Presented By: Department of Economics
Non-Discriminatory Personalized Pricing (joint with Philipp Strack)
Kai Hao Yang, Yale University

A unit mass of consumers with unit demands purchase a product from a monopolist. Consumers have a binary protected characteristic, which is associated with value distributions ranked in the likelihood ratio order, conditional on the cost of serving them. We characterize the revenue-maximizing market segmentation and pricing strategy subject to a non-discriminatory constraint, where consumers with the same cost but different protected characteristics must face the same price distributions. This problem is equivalent to an optimal transport with a non-supermodular objective function. When consumers' value distributions given protected characteristics are different enough, consumers could retain positive surplus under the profit-maximizing pricing rule, although which protected characteristic benefits more is generally ambiguous. Moreover, these surplus are enjoyed by consumers with intermediate values, whereas high-value and low-value consumers do not retain any surplus.