126K - views

Marketing Letters

2001 Kluwer Academic Publishers Manufactured in The Netherlands Always Leave Home Without It A Further Investigation of the CreditCard Effect on Willingness to Pay DRAZEN PRELEC AND DUNCAN SIMESTER Sloan School of Management MIT 38 M

Download Pdf

Marketing Letters




Download Pdf - The PPT/PDF document "Marketing Letters " is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.



Presentation on theme: "Marketing Letters "— Presentation transcript:

MarketingLetters12:1,5±12,2001. # ownersandnon-owners,ordifferencesintheoccasionsonwhichcashandcreditcardsarethepreferredmethodsofpayment.WhileFeinberg'sandSoman'sresultsarefreeofthisconfound,thegeneralityoftheir®ndingsisalsolimited,butbydifferentfactors.Forexample,onlyoneofthesestudies(Feinberg,Study3)involvedrealmoneytransactions;charitabledonationstotheUnitedWay.Althoughthepresenceofcreditcardlogosinthestudydidincreasedonationsize,theabsoluteamountsweresmall(averagedonationswere11¢and33¢forthecashandcreditcardconditions).Second,Feinberg'sstudiesonlymanipulatedexposuretocreditcardstimuliandnotthepaymentmethoditself.Paymentwasleftambiguousintheperceivedvaluestudies,whiletheUnitedWaystudyrequesteddonationsincash.2.PurposeofStudyThisempiricalnotepresentsnewevidencesupportingthepropositionthatconsumersarewillingtospendmoreforaproductwhenusingacreditcard(inatleastsomepurchasecontexts).WeextendFeinberg'sandSoman'sexperimentsintwoseparatestudies.Inthe®rststudy,wemanipulatethepaymentmethod,whileusingreal-moneytransactions.Recallthatonlyoneofthepreviousstudiesusedrealmoneytransactionsandinthatstudythepaymentmethodwasnotmanipulated(allrespondentspaidincash).Furthermore,weusedesirablegoodsofpotentiallyhighvalue.ThisallowsustoinvestigatewhethertheFeinbergresultislimitedtotransactionsofverysmallmagnitudeorwhetheritalsoextendstotransactionsinvolvingmuchlargeramounts.Inthesecondstudywefurtherexplorethenecessaryandsuf®cientconditionsthatsupporttheeffect.Inparticular,weseparatelymanipulatedpaymentmethodandexposuretoacreditcardlogotoinvestigatewhichofthesefactorsstimulatepurchasing.Moreover,weinvestigatewhethertheeffectonlyariseswhenthemarketpriceoftheproductisuncertain.Inthesecondstudyweuseaproductforwhichthemarketpriceisknown.Ourresultsrevealapotentiallylargecreditcardpremium,buttheeffectisnotalwayspresent.Inthe®rststudy,whichusesgoodswithuncertainmarketvalue,weobservealargepremium(morethan100%).However,inthesecondstudy,usingagoodwithaknownmarketvalue(arestaurantgiftcerti®cate)thepremiumdoesnotalwaysarise.3.Study13.1.PrizesThe®rststudywasdesignedtoelicitwillingness-to-payforticketstosportingevents.Twopairsofticketswereseparatelyauctionedoff.Onepairofticketswasforthelastregular-seasongame,betweentheBostonCelticsandtheMiamiHeat.TheCelticsneededthiswintoclinchthedivisiontitle,and,likeallCelticsgamesatthetime,wassold-outwellinadvance.ThegamewasplayedonaSundayafternoon,threedaysaftertheThursdayauction.Theotherpairofticketswastoaregularseasonbaseballgame,betweentheRedPRELECANDSIMESTER SoxandtheTorontoBlueJays.Therewasalsoaconsolationprizeofapairofbanners(onefeaturingtheCelticsandonefeaturingtheRedSox).Wedidnotprovideinformationaboutthemarketvaluesofanyofthethreeprizes.Thespeci®cdescriptionswere:(1)Onepairof3rdrowbalconyticketsforCeltics-Miamigame,SundayApril19,1PM,attheBostonGarden.(2)OnepairofbleacherticketsforRedSox-Torontogame,SundayApril19,1PM,atFenwayPark.(3)ConsolationprizeofoneCelticsandoneRedSoxbanner.3.2.RespondentsandGeneralProcedureTherespondentswereMBAstudents,whorespondedtoaposterpromisinga$2billandanopportunitytopurchaseCelticstickets.Theexperimentwasheldduringlunchtimeinaclassroom.Uponenteringtheroom,studentswerehandedaninstructionsheetthatdescribedtheprizes,therulesforallocatingthetickets,andasheetforindicatingtheirreservationvalues.Theywereinstructednottodiscusstheiranswersoranythingelseaboutthequestionnaire.Respondentshandedintheircompletedsheetsastheylefttheroom.Theyweretoldthatthewinnerswouldbeannouncedat5PMthatafternoon.Weelicitedreservationvaluesusingasecond-pricesealed-bidauction.Inthisprocedure,theprizeisgiventothepersonwhowritesdownthehighestvalue;however,theprizeissoldatapriceequaltothesecondhigheststatedvalue.Undertheserules,participantshavenoreasontoeitheroverestimateorunderestimatetheirtruemaximumwillingness-to-pay(i.e.,itisan``incentive-compatible''elicitationmechanism;Vickrey1961).Respondentswrotedowntheirreservationvaluesforallthreeprizes.Ifthesamepersonwrotedownthehighestvalueformorethanoneprize,theywouldonlywintheprizethattheyvaluedthemost.3.3.TheCashvs.CreditCardManipulationUnbeknownsttotherespondents,twodifferenttypesofelicitationsheetswerehandedout,inarandomfashion.The®rst,cashconditionsheet,stipulatedthatpaymentwasrequiredfromwinnersincash.Respondentsinthisconditionwerealsoaskedtoindicatewhethertheyhad``readyaccesstoalocalcash-machine.''Thesecond,creditcardconditionsheet,stipulatedthatpaymentwastobemadebycreditcard.Respondentsinthisconditionwereaskedtoidentifythetypeofcardthatwouldbeusedtopurchasetheprizes(VISA,Amex,orMasterCard),itsexpirationdate,andtowritedownthelastthreedigitsofthecreditcardnumber.Respondentsinbothconditionsweretoldthattheywouldhavetomakethepaymentby5PMthenextday.Inaslightdeception,weplannedtoselltheticketstothewinnersatnomorethantheirfacevalue,andwouldacceptpaymentbycheckineithercondition.ALWAYSLEAVEHOME 3.4.ResultsWefoundthatrespondentsinthecreditcardconditionwrotedownsigni®cantlyhighervaluesforallthreeprizes(Table1).ThecreditcardpremiumwashighestfortheCelticstickets,andsmallestforthebanners(althoughthesedifferenceswerenotstatisticallysigni®cant).Allthreepremia,andthepremiumfortheCelticsticketsinparticular,weremoresubstantialthancouldbejusti®edbythe®nancialbene®tsofcreditcards.4.Study2Inthe®rststudy,respondentsrequiredtopaybycreditcardofferedhigherbidsonaveragethantheircounterpartsinthecashcondition.Inthesecondstudyweinvestigatethenecessaryandsuf®cientconditionsthatsupportthiseffectbyvaryingthestudyintwoways.First,wemeasurewillingnesstopayforaproductforwhichthemarketpriceiscertain.Recallthatinthe®rstconditionwemeasuredwillingnesstopayforticketstoasold-outsportingevent,forwhichthevaluewasunstated.Inthissecondstudywemeasurewillingnesstopayforagiftcerti®cateofastatedvalue.Thisallowsustoevaluatewhetherthecreditcardpremiumarisesbecausecustomersadjusttheirvaluationsfromdifferentanchoringpointsinthecashandcreditcardconditions.Forexample,inthecashconditionrespondentsmayanchortheirvaluationsusingtheamountofcashthattheytypicallycarryintheirwallets.Inthecreditcardconditiontheymayanchorontheircreditlimitorthesizeoftheirmonthlybill.Ifthisistruewewouldexpecttheeffecttobereducedifthestatedvalueofthegiftcerti®cateprovidesacommonanchorforvaluation.Thesecondstudyalsovariesfromthe®rststudyduetotheseparatemanipulationofpaymentmethodsandexposuretocreditcardlogos.Feinberg'sresultsindicatethatvaryingexposuretoacreditcardlogoissuf®cienttoeffectwillingnesstopay.Thissuggestsanassociationbiasinwhichrespondentshavepositiveassociationsforcreditcardsthatin¯atetheirwillingnesstopayregardlessoftheactualpaymentmethod.Ifthecreditcardpremiumisduetothiseffectalone,wewouldnotexpectalargeeffectwhenvaryingpaymentmethodwithoutvaryingexposuretothelogo.However,varyingexposuretothelogo,whileholdingpaymentmethodconstantshouldbesuf®cienttoproducetheeffect. Table1.Study1:MeanvaluesforCelticstickets,RedSoxtickets,andBanners,bypaymentmethodCelticsRedSoxBannersCashmean(N31)$28.51$9.02$3.32(stderr)(3.25)(1.10)(1.61)Creditcardmean(N33)$60.64$15.92$5.29(stderr)(11.09)(2.66)(1.66)Creditcardpremium113%76%59%t-testt2.71,p.01t2.35,p.05t.85,nsCashmedian$25.00$8.00$1.00Creditcardmedian$41.00$12.00$2.00Wilcoxsonrank-sumtestz2.64,p.01z1.42,nsz1.98,pPRELECANDSIMESTER 4.1.RespondentsandProcedureTherespondentswere®rst-yearMBAstudents,whowereparticipatinginadaylongorientationprogram.Studentswereseatedatseparatetablesinalargemeetingroom.Asabreakbetweentheorientationactivities,theywereinvitedtopurchaseadinnercerti®cateatanearbyrestaurant.Therestaurantisalocallandmarkandiswithin3±4minuteswalkingdistanceoftheschoolcampus.Thecerti®catewasdescribedasfollows:Thecerti®cateisgoodforupto$175,andcanbeusedonasingleoccasionwithinthenextthreemonths.Thecerti®catemaybeusedtopurchaseanyfoodorbeveragesonthemenu,andmayincludepaymentfortaxesandagratuity.Weelicitedreservationvaluesforthecerti®catewiththeBecker-DeGrootprocedure(Becker,DeGroot,andMarschak,1964).Likethesecond-priceauctioninStudy1,thismechanismisalso``incentive-compatible,''i.e.,itprovidesnoincentivesforoverstatingorunderstatingtruemaximumwillingness-to-pay.Themechanismwasexplainedasfollows:Step1:Thepriceofthecerti®catewillbedeterminedbydrawinganumberatrandom,fromzerouptothefacevalueofthecerti®cate(eachnumberisequallylikely).Step2:Awinnerwillbeselectedatrandomfromtherespondents.Step3:IfthevaluestatedbythewinnerisgreaterthanthepricerandomlyselectedinStep1thenthedinnercerti®catewillbesoldatthatrandomlyselectedprice.Ifthevaluestatedbythewinnerislowerthantheprice,thenalternativewinnerswillbeselecteduntilthedinnerissoldtosomeone.Althoughthesecond-priceauctionandBecker-DeGrootprocedurearestrategicallyequivalent(inbothcasesrespondentsshouldsimplywritedowntheirreservationvalues),thelatterproceduretendstoprovidemoresurplustothewinner,whichwasimportantinthecontextofthestudentorientationsetting.Thestudydividedtherespondentsintofourgroups,accordingtoanticipatedmethodofpaymentandtwotypesofpersonalidenti®cationmethods.Thetwopaymentmethodswereunchangedfromthe®rststudy(creditcardversuscashpayment).Thetwoidenti®cationmethodswereeitherasequenceofanyfourcharacters(lettersordigits),ortypeanddigits5±8ofamajorcreditcardownedbytherespondent(Visa,Amex,MasterCard).Thisidenti®cationmanipulationwasdesignedtovarycustomers'exposuretocreditcardlogosbypromptingrespondentsinthelatterconditiontopulloutandexaminetheircreditcards.NotethatwhileinStudy1,themethodofidenti®cationwascongruentwiththemethodofpayment,inthisstudy,wecreateda22design,varyingidenti®cationindependentlyofpaymentmethod.ALWAYSLEAVEHOME 4.2.ResultsThemeanvaluesbyconditionarepresentedinTable2.Thereareseveralinterestingcomparisons.First,thetwodiagonalcells,wherepaymentmethodiscongruentwithidenti®cationmethod,areareplicationofStudy1.However,thistimewe®ndnosigni®cantdifferencebetweenthecreditcardandcashconditions(t.66,ns;WilcoxsonRankSum:z.69,ns).Second,paymentmethodhasasigni®canteffect,butonlywhenanalysisisrestrictedtostudentswhosuppliedthecreditcarddigits(t2.58,p.05;WilcoxsonRankSum:2.45,p.05).Withinthesegroups,respondentswhoanticipatedpayingbycardwrotedownvalues36%greaterthanthosewhodidnot.Forstudentswhodidnotprovidetheircreditcarddigitsasidenti®cation,actualpaymentmethodmadenosigni®cantdifference1.2l,ns;WilcoxsonRankSum:z1.42,ns).Third,valuesinthetwocongruentconditionsaresigni®cantlygreaterthanthevaluesinthenon-congruentconditions(diagonalcellsversusnon-diagonalcells;t2.63,pWilcoxsonRankSum:z2.78,p.01).Weconjecturethatthiseffectmightbeduetotheratherunusualnatureoftherequestforcreditcarddigitsinthecashpaymentcondition.Thecoverstoryfortheupper-rightcellistheleastplausible,andthismayhavedepressedthesebids.Finally,themaineffectsofpaymentmethodandidenti®cationmethodwerenotsigni®cant.5.ConclusionsWeoffernotheoryexplainingthepresenceorsizeofthecreditcardpremiumobservedinStudy1.However,wedobelievethat,takentogether,thestudiesoffersomeinsightastothecauseoftheeffect.First,thetwostudiesappeartoexcludeliquidityconstraintsasacompleteexplanation.Ifthepremiumisduesolelytoliquidityconstraints,thenitshouldincreasewiththesizeoftheloan,andshouldhavebeengreaterinStudy2thanforanyoftheprizesinStudy1.ThefailuretoobserveanycreditcardpremiumbetweenthetwocongruentconditionsinStudy2isnotconsistentwiththisprediction.Furthermore,giventherandomassignmentacross Table2.Study2:Meanvaluesforthe$175dinnercerti®cate,bypaymentmethodandidenti®cationmethodPaymentmethodAny4charactersCreditcarddigitsCashmean$77.08$52.8043,std.err5.90)(N45,std.errCreditcardmean$67.12$71.7846,std.err5.71)(N34,std.errCreditcardpremium13%36%t-testt1.21,nst2.58,pCashmedian$80.00$50.00Creditcardmedian$62.50$69.00Wilcoxsonrank-sumz1.42,nsz2.45,pPRELECANDSIMESTER conditions,wecanpresumethatthe``creditcard''and``cash''populationsinStudy1areapproximatelyequivalent,bothintermsoftheirinterestintheprizesandintheir®nancialcharacteristics.Thismeansthatthe``medianrespondent''whowaswillingtopayupto$25fortheCelticsticketsunderthecashconditionwouldhavebeenwillingtopayupto$41inthecreditcardcondition.Ifthelowervalueinthecashconditionistrulyduetoahardliquidityconstraint,thenthis``medianstudent''wouldhavebeenwillingtopayalenderupto$16($41±$25)fortheabilitytoincreasetheirvaluebyanyamount(upto$41).Evenifweentertainmorecomplex®nancialexplanations(e.g.,maintainingsomeprecautionarycashbalances,andsoforth),itseemsimplausiblethatourrespondentswouldacceptloansatthehighexchangeratesthattheratioofcreditcardtocashvaluesimplies.Itisdif®culttobelievethatanMBAstudentwouldaccepta64%surchargeinexchangefortheprivilegeofpayingbycreditcardinsteadofcash!Second,theinformationprovidedaboutthemarketvalueoftheproductsofferedinthetwostudiesvaried.Inthe®rststudyweusedticketstoasold-outsportingevent,forwhichthevaluewasunstated,andinthesecondstudyweusedagiftcerti®cateofexplicitvalue.Thisdifferencewasdesignedtoinvestigatewhetherthecreditcardpremiumarisesbecausecustomersadjusttheirvaluationsfromdifferentanchoringpointsinthecashandcreditcardconditions.Thestrongerevidenceoftheeffectinthe®rststudyisconsistentwiththistheory.However,theeffectdidnotdisappearentirelyinthesecondstudy.Paymentmethodhadasigni®canteffectamongstrespondentsexposedtothecreditcardlogo.Third,paymentmethodandexposuretocreditcardlogoswereseparatelymanipulatedinthesecondstudy.Thepossibilitythatthecreditcardpremiumisduetoanassociationbiaswiththelogopredictsthatthepremiumdependsuponexposuretothelogoratherthanthepaymentmethod.The®ndingsinthesecondstudyarenotconsistentwiththisprediction.Althoughtherewassomeevidenceofapaymenteffect,therewasnoevidencethatexposuretothelogosincreasedwillingnesstopay.Insummary,thisisthe®rststudythatdemonstratesthatwillingness-to-payisincreasedwhencustomersareinstructedtouseacreditcardratherthancash.Theresultsaresurprisingbothduetothesizeofthepremiumandtheubiquityofcreditcarduse.Thevarianceweobserveinthepremiumprovidesahurdletootherexplanationsandameansofvalidatingalternativehypotheses.1.ThepaperwascompletedwhilePrelecwasaFellowattheCenterforAdvancedStudyintheBehavioralSciences,andsupportedinpartbyNSFGrant#SBR-960123totheCenter.AcknowledgementsWethankGeorgeLoewenstein,RobertMeyer,DilipSomanandNaderTavassolifortheirmanyhelpfulcomments.ALWAYSLEAVEHOME ReferencesAusubel,LawrenceM.(1991).``TheFailureofCompetitionintheCreditCardMarket,''AmericanEconomicReview81,50±81.Becker,GaryM.,MorrisH.DeGroot,andMarschak,Jacob.(1964).``MeasuringUtilitybyaSingle-responseSequentialMethod,''BehavioralScience9,226±232.Feinberg,RichardA.(1986).``CreditCardsasSpendingFacilitatingStimuli:AConditioningInterpretation,''JournalofConsumerResearch12,384±356.Hirschman,ElizabethC.(1979).``DifferencesinConsumerPurchaseBehaviorbyCreditCardPaymentSystem,''JournalofConsumerResearch6,58±66.Soman,Dilip.(1999).EffectsofPaymentMechanismonSpendingBehavior:TheIllusionofLiquidity.WorkingPaper,HongKongUniversityofScienceandTechnology,HongKong.Vickrey,William.(1961).``Counterspeculation,Auctions,andCompetitiveSealedTenders,''JournalofFinance16,8±37.PRELECANDSIMESTER