Sell in May and Go Away? Market Timing and Stock Return Seasonality
by RONNIE R. SHAH, May 2012, Dimensional
Basically this paper asks the question if May, as opposed to picking any of other 11 Months to sell (in other words, small changes to the strategy), makes any difference. It also looks at which month to repurchase again because staying out is not part of the strategy … so it also begs the question when to get back in (and which of the 12 months is that done)!
Conclusion: the phenomena can be explained by chance since the winter – summer returns represent a backtest bias in the data. Removing the bias through bootstrapped simulations suggest the returns differences between periods may occur by chance rather than rule. Out-of-sample testing also supports this claim.
My observation: What may have held true in the past may no longer hold true of the future. And statistically testing the data suggests that the rule is as random as any other possible month should the future decide to change the data (only recognized once the future returns become historical returns … so this too is not predictable).
My conclusion? Develope a prudent plan supported by an academically based philosophy that is flexible as the situation warrants.