41 per cent of people aged 40 to 75 told a government survey they had no idea how much income they would need in retirement. But working out how much is enough to retire isn't really a question about the life you want. It's a question about what your money would have to survive.
In late 2024 the Department for Work and Pensions asked 4,036 people aged 40 to 75 across Great Britain how much income they would need in retirement. Published the following year as Planning and Preparing for Later Life, the survey found that 41 per cent had no idea. Not a rough figure or a range. Nothing at all. For a large share of people approaching the end of their working lives, the question of how much is enough to retire has never been answered, in plenty of cases never even asked.
The comfortable reading is that this describes people without money. The same survey closes that off. It sorts the 40 to 75 population into five groups by how prepared they are, and the biggest of them, at 27 per cent, is labelled 'Equipped but unengaged': homeowners on middling to good incomes, reasonably confident about the standard of living they will get and mostly unable to say when they last looked at what their pension is worth. Nearly three in ten people in this age bracket are financially comfortable and have never checked.
The rest of the survey says much the same. Fewer than half of private pension holders aged 40 to 75, 46 per cent, had reviewed the value of their pension in the previous 12 months. Among those with a defined contribution pot they hadn't yet touched, 77 per cent had no clear plan for getting at it, and 21 per cent didn't know a choice was required of them.
Before anybody can say whether they have enough, somebody has to work out what enough would look like, and then test it against everything that might happen to it.
A number pointed at the wrong target
The people who do have a number are mostly carrying a spending target: the income that would pay for the life they picture. The evidence on what that buys in day-to-day feeling is weaker than the target implies.
Among DWP respondents who had some idea of how much is enough to retire on, the median estimate was £24,000 a year. Set against the Retirement Living Standards, which put a Moderate standard at £32,700 for a single person and £45,400 for a couple and assume no rent or mortgage, that looks light.
The DWP's own reading is more forgiving: it notes that £24,000 is broadly in line with what a median earner would need to reach their Target Replacement Rate. The trouble is how it got there. It's a number people produce when asked, not one they have worked out.
And the exercise behind it, costing up the life you want, rests on an assumption the research doesn't support.
In 2023 Matthew Killingsworth and Daniel Kahneman, who had reached opposite conclusions on whether happiness keeps rising with income, published a joint paper in the Proceedings of the National Academy of Sciences, with Barbara Mellers as arbiter. The famous plateau at around $75,000 was real, but only for the least happy 15 to 20 per cent. For everyone else happiness carried on rising, and for the happiest 30 per cent it accelerated.
The correlation between average happiness and log income is 0.09. And the authors' own comparison: a roughly fourfold difference in income carries about the emotional weight of being a caregiver, twice that of being married, about that of a weekend and less than a third that of a headache.
The caveats matter. This is American data, drawn from 33,391 employed adults aged 18 to 65 and collected between 2009 and 2015. Retired people were excluded by design. It is cross-sectional, so it describes associations rather than causes.
None of which says money doesn't matter. It plainly matters, and the paper found it still matters at high incomes. What it shows is narrower: more income buys far less daily experience than a lifestyle figure implies, which makes a spending target necessary but not sufficient. The more useful question is what the money has to withstand: a retirement date you don't get to choose, a lifespan nobody can specify in advance and a household that might shrink to one.
The date you don't choose
Every retirement number assumes a date. A substantial minority never get the one they planned for.
The DWP survey found a median expected retirement age of 66 against a median ideal of 60, and 72 per cent of people not yet retired expecting to stop later than they would like. That gap gets plenty of attention. It isn't the useful one.
Look instead at the two ends of the same survey. Among people who hadn't yet retired, five per cent expected to retire earlier than their ideal age. Among people in the same survey who had already retired fully, 24 per cent said they had. Those are two different groups at two different stages, not the same people tracked over time. The expectation and the experience are still a long way apart.
The DWP's labour market statistics for people aged 50 and over, published in September 2025, show where some of that comes from. Of roughly 2.2 million people aged 50 to 64 who weren't in work and had left their last job within the previous eight years, 32.6 per cent gave retirement as the reason. Health reasons accounted for 28.3 per cent.
Not every early exit is forced, and a proportion of that group stopped because they could afford to. But nearly three in ten stopped because of their health, and nobody schedules that.
If your response is that you're healthy, it's the normal one, and it's not much of a guide. 28 per cent of people aged 40 to 75 already report a long-term limiting health condition, and health can change faster than most retirement plans get revisited.
How long it has to last
Whatever figure you land on for how much is enough to retire on, it has to cover a stretch of time nobody can specify in advance, and about a decade of it won't be spent in good health.
On the ONS's 2024-based cohort life tables, published in May 2026, a man aged 65 in 2024 could expect a further 20.0 years and a woman 22.7. These allow for the mortality improvements expected across the rest of a life, which suits an individual better than the period tables usually quoted. The ONS is blunt about those: they 'do not give a true estimate of how long someone can expect to live', it says.
Averages hide a great deal. In the ONS's analysis by area deprivation, male life expectancy at 65 runs from 15.3 years in the most deprived tenth of English neighbourhoods to 21.2 in the least deprived. Six years of range around a national figure of 18.7. That is area deprivation rather than individual wealth, so it says nothing about any particular reader. It shows how much variation a single national number is holding.
Then the figure that does most of the work here. Healthy life expectancy at 65 is 10.0 years for men and 11.0 for women, against total period life expectancy of 18.7 and 21.2. Somewhere around nine or ten years of a typical retirement are lived in something other than good health. The ONS measure is self-reported: it counts years spent in 'very good' or 'good' general health, and it isn't a measure of needing care.
A flat annual spending figure describes none of that. Retirement has a healthy first phase and a longer, different second one. A single sum spread evenly across both is what most planning assumes.
When two become one
For most couples, one of them ends up living alone. How comfortable that is was largely settled years earlier, by decisions almost nobody goes back to.
The first assumption to challenge is that spending halves when income does. It doesn't. The Retirement Living Standards put a Moderate standard at £45,400 for a couple and £32,700 for one person, meaning a survivor needs around 72 per cent of what the couple was spending.
Income, on average, falls by less than that. Analysis of the English Longitudinal Study of Ageing carried out for the Centre for Ageing Better found that among people bereaved after 65, mean weekly income dropped from around £400 to around £320. A fall of about a fifth, against a spending requirement that drops by nearer 28 per cent.
But averages conceal the important point. Retirement income is built differently from one household to the next, and the difference matters when one person dies.
The State Pension is not passed from one spouse to another. A couple each drawing the full amount, £241.30 a week in the 2026/27 tax year, becomes a household drawing one. On the FCA's retirement income market data for 2024/25, more than two-thirds of annuities bought that year were single-life, which means the income stops when the policyholder dies.
Whether a survivor is comfortable or stretched often depends on decisions made years earlier, usually around the point of retirement. Unlike many retirement risks, this is one that can be examined before it happens.
Working out how much is enough to retire
Testing the number means running it against the things that would break it, rather than against the life you would like it to pay for.
Three questions get most of the way there: what happens if you stop three years earlier than planned, if one of you dies at 72 and if the first five years of drawing an income land in a falling market.
The trouble is when people get round to asking them. Most start saving in their 20s or 30s and start actively planning in their 50s: 45 per cent of the semi-retired and 40 per cent of the fully retired in the DWP survey began then. That leaves a narrow stretch between planning starting and the big decisions becoming irreversible.
Only 16 per cent of people aged 40 to 75 had used regulated advice or guidance in the previous 12 months.
The answer to how much is enough to retire usually isn't a bigger number. More often it's the same number with the shocks priced in, which is worth a good deal more than a larger one nobody has tested.
The part money was never going to fix
Killingsworth, Kahneman and Mellers make a point in passing worth holding on to. Above a certain income, the miseries that remain aren't the kind money reaches. They name heartbreak, bereavement and clinical depression.
There is a counterweight in the World Happiness Report's 2025 analysis of kindness: believing a lost wallet would very likely be returned is associated with life satisfaction more than three-quarters of a point higher on a 0 to 10 scale, a larger association than that of a doubling of income. People are also too pessimistic about this: across the more than 17,000 wallets in Cohn and colleagues' 2019 wallet-drop experiment in Science, spanning 40 countries, wallets came back rather more often than expected.
So the things that appear to shift life satisfaction most don't appear on anybody's balance sheet. Which is a reasonable argument for settling that question properly, once, then leaving it alone. Knowing how much is enough to retire is worth having mainly because it stops the asking.
If you'd like to talk to Matt Millard, founder of rockwealth Cardiff, about testing the numbers behind your own retirement plan, why not arrange an initial meeting?
