Two countries share a border and, for the most part, the same vehicles. When the same defect is recalled on both sides, each country’s notice says how many vehicles it covers. Are the two numbers in proportion?
Nobody publishes the answer, because answering it means proving that a Canadian notice and an American notice are the same recall. They share no identifier, but they do share the manufacturer’s own campaign code, and this site already joins the registers on it, exactly, for the crosswalk.
Mostly, yes, and that is worth saying first
Across 1,889 matched pairs that carry a unit count on both sides, the median Canadian recall covers 9.8% of the vehicles its American twin does. Canada’s vehicle fleet is usually estimated at about a tenth of America’s. The two registers, in the middle of the distribution, recall in proportion.
| Canadian notice, as a share of the American one | Pairs | Share |
|---|---|---|
| under 5 per cent | 442 | 23% |
| 5 to 10 per cent | 522 | 28% |
| 10 to 20 per cent | 521 | 28% |
| 20 to 100 per cent | 319 | 17% |
| Canada recalled more | 85 | 4% |
One pair in ten sits below 2.3%, and one in ten above 39.9%. The middle behaves. The tails are the page.
85 times, Canada recalled more cars than America
Not more in proportion. More in absolute numbers, from a market a tenth the size, for the same defect under the same manufacturer campaign. 85 of the 1,889 pairs are like this.
| Maker code | American campaign | Canadian recall | US vehicles | Canadian vehicles |
|---|---|---|---|---|
| SC0472 | 25V352 | 2024632 | 5 | 6,855 |
| F1027 | 26V029 | 2025610 | 2 | 2,464 |
| 25S52 | 26V239 | 2025281 | 55 | 40,428 |
| SR14005 | 14V365 | 2014248 | 5 | 3,549 |
| 202301 | 23V408 | 2022605 | 25 | 6,929 |
| 25C69 | 26V301 | 2025691 | 208 | 50,037 |
| CR5512 | 23V781 | 2023622 | 2 | 442 |
| CR1366 | 15V078 | 2009200 | 5 | 359 |
| F1028 | 25V850 | 2025678 | 2 | 131 |
| 14515 | 14V606 | 2014432 | 78 | 5,071 |
| RP1061 | 25V409 | 2025333 | 2 | 113 |
| 25SA4 | 26V012 | 2025554 | 2,403 | 128,886 |
A pair like this is a fact about the two notices, not by itself a verdict on either regulator. A model can sell in one market and barely exist in the other; an American notice can be an expansion of an earlier campaign; a Canadian one can sweep in model years the American one already covered. The table exists so that the question can be asked recall by recall, with both file numbers in hand, which until now it could not be.
And the other tail
At the far end sit pairs where the Canadian notice is a fraction of a per cent of the American one: a handful of vehicles against thousands. Most will be honest differences in what was sold where. The point of listing them is that an owner in one country can now see what the same defect looked like in the other.
| Maker code | American campaign | Canadian recall | US vehicles | Canadian vehicles |
|---|---|---|---|---|
| SC250 | 22V760 | 1995134 | 69,038 | 2 |
| SC138 | 16V815 | 1981079 | 71,704 | 6 |
| 20S15 | 23V775 | 2020131 | 45,173 | 4 |
| FL664 | 14V566 | 2015482 | 6,759 | 2 |
| 14549 | 14V544 | 2014373 | 7,598 | 3 |
| 25S20 | 25V163 | 2025133 | 11,621 | 5 |
| SR25040 | 25V670 | 2025536 | 22,319 | 10 |
| 15C13 | 15V707 | 2015508 | 1,900 | 1 |
What these numbers are, and are not
The two figures are adjacent instruments. Canada publishes units affected; America publishes potentially affected units. They measure the same idea from neighbouring angles, and this page compares them without pretending they are one instrument. Where an American campaign carries more than one figure in the flat file, the larger is kept, so the American side is never understated in the comparisons above.
Nothing is paired on brand and date. A pair exists only when the manufacturer’s own campaign code appears in both registers, normalised the same way and matched exactly. Pairing on brand and date is how a finding gets manufactured rather than discovered.
Where the join failed, it failed openly. 19 matched campaigns carry no positive American unit count and 6 no Canadian one. They are excluded from every figure above rather than counted as zero.
Where the numbers come from
- Transport Canada, Vehicle Recalls Database, the full monthly open file. The unit count is the recall’s own figure, never a sum across its make-and-model rows, which would triple-count it.
- NHTSA flat recall file, post 2010, the copy this site already keeps for the register fingerprint.
- The engine reports itself at its status route, including every figure on this page and the full list of pairs where Canada recalled more.
- The filings belong to the two regulators. The join, the arithmetic and any mistake in them are ours.
Built 8 September 2026. Corrections go on the corrections page.
Canada and America · Europe and America · Download the record · How we work