What the 1994 paper actually did

William Bengen, a financial planner, published Determining Withdrawal Rates Using Historical Data in the Journal of Financial Planning in October 1994. The question he set was narrow and answerable: taking every rolling thirty-year retirement window in United States market data from 1926 onward, what is the largest constant inflation-adjusted withdrawal rate that would have survived even the worst starting year?

The answer was 4.15%. He rounded it down to 4% and called it the SAFEMAX. The worst starting year was 1966, whose retirees walked into the inflation of the 1970s early in their retirement, when a portfolio is least able to absorb it.

A second study is usually bundled with it. In February 1998, three professors at Trinity University — Philip Cooley, Carl Hubbard and Daniel Walz — published Retirement Savings: Choosing a Withdrawal Rate That Is Sustainable in the AAII Journal. They tested allocations from all-bonds to all-stocks and withdrawal rates from 3% to 12% against US data from 1925 to 1995, and reported success probabilities rather than a single survivable maximum. The Trinity study is where the familiar “95% success rate” language comes from; Bengen is where the number comes from.

It is not 4% of your balance

The single most common misreading is that the rule means withdrawing 4% of whatever the portfolio is worth each year. It does not. You take 4% in the first year, and from then on you take the same amount uprated for inflation, whatever the market has done.

The distinction is not academic, because the two strategies behave in opposite ways in a bad year. Take a corpus of ₹2 crore and a first-year withdrawal of 4%, which is ₹8,00,000. Assume inflation of 6% in year one and a market fall of 20% over the same year, so the corpus enters year two at roughly ₹1.53 crore after the withdrawal and the fall.

1. Under the actual rule, year two’s withdrawal is ₹8,00,000 raised by 6% inflation, so ₹8,48,000. Your income is protected; the portfolio absorbs the shock.
2. Under the misreading, year two’s withdrawal is 4% of ₹1.53 crore, so about ₹6,12,000. That is a 23% income cut in the year after a crash, arriving alongside 6% inflation.

The author revised it upward, not downward

Bengen has continued working on the question. In his 2024 book he raises the figure to 4.7%, which he calls a universal SAFEMAX, after widening the asset mix beyond large-cap US stocks and intermediate Treasuries to include small-cap value, microcap and international holdings, and adding an element that responds to market valuations at the point of retirement.

That is worth sitting with, because it cuts against how the rule is usually invoked. The figure most people quote is not Bengen’s original precise result and it is not his current one. It is a rounded intermediate value that acquired the word “rule” somewhere along the way and stopped being revised in the popular telling, while its author kept revising it.

Auditing the assumptions, one at a time

Whether the number transfers is a question about the assumptions, not the number. The useful exercise is to take each condition in the table above and ask whether it holds for the retirement you are actually planning.

The horizon row is the one that does the most damage quietly. Bengen's finding is explicitly a thirty-year result, and the arithmetic of a fifty-year drawdown is not the same problem with a longer number attached — the reasons why are worked through in FIRE in India against FIRE in the US, and the corpus consequences in the FIRE calculator.

The inflation row deserves its own caution. A national target is not what your household experiences; the basket that produces the headline number is not your basket, which is the whole argument of your personal inflation rate. Planning a withdrawal rate against a target you do not personally face is the same category of error as importing the rate without its allocation.

What the India-specific numbers are, and where they come from

You will find figures of 3% to 3.5% quoted widely as the Indian safe withdrawal rate, sometimes as low as 2.5%. These are worth taking seriously and worth attributing accurately.

The more useful move than adopting somebody else’s decimal is to notice how violently the required corpus responds to the rate you pick. The corpus you need is simply your annual spending divided by the withdrawal rate, so at 4% you need 25 times annual expenses, at 3.5% about 28.6 times, and at 3% a full 33.3 times.

Read across any row and the point is unmissable. Moving from 4% to 3% raises the corpus you must accumulate by a third on identical spending — on ₹1,00,000 a month, an extra ₹1 crore. That gap is a far larger planning fact than the rate itself, and it is why the decimal you adopt matters less than knowing how much later you retire if you are wrong about it.

How to use the number honestly

The 4% rule is best treated as a well-documented historical benchmark from a specific market, useful for orientation and useless as an authority. Three things follow from that.

1. Carry the assumptions with the number. If you quote 4%, you are also quoting a thirty-year horizon and an equity allocation between half and three-quarters. Dropping either while keeping the rate is not conservatism, it is a different plan wearing the study’s credibility.
2. Model your own horizon rather than adopting a rate. The rate is an output of a horizon, an allocation and an inflation path. Working backwards from someone else’s output discards the only three inputs you actually control.
3. Test the sensitivity, not the point estimate. The question worth answering is not “is it 3.5% or 4%” but how much later you retire if it turns out to be the lower one, and whether that gap is something you can absorb by saving more, working longer, or spending less.
4. Remember there is another way to answer the question entirely. A withdrawal rate is what you need when your retirement income has to come out of a corpus you manage. An assured payout answers it differently, by fixing the income and taking the corpus decision away from you — which is the trade the Unified Pension Scheme makes against NPS, and it comes with its own formula worth reading before assuming a guarantee is simpler.

If you are at the stage of asking whether you can stop contributing and still finish on time, the Coast FIRE calculator answers that specific question directly, and it is a more tractable one than choosing a withdrawal rate three decades in advance.

Questions worth asking

Who created the 4% rule and what did the original study test?

William Bengen, a financial planner, in “Determining Withdrawal Rates Using Historical Data” published in the Journal of Financial Planning in October 1994. He ran every rolling 30-year retirement window in United States market data from 1926 onward, using the S&P 500 for stocks and intermediate-term Treasuries for bonds with annual rebalancing and no fees, and asked what constant inflation-adjusted withdrawal rate would have survived even the worst starting year.

Does the 4% rule mean withdrawing 4% of your portfolio every year?

No, and this is the most common misreading of it. You take 4% of the portfolio in the first year only, then increase that rupee amount by inflation each year regardless of what the portfolio is worth. Taking 4% of the current balance annually is a different strategy with different behaviour: it can never run out, but it cuts your income sharply in a bad year, which is precisely when spending is hardest to reduce.

Has the 4% rule been revised since 1994?

Yes, by its own author. Bengen’s original analysis produced 4.15%, which he rounded down to 4% and called the SAFEMAX. In his 2024 book he revised the figure upward to 4.7%, which he terms a universal SAFEMAX, after widening the asset mix to include small-cap value, microcap and international holdings and adding a valuation-aware element. The number people quote is neither his original precise figure nor his current one.

What asset allocation does the 4% rule assume?

An equity-heavy one. Bengen found a 50/50 stock and bond split close to optimal, but recommended holding equity as close to 75 per cent as possible and in no case below 50 per cent. That matters because the rule is often quoted to people holding mostly fixed deposits and debt. The withdrawal rate and the asset allocation are a single package; taking the rate without the allocation is not applying the study.

Was the 4% rule ever meant to be a rule?

No. It was the answer to a bounded historical question: what was the largest withdrawal rate that would have survived the worst starting year in one country’s recorded market history. That is a description of the past, not a prediction, and it carries the specific inflation record, asset returns and time horizon of the data it was measured on. Bengen’s own later revision is the clearest evidence that it was never fixed.

Related

Sources

• William P. Bengen, “Determining Withdrawal Rates Using Historical Data”, Journal of Financial Planning, October 1994. The data period, asset classes, annual rebalancing, thirty-year horizon, the 4.15% result and the recommendation to hold equity as close to 75% as possible and never below 50%. Print journal article; no free authoritative copy is published online, so it is cited here bibliographically rather than linked to a third-party reproduction.
• Philip L. Cooley, Carl M. Hubbard and Daniel T. Walz, “Retirement Savings: Choosing a Withdrawal Rate That Is Sustainable”, AAII Journal, February 1998 — the Trinity study, source of the success-probability framing, tested on US data from 1925 to 1995 across allocations from all-bonds to all-stocks. Print journal article, cited bibliographically for the same reason.
• Bengen’s later revision of the figure to 4.7%, and his term for it, are set out in his 2024 book A Richer Retirement.
• Reserve Bank of India, Monetary Policy Framework. The inflation target is determined by the Central Government in consultation with the Bank once every five years under Section 45ZA of the RBI Act, 1934, and stands at 4% CPI inflation with a tolerance band of plus or minus 2%.

This is an explanation, not advice about your money. TLDR Money is not a registered investment adviser and earns no commission on any product mentioned. Figures are illustrative; your own numbers, taxes and circumstances will differ.