구조변화 이후 산업용 천연가스 수요함수의 추정

This study estimates the demand function of industrial natural gas by country and region through error correction models. Specifically, we estimate the price and income elasticities of the nation and the region by applying the cointegration method(FM-OLS) to the time series data since the recent structural change(May 2012). As the impact of macroeconomic variables on industrial natural gas consumption has greatly decreased since the time of structural change, explanatory variables were composed of relative prices of industrial natural gas compared to petroleum, temperature effect, the gas demand industry index, etc. As a result of estimating the national demand function as a cointegration regression model, income elasticity and price elasticity appeared to be inelastic by 0.465 and -0.227, respectively. In the case of demand function by region, the relative price has no effect on the industrial natural gas consumption in the metropolitan area, and the income elasticity in the metropolitan area is smaller than that in the local area. But the temperature effect is larger in the metropolitan area. In addition, the result of an out-of-sample experiment for the aggregate demand function indicated that the mean absolute percentage error(1.7%) of the error correction model was much smaller than that(6.0%) of the conventional multiple regression model.