Based On Copula 's Portfolio Risk Analysis

In recent years, Copula connect function and the VaR and CVaR method has been widely used in portfolio risk measurement. Topics of this article are portfolio and risk analysis. Firstly, We introduced the Copula, pair Copula and vine. Then we introduced the portfolio models and VaR theory. Finally, we gave an empirical analysis. Using Copula-GARCH-partial t model, we fitted the dependency structure between the assets. The portfolio weight under the minimum risk is calculated through the combination between the Copula-GARCH-partial t model and the mean-variance model, the mean-CVaR model. The portfolio risk, VaR, is calculated through the Copula-GARCH-partial t-VaR model.In the empirical analysis, the marginal distributions is fitted by the kernel density estimation and GARCH models. Under the AIC criterion, we compared the AIC of four single Copula models and three vine Copula models. The result of R vine Copula is more accurate in the description of the dependency structure between the assets. In combination with the VaR and portfolio models, we obtained the optimal investment weight, the results of the back test for the estimated VaR. After that, we compared the VaR which the weight is from the two kinds of portfolio models.