Understanding agricultural market dynamics in times of crisis: The dynamic agent-based network model Agrimate

Abstract

The concentration of crop production in a few global breadbaskets and strong import dependencies of many developing countries render global grain markets susceptible to systemic shocks from weather- or conflict-induced supply failures. Often amplified by unilateral policy responses, such as export restrictions, the resulting short-term risks to global food security are substantial but insufficiently captured by established modeling approaches. Here, we present Agrimate, a dynamic agent-based agricultural market model. Explicitly accounting for commercial and strategic stockholding, and endogenously modeling supply- and demand-side responses, Agrimate describes the spreading of supply failures in international grain trade networks and associated price effects with high temporal resolution. For the major food grain wheat, we show that Agrimate can quantitatively reproduce monthly world market price hikes and annual changes in regional supply, consumption, and stocks during the 2007/08 and 2010/11 world food crises. Further, we study potential food security risks arising from multi-breadbasket failures. We find that in a $+2°C$world, the risk of severe (90th percentile) price hikes more than doubles, while the risk of severe regional consumption losses increases by up to 130%, compared to 2006–2015 climate conditions. Our modeling shows that Agrimate can provide policy-relevant insights into the spreading of food security risks.