成交量与收益的关系(1)

TheVolume-ReturnRelationshipunderAsymmetryofInformationandShortSalesProhibitions1TheVolume-ReturnRelationshipunderAsymmetryofInformationandShortSalesProhibitionsbyPongsakHoontrakul♣,PeterJ.Ryan•andStylianosPerrakis♦Thisstudyexaminesthebehaviourofstockpricesinthepresenceofasymmetricinformation,whenmarketparticipantsareprohibitedfromshortselling.Wemodelthevolumebehaviourwhenhighorlowsignalsareobservedbyinsiders,whoareunabletosellunlesstheyownstock.Wedemonstratethatpricesadjusttogoodnewsonhighvolumeandtobadnewsonlowvolume;furthermore,weshowthatthespeedofadjustmentisgreateronbadnews,thatis,thepricefallsmorequicklytoalowerequilibriumpricethanitrisestothepricecorrespondingtogoodnews.Undermildconditionsonthemarketmicrostructure,thereisapositiveintertemporalcorrelationbetweenvolumeandreturn,withvolumeleading;additionalconditionsleadtothespeedofadjustmentresult.TheprimaryvolumeanomalyistestedbyanempiricalstudyofstockstradingontheStockExchangeofThailand,whereshortsalesareprohibited.July2001Donotquoteorreproducewithoutthepermissionoftheauthors.♣SasinInstitute,ChulalongkornUniversity,Bangkok,Thailand•FacultyofAdministration,UniversityofOttawa,Ottawa,Canada♦TheJohnMolsonSchoolofBusiness,ConcordiaUniversity,Montreal,CanadaTheVolume-ReturnRelationshipunderAsymmetryofInformationandShortSalesProhibitions2Insideinformationandrestrictionsontradingopportunitiesaregenerallyconsideredtobedetrimentaltomarketefficiency.Marketmakersreacttotheseconditionsastheyattempttosetpricesandquotespreadsfortheirstocks.Theirdecisionstoadjusttheirpricesandspreadsarebasedonobservationoftheorderstheyreceivefromunidentifiedtraderspotentiallyhavingsuperiorinformation.Boththetypeoforders,asbuysorsells,andthevolumeoftrading,eitherperorderorinaggregate,ispartiallyrevealingofthetruenatureofexistinginformation.Inaddition,theabsenceoforderspotentiallysignalstheoccurrenceofnews.Ifshortsellingisprohibited,theninformedtradersawareofbadnewswillonlybeabletoselliftheyalreadyownstockinsufficientquantitytosatisfytheirdesiretoprofitfromthenews.Hencetheobservationofnotradesmaybeindicativeofthepresenceofinformedwouldbesellers.Themarketmaker’sproblemislessenedifheknowsthecompositionofthemarketforthestockintermsofthepercentageofinformedtradersandoftheholdingsofinformedanduninformedinvestors.Thispermitsbetterestimatesoftheprobabilitiesthatbuys,sellsandnotradesarebeinggeneratedbyeitherinformedoruninformedtraders.ThispapermodifiesandextendstheEasleyandO’Hara[1992a]modelofsecuritiestradingunderasymmetricinformationbytheincorporationofshortsalesprohibitions.Duetotheseprohibitions,thepriceadjustmentprocessdiffersfromthatofEasleyandO’Hara;specifically,themarketmaker’srevisedprobabilitydistributionofthetruepriceisaffectedbyobservationofnotrades,ratherthanbythebuysandsellsalone.Becauseoftheasymmetricinterpretationgiventonotradeswhenshortsalesareprohibited,thereisalsoanasymmetricvolumereactiontogoodandbadnews.Wedemonstratethatundermildconditionsonthemarketstructure,thereisapositiveintertemporalcorrelationbetweenvolumeandreturn,withvolumeleading.Furthermore,additionalmicrostructureconditionsenableustoattributeafasterresponseinthepriceadjustmenttobadnewsthantogoodnews.Hence,wefindtheeffectofashortsaleprohibitiontobebeneficialtoefficiency,incontrasttopriorresults.Ingeneral,weshowthatthecombinationofinformationasymmetryandtheprohibitionofshortsalesprovidesanopportunityTheVolume-ReturnRelationshipunderAsymmetryofInformationandShortSalesProhibitions3togaininformationabouttheintertemporalpatternofstockreturns,therebyofferingajustificationfortechnicalanalysis.Academicliteraturecontainsnumerousattemptstoprovidestructuretothisproblemaspartofmarketmicrostructure.Kyle[1985]analysedorderflowincontinuoustimetoshowthatinformationisgraduallyincorporatedintopricesovertimeasinformedtraderscouldgeneratemoreprofitsfromcontinuoustradingthanstrategicentries.FosterandViswanathan[1990]modifiedthisapproachandfoundarelationshipbetweensizeofordersandtheactivityofamonopolistinsider,withtheresultthattheuninformedcouldgaininformationfromobservingthetradehistory.GlostenandMilgrom[1985]createdasequentialequilibriummodelwithpricesequaltothemarketmaker’sconditionalexpectationofassetvaluebasedontradeflow,findingvolumenottobeincreasinginthevarianceinprices.DiamondandVerrecchia[1987]extendedGlostenandMilgrom’smodeltoincludethreetypesofshortsaleconstraints.Theyexaminedwhethermarketshortsaleconstraintsaffectthetradingpropensity,therebycausingasymmetriesinthespeedofpriceadjustmenttogoodandbadnews.Theyconcludedthatashort-saleprohibitionreducedthespeedofpriceadjustmenttoinsideinformation.EasleyandO’Hara[1987]usedamodified,discretetimeGlostenandMilgromapproachtoincludedifferenttradesizesandinformationuncertainty,fromwhichthemarketmakerattemptedtodetermineboththeexistenceandnatureofnewinformation.EasleyandO’Hara[1992a](henceforthEO)focusedonthetradeprocessinsteadoftheusualpriceprocess,findingthateventuncertaintyprovidesaninformationalrolefortradingvolumegaineddirectlyfromthepropertiesoftheunderlyinginformationstructure,w