基于Realized GARCH模型已实现波动率预测研究——以沪深300指数为例

Research on Realized Volatility Forecasting Based on the Realized GARCH Model--Taking the CSI 300 Index as an Example

为了分析股市高频数据的波动变化规律及其风险值预测,利用已实现波动率和改进后的已实现波动率分析Realized GARCH模型下沪深300指数的5分钟高频数据收益率波动,结果显示:沪深300指数的收益率序列具有波动性和聚集性特征,不论是基于RV还是改进后的RMV,都表明t分布下模型拟合效果最好,正态分布下效果最差;而且改进后的RMV相比RV估计残差降低,估计精度提高,并且在风险价值的预测效果上表现更好。

In order to analyze the fluctuation law of the high-frequency data of the stock market and the prediction of the risk value,based on the realized volatility and the improved realized volatility,this paper analyzes the volatility of the 5-minute high-frequency data return rate of the CSI 300 Index using the Realized GARCH Model.The empirical results show that the yield series of CSI 300 Index has the characteristics of volatility and aggregation.Whether it is based on RV or the improved RMV,the model fitting effect is the best under the t distribution,and the worst under the normal distribution.In addition,compared with RV,the estimated residual of the improved RMV is reduced,the estimated accuracy is improved and the prediction effect of VaR is better.