Learning from Trading with Ambiguous Information
54 Pages Posted: 20 Feb 2014
Date Written: February 19, 2014
In this paper we study long-term learning process under the conditions of ambiguous information. We study a unique empirical setting in which pieces of the ambiguous information are sequentially provided to institutional investors. We use parametric and non-parametric models to show that the actual convergence process is non-linear and takes a specific, sigmoid shape. We show that the learning process is strict and that the convergence is towards an unbiased point that represents the set of prices, which are dictated by the available information. The theoretical and experimental literature on learning processes is rich, but there are only a handful of empirical studies. We contribute direct empirical evidence that supports important theoretical and experimental predictions.
Keywords: Ambiguous Information, Learning process, Trade, Bonds
JEL Classification: G12, G23, G24
Suggested Citation: Suggested Citation