资源描述：
《A Generalized Approach to Portfolio Optimization Improving by Constraining Portfolio.》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、MANAGEMENTSCIENCEinforms®Vol.55,No.5,May2009,pp.798–812doi10.1287/mnsc.1080.0986issn0025-1909eissn1526-55010955050798©2009INFORMSAGeneralizedApproachtoPortfolioOptimization:ImprovingPerformancebyConstrainingPortfolioNormsVictorDeMiguelLondonBusinessSchool,LondonNW
2、14SA,UnitedKingdom,avmiguel@london.eduLorenzoGarlappiMcCombsSchoolofBusiness,UniversityofTexasatAustin,Austin,Texas78712,lorenzo.garlappi@mccombs.utexas.eduFranciscoJ.NogalesUniversidadCarlosIIIdeMadrid,28911Leganes,Madrid,Spain,fcojavier.nogales@uc3m.esRamanUppalLond
3、onBusinessSchool,LondonNW14SA,UnitedKingdom,ruppal@london.eduWeprovideageneralframeworkforfindingportfoliosthatperformwellout-of-sampleinthepresenceofestimationerror.Thisframeworkreliesonsolvingthetraditionalminimum-varianceproblembutsubjecttotheadditionalconstrainttha
4、tthenormoftheportfolio-weightvectorbesmallerthanagiventhreshold.WeshowthatourframeworknestsasspecialcasestheshrinkageapproachesofJagannathanandMa(Jagannathan,R.,T.Ma.2003.Riskreductioninlargeportfolios:Whyimposingthewrongconstraintshelps.J.Finance581651–1684)andLedoit
5、andWolf(Ledoit,O.,M.Wolf.2003.Improvedestimationofthecovariancematrixofstockreturnswithanapplicationtoportfolioselection.J.EmpiricalFinance10603–621,andLedoit,O.,M.Wolf.2004.Awell-conditionedestimatorforlarge-dimensionalcovariancematrices.J.MultivariateAnal.88365–411)
6、andthe1/NportfoliostudiedinDeMigueletal.(DeMiguel,V.,L.Garlappi,R.Uppal.2009.Optimalversusnaivediversification:Howinefficientisthe1/Nportfoliostrategy?Rev.FinancialStud.221915–1953).Wealsouseourframeworktoproposeseveralnewportfoliostrategies.Fortheproposedportfolios,wep
7、rovideamoment-shrinkageinterpretationandaBayesianinterpretationwheretheinvestorhasapriorbeliefonportfolioweightsratherthanonmomentsofassetreturns.Finally,wecompareempiricallytheout-of-sampleperformanceofthenewportfoliosweproposeto10strategiesintheliteratureacrossfiveda
8、tasets.Wefindthatthenorm-constrainedportfoliosoftenhaveahigherSharperatiothantheportfoliostrategiesinJagannathanandMa(2003),L
