How big is the gap between official and real-world mpg?
Around a fifth for a petrol car in Europe. The European Commission's first analysis of data from on-board fuel consumption meters, published in 2024 and covering cars registered in 2021, found petrol cars used about 24% more fuel on the road than their WLTP figure and diesels about 18% more. Plug-in hybrids used roughly three and a half times their official figure, because most owners charge less than the test assumes.
In the United States the gap is smaller, because the EPA's five-cycle test is more demanding. Owner-reported data on fueleconomy.gov typically sits within 5 to 10% of the label, in either direction. An EPA figure is close to a realistic expectation; a WLTP figure is a target you will approach on a good day.
Why is the laboratory figure higher than real driving?
Because the test removes almost everything that costs fuel in real life. The cycle is run at a set temperature, usually 23 degrees Celsius for WLTP, with no wind, no hills, no passengers, no roof box, no air conditioning and no heater. Tyres are at the specified pressure. The car has been warmed through in a way that a school run never allows. The driving itself is smooth, with prescribed acceleration rates that no impatient driver matches.
Each of those conditions is reasonable on its own, since a repeatable test has to fix them somewhere. Together they describe a drive that does not happen. The figure is honest about the car under those conditions, and silent about the conditions you drive in.
There is also a small amount of gaming within the rules. Manufacturers may test the lightest version of a model, fit the narrowest tyres it is sold with, and tape over panel gaps in ways that were permitted under NEDC. WLTP closed most of those loopholes, which is one reason its figures dropped relative to NEDC, but a test is still a thing that can be prepared for.
What in my driving causes the difference?
Short trips, cold weather, speed and load, in roughly that order. A petrol engine uses far more fuel in its first few minutes while it warms up; a trip of under three miles may never leave that phase, and a car used mainly for such trips can return 30 to 40% below its rating. Winter widens the gap, with denser air, stiffer oil, a slower warm-up and a heater drawing on the engine.
Speed matters above about 55 mph, because aerodynamic drag rises with the square of speed. A car that returns its official figure at 60 mph will use 15 to 20% more fuel at 75. Load is the fourth factor: passengers, luggage, a roof box and a trailer each add weight or drag. Air conditioning costs 5 to 10% in hot weather. The improving fuel economy guide puts numbers on each.
Which cars fall furthest short?
Plug-in hybrids, by a wide margin, for the reason above: the official figure assumes a full battery at the start of a short cycle, and a plug-in that is never charged is a heavy conventional hybrid. Small turbocharged petrol engines are next. They score well on gentle test cycles, where they run mostly off boost, and lose more than larger engines in real traffic, where they are on boost often.
Conventional hybrids fall short by the least in urban driving, where they recover energy the test also captures, and by more on motorways, where the battery contributes little. Diesels are the most consistent, because their efficiency changes less with load and temperature. Large naturally aspirated engines are also consistent, at a poor figure.
How do I predict my own figure from the rating?
Start with the official combined figure and apply a discount for the test: none for EPA, about 15 to 20% for WLTP, about 20% for ARAI, more for older NEDC or JC08 figures. Then adjust for your driving: another 10 to 20% off for mostly short urban trips or a cold climate, roughly nothing for a mix, and a small gain if you spend most of your miles at a steady 50 to 60 mph.
A WLTP figure of 55 mpg (UK) for a petrol hatchback therefore becomes an expectation of around 45 to 47 mpg for a mixed driver, or 40 for a city driver in a northern winter. Convert it with the UK mpg to L/100km page if you think in metric: 45 mpg (UK) is 6.28 L/100km. Then measure. The manual calculation guide shows how to turn a few fill-ups into a figure you can trust.
Should the official figures be ignored?
No. They are the only fair comparison between cars, precisely because the conditions are fixed. A car that scores 10% better than another on WLTP will very probably use about 10% less fuel in your hands too, even though neither will hit its number. The rating is a ranking tool, not a forecast.
What deserves scepticism is any single number presented without its test and its unit. An mpg figure with no gallon and no cycle behind it is not comparable with anything, and the US vs UK gallon guide explains how much of the apparent gap between countries is simply the size of the gallon.