Identifying Interbank Loans, Rates, and Claims Networks from Transactional Data

Abstract

Our objective is to identify interbank (i.e., non-collateralized) loans between financial institutions from Colombian large-value payment system data by implementing Furfine’s method. After identifying interbank loans from transactional data, we obtain the interbank rates and claims without relying on financial institutions’ reported data. Contrasting identified loans with those consolidated from financial institutions’ reported data suggests the algorithm performs well, and it is robust to changes in its setup. The weighted average rate implicit in transactional data matches local interbank rate benchmarks strictly. From identified loans, we also build the interbank claims network. The three main outputs (i.e., the interbank loans, the rates, and the claims networks) are valuable for examining and monitoring the money market, for contrasting data reported by financial institutions, and as inputs in models of financial contagion and systemic risk.