Interest groups and majority voting
Description
One of the tasks public choice theorists have attempted to accomplish is the building of a theory of large majority rule elections which is based upon rational behavior and is non-vacuous. Two voting paradoxes have generally stood in the way of the development of such a theoretical framework: (1) the instability of the pure majority rule system and (2) the large turnout of voters despite the probable lack of a rational basis for voting. The purpose of this paper is the formulation of a model of majority voting incorporating the behavior of groups in the voting process which can explain large majority rule elections despite these two paradoxes Chapter I is a review of the literature concerning these two paradoxes with particular attention to previous theoretical attempts toward resolution of each. A formulation of the cost-benefit analysis of voting using interest groups is presented in Chaper II. Constituents maximize expected utility which depends not only upon which candidate wins, but also to what degree the constituent's group voted for the winning candidate. The choice of how many votes to provide is made at the group level where the benefits to the group are compared with the costs of voting. Each group is assumed to have its own production function of votes whereby it enforces its decision upon its members. Uncertainty is introduced in the model through the random influences each group faces in enforcing its voting decision. From the group perspective there is shown to be a 'rational' basis for voting In Chapter III the model is expanded to consider probability maximizing candidates with control over public good levels. It is shown that if constituents are risk averse in public good preferences, then a unique symmetric (both candidates offering identical platforms) exists. In a two group example it is demonstrated that this equilibrium will not necessarily represent the median of voter preferences. Chapter IV analyzes the theoretical and empirical implications for future research of the results of the model