The labour market is generating fewer jobs, but paying better for those who have one


Today’s labour market data paint a picture of a labour market that is generating fewer jobs, but paying better for those who do have one. This picture is hazy, however, because the data are themselves increasingly uncertain.

The employment rate is falling

The 16-64 employment rate in the ONS’s Labour Force Survey (LFS) was unchanged at 74.5% in the three months to March (see chart below, left panel). But it has been flat or falling for the past six months, and is now much closer to the pandemic low point (74.1%) than its pre-pandemic level (76.2%). This stands in sharp contrast to most other high-income countries, where employment rates substantially exceed their pre-pandemic levels.

For some four years now, this has been a story of weak labour supply – i.e. rising labour market inactivity (chart above, right panel). But more recently, weak labour supply has been reinforced by weak labour demand. Vacancies peaked around two years ago, and unemployment has risen by half a percentage point over the past four months (chart above, middle panel).

This weakness in labour demand is in part an echo of the recession we had late last year, which is now thankfully at an end. Unemployment tends to track GDP growth with a lag (see chart below). So it’s possible that unemployment will start to fall again if the economy keeps growing at its current rapid rate.

Real wages are rising

Set against this weakness in labour quantities (at least in terms of rates – see below for our take on data uncertainty), the price of labour (i.e. wages) has been rising strongly, in both cash and real terms. Nominal pay growth was little changed in the latest data (unchanged overall, and actually up slightly, to 6.3%, in the public sector). With inflation continuing to fall, this means real wages grew at a healthy 2 per cent in the 12 months to March 2024, much faster than the post-financial crisis trend. Real wages have grown almost as much in the past 12 months as over the previous 16 years in total (see chart below).

These increases in real pay are even more striking when set against the fall in productivity in 2023 (down 0.6 per cent in heads terms). For the time being, employers can afford these higher real wages because other costs – pension contributions and the price of imports – have been falling. But we’ll soon need to see productivity growth pick up if we want these real wage increases to continue. 

The data are uncertain

This month we have not used any Labour Force Survey data to describe the levels (i.e. absolute numbers) of employment, unemployment and so on. That is because the numbers it describes are quite different to the count of employees in the PAYE tax system (as well as to the ONS’s own Workforce Jobs series, which is based on surveys of employers). In particular, it is the difference between estimates of the change in the number of jobs since 2021 in those datasets which is most surprising. According to the PAYE data, the number of employees has grown by 2.1 million since the start of 2021; in the same period in the LFS data the number of employees is estimated to have grown by 840,000.

Logically speaking, the differences between the PAYE and LFS data can come from three places. First, it could be that they are measuring slightly different things, or people aren’t answering the survey properly. For example, early in the pandemic we saw that lots of people who thought they were self-employed turned out to be employees when they started getting furlough payments. 

Secondly, it could be that the rates the LFS measured are biased by differences in response rates between people with different labour market status – for example, by employed people being less likely to answer the survey – in a manner that the ONS cannot or do not control for. The response rate of the LFS is now around 15%, making it very easy for non-response to muck things up.

Finally, the ONS multiply their estimates of rates in the LFS by their estimate of population numbers to get levels, and it’s possible that these population numbers are too low. As the Bank of England recently noted, the latest population projections point to rather faster population growth than the numbers currently used in the LFS.

We’ll return to this important issue ahead of next month’s Labour Force Statistics release.