Forecasting Prices of Charcoal as Source of Cooking Energy using Sarima Modelling: Evidence from A Developing Country

Abstract

This study was set out to forecast average quarterly prices of charcoal used as source of energy in Uganda using Box-Jenkins Approach of modelling. Data were extracted from Annual abstracts of the Uganda Bureau of Statistics for the years 1998 up to 2014. Analysis was carried out with the help of E-Views 8. Data analysis covered Visualisation, Identification, Estimation, Diagnosis and Evaluation as an exploratory process with rollbacks at different stages. Finally Modelling and Forecasting the trends was made. A best fit model was a first order of seasonally differenced natural logarithms of the prices. According to the model lag values and seasonal errors of the natural logarithms of the prices explained 64% of the current price and its MAPE evaluation forecasting was 45% and later the model was used to forecast prices in 2015. The study extends applicability of Box-Jenkins Approach of modelling and need for data transformation to achieve normalisation before modelling. From the study, recommendations are made to intervene in the general trend of increasing prices requiring government to support the sector; and producers and consumers to review its processing and use because of its environmental and socialeconomic impacts.

Key words SARIMA, Charcoal Prices, Uganda, Box-Jenkins, Correlograms, MAPE,