Measuring dependence structure and extreme risk spillovers in stock markets: An APARCH-EVT-DMC approach

Abstract

A new APARCH-extreme value theory-dynamic mixture copula (APARCH-EVT-DMC) approach is proposed to investigate the dependence structure and risk spillovers. This method can illuminate the dynamic and complex tail dependence among stock markets. By analyzing the dynamic extreme risk spillovers in stock markets from China to G20 countries for the period from January 4, 2007 to February 7, 2023, firstly, our empirical results indicate that the EVT model is considerably better than the alternative model in fitting the tail distribution and the new DMC models outperform available copula models. Furthermore, the Russian and Argentine stock markets present the largest upside and downside risk spillovers for developed and emerging markets, respectively. Upside and downside risk spillovers from China to Canada and Saudi Arabia stock markets are the smallest for the two types of markets. Our results also provide evidence that upside and downside risk spillovers exhibit asymmetry, or more precisely, the downside risk spillovers are larger than the upside risk spillovers with exceptions in Brazil. Finally, the dynamic risk spillovers display heterogeneity and significant differences across countries.