Every few years, a new paper comes out about the rebound effect and the issue receives some short-term attention. (When a consumer or business buys an efficient car or air conditioner, they may use their energy-efficient equipment a little more often or may spend some of their energy bill savings on things that use energy—these are examples of rebound effects.) ACEEE wrote a paper on the rebound effect in 2012, concluding that both direct and indirect rebound effects exist, but they tend to be modest. Overall, we found that rebound may average about 20%, meaning that 80% of the savings from energy efficiency programs and policies register in terms of reduced energy use, while the 20% rebound contributes to increased consumer amenities (for example, more comfortable homes) as well as to a larger economy.
Recently, E2e and the Breakthrough Institute each published a new paper about rebound. E2e’s work is largely consistent with what we’ve written, while Breakthrough’s paper alleges that the rebound effect is much larger than most other research has found. In order to contribute to the ongoing dialog on this issue, this blog post briefly summarizes the findings of both papers and then critiques the Breakthrough work.
E2e, a joint initiative of three universities, released a working paper entitled “The Rebound Effect and Energy Efficiency Policy.” In it, they discuss various types of rebound and ways to analyze it. Much of their data relates to gasoline and oil prices and consumer and market responses to changes in those prices. They find that for developed countries, “most… studies fall […] in the range of 5 to 25 percent” direct rebound effect (where direct captures consumer response but not whole-economy effects). In developing countries, where incomes are lower and impose constraints on miles driven and other energy-consuming behavior, the E2e paper finds the “most common range” is 10-40% demand elasticity (related to but not exactly the same as direct rebound). They also discuss macroeconomic effects, emphasizing studies that show rebound of 11 percent and 21 percent due to economic growth. By way of comparison, the ACEEE paper estimates 10 percent direct rebound on average for the United States, noting the first of the two economic growth studies. In addition, in the case of oil prices, the E2e paper discusses how improvements in fuel economy soften oil prices, which can lead to a 20-30% increase in global oil use due to these price effects. Bottom line: The E2e paper sees modestly higher rebound effects than the earlier ACEEE paper.
The new Breakthrough paper, on the other hand, continues the institute’s longstanding view that rebound effects are very large and often result in “backfire,” meaning that rebound is larger than the efficiency savings and therefore energy use increases rather than decreases. The latest Breakthrough paper discusses three historic case studies, involving lighting efficiency trends over centuries, and improvements in electricity generation and steel production over decades.
Regarding lighting, they rely on work by Fouquet and Pearson that shows that in the 1800s lighting energy use grew much faster than incomes as society switched from low-efficiency candles to higher efficiency kerosene and gas lamps. Breakthrough uses this period as evidence for backfire but does not discuss how in the 1900s, as electric lighting became predominant, Fouquet and Pearson find that lighting energy use increased much more slowly than incomes increased, finding that after 1900, “rebound effects in the lighting market were still strong, although they do not suggest backfire.” Much has changed since the 1800s and the lack of more recent evidence shows how tortured the Breakthrough analysis is.
Regarding electricity use, Breakthrough discusses how electricity use has risen more quickly than generating plant efficiency has increased. The authors call this backfire, even as they acknowledge that these trends are also affected by rising incomes, urbanization, changes in consumer preferences, and other socioeconomic and demographic trends. They provide no evidence on the relative importance of energy efficiency relative to these other factors. Furthermore, they seem to mix up energy efficiency and economic efficiency. They focus on the period of 1900-1950 in the United States, when electricity use per capita increased 30 times, while residential prices decreased about 95% in real terms. These price decreases are substantially greater than the energy efficiency improvements, indicating improved economic efficiency, such as economies of scale. Economic efficiency also explains a good portion of the steep declines in the price of lighting and appliances that contributed to rising electricity use. In other words, a large portion of the increase in electricity use was due to improvements in economic efficiency that go beyond the technological energy efficiency improvements.
In the case of iron and steel, the authors appear to mix up energy and economic efficiency again. To provide just one example, they discuss China’s large increase in energy use for steel production, and do not address how much of this may be due to energy efficiency as differentiated from many broader changes that are profoundly affecting the Chinese economy.
Breakthrough released their new report with an op-ed in the New York Times. The op-ed goes several steps further than the report. First, applying its claims of lighting backfire from the 1800s, it claims that LED lighting, for which the most recent Nobel Prize in physics was awarded, will increase lighting energy use, particularly in developing countries. As I wrote in a letter to the editor of the Times, LEDs are about six times more efficient than incandescent lamps, so in order to reach the backfire point, the average purchaser would need to increase the amount of lighting they use by a factor of six. While such an increase may well happen among the poorest households in developing countries, it is unlikely to be seen in developed countries, or even among the middle class in developing countries.
The Breakthrough op-ed also claims that the International Energy Agency and the Intergovernmental Panel on Climate Change find that “rebound could be over 50 percent globally.” While technically correct, their claim takes the upper end of the ranges found in recent IEA and IPCC studies. For example, IEA states, “Direct rebound can range from 0% to as much as 65%. However, estimates tend to converge between 10% and 30%.” It would be much more accurate if the institute would cite the full range, instead of looking only at the extreme. Applying that logic, I could argue that IEA supports ACEEE’s 10% direct rebound estimate--at least 10% is within IEA’s most likely range of 10-30%. IPCC estimates get similar treatment from Breakthrough.
Bottom line: The E2e analysis is very reasonable, but Breakthrough appears to be more interested in exaggerating to make its case, rather than sticking to the facts. The truth is that for 40 years energy efficiency has had a dramatic effect on worldwide energy consumption. In the United States, if we were to use energy today at the rate we were in 1974, we would be consuming more than twice the amount that we are actually using.
Energy Intensity Declines Due to Efficiency Gains
Source: Updated from ACEEE using data from U.S. EIA & U.S. Census data
And when you add the fact that these savings also have positive economic effects, such as creating jobs and protecting the environment, it’s easy to see why energy efficiency has received bipartisan support in every major energy bill since the 1970s—it works.