Monte Carlo for Retirement Plans
You've spent decades building a portfolio. Now you're about to leave your paycheck behind. The question that keeps you awake isn't whether stocks go up or down this year—it's whether your money will actually last 30 years. Monte Carlo retirement simulations don't predict the future. Instead, they ask a harder question: if we run your plan against thousands of plausible market histories, how often does it survive?
Quick definition
Monte Carlo simulation randomly generates thousands of market return sequences (based on historical volatility and correlation patterns) and runs your retirement spending strategy through each one, measuring what fraction "succeed" (portfolio still has money at your target age).
Key takeaways
- Monte Carlo strips away false certainty by testing your plan against multiple realistic market paths, not just the average path
- Success rates above 90–95% are generally considered sustainable; below 85% signals high depletion risk
- The outputs are sensitive to assumptions about returns, spending, longevity, and inflation—small changes in inputs can swing success rates 10–20%
- Historical return sequences matter more than average returns alone; clustering of bad years can derail an underfunded plan
- Monte Carlo is a screening tool, not prophecy; it cannot predict the next recession but can show whether your strategy tolerates bad timing
How Monte Carlo Retirement Planning Works
Traditional retirement planning asks: "If average returns are 7%, can I safely withdraw 4%?" Monte Carlo asks: "Across 10,000 different paths the market could take, what percentage of those paths leave me solvent at 95?"
Flowchart
The mechanics are straightforward. A Monte Carlo engine:
- Samples historical returns to estimate mean returns, volatility, and correlation between asset classes
- Generates thousands of random but realistic return sequences where some years are +20%, others are −10%, all respecting observed market behavior
- Runs your withdrawal plan through each sequence (e.g., "withdraw $50,000 adjusted for inflation each year")
- Counts successes (portfolio never hits zero) and reports the failure rate
For example, if 9,400 of 10,000 scenarios keep your money alive until age 95, you have a 94% success rate.
The power of this approach is that it captures sequence of returns risk—the reality that a 20% crash in year one hits harder than a 20% crash in year 20, when you've been drawing down your portfolio. A retiree who faces a severe bear market early, when the portfolio is largest, can drain it faster than someone who faces that same bear market at age 90.
Why Success Rate Matters More Than Average Returns
A portfolio might have an "average" 7% return, but the sequence matters enormously. Imagine two return sequences:
Sequence A (Lucky): +8%, +8%, +8%, +8%, +8%
Sequence B (Unlucky): −20%, −15%, +25%, +25%, +25%
Both have roughly the same average return (~8%), but Sequence B is catastrophic for a retiree withdrawing 5% per year in the early years. The early losses compound with withdrawals, so the portfolio never recovers. Sequence A, with steady moderate gains, lets compounding work in your favor.
Monte Carlo reveals how often your plan encounters Sequence B scenarios—and whether those occasional catastrophes blow up your retirement.
Key Inputs and Their Sensitivity
A Monte Carlo model requires several assumptions:
Expected returns – Usually derived from historical data (stocks: 10%, bonds: 5%) or forward-looking estimates. Even 0.5% changes in expected returns can shift success rates 5–10%.
Volatility – Historical standard deviation of returns. Higher volatility means wider swings, more extreme scenarios, lower success rates. U.S. stocks have had roughly 18–20% annual volatility over the past century.
Inflation – The speed at which your withdrawals grow. Assuming 3% inflation instead of 2% meaningfully lowers success rates because it requires larger later-in-life withdrawals.
Time horizon – Planning to age 95 versus 100 changes the math significantly. Women have longer life expectancies than men; for couples, plan to the longevity of the longer-lived spouse.
Withdrawal strategy – Fixed dollar amount, percentage of portfolio, or dynamic (spend more when portfolio is up). Dynamic strategies often show higher success rates because they adapt to market conditions.
Spending shocks – Health costs, long-term care, or unexpected large expenses can model as occasional portfolio hits. Many retirees don't account for these.
Sensitivity analysis—running the model with slightly different assumptions—reveals which inputs matter most. Often, success rates are more sensitive to spending assumptions than return assumptions.
The Danger of False Precision
A model that outputs "94.3% success" feels scientific and precise. It's not. That 94.3% is conditional on every assumption being correct. If actual returns are 1% lower than assumed, inflation is higher, or the retiree spends more than planned, the true success rate drops.
The responsibility of the person running the simulation is to test a range of assumptions (a 5% return scenario, a 6% scenario, a 7% scenario) and report what success rate holds across that range. If your plan has a 90% success rate when returns are 7%, but a 60% success rate when returns are 5%, that's a brittle plan. You're betting on above-average market performance.
Professional advisors often stress-test plans against "safe" return assumptions (5% nominal for stocks) rather than historical averages (10%) to avoid giving false confidence.
The Clustering Problem: Bad Years in a Row
Monte Carlo models sometimes fail to capture the real-world phenomenon of clustered returns—the 1970s were ugly, the 2000s saw two massive crashes, 2008–2009 felt like the world was ending. Historical data shows that returns aren't truly random; they can cluster.
A "pure" Monte Carlo engine draws returns independently, meaning each year's return is unrelated to the previous year's. Real markets have momentum, mean reversion, and regime changes that create correlation across years. A more sophisticated model (using "block bootstrap" or regime-switching parameters) can account for this, but most consumer-facing tools don't.
The practical implication: a standard Monte Carlo might underestimate the odds of facing three consecutive 15% losses. Conservative planners run their models with slightly lower expected returns or higher withdrawal amounts to compensate.
What Happens When Plans Fail
In a Monte Carlo scenario where success rate is 90%, 10% of the simulations show the portfolio running out of money before the target age. The question is: what does "running out" mean?
In some models, it's immediate ruin—you hit zero at age 82 and then you have no income. In other models, it's gradual; you're forced to cut spending. Some retirees can cut 30% of spending in bad years and continue; others can't.
A realistic model should include a "floor" outcome—if the portfolio depletes, Social Security or other guaranteed income covers basics, but discretionary spending cuts dramatically. A 90% success rate might actually mean "portfolio never depletes" (very strict) or "portfolio with Social Security can support at least 80% of planned spending" (more nuanced).
External Validation and Authority
The Trinity Study (Cooley, Hubbard, Walz, 1998, updated through the 2000s) established that a 4% initial withdrawal rate with annual inflation adjustments had a 95% success rate across rolling 30-year periods in U.S. history. That work was largely grounded in historical back-testing, not Monte Carlo, but it provided a benchmark.
More recent work by financial planners and researchers has incorporated Monte Carlo methods with higher rigor. The U.S. Securities and Exchange Commission's Office of Investor Education and Advocacy provides guidance on retirement projections here: <https://www.investor.gov/introduction-investing/investing-basics/getting-started>. The Federal Reserve and academic economists have published research on sustainable withdrawal rates; you can find current data on inflation and long-term yields at the Federal Reserve Economic Data (FRED) database: <https://fred.stlouisfed.org/>.
The Social Security Administration's life expectancy calculator (<https://www.ssa.gov/benefits/retirement/>) allows you to refine assumptions about longevity, a key input.
Real-World Examples
Example 1: The $1 Million Retiree
Sarah has $1 million in a 60/40 stock/bond portfolio, plans to live on $50,000 per year (adjusted 2.5% annually), and wants to know if she'll have money at 95.
She runs a Monte Carlo with:
- 8% stock returns, 4% bond returns (blended: ~6.4% nominal)
- 18% stock volatility, 4% bond volatility
- $50,000 year one, growing with inflation
- 30-year horizon (age 65 to 95)
Result: 92% success rate. She's comfortable—that's in the "sustainable" zone. However, the advisor runs a stress test with 5% blended returns (more conservative) and gets 78% success. That warns her: if returns come in below historical average, she's at risk.
Example 2: The Late-Start Retiree
David is 72, has $400,000, and wants $30,000/year. He also has $20,000/year in Social Security, so total needed is $10,000 from portfolio.
Monte Carlo shows 96% success rate through age 95. But David recalculates for longevity to 100 (he's fit, long family history), and success drops to 82%. He adjusts: he'll delay Social Security to 75 (boosting it 32%), and that raises success back to 91%. The model shows which levers matter.
Example 3: The Early Retiree with Sequence Risk
Marcus is 50, has $500,000, plans to spend $25,000/year (adjusted for inflation), and won't touch Social Security until 67. His horizon is 47 years (to age 97)—long enough that sequence risk bites hard.
Standard Monte Carlo: 73% success rate. Too low. But Marcus adjusts his plan: he'll hold three years of expenses in cash/bonds (reducing the portfolio available for long-term growth, but providing a buffer), and he'll implement a dynamic strategy (cut spending 10% in years when portfolio return is negative). Those changes lift success to 87%.
The lesson: younger retirees with longer horizons need more conservative assumptions or adaptive strategies.
Common Mistakes
Mistake 1: Confusing Success Rate with Probability of Recession
A 90% success rate does not mean "there's only a 10% chance of a recession." Recessions happen regularly. Success rate means "of the 10,000 recession scenarios (and non-recession scenarios) we modeled, your plan survives 90% of them." Recessions are already built in.
Mistake 2: Ignoring Spending Flexibility
Many Monte Carlo tools assume spending is fixed or grows mechanically with inflation. Real retirees cut spending in bad markets. A model that allows 5–10% spending flexibility in down markets will show higher success rates, more accurately reflecting reality.
Mistake 3: Using Only One Set of Assumptions
Running one scenario and trusting the output is risky. The credible approach is stress-testing: how does success rate change if returns are 1% lower? If inflation is higher? If you live to 100?
Mistake 4: Overlooking Inflation Assumptions
Many planners use low inflation assumptions (1.5–2%) because that's been recent experience. But long-term inflation has averaged 3%+. Using overly optimistic inflation will inflate (pun intended) success rates.
Mistake 5: Not Accounting for Concentrated Positions or Significant Expenses
Monte Carlo assumes a diversified, rebalanced portfolio. If you have a concentrated position (heavily weighted to company stock or real estate), or if you're planning for major expenses (home renovation, grandchild's education), those need to be modeled separately.
FAQ
Q: Is 90% success rate "good"?
A: Yes. Financial planners typically aim for 85–95%. Below 85%, the plan is fragile. Above 95%, you're probably over-saving. The choice depends on your risk tolerance and flexibility (can you cut spending if needed?).
Q: Why do different tools give different success rates for the same plan?
A: Different return assumptions, volatility estimates, inflation inputs, and algorithm details. Two tools might use different Monte Carlo engines or different historical data ranges. Always ask the tool designer about their assumptions.
Q: Can Monte Carlo predict the next market crash?
A: No. It doesn't predict timing or magnitude of specific events. It models the statistical likelihood of various scenarios occurring, based on history. If a crash happens tomorrow, your actual success rate might differ from the model.
Q: Should I use more stocks if I'm young?
A: In theory, yes—longer time horizon means more ability to recover from crashes. In practice, Monte Carlo success rates for very young retirees (30s, 40s) planning 50+ year horizons are often lower because sequence risk is acute early on, and spending grows a long time. Dynamic spending (cutting when portfolio is down) or phased retirement might make more sense.
Q: How often should I re-run my retirement Monte Carlo?
A: At least annually, or whenever major life changes occur (inheritance, major expense, job loss, health change, spending changes). Market conditions and your circumstances evolve; the model should too.
Q: What if Monte Carlo says I have a 100% success rate?
A: You're either very wealthy relative to spending, or the assumptions are too optimistic. Test with lower returns (5–6%) or higher inflation (3–4%) to find the real margin of safety.
Related Concepts
- Sequence of returns risk – Why the order of returns matters more than the average
- Withdrawal rate strategy – Fixed vs. dynamic spending in retirement
- Asset allocation in retirement – How to structure 60/40 or 80/20 when you're drawing down
- Longevity risk – The statistical reality of living longer than expected
- Inflation in retirement – Why purchasing power erosion matters more in long retirements
Summary
Monte Carlo retirement planning replaces false certainty with realistic stress-testing. By running your spending plan through thousands of plausible market scenarios, you see not whether you'll succeed in the "average" market, but what fraction of real-world market paths leave you solvent.
A 90–95% success rate is the standard target for sustainable retirement plans. But the output is only as good as the assumptions: returns, volatility, inflation, spending flexibility, longevity, and major expenses must all be defensible. Stress-testing across a range of assumptions reveals whether your plan is robust or brittle.
The real value of Monte Carlo retirement analysis isn't the precise percentage—it's the conversation it forces between you and your planner about what could go wrong, what you'd do in that scenario, and whether your plan has enough margin for error to weather the inevitable surprises of a 30-year or 50-year retirement.