IMF-促进创新:研发补贴的差异影响(英)_第1页
IMF-促进创新:研发补贴的差异影响(英)_第2页
IMF-促进创新:研发补贴的差异影响(英)_第3页
IMF-促进创新:研发补贴的差异影响(英)_第4页
IMF-促进创新:研发补贴的差异影响(英)_第5页
已阅读5页,还剩62页未读 继续免费阅读

下载本文档

版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领

文档简介

PromotingInnovation

TheDifferentialImpactofR&DSubsidies

RedaCherif,FuadHasanov,ChristophGrimpe,andWolfgang

Sofka

WP/22/192

IMFWorkingPapersdescriberesearchin

progressbytheauthor(s)andarepublishedto

elicitcommentsandtoencouragedebate.

TheviewsexpressedinIMFWorkingPapersare

thoseoftheauthor(s)anddonotnecessarily

representtheviewsoftheIMF,itsExecutiveBoard,

orIMFmanagement.

2022

SEP

©2022InternationalMonetaryFund

WP/22/192

IMFWorkingPaper

EuropeanDepartment

PromotingInnovation:TheDifferentialImpactofR&DSubsidiesPreparedbyRedaCherif,FuadHasanov,ChristophGrimpe,andWolfgangSofka1

AuthorizedfordistributionbyIvannaVladkovaHollar

September2022

IMFWorkingPapersdescriberesearchinprogressbytheauthor(s)andarepublishedtoelicitcommentsandtoencouragedebate.TheviewsexpressedinIMFWorkingPapersarethoseoftheauthor(s)anddonotnecessarilyrepresenttheviewsoftheIMF,itsExecutiveBoard,orIMFmanagement.

ABSTRACT:WeinvestigatetheeffectofR&Dsubsidiesonfirms’innovationbyownership,industry,andfirmsizeusingGermanfirm-leveldata.TheimpactofR&Dsubsidiesisheterogeneousacrossindustriesformultinationalcorporations(MNCs)anddomesticfirmswhileitdoesnotdiffersubstantiallybyfirmsize.DomesticfirmshavealargerresponseinR&Dspendinginlow-techmanufacturing,knowledge-intensiveservices,andtechnologicalserviceswhiletheresponseofdomesticandforeignMNCsisbroadlysimilarandisgreaterinmedium-techandhigh-techmanufacturing.ForeignMNCsubsidiaries’responseintermsofpatentsisgreaterthanthatofdomesticMNCsinmostindustries.

JELClassificationNumbers:H25,H32,030,025

Keywords:

Innovation;patents;researchanddevelopment;R&D;subsidies;multinationals;investment;technologypolicy

Author’sE-MailAddress:acherif@;fhasanov@;cg.si@cbs.dk;ws.si@cbs.dk

1

CherifandHasanov:IMF.Grimpe:CopenhagenBusinessSchoolandZEWLeibnizCentreforEuropeanEconomicResearch.Sofka:CopenhagenBusinessSchoolandUniversityofLiverpoolManagementSchool.TheauthorsthankGraziaSantangeloforexcellentfeedbackandDavidAmaglobeli,RodrigoCerda,RuuddeMooij,MauricioSoto,andYueZhouforhelpfulcomments.TheworkonthispaperstartedwhenWolfgangSofkawasaVisitingScholarintheInstituteforCapacityDevelopmentoftheInternationalMonetaryFund.HethankstheIMFforitshospitality.WearegratefultoJocelyneVanderhaegenforformattingthepaper.Allerrorsareourown.

3

Contents

Abstract

2

Introduction Error!Bookmarknotdefined.

Data,DescriptiveStatistics,andEmpiricalMethodology

6

Results

17

Conclusion

21

References

22

Appendixes 7

2

TABLES

1.R&DSpending(inmillioneuros):DescriptiveStatisticsbyOwnership,Industry,andSize

10

2.PatentApplicationsintheNext5Years:DescriptiveStatisticsbyOwnership,Industry,andSize

11

3.R&DSpending:WeightedRegressions,MatchedSample

18

4.Patents:WeightedRegressions,MatchedSample

1

9

FIGURES

1.R&DSpending(log[1+R&D])byOwnership

12

2.R&DSpending(log[1+R&D])byOwnershipandSubsidy(=1)

12

3.R&DSpending(log[1+R&D])bySizeandSubsidy(=1)

13

4.R&DSpending(log[1+R&D]and5-YearAheadPatents(log[1+Patents])byOwnershipandSubsidy

(Red)

13

5.R&DSpending(log[1+R&D])and5-YearAheadPatents(log[1+Patents])byIndustryandSubsidy

(Red)

14

6.R&DSpending(log[1+R&D])and5-YearAheadPatents(log[1+Patents])bySizeandSubsidy

(Red)

14

4

Introduction

Innovationiskeytoeconomicgrowth,andgovernmentsarekeentopromoteinnovationusingvarioustoolsattheirdisposal.Researchanddevelopment(R&D)activitiesrepresentanimportantingredientintheinnovationprocess,andastandardtoolusedbygovernmenttoencourageitisR&Dsubsidies.Inthiscontext,itisimportantforpolicymakerstoknowhoweffectiveR&Dsubsidiesarebeyondtheaverageoraggregateeffect.Indeed,itwouldbebeneficialtoknowifthesubsidyimpactisdifferentialdependingontheownershiptype(e.g.,foreignownedvs.domestic),firmsize,andindustry(Blanes&Busom,2004).Thiscouldhelpfocusscarceresourceswhileincreasingtheoverallimpact.

Inthispaper,usingGermanfirm-leveldata,coveringthe2000s,westudytheeffectofR&DsubsidiesoninnovationinputssuchasR&Dspendingandoutputssuchaspatents.WefurtherstudywhetherR&Dsubsidieshaveadifferentialeffectonthesubsidiariesofforeignmultinationalcompanies(MNCs)vs.domesticMNCsanddomesticfirms,whetherthefirmsizematters,andwhethertheeffectofsubsidiesdiffersbyindustry.WeuseatreatmentmodeltoidentifytheeffectofR&Dsubsidies.Identificationrequiresmodelingthecounterfactual,thatis,howfirmswouldhaveengagedininnovationactivitieshadtheynotreceivedanysubsidy.Weemployapropensityscorematchingandregressionsonamatchedsampletoestimatetheparametersofinterestinacross-sectionoffirms.

TheempiricalevidencesuggeststhatsubsidiesincreaseR&Dspendingcomparedtothecomparablefirmsthatdonotreceivesubsidies2,andmoreover,resultinhigherfuturepatentingactivity.WefindthattheeffectofR&DsubsidiesonR&DspendingisbroadlysimilarforforeignMNCsubsidiariesanddomesticMNCs.Incontrast,therearesignificantdifferencesintheeffectsonMNCsanddomesticfirmsacrossindustries,butnotbyfirmsize.TheimpactonR&DspendingislargerforMNCsinmedium-andhigh-techmanufacturingthanfordomesticfirmsbutislowerinlow-techmanufacturingandsomeserviceindustries.Thesubsidy’seffectonfuturepatentsisbiggerforMNCsthandomesticfirmsandinterestingly,isbiggerforforeignMNCsthandomesticMNCsinmostindustriesexceptformedium-techmanufacturing.

ThereareseveralreasonsastowhyR&Dsubsidieswouldincreasetheinnovationeffortandperformanceoffirms.Thealleviationoffinancialconstraintswould,forexample,increaseR&Dandconsequently,patentingactivity.ThefactthatMNCsincreaseR&Dspendingmoreincertainindustriesandpatentingactivityingeneralthandomesticfirms,suggeststhatcompetingininternationalmarketsinmedium-andhigh-techmanufacturesrequiresalotofinnovation.ThehigherpatentingofforeignMNCscomparedtodomesticMNCscouldsuggestmoreefficientuseorallocationofresourcestowardpatentabletechnologiesorasignalingtowardheadquarterstopursuemoreinnovationinaforeignlocation.Infact,pursuingmoreinnovationatthetechnologicalfrontier,resultinginbetterinnovationoutcomeslikepatenting,couldbetheresultofknowledgespilloversandtechnologytransfersacrosstheMNCglobalnetwork(Niosi,1999;CantwellandMudambi,2005;UnandCuervo-Cazurra,2008;GelmanandImbens,2013;Santangeloetal.,2016;Cantwell,2017).

Economictheoryprovidesajustificationforthegovernment’ssupportoffirms’innovationactivities.FirmstendtounderinvestinR&Dspendingduetoavarietyofmarketfailures(e.g.,Arrow1962;Kletteetal.,2000).Positiveexternalitiessuchasknowledgespillovers,imperfectinformationaboutfuturereturns,andmarketfrictionsaswellasresourceconstraintswouldproducelessthansociallyoptimalinvestmentand

2Giventhattheamountofthesubsidyisnotavailableinthedataset,wehaveevidencethatthemultiplieronR&Dspendingispositivebutnotnecessarilygreaterthanone.

5

innovationactivities(Bozeman,2000;Davidetal.,2000a;FeldmanandKelley,2006;Ceh,2009;RigbyandRamlogan,2012;Lazzarini,2015;Holmesetal.,2016).Inaddition,governmentsmayalsointervenetotacklepotentialsystemicproblemsintheinnovationecosystem(SalterandMartin,2001).Theseincludetechnologicaltransitions,lock-inproblemsduetoanexcessivefocusonexistingtechnologies,andinnovationnetworkdeficiencies(ChaminadeandEdquist,2008).Thesystemofinnovationapproachemphasizestheimportanceofinnovationnetworksinwhichfirmsdonotinnovateinisolationbutcontinuouslyinteractwithotherinnovationagentssuchasotherfirms,universities,orpublicresearchorganizations(Lundvall,1992).

AlthoughthereisadebateabouttheeffectivenessofR&Dprograms,alargeliteratureineconomicsandmanagementhasmostlyfoundapositiveeffectoftheseprogramsoninnovationsuchasR&Dspendingandpatents(CzarnitzkiandToole,2007;Aerts&Schmidt,2008;GonzalezandPazo,2008;Zúñiga-Vicenteetal.,2014;JaffeandLe,2015;Howell,2017).Subsidiescouldnarrowthescopeoffirms’innovationsolutions(Elyetal.,2014),increasethecostsofR&Dinputssuchaswagesofresearchers(Davidetal.,2000b;Clarysseetal.,2009),orpropuplaggingfirmsintermsofinnovationcapacityratherthanfacilitatemovementofskilledlabortohigh-techfirms(Acemogluetal.,2018).

Moreimportant,subsidiescouldpotentiallycrowdoutR&Dspending,resultinginnoincreaseinR&Dspendingfundedbyfirms,buttheempiricalevidenceshowsotherwise.Essentially,azeromultiplierimpliesthattheincentiveeffectofsubsidiestoengageinmoreR&Dandeventuallypatentingactivitiesisoffsetbythereallocationofspendingtowardotheractivitiesthatthefreeingoffirmresourcesallowsafterasubsidyreceipt.Priorliterature,however,suggeststhatoverall,thereisnooronlypartialcrowdingoutofR&Dinvestments(e.g.,JaffeandLe,2015;DimosandPugh,2016).Moreover,theelasticityofR&Dtotaxcreditsisequaltoorgreaterthanone(e.g.,Bloom,Griffith,andvanReenen,2002;Rao,2016),andbenefitsoftaxincentivesexceedtheircosts(LesterandWarda,2020).TargetedR&Dsubsidiescouldalsoencouragelong-termresearchsuchascancerresearch(Budish,Roin,andWilliams,2015).R&DsubsidiescouldalsopotentiallynudgesubsidiariesofforeignMNCstobemoreinnovative(BirkinshawandHood,1998;CantwellandMudambi,2005;Cantwell,2017).

R&Dsupporttofirmshasstrongimpactoninnovationoutcomessuchaspatents.R&Dsubsidieshaveastatisticallysignificantandeconomicallylargeeffectonpatentsandventurecapitalinvestment,enablingproof-of-conceptthatwouldnothavebeenfinanciallypossibleotherwise(Howell,2017).IntheUK,thereisevidencethattaxcreditsincreasebothpatentingandcitations(Dechezleprêtre,Einiö,Martin,NguyenandvanReenen,2016)whileR&DsubsidiesincreasepatentinginItaly(BronziniandPiselli,2016).Taxcreditsalsoincreaseproductandprocessinnovation(Czarnitzkietal.,2011;Cappelenetal.,2012)whileR&Dsupportinducesinnovationcollaborationamongfirmsintheinnovationecosystem(e.g.,OECD,2006;Unetal.,2010;GökandEdler,2012)andincreaseslaborproductivityoffirms(MinnitiandVenturini,2017).

IntermsoftheheterogeneityoftheR&Dsubsidyimpact,Sofkaetal.(2021),usingGermanfirm-leveldata,findthatbothforeignanddomesticMNCsincreaseR&DbythesameamountbutspendmoreonR&Dthandomesticfirms.Inaddition,foreignMNCspatentmorethandomesticMNCsanddomesticfirms.Incontrast,thispaperexplorestheheterogeneousimpactofsubsidiesbyindustryandfirmsizeandassessesthesubsidyimpactonR&Dspendingandpatentsusingmatched-firmregressions.

6

Data,DescriptiveStatistics,andEmpiricalMethodology

GovernmentR&DpoliciesinGermany

GermanyisaninterestingcasestudytoanalyzeR&Dpolicy’simpactoninnovationasitisbothR&Dintensiveandopen.TheR&DintensityoftheGermaneconomyisabout3percentofGDP,andR&Dexpenditureshaveincreasedinboththegovernmentandbusinesssectors(Sofkaetal.,2018).Germanyisalsoamajorforeigndirectinvestorintermsofoutflows(FDIoutflowpositionsofabout44percentofGDPin2017)aswellasinflows(FDIinflowpositionsofabout26percentofGDPin2017)(OECD,2019).Inaddition,MNCshaveoperatedsubsidiariesinGermanyforalongtime—forinstance,investmentsbytheUScarmanufacturersGeneralMotorsandForddatebacktothefirsthalfofthe20thcentury(deFaria&Sofka,2010).

UsingtheGermanfirm-leveldataallowsustobettercontrolforvariousinstitutionalconditions,allowingustofocusoninnovationimplicationsofR&Dpolicy.Germanyhasastableinstitutionalenvironmentsuchasastrongintellectualpropertyrightsregime(Park,2008;PapageorgiadisandSofka,2020)aswellascontinuouspoliticalsupportforinnovation(EFI,2017),whichcouldpositivelyaffectR&Ddecisions.Germanyprovidesinformationandadministrativesupportforpotentialforeigninvestors(GermanyTradeandInvest,GATI.de)butdoesnotprovidetaxincentives.ThecountryalsohaslowrestrictionsforFDI,whichcouldbebeneficialforMNCs.Interestingly,therearenoR&Dtaxcredits(Sofkaetal.,2018),andgovernmentR&Dsubsidiesareapplication-basedgrants.

ResearchandinnovationpolicyinGermanyisthesharedresponsibilityoftheFederalGovernmentandthe16State(“Laender”)Governments.Atthefederallevel,theFederalMinistryofEducationandResearch(BMBF)drivesmostpolicyinitiativeswhiletheFederalMinistryofEconomicsandEnergy(BMWi)isinvolvedinspecificareas(Sofkaetal.,2018).SupportforresearchandinnovationinprivatefirmsisanimportantcomponentofGermany’sHigh-TechStrategyandNationalReformPrograms(NRP,2017).Germanfirmscanapplyforgovernmentsupport,followingaformalapplicationprocessandtheguidelinesoftheEuropeanUnion(BMBF,2016a).Thetypicalsupportinstrumentisagrant.

Thereisabroadvarietyofsupportschemesinplace,whichareoftenoperatedbyspecializedprojectmanagementorganizations(“Projekttraeger”)orassociationsliketheGermanFederationofIndustrialResearchAssociations(AiF)“OttovonGuericke”(SofkaandSprutacz,2017).Otherpolicyinitiativestargetparticularareas,e.g.ITsecurity(BMBF,2016c)orgroupsoffirmslikestartups(e.g.High-TechStartupFund)orsmallandmediumsizedfirms(“Mittelstand”)(RKW,2017).Stategovernmentscomplementthesupportschemesofthefederalgovernmentwiththeirownmeasures(BMBF,2016b).Oftenthesemeasuresreflecttheprioritiesorstructuresofthestate.Forexample,thestateofBaden-Wuerttembergidentifiessustainablemobilityasoneofitsresearchprioritytopics(Sofkaetal.,2018).

7

Data

OurdatasetmergesdatafromarepresentativeinnovationsurveyoffirmsinGermanywithpatentstatisticsfromtheEuropeanPatentOffice(EPO).Theinnovationsurveyusedisthe“MannheimInnovationPanel”(MIP),whichistheGermancomponentoftheCommunityInnovationSurvey(CIS)oftheEuropeanUnion.IncontrasttomanyotherCISsurveys,theMIPisconductedannuallyandallowstheconstructionofanunbalancedfirmpaneldataset.MIPrespondentsareresponsibleforinnovationtopicsintheirfirmssuchasCEOs,headsofR&D,orinnovationmanagement.Theyareaskedtoansweracomprehensivesetofquestionsaboutinnovationinputsaswellasoutputsandtoassignimportanceratings(Criscuoloetal.,2005).TheMIPprovidesastratifiedrandomsample,whichisrepresentativeoffirmsinGermany.Non-responseanalysesshownosystematicdistortionsbetweenrespondingandnon-respondingfirms(Rammeretal.,2005),andEurostat(2009)considersCISdatafromGermanyashighquality.

Weobtainfirm-leveldatafromtheMIPfortheyears2000,2002,2003,2004,and2006.Thesurveywavesfortheyears2001and2005arenotusefulforouranalysissincetheydonotincludequestionsonthereceiptofR&Dsubsidies,whichisthecentralvariableinourstudy.ThefirminformationobtainedfromtheMIPismergedwiththepatentstatisticsfromtheEPOusingassigneenamesandaddresses.Patentstatisticsareavailableforlongertimeperiodsthanthesurveydata.Weusepatentapplicationsbetween1997and2011toconstructthestockofpatentapplicationspriortooursurveyperiodandpatentapplicationsforuptofiveyearsafterthesurveyyear.

Asadependentvariable,weusebothR&Dexpendituresreportedaswellasthenumberofafirm’sEPOpatentapplicationsinthesubsequentfiveyears.SincethereisasignificanttimedelaybetweeninvestinginR&Dandarrivingatapatentableinventionandlengthypatentfilingprocedures,weuseafive-yeartimewindowforthepatentapplicationsvariable.Longertimewindowsformeasuringpatentnumbersincreasetheriskofconfoundingfactors.

OurcentralindependentvariablesofinterestarerelatedtoR&Dsubsidy(recipientvs.non-recipient),firmownership(foreignMNC,domesticMNC,anddomesticfirm)andsize(largevs.small),andindustry(low-,medium-andhigh-techmanufacturingandvariousservices).Weusedummyvariablestoidentifythesegroupsasfollows:ThesurveyrespondentsindicatewhethertheirfirmshavereceivedR&DsubsidyfromtheGermangovernment(stateand/orfederallevel).3Weclassifyfirmsintoadomesticfirm,adomesticMNC,orasubsidiaryofaforeignMNC.Thelattergroupisbasedonwhetherafirmispartofacompanygroupwithheadquartersabroad,inlinewithpreviousresearchontheinnovationactivitiesofforeignMNCsubsidiaries(Sofkaetal.,2014).Thefirmsizeisbasedonthenumberofemployees,andwesplitfirmsintothreecategories:smallwith50orlessemployees,mediumwith51-500employees,andlargewithmorethan500employees.Lastly,wecreate6industrydummiesbasedongroupedtwo-digitNACEcodes,whichhavebeenusedfrequentlyinpreviousinnovationstudies(Grimpeetal.,2017).Thesearelow-techmanufacturing(e.g.,foodandtextiles),medium-techmanufacturing(e.g.,motorvehicles),high-techmanufacturing(e.g.,medicaldevices),distributiveservices(e.g.,logistics),knowledge-intensiveservices(e.g.,consulting),andtechnologicalservices(e.g.,ICT-relatedservices).ThedescriptionisprovidedintheAppendixTable1.

3Thesurveydoesnotprovideinformationonthesubsidyamount.TherelatedquestioninthesurveyintroducesgovernmentsubsidiesassupportforR&Dand/orinnovationwithashortdescription.Then,respondentschoosethefundingsource.Withinourcontext,agovernmentR&Dsubsidycanbeobtainedfromthefederalgovernmentoroneofthe16stategovernmentsinGermanyinwhichthefirmoperates.Firmsmightobtainmultiplegrantsfromoneormultiplegovernmentsources,butthisinformationisnotavailableinthesurvey.

8

Weincludeadditionalcontrolvariablestocaptureotherfactors,whichcouldpotentiallyaffectfirm’sinnovationperformance.Wecontrolfortheoverallsize(numberofemployees)andageoffirms(numberofyearssincefoundedinGermany).Wealsoattempttoaccountfordifferencesinfirms’innovationcapacitiesbyaddingfirm’spatentapplicationsintheyearofobservationsincetheseinventionsarelikelytopredateR&Dpotentiallyinducedbyagovernmentsubsidy.Wealsoaddthenumberofpatentapplicationsoffirmsinthethreeyearsprecedingoursample,thatis,1997,1998and1999,followingBlundelletal.(2002).Wecapturedifferencesintheskillsofemployeesbycontrollingfortheshareofemployeeswithcollegeeducation.WeaddadummyvariableforwhetherthefirmengagesinR&Dcontinuously.ThisvariableisfrequentlyusedtoindicatethepresenceofadedicatedR&Ddepartment(CzarnitzkiandLicht,2006;Koehleretal.,2012).Wecontrolforthedegreeofinternationalizationthroughtheshareofexportsinfirmsalessinceinternationalizationhasbeenfoundtoaffectafirm’sinnovationactivities(CassimanandGolovko,2011).Additionally,weincludeadummyvariableforwhetherafirmengagesinprocessinnovationfromthesamesurveysincesuchactivitiesmayaffectitsabilitytopatent.Lastly,weaddtimedummiestocontrolformacroeconomicconditions.

Ourdatasetcontains6,052firmobservationsafterexcludingfirmsthathavenotengagedineitherproductorprocessinnovationandwithlessthan5employees.Afterdroppingoutliers,thesampleincludes5,717observations,ofwhich5,623canbematchedwithcontrolfirms.4Inthepatentregressions,thesamplewithoutoutliershas5,353observations,whilethematchedsamplecontains5,263observations.Weperformregressionsonthefullsampleof6,052observationsandthematchedsamples.5

Descriptivestatistics

Inthefullsample,firmsspendonaverage€1.6milliononR&Dandfileabout2.3patentsina5-yearperiod.AboutathirdofthemhasreceivedanR&Dsubsidyfromstateorfederalgovernments,irrespectiveofownershiporsize.Largefirmsthatreceiveasubsidycompriseabout12percentofsubsidy-receivingfirms.Smallfirmsaccountforone-halfofsubsidy-receivingfirms.Thisdistributionissimilartothefirmdistributionobserved.Abouthalfoffirmsissmallwithlessthan50employeesandabout13percentislargefirmswithmorethan500employees.Thesubsidyreceiptdiffersbyindustrywithabouthalfoffirmsinhigh-techmanufacturingandtechnologicalservicesreceivingasubsidywhile10-15percentoffirmsreceiveasubsidyinknowledge-intensiveanddistributiveservices(AppendixTables3-5).AbouthalfofthefirmsareengagedincontinuousR&Dactivitiesandabouttwo-thirdsofthemareprocessinnovators.ForeignMNCsubsidiariescompriseabout10percentoffirms,similartothatofdomesticMNCs,i.e.,headquarteredinGermany.Aboutathirdoffirmsisinlow-techmanufacturing,about18percenteachinmedium-techandtechnologicalservices,10percenteachinhigh-techmanufacturingandknowledge-intensiveserviceswhile14percentoffirmsisindistributiveservices.

AcursorylookatR&Dspendingrevealsafewinterestingpatterns(Table1).MNCs,whetherdomesticorforeignMNCsubsidiaries,domoreR&DandfilemorepatentapplicationsthantheirdomesticcounterpartswhiledomesticMNCstendtooutperformtheirforeignMNCsubsidiaries(Figure1).Yetinterestingly,R&Dspendinginlow-techmanufacturingbydomesticfirmsisatparwithforeignMNCs,albeitwithfarsmallerdispersioninthelatter(similarly,indistributiveservices).Incontrast,domesticMNCsdomuchmoreR&Dinlow-techmanufacturingandinfact,onaverageexceedR&Dinalltheserviceindustries,including

4Theextremevaluesofthefollowingvariables—firmsabove100yearsofageaswellasobservationsabovethe99thpercentileofthedistributionsofthenumberofemployees,patentstock,shareofR&Dinsales,andexports—aredropped.

5Theestimationresultsbetweenthematchedsampleandtheunmatchedsamplethatexcludesoutliersaresimilar.

9

sophisticatedones,whetherbydomesticfirmsorforeignMNCs.ThelargestR&DspendingofdomesticMNCsisinmedium-techmanufacturing,andalthoughthesecondlargestspendingindustryishigh-techmanufacturing(bothexceedingotherindustriesbyfar),domesticMNCs,onaverage,slightlyfallbehindforeignMNCsubsidiariesinhigh-techmanufacturing.

Thepatentsstatisticstellusalargelysimilarstory(Table2).Onedifferenceisthatforeignsubsidiariestendtofilemorepatentsinlow-techmanufacturingthandomesticfirms,butthispatternisreversedinknowledge-intensiveserviceindustry.

Largefirms(withmorethan500employees),notsurprisingly,alsooutperformtheirsmallercounterparts(Tables1-2).EvendomesticlargefirmstendtospendmoreonR&DandfilemorepatentsonaveragethansmallerdomesticMNCsandforeignMNCsubsidiaries.Interestingly,smalldomesticMNCstendtofilemorepatentsthanotherfirms,includingevenallmedium-sizedfirms(butthesamplesizeofmedium-sizedMNCsisrelativelysmall).

Inaddition,whetherbyindustry,size,orownership,firmsreceivingaR&DsubsidytendtohavelargerR&Dspendingthanfirmsthatdonotreceiveasubsidy(Figures2-6).Itisnotjustthelargemedianoraveragewithonlyafewfirmsskewingtheimpact;rather,manyfirmshavehigherspendingwithalargemassnotfarabovethemedianofthedistribution.Theimpactismuchmorepronouncedinlargefirms,domesticMNCs,andforeignMNCsubsidiaries.BothdomesticMNCsandforeignMNCsubsidiaries,receivingasubsidy,havegreaterR&Dspendingandpatenting(Figure4).Furthermore,firmsinmanufacturingandtechnologicalservicestendtodomoreR&Dandpatentingwithasubsidyreceipt(Figure5).whileasubsidyseemstoincreaseinnovationactivitiesofsmallfirmssubstantially(Figure6).

10

Table1.R&DSpending(inmillioneuros):DescriptiveStatisticsbyOwnership,Industry,andSize

Industry

Domestic

DomesticMNCs

ForeignMNCSubsidiaries

All

Low-tech

manufacturing

Mean

St.Dev.

Obs.

0.5412.46

1562

2.0016.92

204

0.551.36

172

0.6912.47

1938

Medium-tech

manufacturing

Mean

St.Dev.

Obs.

0.401.44

737

18.9993.15

191

5.4145.97

159

4.4043.30

1087

High-tech

manufacturing

Mean

St.Dev.

Obs.

0.321.34

427

11.8451.21

75

15.9259.41

92

4.1930.16

594

Distributiveservices

Mean

St.Dev.

Obs.

0.497.91

702

1.095.14

72

0.351.42

59

0.537.43

833

Knowledge-intensiveservices

Mean

St.Dev.

Obs.

0.211.25

480

1.428.00

59

0.934.11

41

0.383.00

580

Technological

services

Mean

St.Dev.

Obs.

0.331.68

901

温馨提示

  • 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
  • 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
  • 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
  • 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
  • 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
  • 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
  • 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。

最新文档

评论

0/150

提交评论