|
To print article, click Print on your
browser's File menu.
Go back
Posted
2/13/2003
|
Pre-order
Niederhoffer and Kenner's new book
The Speculator
Recent
articles:
• A nice surprise may
not mean a nice profit, 2/6/2003
• Investing lessons you
can't refuse, 1/30/2003
• Buy the very worst
stocks you can find, 1/23/2003
More...
|
|
|
|
Advertisement
|
|
|
|
The
Speculator
Why a bad golfer
makes a better CEO
Companies run by the best golfers outperform
those run by the fairway-challenged, or so the story goes. But our results
suggest those executive Arnie Palmers might want to spend less time on the
links.
By
Victor
Niederhoffer and Laurel Kenner
Good
golfers make good chief executives -- and good stock returns, the old
theory goes. Put to The Speculator’s test, though, this golden triangle
turns out to be more of a Bermuda triangle.
We
almost hesitate to say so. After all, the purported link between a
company’s stock performance and its CEO’s golf handicap (the lower, the
better) is as widely accepted as the January barometer and the Super Bowl
theory. Yet duty compels us to reveal that all three fall into the realm of
mumbo, and that all three could serve as Exhibit A of Lesson 1 on spurious
correlation analysis.
It’s
unfortunate that this is so. Where but the links can a CEO spend a
leisurely few hours exercising, reaping the rewards of success while
enjoying nature? What better setting than the putting green to size up a
prospective executive’s manners? Which social setting could be more
conducive to negotiating a mutually beneficial deal?
As
CXO Golf, an organizer of golf conferences, says on its Web site, golfing
“allows participants to let their guard down, often becoming more
approachable, communicative and benevolent -- a combination that presents
an ideal opportunity for business relationship building."
“Where
else do you have the opportunity to gather 100 executives together for an
all-day event, which is both fun and laid back, while at the same time an
optimal environment for negotiating, networking and information gathering?”
CXO says.
Even
the humble Sage of Omaha, Warren Buffett, who holds only a 21.4 handicap,
rewards his CEOs, maintaining loyalty with games for Jack Welch in
Nantucket or Bill Gates in Pebble Beach. The saying, “If you really want to
know a man, play a round of golf with him,” has apparently been taken to
heart by executives such as Sun Microsystems (SUNW, news, msgs) CEO Scott McNealy, who
reportedly plays 300 rounds a year.
Executive drive
Numerous
studies link high stock returns with success on the golf course. The New
York Times, in a famous May 31, 1998, Sunday business feature, provided a
study of CEO handicaps and stock returns. Graef Crystal, who writes about
executive compensation, conducted the study for the paper. The article
concluded: “Given the strong correlation between golf handicaps and
performance, executive wannabes can glean an obvious career advancement
tip: Spend more time on the links.”
USA
Today reported in an Aug. 7, 2002, article that the stock performance of
the 30 Fortune 500 companies with the best CEO golfers was substantially
better than the rest of the group during 2001 and 2002. Other studies have
concluded that the companies of CEOs who improve their handicap in a year
perform better in the market than those whose handicaps worsen.
The
usual explanation cited is that if a CEO has enough drive to become a good
golfer, he’ll bring the same competitive spirit to his business.
Just
for the record, here are the CEOs with the five lowest and five highest
handicaps as of July 2002, according to a survey of 270 CEOs by Golf
Digest, which has been publishing a ranking every couple of years. (Golf
Digest’s latest survey is available online; you'll find a link at left
under Related Sites.)
|
Best
|
|
CEO
|
Company
|
Handicap
|
|
Scott McNealy
|
Sun Microsystems (SUNW,
news,
msgs)
|
0.3
|
|
Curt S. Culver
|
MGIC Investment (MTG,
news,
msgs)
|
2.9
|
|
Wm. G. Jurgensen
|
Nationwide Financial (NFS,
news,
msgs)
|
3.8
|
|
Stephen M. Bennett
|
Intuit (INTU,
news,
msgs)
|
4.6
|
|
Michael B. McCallister
|
Humana (HUM,
news,
msgs)
|
4.8
|
|
|
Worst
|
|
CEO
|
Company
|
Handicap
|
|
J. Barry Griswell
|
Principal Financial (PFG,
news,
msgs)
|
35.1
|
|
James L. Dolan
|
Cablevision Systems (CVC,
news,
msgs)
|
33.3
|
|
Peter R. Dolan
|
Bristol-Myers Squibb (BMY,
news,
msgs)
|
31.5
|
|
Michael D. Lockhart
|
Armstrong Holdings (ACKHQ,
news,
msgs)
|
31.2
|
|
Barry W. Perry
|
Engelhard (EC,
news,
msgs)
|
30.0
|
|
Before
adopting Golf Digest’s list as a stock-picking guide, our advice is: Take
all of the above studies with a million grains of salt.
The
New York Times study we cite could be a mainstay in a basic college course
on the fallacies to which decision-makers and the media fall prey. A close
look shows that Crystal excluded from his 51-CEO sample six executives with
low handicaps and terrible stock returns, and one executive (our MSN Money
boss, Bill Gates, who had a 24 handicap and great stock returns) as
“extreme data.”
Watch out for these analytical sand traps
The
main problems with these studies and many others that attempt to forecast
stock returns were detailed with panache by Barry Ritholtz, a lawyer,
technology entrepreneur and U.S. market strategist for Ehrenkrantz King
Nussbaum, a New York investment bank, on his personal Web site (again,
you'll find the link at left.). We list the problems he noted, augmented
with our own skeptical points:
- Unproven
assumption. The studies assume a causal relation where none exists.
- Failure
to consider other explanations. If there were a causal relation,
it might well run the other way. The perks given to successful
executives can include club memberships, and time off to play, and
even private courses. There's the infamous case of the $51 million
golf course that Lucent Technologies (LU,
news,
msgs)
developed for its executives and customers back when it was flying
high.)
- Fuzzy
data. The compilers of the New York Times study were able to come
up with handicaps for only 51 CEOs, and what data they had was often
self-reported. As is well known, amateur golfers are notorious for
underreporting their scores.
- Hothouse
results. The relation might exist, but only for a certain time and
place, such as the bull markets of the 1980s and 1990s.
- Data
mining. If you selected only one horse from a corral with hundreds
of possible choices, some of the horses will seem to have shown good
form for selected years and events by chance alone. The golf
handicap-stock return studies could have considered the relation
between golf in January and the rest of the year, or the similarity to
the high handicaps that existed in 1929, or the relation between
tennis and stock returns.
- Limited
sample. The analysis (what there is of it) is limited to only half
the sample. It is quite likely that the non-reporters would
drastically change the results. A similar but much less excusable
problem occurs in Didier Sornette’s book, “Why Stock Markets Crash:
Critical Events in Complex Financial Systems,” where he focuses on predicting
the left tail of a distribution of stock moves (the big declines)
without reporting the expectations for the rest of distribution,
including the right tail (the big rises).
Regrettably,
the flaws listed above characterize thousands and thousands of stock-market
studies. Such studies mainly serve to put investors on the wrong foot. It
so happens that one of our best and most admired friends, Yale Hirsch, is
the inventor of the January barometer. While we don’t find the barometer at
all compelling, Hirsch is one of the great countists of our day, and we tip
our hat to him while cautioning against blind acceptance of that or any
other indicator of this nature.
How to test goofy golf theories
Today,
we’ll try to show a proper method for testing the golf handicap-stock
performance phenomena -- not only to reveal the fallacy in this particular
correlation, but to serve as a model of how to consider other relations of
this nature.
Starting
with Golf Digest figures as of June 2000, we compiled a list of the
handicaps of 172 top golfers among the CEOs of Fortune 500 companies and
returns for those companies in both 2001 and 2002.
In
considering a relation like this, the first and most important step is to
draw a scatter diagram that lets us look at returns vs. handicaps. The
results from the scatter diagrams let produce this succinct summary:
The
points in the aggregate on the scatter chart and the data that summarizes
the results show a positive correlation between golf handicaps and returns.
In other words: the worse the golfer, the higher the predicted return. The
companies led by the best golfers performed terribly. (The lower the
handicap, the better the golfer.)
|
Handicaps and returns
|
|
Handicap
|
Count
|
2001
return
|
2002
return
|
Average
|
|
<10
|
26
|
-2
|
-22
|
-12
|
|
10-20
|
100
|
13
|
-16
|
-2
|
|
20-30
|
46
|
11
|
-8
|
2
|
|
>30
|
Not reported*
|
-1
|
-22
|
-12
|
|
*The
Speculators assume those Fortune 500 execs who didn't report have high
handicaps.
Forecast equations
By
going one step further, you can use a scatter diagram to derive an equation
for forecasting. (We find that equations, while not perfect by any means,
work much better in trading than guessing, dart-throwing or watching
talking heads on TV.)
We
find that the average expected return for a 0 handicap would be -27%. So
the regression equation for 2002 returns is:
Return
= -27% + 0.8 x handicap
In
other words, if the handicap was 20, the predicted return would be:
-27
+ 0.8 x 20 = -11
Strangely
and ominously for all other studies, since it shows how subject to biases
even a well-designed study like this can be, the 0.8 slope of the handicap
coefficient in 2002 has less than a 4% shot to occur through chance
variations alone.
The
regression equation for 2001 is:
Return
= -3 + 0.9 x handicap
Thus,
if the handicap were 1, the predicted return would be -2.
Real-world results
Not
surprisingly, our conclusions were completely opposite to the conventional
wisdom. As in most relations of this nature, the out-of-test tube results
were much less supportive of the theory than the retrospective contrived
similarities suggest. The principles of ever-changing cycles and “garbage
in, garbage out” both apply.
Remarkably,
not a single one of the eight top-ranked CEO golfers we contacted was
willing to comment. In fact, the only two who bothered to have their people
get back to us were No. 1 CEO golfer Scott McNealy and No. 8-ranked L.
Phillip Humann of SunTrust Banks (STI, news, msgs), both of whom merely
relayed their unwillingness to comment. (We spared 4.3-handicap Richard A.
Snell, former CEO of now-bankrupt Federal-Mogul (FDMLQ, news, msgs), from our inquiries.)
We
will not add the no-commenting CEO golfers to the list of executives who
refuse to talk with us. So far, the list starts and stops with General
Electric (GE, news, msgs), which asked us when its
stock was at 42 to refrain from contacting the company ever again.
We
think a good half-hour or so of vigorous exercise is necessary every day,
but we feel that an executive should not be spending his time at the golf
links, and our study underlines this.
Apparently
the attention that district attorneys and the media have focused on
executive perks is finally making a dent.
“This
has been a tough year,” Richard B. Priory, CEO of Duke Energy (DUK, news, msgs), told Business Week
magazine in November. “My clubs are sitting in a corner, waiting for a
rebound.” Priory told the magazine his handicap had increased from 6.3 to
10.3.
In
today’s belt-tightening atmosphere, it’s more fashionable to be like Reuben
Mark, the present Colgate-Palmolive (CL, news, msgs) CEO, than his predecessor,
David Foster, who instituted the company’s sponsorship of the famous Dinah
Shore golf tournament and bought a golf club near his planned vacation
home. Mark, a non-golfer, dropped the golf sponsorship, shuns publicity and
concentrates on toothpaste.
Golf
is not the game the Specs know most about. But Paul DeRosa, chief bottle
washer at Simpson & Co., one of the best-performing hedge funds and
author of the withering indictment of the faulty logic in Bill Gross’s Dow
5,000 prediction (available by writing to the Spec Duo), is a knowledgeable
golfer with a handicap below 10. He adds: “Several golfing CEOs live in
California, and we know how they've done in recent years. (Former Long-Term
Capital Management CEO) John Meriwether’s handicap hit its all-time low in
1998.”
Our
complete workout of CEOs, handicaps and stock returns for the companies
studied is available on the Spec Duo’s Web
site. We thank Patrick Boyle for his statistical work on this
column. Kindly e-mail us with your knowledge of time and place of
what’s happening out there to so we can improve together. We respond to all
queries not related to individual stocks.
Final note
After
searching for investment opportunities related to the Middle East conflict,
we concluded that Saddam futures probably offer greater rewards (and hence,
greater risks) than defense stocks. These futures are traded on the Irish
betting site Tradesports.com, where you can bet that Saddam Hussein will
leave by the end of March, April, May and June. See the link at left.
At the time of publication, Victor Niederhoffer and Laurel Kenner did
not own or control any of the securities mentioned in this column. Both of
them play racquet sports.
|
