Information, beliefs, and decisions

Information is useful because it can reduce uncertainty and lead to better decisions. According to standard theory, information has value if agents can improve their outcomes by acting on the Bayesian posteriors associated with the information. However, there is ample evidence that belief updating deviates from the Bayesian benchmark, and that demand for–and use of–information is not driven solely by value. We study these issues in a single experimental framework, which combines belief elicitation with choice and use of information. Our contribution is twofold. We show that superficial aspects of the information source are important both for belief updating and for demand for and use of information. Second, we show that beliefs and decisions are linked in a non-trivial way.

Joint with Aniol Llorente-Saguer and Santiago Oliveros.