均值尾部相关系数及其在金融领域的应用

Mean Tail Dependence Coefficient and Its Application in Financial Fields

传统尾部相关系数常被用于刻画变量之间的极值相关性,但这种相关系数存在对相关性信息刻画不完全的问题。为能够捕获更多的非极值相关性信息,本文提出均值尾部相关系数的概念,研究了均值尾部系数同Copula理论之间的关系,并以t Copula为例分析了4种均值尾部相关系数的变化特征。通过将均值尾部相关系数和传统尾部相关系数分别应用于沪深股市的相关性研究,从实证角度验证了这种相关系数的实用价值。

The traditional tail dependence coefficients can be used for describing the extreme correlation between different variables, however, this kind of coefficients lost information in capturing the non-extreme correlation information. In order to capture more correlation information, this paper proposes the concept of Mean Tail Dependence Coefficient. In this paper, we analyze the relationship between Copula theory and Mean tail dependence coefficient, and then we use the t Copula Function to analyze the characteristics of four Mean tall dependence coefficients. Finally, in order to test their practical value, we apply Mean tail dependence coefficients in the Chinese stock markets and analyze the correlation between the stock index of Shanghai and Shenzhen.