-
Author
-
Zvoushe, Moreblessing
-
Title
-
A Time series analysis of maize prices in Zimbabwe
-
Abstract
-
Maize production in Zimbabwe faces significant challenges due to a lack of sustainability in the sector. Farmers are increasingly dissatisfied with the fluctuating prices at which they are able to sell maize, leading to diminishing returns on their investments. The volatility in maize prices undermines the stability of the agricultural sector, as inconsistent pricing creates uncertainty for farmers and discourages long-term investment in maize production. This instability contributes to a cycle of discontent among farmers and hinders the potential for sustained agricultural growth in Zimbabwe. Therefore, understanding the underlying causes of maize price fluctuations is crucial for addressing these challenges and improving farmer livelihoods. This study investigates maize price dynamics through time series analysis, aiming to identify key macroeconomic (Inflation, GDP, Foreign Exchange rate) and climatic variables influencing price behaviour, assess volatility using ARCH and GARCH models, and forecast future maize prices using both GARCH and Feedforward Neural Network (FFNN) models. Monthly data from 2000 to 2024, primarily sourced from the International Monetary Fund (IMF), was analysed using Python and Excel under a quantitative research design. Model performance was evaluated using AIC, BIC, MAE, and RMSE. The results show that the FFNN model outperformed the GARCH model in forecasting accuracy, effectively capturing nonlinear trends and seasonal fluctuations. While the GARCH model was useful for modelling volatility, it consistently underestimated actual price levels. Forecasts generated using the FFNN projected a steady rise in maize prices from 2025 to 2028, reflecting ongoing seasonal demand and market shifts. Based on these findings, policy recommendations include paying farmers in U.S. dollars to preserve value, increasing public awareness of the Zimbabwe Gold (ZIG) currency, ensuring timely and transparent distribution of subsidies, and developing mobile-based market information systems to support farmer decision-making. Future research should explore hybrid models, the impact of climate change, GDP, Inflation, Foreign exchange and cross-country maize price comparisons to enhance forecasting precision and inform sustainable agricultural policy.
-
Date
-
June 2025
-
Publisher
-
BUSE
-
Keywords
-
Forecasting
-
ARCH Models
-
FFNN Models
-
GARCH Models
-
Time Series Analysis
-
Supervisor
-
Mr. B. Kusotera
