Revisiting The Capital Asset Pricing Model.docx

Revisiting The Capital Asset Pricing Model.docx

《Revisiting The Capital Asset Pricing Model.docx》由会员分享，可在线阅读，更多相关《Revisiting The Capital Asset Pricing Model.docx（11页珍藏版）》请在冰豆网上搜索。

RevisitingTheCapitalAssetPricingModel

RevisitingTheCapitalAssetPricingModel

byJonathanBurton

ReprintedwithpermissionfromDowJonesAssetManager

May/June1998,pp.20-28

Forpicturesandcaptions,clickhere

ModernPortfolioTheorywasnotyetadolescentin1960whenWilliamF.Sharpe,a26-year-oldresearcherattheRANDCorporation,athinktankinLosAngeles,introducedhimselftoafelloweconomistnamedHarryMarkowitz..Neitherofthemknewitthen,butthatcasualknockonMarkowitz'sofficedoorwouldforeverchangehowinvestorsvaluedsecurities.

Sharpe,thenaPh.D.candidateattheUniversityofCalifornia,LosAngeles,neededadoctoraldissertationtopic.Hehadread"PortfolioSelection,"Markowitz'sseminalworkonriskandreturn—firstpublishedin1952andupdatedin1959—thatpresentedaso-calledefficientfrontierofoptimalinvestment.Whileadvocatingadiversifiedportfoliotoreducerisk,Markowitzstoppedshortofdevelopingapracticalmeanstoassesshowvariousholdingsoperatetogether,orcorrelate,thoughthequestionhadoccurredtohim.

SharpeacceptedMarkowitz'ssuggestionthatheinvestigatePortfolioTheoryasathesisproject.Byconnectingaportfoliotoasingleriskfactor,hegreatlysimplifiedMarkowitz'swork.Sharpehascommittedhimselfeversincetomakingfinancemoreaccessibletobothprofessionalsandindividuals.

Fromthisresearch,Sharpeindependentlydevelopedahereticalnotionofinvestmentriskandreward,asophisticatedreasoningthathasbecomeknownastheCapitalAssetPricingModel,ortheCAPM.TheCAPMrattledinvestmentprofessionalsinthe1960s,anditscommandingimportancestillreverberatestoday.In1990,Sharpe'sroleindevelopingtheCAPMwasrecognizedbytheNobelPrizecommittee.SharpesharedtheNobelMemorialPrizeinEconomicSciencesthatyearwithMarkowitzandMertonMiller,theUniversityofChicagoeconomist.

Everyinvestmentcarriestwodistinctrisks,theCAPMexplains.Oneistheriskofbeinginthemarket,whichSharpecalledsystematicrisk.Thisrisk,laterdubbed"beta,"cannotbediversifiedaway.Theother—unsystematicrisk—isspecifictoacompany'sfortunes.Sincethisuncertaintycanbemitigatedthroughappropriatediversification,Sharpefiguredthataportfolio'sexpectedreturnhingessolelyonitsbeta—itsrelationshiptotheoverallmarket.TheCAPMhelpsmeasureportfolioriskandthereturnaninvestorcanexpectfortakingthatrisk.

MorethanthreedecadeshavepassedsincetheCAPM'sintroduction,andSharpehasnotstoodstill.AprofessoroffinanceattheStanfordUniversityGraduateSchoolofBusinesssince1970,hehascraftedseveralfinancialtoolsthatportfoliomanagersandindividualsuseroutinelytobettercomprehendinvestmentrisk,includingreturns-basedstyleanalysis,whichassistsinvestorsindeterminingwhetheraportfoliomanagerisstickingtohisstatedinvestmentobjective.TheSharperatioevaluatesthelevelofriskafundacceptsvs.thereturnitdelivers.

Sharpe'slatestprojectischaracteristicallyambitious,combininghisdesiretoeducateamassaudienceaboutriskwithhislongtimeloveofcomputers.Technologyisdemocratizingfinance,andSharpeishelpingtopushthispowerfulrevolutionforward.ThroughFinancialEngines,SharpeandhispartnerswillbringprofessionalinvestmentadviceandanalysistoindividualsovertheInternet.

Whatdoyouthinkofthetalkthatbetaisdead?

TheCAPMisnotdead.Anyonewhobelievesmarketsaresoscrewythatexpectedreturnsarenotrelatedtotheriskofhavingabadtime,whichiswhatbetarepresents,musthaveaveryharshviewofreality.

"Isbetadead?

"isreallyfocusedonwhetherornotindividualstockshavehigherexpectedreturnsiftheyhavehigherbetasrelativetothemarket.Itwouldbeirresponsibletoassumethatisnottrue.Thatdoesn'tmeanwecanconfirmthedata.Wedon'tseeexpectedreturns;weseerealizedreturns.Wedon'tseeex-antemeasuresofbeta;weseerealizedbeta.Whatmakesinvestmentsinterestingandexcitingisthatyouhavelotsofnoiseinthedata.Soit'shardtodefinitivelyanswerthesequestions.

Wouldyouapproachastudyofmarketriskdifferentlytodaythanyoudidbackintheearly1960s?

It'sfunnyhowpeopletendtomisunderstandtheCAPM'sacademic,theoreticalandscientificprocess.TheCAPMwasaverysimple,verystrongsetofassumptionsthatgotanice,clean,prettyresult.Andthenalmostimmediately,weallsaid,let'sbringmorecomplexityintoittotrytogetclosertotherealworld.Peoplewenton—myselfandothers—towhatIcall"extended"capitalassetpricingmodels,inwhichexpectedreturnisafunctionofbeta,taxes,liquidity,dividendyield,andotherthingspeoplemightcareabout.

DidtheCAPMevolve?

Ofcourse.AretheresultsmorecomplicatedshalljustexpectedreturnisalinearfunctionofbetarelativetotheStandard&Poor's500-StockIndex?

Ofcourse.Butthefundamentalidearemainsthatthere'snoreasontoexpectrewardjustforbearingrisk.Otherwise,you'dmakealotofmoneyinLasVegas.Ifthere'srewardforrisk,it'sgottobespecial.There'sgottobesomeeconomicsbehinditorelsetheworldisaverycrazyplace.Idon'tthinkdifferentlyaboutthosebasicideasatall.

WhataboutHarryMarkowitz'scontributiontoallofthis?

Markowitzcamealong,andtherewaslight.Markowitzsaidaportfoliohasexpectedreturnandrisk.Expectedreturnisrelatedtotheexpectedreturnofthesecurities,butriskismorecomplicated.Riskisrelatedtotherisksoftheindividualcomponentsaswellasthecorrelations.

Thatmakesriskacomplicatedfeature,andonethathumanbeingshavetroubleprocessing.Youcanputestimatesofrisk/returncorrelationintoacomputerandfindefficientportfolios.Inthisway,youcangetmorereturnforagivenriskandlessriskforagivenreturn,andthat'sefficiencyalaMarkowitz.

WhatstandsoutinyourmindwhenyouthinkaboutMarkowitz'scontribution?

Ilikedtheparsimony,thebeauty,ofit.Iwasandamacomputernut.Ilovedthemathematics.Itwassimplebutelegant.Ithadalloftheaestheticqualitiesthatamodelbuilderlikes.Investmenttextsinthepre-Markowitzeraweresimplistic:

Don'tputallyoureggsinonebasket,orputtheminabasketandwatchitclosely.Therewaslittlequantification.

Tothisday,peoplerecommendacompartmentalizedapproach.Youhaveonepotforyourcollegefund,anotherforyourretirementfund,anotherforyourunemploymentfund.People'stendencieswhentheydealwiththeseissuesoftenleadtosuboptimalsolutionsbecausetheydon'ttakecovarianceintoaccount.Correlationisimportant.Youwanttothinkabouthowthingsmovetogether.

TellusaboutyourrelationshipwithMarkowitz.

Harrywasmyunofficialdissertationadvisor.In1960,heandIwerebothattheRANDCorporation.MyofficialadvisorattheUniversityofCaliforniaatLosAngelessuggestedIworkwithHarry,butHarrywasn'tontheUCLAfaculty.IintroducedmyselftohimandsaidIwasagreatfanofhiswork.

WithMarkowitz'sencouragement,youdelvedintomarketcorrelation,streamliningPortfolioTheorywiththeuseofasingle-factormodel.Thisbecamepartofyourdissertation,publishedin1963as"ASimplifiedModelofPortfolioAnalysis."

Ididmydissertationunderastronglysimplifiedassumptionthatonlyonefactorcausedcorrelation.TheresultIgotwasinthatsetting,priceswouldadjustuntilexpectedreturnswerehigherforsecuritiesthathadhigherbetas,wherebetawasthecoefficientwith"thefactor."

PortfolioTheoryfocusedontheactionsofasingleinvestorwithanoptimalportfolio.YouwonderedwhatwouldhappentoriskandreturnifeveryonefollowedMarkowitzandbuiltefficientportfolios.

Isaidwhatifeveryonewasoptimizing?

They'veallgottheircopiesofMarkowitzandthey'redoingwhathesays.ThensomepeopledecidetheywanttoholdmoreIBM,buttherearen'tenoughsharestosatisfydemand.SotheyputpricepressureonIBMandupitgoes,atwhichpointtheyhavetochangetheirestimatesofriskandreturn,becausenowthey'repayingmoreforthestock.Thatprocessofupwardanddownwardpressureonpricescontinuesuntilpricesreachanequilibriumandeveryonecollectivelywantstoholdwhat'savailable.Atthatpoint,whatcanyousayabouttherelationshipbetweenriskandreturn?

Theansweristhatexpectedreturnisproportionatetobetarelativetothemarketportfolio.

InapaperIfinishedin1962thatwaspublishedin1964,Ifoundyoudidn'thavetoassumeonlyonefactor.Thatbasicresultcomesthroughinamuchmoregeneralsetting.Therecouldbefivefactors,or20factors,orasmanyfactorsastherearesecurities.InaMarkowitzframework,wherepeoplecareabouttheexpectedreturnoftheirportfoliosandtheriskasmeasuredbystandarddeviationtheresultsheld.Thatpaperwascalled"CapitalAssetPrices:

ATheoryofMarketEquilibriumUnderConditionsOfRisk."EugeneFamacalledittheCapitalAssetPricingModel.That'swherethenamecamefrom.

TheCAPMwasandisatheoryofequilibrium.Whyshouldanyoneexpecttoearnmorebyinvestinginonesecurityasopposedtoanother?

Youneedtobecompensatedfordoingbadlywhentimesarebad.Thesecuritythatisgoingtodobadlyjustwhenyouneedmoneywhentimesarebadisasecurityyouhavetohate,andtherehadbetterbesomeredeemingvirtueorelsewhowillholdit?

Thatredeemingvirtuehastobethatinnormaltimesyouexpecttodobetter.ThekeyinsightoftheCapitalAssetPricingModelisthathigherexpectedreturnsgowiththegreaterriskofdoingbadlyinbadtimes.Betaisameasureofthat.Securitiesorassetclasseswithhighbetastendtodoworseinbad