Global Systemic Risk Dynamic Network Connectedness During the Covid-19 : Evidence from Nonlinear Granger Causality

In this paper, we focus mainly on the effects of COVID-19 on global systemic risk. We apply principal components analysis (PCA) and nonlinear Granger causality to analyze the systemic risk dynamic network connectedness during COVID-19 with the stock, bond, and foreign exchange markets of 14 countries from 2000 to 2021. From the PCA, we find that the commonality among multiple markets is high, and the systemic risk of COVID-19 is slightly smaller than the Financial Crisis of 2007–2009. By comparing the PCAs in each period, the exposure of bond markets to systemic risk is generally bigger than the exchange rate and stock markets. The global financial networks are constructed with the nonlinear Granger causality to capture the nonlinear relationship between markets. After constructing the global financial networks, we measure the connectedness of the global networks with the centrality and eigenvector centrality used in social networks. We find that the connection density of the network during the Financial Crisis of 2007–2009 is bigger than the COVID-19. During the Financial Crisis of 2007–2009, the stock market is the main risk exporter with the strongest spillover, and also most closely connects with the global network, indicating that the stock market is the most sensitive to risk and spreads the risk globally. Simultaneously, the bond market is the main risk receiver. However, during COVID-19, the stock and bond markets are the main sources of risk, but the foreign exchange market has the strongest connection with the global financial network. Although the Financial Crisis of 2007–2009 and COVID-19 have led to systemic risks in the global financial system, they have significant differences in the risk spillovers. Therefore, we need to take different policies to deal with the systemic risk caused by COVID-19