The distribution of power among EU institutions: who wins under codecision and why?

Abstract

The codecision procedure was designed to change the distribution of power among the European Union (EU) institutions. In theory, the codecision procedure, at least the amended version introduced by the Amsterdam Treaty that came into effect in 1999, weakened the Commission and placed the Parliament on an equal footing with the Council. We assess how the codecision procedure works in practice using data on the preferences of legislative actors on a large number of proposals negotiated between 1999 and 2009. We also test theoretical propositions derived from Schelling regarding the effects of policy agreement within each chamber on the relative bargaining success of the Council and EP. Our findings suggest that, in comparison to the consultation procedure, codecision has strengthened the EP and weakened the Commission. However, the Council holds certain bargaining advantages over the EP, and as a result the EP has not achieved parity with the Council under codecision.

Keywords:

• Codecision
• Council of Ministers
• European Parliament
• legislative bargaining

ACKNOWLEDGEMENTS

We thank Fabio Franchino, Anne Rasmussen, Andreas Warntjen and the anonymous reviewers for their comments.

Notes

Tsebelis (2002: 45–50) makes a similar point using a spatial model for a collective legislative body. His analysis indicates that when a chamber is internally divided, there is typically a greater range of alternatives to the status quo on which agreement can be found (p.50). However, this is not always the case: he also points out that there are situations in which greater internal cohesion increases the number of policy alternatives that the collective actor prefers to the status quo.

This assumption does not have significant consequences for our results. We also tested models where all member states were assigned equal power weights; the results for the inter-institutional distribution of power were very similar.

Where amendments were adopted by a show of hands, we classified the EP as not divided.

An alternative measure of Council division was tested, which simply took the standard deviation of member states’ positions, which gave substantively the same results.

Note that this is the effect compared to the reference category – i.e., issues on which there is no status quo position.

This is calculated by running a model which includes only the four key variables from model 1 (i.e., ‘EP closer to status quo’, ‘Council closer to status quo’, ‘EP committee division’ and ‘Council division’), and estimating the value of the dependent variable when the values of these independent variables are fixed at the specified levels.