Abstract
This paper reassesses the way in which the voting behaviour of individual members of the European Parliament (EP) is influenced by their national and European party delegations. It does so within the previously used framework of one agent with two principals. First, it shows that the previous literature was unable to fully fathom the mechanisms through which the tripartite principal–agent relationship works. Second, it develops a model that looks solely at the votes contested between the European and the national group and the results of the test correct many of the findings of the previous literature. The paper also develops a new theoretical framework of vote cohesion in which the national and European groups are motivated by group norms and external incentives. Finally, the analysis of roll-call votes from the sixth EP finds that the new members from Central and Eastern Europe are more likely to stay with their European group than the members from Western Europe.
Acknowledgements
An earlier version of the paper was presented at the Midwest Political Science Association Paper Conference, 2008. The author would like to give special thanks to Gary Marks for useful methodological advice and numerous readings of earlier versions, as well as to the following people who gave their valuable comments on previous drafts of the paper: Lisbet Hooghe, Amie Kreppel, Jonathan Kropko, Marius Radean, Georg Vanberg, Simon Weschle, the participants at the Comparative Working Group at the University of North Carolina at Chapel Hill, and two anonymous reviewers from West European Politics. Finally, the collection of the data was done with the help of a summer grant from the Center for European Studies at University of North Carolina.
Notes
There are two differences between the two analyses. First, Hix et al. (Citation2007) include measurements of ideological differences between the individual MEP and their NPG and EPG. Due to lack of individual data, these variables were dropped. Second, in the current analysis the variable counting the years of experience was replaced by a dummy variable that differentiates between the MEPs that are at their first term and the MEPs who served previously.
2. The initial intention was to include all MEPs affiliated with an EPG and a NPG. However, party data were unavailable for five countries: Belgium, Cyprus, Estonia, Lithuania, Luxembourg and Malta. Also, only the MEPs from national party groups of at least three members were included in the OLS regression of defection from the NPG.
To test this, I regressed the variable ‘degree of defection from the EPG’ on the variable ‘degree of defection from the NPG’, and obtained a highly significant coefficient (t-value >200).
4. GAL stands for Green-Alternative-Libertarian, and TAN stands for Traditional-Autocratic-Nationalist. The introduction of this new dimension is more appropriate after the EU's eastward enlargement. Marks et al. (Citation2006) find that for the parties in Western Europe there is a natural relationship between GAL-economic left and TAN-economic right. This relationship however is completely different for the parties from CEE, where one encounters a GAL-Right and a TAN-Left relationship.
5. Similar levels of EPG voting cohesion were computed with the database used here for the first three semesters after EU expansion. The conclusion derived is that the expansion has not decreased the level of cohesion at the EPG level.
6. For a discussion about how the political leaders from the new EU members seek to be recognised as equals see Vachudova (Citation2005: 248).