Cross-border interbank contagion in the European banking sector

Abstract

This paper studies the scope for cross-border contagion in the European banking sector using true bilateral exposure data. Using a model of sequential solvency and liquidity cascades in networks, we analyze geographical patterns of loss propagation from 2008 to 2012. We study the distribution of contagion outcomes after a common shock and an exogenous bank default over simulated networks of actual long- and short-term claims. We exploit a novel and unique dataset of money market transactions estimated from TARGET2 payments data. Our results suggest the evidence for cross-border contagion with evolving over the years geographical patterns and decreasing potential for contagion. Losses due to defaults of domestic counterparties remain on average more important. Furthermore, our results underline an important effect of the underlying network structure on the propagation of losses. Notably, an econometric analysis shows that a denser network of long-term commitments with a shorter average path is more prone to contagion, while higher clustering (more triangles) in short-term networks reduces network fragility mostly due to better liquidity sharing.