An agent based propagation model of bank failures

Abstract

In this paper, we analyze a network model of banking relationships in the inter-banking market and with clients using an Agent Based approach. In order to study the relationships between different agents, accounting and financial concepts are used. The goal is to understand how propagation of failures in the banking network occurs in a very short run analysis. For this purpose, an outside credit shock on one of the banks is triggered and the cascade effect of failures is simulated. This approach contributes with three new aspects to existing literature. First, three different types of agents are used in the same simulation with their own micro-behaviors – banks, consumers and a central bank; second, both credit and liquidity shocks under market stress conditions are considered; and, third, a scale-free network topology for the inter-banking relationships is adopted, which is more consistent with reality. In order to create the model and run the simulations, Netlogo Software has been used. The simulations show the presence of systemic risk for certain setups and their analysis provide some insights for policy makers on questions about solvability minimum requirements along with market regulation.