Earnings top-up evaluation: effects on low paid workers
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 Resaerch Division, Department for Work and Pensions Security, 4th Floor, Adelphi, 1-11 John Adam Street, London WC2N 6HT'.
Research Report No. 134
Part One: Survey of Low Paid Workers
By Alan Alan Marsh, Augusta Stephenson and Richard Dorsett
Part Two Econometric Analysis
By Peter Elias
Earnings top-up (ETU) was an in-work benefit available to low paid workers without children. ETU was piloted from October 1996 to October 1999 in eight areas across Britain. The evaluation was conducted by the Policy Studies Institute (PSI), the Centre for Research in Social Policy (CRSP) at Loughborough University and the Institute for Employment Research (IER) at the University of Warwick.
This report is one of seven final reports published from the ETU evaluation and is in two parts. The first section, by PSI, presents findings of the surveys conducted with low paid workers and ETU recipients throughout the pilot. It explores the characteristics of these workers and the effect that ETU had on their lives and examines the reasons for non take-up of ETU among eligible workers. The second part, by IER, assesses employment effects using employment history analysis.
The main findings include:
- ETU reached the people it was intended to reach, met a need and went some way to reducing hardship. However take-up was low.
- A large proportion of non-take-up among eligible workers was accounted for by:
- the absence of main-media publicity,
- sparse distribution of eligibility under Scheme A (lower rate ETU), and
- social and economic isolation of older couples and their lack of claiming-skills.
- The much higher take-up rate among those already attuned to the benefit system – younger low-paid single workers and tenants living in higher rate ETU areas – indicated that a new version of the benefit could be more successful if it:
- allowed for the introduction of the National Minimum Wage with adjusted thresholds,
- adopted a shallower taper like Working Families Tax Credit
- was backed by a national advertising campaign
- There was little evidence from the surveys of ETU recipients and workers to oppose the view that most of the expenditure on ETU in the pilot areas went to people who:
- would anyway have done the jobs they did,
- or who took the jobs they would have taken,
- worked the hours they would have worked, and
- worked for the wages they would have otherwise accepted.
- ETU secured in work a group who had had poor work histories throughout the 1990s. These ETU workers showed signs of maintaining this improvement, with rates of labour market participation on a par with existing workers.
Part one – Surveys with low paid workers and ETU recipients
The benefit and the evaluation programme
Earnings top-up - earnings top-up (ETU) was an in-work benefit for people without dependent children, introduced in October 1996 in eight areas for a three-year pilot period. Its main objectives were to improve the incentive for unemployed people to take low-paid work, and to help low-paid workers remain in work by raising their incomes relative to out-of-work support. ETU was paid at a fixed rate for 26 weeks to people working 16 or more hours per week, with a premium for 30 hours or more. The maximum payment was reduced by 70p for each pound of income above the earnings thresholds. It was payable at two different rates, Scheme A and Scheme B, each in four areas, and rates also differed between couples, single people aged 18-24, and single people aged 25 or over.
The ETU evaluation - The eight test areas were matched with four corresponding control areas. All the areas were chosen for their high level of unemployment, high number of job vacancies and high proportion of low-paid vacancies. They represented four types of labour markets: large towns and major urban, seaside, and rural areas. As part of the programme of evaluation research, several surveys of low-paid workers and ETU recipients were carried out.
The progress of ETU - The projected caseload was reached after 14 months and the caseload stabilised in mid-1998 at around 24,000. In March 1999, 46 per cent of recipients were single over 25s, 40 per cent single under 25s, and 14 per cent couples. From April 1999, when the National Minimum Wage (NMW) was introduced, the caseload declined, mostly at the expense of young single recipients, consistent with NMW increasing their wages above their limited ETU entitlement. Women were a slight majority and the proportion of self-employed recipients rose from seven per cent to 12 per cent during the evaluation period. At August 1999, prior to claiming ETU, 18 per cent of recipients had been claiming Jobseekers Allowance or Income Support, a fall of seven per cent since April 1998.
Despite measures intended to balance the caseloads in the two schemes, Scheme B attracted more customers than Scheme A, largely because of the very high caseload in Sunderland (major urban area B) - A recipients only outnumbered those in B in the rural areas. In particular, B attracted far more young single people than A. Furthermore, northern industrial regions contained many more recipients than southern seaside areas.
Hours and wages were relatively stable for most of the pilot, but following introduction of NMW, average gross weekly earnings of employees claiming ETU increased slightly, while average hours worked declined. This indicates that some employees working longer hours had become ineligible for ETU when their hourly rate rose.
Claims and recipients
Key characteristics of ETU recipients – A small majority of the sample were women and ETU receipt was concentrated among young and single people. Half had no formal housing costs, 42 per cent living in their parents’ home. ETU counted as assessable income for Housing Benefit (HB) claims, but there seemed little conflict. Only one-third were tenants and among those who remained eligible recipients of ETU at interview, only 43 per cent would receive more HB if they ceased to receive ETU.
Just 17 per cent had a health problem, but levels of education and training were low. The highest academic qualification of half the respondents was at GCSE level and almost all the rest had no academic qualifications. Almost one-third overall had no qualifications of any kind. Eleven per cent of respondents were unemployed at interview, but 90 per cent of those who remained in work of 16 or more hours had a permanent job, though just five per cent were in a managerial, professional or technical occupation. Forty-four per cent worked part-time, but only eight per cent had changed their hours as a result of claiming ETU.
Some of the ETU recipients had financial problems - 29 per cent were in ‘severe hardship’. Those in non-working households were worse off than those who remained eligible, but those in working households which had become ineligible were better off than the rest.
Claiming ETU – One-third of respondents had become ineligible for ETU by their interview, and 13 per cent of those who remained eligible no longer received ETU. Of those no longer eligible, half were no longer in work of 16 or more hours, but the other half worked longer hours for higher pay than those who remained eligible. Delays in claiming were fairly rare and most delay was caused by uncertainty. The most common reason for applying for ETU was that the respondent had just found out about it – getting a new job came third. Word of mouth was a very important source of information, though over one-third had been told about ETU by an official. However, knowledge levels regarding the rules were low.
Did ETU make a difference? - Half of those taking their job since ETU’s introduction had been aware that they might qualify. Of these, 56 per cent would have accepted the job at the wages offered without ETU. Of the remainder, 43 per cent would have asked for longer hours. Just 15 per cent of the few who discussed it with their employer at the time felt ETU had made a difference to their getting a job, and just six per cent of workers receiving ETU would stop working if they ceased to get it. Half said they would continue their present job, though half of those would want more money or hours. However, 72 per cent of respondents felt that without ETU they would have had to cut back or could not have managed at all. Those who started their job before ETU’s introduction were just as likely to find it indispensable as those who did not.
The impact of Earnings Top-up on workers-in-work
“Britain’s low-paid workers” - Spells of unemployment were common among the low-paid workers and only two-thirds remained in work of 16 or more hours per week at interview. A quarter were ineligible for ETU because neither they (nor a partner if they had one) were working 16 or more hours. However, employment histories were similar in the pilot and control areas. Median pay rates had risen among the workers-in-work since 1996, but hours of work had fallen sharply. One in six workers appeared to be below the National Minimum Wage and net weekly wages were typically between £100 and £140 a week.
Respondents tended to be either young, single and living with their parents or older couples who owned their home outright or still had a small mortgage. Few were tenants, less than a fifth paid rent and only a third of these received Housing Benefit. Overall, only three per cent of those working 16 or more hours per week continued to get Housing Benefit in work.
Women were a small majority. Rates of illness and disability were high - 28 per cent had a long-term health problem - and, in addition, these respondents were poorly educated. More than half had no qualifications, rising to eight out of ten among older workers, while the young rarely had more than basic or vocational qualifications.
The economic and social profile of workers did not differ significantly between Scheme A, Scheme B and Control Areas, nor between the 1996 and 1999 samples, except that the later sample contained more older workers.
Awareness, eligibility and claiming - Awareness of ETU was low and falling – only 29 per cent had heard of a new benefit for working people without dependent children. Even in places where ETU was more popular, two-thirds were unaware of it and fewer could name ETU. The effectiveness of networks as sources of information had risen since 1997 to 44 per cent, but 29 per cent had been told by the DSS, Employment Service or Benefits Agency.
Overall, 47 per cent of current workers in the pilot areas qualified for ETU, 39 per cent in A and 54 per cent in B. However, if B rates were available in Scheme A, B and Control areas, about half the sample would have qualified in each area, underscoring the all-important similarity between them.
Eight out of ten low-paid workers still in work of 16 or more hours at interview had never claimed ETU. Current receipt was concentrated among older single women, while receipt among couples and home-owners was rare. The proportion of “eligible ”workers claiming ETU rose from the 18 per cent found in the 1997 follow-up sample, but remained low at 23 per cent overall: 14 per cent in Scheme A and 30 per cent in Scheme B.
Eligibility and claiming ETU: why did so many low-paid workers fail to take up their entitlement to ETU?
Take-up among couples - The small proportion of couples among the ETU caseload was due to non-take-up among eligible working couples, not to a lower rate of eligibility. Eligible couples were about as numerous as the eligible single workers were, but only 10 per cent of them claimed their ETU compared with 37 per cent of single people. In Scheme B single people had a take-up rate of 43 per cent, which was much more in line with expectations, especially since the new benefit had had so little public promotion.
Modelling take-up rates - Multivariate analysis indicated that, other things being equal:
- Eligible workers who took up their entitlement to ETU“ ”“”were young, single workers, many of them living in Sunderland and Doncaster. They worked shorter hours and so had lower earnings, higher entitlements and felt hard up. Often they had experience of claiming other income-tested benefits such as Housing Benefit and Family Credit or came from families who did.
- Eligible workers who failed to take up their entitlement to ETU were couples, especially dual-earner couples, and older people who either owned their own homes or lived in some other arrangement rather than as homebuyers or tenants. Additionally one of them might be receiving a disability benefit. They lived in a Scheme A area or in Bournemouth. They had no experience of claiming other income-tested benefits, had slightly higher earnings and so were entitled to slightly lower amounts of ETU and they were used to managing on a low income.
In addition, the lack of publicity was crucial. In the first six months it was limited to non-electronic media and then it vanished altogether, leaving it to official and informal networks to inform workers of ETU. Few eligible non-claimants were aware of the benefit. Geographical and social isolation were important factors in inhibiting claims among eligible workers. In geographical terms, eligibility under Scheme A was confined to too small a band of the lowest incomes. This meant that eligibility was too sparsely scattered to support adequate networks of informal information that would prompt others to claim. In social terms, lowest-paid older couples, dual earners, homeowners, people with disabled partners and so on were all groups isolated from streams of information that would prompt claiming a new in-work benefit, even in Scheme B areas.
Skill transfer was also important. It is highly significant that claiming ETU was both need-driven and associated independently with prior experience of claiming income-tested benefits, especially in claiming Housing Benefit and Family Credit. Those picking up their entitlements, as well as being young, single and connected to information sources, were the lowest-paid of eligible workers and sensitised to news of a subsidy by the experience of hardship.
Job retention among ETU workers
Workers interviewed before the introduction of ETU were followed up twice, first in 1997 and again in 1998. Alongside them, ETU recipients interviewed in 1997 were also followed up in 1998.
There was no evidence that workers left their jobs at a slower rate in the Scheme A or B areas compared with the Control areas. Nor was there direct evidence of people claiming ETU and so hanging on to their jobs longer. Furthermore, there was no indirect evidence of job retention being better in the presence of ETU in the Scheme A or B areas compared with the Control areas.
Nor was there any retrospective evidence in the 1999 surveys that workers in the Scheme A or B areas had had better or more secure work histories, either directly as a result of receiving ETU or by working in a more secure employment market that was now underwritten by wage subsidy.
Examining wage and substitution effects using administrative data
Department of Social Security (DSS) records of awards of ETU and of Family Credit were examined for two possible effects resulting from the introduction of ETU:
- A wage effect - The wage bargaining position of everyone at the bottom end of the wage distribution might become weaker. Employers might respond to the increasing numbers willing to work for very low wages by buying labour more cheaply, that is low wages might become lower or remain steady.
- A substitution effect - ETU recipients might move into jobs currently occupied by low-paid workers with children claiming FC if few new jobs were created by the introduction of ETU.
In practice, there are reasons why we might not expect to find such effects as a result of the ETU pilot scheme:
- Around seven in 10 ETU recipients were already in work when they applied for the benefit, giving employers less scope to reduce wages than they would have if they were fixing the wage of a new job.
- Both the number of claims and the proportion of eligible workers who claimed were relatively low. It is unlikely that the wage bargaining and employment position of low-paid workers would have been weakened by the introduction of ETU at this take-up and take-up rate.
- Though FC recipients also occupied the low-paid end of the labour market, they worked for wages which were, on average, higher than the wages of those claiming ETU and they were different populations in many respects. ETU would not necessarily have significant effects on the slightly different labour market they occupied.
Administrative data on the wages of ETU recipients themselves offered no conclusive evidence that employers responded to ETU by holding down the wages of ETU recipients. However the failure of ETU recipients’ wages to rise across the evaluation period to April 1999, particularly among couples, suggests that it should not be ruled out as a possibility.
In the Family Credit wage data, similar trends occurred both before and after the introduction of ETU and in pilot areas, control areas and areas outside the evaluation alike. That is, they showed no significant effect on the wages earned by FC recipients in the same workforces as those chosen for the ETU pilot, compared with the control areas and with the rest of Britain. Wage data for employed lone mothers on FC, who typically earned lower weekly wages than other employees on FC and might therefore be the most vulnerable to an ETU wage effect, gave no indication of a downward pressure on their wages caused by the introduction of ETU. This was true generally and also specifically within Sunderland where ETU had penetrated a significantly larger proportion of the low-paid workers without children than it had elsewhere.
There was no evidence from the data as a whole of a substitution effect occurring among FC recipients in the pilot areas. This was true for both FC recipients as whole and specifically for employed lone mothers receiving FC whose weekly wages were on average lower than those of other employees receiving FC and therefore closer to those of ETU recipients
Conclusions
“Administrative data - ”Overall, evidence taken solely from administrative data would have concluded that ETU was on balance a success. In particular, data on the wages of ETU recipients offered no conclusive evidence that employers had responded to ETU by lowering or holding down the wages of workers in the low paid end of the labour market, although it could not be ruled out. The Family Credit data showed no significant effect of ETU on the wages earned by FC recipients in the ETU pilot areas. Nor was there any evidence of ETU eligible workers replacing FC eligible workers in the pilot areas at lower wages.
Surveys of ETU recipients and the workers-in-work - earnings top-up reached the people it was intended to reach - the lowest paid workers with little education, small housing costs and a record of unemployment - but it did not reach enough of them. However, those it did reach were glad to have it and there was evidence that it met need.
Explaining non-take-up and its implications for policy - ”A large proportion of non-take-up among eligible workers was accounted for by:
- the absence of main-media publicity,
- the sparse distribution of eligibility under Scheme A, and
- the social and economic isolation of older couples and their lack of claiming-skills.
The much higher take-up rate among those already attuned to the benefit system – younger low-paid single workers and tenants living in Scheme B areas – indicated that a new version of the benefit could be more successful, especially one that:
- allowed for the introduction of the National Minimum Wage with adjusted thresholds,
- adopted a shallower taper like Working Families Tax Credit, that is one which was withdrawn at a lower rate against income above the earnings threshold
- was backed by national television and radio advertising which would penetrate the relative isolation of poorly paid older couples, especially dual earner couples and those who have their own homes, or work in residential settings.
Such a benefit or tax credit would extend eligibility to large numbers of lower-paid workers, especially so if the estimates of the take-up rate for ETU from the workers-in-work surveys were accurate. Evidence from the surveys of ETU recipients suggests that they were accurate.
Job retention - No evidence was found to encourage the view that people kept the jobs they had more easily because of support from ETU, though qualitative research found that some people said they would find it hard to cope without ETU (Heaver et al, 2001).
Conclusion - There was little evidence from the evaluation surveys of ETU recipients and workers to oppose the view that most of the expenditure on ETU in the pilot areas went to people who:
- would anyway have done the jobs they did,
- or who took the jobs they would have taken,
- worked the hours they would have worked, and
- worked for the wages they would have otherwise accepted.
PART TWO - Econometric analysis: Assessing Employment Effects - Employment History Analysis.
This part of the report makes use of evidence collected by survey methods to evaluate the wider labour market and longer term effects of the Earnings Top-up Pilot Scheme, an in-work benefit paid to certain categories of low-paid workers and available in eight localities from October 1996 to October 1999. The evaluation methodology involves the collection of information from participants and non-participants in the eight pilot scheme areas and from non-participants in ‘comparison areas’ – localities in which the scheme was not available. A statistical comparison between these various groups allows certain conclusions to be drawn about the wider labour market effects of the scheme.
The term ‘wider labour market effects’ is used to refer to the impact of the scheme on non-participants. A consequence of measures that are designed to promote the employment of one group is that these might reduce employment for persons who are not eligible for them. Comparison of the movements into and out of employment for ETU-recipients and non-recipients as the scheme built up following its inauguration in October 1996 did not reveal any such effects which are statistically significant and which might be attributed to its operation. However, the analysis presented here was restricted to a 10-month period from October 1996 to August 1997. It could well be the case that the effect of the scheme on employment opportunities for non-participants took longer to materialise than 10 months.
The longer-term impacts of the scheme are assessed by examining the early work histories of those who participated in the scheme, comparing them with those who did not participate. Additionally, this part of the study takes advantage of the fact that the ETU Pilot Scheme has now ended, facilitating observation of the subsequent movement out of work for those who were participants. This investigation revealed that ETU-recipients had a weaker degree of attachment to the labour market than a sample of low-paid employees. Their employment profiles in the five years prior to their participation appear similar to persons who had experienced a spell of long-term unemployment in the year preceding the start of ETU. Importantly, they display no sign of returning to their earlier levels of joblessness as the scheme now unwinds.
There are, however, a significant number of caveats that must be attached to these conclusions. First, the so-called ‘comparison areas’ may be fundamentally different from the scheme areas, invalidating their use for statistical comparison. Second, the survey samples used suffer from significant attrition over the period of the study, particularly for the postal follow up survey in 1998. Third, the problem of sample selection is endemic to this study. The original methodology for evaluation of the ETU Pilot Scheme envisaged that samples of low-paid workers and unemployed persons (i.e. potential ETU claimants) would be selected in 1996, then followed at intervals throughout the period that the scheme operated. Due to the extremely low take-up rate among these two groups, this methodology was modified to include a sample of scheme participants and to vary the frequency and method of follow up employed.
Despite these difficulties, the analysis presented here reveals a number of interesting findings. First, there is little evidence from these survey data to support the notion that, in its early stages, the scheme operated to the detriment of non-participants. Among a sample of persons who were low-paid workers in 1994-95, there is no evidence that the scheme operated to reduce their later experiences of employment. While it does appear to be the case that a higher rate of employment is observed among previously low-paid workers in the comparison areas than in the scheme areas, the fact that this higher rate existed for two years prior to the scheme suggests that it is a structural effect and does not relate to the operation of ETU. Secondly, there is no evidence to suggest that previously unemployed persons had more difficulty regaining employment in the scheme areas than in the comparison areas. Multivariate analysis of the duration of spells of employment and non-employment by scheme participants and non-participants also fails to find evidence of any effects on the duration of such spells that can be associated with operation of the scheme. However, it is important to point out that even at its peak the scheme only covered between 1 and 2 per cent of the employed population in the areas in which it was piloted. It is unsurprising therefore that no statistically significant effects of its operation among non-recipients can be detected by survey methods.
In terms of the longer-term impacts, the conclusions presented here must remain tentative, simply because insufficient time has elapsed for longer-term impacts to become apparent. From an analysis of the early work histories of the three groups of survey respondents (two selected because of their participation in a spell of employment and one selected because of participation in a spell of unemployment) it is apparent that the early work histories of ETU participants bear more resemblance to the early employment experience of persons who have had a fairly lengthy spell of unemployment than to persons who were in low paid employment in 1994-95. At the beginning of the decade, and discounting the effects of participation in full-time education on employment rates, the average employment rate of both ETU recipients and those persons who experienced a significant spell of unemployment in 1994-95 was approximately 60 per cent. For those in low-paid employment in 1994-95, the rate was about 25 percentage points higher. Some convergence in these rates is apparent in the period October 1997 to November 1998, but those who experienced a significant spell of employment in 1994-95 do not regain their earlier employment rates, whereas ETU recipients continue to display higher rates of employment than in their earlier work histories, possibly by between five and ten percentage points. Analysis of the benefits received by ETU-recipients as the scheme winds down does not reveal any evidence as yet of a return to the high rates of unemployment that marked the work histories of ETU-recipients prior to their participation. This suggests that ETU did well to locate its target group (persons with a history of poor attachment to the labour market) and may well have an impact upon their attachment that outlives the scheme itself.
Publication details
Marsh, A., Stephenson, A., Dorsett, R and Elias, P., (2001) “Earnings Top-up Evaluation: Effects on Low Paid Workers” Department of Social Security Research Report No. 134, CDS: Leeds (£37.00)
Relevant publications
Finlayson, L., Ford, R., Marsh, A., Smith, A., and White, M., (2000) “The First Effects of Earnings Top-up”, Department of Social Security Research Report No.112) CDS: Leeds
Green, A. (2001) “Earnings Top-up Evaluation: Labour Market Conditions”, Department of Social Security In-house Research Report No. 75, London
Heaver, C. Roberts, S. Stafford, B. and Vincent, J. (2001) “Earnings Top-up Evaluation : Qualitative Evidence”, Department of Social Security Research Report No. 133, CDS: Leeds
Lissenburgh, S., Hasluck, C and Green A., (2001) “Earnings Top-up Evaluation: Employers’ Reactions”, Department of Social Security Research Report No. 132, CDS: Leeds
Marsh, A., Callender, C., Finlayson, L., Ford, R., Green, A and White, M., (1999) “Low Paid Work in Britain”, Department of Social Security Research Report No. 95) CDS: Leeds
Marsh, A., (2001) “Earnings Top-up Evaluation: Synthesis Report, ”Department of Social Security Research Report No. 135, CDS: Leeds
Smith, A., Dorsett, R. and McKnight, A., (2001) “Earnings Top-up Evaluation: Effects on Unemployed People”, Department of Social Security Research Report No. 131, CDS: Leeds
Vincent, J., Abbott, D., Heaver, C., Maguire, S., Miles, A., Stafford, D., (2000) “Piloting Change”, Department of Social Security Research Report No. 113 CDS: Leeds
Vincent J., Heaver, C., Roberts, S. and Stafford, B., (2001) “Earnings Top-up Evaluation: Staff Views”, Department of Social Security In-house Research Report No. 74, London