Why the average is a bad plan

Say the long-run average stock market return is 7 percent per year. Fine. Build a retirement plan around it. Withdraw 4 percent a year, and you can easily calculate that the plan lasts until you are 100.

There is only one problem. No investor has ever earned exactly 7 percent every year. Averages hide an enormous amount of variation, and variation is precisely what makes a retirement plan stand or fall.

Why averages hide half of possible worlds

The S&P 500 has returned about 10 percent nominal per year over the last century, or roughly 7 percent after inflation. But individual years have ranged from around minus 40 percent to over plus 50 percent.

If you use an average in your retirement plan, you are quietly assuming the market delivers the same result every year. It never does. And that is not a cosmetic issue, it changes the answer.

Order matters

Two retirees earn the exact same average return over 30 years. One of them sees the bad years first, the other sees them last. If both withdraw the same amount each year, they end with very different final balances.

The reason is that when you sell stocks to fund your life in a bad market, you sell more shares for the same amount of cash. Fewer shares left means less to work with when the market recovers. It is simple, but brutal, arithmetic.

The phenomenon is called sequence-of-returns risk, and it is the main reason "average plans" fall apart even when the average itself is right.

Same average, different order. One retiree has money left, another has run out. That is not luck. That is math.

What Monte Carlo does with the problem

Monte Carlo simulation sounds technical, but the idea is simple. Instead of assuming one average, you simulate many possible paths. Usually 500 or 1000. Each path draws random returns from a distribution that matches historical returns and volatility.

The output is not a single answer. It is a distribution. In how many of the 500 paths did your plan last? In how many did you run out? What was your average ending balance, and what was the worst tenth percentile?

This is not prediction. It is a way to test how sensitive a plan is, without assuming one particular future.

What a robust result looks like

A plan that survives 95 percent of simulated paths is what most planners call robust. A plan that survives 80 percent is acceptable but fragile. A plan that survives 60 percent is a gamble.

What percentage you should target depends on how much uncertainty you can live with, and how much flexibility you have to adjust. A retiree who can cut spending 10 percent in bad years needs a smaller reserve than one whose costs are fixed.

A concrete worked example

Take two retirees. Both start with $1 million. Both withdraw $50,000 a year. Both see the same sequence of returns: 15 bad years followed by 15 good years, averaging 7 percent overall.

Retiree A gets the 15 bad years first. After 15 years, about $300,000 is left, and the money runs out around year 22, well before the plan expected.

Retiree B gets the 15 good years first. After 15 years, about $1.75 million is left. Even with the bad years that follow, the plan comfortably lasts 30 years and beyond.

Same starting balance, same spending, same average return. Radically different outcomes. That is why Monte Carlo is a better test than any single average.

In reality, returns do not arrive in a neat 15-and-15 sequence. They arrive in messy, unpredictable clusters. But the example captures the underlying mechanic: the combination of withdrawals and negative returns is most destructive in the years immediately around retirement.

Simple ways to make your plan more robust

• Test the plan against multiple return scenarios, not just the average.
• Hold two to three years of spending in safe assets.
• Choose a withdrawal rate on the cautious end, 3.5 to 4 percent.
• Decide now how you would cut spending if the market falls 30 percent in year one.
• Recalculate each year. A plan is an ongoing conversation, not a final contract.

The average is a fine start to a conversation. It is a bad plan. The difference is the whole difference.