Context

One of the most complex choices people need to make is how much to save for their retirement. For most people, this means making trade-offs between present day living standards and an uncertain financial future. While auto-enrolment into workplace pensions starting in 2012 have increased pension savings, there has been substantial debate among policy-makers and retirement industry about how to construct models that provide individuals with a clear definition of ‘retirement savings adequacy’ that they should aim for. Addressing this problem meaningfully can only be achieved with a full understanding of the financial circumstances of an individual saver and their household over the course of their lifetimes.

The report refers to a ‘saver population’, defined as working people who were earning enough from work to be eligible for automatic enrolment, excluding the highest earners.

The study

With support from the think tank Pheonix Insights, Nest Insight undertook this study starting in 2023 to understand retirement adequacy in the context of household balance sheets and circumstances across the lifetime.

The aim was to develop a model of optimal saving rates and in doing so, the study used mixed method study across three strands:

• Review: A formative review of relevant literature including economic research about lifecycle models of saving.
• Qualitative: Formative interviews and a roundtable discussion with experts from across the retirement savings sector.
• Quantitative: Scoping, analysis and model and scenario development using existing data sources to aid understanding of the varied composition of people’s household finance and projecting future financial wellbeing and retirement savings.
  • Analysis was based on panel survey data from UK’s Understanding Society, from 2009 to 2022 (waves 1–13). The analysis used three working-age samples: a main sample representative of the ‘saver population’, a sample of self-employed people with similar incomes, and a sample of paid employees ineligible for auto-enrolment.
  • Supported by NEST’s internal pension projection models, the survey data were used to model lifetime experiences (diverse income and financial outcomes) for 30 individuals and their households. These were illustrated based on a range of individual and household socioeconomic characteristics.

Key findings

• Key themes emerging: Many people were not saving enough for retirement, and there was a strong case for raising their contributions. However, many low-earner incomes were lower than required for minimum benchmarks for saving. Pressures on low-earner household were acute, extending beyond high-inflation to household debt, liquid savings, and difficulty paying bills. Such households could be demotivated to save for retirement and risk greater financial vulnerability if their retirement saving contributions were increased.
• How much a saver should save: The modelling enabled 30 contrasting, fictionalised retirement savings personas to be produced. These represented the diversity of people’s socio-demographic and financial circumstances, and enabled optimal retirement savings choices at each stage of their working lives to be identified. The optimal savings choices for the personas were equally diverse, however there were some observable patterns:
  • For above-median earners, the model often recommended savers increased savings to the maximum allowed in the model, of 15%.
  • Workers with just-below median earnings were recommended to save more than current defaults at varying levels.
  • Earnings at around the level of the full-time living wage resulted in recommendations for saving at the level of 5% at most, and sometimes 0%.
• Implications for the pensions system: It needs to be as adaptive as possible to people’s earnings levels, with universal guidance and messaging structured around income and age, flexible to people’s changing circumstances and needs, and more progressive than it currently use potentially through employer contributions.

Points to consider

Methodological strengths or limitations:

• The report presents the results of a complex and carefully-considered analysis and modelling process.
• Details (methods or results) are not given from the review and qualitative strands of research which informed development of the resulting model.
• Although the study analysed empirical data (data from real individuals), the subsequent modelling relied on certain assumptions being made and the personas were fictionalised, which introduced arbitrary influences on the results.
• The authors note that the optimised approach they identified for each persona was not the only possible optimised approach.

Applicability:

• The study should be of substantial interest to policymakers, regulators, employers, pension-scheme providers and researchers with an interest in workers’ retirement savings adequacy.

Relevance:

• The findings are important given the complexities and diversities of people financial circumstances and wellbeing and the resulting need for policymakers, employers and scheme providers to take a holistic view of people’s circumstances when they set saving policies and defaults.

Generalisability/transferability:

• With the inclusion of personas representing benefits recipients and the self-employed, there is important learning from the study which transfers to the wider working-age population beyond those eligible for auto-enrolment into a workplace pension in the UK.
• Although the study benchmarks against UK workplace pension defaults and thresholds, the study offers substantive and methodological learning that may transfer to other countries with similar economic contexts and workplace pension policies.