The November Effect

The old adage “Sell in May, and then go away” has a corollary that comes into effect this time of year. You don’t go away forever, of course; the idea is that savvy investors get back into the market around the time of Halloween, in order to take advantage of the so-called November Effect. Like a lot of traditional market wisdom, this one has a certain amount of truth to it. But there’s also a surprisingly robust and growing amount of research around the phenomenon.

The November Effect was first observed in Great Britain, but it holds up well here in the U.S., too, as well as in most stock markets around the globe. The canonical study of the November Effect was published in 2001, testing the theory in international markets from 1973 to 1998. The researchers, Ben Jacobsen and Sven Bouman, found that in 36 out of 37 international exchanges, returns were significantly higher from November through April than they were in May through October.

Jacobsen has now revisited the subject in a paper written with Cherry Zhang, his colleague at Massey University in New Zealand, and if anything, the phenomenon is getting stronger and more widespread. This time, they looked at 108 different markets using all available historical data, which meant they had a whopping 300 years’ worth of information regarding the market in the United Kingdom, and more than 200 years’ worth for the United States. The bottom line: Returns from November through April over the past 50 years were on average 6.25 percent higher than returns from May through October.

Adding all the markets around the world didn’t dilute the effect any; although it was more persistent in the more developed economies, the November Effect held true in 81 out of 108 countries. A 2007 study found the effect even taking hold in seven out of nine Arabic stock markets.

And the phenomenon has been stronger in more recent years. For all the data collected by Jacobsen and Zhang, over the course of centuries, the November Effect accounted for an increase of 4.52 percent – a solid enough difference, but significantly less than the effect in the past five decades. Over the last five 10-year sub-periods that the researchers looked at, the November Effect ranged from 5.08 percent to 8.91 percent. In other words, it’s lately been at least 5 percent, and is regularly much more than that.

The awareness of the November Effect dates back to at least 1935 in London, although at that point the “Sell in May” part of it was already being referred to as an old piece of conventional wisdom. An article in Britain’s Financial Times from May 10, 1935, quoted an 88-year-old broker named Douglas Eaton, who was well familiar with the market’s annual cycle. “It was always sell in May,” Eaton said. “I think it came about because that is when so many of those who originate the business in the market start to take their holidays, go to Lord’s [cricket ground] and all that sort of thing.”

That suggests that the effect is rooted in brokers and financial professionals taking summer vacations, but Jacobsen and Cherry stress that they don’t really know why the phenomenon persists. It is certainly clear that trading volume is lower over the three summer months, but that would hardly seem to be enough to keep stocks artificially depressed for half the year, only to spring forward in the autumn. Other researchers have suggested that the shift in stock prices is a result of Seasonal Affective Disorder, but that theory has been largely debunked.