Introduction. Use and Limitations of the Principal–Agent Model in Studying the European Union

  • Tom Delreux
  • Johan Adriaensen
Open Access
Part of the Palgrave Studies in European Union Politics book series (PSEUP)


Given the omnipresence of delegation and control in the EU, the principal–agent model has become a popular analytical framework to study the design and effects of delegation and control. Yet, with the ascendance of governance as a mode of decision-making, the contemporary relevance of the principal–agent model became contested. We argue that the model still retains its relevance to study contemporary EU politics, but it requires researchers to follow a two-step approach. First, the hierarchical, dyadic relationship under study has to be clearly defined amidst a complex web of relations. Second, the conditions that have led to the observed pattern of delegation and/or the consequences of this pattern on the distribution of power between the principals and the agent can be inquired.

1 Introduction 1

Delegation is at the very heart of the European integration process. Since the 1950s, national governments in Europe have delegated ever more rule-making powers to the supranational level. Doing so, they have created and strengthened the European Union (EU) in an effort to solve collective action problems and because they expect delegation to be beneficial. Yet simultaneously, member states have also opted to keep control on European politics and European policies. They remain the masters of the EU budget; hold decisive power in the appointment of Commissioners, judges in the Court of Justice and other influential positions; retain sovereignty over the implementation and transposition of much EU legislation; or put the intergovernmentally organized European Council at the centre of the most salient EU decision-making processes. Today, with growing Euroscepticism and open contestation of European authority by national politicians, the mere issue of delegation to the EU has become more controversial than ever. In contrast, more European integration—and thus delegation of more and deeper powers to the EU—is often advocated as the solution to many of the crises the EU is currently facing, be it in the domains of macro-economic governance, migration or foreign policies. In other words, the division of power between member states and the EU is a matter of contentious debate and is still in flux. This shows that delegation and control are highly political and that their study touches upon the essence of European politics.

Once powers have been delegated to the EU, its institutions receive specific authority, but they are also kept under control. Indeed, when diverting our focus from the EU integration process (and thus delegation to the EU) towards the functioning of the EU political system (and thus delegation to EU institutions and actors), we see similar patterns of delegation and control. The European Commission’s powers in the execution of European legislation, the vexed autonomy of the European Central Bank, the influence of the High Representative and the European External Action Service in the foreign policy realm, and the autonomy of rapporteurs in the European Parliament vis-à-vis their committees, political groups or national constituents are all the result of a decision to delegate and the ensuing quest to maintain control.

Given the omnipresence of delegation and control in the being and the functioning of the EU, it does not come as a surprise that scholars are interested in questions on the design and effects of delegation and control. The principal–agent model has become a popular analytical framework to study such political processes. It offers theoretical insights into the reasons, modalities and consequences of “principals” (e.g. member states) delegating powers to “agents” (e.g. the EU or its institutions). Moreover, the principal–agent model enables us to better understand the divisions of power within the EU. As such its empirical findings may also inform normative debates about the evaluation of the democratic nature of the EU’s institutional structure (Brandsma and Adriaensen this volume).

However, despite the model’s attractiveness in the study of the EU, more and more questions are being asked on the contemporary relevance of principal–agent analyses. With its exclusive focus on hierarchical, dyadic relations, the principal–agent model seems at first sight ill-equipped to study an empirical reality where decision-making is increasingly characterized by large, horizontal networks among a plethora of public and private actors. This book assesses the use and limitations of the principal–agent model in a context of an increasingly complex European political system. It claims that the principal–agent model is still extremely insightful for a better understanding of well-chosen political phenomena in the EU, but that the model has to address a number of conceptual, theoretical and methodological challenges.

This introductory chapter first briefly introduces the principal–agent model. Section 2 traces its origins and evolutions in the study of the EU and its status as a heuristic tool. Section 3 then presents the main puzzle that lies at the source of the book. We portray the major challenges faced by the principal–agent model to grasp contemporary European politics and we suggest a way to deal with it, namely using the model for three well-defined purposes: (1) mapping institutional structures and political interactions; (2) studying the reasons and modalities of delegation; and (3) investigating the consequences of delegation in terms of the resulting balances of power. In Sect. 4, we present the main objectives and the overall approach of the book and we introduce how the remainder of the book is structured.

2 The Principal–Agent Model in the EU Studies

The principal–agent model is a heuristic tool that helps to identify the key factors for understanding and explaining the politics of delegation and discretion. It allows scholars to understand why, how and with which consequences certain actors delegate the authority to execute a particular task to other actors (Tallberg 2002). The former are called principals, and the latter agents. The social interactions between principals and agents point to fundamental political processes: Why is authority given up and put in the hands of others? How is the contract between a principal and an agent designed? What does this mean for the new division of power between them once delegation has taken place?

2.1 A Short Introduction to the Principal–Agent Model

A principal–agent relationship is established through an act of delegation by the principal, which can be considered as a contractual relation between a principal and an agent. Such a contract can be more or less formal. The most formalized acts of delegation in the EU are included in the European Treaties where the member states delegate competences to the EU and different types of powers to the EU institutions. Such macro-delegation is complemented with a variety of acts of micro-delegation that take place in everyday decision-making (Dür and Elsig 2011). This occurs in a formalized manner, for instance when the Commission is granted implementation powers by the European legislators or when it receives the authority to conduct international negotiations from the Council. The act of delegation is then respectively a legislative act or a negotiation mandate. Yet, it also happens in a non-formalized way, for instance when the European Council cuts the final Gordian knots in legislative policy-making when the Council could not (Kroll this volume) or when the Council secretariat drafts proposals in intergovernmental negotiations (Reykers and Beach this volume). Hence, the act of delegation can also be informal or implicit (Hawkins et al. 2006a). As Niemann and Huigens (2011: 421) argue: “[a]lthough a [principal–agent] relationship presupposes the presence of a contract between the actors, this contract need not necessarily be explicit or legalized”. However, irrespective of whether the act of delegation is formal or informal, its mere existence is essential to conceive a social relationship as a principal–agent relationship. To put it clearly, there is no principal–agent relationship without an act of delegation. Consequently, applying the principal–agent model for studying a political or social phenomenon where no act of delegation can be identified is pointless.

But why and when does delegation occur? A principal delegates powers to an agent because the former considers it functional and beneficial to do so (Dijkstra this volume). The main motivations for delegation are the reduction of transaction costs in policy-making or the signalling of a credible commitment to the relevant stakeholders (Epstein and O’Halloran 1999; Majone 2001; Pollack 2003). Delegation occurs because principals find it beneficial that agents execute functions on their behalf. These functions are extremely diverse and range from monitoring compliance and thereby addressing collective action problems; solving problems of incomplete contracting; providing technical and independent expertise; being blamed for unpopular decisions; and reducing instability by locking in political agreements and setting the agenda (Garret 1992; Pollack 1997; Kassim and Menon 2003).

A key aspect of the principal–agent model is that delegation not only implies benefits for the principal, but also agency costs. Yet the benefits initially always outweigh the costs since otherwise delegation would not occur. The costs of delegation relate to the fact that there is a real possibility that the agent will act opportunistically and behave contrary to what the principal wants. If that occurs, it is called “agency slack” or “agency loss”. The costs of delegation are caused by preference divergences and information asymmetry between principal and agent. On the one hand, an agent may have—and try to realize—his proper preferences, which can differ from those of the principal. 2 On the other hand, the agent can be better informed than the principal and use that private information at the principal’s expense. Briefly, the decision to delegate immediately implies a risk of heterogeneous preferences and a risk of asymmetrical information, which can be counterproductive for the principal. The fact that functional delegation simultaneously generates costs for the principal confronts the latter with a dilemma: “the need to delegate authority may give powers to the agents that can be used against the principals” (Nielson and Tierney 2003: 246).

Because the principal is well aware of these heterogeneous preferences and the asymmetrical information risks, he creates mechanisms to control the agent. The aim of control is to mitigate the costs of delegation. Hence, delegation and control are inextricably bound to each other. Principal–agent models emphasize the joint occurrence of both phenomena (Christensen and Lægreid 2007), and they distinguish between different types of control mechanisms. Ex ante control mechanisms are established before the agent starts executing his delegated tasks and determine the scope (what?) and the procedure (how?) of that delegated task. Ex post control mechanisms can be employed during or after the agent acts. It comprises the monitoring of the agent (sometimes referred to as ad locum control, Delreux and Kerremans 2010) and the possibility to sanction him. If the principal directly monitors the agent himself, it is called “police patrol” control. By contrast, if the principal relies on third parties in the hope that they will signal unwanted agent behaviour, it is referred to as “fire alarm” control (McCubbins and Schwartz 1984). Importantly, setting up and activating control mechanisms can be costly as well for the principal. This implies that too much control can undo the initial benefits of delegation. Hence, the principal is constantly facing a delicate balancing exercise between, on the one hand, granting sufficient leeway to his agent to ensure that the functional benefits are realized and, on the other hand, sufficiently controlling his agent to ensure that the potential preference and information costs do not rise too much. One peculiar case in which this balancing exercise is further complicated occurs when delegation is motivated by an effort to signal a credible commitment (Majone 2001; Miller 2005). In such situations, the benefits from delegation are tightly linked with the agent executing his delegated task with considerable leeway.

The result of delegating and controlling is an agent who enjoys a certain degree of discretion. “Discretion” should be understood as the autonomy or the room for manoeuvre the agent has in carrying out the delegated task. There is currently no consensual definition in the literature of discretion or autonomy (for an overview of definitions, see Conceicão-Heldt 2013: 24–25). In more conservative definitions, discretion is limited to the room of manoeuvre as determined in the act of delegation, whereas autonomy corresponds to the discretionary space within the context of post-delegation politics (Hawkins et al. 2006a). However, empirical research continuously shows “how subtle the boundary between the degree of discretion and autonomy enjoyed by bureaucracies can be” (Conceicão-Heldt 2013: 25), making us prefer a broader definition less prone to conceptual confusion. In this book, discretion is considered as the leeway enjoyed by the agent in the execution of the delegated task.

This broad conceptualization allows us to take into account the principal’s actions—i.e. what he “gives” to the agent by delegating and what he “takes back” by controlling, be it ex ante, ad locum or ex post—as well as the agent’s actions—i.e. the additional room for manoeuvre taken or acquired by the agent. Likewise, this broad conceptualization allows for taking into account discretion caused by shirking, where autonomous agent behaviour is caused by the agent’s wish to maximize his own preferences, and discretion caused by slippage, where excessive autonomous agent behaviour can be an unintended consequence of the design of the principal–agent relationship (Kiewiet and McCubbins 1991). Through the concept of discretion, principal–agent models allow to make claims on which actor in the dyadic relationship—the principal or the agent—dominates in a particular policy-making process in the EU. The more discretion the agent enjoys, the more influence he yields over the outcome of the delegated task and the more likely he becomes the prevailing actor over the principal.

2.2 Origins and Evolution of the Model in EU Studies

The principal–agent model as it is currently applied by political scientists was borrowed from the field of economics, where it was initially used to study insurance contracts (Spence and Zeckhauser 1971). Subsequently, economists started to use the model in the 1970s to study the organization of firms and to analyse the relationship between the shareholders, as owners of the company, and the managers, to whom they delegate the responsibility to run the company (Jensen and Meckling 1976). In the early 1980s, political scientists borrowed these insights to explore questions on the creation, design and effects of institutions in the realm of politics. The first applications were made in American politics where the principal–agent model was used to study the relation between the plenary and Congressional committees, as well as legislative delegation from Congress to regulatory agencies and courts (Weingast and Moran 1983; Epstein and O’Halloran 1999). This Congressional politics literature inspired international relations scholars to use principal–agent insights in the study of international organizations (Hawkins et al. 2006b). The principal–agent model provided a powerful tool in order to grasp delegation by states to intergovernmental organizations such as the World Bank (e.g. Nielson and Tierney 2003) or the International Monetary Fund (e.g. Martin 2006).

Simultaneously, the potential for the study of EU integration was quickly acknowledged. It is Mark Pollack’s seminal article (Pollack 1997) and his subsequent book (Pollack 2003) that introduced the model into the study of the EU. Being part of the new-institutionalist—and particularly rational choice institutionalist—turn in the EU studies, Pollack’s work demonstrated the usefulness of the principal–agent model for understanding why and with which consequences national governments delegate powers to supranational institutions. That delegation to an agent takes place in an attempt by the principal to reduce the transaction costs of policy-making is the main argument that travelled across the Atlantic and that spilled over from US Congressional politics to EU studies. Common themes in these first studies were the focus on the member states as principals and the effects of delegation on the empowerment of actors, institutions and agencies at the EU level. The traditional supranational institutions such as the Commission and the Court were obvious agents, but also the member states’ relations with the other “non-majoritarian institutions” such as the Council Presidency or the European Central Bank were rapidly analysed from a principal–agent perspective (Elgie 2002; Tallberg 2002; Thatcher and Stone Sweet 2002; Franchino 2006).

This first wave of studies aimed to understand the autonomy of EU actors vis-à-vis the member states as well as the former’s ability to have an impact on the European integration process and European policies. The focus on EU actors and the question of how important they are in driving European integration and European policy-making was not new, but approaching these issues through the framework of the principal–agent model allowed EU scholars to made progress in our understanding of the EU. The principal–agent model’s entry onto the scene of European integration theory at the end of the 1990s helped to overcome the somewhat artificial and increasingly redundant debate between intergovernmentalists and neofunctionalists on whether respectively the member states or the European institutions were driving European integration. The principal–agent model does not assume that the principals or the agents are dominating the game (for an overview of different answers to this question in the field of EU trade politics, see Dür and Zimmermann 2007: 779–780), but offers the conceptual and theoretical tools to analyse under which conditions agents enjoy more or less autonomy. This entailed an important step in the theorization of the EU, which contributed to the consequent success and attractiveness of the model, leading to a swift growth of studies applying the model to various instances of EU politics (Billiet 2009), yet also to a debate on the way the model was applied and its analytical usefulness in the study of the EU (Kassim and Menon 2003; Pollack 2007; Maher et al. 2009).

Since the pioneering principal–agent literature that examined the general questions of delegation to European institutions and their autonomy vis-à-vis member state governments, the literature has developed in two directions. First, the focus has narrowed down from macro-delegation (i.e. delegation of authority to the EU and its institutions in general, for instance in the European Treaties) to instances of micro-delegation (i.e. delegation of authority to an agent in a particular policy area, in a particular time period or even in a particular policy-making process). EU trade policy-making is a prime example of this development because of its relatively straightforward representational set-up in principal–agent terms. As trade is an exclusive competency of the Union, member states in the Council are obliged to delegate negotiating authority to the Commission as agent (Meunier and Nicolaïdis 1999; Kerremans 2006; Elsig 2007; Conceicão 2010). But also in domains of the EU’s Common Security and Defence Policy (Dijkstra 2013), competition policy (Damro 2006), external environmental policy (Delreux 2011) or in the study of EU agencies (Groenleer 2009; Buess 2015) have we seen principal–agent approaches being used as a theoretical framework.

Second, an evolution has also taken place from an exclusive focus on internal explanatory variables for delegation and discretion in the research design towards a greater appreciation of the context in which the agent operates (i.e. external variables) as an explanation for the core features of the principal–agent relationship. Initially, principal–agent scholars directed their attention to the distribution of preferences within the principal, the heterogeneity of the preferences of principal(s) and agent or the salience of the policy-making process. Studies have shown that preference heterogeneity between the principals affects the agent’s discretion (Elsig 2010; Conceição-Heldt 2011, this volume). Other studies demonstrated that when the preferences of the principal converge with those of the agent, the former will consider the latter as an ally and will feel less inclined to deploy strict control mechanisms, as a result of which the agent’s discretion will increase (Kerremans 2006).

More recently, principal–agent studies have shown that the agent’s discretion also depends upon external variables, which relate to the context in which the agent acts and which are not directly affected or experienced by the principal (Delreux et al. 2012). In other words, the characteristics of the institutional setting in which the agent executes the delegated task—such as the degree to which binding decisions can be taken in this setting, its level of informality, the degree of judicialization or the political pressure to reach an agreement—must be taken into account to fully explain the agent’s discretion (Billiet 2009; Delreux 2011; Niemann and Huigens 2011; Poletti 2011). Agents can also exploit these characteristics to their advantage and use them as leverage to increase their discretion vis-à-vis the principal. This has brought principal–agent scholars to examine strategic behaviour by the agents, showing that discretion is not completely determined by the principal but is also affected by the broader context and agent behaviour (Delreux and Kerremans 2010). As a result, recent principal–agent analyses are taking the agent side of the dyad more and more seriously and are even investigating the relationship from the agent’s point of view (Delreux 2009; Dür and Elsig 2011; Gastinger 2016; Plank and Niemann this volume; Coremans and Kerremans this volume). This evolution meets the criticism by Hawkins and Jacoby of a decade ago that the imbalanced attention for the principals has left the principal–agent literature with a “remarkably thin view of agent behaviour” (Hawkins and Jacoby 2006: 199). It has led to a rising importance of what one might call an “agent-principal” analysis.

2.3 The Principal–Agent Model as Heuristic Tool and the Question of Assumptions

This book considers the principal–agent model as a heuristic tool that helps us to make sense of certain aspects of EU policy-making. The principal–agent model should thus not be seen as a fully fledged theory with predetermined outcomes (Thatcher and Stone Sweet 2002). 3 As a heuristic tool, the principal–agent model reduces complexity in real-life political processes and it allows us to reveal the key factors, which are essential to grasp and to explain delegation and discretion. By taking this point of departure, the book contends that the principal–agent model allows EU scholars to discover political dynamics and phenomena in their empirical data that would probably not have been identified without employing the model. Heuristic tools give meaning to the data, but also frame them in a particular way so that certain dynamics receive particular attention while others are pushed into the background. In that sense, the principal–agent model “cuts through the inherent complexity of organizational relationships by identifying distinct aspects of individuals and their environments that are most worthy of investigation, and it integrates these elements in a logically coherent whole” (Moe 1984: 757).

This brings us to the assumptions of the principal–agent model. We consider the model as a rather flexible heuristic tool, thereby echoing the recommendation by Maher et al. (2009: 409) on its use in the EU studies: “apply liberally but handle with care”. “Apply liberally” implies that we recommend a flexible use of the model in which most features relating to policy-makers (principals and agents) or to a policy-making process are treated as variables rather than constants or assumptions. A liberal application of the principal–agent model is opposed by scholars who advocate a more rigid use of the model as it was initially developed in economics (e.g. Karagiannis 2016). The approach of this book stands out against this “canonical” use of the model, which argues that a political or social situation can only be defined in terms of a principal–agent model if a number of core assumptions are fulfilled, including an information benefit for the agent, heterogeneous preferences between the principal and the agent, and a unified principal (Bendor 1988; Miller 2005). According to this canonical approach, the relationship and the subsequent game between principals and agents unfold through sequential moves, which ultimately leads to rational predictions on actor behaviour. However, although the assumptions advocated by the canonical approach may potentially work in economic contexts, they are very seldom applicable to real-life (EU) politics. Therefore, instead of maintaining that political processes in the EU should not be modelled in (canonical) principal–agent terms, we favour to lessen the assumptions and to examine the effects of variation on these conditions upon the politics of delegation and discretion. This is the choice that the lion’s share of political science applications of the principal–agent model has made in the last decades, albeit often implicitly. Although it implies that the principal–agent model has been “modified in ways that are inconsistent with the original formulations”, it has the major advantage that it is done “in ways that are distinctly advantageous for progress in political science” and thus for a better understanding of EU politics (Miller 2005: 206). Whereas some may regret the conceptual stretching of the model a liberal use implies, this book takes a more pragmatic approach and accepts the flexible manner in which the principal–agent model has been increasingly used in the study of the EU.

For instance, the distribution of preferences or of information is not necessarily given but should rather be treated as a variable. This allows us to examine the effect of preference heterogeneity or homogeneity between principals mutually or between the principals and the agent on delegation and discretion. Likewise, it is equally interesting to analyse the effect of a particular level of information asymmetry, in favour of the agent or of the principals, on the principal–agent relationship. Consequently, we do not consider a priori that diverging preferences between the principals and the agent are an assumption of the model nor a necessary condition to apply the model to a particular situation. It is more interesting and it is likely to provide more explanatory power when it is treated as a variable (Waterman and Meier 1998; Rasmussen 2005; Elsig 2010; Conceição-Heldt 2011; Delreux 2011; Adriaensen 2016).

Next to “apply liberally”, we also advocate “handle with care”. Here we argue that the principal–agent model can only be adequately applied when two necessary conditions are fulfilled: (1) the rationality of actors is accepted and (2) an act of delegation can be identified. First, as the principal–agent model has entered the political science discipline as a branch of rational choice institutionalism, the latter’s assumptions should be accepted (Hall and Taylor 1996; Pollack 2009). This means that the principal–agent model assumes principals and agents to be goal-seekers and utility-maximizers applying an instrumental rationality and basing their behaviour upon a cost-benefit analysis. Likewise, individual actions, choices and behaviour of principals and agents are assumed to be affected by institutional constraints of the context in which they act. Assuming rationality does not necessarily imply that principals and agents only have narrowly defined self-interested preferences. They might also seek broader or collective goals. What matters is that principals and agents seek goal achievement, irrespective of what their goal exactly is.

Second, as mentioned above, the principal–agent model is only useful to better understand a particular political situation when an act of delegation—be it formal or informal—can be identified. That act of delegation implies that we are dealing with hierarchy in a dyadic relation. Hierarchy refers to the ability of principal–agent models to understand and to explain the relative balance of power: Which actor is more powerful vis-à-vis another actor? Is it the actor who controls (the principal) or the actor who enjoys delegated authority (the agent)? The dyadic relation refers to the fact that the principal–agent model captures a relation between two (sets of) actors: a principal and an agent. It is exactly these final two constitutive elements of a principal–agent relationship—hierarchy and dyad—that are under pressure through developments in contemporary politics.

3 The Contemporary Relevance of the Principal–Agent Model

3.1 Identifying the Challenge: EU Politics as a Web of Non-exclusively Hierarchical Relations with Multiple Actors

Studying instances of EU politics through the lens of the principal–agent model is currently challenged in two ways: the main mode of governance is no longer exclusively hierarchical and an exclusive focus on dyadic relationships seems at first sight artificial. A hierarchical approach of social interactions has been a core feature of the principal–agent model. In his seminal article “The New Economics of Organization”, Moe (1984) laid the theoretical foundations for the principal–agent model within political science. He placed hierarchical relations into the limelight by contrasting them to market-based relations between political actors. This hierarchical understanding of political relations constitutes the common core of the principal–agent analyses from the early Congressional politics literature, over Epstein and O’Halloran’s (1999) study on delegation to, and discretion of, American executive agencies to the initial applications in EU politics by Pollack (1997, 2003). As most principal–agent studies of EU politics in the last two decades are based on Pollack’s ground-breaking work, hierarchy is still at the core of today’s principal–agent analyses of the EU.

However, throughout the 1990s, it became ever so clear that the dichotomy between hierarchies and markets, as suggested by Moe, neglected many cooperative, egalitarian and reciprocal relations that exist between political actors (Powell 1990; Thompson et al. 1991). As Börzel (1998: 260) noted: “Governments have become increasingly dependent upon the co-operation and joint resource mobilization of policy actors outside their hierarchical control. These changes have favoured the emergence of policy networks as a new form of governance”. Governance approaches became increasingly popular in international relations and comparative politics alike and the EU was rapidly seen as a fertile ground to apply insights from the governance literature. The greater awareness of horizontal, non-hierarchical, relations led to an increasing focus on (policy) networks as a mode of governance (Peterson 2009). Yet, networks do not completely replace hierarchical relationships, they rather complement them (Héritier and Lehmkuhl 2011). Nonetheless, governance approaches pose a first challenge for the principal–agent model, which has an explicit focus on hierarchies instead of on networks.

A second challenge relates to the fact that the governance literature points to the growing involvement of a multitude of public and private actors in decision-making processes. Consequently, adequately grasping political processes by using a dyadic perspective on political relations becomes more and more difficult. When studying EU politics, these challenges are exacerbated by the EU’s multi-level polity and the bureaucratic apparatus that supports it. Public and private actors are involved not only at the subnational, national and European level but also in various bodies that are inherently transversal. Cooperation rather than conflict is often the norm in these relations, leading to deliberation and consensus-seeking.

These observations give rise to the main challenge we discern for scholars that wish to apply the principal–agent model: How can the principal–agent model contribute to the study of political processes in the EU, which are increasingly taking place in a web of non-exclusively hierarchical relations and which involve a multitude of actors who influence each other?

3.2 Coping with the Challenge: Three Uses for the Principal–Agent Model

Our main answer to the observed challenge is that this does not necessarily mean that the principal–agent model has lost its use. The model may even be more valuable than ever for understanding certain aspects of EU politics as it allows for identifying and explaining key features of political interactions, which risk fading away in a complex networked and multi-actor reality. This book provides a strategy to cope with this challenge and to maximize the utility of the principal–agent model by means of a two-stage inquiry. The first stage consists of distinguishing the relations that can be studied through the lens of the principal–agent model. This means that we need to identify the “ principalagent proof ” hierarchical relations amidst a plethora of other types of relations. Many problems in current applications of the principal–agent model originate from an unclear identification of the principal and the agent, as well as the act of delegation that links them. As a second stage, and only when the principal–agent relationship is clearly determined, we can inquire into the conditions that have led to the observed pattern of delegation (i.e. the politics of delegation) and/or assess the consequences this pattern has on the distribution of power between the principal and the agent (i.e. the politics of discretion).

This implies that the principal–agent model is to be applied for three uses: mapping principal–agent relations; studying the politics of delegation; and studying the politics of discretion. The first use—carefully mapping the principal–agent relationship under study—is a necessary step before the model can be applied for the two other uses. In other words, mapping the various aspects of the “principal–agent proof” relation is a prerequisite for a sound and correct study of the politics of delegation and the politics of discretion. The specific challenges related to each of the three uses of the principal–agent model are elaborated below.

3.2.1 Mapping Principal–Agent Relations

Mapping a decision-making process is often complex in its own right. Differentiating between the roles and relations of the relevant actors adds an additional layer of complexity. We will focus our attention on two challenges: the identification of the principal and the agent (who delegates to whom?) and the specification of the object of delegation (what is delegated?).

Each principal–agent analysis requires the identification of a (set of) principal(s) and a (set of) agent(s). But when can an actor really be conceived of as a principal? The governance turn and the network approaches in the study of EU politics have highlighted the proliferation of public and private actors that seek to influence policy-makers. But the ability or the desire of an actor to influence (or control) another actor is not sufficient to classify the former as a principal. While many actors seek influence, few have actually delegated the authority to the agent through an act of delegation. Only the latter are principals. In other words, the act of delegation rather than the exertion of control is the necessary condition upon which the principal and the agent can be identified. This implies that there is a key difference between an endorser and a principal. An endorser attempts to influence an actor by granting legitimacy to a policy as an approximation of the principal’s interest (Milner 1997: 60), whereas a principal actually delegates authority to another actor. However, differentiating principals from endorsers is not always straightforward in the EU when agents are required to consult with public actors (e.g. the European Parliament under the consultation procedure or the European Economic and Social Committee and the Committee of the Regions under the ordinary legislative procedure) or in the event that a principal delegates part of his monitoring prerogatives to a separate shadow bureaucracy (e.g. the Court of Auditors which audits the accounts of the EU institutions). In a similar sense, an actor’s mere ability to veto a decision is not a sufficient guarantee to make it a principal. The distinction between a veto-player and a principal thus lies in the former’s absence from the original decision to delegate some form of authority.

The dyadic nature of the principal–agent relation gives rise to additional questions when identifying the principal. The frequent occurrence of more than one principal delegating authority either jointly or separately to a common agent poses a first challenge. Joint delegation from a “collective principal” implies a single contract between the group of principals and the agent, whereas separate delegations are characterized by a multiplicity of contracts through which each of the “multiple principals” decides to delegate to the agent (Nielson and Tierney 2003). For instance, when EU secondary legislation delegates executive powers to the Commission, the Council and the European Parliament are rather acting collectively (through the same piece of legislation) than separately (as multiple legislative chambers). But in many of the EU institutions, identifying cases of multiple or collective principals is even more complicated because of the voting rules in force (simple majority in the EP; simple majority, qualified majority, reverse qualified majority or unanimity in the Council). These rules, together with power asymmetries within the collective principal, give (individual) actors credible instruments of (individual) control even if they need to act as a collective (Adriaensen 2016; Dijkstra this volume). This leads to what Sobol (2016) calls “pathological delegation”: individual actions of control by separate actors within a collective principal undermine the collective delegation effort as well as the agent’s work.

Second, also identifying the agent can give rise to challenges. The era in which the traditional supranational institutions (the Commission and the Court) are the only agents in the political system of the EU is clearly over, as we observe a proliferation of actors in the EU who execute tasks on behalf of others. In the last decades, a number of “de novo bodies” have been created to which specific tasks have been assigned, including the European External Action Service (EEAS), the European Stability Mechanism (ESM) or the more than 30 specialized agencies (Brickerton et al. 2015). It also frequently occurs that principals delegate (sub)tasks to multiple agents, which also have a particular relation with each other. For instance, association agreements with third countries are often negotiated by the Commission (for the trade part) and by the EEAS (for the CFSP part). The almost unlimited possibilities in the EU to create new agents has resulted in the existence of multiple agents performing (components of) a single task. Moreover, it has also generated a series of potential agents to execute a task. This leads to questions of why and when certain agents are (not) selected and what the consequences are of having multiple agents (Helwig this volume). Next to the phenomenon of “multiple agents”, an additional complexity emerges in the sense that some of these agents can also be “collective agents”, consisting of various subunits and leading to fragmentation inside the agent (Graham 2013). Opening the black box of intra-agent politics allows for examining how conflict or division of labour within a collective agent affects the principal–agent relationship, yet it also raises questions on the level of analysis where agency should be identified (Laloux this volume; Gastinger this volume).

It is often quite challenging in the EU context to decide whether an actor should be treated as a principal or an agent. In some cases, the “chain of delegation” concept offers a solution, as it allows for simultaneously considering multiple dyadic principal–agent relations that are linked to each other in the sense that one actor is at the same time an agent and a principal (of another agent) (Bergman et al. 2000; Nielson and Tierney 2003). This implies that, empirically, an actor can be an agent while at the same time also being a principal. Analytically, however, whether the actor is treated as a principal or an agent depends on the particular relation a researcher is interested in.

But in other cases, the “chain of delegation” concept is less useful because it cannot grasp the hybrid nature of an actor who has both principal and agent characteristics. Two complexities are worth mentioning in this regard. On the one hand, it often occurs in the EU that an agent is selected among the principals and that, consequently, the agent is a subset of the principals. If member states in the Council delegate a particular task to one or a few member states, these member states become agents but they are simultaneously (part of) their own principal. Such a situation in which the agent is a subset of the principals happens for instance when the rotating Presidency acts on behalf of the Council (e.g. representing the Council in trilogue negotiations) or when a group of member states represents the EU in negotiations with third countries (e.g. the EU-3 in the negotiations with Iran). Second, some principals can be closer involved to the work of the agent than others, blurring the distinction between a principal and an agent as a result. Such an actor is positioned somewhere in between the principals and the agent. For instance, when the European Parliament conducts trilogue negotiations with the Council (or better, with the representative of the Council), the Parliament is represented by the rapporteur (acting as agent of the EP committee) but also the shadow rapporteurs are member of the EP team in trilogues, giving the latter a different status than the “normal” principals in the committee (Laloux this volume).

In addition to the identification of the principal(s) and the agent(s), it is important to specify the object of delegation: What is being delegated by the principal to the agent? The decomposition of the public administration into single-purpose agencies, institutions and coordinative bodies has resulted in a patchwork of principal–agent relations where the actors and their roles can vary based on the specific task being studied. In the extreme case, the principal–agent relation can be reversed across the policy cycle (Brandsma and Adriaensen this volume). If this happens, it not only leads to challenges in the identification of the principal and the agent (as discussed earlier), but it also doubles the object of delegation within a single principal–agent dyad. For instance, in the context of the EU’s development cooperation programmes, the implementation of these programmes can be delegated back to (a consortium of) national development agencies.

3.2.2 Studying the Politics of Delegation

The complex institutional patterns of delegation described in the previous section are the outcome of the (joint) decisions of a principal to delegate (part of) his authority to a specific agent. This section focuses on the why question and the challenge of providing an explanation of the complex patterns of delegation described above. We point here to three patterns in the politics of delegation which have not been given due attention in principal–agent analyses of contemporary EU politics: the occurrence of joint delegation, questions of agent selection versus agent creation, and the observation of partial delegation.

Joint delegation is a first complex pattern of delegation. It implies that multiple actors within a collective principal delegate authority to an agent by means of a single act of delegation (Dijkstra this volume; Conceição-Heldt this volume). But how do they jointly determine whether the benefits of delegation outweigh its costs? The outcome of any cost-benefit analysis is likely to diverge for the various actors within the collective principal. Questions of European integration by their very nature focus on the joint decision by the member states to delegate authority. It becomes even more complicated when authority is delegated by a set of principals that are not (completely) collective. One example is the choice between the applicable regimes for the implementation of secondary legislation: Will it occur through implementing or through delegated acts? As co-legislators, the Council and the European Parliament can jointly delegate executive authority to the Commission, but they dispose of different control mechanisms depending on the applicable implementation regime (the EP has stronger possibilities to control the Commission under the delegated acts regime than under the implementing acts regime). As a result, the preferences of both principals, who are obliged to delegate jointly, with respect to the design of the act of delegation are likely to differ.

Second, as mentioned before, questions of agent selection and agent creation have become increasingly relevant in the EU context. The population of possible agents in the EU has grown considerably with the creation of new institutional entities and actors. This makes agent selection an increasingly important feature of the principal–agent relationship, and it generates questions about why principals select certain agents and why they do not select others (Mansbridge 2009). For instance, when creating the Banking Union, why was it not the European Commission that was given more powers by the member states, but rather the European Central Bank (agent selection) and the newly established European Banking Authority (agent creation)? Another illustration pertains to the EU’s external representation in international negotiations in areas of mixed or member states’ competences (Delreux 2008, 2011). Here, the EEAS, the European Commission, the rotating Presidency as well as individual member states can assume a representation role. Why and under which conditions are some actors preferred as agents to others?

A third phenomenon that warrants our attention is the already mentioned occurrence of partial—or non-exclusive—delegation of authority (Dijkstra this volume). Why has authority been fragmented prior to delegation rather than maintained whole? Which types of authority are commonly separated and what logic underlies such separation? Picking up the example of the creation of a European Banking Union, one can ask why the task of creating (or realizing) the Single Rulebook was separated from the implementation of the Single Supervisory Mechanism.

3.2.3 Studying the Politics of Discretion

Once the principal has determined the amount of authority to be delegated and established the appropriate mechanisms of control, the principal–agent model turns its focus towards the ensuing tug of war between the agent and the principal. In other words, whereas the politics of delegation lay down the rules of the game to be played between principal and agent, the eventual unfolding of the game is subject of the politics of discretion. Just like the outcome of a game depends on the rules by which one plays, the politics of discretion will be deeply affected by the politics of delegation. By consequence, the challenges identified in the previous sections continue to reverberate in this section.

As argued above, the agent’s discretion depends partly on the actions of the principal and partly on the actions of the agent. The two primary ways in which the principal can affect the agent’s discretion are through the amount of authority that is initially delegated and through the creation and activation of control mechanisms. The study of control complicates when we accommodate real-life complexity in a model conceived for simple dyadic hierarchy. A first complexity pertains to the two dimensions of control that can be studied: the range of available control mechanisms and the actual activation of these mechanisms. The available control mechanisms are identified by looking at the institutional set-up determined through the politics of delegation. Yet, when focusing on the politics of discretion, the actual use and activation of the available control mechanisms come into the story. Indeed, while the act of delegation may provide for a plethora of tools to avoid agency slack, this does not necessarily imply that these tools are actually employed (all the time). However, it is clear that the availability and the activation of control mechanisms are related. The likelihood of activation can be understood as a function of the availability of control and the political cost resulting from activation for the principal. A simple threat to activate control by the principal (by sending a political signal to the agent, not by actually activating it) is more influential when its effective use would have dire consequences for the agent.

A classic methodological problem in principal–agent analyses with the measurement (or even the empirical observation) of control is related to observational equivalence (Weingast and Moran 1983; Damro 2007; Kroll this volume; Reykers and Beach this volume). It means that the non-activation of control mechanisms might simultaneously be interpreted as agency slack, but also as effective control. Indeed, the mere availability of a control mechanism and the anticipatory behaviour of the agent can be sufficient to foster compliance when the agent wants to avoid being sanctioned by the principal. In other words, it might be the case that the agent’s anticipation of the available control mechanisms results in the non-activation of the control mechanisms by the principal because the agent wants to avoid being sanctioned. That is why we might classify a specific case as “no control” even though the agent’s behaviour was restrained by the availability (and the risk of activation) of control mechanisms. Moreover, to avoid involuntary defection, such anticipatory behaviour may even push the agent to report beyond the principal’s demands (Coremans and Kerremans this volume).

Control that is executed by a collective principal requires a particularly qualified understanding. The assessment of what is deemed to be a case of control (or agency slack) can vary between individual actors within that collective principal. What might be a gross overstepping of the mandate by the Commission in trade negotiations (and thus a good case for control activation) for one member state might be perfectly acceptable agent behaviour for another (and thus not requiring the activation of control). This different assessment is related to the preference heterogeneity within the principal (or rather the varying preference correspondence with the agent).

From the above, it is clear that control by the principal primarily aims to reduce the unintended consequences of the agent’s discretion. However, discretion is not merely the inverse of control. It is also the result of the agent’s actions. The agent can acquire additional discretion independently of the actions of the principal. Discretion-enlarging behaviour by the agent can be either interest-induced or structure-induced. The first concept, interest-induced discretion, is closely related to “shirking”, which is the result of opportunistic behaviour of the agent who aims to maximize his own preferences even if they are diverging from the principal’s preferences. This relates to the classic agency problem where an agent shows undesired behaviour in the eyes of the principal because the agent is self-interested. The concept of interest-induced discretion is broader than mere shirking as it also incorporates actions by the agent to alter the preferences of the principal, thereby aiming to loosen the principal’s propensity to activate control mechanisms.

The second concept, structure-induced acquired discretion, emphasizes that the agent’s discretion does not necessarily result from the agent’s own interests nor from the principal’s control. It can also be the result of the (institutional) structure in which the agent performs the delegated task (Plank and Niemann this volume). Agents are often exposed to various pressures in the context where they act on behalf of the principals. What is important here is that the principal is excluded from these pressures and that they are only felt at the agent side of the dyad. These pressures can affect an agent’s behaviour and ultimately also his relations with the principal (and thus his discretion). For instance, when the Commission negotiates in international organizations on behalf of the member states, it might be pressured by the external negotiation partners “to take its responsibility” by deviating from its mandate, as a result of which the agent will have to return to the principal defending a deal that was the only feasible one. Likewise, national representatives in the Council can be socialized into Community norms (Lewis 2000). These pressures or norms are not felt by the principals but they can affect an agent’s behaviour and ultimately also his relations with the principals (and thus his discretion). This can lead to situations in which they negotiate beyond their mandate—for instance to facilitate the compromise or as a result of diffuse reciprocity—leading to more discretion vis-à-vis their domestic principals. Importantly, this kind of additional, structure-induced, discretion is not the result of opportunistic behaviour by agents but of compliance to what the external context expects from them.

4 Aim, Approach and Structure of the Book

Despite the challenges it faces, the principal–agent model remains an attractive heuristic tool to study EU politics. It is useful and relevant for understanding and explaining key political dynamics related to delegation (and control) as well as discretion (and distributions of power). But to make full use of the principal–agent model’s potential, it should be applied cautiously and the challenges related to the institutional mapping as well as the study of delegation and discretion should be tackled adequately. This book finds its origins in our assessment that the state of the art of principal–agent analyses in the EU studies has insufficiently addressed the above-mentioned challenges. Under the guise of parsimony, scholars using the principal–agent model seem to have evaded too many of these questions. Others noted the complexities in identifying principals, agents, objects of delegation, complex delegation patterns or sources of discretion, but every so often refrained from evaluating or hypothesizing its implications for the core features of the principal–agent model.

Therefore, the aim of this book is to assess the uses and limitations of the principal–agent model in the context of an increasingly complex EU political system. It intends to advance our knowledge on delegation and discretion in the EU by incorporating the complexities of contemporary European politics into the model. By tackling these challenges, we hope that the book will contribute to a more qualified—yet also improved—use of the principal–agent model in general and in the study of the EU more specifically. While the highly institutionalized nature of the EU proved fertile ground for the use of the model, the theoretical contribution of most studies was rather limited to the application of existing insights to a new context. Instead of borrowing from the literature on US politics or international organizations, we expect that—through this book—scholarship on the EU will become a breeding ground for innovative research on using principal–agent model for the wider discipline.

Each of the following chapters presents a theoretical argument on a particular aspect of the principal–agent model, which contributes to addressing the challenges and questions outlined in the previous section. These theoretical arguments allow for a better understanding of delegation and discretion patterns in the EU, but are intended to be relevant for principal–agent analyses beyond the EU too. The main focus of the contributions is thus theoretical, yet each chapter is based on empirical research that inspires and informs the theoretical argument. In this way, this book presents theoretically inspired empirical research leading to necessary sophistications of the principal–agent model.

As Table 1 shows, the book consists of 12 chapters, which are structured in three parts. Part 1 aims to introduce the reader to the use and limitations of the principal–agent model in empirical research and its subsequent normative interpretation. Whereas this introductory chapter has outlined the main challenges to the state of the art of principal–agent analysis of EU politics, Chap.  2 by Brandsma and Adriaensen clarifies the normative value of the principal–agent model. Starting from the observation that many principal–agent analyses—eventually—serve an implicit normative objective, the chapter essentially argues that the principal–agent model has most value as an empirical model. The reductionist framework of the principal–agent model warrants caution when using it as a normative benchmark for assessing the democratic legitimacy of the EU.
Table 1

Structure of the book



Theoretical focus

Politics of delegation

Politics of discretion

Empirical application

i. Setting the scene


Delreux and Adriaensen

Challenges and uses of the principal–agent model



Brandsma and Adriaensen

Relevance of principal–agent for accountability and legitimacy


ii. Refining theoretical aspects of the principal–agent model



Reasons and consequences of non-exclusive delegation



Foreign policy; EEAS



Principals’ motivations for reforming a single agent into a collective agent



Trilogues; European Parliament



Effect of agent interactions on discretion



Foreign policy; EEAS; European Commission


Plank and Niemann

Effect of institutional characteristics of context in which agent operates on discretion



Conflict resolution



Manifest and latent control mechanisms



Relationship between European Council and Council of the EU



Effect of contestation within a collective agent



Trade policy; European Commission



Effect of preference distribution between collective and multiple principals and of types of control mechanisms on discretion



Trade policy


Coremans and Kerremans

Effect of information asymmetry management by agent on politics of discretion



Trade policy

iii. Conclusions


Reykers and Beach

Studying discretion: how principal–agent can benefit from process-tracing


Council Secretariat


Adriaensen and Delreux



The second—and major—part of the book encompasses eight chapters with a specific theoretical focus regarding both the politics of delegation and the politics of discretion. Table 1 presents the theoretical foci of all chapters, as well as the empirical case that will be used to illustrate the principal–agent argument. This part is structured along the uses of the principal–agent model: Chaps.  3 and  4 only deal with the politics of delegation, Chap.  5 analyses both the politics of delegation and the politics of discretion, and finally Chaps.  6 10 focus on the politics of discretion only.

Examining the non-exclusive character of acts of delegation in the context of the European External Action Service, Dijkstra (Chap.  3) discusses what the implications of non-exclusive delegation are for our understanding of the principal–agent model. Non-exclusive delegation occurs when principals delegate a function to an agent but when some of them also carry out that task themselves. This challenges our understanding of the functional logic driving the politics of delegation and calls for taking the role of individual members of a collective principal seriously.

Laloux (Chap.  4) also focuses on the politics of delegation, and more particular on the question why principals redesign the pattern of delegation by reforming a single agent into a collective one. He starts from the observation that, through multiple reforms of its rules of procedure, the European Parliament has opted to be represented in trilogue negotiations by a collective agent (a “EP negotiation team”) and no longer by a single agent (the rapporteur). By analysing the motivations of the principals to reform their agent, he finds that the design of the agent progressively reflects the desire of the collective principal to mirror its proper dividing lines.

Shifting attention to the politics of discretion, Helwig (Chap.  5) starts his analysis from the perspective of the agent as he examines how the interactions between multiple—and often rivalling—agents affect their discretion. Studying the competition between the High Representative and the European Commission in the area of EU foreign policy, he argues that competition between multiple agents is likely to reduce the agents’ discretion. That may lead to a level of discretion that is lower than originally granted by the principals.

The sources of the agent’s discretion are then further scrutinized by Plank and Niemann (Chap.  6). They look at how agents can acquire more discretion independently of what the principals want. Examining the discretion of the EEAS in the area of conflict resolution, they claim that the external setting in which the agent acts, allows the latter to increase his discretion.

Equally focusing on the politics of discretion, Kroll (Chap.  7) then looks deeper into how latent and manifest controls work. The former implies that a principal controls the agent by revising its decisions, whereas the latter refers to control through anticipation by the agent that makes manifest control unnecessary. Kroll analyses the relationship between the European Council and the Council of the EU through two case studies, one in the context of the patent package and the other on the emergency relocation system of refugees. She finds that the type of control exerted by the principal depends on the level of conflict within the agent and the importance attached to the issue by the principal.

In Chap.  8, Gastinger opens the black box of a collective agent and looks at internal conflict within the agent. He investigates the principal–agent relationship between the Council and the Commission in the context of five bilateral trade negotiations from 1970 to 2007, where he finds different degrees of intra-Commission conflict. His case studies reveal that intra-agent conflict might be cultivated by the agent to exercise discretion and shape outcomes.

Chapter  9 then re-introduces the principals into the study of discretion, as Conceição-Heldt analyses the effect of multiple principals with heterogeneous preferences on the discretion enjoyed by the agent. When the European Commission negotiates free trade agreements for the EU, it acts on behalf of multiple principals: the Council and the EP. Conceição-Heldt examines how the multitude of principals and the types of oversight mechanisms available to these principals affect the agent’s discretion. She argues that preference heterogeneity between multiple principals increases discretion and that agents are able to circumvent centralized control mechanisms when there is disunity within a collective principal.

Finally, in Chap.  10, Coremans and Kerremans analyse how agent discretion also depends on the way agents manage information asymmetry with their principals. Their empirical analysis is also situated in the area of trade policy-making, where informal technical meetings between the Commission and the member states allow for confidential information exchange. Coremans and Kerremans show that these interactions between principals and agent offers an ideal institutional environment for the agent to manage the degree and quality of information asymmetry in such a way that it can prevent unintended shirking but still maintains enough autonomy to bring the negotiations to a good end.

Each of those eight chapters is structured in the same way. They first present the theoretical puzzle by delineating the particular aspect of the principal–agent model that needs to be refined for a meaningful analysis. All chapters include a figure that visualizes the identification of the principals, the agents, the object of delegation and the particular aspect of the principal–agent relationship that it studies (see Box 1). Doing so, each chapter starts by our first use of the principal–agent model, namely mapping the hierarchical dyad in the institutional context under scrutiny. The theoretical puzzle is then addressed through principal–agent analysis of a particular instance of EU politics. As the identified challenges cut across various policy domains and various kinds of decision-making processes, the subject of the empirical application is sourced widely to encompass decisions on, for instance, legislative policy-making, supranational agencies or EU’s external negotiations. Finally, each chapter in this part concludes by presenting its theoretical argument. It discusses what the principal–agent model can tell us about the theoretical puzzle on the basis of its empirical analysis, and it reflects upon the conditions under which this theoretical argument applies beyond that empirical case study.

Box 1.

Visualization of principals, agents and the object of delegation

Since we argued that the mapping of the hierarchical dyad that the principal–agent model will examine is a crucial and necessary first step in any principal–agent analysis, all chapters making a theoretical argument start with such a mapping exercise. The result is presented in a figure in the introduction of each of these chapters. These figures are designed in a systematic way and look like the one presented in this box. First, the principal(s) (visualized in a circle) and the agent(s) (in a square box) are portrayed. Second, the act of delegation is presented as an arrow and the object of delegation is added to that arrow in italics. Finally, the main theoretical puzzle of the chapter is depicted in a black box, possibly complemented by a thick arrow that points to the precise feature in the dyadic relationship that is examined by that chapter.

The third part concludes the book with two more general contributions. Chapter  11 by Reykers and Beach provides some methodological guidance on how discretion can be analysed in a systematic manner. As a reliable measurement of discretion is essential to study this phenomenon, its explanatory factors and the causal mechanisms linking both, particular attention needs to be paid to the method. The authors advocate the use of process-tracing as a methodological tool to open the black box of causality in principal–agent research and to unravel the causal mechanisms that lead to agent discretion. They demonstrate the added value of the method with a case study on the discretion enjoyed by the Council Secretariat when drafting proposals in intergovernmental negotiations.

Finally, in Chap.  12, we present our conclusions by linking insights from the various chapters to the overall challenges identified in this introduction. By applying the principal–agent model to less hospitable environments, a series of opportunities and challenges can be derived which may guide the future use of principal–agent analysis in EU studies.


  1. 1.

    We thank Eugénia da Conceição-Heldt, Dirk De Bièvre, Samuel Defacqz, Hylke Dijkstra, Antoine Feron, Markus Gastinger, Daniela Kroll, Thomas Laloux, Pauline Pirlot, François Randour, Bart Van Ballaert and Yannis Karagiannis for their constructive comments on earlier drafts of this introduction.

  2. 2.

    Throughout the volume, gender denominations (he/his and she/her) should be interpreted in a gender-neutral manner.

  3. 3.

    Clearly, there are various definitions of what a theory exactly is in the social sciences. Our interpretation of the principal–agent model does not correspond to a theory as a “causal argument of universal, trans-historical validity and nomothetic quality, which can be tested through the falsification of a series of hypotheses” (Diez and Wiener 2009: 3). Yet if one sticks to a broad definition of a theory as an “abstraction from a complex reality [that attempts] to provide generalizations about the phenomena under study” (Dunne et al. 2013: 407), our interpretation of the principal–agent model as a heuristic tool comes close to a theory. However, in order to avoid the misunderstanding that we consider the principal–agent model as a theory in the Diez and Wiener sense of the word, this book does not refer to “principal–agent theory” but rather uses “principal–agent model”.



  1. Adriaensen, J. (2016). National administrations in EU trade policy: Maintaining the capacity to control. Houndmills: Palgrave MacMillan.CrossRefGoogle Scholar
  2. Bendor, J. (1988). Review article: Formal models of bureaucracy. British Journal of Political Science, 18(3), 353–395.Google Scholar
  3. Bergman, T., Müller, W., & Strøm K. (2000). Introduction: Parliamentary democracy and the chain of delegation. European Journal of Political Research, 37(3), 255–260.Google Scholar
  4. Börzel, T. A. (1998). Organizing Babylon—On the different conceptions of policy networks. Public Administration, 76(2), 253–273.Google Scholar
  5. Brandsma, G., & Adriaensen, J. (2017). The principal–agent model, accountability and democratic legitimacy. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 35–54). London: Palgrave MacMillan.Google Scholar
  6. Brickerton, C., Hodson, D., & Puetter, U. (2015). The new intergovernmentalism: European integration in the Post-Maastricht Era. Journal of Common Market Studies, 53(4), 703–722.Google Scholar
  7. Billiet, S. (2009). Principal–agent analysis and the study of the EU: What about the EC’s external relations? Comparative European Politics, 7(4), 435–454.CrossRefGoogle Scholar
  8. Buess, M. (2015). European Union agencies and their management boards: An assessment of accountability and demoi-cratic legitimacy. Journal of European Public Policy, 22(1), 94–111.Google Scholar
  9. Christensen, T., & Lægreid, P. (2007). Regulatory agencies. The challenges of balancing agency autonomy and political control. Governance, 20(3), 499–520.CrossRefGoogle Scholar
  10. Conceição, E. (2010). Who controls whom? Dynamics of power delegation and agency losses in EU trade politics. Journal of Common Market Studies, 48(4), 1107–1126.CrossRefGoogle Scholar
  11. Conceição-Heldt, E. (2011). Variation in EU member states’ preferences and the commission’s discretion in the Doha Round. Journal of European Public Policy, 18(3), 403–419.Google Scholar
  12. Conceicão-Heldt, E. (2013). Do agents “run amok”? A comparison of agency slack in the EU and US trade policy in the Doha Round. Journal of Comparative Policy Analysis, 15(1), 21–36.CrossRefGoogle Scholar
  13. Conceição-Heldt, E. (2017). Multiple principals’ preferences, types of control mechanisms, and agent’s discretion in trade negotiations. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 203–226). London: Palgrave MacMillan.Google Scholar
  14. Coremans, E., & Kerremans, B. (2017). Agents as information asymmetry managers in EU trade policy-making. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 227–253). London: Palgrave MacMillan.Google Scholar
  15. Damro, C. (2006). Transatlantic Competition Policy: Domestic and International Sources of EU-US Cooperation. European Journal of International Relations, 12(2), 171–196.Google Scholar
  16. Damro, C. (2007). EU delegation and agency in international trade negotiations: A cautionary comparison. Journal of Common Market Studies, 45(4), 883–903.Google Scholar
  17. Delreux, T. (2008). The EU as a negotiator in multilateral chemicals negotiations: Multiple principals, different agents. Journal of European Public Policy, 15(7), 1069–1086.Google Scholar
  18. Delreux, T. (2009). The EU negotiates multilateral environmental agreements: Explaining the agent’s discretion. Journal of European Public Policy, 16(5), 719–737.Google Scholar
  19. Delreux, T. (2011). The EU as international environmental negotiator. Surrey: Ashgate.Google Scholar
  20. Delreux, T., Drieskens, E., Kerremans, B., & Damro, C. (2012). The external institutional context matters: The EU in international negotiations. In O. Costa & K. Jørgensen (Eds.), The influence of international institutions on the EU (pp. 58–75). Basingstoke: Palgrave Macmillan (When Multilateralism hits Brussels).CrossRefGoogle Scholar
  21. Delreux, T., & Kerremans, B. (2010). ‘How agents weaken their principals’ incentives to control: The case of EU negotiators and EU member states in multilateral negotiations. Journal of European Integration, 32(4), 357–374.CrossRefGoogle Scholar
  22. Diez, T., & Wiener, A. (2009). Introducing the mosaic of integration theory. In A. Wiener & T. Diez (Eds.), European integration theory (2nd ed., pp. 1–22). New York: Oxford University Press.Google Scholar
  23. Dijkstra, H. (2013). Policy-making in EU security and defense: An institutional perspective. Basingstoke: Palgrave Macmillan.CrossRefGoogle Scholar
  24. Dijkstra, H. (2017). Non-exclusive delegation to the European external action service. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 55–81). London: Palgrave MacMillan.Google Scholar
  25. Dunne, T., Hansen, L., & Wight, C. (2013). The end of international relations theory? European Journal of International Relations, 19(3), 405–425.CrossRefGoogle Scholar
  26. Dür, A., & Elsig, M. (2011). Principals, agents, and the European Union’s foreign economic policies. Journal of European Public Policy, 18(3), 323–338.Google Scholar
  27. Dür, A., & Zimmermann, H. (2007). Introduction: The EU in international trade negotiations. Journal of Common Market Studies, 45(4), 771–787.Google Scholar
  28. Elgie, R. (2002). The politics of the European Central Bank: principal–agent theory and the democratic deficit. Journal of European Public Policy, 9(2), 186–200.Google Scholar
  29. Elsig, M. (2007). The EU’s choice of regulatory venues for trade negotiations: A tale of agency power? Journal of Common Market Studies, 45(4), 927–948.CrossRefGoogle Scholar
  30. Elsig, M. (2010). European Union trade policy after enlargement: Larger crowds, shifting priorities and informal decision-making. Journal of European Public Policy, 17(6), 781–798.Google Scholar
  31. Epstein, D., & O’Halloran, S. (1999). Delegating powers. A transaction cost politics approach to policy making under separate powers. New York: Cambridge University Press.CrossRefGoogle Scholar
  32. Franchino, F. (2006). The powers of the union. Delegation in the EU. Cambridge: Cambridge University Press.Google Scholar
  33. Garrett, G. (1992). International cooperation and institutional choice: The European Community’s internal market. International Organization, 46(2), 533–560.CrossRefGoogle Scholar
  34. Gastinger, M. (2016). The tables have turned on the European Commission: The changing nature of the pre-negotiation phase in EU bilateral trade agreements. Journal of European Public Policy, 23(9), 1367–1385.Google Scholar
  35. 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
  36. Graham, E. (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
  37. Groenleer, M. (2009). The autonomy of European Union agencies. A comparative study of institutional development. Delft: Eburon.Google Scholar
  38. Hall, P., & Taylor R. (1996). Political science and the three institutionalisms. Political Studies, 44(4), 936–957.Google Scholar
  39. Hawkins, D., & Jacoby, W. (2006). How agents matter. In D. Hawkins, D. Lake, D. Nielson, & M. Tierney (Eds.), Delegation and agency in international organizations (pp. 199–228). New York: Cambridge University Press.CrossRefGoogle Scholar
  40. Hawkins, D., Lake, D., Nielson, D., & Tierney, M. (2006a). 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. 3–38). New York: Cambridge University Press.CrossRefGoogle Scholar
  41. Hawkins, D., Lake, D., Nielson, D., & Tierney, M. (Eds.). (2006b). Delegation and agency in international organizations. New York: Cambridge University Press.Google Scholar
  42. Helwig, N. (2017). Agent interaction as a source of discretion for the EU high representative. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 105–129). London: Palgrave MacMillan.Google Scholar
  43. Héritier, A., & Lehmkuhl, D. (2011). Governing in the shadow of hierarchy. New modes of governance in regulation. In A. Héritier & M. Rhodes (Eds.), New modes of governance in Europe. Governing in the shadow of hierarchy (pp. 48–74). New York: Palgrave Macmillan.Google Scholar
  44. Jensen, M., & Meckling, W. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305–360.Google Scholar
  45. Karagiannis, Y. (2016). Decision-making in the European Union—Or, the meticulous design of power dispersion. Cuadernos Europeos Deusto, 55, 119–143. Google Scholar
  46. Kassim, H., & Menon, A. (2003). The principal–agent approach and the study of the European Union: Promise unfulfilled? Journal of European Public Policy, 10(1), 121–139.CrossRefGoogle Scholar
  47. Kerremans, B. (2006). Pro-active policy entrepreneur or risk minimizer? A principal–agent interpretation of the EU’s role in the WTO. In O. Elgström & M. Smith (Eds.), The European Union’s roles in international politics (pp. 172–188). Oxford: Routledge.Google Scholar
  48. Kiewiet, D., & McCubbins, M. (1991). The logic of delegation. Congressional parties and the appropriations process. Chicago: University of Chicago Press.Google Scholar
  49. Kroll, D. A. (2017). Manifest and latent control on the council by the European council. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 157–180). London: Palgrave MacMillan.Google Scholar
  50. Laloux, T. (2017). Designing a collective agent for trilogues in the European parliament. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 83–103). London: Palgrave MacMillan.Google Scholar
  51. Lewis, J. (2000). The methods of community in EU decision-making and administrative rivalry in the council’s infrastructure. Journal of European Public Policy, 7(2), 261–289.Google Scholar
  52. Maher, I., Billiet, S., & Hodson, D. (2009). The principal–agent approach to EU studies: Apply liberally but handle with care. Comparative European Politics, 7(4), 409–413.CrossRefGoogle Scholar
  53. Majone, G. (2001). Two logics of delegation agency and fiduciary relations in EU governance. European Union Politics, 2(1), 103–122.CrossRefGoogle Scholar
  54. Mansbridge, J. (2009). A “selection model” of political representation. The Journal of Political Philosophy, 17(4), 369–398.CrossRefGoogle Scholar
  55. Martin, L. (2006). Distribution, information, and delegation to international organizations: The case of IMF conditionality. In D. Hawkins, D. Lake, D. Nielson, & M. Tierney (Eds.), Delegation and agency in international organizations (pp. 140–164). New York: Cambridge University Press.CrossRefGoogle Scholar
  56. McCubbins, M., & Schwartz, T. (1984). Congressional oversight overlooked: Police patrols versus fire alarms. American Journal of Political Science, 28(1), 165–179.Google Scholar
  57. Meunier, S., & Nicolaïdis, K. (1999). Who speaks for Europe? The delegation of trade authority in the EU. Journal of Common Market Studies, 37(3), 477–501.CrossRefGoogle Scholar
  58. Miller, G. (2005). The political evolution of principal–agent models. Annual Review of Political Science, 8(–), 203–225.Google Scholar
  59. Milner, H. V. (1997). Interests, institutions, and information: Domestic politics and international relations. Princeton: Princeton University Press. Google Scholar
  60. Moe, T. (1984). The New Economics of Organization. American Journal of Political Science, 28(4), 739–777.Google Scholar
  61. Nielson, D., & Tierney, M. (2003). Delegation to international organisations: Agency theory and world bank environmental reform. International Organization, 57(2), 241–276.Google Scholar
  62. Niemann, A., & Huigens, J. (2011). The European Union’s role in the G8: A principal–agent perspective. Journal of European Public Policy, 18(3), 420–442.Google Scholar
  63. Peterson, J. (2009). Policy networks. In A. Wiener & T. Diez (Eds.), European integration theory (2nd ed., pp. 105–124). New York: Oxford University Press.Google Scholar
  64. Plank, F., & Niemann, A. (2017). Impact of the agent’s environment on discretion in the field of EU conflict resolution. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 131–155). London: Palgrave MacMillan.Google Scholar
  65. Poletti, A. (2011). World Trade Organization judicialization and preference convergence in EU trade policy: Making the agent’s life easier. Journal of European Public Policy, 18(3), 361–382.Google Scholar
  66. Pollack, M. (1997). Delegation, agency, and agenda setting in the European Community. International Organization, 51(1), 99–134.Google Scholar
  67. Pollack, M. (2003). The engines of European integration. Delegation, agency, and agenda setting in the EU. New York: Oxford University Press.Google Scholar
  68. Pollack, M. (2007). Principal–agent analysis and international delegation: Red herrings, theoretical classifications, and empirical disputes. Bruges: Bruges Political Research Paper.Google Scholar
  69. Pollack, M. (2009). The new institutionalisms and European integration. In A. Wiener & T. Diez (Eds.), European integration theory (2nd ed., pp. 125–143). New York: Oxford University Press.Google Scholar
  70. Powell, W. (1990). Neither market nor hierarchy: Network forms of organization. In B. Staw & L. L. Cummings (Eds.), Research in organizational behavior (pp. 295–336). Greenwich, CT: JAI Press.Google Scholar
  71. Rasmussen, A. (2005). EU conciliation delegates: Responsible or runaway agents? West European Politics, 28(5), 1015–1034.CrossRefGoogle Scholar
  72. Reykers, Y., & Beach, D. (2017). Process-tracing as a tool to analyse discretion. In T. Delreux & J. Adriaensen (Eds.), The principal–agent model and the European Union (pp. 255–281). London: Palgrave MacMillan.Google Scholar
  73. Sobol, M. (2016). Principal–agent analysis and pathological delegation: The (almost) untold story. Governance, 29(3), 335–350.Google Scholar
  74. Spence, M., & Zeckhauser, R. (1971). Insurance, information, and individual action. The American Economic Review, 61(2), 380–387.Google Scholar
  75. Tallberg, J. (2002). Delegation to supranational institutions: Why, how, and with what consequences? West European Politics, 25(1), 23–46.CrossRefGoogle Scholar
  76. Thatcher, M., & Stone Sweet, A. (2002). Theory and practice of delegation to non-majoritarian institutions. West European Politics, 25(1), 1–22.Google Scholar
  77. Thompson, G., Frances, J., Levačić, R., & Mitchell, J. (Eds.). (1991). Markets, hierarchies, and networks: The coordination of social life. London: Sage.Google Scholar
  78. Waterman, R., & Meier, K. (1998). Principal–agent models: An expansion? Journal of Public Administration Research and Theory, 8(2), 173–202.CrossRefGoogle Scholar
  79. Weingast, B., & Moran, M. (1983). Bureaucratic discretion of congressional control? Regulatory policymaking by the federal trade commission. Journal of Political Economy, 91(5), 765–800.CrossRefGoogle Scholar

Copyright information

© The Author(s) 2017

Authors and Affiliations

  • Tom Delreux
    • 1
  • Johan Adriaensen
    • 2
  1. 1.Université catholique de LouvainLouvain-la-NeuveBelgium
  2. 2.Maastricht UniversityMaastrichtThe Netherlands

Personalised recommendations