items in Spiral are protected by copyright and are licensed under a Creative Commons Attribution NonCommercial NoDerivatives License.The Performance Persistence,thesefindingsareheavilycontingentonthestrategyfollowed:thepersistenceofDiscretionaryCTAsisdrivenbysmallfundswhereaslargefundsdrivetheperformancepersistenceofSystematicfunds.Theseresultshaveimportantimplicationsforinstitutionalinvestorswhofacecapitalallocationconstraints.Theyalsosuggestthatcontrarytothepreviousfindings,HWM,thevolatilityofreturnsislessabletopredictfailure.Fundsthatreceivelargerinflowsareabletosurvivelongerthanfundsthatdonot.LargeSystematicCTAshavethehighestprobabilityofsurvival.ThesecondpartstudiestheperformanceandperformancepersistenceofSystematicandDiscretionaryCTAs.Controllingforbiases,minimuminvestment,afterfeestheaverageCTAisabletoaddvalue.TheseresultsarestrongestforlargeSystematicCTAs.IextendthesevenfactormodelofFung-Hsieh(2004a)andfindthatthismodelisbetterabletoexplainthereturnsofSystematicratherthanDiscretionaryCTAs.IfindthreestructuralbreaksintheriskloadingsofCTAsdifferenttohedgefundbreaks:September1998,March2003andJuly2007.UsingthesebreaksIshowthatsystematicCTAswereabletodeliversignificantalphaineverysub-period.Ialsofindevidenceofsignificantperformancepersistence.However,performancepersistence,

ThePerformancePersistence,FlowandSurvivalofSystematicandDiscretionaryCommodityTradingAdvisors(CTAs)

12vs.8years.Ihandcollectinformationonreasonsforexitfromthedatabase.Iproposenewfiltersthatwillbetteridentifyrealfailuresamongfundsinthegraveyarddatabase.SeparatinggraveyardfundsintorealfailureIre-examinetheattritionrateofCTAs.Therealfailurerateis11.1%,fundageandlockup,theCTAindustrydoesnotappeartobeheadingtowardszeroalpha.Thefinalsectionlooksattherelationshipbetweenfund-flowsandperformance.Investorschasepastperformance,survivalandflowofCommodityTradingAdvisors,lowerthantheaverageyearlyattritionrateof17.3%ofCTAs.TheeffectofvariouscovariatesincludingseveraldownsideriskmeasuresisinvestigatedinpredictingCTAfailure.Controllingforperformance,Thisthesisstudiestheperformance,thereisnosmartmoneyeffectintheCTAliterature.IfindnoevidenceofcapacityconstraintsamongSystematicCTAs.InvestorsarethusnotabletosmartlyallocatefundstofuturebestperformersandtakefulladvantageoftheliquiditythatCTAsoffer.Unless otherwise indicated,alsoknownasCTAsorManagedFuturesFunds.OneofthemaincontributionsofthisthesisisthenovelclassificationofCTAstrategies.Thisisobtainedbyhand-collectinginformationfrequentlybydirectlycontactingthefundsinthedatabase.Ithusidentifytwomaintradingstyles:SystematicandDiscretionaryCTAswhicharethemainfocusofthisthesis.IfurtherseparateSystematicCTAsintotrend-followerswithdifferingtradinghorizon.ThisnoveldatasetallowsmetoreconsidermanyhithertostudiedissuesintheCTAspacewithanapplicationtothesesub-strategies.ThefirstsectioninvestigatesthedifferencesinmortalitybetweenSystematicandDiscretionaryCTAs,fundswithhigherdownsideriskmeasureshaveahigherhazardrate.Comparedtootherdownsideriskmeasures,overthelongesthorizonthanofanyintheliterature.AdetailedsurvivalanalysisoverthefullrangeofCTAstrategiesisprovided.SystematicCTAshaveahighermediansurvivalthanDiscretionaryCTAs.thefund-flow-performanceissignificantandconcaveforsomestrategies.Althoughthereissomelong-termperformancepersistenceofSystematicfundswiththehighestinflows,Flow and Survival of Systematic and Discretionary Commodity Trading Advisors (CTAs!