Every comparison anyone publishes between countries is a comparison of volume: how many recalls each one had. That is a fact about the car industry. The more useful fact is about the regulator.
When your country publishes a recall, how much does it actually tell you? Does it say what the danger is, or only which part broke? Does it say how many cars? Does it say when your letter is coming? Does it ever say whether the repair happened?
Nobody has put the registers side by side, so a reader in one country has no way to know what a reader in another is being told. This is four of them.
The four registers, and what we hold from each
| Register | Records held here |
|---|---|
| United States | 15,204 |
| United Kingdom | 8,944 |
| Netherlands | 1,500 |
| Canada | 143 |
These counts are read from each engine every time this page is rebuilt, so it cannot drift away from the record it describes. They are what we hold and read, and they are not a league table. The Dutch figure in particular is the number of actions we keep full wording for, not the size of that register: it publishes several thousand more. Read this table as the reach of our own reading, and nothing else.
What each one tells you
| Does the published record tell you… | US | UK | NL | CA |
|---|---|---|---|---|
| What the part is, and what failed | yes | yes | yes | yes |
| What could happen to you if it fails | yes | no | yes | no |
| What the fix is | yes | yes | yes | yes |
| How many vehicles are affected | yes | no | no | no |
| When owners will be written to | yes | no | yes | no |
| When the manufacturer reported it | yes | yes | yes | yes |
| The manufacturer’s own reference for the recall | partly | no | yes | no |
| Whether the cars were actually repaired | partly | no | no | no |
| What the notice said before it was edited | no | no | no | no |
Every row, in words
What the part is, and what failed
Every register describes the defect. This is the floor, and all four clear it.
What could happen to you if it fails
The consequence is a separate field in the American and Dutch records: “can increase the risk of a fire”. The British and Canadian records we hold describe the fault and leave the reader to work out what it means.
What the fix is
All four say what will be done to the car.
How many vehicles are affected
Only the American record carries a vehicle count. Without it there is no way to tell a recall of forty cars from a recall of four hundred thousand, which is the most basic question anyone has about any recall.
When owners will be written to
Two of the four record an owner-notification date. It is the only field that tells a reader when to start worrying that nothing has arrived, and it is the field that made the cross-border comparison on this site possible at all.
When the manufacturer reported it
All four date the filing. Only the American record pairs that date with a notification date, which is what makes a delay measurable rather than merely suspected.
The manufacturer’s own reference for the recall
A field in the Dutch record. In the American record it is a sentence inside the remedy text, present only when the maker chooses to write it. Nowhere in the British or Canadian records. This one field is the only reason any two registers here can be joined at all.
Whether the cars were actually repaired
The United States publishes completion rates once a year, in a PDF, for a subset of campaigns, not in the register itself. No other register here publishes them at all. A recall can be announced, closed, and half unrepaired, and only one country lets you find out.
What the notice said before it was edited
No register keeps the earlier wording. A register states what is true now, which is a reasonable thing for a register to be. It also means a notice can be rewritten after publication and leave no trace, which is why this site fingerprints every notice on first read and compares it on every read since.
How this was established, and what it does not claim
By reading the records themselves, not the authorities’ documentation about themselves. A field is marked yes only where it carries a value in the records in our store, partly where it exists but not reliably or not in the register proper, and no where it is absent.
That is deliberately a narrow claim. It is about what arrives in the published data (the thing a reader, a journalist or a researcher can actually use), and not about everything an authority holds internally. An authority may well know how many vehicles are affected and simply not publish it. From outside, those two situations look identical, and this page can only speak to the outside.
The field structure of a register changes rarely, so unlike the counts above these rows are not recomputed on every build. They were established by reading the stores on 24 August 2026. If you find one that is wrong, the corrections page is where it will be fixed and said so.
Who gets told first · The same fault, in two authorities’ words · What changed after publication · How we work