Look
ma! No hands!
The roller coaster ride ride called the stock market is back up today.
The Dow Jones Average was up 4.6% Tuesday after a 7.2% drop (the 12th largest
single day loss, and the largest point drop ever)
on Monday. Speculation abounds, as always, as to why the market jumps so
wildly.
Yesterday, after the market closed, a parade of commentators spoke of
the market’s apparent “correction” and speculated about its future. Some
thought that the market was over-valued and that the drop was a natural
“re-alignment” of prices to fundamentals with the Asian Markets acting
as a trigger. Some though that the drop would continue, some not. With
the rebound today, I’d guess that just about half of the commentators were
proven wrong.
Ok. Fine. So like a great baseball player, the market sometimes needs
to adjust itself. The timing and the magnitude of these adjustments are
easy to explain after the fact, but hard to predict beforehand. One side
of the issue that has been neglected is the reason for why these adjustments
always seem to happen during the course of a single day of trading. Why
not a week, or a month?
Part of the answer, surprisingly, can be seen by thinking back to the
last time you saw a good concert.
Standing Ovation
The concert has just come to an end and the applause begins. “Wow”,
you think, “That was a great performance.” The applause gets louder, you
hear “bravo!” from someone in a seat in front of you. More applause…
what do you do? Keep clapping? Stop? Or rise from you chair for a standing
ovation?
If you stand up for the ovation, what happens if you are the only one
who thought the performance was good enough to warrant an ovation? You
might end up the only one standing and looking foolish – not all concerts
end in an ovation. On the other hand, you could just wait in your seat
clapping to see if anyone else is willing to stand up, then join them if
they do. Also, if a standing ovation does occur, you don’t really want
to be the only one sitting.
Now think of an auditorium full of people just like you who might like
to stand (and who would if others were) but are not willing to take a chance
of being the first (and only) one standing. What happens? Typically one
person, or a small group, is finally bold enough to make the first move
and stand, then, very quickly, others in the auditorium follow and do the
same, followed by more and more (including those who didn’t necessarily
like the show but feel too embarrassed to be the only ones still sitting)
and in only a few seconds the entire auditorium is standing.
Information cascades
The parallel with stock market crashes (and rallies) should be obvious.
The end of the concert is like the opening bell on a day in which many
of the participants think stock prices are over-valued. A normal amount
of trading follows at reasonable prices- nothing too extreme until someone
or some small group “stands up” and a sell-off starts. Like the people
who liked the concert, the ovation rapidly encompasses the entire audience
and the market participants are selling like crazy.
No one wants to be the first and only one to start selling since the
market is still rising, but once the cascade begins, no one wants to be
the last to sell. The result is the one-day crash. The same happens in
the one day boom – no one wants to be the only person to by in a falling
market, but one things pick up, no one wants to be left out.
The parallel can be expanded by seeding the audience and the market
with people with better, or different, information. Say you are at your
first Jazz concert and that you know very little about both the music as
well as the ovation habits of typical Jazz audiences. In this case you
are more likely to be a follower and to do what ever the better informed
or more experienced people in the audience do. If they liked the performance,
and stand you will gather that it was a high quality show and that standing
is appropriate, and thus you will be more likely to stand.
In the market case, one a sell-off starts people may be thinking “hey,
maybe they know something I don’t” and will follow the trend, since the
information generated by some trades helps to inform others about what’s
going on. Alternatively, the selling by others may confirm your beliefs
and give you more confidence to act on them.
Micromotives
The above ovation analysis began by looking at “micromotives” – the
actions of a single individual or investor in a larger group – and wound
up explaining “macrobehavior” – a standing ovation or market crash.
This style analysis is certainly nothing new to economics – it’s roots
go at least all the way back to Adam Smith. It is not, however, an easy
process to move from a description of individuals within a group to the
behavior of the group as a whole – especially if there are important interactions
between the individuals. If we wish to further complicate things by introducing
any kind of informational or other kinds of heterogeneity into the population,
it becomes very difficult to understand the macrobehavior of the system.
Computers are helping in this regard as it is becoming possible to simulate
a large group of interacting artificial computer generated “people” to
see what happens in, say, a market or an auditorium. The field of “Computational
Economics” is developing along these lines.
Elvis has left the building, Elvis has left the building.
I think everyone has had the experience of standing too long at the
end of an ovation – the feet start to tire, the hands get sore from too
much clapping. So now that a standing ovation has started, or a crash has
taken hold, how do we stop it? What we need is for some common signal to
coordinate our actions.
It was hoped that the circuit breakers would be able to provide this
signal and the worst of the crashes would be avoided. The fact that the
first circuit breaker did not stop things on monday means we need a better
signal (although, what would have happened without them is still an open
question).
It does appear that turing on the house lights and ending the day at
the market does help, since, as usual, a large rebound followed the cascade.
Or maybe it was just a good night’s sleep.
See Also:
I highly recommend the book Micromotives
and Macrobehavior by Thomas Schelling as a nice look at how individual
level motives can lead to unexpected macro behavior.
The standing ovation
problem is not originally my idea although this particular analogy
to stock markets may be. I first saw the problem at a Graduate
Workshop in Computational Economics at the Santa
Fe Institute run by John Miller
and Scott Page. They should get the credit for the ovation example.
The link between financial markets and information cascades is also not
new; however, the research still remains in the academic literature and
not on the Web!
More information on computational economics can be found through:
The Santa Fe Institute.
The Society for Computational Economics.
Center for Computable Economics Home
page.
Web Site for Agent-Based
Computational Economics.
John Miller also has a nice set of resources on Computational
Economic Modeling.
For a discussion dealing with complexity, chaos, and the stock market see
“Cracking
Wall Street” by Kevin Kelly. The article is based on the work of
The Prediction Company.
| October 97. Not the biggest drop or gain in history. The market has seen considerable swings in the past few days; however, It is true is that the absolute point gain has bounced around Today the market was up by nearly 4% after loosing around 6% on monday. In point terms, a crash of the 1987 magnitude would have meant around The “crash” on monday was in fact only the 12th largest crash in history.
|
Filed under: Finance
John Irons's Blog 