NHTSA campaign numbers encode the year they were issued, so the gap to the model year is measurable. Averaged by brand, it splits the industry in half.
| Brand | Average lag | Issued in year one |
|---|---|---|
| Acura | 6.86 yrs | 17.2% |
| Honda | 6.75 yrs | 19.7% |
| Subaru | 6.33 yrs | |
| Chevrolet | 5.01 yrs | |
| Toyota | 4.32 yrs | |
| Nissan | 3.81 yrs | 45.8% |
| Kia | 3.73 yrs | |
| Buick | 46.8% |
One in five against one in two
The second column is the clearer one. Buick issues 46.8% of its recalls within a year of the model year, Nissan 45.8%, GMC 45.5%. Honda manages 19.7% and Acura 17.2%.
So on a Buick, roughly half the defects that will ever be admitted are admitted while the car is nearly new. On a Honda, four out of five arrive later — often years after the original owner has sold it.
What the number does and doesn't mean
It is tempting to read a long lag as concealment. That isn't supportable from this data, and there's a more mundane explanation available.
The mix of components matters enormously. We measured that airbag inflators take 9.23 years to surface while brakes take 2.18. A brand whose campaigns skew towards slow-degrading parts will show a long average regardless of how it behaves.
Honda and Acura were among the most heavily affected by the Takata inflator recall, which is the slowest-surfacing defect in the entire database. That alone pushes their averages up.
What the number genuinely tells you is when the risk lands on you, not who is at fault. And that timing is what matters to a used buyer.
The practical consequence
A fast-recalling brand is safer to buy used. If most defects were admitted while the car was new, the campaigns are already on record and mostly repaired. What you see is what there is.
A slow-recalling brand carries more unknown. On a Honda, a clean report on a five-year-old car is less reassuring, because most of its eventual recalls haven't been issued yet.
Recheck cars from slow brands. The single cheapest thing on this page. We found that a third of all recalls arrive five or more years after the model year — the check you ran at purchase goes stale.
A VIN check lists the currently open campaigns for a specific vehicle, and the repair stays free at any age. We looked at what's typically still outstanding in 5.9 open recalls per car.
Frequently asked questions
Which brands take longest to issue recalls?
Acura at 6.86 years from model year to campaign, Honda at 6.75 and Subaru at 6.33. The fastest are Kia at 3.73, Nissan at 3.81 and Toyota at 4.32.
Which brands issue recalls fastest after a car is built?
Buick announces 46.8% of its campaigns within a year of the model year, Nissan 45.8% and GMC 45.5%. Honda manages 19.7% and Acura 17.2% — roughly one in five.
Does a slow recall lag mean a brand hides defects?
Not necessarily. Component mix explains much of it — airbag inflators take 9.23 years to surface against 2.18 for brakes, and Honda and Acura were heavily affected by the Takata inflator campaign, the slowest defect in the database.
Why does recall speed matter when buying used?
It tells you how much is still unknown. On a fast-recalling brand most defects were admitted while the car was new and are already on record. On a slow one, a clean report on a five-year-old car means less.
Keep reading
All articles →
One Model Year of the Ford E-350 Carries 38 Separate Recall Campaigns
Recall counts usually get quoted per brand, where volume dominates everything. Counted per model year, the top of the list is a very short list of specific, unlucky vehicles.

62% of Modern Cars Were Never Crash Tested. That Includes Three Quarters of Honda's Range.
Everyone knows what five stars means. Almost nobody notices when there are no stars at all — and that turns out to describe most of the cars on sale.

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.

