資源描述：
《multivariate stochastic volatility models》由會(huì)員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在工程資料-天天文庫(kù)。
1、SingaporeManagementUniversityInstitutionalKnowledgeatSingaporeManagementUniversityResearchCollectionSchoolOfEconomicsSchoolofEconomics1-2006MultivariateStochasticVolatilityModels:BayesianEstimationandModelComparisonJunYuSingaporeManagementUniversity,yujun@smu.edu.sgRen
2、ateMeyerFollowthisandadditionalworksat:http://ink.library.smu.edu.sg/soe_researchPartoftheAppliedStatisticsCommons,andtheEconometricsCommonsCitationYu,JunandMeyer,Renate.MultivariateStochasticVolatilityModels:BayesianEstimationandModelComparison.(2006).EconometricRevie
3、ws.,25(2/3),361-384.ResearchCollectionSchoolOfEconomics.Availableat:http://ink.library.smu.edu.sg/soe_research/360ThisJournalArticleisbroughttoyouforfreeandopenaccessbytheSchoolofEconomicsatInstitutionalKnowledgeatSingaporeManagementUniversity.Ithasbeenacceptedforinclu
4、sioninResearchCollectionSchoolOfEconomicsbyanauthorizedadministratorofInstitutionalKnowledgeatSingaporeManagementUniversity.Formoreinformation,pleaseemaillibIR@smu.edu.sg.EconometricReviews,25(2–3):361–384,2006Copyright?Taylor&FrancisGroup,LLCISSN:0747-4938print/1532-4
5、168onlineDOI:10.1080/07474930600713465MULTIVARIATESTOCHASTICVOLATILITYMODELS:BAYESIANESTIMATIONANDMODELCOMPARISONJunYuSchoolofEconomicsandSocialSciences,SingaporeManagementUniversity,SingaporeRenateMeyerDepartmentofStatistics,UniversityofAuckland,Auckland,NewZealand
6、Inthispaperweshowthatfullylikelihood-basedestimationandcomparisonofmultivariatestochasticvolatility(SV)modelscanbeeasilyperformedviaafreelyavailableBayesiansoftwarecalledWinBUGS.Moreover,weintroducetotheliteratureseveralnewspeci?cationsthatarenaturalextensionstocertain
7、existingmodels,oneofwhichallowsfortime-varyingcorrelationcoef?cients.Ideasareillustratedby?tting,toabivariatetimeseriesdataofweeklyexchangerates,ninemultivariateSVmodels,includingthespeci?cationswithGrangercausalityinvolatility,time-varyingcorrelations,heavy-tailederro
8、rdistributions,additivefactorstructure,andmultiplicativefactorstructure.Empiricalresultssuggestthatthebestspeci?catio
