Related Resources
Build
a screen of real estate stocks with our Stock
Screener
What
stocks are hot or not?
What
are the top-performing
industries?
The Speculator
Recent articles:
• Two strikes
and we're still optimistic, 5/9/2002
• Why the trend
is not your friend, 5/2/2002
• Don't settle
for 1,500,000% profit, 4/25/2002
More...
|
|
| sponsored by: |
 | |
The Speculator
Market momentum? Don't bet on it
If we look at what the
market month for clues to where it's going next month, we find
precious little evidence to support the myth of momentum. Your best
play: Big drops in the Nasdaq.
By
Victor
Niederhoffer and Laurel Kenner
The great forces that explain the trends,
reversals, leads, lags, rotations and signals for individual stocks
and the market are changeable, complex, ephemeral, intertwined and
seemingly fathomless. It’s rather natural to wish for a soothsayer
to interpret which way the momentum is going in instances like
these:
- The Nasdaq Composite ($COMPX)
was down 8.5% in April, but was up 7.4% in the three days ending
at Tuesday's close. Are such momentous swings auguries of good or
evil to come in the S&P 500 ($INX)?
- The Nasdaq fell 72% from its March 10, 2000, close of 5,048 to
1,423 on Sept. 21, 2001. Could investors have escaped this reign
of terror by paying heed to early signals from downward-sloping
moving averages?
- The dividend yield of the S&P 500 has fallen to a mere
1.4%, down from the 2% level that triggered Federal Reserve
Chairman Alan Greenspan’s legendary “irrational exuberance” speech
on Dec. 5, 1996, at the American Enterprise Institute. Is the
yield even more bearish now?
Violent downward moves in the
market are a source of fear and anxiety. The ability to predict
moves up would guarantee great wealth and power. It is therefore not
surprising that a mythology of momentum should be constructed,
replete with stories of great heroes and heroines who have conquered
the mighty forces, to ease our anxieties and fuel our hopes. We’ll
try in this column to bring some science to the subject of momentum,
starting with a survey of the landscape of belief.
A great
deal of scholarly and popular research has been undertaken to help
us understand the what, why, where, and how of the traditional
stories that capture the worldview of pre-scientific societies. We
found Geoffrey Kirk’s definitive book, “Myth: Its Meaning and
Functions in Ancient and Other Cultures,” a good source of insights
as we examined the role of mythology in the market. Kirk describes
cultural, psychological, anthropological, structural and linguistic
approaches to unraveling myths in a commonsense way that lends
itself to generalizations about other fields, like
markets.
The market as
myth
Our view is that stories and studies about market
forces are often myths comparable to those about the task of
Sisyphus, the labors of Hercules or the quest for the Golden Fleece.
Unraveling the influences of myths on markets is as
frustrating and difficult as the task that Sisyphus faced in trying
to roll his rock up a hill in the underworld. Just when it seems
that an approach or study might have validity, the landscape changes
and new studies have to be undertaken to get to the top
again.
Crucial to any consideration of myths is their purpose
in maintaining the social system. Myths about market momentum tend
to serve the interests of those who believe they can unravel the
future, as well as those who can profit from excessive trading by
investors.
The field of market momentum mythology is so rich
with examples of misremembered facts, pre-scientific explanations
and highly embellished stories of great deeds that we can only
scratch the surface in this article. But we feel the approach is
illuminating and hope our own variations on this theme will inspire
the readers' reflections and contribution for our mutual education
and profit.
Greenspan as Greek
hero
Any study of the mythology of stock market momentum
must begin with the story of Greenspan’s famous “irrational
exuberance” speech, in which he presented the idea that evaluating
stock price levels must be an integral part of monetary policy. Who
but a hero of Greenspan’s mythological stature could be capable of
such wise evaluation?
Fittingly, the magic phrase had been
given him the day before by a high priest of economics, Yale
University’s chronically bearish Robert Shiller. The professor had
presented a study to Greenspan purporting to show that high P/Es are
bearish. The study was based on a 10-year moving average that left
out the entire decade of the 1990s, the greatest bull market in
history.
We deflated that myth behind the Shiller study in a
previous article, "Fear, greed
and other reasons to ignore P/Es." A recent paper by Bjorn
Tuypens, also of the economics department of Yale University, titled
“Stock Market Predictability: A Myth Unveiled,” backed up our
conclusions:
"The use of overlapping data for long-term stock returns,
combined with the non-stationary behavior of the dividend yield,
gives rise to spurious regression problems when regressing
long-term stock returns on dividend yields. . . . This regression
(dividend yield vs. subsequent one-year and two-year returns)
provides additional evidence for the argument that the strong
long-run forecasting power of the Campbell-Shiller regressions is
a myth, due to the use of overlapping data in too small a data
set." Greenspan, however, found in the study a justification for
government intervention in the stock market on a Titanic
scale
We recently had the pleasure of interviewing Dr.
Shiller. He was quite humble about the resilience and robustness of
his results. He has since moved on to other fields. However, he
points with pride to the timing of a lunch he shared with Greenspan
at the Fed on Dec. 4, 1996. He had just finished presenting his
overlapping yield and return studies. Seated next to the chairman,
he asked him when the last time was that a Fed chairman had
questioned whether the stock market was too high. The answer was the
Arthur Burns era in the 1970s. The next day, Greenspan made his
now-legendary speech.
A further attempt to study momentum
was recently made by two professors at Southern Methodist
University, Venkat Eleswarapu and Marc Reinganum, in a paper titled,
“The Predictability of Aggregate Stock Market Returns: Evidence
Based on Glamour Stocks.” They conclude that the returns of the
market are negatively correlated with the returns of glamour stocks
in the previous 36 months. They find no evidence that value stocks
predict future returns.
Unfortunately, the study uses data
from 1951 to 1997. Assume for a moment that their conclusions are
not riddled with the same defects as Shiller’s. The study still
doesn’t take into account the regime shift that has taken place in
glamour stocks since 1997. The 72%, 19-month decline in the Nasdaq
documented above created new dragons requiring new powers of
divination in those who would confront them.
Moreover, the
professors’ selective use of a group of companies for which
continuous earnings data were available on S&P Compusat may well
have resulted in the omission of some beasts pertinent to the
results.
What we
asked
Under the circumstances, we took out our swords with
the intention of slaying some of these dragons. Our swords, combined
with pencil and envelope, started out with a simple quest. Do
monthly changes in the Nasdaq tend to continue or to reverse, and do
they lead to changes in the S&P 500 index?
To find out,
we computed the correlation between all leads and lags in the price
histories of the two indexes. We used monthly data on returns from
year-end 1971, the year the Nasdaq began, through April 2002, a
total of 370 monthly observations. We looked at the scatter
diagrams, we computed a variety of statistical tables, and we
assembled tables of co-movements and counter-movements. All of these
tests tend to answer the question of whether large moves in one
series are accompanied by large moves in the same or opposite
direction of the other.
Our conclusion is that the only
relation that has statistical significance is that Nasdaq returns in
one month tend to have a moderate positive correlation of 10% with
the returns in the next month. This means that if the Nasdaq goes
down 15% in one month, the best estimate is that it will fall 10% of
15%, or 1.5%, in the next month. If it goes up 20% in one month, the
best estimate is that it will go up 10% of 20%, or 2%, in the next
month.
That doesn't reduce the uncertainty by much. In fact,
the predictions reduce the squared error of forecasts by a mere 1%.
But that's the best we can come up with. All other correlations are
random and consistent with chance. We ran many correlations, such as
correlating the last two months' return, skipping a month, checking
S&P 500 for various monthly leads and lags, and performing the
same exercise with Nasdaq vs. S&P 500 leads and lags.
Considering the many correlations that we ran, there is a
reasonable likelihood that the one significant correlation that we
did come up with is the kind of chance result that would occur if we
picked the winner of a heads-and-tails coin-tossing
contest.
The best we can come up with after examining the
data is that the tendency for positive correlation tends to be
greatest after big declines in the Nasdaq in a month. The three
largest declines in the Nasdaq along with their moves the next
month, are listed below:
| One big Nasdaq drop leads to
another |
|
Decline in month |
Move next month |
| Feb 2001 |
-22% |
-14% |
| Nov 2000 |
-23% |
-5% |
| Oct 1987 |
-27% |
-6% |
| Average |
-24% |
-8% | |
This
average decline of -8% compares to a 1% per month average gain in
the Nasdaq. It's not much, but it's the best that we can come up
with after examining literally hundreds of relations.
All in
all, our battle with the market momentum dragons did not end in
unqualified victory. And yet, we hope our approach will allow
readers to reflect upon and build beliefs about market momentum that
are closer to science than superstition. There is an infinite poetry
and beauty to market moves with inexhaustible wealth awaiting those
who can unravel the future course of movement.
Final note
We previously reported in our
April 18 article, “Buy companies
buying their own stock” that S&P 500 companies announcing
buybacks returned an average of 30 % more than the average stock
since year end 1999. These seem like good candidates for purchase to
us, and we have been buying them accordingly.
Since the
article, eight more S&P 500 companies have announced buybacks.
These companies are Charter One Financial (COF,
news,
msgs),
Eastman Kodak (EK,
news,
msgs),
Kimberly-Clark (KMB,
news,
msgs),
First Data (FDC,
news,
msgs),
King Pharmaceuticals (KG,
news,
msgs),
Linear Technology (LLTC,
news,
msgs),
Maxim Integrated Products (MXIM,
news,
msgs)
and Waste Management (WMI,
news,
msgs).
In
keeping with our abstemious ways, we will be adding them to our own
portfolio when the market takes a little swoon.
Our article
spoke with some admiration about The Buyback Letter's creator, David
Fried, who has been ranked No.1 by Hulbert in 5-year performance.
However, we queried whether his subjective methodology was capable
of capturing the kinds of regime shifts in buyback performance that
have occurred in the past. He kindly responded by telling us that he
felt that special factors were at work in making the performance of
buybacks among S&P 500 companies in recent years particularly
good. They tended to be concentrated in humble, value stocks, and
these were the ones that the market has been favoring
recently.
Fried questioned whether blind adherence to buying
such companies would do as well as the more qualitative analysis of
his own in the future. He pays particular attention to the reduction
in the number of shares outstanding in such companies and believes
that a reduction, which is his key trigger, is a better acid test.
He has kindly prepared a table of all buyback announcements with
number of shares outstanding for the benefit of our readers. Kindly
request the same by e-mailing The Speculator.
At the time of publication, neither Victor Niederhoffer
nor Laurel Kenner owned any of the equities mentioned in this
column.
|
