金融计量济学入门：模拟方法及其应用.pptx

‘IntroductoryEconometricsforFinance’?ChrisBrooks20021Chapter10SimulationMethods

‘IntroductoryEconometricsforFinance’?ChrisBrooks20022SimulationMethodsinEconometricsandFinanceTheMonteCarloMethod Thistechniqueisoftenusedineconometricswhenthepropertiesofaparticularestimationmethodarenotknown. Examplesfromeconometricsinclude:Quantifyingthesimultaneousequationsbiasinducedbytreatingan endogenousvariableasexogenous.DeterminingtheappropriatecriticalvaluesforaDickey-Fullertest.Determiningwhateffectheteroscedasticityhasuponthesizeandpowerofa testforautocorrelation. Simulationsarealsooftenextremelyusefultoolsinfinance,insituationssuchas:Thepricingofexoticoptions,whereananalyticalpricingformulaisunavailable.Determiningtheeffectonfinancialmarketsofsubstantialchangesinthemacroeconomicenvironment.“Stress-testing”riskmanagementmodelstodeterminewhethertheygeneratecapitalrequirementssufficienttocoverlossesinallsituations.

‘IntroductoryEconometricsforFinance’?ChrisBrooks20023ConductingSimulationsExperimentsInalloftheaboveexamples,thebasicwaythatsuchastudywouldbeconducted(withadditionalstepsandmodificationswherenecessary)isasfollows. 1. Generatethedataaccordingtothedesireddatageneratingprocess (DGP),withtheerrorsbeingdrawnfromsomegivendistribution. 2. Dotheregressionandcalculatetheteststatistic. 3. Savetheteststatisticorwhateverparameterisofinterest. 4. Gobacktostage1andrepeatNtimes.Nisthenumberofreplications,andshouldbeaslargeasisfeasible.ThecentralideabehindMonteCarloisthatofrandomsamplingfromagivendistribution.Therefore,ifthenumberofreplicationsissettoosmall,theresultswillbesensitiveto“odd”combinationsofrandomnumberdraws.

‘IntroductoryEconometricsforFinance’?ChrisBrooks20024RandomNumberGenerationWegeneratenumbersthatare