Research on the Contagion of Systemic Financial Risk Under the Impact of Climate Risks—From the Perspective of Complex Networks and Machine Learning

To systematically examine the impact of climate risks on China’s financial system, this study employs the EGARCH-SGED model to precisely fit financial market volatility based on China’s Climate Change News Index. It then combines the LASSO-CoVaR method to measure tail risk spillover effects within China’s financial system under climate risk shocks, constructs a risk contagion network, and innovatively utilizes the RF-AdaBoost model to establish the risk early warning system. Findings reveal that climate risk is a key driver of dynamic correlation evolution within the financial system, with heterogeneous impacts across different markets. Physical climate risk events intensify short-term risk contagion while generating long-term effects; transition risks undergo a dynamic process, initially amplifying uncertainty before enhancing systemic stability over the long term. The RF-AdaBoost model outperforms traditional machine learning models in risk warning, demonstrating outstanding predictive accuracy and generalization capabilities, thereby providing effective intellectual support for climate risk prevention and financial stability management.