Party Cohesion as Outcome of MP Cooperation

Party cohesion is hard to assess. Recently, scholars turned away from studying party cohesion purely based on legislative behavior to a broadened approach studying members of the parliament by employing surveys. Role call voting bears far too little variance, among other serious shortcomings (Sieberer 2006). To overcome these obstacles, scholars turned to surveying MPs directly (Andeweg & Thomassen 2011). This enabled progress in form of new theoretical and empirical insights, the now famous 'Pathways to Party Unity'. But surveys are problematic in its own regard (Close & Lopez 2013). I propose to complement recent advances employing both a new source of data, as well as furthering the use of relational analysis in the field. The proposed measurement of party cohesion is based on the observable behavior of party members in the parliament, the cosponsorship of bills. Assessing the homogeneity of cosponsorship in a time series is used to quantify cohesion, based on a model of cooperation. The analysis will also account for effects of MP specialization (see Louwerse & Otjes 2015). This design is tested on a new dataset (German Bundestag). The preliminary analysis yields strong evidence that cosponsorship shows episodical disagreement on a baseline that otherwise displays strong cohesion. The data also supports a visible distinction between government and opposition, and furthermore hints to discernible patterns on the individual level (e.g. party leadership). Using cosponsorship data not only circumvents a substantial amount of problems of contemporary studies, but also opens up new perspectives on the phenomena in general. The presented analysis shows clearly visible patterns and performs well on validity measures. Briatte (2016) showed the wide availability of cosponsorship data in a large number of parliaments, a resource that students of party cohesion should not miss.