Bargaining With Endogenous Learning

Abstract

I study a discrete-time dynamic bargaining game in which a buyer can choose to learn privately about her value of the good. Information generation takes time and is endogenous. After learning, the buyer can disclose verifiable evidence of her valuation to the seller. Examples include venture capital negotiations or procurement of new technologies, which sometimes feature significant delay due to endogenous costly learning. The buyer receives informational rents for any period-length only if learning is costly. The high-frequency limits of stationary equilibria result in a folk- theorem type of result about the delay until agreement. Maximal delay is achieved in equilibria with mixed pricing. Near the high-frequency limit, all stationary equilibria feature non-extreme prices and non-extreme payoffs. The analysis allows for closed-form solutions and for comparative statics.