Research On Volatility Measurement And Prediction Of CSI 300 Stock Index Futures In China Based On Realized Range-based Bipower Variation

Stock index futures has been made since the people’s attention, it can facilitate asset management and risk aversion, it is very important to financial derivatives. CSI 300 stock index futures since April 16, 2010 since it was launched by everyone welcome, its turnover by leaps and bounds to become the world’s largest stock index futures. Volatility CSI 300 Index futures prices is an important symbol to reflect its degree of risk. Therefore, the establishment of a predictive model to accurately depict and CSI 300 stock index futures volatility has important theoretical and practical significance. In this paper, the lack of previous studies by mathematical modeling and empirical analysis carried out the following work:1.By not only avoid market microstructure noise, and can retain more information CSI 300 Index Futures 5 minutes high-frequency data were analyzed to verify the CSI 300 Index futures high frequency data also has a typical statistical frequency data characterized in that the peak fat tail, calendar, long memory, autocorrelation.2.From the descriptive statistics, jumping fluctuation characterization, long memory of the three aspects of realized volatility, realized volatility poor, bipower variation and realized volatility deteriorated four poor bipower rates were compared empirical results show that, compared with other estimator volatility, realized Range bipower has a smaller variation of the mean and variance in the descriptive statistics, and the ability to jump more volatility effectively portrayed.3.Based on heterogeneous market hypothesis and realized range bipower deteriorated constructs realized range bipower variation heterogeneous autoregressive model(HAR-RRBV) and consider jumping component realized range bipower change qualitative differences autoregressive model(HAR-RRBV-J).4.Structure based on the realized range-based bipower deteriorated heterogeneous autoregressive model(HAR-RRBV) and consider jumping realized range bipower deteriorated heterogeneous autoregressive model(HAR-RRBV-J), Shanghai and Shenzhen300 stock index futures volatility comparison of predicted and compared with the results of HAR-RV model and HAR-RV-J model empirical comparative study. Mainly through thefit and the sample within the sample outside the predictive ability compared. Wherein the multi-step static prediction sample forecasting using dynamic prediction and rolling time window method, through an intuitive graphical trend analysis and root mean square error(RMSE), mean absolute error(MAE), mean absolute percentage error(MAPE),Heteroskedasticity adjusted mean square error(HMSE) the size of the four kinds of loss function values, the prediction of the effect of different models were compared and found that CSI 300 index Futures volatility research in our country, HAR-RRBV and HAR-RRBV-J model can effectively improve the prediction accuracy of the Shanghai and Shenzhen 300 stock index futures volatility.