What the record actually says · six assumptions
If you read nothing else
| Dial | Page default | What I would set | Why |
|---|---|---|---|
| Inflation | 3% | 3% | The 97-year average, and above today's forecasts |
| Return while investing | 7.9% | 6.5% | Your runway is 10 years and 10-year forecasts are 4% to 6.5% |
| Return in retirement | 6.4% | 6.0% | 70/30 at today's bond yields |
| Withdrawal rate | 3.25% | 3.3% | Between the historical floor and the forecast-based number |
| Plan through age | 95 | 100 | Five more years costs 3.9%. Cheapest insurance on the page |
| Reach the money at | 59.5 | 59.5 | Retiring at 51 means the rule of 55 never applies to you |
Page default 3% · leave it there
Stay at 3%. It matches the long record, it sits above what the forecasters expect, and being wrong high here is the safe direction: it makes the plan ask for more, not less.
One thing the 3% hides. Retiree spending is weighted toward health care, which has run well above headline inflation for decades. If a big share of your later spending is medical, the real number for you is higher than the one on the news.
Page default 7.9% · probably too generous for a 10-year runway
| 1928 to 2024 | Nominal | After 3% inflation |
|---|---|---|
| US large-cap stocks | 9.9% | 6.7% |
| 10-year Treasuries | 4.5% | 1.5% |
| Cash | 3.3% | 0.3% |
That is the number everybody quotes, and it is real. It is also 97 years long, which is longer than the 10 years you actually have before you stop working.
| Source | US stocks | Bonds | Horizon |
|---|---|---|---|
| Vanguard, June 2026 | 4.2% to 6.2% | — | 10 years |
| Verus 2026 | 6.5%* | 4.8% | 10 years |
| Morningstar | ~7.0% | ~4.0% | 30 years |
* an arithmetic average. See the first trap below, it is worth about 1.2 points.
Every one of them lands under the historical 9.9%, and they land there for the same reason: stocks are expensive relative to earnings right now, and starting valuation is the one thing that has any predictive power over a decade.
Page default 6.4% · close, but the bond half is knowable
Bonds are the unusual case in finance where the future is nearly readable. The yield on the day you buy explains most of what a bond portfolio returns over the following decade. Today the 10-year Treasury pays 4.79% and the 30-year pays 5.25%. So the bond side of a retirement portfolio is worth roughly 4.8%, and that is not a guess.
| Mix | Nominal | Real |
|---|---|---|
| 70 stocks / 30 bonds | 6.0% | 2.9% |
| 60 / 40 | 5.8% | 2.7% |
| All bonds, a bridge portfolio | 4.8% | 1.7% |
Built from 6.5% stocks and 4.8% bonds. Your 6.4% default is not wrong, it just quietly assumes stocks do a little better than the forecasters expect.
Cross-check only · but worth understanding
This box does not drive your number. It drives the cross-check, which exists so you can compare against calculators built on the 4% rule. Three separate bodies of research answer it and they do not agree.
Morningstar redoes this every year from current market assumptions rather than history. Their 2026 answer is 3.9% for 30 years at a 30% to 50% stock allocation. Their published series:
| Year | 2021 | 2022 | 2023 | 2024 | 2025 | 2026 |
|---|---|---|---|---|---|---|
| Safe starting rate | 3.3% | 3.8% | 4.0% | 3.7% | 3.7% | 3.9% |
Stretch the horizon and it drops: 3.5% at 35 years, 3.2% at 40. Yours is 44.
Pfau ran the same test across 20 developed countries from 1900 to 2015. Only five supported a 4% withdrawal. In eleven of the twenty it was under 3%. Japan's 1937 retiree could take 0.27%. Austria's 1914 retiree, 0.07%. The United States finished near the top of that table at about 4.0%.
That is the uncomfortable finding buried under the 4% rule. It is not a law of markets. It is the result for the country that won the century.
Page default 95 · go to 100, it is nearly free
By that arithmetic, planning to 95 is already conservative and planning to 100 looks close to paranoid. Which would be the right conclusion if the two errors cost the same. They do not.
Dying at 88 with money left over costs you some spending you could have enjoyed. Living to 97 with the account empty at 95 costs you your independence at the age you can least do anything about it. Those are not the same size mistake, so they do not deserve the same weight.
And here is the part that settles it. In your plan, moving the horizon from 95 to 100 raises the number you need from $4,889,327 to $5,077,987, a difference of 3.9%. Five extra years of certainty for less than four cents on the dollar, because money needed 55 years out discounts to almost nothing today.
You had this set to 55 · it should be 59.5
| Leaving at 51 | Years to bridge | Needed outside the 401(k) |
|---|---|---|
| If the rule of 55 applied | 4 | $690,007 |
| Reality, 59.5 | 8.5 | $1,377,198 |
Today's dollars, at $15,000 a month, at a 6% return.
That is double the bridge, and it is the single most useful thing on this page. It is also exactly what your taxable brokerage is for.
A 72(t) lets you take substantially equal periodic payments from an IRA at any age with no penalty. The catch is real: payments must continue for five years or until 59.5, whichever is longer, and breaking the schedule triggers retroactive penalties plus interest on everything you already took. For a 51-year-old that is an 8.5-year commitment to a fixed withdrawal, locked in before you know what the market does.
Roth contributions, as opposed to earnings, come out any time tax and penalty free. That is bridge money most people forget they have.
Your plan, three sets of assumptions
Retiring at 51 on $15,000 a month, $1,000,000 invested, $80,000 a year going in, Social Security of $6,243 from 67, spending it down by 95. All figures in 2036 dollars.
| Assumptions | Need at 51 | Will have | Short by | Earliest |
|---|---|---|---|---|
| US history, 8.5% / 7.0% | $4,300,392 | $3,634,917 | $665,475 | 53 |
| Page default, 7.9% / 6.4% | $4,637,622 | $3,476,087 | $1,161,535 | 55 |
| My suggestion, 6.5% / 6.0% | $4,889,327 | $3,132,492 | $1,756,835 | 57 |
| Forecasts, 5.0% / 4.8% | $5,804,972 | $2,803,005 | $3,001,967 | 61 |
Nobody knows which row is the real one. What you can know is that the honest answer to "when can I retire" is not a date, it is 53 to 61, and the way you shrink that range is by saving more and rechecking every year, not by picking a better number for the box.
Both of them make things look better than they are
A portfolio that gains 50% then loses 50% has an average annual return of zero and has lost a quarter of your money. The average of the yearly numbers, the arithmetic mean, always overstates what you actually compound at, and the gap grows with volatility. For a stock portfolio it runs about 1.2 points.
So when a firm publishes 6.5% for US stocks, check which one it is. Verus's 6.5% is arithmetic, which compounds at roughly 5.3%. This calculator compounds, so it wants the lower one.
Two retirees can earn the identical average return over thirty years and one runs out of money while the other dies rich. The difference is whether the bad years came first. Selling shares to live on in a down market takes out capital that never comes back, and no average return in any box on this page captures that risk.
It is why Morningstar's optimal withdrawal-rate portfolio is only 30% to 50% stocks rather than 70%, even though 70% has the higher expected return. And it is why hitting your number exactly, on the day, is not a safety margin. It is a coin flip you have arranged to be roughly fair.
Checked September 2026