How to read a jobs number, using today's as the worked example

This is a practical note for anyone who reads a jobs headline and wants to know how much weight it will take. It uses today's number as the example, but the method is general.

One. Ask which of the three sources it came from.

Britain measures work three ways and they are not interchangeable.

PAYE Real Time Information comes from tax records, filed every time an employer runs payroll. Nobody has to agree to take part, so there is no response rate to collapse, and it is the firmest count of people on payrolls there is. It cannot see the self-employed and it cannot see unemployment.

The Claimant Count is people claiming unemployment benefits. Also from records rather than interviews, also firm, and it measures who qualifies for a benefit rather than who wants a job. Change the rules and the number changes.

The Labour Force Survey is a survey of households. It is the only one of the three that can produce an unemployment rate, an inactivity rate or a count of the self-employed. The ONS says so in terms: for those three things it is the only source.

If a figure is about unemployment, it came from the survey. There is no second opinion to check it against.

Two. For a survey number, look for the response rate and who answered.

The ONS publishes both, every quarter, in its Labour Force Survey performance and quality monitoring report. It is free and it is not written to be hard.

Two things in it are worth a minute. The headline response rate, which was 25.1 per cent for Great Britain in the first quarter of this year. And Section 5, where the ONS charts who is actually answering. Its own chart title currently reports more respondents over 55 and fewer under 24 than before the pandemic.

That second one matters whenever the number you are reading is about a particular age group.

Three. Check who answered the question.

Buried in Table 7 of the same report is the proxy response rate: how often somebody else in the household answered on a person's behalf. Overall it is 36.6 per cent. For 16 and 17 year olds it is 95.3 per cent, and for 18 and 19 year olds 87.5 per cent.

So a statistic about teenage employment is largely a statistic about what parents say about their children's employment. That is not a scandal, it is how household surveys work, and it is a reason to be careful with small movements.

Four. Ask whether a second, unrelated source points the same way.

This is the most useful habit. If the survey says young people are struggling and the tax records say the same thing, the finding is sturdier than either source alone. If only one says it, hold the number more loosely.

It works today, partly. The payroll figures show under-25s falling by 115,000 in a year while the 35 to 49 group rose by 87,000. That matches the direction. It does not match the year, and it cannot confirm an unemployment rate, because it does not measure unemployment.

And one thing this note will not tell you.

Whether the gap reaching a record in 2024 means what it appears to mean. We do not know, we say so in the lead, and no amount of method will substitute for the work nobody has done yet.

The Quernal provides general information, not financial advice.