Download this Paper Open PDF in Browser

Optimal Equity Infusions in Interbank Networks

35 Pages Posted: 13 Aug 2012 Last revised: 13 Feb 2013

Hamed Amini

Ecole Polytechnique Fédérale de Lausanne

Andreea Minca

Cornell University

Agnes Sulem

French National Institute for Research in Computer Science and Control (INRIA)

Date Written: August 13, 2012

Abstract

We study optimal equity infusions into a financial system represented as a hierarchical network with two classes of banks, prone to both the risk of insolvencies and the risk of runs by short term creditors. The government seeks to minimize, under budget constraints, the magnitude of the total loss in the system. Under complete information on interbank exposures, the problem can be expressed as a combinatorial optimization problem, tractable when the set of core banks is small. Under partial information on exposures, revealed over time, distress propagation under intervention is shown to become a Markov Decision Process. We reduce the dimensionality of the problem so as to make it numerically tractable. We find the optimal intervention policy as a result of Hamilton-Jacobi-Bellman equations. Our results show that, in presence of a relatively small fraction of banks that use short-term funding, the optimal strategy swiftly changes from a strategy to avoid insolvency to a strategy to avoid runs by short term creditors.

Keywords: systemic risk, liquidity risk, financial contagion, Markov decision process, optimal interventions

JEL Classification: C6, G01, G28

Suggested Citation

Amini, Hamed and Minca, Andreea and Sulem, Agnes, Optimal Equity Infusions in Interbank Networks (August 13, 2012). Available at SSRN: https://ssrn.com/abstract=2128476 or http://dx.doi.org/10.2139/ssrn.2128476

Hamed Amini

Ecole Polytechnique Fédérale de Lausanne ( email )

Station 5
Odyssea 1.04
1015 Lausanne, CH-1015
Switzerland

Andreea Minca (Contact Author)

Cornell University ( email )

222 Rhodes Hall
Ithaca, NY NY 14853
United States

HOME PAGE: http://people.orie.cornell.edu/acm299/

Agnes Sulem

French National Institute for Research in Computer Science and Control (INRIA) ( email )

40 avenue Halley
Villeneuve, 59650
France

Paper statistics

Downloads
428
Rank
55,238
Abstract Views
1,709