> But the real question is whether there were any clustering in the
>attributes of the minority who consistently beat the market. If there is a
>strong clustering of attributes (they ate the same brand of corn flakes for
>many many years or went to the same school or followed the principles of
>the same Guru of investing or whatever ...) obviously then there is some
>causal variable that may explain the phenomenon and it would not be
>scientific to dismiss this clustering by taking the argument that majority
>under performs the market anyway.
There is strong evidence against clustering. Clustering would imply that there should
be a fairly strong correlation between who outperforms the market in one year and who
outperforms the market in the next. A study done a few years ago (NBER working paper -
- I don't recall the authors) showed that there was a statistically significant, but
vanishingly small, correlation in the performance of publicly traded funds from one
year to the next. So you can increase your expected return a tiny bit above the
average for all mutual funds by picking a fund that performed well in the previous
year, but from what I remember it wasn't enough of a gain in expected performance to
overcome the under-performance due to over-management [*whew*].
So then what should we make of the fact that several people who follow a
particular strategy have all done well? Nothing at all. Suppose that I profess to the
world that the thing to do today is to own gold and drug stocks. Suppose that I happen
to get lucky and those two assets do particularly well over the next five years. Would
anyone be surprised if dozens of people from the "golden-drugs" school also did well?
In the example cited above one would need to look deeper. Have all these people done
well picking different stocks using the same principles or does the fact that they
ascribe to the same principles mean that they have all picked mostly the same stocks
and therefore had highly correlated returns? If the latter then there is no more of an
insult to efficient market theory than if one person had done very well for the same
length of time. And in a market with lots of participants that will happen frequently.
All that aside, there is evidence that value investing works and that is an
anomaly. However, as I recall, the increase in expected returns that one gets by
following such a strategy are measured in basis points, not percentage points as some
advocates of this approach would suggest.
One other thing. I very much liked Alex's thoughtful commentary on this. As he
noted, there are lots of anomalies that can mean that stocks are badly mispriced, but
that doesn't necessarily mean that there are guaranteed excess returns out there. This
was the point of Summers' old noise trading model. You can have market equilibrium
with irrational traders dominating the market, but the additional risk that their
behavior induces in the market exactly offsets the increase in expected return that is
created by the mispricing that they cause.
This possibility was made all too clear to me when I took a $20,000 short
position in Amazon.com - - a year too early. I haven't met anyone who will argue that
Amazon wasn't over priced at that time, and if I hadn't been forced to abandon the
position or face bankruptcy I would have made money. However, I couldn't afford the
margin calls and ended up losing a lot of money on the deal. Insult was added to
injury when a year after I was forced to abandon my position I had to sit at a dinner
table listening to someone brag about how much money he had just made shorting
Amazon.com, and about how stupid participants in the stock market obviously are. When
I asked him how he had decided when to take his short position he cited an argument
with another person over market efficiency as the precipitating incident - - in other
words dumb luck. - - Bill Dickens
William T. Dickens
The Brookings Institution
1775 Massachusetts Avenue, NW
Washington, DC 20036
Phone: (202) 797-6113
FAX: (202) 797-6181
E-MAIL: [EMAIL PROTECTED]
AOL IM: wtdickens
