Financial Fragility and Contagion in Interbank Networks
42 Pages Posted: 7 Apr 2013 Last revised: 13 Jun 2015
Date Written: October 19, 2012
In this paper I assess the resilience of different network topologies to an exogenous shock for varying degrees of financial integration, as defined by connectivity. Three different network configurations for the financial system are taken into consideration: random graphs, small world networks and scale free networks. These three network topologies are then stress-tested under four different scenarios: a benchmark case of no liquidity risk and random attack, the case of liquidity risk and two cases of targeted attack (one in a probabilistic sense, the other by setting into default the most connected institution). I find that both the topology and the scenario strongly affect the width of the contagion window and the severity of contagion itself. In particular I show how random graphs provide protection against targeted attacks, but they are particularly fragile to random attacks, while the scale free configuration is extremely fragile to targeted shocks, despite being resilient to random ones. For small world graphs I find instead that the only factor explaining contagion is the average degree, since no substantial difference is found among the four scenarios.
Keywords: Contagion, Systemic Risk, Financial Fragility, Financial Networks, Network Topology
JEL Classification: C63, D85, G01, G21, G32
Suggested Citation: Suggested Citation