An Investigation of the Contagion Effect in Asian Stock Markets under Extreme Rate of Return Using Copula Approach

This study applies copula functions with properties of asymmetric dependence structures and extreme value and the GJR-GARCH model with skewed Student t distribution (GJR-GARCH-ST) to estimate the marginal and joint distributions of stock returns in Taiwan, Hong Kong, Japan, South Korea and Singapore. Furthermore, this study applies the conditional probability of contagion to investigate the possibility of the extreme co-movement among Asian stock markets given that an extreme positive or negative shock occurs in one of them. The empirical results are as follows: (1) both large market rise and fall result in the highest possibility of co-movement between the Hong Kong and Singaporean markets, implying the risk diversification benefits offered by the portfolios with Hong Kong and Singapore are very limited; (2) when Hong Kong faces an extreme fall, Taiwan is less likely affected, and vice versa, implying long portfolios built around Hong Kong and Taiwan represent a better risk diversification; (3) when Japan experiences an extreme positive impact, Taiwan is the less likely affected, and vice versa, implying short portfolios built around the Taiwan and Japanese markets represent a better risk diversification. The conclusions can help bullish and bearish investors realize the co-movement between Asian markets and diversify risks under extreme impact.