A Unified Model of Voting in Different Institutional Contexts
Policy voters in national elections choose either the party whose position in policy space is closest to their own ideal point (proximity voting) or they elect a party which promises to move the status quo furthest in the preferred direction (directional voting). Mixed models take both criteria into account. Henning, Hinich and Shikano (2007) developed a unified model of outcome orientated voting which derives the relative weight of the two criteria (the compositional parameter β) from a formal model of the legislative process. The β-parameter is conceived as a measure of a single voter’s influence on the policy outcome in the next legislative period. The larger this influence, the higher the value of β, that is the higher the relative weight of proximity voting versus directional voting. The aim of the project was the derivation of hypotheses from this unified model for voting behaviour in various institutional contexts and their empirical test by secondary analyses of national election studies. Results: Generally speaking, a single voter will have a higher impact on legislative outcomes in democracies characterized by a high concentration of power than in more fragmented systems. Voters in parliamentary systems with two disciplined parties, plurality elections and a single legislative chamber should, therefore, vote more according to the proximity criterion than voters in consensus democracies with proportional election, a multiparty system and two chambers. Henning et al. (2007:34-35) computed the β-parameters for Belgium at the most fragmented pole, Germany, Canada and New Zealand before the 1996 electoral reform as examples of increasing power concentration and found indeed the expected increasing values of β. The β-parameters can also be computed for single parties so that hypotheses can be tested about the higher relative weight of the proximity criterion for large as compared to small parties. This hypothesis was falsified when applied in a comparative context. But Herrmann (2008) could show for the German federal election of 2005 that voters’ party preferences are formed according to proximity principles whereas directional criteria gain importance when electing a candidate or party. Voters tend to elect more extreme parties than justified by their own ideal point on a left-right scale. In Germany 2005, this mechanism benefitted the FDP at the cost of CDU/CSU and the Linke at the cost of the SPD.