Monte Carlo Retirement Simulation: A Complete Guide

What Is Monte Carlo Simulation?

Monte Carlo simulation is a computational technique that uses repeated random sampling to model the probability of different outcomes in a system influenced by uncertainty. The name comes from the Casino de Monte-Carlo in Monaco, a nod to the role that randomness and chance play in the method. Physicists Stanislaw Ulam and John von Neumann formalized the technique in the 1940s while working on nuclear weapons research at Los Alamos, but the core idea is simple: if you cannot predict the future with certainty, simulate it thousands of times and study the distribution of results.

In the context of retirement planning, Monte Carlo simulation replaces a single assumed rate of return with thousands of randomly generated return sequences. This distinction is critical. A traditional retirement calculator might tell you that earning 7% annually on a $1.5 million portfolio will sustain $60,000 per year of spending for 30 years. That statement is technically correct for a world where you earn exactly 7% every single year. But that world does not exist.

Real markets are volatile. You might earn 22% one year and lose 15% the next. The order in which those returns arrive matters enormously, a phenomenon called sequence of returns risk. Two retirees with identical average returns over 30 years can end up with wildly different outcomes depending on whether the bad years came early or late. Monte Carlo simulation captures this reality by testing your plan against thousands of possible return sequences, not just one idealized average.

Why It Matters for Retirement Planning

During the accumulation phase of your financial life, when you are saving and investing toward financial independence, volatility is relatively forgiving. Dollar-cost averaging means that market dips let you buy more shares at lower prices. A downturn in year three of a 20-year accumulation period is an inconvenience, not a catastrophe. Time and continued contributions smooth out the bumps.

The decumulation phase is a different animal entirely. Once you stop contributing and start withdrawing, the math flips. A portfolio that drops 30% in year one of retirement does not just lose 30% of its value. It loses 30% of its value and you pull out a year of living expenses from the diminished balance. The portfolio now has to recover from a deeper hole with a smaller base, and every subsequent withdrawal digs that hole deeper. This is sequence-of-returns risk in action.

Consider two scenarios. In Scenario A, a retiree experiences strong returns in years one through five, followed by poor returns later. In Scenario B, the same returns arrive in reverse order, with the poor years coming first. Both scenarios produce the same average annual return. But Scenario B, where the bad years hit early, can deplete the portfolio a decade sooner. A simple average-return calculator treats these scenarios as identical. Monte Carlo simulation does not.

This is why Monte Carlo analysis is the standard tool for rigorous retirement planning. It forces you to confront the range of possible outcomes, not just the expected one. It answers the question that actually matters: across the full range of plausible market histories, how often does your plan survive?

How Monte Carlo Simulation Works

A Monte Carlo retirement simulation follows four steps. Understanding each step helps you interpret the results with confidence and skepticism in appropriate measure.

Step 1: Define Your Inputs

Every simulation begins with a set of assumptions about your financial situation and the future. The core inputs are:

Starting portfolio balance. Your total investable assets at the point of retirement or financial independence.
Annual spending. How much you plan to withdraw each year, often adjusted for inflation.
Time horizon. How many years the portfolio must last. For early retirees in the FIRE community, this is often 40 to 60 years, much longer than the traditional 30-year planning window.
Asset allocation. The mix of stocks, bonds, and other assets, which determines the expected return and volatility of the portfolio.

Additional inputs can include Social Security income (and the age it begins), pension payments, tax rates, inflation assumptions, and any planned changes in spending over time.

Step 2: Generate Random Return Sequences

The simulation engine generates thousands of possible future return sequences. Each sequence represents one plausible path that markets could take over your retirement horizon. The returns are sampled from a statistical distribution calibrated to historical market data.

The choice of distribution matters more than many people realize. Anormal (Gaussian) distribution is the simplest option, treating returns as symmetric around a mean. A log-normal distribution prevents impossible negative portfolio values and better reflects the asymmetry of compounding. A fat-tailed distribution, such as Student's t, assigns higher probability to extreme events like market crashes, producing more conservative estimates. Each choice yields different success probabilities for the same inputs, so it is worth understanding which distribution your tool uses.

Step 3: Simulate Year-by-Year Portfolio Values

For each randomly generated return sequence, the simulation walks through your retirement year by year. In each year, it applies the random return to the portfolio, subtracts your spending (adjusted for inflation and your chosen spending policy), adds any income sources, and records the resulting balance. If the balance hits zero before the time horizon ends, that simulation is marked as a failure. If the balance remains positive throughout, it is a success.

Step 4: Aggregate Results

After running thousands of simulations (typically 5,000 to 10,000), the engine aggregates the results. It calculates the percentage of simulations that succeeded (the success probability) and sorts the ending portfolio values into percentiles. These percentiles form the basis of fan charts and other visualizations that communicate the range of outcomes.

Understanding the Results: Fan Charts and Success Probability

Fan Charts

A fan chart displays the spread of possible portfolio trajectories over time. The center line typically shows the median (P50) outcome: the scenario where half of simulations did better and half did worse. Shaded bands expand outward to show other percentiles:

P75 and P25 form the inner band, covering the middle 50% of outcomes. If your plan falls within this range, things are going roughly as expected.
P95 and P5 form the outer band, covering 90% of outcomes. The P95 line shows what happens in an unusually favorable market. The P5 line shows the near-worst case, the scenarios where markets conspire against you.

The P5 line deserves special attention. It represents the portfolio trajectory in the unluckiest 5% of simulations. If the P5 line crosses zero during your planned retirement horizon, it means that roughly 1 in 20 plausible market histories would deplete your portfolio. Whether that risk is acceptable is a personal judgment, but it is a judgment you should make explicitly rather than ignore.

Success Probability

Success probability is the headline metric: the percentage of simulations in which your portfolio lasted the full time horizon. A 95% success rate means that 950 out of 1,000 simulated retirements ended with money remaining.

What does 95% success actually mean? It means that if you could retire 1,000 times into 1,000 parallel universes with different market histories, your plan would fail in about 50 of them. Whether 95% is "safe enough" depends on your flexibility. If you can cut spending, take on part-time work, or adjust your lifestyle when markets falter, then a lower success rate may be perfectly fine. If your spending is truly fixed and non-negotiable, you may want a higher margin.

It is worth noting that 100% success is not the goal. A plan that succeeds in every conceivable market scenario, including the catastrophic ones, almost certainly requires you to save far more or spend far less than necessary. In the median scenario, a 100% success plan often ends with a portfolio several times larger than where it started. That surplus represents years of potential enjoyment that were sacrificed for protection against worst-case scenarios that never materialized. The sweet spot lies in balancing the risk of running out with the risk of needlessly underspending.

Spending Policies: Beyond Constant Withdrawal

The original 4% rule assumes you withdraw a fixed, inflation-adjusted dollar amount every year regardless of market performance. This is the simplest approach, but it is also the most rigid. If markets crash, you keep spending the same amount, accelerating portfolio depletion. If markets boom, you leave potential spending on the table. Modern retirement research has produced several adaptive spending policies that can significantly improve outcomes.

Constant Dollar (The Traditional 4% Rule)

You withdraw a fixed percentage of your initial portfolio value, adjusted annually for inflation. For a $1.5 million portfolio, this means $60,000 per year (in today's dollars) every year for the duration of retirement. The advantage is simplicity and predictable income. The disadvantage is that it ignores market reality entirely. Your spending does not respond to portfolio performance, which means you overspend in bad markets and underspend in good ones.

Variable Percentage Withdrawal (VPW)

Each year, you withdraw a percentage of your current portfolio balance, with the percentage determined by your remaining time horizon and asset allocation. When markets are up, you spend more. When markets are down, you spend less. VPW virtually eliminates the risk of portfolio depletion (you can never withdraw more than the balance allows), but it introduces income volatility. Your spending could drop significantly after a market downturn, which may or may not be tolerable depending on your fixed expenses.

Guardrails (Guyton-Klinger)

The guardrails approach sets upper and lower thresholds around your initial withdrawal rate. If your current withdrawal rate falls below the lower guardrail (because the portfolio has grown), you give yourself a raise. If it exceeds the upper guardrail (because the portfolio has shrunk), you take a pay cut. The size of the raise or cut is predetermined, typically 10% of current spending. This policy offers a compelling middle ground: it adapts to market conditions while keeping spending changes infrequent and bounded.

Floor-Ceiling

You set an absolute minimum spending level (the floor) and an absolute maximum (the ceiling). Within those bounds, spending adjusts based on portfolio performance, often using a variable percentage method. The floor protects your essential lifestyle needs. The ceiling prevents lifestyle inflation from eroding long-term sustainability. This policy is particularly useful if you have a clear distinction between needs and wants in your budget.

CAPE-Based Withdrawal

This policy ties your withdrawal rate to the Cyclically Adjusted Price-to-Earnings ratio (CAPE or Shiller PE) of the stock market. When valuations are high (and expected future returns are lower), you withdraw less. When valuations are low (and expected returns are higher), you withdraw more. The idea is rooted in the empirical relationship between starting valuations and subsequent returns. The trade-off is complexity: you need a reliable CAPE data source, and the relationship between CAPE and future returns, while statistically significant, is noisy over shorter horizons.

A Worked Example

Let us walk through a concrete example to see how Monte Carlo analysis informs real decisions.

Profile: You have a $1.5 million portfolio with a 60/40 stock/bond allocation. You plan to spend $60,000 per year (a 4% withdrawal rate), adjusted for inflation. Your time horizon is 30 years.

You run a Monte Carlo simulation with 10,000 trials using a log-normal return distribution calibrated to historical U.S. market data (roughly 7% real return for stocks, 2% real for bonds, with corresponding volatilities and correlations).

Results with constant dollar spending:

Success probability: 87%. Your plan survives in 8,700 of 10,000 simulations.
Median (P50) ending balance: $1.8 million (in today's dollars). In the typical scenario, you actually end up wealthier than you started.
P5 scenario: portfolio depleted by year 23. In the unluckiest 5% of simulations, you run out of money seven years before your planning horizon ends.
P95 ending balance: $5.2 million. In the luckiest scenarios, your portfolio more than triples.

The headline number, 87% success, might make you uneasy. A 13% failure rate means that roughly 1 in 8 plausible market histories would leave you broke. And the P5 scenario shows the failure is not marginal; you run out a full seven years early.

Now you make one change: switch from constant dollar spending to a guardrails policy. You set a lower guardrail at 3.5% and an upper guardrail at 5.5%, with 10% spending adjustments when guardrails are breached.

Results with guardrails spending:

Success probability: 97%. A ten-percentage-point improvement.
P5 ending balance: $320,000. Instead of depletion at year 23, the worst-case scenarios now end with a positive balance.
Trade-off: In the P5 scenario, your spending dips to approximately $48,000 per year during the worst stretch, a 20% reduction from your target.

This is the power of Monte Carlo analysis. It did not just tell you whether your plan works. It showed you how it fails, revealed that a modest spending flexibility dramatically improves outcomes, and quantified the exact trade-off (a 20% temporary spending cut in exchange for near-elimination of portfolio depletion risk). That is actionable information you can use to make a decision with clear eyes.

Limitations and Honest Caveats

Monte Carlo simulation is a powerful tool, but it is not an oracle. A few important limitations deserve honest acknowledgment.

Returns Are Not Truly Independent

Most Monte Carlo engines assume that each year's return is independent of the previous year. In reality, markets exhibit both mean reversion (extended periods of high returns tend to be followed by lower returns, and vice versa) and momentum (short-term trends tend to persist). These serial correlations are well-documented but difficult to model accurately. Ignoring them means that Monte Carlo simulations may overstate the probability of extremely long winning or losing streaks compared to what historical markets have actually produced.

Calibration Depends on Historical Data

The statistical distributions used to generate random returns are calibrated to historical market data, predominantly U.S. equities. The U.S. stock market in the 20th century was arguably the most successful capital market in human history. Using it as the baseline for future projections implicitly assumes that the future will resemble the past. It might. It might not. International markets, particularly those of countries that experienced wars, hyperinflation, or political upheaval, tell a less optimistic story. Some practitioners address this by using lower expected returns or incorporating international data, but the fundamental issue remains: we are projecting the future from the past.

Garbage In, Garbage Out

The quality of Monte Carlo results depends entirely on the quality of your inputs. If you underestimate your spending by $10,000 per year, the simulation will give you an overly optimistic success probability. If you overestimate your portfolio by including illiquid assets you cannot actually spend, the results will be misleading. The precision of the output (e.g., "93.7% success probability") can create a false sense of accuracy. Treat the results as directional guidance, not as precise predictions. The difference between 93% and 95% is less meaningful than the difference between 75% and 95%.

Behavioral Risk Is Not Modeled

Monte Carlo simulations assume you will execute your plan faithfully through all market conditions. They do not account for the very real possibility that you will panic-sell during a 40% drawdown, chase performance into an overvalued sector, or abandon your asset allocation at the worst possible moment. Behavioral risk is arguably the largest risk in any long-term investment plan, and no simulation can capture it. The best defense is self-awareness: know your temperament, build a plan you can actually stick to, and keep the simulation results in perspective during turbulent markets.

Related Guides

Monte Carlo simulation is one tool in a broader retirement planning toolkit. These guides cover the strategies and concepts it connects to:

Safe Withdrawal Rate explores five withdrawal strategies you can test with Monte Carlo, from guardrails to CAPE-based rules.
Your FI Number defines the portfolio target that Monte Carlo simulation stress-tests.
Lifecycle Asset Allocation explains how to set the glide path inputs that Monte Carlo uses to model your changing stock/bond allocation over time.
Coast FIRE identifies the portfolio threshold where compound growth alone can reach your FI number, a milestone Monte Carlo can validate.
Roth Conversion Ladder covers the tax strategy that determines how you access retirement funds in the early years of your simulation.

Key Takeaways

Monte Carlo simulation tests your retirement plan against thousands of possible market futures, not just one average. This makes it far more realistic than a simple compound-growth calculator, especially for modeling the risk of portfolio depletion during decumulation.
Sequence of returns risk is the central danger in retirement. Two portfolios with identical average returns can produce vastly different outcomes depending on the order of good and bad years. Monte Carlo is the primary tool for quantifying this risk.
Adaptive spending policies significantly improve outcomes. Switching from rigid constant-dollar withdrawals to guardrails, variable percentage, or floor-ceiling strategies can dramatically reduce the probability of portfolio depletion, at the cost of some income variability.
Pay attention to the P5 scenario, not just the success probability. A 95% success rate tells you how often the plan works. The P5 trajectory tells you how badly it fails when it does. Both pieces of information are essential for making informed decisions.
Treat results as directional, not precise. Monte Carlo output is only as good as the assumptions baked into the model. Use it to compare strategies, identify vulnerabilities, and stress-test your plan. Do not mistake a percentage with decimal places for a guaranteed prediction about your financial future.