Price Adjustment to News with Uncertain Precision

Abstract

Bayesian learning provides a core concept of information processing in financial markets. Typically it is assumed that market participants perfectly know the quality of released news. However, in practice, news’ precision is rarely disclosed. Therefore, we extend standard Bayesian learning allowing traders to infer news’ precision from two different sources. If information is perceived to be imprecise, prices react stronger. Moreover, interactions of the different precision signals affect price responses nonlinearly. Empirical tests based on intra-day T-bond futures price reactions to employment releases confirm the model’s predictions and reveal statistically and economically significant effects of news’ precision.