Political Economy of Committee Voting and Its Application

Abstract

This dissertation consists of three essays on information aggregation in committees and its application. The first essay analyzes how the distribution of votes affects the accuracy of group decisions. In a weighted voting system, votes are typically assigned based on the criteria that are unrelated to the voters’ ability to make a correct judgment. I introduce an information aggregation model in which voters are identical except for voting shares. If the information is free, the optimal weight distribution is equal weighting. When acquiring information is costly, by contrast, I show that the accuracy of group decisions may be higher under some weighted majority rules than under unweighted majority rule. I characterize the equilibrium and find the optimal weight distribution to maximize the accuracy of group decisions. Asymmetric weight distributions may be optimal when the cost of improving signal is moderately high. The second essay analyzes how intergenerational family transfers can be sustained. Why are generous transfers from the younger to the older generations made in some families and not in others? My paper argues that differences in intergenerational dependence are due to variation in community networks. My analysis of the sustainability of intergenerational transfers posits game theoretical models of overlapping generations in which breadwinners make transfers to their parents and children. A novel feature of my models is that there is a local community that may supply information about its members past behaviors. I demonstrate that an efficient level of intergenerational transfers can be sustained if neighbors gossip about each other. The third essay, co-authored with Fuhito Kojima, investigates a jury decision when hung juries and retrials are possible. When jurors in subsequent trials know that previous trials resulted in hung juries, informative voting can be an equilibrium if and only if the accuracy of signals for innocence and guilt are exactly identical. Moreover, if jurors are informed of numerical split of votes in previous trials, informative voting is not an equilibrium regardless of signal accuracy.