e. The multilateral trading system and pTas · PDF fileA perennial policy question is how the multilateral trading system is affected by the ... system. This system puts Home in a

WOrld Trade repOrT 2011

164

A perennial policy question is how the multilateral trading system is affected by the rise of preferential trading agreements (PTAs). Is multilateral trade cooperation compromised by burgeoning regionalism? Should we see these different approaches as complementing or competing with each other? Are there synergies, or inevitable conflicts? Building on the analysis of the report so far, this final section examines these questions.

e. The multilateral trading system and pTas

II – tHe Wto AnD PReFeRentIAL tRADe AGReements

165

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

Contents 1. Systemiceffectsofpreferentialtariffliberalization 166

2. DeepPTAprovisionsandthemultilateraltradingsystem 168

3. RegionalismandtheWTO:historicalperspective 182

4. TherelationshipbetweenPTAsandtheWTO 187

Some key facts and findings

• Deep integration is often non-discriminatory in nature.

• Global production networks can result in PTAs with tariff and non-

tariff measures that are more consistent with the principles of the

multilateral trading system.

• A large number of disputes between PTA members are brought to

the WTO dispute settlement system. On average, about 30 per cent

of WTO disputes are between members who are parties to the same

PTA.

• A critical-mass approach to decision-making in the WTO may be

required, at least in the short term, to move forward on an agenda

that creates greater coherence between PTAs and the multilateral

trading system.


166

1. Systemiceffectsofpreferentialtariffliberalization

Inthe late1980sandearly1990s,aseriesofeventsled analysts to focus on the systemic effects ofregional integration (Baldwin, 2009).1 Regionalismrose in North America, where the Canada-UnitedStatesPTAwasfollowedbytheNorthAmericanFreeTrade Agreement (NAFTA) negotiations. It alsoreignited in Europe with the Single European MarketinitiativeandthedisintegrationoftheSovietUnion.Atthe same time, the prospects for a prompt andcomprehensivecompletionoftheUruguayRoundwereshroudedinuncertainty.

Thepossibilityofacausallinkbetweentheexpansionof regionalismanddifficulties incoming toclosure inmultilateral negotiations could not be ignored. Thisturned the regionalism debate into a systemicdiscussion. This section provides a short overview ofthe literature in this area, drawingon several surveysthat have been published recently: Baldwin (2009),FreundandOrnelas(2010)andWinters(2011).

The broad concern of this literature is the relationbetween discriminatory and non-discriminatory tariffliberalization.Thestandardapproachistostudywhetherpreferential tariff cuts lead to a reduction or to anincreaseinthemost-favourednation(MFN)tariff,whichis applied by WTO members on a non-discriminatorybasis.AsdiscussedinSectionC,theevidencesofarisnot conclusive. However, there are some studies thatfocusontheeffectofpreferentialtariffliberalizationonnon-discriminatory tariff liberalization.2 Due to thepaucity of adequate data, opportunities for convincingempirical work are limited. The literature is thereforemostly theoretical, and its predictions are oftensupportedonlybyanecdotalevidence.

(a) DoPTAsfosterorhindermultilateraltariffreductions?

A number of different mechanisms have beenidentified through which PTAs could foster or hindermultilateraltradeopening.

AsdiscussedinSectionC,theKemp-Wantheoremisa theoreticalbenchmarkshowing thatPTAsneednothave adverse effects on multilateral tariff reductions.StartingfromasituationwhereallcountrieshaveMFNtariffs, groups of nations can always raise theircollectivewelfarebyformingatradebloc.Apiecemealenlargement of the bloc will raise bloc members'welfare,andthehighestwelfarewillbereachedwhenallnationsarepartofthebloc(KempandWan,1976).Thistheoreticalresultrestsontwostrongassumptions.First,PTAmembersmustsetexternal tariffsat levelsthatfreezetheirtradeflowswiththerestoftheworld.Secondly, lump-sum transfers between membersensurethattheyallgainfromthePTA.3

Thefearofpreferenceerosionisanimportantaspectof the relationship between preferential andmultilateral tariff opening.4 In a world where moreopen tradewouldbe in the interestof all nationsbutwhere individual nations fearing erosion of theirpreferences would veto it, regionalism can helpachieve global trade opening. Baldwin (2009)illustratestheargumentwithanexamplewhereHomecountrysignsseparatePTAswithPartner1andwithPartner2,therebyformingaso-calledhubandspokesystem. This system puts Home in a favourableposition as it combines opening trade on the importsidewithpreferentialtariffsontheexportside.Home,the hub, is likely to oppose WTO talks aimed atachieving more open trade for fear of preferenceerosion.Despitethis,Homeanditstwopartnerscouldreach global trade opening, not through multilateralnegotiations, but rather through a PTA between thetwospokes.AsBaldwinshows,thetwopartnerswouldalways prefer global trade opening to the hub-and-spokesituation.

Thefearofpreferenceerosioncan,however,constitutea potent force of resistance to multilateral tariffreductions. The economic literature has shown thattwoormorenationscan formaPTAwhich increasestheirjointwelfareattheexpenseofthirdnations.Sucha PTA will hinder multilateral trade opening becauseits removal will be resisted by member countriespreciselytoavoidpreferenceerosion.Thiscanbetruenotonly ifPTAmembers increasetheirexternalMFNtariffs, but also when external tariffs are frozen.Baldwin (2009) provides an example in which at asufficiently lowinitialtariff, thegainsofmaintainingaPTAthatreducesthird-countrywelfareareworthmorethanthestandardgainsofglobaltradeopening.5

Developingcountriesthatweregrantednon-reciprocalpreferential access to developed countries’ marketsare particularly concerned by preference erosion,particularly where reduced advantages frompreferentialtariffsarenotoffsetbythegainsinmarketaccessduetotariffcutsongoodsthatdonotreceivepreferences.6

Politicaleconomy factors canalsoaffect thepaceatwhich preferential tariffs are extended to non-membersonaMFNbasis. IfPTAsaretrade-creating,theywillincreasethesizeofexportsectorsandreducethesizeofimport-competingsectors.Ifpoliticalpoweris proportional to the size of the sector, the PTA willincrease support for trade opening.7 In particular, itcanmake itpoliticallyoptimal forgovernments tocutMFNtariffstolevelsthatwouldhavebeenundesirablewithoutthePTA.8

Along the same lines, if workers have imperfectinformationonhowtheywillbeaffectedbymoreopentrade, they may initially oppose global trade openingbut accept a PTA, which is an intermediate form oftrade barrier reduction (Frankel et al., 1995). A PTA


167

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

may inform workers on how they will be affected byglobal trade opening and make an MFN approachpoliticallyfeasible.

Thepoliticaleconomymodelsdiscussed inSectionC(Grossman and Helpman, 1995; Krishna, 1998),however, offer some insights as to why PTAs mightinhibit multilateral tariff reductions. In such models,interest groups might seek primarily trade-divertingPTAs, i.e. agreements that provide enhancedprotection.9 In Krishna's model the extent of tradediversiondeterminesthedegreeofpoliticaloppositiontoamultilateralagreement thatwouldfindsupport intheabsenceof thePTA. Intuitively, if there is littleorno trade diversion, firms from each member countryobtainhighermarketshares (andprofits) in theothermember’s market but lose domestic profits, with anoverallsmalleffectonnetprofits.However,ifthePTAallows bloc firms to displace those from excludedcountries in each other’s markets, it surely enhancesprofitsforallfirms,attheexpenseofoutsiders(FreundandOrnelas,2010).10

The result that specific interestgroupsmight opposemultilateral tradeopening thatwouldbesupported inthe absence of a PTA is also obtained in a median-votersettingbyLevy(1997).HeshowsthatabilateralPTAmightofferdisproportionately largegains tokeyagentsinacountry,makingthemunwillingtosupporta multilateral agreement, which would therefore beblocked. This might be the case, for instance, if thetwocountrieshavesimilarfactorendowments,sothata lot of trade within the PTA is intra-industry trade,with limited redistributive effects. A move towardsmultilateralopeningwouldalterdomesticfactorprices,creating winners and losers and adding only modestgainsfromincreasedvarietyorspecializationbasedoncomparativeadvantage. Inthiscase,themedianvoterwould oppose such a move, and the PTA acts as anobstacletomultilateraltradeopening.

SomePTAsmaybeconcludedpartlyinpursuitofnon-economic objectives, such as understanding andreconciliation between former enemies (e.g. FranceandGermany),orbetweennationswithformercoloniallinks (Schiff and Winters, 1998). As discussed inSectionC,someauthorshavearguedthatthesenon-economic objectives might lead member countries tooppose further multilateral trade opening. In a modelby Limão (2007), PTAs allow partner countries toextract mutual cooperation on the non-trade issue,using preferential tariffs as bargaining chips. Theprospect of dissipating this possibility via multilateraltrade opening might make countries less likely tofavouraglobalapproach.11

PTAs may also increase the adjustment costsassociatedwithmultilateral tradeopeningwhenfirmshave to make sunk, sector-specific investments toproduce. As shown by McLaren (2002), in such asituationtheex postgainsfrommultilateralreductions

can be reduced relative to those from preferentialtrade opening, and the latter emerges in equilibrium.Thereasonisthefollowing:iffirmsexpectglobaltradeopening to arise, they will invest in sectors ofcomparativeadvantage,soeverycountrywillbecomehighlyspecialized.Inthissituation,theex antegainsofmultilateral trade opening materialize, and suchopening is likely to occur. If, however, firms expect aPTA to be signed, they will invest in goods in whichexcluded countries have a comparative advantage,because external tariffs will render these goodsexpensive. For similar reasons, firms from excludedcountries will invest in goods where PTA membershave a comparative advantage. As PTA countriesbecome specialized relative to each other, and lessspecializedrelativetooutsiders,thegainsfromglobaltrade opening will be reduced. As McLaren (2002)explains, the resulting regionalism is “insidious”because it is an inferior outcome for all participants,and it emerges only because it prompts sunkinvestmentsthatreducethevalueofmultilateraltradeopening.

Finally,oppositiontofurthermultilateraltariffopeningbyPTAmembersmightcomefromexcludedcountries.The logic is as follows: if PTA members reduce theirexternal tariffs for political economy reasons aftersigning an agreement, this might result in pure tradecreation.AsarguedbyOrnelas(2005b),non-membersbenefitfromsuchPTAsbyobtainingincreasedmarketaccesstomembercountrieswithouthavingtoreducetheir own tariffs, as would be required under amultilateral agreement. Therefore, non-members mayturnagainstmultilateraltradeopeningthattheywouldsupportintheabsenceofthePTA.12

The overview of the literature thus suggests that theeffect of regionalism on the prospects of multilateraltradeopeningwilldependonanumberoffactors.Theresults depend on how much members and non-members stand to gain from a PTA, and how muchtheywouldlosefrommultilateraltradeopening,ontheimportance of political economy considerations inpolicyformation,andontheextentoflock-ineffectsofpreferential trade opening. Moreover, results dependon whether regionalism is open or not (Yi, 1996); onthepresenceofdissimilaritiesinendowmentsorcosts(SaggiandYildiz,2009);ontherulesofthemultilateraltrade system (Bagwell and Staiger, 1999; Saggi andYildiz, 2009); as well as on the formal enforcementconstraints(BagwellandStaiger,1997a:1997b).

(b) Evidenceonthesystemiceffectsofregionalism

Whenthetheoryisinconclusive,themostnaturalthingtodoistoturntoempiricalevidence.Afirststrandofliterature tests whether MFN and preferential tariffsare complements or substitutes.13 As discussed inSectionC,differentresultsemergefordevelopinganddeveloped countries. While in the former group of


168

countries preferential trade agreements appear toreduceexternaltariffs,inthelattergroupofcountriestheyseemtoincreasethem.MostofthecontributionsdonotdistinguishbetweenMFNtariffsthathavebeennegotiatedatthemultilaterallevelandunilateraltariffreductions.14ThenotableexceptionsareLimão(2006)and Karacaovali and Limão (2008), who explicitlyconsider the effect of preferential trade opening onmultilateraltradeopeningattheUruguayRoundintheUnitedStatesandtheEuropeanUnion,respectively.15

A second strand of literature investigates thecorrelationbetweenPTAformationandmultilateralism.One often-used example of regionalism promotingmultilateral trade opening is when the United States,which for many years had been advocatingmultilateralism,convertedtoregionalisminthe1990sand thereby revived the Uruguay Round negotiations(Bergsten and Schott, 1997).16 Mansfield andReinhardt(2003)observethatmorePTAsareformedduring multilateral negotiations than at other times.They interpret this result as evidence consistent withmultilateralism promoting PTAs as devices to obtainbargaining leverage within the multilateral regime(pressuring outsiders to open their markets orescapingfromfree-riders).

Ageneralproblemwith theapproachof linkingPTAswithmultilateraltraderoundsisthatthelatterarerareevents. Moreover, the practice of multilateral traderounds is to negotiate multilateral opening with moreor less ambitious scenarios of trade opening, ratherthan opting for full or no multilateral opening.Therefore,adirecttestofwhetherPTAsdecreasethelikelihood of signing multilateral trade openingagreements is impossible (World Trade Organization(WTO),2007).

Anecdotal evidence can be found in support both oftheview thatPTAs facilitate furthermultilateral tradeopening and of the view that they hinder it.17 On theone hand, there is anecdotal evidence that PTAsincreaseexcludedcountries’ incentivetomoveonthemultilateral front to avoid trade diversion. A relatedargument is that the last three rounds of multilateraltradenegotiationshave started in tandemwithmajormoves towards regional integration, which issometimes taken as evidence of the building blockrelationshipbetweenthetwoprocesses.Furthermore,the cost from overlapping PTAs can trigger arationalization of the system – as in the case of thePan-EuropeanCumulationSystem–or a recourse tothe multilateral system – as in the case of the WTOInformationTechnologyAgreement.18

Ontheotherhand,ithasbeenarguedthattheconcernforpreferenceerosionhascontributed to the stallingof multilateral negotiations and has actually beenreflected in less multilateral trade opening, see forinstanceCurtisandVastine(1971).Furthermore,thereis also evidence that the engagement in regional

negotiationsmaystalltheprocessofmultilateraltradeopening by absorbing resources away from themultilateral negotiations (World Trade Organization(WTO),2007).

2. DeepPTAprovisionsandthemultilateraltradingsystem

While the literature on the systemic effects ofpreferential tariffs isrichandveryactive,sofar therehasnot beenmuch researchon the systemiceffectsof other, “deep” integration, provisions. Availableresults suggest that in some deep integration areas,such as technical barriers to trade (TBT), multilateralregulation may not be economically optimal orpolitically feasible. Because deep integration is oftenMFN innature, however, such regulationmayalsobeless necessary. Indeed, the literature has identified anumberofmechanismsthroughwhichdeepintegration“automatically” supports further opening, or at leastdoes not entail negative static effects on themultilateraltradingsystem.

(a) Deepintegrationisoftennon-discriminatoryinnature

Bytheirverynature,somedeepintegrationprovisionsarede facto extended tonon-membersbecause theyare embedded in broader regulatory frameworks thatapply toall.Anexample isprovidedbyservices tradeopening. Barriers to trade in services are generallybehind-the-border, regulatory measures. Even thoughsomeservicesbarrierscould inpracticebeapplied ina differentiated manner depending on the suppliers'countryoforigin(e.g.restrictionsonthemovementofpersons, foreign equity restrictions, or foreign directinvestment screening), one expects that barriersremovedor relaxedasa resultofaPTAbeextendedde facto to non-parties. This also makes mosteconomicsense,andmaylimitanyeconomicdistortionresultingfromservicesPTAs.19

Evidence suggests that in certain cases, preferentialtreatment was granted to PTA parties, but properanalysis of this is made difficult by the absence ofcomprehensive information on the treatment appliedbycountries toservicesandsuppliersof their tradingpartners.Thisiscompoundedbythefactthatanalysisof non-discriminatory treatment in services wouldneed toconsidernotonly treatmentspecified in lawsand regulations, but also de facto treatment − forexample, which suppliers receive operating licences,which are sometimes limited in number. Furthermore,given the importance of first-mover advantage forsuppliers in a number of services sectors,20 whatmatters is whether non-preferential treatment isavailable forall suppliersofdifferentorigins from themomenttradeopeningtakesplace.Whilethismaywellbethesituationmostofthetime,informationislacking.


169

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

The fact that services commitments in PTAs can benon-discriminatoryalsosuggeststhatanytechnicaloreconomic obstacle to the multilateral extension ofsuch PTA commitments as part of the Doha Roundwould be limited. It can be hoped that preferentialcommitments made by several WTO members maketheir way into these members' conditional offers andinject momentum in the Doha services negotiations.Thishasnothappenedinofferscurrentlyonthetable– which for the most part were submitted in 2005 –therefore suggesting that other factors are at play,either within the Doha negotiations or domestically.One such factor may be that, in the context of thegrowing number of preferential trade agreements inrecentyears,anumberofcountriesmaywishtokeepleverage for their PTA negotiations, wherecommitments thatgobeyond theGeneralAgreementon Trade in Services (GATS+ commitments) areexchanged as part of the overall trade-off betweenparties (e.g. against preferential goods access), eventhough the resulting overall outcome is lesseconomically significant than what the Doha Roundcanproduce,includingforthesePTAparties.

Anotherfactortoconsideristhatrulesoforigin(RoOs)for services do not carry the same potential fordistortionastheydoforgoodstrade.RoOsinservicesPTAs are usually liberal, along the lines of GATSArticleV(6),21althoughtherearecertainexceptions.22This reduces the extent of the spaghetti bowl effect(seeSectionC).

For mode 1 (cross-border supply), PTAs generallyfocusontheterritorialpresenceoftheproviderratherthan on its nationality or the origin of the service,according origin status to the services provided byentities located in a PTA partner nation. For mode 2(consumption abroad), the supplier's nationality isunimportant as well; the focus is on the territory inwhich the service is supplied and consumed. Formode3(commercialpresence),RoOstypicallyaccordorigin status to firms with “substantive businessoperations” within the PTA region, irrespective of thenationalityofbusinessowners.Inotherwords,theonlyrequirement is to establish a legal presence and acertain levelofcommercialactivity inoneof thePTAmembers.23

Inotherareas,suchasmutualrecognitionagreements(MRAs)ontesting,RoOsareabsent.Iftwonations(forexample, the United States and Singapore) sign anagreement whereby the United States acceptsproducts tested in Singapore laboratories,independentlyoftheirorigin,Singaporecanbecomearegional hub for testing and conformity assessment.Neighbouring countries can ship their products thereto be certified before being exported to the UnitedStates.ThelackofRoOsautomaticallymultilateralizesthebilateraltestingMRA,reducingthespaghettibowleffect(Baldwinetal.,2009).

Competitionpolicyprovisions inPTAsarealsomostlycharacterizedbynon-discrimination(Teh,2009;Dawarand Holmes, 2010). Competition disciplines usuallyoperatethroughtheuseofdomesticregulations.WhileitisnotimpossiblefortheseregulationstobetailoredtofavourenterprisesoriginatingfromPTApartners, itmay be costly to do so and becomes even moredifficultasthenumberofPTAstowhichacountryisasignatoryincreases.Transparencyandinparticulartheobligation to publish laws promoting competition willprovide information that becomes (simultaneously)availabletoPTAandnon-PTAmembersalike.

The substantive obligations in the competition policychapters of PTAs generally involve applyingcompetition laworsettingupacompetitionauthority.Totheextentthatenforcementofcompetitionlawinacountry reduces the market power of domesticincumbents, the prospects of foreign enterprises,whether they are from a PTA member or not, areimproved. Carrying out the competition obligationsalso opens up opportunities for new foreign entrants(either from PTA or non-PTA members) to challengedomesticincumbents.

Moreover, therearepositiveeffects fromcompetitionprovisions,particularlyiftheyarecontainedinregionalagreements(DawarandHolmes,2010).Therecanbeeconomies of scale from the creation of a regionalcompetitionauthority.Even ifnocentralizedauthorityis established, benefits can come from information-sharing and cooperation among enforcementauthorities. There could be demonstration effects tootherjurisdictionswhenacompetitionauthorityinonePTA member takes action against anti-competitivebehaviour. Eventually, more common competitionnorms and practices within the PTA will preventregulatory arbitrage, where enterprises locatethemselves in a jurisdiction in the PTA with relativelylaxcompetitionpolicy.

Finally,PTAsmaydirectly refer toWTO rules. Lesser(2007)arguesthatthemajorityoftechnicalbarrierstotrade (TBT) provisions in PTAs signed after 1995reaffirm the parties' rights and obligations under theWTO TBT Agreement and make reference to itsobjectives.

Furthermore, most transparency commitmentsincluded in PTAs are similar in nature to the onesincluded in the WTO TBT Agreement. Finally,provisionsthatrequirepartiestoprovideanexplanationin case of non-recognition of standard-relatedmeasures and mechanisms supporting furthercooperationamongparties (e.g. technical assistance,jointstandardization)caninfactsupportandenhancethe implementation of the WTO TBT Agreement,supportingthemultilateraltradingsystem.


170

BoxE.1:Investment provisions in international agreements: is there a potential for third-party discrimination?

Theprocessofgradualopeningofforeigndirectinvestment(FDI)hasbeentheoutcomeofamulti-layeredprocess combining autonomous MFN investment opening, commitments made in the context of bilateralinvestmenttreaties(morethan2,700todate),24andonlymorerecentlycommitmentsmadeinPTAs.Despitethe progress in investment provisions in PTAs, investment remains overwhelmingly regulated by bilateralinvestmenttreaties(BITs).

InvestmentprovisionsaretypicallyincludedinPTAstofosterinvestmentflowsbetweenmembercountries.Some provisions are clearly aimed at protecting investors, without increasing barriers to investment fromthirdcountries(BacciniandDür,2010).TheinvestmentchaptersofPTAsnormallyincludeabsolutestandardsoftreatmentprovidingaminimumlevelofprotectionforinvestors.Inmanycases,theyreflecttheactualstateofdomesticlegislationconcerningFDIandthelevelofcommitmentachievedinearlierBITs.Theprovisionsregarding investment protection are either directly included in the text of the agreement, such as in theagreements signed by the United States, or they are indirectly referred to in agreements providing thatinvestorsshouldbetreatedinaccordancewithcustomaryinternationallaw(Kotschwar,2009).

Ithasbeennoted,however,thatthecreationofaPTAmaybeasourceofinvestmentdiscrimination,wherebypotential investors from excluded countries are put at a disadvantage vis-à-vis investors from membercountries. This can occur through two channels: one direct and the other one indirect (Baccini and Dür,2010).First,investmentdiscriminationcanresultdirectlyfromtheinclusionofprovisionsthatopenupcertainsectorsforinvestmentonlyonapreferentialbasis.AllPTAsincluderelativestandardsoftreatment,namelyMFNandnationaltreatment(NT).25MostrecentPTAs,includingtheonessignedbytheUnitedStatesandtheonesamongAsiancountries,tendtoprovidebothMFNandNTduringallphasesoftheinvestment(pre-andpost-establishment).26Relativestandardsoftreatmentcanprovideacompetitiveadvantagetoinvestorsfrommembercountriesvis-à-visinvestorsfromnon-membercountries,especiallyintheservicessector.Forinstance, the PTA between Australia and the United States relaxes the requirements for governmentscreeningofFDIforUScompaniesinvestinginAustralia(BacciniandDür,2010).

Secondly,investmentdiscriminationcanresultindirectlyfromdiscriminatorytariffreductions.AssumefirmsfromcountriesAandBareengaged inmarket-seekingFDI incountryC.Theysource inputsdomestically,andimportthemintoCattheMFNtariffτC.APTAbetweenAandC,thateliminatestariffsonintermediaryinputs from A, creates investment discrimination by putting investors from country B at a competitivedisadvantage.However,thereisverylittleempiricalevidenceontheactualincidenceofsuchdiscrimination.

The extent of potential investment discrimination also depends on the RoOs included in the PTA. LiberalRoOs in the services sector, for instance, reduce the discriminatory aspects of investment provisions forservicesproviders.Thereis,however,considerablevariationinthestrictnessofrulesoforiginforinvestmentacrossPTAs(BacciniandDür,2010).Moreover,oneshouldconsidertherelationbetweentheprovisionsofPTAsandtheonescontainedinBITs.

BITs are traditionally about the protection of investment that is already established in the host countries(DiMascio andPauwelyn, 2008), guaranteeing compensation in casesof expropriationand repatriationofprofits. In the early BITs, what mattered for host country governments was the flexibility to differentiatebetween national and foreign governments, not so much among foreign investors. Nonetheless, a hostcountrycouldwishtoexerciseselectivescreeningovertheadmissionofforeigninvestorsandthetermsoftheir admissionaspartof itspolicies topromotenational investments.Forexample, it couldwish toofferinvestmentincentivesonlytocertainforeigninvestorsonadiscriminatorybasis.DespiteanimprovementinabsolutestandardsoftreatmentinrecentBITs,mostofthemstilldonotcoverpre-establishmentorentryofinvestments,accordingNTand/orMFNonlyonceinvestmentsareinthecountry.Forthisreason,andalsobecause theydonot cover tariff reductions,Baccini andDür (2010)argue thatBITsarenot very likely tolowerPTAs’potentialforinvestmentdiscrimination.

It should be noted that investment discrimination need not imply a reduction in FDI flows from excludedcountriesintomembercountries.Tariffdiscriminationmayleadtotariff-jumpingFDI(i.e.theestablishmentofaproductionfacility inamembercountry, throughFDI, inordertoavoidatariff).StudiesfindingthatPTAsattractFDIfromthirdcountries,suchasteVeldeandBezemer(2006),donot, therefore,provideevidenceagainstPTA-driveninvestmentdiscrimination.


171

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

(b) SeveralmechanismssupportingfurtherliberalizationarefoundinPTAs

First, PTAs may include “non-party” MFN clauses.Theseclausesstipulate theextension tocurrentPTApartners of preferences or concessions that membercountriesmayhavegrantedinthepastormaygrantinthe future to third nations.27 In the case of servicesand government procurement for instance, suchprovisionsensure that futureandmoreadvantageouscommitmentswithothernon-memberpartnersshouldbe granted to PTA partners as well (Fink andMolinuevo, 2008). Many PTA procurement provisionsrequire third-partyMFNguaranteessoas to limit theextent to which preferential procurement isundermined by subsequent PTAs (Baldwin et al.,2009).28

Secondly, there is a tendency to replicate trade-opening rules in PTAs because template approachesare often used for PTAs. The spread of the NAFTA-style telecommunication competition provision is anexample. Baldwin et al. (2009) argue that the largenumber of countries that have included this provisionin PTAs suggests that it is progressively becoming anorm.Theyfurtherarguethatharmonizationtoasingleregulatory regime, including a common set of rulesthat governments apply to private firms in manynations, tends to foster competition and trade and itcannotbeconsideredpreferential.

Another example is provided by NAFTA's investmentprovisions, in particular performance requirements.These provisions have spread in Latin America andbeyond.Fifteencountrieshaveagreednever toapplyperformance requirements against foreign investorsfrom any jurisdiction. Another 36 countries havecommitted to forgo the application of suchrequirements,howeveronlyagainstCanadianandUSinvestors(Baldwinetal.,2009).

Alongsimilarlines,asarguedbyAndersonetal.(2010),“thegovernmentprocurementprovisionsofRTAshavemade feasible a significant further expansion of themembership of the Government ProcurementAgreement (GPA), in the event that parties decide totakethisstep.”

Thirdly,dominoeffects (Baldwin,1993)pointing in thedirectionofprogressiveextensionofpreferentialmarketaccess might be at play also for deep integrationprovisions.Considertheexampleof theGPA.WiththeEUenlargementfrom15to25members,non-EUGPAmembers started facing more competition ingovernmentprocurementbothinthe15EUincumbents(fromthetennewcomers)andinthetenEUnewcomers(fromthe15incumbents).Asareactiontothisformoftrade diversion, the non-EU GPA members startedpressuring the new EU members to join the GPA.29SimilardominoeffectscanbediscernedinallcasesinwhichcountriesexcludedfromaPTAfindthemselvesin

a position to adopt similar provisions to the onesadoptedbymembercountries toavoid tradediversion.The implementationbycountries intheEuropeanFreeTrade Association (EFTA) of competition policy normsthatmimictheonesofEUcountriescanbeinterpretedasawayofensuringthatfirmsinEFTAcountriesdonotfind themselves at competitive disadvantage vis-à-visfirmsintheEuropeanUnion(Baldwinetal.,2009).

(c) Theeffectsofglobalproductionsharing

The presence of international fragmentation ofproductioncanalterpolitical-economyforcesinfavourof theadoptionof tariff andnon-tariffmeasures thatare less discriminatory, and more consistent with theprinciples of the multilateral trading system. TheunderlyinglogiccanbeexplainedwiththeexampleofthePan-EuropeanCumulationSystem(PECS)ofrulesoforigin(Baldwinetal.,2009).

Firms from EU countries started to relocate labour-intensive stages of production in low-wageneighbouring nations from the 1990s. At the sametime, the European Union engaged in bilateralagreements with a number of countries both fromCentral and Eastern Europe and from the SouthernMediterranean. These agreements contained non-harmonized rules of origin, giving rise to a spaghettibowl effect that restricted firms’ ability to sourceintermediate goods from the cheapest source(Gasioreketal.,2009).

Moreover, the downsizing of production in theEuropean Union, also due to competition fromemergingAsiancountriessuchasChina,reducedthenumberandpoliticalinfluenceofEU-basedproducersof intermediate inputs which benefited from theprotectionist effects of the spaghetti bowl. Thepolitical economy forces thus turned in favour ofharmonizing rules of origin across PTAs, to avoid thecost of different administrative requirements, andpermitting diagonal cumulation (i.e. allowing EU finalgood producers to source inputs from a wider set ofcountrieswithoutfearoflosingoriginstatus).ThiswasaccomplishedwiththesigningofthePECSin1997.30

International fragmentationofproductionmayalsobeadriver of deep integration, and of the multilateralextensionofdeepprovisions.Examplescanbefoundinthefieldoftechnicalbarrierstotrade(TBTs),theopeningof markets for trade in services and the presence ofcontingency measures within trade commitments(Baldwinetal.,2009).InTBTs,unbundlingofproductionmayhelpexplaintheadoptionofinternationalstandards,at least in parts and components, in industriescharacterized by global sourcing (e.g. electronics).Concerningtheopeningofmarketsfortradeinservices,offshoring is likely to create an incentive for nations toapply international standards to improve thecompetitivenessoftheirownexportersandtomaketheirownservicesmarketsmoreattractivetoforeigninvestors.


172

Finally, unbundling of production may create greatersupport for new multilateral rules on contingencymeasures, such as safeguards, anti-dumping andcountervailingmeasures,intradecommitments.Whenfirms engage in outsourcing, they prefer measuresdiscouraging the impositionof contingencymeasuresin as many bilateral trading relationships as possible,ratherthaninanyonebilateraltraderelationship.Thisunderlies the producer support for the spread of acommon or similar set of rules on the application ofcontingencymeasures(Baldwinetal.,2009).

(d) RelationshipbetweentheWTOandPTAdisputesettlementsystems

As noted in Section D, the vast majority of PTAsestablishsomekindofdisputesettlementmechanism.Porges(2010)presentsasurveyofdisputesettlementmechanisms in PTAs. She describes thesemechanisms as generally falling into the followingthree types: (i) diplomatic or political mechanisms(such as the Latin American Integration Association,ALADI); (ii) standing tribunals (such as the EuropeanUnionandtheAndeanCommunity);and(iii)referraltoadhocpanels(suchasNAFTAandotherUSFTAs,EUFTAswithChile,theRepublicofKoreaandMexico,theAssociation of Southeast Asian Nations Enhanced

Dispute Settlement Mechanism, and the SouthernCommonMarket−MERCOSUR).ThesurveyindicatesthatreferraltoadhocpanelsisthedominantmodelforPTA dispute settlement mechanisms. A slightlydifferent classification is used in Ramirez Robles(2006), which classifies the mechanisms as:(i) diplomatic; (ii) quasi-adjudicative; and (iii) “hybrid”,(i.e.mechanismsthathavefeaturesofbothmodels).

The relationship between the WTO and PTA disputesettlement mechanisms has received considerableattention in the trade literature and somecommentators have cautioned about potential risksfrom the coexistence of dispute settlementmechanisms at different levels (multilateral, regionaland bilateral) that may have overlapping jurisdictions.In this subsection, we first describe how thejurisdictions of the WTO and PTA dispute settlementsystems may overlap. We then discuss the concernsthat have been raised and the recommendations thathavebeenmadetoreducetherisksofconflict.ThisisfollowedbyareviewofthehandfulofWTOdisputesinwhich therelationshipof theWTOdisputesettlementsystemandaPTAdisputesettlementmechanismhasbeen raised as an issue. Finally, we present data onthe use of the WTO dispute settlement system bymemberswhoarepartnersinaPTA.

BoxE.2:making rules of origin more compatible with the multilateral trading system

It has been argued in this report that rules of origin (RoOs) are likely to strengthen the “spaghetti bowl”effect of PTAs. In view of this adverse effect, various commentators have argued in favour of reformingRoOs,makingthemmoretransparentandcompatiblewith theprinciplesof themultilateral tradingsystem(see for instance Cadot and de Melo, 2007).31 This box discusses the system of “cap and convergence”proposedbyEstevadeordaletal.(2009a)andsupportedbyBaldwinandThornton(2008),basedonthetwoconceptsof“multilateralization”and“convergence”.

"Multilateralization”ofRoOsreferstotheestablishmentofmultilateralrulesthatlimittherestrictivenessandcomplexityofRoOsinPTAs(Estevadeordaletal.,2009a).Accordingtotheauthors,suchruleswouldensurethat “at least the qualifying production methods in a given sector remain relatively similar across exportmarkets”. They claim that multilateralization should ideally be coupled with “convergence”, which is the“unificationofmultipleoverlappingexistingRTAsintoasinglecumulationzonewithanew,singlelistofrulesoforigin”,likeintheEuropeanPECS.

Theproposedsystemof“capandconvergence”wouldincreasetransparency(oneofthekeyprinciplesofthemultilateral trading system).Moreover, it could be subject toWTOdiscipline.Estevadeordal et al. (2009a)suggestthatthenon-preferentialRoOscurrentlynegotiatedattheWTOcouldserveastheglobalbenchmarkwithwhichtocomparetheoverall restrictivenessofRoOsofagivenPTA.Thiswouldbeanalogousto theGeneralAgreementonTariffsandTrade(GATT)ArticleXXIVrestrictiononacustomsunion'sexternaltariff,whichcapsitattheaverageofthetariffspreviouslychargedbythemembers(BaldwinandThornton,2008).Thisprovidesanotherstrongreasonforconcludingthelong-standingnegotiationsonnon-preferentialrulesoforiginattheWTO.

Therationaleforcouplingconvergencewithcappingisthefollowing:largercumulationzonesincreasetrade,especiallyamongthecurrentspokecountries(seeSectionC).However,observedrestrictivenessofRoOsispositively correlated with the size of the cumulation zone, measured as the combined GDP of members(Estevadeordaletal.,2009b).LargercumulationzonescouldthereforeendupwithhighlyrestrictiveRoOsthat would serve to isolate production within each zone, increasing trade diversion and reducing globalefficiency.TradediversionforthirdnationsjustifiesinvolvementoftheWTOthroughmultilateralizationeffortsaimedatlimitingtheoverallrestrictivenessofRoOswithinagivencumulationzone.


173

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

(i) Overlapping jurisdictions

Article 23.1 of the WTO's Dispute SettlementUnderstanding (DSU) provides that “(w)hen Membersseek the redressofaviolationofobligationsorothernullification or impairment of benefits under thecovered agreements or an impediment to theattainmentofanyobjectiveofthecoveredagreements,theyshallhaverecourseto,andabideby,therulesandprocedures of this Understanding.” The AppellateBody has explained that “Article 23.1 lays down afundamental obligation of WTO Members to haverecoursetotherulesandproceduresoftheDSUwhenseeking redress of a violation of the coveredagreements” and “establishes the WTO disputesettlement system as the exclusive forum for theresolutionofsuchdisputes"32(AppellateBodyReport,US / Canada – Continued Suspension,para.371).

Recourseto theWTOdisputesettlementsystemmaybe had where a WTO member considers that anybenefits accruing to it directly or indirectly under theWTO agreements are being impaired by measurestaken by another member. Thus, in principle, a WTOmember may not have recourse to the WTO disputesettlementsystemtoprosecuteanallegedviolationofa PTA obligation.33 The potential for overlappingjurisdiction arises where an issue is regulated bothundertheWTOandthePTA.Porges(2010)observesthat “(a)lmost all PTAs overlap with the WTOAgreement,asbothPTASantheWTOrequirenationaltreatment and ban quantitative restrictions on trade.Indeed,manyPTAssimplyincorporateGATTArticlesIIIandXIbyreference”.

PTAs takedifferentapproaches tohow they regulatetherelationshipbetweentheirowndisputesettlementmechanism and that of the WTO. Porges (2010)identifies the following four approaches. Most PTAsuse the “fork-in-the-road” approach which allows theparty initiating the dispute to choose between themultilateral or the PTA fora. However, once it hasinitiatedthedisputeinoneforum,theotheroption(beitthePTAmechanismormultilateralone)isnolongeravailable to it. (See, for example, theNAFTAand theColombia-EU PTA.) The NAFTA has a provision(Article 2005(4)) under which the respondent partymayrequireanenvironmentaldisputetobeaddressedattheregionallevel,evenifthecomplainingpartyhasinitially chosen the multilateral fora. This provision isthe subject of apendingdisputebetween theUnitedStates and Mexico (discussed further below). A thirdapproach,whichhasbeenusedinfarfewerPTAs,istoestablish the PTA dispute settlement mechanism astheexclusiveforumwherethematterisoneregulatedunder the PTA. The EU-Mexico and EU-Chile PTAstake the opposite approach, requiring disputesinvolving a breach of a PTA obligation that areequivalent in substance to a WTO obligation to bebroughttotheWTO(Porges,2010).

Therearemanyfactorsthatcaninfluenceacountry'sdecisiontobringadisputetooneforumovertheotherwherethechoiceisavailabletoit.HorlickandPiérola(2007)examinea listoffactorsthatmayberelevant,including:thetypeofmeasurethatisbeingchallenged,theapplicable law, issuesofstanding, the time-frameof the proceedings, the remedies available, and thepossibility of other countries participating in thedispute as third parties. According to Horlick andPiérola(2007),“thecautiousdecision-makingprocessto choose the appropriate forum requires weighingandbalancingofall these factors inaccordancewiththeultimateneedsandobjectivesofthecomplainant”.

(ii) Concerns over the coexistence of the WTO dispute settlement system and PTA dispute settlement mechanisms

TheconcernsraisedaboutthecoexistenceoftheWTOdispute settlement system and the increasing numberof dispute settlement mechanisms of PTAs revolvearoundtwosetsofissues.Thefirstsetofissuesderivefrom the view that the proliferation of PTA disputesettlement mechanisms could undermine the WTOdispute settlement system's status as a public good.ThosewhoholdthisviewconsiderthattheWTOdisputesettlement system has positive externalities formembersthatarenotpartiestoaparticulardispute.

Drahos(2005),forexample,notesthattheinterpretationof the WTO agreements provides greater certainty toWTO rules. He also observes that when a respondentmember brings an infringing measure into conformitywith its WTO obligations, this will be of benefit to themembership at large because of the MFN principle.Thus, Drahos (2005) proposes that where a disputeconcernsamatter regulatedunderboth theWTOandthe PTA, it be brought to the WTO. Davey and Sapir(2009)takeadifferentapproachandproposethattheWTO should require members that do not belong to aPTA to be allowed to participate in the PTA disputesettlementforumasthirdparties.

Theothersetofconcernsrelatestothepossibilitythata dispute is brought under both the WTO and PTAdisputesettlementmechanisms.Herethere isconcernover the inefficiency of litigating similar matters twiceandmoreimportantlyaboutfairnesstotherespondentparty thatwouldhave todefend itself in two fora (seeKwakandMarceau,2006).ThereisalsoconcernaboutthemoreextremesituationinwhichtheWTOandPTAfora issueparallelorconsecutiveconflictingdecisions.One way of reducing the risks of this happening isthrough stricter jurisdictional clauses in PTAs thatprecludeadisputefromgoingtobothforaorforeclosebringingadisputetotheWTOoveramatterregulatedunderthePTA(MarceauandWyatt,2010).Thisraises,however, the question of the extent to which suchclauseswouldbindWTOadjudicatorybodies.


174

At the other extreme, there is the risk that thejurisdiction of the WTO could be gradually “carvedout”. For the moment, it appears that few PTAscompletely close off access to the WTO disputesettlement,butratherleavethechoiceofforumtothecomplaining party. The data discussed below showthat an important number of disputes betweenmembers that are partners in a PTA continue to bebroughttotheWTOdisputesettlementsystem.Somecouldalsoconceiveofmakingchanges to theWTO'sDispute Settlement Understanding to regulate therelationshipwithdisputesettlementforaofPTAs.Thisapproach, however, has not been taken up by WTOmembers in the negotiations to improve the DisputeSettlementUnderstandingcurrentlyunderway.

Theacademicliteraturediscussesothermorecomplexarrangementsthatcouldminimizetherisksofconflictsandpromotemorecoherencebetweenthemultilateraldisputesettlementsystemandthedisputesettlementsystems of PTAs. For example, there has beendiscussion of making exhaustion of PTA disputeresolution procedures a prerequisite to initiation ofWTO dispute settlement (see Kwak and Marceau,2006). Another suggestion is to create a system ofpreliminary references from the dispute settlementsystems of PTAs to the WTO dispute settlementsystemwheretheissueconcernstheinterpretationofprovisionsoftheWTOagreements(Kuijper,2010).

Commentators have also referred to severalinternationallawdoctrinesthatcouldbeusedtoavoidor resolve conflicts between overlappingjurisdictions.34 The doctrine of res judicata or finalityrefers tosituationswhereamatterhasbeendecidedbyacompetentadjudicativebodybarringitsrelitigationin subsequent proceedings. Lis Alibi Pendens, for itspart, refers to parallel proceedings and is a principlepursuant to which once a dispute is pending in oneforum, it cannot be brought before another forum.However, for these doctrines to apply, there must bean “inextricable link”between theproceedings,whichusuallyisunderstoodasanidentityofthepartiesandof the issues (Shany, 2005). Thus, application of thedoctrinescanbeavoidedincertaincircumstances.35

Undertheprincipleofcomityorforum non conveniens,an adjudicative body could seek to avoid exercisingjurisdictionoveradispute if it considers that itwouldbe more appropriate for another tribunal to exercisejurisdiction. There is considerable debate as to theapplicability of these principles to resolve a potentialconflict of jurisdiction involving the WTO disputesettlement system and a PTA dispute settlementmechanism(seeKwakandMarceau,2006).TheWTOdispute settlement system is available to WTOmembersasofright;theydonothavetoseekleavetostart theprocessunder thecurrent rules.Thus,somewouldconsiderthatapplyingtheseprerequisitescouldonlybeeffectedthroughachangeintherules.

As discussed below, questions about the relationshipbetweentheWTOdisputesettlementsystemandPTAdisputesettlementmechanismshavecomeup inonlyahandfulofWTOdisputes.Itshouldbenotedthatsofar concerns over potential conflicts have notmaterializedtotheextentthatsomehadfeared.36Thisis not to say that it is not important to think throughissuesarisingfromthecoexistenceofthemultilateralandPTAsettlementsystems.

(iii) Issues relating to PTA dispute settlement raised in WTO disputes

Asnotedearlier,issuestouchingontherelationshipoftheWTOdisputesettlementsystemandPTAdisputesettlementmechanismshavecomeup inahandfulofWTOdisputes.InArgentina – Poultry,Argentinaarguedthat Brazil was “estopped” from pursuing the disputeat the WTO because Brazil had first challenged theanti-dumpingmeasuresintheMERCOSURforum.ThepanelrejectedArgentina'sargument,notingthattherewas “no evidence on the record that Brazil made anexpressstatementthatitwouldnotbringWTOdisputesettlement proceedings in respect of measurespreviouslychallengedthroughMERCOSUR”.Moreover,thepanelfoundthat:

"Inparticular,thefactthatBrazilchosenottoinvoke its WTO dispute settlement rightsafter previous MERCOSUR disputesettlementproceedingsdoesnot,inourview,mean that Brazil implicitly waived its rightsunder the DSU. This is especially becausetheProtocolofBrasilia,underwhichpreviousMERCOSUR cases had been brought byBrazil, imposes no restrictions on Brazil'sright to bring subsequent WTO disputesettlement proceedings in respect of thesame measure. We note that Brazil signedthe Protocol of Olivos in February 2002.Article 1 of the Protocol of Olivos providesthat once a party decides to bring a caseunder either the MERCOSUR or WTOdispute settlement forums, that party maynot bring a subsequent case regarding thesamesubject-matterintheotherforum.TheProtocolofOlivos,however,doesnotchangeourassessment,sincethatProtocolhasnotyet entered into force, and in any event itdoesnotapplyinrespectofdisputesalreadydecidedinaccordancewiththeMERCOSURProtocol of Brasilia. Indeed, the fact thatparties to MERCOSUR saw the need tointroducetheProtocolofOlivossuggeststousthattheyrecognisedthat(intheabsenceof such Protocol) a MERCOSUR disputesettlementproceedingcouldbe followedbya WTO dispute settlement proceeding inrespectofthesamemeasure.”(PanelReport,Argentina–Poultry,para.7.38)


175

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

Alternatively,ArgentinaarguedthatifBrazilwereentitledtobringthedisputetotheWTO,“thenthePanelisboundbytheearlierMERCOSURrulingonthemeasureatissueinthiscase”as“theearlierMERCOSURrulingispartofthenormativeframeworktobeappliedbythePanelasaresult ofArticle31.3(c) of theVienna Convention”. Thisargumentwasalsorejectedbythepanel,whichexplaineditsreasonsasfollows:

"Rather than concerning itself with theinterpretation of the WTO agreements,Argentina actually argues that the earlierMERCOSURTribunal ruling requiresus torule in a particular way. In other words,ArgentinawouldhaveusapplytherelevantWTO provisions in a particular way, ratherthan interpret them in a particular way.However,thereisnobasisinArticle3.2oftheDSU,oranyotherprovision,tosuggestthat we are bound to rule in a particularway, or apply the relevant WTOprovisionsinaparticularway.Wenotethatwearenoteven bound to follow rulings contained inadoptedWTOpanelreports,soweseenoreason at all why we should be bound bytherulingsofnon-WTOdisputesettlementbodies.”(PanelReport,Argentina – Poultry,para.7.41)

Thepanelreportinthatcasewasnotappealed.

TheissuealsoaroseinMexico – Taxes on Soft Drinks,where the United States was challenging certain taxmeasuresandbook-keepingrequirementsimposedbyMexicoonsoftdrinksandotherbeverages thatusedsweetenersotherthancanesugar.MexicoarguedthattheWTOdisputewas“inextricablylinkedtoabroaderdispute regarding access of Mexican sugar to theUnited States' market under the NAFTA.” Mexicorequested the panel to decline jurisdiction over thedispute. According to Mexico, WTO panels have“implied jurisdictional powers” and these include “thepower to refrain from exercising substantivejurisdiction incircumstanceswhere 'theunderlyingorpredominantelementsofadisputederivefromrulesofinternational law under which claims cannot bejudicially enforced in the WTO, such as the NAFTAprovisions' or 'when one of the disputing partiesrefusestotakethemattertotheappropriateforum'.”

The Appellate Body affirmed the panel's finding that,under the DSU, it had no discretion to decline toexercise its jurisdiction in that case. Before reachingthis finding, however, the Appellate Body noted thatMexicohadnotarguedthatthesubjectmatternortherespectivepositionsofthepartieswereidenticalintheNAFTA and WTO disputes and Mexico had notidentifiedalegalbasisthatwouldallowittoraise,inaWTO dispute settlement proceeding, the marketaccess claims Mexico was pursuing under NAFTA.Furthermore, itwasundisputed thatnoNAFTApanel

hadyetdecidedthe“broaderdispute”towhichMexicohad alluded and Mexico had acknowledged that the“exclusion clause” of Article 2005(6) of NAFTA hadnotbeenexercised.Thus, theAppellateBodydidnot“express any view on whether a legal impediment totheexerciseofapanel'sjurisdictionwouldexistintheevent that features such as those mentioned abovewerepresent.”(AppellateBodyReport,Mexico – Taxes on Soft Drinks,paras.44-57)

Anothercasethathasbeendiscussedintheliteratureis a dispute between Canada and the United Statesover the imposition by the latter of anti-dumping andcountervailing duties on imports of softwood lumberfromthe former.Variousaspectsof thisdisputewerethesubjectof litigation inboth theWTOandNAFTA.At one point an injury determination made by the USinvestigating authority was found to be lacking by aNAFTA panel, while a WTO panel upheld it. Theconflict nevertheless was eventually resolved whenthe decision of the WTO panel was eventuallyoverturned upon review by the Appellate Body(Hillman,2009).37

The relationship between the dispute settlementmechanisms of NAFTA and the WTO has surfacedagaininamorerecentdisputebetweenMexicoandtheUnitedStates. In2009,Mexico requested thataWTOpanelexaminetheconsistencyofcertainrequirementsconcerning the labelling in the United States of tunaproductsas“dolphinsafe”(WT/DS381/4).Inresponse,the United States invoked Article 2005(4) of NAFTA,which it considers to require that in certain types ofdisputes, if thedefendingpartymakessucharequest,NAFTAratherthananyotherforumshouldbethesolevenue of the dispute. The United States initiated adisputeunderNAFTAchallengingMexico'sdecisionnotto move the dispute from the WTO to NAFTA, asrequested by the United States (United States TradeRepresentative (USTR), 2010). Both proceedings arepresentlyongoing.

(iv) WTO disputes between WTO members that are partners in a PTA

Inthissubsection,weexaminedataonWTOdisputesbetween WTO members who are partners in a PTA.Dataon thenumberofdisputes refer to requests forconsultations, which is the first step under the WTOdispute settlement procedures. The data concernparticipationbyWTOmembers(whoarePTApartners)as complainants and respondents, and does notinclude participation as third parties. Moreover, theexercise looks only at WTO dispute settlement anddoes not examine whether the disputes could havebeen brought under the PTA dispute settlementmechanism.Certainlyamorecompleteanalysiswouldrequire looking at whether the disputes could havebeentakentothePTAdisputesettlementmechanism.Notwithstandingthislimitation,thedataprovidesomeusefulinsights.


176

First, the data show that WTO members that arepartners inaPTAcontinuetohavefrequentrecoursetotheWTOdisputesettlementsystemtoresolvetradedisputes (the methodology employed in Tables E.1 toE.3 and Figure E.1 is explained in Box E.3). AsillustratedinTableE.1,82ofthe443disputesbroughttotheWTOupto2010werebetweencomplainantandrespondentmemberswhoatthetimewerepartnersina PTA. Disputes between PTA partners represent19per centof all disputes. The ratio is higherwherethecomplainant isadevelopingcountry(28percent)thanwhenitisadevelopedcountry(13percent).Thisis probably explained by the fact that the UnitedStates, the European Union, Japan and China do nothavePTAsbetweenthem,andtheyhavebeenpartiesinanimportantnumberofdisputes.

The largest share of the disputes between PTApartners brought to the WTO is made up of disputesbetweenparties toNAFTA,but therealsohavebeenWTO disputes between WTO members that arepartnersinotherPTAs,asillustratedinFigureE.1.

As depicted in Table E.2, the share of WTO disputesbetweenPTApartners increasedsteadilysince1995,reaching a peak of 50 per cent in 2005. Since then,thesharehasremainedaround30percent,althoughit was significantly below this number in 2009. ThesteadyincreaseintheshareofdisputesbetweenPTApartnersmaybepartlya reflectionof thenegotiationof new PTAs, but is more likely a reflection of thediversification of parties making use of the WTOdispute settlement system. An interesting point that

TableE.1:Frequency of requests for consultations, by development level and existence of PtAs in force between the parties, 1995-2010 (Totalnumberofpairsofmembers/pairswithaPTAinforce)

ComPLAInAnt

Developed Developing LDC totAL

DE

FE

ND

AN

T

Developed 154/24 115/10 0/0 269/34

Developing 102/8 71/39 1/1 174/48

LDC 0/0 0/0 0/0 0/0

totAL 256/32 186/49 1/1 443/82

Source:WTOSecretariatbasedonLegalDivision'sandRTA'sdatabases.Thetabletakesaccountof419requestsforconsultationsundertheWT/DSdocumentseriesasof31December2010,whichaccountforatotalof443pairsofmembers(i.e.complainant-defendant).SeeBoxE.3.

FigureE.1: PtAs in force at the time of the request for consultations, 1995-2010

Source:WTOSecretariat.

Arm

enia

- U

krai

ne

AS

EA

N

Aus

tral

ia -

New

Zea

land

CA

FTA

-DR

Chi

le -

ME

RC

OS

UR

Chi

le -

Mex

ico

Col

ombi

a-P

anam

a

DR

- C

entr

al A

mer

ica

EC

- N

orw

ay

EC

-Chi

le

EC

-Mex

ico

LAIA

ME

RC

OS

UR

ME

RC

OS

UR

- P

eru

Mex

ico

- N

icar

agua

Mex

ico

- N

orth

ern

Tria

ngle

NA

FTA

Pan

ama-

Mex

ico

SA

PTA

30

35

20

25

15

10

5

0

No.

of p

airs

of m

embe

rs


177

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

comes out of Table E.2 is that the share of disputesbetweenPTApartnersthatadvancetothepanelstage(45 per cent) is very close to the overall average,indicatingthatadisputebetweenPTApartnersisjustas likely tobesettledat theconsultationsstageasadisputebetweennon-PTApartners.

Table E.3 compares the number of times a particularWTO agreement has been the subject of a disputebetweenPTApartnerswiththenumberoftimesithasbeen invoked in all disputes. There are significantdifferences with respect to some of the agreements,thoughitmaybedifficulttodrawconclusionsinmanycases given the small number of disputes involvingcertain agreements. The most frequently citedagreementsindisputesbetweenPTApartnersaretheGATT 1994, the Anti-dumping Agreement, the

Subsidies and Countervailing Measures (SCM)Agreement, the Agreement on Safeguards, and theAgreement on Agriculture. Interestingly, subsidy andsafeguards disputes make up a larger share ofdisputes between PTA partners (intra-PTA) than ofoveralldisputes,whileintra-PTAdisputesinvolvingtheGATT1994representalowersharethanoverall.

Porges (2010) offers some possible explanations forthe continued use of WTO dispute settlement bymembersthatarepartnersinaPTA:theWTO's“familiarinstitutions” and “unblockable” dispute settlementprocedures; thepossibility tosuspendMFNtariffsandother WTO obligations (particularly where the PTA'smarginofpreferenceislow);thebroaderpoolofneutralpanellists; the broader issue scope of the WTO; thepossibility of forming alliances; access to assistance

TableE.2:Requests for consultations, by year and subsequent procedures, 1995-2010

Yearofrequestforconsultations

Request for consultations With a panel established

Totalrequestsfor

consultations

Totalpairsofmembers

Pairsw/aPTAinforceTotalpanelsestablished

Totalpairsofmembers

Pairsw/aPTAinforce

No. Share(%) No. Share(%)

1995 22 25 1 4.0 12 12 0 0.0

1996 42 50 3 6.0 19 24 1 4.2

1997 47 47 2 4.3 20 20 1 5.0

1998 43 43 3 7.0 15 15 1 6.7

1999 31 35 4 11.4 17 17 1 5.9

2000 30 30 7 23.3 11 11 3 27.3

2001 27 36 12 33.3 11 20 7 35.0

2002 34 34 7 20.6 23 23 5 21.7

2003 28 28 9 32.1 16 16 4 25.0

2004 20 20 5 25.0 9 9 1 11.1

2005 12 12 6 50.0 5 5 1 20.0

2006 18 18 6 33.3 13 13 4 30.8

2007 15 15 5 33.3 7 7 4 57.1

2008 17 17 4 23.5 10 10 4 40.0

20091 16 16 2 12.5 n.a. n.a. n.a. n.a.

20101 17 17 6 37.5 n.a. n.a. n.a. n.a.

totAL 419 443 82 18.5 188 202 37 18.3

Note:Thenumbersforeachrowwerecalculatedfortheyearinwhichtherequestforconsultationswasmade(i.e.theyalwaysrefertothesamegroupofrequestsforconsultationsmadeinthatyearandnottothenumberofpanelsestablishedduringaparticularyear).

1Thefiguresrelatingtothenumberofpanelsestablishedfortheperiod2009-2010werenotincludedbecausetheyarenotcomparable(i.e.duetoongoingprocedures).

Source:WTOSecretariatbasedonLegalDivision'sandRTA'sdatabases.SeeBoxE.3.


178

TableE.3: Wto Agreements cited in the requests for consultations, 1995-2010

WTOAgreement

no. of references to the Agreements1

In requests where a pair of members has a PtA in force

FrequencyShareof

references(percent)

Frequency

Shareofreferencesindisputesbetween

PTApartners(percent)

Shareofoverallreferences(percent)

GATT 1994 (adjusted)2 227 31.0 31 23.7 13.7

SubsidiesandCountervailingMeasures 86 11.7 16 12.2 18.6

Anti-dumping 84 11.5 27 20.6 32.1

Agriculture 66 9.0 12 9.2 18.2

TBT 41 5.6 7 5.3 17.1

Safeguards 38 5.2 15 11.5 39.5

SPS 37 5.0 6 4.6 16.2

ImportLicensing 34 4.6 4 3.1 11.8

TRIPS 29 4.0 1 0.8 3.4

TRIMs 27 3.7 1 0.8 3.7

GATS 22 3.0 3 2.3 13.6

ATC 16 2.2 1 0.8 6.3

CustomsValuation 15 2.0 5 3.8 33.3

RulesofOrigin 7 1.0 2 1.5 28.6

Gov.Procurement 4 0.5 0 0.0 0.0

totAL 733 100 131 100 17.9

1ReferencestotheDSUandtheMarrakeshAgreementEstablishingtheWTOwerenottakenintoaccount.

2SeeBoxE.3foradescriptionoftheadjustmentmethodologyused.

Source:WTOSecretariat.

fromtheAdvisoryCentreonWTOLaw;themultilateralsurveillanceprocess;theinstitutionalizedframeworkfortaking countermeasures; and the fact that the cost ofWTO dispute settlement is included in a member'sannualassessment,whileinmostPTAs,thepartiespaythepanellists,orpayforthecostofthetribunal.

(e) Caveats:mechanismsgeneratingnegativesystemiceffects

Some of the deep provisions contained in new-eraPTAs can contain discriminatory aspects, creating atensionwiththemultilateraltradingsystem.Themostprominent examples are the area of contingencymeasures(anti-dumpingandsafeguards).

(i) Discriminatory aspects in anti-dumping rules in PTAs

Recent research suggests that the risk of tradediversion may extend beyond tariffs. Prusa and Teh

(2010)uncoverwhattheycallaprotectionanaloguetothe trade creation-trade diversion impact of PTAs inthe area of anti-dumping. Anti-dumping provisions inPTAs result in members being spared from anti-dumping actions (“protection reduction”) while non-PTAmembersfaceevengreateranti-dumpingscrutiny(“protectiondiversion”).

TheideathatPTAsmayhavethisdistortionaryeffectisnotnew. Inaseriesofpapers,Bhagwati (1992:1993)and Bhagwati and Panagariya (1996) conjecture thatdue to its “elastic” and selective nature, anti-dumpingcanincreasetheriskofprotectiondiversionfromPTAs.According to their explanation, contingency measuresaredrivenbyimportvolume.Whoistargetedintheanti-dumpingpetition is entirely up to thediscretionof thedomesticindustry.

If anti-dumping provisions make PTA members moredifficult to sanction, the domestic industry will simplytarget other sources. As a result, we might see anincrease in anti-dumping protection directed towards


179

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

BoxE.3:methodology

A Data sources The tables and graphs in this section are based on a specialized dataset that was developed based ondatabasesmaintainedbytheLegalAffairsdivisionandtheRegionalTradeAgreementsunitoftheWTO.Thedatasetincludesatotalof419requestsforconsultationssubmittedundertheWT/DSdocumentseriesasof31December2010.

B "Pairs” of members (i.e. complainant-defendant)Sevenrequestsforconsultationsinvolvedmorethanonecomplainant(i.e.DS16,DS27,DS35,DS58,DS158,DS217andDS234),whichmeantitwasnotpossibletoestablishwhetheraPTAwasinforcebetweentheparties without creating a bias in the figures. For this reason, the 419 requests for consultations as of31 December 2010 were re-expressed as 443 pairs of complainants-defendants. Figures relating to theprevalenceofaPTAatthetimeoffilingtherequestforconsultationswerederivedonthisbasis.

C Adjusting the references to the GAtt 1994SantanaandJackson(2011)notedthat,becausecomplainantstendtocitea largenumberofagreementsandprovisionsintheirrequestsforconsultationsundertheDSU,frequencycountsofprovisionscitedtendtooverestimatetheimportanceoftheGATT1994.ThisismainlybecausereferencestocertainGATTArticlestendtobesubsidiaryinnaturewhenmadetogetherwithother“specialized”agreementsorevenArticlesintheGATT.Forexample,thecomplainantinatypicalanti-dumpingcasewillnormallyclaimthatthedefendantis in breach of provisions in the Agreement on Anti-dumping, Article VI of the GATT, and that the anti-dumpingduty imposed is inviolationof the tariffbinding (Article II:1(b)of theGATT)and theMFNclause(ArticleIoftheGATT).

Inspiteof the fourArticlescited, theGATTnormallyplaysasecondary role in thesedisputes.Similarly,arequest for consultations citing both Articles II and XIX of the GATT is almost certainly a case aboutsafeguards and not about tariff bindings. To minimize the incidence of those secondary references, andfollowingtheprincipleoflex specialis,SantanaandJacksonproposedamethodologythatdoesnottakeintoaccount references to certain Articles of the GATT 1994 when cited together with other provisions. Theadjustmentsareasfollows:

1.Article I was excluded when a reference was made in the same dispute to the Agreements on Anti-dumping, Safeguards, SCM (related to countervailing duties - CVD), sanitary or phytosanitary measures(SPS),ortechnicalbarrierstotrade(TBT),orwhenareferencewasmadetoArticleVIoftheGATT(i.e.CVDoranti-dumpingrelated).

2.Article II was excluded when a reference was made in the same dispute to the Agreements on Anti-dumping,CustomsValuations,SafeguardsorSCM(CVDrelated),orretaliationunderArticle22oftheDSU.ItwasalsoexcludedwhenareferencewasmadetoGATTArticlesVI(i.e.CVDoranti-dumpingrelated)orXIX(safeguards).

3.Article IIIwasexcludedwhena referencewasmade in thesamedispute toeither theSPSor theTBTAgreements.

4.ArticleVIwasexcludedwhenareferencewasmadeinthesamedisputetoAnti-dumpingorSCM(CVDrelated)Agreements.

5.ArticleXIwasexcludedwhena referencewasmade in thesamedispute to theSafeguards,SPS,TBTAgreements,aswellasGATTArticlesXIIandXIX.

6.ArticleXVIwasexcludedwhenareferencewasmadeinthesamedisputetotheSCMAgreement(relatedtotheprovisionofsubsidies),ortoArticles3,6-11oftheAgreementonAgriculture.

7.ArticleXIXwasexcludedwhenareferencewasmadeinthesamedisputetotheSafeguardsAgreement

Onthebasisofanadjusteddataset,anagreementisconsidered“cited”ifoneormoreofitsprovisionsarecitedinaspecificrequestforconsultations.


180

non-PTAmemberswheninfacttheinjurytodomesticindustry mostly stems from imports from other PTAmembers.38 The work by Prusa and Teh (2010)provides the first empirical support for thisconjecture.39 Their findings are especially relevantgiven the prominence of anti-dumping in the tradepolicy arena. Anti-dumping has long been thecontingency measure of choice and its prominencehasincreasedoverthepasttwodecades.Thenumberof countries using anti-dumping has increased five-foldandtheannualnumberofanti-dumpinginitiationshasmorethandoubled(Prusa,2005).

Figure E.2 shows a discernible difference in thepattern of anti-dumping activity of countries beforeandafterenteringintoaPTA.MeasuringtimerelativetotheyearthePTAwasenacted,yearzeroistheyearthePTAwasestablished,yeart–1istheyearbeforewhileyear t+1istheyearafter,etc.Noticethatduringthe years prior to the establishment of the PTAenactment,intra-PTAanti-dumpingactivityisgrowing.Thenumberofanti-dumpinginitiationsdropsharplyinthe year of establishment (t = 0) and remain muchlower in subsequent years as compared to the yearsprior to enactment. On average, during the ten yearsprior to establishment there were 29.5 anti-dumpingcases per year and during the ten years followingestablishmenttherewerejust23.6casesperyear.

There is another way to show how PTA membershipchangesthepatternofanti-dumpingactivity.TableE.4depicts anti-dumping filings when countries aredistinguishedbetween(i)thosewhoaremembersofaPTAand(ii)thosewhoarenot,andthetimeperiodisdistinguished between pre- and post-PTAestablishment. As seen, countries file about 58 percentofanti-dumpingcasesagainstnon-PTAcountries

priortoPTAenactmentbutaremarkable90percentfollowingenactment.Again,thisstronglysuggeststhatPTAsarechangingthepatternofprotection.

While illustrative, are these patterns statisticallysignificant (unlikely to have occurred by chance)?Furthermore, there may be other provisions in PTAsthatcanexplainthepattern intheanti-dumpingdata.PTAs often liberalize investment, thus increasing thelevel of FDI flows between PTA partners. The fall inanti-dumping activity between PTA members mightthus arise because imports are sourced frommultinational affiliates. Another concern is that theresults may be entirely driven by the big users(European Union and the United States) or targets(China)ofanti-dumping.

Prusa and Teh's econometric analysis (a methodknown as difference-in-difference regression)establishesthatthepatternsdonotarisesimplyfromchance.40 In addition, they find that PTAs cause asmuch as a 60 per cent reduction in anti-dumpingdisputes between PTA members. This result is notsolelydrivenby thosePTAs thathaveabolishedanti-dumping (forwhom intra-PTAanti-dumpingactivity is

FigureE.2: Intra-PtA anti-dumping initiations

Source:PrusaandTeh(2010).

30

35

40

45

50

20

25

15

10

5

0

Num

ber o

f AD

cas

es

29.5 cases/yr 23.6 cases/yr

Time (relative to enactment of PTA)

PTA enactment

-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10

TableE.4:Anti-dumping initiations by PtA status

target country

Non-PTAcountry PTAcountry

Pre-PTA 506 370

58% 42%

Post-PTA 3,554 375

90% 10%

Source:PrusaandTeh(2010).


181

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

essentially eliminated). When they only look at thosePTAs that have adopted PTA-specific anti-dumpingrules,theyfinda33-55percentreductioninintra-PTAanti-dumpingactivity. Theyfindnosignificantchangeinanti-dumpingactivityforPTAswithoutPTA-specificanti-dumpingrules.

Their econometric estimates also suggest that PTAscausea10-30percentincreaseinthenumberofanti-dumpingfilingsagainstnon-PTAmembers.Takingtheprotection reduction and diversion results together,they find that the reduction in intra-PTA activity ismorethanoffsetbytheincreaseinactivityagainstthefar larger set of non-PTA members. Overall, theyconclude that PTAs increase the number of anti-dumpingfilingsbyperhapsasmuchas10percent.

Theirresultsappeartobeextremelystable.Evenwhenthey excluded the EU, NAFTA and China individuallyfrom their analysis, the results were essentiallyunaffected.Totakeaccountofthepossibleeffectsofother PTA provisions, they included FDI flows and ameasure of the investment liberalization in each PTAbased on work done by Dee et al. (2006) and Dee(2008). While investment provisions in PTAs reducetheincidenceofanti-dumpingdisputes,theycontinuedto find that anti-dumping rules remain a significantindependentexplanationforthereductioninintra-PTAanti-dumpingcases.

(ii) Discriminatory aspects in safeguard rules in PTAs

There are typically two types of safeguard actionswhich are covered in PTAs: “bilateral” and “global”safeguard actions.41 Bilateral safeguard actions aremeanttoapplyonlytothetradeofotherPTAmembers.Theyprovidea temporaryescape formemberswhen,asaresultofundertakingthecommitmentsundertheagreement, increased imports from PTA partnersresultinseriousinjurytothedomesticindustry.Globalsafeguard actions, on the other hand, are triggeredunderGATTArticleXIX(EmergencyActiononImportsof Particular Products) and the Agreement onSafeguards. Multilateral rules require that anysafeguard measures be applied on a non-discriminatory basis. Typically, the PTA provisions onglobalsafeguardactionsspecifytheconditionsunderwhichPTApartnerscouldbeexcludedfrommultilateralsafeguardactionsinvokedbyamember.

While most of these PTAs state that their safeguardprovisionsareinaccordancewithordonotaffecttheirmembers'rightsandobligationsunderthemultilateralagreements,manygoontoexcludetheimportsofPTApartnersfromglobalsafeguardactions.42

The conditions under which imports from PTAmembers can be excluded from a global safeguardaction are if those imports do not account for asubstantial share of total imports and if they do not

contributetoseriousinjurytothedomesticindustryorthethreatthereof.43

TheAgreementonSafeguardsrequiresthatsafeguardmeasures be applied to all imports irrespective ofsource (non-discrimination). Thus, the exclusion ofPTA partners from a safeguard action poses apotential conflict between regional and multilateralrules.ThisconflicthasbeenaddressedinanumberofWTO dispute cases (Argentina–Footwear, United States–Wheat Gluten, United States–Line Pipe and United States–Steel).Inthesecases,theinvestigatingauthority had included imports from all sources inmaking the determination that imports were enteringin such increased quantities so as to cause seriousinjury to the domestic industry. However, instead ofapplying safeguard measures to all importsirrespective of their source, the country invoking thesafeguard action excluded its PTA partners.44 In allfour cases, the Appellate Body has ruled against theWTO member which included its PTA partners in thesafeguard investigation but excluded them in theapplicationofthesafeguardmeasure.

The key concept that underlines all these cases hasbeen called “parallelism”.45 In brief, parallelismprohibits any differences in the application ofsafeguards measures.46 In the case of PTAs,parallelism means that when a WTO member hasconducted a safeguard investigation consideringimports from all sources, it cannot, subsequently,withoutanyfurtheranalysis,excludeimportsfromPTApartners from the application of the resultingsafeguard measure. In order to be able to excludeimportsfromPTApartners,theinvestigatingauthoritymust establish explicitly that imports from non-PTAsources alone caused serious injury or threat ofserious injury to the domestic industry. Theinvestigatingauthority, initscausalityanalysis,shouldfurtherensure that theeffectsof theexcluded (PTA)imports are not attributed to the imports included inthesafeguardmeasure.

Whiletheelaborationoftheprincipleofparallelismbythe Appellate Body in these four cases has clarifiedone issue, WTO jurisprudence has not provided adefinitive ruling to what extent GATT Article XXIVcouldbereliedonbyaWTOmember toexcludePTApartners from the application of a safeguardmeasure.47 The provisions excluding PTA partnersfrom global safeguard actions raises concerns aboutincreased discrimination against non-members andtrade diversion. Although WTO dispute settlementpanelshaveruledagainstexcludingPTApartnersfromsafeguard measures if imports from those PTApartners had been included in the investigation, theyappeared tohavedonesoonquitenarrowgrounds–on the lack of parallelism in the application ofsafeguard measures. So far the Appellate Body hasnotruledonwhethersuchexclusionswillbejustifiableunder GATT Article XXIV. Conceivably, under a


182

different set of circumstances, exclusion of PTApartnersfromsafeguardmeasurescouldpassmuster.

(iii) Other mechanisms

The non-discriminatory nature of deep provisionsmight in principle create adverse systemic effects,namely political-economy and third-countryresistances to further multilateral liberalization. Ifpreferential liberalization is non-discriminatory innature, it might be opposed by political-economyforces, because higher market shares (and profits) intheothermember’smarketmightbemorethanoffsetby the loss of domestic profits vis-à-vis firms frompartnersandnon-members.48

Secondly, the non-discriminatory nature of deepprovisions may undermine the willingness ofdeveloping countries to engage in multilateralnegotiations with developed countries with theobjectiveofexchangingdeepregulatorycommitmentswith market access for goods (Chauffour and Maur,2011).Thisisbecausepreferentialtariffsareboundtobeerodedovertime,whereasregulatorycommitmentsarebothpermanentandMFN;therebytheycannotbeused as bargaining chips over time and vis-à-visdifferentcountries.

Thirdly, it has been argued that lock-in effects ofregulatoryharmonizationwithinagivenPTAmayhavenegative systemic effects (World Trade Organization(WTO), 2007). Competing PTAs with incompatibleregulatory structures and standards may lock-inmembers. This can constitute a threat to themultilateral trading system for two reasons. First, itundermines the principles of transparency andpredictability of regulatory regimes. Secondly, it mayhinderfurthermultilateralliberalization.Arecentstudy(Piermartini and Budetta, 2009) has found evidenceofdistinct “families”ofPTAswithdifferentiated rulesontechnicalbarrierstotrade.Thestudyshowsthatanumber of regional arrangements that have theEuropean Union as the hub include provisions toharmonizethestandardsofthespokepartnercountrytoEUstandards.TotheextentthattheadjustmenttoEuropean standards requires making investments,theseprovisionsmay lock-inacountrytotheregionalarrangement, thus making movement towardsmultilateralliberalizationcostly.

Finally,ithasbeenarguedabovethatthird-partyMFNclauseshavethepotentialtoreducethediscriminatorynatureof preferential agreements.However, a varietyofPTAsdonot contain third-partyMFNclauses (e.g.China – ASEAN). In this case, the provisions of theagreement effectively discriminate vis-à-vis thirdcountries, and there is the risk of discriminatorytreatmentbetweendifferentpartiesofdifferentPTAssigned by the same country (Houde et al., 2007). Intheir services and investment chapters, other PTAsincludesectoralexceptionstotheautomaticextension

ofthethird-partyMFNtreatment.Excludedsectorsdonot therefore automatically benefit from the bettertreatmentoffutureagreements.However,asreportedbyHoudeetal.,veryfewsectorsareconcerned.

Moreover,asarguedbyAdlungandMorrison(2010),anumberofagreementsexcludesomeofthepotentiallymost distortive types of intervention from third-partyMFNobligations(e.g.allsubsidiesareexcludedunderthe Australia-United States Free Trade Agreement –AUSFTA). The Economic Partnership Agreements(EPAs)thattheEUconcludedwithAfrican,Caribbeanand Pacific (ACP) countries contain MFN clausesrequiring that, if an ACP country concludes asubsequent PTA with a major trading economy otherthan theEU, suchas theUnitedStatesorBrazil, theEUshouldautomaticallyreceivethebenefitsconcededinsuchPTA.AsarguedbyPauwelyn(2009),inclusionof thisclause in recentEPAs iscontroversial. Itcouldin fact have a chilling effect on third countriesqualifying as “major trading economies” that werepreviously interested in concluding a PTA with ACPcountries.

3. RegionalismandtheWTO:historicalperspective

The MFN principle is at the core of the multilateraltradingsystem.Nevertheless,fromitsverybeginnings,the multilateral trading system has allowed somespaceformembercountriestogranteachothermorepreferential treatment under free trade areas orcustomsunions.Asonecommentatorhasputit,“(t)hereal thrust of the GATT had been to control andcontaindiscriminationratherthaneliminateit”(Hudec,1990). The rules applicable to free trade areas andcustomsunionsunderArticleXXIVoftheGATThavebeenincorporatedintotheWTOwithlittlechangeandthemanyinterpretativequestionsthatariseunderthatprovision remain intensely debated today.49 Althoughthere are still many observers who would like to seethe rules clarified and strengthened, recent effortshavefocusedonimprovingtransparency.

(a) TheoriginsoftheGATT

Preferentialtradingarrangementswereoneofthemainissues of concern of some of the countries thatparticipated inthenegotiationsfortheestablishmentofan International Trade Organization (ITO), whicheventuallybecamethebasis for theGATT. Inparticular,some countries saw the ITO negotiations as anopportunity to dismantle certain existing preferentialtrade arrangements, such as the preferences betweenterritoriesbelongingtotheBritishCommonwealth,whiletheBritishseemedwillingtodismantlethesepreferencesonlyiftheyobtainedmeaningfulaccesstoothermarkets,particularly the United States (Hudec, 1990). Indeed,several commentators note that this was an importantobjectivefortheUnitedStates,whichmadeaproposalto


183

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

allow preferences only between territories that formedpart of a customs union and later accepted interimarrangements that would lead to a customs union. Agroup of developing countries that included Syria andseveral Latin American countries sought to widen theexceptiontoincludefreetradeareas.

The language adopted at the Havana Conference of1947-48,whichwaslater incorporatedintotheGATT,allowed for free trade areas and customs unions, aswellasinterimarrangementsleadingtotheirformation.Several explanations have been put forward bycommentators to explain the eventual acceptance ofpreferencesunder free tradeareas, especiallyby theUnitedStates,whichinitiallyhadopposedthem.

Inarecenthistoricalstudy,Chase(2006)summarizesthe reasons that were traditionally given for theacceptance of free trade areas within the frameworkof the GATT: the need to compromise to reachagreements (Viner, 1950); discouraging aconsolidation of the Commonwealth preferences(OdellandEichengreen,1998);encouragingEuropeanintegration (Bhagwati, 1991; Odell and Eichengreen,1998); or pressure from certain developing countries(Haight,1972;Mathis,2002;WorldTradeOrganization(WTO), 1995). Chase (2006) disagrees with thesetraditional views and, based on his archival research,suggests that the United States and Canada weresecretly negotiating a bilateral free trade agreementand the United States changed its position on freetrade areas to accommodate this eventuality.AccordingtoChase(2006),theUnitedStatesdidnothave to make a new proposal because it saw anopportunity in the proposal allowing free trade areassubmittedbyLebanonandSyria.

ArticleXXIVof theGATT recognizes “thedesirabilityof increasing freedom of trade by the development,through voluntary agreements, of closer integration”,yetcautions“thatthepurposeofacustomsunionorofafree-tradeareashouldbetofacilitatetradebetweentheconstituent territoriesandnot to raisebarriers tothe trade of other contracting parties with suchterritories.” Article XXIV:5 establishes that theprovisionsoftheGATT“shallnotprevent,asbetweentheterritoriesofcontractingparties,theformationofacustomsunionorofafree-tradeareaortheadoptionofaninterimagreementnecessaryfortheformationofacustomsunionorofafree-tradearea”.

For purposes of Article XXIV, a customs union isunderstood as “the substitution of a single customsterritory for two or more customs territories, so that(i) duties and other restrictive regulations of commerce(except, where necessary, those permitted underArticlesXI,XII,XIII,XIV,XVandXX)areeliminatedwithrespect to substantially all the trade between theconstituentterritoriesoftheunionoratleastwithrespectto substantially all the trade in products originating insuch territories,and, (ii) ... substantially thesameduties

andotherregulationsofcommerceareappliedbyeachofthemembersoftheuniontothetradeofterritoriesnotincluded in the union”. A free-trade area is “a group oftwoormorecustomsterritories inwhichthedutiesandotherrestrictiveregulationsofcommerce(except,wherenecessary, those permitted under Articles XI, XII, XIII,XIV,XVandXX)areeliminatedon substantially all thetrade between the constituent territories in productsoriginatinginsuchterritories”.

Article XXIV sets out additional conditions that mustbe met by customs unions and free trade areas.Generallyspeaking,inbothcases,thedutiesandotherregulationsapplieduponformationmaynotbehigheror more restrictive than previously. In the case ofcustomsunions, thedutiesor regulationsmaynotbe“on thewhole”higher than the “general incidence”ofthe duties and regulations of commerce previouslyapplicable in the constituent territories. Interimagreements for the formation of a customs union orfreetradeareamustinclude“aplanandschedule”fortheformationofthecustomsunionorfreetradearea“within a reasonable length of time”. Certainnotification requirements also apply underArticle XXIV. Furthermore, Article XXIV includesprovisions on frontier traffic (Article XXIV:3) and onobservanceofGATTobligationsbyregionaland localgovernments and authorities (Article XXIV:12).Specific exceptions for preferences between certainneighbouring countries (for example, Lebanon andSyria; Belgium-Luxembourg-Netherlands) wereincludedinArticleIoftheGATT.

(b) DevelopmentsduringtheGATTyears

The creation of the European Economic Community(EEC) and its association agreements were theprincipalfocusofthediscussionsaroundArticleXXIVduring the early years of the GATT. CommentatorsdescribeintensedebatesamongtheGATTcontractingparties on the consistency of the EEC with therequirementsofArticleXXIV.ThecompatibilityoftheTreatyofRomewiththerequirementsofArticleXXIVwasnotresolvedbythecontractingparties.AsLadreitdeLacharrière(1987)notes, in1958,thecontractingparties considered it “more fruitful if attention couldbedirectedtospecificandpracticalproblems,leavingaside for the time being ... debates about thecompatibilityoftheRomeTreaty”withtheGATT.50

Eventually the GATT contracting parties opted forresolving some of the tariff issues surrounding theformation of the EEC as part of the Dillon Round(Hoda, 2001). The EEC association agreements withother countries were also the subject of intensedebates. Here the concern was about the lack of aclearcommitment to full liberalizationormembership.EFTA's notification also gave rise to discussions,particularlybecauseofitsexclusionofagricultureandfisheries (Hudec,1990).Anotheragreement thatwasnotifiedatthetimewasALALC,whichincludedseveral


184

LatinAmerican countries, andwhich raisedconcernsastotheambitiousnessoftheliberalizationprogrammeand its objective of promoting infant industries51

(Hudec,1990).

At the time, therewasnostandingbodyof theGATTthatwasresponsibleforreviewingagreementsnotifiedunder Article XXIV. Instead, these agreements werereviewed by individual working parties. GATTcontracting parties did not adopt definitive reportswithrespecttotheseagreements.Mostcommentatorsagreethat,despitethemanyquestionsraisedbysomecontractingpartieswithrespecttothePTAsthatwerenotified, what essentially developed was a policy oftolerance towards theseagreements.Jackson (1969)observes that generally speaking the practice of theGATT was of “a high degree of tolerance for a widediversity of regional arrangements”. Nevertheless, herecognizes that “legal discussions about criteria inArticleXXIVandconsultationsmayhaveenabled theinterestsofpartiesthatwerenotmemberstoregionalarrangements to influence those regionalarrangementsinawaythatsoftenedtheirdetrimentalimpactonthetradeofnon-members”.

Another importantdevelopmentduringtheGATTwastheadoptionoftheDecisiononDifferentialandMoreFavourable Treatment, Reciprocity and FullerParticipation of Developing Countries, commonlyknown as the “Enabling Clause”. In addition toproviding a basis for unilateral tariff preferences fordevelopingcountries,theEnablingClauseprovidesanexemption from theMFNobligation inArticle Iof theGATT for “(r)egional or global arrangements enteredinto amongst less-developed contracting parties forthe mutual reduction or elimination of tariffs and, inaccordance with criteria or conditions which may beprescribed by the CONTRACTING PARTIES, for themutualreductionoreliminationofnon-tariffmeasures,onproductsimportedfromoneanother”.

A totalof124agreementswerenotified to theGATTbetween1948and1994.Of these,however,only38remained in force in 1995 when the WTO wasestablished.AsexplainedinaWTOSecretariatReport,this reflects “inmostcasestheevolutionover timeoftheagreementsthemselves,astheyweresupersededby more modern ones between the same signatories(most often going deeper in integration), or by theirconsolidation into wider groupings” (Crawford andFiorentino,2005).

Discriminatory treatment underPTAsbecamea topicof increasing concern over the years. In 1983, theDirector-GeneraloftheGATTcreatedanindependentgroupofseveneminentpersonstostudyandreportonthe problems facing the international trading system.ThegroupissueditsreportinMarch1985.Commonlyreferred to as the “Leutwiler Report”, one of itsconclusions is that “(t)he rules permitting customsunions and free-trade areas have been distorted and

abused” and that “(t)o prevent further erosion of themultilateral trading system, they need to be clarifiedandtightened”.

The Report indicated that, while the EuropeanCommunity and EFTA met the conditions inArticleXXIV, “many agreements presentedunder therules, including some agreements between theEuropeanCommunity and its associates, fall short ofthe requirements”. It further cautioned that “(t)heexceptions and ambiguities which have thus beenpermitted have seriously weakened the trade rules,andmake itverydifficult to resolvedisputes inwhichArticle XXIV is relevant”. Accordingly, the Reportproposesthat“GATTrulesoncustomsunionsandfreetrade-areas should be examined, redefined so as toavoidambiguity,andmorestrictlyapplied,sothatthislegalcoverisavailableonlytocountriesthatgenuinelyuse it to establish full free trade among themselves”(Leutwiler,1985).

(c) PTAsintheUruguayRound

During the Uruguay Round, a group of countries thatincludedAustralia,India,Japan,NewZealandandtheRepublic of Korea favoured strengthening thedisciplines of Article XXIV. Japan, in particular,proposed among others, improving the consultationsbefore and after agreements were reached;establishingafirm time limiton “interimagreements”,to ensure that members moved to genuinely opentrade;clearlydefining“general incidence”ofdutiesorother regulations; and limiting the credit that a newcustomsunioncouldclaimif thegeneral incidenceofduties or regulations was actually lower than before.India,for itspart,proposedreviewingtherequirementthat duties and other restrictive regulations beeliminatedon“substantiallyalltrade”betweenthePTApartners(Croome,1995).

Inasecondsetofproposals,Japansoughttoimprovethe procedures for examination of preferential tradeagreements, suggesting the establishment of specialprocedures, separate from GATT dispute settlement,to assess and discuss compensation for damagescausedbypreferentialagreementstothetradeofnon-members. Some of those who opposed this proposalsuggested that surveillance of preferential tradeagreements could be undertaken under the newly-created Trade Policy Review Mechanism (Croome,1995).

AnotherissuediscussedduringtheUruguayRoundinconnectionwithpreferentialtradeagreementswastheobligationinArticleXXIV:12relatingtofederalstates.This point was initially raised by India, but was latertakenupbytheEuropeanCommunity,whichpresentedaproposal to tightenArticleXXIV:12byaffirming thefull responsibility of GATT members for measurestaken by their regional or local governments orauthorities(Croome,1995).


185

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

Ultimately, the discussion coalesced around the ideaof negotiating an Understanding on Interpretation ofArticleXXIV,whichwould focuson thecalculationofthelevelofdutiesbeforeandafteracustomsunionisformed,reasserttheobligationtocompensate,setoutrequirements for interim arrangements, limit the“reasonable period of transition” to ten years unlessotherwise authorized, and acknowledge that mattersarising under Article XXIV could be submitted todisputesettlement.

Despite initial opposition from the EuropeanCommunity(whichwantedfullercreditincompensationnegotiations for tariff reductions made by groupmembers and was dissatisfied with the text onArticle XXIV:12), India (which considered the textdisproportionately weak), and Yugoslavia (whichobjected to the text on Article XXIV:12), theUnderstanding on Interpretation of Article XXIV wasadopted and became part of the Uruguay Roundagreements(Croome,1995).

An additional development of significance during theUruguay Round was the inclusion in the GATS of aprovisiononpreferentialagreements relating to tradeinservices.52

(d) DevelopmentsintheWTO

(i) Committee on Regional Trade Agreements

The WTO Committee on Regional Trade Agreements(CRTA)wasestablishedbytheGeneralCouncilin1996(WT/L/127). It was initially foreseen that the CRTAwouldcarryout theexaminationsof theregional tradeagreements notified to the WTO, thus taking over thefunctionsoftheindividualworkingpartiesoftheGATT.Despite the establishment of the CRTA in 1996, theexamination of RTAs resulted in stalemate. Between1996 and 2001 not a single examination report hadbeen adopted by the CRTA, in part due to continuingdisagreements over the inherent ambiguities in GATTArticleXXIV, the lackof informationsubmittedbyRTAparties, and the fact that the determination ofconsistency was to be made by all WTO members,includingthosewhoseRTAswereunderexamination.

In December 2006, WTO members adopted on aprovisional basis a new transparency mechanism forregional trade agreements (WT/L/671).53 The newmechanism calls on members to provide an “earlyannouncement”oftheirinvolvementinnegotiationsfora regional trade agreement, requires members topromptly notify a newly concluded regional tradeagreement, and sets out a schedule for itsconsideration by WTO members.54 The mechanismprovides that consideration of notified regional tradeagreements should conclude within a year from thedate of notification. For this purpose, parties to a

regional trade agreement are required to submitcertain data to the WTO Secretariat, such as tariffconcessions, MFN duties, rules of origin and importstatistics.

Based on this data, the text of the agreement, andinformation from other sources, the WTO Secretariatprepares a factual presentation that is intended toassist members in their consideration of the notifiedregionaltradeagreement.WTOmembersarecurrentlyreviewing the transparencymechanismwithaviewtomaking it permanent. The transparency mechanismplacesemphasisonthe“consideration”ofRTAsratherthan on their “examination”, which may be viewed bysome as a tacit acknowledgement by members thattheir interests would be better served by focusingeffortsonimprovingtransparency.

WTO members are also engaged in negotiations aspart of the Doha Round aimed at “clarifying andimproving disciplines and procedures under theexisting WTO provisions applying to regional tradeagreements.” Negotiations are to “take into accountthe developmental aspects of regional tradeagreements” and have been taking place in theNegotiatingGrouponRules.55

The CRTA reported that, as of 1 November 2010, 479regionaltradeagreements,countinggoodsandservicesnotificationsseparately,hadbeennotifiedtotheGATT/WTO, 288 of which were in force at the time.56 Thesefigures correspond to 375 “physical” agreements, ofwhich197were inforce(117goods,1servicesand79goodsandservices).Ofthe288notifications,174werenotifiedunderGATTArticleXXIV,31undertheEnablingClause, and 83 under GATS Article V. A total of 92regional trade agreements had been considered underthe Transparency Mechanism since its adoption inDecember2006.57

(ii) Dispute settlement

Despite the concerns expressed by many observersregarding the compatibility of many notified regionaltrade agreements with Article XXIV of the GATT,issuesrelatingtoregional tradeagreementshavenotfigured prominently in WTO dispute settlement. ThemostimportantissuethatcameupwasthequestionofwhethertheconsistencyofaregionaltradeagreementwithArticleXXIVcouldbeexamined inWTOdisputesettlement.InTurkey–Textiles,theAppellateBodyheldthat panels have the authority to examine whether aregional trade agreement meets the requirements ofArticle XXIV. The burden of establishing that theregional agreement meets the requirements ofArticleXXIVfallsontherespondentWTOmembertotheextentthatit invokestheregionalagreementasadefencetojustifyadiscriminatorymeasure.

TheavailabilityofWTOdisputesettlementtochallengeregional trade agreements has given rise to mixed


186

reactions from commentators. Roessler (2000) hasargued that the examination of the consistency ofregional trade agreements was a matter that shouldhavebeen reservedexclusively to theWTO'spoliticalorgans and specifically to the CRTA. By contrast,Davey (2011) has suggested that WTO disputesettlement could be used to further clarify thedisciplinesofArticleXXIV.WTOmemberssofarhavebeen reluctant to use the WTO dispute settlementsystem to enforce the obligations of Article XXIV oftheGATTandArticleVoftheGATS.

Issuesconcerning the relationshipbetween theWTOdisputesettlementsystemandthedisputesettlementsystems of PTAs have been discussed in connectionwithahandfulofWTOdisputes.Thesedisputeswereaddressed in subsection E.2. In this subsection, weaddress the small number of disputes in whichArticleXXIVhasbeenexplicitlyraised.

Asnotedabove, thecasethathasdealtmostdirectlywith the requirements of Article XXIV is Turkey – Textiles.Inthiscase,theAppellateBodyexaminedtherequirementsapplicabletocustomsunionsundersub-paragraph5ofArticleXXIVandexplainedthatapartyinvoking this provision to justify an otherwise WTO-inconsistentmeasuremustestablishthatthefollowingtwo conditions have been fulfilled. First, it “mustdemonstrate that the measure at issue is introducedupontheformationofacustomsunionthatfullymeetsthe requirements of sub-paragraphs 8(a) and 5(a) ofArticle XXIV”. Secondly, it must show that “theformationofthatcustomsunionwouldbepreventedifitwerenotallowedtointroducethemeasureatissue”(AppellateBodyReport,Turkey – Textiles,para.58).

Article XXIV has also been raised in the context ofseveral safeguard cases, where the issue has beenwhether a WTO member could exclude one of itspartners in a preferential trade agreement from theapplicationofasafeguardmeasure indeparturefromArticle 2.2 of the Agreement on Safeguards. ThesecaseswerediscussedinsubsectionE.2.

Ameasure takenpursuant toaPTAbecamerelevantin a dispute in which Brazil invoked the GeneralExceptions in Article XX of the GATT to justify animport ban on retreaded and used tyres on publichealth grounds. As a result of a decision by aMERCOSURtribunal,however,theimportbanwasnotapplied to imports of remoulded tyres fromMERCOSURmembers.

The panel found that “(t)he exception of remouldedtyres originating in MERCOSUR therefore does notseem to be motivated by capricious or unpredictablereasons”andthat“(t)otheextentthattheexistenceofsome discrimination in favour of other members of acustomsunionisaninherentpartofitsoperation,thepossibility that such discrimination might arisebetween members of MERCOSUR and other WTO

Members as a result of the implementation of theMERCOSUR Agreement is not, in our view, a prioriunreasonable”.

Thepanelneverthelessnotedthat“thefactthatwegivedue consideration to the existence of Brazil'scommitments under MERCOSUR in our assessmentdoes not imply that the exemption must necessarily bejustified. Rather, we must now examine the manner inwhichtheimportbanisapplied,takingintoaccounttheexistenceofanexemptionforMERCOSURmembers,inorder to determine whether the discrimination arisingfrom the MERCOSUR exemption is arbitrary orunjustifiable”.Becausethepanelfoundthatthe“volumesofimportsofretreadedtyresundertheexemptionappearnot to have been significant”, it concluded that “themeasure'sabilitytofulfilitsobjectivedoesnotappeartohavebeensignificantlyunderminedbytheoccurrenceofimports fromothersources,even in thepresenceofanexemptionforMERCOSURimports”.

Therefore, thepanelconcluded that “theoperationofthe MERCOSUR exemption has not resulted in themeasure being applied in a manner that wouldconstitutearbitraryorunjustifiablediscrimination”.Thepanel also relied on its analysis of the volume ofimports to conclude that the MERCOSUR exemptiondid not result in the import ban being a disguisedrestrictiononinternationaltrade(PanelReport,Brazil-Retreaded Tyres,paras.7.272-7.289and7.354-7.355).

TheAppellateBodydisagreedwiththepanel'sfinding,explaining that the ruling of the MERCOSUR arbitraltribunal was not an acceptable rationale for thediscrimination, because it bore no relationship to theprotection of public health, the legitimate objectivepursued by the import ban under Article XX(b), and“even[went]againstthisobjective,tohoweversmalladegree”. The Appellate Body held “that theMERCOSURexemptionhasresultedintheImportBanbeingappliedinamannerthatconstitutesarbitraryorunjustifiablediscrimination”.

Moreover,theAppellateBodydisagreedwiththepanel'sconsideration of the volumes of imports. According tothe Appellate Body, the analysis of “whetherdiscrimination is 'unjustifiable' will usually involve ananalysis that relates primarily to the cause or therationaleofthediscrimination”,anddoesnotdependon“the quantitative impact of this discrimination on theachievementof theobjectiveof themeasureat issue”.For thesamereason, theAppellateBodyreversedthepanel'sfindingthattheimportbanwasnotappliedinamanner that constituted a disguised restriction oninternational trade (Appellate Body Report, Brazil–Retreaded Tyres,paras.228-229).

A point emphasized by the Appellate Body was that“before the arbitral tribunal established underMERCOSUR, Brazil could have sought to justify thechallenged Import Ban on the grounds of human,


187

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

animal, and plant health under Article 50(d) of theTreatyofMontevideo”,yetBrazildecidednottodoso.TheAppellateBodyobservedthat“Article50(d)oftheTreaty of Montevideo, as well as the fact that Brazilmight have raised this defence in the MERCOSURarbitral proceedings, show, in our view, that thediscrimination associated with the MERCOSURexemptiondoesnot necessarily result froma conflictbetweenprovisionsunderMERCOSURandtheGATT1994”(AppellateBodyReport,Brazil–Retreaded Tyres,para.234).

4. TherelationshipbetweenPTAsandtheWTO

(a) Coherenceininternationaltradegovernance

The quest for coherence between regionalism andmultilateralism is nothing new. In the early days of themultilateral trading system, economic thinking focusedonthewelfareeffectsofPTAs.AsexplainedinSectionC,themainfindingwasthattheseeffectswereambiguousformembersandgenerallynegativeforthirdparties.AsPTAs were mostly about tariff reductions, multilateralmarket opening which, even if it does not meancompletelyopentrade,reducesdiscrimination,wasseenas superior to preferential opening.58 In this context,ensuring coherence was understood as accepting thatPTAsandthemultilateralsystemcouldcomplementeachotherwhileimposingdisciplinesaimedatminimizingthenegativeeffectsthatPTAscouldhave.

As mentioned above, in the 1990s, the expansion ofregionalism brought the coherence issue back to theforefront.Manyanalysts re-examined the relationshipbetweenthetwoapproaches,thistimefocusingonthesystemiceffectsof regional integration.TheyshowedthatPTAscouldeitherbesteppingstonesorstumblingblocksontheroadtomultilateralmarketopening.Thisliterature,however,didnotprovidemuchguidanceonhowtoimprovecoherence.

WhethertheyviewthemultilateraltradingsystemandPTAs as complementing each other or think that themultilateral system is simply superior to the regionalapproach, observers broadly agree that “the case forfindingwaysofstrengtheningtheabilityoftheWTOtoinfluenceanddisciplinePTAs,oratleasttoblunttheirmoreexclusiveanddistortingfeatures,remainsstrong”(Low,2008).59Subsection3hasshownhowsinceitsinception the multilateral system has accommodatedpreferential trade agreements. GATT/WTO membershave largely taken a non-confrontational and non-litigiousapproach.Approachestoimprovingcoherencehave focused on the weaknesses of multilateraldisciplines and how they could be fixed. This sub-section summarizes the debate and briefly discussesthe main proposals. It appears that feasibility is themainissueandpoliticaleconomyisthekey.

Recentdevelopments inPTAactivitymaywell changethe perspective on coherence. As documented inSection B, PTA activity accelerated noticeably from1990 onwards. The number of PTAs had more thandoubled by 1995 and more than quadrupled by 2010,resulting in close to 300 active PTAs today. Aspreviously discussed, new PTAs – or at least some ofthem–arequalitativelydifferentfromolderones.Whilepart of recent PTA activity has consisted of theconsolidation and rationalization of bilateralarrangements, there has also been a trend towardsbilateral deals across the world. Since 1995, PTAactivity has increasingly crossed regional boundaries.ThecoverageofPTAsintermsofbothpolicyareasandproductshasalsowidenedanddeepenedovertime.

This has led some observers to think that regionalismhas entered a “new era” where the old analyticalframework is no longer valid and where ensuringcoherencenolongermeansmerelyimposingmultilateraldisciplines on discrimination. Baldwin (2010), forinstance,seesrecentPTAsasprovidingtheframeworkto underpin the “production unbundling” thatcharacterizesagrowingshareofworldtrade.Inhisview,twenty-firstcenturyregionalismismoreaboutreducingfrictional tradebarriersandthecostofdoingbusinessandremovingdomesticentrybarriers thanabout tariffpreferences. Given that preferential agreements onsuch behind-the-border measures do not typicallyinduce trade diversion, their systemic implicationscannot be analysed using the traditional stumblingblock/steppingstoneframework(seeSectionC).

The political economy of more recent PTAs is alsoabout a lot more than preferential tariffs. First,accordingtoBaldwin(2010),onlyafewcountriescanplayaleadingroleinsuchagreements.PTAsmotivatedbyproductionsharing,inparticularbetweendevelopedanddevelopingcountries,maybeseenasanexchangeof factories for the relaxation of behind-the-borderbarriers andassurances tooffshoringfirms that theirinvestmentsandintellectualpropertywillbesafe.Fewcountries, inBaldwin'sview,havethesortoffactoriesthatcanbeexchangedfordeepreformofbehind-the-bordermeasures.

Secondly, negotiating behind-the-border reform in theWTOmaynothelptodirectlyfosterinwardinvestment.Thirdly,thenatureofbehind-the-borderpoliciesmakesit difficult to multilateralize PTAs. For example, theprinciple of subsidiarity (see below) may apply in thatsomeareasmaybestbedisciplinedat the regional orbilateral level. These considerations lead Baldwin(2010) to the conclusion that “it is, thus, possible andevenlikelythatthenewdisciplinesformanindependentsystemofgovernancethatdoesnot intersectmuch,orat all, with Marrakesh rules”. If this is the case, thecoherencechallengeposedbyrecenttrendsinregionalagreements may be quite different from that arisingfromdiscriminatorytariffreductions.Itmaybethatnewinternational trade rules are being negotiated and


188

decidedoutsidetheWTOinasettingwheredifferencesin power are greater and in the absence of the basicprinciplesofnon-discriminationandreciprocity.60

Whether and how this new challenge needs to beaddressedisanopenquestion.Furtherresearchwillbenecessary tounderstandbetter thesystemiceffectsofdeep integration. One issue that may require furtherinvestigation is the effects of power asymmetries andoptionsformitigating them.Also,asalreadymentioned,the principle of subsidiarity could be used to assesswhether measures agreed at the bilateral or regionallevelneedtobesubmittedtomultilateraldisciplines.6162Thisprinciplestatesthat“actiontoachieveagreedpolicyobjectives should be taken at the lowest level ofgovernment capable of effectively addressing theproblem at hand” (Sauvé and Beviglia-Zampetti, 2000).Because countries have different tastes, cultures,endowments,orinstitutions,theirsocialchoicesdiffer.Atthesametime,efficiencycriteriasuggestthatregulatoryregimesshouldapplytothelargestpossiblecommunities.

Given this trade-off, the subsidiarity principle statesthatthedeterminationofregulatoryregimesshouldbeas decentralized as possible unless action in onejurisdiction has an impact in others (spillovers) –resulting in cross-border external effects(externalities),orthecreationofeconomiesofscaleorpublic goods, in which case they too should beconsulted.Inotherwords,“unlesstherearesignificantspillovers,thereisnoefficiencycaseforimposingonesetof standardsacrossdifferent regulatorydomains”(RolloandWinters,2000).

A basic rationale for international cooperation onregulation is that the cost of complying with differentstandards may be high. Economies of scale (acrosscountries)andscope(acrossissues)arelikelytoexistinrule-making. However, conflicts of interest can arisebetween countries with permissive regulations andcountries with strict regulations that make multilateralcoordination hard and perhaps in some instancesundesirable. If these factors are sufficiently prevalent,mutual recognitionandharmonizationofproductnormsand testingmayworkbetterbilaterallyandplurilaterally(between relatively similar countries) than multilaterally.Whiletheremaybeconcernsregardingpossiblenegativethird-party effects of common or mutually recognizedstandards and shared conformity assessment in PTAs,empiricalevidencesuggeststhattheEU'ssinglemarketprogrammeincreasedaccessatleastasmuchforthird-partyfirms(MayerandZignago,2005).63

Finally, the fact that PTAs where preferential tariffsare still important have not disappeared means thatboththenewandtheoldcoherencechallengesneedtobetackledatthesametime.Theevidencepresentedin Section D suggests that only a (relatively small)number of the new PTAs have little or nothing to dowithpreferentialtariffs,andthattariffpreferencesstillplay a role in many new agreements. The next sub-

section provides a short summary of the debate onexisting multilateral disciplines. This overview isfollowedbyadiscussionofsomeof themainoptionsforimprovingcoherence.

(b) MultilateraldisciplinesonPTAs

As explained in subsection 3, the multilateral systemhas generated three core provisions to deal withregionalism. The first provision is GATT Article XXIV,whichallowsdeparturesfromMFNforcustomsunionsandFTAs.TheUruguayRoundUnderstandingon theInterpretation of Article XXIV of the GATT seeks toclarify the criteria and procedures for assessing neworenlargedagreementsand to improve transparency.The second provision is the “Enabling Clause”, whichrelaxes (some of) the GATT provisions on PTAs fordeveloping countries in the name of “special anddifferential treatment”forthisgroupofcountries.ThethirdprovisionisArticleVoftheGATS,whichsetsoutthe rules forPTAs in theservicesfield.Asdiscussedabove,WTOmembersmorerecentlyalsoadoptedonaprovisional basis a new transparency mechanism forregionaltradeagreements.

Over the years, a number of concerns regarding theeffectiveness of the multilateral oversight of regionalagreementshaveemerged (Davey,2011;Low,2008).First,ithasbeenarguedthatanumberofArticleXXIVprovisions defy uncontested legal interpretation and,more generally, are deficient.64 The debate hasfocusedontheinterpretationof:

• Paragraphs 5(a) and 5(b) of GATT Article XXIV,whichstatethat“thedutiesandotherregulationsofcommerce” imposedon thirdpartiesshouldnot “onthe whole be higher or more restrictive than thegeneral incidence” of the pre-PTA duties andregulations;65

• Paragraphs 8(a) and 8(b) of GATT Article XXIV,which state that duties and other restrictiveregulations of commerce should be eliminated withrespect to “substantially all the trade” between theconstituent territories,andParagraph1(a)ofGATSArticle V, which states that an RTA should have“substantialsectoralcoverage";

• Paragraph3oftheUnderstandingontheInterpretationof Article XXIV of the GATT, which states that the“reasonable length of time” within which theimplementation of an RTA should take place shouldexceedtenyearsonlyinexceptionalcases.

Secondly, several gaps in the GATT/WTO legal andinstitutional framework have been identified. Theabsenceofdisciplinesregardingrulesoforiginforfreetradeagreements, inparticular, hasbecomean issuewith the multiplication of such agreements and theresulting expansion of a spaghetti/noodle bowl.Similarly, there is no indication regarding howagricultural tariff quotas should be treated in


189

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

preferential agreements, whether members of suchagreementsareallowedtoexcludetheirPTApartnersfromtheapplicationofcontingencymeasuresappliedto the trade of third parties, or whether PTA partiesmay or may not apply safeguards on their trade witheach other. Another question that has been raised iswhether the special and differential treatmentprovisions for developing country PTAs should beextendedbeyondthoseintheEnablingClause.66

Thirdly, while the law of the GATT/WTO may haveinfluenced PTA negotiations, in practice, it has neverbeen used to impose discipline on discriminatoryreciprocaltradeagreements(Davey,2011;Low,2008).Governments have almost never agreed throughestablished procedural arrangements whether anygiven PTA is in conformity with the multilateral rules.Procedural requirements such as notifications havebeenpartiallyobservedatbestanddisputesettlementfindingshavenothelpedaddressexistingweaknessesinthedisciplines.

In theeyesofsomeobservers, it is revealing that theTransparency Mechanism for Regional TradeAgreements is the only result of the Doha Roundnegotiations that has been allowed to go forwardindependently of the full results of the Round.67 Thissuggests both that WTO members are aware of theneed to understand better what regional tradeagreements are about and that they continue toprivilege a cautionary approach (Low, 2008). Othersgo even further and consider that the TransparencyMechanism advantageously substitutes the “old”reviewprocess(Mavroidis,2010).Withtradediversionreduced as a result of multilateral tariff reductions,along with empirical evidence suggesting that PTAscan be welfare improving, and with PTAs covering anumber of issues not covered by the WTO, existingrules are considered to be of limited relevance.Mavroidis (2010) argues that the TransparencyMechanism should become the de jure new forum todiscussPTAswithinthemultilateraltradingsystem.68

(c) Possiblewaystoimprovecoherence

Thisreporthasdiscussedtheideathattheremaybeacaseformaintainingseparateregimesforregionalandmultilateralcooperation.Thiswouldbethecasewhereparticulartypesofcooperationaremoreappropriatelymanaged at the regional rather than the multilaterallevel.Bythesametoken,thereareissuesthatcannotbe addressed adequately at the regional level. Inbetween these two polar realities, the coherencequestionarises.Essentially,thechallengeistoidentifywheretherearegainsfromensuringgreatercoherenceamong PTAs and between PTAs and the multilateraltradingsystem.

Anumberofdifferentapproacheshavebeenproposedfor improving coherence between PTAs and themultilateral trading system (Davey, 2011; Low, 2008;

Sutherland Report, 2004; The Warwick Commission,2007; World Trade Organization (WTO), 2003). Thissubsection reviews theseproposals andgroups themunder four headings: i) accelerating multilateral tradeopening; ii) fixing the deficiencies in the WTO legalframework; iii) adopting a softer approach as acomplement to the existing legal framework; andiv)multilateralizingregionalism.Theseapproachesarenot necessarily mutually exclusive. They all aim atreinforcing compatibility and coherence, whichessentiallymeansmakingsurethatPTAscontributetotrade cooperation and opening in a fundamentallynon-discriminatorymanner.Theydiffermainlyintermsof what they see as a politically feasible strategy toreachthisobjective.

LoweringMFNtariffswouldreducediscriminationandthereby blunt the adverse effects of PTAs. TheSutherlandReport,forinstance,recommendedthatalldeveloped country tariffs should be bound at zero inWTO members' schedules of commitments at someagreed upon time in the future. While a reduction tozero of all developed country tariffs on industrialproductsmaynotseemimpossibletoachieveinanottoo distant future, the Doha Round negotiationssuggestthatthismaynothappenwithoutameasureofreciprocity from emerging economies. As for theelimination of all tariffs on agricultural products, thisdoesnotseemtobepolitically feasible in thecurrentcontext.Also,bindingall tariffsatzeromaytakecareof tariff-induced trade diversion but it would noteliminate all potentially adverse effects of deeperintegrationmeasures.

As for the idea of filling gaps in the WTO legalframework, the Doha Round includes a mandate tonegotiate with a view to “clarifying and improvingdisciplines and procedures under the existing WTOprovisionsapplyingtoregionaltradeagreements”.Thenegotiationshavebeenpursuedalongtwotracks.Onthe one hand, members addressed procedural issuesrelating to the transparency of PTAs. On the otherhand, they tried to identify issues for negotiation,including “substantive” issues, such as systemic andlegal issues.69Asalreadymentioned,negotiationsonthe procedural issues resulted in the adoption on aprovisionalbasisofanewtransparencymechanismforregional trade agreements (WT/L/671). Thenegotiations on the “substantive” issues have so fargeneratedproposalsbyvariousmembersmainlyaimedatclarifyingtheprovisionsofGATTArticleXXIV.Whiletheseproposalscontributeusefullytothedebate,theydo not seem to have converged towards any form ofconsensusonpossiblereformstotherules.70

This should not come as a complete surprise aspreviousdiscussionshavenotledtomuchprogressonsubstantiveissues.71Onepossibleexplanationforthelack of progress is that members who have enteredPTAs in the past may be reluctant to sign off onclarifications in the rules that might suggest that the


190

PTAstheybelongtodidnotcomplywithArticleXXIV(Davey, 2011). Considering that efforts to clarifyconcepts such as “substantially all trade”, “otherrestrictive regulations of commerce”, etc. have hadlimited success so far, it seems unlikely that thesecond option referred to above – that of clarifyingandstrengtheningexistingrules–wouldbeviable.

Moreover, WTO members have been reluctant to usetheWTOdisputesettlementsysteminordertoclarifyexistingrulesanditdoesnotseemlikelythattheywillchange thisposture in thenear future.Thisdoesnotmeanthatrevisedandimprovedruleswillnotonedaybe part of any significant progress towards morecoherence, only that this does not seem to be apromising starting point. In that context, economicanalysiscouldhelpstrengthentheexistingprovisions.It shows, for example, that the condition in GATTArticle XXIV that the protection applicable to non-members should not increase with the creation orextension of a PTA will not necessarily protect thelatterfromawelfareloss.72

Thethirdoptionnotedabovewouldbetoadopta“softlaw” approach to complement the “hard law” and thedisputesettlementmechanism.Thereisnoagreementintheliteratureregardingthedefinitionoftheconceptof “soft law”, although legal scholars often seem todefinehardlawasbindingandsoftlawasnon-binding(ShafferandPollack,2010).Oneexampleofsoft lawwould be the Code of Good Practice for thePreparation, Adoption and Application of StandardsannexedtotheWTOAgreementonTechnicalBarriersto Trade. Following the Code is optional for WTOmembers and WTO dispute settlement is unavailableasa remedyunder theCode.AnotherexamplewouldbeAPEC'sBestPracticesforFreeTradeAgreementsandRegionalTradingAgreements.73Therationaleforusing a soft law approach would be to allow WTOmembers to better understand their respectivepriorities and interests, with a view eventually tounblocking progress towards legal interpretations ofparticularprovisionsthatwouldensurecoherence.

Thesoft lawapproach isnotwithout risk.Aspointedout by Shaffer and Pollack (2010), soft law and hardlaw could become antagonistic to one another if theunderlyingconditionsforcooperationareabsent.Low(2008)argues thatasharedperceptionofobjectivesand the nature of the transition to hard law wouldincrease thechances that soft lawcouldhelp rebuildhardlaw.Inviewoftheseconsiderations,heproposesa three-stageapproach.Thefirststagewould involveincreasedtransparencyandinformationsharingunderthe new Transparency Mechanism. This reinforcedexchange of views would pave the way for theprogressive development of soft law in the form of acodeofgoodpracticesinthesecondstage.Finally,ina third and last stage, when governments becomecomfortable with the soft law, negotiations aimed atimprovingthehardlawprovisionscouldbeundertaken.

The fourth and last proposal is to multilateralizeregionalism (Baldwin, 2006; Baldwin and Thornton,2008). Baldwin (2009) defines a process ofmultilateralization as the extension of existingpreferential arrangements in a non-discriminatorymanner to additional parties, or a fusion of distinctPTAs.Theideaisthat,asaresultofglobalproductionsharing,politicaleconomyforcesthatwerebehindtheproliferationofPTAsandthecreationoftheso-calledspaghetti bowl have weakened and are beingprogressivelyreplacedbynewforcesfavourabletothemultilateralizationofpreferences.Thistranslatesintoanumber of multilateralization initiatives both at theregionalandatthemultilaterallevel.

Examples of initiatives taken at the regional level toreduce the tangle of PTAs include APEC's BestPractices for PTAs or the Pan European CumulationSystem, which reduced the distortions of internationaleconomic production within the zone through theharmonizationofrulesoforiginanddiagonalcumulation.An interesting example of multilateralization at themultilateral level is the Information TechnologyAgreement, which established a mechanism for theelimination of MFN tariffs on information technologyproducts and thus made rules of origin and rules ofcumulationnon-operative.

Recent researchhashighlighted thepotentialcostofoverlapping PTAs and complicated rules of origin totoday'sworldofgeographicallyfragmentedproductionchains(Baldwinetal.,2009).TheremaybearolefortheWTOtoreducethesetransactioncostsbyservingas a forum for the coordination/standardization/harmonizationofpreferentialrulesoforigin.74Anotherway that greater coherence can be established hasalready been discussed and consists of identifying“bestpractices” inPTAs.75AsnotedinSectionD,theextent to which deep integration measures in PTAshavethepotentialtogeneratethesamesortofcostlyspaghetti/noodle bowl as tariff preferences is stillbeing debated. Baldwin et al. (2009) explore sixdifferent areas, discussing for each of them whetherPTAs have created a spaghetti bowl and how PTAprovisionshavebeenorcouldbemultilateralized.

A final thought with respect to moves towards themultilateralization of PTAs concerns decision-makingprocedures. Several authors (Lawrence, 2006;VanGrasstek and Sauvé, 2006; Cottier, 2009; Elsig,2009; Low, 2011) have considered the possibility ofdeveloping a multilateral approach to a modifiedconsensus rule, often referred to as critical massdecision-making. The approach proposed by Low(2011)isverysimilartotheso-called“code”approachthatemergedintheTokyoRoundagreementsonnon-tariff measures, but which was subsequentlyeliminated by the “Single Undertaking” (wherebynothing is agreed until everything is agreed) thataccompanied the creation of the WTO in 1995. Arevivalofthecriticalmassapproachoccurredwiththe


191

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

post-Uruguay Round agreements on basictelecommunicationsandfinancial services,aswellastheInformationTechnologyAgreement.

Theadoptionofacriticalmassapproachwouldmakeit possible to multilateralize trade rules withoutimplicatingtheentireWTOmembership–apropositionthatmaylookattractivewherethereisacaseformorebroadlysharedregulatoryapproachestotradebutnotnecessarily onaglobal basis.Acriticalmassmaybesaid to exist when a sufficiently large subset of theentire membership agrees to cooperate under theauspices of the WTO. An important characteristic ofthe approach is that agreements do not involve anydiscriminationvis-à-visnon-signatorycountries.

Appropriately chosen institutional and proceduralsafeguards could protect the system against the riskof fragmentationanddilutionof themultilateralbasisfor trade cooperation. Regarding the definition ofcritical mass, for example, a simple but effectiveapproachcouldbetoletthecriticalmassdefineitself.Criticalmasswouldbe reachedwhen thoseprepared

togoaheadwithanagreementconsider thatsupportandcommitmentfortheagreementinthemembershipis sufficient. Those left outside would then beconsideredtoosmalltounderminetheagreementandtherewouldnotbeanyreasonforrefusingtoapplytheMFN rule in respect of all the benefits to all non-signatories.

Another important question is whether and whenconsensusdecision-makingwouldneedtobeappliedtocriticalmassinitiatives.Intheabsenceofmultilateralparticipation through a consensus-based process, arisk exists that a sub-set of the membership couldshape rules from which they benefitted, but at theexpenseofmembersthatwerenotpartofthecriticalmass. The suggestion here is that critical massagreementswouldneedtobeapprovedbyconsensusbeforetheyenterintoforce.Notonlywouldtheriskofdamagingtheinterestsofnon-membersofthecriticalmassbeguardedagainst,butcriticalmassagreementswouldalso remainwithin theambitof themultilateralsystem.


192

1 “Systemiceffects”aredefinedforthepurposeofthisreportasthestaticanddynamiceffectsofPTAsonthemultilateraltradingsystem.Anexampleofstaticeffectisthepossibilityofconflictingrules,forinstanceontraderemedies.AnexampleofadynamiceffectistheimpactofaPTAontheprobabilityofengaginginfurthermultilateralnegotiations.

2 ThereissometheoreticalandempiricalworkstudyingtheinversequestionofwhethermultilateralismdrivestheproliferationofPTAs.Ethier(1998)andFreund(2000)buildtheoreticalmodelswherePTAformationisanendogenousresponsetothemultilateraltradingsystem.Usingdataonmultilateraltariffcutsandduty-freeaccessconcessionsgrantedbytheUnitedStatesatthetariff-linelevel,FugazzaandRobert-Nicoud(2010)findempiricalevidenceinsupportoftheclaimthatpastMFNopeningsowstheseedsoffuturepreferentialopening.

3 Therearepracticalproblemswiththisargument.First,assumingtheavailabilityofinternationallump-sumtransfersmaynotberealistic,andintheirabsence,itmayverywellbethat,atsomepoint,someblocmemberswillvetofurtherenlargements.Secondly,nothingforcesPTAmemberstosettheirexternaltariffsasassumedbyKempandWanandtheymayindeedhavereasonstosetthemdifferently(seeSectionC.1).

4 “Preferenceerosion”referstodeclinesinthepreferencemarginthatsomeexportersenjoyinforeignmarketsasaresultofpreferentialtradetreatment.Itcanoccurwhenexportpartnerseliminatepreferences,expandthenumberofpreferencebeneficiaries,orlowertheirMFNtariffwithoutloweringpreferentialtariffsproportionately(AlexandrakiandLankes,2004).

5 ExcludedcountriessufferfromthePTAbecausetheborderpricefacedbytheirexportersfalls.Fromtheperspectiveofmembercountries,thegainsofmovingtoglobalfreetradearebetteraccesstothird-countrymarketsandmoreliberalizationintheirimportmarkets.However,thesegainsaresmallforlowinitialtariffs,givingnoincentivetoPTAmembercountriestomovetomultilateraltariffreductions.

6 However,AmitiandRomalis(2007)arguethatformanydevelopingcountries,actualpreferentialaccessislessgenerousthanitappearsbecauseoflowproductcoverageorcomplexrulesoforigin.Therefore,loweringtariffsatthemultilaterallevel(DohaRound),especiallyonagriculturalgoods,islikelytoleadtoanetincreaseinmarketaccessformanydevelopingcountries.

7 Thisistheso-called“juggernaut”logic(BaldwinandRobert-Nicoud,2008).

8 NotethattheeffectcouldbereversedifthePTAresultedinahigherlevelofprotectionforthehomeimportcompetingsector.Inthiscase,asarguedbelow,thePTAwouldinhibitmultilateralism.

9 Enhancedprotectionisobtainedwhenproducersfromthelow-(external)tariffmembercanexportalltheiroutputtothehigh-tariffmemberwithoutaffectingpricesthere.Inthatcase,producersinthehigh-tariffcountryarenothurtwhileproducersfromthelow-tariffcountryenjoyhigherprotectionrents(FreundandOrnelas,2010).

10 AsdiscussedinSectionC,Ornelas(2005b),(2005a)qualifiestheargumentinmodelswheretheexternaltariffisendogenous.Thepossibilitythattrade-divertingPTAsareformedismorelimited,butcannotberuledout.

11 SchiffandWinters(1998)argue,however,thatPTAsbasedonsuchfactorsarelikelytobetransitory,sinceoptimumtradepreferencestendtodeclineovertime.Intheirmodel,thePTA’sexternaltradepolicybecomesincreasinglyopenovertime.

12 Noticethatthisresultisindependentoftheexistenceofpoliticaleconomymotivationsinexcludedcountries.If,however,thegovernmentsofnon-membercountriesputadisproportionatelyhighvalueontheprofitsofproducers,theyareevenmorelikelytoopposeglobaltradeopening.

13 SinceitisnotpossibletoobservethedegreeofmultilateralliberalizationtowhichacountrythatisamemberofaPTAwouldhavecommittedtoinitsabsence,theseempiricalstudieshavetorelyondifferencesinliberalizationpatternsovertime,acrosscountriesoracrosssectors,makingithardertoidentifythecausaleffectofPTAs.

14 Unilateraltariffreductionshaveaccountedfortwo-thirdsofthe21percentagepointcutsinaverageweightedtariffsofalldevelopingcountriesbetween1983and2003,accordingtotheWorldBank(2005).TariffreductionsassociatedwiththemultilateralcommitmentsintheUruguayRoundaccountedforabout25percent,andtheproliferationofregionalagreementsamountedtoabout10percentofthereduction.

15 BothstudiesfindthatUruguayRoundliberalizationwassmallerinproductswherepreferencesweregranted.

16 Thisinterpretationisstronglycriticizedbyanumberofscholars(Baldwin,2009).AccordingtoBaldwin(2009),itisCanadaandMexico’schangeofmindthattriggeredtheriseofregionalisminNorthAmerica.

17 ThisandthefollowingparagraphdrawonWorldTradeOrganization(WTO)(2007).

18 Asexplainedinmoredetailbelow,thePECSarrangementscameintobeingbecauseindustrialtradewasalmostduty-freeinEurope,buttradeflowswerebesetbycomplexandintertwiningoriginandcumulationrules.Tradeininformationtechnologyproductswasvirtuallydutyfree,buttheimpedimentstoefficiencyarisingfrommultiplepreferentialarrangementsbuiltpressureongovernmentstosimplifyarrangements–hencetheITA.

19 Thepointismoregeneralthanserviceliberalization.Itapplies,forinstance,topoliciesthatreduceoreliminatetechnicalbarrierstotrade(TBTs)acrosstheboard,bywayofregulatoryharmonizationormutualrecognition.EmpiricalevidencesuggeststhattheEU’ssinglemarketprogramme(alargepartofwhichisbasedonnon-discriminatoryregulation)increasedaccessatleastasmuchforthird-partyfirmsasforEUmembers(MayerandZignago,2005).

20 First-moveradvantagedefinescasesinwhichthesupplierthatfirstgetsintothemarketcanbenefitfromalong-lastingadvantage,evenifothersuppliersarenotsubsequentlyprohibitedfromentering.SeeMattooandFink(2004)andManger(2008).

21 GATSArticleV:6mandatestheestablishmentofliberalRoOsforPTAsinvolvingdevelopedcountries.TheArticleestablishesthat“A service supplier of any other Member that is a juridical person constituted under the laws of a party […] shall be entitled to treatment granted under such agreement, provided that it engages in substantive business operations in the territory of the parties to such agreement”.GATSArticleV:3(b)providesthatPTAsinvolvingonlydevelopingcountriesmay“limittradepreferencestoservicesuppliersownedorcontrolledbypersonsoftheparties”.YetmostPTAsamongdevelopingcountrieshavenottakenadvantageofthisoption.Amongthe

Endnotes


193

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

reasonswhycountrieshaveagreedtoincludeliberalRoOsintheGATSandnottousethespecialanddifferentialtreatmentprovisionspecifiedabove,FinkandJansen(2009)mention:i)thefactthatestablishednon-partyservicesuppliersareseenaspartofthedomesticeconomy;ii)inthepresenceofnetworkeconomies,itismoreefficientforservicesproviderstosimultaneouslyserveseveralmarkets,whichismadeeasierbyflexiblerulesoforigin;iii)participationinglobalproductionsharingcreatesanincentivetoabandonidiosyncraticservicestandardsasawayofboostingthecompetitivenessofownexportersandimprovingtheattractivenessofnationstoFDI.

22 Forinstance,theCloserEconomicPartnershipArrangements(CEPA)betweenChinaandHongKong,ChinaandMacao,China,respectively,followthewordingofGATSArticleV:6veryclosely.However,Emch(2006)arguesthatthenecessitytoaccumulativelycomplywithsixrequirements(natureandscopeofbusiness;yearsofoperations;paymentoftaxes;businesspremises;employmentofstaff;exclusionofintra-groupservices)toqualifyforthe“substantialbusinessoperations”requirementmayde factograntaccessonlytoafewservicesuppliers,onaselectivebasis.

23 ItshouldbenotedthatGATSArticleV:6onlyrecognizestheinterestsofjuridical,butnotofnaturalpersonsofthirdcountrieswhosupplyservicesundermode4intheterritoryofoneofthePTAmembers.Forinstance,aJapanesenationalwithadegreefromaFrenchuniversityandalicencetopracticeinFrancewhowantstoworkinGermanywouldnotbeentitledtothetreatmentgrantedtoEUnationals.

24 AccordingtoUNCTAD(2009),2,676BITswereinplaceattheendof2008.Eighty-twoBITsweresignedin2009,andsixduringthefirstfivemonthsof2010(UnitedNationsConferenceonTradeandDevelopment,2010).

25 Inthecontextofinvestment,MFNrequiresthatallinvestorsfromPTA-membercountriesareaccordedthebesttreatmentaccordedtoanyotherforeigninvestor.NTrequiresthatinvestorsfromPTA-membercountriesaretreatedaswellasdomesticinvestors.

26 NAFTA-basedagreementsaccordthebetterofMFNandNT.SeeKotschwar(2009)andthediscussionofinvestmentprovisionsinSectionD.

27 ThebilateralagreementsthatflourishedinEuropefromthemid-nineteenthcenturyuntilWorldWarIincludedsuchunconditionalnon-discriminationclauses.Theendresultwasde factomultilateralnon-discriminatoryliberalization(Lampe,2009).

28 Thereare,however,anumberofcaveatsthatlimittheroleofsuchMFNclausesasautomaticmultilateralizersofpreferentialtreatment.ThesecaveatsarediscussedinSectionE.2(e)below.

29 SeeBaldwinetal.(2009)fordetails.

30 ThetradeeffectsofPECSarediscussedinBoxC.4ofSectionC.Foradiscussionoftheeffectsofthe“multilateralization”ofrulesoforiginonthemultilateraltradingsystem,seeBoxE.2.

31 AradicalsolutionwouldbetheeliminationofMFNtariffsonindustrialgoods,whichwouldrenderrulesoforiginunnecessary.Thisisobviouslypoliticallyunpalatable.

32 Article23.2oftheDSU“prohibitscertainunilateralactionbyaWTOmember”.Morespecifically,underArticle23.2,aWTOmember“cannotunilaterally:(i)determinethataviolationhasoccurred,benefitshavebeennullifiedorimpaired,orthattheattainmentofanyobjectiveofthecoveredagreementshasbeenimpeded;(ii)determinethedurationofthereasonableperiodoftimeforimplementation;or(iii)decidetosuspend

concessionsanddeterminethelevelthereof”.(AppellateBodyReport, US / Canada – Continued Suspension,para.371).

33 SeetheGATTrulinginUnited States – Margins of preference,BISDII/11.

34 Foradetaileddiscussionofjurisdictionofinternationaladjudicativebodiesandofthesedoctrines,seeShany(2005).

35 Thiscanhappen,forexample,wherethecomplainantinoneforumisagovernment,whilethecomplainantintheotherforumisaprivateparty.

36 Foracontraryview,seeKuijper(2010).

37 ItshouldbeclarifiedthattheexistenceofconflictingdecisionswasnotthebasisforthereversaloftheWTOpanelbytheAppellateBody.

38 Noticethatthewelfareeffectsofthisincreaseddiscriminationare,however,unclear,becausethereispotentiallybothtradecreationwithinthePTAandtradediversionawayfromcheapersourcesofimportsfromnon-members.

39 Tehetal.(2009)andPrusaandTeh(2010)maptheanti-dumpingprovisionsofabout80PTAs,coveringalmost50percentofworldwideexports.Becauseanti-dumpinguseisgovernedbytheWTOAnti-dumpingAgreement,theyexpectthatifPTAruleshaveanyimpact,theywillservetomakeADdutiesmoredifficulttoimposeonPTAmembers.Thiscantakeanumberofforms.SomePTAsincreasethethresholdrequiredtoapplyanti-dumpingduties,orintheeventthatadutyisapplied,eitherreducesitbelowthedumpingmarginorshortenstheapplicableduration.OtherPTAsgivearoletoregionalbodiestoconductinvestigationsand/orreviewthefinaldeterminationsofnationalauthorities.

40 Toexplainthemethod,imagineobservinganti-dumpingactivityagainsttwogroupsofcountries(PTAmembersandnon-PTAmembers)fortwotimeperiods(pre-andpost-PTAestablishment).ThePTAcountriesare“treated”tosomeadditionalanti-dumpingrulesthatpossiblyaffectactivityinthepost-PTAperiodbutnotinthepre-PTAperiod.Thenon-PTAcountriesarenotexposedtothetreatmentduringeitherperiod.Thus,anyobserveddifferenceinanti-dumpingactivitybetweenthetwogroupsofcountriescanbecausallyattributedtothetreatment–theanti-dumpingrules.

41 ThediscussioninthissubsectioncloselyfollowsPrusaandTeh(2010).

42 PTAswhichexcludePTApartnersfromglobalactionsincludeAustralia-Thailand,Australia-US,Canada-Chile,Canada-Israel,EU-Chile,GroupofThree,Mexico-Chile,Mexico-Israel,Mexico-Nicaragua,Mexico-NorthernTriangle,Mexico-Uruguay,NAFTA,US-CAFTA-DR,US-JordanandUS-Singapore.

43 MostofthePTAsdescribeverypreciselywhat“substantialshare”oftotalimportsand“contributeimportantlytoseriousinjury”mean.InsomePTAs,“notsubstantialshareoftotalimports”meansifthepartnerisnotamongthetopfivesuppliersduringthemostrecentthree-yearperiod.Thephrase“notcontributeimportantlytoseriousinjuryorthreatthereof”meansthatthegrowthrateofimportsfromthePTApartnerisappreciablylowerthanthegrowthrateoftotalimportsfromallsources.

44 InArgentina–Footwear,ArgentinaincludedMERCOSURimportsintheanalysisoffactorscontributingtoinjurytoitsdomesticindustry.ButitexcludedMERCOSURcountriesfrom


194

theapplicationofthesafeguardmeasure.InUnited States–Wheat Gluten,theUnitedStatesexcludedCanadafromtheapplicationofitssafeguardactionalthoughimportsofwheatglutenfromCanadawereincludedintheinvestigationphase.IntheUnited States–Line Pipecase,theUnitedStatesexcludedimportsfromitsNAFTApartnersfromthesafeguardmeasurewhileincludingthemintheanalysisoffactorscontributingtoinjury.AndinUnited States–Steel,theUnitedStatesincludedallsourcesofimportsinitsanalysisofincreasingimports,seriousinjuryandthecausalnexus.However,itexcludeditsNAFTApartners,IsraelandJordanfromtheapplicationofitssafeguardaction.

45 WhilethewordparallelismisnotfoundinthetextoftheAgreementonSafeguards,theAppellateBodyconsideredthattherequirementofparallelismisfoundinthelanguageusedinthefirstandsecondparagraphsofArticle2oftheAgreementonSafeguards.SeeAppellateBodyReport,US –Steel,para.439.

46 SeePauwelyn(2004)foracritiqueoftheAppellateBody’suseofthisprinciple.

47 Onedispute(betweentheUnitedStatesandtheRepublicofKorea)inwhichthisissuewasgivensomeconsiderationwastheUnited States.–Line Pipecase.TheretheUnitedStatesarguedthatGATTArticleXXIVgaveittherighttoexcludeitsNAFTApartnersfromthescopeofthesafeguardmeasure.ThepanelacceptedtheUSargumentthattheexclusionofitsPTApartnersfromsafeguardactionsformspartoftherequiredeliminationof“restrictiveregulationsofcommerce”on“substantiallyallthetrade”amongthefreetradeareamembers,whichisaconditionrequiredbyGATTArticleXXIV.ThepaneldecisionwassubsequentlyappealedbytheRepublicofKorea.Onappeal,theAppellateBodydeclaredtherulingbythepanelonArticleXXIVasmootandhavingnolegaleffect.ThequestionwhetherArticleXXIVoftheGATT1994permitsimportsoriginatingfromaPTApartnertobeexemptedfromasafeguardmeasurebecomesrelevantonlyintwocircumstances.ThefirstwaswhentheimportsfromPTAmemberswerenotincludedinthesafeguardinvestigation.ThesecondwaswhenimportsfromPTAmemberswereincludedinthesafeguardinvestigationitneverthelesswasestablishedexplicitlythatimportsfromsourcesoutsidethefree-tradearea,alone,satisfiedtheconditionsfortheapplicationofasafeguardmeasure.SinceneitheroftheseappliedtothecircumstancessurroundingtheUnited States–Line Pipecase,theissuewasnotrelevanttothecase.TheAppellateBodywascarefultopointoutthoughthat,intakingthisdecision,itwasnotrulingonthequestionwhetherArticleXXIVoftheGATT1994permitsexemptingimportsoriginatinginamemberofafree-tradeareafromasafeguardmeasure.ThisdecisionthusleavesthequestionofanappealtoGATTArticleXXIVstillverymuchopen.

48 However,Baldwinetal.(2009)arguethatproductionunbundlingislikelytosoftenpoliticaloppositiontonon-discriminatorydeepprovisions.SeeSectionE.2(e).

49 TwominoramendmentsweremadetoArticleXXIVoftheGATTin1955-1957.Theterm“constituentterritories”wasreplacedwith“parties”,andtheterm“included”wasreplacedwith“providedfor”(Jackson,1969).

50 CertainmeasuresthatwerelinkedtotheformationoftheEuropeanEconomicCommunityoritsexpansionwerechallengedinGATTdisputesettlement.(See,forexample,USActionUnderArticleXXIII(ChickenWar)andEECCitrusPreferences(andAssociationAgreements)).Atthesametime,asHudec(1990)notes,theformationoftheEuropeanEconomicCommunitymeantthatdisputesbetweenEECmemberswerenolongerbroughttoWTOdisputesettlement.HefurtherobservedthatforsometimetheEEC

wasreluctanttoinitiatedisputesagainstothercontractingpartiesfearingthatitwouldinvitechallengestoEECmeasures.

51 Hudec(1971)suggeststhatArticleXXIVmaynothavebeen“draftedwiththedevelopingcountriesinmind”.HeexplainsthatwhiletheGATTrecognizestherighttoraisetradebarriersforthepurposesofindustrialdevelopment-thatis,topromoteinfantindustries-therequirementsofArticleXXIVmaylimitthispossibility,astheycallforeliminationofinternalbarriersandastatus quo anteceilingonexternalbarriers.

52 Forahistoryofthisprovision,seeSystemic Issues related to ‘Substantially all the Trade’ ,BackgroundNotebytheSecretariat(Revision),WT/REG/W/21/Rev.1,5February1998.Bycontrast,aprovisiononpreferentialtradeagreementswasnotincludedintheTRIPSAgreement.

53 On14December2010,theGeneralCounciladoptedaDecisiononaTransparencyMechanismforPreferentialTradeArrangements(WT/L/806),whichwasdraftedasaresultofthemandategivenbytheGeneralCounciltotheCommitteeonTradeandDevelopmentin2006.Thismechanismcovers:preferentialtradeagreementsfallingunderparagraph2oftheEnablingClause,withtheexceptionofregionaltradeagreementsunderparagraph2(c);preferentialtradeagreementstakingtheformofpreferentialtreatmentaccordedbyanymembertoproductsofleast-developedcountries;andanyothernon-reciprocalpreferentialtreatmentauthorizedundertheWTOAgreement.Paragraph2(c)oftheEnablingClauserefersto“Regionalorglobalarrangementsenteredintoamongstless-developedcontractingpartiesforthemutualreductionoreliminationoftariffsand,inaccordancewithcriteriaorconditionswhichmaybeprescribedbytheCONTRACTINGPARTIES,forthemutualreductionoreliminationofnon-tariffmeasures,onproductsimportedfromoneanother”.

54 AgreementsnotifiedunderGATTArticleXXIVandGATSArticleVareconsideredbytheCRTA.AgreementsnotifiedundertheEnablingClauseareconsideredintheCommitteeonTradeandDevelopment(CTD).

55 AttherequestoftheNegotiatingGrouponRules,theWTOSecretariathaspreparedacompendiumofissuesrelatedtoPTAsthathavebeengeneratedbyworkwithintheCRTAanddiscussionsinotherWTObodiesupto2002(seeCompendium of Issues related to Regional Trade Agreements ,BackgroundNotebytheSecretariat,TN/RL/W/8/Rev.1,1August2002).

56 Thesefigurescorrespondtonotificationsofnewregionaltradeagreements,aswellasaccessionstoexistingones.

57 Eighty-eightregionaltradeagreementswereconsideredintheCRTAandfourintheCommitteeonTradeandDevelopment.

58 Multilateralismisalsoconsideredsuperiortoregionalismbecauselargecountriescanbehaveinamorehegemonicwaywhentheynegotiatebilaterallywithsmallercountries.

59 SeealsoDavey(2011).

60 AsimilarpointismadebyBrownandStern(2011).

61 Thetraditionaltheoryoftradeagreementsfocusesitsattentiononterms-of-tradeeffects.Interms-of-tradetheory,themotivationforenteringintotradeagreementsdependsonwhetheracountrycaninfluencethepriceofitsimportsthroughitstradepolicy.Iftwolargecountriesenterintoatradeagreementtoescapeaprisoners’dilemma,thisagreementshouldbemultilateralratherthanpreferential.Thisisbecauseiftheydonotextendthebenefitoftheir


195

e. tH

e m

uLtILA

teR

AL

tR

AD

InG

sy

ste

m A

nD

PtA

s

bilateralagreementtoanythirdcountrythroughsomeformofMFNtreatment,oneortheotherofthetwolargecountriescouldindulgein“bilateralopportunism”bymakinganagreementwithathirdpartywhichexcludedtheotherlargecountrypartner(WorldTradeOrganization(WTO),2007).

62 SectionCpresentstheOatesdecentralizationtheorem,whichprovidestheeconomicrationaleforthesubsidiarityprinciple.

63 SeethediscussionofTBTcommitmentsinPTAsinBaldwinetal.(2009).

64 SeeDavey(2011),theoverviewofthedebateintheWTO’sWorld Trade Report(2007)andMarceauandReiman(2001).

65 Boththedefinitionofthe“otherregulationsofcommerce”andthequestionofhowtherequirementthatRTAsshouldnotresultinhigherbarriersagainstthirdpartieswereintenselydebated.

66 ProceduralissuesrelatingtotheadministrationofthePTAprovisionsoftheEnablingClausehavebeenaddressedthroughtheTransparencyMechanismforRegionalTradeAgreements.

67 NotethatinDecember2010theWTOGeneralCounciladoptedaTransparencyMechanismforPreferentialTradeAgreements(WTOdocumentWT/L/806),whichextendstheTransparencyMechanismforRTAstonon-reciprocalpreferences.

68 Evenett(2009)emphasizesthattheWTOGeneralCouncilDecisionestablishingtheprovisionalTransparencyMechanism(WT/L/671)mentions“consideration”ratherthan“examining”oran“evaluation”ofRTAs,which,inhisview,suggeststhatthecollectiveWTOmembershipdoesnotwantthisnewmechanismtohave“teeth”.

69 Notethatsomeissues,suchasforinstancethosepertainingtotheinternalcoherenceofWTOprovisionsthatapplytoPTAs,havebothaproceduralandasubstantiveorlegaldimension.

70 SeeDavey(2011).Whiletheredoesnotappeartohavebeenmuchconsiderationoftheseissuesinrecentyears,thereisnowanewproposalonthetableanddiscussionshaverestarted.Itremainstobeseenwhethertheywillbesubstantive.

71 SeethesummaryofdiscussionspreparedbytheWTOSecretariat(TN/RL/W/8/Rev.1).

72 Foramoredetailedeconomicdiscussionoftheproposals,seeWorld Trade Report 2007(WorldTradeOrganization(WTO),2007).

73 SeeMarceau(2007).

74 Onthemultilateralizationofrulesoforigin,seealsoBoxE.2.

75 A“bestpractice”hasalternativelybeendefinedasarulethatallowsconvergencetosomemultilateralbenchmark.SeePlummer(2006)forapossibleapproach.

Documents

e. The multilateral trading system and pTas · PDF fileA perennial policy question is how the multilateral trading system is affected by the ... system. This system puts Home in a