Abstract We use a survey experiment to generate direct evidence on how people acquire and process information. Participants can buy different information signals that could help them forecast future national home prices. We elicit their valuations and exogenously vary the cost of information. Participants put substantial value on their preferred signal and, when acquired, incorporate the signal in their beliefs. However, they disagree on which signal to buy. As a result, making information cheaper does not decrease the cross-sectional dispersion of expectations. We provide a model with costly acquisition and processing of information, which can match most of our empirical results.

ABSTRACT Innovations in statistical technology in functions including credit‐screening have raised concerns about distributional impacts across categories such as race. Theoretically, distributional effects of better statistical technology can come from greater flexibility to uncover structural relationships or from triangulation of otherwise excluded characteristics. Using data on U.S. mortgages, we predict default using traditional and machine learning models. We find that Black and Hispanic borrowers are disproportionately less likely to gain from the introduction of machine learning. In a simple equilibrium credit market model, machine learning increases disparity in rates between and within groups, with these changes attributable primarily to greater flexibility.