Extremal spillovers in financial markets

We analyze the interdependency between different financial markets by using multivariate extreme value theory. This permits one to focus on the occurrence of simultaneous financial market crises, whereas standard co-variance analysis is less suitable for studying extreme interdependencies. The analysis builds on the so-called stable tail dependence function which measures the amount of interdependency between the tail probabilities of multiple random variables. The empirical implementation of this semipara-metric approach relies on order statistics. With these estimates one can calculate conditional spillover probabilities or other VaR-related multivari-ate risk measures for vectors of asset returns and for chosen crash levels. An empirical illustration shows relatively low stock market spillovers which is not in line with the presumption that stock markets are fairly good in-tegrated and that integration has risen over time.