Umeå universitets logga

A central finding in earlier research is that attitudes to work generally are more positive among older workers than among younger workers. This result has been interpreted in two different ways, by the cultural and the structural hypotheses. The cultural hypothesis sees age differences as outcomes of generational differences. We would expect that different cohorts should hold different work attitudes and that the work values of an age group at an earlier point in time should be different from the work values of the same age group at a later point in time. The structural hypothesis sees age differences as expressions of labour market inequality between older and younger workers. This point of view leads us to expect that age differences in work attitudes will follow changes in the job structure and in working conditions.

Drawing on data from the Swedish survey of living conditions (ULF), attitude change within the Swedish work force during the period 1979–2003 was examined. Three sub-periods, 1986/1987, 1994 through 1996 and 2001 through 2003 were compared to 1979, the year of reference. The main results showed that a consistently lower share of the work force held extrinsic work values in the subsequent periods and this applied to all age groups. The results did not support the assumption that broader cultural differences between generations are central explanations of differences in work values. Older workers held extrinsic work values to a lesser degree than younger workers regardless of period. Most strikingly the gap between the youngest group on the labour market (aged 16

–29) and the older ones widened during the period. Furthermore, class differences in the distribution of the extrinsic attitude were intact throughout the study period; manual employees were consistently more likely to hold an extrinsic attitude than were service class employees. This implies that differences in the probability of extrinsic work attitudes have been identifiable regardless of period, but that their prevalence has decreased since jobs involving features related to extrinsic work values have decreased since 1979.

This study examines the relationship between retirement preferences, expressed as preferred retirement age, and actual retirement age in Sweden. The data were drawn from the Swedish Panel Survey of Ageing and the Elderly (PSAE). The PSAE was fielded in 2002 and 2003 with the aim of mapping living conditions of older people in Sweden. The data, which have a powerful longitudinal component, cover a broad spectrum of welfare indicators such as health, daily activities, social interactions, labour market and working conditions, and attitudes towards and experiences of retirement. Cox regression analysis was employed to explore whether and to what extent retirement preferences had any impact on actual retirement age. The results imply that retirement preferences do represent, in relation to other known key factors, an isolated influence on retirement patterns. The introduction of time-dependent variables strengthened this argument by showing how the “hazard” for the timing of retirement varied during the study period: those who preferred to retire close the end point of the study period were more likely to retire at this time than those who preferred to retire after the end of the study period and those who preferred to retire at the beginning of the study period. The results also indicated that the categories that wished to retire close to the beginning of the study period were more likely to retire at this point of time. The study thus provides empirical support for those researchers, debaters and policymakers who have addressed the importance of changing preferences towards later retirement in order to prolong working life.

This study examines the change of retirement preferences in the Swedish work force between 2002/2003 (t0) and 2010/2011 (t1). In 2003 a new pension system was introduced in Sweden. A central aim was to postpone retirement. Work incentives were strengthened by linking benefits more closely to the individual’s labour market participation. Also, older workers were given the right to work until age 67 which meant that age 65 was abandoned as the statutory age of retirement.

Drawing on cross-sectional data from the PSAE surveys in t0 and t1, the aim of the paper is to examine how retirement preferences developed between the time when the new system was about to be introduced and a time when it had been set in place. The study design has the character of a natural experiment. The main results show that there was substantial change in how retirement preferences were distributed in the two time-points.

In general, the 55–64 year-olds in t1 preferred to retire later than the same age group did in t0. The share of the older workforce which preferred to retire beyond 65 doubled and the increase was clustered around age 67. Most strikingly, this pattern applies to most sub-categories. Even those who reported poor health and poor work environment preferred to retire later in t1 than the corresponding category did in the preceding time-point. The results indicate that the strengthened work incentives and public campaigns to raise awareness of them have had a general impact on the older workforce in Sweden. The strong increase in age 67 as a preferred exit age indicates that the norm of suitable exit age is being delayed from 65 to 67.

The study analyses whether and to what degree specific routes into retirement affect older people, i.e. the relationship between heterogeneous exit patterns and post-retirement health and wellbeing. We used longitudinal data from two points in time; data related to t ₀ were collected in 1993, 1994, 1995 and 1996 and data related to t ₁ were collected in 2002 and 2003 (N = 589). We focused on older people (55+ at t ₁) who were employed at t ₀ and retired at t ₁. We used confirmative factor analysis to identify identical measures of health and wellbeing at both t ₀ and t ₁. Hence, we were able to control for pre-retirement health and wellbeing when evaluating the effects of different exit routes. These routes were defined as dependence on incomes from sickness benefit, disability pension, part-time pension, unemployment insurance and active labour market programmes. Our initial structural equation model showed a clear relation between exit routes and post-retirement wellbeing. People who prior to retirement were pushed into social benefit programmes related to health and unemployment were significantly worse off as retirees, especially those with health-related benefits. However, these relationships disappeared once pre-retirement wellbeing was added to the model. Our main conclusion is that post-retirement wellbeing first and foremost is a consequence of accumulation of advantages and disadvantages during the life course. Both labour market exit routes and post-retirement wellbeing can be seen as outcomes of this process. There are no independent effects of the retirement process. Judging from our findings, there is no reason to believe that involvement in social security programmes allowing early retirement on health grounds has any additional negative consequences for health and wellbeing.