基于Copula函数的CTE研究与实证分析

采用Copula函数结合非对称Laplace分布的方法来刻画股票收益的尖峰、厚尾及偏倚性,计算了以上证指数和深证成指为组合的对数收益率的CTE,与传统的正态假设进行了对比,证实了"在资本收益率不服从正态分布时,用VaR方法来度量风险就不再准确"的结论,Copula函数结合非对称Laplace分布的方法可以较好的计算投资组合的CTE。

Research and empirical analysis of CTE based on Copula

In order to improve the VaR model, a better method of risk measurement, Conditional Tail Expectation （CTE） is adopted. As actual distribution of asset earning rate possesses the characteristics of steep peaks, heavy tails and skew, traditional normal distribution cannot properly describe these characteristics. To solve this problem, a new approach by combining copula function technique with asymmetric Laplace distribution is used. Finally, the VaR and CTE of the portfolios are computed by Monte Carlo simulation. The empirical analysis describe that the Copula method is much better than the Gaussian one.