Energy Efficiency and Price Responsiveness in Energy Intensive Chemicals Manufacturing
Release date: January 11, 2018
This paper presents estimates of the distribution of energy efficiency and price elasticities in the four major energy-using sectors of the upstream, energy-intensive portions of the Chemical industry (inorganic, organic, resins & plastics, and fertilizers). To obtain the estimates, we analyze non-public plant-level data from the Census of Manufacturing (CM) and the Manufacturing Energy Consumption Survey (MECS) separately, since these two data sources have their own strengths and weaknesses. The basic approach is to use an ad-hoc (reduced form) stochastic frontier energy demand function for electricity and fuel use separately. A two-stage approach is used to control for plant-level energy price endogeneity and plant-level heterogeneity. This approach provides a decomposition of the total energy efficiency into a persistent (plant specific) and time-varying component. We find that the dispersion of efficiency is relatively small, consistent with other studies of energy intensive sectors. The CM analysis implies, that if all plants were to perform at the 90 percentile of their corresponding efficiency distribution, the reduction in energy use would range between 4% and 13%. Persistent efficiency is smaller than time-varying efficiency, and new plants tend to have higher persistent efficiency but enter the industry with lower time-varying efficiency that subsequently improves. The CM analysis finds higher energy price elasticities than the MECS analysis; many are near or above unity. The MECS analysis finds elasticities in the range of -0.2 to -1.0. A logit analysis of energy price elasticities in the MECS data using a logit energy share framework a logit analysis finds lower own price elasticities and modest evidence of fuel-electricity substitution. This lower own price effect is expected because it does not account for the overall impact of aggregate energy prices on total energy demand.