Introduction
The retention of older employees in employment, often referred to as EWL (extended working life), is a crucial issue for the funding of the welfare state because of rapidly increasing longevity and the costliness of pensions (Esping-Andersen, Reference Esping-Andersen and Esping-Andersen1996; OECD, 2004). In Britain, it is only if employees can be persuaded to remain in work for additional years that a solution to this problem looks feasible (Pensions Commission, 2005; for some other European countries, see Hofäcker, Reference Hofäcker2010).
Government policy toward older workers in many countries has focused mainly on financial incentives and disincentives. In Britain, there have been numerous government actions of that type. State pension age (SPA) for women is being progressively raised from age sixty and will soon reach sixty-three; by 2020 SPA for both men and women will be sixty-six (for men it currently remains sixty-five). Employers, moreover, are now prevented (in most circumstances) from setting a compulsory retirement age for their employees, who are free to work on, and the scope for combining earnings with the state pension was enlarged in the 2004 Finance Act. Meanwhile gateways to early exit have been narrowed: in 1997, by taxation of company pension funds, a move that, perhaps inadvertently, weakened these funds and restricted their capacity to support early retirement; in 2010, by raising to fifty-five the age at which occupational or private pensions could be realised (it was previously fifty). There have also been recent moves to reduce the level of incapacity benefits and make it more difficult to obtain them. On the side of positive incentives, a substantial addition to the state pension is now (since 2006) awarded for each year by which retirement is delayed beyond SPA.
Government policy, however, is only one influence on older employee retention: employers and workplace conditions are also important. Direct age discrimination was outlawed in 2005, but it is doubtful if this would help older employees if they encounter deteriorating job conditions (see McNair and Flynn, Reference McNair and Flynn2006). Further progress on retention is likely to depend on job quality dimensions as well as on the financial deal provided. Britain, with its ‘voluntaristic’ tradition of workplace regulation, avoids state involvement in ‘quality of working life’ (QWL), unlike, for example, Scandinavian countries. Reliance on employers to sustain older employee recruitment and retention therefore assumes implicitly that employers will take steps to assure older employees an adequate QWL. This is the key issue addressed in the present article.
This article argues that older British employees have, since the 1990s, experienced a less satisfactory work situation than they would have been led to expect from conditions in their earlier careers. Having disappointed expectations, older employees naturally manifest discontent through their attitudes toward their jobs (as predicted by psychological contract theory). As shown here, attitudes in fact declined significantly over a fifteen-year period from the early 1990s to the mid-2000s, and did not recover during the subsequent finance-industry-led recession: in some respects they became increasingly negative. These attitudes matter both for the retention aim and more generally in terms of economic efficiency. Harrison et al. (Reference Harrison, Newman and Roth2006) have concluded that overall job satisfaction (OJS) and organisational commitment (OC), jointly termed ‘overall job attitude’, play a central role in predicting task performance, contextual performance (‘organisational citizenship behaviour’), timekeeping, attendance and propensity to stay/quit. These cover the main aspects of labour cost and employee performance that interest employers.
Following Maltby (Reference Maltby2011) this article broadens the usual perspective on policies toward older workers by highlighting the role of employers and of QWL. It also makes new empirical contributions to this field by providing the first full analysis of overall job attitude of older British employees, and, furthermore, by examining satisfactions with a number of aspects or facets of QWL, with the aim of diagnosing which aspects give rise to deteriorating experience of work and hence might be the focus of QWL policy. Analyses cover a twenty-year period in all, the longest of any study of older British employees, and the article considers what new issues arise for older employees in the period following the financial recession of 2008. The research provides a basis, therefore, for assessing whether declining attitudes among older employees are transitory or persistent, and also for testing whether the recent recession has led to any reversal of trends, as might happen if older employees are grateful simply to have a job.
The treatment of older employees, in context
From the early 1800s, there were British employers who were aware of the incentive value of ‘welfare’, for instance in recruiting and retaining labour under tight labour market conditions. Older employees who had shown ‘loyalty’ by long service to such employers often received special benefits and privileges (Russell, Reference Russell1991), although this is not to deny the poor conditions that widely prevailed at the time. By the inter-War period, during which large employers became dominant in Britain, long-serving employees in the public sector and in major firms could enjoy occupational pensions, sick pay schemes and other benefits (Fitzgerald, Reference Fitzgerald1988). Even though British companies’ personnel practices rarely came to match the internal labour markets of the USA's primary sector (Edwards, Reference Edwards1979), the late career was when British employees could expect payoffs from their service. Nor should it be assumed that these points only apply to large employers. Many small employers rely on long-service employees who are retained and valued: a regime that may help explain the above-average levels of trust and the below-average levels of absence and staff turnover in such establishments (see, for example, Forth et al., Reference Forth, Bewley and Bryson2006: 75–7). These general points do not of course preclude the existence of numerous ‘bad jobs’ in the low-paid sector of employment, whether for small employers or the largest (see, for example, Toynbee, Reference Toynbee2003).
Even in the severe economic recession of the 1980s, large British employers were able to draw on their pension funds to offer older employees early retirement packages that shielded them from the adverse labour market. Early retirement was taken by about one half of British working age men over fifty-five, and one third of women of that age, by around 1990 (Disney et al. Reference Disney, Grundy and Johnson1998). This development was still more marked, and/or took place earlier, in some other European countries, where state welfare benefits and public early retirement policy played a larger role (Hofäcker, Reference Hofäcker2010).
It was in the 1990s that a more serious challenge to the established order became apparent. Another sharp recession for Britain affected the years 1990–93, and although the subsequent economic recovery was rapid, British employers were now encountering that complex of changes that has become known as ‘globalisation’ (Dreher et al., Reference Dreher, Gaston and Martens2008), with increased competitive pressures on Britain and other western industrialised economies. Alongside this, the public sector in Britain (as also in the USA and some other countries) experienced repeated bouts of privatisation and restructuring, leading to reductions in public sector jobs (Morgan et al., Reference Morgan, Allington, Heery, Heery and Salmon2000). External competitive and financial pressures in turn led employers to seek cost reductions through new forms of work organisation and regimes of continuous change, and impoverished pension funds could do less to protect older employees’ future.
The question is whether and in what specific ways there has been a deterioration in the employment experience and in QWL, likely to affect older workers? Here three main ideas can be discerned in debates about QWL. One concerns a rising level of work intensification or work pressure, another insecurity and precarity, and the third a diminution of autonomy in jobs.
Work intensification
Work intensification is a particularly well-documented aspect of declining QWL. Green (Reference Green2006) has analysed increased work pressures not only for Britain but for several other ‘advanced’ economies. He attributes this development primarily to the use of ICT in work control and to associated changes in work organisation. Others have pointed to the pressure of implementing continuous change (Worrall and Cooper, Reference Worrall and Cooper2001), and the prevalence of unpaid overtime (TUC, 2012), a practice earlier established in the USA (Schor Reference Schor1991). The ICT-based monitoring of tasks (Green, Reference Green2006) greatly reduces the costs of control compared with previous methods, and renders deferred rewards, such as service increments and pensions that older long-serving employees have relied on, less attractive as an employer incentive strategy. In both the private and the public sectors, there has been a steady advance of performance-based pay in Britain, and government has recently been moving to cut the costs of automatic service increments in the public services.
Of the three adverse tendencies highlighted in the QWL literature, this is the one with the strongest evidence in regard to impact on older employees. Quantified evidence comes from survey data (Smeaton and Vegeris, 2009) relating to employees leaving a job in their late career, and to those approaching retirement age (Smeaton et al., Reference Smeaton, Vegeris and Sahin-Dikmen2010): stress and overwork is the common message. Furthermore, case study evidence from McNair and Flynn (Reference McNair and Flynn2006), across several branches of both public and market services, reveals the effects of work pressures and stress, often provoked by organisational change, on early retirement. These authors discount the anti-discrimination legislation of 2005 as of relevance to these issues. Roberts (Reference Roberts2006) provides further case evidence concerning older workers in heavy manual occupations. Cost concerns and inflexibility in the treatment of older employees are revealed in a qualitative study of managers by Loretto and White (Reference Loretto and White2006).
Increasing precarity
The evidence for increasing precarity as impacting on older employees is more circumstantial than in the case of work pressure. Indirect evidence comes in part from European trends. The rise in unemployment across most of Europe during the 1980s, and the increase in ‘non-standard’ forms of employment contract, gave rise among sociologists to the ideas of precarious employment and destandardisation of the life-course (Beck, Reference Beck1992). Older workers are affected by increased probability of unemployment or of diminished jobs before retirement can be reached in the precarious labour market (Guillemard and Rein, Reference Guillemard and Rein1993). It has been argued by Standing (Reference Standing2011: 83–6) that as SPA increases and occupational pensions decline (both being the case in Britain), older people may compete for places in a growing sector of part-time, temporary or casual jobs. Britain, it is true, has generally made rather limited use of temporary jobs, at least by comparison with other European countries. Nonetheless, there is evidence from the Labour Force Surveys that in moving jobs some older employees experience a change of status to temporary, part-time or self-employed (Lissenbergh and Smeaton, Reference Lissenbergh and Smeaton2003); this evidence, and the case material of Roberts (Reference Roberts2006), point to a segmentation of older workers, at least at the less-qualified margin, into lower-quality jobs. Also, British employers appear to have adopted redundancy (labour-shedding) as a normal instrument of policy even in times of growth (Turnbull and Wass, Reference Turnbull, Wass, Heery and Salmon2000); add to this the retreat of ‘first in-last out’ redundancy practice, and it is easy to see how older employees can be sorted into those complying with higher work pressure and those who can be let go.
Diminishing autonomy
The third area of declining QWL is harder to encapsulate, and still remains to be linked specifically to the experience of older British employees. There is, however, substantial evidence for general British employee samples (for example, Gallie et al., Reference Gallie, Felstead and Green2001) that the task discretion permitted to employees, an important component of autonomy, was reduced in Britain during the 1990s. Moreover, these authors showed that lower task discretion was associated with reduced organisational commitment. More recent studies have revealed a variety of wider repercussions from shrinking autonomy/discretion (Gallie, Reference Gallie and Gallie2007, Reference Gallie2013; Gallie et al., Reference Gallie, Felstead and Green2012). Despite the lack of a specific focus on older employees in these studies, it remains plausible that employees who can draw on long experience and substantial firm-specific knowledge and skills will be sensitive to any loss of autonomy and discretion.
What further changes in QWL can one envisage as a result of Britain's 2008 finance-led recession? Several developments have been highlighted by commentators. Unemployment has risen, but less than in the two previous recessions, and the main impact has been on youth, so in this respect older employees may even be somewhat grateful for how things turned out. Employment has risen, but employee jobs have tended to offer fewer hours, and real earnings have fallen, both because of this and because of wage cuts or wage freezes. Thus, pay may be an emergent area of dissatisfaction, while work intensification may have been partly replaced by under-employment. Finally, British occupational pensions have gone down the path taken in the USA more than a decade previously (Jacoby, Reference Jacoby1999), constraining the benefits on offer, and putting up the shutters on new entrants. These pension developments were foreseeable, but the recessionary pressures have made pension decline a more prominent issue in public debate.
Hypotheses
The hypotheses suggested by the foregoing discussion can now be summarised. After setting out the hypotheses, we will also outline a method of testing.
H1: There has been a deterioration in the overall job attitude of older employees following the sea-change in the economy, technology and work organisation that took place in the 1990s. Following Harrison et al. (Reference Harrison, Newman and Roth2006), we operationalise this hypothesis as follows:
H1a: There has been a deterioration in the overall job satisfaction (OJS) of older employees.
H1b: There has been a deterioration in the organisational commitment (OC) of older employees.
We further hypothesise that there has been deterioration in attitude of older employees toward one or more of the particular forms in which QWL has been undermined:
H2: There has been a reduction in the satisfaction of older employees toward job security.
H3: Because of declining autonomy, there has been a reduction in the satisfaction of older employees toward the work that they do.
H4: There has been a reduction in the satisfaction of older employees toward work effort, which we operationalise in two parts:
H4a: There has been a reduction in the satisfaction of older employees toward the hours they work (long hours, unpaid overtime, presenteeism, and lack of personal choice may be components).
H4b: There has been a reduction in the satisfaction of older employees toward the amount of work they are required to do.
Additionally, it is possible that feelings of increased or unreasonable work demands will express themselves as dissatisfaction with compensation. Thus we add the following:
H4c: There has been a reduction in the satisfaction of older employees toward their pay (this may have been intensified over the recent recessionary period when employees were often kept on short hours).
To test these hypotheses, we compare attitudes from a baseline in 1992 with year 2006. The comparison provides a strong test of H1–H4, envisaging that the changes in the situation of older employees that took place in the 1990s were so deep-seated that even a long period of apparent prosperity (1995 to 2006) was unable to reverse them. We also investigate whether attitude change up to 2006 was continued or reversed over the recession that began in Britain in 2008. For this purpose, we compare attitudes in 2006 and 2012.
Finally, a further comparison of 1992 with 2012 provides a still stronger test of H1–H4. This comparison covers two decades and the years are reasonably well matched in their position within the business cycle, 1992 being the third year of the 1990s recession, and 2012 being the fifth year of the long recession of the 2000s. Both 1992 and 2012 came two years before the start of an economic upturn.
When making comparisons over time, it is analytically advantageous to treat them not in absolute terms but relatively, with older employees naturally compared to younger. This also has a theoretical rationale. Older employees are likely to experience the adverse trends in the employment relationship more painfully than younger employees for reasons that we have already alluded to. Traditionally they have enjoyed a relatively privileged treatment (Russell, Reference Russell1991), and the motive on the employers’ side for providing such favour is explained by the incentive theory of deferred compensation (Edwards, Reference Edwards1979; Lazear, Reference Lazear1981). This depicts employers under-rewarding employees (relative to their productivity) during their early years and over-rewarding them (again relative to productivity) in their later years. The expectation of late-career payoff for long service strengthens loyalty and commitment and deters opportunism. But when this is replaced by a hardening of the employment relationship (for example, Roberts, Reference Roberts2006), the older workers, with expectations formed in very different circumstances, will experience it as a reneging or ‘contract violation’ (Robinson and Rousseau, Reference Robinson and Rousseau1994; White, Reference White2012), and perceived breaches of the ‘psychological contract’ are typically followed by a negative reaction (for example, Coyle-Shapiro and Kessler, Reference Coyle-Shapiro and Kessler2000).
Data and variables
The research uses three survey datasets representative of British employees: for 1992, the ‘Employment in Britain’ (EIB) survey (Gallie et al., Reference Gallie, White, Cheng and Tomlinson1998), for 2006 the ‘Skills Survey’ (SK06) (Felstead, Reference Felstead2010), and for 2012 the Skills and Employment Survey (SES12). The analysis covers employees aged twenty to sixty. Excluding people aged over sixty reduces the potential problem of selection into or out of the labour force during years when early retirement or incapacity is particularly common.
The surveys, given these age limits, provide samples of 3,458 employees in 1992 (71 per cent response rate), 4,237 from 2006 (57 per cent response) and 2,377 in 2012 (49 per cent response). The decline of response rates is characteristic of British social surveys over this period. The numbers exclude booster samples for various geographical areas available in the 2006 and 2012 surveys that have been disregarded here so as to maintain uniformity across the surveys. All three surveys had a similar sampling design and information was collected from respondents by means of personal interviews in the home. Weights are applied to make the surveys representative of British employees at the specified dates. The datasets and documentation are available in the UK Data Archive.
Dependent variables
To assess overall job attitude (H1), the analysis uses measures of (1) overall job satisfaction (OJS), and (2) organisational commitment (OC); this closely matches the concept proposed by Harrison et al. (Reference Harrison, Newman and Roth2006). OJS is obtained by the familiar single-item measure. Originally developed in applied psychology, single-item measures of satisfaction have also been adopted by economists (see, for example, Clark et al., Reference Clark, Diener, Georgellis and Lucas2008). OJS has been treated as an interval (cardinal) measure by psychologists, and this has now also been widely accepted in analysis by economists (Ferrer-i-Carbonell and Frijters, Reference Ferrer-i-Carbonell and Frijters2004). The cardinality assumption is adopted for the present analysis.
Affective OC measures the degree to which an employee identifies with the goals and values of an organisation and is willing to exert effort to help it succeed (Kalleberg and Berg, Reference Kalleberg and Berg1987). It is here measured as the sum of six items used by Lincoln and Kalleberg (Reference Lincoln and Kalleberg1990). The items have a reliability in these surveys (Cronbach alpha) of 0.79. We sum the items to derive an OC scale. Further details and descriptive statistics are shown in Table 1.
Table 1 Questions and descriptive statistics for overall job satisfaction (OJS) and organisational commitment (OC)
Notes: Statistics are unweighted. 1992 = Employment in Britain survey; 2006 = Skills Survey 2006; 2012 = Skills and Employment Survey 2012.
+ Responses were ‘completely satisfied, very satisfied, fairly satisfied, neither satisfied nor dissatisfied, fairly dissatisfied, very dissatisfied, completely dissatisfied’ with associated response numbers 1–7. For analysis, the scoring has been reversed so that completely satisfied = 7.
* Responses were ‘Strongly agree, agree, disagree, strongly disagree’ with scoring 1–4. The scoring has been reversed so that strongly agree = 4, except in the case of the second item that already has this sense. An answer ‘Don't know’ is treated as missing data.
For hypotheses H2–H4, facet satisfaction measures are available with identical wording in EIB, SK06 and SES12. Each is analysed separately, making the cardinality assumption. We select six facet measures for analysis, because they connect with H2–H4. Thus, satisfaction with security is included to assess the impact of precarity (H2). Satisfaction with work is included to assess the impact of declining autonomy (H3); the item is rather wider in its reference, but there is no satisfaction item that exclusively captures autonomy. Satisfaction with hours worked and satisfaction with the amount of work are included to assess the impact of work intensification (H4a, H4b). Further, since increased work demands might also be reflected in higher wage expectations, we analyse satisfaction with pay (H4c). Finally, satisfaction with benefits is also analysed because of the recent deterioration in occupational pension provision. Details of the six facet satisfaction items are shown in Table 2.
Table 2 Questions and descriptive statistics for facets of job satisfaction
Notes: Statistics are unweighted. 1992 = Employment in Britain survey. 2006 = Skills Survey 2006; 2012 = Skills and Employment Survey 2012. Responses and scoring are as for OJS (see Table 1).
Explanatory and control variables
The chief explanatory variable throughout the analyses is age group, with across-time comparisons. Age fifty is the most widely used cut-off for defining older workers (Felstead, Reference Felstead2010), and in Britain this appears consistent with managerial perceptions (Loretto and White, Reference Loretto and White2006). Comparisons are made with employees in the twenty to forty-nine age bracket. Attitudes are compared in three pairings: 2006 versus 1992 (the basic effect resulting from the new pressures of the 1990s), 2012 versus 2006 (the across-recession effect) and 2012 versus 1992 (the twenty-year persistence test between similar points in the business cycle). We also explored an elaboration with age fifty to sixty split into fifty to fifty-four and fifty-five to sixty. The fifty-five to sixty subgroup tends to yield somewhat larger (more negative) estimates, but the qualitative picture remains the same for the two subgroups, so we confine our report of results to the combined fifty to sixty group.
Control variables address compositional changes over time that potentially vary between age groups and could affect attitudes. The analysis includes controls for the variables that are commonly assumed to affect the participation decision. The individual's earning capacity is represented by highest educational qualification and occupational class (the latter being interpreted as an indicator of achieved skill level: see Tåhlin, Reference Tåhlin and Gallie2007). To indicate family situation, we used questions about whether the respondent had a partner, and whether there was a dependent child. These family questions were interacted with gender in recognition of their different importance for labour market participation between women and men. For the analysis of OJS only, we add hours and the square of hours to the specification. A second group of control variables concerned the type of organisation in which the individual worked, with a five-category variable for workplace size, ranging from ten or less to 500 or over, a six-category industry group variable, and dummies for private sector and for presence of a recognised trade union.
Analysis
The analysis method is robust regression (Berk, Reference Berk, Fox and Long1990), with each attitudinal outcome analysed in turn and using common specifications across outcomes. Robust regression is an OLS-like methodology which takes account of weighting and complex survey design in calculating standard errors.
The analysis design is to pool data for each pair of surveys to be compared and to obtain a second difference (or ‘difference in differences’, see Wooldridge, Reference Wooldridge2002: 130–2) by means of an interaction term between age group and year. The coefficient on this interaction term is equivalent to the (conditional) mean between-age-group difference across times, or the mean between-time difference across age groups. An advantage offered by the design is that it removes unobserved differences between the age groups to the extent that these are constant over time. The design thus helps address the problem of persistent unobserved influences, such as positivity/negativity of outlook, which may vary with age, on expressions of dissatisfaction (see Ferrer-i-Carbonell and Frijters, Reference Ferrer-i-Carbonell and Frijters2004). On the assumption that such age-specific differences are constant across time periods, the design differences them out.
Results
In the following, ‘age effect’ is a shorthand expression meaning the interaction effect of age-group compared across time. In other words, an ‘age effect’ is always relative, and it is always a difference (comparison) between two years. A negative estimate means that older employees have moved toward a more dissatisfied position, on average, relative to younger employees over the time period in question. All hypotheses predict negative effects.
Overall job attitudes
Table 3 presents the estimated age effects for each of the comparisons. All analyses include the full set of controls, but for reasons of space, only the estimates for the age effects are shown (full results are available on request). It can be seen that on both OJS and OC the age effects were negative when comparing 2006 with 1992. In 1992, older employees had much higher levels of OC than younger ones, but by 2006 this difference had all but disappeared (table not shown). The relative change was less marked in the case of OJS; older employees were again significantly more satisfied than younger employees in 1992, and this was still the case in 2006 though with smaller magnitude. There is no ‘recovery’ of older employee attitudes over the recessionary period of 2006 to 2012, but rather a further decline, weakly significant (10 per cent level) on OC, but small and non-significant for OJS. Over the twenty-year period 1992 to 2012, there is a highly significant fall in OC for older employees and a smaller fall in OJS that is weakly significant at the 10 per cent level.
Table 3 Estimated interaction effects of year by age-group (older employee (50+) with reference to younger (under-50) employees) on overall job attitudes (OC and OJS)
Note: All estimates come from models with complete controls (see text). Standard errors are derived by robust regression.
Job facet satisfactions
Table 4 summarises the estimated age effects for each of the six job facets (security: H2; work itself: H3; hours, amount of work, and pay: H4a/b/c; and benefits). The first two columns of results show the age comparisons for year 2006 relative to 1992. All six estimates are negative, but only three significantly so: satisfaction with hours (H4a), satisfaction with amount of work (H4b), and satisfaction with the work itself (H3). The age effect for satisfaction with pay (H4c) is non-significant, and this is also the case for satisfaction with fringe benefits. The estimate for satisfaction with job security (H2) fails to reach significance at the 5 per cent level but is weakly significant at the 10 per cent level.
Table 4 Year and age interaction effects on six facets of job satisfaction
Note: All estimates come from models with complete controls (see text). Standard errors are derived by robust regression.
The results, therefore, provide evidence of deteriorating satisfaction in relation to the demands of work (H4a/b: hours/workload), but there is no evidence of reward (pay) being implicated. The absence of any increase in dissatisfaction with pay also makes it relatively unlikely that dissatisfaction with hours relates to ‘too few hours’ since desire for more paid hours would surely be connected to feelings that pay was inadequate. There is evidence also of deterioration in relation to the nature of work itself (H3) as experienced by older employees; however, this evidence does not link directly to decreasing autonomy, and could reflect other intrinsic job changes. For deteriorating satisfaction with job security (H2), the evidence is weaker. A relatively weak effect is, however, consistent with the restricted development of precarious forms of employment contract in Britain. The proportion of older employees in non-permanent jobs was less than one in ten (8.6 per cent) in 1992, and fell further to 5.8 per cent in 2006; this was lower than the corresponding proportion for younger employees. The analysis of satisfaction with fringe benefits suggests that by 2006 pensions and related welfare policies had not become a major issue for older employees.
The across-recession comparisons (2012 versus 2006) are very different. The age effect for satisfaction with amount of work (i.e. work-load) has become positive (though non-significant) over the recessionary period. Age effects for satisfaction with hours of work and with the work itself remain negative but are non-significant. Age effects for satisfaction with security and satisfaction with benefits (both nonsignificant in the 2006 versus 1992 comparison) are significant at the 5 per cent level in the cross-recession comparison.
Turning finally to the twenty-year comparisons (2012 versus 1992), we find that all estimates for the older employee effect are negative, and significant at the 1 per cent or 5 per cent levels, except in the case of pay, where significance is only weakly present at the 10 per cent level.
The set of results for the job facet measures shows older employees becoming progressively less satisfied across important features of their work situation. Initially the tendency is concentrated on work-load/work-hours and the nature of work itself, but over the recession these sources of dissatisfaction ease somewhat, which is consistent with the ‘short hours’ noted by economic commentators. At the same time, however, feelings of insecurity grow as does dissatisfaction with fringe benefits. Feelings of insecurity probably do not reflect increased exposure to temporary contracts. In 2012, these applied to 5.9 per cent for older employees, virtually the same as the 5.8 per cent in 2006; less than younger employees in both years. Overall unemployment rates, the prevalence of compulsory redundancies, and the financial implications of becoming unemployed are more plausible explanations.Footnote 1 Growing dissatisfaction with benefits can be explained straightforwardly by greater awareness of the parlous state of employer pension provision.
In a linear model, the magnitudes of these twenty-year effects can be legitimately compared with the estimates for other variables. On insecurity, hours worked and amount of work, the age–period interaction has the largest negative shift of any regressor variable. On pay, benefits and the work itself, the age–period interaction effect is exceeded only by the class differential between managers and lower routine work (i.e. low-skill, low-pay) employees, and by the public sector versus market sector contrast (see the authors’ website for full results).
Summary and policy discussion
This research provides strong evidence in support of the hypotheses specified earlier. It reinforces and extends existing evidence that older employees in Britain were becoming more discontent across two decades that span the end of the previous century and the start of the current one, a period marked by globalised competition and technological innovation in the workplace. By 2006, both organisational commitment and overall job satisfaction, the two elements of ‘overall job attitude’, were negatively displaced for older employees relative to younger. This was despite a prolonged period of economic prosperity and improving labour market in Britain, from the mid-1990s. There was also no recovery over the recessionary period starting in 2008, when perhaps older employees might have been glad at least to remain in jobs. The twenty-year comparison between 1992 and 2012, two similarly placed years in the business cycle, confirms a persistent decline in overall job attitude.
There was also substantial evidence of declining relative satisfaction with specific aspects of QWL. Here, however, there were some notable differences between the picture up to 2006 and the recessionary period 2006 to 2012. Initially the largest negative tendencies concerned satisfaction with hours worked, effort demanded, and the nature of the work itself. Subsequently, as the ‘long hours culture’ was replaced with a period of shortened hours and ‘under-employment’, these work-related attitudes stabilised somewhat. At the same time though, dissatisfaction with job (in)security increased; despite unemployment remaining at a moderate level compared with the early 1990s, it appears that anxieties were not allayed (see Footnote note 1). Furthermore, older employees became more dissatisfied with their occupational benefits, a result that probably reflects growing awareness of the deterioration of corporate (including public services) pension provision. Over the full twenty-year comparison period, five of the six facet satisfactions that were analysed declined significantly for older employees relative to younger.
The chief policy implication of these findings is that attention needs to be paid to QWL issues for older employees: partly on grounds of equity, so that people are not experiencing an unpleasant EWL; partly on grounds of efficiency, as per Harrison et al. (Reference Harrison, Newman and Roth2006). The problem area for many older employees is the work situation, its demands (excessive) and its intrinsic rewards (declining).Footnote 2 However, as noted earlier, Britain with its ‘voluntaristic’ workplace regulation has no tradition of public involvement in QWL. Employers and trade unions must be the agents of QWL initiatives. This seems to leave only a residual role for government, but one must also bear in mind the important role of public service employers in the British labour market (see later).
The leading initiative of relevance to QWL on the part of British government is Health for Work and Well-being (HWWB) which involves cooperation between the Department for Work and Pensions (DWP), the Department of Health (DoH), and the Health and Safety Executive (HSE) (Black, Reference Black2008). However, the inherent limitation here is the focus on illness and disability (see Maltby, Reference Maltby2011). Most older employees are in good health and it is not themselves but workplace conditions that are in need of treatment. The possibility of an alternative intervention model can be glimpsed from the WorkAbility initiative that stems from Finland (Ilmarinen, Reference Ilmarinen, Ilmarinen and Lehtinen2004). It emphasises the value of making work fit the abilities of those who perform the work, and its application has been based on widespread absorption into Finnish HR management practice.
The nearest to such an approach in Britain is the field of stress management. There is a wide consensus that work-strain is a general problem of the continuously changing workplace (Worrall and Cooper, Reference Worrall and Cooper2001), and it is one that may have special relevance to older employees. The HSE has addressed the issue by developing a set of ‘management standards’ that can be used in support of HRM activities to monitor stress (HSE, 2012). The crucial point here is the recognition that progress in this area depends on managerial action.
Another area where public policy and workplace practice can complement each other is flexibility. Government has progressively extended coverage of the employee right to make a ‘flexible work request’. Flexible work is defined as a way of working that suits an employee's needs, and examples include variable start and finish times or working from home. Employers now have an obligation to deal with the request in a reasonable manner; in other words, the legislation imposes equitable process. Regulation here opens the door a little wider, since flexibility policy has formed a recognised element in workplace HR practice at least since the early 2000s in Britain. But there may be a need to make (older) employees more aware of their ‘flexibility rights’ if the opportunity is to be taken, especially in smaller firms where professional HRM is often absent. There is also scope, given flexible hours practices, for innovative use of work-breaks: healthy eating initiatives are an example of such thinking (Jabbs and Devine, Reference Jabbs and Devine2006).
Diversity management has become one of the core concepts for employers with well-developed HRM functions, and is built on a firm business case. By the early 2000s, older people were being recruited more widely than other ‘minority’ groups by employers (CIPD, 2002; White et al., Reference White, Hill, Mills and Smeaton2004: 121). ‘Age management’ has emerged as a fairly distinct aspect of diversity policy and this has led to a variety of innovative actions by companies (Walker, Reference Walker2005). In Britain, the provision of training opportunities in the workplace (often highlighted as an area where older employees are neglected) recently shifted toward more equal provision (Felstead, Reference Felstead2010).
This scan of recent developments provides some indications of where employers have been developing a more age-friendly QWL. However, a lesson from the past is that major change requires leadership, and assuming government will not (or cannot) provide it, we have to ask which branch of industry would champion a focus on QWL either as a whole, or specifically for older workers. One might expect the public sector to play this role, in part because it is the largest employer of older workers, in part because of the much stronger position of unions there than in the private sector, and in part on the basis of recent history. Notably, the public sector has been the leading promoter of equal opportunities policies and recruitment diversity (Barnes et al., Reference Barnes, Smeaton and Taylor2009: 33–9), and more generally of HRM practices (White, Reference White2015). However, the capacity of public sector management to take on the role of QWL leadership in regard to older employees may be low in a coming period of further contraction and financial stringency. Public sector management might though be encouraged to take action in this field if government signalled concern for QWL, and this in turn might become more likely if it can be shown that improved QWL would lead to an increase in EWL, an existing British government priority. Establishing such a link remains a task for future research.
Acknowledgements
Data were provided by the UK Data Archive; thanks to Alan Felstead for advice on use of the Skills Surveys, and to three anonymous reviewers and the Editors for their comments and suggestions. The usual disclaimers apply.
