The Study of Nonlinear Correlation between Shanghai, Hongkong and American Stock Returns — An Empirical Analysis Based on MS-VAR Model and MS-DCC-MVGARCH Model

Considering the mean and the volatility correlation of Chinese and foreign stock market may undergo structural changes because of the reform and opening-up, the paper attempts to incorporate Markov state transition mechanism(MS) into both the VAR model and DCC-MVGARCH model at the same time. Based on that, it constructs the MS-VAR model and MS-DCC-MVGARCH model to empirically verify the nonlinear mean spillover effect and the volatility correlation among the Shanghai, Hong Kong and American stock markets. Empirical research shows that: firstly, there exists differentiating character among the correlation of these stock markets. USA stock market has positive spillover effect on Shanghai and Hong Kong stock markets, but it is not obvious conversely; at the same time; the volatility correlation between Shanghai and Hong Kong stock market is the highest, and it presents periodic volatility, while the volatility correlation between HK and US is the lowest, and it presents a stable fluctuant feature. Secondly, the interaction among the effects of Shanghai, Hong Kong and American Stock Market presents the obvious non-linear feature. The mean spillover effect among these stock markets in state 2 is significantly greater than in state 1; at the same time, the effects of volatility among these stock markets in state 1 is significantly higher than that of in state 2