fbpx Sell In May and Go Away...or don’t (23:9&10) | Option Strategist
Home » Blog » 2024 » 05 » Sell In May and Go Away...or don’t (23:9&10)
Sell in May and Go Away?
By Lawrence G. McMillan

This article was originally published in The Option Strategist Newsletter Volume 23, No. 9&10 on May 23, 2014. 

One of the luxuries of publishing a “double issue” is that we can spread our range of topics out a bit. That’s what this article is about. It’s confirming and recounting some technical studies that have made their way around lately. They don’t necessarily have anything to do with the option market directly, but they certainly have to do with market direction and volatility. There are several technical studies presented in this article. Most of them are rather long-term, so we would not necessarily be using them directly in our trades, but they are certainly useful as a guide, and then shorter-term, sentiment-based indicators can be used to refine the timing of these longer-term signals.

An Interesting Footnote to “Sell in May and Go Away”

Not wanting to merely accept someone else’s work without checking, I coded up a study of my own:
If $SPX trades at a new all-time high in May, a) what did the market do from the end of May until the end of the year?, and b) what did the “Sell in May...” period do (i.e., how did $SPX perform from the end of April through the end of October)?

$SPX was only created in its present from in 1957. It was backdated prior to that, and for some reason, most data sets have $SPX historic prices only back to the beginning of 1950. Standard & Poors created its first stock index in 1923, but I used only the post-1949 data for this study.

It turns out that $SPX has made a new all-time intraday high 20 times during the month of May, including this year ($SPX made a new all-time closing and intraday highs last week, on Tuesday, May 13th). In 17 of the previous 19 years (we don’t know the outcome yet this year), the market rose in the 7-month period, from June through December. In other words, this is a new seasonal indicator of sorts:

When $SPX makes a new all-time intraday high in May, then buy “the market” at the close of the last day of May and hold through the last trading day of the year

The average gain for all 19 years was +8.0% over the 7-month period. In the 17 winning years, the average gain was +9.6%.

The best year was 1954, when the 7-month period produced a gain of 23.4% (even I am not old enough to remember that year!). 1995, 1996, 1997, and 1999 produced new highs in May and double-digit gains for the 7-month period.

The only two losing years were 1990, when Iraq attacked Kuwait in the early fall. That year, the market rallied back late in the year, but the 7-month period still ended up with a loss of 8.6%. The only other losing year was 1986, when the 7-month period produced a modest loss of –2.1%.


But merely knowing that one could buy and hold for seven months and make money doesn’t really tell the entire story. For example, what kind of drawdowns were there? It turns out, they were modest. The average drawdown for all 19 years was –4.8% (and that includes the –18.4% drawdown for 1990, which was the worst one). The median drawdown was –3.4% (reflecting the outlier that 1990 is).

In winning years, the average drawdown was –3.8% and the median drawdown was –3.4% (the same as when all years are considered).

The smallest drawdown for the 7-month period was –0.7% in 1967.

Sell in May and Go Away

But even with these seemingly positive statistics, one might wonder how the traditional “sell in May and go away” seasonal trade fared in these years, compared with how in normally fares.

The “Sell in May and Go Away” seasonal trade was established by Yale Hirsch a number of years ago. The complete statistical analysis of this seasonality can be found in Stock Trader’s Almanac each year. Jeff Hirsch, Yale’s son and keeper of the flame, has issued some refinements to the system in recent years – trying to generate better entry and exit points. However, the original system was to sell at the close of trading on the last day of April and to buy back in on the last trading day of October.

Using that latter definition, for the 19 years in which a new all-time intraday was made in May, the average year saw a gain of 5.0% over the 6-month Maythrough- October period. There were 15 winning years and 3 losing years (one year was flat). The best gain was +14.1% in 1997, and the worst was –8.1% in 1990, again.

Table 1: Year-by-Year When All-Time High in May
Year      7 mo P&L    Drawdown    Sell in May...
1950        9.1%       -11.2%         8.9%
1951       10.2%        -2.8%         2.2%
1954       23.4%        -3.4%        12.1%
1959        2.0%        -6.0%         0.0%
1961        7.5%        -3.2%         5.1%
1964        5.5%        -2.1%         6.8%
1965        4.5%        -7.8%         3.7%
1967        8.3%        -0.7%        -0.7%
1968        5.3%        -2.0%         6.2%
1972        7.8%        -3.4%         3.6%
1983        1.6%        -2.0%        -0.5%
1985       11.5%        -4.7%         5.6%
1986       -2.1%        -7.0%         3.6%
1990       -8.6%       -18.2%        -8.1%
1995       15.5%        -1.0%        13.0%
1996       10.7%        -6.4%         7.8%
1997       14.4%        -1.0%        14.1%
1999       12.9%        -4.2%         2.1%
2013       13.3%        -3.5%        10.0%

Sum       152.8%       -90.6%        95.5%
Avg         8.0%        -4.8%         5.0%
Median      8.3%        -3.4%         5.1%
# wins        17                        15
# losses       2                         3
Tot wins  163.5%                    104.8%
Tot loss  -10.7%                     -9.3%
Avg win     9.6%                      7.0%
Avg loss   -5.4%                     -3.1%
Max        23.4%        -0.7%        14.1%
Min        -8.6%       -18.2%        -8.1%

Citing the current edition of Stock Traders Almanac 2014, covering the years 1950 through 2012, the Dow Jones 30 Industrials averaged a mere +0.3% gain from May 1 through Oct 31 over those years. If one had invested $10,000 in 1950 and persisted in being long only in that six-month period each year, he would actually have wound up with a loss of –$1,105 by Oct 31, 2012. There were 37 winning years and 26 losing years. The best gain was +19.2% was back in 1958,and the worst loss was –27.3% in 2008 (1973 also produced a loss of more that 20%).

So, one can see that the results for the May- October period in the 19 years in which $SPX made a new all-time high in May were far superior to the totality of all years (average +5.0% gain vs. average +0.3% gain).

This might be useful, for example, if you exited your positions (or were even short the market) because of the “Sell in May and go away” philosophy. You would have exited your longs (and maybe gone short) at the close of trading on April 30th. However, if during the next month, $SPX traded at a new all-time high, then you could use this new information, to re-establish positions at the end of May, figuring that the remainder of the year would produce good results, on average.

Table 1 is a year-by-year summary, with totals below the year-by-year detail, for the 19 years in which May has seen a new all-time intraday high.

Too Much of A Good Thing?

Alan Kuenhold, a fellow member of AAPTA, found one more morsel of information related to this phenomenon of new highs being made in May. He noted that in six of the years – 1950, 1961, 1964, 1983, 1986, and 2014 – $SPX made a new all-time high in every month from January through May. Table 2 shows when those highs were made this year:

Table 2: New All-Time Highs Made in 2014
    Date       $SPX Close
    20140115   1848.38
    20140227   1854.29
    20140228   1859.45
    20140304   1873.91
    20140306   1877.03
    20140307   1878.04
    20140401   1885.52
    20140402   1890.90
    20140512   1896.65

While it is not a statistically significant subset, the returns for the five years where monthly new highs were made were a bit less than the average of all 19 years.

In summary, all of this is interesting information that augurs for higher prices ahead this year.

This article was originally published in The Option Strategist Newsletter Volume 23, No. 9&10 on May 23, 2014.  

The Option Strategist Newsletter $29 trial