Propagation of Systemic Risk in Interbank Networks

TL;DR

This paper investigates how the structure of interbank networks affects systemic risk propagation, showing that networks with high connectivity and credit concentration are more resilient to contagion, and topological risk indices can predict systemic impact.

Contribution

It introduces a model linking network connectivity and link concentration to systemic risk, analyzing different network types and their resilience to contagion.

Findings

Networks with high credit concentration are more resilient.

Topological indices correlate with systemic risk impact.

Connectivity and link concentration influence contagion dynamics.

Abstract

This work explores the characteristics of financial contagion in networks whose links distributions approaches a power law, using a model that defines banks balance sheets from information of network connectivity. By varying the parameters for the creation of the network, several interbank networks are built, in which the concentrations of debts and credits are obtained from links distributions during the creation networks process. Three main types of interbank network are analyzed for their resilience to contagion: i) concentration of debts is greater than concentration of credits, ii) concentration of credits is greater than concentration of debts and iii) concentrations of debts and credits are similar. We also tested the effect of a variation in connectivity in conjunction with variation in concentration of links. The results suggest that more connected networks with high concentration of credits (featuring nodes that are large creditors of the system) present greater resilience to contagion when compared with the others networks analyzed. Evaluating some topological indices of systemic risk suggested by the literature we have verified the ability of these indices to explain the impact on the system caused by the failure of a node. There is a clear positive correlation between the topological indices and the magnitude of losses in the case of networks with high concentration of debts. This correlation is smaller for more resilient networks.