Copula–GARCH Time-Varying Tail Dependence

Tail-dependence evolution for the symmetrized Joe–Clayton copula is proposed to depend on an exponentially weighted moving average (EWMA) of the absolute difference in probability integral transforms. Using these dynamics, time-varying tail dependence between bank and insurance equity prices is assessed in a parametric copula, generalized autoregressive conditional heteroscedastic framework. The results suggest a relatively long lag and support the EWMA lag structure as an effective estimation vehicle. Tail dependence is shown often to tend higher during periods of market stress.