Abstract: Under the frequent occurrences of external shocks and unexpected events since 2008, cross-market linkages have intensified the transmission of systemic financial risks, making the analysis of extreme risk scenarios increasingly critical. This paper focuses on China's stock market, examining typical stress periods such as the 2008 international financial crisis, the 2015 domestic market crash, and the 2020 major public health event. We construct an integrated "tail risk—correlation—spillover—comprehensive risk" measurement framework. Specifically, the Expected Shortfall (ES) model is used to capture tail risk, the DCC-GARCH model to measure dynamic cross-market correlations, and the ΔCoVaR model to identify external risk spillovers. Four categories of risk indicators—extreme tail risk, financial institution contagion, market instability, and liquidity risk—are synthesized using the CRITIC objective weighting method. Furthermore, a generalized variance decomposition approach is applied to analyze the evolving characteristics of risk spillover intensity. The empirical findings reveal several key insights. First, the integrated multidimensional risk indicator provides a more comprehensive depiction of systemic risk conditions by incorporating dimensions such as tail losses, leverage instability, liquidity constraints, and institutional contagion. Second, the sources of risk in China's market exhibit significant heterogeneity across different extreme events. The 2008 crisis was characterized by strong resonance between internal and external risks, with pronounced spillover effects. The 2015 crash was primarily driven by internal leverage unwinding and trading congestion, with relatively limited external impact. In 2020, the surge in external tail risk and heightened cross-market correlations significantly amplified overall spillovers. Third, the core mechanism of systemic risk lies in the synergistic amplification of multiple risk dimensions. External shocks tend to propagate rapidly through the channel of "external spill-in—correlation amplification," whereas internal vulnerabilities are more likely to diffuse gradually along the chain of "deleveraging—liquidity discount—tail losses." For data and methodology, daily index returns from major markets including China (HS300), the US (SPX), the UK (FTSE), France (CAC), Japan (N225), and India (SENSEX) from January 2006 to November 2020 are analyzed. The study employs EGARCH and DCC-GARCH models for volatility and dynamic correlation estimation, uses ΔCoVaR for bilateral spillover measurement, and applies the CRITIC method for multidimensional indicator integration. This research contributes by offering a holistic and comparable framework for measuring systemic risk under extreme events, addressing the limitations of single indicators. It provides empirical evidence for macroprudential regulation, enhances the monitoring and early warning of risk spillovers, and offers insights for cross-border risk prevention and control. The international comparative perspective further clarifies the exposure characteristics and transmission dynamics of China's market under varying global conditions.

Key words: Extreme Risk Events, Systemic Financial Risk, Integration method, Multidimensional risk indicators