Amory Lovins’ Efficiency Fantasy

Why Rocky Mountain Institute’s Energy Solutions Don’t Add Up


In selling their vision of a world running on efficiency and renewables, Amory Lovins and the Rocky Mountain Institute have ignored a substantial body of research demonstrating the importance of rebound effects. A new analysis from the Breakthrough Institute shows how the growing expert consensus that energy efficiency rebound is real and significant substantially undercuts RMI’s projected gains from efficiency measures and makes their proposals of limited relevance as far as climate policy is concerned.

February 22, 2013 | Breakthrough Staff,

Energy efficiency has often been proposed as a kind of “holy grail” for environmentalists: a means of contributing to climate change goals while also spurring economic growth. Almost no major institution today leaves out energy efficiency measures from climate mitigation plans. Many national governments, the International Energy Agency, and the United Nations Intergovernmental Panel on Climate Change have all recommended efficiency as a business-friendly way to reduce significant quantities of greenhouse gas emissions.

Yet many of the chief proponents of energy efficiency make problematic assumptions about how effective such measures are, neglecting a growing body of expert literature. McKinsey and Company, the National Resources Defense Council, Rocky Mountain Institute, and others assume a near-linear or near-direct relationship between improvements in energy efficiency and reductions in aggregate energy consumption. In other words, they ignore the possibility that a significant portion of the efficiency-related energy savings will be offset as cost savings from efficiency improvements lead to increased energy use, otherwise known as “rebound.”

Rebound occurs when individuals and firms increase energy use as a result of cost savings from efficiency gains. In its most extreme form (“backfire”) rebound can completely offset the initial efficiency-related energy savings. While rebound is difficult to measure, especially at macro-scales and in its indirect effects, there is a growing consensus among academics that it is significant and should be taken into account in energy efficiency modeling.

The Breakthrough Institute was therefore disappointed when we found that Rocky Mountain Institute’s recent publication, Reinventing Fire: Bold Business Solutions for the New Energy Era (2011), appeared to ignore the bulk of the evidence on rebound. As lead author and Rocky Mountain Institute founder Amory Lovins recently wrote in the Huffington Post, Reinventing Fire promises “a 2.6-fold-bigger US economy by 2050 with no oil, coal, or nuclear energy, one-third less natural gas, a $5 trillion dollar net savings, 82-86 percent lower carbon emissions, and no new inventions” solely through profit-motivated business activity. Attractive to anyone frustrated by the current political gridlock, Lovins argues that the “twin transition to efficiency and renewables” “requires no new federal taxes, subsidies, mandates, or laws. The policy innovations needed to unlock and speed it need no Act of Congress.”

Efficiency measures are central to achieving the Reinventing Fire vision. When contacted by staff at the Breakthrough Institute for clarification about how Reinventing Fire treats rebound effects, Lovins informed us that the analysis does take them into account — a departure from previous publications by RMI. Lovins’ defense, unfortunately, does not survive close scrutiny. After a thorough comparison with the established literature on rebound effects across sectors, we found that Reinventing Fire simply fails to reflect a fair assessment of the data. The RMI therefore minimizes the estimated effects of rebound, resulting in deeply flawed and misleading prescriptions for business and policy leaders.

The RMI analysis is all the more unsettling because of the consensus among experts and analysts that has emerged over the past decade. A 2007 review commissioned by the Energy Research Centre in the United Kingdom, drawing from more than 500 studies on the topic, rejects the premise that rebound is too small to be ignored. The authors of the report, led by Steve Sorrell at the University of Sussex, found rebound levels ranging from 10 percent (low, but certainly still significant) to as high as 80 percent, depending on the sector in question. Our own survey of the literature, published in 2011, finds that rebound effects are real and significant, operating at various scales and with the greatest magnitude at the macroeconomic global scale most relevant to climate change mitigation efforts. A 2011 review of the evidence commissioned by the European Commission concludes, “In general, evidence does not support the views that rebound effects are too small to be insignificant.”

Citing these and other publications, analysts and journalists including Matthew Yglesias of Slate, John Tierney of the New York Times, Robert J. Michaels of the Cato Institute, David Roberts of Grist, and Severin Borenstein of the University of California-Berkeley have come to similar conclusions about the significance of rebound. This emerging consensus casts doubt on dreamy visions of efficiency — like the one put forth in Reinventing Fire — that dismiss the importance of rebound.

Our point is not to argue against energy efficiency improvements. Cost-effective energy efficiency improvements make great economic sense and can be a key determinant of future economic welfare, particularly in developing countries. They also have an important role to play in the transition to a low-carbon economy. The reality of rebound effects, including the uncertainty represented by estimated ranges within sectors, does, however, give caution to the sweeping claims of RMI’s “efficiency-renewables transition.”

What follows is a discussion of rebound and how it relates to various sectors detailed in Reinventing Fire, including transportation, heating and cooling, industry, household appliances, and multi-factor productivity improvements, as well as indirect and economy-wide rebound effects. We conclude with a discussion of the implications of rebound for efficiency as a climate strategy.



The Reinventing Fire (RF) analysis cites two studies that estimate direct rebound effects for personal automotive transportation rebound at 3 percent to 22 percent. The RF analysis, however, ultimately decides to waive all automotive transportation rebound effects, arguing that increasing income and a saturation of the amount of time we spend behind the wheel will eliminate any rebound in this sector (Reinventing Fire pp. 43, 45 and note 92). This justification is unsupported in the literature. While it is true that transportation rebound rates tend to decline with rising income, nowhere does the academic literature suggest that the rate will go to zero. Moreover, the assertion that saturated driving time will eliminate transportation rebound is completely unsupported at the global scale, where personal automotive transportation in countries like China and India is far from saturated. Although RF focuses exclusively on the United States economy, personal automotive transport at the global scale will almost certainly experience rebound for years to come, as per-capita driving hours in China, India, and other developing nations continues to increase.

Drawing on 17 studies that use aggregate time series, cross-sectional, or panel data, Sorrell (2007) finds that the long run direct rebound effect for personal automotive transport in OECD nations lies somewhere between 10 percent and 30 percent. Drawing from 13 studies that also use aggregate and survey data, Gavankar and Geyer (2010) find that long term rebound for personal transport is between 20 percent and 65 percent. A recent European Commission report, as well as our own literature review, both of which draw from several independent analyses, place the long-term automotive transport direct rebound effect at between 10 percent and 30 percent. While Reinventing Fire draws on just two estimates and ultimately decides to waive all automotive transportation rebound effects, the consensus in the literature indicates that direct rebound in personal transportation is significant and should not be omitted from analyses.


Heating and Cooling

Again drawing on just a few studies, the Reinventing Fire analysis assumes rebounds of 10 percent and 5 percent for household heating and cooling, respectively (Reinventing Fire General Methodology, pg. 5). Although this is consistent with some selected studies, it is inconsistent with the average across the body of literature. We conclude, based on Sorrell (2007)’s extensive review, that the likely ranges for space heating and cooling direct rebound are 10-30 percent and 1-26 percent, respectively. The RF estimates, drawn from only a few sources, do not align with the average and range estimates supported by a wider body of literature, showing that heating and cooling rebound effects may up to three to six times larger.



The Reinventing Fire analysis refutes the possibility of a direct rebound effect in industrial processes, arguing that because energy usually makes up only a few percent of total manufacturing costs, saving part of that few percent is unlikely to make products far cheaper and strongly stimulate sales (Reinventing Fire, pg. 145). While this is true of many industries (for example, Greening et al. 2000 find that energy is typically less than 10 percent in the total cost of most products), it is certainly not for energy-intensive ones such as chemical production, steel and other primary metals, electric utilities, auto manufacturing, and agriculture. Indeed, a historical analysis by Saunders (in review) finds that the long-term rebound from output in these industries is 33 percent, 30 percent, 25 percent and 21 percent, respectively. Energy efficiency improvements in energy-intensive sectors feature prominently in the Reinventing Fire strategy, so neglect of rebound effects at such levels would call for a significant correction.

Even in non-energy intensive industries, however, rebound from output may be present. Saunders (in review), using a rigorous econometric methodology to analyze historical rebound across 30 producing sectors of the US economy, finds that long-run rebound effects due to increases in output for these industries range from 0 percent to 15 percent.

Moreover, when we include industrial rebound due to substitution — substituting to more energy intensive factors of production in response to energy efficiency gains and cost savings — overall industrial rebound may be much higher. Saunders (in review) finds that long-term substitution rebound across 30 industrial sectors ranges from 10 percent to 90 percent. The RF analysis entirely ignores the possibility of substitution effects, departing from both the literature and long-standing economic theory regarding the optimization of factors of production in profit-maximizing firms.


Household Appliances

The authors of Reinventing Fire flatly state that they don’t believe the technical literature on rebound for household appliances without justifying why (Reinventing Fire General Methodology, pp. 5-6). While the literature tends to support low rebound effects for household appliance efficiency improvements, they are potentially larger when efficiency gains enable entirely new uses for those products. Consider, for example, how higher-efficiency LED lights are now used in a variety of products that were not compatible with incandescent lights or compact fluorescent lamps. The RF analysis fails to address the possibility of such “frontier effects.” (For more on frontier effects, see pages 46-48 of our literature review.)


Multi-Factor Productivity Improvements

As Lovins commonly notes in Reinventing Fire and other works, energy efficiency improvements frequently accompany simultaneous improvements in the productivity of other factors of production, such as labor or capital. For example, Lovins explains that in addition to increasing energy efficiency, better industrial lighting could also lead to more productive factory workers (Reinvening Fire, pg. 151). What Lovins fails to address when he writes about these kinds of “multi-factor” productivity improvements, however, is that they are predicted to drive significantly greater rebound effects than a simple improvement in energy efficiency would. This is because, in addition to significant rebound associated with energy efficiency improvements, neoclassical growth theory predicts that improvements in the productivity of capital, labor, and other materials will drive rebounds in demand for energy. Therefore, the rebound effects associated with multi-factor productivity are likely to be even larger than those associated with “pure” energy efficiency improvements (Saunders, 1992, 2005, in review; Sorrell, 2009; Jenkins et al., 2011, at pages 43-44).


Indirect and Economy-Wide Rebound

Unlike direct rebound effects — which entail an increase in demand from an energy user for the same product or energy service that experiences an increase in energy efficiency — indirect rebound effects drive an increase in demand for other energy services during or after an efficiency increase in an original service. First, energy saving technologies and investments will require energy to manufacture and install, and this energy “embodied” in the efficiency improvements themselves will offset some portion of the net energy savings achieved (Sorrell, 2007). Second, any savings from energy efficiency improvements that are not spent on the same good or service will be spent on some other combination of goods and services.

This re-spending effect may increase energy consumption. Indirect rebound effects are difficult to measure and there is debate about their precise magnitudes due to sporadic and sparse data. Nonetheless, there is evidence that they are significant. Our review, for instance, finds that the embodied energy rebound for most cost-effective efficiency improvements is likely to be less than 15 percent but more significant for efficiency improvements with long payback periods, including some building efficiency measures. Additionally, Sorrell (2007) finds that the re-spending effect for consumers could be on the order of 5-35 percent.

At the economy-wide macroeconomic scale, the aggregate impacts of widespread energy efficiency improvements can lead to substantial rebound effects, as producers and consumers respond in turn to various cascading changes in the price of goods and services, the pace of economic growth quickens, and market prices for fuels may fall, driving a further rebound due to market price effects. Just as microeconomic decisions may lead to macroeconomic effects that are greater than the sum of the aggregate decisions (enshrined in economic literature through the macroeconomic “multiplier” concept), countless industrial, commercial and consumer efficiency improvements may lead to a macroeconomic rebound effect that is larger than the sum of the rebound effects associated with those improvements.

The evidence available for the economy-wide rebound is even more sporadic than that for indirect rebound. Nevertheless, most analyses suggest economy-wide rebound is significant and large. A number of “Computable General Equilibrium” models (see page 34 of our review) generally show rebound at the scale of a national economy of 30-50 percent or greater, with a surprising number predicting rebound greater than 100 percent (aka “backfire”). These studies look at national economies and thus ignore global, macro-economic impacts beyond national boarders, which can add additional rebound in energy consumption. Sorrell and Dimitropoulos (2007), summarizing a set of theoretical literature on the topic, predict economy-wide rebound effect to be in the 35-40 percent range. “Integrative modeling,” a more detailed modeling approach utilized by energy analysts at Cambridge to study macroeconomic rebounds, found that if the world adopted all of the “no regrets” energy efficiency policies suggested by the International Energy Agency, then economy-wide rebound effects would erode more than half of expected savings (52 percent) in the long-term. There are also several reasons to think this is may be a conservative estimate (see pages 39-40 of our review).

Reinventing Fire entirely avoids the literature on indirect and economy-wide rebound. As a Rocky Mountain Institute consultant explained in an email exchange, Reinventing Fire projections are calculated as a percentage of EIA’s forecasted energy demand. EIA’s energy demand projections contain assumptions about the rate of economy-wide energy efficiency improvements, and include a 5 percent economy-wide efficiency rebound effect. The Reinventing Fire analysis assumes much greater economy-wide efficiency improvements than EIA’s baseline scenario, but backs out the 5 percent rebound effect included in the EIA scenario, and ultimately does not include an economy-wide rebound effect (see Reinventing Fire General Methodology, pg. 6 for more information on Reinventing Fire’s treatment of economy-wide rebound). This treatment of economy-wide rebound effect is inappropriate given that most theoretical literature on the topic predicts the effect to be at least 30 percent.


Implications for Climate Strategy

Reinventing Fire’s failure to provide for an accurate accounting of rebound effects for energy efficiency measures makes its prescriptions for domestic and global climate policy of limited relevance. In particular, Rocky Mountain Institute’s decision to waive economy-wide rebound is at odds with the established literature, with serious implications for measuring efficiency gains. Reinventing Fire concludes that by 2050, rigorous energy efficiency improvements in the United States could lead to economy-wide savings in energy consumption (24 percent), CO2 emissions (82-86 percent), and spending ($5 trillion) while sustaining economic growth. If we assume that even the most conservative estimates of economy-wide rebound are accurate, CO2 and aggregate energy savings estimates in Reinventing Fire decrease substantially.

Importantly, rebound effects are almost certainty larger in poorer, developing nations than in developed countries like the United States. This is important, because growing demand in developing nations is the principal driver of worldwide energy demand and CO2 emissions growth. For efficiency in end-use consumer energy services in developing nations, direct rebound effects alone are likely to be much higher than in richer nations. While far more research is needed in this critical area, studies to date have found direct rebounds in emerging economy contexts on the order of 40-80 percent (Schipper and Grubb, 2000; Sorrell, 2007; IAC, 2007). Rebound is higher here because demand for energy services is far from saturated and far more elastic (responsive to changes in price). Moreover, the cost of energy services is often a key constraint on the enjoyment of energy services, and larger re-spending and macroeconomic effects are expected in developing countries because of higher energy shares of both overall GDP and individual firms.

It is surprising that Rocky Mountain Institute, a think tank that has built its reputation around expertise on energy efficiency, failed to take rebound effects into account in Reinventing Fire. The regrettable result is that Lovins’ vision — a prosperous world built on efficiency and renewables — appears too good to be true. While the Breakthrough Institute believes efforts to improve energy efficiency should be pursued, based on sound economic sense and limited climate benefits, the growing consensus on the significance of rebound give good reason to remain skeptical that efficiency can be a primary driver of lasting reductions in climate-destabilizing greenhouse gases.


­— February 2013



Barker, T. and Foxon, T. (2008). The Macroeconomic Rebound Effect and the UK Economy. A Research Report for the UK Energy Research Center, February 2008.

Gavankar, S. and Geyer, R. (2010). The Rebound Effect: State of the Debate and Implications for Energy Efficiency Research. Bren School of Environmental Science and Management, June 2010.

Greening, L., Greene, D. L. and Difiglio, C. (2000) Energy efficiency and consumption - the rebound effect - a survey. Energy Policy 28(6-7): 389-401.

IAC (InterAcademy Council) (2007). Lighting the way: Toward a sustainable energy future. InterAcademy Council Secretariat, Amsterdam, October 2007.

Jenkins, J., Nordhaus, T., and Shellenberger, M. (2011). Energy Emergence: Rebound & Backfire as Emergent Phenomena. The Breakthrough Institute, February 2011.

Lovins, A. B. (1976). Energy Strategy: The Road Not Taken? Foreign Affairs 55(1): 65-96.

Lovins, A. B. (1977). Soft Energy Paths: Toward a Durable Peace. Cambridge, MA: Ballinger.

Lovins, A. B. (2011). Reinventing Fire: Bold Business Solutions for the New Energy Era. White River Junction, VT: Chelsea Green Publishing.

Maxwell, D. and McAndrew, L. (2011). Addressing the Rebound Effect. European Commission DG ENV, April 2011.

Rocky Mountain Institute (2012). Reinventing Fire General Methodology. Accessed October 2012.

Saunders, H. D. (1992). The Khazzoom-Brookes postulate and neoclassical growth. Energy Journal 13(4): 131.

Saunders, H. D. (2005). A calculator for energy consumption changes arising from new technologies. Topics in Economic Analysis and Policy 5(1): 1-31.

Saunders, H. D. (in review). Historical Evidence for Energy Consumption Rebound in 30 US Sectors and a Toolkit for Rebound Analysis.

Schipper, L. and Grubb, M. J. (2000). On the rebound? Feedback between energy intensities and energy uses in IEA countries. Energy Policy 28(6-7): 367-388.

Sorrell, S. (2007). The Rebound Effect: An Assessment of the Evidence for Economy-Wide Savings from Improved Energy Efficiency. UK Energy Research Center, October 2007.

Sorrell, S. (2009). Jevons’ Paradox revisited: The evidence for backfire from improved energy efficiency. Energy Policy 37(4): 1456-1469.

Sorrell, S. and Dimitropoulos, J. (2007). UKERC Review of Evidence for the Rebound Effect: Technical Report 5 - Energy Productivity and Economic Growth Studies. UK Energy Research Center.


Photo credit: Flickr user dsearls

Submit a comment