EPA car mileage calculations before 2009 were wrong, making it harder to sell used cars

The National Journal reports:

That’s because the government’s calculations have changed since 2008, and cars manufactured before that have been downgraded about 2 to 3 miles per gallon from their original advertised fuel economy.

“The older test was biased a little high,” said Doug Greenhaus, chief counsel for environment, health, and safety at the National Automobile Dealers Association. “In almost every instance, [the new test] is a lowering.”

The new calculations add factors such as acceleration, air-conditioner use, and outside temperatures.

Read more…

21 thoughts on “EPA car mileage calculations before 2009 were wrong, making it harder to sell used cars”

  1. E-15 won’t hurt your car. That is just junk science in the name of anti junk science.

    But adding the ethanol is also not designed to help new cars either. Los Angeles and other smoggy areas of California have a 5.6% ethanol requirement so that the fuel can be at least 2% oxygenated. This allegedly allows older cars whose smog equipment is no longer up to snuff, to burn fuel more cleanly with less carbon monoxide and nitrates in the exhaust.

    In Los Angeles especially, lots of cars are driven by folks without licenses who don’t bother to go through registration either. If caught – because they are usually illegal the car is just handed back to them (so the police can’t be accused of being racist) so this might actually help a bit there.

  2. To presume that manufacturers wouldn’t improve their products unless forced to by law is over simplifying the situation. What it does it take the choice of which advances to focus on out of the hands of the interested parties (manufacturers and consumers) and put it into the hands of disinterested parties. That is, if the majority of buyers felt those improvements were a priority than the market would have brought them out without intervention. The fact that intervention was deemed necessary is evidence that few people actually wanted it.

  3. I’m not into advances at gunpoint. Yes, advances have been made, but they don’t justify tyranny.

  4. One last remark – the EPA’s gotten many things wrong, and I’d like nothing better than to see them funded at not too much over their current shutdown-level permanently, but we owe much of the reliability of modern cars to their extended emissions warranty mandates; having to warranty emissions-related components for 50K or more miles was a big driver in the adoption of weathersealed electrical connectors and other components in cars, and that’s been a HUUUUUUUge advance all by itself.

  5. Oh, and the Swedes…Saab and Volvo along with the Germans also adopted D-Jetronic analog EFI in the very early ’70s, the reliability problems were bad enough that in most applications they replaced them with Bosch K-Jetronic mechanical fuel injection, which was a complete evolutionary dead-end and had its own set of issues, most started drifting back to later electronic systems in the mid ’80s but VW persisted with a bastardized electronically-controlled K-Jet derivative until nearly the end of the decade IIRC.

  6. 1970s? The first analog electronic injection systems showed up in the 1950s, Bendix bailed on that one and sold the rights to Bosch, who developed it through the ’60s – first becoming common in late ’60s VWs and a few other Germans. Bosch D-Jetronic fuel injection was common in many European cars mid-’70s, and Hitachi licensed it for the Japanese industry. Sensor and wiring reliability and lack of learning/adaptation were problems.

    By the ’80s you had hybrid analog/digital and the first digital fuel injection systems coming in, as well as broader-scope controls (ignition), etc. By mid-decade everything had oxygen-sensor feedback controls for regulating fuel mixture, most had fully electronic fuel injection, half had integrated ECU-controlled ignition timing, and electronically-controlled automatic transmissions were becoming common.

    The first cheat code I’m aware of was in the late ’80s.

  7. But wait EPA mandates the use of additives to reduce pollution in gasoline, first MTBE then ethanol when MTBE was found to be a pollution problem. See web page –


    Why does EPA not test with the gasoline it mandates? Amazing Autos pass emission test with gasoline without additives EPA mandates on most of us. Additives that EPA wants to increase the percentage used that can damage existing autos (E15).

  8. No no no! The EPA test was not designed to test fuel econom! This is the EPA. The EPA test was to check to see that autos were meeting emision standards. People got interested in fuel economy during the 70’s fuel price jumps. Someone at EPA realised their emision test data could be used to figure how much gasoline was used during their test. Then the EPA released the figures as fuel economy numbers. Now EPA has changed their testing for emisions, so now new fuel economy numbers are figured. The EPA test is not realworld driving. Just look at the numbers for hybrids. Oh, the EPA does not test all cars, remember some car maker got caught doing the test wrong ans so had wrong fuel economy figures.

  9. In the 80’s cars started having small electronic controls for fuel injection, and the instrament panel – Remember the talking cars “The door is ajar” wonder why they don’t do that any more.

  10. It could have been an honest error, but this isn’t the first time time they monkeyed with the fuel efficiency requirements. They did so in the 1990’s as well. In fact I bet they are always changing it.

    And no they don’t crowd source to get the EPA mileage data.

  11. @JEM

    …build ‘cheat’ code into their ECUs…

    Seriously? Just how many 1970’s and 1980’s vehicles do you know of that had ECU’s?


  12. Further, the EPA dyno test was based on an inertia-weight setting on their dynamometer, and getting the vehicle into a lighter weight category would reduce the resistance the vehicle had to operate against on the dyno thereby providing a small MPG boost. Automakers would often make small spec changes to the car – for instance, a smaller gas tank, one case that comes to mind was that the V6 version of the Ford Contour/Mercury Mystique (Mondeo elsewhere in the world) got smaller brakes in the US than in other markets.

  13. This is not an error, it’s an artifact of how the EPA’s test procedure worked.

    They tested under one set of operating conditions originally developed back in the ’70s, now they’ve changed those test conditions to (they believe) better represent current driving conditions.

    The old EPA test protocols were anomalous enough to real-world driving that clever automakers could (and did, on occasion) build ‘cheat’ code into their ECUs that could identify the conditions that made up the test and go into a high-mileage mode (that might not meet emissions-test criteria, and might not be real-world driveable, but it would optimize the vehicle’s fuel economy in that one particular test.)

  14. This could have been an honest error, and I can also imagine how it can depend on the sample. If any of the mileage data was crowd-sourced, the result could be biased depending on the reporter. When I drive my father’s car, I achieve a 50% higher economy. Generally, today, the populations driving the cars that are older than 2009 and those driving the newer cars are different populations. I suspect the former is more concerned with economy, and there is nothing green about it, faux ou authentique.

    I also suspect people’s driving habits can have a much greater effect than ambient factors.

Leave a Reply

Your email address will not be published.