Aninterpretativemodelofforeigndirectinvestment
inChina:Aneconomicpolicyapproach
´MiguelGINER*,GracielaGINER1Jose
DepartmentofAppliedEconomicsandEconomicPolicy,UniversityofAlicante,
P.O.Box99,E-03080Alicante,Spain
Accepted16July2003
Abstract
Theobjectiveofthispaperistoelaborateaninterpretativemodelofforeigndirectinvestment(FDI)inChinabasedonanintegratedviewofeconomicpolicy.Alongwiththeanalysisofpurelyeconomicfactors,othervariableswillbeincorporatedsuchasthoserelatingtosociopoliticalinstability,thebusinessoperatingcontext,andtheexistenceofanexport-orientedpolicy.Themethodologythatwehavechosenrecognisestheimportanceofeconomicelementsaswellasthatofsociopoliticalelements,especiallywithregardtocountriesintheprocessofdevelopment,transition,andreform,suchasChina.
D2003ElsevierInc.Allrightsreserved.
JELclassification:C5;F2;P2
Keywords:Foreigndirectinvestment;Model;Macroeconomicfactors;Sociopoliticalfactors;China
1.Introduction
Thereexistsabroadconsensusthatforeigndirectinvestment(FDI)occurswhenthreedeterminingfactorsexist:thepresenceofspecificcompetitiveadvantages,whicharethepropertyofamultinationalcompany;theexistenceofspecificadvantagesoflocatinginthecountryreceivingtheinvestment;andthepresenceofgreatercommercialadvantagesattheintracompanylevel.
WhilethefirstandthirdconditionsarespecificbusinessfactorsaffectingFDI,thesecondisspecifictolocationandexertsacrucialinfluenceontheflowofdirect
*Correspondingauthor.Tel.:+34-965903958;fax:+34-965903603.
E-mailaddresses:giner@ua.es(J.M.Giner),graciela.giner@ua.es(G.Giner).1Tel.:+34-965903400x3171;fax:+34-965903603.
1043-951X/$-seefrontmatterD2003ElsevierInc.Allrightsreserved.doi:10.1016/j.chieco.2003.07.002
J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280269
investmenttowardsthereceivingcountry.Ifonlythefirstconditionexists,companiestendtofocusonexportation,licensing,orthesaleofpatentstoattendtotheforeignmarket.Ifthethirdconditionalsoexists,FDIbeginstobethepreferredmodel,butonlywhenspecificadvantagesoflocationarealsopresent.Whenallthreeconditionsexist,thefactorsoflocationaretheonlyonesthatcanbedirectlyinfluencedbygovernmentsofthosecountriesreceivingdirectinvestmentfromabroad.
InordertoexplainthedifferencesintheflowsofinvestmentbetweendifferentcountriesandtoformulatepoliciesdesignedtoattractFDI,wemustnotonlyunderstandhowandwhymultinationalcompanieselectcertaincountriesorzonesforsuchinvestmentbutalsothemannerinwhichthesemultinationalsact.
Theanalyticapproachthatwehavechosentouseinthispaperissimilartothatemployedinmacroeconomicstudies.Althoughmanysuchstudieshavediscoveredtheimportanceofeconomicdeterminants,thereexistsnoconsensusaboutthose.Alongwithpurelyeconomicelements,manyauthors,amongthemSinghandJun(1995),haveemphasisedtheimportanceofpoliticalandinstitutionalquestionsasdeterminingfactorsinFDI.AlthougheachtheoreticalandanalyticalapproachprovidesapartialvisionofthecomplexityoftheflowsofFDI,theredoesnotyetexistanintegratedtheorythatcombinestheseelementsinananalyticalmanner.WehaveattemptedtoovercomethissignificantlimitationbyproposinganintegratedanalyticalapproachthatexplainstheFDIdirectedtowardsChinaandbypointingoutnotonlyeconomicfactorsbutalsotheimportantroleofvariablesassociatedwithpoliticalstabilityandwiththeenvironmentwithinwhichcompaniesmustfunction.Theempiricalresearchthatwehaveutilisediscentred,therefore,onproducingmodelsofmacroeconomicandsociopoliticalfactorsthataffecttheflowofFDItoChina.InSection2,wereviewtherecentliteraturedealingwiththeoreticalandempiricalaspectsofFDI.IntheSection3,weproposeaninterpretativemodelderivedfromanintegratedfocusoneconomicpolicyoftheFDIinChina.Togetherwiththeanalysisofpurelyeconomicfactors,weincorporatevariablesassociatedwithsociopoliticalinstability,withtheconditionsinwhichcompaniesmustdobusinessandwithanexport-orientedstance.Finally,wepresentourconclusions.
2.Theoreticalandempiricalperspectives:areviewofrecentliterature2.1.Empiricalapproaches
TheliteraturedescribingthefactorsthatdeterminetheflowofFDI,usingdifferentandvariedmethodologies,isvoluminousandcontroversial.Empiricalstudiesthatevaluatethesedeterminingfactorsaregenerallybasedononeofthreeapproaches.Thesearemicroeconometricstudies,studiesusingdataanalysis,andaggregateeconometricstudies.2.1.1.Microeconometricstudies
Thesestudieshavebeencarriedoutatthefirmandindustriallevel,particularlyinthoseindustriesconsidered‘‘footloose,’’2suchasthosewefindintheelectronicssector.When
2Industriesthatfacenorestrictionsonwheretheychosetolocate.
270J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280
thedecisiontoinvestisacorporatedecision,suchdesegregatedstudiesarevaluablebecausetheycapturethecomplexityofthedeterminingfactorsintheflowofFDIwithinaspecificindustry.However,althoughthemajorityofstudiesuseadependentvariablespecifictothesector(normallytheStandardIndustrialCodeorSIC—anumbercomprisedoftwoorthreedigits),theexplanatoryvariableshavedimensionsthatarespecifictotheareaortothecountry.
Forexample,WoodwardandRolfe(1993)haveanalysedthefactorspertainingtothelocationofFDIorientedtowardsfomentingexportsintheCaribbean.ThedependentvariableusedinthisstudywasderivedfromdataprovidedbytheU.S.DepartmentofCommercewithregardto187instancesofinvestmentinthecreationofnewfactories.Alsousedasexplanatoryvariableswerevariousaggregatevariablesspecifictotheregion,suchaspercapitaGNP,politicalstability,theavailabilityoffree-tradezones,concentra-tionofindustry,andtheavailabilityofland.Asaconsequence,thistypeofstudydoesnothaveatotallymicroorientationandthereforemayfailtoadequatelyexplaininvestmentdecisionsbasedonregionalconsiderations.Studiesdealingwithonespecificcompanyaremoreusefulforincorporatingmicroaspects,buttheyoftenfailtoprovidegeneralisationsthatmaybeappliedtoadevelopingeconomy.
2.1.2.Dataanalysisstudies
Thesestudiesanalysekeyfactorsthatmotivateinvestors;ideally,theyaredesignedtostudythequalitativefactorsthatotherwisemaybedifficulttoincorporateinaneconometricmodel.AgoodexampleofthisapproachisthatofRolfe,Ricks,Pointer,andMcCarthy(1993).
Inthisstudy,aquestionnairewassenttothedirectorsofNorthAmericancompaniesthathadinternationalinvestments.Theywereaskedtoprovidearankedlistof20specificincentivestoinvestment.Theresultsdemonstratedthataveryhighvaluewasplacedonthelackofrestrictionsonpaymentsmadebetweencompaniesandonthepaymentofdividends.Onegeneraldisadvantageassociatedwiththisapproachisthatstudiesbasedonsurveysareinherentlylimited,particularlywithregardtothesensitivityofrepliestothestructureandstyleofsurveyquestions(Wallsten,1980).
2.1.3.Aggregateeconometricstudies
Thatisourempiricalapproachofreferenceinourresearch.Aggregatedstatisticalstudiesattempttoevaluatethespecificdeterminantsexistinginthecountryorregionwithreferencetoglobaltendenciesoftheflowofinvestmentfunds(FDI).Byanalysingdataonamacroeconomiclevel,thesestudiescanprovideavisionofthestructuralcharacteristicsandmacroeconomicpoliciesthatfavourtheflowofFDI.Butbecauseitisdifficulttoincorporatespecificfactorsatthecorporateorindustriallevelinthesemacrostudies,therelativeaggregatevaluesonlycapturegeneralandlong-termtendencies.
WecanfindareviewofFDIdeterminants,usingseveraldifferentmethodologies,inPearce,Islam,andSauvant(1992).
Becauseeachapproachhasitsownadvantagesandlimitations,pluralisminmethod-ologyishighlydesirable.Onewaytoprovethesolidityofourfindingsistofilterthesameproblemthroughdifferentmethodologies.Agarwal(1980),Hein(1992),andSchneider
J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280271
andFrey(1985)havecarriedoutsuchanevaluationofempiricalstudiesbasedonanaggregatedeconometricapproach.
2.2.MacroeconomicandsociopoliticaldeterminantsoftheFDI:empiricalbackgroundTheempiricalbackgroundpresentedinthissectioniscentredonthosemacroeconomicandsociopoliticaldeterminantsthatinfluencetheflowofFDItowardsChina.Althoughtherearenosimplemodelsbasedonsolidtheoreticalprinciplesthatmightguidetheempiricalanalysisofthesequestions,wehavebasedourworkontheresultsofpaststudies,principallyonthoseofSinghandJun(1995),whoexaminethedeterminantsofFDIin30developingcountries.Otherrelevantandrecentstudieswithaspecificfocushavebeenexamined(Liu,Wang,&Wei,2001;Song,Liu,&Jiang,2001;Wu,2001;Zhang&Felmingham,2001;Zhao,2001).
Althoughmanyaggregateeconometricstudieshavebeencarriedout,thereisnever-thelessnoconsensusontheprincipaldeterminantsofFDI.ThislackofconsensuscanbeattributedatleastinparttothelackofpreciseandtrustworthydataontheflowofFDIandonitspotentialdeterminingfactors,particularlyatthesectoriallevel,aswellastothefactthatthemajorityofempiricalstudiescarriedouthaveanalysedthedeterminantsofFDIbygroupingtogetheraseriesofcountriesthatmaybestructurallydifferent.
Inouranalysis,besidesthetraditionaleconomicvariablessuchasGDPpercapita,GDPgrowth,andlabourcosts,wewillincludevariablesofasociopoliticalnatureasfactorsthatmayaffecttheflowofFDItoChina.ThisanalysisisintegratedinaresearchlinedevelopedbyauthorsinthefieldoftheChinaeconomyandFDI(Giner,2000;GinerandGiner,1999,2000,2001,2002a,2002b).
2.2.1.Socialandpoliticalinstability
Althoughpoliticalriskmayhaveanimpactonforeigninvestmentdecisions,empiricalresultsshowthatthisisnotalwaysthecase.Aharoni(1966)considersthatapartfromthepotentialofthemarket,politicalinstabilityisthemostsignificantvariable.
Ontheotherhand,BennetandGreen(1972)believethatdirectinvestmentfromtheUnitedStatesisnotaffectedbythepoliticalinstabilitythatmayexistinthosecountriesreceivingtheinvestment.Levis(1979),usingtwoproxyvariablesforpoliticalinstability,obtainedmixedresults.HefoundthatpoliticalinstabilitywasindeedasignificantdeterminantofFDIintoday’sworld,althoughitwasnotinthepast.Anothervariable,thelegitimacyoftheregimeinpower,wasfoundtobesignificantforpastperiodsbutnotfortoday’sinvestmentdecisions.
Analysesof58developingcountries,carriedoutbyRootandAhmed(1979),haveshownthatthenumberofchangesinleadershipofthecountrybetween1956and1967wasasignificantfactor.Nevertheless,otherpoliticalvariablessuchasthenumberofinternalarmedattacks,thedegreeofnationalism,andthecolonialaffiliationofthecountrywerenotsignificant.
SchneiderandFrey(1985)foundanegativerelationshipbetweenFDIandthenumberofpoliticalstrikesandscandalsincountriesreceivinginvestment.Nigh(1985),usingCOBDAB,adatabasethatprovidesanaggregatemeasureofconflictsandcooperationwithinagivencountryaswellasbetweencountries,hasdemonstratedthatfordeveloped
272J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280
countries,itisthepoliticaleventsthatoccurbetweencountriesthatdetermineFDIandnottheeventsthatoccurwithinagivencountry.Ontheotherhand,fordevelopingcountries,internalpoliticaleventshaveastronginfluenceonFDI.
Morerecently,WheelerandMody(1992)havefoundthatthecomponentmeasuringadministrativeefficiencyandpoliticalriskcouldbestatisticallyinsignificant.Lucas(1993)doesnotdirectlyincorporateproxyvariablesforsociopoliticalrisk.Hedoes,however,discoverthat‘‘goodevents’’suchastheAsiaticandOlympicgamesinKorea,theascensiontopowerofAquinointhePhilippines,etc.,areindeedrelatedinapositivemannertoFDI.Ontheotherhand,‘‘badevents’’suchastheregimeofSuhartoinIndonesia,theassassinationofParkinKorea,andtheimpositionofmartiallawbyMarcosinthePhilippinesallhavenegativeeffectsonFDI.
Empiricalevidenceontheimpactofpoliticalriskisnotunequivocalsinceitisdifficulttoobtainreliablequantitativeestimatesofthisqualitativephenomenonoveralongperiodoftime.Thisisparticularlytruewithregardtosomeaspectsofpoliticalriskthatmaybeseenbyforeigninvestorsasimposingdirectrestrictionsininvestment.Politicalinstabilityisaverycomplexphenomenon.Themajorityofproxyvariablesavailablecanonlycapturecertainaspectsofthisdeterminingfactor.
2.2.2.Businessoperatingcontext
AsHelleiner(1989)haspointedout,government-createdincentivesforinvestment,suchastaxholidays,appeartoplayalimitedroleininvestmentdecisionsbetweencountries.ThemajorityofempiricalliteraturesupportsthenotionthatspecificincentivesdonothaveamajorimpactonFDI,particularlywhentheseincentivesaredesignedtocompensateforothercomparativedisadvantages.Ontheotherhand,itisgenerallybelievedthatbyeliminatingrestrictionsandprovidingabetterenvironmentforcompanies,theflowofFDIwillbepositivelyaffected.
Inthiscontext,thereexistsawidegamutofgovernmentalpoliciesthatcanincreaseorreduceFDIflows.Itisdifficulttoquantifythesepoliciesasexplanatoryvariables;analternativemethodistouseaqualitativeindexrepresentingthejudgementofexpertsinthefield.
Oneveryrelevantpointinthiscontextistheroleplayedbytariffbarriers.ExaminationofthehypothesisassociatedwiththeideathatprotectionisttariffbarriersthatmaystimulateFDIsubstitutesforimportationhasshowninconclusiveresults.Nevertheless,themajorityofavailableempiricalevidencesupportsthishypothesis,particularlytheanalysisofU.S.FDIintheEuropeanUnion.
2.2.3.Orientationtoexport
Alongwiththesizeofthedomesticmarketinthecountryreceivingforeigninvestment,aproclivitytoexportisanotherimportantfactorinfavouringandsuccessfullyattractinginvestment.Recently,Dollar(1992)andHein(1992)havediscoveredthatdevelopingeconomiesthatareorientedtowardsexteriormarkets(thatexpecttofindnewexportmarkets)havebeenrelativelyprosperous.Lucas’(1993)investigationofthecountriesofSoutheastAsiagivessomeevidenceoftherelativeimportanceofpoliciesorientedtowardstheexterior.Specifically,FDIisrelativelymoreelasticwithrespecttoexportdemandthanwithrespecttoaggregatedomesticdemand.Furthermore,ifthoseeconomiesthatare
J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280273
orientedtowardstheexteriorareindeedmoresuccessfulinattractinggreaterquantitiesofFDI,thesizeofthedomesticmarketceasestobeanimpediment.Evensmallcountriesreceivinginvestmentcaninfluencethedecisionsofforeigncompaniesbyfomentingpoliciesorientedtowardsexports.
However,theexistingempiricalliteraturedoesnotestablishwhethertheflowsofFDIareattractedbyeconomiesthatarealreadyorientedtowardsexportation(exportsprecedetheflowsofFDI)orwhethermultinationalinvestmentincreasesexports(FDIprecedesexports).Obviously,fromapoliticalpointofview,itisimportanttoknowwhicheventcausestheother.
3.AninterpretativemodelofFDIinChina3.1.Hypothesis,data,andmodelspecifications
WehaveconstructedamodelbasedpartiallyontheanalysiscarriedoutbySinghandJun(1995).Whiletheirstudyisbasedonamodelthatincorporatesacrosssectionofdevelopingcountries,ouranalysiswasconductedfortheperiod1980–1997(18observations)andforonecountry:thePeople’sRepublicofChina.Thedataperiodisannualandthedataarepresentedasapuretimeseries.Thesourcesofdataareasfollows:-ChinaStatisticalYearbook,StateStatisticalBureau(1996,1997,1998).-InternationalFinancialStatistics,InternationalMonetaryFund.-BusinessEnvironmentRiskIntelligence(BERI).
ThedependentvariableisFDIcalculatedinlocalcurrency(Reminbi)atconstantprices.FDIreferstonetinvestmentmadebyforeigncompanies,excludingrepatriatedprofits.UsingLeamer’s(1985)approach,arangeofspecificationsisestimatedtoassessthesensitivityofthecoefficientsofthehypothesisedvariables.Thehypothesestestingcanberepresentedbyasimpleequation:
FDI¼FðHV;CVÞ
Basedonthediscussionofthelastsectionandonthecontextofourmodel,wetestthefollowingempiricalhypothesisandnoteourresultsinSection3.2:Hypothesis1.PoliticalinstabilitynegativelyaffectstheflowofFDI.
Hypothesis2.ThegeneralperceptionthatafavourablebusinessenvironmentexistsinthecountryreceivingtheinvestmentpositivelyaffectstheflowofFDI.Hypothesis3.TheexportsarerelatedtotheflowofFDI.
Besidesspecifyingtheproxyvariableassociatedwiththehypothesis(HV)—seethenextsections—itisimportanttospecifythevectorofcontrolvariables(CV)inordertocorrectlystatetheequation.Thecontrolvariablesareprogressivelyintroducedintothe
274J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280
modelinordertodeterminetherobustnessofourresultsvis-a-visalternativespecifica-tions.Thestabilityofthecoefficientsofthehypothesisedvariablecanbeassessedasthecontrolvariables.Notethattheprimaryfocusisonthecoefficientsofthehypothesisedvariable,thatis,toobservewhetherthatarerobustasdifferentcontrolsareimplemented.Thecoefficientsofthecontrolvariablesarenotsubjectedtoanysystematicanalysisandthusshouldbeinterpretedwithcaution.
Thecontrolvariablesofouranalysisareasfollows:Sizeofthemarket:GDPpercapita(GDPPC)andtherateofgrowthoftheGDP(RGGDP).
Labourcosts:anindexofrealgainsasacontrolvariableforrealwages(EARN).Exchangerate:therealexchangerateasacontrolvariable(XRATE).
Factorsinthecountryoforigin:theindexofaverageindustrialproductionofthecountriesthatcomposetheG-7(Canada,Japan,UK,Italy,France,theUnitedStates,andGermany)(IPG7),calculatedwithdataprovidedbytheIMF’sInternationalFinancialStatisticsdepartment.
Theimportanceofexports:exportsrelatedtoGDPareincludedasacontrolvariable(EXPTPTR).
Laggeddependentvariable:theregressiveindependentvariableisusedasacontrolvariable(FDI1).
Timeeffect:adummyvariablethatcapturestheeffectsrelatedwithtime(DTIME).Theinclusionofadditionalvariablescanalsodemonstratetheextentofmulticollinearityintheestimatedequation.
Anotherissuethatshouldbeaddressedistheproblemofautocorrelation.Theinclusionofthelaggeddependentvariablereducesautocorrelationconsiderably.Becausewehavealaggeddependentvariableontherighthandside,theDurbin–Watsonstatisticisnotstrictlyapplicable,althoughitdoesgivesomeindicationoftheextentofautocorrelation.Foreachhypothesis,thealternativeDurbinstatisticiscalculated.TheDurbinalternativetestinvolvesregressingtheerrortermoftheprimaryequationwithalltheexplanatoryvariablesandthelaggederrorterm.Asignificanttvalueforthelaggederrortermindicatesthepresenceofautocorrelation.AftercalculatingDurbinalternativetest,theresultsshowusthattheautocorrelationisnotsignificant.
Inthefirststageoftheresearch,allmodelsareestimatedusingordinaryleastsquares(OLS).Thereisapotentialsimultaneityamongtheexplanatorsthatmayrequireaninstrumentalvariables(IV)ortwo-stageleastsquares(TSLS)estimation.IVorTSLSarepreferredtoOLSinthepresenceofsimultaneity.
Afundamentalassumptionofregressionanalysisisthattheright-handsidevariablesareuncorrelatedwiththedisturbanceterm.Ifthisassumptionisviolated,bothOLSandweightedLSarebiasedandinconsistent.Thereareanumberofsituationswheresomeoftheright-handsidevariablesarecorrelatedwithdisturbances.Someclassicexamplesoccurwhen:
-Thereareendogenouslydeterminedvariablesontheright-handsideoftheequation.-Right-handsidevariablesaremeasuredwitherror.
J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280275
Thestandardapproachincaseswheretheright-handsidevariablesarecorrelatedwiththeresidualsistoestimatetheequationusingtheIVregression.TheideabehindIVistofindasetofvariables(termedinstruments)thatareboth(1)correlatedwiththeexplanatoryvariablesintheequationand(2)uncorrelatedwiththedisturbances.Theseinstrumentsareusedtoeliminatethecorrelationbetweenright-handsidevariablesandthedisturbances.TSLSisaspecialcaseofIVregression.Asthenamesuggests,therearetwodistinctstagesinTSLS.Inthefirststage,TSLSfindstheportionsoftheendogenousandexogenousvariablesthatcanbeattributedtotheinstruments.ThisstageinvolvesestimatinganOLSregressionofeachvariableinthemodelonthesetofinstruments.Thesecondstageisaregressionoftheoriginalequation,withallofthevariablesreplacedbythefittedvaluesfromthefirst-stageregressions.ThecoefficientsofthisregressionaretheTSLSestimates.ForadiscussionofthefinitesamplepropertiesofTSLS,seeDavidsonandMacKinnon(1993,pp.221–224)orJohnstonandDiNardo(1997,pp.355–358).TheIVofouranalysesthatarebothcorrelatedwiththeexplanatoryvariables(HV)intheequationanduncorrelatedwiththedisturbancesareDTIME,FDI1,RGGDP,EARN,IPG7,XRATE,andEXPTPTR.
3.1.1.Sociopoliticalinstability
Wetestedthefirsthypothesisusingaproxyvariablethatincorporatesthevariousaspectsofpoliticalinstability.Thisvariableisthepoliticalriskindex(PRI)developedbyBERIinwhichapproximately60specialistsinpoliticsandpoliticalsciencefromallovertheworldevaluateeachcountrywithrespecttointernalcausesofpoliticalriskandthesymptomsofsuchrisk(demonstrations,strikes,streetviolence,etc.).Anaverageofthevaluesdeterminedbythesespecialistsforeachcountryiscalculatedonanannualbasis.Thisindexrangesfrom0(prohibitiverisk)to100(completestability).ThemodelspecificationofHypothesis1is:
FDI¼b0þb1PRIþ
nXi¼2
biCVi
whereb0istheconstantparameter;b1isthehypothesisvariableparameter;biisthecon-trolvariablesparameters;PRIisthehypothesisvariable;andCViisthecontrolvariables.3.1.2.Businessoperatingcontext
EventhoughHelleiner(1989)hasindicatedthatthespecificincentivesestablishedbygovernmentsseemtoplayalimitedroleintheinvestmentdecisionsbetweencountries,thereexistsageneralbeliefthattheconditionsinwhichforeigncompaniesmustoperatearedeterminingfactorsinattractingFDI.Inordertoevaluatethisargument(Hypothesis2),weanalyseoperationriskindex(ORI)asaproxyvariabletothebusinessenvironment.ThemodelspecificationofHypothesis2is:
FDI¼b0þb1ORIþ
nXi¼2
biCVi
whereb0istheconstantparameter;b1isthehypothesisvariableparameter;biisthecontrol
variablesparameters;ORIisthehypothesisvariable;andCViisthecontrolvariables.
276J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280
TheORI,conceivedbyBERI,measuresthedegreetowhichthecomplexconditionsofoperatinginaforeigncountryaffectproductionandprofitsobtainedbytheinvestingcompany.
Thisindexisdesignedtomeasuretheclimateofcommercialoperations.Itiscalculatedusingtwovariablesforeachcountry,positedbyapanelof105expertsfromallovertheworld:
-Thedegreeinwhichapreferentialtreatmentisgiventonationals.
-Thegeneralbusinessclimate,includingpoliticalandadministrativecontinuity.Thisindexwillhavevaluesthatrangefrom0(unacceptablebusinessconditions)to100(unimprovableconditions).
3.1.3.OrientationtoexportandFDI
Inthissection,welookatthelinkbetweenexportsandFDI(Hypothesis3).Lookingatthecorrelationcoefficients,wemustpointouttheimportantcorrelationthatexistsbetweenFDIandtotalexportsinrelationtoGDP(EXPTOTR),whichhasavalueof.95.Thiscorrelationisevengreaterinthecaseofmanufacturedexports(EXPMFR)withavalueof.96thaninthecaseofexportsofprimaryproducts(EXPPPR),whichhaveavalueof.85.TheChineseeconomyclearlydemonstratesatendencytobecomeprogressivelymoreopen,withtherelativeimportanceofexportswithrespecttoGDPshowingatendencytogrow.ThemodelspecificationofHypothesis3is:
FDI¼b0þb1EXPTOTRþ
nXi¼2
biCVi
whereb0istheconstantparameter;b1isthehypothesisvariableparameter;biisthe
controlvariablesparameters;EXPTOTRisthehypothesisvariable;andCViisthecontrolvariables.
3.2.Resultsandinterpretationofthemodel
3.2.1.Sociopoliticalinstability
Table1showsasynthesisoftheresultsobtainedforthedifferentspecificationsofthemodelwithalternativehypothesisvariablesandcontrolvariables.Ascanbeobserved,thevariablehypothesis(PRI)issignificantinthevariousspecificationsofthemodel,itscoefficienttakingapositiveandrelativelystablevalue.
Initially,inModel1,theregressionofFDIwithPRI,adummyvariablethatcapturestheeffectsrelatedwithtime(DTIME)andtheregressiveindependentvariable(FDI1)wascarriedout.PRIcoefficientissignificant.ThevariablesDTIMEandFDI1arealsosignificant.
InModel2,togetherwiththevariablesofModel1,weintroduceacontrolvariable,theexchangerate(XRATE).Ascanbeobserved,thesignremainsthesameandthecoefficientofthevariablehypothesisissignificance.Theestimatedcoefficientassociatedwiththecontrolvariable(XRATE)isstatisticallysignificantandhasapositivevalue.
J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280Table1
Regressionresults(TSLS)ModelCHVDTIMEFDI1IPG7
EXPTPTRXRATEAdjustedR2FstatisticHypothesisvariable(HV)
1
À587,791.10**10,201.26**11,627.78***0.64***
2
3
4
À853,331.00***15,171.19***16,096.60***0.58***
5
6
7
8
9
À504,407.70**8080.69**6286.21***755.2619,421.34***0.57**0.70***
À4744.04***25,792.37*0.97103.22PRI
À788,514.50***13,999.95***6962.48***6504.6221,693.54***0.38***0.71***
À5182.13***
5175.52**0.97112.38ORI
0.97132.49ORI
À31,909.91389,442.707603.00*14,812.76***12,685.25**À5460.91À2040.32À6505.37*0.55***0.45*
À5625.67**
À21,236.68*0.9471.04
EXPTOTR
0.96106.16PRI0.97134.85PRI0.96117.13ORI0.9585.89
EXPTOTR0.9149.38
EXPTOTR
*Significantat10%level.**Significantat5%level.***Significantat1%level.
277278J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280
ThesameoccursinModel3whenweintroduceIPG7asacontrolvariable.Inthiscase,thevalueofthecoefficientissignificantandtakesanegativesign.Accordingtotheexistingliterature,theexpectedsignofthecoefficientoftheIPG7variablecanbeinterpretedinvariousmanners.Thenegativesignobtainedmaydemonstratetoushowahigherlevelofeconomicactivityinthemostdevelopedcountriesimpliesalesserflowofinvestmenttowardsdevelopingcountries(China,forexample)duetogreatervolumesofinvestmentrequiredbytheeconomiesofdevelopedcountriesinexpansion.
3.2.2.Businessoperatingcontext
TheempiricalresultsassociatedwiththevariablehypothesisORI(Models4,5,and6)indicatethatitisstatisticallysignificantinthedifferentspecificationsassociatedwiththemodel(seeTable1).Itisimportanttonotethatthecoefficientofthevariablehypothesisremainsrelativelystableandwithapositivevaluewhendifferentcontrolvariablesareintroduced.ThesignificantcontrolvariablesinthevariousspecificationsforthemodelsareIPG7andEXPTOTR.
3.2.3.OrientationtoexportandFDI
Aswecanseeintheprecedinganalysisofourhypotheses,totalexportsappearasasignificantcontrolvariable.Uponconsideringtotalexports(EXPTOTR)asavariableofthehypothesis,weseethattherelationshipofthatvariabletotheflowsofinvestment(FDI)ispositiveandsignificant,independentlyofthecontrolvariablesintroducedinthedifferentspecifications(Models7,8,and9)ofthemodel(Table1).Theresultssupportthegeneralnotionthatexportsinduceinvestment.
TheseresultswillbeanalysedinthedebateaboutcomplementarilyorsubstitutabilityofinwardFDIandexportsfromhometothehostcountry.SomebidirectionalcausalitymayexistbetweenFDIandexports.Butthequestionremainsastowhetherexport-orientedeconomiesattractedFDI(exportsprecededFDI)orforeigninvestmentencouragedhigherexports(FDIprecededexports).Toanswerthisquestion,weperformtheGrangercausalitytest.AsLeamer(1985)pointsout,Grangercausalitydoesnotimply‘‘causality’’asdefinedinaLayman’stermsbutrather‘‘precedence’’inalead–lagrelationship.Thereisnoevidenceaboutcausalityinthismodel.
4.Conclusions
Theprincipalconclusionsoftheempiricalinvestigationincludedinourstudystemfromamethodologicalfocusinwhichbothmacroeconomicdeterminantsandsociopolit-icalfactorsthatbearupontheflowofdirectforeigninvestmenttowardsChinaaretakenintoaccount.Andalthoughtherearenosimplemodelsorsolidtheoreticalfundamentsthatguidetheempiricalanalysisofthesequestions,wehavebasedourstudyatleastinpartontheresultsofpaststudies,principallyonthestudyofSinghandJun(1995).
Inthisanalysis,besidesthetraditionaleconomicvariables,weincludeothervariablesofasociopoliticalnatureasfactorsthatcanaffecttheflowsofFDIthataredirectedtowardsChina.Amongthese,wecanpointoutsociopoliticalinstability,thebusinessoperatingcontext,andthedegreetowhichtheeconomyisgearedtowardsexportmarkets.
J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280279
Wealsoincludeothervariablesascontrolsinordertoverifyourhypotheses.Theseincludethesizeofthemarket,labourcosts,exchangerates,thecharacteristicsoftheinvestor’scountryoforigin,therelativeimportanceoftheexportssectorinthecountryreceivinginvestment,andtheamountofFDIduringprecedingperiod.
Themethodologythatweelectedrecognisestheimportanceofboththeseeconomicfactorsandthesociopoliticalelementsmentioned,especiallyinacountrythatisgoingthroughaprocessofdevelopment,transition,andreform,asisChina.Fromthepreviousevidence,andaccordingtoourempiricalevidence,wecanpointoutthepoliticalimplicationsthatcanbeextractedfromtheanalysescarriedout,inourcase,forChina.Thefollowingarepoliciesthatmayencouragepotentialinvestors:
-Onaninternallevel:Increasepoliticalstabilityandimprovetheconditionsinwhichcompaniesmustoperate,withoutignoringtheprogramofimprovementofthebasiceconomiccharacteristicssuchastheGDP,theGDPpercapita,thegrowthofGDP,especiallysincethesearethefactorsthatwillcontributetoconsolidatingFDI.
-Onanexternallevel:Intensifyallpoliciespromotingeconomicopenness,fromcommercialpoliciestothoseregulatingFDI.
WemustkeepinmindthatwhicheverinterpretativemodelofFDIweuse,itmaybealteredfromadynamicpointofview.Thisiseventruesointhecaseofacountrythatisintheprocessofdevelopingandwithgoodeconomicprospects,suchasChina.Aninitialinterpretativefocusbasedoninvestinginordertoexportcanbecomplementedwithasecondfocusoninvestinginordertosellintheinternalmarket—thehomemarketoftheforeigninvestorshavingbeensaturated—andinthismannerconsolidateFDIforthefuture.
Acknowledgements
´pezRoaforhiscollaborationandforTheauthorswishtothankProfessorAngelLuisLo
thesupportgiventousduringthepreparationofthispaper.Also,wewishtothankF.T.HanerofBERIforyourcontributiontoprovideusthePRIandORIIndexofChina.Theauthorsareresponsibleforanyremainingerrorsandomissions.References
Agarwal,J.P.(1980).Determinantsofforeigndirectinvestments:Asurvey.WeltwirtschaftlichesArchiv,116,739–777.
Aharoni,Y.(1966).Theforeigninvestmentdecisionprocess.Cambridge,MA:HarvardUniversityPress.
Bennet,P.D.,&Green,R.T.(1972).Politicalinstabilityasadeterminantofforeigndirectinvestmentinmarketing.JournalofMarketingResearch,9,162–186.
Davidson,R.,&MacKinnon,J.G.(1993).Estimationandinferenceineconometrics.London:OxfordUniversityPress.
Dollar,D.(1992).Outward-orienteddevelopingeconomiesreallydogrowmorerapidly:Evidencefrom95LDC’s1976–1985.EconomicDevelopmentandCulturalChange,40,523–544.
´micaenChina:Unmarcointerpretativodelainversio´nextranjera.Giner,G.(2000).Reformayaperturaecono
Tesisdoctoral(PhD).
280J.M.Giner,G.Giner/ChinaEconomicReview15(2004)268–280
´n.IIIJornadasdeGiner,J.M.,&Giner,G.(1999,Abril).ElsistemabancarioenChina:Estructurayevolucio
´ticaEcono´mica(Alicante,Spain).Polı
´micaenChina:Unareflexio´nsobreelGiner,J.M.,&Giner,G.(2000,Diciembre).Aperturayreformaecono
´ticaEcono´mica(Tenerife,Spain).EstadodelBienestar.IVJornadasdePolı
´ndirectaextranjera:Unaaprox-Giner,J.M.,&Giner,G.(2001,Abril).Modelosinterpretativosdelainversio
´´´imacionalcasodeChina.IIIReuniondeEconomıaMundial(Madrid,Spain).
´nextranjeraenChina:Ana´lisisypolı´ticas.VJornadasdeGiner,J.M.,&Giner,G.(2002a,Mayo).Lainversio
´ticaEcono´mica(Bilbao,Spain).Polı
´nextranjeraenChina:Ana´lisisypolı´ticas.Boletı´nICEEcono´mico,Giner,J.M.,&Giner,G.(2002b).Lainversio
2733,43–54.
Hein,S.(1992).Tradestrategyandthedependencyhypothesis:Acomparisonofpolicy,foreigninvestment,andeconomicgrowthinLatinAmericaandEastAsia.EconomicDevelopmentandCulturalChange,40(3),495–521.
Helleiner,G.K.(1989).Transnationalcorporationsanddirectforeigninvestment.Handbookofdevelopmenteconomics,vol.II(pp.1441–1480).Amsterdam:Elsevier.
Johnston,J.,&DiNardo,J.E.(1997).Econometricmethods(4thed.).New-York:McGraw-Hill.Leamer,E.(1985).Sensitivityanalysiswouldhelp.AmericanEconomicReview,73(3),308–313.
Levis,M.(1979).Doespoliticalinstabilityindevelopingcountriesaffectforeigninvestmentflow?Anempiricalexamination.ManagementInternationalReview,19,59–68.
Liu,X.,Wang,Ch.,&Wei,Y.(2001).CausallinksbetweenforeigndirectinvestmentandtradeinChina.ChinaEconomicReview,12(1),40–57.
Lucas,R.(1993).Onthedeterminantsofdirectforeigninvestment:EvidencefromEastandSoutheastAsia.WorldDevelopment,21(3),391–406.
Nigh,D.(1985).TheeffectofpoliticaleventsonUSdirectforeigninvestment:Apooledtime-seriescrosssectionalanalysis.JournalofInternationalBusinessStudies,16,1–17.
Pearce,R.,Islam,A.,&Sauvant,K.(1992).Thedeterminantsofforeigndirectinvestment.Asurveyofempiricalevidence.NuevaYork:UnitedNations.
Rolfe,R.,Ricks,D.,Pointer,M.,&McCarthy,M.(1993).DeterminantsofFDIincentivepreferenceofMNEs.JournalofInternationalBusinessStudies,24(2),335–355.
Root,F.,&Ahmed,A.(1979).Empiricaldeterminantsofmanufacturingdirectforeigninvestmentindevelopingcountries.EconomicDevelopmentandCulturalChange,27,751–767.
Schneider,F.,&Frey,B.(1985).Economicandpoliticaldeterminantsofforeigndirectinvestment.WorldDevelopment,13(2),161–175.
Singh,H.,&Jun,K.W.(1995).Somenewevidencesondeterminantsofforeigndirectinvestmentindevelopingcountries.PolicyResearchWorkingPapers,vol.1531.Washington,DC:WorldBank.
Song,H.,Liu,Z.,&Jiang,P.(2001).AnalisingthedeterminantsofChina’saggregateinvestmentinthereformperiod.ChinaEconomicReview,12(2–3),227–242.
´n:Author.StateStatisticalBureau(1996).Chinastatisticalyearbook1996.Pekı
´n:Author.StateStatisticalBureau(1997).Chinastatisticalyearbook1997.Pekı
´n:Author.StateStatisticalBureau(1998).Chinastatisticalyearbook1998.Pekı
Wallsten,T.(Ed.)(1980).Cognitiveprocessesinchoiceanddecisionbehavior.Hillsdale,NJ:Erlbaum.
Wheeler,D.,&Mody,A.(1992).Internationalinvestmentlocationdecisions:ThecaseofUSfirms.JournalofInternationalEconomics,33,57–76.
Woodward,D.,&Rolfe,R.(1993).Thelocationofexport-orientedforeigndirectinvestmentintheCaribbeanBasin.JournalofInternationalStudies,24(1),121–144.
Wu,X.(2001).Foreigndirectinvestment,intellectualpropertyrights,andwageinequalityinChina.ChinaEconomicReview,11(4),361–384.
Zhang,Q.,&Felmingham,B.(2001).TherelationshipbetweeninwarddirectforeigninvestmentandChina’sprovincialexporttrade.ChinaEconomicReview,12(1),82–99.
Zhao,Y.(2001).Foreigndirectinvestmentandrelativewages.ChinaEconomicReview,12(1),40–57.
因篇幅问题不能全部显示，请点此查看更多更全内容