Systemic risk analytics: A data-driven multi-agent financial network (MAFN) approach

Abstract

Systemic risk from financial intermediaries (FIs) refers to a negative externality problem, which is rife with fallacy of composition-type errors. To ‘see’ why seemingly rational behaviour at the level of an individual FI contributes to system-wide instability is a non-trivial exercise, which requires holistic visualization and modelling techniques. Paradox of volatility inherent to market price-based measures of systemic risk has made bilateral balance sheet and off balance data between FIs and network analysis essential for systemic risk management. There is both a data and a skills gap in implementing large-scale data-driven multi-agent financial network models that can operationalize macro-prudential policy. Different designs for a Pigou-type systemic risk surcharge are discussed with special reference to the Markose eigen-pair method, which simultaneously determines the degree of instability of the network of financial flows of obligors and also the rank order in the centrality of FIs contributing to it.