Fat Tails and Retirement: How Often Do Big Crashes Really Happen?

Many retirement calculators assume market returns follow a normal distribution - the familiar bell curve. It is a reasonable starting point. But it gets one thing wrong that matters a lot in retirement: how often, and how badly, markets crash.

How Often Big Crashes Really Happen

Take a bell curve with an average return of 8% a year and a volatility of 16% - roughly what many planners assume for stocks. It says a loss of 30% or more in a single year has a chance of about 0.9%: once every 110 years or so.

Real history is harsher. From 1928 to 2025, the S&P 500 lost 30% or more in three calendar years: 1931, 1937 and 2008. Try it with the actual data:

The pattern holds at every level: severe losses show up about twice as often as a bell curve with the same average and volatility expects. That extra weight in the extremes is what "fat tails" means.

It's Mostly About the Downside

Looking at yearly returns, the U.S. record is close to a bell curve in the middle. The difference is lopsided: big losses are larger and more frequent than equally big gains. Statisticians call this negative skew. For the S&P 500's yearly returns since 1928, the skew is about -0.5 - markets tend to fall faster and further than they rise.

Monthly and daily returns show much fatter tails than yearly ones. A yearly simulation smooths some of that out, but the crashes that matter for a retirement plan - the ones that take a big bite out of your savings in a single year - still come more often than the bell curve suggests.

Why This Matters More in Retirement

During your working years, a crash is painful but recoverable. You keep contributing, you buy at lower prices, and you have decades for the savings to recover.

In retirement, a crash combines with withdrawals. You sell investments at low prices to pay the bills, and those investments are not there for the recovery. A crash hurts most in the first years of retirement - that is sequence-of-returns risk.

How Simulators Model Fat Tails

Instead of the normal distribution, a simulator can draw returns from a Student's t-distribution, which puts more weight in the extremes. Its "degrees of freedom" set how heavy the tails are: lower means heavier. Adding skew makes sharp falls more likely than equally sharp rises.

Retirement Lab's Pro plan offers this as an option: a Student's t-distribution with two settings, moderate (5 degrees of freedom) and extreme (3), plus an optional skew. It keeps your average return and volatility exactly as you set them and changes only the shape: more very bad and very good years, fewer in between.

The Surprising Part: Less Than You'd Think

Here is what that does to a typical plan - a 65-year-old spending 4% of 1,000,000 a year from a 60/40 mix, raised with inflation, over 30 years (Retirement Lab's engine, Balanced outlook, 50,000 simulated futures):

With the same average and volatility, fat tails move a 30-year success rate by about a point. Two things cancel out: there are more extreme bad years, but also more extreme good years and more calm ones. Over 30 years, those mostly even out. The downside skew - crashes bigger than rallies - is what nudges the result down.

What moves the result far more is the average and the volatility you assume. The same plan scores 59% with a cautious market outlook and 90% with U.S. historical averages. A crash right after you retire moves it a lot too.

So Why Model Fat Tails at All?

To see how deep the bad years go. Fat tails change what a single terrible year looks like, even if they barely change the long-run odds. If you want to know how your plan feels in a 1931 or a 2008, that matters.

To check you're not relying on a bell curve. If your plan holds up with fat tails and skew switched on, it isn't quietly depending on crashes staying small.

To keep perspective. The bigger risks for a retirement plan are assuming returns that are too high, spending too much, and a bad start. Fat tails are a useful second look, not the main event.

The idea is not new - Benoit Mandelbrot wrote about fat tails in market prices in the 1960s. What is new is that you can check your own plan against it in a couple of minutes.