The pitch for financial independence is a single sentence: save 25 times your annual spending, withdraw 4% a year, and never work again.
The arithmetic is right. The evidence behind the 4% is narrower than the pitch suggests.
The 4% rule is not a law of finance. It is the worst outcome in a particular set of historical data, measured for a particular length of retirement.
Where the number comes from
In 1994, financial planner William Bengen studied rolling retirements in US history using stock and bond returns, withdrawing a fixed real amount each year, and found that the worst case in his data supported an initial withdrawal of a little over 4%. A 1998 study by three finance professors at Trinity University examined similar ground and made the idea widely known as the Trinity study.
Three things sit inside that result:
- It uses US data, in a century when the US was an unusually successful market.
- It assumes a 30-year retirement. Someone retiring at 40 plans for 50 or more.
- It assumes you never adjust, which almost no one does.
Later research using data from other developed markets has generally found lower safe withdrawal rates than the US figure. We cite that as a general finding and have not reproduced a specific table here; if you rely on a particular number, read the study for the country and period you care about.
The number you need
The financial-independence figure is annual spending divided by the withdrawal rate. The choice of rate is a choice about how much failure you will accept.
| Withdrawal rate | Multiple of annual spending | Portfolio for US$40,000 a year |
|---|---|---|
| 4.0% | 25.0x | US$1,000,000 |
| 3.5% | 28.6x | US$1,142,857 |
| 3.0% | 33.3x | US$1,333,333 |
Moving from 4% to 3.5% means needing 14% more capital. That is the price of a margin of safety.
Sequence risk: the same returns, a different outcome
The reason a withdrawal rate can fail is not low average returns. It is the order of returns.
Take 20 annual returns, from +25% to -20%, averaging 7.55% a year. Start with a portfolio of 1.0, withdraw 4% of the starting value in year one, and raise the withdrawal by 3% a year for inflation. Now run the same 20 returns in two different orders:
| Order of the same returns | What happens |
|---|---|
| Best years first | After 20 years the portfolio is worth about 2.5 times the starting amount |
| Worst years first | The portfolio is exhausted in year 15 |
Same average. Same withdrawals. The only difference is whether the bad years come early, when the portfolio is largest and withdrawals take the biggest share of what is left, or late.
This is a deliberately extreme pair of orderings, chosen to make the mechanism visible. Real sequences sit between them. But the lesson is general: the first decade of retirement carries most of the risk, because losses early in the withdrawal phase are locked in by the withdrawals themselves.
What changes outside the United States
Financial independence is a global idea, but the plumbing is local. A few differences that came out of our research on six markets:
The bridge. Many retirement systems lock money until a statutory age. If you stop work at 45, state or provident-fund balances may be unavailable for ten or more years, and your taxable portfolio has to carry that bridge. Malaysia's 2024 EPF reform created an Account 3 that is withdrawable at any time, but at the cost of the other accounts' long-run growth.
Moving goalposts. China began raising its retirement age in January 2025, and the minimum contribution period starts rising in 2030. A plan that assumed a fixed date for state benefits needs a bridge that can lengthen. See retirement systems are policies, not assets.
Currency. If your assets are in dollars and your spending is in ringgit, rupiah or Taiwan dollars, the withdrawal rate is also a currency bet. A 10% move in the exchange rate changes your effective spending by about 10%. See the currency you did not choose.
Tax on the way out. Withdrawals are taxed differently by account and country. China's new private pension, for example, taxes withdrawals at 3% in exchange for tax-free growth. The sustainable rate is an after-tax rate.
What the fundamentals add
Plan for ranges, not an average. Sound retirement-spending analysis uses simulations and stress tests, not a single assumed return, because the risk is in the distribution.
Flexible spending beats fixed spending. If you cut spending 10% after a bad year, the plan survives scenarios that a fixed rule fails. The guardrail rule is a simple version: set a lower and upper bound for your withdrawal rate, and adjust when you cross either.
Longevity is a range. A longer retirement lowers the safe rate. The younger you retire, the more conservative the number.
Practical implications
- Pick a rate for your own horizon. 4% is a reference for a 30-year retirement in one dataset. For 40 to 50 years, use a lower one as the starting point.
- Build a bridge account. Count assets you can reach before the statutory age separately from locked ones.
- Hold two to three years of spending in lower-risk assets. The aim is to avoid selling equities during a downturn in the first decade.
- Write guardrails. "If the withdrawal rate rises above 5%, cut spending by 10%. If it falls below 3.3%, raise it by 10%."
- Match currency to spending. The money you spend in a currency should sit in that currency, or be hedged on purpose.
- Test the plan with an ugly first decade. If it only works when returns arrive in a friendly order, it is a hope.
Note on the examples
The sequence illustration uses 20 hypothetical annual returns (+18%, +10%, -12%, +22%, +7%, +15%, -20%, +12%, +9%, +25%, -8%, +11%, +16%, -15%, +19%, +5%, +13%, -5%, +21%, +8%), withdrawals of 4% of the starting value in year one, raised 3% a year, taken at the start of each year. The two orderings sort the same returns from best to worst and from worst to best. It is a demonstration of mechanism, not a forecast or a measure of how often a sequence would occur. The Bengen and Trinity references are from our general knowledge of the literature; verify details against the original papers before relying on a figure. This is education, not retirement advice.