拉丰和Tirole的产业经济学文章经典

(s):Jean-JacquesLaffontandJeanTiroleSource:TheJournalofPoliticalEconomy,Vol.94,No.3,Part1,(Jun.,1986),pp.614-641Publishedby:TheUniversityofChicagoPressStableURL::14/07/200806:29YouruseoftheJSTORarchiveindicatesyouracceptanceofJSTOR'sTermsandConditionsofUse,availableat://=ucpress.EachcopyofanypartofaJSTORtransmissionmustcontainthesamecopyrightnoticethatappearsonthescreenorprintedpageofsuchtransmission.JSTORisanot-for-profitorganizationfoundedin1995tobuildtrusteddigitalarchivesforscholarship.Weworkwiththescholarlycommunitytopreservetheirworkandthematerialstheyrelyupon,andtobuildacommonresearchplatformthatpromotesthediscoveryanduseoftheseresources.FormoreinformationaboutJSTOR,[e-mail]gulateFirmsJean-JacquesLaffontUniversitydeToulouseJeanTiroleMassachusettsInstituteofTechnologyThepaperemphasizestheuseofaccountingdatainregulatoryorprocurementcontractswhenthesupplier(1)hassuperiorinforma-tionaboutthecostoftheprojectand(2)investsincostreduction.Themainresultstatesthat,underriskneutrality,thesupplieran-nouncesanexpectedcostandisgivenanincentivecontractlinearincostoverruns.This(optimal)contractmovestowardafixed-pricecontractastheannouncedcostdecreases.Aninvestmentchoiceisthenintroducedandtheuseofarate-of-returnregulationisstudied.I.IntroductionTheliteratureonthecontrolofpublicfirmsorprivatemonopoliescanbedividedintoaliteraturestudyingthepropertiesofgivenincen-tiveschemesandafewrecentpapersdesignedtocharacterizeoptimalcontrolmechanisms.Themajorinterestoftheearlierliterature'SupportfromtheCommissariatauPlanandtheNationalScienceFoundationisgratefullyacknowledged.WethanktworefereesandP.Picardforhelpfulcommentsonanearlierdraft.1TheliteratureonSovietbonusschemes(see,e.g.,Ellman1973;Fan1975;Bonin1976;Weitzman1976)andontherate-of-returnregulation(see,e.g.,Crampes[1982]forastudywithasymmetricinformation)considerstherevelationofproductionpos-sibilities.Anotherliterature(Domar1974;Tam1979,1981;FinsingerandVogelsang1981,1982;Vogelsang1983)assumesthatthefirmhassuperiorinformationaboutdemand.Althoughthelatterliteratureisrelevantinsomecases,wewouldexpectinformationalasymmetriesaboutproductionpossibilitiestobemoreimportantingen-journalofPoliticalEconomy,1986,vol.94,no.3]?1986byTheUniversityofChicago.Allrightsreserved.0022-3808/86/9403-0008$01.50614COSTOBSERVATION615stemsfromthesimplicityoftheschemesstudied,whichareeasilyrelatedtowhatcanbeobservedinplannedeconomiesorinlargecorporations.However,theyareadhoc.Thereisclearlyaneedforanormativetheorythatwillderiveoptimalincentiveschemes,studytheperformanceoftheseadhocschemes,andtestthesoundnessoftheintuitiononwhichtheyarebased.Recentlyanormativetheoryhasemergedfromthenonlinearpric-ingliteratureandthemoreabstractincentivetheorydevelopedtodealwiththefree-riderproblem.Inthisapproachtheregulator/plannerisviewedasaBayesianstatisticianwhohaspriorknowledgeaboutcostanddemandconditions.Theoptimizationproblemoftheregulatoristomaximizetheexpectedsocialwelfareunderthecon-straintofthedecentralizationofinformation.Theoutcomeoftheanalysisisthecharacterizationofoptimalincentiveschemesgiventheobjectivefunctionsandtheobservationsmadebytheregulator.Papersinthistradition(LoebandMagat1979;BaronandMyerson1982;Sappington1982)studythecontrolofaprivatemonopolywhenthedemandfunctioniscommonknowledgeandthecostfunc-tioncanbeparameterizedbyonerealnumber.2Theoptimalincen-tivemechanismingeneralentailsawelfarelosscomparedwithwhatcouldbeachievedunderperfectinformation.Costsareeasytoobserve,atleastatthefirm'slevel.Thevalueofcostobservationtotheplannerdependsonwhatheattemptstocon-trol.Ifhemonitorsasingleprojectinamultiprojectfirm,thefirmcanshiftexpensestoandfromtheparticularproject,atbothrealandaccountinglevels.Inafirstapproximationitisreasonabletoassumethattheplannerdoesnotperfectlyobservethefirm'scostfortheproject.3Whentheplannercontrolstheentirefirm,aggregatecostinformationbecomesveryvaluable.IfcostobservabilityisintroducedintotheBaron-Myersonmodel,itisthenpossibletoinferthetruecostparameterandtoreachthefirst-bestwithappropriatepenalties.eral(e.g.,theregulatorcanrunaconsumersurveyorusethefirm'soutputandpricedatatoestimatedemand).Bergson(1978)hasstressedtheroleofdistributionalaspects.Ouruseofasocialcostoftransferscanbeviewedasaformalizationoftheconcernforequityinthedesignofincentiveschemes.2Theanalysiscanbeconceptuallygeneralizedtoanynumberofparameters.Inparticular,demandmayalsobeparameterizedinthesamewayascostsiftheregulatordoesnotknowdemand.But