Surprise (Or Not): Saving Energy Saves Energy

Share Button

What do you say to those who claim that saving energy is a waste of time because of the rebound effect? NRDC’s David Goldstein and Sierra Martinez explain research that shows saving energy does in fact save energy.

Stanford and Berkeley researchers recently found that energy-saving devices, like more efficient electrical appliances and better insulated homes and businesses, actually do save us energy. You might be asking – as I often do — why do we need the nation’s top scientists to weigh in on such a statement?

That’s a good question, as many people around the country have had to invest a lot of time and resources to refute a damaging attack on energy efficiency that goes along these lines: maybe all these energy efficient devices actually cause us to consume more energy instead of less of it.

In other words, their theory is that since we know these appliances, electronics, lighting, cars, and other devices aren’t wasting as much energy as they once did, we’ve saved enough money that we’ll go ahead and use more energy in the end.

Unfortunately, these claims have just enough superficial plausibility to do damage to our collective efforts to support programs that combat climate change through smarter energy use.

Energy efficiency improvements (like cars that burn less fuel to get you between home and work) are a key strategy to reduce climate change pollution. And sowing the seed of doubt about their ability to do so sets our country back in efforts to save energy and clean the air we breathe – which is why we need the nation’s top researchers to affirm that, yes, saving energy really does save energy.

How do you prove ’em wrong?

When you need to refute an incorrect report–or a number of them–there are at least two ways of doing it:

  • You can show that the reports were done incorrectly; or
  • You can show that they are taking an approach that is fundamentally wrong, even if the analysis were carried out correctly.

The reports in question claim there is a “rebound effect”–that efficiency could lower the cost of energy services so much that people will demand more of them –so much more as to negate the energy savings.

Taking a closer look

A report that my colleagues and I published disproving the existence of significant rebounds primarily took the second approach, looking at a variety of studies attempting to demonstrate rebounds. We will add to that below.

The recent report from Stanford and Berkeley shows that, independent of the fact that the rebound theory is fundamentally non-scientific, the rebound enthusiasts were using the wrong data in their analysis.

In short, the efficiency doubters were using electricity price data that did not apply to the specific circumstances they were evaluating –circumstances needed to support their claim that energy efficiency would (theoretically) not actually save energy. Specifically, they used national average electricity prices for the whole economy, but the circumstances evaluated require actual prices for the particular sector and state (at least) where the actor theoretically might consume more energy. Well, it turns out that using sloppy price data has big consequences for rebound theory: depending on the location and sector of the economy analyzed, the price data used to support rebound theory could be off by hundreds of percents. See the variance in electricity prices below, showing that in some circumstances the price of electricity is a few hundred percent greater than in other circumstances. (The price of electricity is a principal driver of results in rebound theory.)

NRDC electricityprices-thumb-500xauto-22939

Simple evidence that definitively refutes larger rebounds

We can see the absurdity of the large-rebound hypothesis by looking at the last 45 years of aggregate energy consumption. The United States has more than doubled its energy productivity (GDP per unit energy use) – how much it can produce from each kilowatt-hour of electricity or therm of natural gas — since 1973. We will assume a simple doubling to be conservative.

What does this tell us about rebounds?

Assume that energy productivity is a measure of efficiency.

(Rebound advocates conveniently fail to define “efficiency” so this may or may not be what they mean. Consequently those who are trying to refute it don’t know what data they have to gather to do so. Since rebound enthusiasts have failed to define efficiency, the initial assumption that efficiency means GDP per unit of energy use may be exactly what they mean. Or it may not be. This fuzziness of what rebound enthusiasts actually are talking about and how it would be measured is another reason to be skeptical of their claims.)

For simplicity assume the current energy use is 100 and thus the doubling in energy productivity means that without the observed doubling in efficiency, energy use would have been 200. 100 is measured energy use, so it is AFTER rebound: that is, the 100 is the energy use that was measured directly, and thus includes any rebounds. The first thing we conclude is that 100 percent rebounds are impossible: an observed savings of 50 percent clearly is not the same as savings of zero. If we assumed 50 percent rebounds, the observed savings of 100 would mean a predicted savings of 200 reduced by 50 percent: it means we would have expected 100 percent energy savings. This is evidently absurd.

What if energy productivity cannot be considered efficiency?

Let’s make a conservative assumption: that 70 percent of the savings in energy per GDP cited earlier was due to efficiency and the other half was due to external changes in the economy. (The U.S. National Academy of Sciences (NAS) has estimated that some 70 percent of this improvement was due to efficiency.)

Thus, instead of attributing a reduction from 200 to 100 being due to saving 100 through efficiency, we are now saving 70 percent of 100, or 70 units of energy. The no-efficiency case is 170 and the savings is 70, leaving us with the observed consumption result of 100. (The savings from 200 projected units of energy to 170 are due to other unexplained causes.)

With an assumption of 50 percent rebound, the projected pre-rebound savings would have had to have been twice the observed savings of 70, or 140 (since 140 units of predicted savings would be reduced 50 percent due to rebound). So the case against which rebound is compared would have had us going from the projected 200 units of energy to 60 units of energy (a savings of 140 units). This is a ridiculously mistaken value, far beyond even the most ambitious estimates of what was POSSIBLE to save by efficiency policies, much less what was actually undertaken.

We would have to lower our estimate of rebound dramatically below 50 percent in order to square 1970s predictions of how much we could save through efficiency with what was observed.


More and more evidence points to the fact that efficiency savings are realized at a rate of near 100 percent without observable losses to rebounds. The latest study from the Stanford and Berkeley researchers is further proof that in the end, saving energy does, yes, save energy.

This blog originated on Switchboard,  the staff blog of the Natural Resources Defense Council. David Goldstein is NRDC’s energy program co-director in San Francisco. Sierra Martinez is NRDC’s legal director for California energy projects in San Francisco. 

Share Button

About Elisa Wood

Elisa Wood is the chief editor of and She has been writing about energy for more than two decades for top industry publications. Her work also has been picked up by CNN, the New York Times, Reuters, the Wall Street Journal Online and the Washington Post.


  1. We are an product design and entrepreneurship non-profit education company operating at Stanford (sponsored by the PreCourt Energy Efficiency Center) and other top education venues. Every year, we have an energy efficiency projects. Do you have any practical examples of the research above that we can use to educate our young aspiring entrepreneurs?
    Thank you,
    Mike Gibbs