Efficiency lives — the rebound effect, not so much

Shining some light on bad analysis in The Economist

Today’s guest debunker is Evan Mills, a leading scientist at Lawrence Berkeley National Laboratory.  He is a widely-published expert on energy efficiency, which remains THE core climate solution and the biggest low-carbon resource by far.

Misreading of a new Sandia National Laboratories study on efficient lighting has led The Economist into the dark.  In the process, they inverted the researchers’ findings, perpetuating the myth of the “energy rebound effect,” which has been postulated in theory but never shown empirically to be significant. The assertion is that the too-cheap-to-meter energy savings from energy-efficient lighting will induce people to become ravenous light hogs, increasing their demand for lumens to such an extent that lighting energy use will grow despite factor-ten improvements in efficiency compared to today’s incandescent lighting.

Perhaps a new species of “Efficiency Deniers” will be born of all this”¦.  I can hear it now:  Energy-efficient technologies will fuel global warming and other scourges of the earth.

The authors of the underlying Sandia study have already pointed out one of many egregious flaws in the article:  Comparing today’s mixed-source lighting energy use with that projected for three and a half decades downstream (larger population, GDP, etc.) and attributing the predicted increase to LEDs (solid-state lighting):

SIR – Your surprisingly negative article on energy efficient lighting technologies (“Not such a bright idea”, August 28th) appears to have resulted from a misunderstanding of our paper in the Journal of Physics.

Unfortunately, your writer’s reasoning erred by comparing today’s per capita mixed light consumption with the projected 2030 consumption for all-solid-state lighting (SSL), rather than comparing the projected 2030 consumption with and without SSL. Because of this “apples and oranges” comparison, you drew a number of erroneous conclusions. For instance, you stated that in 2030 a tripling of electricity prices would be required before energy consumption for lighting declined. In fact, our paper shows that, for the two 2030 scenarios (with and without solid-state lighting), a mere 12% increase in real electricity prices would result in a net decline in electricity-for-lighting consumption. This “green” result is obtained while at the same time enabling consumers in 2030 to use three times more light with SSL than without it. Your amusing but hopefully tongue-in-cheek conclusions about the “greenness” of incandescent lighting would be, if serious, off-base and in our view potentially harmful.

Jeff Tsao
Harry Saunders
J. Randall Creighton
Michael E. Coltrin
Jerry A. Simmons
Sandia National Laboratories
Albuquerque, New Mexico

To be more specific, The Economist says that under an LED lighting scenario for 2030 the amount of electricity needed to generate light would more than double (at today’s energy prices) while the Sandia researchers actually say that lighting energy use would be about the same in 2030 with or without LEDs. And, with real energy price increases of about 3%/year, the researchers project a 25% absolute decline in lighting energy use from 2005 levels even with an assumed ten-fold increase in lighting service levels.

It would indeed be worrisome if the energy savings from this game-changing technology were gobbled up by a penchant for cheaper lumens. However, the significant stipulated growth in demand for lighting services is based on an extrapolation of the growth that took place three centuries ago (in the days of whale oil lamps), without regard for contrary indications in the recent and far more relevant past.

Consider the highly analogous case of compact fluorescent lamps (CFLs), which are today replacing Edison’s incandescent lights hand over fist while reducing energy use by five-fold (about the same improvement that LEDs will eventually offer over CFLs on a per-delivered-lumen basis).

CFLs have made a major dent in the overall lighting market in the U.S., replacing about one in four incandescent lamp sales in just a decade’s time.  Importantly, this has been accompanied by no visible change in the overall number of light bulbs sold or illumination desired by their users. By both of these measures (sockets and lumens), the speculative rebound-effect of consumer behavior is invalidated by the facts.  And other well-demonstrated strategies for improving lighting systems – e.g., dimming coupled with controls for daylight harvesting – actually reduce light levels by better tailoring light output to what is needed by users and not otherwise provided by daylight.  These kinds of strategies must figure big in any serious scenario for 2030, irrespective of the underlying lighting technology mix.


Source: U.S. Department of Energy

More is not always better. For rich and poor alike, the sky (i.e., a burning sun in every living room) is not the limit for lighting demand.  Illuminating engineering societies around the world have actually been reducing their lighting-level recommendations for many years running, as overzealous guidelines have been seen to create excessive glare and other problems.  Even granting some pent-up demand for more lumens, LEDs can save energy because their light can be more precisely directed to end-use needs and more easily controlled.

Another problem with the underlying analysis is the lack of allowance for the fact that energy policies promoting the projected expansion of LEDs will endeavor to curb their over-use via over-lighting.  The analysis is also predicated on a very aggressive projection of LED cost reductions (driving their adoption), which may well not be met. In any event, LEDs will find a home in certain markets but other lighting strategies (and perhaps even new technologies not yet discovered) will also be part of a diverse mix.

LEDs stand to have their largest, earliest effect in the developing world. A quarter of humanity still lack access to electricity and instead attempt to meet their appetite for light with kerosene and other fuels, consuming a whopping 20% of global lighting energy in the process. There is vast suppressed demand for illumination in the developing world, and the rebound effect should show up most strongly here.  But even in this context LEDs are already enabling vast increases in light for underserved populations, coupled with major decreases in energy use and carbon emissions.  IFC, the World Bank, and the U.S. Department of Energy are betting on this.

Bottom Line: When people save money from next-generation lighting retrofits, they will have more than enough other things to spend it on than buying a pair of sunglasses and cranking up the light.  The only rebound effect we need to worry about is how public understanding of emerging technologies can rebound from misinformation.

— Evan Mills, author of Building Commissioning: The Stealth Energy Efficiency Strategy

JR:  The efficiency deniers at the Heritage Foundation were quick to jump all over this misreading (here).  Then again they have opposed energy efficiency for a long time and lost their grip on reality a while back, calling science ‘magic’.

Good studies on the rebound effect are hard to find, since here are so many confounding factors, but one of the best recent analyses in the heavily studied area of fuel economy standards, “Fuel Efficiency and Motor Vehicle Travel: The Declining Rebound Effect” finds only a very small effect.

Related Posts:

17 Responses to Efficiency lives — the rebound effect, not so much

  1. mike roddy says:

    Not surprised. I subscribed to The Economist for years, until noticing that when they got outside their comfort zone- international business conditions- they tend to be very careless with methodology. This is especially true for environmental analysis, where there is little rigor at all, intentional or not.

  2. Colorado Bob says:

    More proof that the system is becoming more energetic –

    Huge Windstorm Spawns New Classification: ‘Super Derecho’

    A windstorm that swept across Kansas, Missouri and Illinois in May 2009 was so fierce that it has earned a brand-new name: super derecho.

    A derecho (from the Spanish adverb for “straight”) is a long-lived windstorm that forms in a straight line – unlike the swirling winds of a tornado – and is associated with what’s known as a bow echo, a line of severe thunderstorms. The term “derecho” was first used over a century ago to describe a storm in Iowa. Across the United States there are generally one to three derecho events each year.

  3. Yes and no (being an economist myself).

    I do agree, as you rightly say, that, in isolation, the rebound effects will tend, on a net basis, to be far outweighed by the overall gains.

    But, on the other hand (there goes that economist prattle again), the individual gains from reductions in energy for a specific usage do not exist in isolation.

    Instead, the dollar savings from an individual’s spending less to use lights or drive a car (net of any greater amortized capital expenses) creates higher income for that person.

    In today’s America, are those new dollars going to be invested in even greater capital investments in energy efficiency, where there would be no rebound effect, or are they going to seek an outlet in higher discretionary expenditures, such as trips on airplanes to Europe, or on eco-trips to see the melting glaciers in Greenland?

    The point being new dollars created by lower energy expenditures are not going to just remain under the mattress. They will be spent somewhere, and where that flows determines the aggregate rebound effect that scares the bee gees out of me. For the 800 pound gorilla in the room is that (1) we need to have India and China agree in a carbon reduction treaty to achieve any global savings, and (2) they will not sign anything that consigns their people to perpetual poverty while American’s continue to live the high life, made higher by the surplus produced by wise conservation choices.

    We’ve got a big problem here that, absent a Great Awakening to revisit what THE American dream means (does more make us happier), we’re not going to pull through by changing light bulbs, which, in this sense, ironically, might lessen further the likelihood of making the Third World copacetic with our largesse.

    [JR: No, this theory of yours does not hold water. Money saved on energy gets spent, on average, on typical things that have a typical amount of embedded energy, maybe 5%. So this embedded energy rebound effect is also essentially insignificant.]

  4. catman306 says:

    It was the magic of money, Koch Industries money, that put a spell on the Heritage Foundation some years ago. Don’t expect any reasonable reports about reality to be coming from that organization any time soon. Status Quo money has no use for the changing realities that come with a changing world.. Their mission seems to be to bring back the past. If their magic could only bring back the world wide environment of 1850, I would support them in a heartbeat. But you have to drink the Kool Aide to believe any of their magic is real and I wouldn’t recommend it.

    Efficiency, over the long haul, has always been the way that our Universe and our planet have evolved. If there were a more efficient way for bio-chemical life to turn sunlight into carbohydrates than photosynthesis, our plants would all use that method. The Universe and Gaia smile on efficiency. We should too.

  5. Edward says:

    How much mercury do CFL’s contain. Are they contaminating the environment with mercury?

    [JR: Coal plants are contaminating the environment with mercury. CFLs are reducing the mercury loading you and your children are likely to experience.]

  6. I can’t help but think a valid class action lawsuit on behalf of the young people in this country could be laid at the feet of Koch brothers et al. for their active participation in deceiving America about the dangers of climate change.

  7. darth says:


    The mercury is CFL’s is contained within the bulb. That is why we recycle them instead of tossing them in the regular trash. The power savings translates to less mercury spewed into the air by coal plants so it’s much better for the environment.

    For those worried about the rebound effect, I think the extra dollars will be spent on improved quality items, not quantity. At least for me, I’ll spend money saved on energy on buy stuff like local organic food, not necessarily extravagant high energy activities. Or more likely I’ll use the savings to pay for the probably more expensive bio-based fuels I use for my vehicles and airplane trips.

    As for the rebound effect, it’s a silly argument. I’m not going to add more lights to my house just because they cost me less to run – I already have enough light. I don’t need an indoor tan. Similarly, my commute to work doesn’t get longer because I buy a more fuel efficient car.

  8. more is definitely not better. learned that already.

  9. paulm says:

    “The greatest shortcoming of the human race is our inability to understand the exponential function.”
    Professor Al Bartlett

  10. BBHY says:

    @Peter Anderson: “we need to have India and China agree in a carbon reduction treaty to achieve any global savings,”

    China is the world leader in the production of solar panels and is third in wind power generation.

    What we really need is for the US to get serious about carbon reduction. This attitude that “we can’t do anything until someone else does it first” is not the way that the US became a world leader!

  11. AL says:

    Efficiency hasn’t really worked for cars though. Efficiency of automobile engines has increased dramatically over the last 40 years but CO2 emissions have not dropped. There are more cars on the road, there are bigger cars on the road, and the cars accelerate faster and have more power hungry features. The over all effect from efficiency in automobile engines has clearly resulted in more CO2 emissions, not less.

    But of course, electric cars powered by renewables could reverse that.

    The point is that sometimes gains in efficiency does result in higher consumption rates, specifically in economies where growth is the most important goal, and in cultures where gathering and accumulating more & more stuff seems to be the only reason for being.


  12. Omega Centauri says:

    It was properly called the Jevon’s paradox, from an eighteenth study of coal consumption versus efficiency of steam locomotives. At least in that particular case (immature industry in rapid expansion) greater energy efficiency was accompanied by much greater use of locomotives. So in at least some cases, greater efficiency leads to a big enough decrease in cost that the volume of the activity increases enough to more than makeup for the efficiency gain. I really don’t think Jevon’s applies to lighting however (for all the reasons cited above).

  13. AL says:

    When talking about energy consumption and behavior it is also referred to as the Khazzoom–Brookes postulate. In short, in growth economies increased energy efficiency at the micro-level leads to increased energy consumption at the macro-level.


  14. Very interesting discussion. People should use google power meter!! Here in perth we install PV solar panels and encourage our customers to use google power meter to track and minimise consumption.

  15. spiritkas says:

    I’m always wondering what and how much of the news to trust. It seems that disinformation lasts a lot longer than not once it gets out as people often don’t go back to read corrections. Even in this particular exposed instance of it, will The Economist issue a correction saying their journalist got a tounge lashing from his sources whose work he was reporting on? On the odd chance they did publish some sort of correction, who would read it?

    It is great to catch this stuff where you can, but it is so systemic and common as to really cast doubt upon the whole industry of for-profit journalism.

    I’m still not sure what the balance is between pushing ahead with your own polcies and ideas about the world and spending time fighting off misinformation and refuting fradulent nonsense pushed out by the status quo crowd. A bit of both strategies seems to be the solution.

    The echo effect doesn’t applify progressive ideas from speaciality sources that only the choir reads up onto the main stage, that’s a right reserved for status quo pro business as usual stories. But then again the progressive media doesn’t have 40 years of think tanks and foundations backed by hundreds of millions of dollars working system wide to push their agenda into every facet of life, media, and advertisements. So it makes sense that we have the media we have, a for-profit model doesn’t see itself as able to publish outside of the status quo business models which fund their existence. It is a structural bias to provide fluff and disinformation.

    And lastly to stay on point a little bit in terms of efficiency rather than media coverage. I think it is a gross and obvious mischaracterization to imply people will run out and install more sockets in their houses or businesses and also in the comments the strange feedback idea that saving $50-$100 a year on lighting will suddenly result in a big jet setting european vacation. It’ll probably go to a bank for a mortgage or to a credit card company or in savings or be spent on bread or milk or a fancier cell phone every 2 years.

    Invoking pure speculation about what people might do with a few extra bucks to attempt to invalidate an arugment for efficiency is innane. LED lights are an improvement of an existing technology and wherever they go in the world and displace kerosene in new markets will surely result in a net decrease in green house gases. This isn’t electricity for the first time or the transcontinental railroad or steam power or the automobile or factory farming…it’s a better lightbulb for a mostly already existing need that will save a few dozen bucks a year in energy usage.

    Moutains that are actually mole hills, they sure do keep us busy. We could be out buying LED bulbs right now and burning less coal.

  16. darth says:

    AL @11: Depends on how you measure efficiency when it comes to cars. Greater power per gallon of fuel CAN be used to make a more powerful engine or a bigger vehicle with the same MPG, or it can be used to make a car with higher MPG.

    I would say that vehicle efficiency is measured most practically in MPG (or per person MPG) not in how much raw power a given engine extracts from the fuel. So the examples you give of higher HP and bigger cars are not more efficient at all.

    Again, if I buy a car with higher MPG it doesn’t change the length of my daily commute, therefore it’s more efficient. Looking at it this way the best way to improve your MPG is to carpool – with 2 people you double your effective MPG at zero cost.

  17. AL says:


    I agree Darth. But we must take into account behavior when talking about efficiency in terms of reducing the overall carbon footprint. A fuel efficient engine in Europe actually does result in lower overall emissions because they have higher fule prices and roads that don’t accommodate large vehicles, also Europeans tend to not drive as much in general.

    But in North America our behavior has been the opposite. We have multiple cars per individual, we drive everywhere, design our twon and cities around cars, have enormous roads that accomodate large vehicles, etc… Where i live in Alberta the average individual has 3 times the carbon footprint of any other Canadian province – an increase since the 70’s while at the same time fuel efficiency has increased dramatically.