Modeling Sectoral Power Demand Using Panel Model

摘要

In this paper, a panel model is designed for the analysis of monthly electricity power demand for 35 industrial sectors using data from year 2005 to 2010 in Korea. A panel model of least square dummy variable approach is used to estimate the price elasticity of power demand for each industrial sector composed of 14 different regional data. The model specification varies among each industrial group, depending on its specification of substitution effect (gas and oil prices), lags (power demand, electricity price and production index) and the inclusion of some additional variables (number of customers and production index). The estimation results of all models give us negative price elasticity of power demand with size ranging from —0.067 (group 26, construction) to —1.109 (group 18, medical, measurement, optical, and watch instruments), which show different reaction of power demand with respect to the changes in electricity price. An experiment of price change is conducted to observe its implication to the quantity of power demand and the total revenue for KEPCO (Korea Electric Power Co.), the national power retail company in Korea. This result from industry specific analysis provides new information on power demand behavior in Korea since previous works are all based on aggregated contract type data.