Why is tail risk fatter in China’s A-share market than in the US market? – based on the XGBoost machine learning method

Fat-tailed risk fuels systemic crises, yet its cross-market driving mechanisms (unlike volatility) call for further in-depth exploration. This study aims to explore the structural drivers of divergent tail risk ‘thickness’ between China’s A-share market and the U.S. stock market, uses 2014–2024 daily data of A-shares and U.S. NYSE stocks, and develops an EVT-XGBoost framework to measure tail risk and identify drivers. Key findings include: 1. The framework shows robust predictability, outperforming linear models/LSTM; 2. A-share tail risk is dominated by ‘high return/leverage/sentiment’ behavioral factors, while the American stock market relies on macro-fundamental factors; 3. A-share behavioral factors show threshold effects and synergistic interactions, with risk concentrated in small-cap stocks. These findings enrich the cross-market mechanism research, and offer actionable guidance for investors and regulators.