Short answer: the "4% rule" — draw 4% of the pot in the first year, then raise the amount with inflation — comes from US market history. Run the identical test on UK returns since 1871 and the rate that survived every 30-year retirement start, for a balanced 60/30/10 mix, is about 2.9%; a 4% start failed roughly one UK retirement in four. The median start could have paid over 5% — which is exactly the problem: the rule's answer depends entirely on the decade you happen to retire into, which is why a year-by-year plan with a flexible spending rule beats any fixed percentage.
Where the 4% rule comes from
In 1994 William Bengen asked: for every US retirement start since 1926, what was the largest inflation-adjusted withdrawal a 50/50 portfolio sustained for 30 years? The worst case — retiring into the late 1960s — allowed about 4%. That single worst case became "the 4% rule". The US, though, had the best-performing stock market of the twentieth century. Britain had two world wars, the 1970s, and inflation that touched 25%.
The same test on UK history
For every start year from 1871, the largest starting withdrawal — rising with inflation each year, taken at the start of the year — that a mix sustained. Computed from the Jordà-Schularick-Taylor UK series, 1871–2020, before fees.
Thirty-year retirements, by asset mix:
| Mix (shares/bonds/cash) | Survived every start | Median start could pay | 4% survived |
|---|---|---|---|
| Very cautious 15/65/20 | 2.2% | 4.0% | 49% of starts |
| Cautious 30/55/15 | 2.4% | 4.5% | 62% |
| Balanced 60/30/10 | 2.9% | 5.2% | 76% |
| Growth 75/20/5 | 3.1% | 5.5% | 84% |
| Adventurous 85/10/5 | 3.2% | 5.8% | 87% |
A balanced mix, by retirement length:
| Years the money must last | Survived every start | Median | 4% survived |
|---|---|---|---|
| 20 | 3.8% | 7.2% | 95% of starts |
| 25 | 3.3% | 6.2% | 88% |
| 30 | 2.9% | 5.2% | 76% |
| 35 | 2.7% | 4.7% | 71% |
| 40 | 2.5% | 4.3% | 61% |
The result nobody expects
Read the first table again: the cautious mixes had the worst worst cases. A 15%-shares portfolio could only ever promise 2.2%; an 85%-shares one promised 3.2%. That is because the killer in UK history was never a crash — markets recover — but inflation: the worst starts are 1910 and 1937, each running into a world war and the price rises that followed, and 1947 and the 1970s for the cautious mixes. Bonds and cash have no defence against a decade of high inflation; shares eventually do. Holding less in shares protects the statement value and loses the purchasing power.
Why no fixed rate is the answer
- The gap between worst and median is enormous — 2.9% against 5.2%. Plan on the worst case and three retirements in four leave a fortune unspent; plan on 4% and one in four runs out.
- These rates are before charges. Fund and platform fees come straight off the withdrawal rate: a 2.9% safe rate with 0.8% of charges is really 2.1% of spending power. (This is the strongest argument for cheap funds there is.)
- Nobody spends like a robot. The rule assumes you keep raising spending with inflation while your pot collapses. Real people cut back — and a rule that says when and by how much turns most historical failures into successes.
- The state pension changes everything. £12,548 a year, inflation-linked, for life, per person — a floor the US studies never had. The pot only has to bridge the gap, and a couple with two state pensions needs a far lower rate from their savings.
What actually raises the safe rate
Money Horizon does not use a fixed rate; it solves your plan year by year and stress-tests it. The levers that raise what you can safely spend, roughly in order of power: retiring onto a full state pension (checking your NI record first), a spending rule that flexes — the planner compares constant, percent-of-wealth, spread-over-years-left and ratchet rules, with optional guardrails that cut spending when the withdrawal rate runs high; cutting charges; a cash bucket or gilt ladder covering the first years so a bad start never forces selling; and, for the cautious, accepting more shares than feels natural — the table above is the evidence.
The planner's sustainable income answer is this whole calculation run on your own figures: the most your savings can pay every year to the age you choose, tested against the worst start since 1871, replays of 1973, 2000 and 2008, and hundreds of resampled futures.
Questions people ask
Is the 4% rule safe in the UK? Not as a guarantee. Against UK history since 1871, a 4% inflation-adjusted withdrawal from a balanced portfolio failed about 24% of 30-year retirement starts. It worked in most decades — but "works three times in four" is not what people think a rule means.
What withdrawal rate is truly safe in the UK? The rate that survived every 30-year start since 1871 is about 2.9% for a balanced mix, before charges. But the historically safe rate is not a plan — it is the price of ignoring the state pension, flexible spending and your actual horizon. A year-by-year plan usually supports more.
Should I just use 3% instead? A 3% rule is safer but keeps the same flaw: it ignores your age, the state pension, tax, and what markets do after you start. The median UK start could have paid over 5% — a fixed 3% would have left most retirements hugely underspent.
Does the safe rate include fees? No — the table is before charges. Subtract your total fund and platform charges from any rate you read, here or anywhere.
Where do these numbers come from? Computed by Money Horizon's engine from the Jordà-Schularick-Taylor Macrohistory Database, UK series 1871–2020 (CC BY-NC-SA 4.0): for every start year, the largest starting withdrawal, rising with inflation, that the mix sustained for the stated number of years. The same engine runs the planner.