Key takeaways
- Fidelity's ladder of 1x your salary by 30 and 10x by 67 is one division sum: a 45% income target divided by the 4.5% withdrawal rate those two numbers imply.
- T. Rowe Price's midpoint benchmark is 11x by 65 rather than 10x by 67, because it assumes a 4% withdrawal rate. At 11x and 4%, savings fund 44% of income.
- Rebuilt on UK inputs the multiple is 4.7x, because the £241.30 a week State Pension already replaces 30.7% of April 2026 median full-time pay of £40,911.
- Change only the conversion rate and the same UK target needs 4.7x, 8.1x or 9.1x. Nothing about the saver differs between those three answers.
- 43% of working-age people, some 15 million, are projected to miss the first Pensions Commission's target replacement rates, and 52% in one Triple Lock scenario.
Where the multiples of salary by age come from
Type the question into Google and you get a ladder: 1x your salary by 30, rising to 10x by 67. It comes from Fidelity, and the number at the top of it is a division sum rather than a discovery.
Fidelity's guideline, as published in November 2026, reads: "Aim to save at least 1x your salary by 30, 3x by 40, 6x by 50, 8x by 60, and 10x by 67." The footnote underneath is where the work happens. The same page tells you to "Plan for your savings to provide about 45% of your pretax, preretirement income." The methodology note calls that the "replacement annual income target", defines it as "45% of pre-retirement annual income" and says it "assumes no pension income".
Divide 45% by 10 and you get 4.5%. That is the withdrawal rate hiding inside the benchmark. The 10x is not an observation about how much people need. It's the pot size that produces 45% of a salary at 4.5% a year, and if you move either input the 10x moves with it.
The rest of the footnote is equally specific. Fidelity assumed "age-based asset allocations consistent with the equity glide path of a typical target date retirement fund, a 15% savings rate, a 1.5% constant real wage growth, a retirement age of 67 and a planning age through 93". Its companion page is explicit about where the ladder starts: the 15% goal rests on research "which indicates that most people would need to contribute this amount from an assumed starting age of 25 through an assumed retirement age of 67". The multipliers themselves came from "multiple market simulations based on historical market data, assuming poor market conditions to support a 90% confidence level of success". Nothing there is hidden. It's just never printed next to the 10x.
A second firm publishes a different ladder, and the gap is the withdrawal rate
T. Rowe Price publishes the same kind of chart. Its savings benchmark sheet, whose ranges are built on "federal tax rates as of January 1, 2023", gives midpoints of half a year's income at 30, then "1x 2x 3x 5x 7x 9x 11x" at 35, 40, 45, 50, 55, 60 and 65. So the ladder ends at 11x two years earlier than Fidelity's ends at 10x.
The assumptions explain the gap. T. Rowe Price's note says the saver "retires at age 65 and begins withdrawing 4% of assets (a rate intended to support steady inflation-adjusted spending over a 30-year retirement)". Household income grows at 5% until 45 and 3% after that, investment returns run 7% before tax, and Social Security is included using the SSA's own calculator.
Now do the arithmetic both ways. Fidelity: 10x at 4.5% funds 45% of salary. T. Rowe Price: 11x at 4% funds 44%. Two firms, two ladders, and they agree almost exactly on the income the pot has to produce. They disagree only about the price of buying that income. The published multiples of income by age differ by a tenth because the withdrawal assumption differs by half a percentage point.
If you've ever wondered why the 25x rule for retirement and the 10x rule look like they belong to different planets, it's because one is a multiple of spending and the other a multiple of pay. The conversion between them is the same withdrawal rate, and the safe withdrawal rate literature is where that number is argued over.
One multiple cannot survive contact with household shape
T. Rowe Price prints a second table that quietly demolishes the idea of a single ladder. At a household income of $75,000, a single person's benchmark at 65 is 10½x, while a married dual-income household's is 8x. At $300,000 the single person's figure is 13½x.
The firm gives the reason: "For higher-earning households, Social Security benefits will be a smaller percentage of income in retirement, so they will rely more on personal savings." A roughly flat state benefit covers a shrinking share of a rising salary. The multiple you need isn't a property of your age at all. It's a property of how much of your income the state has already replaced.
Which is why importing a US ladder into a UK plan goes wrong before you reach the first rung.
Britain never published a multiple. It published a replacement rate.
The UK's benchmark came out of the first Pensions Commission in 2004, and it never took the form of multiples of salary by age. It was a target replacement rate that fell as income rose. The Department for Work and Pensions restated those benchmarks in 2013 earnings terms: 80% for pre-retirement earnings below £12,200, 70% from £12,200 to £22,400, 67% from £22,400 to £32,000, 60% from £32,000 to £51,300, and 50% above £51,300. Those thresholds are not 2026 figures: DWP uses the 2004 benchmarks "with the income thresholds which they apply to adjusted for earnings growth".
DWP explains why none of them is 100%: "The benchmarks are less than 100 per cent as adequacy doesn't mean having the same income in retirement as in work. Many costs are likely to be lower in retirement, for example, mortgage payments, travel to work costs and the costs of raising children."
The Commission also set out who was meant to pay for it. Writing in 2026, the Second Pensions Commission describes the original design as three pillars: a State Pension restored over time "to 31% of pre-retirement earnings for a median earner", automatic enrolment "contributing a further 15-18% of earnings", and voluntary saving supplying "the remaining 15-18%". Automatic enrolment on its own was never supposed to be enough. DWP put it bluntly in 2013: the 8% default "would only be sufficient, when combined with state pensions, to achieve a replacement rate of about 45 per cent for the average earner".
Rebuilding the multiple on UK numbers
You can convert a replacement rate into a multiple of salary if you fix three things: the salary, the state benefit, and the rate at which a pot turns into income.
Median gross annual earnings for full-time employees who had been in their job at least a year were £40,911 in April 2026, on the ONS Annual Survey of Hours and Earnings. The full new State Pension for 2026/27 is £241.30 a week, which is £12,548 a year, or 30.7% of that salary. That is almost exactly the 31% the first Commission designed the state pillar to deliver. For the conversion, DWP used 4 December 2025 annuity rates and found a £100,000 pot could buy an annuity income of £7,718 a year at 65.
Run four target incomes through it. Three are the Pensions UK Retirement Living Standards for one person, updated on 3 June 2026 and calculated by the Centre for Research in Social Policy at Loughborough University. They are spending baskets built from what the public considers necessary rather than income targets. The fourth is the 67% replacement rate the Commission set for an earner in the median band.
| Target income a year | Needed from private savings | Pot at 65 | Multiple of a £40,911 salary |
|---|---|---|---|
| Minimum standard, £13,900 | £1,352 | £17,500 | 0.4x |
| 67% replacement rate, £27,410 | £14,862 | £192,600 | 4.7x |
| Moderate standard, £32,700 | £20,152 | £261,100 | 6.4x |
| Comfortable standard, £45,400 | £32,852 | £425,700 | 10.4x |
Two things fall out of that table. The first is that the US 10x, dropped onto a UK median salary, buys something close to the Comfortable standard. It isn't a baseline. It's the top of the published range.
The second is less comfortable. The Commission's own 67% target comes to £27,410 a year, which is £5,290 short of the £32,700 Moderate standard. The replacement rate Britain designed its pension system around does not reach the middle rung of the lifestyle ladder the pensions industry now publishes.
The conversion rate does most of the work
Hold the target fixed at £14,862 a year from private savings and change nothing but the rate at which the pot converts. At the 7.718% annuity rate, the pot is 4.7x salary. At Fidelity's implied 4.5%, it is 8.1x. At T. Rowe Price's 4%, it is 9.1x.
Same saver, same salary, same target. The answer nearly doubles.
And the rate is not a constant. The Second Pensions Commission records that a £100,000 pot bought a single-life level annuity income of around £7,900 in 2008 for a 65 year old, and closer to £6,200 by the end of 2013. At the 2008 rate, the same UK target needed 4.6x. At the end-2013 rate it needed 5.9x. A quarter more savings for the identical retirement, because gilt yields had moved.
A benchmark expressed as multiples of salary by age silently fixes a price that doesn't hold still. That is the single strongest reason to treat the ladder as a rough signal rather than a target.
What the ladder assumes about the years in between
The rungs below the top are what turn one retirement target into multiples of salary by age, and Fidelity's footnote lets you take them apart. Contribute 15% of pay every year from 25 to 66, with 1.5% real wage growth, and the contributions total about 4.6 times your final salary. Reaching 10x therefore needs investment returns to supply the other 5.4x, which is a little over half the pot.
DWP's modelling puts the return share higher, at "around two-thirds of the value of a DC pot", on an assumption of "just under 4% per annum" real returns. A lower contribution rate leaves a bigger share of the pot to returns, and the two estimates differ in exactly that direction. Between them they explain why the ladder is so back-loaded. Getting from 1x to 3x takes a decade of your thirties. Getting from 8x to 10x takes seven years, on a base eight times larger. Our piece on when compound interest takes off works through the crossover in detail.
Where people actually are, measured against the benchmark
The Second Pensions Commission's interim report, published on 19 May 2026, projects that "around 4 in 10 (43%) of the working-age population (15 million people)" are undersaving against the first Commission's target replacement rates. Generation X fares worst, at 46%. The range matters as much as the central number: under the OBR's lower-volatility Triple Lock scenario the share rises to 52%, or 19 million, and under high volatility it falls to 37%, or 13 million.
Against the Minimum standard rather than a replacement rate, 13% of working-age people, nearly 5 million, are not projected to get there. Pensions UK puts the same picture the other way up: it expects "around 82% of the working population to reach the Minimum standard of living in retirement", while "this falls to just 23% reaching a Moderate standard and 9% reaching Comfortable". On a median salary those two standards are the 6.4x and 10.4x rows above.
The stock figures are starker than the flow figures. Of Generation X members who have any defined contribution pension wealth at all, 32% have pots totalling less than £50,000, which the report notes "is not significantly higher than Millennials (38%), who still have a decade or more longer to save for their retirement". On a £40,911 salary, £50,000 is about 1.2x. T. Rowe Price's ladder asks for 3x at 45 and 7x at 55.
DWP's own case-study pots tell the same story from the other end. Its model gives an average earner a pot of around £165,000 over a full career, and a £70,000 earner contributing around 10% of full pay a pot of around £455,000. Measured against the April 2026 median salary, the first is 4.0x, just under the 4.7x that salary needs. The second is 6.5x of £70,000, which is almost exactly the 6.4x a £70,000 salary needs at the same 67% target. Both comparisons are indicative rather than exact. DWP's "average earner" is not defined as the ONS median, and the Commission's targets fall as income rises, so the bar for a £70,000 salary is probably below 67%.
Internationally, the UK system "delivers a net replacement rate of 54.2% for a median earner from mandatory pensions (State Pension and AE minima), which is below the OECD average of 63.2%". The report's own diagnosis is that the "third pillar" of voluntary savings "has not materialised for many".
The strongest objection to throwing the multiple away
There's a serious case against everything above, and the Second Pensions Commission makes it. Its interim report says that "relying on these target replacement rates is too ambitious and identifies the wrong groups for public policy to focus on". The numbers show why. For lower earners the UK system already delivers a net replacement rate of 76.2%, slightly above the OECD's 75.4%, because a flat State Pension replaces most of a small wage. A test they pass is not a test that tells you anything. The Commission says the same of the other end: higher earners "are more likely to miss their target replacement rates", yet "the current pensions system is working well for many higher earners in practice and their absolute incomes in retirement remain high".
A multiple of salary inherits that flaw. The Minimum standard needs 0.4x on a £40,911 salary and 1.2x on a £15,000 one, because the target is a cash amount and the salary isn't. A single ladder cannot be right for both.
So why do the ladders persist? Because their crudeness is the feature. The table earlier in this piece needed an annuity quote, a tax year's State Pension rate, an earnings survey and a lifestyle basket, and every one of those moves. "1x by 30" needs none of them, and a rule nobody can compute is a rule nobody uses. Critics of the arithmetic here would say that a benchmark wrong by a factor of two, available instantly, beats a precise one nobody reaches for. On the evidence of how few people are on track, that argument is hard to dismiss.
What these numbers cannot tell you
The annuity figure is one day's quotes, on 4 December 2025, and the report doesn't state the shape of the policy. That matters: of annuities purchased in 2024/25, "over two-thirds of these are 'single-life' rather than 'joint-life' meaning that an annuity income will end when the policy holder dies, leaving a widow/er without that income stream". Nor is an annuity what most people do: of pots first accessed in 2024/25, only 8% bought one, while 48% were taken as full cash withdrawals. A reader in drawdown faces a rate nobody can quote in advance.
The £40,911 is a median for full-time employees who had been in post at least a year, so it excludes part-time workers and job movers, and a median is not your salary.
The Retirement Living Standards are spending baskets, not incomes. "Housing costs are not included within the Retirement Living Standards, as they vary significantly depending on location and personal circumstances", so a renter needs more. DWP's own conversion for 2023-24 put the pre-tax income needed at £35,982 for the Moderate standard and £50,887 for Comfortable, above the spending figures, which means the multiples in the table understate what a taxpayer needs.
The two US ladders are built on US tax rates and Social Security, and T. Rowe Price's ranges use federal tax rates as they stood on 1 January 2023. Their multiples don't transfer. The undersaving projections are model output across a sample of the working-age population, not a count, and the report's own 37% to 52% range is wider than its headline. The pension wealth distributions come from the 2020 to 2022 Wealth and Assets Survey, before the most recent contribution data.
What would change the conclusion
If annuity rates slid back toward their 2013 level, the required multiple would rise from 4.7x to 5.9x with nobody saving a penny less. That is a quarter more capital for the same retirement, and the 2008 to 2013 fall shows the rate can move that far.
If the State Pension's relative value slips, the flat £12,548 offset covers less of a rising salary, and every multiple in the table climbs. That is precisely the OBR scenario in which projected undersaving reaches 52%.
If retirement comes earlier, both ends of the sum move against you. A £100,000 pot bought £6,476 a year at 55 against £7,718 at 65 on the same December 2025 rates, and DWP's model has a median earner born in 2002 needing to contribute 22% of qualifying earnings to retire at 57 rather than 11% to retire at 65.
The useful thing to watch isn't the multiple. It's the two numbers underneath it: the income you're aiming at, and the rate at which a pot converts into it. A tracker like LedgerTouch can show you the pot; the other two are assumptions you pick, and the official UK version of them gets rewritten when the Second Pensions Commission reports again. Its "final report is planned for Spring 2027".