不同分布条件下金融风险度var的garchtarch建模与实证分析-garch tarch modeling and empirical analysis of var of financial risk under different distribution conditions.docxVIP

AbstractUndertheinfluenceoffactorssuchaseconomicglobalizationandfinanceintegration,competingandrelaxingcontrolandfinanceinnovationandtechnologicalprogress,globalfinancialenvironmentandfinancialmarkethavechangedgreatly.Meanwhile,thefluctuationofthefinancialmarketandsystemriskareaggravatedgreatly.However,inordertograspthecoretechnologyoffinancialriskmanagement.Onlydependingonmeasuringriskstakecorrectly,riskmanagementcouldbeachieved.Asoneoftheriskmanagementtechnology,theVaRmethodisproducednaturally.Thestockmarketofourcountryhavedevelopedmorethan10years,muchsuccessfulexperiencehasalreadymade,buttherearealotofunripeandnonstandardaspects.Sothestockmarketofourcountryoftenfluctuatesradically,andthemarketfluctuationisfarhigherthanthewesternripestockmarketofdevelopedcountry.Anditisalsoinfluencedbycountrypolicyinagreatdegree.soitiscalled“policymarket”.Becausethestockmarketofourcountryisataspecificdevelopingstageatpresent,itisimperativetostrengthenriskmanagement.So,VaRmodelhasagreatmeaningtotheriskmanagementofstockmarketofourcountry.Inthispaper,AdeeperresearchhasbeendoneonVaRsystemincludingitsbackgroundofanalyzing,calculationprinciple,advantageanddisadvantageofVaRmodel.Then,itprofoundlystudiesthetypicalthreemethodsforcalculatingVaRatpresent---theVarianceandCovarianceApproach,theHistorricalSimulationMethod,theMonteCarloSimulation.Andegivesoverallcompareandanalysisaboutthethreemethods.Amongthem,AstocalculateVaR,thetraditionalcalculationmethodsisestablishedinthepremiseofNormaldistribution.ButinChina,Intheempiricalstudywealsofoundthephenomenonoffattaildistribution,conditionalvarianceandasymmetricvarianceofstockindexreturn.SoweunitetheGARCH-modelandTARCH-modelandVaR-modeltogetherandresearchtheheteroscedasticityconditionsinstockmarketatthefollowinghypothsis—theNormaldistribution,theTdistributionandtheGEDdistribution.andthenusetheKupeictotestthosemodel.Finally，Wegetameaningfulconclusion:theGED-distributionisthebesttosimulatetheChinesestockmarketyielddistribution，Andsoimprovetheaccurac