How much do agents in trilogues deviate from their principals’ instructions? Introducing a deviation index

ABSTRACT

This article presents a newly developed ‘deviation index’ to measure, in a quantitative and standardized way, the extent to which the negotiators in trilogues (the rapporteur and the rotating Presidency) deviate from the instructions of the institutions they represent (respectively, the EP and the Council). Based on text-mining techniques, the index is applied to the entire set of trilogue decision-making processes in the 2012–2016 period (N = 111). The article then presents three examples of how the index can generate new insights about legislative policy-making in the EU. These empirical applications show that agents deviate more than minimally required to reach an inter-institutional compromise; that rapporteurs deviate in general more than Presidencies do; and that deviation is not affected by the support for the mandate or by the size of the agent.

KEYWORDS:

• Deviation index
• early agreements
• ordinary legislative procedure
• principal–agent
• text-mining
• trilogues

Acknowledgements

We would like to thank the participants of the panel, ‘The Politics of Trilogue Negotiations’, at the 2017 EUSA Conference, the editor and the two anonymous referees for their helpful comments and encouragements.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Thomas Laloux is a PhD candidate at the Institut de sciences politiques Louvain-Europe (ISPOLE) of the University of Louvain (Louvain-la-Neuve), Belgium.

Tom Delreux is professor of EU politics at the Institut de sciences politiques Louvain-Europe (ISPOLE) of the University of Louvain (Louvain-la-Neuve), Belgium.

Notes

1 If the legislative institutions wish to amend the trilogue deal, the OLP continues to second reading and it no longer results in an early agreement.

2 We divide by the total number of words in the mandate because the entire mandate has to be considered as the principal’s preference and not only the amendments to the Commission proposal. If a principal does not amend a sentence in the proposal, it means that this sentence corresponds to its preference. Consequently, if the agent accepts a modification on this sentence in the trilogue negotiations, the agent deviates from the principal’s preferences.

3 In two specific scenarios, we use other kinds of documents for the EP mandate. First, when a new report is adopted following European elections, we use that new report only when no trilogue agreement was reached before. This new report more adequately measures the instructions of the principals to whom the agent has to defend the final trilogue deal. Second, when the EP plenary is asked to adopt a position before the first trilogue, we use the report adopted by the plenary instead of the committee report.

4 To identify the COREPER position adopted before the first trilogue meeting, we consult various Council documents, such as the COREPER analysis of the final compromise, where the reference or date of adoption of the mandate for the first trilogue is mentioned. In the few cases where there is no reference to a COREPER position, we use the general approach adopted by the Council before the first trilogue.

5 The mandates of the EP and the Council either contain only a list of amendments to the Commission proposal, or they present the amendments and the non-modified part of the Commission proposal. When only the amendments are presented in the mandate, we manually adjust the Commission proposal by adding the proposed amendments in order to construct texts with the same format and structure. This is a necessary operation to adequately compare the mandates mutually, as well as the mandates with the trilogue deal.

6 Comparing deviation scores between different cases is not possible with this method, as the assessments of deviation that we can deduce from qualitative case studies are too dissimilar. As mentioned before, such unsystematic – and thus incomparable – assessment of deviation is the main weakness of studying post-delegation principal–agent relationships by means of qualitative case studies. It is precisely this issue that the deviation index aims to address.

7 This is confirmed by a non-parametric Wilcoxon test.

8 In the limited number of cases (N = 9) where the EP plenary also voted on the mandate, we used the percentage of votes in plenary.