Statistical Validation of Contagion Centrality in Financial Networks

TL;DR

This paper introduces a new impact centrality measure for financial networks to assess contagion and systemic risk, along with a statistical validation method, validated through simulations and empirical data, revealing increased centrality during financial distress.

Contribution

The paper presents a novel impact centrality measure for financial networks and a statistical validation approach for network-based risk assessment.

Findings

Centrality increases during financial crises

Method reliably assesses systemic risk levels

Provides insights into firm and sector risk contributions

Abstract

In this paper, we introduce an impact centrality measure to evaluate shock propagation on financial networks capturing a notion of contagion and systemic risk contributions, permitting comparisons of these risks over time. In addition, we provide a statistical validation method when the network is estimated from data, as is done in practice. This statistical test allows us to reliably assess the computed centrality values. We validate our methodology on simulated data and conduct empirical case studies using financial data. We find that our proposed centrality measure increases significantly during times of financial distress and is able to provide insights into the (market implied) risk-levels of different firms and sectors.