Search

Increases in retirement ages make it particularly pressing to better understand how long people will work. Working life expectancy (WLE) is a useful measure for this and the current paper assesses the tools, that is, software packages, available to assess it. We do this using data from the English Longitudinal Survey on Ageing (ELSA, 2003-2018) and multistate models to estimate WLE stratified by sex and socioeconomic status. Men's versus women's WLEs were slightly higher at all ages. Estimates were similar in ELECT and SPACE by both sex and socioeconomic status. WLEs were comparatively higher from IMaCh, ranging from approximately 0.28 to 1.49 years. Life expectancy estimates from IMaCh were also higher compared to SPACE and ELECT. Using multistate models to estimate WLE provides a useful indication of the actual expected length of working life. More research is needed to better understand why estimates from IMaCh differed from ELECT and SPACE.

Longer life expectancy and fertility decline have increased concerns about the security of old-age pensions. Raising retirement ages is one strategy to offset rising costs, though the option to retire varies considerably by socioeconomic status (SES) and sex. In terms of SES, the level of variation may depend on the measure used. Also, many workers now transition into retirement slowly, e.g., move from full- to part-time work. Thus, retirement age may not sufficiently capture how long people work. Working life expectancy (WLE)—the expected average number of years worked—better measures total working life. We use data from the Swedish Longitudinal Occupational Survey of Health (SLOSH) from 2008 to 2020 (n=4,940 people age 50+ and n=74,093 person-observations) to examine WLE by education and occupation. We estimate a three-state multistate model (i.e., working, not working, dead) and a four-state model (working part-time, working full-time, not working, and dead); both assume a continuous-time first-order Markov process. We estimate two sex-stratified models, cross-classified by: 1) occupation; and 2) education. We find that professionals work full-time 1 year more than routine workers, regardless of sex. The low educated work full-time 1 year less than the highly educated. In our weighted three-state model, where part-time work contributed ½ of full-time work, the difference increased to 1.14 and 1.05 years, respectively. Our unweighted three-state model showed slightly larger education differences. Findings suggest that WLE differs by SES, regardless of sex, and the differences are greater by education than occupation. This has implications for extending working life policies.

Background: In childhood, relations with siblings and friends lie at the core of social interaction. Lacking either type of relationship may reflect lower levels of social support. While social support is known to be negatively associated with premature death, there are still no long-term follow-ups of mortality risks among children without siblings (‘only-children’) and children without friends (‘lonely-children’). The aim of the present study was therefore to examine and compare all-cause mortality in these two groups. Methods: Cox regression analysis was based on a Stockholm cohort born in 1953 (n = 15,117). Individuals were identified as only-children if there were no records of siblings before age 18. Derived from sociometric data collected at age 13, lonely-children were defined as not being nominated by classmates as one of three best friends. The follow-up of all-cause mortality covered ages 20-56. Results: Both only-children and lonely-children had increased risks of premature mortality. When adjusted for a wide range of family-related and individual factors, the risk ratio for only-children increased in strength whereas the risk ratio for lonely-children was reduced. The former finding may be explained by suppressor effects: for example, both only-children and those whose parents had alcohol problems had higher mortality risks but only-children were less likely to have parents with alcohol problems. The latter finding was primarily due to adjustment for scholastic ability. Conclusions: It is concluded that while only-children and lonely-children have similar risks of all-cause mortality, the processes leading up to premature death appear to be rather different. Yet, interventions targeted at improving social learning experiences may be beneficial for both groups. Key messages: Only-children have higher risks of premature mortality but the mechanisms remain unclear. Lonely-children are at risk of premature mortality primarily due to poorer scholastic abilit, Reproduction of inequality through linked lives (RELINK), Ofärdens samgång: Hur klustrar sig sociala, ekonomiska och hälsomässiga villkor i befolkningen över tid och vad påverkas av samt påverkar klustertillhörighet?

Reproduction of inequality through linked lives (RELINK)