Research

Abstract: This paper analyzes a traditional negotiation framework, introduces a neutral third party as an information designer, and explores the limits of efficient dispute resolution. In our negotiation framework, the neutral third party—whether a mediator or arbitrator, referred to as the neutral—facilitates the division of a fixed surplus between two privately informed parties while seeking to minimize the escalation of a costly conflict. Recognizing that neutrals are often experienced professionals in practice, we deviate from the conventional assumption of an uninformed neutral and instead allow the neutral to have imperfect private information about the disputing parties. This enables us to explore scenarios in which neither the neutral party has an informational advantage over the players (as in classic Bayesian persuasion and information design literature), nor do the players have an informational edge over the neutral party (as in Myerson’s communication games and classic mechanism design literature). Our findings indicate that full efficiency is attainable, a conclusion that sharply contrasts with earlier research asserting that full efficiency is impossible with an uninformed neutral. We compare three mediation formats and conclude that directive mediation—where the mediator disregards the reports of the disputing parties—exhibits greater efficiency compared to the other two formats. Moreover, while existing literature suggests that mediation and arbitration are equally inefficient alternative dispute resolution methods, our analysis reveals that this equivalency may break down in favor of arbitration when the neutral possesses some information about the disputing parties.

Abstract: Information is often distributed across multiple senders. This paper explores the extent to which senders with a common utility function but independent and imperfect information can collaboratively persuade a receiver. We suppose that the senders' combined information (almost) fully reveals the true state, and compare these senders' attainable utility to the maximum achievable by a single, fully informed sender. We show that decentralized senders cannot generally attain this maximum, except in cases where fully revealing a state is optimal. On the positive side, we show that they can guarantee a constant fraction of this maximum, and that the maximum itself can be obtained if senders have access to a public randomization device.

Abstract: Consider a Bayesian persuasion problem with one sender and one receiver. If the sender is restricted to sending only verifiable messages, are they strictly worse off than under the optimal information structure with unverifiable messages? We show that the answer is no if and only if, there exists an optimal distribution of posterior beliefs in the Bayesian persuasion problem under which, (i) the support of the distribution is smaller than the number of states, and (ii) a condition we call “pooling verifiability" is satisfied. Specifically, we show that whenever the number of receiver actions is weakly less than the number of states, requiring that messages be verifiable does not hurt the sender. We provide an example with two states and three receiver actions to show that when the optimal distribution of posteriors under general Bayesian persuasion includes two beliefs that both have full support, a restriction to verifiable messages has bite, and the sender is strictly worse off than under Bayesian persuasion.

Abstract: Why do firms rarely use observable output to evaluate employees? This paper explores contracts that allow firms to commit to evaluation schemes and compensate employees based on these evaluations in a dynamic principal-agent setting. We show that introducing evaluations enables the principal to implement stochastic efforts in later periods, which can help mitigate the inefficiencies of punishing the agent for low outputs. This provides a rationale behind the prevalence of performance evaluations in practice. Moreover, these benefits depend on dynamic interactions and the agent’s risk aversion; they disappear when the agent is risk-neutral or in static settings.

Abstract: This paper investigates how to persuade multiple interacting receivers under general equilibrium selection and when receivers’ preferences may exhibit psychological traits. Our main result charac- terizes the conditions under which the sender can categorize information using partitions and then disclose partition-related information without altering the equilibrium outcome. This finding broadens the traditional direct approach based on the revelation principle, which categorizes messages solely based on equilibrium actions. Consequently, our result supports a generalized direct approach, which we apply to the study of information disclosure in fostering fairness within a group where some members may demonstrate altruistic traits.

Abstract: We consider a setting where one sender can communicate with several privately informed receivers through a persuasion mechanism before the receivers play a game. We show that for any potentially randomized persuasion mechanism, under certain conditions, there is an effectively equivalent deterministic persuasion mechanism, and these two mechanisms have the same set of equilibria. We exhibit the usefulness of our result in a new application, where our technique helps to derive the optimal persuasion mechanism. Overall, this paper provides a rationale for the fact that persuasion mechanisms are often deterministic in practice.

Abstract: This paper studies pure strategy perfect and proper equilibria for games with non-atomic measure spaces of players and infinitely many actions. A richness condition (nowhere equivalence) on the measure space of players is shown to be both necessary and sufficient for the existence of such equilibria. The limit admissibility of perfect and proper equilibria is also proved.  

Abstract: This paper studies Bayesian games with general action spaces, correlated types and interdependent payoffs. We introduce the condition of ``decomposable coarser payoff-relevant information'', and show that this condition is both sufficient and necessary for the existence of pure-strategy equilibria and purification from behavioral strategies. As a consequence of our purification method, a new existence result on pure-strategy equilibria is also obtained for discontinuous Bayesian games. Illustrative applications of our results to oligopolistic competitions and all-pay auctions are provided.