Changes across Cohorts in Work-Family Histories and Extended Working Lives: Evidence from the United Kingdom.

Giorgio Di Gessa, King''s College
Rachel Stuchbury, Celsius, UCL
Karen Glaser, King''s College London
Laurie Corna, King''s College London
Loretta Platts, Stress Research Institute Stocholm
Diana Worts, University of Toronto
Peggy McDonough, University of Toronto
Amanda Sacker, University College London
Debora Price, University of Manchester

Introduction: Late life labour force participation is influenced by contemporaneous factors such as health and economic status, but also by earlier work and family experiences.

Aim: We examined changes in the relationship between typologies of work and family histories and working up to and beyond State Pension age (SPa) across three 10-year UK birth cohorts.

Data and Methods: Employing data from the 1988/89 British Retirement Survey (RS), the 1998 British Household Panel Study (BHPS), and the 2008 English Longitudinal Study of Ageing (ELSA) we examined each cohort’s circumstances between the ages of 60 – 69 for men and 55 – 64 for women (i.e. respondents aged 55-69 born 1919-1933 in the RS, 1929-1943 in the BHPS, and 1939-1953 in ELSA). We considered current labour market status, socio-economic conditions, health, as well as retrospective labour market and family histories. The associations between labour market experiences and participation in paid work up to and beyond State Pension age (SPa) were analysed using logistic and multinomial regression separately for men and women.

Results: Respondents (and women in particular) who had a strong attachment to the labour market throughout their lives were more likely to be in paid work in later life compared to those with prior weak attachment. These relationships have strengthened across cohorts: recent cohorts of respondents were more likely to have worked throughout their lives, and to remain in the labour market at older ages.


Introduction

Our research investigates whether the relationship between family and labour market experiences and working until or beyond SPa has changed across cohorts in the UK. This is a critical issue as many industrialised countries are now pursuing polices to extend working lives in response to the fiscal challenges posed by population ageing. Although later-born cohorts are more likely to be in paid work in their 50s and early 60s in comparison to earlier cohorts, little is known about whether and to what extent these changes are driven by variations in the health and socio-economic profiles of later-born cohorts, or by life-course events pertaining to family and work experiences.

Data and Methods

Study Population. We employed three longitudinal nationally representative datasets: (1) the 1988/89 British Retirement Survey (RS); (2) the 1998 British Household Panel Survey (BHPS), and (3) the 2008 English Longitudinal Study on Ageing (ELSA). Each dataset includes a wealth of information on current labour market activity, health (physical and psychological health), socio-economic circumstances and retrospective employment, marital and parenthood histories. Our analytic sample is based on those aged 55-69 at the time of the survey, that is respondents born between: 1919-1933 in RS; 1929-1943 in BHPS; and 1939-1953 in ELSA.

Key indicators. Current labour market activity. We considered whether respondents were in paid work in the respective survey year, distinguishing between full-time (FT) and part-time (PT) employment for women. Lifecourse factors. Combining the annual interviews with the retrospective histories, we derived age-specific state variables indicating a respondent’s labour market activity, marital status and parental status between the ages of 16 and 54 years (for women), and between 16 and 59 (for men). We used optimal matching analysis to derive frequent and meaningful patterns of experience for individuals’ labour market, marital and parental histories, treating the whole life course as the unit of analysis. In our final models, we considered three ideal employment histories for men (employed FT throughout up to 59; not employed throughout; early exit at 49) and five for women (employed FT throughout up to 54; not employed throughout; early exit at 49; employed PT throughout with a career break between 26 and 32; employed FT throughout with a career break between 23 and 34). Parental histories were summarised using three main categories (no children; one or more children at about age 23; one or more children after age 30), whereas marital histories captured whether respondents ever experienced divorce and/or widowhood throughout their lifecourse. Our analyses also adjusted for age, self-reported health, long-standing illness, mobility limitations, highest educational qualification, housing tenure, caring responsibilities, whether respondents were eligible for an occupational pension, and partners’ labour market activity.

Methods. The associations between work-family experiences and participation in paid work in later life were analysed using multinomial and logistic regression techniques. Analyses were performed using STATA 13.

Results

Descriptive Findings. As expected the percentage of men (aged 60-69) and women (aged 55-64) in paid work was higher among those born in the more recent cohorts. The distribution of work and family histories also changed significantly across time. For instance, the percentage of women who were mostly out of the labour market between the ages of 16 and 54 years decreased from around 30% of those born between 1924-34 to 17% of those born twenty years later.

Multivariate analyses. Multivariate models suggest that respondents with a strong attachment to the labour market throughout their lives were more likely to be in paid work (either FT or PT) in later life compared to those with weak attachment to the labour market. For example, women born between 1944-53 with continuous full-time employment histories had a 24% predicted probability of being in paid work at age 64 compared to a predicted probability of 4% among those who had been non-employed or family carers. However, for women who had not been employed, those in the youngest cohort were more likely to be working 20 or more hours per week at age 60 (12%) compared to those in the oldest cohort (4%). Divorced women were disproportionately represented in this group suggesting that those who extend paid work may be compensating for the loss of a spouse’s pension and/or for earlier periods out of the labour market.

Conclusion

Our study provides support for the hypothesis that early work-family histories play an important role in determining who works up to and beyond SPA. Both men and women who had worked most of their lives were more likely to be in paid work compared to those who had weak attachment to the labour market; moreover, this relationship appears to have strengthened across cohorts. However, findings for the most recent cohort of women suggest that those who experienced divorce are more likely to work up to and beyond SPa, perhaps making up for reduced income in later life.

Presented in Session 1133: Life Course