The Halloween Indicator, 'Sell in May and Go Away': Everywhere and All the Time

Abstract

To answer the sceptics, we use all historical data (62962 observations) on all stock market indices worldwide to verify the robustness of the so-called Halloween Indicator or Sell in May effect. The effect seems remarkably robust with returns on average 4% higher during November-April period than during May-October. A new test for the effect offers some additional insights. Worldwide excess returns during summer seem negative (around -1%) and often significantly so suggesting a flat or negative risk return relation. Only for Mauritius do we find a significantly positive risk return relation during May-October. Our dataset also allows for a new (upper bound) estimate for the equity premium of around 4%.

Keywords: seasonal anomalies, sell in May, Halloween indicator, long time series data

JEL Classification: G10, G14