The idea that you should buy stock in October and sell in May is known as the Halloween strategy. When people say “Sell in May and go away”, they mean the same thing. It has become a popular investment strategy because it is so easy to understand and widely believed to yield excess returns. By the latter we mean a higher return than the amount of risk you take would justify. In other words, by selling in May and coming back in October, you can earn a free lunch. But if this strategy is so successful and if so many people are indeed following it, wouldn’t the advantage disappear? And should you buy stocks next month? I dived into the literature on the Halloween effect to get some answers for you.
"By selling in May and coming back in October, you can earn a free lunch"
Evidence for the Halloween strategy
In 2002, scientists for the first time published an article in which they presented evidence for the existence of excess returns during winter months, which is to say from November to April. Sven Bouman and Ben Jacobsen found that the “Sell in May effect” was present in 36 out of the 37 countries investigated. In the United States, returns were on average more than 5% higher in the period from November to April than in the remaining months. In most European countries, the effect is even stronger, with an average difference of 6% during these time periods. Sounds pretty good and easy to obtain!
But how come that stocks perform better in winter? Some argue that it could be the case that people are less risk averse in winter times and take higher risk, but why would that be the case? Another explanation is that people’s liquidity preferences change. When they go on a vacation, they need cash, not stocks. Not unreasonable, but quite far-fetched since most investors won’t have to sell their stock in order to be able to afford a vacation. The truth is, we do not know what causes the Halloween effect. It could be due to data mining or a selection bias, as suggested by some researchers. We just don’t know.
"The truth is, we do not know what causes the Halloween effect"
But what’s even more curious is that it does not seem to disappear. Market anomalies are usually arbitraged away, as expected by the Efficient Market Hypothesis. The Halloween effect is not.
10 years after the publication of the famous article mentioned above, one of the two academics re-examined the Halloween effect and looked at 108 stock markets over their available histories. What he found was interesting: the Halloween effect endured! Stock market returns were higher during winter months in 81 out of 108 countries, whereas it was higher in summer months in only two countries. The effect was statistically significant in 35 countries, while Europe, North America and Asia showed a stronger effect than other regions. Besides, the scientists show that the effect has become stronger over time. Again, there is no clear explanation for this phenomenon, so the puzzle persists. Other researchers also find evidence for the Halloween effect, some even find a difference of 10% in returns in winter and summer months. They found that the Halloween effect was present even during the fourteen years after the publication of the famous article.
So far, following the Halloween strategy seems to be good investment advice. However, a recent study throws a spanner in the works. When taking actually available and adequate investment instruments into account, as well as the publication date of the Bouman and Jacobsen article, the Halloween effect has become weaker over time and has diminished. After all, you need a real investment instrument to put the strategy into practice.
"The Halloween effect has become weaker over time and has diminished"
In line with the Efficient Market Hypothesis, the Halloween strategy seems not to constitute a free lunch when you want to put the strategy to the test. Unfortunately for investors, these findings are mostly confirmed by the newest research.
Ever since the first scientific evidence for the Halloween effect was presented, investors have used the strategy to buy stock around Halloween and sell in May. Researchers show, almost invariably, that this strategy would have yielded excess returns, popularizing the strategy around the world. However, the effects seems to have diminished over time. Moreover, when accounting for the availability of real world investment instruments, rather than just relying on theoretical returns on indices, the Halloween effects weaken further. When Halloween arrives next month, consider the popular investment strategy with care. Fata morganas are nothing more than a mirage.