Working Papers

We study a dynamic principal-agent setting in which both sides learn about the importance of effort. The quality of the agent’s output is not observed directly. Instead, the principal jointly designs an evaluation technology and a wage schedule. More precise performance evaluation reduces current agency costs but promotes learning, which is shown to increase future agency costs. As a result, the optimal evaluation technology is both imprecise and tough: a bad performance is always sanctioned, but a good one is not always recognized. We also study the case in which principal and agent have different priors, for instance because the agent has incorrect beliefs about his abilities. If the agent is overconfident, the principal uses a tough evaluation structure to preserve the agent’s profitable misperception. For an underconfident agent, by contrast, she either uses a fully informative evaluation in order to promote learning and eliminate costly underconfidence, or is lenient if learning is too costly.

R&R The Review of Economic Studies

We analyze screening with frame-dependent valuations. A monopolist principal designs an extensive-form decision problem with frames at each stage. This allows the firm to induce dynamic inconsistency and thereby reduce information rents. We show that the optimal extensive form has a simple three-stage structure and uses only the two highest frames (high-low-high). Some types buy in the first stage, while others continue the interaction and buy at the last stage. The principal offers unchosen decoy contracts. Sophisticated consumers correctly anticipate that if they deviated, they would choose a decoy, which they want to avoid in a lower frame. This eliminates incentive compatibility constraints into types who don’t buy in the first stage. With naive consumers, the principal can perfectly screen by cognitive type and extract full surplus from naifs.

Screening while Controlling an Externality, with Elia Sartori

We propose a tractable framework to introduce externalities into a monopolist screening model. Agents differ both in their payoff type and their influence, i.e. how strongly their action affects the aggregate externality. Applications range from non-linear pricing of a network good, to taxation or subsidization of industries that produce externalities (e.g. pollution and human capital formation). When both dimensions are unobserved (full screening) the optimal allocation satisfies lexicographic monotonicity: within a payoff-type, the monopolist optimally tilts the allocation towards influential agents to increase the externality, while standard IC drives monotonicity across payoff-types. We characterize the solution through a two-step ironing procedure that addresses the nonmonotonicity in virtual values arising from the countervailing impact of payoff-types and influence. The allocation is inefficient if and only if the payoff-type is unobservable. Only influence is observable, equilibrium utility can vary across the latter as it is used as a signal of the payoff-type. We provide sufficient conditions for (expected) rents from influence to emerge.

Acquisition, (Mis)use and Dissemination of Information: The Blessing of Cursedness and Transparency, with Elia Sartori

This paper studies strategic interactions where players observe statistics of others' actions, focusing on: First, the endogeneity of the precision of such aggregate information as signals of the fundamental; and second, agents' well-documented difficulty in making inference based on such signals. We conduct our analysis in a beauty contest game with information acquisition, adapting cursed equilibrium to model agents limited ability to process aggregative information. To discipline information acquisition choices in this setting with incorrect information use, we define a novel notion of cursed expectations equilibrium with information acquisition: Agents assess the value of private information according to a subjective envelope condition, as they correctly anticipate their actions and (incorrectly) deem them optimal. We show that there is inefficiently low acquisition and use of private information in the rational benchmark due to an information dissemination externality. Despite suboptimal use, cursed agents rely more heavily on their private information which pushes information acquisition towards its efficient level and causes an initial increase in welfare.Transparency crowds out private information but always increases the endogenous precision of the aggregative signal and welfare, while other policy instruments can have paradoxical effects due to their interaction with cursedness. Finally, we explore the behavior and welfare of an atomistic rational agent playing against a cursed crowd and demonstrate that transparency may be an elitist policy.

Hidden Costs: Anticompetitive Effects of Transparency

We identify a market characteristic central to the impact of the exploitation of consumer naïvete about hidden costs on profits, incentives to increase price transparency and the impact of transparency on consumer surplus: the difference between cross-price elasticities of demand of naïve and sophisticated consumers. Exploiting consumer naïvete increases profits and persist in the long run only if naïfs are more price sensitive than sophisticates. Targeting of transparency matters. Increased transparency is most valuable to consumers when firms have the least incentives to engage in it, can increase aggregate welfare, but typically redistributes surplus from consumers to firms.


„A note on consequentialism in a dynamic Savage framework: a comment on Ghirardato (2002).“ Economic Theory Bulletin, 2015, 3:2, 265-269, with Joachim Hubmer