RCV, STV, and Party Systems: A New Scaling Method on Ranked Ballots (Pre-Recorded)

• In Event: Reactions to Ranked-Choice Voting

Abstract

Single-winner ranked-choice voting has been an important topic in discourse around electoral reform in the past decade, with several campaigns to expand it at the state level. However, this has been shown to have several drawbacks, including lower voter confidence and lower voting satisfaction, with only a slight increase in non-major party voting. One proposed improvement is to implement multi-winner ranked-choice voting, which is referred to as the “single transferable vote”, a system of proportional representation which utilizes ranked ballots which may provide better representation than both single-winner plurality elections and single-winner RCV elections. In this study, I demonstrate a new scaling algorithm that better incorporates missing data and geographic covariates in order to estimate the ideal points of both candidates and voters using ranked-choice ballots. I then use this new technique on election returns in the city of Cambridge, Massachusetts, which has used STV since the 1940s. In doing so, I demonstrate that while Cambridge ideology can be seen as bipolar, STV better represents the underlying population of voters than both single-winner RCV and plurality elections.

Author

• Jesse Clark, Princeton University