We introduce a decomposition method for factorizing changes in energy
demand or gas emissions over time. This method has the advantage of gi
ving perfect decomposition. It can also handle cases with zero values
in the data set. We compare this new method with three existing method
s and summarize the respective decomposition formulae for various appl
ications. Three application studies using data for Singapore, China, a
nd Korea are presented. In each case, the change of a different energy
or environmental indicator is decomposed using the four methods and t
he results obtained are compared. Our new method is superior to any of
the three existing methods and may be generally applied in energy and
environmental decomposition studies. (C) 1998 Elsevier Science Ltd. A
ll rights reserved.