Communication In Global Games: Theory and Experiment


Abstract

Global games are used to model a variety of environments including speculative attacks, debt crises, bank runs, and political protests. I study the effects of introducing communication and information sharing in these environments. In this paper I characterize the equilibrium for these communication protocols and then test their theoretical predictions in the lab. I find that allowing cheap-talk communication induces a unique informative partially revealing equilibrium, which drastically reduces the inefficiency region present in global games. Communication is welfare improving in two distinct ways: (i) the payoff dominant equilibrium is selected for a larger interval of states, and (ii) there is less waste of opportunity cost since players' actions are more correlated.

A Road to Efficiency Through Communication and Commitment

(with João Ramos)

Abstract | Paper

We examine the efficiency gains of introducing a pre-play phase—allowing agents to communicate their intentions and commit to them—in a game with Pareto ranked equilibria. We focus on a game in which a Pareto inferior equilibrium is usually chosen. We first derive the theoretical conditions under which the efficient equilibrium is unique in the extended game and then we test our theory in the lab. The introduction of the pre-play revision phase increases the coordination on the Pareto dominant equilibrium, restoring over 50% of the efficiency lost in the standard setting. The results shed new light on cheap talk and reveal that a combination of communication and commitment leads to significantly higher welfare.

Attention In Games: An Experimental Study

(with Andrew Schotter)

Abstract | Paper

A common assumption in game theory is that players concentrate on one game at a time. However, in everyday life, we play many games and make many decisions at the same time, and, thus, we have to decide how best to divide our limited attention across these settings. In this paper we ask how players solve this attention-allocation problem and how their decision affects the way players behave in any given game when that game is viewed in isolation. We find that the attention of players is attracted to particular features of the games they play: the maximum payoff in the game, the minimum payoff, the degree of inequality in the game’s payoff, whether the game has zero payoffs, the complexity of the game, and the type of game being played. Moreover, how much attention a subject gives to a particular game depends on the other game that he or she is simultaneously attending to.

Paying For Inattention

(with Giorgia Romagnoli)

Abstract | Slides

We extend the model of costly information acquisition to the case where a decision maker is able to affect his own incentives to pay attention via an ex-ante redistribution of payoffs across states. We use this framework to derive a novel method to elicit the level and the cost of attention solely from observing the choices of payoff redistribution. While existing work in this literature typically involves enriched datasets (e.g., mouse and eye tracking, response times, stochastic choice data etc.) our method requires only standard choice data. We run an experiment and estimate the level and the cost of attention using our elicitation technique.