Should the UK government get active about economic inactivity?

Concerns about a rise in economic inactivity have prompted new policies aimed at discouraging early retirement and encouraging the labour force participation of women with young children. Such interventions would be more cost-effective if targeted at particular disadvantaged groups.

At less than 4%, the UK unemployment rate is still low. But this doesn’t tell the whole story about the state of the labour market. Last week’s fall in the UK’s rate of economic inactivity – the proportion of the working age population that is neither in work or looking for work – opens up new policy questions about inactivity and its effects on the performance of the UK economy.

Between the start of the pandemic in early 2020 and the end of 2022, economic inactivity rose by around 600,000. Although it has since fallen back by around 300,000 (see Figure 1), this fall in labour supply is a potential problem because:

Fewer people offering their labour restricts the capacity of the economy to grow, in the absence of offsetting movements in productivity and/or immigration.
Fewer people available for work could exacerbate skill shortages.

Figure 1: UK economic inactivity, 2019-2023

Source: Office for National Statistics (ONS)

The Bank of England appears to pin rising inactivity on a ‘change to the supply of labour, independent of demand’. According to this view, such a fall in labour supply would then put upward pressure on wages. This could in turn contribute to further cost-push inflation at a time of already double-digit inflation. If true, this can be seen as justification for a policy of raising interest rates.

At the same time, in the March budget, HM Treasury introduced a wave of measures aimed at addressing inactivity, including:

A £1 billion a year package offering tax relief on pension contributions.
A £4 billion a year package offering free childcare places to parents with children aged nine months to 3 years old.
A £0.6 billion a year package encouraging people who are disabled and sick to find and stay in work.

Most students of economic policy are taught about ‘deadweight loss’ – expenditures that are largely wasted because the actions they seek to address would have happened anyway in the absence of the policy. A better and more cost-effective solution would be more targeted or ‘marginal’ policies, aimed at groups that are most likely to benefit from intervention and whose circumstances would not change in the absence of the policy.

Was the Bank right – and did the Treasury introduce the right policies to address inactivity?

The first point to make is that context is everything. Figure 2 shows the UK inactivity rate over the past 50 years. The current inactivity rate isn’t that different from those in the 1970s. It was 25% of the working age population in 1971 and 22% in 2022. Since the 1980s, each recession has been associated with a rise in the aggregate inactivity rate.

But recent history also shows that during an economic recovery, the inactivity rate typically falls. This suggests that the inactivity rate is positively correlated with the level of demand or activity in the economy. The most recent rise, which has garnered much attention, started during the Covid-19 crisis. Despite the recent concern, the current inactivity rate is now only back to where it was in 2016, and it is still below the level seen at the tail end of the Great Recession of 2008-11.

But the path of the aggregate inactivity rate over time masks two contrasting patterns across gender. The inactivity rate has been rising for men (dark blue line) and falling for women (turquoise line) over the past 50 years. Broadly speaking, the two opposing trends cancel each other out, leaving much more muted changes in the aggregate inactivity rate.

Nevertheless, Figure 1 shows that the long-term rise in the share of women entering the labour force is dampened during economic downturns, and that the inactivity rate for men tends to ratchet up after (or during) each recession. Both these features might be expected when job opportunities are scarcer.

Figure 2: UK inactivity rate by gender, 1971-2023

Source: ONS

If rising inactivity were associated with local labour shortages that drive up wages, we would expect to see a positive relationship between wage growth and inactivity growth.

Figure 3 shows the association between wage growth in local areas and the growth in local area inactivity rates over the period between 2014 and 2022. Each dot on the graph plots the change in gross monthly wages and the change in the inactivity rate for 16-64 year olds in each of 41 NUTS2-level areas of the UK (corresponding approximately to areas with a population of around 1.5 to 2 million people).

Figure 3 (red line) shows that there is in fact a negative association between wage growth and inactivity growth. Areas where inactivity has grown most have experienced smaller wage growth. Areas where inactivity has fallen most have experienced larger wage growth. This trend applies both over the medium term and during the pandemic (not shown).

Figure 3: Wage growth and inactivity growth across local areas, 2014-2022

Source: ONS, authors’ calculations

To see why inactivity may not be entirely a labour supply effect, it is helpful to break down the main drivers of inactivity over time.

Figure 4 shows that there are three main drivers of the recent rise in inactivity: more students; more sick people; and more people looking after the home. Sickness and students each now account for around a third of all inactivity.

Looking after the home accounts for around 25% of all inactivity, but it is clearly a vanishing occupation. Just 4% of the working age population now profess to do this as their main activity, down from 8% in the 1990s. So, the recent upturn in looking after the home still looks more like a blip in a long-term downward trend, rather than an economy-wide reversal of activity.

The share of early retirements in the working age population has not risen much at all in recent years. Indeed, it is lower now than it was in the 1990s. The ‘Great Resignation’ is not manifesting itself as early retirement. A rise in inactivity caused by more people in education should be regarded as a good thing. More people staying on to get more skills should raise the productive potential of the economy and help the labour market prospects of those who do stay on.

But Figure 4 only measures the proportion of students who are not in work or seeking work, and not the total number of students (international guidelines record any full-time student with a part-time job as being employed and not a student). So, some of this rise after 2019 is because a larger share of students are not in work or seeking work.

Figure 4: Reasons for inactivity, 1993-2022

Source: ONS, authors’ calculations

One of the government’s responses to rising inactivity was a £1 billion subsidy to occupational pension savings, the idea being presumably to reduce any incentives to retire early to avoid paying tax on an accrued pension pot above £1 million.

But Figure 4 shows clearly that early retirement does not underlie recent rises in economic inactivity. Some deadweight loss from this policy is therefore highly likely. A more targeted, marginal policy, aimed at occupations where early retirement is affecting labour supply is likely to have been a more cost-effective option.

The largest amount of cash announced in the Spring Budget was the £4 billion package for free nursery places for families with children aged from nine months to 3 years.

Figure 5 tracks inactivity rates over time among women aged 16-45 with a (youngest) child aged 1-2. Inactivity is higher among this group than those with a youngest child aged 3-4, although the gap is narrowing over time. What’s more, inactivity rates among those with young children under 3 have been falling for most of the past 20 years – unlike that among women with children aged 10-18, where inactivity has been rising for a long time. So, a policy subsidising childcare may for young children well incur substantial deadweight loss.

There is, however, a large difference in inactivity rates by educational attainment among women with a youngest child the same age (see Figure 6). Inactivity is much higher among women who left school at age 16 than among other women with children the same age. Only among those with less formal education has inactivity risen recently. This suggests that a more targeted (marginal) scheme could again be a more cost-effective use of public funds.

Figure 5: Rates of economic inactivity rates among mothers by age of youngest child, 2001-2022

Source: LFS, authors’ calculations

Figure 6: Inactivity rates among women with a child aged 1-2, by mothers’ educational attainment

Source: LFS, authors’ calculations

How do wider economic conditions affect inactivity levels?

Long-term sickness has been an issue that has dogged the UK economy for at least 30 years. It is concentrated among the less skilled and it is higher in disadvantaged areas. Targeted interventions aimed at helping the long-term sick into better health and work are to be welcomed. But most policies do not operate in a vacuum. The lack of evidence regarding either the minimum wage or immigration leading to widespread job losses is likely to be correlated with the state of the UK economy.

One way to see whether economic performance, and by extension the level of demand, influence changes in economic inactivity is to look at changes in particular inactivity in a region following changes in local area performance. If inactivity falls when the local economy improves, that suggests that demand may have some role to play.

Figure 7 compares changes in sickness rates among those aged 50-64 across 19 areas of the UK with changes in the area unemployment rate for those aged 25-49 (we use the latter group to avoid the issue that a change in inactivity for a given age group will affect the unemployment rate for that same age group as an accounting identity).

The chart highlights a positive association between changes in local area unemployment rates and changes in area sickness rates. Areas where unemployment fell most had larger falls in sickness rates among those aged 50-64 than areas where unemployment fell less. While far from being definitive, this pattern does at least suggest that local area demand may matter.

Figure 7: Changes in area sickness rates for those aged 50-64 and changes in area unemployment rates for those aged 25-49

Source: LFS, authors’ calculations

What have we learned?

Concerns about the recent rise in inactivity look to be somewhat misplaced. Policy-makers should probably avoid devoting resources to issues that may resolve themselves as the economy recovers. It would be better instead to focus them on a longer-standing challenge – namely, a stubborn realisation of low labour force participation among older, less skilled workers with illness in disadvantaged areas. Stimulating demand in these areas is likely to complement targeted interventions aimed at the long-term sick.

Where can I find out more?

Spring Budget 2023: details of the government announcements in March about pension contributions, childcare places and much more.
Economic inactivity: latest ONS data.