What factors contribute to the volatility of food prices? New global evidence

• Published in: Agricultural economics (Praha) Vol. 69; no. 5; pp. 171 - 184
• Main Authors: Zmami, Mourad, Ben-Salha, Ousama
• Format: Journal Article
• Language: English
• Published: Czech Academy of Agricultural Sciences 01.01.2023
• Subjects: causality; food commodity price; glosten, jagannathan, and runkle-generalised autoregressive conditional heteroskedasticity (gjr-garch); quantiles
• ISSN: 0139-570X; 1805-9295
• DOI: 10.17221/99/2023-AGRICECON

The adverse economic and social effects of the recent increases in food prices have prompted policymakers and academics to reconsider the potential causes of such increases. This paper contributes to the existing literature by investigating the causal effects of oil prices, fertiliser prices, global economic activity, and geopolitical risk on international food price volatility between January 1993 and December 2021. The research considers the aggregate food price index and the prices of various specific foods, including cereal, vegetable oils, dairy, meat, and sugar. The Glosten, Jagannathan, and Runkle-generalised autoregressive conditional heteroskedasticity (1,1) [GJR-GARCH(1,1)] model is employed to estimate the food price volatility series, while the causality-in-quantiles test is conducted to identify the drivers of food price volatility for different volatility regimes. The analysis suggests heterogeneous results regarding the significance of causal linkages. More specifically, the aggregate food price volatility is affected by oil prices, global economic activity, and geopolitical risk under different market conditions. The causality analysis also indicates that the volatility of cereal prices is the most sensitive to the four considered variables. Likewise, geopolitical risk is the most critical factor affecting all food commodities during almost all market conditions, while oil prices and global economic activity have limited predictive power. Finally, there is strong evidence that most causal linkages are confirmed during normal market conditions. Policy recommendations are subsequently derived.