The Performance of ARIMAX Model and Vector Autoregressive (VAR) Model in Forecasting Strategic Commodity Price in Indonesia

Rice as one of the strategic commodities has an important role in the life of Indonesian society. This is cause of rice is the main food of the Indonesian nation. Therefore, the stabilization of food prices is one of the priorities of the Indonesian government's policy. It can minimize the impact of the global financial crisis such as inflation and purchasing power of the poor. The sta bility price can be maintained by price forecasting for several periods ahead. It can be used to set up the anticipatory action. In this research, ARIMAX model and VAR model used to forecast the rice price. This model involves several variables including consumer rice price (HKB), production (PROD), dry milled rice (GKP), harvested area (LP), and rice price in Thailand (HD). The results show that ARIMAX model can predict the rice consumer price with MAPE 0.15%. This is 15.27 % better than VAR model. The GKP variable did not significantly affect to the rice price. This is indicated by the MAPE difference between model with GKP and model without GKP is less than 0.01%.