Modelling maize price volatility in sub-Saharan Africa: a GARCH-X approach

We examine maize price volatility in Nigeria, Gambia, the Democratic Republic of Congo (DRC), Burundi, and Burkina Faso using Generalized Autoregressive Conditional Heteroskedastic-X (GARCH-X) models. Monthly price data from 2013 to 2023 are analysed, incorporating exogenous variables such as exchange rates, crude oil prices, fertilizer prices, global maize prices, GDP per capita, and domestic maize supply. The ARCH-LM test validates volatility, justifying the use of GARCH-type models. Model comparison using AIC, BIC, likelihood-ratio and root mean square error tests confirm that the GARCH-X model outperforms the standard GARCH model, demonstrating that exogenous factors enhance volatility modelling. Findings indicate that exchange rates significantly influence maize price volatility in DRC and Burkina Faso. In contrast, crude oil prices impact maize price volatility in the Gambia and Burkina Faso. In Nigeria, GDP per capita and domestic maize supply are key drivers of maize price volatility. Additionally, the GARCH-X model shows lower volatility persistence, suggesting exogenous factors contribute to the stabilization of prices. The results highlight the need for targeted policy interventions, including exchange rate stabilization, energy diversification, and investment in domestic production.