基于时变Copula的股票市场相关性分析

Correlation Analysis of Stock Market Based on Time-Varying Copulas

通过用t-GARCH模型拟合边际分布,和时变Copula方法一起构造联合分布函数,对常相关NormalCopula模型的缺点,以及将时变NormalCopula模型用于上证指数和恒生指数收益率的相关性进行实证研究,结果表明时变NormalCopula模型能精确描述相关性的动态变化过程,上证指数和恒生指数收益率在整体上具有正相关关系,这种相关关系具有明显的时变性,且伴随着国际金融市场一体化的进程,市场间的关联程度也越来越强。

Through fitting marginal distribution with t-GARCH model and constructing joint distribution function with time-varylng Copula to examine the shortage of Normal Copula, this paper applies time-varying Normal Copula model to investigate the correlation of return rates between Shanghai Stock Index and Hengsheng Index. The research shows that time-varying Normal Copula model can indicate the dynamic process of correlation accurately, i.e. there exist positive time-varied correlation in the return rates between Shanghai Stock Index and Hang Seng Index overall. Furthermore, the correlation between markets is closer with the gradual integration of international financial market.