Dr. Karen Arulsamy, Health Economics Research Unit, University of Aberdeen
Prof. Liam Delaney, Department of Psychological and Behavioural Science, London School of Economics
A pdf version is available here.
Key Findings and Policy Recommendations
- Private sector employees with poor mental health are less likely to participate in workplace pension plans in the UK. This mental health gap in pension participation remains even after accounting for a range of other circumstances including income, occupation type, industry, and physical health.
- The implementation of automatic enrolment in 2012 onwards, based on an earnings threshold of £10,000 per year, equivalised the pension participation rates of employees with and without poor mental health, thus closing the mental health gap in pension participation.
- While most employees with poor mental health met the earnings threshold for automatic enrolment, they were significantly less likely to do so. Automatic enrolment policies should pay particular attention to the earnings threshold so as not to inadvertently exclude individuals with poor mental health who tend to earn less income, have lower access to workplace pension plans, and lower general savings.
Introduction
Ageing populations and longer life expectancies mean that more people are retiring and spending more years in retirement. This trend places an enormous burden on governments to ensure that people save adequately for their retirement and are financially secure in retirement. Individuals with mental health conditions are particularly at risk of financial vulnerability in retirement as they earn less income and experience higher unemployment levels[1].
The most commonly experienced mental health conditions i.e., anxiety and depression, can also lead to greater cognitive burden and ‘present bias’. Present bias is the tendency to prioritize immediate smaller rewards as opposed to larger later rewards.[2] These behavioural factors may make it harder for people with poor mental health to navigate the complexity of pensions decisions and delay the decision to participate in a plan.
If people with poor mental health are also less likely to participate in workplace pension plans, they face a greater risk of financial insecurity in retirement as these plans provide a crucial source of retirement income.
This paper, based on a study that was recently published at the Journal of Health Economics, outlines the relationship between mental health and pension participation in the UK and evaluates the impact of pensions automatic enrolment on this relationship. [3] This paper also discusses the implications of these findings and potential points to consider for Ireland’s pension enrolment scheme which will be introduced in 2024.
Automatic pension enrolment in the UK
In 2012 onwards, the UK implemented one of the largest pension policy reforms in the world, by introducing automatic enrolment nationwide.
Automatic enrolment made it mandatory for employers to set up access to workplace pension plans (if they were not previously available) and automatically enrol eligible employees into these plans.
Employees are eligible if they earn at least £10,000 per year (or the equivalent weekly amounts) and are between 21 and the state pension age. Employees who earn less than £10,000 per year are not automatically enrolled but could opt in.
Until 2018 the minimum contribution required under automatic enrolment was 2% of qualifying earnings, with a minimum of 1% contributed from the employer. In April 2018, this minimum increased to 5% of qualifying earnings, with a minimum of 2% from the employer. In April 2019, the minimum contribution increased to 8% of qualifying earnings with a minimum of 3% from the employer. This staggered increase addresses the tendency of people to stick with default contribution rates over time, thus ensuring that savings rates are commensurate with growth in income.[4]
Compliance with the policy was high with strict fines being imposed on employers who failed to enrol their employees by the assigned timelines.
The UK’s Department for Work and Pensions reports that over 10 million workers were automatically enrolled into a workplace pension plan as of April 2019[5]. The largest gains in participation were observed for women, employees in their 20s, and lower-income workers.[6] Approximately 10% of employees have opted-out of their workplace pension plan with older employees more likely to do so.[7]
Dataset and Findings
We use nationally representative panel data from the UK Household Longitudinal Study (UKHLS). The UKHLS dataset surveys the same group of people over time and collects extensive information on their personal, socioeconomic, and health circumstances.
The dataset captures whether the respondent participates in a workplace pension plan before and after automatic enrolment. We measure mental health using the General Health Questionnaire (GHQ-12) which is a clinically validated tool to assess the presence of psychological distress. While the GHQ-12 accurately predicts the risk of a mental health condition, it does not capture the type of condition for example, depression or generalized anxiety disorder.[8] The GHQ-12 assesses the risk of psychological distress via questions on general happiness and the ability to face problems, overcome difficulties, make decisions, and enjoy normal day to day activities.
Based on data from UKHLS as shown in Figure 1, we find that automatic enrolment led to a huge increase in the pension participation rates of employees in the private sector. While employees who are eligible for automatic enrolment experienced the greatest increase in participation, employees who are not eligible but could opt-in were also more likely to participate after the policy came into effect. These increases are consistent with other studies.
Figure 1: Pension participation rates among public sector and private sector (eligible and non-eligible) employees in the UK

Notes: The sample covers private and public sector employees aged 22 to 65 years old
First, we compare the relationship between poor mental health and pension participation before and after automatic enrolment. Male and female private sector employees with poor mental health are significantly less likely to participate in a workplace pension plan than employees with good mental health prior to the implementation of the policy. This mental health gap in pension participation exists even after accounting for income, occupation type, industry, ethnicity, marital status, presence of children and the presence of a physical health condition.
Thus, poor mental health predicts lower workplace pension participation independent of a large range of other circumstances. Although employees with poor mental health earn less than employees with good mental health, they are only slightly less likely to meet the earnings threshold of £10,000 per year. Moreover, fewer employees with poor mental health reported working in jobs that provided them with access to a plan prior to automatic enrolment.[9]
Notably, we find that automatic enrolment closed the mental health gap in pension participation in the UK. Based on Figure 2, private sector employees with poor and good mental health were equally likely to participate in a workplace pension plan after the implementation of automatic enrolment. While the mental health gap in pension participation exists for both male and female employees, it is larger among male employees. Nevertheless, automatic enrolment closes the mental health gap in participation for both male and female employees, as shown in Figure 3.
Figure 2: Pension participation rates by baseline psychological distress in Wave 1 (2009-2010) among private sector employees

Notes: The sample covers private employees aged 22 to 65 years old. Baseline psychological distress refers to individuals scoring 3 and above and no psychological distress refers to individuals scoring below 3 on the GHQ-12 scale in Wave 1 (2009-2010).
Figure 3: Pension participation rates by baseline psychological distress in Wave 1 (2009-2010) for male (left) and female employees in the private sector

Notes: The sample covers private employees aged 22 to 65 years old. Baseline psychological distress refers to individuals scoring 3 and above and no psychological distress refers to individuals scoring below 3 on the GHQ-12 scale in Wave 1 (2009-2010).
Summary and policy implications
Our study highlights the importance of systematically accounting for the effects of mental health when designing policies to increase retirement preparedness. The link between poor mental health and adverse economic outcomes means that people with mental health issues are likely to face greater challenges in attaining financial security. Individuals with poor mental health are particularly at risk of financial insecurity in retirement not only because they earn less income over their working lives but because they are also less likely to have access to and participate in workplace pension plans.[10] Based on the data used in this study, individuals with poor mental health are also less likely report other types of savings when they do not participate in a workplace pension plan implying that they have lower overall savings. Thus, retirement savings policies that do not consider how people with poor mental health might be affected by particular policy features may inadvertently exacerbate existing mental health disparities in economic outcomes. Importantly, our study highlights the role of pensions automatic enrolment in reducing disparities in participation amongst vulnerable individuals in addition to increasing overall pension coverage.
Our findings are currently relevant to the Irish pensions landscape in light of a recent announcement by the government of Ireland to implement pensions automatic enrolment in 2024 onwards. Based on this announcement, employees between 23 and 60 years old and earning over €20,000 will be automatically enrolled into a workplace pension plan. We highlight implications from our findings and key points to consider in the implementation and monitoring of this policy:
- As the earnings threshold for automatic enrolment in Ireland is higher, consideration should be given as to whether people with poor mental health are less likely to meet this threshold and whether they would benefit from being auto enrolled. A thorough assessment on how automatic enrolment might affect employees with poor mental health can be done by analysing linked administrative earnings and mental health data. Qualitative interviews conducted amongst individuals with mental health issues may also shed some light on how the current design and implementation of automatic enrolment may affect them.
- Ireland is adopting a staggered increase in contribution rates such as implemented by the UK. Employer and employee contributions will start at 1.5% of gross salary and increase automatically every three years with a maximum contribution rate of 6% in 2034. An evaluation of how different sub-populations including those with poor mental health are likely to stick with default contribution rates can guide future initiatives to increase savings amongst those who can afford to save more.
- The effects of the pandemic and current cost of living crisis have resulted in severe financial strain on households, particularly those who are already vulnerable. Recent evidence in the UK indicates a long-term increasing trend in opt-out behaviour since 2020. In August 2022, the proportion of newly enrolled employees who opted-out of their pension plan was 10.4 per cent compared to 7.6 per cent in January 2020. However, the proportion of people who are still automatically enrolled in their workplace pension plan but stopped saving (savings cessation) remained relatively low.[11] As people with poor mental health tend to experience financial hardship to a greater extent, they may benefit more from pausing contributions to workplace pension plans during financial strain.[12] Procedures that allow for temporary savings cessation as opposed to opting-out of a plan during periods of financial distress could potentially help vulnerable individuals save more in the long-term.
References:
Arulsamy, Karen, The Impact of Adolescent Psychological Distress on Access and Participation in Employer Sponsored Pension Plans in the US, 2022
Arulsamy, Karen, and Liam Delaney, ‘The Impact of Automatic Enrolment on the Mental Health Gap in Pension Participation: Evidence from the UK’, Journal of Health Economics, 2022, 102673
Benartzi, Shlomo, and Richard H. Thaler, ‘Behavioral Economics and the Retirement Savings Crisis’, Science, 339.6124 (2013), 1152–53
Cella, Matteo, Simon Dymond, and Andy Cooper, ‘Impaired Flexible Decision-Making in Major Depressive Disorder’, Journal of Affective Disorders, 124.1–2 (2010), 207–10
Cribb, Jonathan, and Carl Emmerson, ‘What Happens to Workplace Pension Saving When Employers Are Obliged to Enrol Employees Automatically?’, International Tax and Public Finance, 27.3 (2020), 664–93
Department for Work and Pensions, ‘Automatic Enrolment Evaluation Report 2019’, 2020.
Department for Work and Pensions, ‘Ten Years of Automatic Enrolment in Workplace Pensions: Statistics and Analysis’, 2022.
Department for Work and Pensions, ‘Workplace Pension Participation and Savings Trends of Eligible Employees: 2009 to 2021’, 2022.
Katie Evans, Merlyn Holkar, and Nic Murray, ‘Overstretched, Overdrawn, Underserved’, 2017.
Goldberg, David, and Paul Williams, ‘General Health Questionnaire’ (Granada Learning Group, 1988)
Goodman, Alissa, Robert Joyce, and James P. Smith, ‘The Long Shadow Cast by Childhood Physical and Mental Problems on Adult Life’, Proceedings of the National Academy of Sciences, 108.15 (2011), 6032–37
Maloney, Erin A., Jason R. Sattizahn, and Sian L. Beilock, ‘Anxiety and Cognition’, Wiley Interdisciplinary Reviews: Cognitive Science, 5.4 (2014), 403–11
Mousteri, Victoria, Michael Daly, Liam Delaney, Per Tynelius, and Finn Rasmussen, ‘Adolescent Mental Health and Unemployment over the Lifespan: Population Evidence from Sweden’, Social Science & Medicine, 222 (2019), 305–14
Xia, Lisheng, Ruolei Gu, Dandan Zhang, and Yuejia Luo, ‘Anxious Individuals Are Impulsive Decision-Makers in the Delay Discounting Task: An ERP Study’, Frontiers in Behavioral Neuroscience, 11 (2017), 5
Zhao, Jinling, Jiuqing Cheng, Mary Harris, and Ronaldo Vigo, ‘Anxiety and Intertemporal Decision Making: The Effect of the Behavioral Inhibition System and the Moderation Effects of Trait Anxiety on Both State Anxiety and Socioeconomic Status’, Personality and Individual Differences, 87 (2015), 236–41
[1][1] Alissa Goodman, Robert Joyce, and James P. Smith, ‘The Long Shadow Cast by Childhood Physical and Mental Problems on Adult Life’, Proceedings of the National Academy of Sciences, 108.15 (2011), 6032–37; Victoria Mousteri and others, ‘Adolescent Mental Health and Unemployment over the Lifespan: Population Evidence from Sweden’, Social Science & Medicine, 222 (2019), 305–14.
[2] Matteo Cella, Simon Dymond, and Andy Cooper, ‘Impaired Flexible Decision-Making in Major Depressive Disorder’, Journal of Affective Disorders, 124.1–2 (2010), 207–10; Erin A. Maloney, Jason R. Sattizahn, and Sian L. Beilock, ‘Anxiety and Cognition’, Wiley Interdisciplinary Reviews: Cognitive Science, 5.4 (2014), 403–11; Lisheng Xia and others, ‘Anxious Individuals Are Impulsive Decision-Makers in the Delay Discounting Task: An ERP Study’, Frontiers in Behavioral Neuroscience, 11 (2017), 5; Jinling Zhao and others, ‘Anxiety and Intertemporal Decision Making: The Effect of the Behavioral Inhibition System and the Moderation Effects of Trait Anxiety on Both State Anxiety and Socioeconomic Status’, Personality and Individual Differences, 87 (2015), 236–41.
[3] Karen Arulsamy and Liam Delaney, ‘The Impact of Automatic Enrolment on the Mental Health Gap in Pension Participation: Evidence from the UK’, Journal of Health Economics, 2022, 102673.
[4] Shlomo Benartzi and Richard H. Thaler, ‘Behavioral Economics and the Retirement Savings Crisis’, Science, 339.6124 (2013), 1152–53.
[5] Department for Work and Pensions, ‘Automatic Enrolment Evaluation Report 2019’, 2020.
[6] Jonathan Cribb and Carl Emmerson, ‘What Happens to Workplace Pension Saving When Employers Are Obliged to Enrol Employees Automatically?’, International Tax and Public Finance, 27.3 (2020), 664–93; Department for Work and Pensions, ‘Workplace Pension Participation and Savings Trends of Eligible Employees: 2009 to 2021’, 2022.
[7] Department for Work and Pensions, ‘Ten Years of Automatic Enrolment in Workplace Pensions: Statistics and Analysis’, 2022.
[8] David Goldberg and Paul Williams, ‘General Health Questionnaire’ (Granada Learning Group, 1988).
[9] Within the private sector, 82.4% of employees without poor mental health meet the threshold, while 78.3% of employees with poor mental health do. We also find evidence that men with poor mental health may be more likely to work in jobs that do not provide access to a plan. 64.7% of those without poor mental health worked somewhere which offered a scheme, compared to 61.3% of those with poor mental health.
[10] Goodman, Alissa, Robert Joyce, and James P. Smith. “The long shadow cast by childhood physical and mental problems on adult life.” Proceedings of the National Academy of Sciences 108, no. 15 (2011): 6032-6037.
Karen Arulsamy, The Impact of Adolescent Psychological Distress on Access and Participation in Employer Sponsored Pension Plans in the US, 2022.
[11] Department for Work and Pensions, ‘Ten Years of Automatic Enrolment in Workplace Pensions: Statistics and Analysis’, 2022.
[12] Katie Evans, Merlyn Holkar, and Nic Murray, ‘Overstretched, Overdrawn, Underserved’, 2017.