Each month, I comment on the Office for National Statistics’ labour market overview for the IEA. The data provide insights into the UK's employment, unemployment, earnings, and other related metrics, influencing decisions by the Chancellor and the Bank of England.
These figures are primarily sourced from the Labour Force Survey, a respected and widely used voluntary household survey. However, this month, major problems have arisen, prompting the ONS to delay publication to address and recalibrate the data.
The response rate to the LFS has plummeted, dropping from over 60% a few years ago to 40% just before Covid to around 15% now. While a uniform decrease across the population would probably widen the margin of error without changing the trend, the decline has been uneven. Older individuals and those home all day have higher response rates, introducing sample bias and potentially skewing the overall picture. This bias is evident in the discrepancy between LFS employment figures and other sources like company payrolls.
Now I’m not telling you this because I’m a statistical nerd. Some of you might dispute the need for the LFS, perhaps echoing the great Sir John Cowperthwaite, who as Financial Secretary in Hong Kong famously refused to collect economic data in case it encouraged people to demand more public spending. Perhaps.
But it’s interesting for a couple of reasons. One is that our increasingly diverse society makes collecting reliable data on the population more difficult, with some communities less likely to reply to questions, possibly because they fear government acquiring information about them even though answers are anonymised. We saw this with the Census.
Two, it’s another example of the long-term fallout from Covid. Previously, the survey was first administered in-person (respondents are asked to repeat the survey four times over the next fifteen months). Someone came to your door, had a cup of tea and went through the questions with you.
Covid put the kibosh on this as we couldn’t let anybody in: phone calls had to suffice. This has continued post-Covid – no doubt welcomed by survey administrators who, like an increasing number of government employees, can now work at home rather than trudge round to people’s flats and houses. But relying on phone calls is difficult. Contacting people by phone is increasingly problematic, and those who are available and prepared to answer questions for up to an hour in this way are not typical of the population as a whole.
So, although there are no questions about this in the LFS, the problems that the ONS is having shed some unintended light on the way Britain is changing.
