the rules governing household regis-tration and will point to the implications of these

anomalies for the incidence of the adverse effects ofChina’s accession to the WTO.

Urban is both a spatial and a demographic cate-gory. According to the usual international usage,urban in its spatial sense denotes localities with ahigh population density and heavily dependent onindustry and services as sources of income. In itsdemographic sense, urban denotes the populationresident in an urban locality—that is, the spatialdivision urban–rural coincides with the demo-graphic division of the population into urban andrural, and the two are interchangeable. In contrast,in China the spatial and the demographic divisionsdo not coincide, and the urban and rural popula-tions, as officially designated, are intermixed andnot mutually exclusive as they usually are. Thissituation creates anomalous divisions in the popu-lation and is associated with arbitrary exclusions orinclusions in the coverage of the social securityschemes. Chinese cities are divided into urbandistricts (shi qu) and rural counties (xian). Similarly,the population is classified by personal registrationstatus as agricultural (nongye) and nonagricultural( fei nongye). The spatial and the demographic divi-sions overlap, but only partially, producing thefollowing anomalies (National Bureau of Statistics2001):

• Of the 296 million people with nonagriculturalregistration (thus officially urban) living withinthe boundaries of 663 cities and towns in 2000,14

42.7 percent (126 million) were resident in ruralcounties, rather than in urban districts, andwere covered by social security schemes con-fined to the urban population.

• Just as anomalous, 40.6 percent (116 million) oflong-term residents in urban districts carriedagricultural registration and were thus regardedas part of the rural population, even thoughmost of them no longer had any relationshipwith farming.15 They were still classified asagricultural because the occupational and resi-dential shift from farming to industry and serv-ices does not automatically lead to a change inregistration from agricultural to nonagricultural.These long-term inhabitants of urban districtsare excluded from social security schemes forthe urban population on the grounds of theirregistration status.

Thus the official designation urban in Chinadoes not denote the usual place of residence andwork, such as urban districts of cities. Rather, itcarries the label nonagricultural ( fei nongye) on thepersonal register (hu kou). Urban is a personalattribute or status rather than a spatial designa-tion. A striking example is provided by a compara-tively common anomaly in China where somemembers of a household in an urban district areregistered as nonagricultural and therefore regardedas urban, while other members of the same house-hold are regarded as rural on the grounds of theirregistration as agricultural. Thus a householdreceiving the MLSI allowance may get less than theamount needed to bring its members up to thepoverty line, as prescribed, because some of itsmembers are not registered urban residents and sodo not qualify. A frequent cause of such anomaliesis marriage across the registration line. Similarly,migrants do not qualify for the MLSI allowance.Urban registration, if not inherited, is difficult toacquire, especially for large cities such as Beijingand Shanghai.

The control on labor migration and the rulesgoverning assignment of the labels nonagriculturaland agricultural to individuals produce two oddi-ties at the aggregate levels of the labor force and thepopulation. These oddities are important becausethey affect the pattern of incidence of impacts ofWTO entry and also the coverage of social protec-tion schemes. The first oddity is that whereas thegross domestic product (GDP) ratio of farming is amere 15.8 percent,16 similar to that in middle-income economies, the percentage of the laborforce reported to be engaged in farming is almostthree times higher, 47 percent, which is similar tothat in low-income economies. Put together, thetwo percentages imply that the output per capita infarming, given by the ratio of its GDP share to itsemployment share, is less than a third of the aver-age, or a mere 21 percent of that in the nonfarmingsector.17 A GDP share of farming that is a smallfraction (say less than half) of its labor force shareis a common feature of developing economies andcorresponds to a wide gap between nonfarm (orurban) and farm (rural) per capita incomes inthose economies. But with a GDP share of farmingclose to one-third of its labor force share, whichimplies that the urban GDP per capita is about fourand a half times the rural GDP per capita,18 the gap

between the two ratios in China is wide even for adeveloping economy. What factors explain the verylow relative labor productivity in farming? Twoexplanatory factors are possible: low land produc-tivity and a high labor-land ratio. The first does nothave much explanatory power in the Chinese con-text. As in other East Asian economies, farming inChina is intensive, and productivity per unit ofland (e.g. hectare) is high by international stan-dards. There is scope for raising land productivityfurther through the adoption of best-practice tech-niques, while keeping the labor–land ratio constant.But the principal explanation of the low labor pro-ductivity in farming lies with three impediments tothe transfer of labor out of farming to nonfarmingactivities. The first impediment to rural–urbanmigration is the household registration system,which has diminished in rigor but not disappeared.The second impediment is the land tenure systemin which rural households receive little or no com-pensation for giving up user rights on the allocatedland. This arrangement is in effect tantamount to atax on leaving farming completely. The third is thebias in favor of industry, which has tended to raisethe investment cost of transferring labor into non-farming activities and has therefore led to a higherland–labor ratio than it would otherwise be.

An important feature of the administrative con-trol on rural–urban migration is that the impedi-ment to the transfer of labor out of farming variesaccording to locality. It does not pose a seriousimpediment to the transfer of labor out of farmingin localities with favorable conditions for the devel-opment of TVEs, such as a relatively high per capitaincome and educational level and good transportlinks. In such localities, most of which are in a fewcoastal provinces, controls on migration to citieshave, arguably, spurred the growth of TVEs. In theseareas, the result has been a substantial weakening ofthe link between labor outflow from farming andout-migration from the countryside. However, inlocalities where conditions for the growth and pro-liferation of TVEs are comparatively unfavorable,the control on labor migration bears down heavilyand thus keeps a higher percentage of labor in farm-ing. Such localities, which are poorer than thosewith developed TVEs, tend to be found in the inte-rior provinces. There, migration control functionsas a high barrier to the redeployment of labor outof farming. Given that, on average, income from


farming is a small fraction of that from nonfarmingactivities, any measure that keeps a higher percent-age of labor in farming in effect creates a trap thatperpetuates low incomes and poverty.

What are the implications for the incidence of theadverse impacts of WTO membership? First, thelarger the percentage of the labor force engaged infarming, the larger is the percentage of the popula-tion exposed to the risk of falling income due tocompetition from agricultural imports and there-fore the more widespread are the adverse impacts.Second, among the localities reliant on crops espe-cially vulnerable to imports, the adverse impactswould be variable. In particular, the impacts wouldbe more severe in localities with less-diversifiedeconomies and those with an already low per capitaincome. The two tend to be correlated because less-diversified rural localities tend to be more depend-ent on farming activities that, on average, generate alower income per capita than do nonfarming activi-ties. However, low income per capita is also impor-tant in its own right because that heightens thevulnerability to falling into poverty after even aminor decline in income. On these criteria, theNorthwest is particularly vulnerable. The region isreliant on crops at a disadvantage against imports,such as corn, wheat, and coarse grain; it has a verylow income relative to the national average, and itsrural economies are less diversified. Moreover, theshortage of water in the region limits the possibilitiesof changing the cropping pattern. The fact that thereduction in barriers to agricultural imports wouldbe uneven and heavily concentrated in some regionsthat are already poor suggests the need for introduc-ing assistance schemes that target particular regions.

The second oddity about the Chinese economy isthe wide divergence between the sectoral distribu-tion of the labor force and the division of the popu-lation into agricultural and nonagricultural (orurban and rural). The percentage of the labor forceengaged in farming is 47 percent, but the percentageof the population/labor force designated as rural is64 percent, 17 percentage points higher. The diver-gence has widened over the reform period becauseof the rapid growth of TVEs and the correlatedexpansion of nonfarm employment in rural areas.The divergence matters because the coverage ofsocial security schemes is determined by the sta-tus division nonagricultural–agricultural, not thesectoral distribution. As a result, two individuals

working in the same employment unit may have dif-ferent social security cover simply because of theirdifferent designations in the household register.

Closing Loopholes in SocialProtection

What feasible options are there for reducing, if notabolishing, the arbitrary anomalies in the coverageof social security schemes? An integrated socialprotection system that abolishes the urban–ruraldistinction as a consideration for the provision ofassistance would eliminate the arbitrary anomaliesaltogether and immediately. Such an integrated sys-tem, however, although highly desirable on equitygrounds, is not feasible for the foreseeable futurefor financial reasons. The tax system such a systemrequires is a decade into the future. A piecemealand gradual elimination of the anomalies is theonly possibility. Assuming that the fiscal constraintremains binding, piecemeal reforms along the fol-lowing three lines appear feasible:

1. extending contributory social insurance schemesto regular wage employees regardless of personalregistration status, which is losing significancebecause of changes in the household registrationsystem

2. broadening the coverage of means-tested socialassistance schemes targeting poor households,including the introduction of an integrated so-cial assistance scheme covering urban and ruralpopulations in selected localities with healthypublic finances

3. overhauling the scheme, financed by the centralgovernment, to assist poor rural localities inorder to extend coverage to localities seriouslyharmed by competition from agricultural im-ports, and introducing a similar scheme to assisttowns and cities caught in a spiral of decline orin a stagnation trap.

The first would involve little, if any, extra expen-diture, because employees brought under the socialinsurance net will be making advance contribu-tions toward benefits. Such an extension will belargely limited to regular wage employees becauseof difficulty in collecting contributions and defin-ing unemployment. The main groups broughtunder the umbrella of unemployment insurance


would include migrants in regular employmentand regular wage employees of TVEs. Some citiesand provinces already include these groups in so-cial insurance. For example, Guangdong Provinceextends social insurance to migrants.

The second recommendation covers both theMLSI for the registered population and the assort-ment of social assistance schemes in rural areas. Achange in the coverage of MLSI to include all long-term urban residents regardless of their registrationstatus will remove glaring anomalies in the way thescheme currently functions. The suggested changein the coverage will mean a greater expenditure, butit is likely to be only a small fraction of the currentexpenditure on the scheme. In rural areas, the pri-ority is to institute a national framework for theprovision of social assistance to poor households inorder to set a minimum standard and put in place amechanism to ensure that at least the minimum isreached. Some provinces are taking steps to abolishthe urban–rural segmentation in social security. Forexample, Zhejiang Province has decided to extendthe MLSI scheme, which had been confined to itsurban population, to its rural population as well.Similarly, Shanghai is planning to introduce anintegrated social security system for the wholepopulation—both urban and rural.

The third recommendation is based on theassumption that an assistance scheme should targetparticular localities, thereby supplementing thesocial assistance scheme that targets poor house-holds. As noted earlier, the current scheme to assistpoor localities is of little use, if any, in assistingrural localities adversely affected by WTO acces-sion. Furthermore, there is a strong case for intro-ducing a similar scheme for towns and cities caughtbetween the pincers of high unemployment andpoverty rates, on the one hand, and strained publicfinances, on the other.19

The inclusion of migrants in social insuranceand MLSI is of particular importance, becauselabor migration is one of the principal ways ofadapting to changes in the structure of the econ-omy, including that induced by trade liberalization.In the Chinese context, migrants form a heteroge-neous group, but they differ in their length of stay.Some are visitors or short-term migrants, and somehave resided in China for a substantial length of time.Six months is the cut-off period used in Chinesestatistics to distinguish between short-term and

long-term migrants. Correlated with this, migrantsdiffer in their official status. A small percentage ofmigrants receive a new household registration andbecome indistinguishable from the local popula-tion. A proportion of them, usually long-termmigrants, have temporary household registration.One set of social security schemes that covers allmigrants regardless of differences among them isneither appropriate nor feasible. That wouldrequire an integrated social security system cover-ing both the urban and rural population, which, asargued earlier, is not realizable is the foreseeablefuture. There is no doubt that for the purposes ofsocial security long-term immigrants should beregarded as part of the urban population. However,for short-term migrants, it is far from obviouswhether they should be regarded as part ofthe urban or the rural population. This decisiontakes on a special importance in China, because,according to the current official practice, theurban poverty line, 1,800 yuan per person per year,exceeds the rural poverty line, 635 yuan, by a verywide margin.20 Therefore, a person who is pooraccording to the urban poverty line would be welloff by the rural one.

The disadvantageous position of migrants isassociated with the control on population migra-tion, especially between rural and urban areas. De-pending on the locality, immigrants are restrictedto jobs and occupations that are not attractive tolocal residents. Furthermore, immigrants do nothave the same access to public facilities, such ashousing and schools, as local residents. With a fewnotable exceptions, migrants in cities are excludedfrom social insurance.21 The notable exceptionsare Guangdong Province and, to a lesser degree,Shanghai. The former treats migrants on a par withthe local population for the purposes of socialinsurance, and the latter has begun to includemigrants in particular social insurance schemes.The exclusion of migrants in cities from socialinsurance is not restricted to migrants from ruralareas, but also extends to migrants from other citiesbecause entitlements to social insurance benefitsare not portable across cities. This arrangementreduces the benefits migrants, including those fromanother city, can derive from participating in socialinsurance schemes.

The official control on migration remains, butits rigor has diminished considerably. According a


recent State Council circular, migration to countyand some small or medium-size towns has beencompletely freed, in principle if not yet in practice.Moreover, the last few years have seen the introduc-tion of assisted migration out of rural localities thathave little or no prospect of emerging from poverty.However, while the focus of government policy anddiscussion has largely been rural–urban migration,urban–urban migration is assuming increased im-portance with the rise in urban unemployment andlayoffs.

Two aspects of the participation of migrants insocial insurance schemes are the collection of con-tributions and the disbursement of benefits. Theproblems raised by the first are, in principle, thesame for migrants and residents. The only differ-ence is that, compared with residents, migrants aremore likely to be in jobs outside the coverage ofsocial insurance schemes, such as casual or self-employment. The extension of social insuranceschemes to the labor force in the informal sector,which includes not only migrants but also increas-ing numbers of local residents, would pave the wayfor wider coverage of migrants. But the aim shouldbe not only collecting contributions but also dis-bursing benefits in return for contributions. Unlikecollection, disbursement raises issues that can besolved only through some fundamental changes inthe organization of social insurance. Aside fromimmigrants who receive a change in residence sta-tus, the rest are by definition temporary residents.This definition poses a particular problem forsocial insurance schemes, such as old-age pensionsand unemployment insurance, in which contribu-tions and benefits are separated in time. The prob-lem arises because current social insurance schemesare largely organized and financed at the level ofcities. Arrangements for the transferability of bene-fits between cities even in the same province arenonexistent or rudimentary. An integration ofsocial insurance at the provincial level would makebenefits transferable within a province, and itwould go a long way toward addressing the prob-lem of transferability because most migration isconfined to one province. Such integration isplanned, but it will take some time to implement.The portability of social insurance benefits alsomakes sense between urban areas that have com-parable social insurance schemes, but it begs thequestion of relevance in the case of rural–urban

migrants because of the absence of social insuranceschemes in rural areas.

The realistic options for extending social insur-ance to migrants have to take into account twoconsiderations. First, restrictions on rural–urbanmigration, though considerably relaxed, are likelyto stay for some time, especially for large cities. Sec-ond, urban and rural social security systems arelikely to remain segmented for some time. It wouldthen be inappropriate to apply one formula to allmigrants. It is important to distinguish betweensocial insurance schemes that (1) cover all migrantsand (2) are confined to particular categories ofmigrants, such as those from other cities. The firstset of schemes covers contingencies that seriouslyreduce a person’s capacity to work such as physicalinjury and major illnesses. The implication is thatthe priority should be to extend disability compen-sation and insurance against large health care expenditure to all migrants. In these two sets ofschemes, benefits are not deferred but run concur-rently with the payment of contributions. The issueof the portability of benefits does not arise. Thesecond group of schemes includes old-age pensionsand unemployment compensation. These schemesinvolve deferred benefits and raise the issue ofportability, which can be solved through a clearingmechanism. These are suitable for extension tocity–city migrants, because all cities are covered bysimilar social insurance schemes.

Education and Training

Between 1978 and 2000, the portion of the laborforce employed in the primary sector, most ofwhich is farming, fell at an average rate of over1 percentage point per year. This rate, althoughtwice that in the planned economy period(1952–78), would have been even higher except forthree factors: (1) the household registration system,(2) the land tenure system that provides little or nocompensation for giving up use rights on allocatedland, and, (3) the bias in favor of industry, whichhas tended to raise the cost of transferring laborfrom farming to nonfarming activities. However, asnoted earlier in this chapter, all these factors areundergoing radical changes. Taking all the changestogether, the implication is that the transfer oflabor out of farming over the 20 years to 2020would be at least as rapid as that seen over the



20 years to 2000. Added to this, driven by a moreopen international trade regime spurred by WTOmembership and economic growth, the Chineseeconomy is likely to experience a more rapid redis-tribution of labor between occupations and sectors.In the Chinese context, the methods for adapting tothe changes and reducing their cost in terms po-verty and increased economic inequality fall intothree broad categories:

1. improving the employability of the labor forceby raising the general level of education andskills, especially in the rural population

2. facilitating the growth and proliferation of pri-vate enterprises

3. encouraging the adaptability of existing enter-prises.

The remainder of this section focuses on the firstpoint, education and skills. The literacy rate inChina is substantially higher than that in manylarger developing economies, especially low-incomeones. However, the literacy rate is no longer anappropriate indicator of human development, norare low-income economies relevant comparators inthe field of human development for China. Thanksto the high growth rate of its per capita income after1978, China graduated to lower–middle-incomestatus at the end of 2000. The appropriate referencegroup for setting targets for human developmentand social security is not the present-day low-income economies but the middle-income ones.In terms of the average number of years of basicschooling for the current cohort of children, Chinalags behind many Asian countries, such as theDemocratic People’s Republic of Korea, Malaysia,Thailand, and Sri Lanka, as well as the Republic ofKorea, Singapore, and Japan. Furthermore, as Chinanarrows the technology gap with developed econ-omies, the educational attainment of the labor forceis assuming growing importance in determining thegrowth rate of the economy.

In China, nine years of schooling for all childrenhas been a policy goal since 1986, but it has not yetbeen attained. The target is largely achieved in theurban areas, but is far from realized in rural areas,which is home to 64 percent of the population andan even higher percentage of those in the 15–16 agecohort. In the western and central regions, basic

education in many rural localities falls far short ofthe prescribed nine years of schooling—in fact,about 26 percent of primary schools have a cycle ofless than six years. Moreover, a significant percent-age of children in poor localities drop out beforecompleting the full cycle. The tenth five-year plan(2000–04) envisages that, by the end of 2005, atleast 90 percent of those leaving school will havehad nine years of education. The target date forachieving universal basic education (UBE) is set for2010. However, according to the government’s plan,the UBE does not mean a uniform length of school-ing for all students. Depending on local financialresources and organizational capability, the UBEwill actually mean less than nine years of schoolingin some localities. For the poorest 5 percent of thepopulation, all in the western provinces, the UBEwill mean no more than three to four years ofschooling.

With respect to post–basic schooling, the periodsince 1995 has witnessed a sharp rise in the grossenrollment rates in secondary schools (includingvocational and regular schools) and in tertiaryinstitutions. This trend is expected to continue. Thetenth five-year plan (2000–04) envisages thatalmost 100 percent of children in cities and townswill go on to senior secondary and vocationalschools after nine years of basic education. It isexpected that by 2005 the percentage of the relevantage group in a tertiary institution would rise to15 percent, from less than 10 percent in 2000. Thesetrends will help to raise the average number ofyears of education to the level in East Asia, but theyalso will widen the gap between urban and ruralareas, which has some adverse implications for thefuture.

What factors should be taken into account indeciding the appropriate general level of educationfor the population? Broadly, educational attain-ment influences labor productivity directly and theproductivity of capital indirectly. An importantfeature of investment in formal education is thatmuch of it precedes workers’ entry into the laborforce, and additional educational investment afterentry in the labor market is, for most individuals,small. Moreover, learning through the work experi-ence is crucially dependent on educational attain-ment prior to entering the labor force. Theimplication is that investment in education should


be decided with reference to the likely course overthe whole working life and to some sections of thepopulation, especially poor rural communities,because of the high cost of education relative tocurrent income.

A major determinant of labor market eventsover an individual’s working life is the changes inthe skills or the education level needed to performwork. Assuming a working life of between 40 and50 years, most of the current cohort of school-agechildren in rural areas either would not beemployed in farming or would switch to a job inindustry or services fairly early on during thecourse of their working life. The implication is thatthe requisite level of education for rural children isnot needed currently to get by with life in ruralareas and work in farming, but that it would beneeded in the future to live in an urban environ-ment and take on jobs in industry or services.Viewed this way, the wide gap in educational attain-ment between urban and rural areas (especially thebackward ones) is a serious deficiency that shouldbe corrected quickly through a concerted policy.Comparatively low educational attainment amongchildren in backward rural areas will in time act asan impediment to their out-migration or reducetheir chances of getting or holding jobs when theymigrate. Either way, comparatively low educationalattainment among a group of children or a time lagin achieving the target of nine years of basic educa-tion for all has adverse implications for reducingpoverty or income inequality in future. The policyimplication is the importance of reducing the costof basic education relative to average income forfamilies in rural communities.

Conclusions

On the eve of its entry into the WTO, China pre-sented a mixed picture of notable successes andsome adverse developments. The Open Doorpolicy, first introduced in 1979, has been a remark-able success. It has brought huge benefits and a fewproblems. China has faced no major balance ofpayments or currency crises. In the 24 years from1978 to 2002, China’s exports have grown at anaverage annual rate of 15 percent (in dollar terms),which points to its ability to adapt to and thrive inthe face of foreign competition. With a foreign

trade to GNP ratio of about 45 percent, the Chineseeconomy was on the eve of its WTO accession asopen as any large economy.22

On the social side, China has also been remark-ably successful in reducing poverty. The officialheadcount of the rural poor dropped from about250 million in 1978 (31 percent of the rural popu-lation) to 32 million at the end of 2000 (3.5 percentof the rural population). The headcount of thepoor and the poverty rate are highly sensitive to thepoverty line. Both of these figures are much higherwhen the $1-a-day poverty line is used, but thereduction in poverty still remains dramatic. Unfa-vorable developments over the last decade or soinclude a sharp rise in interpersonal inequality inboth urban and rural areas and a widening urban-rural disparity. Interprovincial inequality remainslarge, though falling in recent years. The urbanunemployment rate has been rising since the mid-dle of the 1990s and is high by historical standards.In recent years, rural household incomes have risenvery slowly and have even fallen in some localities.

All the studies of the impacts of China’s acces-sion to the WTO agree that the overall benefits out-weigh the losses. Although in the aggregate thepopulation would benefit, losses and gains wouldnot be evenly distributed, and there is little doubtthat there would be a significant number of losers,at least in the short run. These will include workersdisplaced by competition from imported manufac-tured goods and farmers suffering a decline inincome because of cheaper agricultural imports.Although it is difficult to foretell the losses pre-cisely, evidence from the United States help informing a rough idea of the magnitude. That evi-dence suggests that workers losing jobs because ofcompetition from imports on average endure a sig-nificant spell of unemployment and also experiencea decline in earnings when they find another job.Arguably, the losses suffered by Chinese workerswho lose jobs because of WTO entry would be noless than those experienced by U.S. workers unem-ployed as a result of competition from imports.

This point raises the issue of compensation forworkers and farmers adversely affected by WTOentry. What should be the magnitude of compen-sation relative to losses suffered? And what schemesor programs should provide the compensation?Full compensation for losses is ruled out on two


grounds. First, it would be very costly to verify thetrue magnitude of losses, because full compensa-tion would provide a huge incentive to exaggeratelosses. Second, even when losses can be verifiedaccurately, the opportunity cost of compensationmay be prohibitive. On the grounds of feasibility,then, any compensation has to be partial, whichraises the issue of the goal of compensation. Theminimum goal for any compensation has to be theprevention of poverty, a term that can be definedin a number of ways. For the present purposes,poverty is defined as “inadequacy of command overeconomic resources,” with “inadequacy” calibratedto take account of personal or household character-istics such as old age, ill health, or schooling age. Abroader aim of compensation is to avoid relativepoverty, which may be defined as a percentage (e.g.,50 percent) of average or median income per capita.

These goals remain valid regardless of the proxi-mate cause of poverty. Adverse events such as a jobloss or a decline in wages or in farm earnings can beproduced by a wide range of causes, both internaland external. There is no general justification forspecial measures to deal with the adverse impacts ofChina’s WTO membership. However, supplemen-tary measures may be needed if the existing onescannot cope with adverse impacts concentratedover a short time period or in particular regions,such as the rural areas of the already impoverishedNorthwest.

China has a range of schemes for income main-tenance and poverty relief, and most of these sharethree characteristics:

• a stark dichotomy between the urban and ruralpopulation

• a goal of doing little or no more than alleviatingabsolute poverty, a feature that applies especiallyto the rural schemes

• a high degree of decentralization in financingand organization, which goes beyond the levelconsistent with equity and maintaining varia-tion in provision within socially acceptablebounds. The income maintenance schemes forthe rural population are supplemented with ascheme that provides each rural household witha land plot. The household land plots haveplayed a crucial role in reducing poverty and inkeeping rural poverty comparatively low, butthey do not guarantee full protection against

poverty and cannot protect households from thelower earnings that might result from competi-tion by agricultural imports.

The crucial question is how effective would theexisting social security schemes be in preventingworkers who have lost jobs and farmers who haveexperienced a decline in income (including theirfamilies) from falling into poverty, which is theminimum expected of any system of social protec-tion? The answer depends crucially on whether theperson is, officially, urban or rural. The (registered)urban population is covered by unemploymentinsurance, subject to the payment of the requisitecontribution, and the Minimum Living StandardScheme. Together, these schemes provide a fairdegree of protection against poverty, althoughthere are serious problems in the coverage of thehealth care and schooling costs of low-incomeurban households. In contrast, the provisions forincome maintenance of the rural population aresparse and patchy. The employees of TVEs areexcluded from unemployment insurance despitethe fact that they number 128 million, which isabout a quarter of the rural labor force. The ruralsocial assistance schemes, which have goals similarto those of the MLSI, vary widely in coverage, ben-efits, and operation. They do provide a social safetynet, but one that is ridden with tears. As presentlyconstituted, the scheme to assist poor rural locali-ties is not relevant because it is directed mostly atremote localities with little connection to the out-side world, although the scheme can be altered tocover rural localities seriously harmed by competi-tion from cheaper imports. However, there is atpresent no scheme to assist towns and cities caughtin a high unemployment and stagnation trap.

Assuming that the aim is to close the major gapsin social protection against the threat of poverty,the priorities in urban and rural areas are different.In urban areas, the priorities are to alter the cover-age of existing schemes, to institute a financingarrangement that ensures an equitable distributionof burden across government tiers, to make thecontributory benefits portable, and to eliminateinadequacies in benefits. As for altering the cover-age of existing schemes, all the major anomalies canbe eliminated by extending the schemes to includeall long-term residents, regardless of their registra-tion status.


In contrast to its urban counterpart, the ruralpopulation needs additional income protectionschemes. First, given the sizable labor force in TVEs,there is a strong argument for extending UI to wageemployees in TVEs. Second, China needs a socialassistance scheme that targets poor rural house-holds, similar to the MLSI, backed by a nationalframework to ensure a minimum level of protec-tion and equitable distribution of financial burdenacross government tiers. Third, the assistancescheme for rural localities needs to be recast so thatit covers a wider range of disadvantaged localities.

There is a broad consensus on the value of rais-ing the general education level and investing inskills, especially of the rural population, in order tomitigate the adverse impacts of WTO entry onlabor and thus the population. However, these goalsare not panaceas and they have their limitations. Aspointed out earlier, most investment in educationand much of the acquisition of skills take placebefore workers enter the labor market. Therefore,raising the level of basic education has little rele-vance for improving the employability of thosewho are already in the labor force, especially olderworkers. Raising the level of basic education shouldbe viewed as a long-term measure that would becomplemented by short-term measures, such asretraining adapted to the immediate circumstancesof adversely affected workers.

Notes

1. That is, those with urban registration. Some rural socialinsurance schemes are in place, but none provides an unemploy-ment benefit.

2. According to the regulation, social insurance applies inprinciple to the whole of the urban labor force of 247.8 million(34 percent of the total), but the actual coverage falls well shortof that prescribed because of difficulties in collecting contribu-tions. Some cities extend unemployment insurance to migrantworkers in regular wage employment.

3. The supplementary condition is usually the obligation totake offered employment or participate in a training program.

4. This situation has begun to change. At the end of 2003, 11provinces had some form of a minimum living standard schemefor the rural population, which creates the possibility of institut-ing one social safety net covering the whole population. On asimilar note, Shanghai is planning to set up a unified social secu-rity system covering both the urban and the rural population.

5. The “8” in the “8-7 Plan” refers to the eighth five-year plan,which was under implementation in 1994, and the “7” refers tothe seven years from 1994 to 2000. The plan expired at the end ofthe ninth five-year plan and has since then been replaced withthe 2005–2010 Rural Poverty Reduction and Development Planwith the same aim.

6. The estimates of the percentage of the rural poor livingin the selected counties vary over time. Recent estimates were50 percent.

7. The program previously targeted the county level. It isgenerally accepted that counties are too large and differentiatedfor targeting. The proposal is to lower the level of targeting tothe village, which would reduce the numbers of false negativesand positives.

8. In the past, the Five Guarantees Scheme also applied tourban areas, but since 1997 the urban Minimum Living Stan-dard Scheme (MLSS) has superseded it.

9. As one might expect, the headcount of the poor is highlysensitive to shifts in poverty. However, the observation that ruralpoverty fell dramatically during the first 10 years of economicreforms remains valid even when a poverty line higher than theofficial one is used.

10. The actual figure depends on the definition of the landarea (numerator) and of the denominator. Here, the land areais taken to be the sown area in 2000, and the total rural popula-tion serves as the denominator. The area of land per person ishigher if the labor force employed in farming is used as thedenominator.

11. These 12 counties include 6 within the administrativeboundaries of Suzhou City and 3 each within the boundaries ofWuxi and Changzhou City.

12. In principle, land plots are leased for 30 years, but theactual duration in many localities is shorter than 30 years.Furthermore, at any particular point in time the remaining leaseis generally a lot less than 30 years.

13. In considering the implications of the redeployment oflabor to a household land plot as a way of coping with a job loss,one should also take into account any hired labor displaced as aresult.

14. About 86 percent of China’s 1.266 billion population in2000 was living within the boundaries of cities and towns.

15. This situation arises largely from the expansion of citydistricts and the associated redesignation of some rural countiesas urban districts.

16. This is the GDP share of the primary sector, which alsoincludes activities such as quarrying.

17. If the share of farming in GDP is α and in employment isβ, then labor productivity in farming relative to the average,which by definition is 1, is given by (β/α). Similarly, the ratio oflabor productivity in farming with respect to that in nonfarmingactivities is (β/α) * {(1 − α)/(1 − β)}.

18. This is based on the assumption that nonfarming activi-ties are confined to urban areas and farming activities to ruralareas. The first part of the assumption is not strictly true forChina. As a result, the urban–rural gap would be less than fourand a half times.

19. The latter matters because social insurance and socialassistance schemes are budgeted at the city level. The plan is toraise the budgeting to the provincial level. However, the plan hasyet to be fully implemented.

20. The rural and urban lines, especially the latter, can varyacross localities. However, the fact of a wide gap between theurban and rural poverty remains valid.

21. Social insurance covers the old-age pension, health careinsurance, unemployment benefits, disability compensation,and maternity benefits.

22. The ratio depends crucially on how GDP is measured.However, the general proposition about the openness of theChinese economy remains valid regardless of how GDP ismeasured.


References

Collins, Susan M., ed. 1998. Imports and Exports and theAmerican Worker. Washington, D.C.: Brookings InstitutionPress.

Ehrenberg, R. G. 1994. Labor Markets and Integrating NationalEconomies. Washington D.C.: Brookings Institution Press.

Fishlow, A., and K. Parker, ed. 1999. Growing Apart—The Causesand Consequence of Wage Inequality. New York: Council ofForeign Relations.

Kletzer, L. G. 2001. Job Loss from Imports: Measuring the Costs.Washington D.C.: Institute for International Economics.

NBS (National Bureau of Statistics). 2001. China Towns andCities Statistics Yearbook 2002. Beijing: China Statistics Press.

———. 2003a. China Labor Statistical Yearbook 2003. Beijing:China Statistics Press.

———. 2003b. Statistical Yearbook of China 2003. Beijing: ChinaStatistics Press.

Wood, A. 1994. North-South Trade, Employment and Inequality—Changing Fortunes in a Skill-Driven World. Oxford: OxfordUniversity Press.

accession of China, 1, 21–28approval and acceptance, 24background prior to, 21–24, 264bilateral negotiations, 22–23economic impacts of, 99–236

See also economic impacts of accessionfuture implications of, 15–16, 26–27household and poverty impacts of, 237–324

See also poverty impacts of WTO accession; welfareimpacts of WTO accession

multilateral steps of, 23–24policy reforms and, 19–97

See also specific sectors (e.g., agriculture)Protocol and Working Party Report, 24, 107WTO’s basic principles and, 24–26

Administration for Industry and Commerce (AIC), 61Agreement on Government Procurement, 26Agreement on Textiles and Clothing. See textilesAgreement on Trade-Related Aspects of Intellectual

Property Rights. See TRIPSagriculture

accession impacts on, 5–13, 101–16, 219background developments since late 1970s, 102–5cap on domestic support for, 24, 26China National Rural Survey (CNRS) on, 92comparison of income, employment, and food output

non-accession vs. post-accession, 112, 113t7.11regional, 107, 108t7.7, 115n10

Decisions of the Communist Party of China CentralCommittee on Major Issues of Agriculture andRural Work, 73

decollectivization of, 81development and policy, 5, 69–79, 106–7environmental issues and, 78exodus to nonagricultural jobs. See labor mobilityFive-Year Plan for the Development of the National

Economy and Society (ninth), 69–71government subsidies for, 84–85, 85t6.3, 113–14,

215, 221“green barriers,” 79GTAP (Global Trade Analysis Project) model,

application of, 14, 102, 107–9, 115n5hukou system, effect of. See hukou system for urban

residenceimport protection measures, 6–7, 6t1.1incentives and distortion, 5–7, 81–97

estimates of nominal rates of protection, 85–90,87t6.4–88t6.5, 95

gradual opening and, 82–85inland areas and, 90–94integration tests, 91–93, 93t6.7price transmission coefficients, 93–94, 93t6.8

income of farmers, 5compared to urban income, 1, 12, 107, 108t7.8, 213decrease in, 110, 111t7.10, 220education and training, effect of, 319–21increases under Five-Year Plan, 69–70need to increase, 71–73poverty level of. See poverty impacts of accessionregional declines in, 72, 72t5.3sources of, 71, 71t5.2spending reductions due to, 72–73, 72t5.4, 73t5.5

infrastructure improvements and, 113land tenure system, 113, 314–15, 323nn12 & 13local advantages in, 76nominal rates of protection (NRPs) and. See nominal

rates of protection (NRPs)optimization of varieties and quality of products, 75Outline for the Development of Agricultural Science

and Technology in the Next 10 Years, 74poverty of agriculture-based regions, 112prices, compared to nonfarm products, 109–11,

110t7.9processing, expansion of, 75–76production increases, 69, 70t5.1, 103t7.1research and development in, 113restructuring of, 5, 73–78, 101self-sufficiency, 104, 112, 114, 114t7.12social assistance and. See social assistancespending on food, 75, 75t5.6tariffs and quotas, 79, 107, 115n2taxation and, 84, 85t6.3, 95n1township and village enterprises (TVEs) and, 76, 102trade and, 79, 83, 83t6.2

liberalization, 84–85, 106rates of protection, 83, 83t6.1sanitary and phytosanitary measures, 115n1

transient employment for farmers, 76–77AIC. See Administration for Industry and Commerceair transport, 152–53

See also logistics sectorantidumping and safeguards, 3, 15, 29–47

current situation under, 3, 8, 30–33measures in force against China, 22, 32,

32t3.3, 46n2sensitivity to “antidumping-intensive” sectors, 35

Index

325

antidumping and safeguards (continued)targets and measures in force, 30–32, 31t3.2,

33, 46n2users and measures in force, 30–32, 30t3.1,

33, 46n2dispute settlement use of, 3EC and U.S. echoing cases, 36, 37t3.5enforcement challenges for China, 39–40

measures in force (as of June 2003), 39, 40t3.6exposure to foreign antidumping, 33–36

Chinese policies to minimize, 35–36Harmonized Tariff System (HTS) and, 33–35,

34t3.4, 43linking liberalization to better treatment of exports by

antidumping users, 42–43negotiations on contingent protection, 41–45non-market economy (NME) status, use of, 29,

42–43, 42t3.7, 46n4regulations

adoption of, 26, 38–39stricter rules, proposal for, 4, 29, 43–44

retaliatory approach, Chinese use of, 3–4, 39, 46Russia and, 46n1safeguard, defined, 29sectors that are “antidumping-intensive,” 33–35,

34t3.4, 46n2steel safeguard measures, 27transitional product-specific safeguard (TPS), 29,

44–45, 46n2, 216apparel. See textilesArgentina and antidumping, 4, 30AT&T, 178n11automobiles. See motor vehicle sectoraveraging tariffs, 35

bankruptcy, 27banks. See financial servicesbeverages

import increases, 224price reduction, 219tariff reductions for, 8, 15, 215trademark protection, 55

bilateral negotiations for accession, 22–23Brazil

antidumping and, 30intellectual property rights and, 55

cable and convergence issues, 173Central Working Committee on Rural Affairs, 73, 74cereals, 23, 26CGE. See computable general equilibrium (CGE)

model, application ofChile and universal access

in services sector, 133in telecommunications, 175b10.3

China Customs data, 212

China Health and Nutrition Survey (CHNS) data used in income and labor study, 242–43,244t14.1, 259n4

China Insurance Regulatory Commission (CIRC), 183China Mobile Communications Corporation, 159,

162, 163China National Rural Survey (CNRS), 92China Netcom, 162, 170China Netcom Communications Group Corporation, 170China Satellite (ChinaSat), 159China Securities Regulatory Commission (CSRC), 183China Telecom

break up of monopoly of, 158–60lowering of long-distance rates and, 162mobile market and, 179n17monopoly of, 157, 158profitability of, 162restructuring as of 2002, 170

China United Telecommunications Corporation.See Unicom

CHNS. See China Health and Nutrition Survey (CHNS)data used in income and labor study

CIRC (China Insurance Regulatory Commission), 183clothing. See textilesCNRS (China National Rural Survey), 92competition

government procurement, fair competition in, 26intellectual property rights and, 57, 62, 66in logistics sector, 151in services sector, 132in telecommunications, 171–72

growth and competition in subsectors, 162–64reforms, effect of, 173

computable general equilibrium (CGE) model,application of

economic impacts of accession, 211, 289–92motor vehicle sector, 197welfare impacts, 262

computer industry, 121, 123t8.2See also technology tariff reductionsoftware protection, 55, 57

construction and engineering services, 121, 124t8.2, 127consumer information, adequate provision of, 9, 132–33cotton

export subsidies for, 84–85NRP for, 88, 94protection for, 88t6.5, 89tariff reductions for, 15

courier services, 155cross-border supply, 118

developing countriesantidumping and, 30, 46n3China’s accession with provisions for, 23–24, 26preferential treatment for, 2, 4

dispute settlement, use of, 3, 27

326 Index

distribution services, 124t8.2, 127, 155n3Doha Development Agenda, 7, 15, 25, 29drought, 71drugs. See pharmaceuticalsdual pricing practices, 2, 25

earnings. See income levelsE-commerce, 169, 172–73, 178n14

See also Interneteconomic impacts of accession, 1, 5–13, 211–36

See also specific economic sectors (e.g., agriculture,financial services, etc.)

assessment of, 219–23, 295, 296t16.5on China, 213–15, 219–21on China’s trading partners, 215–17, 221–23, 224complementary policy reforms, 223–24experimental design for evaluation of, 217–19

changes in China’s key economic indicators,220t13.5, 231t13.A.5–233t13.A.7

projected growth, 217, 218t13.4export processing arrangements, 212intellectual property rights and, 54–60labor market policies. See labor marketlabor mobility and. See labor mobilitymarket access and (1997), 286, 287t16.2, 289methodology for analysis, 212–13rural development. See rural economysectoral effects, 109–11, 110t7.9, 219–21

See also specific sectors (e.g., agriculture)skill levels and. See labor markettrade reform, 81, 213–17

See also trade reformEdge2net Inc., 178n11education and training

cross-border delivery of, 124t8.2, 127Cultural Revolution, effect of, 207for farmers, 319–21income level and, 246, 247on intellectual property rights, 49, 61, 65for logistics services, 148nonagricultural vs. agricultural work and, 259, 285skill level increases as result of, 223–24

8-7 Plan, 313, 323n5electronics. See technology tariff reductionEl Salvador, 22employment. See labor market; labor mobilityengineering services. See construction

and engineering servicesenvironmental issues and agriculture, 78European Community

antidumping“echoing” against China, 36, 37t3.5position on, 29, 46n5

bilateral agreement with, 22European Union patent protection, 66n6exports. See trade reform

Fair Trade Administration for Import and Export (FTA), 39

farmers. See agricultureFDI (foreign direct investment). See foreign

investmentsfinancial services, 10–11, 181–90

accession impacts on, 15compared to other East Asian countries, 189n9foreign banks, 182, 185, 189n12GATS commitments and, 125t8.2, 127, 138n16initial conditions of, 182–85issues needing to be addressed, 187–89nonperforming loans (NPLs), 184policy banks, 185preaccession situation, 183–85, 183t11.1reform, 182–83

sequencing of, 134–36rural system, 184state banks, 10, 184–85

survival of, 187–88, 188b11.1, 189n8WTO commitments in, 186–87, 186t11.2

First Automobile Works (FAW), 193, 197Five Guarantees Scheme, 313, 323n8Five-Year Plan for the Development of the National

Economy and Society (ninth), 69–71fixed-line sector, 162food

See also agriculture; specific productsconsumption trends, 96n4import increases, 224population growth and, 70tariff reductions for, 215trademark protection, 55trade structure of, 104, 105t7.5

food processing, 75–76foreign exchange market, 27foreign investments

accession impacts on, 226growth in, 283in intellectual property rights, 56in services sector, 120, 131, 215in telecommunications, 10, 158, 167–68

forestry, 78forwarding, 154–55

See also logistics sectorfranchising. See distribution servicesfreight services. See logistics sectorFTA (Fair Trade Administration for Import and

Export), 39

GATS (General Agreement on Trade in Services)China’s commitments to, 118–28, 120t8.1

comparison over time, 121, 122t8.2liberalization on faster schedule, 117by subsector, 119–28, 122t8.2, 137t8.A.1

See also specific subsectors (e.g., financial services)

Index 327

GATS (continued)country commitments by mode, 118–19, 119f 8.1logistics sector and, 10WTO reference paper on basic communications,

10, 167, 174responding to, 168–69summary of, 177–78

GATT (General Agreement on Trade and Tariffs)China and, 2, 12, 21principles of. See specific principle

(e.g., nondiscrimination)GATT Working Party on China’s Status, 21, 22GDP. See gross domestic product (GDP) growthGlobal Trade Analysis Project. See GTAP (Global Trade

Analysis Project) model, application ofgovernment procurement, fair competition in, 26government subsides. See agriculturegrain production and consumption

See also specific types (e.g., wheat)demand for and adjustment in quantity, 73–74exports, 83factors in successful harvests, 70–71local comparative advantages and, 76local grain bureaus, 96n2NRP for, 87, 87t6.4reduction in government subsidies for, 113–14self-sufficiency and, 104, 114, 114t7.12variances from year to year, 104, 104t7.4

Great Wall Telecom Corporation, 178n7gross domestic product (GDP) growth, 81

and rural poverty, 242GTAP (Global Trade Analysis Project)

model, application ofagriculture, 14, 102, 107–9, 115n5defined, 234n5economic impacts, 211, 212–13, 217motor vehicle sector, 197welfare impacts, 264–65

Harmonized Tariff System (HTS) and antidumpingmeasures, 33–35, 34t3.4, 43, 46n2

Hong Kongantidumping and, 36Trade Development Council, 149universal access, 175b10.3

household impacts of WTO accession, 13–15, 237–324See also poverty impacts of accession; welfare impacts

of accessionincome levels. See income levels

hukou system for urban residence, 109, 213, 291abolition of, 12–13, 14, 16, 224, 301

ICBC (Industrial and Commercial Bank of China), 182income levels, 239–60

See also poverty impacts of accession; welfare impactsof accession

8-7 Plan and, 313, 323n5factors affecting, 239farmers. See agriculturegain/loss after accession, 269, 270f15.3–272f 15.8, 273gap between poorer and richer households, 240, 247,

248t14.4, 258income generation functions, determination of,

243–48, 245t14.2–246t14.3, 258land plot as insurance, 314–15, 323nn12 & 13maintenance schemes, 311–14Minimum Living Standard Insurance (MLSI),

312–13, 318per capita income by stratum, location, and vingtile,

284–85, 285t16.1survey sample data for study of, 242–43, 244t14.1types of labor income, 265, 301wages and shadow wages, 240, 251–55,

253t14.7–255t14.9, 259India

antidumping and, 4, 30effect of China’s accession on, 223telecom infrastructure in, 132, 138nn8 & 12

Industrial and Commercial Bank of China (ICBC), 182industrial products, 7–8, 15

nontariff barriers and, 226tariff reductions for, 107, 215wholesale price reductions for, 219

infrastructure improvementsfor agriculture, 113for e-commerce and Internet, 173for logistics services, 148

inland areasagricultural prices and, 90–94services sector in, 131, 138n10

inspection system, 24institutional implications of accession, 26–27insurance. See financial servicesintellectual property rights, 49–67

See also TRIPS (Agreement on Trade-Related Aspectsof Intellectual Property Rights)

China’s record on, 4, 57–59, 58t4.2competition and, 57, 62, 66copyright, 54, 55, 57, 58–59

ineffective enforcement, 61reform, 50, 53, 58

costs associated with, 53, 56current situation in China, 60–65

interview findings, 60–62database protection, 59defensive measures used, 62economic development and, 54–60

constraints on economy, 4, 56–57stimulation of economy, 4, 55–56

education and training programs in, 49, 61, 65enforcement, 49, 55, 59–60

ineffective, 60–61, 65

328 Index

foreign companies and, 61–62foreign direct investment in, 56government support for inventions, 54licensing contracts, 56local protectionism and, 50, 61nature of, 53–54patents, 54, 55, 62–65

biological, 59increase in, 62–65, 62t4.3–64t4.7ineffective, 61reform, 50regional statistics, 63–65, 64t4.7software coverage, 57–58

pirating of recorded entertainment and software, 56, 60, 61

plant varieties protection, 50, 54, 59regulatory changes due to WTO commitments,

50–53, 65–66trademarks, 54, 55, 62–65

counterfeit, 56enforcement weakness, 61increase in number of, 64–65ineffective enforcement, 60reform, 50, 52–53, 59

trade secrets, 54, 56, 59, 62types of, 54–55

Internetadministration of, 160competition in services for, 158, 159development of, 172–73increased use of, 163–64, 164f10.2, 178n9regulations regarding, 169

IPRs. See intellectual property rightsirrigation, 71

Japanantidumping and, 36China’s accession, effect on, 224intellectual property rights and, 55

Jiang Zemin, 74Jia Yu Freight Group Company, 148b9.1Jitong, 158, 159, 162, 170joint ventures for services, 120, 131

labor market, 13See also labor mobilityaccession impacts on, 259, 305–8agriculture. See agriculturebackground of, 239defection of employees with technical

knowledge, 61displaced workers as result of accession, 306income and. See income levelslevel and pattern of labor supply, empirical study of,

248–58empirical specification, 249–50

estimation of labor supply functions, 255–58,256t14.10

estimation of production functions, 250–51,251t14.5–251t14.6

estimation of wages and shadow wages, 251–55time allocation by rural household for

labor, 249, 259policies, 212–13

effect of reforms, 223, 224, 225t13.8skill levels

accession impacts on, 307increase in, after accession, 223–24, 225t13.8skilled, defined, 282n12skilled vs. unskilled in services sector, 138n13urban skilled and semiskilled labor, 301

social protection and adaptation in, 308–14See also social assistanceclassification of schemes and measures, 308,

308t17.1decentralization and, 312b17.1, 313differences between social insurance and social

assistance, 310, 310t17.2instruments for, 309–11

types of labor, 265, 295unemployment, 76, 225, 235n25, 306–7, 312, 317–18

labor mobility, 13, 15, 239–60barriers to, 212–13, 223, 224, 235n28farmers and, 76–77, 220, 223, 224, 235n23hukou system, effect of. See hukou system for urban

residencelimited nature of, in rural environment, 240return to agriculture, 295rural-urban migration, 241, 291–92, 295, 316

landagricultural plot as insurance, 314–15, 323nn12 & 13tenure system, 113, 265, 314–15

licensing. See intellectual property rightslocal protectionism

intellectual property rights and, 50, 61logistics sector and, 147, 148b9.1motor vehicle sector and, 209n4

logistics sector, 9–10, 141–56accession implications for, 15, 145–48

direct effects of, 145–47, 145t9.1–146t9.3indirect effects of, 147–48, 147t9.4by subsector, 152–55

See also specific subsector (e.g., air transport)competition in, 151constraints on development of, 143–44cost savings in third-party sector, 148–50, 149t9.5definition of logistics, 142demand for 3PL (specialized) services, 144, 146, 152evolution of, 142–43freight traffic by mode (1970-2000), 143, 143f 9.1GATS commitments and, 126t8.2, 128human resources, 148

Index 329

logistics sector (continued)infrastructure and facilities/equipment, 148in inland areas, 131, 138n10less developed areas, implications for, 150–51, 152local protectionism and, 147, 148b9.1separation of government administration from

enterprises, 147–48long-distance service. See telecommunicationsLong Yongtu, 29, 43

maizeSee also agricultureborder prices and, 93–94, 93t6.8export subsidies for, 84–85NRP for, 83, 88, 94rate of protection, 7tariff reductions for, 15variation in price between imported and domestic,

88, 88t6.5manufacturing. See industrial productsmarket access, 2–3, 25

See also tariffseconomic structure and (1997), 286, 287t16.2, 289GATS obligations, 118, 121, 130–31

Marrakesh Agreement, 21meat exporters, 85medicines. See pharmaceuticalsmergers of telecommunications companies, 170Mexico and antidumping, 4, 30MFA. See Multi-Fiber ArrangementMFN. See most favored nation (MFN) statusMII. See Ministry of Information IndustryMinimum Living Standard Insurance (MLSI),

312–13, 318Ministry of Foreign Trade and Economic Cooperation

(MOFTEC), 39Ministry of Information Industry (MII), 158–61

complaints against telecom companies to, 165September 2001 regulations, designation of, 168universal service fund, 175

Ministry of Posts and Telecommunications (MPT),157–58, 161

MLSI. See Minimum Living Standard Insurancemobile phones. See telecommunicationsMOFTEC (Ministry of Foreign Trade and Economic

Cooperation), 39most favored nation (MFN) status

accession and, 2, 25GATS obligations, 118, 129non-market economy (NME) status and, 43U.S. treatment of China, 22–23

motor vehicle sector, 11, 15, 191–210comparison with other countries, 191, 192t12.1import and domestic shipment values, 194, 196t12.3import increases, 224import share of total auto parts, 203, 203f 12.4

local protectionism, 209n4modeling framework for assessing accession impact,

197–206CGE model used, 197data used, 197–98, 198t12.5experiments and results, 201–6, 202t12.7–203t12.8export impacts, 205–6, 205t12.10–206t12.11scale and average cost, 200, 201f12.2structure of model, 198–200value-added results, 204, 204t12.9

price reduction, 219production

comparison with U.S., 200, 201f12.3location of plants, 208f12.5by plant (1995-2002), 194, 195t12.2since open-door policy adopted, 193b12.1,

194f 12.1regional and sectoral breakdown of model,

198t12.5restructuring after accession, 115n7, 221scale and average cost, 201f12.2smuggling, 209n3summary of developments, 193b12.1tariff reductions for, 215tariffs on vehicles, 26, 196, 196t12.4, 199, 199t12.6

MPT. See Ministry of Posts and TelecommunicationsMulti-Fiber Arrangement (MFA), 2, 3, 33, 283, 306multilateral steps of accession, 23–24

National Conference on Agricultural Science andTechnology, 74

National Conference on Rural Work, 73National Industrial and Commercial Administration, 53national treatment and GATS obligations, 118,

121, 129NME. See non-market economy (NME) status, use ofnominal rates of protection (NRPs)

agriculture and, 85–90, 87t6.4–88t6.5, 106,106t7.6, 234n3

greater protection cuts, 111–12removing protection, 112

challenges and issues in measuring, 95nonagricultural employment. See rural economynondiscrimination, 2, 24, 25

GATS commitments and, 129–30non-market economy (NME) status, use of, 29, 42–43,

42t3.7, 46n4nontariff barriers (NTBs), 82, 83, 84, 85

import coverage, 213–14, 214t13.1,227t13.A.1–229t13.A.3

manufacturing sector and, 226NRPs. See nominal rates of protection

Open Door policy, 307, 321motor vehicle sector production after, 193b12.1,

194f 12.1

330 Index

Organisation for Economic Cooperation andDevelopment (OECD) countries

agricultural reform and, 101, 114

packaging services, 155Patent Cooperation Treaty, 56, 66n8patents. See intellectual property rightsPeople’s Bank of China (PBC), 136, 181, 182Peru and universal access, 175b10.3pharmaceuticals

patent protection, 56–57pricing of, 5, 56, 66trademark protection, 55trade secret protection, 59

Philippines and intellectual property rights, 55pirating of copyrighted material. See intellectual

property rightsplant varieties. See intellectual property rightsPoland and telecommunications, 138n11policy reforms of WTO accession, 2–5, 19–97poverty impacts of accession, 13–15, 16, 27, 237–324

See also welfare impacts of accessionagricultural income and, 285, 289agriculture-based regions, 112, 242background, 283–84base case projections and accession simulation,

292–94, 293t16.3simulation results, 294–301summary of, 294t16.4

CGE model used in study of, 289–92economy-wide impacts, 259n1, 2958-7 Plan and, 313Five Guarantees Scheme and, 313, 323n8gross domestic product (GDP) growth and, 242household impacts, 242, 284–86, 298–300, 299t16.7

rural households, 300f16.2urban households, 300f16.3

land plot as insurance, 314–15, 323nn12 & 13Minimum Living Standard Insurance (MLSI),

312–13, 318model used in study of

computable general equilibrium (CGE) model,application of, 289–90, 289–92

household behavior, 290–91household migration, 291–92production, exports, and imports, 292

per capita income by stratum, location, and vingtile,284–85, 285t16.1

production, trade, and protection, 286–89rural nonfarm poverty, 112sector impacts, 295–98, 297t16.6, 298f16.1social assistance and. See social assistancespecial measures following accession, 322–23

predictability. See transparency and predictabilitypreferential treatment for developing countries,

2, 4, 26

pricingagriculture, 78

compared to nonfarm products, 109–11,110t7.9

inland areas and, 90–94dual pricing practices, 2, 25medicines. See pharmaceuticalsreductions after accession, 219–20, 221t13.6,

230t13.A.4telecommunications, 161–62welfare impacts of accession and price changes,

265–66, 266t15.1–267t15.2private enterprises, 241, 259n2professional services, 121, 122t8.2Protocol and Working Party Report, 24, 107

Railcom, 159rail transport, 154

See also logistics sectorregional comparisons

income gap, 248income of farmers, 72, 72t5.3logistic services, 150–51patents, 63–65, 64t4.7rural income, employment, and food output, 107,

108t7.7, 115n10research and development

in agriculture, 113intellectual property rights for. See intellectual

property rightsretaliatory antidumping, 3–4, 39, 46rice

border prices and, 93t6.8, 94import from Thailand, 89–90NRP for, 83, 88t6.5rate of protection, 7subsidies for, 85

road transport, 153–54See also logistics sector

rural economySee also agriculture; township and village

enterprises (TVEs)8-7 Plan and, 313, 323n5income levels, 239–60labor market in. See labor marketland plot as insurance, 314–15, 323nn12 & 13multiple occupations of workers, 242nonagriculture employment, 241–42regional comparison of income, employment, and

food output, 107, 108t7.7, 115n10restructuring of, 74–78small town development, 76–77social assistance and income maintenance schemes,

309–14telecommunications, 158

Russia and antidumping, 46n1

Index 331

safeguards. See antidumping and safeguardssatellite communications, 159schooling. See education and trainingSCIO. See State Council Information Officesecurities. See financial servicesservices sector, 2, 8–11, 117–39

adjustment costs, 133–34business scope for service, 120comparison with other WTO members, 118–19,

119f 8.1competition in, 132computer and related services, 121, 123t8.2construction and engineering services, 121,

124t8.2, 127consumer information, adequate provision as priority,

9, 132–33distribution services, 124t8.2, 127educational services. See education and trainingfinancial services. See financial servicesform of establishment, 120, 131GATS framework and commitments, 118–28

See also GATS (General Agreement on Trade in Services)

geographic scope of service, 120grandfather provisions for, 127–28import increases, 224liberalization after accession, 115n7, 221

eliminating barriers to entry, 2, 26, 130–31, 215eliminating discretion, 128–29eliminating discrimination, 129–30eliminating foreign ownership restrictions, 131, 215eliminating regional restrictions, 9, 131–32

nontariff barriers, reduction of, 215priorities for reform, 132–36professional services, 121, 122t8.2regulatory requirements for, 120, 132–36telecommunications. See telecommunicationsuniversal service as priority, 9, 133

SETC. See State Economic and Trade CommissionShanghai Banking Association, 181Shenzhen China Motion Telecom, 178n11small town development, 76–77

See also township and village enterprises (TVEs)social assistance, 14–15, 305–24

classification of population and, 315–17classification of schemes and measures, 308,

308t17.1closing loopholes in, 317–19decentralization of, 312b17.1, 313differences from social insurance, 310, 310t17.2,

323n218-7 Plan, 313, 323n5Five Guarantees Scheme, 313, 323n8migrants and, 318–19, 323n2Minimum Living Standard Insurance (MLSI),

312–13, 318

special measures following accession, 322–23types of, 309–14

South Africaantidumping and, 4, 30telecommunications, 134

universal access, 175b10.3soybeans

NRP for, 83, 88, 88t6.5, 94tariff reductions for, 115n2value-added tax on, 84variation in price between imported and domestic,

89, 89t6.6State Council Information Office (SCIO), 132, 160, 169State Economic and Trade Commission (SETC), 39, 158State Information Leading Group (SILG), 160State Market Administration Bureau (SMAB), 91steel safeguard measures, 27stock exchanges, 182

See also financial servicessubcontracting in logistics sector, 148–50, 149t9.5subsidies for agriculture. See agriculture, subheading:

government subsidies forsugar

NRP for, 88, 94price reduction, 220protection for, 88t6.5, 89tariff reductions for, 15

TaiwanChina’s accession, effect on, 217, 221, 224, 225, 234n1counterfeit goods, regulation of, 66n5recognition of, 22

Tariff Commission under the State Council (TCSC), 39tariffs

agricultural products and, 15, 79, 107, 115n2antidumping and, 35averaging, 35motor vehicles and, 26, 196, 196t12.4reduction, 2–3, 25, 26, 27–28, 107

See also specific productsby China’s trading partners after accession, 217comparison of rates pre- and post-accession, 214,

216t13.3experimental design to evaluate accession impact

on, 217, 219for years 1992-2001, 214–15, 215t13.2

structure of, 289taxation. See value-added taxTCSC (Tariff Commission under the State Council), 39technology tariff reduction, 8, 217telecommunications, 10, 132, 157–79

accession impacts on, 15allocation of scarce resources, 168, 178break up of monopoly, 158–60, 159t10.1, 160t10.2cable and convergence issues, 173competition in, 164–66, 171–72, 173

332 Index

complaints against companies, 165defining coverage of, 167, 167b10.2E-commerce, 169fixed-line sector, 158, 159, 162, 163f10.1foreign investment in, 10, 158, 167–68GATS commitments and, 121, 123t8.2growth and competition in subsectors, 162–64in inland areas, 131Internet. See Internetissues not yet addressed, 177long-distance rates, 161–62mergers and consolidation, 170mobile phones, 159, 162–63, 163f 10.1, 171–72new entrants into, 171–72preaccession situation, 157–61pricing

decline, 161–62setting of, 158

profitability, 162quality of service, 164–66reforms, 165t10.3, 176

direct impact of, 170–76new law, 169, 177prior to accession, 161–66sequence of, 134, 135f 8.2

regulatory framework for, 166–70, 166b10.1September 2001 regulations, 167–69

restructuring, 168, 170, 170f 10.3revenue share by company (2001-2002), 172f 10.4satellite communications, 159third-generation (3G) technology, 163, 171township and village enterprises (TVEs) for, 158universal service to poor and isolated areas, 161, 168,

174–76, 175b10.3, 178, 179n19value-added services, 166, 166b10.1, 176,

178n12WTO commitments on, 166–67, 166b10.1WTO reference paper on basic communications

(GATS), 10, 167, 174responding to, 168–69summary of, 177–78

textilesantidumping measures and, 33import increases, 224quotas, 2, 3, 11, 115n7, 215, 219, 221tariff reduction, 8trade expansion after accession, 12, 215–16, 219

tobaccoimport increases, 224price reduction, 219restricted to state trading, 23tariff reductions for, 8, 15, 215

township and village enterprises (TVEs)for agriculture, 76, 102growth of, 316nonagricultural employment in, 241

rural workforce employed by, 312for telecommunications, 158

TPS. See transitional product-specific safeguardtrade expansion, 1, 12trademarks. See intellectual property rightstrade reform, 1, 213–17

agriculture. See agricultureChina’s policy changes, 213–15, 219–21export processing arrangements, analysis of, 212export shares, 206, 207t12.12global trade, effect on, 224nontariff barriers. See nontariff barriers (NTBs)tariff reduction. See tariffsthree-year transition period for full liberalization, 24trade partners’ policies, 215–17, 221–23, 224welfare impacts of, 262–64

Trade-Related Aspects of Intellectual Property RightsAgreement. See TRIPS

trade secrets. See intellectual property rightstraining. See education and trainingtransitional product-specific safeguard (TPS), 29, 44–45,

46n2, 216transparency and predictability, 2, 3, 25–26, 118transport of goods. See logistics sectorTRIPS (Agreement on Trade-Related Aspects of

Intellectual Property Rights), 2, 4–5, 15Chinese implementation of, 26, 49, 57substantive requirements of, 50–53, 51t4.1

TVEs. See township and village enterprises

undistorted trade, 2, 3, 26See also antidumping

unemployment, 76, 225, 235n25, 306–7, 312, 317–18Unicom (China United Telecommunications

Corporation), 158, 159, 162, 163, 178n5Union for the Protection of New Varieties of Plants

(UPOV), 50United States

antidumping and safeguards“echoing” against China, 36, 37t3.5position on, 46n5steel safeguard measures, 27

bilateral agreement with China, 22universal services

as priority, 9, 133telecommunications, 161, 174–76, 175b10.3, 178,

179n19urban areas, migration to. See labor mobilityurban income. See income levelsUruguay Round

accession process and, 22antidumping and, 30preferential treatment for developing countries

and, 26textiles and, 215. See textilesZero-for-Zero program, 217

Index 333

value-added taxagriculture and, 84, 85t6.3, 95n1experimental design to evaluate accession impact

and, 219export processors and, 234nn8 & 9

Vietnam, effect of China’s accession on, 222t13.7, 223VoIP (Voice over Internet Protocol), 158, 164

wages. See income levelswarehousing and storage, 154

See also logistics sectorwater conservation, 78water transport, 153

See also logistics sectorwelfare impacts of accession, 261–82

See also poverty impacts of accessionapproach to measuring, 262–64computable general equilibrium (CGE) model,

application of, 262gains and losses, 221, 222t13.7

by agricultural households, 275, 278t15.8explanation of, 273–78income levels, 269, 270f15.3–272f15.8, 273by rural areas of provinces, 274, 274t15.4–275t15.5,

280t15.A.1by urban areas of provinces, 274,

276t15.6–277t15.7, 281t15.A.2

GTAP (Global Trade Analysis Project) model,application of, 264–65

household income, defined for, 265labor and labor earnings, defined for, 265land prices, effect of, 265measured impacts, 265–73

poverty incidence, 268f15.1–269f15.2price changes, 265–66, 266t15.1–267t15.2summary statistics on, 268t15.3

sampling weights for study of, 264survey data for study of, 264trade reform and, 262–64

wheat, 83, 86–88, 87t6.4, 89–90World Intellectual Property Organization

(WIPO), 53Copyright Treaty and Treaty on Performances and

Phonograms, 59WTO membership, 21–22

See also accession of ChinaWTO Working Party on the Accession of China, 21, 22,

23, 24

Zhu Rhongji, 74

334 Index

Th

e W

orl

d B

an

k

1818

H S

tree

t, N

.W.

Wa

shin

gto

n,

D.C

. 2

04

33

U.S

.A.

CH

INA

AN

D T

HE

WT

O:

ACCESSIO

N,

PO

LIC

Y R

EFO

RM

, AN

D P

OV

ERT

Y R

ED

UCT

ION

ST

RAT

EG

IES

○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○ ○

Ho

w w

ill

Ch

ina’

s ac

cess

ion

to

th

e W

TO

im

pac

t it

s p

eop

le a

nd

its

eco

no

my

? D

raw

ing

on

an

aly

sis

by

lead

ing

sch

ola

rs f

rom

Ch

ina

and

els

ewh

ere,

th

is b

oo

k l

oo

ks

at t

he

po

licy

ref

orm

s n

eed

ed i

n k

ey a

reas

,su

ch a

s a

gri

cult

ure

, ser

vic

es a

nd

in

tell

ectu

al p

rop

erty

, to

en

sure

co

nti

nu

ed e

con

om

ic d

evel

op

men

t an

dre

du

ctio

ns

in p

ov

erty

.

eltiT

#kcot

S*ecir

P.yt

Q$SUlat

oT

nihC

aOT

Weht

dna

htiw

noit

acil

bu

poc

A.s

eire

St

ne

mp

olev

eD

dn

ae

dar

Tk

na

Bdlr

oW

)X-

76

65-

31

28-

0N

BSI(

.s

eg

ap

00

5.

40

02

yluJ

.ss

erP

ytisr

evi

nU

drof

xO

76651D

52$

*ees

esael

P.seirt

nuoc

emos

rof

%57

otpu

-yl

ppa

stnuocsi

dci

hpar

goe

G.st

nuocsi

d/gr

o.knabdlr

ow.

snoit

acil

bup//:

ptth

laut

ca

,desu

siredr

oesahcr

up

afI

.redr

orep

00.

8$S

Uer

asegr

ahc

gnil

dna

Hdna

gni

ppi

hS**

era

segr

ahc

,set

atS

deti

nU

eht

edi

stuo

yrevil

edli

am

ria

roF

.degr

ahc

eblli

wgni

ppi

hs

.meti

rep

00.

6$S

Usul

predr

orep

00.

7$S

U

DOHTE

MTNE

MYAP

.noit

amr

ofni

drac

tider

cro

tne

myap

yb

dei

nap

mocca

eb

tsu

msl

audi

vidni

morf

sredr

Oruoy

hiw

kceh

C.k

na

Bdlr

oW

eht

otdesser

dda

sredr

orof

ylno

det

pecca

sdr

ac

tider

Crot

ubirt

sidlacol

.hsac

dnes

ton

od

esael

P.sdr

ac

tider

cel

bat

pecca

tuoba

atot

buS

l **

H&

S

$S

Ulat

oT

(Ple

ase

prin

t)

Na

me

Titl

e

Org

aniz

atio

n

Add

ress

City

Sta

teZ

ip/P

osta

l Cod

e

Cou

ntry

Pho

ne

Fax

E-m

ail

Car

d no

.

Exp

iratio

n da

te

/

Nam

e(a

s it

appe

ars

on th

e ca

rd)

Sig

natu

re(r

equi

red

for

all

cred

it ca

rd c

harg

es)

Che

ckno

.in

the

amou

nt o

f $is

enc

lose

d. W

hen

orde

ring

dire

ctly

from

the

Wor

ld B

ank,

mak

e ch

eck

paya

ble

in U

.S. f

unds

onl

y dr

awn

on a

U.S

. ban

kto

: The

Wor

ldB

ank.

Ple

ase

send

you

r ch

eck

with

you

r or

der.

Bill

me.

Ple

ase

incl

ude

purc

hase

ord

er.

Cha

rge

my:

Am

eric

an E

xpre

ssM

aste

rcar

dV

isa

By

mai

l:W

orl

d B

ank

Pu

blic

atio

ns

P.O

. B

ox 9

60,

Her

ndon

, V

A20

172-

0960

, U

SA

By

fax:

+1-

703-

661-

1501

By

phon

e: +

1-70

3-66

1-15

80 o

r 80

0-64

5-72

47O

nlin

e:w

ww

.wo

rld

ban

k.o

rg/p

ub

licat

ion

s

Que

stio

ns?

E-m

ail

us a

tb

oo

ks@

wo

rld

ban

k.o

rg

YE

S, p

leas

e se

nd m

y co

py o

f Ch

ina

and

th

e W

TO

SO

UR

CE

CO

DE

DC

WT

O

PR

SR

TS

TD

U.S

.PO

ST

AG

E

PA

IDP

ER

MIT

#4

11

WA

SH

ING

TO

N, D

.C.

WO

RL

D B

AN

K

Pu

blicati

on

sT

he

re

fere

nc

e o

f c

ho

ice

on

de

ve

lop

me

nt

To

Ord

er:In

stit

uti

on

al c

ust

om

ers

in t

he

U.S

. o

nly

:

T R A D E A N D D E V E L O P M E N T S E R I E S

T his book fills a large gap in the analysis of China’s WTOaccession. Evidence-based investigations of the likely effect offurther opening of China’s economy, following the steady

globalization that had occurred since 1979, have been largely unavailable.Moreover, the strategic dimensions of the policy changes implemented byChina’s far-seeing leaders have not been addressed adequately until now, andnot as thoroughly as in this book. This collection of first-rate technicalstudies, drawn from good understanding of Chinese and internationaleconomic institutions, and tracing the likely effects of China’s globalizationon different segments of its economy — especially its poor — is essentialreading for students of public policy.”

Long Yongtu, Secretary General, BOAO Forum for Asia

“This book provides the most comprehensive and up-to-date evaluation ofChina’s entry into the WTO. It provides a succinct summary of China’scommitments to the WTO and a detailed analysis of the economicconsequences of the fulfillment of these obligations. A particular strength ofthe volume is its treatment of the relationship between trade liberalizationand the distribution of income in China.”

Nicholas R. Lardy, Senior Fellow, Institute for International Economics

™xHSKIMBy356678zv,:&:<:':(0-8213-5667-4

THE WORLD BANK

3

“

china and the wto - isbn: 0821356674 - world bank

Documents