Monte Carlo Simulation for Retirement: Why Average Returns Lie
Discover why a single 'average' return can overestimate your retirement success — and what to use instead.
If you've ever plugged numbers into a retirement calculator and felt a warm glow of confidence, you may be building your future on a dangerous assumption. Most traditional calculators use a single "average" rate of return — typically 7–8% for equities — and project your portfolio in a smooth upward line. Reality is nothing like that.
The average is misleading
Two investors can earn the same "average" 7% return over 30 years but end up with portfolios that differ by hundreds of thousands of dollars. Why? Because the order of returns matters enormously when you're making withdrawals.
The Problem with Average Returns
Imagine Investor A gets exactly 7% every year. Investor B suffers a 30% crash in year one, then recovers with strong gains that bring the arithmetic average to the same 7%. Even though the averages match, their ending portfolios can differ dramatically — especially if they're making withdrawals along the way.
When you're drawing down a portfolio, early losses force you to sell more shares at depressed prices, leaving fewer shares to participate in future recoveries. Financial planners call this sequence of returns risk, and it's the single biggest reason deterministic projections can mislead retirees.
What Is Monte Carlo Simulation?
Monte Carlo simulation runs a model thousands of times, each time feeding it a different set of randomly generated inputs. The name comes from the famous casino in Monaco — a nod to the role randomness plays in the method.
Rather than asking "what happens if I earn 7% every year?" it asks "what happens across a wide range of plausible market histories?"
How It Works for Retirement
Define your inputs
Current portfolio, annual contributions or withdrawals, investment horizon, asset allocation, and expected return/volatility for each asset class.
Generate random return sequences
For each run, draw annual returns from a probability distribution (e.g., normal distribution with historical mean and standard deviation).
Simulate year by year
Apply the random return, subtract withdrawals (or add contributions), and carry the balance forward through each year.
Repeat 1,000+ times
Run the entire sequence many times to build a large sample of possible outcomes.
Aggregate and analyze
Sort results into percentiles, success rates, and confidence bands.
Sequence of Returns Risk in Action
Consider a retiree starting with $1,000,000 who withdraws $40,000 per year, adjusted for inflation:
| Scenario | Crash Timing | Portfolio at Year 25 | Outcome |
|---|---|---|---|
| Early crash | Years 1–2 | $0 (depleted) | Ran out of money |
| Late crash | Years 28–29 | $680,000+ | Healthy surplus |
Same total returns, same average — but completely different outcomes. Monte Carlo exposes this vulnerability by generating hundreds of different orderings.
Interpreting the Results
Monte Carlo Percentile Bands (1,000 simulations)
The shaded area shows the range of outcomes. Wider spread = more uncertainty.
The output is not a single number — it's a distribution. Here are the key metrics:
- Success rate — The percentage of runs where your portfolio survived the entire horizon. 85–95% is typically considered acceptable; 100% may mean you're under-spending.
- Median (P50) — The middle-of-the-road scenario. Half of simulations ended above, half below.
- 10th percentile (P10) — The "bad luck" scenario. Only 10% of runs were worse. A useful stress test.
- 90th percentile (P90) — The "good luck" scenario. Useful for legacy goals or deciding when to increase spending.
Why 1,000 simulations?
With fewer runs, the success rate can swing by several percentage points due to sampling noise. At 1,000 runs, the estimate stabilizes — rerunning typically produces a result within one percentage point. Some tools go to 10,000, but the incremental insight is usually modest.
A Practical Example
Suppose you're 35, earn $90,000/year, save 30%, and have $150,000 invested. You plan to retire at 50 and spend $45,000/year. You invest in a global equity portfolio (5.5% real return, 15% standard deviation).
| Method | Success Rate | Worst Case (P10) | Best Case (P90) |
|---|---|---|---|
| Deterministic (7%) | 100% | Same as median | Same as median |
| Monte Carlo (1,000 runs) | 78% | Portfolio depleted at 72 | Surplus of $1.2M |
The deterministic calculator says "you're fine." Monte Carlo reveals that 220 out of 1,000 simulated futures end in failure. That information is actionable: save more, plan part-time income, or adjust your asset allocation.
Making Better Decisions Under Uncertainty
Monte Carlo doesn't predict the future. What it does is replace a false sense of certainty with an honest accounting of the range of possibilities. Instead of asking "will I have enough?" you start asking "how likely am I to have enough, and what can I do to improve those odds?"
Run your own Monte Carlo simulation with 1,000 scenarios. Free, no signup.
Try the FIRE Calculator →See how Monte Carlo results translate into a concrete year-by-year retirement plan with our Wealth Plan tool.
Build Your Wealth Plan →