The hierarchy of critical participants: A clustering approach utilising network-based indicators for payment systems

Identifying critical participants is important for central banks as operators and overseers of payment systems. The binary decision about criticality is not evident, as there are no common criteria that define it. This paper introduces a novel approach using the example of TARGET2 for tackling two crucial aspects of the identification: the selection of indicators and the data-driven calibration of thresholds. It adds network centrality indicators to traffic-based indicators to tackle interconnectedness. It then utilises hierarchical cluster analysis to support the definition of thresholds for the separation into critical and non-critical participants. The results show that this approach effectively facilitates this task.