基于时变Copula方法的股票收益率相关性研究

The Research on the Correlation of Stock Index Returns Based on Time Varying Copula Method

股指收益率相关性问题多数从线性相关和静态相关的角度进行研究,本文预选取上证综合股指和恒生AH股H指实际数据,采用时变条件Copula模型动态研究以上两股指之间的相关性。用经验分布函数作为股指收益率的边际分布函数对条件时变Copula函数进行参数估计,最终采用拟合优度AIC和BIC准则选择最优动态连接Copula函数。实证结果表明,以上两股存在动态的上尾相关性,并且时变Copula模型刻画尾部相关性效果优于静态Copula,同时时变RG-Copula对数据的刻画能力较好。

The correlations of most of the stock index returns are studied from the perspective of linear correlation and static correlation, time varying condition Copula is used to study correlation between Shanghai Composit Index and the Hang Seng AH Index in the paper, based on the above two indexes actual data. Firstly, take empirical distribution function as the marginal distribution to estimate the parameters of conditional time varing Copula function. Finally, the AIC and the BIC criterions are used to select the best Copula function. The empirical results show that the above two stock are upper tair correlation, and time varying Copula is better than static Copula, RG-Copula is goodness fitted to describe the data.