Research 

Working Papers:

A Manipulation-Proof Law of Large Numbers and the Scope of Belief Manipulation (Joint with Alex Frug)​

abstract:  We study long-run learning when data are mediated by an informed agent with limited discretion over how samples are reported. We ask when strategic reporting can prevent learning and, when it can, how far long-run beliefs can be manipulated. We establish a Manipulation-Proof Law of Large Numbers that characterizes when the true state is learned regardless of the agent’s dynamic reporting strategy. We then introduce a  notion of manipulability for sets of states that yields a simple characterization of the distributions of long-run beliefs that can be induced when the Manipulation-Proof Law of Large Numbers fails.

Search, Matching, and Dating Apps: Designed to be Deleted, or Designed to be Repeated? (Joint with Yair Antler and Daniel Fershtman)

abstract: Should an online matching platform utilize new technologies that enable faster search and better screening of proposed matches? On the one hand, the platform wishes to provide high-quality search and matching services in order to become more attractive to its users; on the other hand, high-quality matches reduce both the duration of search and the likelihood of repeat search, and can therefore reduce revenue from repeated use and advertising exposure. We compare the profits of a matching platform under technologies that differ in search speed and the screening of proposed matches. We find that better technology raises profit when the platform relies mainly on fees paid by users, but can lower profit when advertising revenue is sufficiently important, creating an incentive not to use improved matching capabilities fully.