“Privacy and Statistical Discrimination with an Application to (Non-Discriminatory) Personalized Pricing” with Professor Kai Hao Yang -Part 1

This mini-course discusses a notion of privacy, and its connection to statistical discrimination. The lecture will begin by reviewing basic theories of the value of information a la Blackwell (1953) and of statistical discrimination a la Phelps (1972). A notion of privacy will then be introduced to Blackwell’s framework and connected to Phelps’ notion of statistical discrimination. The first part of the lecture will focus on the general theory of such notion of privacy, and will provide characterizations of privacy-preserving signals as well as their values in Blackwell’s sense. Several implications to algorithmic fairness and information disclosure in auctions will be discussed.  The second part of this lecture applies this framework to the context of non-discriminatory personalized pricing, and characterize the optimal (personalized) pricing strategies subject to a constraint that consumers’ protected characteristics cannot be discriminated.