资源描述:
《The Relationship between FDI and Growth under Economic IntegrationIs There One【外文翻译】》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
本科毕业论文外文翻译外文题目:TheRelationshipbetweenFDIandGrowthunderEconomicIntegration:IsThereOne?出处:InternationalConferenceonEmergingEconomicIssuesinaGlobalizingWorld作者:AntonioMarascoLahoreUniversity,Pakistan原文:1.AbstractThisstudyisacontributiontothedebateontherelationshipbetweenFDIandgrowth.TheideathattheallegedlinkbetweenFDIandgrowthisrathertheconsequenceofbothFDIandgrowthrespondingendogenouslytoeconomicintegrationistestedempirically.Theresultsconfirmpreciselythispoint:itisnotFDIassuchbuteconomicintegration,inanyformorshapethatdeterminesgrowth.IntroductionTherelationshipbetweenFDIandgrowthisoneofthemostintensivelyresearchedissuesininternationaleconomics.Thereisafairamountofevidencesuggestinghatthereexistsapositiverelationshipbetweenthesetwoquantities,albeitwithsomequalifications(see,amongothers,Borenzsteinetal.1998).MorecontroversialhasbeentheissuewhetherunderpinningsuchapositiverelationshipthereiscausalityrunningfromFDItogrowthornot.OnerecenttwistonthisdebatehasbeenprovidedrecentlybyTingGao(2005).AccordingtoTingGao’spaper,theoftenobservedpositivecorrelationbetweenFDIandgrowthmightnotimplyanycausalrelationship,sincebothofthemmightrespondendogenouslytoeconomicintegration.Thesituationhesuggestsisliketheoneillustratedinflowchart1below:TheEconometricFrameworkThisstudyaimsattestingtheexistenceofacausalrelationshipthatrunsfromeconomicintegrationthroughFDItogrowth.Withthisobjectiveinmind,the6 followingeconometricspecificationisused:Theeconometricspecificationconsistsofastructuralmodelmadeupoftwoequations.ThefirsthastheratioofFDIflowtoGDP(FDI)asthedependentvariable,whichisregressedoneconomicintegration(Integr),onaninstrumentforFDIandonasetofthreecontrolvariables(controls).Thesecondequationhasthegrowthrateofoutput(g)asthedependentvariable,andthisisregressedonFDI,economicintegrationandthesamesetofcontrolvariables.Estimationisdoneviatwo-stageleastsquares(2SLS),themostcommonmethodusedforestimatingsimultaneous-equationmodels(seeGreene,2003).Thequalityofthisstudyhingesagreatdealonthechoiceofagoodinstrument.ThevariabletobeinstrumentedisFDI,henceinthiscaseaninstrumentisgoodifitishighlycorrelatedwithFDIandweaklycorrelated,ifatall,withgrowth.Thisisahardcall,particularlyingrowthregressions,wheremosteconomicvariableshavesomekindofrelationshipwithgrowth.Inthespecificcase,thevariablechosenasinstrumentisthelaggedvalueofFDI.AnotherimportantissuerelatestothecomputationofthevariableIntegr.TheexistingliteratureonthesubjecthasproducedmeasuresofintegrationwhicharebasedonFDI,tradeandprivatecapitalflows(asanexample,seeIsmihanetal.,1998).Inourcase,relianceonsuchanindexwouldcreateaseriousendogeneityissueinthefirstequation,sinceFDIwouldenterbothsidesoftheequation.Ideally,ourmeasureofintegration.shouldnotincludeFDIatallinitscalculation.Ontheotherhand,anaccomplishedmeasureofintegrationshouldtakefinancialintegrationintoaccount,animportantpartofwhichisofcourseFDI.Thisstudytriestostrikeadelicatebalancebetweenthesetwooppositeconsiderations.Tothisend,thevariableIntegrconsistsofanindexcomputedastheaverageoftwoitems.Thefirstitemisatradeintegrationindexwhichiscomputedasfollows:6 Theseconditemisafinancialintegrationindexwhichiscomputedinalikewisefashionasfollows:Thethirdmeasureofintegrationemployedisjustthetradeopennessvariableassuch,withnofurthermanipulation.Thatis:Onefurtheralternativemeasureofintegrationcouldalsopotentiallybeusedtocheckforrobustnessoftheresults.Suchameasurewouldbebasedonanevaluationofthebarrierstointegration.Inprinciple,thismeasureshouldaccountbothfortariffsaswellnontariffbarriers(NTB).BecauseofseverelackofdataonNTBinthetimedimension,ameasurethataccountbothfortariffsaswellasNTBisnotfeasible.Eveniftheindexweretobebasedontariffs’dataonly,lackofdatawouldstillbesevereenoughtoundermineanykindofcomparisonthatonewouldwanttomakewiththeothermeasuresofintegration.thereforeleavethisoptionasapossibleadditiontobeincludedinfutureresearch,oncedatacoverageontariffsandNTBimproves.2.DataandSampleSelectionIssuesThereisachoiceofsourcesforthedataregardingthemainvariablesofthisstudy.FDIdataweretakenfromtheUNCTADFDIonlinedatabase,GDPdatacamefromtheU.N.NationalAccountsdatabase.Dataontradeopenness(usedincalculatingIntegr)arefromthePennWorldTables,Version6.2.Dataregardingfinancialassetsandliabilities,usedtocalculatethefinancialintegrationindex,arefromtheExternalWealthofNations(EWN)database(seeKoseetal.,2006).Asforthecontrolvariables,dataonpopulationandinflationcamefromtheWorld6 DevelopmentIndicators2005(WorldBank)and,inafewinstances(mainlyfor2004)fromtheWorldDevelopmentIndicatorsonline.Finally,dataforaverageyearsofschooling(myproxyforhumancapital),camefromBarroandLeedatasetoneducationalattainment(2000).Withrespecttosampleselection,thiswasdictatedbyavailabilityofdataforthemainvariables.InitiallyIhadthoughttohaveapanelofbothdevelopedanddevelopingcountriescoveringaslargeageographicalareaaspossibleforthetimeinterval1980-2004.IncludedinthesamplearecountriesfromLatinAmerica,EastAsiaandPacific,SouthAsia,Africa,MiddleEast,EasternEurope,aswellastheOECDcountries.Itsoonbecameclear,though,thatinordertomaintainthecountriesofEasternEuropeinthesample,thetimeintervalhadtobeshortenedtotheperiod1990-2004.Afterrunningtheregressions,breathofgeographicalcoverageseemedtobequalitativelymoreimportantthanthelengthofthetimeintervalchosen,Ioptedforstickingtotheperiod1990-2004andkeepingthecountriesofEasternEuropeinthesample.Asaresultofthisstrategy,thesampleincludes51countries(thefulllistisgivenintheAppendix)covering15years.Intheyear2000,these51countriesaccountedforapproximately65%ofworldGDP3,andfor78%ofworldpopulation.Theregressionwiththefullsample,bothintermsofcountriesincludedandyearscovered,features680observations,insteadofthepotential765(51*15=765),because51valuesarelostwhenlaggingFDIforthefirstyear(1990),andinflationdatainclude34negativerates,whichresultinto34lostvalueswhentakinglogs(51*15=765-51=714-34=680).DetaileddescriptivestatisticsareshowninTables4,5and6.ResultsTheresultsofthe2SLSregressionsaredisplayedinTable1(firststage)andTable2(secondstage).Asdiscussedearlier,resultsaregivenforthreedifferenttypesofintegrationmeasures,andalongthreedifferentlevelsofaggregation(allcountries,developingcountriesanddevelopedcountries).Regressionsareidentifiedbytwodigits,thefirstreferringtotheintegrationmeasureused,andthesecondreferringtothelevelofaggregation.Forexampleregression2.1referstoIntegr2andtoallcountries,andsoon.Table1clearlyshowsthateconomicintegrationisasignificantandpositivelysigneddeterminantofFDI.Suchresultholdsnomatterhow6 onedefinesintegrationorwhichlevelofaggregationischosen.InthecaseofTable2,twopointsemergeinalmostasequallyclear-cutamannerasthemessageconveyedbyTable1.Firstly,integrationisapositivedeterminantofgrowthinallcasesbutregressions1.2and1.3.ThispointisinfullaccordancewithGao(2005).Secondly,anevenmoreimportantpoint,FDIisneverasignificantcontributortogrowth.This(non)resultisveryrobusttoalltypesofintegrationmeasuresandalllevelsofaggregation.ItisalsoperfectlyinlinewiththeargumentthattheallegedrelationshipbetweenFDIandgrowthmightjustbeaclassicalexampleofomittedvariablebias,wheretheomittedvariableinthespecificcasewouldbeeconomicintegration.Tomaketheevidencemorecompelling,Irunafixed-effectsregressionofFDIongrowthwithouteconomicintegration5,whoseresultsarepresentedinTable3.Asbefore,theexerciseisrepeatedforallcountriesinthesample,thedevelopingcountriesandthedevelopedcountriesrespectively.TheevidencethatIgetismixed,sinceFDIissignificantatthe5%levelifIrestrictattentiontodevelopedcountries,notsignificantwhenattentionisrestrictedtodevelopingcountriesandsignificantatthe10%leveliftheentiresampleisincluded.ThisispreciselythekindofmixedevidencethatwouldemergefrompastliteratureonFDIandgrowth.Suchuncertaintyiswipedoutthoughonceeconomicintegrationentertheframe,aswehaveseen.Then,thereissimplynoroleforFDI,singularlyconsidered,asadeterminantofgrowth.3.ConclusionThisstudyhasbeenyetonemoreattemptatsheddinglightontherelationshipbetweenFDIandgrowth.Thenewtwisthere,aftertakinginspirationfromrecenttheoreticalworkbyGao(2005),consistedinaddingthevariable“economicintegration”totheanalysis.Exactlyasexpected,andasclaimedbyGao,theallegedpositivelinkbetweenFDIandgrowthdisappearsonceintegrationisadded.ThisstudysuggeststhatthecurrentfrenzyofcountriesfromallincomebracketstoattractFDIasawaytoimprovetheirgrowthprospects,mightbemisplaced.Whatcountriesthatwanttogrowfastershoulddoistobecomeevermoreintegratedwiththeworldeconomy.Theactualmodeofintegration,whetherthroughtrade,FDIorelse,seemsnottomatter.Thisstudycanbeimproveduponandextendedinseveralways.Firstly,thedatasetofreferenceshouldbeextendedasnewdatabecomeavailable,6 particularlywithrespecttothecountriesofEasternEuropeandthecountriesbelongingtothelowerincomebrackets.Also,theconceptofeconomicintegrationshouldbeaugmentedtoincludelabormarketintegration.Laborofcourse,isaveryimportantdimensionoftheeconomy,andIhaveleftitoutbothforproblemsofdataavailabilityandalackofaneffectiveproxytomeasurelaborintegration.Infutureworkhowever,thelattershoulddefinitelybeincludedifoneistomakeamoreconvincingclaimthat,undereconomicintegration,thereisnolinkbetweenFDIassuchandeconomicgrowth.6