Measures and Models of Legislative Behaviour in the German Bundestag

In analyzing legislative behavior and more specifically, the effects of the institutional setting on parliamentary party group discipline, models utilizing roll-call data have been met with undifferentiated criticism. I propose a model which addresses the persisting, yet neglected problems which involve the undifferentiated use of measures for party unity, cohesion and discipline as well as the absence of a general consensus concerning the appropriate modeling of legislative voting behavior. Taking into account the lack of direct measures for discipline due to its elusive nature, any model attempting to explain party group behavior must at least incorporate a direct measure of party cohesion and unity. In order to meet this requirement, the model I propose makes use of new survey data which allows controlling for preference homogeneity. I also argue that the researcher’s view should be shifted back to the central arena of the legislative process, which becomes manifest in legislative voting; for many parliaments roll-call data remain the closest measure of party unity. Finally, the policy field in which the vote is being held will be taken into account. On the methodological side, the use of individual level data in a hierarchical setting will be discussed; the nesting of votes within policy fields as well as that of MPs within party groups is considered. These theoretical propositions are translated into a model explaining legislative behavior in the 16th German Bundestag. Due to its mixed-member system, the Bundestag has received much attention. Consolidated findings of existing research as well as availability of topical survey and roll-call data of the Bundestag establish the ideal basis for a revision of models on legislative behavior. This within country comparative study will provide the foundation for a broader study based on the project “Parliamentary Party Groups in Comparison” by the University of Zurich.