Why Your Retirement Plan Needs More Than One Future: Understanding Monte Carlo Stress Testing

The Uncomfortable Truth About Retirement Projections

Most retirement calculators ask you a deceptively simple question: What rate of return do you expect? You type in 7%, or maybe 8%, and the calculator draws a single line from today to your death — a clean, uninterrupted arc of wealth. It feels reassuring.

Markets don’t deliver 7% every year. They deliver +32% some years and -38% in others. They go sideways for a decade. They crash 50% right before you retire and recover five years after. The average return might be 7%, but you don’t live the average — you live the sequence.

This is the problem Monte Carlo simulation was built to solve. Instead of one future, it runs thousands. And not all Monte Carlo approaches are created equal.

The Baseline: Geometric Brownian Motion (GBM)

Before we get to the interesting stuff, it helps to understand where most retirement tools start: Geometric Brownian Motion, or GBM.

GBM is the mathematical backbone of most financial models going back to the Black-Scholes options pricing formula in the 1970s. In retirement planning, it works like this: your portfolio grows each year according to a fixed average return (say, 8%) plus or minus a random shock drawn from a bell curve with a fixed standard deviation (say, 15%). Every year is independent. Every year draws from the same distribution.

It’s elegant. It’s mathematically tractable. And it misses something fundamental about how markets actually behave.

What GBM gets right:

• Captures year-to-year randomness
• Easy to reason about (returns are log-normally distributed)
• Provides a useful baseline for comparison

What GBM gets wrong:

• Markets cluster. Bad years tend to follow bad years. 2008 wasn’t a fluke that reset — it was the beginning of a multi-year crisis.
• Volatility isn’t constant. The VIX during the 2008 crash hit 80. During normal markets it sits around 15.
• Returns aren’t bell-shaped — real return distributions have fat tails, meaning catastrophic losses happen far more often than a normal distribution predicts.
• There’s no concept of regime — no “the market is currently in a bull run” vs. “we’re in a grinding bear market.”

Enter Regime Switching: How Markets Actually Work

The alternative is a class of approaches called Markov Regime Switching. The intuition is simple: markets operate in distinct states — regimes — and they tend to stay in those states for extended periods before transitioning to another.

Think about it from lived experience:

• Bull markets don’t last a year — they last three to five, driven by expanding earnings, low rates, or new technology cycles.
• Bear markets don’t resolve in a quarter — they grind for 12 to 18 months, with volatility spiking and correlations breaking down.
• Stagnation periods — the 1970s, or Japan’s “lost decade” — can last years with near-zero real returns.

Regime switching captures this by defining distinct market states, each with its own expected return and volatility, and a transition probability governing how likely the market is to stay in that state or shift to another.

The Seven Models, Explained

Fatboy Financial Planner includes seven distinct Monte Carlo scenarios, each designed to answer a different question about your retirement’s resilience.

1. GBM (Baseline)

Expected avg return: ~8-9% | Volatility: ~15%

The starting point. Single regime, constant drift, no memory. Useful as a benchmark — if your plan fails GBM, it almost certainly fails everything. If it only passes GBM, you haven’t learned much.

2. Regime Default

Expected avg return: ~9.6% | 4 regimes

Calibrated to the US market’s long-run history from 1926–2024. Bears happen about 15% of the time and last just over a year. This is the “realistic historical average” scenario — it captures regime clustering without being pessimistic about the future.

3. Regime Enhanced

Expected avg return: ~9.6% | Same regimes + two advanced features

This adds two important behaviors on top of the default:

Valuation adjustment: When markets are expensive (high P/E ratios), the probability of a bear regime increases. When markets are cheap, bull probability increases. This reflects the well-documented relationship between starting valuations and subsequent 10-year returns.

Correlation dynamics: In normal markets, stocks and bonds tend to move opposite each other. This is the foundation of diversification. But in genuine bear markets, this relationship breaks down — the stock/bond correlation flips from -0.2 to +0.5. Your diversified portfolio isn’t as diversified as you thought when it matters most. This scenario captures that.

4. Crash Model

Expected avg return: ~9.2% | 5 regimes including explicit crash + recovery

This adds two new regimes to the default:

• Crash: Occurs about 5% of the time. Returns average -35% with 40% volatility. Average duration: less than a year. These are the 1929s and 2008s — fast, violent, severe.
• Recovery: Follows crashes with a 60% probability. Returns average +25% as markets snap back.

The crash scenario tests how well your plan handles a sudden catastrophic drawdown. If you’re five years from retirement and the market drops 35%, does your plan survive?

5. Regime Conservative (Stress Test)

Expected avg return: ~6.1% | 4 regimes, structurally pessimistic

This is where the conversation gets serious.

The conservative scenario isn’t just a “lower return” version of the default — it’s a structurally different market environment. Bears are more frequent (18% vs. 15%), deeper (-20% vs. -15%), last longer (1.75 years vs. 1.25 years), and even stagnation periods deliver slight losses. Bull markets are muted (+15% vs. +20%) and happen less often (35% vs. 45%).

The net result: an average annual return roughly 3.5 percentage points lower than the default scenario. Over a 25-year retirement, that gap compounds into a dramatically different outcome.

This scenario targets a 65–75% success rate for well-constructed plans. If you’re seeing 90%+ here, your plan has meaningful margin. If you’re in the 40s or 50s, consider it a yellow flag worth addressing.

6. Historical Recent (2000–2025)

Method: Block bootstrap resampling

Instead of generating synthetic returns, this resamples from actual market data: the 25 years from 2000–2025. This period includes:

• The dot-com bust (2000–2002): -45% peak to trough
• The Global Financial Crisis (2007–2009): -57% peak to trough
• The COVID crash (2020): -34% in 33 days, then full recovery in 5 months
• The 2022 rate-shock bear market: simultaneous stock and bond losses
• The post-COVID bull run of 2023–2024

This is arguably the most relevant historical period for people retiring today or in the next decade — it includes the modern low-rate, high-valuation environment and the subsequent normalization.

The block bootstrap method preserves the sequence of returns, not just the distribution. A block of 2007–2009 stays together — you don’t accidentally insert a recovery year in the middle of the financial crisis.

7. Historical Full (1926–2024)

Method: Block bootstrap resampling | 99 years of data

The full century of US market history: the Great Depression, World War II, the postwar boom, stagflation, the dot-com bubble, the financial crisis, and everything in between.

This includes scenarios modern investors haven’t experienced — like a 90% market decline that takes 25 years to fully recover. It also includes the extraordinary 1950s–60s bull market and the post-WWII expansion.

The full historical scenario is most useful as a lower bound — the absolute worst outcomes in modern financial history. If your plan passes here, it’s been tested against everything the US market has delivered in a century.

What the Model Comparison Is Telling You

When you run a Monte Carlo projection and pull up the comparison overlay, you’re seeing your specific portfolio, your specific spending rate, and your specific retirement timeline run through all seven market environments simultaneously.

The results tell different stories:

• If most scenarios stay above zero: Your plan is broadly resilient across market environments.
• If GBM passes but regime scenarios fail: Your plan is sensitive to sequence-of-returns risk — you’re relying on smooth, average growth, not the lumpy reality of actual markets.
• If the default regime passes but conservative fails: Your margin is thin. You’re one bad decade away from a shortfall.
• If even the full historical scenario passes: You’ve been tested against the Great Depression. That’s a meaningful finding.
• If all scenarios fail early: The issue isn’t market assumptions — it’s fundamental. Spending too high, savings too low, or timeline too long.

A Note on Sequence-of-Returns Risk

There’s one phenomenon no average return can capture: sequence-of-returns risk. Two portfolios with identical average returns over 30 years can have dramatically different outcomes depending on when the bad years occur.

If the market drops 40% in year one of your retirement, you sell shares at rock-bottom prices to fund your living expenses. You lock in those losses permanently. When the market recovers, you have far fewer shares to participate in the recovery.

If the same 40% drop happens in year 25 — after you’ve been drawing down for two decades — the impact is far smaller. The remaining balance is smaller, but you’ve had 24 years of withdrawals before the crash hit.

How to Use This in Practice

When reviewing your Monte Carlo results, a useful mental framework:

The optimistic floor — GBM What happens in a smooth, constant-volatility world. If you fail here, the plan has a fundamental problem independent of market assumptions.

The realistic baseline — Regime Default Calibrated to 100 years of US market history. This should be your primary planning scenario.

The stress test — Regime Conservative Your canary in the coal mine. A 65–75% success rate here is healthy for a conservative stress test. Under 50% is a warning sign worth addressing — more savings, lower spending, a part-time income bridge, or a different asset allocation.

The historical reality check — Historical Recent (2000–2025) Grounds your plan in the actual market environment you’re likely navigating. If this is significantly worse than the default regime scenario, your plan may be over-reliant on the strong decades before 2000.

The worst case — Historical Full + Crash Model The outer boundary of bad luck. Passing isn’t required — failing isn’t a death sentence. But the gap between these and your baseline tells you how much margin you have.

The Bottom Line

Retirement planning isn’t about finding the one true projection — it’s about understanding the range of outcomes and making sure you have a plan that survives enough of them.

GBM is a starting point. Regime switching is a more honest picture of how markets work. The conservative stress test separates plans with genuine resilience from plans that depend on everything going reasonably well. The historical scenarios anchor you to reality.

No scenario predicts the future. But a plan that survives seven different simulations of it — ranging from the optimistic to the historically catastrophic — is a plan you can actually retire on.

Questions? Email: fbfinancialplanner@gmail.com

Related reading:

• Your Heirs Have a Tax Problem. Your Retirement Plan Isn’t Showing It.
• Sequence of Returns Risk: Why Your Retirement Date Matters More Than Average Returns
• The Retirement Red Zone: Why the 5 Years Before and After Retirement Are Make or Break