Network effects on risk co-movements: A network quantile autoregression-based analysis

Abstract

Interconnections and volatilities among financial institutions, notably during periods of extreme risk events, are fundamental factors in studies of financial markets. Utilizing tail-event-driven networks, we depict the risk interdependencies among financial institutions and introduce a novel ranking strategy for institutional influence based on network effects and quantile regression techniques. This interdependence is examined through the prism of joint extreme movements, utilizing bivariate extreme value theory and classification analysis. To validate their effectiveness and applicability in the real world, we deploy these methodologies on Systemically Important Financial Institutions (SIFIs). Additionally, we provide a ranking of the importance of eight U.S. institutions based on tail risks, offering regulators valuable insights and references.