Labour Market Outlook Q3 2021

Prospects for unemployment after the Job Retention Scheme

As the economy has continued to reopen over the summer, the labour market has continued to surprise on the upside. After peaking at 5.2 per cent in Q4 2020, the headline unemployment rate has continued to fall, reaching 4.7 per cent in Q2 2021 – and timelier, though more volatile, single-month data has the unemployment rate at 4.4 per cent, just 0.4 per cent above pre-pandemic levels. Even real-time employee jobs data from HM Revenue & Customs (HMRC) payroll data, which has shown a greater hit to the labour market than the headline unemployment rate, suggests that employee jobs are just 0.7 per cent below the pre-crisis peak. And the outlook for the very short term appears positive: vacancies have recovered to or surpassed pre-crisis levels, even in those sectors hardest-hit by the pandemic.

The next challenge for the labour market, however, will be the end of the Job Retention Scheme in less than four weeks’ time. Official forecasts throughout the pandemic have predicted a rise in unemployment after the scheme ends – but in August, the Bank of England Monetary Policy Report suggested that unemployment hold steady over the next few months, before resuming its fall. In this Outlook’s spotlight, we look at the prospects for unemployment in the months ahead. Our ‘Lifting the Lid’ section explores economic inactivity, the recovery in local labour markets, and the disproportionate fall in employment among people working fewer than 16 hours.


The Coronavirus Job Retention Scheme (JRS) has been a key policy success of this crisis, protecting over 11 million jobs since March 2020. It has held down unemployment, which peaked at just 5.2 per cent last autumn – and through extensions as the pandemic pressed on, the JRS has supported a gradual return to work as the economy has begun to recover, avoiding the unemployment spike risked by a premature end to the scheme.

Over the course of the crisis, official forecasts from the Office for Budget Responsibility (OBR) and Bank of England have become more and more optimistic. In July 2020, for example, the OBR’s central scenario saw unemployment peak at more than 10 per cent; by March 2021, the expected peak had fallen to 6.5 per cent, lower than in any recent recession. But the latest Bank of England forecasts are even more optimistic: according to the August Monetary Policy Report, the Covid-19 unemployment peak could already be behind us, with unemployment reaching just 4.7 per cent in the final quarter of this year.[1] This would be a remarkable, and extremely welcome, outcome for a crisis that saw the largest GDP fall in 300 years. But with an estimated 1.7 million employees still furloughed in late July – just two months before the end of the scheme – this spotlight considers whether this a plausible outcome, and whether policy makers should still be concerned about the prospects for the labour market.[2]

Not only is the outlook for the labour market uncertain, but that uncertainty comes from multiple sources. Firstly, the number of people who will be on furlough by the end of the JRS is unclear. Furlough rates have been steadily falling since January; rising employer contributions since July should, in theory, already have discouraged businesses from keeping people on furlough; and redundancy notifications submitted to the Insolvency Service show no sign of a similar rise to last summer.[3] But survey data suggests that the fall has recently halted – and furlough rates remain elevated in sectors like international travel and in local areas such as London and city centres, where changes in commuting patterns and lower high street footfall could signal longer-term changes in demand. It is certainly possible that further falls in furlough rates, coupled with a step-change move back into work when the scheme ends, will hold back post-JRS job losses – but it is nonetheless likely that at least some workers will be let go when support is withdrawn at the end of September.

Secondly, we do not know how many of those who lose their current job will become unemployed. Prior to the pandemic, more than two-fifths (42 per cent) of people who had been made redundant in the last three months were already employed in a different job,[4] with just over a third (35 per cent) becoming unemployed.[5] Post-redundancy moves into unemployment have risen slightly during the pandemic, but still less than two-in-five (39 per cent) redundancies during the crisis led directly to unemployment.[6] And the unemployment rate has also been declining since the start of the year, helped by record-high vacancies, and may continue to do so: the single-month unemployment estimate (a relatively timely, albeit volatile measure) was just 4.4 per cent in June.

With so many moving parts, there is a wide range of possible unemployment outcomes when the furlough scheme ends. In Figure 1, we present three of these potential scenarios with very different consequences for unemployment, as described in Table 1.

Table 1: The outlook for unemployment later this year could vary widely

Scenario 1 presents the most optimistic view, where the unemployment rate could fall to 4.1 per cent by October. Under the more central Scenario 2, unemployment could reach 4.9 per cent – a rise of around 150,000 people compared to its current level (and slightly higher than the Bank’s August forecast), but not surpassing its Q4 2020 peak. Finally, Scenario 3 is the most pessimistic: in this case, unemployment could reach 5.5 per cent, above the Q4 2020 peak (though still a full percentage point below the 6.5 per cent predicted by the OBR in March). These scenarios represent just three possible outcomes. But they make the point that while the outlook for unemployment looks far more benign than most hoped for earlier in the pandemic, a rise in unemployment is a very real prospect – and it is not inconceivable that unemployment has yet to peak.

Figure 1: Under some reasonable scenarios, unemployment could rise later this year

When considering the likely scale of moves from furlough into unemployment, it is also key to consider the profile of people currently on furlough. Age continues to be a significant factor: while under-25s have been the age group most likely to be on furlough for most of the crisis, over-65s have recently become the group with the highest furlough rates – and most likely to have been furloughed for long periods of time – putting them at higher risk of job losses when the scheme ends. Older workers also suffer more severe consequences from losing their jobs than other age groups: over-55s are less likely than younger workers to return to work within six months of becoming unemployed, and tend to take a substantial pay cut when they do return. And as Figure 2 shows, the likelihood of leaving the labour force entirely after a redundancy increases with age. This means that while unemployment might not rise among older workers, employment is likely to fall nonetheless.

Figure 2: Older workers are the least likely to move into unemployment after being made redundant

Perhaps surprisingly, accounting for the age patterns shown in Figure 2 does not have significant implications for potential moves out of work after the JRS ends.[7] Although the large numbers of older workers on furlough suggests there will be more moves into inactivity (and fewer into unemployment), this is counterbalanced by younger workers – who are more likely to move into unemployment – still having relatively high furlough rates. But while we have not yet seen evidence of the crisis pushing large numbers of older workers into early retirement,[8] policy makers should remain cautious. Workers aged 45+ are more likely than other age groups to have been furloughed for long periods of time, possibly leading to skills depreciation that could put them at a disadvantage on the jobs market, and rising asset prices may make an early departure from the labour force more attractive to those approaching retirement age and struggling to find work.

If significant numbers lose their jobs at the end of this month, will the labour market be able to absorb them all? Vacancies have recovered to pre-pandemic levels, and despite reports of labour shortages in some sectors, jobs are still being filled faster than before the crisis. As Figure 3 shows, total job starts (including both job-to-job moves and inflows from unemployment and inactivity into employment) reached a record high in Q2 2021. If furlough off-flows were to be added to these levels of job starts in October, this would imply job starts surpassing 2 million for the first time since records began in the early 2000s; if other job starts returned to 2019 levels, overall starts (including flows from furlough) would still be higher than the pre-Covid-19 peak. With matching frictions already creating a hiring bottleneck, it may be challenging for the labour market to immediately absorb such high levels of job moves if they materialise.

Figure 3: Job starts are already at their highest levels on record

Unemployed workers in some sectors may also find it more difficult to find a new job than in others. Figure 4 shows that in June, there were still more unemployed workers per vacancy overall than there were in 2019, but the picture varies substantially by sector: in hospitality, for example, a boom in vacancies had pushed the number of unemployed people for each vacancy almost back to pre-pandemic levels by Q2 2021, whereas in transport and storage, there were more than two unemployed people for each vacancy, well above the rate before the crisis.

Figure 4: Despite record vacancies, the labour market remains looser than before the pandemic

Any further rises in unemployment are also likely to be sectorally uneven. The dashed bars in Figure 4 add the number of furloughed workers in June to the number of unemployed people in each sector as a rough measure of the amount of slack in each industry by June. While the levels of furloughed workers will decline before the end of September, the sectoral distribution is likely to continue to vary widely, leaving some sectors with far more unemployed people per vacancy than others. (While vacancies may also increase, experimental ONS estimates of online job adverts suggest that vacancies may have stopped rising, remaining broadly flat over July and August.) This could lead to a higher unemployment rate at least in the short term, while the labour market adjusts. And in some sectors like retail, job losses could reflect a longer-term structural shift, requiring workers to move to a different sector – and since workers in the sectors hardest-hit by the crisis have been more likely to seek work in other hard-hit sectors, this could increase the burden on sectors already experiencing the biggest rises in post-JRS jobseekers.

Finally, the focus of this spotlight has been the impending end of the JRS – but there are other changes in the labour market that will increase labour supply. The end of the 2020/21 academic year brings a new cohort of education leavers into the labour market, including some who have stayed in education for an extra year to ride out the storm of the pandemic. Labour force participation could rise, reversing an increase in inactivity (explored further in this Outlook’s ‘Lifting the Lid’ section), for example as discouraged workers move back into the labour force. And lastly, the Self-Employment Income Support Scheme is also coming to an end; given the disproportionate hit to self-employed workers during the crisis, some of the self-employed whose demand has not recovered may also enter unemployment as support ends.

Overall, then, it certainly seems plausible that we have passed the peak of unemployment, reflecting the extraordinary success of the JRS in keeping people in their jobs over the past 18 months as well as supporting incomes, alongside the strong recovery in the economy. But while the challenge is set to be far smaller than many expected, policy makers should not be complacent: unemployment is likely to rise, at least temporarily, when the JRS comes to an end. Employment support schemes like Kickstart and Restart will remain important to support vulnerable groups, as well as wider job search support and retraining where necessary. Macroeconomic policy should remain supportive, stimulating demand to ensure the recovery continues to take hold. And for those who do lose their jobs, adequate income support through the benefit system will be crucial to maintaining living standards later this year.


[1] The Bank’s relative optimism compared to previous forecasts is the result of both a positive outlook for GDP in their macroeconomic modelling and promising labour market indicators such as a strong recovery in job vacancies. The minutes of the Monetary Policy Committee meeting do, however, acknowledge significant uncertainty around the end of the JRS.

[2] In this spotlight, we do not explicitly consider a scenario where restrictions or lockdowns are re-imposed. This would have significant implications for the labour market – and the Government should extend or reintroduce support if this were the case.

[3] It is worth noting that only businesses planning 20 or more redundancies need to notify the Insolvency Service, while the profile of furloughed workers is increasingly skewed towards smaller businesses that would tend to make fewer redundancies each.

[4] We note that some people who are back in work within three months may spend a short time out of work in the interim, but due to available data we use these rates as a proxy. Given that this data covers those who have been made redundant at any point within the last three months, we assume that the average length of time since redundancy among this group is 1.5 months.

[5] The remaining 22 per cent moved into economic inactivity (proportions do not sum to 100 per cent due to rounding). Analysis covers calendar years 2017-2019. Source: RF analysis of ONS, Labour Force Survey.

[6] A further 36 per cent were employed in another job and 25 per cent became economically inactive. Analysis covers Q2 2020-Q2 2021. Source: RF analysis of ONS, Labour Force Survey.

[7] In fact, the economic activity of people who become redundant is almost identical after reweighting for the age of workers currently furloughed: the shares in employment and unemployment fall by 0.5 percentage point, and the share in economic inactivity rises by 1.0 percentage point.

[8] The ‘Lifting the Lid’ section explores changes in inactivity since the start of the crisis in more detail. Looking ahead to retirement intentions, the Institute for Fiscal Studies has found that while 5 per cent of older workers plan to retire earlier than before the pandemic, 8 per cent plan to retire later. Moreover, Resolution Foundation research based on a survey fielded in January 2021 found that older workers who had been furloughed were less likely to plan to leave the labour force after the pandemic than their counterparts who had not been furloughed.