In this study, we performed a comprehensive out-of-sample analysis of a calendar anomaly studied by previous researchers and identified by the adage “Sell in May and go away.” We found that the adage remains good advice: Reducing equity exposure beginning in May and levering it up beginning in November persists as a profitable market-timing strategy.
We found that the sell-in-May effect not only persists but also maintains the same economic magnitude as in the sample of previous researchers. On average across 37 countries, stock returns are roughly 10 percentage points higher for November–April half-year periods than for May–October half-year periods. This out-of-sample persistence indicates that the effect is enduring and not a statistical fluke.
We also showed how the sell-in-May effect could have been profitably exploited through low-cost trading strategies using extremely liquid securities. Simple market-timing strategies based on the effect deliver high Sharpe and information ratios. For example, a trading strategy with 0% in stocks in May–October half-year periods and 200% in stocks (spot plus futures) in November–April half-year periods would have a Sharpe ratio 40% larger than that of a buy-and-hold strategy. The annualized information ratio would equal 0.34, placing the sell-in-May market timer in the top quartile of the distribution of active equity mutual fund managers in the United States.
We also presented novel evidence consistent with widespread seasonal variation in aggregate risk aversion as the cause of the sell-in-May effect. In addition to its presence in the equity risk premium, as documented by previous researchers, we found an economically large and statistically significant sell-in-May effect in strategies that exploit the size, value, foreign exchange carry trade, equity volatility risk, and credit risk premiums. Because these other trading strategies are outside the realm of typical retail investors, our results indicate that risk aversion seasonality may affect not only retail investors but also professional market participants. Therefore, widespread seasonality in financial markets’ aggregate risk aversion is likely the proximate cause of the sell-in-May effect. To the extent that this seasonality is ultimately irrational, our results suggest that markets may be slower to arbitrage away inefficiencies than previously thought.