Hi Guys,
CXO Advisory Group just released a brief study that examines the efficacy of the "Sell in May" rule that currently enjoys great popularity. The reason for that popularity is obvious: It has worked in recent decades.
CXO has asked a more challenging question. How has that rule performed over a very long timeframe? That firm explored the issue by using the extensive database generated by Professor Robert Shiller. CXO discovered some significant departures from that golden rule buried in the 1930 to 1960 equity return decades. Here is a Link to the study:
http://www.cxoadvisory.com/3873/calendar-effects/sell-in-may-over-the-long-run/CXO provided the following summary statement: "(E)vidence indicates that the conventional wisdom to “Sell in May” (or conversely to “Buy in November”) works well for U.S. stocks on average since the 1950s, but does not consistently work well before then."
When reviewing statistical studies, one constant caution is that selected timeframes are very important, and can be crafted to support or destroy any hypothesis. So always be skeptical. Trends work until they stop working, and forecasting that tipping point is beyond the scope of the common man.
Enjoy the article.
Best Regards.
Comments
Here we see a seasonal strategy that has worked well overall
for some 60 years. But you seem to be saying that data period is suspect
because it is not long enough.
What is your personal opinion of its current value?
Hi Flack,
Thank you for reading and taking time from your busy schedule to respond to my brief posting.
I posted the CXO study simply to inform and/or update members of the MFO board. Sell in May studies have been reported for decades and are included in endless equity investment courses, speculatively even the one that you teach at the community college level. I appreciate that it is well documented, but the CXO study claims to be one of the few (perhaps the only one) that incorporates all the historical database. In general, CXO does good work.
My comments relative to statistical data were generic in both character and intent. I take no issue with the entire data set, or the reported findings for the last six decades. The data is the data if it is honestly and accurately reported.
My cautionary warning was directed at some unscrupulous authors who present carefully selected data sets by design to bolster their particular viewpoint while purposely ignoring conflicting data. That action is fundamentally dishonest and does a disservice to readers while simultaneously discrediting the reporter himself. Shame on him.
I have been familiar with “Sell in May” studies for decades. I have never found their results compelling and have never followed the schedule advocated by these numerous studies. That’s just me. If your fast and slow brain components dictate a contrary conclusion for you, by all means just do it. I have never, and never will, recommend that any investor follow my decisions. Investment decisions are ubiquitously a personal and often a unique matter.
Best wishes for your continuous investment success.
Perhaps MJG targeted this somewhat towards me as I have been one that has used the seasonal strategy for a good number of years and posted and referenced to it frquently. It is also one that my late father used as far back into the 50's. For sure, it does not work every year ... but, it has worked more times than not. Over the past three years I have made more than enough to live off of from just investing alone ... not, counting my earned income. The subject strategy was a contributor to these returns. I agree it is not for everyone ... but, it works for me and I plan to continue to utilize it. Is it one I'd bet the farm on? Absouetly not. But, it is one that I plan to continue to put a certain percentage of my portfolio into each year ... and, by following the technicals that Flack teaches investors ... If it gets too tough and has moves against you, simply stated, just exit it. Save your capital to invest for another day.
I wish you both the very best with your investment endeavors ... and, thank you both for your contributions. From my perspective, we all learn from one another.
Have a Good Day and Good Investing,
Skeeter
A little help please.
I understand the CXO examples to be saying that an investor following the
Sell in May/Buy in November strategy would do so every year – regardless
of the market condition.
In other words, the investor would ignore Bull and Bear conditions (by their
definition: Bull market being above a 1-year Simple Moving Average and
Bear market below that SMA) and continue to exit in May and buy in November
every year.
Yes, I’m ignoring their data using the 5-year SMA. I can’t see where this is significant.
So, am I reading this right?
CXO approaches these questions from an academic/research standpoint and
not from the investor or trader standpoint. That is, they answer the question but
rarely crank up the microscope more than one notch.
OK, so CXO is Dr. Jekyll and I suppose that makes me Mr. Hyde.
An investor seeking to optimize returns and minimize losses would understand
(by looking at the historical data) that the easiest path would be to remain within
the larger market trend.
So, assuming that we’re in a bull market condition, the investor would not sell in May.
He would not buy in November in a bear market.
Instead, he follows the major trend.
This means selling in May only in bear markets and buying in November in bull markets.
Granted, this is not a perfect system.
But it is in sync with the market and should reduce risk. It may in fact increase returns.
What’s more, it seems entirely logical.
But I’m not the mathematician in this group (or any group!), I’m just looking at the charts
and this looks to be the case. If I find the time, I may try to do the math.
I did say TRY.
Another thing that bugs me is that CXO used a 1-year SMA. I didn’t catch their rational.
I do show my classes this market seasonality but only as something that I think they should
be aware of and consider as they observe the market. I don’t consider seasonality as
a stand alone strategy but rather as an adjunct to a broader strategy of timing.
Thanks for the interesting discussion.
And if you’re reading this – Hi, Skeeter.
Hi Flack,
Thanks for your reply. However, I’m not sure I want to thank you for your numerous questions, all pertinent, but some of which require considerable effort for a full response.
I have not convinced myself that it is worthwhile or wise to commit to that effort, so I’ve compromised and prepared an abridged, and likely incomplete, reply. At my age, I’m forever deviled by the specter of “muda”. A few months ago I posted comments on something called the “muda trap”. The “muda trap” is a popular Japanese idiom that translates into English as a waste of time and energy. Everyone should take cautionary steps to avoid this resource depleting trap, especially me. I’m sure you understand.
First, congratulations on introducing your investment students to the Sell in May seasonality indicator. I assume that you also incorporate many of the other seasonality indicators that CXO Advisory Group documents in your classroom lectures. Good for you; good for your students.
You are exactly correct in your interpretation of the CXO Sell in May study. CXO clearly identifies that the returns are based upon a single decision criterion: that criterion is the calendar month. CXO does not claim that this simple criterion generates optimum rewards nor does it minimize downside losses (these are separate and distinctive goals and require careful specification in any mathematical optimization program).
You apparently impose a second criterion in your personal portfolio returns optimization efforts; you insert a market momentum component into the decision matrix to enhance returns. For the purposes of easily reporting their study findings, CXO did not consider a more complex multiple decision tree rule set.
Will multiple criteria enhance performance results? In many instances the answer is in the affirmative since financial markets are complex and governed by many moving parts. However, because of that complexity, sometimes adding additional constraints can do damage. So care must be vigilantly exercised, and the added constraints must be fundamentally sound from a market modeling perspective.
As constraints are added, the prospect of data mining becomes a more relevant issue. At the two-criteria level I would not consider this concern too pressing an issue.
CXO does good work. They carefully define and identify their study purposes, their assumptions, and their shortcomings. The CXO studies do have an academic quality. Perhaps that is why I often reference them. They also document their work carefully and end each article with cautionary warnings about the limitations of their mostly statistical studies.
The limitations of statistically-based studies must always be fully documented and recognized. CXO can only do the first part; the reader must be cognizant of the second part of that pact and interpret findings accordingly. User beware.
Best Wishes.
I’m sorry that I asked so many questions. I’ll attempt to do less of that in the future.
I also have to apologize for the confusion I caused you.
After reading the seasonality report you provided, I moved on the next link following
that report. In this link they mentioned the addition of a second criterion – the identification
of bull and bear markets using moving averages.
I should have pointed out this second report entitled “Sell in May” During Bull and Bear Conditions.
You’re a good sport for putting up with me.
Best to you,
Flack