Analysing contagion and bailout effects with copulae

Abstract

In this paper, a new methodical framework that combines elements of event studies and copula methodology is proposed in the context of the analysis of bank contagion. Furthermore, to the best knowledge of the author, this paper is the first one to analyse changes in the dependence structure of banks around bailout announcements. The results of the empirical study show that significant contagion effects could be detected both in the German banking sector after the onset of the subprime crisis as well as in the Japanese banking sector in the mid-nineties. I find that announcements of crisis at struggling banks induce a significant increase of lower tail dependence in the banking sector. The analysed bailouts and rescue measures by the central bank proved to be effective in reducing this increased lower tail dependence while increasing tail independence of bank stock returns at the same time. In both data samples, I find that the bailout announcements did not simply restore the pre-crisis dependence structure, but rather only decreased the likelihood of a joint crash of bank stocks without increasing the chances of a joint boom.