Superstition or a measurable pattern?
The full moon has long been blamed for strange behavior — supposedly on the stock market, too. Viewed soberly, that makes it a perfect candidate for a data check. So we examined the Dow Jones around every full moon: what happens in the window from five trading days before to five trading days after?
The answer is surprisingly nuanced. Across all months there is a signal — but a weak one. Only in July does it become genuinely striking.
The overall signal: real, but weak
Across 129 full moon events, the Dow Jones gained an average of +0.45 % in the t-5 to t+5 window. The win rate — the share of events with a positive outcome — was 63 %, with a median of +0.63 %. At first glance, that sounds like a clear "full moon bonus."
But the statistics temper the enthusiasm. The t-value is 1.77 and the p-value is 0.08. A result is usually considered significant below a p-value of 0.05. At 0.08, the overall signal is therefore only borderline — it could easily be chance. The dispersion is also considerable, with a standard deviation of 2.74 %.
In short: averaged across all months, the lunar effect on stocks is measurable but statistically unconvincing. Betting on it alone means standing on thin ice.
A look at the event window
The chart below shows the cumulative return of the Dow Jones around the full moon, normalized to zero on the full moon day (t0). This reveals the typical path without different price levels distorting the picture.
The shape is interesting. In the days before the full moon (t-5 to t-3) the index drifts slightly negative. From t-2 the curve turns upward, and after the full moon (t+1 to t+5) it keeps rising steadily — up to around +0.30 %. The strength lies less in the full moon day itself than in the phase just before and after it.
The July outlier: 90 % win rate
Things get truly interesting in the monthly split. Break the 129 events down by calendar month, and the July full moon stands out clearly:
| Metric (July full moon) | Value |
|---|---|
| Events (n) | 10 |
| Win rate | 90 % |
| Avg return | +0.95 % |
| t-value | 3.21 |
| p-value | 0.01 |
A p-value of 0.01 means this pattern is statistically significant — far more robust than the overall signal. In nine of ten cases the Dow Jones closed the July full moon window in positive territory, gaining nearly one percent on average.
Even so, caution is warranted. Ten events are a small sample. One or two unusual years can shift the statistics noticeably. A low p-value on a small n is a hint — not proof. The July full moon effect deserves attention, but not a blind wager.
August confirms the summer picture
July does not stand alone. The August full moon is also notably positive: +1.53 % on average with an 82 % win rate across 11 events — likewise significant. November comes out borderline.
That places the full moon pattern within a larger picture: around the full moon, the high-summer months were historically above average. Whether that is down to the moon or to other seasonal forces cannot be separated from the data alone.
Where might the effect come from?
There is still no solid scientific explanation for a genuine lunar effect on equities — the research is contested. The factors discussed are mostly indirect:
- Liquidity and trading rhythm: Full moon dates often coincide with certain weeks in the monthly cycle that carry their own seasonal patterns.
- Risk appetite and mood: Some studies suspect a slight influence of moon phases on investor behavior — but the evidence is weak and not reproducible.
- Random clustering: With twelve months and small samples, there is almost always one "striking" month without any real underlying cause.
The key point remains: correlation is not causation. The July full moon may mark a real seasonal window — but need not be causally linked to the moon.
Conclusion
The Dow Jones full moon effect is a fine example of testing superstition with data. Across all months the signal is real but weak (p=0.08). In July, by contrast, a 90 % win rate and p=0.01 point to a statistically significant pattern — weakened only by the small sample of ten events.
As a standalone trading signal, the full moon is no help. As a context building block alongside the monthly cycle, turn-of-month and significance tests, it can round out the picture. Check the moon phases yourself at seasonalpha.ai/mondphasen — with p-value, win rate and event window for every ticker.
Frequently Asked Questions
Is there a full moon effect on the stock market?
Measurable yes, convincing no. The Dow Jones gained +0.45 % on average around the full moon (63 % win rate), but with a p-value of 0.08 the overall signal is only borderline significant — not statistically reliable.
Why is the July full moon so striking?
In July the effect reached a 90 % win rate and +0.95 % on average at p=0.01 — statistically significant. Why July in particular stands out cannot be pinned down; the sample is also small at ten events.
Can I trade on moon phases?
Only with caution. The effect is weak, the July signal rests on few observations, and a scientific cause is missing. Moon phases work at best as supplementary context, not as a standalone buy signal.
Where can I verify the full moon effect myself?
On seasonalpha.ai under "Moon Phases." There you can see the event window around the full and new moon, complete with a significance readout (t-value, p-value, win rate, n) for the Dow Jones and many other tickers.