Macro-Financial Information and Aggregate Risk Spillovers: State-Dependent Predictive Evidence from China’s Stock Market

This study examines when real-time macro-financial information improves forecasts of cross-industry risk spillovers beyond persistence. Using eleven Wind first-level industries in China’s A-share market from 2006 to 2025, we construct a monthly connectedness index and evaluate forecasts across multiple horizons. At three months, the relative predictive value of macro-financial information increases as connectedness rises more rapidly. The result is robust to alternative loss functions, persistence benchmarks, uncertainty propagation, and leave-one-industry-out tests. Real activity and price-cycle indicators provide the most consistent information, complemented by financing and external financial conditions, while evidence weakens at longer horizons.