Imported Risk in China’s Commodity Markets: Cross-Border Transmission Under Major External Shocks

Against intensifying geopolitical tensions, trade-policy uncertainty, and recurrent global supply-chain disruptions, understanding cross-border commodity risk transmission is important for managing imported risk and maintaining commodity-market stability and supply continuity. Using daily data for 18 international and Chinese commodity futures markets from 26 March 2018 to 30 December 2022, this paper combines Glosten–Jagannathan–Runkle generalized autoregressive conditional heteroskedasticity (GJR-GARCH), Elastic Net vector autoregression (Elastic Net-VAR), and the Diebold–Yılmaz framework with event-study and transmission-channel analyses to examine the China–US trade dispute, the COVID-19 pandemic, and the Russia–Ukraine conflict. The main findings are as follows. (1) Dynamic spillovers are intermittent and dispersed during the China–US trade dispute, rise abruptly during COVID-19, with international crude oil becoming the dominant risk transmitter, and become broader and more persistent during the Russia–Ukraine conflict, with pronounced risk output from wheat markets; (2) the selected trade-dispute events are not followed by a broad-based increase in imported risk, whereas imported risk peaks on the first post-event trading day during COVID-19, particularly in precious and industrial metals, and around the third trading day during the Russia–Ukraine conflict, with industrial metals showing the clearest increase; and (3) macro-financial associations vary across shocks: exchange-rate risk and international monetary-policy conditions become more prominent during the trade dispute, COVID-19 involves broader changes across multiple channels, and the Russia–Ukraine conflict shows clearer divergence between international and Chinese sentiment and stock-market factors. These findings provide an empirical basis for differentiated risk management aimed at maintaining commodity-market stability, supply continuity, and sustainable functioning under recurrent external disturbances. This study remains limited in terms of data coverage, event-specific analytical depth, and the granularity of transmission-channel indicators, leaving scope for future extensions.