Vector Autoregressive Exogenous Modelling to Forecast Rice Prices Based on Inflation and Rice Production in West Sumatra Province

Abstract

Rice prices in West Sumatra Province tend to be high despite high production levels, making forecasting essential to support food security. This study aims to forecast rice prices in traditional and modern markets using a Vector Autoregressive with Exogenous Variables (VARX) model, incorporating inflation and rice production as exogenous variables. The data used consists of monthly secondary data covering the period from January 2019 to December 2024, sourced from PIHPS and the West Sumatra Provincial Statistics Agency. The analysis includes the Augmented Dickey-Fuller stationarity test, determination of the optimal lag based on the Akaike Information Criterion, parameter estimation using Ordinary Least Squares, as well as tests of stability, parameter significance, and residual diagnostics. Forecast performance is evaluated using the Mean Absolute Percentage Error (MAPE). The results show that the VARX (3,3) model is the best, with an MAPE of 1.95\% for traditional markets and 1.56\% for modern markets, indicating very high forecasting accuracy. This study demonstrates that incorporating external factors into the VARX model improves rice price forecasting accuracy, providing a basis for the government to formulate policies to maintain food price stability in West Sumatra Province.