M Burke's letter to Barron's MailBag
The Future of Options
To the Editor
Michael Santoli's excellent column (The Striking Price, April 19) points out a
fact that has been the source of much discussion in recent years -- notably,
that the United States equity markets exhibit a price distribution with fatter
tails than traditional option pricing models would suggest. No one can doubt
the Black-Scholes-Merton option pricing formulas were revolutionary in their
time and indeed worthy of the Nobel Prize they generated. Their widespread
popularity would attest to the usefulness of these tools.
However, since these theories were first employed the environment has
changed in many ways. A sea change in the availability and processing power
of computer technology for the average trader has occurred. One can now
purchase up 25 gigabytes of disc capacity for prices in the $300 range.
Historical databases of equity prices are available free from numerous sources
on the Internet.
My hope is that a new generation of traders, academics and theorists use
today's technology to create models as groundbreaking as the ones we have
used in these past three decades.
NICK WILLETT
Getzville, New York
To the Editor
I had mixed feelings reading Michael Santoli's column. In some ways, I was
glad he mentioned Larry McMillan's notion that abnormal pricing is fairly
common in options trading. For several decades, when I've told folks I was a
buyer of options, they looked at me like I was Forrest Gump.
Sometimes being a contrarian makes one feel guilty about making money by
doing the wrong thing at the right time. Now, thanks to McMillan's work, I no
longer feel like I should return the outsized profits I've made buying options
over the past two decades-plus.
The mixed part of my feelings comes from the fact that McMillan, Keon and
Santoli are names the public often listens to. If they convince the
dumber-than-ared-brick-fence volatility sellers to leave the marketplace, that
could definitely put a damper on my future profits. If that is the case, I expect
Santoli et. al to compensate me for lost earnings. They broke the cardinal rule
of trading by giving the suckers an even break, and I don't want to pay for
their generosity.
MICHAEL BURKE
Houston
The Article that he Responded to:
April 19, 1999
When "Normal" Isn't
Defying the options probabilities
By MICHAEL SANTOLI
Armchair market watchers aren't the only ones left at a loss these days by the
stunning one-day jumps and dives of so many stocks. Indeed, even the
number-crunchers and the formulas they rely on to make trading decisions fail
to account for just how potentially volatile a given stock is likely to be. And
this systematic lowballing of potential volatility can have serious implications
for options players, particularly sellers.
Larry McMillan of McMillan Analysis offers a primer on this phenomenon
with an important bit of recent research. The crux of the issue is that just about
all probability estimates that go into option-screening and pricing models rely
on a "normal" distribution of a stock's historical price volatility. The standard
bell curve is one sort of representation of normal distribution, which would
suggest that a particularly large price move (in statistical terms not necessary
to explain here, three standard deviations) should occur no more than 1/10th
of 1% of the time. A massive move of close to three times that magnitude
should be nearly impossible -- the probability number has 15 zeros after the
decimal.
Yet by picking a trading day at random, McMillan showed that such moves
happen with regularity, belying their statistical remoteness. On April 5, he
found, four stocks made moves of eight or more standard deviations, including
major names like Ameritrade, Sabre Group and Checkpoint. To ensure that
this wasn't the fluke results of an overly volatile market, he went back and
found that on the least-volatile day of the 1990s -- July 25, 1993 -- 12 stocks
had moves that the normal distribution of price changes would suggest were
blue-moon events. These included such substantial stocks as U.S. Steel and
Novell.
The reason for the underestimation of
possible movements is that the only way
to gauge such things is to use a stock's
price history. And a fundamental rule of
history is that things are "normal" only
until they're not. Not long ago, Ed
Keon, head of quantitative research at
Prudential Securities, illustrated the limits
of quantitative investing using the 1998
major league home-run race between
Mark McGwire and Sammy Sosa. He
figures a historical analysis of the 20,000
batting seasons completed this past
century would produce an average
below 10 home runs a year per hitter, with a standard deviation near 5. That
would make a 70-home run year spectacularly unlikely in a mere 100 years.
As for the chances of two hitters spanking more than 65 in a single year,
Keon says, "in precise statistical terms: No way!"
The upshot of all this for option traders is that bets against volatility -- most
acutely selling naked options -- are even more risky than standard analysis
suggests. Selling calls or puts without owning the underlying stock is a
much-disparaged strategy by conservative types anyway, as it represents a
trade where the upside is limited and the downside potentially infinite if the
stock rushes against you. But in light of McMillan's admonishments, the
danger is even greater.
On the flip side, the probability of profiting from straddle buys -- in which
both a call and put are bought in hopes that any major stock move leads to
profit -- is likely higher than the off-the-shelf analysis systems would have you
believe.
Speaking of straddles, Don Fishback, cited here recently for a couple of trade
ideas, reports that his volatility plays -- in which he looks for underpriced
options -- have been doing well. One of those, a Case Corp. 20 straddle, has
gone deep in the red now that the farm-equipment firm's stock has leapt more
than 50% in a month. Last week, Fishback said, he was setting a straddle on
Boise Cascade.
During Wednesday's technology selloff, the hot new Nasdaq 100 index
shares (QQQ) traded a record 10.6 million shares, meaning more than $1
billion worth of the leading U.S. technology companies changed hands. Worth
noting: The shares can be sold short even on a downtick, unlike stocks.
|
