International organizations are formal entities with states as members and that possess a permanent secretariat, bureaucracy, or some other form of permanent administration (Pevehouse, Nordstrom, and Warnke Reference Pevehouse, Nordstrom and Warnke2004). Students of international organization pay special attention to the interaction between states and their administration. This interaction is believed to be fraught with delegation problems. Specifically, bureaucracies are suspected of using their informational advantages over the member states to deviate from their mandate and act against the states’ best interests. Scholars consequently ask if and how states can design institutional controls in order to prevent this problem of ‘bureaucratic drift’ (Pollack Reference Pollack1997).
While it is useful, I argue that the scholarly focus on bureaucratic drift may obscure other delegation problems that arise among the states themselves. These problems stem from the fact that bureaucracies are not unitary actors. They are permeable institutions that are susceptible to internal manipulation. Individual states may ignore collective control mechanisms and instead seek to influence the bureaucracy unilaterally from within. This need not be a problem per se, as one state’s influence gain does not necessarily equal another state’s loss. However, unilateral influence creates informational advantages that a state can potentially abuse in order to bias policy implementation to the detriment of its cooperating partners. In that case, other states will react by stepping up their own unilateral influence and an inefficient and mutually costly ‘influence race’ (Urpelainen Reference Urpelainen2012, 705) ensures that slowly erodes cooperation. The prominent delegation problem of ‘bureaucratic drift,’ which is based on informational asymmetries between states on one side, and the bureaucracy on the other side, is consequently overshadowed by what I refer to as the problem of ‘interstate control,’ which is rooted in informational asymmetries among the member states themselves.
The central argument of this article is that the problem of interstate control and its potential institutional remedies play an important role in the operation of international organizations. For this purpose, the article complements popular principal-agent approaches to international organization with theoretical models from the literature on coalitions and coalition government (Laver and Shepsle Reference Laver and Shepsle1990). Delegation problems in coalition governments arise from the fact that parties face trouble controlling the actions of their coalition partners within the government. To render the coalition bargain more credible and durable, coalition partners therefore devise mechanisms that allow them to ‘keep tabs’ on each other (Thies Reference Thies2001). Similarly, I argue that when states suspect that a cooperating partner’s unilateral influence within the bureaucracy might infringe their own interests, they will seek to avoid a subsequent erosion of their cooperation by allowing each other to monitor their actions through what I call ‘interstate control mechanisms.’ Just like coalition partners allow ‘junior ministers’ of one party to keep tabs on senior ministers of the opposite party, senior officials with a suspected bias towards one state are regularly shadowed by deputies with the opposite bias. In addition, just like coalition partners use parliamentary scrutiny mechanisms to monitor the other governing party’s executive decisions, states may coopt oversight mechanisms, which are meant for the principal as a whole or the public to control the bureaucracy, in order to reveal information about their cooperating partners’ unilateral activities within the same bureaucracy.
The argument that international organizations are fraught with the problem of interstate control has two empirical implications. First, it implies that interstate control mechanisms will emerge and be used in response to states’ attempts to abuse their unilateral influence to the detriment of others. Second, and in contrast to collective control mechanisms, interstate control mechanisms exist within rather than outside the bureaucracy, and serve to level informational asymmetries among the member states. Because data on unilateral influence within bureaucracies is rare and typically incomplete (McKeown Reference McKeown2009, 288), the article illustrates these implications using the example of the European Commission, the European Union’s (EU) powerful supranational bureaucracy, for which a comparatively large amount of internal data have been made available.Footnote 1 I recommend that my argument be interpreted as a plausibility probe of a theory that promises a more realistic analysis of the relationship between states and international bureaucracies. It thereby seeks to contribute to at least two bodies of literature.
The first contribution is to the literature on international delegation (reviewed below). The article cautions against an indiscriminate application of principal-agent models to the study of international organization, as this carries the risk of exaggerating problems of bureaucratic drift while blinding the researcher to delegation problems that are already rooted in the interaction among the principals themselves. A pessimistic reading of this article would therefore emphasize the fact that formal institutions, such as international bureaucracies, which are created in order to deal with cooperation problems, merely offer a new site for these problems to play out. However, this is but the first part of my argument.
The second contribution of this article is to the literature on informal governance (Stone Reference Stone2011; Kleine 2013a, Reference Kleine2014). Rather than viewing regular deviations from formal rules as institutional pathologies, this literature describes informal practices that are inherent to and often even necessary for the daily operation of formal institutions. Specifically, the article builds on Urpelainen’s concept of ‘unilateral influence contests,’ an ever-intensifying zero-sum competition for greater influence on the bureaucracy (Urpelainen Reference Urpelainen2012, 705). In contrast to Urpelainen, and building on my previous work on national fiefdoms, I posit that unilateral influence need not always be zero-sum and may indeed offer opportunities for mutually beneficial exchanges of control over individual aspects of an international organization (Kleine Reference Kleine2013b). However, lacking information about the nature and reach of each other’s activities within the bureaucracy, states will assume that their partners abuse their unilateral influence to their detriment. In order to prevent these suspicions from turning into a unilateral influence contest, it is crucial for states to keep a watchful eye on each other in order to elicit information about one another’s activities. A more optimistic reading of the article therefore highlights how less formal institutions stabilize cooperation even in the face of a regular circumvention of an organization’s formal rules.
The article begins with a discussion of how the possibility of unilateral influence within international bureaucracies creates problems of interstate control. The following section assesses the strengths and limitations of the analogy between delegation problems in coalition governments and international organizations. The theory developed in this section is subsequently illustrated with examples from the European Commission. The article concludes with a discussion of the theory’s scope, scope conditions, and generalizability.
Beyond bureaucratic drift: the problem of interstate control
While the application of principal-agent models to International Relations has shed invaluable light on the relationship between states and international bureaucracies, it has also obscured inefficiencies that are rooted in the interaction among the states themselves. The standard set up of the principal-agent problem is well known and is only briefly repeated here. An employer intends to delegate some tasks to an employee. Their interests, however, are not perfectly aligned and, if given the chance, the employee will act in ways that are not in the employer’s best interest. Anticipating this problem, the employer seeks to reduce informational asymmetries and devise mechanisms that align their interests. This mundane scenario has spawned a large literature on delegation problems in international politics where, analogous to the workplace, scholars argue that states suspect bureaucracies of acting against their best interest (Bendor, Glazer, and Hammond Reference Bendor, Glazer and Hammond2001; Hawkins et al. Reference Hawkins, Lake, Nielson and Tierney2006). There is wide scholarly agreement that states, like employers, anticipate this delegation problem and limit bureaucratic drift through various control mechanisms that are supposed to align the bureaucracy’s incentives with the states’ collective objectives (Pollack Reference Pollack1997, 129).Footnote 2
It is also widely acknowledged that the analogy to the workplace has its limits in the context of international politics, since neither the principal nor the agent is, in fact, an autonomous person in the Kantian sense with the capacity of self-determination. As a large literature points out, the principal in international politics is typically composed of two or more states with heterogeneous preferences and bargaining power – a heterogeneity that, under certain conditions, enhances the agent’s autonomy as it becomes more difficult for states to agree on the use of control mechanisms (Nielson and Tierney Reference Nielson and Tierney2003, 249; Martin Reference Martin2006, 144; see, however, Lake and McCubbins Reference Lake and McCubbins2006, 361–2).Footnote 3
The bureaucracy, too, is not a unitary actor with the capacity to independently choose its actions. While employees are actual persons whose behavior principals can only seek to control from outside through incentives, bureaucracies are permeable structures that offer plenty of opportunities for states to manipulate the bureaucracy’s interests and strategies from inside (see also Elsig Reference Elsig2011; Kleine 2013a, Reference Kleineb). The literature on informal governance describes a myriad of possibilities to bias policy implementation by an international organization through the manipulation of its staff and its funding. For example, states may press for the recruitment of co-nationals with similar preferences (Kleine Reference Kleine2013b, 329–31; similarly, Novosad and Werker Reference Novosad and Werker2018) or they use career incentives and punishments in order to compel compatriots in the bureaucracy to bias policies in their favor (Wonka Reference Wonka2007, 182–3; Voeten Reference Voeten2008, 420; Chwieroth Reference Chwieroth2013). Major powers can threaten to withhold funds or create complicated legal obstacles to the implementation of unpopular policies (Urpelainen Reference Urpelainen2012, 708; similarly Stone Reference Stone2004).
If we accept the premise that neither the principal nor the agent is an autonomous actor in the philosophical sense, the application of the P-A model to international organizations yields a new implication. Given that the level of control of the agent is a function of the heterogeneity of preferences among the states, situations will arise in which states with outlier preferences are permanently dissatisfied with the actual level of control that is ultimately achieved. These states have an incentive to renege on the agreement about the use of external collective control mechanisms. They can also be expected to bypass these mechanisms and use unilateral influence in order to offset the perceived gap in bureaucratic control (Thompson Reference Thompson2007, 10; for an overview see Kleine Reference Kleine2014).
When states use unilateral influence in addition to or instead of external collective controls, this has important implications for our understanding of delegation problems in international organizations. First, it implies that the problem of bureaucratic drift, which is rooted in bureaucrats’ ability to acquire private information, is potentially less severe than the standard principal-agent model predicts, as states have additional means of controlling the agent from inside (similarly, Stone Reference Stone2009). Second, it implies that, under certain conditions, new informational asymmetries may arise in the process of unilateral influence on the bureaucracy, this time among the member states themselves (Kleine Reference Kleine2013c, 249). These asymmetries create a new delegation problem that poses a distinct threat to durability of international cooperation.
To be clear, the very existence of unilateral influence in an information-rich environment need not become a problem. As argued elsewhere, in international organizations dealing with multiple issues, it may even offer the opportunity for states to engage in the mutually beneficial exchange of control. This exchange brings about national fiefdoms in which (groups of) states tolerate each other’s special influence over those aspects of the international organization that they value the most (Kleine Reference Kleine2013b). In addition, one state may simply hide behind the unilateral influence of another state with similar preferences.
However, if states lack information about just how far their partners take their unilateral influence within the international bureaucracy, they will assume the worst. An inefficient influence race, not unlike an arms race in situations of imperfect information (Downs et al. Reference Downs, Rocke and Sieverson1985, 135), might ensue as each state constantly tries to offset the presumed harmful dominance of others. An influence race like this threatens to waste resources and erode cooperation (Urpelainen Reference Urpelainen2012, 716–9). I refer to this situation as the problem of interstate control.
Keeping tabs on cooperating partners
The previous section asserted that cooperation is potentially more fragile than the P-A literature suggests when international organizations, which we assume to be permeable and susceptible to states’ unilateral influence, become but another site for cooperation problems to play out. It was also argued that unilateral influence becomes a problem under conditions of informational asymmetry among the member states that leads states to suspect cooperating partners of abusing their unilateral influence on the bureaucracy at their expense. This section argues that these two features of the problem of interstate control – the fragility of cooperation and the lack of information about each other’s action within the bureaucracy – bear resemblance to problems in the formation and operation of coalitions.
In International Relations, coalitions are characterized by their ephemeral nature. They are typically created in an ad hoc manner in order to achieve a tangible common goal (Riker Reference Riker1962, 159). In domestic politics, legislative coalitions are transitory by design. They are at the latest dissolved at the end of a legislative term and may also lose the confidence of the parliament at any time. The literature in comparative politics, therefore, raises the question of how long coalition governments persist, and why (Diermeier Reference Diermeier2008)? In a seminal paper, Laver and Shepsle (Reference Laver and Shepsle1990) introduce the idea of the structure-induced equilibrium to parliamentary democracies, arguing that a certain allocation of portfolios can create a stable majority core that allows for the formation of government. A growing literature in this field now discusses additional institutional solutions that allow coalitions partners to render their cooperation more durable.
International organizations as coalition governments
In parliamentary democracies, especially those with proportional representation, governments often consist of a coalition of two or more parties. For the purpose of forming a coalition, parties must agree on common objectives (the coalition treaty) they would like to pursue during the legislative term as well as on the allocation of portfolios (e.g. finance, environment) among party representatives (Laver and Shepsle 1990, Reference Laver and Shepsle1996). Parliamentary supporters of a coalition therefore delegate authority not only to their own party’s cabinet ministers, but also to ministers of their party’s coalition partner.
The delegation of authority to ministers from their own party does not pose a problem in this literature as parties are expected to have ample means at their hand to control their own members. However, the parties in government expect to face difficulties controlling the ministers of their coalition partner. Informational asymmetries about each partner’s reign over their respective departments offer opportunities for all coalition partners to bias government policy beyond what has been agreed in the coalition bargain. In other words, there is an agency problem only with respect to those parts of the bureaucracy over which the parties have less oversight.
A similar logic applies to international organizations if we allow for the possibility of unilateral influence within the bureaucracy. Similar to coalition governments, international organizations are created through an international treaty that identifies common objectives and delegates their implementation to an international bureaucracy. If we acknowledge that this bureaucracy is susceptible to unilateral influence, we can posit a delegation problem similar to that faced by coalition governments. While unilateral influence on policy implementation need not be a zero-sum game, it becomes problematic when states have little information about one another’s activities within the bureaucracy and, therefore, have to suspect that cooperating partners are abusing their unilateral influence at others’ expense. It is not the existence of unilateral influence per se, but the fact that states suspect the abuse of this influence, which opens the door to the problem of interstate control.
Keeping tabs on coalition partners
If states lack information about whether or not cooperating partners abuse their influence to an extent that infringes on their own interests, they will seek to match the presumed gap in control with even more unilateral influence of their own. A costly influence contest follows that slowly erodes cooperation after all. Aware of this permanent threat to the stability of government, the comparative literature identifies a number of mechanisms that reduce informational asymmetries and allow cooperating partners to keep tabs on one other’s actions within the bureaucracy.
A direct approach consists of shadowing one’s cooperating partner. Thies (Reference Thies2001) argues that in order to scrutinize a coalition partner’s governing activity, parties create the position of ‘junior ministers’ within rival ministries. Working in close proximity, junior ministers monitor the senior minister’s action in order to prevent her from building up private information within the respective department. In line with the argument made above about the incentives of preference outliers to circumvent collective control mechanisms, Thies (Reference Thies2001, 585–6) predicts that this mechanism is especially relevant when one of the coalition partners holds more extreme preferences on an issue and, therefore, has incentives to depart from the coalition deal (e.g. social democrats on labor issues, or a green parties on the environment). Müller and Strøm (Reference Müller and Strøm2000) provide evidence for the use of junior ministers in coalition governments.
The similarity of the delegation problems in coalition governments and international organizations suggests that we can expect functional equivalents to junior ministers in international bureaucracies. In other words, states will use loyal staff in order to report back on the activities in those parts of the bureaucracy that are dominated by cooperating partners, especially if these partners’ interests differ from their own. As discussed below, the EU member states used different ways to shadow their partners within the Commission, from the practice of triumvirates of commissioners at the political level to informal quotas about a minimum of national diversity in the composition of departmental units at lower administrative levels.
Martin and Vanberg (Reference Martin and Vanberg2005, 97) argue that the time and resource constraints confronting cabinets hinder ministers, especially those of small parties, from monitoring their partners more directly. They propose a public solution to the coalitional delegation problem (Martin and Vanberg Reference Martin and Vanberg2004). In their view, parliamentary scrutiny mechanisms, such as committee oversight or question time, can be used to counteract problems posed by ministerial autonomy. Martin and Vanberg show empirically that while typically thought of as instruments of the opposition, these mechanisms are often co-opted by coalition partners in order to reduce informational asymmetries among them (Reference Martin and Vanberg2005, 97).
Shadowing a cooperating partner can be costly and therefore an option that is mainly available to large and wealthy states (see, e.g., Dijkstra Reference Dijkstra2015). We would therefore expect smaller states to make do with existing mechanisms that they co-opt in order to elicit information about the reach of other states’ unilateral influence. Although many international organizations lack the scrutiny mechanisms available in parliamentary democracies, there are other more public instruments that would allow smaller states to compel their larger counterparts to reveal some of their activities within the bureaucracy. Further below we shall see how coordination meetings among senior commission officials were repurposed in order to spot implementing decisions that could potentially infringe on other governments’ political interests.
Differences between coalition governments and international organizations
How far does the analogy between coalition governments and international organizations travel? Coalition governments differ from international organizations because the former typically have an expiration date that the latter do not.Footnote 4 Furthermore, since the breakup and formation of governments are constitutive of democracies, the transition process in parliamentary democracies is designed in ways that reduce transaction costs, whereas the breakup of international cooperation is often prohibitively costly.Footnote 5 One might therefore object that information is useless without a credible punishment for an abuse of unilateral influence. If a costly influence contest is the only option in response to an abuse of unilateral influence, states might shy away from delegating authority to an international organization to begin with (Urpelainen Reference Urpelainen2012, 706; cf. Manulak Reference Manulak2017).
It should be noted that the threat to end a coalition is a drastic step that, depending on a party’s electoral prospects, is not always credible in the domestic context either. Some scholars argue that government coalitions therefore require some level of hierarchy among the party leaders in order to mediate conflicts (Andeweg Reference Andeweg2000, 383; Strøm and Müller Reference Strøm and Müller1999, 273–5). A similar option is available in the international context where heads of state, at the side of a major summit, often deal with conflicts over more mundane issues of staffing and funding of an international organization.
Furthermore, the fact that international organizations are typically more durable than coalition governments implies that states should care about their reputation. Stone (2011, 45) argues that in these situations of (infinitely) repeated interaction states have an incentive to build a reputation for restraint. If they manage to level informal informational asymmetries among them, we can expect states to reach an equilibrium where they refrain from abusing their unilateral influence to the detriment of their cooperating partners in order to avoid retaliation and a subsequent influence contest that spirals out of control.
One might further object that it is more difficult for states to exert unilateral influence on international bureaucracies than it is for a coalition party to influence a ministry beyond the coalition agreement. After all, ministries are meant to be responsive to changes in political preferences whereas international bureaucracies are typically designed to resist varying political and popular pressure. In this view, international organizations are more similar to regulatory agencies than to government departments (Majone Reference Majone1994). However, this is a difference in degree, not in kind. The insulation of international bureaucracies from unilateral influence is an empirical question, and the literature on informal governance shows that scholars should not underestimate states’ ingenuity when it comes to finding legal or nonlegal ways around the formal rules (Urpelainen Reference Urpelainen2012, 718).
In sum, the similarities of the delegation problem in coalition governments and in international organizations suggest that we can expect states to devise interstate control mechanisms in order to ‘keep tabs’ on one another’s unilateral action within an international organization. The theory implies that these mechanisms emerge in response to an abuse of unilateral influence that infringes on other states’ interests, and that they exist inside rather than outside of the bureaucracy. The following two sections evaluate these implications using the example of the European Commission and anecdotes from other international organizations.
Problems of interstate control in the EU and beyond
Before turning to the emergence and use of interstate control mechanisms, we first establish that the problem of interstate control within international organizations is real and carries the risk of an erosion of cooperation. For example, a problem of interstate control emerged within the United Nations (UN), especially at the height of the Cold War, as national ‘fiefdoms’ in this bureaucracy (Meron Reference Meron1977, 93–100; Reference Meron1991, 322; Ameri Reference Ameri2003, 20–8) engaged in an influence contest in which more and more states ‘[succumbed] to a vicious circle in that no state is ready to be the first to honor [the formal independence of the Secretariat] and, consequently, lose its stake and power’ (Ameri Reference Ameri2003, 9). Reviewing the activities of Soviet nationals within the UN Secretariat in the early 1980s, the U.S. Senate Select Committee on Intelligence concludes that ‘the Soviets will maintain a distinct advantage in the United Nations until their diplomatic and intelligence efforts are matched by systematic, long-term Western opposition’ (United States Senate 1985, 2). Similar anecdotes of how some countries’ unilateral influence on the political and judicial decision-making process prompts other countries to follow suit exist for the World Trade Organization (Yi-chon and Weller Reference Yi-chon and Weller2004, 265–6; Stone 2011, 80–102).
The problem is most pronounced in the case of the EU where confusion about each member state’s reach inside the Commission triggered a race for unilateral influence that threatened to spiral out of control and paralyze the EU bureaucracy.
The European Commission from a principal-agent perspective
Like any other international organization, the EU is based on interstate agreements that spell out the objectives and rules of cooperation. The EU institutions implement these objectives in a decision-making process which can be characterized as follows: The Commission sets the agenda by submitting a proposal for a legislative act. After submission, governments in the Council of Ministers either adopt the legal act jointly with the European Parliament by majority vote, or they amend it unanimously. Finally, the European Commission, national administrations, or both, implement the legal act. The European Commission is, therefore, the EU’s principal bureaucracy, tasked with the formulation and implementation of EU law (Nugent Reference Nugent2010, 122–33). It is composed of an administrative layer of permanent civil servants that is itself divided into several departments (Directorate-Generals (DGs)), and a political layer composed of commissioners.Footnote 6 Commissioners are nominated by the member states for a 5-year term and responsible for a specific portfolio that falls under the remit of one or more DGs. Similar to a government cabinet, the commissioners form a ‘college’ that is headed by the Commission president. The college acts on the basis of collegiality, meaning that its members are responsible for the Commission’s actions as a whole.
The Commission’s powers in agenda setting and implementation are often explained with reference to the standard principal-agent model, according to which the member states, the principals, delegate authority to an autonomous entity, the Commission (Pollack Reference Pollack1997). Specifically, it has been argued that because governments are at risk of reneging on their commitments due to incentives to respond opportunistically to special interests, they delegate authority to an independent agent that is shielded from ad hoc political influence (Majone Reference Majone1994; Moravcsik Reference Moravcsik1998). Viewed from this principal-agent perspective, the main delegation problem that ensues from this act of delegation is bureaucratic drift: the possibility that the Commission abuses its authority against the member states’ best interests. However, Pollack (Reference Pollack1997, 114–6) argues that states anticipate this problem and consequently devise collective control mechanisms in order to align the Commission’s incentives with their own. Among these collective control mechanisms are so-called comitology committees, composed of state officials that monitor the Commission’s executive functions (Pollack Reference Pollack1997, 114–5).
The limits of Commission autonomy and opportunities for unilateral influence
In theory, the Commission and its employees are not supposed to take instructions from their native country, party or any other group. In reality, however, the Commission is but an institution that is not immune to influence from outside and within its own structures. The literature on informal governance points to two channels of unilateral influence: resources and staffing. Since the Commission can partly draw on its own system of resources, unilateral influence inside the Commission typically takes the route of recruiting likeminded staff and rewarding favorable decisions.
A first way for member states to exert unilateral influence inside the Commission is through the co-option of its political level. As Wonka (Reference Wonka2008) argues, states deliberately nominate commissioners that share their interests and are easy to control through career incentives. There is also evidence that member states place ‘their’ commissioners strategically in areas that are of special sensitivity to them (Kleine Reference Kleine2013b). Insiders agree that governments do not shy away from using their commissioners in order to influence the Commission’s internal decision-making process, although, as a former Commission official admits, some countries are better at masking this influence than others (Interview, 18 February 2014). Another former Commission official speaks of ‘intergovernmentalism under camouflage’ (Interview, 14 February 2014), a system in which commissioners use their position, at a minimum, to defend their home country’s ‘red lines’ in the formulation and implementation of EU law (Interviews, 28 February 2014a and b). In light of these testimonies, it is not surprising that Thomson (Reference Thomson2008, 187) finds that legislative proposals are systematically closer to the position of the home country of the commissioner responsible for this proposal than to the position of other member states.
Unilateral influence is not confined to the Commission’s political level. It penetrates the administrative level, the ‘services,’ as well. When this happens, the pressure on the services to bias policy implementation originates from the commissioner’s personal offices, the cabinets (not to be confused with government cabinets), which came to serve as transmission belts between the political and the administrative level. Typically composed of the commissioner’s co-nationals, drawn from national administrations and the commissioner’s party,Footnote 7 they became a means for the home government to keep itself informed about the service’s activities in order to raise, if deemed necessary, objections against decisions in the making (European Communities 1979, 56). The size and influence of the cabinets grew considerably in the 1960s and 1970s (Poullet and Deprez Reference Poullet and Deprez1976, 53; Michelmann Reference Michelmann1978, 495).
Influence contests and problems of interstate control
As the quotes above illustrate, unilateral influence within the Commission has been an open secret and not always considered a problem. However, we argued above that the lack information about a cooperating partner’s reach and purpose of unilateral influence can trigger a costly influence race and, thus, the slow erosion of cooperation. This is what happened a number of times in the Commission.
The first Commission under the presidency of Hallstein in the 1960s gained notoriety for its conflicts with the French President De Gaulle. It is less known that it was also rocked by conflicts among the member states that were mirrored in rifts within the college itself. Arguably the most important conflict among the member states at that time was the organization of the market for agricultural goods. This conflict pitted France and other countries with important agricultural exports against Germany, a net consumer. Whereas France, Italy, and The Netherlands hoped for a quick organization of this market with substantial state intervention, Germany prioritized a quick liberalization of the trade in industrial goods (Moravcsik Reference Moravcsik2000, 5). These differences among the states were mirrored inside the college of Commissioners, with the French and Dutch commissioners defending an interventionist organization of the common market, and the German commissioner for competition, von der Groeben, defending a more liberal market order (Marjolin Reference Marjolin1986, 308; Bitsch Reference Bitsch2007, 203; van der Harst Reference van der Harst2007, 355).
At about the same time, the member states began to worry about their clout within the Commission. An example of this problem of interstate control is a discussion in the 1960s in an internal meeting of the committee of undersecretaries of German ministries (Staatssekretärausschuss). The German ministry of agriculture complains to its colleagues that other countries remained in close touch with their own citizens within the Commission. It worries quietly that this offered their cooperating partners an unfair advantage.
The free circulation of goods in Europe means that the European Commission’s power will grow considerably, especially with regard to the organization of the common agricultural market… With regard to this situation it is necessary to acquire information about the Commission’s initiatives at an early stage and communicate German views on these matters to the responsible services. It is well known that other countries, France, Italy and the Netherlands in particular, have especially in recent years maintained close ties to their citizens inside the Commission. Although we must not exaggerate the influence that stems from these contacts (…), there can be no doubt that it is essential that the German government improve its contacts to German citizens inside the Commission (Bundesministerium für Landwirtschaft 1967, 1–2, my translation).
This quote demonstrates the similarities between delegation problems in coalition governments and the problem of interstate control. Unilateral influence over parts of the bureaucracy is not considered a problem per se. However, similar to parties in a coalition, the ministry worries about their partners’ abusing their unilateral influence on the Commission at Germany’s expense. Lacking adequate information about other countries’ actions, the Staatssekretärausschuss recommends following suit and, thus, engaging in a unilateral influence race.
Interstate control problems became even more problematic during the presidency of the French socialist Delors from 1985 until 1995. Although this era is generally regarded as the Commission’s heyday, Delors’ second term was marked by strong conflicts among the member states that were mirrored in growing rifts within the college of commissioners. Having brought the Single Market on its way, Delors’ commission faced a new bone of contention – strengthening the market’s social aspects. Reflecting their home countries’ market-liberal stance at that time, the commissioner for competition policy, the Irishman Sutherland, and his successor, the Englishman Brittan, became increasingly critical of Delors’ leadership (Peterson Reference Peterson2010). Delors’ cabinet, which was overwhelmingly composed of French nationals, supposedly became increasingly patronizing toward other commissioners and their personal offices (Ross Reference Ross1995, 63–8; Endo Reference Endo1999, 46). His policy stance and behavior increasingly fueled criticism that Delors was abusing his authority in order to pursue French national interests (Cini Reference Cini1996, 190). In reaction to perceptions of growing French influence within the Commission, other commissioners reacted by strengthening their own personal offices and thereby their grip on the Commission services. The total number of the commissioners’ personal staff consequently spiraled out of control during this time. By 1989, personal staff exceeded 300 in total or on average 25 staff members per commissioner, compared to an average of 14 in the 1970s. This expansion was only possible because the commissioners circumvented official quotas for the size of cabinets, for example, by employing staff on the payroll of national governments (Ludlow Reference Ludlow1991, 93). In other words, the college became the site of another influence contest among the member states. As one senior official explained, ‘[Delors’] cabinets are very active, and other cabinets are responding’ (Peterson Reference Peterson1999, 56).
In light of the Commission’s experience with influence contests, it is not surprising that a large part of the cabinets’ responsibility is to keep track of the activities of other commissioner’s personal offices (Interview with a former senior Commission official, 28 February 2014). If they were mainly tasked with supporting the commissioner in her area of responsibility, we would expect cabinets to be composed of policy experts in this particular field. Yet, cabinets comprise primarily people that are experts in precisely the fields that do not fall under their commissioner’s remit (Spence Reference Spence2006b, 62). Their role has consequently been likened to that of internal spies.
A Commissioner’s cabinet is not only interested in its own portfolio; it also keeps a watchful eye on the portfolios of other Commissioners. [The cabinet member] has to be a kind of internal spy. To do this job, he has to know what is going on in the DG—and this is not always straightforward… Ultimately, a mandarin has to detect draft proposals of a possible competitor far in advance, because information, if timely received, is power (Eppink Reference Eppink2007, 115–6).
In sum, the Commission is fraught with the problem of interstate control, which is rooted in the member states’ unilateral influence over parts of their administration. Several times in the Commission’s existence, this problem spiraled out of control and into a costly influence contest among the member states over their administration. To be sure, some governments are less aggressive than others in pursuing their national interests within the bureaucracy, and some staff might be less susceptible to unilateral influence than others. However, it is apparent that unilateral influence exists within the walls of the Commission and, when combined with a lack of information, it threatens to paralyze the EU bureaucracy through inefficient influence contests.
Interstate control mechanisms in the European Commission
The example of coalition governments suggests that it is possible to contain influence races and level informational asymmetries among coalition partners by shadowing one another (junior ministers) or publicly questioning ministers about their actions (parliamentary scrutiny). Similar mechanisms can be found within the European Commission.
‘Junior ministers’
Coalition partners can shadow the senior minister in order to monitor a potential abuse of her discretion. Functional equivalents of ‘junior ministers’ already existed in the early years of the Commission. In the 1960s, each commissioner was given prime responsibility over one specific policy and ancillary responsibilities for one or two more policy areas (Spence Reference Spence2006b, 62). During the Hallstein presidency, the commissioner for agriculture, the Dutchman Mansholt, would therefore have to consult with the German commissioner von der Groeben and the Luxembourgian Schaus (Lindberg Reference Lindberg1963, 72). This mechanism became impractical and eventually defunct with a growing workload in the late 1960s that made it impossible for commissioners to take on responsibilities beyond their core remit (Spence Reference Spence2006a, 152). It is at about the same time that the aforementioned unilateral influence race gains in speed within the Commission.
A similar mechanism surfaced at the intersection between the Commission’s political and administrative layer. It became an informal norm in the 1960s that if the commissioner was of a certain nationality, the most senior civil servant in his jurisdiction, the Director General, was expected to be of a different nationality (Lindberg Reference Lindberg1963, 72; Coombes Reference Coombes1970, 132). The same norm is applied within the DGs between the Directors General and their deputies. If this norm were merely meant to strike a certain national balance among senior ranks within each DG, we would expect the same norm to apply also between the commissioner and the deputy Director Generals. However, it is not uncommon for co-nationals to serve in the senior ranks of the same or overlapping policy areas.Footnote 8 It seems intuitive that the norm is not primarily meant to guarantee a fair national balance. It is plausible to argue that its primary function is to prevent the buildup of private information about the reach of unilateral national influence within the higher echelons of the bureaucracy.
Administrative reforms in the 1990s served a similar purpose. Commission president Prodi insisted that either the chefs de cabinet (the heads of the commissioners’ personal offices) or her deputy be of a different nationality than the commissioner, and that the commissioner’s personal offices comprise at least three different nationalities (Agence Europe 1999). These quotas are insufficient to attain a national balance within the cabinets, but they appear to be geared towards preventing informational asymmetries about the commissioners’ activities. According to insiders, the reforms led to the emergence of informal networks across the personal offices of the different commissioners that substantially improved the flow of information among them (Interview with a former senior Commission official, 28 February 2014b; Interview with a senior Commission official, 6 March 2014; Interview with a senior member of cabinet, 11 March 2014). Although they did not prevent commissioners from giving the policy of their responsibility a certain national imprint (Thomson Reference Thomson2008), the reforms allow others to remain informed about just how far this bias might go.
‘Parliamentary scrutiny’
Martin and Vanberg propose a second way for parties to police their coalition bargain. They argue that because some coalition parties face resource constraints, they instead use open debates and question times in order to ‘“check up” on their partners and, if necessary, to ‘correct’ the actions of hostile ministers’ (Reference Martin and Vanberg2005, 97). Similarly, some states face limits in monitoring the actions of the Commission’s highest ranks. Each year the Commission prepares several hundred legislative proposals, and it takes many more decisions regarding the implementation of policies. This workload offers ample opportunities for commissioners to frame issues in such a way as to avert close scrutiny (Wille Reference Wille2013, 82). It is therefore not surprising that we can find functional equivalents of the ‘parliamentary scrutiny’ mechanism in the Commission. There are at least two opportunities to vet each commissioner’s actions through open debate before draft texts are submitted to the college of commissioners. In regular meetings (called special chefs), the members of the cabinets that are responsible for a specific item on the agenda have the opportunity to raise questions about all aspects of a draft text. These meetings are followed by weekly meetings (called Hebdos) where the chefs de cabinet discuss whether or not the draft text is ready for the college of commissioners to adopt it without an open debate.
One might object that this mechanism primarily serves the internal coordination of policies within the Commission. This might well be the case, and we should not forget that even parliamentary scrutiny mechanisms in coalition governments are primarily a means of the opposition to control the government that are nevertheless co-opted by coalition partners in order to keep themselves informed about their partners’ actions. Thus, while debates at the cabinet level may serve the coordination of policies, they are also used in order to scrutinize for the degree of unilateral influence that went into them and their likely political repercussions in other member states. A senior cabinet member confesses that it is one of his main tasks in these meetings ‘to spot dossiers that proceed below the radar level’ but might stir up national sensitivities at home (Interview 11 March 2014). David Spence, an intimate observer of the Commission, concurs that it is the purpose of these meetings to avoid intense biases that impinge upon one of the member states’ interests.
It should be remembered that [in these meetings] the cabinets will defend both their own Commissioner’s interests and the Commissioner’s national interest. They will be fully briefed by ‘their’ permanent representation and national officials about what is at stake (Spence Reference Spence2006a, 151).
In sum, several mechanisms within the Commission, akin to the monitoring tools of coalition governments, allow the member states to keep the problem of interstate control in check. By reducing informational asymmetries about the reach and purpose of one another’s unilateral influence, they are able to build a reputation for restraint in the sense that states use their unilateral influence only up the point that it infringes on their partners’ interests. According to a contemporary observer, defending one’s home country’s interest within the Commission is a generally tolerated practice ‘unless it becomes too blatant and obstructive’ (Interview 14 February 2014; Interview 28 February 2014a).
Discussion and conclusion
The objective of this article was twofold. First, it identified a bias in standard principal-agent models toward the analysis of bureaucratic drift and a concurrent blind spot for delegation problems that are rooted in the interaction among the principals themselves. It was argued that the possibility of unilateral influence within an international bureaucracy implies that bureaucracies might be less autonomous than generally believed. At the same time, unilateral influence creates new informational asymmetries among the member states and the challenge for them to identify actions that potentially harm their own interests. Second, the article made the case for approaching international organizations as coalition governments in which cooperating partners seek to distinguish appropriate from inappropriate unilateral influence. It was argued that states stabilize the potentially brittle cooperative equilibrium by keeping a watchful eye on one another’s actions within the shared bureaucratic apparatus. I illustrated the plausibility of these claims using the case of the European Commission and anecdotes from the United Nations.
One might object that states anticipate the problem of interstate control in their institutional design and devise mechanisms to prevent states from exerting unilateral influence in the first place (see, e.g., Manulak Reference Manulak2017). However, it is questionable that institutional designs can ever be perfect (Koremenos, Lipson, and Snidal Reference Koremenos, Lipson and Snidal2001, 1082). If we then accept the fact that states are unable to anticipate and thwart all opportunities for unilateral influence, we are left with the question of when and where the problem of interstate control arises and overshadows the problem of bureaucratic drift.
Are all international organizations then bound to develop problems of interstate control or is this problem limited to the case of the EU? For reasons of simplification, we assumed that any international organization is in principle susceptible to unilateral influence. Given that states’ unilateral influence can but need not be zero-sum, the problem of interstate control emerges when states lose sight of the nature and reach of their partners’ activities within the bureaucracy. In other words, the problem is rooted in informational asymmetries among the member states and is, from a theoretical perspective, not limited to a specific international organization. However, it can be argued that informational asymmetries are a function of the formal institutional architecture. In a bureaucracy as vast and as complex as the European Commission or the UN General Secretariat, we would expect it to be easier for states to hide the true purpose of their actions and more difficult for states to keep track of their partners’ activities. In international organizations that deal with fewer issues and possess a leaner administration, the extent and purpose of a state’s unilateral influence should be more immediately apparent.
Is unilateral influence a privilege of the powerful?Footnote 9 Having discussed institutional opportunities for covert unilateral influence, let us now take a look at states’ incentives and resources. For states to step up their unilateral influence on the bureaucracy, they must be dissatisfied with the actual level of collective control. This happens when their preferences become more extreme relative to the member state mean. Dissatisfied with the actual level of bureaucratic control, they will seek to make up the perceived control gap through influence of their own. We saw this happening in the EU when growing tensions among (groups of) member states translated into rifts within the Commission. This conflict went hand in hand with a contest for greater unilateral control on the bureaucracy.
While all states may develop incentives to step up their unilateral influence, they may not all have the capacity to do so. Indeed, if unilateral influence were cheap, bureaucracies would be expendable. In other words, certain forms of unilateral influence must be considered a privilege of powerful, affluent states. The same holds true for surveillance mechanisms. As argued above, it will be easier but arguably less effective for smaller states to co-opt existing scrutiny mechanisms than to shadow their partners’ actions in other departments. More research is needed on the types of unilateral influence and surveillance mechanisms, but it seems safe to say that interstate control problems largely play out between larger states, with smaller states struggling to remain up to date.Footnote 10
Acknowledgments
The author thanks Tobias Hofmann, Helen Milner, Lukas Obholzer, Robert Thomson, Johannes Urpelainen, the members of the 2013 Workshop on International Secretariats in Oxford, the 2014 ISA workshop ‘Influencing International Relations’, and the participants of the 2015 PEIO, EPSA and APSA conferences for helpful comments that substantially improved this article. The author is grateful to Fergal Treanor for excellent editing. The article was finalized during her time as a Jean Monnet Fellow at the RSCAS of the European University Institute.
