The influence of the moon on human behavior has been featured in many, not only scientific publications. This paper tests the hypothesis that the one-session rates of return of 107 stock indices and commodities, calculated for full moon and new moon phases (considering moon phases falling on Saturdays and Sundays – Regime 1, or ignoring them – Regime 2), are statis-tically different form zero (at the significance level of 95%). Calculations presented in this paper indicate that the one-session average rates of return for the sessions, when the moon was in new phase, are statistically different from zero. For Regime 1, the null hypothesis was rejected for 15 and 39 indices and commodities when the moon was in the full and noon phase (α=5%), respectively, and for Regime 2, the number of null hypothesis rejection was equal to 26 (full moon) and 45 (new moon) – for α=5%. In case of testing equality of average rates of return in two populations, the number of null hypothesis rejection was equal in the Regime 1: 5 and 21 (for full and new moon, respectively), whilst in the Regime 2, it mounted to 5 and 17. In case of testing the null hypothesis regarding equality of one-session average rates of return, computed for each day of the week, the result permit to reject in the Regime 1: 46 and 34 cases, and in the Regime 2: 39 and 54 cases (full and new moon, respectively). Calculations of one-session average rates of return, regarding moon phases falling in a specific month, displayed that they are statistically different from zero – in the Regime 1: in 131 and 70 cases and in the Regime 2: in 120 and 80 cases (for full and new moon, respectively).
Keywords: Calendar Effects, Market Anomalies, Market Efficiency, Moon Influence
