Most retirement calculators show you a single number: 'At your current savings rate, you will have $X at age 65.' That single number is almost certainly wrong. Not because the math is bad, but because it assumes a straight-line return every year — and the market does not deliver straight-line returns.
Monte Carlo simulation solves this by replacing the single projection with thousands of possible futures, each with a different sequence of market returns. Instead of one answer, you get a probability distribution — and that distribution tells you far more about your retirement readiness than any single number ever could.
The Problem with Deterministic Projections
Sequence-of-returns risk is one of the most important concepts in retirement planning. Two investors with identical average returns over 30 years can have dramatically different outcomes if the returns arrive in different order. Bad returns early in retirement, when your portfolio is large and withdrawals are ongoing, do permanent damage. Good early returns give the portfolio a compounding base that bad years later cannot erase.
Consider two retirees, both starting with $1 million and withdrawing $50,000 per year. Both experience an average annual return of 7% over 25 years. Retiree A gets the bad years first (2000–2002 style losses followed by recovery), while Retiree B gets the good years first. Despite identical average returns, Retiree A runs out of money at year 22. Retiree B finishes with over $2 million. The sequence is everything.
What Monte Carlo Actually Does
A Monte Carlo simulation runs thousands of randomized scenarios — each representing a different possible sequence of market returns — and reports what percentage of those scenarios result in the portfolio surviving through your target retirement age. Each scenario draws returns from a statistical distribution calibrated to historical or forward-looking data for each asset class.
The output is a success rate: if 9,200 out of 10,000 scenarios resulted in the portfolio lasting through your plan, the success rate is 92%. You also get a distribution of ending balances, showing the range from worst-case to best-case outcomes — not just the average.
A 90% success rate does not mean your retirement is 90% certain to succeed. It means that 90 out of 100 simulated scenarios succeeded under the specific model assumptions.
How Many Simulations Do You Need?
Statistical convergence matters. With 100 simulations, the success rate can vary by 5–8 percentage points depending on the random seed. At 1,000 simulations, the variance narrows to about 1–2 points. At 10,000 simulations, results are highly stable — running the same plan twice produces nearly identical results.
For most planning purposes, 500 to 1,000 simulations provide a reliable estimate. However, if you are making high-stakes decisions near the boundary (a plan with an 80% success rate where the margin matters), 10,000 simulations give you confidence that the result is not a statistical artifact.
Why Asset Correlations Matter
A naive simulation might generate stock returns and bond returns independently. In reality, asset classes are correlated — stocks and bonds sometimes move together, and sometimes diverge. A proper simulation preserves these correlations using a mathematical technique called Cholesky decomposition, which generates correlated random draws from a covariance matrix.
This matters because diversification benefits depend on correlation. If your simulation ignores the fact that international stocks and U.S. stocks are positively correlated, it will overstate the benefit of international diversification. Accurate correlations lead to more realistic fan charts and more honest success rates.
How RetireWise Runs Its Simulations
RetireWise uses J.P. Morgan's Long-Term Capital Market Assumptions (2026 edition) for expected returns, volatilities, and correlations across nine asset classes. Each simulation draws correlated annual returns using Cholesky decomposition to preserve the realistic co-movement structure of global markets. The free tier runs 500 scenarios; Premium runs 10,000.
Beyond raw market returns, each simulation also models inflation (drawn from a distribution, not fixed), tax brackets (updated annually within the simulation), Social Security COLA adjustments, and required minimum distributions. This means the simulation captures the full complexity of a real retirement — not just investment returns in isolation.
What the Success Rate Tells You — and What It Doesn't
A high success rate (90%+) is reassuring but not a guarantee. The simulations are bounded by the return distributions embedded in the model. Truly unprecedented events — a sustained 15-year bear market, a 1940s-style bond wipeout, or a global depression — may fall outside what the model can represent. Use the success rate as a planning tool, not a promise.
Conversely, a 100% success rate does not mean you should celebrate — it usually means you are being too conservative. A plan that never fails under any simulated scenario is likely leaving significant money on the table. Most financial planners consider 85% to 95% to be the healthy range: high enough to feel confident, low enough to indicate you are actually using your money.
Inflation: The Hidden Variable
Most people focus on investment returns, but inflation is equally important in a 30-year retirement plan. A fixed 3% inflation assumption can dramatically understate the range of real outcomes. In the 1970s, inflation averaged over 7% annually for a decade. In the 2010s, it averaged under 2%. A simulation that draws inflation from a realistic distribution — rather than fixing it — captures this uncertainty and produces more honest results.
Inflation affects every part of a retirement plan: the real value of your withdrawals, the purchasing power of fixed income sources, the growth of tax brackets (which are indexed to CPI), and the COLA on Social Security benefits. A simulation that varies inflation alongside market returns gives you a much more complete picture of the risks you face.
Withdrawal Strategies and Monte Carlo
The classic 4% rule — withdraw 4% of your initial portfolio in year one and adjust for inflation each year — was derived from historical backtesting. Monte Carlo simulation can test this rule under a wider range of scenarios and also evaluate more dynamic strategies: reducing withdrawals in down years, taking a percentage of the current portfolio each year, or using guardrail approaches that adjust spending based on portfolio performance.
Flexible withdrawal strategies consistently show higher success rates in Monte Carlo analysis compared to fixed strategies. Even a modest willingness to cut spending by 10% when the portfolio drops below a threshold can improve success rates by 10 to 15 percentage points. This flexibility is one of the most powerful levers available to retirees.
Common Misconceptions
"Monte Carlo is just random guessing"
The randomness is not arbitrary — it is calibrated to real-world data. Each scenario draws from probability distributions that reflect historical returns, volatilities, and correlations. The randomness captures the genuine uncertainty of market outcomes, not noise.
"A higher success rate is always better"
Not necessarily. A 100% success rate at age 90 might mean you could have retired five years earlier, spent more, or given more to family and charity. The goal is not to maximize the success rate — it is to find the plan that balances security with quality of life.
"I only need to run the simulation once"
Your plan should be a living document. As your portfolio value changes, tax laws shift, and your health and goals evolve, the inputs to the simulation change — and so do the results. Running the simulation annually (or after any major life event) keeps your plan grounded in current reality.
Practical Takeaways
- Target a success rate of 85–95% for most plans; 100% often means you are over-saving.
- Check how the success rate changes if you retire 1–2 years earlier or later.
- Model a spending flex — if you can cut spending 10% in bad years, your success rate rises substantially.
- Run the simulation with a higher equity allocation and a lower one to see how sensitive your plan is.
- Revisit the simulation annually as market conditions and your portfolio value change.
Frequently Asked Questions
What is a good Monte Carlo success rate for retirement?
Most financial planners consider 85% to 95% to be a healthy range. Below 80% suggests meaningful risk that your plan may fall short. Above 95% suggests you may be under-spending or over-saving relative to your goals. The right target depends on your risk tolerance and flexibility to adjust spending.
Is Monte Carlo simulation better than historical backtesting?
They serve different purposes. Historical backtesting ("what if I retired in 1966?") shows how your plan would have fared in real past scenarios, including the worst ones. Monte Carlo simulation generates a wider range of possible futures, including scenarios that have not occurred historically. Monte Carlo is generally more useful for forward-looking planning because it is not limited to the specific sequences history happened to produce.
Does Monte Carlo account for inflation?
It depends on the tool. Some calculators use a fixed inflation assumption (e.g., 3% per year). RetireWise draws inflation from a distribution in each simulation, so inflation itself varies across scenarios — just as it does in the real world. This produces more realistic results than a fixed assumption.
Can Monte Carlo predict a market crash?
No. Monte Carlo does not forecast specific events. Instead, it generates scenarios that include crash-like outcomes as part of the probability distribution. Some simulated sequences will include steep early losses, extended bear markets, or high-inflation periods — not because the model "predicts" them, but because those outcomes are statistically plausible given historical volatility.
Source: J.P. Morgan — 2026 Long-Term Capital Market Assumptions