International Business & Economics Research Journal – January 2009 Volume 8, Number 1 11 Demand For Electricity In Lebanon Salah Abosedra, Lebanese American University, Beirut Campus, Lebanon Abdallah Dah, Lebanese American University, Beirut Campus, Lebanon Sajal Ghosh, Management Development Institute, Gurgaon, India ABSTRACT This paper estimates the demand for electricity in Lebanon by employing three modeling techniques namely OLS, ARIMA and exponential smoothing for the time span January 1995 to December 2005. In- sample forecasts reveal that the forecasts made by ARIMA (0,1,3) (1,0,0) 12 is superior in terms of lowest RMSE, MSE and MAPE criteria, followed by exponential smoothing and OLS. Therefore, the planners in Lebanon could utilize linear univariate time-series models for forecasting future demand of electricity until detailed data on various socio-economic variables are available, which, in the future, may result in other modeling techniques being superior to estimate the demand for electricity in the country. Keywords: Electricity Demand; Exponential Smoothing, ARIMA Modeling 1. INTRODUCTION AND REVIEW OF RELATED STUDIES lectricity is one of the vital ingredients of socio-economic development of modern society. Demand for electricity is increasing rapidly, particularly in the developing countries. This increasing demand needs to be assessed properly to have a proper planning for this key infrastructure, so that the society gets the desired growth rate. The nature of electricity differs from that of other commodities since electricity is a non-storable good and there have been significant seasonal and diurnal variations of demand. Precise estimation and forecasting of electricity demand can help in proper investment for new infrastructure as well as successful operation of all type of electrical utilities so that customer demands are met cost effectively. Accurate demand forecasts are also important for better scheduling of generating utilities and assessing system security. Many studies in the literature have examined electrical energy consumption forecast and related topics, [Bernard et al., 1996 – Hsing, 1994 –Al-Faris, 2002 – Nasr et al., 2000], among others, estimated demand for electricity for various countries using different modeling techniques. Most such studies have investigated the impact of real income, price of electricity, price of substitute source of energy, population, temperature and other related variables, on the consumption of electricity. Forecasting the demand for electricity in a country like Lebanon is very important for two reasons. First, so far, only few studies have dealt with issues related to various economic aspects of electricity consumption in Lebanon . Houri and Korfali (2005) used a sample of 509 households in Lebanon to study the residential energy consumption patterns in the country in relation to income, price, area of residency and number of occupants. Nasr, Badr & Dibeh, (2000) investigated the determinants of electricity consumption (Imports and Degree Days) in Lebanon for the period 1993-1997. Cointegration analysis for the two sub-periods 1995-97 and 1996-97 revealed the existence of a long run relationship between the variables. In another study, Bader & Nasr (2001) investigated cointegrating relationship between electricity consumption and climate factors, temperature, relative humidity and clearness index for the period 1992 – 1999. Saab, Badr & Nasr (2001) used three univariate models namely autoregressive (AR), autoregressive moving average (ARMA) and an AR(1)/highpass Filter model to forecast electricity consumption in Lebanon using sample size Jan 1990 – May 1999. The study finds that the AR(1)/highpass Filter model produced the best forecast. In the present study, we have used a sample size for the time span January 1995 to December 2005, which is more recent than those used in the above-mentioned studies. E