Monday, May 7, 2012

May 7, 2012, Bob Brinker's Moneytalk Summary Part Two

May 7, 2012..................................................................(comments welcome)

Yesterday, Bob Brinker commented on the subject of "seasonality" in the stock market. He called it "fiction" (read his comments in my show summary part one).  David Korn covered this subject in his weekly newsletter that includes a summary of Bob Brinker's Moneytalk. David has generously  given me permission to share this with you: 

David Korn wrote:

A SIX MONTH STOCK MARKET TIMING STRATEGY - A DAVID KORN PRIMER

************************************************************

"Those who study market history are bound to profit from it."

- Jeffrey A. Hirsch, Co-Editor of Stock Trader's Almanac

"In the stock market, those who expect history to repeat itself exactly are
doomed to failure."

- Yale Hirsch, Co-Editor of Stock Trader's Almanac

************************************************************

Introduction:  May marks the beginning of what some investors view as the
seasonally weak part of the year for stock market returns.  I have updated
my research on this market phenomenon and include it below as a David Korn
primer on the subject.  Charter subscribers will be familiar with the
tenants, but there is always something new to consider and I learn a new
fact every time I re-write this piece.  There are a couple of newsletters
and advisors that make a living off of this particular strategy, so they may
have an interest in the data.  Not me.  I am my own show and just try to
tell it like I see it.

Much of the data was originally collected by Yale Hirsch, editor of the
Stock Trader's Almanac, but the phenomenon actually has much older roots.
Like most David K. Primers, I like to utilize a Q&A format in an effort to
anticipate the questions you might have about the strategy.

Question:  David, what the heck are you talking about?

Answer:  I will forgive your informality if you will forgive mine.
Essentially, there is a fascinating historical pattern showing that most of
the stock market's gains occur during the time frame that begins November
1st and ends April 30th of each year.  Conversely, the time period between
May 1st and October 31st have generally not produced favorable returns in
the equity markets.

Question:  Isn't this referred to by another name?

Answer:  Not just one other name, several.  The first I ever heard of it,
was a reference to the "Seasons in the Sun" strategy, presumably because it
refers to the fact that you rotate out of stocks during the summer months
when the sun is shining bright.  Another popular phrase for this strategy is
the "Halloween Indicator" for the obvious reason that the strategy turns on
Halloween each year, at least as celebrated in the United States.  The Stock
Trader's Almanac refers to it as the "Best Six Month's Strategy" and Sy
Harding, another well-known follower of its tenants, calls it the "Seasonal
Timing Strategy." 

Some use an idiom to remind them what to do  -- "Sell in May and Go Away."
Indeed, investors have coupled the "Sell in May and Go Away" phrase with
"but buy back on St. Leger Day."  St. Leger Day refers to the date of a
classic horse race that has been run every September since 1776.   Even our
northern neighbors in Canada have a saying for this phenomenon -- "Buy when
it snows, sell when it goes."

Question:  Do you have any statistics to show if this strategy works?

Answer:  I knew would ask that -- probably because I wrote the question.
According to the Stock Trader's Almanac, if you invested in the stock market
from November through April, and switched to fixed-income investments from
May to October, over a 62-year period (1950-2009), your returns would have
been quite dramatic.  Using this strategy, a $10,000 investment in the Dow
Jones Industrial Average produced a compounded annual return of $619,071
gain in the November - April period, versus a loss of $9,621 for the
May-October period.   That is pretty amazing.

Source, Stock Trader's Almanac.  Yale Hirsch & Jeffrey A. Hirsch, Editors.

http://www.StockTradersAlmanac.com/

A little less biased source, Standard & Poor¹s Equity Research did research
and found that the market has returned 1.3% on average during the
April-September time frame versus the October ­ March time frame when the
U.S. stock market has a 6.7% average return.  Note that their time frame
modifies the sell and buy points by about a month.

Question:  David, has anyone looked at the data for the United States market
beyond the 54-year period referenced by the Hirsch Organization?

Answer:  Yes.  In 1991, Michael O'Higgins looked at data from 1925 through
1989 and observed that 85% of the capital gains, excluding dividends, was
earned in the Dow during the October 31st through April 30th time frame.
Higgins came up with some inventive strategies to optimize it by using
various proxies, such as the five highest yielding Dow stocks.  Mr.
O'Higgins' data was included in his book, "Beating the Dow" which was
updated in 2000. To learn more about it, this url brings you to the book's
profile on Amazon.com:

http://tinyurl.com/b6ssg

Question:  But David, isn't this risky business?

Answer:  I'll answer this as any good Hollywood Scientologist might.  Any
type of market timing involves risk of under performing a fully invested
position.  This particular strategy, however, involves taking half of the
normal market risk since you are only invested in the stock market six
months of the year.  That said, recent years have been a perfect example of
how this strategy can be fantastic but also make you miss out on the good
stuff. 

Recall that the ultimate high of the market ever was October 9, 2007.  That
was just a few weeks before the ³best six months² started.  On the other
hand, if you got out of the market in May 2008 and stayed out through the
end of October 2008, you missed out on a lot of carnage during the
May-November 2008 timeframe.  But then the May-October 2009 was a fantastic
period for the market.  You see how this works huh?

Question:  Has this seasonal strategy produced similar results in other
countries? 

Answer:  It actually has. In fact, there is a study that examined monthly
returns for the U.K. stock market back to 1693 ‹ that¹s 317 years of data!.
Turns out this strategy beat the market 71% of the time over all two year
periods and 82% over five year periods.  Pretty impressive.  Here is a link
to the article entitled, ³Are Monthly Seasons Real? A Three Century
Perspective²

http://tinyurl.com/26soyog

Question:  Any other serious academic research done into this phenomenon?

Answer:  In one fascinating academic study I found on this issue, the
research showed that stock market returns were higher in the November -
April period versus the May - October period in 36 of 37 countries studied.
Did you hear me right?  Or perhaps I should say, did you read me right?  In
36 of 37 countries!  This study concluded that the effect seemed to be
strongly related to the length and timing of vacations, especially in
Europe.  This study was written by Ben Jacobsen from the Rotterdam School of
Management, and AEGON Asset Management and is entitled, "The Halloween
Indicator, 'Sell in May and Go Away': Another Puzzle."  This link will bring
you to an abstract of the article:

http://tinyurl.com/6e4tl

It does bear noting that this study was based on data beginning in 1970
because that is when much of the data for worldwide stock markets became
readily accessible and more importantly reliable.

Question:  David, what do you think could be causing this phenomenon?

Answer:  I have seen many explanations for the phenomenon.  Here are some of
the reasons that are offered:

1)  As noted in the study I previously cited, during the summer months which
occupy the worst 6 months for investing, everyone is on vacation, hence the
lack of "buying" interest in the market.

2) Investors tend to come into extra cash in the time frame between November
1st and April 30th due to year-end bonuses and distributions, income tax
refunds, company contributions to retirements and profit sharing plans.  Of
course, this wouldn't explain why the phenomenon seems to work in other
countries.

3) The Infernal Revenue Code encourages retirement contributions between
January 1st and April 30th.

4) Mutual fund inflows tend to be less during the summer months.  As
reported on theStreetcom, since 1984, 62% of all mutual fund inflows have
occurred between November and April.  In the secular bear market that
occurred during 1966 to 1982, monies invested solely in the May though
October time frame fell nearly 85% on an inflation adjusted basis:

http://tinyurl.com/apq6

5)  Several entities have suggested that seasonal factors that influence
economic data might help explain the seasonal cycles in stocks.  Here is a
link to one such article entitled, aptly, ³Seasonality²:

http://tinyurl.com/nhapbr

6)  The Hirsch organization suggests this seasonality is based on the
"habitual behavior of society, which extends to stocks."  With respect to
the poor performance during the May-November time frame, Jeffrey Hirsch
notes that the second quarter tends to be weaker as the positive effects of
holiday bonuses and the holiday retail sales period fade out, and a "spring
cleaning mentality kicks in."

7)  A final explanation offered is a seemingly obvious one.  This is a
pattern that is widely recognized, and people have decided to follow it.
There have been many publications (including yours truly) that regularly
report on this strategy.

Question:  David, has anyone ever used a combination of technical analysis
and market timing to improve on this strategy?

Answer:  Funny you should ask, I was just preparing a response to that very
question.  It appears that editor and subscriber have reached a symbiotic
relationship.

Sy Harding, editor of Street Smart Report, and author of the book, Riding
the Bear, uses the Moving Average Convergence/Divergence indicator (MACD) to
try and enhance the Six Months Strategy by picking a better entry and exit
point, instead of simply relying on the rigid May 1st and October 31st buy
and sell dates.  Using the MACD, the sell signal could come a day or even
several weeks before or after May 1st, and the buy signal could come in the
days or weeks before or after October 31st.  Mr. Harding's system has been
so successful, that he has calls it, the "Best Mechanical System Ever."
Mr. Harding calls his strategy the Seasonal Timing Strategy or "STS
Portfolio" and he uses the DJIA Index for the model's measurement.  Over the
last 13 years, through the end of December, 2011, Sy claims that his STS
Portfolio has produced compound returns of 190.6% versus 64.4% gain for the
S&P 500 and a 18.8% gain for the Nasdaq.  Sy Harding has a fairly in depth
discussion of this system on his web site which you can access at the
following link:

http://www.streetsmartreport.com/sts.html

The Hirsch Organization over at the Stock Trader's Almanac are well aware
that Sy Harding is marketing the Best Six Months strategy using the MACD
indicator.  In fact, they even reference his use of it on their web site at
this link: 

http://tinyurl.com/yzepqs

Since they know a good marketing tool when they see it, STA now offers their
own market timing newsletter based on their version of the MACD indicator
using the Six Month Strategy.  So, we have competing newsletter writers, all
of whom are desperate to have the better timing on the Best Six Months.  Now
we have some drama and competition.

Question:  David, has the application of the MACD enhanced the Best Six
Months strategy over the long haul?

Answer:  The Hirsch Organization back-tested this method from 1950-2010.
According to their research, applying the MACD signal produced what can only
be characterized as astounding results.  Check it out at this url:

http://tinyurl.com/yfqb8ep

Question:  David, whose system is better, Sy Harding or the Hirsch
Organization?

Answer:  Mark Hulbert actually took up that question and concluded that Sy
Harding's system is superior due to his better timing of the exit and entry
points.  Read his article entitled, "Selling in May and not going away?" at
this url:

http://tinyurl.com/6nr853

Question:  Does anyone find fault in the best six months timing strategy?

Answer:  Yes.  The Schwab Center for Investment Research (SCIR) looked at
this strategy and found that the S&P 500's average MONTHLY return was higher
(1.2%) for the November to April period, and lower (0.8%) for the May to
October period.  Moreover, SCIR concluded that simply remaining fully
invested in the S&P 500 outperformed a seasonal portfolio more than half of
the time.  SCIR concluded that "time in the market" is much more important
than "timing the market" especially when you consider the transaction costs
and tax consequences.  The article used to be available online, but it is
gone now.  And you heard Bob Brinker this weekend on Moneytalk say that the
system is nonsense.  There are definitely skeptics out there.

Question:  If you follow this strategy, and sell your stocks at the
beginning of May, is there any place you should put it other than cash?

EC:  The original advice was cash or Treasuries.  However, there are many
things you can do.  If you really thought the stock market was going down,
you could purchase put options on the market or buy a contra-fund ­ both
very risky if the market moves up.  Here is a good article just recently
published entitled, ³Sell in May and Do This Instead²:

http://tinyurl.com/7tqxabj

Question: How has this strategy done lately?

Answer:  It¹s performed great recently.  Selling in May last year was almost
perfect timing as April 30th marked a high and the market plunged for almost
five months.  Even after huge 10% move up in October, the Dow lost about 7%
form May-November and has gained over 10% since then when the calendar
switched over to a buy signal.  2010 was also a great year for the system.
Stocks gained only 1% from May to November but then took off in the six
months that followed.  The Dow gained 15.2% from the end of October 2010
through the end of April, and then while safely in cash from April 30th on
-- well, you saw how the market performed.  Of course, past performance is
not a guarantee of future results.

Question:  David, what do YOU think about this strategy?

Answer and Conclusion:  Frankly, I believe that there is some merit to this
phenomenon.  Of course, it could just be a statistical anomaly, but if it
is, that's one heck of an anomaly.  There is also a bit of trying to
outguess each other as to when the trigger points to enter/exit the market
will take place.  You should also keep in mind that this timing strategy is
actually pretty conservative on a risk-adjusted basis given that you are out
of the market for half the year.   Finally, like all timing systems, there
is no "Holy Grail" when it comes to investing.  I firmly believe that you
should never rely on one single factor when analyzing any stock, timing
system, or investment strategy in general.  Nevertheless, I also think you
can never have too many tools in your investment arsenal.  All in all, I
think the Best Six Months strategy is worthy of being aware of, and warrants
consideration if you are adopting a timing approach to investing.

David Korn's Stock Market Commentary, Interpretation of Moneytalk (Bob
Brinker Host), Financial Education, Helpful Links, Guest Editorials, and
Special Alert E-Mail Service.  Copyright David Korn, L.L.C. 2012


Honey here:  You can request complimentary issues of David's weekly newsletter and The Retirement Advisor published by Kirk Lindstrom and David Korn.

Bob Brinker was late arriving at the studio yesterday so Jeffchristie has a new Moneytalk Final Exam question: 
 * Bob Brinker was over an hour late for work Sunday.  Some people have suggested that his car broke down on the way to the studio.  Bob  drives:

A) AMC Gremlin

B) Prius

C) Edsel

D) Ford F-150 complete with gun rack

8 comments:

Anonymous said...

Brinker started pumping gold at the top after bashing it the whole way up.

Brinker got out of bonds at the highs saying yields were going higher.

Brinker pounded the table for more nuclear reactors, Japan is radiating the world.

Man if this guy didn't have bad luck, he would not have any luck at all.

Joey

Jim said...

A very thorough analysis by David Korn. When Brinker made the comment that it was fiction, my first thought was: "How can you say that Bob?" While I have never sold because of seasonality, I've always believed that it did indeed exist.

I guess Brinker cannot accept the fact that such a simple system could be better than his own.
The best part is that someone who wants to time the market does not have to pay $185.

Anonymous said...

Frankj --

The seasonal strategy is a market-timing strategy. BB uses his own model (presumably.) So he is unlikely to acknowledge that a system requiring only a calendar could give good results.

Thank you to HB and David Korn for the links.

birdbrain said...

Reading Honey's summary, Mr B criticized the Sell In May approach "the seasonality that the market goes down from May to Oct every year" and "going back six decades 60% of the May-Oct periods the market is either up or even."

No one has ever claimed the S&P 500 falls every year during those six months, and it may be true that more often than not the index is higher. The point is history shows that Nov-Apr has significantly outperformed the other six months. From Stock Trader's Almanac
since 1972 annual average return:

Nov-Apr 7.4%
May-Oct 0.4%

Facts are facts, Bob Sr.

Anonymous said...

...a system requiring only a calendar could give good results...

Oh come on people. Get Real!

Don't you think that every "system" that only required a calendar to give good results would be practiced by every person out there and his dog?

You all talk about "average" returns for the best six months versus "average" returns for the worst six months which mean nothing.

You can drown in a lake whose average depth is only 3 feet. Figure it out.

The Seasons strategy just doesn't work any better than any other market timing scheme.

Noseasons

Anonymous said...

I am really not into explaining why you should not count fairies dancing on pinheads. But sufficient to say if you randomize your holding period you can find period of time where the rule is accurate and find periods where it is not.

Let's take a peak at a significant period of time: 1929 - 2009(I chose this because I have the numbers):

Monthly average stock returns:

Jan 01.49%
Feb 00.06%
Mar 00.72%
Apr 01.67%
May 00.30%
June 01.19%
July 01.83%
Aug 01.30%
Sept(00.75) - yep it is negative
Oct 00.43%
Nov 01.07%
Dec 01.74%

So now...armed with the data above you/we can have all sorts of fun.

Notice July has the highest average increase...in fact June, July, August have pretty nice returns...now where were you going to put that money when you sell in May? Do you have a cash investment that yields over 1.19% a month? Can it compensate for the loss of 1.83%?

Moreover if you add those months up you get 4.51%...guess what, that is more than any other 3 month consecutive period.

Actually what I just wrote is nonsense, though it probably sounded good because it is true. It is nonsense because it is simply based on a averages and amounts to statistical noise.

Take a close look at Sept and put on your thinking hat, it is negative simply because of 1931 and 1937, start your index at 1938 and Sept looks pretty darn good! So maybe a better indicator would be is we weighted the more recent time periods, or threw out he highs and lows outside the range of the norm (all statistically valid approaches btw)

And that is the point, humans evolved (cool tfb finely gets to use his cool anthro background) because we have one huge advantage over other life forms, we excel at pattern recognition. This was a huge evolutionary advantage.

Stepping aside that also means we have this annoying tendency to assign significance to patterns that are essentially random. In other words the pattern may exist but it may exist merely through happenstance.

You can play around with this all you want, start with the numbers I listed and create all sort of rolling periods and see what rules you can create. then if you have the time alter the duration of the recording period for creating the monthly averages. eventually you should realize all you are doing is a crude form of regression analysis, i.e finding a line of best fit. If regression analysis was a valid implementation method for divining the future pensions and endowments would utilize it(where the real money is)!

Lastly consider the frequent friendly sparring point between heavyweights John Bogle and Larry Swedroe. Essentially they have a disagreement,a nd that is over whether "value" stocks really provide a superior return. Larry would say look at the evidence and Bogle would say the holding period you examine it under is not long enough, and it is simply a statistical anomaly.

the point above is, even experts at analysis can disagree over the difference between statistically significant and statistical noise.

Live Large...

tfb

Anonymous said...

OMG - I just sided with Brinker!!!!

tfb

Anonymous said...

Noseasons: Calm down. I didn't say it worked. I gave a reason why BB might throw cold water on it.