Safe Withdrawal Rate for Long Retirements: What 98 Years of Data Say
We ran 98 years of markets through our simulator to measure how far the safe rate falls as retirement gets longer.
- Core Concept: The safe withdrawal rate is the initial share of the portfolio you can withdraw each year, adjusted for inflation, without running out of money in any historical period of the chosen length.
- Why It Matters: The 4% rule was designed for retirements of about 30 years. Retire at 45 or 50 and the money has to last 40 or 50 years, and every tenth of a point changes how much you need.
The result: each extra decade takes off about two tenths of a point
With US data from 1928 to 2025 and 0.2% fees, the highest rate that never ran out with a 60/40 portfolio falls from 3.6% over 30 years to 3.35% over 40 and 3.15% over 50.
For each horizon and portfolio we searched for the highest initial rate that would have lasted in every possible historical period since 1928, withdrawing the same amount in today’s money every year. With a 60/40 portfolio, that rate falls from 3.6% over 30 years to 3.15% over 50.
In money: to spend €30,000 a year for 50 years, 3.15% means saving about €952,000, against the €750,000 the 4% rule asks for.
We use Aswath Damodaran’s (NYU Stern) annual series: the S&P 500 with dividends, 10-year US Treasury bonds and inflation. Withdrawals happen at the start of each year, the portfolio is rebalanced yearly and 0.2% of fees is deducted. You can reproduce any figure in our retirement simulator; details are in the methodology.
Safe rate by horizon and portfolio
The 80/20 portfolio was the most resilient at every horizon, from 3.7% over 30 years to 3.3% over 50. The most conservative mix, 40/60, falls the most as the horizon lengthens.
| Horizon | 40/60 portfolio | 60/40 portfolio | 80/20 portfolio | 100/0 portfolio |
|---|---|---|---|---|
| 30 years | 3.5% | 3.6% | 3.7% | 3.4% |
| 35 years | 3.3% | 3.45% | 3.55% | 3.3% |
| 40 years | 3.15% | 3.35% | 3.45% | 3.25% |
| 45 years | 3.05% | 3.25% | 3.4% | 3.2% |
| 50 years | 2.85% | 3.15% | 3.3% | 3.15% |
Two takeaways. First, more bonds do not mean more safety in long retirements: with 40/60 the safe rate over 50 years is only 2.85%. Second, 100% stocks is not the best either, because a retirement starting in 1929 sinks the portfolio just as withdrawals begin.
Success rates at 3%, 3.5% and 4%
Over 50 years with a 60/40 portfolio, 4% lasted in 63% of historical periods, 3.5% in 88%, and 3% in all of them.
| Initial rate | 30 years | 40 years | 50 years |
|---|---|---|---|
| 3% | 100% / 98% | 100% / 94% | 100% / 90% |
| 3.5% | 100% / 95% | 97% / 88% | 88% / 82% |
| 4% | 94% / 89% | 83% / 79% | 63% / 73% |
| Initial rate | 30 years | 40 years | 50 years |
|---|---|---|---|
| 3% | 100% / 96% | 100% / 93% | 100% / 89% |
| 3.5% | 100% / 93% | 98% / 87% | 92% / 83% |
| 4% | 94% / 88% | 86% / 81% | 78% / 76% |
Monte Carlo randomly combines whole years and gives somewhat worse results at low rates, because it creates sequences that never happened. When both methods agree, the conclusion is more robust.
What guardrails, a pension or a cash buffer change
At 4% for 45 years, guardrails raised historical success from 78% to 96%, a pension from year 20 lifted it to 85% and a 3-year cash buffer lowered it to 57%.
| Option | Historical | Monte Carlo | Rate that never failed |
|---|---|---|---|
| Fixed spending | 78% | 76% | 3.25% |
| Guardrails (80% floor) | 96% | 86% | 3.9% |
| €12,000 pension from year 20 | 85% | 85% | 3.6% |
| 3-year cash buffer | 57% | 69% | 3.15% |
Guardrails are the strongest lever, but they have a cost: in half of the periods spending hit the 80% floor at some point. The pension helps because it reduces what has to come out of the portfolio in the second half. The cash buffer did not help in any case we tested: it avoids selling after a bad year, but idle cash earns little for decades.
Limits of the study
This is US data in dollars, before taxes, and over 50 years there are only 49 historical periods, which overlap. Treat the results as an order of magnitude, not a guarantee.
- US data. It is the longest reliable public series, but the US was one of the best markets of the 20th century; for a European portfolio it pays to be a little more cautious.
- Few independent periods. Over 50 years there are 49 periods and they share most of their years; at 4%, failures cluster in starts from 1929 to 1939 and from 1959 to 1973.
- No taxes. Spending is before taxes on sales.
- The past does not guarantee the future. This is an educational study, not investment advice.
For context on the 4% rule, read our Trinity study analysis and the guide to sequence-of-returns risk.
Frequently Asked Questions (FAQ)
How much money do I need for a 50-year retirement?
At the rate that never failed over 50 years, about 32 times your annual spending with a 60/40 portfolio (3.15%) or about 30 times with 80/20 (3.3%). The 4% rule asks for 25 times.
Why isn’t 100% stocks the safest?
Because the worst starts, such as 1929, sink a portfolio without bonds just as withdrawals begin. In our data, 80/20 had a higher safe rate than 100/0 at every horizon.
Do these results apply to a European investor?
They show how much the horizon and the portfolio matter, but US markets returned more than most. Using a rate a little below the table compensates for that difference.
Keep reading
How to shield your portfolio from Sequence of Returns Risk (SORR) in 2026. Guyton-Klinger guardrails, bond tents, and multi-year cash bucket strategies.
Read the analysis →FIRE & Financial IndependenceThe 4% Rule and Updated Trinity Study (2026): Safe Withdrawal Rates and InflationUpdated 2026 guide to the 4% rule and the Trinity Study. Discover safe withdrawal rates (SWR), sequence of returns risk, and 50-year FIRE simulations.
Read the analysis →Self-Employed & FIRESelf-Employed Retirement in 2026: Closing the 40% Pension DeficitSelf-employed retirement guide 2026. The 40% public pension cliff, net-income contribution systems, tax-advantaged solo plans, and the 4% FIRE rule.
Read the analysis →