Empirical estimates of the direct rebound effect: A review
Introduction
Improvements in energy efficiency make energy services cheaper, and therefore encourage increased consumption of those services. This so-called direct rebound effect offsets the energy savings that may otherwise be achieved. For example, consumers may choose to drive further and/or more often following the purchase of a fuel-efficient car because the operating cost per kilometre has fallen. Similarly, consumers may choose to heat their homes for longer periods and/or to a higher temperature following the installation of loft insulation, because the operating cost per square metre has fallen. The extent to which this occurs may be expected to vary widely from one energy service to another, from one circumstance to another and from one time period to another. But any increase in energy service consumption will reduce the ‘energy savings’ achieved by the energy efficiency improvement. In some circumstances, it could offset those savings altogether—an outcome that has been termed ‘backfire’.
Direct rebound effects are the most familiar and widely studied component of the overall or economy-wide rebound effect (Sorrell, 2007) which also involves various indirect effects (for example, the energy associated with other goods and services whose consumption has increased as a result of the energy efficiency improvement). Beginning with Khazzoom (1980), there have been a series of estimates of the direct rebound effect for different energy services (Greening and Greene, 1998). These studies are extremely diverse in terms of the definitions, methodological approaches and data sources used. Also, despite growing research activity, the evidence remains sparse, inconsistent and largely confined to a limited number of consumer energy services in the United States—notably personal automotive transport and household heating. The main reason for this is the lack of suitable data sources for other types of energy service in other sectors and countries. In addition, interpretation of the evidence is greatly hampered by the use of competing definitions, measures, terminology and notation. Many studies do not mention the direct rebound effect at all, but nevertheless provide elasticity estimates that may, under certain assumptions, be used as proxy measures of that effect. Taken together, these features inhibit understanding of the direct rebound effect and the appropriate methodological approach to estimating its magnitude in different circumstances, as well as making it difficult to identify the relevance of particular studies.
This paper provides an overview of the methodological approaches to estimating direct rebound effects and reviews the evidence that is currently available. It updates an earlier review by Greening et al. (2000) and seeks to clarify a number of issues that were raised therein. The underlying research is reported in detail in Sommerville and Sorrell (2007) and Sorrell and Dimitropoulos (2007a). The paper focuses entirely on energy services in the household sector, since this is where practically all of the research has been undertaken. As a result, the conclusions do not provide guidance on the magnitude of direct rebound effects in other sectors, nor on the economy-wide rebound effect, which is fully discussed by Sorrell (2007) and Sorrell and Dimitropoulos (2007c).
Section 1 describes the operation of the direct rebound effect, highlighting some key issues concerning the measurement of this effect and the conditions under which it may be expected to be larger or smaller. 2 Understanding the direct rebound effect, 3 The quasi-experimental approach describe the ‘quasi-experimental’ and ‘econometric’ approaches to estimating direct rebound effects and the methodological challenges associated with each. 4 The econometric approach, 5 Estimates for personal transport, 6 Estimates for household heating summarise the results of a number of studies that use these approaches to estimate direct rebound effects for personal automotive transport, household heating and a limited number of other household energy services. More information on the studies reviewed is provided in the Appendix A. Section 7 discusses a number of potential sources of bias with econometric estimates that may lead the direct rebound effect to be overestimated. Section 8 concludes.
Section snippets
Understanding the direct rebound effect
Energy services such as heating and lighting are provided through energy systems that involve particular combinations of capital equipment, labour, materials and marketable energy commodities such as electricity. The relevant systems may include primary conversion equipment such as boilers, secondary conversion equipment such as radiators, equipment for distributing energy and manual or electronic controls. For space heating and lighting the relevant energy systems may also include building
The quasi-experimental approach
One approach to estimating direct rebound effects relies upon measuring the demand for the energy service before and after an energy efficiency improvement: for example, measuring the change in heat output following the installation of a fuel-efficient boiler. The demand for the energy service before the energy efficiency improvement could be taken as an estimate for what demand ‘would have been’ in the absence of the improvement. However, various other factors may also have changed the demand
The econometric approach
A more common approach to estimating direct rebound effects is through the econometric analysis of secondary data sources that include information on the demand for energy, the relevant energy service and/or the energy efficiency of that service. This data can take a number of forms (e.g. cross-sectional, time-series, panel) and apply to different levels of aggregation (e.g. household, region, country). Such studies typically estimate elasticities, meaning the percentage change in one variable
Estimates for personal transport
By far the best studied area for the direct rebound effect is personal automotive transport. Most studies refer to the US, which is important since fuel prices, fuel efficiencies and residential densities are lower than in Europe, car ownership levels are higher and there is less scope for switching to alternative transport modes.
In principle, estimates of the own-price elasticity of gasoline consumption for personal transport should provide an upper bound for the direct rebound effect
Estimates for household heating
The next best studied area for direct rebound effects is household heating. Table A9 summarises the main features and results of 12 quasi-experimental studies of household heating, including the reviews by Nadel (1993) and Milne and Boardman (2000). There is a larger ‘grey’ literature on this subject—including evaluations of energy-efficiency programmes by individual utilities—but this is relatively inaccessible.
Taken together, the reviewed studies suggest that standard engineering models may
Estimates for other household services
There are relatively few estimates of the direct rebound effect for other household energy services, owing largely to lack of data. Nadel (1993) reports the results of a number of evaluation studies by US utilities, which suggest direct rebound effects of 10% or less for lighting and approximately zero for water heating, with inconclusive results for refrigeration. We were not able to access these studies, which appear to be small-scale, short-term and methodologically weak. Instead, Table A7,
Potential sources of bias
Most estimates of the direct rebound effect assume that the change in demand following a change in energy prices is equal to that following a change in energy efficiency, but opposite in sign. Most studies also assume that any change in energy efficiency derives solely from outside the model (i.e. energy efficiency is ‘exogenous’). In practice, both of these assumptions may be incorrect.
First, while changes in energy prices are generally not correlated with changes in other input costs, changes
Summary
In summary, the accurate estimation of direct rebound effects is far from straightforward. A pre-requisite is adequate data on energy consumption, energy services and/or energy efficiency which is only available for a subset of energy services. As a consequence, the evidence remains sparse, inconsistent and methodologically diverse, as well as being largely confined to a limited number of consumer energy services in the OECD. It is important to recognise that estimates of the direct rebound
Acknowledgements
This paper is based upon a comprehensive review of the evidence for rebound effects, conducted by the UK Energy Research Centre (Sorrell, 2007). An earlier version of the results is contained in Sorrell (2008). The financial support of the UK Research Councils is gratefully acknowledged. The authors are grateful for the advice and comments received from Manuel Frondel, Karsten Neuhoff, Jake Chapman, Nick Eyre, Blake Alcott, Horace Herring, Paolo Agnolucci, Jim Skea, Rob Gross, Phil Heptonstall
References (87)
Technological progress and sustainable development: what about the rebound effect?
Ecological Economics
(2001)- et al.
Analyzing gasoline demand elasticities—a survey
Energy Economics
(1991) - et al.
The imperfect price irreversibility of non-transportation of all demand in the OECD
Energy Economics
(1995) - et al.
Temperature takeback in the hood river conservation project
Energy and Buildings
(1989) Gasoline demand revisited: an international meta-analysis of elasticities
Energy Economics
(1998)- et al.
Evaluating environmental programs: the perspective of modern evaluation research
Ecological Economics
(2005) - et al.
Residential energy demand in OECD-countries and the role of irreversible efficiency improvements
Energy Economics
(1998) - et al.
The impact of consumer behavior on residential energy demand for space heating
Energy and Buildings
(1998) - et al.
Iindoor temperature changes in retrofit homes
Energy
(1985) - et al.
Effects of the Hood river conservation project on electricity use
Energy and Buildings
(1989)