CFA三级知识点必备50：Equity Portfolio Management_打印版.docx

rEquityPortfolio'卜ManagementJCFA货漂笺亩笛VBobHong吗jX*3斯MM*5Iql1.Investmentstyle冷EquityInvestmentStyleClassificationsThetwomainapproachesinstyleanalysisaretheholdings-basedapproachandthereturns-basedapproachTheholdings-basedapproachlooksattheattributesofeachindividualstockinaportfolioandaggregatestheseattributestoconcludetheoverallstyleoftheportfolio.AcommonapplicationofthisideaistheMorningstarStyleBox.Holdings-BasedStyleAnalysisThestyleboxapproachaimstoclassifyapproximatelythesamenumberofstocksineachofthevalue,blend,andgrowthgroups,essentiallydistributingthemarketvalueofeachrowevenlyacrossthegrid.Theclassificationofstocksintovalue/blend/growthinvolvesassigningastylescoretoeachindividualstock.Forexample,toassignavaluescore,thedividendyieldmaybeused.Stockswouldberankedaccordingtotheirdividendyieldandascoreallocatedtoastockbasedontheirpercentileofthemarketvalueoftheirparticulargroup.Ifthestockisatthelower(high)endofthedividendyieldrange,itwillreceivealow(high)scoreclosetoO(100).Acomprehensivescoringmodelwouldusemanyindicatorsofvalueandcombinethemtogetherinapre-determinedweighting.4-14MH巨亚盅新tenHoldings-BasedStyleAnalysisTheclassificationofstocksThedifferencebetweenthestocksgrowthandvaluescoresiscalledanetstylescoreIfthenetstylescoreisstronglynegative,thestockisclassifiedasvalue.Ifthenetstylescoreisstronglypositivethenthestockisclassifiedasgrowth.Ifthenetstylescoreisclosetozerothenthestockwillbeclassifiedascore.Onceconstructedforastockuniverse,thegridcanbeusedasavisualaidtohelpcategorizeandtrackmanagedinvestmentportfolios.Ataglance,aninvestorcanseewhereamanagerispositionedonthegrid,and,ifhistoricaldataexists,howthisstylehaschangedovertime.5-14HHHHHIM亚&新1值一Returns-BasedStyleAnalysisAreturns-basedstyleanalysisaimstoidentifythestyleofafundthroughregressionofthefundsreturnsagainstasetofpassivestyleindicesByimposingaconstraintontheregressionthatthesumoftheslopecoefficientsshouldsumtoavalueofI1theslopecoefficientscanbeinterpretedasthemanagersallocationtothatstyleduringtheperiod.Forexample,areturn-basedstyleanalysismightconductaregressionoffundreturnsversusfourpassiveindicesasfollows:mt=a+ZSS&+£RSt=thereturnofstyleindexsinthesameperiods=thefundexposuretostyles(withconstraintsEmS=Q=1ands>0foralong-onlyportfolio)=aconstantofteninterpretedasthevalueaddedbythefundmanager产theresidualreturnthatcannotbeexplainedbythestylesusedintheanalysis6-14HB巨业.色新！811_ManagerSelf-IdentificationThefund,sinvestmentstrategyisusuallyself-describedbythemanager.Comparingthatself-descriptionwithreturns-basedandholdings-basedstyleanalysiswilleitherconfirmaconsistentidentificationorindicateaneedforfurtherinvestigationandanalysistoexplainthediscrepancy.Somestylessuchasequitylong/short,equitymarketneutralandshortbiasdonotfittraditionalstylecategoriesandthemanagersdescriptionandfundprospectusbecomesthekeysourceofinformationonstyleofsuchfunds.7-14HM亚鼻新.t1l_Equitystyleanalysis-Comparisonof2tech.AdvantagesDisadvantagesReturn-basedRequiresminimalinformationCanbeexecutedquicklyCosteffectiveMorewidelyappliedMaybeineffectiveincharacterizingcurrentstyleDifficulttodetectmoreaggressivepositionsHolding-basedMoreaccuratethanreturns-basedFacilitatescomparisonsofindividualpositionsCapturechangesinstylemorequicklyMoredataintensivethanreturns-basedanalysisLesseffectiveforfundswithsubstantialpositionsinderivatives.2.PortfolioConstructionPortfolioConstructionPassively-managedindex-basedequityportfolioscanbeconstructedby:FullReplication:fullreplication(holdallofthesecuritiesintheindex)StratifiedSampling:holdasampleofthesecuritiesbasedonstratifiedsamplingOptimization:usemorecomplexoptimizationtomaximizedesirablecharacteristicswhileminimizingundesirablecharacteristics.10-14BlendedApproach:inpracticeablendoftheseapproachesmaybeused.行业&iwitriPortfolioConstructionFullreplicationcanbecostlywhentherearelargenumbersofstockandliquidityislimited.Theportfoliomustberegularlyreconstitutedandrebalanced.Number of Secut) HeM11-14TheadvantageoffullreplicationisthatitcloselymatchestheindexM亚&新！Tl_PortfolioConstructionToavoidthehighcostoffullreplication,itoftenmakesmoresenseforthemanagertousestratifiedsampling,inwhichheholdsasubsetoftheconstituentstocks,withthesampleselectedinsuchawayastoreplicatetheindexreturn/riskcharacteristics.Toimplementstratifiedsampling,themanagercreatesstrataacrosstheconstituentstocksthataremutuallyexclusiveandexhaustiveThemanagermustconsidersizeofthesampleused.Asmorestocksareaddedandtheportfolioapproachesfullreplication;lessliquidstocksareadded,increasingtransactioncostandtrackingerror.PortfolioConstructionOptimizationusesthetoolsofmodernportfoliotheorytoaddresstheproblemofminimizingtrackingerror.Theoptimizerseeksthecombinationofstocksthatwouldhaveminimizedtrackingerrorandpossiblymaximizedreturn.Theadvantagesofoptimizationtechniquesistheytypicallyexhibitlowertrackingerror,andthattheyexplicitlyaccountforthecovarianceamongconstituentstocks.Theobviousdrawbackofoptimizationisthatitisbasedonhistoricalrelationshipsandthosecanchange.Maintaintheoptimizationasthedatachangecanbecostly.Anotherdrawbackisthatitcancreateportfoliosthatarenotmeanvarianceefficientrelativetothebenchmark.Thesolutionistoaddaconstraintthattotalportfoliovarianceisequaltothevolatilityofthebenchmark.13-14HHH目亚.色新HBtIPortfolioConstructionBlendedApproachFullreplicationispreferredforindexeswithsmallnumbersofliquidstocks,whilestratifiedsamplingoroptimizationispreferableforindexeswithlotsofheterogeneous,thinlytradedstocks.14-14ForlargeindexesliketheWilshire5000rtheconstituentstocksrunthegamutfromlargeandliquidtosmallandthinlytraded.Inthatcaseacombinationoftwoapproaches,fullreplicationandstratifiedsamplingoroptimization.雪业1«MH.