Does universalization ethics justify participation in large elections?

Why would voters incur costs to cast votes they believe are inconsequential? We show that a form of universalization ethics can help explain this "paradox of costly voting". The political model has two candidates, a continuum of voters, a known underdog, and two novel elements: core voters (who always vote), and distinct election stakes for the partisan groups. Elections are modeled as power-sharing institutions, winner-take-all being a limit case. Cost-sensitive voters, whose turnout behavior we analyze, consider the election outcome, should their decision be universalized to other voters. The "others" can be limited to co-partisans ("partisan ethics") or not ("non-partisan ethics"). By contrast to most of the literature, we do not impose conditions ensuring existence and uniqueness. The universalization ethics often fosters participation, but coordination problems remain. Equilibrium can fail to exist under partisan ethics, and equilibrium multiplicity can arise under both partisan and non-partisan ethics. Equilibria where the underdog wins may co-exist with equilibria where the topdog wins. The former are sustained because of a boost of cost-sensitive voters' willingness to vote, stemming either from a large core constituency or from a high stake. In this case, the core constituency is a complement rather than a substitute for the cost-sensitive voters.