Research

This paper investigates the impact of artificial intelligence on the interaction between firms and consumers. It focuses on the use of learning algorithms in environments with strategic consumers — where learning must occur in the face of consumers who best-respond and adapt their behavior. An algorithm is transparent if consumers observe its inputs, whereas it is opaque if consumers do not observe its inputs. The main results are as follows. First, opaque algorithms perform better for the firm than transparent ones. In contrast to a transparent algorithm, an opaque algorithm learns the optimal policy and maximizes long-run profits. Second, opaque algorithms outperform transparent ones in terms of consumer welfare in important applications. That is, consumers may benefit from having less information about the algorithm's inputs. Third, whether the firm benefits from using an algorithm instead of behaving strategically depends on consumers' information about the algorithm's inputs. When the algorithm is opaque, it yields higher payoffs than a fully strategic firm.

This paper studies how to combine screening menus and inspection in mechanism design. A Principal procures a good from an Agent whose cost is his private information. The Principal has two instruments: screening menus – i.e., quantities and transfers – and (ex-ante) inspection. Inspection is costly but reveals the Agent's cost. The combination of inspection and screening menus mitigates inefficiencies: the optimal mechanism procures an efficient quantity from all Agents whose cost of production is sufficiently low, regardless of whether inspection has taken place. However, quantity distortions still necessarily occur in optimal regulation; the quantity procured from Agents with higher production costs is inefficiently low. A cost report triggers inspection only if the quantity procured from Agents at the reported cost is inefficiently low.  In contrast to settings without inspection, incentive compatibility constraints never bind locally, but only globally. Nonetheless, the paper characterizes which incentive constraints bind. 

I analyze the value of persistent private information in repeated competitive interactions with short-lived players. To address this question, I study a repeated zero-sum game between a patient, long-lived player who is informed of the payoff functions and a sequence of short-lived, uninformed players. When monitoring of past actions is perfect, Aumann and Maschler’s (1995) seminal “Cav u”-result obtains. The informed player loses her informational advantage when she utilizes her information, even when facing short-lived players; her equilibrium payoff is the same as when facing a long-lived competitor. When monitoring of past actions is imperfect, however, the payoff of the informed player can be strictly higher when facing a sequence of short-lived players instead of a patient long-lived player, depending on the payoff function and the monitoring structure. In this case, the informed player can leverage her private information without revealing it to the short-lived players. I provide a partial characterization of equilibrium payoffs when monitoring is imperfect.