Optimal Refund Mechanism With Consumer Learning

This article studies the optimal refund mechanism when an uninformed buyer learns about their valuation over time. We consider various refund mechanisms including simple return policies (no returns or free returns), and stochastic return policies, which allow the buyer to keep the product with some probability upon receiving a refund. We show that the optimal refund mechanism is deterministic and takes a simple form: either the seller deters buyer learning by offering a low price without returns, or encourages maximal learning by setting a high price with free returns. Interestingly, free returns are optimal only for intermediate prior beliefs.