The use of fuzzy facts and card-stacking has
pushed the discussion of P/E ratios into the realm of propaganda,
rather than rational analysis.
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:
• 3 lessons from ace
investor George Soros, 3/21/2002
• Why earnings optimism
is bad news, 3/14/2002
• Small business or
big, the rules are the same, 3/7/2002
More...
|
|
 |
| sponsored
by: | |
The Speculator
Fear, greed and other reasons to ignore
P/Es
Steering clear or selling out because the market is
'overvalued' and due for a fall, based on high price-to-earnings
ratios? Don't buy that bunk. P/Es say little or nothing about stock
performance.
By Victor
Niederhoffer and Laurel Kenner
The price-to-earnings ratio of the S&P 500 ($INX)
now stands at 29, according to Standard & Poor’s; or 40,
according to Barron’s; or 62, according to Bloomberg. That compares
with a 16.1 average over the past 50 years, or some other number of
your choice. Based on this uncertain and shifting statistic,
professional “bearocrats” are out in force these days telling
investors that the market is still “too high” and thus ripe for
another fall, despite 10% declines in each of the past two years.
What
a glittering generality. We have concluded instead from our own
studies that there is very little evidence that broad-market P/Es
and subsequent stock-market performance are related at all.
Moreover, we were unable to confirm research reported in the most
oft-cited recent work on P/Es: that of Yale University professor
Robert Shiller, author of the 2000 book “Irrational Exuberance.” The
most we were able to come up with was a very slight association of
high P/Es and inferior returns over very long periods -- seven years
or longer.
The use of fuzzy facts and card-stacking has
pushed the discussion of P/E ratios into the realm of propaganda,
rather than rational analysis. The propagandists of the market
specialize in fear and greed just as advertising pros and political
propagandists specialize in envy and regret, and they have made
efficient use of Shiller’s work.
Irrational ratios
Let’s back up for a
moment and remember that it was Shiller who displayed a diagram of
his findings on the relation between P/Es and subsequent 10-year
stock market returns to Alan Greenspan two days before the Federal
Reserve chairman delivered the famous “irrational exuberance” speech
on Dec. 5, 1996. It is widely believed that Shiller sparked this
speech and subsequent boosts in the Federal funds rate in 1999-2000
from 4.75% to 6.5%. Furthermore, he predicted “substantial negative”
performance for the entire first decade of the new century.
We hasten to add that our aim is not to disparage the good
intentions of the news reporters who pass along the latest
fear-and-greed updates. They are fulfilling their function in the
great ecological system of the market, ensuring that investors make
the maximum contribution to the upkeep of brokerages’ skyscrapers
and profit-sharing plans.
However, our simulation studies of
the relation between P/Es and returns from 1950 to the present show
that the results are consistent with nothing but randomness. Note,
for example, that at the beginning of 1970, the S&P 500 P/E was
16 and the subsequent five-year return was -6% per year. At the
beginning of 1994, the P/E was 21.3 and the subsequent five-year
return was 21% annualized. For periods of seven years or longer,
high P/Es are slightly associated with future inferior returns.
The problem with using returns for periods as long as seven
to 10 years is that each observation shares many years of return.
Thus, there are not enough independent observations and too many
degrees of freedom in choosing a beginning period and ending period
for the results to be differentiated from chance.
These
findings are in the same vein as those we reported in a recent
article -- “Why
earnings optimism is bad news” -- on the unexpected inverse
relation between S&P earnings changes and returns. (We are
pleased to provide our yearly scatter plots, regressions and
original data taken from the Standard & Poor’s Securities Price
Record to all readers who write in to gbuch@bloomberg.net.)
As
to why there are such discrepancies as noted above in P/E data,
reasons are plentiful and not susceptible to easy correction. Many
companies on a calendar year don’t report results until mid-April.
One sometimes does not know whether a particular P/E number reflects
the last 12 months of reported earnings or calendar-year earnings.
It’s uncertain whether companies that lost money are included in the
results. And a related problem is that the series may be revised
subsequently, so that data originally reported are quite different
from what eventually appears in summaries, price records and data
sources many years thereafter.
Illusions of validity
We are not alone
in our conclusions about the lack of a strong relation between P/Es
and market returns. Kenneth L. Fisher and Meir Statman, in a
detailed and definitive study covering the years 1872 through 1999,
concluded that P/Es provide unreliable forecasts of future returns.
“There is no statistically significant relation between P/E ratios
at the beginning of the year and returns during the following year,
or during the following two years,” they wrote.
How can we
explain such a divergence between the facts and the propaganda?
Fisher and Statman believe that the answer is the illusion
of validity. People are prone to believe in the accuracy of their
judgments much more than is warranted. Fisher and Statman enumerate
five psychological biases that have been studied extensively in work
by David Kahneman and Amos Tversky as the irrational components of
decision-making that lead to these mistakes. They are the biases of
overconfidence, confirmation, representativeness, anchoring and
hindsight; detailed explanations can be found in most books on
cognitive psychology.
In an e-mail to us, Fisher alluded to
the “weird assumptions” that underlie the Shiller conclusions. “The
Shiller stuff is largely a data mine,” he wrote. “It makes no real
difference how you redefine high and low P/E, regardless. P/E has no
significant predictive power for markets so far into the future that
it is enough to drive most folks crazy, i.e., longer than five
years.”
Fisher’s conclusion as to why people, including
Shiller, believe in the P/E fantasy is that our ancestors found
great survival value in being afraid of heights, and they passed
along their genes to us. “Anything in a heights framework generates
fear,” Fisher wrote. Even high P/Es.
In an interview from
his Yale office on Tuesday, Shiller told us that he is not sure that
the current high P/E levels are actually bearish at all. “The
recession has artificially depressed earnings to an inordinate
extent, thereby making P/Es appear much higher than a reasonable
base,” he said. He also noted that his measures of consumer
confidence, which move slowly, are at bullish levels for now.
However, in the long run he said he is as confident as ever that the
market is in for a period of abnormally low returns -- and he
expressed confidence that "econometric studies will verify the
significance of the negative P/E relationships” that he and his
colleagues have found.
How P/E
propaganda works
We have a different explanation for
investors’ errant concerns over P/Es: Individuals are induced by
market propaganda to take actions that benefit the messenger more
than the listener. To help you see them, we will lean on insights
into the seven standard techniques of propaganda as relayed in “The
Fine Art of Propaganda,” a 1939 classic of the topic. The techniques
were given catchy names so as to be simple enough to teach in public
schools: name-calling, glittering generality, transfer, testimonial,
plain folks, card stacking and bandwagon.
Our take on each,
as they relate to the market:
- Name-calling: “Only a blindfolded bull would buy stocks
at these levels.”
- Glittering generality: “There’s no value in the market
with these stratospheric P/Es.”
- Transfer (invokes a name or symbol): “Until P/Es fall
to Graham & Dodd levels, we’re forecasting a Dow of 100.”
- Testimonial: “The oldest and canniest investor I know
is waiting for P/Es to fall by another 50% before he considers
buying.”
- Plain folks: “I don’t know much, but the last time
dividend yields were this low was year-end 1928.”
- Card stacking: “P/Es were above 20 at the beginning of
1929 and stocks fell 40% in the next two years.” (Forget that P/Es
were the exact same level at the beginning of 1950, when prices
rose 40% over the next two years.)
- Bandwagon: “None of the hedge fund managers at the
conference in Bar Harbor will touch these high P/Es with a 10-foot
pole.”
Examples of propaganda are so widespread on Wall
Street that we have found numerous areas where investors must be
vigilant. Chief executives that guide research analysts’ earnings
estimates down so that their firms can beat them later is a nice
instance of card stacking. The use of sex appeal – as in focus on a
company’s “hot” technology rather than earnings power -- can
transfer excitement from miscreant firms to the dull analyst at the
end of a phone.
Kindly communicate your own examples and thoughts on this
subject to us by writing gbuch@bloomberg.net, so that
we can provide a comprehensive report on this in a future column. We
will send our traditional cane, to memorialize those carried by the
old-time speculators down to Wall Street during panics to buy stocks
at good prices, to the best few.
While propaganda is
ubiquitous in the market and other aspects of our lives, there is
one very simple antidote that will steer you away from manipulation:
science. If theories about the market are framed in a testable
fashion, then they can be verified by counting. This will not only
preclude debates about the meaning of words such as “irrational” and
“exuberant,” but will relate market theories to the real world. An
unintended consequence will be that those theories that are verified
can actually lead to practical profits.
End Note
We are sorry to report a snafu
in our promised effort to provide first-hand information on
General Electric (GE,
news,
msgs).
In the last communication that we received from the company, not
only did they refuse to grant us an interview in person, they asked
that we refrain from contacting the GE offices ever again. Since
that time, several interesting developments have placed the refusal
in perspective.
First, there was the brouhaha over Jack
Welch’s in-person interviews with Suzy Wetlaufer, editor of the
Harvard Business Review. And Welch himself has taken actions that
lead many to believe that he is ready at a moment’s notice to assume
the reins at GE again if Jeffrey Immelt’s work as chief executive is
greeted with further unfavorable market and analyst sentiment.
General Electric shares are down 6% this year, a regrettable
performance versus the Dow Jones industrials’ 3% advance. In 2001,
General Electric failed to beat the Dow’s performance for the first
time in five years, falling 15% versus the index’s 7%
decline.
Our studies of GE yearly returns, including
dividends, from 1899 to the present, show that an investor could
have realized a 7,500,000% return by investing in the company at the
turn of the century and holding through 2001 (reinvesting all
dividends, of course). If any readers have pull with a high-level GE
executive who can evade the "Don’t Talk to Niederhoffer"
interdiction and discuss this apparently touchy topic, we would
appreciate an introduction.
At the time of publication,
neither Victor Niederhoffer nor Laurel Kenner owned any equities
mentioned in this article.
|
