Post-retirement work: A path to extending working life that generates new inequalities? A comparison of Sweden, the USA and Japan
It is increasingly common for people to be in paid work after retirement. This is partly a result of governments in many advanced economies implementing reforms to extend working lives. Previous studies suggest that wealthier and healthier retirees are more likely to have paid work in retirement. However, little research has explored the implications of expansion of post-retirement work for inequalities in relation to education, income and health. Similarly, research is lacking on the extent to which inequalities related to post-retirement work vary between countries with different institutional frameworks.
This project will provide state-of-the-art evidence concerning the impact of post-retirement paid work on inequalities in later life in Sweden, the USA and Japan, three countries where retirement tends to occur late in life. These countries have succeeded in attaining high employment rates for older people despite contrasting work and retirement contexts (pension and employment policies, welfare state arrangements). Data analysis will be carried out using comparable longitudinal ageing surveys of older people from each country. A comparison of these contrasting national settings will shed light on the mechanisms through which post-retirement work generates inequalities. This information may be used to develop more equitable and effective policies, and in this way encourage labour market participation in old age.
Final report
Project goals
The project aimed to answer the following research questions:
• How does gender, health or education level affect access to and the quality of post-retirement work?
• What are the implications of post-retirement paid work for income inequalities in later life?
• Are there intercountry variations in these processes and can they be understood in terms of institutional and labour market factors?
Project implementation
The analyses used survey data from Japan, Sweden and the United States, specifically the Japanese Study of Ageing and Retirement (JSTAR), eight triennial cross-sectional nationally representative surveys for Japan, the Swedish Longitudinal Occupational Survey of Health (SLOSH) and the American Health and Retirement Study (HRS).
The project resulted in four published empirical papers and a theoretical paper which will be submitted for publication in 2023/4. One published paper employed a harmonized set of variables created from information in the JSTAR, HRS and SLOSH datasets in order to examine differences by age in psychosocial and physical job quality and job satisfaction (random-effects modelling). With three countries included, we could look for inter-country differences. In a second published paper using the SLOSH data only, we carried out differencing analyses using more sophisticated scales of work quality in order to examine within-individual change in job quality and job satisfaction for workers in their sixties. In a third published paper using the HRS data, we examined the impact of participation in bridge job employment (taking up a new job) versus retiring directly on wealth inequalities in old age. In a first step, we compared the distribution of pre-retirement wealth among older adults who have bridge jobs and those who retire directly. In a second step, we examined whether participation in a bridge job moved individuals up or down the wealth distribution. The last published paper used eight Japanese triennial cross-sectional nationally representative surveys. Slope and relative indexes of inequality were used to model inequalities in self-rated health based on household income over the period 1986–2013.
Key results and conclusion
The very oldest workers are not segregated into poor quality jobs: We found that job quality was better among people who were working after pensionable age than for those who were still in their fifties in Japan, Sweden and the United States. Specifically, post-pensionable-age jobs were generally less demanding, more rewarding, with better schedule control, and more satisfying than jobs held by younger workers. These findings held regardless of gender and education level. The job quality–age relationship tended to be steeper at pensionable ages, indicating an effect of institutional clocks of pension and retirement rules. There did not appear to be intercountry variations in these processes, which are probably universal in countries with developed welfare states.
Swedish workers in their sixties experienced improvements in working conditions as they aged: Following-up workers in their sixties, we found that their physical and psychosocial working conditions and job satisfaction improved. Improvements were particularly marked for workers who reduced their hours or took up a bridge job (switching sector, becoming self-employed and switching to a temporary contract) but did not differ by gender.
Patterns of selection into post-retirement work suggests that this work may impact income inequalities, but we could not demonstrate impacts on wealth inequalities: The main impact of education and gender was in who remained in work at ages when working conditions tended to be better: rates of participation in these jobs were often lower for women and for the less well educated. Wealth inequalities later in life in the United States, which are pronounced, looked similar between those who transitioned to bridge employment and those who exited the workforce directly. It appears that bridge employment does not significantly exacerbate (or mitigate) wealth inequalities later in life.
Taken together, these findings suggest that two mechanisms may cause better working conditions for the oldest workers: (1) differential selection of older workers with poorer working conditions into retirement and (2) the oldest workers experiencing improvements in their working conditions. For those older workers able to stay in paid work, the agency of these workers is enhanced as they approach and exceed pensionable age. It is possible that, after pensionable age, lessened financial and symbolic penalties of retirement alter the balance of power between employers and employees, giving workers more leverage to craft jobs that suit themselves better.
New research directions
These empirical findings challenge existing perspectives in social gerontological scholarship about the nature of the workplace for late career workers. They undermine the usual conceptualization of the late career as a single period extending from around 50/55 years to retirement. To the contrary, psychosocial working conditions of jobs held subsequent to attaining ages of eligibility for state pensions are distinctive. The empirical findings from this project highlight how workers older than pensionable age are decommodified from the labour market and are no longer obliged to work, which has consequences for the balance of power with employers who may have to make greater accommodations for post-pensionable-age workers.
These results have inspired our theoretical work currently in progress that suggests the late career is better conceptualized in two phases—prior to and post-pensionable age—and that after reaching pensionable age older workers are managing to strategically select better working conditions. The results offer intriguing insights into the nature of trade-offs that may take place in the late career: gains in perceived adequacy of pay provide preliminary evidence arguing against this characteristic being traded off against other job characteristics such as job satisfaction. Some of the project partners are exploring these avenues in a new Swedish Research Council-funded project.
Dissemination and impact
Media coverage
One publication from the project (Platts et al.) was jointly published with a paper from a Forte-funded project. Print, web and broadcast media coverage of these papers included Svenska Dagbladet, SVT Nyheter, Aftonbladet and forskning.se.
Conferences
Findings from the project were presented at conferences in Europe, North America and Asia. Conferences included: International Association of Gerontology and Geriatrics – European Region conference (2019), MIRAI conference (2019), Annual Conference of the Southern Economic Association (2019), Nordic Gerontological Congress (2020 & 2022), The Osaka University Public Health Seminar (2020), Annual Scientific Meeting of the Gerontological Society of America (2020, 2021 & 2023), and the Annual Conference of the Western Economic Association International (2023).
The project aimed to answer the following research questions:
• How does gender, health or education level affect access to and the quality of post-retirement work?
• What are the implications of post-retirement paid work for income inequalities in later life?
• Are there intercountry variations in these processes and can they be understood in terms of institutional and labour market factors?
Project implementation
The analyses used survey data from Japan, Sweden and the United States, specifically the Japanese Study of Ageing and Retirement (JSTAR), eight triennial cross-sectional nationally representative surveys for Japan, the Swedish Longitudinal Occupational Survey of Health (SLOSH) and the American Health and Retirement Study (HRS).
The project resulted in four published empirical papers and a theoretical paper which will be submitted for publication in 2023/4. One published paper employed a harmonized set of variables created from information in the JSTAR, HRS and SLOSH datasets in order to examine differences by age in psychosocial and physical job quality and job satisfaction (random-effects modelling). With three countries included, we could look for inter-country differences. In a second published paper using the SLOSH data only, we carried out differencing analyses using more sophisticated scales of work quality in order to examine within-individual change in job quality and job satisfaction for workers in their sixties. In a third published paper using the HRS data, we examined the impact of participation in bridge job employment (taking up a new job) versus retiring directly on wealth inequalities in old age. In a first step, we compared the distribution of pre-retirement wealth among older adults who have bridge jobs and those who retire directly. In a second step, we examined whether participation in a bridge job moved individuals up or down the wealth distribution. The last published paper used eight Japanese triennial cross-sectional nationally representative surveys. Slope and relative indexes of inequality were used to model inequalities in self-rated health based on household income over the period 1986–2013.
Key results and conclusion
The very oldest workers are not segregated into poor quality jobs: We found that job quality was better among people who were working after pensionable age than for those who were still in their fifties in Japan, Sweden and the United States. Specifically, post-pensionable-age jobs were generally less demanding, more rewarding, with better schedule control, and more satisfying than jobs held by younger workers. These findings held regardless of gender and education level. The job quality–age relationship tended to be steeper at pensionable ages, indicating an effect of institutional clocks of pension and retirement rules. There did not appear to be intercountry variations in these processes, which are probably universal in countries with developed welfare states.
Swedish workers in their sixties experienced improvements in working conditions as they aged: Following-up workers in their sixties, we found that their physical and psychosocial working conditions and job satisfaction improved. Improvements were particularly marked for workers who reduced their hours or took up a bridge job (switching sector, becoming self-employed and switching to a temporary contract) but did not differ by gender.
Patterns of selection into post-retirement work suggests that this work may impact income inequalities, but we could not demonstrate impacts on wealth inequalities: The main impact of education and gender was in who remained in work at ages when working conditions tended to be better: rates of participation in these jobs were often lower for women and for the less well educated. Wealth inequalities later in life in the United States, which are pronounced, looked similar between those who transitioned to bridge employment and those who exited the workforce directly. It appears that bridge employment does not significantly exacerbate (or mitigate) wealth inequalities later in life.
Taken together, these findings suggest that two mechanisms may cause better working conditions for the oldest workers: (1) differential selection of older workers with poorer working conditions into retirement and (2) the oldest workers experiencing improvements in their working conditions. For those older workers able to stay in paid work, the agency of these workers is enhanced as they approach and exceed pensionable age. It is possible that, after pensionable age, lessened financial and symbolic penalties of retirement alter the balance of power between employers and employees, giving workers more leverage to craft jobs that suit themselves better.
New research directions
These empirical findings challenge existing perspectives in social gerontological scholarship about the nature of the workplace for late career workers. They undermine the usual conceptualization of the late career as a single period extending from around 50/55 years to retirement. To the contrary, psychosocial working conditions of jobs held subsequent to attaining ages of eligibility for state pensions are distinctive. The empirical findings from this project highlight how workers older than pensionable age are decommodified from the labour market and are no longer obliged to work, which has consequences for the balance of power with employers who may have to make greater accommodations for post-pensionable-age workers.
These results have inspired our theoretical work currently in progress that suggests the late career is better conceptualized in two phases—prior to and post-pensionable age—and that after reaching pensionable age older workers are managing to strategically select better working conditions. The results offer intriguing insights into the nature of trade-offs that may take place in the late career: gains in perceived adequacy of pay provide preliminary evidence arguing against this characteristic being traded off against other job characteristics such as job satisfaction. Some of the project partners are exploring these avenues in a new Swedish Research Council-funded project.
Dissemination and impact
Media coverage
One publication from the project (Platts et al.) was jointly published with a paper from a Forte-funded project. Print, web and broadcast media coverage of these papers included Svenska Dagbladet, SVT Nyheter, Aftonbladet and forskning.se.
Conferences
Findings from the project were presented at conferences in Europe, North America and Asia. Conferences included: International Association of Gerontology and Geriatrics – European Region conference (2019), MIRAI conference (2019), Annual Conference of the Southern Economic Association (2019), Nordic Gerontological Congress (2020 & 2022), The Osaka University Public Health Seminar (2020), Annual Scientific Meeting of the Gerontological Society of America (2020, 2021 & 2023), and the Annual Conference of the Western Economic Association International (2023).