11-Agency Problems and the Theory of the Firm.pdf

《11-Agency Problems and the Theory of the Firm.pdf》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

AgencyProblemsandtheTheoryoftheFirmAuthor(s):EugeneF.FamaSource:JournalofPoliticalEconomy,Vol.88,No.2(Apr.,1980),pp.288-307Publishedby:TheUniversityofChicagoPressStableURL:http://www.jstor.org/stable/1837292.Accessed:17/10/201107:12YouruseoftheJSTORarchiveindicatesyouracceptanceoftheTerms&ConditionsofUse,availableat.http://www.jstor.org/page/info/about/policies/terms.jspJSTORisanot-for-profitservicethathelpsscholars,researchers,andstudentsdiscover,use,andbuilduponawiderangeofcontentinatrusteddigitalarchive.Weuseinformationtechnologyandtoolstoincreaseproductivityandfacilitatenewformsofscholarship.FormoreinformationaboutJSTOR,pleasecontactsupport@jstor.org.TheUniversityofChicagoPressiscollaboratingwithJSTORtodigitize,preserveandextendaccesstoJournalofPoliticalEconomy.http://www.jstor.org AgencyProblemsandtheTheoryoftheFirmEugeneF.FamaUniversityofChicagoThispaperattemptstoexplainhowtheseparationofsecurityown-ershipandcontrol,typicaloflargecorporations,canbeanefficientformofeconomicorganization.Wefirstsetasidethepresumptionthatacorporationhasownersinanymeaningfulsense.Theentre-preneurisalsolaidtorest,atleastforthepurposesofthelargemoderncorporation.Thetwofunctionsusuallyattributedtotheentrepreneur-managementandriskbearing-aretreatedasnatu-rallyseparatefactorswithinthesetofcontractscalledafirm.Thefirmisdisciplinedbycompetitionfromotherfirms,whichforcestheevolutionofdevicesforefficientlymonitoringtheperformanceoftheentireteamandofitsindividualmembers.Individualpartici-pantsinthefirm,andinparticularitsmanagers,faceboththedisciplineandopportunitiesprovidedbythemarketsfortheirser-vices,bothwithinandoutsidethefirm.Economistshavelongbeenconcernedwiththeincentiveproblemsthatarisewhendecisionmakinginafirmistheprovinceofmanagerswhoarenotthefirm'ssecurityholders.'Oneoutcomehasbeenthedevelopmentof"behavioral"and"managerial"theoriesofthefirmwhichrejecttheclassicalmodelofanentrepreneur,orowner-ThisresearchissupportedbytheNationalScienceFoundation.RogerKormendihascontributedmuch,andthecommentsofA.Alchian,S.Bhattacharya,G.Becker,F.Black,M.Blume,M.Bradley,D.Breeden,N.Gonedes,B.Horwitz,G.Jarrell,E.H.Kim,J.Long,H.Manne,W.Meckling,M.H.Miller,M.Scholes,C.Smith,G.J.Stigler,R.Watts,T.Whisler,andJ.Zimmermanaregratefullyacknowledged.Presentationsatthefinance,laboreconomics,andindustrialorganizationworkshopsoftheUniversityofChicagoandtheworkshopoftheManagerialEconomicsResearchCenteroftheUniversityofRochesterhavebeenhelpful.ThepaperislargelyanoutgrowthofdiscuionscrwithMichaelC(lnsun'JensenandMeckling(1976)quotefromAdamSmith(1776).ThemodernliteratureontheproblemdatesbackatleasttoBerleandMeans(1932).[Journal{fPoliticalEconomy,1980,vol.88,no.21?1980byTheUniversityofChicago.0022-380880/28802-0005$01.65288 THEORYOFTHEFIRM289manager,whosingle-mindedlyoperatesthefirmtomaximizeprofits,infavoroftheoriesthatfocusmoreonthemotivationsofamanagerwhocontrolsbutdoesnotownandwhohaslittleresemblancetotheclassical"economicman."ExamplesofthisapproachareBaumol(1959),Simon(1959),CyertandMarch(1963),andWilliamson(1964).Morerecentlytheliteraturehasmovedtowardtheoriesthatrejecttheclassicalmodelofthefirmbutassumeclassicalformsofeconomicbehavioronthepartofagentswithinthefirm.Thefirmisviewedasasetofcontractsamongfactorsofproduction,witheachfactormoti-vatedbyitsself-interest.Becauseofitsemphasisontheimportanceofrightsintheorganizationestablishedbycontracts,thisliteratureischaracterizedundertherubric"propertyrights."AlchianandDem-setz(1972)andJensenandMeckling(1976)arethebestexamples.TheantecedentsoftheirworkareinCoase(1937,1960).ThestrikinginsightofAlchianandDemsetz(1972)andJensenandMeckling(1976)isinviewingthefirmasasetofcontractsamongfactorsofproduction.Ineffect,thefirmisviewedasateamwhosemembersactfromself-interestbutrealizethattheirdestiniesdependtosomeextentonthesurvivaloftheteaminitscompetitionwithotherteams.Thisinsight,however,isnotcarriedfarenough.Intheclassicaltheory,theagentwhopersonifiesthefirmistheentre-preneurwhoistakentobebothmanagerandresidualriskbearer.Althoughhistitlesometimeschanges-forexample,AlchianandDemsetzcallhim"theemployer"-theentrepreneurcontinuestoplayacentralroleinthefirmoftheproperty-rightsliterature.Asaconsequence,thisliteraturefailstoexplainthelargemoderncorpo-rationinwhichcontrolofthefirmisinthehandsofmanagerswhoaremoreorlessseparatefromthefirm'ssecurityholders.Themainthesisofthispaperisthatseparationofsecurityowner-shipandcontrolcanbeexplainedasanefficientformofeconomicorganizationwithinthe"setofcontracts"perspective.Wefirstsetasidethetypicalpresumptionthatacorporationhasownersinanymeaningfulsense.Theattractiveconceptoftheentrepreneurisalsolaidtorest,atleastforthepurposesofthelargemoderncorporation.Instead,thetwofunctionsusuallyattributedtotheentrepreneur,managementandriskbearing,aretreatedasnaturallyseparatefac-torswithinthesetofcontractscalledafirm.Thefirmisdisciplinedbycompetitionfromotherfirms,whichforcestheevolutionofdevicesforefficientlymonitoringtheperformanceoftheentireteamandofitsindividualmembers.Inaddition,individualparticipantsinthefirm,andinparticularitsmanagers,faceboththedisciplineandopportunitiesprovidedbythemarketsfortheirservices,bothwithinandoutsideofthefirm. 290JOURNALOFPOLITICALECONOMYTheIrrelevanceoftheConceptofOwnershipoftheFirmTosetaframeworkfortheanalysis,letusfirstdescriberolesformanagementandriskbearinginthesetofcontractscalledafirm.Managementisatypeoflaborbutwithaspecialrole-coordinatingtheactivitiesofinputsandcarryingoutthecontractsagreedamonginputs,allofwhichcanbecharacterizedas"decisionmaking."Toexplaintheroleoftheriskbearers,assumeforthemomentthatthefirmrentsallotherfactorsofproductionandthatrentalcontractsarenegotiatedatthebeginningofeachproductionperiodwithpayoffsattheendoftheperiod.Theriskbearersthencontracttoaccepttheuncertainandpossiblynegativedifferencebetweentotalrevenuesandcostsattheendofeachproductionperiod.Whenotherfactorsofproductionarepaidattheendofeachperiod,itisnotnecessaryfortheriskbearerstoinvestanythinginthefirmatthebeginningoftheperiod.Mostcommonly,however,theriskbearersguaranteeperformanceoftheircontractsbyputtingupwealthexante,withthisfrontmoneyusedtopurchasecapitalandperhapsalsothetechnologythatthefirmusesinitsproductionactivities.Inthiswaytheriskbearingfunctioniscombinedwithownershipofcapitalandtechnology.Wealsocommonlyobservethatthejointfunctionsofriskbearingandownershipofcapitalarere-packagedandsoldindifferentproportionstodifferentgroupsofinvestors.Forexample,whenfrontmoneyisraisedbyissuingbothbondsandcommonstock,thebondsinvolveacombinationofriskbearingandownershipofcapitalwithalowamountofriskbearingrelativetothecombinationofriskbearingandownershipofcapitalinherentinthecommonstock.Unlessthebondsareriskfree,theriskbearingfunctionisinpartbornebythebondholders,andownershipofcapitalissharedbybondholdersandstockholders.However,ownershipofcapitalshouldnotbeconfusedwithowner-shipofthefirm.Eachfactorinafirmisownedbysomebody.Thefirmisjustthesetofcontractscoveringthewayinputsarejoinedtocreateoutputsandthewayreceiptsfromoutputsaresharedamonginputs.Inthis"nexusofcontracts"perspective,ownershipofthefirmisanirrelevantconcept.Dispellingthetenaciousnotionthatafirmisownedbyitssecurityholdersisimportantbecauseitisafirststeptowardunderstandingthatcontroloverafirm'sdecisionsisnotneces-sarilytheprovinceofsecurityholders.Thesecondstepissettingasidetheequallytenaciousroleinthefirmusuallyattributedtotheentre-preneur.ManagementandRiskBearing:ACloserLookTheentrepreneur(manager-riskbearer)iscentralinboththeJensen-MecklingandAlchian-Demsetzanalysesofthefirm.For THEORYOFTHEFIRM291example,Alchian-Demsetzstate:"Theessenceoftheclassicalfirmisidentifiedhereasacontractualstructurewith:1)jointinputproduc-tion;2)severalinputowners;3)onepartywhoiscommontoallthecontractsofthejointinputs;4)whohastherighttorenegotiateanyinput'scontractindependentlyofcontractswithotherinputowners;5)whoholdstheresidualclaim;and6)whohastherighttosellhiscentralcontractualresidualstatus.Thecentralagentiscalledthefirm'sownerandtheemployer"(1972,p.794).Tounderstandthemoderncorporation,itisbettertoseparatethemanager,theagentsofpoints3and4oftheAlchian-Demsetzdefini-tionofthefirm,fromtheriskbearerdescribedinpoints5and6.Therationaleforseparatingthesefunctionsisnotjustthattheendresultismoredescriptiveofthecorporation,apointrecognizedinboththeAlchian-DemsetzandJensen-Mecklingpapers.Themajorlossinre-tainingtheconceptoftheentrepreneuristhatoneispreventedfromdevelopingaperspectiveonmanagementandriskbearingasseparatefactorsofproduction,eachfacedwithamarketforitsservicesthatprovidesalternativeopportunitiesand,inthecaseofmanagement,motivationtowardperformance.Thus,anygivensetofcontracts,aparticularfirm,isincompetitionwithotherfirms,whicharelikewiseteamsofcooperatingfactorsofproduction.Ifthereisapartoftheteamthathasaspecialinterestinitsviability,itisnotobviouslytheriskbearers.Itistruethatiftheteamdoesnotproveviablefactorslikelaborandmanagementareprotectedbymarketsinwhichrightstotheirfutureservicescanbesoldorrentedtootherteams.Theriskbearers,asresidualclaimants,alsoseemtosufferthemostdirectconsequencesfromthefailingsoftheteam.However,theriskbearersinthemoderncorporationalsohavemarketsfortheirservices-capitalmarkets-whichallowthemtoshiftamongteamswithrelativelylowtransactioncostsandtohedgeagainstthefailingsofanygiventeambydiversifyingtheirholdingsacrossteams.Indeed,portfoliotheorytellsusthattheoptimalportfolioforanyinvestorislikelytobediversifiedacrossthesecuritiesofmanyfirms.2Sinceheholdsthesecuritiesofmanyfirmspreciselytoavoidhavinghiswealthdependtoomuchonanyonefirm,anindividualsecurityholdergenerallyhasnospecialinterestinpersonallyoverseeingthedetailedactivitiesofanyfirm.Inshort,efficientallocationofriskbearingseemstoimplyalargedegreeofseparationofsecurityowner-shipfromcontrolofafirm.Ontheotherhand,themanagersofafirmrentasubstantiallumpofwealth-theirhumancapital-tothefirm,andtherentalratesfor2DetaileddiscussionsofportfoliomodelscanbefoundinFamaandMiller(1972,chaps.6aind7),Jensen(1972),andFama(1976,chaps.7and8). 292JOURNALOFPOLITICALECONOMYtheirhumancapitalsignaledbythemanageriallabormarketarelikelytodependonthesuccessorfailureofthefirm.Thefunctionofmanagementistooverseethecontractsamongfactorsandtoensuretheviabilityofthefirm.Forthepurposesofthemanageriallabormarket,thepreviousassociationsofamanagerwithsuccessandfailureareinformationabouthistalents.Themanagerofafirm,likethecoachofanyteam,maynotsufferanyimmediategainorlossincurrentwagesfromthecurrentperformanceofhisteam,butthesuccessorfailureoftheteamimpactshisfuturewages,andthisgivesthemanagerastakeinthesuccessoftheteam.Thefirm'ssecurityholdersprovideimportantbutindirectassis-tancetothemanageriallabormarketinitstaskofvaluingthefirm'smanagement.Asecurityholderwantstopurchasesecuritieswithconfidencethatthepricespaidreflecttherisksheistakingandthatthesecuritieswillbepricedinthefuturetoallowhimtoreaptherewards(orpunishments)ofhisriskbearing.Thus,althoughanindi-vidualsecurityholdermaynothaveastronginterestindirectlyoverseeingthemanagementofaparticularfirm,hehasastronginterestintheexistenceofacapitalmarketwhichefficientlypricesthefirm'ssecurities.Thesignalsprovidedbyanefficientcapitalmarketaboutthevaluesofafirm'ssecuritiesarelikelytobeimportantforthemanageriallabormarket'srevaluationsofthefirm'smanagement.Wecomenowtothecentralquestion.Towhatextentcanthesignalsprovidedbythemanageriallabormarketandthecapitalmarket,perhapsalongwithothermarket-inducedmechanisms,disci-plinemanagers?Wefirstdiscuss,stillingeneralterms,thetypesofdisciplineimposedbymanageriallabormarkets,bothwithinandoutsideofthefirm.Wethenanalyzespecificconditionsunderwhichthisdisciplineissufficienttoresolvepotentialincentiveproblemsthatmightbeassociatedwiththeseparationofsecurityownershipandcontrol.TheViabilityofSeparationofSecurityOwnershipandControloftheFirm:GeneralCommentsTheoutsidemanageriallabormarketexertsmanydirectpressuresonthefirmtosortandcompensatemanagersaccordingtoperformance.Oneformofpressurecomesfromthefactthatanongoingfirmisalwaysinthemarketfornewmanagers.Potentialnewmanagersareconcernedwiththemechanicsbywhichtheirperformancewillbejudged,andtheyseekinformationabouttheresponsivenessofthesysteminrewardingperformance.Moreover,givenacompetitivemanageriallabormarket,whenthefirm'srewardsystemisnotre-sponsivetoperformancethefirmlosesmanagers,andthebestarethefirsttoleave. THEORYOFTHEFIRM293Thereisalsomuchinternalmonitoringofmanagersbymanagersthemselves.Partofthetalentofamanagerishisabilitytoelicitandmeasuretheproductivityoflowermanagers,sothereisanaturalprocessofmonitoringfromhighertolowerlevelsofmanagement.Lesswellappreciated,however,isthemonitoringthattakesplacefrombottomtotop.Lowermanagersperceivethattheycangainbysteppingovershirkingorlesscompetentmanagersabovethem.Moreover,intheteamornexusofcontractsviewofthefirm,eachmanagerisconcernedwiththeperformanceofmanagersaboveandbelowhimsincehismarginalproductislikelytobeapositivefunctionoftheirs.Finally,althoughhighermanagersareaffectedmorethanlowermanagers,allmanagersrealizethatthemanageriallabormar-ketusestheperformanceofthefirmtodetermineeachmanager'soutsideopportunitywage.Inshort,eachmanagerhasastakeintheperformanceofthemanagersaboveandbelowhimand,asaconse-quence,undertakessomeamountofmonitoringinbothdirections.Allmanagersbelowtheverytoplevelhaveaninterestinseeingthatthetopmanagerschoosepoliciesforthefirmwhichprovidethemostpositivesignalstothemanageriallabormarket.Butbywhatmecha-nismcantopmanagementbedisciplined?Sincethebodydesignatedforthisfunctionistheboardofdirectors,wecanaskhowitmightbeconstructedtodoitsjob.Aboarddominatedbysecurityholdersdoesnotseemoptimalorendowedwithgoodsurvivalproperties.Diffuseownershipofsecuritiesisbeneficialintermsofanoptimalallocationofriskbearing,butitsconsequenceisthatthefirm'ssecurityholdersaregenerallytoodiversifiedacrossthesecuritiesofmanyfirmstotakemuchdirectinterestinaparticularfirm.Ifthereiscompetitionamongthetopmanagersthemselves(allwanttobethebossofbosses),thenperhapstheyarethebestonestocontroltheboardofdirectors.Theyaremostdirectlyinthelineoffirefromlowermanagerswhenthemarketsforsecuritiesandman-ageriallaborgivepoorsignalsabouttheperformanceofthefirm.Becauseoftheirpoweroverthefirm'sdecisions,theirmarket-determinedopportunitywagesarealsolikelytobemostaffectedbymarketsignalsabouttheperformanceofthefirm.Iftheyarealsoincompetitionforthetopplacesinthefirm,theymaybethemostinformedandresponsivecriticsofthefirm'sperformance.Havinggainedcontroloftheboard,topmanagementmaydecidethatcollusionandexpropriationofsecurityholderwealtharebetterthancompetitionamongthemselves.Theprobabilityofsuchcollusivearrangementsmightbelowered,andtheviabilityoftheboardasamarket-inducedmechanismforlow-costinternaltransferofcontrolmightbeenhanced,bytheinclusionofoutsidedirectors.Thelattermightbestberegardedasprofessionalrefereeswhosetaskistostimulateandoverseethecompetitionamongthefirm'stopmana- 294JOURNALOFPOLITICALECONOMYgers.Inastateofadvancedevolutionoftheexternalmarketsthatbuttressthecorporatefirm,theoutsidedirectorsareintheirturndisciplinedbythemarketfortheirserviceswhichpricesthemac-cordingtotheirperformanceasreferees.Sincesuchasystemofseparationofsecurityownershipfromcontrolisconsistentwiththepressuresappliedbythemanageriallabormarket,andsinceitlikewiseoperatesintheinterestsofthefirm'ssecurityholders,itprobablyhasgoodsurvivalproperties.3Thisanalysisdoesnotimplythatboardsofdirectorsarelikelytobecomposedentirelyofmanagersandoutsidedirectors.Theboardisviewedasamarket-inducedinstitution,theultimateinternalmonitorofthesetofcontractscalledafirm,whosemostimportantroleistoscrutinizethehighestdecisionmakerswithinthefirm.Intheteamornexusofcontractsviewofthefirm,onecannotruleouttheevolutionofboardsofdirectorsthatcontainmanydifferentfactorsofproduc-tion(ortheirhiredrepresentatives),whosecommontraitisthattheirmarginalproductsareaffectedbythoseofthetopdecisionmakers.Ontheotherhand,onealsocannotconcludethatallsuchfactorswillnaturallyshowuponboardssincetheremaybeothermarket-inducedinstitutions,forexample,unions,thatmoreefficientlymonitormana-gersonbehalfofspecificfactors.Allonecansayisthatinacompeti-tiveenvironmentlower-costsetsofmonitoringmechanismsarelikelytosurvive.Theroleoftheboardinthisframeworkistoprovidearelativelylow-costmechanismforreplacingorreorderingtopmana-gers;lowercost,forexample,thanthemechanismprovidedbyanoutsidetakeover,although,ofcourse,theexistenceofanoutsidemarketforcontrolisanotherforcewhichhelpstosensitizetheinter-nalmanageriallabormarket.Theperspectivesuggestedhereowesmuchto,butisneverthelessdifferentfrom,existingtreatmentsofthefirminthepropertyrightsliterature.Thus,Alchian(1969)andAlchianandDemsetz(1972)commentinsightfullyonthediscipliningofmanagementthattakesplacethroughtheinsideandoutsidemarketsformanagers.However,theyattributethetaskofdiscipliningmanagementprimarilytotheriskbearers,thefirm'ssecurityholders,whoareassistedtosomeextentbymanageriallabormarketsandbythepossibilityofoutsidetakeover.JensenandMeckling(1976)likewisemakecontrolofman-3WattsandZimmermian(1978)provideasimilardescriptionofthemarket-inducedevolutionof"independent"outsideauditorswhosefunctionistocertifyand,asaconsequence,stimutilatetheviabilityofthesetofcontractscalledthefirm.Liketheoutsidedirectors,theoutsideauditorsarepolicedbythemarketfortheirserviceswhichpricestheminlargepartonthebasisofhowwelltheyresistpervertingtheinterestsofonesetoffactors(e.g.,securityholders)tothebenefitofotherfactors(e.g.,manage-ment).Liketheprofessionaloutsidedirector,thewelfareoftheoutsideauditorde-pendslargelyon"reputation. THEORYOFTHEFIRM295agementtheprovinceofthefirm'sriskbearers,buttheydonotallowforanyassistancefromthemanageriallabormarket.Ofalltheau-thorsintheproperty-rightsliterature,Manne(1965,1967)ismostconcernedwiththemarketforcorporatecontrol.Herecognizesthatwithdiffusesecurityownershipmanagementandriskbearingarenaturallyseparatefunctions.Butforhim,discipliningmanagementisan"entrepreneurialjob"whichinthefirstinstancefallsonafirm'sorganizersandlateronspecialistsintheprocessofoutsidetakeover.Whenmanagementandriskbearingareviewedasnaturallysepa-ratefactorsofproduction,lookingatthemarketforriskbearingfromtheviewpointofportfoliotheorytellsusthatriskbearersarelikelytospreadtheirwealthacrossmanyfirmsandsonotbeinterestedindirectlycontrollingthemanagementofanyindividualfirm.Thus,modelsofthefirm,likethoseofAlchian-DemsetzandJensen-Meckling,inwhichthecontrolofmanagementfallsprimarilyontheriskbearers,arenotlikelytoallaythefearsofthoseconcernedwiththeapparentincentiveproblemscreatedbytheseparationofsecurityownershipandcontrol.Likewise,Manne'sapproach,inwhichthecontrolofmanagementreliesprimarilyontheexpensivemechanismof-anoutsidetakeover,offerslittlecomfort.Theviabilityofthelargecorporationwithdiffusesecurityownershipisbetterexplainedintermsofamodelwheretheprimarydiscipliningofmanagerscomesthroughmanageriallabormarkets,bothwithinandoutsideofthefirm,withassistancefromthepanoplyofinternalandexternalmonitoringdevicesthatevolvetostimulatetheongoingefficiencyofthecorporateform,andwiththemarketforoutsidetakeoverspro-vidingdisciplineoflastresort.TheViabilityofSeparationofSecurityOwnershipandControl:DetailsTheprecedingisageneraldiscussionofhowpressurefrommanage-riallabormarketshelpstodisciplinemanagers.Wenowexaminesomewhatmorespecificallyconditionsunderwhichthedisciplineimposedbymanageriallabormarketscanresolvepotentialincentiveproblemsassociatedwiththeseparationofsecurityownershipandcontrolofthefirm.Tofocusontheproblemwearetryingtosolve,letusfirstexaminethesituationwherethemanagerisalsothefirm'ssolesecurityholder,sothatthereisclearlynoincentiveproblem.Whenheissolesecurityholder,amanagerconsumesonthejob,throughshirking,perqui-sites,orincompetence,tothepointwheretheseyieldmarginalex-pectedutilityequaltothatprovidedbyanadditionaldollarofwealthusableforconsumptionorinvestmentoutsideofthefirm.Theman- 296JOURNALOFPOLITICALECONOMYagerisinducedtomakethisspecificdecisionbecausehepaysdirectlyforconsumptiononthejob;thatis,asmanagerhecannotavoidafullexpostsettlingupwithhimselfassecurityholder.Incontrast,whenthemanagerisnolongersolesecurityholder,andintheabsenceofsomeformoffullexpostsettlingupfordeviationsfromcontract,amanagerhasanincentivetoconsumemoreonthejobthanisagreedinhiscontract.Themanagerperceivesthat,onanexpostbasis,hecanbeatthegamebyshirkingorcon-sumingmoreperquisitesthanpreviouslyagreed.Thisdoesnotneces-sarilymeanthatthemanagerprofitsattheexpenseofotherfactors.Rationalmanageriallabormarketsunderstandanyshortcomingsofavailablemechanismsforenforcingexpostsettlingup.Assessmentsofexpostdeviationsfromcontractwillbeincorporatedintocontractsonanexantebasis;forexample,throughanadjustmentofthemanager'swage.Nevertheless,agamewhichisfaironanexantebasisdoesnotinducethesamebehaviorasagameinwhichthereisalsoexpostsettlingup.Hereinliethepotentiallossesfromseparationofsecurityownershipandcontrolofafirm.Therearesituationswhere,withlessthancompleteexpostsettlingup,themanagerisinducedtoconsumemoreonthejobthanhewouldlike,giventhatonaveragehepaysforhisconsumptionexante.Threegeneralconditionssufficetomakethewagerevaluationim-posedbythemanageriallabormarketaformoffullexpostsettlingupwhichresolvesthemanagerialincentiveproblemdescribedabove.Thefirstconditionisthatamanager'stalentsandhistastesforconsumptiononthejobarenotknownwithcertainty,arelikelytochangethroughtime,andmustbeimputedbymanageriallabormarketsatleastinpartfrominformationaboutthemanager'scurrentandpastperformance.Sinceitseemstocapturetheessenceofthetaskofmanageriallabormarketsinaworldofuncertainty,thisassumptionisnorealrestriction.Thesecondassumptionisthatmanageriallabormarketsappro-priatelyusecurrentandpastinformationtorevisefuturewagesandunderstandanyenforcementpowerinherentinthewagerevisionprocess.Inshort,contrarytomuchoftheliteratureonseparationofsecurityownershipandcontrol,weimputeefficiencyorrationalityininformationprocessingtomanageriallabormarkets.Indefenseofthisassumption,wenotethattheproblemfacedbymanageriallabormarketsinrevaluingthemanagersofafirmismuchentwinedwiththeproblemfacedbythecapitalmarketinrevaluingthefirmitself.Althoughwedonotunderstandallthedetailsoftheprocess,availableempiricalevidence(e.g.,Fama1976,chaps.5and6)suggeststhatthecapitalmarketgenerallymakesrationalassessmentsofthevalueof THEORYOFTHEFIRM297thefirminthefaceofimpreciseanduncertaininformation.Thisdoesnotnecessarilymeanthatinformationprocessinginmanageriallabormarketsisequallyefficientorrational,butitisawarningagainststrongpresumptionstothecontrary.Thefinalandkeyconditionforfullcontrolofmanagerialbehaviorthroughwagechangesisthattheweightofthewagerevisionprocessissufficienttoresolveanypotentialproblemswithmanagerialincen-tives.Inthisgeneralform,theconditionamountstoassumingthedesiredresult.Moresubstanceisprovidedbyspecificexamples.Example1:MarketableHumanCapitalSupposeamanager'shumancapital,hisstreamoffuturewages,isamarketableasset.Supposethemanagerperceivesthat,becauseoftheconsequentrevaluationsoffuturewages,thecurrentvalueofhishumancapitalchangesbyatleasttheamountofanunbiasedassess-mentofthewealthchangesexperiencedbyotherfactors,primarilythesecurityholders,becauseofhiscurrentdeviationsfromcontract.Then,aslongasthemanagerisnotariskpreferrer,theserevalua-tionsofhishumancapitalareaformoffullexpostsettlingup.Themanagerneednotbechargedexanteforpresumedexpostdevia-tionsfromcontractsincetheweightofthewagerevisionprocessissufficienttoneutralizehisincentivestodeviate.Itisimportanttoconsiderwhythemanagermightperceivethatthevalueofhishumancapitalchangesbyatleasttheamountofanunbiasedassessmentofthewealthchangesexperiencedbyotherfactorsduetohisdeviationsfromcontract.Notefirstthatthemarket'sassessmentofsuchwealthchangesisalsoitsassessmentofthedif-ferencebetweenthemanager'sexpostmarginalproductandthemarginalproducthecontractedtodeliverexante.However,anyassessmentofthemanager'smarginalproductislikelytoincludeextraneousnoisewhichhaslittletodowithhistalentsandefforts.Withoutspecificdetailsonwhatthemarkettakestobethestatisticalprocessgoverningtheevolutionofthemanager'stalentsandhistastesforconsumptiononthejob,onecannotsayexactlyhowfaritwillgoinadjustinghisfuturewagestoreflectitsmostrecentmeasurementofhismarginalproduct.Assumingthemarketusesinformationration-ally,theadjustmentisclosertocompletethelargerthesignalinthemostrecentmeasurementrelativetothenoise,butaslongasthereissomenoiseintheprocess,theadjustmentislessthancomplete.4Althoughhisnextwagemaynotadjustbythefullamountofanunbiasedassessmentofthecurrentcostofhisdeviationsfromcon-'Specificillustrationsofthispointareprovidedlater. 298JOURNAL,OFPOLITICALECONOMYtract,amanagerwithamultiperiodhorizonmayperceivethattheimpliedcurrentwealthchange,thepresentvalueoflikelychangesinthestreamoffuturewages,isatleastasgreatasthecostofhisdeviationsfromcontract.Inthiscase,thecontemporaneouschangeinhiswealthimpliedbyaneventualadjustmentoffuturewagesisaformoffullexpostsettlingupwhichresultsinfullenforcementofhiscontract.Moreover,thewagerevisionprocessresolvesanypotentialproblemsaboutamanager'sincentiveseventhoughtheimpliedexpostsettlingupneednotinvolvethefirmcurrentlyemployingthemanager;thatis,lowerorhigherfuturewagesduetocurrentdevia-tionsfromcontractmaycomefromotherfirms.Ofcourse,changesinamanager'swealthasaconsequenceofanticipatedfuturewagerevisionsarenotalwaysequivalenttofullexpostsettlingup.Whenamanagerdoesnotexpecttobeinthelabormarketformanyfutureperiods,theweightoffuturewagerevisionsduetocurrentassessmentsofperformancemayamounttosubstan-tiallylessthanfullexpostsettlingup.However,itisjustasimportanttorecognizethattheweightofanticipationsaboutfuturewagesmayamounttomorethanfullexpostsettlingup.Theremaybesituationswherethepersonalwealthchangeperceivedbythemanagerasaconsequenceofdeviationsfromcontractisgreaterthanthewealthchangeexperiencedbyotherfactors.Sincemanyreadershavehadtroublewiththispoint,itiswelltobringitclosertohome.Economists(especiallyyoungeconomists)easilyimaginesituationswheretheeffectsofhigherorlowerqualityofacurrentarticleorbookonthemarketvalueofhumancapital,throughenhancementorloweringof"reputation,"areinexcessoftheeffectsofqualitydif-ferencesonthemarketvalueofthespecificworktoanypublisher.Managerscansometimeshavesimilarperceptionswithrespecttotheimplicationsofcurrentperformanceforthemarketvalueoftheirhumancapital.Example2:StochasticProcessesforMarginalProductsThenextexampleofexpostsettlingupthroughthewagerevisionprocessissomewhatmoreformalthanthatdescribedabove.Wemakespecificassumptionsaboutthestochasticevolutionofamanager'smeasuredmarginalproductandabouthowthemanageriallabormarketusesinformationfromtheprocesstoadjustthemanager'sfuturewages-inamannerwhichamountstoprecise,fullexpostsettlingupfortheresultsofpastperformance.Supposethemanager'smeasuredmarginalproductforanyperiodtiscomposedoftwoterms:(i)anexpectedvalue,givenhistalents,effortexertedduringt,consumptionofperquisites,etc.;and(ii) THEORYOFTHEFIRM299randomnoise.Therandomnoisemayinpartresultfrommeasure-menterror,thatis,thesheerdifficultyofaccuratelymeasuringmar-ginalproductswhenthereisteamproduction,butitmayalsoariseinpartfromthefactthateffortexertedandtalentdonotyieldperfectlycertainconsequences.Moreover,becauseoftheuncertainevolutionofthemanager'stalentsandtastes,theexpectedvalueofhismarginalproductisitselfastochasticprocess.Specifically,weassumethattheexpectedvalue,_z,followsarandomwalkwithstepsthatareinde-pendentoftherandomnoise,Et,inthemanager'smeasuredmarginalproduct,zt.Thus,themeasuredmarginalproduct,Zt=Zt+Et,(1)isarandomwalkpluswhitenoise.Forsimplicity,wealsoassumethatthisprocessdescribesthemanager'smarginalproductbothinhiscurrentemploymentandinthebestalternativeemployment.Thecharacteristics(parameters)oftheevolutionofthemanager'smarginalproductdependtosomeextentonendogenousvariableslikeeffortandperquisitesconsumed,whicharenotcompletelyobserv-able.Ourpurposeistosetupthemanageriallabormarketsothatthewagerevisionprocessresolvesanypotentialincentiveproblemsthatmayarisefromtheendogeneityofztinasituationwherethereisseparationofsecurityownershipandcontrolofthefirm.Supposenextthatriskbearersareallriskneutralandthat1-periodmarketinterestratesarealwaysequaltozero.Supposealsothatmanagerialwagecontractsarewrittensothatthemanager'swageinanyperiodtistheexpectedvalueofhismarginalproduct,zt,condi-tionalonpastmeasuredvaluesofhismarginalproduct,withtheriskbearersacceptingthenoiseEt,intheexpostmeasurementofthemarginalproduct.Weshallseebelowthatthisisanoptimalarrange-mentforourrisk-neutralriskbearers.However,itisnotnecessarilyoptimalforthemanagerifheisriskaverse.Arisk-aversemanagermaywanttosellpartoftheriskinherentintheuncertainevolutionofhisexpectedmarginalproducttotheriskbearers,forexample,throughalong-termwagecontract.Weavoidthisissuebyassumingthat,perhapsbecauseofthemoreextrememoralhazardproblemsinlong-termcontracts(rememberthatizisinpartunderthecontrolofthemanager)andthecontractingcoststowhichthesemoralhazardproblemsgiverise,simplecontractsinwhichthemanager'swageisresetatthebeginningofeachperiodaredominant,atleastforsomenontrivialsubsetoffirmsandmana-gers.'Ifwecouldalsoassumeawayanyremainingmoralhazard5Institutionslikecorporations,thataresubjecttorapidtechnologicalchangewithalargedegreeofuncertaintyaboutfuturemanagerialneeds,mayfindthatlong-term 300JOURNALOFPOLITICALECONOMY(managerialincentive)problems,thenwithrisk-aversemanagers,risk-neutralriskbearers,andthepresumedfixedrecontractingpe-riod,thecontractwhichspecifiesexantethatthemanagerwillbepaidthecurrentexpectedvalueofhismarginalproductdominatesanycontractwherethemanageralsosharestheexpostdeviationofhismeasuredmarginalproductfromitsexanteexpectedvalue(see,e.g.,SpenceandZeckhauser1971).However,contractswhichspecifyexantethatthemanagerwillbepaidthecurrentexpectedvalueofhismarginalproductseemtoleavethetypicalmoralhazardproblemthatariseswhenthereislessthancompleteexpostenforcementofcontracts.ThenoiseEtintheman-agersmarginalproductisbornebytheriskbearers.Oncetheman-ager'sexpectedmarginalproductZt(=hiscurrentwage)hasbeenassessed,heseemstohaveanincentivetoconsumemoreperquisitesandprovidelesseffortthanareimpliedinzt.Amechanismforexpostenforcementis,however,builtintothemodel.Withtheexpectedvalueofthemanager'smarginalproductwanderingrandomlythroughtime,futureassessmentsofexpectedmarginalproducts(andthusofwages)willbedeterminedinpartbyEt,thedeviationofthecurrentmeasuredmarginalproductfromitsexanteexpectedvalue.Inthepresentscenario,whereztisassumedtofollowarandomwalk,Muth(1960)hasshownthattheexpectedvalueofthemarginalproductevolvesaccordingtoZt=Zt.l+(1-))Et-1,(2)wheretheparameter4(0<4<1)isclosertozerothesmallerthevarianceofthenoiseterminthemarginalproductequation(1)relativetothevarianceofthestepsintherandomwalkfollowedbytheexpectedmarginalproduct.Infact,theprocessbywhichfutureexpectedmarginalproductsareadjustedonthebasisofpastdeviationsofmarginalproductsfromtheirexpectedvaluesleadstoapreciseformoffullexpostsettlingup.Thisisbestseenbywritingthemarginalproductztinitsinvertedform,thatis,intermsofpastmarginalproductsandthecurrentnoise.Theinvertedformforourmodel,arandomwalkembeddedinrandomnoise,isZt-(I-)Zt-l+(1-))Zt-2+4)2(1-4)Zt-3+.+Et,(3)managerialcontractscanonlybenegotiatedathighcost.Ontheotherhand,institu-tionslikegovernments,schools,anduniversitiesmaybeabletoforecastmorereliablytheirfutureneedsformanagers(andotherprofessionals)andsomaybeabletoofferlong-termcontractsatrelativelylowcost.Theseinstitutionscanthenbeexpectetitoattracttherelativelyrisk-aversemembersoftheprofessionallaborforce,whiletheriskieremploymentofferedbycorporationsattractsthosewhoarewillingtoacceptshorter-termcontracts. THEORYOFTHEFIRM301sothatZt=(1-)Zt-1+0(1-)Zt-2+42(1P-)Zt-3+...(4)(see,e.g.,Nelson1973,chap.4,orMuth1960).Forourpurposes,theinterestingfactisthat,althoughheispaidhisexanteexpectedmarginalproduct,themanagerdoesnotgettoavoidhisexpostmarginalproduct.Forexample,wecaninferfrom(4)thatZt-,hasweight1-/inZt;thenithasweightk(1-/)inZt+,,02(1-inZt+2,andsoon.Intheend,thesumofthecontributionsofZt,tofutureexpectedmarginalproducts,andthustofuturewages,isexactlyZti.Withzerointerestrates,thismeansthattheriskbearerssimplyallowthemanagertosmoothhismarginalproductacrossfutureperiodsatthegoingopportunitycostofallsuchtemporalwealthtransfers.Asaconsequence,themanagerhasnoincentivetotrytoburyshirkingorconsumptionofperquisitesinhisexpostmeasuredmarginalproduct.Sincethemanageriallabormarketispresumedtounderstandtheweightofthewagerevisionprocess,whichinthiscaseamountstoprecisefullexpostsettlingup,anypotentialmanagerialincentiveproblemsintheseparationofriskbearing,orsecurityownership,fromcontrolareresolved.Themanagercancontractforandtakeanoptimalamountofconsumptiononthejob.Thewagesetexanteneednotincludeanyallowanceforexpostincentivestodeviatefromthecontractsincethewagerevisionprocessneutralizesanysuchincentives.Note,moreover,thatthevalueof(Ainthewagerevisionprocessdescribedby(4)determineshowtheobservedmarginalproductofanygivenperiodissubdividedandspreadacrossfutureperiods,butwhateverthevalueof4,thegivenmarginalproductisfullyaccountedforinthestreamoffuturewages.Thus,itisnowclearwhatwasmeantbytheearlierclaimthatalthoughtheparameter4intheprocessgeneratingthemanager'smarginalproductistosomeextentunderhiscontrol,thisisnotamatterofparticularconcerntothemanageriallabormarket.Asomewhatevidentqualificationisinorder.Thesmoothingpro-cessdescribedby(4)containsaninfinitenumberofterms,whereasanymanagerhasafiniteworkinglife.Forpracticalpurposes,fullexpostsettlingupisachievedaslongasthemanager'scurrentmarginalproductis"verynearly"fullyabsorbedbythestreamofwagesoverhisfutureworkinglife.Thisrequiresavalueof4in(4)whichissufficientlyfarfrom1.0,giventhenumberofperiodsremaininginthemanager'sworkinglife.Recallthat4iscloserto1.0thelargerthevarianceofthenoiseinthemanager'smeasuredmarginalproductrelativetothevarianceofthestepsoftherandomwalktakenbytheexpectedvalueofhismarginalproduct.Intuitively,whenthevariance 302'JOURNALOFPOLITICALECONOMYofthenoisetermislargerelativetothatofthechangesintheexpectedvalue,thecurrentmeasuredmarginalproducthasaweaksignalaboutanychangeintheexpectedvalueofthemarginalprod-uct,andthecurrentmarginalproductisonlyallocatedslowlytoexpectedfuturemarginalproducts.SomeExtensionsHavingqualifiedtheanalysis,letusnowindicatesomewaysinwhichitisrobusttochangesindetailsofthemodel.1.MoreComplicatedModelsfortheManager'sMarginalProductThecriticalingredientinenforcingprecisefullexpostsettlingupthroughwagerevisionsonthebasisofreassessmentsofexpectedmarginalproductsisthatwhenthemarginalproductanditsexpectedvalueareexpressedininvertedform,asin(3)and(4),thesumoftheweightsonpastmarginalproductsisexactly1.0.Thiswillbethecase(see,e.g.,Nelson1973,chap.4)wheneverthemanager'smarginalproductconformstoanonstationarystochasticprocess,butthechangesfromperiodtoperiodinthemarginalproductconformtosomestationaryARMA(mixedautoregressivemovingaverage)pro-cess.Theexamplesummarizedinequations(1)-(4)istheinterestingbutspecialcasewheretheexpectedmarginalproductfollowsaran-domwalksothatthedifferencesofthemarginalproductareastationary,first-ordermovingaverageprocess.Thegeneralcaseal-lowstheexpectedvalueofthemarginalproducttofollowanymorecomplicatednonstationaryprocesswhichhasthepropertythatthedifferencesofthemarginalproductarestationary,sothatthemargi-nalproductanditsexpectedvaluecanbeexpressedininvertedformasZt=1Zt-I+1T2Zt-2+**+Et(5)Zt=1TZt-I+72Zt-2+***(6)withi1.(7)Thesecanbeviewedasthegeneralconditionsforenforcingprecisefullexpostsettlingthroughthewagerevisionprocesswhenthe THEORYOFTHEFIRM303manager'swageisequaltothecurrentexpectedvalueofhismarginalproduct.62.Risk-AverseRiskBearersIntheframeworksummarizedinequations(5)-(7),ifthemanagerswitchesfirms,theriskbearersofhisformerfirmareleftwiththeremainsofhismeasuredmarginalproductsnotpreviouslyabsorbedintotheexpectedvalueofhismarginalproduct.Nevertheless,inthewaywehavesetuptheworld,theriskbearersrealizethattheman-ager'snextfirmcontinuestosethiswageaccordingtothesamestochasticprocessasthelastfirm.Sincethisresultsinfullexpostsettlinguponthepartofthemanager,themotiveforswitchingfirmscannotbetoavoidperverseadjustmentsoffuturewagesonthebasisofpastperformance.Onaverage,theswitchingofmanagersamongfirmsdoesnotresultingainsorlossestoriskbearers,whichmeansthattheswitchesareamatterofindifferencetoourpresumedrisk-neutralriskbearers.Itis,however,interestingtoexaminehowtheanalysismightchangewhentheriskbearersareriskaverseandswitchingofmana-gersamongfirmsisnotamatterofindifference.Suppose,forthemoment,thattheriskbearersoffermanagerscontractswhere,asbefore,themanager'swagetrackstheexpectedvalueofhismarginalproduct,buteachperiodthereisalsoafixeddiscountinthewagetocompensatetheriskbearersfortherisksofunfinishedexpostsettlingupwiththefirmasaconsequenceofapossiblefutureshiftbythemanagertoanotherfirm.Suchanarrangementmaysatisfytheriskbearers,butitwillnotbeacceptabletothemanager.Aslongashismarginalproductevolvesaccordingtoequations(5)-(7),bothinhiscurrentfirmandinthebestalternative,themanagerissubjecttofullexpostsettlingup.Thus,anyriskadjustmentofhiswagetoreflectthefactthatthesettlingupmaynotbewithhiscurrentfirmisanuncompensatedlosswhichhewillendeavortoavoid.Themanagercanavoidanyriskdiscountinhiswage,andmaintaincompletefreedomtoswitchamongfirms,byhimselfbearingalltheriskofhismarginalproduct;thatis,hecontractstoaccept,attheendofeachperiod,hisexpostmeasuredmarginalproductratherthanitsexanteexpectedvaluesothatthereis,periodbyperiod,fullexpostsettlingupwithhiscurrentfirm.Thereissuchapresumptionagainst6WhenYtfollowsastationaryprocess,thelong-runaveragevaluetowardwhichtheprocessalwaystendswilleventuallybeknownwithnearperfectcertainty.Thus,thecaseofastationaryexpectedmarginalproductisoflittleinterest,atleastforthepurposesofexpostsettlingupenforcedbythewagerevisionprocess. 304JOURNALOFPOLITICALECONOMYtheoptimalityofimmediate,fullexpostsettlingupintheliteratureonoptimalcontractingthatitbehoovesustoexaminehowandwhyitworks,andisoptimal,inthecircumstancesunderexamination.ContractualSettlingUpTheliteratureonoptimalcontracting,forexample,HarrisandRaviv(1978,1979),Holmstrom(1979),andShavell(1979),suggestsuni-formlythatwhenthereisnoiseinthemanager'smarginalproduct,thatis,whenthedeviationofmeasuredmarginalproductfromitsexpectedvaluecannotbetracedunambiguouslyandcostlesslytothemanager'sactions(talents,effortexerted,andconsumptiononthejob),thenarisk-aversemanagerwillalwayschoosetosharepartoftheuncertaintyintheevaluationofhisperformancewiththefirm'sriskbearers.Hewillagreetosomeamountofexpostsettlingup,butalwayslessthan100percentofthedeviationofhismeasuredmargi-nalproductfromitsexanteexpectedvalue.Inshort,thecontractingmodelssuggestthatwemustlearntolivewiththeincentiveproblemsthatarisewhenthereislessthancompleteexpostenforcementofcontracts.Thecontractingliteratureisalmostuniformlyconcernedwith1-periodmodels.Ina1-periodworld,therecanbenoenforcementofcontractsthroughawagerevisionprocessimposedbythemanageriallabormarket.Theexistenceofthisformofexpostsettlingupinamultiperiodworldaffectsthemanager'swillingnesstoengageinexplicitcontractualexpostsettlingup.Forexample,inthemodelsummarizedinequations(5)-(7),themanager'swageinanyperiodistheexpectedvalueofhismarginalproductassessedatthebeginningoftheperiod,andthemanagerdoesnotimmediatelyshareanyofthedeviationofhisexpostmargi-nalproductfromitsexanteexpectedvalue.However,becauseitcontainsinformationaboutfutureexpectedvaluesofhismarginalproduct,eventuallythemanager'scurrentmeasuredmarginalprod-uctisallocatedinfulltofutureexpectedmarginalproducts.Equiva-lently,inthewagerevisionprocessdescribedbyequations(5)-(7),themanageriallabormarketineffectactsasafinancialintermediary.Itwithdrawsportionsofpastaccumulatedmeasuredmarginalproductstopaythemanageradividendonhishumancapitalequaltotheexpectedvalueofhismarginalproduct,andimplicitlyprovidesthelendingarrangementswhichallowthemanagertospreadhiscurrentmeasuredmarginalproductoverfutureperiodsinpreciselythewaythecurrentmarginalproductwillcontributetoexpectedfuturemar-ginalproducts. THEORYOFTHEFIRM305Lookedatfromthisperspective,however,themanagermightsim-plycontracttotaketheexpostmeasuredvalueofhismarginalproductashiswageandthenhimselfusethecapitalmarkettosmoothhismeasuredmarginalproductoverfutureperiods.Sincethesameasset(hishumancapital)isinvolved,themanagershouldbeabletocarryoutthesesmoothingtransactionsviathecapitalmarketonthesametermsascanbehadinthemanageriallabormarket.Theadvantagetothemanagerinsmoothingthroughthecapitalmarket,however,isthathecanthencontracttoacceptfullexpostsettlingupperiodbyperiod(heispaidhismeasuredmarginalproduct),whichmeanshecanavoidanyriskdiscountinhiswagethatmightbeimposedwhenheispaidtheexpectedvalueofhismarginalproductwiththepossibilityofunanticipatedswitchestootherfirms.7Itisimportanttorecognizethatusingthecapitalmarketinthemannerdescribedaboveallowsthemanagerto"averageout"therandomnoiseinhismeasuredmarginalproduct.Thus,whenheisinsteadpaidtheexpectedvalueofhismarginalproducteachperiod,andwhentheprocessgeneratinghismarginalproductisdescribedbyequations(5)-(7),themanager'scurrentmeasuredmarginalproductiseventuallyallocatedinfulltofutureexpectedmarginalproducts.Thishappily,butonlycoincidentally,resolvesincentiveproblemsbyimposingfullexpostsettlingup.Theallocationofthecurrentmargi-nalproducttofutureexpectedmarginalproductsinfactoccursbecausethecurrentmarginalproducthasinformationaboutfutureexpectedmarginalproducts.Theweightsiriinequations(5)-(7)arepreciselythosethatoptimallyextractthisinformationandsoopti-mallysmoothoraverageoutthepurelyrandomnoiseintheman-ager'smeasuredmarginalproduct.Themanagercanachievethesameresultbycontractingtobepaidthemeasuredvalueofhismarginalproductandthenusingthecapitalmarkettosmoothhismarginalproduct.Thispowerofthecapitalmarkettoreducetheterrorinfullcontractualexpostsettlingupislostinthe1-periodmodelsthatdominatethecontractingliterature.'Withpositiveinterestrates,contractingtobepaidhismeasuredmarginalproductandthenusingthecapitalmarkettosmooththemarginalproductoverfutureperiodsdominatesthecontractinwhichthemanagerispaidtheexpectedvalueofhismarginalproduct.Equivalencecanberestoredbyadjustingtheexpectedmarginalproductx,illeq.(6)foraccumulatedinterestonthepastmarginalproducts,Zt-,Zt-2,...,orbyprepayingthepresentvalueofinterestonthedeferralsofthecurrentmarginalproductoverfutureperiods.Sufficeittosay,however,thateitheraccumulationorprepaymentofinterestcomplicatestheproblemsposedbypossibleshiftsoftheman-agertootherfirmsandsomayleanthesystemtowardcontractsinwhichthemanagerispaidhismeasuredmarginalproductandthenusesthecapitalmarkettoachieveoptimalsmoothing. 3o6JOURNALOFPOLITICALECONOMYConclusionsThemodelsummarizedbyequations(5)-(7)isonespecificscenarioinwhichthewagerevisionprocessimposedbythemanageriallabormarketamountstofullexpostsettlingupbythemanagerforhispastperformance.Theimportantgeneralpointisthatinanyscenariowheretheweightofthewagerevisionprocessisatleastequivalenttofullexpostsettlingup,managerialincentiveproblems-theproblemsusuallyattributedtotheseparationofsecurityownershipandcontrolofthefirm-areresolved.Noclaimismadethatthewagerevisionprocessalwaysresultsinafullexpostsettlinguponthepartofthemanager.Therearecertainlysituationswheretheweightofanticipatedfuturewagechangesisinsufficienttocounterbalancethegainstobehadfromexpostshirk-ing,orperhapsoutrighttheft,inexcessofwhatwasagreedexanteinamanager'scontract.Ontheotherhand,precisefullexpostsettlingupisnotanupperboundontheforceofthewagerevisionprocess.Therearecertainlysituationswhere,asaconsequenceofanticipatedfuturewagechanges,amanagerperceivesthatthevalueofhishumancapitalchangesbymorethanthewealthchangesimposedonotherfactors,andespeciallythefirm'ssecurityholders,byhiscurrentde-viationsfromthetermsofhiscontract.Theextenttowhichthewagerevisionprocessimposesexpostsettlingupinanyparticularsituationis,ofcourse,anempiricalissue.Butitisprobablysafetosaythatthegeneralphenomenonisatleastoneoftheingredientsinthesurvivalofthemodernlargecorpora-tion,characterizedbydiffusesecurityownershipandtheseparationofsecurityownershipandcontrol,asaviableformofeconomicorganization.ReferencesAlchian,AimenA."CorporateManagementandPropertyRights."InEco-tiomicPolicyatdtheReglkatiotnofCorporateSecirities,editedbyHenryG.Manne.Washington:AmericanEnterpriseInst.PublicPolicyRes.,1969.Alchian,ArmenA.,andDemsetz,Harold."Production,InformationCosts,andEconomicOrganization."A.E.R.62(December1972):777-95.Bauimol,WilliamJ.BisintessBehavior,ValbeatndGrowth.NewYork:Macmillan,1959.Berle,AdolphA.,Jr.,andMeans,GardinerC.TheModertiCorporatiotiandPrivateProperty.NewYork:Macmillan,1932.Coase,RonaldH."rTheNatureoftheFirm."Ecotiomtica,n.s.4(November1937):386-405.."TheProblemofSocialCost."J.LawantdEcon.3(October1960):1-44.Cvert,RichardM.,andMarch,JamnesG.ABehavioralTheorvoJtheFirmP.EnglewoodCliffs,N.J.:Prentice-Hall,1963. THEORYOFTHEFIRM307Fama,EugeneF.FoundationsofFinance.NewYork:Basic,1976.Fama,EugeneF.,andMiller,MertonH.TheTheoryofFinance.NewYork:Holt,Rinehart&Winston,1972.Harris,Milton,andRaviv,Artur."SomeResultsonIncentiveContractswithApplicationstoEducationandEmployment,HealthInsurance,andLawEnforcement."A.E.R.68(March1978):20-30.."OptimalIncentiveContractswithImperfectInformation."WorkingPaperno.70-75-76,Carnegie-MellonUniv.,GraduateSchoolofIndus.Admin.,April1976(rev.January1979),forthcominginj.Ecoti.Theorv.Holmstrom,Bengt."MoralHazardandObservability."BellJ.Econ.10(Spring1979):74-91.Jensen,MichaelC."CapitalMarkets:TheoryandEvidence."BellJ.Econl.alndMVIaniagemenitSci.3(Autumn1972):357-98.Jensen,MichaelC.,andMeckling,WilliamH."TheoryoftheFirm:Manage-rialBehavior,AgencyCostsandOwnershipStructure."J.FinancialEcoti.3(October1976):305-60.Manne,HenryG."MergersandtheMarketforCorporateControl."J.P.E.73,no.2(April1965):110-20."OurTwoCorporateSystems:LawandEconomics."VirginiaLawRev.53(March1967):259-85.Muth,JohnF."OptimalPropertiesofExponentiallyWeightedForecasts."J.AittericantStatis.Assoc.55(June1960):299-306.Nelson,CharlesR.AppliedTimeSeriesAnialysisforMatnagerialForecastitng.SanFrancisco:Holden-Day,1973.Shavell,Steven."RiskSharingandIncentivesinthePrincipalandAgentRelationship."BellJ.Ecotn.10(Spring1979):55-73.Simon,HerbertA."TheoriesofDecisionMakinginEconomicsandBehav-ioralScience."A.E.R.49(June1959):253-83.Smith,Adam.TheWealthofNatiotns.1776.Cannaned.NewYork:ModernLibrary,1937.Spence,Michael,andZeckhauser,Richard."Insurance,InformationandIndividualAction."A.E.R.61(May1971):380-87.Watts,RossL.,andZimmerman,Jerold."AuditorsandtheDeterminationofAccountingStandards,anAnalysisoftheLackofIndependence."WorkingPaperGPB7806,Unix'.Rochester,GraduateSchoolofManagement,1978.Williamson,OliverE.TheEconiomicsofDiscretiotnaryBehavior:ManagerialObjec-tivesitaTheoryoftheFirm.EnglewoodCliffs,N.J.:Prentice-Hall,1964.