Slow learning in the market for lemons: A note on reinforcement learning and the winner's curse

A common result of experiments involving the "winner's curse" is that it is very persistent; subjects tend to overbid repeatedly, even when this causes them to lose money. Why do people keep making the same mistakes over and over again? We consider a class of one-player decision problems which are a generalization of Akerlof's [1] market-for-lemons model. We show that if decision makers learn via reinforcement, specifically by the reference point model of Erev and Roth [8], their behavior typically changes very slowly, and persistent mistakes are likely. We also develop testable predictions regarding when individuals ought to be able to learn quickly.