How to Calculate Your FI Number: The 25x Rule and Its Limits

Your FI number is the amount of invested savings that could pay, year after year, the part of your living costs that no pension or other reliable income covers. The usual shortcut multiplies that yearly gap by 25, which is the 4 percent withdrawal rule turned upside down. The best-known US studies behind that rule planned for retirements of about three decades and took no fees or taxes out of the portfolio, so that multiple is a starting estimate that a longer or costlier retirement can outgrow.¹²

This guide works through the arithmetic, one worked example and the assumptions hiding in the multiple. For the bigger picture, including the partial independence that comes long before any full target, start with what financial independence means beyond never working again.

Two numbers make an FI number: the gap and the multiple

An FI number has two parts: the yearly spending your savings must cover, and a multiple that turns that yearly amount into a pot big enough to keep paying it. The popular multiple simply inverts that withdrawal rule: 4 percent of a pot worth 25 years of your gap is exactly one year’s gap.

In practice, with an invented household: take what it spends in a year, subtract the part a pension will pay, and multiply what is left by the multiple. A gap of $30,000 a year (in 2026 US dollars) times 25 gives $750,000. Each later section tests one of those inputs.

The multiple is the part people copy without checking. The 4 percent rate traces to a 1994 paper in which the planner William Bengen replayed US market history: he tested a first-year withdrawal that then rose with inflation each year, and asked whether the money lasted at least three decades.¹ The rule of thumb inherits every assumption of that test, including its length.

So the multiple quietly answers a question nobody asked you: how long must the money last? A rate that suits a new retiree whose pension is already paying may not suit someone who stops work in midlife and waits two decades for a pension, yet both are handed the same multiple.

Keep the gap and the multiple apart on paper, and question each on its own.

How to calculate your FI number in four steps

Calculating your FI number takes four steps: work out your yearly spending, subtract the income you can count on, make sure taxes and fees are inside the spending figure, and choose a multiple that fits how long the money must last. The first three set the gap; the last is where the research matters most.

Start from what you spend, not what you earn

The target covers spending, so it starts from spending. Add up a full year of what actually left your accounts rather than guessing, because every error in that figure gets multiplied along with it. If you have never totted it up, a step-by-step guide to making a budget walks through it.

Then adjust it for life without a job. Money you now save or put into a workplace pension stops; costs your employer pays for you today may become yours. A household spending the same each month but saving a quarter of its pay has a much smaller gap than its income suggests.

Subtract the income you can count on

Only the gap needs a multiple. Morningstar’s 2025 safe-withdrawal estimates for US retirees, for example, leave out Social Security and other income from outside the portfolio: they describe what savings alone must supply.³

Timing is the catch for anyone stopping work early. A US Social Security benefit claimed at full retirement age does nothing for the years before it, so an early leaver’s gap is larger at first and shrinks later. Ignoring the later pension entirely may be overly conservative, a 2026 Morningstar analysis for early retirees notes, and it suggests planning software or an adviser that can vary income year by year.⁴

To find your own figures: in the US, the CFPB’s retirement page (updated January 2026) points out that your monthly Social Security benefit depends on how old you are when you claim it.⁵ In the UK, GOV.UK’s forecast service shows how much State Pension you could get and when.⁶ In other countries, ask the national pension authority for a forecast.

Put taxes and fees inside the spending figure

The classic withdrawal studies left both out. The 1998 Trinity study did not adjust for taxes or transaction costs, and Wade Pfau’s 2010 paper points out that such rates are pre-tax for taxable accounts.²⁷ If withdrawals will be taxed, the spending figure you multiply has to include the tax.

Fees work differently: they lower the safe rate itself. In Pfau’s 2010 illustration, adding the average fund fees of the time cut the highest safe rate in Bengen’s later US calculation by about two-thirds of a percentage point, which moves the implied multiple from about 24 to about 29.⁷ That is why the yearly charge a fund takes belongs in the calculation, not in a footnote.

Choose a multiple for your horizon

The last step is the judgment call: how many years must the money last, and how much safety margin do you want? The table shows how the same gap turns into different targets.

The $30,000 gap is invented, in US dollars as of 2026, and the results are arithmetic, not forecasts or targets for any reader. Future returns may fall short of anything these studies assumed, and after inflation they can stay negative for years at a time. Many households, especially on low incomes or caring for others, will not reach a full target; covering part of the gap still adds choices.

To use the table, run your own gap through two multiples, the usual one and one that fits a longer horizon, and treat the space between the two results as your planning range rather than picking a single number.

The 25x rule rests on three decades of US market history

The usual multiple rests on studies that replayed US stock and bond returns from 1926 onward, chiefly Bengen’s 1994 paper and the 1998 Trinity study by three finance professors at Trinity University in Texas. Both asked whether a withdrawal plan would have survived each past stretch of market history, up to about three decades long.

The Trinity study is easier to read: it gives a success rate for every mix of withdrawal rate, portfolio and length it tried.

The study

Limited evidence

Did 4 percent last 30 years? The Trinity study counted every period (1998)

Philip Cooley, Carl Hubbard and Daniel Walz took a set share of the starting balance in year one, raised it with inflation each year, and counted how often the money lasted. At 4 percent over 30 years, a portfolio split evenly between stocks and bonds lasted in 95 percent of the periods, and a portfolio of 75 percent stocks in 98 percent.²

Those success rates describe the past, not odds for the future. The periods overlap heavily: US data since 1926, Pfau noted in 2010, held fewer than three separate stretches of that length, so the record is thinner than the count suggests.⁷ And a test of three decades says nothing about the years after them.

Readers of our financial independence guide may remember a rosier line: in Bengen’s test of a half-stock portfolio, a 4 percent start outlasted three decades from every starting year.¹ The two results do not clash. Bengen held intermediate-term Treasury notes and reported how many years the money lasted from each start year; the Trinity team held long-term, high-grade corporate bonds and reported the share of three-decade periods that survived.¹² Both say the same thing: in past US markets, before costs, that starting rate lasted three decades in all or nearly all periods for portfolios of half to three-quarters stocks.

Put the three assumptions to your own plan.

If the honest answer to any part is yes, those studies did not test a plan like yours, and the usual multiple is a starting point, not a finish line.

Why a longer retirement needs a bigger multiple

The longer your savings must last, the lower the rate you can safely start with, so the bigger your FI number. Morningstar’s 2026 estimates for US retirees put the safe starting rate for a 40-year retirement at 3.3 percent, which works out to about 30 times spending.⁴

The reason is exposure. More years give the portfolio more time to grow, but also more chances to meet a run of poor returns.⁴ Timing matters too: in Morningstar’s 2025 research, retirees who met poor returns in their first five years and did not cut spending were much more likely to run out of money.³ The Trinity authors drew the horizon lesson from their own tables.

Early retirees who anticipate long payout periods should plan on lower withdrawal rates.

Bengen told his own clients the same: his rate was a ceiling for the early years, above all for those retiring early.¹ Take two invented savers with the same gap: one stops work just before her pension starts, the other in midlife. The second is not asking for the same retirement earlier; she is asking for one that may have to last about two decades longer.

1. About 30 years: a starting rate of 3.9%, or about 26 times the yearly spending gap
2. About 40 years: a starting rate of 3.3%, or about 30 times
3. About 50 years: a starting rate of 2.9%, or about 34 times

Morningstar’s figures are company research, not peer-reviewed, built on its own forecasts of returns and inflation and run through a thousand simulated market paths; Morningstar also sells investment products.⁴ Treat them as one informed estimate.

Where you live matters too. Pfau found in 2010 that the 4 percent rule, even with some overly optimistic assumptions, would have been safe in fewer than a quarter of the 17 developed countries he studied.⁷ A 2025 study of 38 developed countries also put the safe rate for a 65-year-old couple materially below conventional advice, for a rule that takes a fixed share of each year’s balance; we could see only its abstract.⁸

What to take from this: if you expect to stop work in midlife, or your savings sit in markets with a weaker history than the US, the research cited here points to a starting multiple nearer 30 than 25, as a planning estimate rather than a target.

What can justify a smaller multiple

Flexibility is the main argument for a multiple at or below the usual one: willingness to cut spending after bad years, income that arrives later, or some paid work. Morningstar’s 2026 analysis estimates that early retirees who accept lower odds or use a flexible spending method can likely spend as much as one or two percentage points more than its fixed-spending figures.⁴

The mechanism is simple. A fixed rule keeps withdrawing the same real amount after a crash; a flexible one takes less when the portfolio is down and more when it recovers.³ Even the Trinity authors called the two lowest rates they tested exceedingly conservative for stock-heavy portfolios and expected retirees to make mid-course corrections.²

Flexibility has a cost of its own. Cuts are easier to absorb in holidays and dining out than in rent, insurance or food, so a budget made mostly of fixed costs has less room to bend. Morningstar makes the same point: the right amount of flexibility depends on how much of your fixed spending other income covers.³

One way to apply this, our suggestion, not a tested method: split the gap into fixed and flexible costs, and apply the more cautious multiple to the fixed part.

When your numbers need a second opinion

Most of the arithmetic here you can do yourself; a few decisions deserve professional advice first. Leaving a job, changing how savings are invested and taking money from a pension early are hard to undo, and the right multiple depends on circumstances no rule of thumb can see.

We suggest talking to a regulated financial adviser before taking any of them, and asking up front how the adviser is paid. US readers can look an adviser up in BrokerCheck, a free FINRA tool that lists professional background along with customer disputes and disciplinary events.⁹ UK readers can search the Financial Services Register, the FCA’s official public record of the firms and individuals it covers.¹⁰ In other countries, ask the national financial regulator how to check an adviser.

The bottom line

The gap is the part of an FI number you can measure; the multiple is the part you have to judge, and it is the weak link. Twenty-five times comes from three-decade tests of past US markets with no fees or taxes; a longer horizon, higher costs or weaker markets argue for about 30 times or more in the research cited here, while genuine room to spend less in bad years argues for less. Work out your own gap carefully, then choose the multiple as a judgment about time and flexibility, not as a law.