Abstract: The correlation structure breakpoint model is utilized to pinpoint the structural breakpoints of three major extreme events: the 2008 financial crisis, the 2014 oil crisis, and the 2020 oil crisis. CoVaR-based network topology and PAMAE clustering are employed to investigate the risk characteristics and risk structure of 23 types of domestic commodities under extreme events. The research uncovers several key findings: First, the three extreme events differentially impact the risk exposure of China’s commodity markets, exhibiting both anticipatory and delayed effects. Second, there is notable heterogeneity in the risk structure within the commodity futures market. Third, different commodity types exhibit diverse risk patterns, with energy and industrial products showing the highest inherent risks, metal commodities displaying high sensitivity to external
risks, and agricultural products demonstrating considerable variations among different varieties. Furthermore, downstream products within the supply chain exhibit a more pronounced level of risk exposure compared with upstream products. These insights serve as crucial guidance for regulatory bodies and market participants in the realms of risk management and investment decision-making.

Key words: extreme events, risk structure, commodities