基金管理外文文献翻译.docx

基金管理外文文献翻译.docx

《基金管理外文文献翻译.docx》由会员分享，可在线阅读，更多相关《基金管理外文文献翻译.docx（10页珍藏版）》请在冰点文库上搜索。

基金管理外文文献翻译

基金管理外文文献翻译

（含：

英文原文及中文译文）

文献出处：

英文原文

IsMoneyReally“Smart”?

NewEvidenceontheRelationBetweenMutualFundFlows,ManagerBehavior,andPerformancePersistence

RussWermers

Mutualfundreturnsstronglypersistovermulti-yearperiods—thatisthecentralfindingofthispaper.Further,consumerandfundmanagerbehaviorbothplayalargeroleinexplainingtheselongtermcontinuationpatterns—consumersinvestheavilyinlast-year’swinningfunds,andmanagersofthesewinnersinvesttheseinflowsinmomentumstockstocontinuetooutperformotherfundsforatleasttwoyearsfollowingtherankingyear.Bycontrast,managersoflosingfundsappearreluctanttoselltheirlosingstockstofinancethepurchaseofnewmomentumstocks,perhapsduetoadispositioneffect.Thus,momentumcontinuestoseparatewinningfromlosingmanagersforamuchlongerperiodthanindicatedbypriorstudies.

Evenmoresurprisingisthatpersistenceinwinningfundreturnsisnotentirelyexplainedbymomentum—wefindstrongevidencethatflow-relatedbuying,especiallyamonggrowth-orientedfunds,pushesupstockprices.Specifically,stocksthatwinningfundspurchaseinresponsetopersistentflowshavereturnsthatbeattheirsize,book-to-market,andmomentumbenchmarksbytwotothreepercentperyearoverafour-yearperiod.Cross-sectionalregressionsindicatethattheseabnormalreturnsarestronglyrelatedtofundinflows,butnottothepastperformanceofthefunds—thus,castingsomedoubtonpriorfindingsofpersistentmanagertalentinpickingstocks.Finally,atthestyle-adjustednetreturnslevel,wefindnopersistence,consistentwiththeresultsofpriorstudies.Onbalance,weconfirmthatmoneyissmartinchasingwinningmanagers,butthata“copycat”strategyofmimickingwinningfundstocktradestotakeadvantageofflow-relatedreturnsappearstobethesmarteststrategy.

Eighty-eightmillionindividualsnowholdinvestmentsinU.S.mutualfunds,withover90percentofthevalueoftheseinvestmentsbeingheldinactivelymanagedfunds.Further,activelymanagedequityfundsgainthelion’sshareofconsumerinflows—flowsofnetnewmoneytoequityfunds（inflowsminusoutflows）totalled$309billionin2000,pushingtheaggregatevalueofinvestmentsheldbythesefundstoalmost$4trillionatyear-end2000.Whilethemajorityofindividualinvestorsapparentlybelieveinthevirtuesofactivemanagementingeneral,manyappeartoholdevenstrongerbeliefsconcerningthetalentsofsubgroupsoffundmanagers—theyappeartobelievethat,amongthefieldofactivemanagers,superiormanagersexistthatcan“beatthemarket”forlongperiodsoftime.Inparticular,MorningstarandLippercompetevigorouslyfortheattentionofthesetruebelieversbyprovidingregularfundperformancerankings,whilepopularpublicationssuchasMoneyMagazineroutinelyprofile“star”mutualfundmanagers.Inaddition,investordollars,whilenotveryquicktoabandonpastlosingfunds,aggressivelychasepastwinners（see,forexample,SirriandTufano（1998））.

Arethese“performance-chasers”wastingtheirmoneyandtime,orismoney“smart”?

Severalpastpapershaveattemptedtotacklethisissue,withsomewhatdifferingresults.Forexample,GrinblattandTitman（1989a,1993）findthatsomemutualfundmanagersareabletoconsistentlyearnpositiveabnormalreturnsbeforefeesandexpenses,whileBrownandGoetzmann（1995;BG）attributepersistencetoinferiorfundsconsistentlyearningnegativeabnormalreturns.Gruber（1996）andZheng（1999）examinepersistencefromtheviewpointofconsumermoneyflowstofunds,andfindthatmoneyis“smart”—thatis,moneyflowsdisproportionatelytofundsexhibitingsuperiorfuturereturns.However,theexactsourceofthesmartmoneyeffectremainsapuzzle—doessmartmoneycapturemanagertalentor,perhaps,simplymomentuminstockreturns?

1Morerecently,Carhart（1997）examinesthepersistenceinnetreturnsofU.S.mutualfunds,controllingforthecontinuationattributabletopricedequitystyles（see,forexample,FamaandFrench（1992,1993,1996）,JegadeeshandTitman（1993）,DanielandTitman（1997）,andMoskowitzandGrinblatt（1999））.Carhartfindslittleevidenceofsuperiorfundsthatconsistentlyoutperformtheirstylebenchmarks—specifically,Carhartfindsthatfundsinthehighestnetreturndecile（oftheCRSPmutualfunddatabase）duringoneyearbeatfundsinthelowestdecilebyabout3.5percentduringthefollowingyear,almostallduetotheone-yearmomentumeffectdocumentedbyJegadeeshandTitman（1993）andtotheunexplainedpoorperformanceoffundsinthelowestprior-yearreturndecile.2Thus,Carhart（1997）suggeststhatmoneyisnotverysmart.RecentstudiesfindsomewhatmorepromisingresultsthanCarhart（1997）.Chen,Jegadeesh,andWermers（1999）findthatstocksmostactivelypurchasedbyfundsbeatthosemostactivelysoldbyovertwopercentperyear,whileBollenandBusse（2002）findevidenceofpersistenceinquarterlyfundperformance.Wermers（2000）findsthat,althoughtheaveragestyle-adjustednetreturnoftheaveragemutualfundisnegative（consistentwithCarhart’sstudy）,high-turnoverfundsexhibitanetreturnthatissignificantlyhigherthanlow-turnoverfunds.Inaddition,thesehighturnoverfundspickstockswellenoughtocovertheircosts,evenadjustingforstyle-basedreturns.Thisfindingsuggeststhatfundmanagerswhotrademorefrequentlyhavepersistentstockpickingtalents.Allofthesepapersprovideamorefavorableviewoftheaverageactivelymanagedfundthanpriorresearch,althoughnonefocusonthepersistenceissuewithportfolioholdingsdata.Thisstudyexaminesthemutualfundpersistenceissueusingbothportfolioholdingsandnetreturnsdata,allowingamorecompleteanalysisoftheissuethanpaststudies.Withthesedata,wedevelopmeasuresthatallowustoexaminetherolesofconsumerinflowsandfundmanagerbehaviorinthepersistenceoffundperformance.Specifically,wedecomposethereturnsandcostsofeachmutualfundintothatattributableto

（1）managerskillsinpickingstockshavingreturnsthatbeattheirstyle-basedbenchmarks（selectivity）,

（2）returnsthatareattributabletothecharacteristics（orstyle）ofstockholdings,（3）tradingcosts,（4）expenses,and（5）coststhatareassociatedwiththedailyliquidityofferedbyfundstotheinvestingpublic（asdocumentedbyEdelen（1999））.Further,weconstructholdings-basedmeasuresofmomentum-investingbehaviorbythefundmanagers.Together,thesemeasuresallowanexaminationoftherelationbetweenflows,managerbehavior,andperformancepersistence.

Inrelatedwork,SirriandTufano（1998）findthatconsumerflowsreactaboutasstronglytoone-yearlaggednetreturnsastoanyotherfundcharacteristic.Inaddition,themodelofLynchandMusto（2002）predictsthatperformancerepeatsamongwinners（butnotlosers）,whilethemodelofBerkandGreen（2002）predictsnopersistence（orweakpersistence）asconsumerflowscompeteawayanymanagerialtalent.ConsistentwithSirriandTufano（1998）,andtotestthecompetingviewpointsofLynchandMusto（2002）andBerkandGreen（2002）,wesortfundsontheirone-yearlaggednetreturnsformosttestsinthispaper.Whileotherwaysofsortingfundsareattempted.

Data

Wemergetwomajormutualfunddatabasesforouranalysisofmutualfundperformance.DetailsontheprocessofmergingthesedatabasesisavailableinWermers（2000）.ThefirstdatabasecontainsquarterlyportfolioholdingsforallU.S.equitymutualfundsexistingatanytimebetweenJanuary1,1975andDecember31,1994;thesedatawerepurchasedfromThomson/CDAofRockville,Maryland.TheCDAdatasetliststheequityportionofeachfund’sholdings（i.e.,theshareholdingsofeachstockheldbythatfund）alongwithalistingofthetotalnetassetsundermanagementandtheself-declaredinvestmentobjectiveatthebeginningofeachcalendarquarter.CDAbegancollectinginvestment-objectiveinformationonJune30,1980;wesupplementthesedatawithhand-collectedinvestmentobjectivedatafromJanuary1,1975.

ThesecondmutualfunddatabaseisavailablefromtheCenterforResearchinSecurityPrices（CRSP）andisusedbyCarhart（1997）.TheCRSPdatabasecontainsmonthlydataonnetreturns,aswellasannualdataonportfolioturnoverandexpenseratiosforallmutualfundsexistingatanytimebetweenJanuary1,1962andDecember31,2000.FurtherdetailsontheCRSPmutualfunddatabaseareavailablefromCRSP.

Thesetwodatabasesweremergedtoprovideacompleterecordofthestockholdingsofagivenfund,alongwiththefund’sturnover,expenseratio,netreturns,investmentobjective,andtotalnetassetsundermanagementduringtheentiretimethatthefundexistedduringourtheperiodof1975to1994（inclusive）.5Finally,stockpricesandreturnswereobtainedfromtheCRSPstockfiles.

Performance-DecompositionMethodologyInthisstudy,weuseseveralmeasuresthatquantifytheabilityofamutualfundmanagertochoosestocks,aswellastogeneratesuperiorperformanceatthenetreturnlevel.Thesemeasures,ingeneral,decomposethereturnofthestocksheldbyamutualfundintoseveralcomponentsinordertobothbenchmarkthestockportfolioandtoprovideaperformanceattributionforthefund.Themeasuresusedtodecomposefundreturnsinclude:

1.theportfolio-weightedreturnonstockscurrentlyheldbythefund,inexcessofreturns（duringthesametimeperiod）onmatchedcontrolportfolioshavingthesamestylecharacteristics（selectivity）

2.theportfolio-weightedreturnoncontrolportfolioshavingthesamecharacteristicsasstockscurrentlyheldbythefund,inexcessoftime-seriesaveragereturnsonthosecontrolportfolio