Designing a Collective Agent for Trilogues in the European Parliament

  • Thomas Laloux
Open Access
Part of the Palgrave Studies in European Union Politics book series (PSEUP)


Why does the European Parliament delegate the task to negotiate with the Council in trilogues to a collective agent, i.e. a negotiating team, when the literature contests the rationality of such decision? Analysing the reforms establishing the negotiating teams, this chapter suggests that collective agents can be beneficial for principals. When principals are divided into ideologically structured subgroups and when agents are a subset of the principals, delegation to collective agents allows for mirroring the principal’s structure and is therefore likely to increase the agent’s faithfulness toward the entire principal, who can in turn give him more autonomy. These benefits explain why the cost-benefit calculus made by principals and the resulting design of delegation lean toward collective, rather than single, agents.

1 Introduction

Trilogue negotiations have progressively become the major forum for legislative negotiations in the EU. Trilogues are informal tripartite meetings in which representatives of the Council, the European Parliament and the Commission negotiate an informal compromise that can then be rubberstamped formally by the two co-legislators, resulting in so-called early agreements. 1 In trilogues, the interactions between the institutions and their representatives can be modelled as two principal–agent relationships: both the Council and the EP delegate to their negotiators the task to reach an interinstitutional compromise. While there is a new act of delegation for every legislative policy-making process, the rules through which delegation occurs are established independently in each institution and applicable to the negotiation of all legislative files.

In the Council, the delegation pattern is relatively straightforward: COREPER adopts a mandate that designates the rotating Presidency as the negotiator for the Council in trilogues. In the EP, by contrast, these rules have evolved through several reforms modifying the composition of the agent. In the early days of trilogues (1999–2004), the rapporteurs were the negotiator, but today broader negotiating teams are negotiating on behalf of the EP. These EP negotiating teams, here considered as collective agents, are composed of several members of the European Parliament (MEPs) who are collectively delegated the task to negotiate an agreement with their Council counterpart. Fig. 1 visualizes the principal–agent relationship studied in this chapter and the evolution from a single to a collective agent.
Fig. 1

Visualization of the principal–agent relationship and designing a collective agent

From a principal–agent perspective, this evolution is puzzling as scholars suggest rather costly effects of collective agents for the principals (Bendor et al. 2001; Graham 2013; Elsig 2010), raising the question of the principal’s motivation to delegate to a collective agent rather than a single one. It is a particularly problematic observation as the principal–agent model assumes a rational cost-benefit analysis to inform the decision to delegate. This chapter seeks to explain this puzzle. Studying the consecutive reforms that have ultimately resulted in the delegation to a collective agent can indeed grant us a unique insight into the underlying politics of delegation. Hence, this case study seeks to explain why principals may rationally opt to delegate to a collective agent.

The empirical data come from three types of sources. First, official documents from the EP were studied. These documents concern the different rules of procedures and the working documents of the reforms. They allow the mapping of the evolution of the principal–agent relationship as well as the official motivations of the principals. Second, I drew on the academic literature studying the intra-institutional consequences of trilogue negotiations. Third, I conducted several interviews with MEP assistants and EP officials on the current EP organization in trilogue negotiations.

2 Theoretical Puzzle: The Design of Collective Agents

What is a collective agent? Do the EP negotiating teams meet the criteria to qualify as a collective agent and why does the principal–agent literature consider this form of delegation problematic? In addressing each of these questions, this section presents the theoretical puzzle of this chapter.

The occurrence of collective principals and agents has been recognized relatively early by the principal–agent literature, which soon relaxed the “unitarian actor” assumptions both for principals and for agents (Worsham et al. 1997; Miller 2005; see also Gastinger this volume). The notion of collective agents (and collective principals) goes back to the core features of the principal–agent relationship: the nature of the agents (and the principal) and the delegation contract linking principals and agents (Delreux and Adriaensen this volume). Nielson and Tierney (2003: 247) define a collective principal as a principal that has a single contract with an agent and is “composed of more than one actor” who make decisions jointly. Drawing on this definition, Graham (2013: 369) argues that the “critical difference between collective and unitary agents [is] that in the former, authority is dispersed across actors”. Additionally, Elsig (2010) considers an agent as “complex” when it is composed of several kinds of actors linked with a single contract with a principal.

On the basis of these definitions, three features must be fulfilled to define an agent as collective. First, he must be linked by a single contract with the principal. Second, he must be composed of more than one actor with different preferences who are collectively responsible for executing the delegated task. Third, there must be a diffusion of authority between the actors composing the agent. They have to act jointly to execute the delegated task. The exact intra-agent decision-making rules can vary between collective agents. Hence, collective agents are distinguished both from single agents, linked by a single contract but not composed of more than one actors, and from multiple agents, composed of several actors but not linked by a single contract with the principal.

Collective agents are thus what we find when examining the principal–agent relationship in the EP in the context of trilogues. The delegation to negotiating teams is currently regulated by Rule 73 of the EP rules of procedure. Accordingly, the act of delegation for trilogue negotiations is the decision of the responsible EP committee to authorize the opening of interinstitutional negotiations, which explicitly delegates the task to “reach […] an agreement in the course of a legislative procedure” (European Parliament 2016: 45). Hence, the responsible committee is the principal, and the EP negotiating team is the agent (see Fig. 1). According to the three features defining a collective agent, the EP negotiating teams must be modelled as such.

First, negotiation authority is delegated to negotiating teams as a whole. Rule 73 of the rules of procedure state that the “decision [to open negotiations] shall determine the mandate […] of the negotiating team”. They explicitly add that there is no other delegation to negotiate with the Council: “[interinstitutional] negotiation shall not be entered prior to the adoption of the decision [authorizing negotiation]”. Hence, negotiating teams are linked with their principal committee by a single contract.

Second, while the responsible committees formally decide on the composition of the negotiating teams, Rule 73 stipulates that the latter “shall be led by the rapporteur and presided over by the chair of the committee responsible or by a vice-chair designated by the chair [and] comprise at least the shadow rapporteurs from each political group”. As a result, negotiating teams shall comprise at least these actors who are selected by the political groups at the beginning of a legislative process, i.e. before the delegation to negotiate in trilogues is established. As political groups have different preferences and specific organizations (Bressanelli 2012; Yordanova 2013), members of the negotiating teams are supposed to have heterogeneous preferences. Hence, negotiating teams are composed by more than one actor and are fragmented.

Third, during trilogue negotiations, the decisions regarding negotiation strategies and the acceptance of certain compromises are taken collectively by the whole team. 2 None of the actors composing the teams is necessary to perform the delegated task, neither formally nor in practice. Formally, there is no provision in the rules of procedure that gives veto power to a particular member inside the team. This is confirmed by the practice of trilogue negotiations, as Judge and Earnshaw (2011) as well as de Ruiter and Neuhold (2012) give examples of negotiations where the shadow rapporteurs agreed on a deal with the Council Presidency while sidelining the rapporteur. These examples are further supported by the interviews which emphasized the varying balance of power inside negotiating teams between trilogue negotiations (Interview D, E). 3 Hence, there is a diffusion of authority between the actors composing negotiating teams, as they have to execute the delegated task jointly.

In studying collective agents, scholars have almost exclusively focused on the politics of discretion and hardly on the politics of delegation. In this regard, the literature suggests that collective agents have rather negative impacts on the realization of the principal’s preferences (Bendor et al. 2001). 4 Graham (2013) showed that the diffusion of authority inside collective agents inhibits the exercise of the control mechanisms at the disposal of the principal, thus decreasing the agent’s faithfulness. Similarly, Elsig (2010) showed that the complexity of an agent might also be detrimental to his ability to function well and thus to deliver on the delegated task, which decreases the benefit of delegation.

In this context, the following question arises: Why, despite these negative effects, do principals still delegate to collective agents? This question is particularly interesting as the principal’s decision on the design of delegation is based on a cost-benefit analysis which aims to maximize the attainment of their preferences: “principals do not only delegate power, they also attempt to choose the right [agent] to minimize agency cost at an acceptable cost to themselves” (Pollack 2003: 28). Principals conceive the initial delegation contracts and thus decide upon the design of delegation. In doing so, they anticipate potential problems raising the cost of delegation and they try to minimize the risk of opportunistic behaviour from the agent (Doleys 2000; Hawkins et al. 2006). Consequently, principals prefer to delegate to agents who have similar preferences or who proved obedience in previous delegations (Bendor et al. 2001; Nielson and Tierney 2003). In sum, the recurrent observation of delegation to a collective agent implies that it must be more beneficial for principals than delegation to a single agent. By analysing the reasons why the EP made this choice for trilogue negotiations, we may uncover what those benefits are.

3 Empirical Analysis: EP Negotiating Teams in Trilogues

The first trilogue negotiations in the early 2000s were developed outside the treaties and were thus characterized by a high degree of informality. 5 Yet, while the practice of trilogue negotiations has progressively become routine, the degree of formalization has also increased (Roederer-Rynning and Greenwood 2015). Particularly, the EP adopted several internal measures to formalize the delegation for trilogue negotiations. These reforms modified the composition of the EP agents in trilogues: while initially represented by the rapporteur, the responsible committees now delegate to negotiating teams comprising several actors. How did this evolution progressively take place? And what were the EP’s motivations explaining this choice? By consecutively answering both questions, this section addresses the puzzle raised in Sect. 2. The current section is divided into two subsections, one for each of the questions.

3.1 Reforming the Agent

The current delegation design for trilogues in the EP is the result of several consecutive internal reforms. During the first years of trilogue negotiations, there were no EP rules regulating trilogue negotiations. Consequently, the acts of delegation were very informal and their design varied to a large extent. During this period, the rapporteurs were the main actors to conduct the informal negotiations with the Council and were thus the EP agents (Shackleton and Raunio 2003; Farrell and Héritier 2004; Roederer-Rynning and Greenwood 2017). This situation raised several criticisms inside and outside the EP (De Clerck-Sachsse and Kaczyński 2009; Héritier and Reh 2012; Reh 2014; Roederer-Rynning and Greenwood 2015).

Particularly, two main concerns appeared. First, trilogue negotiations were seen as reducing the transparency of the legislative procedure. 6  Second, the restricted participation and the seclusion characterizing trilogue negotiations were seen as empowering the participating actors inside their institutions (Farrell and Heritier 2004). Particularly, rapporteurs were identified as the main winners of informalization (Benedetto 2005; Rasmussen and Shackleton 2005; Costello and Thomson 2010). 7 While this issue of rapporteurs’ empowerment has been challenged in the more recent literature (Rasmussen and Reh 2013), it nevertheless created tensions within the EP (Costa et al. 2011; Héritier and Reh 2012), leading to the successive adoptions of more stringent rules. Three major revisions of the rules of procedure took place.

First, in 2004 the Conference of Presidents of the EP adopted a set of “guidelines for best practice within Parliament in first and second reading” (European Parliament 2004a). Their official purposes were to “establish a uniform way of proceeding within Parliament when seeking an agreement at first or second reading, while at the same time ensuring maximum flexibility, transparency, effectiveness and legitimacy [of trilogue negotiations]”. Although the guidelines stated that the EP participation to trilogues should be decided by the group coordinators of the responsible committees, rapporteurs were considered to be the agents as the guidelines did not mention larger negotiating teams. While the participation of other committee members was not compulsory, information on the negotiations should be provided “either through direct participation of the Committee chair and/or shadow rapporteurs or coordinators, or through prompt and sufficiently detailed information from the rapporteur to the chair and shadow rapporteurs or coordinators”. Vague and without any binding nature, these guidelines had little impact on the conduct of trilogue negotiations (Costa et al. 2011; Héritier and Reh 2012; Obholzer and Reh 2012).

Second, the delegation structure was revised in 2009 in the context of the general reform of the EP rules of procedure. The special working party charged to prepare the reform 8 was “explicitly tasked with looking at the practice of early agreements” (Héritier and Reh 2012: 1147). In the resulting report, the working party recognized the gaps in the implementation of the guidelines and recommended to strengthen the content and to increase the status of the guidelines (European Parliament 2007). As a result, the EP adopted in its rules of procedure a new rule (Rule 70), dedicated to the “Interinstitutional negotiations in legislative procedures” and a “Code of conduct for negotiating in the context of the ordinary legislative procedures” in the annex (Annex XX of the rules of procedure). Regarding the question who should be the EP agent, Rule 70 only referred to the code of conduct. The latter introduced the concept of “EP negotiating team” and specified its composition as follows: “as a general principle, political balance shall be respected and all political groups shall be represented at least at staff level”. However, the code of conduct did not specify the membership any further. In addition, these new rules made the committees, and no longer the group coordinators only, responsible for deciding upon the agents’ composition. Again the impact of this reform on trilogues remained unclear, as its implementation was considered to be dissatisfactory (Héritier and Reh 2012).

Third, the last reform occurred in 2012 and consisted of a modification of the Rule 70 and the addition of a Rule 70bis. This reform was initiated by the Conference of Presidents who invited the AFCO committee to review Rule 70 of the rules of procedure with the aim of “making the procedures more effective, more transparent and more inclusive through the incorporation of some key elements of the Code of conduct for negotiating in the context of the ordinary legislative procedures in the binding part of the Rules” (European Parliament 2012: 9). Especially, the Conference of Presidents asked to focus on the decision for the composition of the negotiating team (European Parliament 2011). As a result, Rule 70 was reformed and it established the current system of negotiating teams, which are now composed of the rapporteur, the committee chair (or vice-chair) and the shadow rapporteurs from each political group. 9 Later, these rules were renumbered Rules 73 and 74.

The evolution of the EP’s rules in the composition of its agent for trilogue negotiations is summarized in Table 1, which presents, for each period, the rules in vigour and the composition of the EP agent in trilogues.
Table 1

Evolution of the EP representation in trilogue negotiations


Main rules in vigour in the EP

EP agent


No fixed rules



Guidelines for best practice within Parliament in first and second reading


• EP participation decided by the coordinators


EP rules of procedure: Rule 70 and Annex XX: Code of conduct for negotiating in the context of the ordinary legislative procedures

EP negotiating team

• As a general principle, political balance shall be respected and all political groups shall be represented at least at staff level in these negotiations


EP rules of procedure: Rule 73 (previously 70): Interinstitutional negotiations in legislative procedures

EP negotiating team

• Shall be led by the rapporteur and presided over by the chair of the committee responsible or by a vice-chair designated by the chair. It shall comprise at least the shadow rapporteurs from each political group

Besides the broadening of the agent, the reforms have also progressively introduced and reinforced several control mechanisms at the disposal of the responsible committee, ex ante as well as ad locum and ex post. Ex ante, the responsible committees must issue a mandate as a basis for the negotiations under the form of committee reports. Ad locum, the agents are progressively compelled to share information with the responsible committee under the form of oral reports and documents sharing. Moreover, committees were given the possibility to update the mandates “in the light of the progress of the negotiations” (European Parliament 2016: 46). Ex post, the responsible committee must now approve the trilogue compromise before tabling it at the plenary, giving also the latter ex post sanctioning power. However, at the same time, the rules have remained explicitly flexible regarding the application of these control mechanisms, as they foresee a potential exception to the application of each control mechanism. As an example, the rules of procedure state that “where it is not feasible to convene a meeting of the committee in a timely manner, the negotiating team shall report back to the chair, the shadow rapporteurs and the coordinators of the committee, as appropriate rather than to the whole committee” (European Parliament 2016: 46).

3.2 Explanation of the Choice of Negotiating Teams

While the established control mechanisms can theoretically be explained from a classic principal–agent point of view (they aimed at decreasing the agent’s discretion which was considered problematic by the principals), the change in the agent structure does not exactly fit this explanation. To understand the establishment of a collective agent, two elements must be taken into account: (1) the fact that the EP committee is composed of different political groups; and (2) the fact that the agent in trilogues is a subset of its principal, i.e. the committee. A single agent, as it was the case when the rapporteurs were agents, would be biased towards one of the political groups, which was likely to affect the balance of power inside the committee. In this perspective, the motivation to introduce negotiating teams was a wish to mitigate the empowerment of one political group in the committee. The solution consisted of allowing all political groups of the EP to be present within the agent negotiating—and not only one of them.

In the single-agent era, rapporteurs could use their discretion to favour their political group by pursuing partisan policies in the execution of their task (Yoshinaka et al. 2010), which affected the balance of power inside the committee. It empowered the largest political groups who not only hold the majority of the rapporteurships but also have the voting power to secure favourable majorities for the informal compromises. This came at the expense of the smaller political groups, which were often excluded from the negotiations as they had fewer rapporteurships (de Ruiter and Neuhold 2012). As witnessed by a green MEP during this period: “we are not part of that informal consensus” (quoted in Farrell and Héritier 2004: 1201).

Hence, to protect the balance of power within the committee and to reduce the empowerment of the political group of the rapporteur, it was deemed necessary to design an agent mirroring the principal’s structure. With all the political groups being present in the collective agent, the information asymmetry between them decreases, and this gives the opportunity to the different political positions to be heard in the negotiation (Reh 2014). This motivation to protect the balance of power within the EP was first officially expressed in 2001 by the Vice-Presidents of the EP who warned against two dangers of trilogue negotiations: the “reduced opportunities for wider societal and political interests to introduce their points of view into the decision making process” and the reduction in the “open and public debate in the plenary with the full participation of all political groups and members” (Imbeni et al. 2001: 2). Consequently, they advised to underline who are the relevant actors at the different phases in the procedure and to ensure a greater role to the committees and the plenary as “it is in these bodies that the political balance of our institution is most fully protected” (Imbeni et al. 2001: 2).

Protecting the balance of power inside the committee was thus the main motivation behind the two reforms that broadened the agent from a single to a collective one. 10 In 2009, the preparatory document highlighted indeed the lack of “political confrontation along clear political lines” due to the small number of representatives of the three institutions participating in trilogues (European Parliament 2011: 3–5). Therefore, the working party on parliamentary reform suggested to “determine the composition of the negotiating team (rapporteur, committee chair, shadow rapporteurs)” and to “clarify the role, rights and duties of shadow rapporteurs and coordinators” (European Parliament 2011: 3–5). As a consequence, the 2009 reform made the mirroring of the principal structure within the collective agent mandatory by imposing the respect of the political balance of the EP committee within the negotiating team.

The objective of the 2012 reform of the agent’s composition was also to provide for a politically balanced composition of the negotiating team and to make trilogue negotiations more “inclusive” (European Parliament 2012). Similarly, the reform strengthened the mirroring of the principal structure by making the presence of all shadow rapporteurs—and thus of all political groups—compulsory. This avoided the possible marginalization of some shadow rapporteurs in the decisive discussions, which was common in the single-agent era (Bunyan 2007). Following the same logic, the reform put the committee chair at the head of the negotiating team. As the chair is supposed to represent the committee as a whole and not his political group, he is considered impartial (European Parliament 2012). The fact that the most neutral committee figure was chosen as the chair of the negotiating team was also prompted by a wish to limit the likelihood that the rapporteur acted in favour of one political group. This role of the chair is further confirmed in the interviews: “the role of the chairs is to make the equilibrium between everyone [as they] represent the whole Parliament and [are] less political in that position” (Interview C). 11

An alternative answer to the concerns of a single agent could have been to increase the control in order to limit the agent’s discretion. Yet, this was not the EP’s choice. On the contrary, each reform of the rules on the conduct of trilogues emphasized the need to guarantee the efficiency of trilogue negotiations. 12 Hence, the EP has defended the necessity of maintaining a space to think in secret, outside full transparency and control (Bunyan 2016). An administrator in the EP phrased it as follows: “It would be unmanageable to take all [these decision] in committee” (Interview E). The desire to guarantee the flexibility of the process is further illustrated by exceptions to the control mechanisms foreseen by the rules (as discussed in Sect. 3.1). It is this space to think (or discretion in a principal–agent vocabulary) that gives the agents the opportunity to shirk by creating information asymmetry and allowing them to emphasize their preferences in the trilogue negotiations (Costello and Thomson 2010; Yordanova 2015).

By changing the composition of the agents, the reforms have allowed the latter to increase their discretion (Reh 2014; Roederer-Rynning and Greenwood 2017). Interviews confirm that decisions on trilogue compromises are often taken “in just the small teams” (Interview E) and thus without intervention of the committee. This can be explained by the structure of the collective agent which mirrors the composition of the principal. As all political groups are represented in the collective agent, the risk of biased outcomes decreases, allowing the committees to reduce the level of control throughout the process: “in general there are no questions from the committee because the shadows are well informed” (Interview F). Yet, the discretion of the negotiating team depends on the capacity of the agent to really mirror the preferences of the committee. For the negotiating teams, this is guaranteed by the procedure of selection of the members of the collective agent, which is done by the political groups (Interviews C, D, E).

Furthermore, the composition of the negotiating teams increases the credibility of the commitment on the trilogue agreement from the rapporteur and the shadow rapporteurs, as they represent their groups and are therefore responsible for them. Thus, participating in trilogues by means of negotiating teams can ensure that a sufficient majority within the EP supports the negotiated compromise without having to consult the committees while negotiating. This way, an involuntary defection can be avoided. This is an important point for the EP as rejecting the informal compromise entails political and transactional costs, in addition to increasing the uncertainty regarding the final output (Costa et al. 2011; Bressanelli et al. 2016). Again, this credible engagement lies in the ability of the members of the negotiating teams to engage their political groups for the vote in plenary (Interviews C, D, E, F).

4 Theoretical Argument

By studying the evolution of the EP agent in trilogue negotiations, this chapter aims at addressing the puzzle of delegation to a collective agent. The previous section showed that, depending on the structure of the principal–agent relationship, such delegation design can actually be beneficial for the principal. Particularly, we are likely to observe delegation to a collective agent when (1) the collective principal is composed of ideologically structured subgroups; and (2) the agent is a subset of the principals.

Collective principals divided into subgroups with heterogeneous preferences often imply the risk of so-called pathological delegation (Sobol 2016). Some subgroups can benefit from agency shirking if the agent’s preferences are more aligned with the subgroup than with other subgroups. This is likely to affect the balance of power within the principal. Particularly, if the agents enjoy discretion, the balance of power may be affected in two ways. First, agents can promote the preferences of a subgroup at the expense of the whole principal. Second, agents can have incentives to give more information to the subgroups to which they are aligned, creating information asymmetry within the principal. This risk is further enhanced when the agent is a subset of the principal because the agent necessarily belongs to one of the subgroups. In such a case of single agents, the latter necessarily has different preferences from the whole principal. Agent shirking then de facto reinforces the subgroup to which the agent belongs, and in such cases delegating to a single agent decreases the benefit of delegation for the whole principal.

Delegating to a collective agent, by contrast, can allow mitigating these negative consequences. The empowerment of some subgroups can be avoided by designing an agent in a way that it mirrors the structure of the principal. But to do so, the agents must be composed of several actors (one for each subgroup) and thus be “collective”. In this case, the two sources of subgroups’ empowerment are challenged. First, as all the subgroups have members in the collective agent, the preferences of the agent and the principal are more closely aligned. This decreases the likelihood of biased outcomes favouring one subgroup. Second, as each member of the collective agent can share information with the subgroup to which he belongs, the information asymmetry between subgroups decreases. This in turn decreases the risk of opportunistic behaviour from the agent and, consequently, the cost of no control for the principals. The latter can therefore give more discretion to the agent and increase the benefits of delegation.

Alternatively, increasing the control mechanisms at the principals’ disposal could decrease both the information asymmetry inside principals and the risk of a biased outcome. Yet, increasing the control on the agent is not always beneficial because the agent might need discretion to execute the delegated task, and therefore limiting discretion might have a cost for the principals too (Thatcher and Sweet 2002). Hence, “the principals have to find a balance between sufficient discretion for the agent on the one hand, and not too much discretion to limit the preference and information advantages of the agent on the other hand” (Delreux 2011: 30). Moreover, even when it would be beneficial for them, the principals are not always capable of exerting tight control, particularly in the case of collective or multiple principals. Because a sufficient majority might be needed to activate the existing control mechanisms, such decisions can be thwarted by members of the principals if they benefit from agency slack (Nielson and Tierney 2003, Sobol 2016). On this basis, if a high discretion is needed or if it is hard to activate control mechanisms, “it is more efficient and productive for the principal to spend time and effort ex ante, choosing an agent with aligned goals, than to expend those resources ex post monitoring and sanctioning an agent who is influenced only by punishment and reward” (Mansbridge 2011: 622).

Furthermore, a collective agent may also decrease the risk of involuntary defection, even without consulting the principal during the execution of the task. Involuntary defections occur when the agent is not able to fulfil the commitments made because they deviated too much from the principal’s preferences (Putnam 1988; Delreux 2011). However, if all the subgroups composing the principals are represented within a collective agent, the latter can ensure more easily that a sufficiently large majority supports his action and will approve the outcome of delegation. Yet, this credible commitment lies in the ability of the members inside the collective agent to properly represent their subgroups.

5 Conclusion

The purpose of this chapter was to examine the motivation of the principals within the EP to design a collective agent for conducting trilogue negotiations. In that sense, it sheds lights on the politics of delegation to collective agents. The analysis of the reforms showed that the evolution of the agent was meant to address the negative effects of delegation to single agents. Particularly, the broadening of the agent aimed at protecting the balance of power between the political groups of the EP, which was threatened by the discretion enjoyed by the rapporteurs as single agents as it allowed them to favour their political groups’ preferences. This issue was addressed by the introduction of the negotiating teams as it allows each group to participate in the execution of the delegated task. This case study enables the analysis of the politics of delegation to collective agents, complementing existing research that has focused rather on the politics of discretion. The findings suggest that, in contrast to the indications of the existing literature, delegation to a collective agent can be beneficial for principals.

This explanation of delegation to collective agents can potentially be applied to other principal–agent relationships within the EU as agents mirroring the principal’s structure are not uncommon. Again in the field of legislative policy-making, this is the case for the delegation to Conciliation Committee, which must correspond to the composition of legislative institutions. Another example occurs in the European Economic and Social Committee (EESC), which is structured in three subgroups (representing, respectively, the employers, the workers and various interests). Similar to the negotiating team, the EESC’s rules of procedure stipulate that its bureau (the organizing and coordinating body) must reflect the principle of balance between these groups (European Economic and Social Committee 2010).

Further, similar motivations might be applied to explain the design of the College of Commissioners, in which each member state has a member. While the Commissioners are officially neutral, scholars have shown that having “their commissioner” allows the member states to protect their interests during the actions of the agent (Wonka 2007; Thomson 2008). Likewise, Kleine (2013) argued that this mirroring nature of the College allows member states to exchange control over policy areas, creating “national fiefdoms” inside the Commission, which increases their grip on the agent. Similar delegation structures, in which all member states have representatives, are also common in the board of the agencies of the EU.

However, the advantages of the collective agents presented above are limited to explain the potential benefit of collective agent in one specific structure of delegation while other forms of collective agents do exist. For instance, Nielson and Tierney (2003) as well as Elsig (2010) studied collective agents who do not mirror their principals’ structure. Likewise, situations occur in which collective principals still prefer to delegate to single agents, such as to the committee chairs in the EP or to the President of the European Council. These observations call for further research regarding the conditions under which delegation to collective agents occurs.

Another line of study regarding collective agents—and negotiating teams in particular—would be to investigate the consequences of the selected design of delegation, linking the intended results of the reforms with the actual results. In that regard, the following questions seem to be promising starting points for further research on collective agents: Was delegation to collective agents actually beneficial for the principals at the end of the day? What was the impact on the relationship between principals and agents? Or how did it affect the balance of power within the principal? Similarly, one could also investigate how this collective agent structure affects the EP bargaining power vis-à-vis the Council during the negotiation, particularly since the interviewees provided a rather contrasted answer on this issue. 13 Finally, there is room to examine how collective agents function internally: How are the relationships between the different actors composing collective agents? And how does intra-agent politics affect the politics of discretion?


  1. 1.

    Such “early agreements” represent now the vast majority of the EU legislation, transforming codecision into “a de facto single-reading legislative procedure” (Roederer-Rynning and Greenwood 2015: 1148).

  2. 2.

    Before and even during breaks of trilogue meetings, the members of the EP negotiating team meet frequently to discuss and adopt such decisions (Interviews A, B, C, D, F, G).

  3. 3.

    Especially, committee chairs increasingly challenge the rapporteur’s leadership in trilogue negotiations.

  4. 4.

    As stated by Bendor et al. (2001: 245): “when authority is delegated to several people, they cannot only shirk but also free-ride on each other’s efforts”.

  5. 5.

    Informality is understood here as defined by Reh et al. (2013: 5): “a restricted, noncodified set of actors operat[ing] in a secluded setting […] structured by informal rather than codified, specific, and enforceable rules”.

  6. 6.

    This reduction in the transparency has been considered detrimental to the EP, as his progressive empowerment aimed, conversely, at increasing the transparency and thereby the democratic legitimacy of the EU legislative process (De Clerck-Sachsse and Kaczyński 2009; Héritier and Reh 2012).

  7. 7.

    The restriction and the seclusion of a trilogue meeting allowed them to “control[ing] the flow of information from their own organization to the other and vice versa” (Farrell and Héritier 2004) and thus acting as gatekeeper (Costello and Thomson 2010).

  8. 8.

    The “special working group on parliamentary reform” was set up in 2007 by the Conference of Presidents to draw up a proposal to reform of the EP rules of procedure and was composed of one member per political group.

  9. 9.

    The reform also added an Rule 70bis regarding the role of the plenary in the decision on the opening of interinstitutional negotiations prior to the adoption of a report in the committee. More precisely, this rule gave to the plenary the possibility to be informed and intervene in such decisions.

  10. 10.

    The 2004 guidelines did not address the issue of agent composition.

  11. 11.

    Yet, in practice this is not always the case. While “it should not be the chair that pushes [for a decision]” (Interview E), “sometimes, the chair really takes over the trilogues from the rapporteur” (Interview D) (see also Roederer-Rynning and Greenwood 2017).

  12. 12.

    To do so, the seclusion and restriction of trilogues were deemed necessary by the EP. Indeed “trilogue negotiators need discretion” and “strict monitoring would decrease negotiators’ flexibility, and the ultimate sanction—voting down the pre-agreed compromise in plenary—would result in sunk costs and reputational loss for a Parliament still struggling to be seen as a responsible and efficient co-legislator” (Reh 2014: 830).

  13. 13.

    On the one hand, some interviewees indicated that negotiating teams can use the preferences of their members against the Council by showing that they are tied. On the other hand, others recognized that disagreements within the negotiating team favour the Council during negotiations.


Cited Interviews


Official from EP Conciliations and Codecision Unit

November 2015



Official from EP Conciliations and Codecision Unit

November 2015



Assistant MEP

March 2016



Assistant MEP

April 2016



Administrator IMCO committee

May 2016



Administrator LIBE committee

May 2016



Administrator LIBE committee

May 2016



  1. Bendor, J., Glazer, A., & Hammond, T. (2001). Theories of delegation. Annual Review of Political Science, 4(1), 235–269.CrossRefGoogle Scholar
  2. Benedetto, G. (2005). Rapporteurs as legislative entrepreneurs: The dynamics of the codecision procedure in Europe’s parliament. Journal of European Public Policy, 12(1), 67–88.CrossRefGoogle Scholar
  3. Bressanelli, E. (2012). National parties and group membership in the European parliament: Ideology or pragmatism? Journal of European Public Policy, 19(5), 737–754.CrossRefGoogle Scholar
  4. Bressanelli, E., Koop, C., & Reh, C. (2016). The impact of informalisation: Early agreements and voting cohesion in the European parliament. European Union Politics, 17(1), 91–113.CrossRefGoogle Scholar
  5. Bunyan, T. (2007). Secret trilogues and the democratic deficit. Retrieved September 2016, from
  6. Bunyan, T. (2016). A missed opportunity to open up secret trilogue decision-making in the EU. Retrieved September 2016, from
  7. Costa, O., Dehousse, R., & Trakalova, A. (2011). La codécision et les “accords précoces”: Progrès ou détournement de la procédure législative? Retrieved September 2016, from
  8. Costello, R., & Thomson, R. (2010). The policy impact of leadership in committees: Rapporteurs’ influence on the European parliament’s opinions. European Union Politics, 11(2), 219–240.CrossRefGoogle Scholar
  9. De Clerck-Sachsse, J., & Kaczyński, P. (2009). The European parliament–More powerful, less legitimate? Retrieved September 2016, from
  10. de Ruiter, R., & Neuhold, C. (2012). Why is fast track the way to go? Justifications for early agreement in the co-decision procedure and their effects. European Law Journal, 18(4), 536–554.CrossRefGoogle Scholar
  11. Delreux, T. (2011). The EU as international environmental negotiator. Aldershot: Ashgate Publishing, Ltd.Google Scholar
  12. Delreux, T., & Adriaensen, J. (2017). Introduction. Use and limitations of the principal–agent model in studying the European Union. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 1–34). London: Palgrave MacMillan.Google Scholar
  13. Doleys, T. (2000). Member states and the European Commission: Theoretical insights from the new economics of organization. Journal of European Public Policy, 7(4), 532–553.CrossRefGoogle Scholar
  14. Elsig, M. (2010). Principal–agent theory and the World Trade Organisation: Complex agency and “missing delegation”. European Journal of International Relations, 17(6), 1–23.Google Scholar
  15. European Parliament. (2004a). Activity report 1 May 1999 to 30 April 2004. PE 287.644.Google Scholar
  16. European Parliament. (2004b). Guidelines for first and second reading agreements under the codecision procedure. Brussels: European Parliament.Google Scholar
  17. European Parliament. (2007). Working party on parliamentary reform: Second interim report on legislatives activities and interinstitutional. PE 406.309 CP.Google Scholar
  18. European Parliament. (2011). Working document on revision of Rule 70 of Parliament’s Rules of Procedure on interinstitutional negotiations in legislative procedures. PE472.201v01.00.Google Scholar
  19. European Parliament. (2012). Report on amendment of rule 70 of parliament’s rules of procedure on interinstitutional negotiations in legislative procedures. PE473.959v04-00.Google Scholar
  20. European Parliament. (2016). Rules of procedure. PE_REGL(2016)07-15.Google Scholar
  21. European Economic and Social Committee. (2010). Rules of procedure. CESE 1083/2010 FR/GW/DS/NT/ms.Google Scholar
  22. Farrell, H., & Héritier, A. (2004). Interorganizational negotiation and intraorganizational power in shared decision making: Early agreements under codecision and their impact on the European parliament and council. Comparative Political Studies, 37(10), 1184–1212.CrossRefGoogle Scholar
  23. Gastinger, M. (2017). Effects of contestation within a collective agent in EU trade policy-making. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 181–202). London: Palgrave MacMillan.Google Scholar
  24. Graham, E. R. (2013). International organizations as collective agents: Fragmentation and the limits of principal control at the World Health Organization. European Journal of International Relations, 20(2), 366–390.Google Scholar
  25. Hawkins, D., Lake, D., Nielson, D., & Tierney, M. (2006). Delegation under anarchy: States, international organizations and principal agent theory. In D. Hawkins, D. Lake, D. Nielson, & M. Tierney (Eds.), Delegation and agency in international organizations (pp. 4–31). New York: Cambridge University Press.CrossRefGoogle Scholar
  26. Héritier, A., & Reh, C. (2012). Codecision and its discontents: Intra-organisational politics and institutional reform in the European parliament. West European Politics, 35(5), 1134–1157.CrossRefGoogle Scholar
  27. Imbeni, R., Proven, J., & Friedrich, I. (2001). Improving the functioning of the co-decision procedure, discussion document presented by the vice-presidents responsible for conciliation. Retrieved September 2016, from
  28. Judge, D., & Earnshaw, D. (2011). “Relais actors” and co-decision first reading agreements in the European parliament: The case of the advanced therapies regulation. Journal of European Public Policy, 18(1), 53–71.CrossRefGoogle Scholar
  29. Kleine, M. (2013). Trading control: National fiefdoms in international organizations. International Theory, 5(3), 321–346.Google Scholar
  30. Mansbridge, J. (2011). Clarifying the concept of representation. American Political Science Review, 105(3), 621–630.CrossRefGoogle Scholar
  31. Miller, G. (2005). The political evolution of principal–agent models. Annual Review of Political Science, 8(1), 203–225.CrossRefGoogle Scholar
  32. Nielson, D., & Tierney, M. (2003). Delegation to international organizations: Agency theory and World Bank environmental reform. International Organization, 57(2), 241–276+i.Google Scholar
  33. Obholzer, L., & Reh, C. (2012). How to negotiate under co-decision in the EU: Reforming trilogues and first-reading agreements. Retrieved September 2016, from
  34. Pollack, M. (2003). The engines of European integration: Delegation, agency, and agenda setting in the EU. New York: Oxford University Press.Google Scholar
  35. Putnam, R. (1988). Diplomacy and domestic politics: The logic of two-level games. International Organization, 42(3), 427–460.CrossRefGoogle Scholar
  36. Rasmussen, A., & Reh, C. (2013). The consequences of concluding codecision early: Trilogues and intra-institutional bargaining success. Journal of European Public Policy, 20(7), 1006–1024.CrossRefGoogle Scholar
  37. Rasmussen, A., & Shackleton, M. (2005). The scope for action of European parliament negotiators in the legislative process: Lessons of the past and for the future. Paper presented at Ninth Biennial International Conference of the European Union Studies Association. Austin, Texas.Google Scholar
  38. Reh, C. (2014). Is informal politics undemocratic? Trilogues, early agreements and the selection model of representation. Journal of European Public Policy, 21(6), 822–841.CrossRefGoogle Scholar
  39. Reh, C., Héritier, A., Bressanelli, E., & Koop, C. (2013). The Informal politics of legislation: Explaining secluded decision making in the European Union. Comparative Political Studies, 46(9), 1112–1142.Google Scholar
  40. Roederer-Rynning, C., & Greenwood, J. (2015). The culture of trilogues. Journal of European Public Policy, 22(8), 1148–1165.CrossRefGoogle Scholar
  41. Roederer-Rynning, C., & Greenwood, J. (2017). The European Parliament as a developing legislature: Coming of age in trilogues? Journal of European Public Policy, 24(5), 735–754. Google Scholar
  42. Shackleton, M., & Raunio, T. (2003). Codecision since Amsterdam: A laboratory for institutional innovation and change. Journal of European Public Policy, 10(2), 171–188.CrossRefGoogle Scholar
  43. Sobol, M. (2016). principal–agent analysis and pathological delegation: The (Almost) untold story. Governance, 29(3), 335–350.Google Scholar
  44. Tallberg, J. (2002). Delegation to supranational institutions: Why, how, and with what consequences? West European Politics, 25(1), 23–46.CrossRefGoogle Scholar
  45. Thatcher, M., & Sweet, A. S. (2002). Theory and practice of delegation to non-majoritarian institutions. West European Politics, 25(1), 1–22.CrossRefGoogle Scholar
  46. Thomson, R. (2008). National actors in international organizations: The case of the European Commission. Comparative Political Studies, 41(2), 169–192.CrossRefGoogle Scholar
  47. Wonka, A. (2007). Technocratic and independent? The appointment of European Commissioners and its policy implications. Journal of European Public Policy, 14(2), 169–189.CrossRefGoogle Scholar
  48. Worsham, J., Eisner, M. A., & Ringquist, E. J. (1997). Assessing the assumptions: A critical analysis of agency theory. Administration & Society, 28(4), 419–440.CrossRefGoogle Scholar
  49. Yordanova, N. (2013). Organising the European parliament the role of committees and their legislative influence. Colchester: ECPR Press.Google Scholar
  50. Yordanova, N. (2015). Collusion in bicameral EU decision-making: The impact of early agreements on representation and committee influence in the European parliament. In F. Müller-Rommel & F. Casal Bértoa (Eds.), Party politics and democracy in Europe: Essays in honour of Peter Mair (pp. 178–193). Abingdon/New York: Routledge Publisher.Google Scholar
  51. Yoshinaka, A., McElroy, G., & Bowler, S. (2010). The appointment of rapporteurs in the European parliament. Legislative Studies Quarterly, 35(4), 457–486.CrossRefGoogle Scholar

Copyright information

© The Author(s) 2017

Authors and Affiliations

  • Thomas Laloux
    • 1
  1. 1.Université catholique de LouvainLouvain-la-NeuveBelgium

Personalised recommendations