ArticlePDF Available

Abstract

This paper examines the macroeconomic rebound effect for the global economy arising from energy-efficiency policies. Such policies are expected to be a leading component of climate policy portfolios being proposed and adopted in order to achieve climate stabilisation targets for 2020, 2030 and 2050, such as the G8 50% reduction target by 2050. We apply the global “New Economics” or Post Keynesian model E3MG, developing the version reported in IPCC AR4 WG3. The rebound effect refers to the idea that some or all of the expected reductions in energy consumption as a result of energy-efficiency improvements are offset by an increasing demand for energy services, arising from reductions in the effective price of energy services resulting from those improvements. As policies to stimulate energy-efficiency improvements are a key part of climate-change policies, the likely magnitude of any rebound effect is of great importance to assessing the effectiveness of those policies. The literature distinguishes three types of rebound effect from energy-efficiency improvements: direct, indirect and economy-wide. The macroeconomic rebound effect, which is the focus of this paper, is the combination of the indirect and economy-wide effects. Estimates of the effects of no-regrets efficiency policies are reported by the International Energy Agency in World Energy Outlook, 2006, and synthesised in the IPCC AR4 WG3 report. We analyse policies for the transport, residential and services buildings and industrial sectors of the economy for the post-2012 period, 2013–2030. The estimated direct rebound effect, implicit in the IEA WEO/IPCC AR4 estimates, is treated as exogenous, based on estimates from the literature, globally about 10%. The total rebound effect, however, is 31% by 2020 rising to 52% by 2030. The total effect includes the direct effect and the effects of (1) the lower cost of energy on energy demand in the three broad sectors as well as of (2) the extra consumers’ expenditure from higher (implicit) real income and (3) the extra energy-efficiency investments. The rebound effects build up over time as the economic system adapts to the higher real incomes from the energy savings and the investments.
ORIGINAL RESEARCH
The macroeconomic rebound effect and the world economy
Terry Barker &Athanasios Dagoumas &
Jonathan Rubin
Received: 22 July 2008 / Accepted: 5 May 2009 / Published online: 28 May 2009
#Springer Science + Business Media B.V. 2009
Abstract This paper examines the macroeconomic
rebound effect for the global economy arising from
energy-efficiency policies. Such policies are expected
to be a leading component of climate policy portfolios
being proposed and adopted in order to achieve
climate stabilisation targets for 2020, 2030 and
2050, such as the G8 50% reduction target by 2050.
We apply the global New Economicsor Post
Keynesian model E3MG, developing the version
reported in IPCC AR4 WG3. The rebound effect
refers to the idea that some or all of the expected
reductions in energy consumption as a result of
energy-efficiency improvements are offset by an
increasing demand for energy services, arising from
reductions in the effective price of energy services
resulting from those improvements. As policies to
stimulate energy-efficiency improvements are a key
part of climate-change policies, the likely magnitude
of any rebound effect is of great importance to
assessing the effectiveness of those policies. The
literature distinguishes three types of rebound effect
from energy-efficiency improvements: direct, indirect
and economy-wide. The macroeconomic rebound
effect, which is the focus of this paper, is the
combination of the indirect and economy-wide
effects. Estimates of the effects of no-regrets efficiency
policies are reported by the International Energy
Agency in World Energy Outlook, 2006, and synthe-
sised in the IPCC AR4 WG3 report. We analyse
policies for the transport, residential and services
buildings and industrial sectors of the economy for
the post-2012 period, 20132030. The estimated direct
rebound effect, implicit in the IEA WEO/IPCC AR4
estimates, is treated as exogenous, based on estimates
from the literature, globally about 10%. The total
rebound effect, however, is 31% by 2020 rising to 52%
by 2030. The total effect includes the direct effect and
the effects of (1) the lower cost of energy on energy
demand in the three broad sectors as well as of (2) the
extra consumersexpenditure from higher (implicit)
real income and (3) the extra energy-efficiency invest-
ments. The rebound effects build up over time as the
economic system adapts to the higher real incomes
from the energy savings and the investments.
Keywords Rebound effect .Energy efficiency .
Macroeconomic modelling .Top-down/bottom-up
modelling .Post-2012 policies
Energy Efficiency (2009) 2:411427
DOI 10.1007/s12053-009-9053-y
T. Barker (*):A. Dagoumas
Cambridge Centre for Climate Change Mitigation Research
(4CMR), Department of Land Economy,
University of Cambridge,
19 Silver Street,
Cambridge CB3 9PE, UK
e-mail: tsb1@cam.ac.uk
e-mail: an325@cam.ac.uk
J. Rubin
School of Economics, University of Maine,
5782 Winslow Hall,
Orono, ME 04469-5782, USA
e-mail: jonathan.rubin@umit.maine.edu
Introduction
This paper explores the macroeconomic rebound effects
for the global economy from climate policies based on
energy-efficiency improvements and programmes
reported by the International Energy Agency (IEA
2005) in the World Energy Outlook 2006 (IEA
2006), synthesised in the 2007 IPCC Report and
discussed elsewhere in this volume. We use a global
New Economics, Post Keynesian model with esti-
mated energy demand equations to illustrate the
potential scale of the rebound problem and suggest
how policy portfolios and strategies can be developed
to manage, monitor and counter the rebound effects.
The rebound effect refers to the idea that some or
all of the expected reductions in energy consumption
as a result of energy-efficiency improvements are
offset by an increasing demand for energy services,
arising from reductions in the effective price of
energy services resulting from those improvements
(Greening et al. 2000 for a survey). As policies to
stimulate energy-efficiency improvements are a key
part of climate-change policies (Geller et al. 2006),
the likely magnitude of any rebound effect is of great
importance to assessing the effectiveness of those
policies. However, the magnitude, the definition and
the scope of rebound effects are controversial
(Brookes 1990; Grubb 1990).
The literature distinguishes between three types of
rebound effect from energy-efficiency improvements:
direct, indirect and economy-wide (Greening et al.
2000):
&Direct rebound effects: Improved energy efficiency
for a particular energy service will decrease the
effective price of that service and should therefore
lead to an increase in consumption of that service.
This will tend to offset the expected reduction in
energy consumption provided by the efficiency
improvement.
&Indirect rebound effects: For consumers, the lower
effective price of the energy service will lead to
changes in the demand for other goods and
services. To the extent that these require energy
for their provision, there will be indirect effects on
aggregate energy consumption.
&Economy wide rebound effects: A fall in the real
price of energy services will reduce the price of
intermediate and final goods throughout the
economy, leading to a series of price and quantity
adjustments, with energy-intensive goods and
sectors gaining at the expense of less energy-
intensive ones. Energy-efficiency improvements
may also increase economic growth, which should
itself increase energy consumption.
Of particular interest for global climate (and energy)
policy is the magnitude of the macroeconomic rebound
effect, which we take to cover the indirect and
economy-wide rebound effects extended to include
effects on consumption from the implicit higher real
income and investment required for the energy-
efficiency policies to be effective. The Khazzoom
Brookes postulate (Khazzoom 1980; Brookes 1990;
Saunders 1992,2000) is an interpretation of the
rebound effect at the macroeconomic level suggesting
that the aggregate energy saving from energy-
efficiency measures might be offset by associated
increases in energy demand. If the energy-efficiency
measures lead eventually to even more energy being
consumed than otherwise, the rebound effect has been
termed a backfireeffect (Saunders 2000, p. 440).
The underlying assumption in our analysis is that the
no-regrets options can be identified by targeted policies
and measures and that they pay for themselves assuming
social discount rates. There will be an investment cost of
the measures, but it is assumed that resources will be
available so that the investment will not replace other
investment or consumption, i.e. there are under-
employed resources in the system sufficient to avoid
inflation. This assumption is more plausible when the
construction industry is working at less than full
capacity as it is after 2008 in many countries as an
outcome of the credit crunchof 2007 and 2008.
The macroeconomic rebound effect considered in
this paper is the combination of the indirect and
economy-wide effects. We start with the estimated
effects of the no-regrets options for final demand for
electricity and fossil fuels that are synthesised by the
IEA in the World Energy Outlook 2006 (IEA 2006).
This report in agreement with the IPCC AR4
considers that electricity savings are found at rela-
tively low costs, and they are, therefore, expected to
be implemented first. The assessed effects cover
energy saving from energy-efficiency policies for
Transport, residential-service Buildings (henceforth
Buildings), and Industry broad sectors of the
economy for the post-2012 period, 20132030.
412 Energy Efficiency (2009) 2:411427
The overall results are decomposed into effects
assuming first that each of the three sectors undertake
the policies unilaterally but implemented for both the
OECD and non-OECD regions and second that the
OECD regions and the non-OECD regions take
unilateral action across the three sectors. Finally, we
have assessed and reported below where the rebound
effects originate by dividing them up into
1. direct rebound effects assumed to be implicit in
the IEA estimates
2. effects from the energy savings per se, reducing
costs and prices for households, businesses and
governments
3. effects from the extra imputed real incomes
accruing to consumers as a result of lower
spending on traditional biomass, oil, gas and
electricity
4. effects from the higher investment required to
generate the energy savings.
Literature reviewof the paper provides a brief
review of the debates on the macroeconomic rebound
effect in relation to energy (and climate) policy.
Modellingdescribes the approach taken here to
modelling the macroeconomic rebound effect. Descrip-
tion of policies and scenariosdescribes the IEA WEO
2006 (IEA 2006) energy-efficiency policies incor-
porated into this modelling and the scenarios used.
Resultsdescribes the results, including the overall
impacts of energy-efficiency policies on energy de-
mand, economic activity and CO
2
emissions and the
sources and magnitude of the macroeconomic rebound
effect. Conclusionsprovides some conclusions.
Literature review
The literature on the rebound effect has developed in
recent years as climate mitigation has moved up the
policy agenda (Herring and Sorrell 2009;Herring2004;
Schipper and Grubb 2000;Vikström,2004; Grepperud
and Rasmussen 2004; International Energy Agency
2005; Sorrell and Dimitropoulos 2007; Sorrell 2007).
The topic has proved controversial, partly through
differences between an energy-engineering approach,
which identifies no-regrets options for energy efficiency
and which is normally adopted in bottom-upenergy
systems models, and a traditional economics approach,
which assumes that no-regrets options do not exist,
except in the case of market failures, and which is
adopted in top-downequilibrium models assuming no
market failures. However, the debate is also about the
source of the energy-efficiency improvements, i.e.
whether they come from the energy-efficiency policies
or from a general improvement in productivity of
energy-using equipment. The different assumptions
about the source have different consequences because
the policies require investment in energy-saving equip-
ment such as more efficient vehicle engines, or home
insulation, to be effective, whereas the energy saving
from technological progress is treated as manna from
heavenin the top-down models.
Brookes (1990) adopted the traditional economic
argument that technological progress has led to
significant increases in energy productivity but that
this has been offset by faster growth in general
productivity and output and so to higher energy use.
Policies for improved energy efficiency may lead to
higher energy use (backfirein the literature) and a
rise in GHG emissions, depending on the source of
energy, unless energy prices increased at the same
time the energy-efficiency policies were introduced.
Grubb (1990) opposed this interpretation, arguing that
there are significant differences between naturally-
occurringenergy-efficiency improvements from
on-going technological change, and energy-efficiency
improvements as a result of targeted policies and
measures. The differences between the two sides
reflect a different view of the efficiency of the market.
If the market is perfectly efficient, then the traditional
view holds and the efficiency improvements come
from exogenous technological change. If there are
market failures, then policies can address them, and
efficiencies can be improved but require additional
investment.
The issue was first raised in The Coal Question
(Jevons 1865/1905). He argued: It is a confusion of
ideas to suppose that the economical use of fuel is
equivalent to diminished consumption. The very
contrary is the truth’…The reduction of the
consumption of coal, per ton of iron, to less than
one third of its former amount, was followed, in
Scotland, by a tenfold increase in total consumption,
between the years 1830 and 1863, not to speak of the
indirect effect of cheap iron in accelerating other coal-
consuming branches of industry.Jevons was one of
the first neoclassical economists, and the issue here is
one of rapid technological change in the whole
Energy Efficiency (2009) 2:411427 413
economy (the Industrial Revolution). He is not
considering either energy-efficiency policies or mar-
ket failures, so his analysis is less relevant to the
effects of energy-efficiency policies today. However,
the general issue of lower cost of energy services and
rapid economic development is relevant in late
twentieth century and in projections to 2030, with
India and China transforming their economies.
One of the main reasons behind the debate is the lack
of a rigorous theoretical framework that can describe the
mechanisms and consequences of the rebound effect at
the macro-economic level (Dimitropoulos 2007).
There exist several models built on different economic
framework, e.g. Post Keynesian models, neoclassical
models of economic growth, computable general
equilibrium models and alternative models for
policy evaluation which were used to evaluate the
rebound effect (Barker et al. 2007; Grepperud and
Rasmussen 2004; Saunders 2008; Small and Van
Dender 2007; Sorrell 2007; Sorrell et al. 2009;Wei
2006). The multi-disciplinary risk analysis carried out
by the Stern Review team (Stern 2006) and the IPCC
4th Assessment Report (IPCC AR4 2007)has
highlighted that important weaknesses of the tradi-
tional, neoclassical approach, especially as regards the
treatment of uncertainty and risks challenges the
validity or confidence that policy makers should place
in policies that have long and (largely) irreversible
consequences. The equilibrium-based models by
themselves are not, in our view, appropriate for
providing an adequate understanding of the climate
change problem (Barker 2008), especially where
energy-efficiency measures constitute basic climate
policies. The rebound effect relevant in the study of
climate change mitigation is essentially a behavioural
response to an improvement in energy efficiency that
comes not as manna from heavenbut from detailed
sectoral policies designed to identify and overcome
market failures. Modelling approaches that fail to
include the apparent market failures arising when
consumer and business behaviours are assessed in
detail (i.e. by assuming that such failures do not exist)
may not properly estimate this effect.
Mandated efficiency improvements (appliance
standards, residential and services building codes,
fuel economy standards) and efficiency improvements
from education or recognition of opportunities (better
business practices) are basically different from price
(tax) or quantity (emission allowances) policies.
Efficiency standards overcome two market failures:
first social rates of return are generally lower than
private rates of return, so more and stronger measures
are justified; and second under conditions of risk
aversion (a particular piece of capital may not deliver
the expected return), society should be risk neutral
with respect to capital improvements and can offset
private risks by collective action, whereas private
individual agents are likely to be more risk averse and
hence less likely to take action. For example, first
purchasers of private automobiles typically hold
these capital purchases for 5 years. If the price of
new automobiles increases due to fuel-economy
technologies, the sales-weighted average value of
automobiles, on average after 5 years discounting at
3%pa, provides an effective residual value of 32.8%
(US DfT 2008, p. VII-42). This is far below the
lifetime social benefits that accrue from the average
new car vehicle lifetime of 15 years. Thus, there are
significant market failures, including principal-agent
failures (or a mismatch between social and private
behaviour) that cause systemic inefficiencies in
energy-capital-investment decisions. These principal
agent problems, plus those ignored by a lack of wide-
spread markets for climate change damages, are not
reflected in market energy costs.
To allow for such market failures, a global New
Economics, Post Keynesian model, namely the
Energy-Environment-Economy Model at the Global
level (E3MG) has been used to confirm the scale and
importance of the macroeconomic rebound effect.
Modelling
The macroeconomic rebound effect arising from IEA
WEO 2006 (IEA 2006) energy-efficiency policies and
programmes is investigated here using E3MG, a
sectoral dynamic macroeconomic model of the global
economy, which has been designed to assess options
for climate and energy policies and to allow for
energy-environment-economy (E3) interactions
(Barker et al.2006; Barker 2008). The model contains
41 production sectors, which enables a more accurate
representation of the effects of policies than is
common in most macroeconomic modelling
approaches. The model addresses the issues of energy
security and climate stabilisation both in the medium
and long terms, with particular emphasis on dynam-
414 Energy Efficiency (2009) 2:411427
ics, uncertainty and the design and use of economic
instruments, such as emission allowance trading
schemes. E3MG is a non-equilibrium model with an
open structure such that labour, foreign exchange and
public financial markets are not necessarily closed. It
is very disaggregated, with 20 world regions, 12
energy carriers, 19 energy users, 28 energy technol-
ogies, 14 atmospheric emissions and 41 production
sectors, with comparable detail for the rest of the
economy. The model represents a novel long-term
economic modelling approach in the treatment of
technological change, since it is based on cross-
section and time-series data analysis of the global
system 1973-2002 (in the version used for this paper)
using formal econometric techniques, and thus pro-
vides a different perspective on stabilisation costs.
The model is based upon a Post Keynesian
economic view of the long-run. In other words, in
modelling long-run economic growth and technolog-
ical change we have adopted the historyapproach
1
of cumulative causation and demand-led growth
2
(Kaldor 1957, Kaldor 1972, Kaldor 1985; Setterfield
2002), focusing on gross investment (Scott 1989) and
trade (McCombie and Thirlwall 1994,2004), and
incorporating technological progress in gross invest-
ment enhanced by R&D expenditures. Other Post
Keynesian features of the model (see Holt 2007, for a
discussion of such features) include: varying returns
to scale (that are derived from estimation), non-
equilibrium, not assuming full employment, varying
degrees of competition, the feature that industries act
as social groups and not as a group of individual firms
(i.e. no optimisation is assumed but bounded ratio-
nality is implied), and the grouping of countries and
regions has been based on political criteria. At the
global level, accounting conventions are imposed so
that the expenditure components of GDP add up to
total GDP and total exports equal total imports at a
sectoral level allowing for imbalances in the data.
For the representation of the electricity generation
and supply sector E3MG incorporates a dynamic
bottom-up simulation submodel, the Energy Technolo-
gy Model (ETM), which implements a probabilistic
theory for the penetration of the energy technologies in
the market (Anderson and Winne 2004). The ETM
submodel is designed to account for the fact that a
large array of non-carbon options is emerging, though
their costs are generally high relative to those of fossil
fuels. However, costs are declining relatively with
innovation, R&D investment and learning-by-doing.
The ETM does not adopt a cost optimization technique
for modelling the electric system expansion and the
dispatch of the different technologies. But it combines
a detailed representation of their economic, technical
and environmental performance with historical data in
order to assess their capability to substitute away from
amarkertechnology. The implementation of differ-
ent policies through time, such as incentives, regula-
tion, and revenue recycling allow low or non-carbon
options to meet a larger part of global energy demand.
The process of substitution is also argued to be highly
non-linear, involving threshold effects. ETM includes
28 representative energy technologies, described by 21
technology characteristics, being less detailed than
bottom-up models such as the POLES (http://upmf-
grenoble.fr/iepe/Recherche/indexe.html), MARKAL
and TIMES (http://www.etsap.org/applicationGlobal.
asp). However, such energy-systems models typically
have no or limited representation of economy-wide
interactions unless they are used as part of an
integrated assessment model. These are captured in
E3MG through the interactions between the different
sectors in the model, with input-output and economet-
ric modelling allowing for complex interactions
between energy demand, output, investment, employ-
ment, incomes, consumption, trade, prices and wages,
without assuming that resources are used at full
economic efficiency.
1
This is in contrast to the mainstream equilibrium approach
(see DeCanio, 2003 for a critique) adopted in most economic
models of climate stabilisation costs. See (Weyant, 2004) for a
discussion of technological change in this approach. Setterfield
(1997) explicitly compares the approaches in modelling growth
and Barker et al. (2006) compares them in modelling
mitigation.
2
The theoretical basis of the approach is that economic growth
is demand-led and supply constrained. Growth is seen as a
macroeconomic phenomenon arising out of increasing returns
(Young, 1928), which engender technological change and
diffusion, and which proceeds unevenly and indefinitely unless
checked by imbalances. Clearly growth can increase only if
labour and other resources in the world economy can be utilised
in more productive ways, e.g. with new technologies and/or if
they are otherwise underemployed in subsistence agriculture or
unemployed. Palley (2003) discusses how long-run supply is
affected by actual growth. In contrast, the modern theory of
supply-side economic growth assumes full employment and
representative agents, and optimises an intergenerational social
welfare function (see Aghion and Howitt, 1998). It goes back to
Solow (1956,1957), with endogenous growth theory developed
by Romer (1986,1990).
Energy Efficiency (2009) 2:411427 415
For energy demand, a 2-level hierarchy is being
adopted. A set of aggregate demand equations on
annual data covering 19 fuel users/sectors and 20
regions is estimated and is then shared out among
main fuel types (coal, heavy fuel oil, natural gas and
electricity) assuming a hierarchy in fuel choice by
users: electricity first for premiumuse (e.g. lighting,
motive power), non-electric energy demand shared
out between coal, oil products and gas. The energy
demand for the rest of the 12 energy carriers is
estimated based on historical relations with the main 4
energy carriers. All energy demand equations use
co-integrating techniques, which allow the long-term
relationship to be identified in addition to the short-
term, dynamic one. A long-term behavioural relation-
ship is identified from the data and embedded into a
dynamic relationship allowing for short-term responses
and gradual adjustment (with estimated lags) to the
long-term outcome. The equations and identities are
solved iteratively for each year, assuming adaptive
expectations, until a consistent solution is obtained.
The economy aggregates, such as GDP, are found by
summation. This enables representation of the wider
macroeconomic impacts of policies focused on par-
ticular sectors, including rebound effects.
These long-run energy demand equations are of the
general form given in equation (1), where X is the
demand, Y is an indicator of activity, P represents
relative prices (relative to GDP deflators for energy),
TPI is the Technological Progress Indicator, the βare
parameters and the εerrors. TPI is measured by
accumulating past gross investment enhanced by
R&D expenditures (Lee et al. 1990) with declining
weights for older investment. The indicators are
included in many equations in the model, but only
those for energy are analysed here. All the variables
and parameters are defined for sector i and region j.
Xi;j¼bo;i;jþb1;i;jYi;jþb2;i;jPi;jþb3;i;jTPIðÞ
i;jþ"i;j
ð1Þ
In the equations, β
2,i,j
are restricted to be non-
positive, i.e. increases in prices reduce the demand
(for energy demand, see surveys in Atkinson and
Manning, 1995 and Graham and Glaister, 2002). In
the energy equations β
3,i,j
are estimated to be
negative, i.e. more TPI is associated with energy
saving. These parameters are constant across all
scenarios.
This approach is in contrast with the treatment of
energy users as representative agents in equilibrium
models. In our approach, each sector in each region is
assumed to follow a different pattern of behaviour
within an overall theoretical structure, implying that
the representative agent assumption is invalid (Barker
and De Ramon, 2005). This means that the behaviour
of each sector-region is not assumed to be the same as
that of the average of the group.
The original energy demand equations are based
on work by Barker et al. (1995) and Hunt and
Manning (1989). The work of Serletis (1992) and
Bentzen and Engsted (1993) has helped in the
cointegrating estimation. Since there are substitutable
inputs between fuels, the total energy demand in
relation to the output of the fuel-using industries is
likely to be more stable than the individual compo-
nents. This total energy demand is also subject to
considerable variation, which reflects both technical
progress in conservation, and changes in the cost of
energy relative to other inputs. Aggregate and
disaggregate energy-demand equationsspecifications
follow similar lines including economic activity,
technology, relative price effects, spending and R&D
investment and are in the process of being respecified
so as to also capture the temperature effect. As an
activity measure, gross output is chosen for most
sectors, but household energy demand is a function of
total consumers' expenditure. The long-run price
elasticity for road fuel is imposed at -0.7 for all
regions, following the research on long-run demand
(Franzén and Sterner 1995; Johansson and Schipper
1997, p. 289). The measures of research and
development expenditure and investment capture the
effect of new ways of decreasing energy demand
(energy-saving technical progress) and the elimination
of inefficient technologies, such as energy-saving
techniques replacing the old inefficient use of energy.
Table 1presents the weighted averages of short-
term and long-term activity and prices elasticities of
demand for aggregate energy, across global energy-
using sectors, with the world average added as the
final row. The equations are estimated from annual
data over the period 1973-2002 and year 2000
weights are used to find the averages. The equations
are estimated as specified above, with further details
in (Barker et al. 2006). In the projections after 2012,
these elasticities are modified to restrict outliers and
to allow for reductions in activity elasticities due to
416 Energy Efficiency (2009) 2:411427
saturation effects and higher responses to relative
prices via emission trading schemes and ad hoc
incentive schemes introduced to accelerate reductions
in energy use.
The modelling undertaken in this study required
the specification of scenarios to reflect the set of IEA
WEO 2006 (IEA 2006) energy-efficiency policies and
programmes for the Transport, Buildings and Industry
sectors of the economy for the period 2013-2030. The
estimated direct rebound effects on electricity and fuel
saving from no-regrets policies were derived from the
literature (Sorrell 2007; Sorrell et al. 2009; Schipper
and Grubb 2000). The investment and other costs to
governments, firms and individuals have been taken
from IEA WEO 2006 (IEA 2006). These estimates are
incorporated exogenously into the macroeconomic
modelling. A set of initial reductions in net energy
demand brought about by energy-efficiency policies
is disaggregated in terms of the models classifica-
tions and imposed on the selected final-demand, fuel-
using sectors with a proportional disaggregation of
the IEA WEO 2006 (IEA 2006) estimates. The effects
of the policies are calculated by comparing model
solutions 20132030 with and without the policies.
Scenarios are developed to allow the calculation of
macroeconomic rebound effects by modelling final
energy demand by 19 fuel-using sectors. The policy
case for the modelling includes implicitly the present
and committed energy-efficiency policies 20132030,
including key assumptions (oil price, and a carbon
price from the EU Emissions Trading Scheme (ETS)).
The fuel price assumptions for the reference case were
based on the ADAM projections from February 2008
(ADAM D-M2.1 2007), considering the outcomes
from the World Energy Technology Outlook 2050
report (http://ec.europa.eu/research/energy/pdf/weto-
h2_en.pdf) using the POLES model (http://upmf-
grenoble.fr/iepe/Recherche/indexe.html).
The methodology of the assessment was developed
in (Barker et al. 2007). The macroeconomic rebound
effect is the response of the economy in terms of
energy demand stimulated, through indirect and
economy-wide effects, following the initial energy
savings arising from energy-efficiency policies. In the
model, the initial effects are treated as exogenous,
from IEA WEO (IEA 2006) 2006 as energy savings
Short-term Long-term
activity relative price activity relative price
Power own use and transformation 0.389 0.113 0.604 0.178
Other energy own use and transformation 0.806 0.172 0.557 0.283
Iron and steel 0.241 0.288 0.457 0.493
Non-ferrous metals 0.420 0.101 0.489 0.480
Chemicals 0.497 0.205 0.569 0.362
Non-metallics nes 0.621 0.201 0.609 0.247
Ore-extra (non-energy) 0.418 0.092 0.683 0.202
Food, drink and tobacco 0.824 0.270 0.134 0.262
Textiles, clothing and footwear 0.429 0.163 0.435 0.267
Paper and pulp 0.215 0.246 0.429 0.221
Engineering, etc. 0.762 0.143 0.157 0.207
Other industry 0.506 0.142 0.618 0.387
Rail transport 0.870 0.311 0.754 0.253
Road transport 0.691 0.213 0.739 0.700
Air transport 0.509 0.128 0.402 0.405
Other transportation services 0.933 0.246 0.923 0.839
Households 0.478 0.244 0.648 0.318
Other final use 0.392 0.141 0.560 0.269
Non-energy use 0.122 0.168 0.001 0.226
World average for all sectors 0.506 0.178 0.591 0.338
Table 1 Weighted averages
(2000 weights) of the
estimated elasticities of
global aggregate energy
demand from the energy-use
equations
Source: E3MG 2.4 and
4CMR
Energy Efficiency (2009) 2:411427 417
and imposed sector by sector. The impacts spread
from the energy-using sectors throughout the rest of
the economy via the inputoutput structure of the
E3MG model to give the macroeconomic and indirect
effects. The total rebound effects are calculated by
taking the difference between the net energy saving
projected by the model, i.e. taking into account the
indirect and economy-wide effects throughout the
economy, and the expected gross energy savings
(after adding back the direct rebound effect) projected
as the effects of energy-efficiency policies by the IEA
in WEO 2006 (IEA 2006), with an additional
calculation (since this is not provided by the IEA
report) of the effects on power generation using
E3MG. This difference is then expressed as a
percentage of the expected gross energy saving from
these studies to give the total rebound effect. The
macroeconomic rebound effect is the difference
between the direct effect, also calculated as the
percentage of the expected gross energy saving, and
the total effect.
These definitions and identities can be expressed as
seven equations:
1. macroeconomic rebound effectindirect rebound
effect+economy-wide rebound effect
2. total rebound effectmacroeconomic rebound
effect+direct rebound effect
3. gross energy savings from IEA energy-efficiency
policiesnet energy savings (taken as exoge-
nous in E3MG)+direct rebound energy use
4. change in macroeconomic energy use from
energy-efficiency policies from E3MGenergy
use simulated from E3MG after the imposed
exogenous net energy savingsenergy use
simulated from E3MG before the imposed exog-
enous net energy savings
5. total rebound effect as %100 times change in
macroeconomic energy use from energy-efficiency
policies from E3MG/gross energy savings from
IEA energy-efficiency policies
6. direct rebound effect as %100 times direct
rebound energy use/gross energy savings from
IEA energy-efficiency policies
From 2, 5 and 6:
7. macroeconomic rebound effect as %total
rebound effect as %direct rebound effect as %
The effect of energy saving in production is to
reduce the costs of industrial energy use, so leading to
reductions in prices and increases in profits of the
industries working more efficiently. These lower
prices are then passed on to reduce costs for other
industries. The process gives rise to a rebound effect
in that the initial savings are (partially) offset by
increases in energy demand due to higher demands
for the exports and outputs of the industries that have
improved their energy efficiency and so reduced their
energy costs. The lower costs will also be passed on
to final consumers, depending on the price behaviour
of the industries. Consumers will substitute spending
towards the lower-priced products. Higher consumer
and labour demand will increase output (and GDP)
more generally and, hence, lead to higher energy
demand.
In the case of extra energy saving in the
residential buildings sector, the reduction in expen-
diture on fuels (assuming that fuel prices are
unchanged) implies an increase in the real income
of consumers. This effect is modelled by assuming
consumers initially maintain the level of energy
services received from the fuels, i.e. cut actual
spending to receive the same services; however, the
further response is more complicated. We assume
that they behave (1) as if fuel prices had fallen, so
that they substitute back towards fuels, depending on
their responses to lower effective prices and (2) as if
they had an increase in real income so that they
increase spending on energy and other activities,
depending on estimated income elasticities. For (2),
the saving ratio is changed so that real expenditures
rise by the appropriate amount. The higher consum-
ersexpenditure on all goods and services, especially
energy-intensive ones such as transport, then raise
energy use more generally.
Description of policies and scenarios
Policies
The majority of the assumed no-regrets options
appear to be aimed at incentivising energy-efficiency
improvements. It is the macroeconomic rebound
effect arising from all these energy-efficiency policy
measures that it is assessed in this paper.
418 Energy Efficiency (2009) 2:411427
Scenarios
The Reference Case is constructed to establish a
counterfactual history of the global economy for the
period 20132030 without the impact of the no-regrets
energy-efficiency policies implemented over this
period. It is a fully dynamic solution of the model over
the period, given the year-by-year profile of exogenous
variables such as population, exchange rates, interest
rates and fiscal policies in general. It includes the impact
of policy measures which are not explicitly targeted at
energy efficiency.
The Policy Case is an alternative fully dynamic
solution but including the sectoral effects on energy
use, year by year 20132030, of all the current and
committed IEA WEO 2006 (IEA 2006)energy-
efficiency policies for the transport, buildings and
industry sectors. The difference between the policy
case and the reference case, thus, gives a dynamic
estimate of the impact of these policies on the global
economy and will enable calculation of the amount by
which the original estimated energy saving of the
policies is reduced through the rebound effect.
Differences between the reference and policy
scenarios are used to assess the impacts of energy-
efficiency policies on energy demand and CO
2
emissions under the different scenario assumptions,
taking into account the macroeconomic effects esti-
mated using the model. By comparing these with the
imposed estimates for energy and CO
2
saving from
the earlier evaluation, which did not take the macro-
economic effects into account, estimates of the mag-
nitude of the macroeconomic rebound effect on
energy demand and CO
2
emissions are calculated.
Scenario assumptions
Within each scenario, the effects of the relevant policy
measures are introduced into the model on an annual
basis. This is done by including the projected direct
energy saving resulting from actions taken as a result of
that policy measure, taking into account any projected
direct rebound effect. The projected electricity and non-
electricity savings for OECD and non-OECD countries,
presented in Table 2, are used as assumptions in E3MG
model, so as to examine their macroeconomic and total
rebound effect. According to IEA WEO 2006 (IEA
2006), 1.1 Gtoe of energy (both electricity and non-
electricity) are projected up to 2030, where one third is
used in OECD countries and the other two thirds in
the non-OECD countries. The projected net energy
savings from the policies (net of the direct rebound
effect) are about 10% by 2030 for both groups of
countries. These savings come almost equally from the
three main examined sectors (Transport, Buildings, and
Industry). Almost two thirds of them concern savings
in fossil-fuel use, while the rest concern electricity
savings, with the exception of transport where all
savings come from fossil fuels, as the electricity use in
this sector is very small.
Table 2shows the projected direct net energy
savings by 2030 allowing for the direct rebound
effects discussed below. The table also converts these
projected savings to the percentage of the total
sectoral energy use and total sectoral emissions,
respectively, to enable an approximate comparison
of the strength of each policy. It should also be noted
that the paper focuses on the macroeconomic impli-
cations of energy-efficiency policies and measures for
final users of energy and does not provide an
evaluation of their likely effectiveness at a micro-
level. The effect of the policies on the power sector is
derived from the IEA final demand data using the
E3MG model and not from IEAWEO 2006 (IEA 2006)
or the synthesis tables in the IPCC AR4 so the results
for energy supply and CO
2
emissions can be regarded
as a check on the IPCC synthesis in Chapter 11.
The required cumulative investment for the post-
2012 period 20132030, presented in Table 3,isbased
on the IEA WEO (IEA 2006, pp. 197), as the relevant
information is not given explicitly in the IPCC AR4.
The direct rebound effects by sector in energy
terms are derived by applying the assumed rebound
percentages to the gross energy savings from the
sector. There are three broad sectors for which these
direct rebound effects have been empirically estimat-
ed in the evaluations of the IPCC AR4 energy-
efficiency policies and reviewed in (Sorrell 2007;
Sorrell et al. 2009). Given that we are applying them
to many different countries and time periods, we have
adopted a stylized treatment: the Transport sector are
assumed at 10%, the residential Buildings sector at
25% and other sectors at low or zero rebound effects.
In the case of energy-intensive processes, no direct
rebound effects are assumed, and the extra efficiency
mainly takes the form of lower unit costs and the
Energy Efficiency (2009) 2:411427 419
Table 2 Projected direct energy savings in 2030 for IEA WEO 2006 energy efficiency policies or measures used in this study as inputs to the modelling
Target sector Projected electricity
savings in 2030 (Mtoe)
% of total sectoral
electricity use in 2030
Projected non-electricity en
ergy savings in 2030 (Mtoe)
% of total sectoral non-
electricity use in 2030
Projected electricity and non-
electricity energy savings in
2030 (Mtoe)
% of total sectoral
energy use in 2030
Residential services
a
201 14.27 248 8.82 449 10.64
OECD 88 12.09 29 3.02 117 6.93
Non-OECD 113 16.59 219 11.83 332 13.11
Industry
b
95 10.11 242 8.09 337 8.57
OECD 33 9.40 58 5.57 91 6.53
Non-OECD 61 10.37 185 9.48 246 9.69
Transport
c
307 9.87 307 9.87
OECD 146 8.80 146 8.80
Non-OECD 162 11.16 162 11.16
Total 295 12.21 827 8.94 1,122 9.62
OECD 119 10.89 245 6.45 364 7.44
Non-OECD 174 13.16 584 10.72 758 11.19
a
Concerns policies/measures on heating, ventilation, air-conditioning, lighting, appliances, office appliances, hot water systems
b
Concerns policies/measures on motors, pumps, compressor systems, irrigation pumping systems
c
Concerns policies/measures on fuel economy, modal shift
Source: International Energy Agency 2006
420 Energy Efficiency (2009) 2:411427
rebound effects are the indirect effects of the lower
costs on sales to other industries and exports, which
we capture in the modelling. Low values (5%) for the
direct rebound are taken for services and other (e.g.
waste, agriculture and forestry) sectors. There are
good reasons for expecting the direct rebound effects
to be small or negligible for these sectors. In the case
of services buildings, indoor temperatures are both
conventionally and legally within acceptable ranges,
and these ranges seem unlikely to change in response
to energy-efficiency measures. In case of the other
sectors, estimated energy savings from no-regret
measures are negligible.
The assumptions used in the modelling for carbon,
oil, coal and gas prices are shown in Table 4.
The growth rates of GDP for the reference case are
shown in Table 5.
To implement the scenarios in E3MG, the effects
of the relevant energy-efficiency policy measures are
introduced into the model by imposing a reduction in
energy use on the estimated aggregate energy demand
equations for the sectors affected (using the projected
energy savings shown in Table 2).
Results
Macroeconomic effects of energy-efficiency policies
Table 6shows the macroeconomic effects of the total of
the energy-efficiency policies as modelled by E3MG
(by comparing the energy-efficiency policy case to
the reference case without policies). These effects
include the macroeconomic rebound effect, which is
Table 3 Projected cumulative investment costs in 20132030 for IEA WEO 2006 energy efficiency policies/measures used in this
study as inputs to the modelling
Target sector Cumulative investment costs
for electricity saving measures
in 20052030 (billion 2005US$)
Cumulative investment costs
for non-electricity saving
measures in 20052030
(billion 2005US$)
Total cumulative investment costs for
electricity and non-electricity saving
measures in 20052030 (billion 2005US$)
Residential services 758 168 926
OECD 546 76 622
Non-OECD 212 92 304
Industry 195 167 362
OECD 121 89 210
Non-OECD 74 78 152
Transport 1076 1076
OECD 661 661
Non-OECD 415 415
Total 953 1411 2364
OECD 667 826 1493
Non-OECD 286 585 871
Source: International Energy Agency 2006
Table 4 EU ETS allowance (Carbon price) and fuel price assumptions, reference case, 20052030
2005 2010 2015 2020 2025 2030
EU ETS allowance price (2005$/tC) 0 82.8 82.8 82.8 82.8 82.8
Crude oil 2005$/bbl 50.62 57.50 55.00 55.00 57.50 60.00
Gas 2005$/MMBTU 7.460 6.750 6.750 7.000 7.318 7.636
Coal 2005$/tonne 60.48 55.00 55.00 57.04 59.63 62.22
Sources: BERR-EWP (2007), BERR-ER (2006), IEA WEO 2007 and International Energy Agency 2006
Energy Efficiency (2009) 2:411427 421
distinguished in Calculation of macroeconomic
rebound effectbelow. Overall the policies lead to
a saving of about 4% of the energy which would
otherwise have been used by 2030 and a reduction in
CO
2
emissions of 5% (or 2.8GtCO
2
)by2030.The
table also shows the effects on GDP, the general
consumer price level and employment for 2020 and
2030. The energy saving shows up as macroeconomic
benefits in two main forms: firstly, lower prices (by
2030), as the production system requires fewer inputs
to produce the same output; and secondly, higher
output, partly the consequence of the lower inflation,
as households spend more in response to their higher
imputed income when their energy bills are reduced
for the same level of energy services provided. The
changes are relatively very small.
Impacts of energy-efficiency policies on energy
demand and CO
2
emissions
Final energy demand
Table 7shows the effect of energy-efficiency policies
on final energy demand only in energy units (mtoe),
grouped by six broad sectors of the economy, again
incorporating macroeconomic rebound effects. Over-
all, the reduction is about 600 mtoe, 4.3% of total
energy demand by 2030. The demand falls over the
period as the energy-efficiency policies gradually
strengthen and their effects accumulate. The table
shows the substantial differences between the sectors,
with Energy supply and Buildings showing the largest
reduction in absolute terms.
Figure 1shows the effects of energy-efficiency
policies on total final energy use for the global
economy 20102030, showing the net energy saving,
after the (exogenously estimated) direct rebound and
(calculated) indirect rebound effects are taken into
account. The figure shows the scale of these effects
and how they accumulate over the period. Figure 2
shows how the energy savings from the policies are
distributed across the main sectors in which they are
implemented.
Impacts on CO
2
emissions
The above reductions in final energy demand,
together with small reductions in own use of energy
in the power generation and other fuel sectors, arising
from energy-efficiency policies, lead to a reduction in
Table 6 Effects of energy-efficiency policies on key macroeconomic variables
Difference in levels 2010 2020 2030
World Final energy demand (%) 0 3.78 4.34
CO
2
emissions (%) 0 4.60 5.50
GDP (%) 0 0.21 0.28
Price index consumersexpenditure (%) 0 0.039 0.003
Employment (%) 0 0.28 0.20
Differences in levels are % difference from reference case. Final energy demand corresponds to Final Consumption, excl non-energy
use. CO
2
emissions refer to whole-economy CO
2
emissions from all anthropogenic sources. In this and subsequent tables, a positive
figure indicates an increase with respect to the reference case, and a negative figure a reduction with respect to the reference case, e.g.
a reduction in final energy demand due to energy efficiency policies is shown as a negative figure
Sources: E3MG 2.4 and 4CMR
Tab l e 5 Average annual growth of key macroeconomic
variables, reference case
20002010 20102020 20202030
OECD
GDP (% pa) 2.39 2.16 1.90
Non-OECD
GDP (% pa) 5.17 4.45 3.65
World
GDP (% pa) 2.97 2.74 2.41
This table shows projections chosen to correspond closely with
the actual outcome and represents a solution of the model
adopted for the study. The projections are not intended to be
forecasts.
Source: E3MG 2.4 and 4CMR.
422 Energy Efficiency (2009) 2:411427
CO
2
emissions. Note that in the E3MG model, CO
2
emissions are allocated at the point of emission so that
reductions in CO
2
emissions from power generation
reflects both reductions in final electricity demand and
reductions in own use of energy in power generation.
Table 8shows the effects of the energy-efficiency
policies on global anthropogenic CO
2
emissions,
grouped into power generation and the final-user
sectors. The contribution from power generation to the
overall reduction in CO
2
from the policies is substan-
tial, about one third of the total 2.8GtCO2 by 2030.
Calculation of macroeconomic rebound effect
Table 9shows the magnitude of the direct, macroeco-
nomic and total rebound effects on energy demand
arising from all energy-efficiency policies, disaggregated
by sector of the economy, with the assumed direct
effects. The effects are calculated by taking the differ-
ence between the energy saving projected by the model
and the expected gross energy saving (including the
direct rebound effect) projected from IEA WEO 2006
(IEA 2006) energy-engineering studies of the policies
(as set out in Table 2above). This difference is then
expressed as a percentage of the expected gross energy
saving from these studies. The macroeconomic results
show that the reduction in energy demand in 2030 is
around 50% less than expected due to several indirect
and economy-wide interactions discussed below, which
are not covered in the IEA WEO 2006 (IEA 2006)or
IPCC energy-engineering studies.
The highly disaggregated nature of the E3MG
model gives detailed insights into the indirect and
economy-wide interactions which give rise to the
macroeconomic rebound effects in addition to the
direct effects. Four potential sources of the total
rebound effects arising from the introduction of
energy-efficiency policies have been identified:
1. Direct rebound effects. These are comfort taking
for residential buildings and increased vehicle use
for transport and other effects as described above.
2. Lowering of energy use and industrial costs. The
lower energy costs for energy consumers enable
them to reallocate spending away from gas and
electricity to a wide range of other goods and ser-
vices, typically with very small energy and carbon
Table 7 Effect of energy policies on final energy demand by
sector difference in mtoe
World 2010 2020 2030
Energy supply industries 0 138.3 168.1
Transport 0 83.6 111.3
Residential/Services Buildings 0 120.9 166.2
Industry 0 108.1 138.2
Agriculture 0 4.7 5.3
Total 0 455.5 589.0
Figures are policy case less reference case. Final energy demand
corresponds to Final Consumption, excl non-energy use
Sources: E3MG 2.4 and 4CMR
Fig. 1 Effects of IEA WEO
2006 energy efficiency
policies on final energy
demand in the period
20002030
Energy Efficiency (2009) 2:411427 423
content. In transport, industry and services, the
targeted reductions in energy and carbon intensi-
ties lead to a reduction in industrial costs and,
therefore, prices and consequently more output
and exports.
3. Higher imputed incomes for private consumers.
The reduction in energy costs implies an increase
in consumer incomes. With the introduction of
tighter building regulations and other policies to
improve efficiency by the domestic sector, market
energy prices are largely unchanged, but gross
energy use falls if the volume of energy services
remains the same. The higher real incomes must
be imputed and allocated to consumers so that
they increase their spending, as if they had an
increase in actual income.
4. Higher investment directly associated with the
energy-efficiency policies. Examples are the cost
of extra insulation of houses or the extra cost of a
fuel-efficient car over another with similar char-
acteristics but lower efficiency. This extra invest-
ment, typically including the costs of the policies
to consumers and business associated with the
energy-efficiency measures, is added to industrial
investment, investment in office buildings and
dwellings and to the investment in road vehicles
by consumers.
Table 10 shows the relative contributions of the
three macroeconomic sources (items 2, 3 and 4
above) to the overall change in final energy demand,
CO
2
emissions, GDP and prices. The table shows that
the lowering of domestic and industrial energy costs
is the main source of reduced CO
2
emissions and a
major contributor to the reduction of prices. If
anything, the effect of the reduction in prices is an
underestimate because the model has a simple
treatment of cost inflation that does not allow for
economies of scale. The extra spending, due to higher
imputed income, leads to slightly higher energy use (a
rebound effect) and emissions and slightly higher
GDP and consumersexpenditure. This shows that
the increased economic activity due to changes in
consumer income mostly occurs in less energy-
intensive areas, i.e. use of energy and carbon is inelastic
to changes in consumer income. Similarly, the extra
investment stimulated by energy-efficiency policies is
itself concentrated on measures which reduce carbon
emissions, whilst increasing economic activity.
Table 10, thus, shows that nearly all the indirect
and economy-wide rebound effects on final energy
use (which are contained within the figure of 4.3%)
are due to the higher output resulting from greater
energy efficiency.
Table 8 Effect of energy efficiency policies on CO
2
emissions
by sector difference in Mt CO2-eq
World 2010 2020 2030
Energy supply industries 0 826.0 1,121.5
Transport 0 460.9 621.4
Residential/Services Buildings 0 348.5 504.6
Industry 0 437.4 583.8
Agriculture 0 0 0
Total 0 2,072.9 2,831.2
Figures are policy case less reference case. Total CO
2
emissions
include emissions from energy intensive industries' own use of
energy, rail transport and water transport
Sources: E3MG 2.4 and 4CMR
-700
-600
-500
-400
-300
-200
-100
0
2000 2010 2020 2030
Mtoe
Agriculture
Industry
Residential/Commercial Buildings
Transport
Energy supply industries
Fig. 2 Disaggregation of
net energy savings from
IEA WEO 2006 Energy
Efficiency policies, in
the period 20002030
424 Energy Efficiency (2009) 2:411427
The rebound effects we find are consistent with the
long-run parameters included in the aggregate energy
equations for the response of energy demand to
economic activity. All these activity elasticities are
below one in the projections to 2030. Energy demand
is, therefore, partly disengaged from activity in the
long run. The low responses are interpreted as the
outcome of several features in future energy use.
Firstly, the activities within each broad sector are
typically shifting over time towards more service-
based and less material-energy-based activities as
incomes rise and quality improves; energy demand
will grow more slowly than activities as a result.
Secondly, technological progress is taking the diffused
form of more control in production and distribution
and more precise use of energy in the form of
electricity rather than fossil fuels directly; aggregate
energy grows less, but the share of electricity rises.
Thirdly, much of energy use for heating and cooling of
buildings (residential and servicesuse of energy) is
largely an overhead cost once comfort levels are
reached; in consequence, energy use will be associated
more with employment and numbers of households
rather than with output and incomes. Employment and
numbers of households grow much less than GDP and
incomes.
Conclusions
We find that the total rebound effect arising from the
IEA WEO 2006 (IEA 2006) energy-efficiency poli-
cies for final energy users over the post-2012 period
20132030 is around 50% by 2030, averaged across
sectors of the economy. Given the large magnitude of
our estimated long-term rebound effects, a priority for
future research should focus on the effectiveness of
complementary policies such as broad-based energy
taxes, educational and other behavioural changes that
lock-infirst-order efficiency gains. There is also an
Table 10 Sources of macroeconomic effects of IEA WEO 2006 energy efficiency policies in 2030 % difference between policy case
and reference case
World Lower energy-use and
industrial costs
Higher imputed
income
Higher energy efficiency
investments
Total
Final energy 4.06 0.0002 0.30 4.34
CO
2
emissions 5.22 0.0001 0.29 5.50
GDP 0.007 0.0004 0.5 0.28
Price index consumersexpenditure 0.029 0.0001 0.026 0.003
The table shows contributions to % difference between policy case and reference case, from scenarios that decompose the total effects
into three components
Sources: E3MG 2.4 and 4CMR
Table 9 Direct, macroeconomic and total rebound effect of energy-efficiency policies (%), % difference between policy case and
reference case
World Direct Macroeconomic Total
2010 2020 2030 2010 2020 2030 2010 2020 2030
Energy supply industries 0 0 0 0 20.8 43.7 0 20.8 43.7
Transport 0 9.1 9.1 0 26.9 43.1 0 36.0 52.2
residential/services buildings 0 20.0 20.0 0 24.3 40.6 0 44.3 60.6
Industry 0 5 5 0 18.3 40.8 0 23.3 45.8
Agriculture 0 5 5 0 11.8 36.1 0 16.8 41.1
Total 0 9.4 9.7 0 22.1 41.6 0 31.5 51.3
Figures are total rebound effects, assumed direct rebound plus projected macroeconomic rebound effects
Sources: E3MG 2.4 and 4CMR
Energy Efficiency (2009) 2:411427 425
important role for the development of policies that are
not focused on saving energy alone but on portfolios
of policies that complement behavioural changes to
ensure reductions in GHG emissions as living stand-
ards improve. For example, a sensible portfolio of
policies for transport may combine (1) tighter engine
efficiency and GHG standards with (2) a switch of
fuel taxes to GHG taxes and (3) requirements that all
new cars and trucks have CO
2
metres visible to
drivers to provide real-time feedback on how driving
behaviour affects fuel use.
The macroeconomic rebound effects arise from the
reduction in energy costs for consumers and pro-
ducers (particularly for energy-intensive industries).
The lower energy costs for consumers lead them to
substitute away from oil, gas and electricity to a wide
range of other goods and services, typically with
relatively small energy and carbon content; hence, the
rebound effect is low. In industry, the targeted
reductions in energy and carbon intensities lead to a
reduction in their industrial costs and, therefore,
prices and consequently more output and exports.
Acknowledgements This paper has been prepared as a
contribution to the research of the UK Energy Research Centre
and the UK Tyndall Centre for Climate Change Research. The
authors are grateful for the support of the Centres and their
funding from the UK Research Councils.
References
Aghion, P., & Howitt, P. (1998). Endogenous Growth Theory.
Cambridge: MIT.
ADAM D-M2.1. (2007). Portfolio of policy and technological
options for P3a case study.
Anderson, D., & Winne, S. (2004). 'Modelling innovation and
threshold effects in climate change mitigation', Working
Paper No. 59, Tyndall Centre for Climate Change Research.
www.tyndall.ac.uk/publications/pub_list_2004.shtml.
Barker, T. (2008). The economics of dangerous climate
change. Editorial for the Special Issue of Climatic
Change on The Stern Review and its Critics.Climatic
Change, 89, 173194. doi:10.1007/s10584-008-9433-x.
Barker, T. S., Ekins, P., & Johnstone, N. (1995). Global
Warming and Energy Demand. London: Routledge.
Barker, T., Pan, H., Köhler, J., Warren, R., & Winne, S. (2006).
Decarbonizing the Global Economy with Induced Tech-
nological Change: Scenarios to 2100 using E3MG. In O.
Edenhofer, K. Lessmann, K. Kemfert, M. Grubb, & J.
Köhler (Eds.), Induced Technological Change: Exploring
its Implications for the Economics of Atmospheric Stabi-
lization Energy Journal Special Issue on the International
Model Comparison Project.
Barker, T., Ekins, P., & Foxon, T. (2007). The macroeconomic
rebound effect and the UK economy. Energy Policy, 35,
49354946. doi:10.1016/j.enpol.2007.04.009.
Bentzen, J., & Engsted, T. (1993). Short- and long-run elasticities
in energy demand: a cointegration approach. Energy Eco-
nomics, 15(1), 916. doi:10.1016/0140-9883(93)90037-R.
BERR ER (2006). Energy review. Overarching initial regulatory
impact assessment, Department for Business & Regulatory
Reform, http://www.berr.gov.uk/files/file32177.pdf.
BERR EWP (2007). Meeting the energy challenge. A white
paper on energy, Department for Business & Regulatory
Reform, http://www.berr.gov.uk/files/file39387.pdf.
Brookes, L. (1990). The Greenhouse Effect: Fallacies in the
energy efficiency solution. Energy Policy, 18, 199201.
doi:10.1016/0301-4215(90)90145-T.
DeCanio, S. (2003). Economic Models of Climate Change: A
Critique. New York: Palgrave-Macmillan.
Dimitropoulos, J. (2007). Energy productivity improvements
and the rebound effect: An overview of the state of
knowledge. Energy Policy, 35, 63546363. doi:10.1016/j.
enpol.2007.07.028.
Franzén, M., & Sterner, T. (1995). Long-run Demand Elastic-
ities for Gasoline. In T. Barker, N. Johnstone & P. Ekins
(Eds.), Global Warming and Energy Elasticities. London:
Routledge.
Geller, H., Harrington, P., Rosenfeld, A. H., Tanishimad, S., &
Unander, F. (2006). Polices for increasing energy effi-
ciency: Thirty years of experience in OECD countries.
Energy Policy, 34,556573. doi:10.1016/j.enpol.
2005.11.010.
Greening, L., Greene, D. L., & Difiglio, C. (2000). Energy
Efficiency and Consumption - The Rebound Effect - A
Survey. Energy Policy, 28, 389401. doi:10.1016/S0301-
4215(00)00021-5.
Grepperud, S., & Rasmussen, I. (2004). A general equilibrium
assessment of rebound effects. Energy Economics, 26,
261282. doi:10.1016/j.eneco.2003.11.003.
Grubb, M. (1990). Energy efficiency and economic fallacies.
Energy Policy,18, 783785. doi:10.1016/0301-4215(90)
90031-x.
Herring, H. (2004). The rebound effect and energy conserva-
tion. In C. Cleveland (Ed.), The Encyclopedia of Energy.
Academic Press/Elsevier Science.
Herring, H., & Sorrell, S. (2009). Energy efficiency and
sustainable consumption. The Rebound Effect, Macmillan
Publishers Limited.
Holt, R. (2007). What is Post Keynesian economics? In M.
Forstater, G. Mongiovi & S. Pressman (Eds.), Post
Keynesian macroeconomics. London: Routledge.
Hunt, L., & Manning, N. (1989). Energy price- and income-
elasticities of demand: some estimates for the UK using
the cointegration procedure. Scottish Journal of Political
Economy, 36(2), 183193. doi:10.1111/j.1467-9485.1989.
tb01085.x.
International Energy Agency (Ed.) (2005). The Experience with
Energy Efficiency Policies and Programmes in IEA
Countries. Paris: IEA.
International Energy Agency (Ed.) (2006). Wor l d E n e rg y
Outook 2006 (IEA WEO 2006). Paris: IEA
International Energy Agency (Ed.) (2007). Wor l d E n e rg y
Outook 2007 (IEA WEO 2007). Paris: IEA
426 Energy Efficiency (2009) 2:411427
IPCC AR4. (2007). IPCC Fourth Assessment Report,http://
www.ipcc.ch/.
Jevons, W. S. (1865/1905). The Coal Question: An Inquiry
Concerning the Progress of the Nation, and the Probable
Exhaustion of our Coal-mines. In A. W. Flux, & A. M.
Kelley (Eds.), 3rd Edition 1905. ed. New York.
Johansson, O., & Schipper, L. (1997). Measuring the long-run
fuel demand of cars. Journal of Transport Economics and
Policy, XXXI(3), 277292.
Kaldor, N. (1957). A model of economic growth. The
Economic Journal, 67, 591624. doi:10.2307/2227704.
Kaldor, N. (1972). The irrelevance of equilibrium economics. The
Economic Journal, 52,12371255. doi:10.2307/2231304.
Kaldor, N. (1985). Economics without Equilibrium. UK: Cardiff.
Khazzoom, J. D. (1980). Economic implications of mandated
efficiency in standards for household appliances. Energy
Journal, 1(4), 2140.
McCombie, J. M., & Thirlwall, A. P. (1994). Economic Growth and
the Balance of Payments Constraint.NewYork:StMartins.
McCombie, J. M., & Thirlwall, A. P. (2004). Essays on
Balance of Payments Constrained Growth: Theory and
Evidence. London: Routledge.
Palley, T. I. (2003). Pitfalls in the theory of growth: an
application to the balance of payments constrained growth
model. Review of Political Economy, 15(1), 7584.
doi:10.1080/09538250308441.
Romer, P. (1986). Increasing returns and long-run growth. The
Journal of Political Economy, 94(5), 10021037.
doi:10.1086/261420.
Romer, P. (1990). Endogenous technological change. The
Journal of Political Economy, 98(5), S71S102.
doi:10.1086/261725.
Saunders, H. (1992). The Khazzoom-Brookes postulate and
neoclassical growth. Energy Journal, 13, 131149.
Saunders, H. (2000). A view from the Macro Side: Rebound,
Backfire and Khazzoom-Brookes. Energy Policy, 28, 439
449. doi:10.1016/S0301-4215(00)00024-0.
Saunders, H. D. (2008). Fuel conserving (and using) production
function. Energy Economics, 30(5), 21842235.
doi:10.1016/j.eneco.2007.11.006.
Schipper, L., & Grubb, M. (2000). On the rebound? Feedback
between energy intensities and energy uses in IEA countries.
Energy Policy, 28,367388. doi:10.1016/S0301-4215(00)
00018-5.
Scott, M. (1989). A New View of Economic Growth. Oxford:
Clarendon.
Serletis, A. (1992). Unit root behaviour in energy future prices.
The Economic Journal, 13(2), 119128.
Setterfield, M. (ed). (2002). The Economics of Demand-led
GrowthChallenging the Supply-side Vision of the Long
Run. Cheltenham: Edward Elgar.
Small, K. A., & Van Dender, K. (2007). Fuel efficiency and
motor vehicle travel: the declining rebound effect. The
Energy Journal, 28(1), 2552.
Solow, R. (1956). A Contribution to the Theory of Economic
Growth. The Quarterly Journal of Economics, 70(1), 65
94. doi:10.2307/1884513.
Solow, R. (1957). Technical Change and the Aggregate
Production Function. The Review of Economics and
Statistics, 39, 312320. doi:10.2307/1926047.
Sorrell, S. (2007). The rebound effect: an assessment of the
evidence for economy-wide energy savings from im-
proved energy efficiency. London: UK Energy Research
Centre.
Sorrell, S., & Dimitropoulos, J. (2007). The rebound effect:
Microeconomic definitions, limitations and extensions.
Ecological Economics, 65, 636649. doi:10.1016/j.ecole
con.2007.08.013.
Sorrell, S., Dimitropoulos, J., & Sommerville, M. (2009).
Empirical estimates of the direct rebound effect: A review.
Energy Policy, 37(4), 13561371.
Treasury, H. M. (2006). Stern Review on the Economics of
Climate Change. London: HM Treasury.
US DfT. (2008). Preliminary regulatory impact analysis:
corporate average fuel economy for my 20112015
passenger cars and light trucks. Washington, DC: Depart-
ment of Transportation, National Highway Safety Admin-
istration, Office of Regulatory Analysis and Evaluation,
National Centre for Statistics and Analysis.
Vikström, P. (2004). Energy efficiency and energy demand: A
historical CGE Investigation on the rebound effect in the
Swedish economy 1957. Umeå: Umeå University.
Wei, T. (2006). Impact of energy efficiency gains on output and
energy use with CobbDouglas production function.
Energy Policy, 35(4), 20232030.
Weyant, J. P. (2004). Introduction and overview: energy
economics special issue EMF 19 study Technology and
Global Climate Change Policies. Energy Economics, 26,
501515. doi:10.1016/j.eneco.2004.04.019.
Young, A. (1928). Increasing returns and economic progress.
The Economic Journal, 38(152), 527542. doi:10.2307/
2224097.
Energy Efficiency (2009) 2:411427 427
... According to Hertwich (2005), the rebound effect is characterized as a behavioral or systemic reaction to efforts aimed at mitigating environmental impacts, which can ultimately offset the intended outcomes of these efforts. This intricate phenomenon suggests that the anticipated energy savings resulting from advancements in energy efficiency may face the challenge of being partially or entirely counteracted by simultaneous upticks in energy consumption, a notion underscored by Barker et al. (2009). ...
... The blue cluster represents basic conceptual studies (e.g., Barker et al., 2009;Berkhout et al., 2000;Binswanger, 2001;Brookes, 1990;Dimitropoulos, 2007;Grepperud & Rasmussen, 2004;Roy, 2000;Saunders, 1992;Schipper & Grubb, 2000;Wei, 2010). In this cluster, pioneering studies that attempt to create a conceptual framework for the rebound effect stand out. ...
Article
Full-text available
The energy rebound effect, characterized by increasing energy use despite enhanced efficiency, has emerged as a complex interdisciplinary topic in the literature, reflecting its significance in both production and consumption sustainability. Recognizing the pivotal role of academic research in shaping energy management strategies and policies, this study conducted a comprehensive bibliometric review, accessing 530 articles from major citation indexes. This analysis provides insights into the evolution of the energy rebound field and potential future directions. Understanding the existing literature on the rebound effect serves as a valuable foundation for forthcoming research endeavors, addressing the intricate challenge of reconciling energy efficiency with sustainable energy consumption and production. The review explores the extensive literature on the energy rebound effect, highlighting a predominant focus on residential (e.g., cooling, heating, lighting, transportation) and industrial energy consumption in the majority of studies. Empirical applications primarily emphasize the measurement of the energy rebound effect, with a notable trend of increasing research concentration in China in recent years. Additionally, there is a growing body of economy‐wide analyses in this field. However, the literature exhibits complexity in both the analyses conducted and the methods employed for measuring the rebound effect. Research gaps are apparent, numerous studies tend to concentrate solely on the rebound effect at the overall energy consumption level, neglecting variations in energy types and the behavioral patterns of economic actors. Furthermore, despite consistent findings of a higher rebound effect in developing nations, there is a noticeable lack of scholarly literature addressing this issue about these countries. This article is categorized under: Sustainable Energy > Energy Efficiency Human and Social Dimensions > Behavioral Science Policy and Economics > Governance and Regulation
... Nevertheless, an interesting, reducing benefits effect called as Circular Economy Rebound (CER) was identified, which may be striking in some circumstances (Zink & Geyer, 2017). The original, more thoroughly researched 'classic' rebound effect typically occurs when beginning boosts in production/consumption efficiency are lost out due to absolute increases in production/consumption (Barker et al., 2009;Brookes, 1990;Jevons & Flux, 1965) also recognised as Jevons' Paradox (Siderius & Poldner, 2021). The concept has also been named as the 'backfire' (Broberg et al., 2015). ...
... The concept has also been named as the 'backfire' (Broberg et al., 2015). The total effect incorporates the direct effect and the influence of (1) the lower energy price on energy demand in the three broad sectors as well as of (2) the extra consumers' expenditure from higher (implicit) real income, and (3) the extra energy-efficiency investments (Barker et al., 2009). The available data on the circular economy with the trend is presented in Table 2. ...
Article
Full-text available
The paper discussed both the benefits and limitations of the circular economy. Furthermore, we discuss that this concept is ambiguously defined.. A circular economy seems to be a satisfactory remedy for expanding environmental pollution and waste overproduction The benefits of the circular economy are believed to ever exist. Conversely, some premises to believe that rebound effects may offset such social benefits have merged in current literature to date by increasing overall consumption or “growing the pie”. We want to check if the circular economy positively contributes to the business model. Therefore, we applied the method of systematic literature review. This paper's value is important since it argues that we must integrate circularity across supply chains rather than limiting them to waste management. Moreover, indicators of the Circular Economy are not yet commonly agreed upon, and data are limited to waste. The paper recommends for further review of the circular concept.
... Environmentally friendly investments could increase output by the income multiplier and increase environmental efficiency, but they could lead to a net increase in pollution (Rezai et al., 2013). According to Barker et al. (2009) As described in the previous section, green government expenditure as an autonomous component of aggregate demand shifts the AD curve from AD to AD' in Figure 3.a, so the economy moves from point 1 to point 2. By stimulating green private investment, green public spending improves the economic system's innovative capacity and environmental and economic performance. As a result, enhanced environmental efficiency reduces the slope of the H curve in Figure 3.b. ...
Article
Full-text available
The ecological transition hinges on ecological structural change, necessitating smart macroeconomic policies. Incorporating ecological considerations into macroeconomic modelling constitutes a pivotal area of inquiry that examines the nexus between environmental sustainability and economic expansion. This article proposes a macroeconomic modelling to establish a foundational analytical framework to address this challenge. Incorporating ecological considerations into macroeconomic modelling constitutes a pivotal area of inquiry that examines the nexus between environmental sustainability and economic expansion. Contributing to the post-Keynesian Ecological Macroeconomics field, this paper proposes an eco-Keynesian cross due to its simplicity and widespreadness for teaching macroeconomics to illustrate the macroeconomic perspective of green targets strategies in a closed economy. Green public and private investments have the potential to enhance output through the multiplier effect and improve environmental efficiency. However, these investments may also induce rebound effects, resulting in a net increase in pollution intensity within the economy. The model underscores the potential trade-off between social and environmental objectives, which an ecological perspective seeks to surmount by advocating for sustainable and inclusive industrial development.
... Wei and Liu (2017) adopted a CGE model to estimate the economy-wide RE, ultimately finding a rebound of 70 % for energy use and 90 % for related emissions in 2040, albeit with sectoral and local disparities. Similarly, Barker et al. (2009) utilised a Post Keynesian non-equilibrium model to again study the economy-wide RE arising from the IEA WEO 2006 energy-efficiency policies for final energy users. The authors arrive at a total RE of 31 % by 2020, rising to 52 % by 2030. ...
... Rebound-Effekte können auch auf einer aggregierten oder "gesamtwirtschaftlichen" Ebene untersucht werden (Barker et al. 2009 Seebauer (2018) fanden psychologische Treiber für Rebound-Effekte, die sich nicht mit ökonomischen Begriffen erklären lassen. Ein Beispiel ist die sogenannte "moralische Lizenzierung" (Dütschke et al. 2018), bei der Verbraucher:innen beispielsweise finden, der Umwelt durch den Kauf energieeffizienterer Beleuchtung einen guten Dienst erwiesen zu haben, und daher berechtigt sind, mehr Lampen länger brennen zu lassen. ...
Article
This study empirically examines the effect of environmental innovation on ecological footprint using a panel smooth transition regression (PSTR) model for 33 OECD countries, over the period 1995–2019. The findings indicate that a rebound effect occurs when the level of environmental innovation surpasses the calculated threshold level (14.254%), suggesting that environmental innovations exert pressure on the environment. Existing empirical results point to a rebound effect, which refers to offsetting the gains from environmental innovation by channeling them into production and consumption.
Chapter
In this chapter, the Jevons paradox is studied in the context of the debate on the limits to Growth. This “Jevons paradox” is part of a more general criticism of William Stanley Jevons to classical economics. For Jevons, when the cost of production declines due to resource efficiency, the marginal utility of commodities that use the given resource declines, increasing directly the consumption of those commodities and indirectly the consumption of other commodities with which they are exchanged. Then, scientific progress and resource efficiency is not a good path to the lesser use of resources. Actually, as coal is a non-renewable energy resource, it may be depleted. Jevons’ line of thought led to new areas in economics that imply that economics cannot be fully split from other sciences. The chapter assesses the current importance of the Jevons paradox in the macroeconomic and the microeconomic level, looking at the relationship between economic growth and energy efficiency. Finally, the chapter comments on the energy policies proposed to avoid the rebound effect, with some concluding remarks on the evolution of the concept.
Chapter
The cycling of materials in natural and semi-natural ecosystems is complex and should only be taken as a model of materials use in human societies with recognition of these complexities. Whilst overall common features can be discerned, in detail there is large variability. This variability is in respect of the relative size of nutrient compartments from ecosystem to ecosystem, the rate of flow between compartments, differences from one element to another and between species of a single element, the degree of overall loss from systems (leakiness under equilibrium and disturbance conditions) and the equilibrium — time interaction of the ecosystem. Analysis of such differences has equal validity in drawing lessons for materials use in human societies as the recognition of the broader generalisations
Article
There are many theories concerning the relationship between the pace of development and economic growth in different countries. This impressive collection explores the relationship between a country's balance of payments and their rate of economic growth. The evidence and ramifications within the book will be of interest to economists and policy-makers in developing countries. © 2004 J. S. L. McCombie and A. P. Thirlwall; individual contributors