Global Monthly Cotton Price Trends: Insights From Arima Time-Series Forecasting

Cotton is a crucial cash crop in the global economy, yet its prices often change due to market trends and seasonal variations. Accurate price forecasting can help farmers, traders, and policymakers make better, timely decisions and promote sustainable production. This study uses a Seasonal Autoregressive Integrated Moving Average (SARIMA) model to analyze and forecast global monthly cotton prices using data from January 1990 to January 2025 obtained from the FRED database. Model selection was performed with the auto.arima function from the forecast package in R, guided by the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC). Following the Box–Jenkins methodology, the final model, ARIMA (4,1,1) (0,0,2) [12], satisfied key assumptions of stationarity, normality, and homoscedasticity, confirmed through the Augmented Dickey–Fuller test, Ljung–Box test, and residual ACF/PACF diagnostics. Forecast accuracy was evaluated using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Percentage Error (MPE), Mean Absolute Percentage Error (MAPE), and Mean Error (ME), showing modest out-of-sample errors consistent with prior studies on commodity price forecasting. A 30-month forecast indicates a steady increase in cotton prices, from 76.60 U.S. cents per pound in February 2025 to 91.66 cents by January 2027, with minor monthly fluctuations. These findings highlight SARIMA as a practical tool for anticipating cotton price trends and supporting informed decision-making. However, forecasts should be interpreted as guidance rather than exact predictors of future prices. Future studies may benefit from integrating SARIMA with hybrid models or machine learning approaches to better capture complex, non-linear relationships in commodity price forecasting.