Monte Carlo Retirement Simulation: Why Your Plan Needs a Stress Test

A single average-return projection for your retirement is misleading. It tells you what happens if returns arrive in a perfectly smooth, predictable pattern. They do not.

Monte Carlo simulation runs your retirement plan through thousands of possible market scenarios. Some are great. Some are terrible. The percentage that end with money still in the account is the "success rate." That number is more useful than any average-return spreadsheet — and often significantly lower.

What Monte Carlo Is Doing

A Monte Carlo simulation takes your portfolio balance, withdrawal rate, and asset allocation, then generates thousands of random return sequences using historical volatility and return data.

It is not predicting the future. It is saying: if the future looks like any random arrangement of historical market returns, here is the distribution of outcomes.

1,000 simulations, $1,000,000 portfolio, $40,000/year withdrawal:

| Outcome | Number of Simulations | |---------|----------------------| | Portfolio lasts 30+ years | 870 | | Portfolio depleted before 30 years | 130 | | Success rate | 87% |

An 87% success rate means the plan works in 870 out of 1,000 historical scenarios. It fails in 130. For a retiree with no fallback income, those 130 represent running out of money with no recovery path.

What "Success" and "Failure" Mean

A simulation "fails" when the portfolio reaches $0 before the end of the specified retirement period. Most tools define failure as literal depletion — not declining below a threshold, but hitting zero.

Success does not mean the plan is ideal. Two plans can both have 85% success rates, but one leaves heirs $2,000,000 on average in success scenarios, while the other ends with $50,000. The median ending balance matters as much as the success rate.

High success rates achieved by extreme frugality (spending $25,000/year on a $1.5M portfolio) are not inherently good plans — they suggest the retiree is leaving quality of life on the table to protect against scenarios that may never occur.

How Withdrawal Rate Moves the Needle

The withdrawal rate is the most powerful variable in Monte Carlo outcomes. Based on standard historical US market return distributions (data sourced from the Shiller CAPE database and CRSP returns):

$1,000,000 portfolio, 60/40 stock/bond allocation, 30-year retirement:

| Annual Withdrawal | Withdrawal Rate | Success Rate | |-----------------|----------------|-------------| | $30,000 | 3.0% | ~99% | | $35,000 | 3.5% | ~96% | | $40,000 | 4.0% | ~87% | | $45,000 | 4.5% | ~79% | | $50,000 | 5.0% | ~71% | | $55,000 | 5.5% | ~62% | | $60,000 | 6.0% | ~53% |

Each $5,000 increase in annual spending drops success rate by roughly 8 percentage points. The 4% rule — derived from the 1994 William Bengen paper and confirmed by the 1998 Trinity Study (Cooley, Hubbard, Walz, Journal of Financial Planning) — targets the 87% zone. It is a reasonable middle ground, not a guarantee.

How Asset Allocation Affects Success

Portfolio composition matters, but less than most people expect. An all-stock portfolio does not maximize success rate because early crashes are more destructive when there is nothing to buffer withdrawals.

Same $1,000,000 portfolio, 4% withdrawal rate, 30-year retirement:

| Stock / Bond Allocation | Success Rate | |------------------------|-------------| | 20% / 80% | ~71% | | 40% / 60% | ~81% | | 60% / 40% | ~87% | | 75% / 25% | ~86% | | 100% / 0% | ~83% |

The 60/40 portfolio produces the highest success rate at a 4% withdrawal. This is because bonds buffer sequence-of-returns risk during early retirement downturns — you sell bonds to fund spending rather than equities at depressed prices. Above 75% equities, sequence risk outweighs the higher expected return.

Sequence of Returns Risk: Why Timing Matters More Than Average

The most important concept Monte Carlo captures — and straight-line projections miss entirely — is sequence of returns risk.

A 40% market crash in year 2 of retirement is catastrophically more damaging than the same crash in year 22, even if the 30-year average return ends up identical.

Why: In year 2, you are selling depreciated shares to fund living expenses. Those shares never recover in your portfolio because you sold them at the bottom. In year 22, your portfolio has had 20 years of growth as a buffer and is far larger — the same percentage drop costs fewer shares and leaves more to recover.

Historical worst-case retirement start years:

| Retirement Start Year | Why It Was Brutal | 30-Year Outcome (4% rule) | |----------------------|-------------------|--------------------------| | 1929 | Great Depression, -89% peak to trough | Failed (~year 20) | | 1966 | High inflation, flat stocks for 16 years | Failed (most scenarios) | | 2000 | Dot-com crash, then 2008 within 8 years | Borderline (depends on allocation) | | 1982 | Best case — 18-year bull market immediately | Massive surplus | | 2009 | Started in trough — strong subsequent decade | High surplus |

The 1966 retiree is the canonical example. The stock market was roughly flat in real terms from 1966 to 1982 — 16 years of inflation eroding purchasing power while flat nominal returns meant real portfolio decline every year. A 4% withdrawal in 1966 depleted most portfolios within 25-30 years.

This is why the 4% rule is derived from the worst historical case (Bengen used rolling 30-year periods from 1926 forward), not the average. Monte Carlo simulates scenarios like 1966 in its random draw of return sequences.

Monte Carlo vs. Historical Backtesting: Which Is More Useful?

Both methods are valid and complementary. Neither is complete on its own.

| Method | Strength | Weakness | |--------|----------|----------| | Monte Carlo | Tests infinite scenarios, not limited to history | Can generate scenarios that never occurred (e.g., 10 consecutive crash years) | | Historical backtesting | Uses real market behavior, actual correlations | Limited sample size (fewer than 60 non-overlapping 30-year periods since 1926) | | Combination | Best confidence — scenarios pass both tests | More complex to interpret |

The CalcMoney FIRE Calculator runs Monte Carlo simulation. For historical backtesting, cFIREsim (free, open-source) runs your plan against every historical starting year back to 1871. Running both and comparing results gives higher confidence than either alone.

If your plan shows 87% on Monte Carlo but only 72% in historical backtesting, the lower number is worth paying attention to — history has been through scenarios Monte Carlo underweights.

Why 90% Is Not Good Enough for Everyone

The "right" success rate is not universal. It depends on your fallback options.

Low flexibility — target 95%+: Retiree with no Social Security, no pension, no ability to return to work. A 90% success rate means 1-in-10 odds of depletion with no recovery path. This person should accept a lower withdrawal rate or work longer.

Medium flexibility — target 85-90%: Retiree with modest Social Security ($18,000-$24,000/year) or part-time work availability. The "failure" scenarios can likely be managed by cutting discretionary spending or returning to part-time income.

High flexibility — 75-85% may be acceptable: Retiree with substantial guaranteed income (pension covering core expenses, full Social Security), significant home equity, or adult children. Portfolio failure is managed by drawing on other assets or reducing spending to the guaranteed income floor.

A 100% success rate is not a realistic target for most people — achieving it requires a withdrawal rate so low (2.5-3%) that the retiree is almost certainly leaving significant quality-of-life spending on the table to protect against scenarios that rarely occurred in history.

Guardrails: How to Improve Success Rate Without Reducing Spending

Rather than targeting a fixed success rate with a fixed withdrawal, dynamic spending rules ("guardrails") significantly improve outcomes while allowing higher average spending.

Example guardrail rule (Kitces/Guyton framework):

• Baseline: withdraw $40,000/year (4% of $1,000,000)
• If portfolio drops below $700,000: reduce spending by 10% ($36,000)
• If portfolio grows above $1,400,000: increase spending by 10% ($44,000)
• Never cut below $32,000; never increase above $52,000

This guardrail strategy achieves approximately a 95% success rate at the same average spending level that achieves 87% with a rigid 4% rule. The trade-off: you must be willing to actually cut spending when the rule triggers. Retirees who set guardrails but do not follow them get the worst of both worlds.

Guardrail success rate comparison:

| Strategy | Average Annual Spending | Success Rate | |---------|------------------------|-------------| | Fixed 4% withdrawal | $40,000 | ~87% | | Guardrails (±10% bounds) | ~$40,000 average | ~95% | | Fixed 3.5% withdrawal | $35,000 | ~96% |

Guardrails achieve near-identical success to a 3.5% fixed rate while preserving more average spending. The cost is behavioral — you need the discipline to cut spending in bad years.

How to Read Your Monte Carlo Results

When you run a Monte Carlo simulation, look at these four numbers, not just the headline success rate:

1. Success rate — the headline. What percentage of scenarios end with money remaining.
2. Median ending balance — the middle outcome in success scenarios. If this is $3,000,000 on a $1,000,000 starting portfolio, your plan is very conservative and you may be under-spending.
3. 10th percentile ending balance — the bad-but-not-worst outcome. If this is negative, 10% of scenarios end in depletion significantly early.
4. Years to depletion in failure scenarios — if you fail, when? Year 28 of a 30-year plan is very different from year 18.

A plan with 85% success rate, $400,000 median ending balance, and year-27 average depletion in failure scenarios is very different from one with 85% success, $2,000,000 median balance, and year-22 depletion in failures.

Frequently Asked Questions

How many simulations are enough?

1,000 simulations is sufficient for reliable Monte Carlo results for retirement planning. Results stabilize beyond 1,000 runs — the difference between 1,000 and 10,000 simulations is typically less than 1 percentage point in success rate. Most institutional financial planning software (MoneyGuidePro, eMoney) uses 1,000-5,000 simulations.

Does Monte Carlo account for inflation?

It depends on the tool. Proper Monte Carlo models adjust withdrawals upward each year for inflation (typically 2-3% annually based on historical CPI from the Bureau of Labor Statistics) and use real (inflation-adjusted) return assumptions. Verify your tool's approach — models that do not inflate withdrawals overstate success rates by approximately 5-10 percentage points.

Is there a better method than Monte Carlo?

Historical backtesting (running your plan against every actual historical starting year) is complementary, not superior. Monte Carlo can generate scenarios that never occurred in history. Historical testing is limited by the number of available historical periods (fewer than 60 non-overlapping 30-year windows since 1926). Using both provides more confidence than either alone.

What does it mean if my success rate is below 80%?

It means your current plan has a meaningful probability of running short of money. Options in order of impact: (1) reduce the withdrawal rate by working one more year and adding to the portfolio, (2) add guaranteed income by delaying Social Security, (3) reduce planned spending, (4) shift to a dynamic withdrawal strategy with guardrails, (5) plan to reduce spending if portfolio drops below a trigger threshold. A sub-80% plan is not a crisis — it is a signal to model alternatives.

You Might Also Like

• FIRE Calculator — model your FIRE number with built-in Monte Carlo
• 72t SEPP Calculator: Access Your IRA Before 59½ Without Penalties
• Barista FIRE Calculator: The Math Behind Semi-Retirement
• Inflation Impact on Retirement Calculator: What Your Savings Are Actually Worth
• Safe Withdrawal Rate by Age: How the 4% Rule Changes Over Time