Foreword

Summary

Acknowledgments

Introduction

Methodology

The manufacturing sector

The services sector

Agriculture, mining and construction sectors

Conclusions

Appendices

Download pdf

2 Methodology

In this report, changes in energy use over the period 1989-90 to 2005-06 have been broken down into three components. A change in energy consumption is divided into the activity effect, the structural effect and the real intensity effect using a technique known as factorisation (also known as decomposition). This report uses an additive version of the log-mean divisia index I (LMDII) method to decompose changes in energy consumption into these three components (figure e). A mathematical framework of this method is provided in appendix A.

The activity effect refers to the changes in energy consumption that arise solely due to the changes in the level of activity in the economy. The activity effect generally accounts for the majority of the observed changes in energy consumption in any particular sector. In this report, the level of activity is estimated using physical production for the iron and steel and basic non-ferrous metals subsectors, and using gross value added for other subsectors. The gross value added used in this report is based on chain volume terms, which removes the effect of price increases and therefore captures only the effect of changes in output. The change in energy consumption associated with shifts in the economy’s structure is referred to as the structural effect. The structural effect captures changes in energy consumption when sectors with different energy intensities grow or decline at different rates, after adjusting for growth in total output. The structural effect can be estimated only where subsectoral data are available. Finally, the change in energy consumption associated with the changes in the amount of energy used to produce each unit of output is referred to as the real intensity effect. This measure provides a good indicator of the progress in reducing energy used to produce goods and services in the economy. This is because the real intensity effect removes the influence of changes in the sectoral composition of the economy, which is already captured through the structural effect. The real energy intensity effect can only be reliably measured at the level of individual production processes or plants where the structural effect can be fully captured. As shown in figure f, the energy intensity indicator can be constructed from the most aggregate level in the economy, in terms of the energy-to-GDP ratio (hierarchy level I), through to the most detailed level for a specific technology or production process (hierarchy level IV). As one moves down the level of hierarchy, the influence of the structural effect on the energy intensity indicator gradually reduces. However, moving down the hierarchy to accurately estimate the energy intensity indicator requires disaggregated data. In the absence of such disaggregated data, energy intensities in this report are estimated at the hierarchy level II. See appendix B for further discussion on the intensity (efficiency) indicators. In this report the analysis is undertaken at the national level for a range of sectors, covering the period from 1989-90 to 2005-06. Changes in energy use associated with each effect are calculated from year to year and then re-based to 1989-90. This means that the activity, structural and real intensity effects can be presented as changes relative to 1989-90. Also, in this report, the amount by which actual energy use is lower than it would otherwise have been as a result of a reduction in real energy intensity is referred to as energy savings. The sectors covered in this report are: manufacturing, comprising 10 manufacturing subsectors; services, comprising seven services subsectors; agriculture, including forestry and fishery; mining; and construction.

Activity and real energy intensity effects can be estimated for each subsector. However, structural effects can only be estimated at the sectoral level because they account for movements between subsectors. If there are no subsectors, then no shift from one subsector to another can be estimated, as is the case for the agriculture, mining and construction sectors. Further, structural shifts can only be estimated when the activity of every subsector (within a particular sector) is measured in identical units. Shifts can then be based on changes in each subsector’s share of total activity from year to year.