Low paid work in Britain: baseline surveys from the Earnings Top-Up pilot evaluation
A hard copy of this report summary can be obtained by contacting Paul Noakes [E-Mail: Paul.Noakes@dwp.gsi.gov.uk] or by writing to him at the 'Social Research Division, Department for Work and Pensions, 4th Floor, Adelphi, 1-11 John Adam Street, London WC2N 6HT'.
Research Report No. 95
By Alan Marsh, Claire Callender, Louise Finlayson, Reuben Ford, Anne Green and Michael White
In October 1996, the Department of Social Security (DSS) launched a pilot of Earnings Top-up (ETU), a new in-work benefit for people (singles and couples) without dependent children. Two different rates of benefit are being piloted: lower rate Scheme A and higher rate Scheme B. As part of the evaluation of the pilot, a series of surveys were commissioned from Policy Studies Institute, and local labour market studies from the Institute for Employment Research. This report is a baseline study that provides data for comparison with subsequent research over the three years of the ETU pilot. It presents the results of the first surveys of low-paid workers, medium-term unemployed people and employers conducted in 1996; and compares the labour markets of the evaluation areas before the introduction of ETU.
- As people with children were screened out of interviews, young people (under 25) and older people (45 plus) were the main age groups represented in the surveys. These age groups are likely to form the main potential customer groups for ETU.
- Educational levels were low among low-paid workers and unemployed people. In both groups, over half had no qualifications compared with less than a third in the working population as a whole. Among medium-term unemployed people (between 26 and 65 weeks), persistent illness was common, reported by a third of unemployed men, and two-fifths of unemployed women. Unemployed people had a weak competitive position in the jobs market having on average spent only one-third of the previous five years in employment and the same proportion unemployed. The remaining time was split into small proportions of education, sickness and non-claimant unemployment.
- One-fifth of potential ETU customers lived in rented households, the majority of whom received only partial Housing Benefit (HB), or none. Nearly three quarters of single people lived with their parents and many had only small housing costs. The majority of couples lived independently.
- Between half and two-thirds of respondents were potentially eligible for ETU under Scheme A or Scheme B respectively. The majority of unemployed people looked for, or took, wages that would not push them beyond eligibility for ETU; and the majority of low-paid workers had household incomes that would qualify them for the new benefit.
- A small group of unemployed people said that they would work for wages that would have made them worse off in work than on benefits. However a large majority would have been better off in work at their preferred wages even without ETU (over four fifths). Among workers, most liked their jobs and only a quarter said they were actively looking out for a new one.
- The employers' survey suggested scope for ETU to influence recruitment and wage-setting. Three-fifths of employers said that the pay for which individuals were willing to work was the most important influence on wage-setting. Half were also influenced by the pay offered by other local employers.
- A high level of awareness of social security benefits was found among employers. Almost all employers had heard about Family Credit when prompted. However, only about a third had any experience of dealing with in-work benefits, or advising employees about them.
Overview
Earnings Top-up (ETU), a new in-work benefit, was introduced by the DSS as a three year pilot in October 1996. ETU is available to people (couples and singles) working 16 or more hours per week, without dependent children. The aim of ETU is to improve incentives for people to take up work or stay in work of 16 hours or more.
Two different rates of benefit are being piloted – lower rate Scheme A and higher rate Scheme B. Each scheme was introduced in four areas reflecting four types of labour market: major urban areas, large towns, rural and seaside areas. Four other matching areas were selected as control groups. Research is being carried out in all twelve areas, which represent about one tenth of the British labour market. In particular, the effects of ETU upon low-paid workers, the unemployed, employers, and the labour market are of interest.
The research reported here is a baseline study that provides data for comparison with subsequent research over three years of the pilot. It presents the results of the first surveys of low-paid workers, unemployed people and employers undertaken by Policy Studies Institute (PSI) in 1996. Comparisons of the labour markets in the evaluation areas prepared by the Institute of Employment Research (IER) are also reported.
The workers-in-work survey
The sample
The sample was selected using National Insurance records from the tax year 1994/95 to identify low earning employed and self-employed workers. 2,400 interviews were achieved in summer 1996, representing a response rate of 79 per cent. Three out of ten of the 'employed' sample, in work in 1994/5, were no longer in work when interviewed in 1996.
Gender
The majority of the low-paid workers were women (56 per cent). Women had persisted longer in their low-earning jobs since 1995: 27 per cent of the men but only 11 per cent of the women in the sample were unemployed at interview in 1996. Among the single-earner couples, women outnumbered men as the breadwinners by two to one. Women were more likely to have partners: nearly half were married or in couples compared with four in ten men.
Age
The sample comprised working-age people, excluding families with dependent children. This resulted in a bi-modal age distribution, with very few aged between 35 and 44 years, while 44 per cent of the workers-in-work were aged 45-plus.
Education
Educational levels were low. More than half the workers-in-work had no educational qualifications and just ten per cent had qualifications at A-level or above. Women had less education than men: 57 per cent had no qualifications compared with 48 per cent of men. Overall, younger respondents were more likely to hold qualifications, with striking differences between younger and older women.
Among the older workers, few had any qualifications at all. But even among the young, educational levels were still low by comparison with higher earners: half the under 25s had no more than a few GCSE or CSE equivalents and fewer than one in seven had an A-level or better. There was however a group of single, well-educated under 25s for whom current low earnings may have been a temporary phenomenon.
Housing
A narrow majority of low-paid workers had no housing costs, or paid only nominal amounts. As many as three-quarters of the under 25s lived this way, paying small contributions, and most of the rest lived rent free. Couples were most likely to be owner-occupiers.
The large numbers of younger workers among single people (four out of ten were under 25), their very low wages and their lack of housing costs, all suggest a solid customer base for ETU among the young.
Occupational group
Though low-paid, not all the workers had the kind of jobs that are usually thought of as low paid. Six out of ten had jobs that lay at the bottom of the income distribution, dividing evenly into four groups: personal services, sales, plant operatives and ‘other unskilled’ jobs. Women were concentrated in service and sales sectors, the men in operative and other unskilled jobs. The remaining four out of ten, however, divided evenly into three categories: craft workers, clerical jobs and even twelve per cent who were 'professional and managerial' workers - though many of the latter were likely to be self-employed and said to be earning little or nothing.
Job satisfaction
Broadly, low-paid workers seemed reconciled to the sort of work they did even without the possible advantage of ETU. Most liked their jobs and only a quarter said they were actively looking out for a new one. These job-seekers, however, said they would continue in the same line of work and sought wages only a little higher than their present wage - typically they sought less than £120 a week.
This suggests that the initial deadweight inflow to the benefit is likely to be composed of a stable population of low-paid workers who have little expectation of a better-paid job. ETU jobs may lack a career structure - 70 per cent of the sample saw their jobs, one way or another, as 'dead-ends'. Added to that, seven out of ten said that, if they were seeking a new job, they would certainly be glad of some kind of 'wage top-up' if one were made available, sooner than hold out for higher wages. This was particularly true of those in work and may have interesting outcomes for eventual take-up of ETU.
Earnings
Earnings were low by sample design, averaging £100 a week for a 34 hour week. They were rarely supplemented by any other income: an average of £3.00 an hour. Even taking out the relatively few part-time workers, average earnings rose only little. The best-paid group were women supporting husbands and even they managed only £3.40 an hour.
The survey of unemployed people
The sample
The sample of (claimant) unemployed individuals was drawn from the Departmental Central Index (DCI) at the DSS. This is a computer database of all DSS customers containing details of recent benefit status. The sample was defined as medium-term unemployed, with 26-65 weeks of unemployment at the point of sampling. 1,991 interviews were achieved, representing a response rate of 81 per cent among eligible sample members. By the time of the interview, four months after sampling, just 64 per cent of the 'unemployed' sample remained unemployed and seeking work.
Gender
The unemployed and workers-in-work samples differed substantially in their composition by gender. Fewer than one in three of the unemployed sample were women compared with nearly three in five of the workers-in-work sample.
Age
As expected, only 30 per cent of the sample were aged 25-44. However the sample was not markedly skewed towards younger people: just under one-third of the unemployed sample were under 25. The shift was towards older age bands: 37 per cent of the sample was aged 45 or over. This age distribution may partly account for higher rates of ill health.
Human capital(1)
Even compared with the lowest-paid workers, the unemployed had a poor accumulation of 'human capital'. They were more likely to lack any qualifications (56 per cent) and, among the under 25s, to recall having persistently truanted from school (46 per cent). They had typically spent only one-third of the past five years in employment and the same proportion unemployed, indicating a very weak competitive position in the jobs market. Both ETU samples suggest a high level of educational disadvantage among low-paid workers and the unemployed.
Health
Health or ill health can also be considered an aspect of human capital since it may limit paid work in various ways, or be a source of discrimination. Persistent illness was common. A third of unemployed men and four out of ten unemployed women reported persistent illness either currently or in the recent past. This compared with a quarter and a fifth among employed men and women.
Households
The great majority of unemployed men were single. Three out of ten of them had partners compared with four out of ten employed men. Among women, even fewer of the unemployed (22 per cent) but more of the employed (48 per cent) had partners. More importantly, the majority of these partners were themselves unemployed. Whereas 61 per cent of the unemployed sample had no other employed person in their household, the proportion was only 42 per cent for the employed sample.
Housing
More than two-thirds of single unemployed people lived with their parents. So many lived at home, especially the men, and so many of the rest were older, that the housing tenure profile of low-income people without children is very different from those with children. Four out of ten of the unemployed and a quarter of the workers-in-work were social tenants, far fewer than low-income families with children.
Job search activities and networks
There was a relatively low level of job search activity among the unemployed sample. 26 per cent had not recently been actively seeking paid employment (including seven per cent who were 'unemployed' but not claiming). 18 per cent were in a job or waiting to take up a job by the time of interview, an average of four months after sampling. 57 per cent were still claiming benefit and actively seeking a job. The inactive proportion was 22 per cent in the case of men but 36 per cent for women. Even after excluding those on disability benefits, there were quite high proportions not seeking work (13 per cent of men and 23 per cent of women).
Of the economically active, there were quite high proportions making only a few job applications or none at all. There were strong indications that the social networks of the unemployed sample contained fewer employed people. If, as might reasonably be assumed, employed friends are more useful for providing information about job vacancies, then the unemployed sample was evidently at a disadvantage since so many of the people they knew were themselves out of work.
Economic activity or inactivity
The extent of economic activity was low compared with other studies of unemployed people. Economic activity was depressed by high rates of ill-health among the unemployed sample. Women were also less active jobseekers, especially those with access to some non-waged income other than partner’s earnings. Men’s activity was increased by a recent history of temporary work and raised further among those reporting frequent contact with friends.
Entry into employment
A minority of the unemployed sample (14 per cent) had found paid employment by the time of their interview (two per cent more were self-employed). Among those free from sickness, the strongest influence on early work-entry was recent employment experience, suggesting that employers rely on work histories rather than ‘human capital’ in judging suitability for employment. Overall, ‘human capital’ factors were weak. Only having an A-level positively influenced job entry chances. Other qualifications, both academic and vocational, were not associated with higher job entry rates. This may have been the result of people who were good at capitalising on their skills no longer being unemployed and claiming benefit.
Women who were economically active tended to obtain work more easily than men, especially if they had a non-working partner. This is the opposite effect to that usually found in nationally based surveys where partners of unemployed people usually have lower participation rates than partners of employed people.
Job search intensity
Moderate job application rates (one to five a week) were associated with younger unemployed people and those who had less access to non-wage income or lower wage expectations. Higher rates of job applications (more than five a week) were, in turn, influenced by better education, experience of training schemes and more continuous employment, and by age.
Wage expectations
Wages sought in new employment averaged £122 a week overall, higher in fact than the average take-home pay earned by the parallel sample of low-paid workers. Controlling for other factors, wage expectations were lowest in ETU Scheme B areas and highest in Scheme A areas (£116 vs £127). Higher wage expectations were associated with better education, people with driving licences, and people in the middle years - who of course are relatively few among a sample of people without children. Debt was also important: an overhang of debt forced people to look for higher wages in work than they may be able to find.
Financial incentives among out-of-work households
The analysis estimated who among the continuing unemployed would be better off in work at their preferred wages; who would then be entitled to apply for ETU and who would not; and what difference ETU would make to the ‘better-off’ estimates for potential applicants currently out-of-work.
Wages and final incomes in work
Even without ETU, most unemployed people (87 per cent) would have been better off in work (all other entitlements considered) at their preferred wages. They would be an average of £44 a week better off. 13 per cent would have made a loss from work. The unemployed were prepared to work full-time hours, on average, for the equivalent of less than half average hourly earnings for their area. They sought a final income gain of around £40 per week if they did not have partners, and about £60 if they did.
The likely impact of ETU
Under Scheme A, half those currently out of work would have qualified for ETU in work at their expected wages. Under Scheme B, the proportion was nearer two-thirds. The addition of ETU to preferred earnings would have left almost everyone considerably better off in work than out of work. This offered a wide margin for further reductions in wage expectations. People's preferred total incomes in work could be maintained at an average wage of £80 a week with ETU. Only £20 a week would still see the average jobseeker better off in work if ETU was added to their in-work income. The relative absence of housing costs meant that ETU would not simply replace in-work entitlements to Housing Benefit (HB) or Council Tax Benefit (CTB). This is not to say that, in every case, housing costs would remain absent if they had got a job.
Sources of income when in work
Prior to ETU, four additional sources of income appeared in people’s calculations of their total incomes in work: partner’s earnings, HB, CTB, and pensions. Expected in-work expenses were high with a median value of £19 a week. Only a quarter of renters expected any Housing Benefit in work. On average, 87 per cent of their total expected income in work was sought from wages alone.
What difference would expected gains from work make?
Gains from work were expected to be £67 a week, though downwardly adjusted to £49 a week after in-work expenses such as travel. Most people judged accurately the extent to which their expected total incomes in work would leave them better off compared with their current out-of-work incomes.
Even low wage expectations left those who attained them relatively much better off than their out-of-work incomes and commonly in scope of ETU. This creates an interesting tension: will they lower their expectations still further when the benefit is actually available? We can only point out now that these wage expectations were designed to reflect a 'rock bottom' position. These levels also reflect some of the latest offers they are likely to encounter from employers. The 1997 levels will show us what they chose in the event.
The availability of ETU
Two thirds of jobseekers said they would accept a ‘top-up’ to their wages if one were available, while the remainder hoped to be free of subsidy-level wages. Among the potential jobseekers who gave a minimum acceptance wage, just under two thirds (63 per cent) said they would accept a wage lower than their earlier income, plus the new top-up, leaving 37 per cent holding out for their minimum wage.
The survey of employers
The sample
The sample of employers was drawn from the British Telecom 'Connections in Business' database. The study consisted of a telephone survey of 2,400 employers, 200 in each of the twelve ETU pilot areas, representing a 78 per cent response rate from those approached. The survey covered all industrial sectors and large as well as small establishments. The re-weighting of the data allowed an analysis of the workforce (employment) as well as the employing unit (establishment) to be undertaken.
Characteristics of the employers
The majority of establishments were in the private sector. Public sector employers tended to be much larger on average. Most employers serviced customers locally and only three in ten were part of a national operation. Half formed part of a larger organisation and a tenth of those interviewed were themselves the Head Office branch.
Job types
The survey concentrated on the employment of three typically low-paid job groups: semi/unskilled; skilled/craft; and clerical/sales. Half the employees in these key categories were women. Women were disproportionately represented in certain occupations. For example, they formed the majority in clerical/sales jobs while men formed the majority of skilled/craft employees. Most employees worked familiar 37-38 hour weeks on average. But about one in ten worked less than the ETU threshold of 16 hours per week: these were predominantly clerical/sales or semi/unskilled workers; those in establishments with less than five employees; and those in hotels/catering, education or 'other service' industries.
Wage-setting and wage levels
The greatest influence on wage-setting mentioned, and one of considerable importance for ETU, was the pay for which individuals were willing to work. Three in five establishments mentioned this. Half were also influenced by the pay offers of other local employers. Overall, about a quarter of employers had to follow Head Office direction in wage-setting, which will tend to blunt any local wage-effect of ETU in their cases.
The average wage offers quoted were much higher than those earned by workers in the employed sample or sought by the unemployed, probably because the sample was not restricted to low-paying industries. Wage offers ranged, on average, from £4.20 an hour for the semi/unskilled to £5.07 for clerical/sales employees and £6.74 for skilled/craft workers. But there were quite wide variations within job groups: half the semi/unskilled employees, three out of ten clerical/sales workers and 14 per cent of the skilled/craft workers were paid less than £4.00 an hour.
Recruitment and turnover
Establishments generally had greater autonomy in their recruitment policies than in wage-setting, though still 43 per cent of local branches had to follow procedures laid down by their Head Office. About a quarter of employers reported difficulties in recruitment over the past twelve months, rising to four out of ten among larger employers. Skilled/craft workers were the hardest to recruit but semi/unskilled workers only a little less so.
Low-paying industries are traditionally associated with high job turnover. There was considerable turnover of unskilled employment, with the majority of firms both recruiting and losing at least some of these staff in the past year. An important factor associated with the turnover rates in all three job types was the proportion of employees paid less than £4 per hour: as the proportion of low-paid employees rose, so did the rate of job termination.
Social Security benefits
There was a high level of awareness of in-work benefits among the employers surveyed: about three-quarters of employers knew about them in principle and, when prompted, almost all employers had heard about Family Credit, for example. On the other hand, only about a third of employers had any actual experience of dealing with in-work benefits or advising their employees about them.
Employers' own attitudes towards benefits were particularly interesting. Half had no doubt that benefits for unemployed people created difficulties for recruitment. In contrast, over a third of those who had some experience of in-work benefits, thought that they had made it easier to recruit to low-paid grades. But over two out of five thought that in-work benefits made employees unwilling to work overtime and made some want to decrease their hours of work. A quarter of employers with experience of in-work benefits thought that such benefits had helped to keep their wages down. Significantly, nearly half of those with no previous experience of in-work benefits thought they could have this effect. Such views were more common among smaller establishments who were more likely to have local autonomy over wage-setting.
Area comparisons: The ETU pilot areas
During the initial phase of the ETU pilot scheme it is likely that most of the recipients would have had their jobs some while. The workers-in-work sample was therefore compared for differences that may influence the take-up of benefit in different areas or that might distort some of the differences in behaviour seen in the pilot areas compared with the control areas.
In terms of the size of the key ETU related categories (working 16+ hours, over or under 25, singles and couples) the pilot and control areas did not differ at all. Nor were there any differences in key background variables such as housing tenure and educational qualifications, present wages or expected/acceptance wages. The only really significant differences to emerge seemed fairly idiosyncratic and not part of any systematic pattern. For example, control areas together had more people claiming disability benefits and more people working in health-related occupations too. Scheme B areas had more people in catering and more trade union members.
The relative similarity between the Scheme A, Scheme B and control areas, however, concealed wide differences between the twelve individual areas. Some differences were to be expected: they were designed into the study in choosing contrasting urban and rural areas, for example. But others remained and some of them were very puzzling.
Southend was not a typical seaside town, as many people commuted to London or to the large Ford factory at Dagenham. Unemployment in Perth also lagged well behind the average and, together with North Wales, had net gains in employment levels contrasting with net losses in the major urban areas like Newcastle. The urban areas and towns still suffered the greater unemployment overall. Newcastle had much lower unemployment levels than the other urban areas. The greatest relative loss of employment occurred in Southend. Here though, it was non-employment, especially among older workers, that was the problem. Among those in work in Southend, earnings were higher than elsewhere, like other ‘Home Counties’ places including the southern control area of Southampton. The lowest earnings were in the rural areas.
Both the field survey data and the IER analyses pointed to the same conclusion. No simple or obvious patterns were found across the four area types, the three ETU groups (A, B or control) or the twelve geographical locations. Some differences did exist between major urban areas and large towns on the one hand, and seaside areas and rural areas on the other, but these were designed into the pilot in the first place.
Conclusions
The baseline surveys were conducted before the introduction of ETU among potential customers of ETU and employers. In summary:
- The samples of low-paid workers and unemployed were young single people, many of them women, or older couples. They had low levels of education and uncertain work histories but were often reconciled to their low-paid jobs, if they had one. Those seeking work seemed prepared to settle for wage-rates below half those typical for their areas.
- Many unemployed people were unwell. Persistent poor health was found to be a barrier to paid work. As medium-term unemployed were sampled, some of the health outcomes were different to those found in surveys of shorter term unemployed people.
- It might be expected to find the great majority of potential ETU customers living as tenants, although relatively few lived in rented households. Young and single people often lived with their parents and many had only small housing costs. Couples were more likely to live independently.
- A large majority of unemployed people would have been better off in work at their preferred wages even without ETU. Among workers, most liked their jobs and only a quarter said they were actively looking out for a new one.
- Large numbers of respondents were potentially eligible for ETU. The majority of unemployed people looked for, or took, wages that would not push them out of eligibility for ETU; and the majority of low-paid workers had household incomes that would qualify them for the new benefit.
The tentative conclusions drawn from the findings suggest that:
- Housing costs would not create a barrier to the work incentives created by ETU: awards of ETU would not simply replace otherwise continuing in-work entitlement to HB.
- The introduction of ETU could increase the scope for people to lower their wage expectations, possibly by significant amounts.
- Deadweight costs would be seen after the introduction of ETU as medium-term unemployed people and low-paid workers were already in a position of being better off in work without an additional top-up.
- Evidence from the employers survey did not oppose the idea that the availability of ETU would influence recruitment and wage-setting.
Views expressed in this report are not necessarily those of the Department or any other Government Department.
Relevant publications
Bottomley, D., McKay S., and Walker, R., (1997), “Unemployment and Jobseeking” (Research Report No. 62), TSO
Bryson, A., and Marsh, A., (1996), “Leaving Family Credit”, (Research Report No. 48), HMSO
Callender, C., Court, G., Thompson, M., and Patch, A., (1995), “Employers and Family Credit”, (Research Report No. 32), HMSO
Elam, G. and Thomas, A. (1997), “Stepping Stones to Employment, ”(DSS Research Report No. 71), TSO
Elam, G., Diffley, M., and Shaw, A., (1998), “Getting the Message Across”, (Research Report No. 85), CDS
Finlayson, L., and Marsh, A., (1998) “Lone Parents on the Margins of Work”, (Research Report No. 80), CDS
Ford, R., Marsh A., and Finlayson, L., (1998), “What Happens to lone parents”, (Research Report No. 77), CDS
Kellard, K. and Stafford, B (1997), “Delivering Benefits to Unemployed People, ”(DSS Research Report No. 69), TSO
McKay, S., Walker, R. and Youngs, R. (1997), “Unemployment and Jobseeking before Jobseeker's Allowance, ”(DSS Research Report No. 73), TSO
Shaw, A., Walker, R., Ashworth, K., Jenkins, S., and Middleton, S., (1996), “Moving off Income Support: Barriers and Bridges”, (Research Report No. 53), HMSO
Snape, D. (1998), “Recruiting Long Term Unemployed People,” (DSS Research Report No. 76), TSO
Stafford, B., Heaver, C., Croden, N., Abel Smith, A., Maguire, S., and Vincent, J“.”, (1998),“ Moving into Work: Bridging Housing Costs”, (Research Report No. 79), CDS?
Trickey, H., Kellard, K., Walker, R., Ashworth, K., and Walker A., (1998) “Unemployment and Jobseeking: Two Years On” (DSS Research Report No. 87), CDS
(1) Human capital is a term used by economists to describe the skills, capacity and abilities possessed by an individual which permit her or him to earn income.