Voting and Social Choice in Information Networks

Abstract

This thesis aims to explore the theoretical and numerical properties of a formal model of voting in which voters are allowed to exchange information. Building on work in social choice theory that deal with non-social voters, the model presented explores the question of how to set up desirable systems of social choice in situations where voting has a social component. After describing social voting as a Markov process, I go on to show the new model reduces to the case where the voters are independent and obtain a lower bound on the spectral gap constant for cases in which the model is reversible. Lastly, the thesis explores numerical results that show that, in situations where the voters are highly social, that majoritarian democracy is one of the most stable systems of social choice.