The Application of the Divisia Index to the Decomposition of Changes in Industrial Energy Consumption *

Abstract

We review a number of methods that have recently been proposed to decompose changes in industrial energy consumption. We then propose two parametric methods based on the Divisia index, where the integral path problem in the Divisia index is transformed into a parameter estimation problem. It is shown that there can be an infinite number of sets of decomposition results, each corresponding to a particular combination of parameter values, and that several recently proposed methods are in fact special cases of these two methods. We then introduce an approach to estimate the parameter values uniquely. Referred to as the Adaptive Weighting Divisia Method, this method is supported by vigorous mathematical analysis and does not involve arbitrary guesses of parameter values as is the case for the existing methods. We also discuss the application and the associated statistical problems of the various decomposition methods, and present the results of a study using the data for Singapore industry.