Clarity Trumps Content: An Experiment on Information Acquisition in Beauty Contests

Abstract

We provide experimental evidence that under strong beauty contest incentives, players ignore signals from an information source with high content if the source has low clarity. Instead, they acquire equally costly signals from a source with higher clarity despite its lower content. Content measures how precisely an information source identifies an economic situation, whereas clarity measures how precisely the source content is commonly interpreted. Low clarity impairs players' ability to coordinate. When signals are provided exogenously, our experimental results are less severe than theoretical predictions, but consistent with level-2 reasoning in a cognitive behavioral model. When players acquire signals endogenously, ignoring a high-content source is more severe than theoretical predictions. Our results imply that when beauty contest incentives are strong (e.g., short-horizon trading), investors can completely ignore a firm's disclosure, despite its high content, if the disclosure is not sufficiently clear.