New fairness criteria for truncated ballots in multi-winner ranked-choice elections

TL;DR

This paper introduces new fairness criteria for multi-winner ranked-choice elections with truncated ballots, analyzing how different voting methods perform under these criteria, highlighting the Chamberlin-Courant rule's robustness.

Contribution

It proposes novel fairness criteria tailored for partial ballots and evaluates popular voting methods against these standards, revealing their relative strengths.

Findings

Chamberlin-Courant rule performs best on new fairness criteria.

Expanding approvals rule performs the worst.

Single transferable vote has intermediate performance.

Abstract

In real-world elections where voters cast preference ballots, voters often provide only a partial ranking of the candidates. Despite this empirical reality, prior social choice literature frequently analyzes fairness criteria under the assumption that all voters provide a complete ranking of the candidates. We introduce new fairness criteria for multiwinner ranked-choice elections concerning truncated ballots. In particular, we define notions of the independence of losing voters blocs and independence of winning voters blocs, which state that the winning committee of an election should not change when we remove partial ballots which rank only losing candidates, and the winning committee should change in reasonable ways when removing ballots which rank only winning candidates. Of the voting methods we analyze, the Chamberlin-Courant rule performs the best with respect to these criteria, the expanding approvals rule performs the worst, and the method of single transferable vote falls in between.