Voluntary Disclosure of Granular Information

Abstract

This paper presents a model of voluntary disclosure in which the manager's information about the firm's value is granular, i.e., consists of a large random number of imprecise signals. Using an argument in the spirit of the Bernstein-von Mises theorem, we show that there exists a simple equilibrium in which the manager discloses all signals exceeding a certain threshold, and the market prices the firm using the same linear rule on and off the equilibrium path. In contrast to models with monolithic information, the manager chooses the optimal disclosure threshold effectively ignoring the effect of disclosure on the investors's residual uncertainty about the firm's value.

Keywords: Bayesian Asymptotics, Voluntary Disclosure, Residual Uncertainty, Granular Information, Bernstein-Von Mises Theorem

JEL Classification: D21, D82, G32, M41