Calculate VaR Based On Skst-copula-garch Model

The research of correlation is playing a very import role in financial quantitative analysis. At present, the measure of single asset risk could no longer meet the needs of investor. The research into risk investment portfolio is gaining more and more attention among the academic world. In such a case, the structure correlation of different assets has significant impact on the measurement of VaR. Traditional portfolio theory and the structure correlation of different assets can be depicted by linearly dependent coefficient. This appearance indicates that correlation is not the useable in reality. In recent years, copula as a new tool for evaluating the structure correlation between financial variables has been developed. It has shows itself as more precise and flexible than linearly dependent coefficient.Through an analysis of the literature review what was conducted on copula researches from domestic, it was foud that the Gaussian copula and Student's t-copula were commonly used in nowadays. However, they only can depict the symmetrical dependence between different assets. Therefore, based on the basic theories of copula, this dissertation employed the skewed Student's t-copula which can depict the unsymmetrical dependence between different assets and builded skst-copula-GARCH model. Then we use this model to analyze two indexs in Chinese capital market. We find that this model is better than other models in depicting the dependence of the two indexs.Finally, we apply this model to estimate portfolio VaR of Chinese capital market. By the back-testing method, we can also see this modle works better than the other models.