Multinational Firms, FDI Flows and Imperfect Capit

MULTINATIONALFIRMS,FDIFLOWS,ANDIMPERFECTCAPITALMARKETS∗POLANTR`ASMIHIRA.DESAIC.FRITZFOLEYThispaperexamineshowcostlyfinancialcontractingandweakinvestorpro-tectioninfluencethecross-borderoperational,financing,andinvestmentdecisionsoffirms.Wedevelopamodelinwhichproductdeveloperscanplayausefulroleinmonitoringthedeploymentoftheirtechnologyabroad.Theanalysisdemonstratesthatwhenfirmswanttoexploittechnologiesabroad,multinationalfirm(MNC)activityandforeigndirectinvestment(FDI)flowsariseendogenouslywhenmon-itoringisnonverifiableandfinancialfrictionsexist.ThemechanismgeneratingMNCactivityisnottheriskoftechnologicalexpropriationbylocalpartnersbutthedemandsofexternalfunderswhorequireMNCparticipationtoensurevaluemaximizationbylocalentrepreneurs.ThemodeldemonstratesthatweakinvestorprotectionslimitthescaleofMNCactivity,increasetherelianceonFDIflows,andalterthedecisiontodeploytechnologythroughFDIasopposedtoarm’slengthtechnologytransfers.Severaldistinctivepredictionsfortheimpactofweakin-vestorprotectiononMNCactivityandFDIflowsaretestedandconfirmedusingfirm-leveldata.I.INTRODUCTIONFirmsglobalizingtheiroperationsandtheassociatedcap-italflowshavebecomemajorfeaturesoftheworldeconomy.Thesecross-borderactivitiesandcapitalflowsspaninstitutionalsettingswithvaryinginvestorprotectionsandlevelsofcapi-talmarketdevelopment.Althoughtheimportanceofinstitutionalheterogeneityindictatingeconomicoutcomeshasbeenempha-sized,existinganalysestypicallyignoretheglobalfirmsandthecapitalflowsthatarenowcommonplace.Investigatinghowglobalfirmsmakeoperationalandfinancingdecisionsinaworldofhet-erogeneousinstitutionspromisestoprovideanovelperspectiveonobservedpatternsofflowsandfirmactivity.∗Thestatisticalanalysisoffirm-leveldataonU.S.multinationalcompanieswasconductedattheBureauofEconomicAnalysis,U.S.DepartmentofCom-merce,underarrangementsthatmaintainlegalconfidentialityrequirements.TheviewsexpressedarethoseoftheauthorsanddonotreflectofficialpositionsoftheU.S.DepartmentofCommerce.TheauthorsthankRobertBarro,fouranonymousreferees,GitaGopinath,JamesMarkusen,AlehTsyvinski,BillZeile,andsemi-narparticipantsatBostonUniversity,BrownUniversity,HitotsubashiUniversity,MIT,theNBERITIprogrammeeting,theNewYorkFed,Oxford,UCBerkeley,UCBoulder,UniversidaddeVigo,UniversitatPompeuFabra,theUniversityofMichi-gan,andtheWorldBankforhelpfulsuggestions.DavinChorprovidedexcellentresearchassistance.C2009bythePresidentandFellowsofHarvardCollegeandtheMassachusettsInstituteofTechnology.TheQuarterlyJournalofEconomics,August200911711172QUARTERLYJOURNALOFECONOMICSThispaperdevelopsandtestsamodeloftheoperationalandfinancialdecisionsoffirmsastheyexploittheirtechnologiesincountrieswithdifferinglevelsofinvestorprotections.Themodeldemonstratesthatmultinationalfirm(MNC)activityandforeigndirectinvestment(FDI)ariseendogenouslyinsettingscharac-terizedbyfinancialfrictions.Themodelgeneratesseveralpredic-tionsregardinghowinvestorprotectionsinfluencetheuseofarm’slengthtechnologytransfers,thedegreetowhichMNCactivityisfinancedbycapitalflows,theextenttowhichmultinationalstakeownershipinforeignprojects,andthescaleofmultinationaloper-ations.Thesepredictionsaretestedusingfirm-leveldataonU.S.MNCs.Themodelconsiderstheproblemofafirmthathasdevel-opedaproprietarytechnologyandisseekingtodeploythistech-nologyabroadwiththehelpofalocalentrepreneur.Avarietyofalternativearrangements,includinganarm’slengthtechnol-ogytransferordirectlyowningandfinancingtheentitythatusesit,areconsidered.Externalinvestorsareapotentialsourceoffunding,buttheyareconcernedwithmanagerialmisbehav-ior,particularlyinsettingswhereinvestorprotectionsareweak.Thecentralpremiseofthemodelisthatdevelopersoftechnolo-giesareparticularlyusefulmonitorsforensuringthatlocalen-trepreneursarepursuingvaluemaximization.Theconcernsofexternalfundersregardingmanagerialmisbehaviorleadtoopti-malcontractsinwhichthedeveloperofthetechnologyisrequiredtoholdanownershipclaimintheforeignprojectand,incertaincases,thisdeveloperisalsorequiredtoprovidefinancialcapitaltothelocalentrepreneur.Assuch,MNCsandFDIflowsariseendogenouslyinresponsetoconcernsovermanagerialmisbehav-iorandweakinvestorprotections.1Extendingthemodeltoallowforasimilarformofmonitoringbyexternalinvestorsdoesnotvitiatetheprimaryresults.Wealsoshowthatalthoughsimplerevenue-sharingagreementsmayalsoprovideincentivesfortech-nologydeveloperstomonitor,thistypeofcontractisgenerallynotoptimal.1.TheexperienceofDisneyinJapan,asdocumentedinMisawa(2005),pro-videsoneexampleofthemechanismthatdrivesthebehaviorofexternalinvestors.In1997,DisneywasevaluatinghowtostructureanewopportunitywithalocalpartnerinJapan.Japanesebanksexpressedastrongpreferenceforequitypartic-ipationbyDisneyoveralicensingagreementinordertoensurethatDisneyhadstrongincentivestomonitortheprojectandensurevaluemaximization.Thecon-cernsoftheselendersandtheintuitionthatDisneywouldhaveauniqueabilitytomonitorlocalpartnersarereflectiveofthecentralideasofthemodel.MULTINATIONALFIRMSANDIMPERFECTCAPITALMARKETS1173ThecharacterizationofMNCsasdevelopersoftechnologieshaslongbeencentraltomodelsexplainingMNCactivity.In