The Effect of Irrelevant Information on Adverse Selection in a Signaling Game

Abstract

We find an effect of irrelevant information on adverse selection in a laboratory signaling game. This effect occurs via two channels: the principal is more (less) likely to adversely reject signals from “good” (“bad”) types. The findings suggest that “perception (or perhaps, misperception) of correlation” is sufficient for people to process information. Failure to recognize information as “irrelevant” is costly: Principals in our experiment are worse off by 3.75 percent. This suggests a “curse of (irrelevant) information.” Our findings could explain why buyers fall “victim” to irrelevant information in markets that are subject to adverse selection such as “lemon’s markets.”

Keywords: signaling game, (irrelevant) private information, adverse selection, principal-agent relationships, laboratory experiment

JEL Classification: C72, C91, D82