WHY NEW ISSUES ARE UNDERPRICED

JournalofFinancialEconomics15(1986)187-212.North-HollandWHYNEWISSUESAREUNDERPRICED*KevinROCKHurvordBwnessSchool,Boston,MA02163,USAReceivedNovember1984,finalversionreceivedAugust1985Thispaperpresentsamodelfortheunderpricingofinitialpublicofferings.Theargumentdependsupontheexistenceofagroupofinvestorswhoseinformationissuperiortothatofthefirmaswellasthatofallotherinvestors.Ifthenewsharesarepricedattheirexpectedvalue,theseprivilegedinvestorscrowdouttheotherswhengoodissuesareofferedandtheywithdrawfromthemarketwhenbadissuesareoffered.Theofferingfirmmustpricethesharesatadiscountinordertoguaranteethattheuninformedinvestorspurchasetheissue.1.IntroductionSeveralyearsago,Grossman(1976)showedthatifoneclassofinvestorshassuperiorinformationabouttheterminalvalueofanasset,theinformationcanbereadbyanyonefromtheequilibriumprice.Thisresultproducesaparadox.Ifanyonecaninferprivateinformationfromtheequilibriumprice,noonepaystocollectinformation.Yetifnoonecollectsinformation,thepricerevealsnone,andanincentiveemergestoacquireit.Thekeytotheparadoxistheassumptionofanoiselessenvironment.Ifnoiseispresentintheequilibriumprice,privilegedinformationissecure.Fortheuninformedcannotbesurewhetherahighpricereflectsfavorableinforma-tionorextraneousfactors,suchasachangeinriskaversionoraneedforliquidity.Thispapertakesanalternativeapproach.Ifprice,whichisobservable,doesnotcorrespondtoauniquelevelofdemand,whichisunobservable,thenthemainchannelbywhichinsideinformationiscommunicatedtothemarketisdestroyed.Untilthechannelisre-established,theinformedinvestorhasanopportunitytoprofitfromhisknowledgebybiddingfor‘mispriced’securities.Inthisway,theinvestoriscompensatedforhiscostlyinvestigationsintotheasset’svalue,andobtainssomeremunerationforshowingwherecapitalshouldbestbeallocated.*ThispaperisbasedonChapter1ofmyPh.D.dissertationattheUniversityofChicagoin1982.IwouldliketothankthemembersofmyCommittee:JonathanIngersoll,MertonMiller(insidemembers),DouglasDiamond,TomGarcia,andWillardZangwill(outsidemembers).IoweaspecialthankstoGeorgeConstantinides(chairman),aswellastoJayRitteroftheUniversityofMichigan.Inaddition,MichaelJensenprovidedmanyhelpfulcommentsandsuggestions.0304-405X/86/$3.5001986,ElsevierSciencePublishersB.V.(North-Holland)188K.Rock,WhynewissuesareunderpricedThesettingforthismodelisthenewissuesmarket,inparticular,themarketfor‘firmcommitmentofferings’.Inafirmcommitmentoffering,thefirmanditsinvestmentbankagreeonapriceandquantityforthefirm’sfirstissuanceofequity.Oncethepriceisset,typicallyonthemorningoftheoffer,nofurtheradjustmentsareallowed.Ifthereisexcessdemand,theunderwriterrationstheshares,sometimesexercisingan‘overallotmentoption’whichpermitsasmanyas10%moretobesold.Ifthereisexcesssupply,theofferconcludeswithunsoldshares.Theinvestmentbankpaysthefirmforthesurplussharesanddisposesofthemlateratmarketprices.Eachcondition-excesssupplyordemand-isnotobserveduntilafterthe‘offeringdate’.Onlythendoesthepresenceorabsenceofinformedtradingbecomeapparent.Thenewissuemarketresemblesanauction,buttheresemblanceisnotexact.Priceisnotdeterminedbythebiddingofinvestors.Inparticular,theinvestorwiththehighestvaluationneednotobtaintheshares,evenifthevaluationexceedstheissuer’sreservation(offer)price.Thatinvestormaysimplynotreceiveanallocationofrationedsharesfromtheunderwriter.Moreover,theissuingfirmisbothabidder,whosubmitsapriceinconsultationwiththeunderwritinginvestmentbank,andaseller,whoexchangesanassetforcash.Nevertheless,thespiritofthemodelanditsmethodologybelongtotheauctionliterature.Thismodelisdirectedtowardanexplanationofananomalyinthenewissuemarket.Newsharesappeartobeissuedatadiscount.Ibbotson(1975)testedthishypothesisandfound,onaverage,an11.4%discountintheofferpricewhichdisappearedwithinweeksintheaftermarket.Usingasimplermodel,IbbotsonandJaffe(1975)founda16.8%averageexcessreturnrelativetothemarket.Bothwereunabletoaccountfortheirfindings.Aftersuggestingseveralexplanations,Ibbotsontermedthephenomenona‘mystery’.Thediscountisanaturalconsequenceofthepresentmodel,whichincorpo-ratesasymmetricinformationandrationing.IbbotsonandJaffethemselvesnoticethatunderpricedsharescanbeseverelyrationed.Theymentionthatitisnotuncommonforunderwriterstoreceive,priortotheeffectivedata,‘indica-tionsofinterest’forfivetimesthenumberofsharesavailable.Thisphenome-nonhasaneffectupontheuninformedinvestor.Ifaninvestorfindsthathereceivesnoneoftheunderpricedissuesduetorationingbroughtonbyinformeddemand,andalloftheoverpricedissues,thentheinvestorrevisesdownwardshisvaluationofnewshares.Hedoesnotparticipateinthenewissuemarketuntilthepricefallsenoughtocompensateforthe‘bias’inallocation.Theanalysisshowsthattheequilibriumofferpriceincludesafinitediscounttoattractuninformedinvestors.Thisresultisnotaforegoneconclusion.Itisnotimmediatelyclearwhatadvantageaccruestotheissuerfromuninformedparticipation.Norisitclearifanydiscountissufficienttoattractthemtotheoffering.ItisconceivablethatreducingtheofferpricecouldelicitgreaterK.Rock,Whynewissuesareunderpriced189informeddemand,exacerbatethebias,andfurtherdisadvantageth