How Climate Shocks Affect Stock Market Risk Spillovers: Evidence from Causal Forest Algorithm

This article categorizes climate risk into transition risk and physical risk. We developed the climate transition risk index and climate physical risk index through manual collection of textual data, and employed high-frequency data and the TVP-VAR-DY model to assess risk spillover in the Chinese stock market. Subsequently, the causal forest method was applied to analyze the causal relationships between these risks and their impact on risk spillovers across various Chinese stock markets. The research findings indicate that both transition risk and physical risk significantly influence stock market risk spillover, demonstrating a suppressive effect. Additionally, this study explored the testing of mechanisms related to public concern about climate, geopolitical risks, the risk aversion index, epidemic uncertainty, and China’s trade policy uncertainty. It also examines the impact of domestic and international policy environments and the dynamics of upward and downward risk spillovers, revealing significant heterogeneity, notably widespread negative risk spillover effects. Finally, we measured risk spillovers in stock markets of varying sizes using TVP-VAR-BK and analyzed the relationship between climate risk and stock market risk spillovers with the causal forest algorithm. We found that climate risk significantly suppresses risk spillovers in stock markets of varying sizes.