Claim check: September is the worst month
“September is the worst month of the year for stocks.”
What the data says
Does not hold Not supported by the data: September returned -0.5% a month on average against +1.4% for the other months in S&P 500 (n = 17 months, 2010-01-04 to 2026-11-10; t = -1.7, likely noise; rank 1 of 12 by average return). The direction points the claimed way, but the result cannot be told apart from chance.
No effect we can tell from chance, or the opposite effect.
Who says it
Stock Trader's Almanac repeats it every August.
Origins are as commonly cited and were not independently verified.
How we tested it
Month-end to month-end returns for SPY. September is compared with the other eleven months (Welch t-stat), must be among the two lowest of the twelve calendar months by average return, and negative relative to the rest in both halves of the history.
Results
| Market | Observations | Mean | Rest | Difference | Up share | t-stat | Rank of 12 | Reading |
|---|---|---|---|---|---|---|---|---|
| S&P 500 | 17 | -0.5% | +1.4% | -1.88% | 53% | -1.7 (likely noise) | 1 | Right way, likely noise |
Average monthly return by calendar month (S&P 500)
Stability across the two halves of history
A saying that only works in one half of the history is not much of a rule. Each row shows the effect in the full history, the first half and the second half; green means the claimed effect was present.
| Market | Full history | First half | Second half |
|---|---|---|---|
| S&P 500 | -1.9% | -0.5% (8) | -3.1% (9) |
Context and cross-checks
Shown for context; not part of the verdict.
Other markets (context, not part of the verdict)
| Market | Observations | Mean | Rest | Difference | Up share | t-stat | Rank of 12 | Reading |
|---|---|---|---|---|---|---|---|---|
| Gold | 17 | -1.5% | +0.9% | -2.46% | 29% | -1.8 (likely noise) | 1 | Right way, likely noise |
| Silver | 17 | -3.0% | +1.3% | -4.27% | 41% | -1.5 (likely noise) | 1 | Right way, likely noise |
| Commodities | 17 | -0.5% | +0.5% | -0.94% | 59% | -0.7 (likely noise) | 3 | Right way, likely noise |
| Agriculture | 17 | -0.9% | +0.3% | -1.15% | 35% | -1.0 (likely noise) | 2 | Right way, likely noise |
| EUR/USD | 17 | -0.4% | 0.0% | -0.40% | 41% | -0.5 (likely noise) | 4 | Right way, likely noise |
| Bitcoin | 12 | -1.1% | +6.3% | -7.40% | 50% | -2.7 (likely noise) | 2 | Too little data |
| Ether | 9 | -5.5% | +5.9% | -11.36% | 44% | -2.5 (likely noise) | 2 | Too little data |
Average monthly return by calendar month (Gold)
The honest caveat
With ~15 Septembers, the ranking among 12 months is highly noisy.
Descriptive statistics from historical data, before costs. Not a forecast and not investment advice.
Questions
September is the worst month of the year for stocks.
Not supported by the data: September returned -0.5% a month on average against +1.4% for the other months in S&P 500 (n = 17 months, 2010-01-04 to 2026-11-10; t = -1.7, likely noise; rank 1 of 12 by average return). The direction points the claimed way, but the result cannot be told apart from chance.
How was this claim tested?
Month-end to month-end returns for SPY. September is compared with the other eleven months (Welch t-stat), must be among the two lowest of the twelve calendar months by average return, and negative relative to the rest in both halves of the history.
Related strategies on Quant Arena
Markets used
S&P 500 · Gold · Silver · Commodities · Agriculture · EUR/USD · Bitcoin · Ether