Data study · risk by age of car
Car repair costs by age
Risk does not climb forever. For everything a car wears out, it peaks around fifteen years and then falls away — and for the one thing it cannot wear out, it never falls at all.
DVSA test recordOpen Government LicenceNo sign-up
A suspension advisory on a fifteen-year-old car comes back as a failure 22.3% of the time — one car in 4.5 — against 10.7% on a five-year-old, according to DVSA MOT data analysed by PriceMyRepair. But past fifteen the rate falls again, and it falls for every worn component we can measure. Corrosion is the exception: it rises from 1.7% at two years to 9.4% at twenty-two and never turns back.
15 years
where the risk of a worn part coming back as a failure peaks — and then starts falling
5.5×
how much corrosion risk grows between a two-year-old car and a twenty-two-year-old — the one curve that never comes down
£24 → £50
what a spring advisory is worth ignoring at five years against fifteen, on the same car
There is no cliff. There are two shapes
In PriceMyRepair’s analysis of DVSA MOT data, the risk of a worn component turning an advisory into a failure peaks at fifteen to seventeen years and then falls, while corrosion risk rises at every age measured. The obvious expectation is a line that climbs: older car, more risk, bigger bills, all the way down. The national test record does not show that. It shows two clearly different curves, and which one a component follows depends on whether it wears out or rusts out.
Things that wear rise, peak, then fall. Suspension, wheel bearings and lamps all follow the same arc — risk climbs steeply through the first decade, tops out somewhere between fifteen and seventeen years, and then eases back. A twenty-five-year-old car has a lower chance of a bearing advisory becoming a failure than a fifteen-year-old one.
Things that corrode only ever go up. Structural corrosion and exhaust condition never turn. Corrosion climbs from 1.7% at two years to 9.4% at twenty-two, the steepest rise of anything measured, and it is still climbing at the last age we can report. The exhaust curve does the same at a shallower angle.
Every curve we can publish
Suspension advisories return as a failure 10.7% of the time on a five-year-old car, 19.0% at ten, 22.3% at fifteen and 20.8% at twenty, according to PriceMyRepair’s analysis. Each figure is the share of advisories in that category, written on a car of that age, that came back as a failure at a later test. Read across, not down — the ages published differ by category.
| Category | Risk by age of car when the advisory was written | Shape |
|---|---|---|
| Suspension | 5y 10.7% · 10y 19.0% · 15y 22.3% · 20y 20.8% | Peaks at 15, then falls |
| Lamps and electrical | 2y 5.2% · 7y 7.4% · 12y 13.5% · 17y 15.5% · 22y 13.5% | Peaks at 17, then falls |
| Axles and bearings | 2y 3.0% · 5y 4.9% · 10y 6.5% · 15y 7.5% · 20y 6.6% · 25y 5.6% | Peaks at 15, then falls |
| Body and exhaust | 6y 7.5% · 11y 10.8% · 16y 13.9% · 21y 14.3% · 26y 14.6% | Never falls |
| Structure | 2y 1.7% · 7y 2.9% · 12y 5.6% · 17y 7.4% · 22y 9.4% | Never falls |
Figures are measured at the ages shown, not averaged across bands, so the columns are not directly comparable between rows and the points must not be summed. Ages beyond the mid-twenties rest on small numbers and are not published. Three of the eight advisory categories in the record do not yet carry an age breakdown and are absent here — see the full category comparison for those.
Why the curve comes down — and why that is not good news
Old cars are not better maintained. There are simply fewer bad ones left. A car that reaches twenty on British roads got there one of two ways: someone looked after it, or it barely moved. Neglected cars with expensive faults leave the fleet long before that, taken off the road or scrapped when a bill exceeds what the car is worth. What remains is a filtered population, and the falling curve is measuring the filter rather than the cars.
Which means the number is not a promise about your car. If you own a twenty-year-old car that has been driven hard and patched cheaply, the 20.8% suspension figure does not describe it — that figure describes the survivors, and survivorship is exactly what you cannot assume about a specific vehicle. The curve is useful for understanding the fleet and dangerous for reassuring yourself.
The wear items also reset, which corrosion never does. Drop links, bearings and bulbs get replaced two or three times across a car’s life and start fresh each time, so their risk reflects how long since the last replacement rather than how old the car is. Rust has no such reset. A careful owner can postpone a bill; nobody can un-rust a car, and every winter adds to a total that never reduces.
That is the sentence that makes the two shapes matter. Careful ownership genuinely protects the wear items — that is most of why the suspension and bearing curves fall after fifteen — but it does not protect the structure in the same way. Underseal, garaging and dry winters slow corrosion; they do not reverse what has already happened.
So the older a car gets, the more its risk profile shifts from things that can be replaced for a known price towards the one thing that cannot be priced at all until it is cut open. Structural corrosion is the only MOT failure we refuse to quote, and it is also the only curve here still rising at the last age we can measure.
What each age is actually worth
Ignoring a corroded coil spring advisory carries £24 of risk on a five-year-old car and £50 on a fifteen-year-old, according to PriceMyRepair’s analysis. A percentage is half the story. Multiplying the risk at each age by what that repair costs gives the figure that matters — the cost you are carrying by leaving that advisory alone, at that age of car.
| Advisory | Repair | Cost of ignoring it, by age of car |
|---|---|---|
| Coil spring corroded | £130–£320 | 5y £24 · 10y £43 · 15y £50 · 20y £47 |
| Exhaust has a minor leak | £60–£300 | 6y £14 · 11y £19 · 16y £25 · 21y £26 · 26y £26 |
| Wheel bearing has play | £120–£300 | 2y £6 · 5y £10 · 10y £14 · 15y £16 · 20y £14 · 25y £12 |
| Headlamp aim | £20–£40 | 2y £2 · 7y £2 · 12y £4 · 17y £5 · 22y £4 |
| Corrosion near a mounting | No honest range | Rising at every age measured — and unpriceable at all of them |
Category risk at that age × the midpoint of the repair range, rounded to the nearest pound. Worked example, the top row at fifteen years: 22.3% × £225 = £50.18, shown as £50. This is not a bill you will receive — it is the average cost of leaving that line alone, across every car of that age that leaves it alone. The same method is used on our category comparison, so the two pages can be read together.
What this means if you are buying
The two shapes give a used buyer a genuinely different checklist depending on which side of fifteen years the car sits.
Under ten years
Risk is real but still climbing, and almost all of it is in parts with known prices. An advisory here is a scheduled cost, not a decision about the car’s future.
Read the advisory list, price it, negotiateTen to eighteen years
The peak. Suspension, bearings and lamps are all at their worst here, and this is where the wear bills cluster. It is also where the most cars are sold with a fresh MOT and a list of things left for the buyer.
Check what was fixed, not just what passedPast eighteen years
Wear risk is falling, but only because you are looking at a survivor. Structure is the question now, and it is the one a test certificate answers least well.
Get it on a ramp before you buyA long clean MOT history is real evidence about how a car has been kept — but it is evidence about the items the test inspects, which are mostly the replaceable ones. Our guide to what the MOT actually checks covers what a clean certificate does not tell you.
Enter your reg — we’ll read its age and its MOT history, show which advisories it is carrying and price each one for your model rather than the national average.
How these curves were produced
Method
Source. The DVSA MOT testing bulk dataset, published under the Open Government Licence v3.0 and analysed by PriceMyRepair, covering 145,103,213 advisories tracked across 216,057,619 pairs of consecutive tests on the same vehicle. Class 4 — cars and light passenger vehicles — in Great Britain. Northern Ireland runs a separate system under the DVA and is not in this data.
What each figure means. Advisories were grouped by the age of the vehicle on the day the advisory was written, then followed forward to that same vehicle’s later tests. The percentage is the share recorded as a defect in the same category at a later test — returned as a failure.
These are measured points, not bands. Each figure is the value at that age, not an average across a range of ages, and the counts behind them cannot be summed. Sample sizes at each point: suspension 890,803 · 2,180,862 · 1,776,192 · 426,288. Lamps 1,717 · 9,702 · 90,315 · 128,087 · 13,650. Axles 9,884 · 58,252 · 149,270 · 159,026 · 55,832 · 17,634. Body and exhaust 176,274 · 493,177 · 460,245 · 117,354 · 27,756. Structure 2,599 · 8,019 · 27,257 · 32,796 · 10,454.
Cost of ignoring. Category risk at that age × the midpoint of our published repair range for that specific advisory, rounded to the nearest pound. Repair ranges are our own independent-garage estimates and are not from the DVSA dataset — the two sources are combined here and nowhere in the official record.
The survivorship caveat, stated plainly. The fall after fifteen years is a property of which cars are still being tested, not of the parts improving. Cars with expensive faults leave the fleet, so the older cohorts are filtered. Any figure past about eighteen years should be read as describing survivors.
Known limitations. No age breakdown is published for three of the eight categories, so they are absent. An advisory and a later failure in the same category are treated as the same fault, which slightly overstates conversion where a different component in that group failed instead. Curves are national and do not adjust for model, mileage or region. Nothing is interpolated between the published points and nothing is extrapolated beyond them. Repair ranges are estimates; final quotes come from garages.
Contains public sector information licensed under the Open Government Licence v3.0.
Car repair costs by age — FAQ
At what age does a car start costing more to keep on the road?
Risk climbs steadily through the first decade and peaks around fifteen years for everything that wears out — suspension, bearings and lamps all top out between fifteen and seventeen. What surprises people is that it then falls again. The exception is corrosion, which rises at every age we can measure and never turns back.
Why does MOT risk fall on cars over fifteen years old?
Because of which cars are still there, not because the parts improve. A car that reaches twenty in Britain was either looked after or barely driven; the neglected ones with expensive faults left the fleet years earlier. The falling curve measures that filter. It says very little about any specific old car.
Which repairs get more likely as a car ages, without ever easing off?
Corrosion-driven ones. Structural corrosion near a mounting climbs from 1.7% at two years to 9.4% at twenty-two, and exhaust condition follows the same shape more gently. Rust does not pause, does not reverse, and does not benefit from the car being cared for in any way that undoes what has already happened.
Is a fifteen-year-old car the worst age to own?
For wear items, yes — that is where suspension and bearing risk peaks. But it is not the worst age overall, because corrosion is still climbing past it. A fifteen-year-old car has the highest chance of a mechanical advisory turning into a bill; a twenty-five-year-old has the highest chance of a structural one.
How much does an advisory cost to ignore on an older car?
It roughly doubles between five and fifteen years. A corroded coil spring advisory carries £24 of risk on a five-year-old car and £50 on a fifteen-year-old, because the same repair becomes twice as likely to be forced on you. The figure is the risk multiplied by the repair cost, not a bill you will receive.
Do older cars fail their MOT more often?
This page measures something narrower and more useful: whether an advisory already on the certificate comes back as a failure. That does rise with age up to about fifteen years. A general failure rate by age is a different figure, and we do not publish one because the analysis behind it is not finished.
Should I avoid buying a car over fifteen years old?
Not on this data alone. The wear risk is actually lower than at fifteen, and that is genuine — old cars still on the road tend to be the good ones. What changes is that the risk shifts towards structure, which is harder to inspect, harder to price and the one thing a clean certificate is worst at telling you about.
Why do the ages differ between categories in your table?
Because each category’s data is published at the ages where its sample is large enough to be meaningful, and those differ. Exhaust figures start at six years because there are too few advisories on younger cars; lamp figures stop at twenty-two. We publish the points as measured rather than forcing them onto a common scale.
Does mileage matter more than age?
For wear items it very likely does, and the reset effect in this data hints at it — links and bearings are replaced several times across a car’s life, so their risk tracks time since the last replacement rather than the car’s birthday. We have not yet published the mileage comparison, so this page does not claim one.
How do I know where my own car sits?
Its MOT history is public and it names every advisory a tester has ever recorded, with the date. Our advisory decoder reads that history from your registration, works out the age at which each advisory was written and prices the repair for your model.
The pages behind these curves
Each category’s age breakdown is published in full on its own page, alongside what a tester actually saw and what the repair involves.