- Email: [email protected]
Final energy use in IEA countries: The role of energy efficiency
Energy Policy 38 (2010) 6463–6474
Contents lists available at ScienceDirect
Energy Policy journal homepage: www.elsevier.com/locate/enpol
Final energy use in IEA countries: The role of energy efficiency Peter G. Taylor , Olivier Lavagne d’Ortigue, Michel Francoeur, Nathalie Trudeau ´de ´ration, 75739 Paris, France International Energy Agency, 9 rue de la Fe
a r t i c l e in fo
abstract
Article history: Received 8 September 2008 Accepted 3 May 2009
Improved energy efficiency is a key policy goal of all International Energy Agency (IEA) member countries, but tracking energy efficiency gains is not straightforward. As part of its contribution to the G8 Gleneagles Plan of Action, the IEA has been developing in-depth indicators—tools that provide data and analysis of energy use and efficiency trends. This paper gives an overview of the IEA indicator methodology and presents examples of how disaggregated indicators can be used to identify the factors that drive and restrain energy demand at the end-use level. A decomposition approach is also used to separate efficiency effects from the impacts of structure and activity. The results clearly show the important role that energy efficiency has played in shaping trends in final energy use in IEA countries for more than 30 years. However, the analysis also reveals that recent gains in energy efficiency have been much lower than in earlier decades. Accelerating energy efficiency improvements is therefore a crucial challenge for IEA governments and indicators have an important role to play in helping to develop and evaluate the policies that will be required. & 2009 OECD/IEA. Published by Elsevier Ltd. All rights reserved.
Keywords: Energy use Energy efficiency Indicators
1. Introduction Governments in many countries are increasingly aware of the urgent need to make better use of the world’s energy resources. The benefits of more efficient use of energy are well known and include reduced investments in energy infrastructure, lower fossil fuel dependency, increased competitiveness and improved consumer welfare. Efficiency gains can also deliver environmental benefits by reducing greenhouse gas emissions and local air pollution. To support better energy efficiency policy-making and evaluation, the International Energy Agency (IEA) has been working for more than 10 years on developing in-depth indicators—tools that provide state-of-the-art data and analysis on energy use and efficiency trends (IEA, 1997a, b, 2004). The latest phase of work has been undertaken in response to a request from G8 leaders for the IEA to support the Gleneagles Plan of Action, launched in July 2005. In this plan, the leaders addressed the global challenges of tackling climate change, promoting clean energy and achieving sustainable development. The G8-related indicators work has progressively extended the geographical coverage of the analysis to 22 IEA member countries, as well as for the first time including indicators for major developing countries. A significant effort has also been made to develop more detailed indicators for the industry sector. The work programme has resulted in a number of
Corresponding author. Tel.: +33 1 40 57 67 37; fax: +33 1 40 57 67 59.
E-mail address: [email protected] (P.G. Taylor).
publications (IEA, 2007a, b, 2008a) and the key messages have been reported to the G8 leaders at their summit in Hokkaido, Japan in July 2008. The methodological framework and data developed under the indicators project also provide input to other analytical activities undertaken by the agency, such as the scenarios found in Energy Technology Perspectives (IEA, 2008b), the World Energy Outlook (IEA, 2008c), and several other energy efficiency and energy technology projects. An important aim of the IEA work on indicators is to increase the availability of good quality, timely, comparable and detailed data on energy use, which build on a sound statistical energy balance. Such data can provide a firm basis for the development, implementation and monitoring of successful energy efficiency and other related policies, as well as facilitating meaningful international comparisons. Many IEA countries have already recognised the importance of a strong statistical foundation to support their energy indicator activities. Their efforts to collect and release more and better statistics have strengthened IEA work. For European countries the analysis has greatly benefited from the ODYSSEE energy efficiency indicators project, which is funded by the European Commission and country governments. Countries in other IEA regions have also established data collection processes for specific sectors. The situation for non-IEA countries is more challenging, with little or no detailed data available for most countries. However, there are a number of promising developments, but these will take time to deliver results. Several non-IEA countries, notably in Asia-Pacific Economic Cooperation (APEC), have embarked upon programmes to develop indicators that reflect their own
0301-4215/$ - see front matter & 2009 OECD/IEA. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.enpol.2009.05.009
6464
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
situations. The IEA, together with the World Bank, is also currently working on projects with organisations in Mexico, China and South Africa to develop energy efficiency indicators for these countries. Furthermore, the IEA is committed to continue working with interested countries on both indicator development and improvements to the underlying statistics.
IEA Aggregated indicators e.g. Statistics E/GDP
Disaggregated indicators
2. The IEA indicators approach1 The IEA indicators approach uses the idea of an indicators pyramid, which portrays a hierarchy of energy indicators from most detailed to least detailed. This illustrates conceptually how the most detailed and disaggregated data and indicators can be combined to give the more aggregated ones higher up on the pyramid (Fig. 1). In the IEA approach, the top element (the most aggregate indicator) is defined as the ratio of energy use to gross domestic product (GDP). Alternatively, it could have been defined as the ratio of energy use to another macro-economic variable, such as population. Energy use per GDP and per capita are both available from IEA statistics for nearly all countries. The second row of elements can be defined as the energy intensity of each major sector, as measured by energy use per unit of activity in each sector. Lower rows represent the sub-sectors or end-uses that make up each sector and progressively provide more details e.g. characterising particular processes or appliances. Joining each level of energy intensities are structural variables that indicate how to weight these intensities to form a more aggregate parameter of intensity or use. Descending lower down the pyramid requires more data and more complex analysis to reaggregate back up to a higher level. However, each descent also provides a better measure of ‘‘technical’’ energy efficiency, defined for a specific technology, process, and/or end-use. This hierarchy is important because it shows how detailed changes (which may be the result of policies, technological progress, structural reform, or behavioural change) can be linked to higher order, more aggregate quantities, showing how the former affects the latter. With this hierarchy, one can better explain more aggregate changes in energy use in terms of components and more carefully chose the depth of analysis required. That choice depends on the questions that need to be answered, as illustrated later. In its current work, the IEA draws on detailed end-use information about the patterns of energy consumption in more than 20 end-uses covering the manufacturing, household, service and transport sectors of 22 IEA countries over the period from 1990 to 2005 (Fig. 2). This information, coupled with economic and demographic data, is used to identify the factors behind increasing energy use and those that restrain it. The IEA energy indicators typically reflect ratios or quantities and, at a disaggregated level, can describe the links between energy use and human and economic activities. The indicators include measures of activity (such as manufacturing output or volume of freight haulage), measures of developments in structure (such as changes in manufacturing output mix or modal shares in transport) and measures of energy intensity (defined as energy use per unit of activity). To separate the effect on energy use of various components over time, the IEA uses a Laspeyres decomposition approach that analyses changes in energy use within a sector using the following equation: X r r E¼A ðS I Þ r
1
The following section is based on the descriptions in IEA (1997b, 2007b).
e.g. sectoral energy intensity
IEA indicator database
e.g. end-use energy intensity Process / appliance indicators
e.g. unit energy consumption
Limited IEA indicators
Fig. 1. The IEA energy indicators pyramid. Source: Modified from IEA (1997b).
In this decomposition, the symbols represent the following parameters: E: A: r: Sr: Ir:
total energy use in a sector, overall sectoral activity, sub-sectors or end-uses within a given sector, share of sub-sector or end-use ‘‘r’’ in a sector and the energy intensity of each sub-sector or end-use ‘‘r’’.
The separation of impacts on energy use from changes in activity, structure and intensity is critical for policy analysis. Most energy-related policies target energy intensities and efficiencies, often by promoting new technologies. Accurately tracking changes in intensities helps measure the effects of these new technologies. The energy intensity effect (EtI), which is used as a proxy for changes in energy efficiency, separates out how changes in the energy intensity of each sub-sector or end-use influence total energy consumption for a particular sector. This is done by calculating the relative impact on energy use that would have occurred between a base year (t ¼ 0) and a future year (t) if the aggregate activity levels and structure for a sector remained fixed at base year values while the sub-sector or end-use energy intensities followed their actual development (Eq. (1)). A similar approach is used to calculate the activity and structure effects. P r r ðS I Þ EIt ¼ A0 r 0 t (1) E0 The hypothetical energy use (HEUI) is then defined as the energy use that would have occurred in year t if the energy intensities in each sector remained constant at their base year values. It is calculated by dividing actual energy use in year t by the intensity effect in that year (Eq. (2)). HEUIt ¼
Et EIt
(2)
Energy savings from reduced energy intensities can be defined as the difference between the hypothetical energy use and actual energy use (Eq. (3)). SAVINGSIt ¼ HEUIt Et
(3)
Regional aggregates for hypothetical energy use are calculated as the sum of hypothetical energy uses across all countries in a particular region. Energy savings for a region are then calculated as the difference between the hypothetical energy use and the actual energy use.
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
6465
Final Energy Consumption
Residential
Services
Space Heating
Total Services
Water Heating Cooking
Passenger Travel Cars & Light Duty Vehicles Motorcycles1 Buses
Lighting Appliances
Passenger Rail Passenger Ships1
Freight Transport Trucks Freight Rail Domestic Shipping Domestic Air Freight1
Manufacturing
Other Industry2
Food, Beverages & Tobacco
Agriculture, Forestry & Fishing
Paper, Pulp & Printing
Mining1
Industrial Chemicals Non-metallic Minerals
Domestic Planes
Construction Electricity, Gas & Water Distribution1
Primary Metals Metal Products & Equipment Other
Fig. 2. Disaggregation of sectors, sub-sectors and end-uses in IEA energy indicators approach. 1Not included in this study due to lack of consistent and reliable data series. 2 Other industry is included in the analysis only in the section aggregate trends and is not analysed separately. Source: IEA (2007b).
150
3.1. Aggregate trends
140
Economy-wide trends during the period 1990–2005 have been analysed for 16 IEA countries (IEA16)2 for which data are available for all sectors (Fig. 3). Over this time, economic activity (as measured by gross domestic product [GDP]) increased by 41%. Total final energy use increased by only 15%. This partial decoupling of energy use from economic growth resulted in a 19% decrease in final energy intensity (measured as total final energy use per unit of GDP). Since 1990, services have been the sector with the fastest growth in final energy consumption in the IEA16 (Fig. 4); in 2005 it reached a share of 14%. Energy use for both domestic passenger and freight transport has also been increasing rapidly, reaching shares in 2005 of 25% and 11%, respectively. Household energy use has shown a more modest rise and, as a result, its share of final energy consumption has remained constant at 22%. In contrast, energy consumption in manufacturing showed a slight decline. Historically, this sector has been the largest final energy user, but its share has fallen since 1990. In 2005, it represented 25% of the total, the same share as passenger transport. When re-grouped by main end-uses, the shares of final energy consumption in the IEA16 in 2005 are as follows: passenger and freight transport—36%; buildings—36% and manufacturing–25%. Total final energy consumption (TFC) in the IEA16 is dominated by oil, with a share of 47% in 2005—the same as in 1990. Oil use for both passenger and freight transport increased strongly between 1990 and 2005, driving the overall rise in oil demand.
130
2 The 16 IEA countries included in the analysis of the aggregate trends are Australia, Austria, Canada, Denmark, Finland, France, Germany, Italy, Japan, the Netherlands, New Zealand, Norway, Sweden, Switzerland, the United Kingdom and the United States. These countries account for 85% of total final energy use in all IEA countries.
Index (1990 = 100)
3. Understanding trends in energy use and efficiency
GDP PPP 2000 Final energy use Energy use per GDP
120 110 100 90 80 1990
1995
2000
2005
Fig. 3. Overview of key economy-wide trends, IEA16. Source: IEA indicators database.
With a share of 22%, electricity has overtaken natural gas as the second most important energy commodity in the final energy mix (Fig. 5). Electricity use increased rapidly in all stationary sectors, with the strongest growth in households and services. Residential electricity demand is largely driven by increased ownership and use of electric appliances. In particular, there has been a substantial increase in consumption from a wide range of smaller appliances such as home entertainment and kitchen equipment. In the service sector, much of the strong growth has been from air conditioning and lighting, and from various kinds of office and information technology equipment. In contrast, coal consumption has declined and now represents only 5% of total energy use by final consumers. Renewable energy use
6466
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
35
100% Manufacturing
90%
30
29%
25%
80% Passenger transport
25
70%
20 EJ
Households*
60%
24%
25%
50% 15
Services
40% 22%
22%
30%
10 Freight transport
20%
13%
14%
10%
11%
1990
2005
5 10% Other** 0 1990
0% 1995
2000
2005
Fig. 4. Total final energy consumption by sector, IEA16. *Corrected for yearly climate variations. **Other includes construction and agriculture, forestry and fishing. Source: IEA indicators database.
60
100% Oil
50
90% 80%
Electricity
40
EJ
Natural gas
30
47%
19%
22%
60% 50%
Coal
20
40% 30%
Renewables
21% 20%
10 District heat
0 1990
47%
70%
10% 0%
1995
2000
2005
8%
21%
4%
5% 4%
1990
2005
Fig. 5. Total final energy consumption by energy commodity, IEA16. Source: IEA indicators database.
(mostly biomass) has grown, but its share remains only 4% of final energy consumption.
3.2. Drivers of energy use Energy underpins nearly every aspect of a modern economy. Thus, increase in economic activity in all sectors tend to put an upward pressure on energy consumption. Fig. 6 shows the evolution of some of the key drivers of energy consumption across each sector over the period from 1990 to 2005, on a per capita basis. All drivers shown have increased between 1990 and 2005, with annual growth rates of between 1% and 2% for the IEA16 as a whole. The service sector showed particularly strong increases in value-added output per capita for all countries, which is one of the reasons why energy use in services is growing most rapidly. Growth in household area
per capita has increased much more strongly in Japan, Australia and New Zealand than in North America and Europe. Other sectors also demonstrate significant differences among countries, with freight transport and manufacturing having the widest variation in activity increases. Freight haulage in Japan showed particularly low growth rates compared to other countries, partly as a result of increased load factors. Consequently, energy use for freight transport in Japan was virtually unchanged between 1990 and 2005. In manufacturing, North America has experienced much higher growth rates per capita than either Japan or Europe. However, the impact of this growth on energy use has been largely offset by significant reductions in energy intensities. The evolution of fossil fuel prices is a further important factor affecting energy consumption (Fig. 7). These prices are strongly influenced by trends in spot oil prices. Most of the 1990s was characterised by decreasing real prices of crude oil, but since 2003
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
Australia
Average annual percent change
3.0%
Canada
New Zealand
Japan
6467
US
EUR13*
2.5%
2.0%
1.5%
1.0%
0.5%
ic e pe s v r c alu ap eita ad d
se
rv
ou
Se
H
M an ad ufa de ct d uri p e ng rc v ap alu i t a e-
ed
r pe
rc pe m -k To n
ne
ss Pa
ho ca ld pi a r e ta a
ap
en ca ger pi -km ta p
er
ita
0.0%
Fig. 6. Changes in key sub-sectoral activity levels relative to population, 1990–2005. *EUR13 includes Austria, Denmark, Finland, France, Germany, Greece, Italy, the Netherlands, Norway, Spain, Sweden, Switzerland and the United Kingdom. Source: IEA indicators database.
Canada US
450
350
Crude oil
8
Steam coal
250 200 150 100 50 0 1990
Italy UK
9
Liquified natural gas 300
Japan France
Gaseous natural gas MJ per USD PPP 2000
USD MER 2000 per toe
400
IEA16 Germany
7 6 5 4
1995
2000
2005 2007
Fig. 7. Fossil fuel import prices in real terms. Source: IEA (2008d).
3 1990
1995
2000
2005
Fig. 8. Final energy use per unit of GDP. Source: IEA indicators database.
prices have risen very steeply. Real prices for natural gas and, to some extent, coal have followed oil price developments, although with less significant fluctuations. Real energy prices decreased between 1990 and 1998; in many cases, they did not return to their 1990 levels until around 2005. Such price developments are unlikely to have provided much incentive for improvements in energy efficiency during the 1990s. Since 2003 prices have risen more steeply and these increases could be expected to have a significant effect on end-use energy efficiency and end-use fuel mix, as a result of behavioural changes across all sectors. However, detailed energy consumption data for years after 2005 were not available to confirm this at the time of writing.
3.3. Changes in aggregate energy intensity The ratio of TFC per unit of GDP provides a measure of final energy intensity for a country, and is one of the most frequently
used aggregate energy indicators. Trends in final energy intensities show that the ratio of TFC to GDP has declined in most IEA countries since 1990 (Fig. 8). However, both the absolute levels of the ratio and the rates at which it has fallen vary significantly. Among the largest IEA economies, the United Kingdom showed the strongest decline, with the ratio of TFC to GDP falling by 24% between 1990 and 2005, mostly due to changes in the structure of the economy. At the other end of the scale, Italy actually showed an increase of 1% over the same period. In a number of IEA countries, the TFC/GDP ratio fell most rapidly during the mid- to late 1990s, a period of rapid economic growth. Trends in final energy intensity are often used to assess the extent to which energy efficiency is improving in countries. However, this can often be misleading as the ratio is affected by many factors such as climate, geography, travel distance, home
6468
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
45
GJ per capita (normalised to 2700 HDD)
40 35 30 25 20 15 10 5
Au
st
ra
19 Au lia 90 2 st ria 005 19 9 C an 20 0 ad 0 a 19 5 D 90 en 20 m 0 ar k 19 5 90 Fi 2 nl 00 an d 19 5 9 Fr 20 0 an ce 1 05 99 G er 20 0 m an 0 y 19 5 90 Ire 20 la 0 nd 19 5 9 20 0 Ita ly 1 05 99 Ja 20 0 pa 0 n 19 5 9 Ko 20 0 re N 05 et a 1 he 99 rla 0 nd 200 N ew s 1 5 99 Ze 20 0 al an 0 d 19 5 9 N or 20 0 w ay 1 05 99 Sp 20 0 ai n 1 05 99 Sw 20 0 e Sw den 1 05 99 itz er 20 0 la nd 1 05 99 20 0 U 1 05 K 9 9 20 0 U 1 05 S 9 9 20 0 05
0
Appliances
Lighting
Cooking
Water heating
Space heating
Total (not normalised)
Fig. 9. Household energy use per capita. Source: IEA indicators database.
size and manufacturing structure, which energy efficiency measures cannot directly influence. As explained earlier, more detailed indicators targeted at the individual end-use sectors are needed to get a better picture of what is happening in relation to energy efficiency. Some examples of these types of indicators are shown in the following sections.
3.4. Buildings
For the IEA19 as a whole, the useful energy per area heated declined by 17% between 1990 and 2005, largely as a result of improvements to the building shell that have reduced heat losses. This is much greater than the fall in space heating per capita, as the upward pressure on energy demand from larger homes and lower occupancy levels has largely offset the benefits of a decline in space heating intensity. Looking at the variation amongst countries, it can be seen that a number of countries with relatively cold climates (e.g. Finland, Canada and Norway) have intensities that are below the average for the IEA19. This can partly be explained by higher levels of insulation in these countries. In contrast, France, Italy and Germany, which have more moderate winters, showed the highest levels of energy intensity. This could be the result of a greater percentage of poorly insulated older buildings.
Buildings account for just over one-third of total final energy use in IEA countries. However, the buildings share of total electricity consumption is much higher at 65%. Space heating is by far the most important energy user in the residential sector of a group of 19 IEA countries (IEA19),3 accounting for 53% of final household energy use in 2005. The significance of space heating varies according to climate. Fig. 9 shows per capita residential energy use with space heating both adjusted and not adjusted to account for climatic differences. The amount of energy used for space heating differs widely among countries, due to differences in house size, expected indoor heating comfort, type of heating equipment and levels of insulation. Overall, for the IEA19, space heating per capita fell by 2.4% between 1990 and 2005. To understand better these variations and trends, it is useful to develop indicators for the components that affect space heating demand. A key indicator in this regard is the level of space heating intensity. This is measured in terms of ‘‘useful energy’’ for space heating per square metre, with useful energy being calculated as final energy minus the losses estimated for boilers. To allow for comparisons across countries with different climates, the space heating intensity is divided by each country’s yearly number of heating degree-days (Fig. 10).
The energy intensity of manufacturing industry (as measured by dividing total manufacturing energy use by total manufacturing value-added) varies widely between sub-sectors. The most energyintensive sub-sector is the production of primary metals. In 2005, energy intensity in this sub-sector was more than 10 times higher than the category of ‘‘other manufacturing’’ (Fig. 11). Yet output from the primary metals sub-sector constituted only 4% of total manufacturing value-added, compared to more than 29% from other manufacturing. Taken together, the four most energy-intensive subsectors (primary metals, non-metallic minerals, chemicals and paper and pulp) accounted for almost 70% of manufacturing energy use in a group of 21 IEA countries (IEA21),4 despite representing only 29% of total value-added output.
3 The 19 IEA countries included in the analysis of buildings are Australia, Austria, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, the Republic of Korea, the Netherlands, New Zealand, Norway, Spain, Sweden, Switzerland, the United Kingdom and the United States.
4 The 21 IEA countries included in the analysis of buildings are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Italy, Japan, the Republic of Korea, the Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, the United Kingdom and the United States.
3.5. Manufacturing
an a Fi da nl a N No nd ew r w Ze a al y Au and st Po ralia rtu Be ga lg l iu m N et he US rla Sw nd ed s Ko en re IE a A2 Sp 1 ai n U Au K st ria Ita Fr ly an c Ja e p D en an G ma er rk m a G n Sw re y itz ece er la nd
C
MJ per USD PPP 2000
& , be to v ba er cc ag o es Pa & pe pr r, p in u tin lp C g he m ic al s N on m m in et er al al lic Pr s im ar y m et M al e s & tal eq pr ui od pm u en cts t m an O uf the ac r tu rin g m an T o uf ta ac l tu rin g
od
Fo
MJ per USD PPP 2000 an y Ita l Fr y an Ire ce la n Ko d re a U K Au Au stria s D trali en a m Sw ar ed k IE en A Fi 19 nl a C n N an d et ad h Sw erla a itz nd er s la nd U Sp S N ain o N ew rw Ze ay al a Ja nd pa n
er m
G
kJ per m2 per HDD
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
Energy intensities
50 100%
40 80%
30 60%
20 40%
10 20%
0 0%
Value-added shares
2005 actuaI intensities
6469
250
200 1990 2005
150
100
50
0
Fig. 10. Useful space heating intensity. Source: IEA indicators database.
Energy shares
Fig. 11. Sub-sector energy intensities, value-added and energy shares, IEA21. Source: IEA indicators database.
18
16 2005 common structure intensities
14
12
10
8
6
4
2
0
Fig. 12. Manufacturing energy intensity at actual and common structures. Source: IEA indicators database.
6470
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
Car fuel intensity (litres of gasoline equivalent per 100 vkm)
12 US 11 Australia
Japan Canada New Zealand
10
9 Switzerland 8
Sweden
Ireland
Norway Germany Austria
7
France
6 0.5
UK
Finland
Denmark
Netherlands
Italy
Greece 0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
Retail average fuel price (USD PPP 2000 per litre) Fig. 13. Car fuel intensity versus average fuel price, 2005. Source: IEA indicators database.
3.6. Transport In IEA countries, the transport sector is taking an increasing share of oil demand and is currently responsible for almost 78% of total final oil consumption. Passenger cars and light trucks account for over half of IEA oil demand for transport, and have been the main driving force behind the growth in oil demand over the last three decades. Reducing the fuel intensity of cars (i.e. increasing the fuel efficiency) is therefore a key way of helping
US
Canada
Australia & NZ
IEA18
Japan
EUR13*
13 Litres of gasoline equivalent per 100 vehicle-km
Large differences are also evident in the level of overall manufacturing energy intensity among IEA countries. Using an indicator approach it is possible to investigate to what extent these differences among countries can be explained by differences in their industrial structure. To examine this question, the effects of structural differences are removed by calculating what the energy intensity of each country would be if they all had a common structure (e.g. the average of the IEA21 countries, based on a breakdown of manufacturing into seven sub-sectors). Fig. 12 shows the results of such a calculation for 2005, and compares these intensities at common structure with actual energy intensities in that year. The results are quite striking; even though they do not account for the detailed structural differences within the sub-sectors of industry (i.e. different product mixes within sub-sectors). In 2005, actual manufacturing energy intensities varied by a factor of more than five amongst countries (3.5–17.8 MJ per USD). However, at common structure, the variation between countries with the highest and lowest intensities falls to a factor of less than three (3.8–10.8 MJ per USD). Thus structural effects are responsible for almost half the variation in manufacturing energy intensities that are observed amongst countries. The results are particularly significant for some countries. For instance, this approach shows that Australia’s very high energy intensity can be largely explained by the structure of its manufacturing industry, which has a high share of very energyintensive industries. If Australia’s industry had the same structure as the average for the IEA21 countries – but kept its actual level of energy intensity in each sub-sector – the country’s aggregate manufacturing energy intensity would be reduced by 45%. Similar, but less dramatic, results are also observed for Belgium, Canada, Greece, the Netherlands, Norway and Spain.
12 11 10 9 8 7 1990
1995
2000
2005
Fig. 14. Average fuel intensity of the cars stock. Source: IEA indicators database.
both to reduce dependence on oil and to limit CO2 emissions from the transport sector. Indicators can help examine some of the factors that impact the fuel efficiency of cars and track the progress of policies to help improve efficiency. Perhaps not surprisingly, fuel prices are an important factor influencing fuel intensity. Fuel prices vary considerably across IEA countries, with gasoline prices in 2005 ranging from USD 0.51 to USD 1.44 per litre (using 2000 PPP exchange rates). Variations in fuel taxes are the main reason for this wide gap. Three groups of countries can be identified: the United States, with the lowest prices (and the lowest tax levels); European countries, with generally high prices (and high taxes); and, in the middle, other IEA countries such as Australia, Canada, Japan and New Zealand. Plotting the fuel intensity of cars against fuel prices shows a clear relationship; the average fuel intensity of cars tends to be higher in countries with low fuel prices (Fig. 13). So the United States has the lowest fuel prices and cars with the highest fuel intensity. In contrast, fuel intensity is lowest in Europe, where fuel prices are much higher.
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
3.7. Factors impacting energy use: the role of energy efficiency The previous sections highlighted how different types of indicators can be used to examine trends in energy use and efficiency, both at an aggregate level and for individual end-use sectors. This section shows how the decomposition approach presented earlier can be used to explain how different factors impact energy use and, in particular, to quantify the role of energy efficiency improvements. Take as an example the large variations among countries in how much energy per unit of GDP has fallen over time. To understand the extent to which this variation reflects differences in energy efficiency developments, it is necessary to separate the impact of changes in sub-sectoral energy intensities (which are used as a proxy for energy efficiency) from the effects of changes in economic structure and other factors that influence the demand for energy. Increased demand for energy services – reflecting increased ownership levels of electric household appliances, bigger houses, more personal travel by car, etc. – drives both energy use and energy use per GDP. Therefore, it is useful to examine how the ratio of energy to GDP is affected by changes in structure (as measured by the ratio of energy services to GDP) and in end-use intensities (such as the energy used to heat a square metre of floor space or car energy use per passenger-kilometre). 5 The 18 IEA countries included in the analysis of transport are Australia, Austria, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, the Netherlands, New Zealand, Norway, Sweden, Switzerland, the United Kingdom and the United States.
100 Energy services per GDP
Index (1990 = 100)
However, fuel prices are not the only factor impacting fuel intensities. A number of other factors such as vehicle technologies, the effect of driving conditions and fuel efficiency policies also play an important role in both the absolute levels and trends observed. The average fuel intensities of cars decreased in most countries between 1990 and 2005 (Fig. 14). In Europe, this was due to a combination of factors. The 1990s were characterised by the widespread diffusion of vehicles equipped with electronic control systems for fuel management and by stronger consumer demand for more efficient cars—a reaction to high fuel prices. Since the early 2000s, intensities declined further in Europe as a result of increased sales of direct-injection diesel cars. Despite showing a decrease, the fuel intensity of cars in the United States remained higher than the average of a group of 18 IEA countries (IEA18),5 at nearly 11.5 l of gasoline equivalent per 100 vehiclekilometres. High levels of fuel intensity also characterised Australia, Canada, New Zealand and Japan. In Japan, until the late 1990s, efficiency improvements of new vehicles had been offset by increase in vehicle weight and congestion-related effects. The increasing weight of vehicles has been another factor offsetting improvements in the underlying efficiency of new car engine technologies. Over the last 15 years, the average size and weight of the stock of cars increased as larger and heavier vehicles, such as SUVs, became more popular. This trend, combined with additional safety features also increasing weight, has tended to raise the energy consumption of cars. Data for Canada show that, in 2005, light trucks accounted for one-third of all cars, up from a share of one-fifth in 1990. In the United States, the stock of passenger cars remained almost stable between 1990 and 2005, while that of light trucks nearly doubled. In European countries, the number of cars with an engine capacity greater than 2 l has more than doubled since 1990. Japan reported a much lower increase in the weight of cars, mainly because of effective regulation of fuel efficiency under the Top Runner programme, which came into effect in the late 1990s.
6471
Decline in energy per GDP from fall in energy services per GDP
95
90
Actual energy per GDP
Decline in energy per GDP from fall in subsectoral energy intensities
85
80 1990
1995
2000
2005
Fig. 15. TFC per GDP and the contribution of changes in energy services per GDP and energy intensities, IEA16. Source: IEA indicators database.
Changes in energy consumption per GDP can be attributed to changes in the ratio of energy services to GDP and to changes in sub-sector energy intensities for more than 20 end-uses. The intensity effect for the whole economy is calculated as the aggregate impact of the sectoral intensity effects. The results of aggregate impact calculations show that the energy intensity effect and changes in structure have both contributed to reduced energy consumption per unit of GDP (Fig. 15). However, declining end-use intensities (the energy intensity effect) have been the most important factor. In the early 1990s, structural effects made almost no contribution to the falling energy to GDP ratio; the economy of the IEA was growing slowly, as was demand for energy services. The contribution of structural changes has grown since then but, by 2005, more than 60% of the total decline in energy per GDP could be attributed to reductions from the energy intensity effect. This equates to an annual average fall in end-use intensities across all sectors of 0.9% per year. The relative contribution of structural and intensity effects to the overall trend varies among countries (Fig. 16). With the exception of Italy, all countries show that the intensity effect contributed to reducing the ratio of energy use to GDP: for most countries, it was the dominant factor. This is particularly true in the case of Canada, the Netherlands, Germany, New Zealand, Sweden and the United States. In contrast, for Norway and the United Kingdom, changes in structure were most important. For those countries having declining sub-sectoral intensities, it is possible to identify three groups according to the magnitude of the intensity effect:
less than 0.5% per year: Austria, Denmark, Japan, Norway; 0.5–1%: Australia, France, Finland, Switzerland and the United Kingdom and
greater than 1%: Canada, Germany, the Netherlands, New Zealand, Sweden and the United States. It can be informative to compare these results with the information on the absolute levels of final energy intensity (Fig. 8). Such a comparison shows that those countries with a high level of energy use per GDP in 1990 tend to have had the largest reductions in intensity. In contrast, those countries that initially had lower energy use per GDP have generally seen smaller declines in intensity. Although the intensity effect for the whole economy declined by an average of 0.9% per year, the rate of reduction was not uniform across end-use sectors. The manufacturing sector showed the largest reductions in intensities (corrected for structure), decreasing by 1.4% per year between 1990 and 2005 (Fig. 17). Building energy use showed lower reductions, with average
6472
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
0.5% Average annual percent change
Intensity effect
Energy services per GDP
Energy per GDP
0.0% -0.5% -1.0% -1.5% -2.0%
6 A1
U K U S
IE
Au
st
ra l Au ia st r C ia an ad D en a m a Fi rk nl an Fr d an c G er e m an y Ita ly Ja pa N et n he rla N ew n Ze ds al a N nd or w a Sw y e Sw de n itz er la nd
-2.5%
Fig. 16. Changes in TFC/GDP decomposed into changes in energy services/GDP and intensity effect, 1990–2005. Source: IEA indicators database.
105
Passenger transport
Total
18
Freight transport
Services
16
Households*
Manufacturing
14
Manufacturing Households* Services Passenger transport Freight transport
EJ
12
Index (1990 = 100)
100
10 8 6 4
95
2 0 1995
90
2000
2005
Fig. 18. Contribution to energy savings from the end-use sectors, IEA16. *Corrected for yearly climate variations. Source: IEA indicators database.
85
80 1990
1995
2000
2005
Fig. 17. Sector intensities and total economy effect, IEA16. *Corrected for yearly climate variations. Source: IEA indicators database.
intensity falls of 0.8% per year in households and 1.2% in the service sector. The passenger and freight transport sectors experienced similar declining trends between 1990 and 2005, but had the lowest falls in intensities, both averaging 0.5% per year (calculated holding the modal mix constant).
3.8. Energy savings from improved energy efficiency The decline in energy intensities (i.e. improvements in energy efficiency) in the various end-use sectors led to energy savings across the whole economy. Energy savings in 2005 were about 16 EJ (370 Mtoe) or 15% of total final energy use in that year, which represents an estimated USD 180 billion of energy cost savings (Fig. 18). The impact of energy intensity improvements on CO2 emissions was also significant, with a saving of 1.3 Gt CO2 (14% of actual CO2 emissions in 2005). These savings are approximately equal to the annual final energy consumption and CO2 emissions of Japan.
Fig. 18 shows the total energy savings since 1990, broken down by sector. In 2005, the largest share was from manufacturing (42%), followed by households and services (both 19%), passenger transport (14%) and freight transport (6%). Initially, there were strong energy savings in households as a result of significant improvements in space heating intensity, which did not continue. In contrast, the savings from services made an impact only in the late 1990s, during a period of high economic growth in this sector. The rates of savings in the other sectors were more constant over time. 3.9. Long-term trends Fig. 19 shows that over the longer term, the savings from improved energy efficiency are even more significant. Without the energy efficiency improvements that occurred between 1973 and 2005, energy use in a group of 11 IEA countries for which longterm times series data are available (IEA11)6 would have been 58%, or 59 EJ, higher in 2005 than it actually was. This makes energy savings the most important ‘‘fuel’’ in the IEA11 for this time period—i.e. the amount of energy saved in 2005 was slightly higher than the actual consumption of oil, or of electricity and natural gas combined. 6 The 11 IEA countries included in the analysis of long-term trends are Australia, Denmark, Finland, France, Germany, Italy, Japan, Norway, Sweden, the United Kingdom and the United States.
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
180
6473
2%
140
Hypothetical energy use without energy efficiency improvements Savings
58%
120
EJ
100 80 Actual energy use
60 40
Average annual percent change
2.0%
160
1% 0.9% 0.8%
0.5%
20 0 1973
0% 1980
Actual energy use
1990
2000
2005
Energy savings due to energy efficiency improvements
1973-1990 1990-2005 Energy efficiency improvements
Fig. 19. Long-term energy savings from improvements in energy efficiency, all sectors, IEA11. Source: IEA indicators database.
A similar pattern emerges from analysis of the individual enduse sectors. Energy efficiency improvements have led to significant savings in all sectors. However, without exception, the rate of improvement since 1990 has been lower than in the period between 1973 and 1990. In manufacturing, the strong energyefficiency improvements between 1973 and 1990 led to a decrease in energy use, but this reduction did not continue into the 1990s. For households and services, lower energy-efficiency improvements since 1990 have led to significantly higher increases in energy use between 1990 and 2005. For both passenger and freight, the reduction in the rate of improvements in energy efficiency since 1990 has been partially compensated by lower increases in energy service demand. For passenger transport, the rate of increase in energy use changed little between the two periods. The rate of increase in energy use for freight transport actually fell after 1990.
4. Conclusions and policy implications This article has highlighted key aspects of the IEA work on energy indicators undertaken in support of the Gleneagles Plan of Action. It has shown the importance of using disaggregated energy indicators for analysing interactions between economic and human activity and energy use. Such indicators provide a powerful set of analytical tools that reveal important trends in energy use and efficiency that are obscured when only more aggregate quantities, such as the ratio of energy use to GDP, are used. Many IEA member countries already use energy indicators and they are also attracting increasing interest from other countries. The IEA is working to assist and internationalise these efforts by developing transparent and consistent international databases and methodologies and by collaborating with governments, industry and regional and international organisations. The results from the indicators analysis highlight that energy efficiency continues to play a key role in limiting increases in energy use in IEA countries. However, they also reveal that the rate at which energy efficiency has improved since 1990 has been much slower than in the previous two decades. This shows that the changes caused by the oil price shocks in the 1970s and the resulting energy policies did considerably more to control
growth in energy demand and reduce CO2 emissions than the energy efficiency and climate policies implemented since the 1990s. More detailed analysis reveals many other developments of importance to policy-makers. Disaggregated indicators have been used to track trends in energy efficiency at the end-use level, as well as quantifying the impact that non-energy factors have over energy consumption. For example, the analysis shows that while household space heating intensities have declined in many countries, the benefits of this improved efficiency have been largely offset by the trend towards larger homes and lower occupancy levels. In the manufacturing sector, differences in the structure of industry can explain almost half the variation observed in the manufacturing energy intensities amongst IEA countries. While in transport, a comparison amongst countries highlights a clear correlation between higher fuel prices and a more efficient stock of cars. Projections published by the IEA in Energy Technology Perspectives 2008 (IEA 2008b) clearly demonstrate that rate of improvement in energy efficiency will need to be substantially increased to have a realistic chance of a more sustainable energy future. The good news is that this is indeed possible; there is still significant scope in IEA countries for energy-efficiency improvements in buildings, industry and transport. However, a key lesson of the past is that the widespread future deployment of more efficient technologies and practices will require very strong government action. The challenge is to find the right mix of market- and regulation-based policies to achieve cost-effective decisions regarding energy efficiency in all sectors. Based on its indicators analysis and other work on best practices in policies and measures, the IEA has presented a list of high-priority energy efficiency policy recommendations to help governments increase rates of energy efficiency improvement in buildings, appliances, lighting, transport, industry, power utilities and cross-sectoral areas (IEA, 2008e). To make these suggested policies a reality, governments will need to provide adequate resources (both financial and human) to the policy agencies responsible for energy efficiency. This should include funding and managing the implementation and monitoring of action plans for energy efficiency. In turn, this will require the establishment and ongoing maintenance of high quality energy indicators and other data; recognising that, while there have been some improve-
6474
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
ments, the availability, timeliness, quality and comparability of data in many countries and sectors are still not sufficient. Building on the success of the Gleneagles Plan of Action, the IEA is committed to support governments and other interested parties in this important work. In the statement issued following their Hokkaido meeting in July 2008, the G8 leaders requested that the IEA ‘‘enhance its work on voluntary sectoral indicators through improved data collection, complemented by business initiatives’’. The IEA recognises that such enhancements will not be easy, but equally believes that failure to provide a robust framework of data collection and indicator development will ultimately undermine the ability of analysts and policy-makers to formulate, implement and monitor successful energy efficiency and other policies.
Acknowledgements The authors would like to thank the governments, organisations, companies and industry associations who have helped with collecting and validating the underlying data used in the IEA
indicator analysis. In particular, we are grateful for the close collaboration of the statisticians and analysts in IEA member countries, including experts from the European Union sponsored ODYSSEE network. References IEA (International Energy Agency), 1997a. The Link between Energy and Human Activity. OECD/IEA, Paris. IEA, 1997b. Indicators of Energy Use and Efficiency—Understanding the Link between Energy and Human Activity. OECD/IEA, Paris. IEA, 2004. Oil Crises and Climate Challenges: 30 Years of Energy Use in IEA Countries. OECD/IEA, Paris. IEA, 2007a. Tracking Industrial Energy Efficiency and CO2 Emissions. OECD/IEA, Paris. IEA, 2007b. Energy Use in the New Millennium: Trends in IEA Countries. OECD/IEA, Paris. IEA, 2008a. Worldwide Trends in Energy Use and Efficiency: Key Insights from IEA Indicators Analysis. OECD/IEA, Paris. IEA, 2008b. Energy Technology Perspectives 2008: Scenarios and Strategies to 2050. OECD/IEA, Paris. IEA, 2008c. World Energy Outlook 2008. OECD/IEA, Paris. IEA, 2008d. Energy Prices and Taxes, 2nd Quarter 2008. OECD/IEA, Paris. IEA, 2008e. Energy Efficiency Policy Recommendations. OECD/IEA, Paris.
Contents lists available at ScienceDirect
Energy Policy journal homepage: www.elsevier.com/locate/enpol
Final energy use in IEA countries: The role of energy efficiency Peter G. Taylor , Olivier Lavagne d’Ortigue, Michel Francoeur, Nathalie Trudeau ´de ´ration, 75739 Paris, France International Energy Agency, 9 rue de la Fe
a r t i c l e in fo
abstract
Article history: Received 8 September 2008 Accepted 3 May 2009
Improved energy efficiency is a key policy goal of all International Energy Agency (IEA) member countries, but tracking energy efficiency gains is not straightforward. As part of its contribution to the G8 Gleneagles Plan of Action, the IEA has been developing in-depth indicators—tools that provide data and analysis of energy use and efficiency trends. This paper gives an overview of the IEA indicator methodology and presents examples of how disaggregated indicators can be used to identify the factors that drive and restrain energy demand at the end-use level. A decomposition approach is also used to separate efficiency effects from the impacts of structure and activity. The results clearly show the important role that energy efficiency has played in shaping trends in final energy use in IEA countries for more than 30 years. However, the analysis also reveals that recent gains in energy efficiency have been much lower than in earlier decades. Accelerating energy efficiency improvements is therefore a crucial challenge for IEA governments and indicators have an important role to play in helping to develop and evaluate the policies that will be required. & 2009 OECD/IEA. Published by Elsevier Ltd. All rights reserved.
Keywords: Energy use Energy efficiency Indicators
1. Introduction Governments in many countries are increasingly aware of the urgent need to make better use of the world’s energy resources. The benefits of more efficient use of energy are well known and include reduced investments in energy infrastructure, lower fossil fuel dependency, increased competitiveness and improved consumer welfare. Efficiency gains can also deliver environmental benefits by reducing greenhouse gas emissions and local air pollution. To support better energy efficiency policy-making and evaluation, the International Energy Agency (IEA) has been working for more than 10 years on developing in-depth indicators—tools that provide state-of-the-art data and analysis on energy use and efficiency trends (IEA, 1997a, b, 2004). The latest phase of work has been undertaken in response to a request from G8 leaders for the IEA to support the Gleneagles Plan of Action, launched in July 2005. In this plan, the leaders addressed the global challenges of tackling climate change, promoting clean energy and achieving sustainable development. The G8-related indicators work has progressively extended the geographical coverage of the analysis to 22 IEA member countries, as well as for the first time including indicators for major developing countries. A significant effort has also been made to develop more detailed indicators for the industry sector. The work programme has resulted in a number of
Corresponding author. Tel.: +33 1 40 57 67 37; fax: +33 1 40 57 67 59.
E-mail address: [email protected] (P.G. Taylor).
publications (IEA, 2007a, b, 2008a) and the key messages have been reported to the G8 leaders at their summit in Hokkaido, Japan in July 2008. The methodological framework and data developed under the indicators project also provide input to other analytical activities undertaken by the agency, such as the scenarios found in Energy Technology Perspectives (IEA, 2008b), the World Energy Outlook (IEA, 2008c), and several other energy efficiency and energy technology projects. An important aim of the IEA work on indicators is to increase the availability of good quality, timely, comparable and detailed data on energy use, which build on a sound statistical energy balance. Such data can provide a firm basis for the development, implementation and monitoring of successful energy efficiency and other related policies, as well as facilitating meaningful international comparisons. Many IEA countries have already recognised the importance of a strong statistical foundation to support their energy indicator activities. Their efforts to collect and release more and better statistics have strengthened IEA work. For European countries the analysis has greatly benefited from the ODYSSEE energy efficiency indicators project, which is funded by the European Commission and country governments. Countries in other IEA regions have also established data collection processes for specific sectors. The situation for non-IEA countries is more challenging, with little or no detailed data available for most countries. However, there are a number of promising developments, but these will take time to deliver results. Several non-IEA countries, notably in Asia-Pacific Economic Cooperation (APEC), have embarked upon programmes to develop indicators that reflect their own
0301-4215/$ - see front matter & 2009 OECD/IEA. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.enpol.2009.05.009
6464
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
situations. The IEA, together with the World Bank, is also currently working on projects with organisations in Mexico, China and South Africa to develop energy efficiency indicators for these countries. Furthermore, the IEA is committed to continue working with interested countries on both indicator development and improvements to the underlying statistics.
IEA Aggregated indicators e.g. Statistics E/GDP
Disaggregated indicators
2. The IEA indicators approach1 The IEA indicators approach uses the idea of an indicators pyramid, which portrays a hierarchy of energy indicators from most detailed to least detailed. This illustrates conceptually how the most detailed and disaggregated data and indicators can be combined to give the more aggregated ones higher up on the pyramid (Fig. 1). In the IEA approach, the top element (the most aggregate indicator) is defined as the ratio of energy use to gross domestic product (GDP). Alternatively, it could have been defined as the ratio of energy use to another macro-economic variable, such as population. Energy use per GDP and per capita are both available from IEA statistics for nearly all countries. The second row of elements can be defined as the energy intensity of each major sector, as measured by energy use per unit of activity in each sector. Lower rows represent the sub-sectors or end-uses that make up each sector and progressively provide more details e.g. characterising particular processes or appliances. Joining each level of energy intensities are structural variables that indicate how to weight these intensities to form a more aggregate parameter of intensity or use. Descending lower down the pyramid requires more data and more complex analysis to reaggregate back up to a higher level. However, each descent also provides a better measure of ‘‘technical’’ energy efficiency, defined for a specific technology, process, and/or end-use. This hierarchy is important because it shows how detailed changes (which may be the result of policies, technological progress, structural reform, or behavioural change) can be linked to higher order, more aggregate quantities, showing how the former affects the latter. With this hierarchy, one can better explain more aggregate changes in energy use in terms of components and more carefully chose the depth of analysis required. That choice depends on the questions that need to be answered, as illustrated later. In its current work, the IEA draws on detailed end-use information about the patterns of energy consumption in more than 20 end-uses covering the manufacturing, household, service and transport sectors of 22 IEA countries over the period from 1990 to 2005 (Fig. 2). This information, coupled with economic and demographic data, is used to identify the factors behind increasing energy use and those that restrain it. The IEA energy indicators typically reflect ratios or quantities and, at a disaggregated level, can describe the links between energy use and human and economic activities. The indicators include measures of activity (such as manufacturing output or volume of freight haulage), measures of developments in structure (such as changes in manufacturing output mix or modal shares in transport) and measures of energy intensity (defined as energy use per unit of activity). To separate the effect on energy use of various components over time, the IEA uses a Laspeyres decomposition approach that analyses changes in energy use within a sector using the following equation: X r r E¼A ðS I Þ r
1
The following section is based on the descriptions in IEA (1997b, 2007b).
e.g. sectoral energy intensity
IEA indicator database
e.g. end-use energy intensity Process / appliance indicators
e.g. unit energy consumption
Limited IEA indicators
Fig. 1. The IEA energy indicators pyramid. Source: Modified from IEA (1997b).
In this decomposition, the symbols represent the following parameters: E: A: r: Sr: Ir:
total energy use in a sector, overall sectoral activity, sub-sectors or end-uses within a given sector, share of sub-sector or end-use ‘‘r’’ in a sector and the energy intensity of each sub-sector or end-use ‘‘r’’.
The separation of impacts on energy use from changes in activity, structure and intensity is critical for policy analysis. Most energy-related policies target energy intensities and efficiencies, often by promoting new technologies. Accurately tracking changes in intensities helps measure the effects of these new technologies. The energy intensity effect (EtI), which is used as a proxy for changes in energy efficiency, separates out how changes in the energy intensity of each sub-sector or end-use influence total energy consumption for a particular sector. This is done by calculating the relative impact on energy use that would have occurred between a base year (t ¼ 0) and a future year (t) if the aggregate activity levels and structure for a sector remained fixed at base year values while the sub-sector or end-use energy intensities followed their actual development (Eq. (1)). A similar approach is used to calculate the activity and structure effects. P r r ðS I Þ EIt ¼ A0 r 0 t (1) E0 The hypothetical energy use (HEUI) is then defined as the energy use that would have occurred in year t if the energy intensities in each sector remained constant at their base year values. It is calculated by dividing actual energy use in year t by the intensity effect in that year (Eq. (2)). HEUIt ¼
Et EIt
(2)
Energy savings from reduced energy intensities can be defined as the difference between the hypothetical energy use and actual energy use (Eq. (3)). SAVINGSIt ¼ HEUIt Et
(3)
Regional aggregates for hypothetical energy use are calculated as the sum of hypothetical energy uses across all countries in a particular region. Energy savings for a region are then calculated as the difference between the hypothetical energy use and the actual energy use.
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
6465
Final Energy Consumption
Residential
Services
Space Heating
Total Services
Water Heating Cooking
Passenger Travel Cars & Light Duty Vehicles Motorcycles1 Buses
Lighting Appliances
Passenger Rail Passenger Ships1
Freight Transport Trucks Freight Rail Domestic Shipping Domestic Air Freight1
Manufacturing
Other Industry2
Food, Beverages & Tobacco
Agriculture, Forestry & Fishing
Paper, Pulp & Printing
Mining1
Industrial Chemicals Non-metallic Minerals
Domestic Planes
Construction Electricity, Gas & Water Distribution1
Primary Metals Metal Products & Equipment Other
Fig. 2. Disaggregation of sectors, sub-sectors and end-uses in IEA energy indicators approach. 1Not included in this study due to lack of consistent and reliable data series. 2 Other industry is included in the analysis only in the section aggregate trends and is not analysed separately. Source: IEA (2007b).
150
3.1. Aggregate trends
140
Economy-wide trends during the period 1990–2005 have been analysed for 16 IEA countries (IEA16)2 for which data are available for all sectors (Fig. 3). Over this time, economic activity (as measured by gross domestic product [GDP]) increased by 41%. Total final energy use increased by only 15%. This partial decoupling of energy use from economic growth resulted in a 19% decrease in final energy intensity (measured as total final energy use per unit of GDP). Since 1990, services have been the sector with the fastest growth in final energy consumption in the IEA16 (Fig. 4); in 2005 it reached a share of 14%. Energy use for both domestic passenger and freight transport has also been increasing rapidly, reaching shares in 2005 of 25% and 11%, respectively. Household energy use has shown a more modest rise and, as a result, its share of final energy consumption has remained constant at 22%. In contrast, energy consumption in manufacturing showed a slight decline. Historically, this sector has been the largest final energy user, but its share has fallen since 1990. In 2005, it represented 25% of the total, the same share as passenger transport. When re-grouped by main end-uses, the shares of final energy consumption in the IEA16 in 2005 are as follows: passenger and freight transport—36%; buildings—36% and manufacturing–25%. Total final energy consumption (TFC) in the IEA16 is dominated by oil, with a share of 47% in 2005—the same as in 1990. Oil use for both passenger and freight transport increased strongly between 1990 and 2005, driving the overall rise in oil demand.
130
2 The 16 IEA countries included in the analysis of the aggregate trends are Australia, Austria, Canada, Denmark, Finland, France, Germany, Italy, Japan, the Netherlands, New Zealand, Norway, Sweden, Switzerland, the United Kingdom and the United States. These countries account for 85% of total final energy use in all IEA countries.
Index (1990 = 100)
3. Understanding trends in energy use and efficiency
GDP PPP 2000 Final energy use Energy use per GDP
120 110 100 90 80 1990
1995
2000
2005
Fig. 3. Overview of key economy-wide trends, IEA16. Source: IEA indicators database.
With a share of 22%, electricity has overtaken natural gas as the second most important energy commodity in the final energy mix (Fig. 5). Electricity use increased rapidly in all stationary sectors, with the strongest growth in households and services. Residential electricity demand is largely driven by increased ownership and use of electric appliances. In particular, there has been a substantial increase in consumption from a wide range of smaller appliances such as home entertainment and kitchen equipment. In the service sector, much of the strong growth has been from air conditioning and lighting, and from various kinds of office and information technology equipment. In contrast, coal consumption has declined and now represents only 5% of total energy use by final consumers. Renewable energy use
6466
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
35
100% Manufacturing
90%
30
29%
25%
80% Passenger transport
25
70%
20 EJ
Households*
60%
24%
25%
50% 15
Services
40% 22%
22%
30%
10 Freight transport
20%
13%
14%
10%
11%
1990
2005
5 10% Other** 0 1990
0% 1995
2000
2005
Fig. 4. Total final energy consumption by sector, IEA16. *Corrected for yearly climate variations. **Other includes construction and agriculture, forestry and fishing. Source: IEA indicators database.
60
100% Oil
50
90% 80%
Electricity
40
EJ
Natural gas
30
47%
19%
22%
60% 50%
Coal
20
40% 30%
Renewables
21% 20%
10 District heat
0 1990
47%
70%
10% 0%
1995
2000
2005
8%
21%
4%
5% 4%
1990
2005
Fig. 5. Total final energy consumption by energy commodity, IEA16. Source: IEA indicators database.
(mostly biomass) has grown, but its share remains only 4% of final energy consumption.
3.2. Drivers of energy use Energy underpins nearly every aspect of a modern economy. Thus, increase in economic activity in all sectors tend to put an upward pressure on energy consumption. Fig. 6 shows the evolution of some of the key drivers of energy consumption across each sector over the period from 1990 to 2005, on a per capita basis. All drivers shown have increased between 1990 and 2005, with annual growth rates of between 1% and 2% for the IEA16 as a whole. The service sector showed particularly strong increases in value-added output per capita for all countries, which is one of the reasons why energy use in services is growing most rapidly. Growth in household area
per capita has increased much more strongly in Japan, Australia and New Zealand than in North America and Europe. Other sectors also demonstrate significant differences among countries, with freight transport and manufacturing having the widest variation in activity increases. Freight haulage in Japan showed particularly low growth rates compared to other countries, partly as a result of increased load factors. Consequently, energy use for freight transport in Japan was virtually unchanged between 1990 and 2005. In manufacturing, North America has experienced much higher growth rates per capita than either Japan or Europe. However, the impact of this growth on energy use has been largely offset by significant reductions in energy intensities. The evolution of fossil fuel prices is a further important factor affecting energy consumption (Fig. 7). These prices are strongly influenced by trends in spot oil prices. Most of the 1990s was characterised by decreasing real prices of crude oil, but since 2003
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
Australia
Average annual percent change
3.0%
Canada
New Zealand
Japan
6467
US
EUR13*
2.5%
2.0%
1.5%
1.0%
0.5%
ic e pe s v r c alu ap eita ad d
se
rv
ou
Se
H
M an ad ufa de ct d uri p e ng rc v ap alu i t a e-
ed
r pe
rc pe m -k To n
ne
ss Pa
ho ca ld pi a r e ta a
ap
en ca ger pi -km ta p
er
ita
0.0%
Fig. 6. Changes in key sub-sectoral activity levels relative to population, 1990–2005. *EUR13 includes Austria, Denmark, Finland, France, Germany, Greece, Italy, the Netherlands, Norway, Spain, Sweden, Switzerland and the United Kingdom. Source: IEA indicators database.
Canada US
450
350
Crude oil
8
Steam coal
250 200 150 100 50 0 1990
Italy UK
9
Liquified natural gas 300
Japan France
Gaseous natural gas MJ per USD PPP 2000
USD MER 2000 per toe
400
IEA16 Germany
7 6 5 4
1995
2000
2005 2007
Fig. 7. Fossil fuel import prices in real terms. Source: IEA (2008d).
3 1990
1995
2000
2005
Fig. 8. Final energy use per unit of GDP. Source: IEA indicators database.
prices have risen very steeply. Real prices for natural gas and, to some extent, coal have followed oil price developments, although with less significant fluctuations. Real energy prices decreased between 1990 and 1998; in many cases, they did not return to their 1990 levels until around 2005. Such price developments are unlikely to have provided much incentive for improvements in energy efficiency during the 1990s. Since 2003 prices have risen more steeply and these increases could be expected to have a significant effect on end-use energy efficiency and end-use fuel mix, as a result of behavioural changes across all sectors. However, detailed energy consumption data for years after 2005 were not available to confirm this at the time of writing.
3.3. Changes in aggregate energy intensity The ratio of TFC per unit of GDP provides a measure of final energy intensity for a country, and is one of the most frequently
used aggregate energy indicators. Trends in final energy intensities show that the ratio of TFC to GDP has declined in most IEA countries since 1990 (Fig. 8). However, both the absolute levels of the ratio and the rates at which it has fallen vary significantly. Among the largest IEA economies, the United Kingdom showed the strongest decline, with the ratio of TFC to GDP falling by 24% between 1990 and 2005, mostly due to changes in the structure of the economy. At the other end of the scale, Italy actually showed an increase of 1% over the same period. In a number of IEA countries, the TFC/GDP ratio fell most rapidly during the mid- to late 1990s, a period of rapid economic growth. Trends in final energy intensity are often used to assess the extent to which energy efficiency is improving in countries. However, this can often be misleading as the ratio is affected by many factors such as climate, geography, travel distance, home
6468
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
45
GJ per capita (normalised to 2700 HDD)
40 35 30 25 20 15 10 5
Au
st
ra
19 Au lia 90 2 st ria 005 19 9 C an 20 0 ad 0 a 19 5 D 90 en 20 m 0 ar k 19 5 90 Fi 2 nl 00 an d 19 5 9 Fr 20 0 an ce 1 05 99 G er 20 0 m an 0 y 19 5 90 Ire 20 la 0 nd 19 5 9 20 0 Ita ly 1 05 99 Ja 20 0 pa 0 n 19 5 9 Ko 20 0 re N 05 et a 1 he 99 rla 0 nd 200 N ew s 1 5 99 Ze 20 0 al an 0 d 19 5 9 N or 20 0 w ay 1 05 99 Sp 20 0 ai n 1 05 99 Sw 20 0 e Sw den 1 05 99 itz er 20 0 la nd 1 05 99 20 0 U 1 05 K 9 9 20 0 U 1 05 S 9 9 20 0 05
0
Appliances
Lighting
Cooking
Water heating
Space heating
Total (not normalised)
Fig. 9. Household energy use per capita. Source: IEA indicators database.
size and manufacturing structure, which energy efficiency measures cannot directly influence. As explained earlier, more detailed indicators targeted at the individual end-use sectors are needed to get a better picture of what is happening in relation to energy efficiency. Some examples of these types of indicators are shown in the following sections.
3.4. Buildings
For the IEA19 as a whole, the useful energy per area heated declined by 17% between 1990 and 2005, largely as a result of improvements to the building shell that have reduced heat losses. This is much greater than the fall in space heating per capita, as the upward pressure on energy demand from larger homes and lower occupancy levels has largely offset the benefits of a decline in space heating intensity. Looking at the variation amongst countries, it can be seen that a number of countries with relatively cold climates (e.g. Finland, Canada and Norway) have intensities that are below the average for the IEA19. This can partly be explained by higher levels of insulation in these countries. In contrast, France, Italy and Germany, which have more moderate winters, showed the highest levels of energy intensity. This could be the result of a greater percentage of poorly insulated older buildings.
Buildings account for just over one-third of total final energy use in IEA countries. However, the buildings share of total electricity consumption is much higher at 65%. Space heating is by far the most important energy user in the residential sector of a group of 19 IEA countries (IEA19),3 accounting for 53% of final household energy use in 2005. The significance of space heating varies according to climate. Fig. 9 shows per capita residential energy use with space heating both adjusted and not adjusted to account for climatic differences. The amount of energy used for space heating differs widely among countries, due to differences in house size, expected indoor heating comfort, type of heating equipment and levels of insulation. Overall, for the IEA19, space heating per capita fell by 2.4% between 1990 and 2005. To understand better these variations and trends, it is useful to develop indicators for the components that affect space heating demand. A key indicator in this regard is the level of space heating intensity. This is measured in terms of ‘‘useful energy’’ for space heating per square metre, with useful energy being calculated as final energy minus the losses estimated for boilers. To allow for comparisons across countries with different climates, the space heating intensity is divided by each country’s yearly number of heating degree-days (Fig. 10).
The energy intensity of manufacturing industry (as measured by dividing total manufacturing energy use by total manufacturing value-added) varies widely between sub-sectors. The most energyintensive sub-sector is the production of primary metals. In 2005, energy intensity in this sub-sector was more than 10 times higher than the category of ‘‘other manufacturing’’ (Fig. 11). Yet output from the primary metals sub-sector constituted only 4% of total manufacturing value-added, compared to more than 29% from other manufacturing. Taken together, the four most energy-intensive subsectors (primary metals, non-metallic minerals, chemicals and paper and pulp) accounted for almost 70% of manufacturing energy use in a group of 21 IEA countries (IEA21),4 despite representing only 29% of total value-added output.
3 The 19 IEA countries included in the analysis of buildings are Australia, Austria, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, the Republic of Korea, the Netherlands, New Zealand, Norway, Spain, Sweden, Switzerland, the United Kingdom and the United States.
4 The 21 IEA countries included in the analysis of buildings are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Italy, Japan, the Republic of Korea, the Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, the United Kingdom and the United States.
3.5. Manufacturing
an a Fi da nl a N No nd ew r w Ze a al y Au and st Po ralia rtu Be ga lg l iu m N et he US rla Sw nd ed s Ko en re IE a A2 Sp 1 ai n U Au K st ria Ita Fr ly an c Ja e p D en an G ma er rk m a G n Sw re y itz ece er la nd
C
MJ per USD PPP 2000
& , be to v ba er cc ag o es Pa & pe pr r, p in u tin lp C g he m ic al s N on m m in et er al al lic Pr s im ar y m et M al e s & tal eq pr ui od pm u en cts t m an O uf the ac r tu rin g m an T o uf ta ac l tu rin g
od
Fo
MJ per USD PPP 2000 an y Ita l Fr y an Ire ce la n Ko d re a U K Au Au stria s D trali en a m Sw ar ed k IE en A Fi 19 nl a C n N an d et ad h Sw erla a itz nd er s la nd U Sp S N ain o N ew rw Ze ay al a Ja nd pa n
er m
G
kJ per m2 per HDD
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
Energy intensities
50 100%
40 80%
30 60%
20 40%
10 20%
0 0%
Value-added shares
2005 actuaI intensities
6469
250
200 1990 2005
150
100
50
0
Fig. 10. Useful space heating intensity. Source: IEA indicators database.
Energy shares
Fig. 11. Sub-sector energy intensities, value-added and energy shares, IEA21. Source: IEA indicators database.
18
16 2005 common structure intensities
14
12
10
8
6
4
2
0
Fig. 12. Manufacturing energy intensity at actual and common structures. Source: IEA indicators database.
6470
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
Car fuel intensity (litres of gasoline equivalent per 100 vkm)
12 US 11 Australia
Japan Canada New Zealand
10
9 Switzerland 8
Sweden
Ireland
Norway Germany Austria
7
France
6 0.5
UK
Finland
Denmark
Netherlands
Italy
Greece 0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
Retail average fuel price (USD PPP 2000 per litre) Fig. 13. Car fuel intensity versus average fuel price, 2005. Source: IEA indicators database.
3.6. Transport In IEA countries, the transport sector is taking an increasing share of oil demand and is currently responsible for almost 78% of total final oil consumption. Passenger cars and light trucks account for over half of IEA oil demand for transport, and have been the main driving force behind the growth in oil demand over the last three decades. Reducing the fuel intensity of cars (i.e. increasing the fuel efficiency) is therefore a key way of helping
US
Canada
Australia & NZ
IEA18
Japan
EUR13*
13 Litres of gasoline equivalent per 100 vehicle-km
Large differences are also evident in the level of overall manufacturing energy intensity among IEA countries. Using an indicator approach it is possible to investigate to what extent these differences among countries can be explained by differences in their industrial structure. To examine this question, the effects of structural differences are removed by calculating what the energy intensity of each country would be if they all had a common structure (e.g. the average of the IEA21 countries, based on a breakdown of manufacturing into seven sub-sectors). Fig. 12 shows the results of such a calculation for 2005, and compares these intensities at common structure with actual energy intensities in that year. The results are quite striking; even though they do not account for the detailed structural differences within the sub-sectors of industry (i.e. different product mixes within sub-sectors). In 2005, actual manufacturing energy intensities varied by a factor of more than five amongst countries (3.5–17.8 MJ per USD). However, at common structure, the variation between countries with the highest and lowest intensities falls to a factor of less than three (3.8–10.8 MJ per USD). Thus structural effects are responsible for almost half the variation in manufacturing energy intensities that are observed amongst countries. The results are particularly significant for some countries. For instance, this approach shows that Australia’s very high energy intensity can be largely explained by the structure of its manufacturing industry, which has a high share of very energyintensive industries. If Australia’s industry had the same structure as the average for the IEA21 countries – but kept its actual level of energy intensity in each sub-sector – the country’s aggregate manufacturing energy intensity would be reduced by 45%. Similar, but less dramatic, results are also observed for Belgium, Canada, Greece, the Netherlands, Norway and Spain.
12 11 10 9 8 7 1990
1995
2000
2005
Fig. 14. Average fuel intensity of the cars stock. Source: IEA indicators database.
both to reduce dependence on oil and to limit CO2 emissions from the transport sector. Indicators can help examine some of the factors that impact the fuel efficiency of cars and track the progress of policies to help improve efficiency. Perhaps not surprisingly, fuel prices are an important factor influencing fuel intensity. Fuel prices vary considerably across IEA countries, with gasoline prices in 2005 ranging from USD 0.51 to USD 1.44 per litre (using 2000 PPP exchange rates). Variations in fuel taxes are the main reason for this wide gap. Three groups of countries can be identified: the United States, with the lowest prices (and the lowest tax levels); European countries, with generally high prices (and high taxes); and, in the middle, other IEA countries such as Australia, Canada, Japan and New Zealand. Plotting the fuel intensity of cars against fuel prices shows a clear relationship; the average fuel intensity of cars tends to be higher in countries with low fuel prices (Fig. 13). So the United States has the lowest fuel prices and cars with the highest fuel intensity. In contrast, fuel intensity is lowest in Europe, where fuel prices are much higher.
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
3.7. Factors impacting energy use: the role of energy efficiency The previous sections highlighted how different types of indicators can be used to examine trends in energy use and efficiency, both at an aggregate level and for individual end-use sectors. This section shows how the decomposition approach presented earlier can be used to explain how different factors impact energy use and, in particular, to quantify the role of energy efficiency improvements. Take as an example the large variations among countries in how much energy per unit of GDP has fallen over time. To understand the extent to which this variation reflects differences in energy efficiency developments, it is necessary to separate the impact of changes in sub-sectoral energy intensities (which are used as a proxy for energy efficiency) from the effects of changes in economic structure and other factors that influence the demand for energy. Increased demand for energy services – reflecting increased ownership levels of electric household appliances, bigger houses, more personal travel by car, etc. – drives both energy use and energy use per GDP. Therefore, it is useful to examine how the ratio of energy to GDP is affected by changes in structure (as measured by the ratio of energy services to GDP) and in end-use intensities (such as the energy used to heat a square metre of floor space or car energy use per passenger-kilometre). 5 The 18 IEA countries included in the analysis of transport are Australia, Austria, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, the Netherlands, New Zealand, Norway, Sweden, Switzerland, the United Kingdom and the United States.
100 Energy services per GDP
Index (1990 = 100)
However, fuel prices are not the only factor impacting fuel intensities. A number of other factors such as vehicle technologies, the effect of driving conditions and fuel efficiency policies also play an important role in both the absolute levels and trends observed. The average fuel intensities of cars decreased in most countries between 1990 and 2005 (Fig. 14). In Europe, this was due to a combination of factors. The 1990s were characterised by the widespread diffusion of vehicles equipped with electronic control systems for fuel management and by stronger consumer demand for more efficient cars—a reaction to high fuel prices. Since the early 2000s, intensities declined further in Europe as a result of increased sales of direct-injection diesel cars. Despite showing a decrease, the fuel intensity of cars in the United States remained higher than the average of a group of 18 IEA countries (IEA18),5 at nearly 11.5 l of gasoline equivalent per 100 vehiclekilometres. High levels of fuel intensity also characterised Australia, Canada, New Zealand and Japan. In Japan, until the late 1990s, efficiency improvements of new vehicles had been offset by increase in vehicle weight and congestion-related effects. The increasing weight of vehicles has been another factor offsetting improvements in the underlying efficiency of new car engine technologies. Over the last 15 years, the average size and weight of the stock of cars increased as larger and heavier vehicles, such as SUVs, became more popular. This trend, combined with additional safety features also increasing weight, has tended to raise the energy consumption of cars. Data for Canada show that, in 2005, light trucks accounted for one-third of all cars, up from a share of one-fifth in 1990. In the United States, the stock of passenger cars remained almost stable between 1990 and 2005, while that of light trucks nearly doubled. In European countries, the number of cars with an engine capacity greater than 2 l has more than doubled since 1990. Japan reported a much lower increase in the weight of cars, mainly because of effective regulation of fuel efficiency under the Top Runner programme, which came into effect in the late 1990s.
6471
Decline in energy per GDP from fall in energy services per GDP
95
90
Actual energy per GDP
Decline in energy per GDP from fall in subsectoral energy intensities
85
80 1990
1995
2000
2005
Fig. 15. TFC per GDP and the contribution of changes in energy services per GDP and energy intensities, IEA16. Source: IEA indicators database.
Changes in energy consumption per GDP can be attributed to changes in the ratio of energy services to GDP and to changes in sub-sector energy intensities for more than 20 end-uses. The intensity effect for the whole economy is calculated as the aggregate impact of the sectoral intensity effects. The results of aggregate impact calculations show that the energy intensity effect and changes in structure have both contributed to reduced energy consumption per unit of GDP (Fig. 15). However, declining end-use intensities (the energy intensity effect) have been the most important factor. In the early 1990s, structural effects made almost no contribution to the falling energy to GDP ratio; the economy of the IEA was growing slowly, as was demand for energy services. The contribution of structural changes has grown since then but, by 2005, more than 60% of the total decline in energy per GDP could be attributed to reductions from the energy intensity effect. This equates to an annual average fall in end-use intensities across all sectors of 0.9% per year. The relative contribution of structural and intensity effects to the overall trend varies among countries (Fig. 16). With the exception of Italy, all countries show that the intensity effect contributed to reducing the ratio of energy use to GDP: for most countries, it was the dominant factor. This is particularly true in the case of Canada, the Netherlands, Germany, New Zealand, Sweden and the United States. In contrast, for Norway and the United Kingdom, changes in structure were most important. For those countries having declining sub-sectoral intensities, it is possible to identify three groups according to the magnitude of the intensity effect:
less than 0.5% per year: Austria, Denmark, Japan, Norway; 0.5–1%: Australia, France, Finland, Switzerland and the United Kingdom and
greater than 1%: Canada, Germany, the Netherlands, New Zealand, Sweden and the United States. It can be informative to compare these results with the information on the absolute levels of final energy intensity (Fig. 8). Such a comparison shows that those countries with a high level of energy use per GDP in 1990 tend to have had the largest reductions in intensity. In contrast, those countries that initially had lower energy use per GDP have generally seen smaller declines in intensity. Although the intensity effect for the whole economy declined by an average of 0.9% per year, the rate of reduction was not uniform across end-use sectors. The manufacturing sector showed the largest reductions in intensities (corrected for structure), decreasing by 1.4% per year between 1990 and 2005 (Fig. 17). Building energy use showed lower reductions, with average
6472
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
0.5% Average annual percent change
Intensity effect
Energy services per GDP
Energy per GDP
0.0% -0.5% -1.0% -1.5% -2.0%
6 A1
U K U S
IE
Au
st
ra l Au ia st r C ia an ad D en a m a Fi rk nl an Fr d an c G er e m an y Ita ly Ja pa N et n he rla N ew n Ze ds al a N nd or w a Sw y e Sw de n itz er la nd
-2.5%
Fig. 16. Changes in TFC/GDP decomposed into changes in energy services/GDP and intensity effect, 1990–2005. Source: IEA indicators database.
105
Passenger transport
Total
18
Freight transport
Services
16
Households*
Manufacturing
14
Manufacturing Households* Services Passenger transport Freight transport
EJ
12
Index (1990 = 100)
100
10 8 6 4
95
2 0 1995
90
2000
2005
Fig. 18. Contribution to energy savings from the end-use sectors, IEA16. *Corrected for yearly climate variations. Source: IEA indicators database.
85
80 1990
1995
2000
2005
Fig. 17. Sector intensities and total economy effect, IEA16. *Corrected for yearly climate variations. Source: IEA indicators database.
intensity falls of 0.8% per year in households and 1.2% in the service sector. The passenger and freight transport sectors experienced similar declining trends between 1990 and 2005, but had the lowest falls in intensities, both averaging 0.5% per year (calculated holding the modal mix constant).
3.8. Energy savings from improved energy efficiency The decline in energy intensities (i.e. improvements in energy efficiency) in the various end-use sectors led to energy savings across the whole economy. Energy savings in 2005 were about 16 EJ (370 Mtoe) or 15% of total final energy use in that year, which represents an estimated USD 180 billion of energy cost savings (Fig. 18). The impact of energy intensity improvements on CO2 emissions was also significant, with a saving of 1.3 Gt CO2 (14% of actual CO2 emissions in 2005). These savings are approximately equal to the annual final energy consumption and CO2 emissions of Japan.
Fig. 18 shows the total energy savings since 1990, broken down by sector. In 2005, the largest share was from manufacturing (42%), followed by households and services (both 19%), passenger transport (14%) and freight transport (6%). Initially, there were strong energy savings in households as a result of significant improvements in space heating intensity, which did not continue. In contrast, the savings from services made an impact only in the late 1990s, during a period of high economic growth in this sector. The rates of savings in the other sectors were more constant over time. 3.9. Long-term trends Fig. 19 shows that over the longer term, the savings from improved energy efficiency are even more significant. Without the energy efficiency improvements that occurred between 1973 and 2005, energy use in a group of 11 IEA countries for which longterm times series data are available (IEA11)6 would have been 58%, or 59 EJ, higher in 2005 than it actually was. This makes energy savings the most important ‘‘fuel’’ in the IEA11 for this time period—i.e. the amount of energy saved in 2005 was slightly higher than the actual consumption of oil, or of electricity and natural gas combined. 6 The 11 IEA countries included in the analysis of long-term trends are Australia, Denmark, Finland, France, Germany, Italy, Japan, Norway, Sweden, the United Kingdom and the United States.
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
180
6473
2%
140
Hypothetical energy use without energy efficiency improvements Savings
58%
120
EJ
100 80 Actual energy use
60 40
Average annual percent change
2.0%
160
1% 0.9% 0.8%
0.5%
20 0 1973
0% 1980
Actual energy use
1990
2000
2005
Energy savings due to energy efficiency improvements
1973-1990 1990-2005 Energy efficiency improvements
Fig. 19. Long-term energy savings from improvements in energy efficiency, all sectors, IEA11. Source: IEA indicators database.
A similar pattern emerges from analysis of the individual enduse sectors. Energy efficiency improvements have led to significant savings in all sectors. However, without exception, the rate of improvement since 1990 has been lower than in the period between 1973 and 1990. In manufacturing, the strong energyefficiency improvements between 1973 and 1990 led to a decrease in energy use, but this reduction did not continue into the 1990s. For households and services, lower energy-efficiency improvements since 1990 have led to significantly higher increases in energy use between 1990 and 2005. For both passenger and freight, the reduction in the rate of improvements in energy efficiency since 1990 has been partially compensated by lower increases in energy service demand. For passenger transport, the rate of increase in energy use changed little between the two periods. The rate of increase in energy use for freight transport actually fell after 1990.
4. Conclusions and policy implications This article has highlighted key aspects of the IEA work on energy indicators undertaken in support of the Gleneagles Plan of Action. It has shown the importance of using disaggregated energy indicators for analysing interactions between economic and human activity and energy use. Such indicators provide a powerful set of analytical tools that reveal important trends in energy use and efficiency that are obscured when only more aggregate quantities, such as the ratio of energy use to GDP, are used. Many IEA member countries already use energy indicators and they are also attracting increasing interest from other countries. The IEA is working to assist and internationalise these efforts by developing transparent and consistent international databases and methodologies and by collaborating with governments, industry and regional and international organisations. The results from the indicators analysis highlight that energy efficiency continues to play a key role in limiting increases in energy use in IEA countries. However, they also reveal that the rate at which energy efficiency has improved since 1990 has been much slower than in the previous two decades. This shows that the changes caused by the oil price shocks in the 1970s and the resulting energy policies did considerably more to control
growth in energy demand and reduce CO2 emissions than the energy efficiency and climate policies implemented since the 1990s. More detailed analysis reveals many other developments of importance to policy-makers. Disaggregated indicators have been used to track trends in energy efficiency at the end-use level, as well as quantifying the impact that non-energy factors have over energy consumption. For example, the analysis shows that while household space heating intensities have declined in many countries, the benefits of this improved efficiency have been largely offset by the trend towards larger homes and lower occupancy levels. In the manufacturing sector, differences in the structure of industry can explain almost half the variation observed in the manufacturing energy intensities amongst IEA countries. While in transport, a comparison amongst countries highlights a clear correlation between higher fuel prices and a more efficient stock of cars. Projections published by the IEA in Energy Technology Perspectives 2008 (IEA 2008b) clearly demonstrate that rate of improvement in energy efficiency will need to be substantially increased to have a realistic chance of a more sustainable energy future. The good news is that this is indeed possible; there is still significant scope in IEA countries for energy-efficiency improvements in buildings, industry and transport. However, a key lesson of the past is that the widespread future deployment of more efficient technologies and practices will require very strong government action. The challenge is to find the right mix of market- and regulation-based policies to achieve cost-effective decisions regarding energy efficiency in all sectors. Based on its indicators analysis and other work on best practices in policies and measures, the IEA has presented a list of high-priority energy efficiency policy recommendations to help governments increase rates of energy efficiency improvement in buildings, appliances, lighting, transport, industry, power utilities and cross-sectoral areas (IEA, 2008e). To make these suggested policies a reality, governments will need to provide adequate resources (both financial and human) to the policy agencies responsible for energy efficiency. This should include funding and managing the implementation and monitoring of action plans for energy efficiency. In turn, this will require the establishment and ongoing maintenance of high quality energy indicators and other data; recognising that, while there have been some improve-
6474
P.G. Taylor et al. / Energy Policy 38 (2010) 6463–6474
ments, the availability, timeliness, quality and comparability of data in many countries and sectors are still not sufficient. Building on the success of the Gleneagles Plan of Action, the IEA is committed to support governments and other interested parties in this important work. In the statement issued following their Hokkaido meeting in July 2008, the G8 leaders requested that the IEA ‘‘enhance its work on voluntary sectoral indicators through improved data collection, complemented by business initiatives’’. The IEA recognises that such enhancements will not be easy, but equally believes that failure to provide a robust framework of data collection and indicator development will ultimately undermine the ability of analysts and policy-makers to formulate, implement and monitor successful energy efficiency and other policies.
Acknowledgements The authors would like to thank the governments, organisations, companies and industry associations who have helped with collecting and validating the underlying data used in the IEA
indicator analysis. In particular, we are grateful for the close collaboration of the statisticians and analysts in IEA member countries, including experts from the European Union sponsored ODYSSEE network. References IEA (International Energy Agency), 1997a. The Link between Energy and Human Activity. OECD/IEA, Paris. IEA, 1997b. Indicators of Energy Use and Efficiency—Understanding the Link between Energy and Human Activity. OECD/IEA, Paris. IEA, 2004. Oil Crises and Climate Challenges: 30 Years of Energy Use in IEA Countries. OECD/IEA, Paris. IEA, 2007a. Tracking Industrial Energy Efficiency and CO2 Emissions. OECD/IEA, Paris. IEA, 2007b. Energy Use in the New Millennium: Trends in IEA Countries. OECD/IEA, Paris. IEA, 2008a. Worldwide Trends in Energy Use and Efficiency: Key Insights from IEA Indicators Analysis. OECD/IEA, Paris. IEA, 2008b. Energy Technology Perspectives 2008: Scenarios and Strategies to 2050. OECD/IEA, Paris. IEA, 2008c. World Energy Outlook 2008. OECD/IEA, Paris. IEA, 2008d. Energy Prices and Taxes, 2nd Quarter 2008. OECD/IEA, Paris. IEA, 2008e. Energy Efficiency Policy Recommendations. OECD/IEA, Paris.