arXiv subjects

Emiliano Catonini

Publications and source records attributed to Emiliano Catonini.

5 recordsLinked to original sources

arXiv2023

Cautious Belief and Iterated Admissibility

We define notions of cautiousness and cautious belief to provide epistemic conditions for iterated admissibility in finite games. We show that iterated admissibility characterizes the behavioral implications of "cautious rationality and common cautious belief in cautious rationality" in a terminal lexicographic type structure. For arbitrary type structures, the behavioral implications of these epistemic assumptions are characterized by the solution concept of self-admissible set (Brandenburger, Friedenberg and Keisler 2008). We also show that analogous conclusions hold under alternative epistemic assumptions, in particular if cautiousness is "transparent" to the players. KEYWORDS: Epistemic game theory, iterated admissibility, weak dominance, lexicographic probability systems. JEL: C72.

econ.TH↗

arXiv2023

Local Dominance

We define notions of dominance between two actions in a dynamic game. Local dominance considers players who have a blurred view of the future and compare the two actions by first focusing on the outcomes that may realize at the current stage. When considering the possibility that the game may continue, they can only check that the local comparison is not overturned under the assumption of "continuing in the same way" after the two actions (in a newly defined sense). Despite the lack of forward planning, local dominance solves dynamic mechanisms that were found easy to play and implements social choice functions that cannot be implemented in obviously-dominant strategies.

econ.TH↗

arXiv2023

On the optimality of full disclosure

A privately-informed sender can commit to any disclosure policy towards a receiver. We show that full disclosure is optimal under a sufficient condition with some desirable properties. First, it speaks directly to the utility functions of the parties, as opposed to the indirect utility function of the sender; this makes it easily interpretable and verifiable. Second, it does not require the sender's payoff to be a function of the posterior mean. Third, it is weaker than the known conditions for some special cases. With this, we show that full disclosure is optimal under modeling assumptions commonly used in principal-agent papers.

econ.TH↗

arXiv2022

A Dutch-Book Trap for Misspecification

We provide Dutch-book arguments against misspecified Bayesian learning. An agent progressively learns about a state and is offered a bet after every discovery. We say the agent is deterministically Dutch-booked when they would accept all bets, but their payoff is ex-post negative under each state. More generally, we say that the agent is Dutch-booked when they would accept all bets, but their expected payoff under each fundamental state is negative. With this, the agent cannot be deterministically Dutch-booked if and only if they update their beliefs using Bayes' rule, even with misspecified likelihoods. The agent cannot be Dutch booked if and only if they update their beliefs using Bayes' rule with a lexicographic prior and using the correct data-generating process. We show that offers of financial instruments and behavior in Monty Hall problems can be viewed as Dutch books that extract a sure expected gain from a misspecified population.

econ.TH↗

arXiv2017

Rationalizability and Epistemic Priority Orderings

At the beginning of a dynamic game, players may have exogenous theories about how the opponents are going to play. Suppose that these theories are commonly known. Then, players will refine their first-order beliefs, and challenge their own theories, through strategic reasoning. I develop and characterize epistemically a new solution concept, Selective Rationalizability, which accomplishes this task under the following assumption: when the observed behavior is not compatible with the beliefs in players' rationality and theories of all orders, players keep the orders of belief in rationality that are per se compatible with the observed behavior, and drop the incompatible beliefs in the theories. Thus, Selective Rationalizability captures Common Strong Belief in Rationality (Battigalli and Siniscalchi, 2002) and refines Extensive-Form Rationalizability (Pearce, 1984; BS, 2002), whereas Strong-$Δ$-Rationalizability (Battigalli, 2003; Battigalli and Siniscalchi, 2003) captures the opposite epistemic priority choice. Selective Rationalizability can be extended to encompass richer epistemic priority orderings among different theories of opponents' behavior. This allows to establish a surprising connection with strategic stability (Kohlberg and Mertens, 1986).

cs.GT↗