Some good news: even more realistic MPG estimates and it seems like they're starting to prominently display tailpipe CO2 (g/mi) emissions for newer cars. Hopefully the European (NEDC?) counterpart follows suit to use more realistic testing procedures.
CARS.COM — Shopping for a 2017 model-year car? Beware: The EPA has changed how it calculates window-sticker gas mileage for 2017 in an effort to better reflect real-world mileage, according to the agency and automakers. And the agency has announced almost nothing about it.
That means that more than a quarter of all new cars could see a reduction in their EPA combined mileage. As a result, new-car shoppers will see mileage numbers on the window stickers of some 2017 model-year cars that are lower than their 2016 predecessors, even though the cars have no mechanical differences. And, at first, shoppers will not be able to make apples-to-apples comparison between 2017 and 2016 versions.
Related: Loopholes and Fuzzy Math: The Tangled Science of MPG Ratings
This recalculation means shoppers cannot compare 2017 mileage numbers to their 2016 predecessors, but Cars.com has learned the EPA plans to address this soon with updates to its website that will allow shoppers to compare the new numbers with models as far back as the 2011 model year.
Real-world mileage shouldn't change for drivers, but the EPA ratings could fall noticeably in the agency's biggest adjustment since it issued sweeping changes to window-sticker mileage ratings nearly a decade ago. Why will the ratings fall? Because the EPA is changing the inputs into their formula.
How Many Cars Could This Affect?
A Cars.com analysis of 2017 model-year EPA data indicates the change could meaningfully affect EPA numbers on nearly 30 percent of all new cars. Among the top 75 best-selling vehicles through July, we identified models with no mechanical or aerodynamic changes between the 2016 and 2017 model years. Of that group, we found 84 variations of individual vehicles (the V-6 versus the four-cylinder, for example).
Twenty-three of those, or 27.4 percent, received a lower EPA combined mileage rating for the 2017 model year. Most of those saw a drop of a single mpg, and included versions of popular cars including the Honda Accord, Toyota RAV4 and Hyundai Sonata. Versions of the Volkswagen Passat and Toyota Camry Hybrid fell by 2 mpg combined.
Our results indicate a lot more changes than the EPA observed in its 2016 model-year analysis. Byron Bunker, director of the compliance division at the EPA's Office of Transportation and Air Quality, told us the agency ran 1,209 fuel-economy results from the 2016 model year through the agency's updated calculations. Just 15 percent of that group — 186 — fell by 1 mpg in combined mileage. Thirteen, or about 1 percent, increased by 1 mpg. The combined mileage stayed the same for the remaining 84 percent.
Of course, EPA figures are seldom whole numbers, so rounding can mask — or exaggerate — actual changes. Some cars may actually see their values increase, the EPA notes, but the difference is "so small and subtle" that the perceived change could be more the result of rounding.
"Some of those changed because they were 26.4 before and now they're 26.6 after," Bunker said. "When you change these methodologies you'll move some more for rounding."
Why the Data Changed
When the EPA issued its broad update of fuel-economy ratings back in 2008, it required automakers to go from two testing cycles to five cycles, adding tests for things such as cold-weather operation, higher speeds, faster acceleration and air-conditioning usage. It also tried to reflect other real-world factors such as road conditions and wind. The additional cycles had been used for federal emissions certification programs since the 1990s but not specifically for mileage tests, the EPA said.
Still, a lot of automakers stuck with two-cycle testing and applied what's called a "curve fit" — a mathematical correction that aligns the results from two cycles to a hypothetical five-cycle test. The EPA developed the math from data it gathered from 2002 to 2006.
"The manufacturer can choose to run all five or run just the city-highway [cycles] and just apply the curve fit," explained Ryan Harty, who manages Honda's environmental business development office. "Not all labs have that [five-cycle] capability, and I think EPA recognized that."
Today, the EPA has a lot more data.
Specifically, the agency has 847 new sets of fuel-economy test results from the 2011 through the 2016 model years. That's what the EPA said in a June 2015 letter to automakers that explained the coming changes.
The agency maintains that the new data better reflects today's fuel-efficiency technologies, and it used that data to change the numbers it used to calculate new ratings. It's an overall improvement over the agency's 2006 analysis and reflects "a high degree of precision," the letter said.
Comparing Old and New
But the shift also means you can't compare window-sticker mileage for cars in different model years, Bunker admitted.
Many consumers might be trying to do just that. Take Bill Meyer, for example. In just one model year, the 2017 Passat he was looking at — with Volkswagen's turbocharged four-cylinder engine — tumbled 4 mpg in its EPA highway mileage rating. The four-cylinder 2016 Passat had a fuel-sipping 25/38/29 mpg city/highway/combined rating from the EPA. For 2017, the same car dropped to a more-lackluster 23/34/27 rating, even though the two cars are mechanically identical.
Meyer, an attorney in Kansas City, Mo., planned to take Volkswagen's pending buyout offer on his 2013 Passat TDI, involved in the automaker's massive emissions scandal. But he liked the Passat and fancied getting a gasoline one, which he thought was pretty efficient — until he saw the 2017 EPA ratings.
"When I looked at the window sticker, I thought, 'Wow, this is taking a big hit,'" Meyer said. "Why is there such a mileage discrepancy? Thirty-four is not nearly as interesting to me as 38. So it's making me rethink my decision."
Mileage labels on 2017 model-year cars are supposed to signal this shift by the EPA in their fine print, the agency said. Indeed, the window sticker for a 2017 Kia Sorento we recently tested said in tiny letters on the fuel-economy portion: "These estimates reflect new EPA methods beginning with 2017 models."
But the practice seems hardly standardized. The fuel-economy label on a 2017 Hyundai Santa Fe Sport we tested around the same time, for example, had no indication that any change had been made.
The EPA handled a larger change differently in 2008. Back then, notes John Voelcker, editor of Green Car Reports, the agency "took their entire set of published numbers, at least back to the '80s, and recast them all," he said. Consumers could compare new cars versus old cars — a big help for shoppers. Not so for the 2017 adjustments.
The EPA says it plans to update its website "over the next few weeks," Bunker said, with a way for consumers to compare mileage across different model years as far back as 2011. The agency "will also include details on the latest methodology update so consumers understand the changes," he added.
The EPA notes that most fuel-economy labels are likely to go unchanged except on higher-mileage vehicles, where the change is "most pronounced." Indeed, Cars.com's analysis shows that the cars that lost the most mileage for 2017 — 2 mpg apiece in EPA combined ratings — were the Toyota Camry Hybrid and four-cylinder Passat. Toyota and Volkswagen attributed the decrease to the new calculations, even as the EPA says the Passat’s drop came from larger mechanical changes.
Those differences shouldn't equate to any change in real-world mileage for owners, but the change in EPA ratings could alter the perceptions of a car shopper considerably. That's because the range between the most- and least-efficient cars in a given segment is often just a few mpg.
After we explained the EPA's changes to Meyer, he looked up a number of other mid-size family sedans and saw similar drops for 2017. Highway numbers "were all down a couple of mpg for 2017," he later wrote in an email. "It looks like the Passat's drop is in line with most other mid-size cars."
EPA Updates Older MPG Ratings to Align With 2017 Calculations
CARS.COM — The EPA has updated its website, Fueleconomy.gov, to display gas mileage for older cars under the same adjustments the agency has applied for the 2017 model year. Consumers previously couldn't make apples-to-apples comparisons between 2017 models and previous model-year vehicles, as the 2017s followed different mileage calculations.
The move comes days after a Cars.com investigation uncovered the incompatibility between the EPA's older methodology and its new calculations, intended to better reflect real-world driving. The calculations affected virtually all cars for 2017 but weren't reflected in previous model years, which rendered the vast majority of mileage comparisons between 2016 and 2017 models inaccurate. Further complicating matters, the EPA allowed automakers to voluntarily adopt the new methodology a year early — i.e., the 2016 model year instead of 2017. A handful of cars did just that, according to the agency.
At the time, EPA officials told us the agency would update its website "over the next few weeks" so consumers could make apples-to-apples comparisons. The agency says it planned the move before Cars.com investigated the matter.
Now the EPA has posted an explanation of the 2017 ratings changes with a separate mileage comparison page for you to compare the old and new mileage ratings on any vehicle. More importantly, the agency has revised existing mileage ratings for the 2011-to-2016 model years to reflect the new calculations. If you compare mileage figures for any cars from 2011 to now on the EPA's website, those figures should come from the same calculations for more accurate comparisons.
Used-car shoppers should note that models from 1984 to 2010 still have numbers using the prior methodology, which the EPA overhauled for the 2008 model year. If you stack up the mileage for an older car against anything from 2011 onward, you won't have an apples-to-apples comparison because the mileage figures come from separate calculations. (But you can go to the EPA's 2017 mileage comparison page, look up the newer model's unadjusted original mileage and get a consistent comparison with the older model.)
The EPA says the update amounts to a slight difference in mileage, with some cars falling by 1 mpg and "a small number" falling 2 mpg. A Cars.com analysis of 84 fuel-economy labels for popular cars that carried over with no mechanical or aerodynamic changes from the 2016-to-2017 model years indicated the same trend, with 23 labels (27.4 percent) falling 1 or 2 mpg in EPA combined mileage.
"Not all changes to MPG values in 2017 models can be attributed to the updated calculations," the EPA cautions. Indeed, many cars have other updates, from mechanical changes to a full redesign, that could have affect mileage for better or worse.
It's unclear if this affects what you'll see on existing window stickers, as new 2016 models that are still on dealer lots could still have window stickers that reflect the old calculations. Asked if the EPA had the authority to force dealers to reprint the stickers, an agency spokesperson did not provide an immediate answer.
Our new fuel testing methods are allegedly proposed for next year too, although some manufacturers, Peugeot and Citroen to name a few, are already compiling real world figures in some sort of loyalty bid to consumers.
Now the UK isn't part of the EU anymore, I'm wondering if we'll still be using the European testing methods or whether the UK will devise it's own ideas, there's alot of uncertainty about everything now, and I noticed our fuel jumped 5 pence per litre this weekend too. Great...
I don't want to sound overly obvious, but when I was shopping for a new car, I went to Fuelly.com to see what real owners were getting in real fuel economy. I based my purchase decision on that, and it turns out I chose wisely.
I'm always playing the "how good fuel economy can I get" game, every trip I take, so I not only handily bested the EPA numbers, I compare favorably to others who drive the same vehicle.
WRT the EPA's new methodology, I'm all for something that gives us more of a "real world" representation. That applies to fuel economy and emissions.
2015 Audi Q5 "Progressiv" + S-Line + Scuba Blue, 3.0L V6 TDI
(Highest fuel economy for all Audi Q5s on Fuelly!)
Yep I tend to look at the lab numbers, then compare them to fuelly users. Or I just take away 25% of the average figure, given that's the average discrepancy between the test cycle and the real world figures. Sadly, many of the cars I want or consider are too rare or too new to have been tracked on fuelly, so I have to make my own assumptions.