Volatility And VaR Measure Of Several RealGARCH Models

Research on financial market volatility has been occupying important position in economic and financial issues, and the harm of financial risk is much accounted. The financial risk assessment and quantitative analysis has become the core of risk management.With the growing availability of high-frequency data, the realized volatility measure of high frequency data has become a frontier problem of volatility. How to get a better analysis of China stock index futures’volatility via the realized volatility measures is yet to test.Under the circumstances, the paper expands RealGARCH model to different forms and introduces a variety of nonparametric methods to implement model estimation. Volatility and VaR(ES) of the CSI 300 index futures are measured via different RealGARCH models.The paper measures volatility and VaR(ES) of the CSI 300 index futures by 5 minutes high-frequency data instead of traditional GARCH models.First of all, parameter RealGARCH models are used to measure the volatility and VaR (ES) of CSI 300 stock index futures’returns. Secondly, the paper extends RealGARCH model to different forms and derives algorithms by reference to nonparametric or quantile regression methods.Then, it further promotes RealGARCH model to multiple situationsto measure portfolio risk. Finally, based on different loss functions, the paper evaluates the effect of different RealGARCH models on the volatility andVaR measurement.According to the empirical analysis of CSI 300 stock index futures, we find the results change obviously when modeling with different yields. Whether the distribution is normal or not, RealGARCH (RV) performs best on volatility measure among parameter RealGARCH models, while the sp-RG(whose second equation is semiparametric) performs best among semiparameter RealGARCH models. To sum up, the simple logarithmic linear RealGARCH and quantile regression RealGARCH is more suitable to measure extreme risk of loss. In addition, highly positive correlation and volatility spillover effect exist between the CSI 300 index futures and spot markets. And there is a relatively strong sustainable joint volatility.In the end, VaR obtained from DCC-RealGARCH can accurately depict risk of theportfolio to some extent. The lower the proportion of investment in the stock index futures market, the higher the accuracy rate for the risk measure.