Information Epidemics

Abstract

This paper extends a growing body of research into the time-series properties of return and trading dynamics caused by direct information ow across a population of investors. The market setting we explore has the virtue of extreme simplicity including atomistic investors and continuous trading, which lead naturally to continuous price and volume processes characterized by dierential equations having analytic solutions. While consistent with prior literature on information dissemination via networks with relatively few agents or in discrete time, we are able to provide sharper asset market implications. With regard to information epidemics, for example, previous predictions of an \S curve in prices and non-monotonicity in returns are shown to depend critically on how markets clear and on details of the information event that lie outside the epidemic process itself. Moreover, these shapes can result from other types of information processes. Our simple structure also allows us to draw connections between return dynamics and the dynamics of trading volume. These connections emerge from a conuence of forces, only one of which is the information process itself. A close cor