Belief Updating and the Demand for Information

Abstract

How do individuals value noisy information that guides economic decisions? Using a novel laboratory experiment, we investigate the willingness-to-pay for useful information. We identify two systematic biases in information demand. Compared to the standard rational agent model, individuals over-value low quality information and under-value high quality information. They also disproportionately prefer information that may yield certainty. By eliciting agents’ posterior beliefs, we find that both these biases are well-explained by non-Bayesian belief updating. Furthermore, individuals differ consistently in their responsiveness to information - the extent that their beliefs move upon observing signals. This single trait explains about 80% of the variation in information demand that is attributable to belief updating. It is also predictive of behavior across different choice environments. Thus, we find that the demand for information is driven by belief updating, and that belief updating behavior is a stable individual trait.