making regionalism complementary to multilateralism

125

6

The emergence of a more proactive stance inmultilateral negotiations by developing coun-tries in parallel with an explosion of preferentialregional trade agreements (RTAs) around theworld raises an important question: Are thesetrends complementary, representing differentpaths to the same desired outcome of fastergrowth, development, and poverty reduction,or are they competing and incompatible? Inprinciple, the best outcome for all countrieswould be a nondiscriminatory trading system;developing countries, in particular, would ben-efit from a nondiscriminatory trading systembecause most poor people and many poor coun-tries might find themselves excluded from pref-erential deals. If the explosion in RTAs impliesa higher probability that the majority ofdeveloping countries would face greater dis-crimination than under an MFN (or nondis-criminatory) regime, the world as a whole willbe worse off, and individual developing coun-tries may lose substantially.1

RTAs can be a complement to multilateralreform, but they are not a substitute. Considera scenario, modeled in the first section of thischapter, in which each developing countrysigned an agreement with the United States,the European Union (EU), Canada, andJapan—the Quad. Each developing countrycould raise its real incomes by individuallysigning bilateral agreements with the Quadcountries; and in some cases they would raisetheir real incomes more than they wouldthrough multilateral accords. But these gains

disappear if all developing countries were tosign such agreements. In fact, all developingcountries would lose relative to a multilateralagreement and even relative to the baseline.This scenario underscores the fact that, whilestalled collective action through multilateralchannels creates a strong incentive for eachdeveloping country to sign a regional agree-ment, if every country does so, they all lose.

RTAs do alter the incentives for countriesto participate in multilateral liberalization.RTAs can be a stumbling block to multilateralarrangements by creating incentives to resistthe preference erosion that can occur throughnew multilateral liberalization. However,because the gains are often substantially largerin multilateral agreements, concerns over pref-erence erosion may be limited to a few smallcountries that could conceivably block a mul-tilateral agreement. Those recalcitrant coun-tries resisting reforms are likely the beneficia-ries of preferences associated with distortedagreements encompassing agriculture or theclothing and textile trade, not RTAs. Large de-veloped countries may gain more from signingindividual bilateral agreements than theywould from a multilateral accord, becausethey can use the carrot of preferential access toextract concessions in nontrade areas fromdeveloping country partners that would be re-sisted in the WTO negotiating framework. Butwe see little evidence that the high-incomecountries have reduced their effort to bring thecurrent multilateral negotiations to fruition.

Making RegionalismComplementary to Multilateralism

From a development perspective, the WTOremains the best-available forum to disciplinethe use of trade-distorting policies. RTAs cancomplement the WTO efforts by cooperatingon behind-the-border policies, especially onregulation-intensive issues such as services,trade facilitation, and the investment climate.Governments pursuing this agenda throughRTAs must adopt rules that are appropriate totheir own level of development. To be effec-tive, the rules must target a priority concern,must be enforceable, and must avoid becom-ing a tax on scarce resources that would bebetter used elsewhere. Getting the rules“right” so that they promote developmenthas implications for negotiations and enforce-ment of the resulting disciplines. The potentialfor inappropriate outcomes is higher inNorth–South RTAs because the asymmetry innegotiating power can overtake real develop-ment priorities.

Reinforcing the complementarity betweenregionalism and multilateralism and minimiz-ing the latent tensions must begin with thecompletion of the Doha Development Agenda.If the Doha Development Agenda is completedin a way that actually promotes development,it would bring down tariffs, enhance the gainsfrom open regionalism, and discipline any ex-clusionary effects of regional accords.

All countries could take steps to promoteopen regionalism—the developing countries,high-income countries, and the internationalcommunity working together through theWTO. Developing countries are likely to havethe greatest success in harnessing trade forgrowth and poverty reduction if they adopta three-pronged strategy that involvesautonomous liberalization, active multilateral-ism, and open regionalism.

High-income countries could promoteopen regionalism by including agriculture inRTAs. They could adopt more common andnonrestrictive rules of origin across agree-ments; and, to the extent that these rules setpatterns common to most agreements, theburden on customs administration would bereduced. They could work with prospective

partners to ensure that new regulations re-garding investment and intellectual propertyare appropriate to the level of development,which would reduce risks of undue enforce-ment costs. Finally, they could provide trade-related technical assistance not only in the im-plementation phase but also in the negotiatingphase, which could promote greater liberaliza-tion and supply response to new marketopportunities in regional or global markets.

The international community, workingthrough the WTO, can reduce discrimination inthe system. The procedures associated withRTA disciplines as currently configured are ill-suited to limit either their proliferation or tocontrol their discriminatory provisions. WTOmembers should establish stronger multilateralsurveillance mechanisms to document, analyze,and monitor the effects of RTAs on nonmem-bers. Expanding the information on the impactof RTAs to stakeholders—firms, consumers,taxpayers—would also help ensure that the po-tential benefits of liberalization are both real-ized and distributed more equitably. Medium-term efforts should focus on implementingWTO disciplines on regional agreements.

Preferential Agreements within the Global Context

To place preferential trade arrangements ina multilateral context, we evaluate how

different collections of trade agreements com-pare with multilateral alternatives. For thisevaluation, we utilize the World Bank’s globaltrade model known as LINKAGE, which hasbeen used in previous Global EconomicProspects. The model is built around theGTAP database that is widely used to assessthe global, regional, and country implicationsof alternative trade liberalization scenarios.However, the results described in this year’sreport reflect an update of the model’s data-base, which has two notable differences. First,it has a new base year, 2001 (the old base yearwas 1997). Second, it has a new protectiondatabase that takes better account of preferen-tial trade access. Box 6.1 provides a brief

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

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

127

The simulation results from the World Bank’sLINKAGE model were presented in the past three

issues of Global Economic Prospects (2002, 2003,and 2004) and were based on the use of variousversions of Release 5 of the GTAP dataset.* An up-dated version of the data has become available; pre-release 6.04 was released in September 2004. Themain innovations of the new dataset include a morerecent base year (2001 instead of 1997), revised na-tional input-output tables, and a new database forestimating the levels of trade protection. This lastinnovation is likely to have a large impact on tradescenarios. First, a brief digression on a technicalissue regarding the new base year: The global data-base, to a large extent, is the outcome of mergingnational data. The national input-output tables havedifferent base years and are virtually always innational currencies. Thus they must be updated tothe given base year—2001 in the case of the newrelease—and converted to the database’s commoncurrency (i.e., the U.S. dollar). This has a practicalimplication because of movements in the value of thedollar. In 1997, the U.S. dollar was relatively strong(for example, $1.13 per euro), and it was muchweaker in 2001 ($0.90 per euro). This means thatthe relative weight of countries will change betweenthe two base years (holding growth constant). Infact, the global U.S. economy as a share of globaloutput was around 32 percent in 1997 and only27 percent in 2001. Thus the change in the value ofthe dollar will have some impact on the reportedgains from trade reform, both globally and by region.

The more important change in the database relatesto the change in protection. The new protection datarelies on the MAcMaps dataset—a collaborative ef-fort of CEPII (Paris) and the International Trade Cen-tre (ITC, Geneva). Among the prominent features ofthe MAcMaps dataset is the incorporation of prefer-ential tariff regimes and the conversion of specifictariffs to ad valorem equivalents; it thus represents amore realistic picture of the bilateral levels of protec-tion. In summary, the new database will capture the

Box 6.1 Impacts of the new GTAP databaseconsiderable reform between 1997 and 2001 (e.g.,continued implementation of the Uruguay Round andChina’s progress towards WTO accession), and animproved treatment of preferential trade agreements.

The overall impact of these changes is that theWorld Bank’s estimate of the global gains fromglobal merchandise trade reform is now around$260 billion** (in 2015, relative to the baselinescenario), compared with around $380 billion usingthe 1997-based results from previous GEPs.*** Thelower number reflects, to a large extent, the impactsof trade reforms achieved between 1997 and 2001and the incorporation of preferences.**** Theallocation of the gains across regions and sectors isbroadly consistent with the previous results. Thus41 percent (instead of 45 percent) of the gains fromglobal reform accrue to developing countries, andagricultural reform generates some 47 percent(instead of 58 percent) of the global gains.

*The GTAP dataset is a product of the Global Trade Analy-sis Project (GTAP), based at Purdue University, with fundingfrom an international consortium of international and nationalagencies (including the World Bank), universities, and researchcenters. Since its initial development in the early 1990s, theGTAP dataset has become the premiere dataset for undertakingglobal trade analysis. The current version includes a full socialaccounting matrix for 87 regions (of which 69 are individualcountries), 57 economic sectors, and fully consistent bilateraltrade flows.

**These results should still be viewed as preliminary as theGTAP consortium is preparing for the final release sometime be-fore the end of 2004. The final release may result in somechanges at the micro level, but will probably only have a rela-tively minor impact at the aggregate level.

***These gains refer to the so-called “static” effects, i.e., nottaking into account dynamic effects such as improvements inproductivity.

****There are other technical differences, including amongothers, the relative change of the value of the U.S. dollar as men-tioned above. The LINKAGE model, apart from the change in thedatabase, is identical to that used in the last Global EconomicProspects.

summary of these changes and the impacts ofusing the new database relative to the resultsoutlined in previous Global EconomicProspects.

To assess the relative impacts of variousRTAs, it is useful to develop a benchmarksimulation (apart from the baseline). Thebenchmark simulation is a global reform

scenario in which all merchandise trade dis-tortions are eliminated (services reform is leftout for lack of sufficient data), domestic dis-tortions in agriculture are removed (input andoutput subsidies, direct payments, and exportsubsidies), and import quotas in the textileand clothing sectors are removed. This sce-nario would be the ultimate long-run outcomeof successful multilateralism. Under this re-form scenario, the global gains in 2015amount to $263 billion, or an increase of 0.8percent in baseline income (table 6.1).2

How much do regional trade agreementsbenefit developing countries? To examine how bilateral agreements affectdeveloping countries, we look at three simula-tions: One in which all developing countriessign a bilateral agreement with Quad-pluscountries (United States, EU, Japan, Canada,plus Australia and New Zealand);3 a secondsimulation—similar to the first simulation—inwhich the large countries, such as Brazil,China, and India are excluded, which is per-haps a more plausible scenario; and a third inwhich each developing country/region signsan individual bilateral agreement with theQuad-plus countries, assuming other develop-ing countries do not sign agreements.4 Notethat these scenarios overstate bilateral andmultilateral effects because they assume thatno sectors are exempt, and rules of origin arenot restrictive. In reality, the United States andEU bilateral agreements usually exclude im-portant sectors, such as sensitive agriculturalproducts, or attach extended phase-in periods

beyond even our 2015 time horizon, and rulesof origin tests often limit preferential marketaccess.

The first simulation, in which all develop-ing countries sign a bilateral agreement withthe Quad-plus countries (columns 2 and 5 intable 6.1), shows that, as a group, developingcountries are substantially worse off than witha multilateral agreement. Instead of gaining$109 billion from global reform, they lose$22 billion relative to a baseline scenario, withno change in protection. If one looks at indi-vidual countries (table 6.2), the effect is nearlyuniversal; only a handful of developing coun-tries—for example, Brazil and China—wouldgain from a full hub-and-spoke system of bi-lateral agreements, and all developing coun-tries would lose compared to full multilateraltrade.

It is interesting that some of the Quadcountries would benefit from this strategy. Al-though the high-income countries would gen-erally lose from this set of bilateral agreementscompared to global reform, the impact is notuniform.5 Both the United States and the EU(the most aggressive advocates of bilateraldeals) would appear to benefit more from pur-suing bilateral agreements with all develop-ing countries than from global reform(table 6.2); the United States would gain anadditional $7 billion (0.1 percent of GDP),while the EU would gain $27 billion (0.4 per-cent of GDP). Although they would have toopen up their agricultural markets to some ex-tent (assuming exemptions are disallowed),they would not have to dismantle domestic


128

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 minusGlobal 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 0.2Low-income countries 16.6 �19.0 �1.9 0.9 �1.0 �0.1Middle-income countries 92.2 �2.6 �4.7 1.2 0.0 �0.1All developing countries 108.8 �21.5 �6.6 1.2 �0.2 �0.1World Total 263.2 112.0 40.3 0.8 0.3 0.1

Source: World Bank simulations.

129

Table 6.2 Comparison of bilateral agreements with global trade reform(change in real income in 2015 compared to baseline)

Bilateral agree- Bilateral agreements Bilateral agree- Bilateral agreements Bilateral agree-ments between between Quad and Bilateral agreements ments between between Quad and ments between

Global mer- Quad and all developing countries between Quad and Global Quad and all developing countries Quad and eachchandise trade developing excluding large each developing merchandise developing excluding large developing

reform countries countries country trade reform countries countries country(1) (2) (3) (4) (5) (6) (7) (8)

$ billion Percent

Australia, Canada, and 8.4 6.0 0.5 0.7 0.5 0.0New Zealand

United States 24.9 32.3 10.7 0.2 0.3 0.1European Union with EFTA 55.0 82.4 33.6 0.7 1.1 0.4Japan 29.7 25.0 4.8 0.9 0.8 0.1Republic of Korea and 26.4 �9.8 �2.3 2.6 �1.0 �0.2

Taiwan (China)Hong Kong (China) and 9.8 �2.4 �0.3 2.8 �0.7 �0.1

SingaporeBrazil 8.0 1.5 �1.7 7.3 1.4 0.3 �0.3 1.3China 14.1 9.7 �7.2 21.8 0.6 0.4 �0.3 1.0India 4.3 �10.0 �3.1 2.1 0.5 �1.2 �0.4 0.2Indonesia 3.6 �2.3 3.0 5.1 1.4 �0.9 1.2 2.1Mexico 0.3 �1.5 �1.3 2.6 0.0 �0.2 �0.2 0.3Russia 2.9 �1.7 �1.3 0.8 0.8 �0.5 �0.3 0.2SACU 2.5 �0.3 0.8 3.7 1.8 �0.2 0.5 2.6Vietnam 2.4 �0.2 0.6 0.9 5.0 �0.5 1.3 1.9Rest of East Asia 19.6 �5.0 �2.8 7.4 4.7 �1.2 �0.7 1.8Rest of South Asia 0.4 �3.2 �1.1 1.2 0.2 �1.3 �0.4 0.5EU accession countries 0.8 �2.0 �0.5 0.9 0.2 �0.4 �0.1 0.2Rest of ECA 2.3 �3.3 �1.3 0.4 0.5 �0.7 �0.3 0.1Middle East 6.1 �2.7 �0.1 1.3 0.9 �0.4 0.0 0.2North Africa 19.1 1.9 4.3 5.7 6.7 0.6 1.5 2.0Rest of Sub-Saharan Africa 2.9 �3.0 �2.5 –0.2 1.1 �1.2 �1.0 �0.1Rest of LAC 16.3 0.9 6.4 9.6 1.6 0.1 0.6 0.9Rest of the World 3.0 �0.3 1.2 4.0 1.3 �0.1 0.6 1.8

High-income countries 154.4 133.6 46.9 0.6 0.5 0.2Low-income countries 16.6 �19.0 �1.9 0.9 �1.0 �0.1Middle-income countries 92.2 �2.6 �4.7 1.2 0.0 �0.1All developing countries 108.8 �21.5 �6.6 1.2 �0.2 �0.1World Total 263.2 112.0 40.3 0.8 0.3 0.1


support programs. Hence the EU and theUnited States improve market access in highlyprotected markets in developing countries, butthey do not face the full force of competitionbetween themselves, particularly in agricul-ture, nor the full brunt of competition in de-veloping countries. For example, when Indiaopens up to Quad country imports, Quadcountry exporters do not simultaneously faceincreased competition from developing coun-try exporters as they would in a multilateralagreement. In agriculture, Quad producerswill face greater competition from relativelyefficient developing country exporters, butsome of the fiercest competition will be amongthemselves. And the terms of trade losses thatthey would suffer from removing agriculturalprotection between Quad countries is mutedwhen signing the bilateral agreements. Notethat this is not the case with Japan, whoseagriculture is relatively more threatened bydeveloping country market access; it wouldgain more from a multilateral agreement, al-though it nonetheless gains significantly frombilateralism. It is the Quad-plus agriculturalexporters—Australia, Canada, and NewZealand—that would prefer multilateralism,because the gains from access to European,Japanese, and the American markets, and thedismantling of distortionary agricultural sup-port programs would be highly beneficial fortheir farmers.

Were the large developing countries6 to beexcluded from the hub-and-spoke bilateralarrangements (perhaps a more plausible sce-nario), the broad conclusions still hold, butare muted (columns 3 and 6). First, many de-veloping regions still lose in absolute termscompared with the baseline scenario. Second,all lose relative to the gains from a global re-form scenario, though for some, the hub-and-spoke gains approach those in the globalreform scenario (e.g., Indonesia, and to alesser extent, the rest of LAC and rest of theworld regions). The gains for the high-incomecountries, on the other hand, are significantlylower when the large developing countries are

excluded—not surprising given their weight inglobal trade with the Quad countries. Finally,the impacts on the excluded countries aremixed: Brazil and China, which would gain ina full hub-and-spoke system, lose when ex-cluded. The other excluded regions—India,Mexico, Russia, rest of East Asia, and rest ofSouth Asia—would see a dampening of theirlosses.

This adverse outcome from bilateral agree-ments with the Quad raises the question ofwhy developing countries are so anxious topursue them. There are a number of possiblereasons, not necessarily mutually exclusive.First, countries may hope to maximize theirbenefits through first-mover advantages. Sec-ond, countries aim to guarantee market accesson a permanent basis. Third, there might bea desire to pre-empt other countries. Fourth, abilateral agreement may be used as leverageto facilitate domestic reforms. And fifth, othercomponents of an agreement (for example,services, trade facilitation, and so on) mayhave significant benefits in addition to simplyimproving market access for merchandisegoods. Focusing on the first of these possiblereasons, we use the model to simulate the im-pact of each developing country signing anagreement with the Quad countries, but withno other country doing so. The results(table 6.2 columns 4 and 8) provide some jus-tification for developing countries’ pursuit ofRTAs with the Quad if they believe they can doso exclusively, or at least capture a “firstmover” advantage by getting there first. Aboutone-half of the developing regions would bebetter off with a bilateral agreement than witha global agreement; the winners (relative toglobal liberalization) include China, Indonesia,Mexico, Southern African Customs Union(SACU), rest of South Asia, and EU accessioncountries. Losers include Brazil (slightly),India, Russia, Vietnam, rest of East Asia, rest ofECA, Middle East, North Africa, rest of Sub-Saharan Africa, and rest of LAC.

A few cases deserve special mention. Therest of the Sub-Saharan Africa region could


130

suffer losses from a bilateral agreement withthe Quad. Because this region already has rel-atively free access to the Quad markets, open-ing up to permit greater imports from theQuad worsens their terms of trade and negatesany gains from the bilateral agreement. Russiaand the Middle East are dependent on energyexports and these face low tariffs in industrialcountries (even if energy is heavily taxed), sothese regions have little to gain from addi-tional market access. These cases highlightone of the key findings from recent GlobalEconomic Prospects, that developing coun-tries have much to gain from greater marketaccess to other developing countries. First, be-cause protection is, on average, higher in de-veloping countries, and second, because of thehigh growth potential of developing countriesover the next decade compared with the in-dustrial countries.

The idea that a single developing countryor region would be able to sign bilateral agree-ments with the Quad countries without otherdeveloping countries doing the same over thenext decade is unrealistic. Indeed, the increasein the number of agreements over the lastdecade means that a portion of any first-mover advantage has been eroded already. Butas a conceptual exercise, it does help illustratewhat may have motivated some developingcountries to aggressively pursue deals withone or more Quad members.

To summarize, developing countries couldgain an (unweighted) average of 1.7 percent inreal income from a global agreement (fig-ure 6.1). But if all developing countries signbilateral agreements with the Quad, creating acomplex hub-and-spoke system, developingcountries actually suffer losses averaging0.4 percent (1.0 percent for the low-incomecountries alone). While some individual devel-oping countries might have gained from enter-ing exclusive agreements with Quad countries,RTA proliferation has already eliminatedthat first-mover advantage. Moreover, RTA-inducedstructural changes couldproducedisin-centives for achieving broad-based multilateral

reforms (box 6.2). The implications are clear:The most development friendly outcome is as-sociated with global reform, and a full set ofbilateral agreements would leave virtually alldeveloping countries worse off than atpresent.

Countries are also pursuing other RTAs,both North-South and South-South. Theseagreements are linked to the ongoing talks onthe Free Trade Areas of the Americas (FTAA),the ASEAN�3 negotiations, and the EU’s vari-ous agreements/negotiations with the EU-accession countries toward the east, partnersaround the Mediterranean, and in Sub-Saharan Africa toward the south. Chapter 2 ofthis report documents the large number of ex-isting or prospective agreements among devel-oping countries. Figures 6.2 and 6.3 summa-rize the overall impacts from simulatingselected North-South and South-South agree-ments.7 Some of these proffer relatively signifi-cant gains (for example, a broad free trade re-gion in East Asia), but are nonetheless clearlyinferior to the gains from global merchandise


131

�1.5 �1.0 �0.5 0 0.5

Percent

1.0 1.5

Note: Global refers to the global merchandise trade reformscenario, JBIL corresponds to the simulation where alldeveloping countries sign bilateral agreements with theQuad-plus, and BILAT corresponds to the simulation wherethe bilateral agreements are signed individually. Resultsreflect un-weighted regional averages.


2.0

Figure 6.1 Global outcomes dominatealternatives

Quad�

Low-income

Middle-income

Developing

Change in real income in 2015 compared to baseline

Global

BILATJBIL

trade reform. Two additional comments re-garding the results of these regional simulationsare suggested by the figures.8 First, whenNorth-South agreements are implemented si-multaneously, the gains are dampened relativeto when they are implemented in isolation—areflection of the impacts of trade diversion. In

the case of the South-South agreements, theweak gains reflect, in part, the preferences al-ready granted in many of these regions as wellas a lack of distinct comparative advantage.9

This again emphasizes that broad South-Southand North-South trade reform is needed to reapsignificant gains.


132

While impact assessment of RTAs and globalagreements often focus narrowly on the real

income impacts, from a political economy point ofview, the main drivers of these agreements typicallywill be at the micro or institutional level, where it iseasier to identify the potential winners and losers.Moreover, the macro analysis typically ignores thetransitional costs. A final issue deals with the com-patibility of partial reforms (as represented by prefer-ential trade agreements) with multilateral reforms,which is arguably where the world economy is head-ing. In other words, how compatible are the struc-tural changes induced by a partial reform with thestructural changes one would anticipate from a mul-tilateral reform? For example, what if a country suchas Vietnam has a regional comparative advantage inagriculture, but a global comparative advantage inapparel? Would a regional agreement then make itmore difficult to achieve a multilateral accord?

The figure above provides a summary indicatorof the congruence or compatibility of the bilateralagreements—both the individual (BIL) and joint(JBIL). The indicator measures the average structuralchange of moving from the baseline to the partialagreement, and then moving from the partial agree-ment to global free trade, relative to the structuralchange induced by the global agreement. If the par-tial agreement is compatible (or congruent) with theglobal agreement, this measure is 1; that is, the two-step structural change is identical to the one-stepstructural change from implementing directly aglobal free trade agreement. For example, a globalagreement of a 50 percent tariff cut is largely con-gruent with a 100 percent tariff cut and would most

Box 6.2 Regional trade agreements, structural change,and congruence

likely lead to a 50 percent change in our structuraladjustment measure.

For most of the developing regions, the joint bilat-eral agreements are broadly consistent with the globalagreement, with the structural index varying fromaround 1 (or near perfect congruence) to a high ofaround 2—for example, for Middle East and NorthAfrica. On the other hand, the individual bilateralagreements are clearly not congruent with a globalagreement. The pattern of structural change inducedby the individual bilateral agreements is markedly dif-ferent from what one would anticipate with a globalagreement. Developing countries therefore face atradeoff. They can get short-term benefits from signinga bilateral agreement with the Quad—assuming othercountries do not—but the longer-term gains from aglobal agreement may be more difficult to attain be-cause the patterns of capital and labor allocationwould be misaligned by the partial agreement.

Index (1 —> fully congruent)

BraChn In

d Idn

Mex Rus Sac

Vnm Xea Xsa Cec XecM

de Naf Xss Rlc

Joint bilateral vs. bilaterals

0

1

2

6

5

4

JBIL

BILAT

3


Building Blocks versus StumblingBlocks

Whether regional agreements are buildingblocks or stumbling blocks to open

global markets—the terms Jagdish Bhagwati(1991) used—remains a central question.10

Proponents of the stumbling block theory em-phasize that: (1) RTAs may promote costlytrade diversion rather than efficient trade cre-ation, especially when sizable MFN tariffsremain—these tariffs create vested interests tomaintain preferential margins in “their” mar-kets; (2) proliferating regional agreements ab-sorb scarce negotiating resources (especially inpoorer WTO members) and crowd out policy-makers attention; (3) competing RTAs (espe-cially different North-South combinations)may lock in incompatible regulatory structuresand standards, and may result in inappropriatenorms for developing country partners; and(4) by creating alternative legal frameworksand dispute settlement mechanisms, RTAsmay weaken the discipline and efficiency asso-ciated with a broadly recognized multilateralframework of rules.

Building block proponents stress that mov-ing forward in smaller steps is often easier toaccomplish, and it creates a certain reformmomentum: (1) regional/bilateral agreementscan help sensitize domestic constituencies toliberalization and keep the stakes lower toallow for incremental progress on trade; (2)expanding the number and coverage of RTAscan erode vested opposition to multilateralliberalization because each successive RTA re-duces the value of the margin of preference,thereby reducing the discriminatory impact;(3) RTAs are often more about building strate-gic and/or political alliances or locking in do-mestic reforms than about actual trade liberal-ization, and so are not necessarily competitivewith multilateral efforts; (4) regional arrange-ments can provide an incubator for develop-ing country firms/producers to learn to tradewith RTA partners without facing full globalcompetition; and (5) for some issues, such asregulatory cooperation, RTAs may be a viable


133

Percent

Note: Global refers to the global merchandise trade reformscenario, separate is when the North/South regional blocksare formed individually and block is when the North/Southregional blocks are implemented simultaneously.


Change in real income in 2015 relative to the baseline

East Asia Europe The Americas

Figure 6.2 Global reform dominatesNorth–South agreements

0.0

1.0

0.2

0.4

0.6

0.8

1.2

1.4

1.6

Global

Separate

Block

Percent

Note: Global refers to the global merchandise trade reformscenario, and separate is when the South/South regionalblocks are formed in isolation.


Change in real income in 2015 relative to the baseline

LAFTA AFTA SSA MNA SAFTACIS

Figure 6.3 Global reform dominatesSouth–South agreements

�0.5

2.0

0.0

0.5

1.0

1.5

2.5

3.0

Global

Separate

and more manageable alternative to theWTO, where “lowest common denominator”outcomes tend to prevail.

One strand of analysis in the literature ex-plores theoretical bargaining models to shedlight on these divergent perspectives. For ex-ample, Saggi (2004) considers how RTA for-mation affects incentives to participate in mul-tilateral trade liberalization. Using a stylizedthree-country oligopoly model of intra-industry trade, he reasons that while forma-tion of RTAs can reduce tariffs in the shortrun, in the long run it hinders multilateraltrade liberalization because it lowers the in-centives for nonmembers to pursue multilat-eral cooperation. In a three-country model,where country size and costs differ, Saggi andYildiz (2004) look at tariff outcomes undertwo different sets of rules: One where coun-tries can form RTAs, and another where theycannot. They find that free trade is less likelyto occur as an outcome when countries canjoin RTAs as well. Moreover, the outcome de-pends on size and cost features: When coun-tries are relatively similar, the RTA optiontends to lower world welfare; whereas withlarger differences among countries, welfare in-creases. In the latter case, gains are somewhatlarger for the small country getting access tothe larger country (because it gains more innew export earnings and loses less in domesticsurplus) than for the country with lower costs(because gains from expanded market accessincrease when their own costs are lower andpartner costs are higher). Aghion and others(2004) construct a dynamic bargaining modelto assess the relative effects of simultaneousmultilateral liberalization or sequential bilat-eral liberalization, and they identify situationsin which the latter can lead to worldwide freetrade. Their model concludes that, as long aslarge coalition payoffs are higher than if allcountries were combined in alternative coali-tion structures, the outcome is likely to end upin global free trade, regardless of sequencing.

The Saggi findings support the empiricalevidence that North-South RTAs with smalldeveloping countries yield the biggest benefits

for developing country participants. This intu-ition comes from the fact that most empiricalmodels derive their impact from the effect ofRTAs on prices and quantities of tradedgoods. But when one partner is much smallerthan the other, its participation in an RTA hasvirtually no impact on prices, and any incre-mental exports it sells are a tiny share of thelarge country trade flows—so the resultinggains or losses for the larger partner are alsosmall. Support for this conclusion comes fromthe computable general equilibrium model(CGE) simulations on the impact of RTAs; inmost cases, the net impact on the Northernpartner has been exceedingly small, whethermeasured in terms of trade flows, terms oftrade, or welfare changes.

This result in turn has implications for howRTA proliferation affects incentives to pursuemultilateral liberalization. If the impact of ad-ditional North-South RTAs on the industrialpartner is indeed quite marginal, then it seemslikely that the economic consequences of theseagreements would not dampen their willing-ness to pursue multilateral deals. As we seebelow, neither the EU nor the United Statesseem to be less disposed toward multilateralnegotiations because of their RTAs or theRTAs of each other.

Other studies show less sanguine results.Limão (2003) notes that North-South RTAscan involve more than just the mutual loweringof tariffs; industrial countries often lower theirtariffs and expand market access in exchangefor cooperation in a variety of nontrade areas,such as labor standards, intellectual property,migration, security, and so forth. Limão mod-els the interaction between such RTAs andmultilateral liberalization, and concludes thatpursuit of RTAs creates a strategic incentive forindustrial countries to maintain their multilat-eral tariffs at a higher level than they otherwisewould—to hold back tariff concessions in themultilateral arena in order to have bargainingroom for negotiating RTAs. In the case of theUnited States, multilateral trade liberalizationcommitments are less deep for products thatare produced by regional firms.11


134

A more compelling argument is that devel-oping countries with preferential access mayhave a vested interest in perpetuating the tar-iff walls that screen out competing countriesfrom preferred markets. In fact, Barbados,Jamaica, and Mauritius have all been outspo-ken opponents of the agricultural liberaliza-tion that would erode their preferences in EUsugar markets, and Bangladesh has advocatedmeasures that would delay the phaseout of thetextile and apparel quotas required under theWTO Agreement on Textiles and Clothing.Nonetheless, this seems confined to a handfulof relatively small countries heavily dependenton a narrow range of preference-benefitingproducts.

An illuminating example of how prefer-ences can be a stumbling block to multilateralliberalization is provided by rum. Low-valuedbottled and bulk rum is one of the largest ex-ports from several Caribbean countries to theUnited States. It enters the United States duty-free under the Caribbean Basin Initiative. In1994 there was a single tariff line for rum in theU.S. schedule (HS 22084000) with a tariff of37 cents/liter. As result of WTO discussions,the United States and EU in 1996 agreed tophase out all tariffs on rum and other “whitespirits” by 2000. Caribbean governments wereconcerned that this would be costly toCaribbean exporters because they would haveto compete with the rest of the world inthe U.S. market. In response to this concern,the United States agreed to substantially liber-alize the duties on expensive rum but to main-tain duties on low-valued rum. As of 2003, theU.S. schedule now has four tariff lines for rum,two for high-valued rum with no MFN tariffcharged, and two for low-valued rum with anMFN tariff of 23.7 cents/liter.12 Since then, theimports of rum from the affected Caribbeancountries have fallen sharply, apparentlyreplaced by imports from Mexico—another bi-lateral RTA partner. So efforts to prevent ero-sion of preferential access may have only lim-ited the impact temporarily.

An increase in RTA activity may be associ-ated with multilateral trade negotiations.

Mansfield and Reinhardt (2003) argue thatmultilateral trade negotiations motivate coun-tries to conclude RTAs; the motivation is forincreased negotiating power and the desire toobtain or maintain preferential access to mar-kets. They also note that as WTO member-ship expands, larger membership reducesindividual countries’ ability to influence thecontent and pace of MFN liberalization andmakes it more difficult to formulate coordi-nated positions. Finally, they find that coun-tries use RTAs to increase leverage againstthird parties with which they are embroiled ina WTO dispute. In their empirical analysisthey conclude that countries are more likelyto form RTAs when: (1) GATT/WTO mem-bership rises; (2) a multilateral round is tak-ing place; and (3) parties have recently beeninvolved in a GATT/WTO dispute in whichthey lost.

Others have argued that RTAs are a mech-anism to enhance the pressure to move on themultilateral front, and they act as laboratoriesfor international cooperation on behind-the-border policy issues. This line of reasoning hasa long history. Winham (1986) and Lawrence(1991) have both argued that the creation andsubsequent expansion of the European Eco-nomic Community (EEC) motivated earlierGeneral Agreement on Tariffs and Trade(GATT) rounds—the objective being to reduceEuropean external protection. More recently,Schott (2004) noted that the United States haspursued bilateral trade agreements over thelast two decades in part to complement andcajole progress at the multilateral level. Heargues that tensions from the GATT meetingsin 1982 were the impetus to pursue bilateraldeals with Canada and Israel; he also claimsthat the start of the NAFTA negotiations in1991 reflected some frustration over failureto conclude the Uruguay Round negotiationson time in late 1990. In the early 1990s, theUnited States also began to pursue broaderumbrella regional initiatives, including AsiaPacific Economic Cooperation (APEC)—whose members have never formally commit-ted to a binding RTA—and the FTAA.


135

In short, there is only limited evidencefrom the theoretical literature and a handfulof empirical studies that the proliferation ofRTAs affects multilateral liberalization eitherway. On balance, the evidence does point totactical behavior in trade negotiations, whichseems to provide mild additional incentivesfor greater liberalization. The exception isfor small countries that suffer from preferenceerosion—a nontrivial obstacle to futureliberalization.

The Competitive LiberalizationGame: The Case of Doha

The rapid proliferation of RTAs may haveaffected the negotiating dynamics of the

Doha Round, and it is unclear whether the in-crease in bilateral deals is in response to slowprogress in the Doha talks. Concerns over RTAproliferation played a role in the launch of theRound and came back full force after the fail-ure of the Cancun ministerial. Consider theactions of the major players in the global andregional game.

The U.S. pursuit of like-minded partnersWith the change in U.S. administration in2001 and the subsequent congressional ap-proval of the Trade Promotion Authority in2002, the pace of bilateral, regional, and mul-tilateral efforts increased. Then, in the absenceof progress on the Doha agenda in Cancun inSeptember 2003, the U.S. Trade Representa-tive (USTR) indicated that the United Stateswould pursue deals with “like-minded” part-ners, and the result has been an unprecedentedspurt of U.S. negotiating activity. The UnitedStates had negotiated RTAs with sixcountries/groups by 2004 (Australia, Bahrain,Central America and Dominican Republic,Chile, Morocco, Singapore), and anotherdozen are under negotiation, including SACU(Botswana, Lesotho, Namibia, South Africa,and Swaziland), Colombia, Ecuador, Panama,Peru, and Thailand. As discussed in chapters 2and 5, the United States is pursuing new rulesin key areas (investment, intellectual property,

and services liberalization)—areas wheremultilateral efforts have gone slowly or haltedaltogether—while developing countries seekmarket access.

Eligibility is not guided by a single criterionor ranking system (box 6.3). Political criteriainclude national security-cum-foreign policyconcerns (a major factor in the recent agree-ments and ongoing discussions with Arabcountries), while others reflect a mixture ofclassic market access goals and a desire to ex-port U.S. regulatory standards (for instance, inthe area of investment and IPRs—seechapter 5). The U.S. decision to pursue bilat-eral arrangements in Latin America also hasthe effect of putting pressure on Brazil andother countries, which are seen to be impedingprogress on the now-stalled negotiations onthe FTAA. Reaching these agreements doesnot appear to have reduced the United States’participation in the WTO negotiations, nor tohave had much effect on the content of itsnegotiating position.

Recent EU regional initiativesHaving been a leading (and early) player in theRTA game, the EU had established manyRTAs prior to the launch of the Doha Round(see chapter 2). Since then, it has pursued mar-ket access agreements with a few individualpartners (notably Chile and Mexico, withSouth Africa predating 2001) as well asMERCOSUR. The EU has preferred to negoti-ate with blocs of countries and has encouragedthe Mediterranean countries to sign agree-ments with one another; the EU made the RTAwith Gulf Cooperation Council (GCC) coun-tries conditional on the adoption of a commonexternal tariff by the GCC and supported thedevelopment of a web of bilateral RTAs be-tween Southeastern European economies. Therecent Economic Partnership Agreements(EPAs)13 were launched to replace the Cotonouagreement with the African, Caribbean, andPacific (ACP) group in a WTO-compatible setof agreements, and the EPAs are seen by the EUas a development-promoting vehicle ratherthan as a way to gain additional market access.


136

These agreements do not appear to have al-tered the EU’s approach to the Doha Round.After the conclusion of the Mexican andChilean bilateral deals, the EU had declared itwould negotiate no new trade agreements,save for the EPAs that were necessary to re-place the Cotonou agreement set to expire in2008. However, this stance may be short-lived. In September 2004, the incoming EUTrade Commissioner announced that hewould review this policy and consider launch-ing new negotiations.

Japan is a latecomer but moving quicklyUntil recently, Japan remained disengagedfrom the RTA trend and instead pursued avoluntary approach that centered on APEC.A distinguishing feature of APEC is that itfocuses primarily on exchanging informationand identifying good practices in the tradepolicy (and other) areas. While it also sets spe-cific targets for trade policy (e.g., free intra-APEC trade by 2020), implementation of the

good practices and targets are left to individ-ual members. The primary enforcement de-vices are mutual surveillance and peer review.More recently, it appears that competitivepressures that emerged from the more inten-sive activity by the United States—also anAPEC member—and the concern that even inthe Asia region there was a risk of being leftout of the new generation of RTAs, haveprompted new Japanese activity. Japan com-pleted the negotiation of its first RTA withSingapore in 2002. Negotiations with Mexicoare advanced. The competitive disadvantagein Mexican markets relative to the UnitedStates and the EU is a strong inducement, withJapanese products facing an average customsduty of 16 percent (Tojo 2004). Negotiationsare beginning with Republic of Korea, whichare considered important because Japan-Korea could form the basis of a future EastAsian economic zone. Talks have also begunwith three individual members of ASEAN(Thailand, Malaysia, and the Philippines), and


137

The U.S. choices of developing country RTA part-ners reflect a range of different objectives, which

makes categorization difficult. In a report preparedfor the U.S. Congress, U.S. General Accounting Of-fice (GAO) (2004) reports that early RTA proposalswere (according to the USTR) based on evaluation of13 different factors (without any formal weightingscheme) that covered five themes: Congressionalguidance, business interests, political will of partners,security and foreign policy concerns, and regionalparity. Following consultations with relevant agen-cies, there is now a shortened list of six factors toguide the choice of future U.S. FTA partners: countryreadiness (political, trade, and legal); economic/com-mercial benefit; benefits to broader trade liberaliza-tion strategy (including success in meeting WTOobligations and support of key U.S. positions inFTAA and WTO negotiations); compatibility withU.S. interests (including foreign policy positions);

Box 6.3 Choosing partners: Selection criteria for U.S. RTAs

congressional/private sector support; and U.S. gov-ernment resource constraints. The report emphasizesthat the selections are not mechanical and arguesthat trade strategy and foreign policy factors domi-nate the selection criteria.

Schott (2004) identifies the same broad criteria,but also notes the role of partner choice in selectionof U.S. FTA partners. Current U.S. law requires thatpotential partners must request negotiations with theUnited States, rather than the other way around. Henotes that most of these requests reflect concernsover discriminatory treatment (from other agree-ments, especially NAFTA) and serve to demonstratecommitment to domestic audiences who may need tobe convinced of the benefits of reform trade anddomestic policies.

Source: Schott 2004.

consultations have been initiated over how tomove forward with a proposed Comprehen-sive Economic Partnership between Japan andthe full ASEAN group. Finally, there are pre-liminary discussions and analysis of a possibleASEAN�3 RTA (including China, Republicof Korea, and Japan) in the context of effortsto strengthen economic relations among thisgroup.

Developing country objectives andnegotiating strategiesDeveloping countries have three general ap-proaches to RTAs. Some have adopted an ag-gressive approach and pursued a serial RTAstrategy; that is, they negotiate a string ofagreements and use the sequence to demon-strate their commitment to trade reforms bylocking these in and increasing the incentivefor excluded countries to negotiate. Harrison,Rutherford, and Tarr (2002) have labeled thisnegotiating dynamic “additive regionalism.”Its most prolific proponents include Chile,Mexico, and Singapore, which have pursuedRTAs with most of their geographic neigh-bors, as well as with many of the other majorplayers (Schott 2004). The idea is that negoti-ating additional RTAs will progressively lowerthe effective average tariff (reducing potentialtrade diversion costs) and assure stability ofmarket access for partner countries. A second,much larger group of countries has pursued astrategy more explicitly regional in focus,which seeks to deepen ties with neighboringcountries; examples include ASEAN, theGCC, MERCOSUR, and SADC. A thirdgroup has focused on negotiating North-South RTAs, often in parallel with their re-gional integration efforts with neighbors;examples are the southern Mediterraneancountries with the EU and the US-CAFTAagreement. The ACP-EU Economic Partner-ship Agreements are another example.

Additive regionalism could create adverseglobal effects by reducing the incentives forcountries to participate in multilateral liberal-ization efforts. That has not been the case withChile, Mexico, or Singapore. The fact remains

that these countries still have incentives to seelower barriers in the many countries withwhich they do not have RTAs and to harnessthe multilateral process to achieve movementin areas outside of the RTAs, such as agricul-ture and anti-dumping. The WTO provides aforum to achieve these objectives at a lowercost than negotiating a plethora of bilateralRTAs. Countries pursuing such strategies arevery much a minority—most developing coun-tries are not in this game (box 6.4). Chile,Mexico, and Singapore are all economies thatmoved substantially toward free trade andthus have already captured most of the poten-tial gains from unilateral trade reforms. Froma global point of view, therefore, these are notcountries that have a vested interest in main-taining high MFN barriers. These countriesare examples of open regionalism.

The determinants of the multilateral nego-tiating stance of the broader group of devel-oping countries are more varied. For a hand-ful of countries, existing preferences underunilateral accords as well as multilateral andregional agreements are openly driving theiropposition to multilateral liberalization. Forsome other developing countries, it is possiblethat RTAs are undermining their interest inmultilateral negotiation, either because theymistakenly see RTAs as an alternative, becausethey feel subsumed in large coalitions withother countries, or because they do not havethe resources to negotiate with both the WTOand with potential regional partners.

This discussion suggests a few impression-istic conclusions. It is hard to argue that com-petitive liberalization through RTAs has muchinfluence on the behavior of the major playersin the WTO, either in their fundamental nego-tiating positions or their willingness to com-promise to achieve a result. As evident in theJuly WTO Framework Agreement, majorplayers are still working with commitmenttoward an agreement in Doha. To be sure, forsmaller countries with scarce negotiating ca-pacity, RTAs do absorb resources that couldbe devoted to multilateral negotiations; butthese countries tend to participate in the WTO


138

as members of coalitions, and it is not clearthat these constraints impede compromise. Ifthe effects on the multilateral round are negli-gible, it does seem that all players appear to bequickening their efforts to seek new preferen-tial agreements. The outcome of the game ofRTA-based competitive liberalization is likelymore RTAs.

One negative incentive effect created bypreferences—reciprocal RTAs or voluntaryprograms—is that members’ desire to safe-guard their preferential access to the regionalmarket may result in less support for MFN-based trade reforms. This incentive effect haslong been recognized; it was one of the argu-ments made against the Generalized System of


139

Debates over trade liberalization have often in-cluded extensive discussions of whether there is

a preferred (or even optimal) sequence for reforms.Theoretical guidance for policymakers in the designof RTAs speaks to two dimensions of the question—the choice (and sequence) of partners over timeand the substantive coverage of an RTA (and thesequencing of inclusion of different elements)—thathave implications for the debate. On the former, thesequencing of partners may affect incentives of mem-bers to pursue MFN liberalization. Thus as discussedelsewhere in this report, the formation of large blocscreates incentives to join for smaller countries thattrade heavily with members—what has been termed“domino regionalism” (Baldwin 1993). The EU is thebest example of this phenomenon: EFTA countriesthat were not in favor of the EU integration model ul-timately concluded that the costs of staying out weretoo high. Whether this process makes RTAs more orless receptive to MFN reforms depends in part on thepreferences and interests of those who join the blocover time—which in turn will affect the willingness ofincumbent countries to accept new members.

On the product/policy coverage issue, the conven-tional wisdom appears to be that agreements shouldfirst focus on trade liberalization and then move onto behind-the-border areas—that is, go from shallowto deeper integration. There is no theoretical justifi-cation for this, however, and there is a well-documented history that, in the case of the EEC,many policymakers were of the view that the twoneeded to be pursued in tandem. The rhetoric ofpolicymakers and their advisors often suggests thatdeeper integration is necessary to attain free trade.During the period leading to the creation of the EEC,

Box 6.4 Sequencing of RTAs:Is there a good practice?

Jelle Zijlstra, the Dutch Minister of Economic Affairsargued that credible tariff removal required commonpolicies on taxes, wages, prices, and employmentpolicy. Similarly, the Belgian government felt thatpolicy harmonization was required to equalize costs,and that without it, a customs union would not befeasible because countries would impose new formsof protectionist policies. French officials persistentlydemanded harmonization in social policies—equalpay for both sexes, a uniform work week—as a pre-condition for trade liberalization—French standardsin this area were higher than in other countries(Milward 1992).

Recent research on the effects of, and interplaybetween, efforts to liberalize trade and investment inservices (and FDI more generally) suggests that coun-tries may be better of pursuing both shallow anddeeper forms of liberalization in tandem. Hoekmanand Konan (2001) and Konan and Maskus (2003),for example, note that not only can this generatemuch greater welfare gains, it can also reduce aggre-gate adjustment costs over time—through avoidingoutcomes in which factors of production must moverepeatedly across sectors (as will, by definition, occurif goods are liberalized first and then services/invest-ment, or vice versa). They also note that becausemany services continue to be less tradable thangoods, there is greater scope for employment oppor-tunities to be created as a result of allowing greatercompetition in services markets, thus helping to ab-sorb labor from other sectors as prices change due totrade reform.

Source: World Bank staff.

Preferences when they were initially proposedin the 1960s. Indeed, unilateral preferenceprograms, especially the more comprehensiveand meaningful schemes adopted in recentyears, such as the EU’s Everything But Arms(EBA) program and the United States’ AfricanGrowth and Opportunity Act (AGOA), po-tentially make matters worse, because theirvalue to recipients depends on the existenceof trade barriers against imports from othercountries—with the greatest rents generatedby programs that distort markets the most,such as in the sugar market. Maintaining pref-erence margins—whether for RTA partners orbeneficiaries under unilateral preferenceprograms—is not the answer. What is neededis a willingness on the part of developed coun-tries to move away from preferential access asan instrument to assist lower-income partnercountries and to move toward greater relianceon direct transfers of technical and financialassistance (Stoeckel and Borrell 2001). Thishas the advantage of allowing high-incomecountries to target their trade-related develop-ment assistance to those countries most inneed, something the system of preferencescannot do.

Multilateral Disciplines onRegional Arrangements

Efforts to deal with the implications ofRTAs within the multilateral trading sys-

tem are long standing. The primary disciplinesare laid out in Article XXIV of the GATT—others are discussed in box 6.5. They permitRTAs if: (1) external trade barriers do not rise(Article XXIV:5); (2) all tariffs and other regu-lations of commerce are removed on substan-tially all exchanges of goods between thepartner countries within a reasonable length oftime (Article XXIV:8); and (3) notification ismade to the WTO Council.

The first criterion is intended to limit thenegative impact of an RTA on nonmembers.14

The second condition might at first seemcounterintuitive—after all, the more extensiveis the liberalization with an RTA, the more

detrimental it is likely to be to nonmembers.But the rationale here was in fact different; asnoted by Finger (1993), the objective was toensure that participants in RTAs are serious.In other words, while more trade in an RTAhurts nonmembers, pursuit of numerous par-tial RTAs can severely undermine the incen-tives for multilateral trade negotiations andcreate opportunities for special interests(farmers, specific industries) to carve out spe-cial arrangements. As a counterexample, if arestrictive interpretation (say 99 percent of alltrade) of this criterion were agreed to and en-forced, it is likely that the appetite for RTAswould be substantially diminished.

Determining whether the GATT or GeneralAgreement on Trade in Services (GATS) testsare met is left for the Committee on RegionalTrade Agreements (CRTA). Before the cre-ation of the WTO, the GATT Council usuallycreated a working party to evaluate whetherits conditions were satisfied. Under the WTO,the CRTA was established for this task. TheGATT experience in testing reciprocal prefer-ential trade agreement (PTAs) against ArticleXXIV was very discouraging. Starting withthe examination of the treaty establishing theEEC in 1957, almost no examination of agree-ments that were notified under Article XXIVled to clear conclusions or specific endorse-ments that the GATT requirements had beenmet. Only one working party could agree thata regional agreement fully satisfied the re-quirements of Article XXIV (the Czech-SlovakCustoms Union).

There had been a conscious political deci-sion made by GATT contracting parties in thelate 1950s not to scrutinize the formation ofthe EEC because the EEC member states hadmade it clear that if the EEC treaty was foundto be inconsistent with Article XXIV, theywould withdraw from GATT (Snape 1993).Given that the EEC did not meet all the re-quirements of Article XXIV—agriculturaltrade was not liberalized—it created a prece-dent. It is also true that the criteria and lan-guage of Article XXIV are ambiguous. Legiti-mate differences of opinion are possible


140

regarding issues such as the definition of“substantially all” trade; how to determinewhether the external trade policy of a customsunion has become more restrictive on average;and what a reasonable time period is for thetransition toward full implementation of anRTA (Hoekman and Kostecki 2001).

In the Uruguay Round some of these issueswere clarified: Specific criteria were adoptedto assess whether a customs union’s externaltariff satisfies Article XXIV; a 10-year maxi-mum for the transition period for implemen-tation of an agreement was imposed; and, asmentioned, a standing committee was createdto oversee RTAs. None of these changes hadany effect on the ability of WTO members toagree on whether an RTA conformed to WTOrequirements. Most observers would agreethat existing WTO disciplines and enforce-ment mechanisms have no teeth, and are not

particularly effective at controlling, limiting,or shaping the growth and coverage of RTAs.At the Fourth Ministerial Conference in Doha,ministers agreed to launch negotiations toclarify and improve the disciplines and proce-dures under the existing WTO provisions thatapply to RTAs, taking into account the devel-opmental aspects of these agreements.

The Doha Agenda negotiations on rules forpreferential trade agreements have been con-ducted on two tracks, one focusing on trans-parency issues and the other addressing thesubstantive disciplines. Transparency is gener-ally less contentious, and includes:

• Administrative overload in reviewingproposals. Review of long, detailed, andoften unclear RTA documents imposesa substantial workload on committeemembers. One possible solution would


141

The fundamental building block of the multilat-eral system is the principle of nondiscrimina-

tion. This is enshrined in Article I of the originaltext of the GATT signed in 1947. But from thestart, there have been exceptions. One such provi-sion is Article XXIV of the GATT, which permitsWTO members to enter into preferential tradeagreements (RTA) that extend trade concessions toRTA participants not offered to nonparticipants, aslong as certain criteria are satisfied—in particular,regarding the scope of the RTA. Second, rules cov-ering preferential agreements that deal with tradein services are set out in Article V of the GATS.Finally, developing countries may, if they wish, invokethe provisions of the 1979 Decision on Differentialand More Favourable Treatment, Reciprocity andFuller Participation of Developing Countries (the so-called Enabling Clause) to exchange tariff and, to acertain extent, nontariff preferences among them.Unlike Article XXIV, the Enabling Clause does notrequire that internal barriers be removed on “sub-stantially all” trade among participants in thosearrangements. MERCOSUR was notified to GATTunder this provision, not under Article XXIV. The

Box 6.5 RTAs and WTO disciplinesEnabling Clause also legitimizes non-reciprocal pro-grams such as the Generalized System of Preferences(GSP).

The task of verifying the WTO compliance ofRTAs notified under GATT Article XXIV and GATSArticle V is entrusted to the CRTA. The CRTA wasestablished in 1996, in particular to (1) oversee,under a single framework, all regional trade agree-ments, and (2) consider the implications of suchagreements and regional initiatives for the multilat-eral trading system and the relationship betweenthem. RTAs among developing countries, when noti-fied under the Enabling Clause, are not, in principle,subject to review by the CRTA. As was the caseunder the GATT, however, the CRTA has been un-able to carry out effectively its functions of examin-ing the consistency of RTAs with the rules, and over-seeing their implementation. The reason for this isessentially a fear of setting a precedent and openingup agreements to dispute settlement proceedings. TheCRTA has thus far been unable to conclude the ex-amination of any of the 110 RTAs currently underscrutiny and has a backlog of about 35 RTA reportsto prepare.

be an expanded “first review” by WTOstaff based on clear and objective criteria;this would require a significant expansionof resources allocated by the Secretariat.

• Notification of RTAs to the WTO. TheWTO has been notified of around 300agreements (of which about 140 are cur-rently in force), and another 60 or so arein advanced stages of negotiation (seechapter 2). Many of these are for RTAsamong developing countries. Moreover,few RTAs have been designated as“interim arrangements,” even thoughmost RTAs have been implemented instages.

• Data and information requirements.There is no clarity on the type, quantity,and level of detail to be provided whenthe WTO is notified of an agreement.Developing countries often lack thecapacity to undertake this task. Whilesome members want more detail, it is un-derstood that unduly straining the capac-ity in developing countries should beavoided. Some suggest that WTO assis-tance could be provided for the notifica-tion process if necessary.

• Review of South-South agreements. TheEnabling Clause does not provide forany consistency examination. Currentpractice is for RTAs formed between de-veloping countries (and notified underthe Enabling Clause) to be reported tothe Committee on Trade and Develop-ment (some developing countries havechosen to notify the WTO of agreementsunder Article XXIV). Countries haveraised the issue of reporting EnablingClause Agreements to a single body, theCRTA.

• Services. Some of the requirements of theprovisions of Article V of GATS are un-clear (e.g., in terms of the departure fromMFN obligations in key areas such astransparency, emergency safeguards, andadministration of domestic regulations).

• Process. Some delegations want notifica-tion of RTAs to occur before they are

implemented. Others support issuing afinal report that does not pass or fail aRTA, but allows members to registerconcerns. Finally, some delegations wantgreater diligence in encouraging RTAparticipants to file biennial reports onthe implementation of RTAs.

Although no early harvest on transparencyissues was undertaken for the 2003 CancunMinisterial, negotiations for a (provisional)application of strengthened surveillance mech-anisms are considerably advanced. Suchmechanisms would require more detail in theRTA notification procedures and might in-volve an enhanced role for the Secretariat inpreparing an assessment of each RTA thatwould be provided to WTO members.

Informal discussions on the substantiverules began in June 2004, and there are manyissues under consideration. The fact thatWTO rules on RTAs relate to several otherregulatory areas (some of which are undernegotiation)—including rules of origin, tradefacilitation rules on trade remedies, theGATS—adds to the complexity. Issues underconsideration include:

• Product coverage of RTAs. One centralissue is whether to make more specificthe requirement that “substantially all”trade be covered in each RTA. Lack ofclarity regarding this criterion is viewedby many as a source of the CRTA’s in-ability to reach clear-cut conclusions onWTO compatibility for most RTAs.Many RTAs exclude agriculture, which isa major problem. Some have proposedthat “substantially all” trade should bedefined as a percentage not only of ac-tual trade but also of all the six-digit tar-iff lines listed in the Harmonized System.This approach could ensure that thestandard is set high, but that there is alsosufficient flexibility to set aside productareas that remain sensitive for one rea-son or another. A related issue concernshow (if at all) such a criterion would be


142

applied to RTAs already notified—wouldthey be subject to the same percentage,or grandfathered in under the existingimprecise criterion? Although inherentlyarbitrary, a more specific coverage crite-rion could help to move the reviewprocess for RTAs along. Schiff and Win-ters (2003) suggest adopting a coveragecriterion of 95 percent of the value oftrade after 10 years of operation of theagreement, rising to 98 percent after 15years. This would require that at leastsome sensitive products be included (e.g.,agriculture), but not all, and wouldimply that over time the set of excludedproducts would decrease.

• Policy coverage of RTAs. GATT ArticleXXIV.5 requires that duties and otherregulations of commerce applied bymembers of a RTA are not more restric-tive than those existing prior to its for-mation. In assessing the impact of aRTA, there is no agreement on what thiscovers. In particular, it is unclear to whatextent it covers policies such as safe-guards, anti-dumping measures, mutualrecognition agreements, or rules oforigin—and if they are covered, howthey should be evaluated. Many RTAscontinue to allow for the use of an-tidumping and safeguards on intra-member trade—that is, conflict with thesupposedly free trade objective of RTAs.Only a few RTAs have abolished thereach of antidumping (the EU,ANZCERTA, Canada-Chile). Alterna-tively, some RTAs preclude the use ofsafeguards on goods originating in part-ner countries—for example, Canada andMexico were exempted from the recentU.S. steel safeguard action. This in-creases the negative effect of the actionfor nonmembers.

Whether rules of origin are a regulation ofcommerce has been a key source of disagree-ment for decades. The rules of origin are rele-vant not just for normal trade flows, but also

play a role in the application of safeguardsand other contingent trade policies. Issues re-lated to rules of origin were already being dis-cussed in the early 1970s in working partiesthat were considering RTAs. Thus in connec-tion with the 1972 free trade agreement be-tween the EEC and European Free Trade inEurope (EFTA) member states, the UnitedStates argued that the rules of origin wouldharm nonmembers.15 Not surprisingly, prefer-ential rules of origin are being discussed in thecurrent Doha talks, although expectations arelow that there will be agreement on commondisciplines. Discussions on harmonization ofnonpreferential rules of origin have beenunder way for almost 10 years and have yet tobe concluded.16 In any event, these harmo-nized rules of origin will not apply to regionalagreements or to GSP schemes. Because rulesof origin facilitate the fine tuning of preferen-tial liberalization at the product level, manycountries do not want to see constraints im-posed on their policy freedoms in this arena.

Many proposals have been made in theWTO for stronger rules for RTAs. One sug-gestion is that all RTA members be required toextend their preferential concessions to therest of the world within a specific time frame(Srinivasan, 1998).17 Another suggestion is tominimize the adverse discriminatory conse-quences of RTAs by requiring (or exhorting)members to allow any developing country to“opt in” on the same terms as existing mem-bers, perhaps after a certain time period. Thisgoes back to Viner (1950). As noted, virtuallyevery existing RTA has geographic restrictionson membership and has features that requirenegotiation, so the practical promise of suchopen regionalism is limited. Other suggestionsdo not make good economic sense (see Schiffand Winters, 2003).

Many observers have concluded that thequest for stronger rules is unlikely to succeedbecause many RTAs will not satisfy the rules,which will, in turn, lead countries to prefer thestatus quo. Schiff and Winters (2003) con-clude their discussion of regionalism and theWTO with a section called “Rules Are Not the


143

Answer.” From this perspective, the primaryfunction of the WTO is to act as a negotiatingforum to bring down the discrimination cre-ated by RTAs. The importance (and feasibil-ity) of this depends, in part, on the motivationof governments to pursue RTAs.

Making Open Regionalism Workfor DevelopmentDefining the role of the WTONumerous observers and analysts of RTAs,who are interested in enhancing their compat-ibility with the global trading system, proposethat WTO disciplines be applied more strin-gently. Years of discussion in the GATT/WTOon the interpretation of existing criteria, andthe many papers proposing more specific cri-teria have had no impact on the spread or con-tent of RTAs. Improving the enforcement ofArticle XXIV or strengthening/changing WTOdisciplines on regionalism is unlikely to fareany better.

Whether it is helpful to articulate more spe-cific criteria delimiting what “substantiallyall” trade means, or tightening up disciplineson “other regulations of commerce,” dependsin large part on the feasibility of attaininga consensus on specific criteria. Given theplethora of RTAs that are in force, and theshare of global trade notionally coveredthrough such agreements, a strong case can bemade that the horse has already bolted fromthe barn—shutting the barn door will makelittle difference. The problem is a political one;any number of countries will oppose strongerrules because they are already members ofRTAs that would violate them. Indeed, even ifgroups of developed countries—such as theEU—make a case that their agreement satisfieswhatever criteria might be proposed, or thattheir model should become the benchmark,other countries could legitimately argue that ifit was acceptable for the EEC to pass musterin the 1960s and 1970s, it would be hypocrit-ical to impose stronger rules on all WTOmembers today.

Nonetheless, WTO members should lookfor ways that existing disciplines on RTAs canbe reinforced and new disciplines introducedto enhance their development impact. Onearea where knowledge is much more extensivenow, compared to when the original GATTrules were written, regards rules of origin. It isclear that complex and restrictive rules of ori-gin limit the benefits of RTAs for developingcountry participants and divert trade in inter-mediate products. Rules of origin that differacross agreements complicate world tradingconditions and contribute to the emergence ofhub-and-spoke patterns. Thus there would besubstantial benefits from an agreement thatpromoted common, simple, less restrictive,and easy-to-apply rules of origin. That beingsaid, the delays and problems in achieving theobjectives defined in the Uruguay Round forthe nonpreferential rules of origin are not pro-pitious in this case.

From a development perspective, the mostuseful and immediate step for the WTO is toimprove transparency. Information and analy-sis are important inputs for a well-functioningtrading order. Greater monitoring and assess-ment of the impacts of RTA-related policieswould allow more informed and proactive en-gagement by civil society (think tanks, non-governmental organizations, consumers, andtaxpayers) in the policy formation and negoti-ation process. It is true that to reduce protec-tion and protectionist pressures, those thatlose (pay) need to be aware of the costs ofsuch policies. The suppliers of and the clientsfor such analysis and information are not nec-essarily governments, but the constituencies inindividual countries who are affected by pol-icy. In order for trade agreements to promotegood policy-making in member countries,stakeholders must be able to be active in thedomestic policy formation process.

This could be achieved by augmenting thecapacity of the WTO Secretariat andthe CRTA to review, document, and analyzethe effects of RTAs. That is, WTO attention inthis area should focus on gathering informa-tion and analysis. Efforts could concentrate on


144

addressing such questions as: Are RTAs beingimplemented? How? What is the effect onmember countries and on nonmembers? Howmuch trade is covered by the RTA (and is it“substantially all”)? Ideally, stronger surveil-lance would involve those responsible forimplementing each RTA and, in the process,empower them to engage policymakers andstakeholders. Such surveillance and analyticalmonitoring should extend to South-SouthRTAs—which should all be notified to theCRTA.

Strengthening information exchange andmutual (multilateral) dialogue, (includingthrough the establishment of formal monitor-ing mechanisms), would facilitate cooperationon new regulatory issues. If existing or pro-posed policies could be evaluated using objec-tive criteria on their ability to achieve thestated national objectives, countries andstakeholders could assess their efficacy and, ifneeded, adjust the policies. Greater analysis ofthe effects of trade discrimination—both uni-lateral preferences and RTAs—should be partof this agenda.

Information of this type would help RTAstakeholders to hold governments accountablefor outcomes and to assist nonmembers byproviding data that could feed into demandsin the context of WTO negotiations.18 Asargued by Hoekman and Kostecki (2001)among others, the primary means throughwhich the WTO can impose limits on RTAs isby providing a venue for negotiation to reduceMFN barriers, which will automatically limitthe discrimination against outsiders that is in-herent in RTAs.

Priorities for the industrial countries The major industrial countries must strike adifficult balance in their pursuit of bilateraland regional deals with developing countries.On the one hand, there are legitimate reasonsto pursue such initiatives, rather than relyingonly on multilateral channels. The ongoingEU effort to deepen relations with a widerEurope is an important example. Cooperationamong a smaller number of countries may

also be the most effective solution to mitigat-ing environmental externalities, addressingnontrade issues such as labor migration (legalor illegal), or attaining national security objec-tives. But it must be recognized that the ag-gressive pursuit of RTAs with developingcountry partners in all corners of the globeserves the cause of global liberalization poorlyif it delays or halts altogether the progresstoward multilateral liberalization.

Fostering development is increasingly iden-tified as a key rationale for North-SouthRTAs. In part this reflects economicinterests—growing markets abroad are ex-panding export markets—and in part it is dueto the recognition that there is a correlation be-tween sustained economic growth in partnercountries and national security. From this van-tage point, several recommendations emerge.

The highest priority should be to ensurethat the Doha Development Agenda is com-pleted in a manner that provides new marketaccess to exporters in developing countries. Inaddition to their interests in development, thelarge countries have an important historicaland systemic responsibility to ensure that theworld trading system remains as open aspossible.

Supporting open regionalism by encourag-ing partner countries to adopt low externalbarriers would create momentum towardintegration with the world, establish a moreefficient development path, and reduce theadverse negative implications of RTAs on non-members. Because these countries are lesslikely to be preferred RTA partners of theUnited States and the EU, they will continue tohave an incentive to engage at the WTO level.

Adopting the widest possible product cover-age and greatest market access expansionwould increase the development impact. Thismeans that agricultural trade policies should beincluded in RTAs. Excluding agriculture fromRTAs—the dominant practice at present—doesnot promote development. Although inclusionof agricultural market access will increase theincentives for small countries to seek RTAs withthe large players, it will also increase the


145

downside of RTAs for large agricultural pro-ducers in the developing world. The latter willthen have a greater incentive to push for WTO-level MFN reforms; the marginal increases incompetition from preferred partners may helpovercome the domestic interests opposed to re-forms through a process of gradual, piecemealexpansion of access to agricultural markets.

Industrial countries negotiating North-South RTAs should adopt liberal cumulationprovisions in their rules of origin. Because the

least-developed countries (LDCs) already havenearly free access to OECD markets, North-South RTAs will erode such preferences. Cu-mulation will ensure that rules of origin donot impose an additional burden on thesecountries. For member countries, liberal rulesof origin will enhance the benefits of North-South RTAs. A demonstrated willingness ofindustrialized countries to put partner countryinterests before those of national industrygroups (who prefer restrictive rules of origin)


146

In July 1995, Tunisia signed the Association Agree-ment with the European Union (AAEU). The agree-

ment, which came into effect in March 1998, wouldliberalize trade for industrial goods over a 12-yearperiod and ultimately create a free-trade zone. Trade inagricultural goods and services was left out for futurenegotiations. By the mid-1990s, Tunisia had alreadybecome a successful exporting country, thanks to theestablishment, in 1972, of a special offshore system forexporting enterprises that mitigated the anti-exportbias of the highly protective trade regime.* Exports ofmanufactured goods had increased from 4 percent ofGDP in 1975 to 20 percent by 1994. Free market ac-cess to EU markets under the AAEU further enhancedTunisia’s export performance, with manufactured ex-ports increasing to 25 percent of GDP in 2002.

The AAEU gave momentum to trade liberaliza-tion in Tunisia. Average MFN tariffs were reducedfrom 33 percent in 1994 to 26 percent in 2003.However, Tunisia still posts the second highest aver-age MFN tariffs in the broader EU neighborhood,including in the EU accession countries. For exam-ple, only Morocco, with a tariff rate of 30 percent,has a higher border barrier, but countries as diverseas Turkey, Bulgaria, Ukraine, Lebanon, and Moldovahave tariffs of 10 percent or lower.

High MFN tariff differentials in the presence ofEU preferential access risks diverting trade awayfrom the lowest cost sources, denying Tunisian pro-ducers and consumers the benefit of less expensiveimports from outside the preferential trade zone withthe EU. As a result of high MFN tariffs and geo-graphical proximity, about 75 percent of Tunisia’s

Box 6.6 Tunisia’s Association Agreement with theEuropean Union

trade is dependent on the European market—especially in EU neighborhood countries.

A trade strategy linked to the domestic reformagenda would help Tunisia fully realize the develop-ment promise of deeper trade integration. First, am-bitious reduction of MFN tariffs for industrial goodswould prevent trade diversion. Second, removingbeyond-the-border obstacles to trade—by reducingthe still-high trade logistics costs—would enhancefirms’ ability to exploit export opportunities and im-prove their competitiveness; liberalization of back-bone services, especially in transport, ICT, and fi-nance, are also essential. Third, once the free-tradezone with the EU is fully implemented in 2007, pref-erential tax treatment of exporting firms will becomemuch harder to justify, because both offshore and on-shore firms will be equally exposed to foreign compe-tition. The regulatory framework of investment in-centives and trade facilitation will thus have tobecome more even. The EU could also liberalize mar-ket access for Tunisian agricultural exports in prod-ucts in which Tunisia is competitive. Unless marketaccess is improved, the scope for farmers to shift intothese products will remain limited, and sectoral ad-justment in agriculture will continue to be impaired.

*This system covers companies located anywhere in thecountry and it grants duty-free imports of capital and intermedi-ate goods, exemptions from the VAT and excise taxes, and ex-emption from the corporate income tax for the first 10 years ofoperation.

Source: World Bank staff.

would show that development objectives arebeing taken seriously in RTAs.

Industrial countries could also enhance thedevelopment credentials of their RTAs by tak-ing action to abolish antidumping and similarinstruments of contingent protection. There isa plethora of evidence that antidumping isstraightforward protectionism and that inso-far as there is a rationale for intervention,other policy instruments can be used (i.e.,competition policy). A number of RTAs—ANCERTA, Canada-Chile, the EU itself—have proceeded down this path, which illus-trates that it is feasible.19

Industrial countries should exercise cautionregarding their demands for new regulatorypolicies in RTAs. Behind-the-border, regula-tory policies are critical for a positive impacton development outcomes, but getting the rulesright—ensuring that rules are calibrated to de-velopment capacities and do not detract fromother, more pressing priorities—is as essentialas it is difficult to orchestrate. Applying regu-latory norms in RTAs on a nondiscriminatorybasis will avoid creating another complex setof discriminatory preferences. In other words,RTAs will be supportive of development if thenegotiation and implementation process is de-signed to ensure that such priorities are set ap-propriately, and the preconditions for benefit-ing development are in place. Many RTAs arefar from satisfying this prescription.

Increasing the effectiveness of developmentassistance for trade can help. Aid is an impor-tant part of the equation motivating RTA ne-gotiations, especially for agreements with theEU. Export growth in many LDCs and othersmall and low-income countries is limited bythe lack of supply capacity and the high-costbusiness environment. Firms in these countriesmay also find it difficult to deal with the regu-latory requirements that apply in exportmarkets. Health and safety standards, for ex-ample, are often regulatory barriers to entry;the standards can be excessively strict and thecompliance costly, which weighs dispropor-tionately on producers in low-income coun-tries. Development assistance can help to build

the institutional and trade capacity needed tobenefit from increased trade and better accessto markets. This assistance will be more effec-tive when it is focused more broadly on supplycapacity, and when it addresses the adjustmentcosts associated with reforms. Recently atten-tion has been given to expanding programsthat provide aid for trade, but only when tradeissues are integrated into a nation’s overall de-velopment priorities. Although priorities willdiffer, in many cases assistance will be neededto address trade-related policy and public in-vestment priorities. More could be done to re-place preferential access as the primary carrotfor RTAs with financial transfers. It hasbeen argued at length (e.g., Hoekman,Michalopoulos, and Winters 2004), that tradepreferences should not be a permanent featureof the global trading system. Appropriatelydesigned aid would offer a similar result topreferences and at a lower overall cost.

Challenges for developing countries: Athree-pronged strategy Developing countries would benefit fromadopting a coherent three-part strategy thatintegrates unilateral, multilateral, and regionalinitiatives. A number of middle-income coun-tries have enunciated a clear set of prioritiesand objectives regarding regional and multi-lateral efforts, and have the technical and ne-gotiating capacity to pursue them; however,many developing countries show less evidenceof a coherent strategy on how to use RTAs tomaximum advantage.

At the multilateral level, the Doha Agendanegotiations are the best instrument for mostdeveloping countries to reduce the discrimina-tion they face from the prevailing web ofRTAs. Doha is also critical insofar as it pro-vides the greatest potential new market accessfor the greatest number of the world’s poor. Itis also the only venue in which key policyareas such as agricultural support policies orantidumping can be negotiated in a compre-hensive and substantive manner. As with thehigh-income countries, completing the Dohadeal is the highest priority.


147

Unilateral reforms can increase the benefitsof RTAs. Governments should generally lowerexternal (MFN) barriers to trade in conjunc-tion with the reduction of trade barriersagainst partners and build this commitmentinto the agreements signed. This not only re-duces the scope for detrimental trade diver-sion, it also helps to support multilateralismand increase global welfare. Such commit-ments can be made in the context of the on-going Doha Round, with the advantage of po-tentially generating additional quid pro quoreductions by other WTO members. Thatsaid, strategies that deliberately delay tradeliberalization and domestic reforms in orderto conserve bargaining chips for either RTA orDoha Agenda negotiations are misguided.Most developing countries have only limitedleverage within any such negotiations, and thecosts of delayed reforms outweigh any transi-tory negotiating advantage. The rapid growthrates realized by countries such as India fol-lowing trade liberalization and relatedreforms in the 1990s illustrate this point.

Pursuing bilateral or plurilateral North-South RTA arrangements at the regional levelmay yield short-term market access benefits toparticipating developing countries. But coun-tries should be aware that preferential access islikely to be eroded as more countries sign suchdeals, reducing the value preferences, if not theaccess. Moreover, the impact of commitmentson nontrade regulatory policies must be takeninto account. As noted above, it is importantto get the rules right. For RTAs to be beneficialin the longer term, there needs to be mecha-nisms through which governments and stake-holders are assisted in implementing a set ofreforms that will help sustain growth andreduce poverty. The establishment of these na-tional mechanisms is an important precondi-tion for benefiting from the RTA game.

South-South RTAs, especially regionalagreements among neighboring countries, cangenerate substantial benefits to participants innon-trade policy areas such as border cross-ings, infrastructure, standards and regulatoryframeworks, and related enforcement institu-

tions. Emphasis should be directed toward en-hancing cooperation in areas where there is astrong public goods dimension—collaborationhere can yield large gains. Such cooperationcan have positive spillovers on the multilateralprocess by helping to identify what type ofmultilateral rules might be beneficial.

Only by adopting a three-part integratedstrategy built around unilateral, multilateral,and regional components can developingcountries ensure that their trade policy willcontribute the most to growth and poverty re-duction. To be fully effective, each trade pol-icy lever must be used to its best purpose, andcoherent trade policy and associated technicalassistance resources have to be integrated intoa national development strategy. In the case oflow-income countries, the Poverty ReductionStrategy Paper (PRSP) is an example of such asinstrument. Governments and stakeholdershave an interest in building such a three-partstrategy into their development strategies. Inthis way, open regionalism can help countriesimprove their standards of living.

Notes1. For a summary of earlier World Bank research

on the linkage and compatibility of regionalism andmultilateralism, see World Bank (2000) or Schiff andWinters (2003).

2. The results reported herein refer to so-called“static” gains, that is, it is assumed that trade reformhas no impact on productivity. Thus the gains aremainly associated with the efficiency gains from re-moving the trade (and perhaps other) distortions.

3. This grouping is referred to as the Quad-plus, orQuad�, because of the inclusion of Australia and NewZealand.

4. Note that under the bilateral agreements, do-mestic distortions are not removed because they arenot specific to any individual trading partner.

5. Note that some of the high-income Asiancountries—for example, the Republic of Korea andSingapore—are excluded from the bilateral agreementsand therefore tend to lose, in part, because of tradediversion effects.

6. The excluded countries/regions are Brazil,China, India, Mexico, Russia, rest of East Asia, andrest of South Asia.


148

7. The three North-South agreements include abroad East Asian region that encompasses both thehigh-income and developing countries, the FTAA in theWestern Hemisphere, and a broad free trade area cen-tered on the EU, including the new accession countries,extending to the Middle East and North Africa and Sub-Saharan Africa. The South-South agreements include aLatin American-wide RTA (LAFTA), a developing EastAsia RTA (AFTA), a Europe and Central Asia RTAexcluding the EU-accession countries (CIS), a Sub-Saharan African RTA (SSA), the Middle East and NorthAfrica (MNA), and a South Asian RTA (SAFTA).

8. In a result not shown in these figures, South Asiais not included in any of the three broad North-Southagreements, but is simulated to create its own RTA. Inisolation, the gains from SAFTA, the South Asian re-gional trade agreement, yields a positive gain of some0.2 percent of baseline income, as shown in figure 6.3.However, SAFTA’s exclusion from the 3-block North-South agreements induces a loss of 0.3 percent of base-line income, largely due to trade diversion and a loss inthe terms of trade.

9. The simulated South/South agreements reflectfree trade among geographic neighbors where the dif-ferences in comparative advantage are slight. Reduc-tions in South/South trade barriers across more distantcountries have significantly more potential for increas-ing trade and incomes.

10. Concerns about the impact of discriminationon multilateralism in trade has a long history, e.g.,Patterson (1965). There is voluminous literature onthis issue, much of it conceptual. Interested readers arereferred to Bhagwati (1991), De Melo, Panagariya andRodrik (1993), Winters (2000, 2001), and Schiff andWinters (2003).

11. Much of the early theoretical literature that an-alyzed the incentive effects of RTAs for multilateralismassumed, for simplicity, that the latter implied freetrade—see Levy (1997), and Krishna (1998). [Winters(2000) provides a thorough review of the theory.] As aresult, the focus was on a binary choice between mul-tilateral free trade and no MFN liberalization. How-ever, in practice countries can choose to conclude amultilateral round with considerable liberalization orwith very little. Thus an empirical perspective that fo-cuses on whether RTAs change the probability of con-cluding a trade round is too narrow. Moreover, it is vir-tually impossible to test this in a systematic waybecause one never observes simultaneously a countrywith and without RTAs. Therefore an empirical assess-ment of the stumbling block hypothesis must useBhagwati’s definition and investigate if RTAs lead toless multilateral liberalization.

12. Testimony before the House Committee onWays and Means, May 8, 2001; accessed June 2003 on

http: / /waysandmeans.house.gov/legacy/trade/107cong/5-8-01/5-8chri.htm; U.S. tariff schedule 1994and 2003. See Limão (2003).

13. These are described in chapter 2.14. A practical problem faced by the drafters of

Article XXIV was that in the case of a customs union,changes in the external tariffs of member countrieswould occur as they adopt a common external tariff.The rule that applies to customs unions is that dutiesand other barriers to imports from outside the unionmay not be, on the whole, higher or more restrictivethan those preceding the establishment of the customsunion (Article XXIV:5a). The interpretation of thisphrase became a source of much disagreement amongGATT contracting parties. However, except for cus-toms unions, (in which a common tariff structurewould be adopted), the rule for RTAs was unambigu-ous. Duties applied by each individual member countrymay not be raised (Hoekman and Kostecki 2001).

15. The RTA was argued to lead to “trade diver-sion by raising barriers to third countries’ exports ofintermediate manufactured products and raw materi-als. This resulted from unnecessarily high requirementsfor value originating within the area. In certain cases ...the rules disqualify goods with value originating withinthe area as high as 96 percent. The rules of origin lim-ited non-originating components to just five percent ofthe value of a finished product of the same tariff head-ing [for] nearly one-fifth of all industrial tariff head-ings” (GATT, 1974:152-53, cited in Hoekman andKostecki 2001).

16. The WTO Agreement on Rules of Origin aimsto foster the harmonization of the (nonpreferential) rulesused by members. The objective is that common rules beapplied equally to all nonpreferential trade policy instru-ments by WTO members—tariffs, import licensing, an-tidumping, and so forth. The agreement calls for a workprogram to be undertaken by a Technical Committee, inconjunction with the WCO, to develop a classificationsystem regarding the changes in tariff subheadings(CTH), based on the Harmonized System (HS); thiswould constitute a substantial transformation. In caseswhere the HS nomenclature does not allow substantialtransformation to be determined by a CTH test, theTechnical Committee is to provide guidance regardingthe use of supplementary tests such as value-added crite-ria. Although the harmonization program was to becompleted by July 1998, deadlines have also beenmissed, in part, reflecting lack of consensus among mem-bers over the formulations to be adopted for sensitiveproducts, especially agriculture, textiles, and clothing.

17. As the implied move to full global free trade isunlikely to be politically feasible, the implicationwould be to impose a ban on most if not all RTAs—also unlikely to be acceptable.


149

18. Clearly there would be budgetary implicationsassociated with a stronger surveillance role, and aprecondition is that WTO members accept that theSecretariat be given the independence to undertake theanalysis and form an explicit judgment of the effects ofspecific agreements. However, if the required resourcesor willingness for the WTO to undertake the task can-not be found, this in itself would be a good indicationof the importance that is accorded by WTO membersto the spread of RTAs.

19. See Hoekman (1998) for a discussion.

ReferencesAghion, Philippe, Pol Antras, and Elhanan Helpman.

2004. Negotiating Free Trade. NBER WorkingPaper 10721, National Bureau of Economic Re-search, Cambridge, MA.

Baldwin, Richard. 1993. A Domino Theory of Region-alism. CEPR Discussion Paper No. 857. Centerfor Economic Policy Research, Washington, DC.

Bhagwati, Jagdish. 1991. The World Trading System atRisk. Princeton, NJ: Princeton University Press.

Bhagwati, Jagdish, and Arvind Panagariya. 1996. TheEconomics of Preferential Trading Agreements.Washington, DC: AEI Press.

Dee, P., and J. Gali. Forthcoming. The Trade and In-vestment Effects of Preferential Trading Arrange-ments. In NBER East Asian Seminar in Economics14 Proceedings. Chicago: University of ChicagoPress.

De Melo, Jaime, Arvind Panagariya, and Dani Rodrik.1993. Regional Integration: An Analytical andEmpirical Overview. In New Dimensions in Re-gional Integration, eds. J. De Melo and A. Panagariya. New York: Cambridge UniversityPress.

Finger, J. Michael. 1993. GATT’s Influence on RegionalAgreements. In New Dimensions in RegionalIntegration, eds. J. De Melo and A. Panagariya.New York: Cambridge University Press.

Harrison, Glenn, Thomas Rutherford, and David Tarr.2002. Trade Policy Options for Chile: The Im-portance of Market Access. World Bank Eco-nomic Review 16(1).

Hoekman, Bernard. 1998. Preferential Trade Agree-ments, In Brookings Trade Forum 1998, ed. RobertLawrence. Washington, DC: Brookings Institution.

Hoekman, Bernard, and Denise Konan. 2001. Deep In-tegration, Nondiscrimination and Euro-Mediterranean Free Trade. In Regionalism inEurope: Geometries and Strategies After 2000,eds. Jurgen von Hagen and Mika Widgren.Kluwer Academic Publishers.

Hoekman, Bernard, and Michel Kostecki. 2001. ThePolitical Economy of the World Trading System,2d ed. New York: Oxford University Press.

Hoekman, Bernard, Constantine Michalopoulos, andL. Alan Winters. 2004. Special and DifferentialTreatment of Developing Countries: MovingForward After Cancún. The World Economy 27:481–506.

Konan, Denise, and Keith E. Maskus. 2003. Quantify-ing the Impact of Services Liberalization in a De-veloping Country. Policy Research Working Paper3193, World Bank, Washington, DC.

Krishna, Pravin. 1998. Regionalism and Multilateral-ism: A Political Economy Approach. QuarterlyJournal of Economics 113: 227–51.

Lawrence, Robert Z. 1991. Emerging RegionalArrangements: Building Blocks or StumblingBlocks? In Finance and the International Econ-omy 5: The AMEX Bank Review Prize Essays,23-25, ed. Richard O’Brien. New York: OxfordUniversity Press.

Levy, Philip. 1997. A Political-Economic Analysis ofFree-Trade Agreements. American Economic Re-view 87(4): 506–19.

Limão, Nuno. 2003. Preferential Trade Agreementsas Stumbling Blocks for Multilateral TradeLiberalization: Evidence for the U.S. Paper. Uni-versity of Maryland, College Park.

Mansfield, and Reinhardt. 2003. Multilateral Determi-nants of Regionalism: The Effects of GATT/WTOon the Formation of Preferential Trading Arrange-ments. International Organization 57(4): 829–862.

Milward, Alan. 1992. The European Rescue of theNation State. Berkeley: University of CaliforniaPress.

Patterson, Gardner. 1966. Discrimination in Interna-tional Trade: The Policy Issues, 1945–1965.Princeton, NJ: Princeton University Press.

Saggi, Kamal. 2004. Preferential Trading Arrange-ments and Multilateral Tariff Cooperation.Southern Methodist University Working Paper.Dallas.

Saggi, Kamal, and Halis Murat Yildiz. 2004. BilateralTrade Agreements and the Feasibility of Multilat-eral Free Trade. Southern Methodist Universityand Ryerson University Working Paper. Dallas.

Schiff, Maurice. 1997. Small is Beautiful: PreferentialTrade Agreements and the Impact of Country Sizeand Market Share. Journal of Economic Integra-tion 12: 359–87.

Schiff, Maurice, and L. Alan Winters. 2003. Regional-ism and Development. New York: OxfordUniversity Press.

Schott, Jeffrey. 2004. Assessing US FTA Policy. In FreeTrade Agreements: US Strategies and Priorities,


150

ed. Jeffrey Schott.Washington, DC: IIE (Institutefor International Economics).

Snape, Richard. 1993. History and Economics ofGATT’s Article XXIV, In Regional Integration andthe Global Trading System, eds. K. Anderson andR. Blackhurst. London: Harvester-Wheatsheaf.

Srinivasan, T. N. 1998. Developing Countries and theMultilateral Trading System: From GATT to theUruguay Round and Beyond. New York: HarperCollins.

Stoeckel, Andrew, and B. Borrell 2001. PreferentialTrade and Developing Countries: Bad Aid, BadTrade Canberra: RIRDC (Rural IndustriesResearch & Development Corporation).

Tojo, Yoshiaki. 2004. The WTO and FTAs. Presenta-tion at the JEF-IIE Conference, 16 March.

U.S. GAO (United States General Accounting Office).2004. International Trade: Intensifying FreeTrade Negotiating Agenda Calls for Better Allo-cation of Staff and Resources. Report to Congres-sional Requesters, Washington, DC.

Viner, Jacob. 1950. The Customs Union Issue. NewYork: Carnegie Endowment for InternationalPeace.

Winham, Gilbert. 1986. International Trade andthe Tokyo Round Negotiation. Princeton, NJ:Princeton University Press.

Winters, L. Alan. 2000. Regionalism versus Multilater-alism. In Market Integration, Regionalism andthe Global Economy, eds. R. Baldwin,Daniel Cohen, André Sapir and Anthony J. Venables. London: Centre for Economic PolicyResearch.

. 2001. Post-Lomé Trading Arrangements:The Multilateral Option. In Regionalism inEurope: Geometries and Strategies After 2000,eds. Jurgen von Hagen and Mika Widgren.Kluwer Academic Publishers.

World Bank. 2000. Trade Blocs. Washington, DC:World Bank.


151

making regionalism complementary to multilateralism

Documents