public administration chapter assignments
O R I G I N A L A R T I C L E
Explaining bureaucratic power in intergovernmental relations: A network approach
Yvonne Hegele
Leibniz Center for Science and Society, Leibniz
University, Hannover, Germany
Correspondence
Yvonne Hegele, Leibniz Center for Science and
Society, Leibniz University Hannover, Lange
Laube 32, 30159 Hannover, Germany.
Email: yvonne.hegele@lcss.uni-hannover.de
The core assumption of the bureaucratic politics model and a large
part of public administration scholarship is that bureaucrats influ-
ence politicians and political decisions via their crucial role in pre-
paring, coordinating and formulating policy. While this influence
has been analysed in a vertical direction, that is, how much do
bureaucrats influence politicians, the horizontal perspective has
been mostly neglected: which bureaucrats are most powerful and
influential during the process of bureaucratic coordination and
decision-making? Deducing hypotheses from bargaining theory and
testing them with a novel network dataset on German Intergovern-
mental Relations (IGR), this contribution finds that bureaucrats
indeed possess varying degrees of power. Jurisdictional and organi-
zational power resources, such as voting, financial and institutional
power, and also party politics, can best explain these variances in
bureaucratic power. Personal characteristics, such as experience
and education, however, are not used as power resources.
1 | INTRODUCTION
The preparation of policy decisions is one of the core tasks of ministerial bureaucracies. Bureaucrats are in charge of
choosing, formulating and coordinating public policy as well as negotiating with actors within and outside the
politico-administrative system. They prepare policy for the head of department and government or take decisions
(e.g., Peters and Pierre 2016). By doing so, bureaucrats carry out political tasks and can potentially influence political
decisions. This influence can be exerted indirectly due to the role, organization and responsibilities of public officials
as administrative working units in government (Mayntz and Scharpf 1975). In many instances, bureaucrats also take
decisions without the involvement of politicians, which is the most direct form of influence on the political process
(Page 2012). The strength of the bureaucratic influence thereby varies according to the stage of the policy process
and the institutional structure of government (Schnapp 2004). Furthermore, the influence and power of bureaucrats
can vary with the structural and procedural role a certain public administration takes on during the process of policy-
making (Hartlapp et al. 2013). Thus, some bureaucratic organizations or even individual bureaucrats can be more
influential and powerful than others. Yet little is known about which bureaucrats are more powerful in influencing
the political process.
Building on these key insights into bureaucratic influence and power, this contribution aims at developing a
framework to explain differences in influence and power between ministerial bureaucrats. Based on the propositions
DOI: 10.1111/padm.12537
Public Administration. 2018;96:753–768. wileyonlinelibrary.com/journal/padm © 2018 John Wiley & Sons Ltd 753
of the bureaucratic politics model, it is argued that bureaucrats from different bureaucratic organizations pursue
varying preferences. To pursue these preferences, they have varying power resources at their disposal. Based on the
bargaining power framework (Bailer 2010), these power resources will be explored.
Measuring power and determining power resources is difficult. Most approaches compare initial positions with
final decisions to estimate how much of the actors’ interests prevailed. Such an approach is problematic for several
reasons, for example strategic signalling (Coddington 1968; Snyder and Diesing 1977), but it is especially problematic
when studying the power of bureaucrats. While politicians’ initial positions are often public, bureaucrats’ are not
because they usually are not public figures. To circumvent these problems, this contribution proposes a new way of
measuring bureaucratic power by using social network analysis. Bureaucrats and their coordination and negotiation
relations can be conceptualized as networks. Those actors who are most central in the coordination and negotiation
process are then assumed to be more powerful during the decision-making process. This conceptualization enables
estimating which factors impact on the power and centrality of bureaucrats.
The question of power and influence of bureaucrats is most pronounced in situations which require intensive
coordination and negotiation and in which a high number of bureaucrats are involved. IGR, meaning processes of
joint decision-making among a number of government executives in multi-level states, represent such an occasion.
For this reason, the power of bureaucratic actors in IGR will be analysed in this contribution. Establishing a frame-
work of analysis and testing it with the most-likely case of Germany is an important step towards a broader and com-
parative analysis of the power of various bureaucrats in decision-making processes.
2 | BUREAUCRATIC POLITICS IN HORIZONTAL INTERGOVERNMENTAL COORDINATION
The bureaucratic politics model (Allison and Halperin 1972; Allison and Zelikow 1999) argues that government deci-
sions can only be properly understood if they are conceptualized as a result of the aggregate of individual decisions
and actions by several actors within this government. A core assumption of the model is that bureaucrats develop
different preferences, objectives and goals, which stem from ‘various conceptions of national … organizational,
domestic and personal interest’ (Allison and Halperin 1972, p. 43). At heart, these various conceptions originate from
the high levels of delegation and specialization which are typical for modern governments (Bouckaert et al. 2010).
These bureaucrats with their varying preferences try to influence politicians in their decision-making. The model pays
attention to the preferences of bureaucrats and the mechanisms through which these are aggregated into a
government decision (Hartlapp et al. 2013, p. 427).
The present contribution focuses not primarily on the interests of bureaucrats per se, but on their power to influ-
ence the decision-making processes. The fact that they possess varying interests thereby makes it necessary for
them to use their power to influence the decision-making process. Power as a concept is defined in the bureaucratic
politics model as ‘effective influence on government decisions and actions’ (Allison and Zelikow 1999, p. 300). Power
thus is the ability of an actor to direct the decision-making process in the desired direction using the available means
to achieve the actor’s preferred outcome (Schneider and Bailer 2002, p. 52). These means are called power resources
in what follows. Bureaucrats possess unevenly distributed power resources which they can use to influence the gov-
ernment decision (Allison and Halperin 1972). These can be resources in the strictest sense, such as financial means,
implicit resources such as veto threats, but also personal characteristics such as experience. Yet, little systematic
knowledge exists about what these resources are and how they are distributed among bureaucrats. To advance
knowledge on this topic, the present contribution develops propositions about power resources which bureaucrats
can make use of when trying to influence decision-making processes. While there are certainly further actors, such
as parties and interests groups, which try to influence political decisions, this analysis concentrates on the process of
bureaucratic decision-making. The power of the bureaucratic actors is analysed in relation to each other and not to
other sources of influence.
754 HEGELE
While classical accounts of bureaucratic politics focus on the political–administrative dichotomy asking how
bureaucrats can influence politicians (e.g., Hood and Lodge 2006; Peters et al. 2016), a similar reasoning holds in a
horizontal perspective, namely among different bureaucrats who find themselves in joint decision-making situations.
Government decisions are formulated and taken in a certain way because a certain department has more power and
influence during the decision-making process than another department. Hence, the question arises why some
bureaucrats or departments are more powerful than others and can better influence decisions. Which power
resources make bureaucrats more influential?
Governments usually consist of a small number of departments which are jointly responsible for the executive
functions of government, among them policy formulation and executive decision-making. Studying bureaucratic
power resources within such a setting is associated with some problems of small-N analysis. In contrast, a situation in
which a higher number of departments and bureaucrats are involved in coordination processes would allow quantita-
tive statistical analysis of power and power resources. IGR in multi-level states represent such a setting. Govern-
ments in multi-level states are interdependent in many instances of policy-making and engage in IGR to deal with
joint problems, create joint solutions or simply exchange best practices (Poirier and Saunders 2015, p. 1). For this rea-
son, executives often exchange information, coordinate joint policy, and also try to protect their autonomy from
encroachment by others and try to influence each other (Behnke and Mueller 2017). Thus, IGR occur whenever two
or more governments in federal states must or want to engage in joint actions and take joint decisions. This is, for
example, the case in processes of reforming the federal state where a joint decision needs to be taken, when sub-
states in federal states want or need to cooperate in implementing federal or European law, or when they install or
finance joint organizations. IGR are omnipresent coordination relations between executives in multi-level states.
In most scholarly analysis, IGR among politicians are considered, but these are usually supported by relations
among public administrations (e.g., Johns et al. 2007; Parry 2012; Hegele 2018). Public officials prepare the political
IGR meetings, and just as in any other political process, exchange information, prepare and even negotiate decisions
or take decisions on their own. Thus, due to the high number of government organizations involved in IGR (core
executives and departments from each government), this offers an ideal setting in which to study the power
resources of public administrations.
The horizontal process of decision preparation and decision-making among departments and government actors
is often analysed under the topic of coordination. While coordination, defined as an output or ‘an end-state in which
the policies and programmes of government are characterized by minimal redundancy, incoherence and lacunae’
(Peters 1998, p. 296), might be achieved to a varying degree by governments and their departments, coordination
defined as a process is one of the key actions in which governments and their departments regularly engage. Coordi-
nation as a process emphasizes the ‘strategies and instruments governments use to coordinate organizations or pro-
grams within the public sector’ (Bouckaert et al. 2010, p. 16). This involves analysing ‘the development of ideas about
joint and holistic working, joint information systems, dialogue between agencies, processes of planning, and making
decisions’ (6 2004, p. 106). Coordination is thus conceptualized as the ‘intervening stage of debate and deliberation
[during which] persuasion, reconsideration, conceivably even coercion takes place’ (Shepsle and Boncheck 1997,
p. 44) before a joint decision is taken. Simplifying Metcalfe’s (1994) nine-step coordination scale, the process of coor-
dination is conceptualized as the exchange of information, the search for allies, and the negotiation of compromise or
decisions. First, actors exchange information and viewpoints in order to reduce uncertainty about the interests and
goals of other actors. Second, they negotiate by trying to form alliances with actors with similar interests and goals
or find a compromise with actors which have different interests or goals. The power of bureaucrats and their power
resources are likely to differ between these two processes. Some actors might be more powerful in the information
exchange network; others when it comes to negotiations. Based on these insights, two coordination networks among
bureaucrats will be conceptualized in this contribution; one network of information exchange and one network of
negotiation.
HEGELE 755
3 | BUREAUCRATIC POWER RESOURCES
What are the power resources of bureaucrats that they can use to exert influence on the decision-making process?
The bargaining literature in political science has put forward a number of theoretical arguments for resources which
make actors more powerful. The most prominent and established power resources of actors are voting power, eco-
nomic power, institutional power and domestic constraints. Furthermore, initial evidence is found for party politics
and personal experience as power resources. While these theoretical arguments have usually been tested on the
political level, they possess a broader scope of validity. Similar power resources are available to bureaucrats when
preparing and negotiating political decisions. Bureaucrats in joint decision situations can make use of jurisdictional or
organizational as well as individual power resources. In IGR, one bureaucrat usually represents one department. Thus,
each bureaucrat represents and speaks for this department and at the same time can make use of the department’s
power resources. Due to the structural similarity between the EU’s Council of Ministers negotiation and IGR (Scharpf
1985; Poirier et al. 2015), this article applies the theoretical assumptions developed by Bailer (2010) and Tallberg
(2008) for the EU Council of Ministers to IGR in multi-level states in general and bureaucratic decision-making pro-
cesses in particular. The present contribution will thus be the first to test all these possible power resources under
one combined framework and with a coherent empirical dataset.
Voting power as a power resource assumes that the more votes an actor has in a joint decision situation, the
more powerful this actor is because his or her voting choices carry more weight than those of actors with fewer
votes (Tallberg 2008, p. 694; Bailer 2010, p. 745). This argument of course only holds for majority-voting situations
with weighted votes. Bureaucrats can similarly use voting power as a power resource because in bureaucratic
decision-making processes, the bureaucrats represent their jurisdiction and thus the number of votes associated.
H1: The higher the number of votes a bureaucrat represents in a majority-voting situation with
weighted votes, the more powerful the actor.
Economic or financial power can be a power resource because joint decisions often have financial consequences
which need to be distributed among the participating actors. Hence, the support by actors with more financial
resources is necessary in order to make the decision meaningful and their implementation realistic. Similarly to voting
power, the bureaucrats represent their jurisdiction in decision situations and thus also the economic or financial
power of the jurisdiction.
H2a: Bureaucrats representing richer jurisdictions are more powerful than those from poorer
jurisdictions.
On an organizational level, the ministries of finance can be assumed to be more powerful than the other departments
within the jurisdiction because they play an important role in allocating resources among ministries and need to agree
on any additional investments and long-term expenditures (Heller 2015). Thus, it can be assumed that bureaucrats
representing the finance ministries also represent this kind of financial power in joint decision situations.
H2b: Bureaucrats representing the departments of finance are more powerful than bureaucrats from
other departments.
Institutional power is ‘the ability to exit, veto and set intuitional agendas’ (Bailer 2010, p. 746). Veto and exit rights
bring the threat of leaving the decision processes and thus have a higher probability of creating concessions from the
other actors. These resources are, however, distributed equally among the actors and only used rarely (for EU Council
negotiations see Tallberg 2008, p. 694). Agenda setting, on the other hand, is likely to be a relevant power resource.
Agenda setters usually have an information advantage and control the procedures through their ‘responsibility to
manage the agenda, broker agreements and represent the decision body vis-à-vis third parties’ (Tallberg 2008,
756 HEGELE
p. 696). This is assumed to be an important power resource for bureaucrats because, as experts with detailed knowl-
edge on the subject, they can act as agenda setters and thus draw on this power resource.
H3a: Bureaucrats representing the agenda setting department or jurisdiction are more powerful than
bureaucrats from the other departments or jurisdictions.
In addition, on the organizational level, the core executives have agenda setting power within their jurisdiction and
thus can be expected to be more powerful (Peters et al. 2003; Dahlström et al. 2011). Thus, bureaucrats representing
the core executives in joint decision situations can use their formal position as power resource.
H3b: Bureaucrats representing the head of government are more powerful than bureaucrats from the
functional departments.
Reflecting on agenda setting and veto power in IGR, the role of the federal government comes to mind. In most fed-
eral states, the federal government has a prominent position. It can put issues on the IGR or public agenda and very
often its agreement or support is needed for a decision to be ratified or implemented successfully. Bureaucrats repre-
senting the federal level could use this as a power resource.
H3c: Bureaucrats representing the federal government are more powerful than bureaucrats from the
sub-state governments.
Domestic constraint is a power resource that is typical for two-level games, which IGR are as well. If actors are con-
strained in their leeway of action at the second (in this case intergovernmental) level by actors at the first (in this case
government) level, this makes them more powerful because the other intergovernmental actors need to make bigger
concessions to get their support or agreement (Putnam 1988; Bailer 2010, p. 747). In bureaucratic IGR, the win-set
of an actor can be limited by the other bureaucrats within its jurisdiction or, in cases of coalition governments, by the
coalition partner.
Several departments of each jurisdiction are involved in coordinating and negotiating IGR decisions. These
departments can choose either to pursue their interests individually or to coordinate among each other. If they
choose to coordinate internally, it is very likely that the result will be an interdepartmental or jurisdictional position.
According to the domestic constraint argument, this strategy gives them more power in IGR because they can credi-
bly argue that a deviation from this position is not possible and their negotiation partner would need to make conces-
sions in order to find an agreement or compromise.
H4a: The bureaucrats representing a jurisdiction where the departments are closely coordinated are
more powerful than bureaucrats from less coordinated jurisdictions.
In coalition governments, the coalition partners could mutually decrease each other’s leeway of action on the inter-
governmental level. Assuming that IGR decisions in the end need the agreement of both coalition partners, this
mutual decrease of leeway is the more likely, the higher the ideological distance between the coalition partners. Coa-
lition partners with a high ideological distance have a lower overlap of win-sets and thus a less flexible common posi-
tion, which they can use to gain concessions from the other IGR actors.
H4b: Bureaucrats representing a jurisdiction where the coalition partners have a high ideological dis-
tance are more powerful than bureaucrats from less ideologically distant coalition governments.
IGR in federal states are the ideal unit of analysis to study party politics on an equal footing with other power
resources. In federal states, clearly defined party families and at least some state-wide parties exist and are involved
in IGR (Bolleyer et al. 2014) while in the study of EU and international relations, party families are more vague and
parties and elections are still based on the national state level (Bailer 2010). The affiliation of the political head of the
HEGELE 757
government to a party family affects the bureaucratic actors in IGR in so far as the bureaucrat represents the depart-
ment and its political head in IGR. Thus, through various forms of politicization, most importantly functional politiciza-
tion (Schwanke and Ebinger 2006), party political considerations play a role in bureaucratic decision-making processes
as well. Assuming that actors whose political superiors are affiliated to the same party family share some ideological
viewpoint, the number of party peers involved in the negotiation process could be used as a power resource.
H5a: Bureaucrats representing coalition partners with higher numbers of political party peers are
more powerful than bureaucrats with lower numbers of peers.
Furthermore, congruence of government composition is important in IGR (Bolleyer et al. 2014; León 2017). Due to
the prominent role of the federal government in many IGR processes, party congruence with the federal level could
be a power resource.
H5b: Bureaucrats representing a government that is congruent with the federal government are more
powerful than bureaucrats from incongruent governments.
Moving now to the power resources of the individual actors, experience can be expected to matter (Tallberg 2008,
pp. 698 f.; Bailer 2010, p. 746). The longer an actor is involved in IGR, the more experience (s)he has. Experienced
actors can be an important source of information because they were involved in prior processes and might know more
about the subject than is written down in any document. Furthermore, they also know the other actors and possibly
their negotiation behaviour. In addition, with growing experience, actors might have developed negotiation strategies
that prove more successful. Experience as a power resource might be even more important for bureaucrats than for
politicians because they usually remain longer in their position and thus can acquire more experience over the years.
H6a: The more years of experience bureaucrats have, the more powerful they are.
Related to this, educational background and training could have an effect on how bureaucrats behave and which
strategies they choose in decision-making processes (Tallberg 2008, p. 701; Bailer 2010, p. 746). Public administra-
tion research has shown that among the demographic variables, educational background has the strongest impact on
a bureaucrat’s attitudes and behaviour. According to Christensen and Lægreid (2009), a legal education leads individ-
uals to be more rule driven and to pay attention to legal requirements while bureaucrats with a political or social sci-
ence education are more sensitive to politics and political signals and work more on coordination tasks. Having
power over the decision by directing the decision-making process in a favourable direction needs political sensitivity
and skill. Thus, it can be assumed that bureaucrats with a political or social science education are more powerful than
those with a legal education.
H6b: Bureaucrats with a political or social science education are more powerful than bureaucrats with
a legal education.
4 | RESEARCH DESIGN, DATA AND OPERATIONALIZATION
The research question of which actors are more powerful than others when conducting IGR will be answered using a
y-centred, quantitative research design, testing the different power resources within a multiple regression model.
4.1 | Case selection and description
In order to test new theories or new applications of a theory, most-likely cases are the appropriate choice because
the theoretically expected phenomenon is most likely to be present there. If it is not, the theory is unlikely to apply
758 HEGELE
to other, less likely cases (Rohlfing 2012). The German case is a most-likely case for the analysis of power and power
resources in bureaucratic IGR because the German ministerial bureaucracy is regarded as one of the most powerful
in comparative perspective (Schnapp 2004) and it has one of the most elaborate and stable systems of
bureaucratic IGR.
The German ministerial bureaucracy plays a crucial role in policy formulation and decision-making. Ministerial
bureaucrats develop and formulate policy initiatives and coordinate and negotiate proposals with other actors inside
the bureaucracy with the aim of including all relevant aspects and interests at an early stage in order to avoid conflict
within government at a later stage (Mayntz and Scharpf 1975, pp. 67–76; Page 2012). When it comes to IGR, the
ministerial bureaucracy plays a similar role in developing, formulating and coordinating policy with their counterparts
from the other sub-states and the federal state. In this respect, Wagener (1979) described the ‘brotherhood of
experts’ (Fachbruderschaften), a vast network of IGR among bureaucrats from the federal state and the sub-states
working within one sector.
The German system of IGR is one of the most institutionalized and stable systems consisting of the Bundesrat
and 18 ministerial conferences (Auel 2014; Lhotta and van Blumenthal 2015; Hegele and Behnke 2017). The ministe-
rial conferences are voluntary meetings of the sub-state governments with the aim of exchanging information and
coordinating among each other and with the federal government. The Bundesrat is a constitutional organ, the second
chamber through which the sub-states’ governments directly participate in federal legislation and represent the inter-
ests of their sub-state in federal decision-making. The Bundesrat is more institutionalized and meets more regularly
than the ministerial conferences. Furthermore, in the Bundesrat, all actors are engaged in inter-sectoral coordination,
while the ministerial conferences are organized sectorally. Thus, the Bundesrat is chosen as the case for analysis here.
The Bundesrat as a whole holds important veto rights in federal legislation. When the administrative power of
the German sub-states (Länder) and their finances are affected as well as for constitutional changes, the consent of
the (qualified) majority of the votes is mandatory for ratification (Article 52, 77, Basic Law). The Bundesrat holds an
absolute veto in these cases, which constitute up to 40 per cent of federal legislation. In all other cases, it can call a
mediation committee and has at least a suspensive veto that can severely delay the process but be overruled with a
Bundestag majority (Article 77, Basic Law). Additionally, the Bundesrat itself can initiate federal law (Article 76 (1), Basic
Law).
In order to prepare the Bundesrat sessions and decisions, a vast bureaucratic apparatus has been set up. Within
each department of the federal state and each sub-state, a section or staff unit is responsible for the preparation of
Bundesrat matters when the department is involved. In addition, there exists a section in each government chancel-
lery at both levels of government (Bundeskanzleramt, Staats- und Senatskanzleien). This section is responsible for the
coordination of the department sections in order to come to a joint Land position. To coordinate Bundesrat matters
with the other Länder and the federal state, a special division of the government chancellery called Land representa-
tion (Landesvertretung) is stationed directly in the capital Berlin (Schrenk 2010). These ministerial bureaucrats are not
usually politicians or political appointees but career civil servants who remain in this coordination position even when
the party composition of the government changes. These bureaucrats meet in a recurring three-week sequence in
order to prepare Bundesrat decisions (Schrenk 2010).
4.2 | The network dataset
To capture the bureaucratic coordination process surrounding the Bundesrat, a network dataset was collected by the
author in a standardized online survey among the Länder government actors in Germany from August to November
2015. The survey was sent out to bureaucrats responsible for multi-level or Bundesrat coordination in all 171 ministe-
rial bureaucracies (i.e., government chancelleries, Land representations and sectoral departments) of all the Länder.
The respondents were shown a list of possible contact partners by position, and were asked, ‘Please indicate, with
whom of the following actors you have contact during the preparation of the Bundesrat.’ Thus, the individual contacts
and coordination behaviour was surveyed. In a second step, they were presented with a table of the chosen actors
HEGELE 759
and asked: ‘You have indicated that you have contact with the following actors. With whom of these actors do you
pre-negotiate final decisions?’ Thus, data for two networks were collected: a network of information exchange and a
network of negotiation. The relations reported can be interpreted as a mean value of typical information exchange
and negotiation contacts over the last year. If respondents have not indicated a contact with another actor, this
either means that no contact takes place or that it is not relevant to the actor. All 16 German sub-states participated
in the survey with a response rate of 65 per cent (112 out of 171), without any systematic missing data (for descrip-
tive statistics of the networks see Table 1).
4.3 | Measuring power in social networks
The dependent variable, power in bureaucratic IGR, is operationalized using the centrality concept of social network
analysis. Social network analysis is an appropriate tool to research power because power as such is relational
(McClurg and Young 2011, p. 39). Power, defined as the ability of an actor to direct the decision-making process in
the desired direction, can be achieved by communicating with other actors during the process. Network data of com-
munication relations are genuine process data which are able to overcome some of the well-known problems of the
analysis of negotiations and power, namely the ‘difficulties of conducting research on a … body that convenes behind
closed doors, whose proceedings are undocumented and whose participants are unusually hard to gain access to’
(Tallberg 2008, p. 686). In empirical studies, these problems are usually addressed by approximating the power of
actors by comparing initial positions of actors and final outcomes of negotiations. This approach, however, has
attracted considerable criticism for various reasons (see Bailer 2010). Using network data on the bureaucratic
decision-making process can help overcome these problems by providing information on the actual conduct of
decision-making processes.
Centrality is repeatedly used as operationalization and measurement for concepts such as ‘power’, ‘(structural)
importance’, ‘strategic significance’ and ‘importance of prominence’ (Hanneman and Riddle 2005; Henning 2009;
Wasserman and Faust 2009, p. 169; Scott 2013, p. 83). Centrality theoretically indicates that actors have a pro-
nounced position in the network. Among the several options which exist to measure centrality in networks (Freeman
1978), indegree centrality is chosen in this analysis. It measures centrality based on the incoming communication
relations of an actor. With the data used here, an actor has an incoming relation if another actor named him as
communication partner.
This measure is chosen due to theoretical considerations as well as consideration of the data-generating process.
Theoretically, a high number of incoming communication ties indicates that the other actors perceive contact with
the actor in question as relevant in the decision-making process, at least relevant enough to report it. If many actors
report that they are in contact with the actor, it can be reasonably assumed that this actor is powerful. In the network
of information exchange, a high indegree centrality means that this actor is powerful because (s)he is in possession of
more information from different sources and can also control the information flow by providing or retaining this or
that piece of information. In the negotiation network, high indegree centrality indicates power because actors with a
high number of negotiation relations can influence several processes of finding agreement and compromise. In addi-
tion, they usually have various options for compromise from which they can choose the one most profitable
for them.
Furthermore, from the viewpoint of the data-generating process, indegree centrality is less prone to manipula-
tion by the actors. First, while the reporting of outdegree centrality can be manipulated by an actor in an attempt to
TABLE 1 Descriptive statistics of the networks
Network No. actors
No. relations
Edgewise reciprocity
Min. indegree centrality
Max. indegree centrality
Mean indegree centrality
Information exchange network 186 2,414 37% 1 28 13
Negotiation network 183 1,329 26% 1 18 7
760 HEGELE
make themselves look more powerful, this is not possible for indegree centrality because it relies on the information
provided by others. Second, the response rate of the survey is 65 per cent, meaning that 35 per cent of the actors
have an outdegree value of zero because they did not report any contacts but could be named by the other actors as
contacts. Other measures such as betweenness centrality, which measures the potential to control the relations
between other actors, or closeness centrality, which measures the distance of one actor from all other actors, capture
theoretically distinct concepts (Freeman 1978). Further, their reliance on incoming as well as outgoing ties would lead
to a biased measurement due to non-respondents with zero values, which indegree does not.
4.4 | Operationalizing power resources
The independent variables need to be specified for the application in the context of German bureaucratic IGR (see
Table 2; Table 3 provides the descriptive statistics for the variables). Voting power as a power resource might be rele-
vant in the German Bundesrat because decisions are taken by weighted majority-vote. The Länder in the Bundesrat
have three to six votes, depending on the number of inhabitants. In terms of financial power, the German federal fis-
cal equalization scheme is oriented towards solidarity (Behnke 2013) and thus levels differences in the financial
endowment by redistributing resources among sub-states. This system, however, creates a group of net-contributors
with higher financial resources and net-beneficiaries with lower financial resources, the former being potentially
more powerful.
Regarding institutional power resources, due to the majority-voting rule, no single actor possesses the right to
individually veto Bundesrat decisions. Similarly, there are no exit-options and because the German Länder
governments are automatically members of the Bundesrat, abstentions are counted as ‘no’ vote. The ability to set the
TABLE 2 Independent variables and their operationalization
Coordination resource Conceptualization Operationalization Data source
Voting power Number of votes in the Bundesrat
Number of votes as numerical variable
Bundesrat homepage
Financial power Net-contributor
Finance ministries
Dummy, 1 for net-contributor
Dummy, 1 for finance ministry
German Federal Ministry of Finance, financial equalization data from 2015
as indicated in survey
Institutional power Bundesrat presidency
Government chancellery
Federal level
Dummy, 1 for presidency (plenary and committees)
Dummy, 1 for government chancellery
Dummy, 1 for federal level
Printed matter of the Bundesrat 452/14
As indicated in survey
As indicated in survey
Experience Experience in multi-level coordination (in years)
Education
Numerical variable
Dummies: law, politics or public administration, others (baseline)
As indicated in survey
As indicated in survey
Domestic constraint Density of intra-governmental network
Ideological distance of parties in government
Numerical, standardized density of the intra-governmental network
Numerical variable, distance between parties
Calculated based on survey data
Bräuninger and Debus (2011), Appendix B, left–right scores, absolute difference
Party politics Party congruence with federal level
No. of party peers Party of the Minister
Dummy, 1 for full congruence
Numerical Dummies: SPD, Greens, Left,
CDU/ CSU (baseline)
Homepages Bundesrat, federal and sub-state governments
HEGELE 761
agenda, however, might be important in Bundesrat negotiations. The presidency of the Bundesrat rotates among the
Länder, on a yearly basis, from the most to the least populated Land. The Land that holds the presidency is responsible
for organizing and preparing the plenary sessions, dealing with the committees, and contact with the federal govern-
ment (Reuter 2007, p. 320). In addition, each Bundesrat committee is also presided over by one of the Länder (printed
matter of the Bundesrat 452/14). Further, the bureaucrat representing the head of government of each Land can be
assumed to have agenda setting power within the Länder based on the prime ministers’ competence for determining
the general policy guidelines (Richtlinienkompetenz). Furthermore, the federal government might be more powerful in
IGR in the Bundesrat even though it does not have any formal voting rights. During the legislative process, the Ger-
man Bundestag in many instances does not need the consent of the Bundesrat for ratification and even if it does, in
some cases it can overrule a rejection (Article 77, Basic Law).
Domestic constraint as power resource will be operationalized in two ways. First, the density of the intragovern-
mental network in information exchange is used to measure the amount of exchange and coordination between the
departments of one jurisdiction. A high density indicates a higher amount of internal coordination. Due to the bias in
the density measure resulting from varying non-response rates among the Länder, density values are standardized by
the percentage of participating actors, assuming that the non-respondents would have, in the mean, reported the
same amount of contacts as the respondents in each Land. Second, the distance between the coalition partners is
measured based on the dataset on parties’ policy positions compiled by Bräuninger and Debus (2011) using the abso-
lute left–right difference among the sub-state parties in government in autumn 2015. Since the distance is quite sta-
ble during the period investigated by the author, it seems legitimate to use the late-2000s data to approximate the
distance in 2015.
In the German Bundesrat, party politics matter (Lehmbruch 2000; Leunig and Träger 2014; Hegele 2018). Con-
gruence with the federal level as power resource exists if the exact same parties constitute the sub-state government
as the federal government. Furthermore, the number of bureaucrats whose ministers are affiliated with the same
party family will be included. Due to a lack of research on party politics as power resource, the theoretical status of
party politics is unclear; therefore, the individual parties are included in the model as a control variable. It might be
that some political parties are more powerful than others for reasons yet to be determined.
TABLE 3 Descriptive statistics of independent variables
Variable Min. Max. Median Mean SD
No votes in Bundesrat 3 6 4 4.36 1.13
Net-contributor 0 1 0 0.27 0.44
Finance ministries 0 1 0 0.09 0.29
Bundesrat presidency 0 1 0 0.10 0.30
Government chancellery 0 1 0 0.09 0.29
Federal level 0 1 0 0.08 0.27
Density of intra-governmental network (in per cent) 47 99 71 70.65 13.73
Ideological distance of parties in government 2 325 159 128.22 102.65
Party congruence with federal level 0 1 0 0.36 0.48
No. of party peers 8 98 98 71.05 30.26
Party of the Minister – CDU 0 1 0 0.29 0.45
Party of the Minister – SPD 0 1 1 0.52 0.50
Party of the Minister – Green 0 1 0 0.15 0.35
Party of the Minister – Left 0 1 0 0.04 0.20
Experience in multi-level coordination (in years) 1 43 6 8.72 7.65
Education – law 0 1 1 0.52 0.50
Education – politics & administration 0 1 0 0.27 0.44
762 HEGELE
The experience and educational background of the actors was surveyed in the online questionnaire. In Germany,
bureaucrats usually have an education in law (Derlien 2002). Some bureaucrats have a background in political science
or public administration. Hence, it will be expected that an education in law decreases while an education in politics
or public administration increases the power of an actor.
5 | EMPIRICAL RESULTS
To test the hypotheses, a multiple regression analysis was performed. One set of models was calculated using cen-
trality in the information exchange network and one using the negotiation network as dependent variable. Several
different model specifications were calculated for each network (see Table 4 and Table 5). The ‘full models’ include
all independent variables; models (a) and (b) were calculated due to the high correlation of the party political vari-
ables. In the ‘jurisdictional and organization model’ the variables measuring experience were excluded because these
were only available for those actors who participated in the survey. The ‘federal model’ only contains those variables
which are also available and meaningful for the federal level. The control variable participation in survey is included
in the models to control for differences between those actors who participated in the survey and those who did not.
The results show that no significant differences exist between these two groups of actors.
In the information exchange network (Table 4) and the negotiation network (Table 5), institutional power
resources, domestic constraints and party politics have a clear impact on power. Bureaucrats from the government
chancellery are more central than the other bureaucrats, confirming Hypothesis 3b. This institutional power resource
has the strongest, significant and positive, effect on power, which is also stable across model specifications and cor-
responds to the expectations. The effects of the other two institutional power resources are less clear. Bundesrat
presidency (H3a) has a positive effect as expected, yet it is not significant. Being from the federal level (H3c) does
not have a clear effect on power in the information exchange network. In the negotiation network, however, bureau-
crats from the federal level are clearly and significantly less powerful, which contradicts the initial expectation.
Bureaucrats from the federal government are apparently included as actors among others when it comes to informa-
tion exchange in the Bundesrat process, but are excluded when it comes to negotiating final positions, which takes
place primarily among the sub-state actors.
The density of the intragovernmental network has a positive and stable effect on centrality in both networks,
confirming Hypothesis 4a. The closer the bureaucrats within one government coordinate, the more powerful these
actors are in IGR. Hence, communicating with the other departments within their own jurisdiction makes the bureau-
crats more powerful. The distance between the coalition partners (H4b), on the other hand, has a negative effect on
power. This contradicts the initial hypotheses as it indicates that governments with a lower distance between the
coalition partners are more powerful. This could be explained by another mechanism, namely that bureaucrats from
governments where the coalition partners are ideologically closer to each other are better able to speak with one
voice in IGR and thus can jointly pursue their interests.
Party politics, indeed, are also a power resource in IGR processes, albeit in a different form than expected. Inter-
estingly, the models show that in the information exchange and the negotiation network, especially the Lefts and
Greens and, to a lesser extent, the Social Democrats are more powerful compared to the Christian Democrats. Nei-
ther the Lefts nor the Greens are part of the federal government and in the Länder are usually the smaller coalition
partners of the Social or Christian Democrats. This indicates that during IGR in the Bundesrat, the smaller parties are
more powerful because their agreement is needed in order to find an intragovernmental decision and avoid absten-
tion. This makes them strongly involved in the information exchange network and gives them power when it comes
to negotiations. On the other hand, neither the number of party political peers nor the congruence with the federal
level impact on the power of bureaucratic actors, contrary to Hypotheses 5a and 5b.
Voting power and financial power as power resources have an effect in the negotiation network but not the
information exchange network, thus partly confirming Hypotheses 1 and 2a. This indicates that when it comes to
HEGELE 763
TABLE 4 Determinants of centrality in the information exchange network
Full model (a)
Full model (b)
Jurisdictional & Organization model (a)
Jurisdictional & Organization model (b)
Federal model (a)
Federal model (b)
Voting power
No. BR votes 0.451 0.510* 0.319 0.223
(0.320) (0.196) (0.422) (0.299)
Financial power
Net-contributor −0.262 −0.174 1.029 0.646
(1.046) (0.805) (1.005) (0.813)
Finance ministry 1.327 1.445 2.179 2.131 2.238* 2.313
(1.530) (1.598) (1.261) (1.376) (1.048) (1.199)
Institutional power
BR presidency 0.875 1.181 1.964 2.096*
(0.741) (0.704) (0.953) (0.853)
Government chancellery 4.251*** 4.441*** 5.253*** 5.357*** 5.831*** 6.124***
(0.960) (1.005) (0.852) (0.876) (0.866) (0.915)
Federal level −1.097 0.148
(0.524) (0.504)
Domestic constraint
Density intragovernmental network
0.083** 0.064* 0.098** 0.0798*
(0.0246) (0.028) (0.0271) (0.0305)
Distance coalition parties −0.210 0.001 −0.572 −0.158
(0.451) (0.439) (0.337) (0.338)
Party politics
Party congruence (fed.) −1.5530 −0.232
(0.909) (1.132)
No. party peers −0.0012 −0.0103 −0.0042
(0.0158) (0.0116) (0.0154)
SPD 2.794** 1.784 2.560*
(0.883) (0.852) (0.876)
Greens 3.550* 2.671 3.773**
(1.411) (1.286) (1.072)
Left 5.329*** 4.496*** 5.898***
(0.963) (0.772) (1.050)
Experience
Years in IGR −0.0507 −0.0401
(0.0436) (0.0416)
Education law −0.629 −0.635
(1.329) (1.323)
Education politics or administration
0.123 0.370
(1.420) (1.258)
Participation in survey 1.184 1.227
(0.652) (0.676)
N 103 103 160 160 186 186
adj. R2 0.182 0.251 0.287 0.324 0.206 0.299
Standard errors in parentheses. *p < .05, **p < .01, ***p < .001.
764 HEGELE
TABLE 5 Determinants of centrality in the negotiation network
Full model (a)
Full model (b)
Jurisdictional & Organization model (a)
Jurisdictional & Organization model (b)
Federal model (a)
Federal model (b)
Voting power
No. BR votes 0.704* 0.540* 0.682** 0.526**
(0.288) (0.216) (0.174) (0.145)
Financial power
Net-contributor 2.629* 1.945 2.317** 1.682**
(1.188) (0.933) (0.644) (0.538)
Finance ministry 1.846 1.899 1.863 1.852 1.743* 1.831*
(1.224) (1.215) (0.888) (0.930) (0.732) (0.817)
Institutional power
BR presidency 1.129 1.201 1.133 1.181 1.129 1.201
(0.898) (0.936) (0.745) (0.701) (0.898) (0.936)
Government chancellery 4.555*** 4.673*** 4.921*** 5.000*** 5.090*** 5.327***
(0.705) (0.643) (0.652) (0.626) (0.589) (0.574)
Federal level −2.756*** −2.131***
(0.330) (0.306)
Domestic constraint
Density intragovernmental network
0.074* 0.0545* 0.0658*** 0.0525**
(0.0265) (0.0209) (0.0140) (0.0136)
Distance coalition parties −0.729 −0.273 −0.694** −0.286
(0.409) (0.285) (0.227) (0.210)
Party politics
Party congruence (fed.) 0.174 0.488
(0.899) (0.598)
No. party peers −0.003 −0.0079 −0.008
(0.010) (0.0064) (0.006)
SPD 1.474 0.826 1.029*
(0.688) (0.695) (0.480)
Greens 2.110 1.570 2.469**
(1.290) (1.152) (0.737)
Left 2.290** 1.881* 2.167*
(0.631) (0.635) (0.833)
Experience
Years in IGR −0.015 −0.014
(0.038) (0.036)
Education law −0.264 −0.480
(0.814) (0.748)
Education politics or administration
0.001 0.080
(0.984) (0.918)
Participation in survey 0.857 0.876
(0.441) (0.441)
N 103 103 160 160 183 183
adj. R2 0.309 0.339 0.428 0.440 0.387 0.433
Standard errors in parentheses. *p < .05, **p < .01, ***p < .001.
HEGELE 765
decision-making, these two hard power resources are more important. In the negotiation network, voting power has
a strong positive and significant effect, as expected. Bureaucrats representing Länder with a higher number of votes
are more central and powerful when it comes to negotiations. Similarly, bureaucrats representing the net-
contributors in the federal fiscal equalization scheme are more powerful in negotiations, confirming the expectation.
Being from the finance ministries also seems to have a positive, yet not always significant, effect on power (H2b).
Contrary to expectations, experience does not have an effect on power (H6a). Years of experience in IGR, if any-
thing, seems to have a negative effect which, however, is not significant. Having an education in law has a negative
and education in politics and public administration a positive but also not significant effect on power, a tendency that
corresponds to the initial expectations (H6b).
6 | CONCLUSION
This study has provided a new approach to research the power and influence of bureaucratic actors in relation to
each other. When involved in joint decision-making processes, bureaucrats are to different degrees central and pow-
erful during the process and thus have different potentials to influence the final decision. This complements the pre-
dominantly vertical study of bureaucratic influence on politicians, by adding a horizontal perspective. It demonstrates
that not all bureaucrats involved in a decision-making and coordination process are equally powerful and that varying
distributions of power resources affect their power and influence.
Organizational and jurisdictional power resources, in contrast to personal power resources, thereby seem to mat-
ter most in bureaucratic coordination processes, even when politicians are not involved. The functional departments’
bureaucrats often contact bureaucrats representing the government chancellery (institutional power). This indicates
that the departments, at least in this context, indeed include them in the coordination process and share information,
questioning a dominant departmental orientation (Fleischer 2011, p. 60).
Bureaucrats whose ministers are affiliated to certain political parties, surprisingly the smaller and federal opposition
parties, play an important and central role in bureaucratic coordination. Thus, party politics and functional politicization
can make even bureaucratic actors more powerful. In the context of German IGR, bureaucratic coordination represents a
mechanism to include minority and federal opposition parties and thus seems to be oriented to inclusion and consensus.
Party politics, hence, should be systematically included in the study of bureaucratic politics (see also Brummer 2017).
Furthermore, close coordination with the other bureaucrats from their own jurisdiction is also an important
power resource. This might indicate that in bureaucratic IGR in federal states, in contrast to international relations,
the two-level logic works with a different mechanism. Actors do not become more powerful through sectoral con-
straints by the other departments or political constraints by their coalition partners, but rather closeness and coordi-
nation among these actors make them more powerful because they can jointly pursue their interests.
Finally, the power resources that actors use differ between processes of information exchange and negotiations.
The hard power resources, voting and financial power, which seem important in international relations, are only rele-
vant for bureaucratic intergovernmental coordination when it comes to negotiations. The process of coordination is
more precisely captured by distinguishing these phases.
An analytical framework for the investigation of power resources of bureaucrats in a horizontal perspective is
proposed here, which is applicable to IGR in any multi-level state. While IGR in multi-level states significantly differ
from each other, the hypotheses developed are sufficiently broad and can be adapted to varying national and inter-
national contexts. Thus, they offer a framework to investigate this aspect of bureaucratic IGR in a comparative per-
spective. Furthermore, these findings might also be transferred to the study of inter-departmental decision-making
processes within one government and thus advance public administration research in general.
ORCID
Yvonne Hegele https://orcid.org/0000-0003-1567-0317
766 HEGELE
REFERENCES
6, P. (2004). Joined-up government in the western world in comparative perspective: A preliminary literature review and
exploration. Journal of Public Administration Research and Theory, 14, 103–138. Allison, G. T., & Halperin, M. H. (1972). Bureaucratic politics: A paradigm and some policy implications. World Politics, 24,
40–79. Allison, G. T., & Zelikow, P. (1999). Essence of decision: Explaining the Cuban missile crisis (2nd ed.). Boston, MA: Little, Brown. Auel, K. (2014). Intergovernmental relations in German federalism: Cooperative federalism, party politics and territorial
conflicts. Comparative European Politics, 12, 422–443. Bailer, S. (2010). What factors determine bargaining power and success in EU negotiations? Journal of European Public Policy,
17, 743–757. Behnke, N. (2013). Was sind Grundsätze für ein gutes Finanzausgleichssystem? Zeitschrift für Politikwissenschaft, 23,
105–115. Behnke, N., & Mueller, S. (2017). The purpose of intergovernmental councils: A framework for analysis and comparison.
Regional & Federal Studies, 27, 507–527. Bolleyer, N., Swenden, W., & McEwen, N. (2014). A theoretical perspective on multi-level systems in Europe: Constitutional
power and partisan conflict. Comparative European Politics, 12, 367–383. Bouckaert, G., Peters, B. G., & Verhoest, K. (2010). The coordination of public sector organizations: Shifting patterns of public
management. Basingstoke: Palgrave Macmillan. Bräuninger, T., & Debus, M. (2011). Parteienwettbewerb in den deutschen Bundesländern. Wiesbaden: Springer-Verlag. Brummer, K. (2017). Governmental politics in consensus democracies. Global Society, 31, 272–292. Christensen, T., & Lægreid, P. (2009). Living in the past? Change and continuity in the Norwegian Central Civil Service. Public
Administration Review, 69, 951–961. Coddington, A. (1968). Theories of the bargaining process. London: Allen & Unwin. Dahlström, C., Peters, B. G., & Pierre, J. (2011). Steering from the centre: Strengthening political control in western democracies.
Toronto: University of Toronto Press. Derlien, H.-U. (2002). Öffentlicher Dienst im Wandel. In K. König (Ed.), Deutsche Verwaltung an der Wende zum 21. Jahrhun-
dert (pp. 229–253). Baden-Baden: Nomos. Fleischer, J. (2011). Steering from the German centre: More policy coordination and fewer policy initiatives. In C. Dahlström,
B. G. Peters, & J. Pierre (Eds.), Steering from the centre: Strengthening political control in western democracies (pp. 54–77). Toronto: University of Toronto Press.
Freeman, L. C. (1978). Centrality in social networks conceptual clarification. Social Networks, 1, 215–239. Hanneman, R. A., & Riddle, M. (2005). Introduction to social network methods. Riverside, CA: University of California,
Riverside. Hartlapp, M., Metz, J., & Rauh, C. (2013). Linking agenda setting to coordination structures: Bureaucratic politics inside the
European Commission. Journal of European Integration, 35, 425–441. Hegele, Y. (2018). Multidimensional interests in horizontal intergovernmental coordination: The case of the German Bundes-
rat. Publius: The Journal of Federalism, 48, 244–268. Hegele, Y., & Behnke, N. (2017). Horizontal coordination in cooperative federalism: The purpose of ministerial conferences
in Germany. Regional & Federal Studies, 27, 529–548. Heller, R. (2015). Haushaltsrecht und Haushaltssystematik. In W. Gatzer & T. Schweisfurth (Eds.), Öffentliche Finanzwirtschaft
in der Staatspraxis (pp. 253–294). Berlin: Berliner Wissenschafts-Verlag. Henning, C. H. C. A. (2009). Networks of power in the CAP system of the EU-15 and EU-27. Journal of Public Policy, 29,
153–177. Hood, C., & Lodge, M. (2006). The politics of public service bargains: Reward, competency, loyalty and blame. Oxford: Oxford
University Press. Johns, C. M., O'Reilly, P. L., & Inwood, G. J. (2007). Formal and informal dimensions of intergovernmental administrative rela-
tions in Canada. Canadian Public Administration, 50, 21–41. Lehmbruch, G. (2000). Parteienwettbewerb im Bundesstaat. Regelsysteme und Spannungslagen im politischen System der Bun-
desrepublik Deutschland (3rd ed.). Wiesbaden: Westdeutscher Verlag. León, S. (2017). Intergovernmental councils in Spain: Challenges and opportunities in a changing political context. Regional &
Federal Studies, 27, 645–665. Leunig, S., & Träger, H. (2014). Landesinteressen, Parteipolitik, 'Parteidruck’ – der Bundesrat im Fokus divergierender Inter-
essen. Zeitschrift für Politikwissenschaft, 24, 55–81. Lhotta, R., & van Blumenthal, J. (2015). Intergovernemental relations in Germany: Complex co-operation and party politics.
In J. Poirier, C. Saunders, & J. Kincaid (Eds.), Intergovernmental relations in federal systems: Comparative structures and
dynamics (pp. 206–238). Don Mills, Ontario: Oxford University Press. Mayntz, R., & Scharpf, F. (1975). Policy-making in the German federal bureaucracy. Amsterdam: Elsevier. McClurg, S. D., & Young, J. K. (2011). Editors' introduction: A relational political science. PS: Political Science and Politics, 44,
39–43.
HEGELE 767
Metcalfe, L. (1994). International policy co-ordination and public management reform. International Review of Administrative Sciences, 60, 271–290.
Page, E. (2012). Policy without politicians: Bureaucratic influence in comparative perspective. Oxford: Oxford University Press. Parry, R. (2012). The civil service and intergovernmental relations in the post-devolution UK. British Journal of Politics & Inter-
national Relations, 14, 285–302. Peters, B. G. (1998). Managing horizontal government: The politics of co-ordination. Public Administration, 76, 295–311. Peters, B. G., Erkkilä, T., & von Maravić, P. (2016). Public administration: Research strategies, concepts, and methods. New York
and London: Routledge. Peters, B. G., & Pierre, J. (2016). Comparative governance: Rediscovering the functional dimension of governing. Cambridge:
Cambridge University Press. Peters, B. G., Rhodes, R. A. W., & Wright, V. (2003). Administering the summit: Administration of the core executive in developed
countries. Basingstoke: Palgrave Macmillan. Poirier, J., & Saunders, C. (2015). Comparing intergovernmental relations in federal systems: An introduction. In J. Poirier,
C. Saunders, & J. Kincaid (Eds.), Intergovernmental relations in federal systems: Comparative structures and dynamics (pp. 1–13). Don Mills, Ontario: Oxford University Press.
Poirier, J., Saunders, C., & Kincaid, J. (2015). Intergovernmental relations in federal systems: Comparative structures and dynam- ics. Don Mills, Ontario: Oxford University Press.
Putnam, R. D. (1988). Diplomacy and domestic politics: The logic of two-level games. International Organization, 42, 427–460.
Reuter, K. (2007). Praxishandbuch Bundesrat. Verfassungsrechtliche Grundlagen, Kommentar zur Geschäftsordnung, Praxis des Bundesrates (2nd ed.). Heidelberg: C. F. Müller.
Rohlfing, I. (2012). Case studies and causal inference: An integrative framework. Basingstoke: Palgrave Macmillan. Scharpf, F. (1985). Die Politikverflechtungs- Falle: Europäische Integration und deutscher Föderalismus im Vergleich. Poli-
tische Vierteljahresschrift, 26, 323–356. Schnapp, K.-U. (2004). Ministerialbürokratien in westlichen Demokratien. Eine vergleichende Analyse. Opladen: Leske + Budrich. Schneider, G., & Bailer, S. (2002). Mächtig, aber wenig einflussreich: Ursachen und Konsequenzen des deutschen Integra-
tionsdilemmas. Integration, 25, 49–60. Schrenk, K. (2010). Die Vertretungen der Länder beim Bund. In K. Schrenk & M. Soldner (Eds.), Analyse demokratischer Regier-
ungssysteme (pp. 359–374). Wiesbaden: VS Verlag. Schwanke, K., & Ebinger, F. (2006). Politisierung und Rollenverständnis der deutschen Administrativen Elite 1970 bis 2005.
Wandel trotz Kontinuität. Politische Vierteljahresschrift (PVS), Sonderheft 37. Politik und Verwaltung, 228–249. Scott, J. (2013). Social network analysis (3rd ed.). Los Angeles, CA: Sage. Shepsle, K. A., & Boncheck, M. S. (1997). Analyzing politics: Rationality, behavior, and institutions. New York and London:
W. W. Norton. Snyder, G. H., & Diesing, P. (1977). Conflict among nations: Bargaining, decision making, and system structure in international
crises. Princeton, NJ: Princeton University Press. Tallberg, J. (2008). Bargaining power in the European Council. JCMS: Journal of Common Market Studies, 46, 685–708. Wagener, F. (1979). Der Öffentliche Dienst im Staat der Gegenwart. Veröffentlichung der Vereinigung der deutschen Staats-
rechtslehrer, 37, 215–266. Wasserman, S., & Faust, K. (2009). Social network analysis: Methods and applications (19th printing ed.). Cambridge: Cam-
bridge University Press.
How to cite this article: Hegele Y. Explaining bureaucratic power in intergovernmental relations: A network
approach. Public Admin. 2018;96:753–768. https://doi.org/10.1111/padm.12537
768 HEGELE
Copyright of Public Administration is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.