股票组合VaR估计:基于动态Copula方法

VaR Estimation of Portfolio Based on Dynamic Copula Methods

已有的使用动态时变Copula估计VaR的研究都仅限于考虑两个资产,对两个资产以上,Copula函数的参数过多,逐一设定参数的动态过程,将使模型复杂化,在计算上也不可行。为解决这一问题,文中使用条件Copula的概念,结合Engle的DCC方法,将椭球Copula的相关系数矩阵动态化,并将t-Copula的自由度设定为一动态过程的Logistic变换,由此得到的动态正态Copula和t-Copula可用于刻画两个以上资产相关结构的动态关系,进而可估计两个以上资产组合的VaR。最后,文章给出了一个经验应用。

The researches on VaR estimation based on time-varying Dynamic Copula methods that been used by many researchers,and have all focused on two assets.As for more than two assets,since there are too many parameters in the Copula function,it will complicate the models and can not be computed by setting every parameter a dynamic process.To solve this problem,this paper uses the concept of conditional Copula,integrates with the DCC method of Engle,and makes the Copula function dynamic to describe the dynamic correlation structures of more than two assets and estimate the portfolio＇s VaR.At last,it gives an empirical application.