The Dark Side of Global Integration: Increasing Tail Dependence

We measure stock market co-exeedances using the methodology of Cappiello, Gerard and Manganelli (2005, ECB Working Paper 501). This method is based on quantile regressions and enables us to measure comovement at each point of the return distribution. First, we construct an annual co-exeedance probability for the 5, 10, 25, 75, 90 and 95 percent return quantiles using daily data from 1974-2006. Next, we explain these probabilities in a panel gravity model framework. This analysis shows that macroeconomic events asymmetrically in uence comovement of upper and lower tail returns. Financial liberalization has a positive impact on comovement across the return distribution, but its eect is strongest on the left tail quantiles. Trade competition weakly impact the 5%, 10% and 95% quantiles, but has a stronger in uence on the other quantiles. Industrial dissimilarity has a strong eect on both tails, but not on the 25% and 75% quantiles. Exchange rate volatilities have a strong eect only on the 5% and 10% quantiles. However, the introduction of the euro has its most pronounced eect on upper quantile comovement.