Clearing payments in dynamic financial networks

Abstract

This paper proposes a novel dynamical model for determining clearing payments in financial networks. We extend the classical Eisenberg–Noe model of financial contagion to multiple time periods, allowing financial operations to continue after initial pseudo defaults, thus permitting nodes to possibly recover and eventually fulfill their liabilities. Dynamic optimal clearing payments in our model are computed by solving a suitable linear program, both in the full matrix payments case and in the pro-rata constrained case. We prove that the proposed model obeys the priority of debt claims requirement, that is, each node at every step either pays its liabilities in full, or it pays out all its balance. In the pro-rata case, the optimal dynamic clearing payments are unique, and can be determined via a time-decoupled sequential optimization approach.