There is an old trading adage that says “Sell in May and go away.” According to it, investors should close all equity positions in May, go to cash and get back into the market in November, thus avoiding a seasonal decline during summer and the first part of fall. However, the last three May months that still reside in my memory were profitable, obviously questioning if this concept is still valid.
If we look over the 20 years course, we can see that there were only 8 losing May months (60% profitable months,) with an average gain of 0.4%. Wouldn’t investors be better off by actually selling in June instead of May and getting back into the market in November? I decided to check this concept on S&P500 and compare it to the original “Sell in May” strategy.
The assumptions for my backtest are the following:
- Testing period: 6/2/95 – 6/1/15
- Position sizing: 100% of equity
- Commission per 1 trade: $10 or $20 per round
- Taxes excluded
As you can see from the chart below, “Sell in June” works pretty well. Initial $10,000 balance has grown to $33,213 in 20 years, what gives us a total return of 232.1% or about 6.2% per year. During the testing period there were 20 trades in total, 1 per each year with a win rate of 80%, max DD of 35%, and a Sharpe ratio of 0.64. Taking into account a total market exposure of 56.6%, I would assume that results are decent even for such a simple idea.
However, the purpose of the post is to check if “selling in June” in superior to “selling in May.” In order to do that, I am comparing the original idea below, which assumes buying on the first day of November and then selling on the first day of May. The assumptions for the test are the same, plus I have added a simple “Buy and Hold” strategy to put the results into perspective.
Despite an average May gain of 0.4% over the last 20 years for S&P 500, tests show that “selling in May” is in fact superior to “selling in June.” Over the testing period initial $10,000 has grown to $35,597 or 256% in total, with an annualized gain of 6.56%. The improvement can also be seen in max DD (33,9% vs. 35%), Sharpe ratio (0.72 vs. 0.64) and even in percent profitable trades (85% vs 80%). Similar results can be seen on DJIA and Nasdaq, with an exception of Russell 2000, where “Sell in June” works better.
Even though a simple “Buy and Hold” strategy performs better, the difference is negligible (annualized return 6.56% vs. 7.18% for B&H). However, by utilizing “Sell in May and Go Away” strategy investors can significantly reduce max DD 33.9% vs. 57%, improve Sharpe ratio (0.72 vs. 0.53 for B&H), and reduce market exposure to 48% vs. 100%, freeing up some cash for other opportunities.
“Sell in May and Go Away” remains a valid and sound strategy which can be applied on a wide range of markets, not only in the U.S. While sacrificing some returns when compared to simple “Buy and Hold” strategy, investors get a chance to drastically decrease their market exposure and risk, while improving risk/return metrics. In addition, by allocation a portion of their total capital, investors may be able to diversify among trading strategies.
Although I have tested the strategy on an index which can’t be invested directly, one can easily utilize ETF, like SPY as a proxy for S&P 500.