Why Your MPG Is Lower Than the EPA Rating

A window sticker says 32. You have been driving the car for a year and the dashboard says 26. The first thing worth knowing is that this is the expected outcome, not a symptom, and that the agency printing the sticker says so in writing. EPA's own guidance is that the ratings “are a useful tool for comparing the fuel economies of different vehicles but may not accurately predict the average MPG you will get.”

The label is a comparison instrument. Every car is measured the same way so that two stickers can be held side by side, which is a genuinely useful thing and not the same thing as a forecast for your commute. This guide covers what the number on the glass actually is, the four ordinary things that thin it, and how to price the gap for your own car rather than wondering about it.

The combined figure is not the average of the other two

Most stickers print three numbers: city, highway, and a larger combined figure in the middle. It looks like the average of the first two. It is not. Under 40 CFR § 600.210-12, combined fuel economy is the harmonic mean of city and highway weighted 55/45:

combined = 1 ÷ (0.55 / city + 0.45 / highway)

city 26, highway 35

  arithmetic average   (26 + 35) ÷ 2                    = 30.5
  the label's method   1 ÷ (0.55/26 + 0.45/35)          = 29.4

A full MPG lower, and lower for a good reason: fuel economy is miles per gallon, but what you actually spend is gallons per mile. Averaging the miles-per-gallon figures over-weights the efficient half of the driving. The harmonic mean averages the consumption instead, which is the thing that costs money.

This matters for your comparison before anything else does. If the tank you measured was almost all highway, compare it against the highway rating. Compare a highway tank against combined and you are measuring your driving mix as much as your car — and 55/45 is very likely not your mix.

The four ordinary things that thin the number

None of these is a fault. All four are published by the DOE and EPA on fueleconomy.gov, with figures attached.

Speed

Fuel economy usually falls rapidly above 50 mph, because drag rises with the square of speed. The DOE puts the penalty in money: each 5 mph over 50 is roughly like paying an extra $0.29 a gallon. A driver who cruises at 75 rather than 60 is paying something close to $0.87 a gallon more than the sticker assumed, and paying it at the pump like everyone else.

How you drive it

Aggressive acceleration and braking cost roughly 15% to 30% at highway speeds and 10% to 40% in stop-and-go traffic. That is the widest band on this list, and the only one entirely under your control.

Cold

A conventional gasoline car loses about 15% in city driving at 20°F versus 77°F, and as much as 24% on short trips of three to four miles. The engine spends the whole trip getting to temperature and never arrives. Anyone comparing a January tank against a sticker is comparing two different seasons.

The fuel itself

Nearly all gasoline sold in the US is E10. It contains less energy per gallon than ethanol-free gasoline, giving 3% to 4% fewer miles per gallon. This one is not a driving habit and not a maintenance item. It is in the pump.

Stacked, they explain most of the gap

These do not compound cleanly — the studies behind them measure different things under different conditions, so multiplying them is arithmetic rather than science. But it shows the scale of what a plausible winter commute does to a 32 MPG rating:

32 MPG rated
  × 0.85   cold city driving at 20°F
  × 0.90   a moderately heavy right foot
  × 0.97   E10 rather than ethanol-free
  = 23.8 MPG   —  26% below the sticker, nothing wrong with the car

A car reading a quarter below its label in February, on short trips, is behaving exactly as the published figures say it should. What deserves investigation is not a gap but a change: a car that returned 28 all year and now returns 22 on the same route in the same weather is telling you something the label never could.

Don't settle the question with the dashboard

The trip computer estimates consumption from how long the fuel injectors are held open and how far the wheels have turned. It never sees a pump. It is a good instrument and it is not a measurement of what you paid.

The measurement is the one the DOE and EPA publish for drivers to do themselves: fill the tank completely and write down the odometer, drive, then fill it completely again and divide.

miles driven = odometer now − odometer at the last fill
MPG          = miles driven ÷ gallons added AT THIS FILL

Which fill you count is the whole trap. The gallons that matter are the ones added at the endof the measured distance. Topping back up to the same point replaces exactly the fuel burned since the last time you topped up to that point — which is why the method works without ever knowing the tank's size or how full it was. The gallons you bought at the first fill measure nothing; they only describe how empty the tank had got before the distance started.

Run it three or four times before trusting the answer. The weakness of the method is entirely in the fill: auto-shutoff triggers at slightly different points depending on the nozzle, the angle of the car and how fast fuel is flowing, and misjudging a 13-gallon fill by a third of a gallon moves the answer by more than 2%. Fill at the same pump, stop at the first click every time rather than topping off, and measure over longer distances. A repeated error cancels out; an inconsistent one does not.

What the gap actually costs

A percentage is not a decision. Dollars are. Take the 32 MPG rating, a real reading of 26.4, gasoline at $3.19 and 12,000 miles a year:

at your 26.4 MPG   $3.19 ÷ 26.4 = $0.1208/mi × 12,000 = $1,450 a year
at the rated 32    $3.19 ÷ 32.0 = $0.0997/mi × 12,000 = $1,196 a year

the gap                                                    $254 a year

$254. Real money, and also not the emergency the 18% shortfall sounded like. That is the useful shape of this answer: a number you can weigh against what it would cost to change anything. If the gap is mostly speed, it is worth something. If it is mostly February, it will fix itself in April.

The True MPG Calculator does exactly this division — your real MPG, the distance from the rating, and the annual cost of that distance at your own pump price. Nothing in it is a national average.

Once you have a real number, use it everywhere

The measured figure is worth more than the correction it makes. Every other cost question about the car takes MPG as an input, and each one has been silently using the sticker:

• Commute Cost — the drive to work, priced on the MPG you actually get on that drive rather than a rating from a lab.
• Trip Fuel Cost — a single drive, split across everyone in the car.
• EV vs. Gas Cost — the comparison where an overstated MPG makes the gas car look better than it is. The matching error on the electric side is covered in EV vs. Gas: What a Mile Really Costs.
• True Cost per Mile — where fuel takes its place beside depreciation, insurance, upkeep, registration and interest, and usually turns out to be smaller than any of the first two.

That last one is worth sitting with. The $254 a year you just found is real, and on most cars it is a rounding error next to what the car loses in value over the same year. Both are worth knowing. Only one of them is worth changing your driving over.