The European financial crisis has shown that the credit risk of large financial institutions is highly interconnected as a results of a number of linkages between entities like exposure to common assets and interbank lending. In this work we propose a novel methodology to study credit risk interdependence in large panels of financial institutions.
We introduce a credit risk model in which bank defaults can be triggered both by systematic economy wide and idiosyncratic bank specific shocks.
The idiosyncratic shocks are assumed to have a sparse conditional dependence structure that we call the bank credit risk network. An estimation strategy based on CDS data and Lasso-type regression allows to estimate the parameters of the model and to recover the bank credit risk network structure. We apply this technique to analyse the interdependence of large European financial institutions between 2006 and 2013. Results show that the credit risk network captures a substantial amount of de pendence on top of what can be explained by systematic factors.
(Co-authors: Christina Hans, Eulalia Nualarte (Universitat Pompeu Fabra).)