The Impact of Trade on Intra-Industry Reallocation

TheImpactofTradeonIntra-IndustryReallocationsandAggregateIndustryProductivityAuthor(s):MarcJ.MelitzSource:Econometrica,Vol.71,No.6(Nov.,2003),pp.1695-1725Publishedby:TheEconometricSocietyStableURL::08/04/201102:27YouruseoftheJSTORarchiveindicatesyouracceptanceofJSTOR'sTermsandConditionsofUse,availableat.://=econosoc..EachcopyofanypartofaJSTORtransmissionmustcontainthesamecopyrightnoticethatappearsonthescreenorprintedpageofsuchtransmission.JSTORisanot-for-profitservicethathelpsscholars,researchers,andstudentsdiscover,use,andbuilduponawiderangeofcontentinatrusteddigitalarchive.Weuseinformationtechnologyandtoolstoincreaseproductivityandfacilitatenewformsofscholarship.FormoreinformationaboutJSTOR,[e-mail]ollaboratingwithJSTORtodigitize,preserveandextendaccesstoEconometrica.(November,2003),1695-1725THEIMPACTOFTRADEONINTRA-INDUSTRYREALLOCATIONSANDAGGREGATEINDUSTRYPRODUCTIVITYBYMARCJ.MELITZ1Thispaperdevelopsadynamicindustrymodelwithheterogeneousfirmstoanalyzetheintra-industryeffectsofinternationaltrade.Themodelshowshowtheexposuretotradewillinduceonlythemoreproductivefirmstoentertheexportmarket(whilesomelessproductivefirmscontinuetoproduceonlyforthedomesticmarket)andwillsimul-taneouslyforcetheleastproductivefirmstoexit.Itthenshowshowfurtherincreasesintheindustry'sexposuretotradeleadtoadditionalinter-firmreallocationstowardsmoreproductivefirms.Thepaperalsoshowshowtheaggregateindustryproductivitygrowthgeneratedbythereallocationscontributestoawelfaregain,thushighlightingabenefitfromtradethathasnotbeenexaminedtheoreticallybefore.ThepaperadaptsHopenhayn's(1992a)dynamicindustrymodeltomonopolisticcompetitioninageneralequilibriumsetting.Insodoing,thepaperprovidesanextensionofKrugman's(1980)trademodelthatincorporatesfirmlevelproductivitydifferences.Firmswithdifferentproductivitylevelscoexistinanindustrybecauseeachfirmfacesinitialuncertaintycon-cerningitsproductivitybeforemakinganirreversibleinvestmenttoentertheindustry.Entryintotheexportmarketisalsocostly,butthefirm'sdecisiontoexportoccursafteritgainsknowledgeofitsproductivity.KEYWORDS:Intra-industrytrade,firmheterogeneity,firmdynamics,selection.1.INTRODUCTIONRECENTEMPIRICALRESEARCHusinglongitudinalplantorfirm-leveldatafromseveralcountrieshasoverwhelminglysubstantiatedtheexistenceoflargeandpersistentproductivitydifferencesamongestablishmentsinthesamenarrowlydefinedindustries.Someofthesestudieshavefurthershownthatthesepro-ductivitydifferencesarestronglycorrelatedwiththeestablishment'sexportstatus:relativelymoreproductiveestablishmentsaremuchmorelikelytoex-port(evenwithinso-calledexportsectors,asubstantialportionofestab-lishmentsdonotexport).Otherstudieshavehighlightedthelargelevelsofresourcereallocationsthatoccuracrossestablishmentsinthesameindustry.Someofthesestudieshavealsocorrelatedthesereallocationswiththeestab-lishments'exportstatus.Thispaperdevelopsadynamicindustrymodelwithheterogeneousfirmstoanalyzetheroleofinternationaltradeasacatalystfortheseinter-firmreallo-cationswithinanindustry.Themodelisabletoreproducemanyofthemostsalientpatternsemphasizedbyrecentmicro-levelstudiesrelatedtotrade.Themodelshowshowtheexposuretotradeinducesonlythemoreproductivefirmstoexportwhilesimultaneouslyforcingtheleastproductivefirmstoexit.Boththeexitoftheleastproductivefirmsandtheadditionalexportsalesgainedby'ManythankstoAlanDeardorff,JimLevinsohn,andElhananHelpmanforhelpfulcommentsanddiscussions.Thismanuscripthasalsobenefitedfromcommentsbytheeditorandtwoanony-mousreferees.FundingfromtheAlfredP.SloanFoundationisgratefullyacknowledged.16951696MARCJ.MELITZthemoreproductivefirmsreallocatemarketsharestowardsthemoreproduc-tivefirmsandcontributetoanaggregateproductivityincrease.Profitsarealsoreallocatedtowardsmoreproductivefirms.Themodelisalsoconsistentwiththewidelyreportedstoriesinthebusinesspressdescribinghowtheexposuretotradeenhancesthegrowthopportunitiesofsomefirmswhilesimultaneouslycontributingtothedownfallordownsizingofotherfirmsinthesamein-dustry;similarly,protectionfromtradeisoftenreportedtoshelterinefficientfirms.Rigorousempiricalworkhasrecentlycorroboratedthisanecdotalevi-dence.BernardandJensen(1999a)(fortheU.S.),Aw,Chung,andRoberts(2000)(forTaiwan),andClerides,Lack,andTybout(1998)(forColombia,Mexico,andMorocco)allfindevidencethatmoreproductivefirmsself-selectintoexportmarkets.Aw,Chung,andRoberts(2000)alsofindevidencesug-gestingthatexposuretotradeforcestheleastproductivefirmstoexit.Pavcnik(2002)directlylooksatthecontributionofm