Abstract: We study coordination in a population of agents with heterogeneous ideal points who interact in local neighborhoods. How do individuals' idiosyncratic preferences and their interaction structure
Abstract: We study coordination in a population of agents with heterogeneous ideal points who interact in local neighborhoods. How do individuals’ idiosyncratic preferences and their interaction structure determine aggregate welfare? How does a planner optimally shape incentives to mitigate the costs of miscoordination? In contrast, how can an adversary optimally sow dissent to reduce a community’s welfare? We characterize the answers to these questions by analyzing the game using a principal component method. The principal component associated to the smallest eigenvalue of the interaction structure turns out to be crucial. We show how the welfare analysis of the coordination game emphasizes network attributes very different from those salient in existing analyses of polarization. Finally, we generalize the spectral approach to other types of games—e.g., joint production or public goods games—and to games of incomplete information. In each case, natural orderings of eigenvalues and eigenvectors determine a planner’s priorities in an intervention problem.
(Thursday) 2:45 pm - 4:00 pm
217 Julis Romo Rabinowitz