Modeling dependence structure of the risk measure on exchange market

This paper examines the structure of dependence in Europe and in Asia international exchange rate markets using extreme value theory and copula functions. We nest copula function to construct the joint distribution so that the linear correlation coefficient, which is the far most used measure to test dependence in financial community, can be well represented by nonlinear dependent construction. And the joint distribution can be divided into dependence pattern and margin distribution. Thus, the margin distribution is described by General Pareto Distribution (GPD) and the dependence structure is used copula functions. The empirical results show that the Copula-GPD model is able to characterize the exchange rate market, and their dependence patterns can be thoroughly described using the mixed copula.