Muraoka, T. and J. Barceló (2017) “The effect of district magnitude on turnout: Quasi-experimental evidence from nonpartisan elections under SNTV”, Party Politics, DOI: 1354068817740337.
The relationship between district magnitude and voter turnout is still a debated issue in the field of research on electoral systems, as previous studies found positive, negative, and nonlinear effects. The article by Muraoka and Barceló aims at testing the direction of the effect of district magnitude on turnout with a quasi-experimental design, using data from municipal council elections in Japan.
Theoretically, as a unifying framework in consideration of the mixed empirical evidence, the authors start from the conceptualization of the relationship between district magnitude and turnout that was put forward by Grofman and Selb (2011) and especially Taagepera et al. (2014). These studies suggest a curvilinear relationship between district magnitude and turnout in form of an inverse U. District magnitude should have a positive effect on turnout up to a certain point (above 4-5 seats) by facilitating finding candidates close to their preference for voters. But after passing this threshold, district magnitude should affect turnout negatively because of the increasing informational complexity, imposing higher costs of information on voters.
The analyses rest upon data from Japanese municipal council elections, which should especially suit to test the second part of the theoretical argument. So, district magnitude is relatively high in these elections and they show two crucial characteristics. On the one hand, parties play only a marginal role, i.e. most councilors do not belong to parties. On the other hand, they apply a single nontransferable vote system (SNTV) where voters can cast only one vote for their most preferred candidate. Accordingly, SNTV encourages the individualization of electoral competition and there is a noticeable chance for voters to finally “waste” their vote if they support a losing candidate. Consequently, the absence of parties and the SNTV system in Japanese municipal council elections require voters to make significant informational efforts and lead to high electoral complexity as district magnitude increases.
The authors use a fuzzy regression discontinuity design, analyzing municipal council elections held between 2006 and 2011. For reasons of data availability and the modelling strategy the final sample includes 918 council elections, district magnitudes ranging from 6 to 46 and thus all above the turning point of the inverse U-curve following the theoretical model. In line with this, the empirical analyses find a negative effect of district magnitude on turnout. Remarkably, the effect is nonlinear: The negative effect of an increasing district magnitude on turnout is stronger in smaller districts compared to larger districts in terms of seats. In the context of SNTV and a weak role of parties, these findings seem to corroborate the hypothesis that a higher district magnitude can lead to an information overload that finally lowers voters’ incentives to go to the polls.
The article is an illuminating contribution to the literature on the effects of electoral institutions on voting behavior as the case selection and research design allow for a most direct look on the impact of the independent on the dependent variable, blending out many confounding factors that previous studies had to cope with. Understandably, the generalizability of the findings in this article will only be proven by further similar fine-grained studies. In the end, this may solve still existing puzzles and help clarifying if the proposed curvilinear relationship between district magnitude and turnout is the final answer.
Taagepera R., Selb P. and Grofman B. (2014) How turnout depends on the number of parties: a logical model. Journal of Elections, Public Opinion & Parties 24(4): 393–413.
Grofman B. and Selb P. (2011) Turnout and the (effective) number of parties at the national and district levels: a puzzle-solving approach. Party Politics 17(1): 93–117.