Research & Projects

Abstract: 

This paper proposes a foundation for ambiguous beliefs in strategic interactions using the von Neumann-Morgenstern approach to ambiguity. A conjecture is a simplex isomorphic to the mixed action set, and it can be interpreted as a multiple-prior belief. "Equilibria in Ambiguous Beliefs" and "Ambiguous Nash Equilibria'' are defined and characterized in finite games and games with compact-convex action sets. Further characterization and comparative statics results are given for 2x2 games.

JEL Codes: C72, D81, D84.

Abstract: 

In this paper, I analyze three related topics around ambiguous information, applying the von Neumann-Morgenstern approach to ambiguity. First, I extend Bayes updating to account for ambiguous prior information and ambiguous data-generating processes, and describe implications for decision-makers. Then, I incorporate this framework into games with communication, analyzing games with vague cheap talk and ambiguous communication equilibria. Finally, I provide a consumption-based asset pricing model that incorporates multiple priors.

JEL Codes: C72, D81, D83, D91.

Abstract: 

I extend the notions of Correlated and Bayes-Nash Equilibria using the von Neumann-Morgenstern approach for ambiguity. Following \citet{myerson_91}'s approach to mechanism design, I define two versions of incentive compatibility under ambiguity to study general principal-agent problems with ambiguous communication.  A version of the Revelation Principle for ambiguous communication is provided. 

JEL Codes: C72, D47, D81, D82.

Abstract: 

This note introduces a notion of general direct recommendations in corporation games. A general recommendation maps a type profile to a simplex of distributions over decisions. I propose two ways to define incentive compatibility, leading to two distinct generalizations of the M-equilibrium. These partially solve the equilibrium non-existence problem. I apply these solution concepts to the games proposed in Myerson (1982) and Forges, Koessler, Salamanca (2024).

JEL Codes: C72, D81, D82.

Abstract: 

A principal with strictly convex preferences designs a rule to allocate a divisible unit of resources to N agent. To win the resources, each agent proposes a lottery over policies. Competition aligns the agents' incentives to the principal's preferences but this is insufficient to ensure the principal's first best platform, which would require coordination from the agents. In general, non-market allocation mechanisms without transfers are inefficient in this class of games. Relevant applications include allocations of public funds in the presence of ambiguity and markets with complementary goods.

JEL Codes: C72, D47, H57.

Abstract

This paper analyzes how governments interact when choosing levels of polluting gas emissions. Each government faces a domestic principal-agent problem: the producing firm has some private information value that determines its efficiency. I analyze how strategic interaction and incompleteness of information impact emissions policies and pollution in this ``game between principals''. Asymmetric information alleviates free-riding problems, lessens the negative externality effect, and ultimately reduces pollution. The paper concludes with some positive implications from the point of view of a social planner.

JEL classification: C72, D82, Q51.