Safe Withdrawal Rate Calculator

• Runs a steady-return projection and a 1,000-trial Monte Carlo simulation side by side, using a fixed seed so the result is the same every time.
• Gives the chance your money lasts to your target age, a percentile chart of balances, and the highest starting withdrawal that still reaches 90%.
• Lets you set the withdrawal as a percentage or a dollar amount, with stock and bond returns, volatility and inflation under your control.
• Takes Social Security or a pension into account so you can see how guaranteed income lowers the risk.

How to use this calculator

Enter the portfolio you expect to have at retirement, then choose whether the first-year withdrawal is a percentage of it or a dollar amount. Set the number of years the money must last and the share held in stocks. The headline is the percentage of simulated retirements in which the money lasted that long.

Open Advanced to change the expected returns and volatility of stocks and bonds, the correlation between them and inflation. You can also enter Social Security or pension income and the year it starts. The calculator subtracts it from the withdrawal, so the portfolio only pays the difference. The seed number picks which set of random markets is used; leave it alone for reproducible results or change it to see that the answer is a range and not a single number.

The 4% rule and what it really said

The 4% rule comes from William Bengen's 1994 paper in the Journal of Financial Planning. He tested retirements that began in every year since 1926 in the US. A retiree who withdrew 4% of the starting portfolio in the first year, then raised the dollar amount each year with inflation, never ran out of money in 30 years in any historical start year, using a portfolio of roughly half to three-quarters in stocks, the rest in bonds. The 1998 Trinity study (Cooley, Hubbard and Walz) turned that into success rates, and found that 4% lasted 30 years in about 95% of historical periods in stock-heavy mixes.

Three limits matter. It is built on US history, which was favourable. It assumes a fixed 30-year period, a constant inflation-adjusted amount, and no taxes or fees. And it is a starting rule of thumb, not a promise. Morningstar's retirement-income research, which uses forward-looking return assumptions, gives 3.9% for 2026 for a 30-year retirement at 90% confidence with a mix of 30% to 50% in stocks. Bengen himself has since argued that a higher figure fits some mixes. The sensible reading is that a starting rate near 4% is a reasonable place to begin, and that your own horizon, mix and flexibility can push it up or down.

How the simulation works

Each year: withdraw first, then the balance earns the year's return
Withdrawal = first-year amount, raised with inflation (constant in today's dollars)
Net draw = max(0, withdrawal - guaranteed income once it has started)
Returns: log-normal, stocks and bonds correlated, rebalanced to the mix yearly
A trial fails if the balance after a withdrawal is zero or less in any year
Success = trials that never fail / 1,000

All balances are shown in today's dollars, so a withdrawal that rises with inflation is a flat amount. The simulation draws 1,000 paths from a seeded random generator. The same inputs always return the same paths, so the answer is repeatable and you can compare changes without noise. The steady-return projection uses the average return every year with no ups and downs, which is what many simple calculators do.

A worked example

Start with $1,000,000, withdraw 4% ($40,000) in the first year, raise it with inflation, hold 60% in stocks and 40% in bonds and plan for 30 years. With stocks returning 7% a year, bonds 4%, inflation 2.5% and typical volatility, the money lasted in 73% of the 1,000 simulated markets. The median ending balance, in today's dollars, was $328,206. In the bad-luck case (10th percentile) the balance ran out before year 30, and in the good-luck case (90th percentile) it ended at $1,681,547.

The steady-return projection tells a rosier story. Earning the same average return every year it ends with $551,696. That gap is the point: a constant return ignores the bad early years that sink real plans. The highest starting withdrawal that still reached 90% success was about $33,903 a year (3.4% of the portfolio).

Those results depend on the assumptions. Lower the stock return to 6% and success at 4% falls to 63%. They are not predictions, they show how sensitive a plan is.

How the starting rate changes the odds

The table under the results re-runs the same 1,000 markets at starting rates from 3% to 6%. For the example above, success was 96% at 3%, 87% at 3.5%, 73% at 4%, 57% at 4.5%, 40% at 5% and 18% at 6%. The curve is steep. Each half-point step in the starting rate moved the odds by roughly 9 to 17 points, which is why small changes to spending matter more than most people expect.

A dollar withdrawal works the same way. Taking $50,000 a year from the same $1,000,000 is a 5% start and succeeds 40% of the time under these assumptions.

Time horizon and the mix of stocks and bonds

The longer the retirement, the lower the safe rate. With the same inputs, a 20-year plan succeeds 98% of the time, a 30-year plan 73% and a 40-year plan 48%. Anyone retiring early, in their 50s, should plan for a 40- or 50-year horizon, which calls for a noticeably lower starting rate than 4%.

The stock share cuts both ways. Here 40% stocks gives 73% success, 60% gives 73% and 80% gives 71%. More stocks usually raises the typical outcome and the chance of keeping up with inflation, but it also widens the spread. In this example the differences between 40%, 60% and 80% are small, because the added return from stocks is offset by the added risk. A very low stock share can fail from inflation and low returns instead. Bengen and the Trinity authors looked at mixes of about 50% to 75% stocks. Try a few mixes and look at the bad-luck line as well as the median.

Sequence-of-returns risk

Two retirees can earn the same average return over 30 years and end up in different places. If the weak years arrive first, the portfolio is smaller when the recovery comes, because withdrawals were taken from a shrunken balance. If the strong years arrive first, the retiree has a cushion. Accumulating, the order barely matters; withdrawing, it matters a great deal.

That is why the first five to ten years of retirement carry so much of the risk. The percentile chart shows it: the bad-luck line falls early and never recovers, while the good-luck line climbs. Ways to reduce the risk include a lower starting rate, holding a year or two of spending in cash or short-term bonds, cutting spending after a bad year, and covering essential costs with guaranteed income.

Social Security and pensions

Guaranteed income lowers how much the portfolio has to pay. Add $20,000 a year of Social Security that starts after five years, and the 4% plan above improves from 73% to 100%. The pension or benefit is entered in today's dollars and assumed to rise with inflation, as Social Security does. Enter the household total, and use the start year to model a delayed claim, for example at 67 for someone who retires at 62.

The Social Security Break-Even Age Calculator shows how the claiming age changes that benefit, and the two tools work together: a larger benefit from waiting reduces the draw on savings in the years you most need it.

How to read the results

Start with the headline percentage, then look at the chart. The dark line is the median: half of the simulated retirements ended above it and half below. The red line is the 10th percentile, the bad-luck case, where nine in ten markets did better. In the example the median ends at $328,206 while the bad-luck line reaches zero, so a retiree who sees the median as the plan is ignoring a real chance of running out.

The year-by-year table shows the share of simulations that have run out of money by each year. That column tells you when the danger sits. In the example, none had run out by year 15, 2% had by year 20 and 27% by year 30. Failures arrive late, even though the damage is usually done by bad returns early on, so a plan can look healthy for two decades and still fail, and by then changing course is hard.

Finally, compare the success rate with the highest starting withdrawal at 90% success. If you are spending well above that figure, the question is not whether you will have to adjust, but when. Small cuts early cost less than large cuts late.

Choosing the return and volatility inputs

The most uncertain inputs are the expected returns. Historical US stock returns before inflation have averaged near 10% a year, but many planners use lower figures for the future because stock valuations are high compared with history and bond yields are lower than in past decades. The defaults of 7% for stocks and 4% for bonds are deliberately cautious. Bond returns depend on the type: a short-term Treasury fund and a long-term corporate fund behave differently. Volatility is the typical size of the yearly swing: about 16% for a broad stock fund and about 6% for an intermediate bond fund are common planning figures. Raising volatility makes the spread wider and the success rate lower even when the average return is unchanged.

Run the calculator more than once. A cautious case, a middle case and an optimistic case show how much of your plan rests on one assumption, which is more useful than any single success percentage.

Flexible spending is the strongest lever

The model assumes a fixed, inflation-adjusted withdrawal, which is the strictest test. Real retirees adjust. Skipping an inflation raise after a down year, trimming discretionary spending by 10%, or working part time for a while can lift the success rate more than any change to the investment mix. Some research suggests that a flexible plan can start with a higher withdrawal than a fixed one for that reason. This calculator does not model that behaviour, so treat its answer as the case where you never change course.

Assumptions and limitations

• Return assumptions are yours. The defaults (stocks 7%, bonds 4%, inflation 2.5%) are planning assumptions, not forecasts. Enter returns after fund fees.
• Random, not historical. The paths come from a statistical model with log-normal returns. Real markets have fatter tails and runs of bad years that this model understates.
• Constant volatility and correlation. They do not change with market conditions.
• Fixed spending. The withdrawal is flat in real terms. Real spending usually changes over retirement.
• No taxes. Withdrawals from pre-tax accounts are taxed, so you may need to take out more than you spend.
• No extra costs. Long-term care, large one-off expenses and advisory fees beyond those in your returns are not included.
• Success is not a promise. A 90% result means 10 in 100 simulated markets ran out of money.

Common mistakes

Treating 4% as a law. It is a historical finding with limits, not a guarantee.

Reading the steady-return line as the plan. The Monte Carlo result is the more honest test.

Using a 30-year horizon for a 50-year retirement. Early retirement needs a lower starting rate.

Counting Social Security twice. Enter it as guaranteed income and leave it out of the withdrawal.

Ignoring tax and fees. A 4% withdrawal from a traditional IRA is not 4% of spendable income.

Chasing a high success number. Aiming for 100% usually means spending much less than you need. Many plans aim for 85% to 90% and adjust as they go.

Frequently asked questions