Title: Incentive and Informational Spillovers in Networks

Abstract

Social and economic interactions are embedded in networks of relationships. Network contacts affect both our incentives --- e.g., to get educated, to vote, to participate in online platforms --- and our information. When network contacts change one's motivations (e.g., motivating someone to share a political message to gain approval) that is an incentive, or strategic spillover. When they change one's information (e.g., making someone believe a piece of false news) that is an informational spillover. This talk surveys some recent models relevant to these phenomena and what they imply for a planner or adversary interested in intervening in the system. We present a new principal component decomposition of incentive spillovers, and use it to find the nudges that best leverage the existing strategic effects. Turning to informational spillovers, we discuss some new, tractable models of dynamic learning by "decentralized statisticians." We discuss some asymptotic results on when communities can learn well and when they are prone to being persistently misled. The talk concludes with some conjectures and open problems.

Bio

Benjamin Golub's research in economic theory focuses on social and economic networks. Since 2015, Ben has been on the faculty at the Harvard Department of Economics, where he is now Associate Professor; prior to that he spent two years as a Junior Fellow (2013-15) at the Harvard Society of Fellows. He has received the NSF CAREER grant (2019) and the Aliprantis Prize (2012) with Matt Elliott for an outstanding paper by a young economic theorist. He was educated at Stanford (Ph.D. in economics) and Caltech (B.S., mathematics).

Ben's work has examined:

- Learning and gossip: what are the dynamics of information in networks? Who is particularly influential? When do agents learn correctly? When are their beliefs polarized?

- Coordination in organizations: how do strategic agents in complex organizations coordinate, and how do the networks that connect them matter?

- Financial contagion: which financial networks are particularly sensitive to sudden breakdowns?

- Public goods and externalities: in complex favor-trading problems such as pollution reduction, we can use a network perspective to understand how much scope for cooperation there is, who is particularly essential to it, and how to organize negotiations.