Herding-behavior-in-online-P2P-lending--An-empiric

HerdingbehaviorinonlineP2Plending:AnempiricalinvestigationEunkyoungLee,ByungtaeLee⇑KAISTBusinessSchool,Seoul,RepublicofKoreaarticleinfoArticlehistory:Availableonline13February2012Keywords:EconomicanalysisEmpiricalresearchHerdingMarketsharemodelingMicrofinanceP2PlendingReverseauctionsSocialnetworksabstractWestudylenderbehaviorinthepeer-to-peer(P2P)lendingmarket,whereindividualsbidonunsecuredmicroloansrequestedbyotherindividualborrowers.OnlineP2Pexchangesaregrowing,butlendersinthismarketarenotprofessionalinvestors.Inaddition,lendershavetotakebigrisksbecauseloansinP2Plendingaregrantedwithoutcollateral.WhiletheP2Plendingmarketsharessomecharacteristicsofonlinemarketswithrespecttoherdingbehavior,italsohascharacteristicsthatmaydiscourageit.ThisstudyempiricallyinvestigatesherdingbehaviorintheP2Plendingmarketwhereseeminglyconflictingconditionsandfeaturesofherdingarepresent.UsingalargesampleofdailydatafromoneofthelargestP2PlendingplatformsinKorea,wefindstrongevidenceofherdinganditsdiminishingmarginaleffectasbiddingadvances.Weemployamultinomiallogitmarket-sharemodelinwhichrelevantvariablesfrompriorstudiesonP2Plendingareassessed.2012ElsevierB.V.Allrightsreserved.1.IntroductionPeer-to-peer(P2P)lendingisabreedoffinancialtransactionsthatoccurdirectlybetweenindividualswithouttheintermediationofatraditionalfinancialinstitution(en.wikipedia.org/wiki/P2p_lending).Ithasashorthistory,buthasrapidlygrowninrecentyears.ThefirstonlineP2PlendingcompanywasZopa(),launchedin2005intheUnitedKingdom.IntheUnitedStates,Prosper()wasthefirstP2Plendingfirm,andopenedtothepublicinFebruary2006.ItisnowthelargestP2Plendingplatform,withoveramillionmembersandoverUS$219millioninpersonalloansfundedasofFebruary2011.P2PonlineexchangesaregrowingintheUnitedStatesandUnitedKingdomasanalternativeplatformtotraditionalsavingandinvestment(Slavin2007).HarvardBusinessReviewreportsthateverymajorbankwillhaveitsownP2Plendingnetworkwithinfiveyears,andthatP2Plendingwillbeamongthemostimportantfinancialserviceinnovationsinthecomingdecade(Sviokla2009).Thisnewphenomenonhasgarneredsignificantattentionfromresearchers.ManyofthemfocusonsocialnetworksinP2Plending(FreedmanandJin2008,Herrero-Lopez2009,Linetal.2011).IntheP2Plendingmarket,transactioncostsarereducedbyeliminat-ingexpensiveintermediaries,butinformationasymmetryprob-lemsbecomemoreseverethanintraditionalmarkets.ThisisbecausemostindividuallendersinonlineP2Plendinglackfinan-cialexpertise,andthelendingexperiencetakesplaceinapseudon-ymousonlineenvironment(Klafft2008).Inthissituation,socialnetworksbetweenindividualsmitigateadverseselectionandleadtobetteroutcomesinallaspectsofthelendingprocess(Linetal.2011).SocialnetworksonProsperrevealsomesoftinformationaboutborrowerrisk,andthereforehavethepotentialtocompen-sateforthelackofhardinformation(FreedmanandJin2008).Be-sidessocialnetworks,borrowers’characteristics,includingdemographiccharacteristics,financialstrength,andeffortpriortomakingarequest,areregardedasdeterminantsoffundingsuc-cessinP2Plending(Herzensteinetal.2008).Despitethosenewexperimentalmechanismdesignsandsystemfeatures,theriskofinformationasymmetrylendersfacemaynotbeerasedeasily.Ithasbeenstudiedthatplayersexhibitherdingbehaviorsinonlinecommercewhentheyfaceriskofuncertaintysuchasinformationasymmetry.Herdingbehaviordescribesmanysocialandeconomicsituationsinwhichanindividual’sdecision-makingishighlyinfluencedbythedecisionsofothers(Duanetal.2009).Therefore,ithasbeentheoreticallylinkedtomanyeconomicareassuchasinvestmentrecommendations(ScharfsteinandStein1990),pricebehaviorofinitialpublicofferings(IPOs)(Welch1992),fadsandcustoms(Bikhchandanietal.1992),anddelegatedportfoliomanagement(MaugandNaik1996).Duanetal.(2009)suggestthatherdingbehaviorcouldbeespe-ciallyprominentontheInternetfortworeasons.Thefirstisinfor-mationoverload.ThereisanexcessiveamountofinformationontheWeb,soonlineusershavedifficultyunderstandingandusingalltheinformation(BrynjolfssonandSmith2000).Doingwhatoth-ersdocouldbeanefficientandrationalwaytomakedecisionsinthiscircumstance.Thesecondreasonisthatpeoplecaneasilyob-serveothers’choicesontheInternet.Mostonlinee-commercewebsitesprovideawaytosorttheirproductsintheorderof1567-4223/$-seefrontmatter2012ElsevierB.V.Allrightsreserved.doi:10.1016/j.elerap.2012.02.001⇑Correspondingauthor.E-mailaddresses:[e-mail],[e-mail](B.Lee).ElectronicCommerceResearchandApplications11(2012)495–503ContentslistsavailableatSciVerseScienceDirectElectronicCommerceResearchandApplicationsjournalhomepage:(2008),thereisaconsiderabledifferencebetweenthenumberoflendersbiddingonfundedloanlistingsandthenumberoflendersbiddingonunfundedloanlist-ings.Theaverageoftheformeris62.6,whiletheaverageofthelat-terisonly1.6.Whatmakessuchsignificantdifference?Isittheout