Pensions auto enrolment is rightly held up as a shining example of a policy intervention that achieved a vast behaviour change across a population. Since the 2010s, most UK workers are saving for retirement automatically, instead of having to opt in. Thanks to this single powerful nudge, millions of people who weren’t contributing to a pension are now doing so. But does this really mean that people are saving more overall?
At first glance, the answer is, obviously, yes. The Department for Work and Pensions estimated that annual pension contributions had increased by over £32bn in 2021 compared to 2012, as a result of the policy.[1] Even so, the overall impact on savings levels depends on what else people did when they were automatically enrolled.
When a worker is enrolled by their employer, they generally start paying employee pension contributions out of their pay. This means their take-home pay goes down a little. There’s an underlying assumption behind the auto enrolment policy, that people account for this reduction in take-home pay by spending a little less. And if they can’t afford to do that, they’ll opt out of pension saving.
But what if, instead, people cover the cost of their contributions by borrowing more, or by reducing the amounts they were previously paying into non-pensions savings? If someone funds 100% of their contributions through one or both of these methods, the net amount they’re saving will be unchanged.
We’ve been working with a group of UK and US researchers who share a common goal of understanding the full effect that auto enrolment has on people’s finances. We’re learning that things are more complex than might have been expected when auto enrolment began. People’s spending, other saving and borrowing can all be affected.
More work is still needed, but it’s already clear that if an auto enrolment system is to work for everyone, it must take into account the wider effects that payroll contributions have on people’s finances, especially for people on lower incomes.
Borrowing to save?
In the most recent of these studies[2], we collaborated with leading researchers from universities including Harvard, Yale, Nottingham and Warwick Business School at the University of Warwick. The research was made possible with the support of the BlackRock Foundation and the Money and Pensions Service, and enabled by Nest Insight’s emergency savings programme that was also supported by JPMorgan Chase.
Together, we looked at the roll-out of auto enrolment at employers with fewer than 30 workers, which took place between 2015 and 2017. The timing of this roll-out was randomised at the employer level, meaning we could see with a high degree of accuracy the effects of enrolment on the financial experiences of the hundreds of thousands of people who were enrolled into Nest at this time. At this phase in the auto enrolment roll-out, the default minimum contribution was only 2% of income, with 1% usually being paid by the employee.
Working with Experian and Nottingham researchers, we used innovative techniques to match up pension and credit records while preserving people’s data privacy. We found that for each additional month of enrolment, on average:
- each worker’s savings in Nest rose by £32 and their total pension savings by an estimated £38
- their unsecured debt, including overdrafts and personal loans, went up by £7
- they became slightly more likely to take out a mortgage (0.05 percentage points), and the average mortgage balance rose by £118
- their credit scores rose a little, and their likelihood of defaulting on a debt went down.
Our findings are consistent with a parallel study[3] by Taha Choukhmane and Christopher Palmer from the Massachusetts Institute of Technology (MIT). They used banking data to explore the impact of increases to minimum auto enrolment contributions that occurred in the two years following our study. They found that, for every £1 reduction in monthly take-home pay due to higher pension contributions:
- people cut their spending by £0.34, mainly on leisure and eating out
- those with deposit account balances saw these go down
- those with credit card balances saw them increase.
Significantly, those with lower initial deposit balances cut their spending the most, while those with significant liquid savings draw down their deposits.
Getting the full picture
It’s clear from these studies that people’s borrowing and other saving were affected by being enrolled. Yet the story is not as simple as ‘people borrowed to save’. The effect on mortgages, for instance, suggests that for some people, being enrolled may have encouraged some to take another step on their financial journeys. This could be an example of a ‘wealth effect’. When people have more assets, such as pension savings, they’re more likely to change their behaviour in relation to money more broadly.
It’s also worth noting that the people who borrowed through overdrafts or loans tended to be those with existing credit balances. They also tended to be those with above average credit scores. In some cases, of course, this will be because they were the ones with access to credit, but this won’t have been the case for all. The results suggest that the people who borrowed were mostly those who could afford to.
Perhaps most importantly, while on average people’s borrowing did go up, our study shows that their average pension contributions increased by a larger amount. Nor should we assume that people went on borrowing more as their pensions saving journeys continued. These studies only looked at the first few years of auto enrolment, at which stage people might not have had time to adjust their spending and savings patterns to a reduced level of take-home pay. It’s reasonable to think that many people’s behaviour has developed over time, and that’s why we are continuing to explore the full range of effects of the policy.
Like a stone cast into a pond, auto enrolment causes ripples that radiate out across people’s household finances. Where these ripples meet other currents, they create interference patterns. Sometimes this could increase existing behaviours, like borrowing, other times it suppresses them. Sometimes it might even be causing something completely new to happen, like getting on the property ladder.
The assumption behind the auto enrolment policy was that people would reduce their spending to cover the cost of being auto enrolled. Thanks to these new studies, we can see that some people do indeed start to do that; but we also see that this doesn’t always fully account for the added cost of making contributions. That is why it’s so important that when we set contribution rates in an auto enrolment system, we take into account the ripples that this can cause in people’s financial lives.
If contribution rates were to increase in the future, it may be that an additional ‘safety valve’, particularly for those on lower earnings, could help protect people against potentially negative outcomes. One option could be to introduce an emergency savings component to any evolution of auto enrolment for some or all workers. This could help them build a financial buffer alongside their retirement savings.[4]
Our own trials of emergency savings are showing how this can work in the real world. We’re excited to see the positive impact that greater take-up of these approaches will make.
Matthew Blakstad, Analysis Director at Nest Insight
If you’re looking to answer a question around the dynamics of household finances, or you have data that could help us do so, please get in touch: insight@nestcorporation.org.uk
[1] https://www.gov.uk/government/news/ten-years-of-automatic-enrolment-achieves-over-114bn-pension-savings. The 2022 data further show a £29bn real-terms increase.
[2] Beshears, J. et al (2024) Does pension automatic enrollment increase debt? Evidence from a large-scale natural experiment (PDF).
[3] https://tahachoukhmane.com/wp-content/uploads/2023/11/CP-Financing-Retirement.pdf
[4] nestinsight.org.uk/research-projects/workplace-emergency-savings/
