How Much Do You Need to Retire?

Start from the income you want, not from a pot size. The pot is just the arithmetic that follows.

Work backwards from income

The useful sequence:

1. Estimate your annual spending in retirement. Start from current spending, then adjust: the mortgage may be gone, commuting and work costs disappear, but healthcare, heating and leisure often rise. A common finding is that spending falls in early retirement, rises again for care costs much later, and does not fall as much overall as people expect.
2. Subtract other guaranteed income. State pension, defined benefit pensions, rental income.
3. The remainder is what your pot must provide.
4. Divide by a withdrawal rate to get the target pot.

Someone wanting £30,000 a year with a £11,500 state pension needs the pot to provide £18,500 — about £463,000 at 4 %, or £617,000 at 3 %. The gap between those two figures shows how much the withdrawal rate assumption matters.

The 4% rule and its caveats

The figure comes from Bengen's 1994 work and the subsequent Trinity Study, which tested historical withdrawal rates against US market data. A portfolio withdrawing 4 % in year one, increased with inflation thereafter, survived 30 years in almost all historical periods.

The caveats are substantial:

• It was US data over a specific period that included exceptional equity returns. Other markets performed worse.
• It modelled 30 years. Someone retiring at 55 may need 40 or more.
• It assumed a specific allocation, roughly 50–75 % equities.
• It ignored fees. A 1 % charge materially reduces the sustainable rate.
• It assumed constant real spending, which no real retiree follows.

Many practitioners now use 3–3.5 % for longer horizons. Others argue flexibility — reducing spending in poor years — improves survival more than any fixed rate, and the evidence supports that.

Sequence risk: the thing that actually matters

Two retirees with identical average returns can have completely different outcomes depending on when the bad years arrive.

A 30 % fall in year one of retirement, while you are withdrawing, forces you to sell more units at depressed prices to fund the same spending. Those units are never recovered. The same fall in year twenty, after two decades of growth, is far less damaging.

This is why averages are misleading and why the years immediately before and after retirement are the most dangerous. Common mitigations:

• Hold one to three years of spending in cash, so you never have to sell into a falling market.
• Reduce equity exposure in the years approaching retirement, then potentially increase it again later.
• Be flexible about spending in poor years, which is the single most effective adjustment.
• Consider partial annuitisation to cover essential expenses with guaranteed income.

Why retiring later helps twice

Working an extra two or three years has a disproportionate effect, because it does three things at once: adds contributions, removes years of withdrawals, and gives the pot more time to compound.

The rough effect of three additional working years is frequently equivalent to a 20 % larger pot at the original date. Very few other adjustments are that powerful.

Part-time work in early retirement does something similar with less of the cost: even modest earnings substantially reduce the withdrawal rate needed in the early years, which is precisely when sequence risk is highest.

Our retirement calculator lets you test these scenarios, and the savings rate calculator shows how contribution rate affects the timeline.