Contagiousness and Vulnerability in the Austrian Interbank Market

The purpose of this paper is to analyze (hypothetical) contagious bank defaults, i.e. defaults not caused by the fundamental weakness of a given bank but triggered by failures in the banking system. As failing banks become unable to honor their commitments on the interbank market, they may cause other banks to default, which may in turn push even more banks over the edge in so-called default cascades. In our paper we distinguish between contagiousness (the share of total banking assets represented by those banks that a specific bank brings down by contagion) and vulnerability (the number of banks by which a bank is brought down by cascading failures). Our analysis consists of three steps: first, we analyze the structure of the Austrian interbank market from end-2008 to end-2011. Second, we run (hypothetical) default simulations based on Eisenberg and Noe (2001) for the same set of banks. Finally, we estimate a panel data model to explain the (hypothetical) defaults generated by these simulations with the underlying structure of the network using network indicators that reflect (i) the network as a whole, (ii) a subnetwork or cluster, and (iii) the node level based on banks’ interbank lending relationships. As a result we find strong correlations between a bank’s position in the Austrian interbank market and its likelihood of either causing contagion or being affected by contagion. Although our analysis is based on a dataset constrained to the interbank market of unconsolidated Austrian banks, we believe our findings could be verified by analyzing other banking systems (albeit with a different model calibration). Given the importance of identifying systemically important banks for the formulation of macroprudential policy, we believe that our analysis has the potential to improve our assessment with regard to second-round effects and default cascades in the interbank market.