Beliefs dynamics in communication networks

We study the dynamics of individual beliefs and information aggregation
when agents communicate via a social network. We provide a general
framework of social learning that captures the interactive effects of
three main factors on the structure of individual beliefs resulting from
such a dynamic process; that is historical factors—prior beliefs,
learning mechanisms—rational and bounded rational learning, and the
topology of communication structure governing information exchange. More
specifically, we provide conditions under which heterogeneity and
consensus prevail. We then establish conditions on the structures of the
communication network, prior beliefs and private information for public
beliefs to correctly aggregate decentralized information. The speed of
learning is also established, but most importantly, its implications on
efficient information aggregation.

Keywords: Learning, social networks, public beliefs, speed of learning,
information aggregation.