Dependence Structure Between The Stock Market And The Foreign Exchange Rate Market Via Copula Approach

The dependence structure between asset returns plays decisive role in financial industry.The main widely used measure is Pearson's linear correlation.However,among other disadvantages it is unable to measure the non-linear dependence across financial assets,which is almost always the case we encounter in reality.To cover shortcomings of ordinary linear dependence,copula functions are used as effective tools.This paper investigates the extreme concurrent movements in stock-currency return relationship in three markets(China,Hong Kong,the United Kingdom)by applying copulas over the period 2005-2017.The equity-currency univariate returns are developed via ARMA-GARCH models and the relationship between them is determined via use of copulas.To capture different forms of dependence,we compare four copulas:Gaussian,Student's t,Gumbel and Rotated Gumbel copulas.The parametric models of copulas are estimated by the Canonical Maximum Likelihood(CML)method.The model selection is based on the AIC,BIC information criterions.The empirical results show that symmetric heavy concurrent movements are present between equity market and currency exchange rate market.This finding is valuable in international investment,risk management and asset pricing as it provides more precise estimation of the relationship across financial markets.