Proving Grounds: Revisiting the “Best Six Months”
By J. Taylor Brown
The “Best Six Months” from November to April has been a cornerstone of Almanac Investing since our founder Yale Hirsch first uncovered the amazing pattern in the 1987 Stock Trader’s Almanac. It spawned the old saw “Sell in May and go away”. But times change and patterns shift, so as we approach the beginning of the “Worst Six Months” it is time to test the seasonality in the crucible of the Proving Grounds and see how it has performed and changed over the decades.
What is the reason or rationale behind all of this seasonality? It is difficult to pinpoint what makes many recurring patterns occur. In the past, the rationale behind this switching strategy was clear. During the quiet summer period traders and managers prefer poolside and beachfront to the trading floor, individuals take family vacations, volume dries up and prices tend to languish. In September— back-to-school, back-to-work—money managers tend to clean house in advance of the end of the third quarter and the October 31st mutual fund deadline.
Then comes November, December and January—the best three-month span. These months are accompanied by an increased inflow of cash into the market and inspired by a holiday spirit and retail spending. January involves a new leaf mentality and reallocation of assets with generally positive outlooks for the New Year. March experiences end-of-quarter window dressing and April is fueled by first quarter earnings expectations.
After the news is out and the gains have been booked, the tendency is for profits to be taken during the spring and cash is accumulated ahead of the summer doldrums. Then the cycle repeats. So there you have it, one big, juicy annual rationalization.
An Impressive Track Record
History may be on our side, but the proof is in the results. Most of you are familiar with the running hypothetical returns of the Six-Month Switching Strategy on page 48 of the 2007 Stock Trader’s Almanac. It shows an initial $10,000 invested in the Dow starting in 1950 only during Best Six Months compared to the Worst Six. At current prices the Best Six Months would have produced a compounded gain of $556,096 versus $341 during the Worst Six Months.
In all fairness, had one put $10,000 into the Dow at the close of 1949 at 200.13 and held it until now it would have gained $618,655. So, you get about the same results being invested half the time with half the risk—and avoided most major drops.
The same $10,000 invested in the NASDAQ since 1971 during the Best Eight Months (November to June) would have produced a compounded gain of $347,853 versus a loss of – $3,568 during the Worst Four Months at current prices. Putting $10,000 in the NASDAQ at the end of 1970 at 89.61 would have gained $266,923 as of April 12, 2007.
Our Best Six Months Switching Strategy remains an excellent starting place for solid long term returns. The strategy capitalizes on the stock market’s propensity to log most of its gains during the six months from November to April while avoiding the historically treacherous months.
In addition to NASDAQ’s Best Eight Months, we have also observed that in recent years the best months have also included October, creating a potential Best Nine Months. The jury is still out on whether or not this is just the effect of last century’s super-bull or a shift in seasonality. The accompanying bar charts compare the results of all three strategies for the NASDAQ and Dow.
In order to analyze the “Best Six Months” in greater detail we have broken the historical performance into four timeframes in the table on page 7. 1950-2006 represents the original dataset the strategy is based on and 1971-2006 corresponds to the inception of NASDAQ and the advent of the “Best Eight Months”. 1982-1999 represents last century’s super-bull market. During this rip-roaring period, a new paradigm was established due to increased participation in the markets and the advent of the Internet. Lastly, we have 2000-2006, a short albeit turbulent period of burst bubbles, war, political instability and unprecedented prosperity.
1950-2006
November which is the third best month for the Dow averaging a 1.7% gain and the best overall month for the S&P which has historically tacked on 1.8% over the 57-year span. December and January make up the second two months of the best three month span. February has always been a hiccup in the strategy ranking 8th for the Dow and 11th for the S&P with a net negative average gain.
Although March has been up far more often then down, the results are somewhat middling. The end of the first quarter ranks 6th on the Dow and 5th of the S&P with both averaging about a 1% advance. The seasonally bullish bias ends on a strong note in April, the Dow’s number one month.
May begins the Worst Sixth Months, barely eking out a gain averaging 0.1% and 0.2% on the Dow and S&P respectively. June is the second worst month for the Dow garnering more losers than winners.
The S&P isn’t quite as bad, ranking 9th, but it has had its fair share of losers especially as of late. July is the bright spot of the summer, but the elusive summer rally is usually squashed in August and September, historically two truly The Best Six Months kicks off in abysmal months. October, often the “bear-killer” has seen more than its fair share of disasters as well. There seems to be a shift underway where October has become a much better month than it has been in the past.
1971-2006
The Best Eight Months, like its brethren kicks off in November and has a best three month run from November to January. November has averaged a 2.2% gain and ranks second for the NASDAQ, December’s 1.9% is good for the third best month and January reigns supreme with an incredible 3.7% average gain. Much in the same as the Best Six, February is a speed bump in the bull-run. The NASDAQ is up almost as often as it is down and it averages an unimpressive 0.6% gain. March has deteriorated recently and now has a worse performance than February hovering just above the flat line with an average gain of 0.3%, the ninth best month. The second quarter is a nice run for the NASDAQ thus resulting in the Best Eight Month Strategy as opposed to the Dow and S&P Best Six. April, May and June all average a better than 1% gain and are up far more of the time than they are down.
NASDAQ’s Worst Four Months start in July. As the mercury rises, the NASDAQ falls as July, the next to worst performer, has been down more often than it has advanced averaging a -0.3% change. August, the 10th best month, and September, the worst, have a slightly better track record in terms of being up or down, but they are no prizes. August manages to be positive overall averaging a 0.3% gain, but September is just flat out awful for the tech-laden index falling almost 1% on average. October for the NASDAQ has improved immensely over the past decade and there may be a shift in the strategy.
1982-1999
During this period, the NASDAQ hit a low of 159.14 on August 13, 1982 and rose to a high of 4069.31 on the last day of 1999 good for a 2557% increase. The Dow and S&P bottomed out the day before. The Dow’s nadir was 776.92 and it surged 1480% to 11497.12 when 1999 ended. The S&P was the laggard of the troika advancing only 1435% from 102.42 to a 1999 close of 1469.25. Although the NASDAQ would move another 1000 points and the S&P 60 or so over the next two and a half months, the writing on the wall was, in retrospect, as clear as day by the middle of January when the Dow faltered. The super-bull officially ended on January 14 of 2000. With that in mind, looking at the Best Six/Best Eight over this period is revealing.
From 1982-1999, the only down month for the Dow and S&P was September. In fact June, the ninth best month for the Dow averaged almost a 1% gain. Only 5 out of 18 years saw a negative Worst Six Months during the super-bull with 1987 representing the lone disaster and 1983 and 1999 were fractional loses.
The bad stretch for both the Dow and S&P was August to September. August was the next to worst month for both indices although it was up more often than down and averaged a slight gain. October was the 10th best month and had a similar track record to August, up 11 and down 7 times. It averaged a 0.6%. May, June and July all had very good records.
The Best Six Months over this period were gangbusters. More important than the average percent changes are the number up compared to number down. For the Dow, the best three-month run saw November up 13 and down 5, December up 14 and down 4 and January also up 14 and down 4. February, historically a contentious month was up twice as often as it was down, as was March. Ironically, April, the second best performer from a percent change standpoint, was down quite often. S&P sported a similar record. The NASDAQ actually adhered to the game plan better than the S&P and Dow.
The NASDAQ saw seven significant declines over the Worst Four Months from 1982-1999 with only 2 loses during the Best Eight. July, August and October were relatively bad months with July and October net negative. September, that perpetually perplexing period, was actually pretty good for the NASDAQ averaging a 1% gain. The Best Eight was amazing with November, December and January off the charts.
2000-2006
Since the top in 2000, the markets have apparently undergone a sea change. With the worst six on the brink of outperforming the best six in 3 of the past 7 years (Dow needs to be at 12839.12 on 4/30 to beat worst six), we are taking this opportunity to evaluate and reassess the switching strategy. While seven years is a not a significant dataset to alter a triedand- true approach such as the Best Six Months, to ignore it would be foolish.
Since 2000, the Street has been powered by the fourth quarter. For the Dow, October, November and December are 1, 2 and 3 in terms of rank and sport impressive returns. The S&P has a similar record. Interestingly, December has been terrible for the NASDAQ. The NASDAQ data must be taken with a grain of salt as the trials and tribulations of this index from 2000 to 2003 make drawing a meaningful conclusion difficult.
January and February have bucked their historical trend becoming somewhat of a trap with several miserable performances, but we discount this as exogenous influences have held sway. We expect Q1 to once again rise to prominence.
Conclusion
What we can deduce from the data is that the Worst Months seem to be starting later and ending earlier. Moreover, September looks like the time to get back into the market. The entire market seems to have fallen in-line and all three major U.S. equity benchmarks are currently exhibiting a Best Eight/Worst Four tendency.
When you sit back and think about that, it all makes sense. The market goes on vacation after Memorial Day and returns on Labor Day. Also, the Street has become a bunch of bonus babies. Everyone wants to hit their target for the year and thus the consistent strength in Q4.
We are not abandoning the Best Six Month Strategy. We will complete this Dow and S&P Best Six Months cycle for the record on April 30. Almanac Investors are currently looking for a MACD sell signal as of April 1. Over the summer, while we are out of the market we will consider making any changes to the strategy.
For more information, please visit StockTradersAlmanac.com
AI
Market Data 
Markets 
