This paper uses the novel quantile coherency approach to examine the tail dependence network of 49 international stock markets in the frequency domain. We find that geographical proximity and state of market development are important factors in stock markets networks. Both the short- and long-run connectedness significantly increased after the global financial crisis and spillover is higher during bearish market states, highlighting the possibility of contagion effect mainly among developed markets. Frontier and emerging markets are relatively less connected. These findings have implications for international equity market diversification and risk management.