Coalizer: A Coalition Tool Combining Office and Policy Motivations of Political Parties

While prior work on coalition theories focuses either solely on office or on policy motivations of parties, more recent theories combine both types of motivation. They are therefore much more appropriate for explaining coalition formation but, at the same time, they are also more complicated. One possibility to make these models broadly applicable are coalition tools. However, existing tools do not incorporate advances from more recent theories and are, for example, confined to the identification of minimal winning coalitions. In this paper, we present a new coalition tool called Coalizer which takes both office and policy motivations into account and reflects the state of coalition theory. As data input, Coalizer needs the seat distribution in a parliament as well as information on party positions. While the latter can come from various sources, Coalizer is explicitly targeted at multi-issue data as it is provided by voting advice applications for many elections. Coalizer is, to the best of our knowledge, the most comprehensive coalition tool and includes, among others, the following features: - identification of (minimal) winning coalitions, - computation of office utility values basing on seat distributions, - computation of policy utility values basing on party positions, - visualization of party distances, - a weighted combination of office and policy utility values, and - showing utility maximizing strategies for parties and equilibria based on the combined utility values. In our paper, we i) sketch the recent coalition theories on which Coalizer is based, ii) describe the functionality of our coalition tool, and iii) evaluate Coalizer at the example of the German federal elections in 2017.