Role of place-of-residence in Provincial and Water Board elections

Abstract

Voting behavior is partly determined by candidate characteristics, one of which is place of residence. The friends-and-neighbors voting hypothesis posits that contenders for office garner disproportionally more support from voters in their hometowns. Voters are more likely to know of or know the candidate personally, they may believe that native candidates will defend their interests better, or they become mobilized through a sense of shared identity. While the bulk of the literature is focused on national elections in majoritarian systems such as the US or UK, this thesis project will investigate the role of the hometown advantage for the Dutch Provincial (Provinciale Staten) and Water Board (Waterschaps) elections. Specifically, it will research whether candidates receive more votes in the voting stations located in their places of residence in either election. These elections are second-order, conducted on the Provincial level, meaning that voters have a lower interest in and awareness of them than in the national elections. As voters seek out less information about candidates’ merits and stances, they may rely more heavily on characteristics such as background to make political decisions. The Dutch elections are especially suitable for such research, as candidates’ place of residence is reported on the ballot, meaning that this information is easily available. Because of the excess presence of zero-values, the analysis is split into a two-part model to investigate firstly what determines whether or not a candidate receives any votes in either election and secondly how the number of votes is influenced. I find that candidates do receive significantly more votes in voting stations located in their hometowns and that this effect is quite large and consistent between the models. Candidates who are the only ones from a certain town running on their party’s ballot receive another vote boost. Furthermore, it appears that this effect is slightly stronger in more rural areas, where community ties tend to be stronger. The roles of wealth and candidate list position on the same-town residence effect remain largely unclear. Similarly, there are no clear convincing differences in the main predictor between either election. Overall, this research offers support for the friends-and-neighbors voting hypothesis in the context of the Dutch Provincial and Water Board elections.