The Efficiency Paradox
If we could make coal plants, natural gas facilities, and gasoline engines three times more efficient so that they were able to extract 90 percent of the energy stored in fossil fuels, would our environmental woes be solved?
Assuming society’s total energy use stayed exactly the same, then we could cut emissions by as much as 66 percent. That would be extremely significant. However, when energy can be extracted more efficiently, that energy becomes cheaper. When energy is cheaper, people consume more of it. This is known as the rebound effect. It means that improving efficiency could lead to greater resource consumption.
The rebound effect has been a hotly contested topic in the academic community for the last 40 years. The intensity has increased recently as climate change has become an important global issue. The consensus is that the rebound effect is real, but the magnitude of the rebound is under serious debate and may be categorized as follows1:
- Super conservation – more resources are saved than expected
- Zero rebound – as many resources are saved as expected
- Partial rebound – fewer resources are saved than expected
- Full rebound – no net resources are saved or consumed
- Backfire – resource consumption has increased
In our example, a 66 percent reduction in energy use would be a zero rebound scenario because as many resources are saved as expected, an ideal outcome. Everyone agrees that some form of rebound occurs. The debate is whether the rebound is small or large, or whether more resources are being used than before.
The rebound concept was first described by William Stanley Jevons in the 1865 book “The Coal Question”. At the time, James Watt had recently invented the Watt Steam Engine which was significantly more efficient than the previous engine in use. According to Jevons, this increase in efficiency resulted in a much broader range of industries using the steam engine and caused an overall increase in the use of coal in the United Kingdom. Thus Jevons’ paradox, as it was coined, was an extreme rebound that reached a backfire condition.2
Unaware of Jevons’ paradox, Dr. Daniel Khazzoom and Dr. Len Brookes came to similar conclusions when researching appliance efficiency and energy efficiency, respectively, in the 1980s. The same thing happened in 1992 when Dr. Harry Saunders created a model incorporating energy efficiency into neoclassical growth theory and was shocked to see the model predict increases in energy use when energy efficiency increased. He then began researching prior knowledge in the subject found that Khazzoom and Brookes had described the same concept ten years earlier. Thus he named it the Khazzoom-Brookes postulate. Only later did he discover that Jevons had beat him by over 100 years.3
Why Rebound Happens
The reasons for the rebound are attributable to three economic considerations: direct, indirect, and economy-wide effects. A direct effect is increasing the use of the specific resource because one can now afford to do so (i.e., traveling more due to cheaper fuel). Indirect effects are increasing spending in other areas because one now has more income (i.e., buying more clothes because one has more money due to cheaper fuel). Economy-wide effects are the increase in societal consumption due to the improved economy that resulted from efficiency improvements.4
It is the attempt to measure the size of each of these effects that results in so much debate around the subject.
Determining the Magnitude of Rebound
The country where the efficient technology is implemented can have a significant effect on rebound because of the maturity of the domestic industry. For example, in America the use of energy is already very high, therefore increasing the efficiency at which energy is generated will have a smaller rebound effect than if the same new energy efficient technology was implemented in a developing country where there is a larger consumer base that will have more novel uses to apply cheaper energy.5
The peculiarities of the customer have a similar effect. If a wealthy individual already puts a thermostat at the optimal temperature, then a decrease in electricity prices will not result in an increase in energy use because the individual was already using the maximum amount necessary. However, a low-income family may adjust the thermostat to a more comfortable level because of reduced energy costs, potentially increasing total energy use.6 It should be noted that this example only considers direct effects. Indirect and economy-wide effects may increase the size of the rebound.
Finally, the sector in which the efficiency improvement occurs affects the magnitude of the rebound. For example, a rebound effect is likely higher for steam engines or electric motors than for thermal insulation.4
The Evidence of Rebound
Those who would say that rebound is significant point to measures such as a 37 percent increase in total national air conditioning energy use between 1993 and 2005 when air conditioners became 28 percent more efficient.7 However, looking at the same set of data, another might say that the increase in the use of air conditioning equipment is due to the fact that income (adjusted for inflation) increased by 30 percent and the average new home was 16 percent larger in 2005 than 1993.8 This is precisely the opinion of Dr. Jim Barrett, Chief Economist at the Clean Energy Development Center. Delving deeper into the problem, some economists, such as Dr. Saunders, say that this is where energy efficiency advocates fail, because they are looking at energy in isolation without considering that materials or labor may be substituted by energy use.3
In another example, David Owens points out that refrigerator size increased in the 1960s and 70s because of an increase in refrigerator efficiency, which made larger refrigerators more affordable.7 David Goldstein, Co-Director of the Energy Program at NRDC, notes that Owens is incorrect because refrigerator size increased even as efficiency decreased in the 1950s and 60s.9 However, both of these analyses have narrowed their scope too far and focus only on direct effects. Perhaps other products became cheaper during the 50s and 60s due to improved manufacturing or energy efficiencies. This would leave consumers with more money in their pockets, which means they could choose to spend more on something such as a refrigerator. This is the exact definition of an indirect effect where people choose to spend more on something else (refrigerators, in this case) because they saved money elsewhere. This is a self-serving argument for rebound enthusiasts that would make it even more difficult to prove anything regarding a rebound effect, but it is important because it highlights the difficulty of rebound analysis.
Governmental Research Efforts
A research effort undertaken by the UK Energy Research Centre in 2007 established that much of the data necessary for quantifying the rebound effect is difficult, if not impossible, to obtain, including effects on capital, labor, and materials used. They concluded that the approximate rebound effect in developed countries is 10 percent and in developing countries is 50 percent or higher. However, it should be noted that their confidence level in their conclusions is very low due to the lack of data.4
As expected, this conclusion is seriously contested by both sides because it strikes more of a middle ground on the issue. Amory Lovins, Chief Scientist at Rocky Mountain Institute, says that rebounds are in the range of zero percent to just a few percent.10 Meanwhile, Dr. Saunders would suggest that the rebound is 50 percent to 60 percent in the US.3 Interestingly, Lovins also says that energy efficiency offers an “economic bonanza because saving fuel is a lot cheaper than buying it”.10 This would seem to support the position that energy efficiency could rebound because of all the extra money consumers and corporations save, yet he does not believe the rebound is significant.
A major part of the argument rebutting Jevons paradox is the fact that energy consumption only accounts for roughly 6 percent of GDP, therefore any direct savings from efficiency spent on other goods would still be saving 94 percent of the claimed savings.7 However, this logic is questionable because as energy efficiency increases, society spends less on energy, but society is even more dependent on it.
The paradox is difficult to prove or disprove because so many studies focus on the effect of one product or one consumer, while the effect is observed at the macroeconomic scale. However, it is clear that some level of rebound exists and that rebounds are different product to product and country to country.
The argument is well summarized by Steve Sorrell, senior fellow at Sussex University, “I think the point may be that Jevons has yet to be disproved. It is rather hard to demonstrate the validity of his proposition, but certainly the historical evidence to date is wholly consistent with what he was arguing.”7
The rebound effect is very similar to a concept called “induced demand” from the transportation industry. In transportation planning, engineers and planners are confronted with problem that when traffic congestion is relieved on one road (by the addition of lanes or constructing trains that run parallel to the road) more people choose to use that road, more people move to that area, and people begin use that road from a longer distance out; all causing that road to return to a similar or greater level of congestion than it previously had. Currently, there is no great solution to this problem, though tolls have begun to show a positive effect. Perhaps the same solution is necessary for energy efficiency, greater taxes on energy use when we use more energy.
Since the rebound effect occurs due to the reduced cost of energy efficient products, a common solution is to manipulate the cost of energy to reduce or eliminate the rebound effect. A carbon tax could reduce the rebound and help counter the adverse environmental impact of fossil fuels. A carbon tax will stimulate development of more efficient fossil fuel technologies and renewable energy. Plus, it will counter the natural rebound effect that occurs. For example, if it cost me $100 to operate my car, then a carbon tax was implemented that increased that cost to $150, then I would be more inclined to purchase a more fuel efficient vehicle or drive fewer miles to bring my cost back down to $100. In that scenario, when I bring my energy use down, I would be spending the same amount as I had been originally, but using 33 percent less fuel. A carbon tax that escalates over time can ensure that the price of energy remains the same even as we are using less of it.
Renewable energy is clearly one of the better solutions if the rebound has any kind of an effect, but even more so if the rebound is significant. If policies and programs designed to improve efficiency end up meaning society uses more energy, then those policies will fail to reduce our carbon emissions.
Interestingly, cheap renewable energy has its own rebound issues based on the same economic principles as Jevons paradox. For example, a solar customer who sees the bill go to zero may increase energy use because the customer can afford to do so. However, it is impossible for the carbon rebound to be 100 percent because the marginal productivity of fossil fuel energy doesn’t increase. Thus, if an individual were paying $100 per month on electricity, and then a new solar system reduces the bill to $20 plus $50 per month on their solar payment plan, the maximum rebound, in carbon emissions, is $30. This would be a rebound of 37.5 percent ( = 30 / 80), which is clearly not ideal, but far from a Jevons scenario of over 100 percent.
Nevertheless, technologies that do not produce emissions will clearly have a beneficial effect on the environment regardless of the rebound because they increase the percentage of environmentally friendly energy in the global mix. Plus, once the carbon emissions problem is solved, energy efficiency can go back to doing what it does best: improving welfare.
- Saunders, Harry D. “Fuel conserving (and using) production function”. Energy Economics. 2007.
- Garrett, Tim. “Rebound, Backfire, and the Jevons Paradox”. University of Utah. 2014. http://www.inscc.utah.edu/~tgarrett/Economics/Jevons_Paradox.html
- The Breakthrough. “Understanding Energy Efficiency Rebound: Interview with Harry Saunders”. The Breakthrough. 23 Jan 2013. http://thebreakthrough.org/index.php/programs/energy-and-climate/understanding-energy-efficiency-rebound-interview-with-harry-saunders
- UK Energy Research Centre. “The Rebound Effect: an assessment of the evidence for economy-wide energy savings from improved energy efficiency”. 2007
- The Breakthrough. “Understanding Energy Efficiency Rebound: Interview with Joyashree Roy”. The Breakthrough. 23 Jan 2013. http://thebreakthrough.org/index.php/programs/energy-and-climate/understanding-energy-efficiency-rebound-interview-with-joyashree-roy
- Milne, Geoffrey and Boardman, Brenda. “Making cold homes warmer: the effect of energy efficiency improvements in low income homes”. Energy Policy. 2000.
- Owen, David. “The Efficiency Dilemma”. The New Yorker. 20 & 27 Dec 2010. http://www.newyorker.com/magazine/2010/12/20/the-efficiency-dilemma
- Barrett, Jim. “Debunking the Jevons Paradox: Nobody goes there anymore, it’s too crowded”. Think Progress. 16 Feb 2011. http://thinkprogress.org/climate/2011/02/16/207532/debunking-jevons-paradox-jim-barrett/
- Goldstein, David B. “Some Dilemma: Efficient Appliances Use Less Energy, Produce the Same Level of Service with Less Pollution and Provide Consumers with Great Savings. What’s Not to Like?”. NRDC. 18 Dec 2010. https://www.nrdc.org/experts/david-b-goldstein/some-dilemma-efficient-appliances-use-less-energy-produce-same-level
- Rocky Mountain Institute. “Jevons Paradox: The Debate That Just Won’t Die”. Rocky Mountain Institute”. 20 Mar 2012. http://blog.rmi.org/blog_Jevons_Paradox