Re: poly: Why interest rates may stay low

From: Perry E. Metzger <>
Date: Fri Feb 27 1998 - 15:43:42 PST

Robin Hanson writes:
> >I'm reminded of all the beautiful results that say that markets should
> >be perfectly efficient and thus speculators like George Soros
> >shouldn't exist or "must be" statistical anomalies. And yet they
> >exist. Why? Because the theories are in fact not perfectly accurate --
> A single example of a company making a giant ROE does not even weakly
> discredit claims about the expected ex-ante ROE.

You forget that this isn't some random company. Cisco Systems is one
of the largest companies in the world, with a market cap in the
stratosphere. (It is largely invisible because it doesn't make
consumer products.)

You also forget that I know quite a bit about the company in
question. It legitimately *is* better than its competitors. It is also
unlikely that you are going to manage to compete with it very
effectively. They own all the best experts in the field, are very well
managed, and very responsive to the customer.

Microsoft is operating substantially out of the normal ballpark,
too. You could easily compete with "Windows". Put together a few
hundred people and write a clone. Nothing illegal about it,
either. However, it hasn't happened. Ditto for their other software,

There are lots of examples of this.

Companies and people aren't a giant liquid slurry that can be measured
in graduated cylinders. They are "lumpy". They are not fungible. Math
designed with lots of simplifying assumptions doesn't work very well
for them.

One can easily produce a very neat sounding theory about idealized
corporate behavior. Unfortunately, when you measure the real world, it
doesn't always fit. As it turns out, rates of return in the high tech
industry, in spite of the general lack of intellectual property, have
been extremely high.

> You might as well point to individual lottery winners and conclude
> the lottery is a good deal.

You are implicitly claiming here that Soros' returns are explained by
random chance.

The theory is that some investors out of millions will be spectacular
losers, and some will be spectacular winners. "They are just
statistical outliers" you say.

Well, no. Thats actually pretty much untrue. George Soros didn't make
his money on one bet -- he made it on a long series. Winning the
lottery once is explicable, but winning it week after week for years
would make one assume the fix was in.

Some of the successfull speculators keep track of their wins and
losses pretty carefully. I've worked with individuals who's runs would
be expected once in trillions of experiments (or even worse) were the
markets actually perfectly efficient. The fact that there are many,
many such individuals makes me highly skeptical of the notion that
they are random flukes.

Sure, you COULD flip "heads" by accident 100,000 times in a row, but
generally speaking, if someone does that we at some point guess that
they either have a two headed coin or are using a trick flip.

Another datapoint: although the vast majority of traders employed by
the larger houses wash out, the number who do not wash out is by no
means one in thousands -- its a much larger number than that, and the
people who did well last year by and large tend to do proportionately
well this year.

I find the "efficient market" theory very seductive in a theoretical
sense, but as a scientifically minded individual I cannot allow myself
to ignore the overwhelming evidence that it isn't true. It is,
certainly, a good first approximation, but it isn't exactly the case,
and the little differences are often important.

Sure, it COULD all just be random luck, but an easier explanation is
contained in the observation that perfectly efficient markets require
perfect and instant dissemination of information. The dissemination of
information, however, isn't free, nor is it instantaneous. Efficient
market theory also depends on actors being perfectly rational, but
they are in the real world only "mostly" rational.

The markets are, overall, very *close* to efficient. With time, as
communications speed up and as information becomes easier to get, the
markets increase in efficiency. However, there is always a tiny gap
between that which would theoretically be perfectly efficient and
reality. The speculators wedge their way in there. In a world where
trillions of dollars transit the global markets per day, they only
need to "wet their beaks" just a little bit in the passing stream to
become spectacularly rich.

The difference between theory and practice is greater in practice than
in theory.

Received on Fri Feb 27 23:46:00 1998

This archive was generated by hypermail 2.1.8 : Tue Mar 07 2006 - 14:45:30 PST