Home

Hi! I’m a Ph.D. Candidate in Economics at Cornell University, with an interest in games, information, and applied theory. My work focuses on practical applications of information design, and my committee consists of David Easley, Tommaso Denti, and Larry Blume.

Research

I consider a model of Bayesian persuasion with observational learning. A sender designs a test to encourage adoption of a new product, which may be either good or bad. The test and its outcome are viewed by an opinion leader, who chooses whether or not to adopt the sender’s product. If the opinion leader chooses to adopt, the test and its outcome are next viewed by an audience who also makes an adopt/reject decision. Both opinion leader and audience strictly prefer to adopt the product when it is good, and reject it when it is bad. The cost of adoption is heterogeneous for each individual, and each individual’s adoption cost is their own private information. The sender’s goal is to maximize the probability of audience adoption. Compared to a benchmark in which the audience views the sender’s test directly, I show that the presence of the opinion leader weakly increases the informativeness of the sender’s test. However, it may lower the probability that the audience views the sender’s test, and so the overall effect on audience welfare is ambiguous. I extend this model to include two forms of collusion. Under information-sharing, the opinion leader informs the sender of her private type. In this case the audience weakly prefers the opinion leader to be present, and they may or may not prefer the opinion leader to share her type with the sender. Under cooperation, the opinion leader adopts the sender’s product regardless of the test result. In this case, the audience receives the same welfare as in the benchmark where the sender communicates with them directly.

A researcher wants to persuade a policymaker to adopt his treatment, which may be either good or bad. The policymaker wants to adopt the treatment when it is good and not when it is bad. The researcher chooses how many subjects to enroll in a trial, under either the sequential sampling regime or the pre-registration regime. Each subject improves with probability ρ when the treatment is good and probability 1-ρ when the treatment is bad. Under pre-registration, the researcher commits to his choice of sample size at the start of the trial, while under sequential sampling the researcher can observe each subject outcome before deciding whether to continue the trial or not. I show that under sequential sampling, as ρ approaches .5, the researcher can achieve his first-best Bayesian persuasion outcome, which minimizes the policymaker’s utility over attainable BP outcomes. At the same time, under pre-registration, the researcher’s optimal trial approaches full revelation, which maximizes policymaker welfare. I also show that under pre-registration, regardless of ρ, the sender is bounded away from his first-best outcome, and when the good state is at least as likely as the bad state ex ante, full revelation is optimal for the sender. However, when the bad state is more likely than the good state, and subject outcomes are very informative, policymaker welfare is higher under sequential sampling than pre-registration.

A Bayes correlated equilibrium of an incomplete information game (G,S) is any outcome that an omniscient mediator could induce with an appropriate information design policy. In this paper I introduce a refinement of Bayes correlated equilibrium called “hear-no-evil Bayes correlated equilibrium” which further requires that no player can make themself better off by publicly committing to ignore the mediator’s message. I provide a simple example in which an agent in a contest strictly prefers not to learn his opponent’s type. I show that the set of hear-no-evil equilibria of an incomplete-information game (G,S) is shrinking in S if and only if the set of hear-no-evil equilibria of (G,S) is the same as the set of Bayes correlated equilibrium of (G,S). I provide an example of a BCE which satisifies the HNE condition but is Pareto-dominated by a BCE which fails the condition. I also show that in the parameterized 2x2x2 coordination game of Taneva (2019), all the symmetric BCE satisfy the HNE condition.

CV