1 Introduction

Since the end of the cold war, more and more tasks have continuously been delegated to International Organizations (IOs) (Lenz et al. 2014: 147), and in order to fulfill these tasks, their administrative bodies have grown considerably (Rittberger and Zangl 2012). International public administrations (IPAs) have become an essential feature of global governance, contributing to what some have described as the ‘bureaucratization of world politics’ (Barnett and Finnemore 2004: 165). While we thus know that IPAs matter somehow for world politics, we neither know exactly to what extent nor how they matter for IOs’ policy-making processes and outputs subsequently. Rather, these aspects are still a matter of intense academic debate (Bauer and Ege 2016; Eckhard and Ege 2016; Cerny 2010). Are IPAs to be considered actors of their own right? If so: Which ones? And how exactly does their role matter for policy-making beyond the nation state?

Assessing IPA influence in a comparative fashion has proven to be tricky. On the one hand, case studies of several IPAs, their involvement in and influence on IO policies and politics have given valuable insights into the various ways IPAs might matter for concrete policies. IPAs can influence IOs’ substantive policies, as for example, economic programs by the International Monetary Fund and the World Bank (Sharma 2013; Vetterlein 2012) or the emergence and change of the World Health Organization’s global AIDS initiative (Harman 2011; Nay 2011), the implementation of IO policies (Stone 2008; Xu and Weller 2008), as well as institutional policies that affect the design and change of IOs (e.g., Weaver and Leiteritz 2005; Hanrieder 2014; Johnson and Urpelainen 2014). Yet, so far there is no concept of analysis that would allow for a proper comparison of IPA influence and agency across different IOs. Rather, its detection as well as its description is assessed idiosyncratically, based on case- or field-specific interpretations of cases in which bureaucratic influence was successful (i.e., actually visible).

On the other hand, in order to fill in this lacuna, a prominent strand of recent research has focused on the formal autonomy of IOs and IPAs (Haftel and Thompson 2006; Hooghe and Marks 2015; Bauer and Ege 2016). However, the implicit assumption in many studies of this kind is that formal discretion is automatically exploited by IPAs in order to pursue own objectives deviating from those of their political masters. In short, it is assumed that IPAs will actually make full use of their formal room for maneuver. In contrast to this conception, we argue that the presence of autonomy does not automatically mean that bureaucracies also seek to exert influence on policymaking. Likewise, the absence of autonomy does not necessarily imply that bureaucracies dispense with the development of informal activities to overcome their formal autonomy limitations. Put simply: knowing about what IPAs are able to do in principle (their discretion) does not tell us much about what they actually do.

To strike a balance between comparability and close attention to the day-to-day workings of IPAs, we propose the concept of administrative styles. We start from the proposition that when an IPA regularly intends to influence the policy making process (and policy outputs subsequently), we should be able to identify patterns of informal administrative behavior directed at accomplishing this objective – even if, at times, the endeavor might fail. We interpret these observable patterns as relatively stable orientations, which gradually emerge over time in response to different internal and external challenges for IPAs. These orientations correspond to an ideal typical characterization of IPAs’ ‘styles’ as entrepreneurial or servant-like. An entrepreneurial style presupposes administrative routines that entail intensive bureaucratic policy advocacy and a strong orientation towards institutional consolidation. A servant style, by contrast, presumes a reactive and instrumental role with the administration following a routine pattern of operating ‘by the books’ and refraining from any attempt to intervene with politics beyond its formal tasks and mandate. By focusing on these informal procedures, we present a novel way for systematically assessing IPA influence across different organizations and different issue areas that goes beyond formal differences.

Previous research did not yet yield a consistent explanation as to why and when IPAs intend to exert autonomous influence. Why are some IPAs more actively trying to influence the policy process than others? We argue that administrative styles vary along two dimensions, namely the internal and external challenges a bureaucracy faces. This approach allows for more systematic and general statements about the factors determining the potential policy influence of IPAs against the backdrop of the structural and institutional context in which they operate. A focus on administrative styles may thus enhance both our theoretical understanding of the determinants of IPA agency and enable the conceptually guided comparison and measurement of IPAs over time and across the IO landscape.

Based on these analytical starting points, this article contributes to the conceptual, theoretical and empirical understanding of IPAs and their influence in the following way: we present a novel conceptualization of administrative styles as bureaucratic agency and develop theoretical explanations for variation in administrative styles across different organization (sections 2 and 3). Second, we introduce our approach of assessing the observable manifestations of administrative styles (section 4). Third, we contribute empirically by exemplifying our concept and theoretical argument in two comparative case studies on four international bureaucracies: the International Monetary Fund (IMF) and the Bank for International Settlements (BIS); and the International Labor Organization (ILO) and the Food and Agriculture Organization of the United Nations (FAO), respectively (section 5).

2 IPA influence and the concept of administrative styles

Administrative styles can generally be defined as informal routines that characterize the behavior and activities of public administrations in the policymaking process (Knill 2001; Knill and Grohs 2015). In this article, we concentrate on those characteristics and behaviors that are associated with IPAs’ ambitions to influence IO policies.

We conceive of administrative styles as relatively stable behavioral orientations characterizing an organizational body. It is an institutionalized informal modus operandi that materializes as a guiding principle over time and by repetition, routinization, and subsequent internalization. Under conditions of uncertainty and complexity, individual bureaucrats develop routines for coping with shortages of knowledge, information-processing capacities, and time (Simon 1997). Similarly and depending on their underlying rationale, administrators can develop and internalize behavioral patterns sought to influence their organization’s policies (Wilson 1989; Knill 2001). We interpret these observable patterns as corresponding to an ideal typical characterization of IPAs’ ‘styles’ in shaping IPA behavior. Three characteristics of this concept stand out:

First, rather than restricting our analytical focus on an IPA’s formal position we are thus interested in the extent to which an IPA developed informal routines that allow it to exert influence beyond formal rules or whether its informal activities remain in line with or even behind its formal position. This focus on informality and informal governance (Stone 2013) allows us to go beyond structural features of bureaucracies, such as their mandate, political autonomy, size, organizational structure, resources, tasks, and staff composition. Instead of solely considering formal autonomy and structure, we hence perceive of bureaucratic agency as the extent to which IPAs make use or even exceed the margin of maneuver that is formally given to them (Martin 2006). In other words, we do not assume a straightforward link between an IPA’s formal autonomy and discretion and its actual actions as usually done in principal-agent models of administrative behavior. Administrative styles describe behavior that is precisely not ‘by the books’. Thereby the extent to which IPAs seek to influence the policy process becomes comparable across a vast array of organizations, which otherwise could not be compared due to idiosyncrasies. If we look at whether or not an IPA regularly exceeds the formal requirements and rules or rather adhere to them closely, the underlying formal rules fade from the spotlight without being rendered irrelevant.

Second, given the predominance of state power in the IO system, the autonomous influence of IPAs may be overshadowed by state interests. It has proven to be very challenging to dissect the concrete influences of different actors on a given output (Zürn et al. 2012). Instead of trying to disentangle those interactions completely and to isolate the share of IPA influence on policy outputs, we depart from the proposition that if international bureaucrats intend to influence the policymaking process (and outputs subsequently), we should be able to identify patterns of regular administrative behavior directed at accomplishing this objective. An IPA’s general inclination towards exerting policy influence renders it a relevant actor in the policymaking process, even if these attempts are more or less successful and even fail from time to time. By conceptualizing IPAs’ styles and influence this way we avoid the selection bias inherent to studies that restrict their focus only to constellations in which IPAs’ influence is clearly visible; i.e., bureaucratic action was found to have made a difference for policymaking beyond the nation state (see Eckhard and Ege 2016).

Third, in contrast to national style typologies (Knill 2001; Richardson 1982; Vogel 1986), we distinguish between two distinctive orientations underpinning administrative styles, namely institutional consolidation and policy effectiveness. Thereby, the concept does not only capture IPAs’ potential policy influence, but also specifies the areas and goals of informal activity IPAs emphasize over others. IPA influence can unfold in different ways according to whether it is driven by positional or functional orientations.

If administrative styles are shaped by a positional orientation, the central objective behind administrative routines is on the question of how a certain policy will affect the autonomy and legitimacy of the IPA. Positional administrative routines are directed towards influencing the institutional conditions under which IPAs operate when interacting with their organizational environment and political principals (Knill and Bauer 2016: 951). The provision of effective solutions to policy problems is less important than positional improvements or – in the extreme case – safeguarding the organizations raison d’être. Effective problem-solving is thus no dominant orientation per se, but mostly considered as a means to the end of positional achievements.

With a functional orientation, IPA behavior is directed at constantly advancing the policy performance of an IO. The dominant focus is on optimizing internal structures and processes as to facilitate the initiation, formulation, and effective implementation of well-designed policies (Meyer et al. 2014). IPAs’ routines and standard operating procedures are thus directed towards effective problem-solving; behavioral elements ensuring functionality are more pronounced than elements directed towards positional achievements. This does not mean that positional orientations are completely absent. However, they are less emphasized and important to the IPA only insofar as they help to improve policy effectiveness.

Based on the distinction and prevalence of positional and functional orientations in IPAs’ behavioral patterns, we suggest an ideal-typical distinction of four administrative styles that apply to administrative routines of shaping the policymaking process. In addition to the extent to which an IPA becomes politically active on its own, our typology thus includes the more specific orientation of their actions.

To begin with, we differentiate between ‘servants’ who do not seek to exert autonomous influence on the IO’s policies and three types of more influence-orientated IPAs. We further specify the three influence-orientated types depending on the extent to which their activities are targeted to influence the IPA’s position within the organization and towards their principals (positional orientation), the policy output of their organization, such as the content, effectiveness, or existence of programs and policies (functional orientation), or both. In conceptualizing administrative styles that way, we arrive at four administrative styles:

  • A servant administration presumes a rather reactive and instrumental role. The administration refrains from any attempt to intervene with politics beyond its formal duties and does not attempt to influence the policy process.Footnote 1 In this case, the bureaucracy follows a routine pattern of operating ‘by the books’, strictly adhering to the formal procedural and legal arrangements that define its tasks and functions.

  • Advocates focus their activities on influencing those aspects that are directly related to the quality, internal consistency, and effectiveness of their policies. Such bureaucracies will place lesser emphasis on behavioral patterns directed at safeguarding or advancing their institutional position, but advocate for their approach towards the achievement of substantive policy goals.

  • Consolidators concentrate primarily on further strengthening their political autonomy, status, size, and competencies. Their main interest lies with the increase of competencies as such, i.e., the growth of the policy portfolio is given priority over policy consistency. Consolidators are hence primarily driven by positional rather than policy interests.

  • An entrepreneurial style is the combination of the two types above and thus presupposes administrative routines that entail intensive bureaucratic advocacy in policymaking as well a strong orientation towards institutional consolidation in order to strengthen the administration’s position.

3 Explanations of administrative styles

The ideal-typical distinction of different administrative styles remains nothing more than an intellectual exercise as long as we cannot come up explanations for variation in administrative styles. Basically, any systematic explanation should be able to account not only for changes in administrative styles over time, but also for variation of administrative styles across different organizations. While the theoretical arguments developed in this section speak to both of these claims, the put particular empirical emphasis on comparing different IPAs rather the investigation of style changes over time. Why are the routines and standard operating procedures in some IPAs more entrepreneurial, while others are more servant-like, as will be shown in the next section?

First, to answer these questions, structural factors such as the polity of the organization, including its mandate, decision rules, resources, size, and constituency were theorized to explain IPAs’ influence (for an overview see Eckhard and Ege 2016). Yet, many of these theories imply that IOs operating in comparatively similar environments and structural arrangements will be equally influential and exhibit the same style. Since one of our main objectives is to show how IPAs work their way around the ‘by the books’-logic, this argument is rendered of limited use. Some authors have pointed to the structure and type of the problems addressed to account for IPA influence on certain policies (ibid.). Yet, this situational, output-centered approach stands in stark contrast to the stable, procedural patterns of behavior we interpret as styles. Although it is well possible that attempts of influence are more or less successful depending on the policies in question, it does not explain the underlying orientation, i.e., the administrative style.

In contrast to structural or situational (Broome and Seabrooke 2012; Chorev 2012) approaches, we argue that IPAs are subject to external institutional challenges but at the same time may have the internal potential to actively influence their environment(Oliver 1991). Actions and behavior are a product of the interplay of organizational agency (internal) and the demands of the surrounding institution, i.e., the organizational environment (external). An IPA’s experience with institutional demands varies depending on how external and internal pressures interact (Greenwood and Hinings 1996). In line with this thought, we argue that administrative styles vary along two dimensions, namely the external and the internal challenges bureaucracies face. While the former refer to the perceived political and domain challenges faced by the administration, the latter capture the degree of cognitive slack and internal epistemic contestedness of an IPA. Table 1 summarizes the determinants of administrative styles, the respective indicators, and operationalization.

Table 1 External and Internal Challenges – Indicators and their Operationalization
Full size table

3.1 External challenges to the bureaucracy

External challenges are defined by the extent to which the IPA perceives its institutional status and operation as subject to intervention and exogenous pressures. It makes a difference for administrative styles whether a bureaucracy is saturated and stable or whether it deems its status as effectively put into question.

Crucially, we contend that an IPA’s perception of external challenges is better suited to explain their style than merely considering factual structural features of their environment. While capturing relevant structural features of the organizational field, this view accounts for the fact that organizational challenges cannot be expected to matter per se, i.e., unfiltered by the organization. As every problem, challenges need to be defined and acknowledged by the organization in order to have an effect. ‘Conflicting institutional demands in a given field are not experienced in a similar way by all organizations since field level institutional processes are filtered and enacted differently by different organizations.’ (Pache and Santos 2010: 11). Even when external pressures are similar, IPAs need not necessarily respond in exactly the same way (Greenwood and Hinings 1996) as they may perceive and assess their environment as well as their own position in it differently. Two kinds of perceived challenges are of importance to explain administrative styles.

First, external challenges are affected by the perception of dynamics within the organizational domain the administration is located in. Within a given domain or jurisdiction, different organizations will compete for resources and competences. Failures to be innovative and adapt may result in budget cuts and a loss of political attention, thus ‘allowing some organizations to flourish and some to languish or, less frequently, to die’ (Peters 2010: 203; Adam et al. 2007).

Second, bureaucratic vulnerability should increase with the extent to which an organization considers itself to be politically challenged. On the one hand, the more an IPA perceives of itself as being in the political spotlight; i.e., in a position highly salient on the agenda of its political principals (e.g., because the IPA deals with a highly salient political problem or is accused of working inefficiently or ineffectively), the more the administration will try to consolidate its status. On the other hand, the more homogeneous the IPA considers the expectations with regard to their activities and performance (Hawkins et al. 2006), the more it has to fear political intervention. Conversely, a pronounced awareness of heterogeneous principal interests, such as, for example, between donor versus receiving or democratic versus autocratic member states, offers the bureaucracy manifold opportunities to play off political actors against each other, thereby effectively making use of ‘institutional complexity’ (Vermeulen et al. 2016).

3.2 Internal challenges to bureaucratic policy capability

The level of internal challenges a bureaucracy faces strongly affects its ability to pursue clear and consistent policy targets. We consider those IPAs as internally challenged whose policy capabilities are limited by internal disparities or cognitive ‘straightjackets’.

In order to fully realize their own informal potential, IPAs need to have sufficient ‘cognitive slack’, i.e., time, space, and resources to communicate, think and innovate. Similarly to the preconditions for bureaucratic expertise and expert authority (Barnett and Finnemore 2004; Beach 2004) and the concept of ‘ideational capacities’ (Broome and Seabrooke 2015: 958), cognitive resources refer to the degree to which an organization prioritizes research and emphasizes and values expertise in the given issue area and beyond. Cognitive slack also presupposes a sufficient amount of time available to professional staff to create and process specialized knowledge. The space to produce high-quality cognitive outputs is diminished the more principals control and interfere in the day-to-day work of the bureaucracy (Boschken 1988). Are policy products developed jointly with a high degree of exchange between bureaucrats and principal representatives or is the role of political masters confined to place an order and control the end-result? Put simply, this relationship can be summarized as follows: the more control the principal is equipped to exert, the more the IPA will be internally challenged with regard to its cognitive slack, i.e., the capability and potential to initiate, design, and promote effective and sound policies for their IO.

Second, we expect internal challenges to increase with the extent to which staff members’ professional backgrounds are characterized by heterogeneous epistemic beliefs (DiMaggio and Powell 1983). The more staff members share the same (professional) background (e.g., in medicine or economics), the more we should expect a strong, stable, and consistent policy orientation (Nelson 2014; Momani 2005). The same holds true for the ratio of professional staff and the permanence of occupation. Conversely, internal heterogeneity can lead to internal contestation of belief systems impairing the IPA’s ability to act concertedly and consistently (Bauer and Ege 2016).

The external challenges dimension primarily affects the positional orientation of an IPA. In the absence of perceived political or domain challenges, we do not expect an IPA to pursue a strong positional orientation towards institutional consolidation. Strengthening the IPA’s institutional position will only be necessary if its standing or existence is seen to be under threat. The more the status of an administration is under political scrutiny and the more its political support is externally challenged, the more consolidation-driven the bureaucracy must act to improve or safeguard its institutional position. The presence or absence of strong internal challenges, on the other hand, determines an IPA’s functional orientation. Without a certain appropriate cognitive resources, space, and time an IPA will be unable to shape policy outputs. It makes a difference for administrative styles whether the bureaucracy is capable of developing an interest in the solution of certain policy problems or whether it is by default restricted to merely ‘administrating’ these policies.

Flowing from this logic, the combination of low external and high internal challenges renders an IPA servant-like. When the administration’s hands are tied and its position is stable, it neither perceives the need for positional consolidation, nor does it command the means to become active in terms enhancing policy effectiveness. When external challenges increase, however, and the status of the organization or its bureaucracy is called into question, the IPA reacts and aims to stabilize its position by attempting to consolidate its institutional standing. Since internal challenges do not allow for the development of administrative routines reflecting a strong functional orientation towards policy effectiveness, the IPA will resort to making the best of its situation in positional terms. The opposite scenario applies when an IPA is saturated and stable internally (low internal challenges) and external challenges are absent. Equipped with sufficient cognitive slack and shared belief systems, the IPA can form a strong functional orientation emphasizing policy effectiveness. Policy advocacy can unfold rather unimpeded internally and no perceived external threat puts them into a position in which they also have to also consolidate themselves. Finally, the combination of high external and low internal challenges lead an IPA to behave as an entrepreneur pursuing consolidation and policy advocacy at the same time. Low internal challenges facilitate the emergence of a strong functional orientation in administrative routines, while at the same time external challenges require the IPA to direct its activities to positional matters of institutional consolidation. Table 2 shows how the four ideal types relate to the two determinant dimensions.

Table 2 Four types of administrative styles
Full size table

4 Empirical manifestations of administrative styles

Empirically, administrative styles can manifest themselves as observable patterns of administrative ambitions to influence policies and institutions. Depending on whether an IPA’s orientations are positional, functional, both, or neither, we will observe different behavioral patterns all the way through the policy making process.

In line with our theoretical argument, we are less interested in IPAs’ formal competencies and activities but concentrate our attention to the informal aspects of day-to-day administrative routines. The question then is not how powerful an administration’s tools are per se, but whether and how the bureaucracy makes use of these possibly scarce formal powers. We hence assess the extent to which informal patterns (given a certain formal status quo) indicate the actual willingness and capacity of an administration to go beyond what is formally required to secure (future) influence.

To identify differences in informal administrative behavior, we focus on behavioral patterns during all central stages of policymaking, namely policy initiation, policy drafting, and implementation. In so doing, we distinguish nine activities that correspond to a positional or functional orientation respectively. IPAs might engage in all these behaviors albeit to varying degrees. We therefore contend that it is the emphasis on a certain activity, not exclusively its presence, which constitutes a certain behavioral pattern. At this, for each of the activities described in more detail below, we assess whether the IPA engages in this kind of behavior on a regular basis or only very rarely. Consequently, we denote a given IPA’s efforts as ‘high’ or ‘low’ with regard to a given activity. In addition, where we do not see a clear behavioral pattern because the activity in question is taken up only on occasion or unevenly across a given department, we assign a ‘medium’ value. Decisive then for the assessment of the IPA’s style is whether ‘high’ or ‘low’ values outbalance the others in each the ‘positional’ or ‘functional’ sets of indicators (summarized in the two blocks of rows in Table 3).

Table 3 Observable manifestations of administrative styles
Full size table

A functional orientation becomes manifest in observable IPA behaviors that are associated with constantly advancing the policy performance of an IO. The central focus is on those activities that are supposed to optimize internal structures and processes and facilitate the initiation, formulation, and effective implementation of well-designed policies (Meyer et al. 2014). First, for IPAs orientated towards policy effectiveness, we should observe active engagement in identifying policy problems and initiating policy developments ‘from the inside’ by accumulating policy-related knowledge. Of central interest here is thus whether or not the IPA routinely places a special emphasis on both working to uncover and make fruitful new problem areas and – subsequently – attempting to bring them on the organizational agenda. Second, with regard to policy quality and content, optimizing strategies of solution search when drafting proposals should be in focus. In the optimizing mode, administrations evaluate a large number of different alternatives and go with the solution they deem best, even if not the easiest. This strategy differs from the satisficing mode, which would merely require the use of simple heuristics and ‘rules of thumb’ (Simon 1997). We would consider an IPA to ‘score high’ at this if it usually makes pronounced efforts to find and pursue the first-best, technically most sound solution. Third, optimizing policy effectiveness requires well-functioning internal coordination mechanisms of the bureaucracy. Optimizing routines demand more of administrations' horizontal coordination mechanisms and require a departure from the default patterns of negative coordination, which are characterized by a highly fragmented drafting process. In negative cooperation, the organizational unit in charge merely has an eye on whether a given option has a decidedly negative impact on other IPA units. By contrast, in its purest form, positive coordination requires determining which option brings most benefits to all units and actors involved (Scharpf 1994). A departure from the negative mode in our sense thus means that informal efforts are made to make cooperation as mutually beneficial as possible on a regular basis, e.g., by actively reaching out to colleagues, bringing everybody to the table, or internal networking. Fourth, IPAs orientated towards policy effectiveness should ambitiously use all conceivable means to evaluate their policies and improve knowledge management beyond the data that is formally required. As opposed to barely following the evaluation guidelines or formal procedure, following the evaluation guidelines strictly while informally seeking to improve knowledge management would thus be considered a sign of a functional orientation. Lastly, in the functional spirit, administrations should engage not only in monitoring and enforcement, but also in enhancing the impact and effectiveness of their programs through capacity-building and policy promotion. We should observe IPA activities to strengthen the implementation capacities of the relevant actors, engaging in proactive communication with member states that goes beyond their formal duties.

By contrast, positional elements of administrative behavior come to the fore for IPAs that are to a lesser extent orientated towards enhancing the quality and outcomes of IO policies per se but primarily pursue routines of securing their institutional standing vis-à-vis their principals. This positional orientation becomes visible in behaviors that aim at stabilizing or enhancing political support for the organization – be it from political principals (i.e., the member states), important societal actors, NGOs or any other actor whose political support may be valuable to the IPA. We suggest the following behavioral patterns in the policymaking process to indicate an underlying positional orientation.

First and foremost, the position-orientated behavior entails patterns of anticipation. In the policy initiation phase, the anticipation of political feasibility of an issue is key to guarantee alignment with political interests of the principals (Aberbach et al. 1981; Mayntz and Derlien 1989). Pronounced anticipation requires IPAs to routinely stay very sensitive of their principals and environment, and very aware of the political implications of their own actions. Keeping track of principals’ preferences informally at an early stage can also be expected in policy formulation. In order to prevent later political rejection by member states, IPAs will proactively anticipate political ‘red lines’ and subsequently make use of political facilitation mechanisms (e.g., strategies of depoliticization). We label this politically sensitive behavior ‘mapping of the political space’. A third way to shape the political agenda and strengthen the institutional position is to mobilize political or societal support. Administrations with a positional orientation should thus be involved in frequent communicative efforts with a broad range of external actors like interest groups, NGOs, policy networks, or the media. Lastly, depending on the nature of the institutional interests, we expect to observe the strategic use of IO power in policy implementation. Regardless of the particular approach chosen, this presumes that the bureaucracy acts ‘politically’ with regard to the aforementioned actual bureaucratic use of formal powers. Enforcement and monitoring routines are then guided by strategic considerations of the administration, either by attempting to go beyond formal duties or deliberately shying away from using their powers.

Table 3 displays how the different observable behavioral emphases relate to the respective ideal-types of administrative styles. Entrepreneurs, according to our definition, should ideal-typically equally incorporate the positional and functional features specified above. Consolidators, by contrast, should predominantly focus on the positional aspects of behavior because they are mainly orientated towards safeguarding their standing vis-à-vis their principals. Exactly the opposite can be expected for advocates who concentrate on the functional aspects of optimizing their work. A servant IPA, finally, will not explicitly engage in any of these strategies since they all run counter its reactive stance.

5 Empirical illustration

In the remainder of this article, we illustrate our conceptual and theoretical considerations by short comparative case studies on the bureaucracies of the IMF and the BIS, and the ILO and the FAO, respectively. We conducted expert interviews with regular staff, middle management and senior management working for the IPAs of these four organizations (see the Appendix Table 4 for an overview of the interviewees). A semi-standardized questionnaire was used to investigate the civil servants’ inclination to informally influence the policy process.

The organizations under study display a broad variation on the dependent variable, administrative styles, i.e., they come rather close to the four ideal types identified above (see Table 2). We opted for this case selection since our main objective is to illustrate the conditions underlying each style. As will be shown in the following, the IMF bureaucracy exhibits a style that comes close to the entrepreneurial ideal type. Whereas the BIS secretariat largely corresponds to the advocatory ideal, its counterpart at the FAO comes close to that of a consolidator. The ILO, finally, has a bureaucracy that predominantly pursues a servant style.

To pinpoint the explanatory relevance of internal and external challenges on administrative styles, we study the four IPAs on the basis of a comparative case design. We present two comparative case studies, in which we focus on the external challenges, while the degree of internal challenges remains constant. Whereas both the IMF and BIS bureaucracies face low internal restrictions to their policy capabilities, exactly the opposite applies to the ILO and the FAO. For both constellations we show that the degree of external challenges makes a difference for administrative styles. The IMF’s IPA deviates from an advocatory style towards an entrepreneurial pattern in view of the strong external challenges it faces in comparison to the BIS. In a similar vein, the FAO’s IPA deviates from the servant style, dominant in the ILO’s, in view of much stronger external pressures it is confronted with.

5.1 Administrative styles in the IMF and the BIS: Entrepreneur versus advocate

We illustrate the difference of advocatory (BIS) and entrepreneurial (IMF) administrative styles in a short comparative case study of the IMF and the BIS. Both organizations operate in the realm of global finance. While the IMF is responsible for overseeing and safeguarding the stability of the international monetary and financial system (IMF 2016), the BIS considers itself the world’s ‘bank for central banks’ (BIS 2016) and is mandated to foster international monetary and financial cooperation. To be able to study the influence of its permanent secretariat, we exclusively focus on the so-called ‘BIS proper’, which drafts policy proposals, research and reports. In case of the IMF, we focus on all IPA activities with the exception of the lending programs to ensure comparability.

While the literature on the BIS is still very limited and has an emphasis on historic accounts (cf. Toniolo and Clement 2005; Bernholz 2009), the IMF’s bureaucracy naturally has attracted a lot more scholarly attention in the past (cf. inter alia Woods 2006; Truman 2009; Moschella 2012; Ban and Gallagher 2014; Chwieroth 2010; Broome 2015). These accounts, however, do not specifically address the bureaucratic influence or inner workings of the IMF but usually focus on the whole of the IMF as an IO. Likewise, the neorealist approaches that merely view the IMF as a negotiation arena for powerful member countries without having an agency of its own (Thacker 1999; Stone 2012; Copelovitch 2010) do not help us to understand the administrative style of the organization.

The latter perspectives neglect the fact of IPA’s agency: The IMF’s administration closely resembles the entrepreneurial ideal-type, with focus on both functional and positional aspects of administrative influence.Footnote 2 Whereas the IMF does indeed face considerably few internal challenges and – at the same time – severe external challenges, the absence of perceived external challenges to the BIS renders political aspects of strategic influence unnecessary to the IPA. The BIS secretariat can therefore be considered an advocate with strongly developed administrative routines orientated towards functional aspects of increased policy effectiveness (Ozgercin 2012).

Especially when compared to other IPAs, both the BIS and the IMF secretariats face no particular internal challenges in terms of strong control mechanisms or cognitive limitations, which is mainly due to staff characteristics; both IPAs can rely on a rather homogenous work force with a shared professional background as economists (Nelson 2014). For instance, the IMF secretariat has very high expertise on monetary and financial policies and typically recruits the best economists from the world’s leading research universities, frequently right after graduation. BIS staff is likewise composed of highly trained experts and forms a comparatively homogeneous work force that shares the common language and methodological background in economics. Thereby, the organizations are able to create a rather homogeneous esprit de corps among their staff, which is further reproduced and reified by hierarchies and socialization (Momani 2005; Chwieroth 2010; IMF 1).

Moreover, the IMF’s and the BIS’s IPAs enjoy a comparatively high degree of procedural discretion (Vaubel 2006). This allows the IMF secretariat to produce drafts that oftentimes will not be substantially changed anymore once the member states receive them.

You spend a lot of time adjusting the texts and the feedback of others until finally it goes to the [Executive] Board and then it basically rubber-stamps it most of the time. […] At the IMF, the staff is supposed to produce the perfect product (IMF 1).

Martin (2006) echoes this statement in her description of how knowledge asymmetries work to the advantage of the IPA’s informal agency and influence vis-à-vis the IMF Executive Board. At the BIS, the central banks allow staff and management handle their business in a very autonomous way as well, with few to none interventions or control mechanisms: ‘The BIS annual report [is] not vetted by national authorities. So our report is not seen by national central banks before it comes out’ (BIS 2). Low internal challenges arising from a combination of homogenous staff composition and high procedural discretion hence allows both IPAs to develop a strong functional orientation in their administrative styles.

Similar internal challenges favor similar administrative routines that emphasize a pronounced functional orientation. The latter manifests itself in a similar emphasis on the managerial aspects of administrative strategy in both IPAs. The IMF and the BIS secretariats both play a pro-active role in setting up and identifying new policy issues. The IMF secretariat either tries to set the agenda by developing original policy proposals and research or by reacting to upcoming issues as early as possible to be able to shape the discussion in the very beginning of a policy debate (IMF 1). IMF staff permanently monitor globally relevant events, such as a sudden economic slowdown in a given country or sector or important monetary policy decisions to ensure that no major issue will be overseen (IMF 2; IMF 3). The BIS secretariat is even entirely free to choose the topics of its annual reports and policy papers and does not have to worry about political interventions as ‘the central banks hardly have any impact on agenda-setting’ (BIS 2). The focus on effective policy development becomes even more apparent when considering the optimizing strategies of solution search, which prevail in both IPAs. Although, strictly speaking, neither organization is a think tank or an academic institution, the production of their reports adhere to standards similar to academia (IMF 3; BIS 3). Staffers at the IMF are constantly concerned with finding and promoting what they deem ‘first-best’ solutions (Chwieroth 2013: 274). At the BIS, there is a widespread understanding that ‘at the end of the day, [their] credibility rests on the credibility of our analysis, which rests on being fair and independent. Doing good analytical work is essential for [the BIS]’ (BIS 1). Also, there is a strong notion of encouraging internal debates about policies in order to ensure that all aspects of an issue have been taken into consideration (BIS 6) in a process similar to that of academic peer-reviewing (IMF 1). This procedure requires a notably high degree of positive coordination efforts in order to optimize the organizational output. Within the BIS and IMF bureaucracies, we find an institutionalized, well-managed process of positive coordination in place (IMF 4), which is even more fiercely pursued informally:

The hierarchies are very flat and we are constantly working together across units in a project-based way. […] This […] means that the official areas of responsibilities and work assignments that you are set within the organizational chart of the BIS do not really matter in reality (BIS 3).

In a similar policy-orientated vein, both policy promotion and policy evaluation are considered important tasks for the IMF. It is not enough to just conduct good analyses based on sound economics for staff to be able to have an impact. The IMF secretariat needs to present findings and recommendations in an appealing way to increase their likelihood of acceptance, given that its target audience may consist of non-economists and people who are generally skeptical about the Fund’s work. The BIS secretariat follows a sometimes even aggressive style of policy promotion and does not shy away from open conflict with the principals and the IMF’s arguments to be ‘able to ignite a lively debate or to steer an existing discourse into a certain direction despite [its] smallness and lack of formal powers’ (BIS 5). In the past, this IPA activism has led to significant reversals of official IMF policy by the member states (Chwieroth 2010). Conversely, other than tracking citation metrics of its research output, the BIS staffs do not employ internal evaluations of their work despite recent attempts to improve this situation (BIS 5). Although the internal evaluation mechanism in the IMF did initially face considerable opposition on the part of the staff, it appears to be well accepted by now (Weaver 2010): The IMF places stronger strategic emphasis on the evaluation of their policies, both in terms of rigidity (IMF 5) and a strong intrinsic motivation to see their policy advice being implemented (IMF 6).

Yet, despite similarities in their functional orientation, the two IPAs differ strongly with regard to the external challenges they face, and hence their positional orientation. As opposed to the advocate-style of the BIS bureaucracy, the IMF secretariat complements its policy advocacy with a pronounced orientation towards institutional consolidation. This variation in administrative styles can be understood against the backdrop varying external challenges both IPAs are facing. Contrary to the BIS, the IMF secretariat is under much political pressure from member states and societal groups as well as the media and therefore has to act strategically in order to both safeguard its institutional interests and to promote its policy initiatives. This is mainly due to the fact that monetary and financial policies, such as the Greek sovereign debt crisis, constitute highly salient and oftentimes contentious topics that attract much political attention (IMF 7, IMF 2010). Although the IMF bureaucracy may hence not have direct competitors in its policy domain, the overall high external challenges induce an interest in institutional policy influence. In the BIS case, by contrast, external challenges are much lower than for the IMF, as it lacks a direct competitor in its policy domain and the central banking community, unlike the community of finance ministers, is a lot less politicized (BIS 3).

The different levels of external challenges become apparent when looking at the strategic stances both IPAs take. The BIS bureaucracy focuses on gaining internal support from its principals rather than trying to influence its external environment. The IMF secretariat, by contrast, actively reaches out to a broad range of other stakeholders, including notoriously critical ones, to mobilize support and to actively map the political environment (IMF 2, Chwieroth 2013). For IMF managers, mapping the political space is part of their job (IMF 2). At this point, the BIS bureaucracy follows a different logic. In the drafting phase, their anticipation approach is two-fold: While policy experts do not have to take political considerations into account and aim to produce a draft of high quality, the political clearance is done in the next step by a manager responsible for checking drafts for political feasibility. While these BIS managers are eager to anticipate political constraints, they do it in a way to be able to ‘choose their battles’ strategically, i.e., only to issue politically controversial policy outputs if they whole-heartedly belief that it is the right thing to do and that it can be backed with a sound analysis. But once the BIS management has made the decision to support a given policy, it will follow through with it regardless of its principals’ preferences (BIS 3; BIS 4). The IMF bureaucracy pursues pronounced strategies to anticipate political constraints when drafting policy proposals not just internally vis-à-vis the Board, but also externally when dealing with member states in country surveillance or program work (IMF 3).

During implementation, the IMF bureaucracy does not shy away from making use of formal powers and moves quite strategically with regard to the degree of potential conflict with its member states it is willing to engage in. Civil servants are well aware of the different boundaries that exist when implementing policies in different countries with different veto power: they would not push as aggressively for their own positions with a large, highly-industrialized country compared to a low-income one (IMF 5; IMF 8). Contrary to the IMF, the formal implementation powers of the BIS bureaucracy are quite limited, as it acts as a hub for sharing information and enhancing cooperation among central banks (BIS 4). However, in terms of making use of its informal power, the BIS senior management displays a high degree of ambitiousness to have an impact by sending clear and coherent policy messages to its member states, which can be very critical (BIS 5).

To conclude, the BIS secretariat is able to afford a purely advocatory style because of its homogeneous staff, secured long-term funding through its banking activities and unique position within the international architecture of financial governance, in which it does not have any real direct competitors (BIS 2). In the absence of external challenges to its institutional status, this IPA does not have to defend or enhance its status, and instead has developed administrative routines orientated towards policy effectiveness (functional orientation). In the IMF case, the combination of managerial and political behaviors indicates an entrepreneurial style of the bureaucracy. In addition to its orientation towards policy effectiveness, it permanently seeks to safeguard its position, which is threatened by several external challenges.

5.2 Administrative styles in the ILO and FAO: Servant versus consolidator

The second comparative case study sheds light on the empirical differences in routines and determinants between an IPA that follows the servant ideal-type (ILO) and one that resembles a consolidator (FAO). The ILO has the exclusive mandate on labor regulation among intergovernmental organizations (Hughes and Haworth 2011), while the FAO is dedicated to the ‘eradication of hunger, food insecurity and malnutrition’ (FAO 2016). Both IOs and their IPAs are located in the area of social policy, perceive similarly high internal challenges, and exhibit – as a consequence – only little functional ambitiousness. However, they differ with regard to their emphasis on positional aspects of their work. While the ILO’s IPA does not actively attempt to influence institutional policies (van Deale 2008), the external challenges the FAO faces favor its positional orientation.

Compared to those of the IMF or the BIS, the FAO and the ILO bureaucracies face many more internal challenges. In the FAO case, this is partly due to the highly hierarchical organizational structure established by the committees and the general modus operandi of country requests and FAO response that does not allow for much cognitive slack. Furthermore, many posts are currently frozen leaving an understaffed organization that is not able to ‘attract the best’ (FAO 1) anymore, which leads to a decrease in its expertise. Because the FAO increasingly employs staff with backgrounds other than agriculture and even agricultural policy, it is not united by a common paradigm and it exhibits considerable epistemic heterogeneity. The ILO’s staff composition is equally heterogeneous as its unique tripartite structure is reflected in the diverse professional backgrounds of its employees and the spectrum of staff’s educational background ranges from social scientists, lawyers, and economists to experts on very specific areas, such as forestry. In this vein, Haas’ now classic finding that the ILO was back then unable to really do politics ‘Beyond the Nation-State’ because its constituents were too heterogeneous with regard to both ideological and material interests and preferences (Haas 1964) is still relevant today in light of their contested belief systems. The ILO’s bureaucracy also displays a low degree of cognitive slack. The approval of the ILO’s larger projects and mission is given ex ante (ILO 2016), which diminishes its leeway for independent action. Moreover, the ILO’s organizational architecture provides many avenues for the constituents to influence the bureaucracy’s work directly and especially the heavyweights among the donors exert considerable (informal) influence on the ILO’s policies (ILO 14).

Several internal challenges thus impede a strong functional orientation towards policy effectiveness in both bureaucracies. Issues rarely emerge from inside both IPAs. The FAO secretariat is strongly driven by its member countries’ needs and interests. In a similar fashion, the ILO bureaucracy ‘react[s] to things’ (ILO 4), relying heavily on the tripartite constituents’ input in identifying problems (ILO 2; ILO 5; ILO 10; ILO 13). A similar pattern of dependence and reactivity dominate the IPAs’ solution search strategies. With regard to the evaluation of alternatives, the scientific sophistication of FAO projects strongly depends on the donor type and project size, and follows a rather satisficing logic in order to save time and resources. This is not necessarily due to a lack of motivation, but donor requirements and a lack of time and financial resources (FAO 1). Staffers at ILO and FAO mostly rely on their own expertise and experiences and follow a satisficing ‘what worked in country A, will also work in country B’-method (FAO 6). In principle, ILO formal requirements in that regard are rather ambitious; yet informal routines are mostly satisficing (ILO 6; ILO 7), since optimizing solution search is considered too expensive and effortful (ILO 3).

Satisficing approaches come with lower coordination demands. In both IPAs coordination within the organization is conceived to be problematic and clearly negative. At FAO, coordination is said to be strongly dependent on individual willingness for cooperation (FAO 4; FAO 5) as ‘the problem is not the structure [but] rather behavioral’ (FAO 13). Moreover, many FAO staffers narrated on severe silo thinking between the various departments (FAO 10; FAO 6; FAO 8; FAO 3). In the ILO, there are detailed formal guidelines on how to coordinate but they are not effectively implemented (ILO 9), which oftentimes leads to a ‘ridiculous amount of silo thinking’ (ILO 14). This goes so far that, at times, country offices might not know that a mission is being launched in their country (ILO 12).

Lastly, with regard to policy promotion and evaluation, positional considerations are perceived as more important than functional concerns for policy effectiveness. Neither promotion, nor evaluation receives much attention. The FAO secretariat’s efforts to promote its outputs and products are perceived as mediocre by many staffers (FAO 3; FAO 5) and in the ILO constituents themselves are the major tools for implementation and promotion (ILO 13). Even though the end-of-assignment-reports of former projects are collected in FAO project databases, staffers held that ‘these feedback mechanisms are not very institutionalized or too widely used’ (FAO 5). The same holds true for the ILO case, where systematic evaluation does not exceed what is formally required in day-to-day practice (ILO 7) and informal knowledge management is considered dysfunctional (ILO 10).

While the FAO and ILO secretariats exhibit similar internal challenges and share a rather weak functional orientation in their administrative styles, they differ substantially when it comes to the perception of external challenges and their positional orientation subsequently. Whereas the FAO bureaucracy explicitly seeks to influence its institutional standing, its ILO counterpart perceives no need for such behavior. The behavioral patterns of the FAO bureaucracy can be traced back to the pronounced external challenges it sees itself subject to. As a Global Environment Fund (GEF) implementing agency it finds itself in permanent competition with other GEF implementing agencies. Because agriculture can be of a highly political character, particularly in developing countries, FAO has to maneuver its course with utmost caution. Under these circumstances, FAO’s staffers consider it necessary to fight for its resources whilst navigating a difficult institutional field.

Conversely, the ‘world of work’ and the ILO are located comparatively low on the global agenda and the IPA is very aware of this. Especially for major donors, labor rights are far from being a ‘hot button’ issue. The ILO’s principals are very heterogeneous as its membership is almost global and the tripartite structure adds further complexity. In the case of the ILO, this situation is seen as an obstacle for independent influence. Because the principals involvement in the policy making process is so remarkably strong and every document needs compromise, the ILO’s bureaucracy too has to be very careful to navigate this structure. Like many UN specialized agencies, the ILO bureaucracy is increasingly facing competition, mostly with regard to funding, albeit to a lesser extent than other UN specialized agencies (ILO 12; ILO 8; ILO 6). However, the ILO thinks itself safe (ILO 11; ILO 4; ILO 14) as the administration’s subjective perception of those threats is strongly mitigated by their self-perception as a ‘separate animal’ (ILO 14).

The differing levels of perceived external challenges constrain and enable the orientation of the two IPAs. ILO’s and FAO’s different orientations empirically manifest themselves with regard to the employment of positional activities. At FAO, significant attention is paid to the mapping of the political space. Given its operational mandate, the FAO bureaucracy has to strike a delicate balance between the often seemingly irreconcilable interests of receiving and donor countries. FAO managers do so by strategically framing and communicating policy initiatives to their stakeholders. As one officer strikingly put it: ‘Words are incredibly important. Sometimes new words just hide the old approach. And in this we are very [good]’ (FAO 1). Owing to excellent informal contacts with members of central decision-making committees (FAO 10; FAO 9), the FAO bureaucracy is well informed about the likely preferences of their member states. This strategy is less pronounced at the ILO. The bureaucracy does have an idea about what might or might not resonate with their constituents (ILO 10). However, these efforts are ‘rather decentered’ (ILO 2) and do not follow any pronounced strategy. More accurately ‘mapping’ – if happening at all – is a by-product of meeting by chance or random talks at the side of formal events (ILO 3; ILO 4).

By contrast, political anticipation constitutes a crucial routine for both the ILO and FAO bureaucracies. In case of the FAO, accommodating member states’ interests is a decisively intentional move as it is deemed essential to generate commitment and continuing demand (FAO 10; FAO 11). For the ILO, pronounced strategies of functional (de-)politicization are absent. This is because anticipation in the ILO is no strategic move but a requirement (Baccaro and Mele 2012: 195). The IPA is forced to detect political ‘red lines’ because of the high consensus demands and complex governance structure of the ILO (ILO 13; ILO 6).

Support mobilization is essential for the FAO bureaucracy. Its country representatives enjoy very good relations with national governments, and local actors such as NGOs, producer organizations, or academia. This type of support allows them to conceptualize project ideas that convey commitment and the impression of an enabling environment in the receiving country. By contrast, the ILO secretariat rarely builds strategic coalitions with civil society or else to promote some sort of issue to its constituents. Since, due to its tripartism, the ILO already includes a relevant share of civil society and – more importantly – the very part it deems most relevant (ILO 13; ILO 7), additional coalitions are almost obsolete to them.

Finally, both ILO and FAO bureaucratic routines reflect a moderate attitude towards the strategic use of their implementation powers. In the FAO, administrative routines prevail that are merely directed at maintaining positive work relations with all member-states instead of fervently pushing for effective implementation (FAO 7). At the ILO, staffers emphasize persuasion instead of putting pressure on governments because ‘this is the one thing [ILO] can do’ (ILO 4, Hartlapp 2007). The ILO’s formal implementation power is even called a ‘paper exercise [that] doesn’t really have any practical implication.’ (ILO 6; see also Douglas et al. 2004).

The ILO bureaucracy is confronted with many internal challenges but rather few external challenges, especially when considering their subjective perception of threats. In this constellation, the IPA does not need to strengthen its position, nor is it sufficiently equipped to functionally enhance the effectiveness and quality of their output beyond what is expected of them. In line with its servant style, it thus pursues no pronounced strategies of policy or institutional influence. Cox’s finding that the ILO’s outputs are based on a consensus that is the ‘lowest common denominator of neutral technical services’ is thus unabatedly modern as it poignantly describes the ILO’s role as its constituents’ ‘servant’ until today (Cox 1973). The FAO bureaucracy, by contrast, perceives not only high internal, but also considerable external challenges. As a consequence, administrative routines in the FAO are strongly orientated towards institutional consolidation, rendering its IPA an almost perfect example of a consolidator style.

6 Conclusion

This article started with the observation that the growing involvement of international bureaucracies in policymaking stands at odds with a lack of systematic knowledge of the role and influence of these administrative bodies. To identify differences in bureaucratic routines and bureaucratic agency, we developed the concept of administrative styles and analytically distinguished between four ideal-typical patterns: servant, advocate, consolidator, and entrepreneur. In addition to this conceptual innovation, we offered a theoretical framework that accounts for variation in administrative styles across different organizations. The analytical use of the concept of administrative styles and the explanatory relevance of our theoretical argument was illustrated in two comparative case studies covering four IPAs, namely those of the IMF, the BIS, the FAO, and the ILO. We have not only presented first empirical evidence that administrative styles and orientations indeed vary considerably across IPAs, but also that the kind of entrepreneurship is strongly affected by the internal and external challenges to the IPA in question.

By introducing this novel concept to the fields of IO and IPA research, we intend to balance the current trade-off between case sensitive knowledge and comparability when looking at IPAs' influence. Instead of merely examining IPAs’ formal and structural characteristics or singular instances of successful bureaucratic influence, we scrutinize their day-to-day workings and routines, guided by a set of indicators ensuring comparability. We look at bureaucratic agency and influence from a bottom-up perspective. Rather than asking what they ought to be doing and extrapolate their influence, we observe what they are actually doing and assess their agency from that. Thereby, we capture an obvious source of bureaucratic influence: the agency IPAs possess in their most basic functions and activities across the policy cycle.

While the concept of administrative styles as developed in this article offers a range of analytical strengths, we are well aware of the fact that our empirical illustration only considers a first plausibility probe of the soundness of our analytical considerations. We consider it a promising avenue for future research to investigate change and variation of administrative styles as well as their conditions in a more systematic manner. Moreover, based on such an endeavor, further studies could shed light on the linkage between IPA styles and the performance of their IOs.