Risks and fragilities in interbank networks.

Summary Through debt contracts, interbank deposits via the markets, or private clearing systems, banks are integrated into true interbank networks. These networks allow banks to manage their liquidity in such a way as to reduce the amount of idle reserves in order to increase the amount of investments. However, the topology - in the sense of architecture - of interbank networks is not neutral as much for the management of liquidity risk as for the question of bankruptcy risk. In fact, in a Diamond and Dybviq (1983) model extended to several banks, it can be shown that the topology influences the capacity of the different networks to distribute liquidity efficiently, i.e., in such a way as to allow participants to face the sequential liquidity service constraint with a minimal level of reserves while protecting themselves from the risk of a run on the bank. To achieve this, the network must exhibit strict topological properties. The speed of liquidity distribution and the rate of decrease of the number of liquid banks vary according to the topology involved. Moreover, the interbank network can be a propagation channel capable of transforming a localized shock in the system into a systemic crisis. The domino effect appears through the cascade of payments linking the participants that the network represents. The speed of propagation of the illiquidity crisis and the number of failed institutions are functions of the distance between the stakeholders in the network, and the degrees of centrality and connectivity of the network.