The likely scale of underemployment in the UK

17 May, 2018


1. The present record rates of employment are misleading because they take no account of the underemployed – those who wish to work more hours but cannot find suitable work. This paper points to evidence that the large inflow of migrant labour which has increased the number of foreign-born workers in the UK by 2 million since 2008 may have contributed to continuing high levels of underemployment which could be around 4 million people.


2. The headline employment rate for the UK is based on the Labour Force Survey (LFS) conducted by the Office for National statistics (ONS) and measures the proportion of all individuals between the ages of 16 and 64 who have done some paid work during the week in which they were surveyed. After falling during the economic downturn in 2008, the rate had recovered to previous levels by the end of 2014 and has increased since then to record levels. Of course, some paid work is not necessarily the same as the amount of work desired.

Figure 1

3. The headline measure of underutilised labour supply is the unemployment rate itself, which is defined by the International Labour Organisation (ILO) as people who are out of work but want a job, are seeking work and are available to start work. In the aftermath of the economic downturn, UK unemployment peaked at over 2.5 million (a rate of over 8%) in 2011. Since then, unemployment has fallen to pre-recession levels and employment has increased by nearly 3 million. The growth in the size of the labour market contributed to unemployment rates that were around record lows of 4.3% by mid-2017- below pre-downturn levels.

Figure 2

How many people in the UK are underemployed?

4. However, unemployment data reveal little about the nature and intensity of work undertaken and whether those who are working are in fact working as much as they would like. Nor do they reveal anything about the reasons why individuals are not working at all and whether they would like to work.

5.While unemployment is conventionally treated as the key measure of spare labour capacity in the economy, it is recognised that this does not provide the full picture. The Office for National Statistics also publishes monthly the number of part-time workers who would have liked a full-time job but could not find one; this is a crude measure of further slack in the labour market over and above the unemployed who are looking for work. As with the unemployed, their number increased sharply over the period of economic recession from 2008, peaking at nearly 1.5 million but, importantly it decreased to a much lesser degree, to around a million people, or nearly twice the pre-recession level, as shown in Figure 3 below.

Figure 3

6. Even this, however, is only a partial measure of underemployment because it identifies only those part time workers who would like a full-time job. It does not include those part-time workers who would like to increase their hours but not to the extent of working full-time, for example a person working 16 hours a week who would like to increase his or her work to 20 hours a week. Further, full-time work is defined as working more than 30 hours a week, so there will be full-time workers under this definition who would also like to work more hours, for example those working 32 hours a week who would like to increase their work to 36 hours a week.

7. The Bureau of Labour Statistics (BLS) in the United States routinely reports a wider range of measures of under-utilised labour which they categorise as discouraged workers, people marginally attached to the workforce and involuntary part-time workers:

  1. Discouraged workers are people who are not counted as unemployed because they are not looking for work and whose reason for not looking for work is that they feel they would not find a job mainly because they do not have the necessary skills or experience (or that employers would not see them as having the necessary skills or experience).
  2. People marginally attached to the workforce are not looking for work because they feel their personal circumstances will prevent them from getting a job, such as childcare problems, study, family responsibilities, or transport problems.
  3. For ‘involuntary part-time workers’, the BLS use a measure similar to the ONS in counting people who are working part-time because they cannot find a full-time job, as described in para 5 above. (The BLS cut off point between part-time and full-time working is 35 hours a week rather than 30 hours).

8. These American measures cannot be exactly reproduced for the UK because of differences in the survey questions asked, but the ONS has sought to apply as similar as possible measures to the UK workforce (Office for National Statistics 2017). The proportion of discouraged workers is barely visible while the component of marginally attached workers does not vary as much as the involuntary part-time workers or, indeed, the unemployed

Figure 4

9. However, even these additional measures do not identify those who

  • would like to work longer hours in their current job if they are already working at least 30 hours a week (up to a maximum 48 hours in a week)
  • were looking for a new job with longer hours than their present job
  • were looking for an additional job so as to obtain more hours would like to work more hours but are not in a position to start within a fortnight

In terms of numbers, rather than the percentages illustrated in para 8 above, these three further categories increase the number of those in work but underemployed to 3 million. When added to the number of those who are out of work and looking for work (the traditional unemployed) this results in over 4 million people in the UK looking for work or for more work than they have at present and available to do so at short notice.

Figure 5

10. To some extent the increase in the level of underemployment results from a workforce that is increasing in size. However, it is notable that on the basis of percentages Figure 6 demonstrates that, whereas underemployment as currently measured closely tracked unemployment both before and during the downturn from 2008 to 2011, the underemployment rate continued to rise even once the unemployment rate started to fall. Furthermore although the unemployment rate is now below the pre-downturn rate, the underemployment rate remains noticeably higher that the pre-downturn rate.

Figure 6

11. Other measures of underemployment tend to address the issue from the perspective of the economy rather than the population. For example, the Bell-Blanchflower underemployment index measures underemployment on a net rather than gross basis. Thus, if one person wants to work 4 more hours a week but another person is working 4 hours overtime there is no underemployment because all of the hours required by employers are being worked.

12. Figure 7 below compares compares the Bell-Blanchflower index with the ONS version. It shows that the former has fallen quite sharply to just over 4%. However, this significantly underestimates the impact of underutilisation on individuals – an important social matter as well as an economic matter.

Figure 7

Which workers are wanting more work?

13. Figure 8 below demonstrates that the main components by number of the population of workers who want more work are full-time male workers whose numbers peak in the 25-29 years band and then diminish with age, together with part-time female workers whose numbers increase with age up to the age of 50.

Figure 8

14. Figure 9 below shows these figures as a percentage of those within each group. Approximately 5% of full-time workers would like more work with relatively small percentage point differences between gender and age group. Approximately 20% of male part-time workers and 25% of female part-time workers seek more hours, with the highest percentages in the latter group being among younger females in the 25-39 year age-bands.

Figure 9

15. The extent to which people wish to work longer hours also varies on a regional basis, but, with the exception of an outlier in Northern Ireland, these differences are not as great as those observed on the basis of age, gender and employment pattern.

Figure 10

16. Figure 11 below shows that there are also differences by occupational level, with workers in lower-skilled jobs being much more likely to want more hours even at their current rate of pay than workers in high-skilled jobs.

Figure 11

The inactive

17. In addition to these 4 million people in the UK looking for work or for more work than they have at present, one has also to consider those who are identified as inactive because they are not working and not looking for work. As an indication of what might be barriers to working, it is notable that (excluding students) the group of those who are not looking for work but who would like a paid job is comprised almost entirely of people looking after the family home or who have a long-term sickness or disability. Between them they account for over a million of the potential workforce.

Figure 12

18. Figure 13 breaks this group down by age band. It can be seen that consistently a quarter of this group who are not looking for work would like a paid job until the rate diminishes rapidly once retirement approaches.

Figure 13

Self employment

19. A further material trend in the UK labour market has been the increase in self-employment, rising from 3.3 million in 2002 to 4.7 million in 2017 (15% of all workers). Research carried out by the Department for Business, Innovation and Skills (2015) found that 1 in 6 were self-employed out of necessity because they could not find an employee job. This amounts to three-quarters of a million workers.

Figure 14

The total number seeking work or more work

20. In total then, the number of individuals who are not working but looking for work or who are working but want to work more hours, is around 4 million people.

More hours in current job (up to threshold of 48 hours) Fig 52,041,000
New job with more hours Fig 5274,000
Additional job for more hours Fig 5214,000
Not working but looking for and available for work Fig 2 (the traditional unemployed)1,470,000

Can immigration have an impact on employment?

21. As the number of people available to employers in the UK has been boosted so significantly by inflows of migrant labour the question arises as to whether this might have had an impact on the amount of work available for the existing population. There is no doubt that over the period during which underemployment on any measure has increased, the number of workers from abroad has increased very substantially. There are clear ways in which the ready availability of workers from abroad can lead to under-employment of existing and life-long residents in the UK. A large, indeed unlimited, pool of potential workers from other countries has meant that where there were barriers to employment for those in the existing population, employers have had little need to seek to overcome them.

22. These barriers are often characterised by employers as the fault of individuals. Such employers may claim that no-one will do the work on offer but the reasons are often because of barriers erected by the employers themselves. For example, where the work process is continuous, an employer might offer jobs only to those are able to work at any time of day or night. That is not because the work demands it, as the employer could, for example, split the work between different shifts, and offer at least some of the jobs working only during the day and some working only y during the night. Clearly this would expand the pool of workers available to the employer, with only modest additional administrative and managerial cost. More obviously, low pay is a barrier both to employment and to changing jobs for those already in the UK but may not be a disincentive for those from much lower-paying countries abroad.

Immigration and unemployment

23. It has been observed that growth in the migrant share in the workforce has not tended to increase unemployment at local level [Wadsworth 2015]. This is used to support arguments that migrants do not ‘take jobs’ from existing residents or as the author puts it “If rising immigration crowds out the job prospects of UK-born workers, we might expect to see joblessness rise most in areas where immigration has risen most”

24. However these observations, comparing 2012 to 2004, are only of unemployment (people without a job who are actively looking for and available for work) not of ‘joblessness’ more broadly, as that measure should include at least the discouraged and marginally attached as described above who are not looking for work. Nor does it include those working below capacity to get more hours. Further, Wadsworth’s own findings are that both migrant share and unemployment were higher in almost every area in 2012 than they were in 2004. It might well be the case that there was no statistical correlation between the increase in unemployment and the increase in migrant share, but this does not answer the question of what might have happened to unemployment or, no less importantly, to underemployment in the absence of the substantial increase in immigration amounting to nearly 2 million over the period.

25. It is very clear that job losses during the post-2008 recession were concentrated among the existing population and that there was substantial growth in the employment of migrant workers while employment levels and rates for the UK-born population remained depressed. Figure 15 below shows that a sharp increase in the unemployment rate for the UK-born population from 2008 was matched by a corresponding increase in the unemployment rate for the non EU-foreign born population, but not in the unemployment rate for the EU born population. The unemployment rate for the EU born population rose only slowly until it eventually matched the rate for the UK born population by 2012.

Figure 15

26. However, there was no material reduction in the number of EU born workers with jobs. The point is illustrated in Figure 16 below. Over the period 2008-2013 during which the unemployment rate for the EU born population rose from under 5% to nearly 8% the number of EU born migrants in work increased by over one quarter of a million. There was similarly no material decrease in the number of non-EU migrants in work. Conversely, the increase in the unemployment rate for the UK born population resulted from a fall from peak to trough of some 800,000 in the number of workers. Thus there was essentially no loss of jobs overall for the migrant population, a population which continued to grow and obtain further jobs in the UK while the number and rate of unemployment for the UK-born population remained high.

Figure 16

Were UK born workers crowded out?

27. Whether this can be described as ‘crowding out’ is a matter for debate. The Migration Advisory Committee has previously found some degree of preference for EU workers over UK workers on the grounds of experience and willingness to accept terms and conditions, including pay, that amount to barriers for UK workers [Migration Advisory Committee 2014] . In its preliminary report on the impact of EEA nationals on the UK labour market, the MAC note that “Some employers do not feel they could improve the supply of resident labour by offering higher wages, that wages are irrelevant to the ability to recruit. The MAC does not think this is credible. Individual employers would almost always be able to recruit resident workers if they paid wages sufficiently above the going rate. This applies even if there are skills shortages at national level – an individual employer should always be able to fill the job if a sufficiently high wage is offered. [Migration Advisory Committee 2018]

28. More generally, employers who can draw freely on a very large pool of labour from elsewhere in the EU have little need to respond to tight local labour market conditions either by offering higher wages to encourage existing workers in their area to change employers or to draw into the labour force those not in work. In these circumstances importing migrant labour would have no impact on wage levels or on unemployment because the very availability of this pool of labour means that employer behaviour need not be responsive to local labour market tightness.

29. Additionally, the availability of such labour means that an employer planning to create new jobs is not faced with the usual choice between carrying out the work in a location where the labour market is tight and having to raise his wage offer or carrying out the work in a location where the labour market is slacker and unutilised labour would be mobilised by his existing wage offer.

30. There are, however, signs that large increases in the number of migrant workers in the UK have led to the supply of labour increasing more than demand, with characteristics of a slack labour market being a restraining influence on wages. Nickell and Salaheen (2015) found that increasing the migrant share could actually depress hourly wages for employees in particular occupational groups - not just compositionally through migrants being paid less, but also by lowering wages for UK-born workers too. Within the semi/unskilled services sector, a 10 percentage point rise in the proportion of immigrants was associated with a 2 percent reduction in pay.

There are also signs that the increasing number of migrant workers may similarly limit employment opportunities for UK-born workers, with underemployment within occupational groups tending to have risen in line with migrant share because there are more workers than there is work available in a particular sector. This can create a double blow for individuals if hourly wages are held back and they cannot make up their wages by working more hours. Results are illustrated in figure 17 below.

Figure 17


31. Record numbers of people working and record rates of employment in the UK do not give a full picture as to whether individuals are able to get work, or if they have some work whether they are able to work as much as they would like or, indeed, if it is the work they want. Underemployment, on a variety of measures, remains considerably elevated compared with the period before the economic downturn. The evidence suggests that during the downturn employers continued to increase their use of migrant labour while a considerable number of those in the UK-born population lost their jobs and were unable to find new work. Headlines that suggest that there has since been recovery to full employment overlook the evidence that very large numbers of people in the UK, quite possibly as many as 5 million, are still unable to get as much work as they would wish to do.


Bell, D N F and Blanchflower, D G (2013). Underemployment in the UK revisited. National Institute Economic Review, Number 234.

Department for Business, Innovation and Skills (2015). “Understanding self-employment”. BIS Enterprise Analysis research report

Forbes, K (2016). A tale of two labour markets: the UK and US. External MPC member, Bank of England speech citing, Mulligan, Casey, 2015. “Fiscal Policies and the Prices of Labor: A Comparison of the UK and US” Journal of Labor Policy.

Haugen, S E (2009) Measures of Labor Underutilization from the Current Population Survey. US Bureau of Labor Statistics Working Papers.

Migration Advisory Committee (2014), ‘Migrants in Low Skilled Work

Migration Advisory Committee (2018), ‘EEA-workers in the UK labour market: Interim Update

Nickell, S. and J. Saleheen (2015) ‘The Impact of Immigration on Occupational Wages: Evidence from Britain’, Bank of England Staff Working Paper No. 574.

Clarke, S (ed. 2017) Work in Brexit Britain: reshaping the nation’s labour market, Resolution Foundation

Office for National Statistics (2017), Economic Review October 2017.

Wadsworth, J (2015) Immigration and the UK Labour Market, Centre for Economic Perormance/LSE Paper EA0019

Walling A and Clancy G (2010) ‘Underemployment in the UK Labour Market’, Economic and Labour Market Review, Volume 4, Issue 2, pages 16 to 24.

Note: except where otherwise referenced, data used is from the ONS’s Labour Force Survey and Annual Population Survey.