Applicability of State-Space Models to Forecast Agricultural Commodity Prices One Year in Advance

Abstract

In Japan, there has been minimal research on the future forecasting of agricultural commodity prices. If “prices in the future harvest season ” can be predicted during the planting planning stage, a more profitable planting can be achieved. This study examines the applicability of a state-space model as a method for forecasting agricultural prices up to a year in advance. According to the analysis results, the state-space model was selected as the best model for 11 of the 14 target commodities. In particular, the state-space model with time-variant parameters was suggested to be effective in forecasting a year in advance.