Systemic financial risk analysis of the U.S. based on the complex network

This paper proposes a new framework for measuring systemic financial risk, which combines the Diebold and Yilmaz spillover index model, complex network, and dimensionality reduction method. This framework simulates the process of risk contagion through network dynamics, accounting for the coupled relationships between internal indicators and addressing the high-dimensional issues of financial networks. We conducted a study of the U.S. financial system from a dynamic perspective using a rolling window approach. The results reveal a strong alignment between total risk spillover and total risk contagion, and the analytical solutions and simulation results of systemic risk are consistent, which indicates the effectiveness of the measurement method for systemic financial risk proposed in this paper. Additionally, we find that hub centrality (HC) and global importance of nodes (GIN) influence the ranking algorithms' performance in identifying the nodes with the greatest influence on total systemic risk within the risk spillover networks. Furthermore, enhancing the risk resistance of a node is highly important for improving the stability and resilience of the financial system. The framework proposed in this paper provides a quantitative tool for real-time measurement of systemic financial risk and offers theoretical tools for risk managers to grasp the direction of risk management, which is crucial for the timely identification and prevention of financial crises.