Skip to the recordSkip to the page
CARSMULTIVERSEWatch a recall

Two countries share a border, a car market and, for the most part, the same vehicles. When a manufacturer finds a defect, both sets of drivers need telling. Are they told at the same time?

Nobody publishes the answer, because answering it means proving that a Canadian notice and an American notice are the same recall, and the two registers share no identifier. They do share one thing: the manufacturer’s own number for the campaign.

The finding

1,825 recalls can be matched between the two countries on that number, exactly. The median gap between the American filing and the Canadian one is 0 days.

Who was told first Recalls Share
America first 511 28%
Canada first 388 21%
Same day 926 51%

38 of the 1,825 are more than ninety days apart, 2 per cent. That is a quarter of a year in which one country’s drivers knew about a defect and the other country’s did not.

The headline, though, is the other way round: half of these recalls were filed in both countries on the same day, and the median gap across all 1,825 of them is zero. North America is, on this evidence, one recall market with two postboxes. The interesting cases are the 38 that are not.

By year

Dated by the American filing, from 2010, which is where the American bulk record begins.

Year Recalls matched Median gap, days
2012 1 657
2013 12 2
2014 213 1
2015 212 3
2016 207 2
2017 3 1
2018 3 1
2019 2 1
2020 4 3
2021 1 1
2022 199 0
2023 233 0
2024 267 0
2025 321 0
2026 147 0

The middle years of that table are thin, and that is about us, not about them. A row can only appear here if the American campaign states its maker code in the remedy text and we hold an American filing date for it. Both of those are patchier for the second half of the twenty-tens than for either side of it. The counts in the middle column are therefore a measure of what can currently be matched, not of how many recalls happened, and no year-on-year trend should be read out of them. The median gap in the right-hand column is the column worth reading: it is computed only from the rows that year actually has.

What the match is, and what it is not

It is the manufacturer’s own campaign code, normalised the same way on both sides and matched exactly. Ford’s Canadian filing and Ford’s American filing both say 26S31, so they are the same recall. Nothing here is paired on brand and date. Pairing on brand and date is how a finding gets manufactured rather than discovered.

The two dates are not the same instrument, and this page will not pretend otherwise. Canada records the date the recall was entered in its database. America’s date is the Part 573 report: the day the manufacturer told the regulator. They measure adjacent moments, not identical ones, so a gap of a day or two means nothing. A gap of months is the thing worth looking at, and that is why the ninety-day figure is quoted above rather than the average.

Where the match failed, it failed openly: 48 matched campaigns have no American filing date held here, and 0 have no Canadian date, 1,589 match only on a code of three characters or fewer, which is not an identifier, and 42 pair two dates more than two years apart, which is far more likely a reused campaign code than one defect recalled twice. All four are excluded from every figure above rather than counted as zero-day gaps, which would drag the median towards a flattering answer.

Where the numbers come from

Source: Transport Canada Vehicle Recalls Database and the United States NHTSA recall record. The filings are theirs; the matching, the counting and any mistake in them are ours, and go on the corrections page.

Europe and America · Who gets told first · Sixty days to tell you · Check your car

A project by A.I.T. Multiverse Consulting Ltd · cars@aitmultiverse.com