Risk Measure Of Asset Portfolio Based On Vine-Copula-Pot Modle

With the process of financial globalization becoming deeper, the connections and the comovement of global financial markets is becoming more strengthened. In recent years, the frequencies of financial crisis and the affecting areas are increasing. All of these are reminding people to strengthen risk management. The first step of risk management is depicting the correlation between assets. Also we should take the devastating effects of extreme risk to financial markets into consideration. The traditional methods assume that asset portfolio follow multivariate Gaussian distributions and use linear correlation coefficient to describe the correlation between asset portfolio. But it doesn’t match the fact which the financial assets represent nonlinear, asymmetric and tail dependence.Based on the shortcomings of traditional methods, we use the Vine-Copula function to depict the correlation between the asset portfolios. Copula is function that joins or couple multivariate distribution functions to their one-dimensional marginal distribution functions. It not only can characterize correlations also be able to describe the dependence structure. So it is suitable to measure the correlation between asset portfolios. Although multivariate Copula functions can describe the correlation well, but the most used multivariate Copula functions are assumed the marginal distribution is normal distribution or t distribution. It has some limitations, so in this paper we will use Vine-Copula functions to depict the correlation and the dependence structure between the asset portfolios. And we will use POT (Peaks over Threshold) model to describe the marginal distribution. This paper done the following works which focuses on the financial asset portfolio risk measurement:Firstly, we introduce the marginal distribution model POT (Peaks over Threshold) model. Then we use it to empirical analyze where the date was picked up from Shanghai Composite Index (From January 1 2000 to December 31th 2014 all trading days closing price data). The results show that the GARHC-POT model is more stable and accuracy compare to GARCH-norm model.Secondly, we introduce the Vine-Copula model, then combined with POT model we use five stock indexes in East Asia and S&P 500 Index in USA as an asset portfolio to empirical analyze. The consequences show that R-Vine-Copula-POT model is the best model of all the models.Finally, we make a conclusion to review the lack of this article and depict the future research.