PurposeThe subject of this paper is seasonal variation in the return on stocks. The phenomenon we analyze here is known as the “Halloween effect” or the trading strategy “sell in May and go away.” The authors test the hypothesis that stock markets tend to return considerably less in the six months beginning in May than in the other half of the year. This effect has shown persistency over time and is seemingly large enough to be a candidate for economic significance.Design/methodology/approachThe authors analyze monthly data from 13 countries for the period 1958–2019, using the Kruskal–Wallis test, t-test and a boot-strap based estimator. In addition, we look a sub-periods for a larger group of countries and include data on both stock returns and interest rates.FindingsThe authors find a strong seasonal effect in a large majority of the markets, with the period from November to April seeing higher returns than the other six months of the year. This result also holds for a larger sample of countries based on data from a shorter period. The effect is found to be economically significant in most countries in the sample. The authors examine one potential explanation for seasonal variation in stock returns, i.e. seasonal affective disorder (SAD). The authors find some, albeit weak, support for this hypothesis.Originality/valueThis paper uses a rich dataset that has not been used for this purpose before and robust tests of statistical and economic significance to shed light on an important aspect of global financial markets.