慕尼黑大學(xué)萊布尼茨經(jīng)濟(jì)研究所（IFO）：美國(guó)的經(jīng)濟(jì)制裁對(duì)目標(biāo)國(guó)家的全球貿(mào)易的影響：目標(biāo)國(guó)家的政治組織是否能起到作用

10910

2024

January2024

TheImpactofUSTrade

SanctionsontheGlobalTradeofTargetCountries:Dothe

PoliticalInstitutionsofthe

TargetsMatter?

SajjadFarajiDizaji,MohammadRezaFarzanegan

Impressum:

CESifoWorkingPapers

ISSN2364-1428(electronicversion)

Publisheranddistributor:MunichSocietyforthePromotionofEconomicResearch-CESifoGmbH

TheinternationalplatformofLudwigs-MaximiliansUniversity’sCenterforEconomicStudiesandtheifoInstitute

Poschingerstr.5,81679Munich,Germany

Telephone+49(0)892180-2740,Telefax+49(0)892180-17845,emailoffice@cesifo.deEditor:ClemensFuest

/en/wp

Anelectronicversionofthepapermaybedownloaded

·fromtheSSRNwebsite:www.SSRN.com

·fromtheRePEcwebsite:www.RePE

·fromtheCESifowebsite:/en/wp

CESifoWorkingPaperNo.10910

TheImpactofUSTradeSanctionsontheGlobal

TradeofTarget

Countries:

DothePolitical

InstitutionsoftheTargetsMatter?

Abstract

WeinvestigatetheeffectsofUS-imposedtradesanctionsontheglobaltradepatternsofsanctionedcountriesbyemployingagravitymodelthatincorporatesdataspanningfrom1980to2020across79nations.TheresultsrevealthatbothpartialandcompleteUSsanctionsleadtosignificantreductionsinbilateraltradebetweentheUSandtargetcountriesaswellasbetweentargetandthirdcountries.AunitincreaseintheintensityofcompletetradesanctionsinplacereducesUSbilateraltradeflowswithitssanctionedtradingpartnersbyabout76percentwhileaunitincreaseintheintensityofpartialUSsanctionsdecreasestradeby16percent.Whencompleteexportandimportsanctionsareimplemented,USbilateraltradeflowswithitssanctionedtradingpartnerswitnessastaggeringdeclineofabout90percentand39percent,respectively.Incontrast,theapplicationofpartialexportandimportsanctionsleadstoadecreaseintradeby13percentand17percent,respectively,allotherfactorsremainingconstant(ceterisparibus).Moreover,weshowthattargetcountrieswithstrongerpoliticalinstitutions,asmeasuredbydemocracyindicators,managetoalleviatesomeoftheadverseeffectsofUSsanctionsonbilateraltradewithboththeUSandthirdcountries.

JEL-Codes:D740,F140,F510,O240,O430.

Keywords:sanctions,trade,import,export,democracy,politicalinstitutions,gravitymodel.

SajjadFarajiDizaji

TarbiatModaresUniversity

Tehran/Iran

s_dizaji@modares.ac.ir

MohammadRezaFarzanegan

Philipps-UniversityMarburg

EconomicsoftheMiddleEastResearch

Group,CenterforNearandMiddleEastern

Studies(CNMS),SchoolofBusinessand

Econmics,Marburg/Germany

farzanegan@uni-marburg.de

2

1.Introduction

AmericanPresidentWoodrowWilsoncharacterizedeconomicsanctionsasa"peaceful,quiet,yetpotentsolution."Wilsonhadconfidenceinthedistinctiveeffectivenessofeconomicwarfare,aforcethatwouldapply"pressure…beyondthecapabilityofanycontemporarynationtowithstand"(Hufbaueretal.2016).Theideaisthateconomicsanctionsimposesuchsignificanteconomicpainonatargetcountry,thatitisforcedtocomplyorchangeitspolicies.Thereisasignificantandgrowingnumberofstudiesthatfocusontheeconomiccostsofsanctionsonthetargets(withanemphasisonUS/UNsanctions)suchas(Farzanegan,Khabbazan,andSadeghi2016;GhasseminejadandJahan-Parvar2021;Dizaji2021;Ghomi2022;DemirandTabrizy2022;LaudatiandPesaran2023)andsanctionsingeneralsuchas(Mulder2022;Hess2023;McDowell2023;FarrellandNewman2023;Demarais2022;Dizaji

andMurshed2024;Dizaji2022;Bergeijk1989).

Inourstudy,wecontributetothesanctionsliteraturebyempiricallyestimatingtheeffectsthatUS-imposedtradesanctionshaveonatargetcountry’strade1.Moreover,weexaminethemoderatingroleofpoliticalinstitutionsintargetcountrieswithrespecttothetradeeffectsoftheUSsanctions.Theimpactthatapoliticalsystemhasontradehasbeendiscussedinsomestudies.Inatheoreticalandempiricalstudy,Mansfield,Milner,andRosendorff(2002)concludethatthetypeofgovernmentcansignificantlyinfluenceastate’swillingnesstocollaborateoneconomicmatters.Leadersindemocraticnationsaremoremotivatedtoengage

ininternationaltradecooperationthantheirnon-democraticcounterparts.

Whathasbeenneglectedsofaristheempiricalanalysisoftheinteractingimpactofsanctionsandatargetcountry’spoliticalsystemontrade.Theautocracy–democracydimensioncouldbeimportantindeterminingthefinalimpactofsanctions(DizajiandvanBergeijk2013).Dizaji(2019)highlightsthatagovernment’sresponsetotradesanctionsalsodependsonthequality

ofpoliticalsystem.

Therehavebeencountrieswithnosignificantdemocracydeficitthathavealsobeentargetedbydifferenttypesofsanctions.Forexample,somecountriesandorganizationshaveimposedsanctionsorcalledforboycottsinresponsetoIsrael'spoliciesintheoccupiedPalestinian

territories.Thequestioniswhethersuchcountrieshavebeenmoresuccessfulinreducingthe

1InarecentstudyconductedbyGutmann,Neuenkirch,andNeumeier(2023),theauthorsemployedtheDifference-in-Differences(DiD)methodalongsideaneventstudytoinvestigatetheimpactofChinaandRussiaonmitigatingtheeffectsofsanctionsimposedbytheUSandEUonthetradeflowoftargetedentities.

3

negativeimpactofsanctionsontrade.Anotherquestioniswhetherthechangesinatarget

country’spoliticalcharacteristicscouldchangetheultimateeffectsofsanctionsontrade.

Asanexample,Iran’spoliticalflexibilityduringtheearlieryearsofPresidentRuhani’sgovernmentcouldreducethenegativeimpactsofsanctionsonIran’stradeflows.ThisislargelyattributedtohiswillingnesstocompromiseonIran’snuclearprogram,whichconsequentlyalleviatedtheeffectsofthesanctionsonIran(Dizaji2018).InJuly2015,followingaseriesofintensenegotiationsbetweenIranandthepermanentMembersoftheUnitedNationsSecurityCouncilandGermany(P5+1),thetwosidesreachedanagreement,knownastheJointComprehensivePlanofAction(JCPOA).ThisagreementsuspendedaportionoftheIslamicRepublic’snuclearprogram.Subsequently,theWesternpowersliftedmanyofthepreviouslyimposedsanctions(Dizaji2019).MostIranianbankswerereconnectedtotheSWIFT(SocietyforWorldwideInterbankFinancialTelecommunication)networkandwereabletoengageininternationalfinancialtransactions,whichwerevitaltoIran’strade.AccordingtotheWorldBank(2023a),theaverageannualgrowthrateofimportsofgoodsandservicesinIran,whichwasabout-15%betweenthesanctionyearsof2012-2015,increasedtoanaveragegrowthrateof6.3%twoyearsaftertheJCPOA(2016-2017).However,theUSadministrationannounceditsunilateralwithdrawalfromtheJCPOAinearly2018,leadingtosignificantinflationandpressureonthecountry’seconomyandbankingsystem.TheUSadministrationclaimedthatliftingthesanctionsassistedIraninusingitsnewfundstobuildnuclear-capablemissiles,supportterrorism,andcausehavocthroughouttheMiddleEastandbeyond(DizajiandFarzanegan2021;Farzanegan2022).TheIranianconservativefaction’sviewswerereinforcedbytheUSdecisiontounilaterallyleavetheJCPOA.There-impositionofsanctionsbytheTrumpadministrationresultedinamajordeclineinIran’saverageannualgrowthrateofimports,droppingtoannualaverageof-30%from2018to2020(forasurveyontheeffectsof

sanctionsontheIranianeconomyseeFarzaneganandBatmanghelidj2023).

Forouranalysis,weemploythegravitymodelapproach,including79countriesduringaperiodof1980to2020.Thegravitymodelofinternationaltradeassumesthattheamountoftradebetweencountriesispositivelyrelatedtotheirsize,asmeasuredbytheirnationalincomes,andnegativelyrelatedtothecostoftransportationbetweenthem,asmeasuredbythedistancebetweentheireconomiccenters(Dell’Ariccia1999;Wall1999).Thestandardformulationofthegravitymodelhasalsobeenextendedtoincludeothervariablessuchascommonlanguages,sharedborders,membershipinregionaltradingblocs,politicalconditions,economicsanctions,

etc.(HufbauerandOegg2003;Hufbaueretal.1997;Rose2004;Yu2010;Dizaji2018).

4

OurmainresultsshowthatbothpartialandcompletetradesanctionsreduceUSbilateraltradewithtargetcountries,controllingforasetofstandarddeterminantsoftradeinthegravitymodel.ThisdecreasingeffectofUSsanctionsisstatisticallysignificantandislargerinsizeinthecaseofcompletetradesanctions.Wealsoshowthatimprovedpoliticalconditionsinthetargetcountries,capturedbydemocracyindicators,mitigatepartofnegativeeffectsofUS

sanctionsonthebilateraltradebetweentheUSandtargetcountry.

Inthisstudy,wealsodifferentiatebetweenpartialandcompleteandexportandimportsanctionsimposedbytheUSandexaminetheireffectsonUStradewithtargetcountries.WefindthatcompleteexportsanctionshavethelargestnegativeimpactonUSbilateraltradewithtargetcountries.Finally,theimpositionofpartialimportsanctionsandcompleteimportandexportsanctionsbytheUSonspecifictargetcountriesleadstoareductioninglobaltrade,affectingbothexportsandimportswiththetargetcountries.Similarly,partialexportsanctionsfromtheUSnegativelyimpactthirdcountries'importsfromsanctionednations,yetthesesanctionsdonotsignificantlyaffecttotaltradebetweenthetargetcountriesandthirdcountries,nordotheynotablyimpactthirdcountries'exportstothesanctionednations.ThefindingsconsistentlydemonstratethatcompleteUStradesanctionsexertamorepronouncedadverseeffectontradeflowsbetweenthirdcountriesandtargetnationscomparedtotheimpactofpartialUSsanctions.However,higherqualitypoliticalinstitutionsintargetcountriescould

mitigatethenegativeimpactofsanctionsontradewiththirdcountries.

Therestofthepaperisstructuredasfollows:Section2presentsareviewofthetheoryonpossiblepositiveandnegativeeffectsofsanctionsontradeintargetcountries.Section3outlinestheresearchmethodologyanddataused.Thefindingsarepresentedanddiscussedin

Section4andSection5providesconcludingremarks.

2.UnderstandingtheImpactofSanctionsonaTargetCountry’sTrade:ATheoretical

FrameworkandLiteratureReview

TradesanctionsimposedbytheUScanhavebothpositiveandnegativeeffectsontradeintargetcountriesandthisinconclusiveimpactjustifiesanempiricalinvestigation.Paradoxically,tradewiththerestoftheworldinsomecasesmayincrease,followingtheimpositionofUStradesanctionsonatargetcountry.Thiscounterintuitiveeffectcanhappenbecauseofseveralfactorsandstrategiesemployedbythetargetcountryandotherinternationalactors.Suchfactorsmayincludediversificationoftradingpartnersfollowingthesanctions.Atargetmay

amplifyitseffortsinsecuringothertradingpartners,evenathighertransactionalcosts.For

5

example,Ioannouetal.(2023)explainhowsanctionshaveforcedRussiatochangeitsglobaltradestrategysinceitsinvasionofUkraine.TheyconcludethatRussiahasbecomeincreasinglyreliantontradepartnersthatarenotsubjecttosanctions,renderingitseconomymorevulnerable.Furthermore,RussiahasbeencompelledtoprovidediscountsonitscommodityexportsinordertoattractnewcustomerstoreplaceEuroAreapartners.Dizaji(2018)highlightsthe“policyofvisiontowardstheEast”anddiscusseshowIranhasdiversifieditsstrategicalliesinthepost-warperiodbyshiftingtowardsAsiancountries.ThemainreasonsbehindthispolicyshiftweretodiversifyIran’stradingpartnersandinternationaleconomicrelationsandtodecreasethenegativeimpactsofthesanctionsimposedbytheUSandotherWesterncountries.ArecentempiricalanalysisbyGutmann,Neuenkirch,andNeumeier(2023)alsoshowthatnon-Westerneconomicpowers,especiallyChina,undermineWesternsanctionsbyincreasing

theirtradewithtargetcountriesinspecificcategoriesofgoods.

AtargetmaychoosetobypassUS-imposedsanctionstomakesuchsanctionsineffective.Thiscaninvolveusingintermediaries,frontcompanies,orfinancialinstitutionsinothercountriestostimulatetrade.Inacompliancenote,theUSDepartmentsofJustice,TreasuryandCommerce(2023)explainedtheuseofthird-partyintermediariesandtransshipmentpointsbytheRussiangovernmenttobypasssanctions.Donovanetal.(2023)alsoprovidemoredetails

onbypassingsanctionsbyIran.

De-dollarizationandalternativecurrenciesmaybeanotherapproachforatargetcountrytomanageUS-imposedsanctions.AccordingtoMcDowell(2021)financialsanctionscreate'politicalrisk,'diminishingtheappealofthedollarforglobalusage.HeprovidesevidencefromthreecountriestargetedbyUSsanctions:Russia,Venezuela,andTurkey.Shagina(2022)explainsRussia’sde-dollarizationplan,whichstartedin2014andintensifiedduring2022.Sheconcludesthat“overuseofsanctionscouldstrengthenrevisionistcountries’desireto

increasinglyconducttheirtradeinnon-dollarcurrenciesinanattempttoavoidUSoversight.”

Targetsmayoptformoreresistancebyincreasingrelianceonitsowneconomyandreplacingtheaffectedexternaltradewithmoreinternalactivity.Cheratian,Goltabar,andFarzanegan

(2023)explainstrategiesthatIranianfirmsselecttomakesanctionslesseffective.

IncreasedsupportfromothercountriesthatdonotagreewiththeUSsanctionsmayalsopositivelyaffectthetarget’sglobalflowoftrade.IncreasingregionaltradeagreementsmaybeanotherchanneltomanageUSsanctions,intensifyingthetarget’stradewithneighboring

countries.Forexample,duringaninterviewinJune2023withthestatenewsagencyTASS,

6

RussianDeputyPrimeMinisterAlexeiOverchukstatedthattalksbetweentheEurasianEconomicUnion-whichincludesArmenia,Belarus,Kazakhstan,KyrgyzstanandRussia-and

Iranareintheirfinalstages(Reuters2023).

IncreasinginformalandunregulatedtradefollowingUSsanctionsmayalsooffsettheeffectofthesanctions.Farzanegan(2013)explainshowsanctionsmayresultintheexpansionofthe

informaleconomyinIran.2

Somefactorsandmechanismsmayalsoresultinthedeclineinatargetcountry’sglobaltradefollowingUStradesanctions.Financialisolationreducesatargetcountry’scapacitytoconductinternationaltradesincethisdependsonthetransferoffundsviaaglobalbankingsystem.AclearexampleishowSWIFTdisconnectsbanksforsanctionspurposes.In2012,inlinewithglobalandcollaborativeeffortstostrengthenfinancialsanctionsagainstIran,EURegulation267/2012wasenacted.ThisregulationbarsspecializedfinancialmessagingproviderslikeSWIFTfromofferingtheirservicestoIranianbanksunderEUsanctions.TheSWIFTnetwork,whichoperatesunderBelgianjurisdiction,isboundtoadheretothisruling,asaffirmedbyitshostcountry'sgovernment.SWIFTfulfilleditsregulatoryobligationbyseveringconnections

withtherelevantIranianbankssubjecttoEUsanctions.

NewmanandZhang(2023)highlightthebroaderconsequencesoffinancialsanctionsintermsoftheimpactedcountries.Theyshowthatthesanctionsreducemultinationalbanks’lendingtoneighboringcountriesofIrananditspoliticallyalignedstates.Banksaimtoprotectthemselvesfrompotentialenforcementbynotonlyrestrictingtheirexposuretodirecttargetsbutalso

establishingcompliancesystemstolimitexposuresacrosstheentiresupplychain.

Dizaji(2021),byemployingtheparametricapproachofstochasticfrontieranalysis(SFA)andusingdatafromdifferentIranianbanks,showsthattheintensityofsanctionsisassociatedwithincreasesincostsforIranianbanks.Hefindsadecreasingtrendinthecostefficiencyscoresof

Iranianbanksovertheperiodofsanctions.

In2022,aspartoftheglobalandcollaborativeefforttoescalatefinancialsanctionsagainstRussia,specializedfinancialmessagingproviders,includingSWIFT,werebarredfromprovidingservicestospecificentitiesasstipulatedinEUCouncilRegulation(EU)833/2014.

SWIFTterminateditsconnectivitywithallthedesignatedRussianentities(alongwiththeir

2Usingprovincelevelpaneldataregressions,FarzaneganandHayo(2019)showthatundereconomicsanctions,thegrowthrateofinformaleconomyinIranwasmoredampenedthangrowthrateofformaleconomy.

7

Russian-basedsubsidiaries).Furthermore,inlinewithEUCouncilRegulation(EU)765/2006,whichimposesasimilarprohibition,SWIFTalsoterminatedconnectionswithdesignatedBelarusianentitiesandtheirBelarus-basedsubsidiaries.3Theeconomicsizeofthetargetanditsinfluenceinglobalenergyandfoodmarketmayalsoaffecttheimplementationofthisfinancialisolation.Forexample,Greene(2022)arguesfortheunwillingnessofWashingtonandBrusselstotakeactionwhichwouldresultinsignificantrestrictionsonthepaymentchannelsofRussia.CompletelydisconnectingRussia’sbanksfromtheSWIFTnetworkmay

resultinhigherenergypriceswhichwouldthenbenefitRussia’sgovernment.

Losingkeytradingpartnerscanalsobeanotherfactorinlowerlevelsoftrade.Often,UStradesanctionsarelinkedtosecondarysanctionswhichpenalizethird-partycountriesandinstitutionsthattransactwithtargetcountries.Forinstance,whenthescaleoftransactionsbetweenaforeignfinancialinstitutionandaUS-sanctionedcountryreachesasubstantiallevel,theforeignentityfacestheapplicationofsecondarysanctions.Upondesignation,thesesecondarysanctionsmayeitherpreventUSindividualsfromengaginginbusinesswiththeforeignfinancialinstitutionorrequireUSbankstocurtailorimposerestrictionsonthe

correspondentaccountsheldbythatforeignfinancialentityintheUnitedStates.4

SupplychaindisruptionsfollowingUSsanctionsmayalsobeanotherreasonforthedeclineintrade.Internationalcompaniesmayvoluntarilyavoiddoingbusinesswithatargetcountry.Theymayhalttheinclusionofspecificcomponentsorpartssourcedfromthetargetcountryinthefinalproductstheymanufacture.ComplexityandtransactioncostsfollowingUSsanctionsmayincreaseanddiscourageinternationaltradewithatargetcountry.Sanctionsmayresultinthedevaluationofatargetcountry’slocalcurrency,raisingthepriceofimportedgoodswhichmayinturnreduceofficialtradeandincreasesmuggling(Zamanietal.2021;Farzanegan

2009).

Higherimportcostsmayalsoincreasetheexportpricesofsectorsthatdependmoreonimportedmaterialsandthus,bothexportsandimportsdecline.Additionally,sanctionscanresultinahigherriskofconflictandinstability,discouragingtradingpartnersfromlong-runcollaboration(DizajiandMurshed2024).AccordingtotheWorldBankGovernanceIndicators(WorldBank2023b),theworstperceptionofinstabilityandconflictinIranareobservedduringthemaximumpressurecampaignbytheUSagainstIranin2019-2020,whentheTrump

governmentre-imposedsanctions.Sanctions,byincreasingtheriskofcorruptionandillicit

3/de/node/11306

4/economic-sanctions-programs/

8

trade,mayalsoresultinahigherriskofinternalconflictinthetargetcountry(Farzaneganand

Zamani2022).

Obtainingtradefinancingsuchaslettersofcredit,insurance,orguaranteescanbemorechallengingfollowingsanctions,negativelyaffectingtheflowoftradeinatargetcountry.Decreasedforeigndirectinvestmentaftersanctionsmayweakenthetarget’sproductioncapacity,resultinginweakerexportcapacity.Finally,DizajiandvanBergeijk(2024)discussthattheimpactofsanctionsdecreasesovertimeaseconomicandpoliticalsystemsadjust.Therefore,sanctionsneedtobecomprehensiveandquick;slowandpartialsanctionsmay

diminishtheirimpactbybeingspreadout,reducingthelikelihoodofsuccess.

3.ResearchDesign

3.1.DataandModel

Inthisstudy,weuseagravityapproachtoexplaintheimpactofUSsanctionsonitsbilateraltrade(exportsplusimports)andexportsandimportsseparatelywithtargetcountries.Moreover,itinvestigateswhetherUSsanctionsaffectthetradeflowsbetweentargetcountriesandothernon-UScountries.Wealsoinvestigatehowahigherqualityofpoliticalinstitutions

inthetargetcountriesmaymitigatethenegativeimpactofsanctionsonbilateraltrade.

Weconstructatime-seriescross-sectiondataset.Thetime-seriescomponentispresentedbyyear,from1980to2020,andthecross-sectionincludes79countries.ControllingforGDPanddistance,themodelallowsustoisolatetheeffectsoftraderestrictionsaswellaspoliticalchanges.Besideseconomicsizeandgeographicaldistance,wealsoincludeothervariablesthatareexpectedtoinfluencebilateraltradeflows,suchastheincomelevels(asmeasuredbyGDPpercapita)andpoliticalcharacteristicsofthetradingpartners,commonlanguage,andlandlock

status.

Ourbasicempiricalmodeltakesthefollowinggeneralformat:

?=+1?+2+3+4?+

5?++6+7+(1)

Where:

?istheloggedbilateraltradebetweentheUSandcountryj.Therearethreemeasures

forthisvariable:exportsfromcountryjtotheUS,importsfromtheUS,andtotaltrade(exports

9

plusimports)betweentheUSandcountryj.ThebilateraltradedataaretakenfromDirectionofTradeStatistics(InternationalMonetaryFund(IMF)).Forestimationpurposes,thelog-linearexpressionforthedependentvariableisused.5ItisexpressedinUSdollarsanddeflated

byUSindustrialproducerprices.6

?istheloggedproductofUSGDPandGDPofcountryj.TheGDPdatainconstant

US$arefromCEPIIdatabase.

?istheloggedproductofUSGDPpercapitaandGDPpercapitaofcountryj.The

GDPpercapitadataaretakenfromtheCEPIIdatabaseandisbasedonPPP.

?istheloggedPopulation-weighteddistancebetweenmostpopulatedcities,measured

inkm,bilateralandcomesfromtheCEPIIdatabase.

isadummyvariablesetequalto1forlandlockedcountries.

isadummyvariablesetequalto1iftheEnglishlanguageistheofficialorprimary

languageinthesanctionedcountries.

,:tocapturetheeffectofUStradesanctions,weaddaseriesofdummyvariablestothemodel.Withrespecttotheextentoftheintervention,weseparateUStradesanctionsintotwocategories:partialtradesanctions()andcompletetradesanctions().TheinformationonthepartialandcompletesanctionsareobtainedfromtheGlobalSanctionsDataBase(Felbermayretal.,2020).Thesevariablestakethevalueof2ifthesanctionsarebilateral(simultaneousimportandexportsanctions),1ifsanctionsaredirectional(exportsonlyorimportsonlysanctions),andzeroiftherearenosanctions.Sanctionsareoftenexpectedtobemoreeffectiveiftakenatacomprehensivelevel.Specifically,becauseextensivecompletesanctionsplacecomprehensiveeconomicandfinancialpressuresonthetarget,theyshouldhavegreaterimpactthanpartialsanctions.?andαaretheerrortermandintercept,respectively.TableA1inAppendixAprovidesasummary

statisticsofusedvariablesintheregressionanalysis.

5-Theregressioncoefficientonacontinuouslogarithmicvariablecanbeinterpretedasanelasticity,thatis,astheratioofthepercentagechangeinthedependentvariableforeach1percentchangeintheindependentvariable.6-AsBaldwin(1994)discusses,oncepaneldataaretobeused,itisnecessarytouserealdata.Therefore,tradedataaretobedeflated.

10

3.2.Methodology

Somepreviousstudieshaveusedordinary-least-squares(OLS)regressiontoestimatethegravityequations(HufbauerandOegg2003;Hufbaueretal.1997;Yangetal.2004).AndersonandvanWincoop(2003)arguethattheregularOLSestimatesmaybebiasedonceresearchersignorecontrollingfor“multilateralresistance”amongtradingpartners,particularlywhenthedatasetisapanel.Standardcross-sectionestimatesofthegravitymodeldonotallowforheterogeneityintheregressionequations.Differenttradingcountry-pairscouldhaveunobservedspecificcountrycharacteristics(ChengandWall,2005).Toaddressthisissue,RoseandvanWincoop(2001)recommendusingcountry-pair-specificfixedeffectstocontrolforotherunobservablefeaturesbetweeneachpairoftradingcountries.Fixedeffectsmodelsallowforunobservedormis-specifiedfactorsthatsimultaneouslyexplaintradevolumebetweentwocountriesand,forexample,theprobabilitythatthecountrieswillbeinthesameregionalintegrationregime(Wall2002and2003;Bayoumietal.2007;Matyas1997).Thefixedeffectsapproachisconditionaluponthevaluesoftheintercepttermsandimpliesthatindividualsinthesampleareconsidered“oneofakind”(Verbeek,2017).Accordingto(ChengandWall2005),fixedeffectsmodelingisaresultofignorance,asitisdifficulttodeterminewhichvariables(cultural,historicalorpoliticalfactors)areresponsiblefortheheterogeneity

bias,simplyallowingeachtradingpairtohaveitsowndummyvariable.

Moreover,sincethereisalong-standingproblemwithdeterminingtheappropriatemeasureofeconomicdistancetocapturetransportationandinformationcosts(seeMayerandHead2002forareviewoftheissue),anaddedbenefitofthefixedeffectsmodelisthatiteliminatestheneedtoincludedistanceintheregression.Thisisalsotrueregardingothertime-invariant

variables.

Thedisadvantagewithstandardfixedeffectsestimationsisthattheyprecludeestimatingthecoefficientsfortime-invariantvariables,suchascountrygroupdummies.Theyalsohavetobedroppedautomaticallybecausethesevariablesaresubsumedintothecountry-paireffectsandarehiddenfromanalysis(WeiandFrankel,1997).7SerlengaandShin(2004)arguethatthefixedeffectsestimatorignoresthepotentialcorrelationbetweenthetime-invariantvariablesandunobservedcountry-pair-specificeffects,whichmayagainleadtobiasedcoefficient

estima