Price discovery in initial public offerings and th

PriceDiscoveryinInitialPublicOfferingsandtheRoleoftheLeadUnderwriterReenaAggarwalandPatConroy*October1999*AggarwalisattheMcDonoughSchoolofBusiness,GeorgetownUniversityandConroyisatfolioTradeLLC.PartofthisworkwasdonewhilebothAggarwalandConroywereattheSecuritiesandExchangeCommission(SEC).WethankseminarparticipantsattheSEC,NASD,GeorgetownUniversity,the1999meetingsoftheEuropeanFinancialManagementAssociation,BillByrnes,PatFishe,ToddHouge,TimLoughran,JayRitter,PietraRivoli,RenéStulz(theeditor),andananonymousrefereeforprovidingveryusefulcomments.ThisresearchwaspartiallysupportedbyresearchgrantsfromGeorgetownUniversityandtheCapitalMarketsResearchCenter.TheSEC,asamatterofpolicy,disclaimsresponsibilityforanyprivatepublicationorstatementbyanyofitsemployees.TheviewsexpressedhereinarethoseoftheauthorsanddonotnecessarilyreflecttheviewsoftheCommissionortheauthors’colleaguesuponthestaffoftheCommission.AbstractWeexaminethepricediscoveryprocessofinitialpublicofferings(IPOs)usingauniquedataset.Thefirstquoteenteredbytheleadunderwriterinthefive-minutepreopeningwindowexplainsalargeproportionofinitialreturnsevenforhotIPOs.Significantlearningandpricediscoverycontinuestotakeplaceduringthesefiveminuteswithhundredsofquotesbeingentered.Theleadunderwriterobservesthequotingbehaviorofothermarketmakers,particularlythewholesalers,andaccordinglyreviseshisownquotes.ThereisastrongpositiverelationshipbetweeninitialreturnsandthetimeofdaywhentradingstartsinanIPO.1ResearchershavedocumentedandtriedtoexplainwhyIPOsjumpupinpriceinitiallybutthenperformpoorlyinthelongrun.1SchultzandZaman(1994)andBarryandJennings(1992)reportthatalmosttheentireinitialreturnisreflectedintheveryfirsttradeprice.However,researchershavenotexaminedhowthepricechangesfromtheofferpricetothepriceofthefirsttrade.Ourempiricalanalysisexplainsthelearningprocessbywhichthepricechangesfromtheofferpricetothefirsttradeprice.Theofferpriceistypicallysetafterthemarketclosesonthedaypriortothefirstdayoftrading.Yet,thereisalargepricerun-upbythenextmorning.Forexample,Amazon.comwentpubliconMay15,1997atanofferpriceof$18andthefirsttradeoccurredat10:30AMonMay16atapriceof$29.25.Thisresearchisalsomotivatedbytheconcernofstockexchanges,regulators,andmarketparticipantsabouttheinitialpricediscoveryandvolatilityofIPOs.PricediscoveryisparticularlyimportantanddifficultfortheopeningofIPOsbecausenotradinghistoryexists.Therefore,theinitialtradingandpricediscoveryinthesestockscanbeverynoisyandhasbecomeacauseforconcern.AsEllis,Michaely,andO’Hara(1999)discuss,theleadunderwriterisalwaysamarketmakerinNasdaqIPOs.Initsroleasamarketmaker,theleadunderwritermustinitiallydecideatwhatpricetostartquotingandtradingthestock.Weuseuniquequotedatawiththeidentityofthemarketmakertoexaminethequotingbehavioroftheleadunderwriterduringthepreopeningperiod;thebehaviorofothermarketmakers;theimportanceofthepreopeningperiodforlearningandpricediscovery;andfactorsthatdeterminethetimeofdaywhentradinginanIPOstarts.Thepaperanalyzeshowaccuratetheleadunderwriter’sstartingquotesareandhowhelearnsfromthequotingbehaviorofothermarketmakersanddecidesthepriceatwhichtobuy/sellthestock.IPOshaveapreopeningperiodthatlastsforamaximumoffiveminutesbeforeactualtradingbegins.Duringthisfive-minutepreopeningperiod,allmarketmakershavetheoptionto2add,revise,orcancelquotesbeforetradingactuallybegins.2Nasdaqisexaminingwhetherthefive-minutepreopeningwindowshouldbelengthenedforsomestockstoachievemoreefficientpriceformationandlowervolatility.Theargumentforalongertimeperiodisstatedbymarketparticipants,“Thefive-minuteperiodwasn’tnearlyenoughtimetogaugethehugelevelsofdemandthathavebuiltformostrecentinternetdeals,andtodeterminewherethestockwouldheadonceitopened.....thenewruleswouldallowNasdaqtradersmoretimetodetermineatwhatpriceanIPOislikelytoopen…”-TheAsianWallStreetJournal,February3,1999Quotesenteredintothesystemduringthepreopeningperiodarenotbinding.Therefore,itispossiblethatmarketmakersdonotshowtheirtrueintentionwhenenteringthesequotes.Theyfaceonlythesmallcostsoforderplacementandhandlingwithouttheriskofexecution.However,marketmakersmayhaveincentivestoproducepricediscoveryevenintheabsenceofbindingcommitments.Theleadunderwriteriscertainlymotivatedtolearnfromthequoterevisionprocessinthepreopening.AllmarketmakersmaycooperateinthepricediscoveryprocessbecausetheopeningofIPOsisarepeatedgame.Therecanbeaconsiderableamountofactivityduringthisfive-minutewindow.Forexample,therewere116quoteentriesinthecaseofAmazon.comduringthefive-minutepreopeningperiod.Weanalyzewhetherthesenon-bindingquoteshaveanyvalueandhowthepricediscoveryprocessworkseventhoughnotradestakeplaceduringthistimeperiod.Thepricediscoveryprocessstartsevenbeforethepreopeningperiodbegins.TheleadunderwriterdecideswhentostarttradinganIPOandsetsthefirstquoteduringthefive-minutepreopening.Thisfirstquoteexplainsalargeportionofinitialreturns.AlimitednumberofrecentstudieshaveempiricallyexaminedhowopeningpricesaredeterminedontheParisBourse,theNewYorkStockExchange(NYSE),andonNasdaqfornon-IPOstocks.Biais,Hillion,andSpatt(1999)findth