THE ENDOWMENT EFFECT英语资料.docx
《THE ENDOWMENT EFFECT英语资料.docx》由会员分享,可在线阅读,更多相关《THE ENDOWMENT EFFECT英语资料.docx(7页珍藏版)》请在冰豆网上搜索。
THEENDOWMENTEFFECT英语资料
TheEndowmentEffect
PraveenKujal,UniversidadCarlosIII.
VernonL.Smith,UniversityofArizona.
TheBackground
Theemergenceofempiricalevidencesuggestingdivergencebetweenthewillingness-to-accept(WTA),forthesaleofanobject,andthewillingness-to-pay(WTP),forthepurchaseofanobject,hasresultedintwoexplanations.OneoftheexplanationswasproposedbyKahneman,KnetschandThaler(1991)(hereafterKKT),thatthediscrepancybetweentheWTPandtheWTAmaybeamanifestationoftheendowmenteffect.Thus,“...endowmenteffectswillalmostcertainlyoccurwhenownersarefacedwithanopportunitytosellanitempurchasedforusethatisnoteasilyreplaceable”(p.1344).Furthertheyarguethattheendowmenteffectwillnotapplywhenthegoodsarepurchasedforresaleandnotforuse;thereisnoendowmenteffectfortheretailfirm,onlyfortheconsumerpurchasingthefirm’sgood.Similarly,theyarguethattheendowmenteffectdoesnotapplytotheexchangeoftokens(orrights)towhichprivateredemptionvalues,orinducedvalueshavebeenassignedbytheexperimenter(Smith,1976).
ThesecondexplanationcamefromHanemann(1991)whoshowedthatwhenclosesubstitutesexistforthegoodinquestion(andgivenpositiveincomeelasticity)WTAandWTPcanbeshowntobeveryclose,andvice-versa.
GiventheexplanationsforthedivergencebetweentheaverageWTAandWTPtheexperimentalliteraturefocusedondirectchoicetestsofthese(theoretical)explanations,andthereexaminationinmarketcontexts.
Thefirstchoice,andexchange,experimentswererunbyKKTestablishingtheendowmenteffectforCornellandother(emblem)coffeemugsbutnotforinducedvaluetokens.ThesewerefollowedbyexperimentsfromShogrenetal.(1994)andFranciosietal.(1996).Shogrenetal.establishthattheHannemanhypothesisunderrepeatplayisrobustandthatthedivergencebetweentheWTAandWTPdisappearswithrepeatinteractionforclosesubstitutesbutnotforimperfectsubstitutes.Further,intheirexperimentalsetting(distinctfromthatofKKT)theyshowthatunderrepeatinteractiontheendowmenteffectdisappears.Franciosietal.showthatwecanobserveundertradinginmarketseveniftheWTA-WTPdiscrepancyisnegligible.ThisistheresultofunderrevelationofintramarginalunitsleadingtoveryflatreportedinversesupplyanddemandsothatverysmalldeviationsinreportedWTAandWTPcanleadtoundertrading.
TheExperiments
Kahneman-Knetsch-ThalerChoiceExperiments:
InthetypicalexperimentofKKTanundergraduateclassisdividedintoequalparts.Halfthesubjectswererandomlyassignedtotheroleofbuyersandtheotherhalfsellers.Universityemblemcoffeemugs(Cornell,SimonFraser,orUniversityofBritishColumbia),costingaround$6inthelocalUniversitybookstore,werethendistributedtothesellers,andallthebuyersweregiventheopportunitytoexaminethemug.Thefollowingformswerethenexecuted.
IWillSellIWillKeep
[Buy][NotBuy]theMug
Ifthepriceis$0____________________
Ifthepriceis$0.50____________________
.....
Ifthepriceis$9.50____________________
Next,arandomprice(KKTusedtheBDMprocedure,inBeckeretal.,1964)wasdrawnfromthelistbetween$0and$9.50,andexchangeswereconductedbytheexperimenteronthebasisofthisprice.Thetypicalresultwasamediansellingpricethatwasdoublethemedianbuyingprice,anobservationthatisconsistentwiththeendowmenteffect.KKT,however,recognizedthatthisproceduredidnotcontrolforanyincomeeffect.Thisproblemwasexacerbatedbythefactthatbuyersintheirexperimentswererequiredtousetheirownfundswhilethesellersweregiventhecoffeemugs.
ToaddresstheneedtocontrolforincomeeffectsKKT(pp.179-80)developedaningeniousvariationontheaboveexperiments.Insteadoftwogroupstheyusedthree:
sellers,buyersandchoosers.Thesellers/buyersmadethesamechoicesasbefore,whilethechooserswereaskedtochooseateachprospectivepricebetweenthemug,orcash.Thus,sellersweregivenamug,andchoosersweregiventherighttoeitheramugorcashastheychose;anyincomeeffectonsellersasdistinctfrombuyers,shouldthenalsoapplytothechoosers.ThedifferenceaccordingtotheKKTimplementationoftheendowmenteffectisthatsellersownthemug,choosersdonot.
Theirresultswereclear:
choosersbehavemorelikebuyersthanlikesellers,althoughchoosersvaluemugssixtypercentmorehighlythanbuyers.(SeeKKT,p.178-80).
Franciosietal.ChoiceExperiments
Franciosietal.conductedfourexperimentseachwith24subjects(8ineachgroup;N=96subjectsintotal)motivatedbythethree-groupdesignwhichcontrolledforanyincomeeffect.However,theymadeseveralinstructionalchangeswhichmightbeimportantinthecontextofthechoiceexperimentsduetotheirframingeffects.Becausetheuseofemotivetermssuchas“buying,”and“selling”mayalterthestrategicbehaviorofmarketparticipants,Franciosietal.(1996)neutralizedtheirinstructionsandremovedallmentionof“buying”or“selling.”Insteadtheyuseexpressionsthatdidnotsuggestanyspecificrolebehavioronthepartofmarketparticipants.First,thethreegroupsweresimplycalledGroup1,Group2andGroup3.ThesubjectsweretoldthateachmemberofGroup1isanownerofanArizonaWildcatMug,andtheirtaskistochoose,foreachamountofmoney(nomentionof“price”),betweenretainingtheirmugor,acceptingtheadditionalamountofmoney.EachmemberofGroup2wasdesignatedashavingtherighttochoosebetweenacceptingamug,orretaininganamountofmoneyoutoftheirearningsinapreviousunrelatedexperimentinthesamesession.Finally,eachpersoninGroup3isdesignatedashavingtherighttochoosebetweenamugoracceptinganadditionalamountofmoney.Thus,allsubjectsweresymmetricallydescribedaschoosers,butunderdifferentinitialconditions.
Alltheirchoiceexperimentswererunattheendoftwosimultaneouspostedoffermarketexperiments(6buyers,6sellersineach),reportedinFranciosietal.(1995).Subjectswereassignedtothethreegroupsrandomly,andwerepaidtheirearningsincashattheendofthemarketexperiments.Earningsrangedfrom$8.75-$62.50providingallGroup2subjectswithsufficientfundstoobtainamugbasedontheexperimentalprices.Themugswerepricedat$9.95(pricetagsremoved)inthecampusbookstore.
FromTable1itcanbeseenthatthesubjectsreportsubstantiallylowerGroup1WTA,asomewhathigherWTP,andahigherGroup3WTA,thandidtheKKTsubjects.SubstitutingachoicetaskforthebuyingandsellingtasksappearstonarrowtheWTA-WTPdiscrepancy.Pairwisestatisticaltests,however,showthatthedatafromallthreegroupscomefromdifferentdistributions(Franciosi,eta.,1996).Hence,thequalitativedifferencesamongthethreegroups,aspostulatedbyKKT,weresupportedasisevidentinthelastrowofTable1.
KKT(1991)Exchangeexperiments:
InadditiontotheirBDMchoiceexperimentsKKTreporttheresultsofseveralexchangeexperiments.Halfthesubjectswererandomlyassignedtheroleofbuyers,theotherssellers.Sellerswereeachendowedwithamug,whiletheformerusedmoneytheyhadbeenaskedtobringtoclass.Buyerseachsubmittedabidpricetobuyamug,sellerseachsubmittedanofferpricetosellthemug.Their‘bids’and‘offers’weresolicitedbyaskingeachsubjecttochoosebetweenapriceandamugforaseriesofpricesasintheBDMprocedureexceptthattherangestartsat$0.25andgoesupinincrementsof$0.50.Theintersectionofthedescendingbidsandascendingoffersdeterminesthepriceandquantityexchanged.Iftherearenoendowmentorincomeeffects,thenduetotherandomallocationofsubjectstothebuyorsellcondition,thesupplyscheduleofthosegiventhemugsshouldbeamirrorimageofthedemandscheduleforthosenotgiventhemugs.Thisleadstothepredictionthatone-halfofthemugsshouldtrade.Consequently,intheirexperimentswith22buyersand22sellers,11mugswerepredictedtotrade.Theyobservethatbetween1and4tradeatpricesbetween$4.25and$4.75.Asbeforeonlyonebid/offertrialischosenatrandom.
MugExchangeExperimentsusingUniformPriceDoubleAuction
Franciosietal.(1996)usedtheuniformpricedoubleauction(UPDA)mechanismtostudymugexchangeduetoitsstrongequilibratingproperties.(SeeMcCabe,RassentiandSmith,1993).Theauthorsfeltthatusinganefficientauctionmarketmechanismmaybecrucialtotestingtheundertradinghypothesis.
Twoseriesofexperimentswererun.Ineachexperiment24differentsubjectswererandomlyassignedtogroupsof12buyersand12sellers.Eachserieswasdividedintoparts1and2.InPart1ofseries1eachbuyerwasassignedavalueandeachselleracostfromtheuniformdistributionon[$0,$9.99atthebeginningofeachof10(or12)tradingperiods.Thisbaselineservedasatrainingsession.Allperiodslasted4minutes.InPart2,Series1and2,eachbuyerwasendowedwitha$9.99cashbalancewhichwastheirstokeepifnomugwaspurchased;eachsellerwasendowedwithaUniversityofArizonaemblemmugpricedat$9.95whichwastheirstokeepifnotsold.EachsubjectwaspaidincashalloftheirearningsfromtheinducedvaluetrainingexperimentsinPart1ofeachofthesessions.Inseries2,Part1usedtheconstantvolumeequilibriumenvironment,butineachperiodarandomconstantwasaddedtoeachvalue,andthevaluesrandomlyassignedtoeachindividual.Part2ofseries2waslikethatofseries1exceptthatthepricetag($9.95)wasleftoneachmug,andthiswaspointedout