Data and explanatory variables

G.E. Nasr et al. r Energy Economics 22 2000 627]640 629 Fig. 1. Electrical energy consumption for the period 1993]1997. Moreover, electricity demand varies during different periods of the year. During the hot and humid summer, substantial use of ArC equipments is required in the coastal and heavily populated areas to maintain comfortable indoor spaces. In this Ž . paper, the impact and statistical significance of TI GDP proxy and DD on electrical energy consumption is investigated over different time spans of the 1993]1997 period. Furthermore, the use of the models in forecasting future electrical energy consumption is investigated.

2. Data and explanatory variables

Monthly data on electricity consumption in Lebanon is obtained from EDL. The Ž . response, electrical energy consumption C in GW-h, for the whole study period, 1993]1997, is shown in Fig. 1. Initially, the GDP is considered for use in this study as an explanatory variable. The GDP is a measurement of the value of final goods and services produced within a country in 1 year. It is an aggregate statistic that measures the total output of a country’s economy and represents all goods produced and services rendered by residents and non-residents within the political boundaries of a country. GDP G.E. Nasr et al. r Energy Economics 22 2000 627]640 630 Fig. 2. Total imports for the period 1993]1997. figures in Lebanon are available only on a yearly basis and are highly unreliable. Ž . Total imports TI are used as a proxy for the GDP for both theoretical and data availability reasons. Imports are a positive function of the GDP with the rate of change given by the marginal propensity to import out of income making it an ideal proxy variable for GDP. Moreover, data on TI, shown in Fig. 2, are available at a higher frequency and are considered to be very reliable. TI data are obtained from Ž . reports of the Central Bank of Lebanon 1993]1997 . Ž . The second explanatory variable used in this study is the degree days DD . Ž . Heating DD HDD were calculated using: Ž . Ž . HDD s 18.6 y T 1 Ý i avg i For any day i of the month with temperature less than 18.68C. Cooling DD Ž . CDD were calculated using: Ž . Ž . CDD s T y 23.3 2 Ý i avg i For any day i of the month with temperature higher than 23.38C. A composite DD for any month of the year is given by: Ž . DD s HDD q CDD 3 G.E. Nasr et al. r Energy Economics 22 2000 627]640 631 Fig. 3. Degree days per month for the period 1993]1997. The daily mean temperatures used to calculate DD are obtained from climato- Ž . logical monthly bulletins Ghaddar, 1993]1997 . Fig. 3 shows the variability of DD throughout each year for the 1993]1997 period.

3. Model specification and methodology