Korean Energy Economic Review ISSN:1599-7057(Print) 에너지경제연구

Unlike in the past, demand for city gas has recently shown a phenomenon of decoupling with a income variable. In this study, we estimated the aggregate demand function and demand functions by use applying the ARDL, error correction, and cointegration model to quarterly time series data from 1995 to 2016. As a result, the price elasticity of city gas demand and the sensitivity to temperature changes in the recent period have been increasing mainly in industrial use, while income elasticity has been decreasing mainly in household and general(or commercial) use. Estimation results by use of the error correction model were similar to those of the ARDL model, but income elasticity of general use and industrial use was somewhat higher. The estimation results of the cointegration regression model are similar to those of the ARDL model, except that the price elasticity and income elasticity of industrial demand are somewhat large. The results of out-of-sample experiments using the ARDL model and the error correction model for the last three years (2015 to 2017) of data indicate that the prediction ability can be further improved by estimating and summing the demand function for each application, rather than directly estimating the aggregate demand function.