Dependence Study Between Chinext And Shanghai Stock Exchange By A Copula Approach

We investigate the dependence of daily returns between the growth enterprises market and the Shanghai Stock Exchange (SSE) in China. We choose the daily closing price data of ChiNext Price Index and Shanghai Composite Index, and build a com-posite model consisted of time-vary ing t-copula function, ARMA-GARCH models and skewed error distributions. Then we calculate full Bayesian estimators for the parame-ters using a Markov Chain Monte Carlo method. The shape of the temporal Kendall rank correlation coefficient of the two markets we obtained supports that there is high posi-tive and time-varying dependence between the two markets. Then we construct similar models for dependence structures in sectoral levels. A compare among the Kendall’s τ curves in main sectors and the entire market shows that the industrials sector is similar to the entire market, and the IT sector has a higher and more stable dependence. Mean-while, we find some pattern in the fluctuation of the Kendall rank correlation coefficient in the entire market and in industrials sector, and try to explain this phenomenon.