Party, candidate, and voter incentives under free list proportional representation

Abstract

Under free list proportional representation voters can: (a) cast preference votes for candidates; (b) cast multiple preferences; and (c) distribute preferences across multiple lists. Alternatively, they can cast a list vote. Our theory shows that office-seeking candidates face incentives to pursue the personal vote, while non-candidate partisans seek the party vote. Voters are in the cross-currents of these forces. Also, since preference voting is so cognitively and informationally demanding, voters have incentives to use shortcuts, especially (a) list voting; (b) casting fewer than their allotment of preferences; and (c) preference voting for well-known or highly placed candidates. We find support for our expectations using linear mixed-effects regression of the proportion of preference votes in candidate-level electoral data from Ecuador. Personal voting is more prevalent as magnitude increases, where the local party is strong, and for candidates that are incumbents, male, high on the list, and in the position of first loser.

Challenging the wisdom on preferential proportional representation

“…Politicians and parties behave as if they anticipate that voters will behave in ways comporting with this assumption (Cheibub and Sin, 2020; Folke and Rickne, 2020). At the same time, the project spotlights the cognitive challenges that voters face (Mustillo and Polga-Hemicovich, 2020), and, what is more, uncovers the elusive, even illusory, notion of sincere voting in any MMD system (Buisseret and Prato, 2020).…”

“…Mustillo and Polga-Hecimovich (2020) look at free list PR, a design under which the number of optional preferences to express hinges on district magnitude. The authors highlight that the free list challenges voters, with cognitive and informational limitations looming ever larger as district magnitude rises.…”

“…For instance, whereas Carroll and Nalepa (2020) define cohesion as preference alignment, Folke and Rickne (2020) use a trio of measures for cohesion, two tapping the politician’s preference alignment with her party, and one, her ideal voting behavior. Among the analyses here, only Folke and Rickne (2020) and Mustillo and Polga-Hecimovich (2020) treat electoral rules as endogenous to the political process (cf., among many, Benoit, 2007; Boix, 2010; Cusack et al, 2010; Kreuzer, 2010; Mares, 2015; Mershon, 1994; 2002: ch. 4; 2016).…”

“…4; 2016). Whereas Mustillo and Polga-Hecimovich (2020) emphasize the cognitive challenges entailed in casting preference votes, only Buisseret and Prato (2020) go so far as to reject the assumption that voters can cast sincere votes under any multi-member district design.…”

“…Consider too that almost all of the project’s participants envision a scenario in which the number of preference votes that a given voter may cast—and hence the number of parties she may support—is one. Mustillo and Polga-Hecimovich (2020) present an exception, as they analyze a context in which the number of preference votes ranges from one to a given district’s magnitude, and the voter is free to cast preferences for as many parties as there are in her district. In still another design, not examined in this project, the voter has the option of casting one or several preferences and is constrained to cast any preference(s) for a single party.…”