See how EPA fuel economy ratings are produced in a laboratory, from the five test cycles and dynamometer runs to the combined window-sticker number.
How does the EPA test fuel economy?
EPA fuel economy ratings come from laboratory tests on a dynamometer, not from driving on public roads. Since the 2008 model year, five test cycles have been combined into the city and highway numbers on the window sticker.
A cycle, also called a schedule, sets the vehicle speed for every second of a test. The driver follows a display that compares the required speed with the actual speed. The dynamometer adjusts the effort needed to turn its rollers to imitate wind resistance and vehicle weight. fueleconomy.gov compares the setup to an exercise bike.
The five cycles behind a modern rating
Each cycle is designed around a different kind of trip. The table below uses the figures published in the detailed test information.
| Test cycle | Top speed | Average speed | Simulated distance | Test time | Lab temperature |
|---|---|---|---|---|---|
| City | 56 mph | 21.2 mph | 11 mi | 31.2 min | 68 to 86 F |
| Highway | 60 mph | 48.3 mph | 10.3 mi | 12.75 min | 68 to 86 F |
| High speed | 80 mph | 48.4 mph | 8 mi | 9.9 min | 68 to 86 F |
| Air conditioning | 54.8 mph | 21.2 mph | 3.6 mi | 9.9 min | 95 F |
| Cold temperature | 56 mph | 21.2 mph | 11 mi | 31.2 min | 20 F |
The city cycle starts with a cold engine and includes 23 stops. Idling takes 18% of its time. The highway cycle has no stops and begins with a warmed-up engine. The high-speed cycle adds harder acceleration and braking up to 80 mph. The air-conditioning test runs at 95 F with the cabin set to cool. The cold test runs at 20 F with heat and defrost on.
Who runs the test, and on which vehicles
Manufacturers test their own vehicles, usually preproduction prototypes, and report the results to EPA. EPA reviews the data and confirms about 15% to 20% of the results with its own tests at the National Vehicles and Fuel Emissions Laboratory. The EPA testing page describes that confirmation share as approximately 15%.
A manufacturer does not have to test every version. It tests one representative vehicle for each combination of loaded vehicle weight class, transmission class and basic engine, according to fueleconomy.gov.
Some vehicles are exempt. From 2011 onward these include pickups with a gross vehicle weight rating over 8,500 pounds and a bed at least 72 inches long, cargo vans over 8,500 pounds, and passenger vehicles rated at 10,000 pounds or more. Before 2011, large vehicles over 8,500 pounds were exempt.
Fuel is measured from the exhaust
For gasoline, diesel and natural gas, a hose collects the exhaust during the test. The carbon in that exhaust is measured and converted into the amount of fuel burned. EPA says this is more accurate than reading a fuel gauge. Vehicles that do not run on carbon-based fuel, such as electric cars and fuel cell vehicles, use a different method.
How five cycles become one window-sticker number
The results are weighted into city and highway estimates. The combined figure on the label weights city driving at 55% and highway driving at 45%, as the label guide explains. The annual fuel cost on the label assumes 15,000 miles a year and a projected gasoline price. Neither figure is a measurement of your driving.
Before the 2008 model year, only the city and highway cycles were used. A 2007 rating and a 2026 rating were therefore not produced in the same way, even when both use the same units.
What the laboratory test leaves out
The test assumes a moderate temperature, flat ground and about 300 pounds of passengers and cargo. Four-wheel-drive vehicles are tested in two-wheel drive. Roof cargo is not fitted. The high-speed cycle stops at 80 mph, so sustained driving above that speed is outside the test.
Those gaps help explain why real numbers vary. fueleconomy.gov’s factor list puts the effect of aggressive driving at roughly 15% to 30% at highway speeds and 10% to 40% in stop-and-go traffic. Running the air conditioner on maximum can cut mpg by about 5% to 25%. Gasoline with 10% ethanol lowers it by 3% to 4%. Cold weather, short trips, poorly inflated tires and extra weight all pull the number down. A new engine may also take 3,000 to 5,000 miles to reach its best economy.
Use the rating to compare vehicles, not to predict a single tank. To measure your own figure, see how to calculate your actual gas mileage. For electric and plug-in hybrid ratings, see what MPGe means. The wider subject sits under Fuel and Efficiency.
Frequently asked questions
Does the EPA test every new vehicle?
No. Manufacturers test a representative vehicle for each weight, transmission and engine combination and report the data. EPA reviews the submissions and confirms a sample, about 15% to 20%, with its own laboratory tests.
Why is my real-world mpg lower than the EPA rating?
The rating comes from controlled cycles. Speed, temperature, air conditioning, cargo, tire pressure and driving style all change the result. The label is an estimate for typical use, not a guarantee for your commute.
Can I compare a 2007 rating with a 2026 rating?
Only with care. Cars from 2007 and earlier were rated on two cycles. From 2008, three more cycles were added, so the two numbers were produced under different procedures.
Do electric cars get an EPA mpg rating?
They get MPGe, which expresses energy use on a gasoline-equivalent basis. EPA uses 33.7 kilowatt-hours as the energy equivalent of one gallon of gasoline. It tests range by driving the vehicle until the battery can no longer meet the cycle.
