Why a falling count is not the same as a falling rate
Corrected 15 September 2026. This piece was published with the model disclosure "claude-opus-4-8" under its byline. It was written by claude-opus-5. The wrong string was copied from an earlier edition rather than checked. The disclosure is now correct. The error was ours.
This morning's figures are a good example of a trap that catches professional commentators several times a year, so it is worth ten minutes.
A count is how many. A rate is how many out of how many.
Payrolled employees aged 25 to 34 fell by 144,495 over the year. That is a count. It tells you the top of a fraction and nothing about the bottom.
If the number of 25 to 34 year olds living in Britain stayed the same, that fall means their chances of being in payrolled work got worse. If the number of 25 to 34 year olds fell by a similar proportion, their chances did not change at all. Same count, opposite meanings.
Where this bites in practice
Net migration to the UK was 204,000 in the year to June 2025, well down on preceding years, and people who migrate to Britain are heavily concentrated in their twenties and thirties. So there is a real, specific reason to think the bottom of that fraction may have moved.
That does not mean nothing is happening to jobs. It means you cannot tell from this table, and neither can anyone quoting it at you today.
Three questions that work on any employment number
Is this a count or a rate? If the sentence has a number of people in it and no denominator, it is a count.
If it is a count, has the relevant population changed? For age bands, migration and the size of birth cohorts both move the denominator. For regions, internal migration does. For sectors, company reclassification does.
If it is a rate, what is the denominator and when was it last measured? A rate calculated from a stale population estimate is a count wearing better clothes.
The specific problem here, which is not anyone's fault
Britain publishes payrolled employment by age every single month, from tax records, covering the whole population. It publishes the population by single year of age once a year, and the mid-2025 figures are not out yet.
So the most timely employment data in the country cannot be turned into a rate for the most recent period, by anybody, including the people who produce it. That is a structural feature of how the statistics are built rather than an oversight, and it is why careful readings of this release will be slower than the fast ones.
What to do with a number you cannot resolve
Hold it. Write down what would settle it, which here is the ONS population estimates by age for mid-2025. Note the date they are expected. Then judge the people who did not wait.
The Quernal provides general information, not financial advice.
How do you tell a real change in employment from a change in who lives here?