Events

Prof. Doron Ravid
Associate Professor in Economics and the College
Kenneth C. Griffin Department of Economics
The University of Chicago

We analyze games preceded by a stage of covert information acquisition, spanning the set of behaviors and welfare outcomes attainable across all specifications of players’ learning abilities. We show the only robust behavioral prediction is a constraint known as obedience, which also characterizes the behavioral predictions attainable in games with prespecified information (Bergemann and Morris 2016). However, the set of welfare outcomes attainable by varying the players’ acquisition technology is larger than those attainable by giving players a fixed information structure. We show this difference is always relevant for some social planner whenever it arises. We also span the predictions attainable across all flexible information acquisition technologies, showing such technologies can only generate behaviors that satisfy a novel separation constraint. We prove this constraint either refines most behaviors satisfying obedience, or refines none of them, with the latter hold for generic games. Finally, we apply our framework to bank runs, informational simplicity, and design with aftermarkets.