Mathematical Modelling and Forecasting of Food Price Inflation in The Gambia: Exchange Rate Pass-Through, Regime Dependence, and Comparative Time-Series and Machine Learning Approaches
Abstract
Katim Touray, Bai Mbye Cham and Madiba Darry
The Gambia experienced significant food price inflation between 2021 and 2024, with food inflation reaching 24.4% in September 2023 amid dalasi depreciation and elevated global food prices. This study examines whether exchange rate pass-through to food prices is stable or regime-dependent and evaluates alternative approaches for forecasting food inflation. Monthly food and headline Consumer Price Index data from the Gambia Bureau of Statistics for January 2012–August 2025 are combined with FAO-sourced GMD/USD exchange rate data. Engle-Granger cointegration results show no long-run relationship over the full sample (p = 0.43), but significant cointegration during the 2021– 2025 depreciation episode (p = 0.0015), indicating that exchange rate pass-through is regime-dependent and more pronounced during periods of rapid depreciation. Forecast evaluation further shows that an exchange-rate shock regressor performs best over an 18-month holdout period. An earlier six-month forecast also achieved a genuine out- of-sample MAPE of 0.56%. However, Granger causality tests find no statistically significant evidence that exchange rate movements Granger-cause food prices, suggesting that cointegration should not be interpreted as evidence of strict causality. The findings highlight the importance of monitoring exchange-rate shocks and food price dynamics for food security, strategic reserves, and tariff policy in The Gambia.