资源描述：
《Analysis of Cross Sectional Equity Models》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、AnalysisofCrossSectionalEquityModelsEdwardQian&RonaldHuaPutnamInvestmentsOnePostOfficeSquareBoston,MA02109February,20021.IntroductionGenerally,twoclassesofquantitativemodelsareusedinactiveportfoliomanagement:timeseriesandcross-sectionalmodels.Thetimeseriesmodelshaveproventobeveryusefulintheareasof
2、tacticalassetallocationandactivecurrencymanagement.Thisisunderstandableforatleasttworeasons.One,therearestrongeconometricbasisforforecastingassetandcurrencyreturn.Two,thereexistslengthycollectionofhistoricaldatainthoseareas–anecessaryingredientforbuildingtimeseriesmodels.However,bothoftheseconditi
3、onsareoftennotmetwhenitcomestoquantitativestockselection.Stockselectionisalmostallaboutstockspecificreturns,whichliesoutsideoftherealmofeconometricmodels.Formanycompanies,therejustisnotenoughtimeseriesdataavailable.Moreover,thesheernumberofstocksinatypicalportfolioshouldpreventanyseriousattempttob
4、uildtimeseriesmodelsonindividualstocks.TheDCFbasedmodels,ontheotherhand,requireearningprojectionanddiscountrateassumption,whicharemostlybasedonfundamentalanalysisratherthanquantitativemodeling.Forthepurposeofthispaper,wedon’tconsiderthemasquantitativemodels.Asaresult,quantitativeequityanalystsandp
5、ortfoliomanagersincreasinglyrelyuponcross-sectionalmodelsinthestockselectionprocess.Theintentofcross-sectionalmodelsisnottoaccuratelyforecastthefuturetotalreturnofindividualstocks,buttopredicttherelativeperformanceofallstocks.Oneadvantageofthesemodelsisflexibility.Onceconstructed,wecanapplythemtoa
6、lmostallcompaniesregardlessoftheirages,sincethemodelinputscomefrommostlycurrentvaluesoffactors.Nevertheless,theconstructionofthemodelsstillrequiresanadequatehistoryofcross-sectionaldata.Despiteitspopularity,thereseemstobealackofbetterunderstandingconcerningconstructionandperformanceofcross-section
7、alequitymodels.Onaverybasiclevel,itisintuitivetoeveryonethatabilitytorankfuturestockreturnsisofcrucialimportanceinacross-sectionperspective.Aquantitativeproxyforthisabilityiscross-sectionalinformationcoefficient(
