Sell in May and Go Away?

It's that time of year again.  As the old phrase claims, now that May is here, it's time to sell stocks and 'go away' as the stock market enters the period of the year which has supposedly been the weakest.  Looking at the historical results, however, shows that the record of this indicator is mixed at best. 

The chart below shows the average performance of the Dow Jones Industrial Average from May 1st through September 1st using various time frames.  Depending on your window, the results vary widely.  Over the last five years, the Summer months have been positive, with an average return of 3.2%.  However, when we look back ten years, the average return over the same time period is a loss of 1.16%.  If we look back 25 years, the average return jumps back up to 2.4%, and then over a fifty-year period it falls back down to 0.84%.  "Sell in May and go away" makes for a nice catch phrase, but it's not really an indicator to bank on.

Comments

[Barry Ritholtz]

Can you combine this with a MACD buy or sell signal?

Some people swear by that!

[Paul Teetor]

Thank you for investigating this topic. This is an important question for stock investors.

I think your analysis might be off the mark, however, so I tested the "Sell In May" idea in a different way.

(1) I collected monthly returns for the S&P500 from January 1986 through April 2008.

(2) I split the data into two groups: Monthly returns for the May to September period, and monthly returns for all other months.

(3) I performed a t test on the average monthly returns, comparing the two groups. I tested the hypothesis that the two groups had equal returns, versus the hypothesis that Summer months had lower average returns.

(4) At the 89.88% confidence level, the Summer months do indeed have lower average returns than all other months.

Now, a Real Statistician would demand 90% or 95% confidence. But this is trading we're talking about, not clinical drug trials, so the 89% confidence got my attention. To measure the possible impact of the "Sell in May" approach, I built a spreadsheet which compared two trading strategies for the S&P500: (A) buy-and-hold versus (B) holding stocks with 1.5 leverage outside the Summer months and being flat during the Summer months. (The 1.5 leverage is necessary to get comparable results, since strategy B is in the market only 2/3 of the time. Alternatively, one could use 1.0 leverage and measure the results using the Sharpe ratio.) The time period was January '86 through April '08.

The results were encouraging. The buy-and-hold strategy gained 561%, but the Sell in May strategy gained 804%. The Sell in May strategy also had a better ratio of profit to maximum drawdown.

My current interest here is answering the question, should I sell calls against my stocks during the Summer months? It seems I should. Another possible application is to hold a leveraged position using futures during non-Summer months, and be flat during the Summer -- which would make for a more restful Summer vacation in any event.

(Technical note: I measured the monthly returns from the 15th of each month to the 15th of the subsequent month, since those days are closer to options expiration dates. I did not consider the impact of dividends, brokerage fees, or other frictional forces.)