Measuring the Time-vary Dependence between Financial Markets by Copula Method

摘要

We measure the time-vary dependence structure between Shanghai and Hong Kong stock markets with Copula functions. We filter return series into I.I.D. series by GARCH model and select t distributions as the marginal distributions. Fitting by the Copula function, we make Monte Carlo simulation to measure the VaR of the portfolio between Shanghai and Hong Kong. It showed that the dependence structure changed at different periods by comparing the empirical results. Time-vary Copula function is more reasonable and accurate than constant parameter Copula to measure the dependence structure.