Accuracy of EPA Gas Mileage Estimates

Consumer Reports just published a series of articles on EPA mileage estimates and how they compare to real-world results. As it turns out, 90% of all vehicles surveyed got worse than expected mileage, with the biggest discrepancies (up to 50% below EPA estimates) occurring in city driving. Even those thrifty hybrids are getting in on the action, with city mileage falling up to 19 MPG short of the EPA estimates. Most troubling of all is that the gulf between the EPA values and real-world results appears to have been growing over time. In short, while the EPA estimates allow you to make comparions amongst makes/models, they’re not very good indicators of actual mileage. Until the system is overhauled, CR suggests that you discount the EPA estimates for city travel by the following amounts: conventional cars and trucks, 30 percent; larger hybrids, 35 percent; diesels, 36 percent; smaller hybrids, 42 percent.

27 Responses to “Accuracy of EPA Gas Mileage Estimates”

  1. Anonymous

    EPA – DOE Issue 2005 Vehicle Fuel Economy Ratings

    Fuel economy estimates are determined by averaging numbers gathered through tests conducted by manufacturers and verified by EPA. Vehicles are tested in a controlled laboratory setting and the results are adjusted to reflect actual driving conditions. All vehicles are tested in the same way so consumers can compare the results when choosing a vehicle type or class. The mpg estimates appear on window stickers on all new cars and light trucks prior to sale. To ensure these estimates continue to remain as reliable as possible, EPA has begun a collaborative process with various stakeholders to update its testing procedures, and plans to propose appropriate changes in the next year.

  2. Anonymous

    Hmmm. This article made a thesis that has not been answered – the creep of estimates over time. Not quite the same as the accuracy issue. We know the govt has been doing this with inflation data (per and we know that big money (Goldman Sachs, etc.) owns govt. This requires that an analysis must disentangle actual economy improvements from data creep. Any thoughts?

  3. Anonymous

    There is a group of us out there that actually NEED a pickup sized vehicle for our work. The best I’ve been able to do is a Chevy WT (work truck) rated at 22/18 with a 4.2 (or 4.3 rounded way up for the market) engine. Now, I only need that size/load capacity/power for about 20% of the time. The other 80% I could use a smaller, more efficient vehicle but the logistics of using two vehicles at the right times is prohibitive due to nationwide travel and cost of ownership of two vehicles. I am strapped to a $3,000/year fuel cost because, if I can’t perform that 20% of the time, I’m out of a job. I strongly invite suggestions.

  4. Anonymous

    FWIW, making sure your tire pressure is at the maximum is very important in the estimate. Along with proper vehicle maintenance.
    Our Sienna experienced much less than what the EPA estimate was. However, after we upped the pressure to the maximum 51 psi, the average mileage has already started climbing (it keeps track of this for you).
    If a tire is not at the maximum psi, the tires will often wear on the outsides of the tire, and if it’s too much pressure, they will wear off faster on the center portion of the tire itself.

  5. Anonymous

    Us too. We actually save our old copies in a file and then pull them out when we have a buying decision. Well worth the subscription (though you can get the info free from the library).

  6. Anonymous

    CR is a great resource for buying almost anything. We’ve used it for years and always refer to it before buying anything. In fact, we just used it this weekend as we’re looking for a new dishwasher.

  7. Anonymous

    I found this at the top of EPA’s fuel economy page:

    IMPORTANT REMINDER: EPA’s Fuel Economy estimates are designed to allow consumers to comparison shop. Your fuel economy will almost certainly vary from EPA’s fuel economy rating. This is based on a number of factors, such as how you drive and maintain your vehicle, use of air conditioning and other accessories, weather, road conditions, and other factors. EPA is currently evaluating ways to improve our fuel economy estimates so they better reflect real-world driving, and we plan to propose changes in the coming year. For more information on how your fuel economy can vary, or tips to improve your fuel economy, please visit Your MPG Will Vary and Gas Mileage Tips on

  8. Anonymous

    I get just as good or better mileage than rated on my 93 Sentra. People are probably complaining about these EPA results because they drive the hell out of their cars. Of course you’re not going to get 30mpg city if you’re flooring it all the time. I can easily get 30mpg city b/c I don’t drive like a moron of the common extreme go to an extreme stop attitude.

  9. Anonymous

    Even the EPA itself doesn’t reccomend their estimates for more than an “apples to apples” comparison. The virtue of those numbers is that they’re based on a standardized set of tests applied the same way to every vehicle. Your mileage will vary based on all the things no statistic on the sticker could ever cover accurately (road conditions, driving habits, vehicle load, etc), but the EPA tests provide a method for an accurate comparison of the factors in fuel economy that are built into vehicle design.

    So yeah…don’t plan your monthly budget down to the penny based on plugging your new car’s EPA mileage into a calculator and doing the arithmetic…but that doesn’t mean the numbers aren’t good for anything.

  10. Anonymous

    I had a Grand Cherokee achieve substantially better mileage than the EPA rating. Do the math before buying a hybrid – at $3.30/gal for fuel, it’ll take 36 years to break even if I replace my ’91 Corvette with a Prius, and that is going by the overly optimistic EPA rating.

    Oh – CHECK THE AIR IN YOUR TIRES. It makes a big difference (and don’t forget the spare).

  11. Anonymous

    To expect a single number (one for each City and Highway) to reflect the driving styles, passenger loads, road conditions, etc. is assinine. People are just looking for faults where I’d say the gov’t is actually doing a good job. As for the hybrids I can only guess that the EPA’s statistical forecast algorithm thingamabobber (technical term) is tuned for gasoline-only cars. It has, after all had over 20 years to tune it.

    My ’00 Saturn SC2 is rated at 27/37. My commute is about 4 miles city, 8 miles highway and I get between about 30 and 35MPG. I drive aggressively some days, and almost always accelerate quickly (no spinning) at a green light. I imagine I could easily get over 35MPG if I didn’t go quickly from a dead stop, or use engine braking on the highway rather than causing traffic by flashing brakelights for a 5MPH speed adjustment.

    My $0.02

  12. Anonymous

    I totally agree with Matt B. I have a 2003 Nissan Altima 2.5, and I am getting EPA or better mileage. You just have to drive carefully. Many (most?) drivers seem to be trying very hard to get the worst mileage they can…

  13. Anonymous

    In my experience with a 93 sentra, 2000 and 97 maxima and 2004 mpv I’ve found the EPA actually quite accurate. I’d say it’s optimistic by as much 10-20% on the MPV, which was the one furthest off epa, but with my 2000 maxima the EPA was, if anything, pessimistic, as I’ve occasionally gotten better than EPA on the highway. The other cars it was reasonably close. Remember, highway would be at 65-70, not at 80 with the AC on. City would be gentle driving, not hammering it.

  14. Anonymous

    With hybrids I’ve heard the EPA estimates are way off and consequently customers are complaining. But with diesel I’ve read, heard, and experienced the opposite. Many drivers of the newer VW TDI’s have experienced better mileage than what the EPA estimates. In my case, a much older VW diesel (1992), my actual mileage is bang on the EPA figures

  15. As inaccurate as they are, there is still some value to these estimates in that you can make relative comparison between different makes/models. Clearly, a car that is rated for 35 MPG will get better mileage than one that is rated for 15 MPG, even if neither one approaches their actual rating.

  16. Anonymous

    Nice to know that a government agency can’t give consumers accurate information. What is the point of epa estimates for fuel mileage if we can’t even take them at face value and have to “discount” them by certain percentages?

    I’ve always liked how Car & Driver (or maybe it was Motor Trend), did long term tests (1 year or so) and had an accurate true cost information, which included the mpg of the car over that time period.

Leave a Reply