This paper demonstrates the importance for energy demand modelling of allowing for trends and seasonal effects that are stochastic in form. Inherent underlying trends may be non-linear and reflect not only technical progress, which usually produces greater energy efficiency, but also other factors such as changes in consumer tastes and the economic structure that may be working in the opposite direction. Using quarterly unadjusted data for various sectors in the UK, it is shown that unless energy demand models are formulated so as to allow for stochastic trends and seasonals, estimates of price and income elasticities could be seriously biased.
|Number of pages||26|
|Publication status||Published - Jan 2003|