Bivariate normal mixture GARCH model: An application to Chinese stock markets

Abstract

The bivariate normal mixture GARCH model is introduced in this paper, and applied to research the dynamic volatility features and the time-varying correlation structure of Shanghai Composite Index and Shenzhen Component Index in Chinese stock markets. Empirical results demonstrate that the bivariate normal mixture GARCH model outperforms other competing GARCH models, in terms of explaining the properties of volatility process and the relation of two markets, which reflects the superiority of the bivariate normal mixture GARCH model. Besides, generalized likelihood ratio test is also used to support this conclusion through making a likelihood ratio statistic.