Supplement to "A Theory of Non-Bayesian Social Learning"

This online supplement contains three parts. Appendix B contains several proofs omitted from the main body of the paper. Appendix C extends the concept of unanimity introduced in Definition 1 and provides a generalization to Theorem 4 in the paper. Appendix D illustrates that if signals are normally distributed and agents’ prior beliefs are normal, Bayesian updating takes a log-linear form regardless of the structure of the underlying social network.

Supplemental Authors: 
Molavi, Pooya - Northwestern University
Tahbaz-Salehi, Alireza - Northwestern University
Jadbabaie, Ali - MIT
Online Appendix