Single-Winner Voting with Alliances: Avoiding the Spoiler Effect
Grzegorz Pierczyński, Stanisław Szufa
TL;DR
The paper tackles spoiler effects in single-winner elections with alliances by introducing alliance-aware voting rules that use alliance structure to prevent vote-splitting among allies. It develops four rules by extending Plurality and Maximin into two rounds, backed by four basic axioms and two individual-oriented axioms (IW and SW consistency), and proves desirable properties such as majority- and Condorcet-consistency where applicable. The approach is shown to retain strong social-welfare performance and to be computationally efficient, with experiments indicating primaries alone often fail to prevent spoilers. The work also provides practical guidance on ballot design and extends naturally to nested alliances, offering a framework with real-world applicability for alliance-based electoral systems.
Abstract
We study the setting of single-winner elections with ordinal preferences where candidates might be members of \emph{alliances} (which may correspond to e.g., political parties, factions, or coalitions). However, we do not assume that candidates from the same alliance are necessarily adjacent in voters' rankings. In such case, every classical voting rule is vulnerable to the spoiler effect, i.e., the presence of a candidate may harm his or her alliance. We therefore introduce a new idea of \emph{alliance-aware} voting rules which extend the classical ones. We show that our approach is superior both to using classical cloneproof voting rules and to running primaries within alliances before the election. We introduce several alliance-aware voting rules and show that they satisfy the most desirable standard properties of their classical counterparts as well as newly introduced axioms for the model with alliances which, e.g., exclude the possibility of the spoiler effect. Our rules have natural definitions and are simple enough to explain to be used in practice.