Count the cars, not the recalls
On cars built from 2016 to 2021, Toyota filed 116 safety recalls in the US. Ferrari filed 11. Count recalls and the $300,000 car wins by a mile. Count the cars each recall reaches and the answer flips: a typical Toyota recall touched 1 in 523 of a year’s cars. A typical Ferrari recall touched a third of a year’s Ferraris.
“Is a Ferrari built better than a Toyota?” sounds like a question about engineering. In the only public record of things going wrong, NHTSA’s recall file, it turns out to be a question about what you count.
The count everyone does
NHTSA publishes every US safety recall since 1966, one row per campaign, make, model and model year. Take every campaign touching model years 2016–2021 and count them: Toyota and Lexus 116, Lamborghini 17, Ferrari 11.
By that count Toyota looks like the problem. But Toyota sold 14,139,369 cars in the US over those six years, and Ferrari shipped 16,554 to the whole of the Americas. A maker that builds 850 times more cars will file more recalls. The count says more about size than about quality.
The count that flips it
So count something that scales: how many cars a recall actually reaches, as a share of one year of that brand’s sales. Do it for every campaign and take the typical (median) one:
| brand | recalls | typical recall | share of a year’s cars | one in |
|---|---|---|---|---|
| Toyota + Lexus | 116 | 4,508 cars | 0.19% | 523 |
| Lamborghini | 17 | 224 cars | 12.4% | 8 |
| Ferrari | 11 | 893 cars | 32.4% | 3 |
Toyota’s typical recall covered 4,508 cars, a rounding error against millions of sales. Ferrari’s covered 893, about a third of everything it ships to the Americas in a year. Lamborghini sits between them, at one car in eight.
A small maker files fewer recalls. Each one lands on a far bigger slice of its cars.
Why a small maker’s mistakes are big
Ferrari builds so few cars that the same parts go into model after model, year after year. When one of those parts is wrong, it is wrong everywhere it was used.
The clearest case is campaign 22V536000. A brake fluid reservoir cap that could leak put 23,555 Ferraris on a single recall, built across 18 model years (2005–2022) and 19 models, from the 430 to the LaFerrari and the Roma. That one cap covered more Ferraris than Ferrari shipped to the Americas in the whole 2016–2021 window. The field of cars at the end of the video is that campaign, at that count, exactly.
What the piles in the video are
The opening scene sorts Ferrari’s 34 recalls on cars built since 2005 by the part each one fixes. One part in a pile is one recall of that part, and every part is drawn at the same size, so a pile’s size follows its count:
| part | recalls | cars covered |
|---|---|---|
| airbags | 9 | 22,077 |
| engine | 5 | 2,947 |
| fuel system | 4 | 2,879 |
| brakes | 3 | 33,580 |
| gearbox | 3 | 586 |
| door and hood latches | 2 | 3,664 |
| glass | 2 | 1,143 |
| wiring and software | 2 | 581 |
| seats | 2 | 98 |
| body structure | 1 | 541 |
| tires | 1 | 85 |
The airbags make the biggest pile, nine recalls. The brakes cover the most cars, because the brake fluid cap is one of them.
The part that argues against us
We don’t claim Ferrari gets more recalls per car. On recall notices per car sold it does, 1.41 against Toyota’s 0.63. But campaigns that run past the window’s edges have to be split across model years, and NHTSA doesn’t publish units per model year. Under the strictest split the two tie, so the video never says it.
Ferrari’s sales are an upper bound. Ferrari reports “Americas” shipments (US, Canada and Latin America), not US sales, so its true US share per recall is larger than shown. That makes our Ferrari numbers the conservative ones.
A recall is a filing, not a failure. It counts what a maker reports and fixes, not how often a car breaks. The point here is the shape, a few recalls that each reach a big slice of a small maker’s cars, not a verdict on which car is built better.
“One in five hundred” is rounded. Toyota’s figure is 1 in 523; rounding to 500 makes Toyota look about 4.6% worse, which works against our own argument, so it stands.
One more: the red car on screen is a stand-in 458 model, and the 458 is not on the brake-cap recall. It’s never counted as one of the 23,555.
Get the data
Free, no email required: the headline table, every one of the NHTSA campaigns behind it, Ferrari’s recalls by part, the sales figures with their sources, and the methodology.
3 brands · every campaign · 11 Ferrari parts · sourced sales · methodology
Sources: NHTSA’s recall flat file (static.nhtsa.gov/odi/ffdd/rcl/), every campaign touching model years 2016–2021 plus Ferrari’s campaigns on 2005+ cars; sales from Ferrari’s SEC 20-F filings (Americas shipments) and from published reports of Lamborghini’s US deliveries and Toyota and Lexus US sales, each row in the kit carrying its URL and the sentence quoted from it. Credits: Ferrari 458 Italia model by vicent091036; generic passenger car pack by Comrade1280 (CC-BY 4.0); “Toyota Auris wagon” scan by thesidekick (CC-BY 4.0), via Objaverse; HDRI, grass and concrete from Poly Haven (CC0); sound from the Sonniss GDC bundle and Mixkit. The prancing-horse mark belongs to Ferrari S.p.A. and appears only to identify the brand. Analysis, figures and video: reelgorithm.py.