Scotland is rusting, and it is showing up in the brakes

The finding that rules out the obvious explanation

The first assumption anyone makes is that poorer or more remote areas run older cars, and older cars fail more. Age does drive failure: in our data a three year old car fails 9.16% of the time and an eighteen year old car fails 35.84% of the time.

That is not what is happening here. The mean age of a car tested in Kirkcaldy is 9.2 years. In East London it is 9.8 years. Scottish cars in the worst areas are, if anything, slightly newer.

So we narrowed the comparison to cars aged 10 to 13 at the date of test, everywhere in the country, and compared like with like. The gap did not close. It widened.

Scottish cars of the same age carry 4.3 times the corrosion failure rate of English ones.

It is the brake pipes

We took the nine worst areas, all of them Scottish, and the nine best, all of them in London, and compared the actual defects recorded on cars of the same age. These are failure items per 1,000 tests.

Every item at the top of that list is a steel component sitting under the car in the spray. Brake pipes run the length of the floor. The prescribed area is the structure around the suspension and seatbelt mountings. Handbrake cables seize. Calipers stick. This is one mechanism, showing up in six places.

Roughly one in twenty Scottish cars aged ten to thirteen failed its MOT on a corroded brake pipe. In East London the figure is closer to one in 1,700. A brake pipe that lets go does not degrade gently, it fails all at once.

Method, and what this does not prove

Source data is the DVSA Anonymised MOT tests and results release for 2023, published under the Open Government Licence. It covers 42,216,721 tests. We kept class 4 vehicles only, which is cars and small vans, and first tests only, discarding retests so a single car cannot be counted twice. That leaves 33,000,722 tests. For the regional comparison we further restricted to cars aged 10 to 13 at the test date, giving 7,360,348 tests across 117 postcode areas with at least 3,000 tests each.

Failure rate is tests resulting in a fail, divided by all tests with a pass, fail, or pass after rectification outcome. Corrosion failures are counted from the separate failure item file, matched to 477 class 4 defect codes whose official description mentions corrosion or rust, and counted only where the item caused a failure rather than an advisory.

Three honest limitations:

• The postcode area is where the test was carried out, not where the car is kept. Most people test near home, so it is a strong proxy, but it is a proxy.
• Tester behaviour may contribute. Testers who see heavy corrosion daily may inspect for it harder. We do not think this explains an eighty fold gap, and the defect pattern is mechanically consistent, but we cannot separate the two effects.
• London's figures may be flattered by the Ultra Low Emission Zone having removed some of the roughest vehicles from the fleet, even within the same age band.

2023 is the most recent year DVSA has published in bulk. Figures are recomputed from the raw release rather than taken from any summary.

Contains public sector information licensed under the Open Government Licence v3.0. Free to reuse with attribution to Bad Drivers UK and a link to this page.