Risk Analysis Of Foreign Exchange Rate Portfolio Based On Pair-Copula Model

On the background of financial globalization, the correlation of financial markets become closer, then, measure this correlation accurately is very important. The traditional linear correlation coefficient can only describe the linear dependencies between variables. At present, the dependent relationships between financial assets represent nonlinear, asymmetric and tail dependence. The original methods could not meet the requirements. Therefore, Copula technique becomes an effective method which can capture the nonlinear dependencies between the variables.This article mainly studies the application of multivariate Copula function. First, it introduces the basic principle and characteristics of Copula, then discuss kinds of Copula-GARCH models and their constructions. Second, we introduce Pair-Copula method to build the multivariate dependencies structure. It allows to apply different Copula functions that to describe relationships among random valuables accurately.Besides, in the selection of Pair-Copula models, we choose three types of copula function:t-Copula、Clayton Copula and SJC Copula. T-copula can capture the tail dependence of variables, Clayton Copula can capture the unconditional dependence of variables, SJC Copula can capture the conditional dependence of variables. Using these copulas, we can pay a full attention to the tail dependence of variables.Last, we use the data picked up from China’s foreign exchange market to empirical analyze the dependences of four variables. Then we combined Monte Carlo simulation technology with Pair-Copula-GARCH model, we calculating the portfolio’s VaR and ES. The VaR of portfolio based on Pair-Copula decomposition is more accurate, the result of goodness of fit proves the superiority of Pair-Copula function.