|
Printable version |
From: | "david.gibson" <david.gibson@k.co.nz> |
Date: | Thu, 18 Sep 2003 17:31:23 +1200 |
I have
been thinking about the point you raised in you email:
The predominant acedemic view in the 80's and early
90's was that the stockmarket obeyed "the law of large numbers" and that
stock time series were a "random walk" phenomenon based on the "efficient
market" theories.
My personal view is that these assumptions are
false.
Conventional models based on the assumption of zero
autocorrelation of stock trajectories are clearly false - yet this is the
dominant view in the acedemic literature.
The problem is that the "herding" and other
psychological effects that drive the short term market are not based on "linear"
processes (this includes periodic "cycles"). The effects are non-linear
and will require a new statistical modelling technique to be
developed.
[D Vidal has a number of very useful links on this
topic].
Finally, getting back to your question - based on the
above - my conjecture is that small cap / large cap obey the same basic
psychologically driven laws. I would expect the same types of models would
apply to either capitalisation.
Best
Rgds
/dbg
|
Replies
References
|