Les cahiers du CREAD
Volume 37, Numéro 2, Pages 67-102
2021-06-30

Forecasting Future Natural Gas Demand In Algeria Using Bayesian Model Averaging

Authors : Kerriche Nawel . Moussi Oum El Kheir .

Abstract

The purpose of this article is to forecast the Algerian natural gas consumption through a combinative method using the Bayesian moving average model (BMA). Four variables for forecasting the natural gas consumption have been chosen, including global domestic product (GDP), electricity demand (ELCD), urban population (UPOP) and industrial structure (INST). The study concludes that among the four variables that have been applied, ELCD is the first most important variable affecting natural gas consumption; the UPOP comes second and then the INST. This reflects the share of the gas use sectors in Algeria: first electricity production, then households, then industry. Based on some pertinent hypotheses and according to BMA estimations of future gas demand, the National demand would be between 62 Bcm and 80 Bcm by 2028 with an average annual growth rate between 3% and 6%.

Keywords

Natural gas consumption ; Bayesian Model Averaging ; Forecasting ; Algeria