Time Series Analysis of Rice Prices using Box-Jenkins ARIMA Methodology: Case Study Hargeisa-Somaliland

Abstract

The international prices of agricultural commodities have been increasing considerably. This upward trend, which may cause a new food crisis, has attracted the attention of the world. Several explanations for these movements in prices have been provided by analysts, researchers, and development institutions. The main purpose of this study was to determine and get forecasts of rice prices in Hargeisa, Somaliland by using Box-Jenkins ARIMA modeling. Rice prices in Hargeisa were examined in order to identify if it is stationary or not. In order to check if it is stationary, we have used time series plot, correlograms and done Augmented Dickey-Fuller test. The results revealed that the data is non-stationary.  We have used some approaches such as taking differences to make the data stationary. After getting it stationary we have determined some time series Box-Jenkins models as candidates.  After that the determined models were compared with respect to the model accuracy criteria such as Akaike Information Criteria (AIC) and Bayesian Information Criteria (BIC). Then we have found the best model fitted well to the data set. After doing diagnostic checking we have calculated the forecasts. And all of the results of whole analysis were presented.   The outcome of this study can aid both Somaliland government and policy makers in making optimal production decisions and in managing overall price risks