Extreme Dependence Across East Asian Financial Markets: Evidence in Equity and Currency Markets

Abstract

Using unconditional and conditional copulas, this paper investigates pairwise extreme dependence across financial markets by directly modeling the tail dependence behavior. We use a dataset consisting of daily stock index returns and foreign exchange rate returns of five East Asian countries (Hong Kong, Indonesia, South Korea, Singapore, and Taiwan). The returns data span from 7/3/1997 to 6/4/2010. GARCH-type models are employed to filter the univariate returns using maximum likelihood method (MLE). Then we use copulas to model the extreme bivariate dependence between the stock index returns and between the exchange rate returns for the countries under investigation. For these two large financial markets, the degree of dependency is found to be quite different. Empirically, we find significant asymmetric tail dependence in equity markets, with a large lower tail dependence coefficient than upper tail dependence coefficient, implying that international diversification is limited since the equity pairs tend to crash together when diversification is most needed. Mixed results are found for currency co-movements. Co-movements in the currency markets are much weaker, in several cases, with larger upper tail coefficients. This study has important implications in portfolio selection and risk management strategies in international diversification.