Recent financial turmoil (e.g., the 2008-2009 global financial crisis) has resulted in financial contagion-induced instability becoming one of the major concerns in the fields of economics and finance. In this paper, we extend the network analysis of financial contagion from three perspectives. First, given that cross-holding of claims and obligations among financial institutions can be viewed as input-output linkages, we model the financial system and the contagion mechanism by introducing the classic Leontief input–output framework. Second, based on this modeling process, we propose a simple contagion algorithm to study how financial system heterogeneity influences its stability. Third, to mitigate financial contagion, we propose several concrete intervention policies based on two widely used prudential approaches—forced mergers and capital injections. The performance of these intervention policies is then evaluated by comprehensive numerical experiments. Our study has significant implications for financial regulation and supervision.

Introduction

The increasing frequency and scope of financial crises has not only made financial stability one of the major concerns of academic scientists and policymakers but also revealed the necessity of changing from a micro-prudential approach to a macro-prudential approach when considering the regulation and supervision of financial risk management (Borio, 2011). One crucial characteristic of such crises is the systemic risk of financial contagion—i.e., the potential for the failure (such as distress, insolvency or default) of one financial institution to propagate through interconnectedness, causing other institutions to fail or even the whole financial system to collapse in an unforeseen domino effect (Brownlees and Engle, 2016). Such interconnectedness is indeed a feature of the modern financial system owing to financial innovation and liberalization; financial institutions are directly interconnected1 due to the bilateral exposures (cross-holding of claims and obligations) created in the interbank market, where institutions with surplus liquidity can lend to those with liquidity shortages. These bilateral exposures are often reflected in the interconnected balance sheets as assets and liabilities. As the financial system can be labeled a network of interconnectedness (cross-holding of claims and obligations), network theory has been used intensively to model it and to analyze financial contagion in general, where the network reflects the interconnectedness of the financial system. Particularly, a large body of network literature has emerged detailing both theoretical studies (including simulation studies) (Acemoglu et al., 2015; Amini and Minca, 2016; Elliott et al., 2014) and empirical applications (Greenwood et al., 2015; Levy Carciente et al., 2014) aimed at analyzing a wide range of issues regarding financial contagion and financial stability. However, there are several shortcomings in the existing literature. First, a particular focus of those works is how the probability and extent of financial contagion are influenced by bilateral exposures and interconnectedness (Caccioli et al., 2015; Elliott et al., 2014), which are often measured using the average of degree and the average of exposures, respectively. However, these are not considered to be sufficient measures. For example, a regular network and a random network may have the same average of degree, but the interconnectedness of the two networks may not necessarily be the same. Second, much of the literature merely uses financial institutions’ book values on balance sheets (Gai et al., 2011; Nier et al., 2007), even though these cannot reflect institutions’ true value due to the ever-present inflation between book value and market value. Third, although financial contagion has become a major concern of financial regulators and supervisors, attempts to understand how to mitigate it are still in the early stage (Galati and Moessner, 2013), and there is a need to design and implement concrete intervention policies when the financial system is under distress.