Extreme Value Copula Analysis of Dependences between Exchange Rates and Exports of Thailand

Abstract

This study aims to investigate a correlation of the dependence structure between USD/THB exchange rate and exports of Thailand, using extreme value copula by combining the bivariate Generalized Pareto Distribution (GPD) extreme value theory and copula. Maximum likelihood method was adopted to fit a parameter estimation based on the GPD extreme value model, and a behavior of dependence was determined by the dependence function. The procedure is suggested for the measurement of the copula function to recover the joint tail distribution by comparing four extreme value copulas. The results of this analysis denote that the Tawn copula analysis is the most appropriate method to best fit extreme value copula because the AIC of this method is the lowest when compared with the other copulas.We applied Value at Risk (VaR) to calibrate the probability of the joint tail that may occur over the threshold.We found that Tawn copula stands the maximum risk of exceeding the threshold. This result could be beneficial for exporters and policy makers to predict the possibility of extreme economical fluctuation in the future.