Monte Carlo simulation for retirement, explained in 600 words

Monte Carlo sounds intimidating. The actual idea is simple enough to explain in the time it takes to make coffee.

The bet you're actually making

When you save for retirement, you're making a bet. You're betting that, over the next 30 to 50 years, your savings plus your future contributions plus market returns minus your spending will leave you solvent.

The problem: you don't know the future. Returns will vary. Inflation will vary. You might live to 85 or 105. A 2008-style crash might happen the year after you retire, or it might not happen at all.

So how do you decide if your bet is reasonable?

Run the bet 10,000 times

Monte Carlo simulation says: instead of guessing the future, replay the past, ten thousand different ways. Take real historical monthly returns and real historical inflation from 1928 onward. Shuffle them. Lay one shuffle against your retirement. Then do another shuffle. Then another, ten thousand times.

Each replay produces a different outcome. Some replays end with you having $5M at age 95. Some replays end with you running out of money at 78. The collection of all ten thousand outcomes is the distribution of your plan.

What you actually get

After ten thousand simulated retirements, you can ask precise questions:

• In how many of them did I have money at age 95? (success probability)
• What's my median ending balance? (50th percentile)
• What's my worst-1-in-10 ending balance? (10th percentile — the "rough seas" case)
• At what age does the 10th-percentile path run out? (depletion age)

These are questions a single-number calculator can't answer. They're also the questions you'd actually want answered before you stop working.

"But what about the next ten years?"

A common objection: history won't repeat exactly. The next thirty years probably won't include another 1929-style crash, or another 1981 inflation spike. Maybe they will. The honest answer is that nobody knows.

What Monte Carlo gives you isn't a forecast. It's a stress test. It says: if the next thirty years look anything like the last hundred, here's the range of outcomes your plan can absorb. If the only paths in which you fail involve ten consecutive 1929-magnitude crashes, your plan is robust. If you fail in the average case, your plan needs work — independent of what the next ten years actually deliver.

When Monte Carlo can mislead

Two honest caveats:

1. The future might be worse than the past. All US historical samples include the greatest economic expansion in human history. Sampling from them assumes the next century will, on average, also be expansionary.

2. Behavior matters. The simulation assumes you stay invested through downturns and follow your withdrawal plan. Real retirees often sell at the bottom or over-spend in booms. No model captures that.

If you're new to Monte Carlo, the best way to build intuition is to try it and watch the percentile band move as you change your inputs. Increase your contribution by $200/month and you'll see the bottom of the band rise visibly. That's the math of patience.