In this note we represent the well known discrete time stochastic volatility (SV) model with a leverage effect and the SV model of Jacquier, Polson and Rossi (JPR) (2002) using Gaussian nonlinear state space forms with uncorrelated measurement and transiti
5SimulationResults
SinceJPR’sspeci cationoftheleveragee ectisneithertheoreticallyappealingnorem-piricallysupported,thesamplingpropertiesofBayesestimatorsreportedintheirpaperarelesspracticallyrelevant.AlthoughthesamplingpropertiesofMCMCestimatesofthecontinuoustimeSVmodelwiththeleveragee ectareexaminedinEraker,JohannesandPolson(2002),tothebestofourknowledge,thesamplingpropertiesofMCMCes-timatesofthediscretetimeSVmodelwithacorrectlyspeci edleveragee ectremainunknown.Ontheotherhand,understandingthe nitesampleperformanceofMCMCestimatesisimportantfromseveralaspects.First,itprovidesthereliabilityofMCMCestimatesoftheA-SVmodels,inparticularofthenewparameter,ρ.Second,sincemanymoreestimationtoolshavebeendevelopedtoestimatethediscretetimeA-SVmodelthantothecontinuoustimeA-SVmodel,itisinterestingtocomparedirectlytheperformanceofMCMCestimateswithotherestimatesinthediscretetimecontext.Inthissection,samplingexperimentsaredesignedtoobtainsamplingpropertiesoftheproposedMCMCestimatesintheconventionallyspeci eddiscretetimeA-SVmodel.3Inthe rstexperimentweusethesimilarparametersettingasinJPR.Wesimulate100samplesof1000observationsfromtheA-SVmodel(2.2).Simulationresultssuchasthesampleaverageandsamplerootmeansquareerror(RMSE)aregiveninTable3.TheevidenceproposedinTable3showsthatBayesestimateshaveverygoodsamplingproperties.
InthesecondexperimentweadoptaparametersettingasinHarveyandShephard(1996)thatenablesustocomparetherelativee ciencyofourBayesestimatetothequasi-maximumlikelihood(QML)estimateofHarveyandShephard(1996).Table4reportsthemeansandRMSEsofalltheestimates.ThesimulationresultsfortheQMLestimatesareobtaineddirectlyfromHarveyandShephard(1996).Ourresultsarecomputedusing100replicationswhereasHarveyandShephard’sresultsarebasedon1000replications.Asexpected,sinceMCMCisafullylikelihood-basedmethod,italwaysperformsbetterthanQML.Forexample,relativee ciencyofQMLtoMCMC
2.intermsoftheRMSE’sare0.5633,0.7071and0.5909respectivelyforρ,φandlnσv
6Conclusions
InthisnoteweproposeaGaussiannonlinearstatespacerepresentationofboththeclassicalA-SVmodelandtheA-SVmodelofJPR.Usingthenewrepresentation,weshowthattheleveragee ectisnotwarrantedinJPR’smodel.Moreover,ourempiricalanalysisdemonstratesthatJPR’smodelisdominatedbytheclassicalA-SVmodel.Combined,theresultsnecessitatearevaluationofthesamplingpropertiesofBayesestimators.
ThesamplingpropertiesofMCMCestimatesfortheSVmodelwiththefat-tailederrordistributionhavebeenobtainedinChibetal.(2002).3
MCMC Methods for Estimat... 暂无评价 11页 免费 Volatility in Financial ....Asymmetry and Leverage in Stochastic Volatility Models-An Exposition 隐藏>> Asym...
MCMC Methods for Estimat... 暂无评价 11页 免费 Volatility in Financial ....Asymmetry and Leverage in Stochastic Volatility Models-An Exposition 隐藏>> Asym...
热门文档