77516 Doha Merchandise Trade Reform: What Is at Stake for Developing Countries? Kym Anderson, Will Martin, and Dominique van der Mensbrugghe The LINKAGE model of the global economy and the latest Global Trade Analysis Project (GTAP) database (version 6.05) are used to examine the impact of current merchandise trade barriers and agricultural subsidies and possible reform outcomes of the World Trade Organization’s (WTO’s) Doha Development Agenda. The results suggest that moving to free global merchandise trade would boost real incomes in Sub-Saharan Africa proportionately more than in other developing countries or in high-income countries, despite the terms of trade loss in parts of that region. Particular attention is given to agriculture, as farmers constitute the poorest households in developing coun- tries but the most assisted in rich countries. Net farm incomes would rise substantially in Sub-Saharan Africa and other developing country regions, alleviating rural poverty. Partial liberalization could move the world some way toward those desirable outcomes, the more so the more developing countries themselves cut applied tariffs, particularly on agricultural imports. This article (a) summarizes the costs of current merchandise trade distortions to developing and other economies; (b) examines some scenarios that might emerge as part of an eventual Doha agreement consistent with the 2005 Hong Kong Ministerial Declaration [World Trade Organization (WTO) 2005], particularly with respect to agriculture; and (c) draws implications for the strategies devel- oping countries might adopt in the WTO’s Doha Round of multilateral trade negotiations. This article estimates what the world economy might look like in 2015 with- out and with a successful conclusion to the Doha Round, how far Doha could take the world toward an outcome with no distortions in merchandise trade, and what contribution various elements of a Doha package could make. The analysis relies on a recursive model of the global economy known as LINKAGE (for details, Kym Anderson, Will Martin, and Dominique van der Mensbrugghe are Lead Economists in the Devel- opment Economics Vice Presidency of the World Bank. Their email addresses are kanderson@worldban- k.org, wmartin1@worldbank.org, and dvandermensbrugg@worldbank.org. The authors are grateful for helpful comments from seminar participants and journal referees, for tariff-cutting data from the staff of Centre d’Etudes Prospectives et d’Informations Internationales (CEPII) in Paris (with special thanks to David Laborde) and for funding from the U.K. Department for International Development. THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2, pp. 169–195 doi:10.1093/wber/lhj009 Advance Access publication May 9, 2006 Ó The Author 2006. Published by Oxford University Press on behalf of the International Bank for Reconstruction and Development / THE WORLD BANK. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org. 169 170 THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2 see supplemental appendix S.2 posted at http://wber.oxfordjournals.org), which has formed the basis of the World Bank’s standard decade-long projections of the global economy and its earlier trade analysis (see, for example, World Bank 2002, 2004). It also uses the latest version (6.05) of the Global Trade Analysis Project (GTAP) database, which includes the tariff preferences enjoyed by many developing countries (http://www.gtap.org). The results distinguish between the effects on developing countries and those on more advanced economies. In doing so, it is necessary to consider both the World Bank’s classification of economies by income level and the WTO practice of self-nominated developing country status, under which even economies as advanced as Hong Kong (China), Singapore, the Republic of Korea, and Taiwan (China) claim developing country status and so are eligible for special and differential treatment, including lesser tariff cuts and longer phase-in periods than are eventually agreed for developed countries following the Doha Round. The analysis suggests that most of the potential gains from multilateral reform would come from agriculture. But because of the large gaps between WTO‘s bound and applied rates of protection, there would be little real agricultural reform globally as a result of the Doha Round—especially by developing countries—unless WTO members make substantial cuts to their bound tariff rates and domestic farm subsidy commitments. This article explores the effects of a more ambitious agricultural reform package over the next decade and of developing countries participating more fully in the Doha Round rather than invoking special and differential treatment to avoid reform. If WTO members insist on classifying even a small number of farm products as ‘‘sensitive’’ and subject to lesser tariff cuts, the gains from agricultural reform could be greatly diminished—and even disappear for developing countries. This article begins with an overview of the key elements of a prospective Doha agreement consistent with the Hong Kong Ministerial Declaration (WTO 2005) and focusing on the agricultural elements. It describes the model of the global economy used to analyze the consequences of such an agreement and of alternative, more-ambitious reforms including a move to complete free trade (which provides a helpful benchmark). The estimates of protection and subsidy rates for each region are a crucial part of the data in the global model, and so they are examined before turning to the key results of the simulations. After discussing some qualifications, this article draws out some implications for developing countries. I. KEY ELEMENTS OF A PROSPECTIVE DOHA AGREEMENT To what extent are trade and subsidy reform commitments likely to emerge from the Doha Round? In addressing that question, it is important to know that WTO trade negotiators are seeking agreement on reductions not to applied tariffs and Anderson, Martin, and van der Mensbrugghe 171 subsidies but rather to members’ legally bound import tariffs, agricultural export subsidies, and commitments on domestic support to farmers. These bound rates are higher than applied rates in nearly all countries, but especially in most developing countries, meaning that cuts in bound rates will have a weaker impact on market access. The Doha Round was launched at the WTO Ministerial Meeting in Doha in late 2001, but the following Ministerial Meeting, in Cancu ´n in September 2003, ended in acrimony and without agreement on how to proceed. At Cancu ´n, developing countries made it abundantly clear that further progress would not be possible without a commitment by developed countries to significantly lower their agricultural subsidies (including, importantly, for cotton, despite its rela- tively minor role in developed country agriculture; see Sumner 2006). The so- called July Framework Agreement (WTO 2004) and the Hong Kong Ministerial Declaration (WTO 2005) reiterate the importance of keeping development at the heart of the Doha agenda and stress agricultural reform as a key to doing that. Annexes to these documents provide guidance on how a Doha agreement might be structured, with frameworks for establishing modalities for agriculture and nonagricultural market access, as well as providing recommendations for trade in services. The following sections highlight the key elements of a prospective Doha agreement focusing on agriculture and the state of negotiations to date. Agricultural Market Access The gap between bound and applied tariffs is the so-called binding overhang and can significantly blunt the impact of any negotiated outcome—so much so that in some Doha scenarios, some countries are not required to change their applied tariffs at all. Jean, Laborde, and Martin (2006) examined the consequences for applied tariff cuts of different bound tariff-cutting formulas, taking into account agricultural tariff rate quotas, the prevalence of preferences for developing countries (as described in Boue ´, and Jean 2006), the need to accom- ¨ t, Fontagne modate ‘‘sensitive’’ and ‘‘special’’ farm products, and the special and differential treatment outlined in the July Framework. Tariff cutting, implemented at the six-digit level of the Harmonized System (HS) of commodity disaggregation, involves a detailed comparison of each country’s bound tariff, which is what negotiations focus on, with the applied most-favored nation tariff on a given bilateral trade flow, which is what affects economic outcomes. The applied tariff cuts vary not only by sector but also by trading partner—and may involve smaller or no cuts on imports from developing countries currently enjoying nonreciprocal preferential access to richer countries’ markets (Hoekman and O¨ zden 2005). Following the detailed tariff analysis, the results were aggregated up to the GTAP and LINKAGE models’ regional and sectoral levels. Jean, Laborde, and Martin (2006) evaluated the consequences for 2001 applied rates of different approaches to liberalization, particularly different 172 THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2 degrees of top-down progressivity in the bound tariff cuts, as well as different degrees to which developing countries participate in reform. They looked first at a proposal similar to the Harbinson progressive reduction formula (WTO 2003b), with marginal tariff rate reductions of 35 percent for tariffs below 15 percent, 65 percent for tariffs above 90 percent, and 60 percent for tariffs within the 15–90 percent bracket.1 Developing countries’ tariff cuts also follow the progressive formula but with four rather than three brackets and with inflexion points at tariff levels of 20, 60, and 120 percent to be consistent with Harbinson’s criterion of cutting by an average of 25, 30, 40, and 45 percent in those four brackets. That set of tariff cuts leads to very little import liberalization, because bound tariffs in many countries exceed applied rates by such large margins. As a result, Jean, Laborde, and Martin focused on another set of reforms that involve cuts in applied agricultural protection rates that are at least 10 percentage points greater: a 45, 70, and 75 percent cutting rule for bound rates for developed countries and a 35, 40, 50, and 60 percent cutting rule for developing countries.2 These cuts are within the (wide) range proposed by key WTO members in the lead- up to the Ministerial Meeting. Jean, Laborde, and Martin then examine the consequences of: . Allowing lesser tariff cuts for self-nominated ‘‘sensitive’’ farm products, assuming that countries would take into account the importance of the commodity, the height of the tariff, and the gap between the tariff binding and the applied rate in deciding as to which products to grant such treat- ment, comparing situations in which countries are allowed to treat 2 percent of agricultural and food tariff lines as sensitive and subject to just a 15 percent tariff cut; . Including ‘‘special’’ agricultural products just for developing countries, by adding another 2 percent of agricultural tariff lines as subject to just a 15 percent tariff cut; . Adding a tariff cap of 200 percent, consistent with the suggestion in para- graph 30 of the July Framework Agreement that the role of a tariff cap be explored; . These scenarios are also modeled in the analysis here. Agricultural Domestic Support Reductions in domestic support have been a particular concern of developing countries. Developed countries are the major providers of such assistance, and 1. This approach provides cuts in average tariffs—without the discontinuities created by the propor- tional cuts involved in the Harbinson formula—that are more or less comparable with those generated by Harbinson’s proportional reductions of 25, 30, and 60 percent, because the larger cuts on higher tariffs apply only on the portion of the tariff above 15 or 90 percent. 2. With no cuts in least developed countries, as specified in the July Framework Agreement and the Ministerial Declaration. Anderson, Martin, and van der Mensbrugghe 173 many developing countries are concerned about the ability of their producers to compete with developed country farmers receiving large amounts of domestic support from their governments. While the marked asymmetry between devel- oped and developing countries is a concern, there is evidence that the benefits to developing countries from reductions in developed country domestic support may be substantially smaller than the potential gains from reductions in market access barriers (Hoekman, Ng, and Olarreaga 2004; Hertel and Keeney 2006; Anderson and Valenzuela forthcoming). Nonetheless, disciplining such support is crucial not only to prevent policy reversals but also to ensure that when tariffs are lowered, import protection is not simply replaced by equally or more distorting domestic measures. The Framework Agreement and the Ministerial Declaration propose tiered reductions in the total bound aggregate measure of support, with larger reduc- tions by members with higher initial levels of support. It turns out that extra- ordinarily large reductions in bound levels of support are required before any reductions in actual support would occur. If all countries with aggregate mea- sure of support notifications above 20 percent of the value of production cut their bound protection by 75 percent and all others by 60 percent, only four members would have to cut applied rates as of 2001: the United States by 28 percent, Norway by 18 percent, the European Union (EU) (the pre-expansion EU- 15) by 16 percent, and Australia by 10 percent. Agricultural Export Subsidies Export subsidies for nonfarm goods are outlawed in the WTO, so eliminating farm export subsidies would simply be bringing agriculture into line with other goods. The empirical analyses summarized in the works of Hertel and Keeney (2006) and Anderson and Valenzuela (forthcoming) show that export subsidies contribute only a small part of the welfare cost of agricultural support programs. That is true even when implicit subsidies in the form of food aid and export credits are included. A phase-out by 2013 of both explicit and implicit forms of farm export subsidies, as agreed at the Hong Kong Ministerial Meeting, should therefore be a politically feasible component of a comprehensive Doha agree- ment. Their elimination in isolation could harm a few food-importing and aid- dependent developing countries, but the poor net buyers of food in those countries can be assisted through other more direct and more cost-effective forms of aid than through these measures. Nonagricultural Market Access Negotiations on nonagricultural tariffs have been lagging behind those on farm products. There has been a clear indication that developing countries wish to make smaller tariff cuts compared with developed countries and that the least developed countries expect not to have to make any cuts. A Doha Round is unlikely to involve cutting all nonagricultural bound tariffs by more than 174 THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2 50 percent, so the assessment assumes a 50 percent cut by developed countries, a 33 percent cut by developing countries, and no cut by least developed countries. However, because cuts in bound rates may lead to very little reduction in applied rates by developing countries, a more ambitious scenario is also explored, with developing countries committing to more reform and in return seeking recipro- city in the form of further cuts in developed countries’ agricultural and textiles tariffs. The most optimistic possibility considered is that developing countries (including least developed countries) agree to cut nonagricultural bound tariffs as much as developed countries (that is, by the 50 percent assumed for the analysis). Services Trade WTO members have been very slow in coming forward with Doha proposals to reform services trade. At this stage, it seems likely that, as with the Uruguay Round, countries will make few meaningful commitments to genuinely open up their services sectors. For that reason, and because services trade is less ade- quately represented in trade models than is goods trade, reductions in this sector are not included in the analysis—despite the fact, as indicated by Hertel and Keeney (2006), that gains from services reform could well be enormous, includ- ing for developing countries. II. THE GLOBAL LINKAGE MODEL FOR ASSESSING EFFECTS OF FUTURE TRADE REFORM The analysis uses the LINKAGE model, a relatively straightforward global compu- table general equilibrium (CGE) model but with some characteristics that distin- guish it from standard comparative static models such as the GTAP model [described by Hertel (1997)]. A key difference is that it is recursive, so that although it starts with 2001 as its base year, it can be solved annually through to 2015. This is important when evaluating a reform that is likely to take a decade or more to be fully implemented, because the structure of the world economy will be quite different in 2015 than it is in 2001. Economic expansion in the model is driven by exogenous population and labor supply growth, savings-dependent capital accumulation, and exogenous labor-augmenting technological progress (as used in the World Bank’s Global Economic Prospects 2004 and as detailed in supple- mental appendix S.2). In any given year, factor stocks are fixed. Producers minimize costs subject to constant returns to scale production technology, consumers maximize utility, and all markets—including for unskilled and skilled labor, which are both intersectorally mobile—are cleared with flexible prices. Also consistent with the focus on long-run adjustment to reform, the aggre- gate supply of farmland is defined by an overall upward sloping supply function, Anderson, Martin, and van der Mensbrugghe 175 with land-abundant countries having a higher land supply elasticity.3 Land is allocated across agricultural activities using a constant elasticity of transforma- tion function. There are three types of production structures: (a) crop sectors reflect the substitution possibility between extensive and intensive farming; (b) livestock sectors reflect the substitution possibility between pasture and intensive feeding; and (c) all other sectors reflect standard capital/labor substitution. There is a single representative household per modeled region, allocating income to consumption using the extended linear expenditure system. Trade is modeled using a nested Armington structure in which aggregate import demand for each sector’s product is the outcome of allocating domestic absorption between domestic goods and aggregate imports, and aggregate import demand is allocated across source countries to determine bilateral trade flows. There are various sources of protection in the model. The most important involves bilateral import tariffs. There are also bilateral export subsidies. Domestically, there are subsidies only in agriculture, applied to intermediate goods, outputs, and payments to capital and land. Household consumption and savings are represented by the extended linear expenditure system, which provides a rigorous framework for mod- eling consumption and savings decisions and the allocation of consumption spending across commodities (Lluch 1973). Government fiscal balances are fixed in any given year, with government spending fixed as a share of GDP and the fiscal objective being met by changing the level of lump-sum taxes on households.4 This implies that losses of tariff revenues are replaced by higher direct taxes on households. The current account balance is fixed, primarily for convenience in this recursive model5 but also consistent with the Feldstein-Horioka finding of limited international capital mobility (Feldstein and Horioka 1980; Ventura 2003). Finally, investment is driven by savings. With fixed public and foreign savings, investment is determined by changes in the savings behavior of households and changes in the unit cost of investment. The model solves only for relative prices, with the numeraire, or price anchor, being the export price index of manufactured exports from high-income countries. Version 6.0 of the LINKAGE model is based on release 6.05 of the GTAP database, which has a 2001 base year instead of the 1997 base year of GTAP version 5, updated national and trade data, and a new source for the protec- tion data (see http://www.gtap.org for details). The new protection data are 3. Key elasticities for the LINKAGE model are in supplemental appendix S.1, available online at http:// wber.oxfordjournals.org. 4. For simplicity, they are fixed in U.S. dollar terms at their base year level, minimizing potential sustainability problems. But this implies that they decrease over time as a percentage of GDP for expanding economies. 5. Only with fixed financial inflows from abroad can utility changes be used to provide a money- metric measure of welfare changes resulting from a reform. 176 THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2 from a joint Centre d’Etudes Prospectives et d’Informations Internationales (Paris)/International Trade Centre (Geneva) project. The product of this joint effort, known as MAcMaps, is a detailed database on bilateral protection at the HS six-digit level that integrates trade preferences, specific and compound tariffs, and a partial evaluation of nontariff barriers such as tariff rate quo- tas.6 The new GTAP database has lower tariffs than the previous database because of the inclusion of bilateral trade preferences, major reforms between 1997 and 2001 such as continued implementation of the Uruguay Round agreements, and China’s WTO accession, which alone caused the ratio of global exports plus imports to GDP to rise from 44 to 46 percent over those four years. The LINKAGE model used for this study comprises a 27-region, 25-sector aggregation of the GTAP database. There is a heavy emphasis on agriculture and food, which account for 13 of the 25 sectors, and on the largest commodity exporters and importers. III. THE SUBSIDIES AND IMPORT PROTECTION DATA SET The main source of protection is tariffs or border barriers, although some countries—particularly high-income countries—also have significant agricul- tural production and export subsidies. The average import tariff for agriculture and food in 2001 was 16.0 percent for high-income countries and 17.7 percent for developing countries, while for manufactures other than textiles and cloth- ing, it was 8.3 percent for developing countries and just 1.3 percent for high- income countries. The averages of course obscure large variations across countries and com- modities. For example, if high-income countries’ tariffs on temperate farm products are at a near-prohibitive 100 percent but zero on tropical products such as coffee, the import-weighted average agricultural tariff could be quite low. Even at a relatively aggregated level, the variations can be quite sharp. For example, India has an average tariff on agriculture and food of 82 percent on imports from East Asia but only 20 percent on imports from Sub-Saharan Africa. Also, high-income countries’ agricultural tariffs are lower on goods from low-income countries than on goods from high- and middle-income countries, while imports of textiles and clothing from low-income countries face a higher average tariff than imports from middle- or high-income countries. 6. More information on the MAcMaps database is available in the study by Boue ¨ t and others (2004) and at http://www.cepii.fr/anglaisgraph/bdd/macmap.htm. For a detailed analysis of the differences between the results presented here and those obtained by using the LINKAGE model and the earlier GTAP version 5 database and those obtained by using the same version 6 GTAP database but the GTAP model, see the works of Anderson, Martin, and van der Mensbrugghe (2006, appendix 12A) and van der Mensbrugghe (2006). Anderson, Martin, and van der Mensbrugghe 177 IV. ESTIMATES OF THE WELFARE IMPACT OF CURRENT PROTECTION POLICIES The LINKAGE model provides a baseline projection of the world economy first to 2005 and then to 2015 assuming no other policy changes. Deviations from that baseline in 2015, due to phased partial or total liberalization from 2005, are then examined. One benchmark against which to measure the prospective benefits of the Doha Round is the gains that would come from completely freeing merchandise trade (including removing all agricultural producer and export subsidies) over the 2005–10 period. That leads to global gains by 2015 of $287 billion7 a year. Another benchmark is the reform incorporated in the presimulation experiment for the period 2001–04, reflecting the final stages of Uruguay Round implemen- tation including the phase-out of the Multifibre Arrangement, the accession of China and Taiwan (China) to the WTO, and the enlargement of the EU from 15 to 25 members.8 The impacts of those reforms on import tariffs are nontrivial. Had those three reforms not already been implemented, the gains in 2015 from freeing global merchandise trade would have been $341 billion instead of $287 billion, or an extra $54 billion a year. Nearly half that difference is due to the removal of export quotas with the phase-out of the Multifibre Arrange- ment and the follow-on Agreement on Textiles and Clothing and so should be considered part of the Uruguay Round’s legacy—assuming safeguards by high- income countries or export restraints by China do not replace textile and cloth- ing quotas after 2005.9 The distribution of the standard economic welfare or real-income (equivalent variation) effects of removing all merchandise trade distortions (including agri- cultural subsidies) shows that two-thirds of the $287 billion gain in income that global reform would generate each year by 2015 would accrue to high-income countries (table 1). However, as a share of national income, developing countries (as self-defined by WTO members) would do twice as well, with an average 7. A billion is one thousand million. 8. These are the key internationally agreed and bound policy changes. Unilateral and unbound policy changes, such as recent reforms in EU and U.S. farm programs, are not included. 9. To get a sense of how important preferences are to developing countries and global welfare, the model was re-run for 2001, before the pre-simulation experiment and without those preferences in place. The estimated global welfare gains from reform are then $382 billion instead of $341 billion, and the developing country gains are $150 billion instead of $113 billion. That is, the inclusion of preferences in the database reduces estimated global welfare gains by 11 percent, developing country gains by 25 percent, and high-income country gains by 2 percent. Much of the difference is attributable to Sub- Saharan Africa, where the reduction is almost 50 percent. The reductions for developing countries are overstated, however, for two reasons. One is that no rules of origin or other impediments are assumed for developing countries fully utilizing their preferences. The second is that importers in the preference- providing rich countries are assumed not to use their power to gain a disproportionate share of the rent from that preferential access. In practice, neither of these assumptions holds, according to recent case ¨ zden and Sharma 2004; Olarreaga and O studies (O ¨ zden 2005). T A B L E 1 . Impacts on Real Income from Removing All Global Merchandise Trade Distortions Including Agricultural Subsidies, without and with Own-Country Participation, by Country and Region, 2015 (Changes from Baseline) Annual income gain due to Total gain as share of Annual real income change in terms of trade baseline income gain ($ billion) ($ billion) (percent) From other From own plus From other From own plus From own plus countries’ other countries’ countries’ other countries’ other countries’ Economy or region reforms reforms reforms reforms reforms Australia and New Zealand 6.2 6.1 4.4 3.5 1.0 EU-25 plus European 40.6 65.2 29.7 0.5 0.6 Free Trade Association United States 21.6 16.2 18.8 10.7 0.1 Canada 1.7 3.8 1.2 À0.3 0.4 Japan 17.4 54.6 13.7 7.5 1.1 178 Korea, Rep. and Taiwan (China) 17.0 44.6 10.0 0.4 3.5 Hong Kong (China) and Singapore 8.7 11.2 6.6 7.9 2.6 Argentina 4.0 4.9 1.9 1.2 1.2 Bangladesh 0.0 0.1 À0.1 À1.1 0.2 Brazil 11.8 9.9 7.4 4.6 1.5 China 16.6 5.6 12.5 À8.3 0.2 India 3.9 3.4 1.3 À9.4 0.4 Indonesia 3.6 1.9 1.5 0.2 0.7 Thailand 9.8 7.7 3.9 0.7 3.8 Vietnam 2.4 3.0 1.5 À0.2 5.2 Russian Federation À1.0 2.7 0.2 À2.7 0.6 Mexico À1.2 3.6 0.9 À3.6 0.4 South Africa 1.1 1.3 0.7 0.0 0.9 Turkey 2.1 3.3 1.3 0.2 1.3 (Continued) T A B L E 1 . Continued Annual income gain due to Total gain as share of Annual real income change in terms of trade baseline income gain ($ billion) ($ billion) (percent) From other From own plus From other From own plus From own plus countries’ other countries’ countries’ other countries’ other countries’ Economy or region reforms reforms reforms reforms reforms Rest of South Asia 1.9 1.0 0.7 À0.8 0.5 Rest of East Asia 2.8 5.3 1.8 À0.91 1.9 Rest of Latin America 11.8 10.3 5.3 0.0 1.2 and the Caribbean Rest of Europe and Central Asia 0.8 1.0 0.7 À1.6 0.3 Middle East and North Africa 5.4 14.0 3.3 À6.4 1.2 Selected Sub-Saharan African countriesa 1.1 1.0 0.8 0.5 1.5 Rest of Sub-Saharan Africa 1.0 2.5 0.9 À2.3 1.1 179 Rest of the World 3.2 3.4 1.5 0.1 1.5 High-income countries 201.6 30.3 0.6 WTO developing countries 141.5 À21.4 1.2 Developing countries (World Bank definition) 85.7 À29.7 0.8 Middle-income countries 69.5 À16.7 0.8 Low-income countries 16.2 À12.9 0.8 East Asia and Pacific 23.5 À8.5 0.7 South Asia 4.5 À11.2 0.4 Europe and Central Asia 7.0 À4.0 0.7 Middle East and North Africa 14.0 À6.4 1.2 Sub-Saharan Africa 4.8 À1.8 1.1 Latin America and the Caribbean 28.7 2.2 1.0 World total 287.3 0.6 0.7 a Countries for which national modules are available in LINKAGE: Botswana, Madagascar, Malawi, Mozambique, Tanzania, Uganda, Zambia, and Zimbabwe. Source: Authors’ World Bank LINKAGE model simulations. 180 THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2 increase of 1.2 percent of national income over the baseline compared with 0.6 percent for high-income countries. The results vary widely, ranging from little impact for Bangladesh and Mexico to 4–5 percent increases in parts of East Asia. As a percentage of national income, Sub-Saharan Africa (excluding South Africa) would gain twice as much as high-income countries, despite the adverse change in the terms of trade for many African countries due in part to the loss of their nonreciprocal tariff preferences.10 Policymakers are frequently interested in the extent to which gains come from their own liberalization compared with the gains from liberalization by their trading partners. This was estimated by solving the model once for each region, considering only liberalization by that region. Column 2 of table 1 summarizes the total gains from multilateral liberalization, while column 1 summarizes the gains that would result without liberalization by the region but from improved access to partners’ markets, which Bagwell and Staiger (2002) suggest is the primary motivation for engaging in multilateral trade reforms. The analysis reveals that for regions with high agricultural protection (Western Europe, Northeast Asia, and Middle East), a large proportion of the gain comes from own-country reform. But for many more open economies, the bulk of the gain comes from increased market access in other countries. Even for Sub-Saharan Africa, 40 percent of the gain ($1.0 billion of $2.5 billion) would come from reforms in other regions, despite the much-lamented losses from preference erosion in industrial country markets. These results suggest that the benefits of a multilateral round could be substantially greater than the benefits from uni- lateral liberalization alone.11 Columns 3 and 4 in table 1 summarize the income effects of changes in the international terms of trade for each country. For developing countries as a group, the terms of trade effect is negative, somewhat reducing the gains from improved efficiency of domestic resource use (especially in China and India). A comparison of columns 3 and 4 reveals that it is mainly own-country reforms that are lowering developing countries’ terms of trade. Under the Armington assumption used in this model, the terms of trade losses are overestimated 10. The gains would be even greater if African reforms were accompanied by complementary domestic policy reforms and investments in trade-facilitating infrastructure and institutions (funds for which may be forthcoming in the proposed aid for trade package that may accompany a Doha agreement; see Nielson 2006). For more detailed disaggregation of the results for Sub-Saharan Africa, see the study by Anderson, Martin, and van der Mensbrugghe (forthcoming-c). 11. This result depends heavily on the size of the models’ Armington elasticities. Those in the LINKAGE model are about one-third larger than in the standard GTAP model on average, reflecting the focus on a longer (decade-long) adjustment period. The Armington elasticities would have to be even larger than those in the LINKAGE model to get more of a WYDIWYG (what you do is what you get) result. There is a fundamental problem with the Armington approach, which assumes that countries export more of the same products following liberalization. Kehoe and Ruhl (2003) show that much of the expansion following liberalization is typically in new products, a response that reduces the adverse terms of trade impacts of export growth. Estimates of response elasticities that take adequate account of this phenom- enon are not yet available. Anderson, Martin, and van der Mensbrugghe 181 because no allowance is made for expansion in the number or quality of exported goods resulting from reform (Hummels and Klenow 2005). There are several other ways to decompose the real-income gains from full global trade reform so as to better understand the sources of the gains for each region. One way is to assess the impacts of developing country liberalization and those of industrial country liberalization in different economic sectors. Another is to decompose by policy instrument. For agricultural reform, decomposing by policy instrument gives results very similar to those from the GTAP–AGR model reported by Hertel and Keeney (2006), who estimate that market access barriers explain 93 percent of the welfare effects of agricultural policies, domestic sup- port just 5 percent, and export subsidies just 2 percent.12 When decomposed by sector, the results suggest that global liberalization of agriculture and food contributes 63 percent of the global gains, which is similar to Hertel and Keeney’s 66 percent (table 2). This is consistent with the higher tariffs in agriculture and food (17 percent global average) than in other sectors but is nonetheless remarkable given the low shares of agriculture in global GDP (4 percent) and global merchandise trade (9 percent). The share of gains from agriculture is even higher for Sub-Saharan Africa, at more than three-quarters of the total welfare gain. Seventy percent of the global gains from agriculture are accounted for by the farm policies of high-income countries, and those policies also account for the bulk of the overall gains to high-income countries. For developing countries, as much of their gain from farm reform would come from agricultural liberal- ization in other developing countries as from getting unrestricted access to markets in high-income countries. The results are nearly the same for manufac- turing in aggregate, despite the large gains from reforms in textiles and clothing markets in high-income countries ($14 billion compared with $9 billion from growth in textile trade among developing countries). Thus, reform by developing countries is as important to the economic welfare gains of developing countries as reform by high-income countries. Notice also that developing country gains from high-income country reform are only half as large from textiles as from agricultural policies. What impact would the removal of cotton trade distortions and subsidies (which raise producer prices by more than 50 percent in the United States and even more in the EU) have in this context of freeing all merchandise trade and agricultural subsidies? The global price of cotton would rise an estimated 21 percent above the 2015 baseline on average because U.S. subsidies would no longer depress prices. However, the volume of U.S. cotton exports would shrink when the subsidies were removed, raising the price and volume of other 12. Hoekman, Ng, and Olarreaga (2004) reach a similar conclusion from estimating the effects of halving each of the three types of agricultural distortions, in their case using partial equilibrium analysis. For an intuitive, nontechnical explanation of this result—which has surprised many observers—see the study by Anderson, Martin, and Valenzuela (2006). T A B L E 2 . Sources of Regional and Sectoral Gains in Real Income from Full Liberalization of Global Merchandise Trade, Developing and High-Income Countries, 2015 (Changes from Baseline) Gains by region ($ billions) Share of regional gain (percent) All developing All high-income World All developing Middle-income Sub-Saharan Africa All high-income World Developing countries liberalize Agriculture and food 28 19 47 33 34 35 9 17 Textiles and clothing 9 14 23 10 12 11 7 8 Other merchandise 6 52 58 7 1 14 26 20 All sectors 43 85 128 50 47 60 42 45 182 High-income countries liberalize Agriculture and food 26 109 135 30 31 43 54 47 Textiles and clothing 13 2 15 15 15 À0 1 5 Other merchandise 4 5 9 5 7 À3 2 3 All sectors 43 116 159 50 53 40 57 55 All countries liberalize Agriculture and food 54 128 182 63 65 78 64 63 Textiles and clothing 22 16 38 25 27 11 8 14 Other merchandise 10 57 67 12 8 11 28 23 All sectors 86 201 287 100 100 100 100 100 Note: Small interaction effects are distributed proportionately, and numbers are rounded to sum to 100 percent. Source: Authors’ World Bank LINKAGE model simulations. Anderson, Martin, and van der Mensbrugghe 183 countries’ exports. In particular, cotton output from Sub-Saharan Africa would be 44 percent larger and cotton exports 73 percent larger under this full liberal- ization scenario, with the value of increases in both output and exports being greater than for any other region including Latin America and Australia (where there are more other agricultural expansion opportunities than in Africa, which would experience preference erosion). Indeed, cotton is so important in Sub- Saharan Africa excluding South Africa that it contributes one-quarter of the region’s net gain in agricultural value added from full global trade and subsidy liberalization. The share of all developing countries in global cotton exports would be 85 percent instead of 56 percent in 2015, further validating the efforts to ensure that cotton receives specific and substantial attention in the Doha negotiations (Baffes 2005; Sumner 2006). All these results are for full trade liberalization. Smaller changes can be expected from partial reforms of the sort currently being negotiated under the Doha Development Agenda. These are explored in the next section. V. ESTIMATING THE CONSEQUENCES OF PROSPECTIVE DOHA SCENARIOS What will the Doha package ultimately contain? Agricultural export subsidies are assumed to be eliminated by 2013 and domestic support for agriculture is assumed to be cut relative to 2001 levels in just four economies: by an average of 28 percent by the United States, 18 percent by Norway, 16 percent by the EU, and 10 percent by Australia. More difficult to determine are the likely nature and extent of reductions in market access barriers, so a number of scenarios are initially considered for agricultural and food products in isolation from nona- gricultural tariff cuts, before some nonagricultural market access reforms are also incorporated.13 Throughout this section, the WTO use of the term ‘‘develop- ing countries’’ applies, which means that Hong Kong (China), the Republic of Korea, Singapore, and Taiwan (China) are all able to enjoy reduced reform commitments under special and differential treatment despite their high-income status. Scenario 1 begins with a progressive or tiered reduction formula of marginal agricultural tariff rate reductions of 45, 70, and 75 percent for developed countries within each of the three bands defined by the Harbinson (WTO 2003b) inflection points of tariff rates of 15 and 90 percent, and reductions of 35, 40, 50, and 60 percent for developing countries within each of their four bands (least developed countries are not required to undertake any reduction commitments). Even with these large cuts to bound tariffs (which are about half way between those proposed by the United States and the EU in late 2005 in the lead-up to the Hong Kong Ministerial Meeting), average applied tariffs on agricultural and food products in 2015 would be only about one-third lower 13. As suggested in the Girard text (see WTO 2003a), a good without a bound tariff is treated as having double the applied most-favored nation rate. 184 THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2 globally, at 10.0 percent instead of 15.2 percent, and 12.5 percent instead of 14.2 percent for developing countries. Scenario 2 examines the consequences of including ‘‘sensitive’’ farm products as allowed for in the July Framework Agreement. Developed countries are allowed to treat up to 2 percent of their agricultural tariff lines at the HS six- digit level as sensitive and subject to just a 15 percent tariff cut,14 and both developing and least developed countries are allowed to treat up to 4 percent of agricultural tariff lines as sensitive, in part to incorporate their demands for ‘‘special’’ treatment of some products.15 Under this scenario, the average agri- cultural tariff falls to only 13.5 percent in both high-income and developing countries. Scenario 3 considers the effects of adding to scenario 2 a tariff cap of 200 percent. Any product with a bound tariff above that limit will be subject to a reduction to that rate, which leads to average cuts in agricultural tariffs of 18 percent for both developed and developing countries. The average agricultural tariff in 2015 would fall considerably more for high-income countries (to 11.5 percent) and only slightly more (to 13.3 percent) for developing countries. Scenario 4 adds to scenario 1 the cuts in nonagricultural tariff bindings of 50 percent in developed countries, 33 percent in developing countries, and zero in least developed countries. That lowers the average tariff on all merchandise from 2.9 percent in the baseline to 1.6 percent for high-income countries and from 8.4 percent to 7.5 percent for developing countries. Finally, scenario 5 makes developing (including least developed) countries full participants in the round, undertaking the same reductions in bound (but not necessarily applied) tariffs as the developed countries in scenario 4. That lowers the average tariff on all merchandise for developing countries from 8.4 to 6.8 percent instead of to 7.5 percent, a cut of almost one-fifth instead of one-ninth, as in scenario 4. Estimated Welfare and Trade Effects of the Scenarios as of 2015 The welfare consequences of implementing these reforms over the 2005–10 period and allowing the global economy to adjust to 2015 are summarized in table 3 in dollar terms and as percentage changes in real income in 2015. Agricultural liberalization using the harmonizing formula (scenario 1) would generate a global gain of $75 billion even without the inclusion of nonagricul- tural tariff reform. But almost all those benefits accrue to the reforming high- income countries (among which are included high-protection Korea and Taiwan (China) as well as Hong Kong (China) and Singapore in this and subsequent tables). Developing countries would gain only $9 billion because their tariff 14. Some proposals involve larger cuts in tariffs on these goods. 15. As described by Jean, Laborde, and Martin (2006), ‘‘sensitive’’ farm products are chosen for each country by taking into account the importance of the product, the size of its existing tariff, and the gap between its bound and applied tariffs. T A B L E 3 . Change in Real Income in Alternative Doha Round Scenarios, 2015 (Changes from Baseline) Billions of dollars Percent Economy or region Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Australia and New Zealand 2.0 1.1 1.2 2.4 2.8 0.35 0.20 0.20 0.42 0.48 EU-25 plus European Free 29.5 10.7 10.9 31.4 35.7 0.29 0.11 0.11 0.31 0.36 Trade Association United States 3.0 2.3 2.1 4.9 6.6 0.02 0.02 0.01 0.03 0.05 Canada 1.4 0.5 0.4 0.9 1.0 0.15 0.05 0.05 0.10 0.11 Japan 18.9 1.8 12.9 23.7 25.4 0.38 0.04 0.26 0.48 0.51 Korea, Rep. and Taiwan (China) 10.9 1.7 15.9 15.0 22.6 0.86 0.13 1.26 1.19 1.79 Hong Kong (China) and Singapore À0.1 À0.1 À0.2 1.5 2.2 À0.02 À0.03 À0.04 0.35 0.52 Argentina 1.3 1.0 1.0 1.3 1.6 0.32 0.26 0.26 0.34 0.39 Bangladesh 0.0 0.0 0.0 À0.1 À0.1 À0.06 À0.03 À0.04 À0.10 À0.09 185 Brazil 3.3 1.1 1.1 3.6 3.9 0.50 0.16 0.17 0.55 0.59 China À0.5 À1.5 À1.1 1.7 1.6 À0.02 À0.06 À0.04 0.07 0.06 India 0.2 0.2 0.2 2.2 3.5 0.02 0.03 0.02 0.25 0.40 Indonesia 0.1 0.2 0.0 1.0 1.2 0.05 0.07 0.01 0.37 0.44 Thailand 0.9 0.6 0.8 2.0 2.7 0.43 0.29 0.38 0.99 1.33 Vietnam À0.1 0.0 À0.1 À0.5 À0.6 À0.20 À0.09 À0.16 À0.83 À0.97 Russian Federation À0.3 À0.7 À0.7 0.8 1.5 À0.06 À0.16 À0.15 0.16 0.31 Mexico À0.2 À0.3 À0.3 À0.9 À0.2 À0.02 À0.04 À0.04 À0.11 À0.02 South Africa 0.1 0.3 0.3 0.4 0.7 0.06 0.17 0.17 0.25 0.49 Turkey 0.6 0.0 0.0 0.7 1.4 0.25 0.02 0.02 0.26 0.55 Rest of South Asia 0.2 0.1 0.2 0.3 0.7 0.13 0.05 0.14 0.17 0.39 Rest of East Asia 0.1 0.0 1.0 0.3 0.6 0.02 0.01 0.36 0.09 0.22 Rest of Latin America 3.7 0.5 0.4 3.9 4.0 0.44 0.06 0.04 0.46 0.47 and the Caribbean (Continued) T A B L E 3 . Continued Billions of dollars Percent Economy or region Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Rest of Europe and Central Asia À0.2 À0.3 À0.2 À0.6 À0.7 À0.06 À0.09 À0.08 À0.22 À0.26 Middle East and North Africa À0.8 À1.2 À1.2 À0.6 0.1 À0.07 À0.10 À0.10 À0.05 0.01 Selected Sub-Saharan African countries 0.1 0.0 0.0 0.1 0.2 0.21 À0.02 À0.05 0.19 0.26 Rest of Sub-Saharan Africa 0.0 À0.3 À0.3 À0.1 0.3 0.02 À0.13 À0.14 À0.02 0.13 Rest of the World 0.4 0.0 0.0 0.6 0.6 0.19 0.00 0.02 0.26 0.28 High-income countries 65.6 18.1 43.2 79.9 96.4 0.20 0.06 0.13 0.25 0.30 WTO developing countries 19.7 1.2 16.8 32.6 47.7 0.17 0.01 0.14 0.27 0.40 186 Developing countries 9.0 À0.4 1.1 16.1 22.9 0.09 0.00 0.01 0.16 0.22 (World Bank definition) Middle-income countries 8.0 À0.5 1.0 12.5 17.1 0.10 À0.01 0.01 0.15 0.21 Low-income countries 1.0 0.1 0.0 3.6 5.9 0.05 0.01 0.00 0.18 0.30 East Asia and Pacific 0.5 À0.8 0.6 4.5 5.5 0.01 À0.02 0.02 0.13 0.16 South Asia 0.4 0.3 0.4 2.5 4.2 0.03 0.03 0.03 0.21 0.36 Europe and Central Asia 0.1 À0.9 À0.9 0.8 2.1 0.01 À0.09 À0.09 0.08 0.21 Middle East and North Africa À0.8 À1.2 À1.2 À0.6 0.1 À0.07 À0.10 À0.10 À0.05 0.01 Sub-Saharan Africa 0.3 0.0 À0.1 0.4 1.2 0.06 À0.01 À0.02 0.10 0.27 Latin America and the Caribbean 8.1 2.3 2.1 7.9 9.2 0.29 0.08 0.08 0.29 0.33 World total 74.5 17.7 44.3 96.1 119.3 0.18 0.04 0.10 0.23 0.28 Source: Authors’ World Bank LINKAGE model simulations. Anderson, Martin, and van der Mensbrugghe 187 binding overhang is so great as to lead to almost no cuts in applied tariffs. Were countries allowed to have smaller cuts for even just 2 percent of farm products declared to be ‘‘sensitive’’ (and another 2 percent in developing countries for ‘‘special’’ farm products), global gains would shrink to just $18 billion and developing countries as a group would be even worse off (scenario 2). If such exceptions are made, it would be important to exploit the opportunity—pro- vided for in the Ministerial Declaration—to cap bound tariffs. Scenario 3 shows that even a cap as high as 200 percent would restore at least half the welfare gain forgone by allowing such exceptional treatment for sensitive and special farm products. The final two scenarios add nonagricultural tariff cuts to the agricultural reforms in the first two scenarios. In scenario 4, smaller cuts are provided for developing countries’ nonagricultural tariffs, as is the case for all agri- cultural tariff cuts in the preceding scenarios. Even so, relative to scenario 1, where only agricultural tariffs are cut, the gains to developing countries double by adding these nonfarm reforms, contributing one-third of the extra boost to global welfare ($7.1 billion of the $21.6 billion difference between the global gains from scenarios 1 and 4). In scenario 5, the devel- oping (including least developed) countries fully engage in the reform pro- cess, forgoing the lesser cuts provided for in scenarios 1–4. Because this leads to considerably larger cuts in applied tariffs, that boosts their welfare substantially, and global welfare as well. Nonetheless, the global average merchandise tariff hardly changes with agricultural reform alone, whereas it falls by almost one-third or 1.5 percentage points when manufacturing is included in the reform package. Retaining lesser tariff cuts for developing countries as in scenario 4 would yield a global gain of $96 billion from merchandise liberalization, which is a sizable one-third of what is in the table (the potential welfare gain from full liberalization of $287 billion, reported in table 1). But for developing countries, the gain would be only $16 billion, less than a fifth of that group’s potential gain from full liberalization ($86 billion; table 1). If developing countries forgo the option of reforming less than developed countries, their gain would rise by 42 percent, or an extra $7 billion (scenario 5). Much of those gains go to the largest developing economies, but in percentage terms Sub-Saharan Africa also gains substantially if it liberalizes more—contrary to the presumptions of many. By contrast, in sce- nario 4, the rest of Sub-Saharan Africa is not liberalizing enough to achieve sufficient efficiency gains to offset the terms of trade losses suffered as net food importers, as recipients of tariff preferences that have eroded with the decline in high-income countries’ most-favored nation tariffs or as a result of the combined export growth of reforming economies with similar export compositions.16 16. Details of the results for Sub-Saharan Africa can be found in the study by Anderson, Martin, and van der Mensbrugghe (forthcoming-b). 188 THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2 What would be the consequences for farm output and employment growth over the Doha implementation period of partial reform? If there were complete free trade, farm output would decline (instead of growing slightly) in just the EU and Japan, grow slower in a few other high-protection countries, and expand in most countries and regions. Doha scenario 4 would involve much less reform than a move to free trade, and hence a much slower loss of farm output for the EU and Japan but also less output growth than under free trade for the vast majority of countries where farm output would grow. For most protected economies, Doha scenario 4 would simply slow the growth of farm output slightly over the coming decade. The farm employment picture is somewhat different. Typically, economic growth leads to declines not only in the relative importance of agriculture but also in the absolute numbers employed in farming once a country reaches middle-income status. Thus, it is not surprising that numerous middle- and high-income countries are projected to lose farm jobs over the next decade in the baseline scenario. For the most protected farm sectors, the rate of farm employment decline would more than double if the world were to move to complete free trade, but it would increase only slightly under Doha scenario 4. For most developing economies, though, farm employment would grow a little faster under scenario 4 than under the baseline (Anderson, Martin, and van der Mensbrugghe 2006, table 12.17).17 How is this reflected in agricultural net income (value added by the farming sector)? Not surprisingly, agricultural value added would fall in regions with the highest agricultural protection—Europe, Northeast Asia, and to a lesser extent the United States (table 4). However, in the Doha reform scenario, none of the developing countries or regions would suffer a decline in agricultural net income, despite the lowering of their own agricul- tural tariffs even though the average agricultural tariff in developing countries is nearly as high as that in high-income countries (14.2 percent compared with 15.9 percent in the baseline). That is because a much larger proportion of developing country agriculture produces exportables, which do not have to be protected from imports. Because as much as 70 percent of the world’s poor are in farm households in developing countries, this result has clear implications for poverty alleviation. The trade consequences of Doha scenario 4 are considerable as well. By 2015, annual exports by developing countries would be $41 billion greater for agricultural products, $25 billion for textiles and clothing, and $12 billion for other manufactures. The total increase of $78 billion is somewhat smaller than that for high-income countries ($135 billion), but the difference is less in percentage terms (a 2.6 percent increase for developing countries compared 17. This finding of only small intersectoral labor movements in response to partial trade reform is consistent with econometric evidence of adjustments to past trade reforms (see, for example, Wacziarg and Wallack 2004). Anderson, Martin, and van der Mensbrugghe 189 T A B L E 4 . Impact of Reform Scenarios on Agricultural Value Added, 2015 (Changes from Baseline) Billions of dollars Percent Full global Full global Economy or region liberalization Scenario 4 liberalization Scenario 4 Australia and New Zealand 6.4 2.4 25.6 9.8 EU-25 plus European Free Trade Association À39.1 À20.4 À26.4 À13.8 United States À18.2 À6.3 À15.0 À5.2 Canada 3.4 0.9 23.3 5.8 Japan À17.7 À7.4 À39.5 À16.6 Korea, Rep. and Taiwan (China) À9.5 À3.4 À33.3 À12.1 Hong Kong (China) and Singapore 0.1 0.0 7.5 1.4 Argentina 6.1 1.7 33.8 9.4 Bangladesh À0.5 0.0 À4.4 0.4 Brazil 15.1 5.5 46.3 16.7 China 0.3 1.8 0.1 0.4 India À17.1 0.4 À8.1 0.2 Indonesia 0.8 0.5 2.7 1.7 Thailand 3.8 1.1 25.0 7.2 Vietnam 0.8 0.0 13.6 0.3 Russian Federation À1.4 À0.2 À6.5 À0.8 Mexico 0.9 1.2 2.5 3.2 South Africa 0.5 0.1 9.6 1.2 Turkey À2.0 À0.1 À7.2 À0.3 Rest of South Asia À0.6 0.8 À1.3 1.8 Rest of East Asia À0.2 0.5 À0.7 1.9 Rest of Latin America and the Caribbean 22.9 8.4 30.2 11.1 Rest of Europe and Central Asia À1.1 À0.1 À1.8 À0.2 Middle East and North Africa 0.3 1.0 0.3 0.9 Selected Sub-Saharan African countries 1.5 0.3 9.1 1.7 Rest of Sub-Saharan Africa 2.3 0.8 5.4 1.9 Rest of the World 3.1 1.0 16.4 5.4 High-income countries À74.6 À34.2 À19.4 À8.9 Developing countries (World Bank definition) 35.6 24.8 2.9 2.0 Middle-income countries 45.3 20.9 5.3 2.4 Low-income countries À9.7 3.9 À2.5 1.0 East Asia and Pacific 5.5 3.9 1.1 0.8 South Asia À18.1 1.2 À6.8 0.5 Europe and Central Asia À4.5 À0.3 À4.0 À0.3 Middle East and North Africa 0.3 1.0 0.3 0.9 Sub-Saharan Africa 4.3 1.1 6.7 1.8 Latin America and the Caribbean 45.0 16.7 27.4 10.2 World total À39.0 À9.5 À2.4 À0.6 Source: Authors’ World Bank LINKAGE model simulations. 190 THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2 with 3.1 percent for high-income countries). This takes global merchandise trade one-fifth of the way to where it would be under complete free trade in merchandise.18 Finally, of interest to those concerned that poor consumers would face higher food bills are the changes that might be expected in average food prices in international markets under the various Doha scenarios (table 5). For agriculture as a whole, prices would rise less than 2 percent over the 10-year phase-in period. The changes could be as high as 12 percent for dairy products, 6 percent for cotton, and 3–4 percent for coarse grains, oilseeds, sugar, and meat but well under 2 percent for other farm products. Thus, the annual change in basic food prices at the retail level would be hardly discernable even to poor consumers, thanks to the supply responsiveness of farmers to the increases in market access opportunities when agricultural subsidies and tariffs are reduced. T A B L E 5 . Impact of Doha Reform Scenarios on Average International Product Prices, 2015 (Percentage Change from Baseline) Product Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5 Rice 0.9 1.4 1.3 1.7 1.1 Wheat 1.8 1.6 1.6 1.8 1.7 Other grains 3.7 3.5 3.5 3.7 3.5 Oilseeds 4.5 3.9 3.9 4.5 4.4 Sugar 2.8 2.4 2.4 2.9 2.7 Cotton 6.0 5.8 5.8 5.9 5.8 Fruit and vegetables 1.2 0.9 0.8 1.3 1.0 Other crops 0.9 0.5 0.5 1.2 0.8 Vegetable oils and fats 0.7 0.6 0.6 1.0 0.8 Livestock 0.8 0.7 0.7 0.9 0.6 Processed meats 3.5 3.3 3.3 3.5 3.4 Dairy products 11.8 11.8 11.8 11.7 11.8 Other food, beverages, and tobacco 0.3 0.5 0.5 0.4 0.1 All agriculture and food 1.8 1.8 1.8 1.9 1.7 All primary agriculture 2.0 1.7 1.7 2.1 1.9 All processed agriculture 1.7 1.8 1.8 1.8 1.6 Textile and wearing apparel 0.2 0.2 0.2 0.2 À0.1 Source: Authors’ World Bank LINKAGE model simulations. 18. It also raises the share of agricultural and food production that is exported globally from 9.5 to 10.0 percent, which is one-seventh of the way toward its share of 13.2 percent under the free trade scenario. Even in the protected countries, this ratio rises a little or, in the case of the EU, falls only slightly. This is because farm resource move from currently protected import-competing subsectors to more competitive farming subsectors (Anderson, Martin, and van der Mensbrugghe 2006, table 12.18). Anderson, Martin, and van der Mensbrugghe 191 Some Caveats Results such as those presented here are always dependent on the underlying assumptions, data, and parameters and so are subject to numerous qualifica- tions. One particularly important qualification concerns the way preferences are treated in the version 6.05 GTAP database. Previous versions of the database included only key reciprocal preferences (notably between members within the EU, North American Free Trade Agreement, Association of Southeast Asian Nations, and Australia–New Zealand regional integration arrangements). Ver- sion 6.05 includes nonreciprocal tariff preferences provided by developed coun- tries to imports from developing countries under arrangements such as the Generalized System of Preferences, the EU’s Africa, Caribbean, and Pacific pro- gram and Everything but Arms agreement and the U.S. Africa Growth and Opportunity Act and Caribbean Basin Initiative. The analysis assumes that there are no rules of origin or other compliance requirements that raise the cost of using these preferences. It also assumes that the full benefits of the preferences flow to developing countries (even though developed country impor- ters often have more market power than developing country exporters of stan- dard commodities, who receive a smaller share of the rents).19 This treatment overstates the extent of preference erosion that would occur, especially for least developed countries, and so understates their gains from multilateral trade reform. If these nonreciprocal preferences were instead excluded from the data- base, the preference-receiving countries’ gains from developed country trade reform would be overestimated.20 Another important issue is the extent to which the model captures the supply- side constraints to adjustment by low-income countries to international price changes. The elasticities are aimed at representing adjustment to long-term changes but are still small compared with those used by some other analysts (for example, Harrison and others 2004). Other models, including GTAP–AGR (Hertel and Keeney 2006), use smaller trade elasticities and generate smaller gains globally and for developing countries, with some regions (including parts of Sub-Saharan Africa) even losing slightly. The uncertainty about the values of 19. Evidence that the preference margin is often eroded by complex rules of origin and that the rent is shared between importing and exporting countries with the exporters getting less the more trade is concentrated on standard commodities can be found in the works of Olarreaga and O ¨ zden (2005) and O¨ zden and Sharma (2004). A recent partial equilibrium study found that in practice export revenue losses from preference erosion are likely to be limited to a small subset of countries, primarily small island economies dependent on exports of sugar, bananas, and, to a far lesser extent, textiles and clothing (Alexandraki and Lankes 2004). 20. The extent of overstatement would not be large though, because the difference in the low-income countries’ estimated benefits even from full liberalization is only $2 billion a year when nonreciprocal preferences are excluded from the LINKAGE model’s database, or $8 billion when middle-income countries are also included (van der Mensbrugghe 2005). A further complication is that the nonreciprocal pre- ference scheme with the African, Caribbean, and Pacific countries is scheduled to be replaced in 2008 with reciprocal Economic Partnership Agreements between regional groupings of those countries and the EU. 192 THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2 these elasticities is a fundamental problem associated with pervasive measure- ment errors and uncertainty about the true model structure (Hummels and Klenow 2005). Also to be kept in mind is that global CGE models necessarily have to aggregate across sectors, thereby reducing the large variance in tariffs that are evident at the HS six-digit or greater levels of disaggregation. Because the welfare cost of a tariff is roughly proportional to the square of its height, this aggregation leads to underestimation of that cost. The analysis here does not include costs of adjustment to reform, but the structural changes that take place over time in the normal course of eco- nomic growth are typically very much larger than the small changes that would accompany gradual and partial trade liberalization [as shown in the study by Anderson, Martin, and van der Mensbrugghe (forthcoming-a)]. Furthermore, adjustment assistance schemes (financed by foreign aid in the case of low-income countries) to help fund adjustment to tariff and subsidy cuts are just one-off payments, whereas the benefits of reform continue into the future. Nor has this analysis taken into account the fact that trade reform typically boosts factor productivity and that not all sectors are subject to constant returns to scale and perfect competition. Most models that allow increasing returns and imperfect competition in some sectors generate higher gains from trade reform, although there is the possibility of the opposite outcome if reform induces resources to move back into an agricultural sector that has sufficiently fewer economies of scale than the rest of the economy.21 VI. IMPLICATIONS FOR DEVELOPING COUNTRIES The good news is that there are still large potential gains from liberalizing merchandise trade under Doha, with global gains on the order of $95 billion– $120 billion a year from an agreement consistent with the Hong Kong Minister- ial Declaration, even if no reforms are forthcoming in services, and with a disproportionately high share of that potential gain available for developing countries (relative to their share of the global economy). Moreover, it is the poorest people in developing countries who appear most likely to gain from global trade liberalization—farmers and unskilled laborers.22 To realize that potential gain, by far the greatest cuts in bound tariffs and subsidies are required in agriculture. However, the political sensitivity of farm support programs, coupled with the complexities of the measures introduced in the Uruguay Round Agreement on Agriculture and the modalities set out in the 21. An example is the study by Francois, van Meijl, and van Tongeren (2005), whose 50 percent global reform scenario yields only a 0.5 percent global income gain despite economies of scale, imperfect competition, and variety effects, and with agriculture contributing less to those gains than in the results here because farming, unlike other sectors, is assumed to be subject to constant returns. 22. For detailed analyses of the poverty consequences of these Doha scenarios, see the study by Hertel and Winters (2006). Anderson, Martin, and van der Mensbrugghe 193 Ministerial Declaration, ensures that achieving consensus on the details of the final Doha agreement will remain challenging. Outlawing agricultural export subsidies is the obvious first step. That will help bring agriculture into line with other sectors and limit the extent to which governments encourage agricultural production by other means (as it would remove one option for surplus disposal). Concurrently, domestic support bindings must be cut substantially to reduce binding overhang. Even more important, agricultural tariff bindings must be slashed so that genuine market opening can occur. Allowing lesser cuts for even a few ‘‘sensi- tive’’ and ‘‘special’’ farm products would greatly reduce the gains from reform, given the tariff peaks currently in place. If it turns out to be politically impos- sible not to designate some products as sensitive and special, the resulting welfare cost could be reduced by imposing a tariff cap at, say, 100 percent. Expanding market access for nonagricultural products at the same time as reforming agriculture would increase the prospects for a successful conclusion to the Doha Round. An essential part of the Doha development agenda is ‘‘concessions’’ between developing countries because that is where half their potential benefits lie. That means reconsidering the extent to which developing countries liberalize. Because developing countries trade so much more with each other now than in the 1980s, they are the major beneficiaries of reforms within their own regions. Even the least developed countries need to consider reducing at least their tariff binding overhang, because doing that in the context of the Doha Round gives them more scope to demand concessions (or compensation for preference erosion or other contributors to terms of trade deterioration) from richer countries than if they hang on to the opportunity not to engage in reform, as provided for in the July Framework Agreement. What ultimately emerges from the analysis is that developing countries would not have to reform very much following the Doha Round because of the large gaps between their tariff bindings and applied rates. But to realize more of the potential gains from trade, they would need to commit to addi- tional trade reforms and complementary domestic reforms and to invest more in trade facilitation. High-income countries could encourage them to do so not only by opening up their markets to developing country exports but also by providing more targeted aid. A new proposal has been put forward to reward developing country commitments to greater trade reform with an expansion of trade-facilitating aid through a major expansion of the Integrated Framework, now operated by a consortium of international agencies for the least developed countries (Hoekman and Prowse 2005; Nielson 2006). This may provide an attractive path for developing countries seeking to trade their way out of poverty. In addition, it is potentially a far more efficient way for developed countries to assist people in low-income countries than the current systems of tariff preferences. 194 THE WORLD BANK ECONOMIC REVIEW, VOL. 20, NO. 2 REFERENCES Alexandraki, K., and H. P. Lankes. 2004. ‘‘The Impact of Preference Erosion on Middle-Income Coun- tries.’’ IMF Working Paper 04/169. International Monetary Fund, Washington, D.C. Anderson, K., and E. Valenzuela. Forthcoming. ‘‘Do Global Trade Distortions Still Harm Developing Country Farmers?’’ Review of World Economics 143(1). Anderson, K., W. Martin, and E. Valenzuela. Forthcoming. ‘‘The Relative Importance of Global Agri- cultural Subsidies and Market Access.’’ World Trade Review 5(3). Anderson, K., W. Martin, and D. van der Mensbrugghe. 2006. ‘‘Market and Welfare Implications of the Doha Reform Scenarios,’’ In K. Anderson and W. Martin, eds., Agricultural Trade Reform and the Doha Development Agenda. New York: Palgrave Macmillan and the World Bank. ———. Forthcoming-a. ‘‘Distortions to World Trade: Impacts on Agricultural Markets and Incomes.’’ Review of Agricultural Economics 28(2). ———. Forthcoming-b. ‘‘Would Multilateral Trade Reform Benefit Sub-Saharan Africa?’’ Journal of African Economies 15. Baffes, J. 2005. ‘‘The ‘Cotton Problem.’ ’’ World Bank Research Observer 20(1):109–43. Bagwell, K., and R. Staiger. 2002. The Economics of the World Trading System. Cambridge, Mass.: MIT Press. Boue ´, S. Jean, and D. Laborde. 2004. ‘‘A Consistent ad valorem Equivalent ¨ t, A., Y. Decreux, L. Fontagne Measure of Applied Protection across the World: The MAcMap-HS6 Database.’’ Centre d’Etudes Prospectives et d’Informations Internationales, Paris. Boue ´, and S. Jean. 2006. ‘‘Is Erosion of Preferences a Serious Concern?’’ In K. Anderson ¨ t, A., L. Fontagne and W. Martin, eds., Agricultural Trade Reform and the Doha Development Agenda. New York: Palgrave Macmillan (co-published with the World Bank). Feldstein, M., and C. Horioka. 1980. ‘‘Domestic Saving and International Capital Flows.’’ Economic Journal 90(358):314–29. Francois, J., H. van Meijl, and F. van Tongeren. 2005. ‘‘Trade Liberalization in the Doha Development Round.’’ Economic Policy 20(42):349–91. Harrison, G. W., T. F. Rutherford, D. G. Tarr, and A. Gurgel. 2004. ‘‘Trade Policy and Poverty Reduction in Brazil.’’ World Bank Economic Review 18(3):289–317. Hertel, T., ed. 1997. Global Trade Analysis: Modeling and Applications. New York: Cambridge Uni- versity Press. Hertel, T. W., and R. Keeney. 2006. ‘‘What’s at Stake: The Relative Importance of Import Barriers, Export Subsidies and Domestic Support.’’ In K. Anderson and W. Martin, eds., Agricultural Trade Reform and the Doha Development Agenda. New York: Palgrave Macmillan and the World Bank. Hertel, T. W., and L. A. Winters, eds. 2006. Poverty and the WTO: Impacts of the Doha Development Agenda. New York: Palgrave Macmillan and the World Bank. Hoekman, B., F. Ng, and M. Olarreaga. 2004. ‘‘Agricultural Tariffs versus Subsidies: What’s More Important for Developing Countries?’’ World Bank Economic Review 18(2):175–204. Hoekman, B., and C. O ¨ zden. 2005. ‘‘Trade Preferences and Differential Treatment of Developing Countries: A Selective Survey.’’ Policy Research Working Paper 3566. World Bank, Washington, D.C. Hoekman, B., and S. Prowse. 2005. ‘‘Policy Responses to Preference Erosion: From Trade as Aid to Aid for Trade.’’ Paper presented at the World Bank Conference on Preference Erosion: Impacts and Potential Policy Responses, June 13–14, Geneva. Hummels, D., and P. Klenow. 2005. ‘‘The Variety and Quality of a Nation’s Exports.’’ American Economic Review 95(3):704–23. Jean, S., D. Laborde, and W. Martin. 2006. ‘‘Consequences of Alternative Formulas for Agricultural Tariff Cuts.’’ In K. Anderson and W. Martin, eds., Agricultural Trade Reform and the Doha Devel- opment Agenda. New York: Palgrave Macmillan and the World Bank. Anderson, Martin, and van der Mensbrugghe 195 Kehoe, T., and K. Ruhl. 2003. ‘‘How Important Is the New Goods Margin in International Trade?’’ Federal Reserve Bank of Minneapolis Staff Report 324. Minneapolis, Minn. Lluch, C. 1973. ‘‘The Extended Linear Expenditure System.’’ European Economic Review 4(1):21–32. Nielson, J. 2006. ‘‘Aid for Trade.’’ In R. Newfarmer, ed., Trade, Doha and Development: A Window into the Issues. Washington, D.C.: World Bank. Olarreaga, M., and C. O ¨ zden. 2005. ‘‘AGOA and Apparel: Who Captures the Tariff Rent in the Presence of Preferential Market Access?’’ The World Economy 28(1):63–87. ¨ zden, C., and G. Sharma. 2004. ‘‘Price Effects of Preferential Market Access: The O CBI and the Apparel Sector.’’ Policy Research Working Paper 3244. World Bank, Washington, D.C. Sumner, D. A. 2006. ‘‘Reducing Cotton Subsidies: The DDA Cotton Initiative.’’ In K. Anderson and W. Martin, eds., Agricultural Trade Reform and the Doha Development Agenda. New York: Palgrave Macmillan and the World Bank. van der Mensbrugghe, D. 2005. ‘‘The Doha Development Agenda and Preference Erosion: Modeling the Impacts.’’ Paper presented at the World Bank Conference on Preference Erosion: Impacts and Potential Policy Responses, June 13–14, Geneva. ———. 2006. ‘‘Estimating the Benefits: Why Numbers Change.’’ In R. Newfarmer, ed., Trade, Doha and Development: A Window into the Issues. Washington, D.C.: World Bank. Ventura, J. 2003. ‘‘Towards a Theory of Current Accounts.’’ The World Economy 26(4):483–512. Wacziarg, R., and J. S. Wallack. 2004. ‘‘Trade Liberalization and Intersectoral Labor Movements.’’ Journal of International Economics 64(2):411–39. World Bank. 2002. Global Economic Prospects and the Developing Countries 2002: Making Trade Work for the Poor. Washington, D.C. ———. 2004. Global Economic Prospects: Realizing the Development Promise of the Doha Agenda. Washington, D.C. WTO (World Trade Organization). 2003a. ‘‘Negotiating Group on Market Access: Report by the Chairman.’’ TN/MA/12 (The Girard Text). Geneva. ———. 2003b. ‘‘Negotiations on Agriculture: First Draft of Modalities for the Further Commitments.’’ TN/AG/W/1/Rev.1 (The Harbinson Draft). Geneva. ———. 2004. ‘‘Decision Adopted by the General Council on 1 August 2004.’’ WT/L/579 (The July Framework Agreement). Geneva. ———. 2005. ‘‘Doha Work Program: Draft Ministerial Declaration: Revision.’’ Ministerial Conference, Sixth Session, December 13–18, Hong Kong, China. WT/MIN(05)/W/3/Rev.2. Geneva.