Energy Indicators for Sustainable Development: Guidelines and Methodologies International Atomic Energy Agency United Nations Department of Economic and Social Affairs International Energy Agency Eurostat European Environment Agency

E = 
∑
i
A
i
 *
∑
j
 S
i,j
* I
i,j
,
(A3.5) 
where the index i denotes the sectors listed in the first column in Table A3.1. By re-
aggregating the decomposition terms to the national level, interesting comparisons can 
be made of developments in energy per unit of gross domestic product (GDP). If both 
sides of Equation (A3.5) are divided by GDP, then 
E/GDP = ((
∑
i
A
i
 *
∑
j
 S
i,j
)/GDP) *
∑
j
i,
 I
i,j
.
(A3.6) 
The product of the activity effect (A) and the structure effect (S) can be defined as the 
energy services effect. Thus Equation (A3.6) helps explain how energy per unit of 
GDP has changed due to shifts in the ratio of energy services to GDP and due to 
changes in end-use energy intensities. The first factor reflects that the structural 
evolution of economies and human activities can cause changes in demand for energy 
services and, therefore, consumption that enhances or offsets shifts caused by changes 
in energy intensities. For example, air travel measured as passenger-km has grown 
faster than GDP in many countries, usually more than offsetting declines in air travel 
intensity (energy per passenger-km), with an increase in energy use for air travel per 
unit of GDP as a result. On the other hand, structural changes away from energy-
intensive manufacturing industries have enhanced the effect of reduced sectoral 
intensities in many places and thus accelerated a decline in energy per unit of GDP. 
Measuring the impact of these changes in the relationship between energy services 
and GDP is therefore crucial to understanding how the ratio of energy use to GDP 
changes over time.
4
The developments in energy services per GDP indicator help to show how much of 
the change in energy per unit of GDP is due to factors other than changes in energy 
intensities. The impact of intensities at the national level is instead captured by the 
4
The decomposition presented in this annex can be extended to address changes in CO
2
emissions by 
introducing the dimension of fuel mix. This approach can be used to assess how changes in CO
2
emissions per GDP can be decomposed into changes in supply efficiency and fuel mix, final energy 
fuel mix, end-use intensity effect, and ratio of energy services to GDP. The approach thus provides 
a framework for quantifying the relative impact each of these factors has on CO
2
emission per GDP 
trends. Since all of these factors, except the ratio of energy services to GDP, are represented by 
some of the ECO indicators presented in this publication, this decomposition approach can help 
weighing the impact on overall CO
2 
emission trends from the relevant ECO indicators. For more 
details on how to decompose CO
2
emissions see: IEA, 2004 Oil Crises and Climate Challenges: 
30 Years of Energy Use in IEA Countries. Paris, France: International Energy Agency. 

---

160 
energy intensity index at a national level (the I term in Equation A3.6). This is 
constructed through weighting the sectoral energy intensity effects (Equation A3.4) at 
the base-year value of energy use. 
The separation between energy services effects and energy intensity effects is 
important from a policy perspective, since restraining energy-service demand is 
seldom a policy objective. This decomposition approach allows for observing the 
impacts of the policy elements related to energy intensity separately from changes in 
the structural and activity components of energy use. This helps both to determine 
where policies can be most effective and to monitor progress once they have been 
implemented. 

---

161