September Is the Worst Month for Stocks. The Math Says Otherwise.

Every August, the same story runs: September is historically the worst month for the S&P 500, so brace for a pullback. The headline number is negative, and the market is currently stretched—up over 13% so far in 2026, trading at 20x earnings, with inflation ticking higher. The setup feels like a fear story with legs.

The only problem is the number doing all the heavy lifting in that narrative tells you almost nothing about what's actually likely to happen.

The Mean vs. The Median

The S&P 500's average September return since 1928 is negative—roughly minus 1.2%. September is the only calendar month with a negative historical average.

That's the end of the story most people hear. But the median September return—the one you're more likely to experience—is plus 0.1%. Slightly positive. In just over half of Septembers over the past century, stocks have closed the month higher than they opened it.

If September were reliably a bad month for stocks, the median wouldn't be positive. The negative average exists because the mean is sensitive to extreme outliers, and September has absorbed the most of them.

The Numbers Dragging the Average Down

Five catastrophe months are doing most of the work pulling September's average negative:

• September 1931: minus 29.6%. Great Depression banking meltdown. The United Kingdom abandoned the gold standard, forcing the Fed to raise the discount rate and constrict liquidity.
• September 1937: minus 13.8%. The Fed doubled bank reserve requirements, causing a credit crunch.
• September 1974: minus 11.5%. Oil-shock recession, double-digit inflation, Watergate.
• September 2002: minus 10.9%. Dot-com bear market still searching for a bottom.
• September 2022: minus 9.2%. Fed rate hikes pushing equity valuations down in real time.These figures span five decades of crisis

These aren't seasonal patterns. They're structural crises that happened to land in September. None of them had anything to do with the calendar. If you remove these five months, September's average becomes essentially flat.

There's also September 2008—the Lehman Brothers collapse caused a nearly 9% drop. That one makes six.

So the story goes like this: over 95 years, six months of genuine financial disasters pulled the average negative enough that financial media can declare September the worst month for stocks. And you're being advised to take that seriously.

Why September Absorbs the Crashes

September isn't inherently bad. It's just the month that has, by coincidence, hosted more crises than any other. One contributing factor: September closes the third quarter, and it's when institutions return from summer, rebalance portfolios, and face tax-loss harvesting deadlines—most U.S. mutual funds have an October 31 fiscal year-end. This creates thin, volatile conditions. When a real crisis hits, it hits harder in a month where trading volume is lower and selling pressure is already elevated.

September is the month where the market's structural plumbing is most exposed. If a crisis lands elsewhere, it plays out normally. If it lands in September, it accelerates.

The Intra-Month Split

Here's the part that makes the "September Effect" even more of a misread: the weakness isn't evenly distributed across the month.

The first three trading days of September—usually clustered around Labor Day—have posted an aggregate positive return over the past century. Everything else in the month is where the losses concentrate.

This means September's "effect" isn't really about September at all. It's about the period after Labor Day, when summer liquidity thins, institutional selling begins, and—historically—crises have happened to land. The first three days are a red herring that makes the month look like one coherent pattern when it's actually two completely different trading environments.

The Strategy Problem

Even if you believed the September Effect was real, acting on it is the wrong move. An investor who skipped September each year from 2021 through mid-2026 would have earned roughly 136% versus 124% for buy-and-hold—on the surface, a win. But that's a narrow window where September weakness happened to hold. An investor who moved to cash for the last four months of every year since 1957 would have earned only 66% versus about 10% annually for buy-and-hold.

The math on seasonal avoidance is unforgiving. You gain by missing a few bad months, but you lose by missing a disproportionate number of the best trading days. September has been positive in 2024 (plus 2.1%), 2025 (plus 3.6%), and 2010 (plus 8.9%). Missing even one of those outweighs avoiding several minus-2% months.

Add taxes on realized gains, transaction costs, and the probability that your re-entry price is worse than the bottom you were trying to avoid, and the strategy goes from "maybe helpful" to "structurally negative for most investors."

What Actually Matters

The current backdrop does have real concerns worth watching. The S&P 500 is up over 13% in 2026 and trading at 20x earnings, above its long-term average of 16x. Inflation ticked to 4.2% in May, and the 10-year Treasury yield climbed to 4.47%. These are concrete pressures on valuation, not calendar superstition.

If the market pulls back, it won't be because September is September. It'll be because earnings don't justify the multiple, rates move higher, or geopolitical risk accelerates. Those are real risks. They don't get more or less likely depending on what month it is.

The Honest Call

The September Effect is a distribution problem, not a seasonal pattern. A handful of crisis months pulled the mean negative while the median stayed positive, and the financial media learned to report the mean. The pattern looks real because the number is real—the mean really is negative. But the mean is also the wrong statistic for predicting what's likely to happen in any given September.

If you're sitting on a strong position, a September-themed pullback is a place to watch prices, not abandon them. If you're in cash waiting to deploy, the calendar isn't the entry signal you're looking for. The actual variables—earnings growth, valuation compression, rate direction, and whether a crisis happens to land—are the ones that move markets. September just happened to host more of them.