Abstract: We analyze many-sender Bayesian persuasion games where receiver can contract their action on senders' choices of Blackwell experiment and realized signals. We characterize robust mechanisms: those that induce the same joint distributions of actions and senders' private types for all possible prior beliefs. These mechanisms take a simple form: they (1) incentivize maximally informative experiments, and (2) depend only on the induced posterior. We prove a grim-trigger-like principle for unilateral deviations, and show robustness is pinned down by a finite set of linear inequalities. We apply our model to allocation problems with either type-dependent outside options or allocation externalities and identify conditions where the efficient outcome is robust. Finally, we uncover a novel justification for deferred acceptance via informational incentives in two-sided matching when school preferences depend on students' unknown ability.

Abstract: I study a repeated binary-action supermodular game with endogenous exit where many short-lived agents attempt to coordinate a revolt against a regime.  The regime undertakes costly actions to increase the short-run players' coordination frictions, though acts only after if the revolt is unsuccessful, inducing a lack-of-commitment problem. In the complete-information repeated game, a folk theorem holds, with payoff multiplicity arising due to both the regime's dynamic incentives and agents' stage-game strategic complementarities. Neither the regime's reputational incentives nor belief dispersion among agents (via global-games type uncertainty) alone meaningfully refine the equilibrium payoff set. Together, though, the interaction between these two forces uniquely select the regime's highest payoff in equilibrium. Furthermore, under a Markov refinement, they select a unique equilibrium where the regime plays their optimal commitment action. Methodologically, I develop tools to analyze repeated games with endogenous exit where the regime's commitment action flexibly varies with their discount rate.

Abstract: What is the optimal order in which a researcher should submit their papers to journals of differing quality? I analyze a sequential search model without recall where the researcher's expected value from journal submission depends on the history of past submissions. Acceptances immediately terminate the search process and deliver some payoff, while rejections carry information about the paper's quality, affecting the researcher's belief in acceptance probability over future journals. When journal feedback does not change the paper's quality, the researcher's optimal strategy is monotone in their acceptance payoff. Submission costs distort the researcher's effective acceptance payoff, but maintain monotone optimality. If journals give feedback which can affect the paper's quality, such as through referee reports, the search order can change drastically depending on the agent's prior belief about their paper's quality. However, I identify a set of assortatively matched conditions on feedback such that monotone strategies remain optimal whenever the agent's prior is sufficiently optimistic.

Description: A non-Bayesian principal interacts repeatedly with one of many potential experts in each period. The expert privately observes the true state of the world and makes a recommendation which is taken at face value by the principal. The expert's preferences are not known to the principal, but is gradually revealed as time goes on. When the principal has the ability to "fire" the expert and re-contract with a new agent, what is the expert's optimal recommendation strategy? Under what circumstances does the principal fire the agent, and what is the value of commitment to the principal?

Abstract: We study reputation formation where a long-run player repeatedly observes private signals and takes actions. Short-run players observe the long-run player's past actions but not her signals. The long-run player can thus develop a reputation for playing a marginal distribution over actions, but not necessarily for playing a particular mapping from signals to actions. Nonetheless, we show that the long-run player can ensure her Stackelberg payoff if the Stackelberg strategy is confound-defeating. This property holds if the Stackelberg strategy in the unique solution to an optimal transport problem. If the long-run player's payoff is supermodular in one-dimensional signals and actions, the Stackelberg strategy is confound-defeating if it is monotone. As an application of our results, we provide a reputational foundation for Bayesian persuasion when the optimal disclosure policy is a stochastic linear partition. 

Description: We study the optimal design of rideshare platforms who jointly control the payout of a trip (transfers) and the information a driver receives before they choose to accept a ride (information). We first characterize the optimal static problem, and show how even though the principal only ever uses pure information and pure transfers at any one prior, their indirect utility function balances both as receiver's prior about the quality of a ride varies. Second, we show how when the platform has dynamic incentives - that is, the rideshare and the driver repeatedly interact about trips of heterogenous prior quaity - the optimal contract has a "credit score" scheme, mirroring the incentive schemes of several prominent rideshare platforms. 

Description: In static communication games with naive receivers, expert senders can obtain maximal information rents because receivers cannot debias their sophisticated signals. When these interactions happen repeatedly, though, receivers can leverage future participation to discipline senders to fully reveal the state. 

Description: Are riots the "language of the oppressed?" When the government can refuse to bargain with protestors and thus keep all surplus, the threat of rioting, or costly destruction of total surplus, may result in welfare-improving equilibrium. The optimal structure of rioting is discontinuous and nonmonotone  in the protestor's bargaining strength, mirroring qualitative evidence on the sudden emergence of disruptive protests following public events. 

Description: Recent work on the complexity of polynomial systems on measure preserving transformations has presented a unified approach to obtaining mean ergodic theorems for complicated dynamical systems. We reviews this complexity structure and show show these arguments provide new, independent proofs of several classical results.