The Information Content of FOMC Minutes

Abstract

In this paper, we analyze the information content of Federal Open Market Committee minutes from 1987-2005. We apply an objective, statistical methodology known as Latent Semantic Analysis to decompose each minutes release into its characteristic themes. We show that these themes are correlated with current and future economic conditions. Our evidence suggests that market participants can extract a complex, multifaceted signal from the minutes. In particular, Treasury yield changes around the time of the minutes release depend on the specific themes expressed, the level of monetary policy uncertainty, and the economic outlook.

Keywords: Monetary policy, minutes, transparency, Latent Semantic Analysis, computational linguistics

JEL Classification: C4, D78, E4, E5