The first effects of Earnings Top-up: interim survey findings from the Earnings Top-up 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. 112
By Louise Finlayson, Reuben Ford, Alan Marsh, Alison Smith 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 were piloted: lower rate Scheme A and higher rate Scheme B. This report brings together the results of surveys of employers, low-paid workers, unemployed people, and recipients of ETU. Administrative statistics and labour market studies are also incorporated. The first field surveys were carried out before the introduction of ETU in 1996. The same respondents were seen again in 1997, together with new ETU customers. The interim results presented here will be supplemented by findings from surveys conducted in 1998 and 1999 in the final evaluation reports. Findings at this stage should be seen as provisional.
- The ETU recipients surveyed were typically young (43% were aged under 25) and single (79%). More than half (56%) had no formal housing costs although those who did reported difficulties in paying for accommodation. The majority of ETU recipients surveyed left school at the minimum school leaving age (62%). 18% of ETU recipients reported a long-term illness or disability. Although in work, 10% of recipients said that illness or disability placed limits on the work they could do or would make moving jobs difficult. Very few were receiving a disability benefit.
- Among low-paid workers more than a third of those eligible for ETU in 1996 had moved out of eligibility by 1997, mostly because of improved incomes, although others had lost their jobs. Many of those unemployed in 1996 had moved into employment by 1997 (38%). Short periods of eligibility may explain in part low estimates of take-up.
- Among low-paid workers only a small proportion of those potentially eligible for ETU were claiming it, with take-up estimated to be 18%. Non-take-up was related to being part of a couple, being an older man, or having a low ETU entitlement.
- Administrative statistics showed that most recipients were already in work when they started claiming ETU (70%). Survey evidence also suggested that those in work aware of ETU claimed it, and for those moving into work ETU did not appear to be a deciding factor.
- No quantitative evidence was found that ETU had affected the rate of job loss in the pilot areas compared to the control areas. Equally, no evidence was found that ETU was encouraging people to reduce their hours, or that earnings were depressed.
- The results from Scheme A suggested that ETU had little or no impact in those areas by 1997. In Scheme B areas awareness and take-up were higher than in Scheme A and wage expectations of jobseekers were lower than in the control areas. If any effect of ETU is to be found, it would expected to be in Scheme B areas. Analysis of surveys conducted in 1998 and 1999 will seek to identify whether these patterns continued.
The Introduction of Earnings Top-up
Earnings Top-up (ETU) was introduced as a pilot scheme in October 1996, extending wage supplementation to low-paid workers without dependent children. Two versions of the new benefit were introduced in each of four areas (Scheme A and Scheme B). These areas were comparable with four control areas. The Policy Studies Institute (PSI), together with colleagues in the Institute for Employment Research (IER) and the Centre for Research in Social Policy (CRSP), is carrying out a four-year evaluation study of the impact of ETU.
This report deals with the first two years of the study. It brings together the findings of several surveys of employers, low-paid workers, unemployed people, and recipients of ETU. Administrative statistics and labour market studies by IER are also incorporated in these findings. The first field surveys were carried out before the introduction of ETU in 1996. The same respondents were seen again in 1997, together with new ETU customers. A complementary report covers interim findings from the qualitative research programme carried out by CRSP (DSS Research Report 113).
In the eight pilot areas, the new benefit was taken up swiftly, exceeding the predicted caseload of 20,000 in a little over a year. Take-up was uneven, though, with far more customers in the Northern industrial areas compared with the South. Scheme B attracted more customers than Scheme A: about 13,000 compared with 10,000, and take up grew fastest among single people, particularly those aged under 25.
ETU claims and claimants
The new customers
The 2,000 ETU recipients interviewed in 1997 proved to be the kinds of people the 1996 baseline surveys had suggested they might be (Marsh et al, 1999). Typically they were young (43 per cent were aged under 25) and single (79 per cent). More than half (56 per cent) had no formal housing costs. For these people, this cleared the way for ETU to act as an additional incentive to work, rather than be lost against a background of in-work entitlements to Housing Benefit (HB). Those who were responsible for paying for their accommodation often found it difficult. Rent arrears were common, and many recalled special difficulties in meeting housing costs when moving off benefit and into work. The majority of the renters, though, were aware of help from HB for those in work.
Many people had characteristics that would have made it difficult for them to move into higher paid work. Though in work, 10 per cent reported persisting health difficulties that limited their employment. One third of the sample (32 per cent) had no qualifications at all, 56 per cent had some academic qualifications and 46 per cent had vocational qualifications. Very few of the older ETU recipients had any formal educational qualifications; and even among younger recipients qualifications above GCSE level were rare. Fewer than half (48 per cent) of ETU claimants held driving licences.
The experience of working in ETU jobs
The high rate of status-mobility that characterises this kind of population, was evident in the sample: on the basis of information given at interview 22 per cent of respondents were calculated as ineligible for ETU. Around half of these were no longer working 16 or more hours a week while the others had seen substantial average wage rises that took them out of eligibility for ETU.
The majority of those remaining in work thought their jobs were permanent and for the most part they liked what they did and had no plans to move. There were predictable concentrations of women in service occupations and of men in routine plant operative and labouring jobs. A quarter of the women were working in the health sector, many of them in care homes. Most of the ETU recipients worked in smaller commercial enterprises and few belonged to Trades Unions. Single recipients typically earned about £84 a week and couples £97 for a 32 hour week, on average. Men tended to work slightly longer hours at lower rates of pay compared with women.
The experience of claiming
When interviewed in 1997, many respondents were settling down in ETU; 41 per cent were into their second claim. As with Family Credit, some respondents were slow to realise their likely eligibility and delayed a claim, often for months, which may have depressed take-up rates. Those in jobs said they had learned about ETU from information in the media or, more often, from informal sources. Those coming in from Jobseekers Allowance (JSA) were more likely to have been prompted by Benefits Agency or Jobcentre sources, though not overwhelmingly so. Overall, the impression was that those who had claimed ETU successfully had claimed easily and only a handful had reported any real setback during the claim process.
Eleven per cent of ETU recipients who entered their job after the introduction of the new benefit had the impression that their new employers had made room for them. But they did not think that it was ETU that had somehow made them more employable. Most of them thought they would have taken the job anyway, particularly since half of them were unaware of ETU when they accepted the offer.
The impact of ETU on existing workers
Labour market change and eligibility for ETU”Existing low-paid workers typically experienced a great deal of household and labour market change. In less than a year, more than a third of those eligible for the benefit in 1996 had moved out of eligibility by 1997, largely by improving their incomes. Overall, 67 per cent of the working sample in 1996 would have been eligible for Scheme B, had it been available, but by 1997 this proportion had fallen to 44 per cent. Some of these left eligibility because they left employment. Of those who were working both in 1996 and 1997 just over half (53 per cent) remained eligible for ETU by 1997. This degree of movement in and out of eligibility suggests that some people may not have remained eligible for the benefit long enough to get around to claiming it which may affect estimates of take-up rates based on point-in-time measures.
Awareness of ETU and take-up rates
Overall, 34 per cent of working respondents in 1997 had some awareness of the introduction of a new benefit but only 18 per cent could name ETU. Awareness was higher in Scheme B areas (39 per cent were aware of the benefit and 24 per cent knew it was called ETU) compared with Scheme A areas (30 per cent and 12 per cent respectively). Workers most commonly heard about ETU from advertising (35 per cent) or from workmates, friends or family (30 per cent).
Take-up rates followed a similar pattern to levels of awareness. At most, it was estimated that just 18 per cent of the workers sample who were entitled to the benefit were actually receiving it when interviewed (11 per cent in Scheme A and 23 per cent in Scheme B areas). In turn this take-up rate was related to the size of the entitlement and the marital status of the potential recipient, as respondents with a partner were less likely to claim ETU despite having above average entitlements. Single women and men aged under 25 were generally more likely to claim ETU than were older men. For single people, having a low entitlement to ETU meant they were less likely to claim. Self-employed people were also more likely to be an eligible non-claimant.
Job loss and ETU
There was little evidence of a strong difference in the rate of job loss between the areas where ETU was available, and where it was not. Furthermore, there was no evidence that receipt of ETU had altered the behaviour of workers as eligibility for ETU captured those who already had the lowest weekly earnings and hours.
The unemployed sample in 1997
Movements into work”Seventeen per cent of the medium-term unemployed people sampled in 1996 had found a job by the time of their first interview in 1996. By 1997, 38 per cent were in work. Over the year as a whole, nearly one half (46 per cent) had held a job at some time. Seventy per cent of those who had obtained jobs had stayed in them through to the 1997 interview, and 28 per cent of those out of work in 1996 were in employment at the 1997 interview. However, only 42 people were working and claiming ETU at the time of the 1997 interview. These were 17 per cent of those who had held a job in the Scheme areas.
Awareness of ETU and take-up rates
Eligibility for ETU among those who had moved into work was 23 per cent in Scheme A areas and 41 per cent in Scheme B areas. Relative to these eligible groups, the take-up rates were 21 per cent in Scheme A areas and 46 per cent in Scheme B areas, somewhat higher than among the existing workers sample. This may reflect greater knowledge of the benefit because they had recent contact with benefit office staff: 43 per cent of those aware of ETU said they had heard about it from official sources. Overall, awareness of ETU among people who remained unemployed was similar to that among existing workers: one third (34 per cent) had heard of the introduction of a new benefit but just 15 per cent could name ETU. Awareness was again higher in Scheme B areas (39 per cent) than in Scheme A (28 per cent).
Earnings and employment
For those who had moved into work, average earnings had risen by about one-third between 1996 and 1997 taking many of the new workers beyond eligibility for ETU. Generally, such wages had risen in all areas: wages in the Scheme A areas were close to, and possibly a little higher than, those in the Control areas. Wages for new workers in the Scheme B areas, however, were estimated to be about 12 per cent lower than in the Control areas, and most of this difference was concentrated in jobs obtained after the introduction of ETU.
Workers in Scheme A and Scheme B areas were significantly more likely to have remained in employment that had commenced before the introduction of ETU than were those in the Control areas. However, there was no indication that either Scheme A or Scheme B areas differed from Control areas in terms of new post-ETU employment. There is, then, no clear indication that ETU had influenced the sustainability of employment for this group.
Wage expectations of unemployed people
Wage expectations had, on average, remained fairly static over the 1996-97 period, despite the improvement in the job market. At the 1996 interview, those in Scheme A areas had significantly higher wage expectations than those in Control areas and in Scheme B areas. Over 1996-97, both Scheme A and Scheme B averages moved down relative to Control areas. At the time of the 1997 interviews, Scheme A and Control areas recorded similar wage expectations among jobseekers. In Scheme B areas, however, wage expectations were significantly lower than in the Control areas by about seven per cent.
Inactivity
While many people had moved into what appeared to be relatively stable employment (38 per cent), another larger group (53 per cent) remained non-employed in both years. This group included a substantial proportion of people who were economically inactive. Whereas 27 per cent of the sample had been inactive at the time of the 1996 interview, this had increased to 30 per cent by 1997.
Scheme A and Scheme B areas had higher rates of persistent inactivity (not working at both 1996 and 1997 interviews) and new inactivity (working in 1996 but inactive by 1997) than did the Control areas but these results were borderline in terms of statistical significance. The strongest factors that increased the probability of being inactive in this sample were having a long-term illness and being female.
The effect of ETU on employers
Growth of workforce
Between 1996 and 1997, there was evidence of an increase in the average size of employers that corresponded to a decrease in unemployment. However, among the job groups where ETU would be expected to have its greatest effect - semi/unskilled and clerical/sales jobs - there was no evidence that employers had increased employment volumes at a faster rate in the pilot areas compared with the Control areas. Nor was there evidence that recruitment difficulties had eased following the introduction of ETU in pilot areas compared with controls.
Hours worked
There was preliminary evidence of a slight change in the hours said to be worked by employees, particularly in the 16-29 hours per week slot that might be associated with the introduction of ETU. For example, a greater proportion of semi/unskilled employees in the pilot areas (Scheme A and B) were employed for 16-29 hours per week in 1997 compared with equivalent workers in 1996. This was not the case for these types of workers in the Control areas where the proportion working shorter hours fell slightly.
Awareness of ETU
Awareness of in-work benefits was high, although only three out of 10 employers reported any actual experience of these benefits. Awareness of ETU rose between 1996 and 1997, particularly in Scheme B areas but only five per cent of employers in the pilot areas could recall any direct experience of ETU by 1997. Among those employers who had had no experience of in-work benefits, there was a lot of uncertainty as to the likely effects of such benefits on their establishment. Just five per cent of employers in the pilot areas who said they had heard of ETU reported that they had tried to make assessment of the likely effects of ETU on their establishment. The majority of employers in the pilot areas thought either that ETU would not affect them or they did not know what the effects might be. There was little evidence of any conscious change in the behaviour of employers in response to the introduction of ETU.
Wage effects
The three most common influences on the levels of pay reported by employers in 1996 - the pay individuals were willing to accept; the pay offered by other local employers; and the availability or scarcity of labour suggested considerable scope for a wage-effect following the introduction of ETU. Small, private-sector employers based in rural and seaside areas would have the greatest capacity to depress their wages in response to the new benefit. However, there was no clear evidence that this had happened by 1997.
The introduction of ETU had no statistically significant effect on the change in hourly wages of clerical/sales and semi/unskilled employees between 1996 and 1997. In fact, between 1996 and 1997 there was an overall reduction“ ”in the proportion of semi-unskilled employees paid at a rate of £4 per hour or below, equally in all areas. The proportion of employers who said they had recently recruited employees at a rate of £4 per hour or below also fell between 1996 and 1997. The introduction of ETU was found to have no significant effect on the change in the proportion of low-paid employment or on the change in wages offered to these low-paid recruits.
Examining the Effect of ETU on Family Credit Recipients“Wage effects”There was no evidence from the Family Credit (FC) administrative data of any ETU wage effect in the 11 months following the introduction of ETU. There was, however, some evidence of a slowing in the rate of increase in mean weekly earnings, though not mean hourly earnings, in the Control areas over the same period. The reasons for this are unclear though they are clearly not related to the introduction of ETU in the pilot areas since a similar change in mean weekly earnings was not seen in the five per cent sample of areas outside the evaluation.
Substitution effects
By November 1997, there was no evidence from the FC administrative data that employers were replacing FC workers with lower paid workers as a result of the introduction of ETU in the pilot areas.
Area Differences in the Impact of ETU“Differences between Scheme A and Scheme B”ETU administrative statistics suggested a few clear differences between areas. Scheme A areas contained proportionately more couples, more self-employed and more older recipients than Scheme B areas. There were higher numbers of recipients in Scheme B areas and these recipients typically had higher weekly earnings and hours (as eligibility conditions allowed). Average awards differed little between areas.
Local labour markets
The detailed labour market studies showed that in 1996 there were few distinguishing differences between Scheme A, B and Control areas, which was intended in the design of the whole pilot. Subsequently, the data showed a large decline in gross levels of unemployment in all the 12 areas, equally in pilot and Control areas, though these falls were much larger from the stock of short-term rather than long-term unemployment. All areas showed large net outflows from unemployment in early 1997 and a corresponding rise in the notification of unfilled vacancies. Overall, there was some evidence that unemployment fell faster in all ETU pilot areas compared with the Control areas. However, there was no evidence of a marked change in unemployment flows after the introduction of ETU. Nor was there evidence of an effect on earnings.
Conclusions
Deadweight costs
There is little in the data to oppose the view that most of the first year of expenditure on ETU in the pilot areas had gone to people who would anyway have done the jobs they did, working the hours they would have worked for the wages they would otherwise have accepted. Some ETU customers (43 per cent of those aware of ETU before taking up employment) said they would not have accepted the wages offered to them without ETU. They had also come to feel that ETU had become an indispensable part of their in-work income. But the majority view was that they would return to coping without ETU if they had to.
On the other hand, the initial deadweight cost was blunted somewhat by the low take-up rate. Indeed, if the predicted take-up“ rate” had been reached at the same speed as the predicted take-up volume was reached, 40-45,000 claims would be in payment. It is also a passive, rather than an active, deadweight cost. For example, there is no evidence that existing workers who claimed ETU smartly pocketed the benefit and then reduced their working hours, buying new leisure with their benefit.
The incentive effect
It is easy to confuse an incentive to work with an incentive to claim. Research to date confirmed that the great majority of people potentially in line for ETU already had a substantial incentive to work. Relatively few were entitled to more than basic income-based JSA payments out of work. Even those earning less than £3.00 an hour already had wages that were getting on for double their out-of-work incomes. Adding typically £20-30 of wage subsidy from ETU on top of this certainly added to an incentive to claim. The key question to answer is why so few of those in the sample who became eligible for the benefit claimed it. Part of that answer is that eligibility is short-lived.
It is also possible that, in the presence of ETU, some people may have moved into work faster than they might otherwise have done, irrespective of whether they became claimants. The evidence is small and based solely on the surveys of movement in all unemployed people in the 12 labour markets covered by the experiment. But this could be quite a plausible effect - that the introduction of ETU may have given a general boost to the most difficult end of the jobseeking market without actually placing every new low-wage job finder directly into the path of the benefit.
Wage effects
Taken together, the labour market evidence that there may have been some acceleration of the unemployed into work that had not also attracted a downward shift in wages is a positive outcome. It would be unusual, not to say alarming, if a clear downward shift in wage rates could be associated directly with the introduction of ETU in the pilot areas. On the other hand, if ETU did have any significant impact, there would be a wage effect of some kind, even if it simply increased the flow of the unemployed into low-wage jobs that had existed as hidden unfilled vacancies. If ETU worked in the way it was supposed to work, more people will have ended up working for the lowest wages than worked for these wages before ETU was introduced. The evidence so far is sparse but this will be investigated further as the evaluation continues.
Summary
A fairly consistent pattern of results for Scheme B suggested that, if there was an ETU effect, it was likely to be in those four areas. Eligibility, take-up and awareness of ETU were higher in the Scheme B areas than in the Scheme A areas. Wage expectations of the jobseekers in Scheme B areas were significantly lower than in the Control areas, and lower wages were recorded for those in Scheme B jobs. Further, those in Scheme B areas were more likely than those in Control areas to have held onto jobs which they got before the introduction of ETU.
Generally, results for Scheme A areas suggested that ETU had little or no Impact. There was relatively low rates of eligibility and take-up, and awareness was somewhat lower than in Scheme B areas. Wages and wage expectations were not significantly different in Scheme A from Control areas. On balance, it appears possible that some early effects of ETU have been detected in the Scheme B areas, but none in the Scheme A areas.
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
Marsh, A., Callender, C., Finlayson, L., Ford, R., Green, A., and White, M., (1999), Low Paid Work in Britain: Baseline Surveys from the Earnings Top-up Evaluation (DSS Research Report 95), CDS.
McKay, S., Smith, A., Youngs, R., and Walker, R., (1999) “ Unemployment and Jobseeking after the Introduction of Jobseekers Allowance” (DSS Research Report No. 99), CDS
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
Thomas, A., Pettigrew, N., Cotton, D., and Tovey, P.,(1999) “Keeping In touch with the Labour Market ”(DSS Research Report No. 96), CDS
Trickey, H., Kellard, K., Walker, R., Ashworth, K., and Smith, A., (1998) “Unemployment and Jobseeking: Two Years On” (DSS Research Report No. 87), CDS
Vincent, J., Abbott, D., Heaver, C., Maguire, S., Miles A., and Stafford, B., (2000) “Piloting Change: Interim Qualitative Findings from the Earnings Top-up Evaluation” (DSS Research Report No. 113), CDS.