"Sell in May and go away." You hear it all the time, year after year as we get closer to the summer. Turn on CNBC and you hear the hosts (otherwise known as talking heads) constantly mentioning "Sell in May and go away" - like it's been pre-programmed into each one of their heads. "Don't worry about the facts, just tell the story." Or go to a financial website and you can't help but find an article talking about the same strategy. But the thing lacking in the arguments are facts. Where and what are they?
Some of you may remember Sergeant Joe Friday of Dragnet, whose favorite line in interviewing somebody was "Just the facts, ma'am." So, I decided to pull historical data from 1988 to present (6 previous election cycles) and look at monthly returns for the S&P 500 to try to determine if there has indeed been a trading pattern from which we can learn and profit (or avoid loss).
But before I give you my findings, a few caveats. This is Historical Data (said very slowly). History is not always representative of the future. I have based my findings on the S&P 500 as a representation of the broad equity market. I'm using averages and CAGRs; the market has a huge standard deviation from those averages so the swings can be very material. In other words, these are facts and probabilities. That's all. There is no guarantee that the future will emulate the past.
OK so here we go. First, my conclusion: You might not really want to sell in May. But June, now that's a different story. (So I guess technically if you sold on the last day of May, you are selling in May.)
I have cut the data a number of different ways and I am going to give you the highlights. For me, my baseline was - was the index up or down in a particular month? If it was up, it's a good month. Here are some interesting tidbits from my findings before I give you the numbers:
- From 1988 thru 2011 (288 months) the market was down 103 months and up 185 months. So, the market was up 64% of the time in general.
- The CAGR during this period was 9.45%, about equal to the long-term annual market gain for the last 100 years. Standard deviation was 18.6%. So the computed range is -9% to +28%, about typical. (Note: CAGR = annualized return)
- During this period, four months produced a negative return over 24 years - February, June, August & September
- Not surprisingly, the best 3 months to invest are Oct thru December (5% CAGR). Maybe surprisingly, the other best 3 month stretch is March thru May (4.8% CAGR).
- The market was down the most times in 24 years during February (10), June(11), July(12), Aug(10), Sept(12
- December has been the best month in terms of probability the month would be up. It was only down 3 times in 24 years. During April (6) and May (7) the market was down the next least number of times.
- During election years, I really see no discernible pattern no matter how I slice it. So regardless of what all the strategists say, There is no election year pattern -- other than political angst.
- How about this obscure fact? - There was only 1 year in the last 24 where each month May thru September was down. And yes, 2011 was it. There were 2 other years where the market was down June-Sept - 1990 & 2001.
OK, enough with the tidbits. Let's go to the meat of my findings and my conclusions.
Yes, there clearly is a pattern and the numbers favor sitting out the months of June through September. However, there is a funny quirk in the numbers. Even though September has historically had a 50% chance of being a down month, in years where August is down, September has often been up.
If you invested January thru May, sold, then invested October through December, the CAGR for the last 24 years is 10.40% compared to a CAGR of 9.45% for investing the whole year. Factoring in commission costs and timing issues, it would be about a wash or slightly better.
The CAGR for June thru September has been about -0.83%. So again, unless you could sell and invest the proceeds in a riskless, income-producing security, the numbers say it is close to a wash after commissions.
During the 7 periods from June thru September where the market was down, the average drop was 13%! So when the market does drop in the summer, it really drops hard. The average annual gain during the June thru September time frames for the 17 up periods was 5%.
You definitely want to be in the market for the final quarter of the year, as historically the market was up 88% of the time.
So boiling it all down, I conclude that if you want peace of mind, you should sell at the end of May and not come back until the end of September. Interestingly, the historical probability actually favors the fact that the market will be up during that period (71% of the time). However, if the market goes down, it really gets hit hard. And this, I believe, is why there is so much talk of sitting out the summer. Especially considering that in the most recent history (2010 & 2011), the market had tough summers. If you in fact are disciplined enough to do this year-in and year-out, over a long period, odds seem to favor that it will produce equal or better returns than staying fully invested through the whole year. And perhaps give you peace of mind.
Now there are other ways to maintain exposure and potentially cut your risk during the summer, but I won't get into that in this article. Following is a summary of some data broken down for you to analyze. I'd be happy to go into more detail with anybody that wants to contact me.
So next time you hear the talking heads asking "Do we sell in May and go away?" You can now say you know "just the facts."
Historical returns for the S&P 500
For the period 1988-2011
CAGR Jan 0.39% Feb -0.08% Mar 1.21% Apr 1.99% May 1.54% Jun -0.42% Jul 0.90% Aug -0.92% Sept -0.39% Oct 1.37% Nov 1.45% Dec 2.08% Total 9.45% Jan - May 5.13% March-May 4.80% Oct - Dec 4.99% May - Sept 0.69% June - Aug -0.45% June - Sept -0.83%
Source: ycharts.com monthly S&P 500 Returns.
Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I am net long the equity market and manage money for clients who may be net long.