Ordinal versus cardinal voting rules: A mechanism design approach

Abstract

We consider the performance and incentive compatibility of voting rules in a Bayesian environment: agents have independent private values, there are at least three alternatives, and monetary transfers are prohibited. First, we show that in a neutral environment, meaning alternatives are symmetric ex-ante, essentially any ex-post Pareto efficient ordinal rule is incentive compatible. Importantly, however, we can improve upon ordinal rules. We show that we can design an incentive compatible cardinal rule which achieves higher utilitarian social welfare than any ordinal rule. Finally, we provide numerical findings about incentive compatible cardinal rules that maximize utilitarian social welfare.