Strategic Voting

Abstract

In a typical voting scenario, a group of voters with diverse preferences need to collectively choose one out of several alternatives. Examples include a committee that selects a candidate for a faculty position or an award, countries in an international forum voting on the adoption of a new environmental treaty, or even automated agents that vote on the preferred meeting time on behalf of their users. As the satisfaction of each voter is determined by the selected alternative, which is in turn affected by the actions (namely, the ballots) of others, casting a vote is in fact playing a strategic game. The study of strategic voting is an effort to utilize game theory, which merits to model and predict rational behavior in a wide range of economic and social interactions, to explain and even direct the strategic decisions of voters. This review paper is a hyper-condensed version of a book on strategic voting that is forthcoming this year. 1 The main purpose of the book is to overview the main approaches to strategic voting, in a way that makes these approaches comparable across fields and disciplines. In this paper I will mention the main directions and lines of work, but almost without going into the technical details. Our starting point will be the seminal Gibbard-Satterthwaite theorem, which states that under a set of natural requirements, one cannot hope to construct a voting rule that is immune to strategic manipulations by the voters. This mean that there will always be situations where some voters have an incentive to misreport their true preferences. From this strong negative result emerged two lines of research. One continues to shape the boundaries and limitations of truthful voting mechanisms, by relaxing some of the assumptions that lead to the G-S impossibility result. The other line forgoes the attempt to elicit truthful votes, and instead applies game theory and equilibrium analysis to understand how strategic voters would vote in existing mechanisms.