A time series analysis of the Uganda’s value of petrol and petroleum import products (1996-2025)

Abstract

This study analyzes the trends, seasonal patterns, and forecasting performance of Uganda's petroleum import values from January 1996 to December 2025 using time series techniques. Petroleum products are a part of the major import commodity in Uganda and play a key role in supporting transport, industrial, and economic activities. However, Uganda's heavy dependence on imported petroleum exposes the economy to external shocks such as global price fluctuations and supply disruptions, making reliable analysis and forecasting essential for planning and policy guidance. The study uses monthly petroleum import values obtained from the Uganda Bureau of Statistics (UBOS) and applies descriptive statistics, seasonal decomposition, stationarity tests, and forecasting models. Specifically, the Autoregressive Integrated Moving Average (ARIMA), Random Walk, and Seasonal Naïve models are estimated and compared using forecast accuracy measures including MAPE, RMSE, and MAE. Results indicate a strong and persistent upward trend in petroleum import values over the study period, with trend effects accounting for over 95% of the total variation in the series. Seasonal effects are present but relatively weak, suggesting that long-term growth dominates short-term seasonal fluctuations. Among the forecasting models, the ARIMA(2,1,3) with drift model outperforms both the Random Walk and Seasonal Naïve models in terms of overall predictive accuracy, making it the preferred model for forecasting Uganda's petroleum import values. Forecasts from the ARIMA model suggest continued growth in petroleum import expenditure through 2027, implying increasing pressure on Uganda's foreign exchange reserves and trade balance. The study recommends that policymakers integrate these forecasts into planning processes and prioritize energy diversification strategies to reduce long-term dependence on imported petroleum products.