NHTSA doesn't test everything. It buys cars and crashes them, which costs money, so it prioritises high-volume models. The result is a gap most buyers never notice.
Of 822 model lines still on sale in 2011 or later, 512 have no overall rating at all.
| Brand | Models untested | |
|---|---|---|
| Ferrari | 100% | 17 of 17 |
| Porsche | 100% | 11 of 11 |
| Land Rover | 100% | 9 of 9 |
| BMW | 90% | 98 of 109 |
| Mercedes-Benz | 83% | |
| Honda | 76% | 37 of 49 |
| Lexus | 53% |
Supercars you'd expect. Honda you wouldn't.
Ferrari and Porsche at 100% makes sense — low volume, high price, nobody is buying a 911 on its star rating.
Honda at 76% does not fit that logic at all. The explanation is that a "model line" in this data includes every variant and trim family, and NHTSA tests the highest-volume configuration and moves on. A Civic gets tested; a Civic Type R, a Civic hatchback in a given year, or a low-volume variant may not.
Across the whole database only 312 models have safety records out of 2,603 — about one in eight.
Why "no stars" reads as bad news and isn't
When a listing shows a rating, buyers use it. When it shows nothing, most people assume the car did poorly or the data is missing by accident.
Neither is true. An untested car has no result because nobody crashed one — it says nothing about how the structure would perform. Meanwhile 81% of cars that were tested score five stars, which we covered in a five-star rating means less than you think. So the rating is nearly binary among the tested, and absent for the majority.
What to use instead
Check IIHS. The insurance industry's programme tests a different set of vehicles and uses tougher criteria — small overlap crashes and headlight performance among them. A car missing from NHTSA is often present there.
Look at the platform, not the badge. Variants share structure. If the standard version of a model was tested and scored well, a rarer body style on the same platform is unlikely to be structurally different.
Weight still matters more than stars. Ratings compare vehicles only against others of similar mass — a point that gets lost every time two cars with the same score are treated as equivalent. We went through that in the heaviest vehicle weighs as much as five of the lightest.
Check the individual car, not just the model. A five-star vehicle with an unrepaired airbag campaign doesn't perform to its rating. A VIN check shows whether the safety recalls on that specific car were ever completed — and given that inflator campaigns take 9.2 years to be announced, that's not a rare situation.
Frequently asked questions
How many cars have no crash test rating?
512 of 822 model lines sold in 2011 or later — 62.3% — have no NHTSA overall rating. Across the whole database only 312 models out of 2,603 have safety records.
Which brands have the most untested models?
Ferrari, Porsche, Aston Martin, Land Rover and Jaguar are at 100%, BMW at 90%, Mercedes-Benz at 83% and Honda at 76% — the last being the surprise, since it's a mainstream high-volume brand.
Does no rating mean the car is unsafe?
No. It means nobody crashed one. NHTSA buys and destroys vehicles to test them, so it prioritises high-volume configurations — an absent rating carries no information about the structure.
What should I check if a car has no NHTSA rating?
IIHS, which tests a different set of vehicles with tougher criteria including small overlap crashes and headlights. Also check whether the standard version of the same platform was tested, since variants share structure.
Denis Kataev is a serial solopreneur and the founder of ForCar. With 15+ years in software engineering and 10 years in SEO, he builds data-driven products end to end — backed by a sharp eye for design. At ForCar he mines proprietary vehicle datasets, turning raw numbers into buying advice you can actually trust.
Keep reading
All articles →
One in Five Recalls Ever Issued Warns About Fire. At Kia It's Two in Five.
Every recall carries a description of what happens if the defect isn't fixed. I read all of them looking for one word, and it turned up far more often than a headline-driven view of the industry would suggest.
Tesla Fixes 74% of Its Recalls Overnight. Kia Fixes 31% That Way.
Two cars can have the same number of recalls and completely different consequences for the owner. The difference is whether the fix requires a workshop or a download.
An Airbag Inflator Defect Takes 9 Years to Be Admitted. A Seat Belt Fault Takes 2.
Some defects get found almost immediately and some hide for a decade. Sorting recalls by how long they took to surface produces a ranking that says a lot about which parts fail slowly.