This paper analyses the excessive comovement between industrial indices in Chinese stock market from Jan 5, 2004 to Dec 28, 2007. The excessive comovement is defined as the correlation between two industries, which is beyond what can be explained by the three Fama-French factors and the GARCH factors. Five industrial indices are illustrated, including energy index, industry index, finance index , consume index and medicine index. Firstly, we use the OLS regress method to estimate the Fama-French three factors, but autoregressive conditional heteroskedasticity is proved to exist in each index yield rate series. Secondly, we use GARCH model to eliminate the the autoregressive conditional heteroskedasticity influence. Finally, rolling OLS regress model is used to get the correlation time-series between any two index yield rate series. The results show that excessive comovement is significant between industrial indices in 80% conditions after excluding the systematic factors and GARCH factors, but the trend is that the excessive comovement is declining with time.In the last, we explain the reason , point out the shortcomings in this paper and describe further research perspective. |