Electoral Systems and the Cost of Campaigns

• In Event: Comparative Electoral Institutions

Abstract

There are reasons to believe that preferential voting, by allowing for intra-party competition, drives up the cost of electoral campaigns, but the evidence on this hypothesis is sparse and does not focus on the true parameter of interest: the marginal cost of votes. Estimating the marginal cost of votes is not trivial, since there are many factors that can simultaneously affect candidates’ campaign spending and electoral performance. An important one is candidate quality. High quality candidates are likely to raise and, thus, spend a lot in their campaigns, but also to perform well at the polls regardless of how much they spend. Thus, failing to account for candidate quality may lead to downward biased estimates of the marginal cost of votes. In order to address this issue and get unbiased estimates of the marginal cost of boats across electoral systems, I explored the adoption of a corporate donation ban in Brazil. Brazil is also an interesting case because municipal elections use two different electoral systems, plurality voting and open-list proportional representation, whereas only the latter involves preferential voting. Using difference-in-differences and instrumental variable regression I find that the marginal cost of the votes in open-list proportional representation elections is more than twice the one in plurality voting races. Using a similar approach I also show that larger district magnitudes, which are thought to intensify intra-party competition, drive up the marginal cost of votes. My findings give credence to critics of extreme forms of preferential voting systems, which point to the disproportionate role of money in elections under these systems as an important reason for abandoning them.

Author

• Pedro Vicente de Castro