A Copula Model Applied to the Correlation among A-share, H-share and N-share

This paper uses bivariate copula model to analyze Dependence relationship between A-share, H-share and N-share Markets from June 1, 2005 to August 31, 2009. First, the parameters of marginal distribution model for yield of A, H and N stock index are estimated by maximum likelihood method. After the most appropriate marginal distribution model selected, the parameters of the Candidate copulas are calculated by maximum likelihood method. Then, using Chi-square test, we can chose a the most appropriate copula model. The Kendall's rank correlation and the tail related coefficient can be acquired with parameter of the most appropriate copula. According to empirical analysis, we find several important characters of these stock markets.