Sell in May is an old stock market adage backed by surprisingly robust historical data. The idea is to exit stocks in late May and return the stock market for Halloween. As such it is sometimes called the Halloween indicator. Following this strategy, proponents argue, can help improve risk-adjusted returns. Does this strategy make sense in 2020?
The Data Suggests A 4%-5% Return Differential
The data over the long-run is pretty clear, as historical stock market data goes. Jacobsen and Bourman performed a comprehensive study over time and countries and found that stock market returns are significantly higher over the November-April period than the rest of the year. An update to the analysis from Jacobsen and Zhang lent further support to the hypothesis with updated data, including examining performance from 1998 to 2017.
In essence, the markets seem to offer superior risk-adjusted returns over winter and spring. Specifically returns are 4%-5% higher for stocks over November-April than May-October over a long run of historical data. Now, this does not happen each year, roughly 2 years out of 3 has seen a positive effect and performances by country differ, but over time the relationship appears historically robust.
There are some implementation issues to consider here. The first is that even though, if history holds you come out ahead, this is not a foolproof rule. It doesn’t work every single year, and even though its held up historically it may not hold going forward. Indeed, in the current volatile market a 4% return differential can amount to a couple of days of market moves.
A second consideration is taxes, trading into and out of stocks in a taxable account may be a drag on performance due to potentially higher taxes. This is true in the U.S. where many investors see a lower tax rate for holding investment gains for over a year. Taxes don’t necessarily derail the strategy, but are a consideration.
Thirdly, we should be clear, the relationship doesn’t typically show stock returns to be negative over the summer, just more muted at around 2% for the summer compared with around 6% or so in the winter period. Therefore, there may not be such a great need to sell in May since returns may still be positive, but more to gear up your stock exposure in November when returns prospects are apparently strongest.
Finally, even though academics have many theories here, we still don’t know quite why this calendar effect happens. The best theory would have statistical support, and a clear explanation. There are competing theories from sunshine to vacations as to why the effect has occurred in the past, but we don’t have alignment on a clear explanation for this phenomena and opinions differ. Without that it’s hard to have rock-solid faith in the anomaly even though the historical data is remarkably strong.
It’s an understatement to say 2020 will be a unique year in many ways. The shape of any COVID-19 recovery, the process toward treatment and vaccines and indeed the U.S. election in November are all major factors that will likely drive the markets over the next several months, aside from any seasonal factors and many other considerations.
However, for those already feeling apprehensive about relatively high stock valuations and the downside risk from more negative COVID-19 outcomes, the evidence to sell or at least trim exposure in May is quite compelling, especially when compared with other stock market rules that see more mixed results, or appear to have a smaller impact after their discovery.
However, do bear in mind that the rule is only implying that stock market returns will be lower through November, not that they will necessarily be negative. Even so it’s one reason to consider taking some risk off the table especially if you have other compelling investment to rotate into that may offer a return in excess of the expected 2% or so that stocks may offer on this theory.
Get The Best Financial Tips
Straight to your inbox
Subscribe to our mailing list and get interesting stuff and updates to your email inbox.
Thank you for subscribing.