34437 Global Economic Prospects 2005 Trade, Regionalism, and Development Global Economic Prospects Trade, Regionalism, and Development 2005 © 2005 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org 1 2 3 4 07 06 05 04 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed herein do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the International Bank for Reconstruction and Development / The World Bank any judgement on the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. 0-8213-5747-6 Cover artist: John Yanson Contents Foreword vii Acknowledgments ix Overview xi Abbreviations xix Frequently Cited Regional Trading Agreements and the Parties to Them xxiii Chapter 1 Global Outlook and the Developing Countries 1 The Global Economy: From Recovery to Expansion 3 Commodity Markets 6 World Trade 8 International Finance 10 Risks and Policy Ppriorities 12 Long-Term Growth, Structural Change, and Poverty 14 Structural Changes over Two Decades 16 Structural Change in the Future 19 Poverty Forecast 21 Concluding Remarks 23 Notes 23 References 25 Chapter 2 Regional Trade and Preferential Trading Agreements: A Global Perspective 27 The Proliferation of Regional Preference Systems 28 Trends in Trade and Growth by Region 42 Changing Export Composition and the Rise of Global Production Networks 46 Preferential Trade and Regional Outcomes 49 Conclusion 53 Notes 53 References 54 Chapter 3 Regional Trade Agreements: Effects on Trade 57 The Impact of RTAs on Merchandise Trade and Incomes 57 Ingredients of Success 66 iii G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Conclusions: Preferential Trade Agreements and Economic Development 72 Notes 74 References 74 Chapter 4 Beyond Trade Policy Barriers: Lowering Trade Costs Together 77 The Costs of Trade 78 Regional Agreements to Facilitate Trade and Transport 79 Standards, Conformity Assessments, and RTAs 85 Trade-Related Regional Cooperation Agreements 91 Conclusions 92 Notes 93 References 94 Chapter 5 Beyond Merchandise Trade: Services, Investment, Intellectual Property, and Labor Mobility 97 Services, Investment, and IPRs in Regional Agreements 98 Economic Consequences of Services, Investment, and IPR Provisions in RTAs 104 RTAs and Provisions for Movement of Labor 111 Conclusions: Beyond Merchandise Trade 117 Notes 119 References 121 Chapter 6 Making Regionalism Complementary to Multilateralism 125 Preferential Agreements within the Global Context 126 Building Blocks versus Stumbling Blocks 133 The Competitive Liberalization Game: The Case of Doha 136 Multilateral Disciplines on Regional Arrangements 140 Making Open Regionalism Work for Development 144 Notes 148 References 150 Figures 1.1 A record year for developing countries in 2004 4 1.2 Strong growth across most regions 5 1.3 Tradable price developments 6 1.4 Terms of trade impacts from higher commodity prices, 2001­04 7 1.5 The oil-important burden for selected countries 7 1.6 Crude oil prices, 1960­2004 8 1.7 World trade rebounds 8 1.8 Export performance, percent change in market share since 2000 9 1.9 Trade balances in major regions 10 1.10 Developing countries' debt and interest payments easing downward since 1999 10 1.11 Rising U.S. dollar reserves 11 1.12 Impact of a 200 basis point increase in interest rates 12 1.13 First year impacts of a $10 increase in oil prices 14 1.14 A rise in services 17 1.15 Rising openness to trade 18 iv C O N T E N T S 1.16 Increased urbanization 19 1.17 Growth rate of labor supply declining 19 2.1 Number of RTAs exploded in the 1990s 29 2.2 Spaghetti and rigatoni: Multiple, overlapping RTAs, 2004 39 2.3 Trade within RTAs 41 2.4 Trade performance has differed across regions 43 2.5 Composition of trade has changed 47 2.6 Evolving trading blocs 50 3.1 Evolution of the share of intra-regional imports in total imports, 1960­2000 58 3.2 The ratio of external and intra-regional trade to GDP 59 3.3 Intra-regional trade grows faster when world trade growth is positive 60 3.4 Simulated welfare impact of various FTAs involving Chile 62 3.5 RTAs that divert imports tend to export less to global markets 64 3.6 Not all regions are open 66 3.7 Preferential tariffs in tandem with all tariffs in Latin America 68 3.8 Rules of origin in North­South agreements are more restrictive than in South­South agreements 70 4.1 More efficient customs are associated with more trade 80 5.1 Share of EU nationals in total population, for selected EU countries and Norway, 1985­2000 112 6.1 Global outcomes dominate alternatives 131 6.2 Global reform dominates North­South agreements 133 6.3 Global reform dominates South­South agreements 133 Tables 1.1 The Global outlook in summary 2 1.2 Current account balances 11 1.3 Long-term prospects: Forecast growth of world GDP per capita 17 1.4 Labor market structure, 2005­15 20 1.5 Regional breakdown of poverty in developing countries 21 2.1 Most countries belong to more than one RTA 30 2.2 RTAs cover many topics besides merchandise trade 35 3.1 Implementation of tariff commitments by type of agreement, 1960­1999 69 5.1 Services, investment, and intellectual property: A comparison 99 5.2 Additional services liberalization in U.S. FTAs 100 5.3 Summary of agreements by degree of labor mobility 111 6.1 Comparison of bilateral agreements to global trade reform 128 6.2 Comparison of bilateral agreements with global trade reform 129 Boxes 1.1 The aggregation paradox 16 2.1 RTAs and types of trade liberalization 28 2.2 Reporting RTAs to the WTO 29 2.3 EPAs become the EU's trade and development instrument: An experiment in "North-South-South" integration 32 2.4 Labor in U.S. FTAs 33 v G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 2.5 Trade agreements and the environment 36 2.6 Can RTAs prevent conflict? 38 2.7 Regional versus multilateral and unilateral liberalization: What's more important? 42 3.1 A primer on modeling of RTAs 61 3.2 Regional trade agreements in gravity models: A meta-analysis 63 3.3 Implementation matters 65 3.4 Restrictive rules of origin under NAFTA--the case of clothing 71 3.5 Monitoring implementation of preferential trade agreements: "Single Market Scoreboard" in the European Union 73 4.1 Trading can be costly 79 4.2 Border delays tax trade 80 4.3 Standardization and simplification can increase trade volumes: The case of the Trans-Kalahari Corridor 82 4.4 Logistics costs in Europe have fallen in the last two decades 83 4.5 The case of the Northern Corridor Stakeholders Consultative Forum 84 5.1 Not all investment is good investment 106 5.2 Do more investor protections mean more investment? Lessons from bilateral investment treaties 107 5.3 Illegal migration: A growing global phenomenon 113 5.4 U.S. temporary admission programs under NAFTA and unilateral policies 116 6.1 Impacts of the new GTAP database 127 6.2 Regional trade agreements, structural change, and congruence 132 6.3 Choosing partners: Selection criteria for U.S. RTAs 137 6.4 Sequencing of RTAs: Is there a good practice? 139 6.5 RTAs and WTO disciplines 141 6.6 Tunisia's Association Agreement with the European Union 146 vi Foreword T HIS YEAR--2004--is shaping up to be the healthiest year for developing countries in the last three decades. East Asia has come out of the crisis of 1997­98 stronger and more vibrant than ever; the countries of Europe and Central Asia are now almost completely out of the long shadow of transition from socialism and are growing more rapidly; and South Asian coun- tries, on the strength of continuing reforms, are performing well. Moreover, countries in Latin America, Sub-Saharan Africa, and the Middle East and North Africa had a much better year. To be sure, some countries within each region have not enjoyed the fruits of this recovery, and these countries remain a source of concern. But in aggregate, economic growth in 2004 was impressive. This performance reflects a fortuitous combination of (1) long-term secular trends built on a foun- dation of better macroeconomic management and (2) an improved domestic investment climate converging with a cyclical recovery of the global economy. This is no time for complacency. Lingering imbalances in the global economy associated with the rising twin deficits in the United States, a delayed recovery in Europe, high and volatile oil prices, and questions about the path of China's economy constitute risks to the pace of growth in developing countries over the medium term. Sustaining this pace is essential for the fate of mil- lions of the world's poor. World trade grew by 10.2 percent in 2004, and has played an important role in this year's ex- emplary performance. On the policy side, the WTO discussions in August recouped the ground that was lost in late 2003 (after the Cancun WTO ministerial), when governments concluded an agreement that provides a framework for pursuing the Doha Development Agenda. The frame- work, however, is only an outline--the actual agreement is still to come. Whether the final agree- ment contains provisions that will provide a meaningful impulse to development remains to be seen. In the meantime, this year's Global Economic Prospects examines the evidence on an impor- tant development that is reshaping the architecture of the world trading system: The dramatic pro- liferation of regional trade agreements (RTAs). These take various forms of preferential reciprocal treaties--they can be bilateral or plurilateral free trade agreements or, less commonly, customs unions. In according preferential access to members, regional arrangements necessarily discrimi- nate against nonmembers. Even though arrangements in some instances can promote develop- ment, it is important to recognize that they also can lead to trade diversion in a way that hurts both member countries and excluded countries. Hence this year's report identifies ways to design and implement preferential trading agreements to maximize their benefits for participants and mini- mize their costs to nonmember developing countries. The key to making regional agreements com- plementary to a nondiscriminatory multilateral system is to strive for "open regionalism"--that vii G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 is, agreements with low external barriers to trade, nonrestrictive rules of origin, liberalized service markets, and a strong focus on reducing transaction costs at borders. The international community working together can leverage the August framework to achieve an ambitious Doha deal. Multilateral liberalization has a greater positive impact on development than do the myriad regional arrangements now being spawned seemingly in every corner of the globe. Moreover, multilateral oversight of inherently discriminatory RTAs must be strengthened, and the first step is to increase transparency by empowering the WTO to collect and regularly make public full details of all arrangements. If Doha can succeed in bringing down border pro- tection in agriculture and manufactures, it will reduce the discriminatory effects of regional agree- ments and lower the probability of costly trade diversion for participating countries. Said differ- ently, a strong Doha arrangement can contribute to open regionalism. Along with other bilateral and multilateral institutions, the World Bank continues to promote the integration of developing countries into the world economy so that the benefits of globaliza- tion can extend to the poor. The Bank now has 91 trade-related projects approved or planned in 75 countries for the three-year fiscal period 2004­06. The actual and projected commitments for new trade operations are at $2.9 billion--larger than the commitments of all ongoing operations approved over the preceding eight-year period of fiscal 1996­2003 ($2.4 billion). For trade fa- cilitation the growth is even larger. Projected commitments over fiscal 2004­06, at more than $1.2 billion, triple the commitments of ongoing trade facilitation operations approved between fiscal 1996­2003. To guide this lending and to provide policy advice, the World Bank research program in trade works for client countries all over the world, and the program continues to be ambitious. There is still a lot to be done. Poverty remains high--with 2.7 billion people living on less than $2 dollars per day--and the global trading system is still riddled with obstacles that prevent the products of the world's poor from reaching markets. Pursuing these challenges through multilat- eral, unilateral, and regional policies can contribute to poverty reduction around the world, which will pay high dividends for generations to come. François Bourguignon Chief Economist World Bank November 2004 viii Acknowledgments T HIS REPORT WAS produced by staff from the World Bank's Prospects Group and from the World Bank Trade Department. Richard Newfarmer was the lead author and manager of the report, under the direction of Uri Dadush. The principal authors of chapter 1 were Andrew Burns and Dominique van der Mensbrugghe. Chapter 2 was written by Paul Brenton, Carolina Diaz Bonilla, Denis Medvedev, and Philip Schuler; chapters 3 and 4 were written by Paul Brenton; chapter 5 was written by Richard Newfarmer; and chapter 6 was written by Bernard Hoekman, Jeffrey Lewis, and Dominique van der Mensbrugghe. Carolina Diaz Bonilla and Denis Medvedev provided research assistance that informed all chapters. The Global Outlook--the on- line twin publication of Global Economic Prospects--was led by Fernando Martel Garcia. Several reviewers, including Michael Finger, Julio Nogues, Guillermo Perry, and Jeffrey Schott, as well as Alan Winters and Jaime De Melo, offered extensive and enormously useful comments. The report was produced under the general guidance of François Bourguignon. Several people contributed substantively to the various chapters. In chapter 1, the Global Trends Team, under Hans Timmer's leadership, was responsible for the projections, with contri- butions from John Baffes, Maurizio Bussolo, Betty Dow, Himmat Kalsi, Robert Keyfitz, Annette De Kleine, Fernando Martel Garcia, Donald Mitchell, Mick Riordan, and Shane Streifel, and for the poverty numbers, Shaohua Chen and Martin Ravallion. For chapter 2, Vlad Manole, Will Martin and Francis Ng, Sherman Robinson and Carolina Diaz Bonilla, Maurice Schiff and Yanling Wang provided background papers and notes. For chapter 3, Souleymane Coulibaly pro- vided statistical analysis for the gravity model work, and Takako Ikezuki contributed background material. In chapter 4, Enrique Aldaz-Carroll, Gael Raballand, and Luc de Wulf wrote back- ground papers. In chapter 5, Ximena Clark and William Shaw, Carsten Fink and Patrick Reichenmiller, Howard Mann and Aaron Cosbey, Aaditya Mattoo, and Beatriz Nofal provided background papers. Maurice Schiff and Alan Winters wrote papers for chapter 6. The Global Outlook benefited from the efforts of several specialists, among them Abhijeet Dwivedi, Sabeera Kulkarni, Bulent Ozbilgin, and Carlos Rossel. Awatif H. Abuzeid was the senior staff assistant, who, together with Katherine Rollins, produced the several drafts of the report and managed the writing stage. Dorota Nowak worked with the World Bank's Office of the Publisher to manage the publication and dissemination process. Many people provided written contributions and/or comments: Frederick Abbott, Hakim Ben Hamouda and staff at the Economic Commission for Africa, Carlos Braga, Harry Broadman, Aaron Cosbey, Robert Devlin and the staff of the Inter-American Development Bank, Peter Draper, Dorothy Dworskin and the staff at the U.S. Treasury and U.S. Trade Representative; ix G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Philip English, Antoni Estevadeordal, Alan Gelb, Caroline Freund, Mary Hallward-Driemeier, Lawrence Hinkle, Elena Ianchovichina, Roumeen Islam, Alejandro Jara, Sebastien Jean, Stephen Karingi, Odin Knudsen, Hans Peter Lankes, Patrick Low, Nuno Limao, Muthukumara Mani, Howard Mann, Daniel Muller-Jentsch, Mustafa Nabli, Caglar Ozden, John Panzer, Carmen Pont-Vieira, Jean-Francois Rischard, Kamal Saggi, Maurice Schiff, Raj Soopramanien, Peter Van Den Heuvel and staff of the EU Commission, Stephen Woolcock, and Gianni Zanini. x Overview T HE PROLIFERATION OF regional trade in control of outcomes. Moreover, RTAs have agreements (RTAs) is fundamentally al- the flexibility to pursue trade-expanding tering the world trade landscape. The policies not addressed well in multilateral trad- number of agreements in force now surpasses ing rules. Trade agreements therefore usually 200, and it has risen sixfold in just two go beyond slashing tariffs to include measures decades. Today more than one-third of global to reduce trade impediments associated with trade takes place between countries that have standards, customs and border crossings, and some form of reciprocal RTA.1 The European services regulations--as well as broader rules Union (EU) and United States are playing a that improve the overall investment climate. prominent role in this proliferation (figure 1). Finally, these agreements often form corner- This report addresses two questions: stones of larger economic and political efforts to increase regional cooperation. RTAs can · What are the characteristics of agree- help motivate and reinforce broader reforms in ments that strongly promote--or domestic policy; they can be designed to con- hinder--development for member coun- tribute to a political environment that is more tries? conducive to stability, investment, and growth. · Does the proliferation of agreements Not all agreements create new trade and pose risks to the multilateral trading investment. Those RTAs with high external bor- system, and how can those risks be der protection are particularly susceptible to the managed? adverse effects of trade diversion (figure 2). In fact, a statistical analysis based on findings from several econometric studies suggests that many Identifying What Works: Open agreements cost the economy more in lost trade Regionalism revenues than they earn, because they discrimi- R TAs are often one component of a larger nate against efficient, low-cost suppliers in non- political effort to deepen economic rela- member countries. Of course, this finding does tions with neighboring countries.2 As such, not take into account the potential dynamic they can create opportunities to expand trade gains, the positive effects associated with ser- through joint action to overcome institutional vices liberalization, or any of the benefits from as well as policy barriers to trade. At a basic adopting new regulations. But it does under- level, it is often easier to motivate reciprocal re- score the point that regional agreements carry ductions in border barriers when the partici- risks that merit close scrutiny by would-be pants are fewer and the policymakers feel more participants. xi G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Figure 1 Regionalism spreads a.The number of RTAs exploded in the 1990s b. EU and U.S. agreements were most important Annual numbers Total numbers Percent of world trade covered 40 350 35 Cumulative number of agreementsa South- 30 (right axis) 300 30 South 25 Cumulative number of agreementsb 250 25 (right axis) U.S. Agreements not 20 200 20 notified to the WTO (left axis) 15 150 15 Agreements notified 10 100 10 to the WTO EU (left axis) 5 50 5 0 0 0 1958 1969 1976 1984 1989 1994 1999 2004 1990 1996 2002 a. EU-25 counted as single country. b. EU-15 counted as single country. Source: World Trade Organization. Figure 2 External protection can differ among agreements a. Only a few agreements have low external tariffs b. And so intraregional trade in some cases grows at the expense of extraregional trade SACU NAFTA NAFTA EU AFTA CACM ECOWAS Overall Overall imports SADC Intraregional trade GCC exports AFTA EAC CAN CEMAC MERCOSUR MERCOSUR SADC COMESA SAPTA CIS ECOWAS EAC WAEMU SAPTA COMESA 0 5 10 15 20 25 4 2 0 2 4 6 8 10 Average weighted tariffs Estimated exponential impact on trade Note: In chart a, tariffs are import-weighted at the country level to arrive at RTA averages. In chart b, the bars show the magnitude of the dummy variables, capturing respectively the extent to which intraregional trade, overall imports, and overall exports differ from the "normal" levels predicted by the gravity model on the basis of economic size, proximity, and relevant institutional and historical variables, such as a common language. Source: World Bank staff using UN TRAINS, accessed through WITS. As agreements proliferate, a single country average Latin America country belongs to often becomes a member of several different seven agreements. This creates a "spaghetti agreements. The average African country bowl" of overlapping arrangements (figure 3). belongs to four different agreements, and the Each agreement has different rules of origin, xii O V E R V I E W Figure 3 RTAs can complicate customs administration a. African agreements are overlapping b. More efficient customs are associated with more trade Nile River Ratio of trade to GDP (percent) AMU Basin IGAD ECCAS CEMAC Somalia 250 Algeria Libya Sao Tomé & Principe Morocco Mauritania Cameroon Egypt Tunisia Central African Rep. Gabon ECOWAS 200 Equat. Guinea COMESA Djibouti Conseil de Rep. Congo Burundi Ethiopia L'Entente Ghana Rwanda Cape Verde Eritrea Nigeria Chad Gambia Sudan DR Congo Kenya 150 Benin Niger Uganda Togo Burkina Faso Côte d'Ivoire Angola R2 0.338 Mali Guinea-Bissau Senegal EAC 100 Liberia Guinea Sierra Leone Mauritius WAEMU Tanzania Malawi Syechelles 50 Zambia Comoros SACU Zimbabwe Mano River Union Madagascar CILSS South Africa Namibia Botswana Swaziland Reunion Lesotho 0 0 5 10 15 20 25 CBI IOC Mozambique Days through customs (imports) SADC Sources: Chart a, Schiff and Winters (2003). Chart b, Investment Climate Surveys data and Global Trends as cited in Subramanian and others (2003). different tariff schedules, and different periods With prerequisites in place, the RTAs most of implementation, and together they compli- likely to increase national incomes over time cate customs administration. Customs agents are those designed with: report that it takes longer to process goods covered by preferential arrangements, and · Low external MFN tariffs, longer processing times drive up the cost of · Few sectoral and product exemptions, trade. In general, the longer the delays in cus- · Nonrestrictive rules-of-origin tests that toms, the smaller the role of trade in GDP. build toward a framework common to So what characteristics lead to expanded many agreements, trade and development? A prerequisite for · Measures to facilitate trade, the success of any trade policy is that it be · Large ex-post markets, integrated into a sound domestic policy frame- · Measures to promote new cross-border work. It is virtually impossible for entrepre- competition, particularly in services, and neurs to take advantage of new opportunities-- · Rules governing investment and intellec- whether they originate in market access tual property that are appropriate to the through an RTA, through a multilateral agree- development context. ment, or other sources--if the domestic invest- ment climate is not supportive. Macroeco- Low external tariffs and wide coverage nomic stability, basic property rights, and minimize the risks of trade diversion, while adequate infrastructure regulation are all key. nonrestrictive rules of origin allow for in- Indeed, trade agreements can reinforce positive creased trade. The practice of excluding many elements in the domestic reform program by agricultural products is common, and it can anchoring policy to the agreement itself. But limit development payoffs. Trade facilitation an RTA cannot substitute for sound domestic measures, though worthwhile in and of them- policies. selves, receive more policymaker attention xiii G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 when they are embedded in an RTA, and they often have positive trade-creating effects for Figure 4 Rules of origin in North-South agreements are more restrictive than in all trade partners. South-South agreements Well designed agreements are of limited value if they are not implemented, and many Index of restrictiveness RTAs have more life on paper than in reality. 6 Weak implementation often afflicts South- 5 South agreements. Monitoring mechanisms 4 are often inadequate and do not receive the 3 sustained high-level political attention neces- sary to drive institutional improvements in, 2 for example, adherence to tariff reduction 1 schedules, customs, and border crossings. 0 Against these benchmarks of success, it is A A xico CM AS difficult to give universally high marks to any AFT W NAFT SADC EU-Chile COMESAECO single category of agreement. In general, EU-Me Chile-CA North-South agreements score better on im- Note: Higher values of the index equals to more restrictive plementation than South-South agreements. rules of origin derived from Estevadeordal and Suominen Because North-South agreements can inte- (2004). grate economies with distinct technological capabilities and other different factor propor- tions, and because they usually result in larger economic similarity. And like the North-South post-agreement markets, the potential gains agreements, South-South agreements rarely are usually greater. However, tighter rules of provide for the temporary movement of labor. origin, more restrictive exclusions for particu- lar sectors (such as agriculture), and a preoc- cupation with rules not calibrated to develop- Consequences for the Multilateral ment priorities can undercut these benefits System T (figure 4). North-South agreements, particu- he development consequences of RTAs are larly those with the United States, have been not limited to their effects on members-- more effective in locking in new services liber- they also have cumulative effects on the multi- alization; they have pressed intellectual prop- lateral system. In one sense, RTAs are a step to- erty rights beyond World Trade Organization ward greater openness in the whole system, by (WTO) rules; and expanded the sphere of in- promoting more trade and generating new do- vestment protections; but they contain few mestic constituencies with an interest in open- provisions to liberalize the temporary move- ness. Moreover, some regional trade policies ment of labor. are effectively nondiscriminatory, such as mea- Some South-South agreements are better sures to improve customs, speed transactions at at focusing on merchandise trade, minimiz- ports or border crossings, or in some cases open ing exclusions, adopting less restrictive services markets. These measures can comple- rules of origin, and lowering the border costs. ment unilateral and multilateral policies. For example, the Caribbean Community However, this view overlooks the effects (CARICOM) and the Common Market of that RTAs can have on excluded countries. Eastern and Southern Africa (COMESA) have Preferences for some countries mean discrimi- had some success in reducing border costs. But nation against others. Indeed, the General in general, South-South agreements have not Agreement on Tariffs and Trade (GATT), adhered to implementation schedules, and borne out of the sad experience of discrimina- they suffer from their small market size and tion in the prewar years, was founded on the xiv O V E R V I E W principle of nondiscrimination. Today, the consequence of the spread of regional agree- adverse consequences for the excluded coun- ments is that many poorer developing coun- tries are much less severe than at GATT's tries have diverted scarce negotiating re- inception, because tariffs and other barriers sources to regional negotiations at the expense have come down sharply, mitigating the ex- of more active participation in the Doha dis- clusionary effects of regional arrangements. cussions. The average developing country be- The exception--and it is not trivial--is agri- longs to five separate RTAs and is negotiating culture. Another mitigating factor is that more all the time. In the future, will countries many countries excluded by trade agreements that now enjoy preferences fight multilateral between the United States and the EU enjoy liberalization, or even oppose further regional some degree of preferential access through liberalization, to keep their privileged market voluntary preference schemes, such as the access? A few small developing countries are Generalized System of Preferences (GSP), indeed likely to lose advantages in preferential America's Growth and Opportunity Act markets, and they may scuttle a deal if their le- (AGOA), and the EU's Everything But Arms gitimate concerns are not addressed. (EBA) program. To be sure, these programs lack the certainty of market access that MFN agreements and RTAs provide, because prefer- The Importance of Doha to Open ences are voluntary and subject to political Regionalism whim, but they do mitigate the effects of ex- he policy solution to these twin con- clusions for selected, very low-income coun- T cerns--the need to design regional agree- tries. Finally, some developing countries--the ments that create trade and regional agree- spokes in the hub-and-spoke analogy--are ments that have minimal exclusionary signing bilateral agreements with each other effects--comes together in the form of low and with other hubs. MFN tariffs and other border barriers. An Inevitably some countries get left out of agreement that lowers border protection trade agreements, either because they are not around the world promotes open regionalism favored politically, because they cannot afford by mitigating trade diversion. At the same the costs of many separate negotiations, or be- time, it would diminish the exclusionary ef- cause their neighborhood is less open. Coun- fects of discriminatory preferences built into tries as diverse as Bolivia, India, Mongolia, regional agreements. The first order of busi- Pakistan, and Sri Lanka do not enjoy the same ness for the international community is to ac- level of access to the United States or the EU celerate progress on the Doha Agenda and to as Chile, Jordan, or Mexico, and they see their fill in the blanks of the August 2004 frame- trade diminished when bilateral agreements work agreement with reductions in protec- are signed. tion, especially for products produced by the RTAs can also undercut the incentives of world's poor. governments to press for multilateral liberal- ization, which would improve global trade rules. This study finds little evidence that major players in the current WTO negotia- For Developing Countries, tions have changed their negotiating positions a Three-Part Strategy D or retreated from the multilateral process, eveloping countries wishing to harness even as they avail themselves of regional trade trade to their development strategy deals. However, as the discussions become should see regional integration as one element politically difficult, the risk is ever present that in a three-pronged strategy that includes unilat- even they will abandon multilateralism in eral liberalization, multilateral liberalization, favor of "satisficing regionalism." One and regional liberalization. xv G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Historically, unilateral liberalization, which with increases in intraregional trade, irrespec- is usually linked to a broader program of do- tive of the presence of an RTA. mestic reform, has accounted for most of the Multilateral liberalization leverages domes- reductions in border protection. Most com- tic reforms into increased market access prehensive trade reforms among large coun- around the world. Developing countries col- tries (Argentina, Brazil, and China in the early lectively stand to gain much more in the WTO 1990s, and more recently, India) were primar- arena than in any smaller regional market. ily unilateral reforms that were undertaken to Moreover, this multilateral forum is the only increase the productivity of the domestic econ- place that developing countries, working to- omy. The same process took place in many gether, can press for more open markets in small countries as well. In fact, of the 21 per- agriculture and can seek disciplines on trade- centage point cuts in average weighted tariffs distorting agricultural subsidies and on con- of all developing countries between 1983 and tingent protection. 2003, unilateral reforms account for roughly Some have argued that RTAs can be an al- two-thirds of the reduction. Tariff reductions ternative to multilateral liberalization. They associated with the multilateral commitments are not. Gains for all developing countries in the Uruguay Round accounted for about from these agreements, even under the most 25 percent, and the proliferation of regional generous of assumptions, are usually only a agreements amounted to about 10 percent of fraction of those from full multilateral liberal- this reduction (see figure 5). ization. Of course, if one of the partner coun- Autonomous liberalization promotes global tries is a high-income, large-market economy, competitiveness by lowering costs of inputs, in- and if most other countries are excluded from creasing competition from imports to drive pro- preferential access, the countries signing the ductivity growth, and integrating the national first trade agreement may benefit individually economy into the global economy. Autonomous and substantially--but those benefits wither trade reform is, ironically, more important than as new countries sign additional agreements. ever in the presence of RTAs; low border barri- In fact, the scenarios in this study show that ers minimize the risks of trade and investment all developing countries would collectively diversion. Low external barriers promote trade lose if they were all to sign preferential agree- in world markets, and this is highly correlated ments with the Quad (Canada, the EU, Japan, and the United States) (figure 6). Therefore, developing countries have a powerful collec- tive interest in an effective Doha Agenda-- Figure 5 Share of total tariff reduction, by even if they all are scrambling to gain prefer- type of liberalization, 1983­2003 ential market access to the Quad. Forging policies on open regionalism is the Regional Agreements third component of trade policy strategy. De- 10% sirable as multilateral liberalization is, the Doha Round is likely to realize only part of its development potential. For some types of pol- Multilateral icy, collective regional actions may be the first, Agreements 25% best course, and may result in effective nondis- criminatory benefits.3 For example, RTAs can Autonomous reduce regional political tensions, take advan- Liberalization 66% tage of scale economies in infrastructure pro- vision, and lead to joint programs to improve Source: Martin and Ng 2004. border crossings or to motivate liberalization in services. But countries should sign on with xvi O V E R V I E W players in the system, they have a special inter- Figure 6 Multilateral liberalization is far est in--and responsibility for--using effective more beneficial than RTAs multilateral reforms to discipline the discre- Change in real income in 2015 compared to baseline tionary aspects of the regional agreements. Allowing developing countries to concen- BILAT Middle-income JBIL trate scarce negotiating resources on the mul- Global tilateral agenda may require that high-income Low-income countries decelerate their efforts at expanding RTAs. Irrespective of the pace of new agree- Developing ments, high-income countries could consider the following rules of thumb when designing Quad agreements to promote development. First, re- ducing the extensive exclusions for agriculture 1.5 1.0 0.5 0 0.5 1.0 1.5 2.0 Percent would transfer the income gains to rural areas in participating developing countries. Second, Note: Global refers to the global merchandise trade reform scenario; JBIL corresponds to the simulation where all adopting more common and nonrestrictive developing countries sign bilateral agreements with the rules of origin across agreements would Quad-plus countries; and BILAT corresponds to the simulation where the bilateral agreements are signed reduce the administrative barriers that often individually. Results reflect unweighted regional averages. undermine agreements and that increase the Source: World Bank simulations with the Linkage model and burden on customs administration. Third, GTAP release 6.04. working with prospective partners to ensure that new regulations regarding investment and intellectual property are appropriate to the their eyes wide open. The lessons of this study level of development would reduce risks of (and others before it4) are that, much as with undue enforcement costs. Finally, providing unilateral or multilateral policies, design and trade-related technical assistance, not only in implementation determine the ultimate effects. the implementation phase but also in the ne- It is important to use trade policy to leverage gotiating phase, would promote greater liber- domestic reforms that promote growth. For alization of services and lower MFN tariffs. South-South agreements, it is essential that the focus be on some combination of full trade liberalization behind low external border protection, greater services deregulation and Acting Collectively to Mute the competition, and proactive trade facilitation Effects of Discrimination T measures that together positively affect both o minimize the discriminatory effects of intra- and extra-regional trade. RTAs at the multilateral level, all countries must assume greater responsibility for main- taining the multilateral system. The interna- High-Income Countries and tional community, working through the WTO, Development should revisit Article V of its charter. If the H igh-income countries, in order to realize stated disciplines cannot be enforced in the their broad development objectives, must near term for collective political reasons, then intensify their efforts to realize the develop- increasing transparency and information ment promise of the Doha Agenda. This has should become a priority. At present, the WTO the potential to open up trade, particularly in collects little if any information updating spe- agriculture, in a way that would benefit low- cific provisions, their implementation, and the income groups around the world. Because trade consequences. It even fails to take ad- the high-income countries are the large vantage of extant public monitoring efforts in xvii G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 specific regions, which could inform their data consequences of regional agreements for trade collection effort. Collecting and publishing creation (chapter 3), trade facilitation specific information on RTAs would allow (chapter 4), and services, investment, intellec- members that find themselves excluded to tual property rights, and labor mobility challenge these agreements in the court of pub- (chapter 5). Chapter 6 returns to the issue of lic opinion. Even the more modest goal of making regional agreements more compatible transparency will require building a new con- with a nondiscriminatory multilateral system. sensus and providing the staff of the WTO with more resources than they have currently Notes available. 1. Negotiated as bilateral or multicountry treaties, Nonetheless, WTO members should con- regional trade agreements grant members assured pref- sider enhancing the existing rules to ensure erential market access, usually at zero tariffs for eligi- that regional agreements have positive devel- ble products. Following WTO convention, the term opment and systemic outcomes. This could in- "regional trade agreement" includes both reciprocal clude (based on a modest tightening of current bilateral free trade or customs areas and multicountry practice) setting quantitative indicators that (plurilateral) agreements. These are distinct from non- reciprocal voluntary agreements, such as the general- define "substantially all trade." It could in- ized system of preferences (GSP). Also, for statistical clude efforts to simplify and harmonize the purposes, unless otherwise noted, intra-EU trade is ex- rules of origin that are applied to both devel- cluded from quantitative trade analysis. The EU is de- oped and developing countries. These items fined as including the 15 countries that belonged to the are on the Doha Agenda and may be ready for union before its enlargement in 2004. action. 2. See Devlin and Estevadeordal (2004) and Schiff and Winters (2003), among others. 3. See Robert Lawrence (1997), who develops the idea of subsidiarity as applied to regional agreements. Organization of This Study 4. See Schiff and Winters (2003). A s is customary, chapter 1 of this study pre- sents the World Bank's view of the global economy. The short-term section analyzes the References main forces shaping the global outlook and the Devlin, Robert and Antoni Estevadeordal. 2004. Trade implications for developing countries; the long- and Cooperation: A Regional Public Goods Ap- term analysis focuses on structural changes proach. In Regional Public Goods: From Theory to Practice, eds. A. Estevadeordal, Brian Frantz, in the global economy that will affect poverty and Tam Robert Nguyen. Washington, DC: Inter- rates and the prospects for attaining the Millen- American Development Bank. nium Development Goals. A novel feature of Lawrence, Robert. 1997. Preferential Trading Arrange- this year's report is the introduction of a com- ments: The Traditional and the New. In Regional panion online feature (see www.worldbank. Partners in Global Markets: Limits and Possibili- org/prospects), where the reader can find addi- ties of the Euro-Med Agreements, eds. A. Galal tional information on regional trends and com- and B. Hoekman. London: Center for Economic Policy Research/Egyptian Center for Economic modity prices, and tools to design scenarios to Studies. his or her own specifications. Martin, W., and F. Ng. 2004. Sources of Tariff Reduc- Chapter 2 introduces the issues associated tions. Background paper. with regional trade agreements and provides Schiff, Maurice, and L. Alan Winters. 2003. Regional an overview of regional trading trends. Integration and Development. Washington, DC: Subsequent chapters focus on the content and World Bank. xviii Abbreviations ACP African Caribbean and Pacific states ACPEU African Caribbean and Pacific states European Union AFTA ASEAN Free Trade Area AGOA African Growth and Opportunity Act ANZCERTA Australia-New Zealand Closer Economic Relations Trade Agreement APEC Asia Pacific Economic Cooperation ASEAN Association of Southeast Asian Nations BITS Bilateral investment treaties CAFTA Central America Free Trade Agreement CARICOM Caribbean Community CEC Commission for Environmental Cooperation CEMAC Economic and Monetary Community of Central Africa CEPR Center for Economic and Policy Research CGE Computable general equilibrium CIS Commonwealth of Independent States COMESA Common Market for Eastern and Southern Africa CRTA Committee on Regional Trade Agreement EAC East African Community EBA Everything but arms EC European Community ECO Economic Cooperation Organization ECOWAS Economic Community of West African States EEC European Economic Community EFTA European Free Trade Association EPAs Economic Partnership Agreements EU European Union FDI Foreign direct investment xix G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 FTAA Free Trade Area of the Americas GAO General Accounting Office GATS General Agreement on Trade in Services GATT General Agreement on Tariffs and Trade GCC Gulf Cooperation Council GDP Gross domestic product GSP Generalized System of Preferences GTAP Global Trade Analysis Project HS Harmonized system IDB Inter-American Development Bank IFC International Finance Corporation IISD International Institute for Sustainable Development IMF International Monetary Fund INS Immigration and Naturalization Services IOM International Organization for Migration IPR Intellectual Property Rights IRCA Immigration and Regularization Control Act IRPA Immigration and Refugee Protection Act LAFTA Latin America Free Trade Area LDCs Least developed countries MDGs Millennium Development Goals MERCOSUR Southern Lone Common Market MFN Most favored nation MRA Mutual recognition agreement NAFTA North America Free Trade Agreement NBER National Bureau of Economic Research OECD Organisation for Economic Co-operation and Development PRSP Poverty Reduction Strategy Paper PTAs Preferential trade agreements RTAs Regional trade agreements SAARC South Asia Association for Regional Cooperation SACU South African Customs Union SAD Single administrative document SADC Southern African Development Community SAPP Southern African Power Pool SAFTA South Asian Free Trade Area SAPTA South Asia Preferential Trade Agreement SPS Sanitary and phyto-sanitary standards xx A B B R E V I A T I O N S TBT Technical barriers to trade TFP Total factor productivity TRIM Trade-related investment measures TRIPS Trade-related aspects of intellectual property rights TTF Transport and trade facilitation UEMOA/WAEMU West African Economic and Monetary Union UNCTAD United Nations Conference for Trade and Development UNESCAP United Nations Economic and Social Commission for Asia and the Pacific USAID United States Agency for International Development USTR United States Trade Representative WCO World Customs Organization WIPO World Intellectual Property Organization WTO World Trade Organization xxi Frequently Cited Regional Trading Agreements and the Parties to Them Agreement Full name Members AFTA ASEAN Free Brunei, Darussalam, Cambodia, Indonesia, Lao Trade Area People's Democratic Republic, Malaysia, Myanmar, Philippines, Singapore, Thailand, Vietnam APEC Asia Pacific Economic Australia, Brunei, Canada, Chile, China, Cooperation Hong Kong (China), Indonesia, Japan, Korea, Malaysia, Mexico, New Zealand, Papua New Guinea, Peru, Philippines, Russia, Singapore, Taiwan (China), Thailand, United States, Vietnam CACM Central American Costa Rica, El Salvador, Guatemala, Honduras, Common Market Nicaragua CAFTA Central America Free United States, Costa Rica, El Salvador, Guatemala, Trade Area Honduras, Nicaragua, Dominican Republic CAN Andean Community Bolivia, Colombia, Ecuador, Peru, República Bolivariana de Venezuela CARICOM Caribbean Community Antigua and Barbuda, Bahamas, Barbados, Belize, and Common Market Dominica, Grenada, Guyana, Haiti, Jamaica, Monserrat, Trinidad and Tobago, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Suriname CEFTA Central European Bulgaria, Czech Republic, Hungary, Poland, Free Trade Agreement Romania, Slovak Republic, Slovenia CEMAC Economic and Monetary Cameroon, Central African Republic, Chad, Community of Republic of Congo, Equatorial Guinea, Gabon Central Africa CER Closer Economic Australia, New Zealand (ANZCERTA) Relations Trade Agreement xxiii G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Agreement Full name Members CIS Commonwealth of Azerbaijan, Armenia, Belarus, Georgia, Moldova, Independent States Kazakhstan, Russian Federation, Ukraine, Uzbekistan, Tajikistan, Kyrgyz Republic COMESA Common Market for Angola, Burundi, Comoros, Democratic Republic of Eastern and Congo, Djibouti, Arab Republic of Egypt, Eritrea, Southern Africa Ethiopia, Kenya, Madagascar, Malawi, Mauritius, Namibia, Rwanda, Seychelles, Sudan, Swaziland, Uganda, Zambia, Zimbabwe EAC East African Community Kenya, Tanzania, Uganda ECOWAS Economic Community Benin, Burkina Faso, Cape Verde, Gambia, Ghana, of West African States Guinea, Guinea-Bissau, Côte d'Ivoire, Liberia, Mali, Niger, Nigeria, Senegal, Sierra Leone, Togo EEA European EU, Iceland, Liechtenstein, Norway Economic Area EFTA European Free Iceland, Liechtenstein, Norway, Switzerland Trade Association EMFTA Euro-Mediterranean EU, Algeria, Cyprus, Egypt, Israel, Jordan, Lebanon, Free Trade Area Malta, Morocco, Syrian Arab Republic, Tunisia, Turkey, Palestinian Authority FTAA Free Trade Area Antigua and Barbuda, Argentina, Bahamas, of the Americas Barbados, Belice, Bolivia, Brazil, Canada, Chile, Colombia, Costa Rica, Dominica, Dominican Republic, Ecuador, El Salvador, Grenada, Guatemala, Guyana, Haiti, Honduras, Jamaica, Mexico, Nicaragua, Panama, Paraguay, Peru, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Suriname, Trinidad and Tobago, United States, Uruguay, Venezuela GAFTA Greater Arab Free Bahrain, Egypt, Iraq, Jordan, Kuwait, Lebanon, Trade Area Libya, Morocco, Oman, Palestine, Qatar, Saudi Arabia, Somalia, Sudan, Syria, Tunisia, United Arab Emirates, Yemen GCC Gulf Cooperation Bahrain, Kuwait, Oman, Qatar, Saudi Arabia, Council United Arab Emirates MERCOSUR Southern Common Argentina, Brazil, Paraguay, Uruguay Market NAFTA North American Canada, Mexico, United States Free Trade Agreement SACU Southern African South Africa, Botswana, Lesotho, Swaziland, Customs Union Namibia xxiv R E G I O N A L T R A D I N G A G R E E M E N T S A N D T H E P A R T I E S T O T H E M SADC Southern African Angola, Botswana, Democratic Republic of Congo, Development Lesotho, Malawi, Mauritius, Mozambique, Community Namibia, South Africa, Swaziland, Seychelles, Tanzania, Zambia, Zimbabwe SAFTA South Asian Free Trade Bangladesh, Bhutan, India, Maldives, Nepal, Area Pakistan, Sri Lanka SAPTA South Asian Preferential Bangladesh, Bhutan, India, Maldives, Nepal, Trade Arrangement Pakistan, Sri Lanka WAEMU West African Economic Benin, Burkina Faso, Côte d'Ivoire, Guinea Bissau, and Monetary Union Mali, Niger, Senegal, Togo xxv 1 Global Outlook and the Developing Countries World growth accelerated sharply in 2004, and in Europe, budgetary policy is expected to with GDP advancing an estimated 4 percent tighten as countries seek to regain control (table 1.1). All developing regions are now over deficits, which in many cases exceed growing faster than their average growth rates Maastricht limits. Finally, efforts in China to of the 1980s and 1990s. The ongoing eco- bring growth down to a more sustainable pace nomic boom in China was a major factor, as should also contribute to weaker, but still were the surges in activity registered in Japan strong, demand over the medium term. and the United States. The economic recovery Given this external environment and espe- was slower to take hold among European cially the less rapid expansion of trade, high-income countries, which contributed to growth in most low- and middle-income coun- the less marked increase in growth rates there. tries is also expected to moderate but remain Meanwhile, very strong import demand-- strong. The extent of the slowdown should be because of the torrid expansion in China and mitigated because of the far-reaching struc- the continued tendency for domestic demand tural reforms carried out in many countries, in the United States to substantially exceed which have contributed to recent gains in mar- production--contributed to an exceptional ket share and economic growth. Recent efforts 10.2 percent increase in world trade volumes. to reduce general government and current ac- Economic growth is expected to slow in count deficits and to pay down debt should 2005 and 2006, expanding by 3.2 percent in enable most developing countries to withstand each year. Several factors are likely to con- the higher interest rates expected over the next tribute to this more moderate pace of activity. few years without excessive adjustment First, the investment cycle in the United States costs. However, there is little room for com- has likely peaked, implying a slowdown in placency--especially for the more highly growth there.1 Second, world demand has indebted countries. outstripped supply, resulting in substantial in- These favorable prospects for the next two creases in oil and other commodity prices that years represent a solid starting point for have cut into incomes, moderating demand in longer-term growth through 2015 and increase many countries. Third, higher interest rates the likelihood that developing countries meet will slow investment growth as central banks the Millennium Development Goals (MDGs). continue shifting monetary policy from a Improvements in macroeconomic fundamen- loose to a more neutral stance. Fourth, the tals, enhanced structural flexibility, a stronger large fiscal impulse that has helped propel the investment climate, and further progress to- U.S. economy in recent years will weaken in ward reducing trade barriers should, if 2004--although the deficit will remain high; sustained, support the ability of developing 1 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Table 1.1 The global outlook in summary Percentage change from previous year, except interest rates and oil prices Forecast 2002 2003 2004e 2005 2006 Global Conditions World Trade Volume 3.7 5.5 10.2 8.4 7.8 Consumer Prices G-7 Countriesa,b 1.0 1.6 1.7 1.4 1.2 United States 1.6 2.3 2.7 2.2 1.7 Commodity Prices (USD terms) Non-oil commodities 5.3 10.2 17.0 3.1 4.2 Oil Price (World Bank average)c 24.9 28.9 39.0 36.0 32.0 Oil price (percent change) 2.4 15.9 35.0 7.7 11.1 Manufactures unit export valued 1.3 7.4 5.2 0.8 0.3 Interest Rates $, 6-month (percent) 1.8 1.2 1.6 3.5 4.7 , 6-month (percent) 3.3 2.3 2.1 2.4 3.6 Real GDP growthe World 1.7 2.7 4.0 3.2 3.2 Memo item: World (PPP weights)f 2.9 3.9 4.9 4.2 4.1 High income 1.3 2.1 3.5 2.7 2.7 OECD Countriesg 1.3 2.0 3.5 2.6 2.6 Euro Area 0.9 0.5 1.8 2.1 2.3 Japan 0.3 2.5 4.3 1.8 1.6 United States 1.9 3.0 4.3 3.2 3.3 Non-OECD countries 2.2 3.1 5.9 4.6 4.4 Developing countries 3.4 5.2 6.1 5.4 5.1 East Asia and Pacific 6.7 7.9 7.8 7.1 6.6 Europe and Central Asia 4.6 5.9 7.0 5.6 5.0 Latin America and the Caribbean 0.6 1.6 4.7 3.7 3.7 Middle East and North Africa 3.2 5.7 4.7 4.7 4.5 South Asia 4.6 7.5 6.0 6.3 6.0 Sub-Saharan Africa 3.1 3.0 3.2 3.6 3.7 Memorandum items Developing countries excluding transition countries 3.2 5.1 5.9 5.4 5.1 excluding China and India 2.1 3.8 5.4 4.6 4.3 Note: PPP purchasing power parity; e estimate. a. Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States. b. In local currency, aggregated using 1995 GDP weights. c. The World Bank average is the unweighted mean of one barrel of West Texas Intermediate, Brent, and Dubai oil. d. Unit value index of manufactured exports from major economies, expressed in U.S. dollars. e. GDP in constant dollars at 1995 prices and market exchange rates. f. GDP measured at 1995 PPP weights. g. Now excludes the Republic of Korea, which has been reclassified as high-income OECD. Source: World Bank. countries to achieve rapid and sustained per Africa, will fail to reduce poverty to this capita growth at a level of 3.5 percent per degree. In Sub-Saharan Africa, per capita annum between 2006 and 2015--double the growth has been slow, and progress to reduce growth rate of the 1990s. Such growth would poverty has been minimal. It would take im- enable many developing countries to halve the plausibly high growth rates during the next incidence of extreme poverty by 2015, which is 10 years to achieve the poverty target along a key development goal. However, even if the with substantial enhancements to pro-poor higher growth of recent periods were sus- policies and significantly more assistance. tained, some regions, notably Sub-Saharan Finally, even if many regions are expected to 2 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S achieve the MDG to reduce poverty, many are failure to rein in the U.S. budget deficit, which off track for reaching other important MDGs, would also tend to reduce its current account such as reducing child and maternal mortality. deficit, could result in an ever increasing stock In many cases economic growth is not enough. of dollar-denominated debt and rising future A more targeted approach and a realignment financing burdens. Moreover, higher interest of spending priorities are also necessary. rates would depress investment levels, provok- Despite the relatively positive picture for ing a prolonged slowing in the rate of increase both medium- and long-term prospects, of potential output. All of these factors downside risks are ever present and could heighten the risk of a resurgence in protection- have negative impacts in the near future and in ist sentiment, which would thwart the pace at the long term. An additional rise in oil prices, which developing countries are able to achieve or a failure of them to moderate, could further their poverty reduction objectives. restrain global demand and reduce incomes in Finally, if current efforts to slow the unsus- most less developed countries. While oil prices tainable pace of growth in China fail, major are expected to decline from present highs, disruptions could result. Currently, investment especially given substantial efforts to increase levels may be unsustainably high, and there supply by oil exporting countries, existing are some signs that rapidly rising food-price demand conditions are such that a significant increases are feeding into production costs, increase cannot be ruled out. Such a rise which could ultimately choke off competitive- would have important negative effects on all ness, (although for the moment there are no oil-importing economies, particularly those clear indications that this is happening). Either of low- and middle-income countries that problem could provoke a much more abrupt face current account constraints. For these slowdown than described in the baseline. countries, difficulties accessing international Given China's growing importance as a driver finance mean that they cannot absorb the in- of world trade growth, such a sharp slow- creased costs associated with higher oil prices down could have a significant damping effect by increasing their current account deficit. on global economic activity, particularly Instead, the additional costs must be accom- among China's major trading partners. modated by lower imports, consumption, and investment volumes--implying a significant real-side adjustment. For the most vulnerable The Global Economy: From of such countries, an additional $10 a barrel Recovery to Expansion T increase in oil prices could reduce domestic in- he world economy accelerated sharply in comes by as much as 4 percent. On average, 2004, expanding by an estimated 4 per- incomes of oil-importing low-income countries cent (figure 1.1). The United States and Japan, would fall by about 1 percent of GDP. whose economies grew by more than 4 percent, Financing requirements of the U.S. current continued to lead Europe in the recovery. Even account and government deficits, and renewed stronger growth was experienced by a number downward pressure on the dollar, may cause of large developing countries, notably China long-term interest rates to rise more than fore- (8.8 percent), Russia (8.0 percent), and India casted. If interest rates rise, short- to medium- (6.0 percent). Their performance helped power term impacts might include a slowing in world developing countries as a whole to an antici- economic growth, sharply increased financing pated 6.1 percent growth rate in 2004--an costs, and economic hardship for heavily in- expansion without precedent over the past debted countries. Increased financial-market 30 years. Moreover, it marks a second year of turbulence might also ensue--especially for very strong growth, and it may be the first time those developing countries most exposed to the that recovery in developing countries preceded, U.S. dollar. Over the medium- to long-term, rather than followed, recovery in high-income 3 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 is under way in Latin America, and there are Figure 1.1 A record year for developing signs of a more modest recovery in Sub- countries in 2004 Saharan Africa. GDP growth, 1995 US$ (percent) Growth should moderate in 2005 and 6 2006, led by a slowing of the expansion among Developing countries World developed countries. In the United States, as 5 the output gap closes, productivity growth is 4 projected to slow and unit labor costs to rise; these factors, in addition to external inflation- 3 ary pressures from commodity prices, likely re- flect the Fed's decision to tighten monetary 2 conditions. This, plus the maturation of the in- 1 vestment cycle, a tailing off of fiscal stimulus, and the impact of higher oil costs, will con- 0 tribute to slowing growth. Similar factors ex- 1980 1983 1986 1989 1992 1995 1998 2001 2004 plain the anticipated slowdown in Japan, where output is expected to increase at about Source: World Bank. trend rates. In contrast, because of its later start and the fact that investment is only now beginning to recover, Europe's growth is countries. In contrast to the United States, expected to continue gaining momentum where the surge was initially led by investment through 2005 and into 2006, notwithstanding and household consumption, exports were the fiscal tightening and a slowdown in the rate of main source of growth in Europe and Japan-- growth of world demand. Overall estimates and much of the increase in external demand suggest that the hike in oil prices already ob- came from developing countries. served can be expected to dampen output in Across the developing world virtually 2005 by about 0.5 percent of GDP. every region enjoyed solid growth, and Moderating growth in the OECD economies rapidly rising trade volumes played an impor- and a soft landing in China should translate into tant role. Even excluding China, India, and slower but still buoyant growth in developing Russia, economic activity in developing coun- countries (figure 1.2). tries is expected to have risen 5 percent in 2004. While easy credit contributed to · In East Asia, efforts to stem the flow of China's remarkable performance, the benefits credits into selected sectors of the Chi- of WTO accession were also a major factor, nese economy are already having observ- and the increase of over 30 percent in Chinese able effects (figure 1.2a). The growth of import demand helped underpin growth imports of raw materials such as steel, among neighboring East Asian countries. copper, and various ores have moderated Russia and the oil-producing countries in the significantly in recent months. Steel im- Middle East and North Africa Region bene- ports have collapsed, although iron ore fited from very strong oil revenues, which import volumes were growing by more were reflected in strong import demand and than 25 percent (year/year) in September. the solid export performance of their trading However, there are indications that partners. Increasing market shares, following consumption demand continues to grow substantial inward investment flows associ- rapidly, and the Chinese authorities ated with the accession of many of the Europe report that GDP increased 9.1 percent in and Central Asian Region's members to the the third quarter. The baseline forecast EU, also contributed to these positive out- predicts that a soft landing (growth comes. Elsewhere, a strong cyclical recovery slowing to 7.1 percent by 2006) will be 4 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S Figure 1.2 Strong growth across most regions a. East Asia and Pacific b. South Asia Growth of real GDP (percent) Growth of real GDP (percent) Forecast Forecast 9 9 8 8 7 7 6 6 5 5 4 4 3 3 2 2 1 1 0 0 1980­2002 2003 2004* 2005 2006 1980­2002 2003 2004* 2005 2006 c. Europe and Central Asia d. Middle East and North Africa Growth of real GDP (percent) Growth of real GDP (percent) Forecast Forecast 8 8 7 7 6 6 5 5 4 4 3 3 2 2 1 1 0 0 1980­2002 2003 2004* 2005 2006 1980­2002 2003 2004* 2005 2006 e. Latin America and the Caribbean f. Sub-Saharan Africa Growth of real GDP (percent) Growth of real GDP (percent) Forecast Forecast 5.0 5.0 4.5 4.5 4.0 4.0 3.5 3.5 3.0 3.0 2.5 2.5 2.0 2.0 1.5 1.5 1.0 1.0 0.5 0.5 0 0 1980­2002 2003 2004* 2005 2006 1980­2002 2003 2004* 2005 2006 Note: * estimate. Source: World Bank estimates. 5 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 achieved and will contribute to slowing and higher payments on debt (see the risks throughout the region. section in this chapter). Here, country- · In South Asia, despite the moderation of specific conditions and the degree to the Chinese and OECD economies, which fiscal consolidation programs are growth is expected to accelerate in 2005, maintained will play an important role. reflecting the enduring impacts of struc- · Sub-Saharan Africa will also benefit tural reforms, market opening, and from the revival in Europe, its main trad- stronger domestic demand as the damp- ing partner, but many oil-importing ening impact of last year's poor crop countries in Africa remain vulnerable fades. As agricultural production and re- due to high oil prices. Notwithstanding lated incomes return to trend growth substantially improved performance, rates in 2006, GDP growth is projected growth in the region will continue to lag to moderate somewhat. the rest of the world by a significant mar- · Output in Europe and Central Asia is fore- gin, implying a further widening of in- cast to remain strong, with still-high oil come gaps. Moreover, the terms of trade prices supporting demand in Russia and appear to be turning against this region the exports of its trading partners. Central as non-oil commodity prices are ex- and Eastern European countries will con- pected to ease. Although additional de- tinue to benefit from rapid investment velopment aid and debt relief would growth following the EU accession of help, continued efforts to improve fun- some of their members. However, policy- damentals and the efficiency of public makers need to prepare for the next down- expenditure are also required to speed turn by pursuing fiscal consolidation to re- the pace at which these countries achieve duce worryingly high government and, in their poverty-reduction objectives. some cases, current account deficits. · Growth in the Middle East and North Commodity Markets Africa region is expected to remain robust, trong world demand and supply shortages but well below the highs observed in 2003, S were responsible for commodity prices re- which were boosted by sharp increases in bounding sharply during the global recovery oil production. All countries, but espe- (figure 1.3). In dollar terms, metals and minerals cially those of the Maghreb, should benefit from the strengthening export demand emanating from Western Europe; but con- Figure 1.3 Tradable price developments sumption demand, reflecting still high oil Cumulative percent change, 2001­2004 incomes, will continue to be the main 60 source of growth for the region as a whole. 50 Nominal · The return to growth in Latin America and 40 the Caribbean is projected to continue, Exchange-rate 30 adjusteda with only Argentina experiencing a signif- icant slowdown as the competitive advan- 20 tage from its depreciation in 2002 wears 10 off. Elsewhere, growth should remain 0 EUN strong, with Brazil expanding steadily at 10 between 3.7 and 3.9 percent. Because Petroleum Metals and Agricultural Manufactures minerals products Latin America and the Caribbean is a heavily indebted region, outturns will ulti- a. The adjusted data show the price change expressed as a trade-weighted average of domestic currency prices. mately depend on the success with which Source: World Bank. policymakers deal with rising interest rates 6 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S developing countries was 5.6 percent of GDP, Figure 1.4 Terms of trade impacts from whereas for oil importers the impact was a higher commodity prices, 2001­04 loss of 0.3 percent. Percent of GDP For the poorest oil-importing countries, 10 Oil exporters high oil prices have dramatically exacerbated 8 already serious poverty. Many of these countries remain particularly vulnerable to 6 high oil prices. Even before the oil price hikes, 4 a number of these countries were spending 2 more than 5 percent of GDP to cover oil im- Oil importers 0 ports. The unweighted average of West-Texas Intermediate, Brent, and Dubai crude oils is 2 estimated to have been $39 in 2004.3 At this and and w income w income income level, it is estimated that as many as seven Lo India) incomeLo w indebted Lo income w w indebted Lo Lo countries will have oil-import bills in excess of Highly Highly middle middle (excluding 10 percent of GDP; these countries would be forced to make substantial cuts in spending Source: World Bank. elsewhere in their economies to compensate for the additional burden (figure 1.5). Indeed, prices have increased the most since 2001 (up al- for the poorest countries the net additional most 60 percent), but the 40 percent hike in burden in 2004 is expected to consume 75 petroleum prices has had the largest economic percent of the World Bank funding they re- effect. In domestic currency terms, the impact of ceive for all development programs, and these price hikes was less important for many countries because of the 15 percent deprecia- tion2 of the dollar over the same period. Figure 1.5 The oil-import burden for Higher commodity prices since 2001 have selected countries boosted incomes of low- and middle-income countries as a whole by an estimated 1.1 per- Ethiopia cent of GDP. However, virtually all of the gain Kyrgyzstan accrued to low- and middle-income oil ex- Burkina Faso porters. Most developing country oil im- Malawi porters suffered net terms of trade losses (fig- Pakistan ure 1.4). The major beneficiaries were the Mali Middle East and North Africa, Europe and Nicaragua Central Asia, and Latin America and the Moldova Caribbean Regions--all of which include Tajikistan major oil exporters. In contrast, the net gains Bosnia and from non-oil commodity prices for low- Herzegovina income countries were modest or even nega- Swaziland tive. This is partly because most of the non-oil Mongolia Pre-hike burden commodity price gains were concentrated Lesotho Additional burden in metals and minerals prices, which restricted Mauritania the benefits to a few resource-rich countries. Lao PDR Moreover, many industrializing low-income 0 5 10 15 20 25 countries, notably India and Pakistan, are Percent of GDP now net commodity importers. The terms-of- Source: World Bank. trade impact on incomes of oil exporting 7 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 localized disruptions in production (fig- Figure 1.6 Crude oil prices, 1960­2004 ure 1.6). Indeed, OPEC excess capacity is esti- Price per barrel ($) mated to have fallen from 4.6 million barrels 80 per day in 2001 to only 1.4 million barrels per 70 Constant dollarsa day in 2004. Moreover, oil prices remain well 60 below past peaks. Corrected for inflation and 50 expressed in 2003 dollars, oil prices averaged 40 more than $72 in 1980, and actually reached 30 more than $100 in November of the previous 20 year. Viewed from this perspective, further 10 Nominal dollars hikes would not be unprecedented. 0 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 a. Nominal petroleum prices deflated by the U.S. GDP World Trade deflator, 2003 1. W Source: World Bank. orld trade growth averaged 10.2 per- cent in 2004, reflecting rapid increases in industrial production and investment activ- ity (figure 1.7). The expansion in trade vol- 12 percent of all the bilateral aid they receive. umes in 2004 is reminiscent of the increase ob- To keep development projects on track, high- served in 2000 and mirrors the rapid recovery income countries will need to increase com- in industrial production that began to take mitments substantially--at least as long as shape in the second half of 2003 and contin- high oil prices continue. ued into 2004. More than 20 percent of the Although a substantial rise in oil prices is increase in world merchandise trade volumes not the most likely scenario, given new was represented by China, whose imports in- sources of supply and reduced oil intensities in creased by 32 percent--reflecting both the the world economy, there remains consider- positive impact of its accession to the WTO able scope for higher oil prices, particularly and unsustainable rates of investment and given the current sensitivity of oil markets to consumption demand. Figure 1.7 World trade rebounds World industrial production (percent change, year/year) Merchandise export volumes (percent change, year/year) 7.5 15.0 Industrial production World trade 12.5 5.0 10.0 7.5 2.5 5.0 2.5 0 0 2.5 2.5 5.0 5.0 7.5 Q1 Q1 Q1 Q1 Q1 Q1 Q1 1991 1993 1995 1997 1999 2001 2003 Source: World Bank. 8 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S Trade in raw materials and investment goods East and North Africa) or lost market share was particularly strong. As discussed above, ro- (Sub-Saharan Africa and Latin America and bust demand for raw materials was an impor- the Caribbean). tant factor underlying the trade expansion in a Within regions the performance of specific number of developing countries. In particular, countries continues to be dictated, in part, by oil, steel, and minerals trade was strongly influ- domestic factors. So notwithstanding very enced by the rapid increase in Chinese manu- strong Chinese import demand, exports in the facturing and construction sectors. Similarly, rest of East Asia failed to increase as quickly, fast-growing global investment expenditures partly because political instability held back were particularly important in spurring export industrial and investment activity in the demand in countries such as Germany and Philippines and Indonesia. In Latin American Japan that specialize in the fabrication of ma- and the Caribbean, export volumes in Brazil chinery and other physical capital. and Argentina grew briskly under the contin- As a whole, developing countries have ued influence of currency devaluations 2 years grown their share in world markets by about ago, while strong world demand for metals 19 percent (figure 1.8), up from 19 to 23 per- and minerals gave special impetus to Chilean cent since 2000. Much of this rise is attributed exports. to China, which has seen its share in world Slower activity throughout the global econ- exports double from 2.9 to 5.8 percent be- omy should translate into less rapid trade ex- tween 2000 and 2004. Excluding China, the pansion in 2005 and 2006. Trade in goods and improvement in the export share of low- and nonfactor services is forecast to expand by middle-income countries has been more mod- about 8.5 percent in 2005, down from an esti- est (from 16 to 17 percent), although develop- mated 10 percent in 2004. Much of the decel- ing countries in the South Asia and Europe eration is conditional on the success of efforts and Central Asia regions have increased their to dampen the pace of activity in China, which market shares considerably. Other regions should be reflected in slower import growth in either maintained their market share (the rest China and slower exports among its trading of the Eastern Asia and Pacific and the Middle partners. Looking to other regions, the easing of activity in the United States, coupled with broadly stable growth in Europe, is expected to Figure 1.8 Export performance, percent result in a somewhat more pronounced decel- change in market share since 2000 eration of trade volumes in Latin America as compared with Africa, the Middle East, and Percent China, percent Eastern European areas. 60 120 Major imbalances in the world trading en- 50 100 vironment persisted during 2004 and will 40 Developing 80 East Asia countries likely continue to play a large role in 2005­06 30 (excluding China) 60 (excluding China) (figure 1.9). Notwithstanding the sharp accel- 20 40 eration in world import volumes, the U.S. cur- 10 20 rent account deficit reached 5.7 percent of 0 0 GDP in the second quarter of 2004, as Amer- 10 20 ican consumption and investment volumes ex- 20 Developed 40 China countries ceeded domestic production by a wide margin 30 60 (higher oil prices represented 0.6 percentage 40 80 points of the 1.4 percentage point deteri- 1989199019911992199319941995199619971998199920002001200220032004 oration in the current account since the first quarter of 2002). The expansion in the trade Source: World Bank. deficit since the mid-1990s has been the main 9 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 imports from other developing countries with Figure 1.9 Trade balances in major whom China has a cumulatively large trade regions deficit (Lau 2003).4 Billions of U.S. dollars, 2004 Failure to address the twin U.S. deficits could 300 260 have significant impacts on developing coun- 200 170 134 tries, especially if that failure leads to an increase 100 52 60 9 in protectionist behavior. This is especially rele- 0 vant because the substantial improvements in 100 living standards, wages, and incomes in many 200 upper-lower and middle-income countries have 300 been the result of expanding their world market 400 500 share in manufactures. An increase in protec- 600 tionism could halt these countries' progress and 700 667 deny other poor countries the same avenue to 800 development. Moreover, a retreat from recent a A pan ica efforts to reduce trade barriers or a failure to States China NAFT Union Ja Ameribbean Other veloping make further progress--especially concerning United Other LatinCar de agricultural subsidies--could have substantial European the and negative consequences on many of the world's poorest countries. a. European Union refers to the 15 member countries, prior to the latest expansion. Source: National agencies. International Finance O ver the past several years, favorable global conditions, strong growth, rapidly factor behind the rise in the U.S. current ac- expanding trade, and domestic reforms (in- count deficit--itself a major factor behind the cluding lower fiscal deficits and inflation) have 15 percent real effective depreciation of the allowed developing countries to substantially currency since February 2002. Barring a sub- improve their financial positions (figure 1.10). stantial increase in domestic savings by, for ex- ample, a tightening of fiscal policy, downward pressure on the U.S. dollar is likely to resume Figure 1.10 Developing countries' debt as U.S. foreign borrowing requirements re- and interest payments easing downward main high, and the already large amounts of since 1999 external debt continue to accumulate (see, for Percent of GNI Percent of GNI example, Bergsten and Williamson 2004). 60 3.0 The U.S. trade deficit is largely a home- Total debt, outstanding and disbursed (left axis) grown problem. While bilateral trade deficits 50 Interest payments 2.5 (right axis) with specific countries are large, notably with 40 2.0 respect to China, the fact that these countries have only small overall surpluses supports the 30 1.5 view that the deficit with the United States is 20 1.0 more a reflection of U.S. trade patterns than an indication of unfair trading practices. For ex- 10 0.5 ample, China's large bilateral surplus with the 0 0 United States (but very small global surplus) reflects its specialization in the production of 19901991199219931994199519961997199819992000200120022003 final consumption goods (sold to the United Source: World Bank. States) based on intermediate and primary 10 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S On average, their debt to GNI ratio has fallen Table 1.2 Current account balances from 44 to 37 percent since its peak in 1999. Percent of GDP in 2004 This progress, plus low interest rates and strong growth, has substantially lowered the East Asia and Pacific 1.5 South Asia 0.5 debt-servicing burden for most countries. Middle East and North Africa 14.4 While the situation of the most heavily in- Sub-Saharan Africa 1.3 debted countries remains serious, they have Europe and Central Asia 0.1 Latin America and the Caribbean 0.7 made the greatest gains--debt to GDP ratios High-income countries 0.8 for these countries are down from 161 to 86 percent since 1994--partly because of debt- Source: World Bank estimates. relief programs instituted over this period. These favorable conditions have also al- private investors' retreat from equity and bond lowed many countries to strengthen their ex- investments in U.S. dollar-denominated assets,5 ternal position. Most countries have succeeded the central banks of these countries have become in improving their structural positions so that, one of the most important sources of financing even in the face of higher oil prices or a more for the large U.S. current account deficit, absorb- moderate pace for growth, their current ac- ing 51 percent of the overall increase in foreign count positions should not deteriorate to the officially-held U.S. treasury bills between March point where financing becomes problematic. 2000 and January 2003. While this has allowed As a whole, the current account position of the these countries to increase their reserves by a sub- major groups of developing countries is close stantial margin, it has been achieved at the ex- to balance or in surplus (table 1.2). pense of increasing their exposure to the U.S. dol- Developing countries have become major lar (figure 1.11). Among these countries, the sources of international capital. Since 2000, the share of U.S. treasury bills in their official re- central banks of some of the largest developing serves has increased by as much as 20 percentage countries have increased their foreign reserves by points and equals almost 70 percent in the case of more than 80 percent. Taken as a group, the re- Mexico, and 58 percent in China. Should these serves of Brazil, China, India, Mexico, Thailand, countries decide to rebalance their reserve port- and Turkey now represent over 45 percent of de- folio by slowing the pace at which they accumu- veloping country reserves. Indeed, following late dollar-denominated reserves, either Figure 1.11 Rising U.S. dollar reserves Developing country reserves Share in increase Billions of dollars 500 450 China Other currencies 400 28% 350 Other developing countries 300 250 200 150 100 U.S. dollar 50 72% 0 2000 2003 Source: World Bank. 11 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 downward pressures on the dollar will accentu- ate, or interest rates will have to rise in order to Figure 1.12 Impact of a 200 basis point increase in interest rates attract sufficient private capital inflow. Notwithstanding robust aggregate perfor- Deviation from baseline (percent of GDP) mance, many countries have been less 0.7 Developing successful in reaping the benefits of the last 0.6 countries Developed few years of strong economic conditions, and 0.5 countries their high current account deficits could im- 0.4 peril their stability--especially in the context 0.3 of slower growth in trade and world economic 0.2 activity. More than 50 developing countries 0.1 have current account deficits that exceed 0 5 percent of GDP. As a result, even the moder- 2005 2006 2007 2008 ate hikes in interest rates, deterioration in Source: World Bank. terms of trade, and the slower export demand projected in the baseline will likely require these countries to undergo significant cuts to imports and domestic consumption in order to probably have to rise in order to motivate maintain external stability. If trade growth private investors to re-enter the market.8 were to slow more than currently predicted, or Simulations suggest that a 200 basis point in- if terms of trade were to deteriorate more be- crease in long-term interest rates could reduce cause of an additional hike in oil prices, the re- world GDP over the short- to medium-term by quired adjustment could be severe. about 0.5 percent per annum;9 the impact would be somewhat stronger for developing countries, because higher rates will raise debt Risks and Policy Priorities servicing burdens, which require additional cuts F orceful steps are required to reduce the to spending and demand(figure 1.12). Over the twin deficits in the United States. As longer term, if the twin deficits in the United the preceding discussion has indicated, over States are not addressed (a tightening of fiscal the past few years, private sector equity and di- policy would reduce both deficits by increasing rect investment financing of the very large U.S. U.S. savings10), the problem is likely to intensify. current account deficit has dried up,6 having Permanently higher long-term interest rates been replaced to a large extent by increased would render a wide range of investment pro- purchases of U.S. bonds by foreign central jects uneconomic and slow the pace of potential banks, notably those of developing countries. output for a considerable time11-- leading, per- While these countries' build up of reserves has haps, to a period of stagflation similar to that helped improve their external financial posi- observed during the 1970­80s. tion, the stock of U.S. dollars that they now While higher U.S. interest rates might hold is very high and represents a dispropor- maintain investor interest in the dollar, they tionate share of their assets. It is not clear that would have serious disruptive impacts on they can or should increase these stocks fur- countries with large U.S. dollar debts. For ther by continuing to absorb the lion's share of countries such as Brazil, Indonesia, the Philip- net new U.S. treasury bills (6 developing coun- pines, Poland, and Turkey, a 200 basis point tries absorbed more than half of net new issues increase in dollar interest rates would signifi- since 2000).7 Assuming their appetite for trea- cantly increase debt-servicing charges. In- suries wanes, downward pressure on the U.S. creased outflows could provoke large depreci- dollar is likely to re-emerge, and yields will ations in their currencies (as much as 9 percent), which would only increase the 12 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S domestic burden of their external debt and levels ($46.8 in October 2004 for this generate further downward pressure on their average--$49.5 for Brent) or rise even currencies. Maintaining stability would, in all further. Simulations suggest that were events likelihood, require a substantial reduction in to temporarily disrupt supply by about 1 mil- imports, consumption, and investment, which lion barrels per day, oil prices could be ex- would result in slower growth and impede in- pected to increase by about $10 a barrel. In creases in poverty reduction. macroeconomic terms, such an increase The risk of such outcomes makes redress- would slow economic growth by about 0.5 ing imbalances all the more pressing. Among percentage points in the following year.13 developed countries, steps need to be taken to However, the resulting terms of trade shock reduce the U.S. government deficit, which would be larger in many poorer countries would lower overall borrowing requirements ( 2.4 percent of GDP for highly indebted and investor's concerns over the long-term poor oil-importing countries versus 0.2 per- financing of the debt. In Europe and other cent of GDP for high-income countries) be- OECD countries, more resolute steps to re- cause of the relatively high share that energy dress government deficits and to create the represents in their imports. And such fiscal room necessary to deal with the fiscal economies tend to be more sensitive to a consequences of aging will be necessary if given terms of trade shock because of their long-term interest rates are to remain low. For limited ability to attract capital flows that developing countries, a gradual appreciation would offset any resulting increases in their of some currencies relative to the dollar could trade deficits. In contrast to high-income help by permitting a further depreciation of countries, which can increase their external the dollar. However, in the absence of fiscal borrowing to offset the real-side impact of tightening in the United States, such measures higher oil prices, low-income countries are are unlikely to have a significant impact. Fis- obliged to absorb most of the shock immedi- cal consolidation is also required in many de- ately. As a result, they undergo a deprecia- veloping countries. Recent steps to lower ex- tion and substantial reductions in consump- isting government deficits move in the right tion and investment spending--adjustment direction and need to be pursued--as do ef- mechanisms that ultimately reduce spending forts to reduce trade barriers so that export on imports by almost the entire amount of opportunities can increase. While these ac- the increased oil bill (figure 1.13). Their in- tions may well imply hardship and impose real ability to defer adjustment (like high-income political costs, the human and political conse- countries do) implies significant costs, both quences of entering into a period of higher in- to individuals who see their consumption terest rates without external and internal fi- possibilities reduced and to the economy, as nances on a firm footing would be even more lower levels of investment feed through to dramatic. Finally, funding for initiatives to re- reduce the capital stock and diminish pro- lieve the debt burden of the poorest countries ductive capacity. needs to be increased. Finally, a failure of current efforts to slow Should oil prices rise even further, the the unsustainable pace of growth in China by economies of low-income countries are likely engineering a soft landing could result in to be among the hardest hit. Oil prices are as- major disruptions. The Chinese authorities sumed to moderate in the base case, falling have put into place a number of specific-- from $39 per barrel (for the average of West- mainly command and control--measures, Texas Intermediate, Brent, and Dubai oils)12 that restrict additional investment and lending in 2004 to $32 in 2006. However, given sup- to the construction and heavy production ply and geopolitical conditions, there is a real sectors. In the World Bank's forecast, this is risk that prices will either remain at current projected to succeed in slowing overall growth 13 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Figure 1.13 First year impacts of a $10 increase in oil prices Low-income net oil importers High-income countries Percent of GDP Percent of GDP 0 0 0.1 0.1 0.2 0.2 0.3 0.3 0.4 0.4 0.5 0.5 0.6 0.6 0.7 0.7 0.8 0.8 Domestic Current account Domestic Current account demand balance demand balance Source: World Bank. to some 7.1 percent in 2006, down from an proportion of the population living on $1 or estimated 8.8 percent this year. less a day by 2015 (compared to 1990 levels). So far, these steps have slowed import de- The strong economic growth in developing mand in a number of sectors, notably metals countries over the last 2 to 3 years, which is and ores,14 while credit restrictions have dra- expected to continue through 2006, albeit at a matically reduced the pace of money creation. somewhat slower pace, is based on solid fun- In contrast, private consumption growth damentals that are likely to carry forward and shows no sign of easing, and inflation has contribute to long-term economic prospects. picked up rapidly. For the moment increased In our base scenario this leads to an annual costs have not found their way into wages, growth of some 3.5 percent in per capita GDP but such a possibility cannot be ruled out. between 2006 and 2015, and contributes to Should overheating contribute to further in- achieving the MDGs. The poverty MDG will creases in inflation, a stronger policy response be met on a global basis, but a large number may be required. Moreover, investment levels of countries will not meet the goal, particu- remain very high, leaving open the possibility larly those in Sub-Saharan Africa. And though of a very rapid correction, especially if bad growth is necessary to make progress toward loans in the banking sector reveal themselves achieving the MDGs, in most countries, to be a serious problem. Either eventuality growth is insufficient without more targeted could provoke a more abrupt slowdown than policies. forecast. At least four factors are responsible for the recent and prospective improvement in growth prospects. As outlined in the first sec- Long-Term Growth, Structural tion of this chapter, among the solid funda- Change, and Poverty mental changes in developing countries is an T his part of the report, as in years past, pre- improvement in macroeconomic conditions sents a long-term growth scenario for (e.g., inflation and indebtedness). The recent the global economy and its implications for World Development Report stresses the meeting one of the MDGs: the halving of the importance of the investment climate, which 14 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S has improved in many countries and has led to vested interests protected by an RTA impede an acceleration in growth. A third factor, progress toward a globally more beneficial explored in more detail below, includes agreement? These questions will be addressed significant structural changes--that is, eco- in chapter 6. The conclusion from this chapter nomic diversification and a move away from is that the challenge for most developing coun- reliance on agriculture, and integration with tries will be the creation of jobs for a rising the global economy; both of these structural work force, rather than how to deal with em- changes involve increased urbanization. A ployment shifts across economic activities. fourth factor, pursued in greater detail in chapter 6, is the reduction in trade barriers. Long-term growth scenario The special focus of the long-term scenario The global economy is currently rebounding in this report is on structural changes, partic- from the downturn suffered in 2001 and ularly as they affect employment. Rapid 2002. Not all regions are benefiting equally growth will, in and of itself, lead to structural from the rebound--Japan and the United changes; that is, a relative decline in agricul- States are leading the way among industrial ture and a rise in the demand for services. economies--but there is fairly solid progress Countries need to think ahead, allocate scarce in all the main developing regions on an ag- public investment in a rational manner, and gregate basis. This year, 2004, is likely to be promote education to better position their the peak in the current upward cycle, with work force for a changing environment. economies drifting toward long-term trend While structural changes are likely to be im- growth in 2005 and beyond. Table 1.3 re- portant, many developing countries face an flects a plausible long-term scenario for the equal challenge in the sheer growth of the high-income countries and the World Bank's labor force. Labor force growth rates are six aggregate developing regions (see box 1.1 likely to decline over the next decade, but in for details concerning aggregation). The sce- many regions they will average between 1.5 to nario reflects current views on potential trend 2.5 percent per annum. For the poor, both growth over the 2006­15 decade. Better growth and structural change are likely to be policies, an acceleration in investment, and beneficial. Growth, to the extent that it lifts other factors could improve the prospects, all incomes, will inevitably lead to a fall in particularly for the slower growing regions. poverty. Structural change can accelerate the There is still a considerable gap in the pro- process of poverty reduction. A decline in the ductivity levels between developing and in- rural population could ease wage pressures. dustrial economies, and a number of develop- A rising urban population provides easier ac- ing countries--particularly in Asia--have cess to essential health and education services demonstrated, over the last 20 to 30 years, a and can lead to a rise in transfers to rural sustained ability for rapid growth. areas. The focus of this forecast section this year is The focus on structural change also links to on anticipated structural changes. These have the broader theme of the report--the shape many dimensions--demographic, rural versus and impacts of regional trade agreements urban, sectoral, employment shifts, openness, (RTAs). The RTAs will undoubtedly lead to ad- and income distribution, among others. While ditional structural shifts, and with associated most of these shifts have long-term positive transitional costs. How do RTAs compare with impacts, they can also be associated with growth-induced structural shifts? Do RTAs short-term transitional costs. Public policies produce structural shifts that are broadly con- can limit the costs of transition, but they can sistent with those induced by a truly open also be significantly reduced--at least in terms global economy (which would emerge from a of duration--in a fast growing economy where multilateral agreement)? And, if not, would the job growth is robust. 15 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 risen (see figure 1.14). Services, according to Box 1.1 The these figures, still represented less than 40 percent of GDP in 2002, well below the aggregation paradox nearly 65 percent share in the high-income countries of East Asia. Thus there is still sig- The per capita growth rate for the world re- nificant scope for further structural shifts. flects the so-called aggregation paradox. The The value added shares also belie the rela- long-term per capita growth rates for high- tive employment share in agriculture, which income and developing countries are, respec- tends to be much higher. Take, for example, tively, 2.4 and 3.5 percent per annum, but Thailand, where agriculture's share of value the global growth rate is only 2.1 percent and is not the average of the growth rates added is below 10 percent, but still employs (weighted or un-weighted). The following around 50 percent of the total labor force. In table highlights the aggregation paradox. the high-income countries, the relevant shares The paradox is explained by the relatively are around 2 percent of value added and less high weight of high-income countries GDP than 4 percent of employment.15 Higher agri- in the world total, but their low weight in cultural productivity and relative wage differ- world population. entials will continue to drive an exodus from agriculture into other sectors. And the change High- Develop can come rapidly. In the Republic of Korea, income -ing World the percent of employment in agriculture dropped from 32 percent in 1982 to 10 per- Population (million) 2006 970 5,340 6,320 cent in 2001. The agricultural transformation 2015 990 5,900 6,900 is present in some of the other developing re- Growth ratea 0.3 1.1 1.0 gions as well; for example, in South Asia the GDP ($billion) percent of employment in agriculture dropped 2006 31,200 8,200 39,400 from 40 percent in 1982 down to 27.2 percent 2015 39,500 12,300 51,800 in 2002, and in Latin America and the Growth ratea 2.7 4.6 3.1 Caribbean, the percent of employment in agri- GDP per capita ($) culture dropped from 14.4 percent down to 2006 32,090 1,530 6,240 10.6 percent over the same two-decade 2015 39,700 2,080 7,510 period. There has been no significant shift in Growth ratea 2.4 3.5 2.1 either the Middle East and North Africa or a. Growth rates are percent per annum. Sub-Saharan Africa regions. At the same time, neither of those two regions witnessed much economic growth, with only 0.4 percent per capita growth per annum in the former, and a loss of 0.3 percent per annum in the latter. Structural Changes over Two In all regions, save East Asia, one can see a Decades climb in the share of services. This is not sur- L ooking back on the last 20 years of devel- prising because services are assumed to be in- opment, many developing regions have al- come elastic and a relative rise in the con- ready witnessed significant structural shifts. sumption share of services is understood. This Perhaps foremost is the decline of agriculture effect is reinforced by the relatively high rate of as a source of income and employment. In productivity growth in manufacturing. All else East Asia and the Pacific, agricultural value being equal, this reduces the price of manufac- added has declined from a 28 percent share in tures relative to services and hence enhances 1982 to only 15 percent in 2002, and manu- the value share of services. Perhaps what is facturing, other industrial, and services have more surprising is the variation across regions. 16 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S Table 1.3 Long-term prospects: Forecast growth of world GDP per capita Real GDP per capita, annual average percentage change 1980s 1990s 2000­06 2006­15 World total 1.3 1.1 1.6 2.1 High-income countries 2.5 1.8 1.7 2.4 OECD 2.5 1.7 1.7 2.3 United States 2.2 1.9 1.8 2.5 Japan 3.5 1.1 1.7 1.9 European Union 2.1 1.8 1.5 2.3 Non-OECD countries 3.5 4.1 1.6 3.5 Developing countries 0.6 1.5 3.4 3.5 East Asia and the Pacific 5.8 6.3 6.0 5.3 Europe and Central Asia 1.0 1.8 5.2 3.5 Latin America & the Caribbean 0.9 1.5 0.8 2.4 Middle East North Africa 1.6 1.1 2.4 2.6 South Asia 3.3 3.2 4.2 4.1 Sub-Saharan Africa 1.2 0.5 1.2 1.6 Note: Aggregations are moving averages, reweighted annually after calculations of growth in constant prices. Source: World Bank. Figure 1.14 A rise in services East Asia and Pacific South Asia Middle East and North Africa Structure of value added (percent) Structure of value added (percent) Structure of value added (percent) 60 60 60 50 50 50 40 40 40 1982 30 1992 30 30 2002 20 20 20 10 10 10 0 0 0 Agriculture Industry Services Agriculture Industry Services Agriculture Industry Services Sub-Saharan Africa Latin America and the Caribbean Europe and Central Asia Structure of value added (percent) Structure of value added (percent) Structure of value added (percent) 60 70 70 50 60 60 50 50 40 40 40 30 30 30 20 20 20 10 10 10 0 0 0 Agriculture Industry Services Agriculture Industry Services Agriculture Industry Services Source: World Bank. 17 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 In the high growth regions--East Asia and growth overall, but are also more dependent South Asia--there are two contrasting on natural resource production, and those patterns. In South Asia the employment in relative prices have been declining over most of agriculture shifted mostly to services, with a the period. Natural resources appear under in- small increase in industrial output. In East dustrial production, so even if volume growth Asia, employment in agriculture shifted more has been positive, with declining relative prices evenly between industry and services. And the natural resources share in output could be there appears to have been a structural break declining. And apart from Latin America and in the 1990s with an acceleration of industrial the Caribbean, these regions have also not re- output. This is consistent with the sharp rise in ally benefited from global production sharing the trade to GDP ratio doubling from 36 per- in the more integrated global economy. The cent in 1982 to 72 percent in 2002, and with Middle East and North Africa Region has East Asia as a hub of assembly and manufac- barely seen any shift in its trade to GDP ratio. turing activities (see figure 1.15). There are, of For both Sub-Saharan Africa and Latin Amer- course, exceptions in each region. The Philip- ica and the Caribbean, however, the ratio has pines, for example, has a sharp rise in services increased markedly, particularly in the and a decline in manufacturing--perhaps as a 1990s--from 50 percent to 69 percent for Sub- result of its regional comparative advantage in Saharan Africa, and from 27 to 47 percent for back office operations, call centers, and other Latin America and the Caribbean. The more services requiring specialized language skills. recent rise in Latin America can be partly ex- In South Asia, India's services dominate, but plained by the implementation of a raft of growth is much lower in Bangladesh and Pak- regional agreements, including NAFTA and istan, where textile and clothing exporters may MERCOSUR. At the same time Latin Amer- be taking advantage of their relatively gener- ica's degree of openness is lower than that of ous quotas to the main importing markets. East Asia, and in general it has been less co- Three of the other regions--Latin America opted into global production networks. and the Caribbean, the Middle East and North The transformation in the economies of Africa, and Sub-Saharan Africa--show less Europe and Central Asia over the last 15 years is a result of an abrupt structural shift. The dominance of industry as part of an economic strategy of planned economies was eliminated. Services in the transition economies quickly Figure 1.15 Rising openness to trade filled the gap, which led to significant disloca- Trade (exports and imports) to GDP ratio (percent) tion for a period, but is now forming the basis 90 1992 of more rational and sustained growth. 80 1982 2002 Looking ahead it is clear that there is the 70 potential for significant change. While the rate 60 50 of urbanization has been persistent over the 40 last two decades, there is a long way to go, 30 particularly in Asia and Africa, before attain- 20 ing the 80 percent level of the industrial coun- 10 tries (figure 1.16). The income gap is also 0 huge, even if incomes are measured in pur- Asia Asia and and and Asia chasing power parity (PPP) terms. In East East PacificSouth East Africa Africa and America CaribbeanEurope Asia, per capita incomes averaged just over North Sub-Saharan Central Middle the Latin $1,000 (1995 dollars) in 2002, compared with nearly $31,000 in the industrial countries-- Source: World Bank. roughly a 30 to 1 differential. Even assuming 18 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S Structural Change in the Future Figure 1.16 Increased urbanization Urban population as a share of total population (percent) T able 1.3 presents the long-term growth rates. This section focuses on some of the 80 2002 consequences of growth and other underlying 70 1992 assumptions of the long-term scenario on 60 1982 structural changes, particularly regarding 50 labor shifts--both in volume terms and across 40 sectors.16 30 In the aggregate, and assuming no change in 20 labor force participation rates, labor supply 10 growth will slow down sharply in most regions 0 after 2010--with the exception of the Middle Asia Asia and and and Asia East and North Africa and Sub-Saharan Africa East PacificSouth East Africa Africa and America CaribbeanEurope regions (figure 1.17).17 In Western and Eastern North Sub-Saharan Central Middle the Latin Europe, Russia, and Japan, the labor supply would most likely shrink (even before 2010), Source: World Bank. putting additional pressure on underfinanced pension schemes. Additionally, it is the regions with the highest labor force growth rates that a PPP exchange rate of 5 would still lead to a also tend to have the lowest per capita growth significant 6 to 1 ratio in per capita incomes. rates, so these regions are on a knife-edge in In Latin America, the richest developing terms of their capacity to absorb high rates of region with a per capita average income of new workers. These same regions typically have $3,700, would have a ratio similar to East relatively low labor force participation rates, Asia using a PPP exchange rate of around 2. particularly of females; thus increases in Figure 1.17 Growth rate of labor supply declining Population between the ages of 15 and 65 (percent) 3.0 2.5 2.0 1.5 2005 1.0 2010 0.5 0 0.5 1.0 2015 1.5 and NIE China India XAS and East Europe Japan States est Asia W Africa Africa America CaribbeanSub-Saharan Middle United East Central North Latinthe and Note: United States+ includes Australia, Canada, New Zealand, and the United States. NIEs are newly industrialized economies. XAS represents other major countries in East and South Asia. Source: World Bank. 19 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Table 1.4 Labor market structure, 2005­15 Growth between 2005­15: percent per annum Growth decomposition Agric Manuf Services Total Structure Expansion Total Australia, Canada & New Zealand 0.7 0.8 0.7 0.4 3.45 3.90 6.88 United States 1.1 0.6 0.8 0.5 2.95 5.62 8.34 Japan 3.0 2.2 0.6 1.0 3.62 9.35 8.70 Korea and Taiwan 1.9 0.6 1.0 0.5 4.18 5.65 9.17 Hong Kong (China) and Singapore 0.0 0.5 1.0 0.7 4.80 7.28 11.65 EU with EFTA 2.6 1.5 0.2 0.2 4.92 2.28 3.12 Brazil 0.8 0.5 1.4 1.2 3.52 13.09 16.28 China 0.5 0.4 1.3 0.8 5.12 8.44 12.02 India 0.1 1.6 2.4 1.7 9.27 19.89 28.02 Indonesia 1.1 1.8 1.6 1.5 8.32 17.18 22.73 Mexico 0.2 1.6 2.5 2.0 7.00 22.78 28.84 Russia 1.0 1.1 0.0 0.4 4.97 4.35 2.43 SACU 0.7 0.2 0.9 0.6 3.82 6.87 10.36 Vietnam 1.7 1.9 2.0 2.0 5.69 22.23 26.20 Rest of East Asia 0.7 1.2 2.1 1.7 4.72 19.36 23.09 Rest of South Asia 1.8 1.3 2.7 2.4 4.60 27.22 31.38 EU accession countries 0.8 1.1 0.2 0.2 6.51 2.41 4.67 Rest of ECA 0.3 0.3 1.1 0.8 6.15 8.88 14.43 Middle East 2.3 1.6 2.5 2.3 5.26 26.43 30.79 North Africa 0.8 2.0 2.5 2.0 7.27 23.74 29.83 Rest of Sub-Saharan Africa 2.0 2.2 3.0 2.6 4.50 30.19 34.08 Rest of LAC 2.1 0.9 2.2 1.8 5.36 20.74 25.41 Rest of the world 0.8 1.2 2.1 1.7 4.15 19.45 23.12 Source: World Bank simulations. participation rates will lead to additional labot "structural" component, which measures the market weakness. quantity of labor force movement across sec- With high-income demand elasticity for ser- tors, assuming no change in the volume of vices and relatively higher labor productivity in labor. The second is an "expansion" compo- manufacturing, labor demand growth will tend nent, measuring the overall growth in the to be higher in services than in manufacturing labor force. In the case of India, for example, and/or agriculture (see table in endnote 17). the numbers suggest that the labor force will This effect is quite pronounced in the industrial grow by about 20 percent between 2005 and countries, where labor demand growth be- 2015, or about 1.8 percent per annum. And tween 2005 and 2015 will be negative, on aver- in each year, about 0.9 percent of the initial age, in agriculture and manufacturing in all labor force will move across sectors. Thus the high-income regions, with all of the net growth total annual movement of 2.5 percent per occurring in services (with the exception of annum is composed roughly of 2/3 expansion Japan, where labor force growth could poten- and 1/3 by intersectoral movements. It should tially decline by 1 percent per annum on aver- be clear from the decomposition that for most age). The shift toward services also occurs in de- of the developing regions, there will be more veloping countries, but with continued high labor movement from the expansion of the growth in manufacturing and less growth in labor force than from structural change, with agriculture. the notable exceptions of Russia and the Table 1.4 also shows a summary measure other countries in Europe and Central Asia-- of the structural changes. It decomposes the and, perhaps somewhat more surprisingly, total change in the structure of the labor force China. For the industrial regions with low or into two components. The first is the declining labor growth, clearly the structural 20 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S shifts will be relatively the same order of mag- with positive aggregate benefits, but hurt the nitude as the expansion component. But the sectors that thrived under the preferential shifts are relatively small on an annual basis, arrangement? perhaps 0.3 to 0.5 percent of the labor force. Chapter 6 of this report will re-address one Poverty Forecast D issue related to structural shifts in the context eveloping country economic performance of RTAs. Do RTAs lead to structural changes has been strong since 2002, and this is that are inconsistent with the structural projected to continue over the next two years changes from a broad multilateral agreement? and beyond (tables 1.1 and 1.3). This pattern For example, a country signing an RTA may of high growth would in all likelihood lead to have a local comparative advantage in a given a halving of the number of poor (i.e., the per- sector, but not a global comparative advantage. centage of poor living on $1 or less a day) in In this case, would the country need to undergo developing countries between 1990 and 2015 two potentially costly adjustments, should a (table 1.5)--one of the key MDGs. At the multilateral agreement be signed subsequent to global level, the target to be achieved in 2015 an RTA? And would the vested interests that is around 14 percent (one-half of 27.9), and benefit from the RTA hamper the ability the forecast is for a headcount index of to achieve a broader multilateral agreement, 10.2 percent. This translates into a forecast of Table 1.5 Regional breakdown of poverty in developing countries Number of people living on less than $1 per day (millions) GEP2004 GEP2005 Region 1990 2000 2015 1990 2001 2015 East Asia and Pacific 470 261 44 472 271 19 China 361 204 41 375 212 16 Rest of East Asia and Pacific 110 57 3 97 60 2 Europe and Central Asia 6 20 6 2 17 2 Latin America and the Caribbean 48 56 46 49 50 43 Middle East and North Africa 5 8 4 6 7 4 South Asia 467 432 268 462 431 216 Sub-Saharan Africa 241 323 366 227 313 340 Total 1,237 1,100 734 1,218 1,089 622 Excluding China 877 896 692 844 877 606 $1 per day head count index (percent) GEP2004 GEP2005 Region 1990 2000 2015 1990 2001 2015 East Asia and Pacific 29.4 14.5 2.3 29.6 14.9 0.9 China 31.5 16.1 3.0 33.0 16.6 1.2 Rest of East Asia and Pacific 24.1 10.6 0.5 21.1 10.8 0.4 Europe and Central Asia 1.4 4.2 1.3 0.5 3.6 0.4 Latin America and the Caribbean 11.0 10.8 7.6 11.3 9.5 6.9 Middle East and North Africa 2.1 2.8 1.2 2.3 2.4 0.9 South Asia 41.5 31.9 16.4 41.3 31.3 12.8 Sub-Saharan Africa 47.4 49.0 42.3 44.6 46.4 38.4 Total 28.3 21.6 12.5 27.9 21.1 10.2 Excluding China 27.2 23.3 15.4 26.1 22.5 12.9 21 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Table 1.5 Regional breakdown of poverty in developing countries (continued) Number of people living on less than $2 per day (millions) GEP2004 GEP2005 Region 1990 2000 2015 1990 2001 2015 East Asia and Pacific 1,094 873 354 1,116 864 230 China 800 600 256 825 594 134 Rest of East Asia and Pacific 295 273 98 292 271 95 Europe and Central Asia 31 101 48 23 93 25 Latin America and the Caribbean 121 136 124 125 128 122 Middle East and North Africa 50 72 38 51 70 46 South Asia 971 1,052 968 958 1,064 912 Sub-Saharan Africa 386 504 612 382 516 612 Total 2,653 2,737 2,144 2,654 2,735 1,946 Excluding China 1,854 2,138 1,888 1,829 2,142 1,812 $2 per day head count index (percent) GEP2004 GEP2005 Region 1990 2000 2015 1990 2001 2015 East Asia and Pacific 68.5 48.3 18.2 69.9 47.4 11.3 China 69.9 47.3 18.4 72.6 46.7 9.7 Rest of East Asia and Pacific 64.9 50.8 17.6 63.2 49.2 14.7 Europe and Central Asia 6.8 21.3 10.3 4.9 19.7 5.2 Latin America and the Caribbean 27.6 26.3 20.5 28.4 24.5 19.6 Middle East and North Africa 21.0 24.4 10.2 21.4 23.2 11.9 South Asia 86.3 77.7 59.2 85.5 77.2 54.2 Sub-Saharan Africa 76.0 76.5 70.7 75.0 76.6 69.2 Total 60.8 53.6 36.4 60.8 52.9 32.0 Excluding China 57.5 55.7 42.0 56.6 54.9 38.6 Source: World Bank. 622 million persons living $1 or less a day in 15 years has been insufficient to be on track to 2015, compared with 1.2 billion in 1990 and achieve the income poverty target in 2015 with- an estimated 1.1 billion in 2001.18 With re- out more rapid growth or policies that are bet- spect to the somewhat higher poverty line of ter targeted to the poor. Within regions, $2 a day, the headcount should improve to progress has also been uneven. Despite the huge 32 percent in 2015--not quite a halving of the overall reduction in East Asia, several countries, estimated 61 percent headcount index in for example, Cambodia, Lao PDR, and Papua 1990--and corresponding to almost 2 billion New Guinea, are off track to meet the goal. In poor. Sub-Saharan Africa, there are only eight However, progress is highly uneven across countries--representing 15 percent of the sub- and within countries. The global target will continent's population--that will potentially largely be achieved because of the significant make significant progress toward achieving the progress on poverty reduction in China and income poverty target. Within countries, such India. Sub-Saharan Africa lags far behind, and as China, there are large pockets of poor though poverty rates are much lower in some of people, and reducing poverty in these pockets is the other regions, for example Latin America difficult because they are often concentrated in and the Caribbean, progress over the last remote, hard-to-reach locations. Links to the 22 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S national and/or global economy are weak, and countries are on track to achieve the target-- provision of public services--education, health, Indonesia, Lao PDR, Malaysia, and the Philip- water and sanitation--is difficult and pines. All other countries are off track, and expensive. two--Cambodia and Papua New Guinea--are This year's poverty forecast, as in years seriously off track. The situation is also dire for past, reflects changes in two key dimensions. births attended (linked to maternal mortality) First, new country surveys lead to a re- and access to safe water. These examples also evaluation of the level of poverty in 1990 and illustrate that the other MDG targets are less in the most recent base year, 2001. At the directly correlated to income levels.21 For ex- global level, the $1/day headcount index for ample, Lao PDR and Indonesia are on track for 1990 has been shaved slightly from 28.3 per- the child mortality target, but Thailand is not. cent in last year's report, to 27.9 percent in this year's report. There is also a very modest decline in the estimated level of poverty for Concluding Remarks 2001. The new surveys also force a re-evalu- he rapid growth of developing economies, ation of the link between income growth and T mostly concentrated in East and South poverty reduction. Using the latest survey in- Asia, has produced a spectacular, if not his- formation and last year's economic forecast, toric, fall in poverty that will enable the the forecasted decline in poverty is somewhat achievement of the poverty MDG on a global more rapid, with the headcount index declin- basis, although many countries will be seri- ing to 10.4 percent (from 21.1 percent in ously off-target. The rapid growth has been as- 2001), instead of 12.5 percent (from 21.6 sociated with large structural shifts--greater percent in 2000).19 The second key dimen- openness, more urbanized populations, and a sion is the change in the long-term economic sharp fall in agricultural employment. These forecast. The changes overall are relatively trends will persist in the future as growth rates modest. However, the somewhat improved remain high, and incomes and productivity lev- performance anticipated between 2003 and els in developing countries are still well below 2006 generates better average growth for the industrial country averages--even taking into forecast period 2001­15 and drops the head- account PPP adjustments. As an example of count index for 2015 from 10.4 percent to potential structural shifts, take China's level of 10.2 percent. urbanization. Its rural population may not ap- While progress on income poverty in parts proach the 20 percent level of industrial coun- of the world, particularly East and South Asia, tries, but a 50 percent share in 2015 could lead has been spectacular if not historic, there is no to a cumulative migration in the range of 140 room for complacency. As mentioned earlier, to 175 million persons between 2005 and there are significant pockets of poverty even 2015. Such large shifts will require consider- within the more successful countries. More- able public and private resources and their effi- over, there are other dimensions of poverty in cient allocation. Chapter 6 addresses a comple- which progress has been more limited, and mentary issue--structural changes induced by almost all developing countries are off track. In changes in trade policies, notably the impacts East Asia, for example, the region scores rela- of preferential trade agreements. tively well for achieving 100 percent primary school completion rates, with China and Viet- nam already having achieved the target and the Notes Philippines on track.20 But Thailand and In- 1. The investment to GDP ratio in the United donesia are off track, as are some of the poorer States is currently 21 percent, close to its peak of countries in the region. For the child mortality 21.5 percent during the Internet bubble, and well MDG, the situation is more worrying. Four above historical peaks of less than 18 percent. 23 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 2. The weighted average of the dollar's fluctuations historical analysis focused on output because of the relative to world currencies. greater availability and reliability of the data. Histori- 3. West-Texas Intermediate was much higher in cal data on employment patterns has many gaps, and October 2004 ($56). The overall average was the data that does exist is often not compatible across depressed by the price of other oil (notably from countries. Dubai), which was lower because producers increased 17. The baseline scenario and the induced struc- the supply of lower quality oil. tural changes are predicated on a number of assump- 4. Lau (2003) estimates that because of the re- tions. First, growth in the labor supply is equated with export nature of its trade, the domestic value-added growth of the working age population. For all regions, content of Chinese exports may be as little as 20 percent. this implies a slowing of labor force growth, albeit with 5. Net equity and foreign direct inflows of foreign high growth in some developing regions. At the same private investors declined by 73 percent between 2001 time, the labor force is assumed to be flexible and thus and 2003. At the same time, net outflows by American will reinforce anticipated structural shifts. Second, sav- private investors increased by 10 percent. As a result, ings are similarly influenced by demographics. In many total flows have reversed, from a significant inflow of developing countries this will translate into a slight $35 billion in 2001 to a $195 billion outflow in 2003. acceleration in savings as the ratio of youth to workers Since then, these trends have continued, with total out- declines, and a decline in industrial countries as the flows representing $267 billion in the second quarter ratio of elderly to workers rise (explored in more detail of 2004. in World Bank 2003a). Investment growth will largely 6. See endnote 5. be driven by domestic savings, as it has in the past; 7. Calculated as the change in U.S. t-bills held by however, with modest increases in net capital flows to- the central banks of these countries divided by the net ward developing countries, with the exception of East increase in t-bills held by official lenders (see http:// Asia, which has been a major source of international www.treas.gov/tic/mfhhis01.txt). capital over the last five years. 8. Mussa (2004) suggests that a further 20 percent Third are the assumptions regarding productivity depreciation might be required to bring the U.S. econ- growth; based on previously observed trends, these are omy into external balance. divided into three broad economic sectors. In agricul- 9. These results are consistent with those published ture, it is assumed that the past growth of roughly by the OECD for developing countries (see Dalsgaard 2.5 percent per annum is maintained through 2015 and others 2001). (see, for example, Martin and Devashish 1999). 10. Even after Ricardian equivalence-based Maintaining this high rate of agricultural productivity changes to private saving. Nevertheless, Brooks and will require continued and perhaps increasing invest- others (2003) show that, taken alone, neither a 2 per- ment in agricultural research and extension, combined centage point cut in fiscal spending, nor a 10 percent with rising investment in agricultural infrastructure, effective depreciation would be sufficient to restore ex- particularly for water resource management. This rate ternal balance in the United States. They argue that a of productivity growth in agriculture is consistent with combination of depreciation, stronger world demand, a modest secular decline in agricultural prices, relative and a larger fiscal contraction would be required. to the general price trend, as observed in the past. The 11. Under higher interest rates, the desired stock of other two broad sectors are manufacturing and services. capital declines, which requires a prolonged period of Again, based on past trends, it is assumed that produc- slower growth before the economy adjusts to the new tivity growth in manufacturing will be higher than in lower levels of output and capital. services. This has two impacts: (1) it reduces the price of 12. In October 2004, this average price was $46.8 manufactures relative to services, all else being equal, comprised of $53 for West-Texas Intermediate, $49.5 and thus enhances the share of services in value terms; for Brent, and $37.7 for Dubai oil. and (2) for the same level of output, it reduces the 13. Dalsgaard and others (2001) estimate similar impacts for OECD countries. Income elasticities in the Linkage model 14. Growth in steel demand fell 36 percent during Ag. and Industrial the 3-month period ending in July, while copper im- food Energy goods Services ports were flat. 15. World Bank 2003b. United States 0.01 0.58 0.78 1.14 16. Unlike the previous section, which focused on Japan 0.04 0.64 0.68 1.24 the structure of value added, this section focuses on Europe 0.08 0.72 0.71 1.29 labor. The focus on labor provides a better perspec- Rest of high-income 0.16 0.86 0.80 1.26 Low-income 0.52 1.40 1.08 1.41 tive on the poverty dimension of structural shift. The 24 G L O B A L O U T L O O K A N D T H E D E V E L O P I N G C O U N T R I E S demand for employment in the manufacturing sectors, of the complementarities across targets, for example the and thus allows for a shift of labor toward services. degree to which improvements in access to safe water Fourth are the demand assumptions--the other side can improve health outcomes. A pilot study is currently of the coin regarding structural changes. High-income being undertaken for Ethiopia and first results will be countries have already witnessed a large decline in described in the Global Monitoring Report 2005. the demand for agriculture and food relative to income. Demand for services has increased relative to income and the demand for other goods. And there is no reason for these trends not to continue in the future. Thus the for- References ward-looking scenarios assume that income elasticities Bergsten, Fred C. and John Williamson, eds. 2004. over the next 10 years will largely reflect their current Dollar Adjustment: How far? Against What? In levels, though highly differentiated across commodities Special Report, No. 17, Institute for International and regions (see table). Economics, Washington DC. 18. The absolute number of poor won't necessarily Brook, Anne-Marie, Franck Sédillot, and Patrice Olli- be halved due to population growth. vaud. 2004. Channels For Narrowing The US 19. A more subtle change in the methodology has Current Account Deficit And Implications For also been incorporated in this year's poverty forecast. Other Economies. OECD Economics Department The poverty forecast is based on the growth of the Working Papers, No. 390, OECD, Paris. survey-based per capita consumption, assuming distri- Dalsgaard, Thomas, Christophe André, and Pete bution neutrality (with some exceptions). However, it Richardson. 2001. Standard Shocks In The has been observed in the past that survey-based con- OECD Interlink Model OECD Economics sumption growth deviates from consumption growth Department Working Papers, No. 306, OECD, as measured in the national accounts. A conversion Paris. factor has been used to adjust for this deviation, which Lau, Larence. 2003. Is China Playing by the Rules? for most countries implied an elasticity of 0.9. In other Free Trade, Fair Trade, and WTO Compliance. words, if national income consumption grows at Testimony at a Hearing of the Congressional Ex- 10 percent, the assumed growth in survey-based con- ecutive Commission on China, Washington, DC. sumption is 9 percent. More recent econometric evi- Martin, Will, and Devashish Mitra. 1999. Productivity dence suggests that the long-run elasticity is 1, but that Growth and Convergence in Agriculture and there are short-term deviations from the long-run Manufacturing. Working Papers, No. 2171, elasticity. Because of the robustness of the long-run re- World Bank, Washington, DC. lationship, the new forecast assumes an elasticity of 1. Mussa, Michael. 2004. Exchange Rate Adjustments Thus, all else being equal, this year's forecast will be Needed to Reduce Global Payments Imbalances. lower than in the past because of higher implied con- Special Report, No 17, Institute for International sumption growth. Economics, Washington DC. 20. See World Bank 2004. World Bank. 2003a. Global Economic Prospects: In- 21. The World Bank, in its effort to improve its abil- vesting to Unlock Global Opportunities. ity to monitor and forecast the other dimensions of the Washington, DC: World Bank. Millennium Development Goals, is developing and test- ------. 2003b. World Development Indicators. ing a new tool to forecast some of the MDGs. The tool Washington, DC: World Bank. will link economic growth with expenditures on health, ------. 2004. Global Monitoring Report. Washington, education, and infrastructure. It will also capture some DC: World Bank. 25 2 Regional Trade and Preferential Trading Agreements: A Global Perspective In the last four decades, developing countries partners on a most favored nation, or nondis- have burst onto the global marketplace. Their criminatory (MFN) basis, at the same time as share of global trade increased from about they have been eliminating barriers preferen- one-fifth in 1960 to about one-third in 2004-- tially through RTAs. In fact, roughly 66 per- at a time when global trade as whole was cent of the decline in average tariffs in devel- increasing to unprecedented levels. In every oping countries during the last two decades region, exports have outpaced the growth of has come from unilateral reductions, as dis- output and increased as a share of GDP. Three tinct from 25 percent coming out of the rounds of multilateral trade negotiations com- Uruguay Round and around 10 percent from bined with structural economic reforms un- RTAs. Moreover, product exclusions and re- dertaken throughout the world ushered in the strictive rules of origin further limit the trade- sustained reduction in border protection that expanding effects of preferences. Nonetheless, made this growth possible. The World Trade the result of this proliferation is an increas- Organization (WTO), formed in 1994, con- ingly complex global trading system where solidated an evolving system of rules based on different countries' access to a given market nondiscrimination among trading partners-- are often governed by very different sets of a cornerstone of the multilateral system. rules. Today a second trend in the trading system This chapter charts the rise of RTAs, exam- is rapidly gaining momentum and establishing ines the different motivations countries have a very different set of rules. This new trend is for pursuing RTAs, and draws attention to the proliferation of regional and bilateral the complexity they generate. It then describes trade agreements (RTAs)--agreements among the evolution of regional trading patterns and a group of countries that reduce barriers to shows how the major developing regions dif- trade on a reciprocal and preferential basis for fered strikingly in their timing of integration, those in the group. The number of these agree- their pace of export growth, their policies ments has more than quadrupled since 1990, toward import competition and foreign in- rising to around 230 by late 2004.1 Trade be- vestment, and the impact of regional trading tween RTA partners now makes up nearly 40 arrangements. It concludes that those regions percent of total global trade, and new agree- that aspired to trade most with the global ments increasingly address issues beyond economy became the most regionally inte- trade. The value of preferences has steadily grated as well. Further, regional trade tends to fallen, however, as most countries have been precede preferential trade agreements rather reducing tariffs across the board to all than the other way around. 27 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Box 2.1 RTAs and types of trade liberalization F ollowing WTO convention, the term regional different MFN barriers on nonmembers. This latter trade agreement encompasses both reciprocal characteristic requires members to develop rules of bilateral free trade or customs areas and multicountry origin to prevent imports from third countries from (plurilateral) agreements. Regional and bilateral trade being transshipped through the member country with agreements provide for one type of trade liberaliza- the lowest tariffs. tion, and they must be seen in a broader context of 2. A Customs Union moves beyond a free trade alternative methods of liberalization. Members of area by establishing a common external tariff on all RTAs liberalize trade on a reciprocal and preferential trade between members and nonmembers. Customs basis. While programs such as the U.S. African unions typically contain mechanisms to redistribute Growth and Opportunity Act (AGOA) and the EU's tariff revenue among members. Everything But Arms (EBA) also liberalize trade 3. A Common Market deepens a customs union preferentially (i.e., different trade partners receive by providing for the free flow of factors of produc- different treatment), the United States and EU extend tion (labor and capital) in addition to the free flow these preferences unilaterally rather than reciprocally. of outputs. In contrast to both of these types of preferential liber- 4. In an Economic and Monetary Union, members alization, countries often lower trade barriers in a share a common currency and macroeconomic nondiscriminatory fashion for all trade partners. They policies. might do so multilaterally--through GATT/WTO The international experience with RTAs is much negotiating rounds--or autonomously, as in the case richer than this simple taxonomy suggests. NAFTA of Pakistan in the late 1990s. The matrix below andother more recent agreements establishing free illustrates this taxonomy of liberalization methods. trade areas contain provisions governing domestic RTAs are commonly divided into several basic labor standards and other regulatory issues, which categories, according the degree of economic integra- one traditionally associated with agreements for tion they provide. The canonical taxonomy of RTAs deeper integration. On the other hand, many free contains the following four levels of integration: trade agreements exclude important categories of 1. In a Free Trade Area, members eliminate goods (notably agriculture) from trade liberalization. barriers to trade in goods (and increasingly services) In some cases customs unions still levy tariffs on among members, but each member is free to maintain trade between members. Method of implementation Scope of beneficiaries Reciprocal Unilateral Preferential: selected countries NAFTA, EU, COMESA, GSP, AGOA, EPAs, and other RTAs EBA, Cotonou Nondiscriminatory GATT/WTO (MFN): all countries multilateral agreements Autonomous liberalization Source: World Bank staff. The Proliferation of Regional estimates that another 60 agreements are in Preference Systems various stages of negotiation. The boom in M ore agreements are being signed. Since RTAs reflects changes in certain countries' 1990, the number of RTAs in force rose trade policy objectives, the changing from 50 to nearly 230 (figure 2.1). The WTO perceptions of the multilateral liberalization 28 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S Figure 2.1 Number of RTAs exploded in the 1990s Cumulative number of agreements (EU-25 counted as single country) 30 300 25 250 20 Agreements not Cumulative number of agreements 200 notified to the WTO (EU-15 counted as single country) 15 150 Agreements notified 10 100 to the WTO 5 50 0 0 1958 1969 1976 1984 1989 1994 1999 2004 Sources: WTO data and WTO staff. process, and the reintegration into the global pursuing two regional arrangements that are economy of countries in transition from social- designated to become customs unions: the ism. This last category accounts for many of the Euroasian Economic Community and the new agreements signed in the early 1990s, when Single Economic Space. Because a consistent countries in Eastern Europe and the former data source covering all RTAs is lacking, data Soviet Union negotiated RTAs with Western are based on the information contained in the Europe [both the EU and the European Free WTO database, supplemented by data from Trade Association (EFTA)] and with each other. the major unreported agreements. Some of the RTAs included in figure 2.1 have never been reported to the WTO, for any Most countries are participating of several reasons. One reason is that the Nearly all countries belong to at least one WTO does not enforce notification (the same RTA,2 and some are party to numerous is true of notification requirements in other agreements (table 2.1). On average, each WTO agreements). Another is that several country belongs to six RTAs, though there is countries that have yet to join the WTO have considerable variation across regions and been quite active in forming RTAs. Russia, for levels of development. East Asian countries example, is in the process of joining the WTO sign fewer agreements than countries in and has signed bilateral free trade agreements other regions. Northern countries have par- (FTAs) with other members of the Common- ticipated to the greatest extent, each signing, wealth of Independent States (CIS). It is also on average, 13 agreements. A substantial Box 2.2 Reporting RTAs to the WTO R TAs represent a fundamental departure from the covering the formation of customs unions and free core WTO principle of nondiscrimination. trade areas in merchandise trade; the General Agree- Nonetheless, the WTO affords its members a large ment on Trade in Services (GATS) Article V on degree of flexibility in entering new RTAs. Within agreements in services; or the Enabling Clause (the the WTO mandate, countries may join agreements 1979 Decision on Differential and More Favorable by meeting the requirements of the General Agree- Treatment, Reciprocity, and Fuller Participation of ment on Tariffs and Trade (GATT) Article XXIV (Box continues on next page) 29 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Box 2.2 (continued) Developing Countries) dealing with trade in goods members have yet to be reported to the organization. between developing countries. These include the Greater Arab Free Trade Area, the Countries are required to notify the WTO Aghadir Agreement in the Middle East and North secretariat of any RTAs they enter into, and their pro- Africa, the India-Nepal and India-Bhutan agreements visions are subject to review by the WTO. However, in South Asia, and several agreements in Africa. Some the review process in practice ends with the commit- agreements, such as EU accessions, are reported more tee's queries of the parties and has not led to a subse- than a year in advance, while other agreements are quent report to the membership or formal WTO notified six or more years after entry into force. The endorsement in any case. Furthermore, even the notifi- average gap is 354 days. Excluding agreements noti- cation requirement, though a formal rule, has enjoyed fied before signature, the average gap rises to 446. The only inconsistent compliance. An examination of the median delays between entry into force and notifica- WTO RTA database reveals very large gaps between tion are 135 and 188 days, respectively. the date agreements are signed and the date they are reported to the WTO. Several agreements with WTO Sources: World Bank and WTO staff. Table 2.1 Most countries belong to more than one RTA Europe Latin Middle East Asia and America East and Central and the and North Sub-Saharan Pacific Asia Caribbean Africa South Asia Africa North Total Number of countries 32 36 39 21 8 48 25 209 North-South bilateral Countries belonging to at least one RTA 4 12 6 10 0 2 10 44 Average number of RTAs per country 2 1 2 1 1 4 2 Maximum number of RTAs per country 4 4 4 3 0 1 24 24 All others Countries belonging to at least one RTA 24 22 33 20 8 47 10 164 Average number of RTAs per country 2 6 8 5 4 4 8 5 Maximum number of RTAs per country 3 12 17 12 9 9 15 17 Total Countries belonging to at least one RTA 26 26 35 20 8 48 11 174 Average number of RTAs per country 2 6 8 5 4 4 11 5 Maximum number of RTAs per country 7 12 19 13 9 9 29 29 Note: Bilateral agreements are defined as an RTA with two members. North is OECD 24 plus Lichtenstein, and South is all other countries. Source: Published WTO data, World Bank staff. number of developing countries (45) have bilateral agreement with a Northern partner. signed bilateral preferential agreements with The enlargement of the EU in May 2004 led a Northern partner. However, this activity is to a fall in the number of North-South RTAs not spread evenly across regions. Most ac- in Europe. tivity has been in Eastern Europe, Northern Since 2000, several major new trends have Africa, and Latin America. There are no emerged in the pattern of regional trade countries in South Asia that have signed a agreements. One unifying characteristic is 30 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S that these take RTAs well beyond agreements (1995), Israel (1995), Morocco (1996), between adjacent countries. For example, Jordan (1997), the Palestinian Authority (1997), Algeria (2001), Egypt (2001), and · The EU's move toward bilateral market Lebanon (2002). In general, services liberal- access FTAs and Economic Partnership ization provisions are limited to the restate- Agreements (EPAs) with the ACP ment of WTO GATS commitments with no countries; new liberalization or with preferential access · The shift in the U.S. position toward reserved for suppliers based in member coun- bilateral preferential agreements; and tries. Dispute settlement is state-to-state based · The effort of a handful of developing on ad hoc arbitration. countries to open markets through RTAs. Partnership and Cooperation Agreements We turn now to a more in-depth investiga- (PCAs) with the Western Balkans, Russia, and tion of these trends. the CIS were designed to help promote stabil- ity on the border of the EU, and in the case of EU Preferential Trade Arrangements Russia, expand trade. The EU has been pro- During the 1990s, the EU was an active spon- viding technical assistance to these govern- sor of bilateral arrangements with individual ments to help implement the institutional re- countries and groups of countries and was the forms that are part of the PCAs. major player in the RTA game. Prior to the Two new agreements have been added to recent accession of 10 new members, the EU this list since 2000. had bilateral or regional agreements with 111 countries. Trade agreements became an inte- · Economic Partnership Agreements gral instrument of European foreign policy, (EPAs) are designed to replace the pref- particularly in the aftermath of the collapse of erential systems embodied in the Conto- the Soviet Union.3 nou Agreement (the successor to the Three types of agreements were intended to Lomé Convention), which had received a stabilize the region after 1989. Europe Agree- waiver under the enabling clause from ments were intended to prepare bordering GATT Article XXIV, a waiver that Eastern European countries for eventual acces- expires in 2007. EPAs are designed to sion into the EU. They involved bilateral agree- promote trade and development in the ments between each other and with the EU to ACP 77 countries in a WTO-consistent reduce tariffs, develop uniform rules of origin, fashion by establishing agreements be- EU-consistent regulatory approaches to ser- tween large groups of countries forming vices, and common treatment of standards as customs unions (box 2.3). well as transition rules in sectors such as agri- · Free Trade Agreements with South culture. These efforts culminated with the full Africa (which entered into force in admission of 10 new countries into the EU in 2000), Mexico (2000), and Chile (2003) 2004--which is why the number of RTAs reg- are designed to open markets and secure istered with the WTO fell for the first time ever. trade. Agreements with the Gulf Cooper- Euro-Mediterranean Agreements were in- ation Council (GCC) and the Common tended to build bilateral trade relations be- Market for the South (MERCOSUR) are tween neighbors, with the objective of form- under active negotiation. These embody ing a NAFTA-like free trade area by 2010. free trade provisions for a range of prod- Launched in 1995, the EU and 12 countries ucts as well as provisions to liberalize at have been involved in talks on "association least some services (Ullrich 2004). agreements" that would subsume some exist- ing bilateral arrangements. To date, bilateral The EU agreements govern services trade in agreements have been signed with Tunisia addition to trade in goods. The agreements 31 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Box 2.3 EPAs become the EU's trade and development instrument: An experiment in "North-South-South" integration E PAs are the most ambitious attempt to harness Southern Africa (16 countries), the Southern African trade, development resources, and technical-legal Development Community (SADC), the Caribbean assistance to the cause of integration-led develop- ACP countries, and the Pacific states (Kiener 2004). ment. The objective is to promote development, The content for the agreements is currently open strengthen regional integration, and ensure compati- for discussion. Reciprocal trade liberalization would bility with WTO principles. By negotiating reciprocal be the centerpiece under the terms of the EPA pro- liberalization with existing South-South regional gram . . . (Most of the EPA countries already enjoy groupings and by providing common rules of origin preferential market access that the EU grants unilat- with cumulative provisions, participants hope to pre- erally under this program.) In addition, the EU has vent the hub-and-spoke effects that plague many stated that it would like to have services liberaliza- bilateral North-South agreements. The EPAs will also tion, investment, competition, government procure- encourage liberalization of services, provide for com- ment, and trade facilitation covered in the agree- mon product standards, and set up the negotiation of ments (Falkenberg 2004). investor protections, based on state-to-state ad hoc Several issues will determine the ultimate effec- arbitration of disputes. tiveness of the EPAs in promoting development: the After a one-year clarification phase by the African degree of additional MFN liberalization in goods Caribbean and Pacific states (ACP), the first negotia- and services markets in both the RTAs and in the tions were launched in October 2003. The EU initi- EU; the restrictiveness of the rules of origin for ated discussions with the Economic Community of goods; and the extent of trade diversion that could West African States (ECOWAS) plus Mauritania, the occur in the event that there are no concomitant re- Economic and Monetary Community of Central ductions in MFN border protections (see Hinkle and Africa (CEMAC) plus São Tomé, Eastern and Schiff 2004). with Mexico and Chile provide for specific Exceptions included only Canada and Israel in liberalization commitments in the financial the 1980s and NAFTA in the early 1990s. In- sector over and above those included in deed, many U.S. trade observers contend that GATS, with the Chilean agreement adding opening NAFTA talks was designed primarily telecommunications and maritime services to support multilateral trade negotiations--to (see Ullrich 2004). The South African agree- spur the Europeans and others into acting on ment alludes to possible services liberaliza- theUruguayRound.Twoyearslater,theClinton tion, but without commitment. The EU agree- administration announced its desire to form a ments differ in important respects from the Free Trade Area of the Americas (FTAA), and it U.S. agreements in that they are generally less signed an FTA with Jordan in 2000. comprehensive, provide less market access in Since the approval of trade promotion au- agriculture, and do not provide for investor- thority in 2002, however, the United States has state dispute resolution (see chapter 5). given much greater emphasis to securing bilat- eral FTAs in tandem with its efforts to achieve The U.S. embraces bilateralism multilateral liberalization through the WTO. Prior to the present administration, the U.S. had Since 2002 the United States has signed bilat- generally eschewed reciprocal preferential eral accords with Australia, Bahrain, Central trade agreements, whether regional or bilateral. America plus the Dominican Republic, Chile, 32 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S Morocco, and Singapore. The United States peanut butter, sugar, and tobacco for the appears to have intensified its pursuit of RTAs United States. Some exclusions are due since the Cancun WTO Ministerial (September to be phased out according to lengthy 2003). Negotiations are officially4 under way timetables; in the Chile-U.S. FTA, for ex- with Colombia, Ecuador, Panama, Peru, ample, all duties will be phased out in SACU, and Thailand. Other economies deemed 12 years (USTR 2004). to be in the queue are Bolivia, Egypt, New · Intellectual property rights are conven- Zealand, Pakistan, the Philippines, South tionally accorded stronger protections Korea, Sri Lanka, and Taiwan (China), and than under the WTO's TRIPS agreement, Uruguay (Schott 2004). This intensified pace with investor-state suits permitted in the may reflect the intention to prod both the mul- event of disputes. tilateral negotiations and the FTAA, as well as · Investment protections, with provisions to respond to U.S. businesses that fear being for national treatment and nondiscrimi- shut out of export markets by a growing num- nation in pre-establishment provisions ber of RTAs in which the United States is not a for companies based in each others member. markets (though liberal rules of origin In the broadest of terms, developing coun- indicate foreign subsidiaries located in tries seek to provide access to their services member countries qualify for eligibility). markets and guarantees in many nontrade · Services trade are to be open except for areas in exchange for assured access to U.S. those excluded in a negative list; notably goods markets. Key facets of these agreements excluded are labor service providers, ex- include:5 cept for the provisional visas held by pro- fessionals associated with investing firms. · Tariff rates on most nonagricultural · Labor and environment issues are in- products are bound at zero; for example, cluded in recent agreements, with signa- the U.S.-Chile FTA will bind duties at tory countries undertaking commitments zero for 85 percent of trade. to enforce their own environmental and · Exclusion or delayed liberalization of labor laws. Dispute settlement panels are sensitive products, commonly including empowered to impose monetary fines agricultural products such as dairy prod- rather than using trade sanctions to force ucts, cotton, ethyl alcohol, peanuts and compliance (box 2.4.) Box 2.4 Labor in U.S. FTAs U ntil NAFTA, the United States did not attempt tion that each member's existing laws are satisfactory to include provisions on labor in trade agree- and therefore any trade distortions that might arise ments that it negotiated. As a presidential candidate, are caused by a lack of enforcement. The agreements Bill Clinton promised to negotiate new side agree- enumerate five core standards: the right of associa- ments to NAFTA on labor in order to secure suffi- tion; the right to organize and bargain collectively; cient political support for NAFTA. Since then labor prohibitions on forced labor; a minimum age for em- issues have featured prominently in Congressional ployment of children; and acceptable working condi- debates on granting the president negotiating tions. The FTAs establish a procedure for making authority and the resulting trade agreements. complaints, encouraging resolution first through con- All recent FTAs negotiated with the United States sultation and, if this fails, by establishing a panel of contain provisions requiring parties to enforce their experts to hear the dispute. own labor laws. These are premised on the assump- (Box continues on next page) 33 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Box 2.4 (continued) The labor provisions break new ground in how a Using fines rather than trade sanctions has several dispute settlement panel's decisions are enforced. advantages: while trade sanctions penalize both pol- Rather than using trade remedies (i.e., granting the luting and clean exporters, fines target the polluters; injured party the right to withhold trade conces- increased trade sanctions hurt all workers in export sions), a panel can impose a monetary fine of up to industries, but fines help restructure plants and $15 million per year (adjusted for inflation). Pay- maintain employment; and fines build in targeted ments of the fines would go into a fund to support solutions to the problem rather than present pro- appropriate labor initiatives, which may include tracted trade disputes. efforts to improve enforcement of labor laws. This mechanism appears in the agreements that the United States has signed with Australia, Bahrain, Chile, Central America and Dominican Republic, Sources: Destler and Balint 1999, texts of FTAs on the USTR Morocco, and Singapore. web site (www.ustr.gov), and Weintraub 2004. The United States indicated it would not nego- over 60 percent of its exports with duty-free tiate changes in its antidumping statutes or on access to markets around the world (see Devlin its agricultural subsidies, insisting on address- and Estevadeordal 2004). ing both through the WTO's multilateral nego- Many existing regional organizations in tiations. In chapter 5, we return to a deeper dis- Africa also moved aggressively to intensify cussion of provision for services, investment, preferential trade liberalization during the and intellectual property rights (IPR). 1990s. For example, the treaty establishing the Common Market for Eastern and South- Developing countries actively pursue ern Africa (COMESA), which was signed in major markets 1993 to replace the Preferential Trade Area, The launching of NAFTA spawned a new called for a free trade area by 2000 and a flurry of interest among developing countries customs union by 2004. The East African eager to use RTAs to secure market access. Community was formed in the mid-1990s to Mexico and Chile have been at the forefront of accelerate economic integration among three these developments. Mexico, having created a COMESA members (Kenya, Tanzania, and world-class trade negotiating team for NAFTA, Uganda). The SADC Trade Cooperation Pro- turned its attention to Central America and tocol was signed in 1996 as part of an effort to other countries in Latin America. It established reintegrate South Africa into the regional econ- arrangements with Costa Rica (1995), Bolivia omy after the end of apartheid. (1995), Nicaragua (1998), the EU (2000), Asian countries have launched similar nego- EFTA (2001), and Japan (2004). After NAFTA tiations since 2001. India has concluded or is was signed, Chile immediately solicited entry negotiating limited arrangements with MER- into the accord. Rebuffed initially, the country COSUR and Thailand; MERCOSUR is negoti- embarked on a wider strategy. Chile estab- ating with the Andean countries; China has lished agreements with MERCOSUR (1996), launched bilateral accords with members of the Canada (1997), Peru (1998), Mexico (1999), Association of Southeast Asian Nations Central America (2002), the United States and (ASEAN), to mention a few. In 2004, India, EU (2003), and EFTA (2004). By 2004, Chile Pakistan, and other South Asian countries an- had signed free trade agreements that provided nounced the South Asian Free Trade Agreement 34 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S Table 2.2 RTAs cover many topics besides merchandise trade Customs Intellectual Dispute Standards Transport Cooperation Services Property Investment Settlement Labor Competition U.S. U.S.-Jordan No No Yes Yes Yes Yes Yes Yes Yes U.S.-Chile Yes No Yes Yes Yes Yes Yes Yes Yes U.S.-Singapore Yes No Yes Yes Yes Yes Yes Yes Yes U.S.-Australia Yes No Yes Yes Yes Yes Yes Yes Yes U.S.-CAFTA Yes No Yes Yes Yes Yes Yes Yes No U.S.-Morocco Yes No Yes Yes Yes Yes Yes Yes No NAFTA Yes No Yes Yes Yes Yes Yes Yes Yes EU EU-South Africa No No No No Yes No Yes No Yes EU-Mexico Yes* Yes Yes* Yes Yes Yes Yes No Yes EU-Chile Yes* Yes Yes Yes Yes Yes Yes No Yes Euro-Med. Agreements No No No No Yes No Yes No Yes* South-South MERCOSUR Yes Yes Yes Yes No Yes Yes Yes Andean Community Yes Yes Yes Yes No Yes Yes Yes Yes CARICOM Yes Yes Yes Yes No Yes Yes Yes Yes AFTA Yes Yes Yes Yes No Yes No No No SADC Yes Yes Yes No Yes COMESA Yes Yes Yes Yes No Yes Yes Yes Yes Other Japan-Singapore Yes No Yes Yes Yes Yes Yes Yes Yes Canada-Chile No No Yes Yes No Yes Yes Yes Yes Chile-Mexico Yes Yes Yes Yes Yes Yes Yes Yes Source: World Bank staff. While EU agreements mention cooperation in most of the subject areas, only those in which specific commitments are under- taken receive a "Yes" rating. *Implementation steps are to be agreed on at a later date. (SAFTA), which is intended to encompass all of agreements, but in nearly all cases politics is as the countries of the region (see Baysan 2004; important as economics. The classic North- Newfarmer 2004). North agreement, the European Union, had its origin in politics (see Schiff and Winters Many RTAs, diverse provisions 2003). The fathers of the European Commu- RTAs have increasingly been designed to cover nity, Robert Schumann and Jean Monnet, much more than liberalization of tariffs and clearly believed that Franco-German integra- quotas. New provisions on enforcement of do- tion through trade and investment would mestic labor and environmental laws have al- produce a new constellation of common eco- ready been mentioned. Table 2.2 gives a flavor nomic interests that would attenuate historic of the range of services and intellectual prop- military animosity. As a first step, they felt erty rights issues that are addressed in current that placing French and German coal and steel agreements. Many of these issues, which are industries under a single authority, the dealt with in more detail in later chapters, European Coal and Steel Community, would have implications for trade, although the pre- make it impossible for either of these histori- cise mechanisms by which trade is affected are cal enemies to use these resources for military not always well defined. purposes against the other. Today, for North-South agreements, Many RTAs, many rationales Northern partners often have a complex mix These recent trends highlight different ratio- of rationales--rooted in foreign policy, com- nales that drive the quest for preferential mercial diplomacy, and development policy. 35 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 "Trade policy has always been the principal motivated countries making the transition instrument of foreign policy for the European from socialism in the 1990s. Mexico under Union" (Sapir 1998). The United States now NAFTA was motivated by a similar objective. appears to be using preferential agreements Guaranteed market access combined with for reasons that are similarly broad. Both the credible domestic reforms can attract foreign EU and the United States seek trade agree- direct investment (see chapter 5). ments that go beyond simple tariff removal to South-South agreements often reflect a po- include rules governing services, protection of litical desire to form or join a broadly based intellectual property, and adherence to health, regional initiative, such as ASEAN, COMESA, labor, and environmental standards. or MERCOSUR. The drive for economic One goal of developing countries seeking integration often begins with political objec- an RTA with a large market, such as the EU6 tives. Like France and Germany in the 1950s, or the United States, is simply to secure market the newly established democracies of the access. One should note, however, that most Southern Cone formed MERCOSUR in the developing countries, especially the least de- mid-1980s in the hopes of damping the tradi- veloped countries (LDCs), already enjoy con- tional military hostility between major re- siderable access to these markets for most gional powers--Argentina and Brazil. SADC manufactured products (whether through uni- originated in the 1980s as a coalition opposed lateral preference programs or because MFN to apartheid in South Africa and has more tariffs are already quite low), and RTAs with recently turned to creating a free trade area. these countries often exclude agriculture and Some observers note that African customs other politically sensitive products. Neverthe- unions and free trade areas are as active in less, RTAs provide some insurance against areas such as conflict resolution as in trade future protectionist policies, and by reaching liberalization. Finally, many see relaxed ten- an agreement "preemptively," they seek to sions between India and Pakistan as the real avoid being left out of a future agreement. payoff from the proposed SAFTA agreement, A second objective is to reinforce internal regardless of what happens to trade barriers regulatory reforms through external treaty in the region. The tentative conclusion of ex- obligations and visible political commitments. isting studies is that RTAs that expand trade Locking in domestic reforms through a for- flows appear to have a substantial dampening eign trade agreement with the EU clearly impact on conflict (box 2.6). Box 2.5 Trade agreements and the environment I t is important to establish coherent relationships WTO. However, in the absence of agreed-on interna- between environmental policies and the trade tional standards (e.g., fisheries), the risk of disguised obligations set out in various RTAs. The following protectionism has prevented further consensus on the examples illustrate the various ways that environ- way forward. Long-standing disputes between the mental issues are handled in these trade agreements. United States and other countries on tuna fishing and WTO. Within the WTO, environmental provisions dolphin or turtle protection are cases in point. are limited to the adoption of product-related mea- NAFTA. The environmental agreement under sures as "necessary to protect human, animal or plant NAFTA created the Commission for Environmental life or health," or "relating to the conservation of ex- Cooperation to promote environmental cooperation haustible natural resources." Process-related require- among the three members. The commission itself does ments continue to remain outside the scope of the (Box continues on next page) 36 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S Box 2.5 (continued) not set standards in the various countries, though part mechanisms for implementing the protocol. The of its mandate is to help harmonize them upward. If a regime is still evolving, and the challenge at hand is country persistently fails to enforce environmental to ensure that the promise of the protocol leads to laws that have conferred a trade benefit, dispute settle- effective regional cooperation and action. ment provisions can be invoked. The commission's Bilateral agreements. A number of recently con- role in the disputes is to see that enforcement of exist- cluded bilateral FTAs, including the U.S.­Singapore ing laws takes place. In addition, it is charged with FTA and the Japan­Singapore Economic Agreement monitoring the environmental effects of NAFTA. Arti- for a New Age Partnership, contain environmental cles of agreement also dictate that countries will not provisions. The U.S.­Singapore FTA establishes an try to attract investment by relaxing or ignoring do- important precedent for dealing with environmental mestic health, safety, or environmental regulations. In- issues by including a chapter specifically on the envi- ternational environmental agreements recognized by ronment. As discussed in box 2.4 on labor laws, this the three parties take precedence over national rules. agreement ensures that countries effectively enforce MERCOSUR. Environmental concerns are cur- their environmental laws, and it provides for en- rently being dealt with in MERCOSUR by a working forcement mechanisms, including fines. group. This group has discussed issues such as the Even in the absence of such special provisions, environment, competitiveness, non-tariff barriers to however, trade agreements can contribute to a trade, and common systems of environmental cleaner environment simply by making trade more information. A draft agreement from this working responsive to market forces. In general, countries group provides for upward harmonization of envi- that are more open to trade adopt cleaner technolo- ronmental management systems and increased coop- gies more quickly, and increases in real income are eration on shared ecosystems, in addition to mecha- often associated with greater demand for environ- nisms for social participation. It also includes mental quality (WTO 1999). Opening up domestic provisions on instruments for environmental man- markets also encourages cleaner manufacturing, agement, including quality standards, environmental because protectionist countries tend to shelter impact assessment methods, environmental monitor- pollution-intensive heavy industries. The incentives ing and costing, environmental information systems to over-exploit or deplete resources are more directly and certification processes, provisions for protecting related to policies and institutions within the sector health and quality of life, and other general than to trade openness per se (World Bank 1999). Not all political objectives involve war and movement of labor, customs, and standards-- peace issues; some South-South agreements that are difficult to broach at the global level. are designed to pool resources for trade nego- RTAs among CIS countries are arguably an tiations and trade policymaking. Much as the attempt to reconstruct some of the economic European Union established a common trade linkages that were severed with the disintegra- policy with a common commissioner in charge tion of the Soviet Union and the disorganiza- of trade (in part to negotiate more forcefully tion caused by the collapse of central planning. with the United States in the GATT), so too a Although many of these regional externalities driving force for MERCOSUR was to estab- can be handled without a trade agreement, lish a common trade policy relative to the mul- RTAs may provide institutions and a frame- tilateral and hemispheric system. work through which to make progress on these Entering into a regional agreement may also issues (see chapter 4). reflect a desire to deal with region-specific is- The wide variation in RTAs flows from the sues--such as transit, water, energy, migration, very different motivations countries have for 37 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Box 2.6 Can RTAs prevent conflict? D oes trade inhibit or increase hostilities between external tariff structure that deprived it of access to states? Greater contact among traders and con- cheaper inputs from the global market and diverted sumers across borders may stimulate mutual respect trade to Pakistan (Schiff and Winters 2003). and more harmonious relations, and high levels of Mansfield and Pevehouse (2000) attempt to iden- trade can create economic interdependence, which, in tify empirically the role of RTAs in ameliorating turn, raises the cost of political disputes and military conflict. They find that, on average, the likelihood conflict. that a pair of states will see the outbreak of a mili- In 1889, Wilfred Pareto suggested that "customs tarized interstate dispute declines by around 50 per- unions and other systems of closer commercial rela- cent if both belong to the same RTA. However, tions [could serve] as means to the improvement of only RTAs that expand trade flows appear to have political relations and the maintenance of peace." In a substantial impact on conflict. When evaluated at 1919, John Maynard Keynes wrote that "a Free the lowest level of trade between partners, it ap- Trade Union, comprising the whole of Central, pears that membership in a RTA reduces the chance Eastern and South-Eastern Europe, Siberia, Turkey, of dispute by just 15 percent. Other studies have and (I should hope) the United Kingdom, Egypt suggested that RTAs that have little impact on trade and India, might do as much for the peace and pros- may actually exacerbate conflict (see Powers 2003). perity of the world as the League of Nations itself." If the gains from trade are not distributed evenly, RTAs also can provide institutions and a forum for example, then the subsequent change in inter- for bargaining and negotiation--to address tensions state power relations can be a source of increased before they erupt in conflict. European integration, tension. Also, rising interdependence may be seen as ASEAN, and MERCOSUR are often cited as venues a source of increasing vulnerability, making expan- for improving political-military relations. Regional sion through military force appear more attractive. trade agreements do not ensure positive political out- These results, which suggest that RTAs could con- comes, however. The U.S. civil war (1861­65) was tribute to a reduced risk of military conflict, should be fought--at least in part--over high protection of treated with a high degree of caution, due to problems northern manufactures and trade restrictions on cot- of causality and omitted factors, such as the broader ton. Similarly, the Central American soccer war of institutional framework governing relations between 1969 emerged out of lingering hostility over trade particular pairs of countries. In Africa, for example, arrangements that created advantages for El Salvador RTAs that address the management of cross-border at the expense of Honduras. And one reason resource issues (such as water) are more effective in Bangladesh seceded from Pakistan was the common reducing military conflict than other RTAs. entering into the arrangements. As we will see strains institutions charged with administer- in subsequent sections, these motivations ing trade agreements. A web of differing trade contribute to greater complexity in rules arrangements can tangle administrative pro- governing world trade. cedures--customs procedures, technical stan- dards, rules of origin, and so on--and thereby Many RTAs can complicate raises the costs for both enterprises and gov- administrative procedures ernments. This complexity undermines work An important feature of the rise in the num- toward greater trade facilitation in developing ber of RTAs is the growing number of over- countries. lapping agreements and the so-called Many agreements between country pairs "spaghetti bowl" that has emerged from the are duplicated by other agreements to which proliferation of bilateral agreements (fig- the same two countries are parties. In Sub- ure 2.2). The associated myriad of rules Saharan Africa, for example, about one-half of 38 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S Figure 2.2 Spaghetti and rigatoni: Multiple, overlapping RTAs, 2004 a. African agreements are overlapping AMU Nile River Basin IGAD ECCAS CEMAC Somalia Algeria Libya São Tomé & Principe Morocco Egypt Mauritania Cameroon Tunisia Central African Rep. Gabon ECOWAS Equat. Guinea Djibouti COMESA Rep. Congo Burundi* Conseil de Ethiopia Ghana Rwanda* L'Entente Cape Verde Chad Eritrea Nigeria Gambia Sudan DR Congo Kenya* Benin Niger Uganda* Togo Burkina Faso Angola Côte d'Ivoire Guinea-Bissau Mali Senegal EAC WAEMU Liberia Guinea Sierra Leone Mauritius* Tanzania* Malawi* Syechelles* Zambia* Comoros* SACU Mano River Union Zimbabwe* Madagascar* CILSS South Africa Namibia* Botswana AMU: Arab Maghreb Union Swaziland* Reunion Lesotho CBI: Cross Border Initiative CEMAC: Economic and Monetary Community of Central Africa CILSS: Permanent Interstate Committee on Drought Control in the Sahel *CBI IOC Mozambique COMESA: Common Market for Eastern and Southern Africa EAC: East African Cooperation SADC ECOWAS: Economic Community of Western African Studies IGAD: Inter-Governmental Authority for Government SADC: Southern African Development Community IOC: Indian Ocean Commission WAEMU: West African Economic and Monetary Union SACU: Southern African Customs Union * Indicates membership in CBI regional grouping Source: Schiff and Winters 2003. b. "Spaghetti Bowl" of RTAs in the Americas and Asia-Pacific (Agreements signed and in force in Latin America as of May 2004) Hong Kong FTAA PR China Taiwan Canada Brunei CACM Costa Panama Rica Cambodia El Salvador Korea Guatemala MERCOSUR Thailand Honduras Paraguay Nicaragua Argentina Brazil Malaysia Japan USA Uruguay Philippines Mexico Myanmar ASEAN Bolivia Singapore Colombia New Zealand Chile Venezuela Laos Ecuador Bahamas Indonesia Australia Andean Peru Vietnam Dominican Community Republic Papua New Guinea APEC CARICOM Dominica Trinidad and Tobago Suriname Grenada Barbados Russia Jamaica St. Vincent and Grenadines Guyana Antigua and Barbuda St. Kitts and Nevis Belize Haiti St. Lucia FTAA APEC Intra-Americas in force Intra-Asia-Pacific in force Intra-Asia-Pacific signed Trans-Pacific signed Source: Devlin and Estevadeordal 2004. 39 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Figure 2.2 (continued) c. In Eastern Europe and Central Asia, bilateral agreements burden customs officials* CEFTA Bulgaria Turkey CIS Czech Republic Latvia BAFTA Slovak Republic Lithuania Poland Estonia Belarus Romania Moldova Tajikistan Hungary Kazakhstan Uzbekistan Slovenia Macedonia Kyrgyz Republic Armenia Russia Croatia Bosnia Ukraine Georgia Azerbaijan BAFTA: Baltic FTA Turkmenistan CEFTA: Central European FTA CIS: Commonwealth of Independent States * Prior to EU expansion. Source: World Bank staff. the pairwise trade relationships covered by an advantageous than in the hub, which enjoys RTA are also covered by another agreement. improved access to all of the spokes. In com- In other regions, overlapping agreements also parison with a broad preferential trade agree- comprise a substantial share of the total num- ment, a hub-and-spoke approach in theory ber of agreements. There would be significant generates lower gains, which accrue mainly to benefits, in terms of lower administrative costs the hub (Wonnacott 1996). Hubs and spokes and more effective implementation, from a are already clearly discernible as the EU and rationalization of the current structure of United States extend restrictive rules of origin overlapping agreements. from one bilateral agreement to another.7 Uneven terms--hub-and-spoke integration Trade within RTAs is rising but The substantial number of bilateral agree- preferential trade is less important ments involving large northern countries, most Trade between RTA members is growing as of which have been signed since 1990, suggests the number of agreements increases, and one- that a hub-and-spoke structure in world trade third of world trade now takes place between is emerging. Of the 109 North-South bilateral RTA members (figure 2.3). (Here we cover agreements, 86 have been created since 1990. only reciprocal agreements and exclude trade In a hub-and-spoke trading system, the largest under the Generalized System of Preferences, markets sign individual agreements with a Cotonou Agreement, and AGOA.) Disregard- wide range of peripheral countries among ing intra-EU trade, bilateral flows between which market access remains restricted. Such RTA members have been growing at a rate agreements can marginalize the spokes, where similar to the growth rate of agreements market access conditions are usually less themselves, as shown in figure 2.3. This figure 40 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S Examining total trade flows between RTA Figure 2.3 Trade within RTAs partners overstates the amount of trade that a. Number of RTAs signed takes place on a preferential basis. However, 250 tariff schedules of many RTA members increasingly contain duty-free MFN rates 200 S-S on which no preference can be given. We estimate that the amount of preferential trade 150 among RTA members, after accounting for 100 MFN rates of zero, is much lower at 21 per- USA cent of world trade (figure 2.3). Furthermore, 50 EU it is often more profitable for enterprises to 0 pay a low MFN tariff when there are high 1990 1996 2002 costs to satisfy rules of origin or other b. Share of world trade covered administrative procedures that a trader must follow to qualify for preferential treatment 35 S-S under an RTA. If we exclude trade covered 30 by facing tariffs of 3 percent or less, we con- 25 USA clude that, at present, the amount of global 20 trade taking place under an economically 15 meaningful tariff preference is around 15 EU 10 percent. An earlier estimate on a similar, but 05 not directly comparable, basis suggests that 0 in the mid-1990s, trade on a preferential 1990 1996 2002 basis amounted to 27 percent of global trade.8 c. RTAs may be less than they appear While the number of RTAs has been in- Share of world trade covered (2002), % creasing, the importance of preferential trade 35 has been falling, reflecting lower tariff barri- 30 ers, especially in OECD countries. Since 1996 25 the number of zero duty lines in the EU tariff 20 schedule has increased from 13 to 21 percent 15 of the total number of tariff lines and from 18 10 to 32 percent for the United States. In 2002 5 about 45 percent of the tariff lines in the EU 0 and United States schedules had duties of Including Excluding Excluding 3 percent or less. This reflects the impact of all trade 0% MFN 3% MFN multilateral liberalization under the Uruguay Note : EUN is EU-15. and earlier trade rounds. Thus a large and Source: World Bank staff. growing proportion of EU and U.S. imports from preferential trade partners is unlikely to actually receive preferential access relative to also reveals that the rapid increase in the num- other countries. ber of South-South RTAs signed in the past For many developing country agreements, decade has not been matched by much change the situation is different because the number of in trade flows among parties to these agree- low-duty tariff lines is small. In 2002, 6 percent ments. This discrepancy highlights a point of Brazilian tariff lines had MFN tariff rates of made earlier: many new South-South agree- zero, as did 1 percent of Indian tariff lines. We ments overlap with existing agreements. estimate that 88 percent of trade between 41 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Box 2.7 Regional versus multilateral and unilateral liberalization:What's more important? H ow does liberalization in RTAs compare to au- chart below shows how trade liberalization is allo- tonomous and multilateral liberalization? The cated according to these different sources. rapid expansion of RTAs has occurred during a pe- riod when developing countries were undertaking autonomous liberalization and also fulfilling commit- ments made during the Uruguay Round of the Share of total tariff reduction, by type of GATT. An examination of tariff reductions by devel- liberalization, 1983­2003 oping countries finds that neither RTAs nor multilat- Regional eral negotiations represent the largest driver of liber- Agreements alization. Autonomous liberalization accounts for the 10% lion's share of trade liberalization since the 1980s. The trade-weighted average MFN tariff rate levied by the 33 largest developing country importers (which collectively account for 90 percent of all de- Multilateral Agreements veloping country imports) was 29.9 percent in the 25% Autonomous 1980s. By 2003 the average MFN rate had dropped Liberalization to 11.3 percent. Based on tariff concessions granted 66% during the Uruguay Round, multilateral negotiations account for 5.1 percentage points of the total decline in MFN tariffs, and the remaining 13.5 percentage points resulted from autonomous liberalization. If the RTAs that these 33 countries have signed were Source: Martin and Ng 2004. fully implemented, the trade-weighted average ap- plied tariff would fall further to 9.3 percent. The countries in Latin America is potentially eligi- Trends in Trade and Growth ble for preferential treatment under an RTA.9 by Region For the Middle East and North African coun- hese agreements were superimposed in a tries it is 83 percent of the total. The new T context of deep changes in global trading SAFTA will lead to three-quarters of the trade patterns.10 The postwar period has seen major between members taking place on a preferen- global shocks and changes in the economic tial basis (assuming all products are included). environment, including oil crises in the 1970s East Asia is an exception, where, for example, and financial crises in the 1980s and 1990s. In 22 percent of Indonesian and 59 percent of the past 20 years there have also been major Malaysian tariffs are zero. Thus the amount of changes in policy regimes. Socialist countries trade between East Asian countries receiving across the world restructured their economic tariff preferences is very small. Like OECD systems and started the process of reorienting countries, however, developing countries have their trade to the world economy. In the former taken great strides to reduce MFN tariffs dur- Soviet Union, this meant collapse and recon- ing the past two decades. Most of this liberal- struction; in East Asia it meant progressive, ization has come from autonomous reductions sustained, and profound institutional change. and not through trade agreements--either Latin America went through its own, if less RTAs or multilateral trade negotiations (see dramatic, transition from import-substitution box 2.7). 42 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S industrialization to a strategy of outward- depressed growth throughout Africa. It has oriented growth. Apartheid in South Africa, been a period of major transitions. political strife in various parts of the conti- The growth performance of developing nent, and the struggle against HIV/AIDS de- countries in the past 20 years reflects this layed the establishment of stable policies and varied experience (figure 2.4). Only the East Figure 2.4 Trade performance has differed across regions a. East Asia and South Asia grew rapidly in 1980­2000, b. Exports became the driving force of growth everywhere but other regions slowed GDP growth rates by region (percent) Non-oil export share of GDP (percent) 8 40 1961­1970 7 35 1970­1980 6 30 5 25 4 2000 3 20 2 1990 15 1 1980 10 0 5 1 1990­2000 1980­1990 2 0 ica an ica an Asia and andica Asia ica Asia and andica Asia ica East Pacific Asia Afr Asia Amer Afr al Ameribbean East Afr East Pacific al ribbean East Afr Europe th South Europe th South and LatinCar Sub-Sahar and LatinCa Sub-Sahar Centr and MiddleNor Centr and MiddleNor Source: World Bank staff. Source: World Bank staff using WITS for non-oil/total exports ratio. c. East Asian exports grew fastest, and all but South Asia d. Integration with the world proceeded hand in hand and Africa increased their share of the global market with regional integration Regional exports as share of world exports (percent)(oil excluded) Intra-regional trade as a share of GDP (percent), 2002 14 30 26.5 12 East Asia 25 and Pacific 10 20 8 Latin America Europe and Middle East 15.3 and Caribbean Central Asia and North Africa 15 6 10 4 6.4 5.3 Sub-Saharan Africa 5 2 3.5 0.8 South Asia 0 0 ica an 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 Asia and andica Asia ica East Pacific Asia al Ameribbean East Afr Afr Europe th South Note: Data obtained from GTAP release 5.4 up to 1998, the and Car Sub-Sahar Centr Latin MiddleNor following years are projected from 1998 using regional growth and rates calculated from COMTRADE. Sources: GTAP 5.4 and COMTRADE. Source: COMTRADE. 43 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Figure 2.4 (continued) e. One reason: East Asia liberalized early in trade f. Strong FDI flows coincided with relocation of production processes Applied average unweighted tariffs FDI, net inflows (percent GDP) 80 3.5 70 3.0 60 1989­1990 2.5 50 2.0 1980s 40 1996­1998 1.5 1990s 30 2003 2000s 1.0 20 10 0.5 0 0 ica an ica an Asia and andica Asia ica Asia and andica Asia ica East Pacific Asia Afr Asia Afr al Ameribbean East Afr East Pacific Ameribbean Afr South al East South Europe th Europe th and LatinCar Sub-Sahar and LatinCar Sub-Sahar Centr and MiddleNor Centr and MiddleNor Source: TRAINS. Note: Period averages for 1980­1989, 1990­1999, and 2000­ 2002 calculated from annual data. For all regions except LAC, some observations are missing. Sources: World Bank, WDI. Asia and Pacific and South Asia regions expe- performance. These trends reflect different ini- rienced higher GDP growth rates in the tial conditions and external shocks, changes in 1980­2000 period than during 1960­1980. development strategies, and policies toward The other four regions fared worse in the last trade liberalization. two decades, with GDP growth rates that East Asia generally followed a strategy of were one-half to two-thirds smaller than export-led growth, through elimination of between 1960 and 1980. anti-export bias and sustained nondiscrimina- Nonetheless, trade grew. The share of trade tory (MFN) liberalization. This trend was in GDP grew in all regions in the 1990s. East most dramatic in China, where border barri- Asian exports grew faster than the other re- ers at the start of the decade were prohibitive gions, and the region increased its share of for most tariff lines, and by the end of the total world exports throughout the 1980s and decade averaged less than 12 percent, with a 1990s. Latin American exports also grew con- plan to go to less than 10 percent by 2004. sistently as a share of the world market during But virtually all countries were liberalizing. the 1990s, but not as steeply as for East Asia. Average tariffs for the region as a whole fell In South Asia, however, although GDP growth from about 23 percent in 1989­90 to 8 per- increased in the 1980­2000 period and the ex- cent in 2003. port share of GDP rose in the latter part of In Europe and Central Asia the disintegra- that period, the trade growth is from a much tion of the Soviet Union after 1990 caused the smaller base. South Asia still has the lowest collapse of growth and required profound re- trade shares of any region. Sub-Saharan structuring of the region's economies and a Africa has also had disappointing growth redirection of its trade. Roughly half of the 44 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S region was drawn to the magnet of the EU's in the last decade, but still remain quite low at large and stable market. The EU responded less than one percent. with technical assistance and a political will- South Asian countries other than Sri Lanka ingness to admit its Eastern European neigh- neither liberalized trade rules nor the rules gov- bors as full members. The combination of a erning inflows of foreign direct investment political framework, trade, investment, and until the 1990s. Removal of the most egregious technical assistance led to an unprecedented forms of anti-export bias and gradual domestic pace of reforms and economic integration that reforms, together with textile preferences, culminated with 10 states joining the EU on produced a rapid expansion in garment/textile May 1, 2004. With their eyes turned toward exports, and led to high growth rates for ex- markets in the EU, the Central European and ports in the 1990­2000 period and an increas- Baltic countries achieved more extensive inte- ing share of exports in GDP. Since growth was gration and higher trade and FDI flows, which from a low base, South Asian exports as a is evident in the rapid export growth of the share of world trade have remained low region as a whole. throughout the 1980­2000 period. South Asia The CIS has moved much more slowly in maintained the highest levels of average ap- its process of reform and reorientation, partic- plied tariffs, even compared to the import-sub- ularly in Central Asia and the Caucasus. stitution industrialization period of other re- Under the CIS-7 initiative, trade regimes have gions. However, this is changing. Nepal been generally liberalized, but have been lim- launched trade liberalization in the early ited by regional trade and transit barriers. 1990s. Sri Lanka and then Pakistan in 1997 Latin America reversed its trade policy began to reduce their border barriers and in- stance, and during the 1990s average tariffs in crease their trade with the world economy. the region declined from over 30 percent to India began to reduce border protection from 12 percent. The region's share of the world very high levels in the early 1990s and has con- markets increased and net inflows of foreign tinued doing so; in early 2004, India an- direct investment (FDI) as a percent of GDP nounced tariff cuts of roughly one-third, steadily climbed, reaching 5 percent of GDP in reducing the average tariff rate to about 22 per- 1999--higher than East Asia. Overall, FDI net cent. Bangladeshi border protections are still inflows in the latter half of the decade more among the highest in the world, but they too than doubled from an average of 1.4 percent announced reductions in 2004.11 The region of GDP in the first half to an average of 3.6 remains only minimally integrated in world percent in the second half. capital markets. Net inflows of FDI, although In the Middle East, policy and economic higher than in the early 1980s, are less than 0.8 barriers, together with a reliance on oil for percent of GDP--the lowest of all the regions. several countries, prevented rapid growth in As with the Middle East and North Africa trade. High tariff rates, restrictions on services and South Asia regions, Sub-Saharan Africa entry, and controls on agriculture interacted remains weakly integrated into the global with poor investment climates to impede trade market. Although exports as a share of GDP and keep transactions costs high. A large in Sub-Saharan Africa increased in 2000, ex- state-led sector also shaped a noncompetitive ports as a share of world exports have re- industrial policy that discouraged trade. Aver- mained flat throughout the last decade and are age tariff rates were almost 30 percent in the lower than in the early 1980s. GDP growth late 1990s, mirroring the import substitution has also been worse than in the earlier policies early in Latin America and more decades. Many countries in Sub-Saharan recently in India. Flows of foreign direct in- Africa are dependent on only a handful of vestment as a percent of GDP have recovered commodities with highly volatile prices; most 45 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 face very high transport costs and have weak strategies that have led to increasing exports, institutions to facilitate trade. These countries especially of manufactures. The share of man- have also experienced a number of armed ufacturing in exports from East Asia, for ex- conflicts throughout the previous decades and ample, increased from about 52 percent in are plagued by endemic diseases such as 1981 to 88 percent in 2001, while the share in malaria and HIV/AIDS, which have major Latin America tripled from about 20 percent impacts on their economies and societies. All to 60 percent. these factors hobble trade performance. Trade has allowed manufacturers to ex- ploit economies of scale, specialization, and scope. This is reflected in the growing share Changing Export Composition of parts and components in total exports. In and the Rise of Global the three more open regions--East Asia and Pacific, Europe and Central Asia, and Latin Production Networks America and the Caribbean--parts and T he differential in trade and growth per- components trade has surged. This interna- formance reflects the fact that certain re- tional segmentation of production--"pro- gions have been better placed--in part duction chains" in which intermediate in- through the policies they adopted--to take puts are traded and transformed into more advantage of new technologies and changes processed intermediate inputs, which are in the nature of world trade. Not only has the then moved across borders to the next stage volume of international trade expanded in the in production--has been a major factor dri- postwar period, but also its structure has ving the surges in intra-regional trade in changed in three fundamental ways. First, ex- those areas. ports of manufactured products from devel- One indicator of specialization is the im- oping countries, and trade in manufactures port content of exports. To measure the role among them, have become increasingly im- of imported intermediates in trade, we calcu- portant for all regions. Second, trade integra- lated an index of vertical specialization, tion has allowed developing countries to spe- which measures the share of the value added cialize (most evident in the emergence of of an export accounted for by imported inter- production chains), with trade in intermedi- mediate inputs, either directly as imported in- ates becoming more important. This trend is puts in the exporting sector or indirectly also evident in the role that new products through the use of imported inputs in the do- play in production. Finally, foreign direct in- mestic production of intermediate goods used vestment is playing an ever-increasing role in by the exporting sector.12 Vertical specializa- the integration process. These developments tion is most important in East Asia, and least have facilitated the integration of countries important in South Asia and Sub-Saharan that have adopted relatively open trade poli- Africa (figure 2.5c). cies, and have increased the disadvantages The evolution of production chains and facing countries that have segmented them- finer division of the production processes selves from global markets. across countries, including developing coun- tries, allows producers to exploit potential Specialization in manufactures efficiency gains from: (1) local increasing Manufactured products as a share of exports returns to scale in the production of interme- increased strongly between 1981 and 2001 for diate inputs, (2) regional differences in factor all regions (figure 2.5). Countries in East Asia costs for different parts of the production and later, Latin America and Eastern and Cen- process, (3) increased competition from a tral Europe, have followed open development wider market, and (4) technology transfer 46 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S Figure 2.5 Composition of trade has changed a. Manufactures have become increasingly important Manufactures: Share of exports by region, % 1981 East Asia and Pacific 2001 Europe and Central Africa Latin America and Caribbean Middle East and North Africa South Asia Sub-Saharan Africa 0 10 20 30 40 50 60 70 80 90 100 Source: UN COMTRADE. b. Parts and components have increased for all regions c. Vertical integration with the global economy has increased Exports of parts and components as a share of total exports (percent) Import content of exports (percent) 16 50 14 45 Indirect 40 12 35 10 30 8 25 6 20 1980 2002 15 4 Direct 10 2 5 0 0 ica an ica an Asia and Eastica Asia ica Asia and Eastica Asia ica East Pacific Asia Afr Asia Afr al Ameribbean Afr East Pacific Ameribbean Afr South al South Europe Middleth Europe Middleth and LatinCar Sub-Sahar and LatinCar Sub-Sahar Centr Nor Centr Nor and and and and Source: UN COMTRADE. Sources: Data from GTAP (ver. 6.03) and World Bank for 2001. d. New products propelled export diversification Share of 1980's bottom 10% of manufactured exports in total intra-regional exports 50 Middle East and North Africa 45 Europe and Central Asia 40 East Asia and Pacific 35 Sub-Saharan Africa 30 25 20 South Asia 15 Latin America and Caribbean 10 5 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 Source: UN COMTRADE. 47 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Figure 2.5 (continued) e. FDI flowed mostly into East Asia and Latin America f. A positive trend between FDI and parts exports for the regions Share of region's net FDI inflows Parts and components, FDI, GDP: 2000 in the region's total investment (percent) FDI, percent of GDP 18 4.5 LAC 16 4.0 14 3.5 ECA 12 3.0 10 2.5 EAP SSA 8 2.0 6 1.5 4 1.0 Y 0.2333x 0.3067 2 0.5 SAS R2 0.6801 MNA 0 0 ica an 0 2 4 6 8 10 12 14 16 Asia and Eastica Asia ica East Pacific Asia Parts and components, percent of exports Afr al Ameribbean Afr South Europe Middleth and LatinCar Sub-Sahar Centr Nor and and Note: Period averages for 1980­1989, 1990­1999, and 2000­2002 calculated from annual data. Both net FDI (BoP) and gross fixed capital formation are in current US dollars. For 1980­1989 in ECA, GFCF was estimated using a three-year average share of GFCF in GDP over the 1990­1992 time period and back-calculating GFCF from GDP for the entire 1980­1989 period. GDP time series (for 1980­1989 ECA only) obtained from UN national accounts. Sources: World Bank, WDI. Sources: World Bank, WDI, and UN COMTRADE. from developed countries embodied in im- may have the important advantage of allow- ported intermediate inputs and backward ing countries to escape the deterioration in the linkages through exports. The magnitude of terms of trade that would come from trying to these links between increased trade in inter- increase market share in existing products.15 mediates and productivity growth in develop- To assess this phenomenon, we reviewed the ing countries has been studied in both cross- trade performance of the least traded decile of country analysis and country case studies.13 product categories. In East Asia and Pacific, While causation is difficult to establish, the those products that figured in the lowest 10 evidence indicates that such links are impor- percent of all EAP manufactured exports to the tant, and productivity growth associated with world in 1981 grew to almost 40 percent by increased trade in intermediates is a poten- 2001 (figure 2.5). For the other five regions, tially important source of growth. the performance of products among their low- est initial 10 percent was also noteworthy. Trade in new products Countries are building dynamic new markets A large part of the expansion in exports in for their existing exports and developing new countries undergoing liberalization and suc- variations of old products to replenish the cessful trade expansion comes from products product cycle. This trend is also associated that were not traded--or minimally traded-- with increased trade in intermediates; detailed prior to liberalization (see Kehoe and Ruhl analysis indicates that many of the new export 2002).14 Growth in trade in new products goods are intermediate inputs. Increased trade 48 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S in new products is thus part of the virtuous cir- Preferential Trade and cle linking trade and growth. Regional Outcomes any historical factors, not just preferen- Investment, handmaiden of trade M tial trading arrangements, contributed Foreign investment has been a driver of to these trends. In the next chapter we provide integration, increased trade in manufactures, a more detailed analysis of the impact of RTAs and vertical specialization. As tariff barriers on trade. Here we simply highlight that the na- have come down in manufactures, market- ture of RTAs and the context in which they seeking, horizontal FDI that once led the way in have been applied have varied enormously the import-substitution process has faded in im- across regions and that regional agreements portance relative to efficiency-seeking, vertical often follow, rather than determine, changes in FDI that looks to locate segments of production regional trade patterns. This suggests that pref- in the lowest-cost site. This form of investment erential trade agreements are just one of many is associated with the rise in production chains factors affecting trade outcomes and that when and trade in components and parts. implemented in a highly restrictive economic FDI has increased as a share of GDP in all re- environment, they are usually inconsequential. gions. This trend abated somewhat since the East Asia crisis in 1997­98 and the global re- cession of 2001­02, but FDI growth is likely to History shapes trading patterns resume with the recovery of the global econ- Differing regional performances in trade and omy in 2004. East Asia and Latin America-- growth have roots that go deeper than just the largest markets--have had and have boundaries on a map. Trade patterns--who retained by far the largest shares of FDI trades with whom--have grown out of long throughout the period (figure 2.5). political and economic histories that preceded In East Asia and Pacific, the increase in FDI the trends evident in the last two decades. The supported the pattern of segmentation and re- clusters of trading partners often bear little re- location of production processes within the re- lationship to arbitrary definitions of regions gion. In the 1990s, a large part of the FDI into (see Anderson and Blackhurst 1993, for an ear- Latin America was due to the privatization lier analysis). Major trading blocs--that is, process the region underwent during this pe- those countries that trade more with each other riod. There is broad correspondence between than with those outside their group--emerge FDI trends by region and the share of parts and from a cluster analysis. These blocs are not de- components intermediates in regional exports. fined as traditional geographic political re- Technology transfer from developed to de- gions, but rather by statistical patterns in trade veloping countries is linked to trade, especially flows over decades. through trade in manufactures and intermedi- ates, and also through foreign direct invest- The bipolar world of the 1960s ment.16 The better economic performance of Coming out of the postwar period, the struc- East Asia and Pacific can be seen as resulting ture of world trade by the 1960s reflected a from the emergence of a "virtuous circle" or bipolar world, in which Europe and the synergy between increased specialization in United States had effectively formed blocs production, increased trade in intermediates, with some of their close neighbors, former increased foreign investment, increased factor colonies, and/or cold war partners; and with productivity, and increased growth. This hub-and-spoke links to most of the developing region started earlier and appears more suc- countries. The two leading world trade blocs cessful than other regions in achieving and sus- effectively accounted for 80 percent of global taining this virtuous circle. trade (figure 2.6). The European bloc was 49 Figure 2.6 Evolving trading blocs In the 1960s, the European Union and United States dominate trade . . . 1960s Percent 60 Asia-U.S. 52.7 50 Asia-UK U.S. 40 Europe 29.9 30 20 10.3 10 7.1 0 Europe U.S. Asia-UK Asia-U.S. . . . but by the 1970s, Japan and Korea begin to lead an East Asian bloc . . . 1970s Percent 50 44.9 Rest 40 South America East Asia North America 30 Europe 24.1 19.7 20 10 2.7 0 Europe North East and South America South America East Asia . . . a decade later, the East Asian Tigers, ASEAN countries, and Australia consolidate the East Asia bloc . . . 1980s Percent 60 48.3 50 Rest South America 40 East Asia North America 30 Europe 23 18.8 20 10 2.8 0 Europe North East and South America South America East Asia . . . and in the 1990s, ECA emerges and East Asia trades more with itself than with the U.S. and EU. 1990s Percent 50 Rest Andean 45.8 South Africa 40 East Asia MERCOSUR North America 30 Europe 27.2 19.9 20 10 1.6 0 Europe North East and MERCOSUR America South East Asia Source: Robinson and Diaz Bonilla 2004. 50 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S closely linked with countries in Africa and for- blocs. To the contrary, trade among these blocs mer colonies. The United States was closely is intensifying, becoming more diversified, and linked to Latin America.17 Britain still re- linked with a web of business ties across tained leadership in a small Asian cluster con- oceans that bind the world market together. sisting of former colonies, China, and the rest Second, in the 1990s, the emergence of new of the Middle East, much as the United States poles of commerce--MERCOSUR and South led a cluster of countries closely linked to the Africa--indicates that the process of segmenta- post-WWII political order in the Pacific-- tion and new bloc formation in world trade is Indonesia, Japan, Korea, and Taiwan. The de- still evolving. Third, it is clear that many of the pendent developing countries traded much emerging preferential arrangements have deep more with Europe or the United States, and roots in historical trading patterns, but that not as much among themselves. some of the more recent bilateral FTAs are going beyond these historical patterns. The realignment and the emergence of East Asia in the 1970s and 1980s RTAs in different regions have In the 1970s and the 1980s, a realignment of different impacts world trade began. The European and U.S. Patterns of regional integration tend to confirm blocs fragmented in the 1970s, and their dom- the view that usually trade has preceded trade inance dissipated to 65 percent of global trade. agreements. For East Asia, integration with the The East and Southeast Asian countries left the global economy was a strong impetus for re- European bloc and effectively formed a new gional integration. Exports to the world pro- bloc (East Asia).18 This bloc consolidated in duced demand for imports from neighboring the 1980s, increasing the share of within-bloc countries. As Korea matured and China trade and expanded membership to include opened, internal regional growth assumed its Australia and New Zealand. Its export share own dynamic. As a share of GDP, intra-regional shifted toward the United States (36.2 percent trade in 2002 was 26.5 percent, twice as high as in the 1980s compared to 26.4 percent in the in the next highest region, yet regional trade 1970s). It also represented a growing share of preferences were very modest at best. total world trade--23 percent in the 1980s In fact, regional preferential trading compared to 16 percent in the 1970s. arrangements followed rather than preceded this regional integration. The ASEAN Free Consolidation and diversification Trade Area (AFTA), established in 1992, ac- in the 1990s celerated these tendencies and contributed to By the 1990s, earlier trends blossomed into a Southeast Asia's integration, but much of the tri-polar world. One new element was the tariff reduction has accompanied rather than breakup of the Soviet Union and Moscow's preceded these patterns of regional integra- central management of trade in Eastern tion. However, ASEAN leaders accepted in Europe. Trading patterns began gravitating the Bali Declaration the need to pursue deeper toward the EU. A second new element was integration and to create a single market to the emergence of a new grouping: Argentina, enhance the competitiveness of the region. Brazil, Paraguay, and Uruguay (i.e., MERCO- The importance of preferential trade in the re- SUR).19 Finally, detailed analysis indicates a gion was dramatically increased by the sign- nascent bloc forming around South Africa, ing of a FTA between ASEAN countries and including the SACU countries. China. These global trading patterns are revealing: Regional arrangements were critical for the First, today's tri-polar world of global com- successful integration of Eastern Europe into merce does not signify that the world is evolv- the world economy, but have not been as ing into three disparate, autarchic trading successful in the CIS countries. With the 51 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 assistance of the EU in its Europe Agreements entered into the Pan-Arab Free Trade Agree- and with the aspiration of WTO membership, ment and most participated in the sub-re- the Eastern European countries have moved gional customs union, the Gulf Cooperation swiftly toward integration. The CIS has been Council. Even so, these agreements have not burdened with incomplete reforms, a poor in- been sufficient to overcome the effects of high vestment climate, and a plethora of trading border barriers and restrictions on services. arrangements that have been implemented The Euro-Med agreements with the EU have only partially. The combination has weighed fallen short in their aspirations because of re- down the subregion's performance. strictions on trade in agriculture, services, and In Latin America, the intra-regional share labor; lack of harmonization of standards; of exports in GDP in 2002 remains only one- and stringent rules of origin.21 fourth of East Asia's share. Since the 1960s, Regional agreements in South Asia, as with with the formation of the Latin America Free Latin America in the 1960s and 1970s, floun- Trade Area (LAFTA), the region has struggled dered on the shoals of high protection. The to expand trade. This proved futile as long as 1993 South Asia Preferential Trade Agreement import substitution policies were in place and (SAPTA) was stillborn, given continuing high state enterprises were used as instruments of levels of protection, a lack of meaningful con- industrial policy. Early attempts in Central cessions, domestic political problems, hostility America and the Andean countries failed be- between India and Pakistan, India's ban on cause of the inherent difficulty in managing imports of all consumer goods (from SAPTA potential trade diversion and location of in- countries until 1998 and from the rest of the dustries within the regions behind high exter- world until 2001), and India's control over nal protection barriers.20 major primary goods (Baysan 2004). This situation changed with the wave of Recently, however, unilateral trade reforms unilateral reforms in the 1985­95 period. in India and Pakistan, political rapprochement, Mexico's reforms paved the way for the later as well as concerns about rising preferential creation of NAFTA in 1994, and reforms in tariff arrangements in other parts of the world, Brazil and Argentina led to the creation of led to the formation of the SAFTA Agreement MERCOSUR in the same year. Similarly, the in January 2004 (Newfarmer 2004). Central American countries, with a second go, In Sub-Saharan Africa, regional trade managed to put in place a successful common agreements are common and reflect an aspi- market in the 1990s. As a result, intraregional ration to overcome the limits that small sov- regional trade has proceeded pari pasu with ereign states impose. These include SACU-- growth in the external markets. one of the oldest customs unions in the In the Middle East, intra-regional trade has world--CEMAC, COMESA, ECOWAS, and failed to gain dynamism. Because countries the East African Community. Although aver- begin with broadly similar production and age applied MFN tariffs were cut by half be- export structures, the scope for using regional tween the 1990s and 2003, non-border barri- trade to establish patterns of specialization ers restrict internal trade. The recent regional and diversification in manufacturing produc- trade agreements have had more impact on tion is limited. Intra-regional trade cannot be outward-looking MFN trade liberalization, a substitute for extra-regional trade. and thus on external trade, than on intra- Several countries signed bilateral trade regional trade. The economic impact of these agreements with the EU as part of the Euro- agreements appears to have been small, espe- Mediterranean agreements. Jordan and cially compared to pre-independence arrange- Morocco also signed agreements with the ments that essentially validated existing eco- United States. All countries in the region nomic links (SACU, the West African 52 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S Economic and Monetary Union, and are considerably more ambitious in content CEMAC). The EPAs under negotiation with and coverage than South-South arrangements the EU may reinforce this outward-looking and reach deep behind the border to include pattern of trade integration, but the hope is services, protection of investment rules, and that they will also aid Africa's own regional intellectual property rights. integration if they succeed in fostering eco- In general, the wave of preferential trading nomic reform and performance.22 arrangements followed--rather than preceded The situation is different for agriculture --an intensification of regional trade. Regions and food products. Here margins of prefer- with the lowest external (MFN) border barri- ence are more substantial in all regions except ers ironically have developed the deepest South Asia. The average margin of preference intra-regional links and have been best posi- in the high-income region is similar to that in tioned to diversify and exploit the emergence East Asia, Eastern Europe, and Latin America. of global production chains in the manufac- Again, preferences are greatest in those re- turing sector. East Asia is the starkest exam- gions showing the lowest degree of regional ple, but Eastern Europe, in the wake of the integration--the Middle East, and Sub- dissolution of the Soviet Union, and Latin Saharan Africa. America, with the end of import-substitution Trade preferences have had very little im- industrialization, are not far behind. pact on the high levels of intra-regional trade What are the economic consequences of in manufactures in East Asia and Eastern these arrangements? That is the subject of the Europe. Regions that offer substantial trade next three chapters. preferences behind high external barriers have not fared well in stimulating the growth of Notes intra-regional trade. 1. The recent accession of 10 new members to the European Union reduced the total number of RTAs in force from 285 to 229. 2. In fact there are only 12 countries that are not Conclusion recorded as being party to a RTA, and many of these D eveloping countries have increased their are small islands and principalities. The 12 are Ameri- share of global trade as multilateral trade can Samoa, Bermuda, Channel Islands, Guam, Isle of Man, Monaco, Mongolia, N. Mariana Islands, Palau, negotiations have led to sustained reductions Puerto Rico, Timor-Leste, and the Virgin Islands. in border protection for manufactured prod- 3. We are indebted to Gaspar Fontini of the Euro- ucts. At the same time, and for a variety of rea- pean Commission's Directorate General for Trade for sons, the preferential trade agreements have this formulation. proliferated. While the number of preferential 4. The U.S. Trade Representative is required to of- agreements has increased rapidly, their trade ficially notify the U.S. Congress of its intent to negoti- coverage is substantially less than their official ate FTAs. 5. See Weintraub (2004) and Schott (2004) for a span of influence. Because many tariffs have more detailed discussion. come down close to zero, rules of origin re- 6. In the course of this discussion, the EU is treated strict preferential access, and many products as a single entity. For example, an EU-Mexico agree- within agreements are excluded. Nonetheless, ment is classified as bilateral. RTAs are leading to a more complex trading 7. The provision for regional cumulation in the system and inefficiencies in customs adminis- rules of origin, particularly full cumulation, will tend tration; high tariffs in certain regions still risk to offset the hub-and-spoke system. See Brenton and Imagawa (2004). The EU, following substantial criti- significant trade diversion. cism of the hub-and-spoke system that emerged in Notable differences are emerging between Europe in the late 1990s, moved to create pan- North-South bilateral agreements and South- European cumulation, although in terms of the more South arrangements. North-South agreements limited partial cumulation. 53 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 8. See Grether and Olarreaga (1999), who include 17. The definition of "closely linked" is countries GSP preferences but a smaller sample of countries. that have such large trade shares with their bloc part- 9. In other words, 88 percent of intra-regional ners, particularly the United States and Europe, that trade takes place among RTA partners. adding them to the bloc increases the average within- 10. This section benefited from the World Bank's bloc trade share. regional chief economists at a workshop titled Region- 18. Plus one region, "rest of Middle East and North alism, Trade and Development, May 5, 2004. Sadiq Africa" (MENA), which is closely linked to Europe, but Ahmed, Harry Broadman, Alan Gelb, Homi Kharas, has significant trade with East and Southeast Asia. Mustafa Nabli, and Guillermo Perry made presenta- 19. The within-bloc trade share for MERCOSUR tions that became the basis of this discussion. of 23 percent is lower than its trade share with Eu- 11. The announced reforms were to reduce "sup- rope (27 percent), but given the increasing trend of its plementary duties" that were, in legal terms, additional within-cluster trade share and the legal formation of to ad valorem tariffs; the economic effect is to reduce MERCOSUR in 1995, it makes sense to designate it effective tariffs. as a bloc. Latin American countries fall into three 12. Hummels, Ishii, and Yi (2001) define an index groups. One group is linked closely with North Amer- of vertical specialization (VS). For direct effects for ica, including RB de Venezuela, Colombia, and rest of country k: Central America. The MERCOSUR countries define a uAMX second group, a trade bloc with diversified exports VSk Xk (a slightly higher share to the EU than to the United where u is a 1 x n vector of 1's, Am is the n x n im- States). Finally, the third group is a heterogeneous col- ported coefficient matrix, X is an n x 1 vector of exports, lection of countries (Peru, Chile, and Other Andean n is the number of sectors, and Xk is the sum of exports Pact) with exports largely to the EU, partly to East across the n sectors. When indirect effects are added, and Southeast Asia, and with little trade among themselves. uAM[I AD] 1X VSk 20. IDB, Beyond Borders, 2002. Xk 21. Chief Economist, MENA, 2004. the index becomes: , where I is the identity matrix and 22. Chief Economist, Africa, 2004. AD is the n x n domestic coefficient matrix. Typically, given data limitations, measures of vertical specializa- tion are imperfect. For example, a sector which pro- References duces both for exports and domestic markets, which is Acemoglu, D., S. Johnson, and J. A. Robinson. 2001. common given the aggregation available for sectoral The Colonial Origins of Comparative Develop- data, is assumed to have the same production technol- ment: An Empirical Investigation. American ogy, particularly use of imported intermediates, for Economic Review 91, no. 5: 1369­1401. goods sold in either market. Even with these limita- Anderson, Kym, and Richard Blackhurst, eds. 1993. tions, the measures provide a good picture of the Regional integration and the global trading sys- changing role of trade in intermediates. tem. New York: St. Martin's Press. 13. Winters (2004, 10­11) reviews the literature Baldwin, R. 2003. Openness and Growth: What's the on links between trade in intermediates and productiv- Empirical Relationship? National Bureau of Eco- ity growth. nomic Research, Working Paper no. w9578. 14. See Kehoe and Ruhl (2002). They present lim- Baldwin, R., and A. Venables. 1995. Regional Eco- ited empirical evidence and then suggest a theoretical nomic Integration. In Handbook of International model that captures the phenomenon. Increased trade Economics, vol. 3, eds. Grossman and Rogoff, in new products is difficult to capture in standard mod- Amsterdam: North Holland Press. els of world trade, so such models will tend to over- Baysan, T. 2004. South Asia: Lessons and the Way state terms-of-trade effects from changes in trade. Forward. Mimeo, World Bank, Washington, DC. 15. See Hummels and Klenow (2002). Bhagwati, J., and A. Panagariya, eds. 1996. The 16. Winters (2004) surveys work on the links be- Economics of Preferential Trade Agreements. tween trade, productivity, and growth in developing Washington, DC: American Enterprise Institute countries. See also Nishimizu and Robinson (1984), Press. and Esfahani (1991) who consider the links in semi- Brenton, Paul, and Hiroshi Imagawa. 2004. Rules of industrial countries. Schiff and Wang (2004) consider Origin, Trade, and Customs. In Customs Mod- the link between TFP growth and regional trade ernization Handbook, Conference edition, eds. links. Keller (2002) considers the mechanisms Luc De Wulf and Jose Sokol. Washington, DC: involved. World Bank. 54 R E G I O N A L T R A D E A N D P R E F E R E N T I A L T R A D I N G A G R E E M E N T S Burfisher, Mary, Sherman Robinson, and Karen the EU-World Bank Retreat on Preferential Trade Thierfelder. 2004. Forthcoming. Regionalism: Agreements, July. Old and New, Theory and Practice. In Agricul- Mansfield, Edward D., and Jon C. Pevehouse. 2000. tural Policy Reform and the WTO: Where Are Trade Blocs, Trade Flows, and International We Heading? Eds. G. Anania, M. E. Bohman, Conflict. International Organization 54 (Au- C. A. Carter, and A. F. McCalla. Cheltenham, tumn) (4): 775­808. UK & Northampton MA, US: Edward Elgar. Martin, W., and F. Ng. 2004. Sources of Tariff Reduc- Destler, I. M., and Peter J. Balint. 1999. The New Poli- tions. Background paper. tics of American Trade: Trade, Labor, and the Melo, J. de, and A. Panagariya, eds. 1993. New Environment. Washington, DC: Institute for In- Dimensions in Regional Integration. Cambridge: ternational Economics. Cambridge University Press. Esfahani, H. S. 1991. Exports, Imports, and Economic Newfarmer, R. 2004. SAFTA: Promise and Pitfalls of Growth in Semi-Industrialized Countries. Journal Preferential Trade Arrangements. Background of Development Economics 35. paper for Global Economic Prospects 2005, May, Falkenberg, Karl F. 2004. EPAs and Doha Develop- World Bank, Washington, DC. ment Agenda--Parallelism or Crossroads Nishimizu, Mieko, and Sherman Robinson. 1984. Trade Negotiations Insights 3, (July) (4). Trade Policies and Productivity Change in Semi- www.ictsd.org. Industrialized Countries. Journal of Development Frankel, J. A., E. Stein, and S-J. Wei. 1996. Regional Economics 16, (1­2): 177­206. Trading Arrangements: Natural or Supernatural? OECD. 2003. Regionalism and the Multilateral American Economic Review 86, (2): 52­56. System. OECD, Paris. Grether, Jean-Marie, and Marcelo Olarreaga. 1999. Robinson S., and C. Diaz-Bonilla. 2004. The evolution Preferential and Nonpreferential Trade Flows in of global trading blocs in the past forty years: World Trade. In Trade rules in the making: Who trades what with whom, and why. Back- Challenges in regional and multilateral negotia- ground paper. tions, 159­79. Washington, DC: Brookings Robinson, S., and K. Thierfelder. 2002. Trade liberal- Institution Press and Organization of American ization and regional integration: the search for States. large numbers. Australian Journal of Agricultural Hinkle L., and M. Schiff. 2004. Economic Partnership and Resource Economics. 46, (4) 585­604. Agreements between Subsaharan Africa and Sapir, Andre. 1998. The Political Economy of EC Re- the EU: A Development Perspective on the Trade gionalism. European Economic Review. 42, Components, Africa Region Working Paper (3­5): 717­32. (draft). February, World Bank, Washington, DC. Schiff, M., and L. A. Winters. 2003. Regional Integra- Hummels, David, Jun Ishi, and Kei-Mu Yi. 2001. tion and Development. Washington, DC: World The nature and growth of vertical specialization Bank. in world trade. Journal of International Schiff, M., and Y. Wang. 2004. North-South Technol- Economics (Netherlands) 54 (June) (1): 75­96. ogy Diffusion, Regional Integration, and the Hummels, David, and Peter J. Klenow. 2002. The Dynamics of the `Natural Trading Partners' variety and quality of a nation's trade. National Hypothesis. World Bank, June, Washington, Bureau of Economic Research, Working Paper DC. Series 8712, January. Schott, Jeffrey J. 2004. Free Trade Agreements: Boon IDB. (Inter-American Development Bank). 2002. Be- or Bane of the World Trading System. In Free yond Borders: The New Regionalism in Latin Trade Agreements: US Strategies and Priorities, America. Economic and Social Progress in Latin ed. J.J. Schott. Washington: Institute for Interna- America. Inter-American Development Bank, tional Economics; Song and Xu. 2000. Washington, DC. Ullrich, Heidi. 2004. Comparing EU Free Trade Kehoe, Timothy J., and Kim J. Ruhl. 2002. How Agreements­Services. European Centre for Devel- Important is the New Goods Margin in Interna- opment Policy Management. July. tional Trade? University of Minnesota. USTR. 2004. Free Trade with Chile, Trade Facts. Keller, Wolfgang. 2002. Trade and the Transmission of www.USTR.gov. Technology. Journal of Economic Growth 7, (1): Weintraub, Sidney. 2004. Lessons from the Chile and 5­24. Singapore Free Trade Agreements. In Free Trade Kiener, Christophe. 2004. Economic Partnership Agreements: US Strategies and Priorities, ed. J. J. Agreements: A New Approach to ACP-EU Eco- Schott. Washington: Institute for International nomic and Trade Cooperation. Paper presented to Economics. 55 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Winters, L. Alan. 2004. Trade Liberalization and World Bank. 1999. Trade, Global Policy, and the Economic Performance: An Overview. The Eco- Environment. World Bank Discussion Paper No. nomic Journal, 114 (February): F4­F21. 402. Washington, DC: World Bank. Winters, L. Alan, Neil McCulloch, and Andrew World Bank. 2003. Trade, Investment, and Develop- McKay. 2004. Trade Liberalization and Poverty: ment in the Middle East and North Africa: En- The Evidence So Far. Journal of Economic gaging the World. Washington DC: World Bank. Literature, XLII (March): 72­115. .2004. East Asia Integrates: A Trade Policy Wonnacott, Paul. 1996. Beyond NAFTA--The Design Agenda for Shared Growth. East Asia and Pacific of a Free Trade Agreement of the Americas. In World Bank Region, World Bank, Washington, The economics of preferential trade agreements, DC. 79­107. College Park: University of Maryland, WTO (World Trade Organization). 1999. Trade and Center for International Economics; Washington, Environment. Geneva: WTO. DC: AEI Press. 56 3 Regional Trade Agreements: Effects on Trade Regional trade agreements (RTAs) can have expand both regional exports and exports to positive or negative effects on trade depending the rest of the world.1 on their design and implementation. Analysis RTAs can be a springboard to global mar- in this chapter confirms that gains from a pref- kets, but here too, low MFN trade barriers are erential trade agreement cannot be taken for necessary for success. RTAs can help countries granted; moreover, even in agreements with integrate with global markets, but no agree- positive impacts on average incomes, not all ment provides guarantees, so design and im- members are assured of increases. The inter- plementation matter. esting policy question then is not whether RTAs are categorically good or bad, but what determines their success? The broader policy context in which an The Impact of RTAs on RTA is designed and implemented is crucial. Merchandise Trade and Incomes Agreements that have been designed to com- RTAs cover much more than trade barriers plement a general program of economic re- RTAs have increasingly been designed to cover form have been most effective in raising trade. much more than formal trade policies (see When RTAs have tended to be fruitless, it is chapter 2), and RTAs are signed for a variety often because of the lack of a coherent pro- of reasons. The impact of these agreements on gram of reform. trade determines the extent to which broader For an RTA itself, the most important political and social objectives are achieved. It ingredient for success is low trade barriers is difficult to identify an agreement that has with all global partners. Most-favored-nation fostered wider political objectives without (MFN; i.e., nondiscriminatory) liberalization, achieving economic integration. It is clear that which creates more trade, is the fastest and the political context and broad economic en- most efficient way to increase intraregional vironment in which integration takes place are trade. In addition, agreements that minimize crucial for determining the trade impact. Suc- excluded products expand the scope for posi- cess derives from a strong willingness to liber- tive net benefits through competition and alize and to accept the subsequent economic trade creation. adjustments, accompanied by intense mutual Recent research has added nonrestrictive economic dialogue and communication and rules of origin to the list of successful factors; genuine efforts toward mutual understanding. local firms must be able to effectively source Severe macroeconomic disturbances and a tur- materials at the lowest cost. Such rules of ori- bulent investment climate can easily disrupt gin are an essential element of agreements that trade and derail an agreement. 57 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 The simplest measure of integration is the 50 percent in 1999. Over the same period, the trend in the share of imports from regional part- importance of trade between MERCOSUR ners in the total imports of a region. Successful members doubled from 10 to 20 percent. regional agreements might be expected to in- For many of the agreements signed in the crease trade between partners relative to those 1990s, intra-regional trade shares were countries' trade with the rest of the world. But growing strongly before the agreements were three important caveats need to be understood. signed (NAFTA, MERCOSUR, SAPTA, First, successful regional integration is typ- SADC). There may have been some anticipa- ically accompanied by reductions in tariffs for tion effect in the year or two before signing, but all partners. Hence, regional trade shares may this doesn't explain trend increases in shares not rise even though the volume of regional commencing five or more years previous, as in trade is increasing. Second, regional trade the case of MERCOSUR. In many cases this in- agreements that provide for the removal or crease in regional trade reflects the impact of reduction in trade costs other than those asso- unilateral, multilateral, as well as regional trade ciated with formal trade policies (such as im- liberalization and the fact that agreements proved customs procedures), may stimulate often follow growing trade relationships. trade from all sources. Third, many agree- In Africa, the picture is mixed. The extent ments cover nontrade issues such as invest- of regional integration among the Common ment, services, and labor, and these can have Market for Eastern and Southern Africa important consequences for growth and (COMESA) members has been relatively static incomes. These are analyzed in subsequent over the past two decades. In contrast the chapters, but it is important to bear in mind share of intra-area trade has increased sub- here that an agreement may be successful even stantially for Economic Community of West if the propensity for members to trade among African States (ECOWAS) since the early themselves does not increase markedly. 1980s and for SADC since the late 1980s. In Trade performance in several regional trade East Asia, a region that has experienced sub- agreements shows that the increase in intra- stantial economic progress over the past 20 regional trade shares of agreements signed in years, there has been little increase in intra- the 1990s has been substantial (figure 3.1). regional trade shares. The share of intra-NAFTA (North America Given these disparate results, it is necessary Free Trade Agreement) trade rose from less to go beyond simple trade shares to identify than 35 percent in the late 1980s to almost the economic impact of regional trade Figure 3.1 Evolution of the share of intra-regional imports in total imports, 1960­2000 Percent Percent 70 14 EC SADC 60 12 ECOWAS 50 NAFTA 10 40 8 COMESA 30 ASEAN 6 CACM 20 4 10 2 MERCOSUR SAPTA 0 0 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 Source: World Bank staff. 58 R E G I O N A L T R A D E A G R E E M E N T S: E F F E C T S O N T R A D E agreements. Because a decline in the share of With the exception of MERCOSUR, all extra-regional trade in total trade will be of regions that have experienced an increasing less significance if the total value of trade is share of intra-regional trade in total trade increasing, a logical (and commonly used) have also seen the ratio of extra-regional measure is the share of extra- and intra- trade in GDP increase. The Association of regional trade in regional GDP (figure 3.2). Southeast Asian Nations (ASEAN) is an Figure 3.2 The ratio of external and intra-regional trade to GDP ASEAN CACM Percent Percent 1992 1961 1991 50 35 Revised 45 30 Original 40 CACM extra 35 25 30 ASEAN extra 20 25 20 Year of entry 15 into force 15 10 CACM intra 10 ASEAN intra 5 5 0 0 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 COMESA ECOWAS Percent Percent 1994 1975 ECOWAS extra 16 COMESA extra 40 14 35 12 30 10 25 8 20 Year of entry 6 Year of 15 into force entry 4 COMESA intra into 10 ECOWAS intra 2 force 5 0 0 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 MERCOSUR NAFTA Percent Percent 1991 1994 16 7 NAFTA extra 14 6 12 Year of MERCOSUR extra entry 5 10 into force 4 8 3 6 Year of 2 4 entry MERCOSUR intra NAFTA intra into 2 1 force 0 0 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 Source: World Bank staff. 59 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Figure 3.3 Intra-regional trade grows faster when world trade growth is positive Annual growth of intra-regional exports (percent) Annual growth of intra-regional exports (percent) 35 40 MERCOSUR 30 35 25 30 SADC 20 25 15 ECOWAS 20 ASEAN 10 15 COMESA CACM 5 10 NAFTA 0 5 5 0 10 5 0 5 10 15 20 25 0 5 10 15 20 25 Annual growth of world exports (percent) Annual growth of world exports (percent) Source: World Bank staff. interesting example. The share of intra- approaches are available that we can crudely regional trade remained fairly flat during the separate into ex ante general equilibrium sim- 1990s. However, the ratios of intra-ASEAN ulation studies and ex post econometric trade to GDP and ASEAN imports from the analyses by using the gravity model (box 3.1). rest of the world to GDP have both increased The broad results3 from general equilib- strongly. ASEAN appears to have been very rium exercises are that, first, excluded successful. countries almost always lose. Second, for de- In general, this suggests that external open- veloping countries the bottom line determi- ness and the expansion of intra-regional trade nant of positive income effects is the increase go together. To take this analysis a little further, in market access. Third, in Free Trade Areas we plot the estimated relationship between an- (FTAs) each country can always lower its tar- nual changes in intra-regional trade and annual iff to ensure gains. This may be more difficult changes in the total volume of world trade, and in a customs union. Finally, regional trade we find a positive association in all cases (fig- agreements are typically expected to create ure 3.3). Although crude, this analysis suggests more trade than they divert, although this is that the successful expansion of trade among not always the case. the members of a regional trade agreement These points are highlighted in figure 3.4, tends to be associated with increasing extra- which summarizes model estimates of the im- regional imports as a share of GDP and with pact of Chile signing FTAs with different re- the growth of world trade.2 gional groupings. Excluded countries lose in every case. Chile loses from an FTA with Do regional trade agreements MERCOSUR. FTAs with larger markets stimulate trade? bring bigger gains for Chile but also tend to The analysis just discussed provides useful in- entail larger losses for excluded countries. formation, but it does not directly measure Large northern countries gain little from the impact of regional trade agreements. To FTAs with substantially smaller southern isolate the role of policy--that is, RTAs-- partners. from other factors influencing trade patterns A number of analysts have concluded that requires more sophisticated economic the numerous estimates from the gravity modeling. Different, yet complimentary, model generally support the contention that 60 R E G I O N A L T R A D E A G R E E M E N T S: E F F E C T S O N T R A D E Box 3.1 A primer on modeling of RTAs A. Simulation studies: Looking forward to potential the friction between them (due to trade and other gains costs, which is proxied by distance). It is also com- The ex ante studies are based on a specific general mon to use so-called dummy variables to capture ge- equilibrium model structure that allows a rich analy- ographical effects (such as whether the two countries sis of the impact of RTAs at both the aggregate and share a border, or if a country has access to the sea), sectoral levels. A key strength of this approach is its cultural and historical similarities (such as if two ability to highlight which sectors may expand and countries share a language or were linked by past which may contract in the face of given resource colonial ties), and regional integration (such as be- constraints. The richness of the model structure, longing to a free trade agreement or sharing a com- however, requires that many key parameters be mon currency). A disadvantage of using dummy vari- selected, (often on the basis of an extensive literature ables is that they may capture the impact of a range search), with others being derived by a process of of other effects that occurred during the same time calibration to a single base- year observation; that is, period as the RTA. For example, most applications the remaining parameters are derived such that the do not distinguish the extent of multilateral trade lib- model replicates the situation in the base year. To a eralization. Ideally, specific trade policy variables large extent the results of the impact of RTAs are would be included in the estimating equations, such determined by the choice of value for key relevant the level of multilateral and preferential tariffs. How- parameters (in this case the price elasticity of ever, the complexity of preferential trade arrange- demand for exports). Also, given that parameters are ments precludes such an approach. A notable excep- chosen and not estimated, the statistical properties of tion is the study done by Estevadeordal and the results are unknown. Robertson (2004), who included a measure of prefer- The characterization of RTAs is often simple, ential tariffs in their analysis of the impacts of RTAs with most studies focusing on the removal of tariffs on regional trade in Latin America. but ignoring issues such as the rules of origin, prod- Although widely used because of its empirical uct exclusions, and services. These simulation exer- success, the gravity model had lacked rigorous theo- cises answer the question, "What would be the im- retical underpinnings and was long criticized for pact of the preferential removal of tariffs against a being an ad hoc model. Recent theoretically limited set of trade partners, given the assumed grounded gravity equations are derived from models model structure?" But they do not tell us whether with strong constraints on preferences and technol- particular agreements have actually created or ogy, which undermines a straightforward interpreta- diverted trade. tion of some of the estimated coefficients. Anderson B. Econometric studies using the gravity model: and van Wincoop (2004) provide a good overview of Looking back at actual performance this debate. Another weakness of many applications The gravity model provides a useful framework of the gravity model is the proxying of trade costs by for assessing the impact of policy variables on the distance, and the implicit assumption that cargoes behavior of bilateral flows between countries. Its traveling 1,000 miles in Africa face exactly the same name is derived from its passing similarity to trade costs as similar cargoes traveling 1000 miles in, Newtonian physics, in that flows between two say, Europe. countries increase in proportion to their economic mass (as measured by GDP) and are constrained by Sources: Inter-American Development Bank 2002 and Bank staff. RTAs create trade.4 This merits further analy- enlargement of the European Union (to in- sis. Differing studies have produced sharply clude Greece, Portugal, and Spain), whereas different results for the same agreement. For Wei and Frankel (1995) find "massive trade example, Bayoumi and Eichengreen (1997) diversion." One way to digest this contradic- find no evidence of trade diversion from tory literature is to combine and assess these 61 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 For none of the agreements do we find un- Figure 3.4 Simulated welfare impact of ambiguous evidence of a net trade-creating various FTAs involving Chile effect extending to all members.5 Thus even if Change in welfare ($) an agreement as a whole creates trade, it is 2500 important that there are mechanisms to ensure 2000 Other that all members benefit. 1500 included Chile Regional trade agreements and exports to 1000 the world 500 So far the analysis has concentrated on 0 whether increases in intra-regional trade fol- 500 Sum for excluded lowing the signing of a RTA are associated 1000 with falling imports from the rest of the world A EU relative to a scenario in which the RTA was NAFT Nafta Nafta Mercosur Mercosur not signed. It is equally important to ask how Mercosur regional agreements can be used as part of a broad approach to openness and especially Source: Harrison et al (2004). whether they can provide a springboard to global markets for local exporters. results in a single statistical analysis, called Applying the gravity model with an addi- meta-analysis (box 3.2). This meta-analysis of tional variable to capture overall exports of a the literature on the impact of regional trade member of a particular set of RTAs, we can agreements on intra- and extra-regional trade assess whether these countries tend to export indicates that although agreements typically proportionately more than would "normally" have a positive impact on intra-regional trade, be the case for a similar country that was not their overall impact is uncertain. Actual party to the agreements.6 experience reinforces that there can be no These results, based on a sample period of presumption that a preferential trade agree- 1948 to 2000, show that different agreements ment will be trade creating. are associated with different propensities for higher-than-"normal" overall exports Do regional trade agreements benefit (figure 3.5). AFTA, EC, GCC, MERCOSUR, all members? NAFTA, and SACU all appear to export signif- The attention in most of the econometric stud- icantly more than they would have done in the ies is on the impact of particular RTAs. Few absence of the agreement. The countries that studies have sought to estimate the impact of comprise these regional groups appear to have RTAs on individual members. This is despite adopted policies that led them to be more the fact that studies of agreements that failed export-oriented than they otherwise would in the 1960s typically identify the lack of have been. We cannot say, however, that it was mechanisms for redistribution in the presence the RTA alone that led to these policies. The of asymmetric impacts as a crucial factor cre- variables that pick up changes in trade flows ating political tension and undermining com- may be capturing the effects of unilateral and mitment to the agreement (Greenaway and multilateral trade policies. Other agreements-- Milner 1990). We estimated gravity equations CEMAC, CIS, COMESA, EAC, ECOWAS, that identify impacts for individual members and WAEMU--show a propensity to export for each of 17 different regional trade agree- significantly less than "normal." The Andean ments to determine whether the statistical evi- Community and SADC appear to export less dence suggests that the agreement has created when the whole sample period is considered, trade, diverted trade, or had no significant net but not when the analysis is confined to the effect on trade for each country. more recent sub-period, from 1980 to 2000. In 62 R E G I O N A L T R A D E A G R E E M E N T S: E F F E C T S O N T R A D E Box 3.2 Regional trade agreements in gravity models: A meta-analysis M eta-analysis provides a means of assessing and 17 research studies. The table below reports the combining empirical results from different mean value of the overall and internal impacts, the studies. The approach takes as individual observa- standard deviation, the number of statistically signifi- tions the point estimates of relevant parameters from cant estimates, and the total number of estimates of different studies. This set of observations is then used each impact. to test the hypothesis that the relevant coefficient is Of the estimates of the overall impact, 76 percent statistically different from zero. Here we are con- are statistically significant, 42 percent are negative cerned with two parameters. The first measures the and significant, and 34 percent are positive and sig- impact of the agreement on total imports (which we nificant. For the internal impact, 66 percent of the label overall impact); a negative value for this para- estimates are statistically significant, 54 percent are meter suggests that for the agreement concerned, the positive and significant, and only 12 percent are neg- level of trade between a member and any other coun- ative and significant. The mean estimate of the over- try is less than the normal level of trade that one all impact is negative. The most robust estimates of could expect. Thus a negative value is evidence of the overall impact are negative. The mean value of trade diversion. The second parameter captures the the internal impact is positive. For both parameters impact of a regional trade agreement on the level of there is a high degree of variance about the mean trade between partners (internal impact). In our values. Within this analysis the estimates of 19 re- analysis we have included 254 estimates of overall gional agreements were assessed; 10 exhibited on impact and 362 estimates of internal impact from average net trade diversion. Summary of the estimates by regional trade agreement Overall impact Internal Impact Mean Standard Significant Total Mean Standard Significant Total Value error estimates estimates value error estimates estimates Total 0.31 1.12 194 254 0.79 1.30 238 362 Source: World Bank staff. that period, the Andean Community also ments that have been relatively open to im- appears to have been more export-oriented ports have shown higher propensities to export than they otherwise would have been, perhaps to the global market than would otherwise be reflecting substantial trade policy revisions. expected. Elsewhere, intraregional trade has Most of the agreements in which export been initiated, but imports have been diverted propensities are lower also appear to generate and exports suppressed. fewer imports than would "normal" countries not participating in the agreements (CEMAC, The potential gains from larger markets CIS, COMESA, EAC, WAEMU). At the same and higher growth time, those agreements that appear to be more Trade of RTA members will be affected export-oriented tend to be more open to im- through the changes in trade policies that take ports (AFTA, EC, GCC, MERCOSUR since place, but will also change if there is an im- 1980, NAFTA, SACU). In many cases, there provement in technology, higher investment, has been a strong impact on intra-regional and a higher rate of growth. By crudely using trade. In general, members of regional agree- dummy variables, gravity models provide a 63 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Figure 3.5 RTAs that divert imports tend to export less to global markets SACU NAFTA EC CACM Overall imports ECOWAS Overall exports Intra-regional trade GCC AFTA ANDEAN CEMAC Mercosur SADC SAPTA CIS EAC WAEMU COMESA 4 2 0 2 4 6 8 10 Estimated exponential impact on trade Note: The bars show the magnitude of the dummy variables capturing respectively the extent to which intraregional trade, overall imports, and overall exports differ from the "normal" levels predicted by the gravity model on the basis of economic size, proximity, and relevant institutional and historical variables, such as a common language. Source: World Bank staff. measure of RTAs, which catches all of these RTAs can also affect growth through tech- factors through their impact on trade but can- nological transfer. Trade raises total factor not distinguish the precise mechanisms. Com- productivity by providing access to a wider plementary approaches look at the impact of and more advanced range of technologies. The RTAs on these other factors. productivity of an importing country can in- Berthelon (2004), using cross-country regres- crease through the importation of intermedi- sions to estimate the effects of RTAs on growth ate goods, which as a result of R&D in the during 1960­99, found that RTAs that enlarged exporting country, are either new and/or of the market substantially had substantial posi- better quality relative to existing products. In tive effects on growth. The results suggest, for this way a country that is open to trade can example, that the FTA signed between Chile benefit from R&D activities undertaken over- and the EU might be expected to increase the seas. RTAs will have a positive effect if they growth rate in Chile by 0.6 percentage points stimulate imports from technological leaders. and in the EU by 0.005 percentage points. The On the other hand, if the trade agreement larger market permits wider competition, larger leads to trade diversion away from more scales, and greater specialization, all of which technologically advanced sources of inputs, increase productivity and growth. South-South then there could be a negative impact on agreements face an uphill struggle in two re- productivity growth. spects: they generally entail much smaller mar- Schiff and Wang (2003) found that, for Mex- kets, and they have less scope for realizing the ico, trade with NAFTA partners had a large and gains from comparative advantage that differ- positive impact on Mexico's total factor pro- ent factor intensities would otherwise bring. ductivity (TFP), while trade with the rest of the 64 R E G I O N A L T R A D E A G R E E M E N T S: E F F E C T S O N T R A D E Box 3.3 Implementation matters The European Union and agriculture for tariff cuts included a number of rubber products The founding treaty (the Treaty of Rome), and subse- of which it was a major exporter. Indonesia, which quent replacements, commit the European communi- lies on the equator, offered a 10 percent cut in the ties to "the harmonious development of world trade, duty on snow plows (Balasubramanyam 1989). the progressive abolition of restrictions on interna- More recently, the trade elements of the agreement tional trade, and the lowering of customs barriers." have been intensified with the launching of the AFTA The European Union (EU) has failed to meet these (ASEAN Free Trade Area), the aim being to create a objectives for agricultural products. There is little genuine free trade area. In 1995 the deadline for doubt that EU agricultural policy has been the source fully implementing AFTA was reduced from 15 to of considerable disharmony among trading partners. 10 years, although there has been some backsliding recently from agreed tariff-reduction schedules. Movement of Moldovan wine through Ukraine Moldova is a major producer of wine. Although it SADC and rules of origin has a free trade agreement with Russia, its main SADC initially agreed to simple, general, and consis- market, it costs more to ship a case of wine from tent rules of origin similar to those of neighboring Chisinau to Moscow than from Australia to Moscow and overlapping COMESA. The initial rules required (UNECE 2003). Why? Moldovan wine must pass either a change of tariff heading, a minimum of through Ukraine, usually by rail. Although the two 35 percent of value-added within the region, or a countries are party to the CIS free trade agreement, maximum import content of 60 percent of the value which provides for fair treatment in transit, the of total inputs. Subsequently, however, the rules were Ukrainian authorities, in addition to imposing delays revised to include more restrictive sector- and and requiring unofficial payments, recently intro- product-specific rules. The requirement concerning duced an additional requirement that bulk wines change of tariff heading has been supplanted by de- must be transported in specially heated railway tailed technical-process requirements, a much higher wagons, although a clear rationale for this is difficult domestic value added requirement, and lower to ascertain (World Bank 2004). permitted import contents. The rules became much more similar to those of the EU and of NAFTA, ASEAN and exclusions from preferences reflecting in part the influence of the recently negoti- ASEAN members initially were allowed to exclude ated EU-South Africa agreement and the rules of certain products from tariff reduction, a right that origin governing EU preferences to ACP countries they exercised liberally. In many cases the tariff (Flatters and Kirk 2003). This example illustrates reductions offered were of very limited value to other how sectoral interests and misperceptions of the role members. Thailand's offers, for example, included and impact of rules of origin can undermine RTAs. wood products that it did not import and that other members did not produce. Malaysia's list of products Source: World Bank staff. OECD did not. They suggest that this is because find that it has led to a permanent increase in Mexico not only benefited from the content of TFP in Mexican manufacturing of between trade with the NAFTA partners, the country 5.5 percent and 7.5 percent. also experienced closer contact and more infor- In a later study, Schiff and Wang (2004) mation exchanges, especially among subcon- look at the dynamic impact of North-South tracting firms, which are more integrated into trade on technology diffusion to Korea, the production networks of their Northern part- Mexico, and Poland from the EU, Japan, and ners than was the rest of the OECD. They simu- North America. Using industry level data, late the impact of NAFTA as a consequence and they found that technology diffusion and 65 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 productivity gains tend to be regional. A possible reason for this being that knowledge Figure 3.6 Not all regions are open diffusion is also governed by close contacts NAFTA and the hands-on relationships that are more AFTA likely with neighbors. Nevertheless, for all SADC countries the biggest impact of trade on EAC TFP can be guaranteed by removing trade MERCOSUR barriers on knowledge-intensive goods from COMESA all countries. ECOWAS SAPTA Ingredients of Success 0 5 10 15 20 25 Open regions do better Average tariff RTAs are only effective for developing coun- Note: Tariffs are import-weighted at the country level to tries if implemented in conjunction with more arrive at PTA averages. comprehensive domestic reforms. At the same Source: UN TRAINS, accessed through WITS. time, a successful RTA will contribute to the overall economic impact of that reform pro- gram. In Europe, the eight Central and East- partners in the agreement, then it will experi- ern European countries that recently joined ence an improvement in economic welfare. the EU experienced strong growth in trade Therefore, countries forming preferential and investment inflows during the 1990s; yet trade areas should simultaneously reduce the two countries in the region, Bulgaria and level of external protection facing nonmember Romania, having almost identical trade agree- trading partners. Risks of trade diversion are ments with the EU but much less extensive do- particularly high in the newly proposed South mestic reform programs, saw a much weaker Asian Free Trade Area (SAFTA); (figure 3.6). trade and investment response. Regional inte- Second, where there are asymmetries in the gration initiatives in Latin America in the level of external protection, it is important 1990s have been much more effective than that the high-duty country reduce tariffs to early efforts, reflecting broad and credible avoid an adverse terms-of-trade shock. This is structural reforms in many countries (Devlin particularly relevant for developing countries and French-Davis 1999). Given this context, seeking to sign agreements with the EU or the there are a number of key features of RTAs United States. In developed countries where that are likely to contribute to favorable trade tariffs on manufactured products are rather outcomes. low (and high-duty agricultural products are The external trade regime is a crucial de- typically excluded from regional preferences), terminant of the success of RTAs for several trade diversion and trade creation are less reasons. First, trade diversion tends to fall likely to be significant. Thus with no trade with the level of the external tariffs main- being created in the developed market, the de- tained by member countries after they form a cline in domestic sales by firms in the high- preferential trade agreement. The negative tariff developing country may not be offset by effects of trade diversion are offset or over- a rise in exports to the developed country. come if the preferential removal of trade bar- Overall, the demand for goods produced in riers against some countries is accompanied the high-tariff country may fall, and its terms by a degree of liberalization to all countries, of trade could worsen. whether undertaken unilaterally or through Third, low MFN tariffs (and nonrestrictive multilateral negotiations. If a country that en- rules of origin) ensure that producers within ters into a free trade agreement increases its the regional trade agreement will have access imports from all countries, not just its to competitively priced inputs. In today's 66 R E G I O N A L T R A D E A G R E E M E N T S: E F F E C T S O N T R A D E globalized market, policies that significantly and deepening of a meaningful RTA, whereas raise the input costs of producers will con- belonging to a regional scheme constitutes nei- strain their exports to both regional and ther a necessary nor a sufficient condition for global markets. Regional integration is more an open and liberal trade regime. likely to be successful if it is achieved on the In the 1960s and 1970s, preferential agree- basis of strong competition and ease of access ments among developing countries were typi- to low-cost inputs. cally accompanied by high external tariff bar- Trade liberalization is a crucial mechanism riers as part of an import substitution strategy. for increasing competition in domestic mar- In contrast, agreements among more devel- kets. Where it is not politically feasible to oped countries in the same period were more open up broadly to all external suppliers, a re- often associated with declining external barri- gional approach can provide a stepping stone ers. For example, the simple average external toward the benefits of comprehensive liberal- tariff of the original six members of the Euro- ization. However, it is important to take the pean Union fell from 13 percent in 1958 to second step: Even in a large region such as the 6.6 percent after the Kennedy Round of Gen- EU, competition from within the region has eral Agreement on Tariffs and Trade (GATT) been found to be much weaker than that pro- negotiations. Agricultural products were ex- vided by external imports. Jacquemin and cluded from these reductions, reflecting their Sapir (1991), for example, found that profit exclusion from GATT negotiations until the margins in European countries were signifi- Uruguay Round. The failure to reduce agricul- cantly dampened by external imports but not tural tariffs in Europe led to substantial trade by intra-regional imports. And collusive diversion in agriculture with significant wel- agreements are more difficult to enforce for fare losses for European consumers, especially companies based in distant locations. Firms the poorest, and a considerable hardship for that face little competition in local and re- poor farmers in developing countries. gional markets will have low incentives to Many developing countries have since re- achieve the efficiency necessary to compete in duced external tariff barriers both unilaterally world markets. and through multilateral negotiations. As a Clearly RTAs may affect the setting of ex- result, recent preferential agreements among ternal tariffs. This is true by definition in the many developing countries have been intro- case of a customs union and indirectly true in duced or revamped with lower external barri- the case of a free trade area. Recent research ers. This is particularly true in Asia and Latin finds that World Trade Organization (WTO) America, where preferential and MFN tariffs members do not appear to have more liberal declined in tandem after 1985, so that margins external trade policies than non-WTO mem- of preference remained stable or were slightly bers (Rose 2004), and that membership in a compressed (figure 3.7). RTA has, on average, no clear effect on a coun- try's trade policy (Nitsch and Sturm 2003). Paper agreements are not enough Foroutan (1998), on the other hand, concludes Important aspects in the assessment of RTAs that countries in effective regional groupings, are whether their members have implemented distinguished by the growth of intra-area their objectives under the agreement and the trade, have undertaken more far-reaching extent to which the objectives in the agree- trade liberalization. However, there are cases ment have been met. Often the objectives in an of liberalizing countries that did not belong to agreement are defined by foreign ministers or an RTA and of countries in an effective RTA even prime ministers, while the way that those that did not liberalize trade policy. The conclu- objectives are to be carried out is determined sion is that the acceptance of a liberal trade later in negotiations between ministries. If tar- policy may be a requirement for the survival iff concessions are subsequently negotiated 67 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 North-South agreements appear to have a Figure 3.7 Preferential tariffs in tandem better track record than South-South agree- with all tariffs in Latin America ments. The comprehensive tariff objectives of MFN and preferential tariff liberalization most North-North agreements signed before Latin America, 1985­1997 the mid-1980s were implemented on or ahead Average tariff rates (percent) of schedule (table 3.1). In contrast, South- 50 South agreements reached during this period--most based on limited positive lists of 40 products for tariff liberalization--had a very MFN tariffs weak record of implementation. The delays in 30 implementing initial regional tariff commit- ments "generally reflected a basic incompati- 20 bility between the inward-oriented develop- Preferential tariffs ment strategies of most members and regional 10 liberalization"(De la Torre and Kelly 1992). A larger number of South-South agree- 0 ments signed or substantially revised in the 1985 1987 1989 1991 1993 1995 1997 late 1980s and early 1990s have sought a Note: MFN tariffs represents the simple average of the much broader degree of internal tariff liberal- most-favored-nation tariffs applied by the following 11 Latin American countries (based on country averages): Argentina, ization, have been more effective in imple- Bolivia, Chile, Colombia, Ecuador, Mexico, Paraguay, Peru, menting agreed-on tariff reductions, and have Uruguay, and Venezuela. Preferential tariffs represent the average preferential tariff that each country applies to the tended to reduce external tariffs. For exam- other countries in this sample under different regional trade ple, the GCC, launched in 1982, was origi- agreements. Calculations include only ad valorem tariffs. nally intended to become a free trade area--a Source: Estevadeordal and Robertson 2004. goal achieved by 1983. By the late 1980s, however, the objective evolved into formation sector by sector or item by item, the process of a customs union, which was established in becomes cumbersome and open to capture by 2003 (see World Bank 2003). However, domestic interests. The distinction is often table 3.1 also reports that substantial prob- made between agreements that reduce duties lems with implementation remain in many of only on products specified in a positive list the regional agreements involving developing and other agreements, typically more liberal, countries. implemented on the basis of a negative list of products excluded from tariff reduction. Nonrestrictive rules of origin are integral Sectoral accords within RTAs can curb to success market forces and limit the benefits from Preferential rules of origin are integral to pref- competition. For example, Ozden and Parodi erential trade agreements. However, it has be- (2003) found that the auto agreement embed- come increasingly clear that rules of origin ded in MERCOSUR between Argentina and can be designed in a way that restricts trade Brazil compelled companies in both countries beyond what is necessary to prevent trade to balance trade, ensuring that production deflection or the transshipment of products would not be reallocated to the lowest cost from third countries through a member for producer (Brazil); this move secured the the purpose of obtaining preferential duties. support of the companies for the agreement. In addition, the proliferation of free trade Because a new entrant would have to build agreements with accompanying rules of origin plants in both countries (not just one), the is increasing the burdens on customs services agreement acted as a barrier to competition in many countries, and these burdens have that favored insiders. consequent implications for trade. 68 R E G I O N A L T R A D E A G R E E M E N T S: E F F E C T S O N T R A D E Table 3.1 Implementation of tariff commitments by type of agreement, 1960­1999 Agreement Objective on intra-bloc tariffs Implementation record North-North agreements reached from 1960­89 ANZERTA (signed 1983) Eliminate all tariffs by 1988 On schedule European Economic Eliminate all tariffs by 1968 Ahead of schedule Community (signed 1957) U.S.-Canada FTA (1988) Eliminate all tariffs by 1999 Ahead of schedule EFTA (1960) Eliminate all tariffs on manufactures by mid 1967 On schedule South-South agreements, 1960­89 Andean Pact (1969) Eliminate tariffs on positive list Postponed several times Central American Common Elimination of tariffs Initially on schedule, most duties Market (1960) removed in the early 1970s, but restrictions reintroduced in the 1980s EAC (1967­1977) Establishment of a common market The Community was dissolved Latin American Integration Liberalization of common lists of products by 1972 Common lists not liberalized on Association schedule ECOWAS Tariff liberalization by 1990 Progress negligible ASEAN (1967) FTA based on positive lists Repeatedly postponed GCC (1982) FTA Virtual elimination of all tariffs in 1983 South-South agreements, 1990­99 AFTA (1992) Gradual reduction of tariffs over 12­15 years Liberalization took place ahead of according to member-specific schedules original schedule CACM--Revised (1991) Customs union Implementation postponed; progress uneven among members GCC Customs union begins in 2003; completed by 2005 Customs union established on schedule COMESA Progressive tariff elimination to be completed Implementation varies by country, 9 by 2000 out of 20 members have moved to duty-free trade MERCOSUR Elimination of all tariffs by 1995 All lines, with the exception of sugar and automobiles, have been liberalized SAPTA Limited tariff concessions from a country-specific No formal schedules have been adopted positive list SADC Tariff liberalization by 2008, with sensitive lists Implementation delayed in some sectors eliminated by 2012 due to lack of agreement on rules of origin CEMAC (1999) Economic Union Tariffs liberalized according to schedule in nearly all lines WAEMU (2000) Economic and monetary union Tariff liberalization mostly on schedule North-South agreements, 1990­99 Europe Agreements Country-specific tariff removal schedules in Bulgaria mostly on schedule, Romania (Bulgaria, Romania) preparation for the EU membership continues to have some unresolved issues NAFTA Tariff elimination in stages to be complete by 2008 On schedule EU-Mexico Progressive tariff elimination by 2010 On schedule EU-South Africa FTA establishment by 2012 Partial implementation pending official ratification U.S.-Chile Progressive tariff elimination by 2015 N/A Source: World Bank staff. In general, the rules of origin in North- sometimes the rule may be a change of tariff South agreements are more restrictive than heading, sometimes a change of tariff chapter; those adopted by South-South agreements for other products there will be a value-added (Figure 3.8). A feature of both EU and NAFTA requirement; and in others the rules of origin agreements is the high degree of variation in may specify a particular technical process. rules of origin across product categories. Dif- The amount of the required value added ferent rules are specified for different products: can vary across products. The change of tariff 69 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Anson and others (2004) and Carrere and Figure 3.8 Rules of origin in North­South de Melo (2004) estimate that the administra- agreements are more restrictive than in tive costs of providing the documentary evi- South­South agreements dence to support the certificate of origin under Index of restrictiveness NAFTA are in the region of 1.8 percent of the 6 value of exports. The distorted impact of the 5 rules, resulting from the need to use local and higher cost inputs to qualify, may be equiva- 4 lent to an average duty of around 4.3 percent. 3 Thus, restrictive rules of origin can very easily 2 wipe out any margin of preference generated 1 by a trade agreement. Other things being equal, compliance costs are lowest for rules in- 0 A A volving a change of tariff heading, followed by xico CM AS AFT W NAFT SADC value-added rules. Rules requiring a specific EU-Chile EU-Me COMESAECO Chile-CA technical process have the highest compliance costs. Note: Higher values of the index equal more restrictive rules of origin [derived from Estevadeordal and Suominen Estevadeordal and Suominen (2004) (2004)]. introduce a synthetic measure of the restrictiveness of rules of origin (the basis for figure 3.8) into a standard gravity model of bi- classification can be used to provide a positive lateral trade flows. Their econometric analysis test of origin by stating the tariff classification leads them to conclude that restrictive prod- of imported inputs that can be used in the uct-specific rules of origin undermine overall production of the exported good. Or it may trade between preferential partners and that be defined to provide a (more restrictive) neg- provisions such as cumulation7 and de minimis ative test by stating cases where a change of rules,8 which increase the flexibility of apply- tariff classification will not confer origin. For ing a given set of processing requirements, example, in the EU rules of origin, bread, bis- boost intraregional trade. Applied at the sec- cuits, and pastry products (heading 1905 of toral level, this approach yields support for the the Harmonized System) can be made from hypothesis that the restrictiveness of rules of imported products of any other tariff heading origin for final goods stimulates trade in inter- except those of chapter 11, which includes mediate products between preferential part- flour, the basic input to these products. ners and diverts trade away from nonmem- Specifying rules of origin on a product by bers. Cadot and others (2002) find that for product basis offers opportunities for sectoral sectors where tariff cuts are larger than aver- interests to influence the specification of the age, the rules of origin are more restrictive and rules in a protectionist way. The outcome of the rate of use of preferences by Mexican ex- highly detailed product-specific rules of origin porters lower than average. They conclude is typically a complex set of rules, which can that rules of origin are the "prime culprit" for be highly restrictive. Box 3.5 provides an ex- the very modest impact of NAFTA on Mexican ample of the sort of complexity that can arise. exports identified by other researchers. Many agreements involving developing coun- tries, on the other hand, tend to specify gen- Deeper agreements can lead to larger trade eral rules that apply to all products. The and income effects AFTA, COMESA, and ECOWAS, for exam- In principle, agreements that address a wider ple, have a single value-added rule applicable range of barriers can have a greater impact on to all products. trade flows and incomes. 70 R E G I O N A L T R A D E A G R E E M E N T S: E F F E C T S O N T R A D E Box 3.4 Restrictive rules of origin under NAFTA--the case of clothing H ere is an example of what rules of origin look through 540784, 540792 through 540794, like; the following pertains to men's or boys' 540822 through 540824 (excluding tariff item overcoats made of wool (HS620111). 540822aa, 540823aa or 540824aa), 540832 through 540834, 551219, 551229, 551299, A change to subheading 620111 from any other 551321 through 551349, 551421 through chapter, except from heading 5106 through 5113, 551599, 551612 through 551614, 551622 5204 through 5212, 5307 through 5308 or through 551624, 551632 through 551634, 5310 through 5311, Chapter 54 or heading 5508 551642 through 551644, 551692 through through 5516, 5801 through 5802 or 6001 551694, 600110, 600192, 600531 through through 6006, provided that: The good is both 600544 or 600610 through 600644. cut and sewn or otherwise assembled in the territory of one or more of the Parties. This stipulates that the visible lining used must be produced from yarn and finished in either party's The basic rule of origin stipulates change of chap- location. This rule may well have been introduced to ter but then provides a list of headings and chapters constrain the impact of the tolerance rule, which from which inputs cannot be used. Thus in effect, would normally allow 7 percent of the weight of the the overcoat must be manufactured from the stage of article to be of nonoriginating materials. In overcoats wool fibers forward, because neither imported and suits, the lining is probably less than 7 percent woolen yarn (HS5106-5110) nor imported woolen of the total weight. Finally, it is interesting to note fabric (HS5111-5113) can be used. However, the that the rules of origin also provide very specific rule also states that imported cotton thread exemptions for materials that are in short supply or (HS5204) or imported thread of man-made fibers are not produced in the United States. In this regard, (HS54) cannot be used to sew the coat together. This the rule reflects firm-specific lobbying to overcome rule in itself is very restrictive; however, the rule is the restrictions of these rules of origin when the orig- further complicated by requirements relating to the inal NAFTA rules were defined. The most extreme visible lining: example is the following, where the apparel will be Except for fabrics classified in 54082210, deemed eligible for tariff preferences if assembled 54082311, 54082321, and 54082410, the fabrics from imported inputs of: identified in the following subheadings and head- Fabrics of subheading 511111 or 511119, if ings, when used as visible lining material in cer- hand-woven, with a loom width of less than tain men's and women's suits, suit-type jackets, 76 cm, woven in the United Kingdom in accor- skirts, overcoats, car coats, anoraks, windbreak- dance with the rules and regulations of the Harris ers, and similar articles, must be formed from Tweed Association, Ltd., and so certified by the yarn and finished in the territory of a party: 5111 Association. through 5112, 520831 through 520859, 520931 through 520959, 521031 through 521059, Clearly, the job of the relevant official to check 521131 through 521159, 521213 through consistency and compliance with such rules is not a 521215, 521223 through 521225, 540742 simple one. through 540744, 540752 through 540754, 540761, 540772 through 540774, 540782 Source: World Bank staff. Subsequent chapters elaborate on the po- whether deeper agreements have a signifi- tential economic impacts of dealing with cantly greater impact on aggregate merchan- many of the regional agreement issues intro- dise trade than more narrow trade agree- duced in chapter 2. Here we simply ask ments. Two studies9 assume a productivity 71 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 response to trade liberalization when other part of its monitoring of the implementation measures are included and ascribe the results of the single market, has introduced the to the deep integration measures. The calcula- "Single Market Scoreboard" (box 3.5). In tions of these ex ante simulation studies illus- addition to providing vital information, the trate the potential that deeper agreements scorecard is useful as a disciplinary measure-- may hold when they produce a productivity to shame governments with a record of poor response, with changes in trade flows and in- implementation into action and to empower comes being a multiple of that under pref- governments with good records of implemen- erential tariff removal. However, because tation to challenge those members who are this result is inevitable from the way that not meeting their commitments. deeper integration is modeled (i.e., inducing More extensive monitoring could make an economy-wide increases in productivity), important contribution to the implementation these results from one or another deep inte- of many trade agreements. Lack of effective gration measure should be seen as indicative implementation has been a major factor lim- of potential rather than evidence of success. iting the impact of many trade agreements in Ex post exercises based on the gravity Africa, South America, South Asia, North model will tend to capture all of the policy re- Africa, and the CIS. lated impacts of a regional trade agreement on trade, not just the removal of trade policy variables. Several authors have tried to Conclusions: Preferential Trade capture in an index the differences of depth between agreements; it is then used as the Agreements and Economic dummy variable in the gravity model to cap- Development T ture the impact of RTAs [Li (2000), Adams his review of the experience of preferen- and others (2003)]. However, this approach tial trading agreements over the past 40 presents the issue of how to weight different years offers the following conclusions: policy measures--for instance, should ser- vices liberalization get more weight than cus- · There is no strong evidence to support toms cooperation? Thus the value of the the claim that a preferential trade agree- index would be dependent on the subjective ment will be net trade creating or that all weights that are assigned. The weights chosen members will benefit. Positive outcomes by Adams and others lead to EFTA being depend on design and implementation. ranked as much more restrictive than the An- · When embedded in a consistent and dean Pact or NAFTA. Further, many agree- credible reform strategy, the key deter- ments appear extensive on paper but have minant of regional trade agreements' accomplished little in practice. success is low levels of external trade barriers. While many developing coun- Extensive monitoring of agreements tries have reduced tariffs, they remain is crucial to ensure effective high in many countries and regions, and implementation the risk of trade diversion remains sig- In order to assess the impact of RTAs, infor- nificant. Further reductions in applied mation is needed on the extent to which the MFN tariffs will be required to ensure agreement's provisions are being implemented that regional agreements are beneficial and how they are affecting decisions by pro- for those participating in them and to ducers and consumers. Given the need for minimize the impact on the countries monitoring, an implementation scorecard that are left out. would be useful--such an approach has been · Trade agreements that provide for com- adopted by the EU Commission which, as prehensive liberalization of trade across 72 R E G I O N A L T R A D E A G R E E M E N T S: E F F E C T S O N T R A D E all major sectors and nonrestrictive rules product specific rules of origin limit the of origin are more likely to be successful. scope for gains. Agreements that devote considerable re- · Effective implementation is crucial to pos- sources to negotiating limited positive itive outcomes, yet implementation is lists or large negative lists and detailed compromised by proliferation. If different Box 3.5 Monitoring implementation of preferential trade agreements: "Single Market Scoreboard" in the European Union T he Single Market Scoreboard measures (1) the substantial improvement in implementation since extent to which Single Market directives have 1997; it also shows that the original six members of been transposed into national law by each member the EU and Greece are currently the worst offenders. state, (2) the average time it takes each member to Effective monitoring of implementation also requires transpose directives, and (3) the extent to which that clear targets be established. In 2001, the EU members are cooperating with enforcement and Heads of State established an interim target of a problem solving. This analysis by the European 1.5 percent implementation deficit. As of July 2004, Commission is supported by regular surveys of busi- only five members had achieved this target. nesses and individuals on perceptions of the Single A further measure of implementation is the extent Market and where it is not working. The Commis- to which agreed-on rules are being properly applied. sion also monitors differences in prices of identical In Europe, the Commission is charged with monitor- goods for indications that integration is leading to ing when Single Market rules are not being applied convergence. correctly; the Commission also takes infringement The left figure below shows the implementation cases against member countries that are breaking EU deficit for each member; that is, the proportion of laws. In terms of the number of infringement cases directives that have not been notified as having been open in May 2004, Italy and France are the worst transposed into national law. The figure shows a offenders (lower right figure). Single Market Scoreboard a. Implementation of single market directives by EU members b. EU law breakers Percentage rate of nonimplementation Open infringement case (May 2004) 12 160 140 10 1997 120 8 100 6 80 2004 60 4 40 2 20 0 0 y k y k FranceGreece stria UK stria man Italylands tugal UKSpain Italy eden Belgiumor rtugal SwedenAu FinlandIreland Fr ance Spain manGreeceBelgium IrelandlandsAu Sw Finland Ger embourg P Denmar Ger Po embourg Denmar Lux Nether Nether Lux Source: http://europa.eu.int/comm/internalmarket/score/index en.htm#score. 73 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 agreements have different product cover- 7. The basic rules of origin define the processing that age, different liberalization schedules, has to be done in the individual beneficiary or partner to and different rules of origin, the ability of confer origin. Cumulation is an instrument allowing pro- ducers to import materials from a specific country or re- agencies such as customs to apply the gional group of countries without undermining the origin agreements is severely undermined. The of the final product. In effect the imported materials from capacity to effectively implement is a the identified countries are treated as being of domestic crucial issue that countries should con- origin of the country requesting preferential access. There sider before signing an RTA. are three types of cumulation, bilateral, diagonal (or par- · Monitoring can play an important role tial), and full. The most basic form of cumulation is bi- in providing for effective implementa- lateral cumulation, which applies to materials provided by either of two partners of a preferential trade agree- tion, but often there is insufficient moni- ment. In this case originating inputs, that is materials, toring as well. Technical reviews are fre- which have been produced in accordance with the rele- quently not done, and when reports are vant rules of origin, imported from the partner, qualify as made, senior officials fail to act on their originating materials when used in a country's exports to recommendations. that Partner. Second, there can be diagonal cumulation on a regional basis so that qualifying materials from any- where in the specified region can be used without under- Notes mining preferential access. Finally, there can be full cu- 1. Flatters and Kirk (2003). mulation whereby any processing activities carried out in 2. The conclusion is unchanged if intra-EU and any participating country in a regional group can be intra-NAFTA trade are excluded from the total of counted as qualifying content regardless of whether the world exports. processing is sufficient to confer originating status to the 3. Drawn from Burfisher and others (2004) and materials themselves. Under full cumulation all of the Harrison and others (2004). processing carried out in participating countries is as- 4. For example, Ghosh and Yamarik (2004) sug- sessed in deciding whether there has been substantial gest that "a consensus has emerged among researchers transformation. Hence, full cumulation provides for that RTAs are trade creating." deeper integration among participating countries. 5. In this exercise, where the counterfactual is based 8. De Minimis or tolerance rules allow a certain on the historical pattern of trade flows, we assess how percentage of nonoriginating materials to be used the regional trade agreement affected trade flows after without affecting the origin of the final product. Thus, its introduction. As a measure of robustness of the ef- the tolerance rule can act to make it easier for products fects, we used three different estimation methods. with nonoriginating inputs to qualify for preferences Effects are considered statistically robust only if all under the change of tariff heading and specific manu- three methods generate a significant impact of the same facturing process rules. This provision does not affect sign. The three methods are pooled OLS with robust value added rules. standard errors. The second estimation method includes 9. For example, Hoekman and Konan (1999) find country-pair fixed effects using a specific OLS method. that a free trade agreement between the European The third approach is a pooled Tobit estimation. Union and Egypt limited to goods (but with substantial 6. Here we follow Soloaga and Winters (2001), progress on removing regulatory barriers) could raise who include an additional dummy variable to assess welfare by around 4 percent while an agreement that the impact on the exports of members of regional trade reduced barriers to services in Egypt could raise agreements, although their focus is on the welfare economic welfare by over 13 percent. Similarly, effects of RTAs. However, here we apply a panel Brenton, Tourdyeva, and Whalley (2002) find that an approach to a sample period of 1948 to 2000, cover- EU-Russia FTA limited to tariff removal would in- ing bilateral trade between 178 countries with country- crease welfare by around one-tenth of one percent pair fixed-effects, which a number of authors, although while a comprehensive agreement removing technical not all, suggest is the preferred method. We apply the barriers to trade and barriers to trade in services would above equation. The regional dummies are time sensi- raise welfare by more than 13 percent. tive; that is, they are relevant only after the agreement has been signed. Using a different estimation tech- nique, such as Tobit and OLS, and a different sample References period can lead to different results for a particular Adams, Richard, Philippa Dee, Jyothi Gali, and Greg agreement but the overall conclusion remains firm. McGuire. 2003. The Trade and Investment Ef- 74 R E G I O N A L T R A D E A G R E E M E N T S: E F F E C T S O N T R A D E fects of Preferential Training Arrangements--Old America in the 1990s. The World Economy 22: and New Evidence. Productivity Commission 261­90. Staff Working Paper, Canberra, Australia. Estevadeordal, A., and R. Robertson. 2004. Do Pref- Anson, Jose, Olivier Cadot, Antoni Estevadeordal, erential Trade Agreements Matter for Trade? In Jaime de Melo, Akiko Suwa-Eisenmann, Integrating the Americas: FTAA and Beyond, and Bolorma Tumurchudur. 2004. Rules of Ori- eds. A. Estevadeordal, D. Rodrik, A. M. Taylor, gin in North-South Preferential Trading Arrange- and A. Velasco. Cambridge, MA: Harvard Uni- ments with an Application to NAFTA. Research versity Press. Unit Working Paper 0406, Laboratoire d'E- Estevadeordal, A., and K. Suominen. 2004. Rules of conomie Appliquee, INRA, Paris. Origin: A World Map and Trade Effects. In The Balasubramanyam,V. N. 1989. ASEAN and Regional Origin of Goods: Rules of Origin in Preferential Trading Co-operation in Southeast Asia. In Trade Agreements, eds. A. Estevadeordal, Economic Aspects of Regional Trading Arrange- O. Cadot, A. Suwa-Eisenmann, and T. Verdier., ments, eds. D. Greenaway, T. Hyclak, and R. J. DC: Inter-American Development Bank. Thornton. London: Harvester Wheatsheaf. Evenett, S. J., and W. Keller. 2002. On Theories Ex- Berthelon, Matias. 2004. Growth Effects of Regional plaining the Success of the Gravity Equation. Integration Agreements. Processed. Journal of Political Economy 110 (2). Brenton, Paul, Natalia Tourdyeva, and John Whalley. Foroutan, Faezeh. 1998. Does Membership in a 2002. Economic Impact of a FTA between Russia Regional Preferential Arrangement Make a and the EU: Numerical Simulations Using a Country More or Less Protectionist? The World General Equilibrium Trade Model. Report pre- Economy 21(2):305­35. pared for the European Commission, Brussels. Flatters, F., and R. Kirk. 2003. Rules of Origin as Brenton, Paul, and H. Imagawa. 2004. Rules of Ori- Tools of Development? Some Lessons from gin, Trade and Customs. In The Customs Mod- SADC. Paper presented at INRA conference on ernisation Handbook, ed. L. De Wulf, and Rules of Origin, Paris, May 2003. J. Sokol. Washington, DC: World Bank. Ghosh, S., and S. Yamarik. 2004. Are Regional Trad- Burfisher, Mary, Sherman Robinson, and Karen ing Arrangements Trade Creating? An Applica- Thierfelder. 2004. Regionalism: Old and New, tion of Extreme Bounds Analysis. Journal of Theory and Practice. Forthcoming in Agricul- International Economics 63: 369­95. tural Policy Reform and the WTO: Where Are Glick, Reuben, and Andrew Rose. 2002. "Does a We Heading? eds. G. Anania, M. E. Bohman, Currency Union Affect Trade? Time Series Evi- C. A. Carter, and A. F. McCalla. Cheltenham, dence." European Economic Review 46: UK: Edward Elgar. 1125­51. Cadot, O., J. de Melo, A. Estevadeordal, A. Suwa- Greenaway, David, and Chris Milner. 1990. Eisenmann, and B. Tumurchudur. 2002. Assess- South­South Trade Theory, Evidence, and Pol- ing the Effect of NAFTA's Rules of Origin. icy. The World Bank Research Observer 5(1): Processed. 47­68. Carrere, Celine. 2004. "Revisiting Regional Trading Harrison, Glenn W., and David G. Tarr. 2003. Rules with Proper Specification of the Gravity Model." of Thumb for Evaluating Preferential Trading European Economic Review (forthcoming). Arrangements: Evidence from Computable Carrere, Celine, and Jaime De Melo. 2004. "Are Dif- General Equilibrium Assessments. The Policy ferent Rules of Origin Equally Costly? Estimates Research Working Paper 3142, World Bank, from NAFTA." CEPR Discussion Paper Washington, DC. No. 4437. London. Hoekman, Bernard, and Denise Eby Konan. 1999. De la Torre, A., and M. Kelly. 1992. Regional Trade Deep Integration, Nondiscrimination, and Euro- Agreements. Occasional Paper 93, Washington, Mediterranean Free Trade. Policy Research DC: IMF. Working Paper 2130, World Bank, Washington, Devlin, R., and A. Estevadeordal. Forthcoming. Trade DC. and Cooperation: A Regional Public Goods Ap- IADB (Inter-American Development Bank). 2002. Be- proach. In Regional Public Goods: From Theory yond Borders: The New Regionalism in Latin to Practice, eds. A. Estevadeordal, Brian Frantz, America. Washington, DC: IADB. and Tam R. Nnguyen. Washington, DC: Inter- Li, Quan. Institutional Rules of Regional Trade Blocs American Development Bank. and Their Impact on Trade. Processed. Devlin, R. and R. French-Davis. 1999. Towards and Ozden, Caglar, and Francisco Parodi. 2003. Customs Evaluation of Regional Integration in Latin Union and Foreign Investment: Theory and 75 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Evidence from Mercosur's Auto Industry. Work- Schiff, Maurice, and Yanling Wang. 2003. Regional In- ing Paper, Emory University. tegration and Technology Diffusion The Case of Palmeter D. 1997. Rules of Origin in Regional Trade the North America Free Trade Agreement. Policy Agreements. In Regionalism and Multilateralism Research Working Paper 3132, World Bank, after the Uruguay Round: Convergence, Diver- Washington, DC. gence, and Interaction, eds. P. Demaret, J.F. Bellis, Volker Nitsch, and Daniel Sturm. The Trade Liberal- and G. Garcia Jimenez. Brussels: European In- ization Effects of Regional Trade Agreements. teruniversity Press. UNECE (United Nations Economic Commission for Pelkmans, J., and Brenton, P. 1999. Bilateral Trade Europe). 2003. Economic Survey of Europe. Agreements with the EU: Driving Forces and Processed. Effects. In Multilateralism and Regionalism in Wonnacott, R. 1996. Trade and Investment in a Hub- the Post-Uruguay Round Era: What Role for the and-Spoke System Versus a Free Trade Area. EU? eds. O. Memdovic, A. Kuyvenhoven, and W. World Economy 19: 237­52. Molle. Boston: Kluwer. World Bank. 2003. Trade, Investment, and Development Soloaga, Isidro, and Alan Winters. 2001. Regionalism in the Middle East and North Africa: Engaging in the Nineties: What Effect on Trade? North with the World. Washington, DC: World Bank. American Journal of Economics and Finance World Bank. 2004. DTIS Moldova, Case Study of the 12 (1). Wine Sector. 76 4 Beyond Trade Policy Barriers: Lowering Trade Costs Together The removal of tariffs and quotas is a key fea- This chapter focuses on three key issues re- ture of regional trade agreements (RTAs), but lated to trade facilitation: customs clearance, modern RTAs can, and are, being designed to transport, and standards and their conformity achieve much more than that. Trade policies assessment. Coordinated action among devel- are only one element--and often a relatively oping countries is likely to be greatest with the minor one--of the overall costs of trade. Be- first two issues, customs clearance and trans- cause logistical, institutional, and regulatory port; examples of best practices in these areas barriers are often more costly than tariffs and are available and can be followed in regional generate no offsetting revenue, cooperative trade agreements. Progress in reducing barri- governmental efforts to improve customs pro- ers is likely to facilitate trade to and from all cedures, minimize the trade distorting impact trading partners with little or no scope to be of standards, and reduce transport costs may discriminatory. And while cheaper, faster, and have a higher payoff than reciprocal reduc- more predictable customs clearance and im- tions in overt trade policy barriers. proved transport services have a direct impact When RTA membership is part of a broad on trade, they are also crucial elements of the program of economic liberalization in which investment climate. the objective is to attract international invest- Initiatives to deal with standards and con- ment as much as to promote trade, a broad set formity assessment on a regional basis are of regulatory issues becomes paramount. scarce; the most successful agreements Which are the most appropriate institutions to have been between rich countries that are address these regulatory barriers? In certain undertaking a deep integration process, as in cases, institutions at the regional level will the case of the European Union (EU). provide for the most effective solutions, rela- Nonetheless, systems of standards, quality tive to both the multilateral and national assurance, accreditation, and measurement levels.1 RTAs can effectively promote dialogue are crucial to competitiveness and sustained and implement coordinated responses. growth. Regional interventions can be use- However, most RTAs have contributed little ful if developed in a transparent way and to reducing the associated trade costs, espe- with the participation of private groups (to cially RTAs among developing countries. ensure that procedures are not manipulated Many regional policy initiatives have to serve a protectionist end). Initiatives foundered because of the lack of effective im- targeted at a small number of key sectors plementation, and crossing borders between and toward improving the quality of confor- most developing countries is still a major mity assessment are likely to be the most impediment to trade. useful. 77 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Agreements that involve large markets and tend to diverge with distance, and this differ- have differing levels of institutional capacity ence is much higher when the two locations generally appear to have had the greatest suc- being compared lie on either side of a national cess in dealing with these trade facilitation border. If trade costs are low, then arbitrage issues. This is because the more advanced should constrain such price variation (Engel partner tends to drive institutional improve- and Rogers 1996). ments among the less advanced partners. The tax equivalent of trade costs can range (However, a real danger is that, in seeking from 30 to 105 percent, depending on the sec- greater access to industrial country markets tor, according to estimates for imports by the through bilateral trade agreements, develop- United States (Anderson and van Wincoop ing countries agree to apply a set of rules and 2004; Evans 2001). High trade costs discour- regulations defined by the advanced country age investment and constrain the ability of local that are inappropriate for their level of devel- firms to integrate into global production chains opment.) For many developing countries, (Faini 2004). Given the magnitude of these agreements with industrial countries alone barriers, ex ante simulation studies suggest that will not be sufficient, because the main source the benefits of lowering transaction costs, re- of higher trade costs are the borders and the ducing insecurity, integrating services sectors, weak transport systems they share with their and increasing competition are multiples of developing country neighbors. reducing tariffs (Hoekman and Konan 1999). Progress often requires coordinated However, there is very little convincing ex post actions; for example, joint customs inspec- evidence of significant returns to regional ini- tions must be allowed, common rules for tiatives that must deal with these issues, sug- transport must be established (including vehi- gesting that substantial progress is difficult to cle weight restrictions), and test results from achieve. partners' laboratories must be accepted. RTAs Ignoring institutional barriers during a tar- can provide a forum to enhance trust among iff reform may undermine the objectives of trade partners that genuinely wish to move reform--and indeed produce perverse results. forward on these and other fronts. For example, tariff liberalization in the face of border delays and customs corruption may have no impact on imports and may even re- The Costs of Trade duce welfare if tariff revenues are replaced by D espite globalization and the rapid increase longer waits to clear customs (Cudmore and in trade over the past 40 years, the costs Whalley 2003). In the absence of competition of trading remain substantial--particularly for in the domestic transport sector, trade liberal- developing countries (box 4.1). Because of ization may simply lead to a transfer of rev- those costs, the actual volume of international enue from the government to monopolistic trade is far less than economic theory would transport owners. On the other hand, progress predict in the absence of significant barriers to on many issues is not possible while high tariff trade [the case of the "missing trade" (Trefler barriers remain in place. 1995)]. And trade within countries is much Cost raising barriers may be linked in cir- more intense than between countries. If trade cles of causation, with significant impacts due costs were insignificant, the propensities to to scale economies in transport. For example, trade nationally and internationally would be a reduction in tariffs or a decline in costs at the equal. In fact, crossing a national border port may stimulate trade that can offer oppor- appears to dampen trade flows even in regions tunities for transport companies to operate at such as the EU, where formal trade barriers more efficient levels of scale. And if there is ef- and customs posts have been removed.2 fective competition in the transport sector, this Finally, the retail prices of particular goods could lead to lower transport prices and more 78 B E Y O N D T R A D E P O L I C Y B A R R I E R S Box 4.1 Trading can be costly V arious policies and factors isolate national · Costs incurred when crossing a border due to economies from world markets and thereby raise documentation, delays, and bribes to corrupt the cost of international trade: officials. · Compliance with national product standards · Tariffs, quantitative restrictions, and other bor- and technical regulations. der barriers--such as taxes on trade that raise · Insurance against risk, especially credit risk, the prices of imported goods relative to those and uncertainty associated with macroeconomic produced domestically. instability, lack of effective institutions, and · Transport costs, both direct (freight and unpredictable politics. insurance) or indirect (inventory costs). trade, and so on.3 A reduction in corruption more important--and more costly. To com- and delays at the border may stimulate trade, pete in international markets and function add to government revenues, and allow for a within global production chains, firms need reduction in tariffs to achieve a given revenue not only low transport costs and efficient target, which again stimulates trade. ports, but also short transit times, reliable de- Landlocked countries that face high barriers livery schedules, appropriate storage facilities, in moving their imports and exports through and security (Carruthers and others 2003). neighboring countries have no choice but to High transport costs, inefficient or corrupt pursue bilateral or regional solutions.4 These customs, and long delays at borders reduce need not be embedded in a regional preferential the trading opportunities available to many trade agreement (PTA), but to be effective for developing countries and can have significant small countries, agreements must provide for economic and social costs (box 4.2). Con- the settlement of disputes. Such provisions are versely, better conditions tend to be related to likely to be more effective if they are part of a higher levels of trade (Wilson and others broad and comprehensive agreement. 2003). Increasing the efficiency of customs, Finally, removing institutional obstacles to for example, can reduce costs and increase crossing borders has a more certain benefit trade (figure 4.1). High transport and border than reducing intra-regional barriers, because crossing costs thwart, in particular, the poor it saves real resources. Trucks that make more landlocked developing countries.5 deliveries to the port are more productive. In- terventions that lead to higher productivity Regional integration can help promote have the greatest impact on trade and more efficient and effective customs welfare--and on further increases in productiv- operations ity. In contrast, removing revenue-generating Unlike many other factors that raise trade tariff barriers on a preferential basis can lead costs, there is broad agreement on what con- to trade diversion and reductions in welfare. stitutes good customs procedures. Since its inception, the World Customs Organization (WCO) has developed best practices of cus- Regional Agreements to Facilitate toms policies and procedures. The Kyoto Con- Trade and Transport vention commits its signatory members to im- A s countries develop their trade beyond the plement these best practice principles and export of basic agricultural and extracted provides them with guidance in their efforts to commodities, logistics requirements become improve national practices. While there is 79 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Box 4.2 Border delays tax trade E ach day lost in transport delays is equivalent to a · Crossing a border in Africa can be equivalent tax of about 0.5 percent (Hummels 2000). The to the cost of more than 1,000 miles of inland situation in crossing borders between developing transportation; in Western Europe the countries can be much worse. equivalent is 100 miles (Arvis 2004). · In Southern Africa, delays at the main border- Border delays are associated with other trade crossing between South Africa and Zimbabwe costs as well, especially corruption in customs--and (Beit Bridge) amounted to six days in February have been linked to the spread of HIV/AIDS. The 2003, leading to an estimated loss of earnings World Bank has recently initiated a project in West- per vehicle of $1,750, equivalent to the costs of ern Africa to reduce, by the end of 2006, the average a shipment from Durban to the United States. time for commercial vehicles to clear border formali- · In Central Asia, on average, it takes more ties along the Lagos­Abidjan corridor by at least than 100 hours to cross the border between 20 percent, and average delays at the Nigeria­Benin Uzbekistan and Turkmenistan. A truck travel- border by at least 50 percent. These reductions are ing from Tashkent to Berlin, passing through critical for this project and its mandate to reduce the Turkmenistan, Iran, and Turkey, will spend, on incidence of sexually transmitted infection among average, a third of total transport time waiting commercial vehicle drivers. at border crossings (UNESCAP 2003). · In the Andean Community, trucks spend more Source: World Bank staff. than half of the total journey time at border crossings (Pardo 2001). other's problems and difficulties and can en- Figure 4.1 More efficient customs are gender the sharing of best practices and posi- associated with more trade tive experiences among members. This ex- Ratio of trade to GDP (percent) change is likely to be more relevant and better 250 accepted inside the regional group of develop- 200 ing countries than examples from countries that are much more advanced and face very 150 different implementation issues. On the other hand, when regional units are 100 R2 0.338 made up of developed and developing coun- 50 tries, there is scope for financial support and technical assistance for less developed coun- 0 0 5 10 15 20 25 tries in their modernization efforts. The EU, Days through customs, imports for instance, provides assistance to the Source: Investment Climate Surveys data and Global Trends African, Caribbean, and Pacific (ACP) coun- as cited in Subramanian and others 2003. tries and incorporates customs technical assis- tance provisions in its Euro-Mediterranean Initiative. Such assistance will also be available much that countries can do individually to im- under the Economic Partnership Agreements it prove customs procedures, there is also scope intends to establish with regional groupings in for regional initiatives to modernize customs. Africa, such as ECOWAS. Similarly, Japan Contacts fostered by regional agreements provides funding for capacity-building initia- can generate a mutual understanding of each tives in APEC member countries. 80 B E Y O N D T R A D E P O L I C Y B A R R I E R S For a variety of reasons, tackling customs examinations and lengthy border cross- issues autonomously may be too daunting a ing procedures. A simple first stage task, and cooperation with trading partners would coordinate hours of operation and may create the necessary momentum to over- provide compatible computer systems on come reluctance and opposition from domestic both sides of the border; these efforts policymakers, customs officials, and traders. would increase efficiency significantly. A review of a number of regional initiatives to · Joint training centers. Countries can join modernize customs suggest the following areas forces to operate regional training cen- in which RTAs can lead to improvements: ters that can benefit from leveraged-up resources and can build cohesion be- · Align customs codes with international tween the customs officers of different standards. A good customs code sup- customs services in the region. ports efficient customs operations. It es- tablishes the competence of the relevant Transport and trade facilitation authorities, promotes transparency and initiatives raise productivity predictability of operational procedures In recent years there has been a development and enforcement, encourages coopera- of a web of transport and trade facilitation tion with the private sector, provides for (TTF) agreements aimed at easing the move- effective appeals procedures, and en- ment of goods and services across borders. hances integrity. It would be advanta- Most of these agreements have been reached geous for all countries to align their as part of, or in parallel with, an RTA.6 customs codes with international Effectively implemented, TTFs can improve standards. access to global markets for developing · Simplify and harmonize procedures. The countries with poor transport systems-- recommendation here is to introduce a particularly landlocked countries. single customs document that limits the TTFs often contain provisions to standard- data requirement to a single set and ize customs procedures at borders and to har- adopts e-commerce techniques. monize customs documentation. TTFs can · Bring all tariff structures in line with further facilitate trade by providing for the the international harmonized tariff clas- interoperability of transport resources and by sification (HS). Many disputes can be fostering market access and competition in avoided if all members of the grouping the transport sector. adhere strictly to an identical tariff Divergent national regulations for truck size classification. and weight require vehicle checks on both sides · Strive for transparency. Increase the of the border and often lead to overload availability and accessibility of the legal charges and costly delays (box 4.3). In text and regulations that traders and Southern Africa, for example, axle-load regu- customs officials require and include lations are different in Namibia, Botswana, other relevant information such as trade and Zambia (Röschlau 2003). Truckers who statistics. are in full compliance in one country can be · Adopt and effectively implement the prosecuted and fined across the border. WTO Valuation Agreement. Member Regulations concerning insurance, driver's countries can assist each other through licenses, and other documentation provide effective mutual assistance agreements ample opportunities for cost savings. For ex- and shared databases. ample, the COMESA carrier's license system · Work together toward customs integrity. allows companies to operate regionally with- · Establishment of joint border posts. out having to pay for multiple licenses. And Joint border posts preclude multiple COMESA's vehicle insurance scheme enables 81 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Box 4.3 Standardization and simplification can increase trade volumes:The case of the Trans-Kalahari Corridor T he Trans-Kalahari Corridor (TKC), the road resulted in agreements (October 2000) to extend the route between Gauteng province (South Africa) operating hours of customs at the Namibia/Botswana and Walvis Bay (Namibia) via Botswana was opened border from 22 to 24 hours to enable loading and in 1998, replacing the traditional longer route unloading in Windhoek and crossing the border in through western South Africa. Despite major road the same day. rehabilitation in 1999, traffic reached only 15 per- In August 2003, the TKC started a pilot phase to cent of the expected capacity. The major obstacles replace all existing transport documents with a single occurred at the border crossings. This led the TKC administrative document (SAD). To complement this Corridor Management Group to seek a partnership effort, South African Customs developed a website with the customs administrations of Namibia, with details on the SAD process. Border processing Botswana, and South Africa. This partnership times were cut by more than half, from an average time of 45 minutes to 10­20 minutes. According to the United States Agency for International Develop- ment (USAID) estimations, reduced border delays created savings of $2.6 million per year along the Zimbabwe corridor. As a result, the route had become economi- Namibia cal, and traffic flows increased. Operators were mov- Botswana Walvis Trans-Kalahari ing about 620,000 tons annually along the TKC, Bay Mamuno about 65 percent of expected capacity, until the Windhoek Gaborone Ramatlabama Trans-Kalahari Pioneers Gate Mozambique Botswanan government increased road user charges Maputo corridor offices/ Skilpadshek in February 2004. In some cases, road charges were border posts: Ramatlabama Walvis Bay Johannesburg Swaziland multiplied by a factor of 10. The customs problem Windhoek Lesotho had been settled, but following this unilateral deci- Trans-Kalahari Mamuno South Africa sion affecting the transport sector, traffic decreased Gaborone Durban Pioneers Gate significantly. Skilpadshek Ramatlabama Cape Town Ramatlabama Johannesburg Source: World Bank staff. transport operators to comply with insurance domestic transport monopoly, the gains from obligations throughout the region with a single lower operating costs and more efficient cus- policy. Similarly, ECOWAS's brown card sys- toms procedures may accrue to the transport tem, introduced in 1982, has helped to reduce company in the form of greater monopoly settlement time significantly. The success of profits--with little impact on trade and such initiatives requires effective cooperation poverty. Equally important is competition be- between different ministries (transport, inte- tween routes and between different modes of rior) in the member states. transport. For example, Lao goods were The impact of harmonizing customs proce- almost exclusively exported through Vietnam dures and transport rules may be limited until the Lao People's Democratic Republic unless there is competition in the domestic developed alternative transit routes through transport sector. In the extreme case of a Thailand. It now takes one day for an export 82 B E Y O N D T R A D E P O L I C Y B A R R I E R S container loaded with garments from customs frameworks, rather than regional Vientiane to reach a main international trans- arrangements (UEMOA Commission 2000).8 port node at Bangkok, compared with three to Many transport companies oppose the adop- four days to reach Danang (Banomyong tion of regional frameworks for fear of 2000). upsetting the "tour de role" system.9 Bilateral Although competition is often included in transport treaties often predefine the transport regional treaties (such as the 1982 ECOWAS share of both countries (normally 50-50); the convention regulating interstate road trans- exporter therefore has no choice in selecting a portation or the 1993 COMESA Kampala transport operator. The effect of this system is Treaty),7 effective implementation is rare. Be- to protect less efficient operators. Even the cause national authorities often fear a loss of more sophisticated North America Free Trade sovereignty if they allow foreign operators in Agreement (NAFTA) has experienced difficul- the market, a regional legal framework and ef- ties in implementing cross-border trucking fective enforcement mechanisms are necessary competition, following protectionist pressures for successful implementation. In the case of from U.S. unions, concerns about truck the EU, full implementation was not achieved safety, and Mexican driver qualifications and until 1985--28 years after the signing of the competence. Treaty of Rome. Since then, the benefits have MERCOSUR countries implemented the been substantial (box 4.4). "International Common Manifesto Cargo and In several West African countries, transit Customs Transit Dispatch" (IMC/CTD) in regimes are governed by national, bilateral, or 1991. This form harmonized and unified all Box 4.4 Logistics costs in Europe have fallen in the last two decades S ince the late 1980s the proportion of company Logistics cost in Europe revenues spent on logistics in Europe has declined from 14.3 percent to 6.8 percent, a reduction that Percentage of revenue far exceeds the average level of EU external tariff on 15 manufactured goods (Mentzoni 2003). 1.5 Administration and planning This development reflects the following important Warehousing trends in the logistics industry: 10 5.8 1.4 · Centralization of inventory through a smaller 1.2 Transport 4.1 1.1 number of warehouses; 5 3.4 2.6 · Increased outsourcing of logistics services to 7 specialized companies; and 4.6 3.1 3.1 · Just-in-time supply policy (Ruijgrok 2001). 0 1987 1993 1998 2003 These changes have followed key policy initia- Source: ELA/ATKearney. tives. The removal of borders within the EU reduced uncertainty and the costs of international transport. domestically in all member countries, regardless of At the same time, the EU acted to increase competi- their country of registry. tion in the industry through the adoption of cabotage--a policy that enables carriers to operate Source: World Bank staff. 83 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 information required by different border dismantling of trade barriers among neighbor- control institutions (customs, migratory; see ing countries, regional cooperation can create Nofal 2004). a broader constituency for facilitating trade flows. Integrating fragmented markets can Regional cooperation reinforces transport make infrastructure projects more viable, and and trade facilitation programs thus promote a virtuous cycle of integration Trade liberalization--whether unilateral, mul- and growth. tilateral, or regional--may have a very muted Realizing the inherent potential of trans- economic impact in the presence of very high port and trade facilitation requires both a transport costs, weak logistical services, and sound institutional environment and a con- long delays to clear customs. Conversely, in ducive economic one. RTAs can help address the presence of high trade barriers there may institutional gaps and reinforce the corridor be little reason for traders to lobby for im- approaches that are common in agreements provements in transport. And trade restric- between developing countries. At the same tions that limit quantities may undermine the time, RTAs can provide a forum for the dis- incentive to invest in improved transport and cussion and definition of norms and harmo- trade facilitation services. Hence, actions to nized rules that are often necessary for effec- improve transport and reduce trade barriers tive implementation. Involvement of the are often complementary. private sector in these discussions is often a Trade facilitation can be effective even in prerequisite for effective action (box 4.5). the absence of a formal RTA. However, a for- Finally, negotiating transport issues in a re- mal agreement may help to entrench and en- gional forum can act to depoliticize issues hance facilitation initiatives beyond what is (Schiff and Winters 2002). possible through cooperation alone. In princi- Despite the enormous potential for gains ple, unilateral and multilateral liberalization from regional initiatives to improve customs by itself should lead to larger trade volumes-- and transport services, progress in many cases and hence raise incentives to invest in trade has been slow. Trade costs remain very high for facilitation. By inducing a more thorough many developing countries. Many initiatives Box 4.5 The case of the Northern Corridor Stakeholders Consultative Forum S ince 1999, officials dealing with transport, transit, · Development of a one-stop processing center; and private operators along the Northern Corri- and dor (including Ministry of Transport, Ministry of · Reduction of the number of required stamps to Trade, customs agencies, exporters, and importers go through Mombasa port (from 21 to 11). associations, etc.) have been regularly meeting twice a year to discuss transit issues. This private-public As a result of this forum, national transit and trade sector alliance has produced the following positive facilitation committees are being established in the re- developments: gion. Private sector participation has been extended to include insurance clearing agents, bank associations, · Elimination of charges on imports routed shippers' council, and the like. Public/private partner- through the port of Mombasa (by Kenya ships to tackle trade and transport facilitation are also Bureau of Standards and the Kenya Plant being established in West Africa. Health Inspectorate Service); Source: World Bank staff. 84 B E Y O N D T R A D E P O L I C Y B A R R I E R S to facilitate trade have suffered from a lack of standards can constitute substantial barriers effective implementation. In several agree- to trade, but these may only become clear ments, disputes over implementation can only after other barriers have been addressed. Re- be raised at the political level, which often ducing tariffs and improving customs and means that small landlocked countries have transport can be likened to reducing the water great difficulty in securing the necessary com- level in a swamp only to find a range of previ- pliance from larger neighbors. In these cases ously covered "snags and stumps that need to implementation is very much a function of po- be cleared away" (Baldwin 1970). litical will. When producers must alter their product to The possibility of taking legal action meet divergent standards in foreign markets, under regional treaties can help drive imple- they lose some of the benefits of larger scales mentation of transport facilitation initiatives. of production. When the foreign government The European Court of Justice has played an does not recognize standards-compliance tests important role in the implementation of a performed in the exporter's home market, or common transport policy in Europe (Funck the home country does not have the facilities 1998). A regional court of justice has re- to test the product, the exporter must foot the cently been established in the Eurasian Eco- bill for additional tests in the foreign market. nomic community of the CIS; another re- For example, in Moldova the certification of gional court exists in UMEOA in West Africa. organic nut production exported to Germany While a regional court of justice does not has to be renewed every 6 months, and each guarantee implementation, it does create po- visit from an international certifying company tential for more efficient enforcement than is costs $5,000 plus $2,000 per production available through less formal dispute settle- test--once before processing and once after ment channels. processing. This can amount to $18,000 per year, which is a heavy burden for firms in an economy such as Moldova, an economy trying Standards, Conformity to compete in international markets. Upgrad- Assessments, and RTAs ing testing facilities and measuring equipment T he construction and implementation of is essential for reducing the costs of confor- systems of standards, quality assurance, mity assessments.11 accreditation, and metrology are crucial to To reduce the damping effect of divergent competitiveness and sustained growth--and standards on international trade, WTO mem- hence to development. Standards have become bers have agreed to discipline the use of key elements for facilitating transactions and mandatory standards by governments. These trade both within countries and in interna- are relatively modest provisions--they deal tional exchange between countries. Standards with transparency of standards regimes, equal support markets and provide for efficient treatment, and the need to justify standards transactions. Standards and technical regula- that differ from internationally agreed-on tions stipulate what can or cannot be ex- norms. Efforts to reduce barriers to trade changed, and they define the procedures that caused by standards and conformity assess- must be followed for exchange to take place.10 ment have been more extensive in a small The ability of would-be exporters to com- number of RTAs, although empirical evidence ply with mandatory health and safety stan- identifying the benefits of these interventions dards, as well as market-driven voluntary is scant at best. The issue for developing coun- standards in overseas markets, is a major fac- tries is whether regional initiatives can provide tor determining access to those markets. for more efficient and effective standards and Divergent product standards and duplicative conformity assessment systems. These im- systems for assessing conformity with those proved systems would allow governments to 85 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 meet domestic objectives to raise health and different levels of development and institu- safety levels, and at the same time, facilitate tional capacities. trade. We start by discussing the EU experience, where integration has proceeded the furthest. Different paths to better standards systems The basis for the free movement of goods in Different approaches are available to raise the EU is the principle of mutual recognition standards and to address technical barriers to of the regulations of partners or the recogni- trade. Countries can unilaterally upgrade tion of equivalence. Although standards vary standards by adopting international stan- from one country to another, it is presumed dards. However, the technological content and that they are designed to meet the same regu- the health, security, and environment objec- latory objectives and to offer equivalent levels tives of the international standard may not be of protection to the public. Thus products appropriate for developing countries, because produced in partner countries can be accepted the international standards are strongly influ- with the assumption that those products will enced by the OECD countries. Further, some not undermine basic regulatory objectives of the returns to adopting the international concerning health, safety, and the environ- standard--in terms of greater market access-- ment.12 Mutual recognition of regulations is only materialize if the country's trading part- the simplest approach to differences in stan- ners also accept products produced to that dards: it is a powerful tool for removing bar- standard. riers to trade in goods and services, and with A second approach requires cooperation this approach the difficulties of detailed har- between countries to upgrade standards. monization measures, which intrude on na- Countries agree that products satisfying par- tional policy making, can be avoided.13 How- ticular standards will be accepted in each ever, mutual recognition of standards requires other's markets. However, cooperation agree- a high degree of trust between regulatory au- ments do not discipline other market access thorities (essentially the responsibility for pro- barriers, so that returns from the upgrading of tection of domestic consumers is, in part, standards may be undermined if other barriers transferred to the overseas partner). As such, are raised to protect a particular sector once mutual recognition can only work in regions standards are harmonized. Typically, the dis- comprising countries of similar levels of in- pute panels in cooperation agreements have a come that have comparable standards. mediation role, not an arbitration role. For Effective institutions are also important. In these reasons, and unless all parties are com- the EU, governments can defer from nondis- mitted to the upgrading process, the process of crimination and the free circulation of goods standards upgrading could have important for reasons of "public policy or public security" obstacles. and protection of health as long as such re- The upgrading of standards within a RTA strictions are not a disguised restriction on is characterized by more formal institutions, a trade. To ensure the latter, the EU has devel- higher degree of enforcement, and greater oped the following mechanisms for disciplin- trust originating from the frequent interac- ing national regulations and interventions into tions between members and the comprehen- product markets:14 sive nature of the agreement. Members cannot use tariffs to prevent the entry of a product · Infringement procedures, whereby the satisfying the regional standard; this increases European Commission acts to enforce the certainty that a country's upgrading efforts community law, although such proce- will be translated into greater market access. dures are very time consuming and costly, Within PTAs different approaches to stan- have an impact only after the event and dards have been followed, which reflect the are ad hoc. 86 B E Y O N D T R A D E P O L I C Y B A R R I E R S · Notification procedures, whereby mem- apparent in the 1970s and early 1980s, when ber states are required to notify all draft new national regulations were proliferating at technical regulations for scrutiny by an a much faster rate than the production of EU Committee, whose objective is to pre- European directives harmonizing regulations vent new regulatory barriers to trade. In (Pelkmans 1987). practice, all new national regulations of It became clear that the degree of interven- EU member states have to pass an EU tion by the public authorities before a product test regarding their impact on the free was placed on the market needed to be re- movement of goods. duced, and that changing the decision-making · Notification of derogation procedures, procedure to allow the adoption of harmo- which require member states to notify nization directives by a qualified majority in authorities of cases in which they wish to the Council was needed. The "new approach" prevent the sale of goods lawfully pro- regulations indicate only "essential" health duced or marketed in another member and safety requirements, allowing greater free- state, on the grounds of nonconformity dom to manufacturers to satisfy the essential and nonequivalence with domestic re- requirements and to industries to flesh out quirements. This ensures that any dero- product specifications in the form of volun- gations from the principle of mutual tary standards. The new approach makes recognition are transparent and subject good use of established standardization to scrutiny. bodies--European Committee for Standard- ization (CEN), European Committee for Elec- While mutual recognition of regulations trotechnical Standardization (CENELEC), underpins the EU Single Market, it has been and the national standards bodies. Standard- apparent for a long time that for certain prod- ization work is achieved in a more efficient ucts and for certain risks (when consumers are way, is easier to update, and involves greater directly exposed to hazards), equivalence be- participation from industry. Products that tween levels of regulatory protection embod- conform to the standards promulgated by the ied in national regulations cannot be assumed. European standards agencies are presumed to In these cases the EU seeks agreement among comply with the essential requirements of the members on a common set of legally binding regulations. However, these standards are vol- requirements. EU legislation harmonizing untary, and firms can produce to different technical regulations has involved two distinct standards if they can prove compliance with approaches, the "old" and the "new." the requirements of the regulation. The old approach mainly applied to prod- MERCOSUR has followed the old ucts (chemicals, motor vehicles, pharmaceuti- approach of the EU and focused its limited cals, and foodstuffs), involved extensive resources on harmonizing national stan- product-by-product or even component-by- dards at the regional level (see Nofal 2004). component legislation, and was carried out by MERCOSUR has formulated 366 common detailed directives. Achieving this type of har- technical regulations and some 300 voluntary monization was slow for two reasons. First, standards. The Andean Community has re- the process of harmonization became highly cently decided to focus regional harmoniza- technical, with attention given to very detailed tion on a targeted number of standards--those product categories. Consultations were often of the products most traded. Only 40 regional drawn out. Second, the adoption of directives standards were created, although they cover required unanimity in the Council, which 60 percent of trade. meant that they were slow to be adopted. The When harmonized regulations are pursued, limitations of the old approach as a broad tool it is important to avoid overly bureaucratic for tackling technical barriers to trade become mechanisms. Harmonization through the use 87 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 of detailed regulations can lead to excessive in- to trade can be removed when standards from tervention by public authorities before a prod- different members are shown to be compati- uct can be placed on the market and have a ble. This is the approach of NAFTA and also chance to prove its viability. Regulators should tends to be applied in bilateral trade agree- concentrate on defining essential health and ments that have a standards component, such safety requirements while allowing firms the as those between Chile and the EU and Chile flexibility to meet those requirements and not and the United States.15 The compatibility ap- stifle technological change and competitive- proach is the converse of the mutual recogni- ness. The CIS, MERCOSUR, and Andean tion of regulations. Under the compatibility Community countries still apply an approach approach, the standards of a trading partner based on very detailed harmonized regula- are assumed to be insufficient in their ability tions. When a company wishes to introduce a to satisfy the importer's regulatory objectives, new product, it is often necessary to change unless proven otherwise. the existing regulations or wait for a new tech- nical regulation to be promulgated, which can Recognizing the results of conformity take considerable time and be very costly. A re- assessment in partners cent Peruvian technical regulation for gas con- Mutual recognition of conformity assessment tainers specifies the minimum width of the [usually negotiated in the form of a mutual walls, stating the exact thickness, which effec- recognition agreement (MRA)] is necessary if tively prevents the use of new materials that nontariff barriers are to be fully removed. might be lighter but thicker. This ensures that the test results from labora- ASEAN is also following a policy of tar- tories in the exporter's home market are ac- geting key sectors, but it is harmonizing cepted by the importer so that the costs of around international standards rather than duplicative testing can be avoided. This agree- promulgating its own standards. For 20 key ment does not require that both countries product groups, members should adopt, as have the same standards nor that both coun- national regulations, the agreed-on interna- tries be members of a PTA. For example, the tional norms. Members that do not adopt any EU and the United States, for certain sectors, of the identified international standards as accept the results of product tests (for com- their national standards still need to accept patibility with their own standards) that have products from partners that comply with been completed in the partner's laboratories. these international standards--unless they If conformity assessment institutions are can demonstrate an inability to adopt the in- relatively weak, however, harmonization of ternational standard due to "climatic condi- standards may be a necessary step to facilitate tions or infrastructural reasons." mutual recognition of conformity assessment. Here the contribution of the RTA has been This is the approach being followed in to provide an enforcement mechanism ASEAN. The Andean Community established through dispute settlement procedures, such a regulation for compulsory mutual recogni- that members who do not accept from part- tion in 2003 for sectors covered by regional ners products that satisfy an ASEAN standard standards. MERCOSUR will proceed with are ultimately liable to fines for compensation mutual recognition of conformity assessment or removal of concessions. Therefore, ASEAN procedures in the near future. However, countries can adopt the ASEAN standards MERCOSUR is a perfect example of how the with confidence that incurring the associated conformity assessment infrastructure is lack- costs will not be undermined by subsequent ing: Policymakers prefer to harmonize stan- denial of access to partners' markets. dards before moving to mutual recognition of In agreements where regional institutions conformity assessment, but firms do not show are weak, especially free trade areas, barriers much interest in standards because, in the 88 B E Y O N D T R A D E P O L I C Y B A R R I E R S absence of mutual recognition of conformity sometimes even between the institutions of assessment, the returns to investment in stan- individual countries. These close relationships dards are low. This suggests that improve- allow obstacles to be overcome in informal ments in the conformity assessment infra- ways, circumventing cumbersome formal in- structure are necessary. terventions. MERCOSUR has yet to adopt Singapore has signed a bilateral trade mutual recognition of standards, but many agreement whereby the United States recog- conflicts over standards have already been nizes certifications provided by Singapore to solved by telephone between relevant officials some of its East Asian partners. This high- in member countries. lights that rules of origin can be an important A degree of trust between public institu- element in an MRA. If Singapore has a com- tions and the private sector is also important parative advantage in the region in testing if more flexible approaches, such as mutual and laboratory facilities and is well endowed recognition of conformity assessment and/or with professional staff in this activity, then the of regulations, are to succeed. Strong, central- U.S. agreement with liberal rules of origin ized regulatory cultures tend to produce tech- (whereby the United States accepts tests from nical regulations that are too detailed and dif- Singapore labs of products from other coun- ficult to change. It is important to ensure the tries), may help to establish or enhance the effective participation of the private sector position of Singapore as a regional hub for and consumers so that the setting of stan- testing and conformity assessment. Rules of dards and their enforcement reflect broad origin that restrict the testing and conformity rather than narrow interests. activities to products produced only in Singapore would tend to constrain such a de- Successful cooperation in harmonizing velopment. EU MRAs tend to have these standards depends on simple principles restrictive rules of origin.16 To date, RTAs in the developing world have not realized their full potential for overcom- Regional trade agreements can facilitate ing standards-related obstacles to regional or mutual recognition of standards global trade, although some slow progress is Effective solutions to problems arising from evident, such as in MERCOSUR. That is different standards require a high degree of di- likely to change as the WTO agreements on alogue and trust among trading partners. Technical Barriers to Trade (TBT) and Sani- RTAs, while not the only path to trust, tend to tary and Phytosanitary Measures (SPS) come promote dialogue and communication, which into full practical application, and as the im- in turn build trust. This has been the case for portance of reforming standards systems in member countries of MERCOSUR and the developing countries gains prominence. In the Andean Community, in which trust has grown meantime, several principles can contribute to as integration has deepened (Nofal 2004). successful cooperation in standards and con- Such trust needs to be nurtured through open- formity assessment procedures. ness and transparency when new national A first step for developing countries is to regulations are being considered. identify priority sectors for reform to keep RTAs also can provide a favorable negoti- costs low and gather momentum for further ating environment and so reduce politicization reform. The sectors to prioritize are those in standard disputes among members, making where trade costs resulting from differences in it easier to find common solutions for the re- standards and conformity assessment proce- moval of non-tariff barriers. The interactions dures are higher and where trade between that take place in an RTA often improve insti- members is large. tutional relationships between the different Second, if international standards exist for standards bodies of member countries--and these products and they are appropriate for 89 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 all members given their level of development, product costs (World Bank 2003). This hurts then the simplest approach is to harmonize MENA exporters and prevents MENA coun- around these international standards. This tries from joining global production chains will only be relevant if all members have the destined for EU markets. capacity to implement them. It is important to Many developing countries are too small to clearly define objectives before harmonizing efficiently offer a full range of these confor- standards so as to avoid overregulation. For mity assessment services. Often there are too example, requiring information on labels that many laboratories offering poor quality the consumer is unlikely to understand will services. increase costs and contribute little to the RTAs can contribute to better conformity objective of making information available to assessment first, and most simply, by facilitat- consumers.17 ing dialogue and the sharing of technical An open and transparent system, with knowledge. More ambitious initiatives can standards published in an accessible official build regional accreditation bodies to increase bulletin before being implemented, is essential efficiency and enhance the reputation of local if regional initiatives are to facilitate trade certification bodies in the global market. An between members and preclude the difficul- open regional market for laboratory services ties facing exporters from outside of the re- can lead to cheaper yet higher quality testing gion. Ensuring flexibility is also important; on the basis of specialization and economies of thus, regulations that set minimum standards scale. However, it must be stressed that while rather than detailed requirements are less re- the potential gains are large, there are very few strictive for firms. When countries face inter- successful initiatives that can provide useful nal resistance to modernizing and adopting guidelines. ASEAN provides an example where new standards, the compliance with stan- members are pushing forward with a number dards can be offered on a voluntary basis. of initiatives, including cooperation on legal This allows those firms that are able and will- metrology, and efforts to enhance conformity ing to satisfy the new standards to progress. assessment bodies to facilitate mutual recogni- Daskalov and Hadjikolonov (2002) show tion of test reports and certifications. how such an approach made it possible for RTAs can also provide a framework for more advanced and competitive local pro- collaboration that increases the effective par- ducers in Bulgaria to quickly adopt European ticipation of developing countries in interna- standards before they were formally intro- tional standards organizations and at the duced as mandatory Bulgarian standards. WTO. This is important if international stan- In the short run, the greatest gain for many dards and conformity assessment measures developing countries is likely to come from are to reflect the interests of developing improvements in the testing, certification, and countries--and therefore make the TBT and accreditation institutions to underpin greater SPS agreements relevant for the majority of enforcement capacity. Initiatives that improve WTO members. For this participation to be these institutions are likely to have large pay- successful, the structure of international stan- offs by allowing governments to achieve more dards institutions needs to be modified to effectively their existing objectives concerning reduce the costs of representation of develop- health and safety and facilitating greater ex- ing countries. The International Standards ports on both a regional and a global basis. Organization (ISO) has moved one step in this For example, most MENA countries require direction by being the first standard institu- that testing be done at their national labora- tion to allow electronic voting. This could be tories, which are usually less sophisticated extended to other institutions. Another possi- than European testing centers and saddled bility would be to allow RTAs to represent with cumbersome procedures, pushing up their members in standard-setting committees, 90 B E Y O N D T R A D E P O L I C Y B A R R I E R S which would reduce members' costs of power, and the like--allow countries to trade representation. off gains in one area against losses in another, Attempts to remove barriers caused by dif- reducing or even eliminating the explicit com- ferences in standards and conformity assess- pensatory schemes that would otherwise be ment requirements will be more effective in a needed (Schiff and Winters 2002). climate of trust and mutual understanding. Consider some examples. The Southern Such a climate requires a genuine willingness to African Development Community (SADC) liberalize and is unlikely to result in agreements provided the coordination point for regional with many exceptions, frequent recourse to integration in a regional power cooperation safeguard measures, and high barriers at the agreement. The Southern African Power Pool border that can be due to customs delays and (SAPP), launched in 1995, was designed to inefficient port and transport services. take advantage of power resources in the re- The lack of relevant examples of successful gion and was the first formal international intervention at the regional level to deal with power pool outside of North America. The standards issues makes it difficult to derive 12-country region has abundant hydropower clear proposals based on best practices. In this resources, especially the Inga Reservoir, large light the best recommendation is for develop- reserves of cheap coal in South Africa, and the ing countries to proceed with caution and Karriba Dam on the Zambia/Zimbabwe bor- concentrate on targeted coordinated action der. The pool covers 6 million square miles for which the institutional requirements are and serves 200 million people. Utilities in the not extensive and the gains are clear. region had been trading electricity for decades through bilateral contracts, but these were cumbersome to administer. The objective for Trade-Related Regional shifting to the pool was to create a more effi- Cooperation Agreements cient regional market. The SAPP is modeled C ountries can benefit from other forms of on the "loose" pools in Western Europe and cooperation that are linked to trade di- the United States, which emphasize constant rectly through RTA arrangements or indirectly exchange of information to maximize the cost when they influence trade-related inputs or and reliability benefits from trading and sys- outputs.18 Such trade-related cooperation can tem autonomy. Rather than relying on central deal with shared resources, such as water, fish- dispatch, loose pools rely on long-term bilat- ing areas, power, railroads, or the environ- eral contracts drawn up with common designs ment. Schiff and Winters (2002) make the case and security standards plus some central ser- that in the presence of economies of scale or vices. Unlike in the developed world, SAPP inter-country externalities, market solutions to membership is limited to national utilities. problems are not necessarily the best, and Each member must meet its Accredited regional cooperation can often pay large Capacity Obligation, a requirement that each dividends. utility have capacity to cover its forecast When regional cooperation arrangements monthly peak. Each member is also obliged to are embedded in RTAs, it may be easier to cover emergency energy up to six hours, to conclude and implement these arrangements. provide automatic generation control and Increasing trade raises the level of salience of other facilities in its control area, and to allow all aspects of regional cooperation and may wheeling through its system. SAPP includes foster greater high level attention to the re- most Southern African Development Commu- gional arrangement and allow for more effec- nity (SADC) members and is predicated on tive and informal dispute resolution. More- the latter's institutions, including the SADC over, agreements that cover more policy Treaty, the SADC Dispute Resolution domains--for example, trade, transport, Tribunal, the SADC energy ministers, and the 91 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Technical and Administrative Unit. The en- measures to protect the monarch butterfly. ergy ministers are responsible for resolving Currently there is an active task force working major policy issues. to stop the smuggling of endangered species. Though still in its early stages, the pool's po- Under this program, U.S. Fish and Wildlife tential benefits include reducing or postponing officers are training Mexican officers. new requirements for generating capacity and While the trade agreements underpinning reserves, lower fuel costs, and more efficient these regional cooperation initiatives were not use of hydroelectricity. A SADC electric power essential to the actual activities, it is clear that study conducted in 1990­92 estimated a sav- they have provided useful political and institu- ings of 20 percent ($785 million) in costs over tional synergies. 1995­2010. Three factors were critical to the develop- Conclusions O ment of the regional agreement: The availabil- ne advantage of regional preferential ity of complementary power sources, an active trade arrangements is that they create regional organization for economic coopera- opportunities to lower trade costs in areas tion, and the political will to support in- other than tariffs and non-tariff barriers to creased regional energy trade. SADC and its trade. This review of trade facilitation, stan- predecessor, the Southern African Develop- dards administration, and regional coopera- ment Coordination Conference, served as tion agreements points to several conclusions. focal points for promoting regional integra- The potential to expand trade by lowering tion and facilitating investments in the needed trade costs other than policy border barriers is interconnection projects. great--and it may have a higher payoff to co- NAFTA offers another example. NAFTA operative governmental efforts than reciprocal has also fostered regional cooperation for the reductions in border barriers. This is because environment by tying essentially extraneous the costs of institutional obstacles, informal environmental issues to the trade and invest- barriers, and sub-optimal regulatory scales are ment deal. This link helped to create the nec- often higher than the costs associated with essary political support for NAFTA in the policy border barriers. Further, many of these United States, and it encouraged Mexico to barriers do not generate revenues but simply accelerate their environment program in order waste economic resources and directly con- to close the deal. The North American strain productivity. These issues are also Agreement on Environmental Cooperation important elements defining the investment (NAAEC) was signed as one of the side agree- climate. ments appended to NAFTA at the last mo- RTAs can precipitate cooperation to lower- ment. It created the Commission for Environ- ing trading costs in these areas because RTAs mental Cooperation (CEC) in Montreal in raise the level of policy salience, spread infor- early 1994 to carry out the provisions of the mation about members and about interna- agreement. The CEC has a young but growing tional markets, improve the institutional effi- conservation portfolio, focused mainly on pro- ciency of countries (better coordination tecting habitats and species. A broad program between the different institutions within a of cooperation to protect North American country and between countries), provide "in- birds is in place, aimed at identifying impor- stitutional homes" for joint initiatives, and tant bird areas across the three member coun- may facilitate dispute resolution across tries and tying them into a protected network. multiple areas. A Biodiversity Information Network is under Countries need not act in concert to reap creation, and strategies are being developed the benefits of unilateral reforms; the chances for cooperation to protect marine and coastal of unilateral success are much improved, ecosystems. The CEC has also coordinated however, when policymakers are well 92 B E Y O N D T R A D E P O L I C Y B A R R I E R S informed about international standards and and ASEAN. Two RTAs have no associated TTF: the trade-facilitation activities of other coun- NAFTA and GCC. Three examples of TTFs existing tries. In the absence of such information, and independently of RTAs are ECO, ECOWAS, and the Northern Corridor Transit and Transport Agreement. of the capacity to act on it, it is unlikely that a 7. In treaties, competition is literally ensured country, acting alone, will be able to match through "equal treatment of carriers" or non- the benefits from participating in an RTA. discrimination regarding carrier's nationality, known Finally, it seems clear that many RTAs are as respect of the third party rule. not realizing their potential as a forum for 8. According to UMEOA Commission (2000), reducing trade costs. North-South agreements 73 percent of the legal rules and customs governing appear to have had somewhat greater success, transport and transit regimes were derived from bilat- eral treaties (34 percent), national legislation (24 per- perhaps because of the institutional interests cent), and customs (15 percent); 27 percent were de- and strength of the more advanced partner. rived from regional treaties. 9. Collusion between transport operators leads to Notes agreement on price setting. National associations play 1. Lawrence (1997). a role in determining which goods a company will 2. For example, McCallum (1995) reports results transport. suggesting that Canadian provinces are more than 20 10. Standards can be mandatory as defined by gov- times more likely to trade among themselves than they ernments (through technical regulations) so as to meet are to trade with U.S. states after controlling for the main their objectives regarding health, safety, and environ- economic determinants of trade. Subsequently, Nitsch mental issues; as well as voluntary, reflecting the de- (2000) found evidence of substantial border effects in mands and tastes of consumers or the technological re- Europe, with internal trade being, on average, larger by quirements of industrial purchasers. In addition to the a factor of 10 than trade with EU partners, although the writing of standards, an essential element of the system magnitude of this effect did decline during the 1980s. of standardization is conformity assessment, the tech- 3. See, for example, Hummels and Skiba (2002). nical procedures such as testing, verification, inspec- 4. GATT Article V mandates freedom of transit and tion, and certification, which confirm that products national treatment of products in transit. However, this fulfill the requirements laid down in regulations and provision has never been invoked. The WTO frame- standards. work provides little leverage for poor, landlocked coun- 11. From The Republic of Moldova Trade Diag- tries to improve transit conditions. nostic Study, World Bank, 2004. 5. Trade facilitation is of particular importance 12. The principle of mutual recognition was devel- to landlocked countries, whose products must pass oped on the basis of European Court of Justice case through numerous border crossings and checkpoints. law, specifically, the Cassis de Dijon and Dassonville Of the 50 least developed countries, 16 are landlocked: judgements. In the former case, cassis from France was Afghanistan, Bhutan, Lao People's Democratic Repub- prevented from being sold in Germany because it did lic, Nepal, Burkina Faso, Burundi, Central African Re- not contain enough alcohol! public, Chad, Ethiopia, Lesotho, Malawi, Mali, Niger, 13. Mutual recognition of regulations has also Rwanda, Uganda, and Zambia. Even coastal develop- been used within federal countries to remove barriers ing countries may be effectively landlocked if they are to inter-state trade. For example, Australia formally not on major shipping routes and are served by ineffi- adopted mutual recognition in 1993 to remove regula- cient and high cost coastal feeder services to main tory barriers to the free flow of goods and labor be- ports. Being landlocked has a significant and depress- tween Australian states and territories. As in the EU, ing effect on trade. For these countries, a regional ap- some harmonized regulations are promulgated at the proach may be the only way to improve access to federal level. global markets, since there seems to be little scope at 14. See Pelkmans and others 2000. present for solving transit issues within the WTO. Cor- 15. Bilateral trade agreements with the EU tend ridor solutions are efficient responses to the transport to contain support for developing capacity in the problems of landlocked economies with deficient infra- developing country, whereas those with the United structure. By definition they require bilateral or re- States provide little such support. gional intervention. 16. Chen and Mattoo (2004) show evidence that 6. RTAs with TTF approaches include the MRAs promote trade, but that restrictive rules of ori- European Union, MERCOSUR, Andean Community, gin lead to trade diversion, especially against develop- SADC, COMESA, EAC, UMEOA, SAFTA, Eurasec, ing countries. 93 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 17. Large companies in Brazil used their influence Barriers to Trade in Europe, eds. P. Brenton, and to their advantage in setting a voluntary labeling stan- S. Manzocchi. Cheltenham, UK: Edward Elgar. dard that requires water bottlers to include the results Engel, C., and J. Rogers. 1996. How wide is the bor- of numerous tests that are not understood by, or par- der? American Economic Review 86: 1112­1125. ticularly relevant to, the consumer, but that constitute Evans, C. 2001. Home Bias in Trade: Location or an effective barrier to entry to small companies that Foreign-ness? Staff Report 128, Federal Reserve cannot afford the battery of tests required to provide Bank of New York. the information. SEBRAI, a private institution that Faini, R. 2004. Trade Liberalization in a Globalizing provides support services for small enterprises in World. Paper presented at Annual World Bank Brazil, is working to improve access to certification Conference on Development Economics, and metrology for its client firms. Through "solidarity Washington, DC, May 2004. certification," for example, SEBRAI helps groups of Funck, R. 1998. Integration of Transport Systems- the small enterprises become certified at subsidized group European Experience. Report for the EU-SADC rates. SEBRAI also provides bonds that subsidize small Conference on Transport, Maputo, October. enterprises' expenditures for metrology services. Hoekman, B., and Konan, D. 1999. Deep Integration, 18. This section draws heavily from Schiff and Nondiscrimination, and Euro Mediterranean Free Winters (2002). Trade. Discussion Paper 2095, CEPR, London. Hummels, D., and A. Skiba. 2002. A Virtuous Circle? Regional Tariff Liberalisation and Scale Economies in Transport. In Integrating the Americas: FTAA and Beyond, eds. A. Estevade- References ordal, D. Rodrik, A. M. Taylor, and A. Velaso. Anderson, J., and E. van Wincoop. 2004. Trade Costs. MA: Harvard University Press. NBER Working Paper 10480, National Bureau of Hummels, D. L. 2000. Time as a Trade Barrier. GTAP Economic Research, Cambridge, MA. Working Paper 18, Center for Global Trade Arvis, J. F. 2004. Transit and the Special Case of Land- Analysis, Department of Agricultural Economics, locked Countries. In Handbook of Customs Mod- Purdue University. ernization, eds. L. de Wulf and J. Sokol. Wash- Lawrence, R. 1997. Preferential Trading Agreements: ington, DC: World Bank. The Traditional and the New. In Regional Part- Banomyong, R. 2000. Multimodal Transport in South ners in Global Markets: Limits and Possibilities East Asia: A Case Study Approach. PhD thesis, of the Euro-Med Agreement, eds. A. Galal, and University of Cardiff, www.bus.tu.ac.th/usr/ruth. B. Hoekman. London: CEPR. Baldwin, R. 1970. Non-Tariff Distortions of Interna- McCallum, J. 1995. National borders matter: Canada- tional Trade. Washington, DC: Brookings U.S. Regional Trade Patterns. American Eco- Institution. nomic Review 85: 615­23. CEC (Commission of the European Communities). Mentzoni, J. T. 2003. Trade Logistics--A Study of 1998. Technical Barriers to Trade, Volume 1. Logistics Costs in Norwegian Wholesaler Indus- Subseries III: Dismantling of Barriers of The Sin- try. Processed. www.transportbrukerne.no. gle Market Review. Luxembourg: CEC. Nofal, Beatriz. 2004. Constructing a deeper integra- Cudmore, E., and J. Whalley. 2003. Border Delays and tion in MERCOSUR. Background paper for the Trade Liberalization. NBER Working Paper Global Economic Prospects 2005. Washington, 9485, National Bureau of Economic Research, DC: World Bank. Cambridge, MA. Nitsch, V. 2000. National Borders and International Carruthers, R., J. Bajpai, and D. Hummels. 2003. Trade: Evidence from the European Union. Cana- Trade and Logistics: an East Asian Perspective. In dian Journal of Economics 33: 1091­105. East Asia Integrates: A Trade Policy Agenda for Pardo, M. 2001. Pasos Fronterizos en la Comunidad Shared Growth, eds. K. Krumm, and H. Kharas. Andina. IADB. Processed. Washington, DC: World Bank. Pelkmans, J. 1987. The New Approach to Technical Chen, M., and A. Mattoo. 2004. Regionalism in Stan- Harmonisation and Standardisation. Journal of dards: Good or Bad for Trade? World Bank. Common Market Studies 25. Processed. Pelkmans, J., E. Vos, and L. Di Mauro. 2000. Reform- Daskalov, S., and D. Hadjinikolo. 2002. The Impact of ing prodict regulation in the EU; a Painstaking Technical Barriers to Trade on Bulgaria's Exports Two-level Game. In Regulatory Reform and Com- to the EU and to the CEFTA Countries. In En- petitiveness in Europe, Volume I, eds. G. Galli and largement, Trade and Investment: the Impact of J. Pelkmans. Cheltenham, UK: Edward Elgar. 94 B E Y O N D T R A D E P O L I C Y B A R R I E R S Röschlau, F. 2003. Double Standards Cause Regional Bailleurs de Fonds sur les Infrastructures et le Transport Nightmares, e-corridor, Walvis Bay Transport Routier des Etats Membres de Corridor Group, number 3. l'UMEOA. Ouagadougou. Ruijgrok, C. 2001. European Transport: Insights and UNESCAP (United Nations Economic and Social Com- Challenges. In Handbook of Logistics and Supply- mission for Asia and the Pacific). 2003. Transit Chain Management, eds. A. Brewer, K. Button, and Transport Issues in Landlocked and Transit De- D. Hensher, chapter 3. New York: Pergamon Press. veloping Countries. New York, ST/ESCAP/2270. Schiff, M., and A. Winters. 2002. Regional Coopera- Wilson, J., C. Mann, and T. Otsuki. 2003. Trade tion, and the Role of International Organizations Facilitation and Economic Development; Measur- and Regional Integration. Policy Research Work- ing the Impact. Policy Research Paper 2988, ing Paper 2872, World Bank, Washington, DC. World Bank, Washington, DC. Trefler, D. 1995. The Case of the Missing Trade and World Bank. 2003. Trade, Investment, and Develop- Other Mysteries. American Economic Review 85: ment in the Middle East and North Africa: 1029­46. Engaging with the World. Washington, DC: UEMOA (West African Economic and Monetary World Bank. Union) Commission. 2000. Projet de Rapport de World Bank. 2004. The Republic of Moldova Trade Synthèse Préparatoire à la Table Ronde des Diagnostic Study. World Bank, Washington, DC. 95 5 Beyond Merchandise Trade: Services, Investment, Intellectual Property, and Labor Mobility As barriers to merchandise trade have come From a development perspective, the most down and trade has expanded, policymakers potentially beneficial components of this set of and trade negotiators have turned their atten- issues are provisions that open services mar- tion to services and trade-related regulatory kets to additional potential suppliers through issues. Of these, services, investment, intellec- foreign subsidiaries (in GATS terminology, tual property, and temporary movement of Mode 3) and the temporary movement of labor arguably have the greatest potential for workers (Mode 4).1 Services liberalization in affecting incomes and trade in developing preferential arrangements can enlarge the countries. Agreements on these four issues are number of competitors and carries fewer risks now becoming common in bilateral and some of income losses than preferential merchan- preferential regional trade agreements (RTAs). dise trade because lifting most common re- North-South agreements, notably the strictions does not cost the government rev- bilateral free trade agreements of the United enue. Though multilateral liberalization is States and of the European Union (EU), have usually preferable even in services,2 RTAs in been the important drivers for services, in- services can be predicted, in general, to in- vestment, and intellectual property rights crease welfare. Similarly, preferential agree- (IPRs). In broad terms, the United States, for ments that widen the scope for the temporary example, offers access to its large market for movement of workers have the potential to goods in exchange for access to services mar- raise incomes. kets in developing countries and their accep- In both services and labor mobility, how- tance of rules governing investment and ever, agreements have yet to fulfill their devel- intellectual property rights. The EU market opment potential. Many of the North-South access agreements also cover many of these agreements are between countries with unusu- topics, if less specifically. Labor services-- ally open service sectors, so the additionality to that is, the temporary movement of the various parties is limited to a handful of workers--are largely confined to profes- relatively small sectors and to the credibility sional and skilled workers, often intra- effects of locking in openness via treaties and corporate transfers. South-South agreements "seals of approval" that investors might take as tend to feature services liberalization less a sign of lower risk. Meanwhile, in many of the prominently, and their rules governing South-South agreements, where the potential investment, intellectual property, and even scope for liberalizing measures is often far the temporary movement of workers, are greater, RTA-driven additional liberalization commonly weak or absent altogether. has been sporadic. For labor services, both the 97 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 North-South and South-South agreements are Services, Investment, and IPRs confined to intra-firm movement of profession- in Regional Agreements als, and neither agreement has substantially orth-South agreements differ sharply in widened market access for the temporary N their coverage of services, investment, movement of labor. and intellectual property. At one end of the By contrast, North-South agreements re- spectrum, U.S. FTAs usually involve the most garding investment and IPRs have succeeded explicit negotiations for market access in ser- in promulgating comprehensive new rules that vices and U.S.-style rules for investment and go beyond multilateral rules in the agreement intellectual property. The EU market access on Trade Related Intellectual Property Rights agreements similarly contain market access (TRIPs). The United States and EU bilateral provision in services, but tend to reinforce pre- agreements have enhanced market access vailing international rules for intellectual prop- through negative-list and positive-list (respec- erty; its Economic Partnership Agreements in tively) pre-establishment rights, and the Africa use development assistance in combina- United States has implemented investor-state tion with trade preferences to promote rules dispute settlement mechanisms that empower beyond international agreements, including foreign investors to seek arbitration awards in EU-style concerns for competition policy and cases of uncompensated expropriation or geographical indications. At the other end of other violations of treaties. the spectrum, most South-South agreements Ironically, of the four areas, investment and are focused primarily on merchandise trade, IPRs are the two where the development po- and tend to treat services, investment, and tential is largely unproven. The investment IPRs unevenly, if at all. These distinctions provisions that enhance investor's rights have should become clearer when we consider the not been shown to increase the flow of invest- U.S., EU, and South-South approaches in turn. ment to developing countries. Nor have stronger IPRs embedded in the TRIPS-Plus agreements been shown to accelerate techno- U.S. FTAs are rule intensive logical flows to low-income countries-- Key features of the U.S. FTAs that cover ser- though it may do so for middle-income coun- vices, investment, and intellectual property tries. On the other hand, because free trade rights include: areas that result in larger markets do attract additional investment flows, it may be that in · Opening services markets to competition combination with large, preferential trade from foreign suppliers or locking in areas, enhanced investor protections and IPRs prior autonomous liberalization, except do have a positive impact--but agnosticism in those sectors excluded (i.e., on a nega- seems warranted. tive list). Because most of the countries This chapter begins with a synoptic com- with which the United States has con- parison of agreements and a focus on the cluded bilateral FTAs are already open in regulation-intensive bilateral U.S. and EU most sectors, the agreements generally free trade agreements (FTAs). Understanding lock in prevailing openness and affect the diversity and reach of these agreements changes in only a few still-restricted permits us, in a subsequent section, to activities. Significant market openings review the economic consequences of provi- took place in the Costa Rican telecom- sions that deal with services, investment, munications and insurance sectors and and intellectual property. A final section less dramatic market openings occurred examines the treatment of movement of in the banking sector in Bahrain. Provi- temporary labor. sions range from inclusion of insurance, 98 h i s i s i s j k 1 m esY Ye Ye Ye No No No No No esY Property Property Intellectual Intellectual TRIPS+ TRIPS+ TRIPS+ TRIPS+ TRIPS+ TRIPS+ TRIPS information -State n n n esY esY esY No esY esY esY No No No esY No esY No No No esY esY No Dispute 2003; Settlement Investor OECD on No No No No No No No agreements. Ban TRIMS+ TRIMS+ TRIMS+ TRIMS+ TRIMS+ TRIMS+ TRIMS+ TRIMS+ TRIMS+ 2004b; Performance Requirements trade nda free laws. Pre- No No U.S. None IP 2004a Investment establishment Limitations of Negative-list Negative-list Negative-list Negative-list Negative-list Negative-list Negative-list Positive-list Negative-list Positive-list Positive-list Positive-list Positive-list equent Abbott f subs No No No No No No esY No No 2004b; other Ownership Limitations harmonization Negative-list Negative-list Negative-list Negative-list Negative-list Negative-list Negative-list than on 2004a, work g countries. limitations. Post- esY esY esY esY esY esY esY No No esY esY ­ No esY Szepesi No No National begin reatment/MFT establishment member to 2004; comprehensive e the and No esY esY esY esY esY esY No No No No No No No No esY esY of pre-establishment Cosbey Ratchet one committees Mechanism and in comparison from specific d A less to w/o is Mann ty: No esY esY esY esY esY esY No No No No No No No No esY esY business" service. non-members. Right Provide a excluded to 2003; Services oper establishment those not coverage pr supply agreement. interparliamentary resolution. c but to the than for Access "substantial chapter Fahnbulleh dispute Services Limitations into Exceptions other & Market Negative-list Negative-list Negative-list Negative-list Negative-list Negative-list Negative-list No Standstill Positive-list Positive-list Positive-list Negative-list Positive-list Positive-list Positive-list Negative-list Negative-list TS. the and Pre-establishment conduct GA required members, provides -state b la to intellectual not à sectors but, eldeV IP te Origin in However incorporated laws. investor and of esY esY esY esY esY esY esY No esY esY esY esY esY esY No No esY esY esY do that treatment Plus. conventions. for 2004; Rule establishment include (Nonrestrictive) patent no not a commitments companies person is TRIPs patents. new Sauve of and for provide automatically equitable of does all vestment, and reatment Access in esY esY esY esY esY esY esY international No esY esY esY No specified esY No esY No esY esY that treatment. is and juridical to agreement. treaties National considered MFN/T Market Not to fair negotiation regulates adoption Mattoo vices, equitable are agreement only shareholdings grant for and adherence framework future liberalization Ser a calls investment treaties; A Africa does governments. Community fair for non-establishment, equity only benefits community has 5.1 A of on provisions Legal by IP MERCOSUR 128(e) le bilateral abT U.S.-Jordan U.S.-Chile U.S.-Singapore U.S.-Australia U.S.-CAFT U.S.-Morocco NAFT EU-South EU-Mexico EU-Chile MERCOSUR Andean CARICOM ASEAN SADC COMESA Japan-Singapore Canada-Chile Chile-Mexico Includes Denial Provides Right Autonomous Limits COMESA The Requires The Andean Act ASEAN EU Agreements U.S. EU South-South Other a. b. c. d. e. f. g. h. i. j. k. l. m. n. Sources: provided 99 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 financial advisory services, and selected Other noteworthy provisions (not the sub- telecommunications services to arguably ject of this chapter) include labor protections relatively minor changes to the already and environment issues that figured promi- open regimes, such as the commitment of nently in the CAFTA, Chile, and Singapore Singapore to cease cross-subsidies in ex- agreements, among others. Signatory coun- press mail delivery or the commitment of tries committed to enforcing their own labor Chile to open selected insurance services laws in five areas: right of association, the (table 5.2). right to organize and bargain collectively, pro- · Ratchet provisions and negative-list ex- hibitions on forced labor, a minimum age for clusions. The ratchet clauses mean that employment of children, and acceptable new autonomous liberalization will au- working conditions. Complaints can be filed, tomatically be subsumed under the terms and if the agreed-on procedures to mediate the of the agreements. Negative lists ensure dispute fail, a panel of experts would review that yet-to-be-invented new service areas the case and, if warranted, impose a fine to be are guaranteed to be covered by the used for the enforcement of labor rights; that treaty. Notable for their absence is the is to say, trade sanctions are not an agreed-on exclusion of labor services, except provi- remedy (Weintraub 2004). sional visas for professionals associated The FTAs involve innovations in trade law with investing firms (discussed in the in two important areas: investment and IPRs: penultimate section of this chapter). Investment Access and Protections. The · Investment rights. Investment rights, with FTAs have incorporated the provisions of provisions for national treatment, bilateral investment treaties (BITs), and in nondiscrimination, pre-establishment some cases, provided new measures cover- provisions for companies based in each ing investment (table 5.1). Agreements, espe- others markets, bans on trade-related cially post-NAFTA ones, include broad investment measures (TRIM), and definitions of investment, comprising not investor-state arbitration of dispute lim- only foreign direct investment (FDI), but also ited only by a negative list of exclusions. portfolio flows, private debt, and even sover- · TRIPs-Plus provisions that provide eign debt issues as well as intellectual prop- stronger protections for IPRs than under erty (Mann and Cosbey 2004; Vivas-Eugui the TRIPS agreement, with investor-state 2003). The inclusion of short-term debt, to- arbitration dispute settlement permitted gether with pre-establishment rights, led the in the event of disputes (subject to U.S. Treasury to demand that Chile modify certain limitations). its controls on capital inflows that were Table 5.2 Additional services liberalization in U.S. FTAs Chile Australia Bahrain CAFTA Morocco Singapore Banking Insurance Telecommunication Broadcasting & Audiovisual Financial Advisory Services and Data Retail/Wholesale Distribution Restrictions on Foreign Directors & Managers Express Mail Delivery Real Estate Legal Services Source: Legal treaties. 100 B E Y O N D M E R C H A N D I S E T R A D E designed to curtail destabilizing hot money procedures; the operation of the agreement inflows.3 Such broad definitions expose was to be reviewed by January 1, 2000, but countries to dispute settlements across a range so far the review has not occurred (Bora of assets that go far beyond multilateral 2003). All of the bilateral FTAs ban, in some commitments. form, trade-related investment requirements, All agreements provide for treatment of such as by local content rules, value-added re- foreign investors on the same basis as domes- quirements, and restrictions on management. tic investors (national treatment) and have The U.S. bilateral agreements have, in effect, provisos banning discrimination among in- established a "TRIMs-Plus" set of obligations vestors from member countries (MFN, or that includes outright bans on certain perfor- nondiscriminatory treatment). For many of mance requirements, including exports, mini- the initial FTA countries, these stipulations mum domestic content, domestic sourcing, had been included in national legislations trade balancing, and technology transfer. and/or had been incorporated into bilateral In general, government procurement, investment treaties, mainly on a post- environmental standards, some health mea- establishment basis. sures, and requirements for local research and What is new is the extension of the pre- development (R&D) are all exempt (Te Velde establishment right to invest in businesses and and Fahnbulleh 2003). activities in all sectors, except where expressly Freedom to make transfers is a nontrivial prohibited via a negative list.4 These pre- investment right granted under the investment establishment rights lock in the right of agreements. This assures investors that Mexican and Canadian investors under they will be able to transfer profits, make NAFTA to invest in all activities in the United investments, or lend without government States. Exceptions for the United States in- interference. clude foreign investment with NAFTA guaran- Finally, all U.S. agreements except the tees in selected areas of communication, Australian FTA create an investor-state media, transportation, and social services. dispute resolution provision that permits in- Pre-establishment rights mark a broad expan- vestors to take foreign governments to dispute sion of market access by foreclosing future resolution for violation of the treaty's national government policies that would raise barriers treatment, nondiscrimination, or expropria- to foreign investment. The rationale for ac- tion provisions, among others. NAFTA's cepting such disciplines is that it provides cer- Chapter 11 and Chile's Chapter 10 are the tainty on the rules of the game, which most widely known mechanisms, but these will in turn translate into increased investment mechanisms are contained in the other bilat- inflows. eral agreements as well. Another discipline more expansive than Intellectual Property Rights. The IPR pro- multilateral accords is in trade-related invest- visions embedded in all recent U.S. FTAs go ment measures (TRIMs). The WTO TRIMs beyond the multilateral IPR standards estab- agreement of 1995 attempted to clarify disci- lished in the WTO's TRIPS Agreement. plines on government policies that require "TRIPS-Plus" elements found in many--but foreign companies to establish joint ventures, not all--of the IPR chapters include:5 export in a certain portion of its sales or bal- ance trade, use local inputs to achieve value- · Extension of the patent term for delays added objectives, or hire local staff. However, caused by regulatory approval processes; the agreement failed to provide adequate def- extension of the term of copyright pro- initions of disciplines, and it presented poorly tection to life of author plus 70 years formulated implementation periods and (compared to life of author plus 50 years inadequate notification and monitoring in TRIPS). 101 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 · A requirement to provide patent protec- The inclusion of these services, investment, tion for plants and animals. and IPR issues was a contributing factor to the · A limitation on the use of compulsory breakdown in negotiations in the Free licenses for national emergencies and as Trade Area of the Americas (Nogues 2004). antitrust remedies, and for public non- We return to these issues below when consid- commercial use. ering the economic consequences of these arrangements. In the area of pharmaceuticals: EU FTAs take a different approach · An obligation to prohibit the marketing In addition to market access in merchandise, approval of generic drugs during the the EU has focused heavily on services in its term of the drug patent. bilateral FTAs. The earliest (and least specific) · A five-year period of marketing exclusiv- is the South African agreement (1999) that ity following the submission of safety contains only the promise of potential liberal- and efficacy data to drug regulatory ization after discussions transpire in 2004 and authorities (so-called data exclusivity). 2005. In the EU-Mexico FTA, several general In addition, marketing exclusivity effec- provisions were included (many ratifying tively applies across borders, so that mar- GATS arrangements), as well as specific liber- keting approval in one market--say, the alization commitments in the financial sector. United States--impedes registration of The EU-Chile agreement went further than the competing products in another market. other two and included liberalization of · An additional three-year period of mar- telecommunications and maritime services keting exclusivity based on the submis- (Ullrich 2004). sion of new clinical data with respect to The EU agreements with Mexico and Chile new uses of previously approved drugs. differ from the U.S. agreements in important Exclusivity would also apply to drugs for respects. First, the trade provisions are which the patents have expired (although phrased on the basis of a positive list and generic competition for previously ap- implicitly exclude new products. Second, the proved uses would remain unaffected). treatment of intellectual property effectively · Imposition of restraints on parallel im- reaffirms a multilateral approach to IPRs, portation, impeding the possibility that because the agreements provide only the list parties to the agreements open their of conventions that signatory countries have markets to the import of products that already ratified, those it intends to ratify, and have already been sold--possibly more those that it will consider ratifying in the cheaply--in foreign markets. future.6 This approach differs from that taken In the area of digital works: toward the EU-accession countries, in which new entrants were required to apply the rigor- · An obligation against circumventing ous EU standards on data protection and mar- so-called technological protection keting exclusivity; these have a major impact measures--devices and software devel- on generic producers. oped to prevent unauthorized copying of The treatment of investment and capital digital content. Rules on the liability of flows in both agreements does not appear to Internet service providers (ISPs) when be extensive. For example, the EU-Mexico copyright infringing content is distrib- agreement simply states that the existing re- uted through their servers and networks. strictions on investment will be progressively These provisions are based on standards eliminated and no new restrictions adopted; found in the U.S. Digital Millennium the agreement did not specify particular sec- Copyright Act of 1998. tors or set a timeline for liberalization. The 102 B E Y O N D M E R C H A N D I S E T R A D E language in the EU-Chile agreement calls for (MERCOSUR) that have delivered services the "free movement of capital relating to di- liberalization additional to levels negotiated rect investments made in accordance with the multilaterally or determined unilaterally. Re- laws of the host country." In both instances, gional agreements in services have competed the agreements allow for the use of safeguards to create complex structures of rules and com- in the event of monetary or exchange rate dif- mitments. But in many cases, the sound and ficulties, and although the time limit is set fury of the negotiations has signified limited at 6 months for Mexico and 12 months for liberalization. Many agreements do not pro- Chile, it would allow for continuation of the vide new market access beyond what coun- safeguard after the time limit through its for- tries have already scheduled with the GATS. mal reintroduction. That said, many of the In telecommunications and financial services, same investor protections found in U.S. FTAs the GATS has in fact achieved a higher level of are also found in EU bilateral investment bound liberalization than that offered in most treaties with developing countries. RTAs. The treatment of dispute settlement is sim- All South-South agreements have a rela- ilar in both agreements. In general, the EU has tively nonrestrictive definition of preferential no special provisions pertaining to investment, access; by allowing firms from nonmember but these are covered under the general dis- countries that have "substantial business" in pute settlement provisions for all matters in member countries to invest through sub- the agreements (Szepesi 2004a and b). Dispute sidiaries based in member countries, the num- settlement is covered on a state-to-state level ber of potential competitions in the market is and is first attempted through consultations enlarged. with a Joint Committee (Association Com- Agreements with negative lists have several mittee in the case of Chile) within 30 days of advantages in terms of market access: they a party's request. If this step of "dispute avoid- permit automatic liberalization of new service ance" proves unsuccessful, the concerned industries; they establish a stronger floor for party can forward its request to an arbitration liberalization by locking in the status quo; panel comprised of representatives of both they are more transparent; and they may lead parties. The arbitration panel's decisions are to a more productive internal dialogue with binding, and the panel can also rule on the sectoral private interests (Mattoo and Sauve conformity of any measures undertaken as a 2004). Ratchet mechanisms that allow new result of its decision with the original ruling. autonomous liberalization to be incorporated Both agreements provide extensive detail on automatically into treaties are most likely to the process of appointing members to the co-exist with negative list provisions. arbitration panel, timelines for the panel's However, most of the South-South agreements ruling, and compliance with the panel's have not liberalized many sectors, and some, decisions. like MERCOSUR, have not implemented accords in the way that was anticipated at South-South agreements focus signing (Nofal 2004). on expanding trade Investment provisions have differed as Virtually all of the other major agreements well. South-South RTAs generally have been contain references to services liberalization. less ambitious with respect to investor protec- Most agreements allow for national treat- tions. This is true for the right to provide ser- ment, post-establishment nondiscriminatory vices without establishing local affiliates. It is provisions (table 5.1). At the other extreme, also true for investor-state dispute settlement. there are more limited agreements like Associ- For the most part, only the United States and ation of Southeast Asian Nations (ASEAN) EU bilaterals have established sophisticated and Southern Common Market Agreement mechanisms to deal with disputes on 103 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 investment. Some agreements provide for investment climate--but the potential gains investor-state dispute resolution, though these are large (World Bank 2001). protections are less strong than in the North- Realizing these gains requires allowing for- South Agreements. eign investors greater market access, and this Intellectual property rights, while men- is the most important provision in a preferen- tioned in South-South agreements, rarely go tial arrangement. Countries can open previ- beyond disciplines negotiated at the multilat- ously closed sectors to RTA partners as part of eral level, and they do not have the tightly for- an agreement. Since today most countries ac- mulated provisions that characterize the cept, indeed clamor for, foreign investment in North-South agreements, notably those with manufacturing and natural resources, RTA- the United States. MERCOSUR, for example, driven reductions in entry barriers affect has agreed to establish a commission to exam- mainly services. Moreover, services now play a ine areas of intellectual property harmoniza- larger role in investment flows, and for some tion, while ASEAN has a framework agree- countries, such as Mexico, they have dwarfed ment. The Andean community has more investments in manufacturing. The great bulk detailed restrictions, but these are written less of services investment are market-seeking, with the view of protecting intellectual prop- horizontal investments. These cover a vast erty than of eliminating territorial restrictions range of large multinationals: Deutsche Bank, on the use of patented technology; the res- WalMart, Starbucks, Microsoft, and so on. trictions were designed to end the restraints These "mode 3" services require the commer- multinational companies put in their technol- cial presence of affiliates, branches, or fran- ogy contracts with their foreign affiliates, chises to deliver the service. To be sure, some which explicitly prevented them from using countries (such as India) have experienced the patents in export production. substantial flows associated with call centers and data processing, and this new investment Economic Consequences of accompanies these cross-border supply ("mode 1") activities, though these activities Services, Investment, and IPR remain small in comparison to trade through Provisions in RTAs commercial presence. New market access in services could Because preferential arrangements permit promote growth more suppliers to compete in the market, a Services liberalization with proper regulation country is almost certain to gain from prefer- can be a powerful driver of economic growth ential liberalization of the services trade, irre- and poverty reduction. At the sectoral level, spective of the supplier. This is in sharp con- removing barriers to competition can lower trast to merchandise trade, where the income prices, improve quality, and add variety. Be- loss associated with trade diversion can occur cause of the linkage effects--the fact that pro- with the loss in tariff revenue. In services, bar- ducers require telephones, use finance, need riers to entry usually take nonmonetary forms adequate transportation services, and benefit such as regulatory restrictions on entry, for- from business services--improving service sec- eign equity limitations, quotas on outputs and tor performance can generate huge economic foreign service workers, and requirements on gains. Mattoo, and others (2001) show that legal form of establishment. None of these countries with fully liberalized financial and generate revenue for the government, so re- telecommunications sectors grew annually on moving these restrictions is less likely to pro- average about 1.5 percentage points faster duce income losses (with merchandise trade, than other countries, controlling for other fac- income losses associated with trade diversion tors. These gains are not automatic--they occur because the government loses the tariff require adequate regulation and a supportive revenue as trade is diverted to higher-cost 104 B E Y O N D M E R C H A N D I S E T R A D E sources of imports). Moreover, the scope for effects on economy-wide performance. Because increased competition and exploitation of of the strong potential links to growth (World scale economies, as well as the possibility of Bank 2002), the additional market access pro- inducing knowledge spillovers, strengthens the vided through RTAs could be important. Un- presumption that a country would gain from a fortunately, the actual additional liberalization preferential agreement in services. has not yet matched this promise. Multilateral, nonpreferential liberalization Nonetheless, restrictive rules of origin can is likely to produce even larger gains than limit the potential benefit to liberalization. preferential regional agreements. This is be- Participants who seek to benefit from prefer- cause multilateral liberalization opens the ential access to a protected market and deny market to the largest number of competitors benefits to third country competitors are likely and permits consumers maximum choice; it to argue for the adoption of restrictive rules of allows imports from the most competitive origin, based on criteria such as ownership or source. It also leads to a less complex policy control considerations. This could be the atti- regime than a preferential arrangement, and tude of regionally dominant but globally non- therefore implies lower administration costs competitive service providers toward third- for government agencies and lower transac- country competition within a regionally inte- tion costs for the private sector. Finally, it is grating area. possible that preferences could lead to a Examples of restrictive rules of origin for higher-cost firm gaining a competitive advan- services and investment can be found in tage relative to investors from outside the re- MERCOSUR and the Andean Pact, both of gion. And first mover advantages and barriers which limit benefits to juridical persons that to entry can make it difficult for lower-cost are owned and controlled by natural persons suppliers from third countries to enter the of a member country. The Hong Kong-China market. Inefficient suppliers from member Free Trade Agreement, for example, features a countries might establish positions behind detailed annex spelling out the set of criteria market barriers sufficiently high that new and by which Hong Kong service suppliers may even more efficient potential competitors benefit from the terms of the agreement. would not choose to pay the cost of entry. Rules of origin for services, as with mer- Do RTAs attract more investment? chandise trade, can play a significant role in de- RTAs can, in theory, promote more invest- termining the degree to which regional trading ment through new trade rules that create a arrangements discriminate against nonmember larger market, new investment rules that countries, and hence the degree of competition permit market access by relaxing restrictions in services associated with an RTA. For exam- on market entry (such as discussed above for ple, if one participant has a fully liberalized services), and new investor protections. market, the adoption of a nonrestrictive rule of New trade rules that eliminate internal bar- origin by the other participants can be likened riers create a larger internal market, which can to MFN liberalization. Service suppliers can raise the return to investment and create an enter the liberal jurisdiction and from there incentive to invest for members and for third move to the other partner countries. Many gov- countries. Firms investing in the RTA coun- ernments take the liberal rules of origin one step tries can achieve economies of scale and scope further and extend regional preferences on an in serving a larger market of potential buyers, MFN basis under the GATS. This widens the may experience reductions in transactions number of competitors in the market and offers costs, and if services are included, benefit from greater opportunities for securing access to the more efficient financial, telecommunications, most efficient suppliers--particularly of infrast- and other services (Schiff and Winters 2003). ructural services likely to exert significant Trade rules can induce greater efficiency in 105 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 transactions with the global economy. activities that are not internationally Markusen (2004), for example, notes that competitive. Latin America's early experiments inward investment to reach the local market with admitting FDI behind high border barri- may also include tapping into lower cost pro- ers produced inefficient investment and a pro- duction sites within the new RTA to serve the tectionist political economy that took decades wealthier parts. Japanese multinational com- to unwind (box 5.1). The formation of RTAs panies might well locate in Mexico to reach may increase both internal and external in- the U.S. market, though the net effects might vestment, but the resulting market size may be diluted as U.S. firms also set up in Mexico. not be sufficient to realize modern scales, and Frischtak (2004) found that MNCs in autos, the high external tariffs drive up the costs textiles, and electronics reallocated their pro- of imported inputs. Indeed, the first Andean duction to Mexico to serve the U.S. market. Pact in the early 1970s and the Central Amer- The larger market can also increase pro- ican Common Market in the 1960s failed ductivity in other ways. Aside from economies to generate investment-related productivity of scale, a larger market can increase competi- gains. tion among a potentially larger set of suppli- RTAs may include new investment rules to ers, and take advantage of differing regional facilitate market access. As with services, the factor prices to drive productivity increases decision to lift an administrative barrier im- and hence more rapid growth; and the more peding manufacturing investment or an in- rapid growth provides a dynamic attraction vestment in natural resources can create an to intra-bloc and extra-bloc investment. If the opportunity for investors and hence prompt RTA reduces border protection on investment new investment. Most remaining restrictions goods and allows domestic producers to source today--equity ownerships limitation and bans cheaper and higher technology capital goods, on foreign investment in particular activities-- members may benefit. are not restrictions on manufacturing, but However, efficient results are not automatic. rather on services (e.g., broadcasting, Even though investment may be destined for a telephony, and airlines in the United States, larger market, border barriers may create incen- among other countries) and natural resources tives to invest in high-cost import-substituting (e.g., oil in Mexico). RTAs that reduce market Box 5.1 Not all investment is good investment T rade barriers in preferential arrangements can prices, even though they could have imported much divert trade to higher-cost sources. No less im- more cheaply. (It turns out domestic firms performed portant is the investment undertaken to produce the even more poorly.) Encarnation and Wells (1986) new trade. If trade policies provide high protection, found that between 25­45 percent of 50 projects they will limit competition, allow for shared monop- studied (depending on analytical assumptions) re- oly pricing, and incur the inefficiencies of price dis- duced national income; again the main culprit was tortions. This pattern of investment was common-- high protection. As average tariff levels have come and costly--in the period before high protection was down, these low-quality type of investments have brought down. Lall and Streeten (1977) who studied faded in importance. Trade and tax policy often in- some 90 foreign investments using a cost-benefit teract in ways that magnify their competition-re- methodology, found that more than 33 percent stricting effects. reduced national income; this was mainly from ex- cessive tariff protection that allowed high cost firms Sources: Lall and Streeten 1997; Encarnation and Wells 1986; to produce for the local market at very high Newfarmer 2001. 106 B E Y O N D M E R C H A N D I S E T R A D E access barriers associated with restrictions on ment climate, and, other things being equal, foreign entry are likely to have greater im- stronger rights would lower risk and entice pact through this channel than through more investment at the margin. Since investors manufacturing. put money at risk against the promise of re- Granting new investment rights may also turns in subsequent periods, predictable regu- attract additional investment, though here the lation and protection of property rights are in- case is more contentious. In general, the tegral to the investment decision. However, the strongest investor protections entail nondis- evidence for many of the same protections con- crimination among all investors, provisions tained in the bilateral investment treaties is that against expropriation, dispute settlement with these additional protections have no significant eligibility for investor-state suits, and indepen- effects on inflows of FDI (box 5.2). To be sure, dent arbitration. The legal power granted to signing an FTA with new investor protections investors to sue governments under terms of may enhance the credibility of a reform pro- the bilateral or regional agreements is ar- gram, but evidence that these have observable guably the strongest new protection in the consequences is scarce. trade agreements. These provisions differ in While the benefits of these protections in detail, but they closely mirror the bilateral in- the form of new FDI inflow are open to ques- vestment treaties, even though they are an- tion, the costs in the form of investor suits chored to the trade agreement. are nontrivial and growing. In NAFTA, for Despite the proliferation of new protections example, as of July 2004, there were 31 to foreign businesses, the positive economic cases brought under Chapter 11 (including 14 consequences have yet to be demonstrated. against Mexico, 9 against Canada, and 8 Theory would suggest that sound property against the United States). Six cases have been rights are a foundation of any country's invest- decided in favor of the investor, but the Box 5.2 Do more investor protections mean more investment? Lessons from bilateral investment treaties D oes increasing investor protections produce the BITS do not add much benefit of high investment? One test of this proposition was Hallward-Dreimeier's (2003) study Share of annual FDI flow of the enhanced investor protections through bilat- 0.3 eral investment treaties for flows of FDI among sig- natory countries.6 Analyzing bilateral flows of OECD members to 31 developing countries over two 0.2 decades, she found that, controlling for a time trend and other factors, BITs had virtually no independent 0.1 effect in increasing FDI to a signatory country from a home country. Said differently, countries signing a BIT were no more likely to receive additional FDI 0 than countries without such a pact. Even comparing 3 2 1 Year 1 2 3 flows in the 3 years after a BIT was signed to the Years before signing signed Years after signing 3 years prior, there was no significant increase in Source: World Bank Global Economic Prospects 2003. FDI. This agrees with the findings of UNCTAD (1998) that the number of BITs signed by the host was uncorrelated with the amount of FDI it received. Sources: Hallward-Driemeier 2003; UNCTAD 1998. 107 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 amount awarded has been small compared to the government and creditors. By defining initial--and inflated--claims. Tribunal awards intellectual property as an "investment," a have totaled $35 million, compared to claims foreign investor who claims his intellectual of $1015 billion. Under the similar BITs, 48 property rights have been abrogated has re- alleged BIT violations are under review arbi- course beyond the national court system to in- tration at the International Center for Dispute ternational arbitration proceedings under the Resolution. Cases have arisen out of the investment provisions of the bilateral agree- Argentine devaluation, the changes in tax pol- ments.7 These provisions have unquantifiable icy perceived as adverse by investors, expro- development benefits--and bring risk, which priations following conflict or coups, irregu- incurs uncertain costs. larities in bidding processes, and others (Peterson 2003a). In perhaps the most signifi- The combination of changes in RTAs may cant case to date, a tribunal in Stockholm or- have an effect on investment dered the government of the Czech Republic Even if protections by themselves contribute to pay one company, Central European Media little additional inflows, evidence is mounting (CME), $350 million for violation of a bilat- that RTAs--that is, the combination of appro- eral investment treaty that deprived CME priate trade rules, liberalized market access, from a stake in an English language TV and investor protections--can have positive station in Prague. effects on inflows of foreign investment, pro- This amount was 10 times higher than pre- vided that the investment climate is supportive viously known awards under arbitration cases and the size of the newly created market is and about equal to the entire public sector attractive. deficit of the Czech Republic (Peterson Indeed, Lederman, and others (2004) 2003b). found that RTAs that formed large markets at- The legal and macroeconomic conse- tracted FDI, (controlling for other factors that quences of investment rights in treaties are influence location), but that small markets largely unknown. They have not been thor- had no effect. They also found positive effects oughly analyzed and tested in arbitration for NAFTA, although the flow of FDI, even cases, and are without precedent. One could controlling for privatizations, appears to have certainly speculate about adverse outcomes. surged in the first years but has not been sus- For example, new rules for Chile could com- tained. Waldkirch's (2001) study, with less plicate management of short-term capital complete annual data, found that NAFTA in- flows; in fact, the IMF expressed reservations creased FDI substantially, mostly from the to the U.S. government that the limitations United States and from Canada. Chudnovsky that the U.S.-Chile bilateral FTA imposed on and Lopez (2001) found that FDI increased in the Chilean government regarding short-term the MERCOSUR, largely from outside capital inflows reduced the government's abil- sources, but that it often entered via acquisi- ity to manage a macroeconomic crisis. Simi- tion, displaced domestic investment, and was larly, the breadth of definition of investment tariff-hopping to produce for the local coverage opens the government to investor- market--so it probably contributed to less to state arbitration in event of default on debt or growth less than it otherwise would have. suspension of payments in emergencies-- They also found that FDI inflows tended to which may ultimately be unenforceable. For locate in the larger countries, underscoring the instance, Argentina's default has led investors need for stronger institutions and policies in to file nearly 30 arbitration cases; none of the smaller countries. these appear to have been associated with Levy Yayati, Stein, and Daude (2004) used nonpayment of debt. However, these debts are a gravity model to analyze the effects of RTAs also subject to ongoing discussions between on FDI inflows in 13 major agreements, and 108 B E Y O N D M E R C H A N D I S E T R A D E then applied their findings to a simulation for investment flows, controlling for other deter- the FTAA. They found that RTAs have a minants of FDI. These wash out the otherwise strong positive impact on inflows, and that if positive effects of RTAs. If the economy suf- these average magnitudes hold after the sign- fers from poor macroeconomic management, ing of an FTAA, the results would be substan- high levels of corruption, and poor infrastruc- tial increases in flows to FTAA countries. ture, an RTA by itself will not offset the dis- However, the distribution is uneven, and advantages. To be sure, an RTA may help gov- countries with larger post-RTA market size, ernments through their collective action to low inflation rates, strong domestic institu- improve the investment climate and bring in tions, and open economics are likely to be the more investment; but an RTA is no substitute biggest beneficiaries. for an adequate investment climate (World To investigate further whether RTA forma- Bank 2004). Second, creation of an RTA will tion can affect FDI flows in a consistent fash- not have much effect on investment inflows ion, we examine the effects of RTA member- from outside the region if restrictions on mar- ship and other variables on FDI inflows for a ket access are severe and remain unchanged. panel of 152 countries over the 1980­2002 period. The sample takes into account 238 How do IPRs affect the price RTAs (both regional and bilateral), many of of technology? which overlap, that encompass the vast ma- Creation and enforcement of IPRs have an im- jority of sample countries.8 In general, coun- portant role to play in development, but nei- tries that are more open (measured as the sum ther theory nor available studies provide much of exports and imports over GDP), growing guidance on the likely outcomes of imple- more rapidly, and are more stable (captured menting in trade agreements the strongest of in less volatile inflation rates), attract greater the IPRs or none at all. On the one hand, quantities of FDI, controlling for growth rates stronger IPR enforcement in general is likely of FDI to all countries and the world growth to enhance the overall investment climate, rate.9 RTAs that result in larger markets do at- especially for high technology firms. On the tract greater FDI. The interaction of an RTA other, recognizing full patent protection for signing and additional market size associated firms may require poor countries to pay with the integrated markets is significant and higher prices, with little additional incentive positively related to FDI. On average, a 10 either to innovate or to make investments in percent increase in market size associated with the local market.11 Full enforcement of an RTA produces an increase of 5 percent.10 patents12 could produce substantial financial This has important policy implementations: If flows, estimated roughly at $19 billion to a country seeks to use an RTA to attract in- the United States and $7 billion to Germany vestment, it should seek to amalgamate with (World Bank 2001). Moreover, the adminis- the largest possible markets; RTAs among trative costs of upgrading IPR systems are not small market countries have little effect. trivial (Finger and Schuler 2004). Two important caveats to this conclusion It was the prospect that developing coun- are worth underscoring: First, a preferential tries would have to pay higher prices for trading arrangement cannot compensate for patented drugs that motivated the interna- an inadequate investment climate. Stein and tional community to agree to clarify flexibili- Daude (2001) have shown that institutional ties embedded in the TRIPS Agreement at the variables that make up the whole of a coun- WTO Ministerial Meeting in 2001. The try's investment climate--including political resulting Doha Declaration on TRIPS and stability, government effectiveness, rule of Public Health reaffirmed the right of WTO law, and lower risks of expropriation--are all members to use the flexibilities of TRIPS in significantly associated with increases in the areas of compulsory licensing and parallel 109 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 importation to ". . . promote access to medi- On the one hand, most countries have cines for all."13 In August 2003, WTO mem- industries that rely on copyright protection and bers created a special mechanism under the that may benefit from strengthened protection. TRIPS Agreement that allows countries with And new technologies that greatly facilitate the insufficient manufacturing capacity to effec- copying of digital works pose challenges that tively use compulsory licenses by importing policymakers need to address. On the other generic drugs. hand, copyright laws have historically sought At first blush, the TRIPS-Plus portions of the to strike a balance between the interests of U.S. FTAs seem to circumscribe the policy space copyright producers and the interests of the provided in the Doha Declaration. In particu- general public. So-called fair use exemptions lar, provisions on the link between patent sta- allow the copying of protected works for edu- tus, marketing approval, and data exclusivity cational or research purposes. There are ques- appear to put limits on the spirit of the Doha tions that new rules on the technological pro- Declaration, because countries may be pre- tection measures and the liability of Internet vented from effectively employing compulsory service providers could diminish the rights of licenses to introduce competition from generic consumers and the general public (CIPR 2002). drug producers.14 To address these concerns, Ensuring fair use of copyrighted material is the U.S. bilateral agreements with Bahrain, particularly important for educational mater- CAFTA-DR, and Morocco contain side letters ial. The opportunities and gains from the use of that share the understanding that the intellec- digital libraries, Internet-based distance learn- tual property chapters do not affect the ability ing programs, or online databases would be of governments to ". . . take necessary mea- limited if access to such tools became unaf- sures to protect public health by promoting fordable or otherwise restricted by copy- medicines for all [. . .]."15 In other words, right law. government can be justified as protecting Evidence is inconclusive about the respon- public health, as permitted under the three siveness of FDI to intellectual property FTAs. The United States Trade Representative regimes. Although surveys of foreign investors office recently clarified: ". . . if . . . a drug is typically indicate concerns for IPRs, this is produced under a compulsory license, and it is often of secondary priority (Mansfield 1995). necessary to approve that drug to protect pub- Maskus (2000) concludes that countries (espe- lic health . . . the data protection provision in cially low-income countries) should focus on the FTA would not stand in the way."16 their overall investment climate to attract Notwithstanding the potential flexibilities more and high technology investment, rather provided by these side letters, they raise sev- than to fine-tune their IPRs. Nonetheless, eral questions. How widely will the parties to some multinational companies (MNCs), when the three agreements define the "protection of selecting an investment location among public health"--or, what definitions would an middle-income countries, clearly take into ac- arbitration panel use? Uncertainty in this re- count the laws governing intellectual property. spect may become itself a barrier to making Other studies have found that weak laws and use of the flexibilities and may open the door weak enforcement deter investment in middle- for restrictive interpretations by vested inter- income countries,17 but the results for low- ests. Also, several of the other U.S. FTAs do income countries are inconclusive. Finger and not contain comparable side letters, raising Nogues (2002), in fact, argue that the intro- questions about conflicts between intellectual duction of patent protection for drugs in Chile property obligations and public health objec- made several multinational pharmaceutical tives in at least some of the affected countries. companies stop production and investment The welfare effects of stronger and new and source this market from other locations. copyright protection standards are ambiguous. In summary, Fink and Maskus (2004) in their 110 B E Y O N D M E R C H A N D I S E T R A D E review of the evidence conclude: ". . . coun- temporary movement of workers--that is, by tries that strengthen their IPR regimes are un- going beyond the relatively limited scope of likely to experience a sudden boost in inflows the GATS Mode 4 provisions. Regional agree- of foreign direct investment," but that IPRs ments are often between countries with histor- can stimulate formal technology transfer ical ties, and former migrants (or their chil- through FDI and licensing. dren and grandchildren) tend to support All in all, the general conclusion is that coun- increased opportunities for citizens of their tries have to develop an IPR strategy appropri- home countries. Working on bilateral or ate to their level of development, and then ana- regional levels may provide receiving coun- lyze carefully which if any IPR provisions ought tries with greater control over the numbers to be contained in trade treaties or RTAs. and nationalities of temporary workers. That is, receiving countries that wish to ensure maximum control in immigration decisions RTAs and Provisions for may prefer to design their policies without Movement of Labor having to negotiate the terms with countries Using RTAs to promote movement outside the region. Given both labor market of unskilled workers and security concerns, this control may make Walmsley and Winters (2003) estimated that if it easier for countries to implement temporary a temporary visa system were introduced in programs for unskilled workers. Close neigh- those developed countries that permitted the bors tend to have a high proportion of immi- movement of up to three percent of their labor grants in receiving countries. To what extent force, world incomes would rise some $160 have RTAs facilitated labor movement? billion. Some 70 percent of the global welfare Regional agreements treat labor movement gains from increased migration would come in varying ways, ranging from full labor mo- from the movement of unskilled workers. To bility to no provisions at all. As shown in date, progress under the GATS Mode 4 nego- table 5.3, RTAs treat labor in one of four dif- tiations in easing restrictions on the temporary ferent ways:18 entry of workers has been limited; the agree- ment has generally been used for skilled · free labor mobility, with limited excep- workers, not unskilled. tions; Regional agreements might offer a more · temporary market access for certain promising venue for realizing the gains from (usually skilled) groups of workers; Table 5.3 Summary of agreements by degree of labor mobility Degree of labor mobility under agreement Agreements Full labor mobility European Union, Agreement on the European Economic Area, European Free Trade Association, Australia-New Zealand Closer Economic Relations Market access for certain groups Caribbean Community, North American Free Trade Agreement, Europe Agreements, Group of Three, and Canada-Chile, U.S.-Singapore, U.S.-Chile, Japan-Singapore Free Trade Agreements Based on GATS Mode 4, with ASEAN Free Trade Area, Euro-Med Association Agreements, New Zealand- additional provisions or limitations Singapore Closer Economic Partnership, Southern Common Market Agreement, and EU-Mexico, EU-Chile, MERCOSUR, and US-Jordan Free Trade Agreements No effective provisions for labor mobility Asia Pacific Economic Co-Operation Forum, South Asian Association for Regional Cooperation, Central European Free Trade Agreement, and Common Market for Eastern and Southern Africa Source: World Bank staff; Nofal 2004. 111 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 · temporary movement based on the three months. The European agreements in- GATS Mode 4 model, often with addi- clude exceptions for public services, public se- tional provisions or limitations;19 and curity, and/or public health works. · no provision in place for market access ANZCERTA provides for both full market (beyond facilitating entry visas), or plans access and national treatment for all service designated to be realized only in the suppliers, excluding a few sectors. future. The migration provisions of these agree- ments facilitated regional integration by mak- Most agreements do not override migration ing it easier for firms from one country to legislation, and parties retain broad discretion transfer personnel to their operations in other to grant, refuse, and administer residence per- countries in the region. The potential for mits and visas. It should be noted also that the workers to move to higher-paying jobs in right of labor mobility does not automatically other countries may have improved discipline entail the right to practice a certain profession; in some labor markets. However, the EU national regulations regarding licensing and agreements had almost no impact on stocks of recognition of qualifications are still permanent migration from other EU countries applied.20 (figure 5.1), in part because several EU coun- tries had already provided for free labor mo- RTAs with full labor mobility bility. In addition, most of these countries are The EU, the European Economic Area and of similar income levels, so the incentive to European Free Trade Association, and the disrupt family and personal relationships by Australia-New Zealand Closer Economic moving for higher incomes was limited. The Relations Trade Agreement (ANZCERTA) entry of less wealthy countries into the EU in allow for free labor mobility, with very limited the 1980s also did not result in greatly in- exceptions.21 The EU also allows the right to creased migration. reside (with family), although residence per- Several factors may explain this: among mits are not required for stays of less than these, the free movement of workers from the Figure 5.1 Share of EU nationals in total population, for selected EU countries and Norway, 1985­2000 Percent 1990 1995 6 1985 2000 5 4 3 2 1 0 k ia y ay eden Ireland Austr lands man Finland France Italy Spain rtugal Norw Denmar Sw Kingdom Belgium Ger Po Nether United Highest income Lowest income Source: OECD. 112 B E Y O N D M E R C H A N D I S E T R A D E new member states were subject to transi- of provisions that delay this migration for a tional periods; the EU's regional funds aimed limited period (renewable up to 7 years). Most at developing less prosperous regions com- studies find that the accession agreements are bined with the prospect of a positive economic unlikely to greatly boost migration to Western development in the new countries are likely to Europe. Forecasts of the additional migration have thwarted a significant migration of labor. due to expansion of the EU, which relies on Greece, Portugal, and Spain underwent both econometric models and opinion polls, 6 years of transition, which limited the free generally find that migration will be limited to movement of people after their entry to the about 3­4 percent of residents of the 10 first- EU. Even the richest EU countries (Denmark round East European nations within a decade and Sweden) experienced only a slight rise in of freedom of movement; this amounts to EU nationals as a share of population after the about 4 million people or 1 percent of the cur- transition periods of the poorest EU countries rent EU population. Additionally, about half of (Spain and Portugal) expired.22 these laborers are likely to return home within It remains to be seen whether the recent ac- the 10-year period, meaning a net migration of cession of Central and Eastern European coun- 2 million persons (Martin 2003). However, tries to the EU will result in substantial migra- some of the EU countries, particularly those tion. While this agreement will ultimately bordering the Eastern European countries like provide for full labor mobility, most of the Germany and Austria, could experience larger original EU countries have taken advantage flows relative to population. Box 5.3 Illegal migration: A growing global phenomenon T he number of migrant workers without proper workers. Thailand is one of the countries most af- work permits, so-called undocumented or illegal fected by undocumented workers, with nearly 5.5 workers, is increasing. The International Organiza- undocumented workers for each registered foreign tion of Migration estimates that each year, some- worker (IOM 2003). In 2000­01, some 73,000 where between 700,000 and 2 million people cross Chinese irregular migrants were assumed to be living borders to take jobs without legal permission.* in South Korea, and, in turn, China hosted some Of this number, up to 500,000 seek entry into 50,000 irregular migrants. Western Europe and 500,000 into the United States, Stricter enforcement of immigration rules may not Canada, Australia, and New Zealand. In the United have been successful in cutting off illegal migration, States, the stock of undocumented workers increased but it may have contributed to the very high costs from an estimated 3.5 million in 1990 to 7 million in involved, particularly between countries that do not 2000, of which 69 percent were Mexicans (USCIS have a common border. Today, irregular migrants 2003). In Europe, illegal migrants numbered as many from China wanting to enter the United States pay as 3 million by the end of the 1990s.* Countries in up to $35,000 to smugglers, to Europe they pay in Asia have also experienced an increased number of ir- the range of $10,000­$15,000, and those seeking regular foreign workers in their labor force. In Singa- entry into Japan up to $10,000. The "fee" for mov- pore, undocumented workers accounted for 4.5 per- ing from Lebanon to Germany varies between cent of the total labor force in 2000, and in Hong $5,000­$10,000; from India to the United States it is Kong (China), it was 2.3 percent. In Hong Kong around $25,000, and from North Africa to Spain it (China), Korea, and Taiwan, there is one undocu- is between $2,000­$3,500. mented foreign worker for every three legal foreign (Box continues on next page) 113 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Box 5.3 (continued) The vast majority of undocumented workers are Regularization programs for undocumented are unskilled for several reasons. Policies in most receiv- also present across developing countries. For ing countries offer greater opportunities for skilled example, in the 1990s, Argentina regularized undoc- workers to migrate legally than for unskilled. Many umented workers from Bolivia, Paraguay, and Peru skilled workers, coming from middle- or upper- under bilateral agreements. Thailand had a similar income households, may be less willing to experience program. Other countries have opted for more dras- the uncertainty inherent in breaking the immigration tic measures to address undocumented workers. For law (including the prospect of being jailed temporar- example, Malaysia in 2002 deported hundreds of ily), than persons from lower-income households. thousands of irregular migrants back to Indonesia Also, employers are more likely to offer unskilled and the Philippines, in reaction to rising levels jobs to undocumented workers, whose tenure is rela- of criminality; however, there was a slowdown in tively uncertain, as skilled jobs may require a greater its economy given its dependency on foreign investment to apply technical backgrounds to the workers. demands of specific jobs. Efforts to control illegal entry have varied in effectiveness. Receiving countries have addressed un- *IOM (2003). Unless otherwise noted, the data in this box documented workers by offering amnesties that gen- are based on this document. erally involve the promise of regularization for un- **The number may be even higher today. The United King- documented workers who have been in the country dom may have up to one million irregular migrants (UK Immi- for a period of time. For example, the U.S. Immigra- gration Service), France 500,000, Belgium 90,000 (Belgium Anti- tion and Regularization Control Act (IRCA) of 1986 racist Centre), and Ireland 10,000 (Irish Police). In terms of regularized the status of undocumented workers who flows, some 100,000 irregulars are smuggled into Germany each year (German Police Trade Union), and some 95,000 irregulars had been living in the country since before 1982, as from Albania, Romania, and Iraq alone enter Greece each year. well as undocumented agricultural workers. This pro- Annually, some 25,000 to 50,000 Chinese irregular mi- gram, which took effect mostly in 1989­91, granted a grants enter the United States. regular status to more than 2.5 million people, mostly Hanson and Spilimbergo (1999) find that additional re- from Mexico and Central America. However, this sources devoted to border enforcement in the United States have massive regularization program did not prevent Mex- yielded only limited results, and Boeri and others (2002) find that icans or other undocumented immigrants from con- tighter controls at a given entry point can be effective, but divert tinuing to cross the U.S. border. Similarly, in Europe, migrants to other points of entry. Worksite inspections can have approximately 1.5 million undocumented migrants a greater impact, but in the United States such efforts are very limited. saw their status regularized under amnesty programs § implemented by Belgium, France, Greece, Portugal, Italy regularized 716,000 irregular migrants in three waves; and Spain during the 1990s.§ Greece accepted 370,000 people (in 1997­98) mostly from the Balkans and Eastern Europe; Spain regularized 260,000 irregular immigrants mostly from Africa and Latin America; and Portugal Source: International Organization for Migration (IOM) 2003. regularized 61,000. Some countries have recognized the need to the movement of undocumented workers. manage both regular and irregular migration However, most of these groups are informal on a regional basis. Regional consultative ones that mostly share information on processes, such as the Manila Process (1996), migration-related issues and generally do not the Migration Dialogue for Southern Africa impose requirements on the immigration poli- (MIDSA 2000), the Puebla Process (1996), and cies of the participating countries. It appears others have emerged--and all of them include that neither unilateral policies nor consultative consultations to deter human trafficking and arrangements have had significant success in 114 B E Y O N D M E R C H A N D I S E T R A D E controlling the substantial pressures for managers and/or highly qualified employees, migration to industrial countries. and company representatives negotiating for the sale or supply of services. Transition peri- Agreements that permit temporary access ods as well as restrictions for public service for certain groups works and sectoral exclusions applied. They Several agreements provide for market access are still in force for the nonaccession coun- of certain groups of workers, usually the tries, Bulgaria and Romania. The Japan- highly skilled. The most recent agreements Singapore FTA provides for the temporary have focused heavily on intra-corporate trans- movement of natural persons for business pur- ferees, including managers and skilled techni- poses, including investors, subject to condi- cal staff. tions for entry (such as pre-employment for at The Central America and Caribbean Com- least one year for intracorporate transferees) munity (CARICOM) allows university gradu- and time limits of stay. The Group of Three ates to move among member countries (Colombia, Mexico, and Republica Bolivari- without passport requirements and allows ana de Venezuela) facilitates temporary entry university graduates, professionals, skilled for business persons. As in other agreements, persons, and workers from some selected oc- GATS Mode 4 restrictions regarding access cupations to work without a permit. to the employment market or permanent em- NAFTA, along with the Canada-Chile, ployment also apply. MERCOSUR also has U.S.-Chile, and the U.S.-Singapore Free Trade provisions to facilitate the temporary entry of Agreements, do not provide for permanent business persons and the exercise of temporary migration, but allow for the temporary move- professional practices, and it is working on the ment of business visitors, traders and in- issues relative to a "MERCOSUR National vestors, intra-corporate transferees, and pro- Residency" and a more flexible migratory reg- fessionals.23 Visas are still required for all four ulation for MERCOSUR citizens (Nofal categories, and work permits are required for 2004). Other agreements among Latin Ameri- all except business visitors.24 Numeric restric- can countries (such as the Mexico-Nicaragua, tions on temporary entry are not allowed for Mexico-Chile, Mexico-Bolivia, Mexico-Costa the first three categories, but were imposed by Rica, and the agreement between Central the United States on professionals from Chile, America and Dominican Republic) contain Mexico, and Singapore, but not from Canada. similar provisions. U.S. professionals seeking to enter either mar- ket are not subject to numerical limits. In the Agreements based on GATS Mode 4 Canada-Chile agreement there are no numeri- Several bilateral agreements, along with the cal limits to any of the four categories. Fol- ASEAN Free Trade Agreement (AFTA) and lowing the NAFTA agreement, the number of MERCOSUR, base labor mobility provisions professionals entering the United States from on the principles of Mode 4 of the GATS. How- Canada and Mexico increased substantially, ever, most of these agreements add provisions and although the gap in favor of Canadians that allow for labor mobility in categories not remained high, it narrowed considerably by covered by the GATS. For example, the U.S.- 2002. Similarly, in the treaty traders and in- Jordan agreement facilitates visa arrangements vestors category, the ratio of Mexican over for independent traders and persons linked to Canadian admittances into the United States investment. The EU-Mexico agreement pro- increased from 0.4 in 1996 to 1.1 in 2002.25 vides for a standstill and sets common regula- The EU agreements with Central and East- tions of work, labor conditions, and residency ern European countries (prior to their 2004 permits for temporary workers in each accession to the EU) allowed for temporary country. The AFTA since 1998 has covered all entry of workers providing a service, modes of supply, including services sectors not 115 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 previously covered by the GATS. The Frame- natural persons as services suppliers (Nofal work Agreement on the ASEAN Investment 2004). Area (1998) commits members to support freer flows of skilled labor, professionals, capital Agreements without effective provisions and technology among ASEAN members. By for market access contrast, the Euro-Med agreements with The Asia Pacific Economic Cooperation Morocco and Tunisia do not provide for pref- Forum (APEC), the South Asia Association for erential access beyond GATS (the case for other Regional Cooperation (SAARC), and Com- Mediterranean agreements as well). The New mon Market for Eastern and Southern Africa Zealand-Singapore agreement generally fol- (COMESA) do not provide for labor mobility. lows the GATS model regarding labor mobility However, the first two have taken measures to for service suppliers. MERCOSUR (and its facilitate business travel. Most of the members agreements with Bolivia and Chile) is limited to of APEC (except the United States and the GATS provisions on the movement of Canada) participate in the Business Travel Box 5.4 U.S. temporary admission programs under NAFTA and unilateral policies T he U.S. has several visa categories for the admis- 1996 to about 66,000 in 2002. Moreover, the largest sion of temporary workers. NAFTA provided temporary business visitor program (B1 visa) is lim- for the temporary admission of professionals under ited to six months (renewable, but total time cannot TN visas and made it easier for Mexicans to take exceed one year), and generally is used for short- advantage of already existing visa categories for term trips to the United States. The other visas can skilled workers and professionals. However, NAFTA be granted for one year or more, often with exten- made no provision for increasing admissions of un- sions. Thus more long-term unskilled Mexican work- skilled workers to the United States. In contrast, the ers have been admitted to the United States under United States has unilateral programs that do allow unilateral programs than under programs initiated for the temporary admission of unskilled workers, under, or supported by, NAFTA. and these have grown from from about 14,000 in Temporary entry of Mexican workers to the United States (thousands of persons) 1996 2002 Program initiated under NAFTA Professional (TN visa) 0.2 1.8 Programs supported by NAFTAa Intracompany transferees (L1 visas) 4.8 15.3 Treaty traders and investors (E1, E2 visas) 1.0 4.0 Business visitors (B1 visas) 309.0 475.0 Unilateral programs Professionals with specialty occupations (H1B visas) 5.3 15.9 Agricultural temporary workers (H2A visas) 8.8 12.8 Non-agricultural temporary workers (H2B visas) 5.5 53.0 aThese visa categories existed before NAFTA and apply to many countries. However, provisions of NAFTA make it easier for Mexicans to use them. Source: USCIS 2002, INS 1997. 116 B E Y O N D M E R C H A N D I S E T R A D E Card Scheme, under which holders of the card parallel programs, many of which predated receive expedited entry at the airport, and are the RTA, such as programs in the United not required to submit separate applications States with Mexico and in the EU with non- for business visas. Members of the SAARC members of Southern Europe. Where indus- adopted a Visa Waiver Scheme in 1992, which trial countries have allowed for some admit- exempts 21 categories of persons from visa re- tance of unskilled workers, the rules differ quirements. Free labor movement is envisaged significantly from those governing admission as a long-term objective of COMESA, to be of skilled workers. Unskilled workers legally accomplished by 2025. However, progress has entering host countries usually do so under been limited to date.26 seasonal work agreements, project (or guest) worker agreements, or specific provisions To date: Limited labor integration such as the working holiday maker programs Most regional agreements have had little im- (WHMP). Seasonal employment usually al- pact on increasing migration. First, the agree- lows foreign workers to stay in the host coun- ments with full labor mobility have been be- try for periods between three months and one tween countries of similar income levels, so year, with work permits provided to foreign there has been little incentive for migration. workers only in the event that no domestic Most agreements, and particularly those labor can do the job and only in some specific agreements that include both industrial and sectors (such as agriculture, forestry, and developing country members, do not allow for tourism). Project worker agreements allow permanent migration. foreign workers entry for specific projects and Second, while these agreements often pro- usually include limits on the maximum stay vide for some temporary labor mobility, par- and quotas (often determined on the basis of ticularly for the service sectors, the provisions labor market conditions). WHMP allows are generally restricted to higher-skilled work- young people (roughly 18­30 years old) to ers. This is consistent with the trend in indus- holiday and work for short periods, provided trial countries of changing unilateral migra- there is a prior bilateral working holiday tion policies to attract higher-skilled agreement among the involved countries. workers--in principle on a temporary basis, Developing countries' regional agreements but in practice allowing for the possibility to also tend to discriminate in favor of skilled settle after some period of time. For example, workers when providing for labor mobility. As the Temporary Immigration Program of Aus- mentioned above, CARICOM allows for the tralia allows for unlimited visa renewal for free movement of university graduates, other skilled workers. In the United Kingdom, professionals, skilled persons, and workers highly skilled temporary workers in certain from selected occupations. Several Latin occupations can settle after 4 years of contin- American agreements (Group of Three, Cen- uous work. In Norway, temporary workers tral America-Dominican Republic, and agree- with special skills can be issued a permanent ments between Mexico on the one hand, and work permit after 3 years of stay. In Canada, Chile, Nicaragua, Bolivia, and Costa Rica on the 2002 Immigration and Refugee Protection the other) also provide for the movement of Act (IRPA) placed more emphasis on educa- some skilled, not unskilled, workers. tion, job experience, and language ability in allowing entry. This trend toward favoring higher-skilled workers limits the potential Conclusions: Beyond Merchandise global gains from migration. Trade A To the extent that developed countries have greements differ markedly in their treat- permitted entry for unskilled workers from ment--to say nothing of their implementa- RTA countries, it has occurred through tion--of non-merchandise provisions for 117 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 services, investment, intellectual property, and the erosion of monopolistic returns to the temporary movement of workers. Differences owners of technology through enhanced IPRs between the United States and EU bilateral FTAs is of doubtful development benefit for the av- and other agreements are particularly vast. erage developing country. In those areas where RTAs could seriously Moreover, the downside risks of misjudg- promote development--services liberalization ments in terms of adverse legal and economic and temporary movement of workers--results ramifications are nontrivial, especially for have ranged from mixed to missed opportuni- unsophisticated governments. International ties. In general, the U.S. FTAs have prompted treaty law in these areas is evolving fast and some additional services market openings in is being set through case laws of arbitration Bahrain, CAFTA, and Morocco, some changes panels, whose judgments at times conflict in Singapore, but relatively few changes in (Ewing-Chow 2001). Governments may find Australia and Chile. To be sure, even those themselves hauled before arbitration panels agreements that required no additional mar- and compelled to pay large amounts of com- ket opening could benefit developing coun- pensation for enacting regulations they had tries, because investors may attach credibility considered in their sovereign domain. to the lock-in effects of the treaty. Broadly defining investment to include all In the South, agreements that have sub- capital flows and assets, including intellec- stantially improved services access are rela- tual property, carries risks. For example, one tively limited, and those with greater market potential risk, is the set of provisions associ- access often have the most restrictive rules ated with the U.S.-Chile arrangement. These of origin for investor nationality. More provisions could limit a less sophisticated common is the lack of progress on services government's ability to deal with a financial liberalization. crisis. More disappointing from a development perspective is the minimal attention given to Realizing the promise of services, creating opportunities for the temporary investment, IPRs, and labor mobility movement of workers, particularly unskilled Using RTAs as a lever to liberalize services has workers. In neither North-South nor South- several advantages. The potential for improv- South agreements is there evidence of much ing economy-wide performance is often great. activity. In the wake of the September 11, Moreover, most services liberalization is in- 2001, attacks on the United States, concerns herently multilateral. This is because member for security have made cross-border move- governments that open markets want to real- ment of all persons subject to greater controls ize the immediate gain of full competition and and scrutiny. This atmosphere does not bode so open markets to all comers, or because well for expanding programs for temporary members adopt lenient rules regarding domi- workers. cile of investors, thereby permitting external At the same time, in those regulatory areas investors to invest in the region through sub- where development benefits are largely un- sidiaries located in member countries. Even if proven, the North-South bilateral FTAs are restrictions impede full MFN access, regional strengthening their rules. Strengthening the agreements can be cost-free stepping stones rules governing investment and intellectual to open markets insofar as they do not carry property may contribute to better institutional the trade-diversion costs of lost revenues asso- environment, but the greatest gain is to be ciated with tariff reductions for goods trade. found in services. Enhancing protections of- More could be done in RTAs, particularly fered to investors has not been shown to in- in South-South agreements, to realize these crease the flow of investment, and preventing benefits. 118 B E Y O N D M E R C H A N D I S E T R A D E Countries have to design strategies toward might adopt common standards that facilitate investment and intellectual property that are cross-border competition in services and in- appropriate to their development priorities vestment. Adopting common technical stan- and then analyze carefully which elements, if dards for telecommunications, for example, any, ought to be contained in trade treaties. has helped integrate markets and opens the For some countries, improving other property way for competition. To adopt international rights--for example, land rights or small busi- norms regulating technology, new agreements ness assets--may have a higher priority than could adopt their own IPR standards, with the establishing market access rights for investors advantage of foreclosing additional restric- or rights for patent holders; for other coun- tions motivated by private interest groups in tries, especially middle-income countries, the North. At the same time, some regions these may indeed be a priority. Once a strategy may find opportunities to agree on common is in place, it is possible to ascertain which of administration of patent and copyright law. the new rules customarily associated with regional trade agreements make sense for Notes development. 1. Agreements can also increase the number of Evaluating the development benefit of any competitors by allowing cross-border provision of given RTA cannot rest solely on one compo- services, Mode 1, such as supplying back-office ser- nent because often rules are accepted in one vices. Restrictions on these tend to be less common area to achieve market openings in partner than on supplying through commercial presence of a countries in another. For example, the net de- foreign subsidiary, Mode 3. One exception is supplying velopment benefits of the investment and IPR various types of insurance and reinsurance, which can be done cross-border, though a sales affiliate is nor- components hinge critically on their appropri- mally required. ateness for development and the market access 2. RTAs that restrict service providers to member granted in product markets of partner countries by definition limit the number of potential countries. competitors relative to multilateral liberalization. Even A corollary lesson is that multilateral liber- if later followed with multilateral opening, RTAs may alization in goods markets is essential for confer on members first mover advantages, and, if reaping any gains from RTAs that contain new some market barriers to entry remain, they will not be readily competed away; the result is higher prices to rules. Governments wishing to maximize consumers. gains from RTAs should ensure that new rules 3. According to the agreement, if Chile chooses to are consistent with extant border protection, impose restrictive measures on capital flows it consid- and if not, they should consider lowering that ers speculative, then special dispute settlement rules protection. If a country has high tariff barriers will apply. and forms an RTA with a partner requiring in- 4. For some FTA countries, pre-existing BITs had vestor protections and TRIPS-Plus, it may end virtually the same rights; however, about half of the countries had no BIT prior to the FTA, and where BITS up entrenching the businesses of that partner did exist, pre-establishments were often fewer and less behind new barriers. It may confer first mover extensive. advantages in services and IPR restrictions in 5. See Fink and Reichenmiller (2004) for a more drug or other high technology manufacture, detailed review. with reinforced barriers to parallel imports or 6. These include, for example, TRIPS, the Paris new entry from third parties. Convention for the Protection of Industrial Property, Finally, RTAs--and other regional cooper- the Berne Convention for the Protection of Literary and Artistic Works, the Rome Convention for the Pro- ation agreements--do offer some important tection of Performers, Producers of Phonograms and opportunities for countries to collaborate, es- Broadcasting Organizations, and others also contains a pecially in South-South agreements. To stimu- list of new conventions that parties are expected to rat- late investment, particularly in services, they ify within a specified timeline, such as the Budapest 119 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Treaty on the International Recognition of the Deposit the host country itself. Thus if we consider Brazil as the of Micro-organisms for Mexico and the World Intel- host country and MERCOSUR as the relevant RTA, lectual Property Organization (WIPO) Copyright the variable lftagdp would be the log of the sum of Treaty for Chile. Finally, the list includes several con- GDP of Argentina, Paraguay, and Uruguay. This vari- ventions that the member states are expected to ratify able serves a twofold purpose in the estimation routine. "as soon as possible" without mentioning a specific First, since it is equal to zero prior to signing an RTA deadline, such as the WIPO Copyright Treaty for and carries a positive value afterwards, it measures Mexico or the Madrid Agreement concerning the In- whether signing an agreement has an effect on FDI in- ternational Registration of Marks for Chile. flows (i.e., including a dummy variable for RTA mem- 7. BITs customarily provide a definition of invest- bership would be counterproductive in the presence of ment coverage, provide investor protections such as this variable, since it will capture the "threshold effect" against expropriation, require national treatment for of signing an RTA). Furthermore, this variable also post-entry establishments, stipulate compensation captures the effects of participating in a larger market for the expropriation of their investments, and provide following the signing of an agreement. This is particu- for a dispute resolution mechanism. The latter usually larly important if a country is party to more than one permit the investor to sue the state for breach of treaty agreement--the variable will then be a sum of all of its under binding arbitration. In some cases, treaties pro- partners' GDP, reflecting the fact that the country has scribe any government action that would reduce the now created a larger market. The fact that this variable value of the private investment, even if it were envi- is positive and significant shows not only that signing ronmental, and establish grounds for compensation. an RTA will generally bring benefits in terms of greater Such compensation could either entail extensive liabil- FDI inflows, but also that larger market size of the ities for the host government or compel them to refrain country's partners tends to generate more incoming from making certain policy choices. FDI. 8. The United States initiated the practice of defin- 12. See Finger and Schuler (2004) for a richer dis- ing investment to broadly include intellectual property cussion of the the asymmetry in TRIPS; they argue that in the late 1990s in negotiations of its bilateral invest- it protects the knowledge that businesses and individu- ment treaties. See Vivas-Eugui (2003). als have in rich countries and that poor people buy, but 9. Some of the problems include the absence of not the knowledge that poor people generate and sell data on implementation and the variable coverage of to the world. FDI provisions across agreements, which makes it dif- 13. In this sense, full enforcement is equivalent to a ficult to distinguish the effects of investment rules from TRIPs standard (Maskus 2000, 183­85). trade rules. Moreover, the absence on FDI data that 14. See paragraph 4 of the Doha Declaration on would enable us to distinguish the effects of RTAs on TRIPS and Public Health, available at http://www. differing type of investment--vertical or horizontal­ wto.org. limits the analysis. Nonetheless, the regressions are 15. Technically, the Doha Declaration does not robust to variations in specifications. address questions of marketing approval during the 10. The regression with fixed effects estimation of patent term and test data exclusivity. However, the pro- net FDI inflows is: visions of the bilateral FTAs in these areas can still be lfdi | Coef. Std. Err. t P |t| seen as being at odds with the spirit of the Doha Dec- lgdp | .9404982 .2065772 4.55 0.000 laration, to the extent that they preclude the effective lgnppc | .1228465 .2008249 0.61 0.541 use of compulsory licenses. open | .0051226 .0011387 4.50 0.000 16. See the letter from USTR General Counsel growth | .0198651 .0040816 4.87 0.000 John K. Veroncau to Congressman Levin dated cpi | .0196485 .0060906 3.23 0.001 July 19, 2004, available at Inside US Trade. Moreover, lfdiwld | .4472645 .0719058 6.22 0.000 the side letters refer "in particular [to] cases such as growld | .0611576 .0432152 1.42 0.157 HIV/AIDS, tuberculosis, malaria, and other epidemics lftagdp | .0518633 .0163279 3.18 0.002 as well as circumstances of extreme urgency on na- R-sq: within 0.3973 corr(u_i, Xb) tional emergency." The language chosen mirrors word- -0.0410 ing in the Doha Declaration on TRIPS and Public between 0.7469 F(28,2003) 47.16 Health, except that the latter employs the term "espe- overall 0.6690 Prob F 0.0000 cially" instead of "in particular" and does not mention F test that all u_i 0: F(143, 2003) 12.79 "circumstances of extreme urgency or national emer- Prob F 0.0000 gency." This difference in language may imply a nar- 11. As mentioned earlier, this variable contains the rower definition of public health and therefore flexibil- sum of the host country's RTA partners GDP, excluding ity to issue a compulsory license, but this is subject to 120 B E Y O N D M E R C H A N D I S E T R A D E legal interpretation that must await actual cases of dis- 26. Mexicans entering through this category into pute settlement. the United States increased to 4,000 business persons 17. See the letter from USTR General Counsel in 2002, from 980 in 1996. John K. Veroneau to Congressman Levin dated 27. Nonetheless, migration flows within COMESA July 19, 2004. have been substantial. The annual inflow of workers to 18. See the excellent studies in Fink and Maskus South Africa from other African countries is estimated (2004) including particularly Fink (2004) for the U.S. to have increased from roughly 500,000 in 1990 to and German MNCs; Smarzynska Javorcik (2004) for more than 3.5 millions in 1995; most were temporary Eastern Europe; and Maskus, Dougherty, and Mertha workers in mining and farming. (2004) for China. 19. The description on agreements that follows is strongly based on Nielson (2003) and a complement, in some cases, to the text of the agreements. References 20. GATS (under "Mode 4") covers the movement Abbott, Frederick M. 2004a. The Doha Declaration on of some temporary foreign workers among WTO the TRIPS Agreement and Public Health and the members, specifically individual service suppliers. Al- Contradictory Trend in Bilateral and Regional though in theory GATS covers services suppliers at all Free Trade Agreements. Quaker United Nations skill levels, in practice WTO members' commitments Office Occasional Paper 14. have been limited to the higher skilled. . 2004b. Towards a New Era of Objective 21. Some agreements include provisions facilitat- Assessment in the Field of TRIPS and Variable ing mutual recognition (e.g., EFTA), and others have Geometry for the Preservation of Multilateralism. complementary arrangements (e.g., the ANZCERTA Paper for The World Trade Forum 2004, Bern, Services Protocol, the Trans-Tasman Travel Arrange- Switzerland, June 4­5, 2004. (also forthcoming in ment, and the Trans Tasman Mutual Recognition Journal of International Economic Law No. 1 Arrangement together provide that persons registered 2005 Oxford.) to practice an occupation in one country can practice Bora, Bijit. 2003. The Agreement on Trade Related an equivalent profession in the other country). Investment Measures, 1995­2002. UNCTAD 22. ANZCERTA does not cover labor mobility Discussion Paper Series No. 15, Geneva. other than for services suppliers, but under the Trans- Boeri, Tito, Gordon Hanson, and Barry McCormick, Tasman Travel Arrangement (signed in 1973), nation- eds. 2002. Immigration policy and the welfare als of each country have the right to visit, reside, and system. New York: Oxford University Press. work in each other's country without time restrictions. Chudnovsky, Daniel, and Andres Lopez. 2001. La 23. Between 1987­97, there was a total increase of Inversión Extranjera Directa en el MERCOSUR: 102,000 Greeks in the rest of the (11) European coun- Un Análisis Comparativo. In El Boom de Inver- tries, which is an annual average of 10,000 only. In sión Extranjera Directa en el MERCOSUR, Siglo the case of Portugal, the annual average increase of XXI, Buenos Aires 2001. Portuguese immigrants in the rest of the EU countries Commission on Intellectual Property Rights (CIPR). during 1986­97 was only 7,700. 2002. Integrating Intellectual Property Rights and 24. In NAFTA, definitions for business persons Development Policy. Report of the Commission and, in particular, a list of professionals are provided, on Intellectual Property Rights established by as well as the minimum academic conditions that the the UK Secretary of State for International latter must satisfy (in general, at least a baccalaureate Development. degree, and sometimes complemented with some Encarnation, Dennis, and Louis T. Wells. 1986. Evalu- years of experience). The other agreements provide ating Foreign Investment. In Investing in Devel- for a more flexible requirement, in that instead of opment: New Roles for Private Capital? ed. listing the professions, they specify the academic Theodore Moran. New Brunswick, NJ: Transac- and/or experience requirements that professionals tion Books. must satisfy. Ewing-Chow, Michael. 2001. Investor Protection in 25. Business visitors need to demonstrate though Free Trade Agreements: Lessons from North that the proposed activity is international in scope and America. Singapore Journal of International and that they are not seeking to enter the local labor mar- Comparative Law 748. ket. They need also to demonstrate that the primary Finger, Michael, and Julio Nogues. 2002. The Unbal- source of remuneration, their principal place of busi- anced Uruguay Round Outcome: The New Areas ness, and actual place of accrual of profits are from in Future WTO Negotiations. The World outside the territory they are seeking to enter. Economy 25. 121 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Finger, Michael J., and Philip Schuler. 2004. Poor Peo- Mann, Howard, and Aaron Cosbey. 2004. International ple's Knowledge: Promoting Intellectual Property Investment Agreements: Trends and Impacts for Rights in Developing Countries. Washington, Developing Countries Background paper for DC: World Bank. Global Economic Prospects 2005, World Bank, Fink, Carsten. 2004. Intellectual Property Rights and Washington, DC. U.S. and German International Transactions in Mansfield, Edwin. 1995. Intellectual Property Protec- Manufacturing Industries. In Intellectual Property tion, Direct Investment and Technology Transfer. and Development: Lessons from Recent Research, IFC Discussion Paper No. 27, World Bank, eds. Fink, Carsten and Keith Maskus. Washington, Washington, DC. DC: World Bank, forthcoming. Markusen, James. 2004. Regional Integration and Fink, Carsten, and Keith Maskus, eds. 2004. Intellec- Third-Country Inward Investment. Business and tual Property and Development: Lessons from Politics 6 (1). Recent Research. Washington, DC: World Bank, Martin, Philip. 2003. Economic Integration and forthcoming. Migration: The Mexico-US Case. Wider Discus- Fink, Carsten, and Patrick Reichenmiller. 2004. sion Paper No. 2003/35. United Nations Univer- Tightening TRIPS: the Intellectual Property Provi- sity World Institute for Development Economics sions of Recent US Free Trade Agreements. Research. Helsinki. Report. World Bank. Maskus, Keith. 2000. Intellectual Property Rights in Fink, Carsten, and Aaditya Mattoo. 2002. Regional the Global Economy. Washington, DC: Institute Agreements and Trade in Services. Policy Re- for International Economics. search Working Paper 2852, World Bank, Maskus Keith, Sean Dougherty, and Andrew Mertha. Washington, DC. 2004. Intellectual Property Rights and Economic Frischtak, Claudio. 2004. Multinational Firms' Development in China. In Intellectual Property Responses to Integration of Latin American and Development: Lessons from Recent Markets. Business and Politics 6:1. Research, eds. Fink, Carsten, and Keith Maskus. Hallward-Driemeier, Mary. 2003. Do Bilateral Invest- Washington, DC: World Bank, forthcoming. ment Treaties Attract Foreign Direct Investment? Mattoo, Aadittya, R. Rathindran, and A. Subramanian. Only a Bit . . . and They Could Bite. Policy 2001. Measuring Services Trade Liberalization and Research Working Paper 3121, World Bank, its Impact on Economic Growth: An Illustration. Washington, DC. Policy Research Working Paper 2655, World Hanson, Gordon H., and Antonio Spilimbergo. 1999. Bank, Washington, DC. Illegal Immigration, Border Enforcement and Mattoo, Aadittya, and Pierre Sauve. 2004. Regional- Relative Wages: Evidence for Apprehensions at ism and Trade in Services in the Western Hemi- the U.S.-Mexican Border. American Economic sphere: A Policy Agenda. In Integrating the Review 89 (5): 1337­57. Americas, eds. A. Estevadeodal, Dani Rodrik, Immigration and Naturalization Service (INS). 1997. Alan Taylor, and Andres Velasco. Cambridge, Statistical Yearbook of the Immigration and MA: Harvard University Press. Naturalization Service. Table 40. http://uscis.gov/ Newfarmer, Richard. 2001. Foreign Direct Investment: graphics/shared/aboutus/statistics/Nim97list.htm. Policies and Institutions for Growth. In New International Organization for Migration (IOM). Horizons for Foreign Direct Investment. Paris: 2003. World Migration 2003: Managing Migra- OECD. tion Challenges and Responses for People on the Nielson, Julia. 2003. Labor Mobility in Regional Trade Move. Geneva. Agreement. In Moving People to Deliver Services, Lall, Sanjaya, and Paul Streeten. 1977. Foreign Invest- eds. Mattoo, Aaditya, and Antonia Carzaniga. ment, Transnationals, and Developing Countries. Washington, DC: World Bank. Boulder, Colorado: Westview Press. Nofal, Beatriz. 2004. Construyendo una intergracion Lederman, Daniel, William Maloney, and Luis Serven. mas profunda en Mercosur. Background paper 2004. Lessons from NAFTA for Latin American for the Global Economic Prospects 2005, World and Caribbean Countries: A Summary of Re- Bank, Washington, DC. search Findings. World Bank, Washington, DC. Nogues, Julio. 2004. The Roles of Multilateralism and Levy Yeyati, Eduardo, Ernesto Stein, and Christian Regionalism for Latin America. Paper prepared Daude. 2004. The FTAA and the Location of FDI. for the Economic Commission for Latin America Paper for Central Bank of Chile and World Bank (ECLA). Santiago, Chile. Conference on The Future of Trade Liberalization OECD. Trends in International Migration, various in the Americas, Santiago, Chile, March 22­23. years. 122 B E Y O N D M E R C H A N D I S E T R A D E OECD. 2003. Regionalism and the Multilateral Ullrich, Heidi. 2004. Comparing EU Free Trade Agree- Trading System. Paris: OECD. ments: Services. ECDPM In Brief. Ozden, Caglar, and Francisco J. Parodi. 2004. Custom UNCTAD (United Nations Conference for Trade and Unions and Foreign Investment: Theory and Evi- Development). 1998. Bilateral Investment Treaties dence from Mercosur's Auto Industry. Paper for in the Mid-1990s. New York: United Nations. Central Bank of Chile and World Bank US Citizens and Immigration Services (USCIS, formerly Conference on The Future of Trade Liberalization the INS). 2002. Yearbook of Immigration Statis- in the Americas, Santiago, Chile, March 22­23. tics (formerly Statistical Yearbook of the Immi- Peterson, Luke. 2003a. Emerging Bilateral Investment gration and Naturalization Service, INS). Table 27. Treaty Arbitration and Sustainable Development http://uscis.gov/graphics/shared/aboutus/statistics/ Research Note, Invest-SD News Bulletin, Interna- TEMP02yrbk/Temp2002list.htm. tional Institute for Sustainable Development US Citizens and Immigration Services (USCIS, formerly (IISD). April. http://www.iisd.org/pdf/2003/ the INS). 2003. Estimates of the Unauthorized trade_bits_disputes.pdf. Immigrant Population Residing in the United . 2003b. Czech Republic Hit With Massive States: 1990 to 2000. Office of Policy and Plan- Compensation Bill in Investment Treaty Dispute. ning. Table B. http://uscis.gov/graphics/shared/ Invest-SD News Bulletin, International Institute aboutus/statistics/Illegals.htm. for Sustainable Development (IISD). March 21. USTR (United States Trade Representative). 2004. Schiff, Maurice, and Alan Winters. 2003. Regional USA--Chile Free Trade Agreement. Integration and Development. Washington, DC: Vivas-Eugui, David 2003. Regional and bilateral agree- World Bank. ments and a TRIPS-plus world: the Free Trade Smarzynska Javorcik, Beata. 2004. The Composition Area of the Americas (FTAA). TRIPS Issue of Foreign Direct Investment and Protection of Papers 1. Geneva: Quaker United Nations Office. Intellectual Property Rights: Evidence from Waldkirch, A. 2001. The `New Regionalism' and Transition Economies. In Intellectual Property Foreign Direct Investment: The Case of Mexico. and Development: Lessons from Recent Re- Oregon State University Working Paper. search, eds. Fink, Carsten, and Keith Maskus. Walmsley, Terri Louise, and L. Alan Winters. 2003. Re- Washington, DC: World Bank, forthcoming. laxing the Restrictions on the Temporary Move- Stein, Ernesto, and Christian Daude. 2001. Institu- ment of Natural Persons: A Simulation Analysis. tions, Integration, and the Location of Foreign Centre for Economic Policy Research Discussion Direct Investment. In New Horizons for Foreign Paper Series No. 3719. January. Direct Investment. Paris: OECD. Weintraub, Sidney. 2004. Lessons from the Chile and Szepesi, Stefan. 2004a. Coercion or Engagement? Singapore Free Trade Agreements. In Free Trade Economics and Institutions in ACP-EU Trade Agreements: US Strategies and Priorities, ed. J. J. Negotiations. European Center for Development Schott. Washington, DC: Institute for Interna- Policy Management (ECDPM) Discussion tional Economics. Paper 56. Maastricht, Netherlands. World Bank. 2001. Global Economic Prospects and Szepesi, Stefan. 2004b. How Did David Prepare to Talk the Developing Countries 2002: Making Trade to Goliath? South Africa's experience of trade Work for the World's Poor. Washington, DC: negotiations with the EU. European Center for World Bank. Development Policy Management (ECDPM) . 2002. Global Economic Prospects and the Discussion Paper 53. Maastricht, Netherlands. Developing Countries 2003: Investing to Unlock te Velde, Dirk Willem, and Miatta Fahnbulleh. 2003. Global Opportunities. Washington, DC: World Investment Related Provisions in Regional Trade Bank. Agreements. Paper prepared for the project on Re- . 2004. World Development Report 2005: gional Integration and Poverty. UK Department A Better Investment Climate for Everyone. for International Development (DFID). London, Washington, DC: World Bank. October. 123 6 Making Regionalism Complementary to Multilateralism The emergence of a more proactive stance in disappear if all developing countries were to multilateral negotiations by developing coun- sign such agreements. In fact, all developing tries in parallel with an explosion of preferential countries would lose relative to a multilateral regional trade agreements (RTAs) around the agreement and even relative to the baseline. world raises an important question: Are these This scenario underscores the fact that, while trends complementary, representing different stalled collective action through multilateral paths to the same desired outcome of faster channels creates a strong incentive for each growth, development, and poverty reduction, developing country to sign a regional agree- or are they competing and incompatible? In ment, if every country does so, they all lose. principle, the best outcome for all countries RTAs do alter the incentives for countries would be a nondiscriminatory trading system; to participate in multilateral liberalization. developing countries, in particular, would ben- RTAs can be a stumbling block to multilateral efit from a nondiscriminatory trading system arrangements by creating incentives to resist because most poor people and many poor coun- the preference erosion that can occur through tries might find themselves excluded from pref- new multilateral liberalization. However, erential deals. If the explosion in RTAs implies because the gains are often substantially larger a higher probability that the majority of in multilateral agreements, concerns over pref- developing countries would face greater dis- erence erosion may be limited to a few small crimination than under an MFN (or nondis- countries that could conceivably block a mul- criminatory) regime, the world as a whole will tilateral agreement. Those recalcitrant coun- be worse off, and individual developing coun- tries resisting reforms are likely the beneficia- tries may lose substantially.1 ries of preferences associated with distorted RTAs can be a complement to multilateral agreements encompassing agriculture or the reform, but they are not a substitute. Consider clothing and textile trade, not RTAs. Large de- a scenario, modeled in the first section of this veloped countries may gain more from signing chapter, in which each developing country individual bilateral agreements than they signed an agreement with the United States, would from a multilateral accord, because the European Union (EU), Canada, and they can use the carrot of preferential access to Japan--the Quad. Each developing country extract concessions in nontrade areas from could raise its real incomes by individually developing country partners that would be re- signing bilateral agreements with the Quad sisted in the WTO negotiating framework. But countries; and in some cases they would raise we see little evidence that the high-income their real incomes more than they would countries have reduced their effort to bring the through multilateral accords. But these gains current multilateral negotiations to fruition. 125 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 From a development perspective, the WTO partners to ensure that new regulations re- remains the best-available forum to discipline garding investment and intellectual property the use of trade-distorting policies. RTAs can are appropriate to the level of development, complement the WTO efforts by cooperating which would reduce risks of undue enforce- on behind-the-border policies, especially on ment costs. Finally, they could provide trade- regulation-intensive issues such as services, related technical assistance not only in the im- trade facilitation, and the investment climate. plementation phase but also in the negotiating Governments pursuing this agenda through phase, which could promote greater liberaliza- RTAs must adopt rules that are appropriate to tion and supply response to new market their own level of development. To be effec- opportunities in regional or global markets. tive, the rules must target a priority concern, The international community, working must be enforceable, and must avoid becom- through the WTO, can reduce discrimination in ing a tax on scarce resources that would be the system. The procedures associated with better used elsewhere. Getting the rules RTA disciplines as currently configured are ill- "right" so that they promote development suited to limit either their proliferation or to has implications for negotiations and enforce- control their discriminatory provisions. WTO ment of the resulting disciplines. The potential members should establish stronger multilateral for inappropriate outcomes is higher in surveillance mechanisms to document, analyze, North­South RTAs because the asymmetry in and monitor the effects of RTAs on nonmem- negotiating power can overtake real develop- bers. Expanding the information on the impact ment priorities. of RTAs to stakeholders--firms, consumers, Reinforcing the complementarity between taxpayers--would also help ensure that the po- regionalism and multilateralism and minimiz- tential benefits of liberalization are both real- ing the latent tensions must begin with the ized and distributed more equitably. Medium- completion of the Doha Development Agenda. term efforts should focus on implementing If the Doha Development Agenda is completed WTO disciplines on regional agreements. in a way that actually promotes development, it would bring down tariffs, enhance the gains from open regionalism, and discipline any ex- Preferential Agreements within clusionary effects of regional accords. the Global Context T All countries could take steps to promote o place preferential trade arrangements in open regionalism--the developing countries, a multilateral context, we evaluate how high-income countries, and the international different collections of trade agreements com- community working together through the pare with multilateral alternatives. For this WTO. Developing countries are likely to have evaluation, we utilize the World Bank's global the greatest success in harnessing trade for trade model known as LINKAGE, which has growth and poverty reduction if they adopt been used in previous Global Economic a three-pronged strategy that involves Prospects. The model is built around the autonomous liberalization, active multilateral- GTAP database that is widely used to assess ism, and open regionalism. the global, regional, and country implications High-income countries could promote of alternative trade liberalization scenarios. open regionalism by including agriculture in However, the results described in this year's RTAs. They could adopt more common and report reflect an update of the model's data- nonrestrictive rules of origin across agree- base, which has two notable differences. First, ments; and, to the extent that these rules set it has a new base year, 2001 (the old base year patterns common to most agreements, the was 1997). Second, it has a new protection burden on customs administration would be database that takes better account of preferen- reduced. They could work with prospective tial trade access. Box 6.1 provides a brief 126 M A K I N G R E G I O N A L I S M C O M P L E M E N T A R Y T O M U L T I L A T E R A L I S M Box 6.1 Impacts of the new GTAP database T he simulation results from the World Bank's considerable reform between 1997 and 2001 (e.g., LINKAGE model were presented in the past three continued implementation of the Uruguay Round and issues of Global Economic Prospects (2002, 2003, China's progress towards WTO accession), and an and 2004) and were based on the use of various improved treatment of preferential trade agreements. versions of Release 5 of the GTAP dataset.* An up- The overall impact of these changes is that the dated version of the data has become available; pre- World Bank's estimate of the global gains from release 6.04 was released in September 2004. The global merchandise trade reform is now around main innovations of the new dataset include a more $260 billion** (in 2015, relative to the baseline recent base year (2001 instead of 1997), revised na- scenario), compared with around $380 billion using tional input-output tables, and a new database for the 1997-based results from previous GEPs.*** The estimating the levels of trade protection. This last lower number reflects, to a large extent, the impacts innovation is likely to have a large impact on trade of trade reforms achieved between 1997 and 2001 scenarios. First, a brief digression on a technical and the incorporation of preferences.**** The issue regarding the new base year: The global data- allocation of the gains across regions and sectors is base, to a large extent, is the outcome of merging broadly consistent with the previous results. Thus national data. The national input-output tables have 41 percent (instead of 45 percent) of the gains from different base years and are virtually always in global reform accrue to developing countries, and national currencies. Thus they must be updated to agricultural reform generates some 47 percent the given base year--2001 in the case of the new (instead of 58 percent) of the global gains. release--and converted to the database's common currency (i.e., the U.S. dollar). This has a practical implication because of movements in the value of the *The GTAP dataset is a product of the Global Trade Analy- dollar. In 1997, the U.S. dollar was relatively strong sis Project (GTAP), based at Purdue University, with funding (for example, $1.13 per euro), and it was much from an international consortium of international and national weaker in 2001 ($0.90 per euro). This means that agencies (including the World Bank), universities, and research the relative weight of countries will change between centers. Since its initial development in the early 1990s, the the two base years (holding growth constant). In GTAP dataset has become the premiere dataset for undertaking fact, the global U.S. economy as a share of global global trade analysis. The current version includes a full social output was around 32 percent in 1997 and only accounting matrix for 87 regions (of which 69 are individual countries), 57 economic sectors, and fully consistent bilateral 27 percent in 2001. Thus the change in the value of trade flows. the dollar will have some impact on the reported **These results should still be viewed as preliminary as the gains from trade reform, both globally and by region. GTAP consortium is preparing for the final release sometime be- The more important change in the database relates fore the end of 2004. The final release may result in some to the change in protection. The new protection data changes at the micro level, but will probably only have a rela- relies on the MAcMaps dataset--a collaborative ef- tively minor impact at the aggregate level. fort of CEPII (Paris) and the International Trade Cen- ***These gains refer to the so-called "static" effects, i.e., not tre (ITC, Geneva). Among the prominent features of taking into account dynamic effects such as improvements in the MAcMaps dataset is the incorporation of prefer- productivity. ential tariff regimes and the conversion of specific ****There are other technical differences, including among others, the relative change of the value of the U.S. dollar as men- tariffs to ad valorem equivalents; it thus represents a tioned above. The LINKAGE model, apart from the change in the more realistic picture of the bilateral levels of protec- database, is identical to that used in the last Global Economic tion. In summary, the new database will capture the Prospects. summary of these changes and the impacts of To assess the relative impacts of various using the new database relative to the results RTAs, it is useful to develop a benchmark outlined in previous Global Economic simulation (apart from the baseline). The Prospects. benchmark simulation is a global reform 127 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Table 6.1 Comparison of bilateral agreements to global trade reform (change in real income in 2015 compared to baseline) Bilateral with Bilateral minus Bilateral with Bilateral minus Global Quad large countries Global Quad large countries (1) (2) (3) (4) (5) (6) $ billion Percent High-income countries 154.4 133.6 46.9 0.6 0.5 0.2 Low-income countries 16.6 19.0 1.9 0.9 1.0 0.1 Middle-income countries 92.2 2.6 4.7 1.2 0.0 0.1 All developing countries 108.8 21.5 6.6 1.2 0.2 0.1 World Total 263.2 112.0 40.3 0.8 0.3 0.1 Source: World Bank simulations. scenario in which all merchandise trade dis- beyond even our 2015 time horizon, and rules tortions are eliminated (services reform is left of origin tests often limit preferential market out for lack of sufficient data), domestic dis- access. tortions in agriculture are removed (input and The first simulation, in which all develop- output subsidies, direct payments, and export ing countries sign a bilateral agreement with subsidies), and import quotas in the textile the Quad-plus countries (columns 2 and 5 in and clothing sectors are removed. This sce- table 6.1), shows that, as a group, developing nario would be the ultimate long-run outcome countries are substantially worse off than with of successful multilateralism. Under this re- a multilateral agreement. Instead of gaining form scenario, the global gains in 2015 $109 billion from global reform, they lose amount to $263 billion, or an increase of 0.8 $22 billion relative to a baseline scenario, with percent in baseline income (table 6.1).2 no change in protection. If one looks at indi- vidual countries (table 6.2), the effect is nearly How much do regional trade agreements universal; only a handful of developing coun- benefit developing countries? tries--for example, Brazil and China--would To examine how bilateral agreements affect gain from a full hub-and-spoke system of bi- developing countries, we look at three simula- lateral agreements, and all developing coun- tions: One in which all developing countries tries would lose compared to full multilateral sign a bilateral agreement with Quad-plus trade. countries (United States, EU, Japan, Canada, It is interesting that some of the Quad plus Australia and New Zealand);3 a second countries would benefit from this strategy. Al- simulation--similar to the first simulation--in though the high-income countries would gen- which the large countries, such as Brazil, erally lose from this set of bilateral agreements China, and India are excluded, which is per- compared to global reform, the impact is not haps a more plausible scenario; and a third in uniform.5 Both the United States and the EU which each developing country/region signs (the most aggressive advocates of bilateral an individual bilateral agreement with the deals) would appear to benefit more from pur- Quad-plus countries, assuming other develop- suing bilateral agreements with all develop- ing countries do not sign agreements.4 Note ing countries than from global reform that these scenarios overstate bilateral and (table 6.2); the United States would gain an multilateral effects because they assume that additional $7 billion (0.1 percent of GDP), no sectors are exempt, and rules of origin are while the EU would gain $27 billion (0.4 per- not restrictive. In reality, the United States and cent of GDP). Although they would have to EU bilateral agreements usually exclude im- open up their agricultural markets to some ex- portant sectors, such as sensitive agricultural tent (assuming exemptions are disallowed), products, or attach extended phase-in periods they would not have to dismantle domestic 128 each agree- between and (8) 1.3 1.0 0.2 2.1 0.3 0.2 2.6 1.9 1.8 0.5 0.2 0.1 0.2 2.0 0.1 0.9 1.8 country developing Bilateral ments Quad and large countries agreements Quad (7) 0.0 0.1 0.4 0.1 0.2 0.1 0.3 0.3 0.4 1.2 0.2 0.3 0.5 1.3 0.7 0.4 0.1 0.3 0.0 1.5 1.0 0.6 0.6 0.2 0.1 0.1 0.1 0.1 countries excluding Bilateral between developing Percent all agree- between and (6) 0.5 0.3 1.1 0.8 1.0 0.7 0.3 0.4 1.2 0.9 0.2 0.5 0.2 0.5 1.2 1.3 0.4 0.7 0.4 0.6 1.2 0.1 0.1 0.5 1.0 0.0 0.2 0.3 developing countries Bilateral ments Quad reform (5) 0.7 0.2 0.7 0.9 2.6 2.8 1.4 0.6 0.5 1.4 0.0 0.8 1.8 5.0 4.7 0.2 0.2 0.5 0.9 6.7 1.1 1.6 1.3 0.6 0.9 1.2 1.2 0.8 Global merchandise trade and agreements Quad (4) 7.3 2.1 5.1 2.6 0.8 3.7 0.9 7.4 1.2 0.9 0.4 1.3 5.7 developing 21.8 ­0.2 9.6 4.0 country orm each Bilateral between ref and trade large countries agreements Quad (3) 0.5 4.8 2.3 0.3 1.7 7.2 3.1 3.0 1.3 1.3 0.8 0.6 2.8 1.1 0.5 1.3 0.1 4.3 2.5 6.4 1.2 1.9 4.7 6.6 10.7 33.6 46.9 40.3 countries global excluding Bilateral between developing billion $ with baseline) all to agree- between and 6.0 9.8 2.4 1.5 9.7 2.3 1.5 1.7 0.3 0.2 5.0 3.2 2.0 3.3 2.7 1.9 3.0 0.9 0.3 2.6 (2) 32.3 82.4 25.0 10.0 19.0 21.5 133.6 112.0 developing countries Bilateral ments Quad greementsa compared - mer trade 0.8 2.3 6.1 2.9 3.0 (1) 8.4 9.8 8.0 4.3 3.6 0.3 2.9 2.5 2.4 0.4 24.9 55.0 29.7 26.4 14.1 19.6 19.1 16.3 16.6 92.2 bilateral 2015 reform 154.4 108.8 263.2 Global of in chandise A income EFT simulations. and Africa and and real with countries Comparison Asia countries Bank in Korea (China) Asia countries orld countries countries W e Canada, Union orld 6.2 of (China) Zealand States East South ECA East Africa Sub-Saharan LAC the W Kong otalT le of of of of of of abT hang New aiwanT accession developing Singapore (c Australia, United European Japan Republic Hong Brazil China India Indonesia Mexico Russia SACU ietnamV orld Rest Rest EU Rest Middle North Rest Rest Rest High-income Low-income Middle-income All W Source: 129 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 support programs. Hence the EU and the excluded--not surprising given their weight in United States improve market access in highly global trade with the Quad countries. Finally, protected markets in developing countries, but the impacts on the excluded countries are they do not face the full force of competition mixed: Brazil and China, which would gain in between themselves, particularly in agricul- a full hub-and-spoke system, lose when ex- ture, nor the full brunt of competition in de- cluded. The other excluded regions--India, veloping countries. For example, when India Mexico, Russia, rest of East Asia, and rest of opens up to Quad country imports, Quad South Asia--would see a dampening of their country exporters do not simultaneously face losses. increased competition from developing coun- This adverse outcome from bilateral agree- try exporters as they would in a multilateral ments with the Quad raises the question of agreement. In agriculture, Quad producers why developing countries are so anxious to will face greater competition from relatively pursue them. There are a number of possible efficient developing country exporters, but reasons, not necessarily mutually exclusive. some of the fiercest competition will be among First, countries may hope to maximize their themselves. And the terms of trade losses that benefits through first-mover advantages. Sec- they would suffer from removing agricultural ond, countries aim to guarantee market access protection between Quad countries is muted on a permanent basis. Third, there might be when signing the bilateral agreements. Note a desire to pre-empt other countries. Fourth, a that this is not the case with Japan, whose bilateral agreement may be used as leverage agriculture is relatively more threatened by to facilitate domestic reforms. And fifth, other developing country market access; it would components of an agreement (for example, gain more from a multilateral agreement, al- services, trade facilitation, and so on) may though it nonetheless gains significantly from have significant benefits in addition to simply bilateralism. It is the Quad-plus agricultural improving market access for merchandise exporters--Australia, Canada, and New goods. Focusing on the first of these possible Zealand--that would prefer multilateralism, reasons, we use the model to simulate the im- because the gains from access to European, pact of each developing country signing an Japanese, and the American markets, and the agreement with the Quad countries, but with dismantling of distortionary agricultural sup- no other country doing so. The results port programs would be highly beneficial for (table 6.2 columns 4 and 8) provide some jus- their farmers. tification for developing countries' pursuit of Were the large developing countries6 to be RTAs with the Quad if they believe they can do excluded from the hub-and-spoke bilateral so exclusively, or at least capture a "first arrangements (perhaps a more plausible sce- mover" advantage by getting there first. About nario), the broad conclusions still hold, but one-half of the developing regions would be are muted (columns 3 and 6). First, many de- better off with a bilateral agreement than with veloping regions still lose in absolute terms a global agreement; the winners (relative to compared with the baseline scenario. Second, global liberalization) include China, Indonesia, all lose relative to the gains from a global re- Mexico, Southern African Customs Union form scenario, though for some, the hub-and- (SACU), rest of South Asia, and EU accession spoke gains approach those in the global countries. Losers include Brazil (slightly), reform scenario (e.g., Indonesia, and to a India, Russia, Vietnam, rest of East Asia, rest of lesser extent, the rest of LAC and rest of the ECA, Middle East, North Africa, rest of Sub- world regions). The gains for the high-income Saharan Africa, and rest of LAC. countries, on the other hand, are significantly A few cases deserve special mention. The lower when the large developing countries are rest of the Sub-Saharan Africa region could 130 M A K I N G R E G I O N A L I S M C O M P L E M E N T A R Y T O M U L T I L A T E R A L I S M suffer losses from a bilateral agreement with the Quad. Because this region already has rel- Figure 6.1 Global outcomes dominate alternatives atively free access to the Quad markets, open- ing up to permit greater imports from the Change in real income in 2015 compared to baseline Quad worsens their terms of trade and negates any gains from the bilateral agreement. Russia BILAT Middle-income JBIL and the Middle East are dependent on energy Global exports and these face low tariffs in industrial Low-income countries (even if energy is heavily taxed), so these regions have little to gain from addi- Developing tional market access. These cases highlight one of the key findings from recent Global Quad Economic Prospects, that developing coun- 1.5 1.0 0.5 0 0.5 1.0 1.5 2.0 tries have much to gain from greater market Percent access to other developing countries. First, be- cause protection is, on average, higher in de- Note: Global refers to the global merchandise trade reform veloping countries, and second, because of the scenario, JBIL corresponds to the simulation where all developing countries sign bilateral agreements with the high growth potential of developing countries Quad-plus, and BILAT corresponds to the simulation where over the next decade compared with the in- the bilateral agreements are signed individually. Results reflect un-weighted regional averages. dustrial countries. Source: World Bank simulations. The idea that a single developing country or region would be able to sign bilateral agree- ments with the Quad countries without other developing countries doing the same over the reforms (box 6.2). The implications are clear: next decade is unrealistic. Indeed, the increase The most development friendly outcome is as- in the number of agreements over the last sociated with global reform, and a full set of decade means that a portion of any first- bilateral agreements would leave virtually all mover advantage has been eroded already. But developing countries worse off than at as a conceptual exercise, it does help illustrate present. what may have motivated some developing Countries are also pursuing other RTAs, countries to aggressively pursue deals with both North-South and South-South. These one or more Quad members. agreements are linked to the ongoing talks on To summarize, developing countries could the Free Trade Areas of the Americas (FTAA), gain an (unweighted) average of 1.7 percent in the ASEAN 3 negotiations, and the EU's vari- real income from a global agreement (fig- ous agreements/negotiations with the EU- ure 6.1). But if all developing countries sign accession countries toward the east, partners bilateral agreements with the Quad, creating a around the Mediterranean, and in Sub- complex hub-and-spoke system, developing Saharan Africa toward the south. Chapter 2 of countries actually suffer losses averaging this report documents the large number of ex- 0.4 percent (1.0 percent for the low-income isting or prospective agreements among devel- countries alone). While some individual devel- oping countries. Figures 6.2 and 6.3 summa- oping countries might have gained from enter- rize the overall impacts from simulating ing exclusive agreements with Quad countries, selected North-South and South-South agree- RTA proliferation has already eliminated ments.7 Some of these proffer relatively signifi- that first-mover advantage. Moreover, RTA- cant gains (for example, a broad free trade re- inducedstructuralchangescouldproducedisin- gion in East Asia), but are nonetheless clearly centives for achieving broad-based multilateral inferior to the gains from global merchandise 131 G L O B A L E C O N O M I C P R O S P E C T S 2 0 0 5 Box 6.2 Regional trade agreements, structural change, and congruence W hile impact assessment of RTAs and global Joint bilateral vs. bilaterals agreements often focus narrowly on the real Index (1 --> fully congruent) income impacts, from a political economy point of view, the main drivers of these agreements typically 6 will be at the micro or institutional level, where it is 5 easier to identify the potential winners and losers. 4 BILAT Moreover, the macro analysis typically ignores the transitional costs. A final issue deals with the com- JBIL 3 patibility of partial reforms (as represented by prefer- 2 ential trade agreements) with multilateral reforms, which is arguably where the world economy is head- 1 ing. In other words, how compatible are the struc- 0 tural changes induced by a partial reform with the a x Br Chn Ind IdnMe Rus SacVnmXea Xsa Cec XecMde NafXss Rlc structural changes one would anticipate from a mul- Source: World Bank simulations. tilateral reform? For example, what if a country such as Vietnam has a regional comparative advantage in likely lead to a 50 percent change in our structural agriculture, but a global comparative advantage in adjustment measure. appa