Peter Gibson and Jeff Evans at CIBC World Markets have an interesting twist on an old investing saw: "Would 'sell in May' please go away?" they ask in a note to clients.
"As the month of May begins, one can usually count on several things: the days being longer, the weather being warmer... and someone, somewhere, pitching 'sell in May' seasonality research," they said, referring to the oft-discussed notion that it is best to ditch stocks at the start of May and stay out of the market until autumn – usually the start of November.
Any market-timing approach to investing can be tantalizing because it reduces the complexities of investing to a simple in-and-out move. Unfortunately, they don't seem to work – or, at least, they don't seem to work after everyone discovers them. And the "sell in May" timing strategy might be among the most suspect.
Mr. Gibson and Mr. Evans get technical (ie, above my head) in their discussion of the strategy, referring to a "median 30-year rolling information coefficient." But the conclusion is simple enough to understand: The predictive accuracy rate of a sell in May strategy is close to zero. What's perhaps even worse is that there is no apparent reason behind the strategy.
"It is our strong belief that investment strategies should, wherever possible, be motivated by a clear causal link to some underlying 'driver,' they said. "In other words, if sell-in-May works, why? It is not good enough, in our view, to simply highlight the correlation between a signal and subsequent returns – we should always seek to identify the true cause of the correlation. If a phenomenon survives after a nuanced and detailed analysis, it should provide greater comfort that the phenomenon is "real" and also that it is likely to repeat. If not, it should be avoided at all costs...."