《CFA三级知识点必备50：Equity Portfolio Management_打印版.docx》由会员分享，可在线阅读，更多相关《CFA三级知识点必备50：Equity Portfolio Management_打印版.docx（6页珍藏版）》请在三一办公上搜索。
1、rEquityPortfolio卜ManagementJCFA货漂笺亩笛VBobHong吗jX*3斯MM*5Iql1.Investmentstyle冷EquityInvestmentStyleClassificationsThetwomainapproachesinstyleanalysisaretheholdings-basedapproachandthereturns-basedapproachTheholdings-basedapproachlooksattheattributesofeachindividualstockinaportfolioandaggregatestheseatt
2、ributestoconcludetheoverallstyleoftheportfolio.AcommonapplicationofthisideaistheMorningstarStyleBox.Holdings-BasedStyleAnalysisThestyleboxapproachaimstoclassifyapproximatelythesamenumberofstocksineachofthevalue,blend,andgrowthgroups,essentiallydistributingthemarketvalueofeachrowevenlyacrossthegrid.T
3、heclassificationofstocksintovalue/blend/growthinvolvesassigningastylescoretoeachindividualstock.Forexample,toassignavaluescore,thedividendyieldmaybeused.Stockswouldberankedaccordingtotheirdividendyieldandascoreallocatedtoastockbasedontheirpercentileofthemarketvalueoftheirparticulargroup.Ifthestockis
4、atthelower(high)endofthedividendyieldrange,itwillreceivealow(high)scoreclosetoO(100).Acomprehensivescoringmodelwouldusemanyindicatorsofvalueandcombinethemtogetherinapre-determinedweighting.4-14MH巨亚盅新tenHoldings-BasedStyleAnalysisTheclassificationofstocksThedifferencebetweenthestocksgrowthandvaluesco
5、resiscalledanetstylescoreIfthenetstylescoreisstronglynegative,thestockisclassifiedasvalue.Ifthenetstylescoreisstronglypositivethenthestockisclassifiedasgrowth.Ifthenetstylescoreisclosetozerothenthestockwillbeclassifiedascore.Onceconstructedforastockuniverse,thegridcanbeusedasavisualaidtohelpcategori
6、zeandtrackmanagedinvestmentportfolios.Ataglance,aninvestorcanseewhereamanagerispositionedonthegrid,and,ifhistoricaldataexists,howthisstylehaschangedovertime.5-14HHHHHIM亚&新1值一Returns-BasedStyleAnalysisAreturns-basedstyleanalysisaimstoidentifythestyleofafundthroughregressionofthefundsreturnsagainstase
7、tofpassivestyleindicesByimposingaconstraintontheregressionthatthesumoftheslopecoefficientsshouldsumtoavalueofI1theslopecoefficientscanbeinterpretedasthemanagersallocationtothatstyleduringtheperiod.Forexample,areturn-basedstyleanalysismightconductaregressionoffundreturnsversusfourpassiveindicesasfoll
8、ows:mt=a+ZSS&+RSt=thereturnofstyleindexsinthesameperiods=thefundexposuretostyles(withconstraintsEmS=Q=1ands0foralong-onlyportfolio)=aconstantofteninterpretedasthevalueaddedbythefundmanager产theresidualreturnthatcannotbeexplainedbythestylesusedintheanalysis6-14HB巨业.色新！811_ManagerSelf-IdentificationThe
9、fund,sinvestmentstrategyisusuallyself-describedbythemanager.Comparingthatself-descriptionwithreturns-basedandholdings-basedstyleanalysiswilleitherconfirmaconsistentidentificationorindicateaneedforfurtherinvestigationandanalysistoexplainthediscrepancy.Somestylessuchasequitylong/short,equitymarketneut
10、ralandshortbiasdonotfittraditionalstylecategoriesandthemanagersdescriptionandfundprospectusbecomesthekeysourceofinformationonstyleofsuchfunds.7-14HM亚鼻新.t1l_Equitystyleanalysis-Comparisonof2tech.AdvantagesDisadvantagesReturn-basedRequiresminimalinformationCanbeexecutedquicklyCosteffectiveMorewidelyap
11、pliedMaybeineffectiveincharacterizingcurrentstyleDifficulttodetectmoreaggressivepositionsHolding-basedMoreaccuratethanreturns-basedFacilitatescomparisonsofindividualpositionsCapturechangesinstylemorequicklyMoredataintensivethanreturns-basedanalysisLesseffectiveforfundswithsubstantialpositionsinderiv
12、atives.2.PortfolioConstructionPortfolioConstructionPassively-managedindex-basedequityportfolioscanbeconstructedby:FullReplication:fullreplication(holdallofthesecuritiesintheindex)StratifiedSampling:holdasampleofthesecuritiesbasedonstratifiedsamplingOptimization:usemorecomplexoptimizationtomaximizede
13、sirablecharacteristicswhileminimizingundesirablecharacteristics.10-14BlendedApproach:inpracticeablendoftheseapproachesmaybeused.行业&iwitriPortfolioConstructionFullreplicationcanbecostlywhentherearelargenumbersofstockandliquidityislimited.Theportfoliomustberegularlyreconstitutedandrebalanced.Number of
14、 Secut) HeM11-14TheadvantageoffullreplicationisthatitcloselymatchestheindexM亚&新！Tl_PortfolioConstructionToavoidthehighcostoffullreplication,itoftenmakesmoresenseforthemanagertousestratifiedsampling,inwhichheholdsasubsetoftheconstituentstocks,withthesampleselectedinsuchawayastoreplicatetheindexreturn
15、/riskcharacteristics.Toimplementstratifiedsampling,themanagercreatesstrataacrosstheconstituentstocksthataremutuallyexclusiveandexhaustiveThemanagermustconsidersizeofthesampleused.Asmorestocksareaddedandtheportfolioapproachesfullreplication;lessliquidstocksareadded,increasingtransactioncostandtrackin
16、gerror.PortfolioConstructionOptimizationusesthetoolsofmodernportfoliotheorytoaddresstheproblemofminimizingtrackingerror.Theoptimizerseeksthecombinationofstocksthatwouldhaveminimizedtrackingerrorandpossiblymaximizedreturn.Theadvantagesofoptimizationtechniquesistheytypicallyexhibitlowertrackingerror,a
17、ndthattheyexplicitlyaccountforthecovarianceamongconstituentstocks.Theobviousdrawbackofoptimizationisthatitisbasedonhistoricalrelationshipsandthosecanchange.Maintaintheoptimizationasthedatachangecanbecostly.Anotherdrawbackisthatitcancreateportfoliosthatarenotmeanvarianceefficientrelativetothebenchmar
18、k.Thesolutionistoaddaconstraintthattotalportfoliovarianceisequaltothevolatilityofthebenchmark.13-14HHH目亚.色新HBtIPortfolioConstructionBlendedApproachFullreplicationispreferredforindexeswithsmallnumbersofliquidstocks,whilestratifiedsamplingoroptimizationispreferableforindexeswithlotsofheterogeneous,thinlytradedstocks.14-14ForlargeindexesliketheWilshire5000rtheconstituentstocksrunthegamutfromlargeandliquidtosmallandthinlytraded.Inthatcaseacombinationoftwoapproaches,fullreplicationandstratifiedsamplingoroptimization.雪业1MH.
