Contrastive Study The Method Of Measure VaR

Risk measurement is the basis of financial risk management, and the method ofthe measuring VaR is the focus nowadays. As so far, there are many methods ofmeasuring value-at-risk, including the parametric methods and the nonparametricmethods. The parametric methods are the equal weighted moving averagemodel(EQMA),the exponentially weighted moving average model(EWMA),ARCHsmodels and the SVs models are, which calculate the VaR by measuring the volatility.In this paper, we studied the problem of contrastive analysis the ARCHs models andthe SVs models theoretically and empirically, which of them can capture theleptokur-tosis and fat tail,the volatility clustering,persistence and leverage effect offinancial time series data, so we need to study the connection and the difference ofthem, especially to China's financial market. First of all, we introduce those two kindsof models separately and the connection of them theoretically. Then, we chose theSZZH index and the SZCZ index as objects and show by empirically analysis theyhave the statistical feature of leptokurtosis and fat tail and heteroskedastic. At last,through we analysis the ARCHs models and the SVs models contras-tively, separatelya parametric estimate on the SVs model and the GARCHs models were conduced byusing MCMC and MLE. The results indicate that the SVs model and the GARCHsmodels are good at depicting volatility clustering. However, a comprehensivecontrastive analysis of their performance of residuals fit degrees and model predictingshowed that, compared with GARCHs models, SVs models are the better at capturethe volatility of the financial returns data on China's stock market.