A Free-Market Energy Blog

Rebound Redux: IEA Gets Energy Efficiency Wrong

By Michael Shellenberger -- April 29, 2013

“The notion that energy efficiency measures might bring deep emissions reductions along with greater economic growth has long promised an attractive ‘win-win’ for policy makers and business leaders. The extensive literature on rebound presents an uncomfortable challenge to this view, undermining the idea that efficiency leads to decreased energy use on anything approaching a one-to-one basis.”

The International Energy Agency’s World Energy Outlook 2013 grossly overestimates how much energy efficiency can reduce greenhouse gas emissions, according to scholars cited in this very study. As such, IEA’s claim that demand-side measures can account for almost three-quarters of emission reductions by 2020 could result in global-warming mitigation efforts that overinvest in efficiency and underinvest in low- and zero-carbon energy technologies.

The Paris-based agency’s projections of emissions reductions from energy conservation and efficiency are based on questionable assumptions, experts said, pointing out that rebound effects frequently take back much or all of the energy savings that efficiency policies attempt to capture.

Though the “rebound effect” is little known to policy makers, researchers around the world have published well over 100 peer-reviewed articles documenting empirically that energy savings from increased efficiency of factories, cars, and houses can result in rebounding energy consumption. Depending on the sector in question and the efficiency policies that are implemented, estimates of rebound range from just 5 percent to well over 100 percent.

Yet in its World Energy Outlook the IEA claims that total rebound from all energy efficiency policies is just 9 percent overall — without accounting for differences across sectors or economies. Experts on rebound said a rate of 9 percent would be an underestimate for developed economies and would have little to do with developing economies, where rebound is known to be far more significant.

The IEA report cites two leading rebound researchers to justify its rebound estimate, Karen Turner of the University of Strathclyde, Scotland, and Steve Sorrell of the University of Sussex. But both Turner and Sorrell rejected the IEA’s 9 percent figure and expressed confusion as to how it had been derived.

“I don’t know how they can say that,” Turner told the Breakthrough Institute. “A 9 percent average seems too low given what we know about macroeconomic rebound effects,” said Sorrell, “as well as the much higher rebound effects in developing countries.”

The IEA’s downplaying of rebound also clashes with a 2011 European Commission report that found that rebound effects are significant, as well as two new reports release this month, by the Asian Development Bank and the Germany-based International Risk Governance Council, which both called for greater attention to the issue.

What’s at Stake

If the IEA’s analysis of rebound is as flawed as experts suggest, it will have major implications for policy makers who have to make tough decisions about how to meet growing energy demand while combating global warming.

With a global rebound rate of 9 percent incorporated into their climate stabilization scenario, which projects carbon dioxide at 450 parts per million, the IEA estimates that various energy efficiency measures will reduce carbon dioxide by roughly 2.3 billion tons per year by 2020, and 7.1 billion tons annually by 2035.

The best available research, however, suggests that rebound is far more significant.

Using macroeconomic models, Terry Barker, an economist and director of the Cambridge Centre for Climate Change Mitigation Research, estimated the total global rebound effect — comprising direct, indirect, and economy-wide rebound — to be 31 percent by 2020, rising to 52 percent by 2030, if all IEA efficiency recommendations are implemented.

In other words, the Barker study shows that more than half of the carbon dioxide savings that the IEA projects will result from efficiency policies would be negated unless increased energy consumption is met by zero-carbon sources. A rebound rate of 52 percent would eliminate the IEA’s projected emissions reductions from efficiency by about 3.2 billion tons annually by 2035 — it would require the construction of about 530 one-gigawatt nuclear power plants to offset this amount of emissions.

Not taking into account high levels of rebound thus results in underestimates of just how much low carbon power will be required to meet global development needs and reduce emissions.

Obscuring the Energy Challenge

The IEA report ignores the major reviews of rebound, including the largest one ever conducted, led by Sorrell for the United Kingdom Energy Research Centre in 2007. It also does not mention the 2011 European Commission review, which concluded that “the evidence does not support views that rebound effects are too small to be significant.”

The IEA does, however, cite a white paper by an energy efficiency advocacy group, the American Council for Energy Efficient Economies. The ACEEE paper includes a highly selective review of the research and concludes that direct and indirect rebound “can be combined to estimate total rebound of about 20 percent.” Yet even that rate is more than double the one provided by the IEA.

In our own review of the literature, the Breakthrough Institute found that direct rebound for consumer energy services — such as indoor lighting, appliances, or more efficient cars — ranges from about 10 to 30 percent, a finding confirmed in the European Commission report. Indirect re-spending effects and macro-economic rebound from those efficiency savings, however, can be much higher. Re-spending effects alone can add a further 5 to 25 percent on top of direct rebound effects, driving the total rebound effect above 50 percent.

In productive sectors, such as electric utilities, construction, agriculture, manufacturing, and communications, rebound is even more significant. In a major study of 30 such sectors in the United States, using the most accurate available data, Harry Saunders found an average direct rebound rate of 62 percent over a 20-year time period, 1980-2000.

Similarly, the European Commission found direct rebound of 30 to 80 percent for fuel efficiency in commercial transport.

Moreover, a number of modeling studies reviewed by the European Commission report as well as the Breakthrough Institute found macroeconomic rebounds of 30 to 50 percent or greater among a range of national economies, including some exceeding 100 percent.

All these findings suggest that the IEA’s estimate of a 9 percent negation of savings from efficiency by 2035 due to rebound is well off the mark even in developed economies. As a global rate, 9 percent appears to be even less defensible.

In a report released this month, “Asia’s Energy Challenge,” the Asian Development Bank acknowledges that “a large and growing body of research has shown that energy efficiency gains do not bring one-for-one reductions in energy consumption because of the so-called ‘rebound effect,’” which, it notes, is more significant in developing economies. By reducing the effective cost of energy, efficiency allows individuals and firms to consume more — a clear benefit to the 2.8 billion Asians who currently live without access to modern electricity, but also a challenge to the view that efficiency will be as major a source of emissions reductions.

The ADB cites a study last year by Li Li and Han Yonglei, who found that rebound in China from 1997 to 2008 was high, reaching levels known as “backfire” in three of those years, meaning that efficiency gains led to increased, rather than decreased, overall energy consumption. Two other scholars cited by the ADB, Boqiang Lin and Xia Liu, found overall rebound in the Chinese passenger transportation sector of 107 percent between 1994 and 2010 — also suggesting backfire.

Similarly, a report released this week by the Germany-based International Risk Governance Council acknowledged the fact that rebound tends to be far higher in developing countries, concluding: “there is a clear and urgent need for studies of rebound in developing economies. The magnitude of rebound that is likely to be observed in these studies will probably be much larger in some cases than the analogous values observed in the industrialized world.”

IEA Misrepresents Rebound Research

In response to a query from the Breakthrough Institute, IEA staff explained that their low rebound estimate was based on the assumption that most studies of direct rebound have examined rebound in the personal transportation and residential sectors. IEA’s modelers assume that rebound is lower in industrial and service sectors.

But studies of rebound in the production sectors of the economy find exactly the opposite. Direct rebound is higher in production sectors of the economy than in end-use sectors such as residential and personal transportation sectors. This is particularly true of energy intensive productive sectors such as electric utilities, chemicals production, and agriculture, as demonstrated by Saunders’ research.

Even in end-use sectors rebound effects are far higher in developing countries, where demand for energy services is much less saturated than in developed economies. Pioneering research in India by Joyashree Roy found that efforts to replace diesel generators with more efficient solar lanterns resulted in rebounding energy use, with energy-hungry consumers supplementing, not replacing, their diesel generators.

Roy agreed that the IEA rebound number significantly understates the rebound effect. “It would be very difficult to say that rebound is just 9 percent. That’s not a representative value,” said Roy, a professor of economics at Jadavpur University in Kolkata, India. “If you really want to tell the truth you need to incorporate rebound and show how hard it is going to be to achieve sweeping energy savings from efficiency improvements.”

A Seductive Message

The notion that energy efficiency measures might bring deep emissions reductions along with greater economic growth has long promised an attractive “win-win” for policy makers and business leaders. The extensive literature on rebound presents an uncomfortable challenge to this view, undermining the idea that efficiency leads to decreased energy use on anything approaching a one-to-one basis.

“The argument that you can boost the economy, reduce energy consumption, reduce emissions, and boost welfare all at the same time is very seductive,” said University of Sussex’s Sorrell. “If you start raising questions about whether that is really the case, it doesn’t go down too well.”

Turner, of the University of Strathclyde in Scotland, went further:

IEA is clearly pushing for energy efficiency, and is making the argument that energy efficiency is leading to economic growth.” But, she added, “introducing efficiency improvements is not a free lunch. I think policy makers are looking at energy efficiency as double dividend (economic benefits coupled with environmental benefits, or decoupling of economic growth and environmental damage). You can see why it’s appealing.

All of the researchers offered ways for the IEA to improve its modeling.

“I don’t think a global across the board figure is appropriate,” Sorrell said. “You should try to disaggregate the assumption as far as the evidence permits,” he added, pointing to the Barker study, which assumed a certain level of magnitude for direct rebound and then modeled various indirect effects.

“It’s hard to argue about a number, given that the evidence is so mixed,” Sorrel said. “But the best way to deal with this uncertainty is to emphasize it and to conduct sensitivity tests, rather than just picking a low number with limited justification.”


Michael Shellenberger is President and cofounder of the Breakthrough Institute of Oakland, California. This commentary, a staff effort, was originally posted at the Breakthrough blog.


  1. Ed Reid  

    People with the freedom to make their own decisions, whether in their personal lives or as owners / managers / employees of businesses, will make decisions regarding their activities, including their use of energy, which they perceive to be in their best interests. Therefore, if an increase in the efficiency of energy use results in the net availability of increased discretionary funds, those funds will be used to advance the interests of those who have increased efficiency, which frequently will not align with the interests of those interested in reducing energy consumption.

    However, it is important to recognize that increased energy efficiency typically requires investment in new, more sophisticated and more expensive facilities and equipment, which may also be both more expensive to install and more expensive to maintain than the less efficient facilities and equipment they replace. Rebound would tend to be limited by the net economic benefit achieved in each individual case.

    Rebound, as noted above, tends to be greater in developing societies, in which the increased efficiency permits the reduction of deprivation. Rebound also tends to be greater in situations in which voluntary or involuntary conservation has been in effect. For example, if an individual has been maintaining the temperature in his or her home at lower than desired temperatures during the winter to reduce operating costs, the installation of a more efficient furnace would be more likely to lead to rebound if the desired temperature could be maintained at the same or lower cost. However, if an individual has been maintaining his or her home at the desired temperature with the less efficient furnace, the installation of a more efficient furnace is less likely to result in a direct rebound, though other uses of the discretionary funds might also result in an indirect rebound.

    Rebound can be minimized or avoided through coercion; or, through the imposition of uneconomic efficiency mandates which compel actions whose costs cannot be recovered through energy cost savings during their useful lives. Neither of these is a particularly attractive option.


  2. archaeopteryx  

    Hate to say it, but the IEA is only a notch more serious on energy matters than Greenpeace, the WWF, the IPCC etc. Behind a veneer of “energy related matters” lies a Green (or Euro-Green) Propaganda machine. They may be lobbying for the European version of a gas lobby (which unlike the US version relies on artificially pumped up electricity prices to justify its existence).


  3. T. Caine  

    I think Ed’s summary is largely accurate, and even as a proponent of board-based, increased energy efficiency, I think rebound is an important factor and very real. That being said, even outside of Ed’s voluntary/involuntary categories, rebound does not affect all efficiency measures equally.

    There are plenty of things that we have and use without a regard to their energy consumption. Dishwashers, washing machines, dryers, any cooking appliance… I’d even argue lighting falls into this category. I do not think there is any data to support that people do more loads of laundry with a more efficient washing machine. Improving the efficiency of these staples is probably a net gain.

    However, when it comes to things like air conditioning & heating, refrigerators and definitely automobiles, rebound has a lot of traction.

    In the end, the most notable gains in lowering consumption is usually not vying for a more efficient gizmo, but just using our gizmos less.


Leave a Reply