Tools/Retirement withdrawals & early access/Retirement Monte Carlo Simulator✓ CHECKED AGAINST WORKED EXAMPLES · SEP 29, 2026

Will your money last?

Retirement calculator with Monte Carlo simulation: 5,000 randomized market paths against your spending.

Edit on my map
Return assumptions
Assumed yearly real returns: stocks 7% (±17), bonds 2.5% (±4.5). The middle path compounds lower than the average.
CHANCE YOUR MONEY LASTSAT RISK
<1%
2 of 5,000 simulated 30-year retirements ended with money left. Well below 85%. Spending, length or balance would need to change to reach it.
Withdrawal rate
14.3%
Median ending
$0
10th pct ending
Depleted
Median failure
Year 9
UNDERSTAND YOUR RESULT
LIBRARY CHAPTERTurning Savings into IncomeHow much a portfolio can pay each year, why the order of returns early in retirement matters, ways to make a withdrawal plan sturdier, and a basic order for drawing from accounts.LIBRARY CHAPTERLife After FI: Making the Money LastSequence-of-returns risk and the ways to manage it, how spending really behaves in retirement, how retiring early changes Social Security, a yearly review routine, and using the freedom, including giving.QUICK ANSWERWhat is the 4% rule?The 4% rule says that if you withdraw 4% of your portfolio in the first year of retirement and raise that amount with inflation each year, your money lasted at least 30 years in every period of US market history that its author tested from 1926. It is a finding about the past, built for 30-year retirements, not a guarantee.
Terms:4% ruleSequence of returns riskMonte Carlo simulation

Where 5,000 futures go

10–90th pctMedianDepleted
$0$132k$265k$397k$529kYr 0Yr 10Yr 20Yr 30
ENDING BALANCES
Depleted$1.38M+
4,998 paths end at zero (red). The median ends at $0, 0.0× the start.

The middle line is the median path: it ends at $0 after 30 years. The shaded band holds the middle 80% of outcomes, and its lower edge ends at zero. Each bar counts paths by ending balance; balances above $1.38M share the last bar.

What moves the needle

Each row re-runs the calculation with one change. Click to apply.

How it's computed

FORMULA
B(t+1) = (B(t) − W) × (1 + r(t))
r(t) ~ Normal(μ, σ), capped at −50% and +80%
P = #{ B(T) > 0 } ÷ N
  • r(t) ~ Normal(μ = 5.7%, σ = 13.3%) for a 70/30 mix, in real (after-inflation) terms. μ is the arithmetic average of the yearly returns, so the middle (median) path compounds lower: swings cost growth.
  • Stock and bond volatility is interpolated linearly with the mix, as if stocks and bonds moved together. Real portfolios diversify a little, so a mix looks somewhat riskier here than in practice, and shifting toward bonds helps a little less.
  • W is flat in today’s dollars: it rises with inflation in nominal terms.
  • Each year: withdraw first, then apply that year’s return, which is capped between −50% and +80%.
  • N = 5,000 paths, seeded — the same inputs always give the same answer, which can differ by about a point from the true figure.
  • No taxes, fees, Social Security, pension or other income, and no change in spending with markets or age.
WORKED EXAMPLE · SAMPLE NUMBERS
Start with $420,000 and withdraw $60,000 in year one. If that year’s sampled real return is −12%, the balance becomes ($420,000 − $60,000) × 0.88 = $316,800. Repeat for 30 years with a fresh draw each year; that is one path. Run 5,000. Here 2 finished above zero, so P = 2 ÷ 5,000 = <1%.
Keep this number honest as your life changes.
Put it on your Money Map and it re-runs as you change the seven numbers. It stays in this browser, and the calculator stays free.
Open your Money MapTell me when bank sync opens

Questions about this result

There is no official cutoff. This page marks 85% and above as robust, a common planning convention rather than a rule. A simulated path that runs out is not the same as a retiree who is ruined: the simulation never adjusts spending, and real retirees usually can. Kitces.com (Tharp, 2021) argues that lower probabilities can be reasonable for people willing to adjust, and decision rules that cut spending after bad years (Guyton and Klinger, 2006) let a portfolio start at a higher rate. This simulator holds spending flat, so it does not show that effect.
Mostly three choices: the return assumptions, whether each year’s return is drawn independently, and whether spending stays fixed. Taxes, fees and other income also differ between tools. Every choice made here is stated in the methodology on this page.
No, and neither a pension nor any other income, so the result is conservative for anyone who expects them. Spending is one flat figure for every year, so Social Security cannot be netted off correctly when it starts part-way through: lowering spending for all years would overstate the chance for anyone who retires before they claim. The Social Security break-even calculator on this site estimates a benefit for a claiming age; a simulation that starts at that age can lower spending by the after-tax benefit.
The base case assumes stocks return 7% a year after inflation on average and bonds 2.5%, with the swings shown above; these are the round-number assumptions built into the engine, not a fit to any one period of history. High starting valuations (a high Shiller CAPE ratio) have historically been followed by lower returns, so the Conservative and Stress cases are the more cautious tests in that setting. Fees come off returns point for point, and the Conservative case lowers the averages by 2 points for stocks and 1 point for bonds, so it is also a rough test of lower returns plus fees.
No. Withdrawals from pre-tax accounts are taxed, so the amount you can spend is lower than the amount withdrawn, and investment fees come off returns point for point: a 1% fee is like earning 1 point less than the assumed average. Lowering spending by the tax you expect, or using the Conservative case, shows how sensitive the result is.
THE LEDGER · 10 min5 Retirement Mistakes That Cost $100K+ (Part 3 of 3)
THE LEDGER · 28 min5 Retirement Mistakes That Cost $100K+ (Part 2 of 3)
GUIDE · $39The Retire-Early Playbook
YOUR MAP · 0 of 7 doneNext: FIRE CalculatorContinue →