The Labour Force Survey asks people whether they have done any work and, if not, whether they have looked for work and are available to start. The classification of employed, unemployed or inactive is derived from that

People out of work who are not looking for work.

More precisely, those out of work who have looked for work in the last four weeks and are available to start within two weeks are classified as 'unemployed'. Everyone else out of work is classified as 'inactive'.

Overall summary: when it comes to inactivity, yes of course we should do better, and many people would benefit if we do, but this is not a new crisis.

The inactivity rate for people aged 16 to 64 in July 2025 was 21%, lower than at any point before Nov 2018. Since 1971, there has been only one 15-month period – Dec 2018 to Feb 2020 – when inactivity was lower than today.

Personally, I think the London Science Museum has always been really poor: over some years, we've taken our daughter to science museums in Copenhagen, Paris, Athens, Boston, Glasgow - and loved them - all better for narrative, explanation of the underlying science - more interesting for her (and us)

The DWP (and I think HMRC) had a major statistics hub in Newcastle (at Longbenton) and so would have made a much more sensible relocation than Newport. Would have allowed movement between organisations as part of career progression

We don't really have a definitive explanation for inactivity being higher than pre-pandemic. Other possible explanations include healthcare backlogs so more people too ill to work or that, post-pandemic, more people decided to accept a lower income in return for more time with their grandkids

But it is VERY far from a jobs crisis (as one journalist put to me in an interview this afternoon). 🧵 /end

Finally, average wage growth has slowed to 5% a year. This is still comfortably above price growth (inflation) but by slightly less so. This is evidence of a degree of restraint from employers as they deal with minimum wage rises and employer national insurance increases. 🧵 /11

So we have two contradictory pictures: employment up according to the Labour Force Survey but number of employees down according to HMRC data: that’s the key info to report, not the unemployment rate whose importance depends on what’s happened to inactivity/employment. 🧵 /10

So they recommend looking at stats on the number of employees reported by their employers to HMRC for shorter-term trends. This shows a small fall in the number of employees in the most recent period. 🧵 /9

But – yes there is a but – there have been problems with the Labour Force Survey in recent years. Not enough people have responded to the survey, increasing uncertainty around its results. The ONS have been working on this but short-term comparisons still suffer from volatility. 🧵 /8

Because of these complexities, the unemployment rate should NOT be the focus of reporting every month – the employment rate is much more useful. What matters most is if people are working, not how hard they are trying to find work, which is what the unemployment rate measures. 🧵 /7

Essentially, according to the Labour Force Survey, the rise in unemployment came from a drop in inactivity, not a drop in employment. If more people who were previously not doing so start looking for work, that’s a good thing. 🧵 /6

When you do this, despite unemployment rising a bit, the Labour Force Survey employment rate rose a bit as well. How is this possible? Because the inactivity rate fell quite a bit – by 0.4 percentage points. 🧵 /5

That’s why the government has made a priority of supporting people who are classified as ‘inactive’ to look for work. So, reporting of the monthly labour market stats should NOT focus just on the unemployment rate but on the employment rate so that the inactivity issue doesn’t get ignored 🧵 /4