Alternatives
The arithmetic is one division. The controversy is entirely in the denominator, and nobody has settled it. Here is the table, and an honest account of how contested the input is.
Most people who imagine winning a jackpot are not imagining a number. They are imagining an income — never working again, at a particular standard of living. Those two things are related by a single equation, and it is worth being able to run it in your head.
capital = annual income ÷ withdrawal rate
That is it. It is the rearrangement of income = capital × withdrawal rate, which is exactly what the site's finance module computes, and the whole difficulty lives in choosing the rate.
| Target annual income | at 2.5% | at 3% | at 3.5% | at 4% | at 5% |
|---|---|---|---|---|---|
| $50,000 | $2,000,000 | $1,666,667 | $1,428,571 | $1,250,000 | $1,000,000 |
| $75,000 | $3,000,000 | $2,500,000 | $2,142,857 | $1,875,000 | $1,500,000 |
| $100,000 | $4,000,000 | $3,333,333 | $2,857,143 | $2,500,000 | $2,000,000 |
| $250,000 | $10,000,000 | $8,333,333 | $7,142,857 | $6,250,000 | $5,000,000 |
| $1,000,000 | $40,000,000 | $33,333,333 | $28,571,429 | $25,000,000 | $20,000,000 |
Read across a row before reading down a column. The choice of rate moves the answer by a factor of two. A $100,000 income needs $2 million or $4 million depending on a decision nobody can make for you with certainty. That sensitivity is the actual subject of this article; the division is trivial.
The figure everyone quotes traces to William Bengen's October 1994 Journal of Financial Planning article "Determining Withdrawal Rates Using Historical Data", and to the 1998 study by Philip Cooley, Carl Hubbard and Daniel Walz of Trinity University, published in the AAII Journal, which backtested stock-bond portfolios against 1925 to 1995 data over withdrawal periods of 15 to 30 years.
Their conclusion, quoted rather than paraphrased: for level withdrawals, "withdrawal rates of 3% and 4% are extremely unlikely to exhaust any portfolio of stocks and bonds", and for inflation-adjusted withdrawals, "withdrawal rates of 3% to 4% continue to produce high portfolio success rates for stock-dominated portfolios."
Now the qualifications, which are as important as the finding and which the authors themselves supplied. They stressed that "selection of a withdrawal rate is not a matter of contract but rather a matter of planning" and that "mid-course corrections likely will be required." Later critics went further: the approach has been attacked for creating "a constant, non-volatile spending plan using a risky, volatile investment strategy", and the economist Laurence Kotlikoff has argued the rule "has no connection to economics".
Four specific reasons the number is not a law:
1. It is a backtest of one country over one period. US markets from 1925 to 1995 were unusually good by international standards. A rule derived from the single best-performing large equity market of the twentieth century is a rule with survivorship built in.
2. It assumes a 30-year horizon. Win a jackpot at 30 and you may need 60 years of income. The sustainable rate falls as the horizon lengthens, and at a long enough horizon it approaches the portfolio's expected real return — which for equities over 1928 to 2025 was well below the 10.02% nominal geometric mean, because that figure is before inflation.
3. It ignores tax and fees entirely. Both come off the top. A 4% gross withdrawal with 1% of fees and tax on the income is not a 4% net withdrawal.
4. Sequence risk is asymmetric when you are drawing down. A crash early in retirement is far more damaging than the same crash late, because you sell units at the bottom to fund living costs and those units never recover. Two portfolios with identical average returns can produce opposite outcomes purely on ordering.
The reasonable position: treat the rate as a planning band, not a constant. Something between 3% and 4% for a long horizon, revisited, with spending flexible enough to fall in bad years.
This is where the table stops being abstract, because the numbers a lottery advertises and the numbers this table needs are further apart than people expect.
A $1,000,000 Match-5 Powerball prize. In the US this is ordinary income; at the top federal marginal rate of 37% and ignoring state tax entirely, $630,000 remains. At 4%:
$630,000 × 0.04 = $25,200 a year
A million dollars, in the sense most people mean it, is a modest supplementary income — not a retirement.
A $100,000 income for life. From the table, that needs $2,500,000 at 4%. Working backwards through the standard deductions — cash value at roughly 50% of the advertised annuity, then 37% federal tax — the advertised jackpot required is:
$2,500,000 ÷ (0.50 × 0.63) = about $7.94 million advertised
So the headline jackpot that genuinely buys a $100,000 income for life is roughly eight million dollars, not two and a half. That gap is the entire content of what the advertised jackpot really means, and the after-tax prize calculator applies the deductions for your jurisdiction.
A $1 billion advertised jackpot. $500m cash, $315m after 37% federal, and at 4% that is $12.6 million a year indefinitely, without ever touching the capital. This is the one case where the arithmetic matches the fantasy.
The equation is symmetric, and the other direction is the useful one for anyone not holding a winning ticket:
income = capital × withdrawal rate
A portfolio built to $500,000 supports $20,000 a year at 4% — not freedom, but a material change to what work has to earn. At $1,250,000 it supports $50,000. And $1,250,000 is a number reachable by the arithmetic in invest instead of tickets and realistic paths to a windfall, given enough decades and a large enough contribution.
One final honesty note. Everything above is nominal. To hold real purchasing power for thirty or sixty years, the withdrawal must rise with inflation, which is precisely the case the Trinity study found harder to sustain. If you want a single sentence: the capital figure in that table is a floor, not a target, and the rate you pick is a bet you are making about the next forty years.
Last verified: 2026-08-29