基于Copula方法深圳A股、B股投资组合风险值实证分析

运用Copula函数代替相关系数表示风险因子之间的依存关系，进而从风险因子依存性角度探讨投资组合风险值(VaR)的计算。采用最大似然估计方法对Copula函数参数进行估计，通过Monte Carlo似然估计投资组合收益风险值。结果表明，基于Copula方法估计的投资组合风险值比方差协方差方法估计的投资组合风险值更接近实际数据。

A Demonstration of the Portfolio Value-at-risk for Share A ＆ B in Shenzhen by a Copula Based Approach

The Copula function is taken to replace the relevant coefficient and characterize the mutual relationship between risk factors. On such a basis, the calculation of portfolio value-at-risk is discussed from the foregoing relationship. The Copula functional parameters are estimated by the maximum estimation method, and a Monte Carlo Simulation based procedure to estimate portfolio value-at-risk is presented hereafter, which assumes the risk factors to be a multivariate log-return distribution different from the conditional normal one. Results show that the Copula based method produces better backtest results than the traditional variance-covariance method.