The biggest bank may not be the most interconnected: A refined entropy-based approach for indicating the direction of interbank flow

The maximum entropy method is widely used to estimate the bilateral exposure in the interbank system. However, this method tends to overemphasize the “too-big-to-fail” issue. It could mislead policymakers into overregulating big banks. This paper proposes the Directed MaxEnt Method to measure the direction of interbank flow by integrating information from interbank deposit and Interbank offered credit. Using data from China's interbank system, we analyze the interbank data of 85 banks from 2018Q1 to 2023Q3 covering the onset of COVID-19. Our findings indicate that the largest bank in China is not the most connected entity within the network. Instead, semi-large banks emerge as pivotal hubs facilitating interbank transactions during the harsh time. Our research provides a new method for examining the short-term dynamics of interbank networks and aims to provide insights into the liquidity transmission of the interbank system.