Increased ecoefficiency and gross rebound effect: Evidence from USA and six European countries 1960–2002

Abstract

Despite increased efficiency in the use of natural resources, the use of these resources continues to increase in most societies. This paper examines the discrepancy between the potential decrease of use of natural resources, as an effect of increased efficiency, and actual use. During the period 1960–2002, this difference was found to grow faster in the USA than the mean for six West European countries. Possible reasons for this difference between the two regions are analysed. To reduce the anthropogenic flows of energy and material, and the consequent deleterious effects on the biosphere, it will become necessary to adapt consumption to degree of efficiency in the use of natural resources. Based on the comparison between the two regions, some economic aspects of this issue are discussed.

Full text

ANALYSIS
Increased ecoefficiency and gross rebound effect: Evidence
from USA and six European countries 1960–2002
Stig-Olof Holm⁎, Göran Englund
Department of Ecology and Environmental Science, Umeå University, S-90187 Umeå, Sweden
ARTICLE INFO ABSTRACT
Article history:
Received 24 January 2008
Received in revised form 9 June 2008
Accepted 10 July 2008
Available online 15 August 2008
Despite increased efficiency in the use of natural resources, the use of these resources
continues to increase in most societies. This paper examines the discrepancy between the
potential decrease of use of natural resources, as an effect of increased efficiency, and actual
use. During the period 1960–2002, this difference was found to grow faster in the USA than the
mean for six West European countries. Possible reasons for this difference between the two
regions are analysed. To reduce the anthropogenic flows of energy and material, and the
consequent deleterious effects on the biosphere, it will become necessary to adapt
consumption to degree of efficiency in the use of natural resources. Based on the comparison
between the two regions, some economic aspects of this issue are discussed.
© 2008 Elsevier B.V. All rights reserved.
Keywords:
Ecological footprint
Economic growth
Global change
Environmental Kuznets Curve
Gross rebound effect
Biodiversity
1. Introduction
Globally, human use of natural resources has steadily in-
creased. Over the last 200 years, most use patterns can be re-
presented by J-shaped curves (Cohen, 1995; Noble and Dirzo,
1997; McNeill, 2000; Gleick, 2000; Tilman et al., 2001; Hails
et al., 2006). By the late 20th century, people had affected about
50% of the entire global land area (Hannah et al., 1994) and
were using, directly and indirectly, about 38% of the yearly
biomass production in terrestrial ecosystems (Vitousek et al.,
1986; Vitousek, 1994; Vitousek et al., 1997). The increasing
mobilization of natural resources by human society is a threat
to global biodiversity and to future supplies (Laurance, 2001;
IUCN, 2002; UNEP, 2002).
At the beginning of the industrial expansion it was debated
whether or not it was possible to stabilise the use of natural
resources through increased efficiency. For example William
Jevons stated: ”an increase in efficiency in using a resource
leads, in the medium to long term, to an increased usage of
that resource rather than to a reduction in this use”(Jevons,
1865). This contention has been called the “Jevons paradox”.
After the Second World War, the debate focused more on
increased economic growth. The “technology factor”was
considered pivotal for managing production increases, and
limits to expansion were taken off the economics agenda
(Friman, 2002). In the late 20th century, as a consequence of
the environmental debate, ways to decouple material mobi-
lization from economic growth were hypothesized. One of
these hypotheses, called the environmental Kuznets curve
(EKC), predicts a hump shaped relationship between degree of
material flow, following environmental impact, and GDP/ca-
pita. (Stern et al., 1996; Cavlovic et al., 2000; Gawande et al.,
2001). In some cases, this hypothesis accurately describes
reality. For example, the consumption of phosphate rock and
crude steel decreased in many European countries during the
period 1970–1990 as GDP/capita increased (Kågeson, 1997). In
ECOLOGICAL ECONOMICS 68 (2009) 879–887
⁎Corresponding author. Tel.: +46 90 786 55 46; fax: +46 90 786 76 64.
E-mail address: Stig-Olof.Holm@eg.umu.se (S.-O. Holm).
0921-8009/$ –see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.ecolecon.2008.07.006
available at www.sciencedirect.com
www.elsevier.com/locate/ecolecon

other cases, especially at the global level, the material flow
continuesto increase with increasing GDP/capita, eg., emissions
of carbon dioxide and use of energy, forest products, fish, and
cement (Loh et al., 1998; Hails et al., 2006; IEA, 2003).
Typically, the material flow per unit of product or service
created tends to decrease with increasing GDP/capita (Kåge-
son, 1997). However, this “decoupling”effect is often not
strong enough to curb the total anthropogenic flow of
materials, because total consumption continues to increase
as a result of the consumption of more units. The two
phenomena have been named respectively, “relative”and
“absolute”decoupling (Spangenberg, 1995). In a situation
where only a “relative”and not an “absolute”decoupling
occur, resource depletion and consequent destruction of
ecosystems continues.
The gap between the decreased use of resources that is
expected from increased “eco-efficiency”and the actual uti-
lisation has been called the “gross rebound effect”(GRB)
(Vehmas et al., 2004). One example of a GRB is that although
efficiency in the usage of energy per unit of product or service
in the OECD-countries improved by about 30% between 1970
and 1991 (Schipper et al., 1997), there was no corresponding
decrease in energy use. In fact, the use of energy in these
countries increased by about 20% (IEA, 1996). Another
example is that when global energy efficiency increased by
2% per year between 1973 and 1990, consumption during the
same period increased by 2.7% per year, resulting in a yearly
net increase of energy flow of 0.7% (Sun, 1998). A third ex-
ample is that since the mid 19th century, the amount of
carbon used per unit of production has decreased by 1.3% per
year although carbon dioxide emissions have risen about
1.7% per year (Kates, 1994).
Note that the definition of the term “rebound effect”is
sometimes limited to the increase of consumption of a pro-
duct that occurs when increased production efficiency results
in a decreased price. This effect is often relatively small,
typically amounting to a few percent of the effect of increased
efficiency (Greening et al., 2000; Berkhout et al., 2000). The
GRB is more close ly related to the Khazzom–Brooks postulate,
which suggests that energy efficiency improvements at the
micro level lead to increased economic growth and thus
higher energy use at the macro level (Herring, 1998). However,
the “gross rebound effect”is a broader concept, applicable not
only to energy use, but also to other issues, such as the net
effect of societies overall resource use/environmental im-
pact due to consumption patterns, population development
and degree of eco-efficiency (Jokinen et al., 1998; Malaska
et al. 1999). These concepts can be formalized using the
Holdren/Erlich equation I=PAT,whereIrepresents environ-
mental impact, Pis population size, Ais affluence per capita
and Tis the effect of technology (Ehrlich and Holdren, 1971).
This model illustrates that resource use (I)willcontinueto
grow if the effects of increased efficiency (reduced T)isout-
weighed by increased population size (P) and/or affluence per
capita (A).
Some limitations of the IPAT approach have been pointed
out. The equation does for example not include the fact that
the life style of individuals and groups may be more or less
harmful to the environment. A factor describing behaviour (B)
have therefore been added, creating I=PBAT (Schulze, 2002).
Non-proportional effects and elasticity coefficients have also
been included into the model (for a review of different variants
of the IPAT equation, see e.g. Fan et al., 2006). However, even if
numerous sub-factors influence the P,Aand T, and various
degrees of elasticity between a certain kind of environmental
impact and the three explaining variables may exist, the IPAT
equation can be considered as a basis for a description of
human impact on the ecosystems.
The main objective of this study is to identify possible
approaches to overcome the GRB. Three different analyses are
presented. First we analyze the relationship between GDP/
capita and the global environment footprint per capita for 135
countries as a test of the EKC hypothesis at a global scale.
For the second analysis we selected two regions with si-
milar material wealth, but with differences in impact/capita
on the global environment. The two regions are (1) the USA
and (2) a group of six West European countries: the United
Kingdom, France, Germany, Italy, Austria and Switzerland. For
these regions the IPAT equation is used to describe the con-
tribution of population size, affluence and technology (P,A,
and T) to the observed difference in environmental impact
during the period 1960–2002. Total energy use is used as a
proxy for the total environmental impact (I), affluence (A)is
defined as GDP/capita, and technology (T) is defined as the
“ecoefficiency”(energy use/GDP). This analysis is also used to
examine the development of the gross rebound effect in the
two regions. Our hypothesis is that a larger gross rebound
effect will be found in the society having the largest environ-
mental impact.
The IPAT analysis indicates that the energy use per capita is
a key to understanding differences between the two regions.
Thus we also perform a more detailed analysis of the
relationship between energy use/capita and 10 descriptors of
society's structure that are expected to correlate with energy
use per capita. The objective is to identify factors that a policy
aimed at reducing environmental impacts should target.
Although the IPAT approach may be useful for devising
strategies for a more sustainable resource use, it is of little
Fig. 1 –The relationship between global ecological footprint
per person and GDP per person for 135 countries 2001. GDP is
given in current international dollars re-calculated as
purchasing power parity (PPP). Sources: Loh and
Wackernagel (2004) and WRI (2006).
880 ECOLOGICAL ECONOMICS 68 (2009) 879–887

guidance for determining sustainable levels of resource use.
Therefore we end the paper with a discussion of a “two step”
solution that involves the identification of sustainable levels
of anthropogenic energy and material flows, and the imple-
mentation of measures to reach those levels.
2. The connection between economic activity
and material mobilization at a global scale
The ecological footprint is an accounting framework which
tracks energy and resource throughput by people and
translates it into the area of biologically productive land
necessary to produce such resource flows, and to absorb
resulting pollution (Wackernagel et al., 2000). Fig. 1 shows
the relationship between the ecological footprint per
capita and GDP/capita in the year 2001 for 135 countries.
There are many countries with a small ecological foot-
print/capita and low GDP/capita and a few with a large
footprint per capita and a large GDP/capita. To test
whether the relationship is non-linear, we fitted the data
to a power model of the form y=a⁎x
b
where yis global
ecological footprint/capita and xis GDP/capita. The linear
form, ln y=a+b⁎ln x, was used to avoid problems with
heteroscedasticity. Both coefficients were significant and b
was significantly smaller than one (estimate+ CI 95%:
a=0.0137± 0.0059, b=0.58± 0.050), which suggests that the
ecological footprint/capita increases at a decelerating rate
with increasing GDP/capita. Including a quadratic term
Fig. 2 –Energy use (solid lines) and an index based on energy
used per GDP (dashed lines) 1960–2002. The index illustrates
the energy consumption expected if only affected by
decreasing energy use per GDP. Data for Europe are mean
values for six countries; France, Germany, Italy, the United
Kingdom, Austria and Switzerland. Source: IEA (2006).
Fig. 3 –Total energy use and indices illustrating the effects of changes in population size and energy use per capita on total
energy use 1960–2002 for A) the USA and B) six European countries. C) and D) show corresponding indices for GDP per capita,
energy use per capita, and the energy used per GDP. Source: IEA (2006).
881ECOLOGICAL ECONOMICS 68 (2009) 879–887

produced a significant positive coefficient (t=4.2, pb0.001)
rather than the negative one expected on the basis of the
EKC hypothesis. Thus we conclude that the ecological
footprint does not decrease at high GDP/capita. Next we
show that similar patterns are found if instead time series
of economic growth and anthropogenic material flow are
used.
3. Analyzing patterns of resource use in
Western Europe and the USA
USA and Western Europe have similar population sizes, si-
milar climate and other out of nature given premises, and the
cultures are, compared with many other regions, very similar.
Both regions have high GDP/capita but the global material
flow/capita is much higher in the USA. To obtain a more
detailed understanding of this difference, we used the IPAT
equation to analyze how changes in population size (P), af-
fluence per capita (A) and degree of efficiency (T) have shaped
the environmental impact (I) in the two regions during the
period 1960–2002. As representative data for Western Europe
we used mean values for France, the United Kingdom, Italy,
Germany, Austria and Switzerland.
The environmental impact (I) could not be measured as an
ecological footprint because detailed temporal data are
lacking. Instead we used energy use, which is strongly cor-
related with the ecological footprint (Hails et al., 2006). Pin the
IPAT equation was represented by population size, Aby the
GDP/capita, and Tby the energy use/GDP (data given in Ap-
pendix A). To obtain a decomposable index we standardized
P, A and T by their values in 1960. Each variable were then
multiplied by the energy use in 1960, in order to illustrate
the role of initial differences in energy use. The resulting
indices, which have the unit exajoule, describes the effect of
each component (P,A, and T) on energy use. For example the
dotted lines in Fig. 2 show how energy use would have
changed due to changes in ecoefficiency (T) if P and A had
remained constant.
We first compare the actual energy use (I) with the scaled
index of ecoefficiency (T) (Fig. 2). Energy use in the USA
was higher in 1960 and increased faster during the period
1960–2002 than in the six European countries. The ecoeffi-
ciency index (dotted line in Fig. 2) shows that the increase in
energy efficiency at the society level was greater in the USA
than in Europe. We can also see that the increased efficiency
was not sufficiently great in either region to off-set the ef-
fect of changes in population size and affluence.
Next we break down the total energy use into effect of
population size (P) and effect of energy use per capita (A⁎T)
(Fig. 3, A, B). The effect on total energy use caused by changes
in energy use per capita is similar in the two regions, whereas
the effect of change in population size is much larger in the
USA. A further decomposition of the effect of changes in
energy use per capita into A(GDP/capita) and T(ecoefficiency
measured as energy/GDP) shows that the USA had a larger
increase in GDP/capita that was balanced by greater improve-
ments in ecoefficiency (Fig. 3C,D). Thus we can conclude that
energy use grew faster in the USA due to faster population
growth.
The analysis above concerns differences in the growth of
energy use between the two regions. However, since the ab-
solute energy use is approximately t wice as high in the USA, it
is informative to analyze absolute differences between the
two regions. A comparison of absolute population size and
per capita energy use (Fig. 4A,B) shows that it is greater per
capita consumption rather than a larger population size that
causes higher total energy use in the USA. The per capita
energy use in the USA was three times as high in 1960 and
twice as high in 2002 compared to Western Europe (Fig. 4A).
Thus it is pertinent to ask which descriptors of society's
structure are associated with a high per capita consumption
rate. Such associations were investigated using PLS regres-
sion (Martens and Naess, 1989), a method that combines
ordination and regression. An ordination of the matrix of
predictor variables is constructed, such that the covariance
between the ordination axis and the dependent variable is
maximized. Orthogonal ordination axes are used as predic-
tors, which means that PLS regression, in contrast to ordinary
least squares regression, is robust against collinearity among
the independent variables. The energy use per capita in the
USA and each of the six European countries for each of the
years 1966–2002 was used as the dependent variable. As
predictors we used 10 variables that are hypothesized to
Fig. 4 –A) Energy use per capita during 1960–2002 for the USA
and for six European countries (France, Germany, Italy, the
United Kingdom, Austria and Switzerland). B) Population in
the USA and in the six European countries during the period.
Source: IEA (2006).
882 ECOLOGICAL ECONOMICS 68 (2009) 879–887

indicate energy use. One group of variables are expected to be
correlated with energy use in the transport sector, i.e., price of
gasoline, tax on gasoline, the number of vehicles per capita,
and population density (see, e.g. Schipper et al., 1997;
Carlsson-Kanyama and Lidén, 1999; Bottrill, 2006). Population
density may also influence energy use through other
mechanisms (Weisz et al., 2006). A second group of variables
are included as indicators of a post industrial “information
society”, which according to the EKC hypothesis should be
characterized by high technological level and de-
materialization of the economy (see, e.g. Jänicke et al., 1989;
Picton and Daniels, 1999, Hinterberger and Schmidt-Bleek,
1999). This group includes: the proportion of GDP that can be
attribute d to the service sector, the nu mber of TVs/capita, and
the proportion of energy produced by wind, geothermal,
solar, wood, and waste. Inclusion of the latter variable was
motivated by the hypothesis that societies which put an effort
tousemoreof“alternative”energy also pay attention to the
importance of decreasing total energy use. We also included
the proportion of the national energy supply that is con-
sumed as a measure of efficiency of the energy system. A low
efficiency (consumptionbsupply) may be caused by losses
during transport of electricity, heat loss due to cooling of
nuclear power plants, etc. Finally we included two variables
which generally affect society's economy, the world market
price of oil and the amount of household saving. The latter
variable was used based on the assumption that a more direct
purchasing of products and services instead of saving money
indicates a higher rate of energy use. The PLS analysis pro-
duced one significant component that explained 83.1% of
the variation in the energy use per capita between years and
countries (Q
2
=0.83, pb0.05, significance tested with cross
validation (Wold, 1978)). Fig. 5 shows that high tax on fuel and
a high gasoline price were negatively associated with energy
use. Weaker ne gative relationship s were found for populatio n
density, household savings, and national energy consump-
tion in relation to energy supply. The strongest positive
relationships were found for number of vehicles per capita,
energy use per GDP and the number of TV sets per capita. A
weaker positive relationship was found for the proportion of
GDP derived from the service sector of the economy. The
effect of world oil price changes between 1973 and 1979 that
can be seen in Fig. 2 showed no apparent relationship to the
overall variation in per capita consumption. This is also true
for the proportion of total energy consumption derived from
renewable so urces, such as wind, solar , waste and biomass. In
an alternative analysis we added a dummy variable as an
additional predictor, contrasting the USA and the six Eur-
opean countries (first component: Q
2
=0.88, 88.5% of variance
in energy/capita explained). The dummy variable showed the
strongest correlation with energy use but the rank order of
the other variables was unchanged. This illustrates that the
PLS analyses mainly reflected differences between the two
regions.
4. Discussion
This study indicates that improved efficiency in the use of
natural resources is insufficient to prevent further increases in
global resource use. The ecological footprint across 135
countries did not decrease at high levels of GDP per capita,
and the energy use in the USA and six European countries
during the period 1960–2002 did not decrease as GDP/capita
increased. The existence of an environmental Kuznets curve is
therefore not supported. Instead a rebound effect may occur;
the improved ecoefficiency, which is seen in many wealthy
countries, (e.g., Jänicke et al., 1989), is associated with eco-
nomic growth which increases the global ecological footprints
of these countries. One example of this apparent paradox is
that Finland, a wealthy and ecoefficient country has been
classified as both the most sustainable country (Devitt and
DeFusco, 2002) and the country that caused the fifth largest
per capita footprint in the world (Loh, 2002).
The USA appears to have departed from the stabilization of
its ecological footprint more rapidly than the West European
countries studied, mainly due to a higher population growth
rate. Both a higher birth rate and greater net migration con-
tribute to this pattern (WRI, 2006).
In addition to faster growth in energy use in the USA, there
was a higher absolute level of energy use during the study
period. This difference could be ascribed to higher energy use
per capita in the USA. Since the PLS-analysis of energy per
capita largely reflected differences between the USA and the
six European countries it can be used to obtain insights into
the mechanisms underlying this pattern. High correlations
between per capita energy use and the price of gasoline, tax on
gasoline, and the number of vehicles per capita, point to the
importance of the transport sector. Taxes on gasoline were
two to three times higher in the Western European countries
than in the USA during the study period (IEA, 2002), and the
transport energy consumption per capita between 1970 and
1990 was about twice as high in the USA compared to Western
Fig. 5 –Loadings from a PLS regression analysis using per
capita energy use in the USA, France, Germany, Italy, United
Kingdom, Austria and Switzerland, during 1966–2002 as the
dependent variable. Data sources are provided in
Appendix B, Table 2.
883ECOLOGICAL ECONOMICS 68 (2009) 879–887

European countries (DOE/EIA, 1994). The total volume of
transport in 1998 was 4223 million vehicle km in the USA,
and 2131 vehicle km in the six European Countries (OECD,
2001). Transportation has increased steadily and in 1999
accounted for about two-third of total US petroleum demand
compared with about 50% before 1973 (Fenn, 2000). It is the
greater frequency of vehicle use, not the differences in length
of journeys that determines the larger total transport energy
use in the USA (Schipper et al., 1997). The negative correlation
between population density and energy use per capita may
also reflect the fact that transportation systems differ between
the regions, since low-density settlement favours cars over
more energy efficient transportation modes (Schipper et al.,
1997). There might also be more general causes for this
population density factor; like a high net import to densely
populated parts of the study regions, and hence the externa-
lisation of material and energy intensive processes (Anders-
son and Lindroth, 2001; Weisz et al., 2006).
An important assumption underlying the EKC hypothe-
sis is that movement towards an “information society”,
characterized by a high level of technology and increased
importance of the service sector, will lead to reduced en-
vironmental impacts (Jänicke et al., 1989; Daniels, 1992;
Picton and Daniels, 1999). Thus, we expected to see a negative
relationship between per capita energy use and the propor-
tion of GDP that can be attributed to the service sector. The
same relationship was expected for the number of TVs/ca-
pita, which was included as a technology indicator. However,
we found positive relationships, contradicting the assump-
tion that movement towards a highly efficient “information
society”leads to decreased use of natural resources. This
finding is supported by other studies. For example Giampie-
tro (1999) found, in a study of 107 nations, a negative rela-
tionship between energy consumption per capita and the
proportion of the population working in the productive sec-
tor. York, Rosa and Dietz (2005) showed in a study of 139
nations that service sector development does not have a
significant relationship with ecological footprint intensity
and Salzman (2000) concluded that the growth of the service
sector during the last decades in the world's wealthier
countries has increased their overall economic activity and
thus overall resource consumption.
In summary, our comparison between the two regions
shows that variations in the size of the rebound effect can be
explained by differences in population growth rate and in
absolute levels of per capita use that were already present at
the beginning of the study period (1960). How this difference
arose could thus not be studied directly, but our comparison
during the study period indicates that the organization of the
transportation sector is likely to be important. Next we discuss
a more general strategy for reducing environmental impacts
to sustainable levels.
Numerous promising environmental techniques are in
various stages of development. One example is the use of
solar energy techniques in combination with hydrogen as an
energy transport medium (Jaramilli et al., 1998; Green, 2000).
Another example is better recycling (Recycling international,
2002). To determine whether such technical improvements
can be sufficient to curb the “rebound effects”,methodsfor
measuring the mobilization of material and energy per capita
(Bringezu et al., 1995), at the municipal level (Burström, 1998),
and in different regions (Jänicke et al., 1993; Wackernagel
et al., 1999; Hanley et al., 1999) could be useful approaches.
When such measurements are made, it is important to in-
clude quality aspects (e.g. Odum, 1996; Cleveland et al., 2000).
Furthermore, it is important to ensure that the global level is
included. For example, investments in railways can lead to a
local decrease in air pollution. But such investments can also
increase the economic activity in a region, which in turn lead
to a net increase in the human global ecological footprint.
One way to avoid such mistakes would be to include
calculations of ecological footprints in local environmental
impact assessments (EIA as defined by Morris and Therivel,
2001). However, the overall finding in this study indicates
that increased “ecoefficiency”alone will be insufficient to
counterbalance the effects of increasing affluence and
increasing population. To overcome this “rebound effect”
problem, to find a way to sustainability, it will be justified to
implement a new policy directed more to the driving forces
behind the “rebound effect”. We suggest a “two-step
strategy”.
A first reasonable step would be to determine sustainable
levels of anthropogenic material and energy flows locally–
globally. There have been some attempts to calculate such
levels, for nations (Moran et al., 2008), for regions (Spangen-
berg, 1995), and for the entire globe (Pimentel et al., 1994;
Benking et al, 1995). The second step will be the adaption of P,
A and T in the IPAT equation so that sustainable material and
energy flows are reached. A prerequisite for a successful
adaption of the P factor is an achievement of a demographic
transition without the concomitant increase in the use of
natural resources that have preceded most demographic
transitions until now. Benking et al. (1995) propose that this
may require that gains resulting from increased eco-effi-
ciency in developed countries are transferred to poor coun-
tries for development of pension systems and education. To
change individual consumption, the A-factor, actions aimed
at bringing about higher-level changes in the socio-eco-
nomic-cognitive system –i.e. changing cultural values and
worldviews –will be most effective (Brown and Cameron,
2000). To find ways of promoting an alternative value orien-
tation also on the society level, greater efforts are required
with respect to relevant research and education, as suggested
by Stern (1993) and Kates et al. (2001). An implementation of
“the two step strategy”outlined above, i.e. determination of
globally sustainable material flows and adapting our activ-
ities in accordance with the IPAT equation, may lead to
reduced extraction of natural resources. If such an interna-
tional approach cannot be performed, there is an apparent
risk that no n-renewable resource s will be exhausted, and that
the use of renewable resources, including much of the earth's
biodiversity, will continue until their renewable capacity is
lost.
Acknowledgments
We would like to thank Mathis Wackernagel for his permis-
sion to use earlier published data.
884 ECOLOGICAL ECONOMICS 68 (2009) 879–887

Appendix A
Data used for the IPAT analysis of energy use. GDP is given as
billion US dollars at 2000 prices and PPPs, population is given
in millions, and energy use is expressed as EJ.
Year USA Western Europe
GDP Population Energy
use
GDP Population Energy
use
1960 2553.6 180.7 42.7 2203.6 232.7 18.4
1961 2616.6 183.7 43.4 2315.3 234.7 19.0
1962 2751.7 186.6 45.1 2417.9 237.5 20.4
1963 2861.0 189.3 47.6 2524.3 239.9 21.9
1964 3020.7 191.9 49.4 2657.0 242.2 22.8
1965 3189.8 194.3 51.4 2764.7 244.4 23.7
1966 3379.6 196.6 54.4 2868.5 246.3 24.1
1967 3471.6 198.7 56.8 2953.6 247.7 25.0
1968 3617.1 200.7 59.6 3100.5 249.1 26.5
1969 3713.7 202.7 62.5 3278.2 251.0 28.5
1970 3721.7 205.1 65.2 3433.4 252.8 33.9
1971 3850.5 207.7 66.7 3541.5 254.6 34.8
1972 4065.8 209.9 70.0 3679.8 256.0 36.1
1973 4304.8 211.9 72.7 3887.5 257.3 38.3
1974 4284.4 213.9 71.0 3954.9 258.0 37.4
1975 4276.9 216.0 69.5 3907.0 258.2 35.6
1976 4507.0 218.1 74.2 4078.9 258.3 37.8
1977 4717.0 220.3 76.7 4195.0 258.5 37.8
1978 4981.9 222.6 78.9 4324.3 259.0 38.5
1979 5140.4 225.1 78.7 4494.5 259.4 40.2
1980 5128.0 227.7 75.8 4553.1 260.1 39.0
1981 5257.4 230.0 73.8 4549.5 260.7 37.8
1982 5153.6 232.2 70.7 4610.5 261.0 36.9
1983 5386.3 234.3 70.7 4702.3 261.1 37.2
1984 5774.0 236.4 73.8 4816.0 261.2 38.2
1985 6011.0 238.5 74.7 4942.9 261.5 39.7
1986 6217.2 240.7 74.8 5076.2 262.0 40.1
1987 6425.1 242.8 77.8 5206.7 262.5 40.7
1988 6690.0 245.1 80.9 5422.9 263.4 41.1
1989 6926.3 247.4 82.0 5602.3 264.6 41.4
1990 7055.0 250.2 80.8 5775.6 266.0 41.6
1991 7041.3 253.5 81.4 5883.9 267.3 42.4
1992 7276.2 256.9 82.9 5966.2 268.6 41.9
1993 7472.0 260.3 84.7 5955.8 269.8 42.0
1994 7775.5 263.5 86.5 6118.8 270.7 41.6
1995 7972.8 266.6 87.5 6263.0 271.4 42.7
1996 8271.4 269.7 89.5 6352.4 272.1 44.2
1997 8647.6 273.0 90.6 6496.3 272.8 43.7
1998 9012.5 276.2 91.5 6668.0 273.2 44.3
1999 9417.1 279.3 93.9 6832.6 273.9 44.1
2000 9764.8 282.4 96.6 7075.5 274.7 44.5
2001 9838.9 285.4 94.6 7197.0 275.7 45.5
2002 9997.6 288.2 95.9 7257.2 276.7 44.9
Appendix B
Predictor Coefficient SE
Constant 2.49 0
Vehicles/cap 0.19 0.003
Alternative energy −0.03 0.007
843474
(continued)
Predictor Coefficient SE
Price of gasoline −0.17 0.005
Tax on fuel (%) −0.20 0.005
Energy consump./supply −0.10 0.003
Oil price −0.02 0.008
Service (%) 0.10 0.004
TV/cap 0.17 0.007
Population density −0.13 0.005
Household savings −0.13 0.003
The table shows PLS regression coefficients corresponding to
centered and scaled X-values, and scaled but uncentered Y-values, and
jack knife standard errors computed by cross validation.
Variable Data source
Energy use/capita IEA (2006)
Vehicles/cap IRF (1966–2005)
Alternative energy EIA (2003)
Price of gasoline OECD (2005)
Tax on fuel (%) OECD (2005)
Energy consump./supply IEA (2006)
Oil price BP (2004)
Service (%) WRI (2006)
TV/cap WRI (2006)
Population density WRI (2006)
Household savings OECD (2003)
REFERENCES
Andersson, J., Lindroth, M., 2001. Ecologically unsustainable trade.
Ecological Economics 37, 113–122.
Benking,H.,Brauer,G.W.,Fliedner,T.M.,Greiner,C.,Malaska,P.,
Morath, K., Pestel, R., Radermacher, F.J., 1995. Robust paths to
global stability: tough but feasible. The XIV World Conference
of World Futures Studies Federation, 25–30 July 1995, Nairobi,
Kenya.
Berkhout, P.H.G., Muskens, J.C., Velthuijsen, J.W., 2000. Defining
the rebound effect. Energy Policy 28, 425–432.
BP, 2004. Statistical Review of World Energy 2004. British
Petroleum.
Burström, F., 1998. Municipal materials accounting and
environment management. Dissertation. The Royal Institute of
Technology, Stockholm.
Bottrill, C. 2006. Personal Carbon Trading: Excluding Public
Transportation. UK Energy Research Centre. Working paper,
December 2006. UKERC/WP/DR/2006/007.
Bringezu, S., Schutz, H., Behrensmeier, R., Shmidt-Bleek, F., 1995.
Indicating environmental pressure for national economies
and industrial sectors on the basis of integrated economic
and environmental accounting. Wuppertal Institute, Wuppertal.
Brown, P.M., Cameron, L.D., 2000. What can be done to reduce
overconsumption? Ecological Economics 32, 27–41.
Carlsson-Kanyama, A., Lindén, A.-L., 1999. Travel patterns and
environmental effects now and in the future: implications of
differences in energy consumption among socio-economic
groups. Ecological Economics 30, 405–417.
Table 1  Regression statistics for the PLS analysis of
relationships between energy per capita and 10 predictor
variables
Table 2  Data sources for the PLS analysis of relation‐
ships between energy use per capita and 10 predictor
variables
Table 1 (continued )
885ECOLOGICAL ECONOMICS 68 (2009) 879–887

Cavlovic, T.A., Baker, K.H., Berrens, R.P., Gawande, K., 2000. A
meta-analysis of environmental Kuznets curve studies.
Agricultural and Resource Economics Review 29, 32–42.
Cleveland, C.J., Kaufmann, R.K., Stern, D.I., 2000. Aggregation and
the role of energy in the economy. Ecological Economics 32,
301–317.
Cohen, J.E., 1995. Population growth and human carrying capacity.
Science 269, 341–346.
Daniels, P., 1992. Barriers to sustainable development in natural
resource-based economies: Australia as a case study. Society &
Natural Resources 5, 247–262.
Devitt, J., DeFusco, D., 2002. Environmental sustainability index.
An initiative of the Global Leaders of Tomorrow Environmental
Task Force, World Economic Forum. Annual Meeteing 2002.
Yale university, Columbia university, USA.
DOE/EIA, 1994. Energy use and carbon emissions: some
international comparisons. Energy Information Administration.
Office of Energy markets and End use. U.S. Department of
Energy, Washington, DC.
Ehrlich, P.R., Holdren, J.P., 1971. Impact of population growth.
Science 171, 1212–1217.
EIA, 2003. International Energy Annual 2003. Energy Information
Administration, USA.
Fan, Y., Liu, L.-C., Wu, G., Wei, Y.-M., 2006. Analyzing impact
factors of CO
2
emissions using the STRIPAT model.
Environmental Impact Assessment Review 26, 377–395.
Fenn, M. (Ed.), 2000. Transportation Statistics Annual Report 2000.
Bureau of Transportation Statistics. U.S. Department of
Transportation, Washington D.C.
Friman, E., 2002. No Limits. The 20th Century Discource of
Economic Growth. Doctoral Thesis. Department of Historical
Studies. Umeå University, Sweden.
Gawande, K., Berrens, R.P., Bohara, A.K., 2001. A
consumption-based theory of the environmental Kuznets
curve. Ecological Economics 37, 101–112.
Giampietro, M., 1999. Economic growth, human's disturbance to
ecological systems, and sustainability. In: Walker, R. (Ed.),
Ecosystems of Disturbed Ground. Ecosystems of the World,
vol. 16. Elsevier, Amsterdam.
Gleick, P.H., 2000. The changing water paradigm. A look at
twenty-first century water resources development. Water
International 25, 127–138.
Green, M., 2000. Power to the people: sunlight to electricity use in
solar cells. UNSW Press, Sydney.
Greening, L.A., Greene, D.L., Difiglio, C., 2000. Energy efficiency and
consumption –the rebound effect –a survey. Energy Policy 28,
389–401.
Hails, C., Loh, J., Goldfinger, S. (Eds.), 2006. Living Planet Report
2006. World Wildlife Found for Nature (WWF).
Hanley, N., Moffatt, I., Faichney, R., Wilson, M., 1999. Measuring
sustainability: A time series of alternative indicators for
Scotland. Ecological Economics 28, 55–73.
Hannah, L., Loshe, D., Hutchinson, C., Carr, L., Lankerani, A., 1994.
A preliminary inventory of human disturbance of World
ecosystems. Ambio 23, 246–250.
Herring, H., 1998. Does Energy Efficiency Save Energy: The
Economists Debate. EERU Report No 074-July 1998.
Hinterberger, F., Schmidt-Bleek, F., 1999. Demateralization, MIPS
and Factor 10. Physical sustainability indicators as a social
device. Ecological Economics 29, 53–56.
IEA, 1996. Energy balances of OECD countries. International Energy
Agency, Paris.
IEA, 2002. Energy prices and taxes 2002. International Energy
Agency, Paris.
IEA, 2003. Key World Energy Statistics. International Energy
Agency, Paris.
IEA, 2006. International Energy Agencies database, 2006.
IRF, 1966–2005. World Road Statistics. International Road
Federation (IRF).
IUCN, 2002. The world conservation strategy. International Union
for Conservation of Nature (IUCN).
Jaramilli, O.A., Arriaga, L.G., Sebastian, P.J., Fernández, A.M., Del
Rio, J.A., 1998. Application of fibre optics in the hydrogen
production by photoelectrolysis. International Journal of
Hydrogen Energy 23, 985–993.
Jevons, W.S., 1865. The Coal Question: An Inquiry Concerning the
Progress of the Nation and the Probable Exhaustion of our
Coal-mines, (Third Edition 1906). Macmillan & Co, London.
Jokinen, P., Malaska, P., Kavio-oja, J., 1998. The environment in an ´
information society´. A transition stage towards a more
sustainable development? Futures 30 (6), 485–498.
Jänicke, M., Mönch, H., Ranneberg, T., Simonis, U.E., 1989.
Structural change and environmental impact: evidence on
thirty-one countries in east and west. Environmental
Monitoring and Assessment 12, 99–114.
Jänicke, M., Mönch, H., Binder, M., 1993. Ecological aspects of
structural change. Intereconomics 159–169 July/August.
Kates, R.W., 1994. Sustaining life on the earth. Scientific American
92–99 October.
Kates, R.W., Clark, W.C., Corell, R., Hall, M., Jaeger, C.C., Lowe, I.,
McCarthy, J.J., Schellnhuber, H.J., Bolin, B., Dickson, N.M.,
Faucheux, S., Gallopin, G.C., Grubler, A., Huntley, B., Jäger, J.,
Jodha, N.S., Kasperson, R.E., Mabogunje, A., Matson, P.,
Mooney, H., Moore, B., O'Riordan, T., Svedin, U., 2001.
Sustainability science. Science 292, 641–642.
Kågeson, P., 1997. Growth versus the Environment —Is There a
Trade-off? Doctoral Dissertation. Department of Environmental
and Energy Systems Studies, Lund University, Sweden.
Laurance, W.F., 2001. Future shock: forecasting a grim fate for the
earth. Trends in Ecology & Evolution 16, 531–533.
Loh, J., Randers, J., MacGillvray, A., Kapos, V., Jenkins, M.,
Groombridge, B., Cox, N. 1998. Living Planet Report 1998.
Overconsumption is driving the rapid decline of world's
natural environments. World Wildlife Fund (WWF).
Loh, J. (Ed.), 2002. Living Planet Report 2002. World Wildlife Fund
(WWF).
Loh, J., Wackernagel, M., 2004. Living planet report 2004. World
Wildlife Fund (WWF).
Malaska, P., Kavio-oja, J., Luukanen, J. 1999. Sustainability and
economic growth: A theoretical framework and empirical
demonstrations. FUTU-publication 4/99. Turku school of
economics and business administration. Finland futures
research centre.
Martens, H., Naess, T., 1989. Multivaraite calibration. John Wiley
and Sons, New York.
McNeill, J., 2000. Something new under the sun. An Environmental
History of the Twentieth Century. Allan Lane, London.
Moran, D.D., Wackernagel, M., Kitzes, J.A., Goldfinger, S.H., Boutaud,
A., 2008. Measuring sustainable development —Nation by
nation. Ecological Economics 64, 470–474.
Morris, P., Therivel, R., 2001. Methods of Environmental Impact
Assessment. Spon, London.
Noble, I.R., Dirzo, R., 1997. Forests as human dominated
ecosystems. Science 277, 522–525.
Odum, H.T., 1996. Environmental Accounting: Emergy and
Environmental Decision Making. John Wileys & Sons,
New York.
OECD, 2001. OECD Environmental Indicators 2001. Towards
Sustainable Development.
OECD, 2003. OECD Economic Outlook.
OECD, 2005. Energy Prices and Taxes. OECD, Paris.
Picton, T., Daniels, P.L., 1999. Ecological restructuring for
sustainable development: evidence from Australian economy.
Ecological Economics 29, 405–425.
Pimentel, D., Harman, R., Pacenza, M., Pecarsky, J., Pimentel, M.,
1994. Natural resources and an optimum human population.
Population and Environment 15, 347–369.
Recycling international., 2002. www.recyclinginternational.com.
886 ECOLOGICAL ECONOMICS 68 (2009) 879–887

Schipper, L., DesRosiers, J.-P., Justus, D., Cornell, R., Sullivan, S.,
1997. Indicators of energy use and efficiency. Understanding
the link between energy and human activity. OECD/IEA,
Paris.
Salzman, J., 2000. Environmental protection beyond the
smokestack: addressing the impact of the service economy.
Corporate Environmental Strategy 7, 20–37.
Schulze, P.C., 2002. I=PBAT. Ecological Economics 40, 149–150.
Spangenberg, J.H. (Ed.), 1995. Towards Sustainable Europe, the
study. Environmental space, sustainable Europe. Friends of the
earth Europe.
Stern, D.I., Common, M.S., Barbier, E.B., 1996. Economic growth
and environmental degradation: the environmental Kuznets
curve and sustainable development. World Development 24,
1151–1160.
Stern, P.C., 1993. A second environmental science:
human-environment interactions. Science 260, 1897–1899.
Sun, J.W., 1998. Change in energy consumption and energy
intensity: a complete decomposition model. Energy Economics
20, 85–100.
Tilman, D., Fargione, J., Wolff, B., Dántonio, C., Dobson, A.,
Howarth, R., Schindler, D., Schlesinger, W., Simberloff, D.,
Swackhamer, D., 2001. Forecasting agriculturally driven global
environmental change. Science 292, 281–284.
UNEP, 2002. Global Environmental Outlook 3. Past, present and
future perspectives. United Nations Environmental
Programme (UNEP).
Vehmas,J., Luukanen, J., Kavio-oja, J. 2004. Technology development
versus economic growth —an analysis of sustainable
development. EU–US seminar: New Technology Foresight,
Forecasting & Assessment Methods—Seville 13–14 May 2004.
Vitousek, P.M., Erlich, P.R., Matson, P.A., 1986. Human
appropriation of the products of photosynthesis. BioScience 36,
368–373.
Vitousek, P.M., 1994. Beyond global warming: ecology and global
change. Ecology 75, 1861–1876.
Vitousek, P.M., Mooney, H.A., Lubchenco, J., Melillo, J.M., 1997.
Human domination of Earth's ecosystems. Science 277,
494–499.
Wackernagel, M., Lewan, L., Borgström, C., 1999. Evaluating the
use of natural capital with the ecological footprint. Ambio 28,
604–612.
Wackernagel, M., Linares, A.C., Deumling, D., Schultz, N.B.,
Sánchez, M.A.V., Falfán, I.S.L., 2000. Ecological footprint In:
Loh, J. (Ed.), Living Planet Report 2000. World Wide Found For
Nature (WWF).
Weisz, H., Krausmann, F., Amann, C., Eisenmenger, N., Erb, K.-H.,
Hubacek, K., Fisher-Kowalski, M., 2006. The physical economy
of the European Union: cross-country comparision and
determinants of material consumption. Ecological Economics
58, 676–698.
WRI, 2006. World Resources Institute, 2006. www.wri.org.
Wold, S., 1978. Cross validatory estimation of the number of
components in factor and principal components analysis.
Technometrics 20, 397–406.
York, R., Rosa, E.A., Dietz, T., 2005. The ecological footprint
intensity of national economies. Journal of Industrial Ecology
8 (4), 139–154.
887ECOLOGICAL ECONOMICS 68 (2009) 879–887