There are several studies that have analyzed commodity prices and applied ARIMA models for forecasting commodity prices in the short run. Bessler and Brandt (1979) have compared ARIMA and Vector Autoregressive (VAR) models for livestock and corn prices and concluded that simple ARIMA models perform better compared to more sophisticated VAR style models. Kaminsky and Kumar (1990) have used futures prices to forecast future spot prices upto a horizon of nine months. Seven commodities were selected for the study.