Working past 65 in the UK and the USA: segregation into ‘Lopaq’ occupations?

Abstract

A prominent business case for employing older people in the 2000s suggests diverse employment opportunities existed for Britons over 65, despite their limited employment rights. However, it is hypothesized that employees over 65 were disproportionately segregated into less desired ‘Lopaq’ occupations: these were low paid, required few qualifications and were often part-time. The UK is contrasted with the USA, a country with long-established age discrimination legislation; the Labour Force Survey and Current Population Survey are analysed. A greater UK concentration in Lopaq occupations suggests employers, working in a context of limited employee rights, selectively retained and recruited people in their 60s to these jobs. An alternative explanation, that Lopaq employment levels reflected the characteristics of those choosing to work, is unsupported by logistic regression analysis. US evidence suggests that the 2011 default retirement age abolition will weaken UK Lopaq occupational segregation after 65 more than voluntaristic commitments to ‘age-diversity’.

Keywords

age and employment age discrimination age-diversity equality and diversity labour market segmentation older workers post-retirement work working past 65

Introduction

Since the 1990s a business case for employing older people has assumed prominence, linked to broader arguments regarding ‘managing diversity’ within organizations. Farnham (2010: 195) argues:

In organizations, the emphasis now is on managing diversity (or valuing differences between people) rather than on equal opportunities […] to harness the advantages to be gained from hiring a diverse workforce […] not just those covered by anti-discrimination legislation.

From a ‘business case’ perspective, employers should realize the benefits of having an ‘age-diverse’ workforce that draws on older people’s wide range of experiences alongside the positive attributes of younger people (Trapp, 2004). ‘Age-diversity’ prominence is evident in the number of high-profile employers joining organizations like the Employers Forum on Age,¹ and its extensive discussion in practitioner publications (for example, Trapp, 2004) and the wider press (Freudenheim, 2005; Peters, 2006). Recognition of the benefits of diverse workforces, rather than legislation, is therefore seen by some as the central driver for integrating under-represented groups (Noon, 2007).

Because Britons over 65 have historically had few employment rights they provide an interesting test-case for the ability of the business case to integrate older people into the labour market. Clearly there has been a willingness to work among a significant minority (Smith, 2000), UK policy encourages delayed retirement, and there will be increasing need to work given state pension age increases above 65 from 2018. However, employers have been free to decide whether they will allow continued employment past 65, and, if so, who they grant this permission to. The Employment Equality (Age) Regulations 2006 gave the right to request continued employment past 65. However, managerial discretion over this decision has remained until 2011.

Some writers have pessimistically argued that a business case for diversity is insufficient and emphasized the importance of legal rights (Kirton and Greene, 2010; Noon and Ogbonna, 2001). What if a business case for employing a certain group is weak, or managers are unable/unwilling to recognize or act when it is strong (Noon, 2007)? Kirton and Greene (2010: 285) have argued that ‘the diversity approach […] needs to be underpinned by increased legal protection and rights [if] the diversity of the labour market is to be reflected in organizational workforces to begin with.’

Loretto and White’s (2006: 319) interviews with Scottish employers revealed that although ‘employers expressed enthusiasm for “an age-diverse” workforce in abstract terms, there was extensive evidence that younger “prime-age” workers were the preferred workforce’. Where older people had been employed this was often because of difficulties recruiting core-age workers. This suggests that job opportunities open to over-65s may be the disproportionately less desirable employment rejected by core-age workers. Writing in the mid-1990s Taylor and Walker (1994: 588) argued that:

The labour market [may] become increasingly age-segmented with available employment for older workers generally consisting of low-skill, low-responsibility and low-paid work with fewer opportunities for career advancement.

This article therefore examines whether over-65s have faced occupational segregation, i.e. being ‘more likely to hold lower status, lower paid jobs with less chance of promotion’ (Kirton and Greene, 2010: 55). Consequently, ‘Lopaq’ occupations are focused on, a categorization developed here for low paid occupations requiring few qualifications and that are often part-time. Previous research focuses predominantly on supply-side factors influencing employment past state pension age (see Lain, 2011); studies have not explored the occupations performed by men and women over 65.²

In the coming years the influence of legislation on occupational integration is likely to become clearer, given the 2011 default retirement age abolition. This extends employment rights to over-65s, thereby outlawing automatic forced retirement at 65. This article examines legislative influence by comparing the UK with the USA, a country where over-65s have long been protected by age discrimination legislation.

The article has three aims:

To outline the rationale for employers segregating over-65s into Lopaq occupations;

To explore the extent to which weaker segregation in the USA is influenced by age discrimination legislation; and

To examine alternative explanations for strong UK occupational segregation, including the characteristics of those choosing to work and local labour market conditions.

Theorizing segregation into Lopaq occupations past 65

Queuing for jobs

To theorize why a concentration in low paid, Lopaq occupations may occur after 65 Reskin and Roos’s (1990) model of occupational stratification is drawn upon. They argue that ‘occupational composition [is] the result of a dual queuing process: labour queues order groups of workers in terms of their attractiveness to employers, and job queues rank jobs in terms of their attractiveness to workers’ (1990: 29). Regarding job queues, ‘most workers try to maximize income, social standing, autonomy, job security, congenial working conditions, interesting work, and chance for advancement; and they rank occupations accordingly’ (1990: 38). Most employees would rank low paid Lopaq jobs towards the end of the job queue.

Little evidence exists that older people differ from the general population by ranking low paid work requiring few formally recognized skills highly (Christensen, 1990; Flynn and McNair, 2009). Although some might value a reduction in working hours (Vickerstaff et al., 2008), occupational preferences appear to be fairly stable as people reach older age (Christensen, 1990; Gobeski and Beehr, 2009). However, research suggests employers rank older people lower down the labour queue than core-age workers. Regression analysis by Aaronson and Housinger (1999) found involuntarily displaced older US workers were less likely to be recruited to new jobs. This mirrors Loretto and White’s (2006) finding that employers look to older workers when core-age workers are unavailable, and research suggesting older people remain unemployed longer (Gangl, 2003: 166–8).

A particular reluctance to employ older workers is expected for jobs incurring significant training/experience costs, as employers want to maximize investment returns and usually expect shorter employment from older workers. UK and US recruitment evidence appears to confirm this (Daniel and Heywood, 2007; Hirsh et al., 2000). Core-age workers are likely to be attracted to these jobs, because they offer career possibilities.

This restricts job opportunities for older people disproportionately to lower paid, Lopaq occupations requiring few formally recognized skills. Table 1 identifies these as lower-level sales, service and elementary occupations based on their average weekly wages and the qualifications held by those doing the jobs (see Data and methods section).³ Pay is particularly low because these jobs are disproportionately part-time (see Table 1), they account for half of all part-time jobs in the UK and two-fifths in the USA. High part-time rates reflect the temporal needs of those they serve: shop sales persons and cashiers, waiters, bar tenders and food preparers often work part-time to cover peak periods of customer demand; cleaners to ensure work-places are functional before the working day; and domestic helpers and personal health care workers (‘home help’) to cover peak periods of the day, getting clients up, lunch time and bed time (Patmore, 2003: 15). For Lopaq occupations likely to have fewer part-timers, such as building supervisors, unsocial hours of employment may nevertheless be common. Within retail and hospitality there are likely to be better paid full-time managers in addition to Lopaq workers, emphasizing the need to focus on occupation.

Table 1.

Characteristics of different occupations: UK and USA

		UK			USA
		Weekly wage (£)	Education (ISCED)	Part-time rate (%)	Weekly wage ($)	Education (ISCED)	Part-time rate (%)
	Managers	528.89	1.65	6.2	925.13	1.50	3.66
	Professionals	472.82	1.21	14.1	888.81	1.17	9.62
	Technicians/Assoc. Prof.	378.64	1.66	17.5	652.92	1.59	12.66
	Craft	326.17	2.06	4.3	633.32	2.04	2.76
	Operatives	293.63	2.18	6.5	517.38	2.12	5.57
	Clerks	224.88	1.95	28.7	405.19	1.89	20.69
Lopaq Occupations	Service/Sales	157.39	2.07	53.2	379.89	2.05	26.78
	Elementary	163.48	2.32	43.1	337.76	2.24	20.95

Notes: For data and occupational classification information see the Data and methods section.

The ISCED measure ranks from 1 (high education) to 3 (low education).

Part-time work is classified as under 30 hours a week.

The Table covers employees aged 16 and over.

Source: Author’s analysis of US CPS 2000 and UK LFS 1999/2000

Lopaq jobs are not necessarily precarious ‘secondary’ employment (Atkinson and Meager, 1986), for example, retailers might want to reduce shop-worker turnover if demand fluctuates over the course of a week. Nor are they necessarily ‘low skilled’, personal health care work at this level, for example, will require many tacit skills. The point is that these have traditionally been insufficiently recognized through qualifications or training. Likewise, experiential skills are likely to be highly prized by hardware stores, although financially under-recognized.

Very low weekly wages, and often unsociable hours of employment, mean employers’ preferred choice, core-age men, will not be in the queue for these jobs, employers will therefore use young people, including students, to fill many Lopaq jobs. Young people are arguably excluded from the desired core-age workforce because employers in general value experience and qualifications, hence the common assumption that the core-age workforce starts at 25 (e.g. Luffman, 2006). Given a choice between the youngest and oldest workers, anecdotal evidence suggests employers in sectors such as retail often prefer older people for Lopaq jobs if they can be attracted. This is not always the case, and Warhurst and Nickson (2007) demonstrate that students’ desirability is increased by class and aesthetic considerations, particularly at the more fashionable end of retail/hospitality. Nevertheless, employers often say they prefer older workers for being harder working, more honest, mature and polite with customers (Macnicol, 2006: 110–11), probably because they are visualized in Lopaq occupations alongside the very young.

Research on the job seeking behaviour of those at or around retirement age is almost non-existent (Adams and Rau, 2004).⁴ However, those wanting to work may seek employment in Lopaq occupations because more desirable jobs are taken by core-age workers. Workers beyond ‘retirement age’ are not those we would necessarily associate with Lopaq jobs, being better educated and healthier than non-workers on average (Haider and Loughran, 2001; Hayward et al., 1994; McNamara and Williamson, 2004; Parnes and Sommers, 1994; Smeaton and McKay, 2003) and less likely to be in the poorest wealth quintile (Lain, 2011). Many older people in such a position would be unwilling to take a Lopaq job, given working identities linked to previous higher status jobs (Riach and Loretto, 2009: 115). However, the oldest workers are a minority who often report strong work orientations (Barnes et al., 2004; Parnes and Sommers, 1994) and may therefore accept Lopaq employment. Older people with a pension, a paid mortgage and no dependents may also find it easier to accept such low earnings, as will young employees supported by parents or educational payments. In both countries a disproportionate concentration in Lopaq occupations after 65 is therefore hypothesized.

Is occupational segregation weaker in the USA?

While a concentration in Lopaq occupations after 65 is anticipated in both countries, this is expected to be weaker in the USA than the UK. This is not because Americans in higher-level jobs had greater financial need to work than Britons. In both countries individuals had access to a ‘normal’ state pension at, or by, age 65 in the early 2000s (the period covered here).⁵ Some had greater medical expenses because US Medicare health insurance, while near universal after 65, excluded prescription charges (Salanicoff et al., 2009). However, counteracting this, Americans past 65 were wealthier on average than Britons (Disney and Whitehouse, 2002: 3). Nor were Americans more intrinsically work-orientated: Smith (2000) found two-fifths of over-65s in both countries reported a preference for employment irrespective of financial need. Instead, it is hypothesized that age discrimination legislation reduced segregation into Lopaq jobs in the USA (relative to the UK). Legislation is expected to have a strong influence on employer treatment of those continuing in their jobs in older age (retention), rather than those seeking employment (recruitment).

Continued employment in older age

Until October 2006 there was no UK age discrimination legislation, and basic employment rights, such as the right to claim unfair dismissal or redundancy payments, ended at 65. Qualitative research suggests that line managers often decided who continued working past 65, based on departmental needs (Vickerstaff, 2006). Employers used ‘normal’ retirement ages to sift out staff they had less demand for, making exceptions for desired workers. The Employment Equality (Age) Regulations 2006 gave employees the right to request continued employment past 65, but did not remove the managerial focus of these decisions (Flynn, 2010). If older people were in greater demand for Lopaq occupations this should be reflected in the retention of older workers in the UK in the 2000s.

In contrast, the US Age Discrimination in Employment Act (ADEA) has covered over-65s since 1978 (with the upper age limit of 70 abolished in 1986) (Neumark and Stock, 1999). By international standards, the rules governing general dismissal of US employees are not strict (OECD, 1999). However, employers are forbidden from getting rid of staff on the basis of their age and mandatory retirement ages within employment contracts are unlawful (Neumark, 2009). The legislation covers almost all occupations, with a very small number of public safety exemptions – ‘basically, airline pilots, fire-fighters and law enforcement officers’ (Macnicol, 2006: 237).⁶

As with any legislation the ADEA will not entirely remove work-place discrimination (Neumark, 2009). However, research suggests that ‘age discrimination legislation has succeeded at boosting the employment of older individuals through allowing them to remain in the workforce longer’ (Adams, 2004: 240; see also Issacharoff and Harris, 1997; Neumark and Stock, 1999). This conclusion is reached by comparing employment between states with and without legislation prior to ADEA reforms. US age discrimination legislation also forces employers to continue contributing to an employee’s occupational pension if they work after state pension age (Quadagno and Hardy, 1991: 473). This removes another means of inducing retirement for those typically in higher-level jobs with defined benefit occupational pensions. Overall, then, US employers are expected to be less able to constrain opportunities to continue working past 65 to those in Lopaq jobs.

Recruitment in older age

While US discrimination legislation may enhance the ability of people across occupations to remain in their jobs, research has not demonstrated a positive recruitment effect (Adams, 2004; Issacharoff and Harris, 1997; Neumark and Stock, 1999). This is perhaps because recruitment is less transparent than managing employee exit (Issacharoff and Harris, 1997: 795). Most US age-related discrimination cases relate to dismissal, not recruitment (Neumark, 2009). Adams (2004: 236) even finds some evidence from the 1960s and 1970s that state level age discrimination legislation had a negative impact on the recruitment of over-65s. Heywood and Siebert (2008) argue that discrimination legislation made employers wary of recruiting older people, although the longer-term impact of the ADEA is unclear. In any case, US employers, as with their UK counterparts, have a great deal of discretion over the recruitment process. Consequently, employees recruited in older age are expected to be disproportionately in Lopaq occupations in both countries in the 2000s.

In sum, a disproportionate concentration in Lopaq jobs after 65 in both countries is expected, but this should be weaker in the USA than in the UK. This is because of Americans across occupations opting to continue in their jobs, rather than employers recruiting older people to a wide range of jobs.

Data and methods

To understand whether older people are disproportionately recruited to, or retained in, Lopaq jobs it is important to know about the jobs held at an earlier period. Using panel surveys to compare occupational transitions between countries is not possible. The British Household Panel Survey includes insufficient employees over 65, fewer than 130 in 2000 (see Smeaton and McKay, 2003: 39). The English Longitudinal Study of Ageing opens up some future possibilities, containing 288 employees over 65 in the first 2002 survey,⁷ with occupational data available at the detailed four-digit level. However, it was not possible to recode this and the equivalent US Health and Retirement Study (HRS) into a meaningful, comparable occupational classification across countries. The HRS provides occupational data at a two-digit level, and some of the 17 occupational categories arguably reflect sector more than occupation.

Instead a ‘pseudo-cohort’ approach is used to reveal aggregate shifts in the composition of employment for a single age cohort. The occupations of employees aged 58–9 in 1992 are compared with those aged 65–9 in 2000. This approach means the large, representative occupationally detailed UK Labour Force Survey (LFS) and US Current Population Surveys (CPS) can be used. The period 1992 to 2000 is the longest spell of consistent UK occupational data, and covers a period before the US state pension age rose above 65. However, the restricted data period means the samples overlap in terms of age cohorts, rather than matching completely.

Because the same people are not followed over time it is impossible to conclusively determine whether individuals make transitions from other occupations into Lopaq employment because of age. Nevertheless, logistic regression analysis is used to control for a range of disadvantage-based differences between the samples that might explain shifts towards Lopaq jobs. Furthermore, recruitment and retention in older age is examined using surveys for 2000 that include job tenure information: the February CPS and the October-December LFS (combined with the July-September 1999 survey to increase the sample). For 1992 the April-June LFS and the January CPS are used. For each country these combined files yield samples of around 2000 individuals aged 58–9/65–9. The surveys from 2000 are also used to present occupational profiles for all age groups (supplemented by LFS data from January-March 2009). Occupational variables were recoded into the International Standard Classification of Occupations (ISCO)’s ISCO-88 classification using Ganzeboom’s US recode (Ganzeboom and Treiman, 1996) and UK syntax produced by the Occupational Information Unit of the Office for National Statistics.⁸

Results

Comparing employment across age groups

Table 2 presents occupational employment across all age groups in 2000, as a prelude to the narrower but more refined pseudo-cohort analysis of people aged 58+. The results imply an increasing segregation into Lopaq jobs after 65, with the trend strongest in the UK as expected. In both countries a similar occupational pattern prior to 65 is evident: two-fifths were in Lopaq jobs at age 16–24, falling to around a fifth for core-age workers aged 25–49. By age 65–9 the proportion in Lopaq occupations had grown in both countries, although to a lesser degree in the USA. Two-fifths of 65–9-year-olds were in Lopaq jobs in the UK, the same share as at 16–24, compared with just over a quarter in the USA.

Table 2.

Occupational profile of employees by age

USA 2000	16-24	25-49	50-54	58-9	60-64	65-9	70+	Total
Managers	5.5	15.2	16.9	15.2	13.5	11.0	11.5	13.6
Professionals	7.6	18.7	20.6	19.2	17.5	16.6	13.0	17.0
Technicians / Assoc. Professional	10.3	12.7	11.3	11.2	9.6	10.5	8.4	11.9
Crafts	8.9	12.0	10.2	10.3	8.8	6.1	5.2	10.9
Operatives	7.3	10.2	9.3	10.6	12.5	8.6	9.9	9.7
Clerks	21.3	13.3	14.5	13.5	16.3	19.1	17.0	14.9
Lopaq	39.1	18.0	17.2	19.9	21.8	28.1	35.0	21.9
Total per cent	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0
UK 2000	16-24	25-49	50-54	58-9	60-64	65-9	70+	Total
Managers	5.7	16.7	16.0	13.9	12.0	11.4	12.8	14.5
Professionals	6.5	18.4	17.5	14.2	11.5	11.4	9.7	15.8
Technicians / Assoc. Professional	7.3	9.3	7.9	7.5	7.3	5.5	4.6	8.6
Crafts	10.5	10.2	9.7	10.6	12.8	7.0	6.3	10.3
Operatives	5.4	7.9	9.0	10.4	11.4	7.9	5.9	7.9
Clerks	22.1	17.3	18.2	18.8	15.5	14.5	15.3	18.1
Lopaq	42.5	20.2	21.7	24.5	29.4	42.3	45.3	24.7
Total per cent	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0
UK 2009	16-24	25-49	50-54	58-9	60-64	65-9	70+	Total
Managers	3.8	17.7	17.3	14.3	12.9	12.5	10.9	15.2
Professionals	5.3	16.3	15.0	16.0	11.7	12.3	7.6	14.3
Technicians / Assoc. Professional	8.9	14.2	11.5	11.2	9.5	6.8	9.5	12.5
Crafts	9.6	6.4	7.2	7.8	8.6	5.5	4.4	7.1
Operatives	3.2	6.1	7.1	7.5	10.0	8.7	9.5	6.2
Clerks	15.1	14.0	15.5	16.4	15.8	15.4	15.3	14.7
Lopaq	54.2	25.3	26.5	26.8	31.7	38.7	42.9	29.9
Total per cent	100.0	100.0	100.0	100.0	100.0	100.0	100.0	100.0

Source: Author’s analysis of US CPS 2000 and UK LFS 1999/2000 and 2009

The greater concentration in Lopaq occupations in the UK appears to result from employers restricting access to other occupations, rather than employing large numbers of Lopaq workers. For the whole theoretical employee workforce aged 65–9 (including the economically inactive) similar proportions worked in Lopaq occupations in the UK and the USA (2.9% versus 4.3%; analysis not shown). For other occupational groups the employment rates in the USA were much higher: 11.3 per cent compared with only 4 per cent in the UK.

The final panel of Table 2 shows that two-fifths of UK employees aged 65–9 were in Lopaq jobs in 2009, suggesting little has changed. It should be noted that the overall share of Lopaq employment is reported to have risen over the period. This, in part, reflects a change in the occupational classification used in the survey, from which the ISCO-88 schema was recoded. The 2009 results therefore slightly over-estimate Lopaq jobs relative to 2000. Nevertheless, a strongly disproportionate concentration in Lopaq occupations at 65–9 is evident in the UK in 2009.

Comparing pseudo-cohort employment over time

To explore whether high Lopaq employment results from recruitment and retention, rather than cohort effects, the pseudo-cohort analysis is now shown. Table 3 gives the percentage of employees in each occupation at ages 58–9 and 65–9 (in bold); the columns between these break down the results for employees aged 65–69 into ‘continuers’ and ‘recruits’. ‘Continuers’ are in the same job as at age 58–9, whereas ‘recruits’ were taken on in their 60s. In both countries ‘recruits’ represented around 45 per cent of employees aged 65–9, and ‘continuers’ 55 per cent.

Table 3

Occupational breakdown of employees at ages 58–9 and 65–9 in the UK and the USA (percentages)

UK	58–9(1992)	65–9 (2000)			Percentage point shift 58–9 to 65–9	Shift attributable to overall occupational change
		Continuers	Recruits	All 65–9
Managers	12.2	17.5**	5.5*	11.5	−0.7	1.1
Professionals	14.2	12.7	10.0	11.3	−2.9	0.6
Technicians / Assoc. Professional	7.2	5.2	5.8	5.5	−1.7	0.7
Crafts	12.5	8.1*	6.1***	7.1 ***	−5.4	−2.0
Operatives	9.0	5.0**	10.2	7.6	−1.4	−0.8
Clerks	16.9	15.4	13.7	14.6	−2.4	−0.8
Lopaq occupations	27.9	36.0***	48.7***	42.4 ***	14.4	1.3
Total per cent	100	100	100	100
Base n	1367	430	432	862
US	58-9(1992)	65-9 (2000)			Percentage point shift 58-9 to 65-9	Shift attributable to overall occupational change
		Continuers	Recruits	All 65-9
Managers	11.6	13.3	8.5	11.0	−0.6	2.0
Professionals	15.1	20.5*	13.6	16.9	2.2	1.8
Technicians / Assoc Professional	11.9	10.1	10.3	9.9	−1.7	0.2
Crafts	13.3	6.9***	5.7***	6.2 ***	−6.9	−0.3
Operatives	12.2	8.8	7.9*	8.6 **	−3.9	−1.2
Clerks	14.8	17.0	21.5**	19.4 *	4.3	−1.9
Lopaq occs	21.1	23.4	32.6***	28.0 **	6.6	−0.4
Total per cent	100	100	100	100
Base n	1221	376	331	707

Note: *** = p<0.001, ** = p<0.01, * = p<0.05 (difference from percentage at 58–9)

Source: Author’s analysis of US CPS 1992 & 2000 and UK LFS 1992 & 1999/2000

Table 3 shows statistically significant shifts towards employment in Lopaq occupations after 65 in both countries, with a bigger UK shift as anticipated. In the UK 27.9 per cent of employees were in Lopaq occupations at age 58–9, compared with 42.4 per cent at age 65–9. In the USA 21.1 per cent were employed in Lopaq occupations at 58–9, and 28 per cent at ages 65–9.

UK employers appear to have selectively retained workers in Lopaq jobs, as the percentage of ‘continuers’ in Lopaq occupations, 36 per cent, was considerably above that of the cohort at age 58–9 (27.9%). In contrast, US employers appear less able to focus on retaining Lopaq workers, as continuers had a similar likelihood of being in a Lopaq occupation to the cohort at 58–9 years (23.4% compared with 21.1%).

Interestingly, US employers seem to be more willing than the British to recruit older people to ‘non-Lopaq’ occupations. The share of recruits in Lopaq jobs in the UK was higher, at 48.7 per cent, than in the USA (32.6%). That two-thirds of US recruits were not in Lopaq jobs perhaps suggests that discrimination legislation has normalized employment of older people to a degree. This is discussed in the conclusion.

Accounting for disadvantage and other explanations: logistic regression analysis

While UK employers appear to restrict employment opportunities to older people, alternative explanations for higher levels of Lopaq employment need to be explored. From the supply side, middle-class Americans may have been more inclined to work because defined benefit pensions were less prevalent than in Britain. These pensions are said to discourage employment (Meadows, 2003), which may have contributed to US employment in higher-level jobs. However, US age discrimination legislation will have reduced this effect by compelling employers to continue pension contributions for workers over 65. Indeed, there was no statistically significant shift in the proportion working in the US public sector between 58–9 and 65–9⁹ despite the dominance of defined benefit pensions.¹⁰

Demand-side explanations are also limited. Lopaq jobs grew in number between 1992 and 2000. However, only 1.3 of the 14.4 percentage point shift to Lopaq jobs after 65 is attributable to occupational growth (see the final column of Table 3). Reluctance to employ over-65s in anything other than Lopaq occupations might alternatively be explained if unemployment were high in 2000, but at 5.5 per cent UK unemployment was comparatively modest (albeit above the USA’s 4%).¹¹ A legacy of higher 1990s unemployment may have discouraged recruitment/retention of the over-60s (Lain, 2011). However, as was shown above, disproportionate Lopaq employment past 65 remained until the end of the 2000s, despite similarly favourable labour market conditions to those of the USA for much of the period.¹²

An alternative explanation is that UK employees over 65 were disproportionately in Lopaq occupations because of the characteristics of those choosing to remain in employment. If those delaying retirement did so for financial reasons, the result of disadvantaged employment histories, it would be unrealistic to expect employment in anything other than Lopaq jobs. This would be surprising, given research suggesting workers over 65 were not a particularly disadvantaged group in general relative to retirees (Lain, 2011). However, disadvantage is theoretically possible because a different sample to Lain (2011) is used here (excluding the self-employed, over-70s and non-workers). The following logistic regression analysis therefore compares the probability of being in a Lopaq occupation at ages 58–9 and 65–9, controlling for differences between the employee populations that might explain the higher rate of Lopaq employment after 65.

Unfortunately, the surveys have limited information on wealth or previous occupations to explore disadvantage. Instead, the following ‘control’ regression variables are used:

Education, because over-65s are more likely to be in Lopaq occupations if they have low qualifications;

Home ownership, because renting is likely to reflect a history of employment disadvantage;

Ethnicity, because on average belonging to a minority ethnic group is expected to increase the likelihood of prior labour market disadvantage;¹³ and

Marital status and gender, because women who are or have been married may have disrupted career histories constraining later employment opportunities.

Regional unemployment levels are also controlled for, in case workers over 65 were in areas with high unemployment restricting job opportunities. Finally, the regional age profile is controlled for, because employers may look more favourably on older employees when the core-age population (i.e. the proportion of the population aged 25–54) is comparatively small.

The regression analysis is presented in Table 4. The dependent variable is whether an employee is in a Lopaq occupation (coded to 1) or not (coded to 0). The odds ratios represent the likelihood of someone with a particular characteristic being in a Lopaq occupation. The regression models include employees aged 58–9 alongside recruits aged 65–9 in Model 1, continuers aged 65–9 in model 2, and all employees aged 65–9 in Model 3.

Table 4.

Logistic regression analysis: influence of age and other factors on working in a Lopaq occupation (odds ratios)

	UK			USA
	Model 1: Continuers 65–9 vs 58–9	Model 2: Recruits 65–9 vs 58–9	Model 3: All 65–9 vs 58–9	Model 1: Continuers 65–9 vs 58–9	Model 2: Recruits 65–9 vs 58–9	Model 3: All 65–9 vs 58–9
Age 58–9 (ref)	1.00	1.00	1.00	1.00	1.00	1.00
65–9	2.32**	4.67***	3.38***	1.16	1.80**	1.40
High education (ref)	1.00	1.00	1.00	1.00	1.00	1.00
Medium education	4.20***	5.20***	4.59***	4.79***	4.33***	4.21***
Low education	11.75***	13.39***	11.92***	11.13***	8.57***	8.71***
House owner (ref)	1.00	1.00	1.00	1.00	1.00	1.00
Renter	1.64***	1.74***	1.69***	1.86***	1.63**	1.82***
Ethnicity: white (ref)	1.00	1.00	1.00	1.00	1.00	1.00
Ethnicity: non-white	1.95*	1.36	1.33	1.81***	1.75**	1.79***
Male (ref)	1.00	1.00	1.00	1.00	1.00	1.00
Female	4.72***	4.55***	3.99***	1.62**	1.51**	1.48**
Married (ref)	1.00	1.00	1.00	1.00	1.00	1.00
Single	1.12	1.15	1.36	0.57	1.65	1.38
Divorced	1.85	1.16	1.56	0.70	0.63	0.62
Widowed	1.40	2.35*	2.16*	0.60	1.03	1.01
Female × single	0.43	0.43	0.35*	1.75	0.56	0.71
Female × divorced	0.44	0.69	0.58	1.42	1.61	1.45
Female × widowed	0.67	0.30**	0.40*	1.45	0.84	0.84
Local unemployment	1.04	1.03	1.04	0.96	0.96	0.95
Local age profile	0.90*	0.93	0.91**	0.99	0.97	0.97
Constant	3.10	0.73	2.20	0.10	0.36	0.47
Nagelkerke r2	0.31	0.32	0.30	0.18	0.16	0.16
−2 Log likelihood	1747.02	1799.16	2285.37	1467.39	1518.39	1880.96
Base (n)	1792	1794	2220	1597	1552	1928

Note: *** = p<0.001, ** = p<0.01, * = p<0.05

Source: Author’s analysis of US CPS 1992 & 2000 and UK LFS 1992 & 1999/2000

For the UK, education, home ownership, ethnicity and gender (but not unemployment or the local age profile) had the expected influences on being in a Lopaq occupation in Models 1–3. Gender had a strong effect with widowhood further increasing female Lopaq employment. After controlling for these factors, Model 1 shows that a continuer aged 65–9 was 2.3 times more likely to be in a Lopaq occupation than an employee aged 58–9. This suggests continuers over 65 were not disproportionately in Lopaq jobs because they were a disadvantaged subset of the cohort at 58–9. If this were the case controls for individual characteristics would be expected to reduce, not increase, the likelihood of being in a Lopaq job at age 65–9.

In Model 2 age has an even stronger influence on being in a Lopaq job: 65–9-year-old recruits were almost five times more likely to be in a Lopaq job than 58–9-year-olds. This big odds ratio increase for 65–9-year-olds suggests recruits had more advantaged personal characteristics than employees at age 58–9. That they entered Lopaq jobs in comparatively large numbers is consistent with occupational downgrading (although this is not proven). Downgrading would partly explain why Lopaq employment increases after 65 while, according to Lain (2011), the poorest are least likely to work.

It should be noted that odds ratios for variables such as education are likely to be inflated to a degree because they will pick up the effect of missing variables also associated with disadvantage (such as wealth, pensions and work attitudes). Unfortunately it proved impossible to fully correct for selection bias using this data.¹⁴ However, the evidence arguably gives no reason to believe that the shift towards Lopaq jobs after 65 was related to the characteristics of those working.

For the USA, education, housing status and ethnicity were important influences on Lopaq employment generally. Gender and marital status were less important than in the UK, an intriguing finding for future research. Overall, continuers aged 65–9 had a similar likelihood of being in a Lopaq occupation to those aged 58–9 (the difference was insignificant). Recruits aged 65–9 were 1.8 times more likely to be in Lopaq jobs than those aged 58–9, which was statistically significant, but nonetheless a smaller difference than in the UK. This reflects the fact that US employers appear to be more willing to recruit older people to non-Lopaq occupations.

Discussion

Despite a prominent UK business case for employing older people, little is known about the occupations open to over-65s in the 2000s (a group with few employment rights at this time). This article makes two major contributions. First, drawing on dual queuing theory, it examines UK employers’ use of employees over 65 in the 2000s. It was hypothesized that people in their 60s were disproportionately used to fill Lopaq occupations that were low paid, required few qualifications and were more difficult to fill with core-age workers. Pseudo-cohort analysis suggests UK employees were disproportionately segregated into Lopaq jobs after 65 because of employer recruitment and retention. Strikingly, half those recruited in their 60s were in Lopaq jobs. Logistic regression analysis suggests that this was not simply a reflection of disadvantaged-worker characteristics. Indeed, those recruited in their 60s appear to have more advantaged characteristics than employees at age 58–9, while not proven, this suggests that occupational downgrading occurs. The critical perspectives of Loretto and White (2006) and Taylor and Walker (1994) discussed in the introduction seem well grounded.

The second contribution is the theoretical and empirical comparison with the USA, a country with age discrimination legislation covering the over-65s. Managing diversity arguments often down-play the importance of legislative rights (Kirton and Greene, 2010), and some argue that age discrimination legislation makes employers less likely to recruit older people (Heywood and Siebert, 2008). For those continuing in jobs, however, it was suggested that legislative rights weaken the concentration in Lopaq occupations. Correspondingly, the proportion in US Lopaq occupations remained almost identical between ages 58–9 and 65–9 for those in the same job over the whole period.

Because US employers have more recruitment discretion it was predicted that over-65s recruited in their 60s would be segregated into Lopaq jobs to a similar degree to the UK. However, US employers appear more willing than their British counterparts to recruit to higher-level jobs: two-thirds recruited in their 60s were in non-Lopaq jobs, compared with half in the UK. Over time US age discrimination legislation may have normalized older peoples’ employment to a degree. Recruitment of older workers might occur because of positive employer experience of staff continuing past 65. Employer perceptions of how long older recruits can usefully work may therefore be extended. If so, legislation may enhance the business case for employing older people, supporting Kirton and Greene’s (2010) assertion that the ‘diversity approach’ requires legal underpinning to integrate under-represented groups.

This has implications for the 2011 default retirement age abolition, although it will take some time for these consequences to emerge. Employment rights for over-65s are not a panacea for retirement timing choice: some will find it particularly hard to continue working as state pension ages rise (Lain, 2011). However, US experience calls into question predictions of an increased unwillingness to recruit older people, at least in the medium term. More broadly, this article suggests that the default retirement age removal will reduce Lopaq segmentation to a much greater degree than voluntaristic ‘age-diversity’ commitments.

Footnotes

Acknowledgements

For comments at various stages the author thanks Professor Jacqueline O’Reilly, Professor Sarah Vickerstaff, Dr Ruth Woodfield, Dr Maura Sheehan, Ray Bacchan and the two anonymous reviewers. The Labour Force Survey was obtained from the UK Data Archive, and the Current Population Survey from the National Bureau of Economic Research. Earlier stages of this research were funded by an Economic and Social Research Council postgraduate studentship (grant number PTA-030-2002-00958).

Notes

References

Aaronson

Housinger

(1999) The impact of technology on displacement and de-employment. Federal Bank of Chicago Economic Perspectives 23(2): 14–30.

Adams

Rau

(2004) Job seeking among retirees seeking bridge employment. Personnel Psychology 3(57): 719–44.

Adams

(2004) Age discrimination legislation and the employment of older workers. Labor Economics 11(2): 219–41.

Atkinson

Meager

(1986) Changing Working Patterns: How Companies Achieve Flexibility to Meet New Needs. London: NEDO.

Barnes

Perry

Taylor

(2004) Working after State Pension Age: Qualitative Research, DWP Research Report, 208. Leeds: HMSO.

Cebulla

Butt

Lyon

(2007) Working beyond the state pension age in the United Kingdom: the role of working time flexibility and the effects of the home. Ageing and Society 27(6): 849–67.

Christensen

(1990) Bridges over troubled water: how older workers view the labor market. In: Doeringer

(ed.) Bridges to Retirement: Older Workers and a Changing Labor Market. Ithaca, NY: ILR Press.

Daniel

Heywood

(2007) The determinants of hiring older workers: UK evidence. Labour Economics 1(14): 25–51.

Disney

Whitehouse

(2002) The Economic Well-Being of Older People in International Perspective, Working Paper 306, Luxembourg Income Study, Luxembourg.

10.

Farnham

(2010) Tackling ageism in employment: age the final frontier of discrimination? In: Moss

(ed.) Profiting from Diversity: The Business Advantages and the Obstacles to Achieving Diversity. Basingstoke: Palgrave Macmillan, 194–214.

11.

Flynn

(2010) The United Kingdom government’s ‘business case’ approach to the regulation of retirement. Ageing and Society (30)3: 421–43.

12.

Flynn

McNair

(2009) What would persuade older people to stay longer in work? In: Chiva

Manthorpe

(eds) Older Workers in Europe. Maidenhead: Open University Press, 24–37.

13.

Freudenheim

(2005) More help wanted: older workers please apply. The New York Times, 23 March 2005. Available (consulted 7 November 2011) at: http://www.nytimes.com/2005/03/23/business/23older.html

14.

Gangl

(2003) Unemployment Dynamics in the United States and West Germany. Heidelberg: Physica-Verlag.

15.

Ganzeboom

Treiman

(1996) Internationally comparable measures of occupational status for the 1988 International Standard Classification of Occupations. Social Science Research 25(3): 201–39.

16.

Gobeski

Beehr

(2009) How retirees work: predictors of different types of bridge employment. Journal of Organizational Behavior 30(3): 401–25.

17.

Haider

Loughran

(2001) Elderly Labor Supply: Work or Play? CRR Working Paper 2001–04, Center for Retirement Research at Boston College, Chestnut Hill, MA, USA.

18.

Hayward

Hardy

Liu

(1994) Work after retirement: the experiences of older men in the US. Social Science Research 23(1): 82–107.

19.

Heywood

Siebert

(2008) Understanding the Labour Market for Older Workers, IEA Discussion Paper 23, Institute for Economic Affairs, London, UK.

20.

Hirsh

Macpherson

Hardy

(2000) Occupational age structure and access for older workers. Industrial and Labor Relations Review 53(3): 401–18.

21.

Issacharoff

Harris

(1997) Is age discrimination really age discrimination? The ADEA’s unnatural solution. New York University Law Review 72(4): 780–840.

22.

Kilgour

(2006) Public sector pension plans: defined benefit versus defined contribution. Compensation and Benefits Review 38(1): 20–8.

23.

Kirton

Greene

(2010) The Dynamics of Managing Diversity. Oxford: Butterworth-Heinemann.

24.

Lain

(2009) Healthy, wealthy and wise? Working past age 65 in the UK and USA. Unpublished DPhil thesis, University of Sussex, Brighton.

25.

Lain

(2011) Helping the poorest help themselves? Encouraging employment past 65 in England and the USA. Journal of Social Policy 40(3): 493–512.

26.

Loretto

White

(2006) Employment of older workers: employers’ attitudes, policies and practices. Human Resource Management Journal 16(3): 313–30.

27.

Luffman

(2006) The core-age labour force. Perspectives on Labour and Income 7(9): 5–11.

28.

McNamara

Williamson

(2004) Race, gender, and the retirement decisions of people ages 60 to 80: prospects for age integration in employment. International Journal of Aging and Human Development 59(3): 255–86.

29.

Macnicol

(2006) Age Discrimination: An Historical and Contemporary Analysis. Cambridge: Cambridge University Press.

30.

Meadows

(2003) Retirement Ages in the UK: A Review of the Evidence. Employment Relations Research Series 18. London: Department for Trade and Industry.

31.

Neumark

(2009) The Age Discrimination in Employment Act and the challenge of population aging. Research on Ageing 31(1): 41–68.

32.

Neumark

Stock

(1999) Age discrimination laws and labor market efficiency. Journal of Political Economy 107(5): 1081–1110.

33.

Noon

(2007) The fatal flaws of diversity and the business case for ethnic minorities. Work, Employment and Society 21(4): 773–84.

34.

Noon

Ogbonna

(2001) Equality, Diversity and Disadvantage in Employment. Basingstoke: Palgrave.

35.

OECD (1999) Employment protection and labour market performance. In: OECD Employment Outlook 1999. Paris: OECD, 48–132.

36.

Parnes

Sommers

(1994) Shunning retirement: work experiences of men in their seventies and early eighties. Journal of Gerontology: Social Sciences 49(3): S117-S124.

37.

Patmore

(2003) Understanding Home Care Providers. York: Social Policy Research Unit.

38.

Peters

(2006) Second careers and the third age: you’re only as old as your new job. Guardian, 19 January 2006. Available (consulted 7 November 2011) at: http://www.guardian.co.uk/business/2006/jan/19/workandcareers.changingjobs

39.

Quadagno

Hardy

(1991) Regulating retirement through the Age Discrimination in Employment Act. Research on Aging 13(4): 470–75.

40.

Reskin

Roos

(1990) Job Queues, Gender Queues: Explaining Women’s Inroads into Male Occupations. Philadelphia, PA: Temple University Press.

41.

Riach

Loretto

(2009) Identity work and the ‘unemployed’ worker: age, disability and the lived experience of the older unemployed. Work Employment and Society 23(1): 102–19.

42.

Salanicoff

Cubanski

Ranji

Neuman

(2009) Health coverage and expenses: impact on older women’s economic well-being. Journal of Women, Politics and Policy 30(2): 222–47.

43.

Smeaton

McKay

(2003) Working after State Pension Age: Quantitative Analysis, DWP Research Report 182. London: HMSO.

44.

Smith

(2000) A Cross-National Comparison on Attitudes to Work by Age and Labour Force Status. Paris: OECD.

45.

Taylor

Walker

(1994) The ageing workforce: employers’ attitudes towards older workers. Work, Employment and Society 8(4): 569–91.

46.

Trapp

(2004) Older and Wiser. People Management, 23 December, p. 40–1.

47.

Vickerstaff

(2006) Entering the retirement zone: how much choice do individuals have? Social Policy and Society 5(4): 507–17.

48.

Vickerstaff

Loretto

Billings

Brown

Mitton

Parkin

White

(2008) Encouraging Labour Market Activity among 60–64 Year Olds, DWP Research Report 531. Norwich: HMSO.

49.

Warhurst

Nickson

(2007) Employee experience of aesthetic labour in retail and hospitality. Work Employment and Society 21(1): 103–20.