\documentclass[12pt]{article}

\begin{document}

\title{Interacting Brownian Motions and the Distribution of Equity Capital}
\author{Soumik Pal, Cornell University}
\date{January 26, 2007 }

\maketitle

The influence of a stock on the entire equity market is often measured
by its market weight, i.e., the proportion of the total market capital
that belongs to that stock. Empirical evidence often suggests that the
growth and volatility of stock capitals depend less on the name than
on the rank of its market weight among the market weights of all
stocks. In fact, the distribution curve obtained by plotting ordered
log-capitals in American stock exchanges against log of their ranks
demonstrates remarkable stability over decades. In order to
mathematically model such phenomena, Bob Fernholz, together with his
co-authors A. Banner, and I. Karatzas, introduces multivariate
interacting geometric Brownian motions whose parameters depend on the
random time-varying order in which the coordinates are
ranked. Analysis of such rank-based models are difficult, and previous
attempts use the equally intractable world of reflecting Brownian
motions. We develop novel tools based on polytope geometry to study
such processes and their generalizations, and derive long-term
properties which are consistent with empirical observations. Our
results also settle some of the open questions raised by Banner,
Fernholz, and Karatzas. One of the major objectives has been the study
of market diversity as implied by the rank-based models. We prove that
when the number of stocks in the market is large, most of such models
exhibit a phase transition phenomenon which drastically affects the
concentration or spread of capital in the equity market.

This is based on a joint work with Sourav Chatterjee of
U. C. Berkeley.

\end{document}

%%% Local Variables: 
%%% mode: latex
%%% TeX-master: t
%%% End: 
