Research On Volatility Of China 's Stock Market Returns Based On GJR Model

In China, stock market return series exhibit four characteristics, that is,the non-normality distribution and fat tail heteroskedasticity; strong volatility clustering of the variance of return; strong autocorrelation of return series;leverage effect of Chinese stock market. We focus on constructing the model which can comprehensively characterize time series features. The feature of fat tail and volatility clustering can be better to describe by the GARCH model. The GJR model, which has similar features with the GARCH model,also can effectively portray the leverage effect of financial market volatility.In order to make the yield proportional to the risk, we construct ARMA-GJR-M model with student-t distribution to study the volatility of Chinese stock market by introducing the volatility term into mean equation,combining with the ARMA model. With the closing price of Shanghai Composite Index and Shenzhen Stock Index from December 9, 1996 to March 14, 2014, we do an empirical analysis for China’s stock market returns using the method of mixture-of-models from the perspective of analysis of time series. And the model can fit the series precisely. The results showed that the ARMA-GJR-M model can be applied to the stock market in mainland of China. Furthermore, ARMA-GJR-M model is not suitable to analyze and describe the stock market of Hong Kong.