Forecasting Tomato Prices in the United States Using a Seasonal ARIMA Model

Abstract

The price of tomatoes often changes with seasonal patterns, supply conditions, and shifting consumer demands. As a globally important crop, these variations can create uncertainty for producers and market planners, making price forecasting important. To better understand how prices might change in the near future, we examined the monthly retail prices of field-grown tomatoes in the U.S. (May 2020 - August 2025) using the data reported in the Federal Reserve Economic Data (FRED). Applying the Box-Jenkins methodology found that a SARIMA (1,0,1) (0,0) [12] model was the best-fit model based on the Akaike (AIC) and the Bayesian (BIC) Information Criteria. Model diagnostics showed that the residuals behaved like white noise (Box-Ljung test, p = 0.9734), confirming that the model was appropriate. Also, the model demonstrated high predictive accuracy (RMSE = 0.0448; MAPE = 1.79%). The price was forecasted from September 2025 to July 2027, which indicated a relatively stable price ranging from $1.90 to $1.94 per pound. The results of our study demonstrated that the SARIMA model can effectively capture short-term patterns in tomato prices, which are valuable for market planning and risk management. However, it is important to note that these predictions indicate the expected values and should be seen as guidance rather than exact figures.