Skip to content

The MPG Illusion

It’s the beginning of summer, which means the beginning of driving season. Perhaps anticipating summer driving, many people bought new vehicles last month, putting automakers on track to have the best year since 2007.

So, here’s a question, particularly for readers who were part of this vehicle-buying wave:

Which of the following two choices leads to more fuel savings:

(A) buying a 15 mpg Cadillac Escalade instead of a 12 mpg Chevrolet Suburban, or

(B) buying a 50 mpg Toyota Prius instead of a 29 mpg Toyota Corolla?

Assume you would drive the same distances and speeds no matter which car you have.

MPG Ratings for Corolla and Prius

The answer is (A), though several academics have discovered that many people get answers to questions like this wrong.

In case you were one of them, here’s a quick explanation. If you were going to drive 100 miles, it would take 8.3 gallons in the Suburban and only 6.7 gallons in the Escalade, so you’re saving 1.6 gallons for every 100 miles driven. By contrast, buying the Prius instead of the Corolla means using 2 gallons for every 100 miles instead of 3.4, so only saving 1.4 gallons over 100 miles.

But, mistakes – like answering (B) instead of (A) – appear so pervasive that academics have labeled the phenomenon the “MPG Illusion.”

In an article published in 2008 in the journal Science, two Duke University business school professors, Richard Larrick and Jack Soll, reported results from surveys they administered to undergraduates. They asked the undergraduates to rank five comparisons, like the two above, and found that only one respondent out of 77 got the correct ranking.

Larrick and Soll’s research suggests that people think that improvements in MPG ratings translate linearly into reductions in gasoline consumption. People think, for instance, that increasing fuel efficiency from 12 to 13 MPG is the same as increasing from 22 to 23 MPG. But, the relationship does not scale linearly.

In a blog post on their Science article, Larrick provides this graph plotting gasoline used per 10,000 miles against MPG. The large gas savings from small changes in MPG at low MPG levels are clear.

In a recent paper, NYU economics professor Hunt Allcott broadens the surveyed population beyond undergraduates to a representative sample of Americans, and asks them to compare the gas consumed by cars they own to gas consumed by cars they recently contemplated purchasing. Again, he finds evidence of MPG Illusion.

MPG Illusion and Energy Policy?

So, if vehicle purchasers suffer from MPG Illusion, what are the repercussions for energy consumption and energy policy? In general, MPG Illusion suggests that consumers are likely purchasing both too many gas guzzlers and too many super-efficient plug-in hybrids, and avoiding vehicles with mid-range MPG ratings more than they would if they weren’t under the MPG Illusion.

In other words, consumers are mistakenly thinking that a Suburban is not that much worse in terms of fuel consumption than an Escalade, so purchasing too many Suburbans.

From a policy perspective, though, addressing the MPG Illusion would seem relatively easy. The EPA already mandates labeling on new vehicles, and these feature MPG ratings prominently. What if instead, labeling encouraged consumers to think about gallons of gas used per 10,000 miles driven?

In ongoing research, Hunt Allcott and Chris Knittel are running an experiment to test whether providing more information to prospective buyers in vehicle dealerships impacts which car they purchase. They are showing information on gas consumption, like what’s contained in the screenshot below, to a randomly selected set of prospective new vehicle buyers. Then, they will compare car purchases between prospective purchasers who see the screenshots and purchasers who were randomly selected into a control group and don’t see the new information.

Allcott Knittel Screen Shots1

My prediction is that there are enough people who are buying the super-efficient vehicles just because they are the best available, and not necessarily because they’re doing a calculation (perhaps wrongly) on the likely fuel savings relative to something else. After all, Prius drivers like Cameron Diaz and Larry David are unlikely to opt for the Corolla once dispelled of the MPG Illusion.

Cameron Diaz and her Prius

So, my guess is that Allcott and Knittel’s screenshots will keep people out of Suburbans more than it will keep them out of Priuses (actually, “Prii,”according to Toyota). If I’m right, providing information like Allcott and Knittel’s may turn out to be a highly cost-effective energy-efficiency policy. Stay tuned for the actual results.



Catherine Wolfram View All

Catherine Wolfram is Associate Dean for Academic Affairs and the Cora Jane Flood Professor of Business Administration at the Haas School of Business, University of California, Berkeley. ​She is the Program Director of the National Bureau of Economic Research's Environment and Energy Economics Program, Faculty Director of The E2e Project, a research organization focused on energy efficiency and a research affiliate at the Energy Institute at Haas. She is also an affiliated faculty member of in the Agriculture and Resource Economics department and the Energy and Resources Group at Berkeley.

Wolfram has published extensively on the economics of energy markets. Her work has analyzed rural electrification programs in the developing world, energy efficiency programs in the US, the effects of environmental regulation on energy markets and the impact of privatization and restructuring in the US and UK. She is currently implementing several randomized controlled trials to evaluate energy programs in the U.S., Ghana, and Kenya.

She received a PhD in Economics from MIT in 1996 and an AB from Harvard in 1989. Before joining the faculty at UC Berkeley, she was an Assistant Professor of Economics at Harvard.

13 thoughts on “The MPG Illusion Leave a comment

  1. An interesting extension of this analysis is what is the cost penalty for driving faster and using more gasoline but saving on time. How much in $/hr does it cost and how does that compare to your personal value of time? I found a couple of resources that reported mpg or %mpg decrease at changing speeds. One, by some colleagues for the DOE site and reported at gives the average percentage decrease at different speeds. If I assume 33 mpg at 50 mph and $4/gal gasoline, then going at 60 mph instead of 50 mph for 100 miles will drop the car to 28.9 mpg, require an extra .46 gal ($1.72) but save me 20 minutes, so the saved time costs me $5.15/hour. If I choose 70 mph instead of 60 mph then mpg goes to 24.9 mpg, gas use goes up an additional .56 gal but I save another 14 minutes. This works out to $9.46/hour. Lastly, changing from 70 mph to 80 mph reduces mpg to 21, gas increases 0.73 gal, and time saved is another 11 minutes, resulting in $16.4/hour. Using a less efficient car raises the cost/hour, as to be expected. Halving the initial mileage to 16.5 mpg doubles the cost/hour so you could be looking at $33/hr with a vehicle that gets 16.5 mpg at 50 mph.

    So, what is your value on time saved? And what if there are several people in the car? And does the value depend on where you’re going (work versus vacation)? The hourly cost savings described above for slow driving are significant, but might not be worth it to many people even though they haven’t explicitly quantified the cost. However, safety, flow of traffic, and the chance of a ticket are other incentives for keeping speeds down.

    • except of course at the 70 and 80 or even the 60 and 50mph case you might get pulled over for speeding; and there goes any time savings, fines, and insurance rate going up 😉

  2. Er, savings as defined above is a misleading metric. As a commentator pointed out, Liters per 100 km removes the confusion for consumers. The graphic about “savings by switching” is so self-serving. OF COURSE the worse vehicle saves more.

    Yes, **savings** per 100 km driven can be higher ONLY because of the pathetic base. IF the bottom line were simply to save total fuel, then it begs the question why we would give subsidies to a massive SUV that turned hybrid and saved some fuel (maybe saving even MORE fuel on a relative basis) instead of either the el-cheapo smaller vehicle or a mid-sized hybrid and quite efficient vehicle.

    The below anecdote is one I use a lot in lectures (disclaimer: I am a scholar and advisor on technology-policy, esp. energy):
    A man comes home an excitedly tells his wife, I saved $2 and also got some exercise. I chased after a bus but didn’t get on. She replies, “You fool, you could have chased after a taxi and saved $20.”

    I hope Ms. Wolfram can also elucidate the broader question above, instead of the narrow, technical logic about savings (which are always relative!!!).

    Rahul Tongia

    • Thanks for the comment. Don’t forget that many people like to drive SUVs, so there are real savings from making them more fuel efficient.

      • IF the base assumption is that an SUV person will stay with an SUV then of course there is fuel saving. But in fact this is a red-herring-like comparison, and the MPG illusion is too. If a consumer has decided that (s)he ‘must have’ an SUV-like vehicle then pick the best MPG in that category that meets the requirements. To even pose the question in the way that is done here is misleading.
        As another post pointed out, upgrading an existing older-house, with poor energy conservation capabilities is generally more sensible than tearing it down and building a super efficient house.
        It is the bogus arguments that confuse people. Academics: PLEASE do your research on realistic scenarios, not contrived and meaningless ones. [a recent article in the economist regarding pay was based on undergraduate student ‘volunteer’ as representing ‘human behavior’ of highly-paid people. Invalid. The underlying experience and situation will make the outcome non-representative/ meaningless for testing a hypothesis]

      • Hi Catherine,

        Thanks for the reply. I think you’ve raised a much broader and (IMHO) more important question – not re. the MPG illusion but WHY people want to drive a huge SUV. YES, you are right, there are real fuel savings in big vehicles, but that is *relative* savings. We really should be interested in ABSOLUTE consumption. Thus, we should give tax breaks to the person who buys a Smart4two, or Honda Fit, etc. Not a hybrid Escalade/Excusion/etc.

        Do you think a schema of target MPG plus/minus subsidy/penalty makes sense?

        E.g., All vehicles must be, say 35, MPG. Those better than this, get credits, those worse, pay a penalty. And it would be a moving target over time, based on both state of technology and actual vehicle (fleet) mix. Of course, there are a few issues. First, people who *need* a truck would complain. We could make exemptions for registered vehicles used to haul goods, i.e., true construction vehicles. Second, we need to avoid the massive loopholes ala CAFE, such as E-85, “light trucks”, model availability by manufacturer (rather than sales), etc. If people with minivans complain, since they may legitimately be carrying 6-7 passengers (i.e., we want to encourage car-pooling), then there would be two ways to deal with that. We technologically could start measuring vehicle use based on kilometers-persons per liter (but that would be intrusive, and require sensors on seat occupancy), or we could have a **small number of** categories of vehicles and set different targets, e.g., one for 5 seaters, 7 seaters, cargo of 1 ton, etc. This avoids the entire my sedan is smaller/bigger than yours debate.



  3. In France they measure mileage as liters per 100 km, not km per liters. Then you have no MPG illusion.

    • But I’m thinking now that in the discussion currently in France about home renovation the same effect happens.
      There’s more gain in enhancing a little a very badly insulated house that by trying to update a more recent one to the best standard.

  4. EPA has already solved the problem in a much more direct and common sense way. The new labels estimated annual expenditure on gasoline, so any consumer who can do subtraction will know the gaoline savings. Since they also have the sticker price at hand, they can decide whether the gasoline savings is worth the differece in the cost of the vehicle (if there is one).

    • Thanks for the comment. There’s definitely an open question about whether the new EPA labels have solved the MPG Illusion. They still feature MPG numbers and many people still think in terms of these numbers. We’ll see!

  5. Yes, but your assumption that the two vehicles would be used the same is incorrect. A person would likely not drive both the same number of miles per year.

    I have gone through the same logic and reached the same conclusion that you do. From a societal perspective, it is better to get people to choose to upgrade their “gas guzzlers” than to replace an efficient vehicle with a more efficient one. But from a consumer’s perspective, if a consumer has a vehicle that gets really good miles/gallon, they going to use it more.

%d bloggers like this: