Multilayer networks in the frequency domain: Measuring extreme risk connectedness of Chinese financial institutions

Abstract

We propose multilayer networks in the frequency domain, including the short-term, medium-term, and long-term layers, to investigate the extreme risk connectedness among financial institutions. Using the conditional autoregressive value at risk (CAViaR) tool to measure the extreme risk of financial institutions, we construct extreme risk networks and inter-sector extreme risk networks of 36 Chinese financial institutions through the proposed approach. We observe that the extreme risk connectedness across financial institutions is heterogeneous in the short-, medium-, and long-term. In general, the long-term connectedness among financial institutions rises sharply during times of financial stress, such as the 2015 Chinese stock market turbulence and the 2020 COVID-19 pandemic. Moreover, we note that the insurers are key players in driving the inter-sector extreme risk networks, because the inter-sector systemic importance of insurance institutions is dominant. Finally, our conclusions provide valuable information for regulators to prevent systemic risk.