Do open lists increase turnout? Probably not, but they increase rates of voter error: New evidence from Spain Leonardo Carella Abstract This article challenges the claim that open-list systems are beneficial for electoral participation, by reassessing and extending the analysis in a notable empirical paper that advances this argument. The paper (Carlos Sanz, “The effect of electoral systems on voter turnout: Evidence from a natural experiment”, PSRM, 2017) leverages a population-based discontinuity in Spanish municipal elections (1979–2011), where towns with fewer than 250 residents employ open lists whereas larger towns employ closed lists. Through a series of statistical tests and the inspection of alternative data sources, I show that the positive effect of open lists on turnout estimated in the paper is dubious, for two reasons: (1) non-random missing data, due to inconsistencies in how non-valid votes were recorded above and below the threshold, and (2) compound treatment issues, due to changes in list-length requirements at the threshold. I then proceed to show that, rather than improving turnout, the more complex open-list ballot actually hinders voters’ ability to express their preferences, by increasing the incidence of voter errors relative to closed lists (reflected in higher rates of ‘null’ voting). To support a causal interpretation of this relationship, I present evidence from the analysis of heterogeneous treatment effects, and show that a similar pattern obtains in Spanish general elections, where open and closed lists are used concurrently for the election of the country’s bicameral parliament. I conclude by discussing the implications of the analysis for implementing population-based regression discontinuities and evaluating electoral system effects.
New paper out at @electoralstudies.bsky.social.
I show that - contrary to claims that personalised electoral systems are good for participation - Open Lists have no effect on turnout relative to Closed Lists; in fact, they increase rates of voter error. 🗳️
www.sciencedirect.com/science/arti...
