Abstract
Given the importance of the global environmental agenda, never before has it been so important to understand the determinants of industrial energy demand in the developed world to aid international policy makers in their deliberations. They need sound and dependable models to support their projections of future industrial energy demand to underpin policy; for example to grant emission trading permits. Therefore, never has it been more important to obtain 'good' estimates for energy demand as well as of the best way to specify and understand technical change. Previous research has attempted to resolve the debate on finding the appropriate way to measure technical progress in energy demand models without a definite consensus. Arguably more recent research has narrowed the debate to two competing methodologies - Asymmetric Price Responses (APR) and Underlying Energy Demand Trends (UEDT) - as a means of accounting for technical progress in energy demand models. This research has therefore used these competing methodologies within a panel data and time series modelling context. The time series methodology attempts to uncover the UEDT using the structural time series modelling (STSM) framework, while the panel data methodology seeks to uncover technical progress through the use of APR and fixed time dummies. This thesis, as part of this debate, uses both methodologies to investigate OECD industrial energy demand, with a view to obtaining an 'appropriate' model for energy demand in the industrial sector. In order to achieve this objective, three models are used to evaluate OECD industrial energy demand in a panel data context and a time series context. These are the asymmetry model without stochastic trend, the symmetric model with stochastic trend, and the asymmetric model with stochastic trend. Results from the time series analysis indicate that the data on OECD industrial energy demand strongly support the inclusion of a stochastic trend in the models. Although it is very difficult to obtain consistent estimates across different data sets, the results and tests from the panel data modelling framework supports the use of APR in explaining technical progress in the OECD industrial sector in the longer data sets while it incorporates an exogenous trend in the shorter data sets.