Why Do Firms Train,Empirical Evidence on the Firs

IZADPNo.319WhyDoFirmsTrain?EmpiricalEvidenceontheFirstLabourMarketOutcomesofGraduatedApprenticesRobEuwalsRainerWinkelmannDISCUSSIONPAPERSERIESForschungsinstitutzurZukunftderArbeitInstitutefortheStudyofLaborJune2001
        !
 ! 

 DiscussionPaperNo.319June2001IZAP.O.Box7240D-53072BonnGermanyTel.:[phone]Fax:[phone]Email:[e-mail]issuedwithintheframeworkofIZA’sresearcharea   .Anyopinionsexpressedherearethoseoftheauthor(s)andnotthoseoftheinstitute.ResearchdisseminatedbyIZAmayincludeviewsonpolicy,buttheinstituteitselftakesnoinstitutionalpolicypositions.TheInstitutefortheStudyofLabor(IZA)inBonnisalocalandvirtualinternationalresearchcenterandaplaceofcommunicationbetweenscience,politicsandbusiness.IZAisanindependent,nonprofitlimitedliabilitycompany(GesellschaftmitbeschränkterHaftung)supportedbytheDeutschePostAG.ThecenterisassociatedwiththeUniversityofBonnandoffersastimulatingresearchenvironmentthroughitsresearchnetworks,researchsupport,andvisitorsanddoctoralprograms.IZAengagesin(i)originalandinternationallycompetitiveresearchinallfieldsoflaboreconomics,(ii)developmentofpolicyconcepts,and(iii)disseminationofresearchresultsandconceptstotheinterestedpublic.Thecurrentresearchprogramdealswith(1)mobilityandflexibilityoflabormarkets,(2)internationalizationoflabormarketsandEuropeanintegration,(3)thewelfarestateandlabormarkets,(4)labormarketsintransition,(5)thefutureofwork,(6)projectevaluationand(7)generallaboreconomics.IZADiscussionPapersoftenrepresentpreliminaryworkandarecirculatedtoencouragediscussion.Citationofsuchapapershouldaccountforitsprovisionalcharacter.IZADiscussionPaperNo.319June2001#$!%
        ∗Theapprenticeshipsystemisthemostimportantsourceofformalpost-secondarytraininginGermany.Ourpapercontributestotheongoingdebateastowhyfirmsarewillingtoinvestinsuchtrainingeventhoughmanyapprenticeswillleavethetrainingfirmsoonaftercompletionoftheapprenticeship.UsingGermanregisterdata–theIABEmploymentSample–wefindthatapprenticesstayingwiththeirtrainingfirmaftergraduationhave(1)higherwagesand(2)longerfirst-jobdurationsthanapprenticesleavingthetrainingfirm.Theseresultssupporttheoriesaccordingtowhichfirmsusetheapprenticeshipsystemtoselectandretainthemoreableapprentices,therebyrecoupingthecostsofinvestinginskillsthatareportableinprinciple.JELCodes:C24,C41,J24,J31,J44Keywords:Training,wages,job-durationRainerWinkelmannIZAP.O.Box7240D-53072BonnGermanyTel.:[phone]Fax:[phone]Email:[e-mail]∗WewishtothankWijiArulampalam,JanvanOurs,StefanWolter,andourcolleaguesatIZAforvaluablecommentsandhelpfuldiscussions.TheauthorsgratefullyacknowledgetheGermanInstituteforEmploymentResearch(IAB)forprovidingthedata.21.IntroductionIninternationalcomparisonsoneducation,vocationaltraining,andlabourmarketperformance,Germanyshowsahighinvolvementoffirmsintheeducationandtrainingoftheyoungbymeansofawell-developedapprenticeshipsystem.ThisfactisseenasapositivecharacteristicoftheGermaneducationandlabourmarketsystem,asitcontributestothelowyouthunemploymentratesandthehighgeneralskilllevelsoftheworkforce.ApeculiarfactaboutGermanapprenticeshiptraining(GAT)isthatitmainlyprovidesgeneraltrainingandthusportableskills.Inaddition,thereisampleempiricalevidencethatfirmsactuallymakesubstantialnetinvestmentsintoGAT.Thisisapuzzleastheexistenceofa(partially)firm-financedapprenticeshiptrainingisnoteasilyreconcilablewithstandardhumancapitaltheoryaccordingtowhichprofitmaximizingfirmsshouldnotpayforgeneraltrainingbutratherfree-rideontheinvestmentofothers.Mostexplanationsintheliteraturearebasedontheideathatthetrainingfirmsareabletorecoupthecostsofthetrainingbyextractingrentsfromthegraduatedapprenticesthatstaywiththetrainingfirm.Examplesofsuchexplanationsarebasedonasymmetricinformationontheabilityoftheapprentices(AcemogluandPischke,1998),highmobilitycostsandlowgenerallabourturnoverintheGermanlabourmarket(HarhoffandKane,1997),or,moregenerally,marketimperfectionsandcompressedwagestructures(AcemogluandPischke,1999a,1999b).Inthesemodelsfirmsareabletopaypost-apprenticeshipwagesbelowmarginalproductivityandthusrecoupthecostofinvestment.Concurrently,asubstantialrelatedempiricalliteraturehasemergedthatinvestigatesvariousaspectsoftheGAT,amongthem,butnotexclusively,thequestionofwhyfirmstrain.Tworecentpapersusefirmleveldatatodirectlyestimatethedeterminantsoffirms’demandforapprentices(Dietrich,2000,FougèreandSchwerdt,2001).Otherpapershaveusedindividuallevelsurveydatatoestablishthewagereturnstoapprenticeshiptraining(Winkelmann,1996b),theeffectofpost-apprenticeshipmobilityonwagesandwagegrowth(Werwatz,1996,Winkelmann,1996a,Clark,2001,BougheasandGeorgellis,2001),thewageeffectsofmovesoutofthetrainingoccupation(Werwatz,1997,Clark,2000),andtheincidenceofunemploymentandno