Multi-scale dependence and risk contagion among international financial markets based on VMD-Vine copula-CoVaR

Considering the frequency domain and nonlinear characteristics of financial risks, we propose a VMD-Vine copula-CoVaR framework to study the dependence structures and risk spillovers among international financial markets at different time scales. Furthermore, a mean-variance optimization technique has been applied to evaluate the performance of the optimal cross-market portfolios. The empirical results show that: (1) The R-Vine copula is superior to the C- and D-Vines; (2) Compared with the Asian stock markets, the European stock markets are more correlated with each other. The correlations between the markets at the long-term scale are greater than those at the short- and medium-term ones; (3) The optimal pair copulas are combined with the CoVaR to estimate the systemic spillover risks among the markets. The downside and upside spillover risks across the markets are not always symmetric. Moreover, the ΔCoVaRs at the medium-term scale are the highest and the lowest at the short-term one; and (4) The optimal cross-market portfolio constructed by our framework perform better than the benchmarks by annualized return, Sharpe ratio, and maximal drawdown. The findings have direct implications for portfolio managers and investors to prevent extreme risks and take better international investment decisions based on the knowledge of optimum portfolio.