May Madness

With May roughly a week away, headline admonitions to “sell in May” and avoid the “summer doldrums” may be on their way—particularly because the last couple summers featured some big volatility. Yet, as we frequently write, a calendar turn alone has little relation (causation or correlation) to market performance. There’s nothing fundamental about any set of months or time of year that causes markets to inherently want to move higher or lower.

What’s more, the aphorism is at best half-advice. It loosely tells you when to sell, but it’s conspicuously silent on when to buy back in. (Although the question of when to sell in May does also come up—the 1^st? the 15^th? the 17^th? April 30?) Among MarketMinder editorial staffers, there seems to be little consensus as to when the adage implies folks should re-enter stocks. Some thought June. Others thought after the summer. Still others thought it advised to be out fully six months—through October. And a few even noted the British version, “Sell in May and go away; don’t come back till St. Leger’s Day.” (Though, even this isn’t very instructive. Why a British horse race—especially one that changes dates every year—matters is a headscratcher.) Googling the phrase revealed enough additional derivations to make one’s teeth hurt.

Whichever version of the admonition you subscribe to, this remains true: To be repeatedly successful at such short-term moves requires a cooperative market and perfect exit and reentry points. And there’s little supporting evidence to show seasonal patterns exist, creating such opportunities. Using monthly S&P 500 returns from 1926 to 2012, May, on average, has been a positive month. In fact, only one month is actually negative on average (September). And September is skewed down by some isolated big negatives during the Great Depression. Likewise, cutting the data other ways—three and six months—fails to reveal any discernible advantage. For example, average returns for June, July and August are actually the best of any three-month period in the calendar year. And since 1926, summer returns are positive roughly 70% of the time. None of this seems like sure-fire negativity to try to sidestep. That’s not to say it won’t happen, but betting on the calendar without question seems a strategy with gaping holes.

Now, if you adhere to the six-month time frame, it is true the summer half has a lower average than the winter half. But the average is still positive, and May-October is positive, again, about 70% of the time. But there’s little saying whatever you buy (cash, bonds, etc.) outperforms stocks or even covers the transaction costs, tax implications, etc. of selling out in May—whether you sit out for one, three or six months or until St. Leger’s Day.

Theory poses additional problems for sell in May adherents. If “sell in May” worked, wouldn’t rational market participants sell before May—beating the crowd? And as more investors shifted to selling earlier, before April and so on? Ultimately, one could be in the rather ridiculous position of selling in June—“because May’s only 11 months away.”

So, ignore “sell in May” and other seasonal sayings about selling automatically based on dates, months, pro football teams, horse races, Santa Claus or anything else based not on fundamentals or portfolio theory, but on arbitrary collections of days.