The importance of digging beneath the headline measure on employment

It may have reversed very slightly in recent months, but the strong employment growth recorded over the last year or two in the UK remains little short of remarkable. The number of people in work topped 31 million in March, establishing a headline employment rate of 73.5 per cent – both historic highs. If not quite surpassing these records, this week’s updates for June should at least be close to these landmark values. Following a particularly severe and protracted economic downturn these numbers make for impressive reading. But just how meaningful are they?

The absolute figure of roughly 31 million gives a clear sense of the size of the British workforce but it would be wrong to read too much into it, affected as it is by population growth. The employment rate – which captures the proportion of 16-64 year-olds in employment – provides a better means of assessing the health of the labour market over time. Accordingly, much has been made of the recent record breaking performance on this measure. Yet we should show some caution because somewhat different pictures can be painted if we alter the population across which we measure the employment rate.

Previously the headline rate of employment was 16-59/64, with the upper age limits reflecting the differing state pension age (SPA) for women and men. With the female SPA rising gradually from 2010 (it currently stands at 62-63), the ONS opted to engage in some form of future-proofing by moving to a 16-64 measure. This has the merit of being clear, and was the product of a considered review of options. But it will date in the coming years, particularly once the SPA for both men and women moves beyond 65.

Even in the shorter term it has the potential to miss some important trends. At the lower end of the age scale it fails to account for a steady rise in the age at which people leave education – in part because of legislative changes and in part because of increased participation in higher education. And at the higher end it fails to capture increasing employment among adults above the SPA.

And of course, increases in the female SPA since 2010 are likely to have had the effect of increasing participation and employment among older women (and indeed among older men who are less likely to retire while their partners remain in work). By sticking to a 16-64 measure throughout this period, some of the employment strength we’ve seen is likely to reflect this change in policy rather than changes within the labour market.

To better capture underlying factors, we can instead look at the employment rate for the 16-SPA population, where the level of the SPA is gender specific and alters over time (effectively reverting to the old headline measure, but shifting the coverage each year to reflect the rising female SPA). Focusing on this measure of course has the effect of increasing the overall employment rate (because we are removing women aged between the SPA and 64, among whom employment is inevitably lower than average, from both the numerator and denominator).

Another approach is to consider the entire 16+ population. The logic here is that the abolition of the default retirement age means there is less reason than in previous periods to assume the SPA should be synonymous with leaving the labour force. This focus provides a better sense of the tax base and dependency ratio. For example, a dramatic fall in the 16+ employment rate might occur because of a sudden shift into retirement among baby boomers – not a comment on the labour market per se, but an important consideration for policymakers nonetheless. Clearly though, the overwhelming majority of those aged 65 and over are retired, so stretching the employment rate measure in this way will inevitably lower it.

These three options – the headline 16-64 rate, the 16-SPA rate and the 16+ rate – are all shown on the chart below. What’s clear is that, despite differences in their levels, all show similar trajectories. The recessions of the early 1980s, early 1990s and late 2000s in particular are very clearly visible.

As noted above, the 16-64 rate stood at 73.5 per cent in Q1 2015, a record high. While higher than the 16-64 rate, the 16-SPA figure is still below its 2000-2008 average, and considerably below its early-1970s peak. The 16+ rate is back to what looks like a historic norm of 60 per cent, though this is 1.1 percentage points below its peak. Whichever approach we take however, the sharp improvement in the employment picture in recent years is clear to see.

Demographic and other changes mean no headline measure will ever be perfect and international comparisons in particular (which typically focus on the 15-64 rate) are difficult to draw. Each of the options set out above have their own merits, and shortcomings. But it’s likely that we’ll need to adopt a ‘horses for courses’ approach, using different measures for different policy purposes. There are perhaps three obvious variants.

First, when considering labour market strength it’s important to retain a measure that sets sensible lower and upper age limits. In the short-term, 16-64 is likely to be as good a measure as any, but clearly this will need to evolve. Removing – or counting separately – those aged 16-18 would help at the lower end of age spectrum; while raising the upper boundary will quickly become necessary. The 16-SPA measure provides a useful supplement in the short-term by better accounting for the gradual change in the female SPA.

Secondly, in order to provide an indication of fiscal sustainability and account for changing demographics we can focus on a measure with no upper limit. For the moment that means looking at the 16+ employment rate, though again we might want to raise the lower age threshold over time.

Finally, for those wanting to push beyond ‘strong’ employment growth towards some measure of ‘full employment’, it is important to better understand the performance among sub-groups. As we showed last month, boosting participation among single parents, those with ill health and disabilities and the low skilled could go a long way to securing the Chancellor’s ambition of generating an extra 2 million jobs by the end of the decade. Shining a light on the labour market performances of these groups can therefore play an important role.

More generally, from the perspective of achieving full employment there are further opportunities to be creative; aligning the construction of statistics with the significant policy challenge. For example, we’ve taken the approach of excluding full-time students from some of our analysis, something that could become more standardised. We might similarly choose to include workers aged 65 and over in our employment count, but exclude the 65+ population from the denominator. This would of course create something of an artificial construct – boosting the reported employment rate – but it would have the merit of describing just how important this group of workers is becoming to the overall total. Other options exist.

What’s clear is that there are many different ways of counting employment and that scratching beneath the headline can be important. As with the approach to defining full employment, it’s a question to which we will return over the course of our project.