Does the full moon affect the stock market? We tested 14 indices since 2004

What the study found

In 2006 the economists Kathy Yuan, Lu Zheng and Qiaoqiao Zhu published “Are investors moonstruck?” in the Journal of Empirical Finance. Across 48 countries from 1973 to 2001, stock returns were lower in the 15 days around each full moon than in the 15 days around each new moon. For an equal-weighted world portfolio the gap was 3.53 basis points a day (a basis point is 0.01%), and switching between the two halves earned 3 to 5% a year before costs. Their explanation was mood: people feel worse around the full moon and pay less for shares.

It is a famous result, and it has a weakness every old finding shares: the market has had decades to learn about it.

How we tested it

We took every trading day of 14 stock indices from 2004 to September 2026: the S&P 500, the Dow, the Nasdaq 100, the Russell 2000, the FTSE 100, the DAX, the CAC 40, the Euro Stoxx 50 (from 2007), the AEX, the IBEX 35, the SMI, the Nikkei 225, the Hang Seng and the ASX 200. None of these days was in the study's data, so this is a clean second look.

Each day went into the full-moon half if it fell within 7 days of a full moon, and into the new-moon half otherwise, exactly as the study did it. The moon's dates come from standard astronomical formulas, which we checked first against six known eclipses: each was right to within two minutes. To measure luck, we shifted the whole lunar calendar by 2,000 random amounts and counted how often a fake calendar showed a gap as large as the real one. The rules for a pass were written down before any return was looked at.

What we found

The new-moon half did a little better than the full-moon half: +3.17 basis points a day against +2.35, a gap of +0.82. That is about a quarter of the study's gap. And 40% of the random calendars produced a gap at least that large, so the real moon does not stand out from made-up dates. A calendar would have needed a gap of +5.20 to beat 1 in 20 of the fakes.

The effect looked alive in the first half of our period, from 2004 to 2014 (+3.81). Then it reversed: from 2015 to 2026 the full-moon half was the better one (−1.99). Nine of the 14 indices leaned the study's way, short of the 10 we had set in advance as the bar.

Could you trade it?

A trader who owned the 14 indices in the new-moon half and sold them short in the full-moon half would have made +0.77% a year before costs from 2004 to 2026. After a broker's spread and the cost of holding positions overnight, the same switch lost 2.20% a year: it made +1.62% a year until 2014 and lost 5.78% a year after.

The gentler version, owning stocks only in the new-moon half and holding cash in the other, earned +3.90% a year after costs: about half of the +7.36% from simply staying invested.