基于SV模型和Copula函數(shù)的VaR度量方法和實(shí)證分析.pdf

1、中國科學(xué)技術(shù)大學(xué)碩士學(xué)位論文基于SV模型和Copula函數(shù)的VaR度量方法和實(shí)證分析姓名：周嶠申請學(xué)位級別：碩士專業(yè)：金融工程指導(dǎo)教師：@20110410ABSTRACTIIABSTRACTTherearethreecommonlyusedmethodsfcalculatingVaR:VarianceCovarianceApproachHisticalSimulationApproachMonteCarloSimulation.Howe

2、verwhenitcomestopracticetwoproblemsweoftenmeetarethevolatilitycalculatingthedependentstructureofaptfolio.SotheimprovementofVaRcalculatingmethodisanimptanttopicinfinance.Manymodelscanbeusedtodescribetheacteristicsofvolati

3、lity.InthispaperwefocusonStochasticVolatilitymodel.ThesampleweincludesexchangeratedataofRMBfromJanuary2008toApril2010.AftertheanalysisofstatisticacteristicsofUSDRMBEuroRMBweuseMCMCmethodunderstochasticvolatilitymodeltoca

4、lculateVaRES.TheresultshowstheSVmodelcanproperlydescribetheacteristicsofexchangeratevolatility.Therefebasingonthismodelwecanestimatetheriskofexchangerate.Inthepast，weassumeeachassetisindependentconfminglineardependencewh

5、endealingwithaptfolio.InthispaperweuseCopulatodealwiththecrelation.ThesampleweincludeslossdataofoperationriskofChinesebanksfrom1998to2008.AssumingeverykindoflosssatisfiesthesamedistributionweuseMonteCarlomethodtocalculat

6、etheVaRES.ThebacktestingshowsCopulacandowiththecrectionsamongassetswelltheVaRbasedonCopulacanreflecttheoperationriskwell.Sobankswouldthenproperlysetthereserve.KeyWds:VaRSVmodelMCMCCopulaMonteCarloSimulationEmpiricalstudy