高微文章job-market-signaling

-MichaelSPENCEJOBMARKETSIGNALINGAndrewMickaelSPENCE(Americaneconomist)BornNovember7,1943Montclair,NewJersey,USAFieldMicroeconomicsMarketanalysisUniversityHarvardUniversity(Ph.D.)UniversityofOxford(B.A.)PrincetonUniversity(B.A.)ContributionSignalingtheoryAwardsJohnBatesClarkMedal(1981)NobelMemorialPrizeinEconomics(2001)alongwithGeorgeA.AkerlofandJosephE.Stiglitz,fortheirmarketanalysesinsituationofasymmetryofinformationContent1.Introduction2.Hiringasaninvestmentunderuncertainty3.Applicantsignaling4.Informationalfeedbackanddefinitionofequilibrium5.Propertiesofinformationalequilibrium:anexample6.Informationalimpactofindices7.Conclusion1.IntroductionWhatisthepurposeofthisessay?•analyzethephenomenaofsignalingwhichaffectthemarket•outlineamodelinwhichsignalingisimplicitlydefined•Definethepropertiesofthesignalingequilibria,theinteractionofpotentialsignalsandtheallocativeefficiencyofthemarketWhychoosingtheJobMarketexemple?«Thisessayisaboutmarketsinwhichsignalingtakesplaceandinwhichtheprimarysignalersarerelativelynumerousandinthemarketsufficientlyinfrequentlythattheyarenotexpectedto(andthereforedonot)investinacquiringsignalingreputation»=itcorrespondexactlytotheJobMarketsituation•Alargenumberofpostulant•Whicharesufficientlyinfrequentlyinthemarketsothattheyarenotexpectedtoinvestinacquiringsignalingreputation2.HiringasaninvestmentunderuncertaintyLet’sdefinethesituationintheJobMarketThehiringdecision•Identifytheproductivecapabilitiesofanindividual•AssociatealevelawagetothisexpectedcapabilitiesProblem•Unknownbeforehiring•Evenafter,itcantaketimetoevaluateHiringaspurchasingalottery•SPENCEcomparehiringtoalotterywhere«theemployerpaysthecertainmonetaryequivalentofthelotterytotheindividualaswage»•Employersareriskneutral=Wage=individual'smarginalcontributiontothehiringorganization•=Uncertainty=aplethoraofpersonaldata=assessmentofthelottery(expectedmarginalcontribution)SPENCEintroduceadistinctioninthekindofdataINDICESvs.SIGNALS•Indicesarepersonalattributethatareimmutablyfixed,unalterableorthatdochange,butnotatthediscretionoftheindividual.RaceNationalityAgesex•Signalsarealterable,subjecttomanipulationbytheindividual.Hecaninvesttimeandmoneytochangethem.Thereisapossibilityofcontrolinsignals.education3.Applicantsignaling•Theemployerwererisk-neutral.•Foreachsetofsignalsandindicesthattheemployerconfronts,hewillhaveanexpectedmarginalproductforanindividualwhohastheseobservableattributes.Thisistakentobetheofferedwagetoapplicantswiththosecharacteristics.•Thereisnotmuchthattheapplicantcandoaboutindices.Signals,ontheotherhand,arealterableandthereforepotentiallysubjecttomanipulationbythejobapplicant.•Individualsareassumedtoselectsignalssoastomaximizethedifferencebetweenofferedwagesandsignalingcosts.ACriticalAssumption:signalingcostsarenegativelycorrelatedwithproductivity•Itismostappropriatelyviewedasaprerequisiteforanobservable,alterablecharacteristictobeapersistentlyinformativesignalinthemarket.•Signalingcosts:psychic,monetary,time….4.InformationfeedbackandthedefinitionofequilibriumInformationalfeedbackloopSelf-confirming•Thesystemwillbestationaryiftheemployerstartsoutwithconditionalprobabilisticbeliefsthatafteroneroundarenotdisconfirmedbytheincomingdatatheygenerated•Weshallrefertothebeliefsasself-confirming.ASignalingEquilibrium•Givenanofferedwageschedule,onecanthinkofthemarketasgenerating,viaindividualoptimizingdecisions,anempiricaldistributionofproductivecapabilitiesgivenobservableattributesorsignals,theemployerhassubjectivelyheldconditionalprobabilisticbeliefswithrespecttoproductility,givensignals.•thesubjectivedistributionandtheoneimplicitinthemarketmechanismareidentical,overtherangeofsignalsthattheemployeractuallyobserves5.Propertiesofinformationalequilibria:Anexample•1)Aspecificnumericalexample:GroupMarginalproduct(wage)ProportionofpopulationCostofeducationlevelyI1q1yII21-q1y/2*Forthetimebeing,indicesplaynopartSupposethattheemployerbelievesthatthereissomelevelofeducation:ifyy*thenproductivityisonewithprobabilityone;ifyy*thenproductivityistwowithprobabilityone.ExpectationA:GroupIsetsy=0if:1-02-y*GroupIIsetsy=y*if:2-y*/21-01y*2Conditionsonbehavior:SignalingEquilibrium≠WelfareEquilibriumSignalingequilibriuminwhichgroupIIisbetteroffthanno-signaling:2-y*/22-q1,thusy*2q1Ifq10.5(groupIIisaminority),thereexistsasignalingequilibriuminwhichthemembersofgroupIIimprovetheirpositionovertheno-signalingcase.Ifq10.5(groupIIisamajority),groupIIwillloseduetotheexistenceofsignaling.SignalingcaseNo-signalingcaseNetbenefitsofgroupI1q1+2(1-q1)=2-q1NetbenefitsofgroupII2-y*/22)GeneralizethisexampleGroupMarginalproduct(wage)ProportionofpopulationCostofeducationlevelyI1q1a1yII21-q1a2ySignalingequilibriuminwhichgroupIIisbetteroffthanno-signaling:Conditionony*:SohowsmallaminoritygroupIIhastobetohavethepossibilityofbenefitingfromsignalingdependsupontheratioofthemarginalsignalingcostsofthetwogroups12-a1y*,2-a2y*1thus1/a2y*1/a12-a2y*2-q1q1a1/a23)OtherequilibriainthesystemwithquitedifferentpropertiesExpectationB:•Supposethattheemployer’s