The short answer
Emissions fails are rare on young cars and climb steadily with age. Across models with 500 or more tests, about 1 in 250 first tests on cars first used in 2021 (0.4% of 2,021,332 tests) failed with an exhaust, emissions or fuel item, in tests carried out 2021 to 2025. For cars first used in 2008 it was about 1 in 30, or 3.5%, with 2006 cars at 3.4%. In between: 0.8% for 2018 cars, 1.5% for 2015, 2.1% for 2012.
The item recorded most, at every age, is not a smoke or gas reading. It is the engine warning light: a malfunction indicator lamp that stays on, or does not light at all. On cars aged 15 and over it appears on 2.0% of first tests; on cars aged 3 to 5, 0.3%. The DVSA manual treats a warning-light fault as a major defect on its own, before the analyser is plugged in. If the light is on, that is the fail to fix first.
Emissions fails by item and car age
Each cell is the share of all first tests in that age band that failed with the item, pooled across 2021 to 2025. The warning light appears twice because the manual lists it once for petrol and once for diesel engines; the store does not say which row is which, so they are shown separately and not added.
| Item recorded | 3-5 years | 6-9 years | 10-14 years | 15 and over |
|---|---|---|---|---|
| Engine warning light on or not working (first entry) | 0.3% | 0.6% | 1.1% | 2.0% |
| Engine warning light on or not working (second entry) | 0.3% | 0.5% | 0.7% | 0.4% |
| Petrol: lambda reading outside limits | 0.1% | 0.2% | 0.6% | 1.4% |
| Petrol: carbon monoxide above the maker's limit | 0.0% | 0.1% | 0.2% | 0.6% |
| Diesel: smoke above the limit for its age | 0.0% | 0.1% | 0.3% | 0.4% |
| Emissions test could not be completed | - | 0.1% | 0.2% | 0.5% |
| Engine oil leaking excessively | 0.0% | 0.1% | 0.2% | 0.3% |
The shape is the same as the cohort figures: nearly nothing under six years, then a steady rise. The petrol gas readings and the diesel smoke readings sit well below the warning light in every band. An exhaust leak that is likely to affect the reading is recorded in its own right on 0.2% of tests on cars aged 10 to 14, and a test that could not be completed, for example because the engine would not hold the required speed, is itself a fail item.
What the test actually measures
The inspection manual sets the limits. A petrol car gets a gas test: under the basic test, carbon monoxide up to 0.2% and hydrocarbons up to 200 parts per million at fast idle, carbon monoxide up to 0.3% at idle, and a lambda reading between 0.97 and 1.03, with the maker's own limits used where they exist. A diesel gets a smoke meter, and the limit depends on when the car was first used: 2.5 for a non-turbo and 3.0 for a turbo engine before July 2008, 1.5 from July 2008 to the end of 2013, and 0.7 from January 2014. Dense blue or clearly visible black smoke for five seconds at idle is a major defect; so is any visible smoke from a car fitted with a diesel particulate filter. Smoke thick enough to obscure other road users' vision is dangerous, and a car that fails on a dangerous item cannot be driven away except for repair.
A clean air zone is a different test
People mix these up. Clean Air Zones in Bath, Birmingham, Bradford, Bristol, Portsmouth, Sheffield and Tyneside charge some vehicles a daily fee if they do not meet an emissions standard. That standard is fixed by the car's type approval when it was built. It has nothing to do with the MOT result: a car can pass every emissions test it ever takes and still be charged, and a car can fail its MOT on smoke and not be. The only way to know is the GOV.UK vehicle checker, which needs the registration number.
What to check before the test
- The warning light. It should light with the ignition and go out once the engine runs. If it stays on, have the fault code read and the cause fixed. Clearing the code without a fix usually brings the light back, and a lamp that never lights at all is also a fail.
- Service items. Old oil, a blocked air filter and tired spark plugs push the gas readings the wrong way on a petrol car.
- Smoke on a cold start. Blue smoke is oil; black is unburnt fuel. Either is worth a look before the tester sees it.
- Exhaust leaks. A blow near the front of the system lets air into the gas sample and can fail the reading on its own.
- Oil leaks. An engine leaking excessively is a fail item in this family, recorded on 0.3% of tests on cars aged 15 and over.
- Arrive warm. A short motorway run before the test helps a diesel particulate filter and brings the engine to the temperature the test expects.
Last year's emissions result and any advisory are on the free MOT history service.
What this does not mean
These are fleet figures across every model with 500 or more tests, and the store has no fuel type: the petrol and diesel rows are the items for each engine type, but the base under every cell is all cars in the band. Nothing here says diesels fail more than petrols or the reverse. A rising line with age says that older engines, as a population, more often had a lit warning light or an out-of-limit reading on the day; it does not say why, and it says nothing about the car on your drive. A well-serviced fifteen-year-old can pass clean; a neglected five-year-old can fail on a lamp.
Frequently asked questions
Yes. A malfunction indicator lamp that stays on, or does not light at all, is a major defect under the inspection manual, and the car fails whatever the gas reading says. It was recorded on 2.0% of first tests on cars aged 15 and over and 1.1% on cars aged 10 to 14.
They depend on the car's first-use date: 2.5 for a non-turbo and 3.0 for a turbo engine before July 2008, 1.5 from July 2008 to the end of 2013, and 0.7 from January 2014. Visible smoke from a car with a particulate filter is a major defect regardless of the reading.
Not necessarily. Clean air zone charges depend on the emissions standard the car met when built, not on its MOT result. Check the registration on the GOV.UK clean air zone service.
Yes, steadily. The share of first tests failing with an exhaust, emissions or fuel item was 0.4% for cars first used in 2021, 1.5% for 2015 and 3.5% for 2008, in tests carried out 2021 to 2025.
Sources and method
Cohort shares are class 4 first tests, 2021-2025, across models with 500 or more tests, by first-used year (4,746,066 tests first used 2006 to 2,021,332 first used 2021), shown as the share of all first tests that failed with at least one item in the exhaust, emissions and fuel family. Item shares are each item as a share of all first tests in the age band (33,793,352 tests aged 3-5, 50,103,915 aged 6-9, 44,980,002 aged 10-14, 31,405,150 aged 15 and over). The malfunction indicator lamp is listed under two headings in the DVSA manual with identical wording; the two rows are shown as recorded and not added. The diesel smoke row shows, for each band, the smoke item whose first-use dates fall in that band (from January 2014 for the two youngest bands, from July 2008 for 10 to 14, before July 2008 turbo for 15 and over). The store carries no fuel type, so every base is all cars in the band. A dash means the item was outside the eight most recorded for that band. National context: 72.3% of 33,210,951 first tests passed in 2025.
- DVSA anonymised MOT results, class 4 first tests, 2021-2025 (data.gov.uk)
- GOV.UK: MOT inspection manual, section 8, nuisance (emissions)
- GOV.UK: Driving in a clean air zone
- GOV.UK: MOT test result
- GOV.UK: Check the MOT history of a vehicle
About these figures. Figures are computed by SortedCars from the DVSA anonymised MOT results dataset (class 4, first tests only, test years 2021-2025), published under the Open Government Licence v3.0. Pass rate = passes divided by first tests, with pass-after-rectification counted as a fail. No figure is shown for any group with fewer than 500 tests. Test class 4 only, first tests only; retests are excluded. Pass rates describe fleets, not individual machines: they reflect how cars were driven, maintained and used as well as how they were built, and beyond recorded mileage none of that is in the data. A well-kept example of a model with a weak average can still pass; a neglected example of a strong one can still fail. Alias spellings of a make are counted under one make name and all trims of a model under one model name, using SortedCars' grouping tables.
This article is general information drawn from published test records. It is not professional, mechanical or legal advice, and an MOT pass or a clean MOT history does not guarantee that a car is roadworthy today. If you are unsure about a car, have it inspected by a qualified mechanic. You can check any vehicle's MOT history free on the official GOV.UK MOT history service.
Check a car's MOT history free
Enter a registration to see its DVLA record and full MOT history, free, no sign-in.