Frequency heterogeneity of tail connectedness: Evidence from global stock markets

Abstract

The propagation of tail risks poses a significant threat to global financial stability. Although a growing body of literature studies tail connectedness across global stock markets, little attention has been given to the frequency heterogeneity of tail connectedness. To fill this gap, we develop the frequency-heterogeneous tail connectedness measures using spectrum analysis and LASSO quantile regression within the generalized variance decomposition framework. The results show that tail connectedness is more informative than mean-based connectedness, and the importance of heterogeneous tail connectedness in predicting market downside risk depends on market conditions. Specifically, tail spillovers are largely driven by the short-term component during non-crisis periods, while the long-term component will rise sharply and dominate during crisis periods. This suggests that frequency-heterogeneous tail connectedness can be taken as a refined early warning signal for systemic risk.