Implications of the National Employment Savings Trust for vulnerable sectors of the UK labour market: a reduced-form statistical evaluation.

van de Ven, Justin

The National Employment Savings Trust (NEST) is a new pension scheme that will be introduced for employees on modest incomes in the UK from 2012. This paper draws out the implications of the NEST, focusing upon low-paid employees and their employers using data from the Annual Survey of Hours and Earnings. The results suggest that the NEST will increase labour costs by between 0.6 and 0.8 per cent on average, and have a disproportionate effect on low pay industries and private sector firms employing fewer than 25 employees. The analysis highlights the potential role of the minimum wage to shield low pay workers from paying a share of the subsidies that the NEST will afford to its members.

Keywords: Pension saving; low pay; employer pension

JEL Classifications: D31; H22

I. Introduction

The National Employment Savings Trust - NEST - is a new Defined Contribution (DC) pension scheme that will be introduced for employees with modest earnings in the United Kingdom this year. (1) The objective of the scheme is to increase private savings, and its success or failure is likely to have a profound bearing upon the way in which the UK pensions system evolves, with important implications for the wider group of countries that face similar pension challenges to those of the UK. This paper focuses upon the financial implications of the NEST for employers and their employees, which are likely to be important in determining how favourably the scheme is received.

The introduction of the NEST reflects a contemporary trend toward greater reliance on DC pension provision in the private sector in the UK, and a similar trend among OECD countries more generally. (2) It is being introduced following recommendations made by the Pensions Commission (2005), which found that administration costs made it unprofitable for existing private sector pension providers to serve employees on modest incomes. The NEST is consequently designed to improve savings incentives by reducing management charges, and by requiring all employers to offer a matching pension contribution on banded earnings to eligible employees. It has been forecast that the scheme will serve between 5 and 11 million people, equal to between one fifth and half of the employee population, and will receive contributions worth 9 billion [pounds sterling] annually (2009 prices), 50 to 70 per cent of which is projected to be new saving. (3)

All employees with total annual earnings (including wages, overtime, bonuses and commission) in excess of 5720 [pounds sterling] and between 16 and 74 years of age will have the right to participate in the scheme. An employee who participates in the NEST will be required to pay an annual contribution into the scheme worth at least 5 per cent of their gross earnings between 5720 [pounds sterling] and 33540 [pounds sterling] (in 2010), and their employers will be required to pay a matching contribution worth at least 3 per cent of banded gross earnings. All pension contributions are exempt from income tax, so that an employee who pays tax at the marginal rate of 20 per cent (the basic rate) will effectively pay a contribution out of post-tax income that is worth 4 per cent of gross earnings, and will receive tax relief worth 1 per cent. Contributions to personal pensions in the UK are currently accessible from age 55, at which time up to 25 per cent of the accrued fund can be taken as a tax free lump sum, with the remainder used to finance an income stream in retirement. The income stream that is generated by the pension is subject to income tax.

The NEST is designed to economise on management charges and minimise employee decision costs. The former of these objectives is to be met by shifting the focus of competition among private sector providers from individual pension members to pools of managed funds. It is hoped that this centralisation will reduce the effective annual management charges applied to pension saving for the target population, from a capital charge in excess of 1 per cent to the 0.3 per cent charge that is currently levied by large occupational pension schemes. To meet the second objective, the scheme has been explicitly designed to minimise the perceived welfare costs experienced by an employee who chooses not to make an active decision. Employers will consequently be required to auto-enrol into the NEST (or other qualifying pension) any employee who is aged between 22 years and state pension age, and whose annual earnings exceed 7475 [pounds sterling] (in 2011/12). All employees will retain the right to opt out of the NEST, so that the long-term viability of the scheme will depend upon rates of voluntary participation. (4) Similarly, contributions to the NEST will be allocated to a default investment fund, unless the member chooses to make an active investment decision.

The Pensions Commission relied upon a detailed set of statistical analyses when formulating its recommendations for reform, and this evidence base was extended during the ensuing public debate. When the Government enacted the reforms that will implement the NEST, however, it identified a need for further research into how the NEST is likely to bear upon the low-pay sectors of the labour market, and the associated implications for the regulatory environment (DWP, 2006a, p. 76). This paper reports analysis that is designed to help meet that information gap.

The importance of understanding the likely effects of the NEST on low-paid employees and their employers is heightened by the weak economic environment into which the scheme will be introduced. Ideally, we would have directly observed data--possibly derived from relevant pilot studies--upon which to conduct such an analysis. As the policy remains a counterfactual, however, we must rely upon indirect methods that project forward from the available evidence base. This paper consequently reports results from a simulation analysis of the likely effects of the NEST, using reduced form statistical projections estimated on contemporary survey data. A deliberately stylised analytical approach has been adopted to enhance transparency, which is useful given the significant uncertainties associated with the policy reform.

The effects that NEST contributions will have on total compensation and employer costs are likely to be an important factor in determining whether employees and employers engage positively with the scheme, and is central in determining how other aspects of the regulatory environment - such as the minimum wage should respond to the scheme. The analysis is structured around three key margins: the scale and scope of employee participation in the NEST, the relation between employer pension contributions to the scheme and the existing wage bill, and the associated bearing that the increased labour costs could have on the wages earned by employees.

Section 2 describes the existing state of employer sponsored pension provisions in the UK, which forms the policy background to the introduction of the NEST. The effects of the NEST on employees and their employers are explored in Section 3, and Section 4 concludes.

2. Existing private pension arrangements in the UK

The NEST will augment the existing system of private pension provisions in the UK, and these provisions are explored here. (5) For a detailed analysis of private pension arrangements in the UK that extends beyond employer pension involvement, see Emmerson and Wakefield (2009).

We begin by considering the prevalence of employer sponsored pension arrangements in the UK, before moving on to discuss the scale of employer pension contributions.

2.1. Prevalence of employer sponsored pension arrangements

Any individual under age 75 and normally resident in the UK has a statutory right to decide whether or not to contribute to a private pension administered outside the public sector. Contributions to a private pension cannot be drawn on until age 55, and can be offered on a defined benefit (final salary), and/or a defined contribution (money purchase) basis. Although most employers have a legal obligation to offer their employees access to some form of private pension, they currently have no obligation to make matching pension contributions--this will change with the introduction of the NEST. Contributions to private pensions by both employees and employers are, nevertheless, currently encouraged through a system of publicly funded tax incentives. These incentives have seen private pensions play a central role in the system of retirement provisions in the UK.

The decline in occupational pension provisions that has occurred over the past three decades is well documented, with a sustained downward trend in both the proportion of employees participating in such schemes, and a migration away from pensions that have typically afforded high rates of employer matching contributions. (6) The prevalence among employees of a pension arrangement with their employer has fallen from 65 per cent in 1995 to just over 50 per cent in 2009. At the same time, the share of employees with Defined Benefit occupational pensions--widely recognised as a gold-standard among employer sponsored schemes relative to all those with employer sponsored pensions has fallen from in excess of 80 per cent to just over 60 per cent. (7) Furthermore, restrictions to pension eligibility of new employees that are currently imposed by private sector employers are likely to see these trends continue well into the future (e.g. Forth and Stokes, 2008).

Table 1 reports the prevalence of employees that receive an employer pension contribution, disaggregated into labour market sub-categories that have an important influence on the prevalence of employer pension contributions. Note that whereas most studies of low-paid workers focus upon measures of hourly pay, the earnings distribution reported in the table is specified with respect to weekly wages. This reflects our current concern with the NEST, eligibility to which will depend upon annual earnings, and not the hours worked.

Table 1 reveals that the prevalence among employees of pensions that receive an employer contribution rises with wages, with the possible exception of the public sector where employer pension contributions are particularly common. Consistent with this observation, employees in low pay industries are, on average, half as likely to have a pension that receives a contribution from their employer relative to employees in non-low pay industries. What is interesting, however, is that controlling for employee earnings meets only a fraction of the gap in the prevalence of employer sponsored pension schemes between low pay and non-low pay industries; half of the employees in the highest earnings quintile receive an employer pension contribution within low pay industries, relative to 70 per cent in non-low pay industries.

The top half of table 1 indicates that employer pension contributions are practically non-existent in the low pay industry of hairdressing, and most prevalent in social care. Hospitality lies at the lower end of this scale, whereas retail--which accounts for almost 40 per cent of all low-pay sector employees by head-count--lies in the middle of the list.

Table 1 also indicates that the proportion of employees who receive an employer pension contribution tends to rise with firm size, both on average and throughout the earnings distribution. This is important because employees are concentrated in large firms; 64 per cent of all employees are identified as working in firms with more than 500 employees in the data considered here. Approximately half of both male and female employees are reported as receiving an employer pension contribution. However, women are disproportionately employed in low pay industries and in low paying positions relative to men, (8) which tend to be associated with lower rates of employer pension participation on average. The apparent contradiction arises because women tend to be more likely than men to receive an employer pension contribution after wages and industry are controlled for.

These observations suggest that the introduction of the NEST will disproportionately affect employees on low earnings, working for small employers, in low pay industries in the private sector. All else being equal, male employees are also likely to be disproportionately affected by the NEST.

2.2. Scale of employer pension contributions

Statistics regarding the distribution of employer pension contributions are reported in figure 1. This figure indicates that, for employees who receive an employer pension contribution, there is surprisingly little disparity between the distributions of contribution rates in low pay industries and the wider labour market. The distributions for employer pension contribution rates in both low pay industries and the wider labour market are dominated by a single mode between 12.5 and 15 per cent of employee wages. (9) The principal disparity between the distribution of employer pension contributions reported for low pay industries and the wider labour market is that the prevalence of employer contributions in excess of this mode is slightly lower among employees in low pay industries, with the difference contributing to a higher peak at the mode for low paid workers and spilling over into the 10 to 12.5 per cent band for employer contribution rates.

Six per cent of all employees in receipt of an employer pension contribution are identified as receiving a contribution worth less than 3 per cent of earnings, rising to 7 per cent of employees in low pay industries. This is important because it reveals that the additional financial burden consequent on the introduction of the NEST will predominantly operate through the extension of employee eligibility to any employer pension contribution; the NEST is unlikely to increase significantly labour costs for employers that currently provide pension contributions to their employees.

Disaggregating the statistics reported in figure 1 by population subgroup (not reported separately here) reveals that the scale of employer pension contributions--specified as a percentage of employee wages--is broadly insensitive to employee wage rates. Furthermore, although employer pension contributions do tend to be relatively low in selected low pay industries, (10) even then the average rates of employer pension contributions, where such contributions are made, remain in excess of twice the contribution rate that will be required under the NEST. These statistics underscore the risks posed by the potential for a general 'levelling down' of existing employer pension contributions to the minimum rate required under the NEST.

[FIGURE 1 OMITTED]

3. Financial implications of the NEST

This section reports projected effects of the NEST, focussing upon the implications for low-paid workers and their employers over the short to medium term. There are four key components that underlie the projections that are reported here: the labour market context, projected participation rates in the NEST, contribution rates of participants to the NEST, and employer responses to the scheme's introduction. It is beyond the scope of this paper to perform an exhaustive sensitivity analysis of the effects of the NEST. Rather, assumptions were selected that reflect the current focus of analysis, discussed in Section 3.1. Results of the analysis are reported in Section 3.2.

3.1. Analytical assumptions

Labour market context

The analysis takes as its starting point the distributional arrangements--regarding employment, wages, and other employer sponsored pension arrangements--that are described by micro-data from waves 2005 to 2009 of the Annual Survey of Hours and Earnings (ASHE, see Appendix A for details). This starting point is consistent with our focus on the short- to medium-term effects of the policy change.

Participation rates

Given the labour market context described by ASHE data referred to above, the rules applied by the NEST permit identification of those individuals who would be eligible to participate in the scheme. (11) The voluntary nature of the decision to participate in the NEST, however, calls into question the eventual rates of scheme membership. The analysis is based on the assumption that the NEST will receive the same rates of take-up that are reported for employer sponsored pensions in contemporary survey data, taking care to reflect distributional differences.

Although the ASHE does not provide information about any pensions to which an employee was eligible, but in which they did not participate, from 2006/7 the Family Resource Survey (FRS) has asked survey respondents whether they were eligible for an employer sponsored pension. These data reveal that 61 per cent of all employees believed that they were eligible to belong to a pension scheme run by their employer in 2006/7 and 200718. The same data, however, indicate that only 46 per cent of employees chose to participate in their employer's sponsored pension, with the vast majority of the remainder choosing not to contribute to a private pension at all (2.6 per cent of employees substituted a privately arranged personal pension for their employer sponsored scheme). Hence, just under one in every four employees who believed that they were eligible to participate in their employer's sponsored pension chose not to take up their option. Similarly, although any individual under age 75 and normally resident in the UK can choose to contribute to a private pension, the FRS data indicate that 45 per cent of all employees between ages 16 and state pension age chose not to participate in any form of private pension in 2006/7 and 2007/8.

Figure 2 reports the proportion of employees who were eligible for an employer sponsored pension, but were not participating in any type of private pension, by age and earnings quintile. This figure reveals a pattern of pension scheme participation that is consistent with the implications of the life-cycle model of behaviour, which suggests that savings motives are weakest at younger ages and among individuals with the lowest earnings. From quite high rates of non-participation among young employees, figure 2 indicates that private pension participation climbs with employee age for all earnings quintiles to age 40. Thereafter, there is some evidence of falling pension participation among individuals in the lowest earnings quintile, which is not evident among higher earning employee subgroups. By age 45, very few employees in the top earnings quintile who were eligible for an employer sponsored pension did not contribute to some form of private pension, in contrast to the bottom earnings quintile for which just over 30 per cent chose not to contribute.

The analysis assumes that rates of take-up of the NEST are consistent with the age and quintile specific statistics that are reported in figure 2. It is important to recognise, however, that participation in the NEST may depart from this assumption. One reason to suppose that this will be the case is that employees will be auto-enrolled into the NEST, which is not common among existing employer sponsored pensions in the UK. (12) International evidence suggests that auto-enrolment tends to result in higher rates of pension scheme participation, so that take-up rates may be higher in the NEST than currently observed among occupational pensions. (13) This proposition is also supported by evidence relating to the KiwiSaver in New Zealand, which is qualitatively similar to the NEST, and has received strong public support. (14) On the other hand, the NEST will require a smaller employer contribution than most employers currently make on behalf of their employees (where such contributions are made), which will depress incentives to participate.

Contribution rates

The analysis focuses upon the lower bounds imposed on contribution rates for participants in the NEST. This focus recognises that any additional pension contributions provided by employers (and employees) will be voluntary. The analysis consequently considers the case in which employers choose to minimise labour costs, given their employees' decisions to participate in the NEST. This seems a sensible starting point when considering the impact that the NEST is likely to have on the existing wage bargain, and upon the associated regulatory setting (including specification of the minimum wage).

Employer responses

Employer responses to the change in pensions policy are limited in the analysis to wage adjustments that offset the additional costs imposed by the NEST. These adjustments relate closely to considerations associated with setting of the minimum wage, and for low paid employees more generally. Two important issues do, however, stand out in this regard.

First, the analysis does not reflect the persistent downward trends in employer sponsored pension provisions that are discussed in Section 2, nor the bearing that the NEST may have on these trends. This issue, which is closely related to the assumed labour market context that is discussed above, is designed to focus upon short- to medium-term effects.

Second, no attempt is made to consider the potential effects of levelling down of existing employer pension contributions to the rates imposed under the NEST (identified in Section 2.2). Although levelling down has been singled out as an important risk associated with introduction of the NEST, (15) it is not considered here due to our focus upon low pay employees and their employers, and the associated implications for setting the minimum wage. As noted in Section 2.2, NEST will tend to extend employee eligibility to any employer pension contribution, rather than increase the contributions to which employees are eligible. Furthermore, the terms of employer sponsored pensions available to employees tend to vary more between employers than they do within employers. Hence, as low paid employees tend to work for employers that offer few pension benefits, levelling down is likely to have relatively little bearing upon their circumstances.

[FIGURE 2 OMITTED]

3.2. Projected effects (16)

The analysis suggests that 46 per cent of all employees, equal to 11 million people, will be eligible for higher employer pension contributions under the NEST, (17) with the vast majority of these employees currently not in receipt of any employer pension contribution. The distributional analysis bears out the suppositions made in Section 2.1; the NEST is likely to have a disproportionate effect on male employees in low pay industries, employed in small private sector firms. The principal statistics supporting these last four conclusions are reported graphically in figure 3.

Panel A of figure 3 indicates that the NEST will offer higher employer pension contributions than currently received by employees throughout the wage distribution, with men benefitting disproportionately relative to women. In contrast to the observation that more women than men currently benefit from employer pension contributions--reported in Section 2.1-98 per cent of all male employees relative to 91 per cent of female employees will be eligible for some employer pension contribution following introduction of the NEST. The disparity between the sexes will be highest--after controlling for income--among employees in the second wage quintile, who currently have relatively weak employer pension provisions and earn a sufficient wage to make them eligible to participate in the NEST.

Panel B of figure 3 reveals the impact that industry has on eligibility for higher employer pension contributions under the NEST. This panel indicates that eligibility will be higher for employees in low pay industries than nonlow pay industries throughout the wage distribution, but particularly so for those on high wages. This is due to the relatively thin employer pension provisions that are currently made in low pay industries throughout the wage distribution and the minimum wage threshold that is required to receive an employer pension contribution under the NEST.

Panel C of the figure indicates that the NEST has significant potential for closing the wide differences that currently exist between the prevalence of employer sponsored pensions in the private relative to the public sector, and in small firms relative to large firms.

Participation rates in the NEST, and employer contributions to the scheme, which were projected on the assumption of the participation observed for contemporary employer sponsored pensions are reported in figure 4.

Panel A of figure 4 indicates that adjusting for imperfect take-up has a more pronounced impact on participation rates among employees in the low pay industries, relative to the wider labour market. This is particularly the case in the low pay industries of hairdressing and employment agencies, where wages tend to be disproportionately low and there is a relatively high prevalence of younger workers. In the case of hairdressing, for example, participation in the NEST falls by 35 percentage points to 42 per cent, so that the proportion projected to participate in the scheme is almost halved as a result of the adjustment for imperfect take-up. Approximately two thirds of employees who are eligible for increased employer pension contributions under the NEST in the retail industry, the largest low pay industry by employee headcount, are projected to participate in the scheme. This is representative of the projected impact of employee take-up of the NEST in the low pay industries more generally.

[FIGURE 3 OMITTED]

Furthermore, even at the population aggregate level, imperfect take-up has an important bearing upon participation in the NEST; whereas the analysis indicates that 46 per cent of all employees are eligible for increased employer pension contributions under the NEST, only 33 per cent are projected here to participate in the scheme. This estimate suggests that 8.25 million individuals will participate in the NEST, which is towards the high end of current DWP (2010) projections. (18)

Panel B of figure 4 reveals the extent to which adjusting for imperfect take-up of the NEST reduces the value of employer contributions (per eligible employee) to the scheme. Comparing Panels A and B of the figure indicates that adjusting for imperfect take-up reduces the prevalence of participation in the NEST by a larger fraction than it does associated employer contributions, which is a logical implication of the higher rates of nonparticipation that are assumed among low paid workers (as reported in figure 2). On average, imperfect take-up is projected to reduce employer pension contributions to the NEST by just over one fifth, from an average of 8.21 [pounds sterling] per week per eligible employee to 6.46 [pounds sterling] per week per eligible employee. This last statistic implies that annual contributions to the NEST will be worth approximately 10 billion [pounds sterling] (in 2009 prices), which exceeds the DWP estimate by 1 billion [pounds sterling].

[FIGURE 4 OMITTED]

Beyond the population aggregates that are reported above, Panel B of figure 4 indicates two distinct differences between the employer contributions projected for low pay industries, relative to the wider labour market. First, the costs to employers of mandatory contributions to the NEST per eligible employee are appreciably lower in low pay industries than the wider labour market. And second, the differences between the mandatory employer contributions to the NEST in low pay and non-low pay industries are likely to widen in response to imperfect take-up of the scheme by eligible employees. Although the reduction in employer contributions among non-low pay industries varies within a fairly tight band about 20 per cent, among low pay industries it shows a relatively high degree of heterogeneity, varying between 20 and 40 per cent. This observation--and the coincident observation in relation to rates of participation underscores the uncertainty that is associated with the financial impact that the NEST will have within its target population.

Wage responses and implications for the minimum wage

What are the proportional wage reductions required to offset the increase in labour costs due to employer pension contributions identified under the NEST, subject to the lower bounds defined by the statutory minimum wage? Given the focus of the NEST on employees with modest incomes, the labour market is divided into the low pay industries that are considered in the preceding sections of this report. For non-low pay industries, the analysis distinguishes between employees in the public and private sectors, and also disaggregates private sector employees by firm size. Firm size is not taken into account for public sector workers due to their high degree of concentration in very large organisations. The results of this analysis are reported graphically in figure 5.

[FIGURE 5 OMITTED]

Panel A of figure 5 indicates that wages would have to fall by between 0.6 and 0.8 per cent on average to offset the additional employer pension contributions under the NEST. The wage offsets identified for low-wage industries tend to be higher than those in the wider labour market, with hairdressing representing a relatively extreme example. The sharp fall in the wage offset identified for hairdressing after accounting for imperfect take-up of the NEST- from 1.7 to 1.1 per cent--is due to the relatively low wages and young ages of employees in that industry. In contrast, the wage adjustments for hospitality, between 1.0 and 1.4 per cent, are at the higher end of the scale, while retail is slightly lower (between 0.9 and 1.2 per cent).

At the other extreme, a negligible wage adjustment to offset the NEST is identified for public sector workers, due to the prevalence and generosity of their existing employer pension contributions. Interestingly, some of the highest wage adjustments estimated for the entire labour market are among firms with fewer than 25 employees and outside of the low pay industries. This is due to the sparse pension provisions that these firms currently provide.

The wage offsets that are reported here have interest beyond simply revealing the wage pressure that the NEST can be expected to exert following its introduction. Like most pension schemes, the NEST will be a vehicle that subsidises saving, with the subsidies to participants paid for by non-participants. The distributional effects of the scheme will consequently depend crucially upon how the subsidies to participants are paid for, and the analysis that is reported here reveals the implications of one plausible alternative. In this context, the minimum wage plays a crucial distributional role by shielding the lowest paid workers--many of whom will not be eligible for the NEST from having to contribute toward the costs of employer contributions to the scheme.

An indication of the scale and scope of the potential role of the minimum wage to shield low paid workers from subsidising employer pension contributions consequent on the NEST is reported in Panel B of figure 5. This figure reveals that the minimum wage may have a very important bearing upon offsetting wage adjustments to the NEST in the low pay industries generally, and hairdressing, hospitality, and employment agencies in particular. The largest jump in the share of employees earning the minimum wage due to the wage offset to the NEST is reported for hospitality, followed very closely by hairdressing. In both of these low pay industries, existing employer pension provisions are relatively scarce and a large share of employees earn very low wage rates.

4. Conclusions

The introduction of the National Employment Savings Trust from 2012 will mark an important transition in the UK pensions system. For the first time, the NEST will require all employers to make a matching contribution in respect of eligible employees who choose to participate in a pension scheme. As up to 46 per cent of all employees could be eligible for increased employer pension contributions under the NEST, the scheme is likely to have an important bearing on aggregate labour costs, savings rates, and ultimately upon private provisions for retirement. The importance of these considerations is exaggerated by the current fragility of the UK economy, and the possibility that this fragility may extend well into the medium term.

As the NEST has not yet been introduced, the analysis reported in this paper uses reduced-form statistical relationships to project forward from the contemporary evidence base. This analysis focuses upon the implications of the NEST for employees and the wage bargain, over the short to medium term. The results obtained suggest that, consistent with the objectives of the scheme, the NEST will have the greatest impact on low pay industries and small private sector firms. This reflects the thin employer pension provisions that are currently afforded in these segments of the labour market. Furthermore, the analysis identifies substantial heterogeneity among low pay industries, with the NEST set to have a substantial impact on agricultural and employment agency industries, and a small impact on social care.

One of the key results of the analysis concerns the implications for low pay workers, who are least likely to participate in the NEST, but who are also employed in firms that will experience a disproportionate increase in costs due to the introduction of the scheme. This is because many firms in low pay industries currently do not provide pension contributions to any of their employees, including those on relatively high wages. A worrying implication is that affected employers may seek to offset the pension contributions that are mandated under the NEST by reducing wages to all of their employees, resulting in a cross-subsidisation from employees on the lowest wages to those who are better paid.

The results obtained highlight the countervailing considerations that introduction of the NEST is likely to have on setting of the minimum wage, and alternative aspects of the regulatory environment more generally. On the one hand, it is important not to lose sight of the fact that the NEST will imply higher labour costs, particularly for small private sector firms in low pay industries. For many of these firms, the minimum wage is likely to limit downward wage flexibility. But on the other hand, the minimum wage may play an important role in shielding the lowest paid workers from suffering as a result of the impact that the NEST has on the wage bargain. The current analysis highlights the potential importance of the minimum wage in balancing the needs of employers following introduction of the NEST against those of low paid employees.

doi: 10.1177/002795011221900108

Appendix A. Data Sources

There are a wide range of micro-data sources available for studying pension arrangements in the UK. This study relies predominantly upon data drawn from the Annual Survey of Hours and Earnings (ASHE) reported between 2005 and 2009. These data were selected because they provide a great deal of detail regarding employee industry of occupation, wages, and the nature of employer sponsored pension arrangements, including employer pension contributions. (19) Data from the Family Resources Survey were also considered for analysis, both to validate the representative nature of the ASHE data, and to supplement the temporal dimension of the data series reported by the ASHE.

Annual Survey of Hours and Earnings (ASHE)

The ASHE replaced the New Earnings Survey (NES) in 2004, but comparable data imputed from the NES are available from 1997. The survey reports data that are primarily based upon a 1 per cent sample of National Insurance numbers for February, supplemented by additional samples drawn from the Inland Revenue PAYE register in April, to cover employees that have either moved into the job market or changed jobs between the time of selection and the survey date.

The survey is addressed to employers, and provides information on the earnings and hours paid for employees within industries, occupations and regions. From 2005, the ASHE was augmented to include detailed information regarding the pension arrangements of employees, including the pension contributions of both employees and employers.

Family Resources Survey

The FRS was introduced in 1992, and reports data regarding the demographic, employment, income and financial circumstances of households in the United Kingdom (Great Britain prior to 2002). Data for the FRS are collected on a continuous basis from a cross-section sample of approximately 24,000 voluntary participating households that is designed to be representative of the wider United Kingdom population. The survey reports data regarding the nature of employment, including hours worked, occupation, industry and firm size. The survey also reports individual-level data regarding pension eligibility, actual membership, and the magnitude of private contributions made to pension schemes. Importantly, however, the survey does not report the size of any employer pension contribution. Furthermore, the FRS is ill-suited for considering low-paid workers because the Standard Industrial Classification is reported only to 2 digits, the Standard Occupational Classification is reported to 1 digit, and the survey is not designed to provide hourly wage rates.

REFERENCES

Brewer, M., Browne, J., Emmerson, C., Goodman, A., Muriel, A. and Tetlow, G. (2007), 'Pensioner poverty over the next decade: what role for tax and benefit reform?', IFS Commentary 103.

Choi, J.J., Laibson, D., Madrian, B.C. and Metrick, A. (2002), 'Defined contribution pensions: plan rules, paticipation choices, and the path of least resistance', Tax Policy and the Economy, 16, pp. 67- I 13.

DWP (2006), Personal Accounts: a new way to save, Norwich, The Stationary Office.

--(2006a), Security in Retirement: towards a new pensions system Norwich, The Stationary Office.

--(2009), 'Saving for retirement: implications of pensions reforms on financial incentives to save for retirement', DWP Research Report 558.

--(2010), Workplace Pension Reform Regulations Impact Assessment, 12 January.

Ernmerson, C. and Wakefield, M. (2009), 'Amounts and accounts: reforming private pension enrolment', IFS Commentary I I0.

Forth, J. and Stokes, L. (2008), 'Employers' pension provision survey 2007', Department for Work and Pensions Research Report 545.

Levell, P., May, R., O'Dea, C. and Phillips, D. (2009), 'A survey of the UK benefit system', Institute for Fiscal Studies Briefing Note, 13.

Madrian, B.C. and Shae, D.F. (2001), 'The power of suggestion: inertia in 401 (k) participation and savings behaviour', Quarterly Journal of Economics, 116, pp. 1149-87.

McKay, S. (2006), 'Employers' pension provision survey 2005', Department for Work and Pensions Research Report 329. OECD (2009), Pensions at a Glance 2009, Paris, OECD Publishing.

PPI (2007), 'Will personal accounts increase pension saving', Briefing Note 42.

Pensions Commission (2004), Pensions: Challenges and Choices, Norwich, The Stationary Office.

--(2005), A New Pension Settlement for the Twenty-First Century, Norwich, The Stationary Office.

van de Ven, J. and George, A. (2011), 'Financial implications of the National Employment Savings Trust', Report to the Low Pay Commission, www.lowpay.gov.uk/Iowpay/research/pdf/ NIESRFinal.pdf.

NOTES

(1) Previously known as Personal Accounts (DWP, 2006a), and before that the National Pension Savings Scheme (Pensions Commission, 2005).

(2) See OECD (2009) for details regarding the contemporary pension arrangements in OECD countries.

(3) DWP (2010), paragraphs 1.35, 2.3 and 2.4. Assumes 24.9 million employees, equal to the estimated employed population in the UK in 2009 (Office for National Statistics code MGRN). These figures update previous DWP (2006), p. 43, estimates that the NEST would serve between 6 and 10 million people, and would receive annual contributions worth 8 billion [pounds sterling] (2006 prices).

(4) Employees who have been auto-enrolled into NEST and those with qualifying earnings can choose to opt out of the scheme within one month of commencement, in which case their NEST account is closed and any contributions are returned to the respective payees. Thereafter, all members of NEST have a right to stop contributing to the scheme, but any accrued funds will remain in their NEST account.

(5) The interested reader is referred to Levellet al. (2009) for details regarding the wider state administered system of retirement provisions in the UK. See, also, Brewer et al. (2007) for an impact analysis of how the broader system of pension reforms will affect pensioner incomes, DWP (2009) for associated incentive effects, and the IFS review (http:// www.ifs.org.uk/projects/346) for further analysis.

(6) See, for example, the Pensions Commission (2004), Chapter 3.

(7) Statistics calculated on data from the ASHE and the FRS--see Appendix A for details.

(8) Just under half of all women are identified as working in a low-pay industry relative to less than 30 percent of men. And while the proportion of male employees in the top earnings quintile is just over twice that of women, the proportion of male employees in the bottom earnings quintile is less than a third that of women.

(9) These population average statistics are consistent with those calculated using the Employers' Pension Provision Survey reported by Forth and Stokes (2008).

(10) Hairdressing, employment agencies, and textiles stand out in this respect.

(11) An important caveat is that the base data do not describe labour market turnover. The analysis is consequently likely to overstate eligibility to payments under the NEST for employees who experience frequent labour market transitions, with associated implications for their respective employers.

(12) McKay (2006), for example, reports that 16 per cent of employees in 2005 were employed in firms that auto-enrolled employees into their principal work-based pension.

(13) See, for example, Madrian and Shae, 2001, and Choiet al., 2002, on US data, and McKay, 2006, for the UK.

(14) Total membership (net of opt-outs and closures) in the KiwiSaver was 1.8 million in September 2011, when the employed population comprised 2.2 million people.

(15) See, for example, Pension Reform, House of Commons Work and Pensions Committee, fourth report of session 2005-06, Volume II, p. Evl95.

(16) The interested reader is referred to van de Ven and George (2011) for further discussion.

(17) This result is consistent with DWP (2010, para 1.35) estimates.

(18) PPI (2007) also suggest that between 4 and 9 million new people will save in a work-based pension scheme when the reforms are introduced in 2012, contributing around 10 billion [pounds sterling] annually to the scheme.

(19) The British Household Panel Survey has a sample size that is appreciably smaller than that of either the ASHE or the FRS, making it of limited use here. The Expenditure and Food Survey identifies employee contributions to pension schemes but does not establish the nature of these schemes, whether employers contribute, or employee occupational classifications. The Labour Force Survey does not provide information on pension provision by employers.

Justin van de Ven, National Institute of Economic and Social Research. e-mail: jvandeven@niesr.ac.uk. I am grateful to the Low Pay Commission for financial support to conduct this study. I received useful comments regarding the research from attendees at the Low Pay Commission research Workshop in October 2010, and from Gerry Franks, Tim Butcher and two anonymous referees on earlier drafts. Anitha George conducted preliminary analysis in relation to the work presented here.

Table 1. The prevalence of pensions attracting an employer
contribution, by employee subgroup (cell %)

 Earnings quintile

Sex Employment characteristics Lowest 2 3 4

Men Low pay 10 21 28 36

 hairdressing 0 3 0 5
 employment agencies 1 3 4 7
 hospitality 2 5 8 13
 agriculture 8 15 25 26
 cleaning 11 29 31 27
 textiles and clothing 14 38 38 40
 retail 11 18 25 31
 leisure / travel / sport 6 20 24 31
 childcare 18 34 48 41
 office work 13 23 30 31
 food processing 14 28 44 51
 security 16 29 34 26
 social care 22 44 56 71

Women Low pay 24 33 39 45

 hairdressing 3 4 4 3
 employment agencies 3 4 7 15
 hospitality 15 15 11 13
 agriculture 6 16 22 28
 cleaning 20 26 23 22
 textiles and clothing 16 27 36 24
 retail 20 24 27 27
 leisure / travel / sport 12 18 21 29
 childcare 32 41 46 31
 office work 39 43 41 23
 food processing 22 33 42 43
 security 25 44 53 43
 social care 34 52 62 71

 Low pay 21 28 33 40

Men Non-low pay 22 38 53 61
Women Non-low pay 51 60 66 73

 Non-low pay 44 49 58 65
Men 14 31 46 56
Women 32 46 56 65

 Public sector 65 81 86 87
 Private sector 15 25 38 48

 1-9 employees 4 7 10 13
 10-24 employees 12 17 20 25
 25-249 employees 13 25 34 42
 250-499 employees 20 35 48 57
 500+ employees 37 50 63 72

Total population 28 39 50 60

Sex Employment characteristics Highest Average

Men Low pay 48 26

 hairdressing 0 1
 employment agencies 20 6
 hospitality 26 8
 agriculture 35 19
 cleaning 32 23
 textiles and clothing 51 37
 retail 39 23
 leisure / travel / sport 34 21
 childcare 76 34
 office work 21 23
 food processing 60 38
 security 30 29
 social care 85 58

Women Low pay 50 32

 hairdressing 1 3
 employment agencies 15 7
 hospitality 20 14
 agriculture 27 14
 cleaning 25 22
 textiles and clothing 21 24
 retail 30 23
 leisure / travel / sport 33 18
 childcare 24 36
 office work 14 39
 food processing 61 35
 security 57 42
 social care 77 53

 Low pay 49 29

Men Non-low pay 69 56
Women Non-low pay 75 65

 Non-low pay 70 60
Men 66 48
Women 70 50

 Public sector 87 81
 Private sector 60 37

 1-9 employees 20 9
 10-24 employees 34 21
 25-249 employees 53 34
 250-499 employees 66 47
 500+ employees 76 60

Total population 67 49

Source: Author's calculations on data from ASHE data from 2005/6 to
2009/10--see Appendix A for details.

Notes: Earnings quintile based on weekly wage excluding overtime,
shift and premium payments. Low pay industries as defined by the Low
Pay Commission.