We investigate a possible change in conditional dependence between the Hong Kong and Chinese stock markets as a result of the Closer Economic Partnership Arrangement (CEPA) that took effect in 2004. Three types of conditional copula, Gaussian,
Gumbel, and Clayton, are employed to measure the change in conditional dependence between these two markets. Our results indicate that the conditional dependence derived from the Gumbel copula increases significantly in the post-CEPA period. Conditional dependences derived from the Gaussian and Clayton copulas display a tendency to increase, but do not reach a significant level. These results imply that, in the post-CEPA period, conditional dependence increases significantly only when the Hong Kong and Chinese stock markets are both rising, and not when they are declining or stable. This finding contradicts some previous studies that stock markets are increasingly correlated when in decline.