Context

Automatic enrolment into pensions is a policy in many countries which is intended to increase pension provision and household net wealth during the accumulation phase. The UK Pensions Act 2008 provided for the roll-out of automatic enrolment into workplace pensions for UK employees aged 22 to State Pension Age. It started with the largest firms in 2012, and included the randomised roll-out for around 160,000 UK employers with 2-29 employees (the focus of the current study) between June 2015 and April 2017.

Though workers have the right to opt-out of their pension, previous research has shown that, through low opt-out rates, automatic enrolment increases pension participation rates and average pension savings. Nonetheless, this could be offset by changes in financial behaviour in other domains, including borrowing.

The study

This study, undertaken for Nest Insight, analysed the impact of automatic pensions enrolment on the borrowing behaviour of households using linked administrative pension from Nest (the UK’s largest provider of automatic enrolment pensions) and credit-file data from Experian UK. The analysis exploited the roll-out of automatic enrolment to UK employers with 2-29 employees which, being randomised, provided a large-scale natural experiment with a baseline sample of 649,747 employees from 161,707 employers.

Econometric analysis was undertaken, using regression methods with robust standard errors clustered at the employer level and variables obtained for individual employees:

• Age, gender, estimated individual gross annual income, observation date.
• Monthly pensionable pay, employer contributions, employee contributions, tax relief, pension balances, monthly mortgage payment due date.
• Debt outcomes: total unsecured debt (comprising revolving and non-revolving debt), total mortgage debt, , total vehicle loan debt.
• Credit worthiness outcomes: recent default, credit score, recent bankruptcy.

Key findings

• Benchmark impacts on pension savings: over the first 41 months after auto-enrolment, each additional monthincreased the average savings automatically enrolled in workplace pensions by £32-£38 (p<.01), of which £16-£19 were employer contributions, £13-£15 were employee contributions, and £3-£4 were tax credits. The ranges given reflect model variations with different samples.
• There were negative impacts on some measures of debt such that each additional month increased:
  • the amount in total unsecured debts by £7 (p<.01), which was 19-23% of the increase in total pension savings;
  • the likelihood of having a mortgage by 0.05 percentage points (p<.01), against a baseline prevalence of 38% and a cumulative 1.9 percentage points after 41 months; and
  • mortgage balances by £118 (p<.01).
• There were positive impacts on other measures such thateach additional monthimproved:
  • the likelihood of loan default in the last six years, which reduced by 0.04 percentage points (p<.01), a cumulative decrease of 1.6 percentage points after 41 months; and
  • credit scores, which increased by 0.3 points (or 0.002 standard deviations, p<.01) per month, a cumulative increase of 0.07 standard deviations after 41 months.
• There were no significant effects on other measures: auto-enrolment on vehicle debt or bankruptcy (p>.10).

Points to consider

Methodological strengths or limitations

• The authors note that there is a small amount of inaccuracy in the timing of actual staging of firms into the policy compared with the staging dates available in the data but expect the effect of this on the analysis to be random and minimal.
• The authors also note that the empirical nature of the study makes it difficult to distinguish the reasons for why the observed changes arose.
• The study uses large scale samples and robust multivariate methods of analysis, and natural randomisation of enrolment into the pensions scheme supports the attribution of the observed differences to auto-enrolment.

Applicability

• This study should interest policy makers in the UK and internationally, especially in relation to the potential unintended financial consequences for individuals of auto-enrolment into workplace pension schemes.

Relevance

• The study is relevant given the extensive reach of workplace pension auto-enrolment policy.
• The study is also relevant given the Money and Pensions Service’s national goal, set out in the 2020 UK Strategy for Financial Wellbeing, for five million more people by 2030 to feel they understand enough to plan for and in later life.

Generalisability/transferability

• Generalisability of the findings to other socio-economic groups is limited in so far as the policy applies only to employees.
• However, learning from the study is likely to have transferability to employees of working age in larger firms in the UK and in similar auto-enrolment workplace pension schemes in countries with comparable economies.