Dynamic risk spillovers across China’s multi-category agricultural food system: evidence from time-frequency decomposition

Frequent extreme events and evolving domestic market conditions have intensified price risk transmission within China’s agricultural food system. Using high-frequency price data for 26 major agricultural commodities, this study constructs a time-frequency volatility spillover network based on the LASSO-VAR framework, and further employs quantile Granger causality tests and a factor-augmented vector autoregression model to examine cross-commodity risk transmission and its driving forces. The results show that agricultural commodity markets in China exhibit substantial systemic connectedness, with an average total spillover index of 47.29%. Risk spillovers are mainly concentrated in the short term, where homogeneous commodities tend to form clustered transmission structures. Sugar, oil-related products, and live hogs play prominent roles in risk transmission, while pork and beef are more likely to act as risk receivers over longer horizons. Major shocks, including extreme weather events, African Swine Fever, COVID-19, and Sino-US trade frictions, significantly reshape spillover intensity, but their effects differ across time horizons. Short-term spillovers are more sensitive to consumer sentiment and pandemic-related shocks, medium-term spillovers respond more strongly to trade frictions, and long-term spillovers are mainly associated with major animal epidemics. These findings provide evidence on the formation of systemic price risk in China’s agricultural markets and offer implications for horizon-specific and commodity-specific risk monitoring and policy intervention.