National Statistical

The ONS uses various data sources to measure changes in the UK labour market. One of those, the ‘Real Time Information’ from the HMRC PAYE system, shows a sharp drop in the number of employees, but the Labour Force Survey does not. Here Jonathan Athow looks at a possible explanation for this seeming contradiction.

We currently collect our main labour market data through the Labour Force Survey (LFS), which covers around 35,000 households a quarter. Our team of skilled interviewers ask people a range of questions: are they working? And if so, how many hours do they work? Are they self-employed or an employee? And much more besides.

Most of the time, this approach gives us a very good and detailed picture of what is happening with our statistics based on the experiences of people across the country. However, the coronavirus pandemic has proved challenging to the collection and production of statistics. We have had to measure new concepts for the UK such as furloughing, where people are temporarily away from work on full or partial pay. They remain contracted to an employer and in receipt of pay. We count these people as in work, which is in line with internationally-agreed definitions

But the PAYE data showed that between March and June, the number of employees being paid fell by around 650,000. The number of people in work as measured by the LFS in May was only around 180,000 lower than March, and if we look on a quarterly basis the number of employees has edged up. How can this seeming contradiction be explained?

In fact, thanks to an extra question added to the LFS at the start of the pandemic, we have been able to identify a group of people – almost half a million for May  – who tell us they are in employment, temporarily away from work because of the pandemic  but not being paid at all. This is confusing, as this is outside the Government’s Coronavirus Job Retention Scheme, and raises the question of whether these people are really employed as we would normally understand it. From the data we have, we cannot really answer that question. However, we think they might be workers who for one reason or another fall outside the scheme but who believe they will have a job for them to come back to, such as informal bar staff previously working in a now closed pub, and so they are saying they still have a job.

Our approach is therefore to take these people’s answers to our surveys at face value: after all, people will be best placed to understand their own circumstances. So we include these people in our LFS measure of employment.

However, as these people do not meet many of the traditional definitions of being employed,  we have separately identified them in our statistics.

The size of this group of people helps to explain much of the difference movements  in the number of employees paid through the payroll, measured using the PAYE  data, and the number of employees measured using the LFS. It could also explain differences between the LFS and the claimant count measure, as these people could be counted as employed by us but eligible for Universal Credit.

Measuring large changes in the labour market through a global pandemic is clearly a challenge but we are working hard to paint the clearest picture possible.

Jonathan Athow is Deputy National Statistician for Economic Statistics at the ONS