Dependence patterns among Banking Sectors in Asia: A Copula Approach

The bitter experience of the subprime crisis of 2007, the Global Financial crisis of 2008, and the extremely slow and painful ensuing recovery, has raised systemic risk to the center stage of global economic discourses. The crisis has brought home the urgent need for a thorough assessment of the dependence and interaction between banking sectors, from which most of the trouble began. This study investigates patterns and trends in absolute and tail dependence over time using daily returns for banking sectors from 12 Asian economies during the period 2000-2012. Static and time-varying Copula models, Gaussian copula, Symmetrized Joe-Clayton copulas, are employed to study the tail co-movements among the selected markets. The paper assumes a skew-t distribution for the innovation process of the marginal models. The results of the marginal models suggest strong volatility persistence in all twelve markets. There is high persistence in the absolute dependence of among market pairs. The evidence from the empirical analysis suggests that the cross-sectional average copula correlations generally remain at moderate levels with slight upward trend for all twelve markets. Correlation among the banking sectors of the advanced Asian markets economies are generally higher compared with the Emerging markets economies. The tail dependence is asymmetric across most of the market pairs; tail dependence at the lower side of the joint distributions is mostly higher than tail dependence at the upper side of the joint distributions. The results show that tail dependence is not upward trending for most of the pairs examined. However, the tail co-movements show significant spikes in response to financial stress in the global economy, which implies that there could be joint crashes in the regional banking system during extreme negative events. The fact that the region has not been the epicenter for most of the crisis periods covered in this study, yet responds significantly, makes it necessary for adoption of policies that maximize resilience to shocks. The study concludes that time-varying copulas are best suited for modeling the dependence structure of the Asian banking sector indices compared with static copula.