Content uploaded by Andreas Hofmann
Author content
All content in this area was uploaded by Andreas Hofmann on Aug 21, 2015
Content may be subject to copyright.
Strategies of the Repeat Player
Andreas Hofmann
Strategies of the Repeat Player
The European Commission between
Courtroom and Legislature
Andreas Hofmann
One of the startling empirical insights from the
growing body of research on the workings of the
EU judicial system is that the European Commis-
sion almost always wins the cases that it argues.
Since the European Court of Justice has been
the author of many important decisions shaping
the content of EU policies, being a frequent and
frequently successful litigator provides the Com-
mission with an alternative means of achieving
policy objectives, at least when all else fails.
This book systematically assesses how the
Commission makes use of its access to both
legislation and litigation in order to advance its
policy interests.
ISBN 978-3-8442-6195-0
Andreas Hofmann
Strategies of the Repeat Player
The European Commission between
Courtroom and Legislature
Imprint
Copyright: © 2013 Andreas Hofmann
Druck und Verlag: epubli GmbH, Berlin, www.epubli.de
ISBN 978-3-8442-6195-0
Cover picture:
Jost Amman (1539-1591): Der Bierbreuwer.
Eygentliche Beschreibung aller Stände auff Erden hoher und nidriger,
geistlicher und weltlicher, aller Künsten, Handwerken und Händeln.
Frankfurt am Main 1568. Artwork in the public domain.
Für meine Eltern
Contents
Preface xi
Introduction 1
1.1 Judicial politics in the European Union 7
1.2 The structure of my study 11
Studying the European Commission 18
2.1 The European Commission in the study of European Union
politics 18
2.2 Who is ‘the Commission’ and what does it want? 32
The European Commission in the policy-making
process 43
3.1 Modes of policy-making 43
3.2 The Commission and the intergovernmental mode 52
3.3 The Commission and the joint-decision mode 57
3.4 The Commission and the supranational-hierarchical mode 66
3.5 Comparing the Commission’s position in the three policy
modes 92
Predicting strategy choice 95
4.1 A default strategy 95
4.2 Factors influencing the Commission’s choice of strategy 99
4.3 Data 101
4.4 Results and discussion 133
Removing barriers to trade 138
5.1 Overview 139
5.2 Quantitative restrictions and measures having equal effect 141
5.3 Mutual recognition and the harmonization of laws regulating
trade 158
5.4 Mutual recognition and the single market programme 178
5.5 Mutual recognition after 1992 186
5.6 Conclusion 193
From market citizen to European citizen 200
6.1 Overview 201
6.2 Cooperation and emerging friction 204
6.3 Moving beyond the economic base of free movement 213
6.4 Towards a general right of residence 219
6.5 Defining the substance of European citizenship 233
6.6 Conclusion 241
Conclusions 247
7.1 Reviewing my hypotheses 249
7.2 Preliminary references and judicial strategy 254
7.3 Some thoughts on the wider legal environment 257
References 261
Figures and Tables
Figures
3.1 Commission strategies in three modes of policy-making 47
3.2 Agenda setting in EU legislation 61
4.1 Commission legislative initiatives 104
4.2 Commission initiatives for regulations and directives 105
4.3 Infringement proceedings referred to the Court 108
4.4 Infringement proceedings referred to the Court – by accession
round 109
4.5 Infringement proceedings referred to the Court – by subject
matter 110
4.6 Infringement proceedings referred to the Court and legislative
proposals 116
4.7 Mean of dependent variable over time 119
4.8 Mean of variables EP veto and Council voting rule over time
124
4.9 Mean of variables controversy and prevalence of directives
over time 127
4.10 EU legislative institutions’ mean party political position on a
left-right scale 129
4.11 Annual absolute ideological distance between EU legislative
institutions 130
5.1 Legislative initiatives and cases referred to the Court 195
6.1 Legislative initiatives and cases referred to the Court 244
Tables
4.1 Cross-tabulation of policy field and responsible DG 114
4.2 Type of Commission action by policy field 115
4.3 Information on Council voting rule 121
4.4 Information on EP participation 121
4.5 Descriptive statistics on all variables in the analysis 132
4.6 Results of OLS regression (with time fixed effects) 134
Preface
One of the startling empirical insights from the growing body
of research on the workings of the EU judicial system is that
the European Commission almost always wins. This finding
is remarkably robust – every empirical study I am aware of
that deals with success and failure in proceedings before the
European Court of Justice has so far replicated this result.
Whether the Commission litigates itself or lodges an ‘obser-
vation‘ (a legal brief outlining its position) in a case brought
by a different party, the Court follows the Commission’s
opinion in the vast majority of cases. Since empirical data on
judicial behaviour at the Court of Justice is sparse – judges
do not publish votes or record dissents – the source of this
‘special rapport’ is difficult to establish. But whatever the
reason may be, it seemed to me that this crucial ‘asset’ would
have to have some bearing on the Commission’s overall
stance in the EU’s policy-making process. Since the Court
has been the author of so many ‘history-making’ decisions
shaping the content of EU policies, being a frequent and fre-
quently successful litigator should provide the Commission
with an alternative means of achieving policy objectives, at
least when all else fails. Hence, this book looks at how the
Commission makes use of its ‘special rapport’ with the Court
of Justice and what factors influence its use of court proceed-
ings vis-à-vis its alternatives in shaping EU policy. In pursu-
ing these questions, I am hoping not only to learn more
about the interaction of different modes of EU policy-
making, but also to contribute to a body of research that
takes a less ‘court-centric’ approach to ‘law and politics’ in
the EU. Like all courts, the European Court of Justice needs
to be ‘activated’ by litigants pursuing their interests couched
in legal terms. Learning more about the motives of litigants,
their alternatives to litigation, and their reasons for choosing
this course of action should contribute to a better under-
standing of the political relevance of the legal system.
Looking back at the long process of which this publication is
a preliminary culmination, I realise that I am indebted to
many sources of inspiration and support, only some of which
I can actually name. First of all I would like to thank my the-
sis supervisor, Professor Wolfgang Wessels, who both sup-
ported my fixation on the Commission and the Court and
urged me to maintain a broader interest in EU politics. I
would further like to thank Professor André Kaiser, whose
seminar on constitutional courts in established democracies
provided the nucleus of my idea for this study, and Professor
Ingo Rohlfing for making me much more self-conscious
about methodology. Between 2010 and 2011, I had the good
fortune to be able to spend some time as visiting graduate
student at the University of Chicago, where I learned a great
deal about the state of research on ‘law and politics’. I am
greatly indebted to Professor Gerald Rosenberg, first for giv-
ing me this opportunity, and second for making me think
much more broadly about the role of courts in society. I
would also like to thank Professor Susanne Schmidt for her
support of my work, which borrows many insights from hers.
Anette Fasang patiently answered my many questions about
the appropriate use of statistics, and I am very grateful for
her help. Naturally, she bears no responsibility for the re-
maining defects.
My deepest gratitude goes to Franzi Bedorf, who not only
read the entire manuscript, but endured my increasingly er-
ratic schedule and supported me at every step. Thank you so
much for being part of my life.
Finally, I dedicate this book to my parents, Gabriele and
Joachim Hofmann, whose unconditional support made all
this possible.
Preface xii
Chapter 1
Introduction
When Duke Wilhelm IV of Bavaria decreed a purity law for
beer on April 23, 1516, he had in mind a late medieval
equivalent of consumer protection. Knowledge about the bio-
logical details of the fermentation process was not widely
available, and many contemporary beers must have tasted
quite foul, particularly during the summer, when high tem-
peratures frequently caused brews to go off. Since beer was
still preferable to unsanitary fresh water, and its production
lucrative business, brewers frequently made use of all sorts of
questionable and not rarely toxic additives to cover up the
off-tastes of their products. The more fanciful of these in-
cluded fly agaric mushrooms, ivy, chalk, soot, snake juice and
oxen bile (cf. Thomas 2006: 32; Dornbusch 2011: 104-105).
To reign in such creativity, Wilhelm IV’s decree restricted the
ingredients allowed in the brewing process to barley, hops
and water.1
Originally only a Bavarian statute, it was not until four cen-
turies later that the German Reichstag established the purity
law as a nation-wide rule, in 1906, much to the detriment of
some northern German brewing traditions that had included
various herbs and spices to vary the flavour of their product
(cf. Thomas 2006: 32). The term ‘Reinheitsgebot’, as the law
is known today, was coined shortly thereafter, in the wake of
the foundation of the first German republic in 1919. Bavarian
delegates used the term in their insistence that the Bavarian
law remain the law of all of Germany, in which they suc-
1 Yeast was not included in the original list since its properties were un-
known at the time. The use of malted wheat for wheat beers was reserved for
families of the nobility, and subsequently became a very lucrative niche
(Dornbusch and Oliver 2011: 829).
ceeded – the newly assembled Reichstag retained the prior
‘Biersteuergesetz’ codifying the purity law. By the time it
celebrated its 450 year anniversary, the purity law was con-
sidered the world’s oldest continuously existing food quality
standard (cf. Dornbusch and Heyse 2011: 692).
Sometime in late 1981, a German distributor announced to
the French brewery Brasserie du Pêcheur, based in Schil-
tigheim in Alsace, that it would cease to carry its products.
German authorities had repeatedly removed Brasserie du
Pêcheur beers from retail shelves, on the grounds that the
Brasserie’s use of additives in the brewing process ran afoul
of Germany’s most recent codification of the purity law, the
revised Biersteuergesetz of 1952, which allowed for bottom
fermenting beers such as the Brasserie’s only the use of
malted barley, hops, yeast and water. Products that con-
tained other ingredients could not be sold in Germany as
‘beer’. Excluded from the lucrative German beer market, the
Brasserie du Pêcheur complained to the European Commis-
sion. In restricting its market to beers brewed according to
the purity laws, it argued, Germany was infringing the Euro-
pean Community’s free market principles, in particular Arti-
cle 30 of the EEC Treaty, which forbade quantitative restric-
tions to trade, and all measures having equivalent effect (cf.
Clark 1988: 769-770).
The removal of such barriers to trade had been a priority of
the Commission for quite some time. Its principal approach
to this issue throughout the 1960s and 1970s had been to
propose legislation introducing common product standards
that would preclude member states from applying their own
rules to imported products. This process was cumbersome,
due on one hand to the complicated nature of detailed stan-
dards and the constant need to update rules to technological
progress, and on the other to decision-making rules that gave
every member state a veto over individual pieces of legisla-
Introduction 2
tion. Product standards can serve ambiguous purposes; they
are obviously necessary to protect consumers and the envi-
ronment, but they can also be a subtle form of protectionism,
especially where a certain standard is basically a codification
of prevalent practices in one country, drawn up by the lead-
ing domestic producers. It is not entirely surprising that
member state governments were reluctant to relinquish such
standards, especially where they affected important domestic
industries.
Among the pieces of legislation proposed by the Commission
in the field of product standards had been one concerning
the characteristics of beer produced within the Community
(OJ 1972, No. C 105/17). It suggested that European stan-
dards on the content of beers follow the example of French
laws, which allowed the use of up to 30 percent of raw mate-
rials other than malted barley or wheat for the production of
fermentable wort.2 That piece of legislation, and subsequent
amendments, went nowhere in the Council, owing mainly to
resistance by the German government (cf. Schweitzer and
Streinz 1984: 42).
By the time the Brasserie du Pêcheur lodged its complaint,
the Commission had announced a change of strategy. Rather
than introduce more legislation, the Commission would rely
on existing laws on the free movement of goods, as it itself
interpreted them, and bring enforcement actions before the
Court of Justice where such laws were infringed by member
states. The Court of Justice’s recent case law had proven fa-
vourable to this approach, with a long series of cases, most
notably Dassonville (Case 8/74, ECR 1974: 838) and Cassis
de Dijon (Case 120/78, ECR 1979: 650),3 restricting member
state discretion to autonomously regulate trade. Prior to the
Introduction 3
2 Cf. “Eurobrew”, The Economist, 29 December 1973, Survey, p. 10.
3 Cf. chapter 5.
beer purity law case, the Commission had taken the Italian
government to court for restricting the designation ‘aceto’
(vinegar) to products derived from wine (Case 2193/80, ECR
1981: 3019). The Court supported the Commission’s position
that this was an illegal restriction to trade, as products made
from other raw materials, like apple must or malt, could
freely be marketed as ‘vinegar’ in other member states. The
analogy to beer is obvious, but the German beer case was
special, as it affected both the largest member state and one
of its most coveted products, protected by a rule that dated
back almost half a millennium. The German beer market in
the 1980s was second only to the US in both production and
revenue, and Germans consumed an annual average of about
150 litres per head (240 litres in Bavaria), by far the highest
rate of consumption in the EEC (cf. Schweitzer and Streinz
1984: 47).4 In fact, the German government claimed that
beer was the single most consumed foodstuff in Germany,
and its protection warranted particular measures (Case 178/
84, ECR 1987: 1236-1237, German edition). The German
laws, which effectively functioned as an import ban on non-
Reinheitsgebot beers, helped secure domestic producers a
market share of about 99 percent (cf. Kohler 1987: 13; Pal-
trow 1987).
Following its revised strategy, the Commission took no heed
of German protestations and initiated an infringement pro-
cedure against Germany with a formal letter to the German
government in February 1982. Since no agreement could be
reached in the early stages of the procedure, the Commission
referred the case to the Court of Justice in July 1984. (Inci-
dentally, the Commission opened parallel proceedings
against Greece, the only other European country with a pu-
rity law for beer, the Greek king Otto, who decreed it during
Introduction 4
4 Cf. also “Bier - hierzulande mehr als ein Getränk”, Frankfurter Allgemeine
Zeitung, 2 February 1986, p. B5.
his reign from 1832-1862, having been of Bavarian origin.)5
The Commission argued that the German laws distorted
trade in two ways. First, by banning the import of non-
Reinheitsgebot beers, it protected its domestic market from
foreign competition. Second, by exempting from the Rein-
heitsgebot beers brewed in Germany exclusively for export, it
simultaneously allowed German brewers to compete on the
world market. By virtue of this ostensibly protectionist regu-
lation, German exports of beer exceeded imports by a factor
of four to one (Case 178/84, ECR 1987: 1234, German edi-
tion). Moreover, the Commission argued that the import ban
on non-Reinheitsgebot beers could not be justified on
grounds of consumer protection. Even where the prohibited
raw materials and additives commonly used in foreign beers
may constitute health risks, these same substances could le-
gally be used in the production of other foodstuffs in Ger-
many, most notably in wine (Case 178/84, ECR 1987: 1236,
German edition).6 In the Commission’s view, there was
therefore no reason that beer that could be legally be in other
member states should not also be sold in Germany.
In its reply, the German government argued that in the ab-
sence of harmonizing legislation, member states retain the
right to autonomously regulate their domestic markets. As
long as this was so, different national regulations of necessity
lead to trade barriers, but these could only be addressed by
Community legislation (Case 178/84, ECR 1987: 1235, Ger-
man edition). Following the Commission’s opinion would
moreover lead to a dangerous increase in the consumption of
additives, as every state would have to adapt its legislation to
Introduction 5
5 Greek beers not being known for their excellence, some observers view this
as evidence falsifying the assumption that a purity law is a sufficient condi-
tion for high quality beers. A trip to Belgium may also convince that neither
is it a necessary one.
6 The German Association of Consumer Organizations quipped that ''The
only ingredient in domestic and foreign beer that has so far been proven to
be potentially damaging is alcohol'' (Quoted in Markham 1987: A7).
the position of the most lenient member state (Case 178/84,
ECR 1987: 1240, German edition). Finally, it argued that the
concept of ‘beer’ was “inseparably linked” to the Reinheitsge-
bot in the mind of he German consumer, so that for the sake
of consumer protection any product differently produced
should not carry that designation (Case 178/84, ECR 1987:
1270).
The Court did not accept the German government’s argu-
ments. To the latter, it replied that consumer concepts of
products evolve over time and that such an evolution must
not be precluded by marketing regulations. Quoting an ear-
lier judgement, the judges stated that “the legislation of a
Member State must not 'crystallize given consumer habits so
as to consolidate an advantage acquired by national indus-
tries concerned to comply with them'” (Case 178/84, ECR
1987: 1270-1271). Rather than a ban on imports, a simple
labeling requirement would sufficiently serve the purpose of
consumer protection in this regard. Moreover, the Court did
not follow the German government’s interpretation of the
discretion available to national governments in regulating
domestic markets where there was no common Community
legislation. While it agreed that, in the absence of harmoniza-
tion, member states retained some regulatory powers, these,
it said, may not be used to pursue unjustified goals, or, where
goals such as public health and consumer protection are jus-
tified, result in measures that go beyond what is necessary to
attain these goals. Consequently, it concluded that “in so far
as the German rules on additives in beer entail a general ban
on additives, their application to beers imported from other
Member States is contrary to the requirements of Commu-
nity law as laid down in the case-law of the Court, since that
prohibition is contrary to the principle of proportionality”
(Case 178/84, ECR 1987: 1276).
Introduction 6
Another precedence was set, and the Commission could
claim an essential vindication of its new, judicial, strategy to
remove barriers to trade within the Community, following
the principle that a product lawfully produced and marketed
in one member state should equally be marketable in all
other member states. After its initial legislative strategy had
failed, the Commission initiated judicial proceedings, seizing
upon legal innovations, and succeeded in establishing a prin-
ciple that allowed it forthwith to minimize its legislative ac-
tivity to cover only minimum standards that member states
could more easily agree on.
This Commission strategy, this interplay between legislative
politics and the judicial sphere is what this book is about.
What I attempt is to provide a more inclusive picture of how
the European Commission acts within the policy process, a
picture that spans various modes of producing policy and
indicates which factors account for the choice of strategy. In
short, I want to systematically explore a connection between
judicial and legislative politics that has as yet not been exten-
sively addressed (cf. Schmidt 2011a: 43; Schmidt 2011b: 37).
1.1 Judicial politics in the European Union
The proximity of law and politics in the European Union, as
demonstrated by the Reinheitsgebot case, has certainly not
escaped the attention of observers from the social sciences,
although they have been late in catching on to the connec-
tion. The central legal principles had all been long estab-
lished when political scientists began debating the role of the
European Court of Justice in EU politics. Early debates cen-
tred around the question if the Court independently set poli-
cies, or whether it was influenced significantly by the prefer-
ences of the largest member states (cf. Burley 1993; Burley
and Mattli 1993; Garrett 1995; Mattli and Slaughter 1995;
Alter 1998; Garrett, Kelemen et al. 1998; Mattli and Slaugh-
Introduction 7
ter 1998; Stone Sweet and Brunell 1998). A satisfying empiri-
cal answer to this question proved elusive, not least because
data on judicial decision-making in the Court of Justice was
lacking (cf. Scharpf 2011: 229). The judges do not publish
their ‘votes’, and there are no dissents or concurring opin-
ions. Although for this reason the question was never conclu-
sively settled, some central insights emerged. For one, the
Court appears to act strategically. The more member states
argue for a specific norm, the less likely the Court is to strike
it down, although the Court shows no preference for the posi-
tion of large over small states (cf. Conant 2007: 51). On the
other hand, the most successful litigant before the Court is
not a member state, but the European Commission; it wins
the overwhelming majority of the cases it argues (cf. Conant
2007: 53).
The new found attention to the judicial sphere of European
Union politics spawned a whole range of research questions
that went beyond the sole issue of the Court’s independence,
and sought to locate the legal system in the broader political
context. Gradually, the focus shifted away from the Court of
Justice as the focal institution to the wider legal environ-
ment, including the role of national courts in the evolution of
EU politics (cf. Slaughter, Stone Sweet et al. 1998), and the
position of individual litigants and the goals they wish to
pursue through litigation. A significant part of the latter lit-
erature looks at the way the European legal system empow-
ers individuals to challenge the domestic status quo (cf. Alter
and Vargas 2000; Cichowski 2006; Kelemen 2006). Daniel
Kelemen’s recent book-length study of the evolving legal in-
frastructure in the European Union points to an increasingly
adversarial system of confrontation between regulators and
the regulated claiming rights derived from the EU legal sys-
tem, a process that he considers akin to American forms of
regulatory politics and dispute resolution, away from more
inclusive corporatist structures formerly prevalent through-
Introduction 8
out Europe (Kelemen 2011: 4-7). Many such studies are
keenly aware that this empowerment does not affect all ac-
tors equally, but favours strong organised interests over indi-
viduals who either cannot muster the resources to initiate the
legal process or are unaware of their rights (cf. Alter and
Vargas 2000; Börzel 2006). In particular, actors who fre-
quently engage in litigation, so called ‘repeat players’ (cf.
Galanter 1974), enjoy specific advantages in the long run,
being able to ‘play for the rules’: they can discount the costs
of individual judgements against future rewards where they
succeed in establishing legal principles that enhance their
overall position. The European Commission is surely the sin-
gle most frequent litigator before the Court of Justice, and
the prototypical repeat player on the European legal stage.
Comparatively few works have sought to combine a focus on
judicial politics with policies pursued in other spheres. An
early example of such an approach is Karen Alter and Sophie
Meunier Aitsahalia’s article on the impact of the Court’s im-
portant ruling in the Cassis de Dijon case, which laid out the
groundwork for the principle of ‘mutual recognition’ (Alter
and Meunier Aitsahalia 1994). They investigate how the
Court’s judgement, by altering the status quo with regard to
permissible trade barriers, provided novel opportunities for
various individual and institutional actors to pursue their
preferences in national and European fora. In particular,
they demonstrate how the Commission seized upon the
newly formulated doctrine and used it as a basis for a new
approach to remove intra-European trade barriers. While
Alter and Meunier bridge the gap between the judicial
branch and other spheres of action, theirs is foremost a study
of how organisations such as the Commission react to Court
rulings they can use to their advantage. Judicial processes,
however, can be strategically induced. Individuals can chal-
lenge national norms in national courts, and the Commission
Introduction 9
can litigate against member states before the Court of Jus-
tice, all in order to influence policy.
The Commission’s use of the infringement procedure has
been closely scrutinised in studies focussing on the national
implementation of EU legislative acts (cf. Mendrinou 1996;
Börzel 2001; Panke 2009; Börzel, Hofmann et al. 2012). As a
result, there is a rich body of insights about what factors in-
fluence the distribution of infringement proceedings across
member states. Prominent among these are national gov-
ernments’ administrative capacity and political clout (cf.
Börzel, Hofmann et al. 2010: 1381). However, these studies
rarely address the policy consequences that judicial proceed-
ings entail. In focussing on national compliance, they de-
emphasise the conflicts over policy content that are fought
out before the Court of Justice and their implications for fu-
ture policy-making.
The way the Commission can employ its enforcement powers
to influence legislative processes has been explored in the
works of Susanne Schmidt. She demonstrates the intercon-
nectedness between litigation and legislation by focussing on
the Commission’s ability to strategically alter the status quo
of legislative procedures through enforcement actions before
the Court of Justice. In particular, she outlines two courses of
action the Commission takes in order to apply pressure on
Council positions. One targets the status quo of legislation as
a whole, by threatening judicial procedures with an outcome
potentially disliked by all Council members. The other strat-
egy consists of targeting individual member states, particu-
larly those pivotal in majority voting. Based on these in-
sights, Fritz Scharpf feels confident to declare that “the
Commission is able to use infringement and annulment pro-
ceedings in order to place specific issues on the Court’s
agenda, and there is no question that it is able to use its
power strategically in order to influence the outcome of legis-
lative processes in the Council” (Scharpf 2011: 229-230).
Introduction 10
My study takes these insights as a starting point to system-
atically assess how the Commission makes use of its access to
legislation and litigation in order to advance its policy inter-
ests. My contribution is twofold. On the one hand, I aim to
analyse what variables influence the Commission’s decision
to resort to infringement proceedings when pursuing policy.
I start from the assumption that the Commission is more
likely to pursue a judicial strategy when the obstacles to suc-
cessful legislation are high. I test this proposition with the
help of a dataset that combines information on legislative
processes, infringement proceedings and the political posi-
tions of the actors involved. On the other hand, I expand this
approach beyond the focus on infringement proceedings to
the Commission’s use of litigation in general in two longitu-
dinal case studies. While Susanne Schmidt argues that in-
fringement proceedings are “the most decisive means for the
interaction of judicial and legislative politics” (cf. Schmidt
2011a: 50), I aim to show that other forms of litigation, in
particular the Commission’s intervention in preliminary ref-
erence procedures, also present viable channels of combining
the two modes of policy-making. In essence, I incorporate a
judicial politics approach in an overall analysis of policy-
making in an attempt to achieve a more comprehensive pic-
ture of how the Commission influences policy development
in the European Union.
1.2 The structure of my study
I structure my study in the following way. Chapter 2 provides
an overview of theoretical approaches to the Commission’s
position in the policy-making process of the European Un-
ion. The question of how much the Commission matters has
traditionally formed one of the focal points of theorising the
dynamics of European integration. I trace the ebb and flow of
scientific attention to this topic, which has been closely
linked to actual successes and failures of the European politi-
Introduction 11
cal system, and outline the central tenets of the opposing
camps. I describe how in more recent time the focus of theo-
retical approaches has moved away from dynamic accounts
of the integration process to a more cross-sectional, com-
parative outlook on the constraints of policy-making. Rather
than aligning squarely with the traditional factions of neo-
functionalism and intergovernmentalism, these newer ap-
proaches analyse the factors and circumstances that allow
the Commission to influence policy content in some in-
stances, and prevent it from doing so in others. I highlight
two approaches, the ‘agenda-setting’ literature and the
‘principal-agent’ approach, as particularly conducive to a
more detailed understanding of the Commission’s position in
European politics. Chapter 2 moreover reviews theoretical
assumptions and empirical findings concerning the Commis-
sion’s ‘actorness’. I start by outlining the ‘traditional’ aca-
demic assumptions on the type of preferences the Commis-
sion pursues. Until quite recently, the Commission was regu-
larly treated as a preference outlier, usually with an ‘extreme’
disposition towards ‘more integration’, however defined.
Some explanatory approaches draw on ‘bureaucratic politics’,
positing that the Commission’s preferences are a conse-
quence of organisational self-interest. I contrast these as-
sumption with more recent empirical data on internal pref-
erence distributions and decision-making dynamics. These
newer accounts challenge the assumptions that the Commis-
sion necessarily pursues ‘more extreme’ preferences than the
member states, and that it can be treated as a unitary actor.
Rather, such studies demonstrate distinct clusters of prefer-
ences across services, and point to procedural structures that
systematically favour certain internal departments, in par-
ticular the respective lead Directorate General (DG) and the
Secretariat General. I conclude by outlining which of these
insights can fruitfully inform my approach to Commission
strategy choice.
Introduction 12
Chapter 3 outlines the Commission’s position in the inter-
institutional policy-making process. I distinguish between
three different modes of policy-making that share the pro-
duction of legal norms as output (cf. Scharpf 2006): an inter-
governmental mode, characterised by bargaining between
member state governments, a ‘joint-decision’ mode, charac-
terised by the Commission’s monopoly of initiative in legisla-
tion, and a ‘supranational-hierarchical mode’, characterised
by the Commission’s ability to influence policy without the
involvement of the legislative institutions, by executive ac-
tion and initiating court proceedings. Each of these modes
affords the Commission different degrees of influence over
policy production, and I discuss the various constraints on
Commission policy-making strategies in some detail. On one
end of the scale, the Commission’s ability to influence out-
comes is weakest in situations of intergovernmental bargain-
ing. On the other end, its policy-making ability is greatest in
those areas where it has been delegated extensive executive
authority. This is primarily so in the field of competition pol-
icy, including the contentious areas of state-aid and public
utilities. While the Commission has the ability to act unilat-
erally, its authority is at the same time closely circumscribed
to a limited policy field. The Commission’s ability to influ-
ence the policy process is broadest in legislation and litiga-
tion – the content of legislative initiatives and court cases
being barely limited in scope. I conclude the chapter by justi-
fying my choice to concentrate on the latter strategies for the
remainder of my study.
The following three chapters constitute the empirical part of
this study. I proceed by developing a series of propositions
which I first test statistically (chapter 4), and then follow up
with two longitudinal case studies (chapters 5 and 6). This
mixed method design is akin to a “nested analysis” (cf. Lie-
berman 2005; Rohlfing 2008), although I do not claim to
follow the model rigorously. My strategy is to start with strict
Introduction 13
assumptions derived from previous works, and then, based
on the results of the statistical test, use case studies to add
complexity and derive additional insights. The case studies
therefore have both a hypothesis-testing and a hypothesis-
generating aspect. They provide additional empirical mate-
rial to test prior assumptions against observed reality, and
their results serve to further refine the theory (cf. Levy 2008:
5).
Chapter 4 concentrates on the Commission’s use of in-
fringement proceedings as a policy tool. I develop assump-
tions about the Commission’s default strategy in policy-
making and formulate a series of hypotheses about the fac-
tors that are likely to influence the Commission’s subsequent
strategy choice. In particular, I propose that the Commission
is more likely to resort to infringement proceedings when the
likelihood of initiating successful legislation is low. This is
the case where voting rules in the Council stipulate unanim-
ity, and where the European Parliament has veto powers.
Obstacles to legislation should also rise, the larger the ideo-
logical difference between the Commission and the Council
and the Commission and the EP, respectively, and in situa-
tions where there is considerable disagreement within the
Council. I also include a control variable that measures the
degree of ‘market orientation’ of a policy area. This allows
me, to some degree, to respond to the frequently voiced
proposition that judicial politics in the EU has a distinctly
market-making bias (cf. Scharpf 2006: 854). It moreover
enables me to draw some, albeit limited, conclusions about
possible internal differences in strategy choice between
Commission portfolios generally held to pursue a form of
interventionist ‘regulatory capitalism’ and more market ori-
ented portfolios (cf. Hartlapp and Lorenz 2012: 11). I then
proceed to test these propositions employing a dataset I have
constructed from data based on three different sources. One
is a dataset on legislative procedures, compiled by Thomas
König, Brooke Luetgert and Tanja Dannwolf (König, Luetgert
Introduction 14
et al. 2006). I combine this dataset with information on in-
fringement proceedings referred to the Court by the Com-
mission, based on data compiled by Alec Stone Sweet and
Thomas Brunell (Stone Sweet and Brunell 2007). Addition-
ally, I use data on the ideological positions of the Commis-
sion, the Council and the European Parliament, as compiled
by Andreas Warntjen, Simon Hix and Christophe Crombez
(Warntjen, Hix et al. 2008). I designed my dataset so that the
dependent variable is the ratio of litigation to legislation in a
particular policy area across time. The dearth of available
data over longer time spans caused me to restrict this analy-
sis to the time period from 1984 to 1998.
Chapters 5 and 6 take up the results of the statistical test in
two longitudinal qualitative case studies, concentrating on
the Commission’s efforts to remove barriers to trade in goods
within the Community (chapter 5) and the Commission’s
support of the free movement of persons (chapter 6). These
case studies expand the scope of the analysis to include the
Commission’s intervention in preliminary reference proce-
dures. Covering a broad time period, I trace the process (cf.
George and Bennett 2005: 205) of Commission policy initia-
tives in the two fields from the early 1960s to (roughly) the
present day. From this series of “causal process observa-
tions” (Collier, Brady et al. 2010: 184) – as a form of “theo-
retically oriented narrative” (cf. George and Bennett 2005:
205) – I attempt to uncover ‘causal mechanisms’ that link
legislative barriers to the Commission’s preferred policies to
an increased involvement of the Commission in judicial pro-
cedures. In this manner, I am able to “examine empirically
the alternative causal mechanisms associated with observed
patterns of covariation” (Levy 2008: 11).
I chose the two ‘cases’ (free movement of goods and free
movement of persons) with a number of objectives in mind.
First, it is never particularly easy to define what constitutes a
distinct ‘case’ in policy studies. Isolating particular events
from their context in the policy cycle necessarily increases
Introduction 15
the possibility of selection bias. Focusing on single legislative
or judicial events, such as a particular piece of legislation or
court case, usually provides incentives to select those that
stand out in some way or the other, as most such events (leg-
islation or litigation) are uneventful. The primary reason why
some events stand out is the presence of conflict, and focus-
ing on those cases alone allows for no variance on one of my
central explanatory variables. Primarily for this reason I
chose a long time span for my studies, as this would enable
me to observe a whole sequence of events and introduce
large within-case variance on the side of the independent
variables I identified in chapter 4 (cf. Levy 2008: 10). Legis-
lative procedures change over time, and so does the ideologi-
cal distance between the actors involved, of which conflict is
usually the result. Also, observing long time periods, particu-
larly from an early point in the development of Community
policies, allows me to identify potential learning effects. Con-
versely, the long time span prohibits a greater attention to
detail. Second, I chose to conduct two case studies, covering
two distinct policy fields in order to add cross-case variation
to the longitudinal comparison (cf. Levy 2008: 10). Both
cases differ with regard to Commission’s mandate and the
degree of conflict. After the Single European Act, moreover,
the policy fields are governed by different legislative proce-
dures – trade in goods being subject to majority voting in the
Council, whereas unanimity remained in place for many
pieces of legislation concerning the free movement of per-
sons. Both cases also show a different affinity to market inte-
gration. Whereas the removal of barriers to trade in goods is
a distinctly market-making enterprise, the Commission’s ef-
forts concerning the free movement of persons follows a dif-
ferent trajectory. While, certainly, motivated initially by a
better allocation of the factors of production, establishing a
right to free movement for all member state nationals was
informed by the objective of giving effect to the emerging
principle of Union citizenship, a principle that is not congru-
Introduction 16
ent with market-making activity and the logics of market
integration alone. In this way, I will be able to test the causal
linkage between judicial activity and the degree of market-
orientation of a policy field, adding onto the results of the
statistical test. Finally, both cases represent the core of the
Commission’s activity. The free movement of goods and the
free movement of persons form part of the fundamental
‘market freedoms’ which the internal market is based on. As
such, they are ‘crucial cases’, constituting a sizable portion of
the Commission’s everyday activities for the period observed.
If I want to test my assumptions about how I typically expect
the Commission to behave, these are the subject areas I
should look to.
Chapter 7, the final chapter, discusses the results of the case
studies with regard to the hypotheses developed in chapter 4.
It outlines some general considerations regarding the role of
the preliminary reference procedure for the Commission’s
choice of strategy and argues for an amendment to the origi-
nal propositions. I conclude with some thoughts about the
relevance of the wider legal environment for the results of my
study.
Introduction 17
Chapter 2
Studying the European Commission
This chapter has two objectives. The first part (2.1) outlines
theoretical approaches to the study of the Commission. I
sketch the positions of the original opposing camps of neo-
functionalism and intergovernmentalism and trace their
varying prominence over time. I also include a summary of
the main characteristics of the ‘institutionalist’ turn in the
study of European Union politics, and conclude by locating
my own position among the various approaches. The second
part (2.2) looks more closely at theoretical assumptions
about what motivates Commission action, and describes
more recent empirical studies about internal preference dis-
tributions and decision-making dynamics within the Com-
mission. These new approaches cast some doubt on the pre-
viously rarely questioned assumption that the Commission
constitutes a unitary actor. I summarise the main findings of
this research and conclude by outlining my own assumptions
about the ‘actorness’ of the Commission and the types of
preferences it pursues.
2.1 The European Commission in the study of Euro-
pean Union politics
The plethora of names and attributes that have been assigned
to the Commission – ‘honest broker’, ‘engine of integration’,
or ‘guardian of the treaties’, to name some of the most fre-
quently used – indicates the difficulty of neatly identifying its
central function in the political system of the European Un-
ion. Simultaneously its executive as well as its public admini-
stration, the Commission engages in all phases of the EU pol-
icy cycle – from initiation via negotiation and implementa-
tion to monitoring (cf. Lindberg 1963: 71). As Neill Nugent
puts it: “The Commission is centrally involved in EU
decision-making at all levels and on all fronts” (Nugent 2010:
105), or, to use the title of one of his books on the subject, the
Commission finds itself “at the heart of the Union” (Nugent
1997b).
At the same time, its actual means of autonomously asserting
influence are rather limited. Its central decision-making
body, the College of Commissioners, is appointed by member
state governments, who at least theoretically have every in-
centive to nominate candidates they believe will not stray too
far from their interests. Its actions, particularly in the proc-
ess of implementing EU policies, are subject to close over-
sight procedures on the part of national governments. Most
implementation, moreover, is not done in Brussels, but car-
ried out by member state administrations, who have fairly
wide leeway to shirk from their duties. They are able to do
this not least because the Commission does not have the re-
sources to monitor implementation ‘on the ground’, but re-
lies instead on member state sources for information and
manpower. Always in danger of being stretched thin by its
multitude of obligations, “if the Commission is to play a role
it must make creative use of the resources it has for influenc-
ing behavior of the governments” (Lindberg and Scheingold
1970: 93). What, then, is the impact of Commission actions
on European Union policy?
The European Commission’s ability to influence policy out-
comes in the European Union has traditionally formed one of
the central points of contention in theories of European Un-
ion politics. How much does the Commission matter? Does it
have an independent effect on policy outcomes (cf. Pollack
2003: 4)? The classical ‘grand theories’ of European integra-
tion reached opposite answers to this question. On the one
hand, ‘neo-functionalism’, as promulgated by Ernst Haas,
Studying the Commission 19
Leon Lindberg and others, centrally underscored the role of
‘supranational’ institutions, such as the Commission, in ad-
vancing the process of European integration. Integration,
Haas argued, could be measured by reference to the preva-
lence of certain types of compromise in international nego-
tiations, of which he identified three. The first of these con-
stitutes an agreement on the basis of a ‘minimum common
denominator’, where the degree of integration never pro-
ceeds past the reservations of the least cooperative bargain-
ing partner (Haas 1961: 367). Second, agreements by “split-
ting the difference” allow for an outcome that lies somewhere
between the final bargaining positions of the partners, aided
by the “mediatory services of a Secretariat-General or an ad
hoc international expert study group” (Haas 1961: 367, origi-
nal emphasis). The third type of compromise, finally, where
agreement would “upgrade the common interest“ of the par-
ties, “takes us closest to the peaceful change procedures typi-
cal of a political community with its full legislative and judi-
cial jurisdictions, lacking in international relations” (Haas
1961: 368). Parties would succeed in “redefining their con-
flict”, maximising the ‘spill-over’ effect of international deci-
sions. This solution, however, would only be achieved
through the competent arbitration by an “institutionalized
mediator” (Haas 1961: 368) such as the Commission (cf. also
Tranholm-Mikkelsen 1991: 6). Leon Lindberg took up this
line of thought, arguing that such a solution “depends on the
participation of institutions or individuals with an autono-
mous role that permits them to participate in actually defin-
ing the terms of the agreement” (Lindberg 1963: 12). He fur-
ther suggested that “the central institutions of the EEC, by
isolating issues and identifying common interests, may play a
crucial role here in ‘precipitating unity’” (Lindberg 1963: 8).
As an arbiter of the ‘common interest’, as opposed to narrow
national interests, he saw the Commission in a central posi-
tion to facilitate compromise and encourage further integra-
tion: “The Commission enjoys some unique advantages by
Studying the Commission 20
virtue of its ability to embody the authority of a Community
consensus” (Lindberg 1963: 284). “Moreover, it has proved
far easier for Member States to give in to the Commission
than it would have been for the Germans to give in to the
French or vice versa; in other words, in justifying their ac-
tions, both to themselves and to their respective govern-
ments, Ministers have been able to defend major concessions
on the ground that they were made in the interest of the
Community” (Lindberg 1963: 286). Precisely because the
member states relied on the Commission to enable consen-
sus was it so difficult for them to resist the proposals made
by the Commission, the argument went. The Commission’s
independent influence on the outcome of policy-bargains was
therefore seen as an essential condition for political integra-
tion per se.
While the original neo-functionalists saw the European
Commission essentially as an “honest broker” (Lindberg and
Scheingold 1970: 94), they nonetheless acknowledged that it
would simultaneously strive to expand its own position of
authority. More than just facilitating agreements, the Com-
mission was expected to become a central agent of spill-
over7: “the initial task and grant of power to the central insti-
tutions creates a situation or series of situations that can be
dealt with only by further expanding the task and the grant of
power” (Lindberg 1963: 10). The Commission “has vigorously
defended its own role as spokesman for Community inter-
ests, and has sought to expand this role in its specific pro-
posals” (Lindberg 1963: 284). Although he was skeptical
about the High Authority’s ambition to lead the process of
integration in the 1950’s, Haas nonetheless asserted that
“spill-over can be accelerated in the face of divisions of opin-
ion among the governments and in the absence of an articu-
late consensus toward unity as an end in itself. All that is
Studying the Commission 21
7 Jeppe Tranholm-Mikkelsen later introduced the term ‘cultivated spill-over’
to describe Commission involvement in the process of European integration
(Tranholm-Mikkelsen 1991: 6).
needed is the effective demonstration by a resourceful su-
pranational executive that the ends already agreed upon
cannot be attained without further united steps” (Haas 1958:
483-484, original emphasis).
Stanley Hoffmann’s ‘Intergovernmentalism’ on the other
hand denied an independent causal role of supranational
institutions and emphasised the dominance of national gov-
ernment preferences in determining the outcomes of Euro-
pean politics. His “logic of diversity [...] sets limits to the de-
gree to which the ‘spill-over’ process can limit the freedom of
action of the governments” (Hoffmann 1966: 882). European
integration would be restricted to areas of ‘low politics’, such
as economic and social policies (the "area of welfare", Hoff-
mann 1966: 882), but political integration (in the sense of
the ‘high politics’ of foreign affairs and defense) would re-
main elusive due to the fundamental differences in member
state interests with regard to the global situation – the Cold
War, American influence in Europe and the threat of Soviet
invasion: “In areas of key importance to the national interest,
nations prefer the certainty, or the self-controlled uncer-
tainty, of national self reliance” (Hoffmann 1966: 882).
While economic and social integration also bears the poten-
tial to limit government’s leeway in conducting foreign pol-
icy, Hoffmann predicted that member states would resist
such tendencies even if this produced short term costs for
certain parts of their constituencies (Hoffmann 1966: 884).
The European Commission, in this view, is powerless in the
face of national government opposition: Facing the govern-
ments of France and Germany, “the supranational civil ser-
vants, for all their skill and legal powers, are a bit like
Jonases trying to turn whales into jellyfish” (Hoffmann 1966:
884). Summing up his position, Hoffmann stated that “If we
look at the institutions of the Common Market as an incipi-
ent political system for Europe, we find that its authority re-
Studying the Commission 22
mains limited, its structure weak, its popular base restricted
and distant” (Hoffmann 1966: 885).
Both of these approaches have over the years been signifi-
cantly refined and expanded, but the central disagreement as
to the influence of European institutions remained (cf. Eil-
strup Sangiovanni 2006: 3). The period following the publi-
cation of Hoffmann’s article, however, seemed to vindicate
his viewpoint. Academic interest in the development of the
European Communities and the actions of its institutions
waned in the 1970s and early 1980s, a period that despite
some institutional and policy innovation was generally char-
acterised as a European “dark ages”, or time of “Eurosclero-
sis” (Keohane and Hoffmann 1991a: 8). The region’s eco-
nomic and political stagnation and its failure to adequately
respond to the oil crisis of 1973 led Ernst Haas to speculate
about the apparent obsolescence of regional integration the-
ory, admitting that “Regional integration in Western Europe
has disappointed everybody: there is no federation, the
nation-state behaves as if it were both obstinate and obso-
lete, and what once appeared to be a distinctive ‘suprana-
tional’ style now looks more like a huge international bu-
reaucratic appendage to an intergovernmental conference in
permanent session” (Haas 1975: 6). As late as 1982, political
developments gave Stanley Hoffmann cause to repeat and
reinforce his argument about the ineffectiveness of EU insti-
tutions and the prevalence of nation states as the decisive (if
embattled) units in the international system, asserting that
the EEC’s “‘Federal’ institutions have weak powers, and their
main organs often paralyze one another; these troubles, in
turn, have depressing effects on the other components of the
emerging or expected ‘central’ political system – community-
wide interest groups or party alignments” (Hoffmann 1982:
31). They are weak, he says, “because they lack autonomy
(from the member states) and because their capacity to act is
small” (Hoffmann 1982: 32). One year later, Paul Taylor
Studying the Commission 23
wrote of the limits of European integration, underscoring
that the “member states have not been absorbed into a new
Euro-federation, nor have they become the vassals of a su-
pranational Commission” (Taylor 1983: 60). He argued that
in the course of the 1970s, the establishment of the Council
presidency and the European Council as central institutions
of the Community had significantly reduced the powers of
the Commission to exert an independent influence on Euro-
pean policy. Moreover, “As the level of tension between gov-
ernments in the Council and their committees increased and
the expectations of governments of getting what they wanted
declined, so it became increasingly difficult and, indeed, fu-
tile for the Commission to spell out detailed European solu-
tions” (Taylor 1983: 82). He concluded that “There seemed
little prospect in the early Eighties of any strengthening of
the Communities’ institutions, and if anything, it seemed
likely that they would be further weakened: the Communities
were in a period of entrenched intergovernmentalism” (Tay-
lor 1983: 83). Finally, writing in 1984, Stephen George, while
shying away from broader predictions, asserted that “the
prospects for the creation of more common policies are not
good” (George 1985: vii). He, too, doubted the ability of the
Commission to assume a strong leadership role, pointing to
its declining room for maneuver through the establishment
of the Committee of Permanent Representatives (CORE-
PER), the Management Committee procedure (what was to
develop into the system of ‘Comitology’), the inception of the
European Council, and an increasing ‘bureaucratization’ of
the Commission, reducing its sense of identity and commit-
ment (George 1985: 12-14).
The in some ways unexpected signing of the Single European
Act (SEA), however, challenged these assumptions and pre-
cipitated a profoundly renewed interest in European institu-
tions. Academically, the apparent success of the SEA and the
Single European Market programme led to a spate of publi-
Studying the Commission 24
cations about the dynamism of the “new” Community (cf.
Taylor 1989; Keohane and Hoffmann 1991b; Tranholm-
Mikkelsen 1991; Sbragia 1992) and heralded a revival of at-
tention to ‘supranationalism’ and central tenets of neofunc-
tionalism. By contrast, the fact that no scientific theory had
predicted this institutional change led some observers to look
for other theoretical approaches to explain this development:
“In view of our failure to predict developments using older
theories, perhaps a new interpretation of joint European de-
cisonmaking should be invented, discarding loaded terms
such as ‘supranationalism’ and ‘spillover’, and drawing in-
stead on contemporary theories of strategic choice in collec-
tive situations, or recent attempts to understand institutional
innovation” (Keohane and Hoffmann 1991a: 9).
The new proponents of a revived neofunctionalism engaged
in detailed studies of the progress of the Single European
Market programme and frequently concluded that the Com-
mission must be seen as a driving factor bringing about pro-
found policy change that could not be explained by reference
to intergovernmental bargaining alone (cf. Tallberg 2006:
196). This line of work has produced a wealth of case studies
outlining the influence of Commission ‘policy entrepreneur-
ship’ on developments in individual policy areas. In one of
the first studies of this kind, Wayne Sandholtz and John
Zysman concluded that the leadership of the Commission
was decisive in bringing about the Single European Market
programme. In their words, “the renewed drive for market
unification can be explained only if theory takes into account
the policy leadership of the Commission“ (Sandholtz and
Zysman 1989: 96). In a similar vein, Laura Cram described
the Commission as a “purposeful opportunist” in bringing
about social policy legislation (Cram 1993: 143), while Sonia
Mazey highlighted the entrepreneurial role of the Commis-
sion in fostering policy networks in support of equality of
opportunity for men and women at a time of strong govern-
Studying the Commission 25
ment opposition (Mazey 1995). Other studies reached similar
conclusions about the efficacy of Commission action in re-
gional policy (Hooghe and Keating 1994: 387-388) and IT
policy (Cram 1994: 201-207).
The renewed interest in the Commission also brought with it
a number of studies focusing on the internal functioning of
the Commission and its Directorates General, attitudes of its
staff and administrative culture (cf. Ludlow 1991; Cram 1994;
Cini 1996; Nugent 1997a). Often without direct reference to
neofunctionalism, their optimistic view of the Commission’s
position in the policy-making process leads Jonas Tallberg to
characterise them as “closet neofunctionalists” (Tallberg
2006: 197).
At the same time, the renewed importance of European poli-
tics attracted a range of scholars with a background in com-
parative politics. While these approaches have not gone en-
tirely beyond the classical supranationalist-
intergovernmentalist divide, their origin in the study of ‘tra-
ditional’ political systems led them to focus on the European
Union as a polity, producing a distinct “allocation of values”
(Easton 1953: 129), rather than as a process of political inte-
gration (cf. Rosamond 2006: 14). This focus allowed research
to concentrate on the factors influencing policy output, and
less on factors influencing the design of the European insti-
tutional framework. Prominent among these approaches
have been those that adopt a theoretical framework drawn
from rational choice institutionalist accounts of legislative,
executive and judicial politics. Originating mainly from the
study of congressional politics in the United States, rational
choice institutionalism posited that the institutional frame-
work of political action has a distinct effect on political out-
comes. The rational choice part of rational choice institu-
tionalism connotes a set of prior assumptions: that a) collec-
tive outcomes can be explained by references to the actions
of individuals (‘methodological individualism’); and that b)
Studying the Commission 26
individuals are self-interested and seek to maximise their
preference for certain states of the world (‘utility maximisia-
tion’) (cf. Pollack 2006: 32). The institutionalism part posits
that individual actors face constraints in their attempts to
achieve outcomes – not only due to multiple actors trying to
achieve conflicting ends, but also by the need to adhere to
certain formal or informal norms – and that taking account
of these constraints (the “rules of the game”, North 1990: 3)
would be necessary to explain outcomes (cf. Pollack
2006:32-33). In a nutshell, rational choice institutionalism
not only postulated that ‘institutions matter’, but provided a
set of hypotheses about how these institutions would matter
in relation to policy outcomes. Similar accounts of the effects
of, say, the European Commission’s exclusive right of initia-
tive on the outcome of bargaining between member state
governments have certainly existed previously, but rational
choice institutionalism provided a coherent theoretical
framework for the analysis of the institutional interplay. Spe-
cifically, these efforts allowed to specify the conditions under
which different institutions wield different forms of influ-
ence.
For the case of the Commission, research within the rational
choice institutionalist framework has primarily introduced
two approaches to the question of how much influence the
Commission has over policy production in the European Un-
ion. The first, originating in the literature on legislative poli-
tics, highlights the function of the Commission as an
‘agenda-setter’ (cf. Romer and Rosenthal 1978): by virtue of
its monopoly on legislative initiatives the Commission has
the opportunity to ‘set’ the legislative agenda (cf. Pollack
2006: 38). In theory, this position would allow the Commis-
sion to emphasise certain policy fields while withholding leg-
islation from others – in practice, this process is much more
cooperative and the Commission is sensitive to requests from
the European Council, the Council and the European Parlia-
Studying the Commission 27
ment. More importantly, when formulating an initiative, the
Commission can take account of possible ‘minimum winning
coalitions’ within the Council and the European Parliament
and move the outcome close to its own policy preferences
within the ‘solution space’. Its power to do so depends pri-
marily on the legislative procedure determining majority re-
quirements and possible ‘veto players’ (cf. Tsebelis 2002). I
will give a more detailed account of the Commission’s agenda
setting powers in chapter 3.
The second prominent approach to the study of the Commis-
sion within the rational choice institutionalist framework is
the so called ‘principal-agent’ model. Largely derived from
organisational theory and the theory of the firm (cf. Coase
1937), it posits that when exchanges between actors under
market conditions are no longer economical due to high
transaction costs, actors will prefer forms of hierarchical in-
tegration (cf. Williamson 2000: 602-604). Within the hierar-
chy, the ‘principal’ (the owner, shareholder, or sovereign)
assigns competences to an ‘agent’ who will conduct tasks on
her or his behalf. This ‘delegation contract’ delimits the
boundaries of agent competences and defines certain
mechanisms of oversight. The principal-agent model derives
its explanatory value from its analysis of possible conflict.
Since all actors have potentially divergent preferences (‘util-
ity functions’), it is likely that principal and agent do not
share goals. The agent has incentives to use his discretion to
‘shirk’ from his assigned duties and pursue his own prefer-
ences instead. He is able to do this on account of his informa-
tional advantage vis-à-vis the principal: While the principal
will want to exercise narrow oversight, the act of delegation
ceases to be functional if the agent can no longer pursue his
or her tasks independent of the principal. If the act of delega-
tion is based on concerns about ‘credible commitment’,
commitment to an agreement among the principals would no
longer be credible if the agent cannot independently police
implementation. If delegation is a result of concerns about
Studying the Commission 28
efficiency, the costs of policing may rise above costs saved by
the act of delegation in the first place, in which case the prin-
cipal might as well exercise the tasks her or himself (cf. Tall-
berg 2002: 25-29). Delegation contracts therefore bear the
possibility of ‘agency drift’, where agents succeed in pursuing
preferences that differ from that of the principal. In essence,
this term encapsulates the notion of ‘supranational entrepre-
neurship’ that is central to most neo-functionalists notions of
a “cultivated spillover” (Tranholm-Mikkelsen 1991: 6).
Agents such as the Commission can use their delegated pow-
ers to pursue their own policy preferences or achieve an even
greater accumulation of competences at the European Union
level (cf. Tallberg 2002: 34-35). Since member state govern-
ments as the ‘high contracting parties’ are usually treated as
a collective principal, intergovernmentalists on the other
hand point to the diverse oversight mechanisms at their dis-
posal which may effectively prohibit such agency drift (cf.
Garrett 1992: 552). Principal-agent analysis is therefore not
predisposed to any particular answer to the question of who
ultimately controls outcomes (cf. Tallberg 2002: 34). Such
studies of European politics typically analyse the amount of
discretion for, e.g., independent Commission action con-
tained in the delegation contract – the European Treaties
and relevant acts of secondary legislation – and predict
which institutional constellation is more conducive to agency
drift than others (cf. Pollack 2006: 39-40).
While rational choice institutionalism is commonly regarded
as the dominant institutionalist approach in the study of the
Commission as the European executive (cf. Tallberg 2006:
198), some authors prefer to base such research on a more
‘sociological’ account of institutional constraints on human
behaviour. As an alternative institutionalist approach, socio-
logical institutionalism (most closely associated with James
G. March and Johan P. Olsen) shares with rational choice
institutionalism the assumption that ‘institutions matter’ in
Studying the Commission 29
shaping human interaction. Where it deviates from rational
choice institutionalism is in its underlying assumptions
about what motivates human behaviour. In particular, it fa-
vours the concept of a ‘logic of appropriateness’ as the moti-
vating factor for human action rather than the ‘logic of con-
sequentiality’, which describes the goal seeking (or utility
maximising) behaviour assumed by rational choice institu-
tionalism. According to March and Olsen, “the processes of
reasoning are not primarily connected to the anticipation of
future consequences as they are in most contemporary con-
ceptions of rationality” (March and Olsen 2006: 690). The
logic of appropriateness stresses that humans adhere to rules
not because they see them strategically as a means to an end,
but rather because they hold them to be ‘good’ or ‘right’ in a
normative sense: “Rules are followed because they are seen
as natural, rightful, expected, and legitimate. Actors seek to
fulfill the obligations encapsulated in a role, an identity, a
membership in a political community or group, and the
ethos, practices, and expectations of its institutions. Embed-
ded in a social collectivity, they do what they see as appropri-
ate for themselves in a specific type of situation” (March and
Olsen 2006: 689).
For the most part, this conception has been interpreted as a
challenge to rational choice institutionalism (cf. Tallberg
2006: 200). As outlined above, sociological institutionalists
do reject the rationalist assumption of goal seeking as the
exclusive motivation for human behaviour. However, neither
do they claim that the logic of appropriateness is the only
logic of human action (cf. March and Olsen 2006: 702), but
stress that it is a dimension of human motivation that is si-
multaneously important and scientifically overlooked. Two
approaches are conceivable to reconcile the two positions.
One useful solution would be to understand considerations
of appropriateness as constraints on human action, while
maintaining the assumption that human behaviour is essen-
tially goal driven. This is to assume a hierarchy between dif-
Studying the Commission 30
ferent logics of action, where the logic of consequentiality
operates at a more basic level of human motivation than the
logic of appropriateness (cf. March and Olsen 2006: 703). In
other words, goal seeking is constrained by norms that may
include internalised, ‘culturally learned’, ideas about appro-
priate behaviour. March and Olsen suggest a more ‘egalitar-
ian’ approach. They differentiate logics of action according to
their “prescriptive clarity”, where “a clear logic will domi-
nate a less clear logic” (March and Olsen 2006: 703, original
emphasis). Codes of appropriate behaviour, as well as indi-
vidual preferences and strategic alternatives may be more or
less apparent and therefore constitute more or less stringent
behavioural motivations. The expectation is that when indi-
vidual preferences for outcomes are weak or the alternative
courses of action not well understood, rule following is likely
more prevalent. Conversely, where rules of appropriate be-
haviour are not clearly defined or widely shared, goal seeking
may become the dominant behavioural logic. The observable
implications of both solutions should be quite similar. In this
fashion, strict assumptions of goal-seeking rationality can be
relaxed without jettisoning their analytical merits.
I locate my study in the institutionalist tradition. In chapter 3
I will draw more heavily on an ‘agenda-setting’ approach,
and my analysis is broadly compatible with the principal
agent vocabulary, although I do not explicitly use the con-
cept. I start with no preconceived disposition toward ‘supra-
nationalism’ or ‘intergovernmentalism’ – although their fun-
damental imprint on the study of European politics makes it
close to impossible to steer clear of this debate. Also, my
choice of topic predisposes me to assume that the Commis-
sion is worthwhile investigating; I start from the assumption
that the Commission is not entirely irrelevant. The task is
less to prove who is right or wrong about the role of the
European Commission in European integration, but rather to
determine the conditions under which the Commission can
Studying the Commission 31
wield influence over policy outcomes, and the factors which
determine its strategy to do so (cf. Schmidt 1998: 23).
2.2 Who is ‘the Commission’ and what does it want?
The Commission is a large institution employing about
33,000 staff, mainly based in Brussels, headed by 27 Com-
missioners who are supported by Cabinets comprising six or
seven close aides, and lead an administration subdivided into
more than 40 Directorates General (DGs) and across-the
board services. Each Commissioner holds a distinct policy
portfolio, but these portfolios do not necessarily match up
neatly with the corresponding DGs (cf. Chalmers, Davies et
al. 2010: 57). Speaking of actions by ‘the Commission’ there-
fore constitutes a crude simplification. Such a reference can
mean anything from the organisation as a whole (as in ‘the
Commission is based in Brussels’), to a decision by the Col-
lege of Commissioners, the Commission’s central decision-
making body (as in ‘the Commission decided to initiate in-
fringement proceedings against Germany for its insistence
on beer purity laws’), or an action by a single Directorate
General (as in, ‘the Commission is investigating the merger
between Honeywell and General Electric’) (cf. Cram 1994:
198; Cini 1996: 101; Nugent 1997a: 1). This linguistic prob-
lem is exacerbated by the ‘principle of collegiality’ (cf.
Chalmers, Davies et al. 2010: 55), which holds the Commis-
sion collectively responsible for all decisions and requires all
Commission members to publicly support decisions even
where they have been outvoted.
The described linguistic simplification harbours the danger
of glossing over significant variables explaining Commission
action. In particular, in treating the Commission as a single,
coherent actor, it directs attention away from possible con-
flict within the Commission. Until quite recently, very little
empirical material was available about decision-making
Studying the Commission 32
processes within the Commission hierarchy.8 For a long time,
literature on the European Commission has held, somewhat
unquestioningly, that the Commission as a single entity pur-
sues objectives that are distinctly different from those of the
member states. More recently, a new body of research has
put this long standing assumption into question and invested
considerable effort in unveiling internal preference distribu-
tions and decision-making dynamics. In the following over-
view of the literature on Commission preferences, I will start
at a general level of abstraction, and then proceed to cover
newer insights into where such preferences come from and
how they are aggregated.
The traditional view: What motivates the Commission?
If I start from the assumption that the Commission possesses
‘actorness’, what does this actor strive for? There are two ref-
erence points to answering this question: one is based on the
literature on European integration, which has always held
very specific assumptions about the preferences of the Com-
mission in the integration process; the other starts from the
observation that the Commission could be fruitfully concep-
tualised as a form of government bureaucracy and therefore
looks to general theories of bureaucratic politics to derive
assumptions about its preferences.
To start with the latter, public choice models of bureaucracy
typically assume public officials to be self-interested utility-
maximisers (cf. Hix 2005: 28). While there are multiple con-
ceptions about what exactly bureaucrats maximise, a number
of such objectives are frequently invoked. Broadly speaking,
such approaches assume that bureaucracies are “constantly
seeking to increase their size, staffs, financing, or scope of
operations” (Dunleavy 1991: 147). Among these objectives,
William Niskanen has prominently proposed ‘budget maxi-
Studying the Commission 33
8 An early exception is Laura Cram’s account of the Commission as a “Multi-
Organization” (Cram 1994).
mising’ to be prevalent (Niskanen 1971). In this view, larger
budgets are key in allowing bureaus to secure their survival,
expand staff numbers, increase salaries, or generally gain
prestige (cf. Dunleavy 1991: 155; Hix and Høyland 2010: 24).
Competition for limited available resources lead public offi-
cials to overstate their institution’s budgetary needs and
oversupply agency output (by spending beyond immediate
necessity) to underscore the inadequacy of their budget and
secure future payment (cf. Hix and Høyland 2010: 24; Peters
2010: 13-14). Other accounts of bureaucratic behaviour de-
emphasise the centrality of budget maximising and instead
highlight bureaucrats’ interest in policy influence and the
absence of direct supervision, especially among more senior
officials: “Rather than maximizing budgets, then, senior bu-
reaucrats (particularly in regulatory agencies) will seek to
maximise their independence from control and their oppor-
tunities to determine policy outcomes” (Hix 2005: 28). Even
when the strict assumptions of utility maximisation are re-
laxed it is commonly accepted that bureaucrats pursue policy
oriented interests, “having both expert knowledge and some
interest in the expansion of their agencies“ (Peters 2010:
199). The extent to which a bureaucracy engages in policy
advocacy is thought to be a function of its relative independ-
ence from government control (cf. Peters 2010: 199). Ex-
pressed in the terminology of the principal-agent approach,
all these views share the assumption that bureaucracies as
agents face incentives to pursue aims that do not correspond
to that of the principal(s).
Looking at the European integration literature, the Commis-
sion has traditionally –almost unquestioningly – been held
to favour policies that represent a ‘European’ perspective and
increase the competences of EU institutions (cf. Tallberg
2002: 44-45; Hooghe 2012b: 87-88). This view may be in-
formed by the Treaty mandate, which stipulates that “The
Commission shall promote the general interest of the Union
Studying the Commission 34
and take appropriate initiatives to that end” (article 17(1)
TEU). Variants of this view range from the comparatively
tame characterisation as the “conscience of the Community”
(Cini 1996: 16), safeguarding collective EU interests against
self-interested national governments, to the more forceful
proclamation that “The members of the Commission are firm
advocates of a maximum economic and political integration,
as well as of the principle of delegating national powers to
the Community institutions” (Lindberg 1963: 67). In the
manner of the latter, Wayne Sandholtz and John Zysman, in
their 1989 article that revived a neofunctionalist line of ar-
gumentation, plainly summarise that “The Commission itself
is an entrenched, self-interested advocate of further integra-
tion [...]” (Sandholtz and Zysman 1989: 108).
These accounts of ‘integrationist’ preferences of the Commis-
sion do not conflict with the assumptions proposed by the
public choice literature on bureaucratic politics. To support
an ongoing transfer of competences to European institutions
in most circumstances means increased Commission compe-
tences, greater policy autonomy and less supervision. Al-
though a view of Commission officials as single-minded
budget maximisers will hardly find corroborating empirical
evidence, increasing the workload of the Commission could
in turn form the basis for demands for greater staff numbers
and greater financial resources.
A more nuanced picture: recent empirical research
Since the ‘second wave’ of interest in the European Commis-
sion there has been a growing number of empirical studies
about the preferences of individual Commission officials that
provide a much more nuanced picture of what ‘the Commis-
sion’ as a whole may strive for (cf. e.g. Hooghe 2000; Hooghe
2005; Egeberg 2006; Wonka 2008; Kassim, Peterson et al.
2013). In particular, these studies demonstrate an internal
Studying the Commission 35
heterogeneity of motivations that underscore the difficulty of
assigning distinct actorness to a complex organisation. The
most recent and most ambitious of these efforts to date,
“Commission in Question (EUCIQ)”, undertakes a large scale
attitudinal survey of Commission officials (n=1901) (Kassim,
Peterson et al. 2013). The project finds that while the com-
monly assumed ‘supranationalist’ preference (embracing the
policy advocacy of the Commission) is indeed prevalent,
large minorities of the respondents adhere to role concep-
tions closer to ‘intergovernmentalist’ attitudes (downplaying
the policy advocacy of the Commission and promoting defer-
ence to the Council) and something that Liesbet Hooghe
terms an “institutional pragmatism”, presenting a ‘third way’
“which conceives the Commission and Member States as in-
terlocking and complementary institutions” (Hooghe 2012b:
91).
The distribution of these preferences is not random across
Commission officials. The EUCIQ project identifies a number
of patterns. For example, ‘supranationalist’ preferences are
more prevalent among senior officials and those coming
from ‘older’ member states (although the two often overlap),
in particular from states that are either smaller, catholic, fed-
erally organised, or inefficiently governed (cf. Hooghe 2012a:
9-10). Preferences also cluster across DGs, but less pro-
nouncedly so. Liesbet Hooghe reports that ‘institutional
pragmatists’ are more prevalent in DGs “with technical con-
tent” (Hooghe 2012a: 11), although her distinction between
‘technical’ and ‘political’ DGs may be questionable.9 Moreo-
ver, personal characteristics such as prior experience in na-
tional administrations (which favours ‘intergovernmentalist’
attitudes), the stated motivation for joining the Commission
(commitment to ‘Europe’ unsurprisingly favouring a ‘supra-
nationalist’ attitude), and gender (women being less likely to
Studying the Commission 36
9 Hooghe classifies Environment, Development and Fisheries as technical
policy areas without justifying this choice.
express ‘supranationalist’ attitudes) have a statistically sig-
nificant impact on preferences (cf. Hooghe 2012a: 11-12).
Apart from the supranational-intergovernmental dimension,
EUCIQ also includes questions about Commission officials’
political positions on a left-right scale. Here, nationality
seems to have a significant impact. Officials from the ‘new’
member states in particular tend to be more ‘pro-market’ in
an economic sense and ‘conservative’ in a socio-cultural di-
mension (cf. Hooghe 2012a: 15). Hooghe also finds clusters
across DGs, with officials with more ‘interventionist’ portfo-
lios expressing less market-friendly attitudes than officials in
market-oriented DGs: “On economic ideology, DG location is
a surer predictor than nationality” (Hooghe 2012a: 18).
Whether this is due to self-selection or socialisation is impos-
sible to tell from the data (cf. Hooghe 2012a: 15).
Finally, the two preference dimensions (supranational-
intergovernmental and left-right) combined accurately pre-
dict Commission officials’ preferred allocation of compe-
tences between the EU and the member states, differentiated
by policy area. This finding, to some extend, casts doubt on
accounts of the Commission as a ‘bureau-maximiser’:
“Commission officials’ attitudes on policy scope in general,
and on the kind of policies that should be centralized are
guided by ideology and EU governance views rather than by
career interests” (Hooghe 2012a: 21).
While EUCIQ provides a very detailed snapshot of the Com-
mission at the time of the survey in 2008, Mariam Hartlapp
and Yann Lorenz have compiled a longitudinal dataset com-
prising information on nationality, party-political and pro-
fessional background on all Commissioners and Directors
General from the first EEC Commission to the Barroso
Commission that completed its term in 2010 (Hartlapp and
Lorenz 2012). In a first descriptive overview of the data,
Hartlapp and Lorenz discern patterns of nationality and
party-political affiliation in relation to particular policy port-
Studying the Commission 37
folios that corroborate some of EUCIQ’s findings. In par-
ticular, they, too, find portfolios relating to more ‘interven-
tionist’ policies to be dominated by officials with social-
democratic party backgrounds, while liberal and conservative
officials tend to cluster in market-related portfolios. They
also indicate that individual member states at times express
preferences for certain portfolios, although such preferences
have no basis in the relevance of these portfolios at the na-
tional level (cf. Hartlapp and Lorenz 2012: 32). However,
Hartlapp and Lorenz do not offer propositions as to how
these patterns may be connected to Commission action.
Overall, they find no periods of party-political dominance
that could be linked to particular policy initiatives (cf. Hart-
lapp and Lorenz 2012: 31). Moreover, a longitudinal analysis
of the varying influence of individual portfolios on Commis-
sion action is hindered by the fact that both content and
number of these portfolios have significantly changed over
time (cf. Hartlapp and Lorenz 2012: 17).
Apart from these empirical analyses of preference distribu-
tions, there is now also a growing body of empirical work on
internal dynamics in Commission decision-making, focusing
on potential conflict among the top tier of Commission offi-
cials. Arndt Wonka, for example, focuses on political dynam-
ics within the College of Commissioners as the Commission’s
central decision-making body. He proposes four possible ra-
tionales for individual Commissioners’ behaviour: a “national
party” scenario, where the Commissioner acts according to
the preferences of her or his national party, which can in turn
be influenced by considerations of domestic party competi-
tion; a “transnational party” scenario, where the Commis-
sioner is informed by broader ideological positions reflected
in the preferences of trans-national party families; a “na-
tional agent” scenario, where the Commissioner acts accord-
ing to cross-party domestic interests of his home country;
and finally a “portfolio” scenario, in which the Commissioner
Studying the Commission 38
seeks to advance the policies delegated to him (cf. Wonka
2008: 1148-1151). Although Wonka admits that some of the
proposed scenarios might be difficult to distinguish empiri-
cally, the evidence he finds in two case studies concerning
the “REACH” and “Takeover” directives points to a domi-
nance of the “national agent” and the “portfolio” scenarios
(cf. Wonka 2008: 1158). He finds no evidence for partisan
dynamics. This insight is consistent with the results of Mor-
ton Egeberg’s 2006 study of Commissioners’ decision behav-
iour in a range of policy types, as reported by officials located
in the Commission’s Secretariat General, who sat in on de-
bates in the College. Egeberg finds that in both sectoral poli-
cies and budgetary matters the “portfolio role” is the most
frequently reported behavioural pattern. The “country role”,
his equivalent to Wonka’s “national agent” scenario, is also
reported, but less frequently so, while partisan affiliation is
very rarely invoked (cf. Egeberg 2006: 11).
Robert Thomson conducts a further test on the importance
of officials’ personal characteristics in Commission decision-
making. His 2008 study connects Commissioners’ national-
ity and party affiliation with their legislative behaviour, in-
vestigating more closely the “the conditions under which the
commission’s positions on legislative proposals agree with
those of different member states” (Thomson 2008: 170). He
finds that the average distance between the policy position of
a member state and that of the Commission is smaller if the
Commissioner responsible for the legislative proposal is a
national of that member state – but only if the voting rule in
the Council allows for a qualitative majority. Under unanim-
ity, no such effect is evident. This finding remains robust re-
gardless of the Commissioner’s party political affiliation –
the responsible Commissioner being a member of the home
state’s governing party does not affect the proximity of that
member state’s position to that of the Commission. The re-
sults of Thomson’s study dovetail with previous insights in-
sofar as they emphasise that nationality has an influence on
Studying the Commission 39
decision-making in the Commission, and that party affilia-
tion is marginal (cf. Thomson 2008: 188). They also under-
score that voting rules in the Council have an effect on the
content of the Commission’s legislative proposal. This is
valuable evidence suggesting that the Commission is a stra-
tegic actor, taking account of the preferences of other actors
involved in the inter-institutional decision-making process
and the institutional rules of the game.
A central problem in analysing decision-making dynamics in
the College, much like the obstacles to the study of judicial
behaviour in the Court of Justice, is the dearth of empirical
material about the positions taken by individual Commis-
sioners, as neither debates or votes are recorded (cf. Wonka
2008: 1152). Egeberg speculates that with the growing size
of the College, there appears to be a greater proclivity to-
wards non-interference with other Commissioners’ projects,
and that “bilateral relations between the President and the
particularly affected commissioner(s) might come to partly
replace collective decision-making” (Egeberg 2012: 945).
This in turn lends credence to the implications of Thomson’s
study, in that the position of the responsible Commissioner
has an important influence on the Commission’s overall posi-
tion.
Miriam Hartlapp, Julia Metz and Christian Rauh provide
some further support for this supposition. Their analysis of
the administrative set-up of the Commission indicates that
procedural structures systematically privilege the lead de-
partment in the formulation of a legislative proposal (cf.
Hartlapp, Metz et al. 2010: 13). They also point to the in-
creasingly important role of the Secretariat General (under
the political leadership of the Commission president) in des-
ignating the lead department and coordinating legislative
proposals: “the SG [Secretariat General] is in a distinguished
position in setting policy priorities on the internal agenda
and influencing how much say a policy portfolio has in a cer-
Studying the Commission 40
tain legislative drafting” (Hartlapp, Metz et al. 2010: 14).
Based on these insights, they propose that “the overall legis-
lative output of the Commission may be skewed towards
those DGs that act more frequently as the lead department
than others” (Hartlapp, Metz et al. 2010: 17). Hartlapp, Metz
and Rauh moreover find evidence that there is significant
sectoral variance as to which proposals are subject to nego-
tiations in the College. The assumption is that proposals that
are negotiated in College meetings are less likely to represent
the lead department’s position than those decided at a previ-
ous stage (cf. Hartlapp, Metz et al. 2010: 22). However, they
do not test the implications of their findings against actual
legislative proposals.
All in all, this new research agenda has produced a rich body
of material on the factors influencing Commission policy po-
sitions. I would sum up the central points as follows:
Judging from EUCIQ’s extensive survey of Commission offi-
cials, it seems reasonable to assume a predominant tendency
within the Commission to support policy advocacy. 36.6% of
the respondents embraced a ‘supranationalist’ attitude that
largely conforms to a ‘traditional’ understanding of Commis-
sion priorities. Another 28.9% were labeled as “institutional
pragmatists” (Hooghe 2012b: 92). While this attitude is more
respectful of the interlocking authority of both Commission
and Council, it is nonetheless clearly in favour of policy im-
plementation over management. In fact, its prime distinction
is a clear rejection of sharing powers of initiative with the
European Parliament – a position that most ‘supranational-
ists’ support (Hooghe 2012b: 93-97). This indicates that
about two out of three Commission officials support Com-
mission policy advocacy. Coupled with the lack of evidence
for an across-the-board bureau-maximising attitude, it
seems reasonable to assume on this basis that Commission
officials on average pursue distinct policies. Policy prefer-
ences on an economic left-right scale moreover cluster with
Studying the Commission 41
portfolios. ‘Interventionist’ DGs lean towards the social-
democratic spectrum, whereas market-oriented DGs lean
towards the liberal/conservative side. Nationality seems to
have an influence on policy positions, too. Internal proce-
dures for policy formulation and decision-making, moreover,
appear to favour the positions of the lead department, as
long as proposals are not subject to intense negotiation in the
College. Finally, there is evidence that the Commission acts
as a strategic goal seeker in inter-institutional processes, as
Thomson’s study has shown the Commission’s position to
vary with voting rules in the Council.
These insights inform my own conception of the Commission
to a large degree. In chapter 4 I will develop a proposition
about the effects of portfolio characteristics on strategy
choice, which I will also test in my case studies (chapter 5
and 6). However, the described lack of readily available em-
pirical data on internal positions within the Commission pre-
cludes me from pursuing this question in greater detail. For
the most part, I will therefore maintain the simplifying as-
sumption of the Commission as a largely unitary actor, inso-
far as I assume that once a position has been internally
agreed upon, the Commission will pursue it employing the
most promising strategy available. Formally, whenever I
speak of Commission action as an aggregate, as the appro-
priate unit of analysis I am actually referring to the majority
of Commissioners taking a decision in the College.
The following chapter describes the Commission’s position in
inter-institutional policy-making processes and compares the
relative merits of its strategic options.
Studying the Commission 42
Chapter 3
The European Commission in the policy-
making process
In the previous chapter I outlined the theoretical approach to
my study of the Commission and my assumptions about its
‘actorness’ and preferences. In this chapter, I will look more
concretely at the position of the Commission in the political
system of the European Union. I will outline the central
characteristics of Commission policy-making strategies in
three policy modes. I argue that the outcome of these strate-
gies is functionally equivalent, but they place varying con-
straints on the Commission’s influence over outcomes. I will
conclude with a comparison of the Commission’s position in
various modes of policy-making and argue that the Commis-
sion’s ability to influence the policy-making process is broad-
est in legislation and litigation. The detailed comparison be-
tween the two strategies will allow me to formulate a series of
hypotheses about the factors influencing the Commission’s
choice of strategy in the following chapter.
3.1 Modes of policy-making
Assuming that the Commission pursues distinct policy pref-
erences strategically, its ability to assert an influence on pol-
icy outcomes is dependent on the institutional constraints of
policy-making in the EU. These constraints in turn depend
on the other actors present in the policy-making process, and
the legally prescribed decision-making procedures. In the
social science literature on EU politics, such distinct bundles
of procedures are described as ‘policy modes’ or ‘modes of
policy-making’.10 Research has identified a number of such
modes at the European level, each of which offers the Com-
mission varying degrees of influence over policy output, but
there is generally no agreement on any particular typology.
Helen Wallace, for example, proposes a typology consisting
of five policy modes. The first three, the “classical Commu-
nity method” (as employed in the original Common Agricul-
tural Policy), the “EU regulatory mode” (as employed for
most internal market regulation) and the “EU distributional
mode” (as employed in distributive policies such as cohesion,
structural funds, and research and development) all entail a
significant amount of delegation of policy-making compe-
tences to the Commission, but vary according to the depth of
these competences and the involvement of other actors, in
particular member state governments and the European Par-
liament, but also sub-national (or “infranational”) regional
actors in the case of cohesion policy. The other two policy
modes in this typology, “policy coordination” and “intensive
transgovernmentalism”, assign only a minor role to the
Commission, either in setting benchmarks and facilitating
dialogue and policy learning among national governments in
the coordinating mode, or as a relative bystander in “trans-
governmental” bargaining in foreign policy and earlier forms
of cooperation in monetary policy and justice and home af-
fairs (cf. Wallace 2010: 92-93).
Fritz Scharpf has suggested an alternative typology. Revisit-
ing his influential article about the European ‘joint decision
trap‘, Scharpf calls to attention three different modes of
policy-making that he sees as central to European Union
politics (Scharpf 2006).
1. The “intergovernmental mode” of policy-making lies at the
heart of Scharpf’s ‘joint decision trap’. This mode is preva-
lent in all those circumstances where decision-making
The policy-making process 44
10 Jonas Tallberg also uses the term ‘process of rule-creation’ (Tallberg
2000: 848). The term ‘mode of governance’ has largely been used in relation
to the so-called ‘new modes of governance’.
takes place between constituent governments deciding
unanimously on outcomes. (In this regard it is akin to
Wallace’s “intensive transgovernmentalism”.) The Euro-
pean Commission’s role in facilitating consensus in this
mode is strictly limited, as it lacks resources to “design and
pursue bargaining strategies” (Scharpf 1988: 255), primar-
ily against the European Council, but also against the
Council in areas like Justice and Home Affairs before the
Treaty of Lisbon or the Common Foreign and Security Pol-
icy. Scharpf’s ‘trap’ describes the situation where the con-
tinuation of existing common policies under changed con-
ditions leads to sub-obtimal outcomes, while the institu-
tional framework forecloses exit and the decision mode
precludes policy change as long as a single member prefers
the status quo (Scharpf 1988: 257). Lacking powerful
mechanisms of consensus formation, Scharpf predicts a
“systematic deterioration of the ‘goodness of fit’ between
public policy and the relevant policy environment”
(Scharpf 1988: 257).
2. The “joint decision-mode” of policy-making, according to
this typology, is employed in most legislative acts at the
EU level. It provides a central position for the Commission
in facilitating policy solutions through its monopoly on
legislative initiatives (Scharpf 2006: 849). This mode best
corresponds to the first three modes described by Wallace,
in particular to the form of co-decision that is now en-
shrined as the ‘Ordinary Legislative Procedure’. The fact
that the Commission has the sole authority to introduce a
legislative proposal (in other words, ‘set an agenda’)
greatly reduces the complexity of bargaining between 27
constituent governments and the European Parliament.
Provided there is a potential overlap of preferences among
the veto players, the Commission wields extensive power
over the eventual outcome as it can choose among the pos-
sible solutions that which is closest to its own preferences.
The policy-making process 45
3. The “supranational-hierarchical mode” of policy-making,
absent in Wallace’s typology, describes all those acts “in
which the Commission, the European Court of Justice or
the European Central Bank are able to exercise policy-
making functions without any involvement of politically
accountable actors in the Council or the European Parlia-
ment” (Scharpf 2006: 851). With regard to the Commis-
sion, this mode involves those areas of competition policy
where the Commission has been delegated competences to
issue autonomous legal acts, or situations in which the
Commission makes use of its function as ‘guardian of the
treaties’ to take a member state to court for an alleged in-
fringement of European law. This policy mode permits an
exit from the ‘joint-decision trap’ as a single actor (or a
small number of actors) can produce policy change with-
out becoming involved in complex bargaining procedures.
At the same time, policies produced in this mode are ex-
tremely difficult to alter (becoming ‘locked in’), as legisla-
tive override is exceedingly difficult to negotiate (cf.
Scharpf 2006: 852-3).
Commonality: legal norm as output
All in all, these different policy modes offer the Commission
different channels of influence over policy outcomes. Figure
3.1 outlines the various strategies through which the Com-
mission can influence the production of legal norms. This
overview is certainly not exhaustive of all modes of policy-
making the Commission is involved in. In particular, I do not
cover Commission involvement in so called ‘new modes of
governance’, essentially captured by Wallace’s “policy coor-
dination” mode, emphasising soft-law mechanisms of ‘nam-
ing and shaming’, information exchange and common learn-
ing (cf. Bartolini 2011: 5), often epitomised by the ‘Open
Method of Coordination’ as introduced by the ‘Lisbon Strat-
egy’ in 2000 and subject to a whole separate body of litera-
The policy-making process 46
ture. I essentially focus on such modes of policy-making that
result in legally enforceable norms. I find this focus justified,
as “Law is one of the central products of politics and the prize
over which many political struggles are waged” (Whittington,
Kelemen et al. 2008: 3). Law, moreover, is the principal
output of the European Union: “Law has always been a basic
instrument and a central symbol of European integration“
(Snyder 1993: 19).11
Figure 3.1 Commission strategies in three modes of policy-
making
The following section demonstrates in more detail that the
outcome of each of the depicted strategies is indeed a legal
norm. I will also defend the argument that no matter the
The policy-making process 47
11 It might be more accurate to refer to “law production” (cf. Tsebelis 1999),
rather than “policy-making”, but the term is not very widely used. Moreover,
“law-production” would include constitutional law-making, i.e. the process
of treaty revision, which I do not cover. I choose to focus solely on policy-
making, that is, that part of “law production” that has a distinct policy con-
tent.
concrete strategy, the outcome is functionally equivalent.
This is easier to demonstrate for legislative and executive
policy-making, but holds for judicial policy-making as well.
There is no need to explain that law is the central outcome of
legislation (which, after all, is Latin for ‘law-making’). But
legislation is only a subset of all possible acts resulting in le-
gal norms, and, in a formal sense, does not even constitute
the majority of all legal acts issued by Union institutions.12
The Treaty of Lisbon attempts to clarify the nature of differ-
ent legal acts by introducing a novel typology, intended as a
hierarchy of norms (cf. Craig and de Búrca 2011: 108). Art,
289(3) TFEU now defines legislation, or a ‘legislative act’, in
formal terms as a legal act adopted by a ‘legislative proce-
dure’ (be it the ordinary legislative procedure or a special
legislative procedure). All other acts are by definition ‘non-
legislative’. Secondary legal acts based on legislation will now
be termed either ‘delegated acts’ or ‘implementing acts’, sub-
ject to an ill-defined distinction that will be difficult to im-
plement in practice (cf. Craig and de Búrca 2011: 117). Such
secondary legal acts will be executive acts carried out by the
Commission (art. 290 and 291 TFEU).
Other EU agencies and bodies with executive tasks also issue
legal norms, but theses are not explicitly captured by the new
typology. The revised article 263 TFEU on judicial review of
EU acts explicitly includes acts of “bodies, offices or agencies
of the Union intended to produce legal effects vis-à-vis third
parties” as amenable to legal challenge before the Court of
Justice. The fourth paragraph of this article moreover intro-
duces the category of “regulatory acts” in regard to which the
rules of standing have been relaxed for natural and legal per-
sons. It is unclear what constitutes a regulatory act and what
its relation is to legislative, delegated and implementing acts
(cf. Craig and de Búrca 2011: 508). Other legal acts, such as
The policy-making process 48
12 Autonomous legal acts by the Commission, which will no longer be for-
mally referred to as legislation, consistently outnumber legislative acts by a
factor higher than 2:1 (cf. König, Dannwolf et al. 2012: 25).
those issued by the European Central Bank, or by the Council
in the context of foreign policy, also do not fit the new typol-
ogy, leading Paul Craig and Gráinne de Búrca to conclude
that “the hierarchy of legal acts composed of theses catego-
ries does not capture the totality of the ways in which legal
norms are made in the post-Lisbon world” (Craig and de
Búrca 2011: 118).
The legal nature of decisions taken by the European Council
is something of a grey area (cf. Werts 2008: 28-29). The Lis-
bon Treaty’s caveat that the European Council “shall not ex-
ercise legislative functions” (art. 15(1) TEU) does not pre-
clude it from issuing acts of legal relevance, or even legal acts
(since Treaty defines legislative acts as those legal acts
“adopted by a legislative procedure”). Many procedures with
legal effect require a European Council decision even where
the formal legal act is carried out by a different institution
(see for example the procedure to suspend the membership
of a member state in “serious and persistent breach” of Un-
ion values, art. 7 TEU). Its position as final arbiter of conflict
in the Treaties’ ‘emergency brake’ procedures moreover di-
rectly infuses the European Council into the legislative proc-
ess, even though it does not take a final decision on the draft
(cf. Chalmers, Davies et al. 2010: 78). A direct legal effect of
European Council acts is most visible in the area of ‘constitu-
tional’ politics, in particular the simplified procedure for
Treaty revision (cf. Chalmers, Davies et al. 2010: 76). While
European Council decisions “amending all or part of the pro-
visions of Part Three [internal policies] of the Treaty on the
Functioning of the European Union” are still subject to na-
tional ratification (art. 48(6) TEU), the so-called ‘passerelle’
clause allows the European Council to unilaterally amend
voting procedures from unanimity to qualitative majority,
and legislative procedures from special to ordinary, as long
as no national parliament objects (art. 48(7) TEU).
The policy-making process 49
The inclusion of the European Council in the list of Union
institutions (art. 13 TEU) entails the possibility of a judicial
review of European Council acts, and article 263(1) TFEU
quite explicitly grants powers of review to the European
Court of Justice where European Council acts are “intended
to produce legal effects vis-à-vis third parties”. In principle,
this could empower individuals to challenge parts of Presi-
dency Conclusions (cf. Chalmers, Davies et al. 2010: 414).
The same applies to a failure of the European Council to act
where such an obligation exists (art. 265 TFEU). While this
certainly constitutes an innovation, it is reasonable to as-
sume that the innovation lies not in the admission of the le-
gal nature of European Council acts, but the existence of le-
gal remedies.
While the outcome of litigation is obviously of legal nature,
arguing that judicial processes are functionally akin to legis-
lation is not as straightforward and may require more per-
suasion. Traditionally, the spheres of legislation and adjudi-
cation have been treated as analytically distinct. Lawyers in
particular have been reluctant to concede that judges ‘make
law’ rather than merely apply it. To the contrary, I argue that
the production of binding norms is central to both processes.
Research on ‘judicial politics’ in many contexts has high-
lighted the wider policy impact of court decisions, even as
they are seemingly addressed to the parties in court only (cf.
Chayes 1976: 1281). Much of the literature on ‘law and poli-
tics’ implicitly assumes the two to be more or less the same,
or similar processes to a common end. In this vein, Stuart
Scheingold proposes that there is “no bright line” between
law and politics – the two being inextricably linked to one
another (Scheingold 2008: 740). I will not have to go so far
as to speak of the law-politics distinction as “a kind of his-
torical curiosity”, as Scheingold does (Scheingold 2008:
740), or as Harold Spaeth, who describes the insistence on a
conceptual distinction between judging and policy-making as
The policy-making process 50
a “morass of legalistic doublespeak” (Spaeth 2008: 753). But,
disregarding this rhetorical barrage, I find the underlying
argument essentially sound and use it as important starting
point for my study. As Spaeth agues, “the assertion that judg-
ing is different from the free choices of congresspersons or
administrators is simply false” (Spaeth 2008: 753). All politi-
cal actors act under constraints imposed by law and concepts
of legality (Shapiro 2008: 770-771). What constitutes the
boundaries of this legality is frequently contested and is itself
decided through legally prescribed means. In this argument,
judges are in fact the least constrained actors as it is their
prerogative to define what is legal and what is not. Judges,
according to Spaeth, “cloak the reality of choice with the
rhetoric of analogical legal phraseology. [...] The ultimate
choice, then, becomes no more than the choice of words”
(Spaeth 2008: 753-754). While this again is strong rhetoric,
Martin Shapiro formulates a similar idea somewhat less
harshly: “[Courts] implement a lot of law, and in the course
of doing so, they make a lot of law” (Shapiro 2008: 769). Or,
in the more nuanced words of Stuart Scheingold, “The appli-
cation of rules is difficult to distinguish in practice from the
making of rules. Each exception can be viewed either as an
application or a deviation depending on one’s perspective”
(Scheingold 1974: 31). Judges face weak constraints on
decision-making by the requirement to reach jurispruden-
tially defensible decisions (legal reasoning can support a va-
riety of different outcomes – after all both sides to the court
case at hand couch their conflicting arguments in legal
terms). However, judges have been empirically shown to be
constrained by strategic concerns for the wider institutional
environment (cf. i.a. Epstein and Knight 1998) and the par-
ticular characteristics of the parties before them (cf. i.a. Gal-
anter 1974). Without launching unnecessarily deep into the
question of judicial behaviour, the main argument here is
that the actions of judges are conceptually difficult to differ-
entiate from policy-making. In more technical terms, I argue
The policy-making process 51
that the interpretation of a legal norm in court is in most
cases equivalent to a (quasi-legislative) recasting of said
norm with universal impact and future relevance as prece-
dent (cf. Dehousse and Weiler 1990: 246; Dehousse 1998:
72; Shapiro and Stone Sweet 2002: 90; Vanberg 2005: 184).
Having established the commonality of the described modes
of policy-making, I will proceed by providing a detailed over-
view of the variation in the Commission’s influence over out-
comes in the three modes. I exclude from this overview
budgetary procedures and the management of EU finances,
as well as the negotiation of international agreements. By
and large, these procedures are special cases of the joint-
decision mode that, while surely essential to the day-to-day
operations of the EU and highly policy-relevant, would
somewhat complicate a concise overview.
3.2 The Commission and the intergovernmental
mode
The intergovernmental mode of policy-making is prevalent in
those areas outside the internal market that have been
gradually added to the list of Union competences over time.
This has been most prominently so in foreign policy and jus-
tice and home affairs, although the latter has been increas-
ingly subject to the ‘Community method’ since the entry into
force of the Treaty of Amsterdam. Cooperation in these areas
had developed outside the formal Treaty framework – or, as
in the case of the Schengen agreements, outside the EU con-
text altogether – with initially very little formal Commission
involvement. It moreover applies to bargaining in the Euro-
pean Council, which is not confined to any particular policy
area.
Although formally without any legislative authority, the
European Council has successively assumed a central “pre-
The policy-making process 52
legislative function” (Wessels 2008: 165) in the policy-
making process, essentially acting as an important informal
agenda setter and the Union’s primary forum for crisis re-
sponse. Now included formally in the list of Union institu-
tions (art. 13 TEU), the Treaty assigns the European Council
the tasks of providing the Union “with the necessary impetus
for its development” and defining its “general political direc-
tion and priorities” (art. 15 TEU). Beyond the mere wording
of the legal text, the European Council represents the Union’s
“focus of authority at the highest political level” (Craig and de
Búrca 2011: 48), and access to this institution is therefore a
valuable political asset. The Commission is the only EU insti-
tution that is formally involved in European Council negotia-
tions, the president of the Commission being a (non-voting,
art. 235(1) TFEU) member. Apart from such participation in
negotiations, however, the Commission has very limited abil-
ity to influence European Council decisions. Without formal
veto-power, the other European Council members have little
incentive to accommodate its views. Its action in this regard
is primarily preparatory and rhetorical. European Council
meetings are formally prepared by the General Affairs Coun-
cil, “in liaison with the President of the European Council
and the Commission” (art. 16(6) TEU). More precisely, arti-
cle 3 of the European Council’s rules of procedure stipulate
that the president of the European Council, in cooperation
with the rotating Council presidency and the Commission
president, present a draft agenda to the General Affairs
Council four weeks before the scheduled European Council
meeting. This same group of persons prepare draft guidelines
for European Council conclusions, draft conclusions and
draft decisions of the European Council, “which shall be dis-
cussed in the General Affairs Council” (OJ 2009, No. L 315/
52). However, the provisional agenda, to be adopted by the
European Council itself at the beginning of its meeting, is
drawn up by the president of the European Council alone (OJ
2009, No. L 315/53). In any formal sense, the Commission
The policy-making process 53
therefore holds very limited agenda setting powers over
European Council decisions. Nonetheless, Paul Craig and
Gráinne de Búrca hold that “The European Council has been
the institutional mechanism whereby the Commission can
secure broad agreement from Member States for major ini-
tiatives [...], and many European Council initiatives are the
result of Commission suggestions fed into the agenda pre-
pared by the GAC [General Affairs Council]” (Craig and de
Búrca 2011: 49).13 Given that the Commission is responsible
for the Union’s multi-annual legislative programming, it
faces strong incentives to match its priorities to those of the
European Council: “Winning the European Council’s ap-
proval for the general direction of policy in a particular area
facilitates the Commission’s task when fashioning more spe-
cific legislation to put that policy into effect” (Craig 2010:
107).
There is as yet preciously little research about the Commis-
sion’s ability to ‘enlist’ the European Council and to use the
support of the heads of state or government as a lever in sub-
sequent legislative negotiations with the Council. Isolated
accounts of such a link dot the literature on policy-making,
particularly with regard to the development of the single
market programme during the 1980s (cf. Armstrong and
Bulmer 1997: 18-19; Craig 2002: 37), but there are so far no
systematic studies on the factors influencing the success of
such a strategy.14 This question could present an interesting
framework for a test of competing rational choice and socio-
logical institutionalist approaches – the former proposing
congruence of preference or pressure from other venues as
explanatory variables for Commission success, the latter the
The policy-making process 54
13 Similarly, Paul Craig asserts elsewhere that “The Commission has fre-
quently fed policy initiatives that it wishes to advance to the European
Council, and gained its imprimatur” (Craig 2011b: 222).
14 The situation is somewhat different with regard to the Commission’s in-
fluence on Treaty change, although this discussion is still largely mired in a
dichotomous, either-or type of argument between intergovernmentalists and
supranationalists.
dynamics of negotiation, issue framing and rhetorical en-
trapment. It will also be interesting to see if the establish-
ment of the permanent president of the European Council
has an effect on the Commission’s ability to ‘upload’ policy
initiatives onto the European Council’s agenda. The fact that
he or she is appointed for a period of up to five years allows
for a greater planning perspective than previous presidents
of this institution and may serve to undermine the institu-
tional advantages previously held only by the Commission
(cf. Craig 2010: 105).
The Commission’s position in those areas of foreign policy
and justice and home affairs that still retain some of their
former (second and third) ‘pillar’ features is somewhat akin
to its relation vis-á-vis the European Council. The Treaty of
Lisbon has incorporated most parts of justice and home af-
fairs into the joint-decision mode of policy-making (with
slight exceptions in the case of criminal law and policy coop-
eration), but little has changed from the pre-Lisbon era with
regard to Commission competence in the policy-making fea-
tures of foreign and security policy. Both policy areas are
characterised by a plethora of preparatory committees, semi-
autonomous agencies and intergovernmental fora in- and
outside of the Treaty structure (cf. Giegerich and Wallace
2010: 441-444; Lavenex 2010: 466-468). While the Treaty of
Lisbon has formally abolished the former pillar structure,
article 24(1) TEU states that the Union’s Common Foreign
and Security Policy is governed by “specific rules and proce-
dures” that differ from other policy areas. The Commission’s
role in this area “is defined by the Treaties”, i.e. it has no
across-the-board competences and remains marginally in-
volved in decision-making procedures (cf. Craig and de
Búrca 2011: 327). The main decision-making bodies in this
field are the European Council and the Council, the former
determining objectives and general guidelines, while the lat-
ter defines and implements concrete decisions (art. 26 TEU).
The policy-making process 55
The Commission, through the Union’s High Representative
for Foreign Affairs and Security Policy, who is a vice presi-
dent of the Commission and replaces the former external
relations Commissioner, shares formal agenda setting pow-
ers with the member states (art. 31 TEU). Its main avenues of
influence, however, are through its access to the European
Council and through cross-issue linkages that tie foreign pol-
icy to policy areas such as trade and development policy, in
which the Commission has greater policy-making abilities
(cf. Giegerich and Wallace 2010: 442).
Prior to the Lisbon Treaty, this situation was similar in the
field of police and judicial cooperation in criminal matters,
that part of justice and home affairs that constituted the Un-
ion’s ‘third pillar’. Since December 2009, little remains of the
intergovernmental mode in this policy field, with the excep-
tion of the right of initiative in police and judicial coopera-
tion in criminal matters, where the Commission shares its
usual monopoly with “a quarter of the member states” (art.
76 TFEU). Apart from this deviation, however, the ordinary
legislative procedure (with ‘emergency brakes’) now applies
to much of this field as well.
Summary
The preceding overview has demonstrated the limited means
for the Commission to influence the content of policy set in
the intergovernmental mode. Although the relevant litera-
ture does indicate some instances of the Commission’s suc-
cess in using this policy mode to its advantage, there is little
to suggest that this is systematically so. Its formal agenda
setting power is weak and frequently shared with other ac-
tors, and little is known about its ability to use informal
means of influence.
The policy-making process 56
3.3 The Commission and the joint-decision mode
One of the central functions of the Commission in the politi-
cal system of the EU is initiating legislation. In the majority
of policy areas within EU competences (and all those rele-
vant to this study), the Commission has a monopoly on legis-
lative proposals (cf. Diedrichs and Wessels 2006: 222). For-
mally, a refusal of the Commission to issue a legislative pro-
posal would preclude the EU’s legislative bodies from becom-
ing active. In practice, the Commission typically responds to
other institutions’ requests for legislative proposals (cf. Die-
drichs and Wessels 2006: 221). Barring unexpected events or
crises, the Commission follows a medium term legislative
programme that it publishes at the beginning of its term of
office. Whereas these programmes are formulated at a highly
general level of abstraction, outlining broad policy priorities,
it issues more concrete Annual Policy Strategies specifying
the legislative agenda for the following year. This serves as a
basis for inter-institutional debate, as the result of which the
Commission publishes a detailed annual Work Programme,
listing specific decisions and legislative proposals the Com-
mission intends to pursue (cf. Craig and de Búrca 2011: 145).
The concrete origins of policy priorities and initiatives are
not always easy to identify. At all stages throughout the
preparation and formulation of policy initiatives the Com-
mission closely cooperates with other actors in the political
system of the EU, including both formal institutions like the
Parliament, the Council and the European Council, and eco-
nomic and social interest groups providing expertise and
seeking influence over the policy agenda. In many instances,
detailed requests for policy initiatives are formulated by the
European Council in its Presidency Conclusions, especially
when responding to unexpected events and crises. The Coun-
cil operates numerous expert groups exercising exploratory
functions in parallel to similar groups under Commission
supervision (cf. Diedrichs and Wessels 2006: 223). The
Treaty of Lisbon introduces the possibility of a citizen’s ini-
The policy-making process 57
tiative (art. 11(4) TEU), albeit with significant constraints.
Part of the legislative agenda is moreover made up by man-
datory revisions of existing legislation. In effect, Commission
data suggests that less than 10 percent of all legislative pro-
posals emanate from genuine Commission initiative (cf. Die-
drichs and Wessels 2006: 223). The reliability of such data is
unclear, as the Commission has strategic incentives to down-
play its entrepreneurial role in order to safeguard its percep-
tion as an ‘honest broker’. Nonetheless, it is obvious that pol-
icy initiatives originate in multiple fora and take account of a
multitude of interests. However, despite this multitude of
channels, the concrete formulation of specific legislative pro-
posals leaves plenty of space for inter-institutional contesta-
tion.
The formulation of an individual legislative proposal is typi-
cally assigned to the sectoral DG that is most closely aligned
with the matter at hand. The DG in turn designates a ‘rap-
porteur’ as lead author who formulates a draft, consulting,
where applicable, with other DGs if cross-cutting issues are
involved (and they often are). Whereas this horizontal coor-
dination of legislative proposals has long been practiced in a
rather haphazard fashion, recent reforms of the process
within the Commission have introduced mandatory cross-
sectional coordination in many areas, with the Secretariat
General in the enforcing position (cf. Hartlapp 2011: 187).
Other DGs then have the opportunity to demand changes to
the draft or to express their opposition to the proposal. All
proposals moreover have to pass the muster of the Commis-
sion’s Legal Service, which assesses each draft’s vulnerability
to legal challenge. Following this technical stage, the pro-
posal is passed upwards to the political level of the Commis-
sioners and their cabinets, where it proceeds through the
weekly meetings of the responsible members of the Cabinets,
the Heads of Cabinet, and, finally, the College of Commis-
sioners, which ultimately decides by a majority of its mem-
The policy-making process 58
bers. In practice, most decisions in the College are reached
by consensus. While the College has to formally endorse all
legislative proposals, agreement is in most cases reached at
an earlier stage in the process. Since 1994, only between 13
and 18 percent of proposals were discussed in the College (cf.
Hartlapp 2011: 187).
Besides mandating increased horizontal coordination be-
tween DGs, recent reforms have introduced a number of
measures to this process to increase the quality of Commis-
sion proposals as part of its ‘better regulation’ approach. An
important innovation has been the introduction of manda-
tory ‘impact assessment’, which outlines the likely impact of
the proposed piece of legislation with regard to environ-
mental, economic and social matters. As these procedures
are coordinated by the Secretariat General, the reforms have
reinforced its position as an important actor in the formula-
tion of legislative initiatives (cf. Hartlapp, Metz et al. 2010:
13-14).
Literature analyzing the Commission’s position in legislative
processes has focused on its power to set the agenda. While,
as I have shown above, many of the EU’s policy priorities
originate from other sources, the legislative process grants
the Commission significant room for maneuver in formulat-
ing the precise content of legislation. The Commission’s mo-
nopoly on legislative initiatives is crucial: “When the setter
has monopoly power, voters are forced to choose between the
setter's proposal and the status quo or fall-back position”
(Romer and Rosenthal 1978: 27-28). The extent of the Com-
mission’s ability to influence legislative outcomes is depend-
ent on two factors in particular: the distribution of prefer-
ences (vis-à-vis the status quo) among the actors involved,
and the legislative procedures prescribed by the Treaties. The
impact of these factors on the Commission’s agenda setting
position is best demonstrated by a starkly simplified model
of the legislative process (figure 3.2; cf. Tsebelis and Garrett
The policy-making process 59
2000: 15; Hix and Høyland 2010: 70). Consider a policy area
located on an economic left-right dimension (say, regulation
of chemicals, or the access of migrant workers to social secu-
rity benefits). The points on the scale represent the ideal
points for new legislation (assumed to be exogenously given)
of each of the actors involved. This model assumes an envi-
ronment of perfect information, meaning that all actors
know the preferences of the other actors involved as well as
the location of the status quo (cf. Romer and Rosenthal 1978:
28).15 The position of the Council is disaggregated into the
individual positions of its members (for ease of presentation,
reduced to seven members with equal voting rights: M1-M7).
SQ represents the status quo. All actors prefer new legisla-
tion to the status quo, but differ as to the extent of desired
reform.16 It is the sole prerogative of the Commission to in-
troduce a piece of legislation. It is in its interest to place it
strategically so that the outcome will be as close to its ideal
point as possible. Given the depicted constellation of prefer-
ences, its ability to do so depends on the legislative proce-
dure specifying veto players and voting thresholds.
In its earliest form, legislation was passed without substan-
tial involvement of the European Parliament. Procedural
rules merely required consultation, giving the EP nothing but
the power to delay legislation (and only so after the Euro-
pean Court of Justice decided that legislation would be inva-
lid unless the EP had given its opinion, Case 179/80, ECR
The policy-making process 60
15 This of course is a very unrealistic assumption. Under perfect information,
all actors would anticipate the other actors’ actions – legislative proposals
would always represent the single possible outcome of the process and
would be adopted at first reading.
16 Other demonstrations often represent the Commission and the European
Parliament as preference outliers, typically because the dominant policy
dimension is assumed to be ‘more integration’ or less. When looking at day-
to-day politics, however, more traditional left-right dimensions appear more
relevant and it is reasonable to assume that Commission and EP pursue
more centrist and potentially conflicting goals (cf. the ‘regulation scenario’
in Tsebelis and Garrett 2000: 30-31).
1982: 3623). If Treaty rules stipulate the Council to decide
unanimously, the Commission could expect the passage of a
proposal at point U, the most ‘progressive’ point (in relation
to the status quo) that the most ‘conservative’ veto player,
M1, would accept vis-à-vis the status quo (cf. Hix and Høy-
land 2010: 70).17
Figure 3.2 Agenda setting in EU legislation
If Treaty rules stipulate a vote by qualitative majority (here
simply defined as 5 out of 7), the Council still has the possi-
bility to alter the Commission proposal, but only unani-
mously so. In this case the Commission could expect the pas-
sage of a proposal at point Q, given that the pivotal member
of the Council for a qualitative majority vote, M3, would ac-
The policy-making process 61
17 Such spatial models of policy-making typically assume ‘circular’ (‘Euclid-
ean’) preferences, meaning that M1 is indifferent to which direction the
status quo deviates from its ideal point, as long as policy change does not
move it further away. This, also, is not always a realistic assumption.
cept any point between Q and U, which would be the out-
come of a unanimous vote (cf. Tsebelis and Garrett 2000:
18).
Things again change when the EP becomes more than just a
bystander in legislative proceedings. The first procedure to
allow for more EP involvement was the so-called cooperation
procedure which was introduced with the Single European
Act and used for most legislation relating to the Single Mar-
ket programme. Here, the Parliament still did not constitute
a veto player in the process, but could impose costs on the
Council when ignoring its opposition (George Tsebelis there-
fore talks of the EP as a "conditional agenda setter" under
cooperation, cf. Tsebelis 1994). Most importantly, if the EP
disagreed with the Commission proposal, the Council could
only override this opposition by a unanimous vote, depriving
the Commission of much of its agenda setting power. In or-
der not to trigger such EP opposition, the Commission could
then introduce a proposal at point CP, as this is the most
‘progressive’ point the EP (or rather, the median voter in the
EP) would accept over U, the default outcome of a unani-
mous vote.
The Single European Act also introduced the assent (now
referred to as consent) procedure, which was less frequently
used but elevated the Parliament to a veto player position.
The procedure is similar to the cooperation procedure except
that the EP now has the power to reject the Commission pro-
posal without the possibility of a Council override.18 Moreo-
ver, the Council cannot alter the proposal once adopted by
the Parliament (cf. Hix 2010: 70-71). Here, the Commission
could expect the passage of a proposal at point A, as this is
the most ‘progressive’ point the EP will accept over the status
quo that will also be supported by a qualitative majority in
the Council.
The policy-making process 62
18 In many instances, predominantly in areas outside the internal market,
the assent procedure does not give the Commission a monopoly of legislative
initiative, in which case the present model is not applicable.
The European Parliament moved towards becoming an equal
legislator with the introduction of the co-decision procedure
in the Treaty of Maastricht. Here, if disagreement between
the Council and the EP about a Commission proposal per-
sisted beyond the second reading, a conciliation committee
consisting of an equal number of members of the Council
and the Parliament would convene to reach a solution. Under
the Maastricht rules, in the event of no agreement in con-
ciliation, the Council could still unilaterally present, by quali-
tative majority, a final proposal that the EP could either
adopt or reject by an absolute majority (which, due to high
absenteeism in the Parliament, amounts to a significant su-
permajority, cf. Tsebelis and Garrett 2000: 29). In effect, at
this stage of the procedure the initial Commission proposal
had become irrelevant and the ‘endgame’ consisted of the
Council acting as agenda setter and the EP acting as veto
player (cf. Tsebelis and Garrett 2000: 23; Hix and Høyland
2010: 72). Stripped of its agenda setting power, a rational,
strategic and omniscient Commission could therefore only
expect a proposal to be accepted in the vicinity of M3, which
is the point that the Council as agenda setter, voting by quali-
tative majority, and the EP would agree on without Commis-
sion involvement.
The Treaty of Amsterdam revised the co-decision procedure
to eliminate the final stage, partly due to the fact that the EP
had adopted a strategy of systematically refusing all subse-
quent Council proposals if the conciliation committee had
not reached a common position, regardless of policy content
(cf. Tsebelis and Garrett 2000: 24). Both Council and EP
now vote on the text produced by the conciliation committee.
If the committee fails to produce a text, the legislation fails.
The Treaty of Lisbon has codified this procedure as the Un-
ion’s ‘Ordinary Legislative Procedure’ (art. 294 TFEU).19 Un-
The policy-making process 63
19 The Lisbon Treaty slightly but not insignificantly changed this procedure
by lowering the voting threshold in the Parliament for adopting the outcome
of the conciliation committee from an absolute to a simple majority.
der this procedure, the Commission can expect a proposal to
pass at a point somewhere between the pivotal member of
the Council (M3) and the EP, depending on the bargaining
dynamics between the EP and Council delegations. Since the
Commission is present as an arbitrator in the conciliation
process, it is reasonable to expect an outcome somewhere in
the centre (cf. Diedrichs and Wessels 2006: 227). While the
EP has now emerged as a truly co-equal legislator, and the
Conciliation Committee as the final agenda setter, the Com-
mission’s ability to influence policy content remains signifi-
cantly curtailed (cf. Tsebelis and Garrett 2000: 23-24).
The model obviously excludes large chunks of the empirical
reality of EU legislative decision-making in general, and by
choice does not capture the full complexity of agenda setting
in particular. Beyond formal powers to introduce a specific
piece of legislation, ‘political entrepreneurship’ involves a
much more informal rallying of support behind a particular
initiative – including questions of issue framing, focal points
and advocacy coalitions – that can be exercised by actors not
formally present in the decision-making process (cf. Pollack
2003: 50-51). When it comes to the Commission, the model
omits informal processes of consensus building that take
place outside the legally mandated procedure, such as the
trilogues between the Commission, the Council and the Par-
liament to secure an early adoption of legislative proposals
(cf. Diedrichs and Wessels 2006: 227; Chalmers, Davies et al.
2010: 108-109; Hix and Høyland 2010: 73), or the protracted
power-games and inter-institutional conflicts surrounding
the choice of legal basis for a legislative act (where, again, the
Commission enjoys certain agenda setting powers) (cf. Brad-
ley 2011). However, reducing empirical complexity is the
principal advantage of the agenda setter model, assuming
that all other factors do not vary systematically with the final
outcome. As such, it is able to demonstrate the systematic
constraints on the Commission’s influence on policy out-
The policy-making process 64
comes in legislative procedures. The simplified model pre-
dicts substantial variation in the Commission’s ability to ex-
ert such influence. In the constellation depicted above, the
outcome ranges from close to the status quo (U) to close to
the Commission’s ‘progressive’ preferences (Q). Naturally, all
of these predicted outcomes depend on the accuracy of the
constellation of actor preferences, which has to be empiri-
cally established.
Empirical applications have at least corroborated the under-
lying assumptions about the individual actors’ ability to in-
fluence the outcomes. Thus, the move from consultation to
cooperation and codecision has indeed led to a growing im-
portance of EP positions in EU legislation (cf. Steunenberg
and Selck 2006: 81). Simultaneously, studies have pointed to
a gradual weakening of the Commission’s influence (cf.
Shackleton 2000: 336; Tsebelis and Garrett 2000: 34; Pol-
lack 2003: 228; Young 2010: 60).
Summary
To summarise, it is important to keep in mind that “the
proximate causes of agreement and disagreement are specific
constellations of actor preferences” (Scharpf 2011: 223). If
there is no ‘win-set’ among the veto-players, the Commission
will not be able to propose successful legislation at all. If it
prefers the status quo to the possible outcomes of legislation,
it might attempt to stall the process even when it is asked to
produce a proposal. But while the potential outcome of legis-
lation changes with the distribution of ideal points among
the central actors, the degree of influence of the Commission
over the outcome remains constant within the respective
decision-making procedures. This capacity is, all else equal,
greatest in policy areas where majority voting applies, since it
is more costly for the Council to amend a Commission pro-
posal than to adopt it (cf. Lindberg 1963: 32; Schmidt 2000:
38). Rather than having to account for the preferences of the
The policy-making process 65
most status quo oriented members in the Council, the Com-
mission merely has to respect the preferences of the pivotal
members (cf. Tsebelis and Garrett 2001: 374). While the ex-
pansion of qualified majority voting thus increased the lee-
way for the Commission in initiating potentially successful
legislation, the concomitant expansion of the involvement of
the European Parliament in legislation has rendered this
procedure more complicated and curtailed its agenda setting
powers to some degree (cf. Tsebelis and Garrett 2001: 374).
The introduction of co-decision further amplified this trend
(Tsebelis, Jensen et al. 2001).
3.4 The Commission and the supranational-
hierarchical mode
The defining characteristic of the supranational-hierarchical
mode is the fact that policy can be made without the in-
volvement of the EU’s legislative bodies – the Council and
the European Parliament (cf. Scharpf 2006: 851). The Com-
mission has access to several strategies within this mode
which can be described as judicial and executive in character.
Among the judicial strategies, the Commission can initiate
infringement proceedings against a member state for failing
to conform with legal obligations,