ra~~~~~~~~~~~ |~~ RI-W = _j v C THE WORLD BANK ECONOMIC REVIEW EDITOR L. Alan Winters, University of Suissex, UK PAST EDITOR FranSois Bourguignon, World Bank EDITORIAL BOARD Abhijit Banerjee, Massachusetts Institute of Ravi Kanbur, Cornzell University, USA T. i. 1, USA Elizabeth M. King, World Bank Kaushik Basu, Cornlell UniversitY, USA Justin Yifu Lin, China Center for Econonmic Tim Besley, London School of Economics, UK Research, Peking University, Cbhina Anne Case, Princeton UL :, '.:1. USA Mustapha Kamel Nabli, World Bank Stijn Claessens, University ofAnmsterdain, Juan Pablo Nicolini, Universidad di Tella, The Netherlands Argentina Paul Collier, World Bank Howard Pack, University ojfPennsylvania, USA Augustin Kwasi Fosu, African Economic Jean-Philippe Platteau, Eacultes Universitaires Research Council, Kenya Notre-Dame de la Paix, Belgiumn Mark Gersovitz, The Johns Hopkinzs Boris Pleskovic, World Bank University, USA Martin Ravallion, World Bank Jan Willem Gunning, Free University, Mark R. Rosenzweig, Harvard University, USA Amnsterdain, The Netherlands Joseph E. Stiglitz, Columbia U' '.. '.,'I USA Jeffrey S. Hammer, World Bank Moshe Syrquin, University of Miami, USA Karla Hoff, World Bank Vinod Thomas, World Bank The Worlk Bank Economic Review is a professional journal for the dissemination of World Bank-sponsored and outside research that may inform policy analyses and choices. It is directed to an international readership among economists and social scientists in government, business, and international agencies, as well as in universities and development research institutions. The Review emphasizes policy relevance and operational aspects of economics, rather than primarily theoretical and methodological issues. It is intended for readers familiar with economic theory and analysis but not necessarily proficient in advanced mathematical or econometric techniques. Articles will illustrate how professional research can shed light on policy choices. Inconsistency with Bank policy will not be grounds for rejection of an article. Articles will be drawn from work conducted by World Bank staff and consultants and from papers submitted by outside researchers. Before being acceptcd for publication, all articles will be reviewed by two referees who are not members of the Bank's staff and one World Bank staff member. Articles must also be recommended by a member of the Editorial Board. Non-Bank contributors are requested to submit a proposal of not more than two pages in length to the Editor or a member of the Editorial Board before scnding in their paper. Comments or brief notes responding to Review articles are welcome and will be considered for publication to the extent that space permits. Please direct all editorial correspondence to the Editor, The World Bank Economic Rrq,i¼,, The World Bank. 1818 H Street. Washington, DC 20433, USA, or wber@worldbank.org. SUBSCRIPTIONS: A subscription to The World Bank Economic Review (ISSN 0258-6770) comprises 3 issues. Prices include postage; for subscribers outside the Americas, issues are sent air freight. Annual Subscription Rate (Volume 18, 3 Issues, 2004): Academic libraries-Print edition and site-wide online access: US$113/C79, Print edition only: USS107/C75, Site-wide online access only US$101/C71; Corporate-Print edition and site-wide online access: US5137/£94, Print edition only USS130/C89, Site- wide online access only US$123/£84; Personal-Print edition and individual online access: US$42/£32. Please note: £ Sterling rates apply in Europe, USS elsewhere. There may be other subscription rates available; for a complete listing, please visit www.wbro.oupjournals.org/subscriptions. Readers with mailing addresses in non-OECD countries and in socialist economies in transition are eligible to receive complimentary subscriptions on request by writing to the UK address below. Full prepayment in the correct currency is required for all orders. Orders are regarded as firm, and payments are not refundable. Subscriptions are accepted and entered on a complete volume basis. Claims cannot be considered more than four months after publication or date of order, whichever is later. All subscriptions in Canada are subject to GST. Subscriptions in the EU may be subject to European VAT. If registered, please supply details to avoid unnecessary charges. For subscriptions that indude online versions, a proportion of the subscription price may be subject to UK VAT. Personal rates are applicable only when a subscription is for individual use and are not available if delivery is made to a corporate address. BACK ISSUES: The current year and two previous years' issues are available from Oxford University Press. Previous volumes can be obtained from the Periodicals Service Company, 11 Main Street, Germantown, NY 12526, USA. E-mail: psc@periodicals.com. Tel: (518) 537-4700. Fax: (518) 537-5899. CONTACT INFORMATION: Journals Customer Service Department, Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, UK. E-mail: jnls.cust.serv@oupjournals.org. Tel: +44 (0)1865 353907. Fax: + 44 (0)1865 353485. In the Americas, please contact: Journals Customer Service Department, Oxford University Press, 2001 Evans Road, Cary, NC 27513, USA. E-mail: jnlorders@oupjournals.org. Tel: (800) 852-7323 (toll-free in USA/Canada) or (919) 677-0977. Fax: (919) 677-1714. In Japan, please contact: Journals Customer Service Department, Oxford University Press, 1-1-17-SF, Mukogaoka, Bunkyo-ku, Tokyo, 113-0023, Japan. E-mail: okudaoup@po.iijnet.orjp. Tel: (03) 3813 1461. Fax: (03) 3818 1522. POSTAL INFORMATION: The World Bank Economic Review (ISSN 0258-6770) is published by Oxford University Press for the International Bank for Reconstruction and Development/THE WORLD BANK. Send address changes to The World Bank Economic Review, Journals Customer Service Department, Oxford University Press, 2001 Evans Road, Cary, NC 27513-2009. Communications regarding original articles and editorial management should be addressed to The Editor, The World Bank Economic Review, The World Bank, 1818 H Street, NW, Washington, D.C. 20433, USA. PERMISSIONS: For information on how to request permissions to reproduce articles or information from this journal, please visit www.oupjournals.org/permissions. ADVERTISING Inquiries about advertising should be sent to Helen Pearson, Oxford Journals Advertising, PO Box 347, Abingdon OX14 1GJ, UK. E-mail: helen@oxfordads.com. Tel/Fax: +44 (0)1235 201904. DISCLAIMVER: Statements of fact and opinion in the articles in The World Bank Economic Review are those of the respective authors and contributors and not of the International Bank for Reconstruction and Development/THE WORLD BANK or Oxford University Press. Neither Oxford University Press nor the International Bank for Reconstruction and Development/THE WORLD BANK make any representation, express or implied, in respect of the accuracy of the material in this journal and cannot accept any legal responsibility or liability for any errors or omissions that may be made. The reader should make her or his own evaluation as to the appropriateness or otherwise of any experimental technique described. PAPER USED: The World Bank Economic Review is printed on acid-free paper that meets the minimum requirements of ANSI Standard Z39.48-1984 (Permanence of Paper). INDEXING AND ABSTRACTING: The World Bank Economic Review is indexed and/or abstracted by CAB Abstracts, Current Contents/Social and Behavioral Sciences, Journal of Economic Literature/EconLit, PA4S International, RePEc (Research in Economic Papers), and Social Services Citation Index. COPYRIGHT (U The International Bank for Reconstruction and Development/THE WORLD BANK 2004 All rights reserved; no part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise without prior written permission of the publisher or a license permitting restricted copying issued in the UK by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London WlP 9HE, or in the USA by the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. THE WORLD BANK ECONOMIC REVIEW Volume 18 * 2004 * Number 1 ECONOMIC EFFECTS OF CHINA'S ACCESSION TO THE WORLD TRADE ORGANIZATION China's Accession to the World Trade Organization, Policy Reform, and Poverty Reduction: An Introduction 1 Deepak Bhattasali, Li Shantong, and Will Martin Impacts of China's Accession to the World Trade Organization 3 Elena Janchovichina and Will Martin Welfare Impacts of China's Accession to the World Trade Organization 29 Shaohua Chen and Martin Ravallion Tracking Distortions in Agriculture: China and Its Accession to the World Trade Organization 59 Jikun Huang, Scott Rozelle, and Min Chang Regulated Efficiency, World Trade Organization Accession, and the Motor Vehicle Sector in China 85 Joseph F. Francois and Dean Spinanger China in the World Trade Organization: Antidumping and Safeguards 105 Patrick A. Messerlin Keep up-to-date with the latest contents of The World Bank Economic Review by registering for our eTOC service at http://www3.oup.co.uk/jnls/tocmail. This service is freely available to all, no subscription required. China's Accession to the World Trade Organization, Policy Reform, and Poverty Reduction: An Introduction Deepak Bbattasali, Li Shantong, anzd Will Martin China's accession to the World Trade Organization (WTO) was a watershed event for both China and the WTO. After 30 years of effective isolation from the world economy, and close to a quarter century of autonomous reforms, China joined the legal framework of the world trading system. In doing so China made an extraordinarily wide-ranging set of commitments to reform of its own legal and administrative system and to thorough-going liberalization of trade in goods and services. This issue contains five studies from a inajor project undertaken by the World Bank and the Development Research Centre of China's State Council. A key objective of the studies was to assess the impact of the reforms associated with WTO accession on poverty in China, particularly in rural areas, which now lag so badly behind urban areas. A key underpinning of this work was an accurate assessment of the distor- tions in agricultural prices created by China's opaque system of pre-WTO agricultural trade policies. The results of previous studies were widely divergent, with some studies suggesting that China's agriculture was highly protected and others that it received low or negative protection. Huang, Rozelle, and Min attempt to resolve some of this contradiction by conducting detailed price comparisons and interviews with participants in key markets. They conclude that China's agricultural protection was generally modest, although some com- modities, such as sugar and maize, received substantial protection. Deepak Bhattasali is Lead Econiomist in PREMI Sector Department, World Bank Country Office, Beijing. His e-mail address is dhhattasali@a)worldhank.org. Li Shantong is Director-General, Research Department of Development Strategy and Regionial Econonmy, Development Research Centre of State Council, China; her e-mail address is shantong(adrc.gov.cn. Will Martin is Lead Economist, Development Research Group at the World Bank; e-miiail: wnmar-tinl('worldhanik.org. Particular thanks are due to the UK Department for Internationial Development for genierous support to this project. A companion hook. including a broader set of studies undertaken In this project, is forthcoming from Oxford University lPress and the World Bank (D. Bhattasali, Shantong Li, and W. Nlartin, eds., China a,znl the WTO: Accessionl, Policy Reforni, anid Potertv Reduiction Strategies). THE WORID BANK ECONOMiC RLVIENiE. V(L l8, No I, ©Cc The International Bank for Reconstructioni and Development / HOL WORLD BANK 2004; all rights reserved. DOI: 10.1093/wber/lhhO35 18:1-2 l 2 THE WORLD BANK ECONOMIC REVIEW, VOL. 1 8, NO. I lanchovichina and Martin draw on the estimates of protection provided by Huang, Rozelle, and Min and their own estimates of liberalization in other sectors to assess the impacts of reform on key economic outcomes such as wages, prices, and volumes of output, imports, and exports. A key finding is the importance of the abolition of the textile and clothing quota system, which will allow China to more fully exploit its comparative advantage in this area. In undertaking this analysis, they built on the study by Francois and Spinanger of adjustment in the motor vehicle sector-a sector that was highly protected and inefficient before accession. Francois and Spinanger conclude that this sector will substantially increase its efficiency, a result that has important implications for the economy as a whole, as well as for this sector's ability to meet import competition. Ravallion and Chen draw on a survey of 85,000 Chinese households to assess the impact of the WTO reforms on the distribution of income, particularly for the poor. They find negative impacts on average in rural areas, where the overwhelming majority of China's poor live, and benefits on average in urban areas. While the average losses to poor rural people are small, impacts on some types of households and certain regions are more substantial, with implications for social protection policies and for complementary policies to raise the incomes of poor families. Messerlin examines one of the controversial, and potentially risky, aspects of the agreement-the provisions on antidumping and safeguards. He points out that China is particularly vulnerable to antidumping actions because of its export product mix and because of the nonmarket economy provisions that other WTO members are allowed to apply to China for up to 15 years from its accession to the WTO. Possibly even more troubling are the product-specific safeguards allowed China's trading partners for up to 12 years from accession. Offering entirely new protection, these measures involve very lax triggers that could lead to a domino-effect closing of many markets for China's exports. Impacts of China's Accession to the World Trade Organization Elena Ianchovichina and Will Martin This article presents estimates of the impact of China's accession to the World Trade Organization. China is estimated to be the biggest beneficiary (US$31 billion a year from trade reforms in preparation for accession and additional gains of $10 billion a year from reforms after accession), followed by its major trading partners that also undertake liberalization, including the economies in North America, Western Europe, and Taiwan (China). Accession will boost manufacturing sectors in China, especially textiles and apparel, which will benefit directly from the removal of export quotas. Developing economies competing with China in third markets may suffer small losses. Accession will have important distributional consequences for China, with the wages of skilled and unskilled nonfarm workers rising in real terms and relative to those of farm workers. Possible policy changes, including reductions in barriers to labor mobility and improvements in rural education, could more than offset these negative impacts and facilitate the development of China's economy. Trade policy reforms such as those flowing from accession to the World Trade Organization (wTo) lead directly to changes in policy instruments, such as tariffs, nontariff barriers, and coverage of trade rules. The main policy concerns, however, are with the impacts on such economic variables as prices; output, employment, and trade volumes; factor returns; and household incomes. This article estimates the impacts on these key economic variables of China's accession to the wro as a guide to policy and as a basis for subsequent analysis at the household level.' It is part of a joint World Bank-Development Research Centre study reported in full in Bhattasali and others (2004). Elena lanchovichina is Economist with the Economic Policy Unit of the Poverty Reduction and Eco- nomic Management Network at the World Bank; her e-mail address is eianchovichina@worldbank.org. Will Martin is Lead Economist, Development Research Group at the World Bank; his e-mail address is wmartinl@worldbank.org. The authors thank Kyym Anderson, Hana Polackova Brixi, Louise Fox, Thomas Hertel, T. N. Srinivasan, Alan Winters, and three anonymous reviewers for helpful comments and Prashant Dave and Zhi Wang for their generosity in providing data. 1. Because China's accession was a necessary condition for that of Taiwan and because of the strong trade linkages between the two economies, the impact of Taiwan's accession to the wro is also considered. THE WORLD BANK ECONOMIC REVIEW, VOL 18, NO. 1, © The International Bank for Reconstruction and Development / THE WORLD BANK 2004; all rights reserved. DOI: 10.1093/wber/lhhO3O 18:3-27 3 4 TIlE WORI I) BANK ECONO.MIC RL\IEW, VOI.. I 8, NO. I The obvious instrument for performing this type of analysis is the computable general equilibrium model. Many such models now exist, and a cottage industry has emerged in estimating the impacts of trade reform in China (Gilbert and Wahl 2001).2 The availability of the internationally standard database of the Global Trade Analysis Project (GTAP) has facilitated such modeling and reduced the burden involved in obtaining estimates of basic information, such as trade flows and patterns of production and consumption. What standard models such as GTAP (Hertel 1997; www.gtap.org) do not do, however, is incorporate the nonstandard features of China's partially reformed economy, where many imports enter duty-free if used in the production of exports and labor market policies result in serious barriers between urban and rural areas. Like lanchovichina and Martin (2001) and Wang (2003), the analysis here explicitly allows for the duty exemption arrangements that result in close to half of China's imports entering as duty-free inputs into the production of exports. Recent work by Sicular and Zhao (2004) is drawn on to represent imperfect labor mobility and labor market distortions. This article extends the earlier work reported in lanchovichina and Martin (2001) by moving to the GTAP version 5 database (Dimaranan and McDougall 2002) based on 1997 data rather than 1995 data; by incorporating improved estimates of protection in agriculture (Huang and others 2004)3 and services (Francois and Spinanger 2004); by using measures of liberalization based on the final, multilateral agreement; by taking into account the restructuring of the automobile sector (Francois and Spinanger 2004); and by simulating the consequences of major labor market reform. This last issue is a particularly critical area for China, and there have been few simulation studies.4 This article first discusses the methodology and then the policy changes associated with China's WTO accession and the results of the simulation analysis. 2. Among the studies using a general equilibrium approach to quantify the impact of China's WTO accession are Lejour (2000), Zhai and Li (2000), Li and others (2000), McKihbin and Tang (2000), lanchovichina and Martin (2001), Walmsley and Hertel (2001), Deutsche Bank (2001), Wang (2002), Zhai and Wang (2002), Walmsley and others (forthcoming), and lanchovichina and Walmslev (2003). 3. Huang and others (2004) find that nominal rates of protection on important agricultural com- modities (rice, vegetables and fruits, livestock and meat) were negative in 2001 and are likely to remain unchanged in the postaccession period. Consequently, the reduction in agricultural protection is likely to he far less than presented in earlier studies. Nonetheless, greater scope for imports is likely for a range of agricultural products (wheat, oilseeds, sugar, and dairy products) protected by tariffs that are scheduled to be reduced substantially and for products (cotton and feedgrains) for which export subsidies are ruled out. These important findings were not incorporated in earlier studies. 4. The current treatment differs from that in Zhai and Wang (2002), who represent imperfect labor mobility in a single-country model with endogenous urban unemployment but do not differentiate between skilled and unskilled labor and employ a low level of elasticity of labor mobility between rural and urban areas (0.25). The results here are similar to those in Zhai and Wang (2002) in the case of wro accession with high labor mobility (elasticity of labor mobility is doubled) and fixed urban unemployment (flexible urban wvage). lanchovichina and Martin 5 It then considers some possible complementary policy actions, such as reducing barriers to rural outmigration and expanding access to education. I. METHODOLOGY The standard GTAP model was adjusted to incorporate China's important export processing arrangements.5 lanchovichina (2003) documents the approach used and shows that failing to account for China's duty exemptions in analyzing wro accession overstates the increase in China's trade flows by 40 percent and the increase in exports of selected sectors by 90 percent. The adjusted model (GTAP-DD) also incorporates some of China's key labor market mechanisms and institu- tions that related research has shown may have a major influence on the impacts of wro accession (Sicular and Zhao 2002, 2004). Export Processing Arrangements Export processing arrangements in China take many forms. Most arrangements allow firms producing goods for export to import intermediate inputs at world prices. These arrangements were incorporated in the GTAP-DD model used in this study by creating two activities for each sector. For sectors covered-or poten- tially covered-by export processing arrangements, one activity is specialized in production for export, and one is specialized in production for the domestic market. This separation is preferable to representations based on a single sector producing differentiated products for domestic and export markets because it allows the two sectors to use different input mixes, and it allows export-oriented activities to use much more import-intensive means of production. This provides a reasonably realistic depiction of China's trade regime in the 1990s, when duty exemptions were used to facilitate exports while protection in the rest of the economy remained fairly high.6 The tax arrangements for export processing (duty and value-added tax [VAT] exemptions on imported intermediate inputs and VAT refunds on exports) dis- couraged export-oriented firms from using domestic intermediate materials and 5. GTAP is a standard global applied general equilibrium model with perfectly competitive markets and constant returns to scale technology. The model represents consumer demands through a constant difference of elasticities functional form and on the supply side emphasizes the role of intersectoral factor mobility in the determination of sectoral output. Product differentiation between imported and domestic goods and among imports by region of origin allows for two-way trade in each product category, depending on the ease of substitution between products from different regions. Land, capital, skilled and unskilled labor, and in some sectors a natural resource factor are used in production and are fully employed. 6. In a deterministic world, a producer of exports will always take advantage of duty exemptions or rebates unless the administrative costs are excessive, which does not appear to be the case in China. Many studies have either ignored the problem or, as in Lejour (2000), treated duty exemptions as simple reductions in initial tariffs instead of exemptions on imports used specifically in the production of exports. 6 THE WORLD BANK ECONOMIC REVIEW, VOL. I8, NO. I selling in the local market. Furthermore, the vast majority of exports were produced using imported intermediates that were either exempt from duties or eligible for refunds on taxes paid.7 China customs data for 2000 show that 60 percent of imports entered duty-free, 41 percentage points of which were imports used for export processing, 13 percent were capital goods, and 6 percent were goods in special categories, such as materials used by research institutes. Rough calculations based on input-output information from the GTAP version 4 database (McDougall and others 1998) and data from China customs suggest that 23 percent of imports were used to produce for the domestic market and only about 3 percent were used to produce ordinary exports.8 Intermediate inputs for domestic and export-oriented activities were initially estimated by allocating them to each sector in proportion to sales in export and domestic markets. However, this yielded unsatisfactory results, with the data- base showing much less use of imported inputs in the export sector than the reported imports of duty-free intermediate inputs for export production obtained from China customs (Li Yan, personal communication). To deal with this, increased use of imported intermediates was allowed for in export activities in accordance with the price changes involved in providing duty exemptions, and the elasticities of substitution between domestic and intermedi- ate goods in the model.9 This more than doubled the import intensity of the exporting activities and reduced that of the domestically oriented activities. Labor Market Policies Perhaps the central labor market issue for the analysis is the barriers to mobility between rural and urban activities. Taking up employment in an urban area is inhibited by the need to obtain an urban residence permit (hukou). In addition, workers tend to be reluctant to permanently cut their ties with the rural sector because it is not generally possible to sell the land to which a rural family has usage rights (Hussain 2004). Many workers move temporarily from rural to urban areas, although restrictions are frequently imposed on such movements, and social welfare benefits (such as health care and schooling for children) 7. The export processing arrangements did not prevent firms producing 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 (lanchovichina 2003). 8. According to GTAP version 4, 14 percent of imports were for final consumption and according to China's customs 40 percent of imports were ordinary imports that were not duty-exempt. This means that about 26 percent were ordinary imports used as intermediates. Also according to GTAP version 4, China's firms exported an average of 10 percent of their output, implying that only about 3 percent of imports were used for production of ordinary exports. 9. The GTAP version 5 database (Dimaranan and McDougall 2002) is the source for the elasticities of substitution between domestic and composite imported commodities in the Armington produc- tion structure of a sector. The values for these elasticities are shown in column 1 of appendix table A.4 in the World Bank Policy Research Working Paper version of this article, available online at http://econ.worldbank.org/files/26864_wps3053.pdf. lanchovichina and Martin 7 enjoyed by urban residents are typically not available to temporary migrants. Although it is sometimes possible to overcome these problems by purchasing an urban residence permit, this imposes an additional cost on migrants from rural to urban areas, a group with particularly limited access to capital. As in all countries, rural-urban labor mobility is also inhibited by the sector-specific nature of farmers' human capital and a reluctance to cut family ties by migra- tion to urban areas. The per capita income of agricultural workers is only about one-third that of urban workers (World Bank 2002). Not all of this difference can be attributed to barriers to mobility between rural and urban areas, however. Urban workers typically have higher skills, work more intensively, and face higher costs of living than rural workers (Sicular and Zhao 2004). To capture the effects of barriers to mobility, the model allows for both imperfect transformation between unskilled workers in agricultural and unskilled nonagricultural employment, and an implicit tax on nonagricultural employment. The imperfect transformation is designed to reflect the substantial differences in the characteristics of unskilled workers in rural and urban areas, and the ability to transform (at a cost) agricultural workers into nonagricultural workers through training, experience, and the creation of nonagricultural jobs in rural areas. The tax is designed to reflect the pure policy-induced barriers between rural and urban workers, such as the requirement for a residence permit in urban areas and the barriers to mobility created by the inability to sell farm land. It is specified as a barrier that raises the cost of labor to urban employers, with urban workers receiving the tax-inclusive wage. The imperfect transformation between agricultural and nonagricultural workers is represented using a constant elasticity of transformation between workers in agriculture and workers in other sectors in the following simple manner: (1) LNF/LF = OS(WNF/WF)'. where LNF is the number of nonfarm unskilled workers, LF is the number of farm unskilled workers, a. is a constant term, WNF and WF are nonfarm and farm wages, and a is the elasticity of transformation. The value of a was set at 1.32, based on Sicular and Zhao's (2002) estimates of this parameter.?0 The tax 10. In a more recent work Sicular and Zhao (2004) estimate the responsiveness of rural labor supplv to changes in agricultural returns. They present two "push" elasticities-2.67 for nonagricultural wage employment and 0.24 for nonagricultural nonwage employment. Focusing on the push elasticity for nonagricultural wage employment and testing the sensitivity of the results bv replacing the elasticity of 1.32 used in this analysis with 2.67 leaves the aggregate results largely unchanged (appendix table A.7, available online at http://econ.worldbank.org/files/26864 wps3O53.pdf). The greater responsiveness of labor movement implied by the larger elasticity of transformation (2.67) translates into better poverty and inequality outcomes because farm wages remain nearly unchanged and an additional 1 million farm workers leave farming. 8 THE WORLD BANK ECONOMIC REVIEW, VOL. 1 8, NO. I reflecting pure policy-induced barriers between rural and urban wages was estimated at 34 percent by Shi (2002). Although there is some level of unem- ployment in China, particularly associated with state enterprise reform and rural-urban migration, much of this unemployment is transitional because of the weakness of the social safety net (Hussain 2004). Given the long-run focus of the analysis here, total employment is treated as exogenous. II. TRADE POLICY REFORMS This section examines the implications of trade policy reforms in China and in its trade partners in the years leading up to accession. Changes in China's Trade Policies During the 1990s China made substantial progress in reducing the coverage of nontariff barriers, lowering tariffs, and abolishing the trade distortions created by the exchange rate regime. Lardy (2002) estimates that the number of tariff lines subject to quotas and licenses fell from 1,247 in 1992 to 261 in 1999. By 2001, 257 tariff lines were covered by a combination of licenses and quotas and 47 by licenses only, whereas 245 were subject to designated trading and 84 to state trading. Tendering and other registration requirements, primarily for machinery and electrical products, covered an additional 120 tariff lines. By 2001, nontariff barriers covered 664 tariff lines, or less than 10 percent of tariff lines,'1 with over a third of these being subject to designated trading, one of the less intrusive forms of quantitative restriction employed in China. Data on nontariff barriers frequency alone may be misleading because of the enormous variations in the importance of tariff lines. To give some indication of the potential importance of nontariff barriers, the import coverage of the key nontariff barriers was calculated using data on nontariff barrier coverage of tariff lines and on import data by tariff line. The import coverage of all nontariff barriers in China fell from 32.5 percent in 1996 (World Bank 1997, p. 15) to 21.6 percent in 2001 (see appendix table A.2). Coverage of import licensing and quotas fell from 18.5 percent in 1996 to 12.8 percent in 2001, and coverage of state trading from 11 to 9.5 percent (table 1). The import coverage of tendering requirements fell particularly rapidly, from 7.4 percent in 1996 to 2.7 percent in 2001. Oil was by far the most important import subject to nontariff barriers, accounting for almost half the value of imports subject to any nontariff barriers (appendix table A.3). Ferrous metals, subject to designated trading arrange- ments, were the second most important category. Imports of oil and oil products accounted for 84 percent of total imports subject to state trading. 11. See appendix table A.1, available in the World Bank Policy Research Working Paper version of this article, accessible online at http://econ.worldbank.org/files/26864_wps3053.pdf. lanchovichina and Martin 9 TABLE 1. Import Coverage of Nontariff Barriers (%) Barrier 1996 2001 Licenses and quotas 18.5 12.8 Tendering 7.4 2.7 Licensing only 22 0.5 State trading I 1.0 9.5 Designated trading 7.3 6.2 Any nontariff barriers 32.5 21.6 No nontariff barriers 67.5 78.4 Total 100 100 Note: For 1996 nontariff harriers coverage, the trade weights used were for 1992, whereas for 2001 the trade weights used were for 2000. Souirce: For 1996, Lardy (2002); for 2001, Mei Zhen of the Ministry of Foreign Trade and Economic Cooperation during an internship at the World Bank, using data from the vTo. The average protective impact of the complete set 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 protective effect arising from licensing and quota- constrained goods. The protective effect of these nontariff barriers has clearly declined since then because of a number of factors, including the progressive phase- out of nontariff barriers, a standstill on new nontariff barriers during the accession process, the general wro prohibition against nontariff barriers after accession, and a likely reduction in the severity with which many of these measures have been administered. A simple rule of thumb that protection provided by nontariff barriers declines with their import coverage would suggest that their protective impact has fallen to about 5 percent. But given the large margin of uncertainty associated with this measure, the analysis here focuses only on tariff liberalization, implying that the results should be taken as a lower bound to the overall impact of liberalization. The pace of tariff reform in China was also rapid during the 1 990s. Although average tariffs were very high in the early 1990s, they fell sharply after 1994 (table 2). A significant tariff reform in October 1997 reduced average tariffs well below 20 percent. Three subsequent reductions at the beginning of 1999, 2000, and 2001 further reduced tariffs on a wide range of items. The progres- sive reductions in tariffs between 1992 and 2001 lowered average tariffs by two- thirds, with larger than average cuts in the manufacturing sector, thereby ensuring that the future reductions in tariffs required under the WTO accession agreement are much smaller than the reductions occurring before accession. Another important feature of the reforms has been a substantial reduction in the dispersion of tariff rates-with the standard deviation falling from 32.1 percent in 1992 to 10 percent in 2001. Examination of weighted average applied tariffs for 1995 and 2001 and after implementation of the final tariff bindings agreed in the accession schedule suggests that substantial merchandise trade liberalization occurred in China 10 THE WORLD BANK ECONOMIC REVIEW, VOL. I 8, NO I TABLE 2. China's Average Statutory Tariff Rates (%) All products Primary products Manufactures Year Simple Weighted Simple Weighted Simple Weighted 1992 42.9 40.6 36.2 22.3 44.9 46.5 1993 39.9 38.4 33.3 20.9 41.8 44.0 1994 36.3 35.5 32.1 19.6 37.6 40.6 1996 23.6 22.6 25.4 20.0 23.1 23.2 1997 17.6 18.2 17.9 20.0 17.5 17.8 1998 17.5 18.7 17.9 20.0 17.4 18.5 1999 17.2 14.2 21.8 21.8 16.8 13.4 2000 17.0 14.1 22.4 19.5 16.6 13.3 2001 16.6 12.0 21.6 17.7 16.2 13.0 After accession 9.8 6.8 13.2 3.6 9.5 6.9 Source: World Bank (1 999, p. 340) to 1998; authors' calculations for tariff lines with imports in 1999; CDS Consulting for protection data for 1999-2001; and after accession data from China's WTO final offer. Trade data come from COMTRADE. over the period 1995-2001 (table 3). Weighted average tariffs dropped substan- tially for wheat, beverages and tobacco, textiles, apparel, light manufactures, petrochemicals, metals, automobiles, and electronics. Analysis by Huang and others (2004) suggests that some agricultural commodities (such as vegetables and fruits, livestock and meat, and rice) faced negative protection in 1995, generally as a result of restrictions on exports. Protection on these commodities rose (or negative protection fell) over the period 1995-2001. Accession is not expected to lead to a significant fall in protection on most agricultural com- modities after 2001. Import protection is expected to remain unchanged for most commodities except oilseeds, sugar, and dairy products. Protection will continue to fall for all other merchandise commodities, with especially big cuts for processed food, beverages and tobacco, automobiles, electronics, and other manufactures. Francois and Spinanger (2004) conclude that the automobile sector liberalization will be accompanied by massive restruc- turing to realize economies of scale and improve structural efficiency, perhaps increasing productivity by 20 percent during the accession period (2001-07).) A key element of China's accession agreement is the abolition of agricultural export subsidies. This required some significant changes. Huang and others (2004) estimate that there was a 32 percent export subsidy on feedgrains and a 10 percent export subsidy on plant-based fibers in 2001 (particularly cotton). 12. Francois and Spinanger base their estimate of the 20 percent productivity increase on the distribution of current plants in China and apply the formula Aln(AC) = CDR * Aln(Q), where AC is average cost, MC is marginal cost, Q is the quantity produced, and CDR is the inverse elasticity of scale, defined as CDR = -(AC -MC)/AC, and varies between 0.125 and 0.135 (the range of values found in engineering studies). Then they calculate an average cost index for the industry. Assuming that the index is 100 at 350,000 units per plant, current plant structure yields a cost index of roughly 120. lanchovichina and Martin 11 TABLE 3. China's Import Protection before and after wro Accession (tariff or tariff equivalent, %) China Taiwan Product 1995 2001 Postaccession' 1997 2001 Postaccessiona Agriculture Rice -5.0 -3.3 -3.3 2.2 0.0 0.0 Wheat 25.0 12.0 12.0 6.5 6.5 6.5 Feedgrains 20.0 32.0 32.0 1.0 1.0 0.0 Vegetables and fruits -10.0 -4.0 -4.0 35.7 36.9 16.0 Oilseeds 30.0 20.0 3.0 1.8 0.8 0.2 Sugar 44.0 40.0 20.0 21.9 25.8 22.7 Plant-based fibers 20.0 17.0 20.0 0.0 0.0 0.0 Livestock and meat -20.0 -15.0 -15.0 7.5 6.5 4.0 Dairy 30.0 30.0 11.0 16.6 9.3 5.9 Total 4.8 7.6 3.6 9.1 6.9 4.6 Manufacturing Processed food 20.1 26.2 9.9 14.9 14.2 9.9 Beverages and tobacco 137.2 43.2 15.6 48.1 22.0 13.0 Extractive industries 3.4 1.0 0.6 5.5 5.5 4.1 Textiles 56.0 21.6 8.9 6.1 6.3 5.6 Apparel 76.1 23.7 14.9 12.8 13.4 11.2 Light manufactures 32.3 12.3 8.4 4.0 4.1 3.4 Petrochemicals 20.2 12.8 7.1 4.2 4.2 2.9 Metals 17.4 8.9 5.7 4.0 3.8 1.5 Automobiles 123.1 28.9 13.8 23.9 21.5 13.3 Electronics 24.4 10.3 2.3 2.9 0.5 0.3 Other manufactures 22.0 12.9 6.6 4.4 3.3 2.1 Total 25.3 13.5 6.9 6.3 5.2 3.5 Total merchandise tradeb 24.3 13.3 6.8 6.5 5.2 3.6 Services Trade and transport 1.9 1.9 0.9 1.3 1.3 0.7 Construction 13.7 13.7 6.8 5.9 5.9 2.9 Communications 9.2 9.2 4.6 9.2 9.2 4.6 Commercial services 29.4 29.4 14.7 3.7 3.7 1.9 Other 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 'Applied rates at the end of the implementation period were estimated as the lesser of the bindings and 2001 applied rates. In virtually all cases the bindings were lower than the applied rates. bEstimates 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 2007 are 12.2 percent in 2001 and 6.3 percent in 2007 for China and 4.5 percent and 3.1 percent for Taiwan. Source: Authors' calculations based on agricultural protection data from Huang and others (2002); manufacturing protection data from GTAP for 1995, from CDS Consulting for 2001, and from China's wTo final offer for protection after accession; and services protection data from Francois and Spinanger (2004). In addition to China's barriers on merchandise trade, border measures and domestic regulations on domestic service sectors and trade in these services have reduced the efficiency of services sectors. China has made substantial commitments 1 2 THE WORI.I) LiANK ECONO.II C RFV IFW, VO [. r 8, NO. I to open its services sectors,' and critical services such as telecommunications, logistics, and finance are likely to benefit from inward foreign direct investment and rising productivity as they are restructured. Based on work by Francois and Spinanger (2001) reported in Francois and Spinanger (2004), these measures are represented here as barriers to trade in services expressed in ad valorem terms. Following Francois and Spinanger (2004), the impact of accession is represented as halving the barriers to services trade. Efficiency improvements in the services sectors are not modeled because there were no reliable estimates of the likely productivity gains at the time this research was conducted. 4 Chaniges in China's Trade Partizers' Policies The arrangements for textiles and clothing are a particularly important element of China's accession. Unlike most other developing economy exporters, China was excluded from the liberalizing elements of the UJruguay Round Agreement on Textiles and Clothing. This means that prior to accession China did not benefit from the integration of textile and clothing products into the General Agreement on Tariffs and Trade (GAFT-) or from the increases in quota growth rates provided for under this agreement. That placed upward pressure on the prices of these quotas in China, raising the costs of exporters just as an equiva- lent export tax would.l5 Under its accession agreement, China benefited imme- diately from the integration of textiles and clothing into the GATT, leading to the abolition of quotas and increases in quota growth rates that have occurred since 1994 (WTO 1994). All quotas are to be phased out by 2005. Importing econo- mies will be allowed to introduce special textile safeguards during the period 2005-07, but these will be effective for only one year at a time. The accession agreement includes a transitional product safeguard mechanism that allows China's trading partners to take safeguard actions under rules that are more permissive of protection than the usual WVTO safeguard rules (Messerlin 2004). These provisions have the potential for introducing a new form of protec- tion against China. This potential danger needs to be weighed against the sub- stantial gains to China from her trading partners being required to follow WTO rules in implementing contingent protection measures against China. For simpli- city, these gains and losses are assumed to cancel each other out. 13. Nlattoo (2004) argues that China's commiiitmenits o*1 services were the most comprehensive ever made in the \xTO. 14. In a recent paper Mviai and others (2003) estimiiate productivity increases over 10 years of 1.8 percent a year for the strategic manlufacturing industries and 2.7 percent a year for the services sectors as reforms take place under \\'To accessioni. 15. These quotas are represented in the analysis as an export tax. In some cases the proceeds of this implicit export tax are redistributed to quota holders, who may he quite different from the producers and exporters of the goods. In other cases the quotas are aUctioned, with the quota rents accruing to the government. In either case, the marginal return froni additional output of textiles and apparel is net of the quota rent/export tax. lanchovichina and Martin 13 China's accession also triggered a liberalization of its partners' trade policies toward Taiwan (China) with average tariffs estimated to have fallen by almost 1.5 percentage points, from 4.5 percent in 1997 to 3.1 percent in 2001. Taiwan committed to tariff reductions on thousands of industrial and agricultural product lines, a phase-out of tariffs on a number of products as part of the Zero-for-Zero program of the Uruguay Round, and reductions in tariffs on chemical products as part of the Chemical Harmonization program. Tariffs on the vast majority of products related to information technology were also reduced in 2000 and, once WTO accession commitments are implemented, the tariff on electronic products will fall to 0.3 percent (see table 3). Taiwan, China, made horizontal and sector-specific commitments in the following service sectors: business, communication, construction, engineering, distribution, education, environmental, financial, health, social, transport ser- vices, and tourism and recreation. Following Francois and Spinanger (2001), Taiwan's wTo accession commitments are represented here by a halving of nontariff barriers to trade in services. III. EXPERIMENTAL DESIGN The impact of accession is evaluated here in the dynamic context of the growth and structural change expected in China and its trading partners during the period up to 2007, when almost all of the changes associated with accession will have come into effect. A baseline scenario is constructed under which the economies of the world grow and experience the manifold structural changes associated with economic growth up to 2007 (table 4 and appendix table A.6). The GTAP model includes key elements, such as changes in demand patterns as incomes rise, changes in the industrial structure associated with changes in the stock of capital per worker, and changes in world prices resulting from changes in world supply and demand, that allow the model to capture key changes in the world economy over this period. The baseline broadly replicates World Bank projections for overall growth in each region and uses projections of factor input growth and a residually determined level of total factor productivity growth to ensure consistency between the two. The model considers the effects of the WTO accession agreement signed at the Doha Ministerial Meeting in November 2001 on protection prevailing in 2001. In addition, it separately takes into account the liberalization in China during 1995-2001 because much of the liberalization during the 1990s was influenced by China's desire to prepare for the type of trade regime needed for WTro accession and to establish the credibility of its commitment to an open economy. These sharp reductions in protection are unlikely to have occurred without the prospect of accession to wro, and they have been locked in by China's wro commitments. Thus results are presented for both the accession period (2001-07) and the entire liberalization period (1995-2007). The analysis starts with 1995 tariff levels because 1995 was a major turning point in the negotiations-the closing of the door on China's attempt to enter 14 THE WORLD BANK ECONOMIC REVIEW, VOL. i 8, NO. I TABLE 4. Projected Growth in Factor Inputs and Total Factor Productivity in China's Trading Partners during 1997-2007 (percent) Unskilled Skilled Manufacturing Trading partner Population labor labor Capital TFP' North America 11(1.05) 11(1.08) 12 (1.11) 49 (4.07) High Western Europe 0 (0.03) -1 (-0.08) 1(0.07) 30 (2.69) High Australia and New Zealand 10 (0.98) 12 (1.14) 10 (0.99) 55 (4.45) High Japan 1(0.06) -2 (-0.19) -7 (-0.71) 35 (3.02) Medium China 8 (0.81) 13 (1.26) 50 (4.15) 174 (10.62) High Taiwan, China 9 (0.86) 11 (1.05) 14 (1.36) 96 (6.97) High Other newly industrialized 10 (0.93) -1 (-0.10) 55 (4.47) 88 (6.53) Medium countries in Asia Indonesia 16 (1.50) 17 (1.59) 123 (8.36) 25 (2.27) Low Vietnam 15 (1.40) 32 (2.79) 36 (3.10) 111 (7.78) Medium Other Southeast Asia 18 (1.70) 22 (2.04) 134 (8.87) 60 (4.83) Low India 18 (1.67) 23 (2.10) 78 (5.92) 88 (6.54) Medium Other South Asia 25 (2.22) 30 (2.69) 80 (6.06) 72 (5.55) Medium Brazil 14 (1.31) 19 (1.77) 72 (5.60) 31 (2.75) Medium Other Latin America 18 (1.68) 6 (0.57) 90 (6.65) 54 j4.42) Low Turkey 16 (1.47) 19 (1.75) 107 (7.55) 55 (4.46) Low Other Middle East 24 (2.16) 37 (3.23) 67 (5.24) 28 (2.50) Low and North Africa Economies in transition -1 (-0.11) 6 (0.56) 9 (0.90) 33 (2.88) High South African Customs Union 15 (1.39) 31(2.76) 47 (3.92) 34 (2.94) Low Other Sub-Saharan Africa 30 (2.65) 40 (3.42) 54 (4.42) 38 (3.26) Medium Rest of world 18 (1.63) 23 (2.10) 35 (3.05) 68 (5.32) Low Note: Numbers in parentheses are annual growth rates. a"Low" corresponds to average annual growth rates of 0.1 percent, "medium" to 1.0 percent, and "high" to between 2 percent and 4 percent. TFP =total factor productivity. Souirce: World Bank and GTAP data. the world trading system by resuming its status as a contracting party to the GATT. As Long (2000, p. 43) emphasized, China focused more strongly on commercial considerations after 1995 than it had previously-and its trading partners also strongly emphasized the commercial aspects of the negotiations. To capture the implications of WTO accession, 1997 protection data for China in the benchmark data (GTAP version 5) are adjusted to 1995 levels to obtain the initial base.16 For Taiwan, liberalization is considered to have begun in 1997, the year for which tariff data are available in GTAP version 5. Two experiments are conducted to evaluate the impact of WTO accession. The first assesses the impact of the fall in tariffs from 1995 to 2001 levels and the restructuring of the automobile sector accompanying the reductions in tariffs on automobiles and automobile parts during this period. The second assesses the 16. This adjustment was made with ALTLRTAX (Malcolm 1998), so that the consistency and the shares in the GTAP database would be preserved. lanchovichina and Martin 15 impact of the fall in tariffs from 1995 to postaccession (2007) tariff levels, liberalization of the services sectors, continued restructuring of the automobile sector, removal of quotas on China's clothing and textiles exports, and removal of China's agricultural export subsidies. The productivity shock designed to capture the restructuring of the automobile sector is proportionate to the fall in tariffs on automobiles in each simulation. The same macroeconomic closure is used for all experiments-full employ- ment, I perfect mobility of skilled and unskilled workers between nonagricultural sectors, and perfect mobility of unskilled workers within agriculture. Based on the working assumption of little induced change in net international capital flows, trade balances as shares of gross domestic product (CDP) were fixed for China and Taiwan. Although trade balances can be expected to vary during the transition, particularly if there are substantial changes in foreign investment levels, foreign investment levels are not determined within the model. 18 Taxes lost because of trade liberalization are assumed to be replaced by a uniform, nondistortionary consumption tax affecting both private and govern- ment final consumption of all goods and services. This hypothetical tax is included to ensure that any adverse impacts of trade reform on government revenues, and hence on the ability to provide income transfers or public services, are allowed for in the analysis of impacts of the reform on households. Because the GTAP version 5 database appears to represent the VAT on domestic production as an output tax, the model generates tax losses from the contraction of such industries as tobacco and alcohol. These inward-oriented industries have higher VAT rates than export-oriented sectors, such as apparel, because exports are exempt from the VAT. When the export-oriented sectors expand, the net impact of WTO accession is a sharp contraction in tax revenues. In reality, such a contraction will not be observed because VATS of the same magnitude are levied on imports. To offset this impact, particularly in the poverty analysis, the consumption tax had to be adjusted downward. This was done by first comput- ing the consumption tax that compensates for the loss in output taxes (this tax as a share of the total replacement tax is equal to the share of the output tax loss in the total tax losses) and then adjusting the consumption tax rate to eliminate the component due to the change in output taxes. To reflect the long-run change in the stance of trade policy, phased in over many years, involved in WrTo accession, most of the analysis uses a standard long-run specification, with capital and labor freely mobile between industrial sectors and within agriculture, although there are barriers to labor mobility between rural and urban employment. 17. The fixed employment assumption may understate the costs of accession to some degree. Zhai and Wang (2002), who explore the impact of W'TO accession on migration and unemployment, conclude that structural unemployment may rise following China's WX'TO accession as farmers move to urban areas. 18. The assumption of fixed trade balance as a share of GuM is required when evaluating welfare impacts using a static trade model such as GTAP. 1 6 THE WORLD BANK ECONOMIC REVIEW, VOL. I 8, NO. I IV. ASSESSMENT OF CHINA'S ACCESSION This section assesses the impacts of China's accession to the wTro on China and its trading partners. Impacts on China The focus here is on the impacts of the trade policy changes remaining after 2001 (detailed results for the period before 2001 are presented in appendix table A.5). A key feature of the period after 2001 concerns the effects of removing the quotas on apparel and textiles imposed on China and other developing economy exporters by major industrial country importers. These quotas are scheduled for abolition in January 2005 for all WTro members. Abolition gives a significant boost to the textile and apparel sectors in China, which had been one of the countries most tightly restricted by the quotas.19 Output in these sectors rises 16 percent and employment 57 percent (table 5). That in turn stimulates the production of plant-based fibers (mainly cotton), which increases by 16 percent. Output and employment in the other agricultural sectors, with the exception of livestock and meat, are expected to fall as unskilled agricultural labor moves into the textile and apparel sectors and unskilled nonfarm real wages rise (table 6). Oilseeds and sugar contract more than other agricultural sectors as a result of falling protection. Tariffs on oilseeds fall from 20 percent to 3 percent, and tariffs on sugar fall from 40 percent to 20 percent. Protection in other agricultural sectors is assumed to remain almost unchanged. The automobile and electronics sectors also expand slightly, creating employment opportunities, particularly for skilled labor.20 Results suggest that approximately 6 million farm workers in China will leave their farm jobs as a result of wro accession reform after 2001 in pursuit of employment in the nonagricultural sectors.21 For most merchandise goods, real wholesale prices fall as a result of trade liberalization after accession. Retail prices reflect a uniform consumption tax increase of about 1.9 percent levied to compensate for the loss of tariff revenue. For some products, such as beverages and tobacco, automobiles, and sugar, the fall in real retail prices reflects a larger than proportionate drop in protection. Increased demand for nonagricultural labor means higher real nonfarm wages and higher returns to nonagricultural labor relative to agricultural 19. This is a consensus finding supported by lanchovichina and Martin (2001), Deutsche Bank (2001), Wang (2002), and lanchovichina and Walmsley (2003). 20. The model underestimates the potential expansion and efficiency increase in the service sectors. With its promise to eliminate over the next few years, most restrictions on foreign entry and ownership, as well as most forms of discrimination against foreign firms (Mattoo 2004), China has set the stage for increases in foreign investment and productivity in these sectors. This in turn could lead to much larger income gains from wTo accession and larger increases in wages of skilled workers than shown here (see Walmsley and others forthcoming). 21. This estimate represents the number of "effective" farm workers likely to migrate from rural to urban areas based on employment data for 2000 from Cbina Statistical Yearbook (NBs 2001, pp. 111-12). lanchovichina and Martin 17 TABLE 5. Changes in China's Key Economic Indicators after 2001 as a Result of wro Accession (% unless otherwise indicated) Trade balance Wholesale Consumer Product Output Employment Exports Imports (US$ million) 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 -3.4 -3.7 14.6 -6.3 214 -1.9 -0.1 and fruits 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 Manufacturing Processed food -5.9 -6.4 11.4 62.6 -3,460 -1.7 -1.8 Beverages and -33.0 -33.1 9.7 112.4 -14,222 -1.8 -6.9 tobacco Extractive -1.0 -1.3 7.5 -4.4 2,088 -0.7 1.2 industries 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 3.7 3.7 5.9 6.8 1,786 -0.9 0.0 manufacturing 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 -2.1 -2.2 4.1 18.9 -11,291 -0.5 0.8 manufactures Services Trade and 0.0 0.0 0.8 -0.4 493 -0.2 1.6 transport 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 -2.0 -2.0 -0.4 35.4 -1,749 0.2 1.9 services Other services -1.7 -1.8 1.4 33.6 -1,525 -0.1 1.6 Total 1.0 0.Oa 16.8 17.3 717 -0.7 -0.2 'Reflects the fixed labor supply assumption. Source: Authors' simulations with modified GTAP model; see details in text. labor. Removal of protection on some agricultural sectors additionally lowers the attractiveness of farming and implies falling returns to farm labor and land. Real farm wages fall 0.7 percent, and the real rental price of land falls 5.5 percent. The decline in farm incomes and the rise in the real retail price of many nonfarm products mean that some farmers may be hurt by w-ro accession. Nonfarm wages rise 1.2 percent and skilled labor wages rise 0.8 percent, implying that workers in 1 8 THE WORLD BANK ECONOMIC REVIEW, VOL. i 8, NO. I TABLE 6. Change in Real Factor Prices in China as a Result of Accession, 2001-07 (%) Accession Accession Accession with with labor with increase labor market Accession market in skill reform and increase Item alone reform only level only in skill level Farm unskilled wages -0.7 16.8 1.6 19.4 Nonfarm unskilled wages 1.2 -3.8 2.7 -2.5 Skilled labor wages 0.8 -1.7 -6.3 -8.7 Rental price of land -5.5 -9.7 -6.4 -10.5 Rental price of capital 1.3 -1.4 0.9 -1.8 Price of capital goods -0.9 -3.6 -1.1 -3.9 Migration from rural 6 28 10 32 to urban jobs (millions) National welfare 10.0 11.0 10.0 11.0 (1997 US$ billion) Source: Authors' simulations with modified GTAP model; see details in text. urban centers-and farmers who are able to engage in nonfarm employment-are more likely to be better off as a result of WTO accession.22 Accession will make China a much bigger player in world markets through three channels-the rapid growth and structural change of its economy, the liberal- ization undertaken in preparation for wro accession, and the liberalization under- taken after accession in 2001. The liberalization undertaken after 2001 contributes to an increase in China's share in world exports from 4.4 percent to 7.8 percent on completion of accession. Similarly, China's share in world import markets rises from 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 an increase in export volume of about 106 percent, followed by textiles and automobiles. The dramatic fall in protection of beverages and tobacco results in imports more than doubling, followed by increases in imports of food products, textiles, agricultural products, automobile parts, and commercial services. China's total welfare gain from WTO accession is estimated at $40.6 billion per year (in 1997 dollars), or 2.2 percent of per capita real income (table 7).23 Most of 22. High unemployment due to the restructuring of state-owned-enterprises, privatization, and fierce competition in China imply that W'ro accession may dampen the effect on the wages of unskilled workers. By assuming full employment, the model overestimates the increase in wages of workers in the nonfarm sectors, underestimates the fall in the wages of farm workers, and overestimates the increase in total welfare. 23. These estimates are in agreement with findings in lanchovichina and Martin (2001) and Wang (2002). These are conservative estimates because they do not reflect income increases resulting from trade- and foreign direct investment-induced productivity gains, especially gains associated with liberal- ization of China's service sectors (Walmsley and others forthcoming). Transaction cost savings from developing institutions compatible with an open and modern market could be very large as well, but were not factored into the analysis. TABLE 7. Welfare Change and Sources of Welfare Change as a Result of China's WTO Accession (millions of 1997 US$) Export subsidy Liberalization Auto sector Country or group Total, 1995-2007 Tariff cuts Quota reductions reductions of services restructuring Impact 2001-07 North America 6,072 (0.0) 3,207 2,713 24 172 -44 5,259 Western Europe 18,189 (0.2) 9,724 8,285 -51 338 -107 14,200 Australia/New Zealand 136 (0.0) 175 -47 2 18 -12 152 Japan 5,694 (0.1) 5,522 291 -22 5 -102 2,553 Chinaa 40,552 (2.2) 29,452 2,389 275 1,160 7,276 9,563 Taiwan, China 2,985 (0.6) 2,300 338 -4 265 85 1,376 Other newly industrialized 6,831 (0.7) 6,539 -82 -185 49 511 1,456 countries Indonesia -408 (-0.2) -167 -216 -10 1 -16 -310 Vietnam -453 (-1.4) -63 -395 0 6 0 -405 Other Southeast Asia -585 (-0.1) -109 -464 -46 16 18 -268 India -3,357 (-0.4) -1,087 -2,338 -5 -23 96 -2,999 Other South Asia -1,622 (-0.8) -176 -1,427 -7 1 -12 -1,619 Brazil -76 (-0.0) -76 3 4 5 -12 359 Other Latin America -32 (-0.0) 59 -171 20 32 29 -36 Turkey -338 (-0.1) -50 -295 -2 7 2 -327 Other Middle East and 368 (0.0) 675 -467 -13 57 116 -365 North Africa Economies in transition 19 (0.0) 318 -321 4 15 3 -185 South African Customs Union 78 (0.0) 89 -18 0 5 2 13 Other Sub-Saharan Africa -45 (-0.0) 71 -159 4 15 24 -78 Rest of world 155 (0.0) 330 -210 -15 27 23 -78 World 74,166 56,733 7,409 -27 2,171 7,880 28,261 Note: Numbers in parentheses are percentage changes in per capita utility. The impact for 1995-2001 is the difference between the impact for 1995-2007 and the impact for 2001-07. aImpacts exclude output tax losses because of a compensating value added tax levied uniformly on both imported and domestic goods. Souirce: Authors' simulations with modified GTAP model; see details in text. 20 I H}F W 0 RI D B ANK FI ONONI(I(C RE VIEYEW, V ()1. 1 8, N(). I the gain ($31 billion) was realized following the massive liberalization between 1995 and 2001 and the ongoing restructuring of the automobile industry. The remaining reforms will lead to an additional welfare gain of $9.6 billion per year. The largest part of this gain in welfare will come from further 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 textiles and apparel and $1.2 billion (12 percent) from services liberalization. Continuing automobile sector restructur- ing will generate $1.1 billion (11 percent), and the removal of agricultural export subsidies will provide only $275 million (3 percent) in additional benefits. Impacts on Chiina's Trading Partners Among China's trading partners the largest absolute gains accrue to North Amer- ica and the Western Europe, with close to half of the gains coming from elimina- tion of the quotas they impose on China's exports of textiles and clothing-and thus elimination of the efficiency and rent transfers to China. North America, Western Europe, and Japan also gain from China's cuts in protection, which increase China's efficiency as an export supplier and its demand for their exports. Taiwan's welfare gain from its and China's accession to the WTO is estimated at $3.0 billion per year-the second largest gain relative to the size of the economy after China's (see table 6). About half of the gain ($1.6 billion) was realized as a result of the liberalization in China and in Taiwan during 1997-2001. Remaining reforms will lead to an estimated real income gain of $1.4 billion a year after 2001. Other newly industrialized economies also benefit from China's accession. Most of these benefits are associated with trade liberal- ization and removal of quotas on textile and apparel, which translate into gains from terms of trade improvements after 2001. The world as a whole and key developing economies that trade directly with China benefit from China's accession, but developing economies in Southeast Asia, South Asia, and Latin America that compete with China in third markets mav lose from the removal of textile and apparel quotas after 2001. The losses will be largest for Vietnam-an economy that is following in China's footsteps and has a similar pattern of comparative advantage in labor-intensive products. The welfare loss for Vietnam is estimated as a 1.4 percent drop in per capita income (see table 7). The loss to India is estimated to be considerably smaller as a share of per capita income, at 0.4 percent, whereas the percentage losses to other countries are very small. V. COMPLEMENTARY POLIC.Y REFORMIS Although the overall impacts of WTo accession on China's economy are generally positive, there are some concerns that declines in real returns to farm labor may exacerbate poverty in rural areas. Approaches that deal directly with these problems are more likely to succeed than approaches that attempt to water down China's trade policy reforms. Two policy tools that lend themselves lanchovichina and Martin 21 to analysis within the model framework used here are relaxation of the barriers to labor migration from rural to urban areas and skills upgrading for workers in rural areas. Impact of Reducing the Policy Barriers to Labor Mobility Abolishing policy barriers to labor mobility from rural to urban areas-such as residence permits, differences in social insurance, and the inability to sell agri- cultural land-in conjunction with accession leads to a nearly 17 percent increase in real returns to rural workers (see table 6).24 This contrasts sharply with the 0.7 percent reduction in real farm wages for accession without labor market reform. Rents to farmland would decline, with higher farm wages leaving a smaller residual return to farmland. Real urban unskilled wages would decline by an estimated 3.8 percent. Clearly, there would be scope for partial reform of these arrangements that could leave both farm and nonfarm unskilled workers better off than in the absence of labor market reform. These results suggest that this reform would have significant impacts on the number of people leaving their farm jobs for jobs in the nonfarm sectors and on the industry composition of China's economy. Some 28 million people would leave their farm jobs if the government removed the policy barriers to labor movement from rural to urban areas25_several times the estimated 6 million people who would move as a result of WTO accession reforms alone between 2001 and 2007. The impact on the composition of Chinese industrial output would also be substantial (table 8). This would allow not only apparel produc- tion to expand more but also metals, automobiles, electronics, machinery, other manufactures, and construction, all at the expense of reductions in some agri- cultural sectors. Impact of an Increase in Skill Level A key problem facing most rural workers is their low levels of education. One way to get a sense of the likely impacts of improving access to education is to consider the impact of resultant increases in the skill levels of rural workers on the perform- ance of the Chinese economy. This experiment looks only at the impact of improvements in education on the skills of rural workers. It ignores any potential benefits to rural households from improvements in access to education for their children-such as reductions in school fees-and any changes in the government budget associated with increases in government spending on education.26 24. Zhai and Wang (2002) obtained similar results for a combination of \W'TO accession and full labor market reform. 25. Because the tax on nonfarm employment of 34 percent represents a bundle of policies that act as a barrier to rural to urban migration, this estimate is representative of the likely impact and could change depending on the policy mix the government adopts. 26. The model does not track education spending as a component of the government budget constraint. TABLE 8. Change in Output and Employment in China as a Result of WTO Accession and Other Reforms (percent change over the period 2001-07) Output Employment Without With With With labor market Without With With With labor market biikou labor market increase in reform and increase labor labor market increase in reform and increase Product removal reform skill level in skill level market reform reforn skill level in 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 i'-) Maufactturing Processed 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 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. Ianchovichina anid Martin 23 An increase in the provision of education that would boost the annual growth rate for skilled labor from 4. 15 percent to 5 percent and would lead to a decline in the annual growth rates for unskilled labor from 1.26 percent to 1.1 percent was considered. This was found to have important impacts on the structure of the Chinese economy. An increase in skilled labor leads to a stronger expansion, or a smaller contraction, in the manufacturing sectors that are skilled labor- intensive than does accession with labor market reform but no change in education spending (table 8). Metals, automobiles, electronics, and other manufactures all expand. Although output in some sectors expands, the real wages of skilled workers fall as the supply of skilled workers increases (see table 6) and world prices of the outputs they produce decline. This contrasts with the case of accession alone, which results in an increase in the real wages of skilled workers. How- ever, the real wages of generally much poorer unskilled workers rise with increased education, with the wages of unskilled nonfarm workers rising more than those of unskilled farm workers (see table 6). Of course, those who are able to transfer from agricultural to nonagricultural employment as a result of increased educational opportunities are likely to be substantially better off. Overall, it is clear that increased education spending will generally induce propoor growth and decrease poverty. It certainly has the opportunity to sub- stantially offset the adverse impacts on rural labor of the trade reforms asso- ciated with accession. Finally, increased education boosts the need for migration as demand for unskilled workers increases in large urban areas. An estimated 10 million farm workers are expected to exchange farm jobs for nonfarm ones. The impact on consumer prices is small-with falling prices for farm products and rising prices for manufactured commodities (see table 8). Impact of Labor Market Reform and an Increase in Skill Levels The combination of removing labor market barriers and increasing education spending creates the most favorable scenario for unskilled farm labor, leading to the largest increase in real farm wages (19.4 percent; see table 6). Farm output contracts more than in the case of labor market reform alone, whereas skilled labor-intensive industries such as metals, automobiles, electronics, other manu- factures, and services expand more than in the case of labor market reform alone or increased education spending alone (see table 8). Under this scenario, an estimated 32 million farm workers would leave their farm jobs for jobs in urban areas (see table 6). These results suggest that to generate propoor growth over the next decade, the government should consider both removing policy barriers to labor movement and changing the composition of spending to favor education. Not only would these policies facilitate the transformation of China's eco- nomy toward services and high-tech manufacturing sectors, but they also have the potential to more than offset any negative impacts of accession on rural wages and incomes. 24 THE WORLD BANK ECONOMIC REVIEW, VOL. i 8, NO. I VI. CONCLUSION The analysis suggests that the reforming economies and their close trading partners will be the biggest beneficiaries of accession to the WTO. China is undertaking the greatest reform and will gain the most. The North American and Western European economies that abolish their export quotas on textiles and clothing and increase their direct trade with China will gain the most in absolute terms. Taiwan will benefit substantially, both as a consequence of its own liberalization and through strengthened trade links with China. Japan will gain substantially because of increased export opportunities in China and China's increased competitiveness as a supplier. Other industrializing and indus- trialized economies that are China's largest trading partners will also be sub- stantial gainers. China's WTO accession will have a noticeable impact on the level and pattern of global trade. With accession, China is becoming a much bigger player in world markets. Apparel exports will lead China's export expansion, followed by textiles and automobiles. In addition to being an important source of traded goods, China will become an important destination for other economies' pro- ducts. Imports of beverages and tobacco will more than double, followed by imports of food products, textiles, agricultural products, automobile parts, and commercial services. The expansion of textiles, light manufactures, petrochem- icals, and equipment exports from Taiwan will be driven almost entirely by demand for these products in China. Accession will have important distributional consequences for China. The wages of skilled workers and unskilled nonfarm workers will rise in real terms and relative to the wages of farm workers. An estimated 6 million people will leave their farm jobs in 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 real retail prices of many nonfarm products suggest that some farmers may be hurt by WTO accession after 2001 (see Chen and Ravallion 2004). To help offset these adverse impacts on farmers, the Chinese government might make changes in its labor market policies. Abolition of the bukou system and reform of the labor market more generally would raise farm wages and allow 28 million people to migrate to nonfarm jobs in search of a better life. It would lead to an even bigger expansion of the labor-intensive manufacturing sector. An increase in skill levels would have a positive impact on the structure of the Chinese economy. As the supply of skilled workers increased, the real wages of skilled workers would fall, while the real wages of unskilled workers would rise with increased education spending. Thus, on the income side, increased education spending would induce propoor growth and decrease poverty and inequality. A number of caveats are important. The gains to China and Taiwan are probably understated because tariff aggregation in the GTAP model hides much of the variation in tariffs and the welfare gains from reducing this variation within the product aggregates used in the analyses (Bach and Martin 2001; lanchovicbina and Alartin 2i Martin and others forthcoming). When Bach and others (1996) adjusted for this in a partial equilibrium context, gains to China almost doubled. The analysis here assumes flexible wages and full employment. However, trade liberalization and foreign competition may worsen unemployment and put downward pres- sure on the wages of unskilled workers in the short run. Furthermore, although the analysis here improves on lanchovichina and Martin (2001), with better treatment of the extent of liberalization in agriculture and services and the changes in the automobile sector, there are still areas that have been ignored. One is nontariff barriers in the manufacturing sectors other than the quotas on apparel and textiles. Another is the impact of accession on foreign direct investment and the hard-to-measure efficiency gains in services that are asso- ciated with this increased investment. Foreign direct investment has contributed significantly to China's economic growth and will play an important role as China continues to reform its economy. wTo accession is likely to increase foreign direct investment in China, as trade liberalization improves returns to investment and the liberal- ization of rules on investment eases financial flows into previously restricted sectors such as services and automobile production. The substantial productiv- ity gaps between local and foreign firms imply that new foreign direct invest- ment will raise productivity.2 Walmsley and others (forthcoming) take into account both the impact of foreign direct investment and increased productivity growth in services. They find that the impacts of accession are far larger than those predicted by earlier studies, including this one, which ignore the potential productivity gains in the services sector and abstracted from capital accumula- tion and foreign investment. REFERENCES Bach, C., and W. M%artin. 2001. "Would the Right Tariff Aggregator for Policy Analksis Pleasc Stanid Up?" Journal of Policy Mfodelinzg 23:621-35. Bach, C., W. Martin, and J. Stevens. 1996. "China anid the wro: Tariff Offers, Exemptionis anld Welfare Implications." Weltwirtschaftliches Arcbiv 132(3):409-3 l. Bhattasali, D., Shantong Li, and W. Martin, eds. Forthcoming. China and the vT-(:: Accesaino, Policy Reform, anid Poverty Reduicti(n Strategies. Washington, D.C.: Oxford University Press and the World Bank. Chen, S., and M. Ravallion. 2004. "Household Welfare Impacts of China's Accession to the WVorld Tracde organization." World Bank Economic Review i8:29-58. Claro, S. 2001. "Tariff and FDI Liberalization: What ro Expect from China's Entry into 55w--?" Paper presented at the Eighth Annual Conference on Empirical fnvestigations in International Trade, PUrdUe University, West Lafayette, Ind., November 9-11. 27. Claro (2001) estimates that in the apparel and footwear industries, the adoption of foreign technology raises productivity by 30-62 percent in collective enterprises and 20-59 perceint in state enterprises. 26 THE WORLD BANK ECONOMI.C REVIEW, VOL. I8, NO. I Dimaranan, B., and R. McDougall. 2002. Global Trade, Assistance, and Production: The GTAP 5 Data Base. West Lafayette, Ind.: Purdue University, Center for Global Trade Analysis. Deutsche Bank. 2001. "Quantifying the Impact of China's wTFo Entry." DB Global Market Research. Online document available at www.deutsche-bank.de/index_e.htm. Francois, J., and D. Spinanger. 2001. "Greater China's Accession to the WTO: Implications for Interna- tional Trade/Production and for Hong Kong." Paper prepared for the Hong Kong Trade Development Council, Hong Kong, December. - 2004. "Regulated Efficiency, World Trade Organization Accession, and the Motor Vehicle Sector in China." World Bank Economic Review 18:XX-XX. Gilbert, J., and T. Wahl. 2001. "Applied General Equilibrium Assessments of Trade Liberalization in China." World Econonmy 25(5):697-731. lHertel, T., ed. 1997. Global Trade Analysis: Modeling and Applications. Cambridge: Cambridge Uni- versity Press. Huang, J., S. Rozelle, and C. Min. 2004. "The Nature of Distortions to Agricultural Incentives in China and Implications of Accession to the World Trade Organization.' World Bank Economic Review 18:XX-XX. Hussain, A. 2004. "Impact of wTo Membership on Inequality and Poverty." In D. Bhattasali, Shantong Li, and W. Martin, eds., China and the W'TO: Accession, Policy Reform, and Poverty Reduction Strategies. Washington D.C.: Oxford University Press and the World Bank. lanchovichina, E. 2003. "`(TAP-DD: A Model for Analyzing Trade Reforms in the Presence of Duty Drawbacks." (GTAI' Technical Paper 21. Purdue University, West Lafayette, Ind. Available online at www.gtap.org/resources/tech-papers.asp and forthcoming in Journal of Policy Modelling. lanchovichina, E., and W. Martin. 2001. "Trade Liberalization in China's Accession to WTO." Journal of Economic Integration 16(4):421-45. lanchovichina, E., and T. Walmsley. 2003. "Impact of China's WTro Accession on East Asia." Policy Research Working Paper 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 and the World Economy." Netherlands Bureau for Economic Policy Analysis. The Hague, Netherlands. Li, S., Z. Wang, F. Zhai, and L. Xu. 2000. wTO: China and the World. Beijing: China Development Press. Long, Yongtu. 2000. "On the Question of Our Joining the World Trade Organization." Chinese Economy 33(1):5-52. Mai, Y., M. Horridge, and F. Perkins. 2003. "Estimating the Effects of China's Accession to the World Trade Organization." CoPS General Paper, no. G-137. Centre of Policy Studies and Impact Project, Clayton, Australia. Malcolm, G. 1998. "Adjusting Taxes in the (TAP Data Base." GTAP Technical Paper 12. Center for Global Trade Analysis, Purdue University, West Lafayette, Ind. Available online at www.gtap.org/resources/ tech_papers.asp. Martin, W. 2001. "Implications of Reform and wTo Accession for China's Agricultural Trade Policies." Economics of Transition 9(3):717-42. Martin, W., D. van der Mensbrugghe, and V. Manole. Forthcoming. "Keeping the Devil in the Details: Aggregation and the Benefits of Trade Liberalization." Policy Research Working Paper. World Bank, Washington, D.C. Mattoo, A. 2004. "China's Accession to the WT-o: The Services Dimension." In D. Bhattasali, Shantong Li, and W. NMartin, eds., China and the wX'TO: Accession, Policy Reform, and Poverty Reduction Strate- gies. Washington D.C.: Oxford University Press and the World Bank. McDougall, R., A. Elbehri, and T. Truong. 1998. Global Trade, Assistance, and Production: The GTAP 4 Data Base. Purdue University, Center for Global Trade Analysis, West Lafayette, Ind. McKibbin, W., and K. Tang. 2000. "Trade and Financial Reform in China: Impacts on the World Economy." World Economy 23(8):979-1003. Ianchovichina and Martini 27 Messerlin, P. 2004. "Contingent Protection." In D. Bhattasali, Shantong Li, and W. Mlartin, eds., China and thewTo: Accession, Policy Reform, and Poverty Reduction Strategies. Washington D.C.: Oxford University Press and the World Bank. NBS (National Bureau of Statistics of China). 2001. Chinza Statistical Yearbook. China: China Statistics Press. Shi, Xinzheng. 2002. "Empirical Research on Urban-Rural Income Differentials: A Case of China." Working Paper, Peking University, Beijing. Sicular, T., and Y. Zhao. 2002. "Employment, Earnings and Labor Market Responses in Rural China." Working Paper, Peking University, Beijing, and University of Western Ontario, London. - 2004. "Labor Mobility and China's Entry to the WTO." In D. Bhattasali, Shantong Li, and W. Martin, eds., China and the WTO: Accession, Policy Reform, and Poverty Reduction Strategies. Washington D.C.: Oxford University Press and the World Bank. Walmsley, T. L., and T. W. Hertel. 2001. "China's Accession to the wro: Timinig Is Everything." World Economy 24(8):1019-49. Walmsley, T. L., T. Hertel, and E. lanchovichina. Forthcoming. "Assessing the Impact of China's wTo Accession on Foreign Ownership." Pacific Economic Review. Wang, Z. 2002. "W'ro Accession, 'Greater China' Free Trade Area and Economic Relations across the Taiwan Strait." Paper presented at the Fifth Conference on Global Economic Analysis, June 5-7, Taipei. - 2003. "The Impact of China's xro Accession on Patterns of World Trade." Journal of Policy Modeling 25(1):1-41. World Bank. 1997. China Engaged: Integration wvith the World Economy. Washington, D.C. - 1999. World Development Indicators. Washington, D.C. - 2002. China: Country Economic Memorandumiz. Washington, D.C. WTO (World Trade Organization). 1994. The Results of the Uruguay Rounzd of Multilateral Trade Negotiations. Geneva. Zhai, Fan, and S. Li. 2000. "The Implications of Accession to WTO on China's Economy." Paper presented at the Third Annual Conference on Global Economic Analysis, June 28-30, Monash University, Melbourne, Australia. Zhai, Fan, and Z. Wang. 2002. "wTo Accession, Rural Labour Migration and Urban Unemployment in China." Urban Studies 39(12):2199-217. Welfare Impacts of China's Accession to the World Trade Organization Shaobua Chen and Martin Ravallion ; Data from China's national rural and urban household surveys are used to measure and explain the welfare impacts of changes in goods and factor prices attributable to accession to the World Trade Organization. The price changes are estimated separately using a general equilibrium model to capture both direct and indirect effects of the initial tariff changes. The welfare impacts are first-order approximations based on a household model incorporating own-production activities calibrated to household-level data and imposing minimum aggregation. The results show negligible impacts on inequality and poverty in the aggregate. However, diverse impacts emerge across household types and regions, associated with heterogeneity in consumption behavior and income sources, with possible implications for compensatory policy responses. There has been much debate about the welfare impacts of greater trade openness. Some argue that external trade liberalization is beneficial to the poor in devel- oping economies, whereas others argue that the benefits will be captured by people who are not poor. Expected impacts on relative wages (notably between * skilled and unskilled labor) and relative prices (such as between food staples and luxury imports) have figured prominently in debates about the welfare impacts. What does the evidence suggest? One might hope to provide a conclusive answer by comparing changes over time in measures of inequality or poverty between countries that are open to external trade and those that are not. A number of attempts to throw empirical light on the welfare effects of trade liberalization have been made using aggregate cross-country data sets that combine survey-based measures of inequality or poverty with data on trade Shaohua Chen is senior information officer at the World Bank; her e-mail address is schen@worldbank. org. Martin Ravallion is research manager at the World Bank; his e-mail address is mravallion@a worldbank.org. The authors thank Tamar Manuelyan-Atinc and Will Martin, who were instrumental in encouraging them to undertake this task and provided many useful comments along the way. Elena lanchovichina and Will Martin provided the estimates of the price impacts of China's trade reform that are crucial to this study, as well as many helpful comments. Yan Fang, Honge Gong, Liqun Peng, Pingping Wang, and Min Yuan helped greatly in matching variables from China rural and urban household surveys to the categories of the general equilibrium model. The comments of Francois Bourguignon, John Cockburn, Neil McCulloch, Johan Mistiaen, Sangui Wang, participants at the Fourth Asian Development Forum in Seoul, the National Bureau of Statistics in Beijing, and the journal's three anonymous referees are gratefully acknowledged. THE WORLD BANK ECONOMIC REVIEW. VOL 18, NO. 1, ©3 The International Bank for Reconstruction and Development / THE WORLD BANK 2004; all rights reserved. DOI: 10.1093/wber/lhhO31 18:29-57 29 30 THE W'ORLD BANK ECONOMIC REVIEW, VOL. 18, NO. 1 openness and other control variables (see Bourguignon and Morisson 1990; Edwards 1997; Barro 2000; Dollar and Kraay 2002; Lundberg and Squire 2003). However, there are reasons to be cautious in drawing implications from such cross-country comparisons. There are concerns about data and econometric specifications. Differences in survey design and processing between countries and over time within countries can add considerable noise to the measured levels and changes in poverty and inequality. It is unclear how much power cross-country data sets have for detecting any underlying effects of greater openness or other covariates. There is also the issue of whether trade volume can be treated as exogenous in these cross-country regressions; it is clearly not a policy variable and may well be highly correlated with other (latent) attributes of country performance independently of trade policy. The attribution of inequality impacts to trade policy reforms themselves is clearly problematic. The correlations (or their absence) found in cross-country studies can also be deceptive because starting conditions can vary so much between reforming countries. Averaging across this diversity in initial conditions can readily hide systematic effects of relevance to policy (Ravallion 2001). In principle, such problems in cross-country comparative work can be dealt with by better data and methods. However, the concerns go deeper. Aggregate inequality or poverty may not change with trade reform even though there are both gainers and losers at all levels of living. Survey data tracking the same families over time commonly show considerable churning under the surface.1 The data show that many people have escaped poverty, whereas others have fallen into poverty, even though the overall poverty rate is unchanged. Numerous sources of such diverse impacts can be found in developing econ- omy settings. For example, geographic disparities in access to human and physical infrastructure affect prospects for participating in economic growth.2 For China, the economic geography of poverty and how this interacts with geographic diversity in the impacts of policy reforms are high on the domestic policy agenda. A policy analysis that simply averaged over such differences would miss a great deal of what matters to the debate on policy. This article follows a different approach, in which the attribution to trade policy changes is unambiguous and the diversity of welfare impacts is not lost. The article examines the welfare impacts at the household level of the changes in commodity and factor prices attributed to a specific trade policy reform, namely, China's accession in 2001 to the World Trade Organization (wro). For China, this meant a sharp reduction in tariffs, quantitative restrictions, and export subsidies, with implications for the domestic structure of prices and wages and thus for household welfare. Drawing on estimates by lanchovichina and Martin 1. Jalan and Ravallion (1998) report evidence of such churning 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). Cbest and Ral'allion 3 1 (2002) of the impacts of reform on prices (for both commodities and factors of production), the following analysis applies standard methods of first-order welfare analysis to measure the gains and losses at the household level using large sample surveys collected by China's National Bureau of Statistics (NBS). I. MEASURING THE WELFARE IMPACTS OF TRADE REFORM Previous approaches to studying the welfare impacts of specific trade reforms have tended to be either partial equilibrium analyses, which measure household- level welfare impacts of the direct price changes due to tariff changes using survey data (typically) covering many thousands of randomly chosen households, or general equilibrium analyses, which use a computable general equilibrium (CGE) model to capture second-round responses.-3 Although partial equilibrium analysis requires little or no aggregation of the primary household data, it misses poten- tially important indirect effects on prices and wages. General equilibrium analy- sis has the power to capture these effects by simulating economywide impacts on markets. However, standard CGE models entail considerable aggregation across household types, with rarely more than six or so representative households. Such models are crude tools for welfare-distributional analysis. The challenge for applied work is to find an approach that respects the richness of detail available from modern integrated household surveys while ensuring that the price changes attributed to reform are internally consistent with economywide equilibrium conditions. In principle, the CGE model could be built onto the household survey, so that the number of households in the model is the number sampled in the survey.4 For this study, that degree of integration would require an extraordinarily high dimensional CGE model, with 85,000 households. This is currently not a feasible route. The intermediate approach used here carries the reform-induced commodity and factor price changes simulated from a general equilibrium model to the level of all the sampled households in the survey.5 The welfare impacts are measured using standard tools of analysis familiar from prior work on the welfare effects of price changes associated with tax and trade policy reform. This approach imposes minimal aggregation conditions on the survey data within unavoidable 3. Examples of partial equilibrium analysis of the welfare distributional effects of price changes include King (1983), Deaton (I1989), Ravallion (I 990), Ravallion and van de Walle (1991), and Friedman and Levinsohn (2002). OnI applications to tax policy reform, also see Newbery and Stern (1987). On CCCE models see Decaluwve and MTartens ( 1988) and Hertel ( 1997). 4. The only known example of this full integration is Cockburn (2002), who built a classic trade- focused CCo model onto the Nepal Living Standards Survey covering about 3,000 households. 5. In an antecedent to the approach take here, Bourguignon and others (2003) also take price changes generated by a (,F model to survev data (for Indonesia). Methodologically, the main difference is that they generate income impacts at the household level from a microeconometric model of income determi- nation, whereas this study derives first-order welfare impacts analytically from a standard competitive farm-household model. 32 THE WORLD BANK ECONOMIC REVIEW, VOL. I 8, NO. 1 data limitations. In addition to calculating the trade reform's overall effects on poverty and inequality, this approach provides a detailed socioeconomic map of impact, showing how it varies with other nonincome characteristics, such as location. This generates better insights to the questions policymakers ask about who gains and who loses from reform. The general equilibrium analysis generates a set of price and wage changes. These embody both the direct price effects of the trade policy change and second- round, indirect effects on the prices of nontraded goods and on factor returns, including effects operating through the government's budget constraint. lanchovi- china and Martin (2002) use a competitive market-clearing model from the Global Trade Analysis Project (GTAP).6 The revenue implications of the trade policy change are reflected in changes in indirect tax rates.7 Because the price changes are based on an explicit model, their attribution to the trade policy reform is unambiguous, thus avoiding the identification problems common to previous attempts to esti- mate distributional effects of trade policy reform using cross-country comparisons. The approach can be outlined as follows. Each household has preferences over consumption and work effort (under the standard assumption that goods have positive marginal utilities, whereas labor supply has negative marginal utility) represented by the utility function ui(qd, Li), where qid is an m-dimension vector of the quantities of commodities consumed by household i and L, is a vector of labor supplies by activity, including supply to the household's own production activities. The household is assumed to be free to choose qi and Li subject to its budget constraint. Consistently with the general equilibrium model that generated the price and wage changes, there is no rationing at the house- hold level; for example, involuntary unemployment is ruled out. The indirect utility function of household i is given by (1) ''i[pj, wi, 7ri] = max [u,(q1 L,) Pq= wiL, + 7i] (q,L,) where p41 is the price vector (of dimension m) for consumption, wv, is the vector of wage rates, and it, is the profit obtained from all household enterprises as given by (2) iri(p,, pd, 'V ) = max[psq - p zi- wiL7 qs, < f,,(zij, LL), j=l ..,m; Zzij < z,,Eli < LO] I i where ps is the m-vector of supply prices, q' is the corresponding vector of quantities supplied, Lo is the labor input to own-production activities, of which II~~~~~~~~~~~~~~~~~~ Lj°j is used in producing good j, fj, is the household-specific production function 6. Hertel (1997) contains descriptions of the standard GTAP model with applications. 7. A full discussion of the assumptions of the general equilibrium model and the results of its application to China's accession to the WTro can be found in lanchovichina and Martin (2002). Cben and );.' .ll. 33 for good j (embodying fixed factors), and the zi terms are the commodities used as production inputs, of which zii is used in producing good j. Measurement of the welfare impacts is of course constrained by the data, which do not include initial price and wage levels.8 However, this data limita- tion does not matter in calculating a first-order approximation to the welfare impact in a neighborhood of the household's optimum. Taking the differentials of equations 1 and 2 and using the envelope property (whereby the welfare impacts in a neighborhood of an optimum can be evaluated by treating the quantity choices as given), the monetary value of the change in utility for household i is given by (3) g,- d p Pd, ps - P(qi-t+zii) p di E WkL W where v,j is the marginal utility of income for household i (the multiplier on the budget constraint in equation (1) and L' = L,k - L' is the household's "external" labor supply to activity k. (Notice that gains in earnings from labor used in own- production are exactly matched by the higher cost of this input to own-production.) Equation 3 is the key formula used in calculating the household-level welfare impacts of the price changes implied by the general equilibrium analysis of the trade policy reform. The proportionate changes in all prices and wages are weighted by their corresponding expenditure and income shares. The weight for the proportionate change in the jth selling price is p',q~, the revenue (selling value) from household production activities in sector j. Similarly, - pd(qd z) is the (negative) weight for demand price changes, and wkL,k is the weight for changes in the wage rate for activity k. The term ps qs - p ±d(qd + z,j) is referred to as net revenue, which (to a first-order approximation) gives the welfare impact of an equiproportionate increase in the price of commodity j. With the gain (or loss) to each household calculated based on equation 3, the covariates of those gains can now be examined. One covariate of obvious interest is income, needed to assess impacts on aggregate poverty and inequality. Ideally, one would use a money metric of utility based on equation 1. However, that would require an explicit model of the demand and supply system (that can be integrated back to obtain the indirect utility function). Again, feasibility becomes an issue because of the absence of complete data on price and wage levels. Thus there is little choice but to use income as the money metric of utility, in effect ignoring all geographic differences in the prices faced or in the extent to which border price changes are passed on locally. However, we make a seemingly plausible allowance for urban-rural cost of living differences in this setting. 8. For food items, unit values can be calculated (expenditure divided by quantitv) from the survey data, but there is no such option for food inputs to production, nonfood commodities consumed or used in production, or wages (the survey data do not include labor supplies or quantities consumed of nonfood goods, including production inputs). 34 THE WORLD BANK ECONOMIC REVIEW, VOL. I8, NO. I Two further limitations of this approach should be noted. First, applying the calculus in deriving equation 3 implicitly assumes small changes in prices. Relaxing this requires more information on price levels and the structure of the demand and supply system.9 This would entail considerable further effort, and the reliability of the results will be questionable given the problem of incomplete price and wage data. Second, as already noted and consistent with the general equilibrium analysis, this approach also rules out rationing in commodity or factor markets or nonconvexities in consumption or production. In principle, these problems can also be handled through a completely specified demand model (which can be used to estimate the virtual prices at which the rationed demand or supply would be chosen). This is not feasible without data on price and wage level. II. SETTING AND DATA Although the official date of China's WTO accession is 2001, it is clear that the Chinese economy had already started to adapt to this expected change well before that time. The trade reform can thus be thought of as having two stages-a lead-up period in which tariffs started to fall in anticipation of wro accession and the period from 2001 onward. lanchovichina and Martin (2002) argue for 1995 as a plausible beginning of the lead-up period. Their estimates of the price changes induced by WTO accession for the periods 1995-2001 and 2001-07 are used in this analysis. Although the primary focus is on the second period, welfare impacts are also estimated for the lead-up period. The measure of welfare impacts given by equation 3 is calibrated to survey data for 1999, two years before official wro accession and a few years after the likely beginning of the lead-up period. The choice of 1999 was partly made for data reasons, because it was the most recent year for which the micro-data were available. Choosing a year near the middle of the lead-up period (rather than a survey at the beginning or end) should also diminish biases due to any non- linearity in the welfare impacts of price and wage changes. Survey Data The survey data used in this study are from the 1999 Urban Household Survey and the 1999 Rural Household Survey by China's NBS. The sample size is 67,900 households for the rural survey and 16,900 households (out of the survey total of 40,000 households) for the urban survey.'0 Over the past 15 years, the NBS has worked to improve both surveys, focusing on sample coverage, questionnaire 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 40,000 households, but until 2002 the central NBS office kept individual record data for only 16,900 households. Cben and Ravallionz 35 design, methodology, and data processing."1 The number of variables in the surveys has increased dramatically, with additional details on income, expendi- ture, savings, housing, and productivity, among others. The NBS also provided micro-data for three provinces (Liaoning, Guangdong, and Sichuan-the test provinces). A computer program to implement the estimation method was written for these data, after which the program was run by NBS staff on the entire national data set. A number of problems remain in the 1999 surveys. For a sample frame, the rural survey relies on its sampled counties from 1985, which may no longer be representative. The urban survey excludes rural migrants, because the base of the sample frame is the legal registration system (hukou). As in other countries, the rural survey gives data on the remittances of migrant workers, but it does not provide information about the migrant workers themselves, who (unlike in other countries) are not sampled in the urban survey either. This makes it difficult to measure impacts through labor mobility and rural-urban transfers. Comparisons between the rural and urban surveys also present problems. For example, income in the rural survey includes in-kind income (such as from own- farm production and other household enterprises), but income in the urban survey ignores some in-kind components, notably subsidies from the govern- ment. Sampling Weights The population census puts the 1999 urban population share at 34 percent, whereas the sample-based urban population share is 20 percent. To correct the rural and urban sampling weights, the urban population share from the China Statistical Yearbook (NBs 2000) was used to replace the survey sample weights to form the national figures. Matching the Global Trade Analysis Project Model and the Surveys There are 57 sectors in the 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 and meat, dairy, processed food, beverages and tobacco, extractive industries, textiles, apparel, light manufac- tures, petrochemicals, metals, automobiles, electronics, other manufactures, trade and transport, construction, communications, commercial services, and other services. To these are added land, capital, and three types of labor (see later discussion). China's rural and urban surveys have about 2,000 categories for consump- tion and production. The variables from the household surveys are matched to the closest category in the GTAP model. For example, corn, millet, and potatoes 11. For further discussion in the context of the Rural Household Survey, see Chen and Ravallion (1996). 36 THE WORLD BANK ECONOMIC REVIEW, VOL. 1 8, NO. I are placed in the category feed grains and cotton and fiber crops are placed in the category plant-based fibers. (The working paper version of this article, Chen and Ravallion 2003, gives details on how the variables from the surveys are matched to the GTAP model sectors.) Definitions of Labor and Labor Earnings The China GTAP model defines three types of labor: unskilled farm labor, unskilled nonfarm labor, and skilled nonfarm labor.'2 Because the rural and urban surveys have different questionnaires, rural and urban labor earnings are treated differently. In the urban survey three variables-sector, occupation, and education-are used to determine labor types. But sector or occupation alone cannot indicate whether a person should be classified as skilled labor. For example, the financial sector may hire some unskilled labor and the services sector may hire some skilled labor. Similarly, a train driver in the occupation category "workers and staff-members in production and transportation" counts as skilled labor. Therefore, education is also taken into account. Workers who have received education at the senior high school level or higher are considered skilled labor. Others are classified as unskilled labor. It is more difficult to determine the type of labor income for rural areas. There is no information on how much each person earns and from what work. Consequently, labor earnings can be classified only roughly by income source. For instance, all labor remuneration from agriculture is considered income from unskilled farm labor; earnings from industry or construction, grain processing, and the like are considered income from unskilled nonfarm labor; earnings from the services sector, transportation and trade, and the like are considered income from skilled nonfarm labor. Land Under China's economic reforms, which began in 1978, all farmers have land- use rights but not the right to sell, although they can subcontract the allocated land to other farmers. Therefore, the change in land prices from the GTAP model affects only the value of land rentals paid and received. Household Income For assessing the overall impacts on poverty and inequality, rural and urban households are combined. There is no cost of living index between urban and rural areas of China. (Urban and rural consumer price indexes are both indexed to 100 at the base date.) The urban price level is assumed to be 15 percent higher than the rural price level. This differential is less than that for other 12. By the International Labour Organization's definitions, skilled labor consists of managers and administrators, professionals, and para-professionals, and uinskilled labor consists of tradespeople, clerks, salespeople and personal service workers, plant and machine operators and drivers, laborers, and related workers and farm workers. Chen and Ravallion 37 developing economies because subsidies to urban households in China help compensate for higher housing and food costs than in rural areas. Income per person is used as the welfare indicator (so that all households are ranked by per capita income, from the poorest to the richest). This is termed "net income" in the rural survey and "disposable income" in the urban survey. Postreform income is then income plus the estimated gain defined by equation 3. III. MEASURED WELFARE IMPACTS OF WTO ACCESSION Based on the predicted relative price and wage changes from the GTAP model for 1995-2000 (table 1) and 2001-07 (table 2) and production and consumption shares from the 1999 rural and urban household survey data, equation 3 can be used to compute the net gain for each household. The first panel in table 3 gives the mean gains for 1995-2001 and 2001-07, split by urban and rural areas. The second panel gives the Gini indices, both actual (for the baseline year, 1999) and simulated. The two simulated income distributions are obtained by subtracting the estimated gains over 1995-2001 from the 1999 incomes at the household level and by adding the household-specific gains from 2001-07 to the 1999 incomes. Thus the first simulation shows the distributional impact of the price changes during the first stage of the trade reform (what the baseline distribution would have looked like without the reforms) and the second shows the impact of the post-2001 price changes (how the changes are expected to affect the baseline distribution, looking forward). The third panel gives the headcount index of poverty for the official poverty line based on the poverty lines used by China's NBS and for the $1/day and $2/day poverty lines from Chen and Ravallion (2001). There is an overall gain of about 1.5 percent of mean income. All of this gain is in the period leading up to wTo accession. There is almost no impact on inequality, either in the period leading up to WTO accession or predicting forward. The aggregate Gini index increased slightly, from 39.3 percent without wro accession to 39.5 percent after accession. The incidence of poverty would have been slightly higher in 1999 if not for the trade policy changes in the lead-up to wro accession, whereas poverty is predicted to increase slightly during 2001-07 due to the expected price changes induced by the remaining tariff changes during that period. The impacts on rural and urban poverty for a wide range of poverty lines can be seen in figure 1, which gives the cumulative distributions of income for both the baseline and the two simulated distributions for the poorest 60 percent in rural areas and 40 percent in urban areas. Although there is virtually zero aggregate impact when predicting forward from wro accession, the disaggregated results show a more nuanced picture. The analysis focuses on three measures of impact at the household level: the absolute gain or loss, gj; the proportionate gain or loss, g,/yi; and whether there is a gain or not, I(g,), where I is the indicator function. This third measure helps determine where there might be high concentrations of losers in specific areas or socioeconomic groups. TABLE 1. Predicted Price Changes from GTAP Model and per Capita Net Gain or Loss for Rural and Urban Households, 1995-2001 Rural Urban Expenditures and Wholesale Consumer Net revenue Mean welfare Net revenue Mean welfare income sources prices (%) prices (%) (yuan) change (yuan) (yuan) change (yuan) Expenditures Rice 0.5 1.5 73.66 0.15 -109.33 -1.64 Wheat -1.7 -1.5 40.86 -0.74 0.00 0.00 Feedgrains 2.6 10.7 117.04 2.15 0.00 0.00 Vegetables and fruits 0.5 1.5 123.41 0.13 -378.69 -5.68 Oilseeds -0.6 -0.8 37.05 -0.24 -1.04 0.01 Sugar 0.7 1.4 13.74 0.05 -174.06 -2.44 Plant-based fibers -3.6 -1.9 36.84 -1.34 (.00 0.00 Livestock and meat 2.0 3.1 194.62 2.59 -500.65 -15.52 Dairy 1.5 2.5 2.50 0.02 0.00 0.00 Other food 1.2 3.1 -81.60 3.39 -343.13 -10.64 Beverages and tobacco -4.6 -7.2 -72.98 5.25 -197.20 14.20 Extractive industries -0.2 0.8 17.99 0.44 -173.03 -1.38 Textiles -5.0 -8.9 -11.08 0.99 -53.50 4.76 Apparel -2.7 -7.4 -64.13 4.75 -394.30 29.18 Light manufacturing -0.3 -2.5 -16.15 0.40 -82.96 2.07 Petrochemical industry -0.7 -0.1 -325.39 0.33 -398.23 0.40 Metals -0.7 --0.1 - 15.30 0.02 -24.02 0.02 Autos -17.7 -20.4 -52.27 10.66 -37.76 7.70 Electronics -1.5 -4.0 -24.27 0.97 -162.69 6.51 Other manufactures -0.6 -0.3 -264.61 0.79 -431.16 1.29 Trade and transport 0.2 1.3 -18.70 -0.24 -110.53 -1.44 Construction 0.1 1.1 0.00 0.00 -31.11 -0.34 Communication 0.9 1.9 -16.72 -0.32 -152.04 -2.89 Commercial services 0.8 1.8 -61.37 -1.10 -533.33 -9.60 Other services 0.1 1.1 -414.45 -4.56 -680.99 -7.49 Income souirces Farm unskilled labor 1.7 1.7 313.58 5.22 0.00 Nonfarm unskilled 1.7 1.7 287.19 4.78 1,227.51 20.44 Skilled labor 2.0 2.0 360.87 7.09 3,391.11 66.64 Land 1.3 1.3 17.08 0.22 0.00 Capital 1.3 1.3 21.14 0.27 126.01 0.77 Source: lanchovichina and Martin (2002) and author's computations based on China NBS 1999 Rural Household Survey and 1999 Urban Household Survey. TABLE 2. Predicted Price Changes from GTAP Model and per Capita Net Gain or Loss for Rural and Urban Households, 2001-07 Rural Urban Expenditures and Wholesale Consumer Net revenue Mean welfare Net revenue Mean welfare income sources prices (%) prices (%) (yuan) change (yuan) (yuan) change (yuan) Expenditures Rice -1.4 0.7 73.66 -1.39 -109.33 -0.75 Wheat -1.5 0.7 40.86 -0.92 0.00 0.00 Feedgrains -3.7 2.1 117.04 -4.90 0.00 0.00 Vegetables and fruits -2.6 -0.6 123.41 -4.02 -378.69 2.24 Oilseeds -5.7 -5.9 37.05 -2.10 -1.04 0.06 Sugar -2.8 -3.5 13.74 -0.34 -174.06 6.01 Plant-based fibers 1.6 4.1 36.84 0.56 0.00 0.00 Livestock and meat -1.5 0.7 194.62 -5.21 -500.65 -3.40 Dairy -2.4 -0.5 2.50 -0.09 0.00 0.00 Other food -3.1 -2.7 -81.60 2.04 -343.13 9.32 Beverages and tobacco -5.6 -7.7 -72.98 5.62 -197.20 15.09 Extractive industries -0.4 1.7 17.99 -0.86 -173.03 -2.92 Textiles -0.2 -1.5 -11.08 0.17 -53.50 0.82 Apparel 2.6 0.8 -64.13 -0.51 -394.30 -2.98 Light manufacturing -0.6 0.5 -16.15 -0.08 -82.96 -0.43 Petrochemical industry -1.1 0.8 -325.39 -2.60 -398.23 -3.19 Metals --0.6 1.3 -15.30 0.20 -24.02 -0.31 Autos -3.8 -4.0 -52.27 2.09 -37.76 1.52 Electronics -1.2 -1.4 -24.27 0.34 -162.69 2.20 Other manufactures -0.8 0.8 -264.61 -2.12 -431.16 -3.46 Trade and transport -0.4 1.7 -18.70 -0.32 -110.53 -1.85 Construction -0.4 1.7 0.00 0.00 -31.11 -0.52 Communication -0.4 1.7 -16.72 -0.28 -152.04 -2.54 Commercial services -1.1 0.9 -61.37 -0.55 -533.33 -4.72 Other services -0.7 1.3 -414.45 -5.39 -680.99 -8.76 Income sources Farm unskilled labor -0.3 -0.3 .313.58 0.85 Nonfarm unskilled 1.0 1.0 287.19 2.96 1,227.51 12.64 Skilled labor 0.4 0.4 360.87 1.55 3,391.11 14.58 Land -4.7 -4.7 17.08 -0.8( Capital 0.6 0.6 21.14 0.13 126.01 (.80 Souirce: lanchovichina and Martin (2002) and authors' computations based on China NBS 1999 Rural Household Survey and 1999 Urban Household Survey. 42 THE WORLD BANK ECONOMIC REVIEW, VOL. I8, NO. I TABLE 3. Summary Statistics on Aggregate Welfare Impacts, 1995-2001 and 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 39.31 Simulated: Less gains 1995-2001 33.90 29.68 39.27 Simulated: Plus gains 2001-07 34.06 29.65 39.53 Poverty impacts (headcount index, I. . L .;. Official poverty line Baseline, 1999 4.38 0.08 2.92 Simulated: Less gains 1995-2001 4.56 0.08 3.04 Simulated: Plus gains 2001-07 4.57 0.07 3.04 $1/day (1993 purchasing power parity) Baseline, 1999 10.51 0.29 7.04 Simulated: Less gains 1995-2001 10.88 0.28 7.28 Simulated: Plus gains 2001-07 10.81 0.28 7.23 $2/day (1993 purchasing power parity) Baseline, 1999 45.18 4.07 31.20 Simulated: Less gains 1995-2001 46.10 4.27 31.88 Simulated: Plus gains 2001-07 45.83 3.97 31.60 'Percentage of mean income. bOfficial poverty line is from China NBS; $1/day and $2/day poverty lines are from Chen and Ravallion (2001). Source: Authors' computations based on China National Bureau of Statistics 1999 Rural Household Survey and 1999 Urban Household Survey. The results by provinces ranked by mean income per person are plotted in figure 2a for mean absolute gains (gi in yuan per capita), in figure 2b for proportionate gains (g,Iyi, as a percentage), and in figure 2c for the proportion of households that registered positive gains. (The average gain or loss by province for urban and rural areas and the number of gainers in each case are shown in appendix tables A.1 and A.2; Chen and Ravallion 2003 gives the province rankings.) The same results are also plotted in figure 3 against percentiles of the income distribution. So, for example, to see the mean impact in yuan per capita at the median income, one looks at the 50th percentile of figure 3a. (Notice that figure 3a gives the horizontal differences in figures I a and l b plotted against the point on the vertical axis.) In the aggregate, about three-quarters of rural households and one-tenth of urban households will experience a real income loss. Farm income is predicted to drop by 18 yuan per person, whereas urban income rises by 29 yuan per person. The breakdown by sectors in table 2 shows that the decline in rural Chen and Ravallion 43 FIGURE 1. (a) Rural and (b) Urban Poverty Incidence Curves A 60 _ 60 ~~~~~~~~~Pre WTO > /(middle line) 20 t 20 400 600 800 1000 1200 1400 1600 1800 2000 B 40 c) Pre WTO income C- ~~~~~~~~~Baseline distribution 30 in 1999 (middle line) Post WTO C, DI 0 .0 O- 1000 2000 3000 4000 Annual per capita income (Yuan) Soturce: China National Bureau of Statistics and Chen and Ravallion 2001. income is due to the drop in wholesale prices for most farm products, plus higher prices for education and health care. Farmers will also benefit from the drop in some consumer prices and from the increase in nonfarm labor wages. In urban areas residents will enjoy lower prices for most farm products and higher wages, but they will also be hit by increases in service fees for education and health care. Impacts differ considerably across regions (see figure 2 and appendix tables A.1 and A.2). The mean absolute gains tend to be highest in the richest provinces in both urban and rural areas (figure 2a), though there is no correla- tion between the proportionate gains and mean income of provinces (figure 2b). One spatially contiguous region-the northeast provinces of Inner Mongolia, Liaoning, Jilin, and Heilongjiang-stands out as having the largest loss from the reform. Both absolute and proportionate impacts are highest in this region- more than 90 percent of farmers in Heilongjiang and Jilin are predicted to experience a net loss. 44 THE WORLD BANK ECONOMI'C REVIEW, VOL. i 8, NO. I FIGURE 2. (a) Mean Absolute Gain or Loss by Province Ranked by per Capita Income. (b) Mean Proportionate Gain or Loss by Province Ranked by per Capita Income. (c) Mean Percentage of Gainers by Province Ranked by per Capita Income A 60 4 0 , -nN/- \ rUrban 40 - 60 -80 1 6 11 16 21 26 31 B m0 T t I ty, X6 -2 - . Rural \ I z -3- 1 6 11 16 21 26 31 C 100 80 o~~~~~~~oa 40 20 20 1 6 11 16 21 26 31 Provinces ranked by provincial per capita income Source: Authors' computations based on data from China National Bureau of Statistics 1999 Rural Household Survey and 1999 Urban Household Survey. Chen and Ravallion 45 FIGURE 3. (a) Mean Gain or Loss by Population Ranked by per Capita Income Percentile (Yuan). (b) Mean Percentage Gain or Loss Ranked by per Capita Income Percentile. (c) Percentage of Gainers by Income Percentile A 40 20 Urban 50 -S~~~~~~~~~~~~~~, X 0- ~~~~~~~~~~National ~ Rural : -40 -60 () 20 40 60 80 100 80 20 460 0 10 20 '- ,,^/ <2\8R Rural 0 20 40 60 80 100 Percenit of population ranked by per capita income (assumzes urban price Is 15°o bigher) Note: Urban prices are assumed to be 1S percent higher than rural prices. Souerce: Authors' compilation based on China National Bureau of Statistics 1999 Rural Household Survey and 1999 Urban Household Survey. 46 THE WORLD BANK ECONOMIIC RFVIEW, V'OL. 18, NO. I Notice that these geographic differences in welfare impacts arise entirely from differences in consumption and production behavior. In reality, there are also likely to be differential impacts on local prices due to transport or other impediments to internal trade. The analysis here does not incorporate such differences, and doing so would pose a number of data and analytic problems. This might, however, be a fruitful direction for future work where the necessary data on prices and wage levels are available by geographic area. When households are ranked by initial income, there is a notable difference between urban and rural households, with absolute gains tending to be higher for higher-income households in urban areas but lower for higher-income households in rural areas (see figure 3a). Nationally (combining urban and rural areas with the corrected weights), there is the hint of a U-shaped relation- ship, but still with the highest absolute gains 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 nationally (see figure 3b). In the aggregate the proportion of gainers rises with income, a result that is driven by the rise in the number of gainers as income increases in rural areas (see figure 3c). IV. EXPLAINING THE INCIDENCE OF GAINS AND LossEs The way the problem of measuring welfare impacts was formulated in section II allows utility and profit functions to vary between households at given prices. To explain the heterogeneity in measured welfare impacts, these functions can instead be supposed to vary with observed household characteristics. The indirect utility function becomes (1')~ ~ ~~V Pi(p ,IV,,7ri Vpi ,wi,7[ij,xIi where (2') 7ri = 7 (pi, p,ti, l 2ix)i for vectors of characteristics x1i and x2i that shift the utility functions in equation 1 and the profit functions in equation 2. Note that the characteristics that influence preferences over consumption (xl,) are allowed to differ from those that influence the outputs from own-production activities (x2,). The gain from the price changes induced by trade reform, as given by equation 3, depends on the household's consumption, labor supply, and pro- duction choices, which in turn depend on prices and characteristics, xti and x2i. For example, households with a higher proportion of children will naturally spend more on food, so if the relative price of food changes, the welfare impacts will be correlated with this aspect of household demographics. Similarly, there Chen and Ravallion 47 may be differences in tastes associated with stage of the life cycle and education. There are also likely to be systematic covariates of the composition of income. Generically, the gain can be written as (4) g, = g (Pi ,p11,s'xliiX21) = E [ppqS ( pd l psitvix2i) s - pd [qP (Pi Wt'7rixii) +Z11 (P P.P, xv> P +Zvk LLk (piw117rixii) -L, (pc,',ps,wjx2,,)] dk) Notice that the gain from reform is inherently nonseparable, in that it cannot be written as a function solely of p, xi,, and 7ri because the gain also depends on production choices. However, as noted in section II, household-specific wages and prices are not observed, so further assumptions are required. In explaining variations across households in the predicted gains from trade reform, wage rates are assumed to be a function of prices and characteristics as lvi = tv(pM,pfx1, 1X2i), and differ- ences in prices faced are assumed to be adequately captured by a complete set of county-level dummy variables. Under these assumptions, and the linearization of equation 4 with an additive innovation error term, the following regression model applies for the gains: (5) gl = Oixii + 32x2i + ^ ikDki + E, k where Dki= 1 if household i lives in county k and 0 otherwise and i, is the error term. The characteristics considered include age and age squared of the household head, education and demographic characteristics, and land (interpreted as a fixed factor of production because it is allocated largely by administrative means in rural China). Also included are dummy variables describing some key aspects of the occupation and principle sector of employment, such as whether the household is a registered agricultural household, whether there is wage employment, whether there is state-sector employment, and whether there is participation in a township and village enterprise. There are endogeneity concerns about these variables, but they appear to be minor in this context, especially when weighed against the concerns about omitted variable bias in estimates that exclude these characteris- tics. Under the usual assumption that the error term is orthogonal to these regres- sors, 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 which complete micro-data are available. There are some differences in the explanatory variables between urban (tables 4 and 5) and rural areas (tables 6 and 7). Results are presented for both absolute 4 8 THE WORLD BANK FCONON1IC RBVILW, Vol . I 8, NO. I TABLE 4. Regression Results for Level of Gain (Yuan) in Rural Areas of Three Provinces, 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 employees/household size 12.665 (4.10) -6.932 (-0.29) 12.652 (3.02) Number township and village enterprise 10.768 (3.13) 29.466 (3.06) 15.327 (4.26) workers/household size Number of migrant workers/household 5.399 (1.73) 7.798 (2.35) 7.067 (3.79) 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 level Illiterate 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) 1M1iddle 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 size 46.999 (2.71) 8.109 (0.35) -2.291 (-0.45) Children 6-11/household size 1.414 (0.11) 2.247 (0.10) -9.011 (-1.50) Children 12-14/household size -0.155 (-0.01) -24.489 (-1.20) -9.606 (-1.51) Children 15-17/household size -2.592 (-0.22) -23.390 (-1.02) -5.485 (-0.73) Constant -17.851 (-0.82) -17.742 (-0.65) -17.220 (- 1.43) R2 0.278 0.116 0.116 Note: Numbers in parentheses are t-statistics. Souirce: Authors' computations based on China NBS 1999 Rural Household Survey and 1999 Urban Household Survey. gains (gi) and proportionate gains (glIyi). Recall that these are averages across the impacts of these characteristics on the consumption and production choices that determine the welfare impact of given price and wage changes. This makes interpretation difficult. These regressions are mainly of descriptive interest to help isolate covariates of potential relevance in thinking about compensatory policy responses. For rural areas, the results show that the predicted gain from trade reform tends to be larger for larger households 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 (47 in Liaoning, 52 in Guangdong, 55 in Sichuan). The gains are lower for agricultural households and higher for house- holds with more employees and more township and village enterprise workers, with more migrant workers, and with less cultivated land (though significant only in Liaoning). The only strong demographic effect is that younger households Chen and Ravallion 49 TABLE 5. Regression Results for Percentage Gains in Rural Areas of Three Provinces, 2001-07 Variable Liaoning Guangdong Sichuan 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) No. employees/household size 0.630 (2.76) 0.271 (2.57) 0.444 (3.61) No. township and village enterprise 0.669 (4.27) 0.585 (4.47) 0.548 (6.11) workers/household size No. migrant workers/household size 0.655 (3.59) 0.187 (3.59) 0.346 (7.08) Area 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 level Illiterate 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 6/household size 3.730 (3.61) 0.461 (1.49) -0.169 (-0.78) Children 6-11/household size 1.557 (1.41) 0.173 (0.72) -0.275 (-1.48) Children 12-14/household size 1.625 (1.54) -0.477 (-1.60) -0.343 (-1.85) Children 15-17/household size 1.325 (1.80) -0.289 (-0.91) -0.192 (-0.88) Constant 0.788 (0.69) -0.709 (-1.39) -0.584 (-1.68) R2 0.108 0.217 0.171 Note: Numbers in parentheses are t-statistics. Source: Authors' computations based on China nbs 1999 Rural Household Survey and 1999 Urban Household Survey. (with a higher proportion of children under age six) tend to be gainers in Liaoning. Although the results for the county dummy variables are not shown (to save space), losses were significantly higher than average in six counties in Liaoning, seven in Guangdong, and six in Sichuan. table 8 gives the mean losses in these counties for agricultural households. In urban areas the gains tend to be higher for smaller households (except in Guangdong). As in rural areas there is a U-shaped pattern (except for Liaoning), with lowest gains at 66 years of age in Guangdong and 51 in Sichuan. Although there is no pattern in the relationship between education and welfare gains in rural areas, the gains in urban areas tend to be larger for less well-educated households. However, this may be biased by the fact that education was used in identifying skilled labor (noting that unskilled nonfarm wages are predicted to increase relative to skilled labor; see table 2). There are some signs of sectoral effects, though only significantly so in Liaoning, with higher gains for those with government jobs. Retirees tend to have lower gains than others. 50 THE WORLD BANK ECONOMIC REVIEW, VOL. i 8, NO. I TABLE 6. Regression Results for Level of Gain (Yuan) in Urban Areas of Three Provinces, 2001-07 Variable Liaoning Guangdong Sichuan 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 water -27.422 (-2.62) 20.089 (0.92) -16.585 (-0.83) conservancy Transportation and 6.368 (1.52) 16.525 (1.24) 1.644 (0.25) telecommunications Wholesale 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 and 21.469 (3.59) 9.637 (0.34) 16.117 (2.35) telecommunication Other 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) R2 0.265 0.131 0.181 Note: Numbers in parentheses are t-statistics. Source: Authors' computations based on China nbs 1999 Rural Household Survey and 1999 Urban Household Survey. Chen and Ravallion 51 TABLE 7. Regression Results for Percentage Gains in Urban Areas of Three Provinces, 2110-07 Variable Liaoning Guangdong Sichuan 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 college) 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 and water -0.407 (-3.06) 0.178 (1.03) -0.228 (-0.53) conservancy Transport and telecommunications 0.206 (2.93) 0.060 (0.79) -0.036 (-0.4) Wholesale 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 telecommunications 0.209 (2.02) -0.018 (-0.09) 0.130 (1.03) Other 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) R2 0.401 0.290 0.359 Note: Numbers in the parentheses are t-statistics. Source: Authors' computations based on China nbs 1999 Rural Household Survey and 1999 Urban Household Survey. 52 THF WORLD BANK ECONOMIC REVIEW', VOL. i8, NO. I TABLE 8. Average Impacts for Agriculture Households in Selected Counties, 200 1-07 Gain Provincial mean Province County identifier Yuan % Yuan % Liaoning -32.34 -1.29 210181 -73.72 -3.07 210212 -145.40 -2.99 210381 - [72.01 -5.57 210921 -57.70 -5.21 211321 -45.58 -3.78 211322 -53.60 -3.23 Guangdong -29.34 -0.81 440111 -107.31 -2.74 440126 -183.63 -2.64 440223 -102.33 -3.53 440523 -148.90 -2.55 440620 -227.23 -3.11 440621 -109.59 -2.64 441425 -316.49 -5.34 Sichuan -12.31 -0.67 510121 -130.46 -2.86 510125 -63.19 -3.81 512425 -138.34 -5.71 512610 -52.23 -3.11 512825 -40.44 -2.80 513021 -93.02 -4.07 Note: A negative sign means a net loss. Agriculture household means that more than 75 percent of income is from agriculture. Source: Authors' computations hased on China nbs 1999 Rural Household Survey and 1999 Urban Hotusehold Survey. V. CONCIUSIONS In the aggregate, the analysis finds that China's trade reforms have had only a small impact on mean household income, inequality, and poverty incidence. There is, however, a sizable (and at least partly explicable) variance in impacts across household characteristics. Rural families tend to lose; urban households tend to gain. There are larger impacts in some provinces than in others, with the highest impacts in the northeast region of Inner Mongolia, Liaoning, Jilin, and Heilongjiang, where rural households are more dependent on feed grain pro- duction (for which falling prices are expected from WTO accession) than else- where in China. Within rural or urban areas of a given province, the gains from trade reform vary with observable household characteristics. The most vulnerable households tend to be in rural areas, dependent on agriculture, with relatively fewer workers Chen and Rat'allion 53 and with weak economic links to the outside economy through migration. There are also some strong geographic concentrations of adverse impacts. For example, agricultural households in some counties incur welfare losses of 3-5 percent of their incomes. Naturally, the approach taken here has limitations. For example, there may well be dynamic gains from greater trade openness that are not captured by the model used to generate the relative price impacts. Trade may facilitate learning about new technologies and innovation, bringing longer-term gains in produc- tivity. Trade reform may also come with (and possibly help induce) other policy reforms, such as in factor markets. The approach here has attempted to capture only the static welfare effects of wro accession. A further limitation was the need to make linear approximations in the neighborhood of an initial optimum for each household. In other applications this could be deceptive if price or wage changes are large or if the household was initially out of equilibrium, due to rationing (including involuntary unemploy- ment), for example. In principle, there are ways of dealing with these problems by estimating complete demand and supply systems that allow for rationing. This may prove a fruitful avenue for future research, though it should be noted that these methods generate their own problems, such as those arising from incomplete data on price and wage levels at household level. Despite these limitations, the type of approach followed here can usefully illuminate the range of welfare impacts to be expected from economywide reforms. By avoiding unnecessary aggregation of the primary household-level data, these relatively simple tools can also offer insights into the sorts of policy responses that might be called for to compensate losers from reform. APPENDIX TABLE A-1. Rural Gains and Losses by Province, 2001-07 Number Original Post-wTo sampled Number income income Gain or loss Change Share of losers Province households gainers (yuan) (yuan) (yuan) (%) (%) Beijing 750 381 4,221.05 4,210.08 --10.96 -0.26 49.20 Tianjin 595 219 3,401.71 3,380.48 -21.22 0.62 63.19 Hebei 4,200 1,310 2,441.50 2,426.82 -14.68 -0.60 68.81 Shanxi 2,100 926 1,772.62 1,765.13 -7.49 0.42 55.90 Inner Mongolia 2,198 206 2,055.49 2,011.26 -44.22 -2.15 90.63 Liaoning 1,886 353 2,501.98 2,469.64 -32.34 1.29 81.28 Jilin 1,598 132 2,260.12 2,210.46 -49.66 -2.20 91.74 Heilongjiang 1,997 115 2,166.59 2,114.18 -52.41 2.42 94.24 Shanghai 600 416 5,409.11 5,428.79 19.68 0.36 30.67 Jiangsu 3,400 1,209 3,495.20 3,486.78 -8.42 0.24 64.44 Zhejiang 2,693 1,148 3,946.44 3,934.92 -11.52 -0.29 57.37 Anhui 3,095 676 1,900.76 1,885.79 -14.97 -0.79 78.16 Fujian 1,750 469 3,091.39 3,071.40 -19.99 -0.65 73.20 Jiangxi 2,450 553 2,129.45 2,117.26 -12.19 -0.57 77.43 Shandong 4,200 822 2,520.76 2,494.89 -25.87 -1.03 80.43 Henan 4,200 828 1,948.36 1,931.70 -16.66 -0.86 80.29 Hubei 3,188 755 2,212.71 2,200.04 - 12.68 -0.57 76.32 Hunan 3,700 1,181 2,102.98 2,095.39 -7.60 -0.36 68.08 Guangdong 2,560 514 3,628.95 3,599.61 -29.34 -0.81 79.92 Guangxi 2,310 309 2,048.33 2,025.75 -22.58 -1.10 86.62 Hainan 718 28 2,086.40 2,057.85 -28.55 -1.37 96.10 Chongqing 1,500 404 1,736.63 1,730.20 -6.43 -0.37 73.07 Sichuan 3,998 879 1,843.23 1,830.92 -12.31 -0.67 78.01 Guizhou 2,240 417 1,363.07 1,354.03 -9.04 -0.66 81.38 Yunnan 2,397 399 1,438.34 1,421.34 -17.00 -1.18 83.35 Tibet 480 143 1,309.46 1,307.41 -2.05 -0.16 70.21 Shaanxi 2,217 446 1,456.48 1,442.09 -14.39 -0.99 79.88 Gansu 1,800 479 1,357.28 1,350.34 -6.95 -0.51 73.39 Qinghai 600 135 1,466.67 1,452.61 -14.06 -0.96 77.50 Ningxia 600 108 1,754.15 1,729.05 -25.11 -1.43 82.00 Xinjiang 1,495 312 1,471.11 1,447.57 -23.55 -1.60 79.13 Rural total 67,515 16,272 2,257.15 2,239.08 -18.07 -0.80 75.90 Note: The ordering of provinces is the traditional administrative ordering as used (for example) in Chiina Statistical Yearbook (NBs 2000). Source: Authors' computations based on lanchovichina and Martini (2002) and China NBS 1999 Rural Household Survey and 1999 Urban Household Survey. TABLE A-2. Urban Gains and Losses by Province, 2001-07 Number Original Post-wro sampled Number income income Gain or loss Change Share of losers Province households gainers (yuan) (yuan) (yuan) (%) (%) Beijing 500 430 9,388.88 9,431.72 42.84 0.46 14.00 Tianjin 500 451 7,323.57 7,358.47 34.91 0.48 9.80 Hebei 650 591 5,673.46 5,702.35 28.89 0.51 9.08 Shanxi 650 598 4,519.20 4,549.94 30.74 0.68 8.00 Inner Mongolia 550 495 4,491.87 4,516.19 24.32 0.54 10.00 Liaoning 1000 916 5,257.42 5,285.65 28.23 0.54 8.40 Jilin 700 610 4,630.13 4,650.46 20.33 0.44 12.86 Heilongjiang 1000 887 4,798.92 4,820.50 21.58 0.45 11.30 Shanghai 500 458 10,927.18 10,984.16 56.98 0.52 8.40 Jiangsu 800 723 6,933.07 6,968.78 35.71 0.51 9.63 Zhejiang 550 498 9,044.40 9,098.28 53.87 0.60 9.45 Anhui 500 458 5,159.46 5,190.37 30.91 0.60 8.40 Fujian 550 516 7,521.52 7,569.70 48.18 0.64 6.18 Jiangxi 550 498 4,762.78 4,783.38 20.60 0.43 9.45 Shandong 650 602 5,689.90 5,720.69 30.78 0.54 7.38 Henan 600 565 4,689.43 4,717.89 28.46 0.61 5.83 Hubei 750 619 5,743.18 5,765.29 22.11 0.38 17.47 Hunan 700 612 5,727.42 5,750.43 23.00 0.40 12.57 Guangdong 600 490 10,871.06 10,903.85 32.79 0.30 18.33 Guangxi 600 496 6,011.10 6,033.40 22.30 0.37 17.33 Hainan 200 172 5,766.33 5,787.64 21.31 0.37 14.00 Chongqing 300 239 5,910.18 5,931.90 21.72 0.37 20.33 Sichuan 800 691 5,610.29 5,634.60 24.30 0.43 13.63 Guizhou 450 383 5,324.43 5,347.71 23.27 0.44 14.89 Yunnan 650 566 5,939.69 5,973.23 33.54 0.56 12.92 Tibet n.a. Shaanxi 500 427 4,768.99 4,788.25 19.26 0.40 14.60 Gansu 400 372 4,610.86 4,641.27 30.41 0.66 7.00 Qinghai 250 240 3,759.53 3,788.65 29.12 0.77 4.00 Ningxia 200 177 4,472.43 4,493.27 20.84 0.47 11.50 Xinjiang 250 214 5,277.25 5,295.94 18.69 0.35 14.40 Urban total 16,900 14,994 6,046.13 6,075.60 29.45 0.49 11.28 Note: The ordering of provinces is the traditional administrative ordering as used (for example) in China Statistical Yearbook (NBS 2000). Source: Authors' computations based on lanchovichina and Martin (2002) and China NBS 1999 Rural Household Survey and 1999 Urban Household Survey. Cbeo aind Ravallion .57 REFERENCES Barro, Robert. 2000. "Inequality and Growth in a Panel of Countries." Jouirnal of Economizic Growth 5:5-32. Baulch, Bob, and John Hoddinott. 2000. "Economic Mobility and Poverty Dynamics in Developing Countries." Journal of Development Studies 36(6):1-24. Bourguignon, Francois, and C. Mvorisson. 1990. "Income Distribution, Development and Foreign Trade." European Economic ReviewZ 34:1113-32. Bourguignon, Francois, Anne-Sophie Robilliard, and Sherman Robinson. 2003. "Representative versus Real Households in the Macro-Economic Modeling of Inequality." Working Paper 2003-05. DETA, Paris. Chen, Shaohua, and Martin Ravallion. 1996. "Data in Transition: Assessing Rural Living Standards in Southern China." China Economic Review 7:23-56. .2001. "How Did the World's Poor Fare in the 1990s?" Rcvieu' of Income and Wealth 47(3):283- 300. - 2003. "Household Welfare Impacts of China's Accession to the World Trade Organization." Policy Research Working Paper 3040. World Bank, Washington, D.C. Cockburn, John. 2002. "Trade Liberalization and Poverty in Nepal: A Computable General Equilibrium Micro Simulation Analysis." Working paper, University of Laval. Deaton, Angus. 1989. "Rice Prices and Income Distribution in Thailand: A Non-Parametric Analysis." Economic Journal 99:1-37. Decaluwe, B., and A. Martens. 1988. "(G. .Modeling and Developing Economies: A Concise Empirical Survey of 73 Applications to 26 Countries." Journzal of Policy Modeling 10(4):529-68. Dollar, David, and Aart Kraay. 2002. "Growth Is Good for the Poor." Journal of Econ(omic Growth 7(3):195-225. Edwards, Sebastian. 1997. "Trade Policy, Growth and Income Distribution." American Economnic Review 87(2):205-10. Friedman, Jed, and James Levinsohn. 2002. "The Distributional Impacts of Indonesia's Financial Crisis on1 Household Welfare: A 'Rapid Response' Methodology." World Bank Ecoonoic Revieiw 16(3):397-424. Hertel, T., ed. 1997. Global Trade Analvsis: Modeling and Applications. Cambridge: Cambridge Uni- versity Press. lanchovichina, Elena, and Will Martin. 2002. "Economic Impacts of China's Accession to the A'l'o. Working paper, Development Research Group, World Bank. Jalan, Jyotsna, and Martin Ravallion. 1998. "Transient Poverty in Post-Reform Rural China." lournal of Comparative Economnics 26:338-57. 2002. "Geographic Poverty Traps? A Micro Model of Consumption Growth in Rural China." Journal of Applied Econcometrics 17(4):329-46. King, NMervyn A. 1983. "Welfare Analysis of Tax Reforms Using Household Level Data." Journal of Public Economiics 21:183-214. Lundberg, Mattias, and Lyn Squire. 2003. "The Simultaneous Evolution of Growth and Inequality." Economic Journal 113:326-44. NBS (China National Bureau of Statistics). 2000. Cbhia Statistical Yearhook. Beijing. Newbery, David, and Nicholas Stern, eds. 1987. Tlhe Tbeory of Taxation for Developing Coluntries. Oxford: Oxford University Press. Ravallion, Martin. 1990. "Rural Welfare Effects of Food Price Changes with Induced Wage Responises: Theory and Evidence for Bangladesh." Oxford Economic Papers 42:574-85. 2001. "Growth, Inequalitv and Poverty: Looking Beyond Averages." W'orlcd Developmient 29(11):1803-15. Ravallion, Martin, and Dominique van de Walle. 1991 . "*The lmpact of Food Pricing Reforms on Povcrtv: A Demand Consistent Welfare Analysis for Indonesia." Joumral of Policyv Modeling 13:281-300. Tracking Distortions in Agriculture: China and Its Accession to the World Trade Organization Jikun Huang, Scott Rozelle, and Min Chang This article examines the impacts of China's accession to the World Trade Organiza- tion (WTO) on prices in its agricultural sector. The analysis uses a new methodology to estimate nominal protection rates in China's agricultural sector before its accession to the wro. These new measures account for differences in commodity quality within China and between China and world markets. The analysis shows that some of China's agricultural commodities are well above world market prices and others are well below. The article also assesses market integration and efficiency in China. It finds high degrees of integration between coastal and inland markets and between regional and village markets. The remarkable improvements in market performance in recent years mean that if increased imports or exports affect China's domestic price near the border, producers throughout most of China will feel the price shifts. Trade liberalization affects rural populations in a number of offsetting ways (OECD 2001). On one hand, increases in the demand for a nation's industrial goods through higher exports can increase the employment and wages of workers in rural areas. Farmers benefit from new opportunities to export agricultural goods and from better access to more affordable inputs. Rural consumers gain from access to cheaper food. On the other hand, rising imports of lower-priced com- modities reduce farm profits, and improved access to export markets raises prices to domestic consumers and to producers that use agricultural goods as inputs. Although all the effects are important, trade officials concerned about the profitability of domestic producers are frequently interested in the impact of trade liberalization on agricultural prices (Martin 2001). Knowing what to expect is particularly important for countries with many small farmers that Jikun Huang is with the Chinese Academy of Sciences; his e-mail address is jikhuang@public.bta.ner.cn. Scott Rozelle and Min Chang are with the University of California, Davis; Rozelle's e-mail address is rozelle@primal.ucdavis.edu, and Chang's e-mail address is chang@primal.ucdavis.edu. The authors thank Yuping Xie for research assistance; this work could not have heen done without him. They acknowledge the financial support of the World Bank's Trade and Rural Development units and helpful comments and suggestions by Kym Anderson, Fred Crook, Tom Hertel, Elena lanchovichina, Will Martin, Francis Tuan, and participants in the World Bank-sponsored conference on the Impact of wro Accession on China. THE WORLD BANK ECONOMIC REVIEW, VOL 18, NO. 1. © The International Bank for Reconstruction and Development / THE WORLD BANK 2004; all rights reserved. DOI: 10.1093/wber/lhhO33 18:59-84 59 60 THE WORLD BANK ECONOMIC REVIEW, VOL. i8, NO. I produce not only for their household but also for commercial markets.' Government officials know that agricultural price shifts can have important effects on domestic food production, farm household incomes, national poverty rates, and overall rural stability. Thus, in determining positions for trade negotia- tions, officials must solve a complicated political economy equation. But there is often confusion about how trade liberalization will affect produ- cer prices. First, there may be confusion about the level of protection (or implicit taxation) for individual commodities. Studies sometimes reach contradictory conclusions, with some claiming that a commodity is being protected and others that the commodity is receiving negative protection.2 One reason for the dis- crepancies is shortcomings in traditional methods of measuring distortions. Studies often assume that price differences between domestic and foreign pro- ducts equal the tariff rate or calculate differences between an average domestic price, taken from surveys, and an average border price, taken from trade statistics. Such methods may not capture the real protection rate because of aggregation issues and because (if some prices are from interior areas) transport costs are confounded with protection. Also, most analyses assume that trade liberalization affects all prices in the same way, but some internal markets may not function well. In short, the confusion in the literature about the real level of trade protection may lie in the way researchers have measured it. Second, the source of protection is sometimes unclear (Garcia 2003). Trade negotiations have tarrified most traditional quotas and reduced average tariff rates. In this environment the main barriers to the flow of many agricultural commodities are nontraditional ones, such as domestic tax policies, export subsidies, and tariff rate quotas (instead of traditional tariffs and quotas). Finally, studies on the effects of trade liberalization can arrive at different answers if assumptions about domestic market integration are not considered. There is often disagreement about the effect of trade liberalization on subsets of producers that produce different commodities, belong to different income groups, or live in certain geographical regions. For example, before the North American Free Trade Agreement (NAFTA) there was a concern that reducing Mexico's tariff on 1. There is also demand for information about the effect of liberalization on prices in more developed economies, where producers exercise considerable political influence. Although the discussion in the article is cast in terms of producer prices, similar arguments can be made for a nation with many poor, landless rural residents. In that case, however, officials are also interested in the effect of liberalization on agricultural prices, but their concerns would focus on how such effects would alter the cost of the average household's consumption bundle. In countries with large populations of smallholders (such as China), the concern of officials with agricultural prices is on maintaining producer prices to keep the income of poor farmers from falling. 2. In China, for example, some researchers argue that the agricultural impact of China's World Trade Organization accession will be substantial, with sharply lower prices adversely affecting hundreds of millions of farmers (Carter and Estrin 2001; Li and others 1999). Others believe that the overall effect on agricultural prices will be modest although there will be substantial impact on prices in some specific areas and for some specific commodities (Anderson and Peng 1998; Anderson and others forthcoming). Huianzg, f. I.. and Cha72g 61 maize imports would hurt maize producers, especially those who lived in poor, remote regions (see Taylor 1998 for details of the debate). Some researchers claimed that by lowering maize prices, liberalization would bankrupt these farmers and force them into the migrant labor force, with many of them eventually entering the United States illegally. Others believed that fragmented markets isolated poor, small farmers in many regions of the nation from the direct effects of the NAFTA- induced downward pressure on prices. Today, more than a decade after NAFTA, research shows the importance of market characteristics in determining the effect of trade liberalization on farm producer prices. When commodity markets do not operate well and there is poor integration, the effects of trade liberalization on producer prices in isolated areas are greatly attenuated (Taylor 1998).3 To improve understanding of how trade liberalizarion will affect agricultural prices and how price changes will be experienced in different parts of country, this article describes ways to create more accurate, disaggregated measures of protection (nominal protection rates) that can be used in two ways. They can be used to analyze the expected effects of liberalization and identify remaining trade barriers by matching up different sources of protection to observed levels of protection. After liberalization, they can be used to assess the effectiveness of policy implementation. Price determination and market integration analyses are then used to study domestic markets to assess how price shifts at the border arising from trade liberalization affect different producers in different parts of the country. The main contribution of this study is the way it combines a series of analytical exercises to improve understanding of how trade liberalization will affect the level of agricultural prices and the distribution of their effects. The impact on agricultural prices of China's accession to the World Trade Organization (\'TO) was chosen for study because of the intense interest by officials and academics in how China's WTO agreements on agriculture would affect the prices received by farmers (RCRE 2000; Huang and Chen 1999). China was also selected because of the lack of agreement in other studies about how liberalization will affect farmers. Finally, because many of China's poor rural households live in remote regions far from the coast and rely more than other groups on income from cropping, if China's markets were integrated, poor households are the most likely to be affected by liberalizing measures (Chen and Ravallion forthcoming). This case study focuses only on the effects of WTO accession on agricultural prices even though other effects of accession on the rural population will likely 3. In his case study, Taylor ( 1998) finds that the impacts of NAFTA on Mexican farmers in border regions and those in more remote regions, who face higher tranisaction costs for marketing their output and buying inputs, differ dramatically. He finids that NAFI.\ has little impact on those in the poorest areas mainly because they are insulated hy high transaction costs. Because economic activities in remote areas are mostly within the household, village, or township, the prices of goods are determined locallv and not affected hy what happens far away in the nation's border areas. That is not to say that trade liberalization policies do not affect welfare in these areas. But the complicated wa's farmers in these economies respond to changes in prices and marketing opportunities usually meani that the effects are much smaller than they would be on households that live and work in completely commercialized economies. 62 THE WORLD BANK ECONOMIC REVIEW, VOL. TS, NO. I be at least as important (Zhao and Sicular 2002). The analysis does not quantify the total welfare effect, but rather considers the qualitative effects on China's farmers to illustrate one way of conceptualizing the effects of trade liberal- ization on agricultural prices. The article reviews China's trade policy liberalization before and after WTO accession, looking at traditional reforms, such as tariffication and tariff reduction, and nontraditional reforms, such as taxation policy, export subsidies, and tariff rate quotas. It then describes how the new measures of nominal protection rates were created and examines how these distortions might change with WTO acces- sion. Finally, it analyzes the nature of China's agricultural markets to see how the price effects of trade liberalization might affect different types of farm households. I. TRADE LIBERALIZATION AND REMAINING DISTORTIONS IN AGRICULTURE Partly because of the vulnerability of parts of the rural economy and partly because of the prominence in China's political economy, agriculture has been at the center of discussions of China's entry into the WTO. Yet the likely shifts in China's trade policy and their impacts on agricultural prices are not well under- stood. Debates on the future of China's agriculture and the price level in the sector remain unresolved. Some argue that the agricultural impact of WTO accession will be substantial (Carter and Estrin 2001; Li and others 1999); others disagree (Anderson and Peng 1998; Huang and Anderson 2003). Traditional Souirces of Protection Some of this divergence can be traced to a lack of understanding of the policy changes that may be induced by China's WTO accession (Martin 2001). Tradition- ally, analysts have focused on the measures most frequently used by other countries to protect their agricultural sectors. Most previous work (for example, CARD 2001; Tuan and Cheng 1999; and OECD 2001) focuses on tariffs, quotas and licensing, state trading, and traditional nontariff barriers. Some of these studies implicitly assume that WTO agreements are concerned solely with these policies, that these policies provided most of the protection China enjoyed before accession, and that accession represents China's initial assault on protection at the border. In fact, after nearly two decades of reform, some of the worst distortions caused by traditional policies have already disappeared. In the late 1970s and early 1980s, the domestic wholesale price of China's four major commodities (rice, wheat, maize, and soybeans), converted at the official exchange rate, far exceeded the world price, measured at China's border. China's main food and feed grain and soybean prices, for example, were 10-92 percent above world market prices (table 1). Over the next 15 years the nominal protection rate became negative for rice and fell to around 15-30 percent for wheat and maize. Intervention by state traders and the use of nontariff barriers also gradually fell (Martin 2001). Huang, Rozelle, and Chang 63 TABLE 1. Changes in Nominal Rates of Protection for China's Major Agricultural Commodities, 1978-2000 (%) Period Rice Wheat Maize Soybeans 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 24 1998-00 -6 26 32 28 1998 -6 22 40 34 1999 -9 30 33 27 2000 -5 15 33 20 Note: Nominal rates of protection are measured as the difference (in percentage terms) between average border prices and average domestic wholesale (market) prices. Source: Huang (2001). Falling protection and changes in international trade and domestic marketing policies have resulted in dramatically shifting price trends and trade patterns. Depreciation of China's currency explains a big part of China's changing protection during the 1980s and 1990s (Huang and Chen 1999). Huang and others (2002) also trace the changes in prices following implementation of trade liberalization policies. Between 1985 and 2000 the real price of agricultural commodities (measured by the agricultural price index divided by the rural consumer price index) fell 27 percent (State Price Bureau various years). These policy reforms led to a large decline in price distortions over the past 20 years. Current policy reforms accompanying China's accession to the WTO should be considered an extension of these past efforts. Much of the falling protection has come from relaxing licensing procedures, reducing the scope of nontariff barriers, reducing tariffs, and tariffying quotas (Huang and Chen 1999). This likely explains why so much research on China's entry into the w«o focuses on these traditional policies. And, as is argued here, while nontraditional policies may be even more important in assessing the effect of trade liberalization on agricultural prices, changes in China's tariff regimes, state trading system, and nontariff barriers undoubtedly will remain at least somewhat of an influence on price distortions in China's agriculture. Nontraditional Sources of Protection With many of the gains from traditional trade reforms already achieved, China may need other, less discussed policies to push forward additional trade liberal- ization. For example, China has used tax policy to protect agriculture, especially for commodities such as soybeans and barley that have been most liberalized in terms of traditional forms of protection. In the early 1990s, leaders radically revised China's fiscal system, making revenue generation more reliant on a value added tax (Nyberg and Rozelle 1999). A 13-17 percent tax is assessed on value added for all goods through all stages of their manufacture and sale. 64 THE WORLD BANK ECONONMIC REVIEW, VOL. I 8, NO. I Because many other countries do not have a value-added tax, national regula- tions state (and the wro allows) that the tax should also apply to imported goods that are not for immediate reexport. For a variety of political and tax collection reasons, however, farmers were initially exempt from the tax when they sold their products to traders from their farms or in local markets. When the good is resold in a downstream wholesale market, the trader owes the tax only on the amount of the marketing margin, or the difference between the procurement price and sales price. With marketing margins in China's competitive grain markets at about 5 percent (ranging between 1 and 10 percent, according to Xie 2002), the real value-added tax rate on domestic agricultural goods is only 5 percent of that of the tax rate on imported goods (or about 1 percent of the value of the domestic good, compared with 13 percent of the value of the imported good). When assessed at the border but not at the farm gate, the value-added tax provides producers with rates of protection greater than the official tariff rate. For example, the published tariff rate on soybeans is 3 percent. Theoretically, then, soybean imports when they arrive at China's borders should cost only 3 percent more than China's domestic soybeans. However, when soybeans cross the border, importers must also pay a 13 percent value-added tax. Domestic soybeans, in contrast, are taxed at less than 1 percent on average. As a con- sequence, the use of a value-added tax at the border gives China's soybean producers more than 10 percent of additional price protection.4 China also aggressively used export subsidies in the years leading up to its accession to the wTo to increase exports of some commodities, thereby in- creasing protection by raising the price of certain domestic commodities (table 2; Rozelle 2003). Maize and cotton have received the largest export subsidies. Interviews in the field during 2001 revealed that maize exporters, especially in northeast China, received subsidies averaging 34 percent of the export price and cotton exporters received subsidies averaging 10 percent of the export price. Although there are no subsidies for meat exports (that is a more difficult transaction because there are many meat exporters and most of them are private or commercialized public firms, unlike maize and cotton traders that are mostly associated with formal, public state trading firms), tax policies also favor exporters of many livestock products. For example, when meat producers execute an export contract, the company can receive a tax rebate. In 2001 pork and beef exporters received a rebate equal to 5.2 percent of the value of their transaction and poultry exporters received a 13 percent rebate. Because domestic wholesalers are not eligible for the rebates, such policies encourage traders to export. 4. Some scholars in China have also pointed out that because part of the value of agricultural commodity production derives from inputs on which the value-added tax has been assessed, the real tax rate on agricultural commodities is actually higher. Although this is so, the most that could be added would be 2-4 percentage points (15 percent times the share of the inputs that were taxed-about 10-30 percent-depending on the commodity, the technology, and the region of production). Huang, Rozelle, and Chang 65 TABLE 2. Subsidies and Tax Rebates for Exports of Selected Agricultural Commodities in China, 2001 (%) Rebate of value-added Commodity Export subsidy tax for exports Rice <1 0 Cotton 10 0 Maize 34 0 Pork 0 5.2 Beef 0 5.2 Chicken 0 13.0 Source: Authors' survey. In summary, then, as China enters the wro there are still a number of challenges in liberalizing its trade. In addition to using traditional trade poli- cies-tariffs, quotas, state trading, and nontariff barriers-to manage agricul- tural prices in the domestic economy, China has protected or may decide to protect agriculture with a number of other policy measures, such as taxation policy, export subsidies, and rebates. China may try to use such policies to protect or further open its agricultural sector. II. NEW ESTIMATES OF CHINA'S NOMINAL PROTECTION RATES IN AGRICULTURE This section illustrates how to estimate nominal protection rates that avoid some of the common problems of past estimates and that permit more accurate assessment of the impact of China's implementation of its WTO obligations. They show, in a more disaggregated way, the level at which China is protecting agricultural commodities or parts of certain markets. Then, by aggregating these nominal protection rates into single crop-specific rates, these estimates allow assessment of how these methods can be compared with traditional methods of estimating nominal protection rates. (The appendix summarizes some of the difficulties of trying to estimate nominal protection rates for China's agriculture using traditional methods.) The new nominal protection rates depend on the collection of a new type of data. Interviews and surveys were used to gather information on prices of agricultural commodities to identify price gaps between an imported good on one side of the border (outside China) and a domestic good on the other side (inside China) and between exportable domestic goods as they leave the country and the same goods from other countries that are being traded in international markets. Between August and November 2001, the survey team visited seven coastal cities (Dalian, Guangzhou, Lianyungang, Ningbo, Qinghuangdao, Shanghai, and Shenzhen) and two inland cities (Beijing and 66 THE WORLD BANK ECONOMIC REVIEW, VOL. I8, NO. I Changchun).5 Information was collected from samples of domestic traders, importers and exporters, wholesalers, grain and oilseed users, trade regulators, agents, and other grain and fiber officials. More than 100 people were inter- viewed.6 Less than 10 percent of those contacted refused to be interviewed. The survey was particularly concerned with understanding price gaps between the international and domestic markets of commodities in which interviewees were trading or otherwise participating. The survey recorded char- acteristics (qualities, grades, varieties) of commodities being traded in the immediate marketing area. For imported commodities interviewees were asked about the current international cost, insurance, and freight (cif) price of the good (for a ship docked in their home port) and then about what the good would sell for in a competitive auction. This yielded a series of price gaps for a carefully defined set of goods. Because each interviewee had information on a number of commodities, this process yielded several thousand observations. A similar set of questions was asked about exportable goods, including rice, fruits and vegetables, and meat products. For exported goods that were being sub- sidized, interviewees were asked how much they would lose if they sold a shipment onto the international market without any financial assistance from the government. Disaggregated Nominal Protection Rates for Selected Agricultural Commodities The analysis here illuminates the problems with traditional nominal protection rate estimates of a single rate of protection for a commodity using the typical types of secondary data that are available for most countries. For example, it would be difficult to provide just one nominal protection rate for wheat in China, one of the world's largest wheat importers over the past two decades (table 3). Traders reported that the price of very high-quality wheat from North America was 20-50 percent higher in the domestic markets of China's major ports than when it was sitting on a ship in China's port ready to be brought into 5. Although Beijing and Changchun are inland cities, firms from the two areas are still actively engaged in international and domestic trade. The prices that they quote used the same basis as those quoted by firms from the coast. 6. Because of the absence of a single central authority that manages grain flows, the enumeration team chose their sample by first visiting the local grain bureau in each area to obtain a list of the firms that they were running on a commercial basis and their subsidiaries. Officials in the grain marketing division and transportation division were interviewed. Three firms owned by the grain bureau and three affiliated with the grain bureau were selected. In several cities, the grain bureau had a list of large grain- trading and grain-using firms (such as mills and feed lots). In others, lists were obtained from the market administration bureau. Five firms were chosen on the basis that they were private and had yearly sales of more than 1 million yuan. Representatives of at least two flour or rice mills and feed mills in each location were interviewed. Five stalls at the wholesale market were randomly chosen for interviews. Questions about the grain trade were asked at a number of other entities, such as the grain reserve, the local COFCO agency, and supermarket chains. Huiang, Rozelle, and Chang 67 TABLE 3. Disaggregated Nominal Protection Rates for Selected Agricultural Commodities in China, October 2001 Comparable Border price domestic price per ton per ton (US$) Commodity and Nominal protection variety Yuan US$ cif fob rate (D) Rice Weighted average - - - - Thai jasmine rice 3,690 446 380 - 17 High-quality japonica 2,930 354 - 398 -11 Medium-quality indica 1,519 184 - 185 -0.5 Wheat Weighted average - - - - 12 U.S. DNS 2,350 284 190 - 49 Canadian number 3 1,800 218 181 - 20 Australian soft 1,625 196 175 - 12 U.S. hard red 1,550 187 169 - 11 U.K. 1,350 163 145 - 12 China high quality 1,350 163 145 - 12 China medium quality 1,250 151 140 - 8 China low quality 1,100 133 133 - -0.1 Sovbeans Common varietv 1,950 236 205 - 15 Maize Common variety 1,150 139 - 105 32 Note: Estimated at the official exchange rate of 8.28 yuan to the U.S. dollar.-, not available. Soutrce: Authors' survey. the country. More precisely, if a ton of imported Canadian number 3 hard white wheat were costlessly brought across the border and auctioned off in China's domestic market in October 2001, the competitive bid price would have been 20.5 percent higher on average than the international price on a cif basis. This price gap would imply that China's protection rate is high, and that if China were to open its markets completely domestic wheat prices would fall and the import volume would rise. Interviewees were quick to point out, however, that they did not think that even with open markets China's overall wheat price would fall anywhere near 50 percent (even if there were no effect on the world price-they were not con- sidering the impact of China's imports on the world price). They noted that the market for baking-quality wheat, the main use for hard white wheat from North America, is small in China, at most only several million metric tons. Few wheat users in China outside those who demanded flour for making cakes, pastries, and high-quality breads would use this type of wheat even if were available at a cheaper price. On the supply side only a small group of farmers and processors inside China were able to competitively produce and market this type of wheat. 68 THE WORLD BANK ECONONMIC REVIEW, VOL. 18, NO. I If these supply and demand dynamics are accurate, this would mean that even in a world free of trade restrictions, imports of hard white wheat would con- tinue only until demand for that variety was met and the domestic price fell to international levels. Most of the production of that variety of wheat would shift outside of China, but because of the limited demand for such wheat varieties, only the small number of domestic farmers who had been producing these varieties at the trade barrier-protected price would have to abandon their production following liberalization. Moreover, in this specific case-one with few domestic suppliers and little or no substitution of baking-quality wheat for other domestic uses-there should be only a small price impact on other domestic wheat producers. In short, growers of the high-quality wheat would lose; they would either have to keep growing at a lower price or switch to another wheat variety or some other crop. The overall price impact would be minimal, however, because such specialized wheat varieties fill such a special niche and the quantities involved are small. Although not as extreme as the case for North American baking-quality wheat, traders reported arbitrage possibilities in other wheat markets (see table 3). With a remarkable degree of consistency, the cif price of medium- quality wheat imports from Australia, the United Kingdom, and the United States (hard red) was reported to be 10 percent lower than the price that interviewees believed the same wheat would command in China's domestic market. Used for more common breads, cheaper pastries, and high-quality noodles, this wheat accounted for an estimated 10-15 percent of China's wheat demand, according to interviewees. However, unlike the case for the highest-quality baking wheat, there was more production in China. During 2001 domestic producers supplied most of this quality of wheat to China's wheat market, enjoying protection provided by state trading or the value- added tax policy (a 13 percent tax is enough to keep these varieties from being competitive inside China's domestic market).7 Finally, China's medium-quality wheat, by far the biggest part of China's production (estimated to be more than 60 percent) appears at most to be only marginally protected (see table 3). Interviewees believed if China's medium- quality wheat were sold on the international market in late 2001, it would sell at a discount of about 8 percent. In other words, if international traders could ship this quality of wheat to China, which they have not done to date, it would command a premium of 8 percent. Because this wheat constitutes the largest 7. In China's domestic market, medium-quality wheat from international markets was considered to be equal to high-quality wheat from China's domestic suppliers. Interestingly, evidence that medium- quality wheat on international markets is the same as high-quality wheat supplied by China's farmers is found in the answer to the question asked of interviewees: "If China's higher-quality wheat were sold on international markets, how much loss would a trader incur?" This rate, 10 percent, was almost exactly the same as the premium importers would make from bringing in medium-quality grain from the international market. Huang, Rozelle, and Chang 69 part of China's wheat crop, a price gap of this size would likely mean that China's accession to the WTO would lead to imports of this type of wheat (in the absence of a value-added tax or other nontariff barrier). China's lowest-quality wheat (about 10-15 percent of its harvest) is at the world's feed wheat price. China did export some feed wheat into international markets in 2001 (mostly to Asia, according to an interview). Similar differences in the size of the price gap among varieties of a single type of commodity are found for rice, though not for soybeans and maize, which are more homogen- eous products (see table 3). New Nominal Protection Rates and Sources of Protection for Agricultural Commodities Although there are differences among major types of any individual agricultural commodity, more traditional aggregate nominal protection rates can be created by weighting the rates by sown area for crops or production shares for meats (table 4). When the individual nominal protection rates in table 3 are weighted by their area shares, wheat, for example, has an aggregate nominal protection rate of 12 percent. Rice, on the other hand, is implicitly taxed at 3 percent. The aggregate figures, though helpful (especially for analysis that is disaggregated only to the crop level), provide much less insight about which groups of farmers in which areas producing which varieties will be hurt or helped if trade liberal- ization were to reduce trade-related distortions. TABLE 4. Average Nominal Protection Rates for Major Imports and Exports in China, October 2001 Domestic price Nominal protection Commodity (yuan per ton) rate (o/o) Imports' Wheatb 1,250 12 Soybeans 1,950 15 Cotton 9,500 17 Sugar 2,612 40 Exportsa Riceb 1,954 -3 Maizeb 1,150 32 Porkh 11,442 -30 Beefh 13,743 -10 Poultryb 9,904 -17 Fresh fruit 5,472 -4 almports commodities are compared with international prices cif China, and exports are compared with international prices fob, China. bAverage nominal protection rates are created by summing the nominal protection rates of individual varieties weighting by share of area sown (production). Source: Authors' survey. 70 THE WORLD BANK ECONOMIC REVIEW, VOL. I 8, NO. I For commodities with fewer quality differences (such as maize, soybeans, sugar, and cotton to a lesser extent), the aggregate measures have more inherent interest. Maize and soybeans are rarely consumed directly (unlike rice and wheat, which are staple food grains that are sensitive to human tastes and preferences) and are used mostly as a feed or are otherwise processed. Thus, in the next part of the analysis, in which the observed protection rate is matched with the source of protection, only the aggregate nominal protection rates for maize, soybeans, sugar, and cotton are examined.8 The findings show not only that significantly positive rates of protection exist for a number of China's major field crops but also that they vary across the country and according to China's position as a net importer or net exporter. Maize prices, according to exporters, averaged more than 30 percent above world prices, meaning that traders would have lost more than 30 percent of the value of their exports without government subsidies. Protection rates differed across regions, however. For example, interviewees in the northeast said that if they were not exporting and foreign maize were to come into China, the importer could make 21 percent on average. In south China, however, the price gap between imported maize (cif China) and maize traded in the domestic market in and around Guangzhou was more than 35 percent. Aggregated across areas and weighted by maize consumption shares, the nominal protection rate on maize was 32 percent in 2001 (see table 4). The level of protection for maize corresponds almost exactly to the subsidies being paid to maize exporters during fall 2001. Interviewees also reported that despite the large increase in the volume of soybean imports in recent years, there is still an average 15 percent difference between the cif price and the domestic price in the port (see table 4). That a price gap remains seems remarkable considering that China imported almost 15 million tons of soybeans in 2001, the official tariff is only 3 percent, and the commodity can be traded by any foreign trade company. The remaining price gap shows that there may be other reasons for distortions beyond tariffs and state trading. In fact, the gap between the domestic and international price is almost certainly a result of China's policy of assessing a value-added tax on imported soybeans at the border. As already shown, the difference in the tax rate between imported and domestic soybeans is about 12 percent. Because this is the difference between the price of imported soybeans after paying the 3 percent tariff, this suggests that the main distortion in China's soybean price in fall 2001 was the value-added tax. 8. The survey was conducted the same way. In most cases, interviewees reported that there was not much difference in quality among maize varieties and that there was only a slight (around 2-3 percent) price difference between imported and domestic soybeans. Hence, questions were asked both ways. "What was the price difference if imported soybeans (cif China) were auctioned off in the domestic market with no taxes or tariffs added? What was the price difference if domestic soybeans (fob China) were auctioned off in the international market with no subsidies?" Huang, Rozelle, and Chang 71 Cotton and sugar were fairly highly protected in October 2001 (see table 4). Traders in south China and Shanghai reported that they could have sold a ton of imported sugar (cif China) for almost 40 percent more on China's domestic market if they did not have to pay any fees. In fall 2001 the official tariff for sugar was 40 percent (MOFTEC 2002). Thus for sugar the main distortion was the official tariff rate, with the value-added tax having almost no role. Cotton demonstrates one of the shortcomings of this new approach (and other approaches) to calculating nominal protection rates. In fall 2001 the average gap between the international and the domestic price of cotton was 17 percent. Inter- viewees told us that if they could costlessly bring imported cotton across the border in late September and early October 2003, they could earn 17 percent if they auctioned their shipment immediately. When the survey team did follow up work at the end of November, however, the domestic price of cotton had fallen from 9,500 yuan per ton to less than 8,000 yuan per ton, bringing the nominal protec- tion rate to around zero. Then by the end of the year (late December 2001), the international price of cotton also fell. Although the December follow-up surveys by phone covered only a few cotton traders and users, the information from that abbreviated survey indicated that the nominal protection rate was positive again. The domestic and international prices of other crops varied less than those of cotton in 200 1, but cotton provides a cautionary lesson of how nominal protection rates can change rapidly, even over a short period. Thus, if a statistical bureau were to adopt this method as a way to track nominal protection rates for a variety of commodities, the surveys would have to be repeated at periodic intervals. Assessing the New Methodology Because one objective was to use the new data and methods for aggregating variety-specific nominal protection rates to generate crop-specific nominal pro- tection rates, this section compares the nominal protection rates created by this time- and data-intensive approach with those using traditional methods, data sources, and assumptions.9 Although the two approaches yield similar results for some commodities, such as soybeans and maize, the results vary consider- ably for other commodities. For example, the average price of wheat imports in 2001, calculated by dividing total import value by total import quantity, was 1,393 yuan per ton, whereas the national average price for domestic wheat as 9. These are computed by comparing the domestic wholesale price with the average implicit inter- national price. For tradables it is total value of the import or export divided by the total volume. For some commodities, comparisons using these traditional methods and data may not be comparable to the estimates here because the traditional measures are calculated on an annual basis and those for the estimates here are for a single quarter (fall 2001). In this particular case, because the international and domestic prices of rice, wheat, and maize were fairly constant across the year, there is little bias. But the proper method would be to compare the fall in nominal protection rates from both methods or for nominal protection rates calculated for the entire year from both methods. 72 ITHE WORLD BANK ECONOMI-C REVIEW, VOL. I 8, NO. r reported in the Ministry of Agriculture's reporting system was 1,113 yuan per ton, or about 21 percent below the cif price of imports. Thus, the standard methodology would imply that wheat, rather than being protected (by 12 percent-see table 3), was actually being taxed by trading policies. Yet as the previous analysis shows, the main reason for the negative rate of protection is that China imports almost exclusively high-grade, baking- quality wheat, whereas its domestic consumers use mostly medium and lower qualities of wheat. Hence, the wrong conclusion is reached when using the specialty prices for imports as an international reference price for types of wheat that are of much lower quality and price. The same problem is found for rice.10 This shows the importance of estimating nominal protection rates more carefully, at least for certain commodities. The traditional approaches work reasonably well for commodities that are fairly homogeneous in quality, such as maize and soybeans. But for wheat and rice in China in 2001, comparing average prices inside and outside the country can yield misleading results. III. AWAY FRONM THE BORDER EFFECTS The entire effect of trade liberalization on agricultural prices (and the distribu- tion of the effect) depends not only on the size of the distortion but also on how the effects are distributed, which is largely a function of the nature of China's markets. At least three factors play a role: policy safeguards that prevent market forces from fully equilibrating domestic and international prices; household responses, which include shifting away from commodities whose price falls to production of higher profitability commodities; and high transaction costs, which can buffer the effects of liberalization policies in rural areas. This section focuses on the nature of markets. Policy safeguards are discussed in the conclu- sion. Household responses are discussed in Taylor (1998), OECD (2001), and Huang and others (2002). If large areas of the country are isolated from the coastal markets where imports enter the country, WTO accession would not be expected to have highly adverse impacts on the poor, most of whom live in inland areas far from major coastal cities. While being isolated from negative external shocks is a benefit, there are also costs. Those living in poor, isolated areas would not benefit from price rises when there are enhanced opportunities to export. Living in isolated markets also makes households more vulnerable to regional 10. Because China imports only high-quality jasmine rice from Thailand, the international price of rice, 3,908 yuan per ton (calculated by total import value divided by total import quiantity), appears to be more than 150 percent higher than the average domestic price, 1,464 yuan per ton. In fact, China's average price protection (tax) rate, calculated on a variety by variety basis, is almost zero (-3, see table 3). Huanzg, Rozelle, and Chaang 73 price shocks (which may be caused by regional production or consumption shifts). In contrast, with well-functioning markets, increased import or export flows concentrated in coastal cities will affect the prices that small, poor households face even when they live thousands of kilometers away. Moreover, if markets link inland and coastal areas, local shocks to supply or demand in inland regions are less likely to have any large effect on the prices local producers receive (which in most cases means that price variability will be less). To the extent that there are high transaction costs inside China and some domestic markets are isolated from others, the impacts of WTO accession policies may not be evenly distributed. Previous work on China's agricultural markets (for example, Park and others 2002) found that markets had become fairly integrated by the mid-1990s. However, certain qualifications apply. First, although markets improved greatly during the early 1990s (compared with the 1980s), the analysis found that in some years large regions of the country, especially poorer areas, were not completely integrated into national markets. Moreover, the study ends with 1995. It is unclear from the literature (although there is no rigorous national study of market integration in the late 1990s) and government market policies during the late 1990s whether markets were likely to have become more or less integrated since the mid-1990s (Nyberg and Rozelle 1999). Assessing Interregional Market Integration To assess market integration in rural China in the late 1990s and in 2000, data from China's State Market Administration Bureau were used to see how well prices in different markets moved together and how well integrated prices were between market towns and China's villages. DATA. The data come from a unique set of price data collected by the State Market Administration Bureau. Nearly 50 sample sites from 15 provinces report the prices of agricultural commodities every 10 days. The prices are averages of transactions that day in the local rural periodic market. The Min- istry of Agriculture (2001) assembles the data in Beijing and makes them available to researchers and policymakers. Price data were examined for rice, maize, and soybeans for 1996-2000 (price data for maize were available only through 1998). The three crops are produced and consumed in nearly every province in China. Because of quality differences among rice varieties in different regions of China, price integration was exam- ined among markets within four regions, South China, the Yangtze Valley, the North China Plain and Northwest China, and Northeast China. Prices for rice are available for more than 90 percent of the time periods for the provinces included in the sample. Prices of maize are available for 13 markets and prices 74 THE WORLD BANK ECONO.MIC REVIEW, VOL. i8, NO. I of soybeans for 20 markets." Product homogeneity for maize and soybeans makes it possible to examine price integration among markets across a broader geographic range. The results for 1996-2000 are compared with the results for 1988-1995 in Park and others (2002).2 INTEGRATION TESTS. This section uses more formal tests of market integration. Cointegration means that although many developments can cause permanent changes in the individual elements of a tested series (grain price here), there is some long-run equilibrium relation tying the individual components together, represented by the linear combination, as in equation 1. Here, the Engle-Gran- ger cointegration approach is applied to test China's market integration. The basic intuition behind the approach is that if one can write two price series in the following way: (1) U, = P't- W, and if each price series is stationary of order zero, I(O), then this condition implies the existence of a long-run equilibrium. In other words, in the long run the two series will eventually return to a constant mean. Moreover, a linear combination of the two prices shows that it is efficient to predict one market's price based on the information from another market's price. Equivalently, these two price series are cointegrated and the two markets are integrated.'3 If the price series are not stationary of order zero, then a unit root test is applied to determine whether each element of the price series is stationary of order one, I(1). The analysis shows that all price series for the commodities in China's grain markets in the late 1990s are stationary of order one. Using the stationary price series, one price series is then regressed on another using ordinary least squares: (2) P' = o + At + 4P{ + et 11. Because time-series data are used, prices must be converted to a real basis. Nominal prices from the data set are deflated using monthly consumer price indices calculated and reported by the China National Statistical Bureau. Deflation facilitates transaction cost comparisons across time and allows transaction cost increases within periods associated with inflation to be disregarded. 12. To produce the results, cointegration tests are run on each pair of markets using the data for each year. In other words, 36 observations are used (because the price data are available every 10 days) and the number of pairs of markets that are cointegrated in a statistically significant way are counted (see note 13 and text for explanation of testing). For example, for the case of soybeans for the late I 990s (1996- 2000), this means that being examined is the extent of integration between 190 (20 * 19/2) pairs of markets in each of five years, which equals 950 pairs of markets. So because prices in 646 markets were found to be integrated (according to the testing procedure), 68 percent of markets are reported as integrated in the late 1990s. Because only 36 observations are used per test, and because cointegration tests typically perform better with longer time series, by splitting data into annual increments, the results are biased against integration. This makes the analysis comparable to Park and others (2002), which follows a similar procedure. 13. Note that the b coefficient need not be unity to conclude cointegration and integrated markets (only needed for applying the much more restrictive criteria of the Law of One Price). Huang, Rozelle, and Chang 75 where t is the common trend of the two price series and et is the error term. The residual, et, is then used in the augmented Dickey-Fuller test: N (3) Aet = 6et- I + E -yAet-i + t . i-2 If the test statistic on the ( coefficient is less (more negative) than the relevant critical value from the Dickey-Fuller table, the null hypothesis is rejected and the two series are said to be cointegrated of order (1,1). According to Engles and Granger, this implies that the two markets are integrated. The analysis assumes that markets are integrated when the absolute value of the test statistic is greater than 3 (implying significance at the 10 percent level). RESULTS. The cointegration analysis shows that China's markets have continued to develop in the late 1990s, especially when the results are compared with the market integration research of the late 1980s and early 1990s (table 5). In the middle part of the reform era (1988-95), a time when markets were starting to emerge, some 20-25 percent of markets showed signs of prices moving together (Park and others 2002). Using the results from the early 1990s as a baseline, the current analysis shows that during the late 1990s China's markets continued along their path of matura- tion. The comovement of prices among pairs of markets in the sample shows significant increase in the share of market pairings that are integrated. In the case of maize, for example, prices in paired markets moved together in 89 percent of the cases, up from 28 percent in the early 1990s (table 5). The share of market pairs showing price integration also increased for soybeans, japonica rice, and indica rice. The integration is especially notable because in many cases the paired markets are more than a 1,000 km apart. For example, in many years soybean and maize prices were found to be integrated between markets in Shaanxi and Guangdong Provinces and between Sichuan Province and southern Jiangsu. Despite significant progress in integration, the results also show pairs of markets that are not integrated. For example, in a third of cases japonica rice prices moved in one market but not in another. One explanation is an TABLE 5. Percentage of Market Pairs in Rural China that Test Positive for Integration Based on Dickey-Fuller Test, 1988-2000 Commodity 1989-95 1996-2000 Mlaize 28 89 Soybeans 28 68 Japonica rice (Yellow River Valley) 25 60 Indica rice (Yangtze Valley and South China) 25 47 Note: Results are for two periods from same data set. For results for 1989-1995 for maize and rice, see Park and others (2002). Rice results are for the whole country in 1989-95. Results for soybeans for 1989-95 and all results for 1996-2000 are from the authors. 76 THE WORLD BANK ECONOMIC REVIEW, VOL. r 8, NO. I institutional breakdown or infrastructure barrier (a policy measure or a weak link in the transportation or communication infrastructure) that is fragmenting China's markets for certain commodities, as shown in Park and others (2002). But because every province in China produces and consumes rice, it is also the case that if supply in one region during one period is just equal to demand and if regional price differentials stay within the band between regional "export" and "import" prices, then moderate price movements in another area might not induce a flow into or out of the region that is in equilibrium. For that reason, despite the nontrivial number of cases in the late 1990s in which market prices in pairs of markets do not move together, it must be concluded that the impacts of wro accession on China's agriculture will increasingly be experienced across wide regions of the nation from coastal to inland areas. Assessing Village Integration into Regional Markets The interregional integration of markets is only half of the story, however. The remarkable degree of integration between coastal and inland markets is still not sufficient to state confidently that village households are integrated into the nation's marketing network. That requires analysis of the extent to which villages are integrated into regional markets. This integration test looks at whether farmers are price takers or whether they reside in isolated villages in which local prices are determined by local supply and demand. The equation to test for village-regional market integration is: (4) P, = aO+al*Ai + b1*Ti +dl-ID +e,. In brief, if variables that affect local grain availability, Ai, in village i signifi- cantly affect the village's price, Pi, villages are assumed to be isolated from markets. If the variables that affect local availability do not affect the price, villagers are assumed to be price takers and markets can be thought to be integrated. Availability in each village during the survey year is measured as the sum of production, PS,, and storage, Si. If markets are isolated, a rise or fall in availability would be expected to negatively or positively affect the village's price. In contrast, if markets are integrated, changes in local availability would be expected to have no affect on the village's price. Because it is total availability (production plus storage at the beginning of the period) that matters (Ai = PS, + Si), total availability should enter equation 4. Equation 4 is solved separately for rice, wheat, maize, and soybeans. In examining the impact of local grain availability on the household's grain price in equation 4, other factors, Di, need to be controlled for in the cross- sectional analysis. In equation 4, D, is assumed to include two components, one spatial (the distance of the village from the county seat, the typical site of the regional market) and one temporal (the timing of the grain sale). The further the village is from the county seat and the closer the grain sales are to the harvest (within the first three months), the lower the price is likely to be. Because village Huang, Rozelle, and Chang 77, price levels in different provinces are expected to vary according to each pro- vince's location (with respect to the port) and infrastructure (the quality of its road and rail network), a provincial dummy variable is also included. For rice, because quality varies so much from region to region, dummy variables for regional quality are included (for South China, the Yangtze River Valley, and North/Northeast China). The data were collected in a randomly selected, nearly nationally representa- tive sample of 60 villages in six provinces of rural China (the China National Rural Survey). To accurately reflect varying income distributions within each province, one county was randomly selected from each income quintile for the province, as measured by the gross value of industrial output. Two villages were randomly selected within each county. The survey teams used village rosters and their own counts to randomly choose 20 households, both those with residency permits (bukoii) in the village and those without. A total of 1,199 households were surveyed. 14 A number of variables were constructed that might affect the price that farmers received in the village. The survey team gathered detailed information on both production and marketing behavior of all farmers in the sample and the char- acteristics of each village and its relationship to the nearest regional market. Individual respondents provided information on the price and timing of sales for each commodity. The prices for all household sales in the village were averaged, with each sale weighted by its volume in kg. From the information on timing a set of variables was constructed that measures the proportion of village sales occurring within each of the first three months after the harvest. A community questionnaire provided information on how far the village's center is from the nearest paved road and the distance to the county market. Finally, for each crop there was information on any shocks, both their incidence and the percentage by which the yield fell. These were aggregated to the village level. There are no variables that control for the presence of a community buffer stock system, primarily because such institutions are rare in modern China. However, farmers, at least in the past, have been known to hold large stores of grain. It is possible that in an isolated village hit by a production shock that caused local prices to rise, farmers could draw on their own stocks so that the local price would exhibit no net change, thus making the village look as though 14. The China National Rural Survey project team gathered detailed information on the production and marketing behavior of all of the farmers in the sample, the characteristics of each village, and its relation- ship to the nearest regional market. Each individual respondent in the survey in each village gives the price and timing of the sale for each commodity. From these data, an average village price for each month are constructed in yuan per kilogram. In a community questionnaire, how far the village's center is from the nearest paved road and the distance to the county market is determined in kilometers. Finally, for any shocks to the farmer's crop, the incidence and the percentage by which the yield fell are known for each crop that the farmer cultivated. No variable that controls for the presence of a community buffer stock system is included, primarily because such an institution is almost never observed in modern China. In addition, sales among farmers wNithin a village are rare (according to data, less than 5 percent of sales). 78 THE WORLD BANK ECONOMIC REVIEW, VOL. 1 8, NO. I it were integrated into the regional market, when it was not. Beginning year stocks, aggregated to the village level, are used to measure the potential that households stocks might have for increasing availability. To test the hypothesis, grain price, Pi, is a function of total grain availability, Ai, for each of the main staple crops i (where i = 1), holding the other variables, T, and D, constant. Total grain availability is measured in three ways: as the production shock, PS,, alone; as the production shock, Pi, and grain storage, Si; and as the interaction between the grain storage variable and the production shock variable (or a direct proxy of Ai = PSi + Si). Because the interaction effect is the most intuitive (it captures total grain availability of the village in one vari- able), the regressions that use this version of the variable are reported in table 6. (Results of regressions using the alternative variables are reported in appendix tables A.1 and A.2.) If villages are isolated from regional markets, the coefficient on the interaction term should be negative and significant when there is positive production shock and high levels of grain storage-that is, when the interaction term is large. If markets are integrated into China's larger marketing networks, the coefficient should be insignificant. The analysis clearly shows that markets in China are integrated down to the village level (see table 6). The signs on the coefficients (and levels of significance in some cases) on the variable measuring the distance of a village from the regional marketing center demonstrate that the further a village is from a market, the lower the price the farmer receives. More important here, the t-ratios of the coefficients of the village supply shock variables are all small, signifying that the output of the local village's crops does not affect the local price. The implication is that factors outside the village are the primary influence on the prices that farmers receive, making them price takers. Moreover, when the main variables of interest are interacted with a dummy variable for village income level (equaling 1 when a village is in the bottom two income deciles), the coefficient is still insignificant. In other words, even farmers in China's poor, remote villages are linked to China's regional markets. IV. CONCLUSIONS This article looked exclusively at the effect of China's accession to the wTo on agricultural prices (although other effects may be equally large or larger). The analysis found that there will be an impact on most farmers in the economy-both those in coastal areas and small, poor farmers in inland areas. The findings, based on new methods to collect data and create more accurate nominal protection rates, show that for some crops, WTO accession will likely lead to a fall in prices and a rise in imports. Maize and cotton prices may be most affected. Soybean and sugar prices could also fall. However, not all effects are negative. There are also com- modities in which China has considerable comparative advantage (rice, meats, and horticulture products) and for which wTo accession could provide benefits to those engaged in these activities to the extent that markets in other countries become TABLE 6. Ordinary Least Squares Regression Explaining Effect of Local Grain Availability on the Price of Major Crops in China's Villages in 2000 Variable Rice Wheat Maize Soybean Explanatory variable Local grain availability Village-level climate shocks (production shock)a Village-level grain storage at the beginning of year (grain storage)a Interaction: production shock -3.15e-06 (1.31) 7.50e-07 (0.37) -3.91e-07 (0.33) 0.000045 (0.15) and grain storage' Control variables Distance to the nearest county (km) -0.00074 (0.74) -0.0079 (2.1) -0.0005 (0.55) -0.032 (2.76)' Variables representing proportion of Included Included Included grain marketed during each of the first three months after harvest Quality dummy variables Included Provincial dummy variables Included Included Included Included Adjusted R2 0.16 0.38 0.50 0.15 Number of observations 31 30 28 17 Note: Village-level price is the dependent variable. t-ratios in parentheses. *Significant at the 5 percent level. aIndependent measures of production shocks and grain storage are not included in this version. See appendix tables A.1 and A.2 for versions that includes these variables. 80 THE WORID BANK ECONOMIC REVIEW, VOL. I8, NO. I more open to China's exports. The prospect of increased imports of feed grains (maize and soybeans) at lower prices means that livestock producers could become even more competitive. How much prices fall because of rising imports or increase because of rising exports in part depends on how China executes its WTO obligations, especially the agreements affecting some of the more nontraditional barriers that were shown to be protecting China's farmers before accession to the wTo. Although there may be room for delay, which could slow the negative effects (for exam- ple, China was subsidizing maize exports in 2003 to keep domestic maize prices from falling and hurting maize producers; Rozelle 2003), the agreement also contains several provisions to limit downside effects. For example, although the tariff rate quotas section of the accesses agreement lowers tariffs and provides access for nonstate traders to import commodities such as cotton, sugar, and edible oil, it also caps the quantity that can be imported at the low tariff rates. Likewise, the size of benefits to China's producers will depend on how well its trading partners honor their commitments to provide China with better access to global markets for products in which China has a comparative advantage. Unlike the case of Mexico after NAFTA it appears as though most of China's villages, even those in remote, inland regions, may be well integrated into the economy. This is good news and bad news for poor farmers. The good news is that they can benefit from falling input prices and rising export opportunities. The bad news is that if the results here are correct and maize and cotton prices fall for large parts of China, poor farmers will be affected. The problem, although a short-run one, may affect the poorest households most-households most dependent on agriculture and least able to adjust their cropping structure. Thus, the findings should signal to government leaders the need to consider the welfare effects on these susceptible groups. APPENDIX: CHALLENGES AND ISSUES IN MEASURING NOMINAL PROTECTION RATES The wide range of nominal protection rate estimates for China demonstrates that measuring differences between an economy's domestic price and the inter- national price is not straightforward. Several issues complicate such measure- ment. First, confusion may stem from the way analysts have asked their question about nominal protection rates. Policymakers and researchers have sought to summarize the impact of various commodities in a single number. Trade modelers need a single number to make their analytical frameworks tractable. People want to know the price of wheat in China and to be able to compare that to the world price of wheat. With this information, the nominal protection rate of a commodity is simply the difference between these two numbers. However, more careful observation shows that the search for a single number may be one of the main reasons why analysts reach so many different conclusions. TABLE A-1. Ordinary Least Squares Regression Explaining Effect of Local Grain Availability on the Price of Major Crops in China's Villages in 2000 Variable Rice Wheat Maize Soybean Explanatory variable Local grain availability Village-level climate shocks -0.108 (1.05) 0.06 (0.61) 0.109 (1.23) -0.11 (0.49) (production shock) Village-level grain storage at the beginning of year (grain storage)' Interaction: production shock and grain storage' Control variables Distance to the nearest county (km) -0.00069 (0.69) -0.0081 (2.15); -0.0007 (0.79) -0.031 (2.75)* Variables representing proportion of Included Included Included grain marketed during each of the first three months after harvest Quality dummy variables Included Provincial dummy variables Included Included Included Included Adjusted R2 0.14 0.38 0.53 0.16 Number of observations 31 31 28 17 Note: Village-level price is the dependent variable. t-ratios in parentheses. "Significant at the 5 percent level. 'The grain storage variable and interaction variable are not included in this table. See appendix table A.2 for version that adds grain storage variable. See table 6 for a version that includes that interaction term. TABLE A-2. Ordinary Least Squares Regression Explaining Effect of Local Grain Availability on the Price of Major Crops in China's Villages in 2000 Variable Rice Wheat Maize Soybean Explanatory variable Local grain availability Village-level climate shocks -0.108 (1.02) 0.06 (0.6) 0.132 (1.3) -0.206 (0.96) (production shock) Village-level grain storage at the -8.21e-08 (0.18) 1.12e-07 (0.14) -4.04e-07 (0.5) 0.00018 (1.66) beginning of year (grain storage) Interaction: production shock and grain storage' Control variable Distance to the nearest county (km) -0.00069 (0.66) -0.0082 (2.11):- -0.0005 (0.51) 0.034 (3.19)^ Variables representing proportion of Included Included Included grain marketed during each of first three months after harvest Quality dummy variables Included Provincial dummy variables Included Included Included Included Adjusted R2 0.10 0.36 0.51 0.29 Number of observations 31 30 28 17 Note: Village-level price is the dependent variable. t-ratios in parentheses. Significant at the 5 percent level. 'The interaction variable is not included in this table. See table 6 for a version that includes interaction term. Huiaing, Rozelle, and Chang 83 There are many prices for wheat in China. Prices vary within a year. They vary across regions. What price should be used in calculating the nominal protection rate? Should it be the price of corn in a Guangzhou feedlot or the price of corn sitting in storage in a farmer's homemade silo in Northeast China? Moreover, not all rice is the same. There are many different varieties and types, all commanding different prices at different places and times during the year. The same sets of issues faces analysts when they attempt to choose a price series (or more difficult yet, the single price) to represent the international price. Should it be fob or cif? Should it be the average annual price or a price during a particular period? If there are many different types of imported varieties, which type should be chosen? In part because previous studies have not dealt with these issues (at least explicitly), it is unsurprising that different research efforts have generated dif- ferent estimates of nominal protection rates. For example, Tuan and Cheng (1999) estimated high and variable nominal rates of protection for agricultural commodities: 62 percent for wheat, 15 percent for maize, and 140 percent for soybeans in 1997. Carter and Estrin (2001) find generally negative price distor- tions. Huang (200 1) provides estimates that show that some products are highly protected and others have negative rates of protection. REFERENCES Anderson, K., and C. Y. Peng. 1998. "Feeding and Fueling China in the 21st Century." World Develop- ment 26(8):1413-29. Anderson, Kym, Jikun Huang, and Ilena lanchovichina. Forthcoming. "Impacts of WI'To Accession on Chinese Agriculture and Rural Poverty." In Will Martin, ed., China and the WTt: Accession, Policy Reform, and Poverty Reduction Strategies. New York: Oxford. CARD (Center for Agricultural and Rural Development). 2001. "Effect of wTo on Agriculture." Paper presented at the 2001 annual meeting of the American Agricultural Economics Association, August 5-8, Chicago, Ill. Carter, C. A., and A. Estrin. 2001. "China's Trade Integration and Impacts on Factor Markets." Working Paper, University of California, Davis. Chen, Shaohua, and Martin Ravallion. Forthcoming. "Welfare Impacts of China's Accession to the wTno. In Will Martin, ed., China and the WTO: Accession, Policy Reform, and Poverty Reduction S.. New York: Oxford. Garcia, Roberto. 2003. "A Case against the Simultaneous Use of Mlarket Access Restrictions, Domestic Support, and Export Subsidies." Paper presented at the 2003 Meeting of the International Association of Agricultural Economics, August 16-23, Durban, South Africa. Huang, Jikun. 2001. "Agricultural Policy and Food Security." Working Paper, Chinese Academy of Sciences, Center for Chinese Agricultural Policy, Beijing. Huang, J., and C. Chen, 1999. "Effects of Trade Liberalization on Agriculture in China: Institutional and Structural Aspects." UN Economic and Social Commission for Asia and the Pacific, Centre for Coarse Grains, Pulses, Roots, and Tubers, Bogor, Indonesia. Huang, Jikun, Scott Rozelle, and Yuping Xie. 2002. "Distortions at the Borders; Integration Inland: Assessing the Effect of wTo Accession on China's Agriculture." Journal of Chinese Economic and Bnsiness Studies 1(1):97-116. Li, S., F. Zhai, and Z. Wang. 1999. "The Global and Domestic Impact of China Joining the World Trade Organization." Project report, Development Research Center of the State Council of PRC, Beijing. 84 THE WORLD BANK ECONONMIC REVIEW, VOL. I 8, NO. I Martin, W. 2001. "Implications of Reform and WTro Accession for China's Agricultural Trade Policies." Economics of Transition 9(3):717-42. Ministry of Agriculture. 2001. China Statistical Yearbook. China Agricultural Press: Beijing. MOFTEC (Ministry of Foreign Trade and Commerce). 2002. Import and Export Trade Management Measures. Beijing: Industrial and Business Press. Nyberg, A., and S. Rozelle. 1999. Accelerating China's Rural Transformation. Washington, D.C.: World Bank. OECD (Organisation for Economic Co-operation and Development). 2001. "Implications of Trade and Investment Liberalization for China's Agriculture." Working paper. Paris. Park, A., H. Jin, S. Rozelle, and J. Huang. 2002. "Market Emergence and Transition: Transition Costs, Arbitrage, and Autarky in China's Grain Market." Anmerican Journal of Agricultural Economics 84(1):67-82. RCRE (Research Center for Rural Economy). 2000. "Agricultural Policy in China, Annual White Paper." Working Paper, Research Center for Rural Economy, Ministry of Agriculture, Beijing. Rozelle, Scott. 2003. "After Accession to the wro: Corn Trade within China and between China and the Rest of the World." U.S. Grains Council, Beijing. State Price Bureau. Various years. China National Price Yearbook. Beijing: State Price Bureau Press. Taylor, J. Edward. 1998. "Trade Liberalization and the Impact on Smallholders in Rural Mexico." Working Paper, Department of Agricultural and Resource Economics, University of California, Davis. Tuan, F. C., and G. Cheng. 1999. "A Review of China's Agricultural Trade Policy." Paper prepared for the iatrc Summer Meeting "China's Agricultural Trade and Policy: Issues, Analysis, and Global Consequences," June 25-26, San Francisco, Calif. Xie, Yuping. 2002. "wro and China's Agricultural Trade and Impact on Domestic Markets." Unpublished masters' thesis, Chinese Academy of Sciences, Center for Chinese Agricultural Policy, Beijing. Zhao, Yaohui, and Terry Sicular. 2002. "Effects of wro Accession on China: Labor Market Responses." Working Paper, Peking University, Center for Chinese Economic Reform, Beijing. Regulated Efficiency, World Trade Organization Accession, and the Motor Vehicle Sector in China Josepb F. Francois and Dean Spinanger This article is concerned with the interaction of regulated efficiency and World Trade Organization (wTo) accession and its impact on China's motor vehicle sector. The analysis is conducted using a 23 sector-25 region computable general equilibrium model. Regulatory reform and internal restructuring are found to be critical. Restruc- turing is represented by a cost reduction following from consolidation and rationaliza- tion that moves costs toward global norms. Without restructuring, WTO accession means a surge of final imports, though imports of parts could well fall as production moves offshore. However, with restructuring, the final assembly industry can be made competitive by world standards, with a strengthened position for the industry. Producing automobiles has often been a symbol of economic prestige in the developing world. Brazil, China, Indonesia, Malaysia, and others have all promoted and sometimes even showcased the development of a domestic motor vehicle industry. In China, with its huge population and a surface area roughly as large as the United States and almost 15 percent larger than Brazil (table 1), almost every province has its own motor vehicle factory and satellite factories. But despite all the factories, China has the largest number of people per vehicle among major economies in the world. Even Indonesia, with a 30 percent lower per capita income, has only half as many people per automobile. China's status as a country with one of the highest number of people per vehicle is the outcome of a series of policy measures, dating as far back as the inception of the People's Republic of China, that have distorted the structure of the automobile industry (table 2). Internal measures limited and even pro- hibited trade through local protectionism (analogous to former interprovincial trade restrictions in Canada). The government has also set prices and limited competition through a barrage of import restrictions, which have included Joseph F. Francois is professor of economics and research fellow at the Tinbergen Institute and research fellow at the Centre for Economic Policy Research; his e-mail address is francois@few.eur.nl. Dean Spinanger is senior research economist at the Institute for World Economies in Kiel; his e-mail address is dspinanger@ifw.uni-kiel.de. The authors thank Zhang Wenkui for help with data, Will Martin and Alan Winters for detailed comments on an earlier draft, and three anonymous referees who offered valuable suggestions. They also thank participants in a World Bank-sponsored conference in Beijing for helpful discussion. THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. I, © The International Bank for Reconstruction and Development / THE WORLD BANK 2004; all rights reserved. DOI: 10.1093/wber/lhhO34 18:85-104 85 86 THE WORLD BANK ECONOMIC REVIEW, VOL. i8, NO. I TABLE 1. GNP, Population, and Stocks of Automobiles in China and Selected Countries, 2000 GNP per Vehicle stock, capita Population, 2001 (millions) (2002 2001 People per Surface area Economy ppp $)a (millions) Automobiles Trucks automobile (1,000 sq. km) Low and middle 4,682 3,274.4 140.6 54.9 23.3 49,263 income India 2,570 1,032.4 6.3 5.9 163.2 3,287 Indonesia 2,990 209.0 3.0 2.4 68.8 1,905 China 4,390 1,271.8 8.5 15.4 149.0 9,598 Colombia 5,870 43.0 1.8 0.8 23.4 1,139 Turkey 6,120 66.2 4.5 1.6 14.6 775 Thailand 6,680 61.2 2.9 4.1 21.4 513 Brazil 7,250 172.4 15.8 4.0 10.9 8,547 Russia 7,820 144.8 21.2 5.1 6.8 17,075 Malaysia 8,280 23.8 4.2 1.0 5.6 330 Mexico 8,540 99.4 12.2 5.6 8.2 1,958 South Africa 9,870 43.2 41.0 2.5 1.1 1,221 Argentina 9,930 37.5 5.4 1.6 7.0 2,780 High income 29,248 742.7 351.6 129.4 2.1 21,937 Korea, Rep. of 16,480 47.3 8.9 4.0 5.3 99 Taiwan, China 17,730 22.4 4.8 0.9 4.6 36 Spain 20,460 41.1 18.2 4.2 2.3 506 Italy 25,320 57.9 33.2 3.8 1.7 301 United Kingdom 25,870 58.8 27.8 3.4 2.1 243 Japan 26,070 127.0 53.5 19.9 2.4 378 France 26,180 59.2 28.7 5.9 2.1 552 Germany 26,220 82.3 44.4 3.6 1.9 357 Canada 28,070 31.1 17.1 0.7 1.8 9,971 United States 35,060 285.3 128.7 88.0 2.2 9,629 appP is purchasing power parity. Source: World Bank, various years, World Development Indicators; Verband der Automobilindustrie, various issues. quotas, high tariffs, and differential taxes favoring local suppliers. The restric- tions on trade have encouraged inefficient production and allowed for market segmentation. China's integration into the World Trade Organization (wro), and thus into most favored nation principles, has important implications for its economy, especially the motor vehicle sector. Accession agreements define major changes in tariffs, quotas and local content requirements, and rules on foreign investment. There has already been a change in market perceptions by outside investors, as the application of wro rules on treatment of foreign firms has reduced uncertainty about the general economic climate, inducing notable increases in investment and prompting new decisions about entering the market. Francois and Spinanger 87 TABLE 2. Summary of Developments in the Chinese Automotive Sector Period Characteristics 1953-65: Self-reliance policy Roughly 60,000 vehicles produced per year. Relied on Soviet technologies. No other international contacts. Provincial governments set up production units. By 1960, 16 auto producers and 28 assembly companies. 1966-80: Security oriented Government invested heavily in western regions (Sichuan, Shanxi, and Hubei). Remote locations caused severe problems and overcapacity. Focus on heavy military vehicles. Car demand increased rapidly and capacities expanded to 160,000 units a year. By 1980, 58 carmakers, 192 assembly companies, and 2,000 spare parts producers. 1981-98: Initial fruits of open-door policy Open-door policy in 1978 kick-started industry. VW already started in 1978. Number of companies almost doubled during 83-85, from 65 to 114 units. By 1998, roughly 2,500 production units. Provincial governments further regionalized production. Major international firms begin to invest and then stop rapidly. Joint ventures accounted for about 60 percent of production. 1999-present: opening up and beyond Major investments by foreign companies. All major Japanese and German companies in China. French, Italian, and U.S. producers nominally present. Rapid expansion; capacity now near 2.5 million units. Growing capacity in costal areas. Source: Summary produced by authors from various sources. This article is concerned with the impact of these broad changes on the Chinese motor vehicle sector. It emphasizes the role of administratively imposed inefficiencies ("regulated efficiency") within the sector and the role of such regulated efficiency in structural adjustment. The industry itself anticipates significant change. In recent years the sector has grown rapidly, with output expanding at an annualized rate of 13 percent in the four years ending in 1999, at a rate of 26 percent in the three years to 2002, and at more than double that in 2003. With modern plants having come on line in 2001 and 2002, and 8 8 THE WORID BANK ECONOMIC REVIEW, VOL. 1 8, NO. I additional facilities expected to increase capacity by more than 150 percent from 2002 to 2005, a large, discrete change in production levels is expected. At the same time, WTO membership implies lower prices and steeper foreign competition in the sector. Response to this shift in the competitive landscape will be shaped by continuing problems with local government protection, lack of automobile infrastructure (roads, parking, service facilities), and related factors that act as constraints on growth of the sector. Even so, the industry itself expects continued strong growth.' Notwithstanding industry expectations, what can realistically be expected once the competitive landscape has changed in critical ways? This question is explored here using a computable general equilibrium (cGE)model. I. THE AUTOMOBILE INDUSTRY IN CHINA National and regional policies in China have resulted in a highly fragmented and inefficient motor vehicle industry by global standards. This was not only the result of the introduction of Soviet-style industrialization beginning in the 1950s, with firms viewed merely as production units producing according to plan, making questions about efficiency irrelevant.2 It was also the result of import substitution policies and cooperation agreements with foreign com- panies beginning in the 1980s that were meant to fill the increasing gap between supply and rapidly expanding demand for automobiles. The major thrust of policies was to build trucks, not passenger cars (figure 1). Motor vehicle companies are thus operating with cost structures that are well within the global frontier, with plants that are producing considerably below global standards for efficient scale (table 3). For plants producing a single model, minimum efficient scale for final assembly of cars has been estimated 1. See, for example, China Online (2000). As Wro membership approached, the opinions of the industry and related ministries, as reflected in the Chinese press, hinged critically on whether restructur- ing of the domestic industry would be allowed to proceed. Thus a report in Touzi Yu Hezuo (summarized in China Online 2000) stressed expected injury to the industry, whereas the industry itself was expressing optimism that it could realize significant cost reductions and remain competitive with imports (Feenstra and others 2001). In the meantime, price cuts by foreign producers in China are becoming common, some of them induced by increased import competition and others by more intense domestic competition. Buick, for instance, reduced prices on its domestically produced models by 12 percent, and Volkswagen lowered Passat prices by 6.5 percent (indiacar.net, May 3, 2002). Even more important, nearly all major foreign producers have announced plans to establish or sizably increase production capacities. A recent major manufacturer to do so was DaimlerChyrsler in September 2003, finally ratifying plans to establish facilities to produce C and E models in China (International Herald Tribune, September 9, 2003). 2. As noted by Zhang and Taylor (2001, pp. 261ff.), First Automobile Works (FAW) provides ample evidence of the impact of various policies over the past 50 years. Between 1959 and 1981 FAW produced a mere 1,542 units, an average 67 units annually. In 1970 the production cost of a particular model (the CA72) was 220,000 yuan, but "the sales price was only 40,000 yuan.... In the absence of competition, all production units ran at low levels of productivity and efficiency. ... By 1980 the number of automotive enterprises had risen to 2,379, consisting of 56 vehicle manufacturers ... [producing] 5,418 cars." Francois and Spinantger 89 FIGURE 1. China's Production of Motor Vehicles, 1984-2002 (thousands) 3,500 3,000 2,500 2,000 l ~~~~~~~~~~~Total motor ve h c / / 1,500 - 1,000 - Trucks and buses - I 500 198 - - - - -Passenger cars 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Source: Bessum (2002) and Verband der Automobilindustrie (various issues). at more than 200,000 units per plant per year (Huang 2002, p. 543). China's entire sedan production in 1998 was 507,000 vehicles produced in 13 factories. Of these, only one factory produced more than 200,000 sedans. Several plants had production runs of fewer than 20,000 sedans.3 In 1998 China had 122 motor vehicle manufacturing plants, 520 automobile refitting factories, 130 motorcycle factories, 62 car engine factories, and 1,589 auto- mobile and motorcycle spare parts factories. Annual production capacity now exceeds 2.3 million motor vehicles and 10 million motorcycles. Since 1995 the general pattern has been to shut down smaller plants (generally relegated to the "other" category in table 3), and expand production runs in the larger plants. With foreign investment and rapid growth in the industry, the number of plants producing at least 25,000 vehicles rose from 3 in 1995 to 11 in 2002. Import and domestic shipment data in value terms for 1997, the preaccession reference point, are summarized in table 4. Reflecting relative tariff differences, imports are concentrated in parts rather than in vehicles. China's preaccession average tariff on automobile products (vehicles and parts) was 35 percent (table 5). The rate for vehicles averaged 70 percent, with sedans subject to tariffs of 80-100 percent. Parts were subject to an average tariff of 23 percent. Import shares were 3. There are strong parallels to the situation in Mexico before the North American Free Trade Agreement (Lopez de Silanes and others 1994), where protected, inefficient factories operated well within the global technology frontier. TABLE 3. Number of Passenger Cars Produced by Plants in China, 1995-2002 Rank 2002/1995 Plant 1995 1996 1997 1998 1999 2000 2001 2002a 1/1 Shanghai-Volkswagen 160,070 200,222 230,443 235,000 230,946 221,524 230,378 248,000 2/4 FAW-Volkswagen 24,553 44,825 46,405 66,000 81,464 94,147 101,622 131,000 3/NA Shanghai-General Motors - - - - - 30,024 58,548 106,000 4/2 Tianjin Xiali (Daihatsu) 65,258 88,232 95,155 100,021 101,828 81,951 41,703 93,000 5/5 FAW-Audi-Hongqi 19,350 15,000 15,731 31,225 52,667 78,000 6/9 Shenlong (Citroen) 3,797 9,228 30,035 36,240 40,200 53,900 52,850 68,000 7/6 Changan (Suzuki) 17,770 16,420 35,160 36,239 44,583 48,235 50,573 64,000 8/NA Guangzhou-Honda - - - 2,246 10,008 32,228 51,153 60,000 9/NA Shanghai-Qirui - - - - - 2,767 30,085 47,000 10/NA Geely Group - - - - - 14,594 21,702 38,000 11/NA Dongfeng Fengshen - - - - - 3,159 8,000 32,000 12/NA Haima (Nainan-Mazda) - - - - - 3,059 7,800 20,000 13/NA Yuedo-Kia - - - - - 2,423 6,210 16,000 14/NA Qinchuan - - - - - 5,380 5,686 16,000 15/NA Nanya - - - - - 1,000 8,000 13,500 16/3 Beijing (Jeep) 25,127 26,051 19,377 8,344 9,294 4,867 4,663 4,400 17/7 Guizhou Yunque (Subaru) 7,105 798 1,000 - - 859 1,253 2,100 18/NA Tianjin-Toyota - - - - - - - 2,000 NA/8 Guangzhou-Peugeot 6,698 2,416 1,557 - - - - - Other 22,570 - 22,479 8,013 31,312 17,930 1,900 Total 352,298 388,192 481,611 507,103 565,366 649,272 732,883 1,040,900 Number of plants 3 4 5 5 5 8 9 11 producing > 25,000 cars Number of plants 2 2 2 3 3 4 7 8 producing > 50,000 cars Number of plants I 1 1 2 2 1 2 3 producing > 100,000 cars NA, not applicable. -, Not available or plant did not exist. aValues are based on company projections. Souirce: Bessum 2002; Chinese Motor Vehicle Documentation Center 2002. Francois and Spinaniger 91 TABLE 4. China's Motor Vehicle Industry before World Trade Organization Accession, 1997 (millions US$) Sector Amount Imported motor vehicles and parts, world prices 3,607.7 Imported motor vehicles and parts, internal prices 4,849.3 Imported parts 3,239.5 Imported motor vehicles 1,609.9 Domestic motor vehicles, intermediates, and parts 32,812.5 Intermediates and parts 10,896.2 Industry consumption of motor vehicles 21,625.5 Final consumption of motor vehicles 290.8 Souirce: McDougall 2001. TABLE 5. Tariffs on Motor Vehicles in China (percent) Item 1997 rate Final rate Finished motor vehicles 71 25 Mlotor vehicle parts 23 10 Electronic parts 12 10 Average motor vehicles and parts 35 1 5 Soufrce: China wTo accession schedule, GTAP data, and Office of the U.S. Trade Representative. relatively low, averaging perhaps 3 percent during 1995-2002. Officially, only 20,000 sedans were imported, though many more were likely smuggled into the country.4 Official policy encouraged the use of domestic parts and favored locally (regionally) produced parts. Domestic content rules applied to new investments, stipulating 80 percent domestic content by the third year. The effects of these policies are reflected in the low share of imported automotive parts imports in total production. Even after China's completion of WTO accession, foreign ownership will be limited to 50 percent.5 Tariff rates are scheduled to come down substantially as part of the WTO accession process: by 25 percent on vehicles and 10 percent on parts on a most favored nation basis. Quotas will be phased out by 2006 and will be reduced by 15 percent a year until then. Domestic content requirements have already been removed. (Both of these nontariff barriers violate basic WTO rules.) Other WTO obligations imply free movement of imported automobiles (free of import quotas) 4. Unofficial estimates (based on interviews) are that 1 00,000 or more sedans have been imported in recent years. Many smuggled cars are luxury models. 5. In the past, finding partners often meant having to go to provinces other than those on the coast. These provinces often tried to ensure that '"huy local" conditions prevailed. In the case of taxis in Shanghai, regulations stipulated specifications that could be filled only by a Volkswagen model. 92 THE WORLD BANK ECONOMIC REVIEW, VOL. i 8, NO. I within the Chinese market. The internal barriers to trade simply cannot be sus- tained if China's new wro obligations are to be taken seriously. All these changes in the structure of protection imply tremendous pressure for a breakdown of internal barriers for domestic production and for rationalization of the domestic industry. The government realizes the implications for the structure of the automotive sector. Official and industry sources indicate an intention to support only a small number of domestic production groups, perhaps including the Shanghai group (Volkswagen), China First Auto Works (Volkswagen), Shanghai GM (Buick), and the Dongfeng Group (Citroen). These groups with their foreign partners already account for more than 70 percent of production in China. Such a sharp rationalization will undoubtedly be painful, but it could allow the industry to consolidate production and work its way down the average cost curve for vehicle production. II. THE MODELING FRAMEWORK A CGE model is used to assess the possible impact of China's accession to the w-ro. (More technical details and references for the model are provided in Francois and Spinanger 2001 and in the technical annex available for download with the model files.6) CGE models have become a standard approach for analysis of multisector policy initiatives, such as the accompanying w-ro acces- sion (Francois 2000). Although the exercises are hampered by both the neces- sary assumptions and the quality of available data, their estimates of the direct and indirect impact of broad policy changes have proved helpful for assessing existing economic policies and formulating new ones. The Model Data The data come from a number of sources. They are organized into 23 sectors and 25 regions (table 6). Details on the value-added chain linking fibers into textiles and clothing production are included to better capture the initial impact on the base scenario of the Agreement on Textiles and Clothing (ATC), which is scheduled to phase out the remaining textile and clothing quotas by 2005. Data on production and trade are based on national accounting data linked through trade flows and drawn directly from the Global Trade Analysis Project (GTAP) version 5 dataset (McDougall 2001). The dataset is benchmarked to 1997 and includes detailed data on national input-output, trade, and final demand structures. The basic database was updated to better reflect actual import protection for goods and services. Basic data on current tariff rates come from UN Conference on Trade and Development and WTO data on the schedules of applied and bound tariff rates. 6. The model files, along with the technical annex describing the model, can be downloaded from www.intereconomics.com/francois. The model is implemented in GEMNPACK. Francois and Spinanger 93 TABLE 6. The Regional and Sectoral Breakdown of the CGE Model Region Sector Primary Hong Kong, China Wool People's Republic of China Natural fibers Taiwan, China Primary food production Japan Other primary production Korea, Rep. of Sugar ASEAN5' Processed food, tobacco, and beverages Vietnam Manufacturing India Textiles Bangladesh Wearing apparel Other South Asian economiesh Leather products Australia Chemicals, refinery products, rubber, plastics New Zealand Steel refinery products Canada Nonferrous metal products United States Motor vehicles and parts Mexico Electronic machinery and equipment Brazil Other machinery and equipment MERCOSURC Other manufactured goods Caribbean Basin Initiative economiesd Services Andean Trade Pact economiesd Wholesale and retail trade services Chiled Transportation services (land, water, air) Other Latin Americad Communications services European Union, 15 economies Construction Turkey Finance, insurance, and real estate services Africa and the Middle East Other commercial services Rest of world Other services aIndonesia, Malaysia, Philippines, Singapore, and Thailand. bNepal, Pakistan, and Sri Lanka. cIncludes Argentina, Paraguay, and Uruguay. Brazil is represented separately. dNot treated in tables and diagrams. Source: Database aggregation produced by authors. These are integrated into the core GTAP database. They are supplemented with data from the Office of the U.S. Trade Representative and the U.S. International Trade Commission on regional preference schemes in the Western Hemisphere. Data on agricultural protection, as integrated into the GTAP core database, are based on estimates by the Organisation for Economic Co-operation and Devel- opment and U.S. Department of Agriculture. Estimates on tariffs and nontariff barriers are further adjusted to reflect remaining Uruguay Round commitments, including the phase-out of textile and clothing quotas under the ATC. Data on post-Uruguay Round tariffs are from recent estimates reported by Francois and Strutt (1999), which come primarily from the WTO'S integrated database, with supplemental information from the World Bank's recent assessment of detailed pre- and post-Uruguay Round tariff schedules. All this tariff information has been matched to the current model sectors. Services trade barriers are based on 94 THE WORLD BANK ECONOMIC REVIEW, VOL. 1 8, NO. I the estimates described in the technical annex and are shown in table 7 (the basic GTAP database includes no information on trade barriers for services, for example). The basic GTAP dataset is benchmarked to 1997 and reflects applied tariffs in place in 1997. Because the interest here is with the post-Uruguay Round world, a "pre-experiment" was run on the model to implement the remaining Uruguay Round tariff cuts. Most of these cuts were already in place in the 1997 benchmark TABLE 7. China's Tariff Rates before and after World Trade Organization Accession, as Modeled (percent) Sector Model base rates Accession rates New bound rates Merchandise Wool 14.8 42.0 38.0 Natural fibers 3.1 17.4 13.6 Primary food production 58.8 58.1 46.8 Other primary production 0.5 6.9 5.0 Sugar 29.5 30.0 20.0 Processed food, tobacco, 37.7 40.7 23.2 and beverages Textiles 25.1 25.4 10.2 Wearing apparel 31.8 32.8 16.1 Leather products 12.1 20.9 17.0 Chemicals, refinery products, 12.6 14.9 7.2 rubber, plastics Steel refinery products 9.7 8.9 5.1 Nonferrous metal products 7.8 8.2 5.5 Motor vehicles and parts 34.4 38.7 15.4 Motor vehicles 70.5 70.5 25.0 Parts 23.4 23.4 10.0 Electronic machinery and equipment 11.9 16.9 9.6 Other machinery and equipment 12.8 15.4 10.1 Other manufactured goods 14.5 22.0 16.3 Services Wholesale and retail trade services 0.0 NA 0.0 Transportation services 4.0 NA 2.0 (land, water, air) Communications services 9.2 NA 4.6 Construction 13.7 NA 6.8 Finance, insurance, and real 8.1 NA 4.0 estate services Other commercial services 48.0 NA 24.0 Other services 25.7 NA 13.0 NA, not applicable. Note: Service barriers are based on gravity equation estimates. Accession rates reflect an assumed 50 percent drop in cross-border trading costs. Source: China wTo accession schedule, GTAP data, and Office of the U.S. Trade Representative. Gravity estimates are based on trade and macroeconomic data and cross-country regressions; see Francois and Spinanger 2001. Francois and Spinanger 95 dataset. The data were also adjusted to reflect regional preference schemes in Latin America (not represented in the core GTAP database). The dataset used for actual policy experiments is therefore a representation of a notional world economy (with values in 1997 dollars) with full Uruguay Round tariff cut implementation. Experiments consider both the ATC phase-out and China's wTo accession, with reference to this post-Uruguay Round tariff benchmark. Model Structure Except for the automobile sector, the CGE model structure is standard. On the production side, firms in all sectors minimize costs, employing domestic factors of production (capital, labor, and land) and intermediate inputs from domestic and foreign sources to produce goods and services. Technologies are modeled as constant elasticity of substitution processes defined over primary inputs and Leontief processes defined over intermediate inputs. Products from different regions are assumed to be imperfect substitutes in accordance with the Armington assumption. Prices on goods and factors adjust until all markets are simul- taneously in (general) equilibrium-all markets clear. Although changes are modeled in gross trade flows, changes in net international capital flows are not (this does not preclude changes in the level of gross capital flows). Trade liberalization in the goods sectors involves reduction of tariffs and a shift from model base rates to the new bound rates. The new bound rates are generally quite close to the calculations of average accession rates. Liberalization in the service sector is modeled as a reduction in trading costs, reflecting the barrier reductions in barriers reported in table 7. These are Samuelson iceberg costs. To reflect the status quo in the motor vehicle sector in a stylized, though representative way, one option was to implement imperfect competition in the model. This was rejected, however, because it does not adequately reflect the primary issue at hand. Government policy has certainly resulted in market segmentation, but there is also price setting and regulation. The choice was made to focus on realized cost efficiency for the sector. The cost structure of the industry reflects the net effect of a basket of policies. Like clothing in India or automobiles in Mexico before the North American Free Trade Agreement, the structure of the automobile sector in China reflects regulated efficiency-the impact of the general regulatory and administrative environment. The critical issue is thus these collective inefficiencies, which follow from the full set of industrial policies. At the same time, an implication of intended public policy seems to be restructuring and consolidation, leading to an improvement in regulated efficiency. What shape will the gains from changes in regulated efficiency take? Through rationalization, the industry may collectively move down relevant cost curves. Although minimum efficient scale for some models is about 200,000 units per plant (Huang 2002), a global norm, based on comparisons with plant sizes in North America and Europe, is closer to 350,000 units. Further comparison of current plant scale in China (see table 3) with such a norm implies that average 96 THE WORLD BANK ECONOMIC REVIEW, VOL, I8, NO. I costs are roughly 20 percent higher simply because of inefficient scale.7 Data from interviews with industry representatives (Feenstra and others 2001) point to similar cost savings, with expectations of even higher cost reductions in the range of 25-30 percent. A World Bank study (1993, p. 57) describes quite succinctly the expected gains from reaching minimum efficient scale (MES): "If this cost-volume relationship is applied to the Chinese automotive industry, the passenger car segment has a cost disadvantage of 20 to 30 percent compared with the international producers having MES. This cost disadvantage could be an understatement, however, as there are already eight producers in the market." This net cost effect is stressed here and sets the treatment of motor vehicles apart from that of other sectors in the model. The lower bound of these cost effect estimates is used. In particular, the focus is on potential cost savings in the final assembly of automobiles due to a higher regulated efficiency level for the industry, resulting from consolidation and rationalization of policy.8 In addition, the differ- ential treatment of parts and finished vehicles in the tariff schedule is also tracked. That large gains can be achieved by rationalizing production and reducing costs was clearly demonstrated in the United States at the beginning of the twentieth century (figure 2). In 1914, "13,000 workers at Ford were producing 260,720 cars. By comparison, in the rest of the industry, it took 66,350 workers to make 286,770" cars.9 Such dichotomies also exist across the spectrum of production possibilities in China today, with new foreign-built modern plants coexisting with Mao-era facilities. Similar demand factors also prevailed. As a result of Ford's new production methods, cars in the United States moved from being scarce goods to goods affordable by large segments of the population. China is already moving into this phase. The similarity between China's motor vehicle production from 1984 to 2002 and that of the United States between 1900 and 1924, as shown in figure 2, would seem to justify such an analogy. III. EXPERIMENTS AND RESULTS The experiments involve full accession for China and Taiwan, China. The basic accession package involves the changes in tariffs detailed in table 7. For auto- mobiles the following effects are modeled: * Tariffs on motor vehicles decrease to 25 percent. * Tariffs on automobile parts are phased down from an average of 23.4 percent to an average of 10 percent. 7. The 20 percent figure is based on the distribution of current plants shown in table 3. An average cost index for the industry can be calculated by applying the formula Aln(Average Cost) = CDR . Aln(Quiantity), where CDR is the inverse elasticity of scale, defined as CDR=(Average Cost- Marginal Cost)lAuerage Cost, and is between 0.125 and 0.135 (the range of values found in engineering stLdies). If the index is 100 at 350,000 units per plant, current plant structure yields a cost index of roughly 120. 8. In other words, cost savings are modeled at the assembly level. 9. See www.wiley.com/products/subject/business/forbes/ford.html. Francois and Spinanger 97 FIGURE 2. Annual Motor Vehicle Production in China and the United States (thousands) 4,500 - -- - 3,500 4,000 300- 3,500 -J 3,0001 , 2,500 2,000 - 2,000-~~~~~~~~~~ ~1,500 - 1.5001 1,000 -00 5001 500-5 0 _ _ -L _ _ _ _ _ _ _ _ _ _ _ 1901 1905 1909 1913 1917 1921 1925 1984 1988 1992 1996 2000 United States 1901-1925 China 1984-2002 Source: Bessum (2002) and Verband der Automobilindustrie (various issues). * The industry is rationalized. Implicitly, this involves elimination of internal regional barriers and consolidation within the domestic market. Small, inefficient factories close. To quantify this effect, sedan production is taken as representative. Given the typical scale of domestic production, automobile plants are assumed to realize a 20 percent cost savings in assembly if they move to efficient scale. (See note 2 and the discussion in section II). This savings is modeled at the assembly level. The overall sectoral impacts of the experiments are presented in table 8, which reports changes in the quantity of output under alternative scenarios. Extending the ATC phase-out to China and Taiwan implies a dramatic expansion in the textile and clothing sectors, with textiles growing 14 percent and clothing 50 percent. There are important general equilibrium effects, as the resources needed for this experiment are drawn from other parts of the economy, includ- ing the motor vehicle sector. Especially important for the motor vehicle sector are the results reflecting the incremental impact of China's market access commitments made as part of wfo accession and shown in columns B and C of table 8. Column B is a business as usual scenario, without restructuring. It reflects a domestic motor vehicle 98 THE WORLD BANK ECONONII(: REVIFW, VOL. I 8, NO. I TABLE 8. Impact on Output of World Trade Organization Accession by China and Taiwan, China (percentage change) Accession Total impact Total Elimination without Accession without impact with of textile automobile with automobile automobile automobile and apparel sector sector sector sector quotas restructuring restructuring restructuring restructuring Sector A B C D=A+B E=A+C Wool 12.8 18.3 16.8 33.4 31.8 Other natural fibers 12.1 17.9 16.4 32.1 30.5 Primary food -0.4 -1.0 -0.9 --1.5 -1.3 Other primary -2.6 -3.6 -3.3 -6.1 -5.8 production Sugar -2.3 -7.9 -8.5 ---10.0 -10.6 Processed foods -1.0 -4.7 -4.7 -5.6 -5.7 Textiles 13.9 32.0 30.6 50.4 48.8 Clothing 50.3 75.5 73.0 163.7 160.0 Leather goods -7.2 5.4 3.5 -2.2 -3.9 Chemicals, rubber, -2.0 -4.5 -4.3 -6.5 -6.2 and refineries Primary steel -4.0 -9.1 -7.9 -12.8 -11.5 Primary nonferrous -5.4 -9.2 -8.9 -14.2 -13.9 metals Motor vehicles -4.1 -36.7 8.0 -39.3 3.5 and parts Electronics -5.1 -3.9 -4.4 -8.8 -9.3 Other machinery -3.8 -5.4 -4.8 -9.0 -8.5 and equipment Other manufactures -2.2 -0.3 0.1 -2.5 -2.0 Wholesale and -0.3 1.4 1.9 1.1 1.7 retail trade Transport services -1.9 -2.0 -1.4 -3.9 -3.3 Communications -0.5 0.1 1.0 -0.5 0.5 Construction 0.8 2.8 4.2 3.6 5.0 Finance, insurance, -0.7 -0.4 0.2 -1.1 -0.4 and real estate Commercial services -0.8 -5.9 -5.4 -6.6 ---6.2 Other services 0.0 0.5 1.2 0.5 1.2 Source: Model estimates; see table 7. industry that continues to be fragmented, with favored producers in each region, small production runs, and high costs. Such an industry is simply unable to compete with imports. It is hit very hard by imports, with domestic production falling 37 percent. Combined with the initial impact of the ATC phase-out, there is a dramatic retrenchment of the uncompetitive domestic industry in the face of imports (column D). By contrast, the scenario with elimination of internal barriers, rationalization of plants (with smaller plants being closed), and an efficiency gain of roughly Francois and /' n.i:. 99 20 percent as scale economies are realized, production rises slightly (3 percent) and the industry emerges as relatively competitive, despite the loss of protection (columns C and E). The most striking difference between the two scenarios is in the impacts on intermediate parts production and final automobile production (table 9). Under the scenario of business as usual, imports of parts rise slightly, whereas their share of the domestic market rises substantially. There is a dramatic surge in imports of motor vehicles, which displace more than a third of domestic production. There is a drop in the overall market for parts because of the decline in domestic vehicle production. Under the second scenario of rational- ization of the final assembly sector, which allows the sector to compete more directly with imports, there is a shift to imported intermediates (rising to a market share of more than 50 percent), a fall in domestic parts production (displaced by imports), and steady overall demand for parts. Although ground is lost to parts imports, sales of domestic vehicles remain steady in the face of imports. China's wTo accession also affects value added and trade. It is logical to expect some export response, both because of the general liberalization in trade TABLE 9. Impact of World Trade Organization Accession by China and Taiwan, China, on China's Motor Vehicle Market Accession without Accession with automobile automobile Benchmark sector sector Item 1997 restructuring restructuring Value (millions of 1997 U.S. dollars) Imported motor vehicles 3,607.7 10,595.7 6,968.0 and parts, world prices Imported motor vehicles 4,806.4 12,080.7 7,995.7 and parts, internal prices Imported parts 1,609.9 2,827.9 5,535.2 Imported motor vehicles 3,196.5 9,252.8 2,460.5 Domestic automobiles, 32,812.5 19,401.9 24,249.6 intermediates, and parts Intermediates and parts 10,896.2 4,494.0 5,189.1 Industrv consumption of motor 21,625.5 14,698.8 18,785.0 v ehicles Final consumption of motor vehicles 290.8 209.2 275.4 Index and share Import share of total 12.9 38.6 51.6 automobile parts (percent of value) Index of vehicle production 100.00 68.0 102.8 Index of parts production 100.00 41.2 56.3 Source: McDougall 2001 (haseline) and authors' model estimates (impact). 100 THE WORLD BANK ECONOMIC RFVIEW, VOl.. I 8, No. I T A B L E 1 0. China's Export Shares, Baseline and Two Scenarios Total impact Total impact without with automobile automobile Baseline sector sector Sector 1997 restructuring restructuring Primary 0.046 0.033 0.033 Textiles 0.084 0.098 0.097 Wearing apparel 0.102 0.303 0.298 Motor vehicles and parts 0.006 0.004 0.019 Electronic machinery and equipment 0.133 0.100 0.099 Other machinery and equipment 0.146 0.104 0.103 Other manufactured goods 0.397 0.294 0.290 Services 0.087 0.062 0.062 Source: McDougall 2001 (baseline) and authors' model estimates (impact). and because pressure from imports may force firms to seek other markets. China exports less that 4 percent ($1.3 billion of production of $32 billion) of its production in the sector based on 1997 values. To put this in perspective, Australia has a comparable level of exports with an industry only one-third the size of China's. The Republic of Korea's export share is 10 times as large. China's trade is therefore well below global integration standards, measured by exports. The model experiments show that restructuring accelerates the export orien- tation of the automobile industry, with a rapid growth in exports (table 10). Exports rise by roughly $3.8 billion, or 300 percent, reaching roughly 10 percent of production by value. Although this seems dramatic, it needs to be kept in perspective. Automobiles and parts are a small share of exports (0.6 percent in 1997) and remain small (up to 2 percent) even with the growth in automobile exports. Most of the restructuring remains focused on the domestic market. IV. SUMMARY AND CONCLUSIONS Regulatory reform and internal restructuring are critical to the impact of w-ro accession on China's motor vehicle sector. Such restructuring is represented here by a cost reduction following from consolidation and rationalization. This representation is supported by a comparison of scale in a typical automobile plant in China to that in typical plants in North America or Europe and also by firm survey responses. It is also supported by earlier estimates of the benefits from achieving minimum and efficient scale and radical restructuring to improve production efficiency. The net result is a movement of costs toward global norms. With restructuring, the final assembly industry can become competitive by world standards, while the parts industry further integrates Francois and Spinanger 101 with the global industry through exports (and through higher imports). Without such restructuring, however, the domestic industry remains uncompetitive, and wTo accession means that imports of final vehicles will surge, though imports of parts will fall as production moves offshore. Viewed in total, what do the results show? They highlight the importance of incorporating the impact of regulatory regimes on costs when assessing the impact of changes in trade policy. For China, restructuring within the domestic market results in a qualitatively different impact from tariff reductions. Without such restructuring, the industry fails to compete and contracts dramatically. However, with restructuring, the final assembly industry can be made inter- nationally competitive. In addition, the industry shifts to local assembly, with high import content for domestic vehicles. Two additional issues need to be raised. The first concerns China's popula- tion to motor vehicles ratio, which is far higher than that in many other countries with similar income levels (see table 1). Because this reflects the impact of existing policies, a significant change in policies could shift demand closer to a normal pattern of consumption, given China's geographic attributes. The second issue concerns further strengthening of demand for cars through better access to financing. Roughly 75 percent of U.S. and European automobile purchases are financed through loans, but only 15 percent of automobile purchases in China are financed this way. Although China's protocol of acces- sion to the WTO stipulates that automobile finance will be liberalized, only draft legislation has been presented to date.'0 To the extent that this potential demand can be tapped, the pressure on firms to be more productive and thus more competitive will be all the greater. This would be another factor helping ensure that the calculated welfare gains will come about. The shortcomings of the analysis also need to be highlighted. The model applied here is very stylized, although it widely captures the real world. Although restructuring has positive overall implications for the industry, there are bound to be adjustment costs that are not pointed to in the model. Even if value added is preserved within the sector, there will likely be a dramatic relocation of jobs toward a limited number of plants, with job losses in the other, smaller plants. The current regional scattering of final automobile pro- duction (table 11) will be replaced by a more geographically concentrated pattern. Parts production will also tend to concentrate. To the extent that parts suppliers are able to supply regional markets, this is likely to mean that existing clustering in coastal regions will intensify, with parts shipments to Japan, the Republic of Korea, the United States, and other regional centers of 10. Nonetheless, some major car companies (Volkswagen and Ford) did reach agreements with Chinese banks earlier this year (KP.xz; 2003, p. 7). According to the International Herald Tribune (October 6, 2003) China has opened up this sector in line with its wro commitments. 102 THE WORLD BANK ECONOMIC REVIEW, VOL. 1 8, NO. I TABLE 1 1. Location of Automobile Production in China, 2002 Foreign production Producer Foreign producer Capacity (cars/year) Production 1 SAIC VW Volkswagen 450,000 278,890 2 SAIC GM General Motors 100,000 111,623 3 FAW VW Volkswagen 270,000 158,654 4 FAW Toyota Toyota/Mazda 70,000 30,165 5 Dongfeng PSA PSA/Citroen 150,000 84,378 6 Dongfeng Honda Honda 60,000 59,024 7 Dongfeng Yulong NissanNYulong 60,000 38,897 8 Tianjing Toyota Toyota 30,000 2,147 9 Jlangsu Nanya Fiat 100,000 23,393 10 SAIC Chery Daewoo 60,000 49,397 11 Zehjiang Jili Daewoo (geplant) 150,000 47,443 12 Chongqing Chang'an Suzuki Suzuki/Yanjin 150,000 67,846 13 Chang'an Ford Ford 50,000 NA 14 Dengfeng Yueda Kia Kia 50,000 20,080 15 FAW Hainan Mazda 50,000 11,989 16 Beijing Hyundai Hyundai 30,000 1,356 17 China Guizhou Aviation Ind. Wanhong/Chenchang 10,000 1,831 18 Shenyang Brilliant Junbei BMW (by mid-2003) 200,000 NA 19 Harbin Hafei Mitsubishi 30,000 14,577 20 Shangdong Yantei General Motors 50,000 NA 21 Southeast Zhonghua 60,000 16,935 22 Beijing Jeep Daimler-Chrysler 85,000 9,052 23 Jinbei General Motors General Motors 30,000 3,751 24 Hunan Changfeng Mitsubishi 30,000 15,067 25 Zhengzhou Nissan Nissan 30,000 NA 26 Rongcheng Huatai Hyundai 20,000 NA 27 Jiangxi Fuqi Golden Lion 20,000 NA 28 Tianjing Huali Golden Lion 20,000 NA 29 SAIC GM Wuling General Motors 150,000 NA 30 Sanjiang Renault Renault 30,000 NA 31 Chengdu FAW Toyota 5,000 NA 32 Yizhong SAIC/RDS 10,000 NA (Continued) production.1' From an employment perspective, output and value added results closely track the impact on employment. The results point to a range of effects on employment, from -40 percent without restructuring to -3 percent with restructuring. This range of effects highlights the importance of rationalizing the structure of plants. 11. European manufacturers have already established 12 plants in China, and one large U.S. company (Delphi) is shifting from Mexico. Francois and Spinanger 103 TABLE 1 1. Continued Production capacity in provinces Province Capacity (cars/year) Production Anhui 60,000 49,397 Beiing 115,000 10,408 Fujian 80,000 16,935 Guandong 120,000 97,921 Guangxi Zhuang 150,000 NA Guizhou 10,000 1,831 Hainan 50,000 11,989 Heilongiang 30,000 14,577 Henan 30,000 NA Hubei 180,000 84,378 Jiangsu 130,000 38,460 Jilin 340,000 188,819 Liaoming 230,000 3,751 Shandong 80,000 NA Shanghai 550,000 390,513 Shanxi 50,000 20,080 Sichuan 205,000 67,846 Tianjing 50,000 2,147 Zehjiang 150,000 47,443 Total 2,380,000 1,046,495 Other foreign companies Number of employees Number of plants Bosch 3,600 6 Kolbenschmidt 1,500 2 Michelin 4,000 2 ZF/Sachs 2,100 2 Total 11,200 12 NA, actual production was not yet available. Source: Bessum 2002; Chinese Motor Vehicle Documentation Center 2002. REFERENCES Bessum, F. 2002. "Global Car Production Statistics." Online document available at www.geocities. com/motorcity/speedway/4939/carprod.html. China Online. 2000. "How wro Membership Could Affect China's Auto Industry." January 5, 2000. Available at www.chinaonline.com. Chinese Motor Vehicle Documentation Center. 2002. Catalogue of the Present Chinese Motor Car Production, 2nd ed. Aldeboarn, Netherlands. Feenstra, R., D. Sperling, L. Branstetter, E. Harwitt, and W. Hai. 2001. "China's Entry to the wro: A View from the Auto Industry." Working Paper, University of California, Davis.. Francois, J. F. 2000. "Assessing the Results of General Equilibrium Studies of Multilateral Trade Negotiations." Policy Issues in International Trade and Commodities Study 3. UN Conference on Trade and Development. Available online at http://192.91.247.38/tab/pubs/itcdtab4_en.pdf. 104 THE WORLD BANK ECONONfIC REVIEW, VOL. I 8, NO. I Francois, J. F., and D. Spinanger. 2001. "Greater China's Accession to the WTO: Implications for International Trade/Production and for Hong Kong." Paper prepared for the Hong Kong Trade Development Council. Francois, J. F., and A. Strutt. 1999. "Post Uruguay Round Tariff Vectors for GT-AP Version 4." Working Paper, Erasmus University, Rotterdam, Netherlands. Huang, Y. 2002. "Between Two Coordination Series: Automotive Industrial Policy in China with a Comparison to Korea." Revieu' of International Political Economy 9 (3):538-73. KPNIG. 2003. "China Automotive and Component Parts Market." Online document available at www.kpmg.com.cnVPress'/.20releases/KPMG China_autojindustry-reportO3.pdf. Lopez de Silanes, F., J. Markusen, and T. F. Rutherford. 1994. "Complementarity and Increasing Returns in Intermediate Inputs." Journal of Development Economics 45:133-51. McDougall, R., ed. 2001. "The GTAP Database Version 5." Working Paper, Purdue University, Global Trade Analysis Center, West Lafayette, Ind. Verband der Automobilindustrie. Various issues. Tatsachen und Zablen. Frankfurt, Germany. World Bank. 1993. "China Industrial Organization and Efficiency Case Study: The Automotive Sector." 12134-CHA. Washington, D.C. World Bank. Various years. World Development Indicators. Washington, D.C. Zhang, W., and R. Taylor. 2001. "EU Technology Transfer to China: The Automobile Industry as a Case Study." Journal of Asia Pacific Economy 6(2):261-74. China in the World Trade Organization: Antidumping and Safeguards Patrick A. Messerlin China finds itself in a unique situation on antidumping and safeguard issues. It is by far the main target of antidumping measures, but (so far) one of the smallest users of such measures. China's World Trade Organization (WTO) accession protocol includes stringent antidumping and safeguard provisions that its trading partners may use against its exports. The article examines three related concerns: how quickly large developing economies can become intensive users of antidumping measures, an evolution raising concerns about China's recent antidumping enforcement, how China could minimize its exposure to foreign antidumping cases, a recipe for both improving trade outcomes and for China's taking a leading role in reforming WTO antidumping; and the opportunities that the Doha Round of trade negotiations offer to China for negotiating stricter disciplines both on WTO contingent protection and on the use by China's trading partners of the special provisions included in China's accession protocol. On November 10, 2001, China was accepted as a full member in the World Trade Organization (wro). A few weeks earlier, China's chief trade negotia- tor, Long Yongtu, had put "stricter rules on antidumping" second among China's priorities in the WTO. At that time the United States was still fighting to exclude antidumping from the topics to be discussed at the WTO Doha Ministerial Meeting and the European Union was adopting an ambiguous position. In the early stages of the negotiations under the Doha Round development agenda, China finds itself in a unique situation on antidumping and safeguard issues. China's WTO accession protocol includes special provisions on antidumping and safeguards that its trading partners may use against Chinese exports. These include continuing use of "nonmarket economy" status in antidumping investiga- tions for 15 years and use of a special "transitional product-specific safeguard" Patrick A. Messerlin is professor of economics at Sciences Po and director at Groupe d'Economie Mondiale de Sciences Po; his e-mail address is patrick.messerlin@sciences-po.fr. The author would like to thank Nlichael Finger, Will Martin, Edwin Vermulst, and two anonymous referees for useful comments on earlier drafts of this article. THE WORLD BANK ECONOMIC REVIEW, VOL 18, NO I. © The International Bank for Reconstruction and Development / THE WORLD BANK 2004; all rights reserved. DOI: 10.1093/wber/lhh032 18:105-130 105 106 THE WORLD BANK ECONOMIC REVIEW, VOL. i8, NO. I provision for 12 years.' China is by far the main target of antidumping measures even though it is one of the smallest users of such measures. But the past decade has shown how quickly large developing economies can become intensive users of this instrument, and the evolution of China's antidumping enforcement in 2002 and early 2003 raises concerns in this respect. Section I of this article describes the current situation with respect to anti- dumping. It is used massively by only 10 countries (4 industrial and 6 develop- ing), and there is strong asymmetry, best illustrated by China, between countries enforcing antidumping and those targeted by antidumping measures. Section II examines how China could minimize its exposure to foreign antidumping cases-an option that would be a recipe for both improving trade outcomes and for China taking a leading role in reforming wro antidumping. Section III analyzes China's antidumping regulations and its first cases, including their crucial relationship with the existing web of the U.S. and EU antidumping cases. Section IV examines the opportunities the Doha Round offers to China for negotiating stricter disciplines both on wTO contingent protection and on the use of the nonmarket economy and transitional product-specific safeguard provisions by China's trading partners. The conclusion summarizes the crucial choices to be made by China in antidumping and safeguard policy. I. THE CURRENT SITUATION During the November 2001 WTO Doha Ministerial Meeting, antidumping was perceived as an issue pitting developing economies, anxious to discipline the use of this instrument, against the United States, which was (and still is) reluctant to change its own antidumping regulations. However, a much more complex picture emerges from a close examination of the antidumping measures in force at the end of each year during 1995-2002, which are notified to the WTO Secretariat by wTo members.2 Antidumping Users and Targeted Countries: A Key Asymmetry Examination of the stock of antidumping measures in force shows two main results (table 1). First, the top 10 antidumping users enforce 90 percent of the antidumping measures notified to the wro, whereas they represent 70 percent of world gross 1. In spring 2002 the European Union and the United States declared that they would consider Russia a market economy for purposes of antidumping. However, the case for introduction of a transitional product-specific safeguard provision in Russia's accession protocol seems open. It would be interesting to make a parallel between the conditions imposed on Japan's accession to the General Agreement on Tariffs and Trade (iAlT) and those imposed on China in its accession to the wTo. 2. Tables 1-3 treat measures taken against individual EU member states as one aggregated measure if adopted at the same time and for the same product (data for 200)2 are still not complete). Table 4 follows the notifications of EU trading partners, which vary in their treatment of the European Union (as one entity or as a set of distinct member states). Messerlin 107 TABLE 1. Top 10 Antidumping Users, 1995-2001 (number of measures in force) Average number Average Country by value applied or group 1995 1996 1997 1998 1999 2000 2001 2002 of imports tarrif Industrial countries Australia 78 46 40 49 39 44 59 38 0.77 5.8 Canada 79 78 78 65 72 71 85 83 0.38 4.8 European Union 140 138 138. 139 159 175 175 183 0.19 4.6 United States 265 271 271 281 282 202 227 239 0.29 4.3 Developing economies Argentina 15 30 33 39 45 42 45 60 1.17 13.7 Brazil 21 28 23 28 38 43 49 54 0.51 12.5 India 13 15 24 44 58 94 115 181 1.28 39.6 Mexico 93 92 81 86 80 77 61 55 0.72 12.6 South Africa 12 29 42 56 87 96 94 80 1.81 15.0 Turkey 37 37 34 34 35 14 16 30 0.61 12.7 All other 50 59 84 102 122 117 97 - 0.04 - China - - 0 0 4 8 11 17 0.03 15.8 All countries 803 823 848 923 1,017 975 1,023 - 0.21 - Note: Measures in force include antidumping duties and undertakings in force as of December 31 of the year. -, Not available. aAverage number per $1,000 of 1997 imports of the user country. Source: Author's computations based on WTO Reports on Antidumping (G/ADP/N series at www.wto.org), WTO trade data, and wTo (2001). domestic product (GDP) and 50 percent of the world trade. Worldwide antidumping enforcement is thus highly concentrated in fewer than a dozen countries. Second, the situation prevailing during the Uruguay Round-antidumping users were almost exclusively industrial countries-is no longer true. Six new intensive antidumping users are developing economies (Argentina, Brazil, India, Mexico, South Africa, and Turkey), and they have almost caught up with the four major traditional users. These new users implemented more than a third of the antidumping measures in force in 2002, up from less than a fourth in 1995. Meanwhile, the share of measures of the four traditional users declined from more than two-thirds to half the total number of antidumping measures in force during the period. In another worrisome sign of increasing use of antidumping measures, the remaining developing economies, though still small users individ- ually, together doubled their global share of measures in force during the observation period. Examination of the stock of antidumping measures in force by targeted country shows a marked asymmetry between antidumping users and targets (tables 1 and 2). The top 10 users are the targets of less than a third of all the measures in force, and the gap between using antidumping measures and being a 108 THE WORLD BANK ECONOMIC REVIEW, VOL. i8, NO. i TABLE 2. Top 10 Antidumping Targets, 1995-2001 (number of measures in force) Average number Country by value or group 1995 1996 1997 1998 1999 2000 2001 2002 of exportsa Indutstrial cotuntries Australia 5 6 5 6 5 5 5 5 0.08 Canada 19 19 19 19 20 18 8 8 0.08 European Union 77 88 89 102 132 149 99 98 0.13 United States 60 66 66 68 66 62 57 67h 0.09 Developing economies Argentina 9 8 7 7 7 7 9 8 0.30 Brazil 48 51 52 45 42 43 34 51 0.85 India 15 15 15 21 29 35 42 44 0.72 Mexico 11 15 17 17 19 21 17 17 0.15 South Africa 7 10 11 11 12 15 16 24 0.39 Turkey 9 9 6 8 10 13 12 18 0.37 All other 543 536 561 619 675 611 724 - 0.29 China 143 148 180 193 202 207 199 212 0.99 All countries 803 823 848 923 1,017 979 1,023 - 0.22 Note: Measures in force include antidumping duties and undertakings in force as of December 31 of this year. -, Not available. aAverage number per $1,000 of 1997 exports of targeted country. bIncomplete estimate. Source: Author's computations based on WTO Reports on Antidumping (G/ADP/N series at www.wto.org) and wTo trade data. target widened in 2001 and 2002. In sum, antidumping is currently an instru- ment enforced by a few large countries against the smaller economies of the rest of the world. Thus there is little pressure coming from the rest of the world to urge intensive antidumping users to restrain their actions. For the top 10 antidumping users, with the exception of Brazil, the domestic interests that are hurt by foreign antidumping measures are smaller than the interests that benefit from antidumping protection. This reflects the well-known economic proposition that views protection more as a conflict between domestic export interests and import-competing interests than as a conflict between countries. To capture this aspect, the number of foreign antidumping measures in force against the exports of a top user can be adjusted by the size of the country's exports (in thousands of U.S. dollars; see table 2). These trade- adjusted measures mirror the intensity of foreign pressures imposed on the export interests of a country, thus giving an indication of the incentives of these export interests to contribute to the opening of domestic markets. These numbers can then be compared with the trade-adjusted antidumping measures in force by the country in question (see table 1), which can be interpreted as an Messerlin 109 indication of the strength of the incentives of import-competing interests to induce their government to use antidumping. The observed imbalance between export interests and import-competing antidumping beneficiaries in the top 10 antidumping users suggests that it is unlikely that domestic coalitions in these key users, which are also key WTO players, are strong enough to support anti- dumping reforms in the WTO. This situation raises a question that needs to be carefully examined in the future. The data on antidumping measures in force suggest that antidumping measures by the six major developing economy antidumping users impose welfare costs on their own domestic economies that are higher than the costs imposed on industrial economies by their antidumping measures, for two rea- sons. First is the marked difference between the number of measures imposed by developing and industrial countries once adjusted for trade size. The average number of measures in force per $1,000 of goods imported (in 1997) by an antidumping user is a better indicator of the potential harm done by antidump- ing to the domestic economy than the absolute number of measures. This indicator is much higher for developing economies than for industrial countries, ranging from 0.5 for Brazil to 1.8 for South Africa and from 0.2 for the European Union to 0.4 for Canada (with an exception, Australia, at 0.8). These differences would be even larger if the number of antidumping measures were adjusted for the number of tariff lines concerned because developing economies tend to cover many more tariff items with antidumping cases than do industrial countries. The second reason for higher welfare costs is that available information (though not systematic) suggests that antidumping duties enforced by developing countries are, on average, more severe than those imposed by industrial countries-and economic analysis shows that welfare costs increase more rapidly than tariffs do. China's Special Situation China has been the main target of antidumping measures-18 percent of anti- dumping cases in 1995 and almost 20 percent in 2001 and 2002 (see table 2). Examination of the raw number of antidumping measures imposed on Chinese exports by the top 10 antidumping users and the number of measures adjusted for trade value between each trade partner and China (the average number of cases per $100,000 of exports from China to these users) shows that China is targeted much more by developing economies than by industrial countries (table 3). China is almost exclusively targeted by the top antidumping users, all of them being relatively large economies. All this raises a key question about China's role in future antidumping activities. Will China follow the same path as other large developing economies, rapidly increasing the number of antidumping cases against other countries? Or will China adopt a different approach, minimizing the use of antidumping 1 10 THE WORLD BANK ECONOMIC REVIEW, VOL. I 8, NO. I TABLE 3. Share of Antidumping Measures in Force against Imports from China, 1995-2002 (% of total antidumping measures in force by user country) Average number Average by value Country share of imports or group 1995 1996 1997 1998 1999 2000 2001 2002 1995-2001 from Chinaa Industrial countries Australia 9.0 4.3 10.0 8.2 7.7 4.5 5.1 7.9 7.0 0.7 Canada 7.6 7.7 10.3 10.8 8.3 8.5 10.6 10.8 9.1 1.8 European Union 20.7 21.7 23.2 23.0 20.8 19.4 19.4 15.8 21.2 1.0 United States 12.8 13.7 15.1 14.6 14.5 16.5 18.5 18.0 15.1 1.7 Developing economies Argentina 33.3 20.0 30.3 35.9 31.1 21.4 15.6 20.0 26.8 10.0 Brazil 14.3 14.3 21.7 28.6 28.9 25.6 22.4 20.4 22.3 4.7 India 38.5 46.7 33.3 27.3 32.8 22.3 25.2 27.1 32.3 10.7 Mexico 33.3 28.3 40.7 38.4 36.3 35.1 44.3 41.8 36.6 60.3 South Africa 8.3 27.6 28.6 23.2 19.5 18.8 19.1 22.5 20.7 12.0 Turkey 13.5 13.5 14.7 17.6 17.1 14.3 25.0 40.0 16.5 4.4 All other 34.0 28.8 26.2 22.5 18.9 12.8 15.5 - 22.7 - Total number 143 148 180 193 202 179 199 - 174 Percent of all anti- 17.8 18.0 21.2 20.9 19.9 18.3 19.5 - 19.4 dumping measures -, Not available. aPer $1,000 of imports from China. Source: Author's computations based on wro Reports on Antidumping (G/ADP/N series at www.wto.org) and wro trade data. The total number of measures is based on Lindsey and Ikenson (2001). measures and invest its negotiating strength in the wTo in working for stricter antidumping rules, as its chief trade negotiator announced? Clearly, China's decision will have a decisive impact on the evolution of world antidumping enforcement and on wro trade disciplines more generally. II. MINIMIZING CHINA'S EXPOSURE TO FOREIGN ANTIDUMPING The slow pace of Wro negotiations means that China would likely not be able to get reforms of wro antidumping rules into play for at least two (more likely four) years. Meanwhile, it will be hard for Chinese authorities to resist pressures from import-competing firms in China that demand more intensive use of antidumping procedures. One possibility: Could China adopt measures minimizing as quickly as possible its exposure to foreign antidumping, alle- viating the political costs of playing a reforming role in wTo antidumping rules? Messerlin 111 As noted, trade problems are fundamentally domestic conflicts between firms- between export-oriented and import-competing industries. The country-based data discussed sheds no light on this deeper aspect of protection and can even be misleading. For instance, the strong asymmetry between antidumping users and targets shown in tables 1 and 2 could suggest that wTo members would be induced to bring more antidumping cases until the situation is so bad that wro members collectively adopt stricter disciplines on antidumping. Arguments have already been presented suggesting that the top antidumping users are unlikely to follow this path. Taking firms' behavior into account suggests a darker scenario. Petition- ing firms-the driving forces in antidumping enforcement-may lodge antidump- ing complaints against several key countries to segment the world market in their products (evidence supporting this hypothesis is shown later). In that case the growing use of antidumping measures becomes not an incentive to discipline antidumping use but an incentive for firms to use the measures ever more inten- sively. Examining these deeper aspects of antidumping protection requires looking at the distribution of antidumping measures in force by sector or product rather than by country. A Few "Antidumping-Intensive" Sectors Antidumping measures are concentrated in a handful of Harmonized Tariff System sections (table 4). Metals, chemicals, machinery and electrical equip- ment, textiles and clothing, and plastics account for 75 percent of antidumping measures, even though these sectors account for less than half of world trade. These sectors are key sources of exports for dynamic developing economies in the first stages of industrial development, and they tend to have a high propor- tion of relatively standard products and oligopolistic market structures. Although the metals and chemicals sectors clearly have these features, the other sectors require a closer look, with greater disaggregation. There are few antidumping actions in the machinery and clothing subsectors that are charac- terized by many firms producing highly differentiated products. Most anti- dumping actions are in the electrical equipment and textiles subsectors that are characterized by relatively standard products produced by oligopolistic firms. This pattern strongly suggests that firms use antidumping as a cheap and powerful instrument for segmenting the markets that ongoing or scheduled trade liberalization is making more competitive. It also suggests that antidump- ing cases are likely to spread to such sectors as clothing that will increasingly experience product differentiation and imperfect competition (based on trade- marks, goodwill, distribution channels, and the like). In sum, the observed sectoral pattern of antidumping reflects the increasing "privatization" of trade policy by firms that have enough initial oligopolistic power to use the "procollusion" bias is embedded in antidumping regulations-a key lesson that should be kept in mind when implementing these regulations, in China as elsewhere. TABLE 4. Antidumping Measures in Force in the World by Section of Harmonized System, 1995-2000 Number antidumping measures China Tariff lines Average with tariffs Growth index applied <10% as % Harmonized System Section 1998-2000/ Export Import tariff of total section 1995 1996 1997 1998 1999 2000 share (%) 1995-97 pattern pattern 2001 tariff lines 15 Metals, metal articles 276 290 305 316 372 401 33.5 117.0 6.7 9.2 9.6 68.6 6 Chemical, allied 201 203 204 210 217 228 21.6 104.1 4.7 8.0 9.4 78.2 industry products 16 Machinery, 81 85 94 95 100 81 9.2 116.1 29.2 37.8 14.4 30.0 electrical equip. 11 Textiles, clothing 48 60 61 54 70 79 6.4 114.2 19.8 7.4 21.0 12.6 7 Plasters, rubbers 53 54 55 53 66 74 6.1 108.4 3.2 7.3 15.9 18.0 17 Vehicles, 42 40 40 42 43 30 4.0 101.6 3.7 2.8 23.8 43.4 transport equipment 13 Articles of stone, 34 37 37 42 44 40 4.0 115.5 1.6 0.8 17.7 12.1 plaster, cement 10 Woodpulp, paper, 18 17 21 44 44 39 3.1 207.6 0.7 3.1 13.4 26.1 paperboard 4 Prepared foodstuffs 35 31 29 31 28 27 3.1 88.9 2.1 0.8 28.4 15.7 2 Vegetable products 15 19 19 24 23 22 2.1 129.4 2.1 1.9 21.3 29.2 12 Footwear, headwear 21 16 9 17 13 16 1.6 70.3 4.8 0.2 23.2 0.0 18 Optical, cinema 14 11 10 9 20 23 1.5 104.0 3.4 3.6 14.3 30.1 instruments 20 Miscellaneous 14 15 14 13 12 14 1.4 89.7 7.1 0.4 20.2 1.2 manufactured articles 9 Wood, articles of wood 6 6 6 5 11 11 0.8 122.2 1.1 1.7 9.8 61.9 1 Live animals, animal 4 3 4 6 7 7 0.5 161.9 1.7 1.0 18.9 21.5 products 3 Animal, vegetable fats, 9 9 9 2 1 1 0.5 44.4 0.1 0.5 40.6 5.9 oils 5 Mineral products 0 0 0 9 9 9 0.5 - 3.7 10.9 4.1 96.3 19 Arms, ammunitions 3 3 3 3 0 0 0.2 66.7 0.0 0.0 13.0 0.0 8 Raw hides, skins, 0 0 1 1 1 1 0.1 - 3.0 1.4 16.4 25.0 leather products 14 Gems, jewelry 0 0 0 0 0 0 0.0 - 1.0 0.5 12.6 67.2 21 Art objects 0 0 0 0 0 0 0.0 - 0.2 0.8 10.9 25.0 Total 874 899 921 976 1,081 1,103 100.0 112.0 100.0 100.0 15.8 38.7 Source: Lindsey and Ikenson (2001). 114 THE WORLD BANK ECONOMIC REVIEW, VOL. I 8, NO. I China's Sensitivity to "Antidum ping-Intensive" Sectors Does China import a high proportion of products from antidumping-intensive Harmonized Tariff System sections? If so, Chinese authorities are likely to be under strong pressure to impose antidumping measures on such goods, thereby participating in the worldwide segmentation game that is the ultimate goal of antidumping enforcement. The five most antidumping-intensive Harmonized Tariff System sections in table 4 make up almost 70 percent of Chinese imports, opening the possibility that Chinese firms or foreign firms producing in China could present antidumping complaints in order to segment world markets. China's sensitivity to antidumping-intensive sectors can also be assessed from an export perspective. China's machinery and electrical equipment and textiles and clothing exports are particularly sensitive to worldwide antidumping activ- ity (see table 4). China would appear to have been a target of foreign antidump- ing measures much more because of intrinsic economic features of its exported products (differentiation level and oligopolistic markets) than because China was not yet a WTO member. Chinese Policies to Minimize Exposure to Foreign Antidumping This conclusion implies that China should not expect to face fewer antidumping measures in the coming years because of its WTO accession. Rather, China should expect to continue to face a large number of antidumping cases. Of course, WTO membership gives China access to the WTO dispute settlement mechanism, which could provide some relief to Chinese exporters harassed by foreign antidumping. But this relief will probably be only marginal and short- lived. For instance, the 2001 WTO dispute settlement ruling banning use of the averaging method is already being circumvented by alternative procedures (hastily developed by creative petitioners). A more promising route for minimizing China's exposure to foreign antidump- ing measures may be China's own policies, beginning with its trade policy. If China's tariffs in antidumping-intensive sectors remain high at the end of its accession period, it may not adopt as many or as severe antidumping measures on imports from the rest of the world in these products (though the analysis in section I shows that the top developing economy antidumping users have not hesitated to add high antidumping measures to still unliberalized trade regimes). But such a tariff policy, far from ensuring fewer antidumping cases against Chinese exports, will facilitate new antidumping measures against exports in antidumping- intensive sectors. High Chinese tariffs will permit high prices in domestic markets in China, making dumping claims against Chinese exporters easier to prove all the more because the exports in question consist mostly of basic products (for which Chinese exporters are likely to align their prices to those prevailing in foreign markets because there are no or small premia for differentiation). One way to minimize exposure to foreign antidumping would thus be to adopt uniform and moderate tariffs to reduce distortions in the domestic Messerlin 115 production pattern (foreign antidumping investigators interpret such distortions as signs of dumping). Although average tariffs for Harmonized Tariff System sections give very imperfect information on the tariff structure (peak tariffs within each section are eroded by low tariffs), they suggest that tariff peaks in China are concentrated in sectors that are not antidumping-intensive activities, with the exception of textiles and clothing (see table 4). A uniform tariff policy may help China shift the composition of its exports away from antidumping-intensive sectors for another reason. Economies such as Hong Kong (China), Japan, the Republic of Korea, Singapore, and Taiwan (China) reduced their exposure to foreign antidumping measures by upgrading their exported products. One could even argue that foreign antidumping meas- ures accelerated the economic development of these economies by inducing them to shift production more quickly to highly differentiated products in which they anticipated having comparative advantage. EU and U.S. Antidurn ping "Echoing" against China's Exports Capturing antidumping protection as a market segmentation strategy by a few large firms requires information at the product level, rather than at the aggre- gated section level of table 4. Table 5 provides such detailed information for EU and U.S. antidumping actions against China. More precisely, it lists EU and U.S. "echoing" cases against China, antidumping cases that targeted the same Chi- nese exports. The European Union and the United States are the top antidump- ing users in the absolute number of measures in force, and they are the two largest markets for Chinese exports (15 and 21 percent, respectively). Echoing cases (58 altogether) constitute 75 percent of antidumping cases initiated against Chinese exports by the United States and 68 percent by the European Union. These cases generally echoed each other within a year or less. All but three of these cases (cycles, hammers, and pocket lighters) resulted in antidumping measures of some kind. Such a large proportion of echoing cases and the similarity of their outcomes are signs that antidumping is a protectionist instrument that petitioners are using in a strategic way to segment the two largest world markets. Thus, assessment of the welfare costs of antidumping measures should take into account not only the severity of the measures (the high level of antidumping duties or the restrictiveness of quantitative restrictions) but also the dramatic reduction in competition in importing markets. These indirect welfare costs generated by antidumping-caused collusion compound the direct welfare costs generated by the antidumping duties. Examination of echoing cases also shows that U.S. antidumping duties are on average higher than EU duties: 104 percent compared with 38 percent. How- ever, comparing the measures closely is difficult because of differences in regulations and the lack of sufficiently detailed information. For instance, some EU cases are terminated by the withdrawal of complaints by petitioners. The effective impact is hard to ascertain. It may be limited to the chilling effect TABLE 5. EU and U.S. Echoing Antidumping Cases against Chinese Products, 1980-99 Positive decisions Negative decisions Antidumping duties (%) Initiating country Dumping Year or group Product margin, EU (%) EU U.S. EU U.S. EU U.S. 1992 EU Antimony trioxide 43.2 noi 1991 U.S. Antimony trioxide N 1988 EU Barium chloride 50.1 D 25.8 1982 EU Barium chloride 75.0 Ds 1983 U.S. Barium chloride N 1983 U.S. Barium chloride A 14.5 1999 EU Brushes, hair ep 1992 U.S. Brushes, hair T 1986 EU Brushes, paint 100.0 Und 1992 EU Brushes, paint noi 1985 U.S. Brushes, paint A 127.1 1999 U.S. Brushes, paint 1994 EU Coumarin 50.0 Ds 1994 U.S. Coumarin A 160.8 1991 EU Cycles 30.6 D 30.6 1995 U.S. Cycles N 1984 EU Cycles, chains 45.0 Und 45.0 1999 EU Cycles, forks withdraw 1999 EU Cycles, frames withdraw 1996 EU Cycles, parts 30.6 D 30.6 1999 EU Cycles, wheels withdraw 1992 EU Ferrosilicon 49.7 D 49.7 1992 U.S. Ferrosilicon A 137.7 1995 EU Furfuryl alcohol withdraw 1994 U.S. Furfuryl alcohol A 45.3 1999 EU Glycine ep 1994 U.S. Glycine A 155.9 1985 EU Handtools: hammers no 1990 U.S. Handtools: hammers A 45.4 1994 U.S. Lighters, disposable N 1990 EU Lighters, pocket 16.9 D 16.9 1991 EU Magnesium oxide 27.7 Und 1994 U.S. Magnesium, alloy A 79.4 1994 U.S. Magnesium, pure A 108.3 1997 EU Magnesium, unwrought 31.7 D + und 31.7 1994 U.S. Manganese, metal A 143.3 1992 EU Manganese, unwrought withdraw 1996 U.S. Persulfates 1994 EU Persulfates, peroxodisulfates 110.1 D 83.3 1982 U.S. Polyester, cotton cloth A 36.2 1990 EU Polyester, yarn 23.5 D 23.5 1986 EU Potassium permanganate 94.5 D + und 28.0 1983 U.S. Potassium permanganate A 39.6 1984 EU Silicon carbide 31.5 Und 1993 U.S. Silicon carbide 1989 EU Silicon, metal 38.7 Ds 18.7 1991 EU Silicon, metal 178.0 Ds 1990 U.S. Silicon, metal A 139.5 1994 EU Steel, pipe or tubes fittings 58.6 D 58.6 1999 EU 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 EU Tungstate, ammon. Para. 75.7 noi 1988 EU Tungsten, carbide 73.1 D+und 1988 EU Tungsten, metal powder noi 1991 U.S. Tungsten, ore A 151.0 1989 EU Tungsten, ofes 50.3 D+und 1988 EU Tungstic, oxide and acid 85.8 D+und Total 58 Average margins and duties 59.2 29 15 5 7 37.8 104.5 Note: EU 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. Source: Bloningen (2001) for the United States and Official Journal for the European Union. 1 18 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. I on Chinese exporters, forced to export less, or at higher prices, or both, to limit the risk of facing new antidumping complaints. Or if withdrawals reflect merely a lack of cooperation by the domestic industry or a failure to get the minimum number of EU member states to support a proposal to impose measures, they may reflect the fact that EU petitioners have been able to impose quantity or price restraints on Chinese exports on a "private" basis-with the corresponding full-fledged impact that is to be expected from these hidden restraints. III. CHINA'S ANTIDUNMPING ENFORCENIENT: AT THE CROSSROADS The following assessment of China's antidumping regulations and enforcement is provisional. It is constrained by the limited number of ongoing cases (a substantial number of cases are often needed for a robust assessment). China's Antidumping Regulations China adopted its first antidumping regulations on March 25, 1997 and the guidelines necessary for implementing the law later the same year. Following China's accession to the wTo, these old regulations were replaced by new regulations in January 2002, along with another set of guidelines (see G/ADP/ N/1/CHN1 and G/ADPIN/1/CHN2 at www.wto.org and Wang 2003). China's regulations follow the usual structure of antidumping legislation: proof of the existence and estimate of the magnitude of dumping and of material injury and proof of the causal relation between dumping and injury. However, they have four striking features. First, many details are left to the detailed guidelines or to case-by-case practice. Although this is common for countries that are adopting their first antidumping regulations, this lack of detail gener- ates legal uncertainty. Second, China's regulations include all the (well-known) protectionist biases of the WTO antidumping provisions. Among these are use of the concept of a major proportion of the industry as the threshold level for accepting complaints (a condition that domestic monopolies, oligopolies, or cartels fit much more easily than competitive industries); possibility of ex officio initiation of cases by the Chinese authorities; screening of complaints by the antidumping office, exposing the office to strong and hidden pressures by vested interests; possibility of withdrawal by petitioners, facilitating private collusion between petitioners and defendants; cumulation of imports, facilitating demonstration of injury and widening the geographical scope of protective measures; recourse to constructed normal values when comparable prices are unavailable in the exporting country, enabling manipulation of costs and reasonable profit; a broad definition of the confidentiality of information, limiting the rights of defendants; the possibility of imposing undertakings as antidumping measures and the requirement that antidumping duties be borne by importers (the so-called no absorption provi- sion); the possibility of imposing retroactive antidumping duties where there is a "history" of dumping (that is, recurrent antidumping complaints); and the Messerlin 119 possibility of taking "appropriate" measures should foreign firms try to circum- vent the antidumping measures. Third, prior to the 2003 government reform, coordination of the many admin- istrative agencies involved in antidumping investigations was extremely complex. The Fair Trade Administration for Imports and Exports of the Ministry of Foreign Trade and Economic Cooperation (MOFTEC) received complaints, decided which to accept, and was involved in the entire process, including investigation. The State Economic and Trade Commission joined MOFTEC in determining the existence of injury at the preliminary stage and conducting the final investigations. The Cus- toms General Administration joined MOFTEC for some parts of the investigations. For the imposition of antidumping duties, MOFTEC presented its proposals to the Tariff Commission under the State Council, which made the decisions. The merger of the State Economic and Trade Commission and MOFTEC in early 2003 brought these two functions into one ministry. Fourth, the old regulations included the following Article 40: "In the event that any country or region applies discriminatory antidumping or countervailing meas- ures against the exports from the People's Republic of China, the People's Republic of China may, as the case may be, take counter-measures against the country or region in question." It is not known whether this provision led to any cases, but it clearly opened the possibility for China to use antidumping rules as a retaliatory instrument. Article 56 of the new regulations is only slightly more diplomatic: "Where a country (region) discriminatorily imposes antidumping measures on the exports from the People's Republic of China, China may, on the basis of the actual situations, take corresponding measures against that country (region)." Antidumping Enforcement by China According to information provided by MOFTEC for antidumping cases under the old regulations and on China's latest notification (G/ADP/N/105/CHN) to the WTO for the cases under the new regulations, 69 cases were initiated between 1997 and May 2003 (table 6). After a slow start in 1997-98, the number of cases increased rapidly, reaching 24 in 2002 and 11 for the first five months of 2003. It is unclear yet whether this simply reflects cases that were in the pipeline for a long time or whether China has begun to follow in the footsteps of the six top developing economy antidumping users (see section I) and will soon become another anti- dumping-intensive user, endangering its so far successful liberalization. Although it is too early to know whether China's initial antidumping meas- ures are representative of future antidumping enforcement, a few observations are in order. First, antidumping measures have been taken in almost all cases, a very high percentage compared with what is generally observed (60-70 percent in industrial countries). Second, the level of the measures adopted by the Chinese authorities is relatively high, although it seems that the most recent measures are less severe than the ones taken under the old regulations. The main countries targeted are industrial and advanced developing economies-not the TABLE 6. China's Antidumping Measures in Force and Investigations, as of June 30, 2003 Provisional antidumping Definitive antidumping duties (%) duties (%) Initiation year Country Product Minimum Maximum Minimum Maximum Cases initiated under old regulations 1997 Canada Newsprint 17.1 78.9 9.0 78.0 1997 Korea, Rep. of Newsprint 17.1 78.9 9.0 78.0 1997 United States Newsprint 17.1 78.9 9.0 78.0 1999 Russia Steel, cold-rolled silicon 11.0 73.0 0.6 62.0 1999 Korea, Rep. of Polyester film 21.0 72.0 13.0 46.0 1999 Japan Steel, cold-rolled stainless 4.0 75.0 17.0 58.0 1999 Korea, Rep. of Steel, cold-rolled stainless 4.0 75.0 17.0 58.0 1999 Germany (EU) Acrylates 24.0 74.0 0.0 0.0 1999 Japan Acrylates 24.0 74.0 31.0 69.0 1999 United States Acrylates 24.0 74.0 31.0 69.0 2000 U.K. (EU) Methylene chloride 7.0 39.0 6.0 39.0 2000 France (EU) Methylene chloride 28.0 75.0 28.0 75.0 2000 Germany (EU) Methylene chloride 67.0 67.0 66.0 66.0 2000 Netherlands (EU) Methylene chloride 10.0 58.0 9.0 57.0 2000 Korea, Rep. of Methylene chloride 7.0 28.0 4.0 28.0 2000 United States Methylene chloride 49.0 58.0 49.0 58.0 2001 Japan Polystyrene a a a a 2001 Korea, Rep. of Polystyrene a a a a 2001 Thailande Polystyrene a a a a 2001 Indonesia Lysine a a a a 2001 Korea, Rep. of Lysine a a a a 2001 United States Lysine a a a a 2001 Korea, Rep. of Polyester, chips a a a a Cases initiated under the new regulations 2001 Korea, Rep. of Polyester, staple fiber 4.0 48.0 2.0 48.0 2001 Korea, Rep. of Pet chips 6.0 52.0 5.0 52.0 2001 Indonesia Acrylates 11.0 24.0 11.0 24.0 2001 Korea, Rep. of Acrylates 11.0 20.0 2.0 20.0 2001 Malaysia Acrylates 13.0 38.0 4.0 38.0 2001 Singapore Acrylates 46.0 49.0 30.0 49.0 2001 Belgium (EU) Caprolactam 6.0 16.0 6.0 16.0 2001 Germany (EU) Caprolactam 28.0 38.0 28.0 38.0 2001 Netherlands (EU) Caprolactam 9.0 18.0 6.0 18.0 2001 Japan Caprolactam 5.0 21.0 5.0 18.0 2001 Russia Caprolactam 6.0 29.0 7.0 16.0 2002 Korea, Rep. of Polyester, film - - 0.0 0.0 2002 India Anhydride, purified 33.0 33.0 b b 2002 Japan Anhydride, purified 66.0 66.0 b b 2002 Korea, Rep. of Anhydride, purified 14.0 33.0 b b 2002 Japan Styrene butadiene 0.0 33.0 b h 2002 Korea, Rep. of Styrene butadiene 10.0 27.0 b b 2002 Russia Styrene butadiene 16.0 46.0 b b 2002 Kazakhstan Steel, cold-rolled products 21.0 48.0 b b 2002 Korea, Rep. of Steel, cold-rolled products 9.0 40.0 b b - 2002 Russia Steel, cold-rolled products 9.0 29.0 b b 2002 Taiwan, China Steel, cold-rolled products 8.0 55.0 b b 2002 Ukraine Steel, cold-rolled products 12.0 22.0 b b 2002 Japan Polyvinyl chloride 32.0 115.0 b b 2002 Korea, Rep. of Polyvinyl chloride 10.0 76.0 b b 2002 Russia Polyvinyl chloride 34.0 82.0 b b 2002 Taiwan, China Polyvinyl chloride 10.0 27.0 b b 2002 United States Polyvinyl chloride 25.0 83.0 b b 2002 Japan Toluene 19.0 49.0 b b 2002 Korea, Rep. of Toluene 6.0 22.0 b b 2002 United States Toluene 23.0 28.0 b b 2002 Japan Phenol 7.0 144.0 b b 2002 Korea, Rep. of Phenol 10.0 10.0 b b 2002 Taiwan, China Phenol 7.0 20.0 b b 2002 United States Phenol 29.0 29.0 b b 2003 Germany (EU) Ethanolamine b b b b 2003 Iran Ethanolamine b b b b (Continued) TABLE 6. Continued Provisional antidumping Definitive antidumping duties (%) duties (°%) Initiation vear Country Product Minimum Maximum Minimum NMaximum 2003 Japan Ethanolamine b b b b 2003 Malaysia Ethanolamine b b b b 2003 Mexico Ethanolamine b b b b 2003 Taiwan, China Ethanolamine b b b b 2003 United States Ethanolamine b b b b 2003 European Union Chloroform b b b b 2003 India Chloroform b b b b 2003 Korea, Rep. of Chloroform b b b b 2003 United States Chloroform b b b b 69 All cases Average antidumping duty 17.2 43.4 9.8 23.7 -, Not available. aStatus unknown. hOngoing investigations. Souirce: China Ministry of Foreign Trade and Economic Cooperation, WTO Semi-annual report (G/ADP/N/I05/CHN, August 22, 2003, www.wto.org). Messerlin 123 same pattern observed for the other developing economy antidumping users. Third, the pattern for cases initiated since 2001 is closer to that observed for the other intensive antidumping users. This is most clearly illustrated by the steel cases (echoing the EU and U.S. safeguards) and the ethanolamine cases (observed in several other antidumping users). The increasingly similar product pattern suggests that China's antidumping enforcement is beginning to join the ongoing process of segmenting world markets through antidumping activity. It also raises issues of the progressive capture of China's trade policy by firms, similar to what is observed in the 10 major antidumping users. In this context it would be important to know whether petitioners are Chinese firms (private or state-owned) or firms with strong links with foreign firms (joint ventures, technical relations, vertical integration) that are experienced in the "art" of antidumping. IV. CHINA'S OPTIONS IN THE WTO NEGOTIATIONS ON CONTINGENT PROTECTION In its antidumping enforcement, China needs to take into accoant an issue that the six top developing economy antidumping users have not had to deal with: China's wTo accession protocol incorporates specific provisions on antidumping and safeguards (for a legal analysis, see Vermulst 2000). Seemingly a handicap, these special provisions can be turned into an instrument of positive change. The special provision on antidumping could induce China to negotiate in the Doha Round a more economically sound interpretation of the specific provisions on antidumping. That could create strong incentives in China for restraining its own use of antidumping and for fighting for stricter WTO rules on antidumping. It could give China's trading partners strong incentives to ease their transition period of accession. The special provision on safeguards is more difficult to deal with. This provision is so much at odds with the spirit and rules; of the WTO that its use will raise a large systemic risk for the entire WTO. Linking China's Effective Liberalization to Better Treatment of Its Exports by Antidiunmping Users China's protocol of accession allows its trading partners to use "nonmarket economy" status in their investigations against allegedly duamped Chinese exports for 15 years (until 2017). That status allows antidumping investigators to use proxies for estimating the home market prices or costs of Chinese exports in determining whether dumping took place. Such proxies make the existence of dumping much easier to prove than the rules for market economies, and they inflate the magnitude of the estimated antidumping margins cornpared to those (already high) imposed on market economies. A summary of the information available for 208 EU and U.S. antidumping cases initiated between 1995 and 1998 gives a sense of the intrinsic biases of the nonmarket economy procedure (table 7; Lindsey 1999; Messer]lin 2000b). The 124 THE WORLD BANK ECONOMIC REVIEW, VOL. 18, NO. I TABLE 7. Do Antidumping Investigations Really Look at Dumping? U.S. cases (1995-98) EU cases (1995-97) Basis of estimated Number Average Number Average normal values cases dumping cases dumping of exports examined margins (%) examined margins (%) Price comparisons only 5 3.2 8 22.7 Price comparisons and 33 14.2 33 24.4 constructed values Constructed values 20 25.1 12 25.1 Nonmarket 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 Source: Lindsey (1999) for U.S. cases; Messerlin (2000a) for EU cases. further from pure price comparisons the methodology used for estimating dumping margins, the higher are the estimated margins: from 3 percent (United States) and 22 percent (European Union) under pure price comparisons to 25 percent under various constructed-value methods. Use of nonmarket economy status is clearly linked to the highest dumping margins found (40 percent in the United States and 46 percent in the European Union).3 It is almost impossible to eliminate a provision included in a country's protocol of accession. But during the Doha Round it may be possible to negotiate an economically sound interpretation of the use of the nonmarket economy status by wro members. China and its trading partners have a com- mon interest in establishing the strongest possible link between China's effective liberalization and elimination of the use of nonmarket economy status by foreign antidumping authorities. What is at issue is not the elimination of the nonmarket economy provision itself (impossible to obtain) but its effective use in the future.4 The argument aims at mobilizing export interests in both China and the rest of the world during China's period of accession to the WTO. Chinese exporters, 3. For instance, under nonmarket economy status, it is possible to use industrial countries (such as the United States or Sweden) as reference countries for China. That introduces systemic errors about the product and 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 metal produced by China and Russia in large quantities for sales on international markets. The U.S. product is likely to have characteristics in terms of quality and availability that make it very different from the Russian or Chinese calcium metal, and it is sold and bought in a market structure that is very different from the markets of its Russian and Chinese counterparts. In the same vein, trying to estimate production costs by combining input prices in industrial countries and input quantities used in a developing economy makes little economic sense. 4. EU and U.S. antidumping authorities have already adopted more liberal interpretations in some cases than the protocol allows. In the European Union, for example, the authorities have accepted individual treatment or market economy status with respect to some Chinese exports. Messerlin 125 knowing that they will face less unfair treatment (no use of the nonmarket economy status) in foreign antidumping cases if China is effectively opening its markets in accordance with its accession protocol, will be motivated to monitor China's liberalization more closely and support it more strongly, including through stricter use of antidumping regulations. Foreign exporters, for their part, will support stricter use of antidumping regulatons by their own authorities, especially with respect to use of nonmarket economy status, if they believe that they will get more effective and stable access to Chinese markets. This could be achieved by implementing the following simple rule. Foreign antidumping investigators will automatically grant market economy status to Chinese exporters of any product that meets the three following conditions. * The Chinese most-favored-nation tariff on the product involved is moder- ate (say, 10 percent or lower). This threshold tariff will be one of the core components of a more economically sound interpretation of China's non- market economy status to be agreed on during the Doha Round. It could be stable over time, or it could increase as time passes-showing an increasing confidence in the ongoing liberalization process among China's trading partners. * No "core gray-area measures" are imposed on the product by the Chinese authorities. The list of core gray-area measures to be introduced in the interpretation agreement should also be negotiated during the Doha Round. The list should be short (say, specific tariffs, quantitative restric- tions, and minimum prices), and only the listed measures should be con- sidered part of the conditions. * State-owned monopolies shall not engage in distributing the competing foreign and domestic varieties of the product in question. Chinese state- owned monopoly producers are acceptable because as economic analysis shows, a protection granted exclusively by a moderate tariff eliminates the risk of monopoly power of the domestic sole producer. In 2001, 38 percent of China's tariff lines had ad valorem tariffs lower than 10 percent (see table 4). (The fact that the Chinese tariff schedule has only roughly 7,000 tariff lines suggests that it does not offer many opportunities to create narrow niches of protection for carefully defined tariff items.) Applying the three conditions would thus substantially reinforce the rights of Chinese exporters in the antidumping cases lodged in the two Harmonized Tariff System sections with the largest number of antidumping cases. (But China should take the initiative to improve the situation in the other antidumping-intensive Har- monized Tariff System sections, in particular in textiles and clothing.) These three conditions make it unlikely that Chinese exporters would dump except for economically sound reasons (differences in demand pattern, need to meet foreign demand, or to make Chinese products known in foreign markets). Thus, China's WTO trading partners should grant China at least the unconditional 126 THE WORLD BANK ECONOMIC REVIEW, VOL. i 8, NO. I benefit of market economy status in any antidumping investigations faced by Chinese exports meeting these conditions.5 Negotiations on improved implementation of nonmarket economy status for China should be as swift as possible. These conditions can easily be defined on a tariff line (Harmonized Tariff System) basis. For instance, China could notify the WTO on a regular basis of the tariff lines for which these conditions are met (this can easily be included in the general monitoring procedures for China's accession). Cross-notifications by China's trading partners could be added to the process, under the condition that they not slow the process. Finally, weaker variants of the suggestions could be considered, if necessary. For instance, the nonmarket economy status could be eliminated for notified goods only after, say, one year, instead of immediately. However, it is worth noting that any weakening of the suggested approach may have huge costs in terms of decreas- ing incentives for export interests in both China and the rest of the world to support the transition process of China's accession to the wro. Stricter Rules on Antidumping The desire of China's chief trade negotiator, Long Yongtu, to introduce stricter rules on antidumping is a natural extension of the negotiations on use of the nonmarket economy status proposed here because of the focus on the anti- dumping rules faced by allegedly dumped exports from market economies. China's efforts to introduce stricter rules on wTo antidumping could follow either of two very different approaches. A cautious approach would be to table a series of proposals or to support those already tabled in Geneva for improving WTO-based antidumping regula- tions at the margin. For instance, the following suggestions, derived from proposals tabled in 1999 by the Swedish Kommerzkollegium (1999), could receive China's support: * Dumping should be the principal cause of material injury. * Double protection (for instance, antidumping measures imposed on top of quantitative restrictions) should not be allowed. * Measures should last five years at most (implying stronger limits to review). * Repeated initiations in a short period of time should not be allowed. * Cumulation of imports from different countries should be banned or severely restricted, unless they come from the same firms or subsidiaries of the same firms. * Aggregation of products under the one single product procedure should be severely restricted. 5. It could be argued that market forces in China for these relatively unprotected products could be distorted by Chinese regulations on inputs for such goods (subsidies, for instance). But there are WTO instruments for addressing such an argument (which could be applied to most China trading partners). Messerlin 127 * All zeroing practices (using only export transactions that have been found to be dumped in calculating dumping margins) should be banned (all export transactions should be included in the investigation). * Antidumping authorities should produce disclosure documents. * Use of the de minimis rule should be expanded in an economically sound way. Alternatively, China could adopt a bolder approach. Antidumping could take, as often as possible, the form of negotiated "quantitative thresholds" (Messerlin 2000a). wTo members could agree that no antidumping measure should be imposed in cases where the level of injury losses is less than an agreed threshold of the complainants' revenues for the year(s) used as the reference (predumping) period. An approach based on quantitative thresholds is concep- tually equivalent to tariffication. It tends to give a sense of the magnitude of the concessions granted by both sides, bringing antidumping more in line with the usual WTO negotiating techniques. It is also flexible enough to permit incremen- tal reforms, to deliver the progressive liberalization that wrio members are looking for, through progressive increases in the thresholds. This would avoid the current deadlock of binary choices between fully enforcing antidumping regulations and rejecting them totally. The Transitional Product-Specific Safeguard Provision China's protocol of accession includes a transitional product-specific safeguard mechanism, which makes it much easier for wTo members to impose safeguard measures against China's exports until 2014 (for a detailed description from a legal perspective see Andersen and Lau 2001). In China's case, all the terms defining the use of safeguard actions in the traditional GATT-WTCO context (under Article XIX) have been weakened. There is no requirement of unforeseen circumstances and no most-favored-nation requirement. Only "material" rather than "serious" injury needs to be demonstrated. There are fewer factors related to the condition of the domestic industry, and the causal link between increased imports and injury is weaker. The most important difference-and potentially the most devastating for the wro-is that wro members are given the right, never before offered, to use a trade-diversion clause. As soon as one wro member implements a transitional product-specific safeguard measure against Chinese exports, all other members can enforce a similar measure at almost no procedural cost (no investigation, no prior notification, no input from Chinese parties). The trade-diversion clause thus means that countries do not have to provide proof substantiating the allegation that Chinese exports will be diverted from the first closed market to the rest of the world. All these features put the transitional product-specific safeguard at odds with the usual WTO concerns about a fair balance between rights and obligations. The provision is so unbalanced that its use could easily trigger in China feelings similar to those provoked by the unequal treaties of the nineteenth century. The 1 28 THE WORLD BANK ECONOMIC REVIEW, VOL. 1 8, NO. I transitional product-specific safeguard represents a serious systemic risk to the WTO regime. Although it could be argued that the transitional product-specific safeguard is such a politically aggressive instrument that it would never be used, this seems unlikely. Because of its politically explosive content, this safeguard is unlikely to be initiated by any but a very large industrial country WTO member (the United States, the European Union, or Japan). But once the provision has been trig- gered, other WTO members that could benefit from the trade-diversion clause will likely do so. How can China reduce the risk that the transitional product-specific safeguard provision will be invoked? One possibility would be for them to negotiate the same kind of preemptive approach as recommended for the nonmarket economy provi- sion during the Doha Round. WTO members could agree not to use the transitional product-specific safeguard provision as long as Chinese products meet the three conditions presented for use of the nonmarket economy option and to use the normal wro safeguard provision under Article XIX instead. Unifying Contingent Protection in the wro Putting antidumping and safeguards on a par would makes a lot of sense from the perspective of the global wro architecture. Most wro members use anti- dumping measures as a substitute for safeguard actions for dealing with indus- tries in difficulty. The transitional product-specific safeguard provision strengthens China's stake in seeking substantial improvement in the whole WTO contingent protection regime-both antidumping and safeguards. During the Doha Round, China could try to expand the negotiations on antidumping to safeguards (so far not explicitly included in the Doha negotiating program) to make the entire contingent protection regime of the wro more consistent. One promising approach would be to tie together the concept of temporary protection embedded in safeguards and the basic concept of renegotiation under GATT Article XXVIII (Messerlin 2000a). Thus, for instance, at the end of the second period of enforcing a safeguard measure under the current safeguards agreement (based on GATT Article XIX), the country would be required either to renegotiate the tariff on the product subjected to safeguard measures or elim- inate the safeguard measure (shifting to antidumping or other trade remedies should be prohibited, in recognition that all instruments of contingent protec- tion are substitutable). This mandatory aspect would help reform safeguard and antidumping procedures back to the transitory protection they were meant to be instead of the permanent protection they have become. The possibility remains that an integrated approach to antidumping, safe- guards, and transitional product-specific safeguard measures would face entrenched hostility from wTo members, in particular from the top 10 users of antidumping provisions. China would then be forced to rely on threats of some kind of retaliation. The least aggressive approach would be for China to announce its intention to systematically use the dispute settlement mechanism Messerlin 129 as soon as a wTo member notifies the WTO Secretariat of its intention to use the transitional product-specific safeguard provision. Lawyers tend to overstate the benefits of such an approach by ignoring the full development of the trade conflicts-that is, the political bickering that follows dispute settlement cases (it is almost certain that dispute settlement cases dealing with the transitional pro- duct-specific safeguard would leave the two parties in a particularly difficult political situation). A more aggressive approach by China would be based on Article 56 of the new regulations, as discussed earlier. However, such an approach should not ignore the basic principle of deterrence: trade deterrence, like nuclear deterrence, works as long as it remains a threat-it must stop short of action. V. CONCL[JSION Two findings of this analysis are particularly important for China. First, the countries that stand to gain most from better discipline on antidumping are the handful of developing economies that have been intensive users of antidumping since the 1995 Uruguay Round. Because the antidumping measures imposed by these developing economies tend to be more frequent and severe than those imposed by industrial countries, these actions hurt developing economies much more than industrial country-imposed antidumping measures harm industrial countries' economies. If China wants to continue to enjoy successful liberal- ization, it must avoid becoming an intensive user of antidumping. Second, there are few economic or political forces to act as automatic restraints on the current situation of antidumping enforcement. Major users of antidumping measures have few incentives to reform their very discriminatory use of the antidumping instrument, whereas smaller countries have few incen- tives to use the antidumping instrument in a retaliatory way. As a result, China is at a crossroads. One way leads to more intensive use of antidumping for several reasons: as a retaliatory instrument against foreign anti- dumping, as a tool for China's progressive integration into the worldwide collusive dimension of antidumping (used as an instrument for segmenting world markets for the benefit of large firms), and as a backdoor entry to old-fashioned protection, even at the risk of unraveling its scheduled trade liberalization. Another way leads to a guiding role for China in arguing for stricter rules on the use of antidumping. As a small antidumping user and a key target of foreign antidumping, China will be one of the main beneficiaries of such a move, which will also help them negotiate an economically sound interpretation of the special provisions on antidumping and safeguards included in its WTO accession proto- col. This new interpretation should be based on China meeting a few key and economically sound conditions: low tariffs, no core gray-area measures, no distribution monopolies (see section IV). This interpretation is motivated by strong economic and political arguments. China and its trading partners have a common interest in establishing the strongest possible links between China's effective trade liberalization and agreement not to use these special provisions 1 30 THE WORID BANK ECONOMIHC REVIEW, VOL. 18, NO. I against Chinese exporters. This interpretation seeks to mobilize export interests in both China and the rest of the world to their mutual gain during the difficult implementation period of China's accession to the WTO. REFERENCES Andersen, Scott, and Christian Lau. 2001. Hedging Hopes with Fears in China' Accession to the World Trade Organization: The Transitional Special Product Safegziard for Chinzese Exports. Geneva: Powell, Goldstein, Frazer and Murphy. Bloningen, Bruce. 2001. "Database on U.S. Antidumping Cases." Online document available at www.uoregon.edu. Lindsey, Brink. 1999. The U.S. Antidumping Law: Rhetoric versus Reality. Washington, D.C.: Cato Institute. Lindsey, Brink, and Dan Ikenson. 2001. Coming Home to Roost: Proliferating Antidumping Laws and the Growing Threat to U.S. Exports. Washington, D.C.: Cato Institute. Kommerzkollegium. 1999. Further Actions within the EU in the Field of Antidumping. Stockholm. Messerlin, Patrick A. 2000a. "Antidumping and Safeguards." In Jeff Schott, ed., The WTO After Seattle. Washington, D.C.: Institute for International Economics. - . 2000b. "An Economic Perspective on Antidumping and Safeguards." Seminar on Antidumping and Safeguards. Ministry of Foreign Trade and Economic Cooperation, Department of International Trade and Economic Affairs, Beijing. Vermulst, Edwin. 2000. "Contingent Protection Provisions in China's Draft Protocol on Accession to the wTo." Seminar on Antidumping and Safeguards, Ministry of Foreign Trade and Economic Coopera- tion, Department of International Trade and Economic Affairs, Beijing. Wang, Lei. 2003. "China's New Antidumping Regulations: Improvement to Comply with the wIo Rules." East Associates, Beijing. [Iwang@ealawfirm.com.] wro (World Trade Organization). 2001. "Market Access: Unfinished Business, Post-Uruguay Round Inventory and Issues." Special Studies 6. Geneva.