A common adage in Wall Street circles is “sell in May and stay away,” which is based on the notion that the stock market tends to perform poorly between May and October relative to other portions of the calendar year. This trading advice is also known as the “Halloween Indicator” because the period to avoid equity markets ends in October. I created two charts that show the typical evidence assembled in support of the Halloween effect. These charts summarize stock performance, as represented by the S&P 500, over the period of 1950 through 2012.
The second chart plots what are called “Sharpe ratios,” which is a measure of an asset’s returns relative to its riskiness (i.e. volatility). Higher Sharpe ratios indicate better risk-adjusted returns.
Both charts seem to suggest that the Halloween effect is real (and scary). Statistical tests that compare the mean daily returns for the two periods confirm that the mean return for the Halloween period (1.1 basis points) is significantly (at the 99% level of significance) lower than the mean return during the “normal” period (5.6 basis points).
In spite of the evidence above, as well as prominent academic studies supporting the Halloween effect (e.g here), we should remain cautious before we sell all of our stocks and hunker down until November 1. Patterns observed in the past don’t necessary predict the future well. Moreover, the Halloween effect might just be an artifact of extreme periods in the stock market (e.g. the stock market crash of October 1987) that don’t occur with any regularity, but by chance happen to be in the Halloween period. Another academic study makes this point and goes on to show that the Halloween effect strategy underperforms a simple buy-and-hold alternative.