Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
The UK’s main employment dataset is infamously borked, with the Labour Force Survey that underpins it having fallen victim to a slow, then quick decline in response rates.
It’s been an ugly episode for the Office for National Statistics, which was apparently unable to correct course before disaster. But if there’s one thing to be said in their credit, it’s that they were totally transparent about things going wrong. Every monthly LFS publication was accompanied by data showing broadly dropping response rates, until they dropped too far and the whole thing flopped:
Various hypotheses have been put forward as to why things got so bad. Popular among them are that the LFS survey process is too arduous, and that people don’t pick up the phone anymore. Here’s a quote:
The response rates for social surveys in general, but especially for the LFS as the UK’s largest continuous survey, may well be suffering from widespread apathy and mistrust from the British public… a large and increasing proportion of non-response is due to potential respondents not believing in surveys, not trusting the government and generally not being bothered
…from 2008, which gives a small sense of how long this has been an issue.
Which is why we’ve always been curious about the comparatively excellent response rates European statistics agencies have been able to maintain, given they face the same notional challenges. Are British people atypically unhelpful when the stats agency comes a-knocking?
Publication of these response stats across individual agencies is mixed at best, and as far as we can tell none are as transparent as the ONS. So the easiest way to look at these stats en masse is by waiting for Eurostat — the European Commission’s stats body — to collate them. As we whinged earlier this year, this can take a long time.
Last month, Eurostat (finally) released the 2025 edition of its LFS quality report, which covers 2021 and 2022. (The delay was somewhat understandable — this edition was especially tricky because of changes to the surveys in 2021.)
Creating the UK’s non-response levels by taking the average of the inverse of the total response (including imputations), over the year in question, here’s how it sat relative to the rest of geographical Europe in 2021 and 2022:
Badly. It sat badly, even before the tipping point into actual failure. There are various caveats to be added here, most importantly:
Belgium, Germany, Spain, France, Italy, Cyprus, Luxembourg, Malta, Austria, Portugal, Slovakia, Norway, Serbia and Türkiye enforce compulsory LFS participation.
And less importantly:
Denmark, Estonia, Luxembourg, Netherlands, Finland, Sweden, Norway and Switzerland all track non-response at the individual level, rather than the household level.
In one respect, it confirms what we already knew: the UK has a uniquely acute LFS response problem. Europeans, it seems, simply do it better — in part, by often not really offering much of a choice.
But compulsory participation is an incomplete excuse: all the other countries with voluntary LFS participation are also doing better than Britain. What makes the ONS so much worse at this than every European national stats agency?
Another dataset within the Eurostat quality report deals with annual averages of the reason given for non-response. Frustratingly, the ONS and Eurostat non-response categories vary:
The ONS tracks: Circumstantial refusal, Outright refusal, Refusal to headquarters (when a household contacts the local field office ahead of receiving to survey to refuse to respond), Refusal to re-interview and Non-contact
Eurostat tracks: Refusals, Non-contacts, Other reasons.
This renders the two datasets only somewhat comparable. But we can sort the countries with voluntary participation by refusals as a proportion of non-responses…
…and see the UK may have a distinctly bad refusal problem: Brits, offered the chance to help their government gather stats, are the most likely to straight-up refuse (and by some distance).
This needs a pinch of salt: those “other reasons” could mean almost anything, and the distribution across the categories is inconsistent enough by countries that we’re sure this data is a bit tricksy. But the trend is there.
All of which is a medium-length-winded way of saying: at the last point in time for which we have passable data, the UK was indeed uniquely bad (within Europe) at conducting the LFS. And maybe, possibly, a factor in this is that British people are particularly uncooperative.
Further reading:
— Eurostat vs Eurostat stats
†OK fine the Dua Lipa reference here doesn’t make much sense because refusing to pick up the phone would presumably be treated as a non-contact, but we couldn’t find a high-res screenshot of the “This is my own private domicile and I will not be harassed” bit from Breaking Bad. Don’t let anyone tell you blogging is easy.