“Sell in May” Over the Long Run

Does the conventional wisdom to “Sell in May” (and “Buy in November”) work over the long run, perhaps due to biological/psychological effects of seasons (such as Seasonal Affective Disorder)? To check, we turn to the very long run data set of Robert Shiller, which offers monthly levels of the S&P Composite Index since 1871. We split the investing year into two half-years (seasons): November through April, and May through October. Using nominal April and October levels for the S&P Composite Index from the end of April 1871 through October 2011 (281 seasons), we find that:

Over the entire sample period, the average seasonal returns are 1.8% for May-October and 3.7% for November-April, with standard deviations 12.8% and 11.5%, respectively. The higher return during November-April is not associated with higher risk. However, before (since) 1951, May-October averages 2.1% (1.4%), compared to 1.4% (6.6%) for November-April. Inconsistency in the best season across subperiods undermines belief in a reliable anomaly.

The following chart shows on a logarithmic scale cumulative values of $1.00 initial investments for three scenarios over the entire sample period:

1. Buy and hold stocks.
2. In stocks during May-October and in “dead” cash (0% return) during November-April.
3. In stocks during November-April and in dead cash during May-October.

November-April in stocks generates a much higher terminal value than May-October in stocks ($62.87 versus $4.05), but does not always outperform. May-October has the higher cumulative value during the 1940s and 1950s.

Moreover, the $254.69 terminal value of buy-and-hold easily beats the November-April seasonal. Note that fairly comparing seasonal results to buy-and-hold would (problematically) require contemporaneous estimates of the yield on cash and semiannual trading frictions.

Investors operating in real time would not know the full history shown in the chart. To see what investors may have believed about seasonal investing during the sample period, we look at the seasonal return difference in two ways.

The next chart plots the difference between the average November-April return and the average May-October return since 1891 based on both inception-to-date and rolling 20 years of historical data (as plausible alternative ways investors may have formed beliefs about seasonality). Over most of the sample, investors would reasonably believe that November-April outperforms May-October. However, there are three or four decades in the middle of the sample (out of 12 total decades) during which investors might reasonably believe that there is no seasonal effect or that May-October outperforms April-November.

The existence of long subperiods during which “Sell in May” does not outperform weakens the argument for a reliable biological/psychological explanation of seasonal returns.

For a different perspective on time variation of returns, we look at average seasonal returns by decade.

The final chart compares average seasonal returns by decade across the entire sample period. From this perspective, “Sell in May” works well since 1951, but not so well before. Again, existence of long subperiods during which “Sell in May” underperforms weakens the argument for a reliable biological/psychological explanation of seasonal returns.

In summary, evidence indicates that the conventional wisdom to “Sell in May” (or conversely to “Buy in November”) works well for U.S. stocks on average since the 1950s, but does not consistently work well before then.

Cautions regarding findings include:

• Assembling a portfolio that tracks the S&P Composite Index may have been difficult and costly during much of the sample period.
• The data set establishes monthly S&P Composite Index levels by averaging daily values during the month rather than using monthly closes. This approach blurs seasonal effects to some degree.
• Results do not include dividends, which may not be equal for the two seasons.
• Older data may be less reliable than recent data.

You May Also Enjoy...

• An Annual Worldwide Optimism Cycle (Sell in May)?
• “Sell in May” During Bull and Bear Conditions
• Sell Index Put Options Only When Above Long-term SMA?
• The (Dynamic) Meanings of Buy, Hold and Sell
• Buy at the Close and Sell at the Open?