Intervention On Default Contagion Under Partial Information

TL;DR

This paper models default contagion in large heterogeneous financial networks under partial information, deriving optimal intervention policies and analyzing their impact on contagion magnitude.

Contribution

It extends existing models to include partial information and interventions, providing analytical results for optimal policies in heterogeneous networks.

Findings

Optimal intervention policy is monotonic with respect to cost and network vulnerability.

Intervening repeatedly on a bank is optimal once intervention has begun.

Simulation results align well with theoretical predictions.

Abstract

We model the default contagion process in a large heterogeneous financial network under the interventions of a regulator (a central bank) with only partial information which is a more realistic setting than most current literature. We provide the analytical results for the asymptotic optimal intervention policies and the asymptotic magnitude of default contagion in terms of the network characteristics. We extend the results of Amini et al. (2013) to incorporate interventions and the model of Amini et al. (2015); Amini et al. (2017) to heterogeneous networks with a given degree sequence and arbitrary initial equity levels. The insights from the results are that the optimal intervention policy is "monotonic" in terms of the intervention cost, the closeness to invulnerability and connectivity. Moreover, we should keep intervening on a bank once we have intervened on it. Our simulation results show a good agreement with the theoretical results.