Multi-Level Governance and European Integration

Full text

Governance in Europe
Gary Marks, Series Editor
The New Political Economy
of
EMU
edited by Jeffry Frieden, Daniel Gras, and Erik Jones
Regional Integration and Democracy: Expanding on the European Experience
edited by Jeffrey 1. Anderson
Contentious Europeans: Protest and Politics in an Integrating Europe
edited by Doug Imig and Sidney Tarrow
How to Democratize the European Union . . . and Why Bother?
by Philippe Schmitter
A Wider Europe: The Politics
of
European Union Enlargement
by Michael J. Baun
Transatlantic Governance in the Global Economy
edited by
Mark
A Pollack and Gregory C. Shaffer
Forthcoming
Wiring Europe: Reshaping the European Telecommunications Regime
by Giorgio Natalicchi
Between Global Economy
and
Local Society: Political Actors and Territorial
Governance
edited by Jeanie Bukowski, Simona Piattoni, and
Mark
E. Smyrl
Differential Europe: New Opportunities
and
Restrictions
for
Policymaking in the
Member States
by Adrienne Heritier, Dieter Kerwer, Christophe Knill, Dirk Lehmkul,
Michael Teutsch, and Anne-Cecile DouiIlet
Democracy beyond the State?: The European Dilemma and the Emerging Global
Order
edited by Michael Th. Greven and Louis W. Pauly
Europe and America: Partners and Rivals in International Relations, 3
rr1
Edition
by John Peterson
Policy and Process in the New Europe
by Jessica R. Adolino and Alan N. Katz
Politics as Usual in the European Union
by Mark E. Smyrl
,.---
Multi-Level
Governance and
European Integration
Liesbet Hooghe and Gary Marks
ROWMAN
&
LITTLEFIELD
PUBLISHERS,
INC.
Lanham • Boulder » New York • Oxford

lI"'"
ROW MAN & LITTLEFIELD PUBLISHERS, INC.
Published in the United States of America
by Rowman & Littlefield Publishers, Inc.
4720 Boston Way, Lanham, Maryland 20706
http://www.rowmanlittlefield.com
12 Hid's Copse Road
Cumnor Hill, Oxford OX2 9JJ, England
Copyright © 2001 by Rowman & Littlefield Publishers, Inc.
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of the publisher.
British Library Cataloguing in Publication Information Available
Library
of Congress Cataloging-in-Publication Data
Hooghe, Liesbet.
Multi-level governance and European integration / liesbet Hooghe and Gary Marks.
p.
cm.-(Governance
in Europe)
ISBN 0-7425-1019-0 (alk.
paper)-ISBN
0-7425-1020-4 (pbk. : alk. paper)
I. European Union. 2. Decentralization in
government-European
Union countries 3.
Regionalism-European
Union countries. I. Marks, Gary, 1952- II. Title. III. Series.
JN30 . H675 2000
320A'049'094--dc21
00.059224
§TM
The paper used in this publication meets the minimum requirements of American
National Standard for Information
Sciences-Permanence
of Paper for Printed Library
Materials, ANSIINISO Z39.48-1992.
List of Figures
List of Tables
Preface
Abbreviations
1 Multi-Level Governance in the European Union
Part
I: Sources of Multi-Level Governance
2 A Historical Perspective
3 Multiple Identities
4 Why National Leaders Diffuse Authority
,
Part
II: Multi-Level Governance with the Regions
5 Channels to Europe
6 Variations in Cohesion Policy
7 Cohesion Policy under Threat
Part
III: Contestation in a Multi-Level Polity
8 The Struggle over European Integration
Contents
vii
IX
xi
xv
33
51
69
81
93
105
119

pi
VI Contents
9 Supranationalism Contested in the Commission 143
10 Political Parties Take a Stand (coauthored with Carole J. Wilson) 163
Appendices 187
Bibliography 213
Index 241
About the Authors 251
Framing Territorial Identity
Territorial Attachment
Types of Decision Making
Dimensions of Contestation
The Social Democratic Possibility Curve
Social Democratic Parties
Liberal Parties
Christian Democratic Parties
The Neoliberal Possibility Curve
Conservative Parties
,
Figures
54
57
121
124
171
172
176
179
181
182
vii
3.1
3.2
8.1
8.2
10.1
10.2
10.3
lOA
10.5
10.6

XVI
1

ENA
EP
EPP
ERT
EU
FF
FPO
GATT
GDP
JHA
MERCOSUR
NAFTA
NATO
OECD
PASOK
QMV
RET!
RPR
SEA
TEU
UNICE
WTO
Abbreviations
EcoLe nationaLe d' administration (training school for French
civil servants)
European Parliament
European People's party (confederation of Christian Democra-
tic and some conservative parties)
European Round Table (lobbying group for multinational cor-
porations in Europe)
European Union (°1993)
Fianna Fail (Irish national-conservative party)
Freiheitliche Partei OSlerreichs (Austrian Freedom party)
General Agreement on Tariffs and Trade
Gross Domestic Product
Justice and Home Affairs (also called Pillar III)
Mercado Comun
deL
Sur, established in 1991 to create a com-
mon market among Brazil, Bolivia, Chile, Argentina, Paraguay,
and Uruguay
North American Free Trade Agreement, established in 1994
North Atlantic Treaty Organization
Organization for Economic Cooperation and Development
Greek socialist party
Qualified Majority Voting
Association of European Regions of Industrial Technology
RassembLement
pour
La
Republique (French Gaullist party)
Single European Act (1986)
Treaty of European Union (initially the formal name for the
Maastricht Treaty of 1993, and updated with the Amsterdam
Treaty of 1999)
Union of Industrial and Employers' Confederations of Europe
World Trade Organization
Multi-Level Governance in the European Union
Developments in the European Union (EU) over the last two decades have re-
vived debate about the consequences of European integration for the autonomy
and authority of the state in Europe.' The scope and depth of policy making at the
EU level have increased immensely. The European Union completed the internal
market on schedule in 1993, and eleven of the fifteen member states formed an
economic and monetary union (EMU) in 1999, with a European central bank and
a single currency, the euro. These policy-making reforms have been accompanied
by basic changes in European decision making. The Single European Act (1986),
which reduced nontariff barriers, also established qualified majority voting in the
Council of Ministers and significantly increased the power
of
the European Par-
liament. The Maastricht Treaty (1993) increased the scope
of
qualified majority
voting in the Council and introduced a codecision procedure giving the European
Parliament a veto on certain types of legislation. The Treaty of Amsterdam (1999)
extended codecision to most areas of policy making in the European Community,
except for HMU.
Our aim in this chapter is to take stock of these developments. What do they
mean for the political architecture of Europe? Do these developments consolidate
national states or do they weaken them? If they weaken them, what kind of po-
litical order is emerging? These are large and complex questions, and we do not
imagine that we can settle them once and for all. Our strategy is to pose two basic
alternative
conceptions-state-centric
governance and multi-level
governance-
as distinctly as possible and then evaluate their validity by examining the Euro-
pean policy process.
The core presumption of state-centric governance is that European integration
does not challenge the autonomy of national states. State-centrists contend that
state sovereignty is preserved or even strengthened through EU membership.
They argue that European integration is driven by bargains among national gov-
ernments. No government has to integrate more than it wishes because bargains

2 3
Chapter 1
rest on the lowest common denominator
of
the participating member states. In
this model, supranational actors exist to aid member states, to facilitate agree-
ments by providing information that would not otherwise be so readily available.
Policy outcomes reflect the interests and relative power
of
national governments.
Supranational actors exercise little independent effect.
An alternative view is that European integration is a polity-creating process in
which authority and policy-making influence are shared across multiple levels
of
government-subnational,
national, and supranational. While national govern-
ments are formidable participants in EU policy making, control has slipped away
from them to supranational institutions. States have lost some of their former au-
thoritative control over individuals in their respective territories. In short, the
locus
of
political control has changed. Individual state sovereignty is diluted in
the EU by collective decision making among national governments and by the au-
tonomous role
of
the European Parliament, the European Commission, the Euro-
pean Court
of
Justice, and the European Central Bank.
We make this argument in this chapter along two tracks. First, we analyze the
variety
of
conditions under which national governments will voluntarily or in-
voluntarily lose their grip on power. Second, we examine policy making in the
EU across its different stages, evaluating the validity
of
contending state-centric
and multi-level models
of
European governance.
TWO
MODElS OF THE
EUROPEAN
UNION
The models that we outline below are drawn from a large and diverse body
of
work on the European Union, though they are elaborated in different ways by dif-
ferent authors. Our aim here is not to replicate the ideas
of
any particular writer,
but to set out the basic elements that underlie contending views
of
the EU so that
we may evaluate their validity.
The core ideas of the state-centric model are put forward by several authors,
most
of
whom call themselves intergovernmentalists (Hoffmann 1966, 1982;
Taylor 1991, 1997; Moravcsik 1991, 1993, 1998; Garrett
1992,1995;
Milward
1992; for an intellectual history, see Caporaso and Keeler 1995; Caporaso 1998).2
This model poses states (or, more precisely, national governments) as ultimate de-
cision makers, devolving limited authority to supranational institutions to achieve
specific policy goals. Decision making in the EU is determined by bargaining
among national governments. To the extent that supranational institutions arise,
they serve the ultimate goals
of
national governments. The state-centric model
does not maintain that policy making is determined by national governments in
every detail, only that the overall direction
of
policy making is consistent with
state control. States may be well served by creating a judiciary, for example, that
allows them to enforce collective agreements, or a bureaucracy that implements
those agreements. but such institutions are not autonomous supranational agents.
,....
r

Multi-Level Governance
Rather, they have limited powers to achieve state-oriented collective goods (Keo-
hane 1984; Keohane and Hoffmann 1991).
EU decisions, according to the state-centric model, reflect the lowest common
denominator among national government positions. Although national govern-
ments decide jointly, they are not compelled to swallow policies they find unac-
ceptable because decision making on important issues operates on the basis
of
unanimity. This allows states to maintain individual as well as collective control
over outcomes. While some governments are not able to integrate as much as they
would wish, none is forced into deeper collaboration than it really wants.
State decision making in this model does not exist in a political vacuum. In this
respect, the state-centric model takes issue with realist conceptions
of
interna-
tional relations, which focus on relations among unitary state actors. National
governments are located in the domestic political arena, and their negotiating po-
sitions are influenced by domestic political interests.
But-and
this is an impor-
tant
assumption-those
arenas are discrete. That is to say, national decision mak-
ers respond to political pressures that are nested within each state. The fifteen
national governments bargaining in the European arena are complemented by fif-
teen separate national arenas that provide the sole channel for domestic political
interests at the European level. The core claim
of
the state-centric model is that
policy making in the EU is determined primarily by national governments con-
strained by political interests nested within autonomous national arenas."
One can envision several alternative models to this one. The one we present
here, which we describe as multi-level governance, is drawn from several sources
(Scharpf 1988, 1994, 1999; Marks 1992, 1993; Schmitter 1992, 1996a; Tarrow
2000; Sbragia 1992, 1993a; Hooghe 1995b, 1996c; Jachtenfuchs and Kohler- Koch
1995; Leibfried and Pierson 1995; Pierson 1996; Risse-Kappen 1996b; Borzel
1998; Tarrow 2000; see also Caporaso and Keeler 1995. or Caporaso I 996a for an
overview). Once again, our aim is not to reiterate
anyone
scholar's perspective,
but to elaborate essential elements of a model drawn from several strands
of
writ-
ing, which

the case that European integration has weakened the state.
The multi-level governance model does not reject the view that national gov-
ernments and national arenas are important, or that these remain the most impor-
tant pieces
of
the European puzzle. However, when one asserts that the state no
longer monopolizes European-level policy making or the aggregation
of
domes-
tic interests, a very different polity comes into focus. First, according to the multi-
level governance model, decision-making competencies are shared by actors at
different levels rather than monopolized by national governments. That is to say,
supranational
institutions-above
all. the European Parliament, the European
Commission, and the European
Court-have
independent influence in policy
making that cannot be derived from their role as agents
of
national executives.
National governments play an important role but, according to the multi-level
governance model, one must analyze the independent role
of
European-level ac-
tors to explain European policy making.

4 5
r
Chapter 1
Second, collective decision making among states involves a significant loss of
control for individual national governments. Lowest common denominator out-
comes are available only on a subset of EU decisions, mainly those concerning
the scope of integration. Decisions concerning rules to be enforced across the EU
(e.g., harmonizing regulation of product standards, labor conditions, etc.) have a
zero-sum character and necessarily involve gains or losses for individual states.
Third, political arenas are interconnected rather than nested. While national are-
nas remain important arenas for the formation of national government preferences,
the multi-level governance model rejects the view that subnational actors are
nested exclusively within them. Instead, subnational actors operate in both na-
tional and supranational arenas, creating transnational associations in the process.
National governments do not monopolize links between domestic and European
actors. In this perspective, complex interrelationships in domestic politics do not
stop at the national state but extend to the European level. The separation between
domestic and international politics, which lies at the heart of the state-centric
model, is rejected by the multi-level governance model. National governments are
an integral and powerful part of the EU, but they no longer provide the sole inter-
face between supranational and suhnational arenas, and they share, rather than mo-
nopolize, control over many activities that take place in their respective territories.
FROM
STATE-CENTRIC TO MULTI-LEVEL GOVERNANCE
Has national government control over EU decision making has been compro-
mised by European integration? In this section we argue that state sovereignty has
been diminished by restrictions on the ahility of individual governments to veto
EU decisions and by the erosion
of
collective government control through the
Council
of
Ministers.
limits
on Individual National Government Control
The most obvious constraint on the capacity of a national government to deter-
mine outcomes in the EU is the decision rule
of
qualified majority voting in the
Council of Ministers for a range of issues from the internal market to trade, re-
search policy, and the environment. In this respect, the European Union is clearly
different from international regimes, such as the UN or World Trade Organiza-
tion, in which majoritarian principles of decision making are confined to sym-
bolic issues.
State-centrists have sought to blunt the theoretical implications
of
collective
decision making in the Council
of
Ministers by making two arguments.
The first is that while national governments sacrifice some independent control
by participating in collective decision making, they more than compensate for
this by their increased ability to achieve the policy outcomes they want. Andrew
Multi-Level Governance
Moravcsik has argued that collective decision making actually enhances state
control hecause national governments will only agree to participate insofar as
"policy coordination increases their control over domestic policy outcomes, per-
mitting them to achieve goals that would not otherwise be possible" (1993, 485).
By participating in the European Union, national governments are able to provide
policy outcomes, such as a cleaner environment, higher levels
of
economic
growth, and so forth, that they could not provide on their own. But two entirely
different conceptions of power are involved here, and it would be well to keep
them separate.
On the one hand, power or political control may be conceptualized as control
over persons. A has power over B to the extent that she can get B to do something
he would not otherwise do (Dahl 1961). This is a zero-sum conception: if one
actor gains power, another loses it. By contrast, power conceived as the ability to
achieve desired outcomes entails power over nature in the broadest sense. Ac-
cording to this conception, I have power to the extent that I can do what I wish to
do. A government that can achieve its goal
of
low inflation and high economic
growth is, from this standpoint, more "powerful" than one that cannot."
The latter way
of
conceiving power is not "wrong," for concepts can be used
in any way one wishes to use them. But it does confuse two things that are sen-
sibly regarded as separate: who controls whom, and the ability
of
actors to
achieve their goals. We argue in chapter 4 that one reason why government lead-
ers shift authority away from the central state is precisely because this may en-
able them to achieve substantive policy goals.
A second line of argument adopted by state-centrists is that majoritarianism
in the Council
of
Ministers camouflages, rather than undermines, state sover-
eignty. They argue that treaty revisions and new policy initiatives remain sub-
ject to unanimity, and that the Luxembourg compromise gives national govern-
ments the power to veto any policy that contravenes their vital national interests.
UltimatelY"they emphasize, a national government could pull out
of
the EU if it
so wished.
However, the Luxembourg veto is available to national governments only
under limited conditions, and even then, it is a relatively blunt weapon. As we de-
tail below, the Luxembourg veto is restricted by the willingness of other national
governments to tolerate its use.
From the standpoint
of
physical force, member states retain ultimate sover-
eignty by virtue
of
their continuing monopoly
of
the means
of
legitimate coercion
within their respective territories.
If
a national government breaks its treaty com-
mitments and pulls out of the EU, the EU itself has no armed forces with which
to contest that decision.> In this respect, the contrast between the European Union
and a federal system, such as the United States, seems perfectly clear. In the last
analysis, national states retain ultimate coercive control over their populations.
But monopoly
of
legitimate coercion tells us less and less about the realities
of political, legal, and normative control in contemporary capitalist societies.

7
6 Chapter J
A Weberian approach, focusing on the extent to which states are able to monop-
olize legitimate coercion, appears more useful for conceptualizing the emer-
gence and consolidation of states from the twelfth century than for understand-
ing changes in state sovereignty from the second half of the twentieth century
(see chapter 2 for a comparison between state building and European integra-
tion). Although the EU does not possess supranational armed forces, a member
state is constrained by the economic and political
sanctions-and
consequent
political/economic
dislocation-that
it would almost certainly face if it revoked
its treaty commitments and pulled out of the European Union.
Limits on Collective National Government Control
We have argued that national governments do not exert individual control over
decision making in the Council of Ministers. State-centrists may counter that
states still retain collective control over EU decision making through the Council
of Ministers and the treaties.
In this section, we argue that neither the Council of Ministers nor the treaties
give national governments full control over EU decision making. The Council is
the most powerful institution in EU decision making, but it exists alongside a di-
rectly elected European Parliament (EP) that has a veto on legislation relating to
a third of all treaty provisions. The power of the EP in the European political
process has grown by leaps and bounds over the past twenty years, and collective
national control of decision making has declined as a result.
The treaties are the main expression of national authority in the process of Eu-
ropean integration. Because representatives of national governments are the only
legally recognized signatories of the treaties, one may argue that state authority
is enhanced in the process of treaty making.
If
a domestic group wishes to influ-
ence a clause of a formal EU treaty, it must adopt a state-centric strategy and
focus its pressure on its national government.
To evaluate treaties as a vehicle for national government control, one needs to
ask two questions: first, to what extent do national governments control the
process of treaty negotiation and ratification; and second, to what extent do
treaties determine European policy making.
National governments are the key actors in negotiating treaties, but since the
tumultuous reception of the Maastricht Treaty in 1993, they have had to contend
with the participation of many kinds of domestic actors. In Britain, opposition
and back-bench Members of Parliament almost derailed the Treaty in the House
of Commons. Just at a time when some observers were claiming that treaty mak-
ing was strengthening national governments at the expense of parliaments, events
in the United Kingdom were proving exactly the opposite. A Conservative gov-
ernment was held ransom by back-benchers, and a split developed within the
party on the issue of European integration that fatally weakened the government
during the remainder of its term and in the subsequent general election of 1997.
Multi-Level Governance
In Germany, ratification of the Maastricht Treaty mobilized German regional
governments who tried to block the Treaty in the constitutional court. In France,
ratification was fought out in a popular referendum in September 1992, and the
result was a hair's-breadth win for the government (51 percent in favor; 49 per-
cent opposed). In each of these countries, and across the EU, public opinion was
mobilized in ways that placed national governments on the defensive.
Tensions, and sometime outright splits, have arisen within major parties. The
British Conservative party is deeply divided on the question of European mone-
tary integration, as revealed in public squabbles and in a survey of MPs (Baker,
Gamble, Ludlam, and Seawright 1997). In France, the Gaullist party split into
two independent factions in the European election of 1999. In Germany, fissures
are evident within the Christian Democratic party, and between the Christian
Democrats and their Bavarian sister party, the Christian Social Union. These ten-
sions are not random, but can be explained systematically in reference to party
ideologies, as we set out to do in chapter 10.
So while it is true that national governments have a formal monopoly in mak-
ing treaties, it is not at all clear that treaty making, or the process of European in-
tegration in general, has strengthened national governments against parliaments,
regional governments, or public pressures.
To what extent do treaties allow national governments to determine institution
building? The treaties are the ultimate legal documents of the European Union,
so it may seem strange to pose the question. But a moment's thought suggests that
the question is worth asking after all. To what extent are American, French, or
German political institutions determined by their respective constitutions?
Treaties, like constitutions, are frameworks that constrain, but do not determine,
institutional outcomes. We would regard a study of American politics that fo-
cused exclusively on the development of the U.S. Constitution as strangely
skewed. Treaties, like constitutions, are sensibly regarded as points of departure,
not final destinations, in understanding the workings of a regime because they do
not capture the way in which actors adapt
to--and
exploit-formal
rules.
EU treaties have been reformed more frequently than most constitutions, and
they lie closer to the ground of policy making. However, national government
control is, to some extent, handcuffed by unanimity. Treaties have to surmount
the highest conceivable decisional barrier: unanimous agreement among the prin-
cipals. This not only makes innovation difficult but also makes it difficult for na-
tional governments to rein in institutions, as we discuss below.
The extent to which treaties constrain EU institutions is diminished because the
treaties themselves tend to be vaguely written. The treaty-making process is heav-
ily biased towards diffuse agreements that avoid contentious issues and allow
politicians from all countries and of all ideological stripes to claim success at the
bargaining table. The principals in treaty negotiations are not simply representa-
tives of national preferences but are flesh and blood politicians who have private
preferences that include a desire to perform well at the next general election

9
8 Chapter J
(a line of analysis we develop in chapter 4). In this respect, the principals sitting
around the European bargaining table, no matter how zero-sum their preferences,
have a col1ective desire to agree to something so that the negotiation itself is not
perceived as a failure. Ambiguity can serve rational political purposes. When in-
dividual or col1ective national control bumps up against electoral considerations,
we expect that electoral considerations will usually emerge the winner.
These considerations suggest that the control of national governments over
EU policies and institutions is highly imperfect. National governments no longer
monopolize EU decision making, partly because the European Parliament has
hecome a co-legislator over much EU policy; increased public scrutiny of EU
decision making increases the weight of public opinion on government policy;
and national governments have limited control over supranational agents, such
as the European Commission and the European Court of Justice. We examine
these in turn.
European
Parliament
The empowerment of an autonomous and directly elected Parliament over the
past two decades presents a fundamental problem for accounts that conceive au-
thority as delegated or pooled by member states.
It
seems forced to conceive the
Parliament as an agent designed by national governments to realize their prefer-
ences. The EP increases, rather than reduces, transaction costs of decision mak-
ing in the EU. The development of the EP cannot be explained as a functional re-
sponse on the part of national governments to problems of intergovernmental
bargaining. On the contrary, the EP is better explained in terms of the response of
national governments to domestic pressures for greater democratic accountability
in the European Union.
The EP does not fit wel1 into an intergovernmental account of European inte-
gration. One line of response is that while Parliament has been strengthened, this
has not been at the expense of other institutions. Andrew Moravcsik writes that
the cooperation procedure "increased the participation of the Parliament without
infringing on the formal powers of either the Commission or the Council" (1998,
366). As we have noted above, the notion that power can be supplied as a nor-
mal good, so some people can have a bit more without anyone having less, con-
fuses political power with ability. The rules comprising the basic governing in-
stitutions of the EU are interwoven, as we explain in the fol1owing section, so
that a reform of one involves change for all. In the case of the cooperation pro-
cedure adopted in 1986, the increase in the formal power of the European Par-
liament came mainly at the expense of the Council of Ministers. The EP could
offer amendments to certain legislation that could only be rejected by the Coun-
cil
if
it did so unanimously. Formal theorists of EU rules agree with almost
al1
participants and observers that this was an important increase in the EP's au-
thority (Crombez 1996; Steunenberg 1994; Tsebelis 1994). While there has been
Multi-Level Governance
disagreement about the relative consequences of cooperation and codecision
(adopted in 1993), it is plain that the combined effect of these reforms has been
to significantly strengthen the Parliament (for a discussion of this debate, see
Hix 1999b, 88-94).
It
is true that the European Parliament is elected nationally and can be con-
ceived as a forum "in which national representatives, generally organized in po-
litical parties, can influence the legislative process" (Moravcsik 1998, 67-68).
Members of the European Parliament, like those in the United States and most
other democracies, represent those living in particular territories, but they do not
represent the governments of those territories. Most members have interests and
ideologies that
mayor
may not lead them to preserve the authority of central gov-
ernments, and these preferences are usual1y consistent with the political party to
which they owe their election. Party membership is often a more powerful influ-
ence on parliamentarians' behavior and attitudes than country of origin (Hi x
1999a; Thomassen, Noury, and Voeten 2000; Thomassen and Schmitt 1999;
Schmitt and Thomassen 1999; Raunio 1998; Scul1y forthcomingj.?
The emergence of the European Parliament as a powerful European player has
altered the institutional balance in the European Union, as we argue in detail
below.' The authoritative competencies of the European Parliament are more nar-
rowly circumscribed than those of the Council, but the Parliament is nonetheless
a weighty player, As a result, national governments cannot impose their collcc-
tive will in many areas of policy making."
Public scrutiny
EU decision making has come under greater public scrutiny. Prior to the Sin-
gle European Act, European integration was essentially a technocratic process in
which national governments coordinated around limited policy goals. European
integration was pragmatical1y oriented, rather than politicized, and national gov-
ernments dohiinated decision making to the virtual exclusion of other domestic
actors. On the occasions when conflict did flare
up-usually
in the form of col-
lective protest by farmers, coal miners, or
steelworkers-national
governments
sought to buy off opposition through sectorial deals. EU bargaining was largely
insulated from public pressures.
This changed with the introduction of the single market in the mid-1980s (see
chapter 8). As the reach of European policy making broadened, and as the stakes
in most issue areas grew, so domestic groups were drawn directly into the Euro-
pean arena (Greenwood, Grote, and Ronit 1992; Fligstein and McNichol 1998).
Such mobilization has created new linkages between supranational institutions
and subnational groups, and it has induced citizens with similar interests or ide-
ological convictions to organize transnationally. EU decision making is no longer
insulated from the kind of political competition that has characterized democratic
politics in the member states.

Multi-Level Governance
II
Chapter 1
10
As noted earlier, one consequence of multiple contending principals in the EU
In the pre-Maastricht era, treaty ratification was dominated by national govern-
is that the treaties provide ample room for interpretation. The treaties are ham-
ments through party control of their national legislatures. Not only did they deter-
mered out in interstate negotiations, in which there is a powerful incentive to
mine the content of treaties but they could be reasonably confident that those treaties
allow ambiguity on points of contention so that each government can claim suc-
would be accepted in their respective domestic arenas." The Maastricht Treaty
cess in representing national interests.
changed all that. The rejection of the Maastricht Treaty in June 1992 by Danish cit-
izens sent a shock wave through European elites, and their anxiety was enhanced by
a near-replay in the French referendum of September 1992. Moreover, public opin-
ion polls indicated that German and British voters too might have rejected the Treaty
if they had been given the opportunity (Nugent 1999). The fact that the Danes re-
versed their decision a year later did not put to rest fears that the process was out of
control. Public scrutiny has changed the rules of the game of treaty negotiation. The
action has shifted from national governments and technocrats in semi-isolation to
domestic politics in the broad and usual sense: party programs, electoral competi-
tion, parliamentary debates and votes, public opinion polls. and public referenda.
10
Principal-agent dynamics
Even if national governments operated in a world without a European Parlia-
ment and without public pressures, it is likely that EU decisions would only im-
perfectly reflect the preferences of national governments. As governments have
agreed to collaborate on more and more issues in the EU arena, so they have
turned to supranational agents, particularly the European Commission and the I
European Court of Justice, to make collaboration work, and by so doing they risk
diluting their control over decision making.
Principal-agent theory builds on the insight that
principals-national
govern-
ments, in this
case-are
not able to plan for all possible future ambiguities and
sources of contention, and so they create
agents-such
as the European Com-
mission and the European Court of
Justice-to
ensure compliance to interstate
agreements and adapt them to changing circumstances (Keohane and Hoffmann
1991; Majone 1996; Pierson 1996; Pollack 1997). According to this line of theo-
rizing, principals exert control over agents by creating the necessary incentives
(Williamson 1985).
If
a principal discovers that an agent is not acting in the de-
sired way, the principal can fire the agent or change the incentives.
Scholars who have applied principal-agent theory to American political insti-
tutions have found that the incentives available to principals are often ineffective
(Moe 1990). There are grounds for believing that limits on principal control in the
EU are particularly severe.
Multiple principals. In the European Union there are as many principals as
there are member states. Each has a veto over basic institutional change. This
vastly complicates principal control. The more hands there are on the steering
wheel, the less control any driver will have. The consequences of this in the EU
are particularly severe because national governments have had widely different
preferences concerning supranational agents.
The basic treaties of the EU have legitimated Commission initiatives in several
policy areas, yet they are vague enough to give the Commission wide latitude in
designing institutions. This has been described as a "treaty base game" in which
the Commission legitimates its preferences by referring to a prior treaty commit-
ment (Rhodes 1995). This was the case in structural (or cohesion) policy, which,
in the wake of the Single European Act, was transformed by the Commission
from a straightforward side payment transferring money from richer to poorer
countries to an interventionist instrument of regional policy (Hooghe 1996c).
The European Court of Justice (ECJ) does not merely act as an agent in
adapting member state agreements to new contingencies. Through its rulings,
it has engineered institutional changes that escape, and transcend, treaty norms.
Supranational authority in the ECJ deepened from the 1960s, with the estab-
lishment
of
principles
of
supremacy and direct effect, as a result of Court rul-
ings, not because of treaty language. The constitutionalization of EU treaties is
the product
of
Court activism, not of national government preferences (Alter
1998; Burley-Slaughter and Mattli 1993; Volcansek 1992; Weiler
IYY
I; Stone
Sweet and Brunell 1998).
Hurdles to change: Unanimity is a double-edged sword for supranational in-
stitutions in the EU.
It
raises the bar for any kind of major institutional change in
the EU, whether it empowers supranational institutions or reins them in. A supra-
national actor need only dent a united front of national governments in order to
block change. For example, the Commission sidestepped an attempt by a power-
ful coalition
of
national governments, including the U.K., Germany, and France,
to renationalize cohesion policy in 1993 because it managed to gain the support
of just three' small member states: Ireland, Portugal, and Belgium.
Informational symmetries. Principal control may be weakened if an agent has
access to information or skills that .are not available to the principal (Majone
1994, 1996; Eichener 1992). As a small and thinly staffed organization, the Com-
mission has only a fraction of the financial and human resources available to na-
tional governments, but its position at the center of a wide-ranging network in-
cluding national governments, subnational governments, and interest groups
gives it a unique informational base for independent influence on policy making.
Mutual distrust. It is the collective interest
of
a national governments to enact
certain common regulations, but each may be better off if others adhere to them
While it defects. One response is to establish a court that can contain defection.
Another is to have very detailed legislation. The reverse side of ambiguity in
the treaties has been a willingness on the part of national governments to allow
the Commission to formulate precise regulations on specific policies so as to

13
12 Chapter 1
straitjacket principals and reduce their scope for evasion (Majone 1996). The
Commission likes to legitimate its role in technocratic terms, as the hub of nu-
merous specialized policy networks of technical experts designing detailed reg-
ulations (Peterson and Bomberg 1999; Richardson 1996, 1998).
POLICY MAKING IN THE
EUROPEAN
UNION
Who are the key actors in European Union policy making?
If
the state-centric
model is valid, one would expect to find that national governments dominate.
This entails three conditions. First, each state should maintain its sovereignty in
the process of collective decision making. Second, national governments, by
virtue of the European Council and the Council of Ministers, should be able to
impose their preferences collectively on other European institutions, i.e., the Eu-
ropean Parliament, the European Commission, and the European Court of Justice.
Third, national governments should control the access of subnational groups in
the European arena. If, however, the multi-level governance model is valid, we
should find that state sovereignty is compromised in collective national decision
making, that collective national decision making does not determine policy out-
comes, and that subnational interests mobilize beyond the reach
of
national gov-
ernments directly in the European arena.
To make headway with this issue, it makes sense to disaggregate policy mak-
ing. We divide the policy-making process into four sequential phases: policy ini-
tiation, decision making, implementation, and adjudication. We lean on analyses
of formal rules where they bear on these phases, but we also pay attention to in-
formal practices that shape the way actors interpret and exploit formal rules.
Policy Initiation: Commission as Conditional Agenda Setter
In political systems that involve many actors, complex procedures, and multi-
ple veto points, the power to set the agenda is extremely important. The Euro-
pean Commission alone has the formal power to initiate and draft legislation,
which includes the right to amend or withdraw its proposal at any stage in the
process, and it is the think tank for new policies (Article 221 TEC, ex-155). In
this capacity, it annually produces two to three hundred reports, white papers,
green papers, and other studies and communications (Ludlow 1991). Some are
highly technical studies about, say, the administration
of
milk surpluses. Oth-
ers are influential policy programs, such as the 1985 white paper on the inter-
nal market; the 1990 reform proposals for the common agricultural policy,
which laid the basis for the European position in the GATT negotiations; the
1993 white paper Growth, Competitiveness,
and
Employment, which argued
for labor market flexibility; or the 1997 Agenda 2000, which shaped the debate
on enlargement to Central and Eastern Europe.
Multi-Level Governance
To be able to play its policy-initiation role, the Commission needs access to in-
formation. It has superior in-house knowledge concerning agriculture, where one-
fifth of its staff is concentrated, and it has formidable expertise in external trade
and competition, the two other areas where Commission competence is firmly es-
tablished. In other fields, the Commission relies upon member state submissions,
its extensive advisory system of public and private actors, and paid consultants
(Laffan I996c; Nugent 1999; Peterson 1997).
Does the European Commission make a real difference? Does it exert signifi-
cant autonomous influence over the agenda, as a multi-level governance per-
spective would suggest? Or is it largely a decorative institution that draws up leg-
islation primarily to meet the demands
of
national governments, as a state-centrist
might suggest?
In recent years, the Commission itself has understandably stressed its lack of
autonomy from more democratically accountable institutions. In an internal ac-
counting exercise in 1998, the Commission estimated that only 5 to 10 percent of
legislative proposals arose spontaneously within the Commission itself.
I I The rest
were a response to international obligations (35 percent), amendments to or cod-
ification
of
existing law (25 to 30 percent), requests from other EU institutions,
national governments, or interest groups (20 percent), or required by prior treaty
(10 percent) (Peterson 1999,59; Peterson and Bomberg 1999, 38).
The proportions are revealing, but not quite in thc way the Commission intended.
Within each of these categories, except perhaps for treaties, the Commission has a
measure of influence. With respect to international obligations, the Commission it-
self negotiates on behalf of the EU on trade and, since the I990s, the environment.
So, for example, the Commission represents the EU in the World Trade Organiza-
tion, and while it must be in close contact with national governments on sensitive
trade issues, it plays a central role in negotiations. The Commission also takes the
lead in negotiating with countries that wish to join the EU and with countries seek-
ing economic or cultural cooperation with the ED. The second and third categories
listed here-t-amendmenrs and codification of existing law, and Commission re-
sponse to other actors--encompass widely varying situations. In some, the Com-
mission merely codifies agreements worked out among national governments, as is
the norm in transport, energy, and fisheries policies. In others, such as the annual
renegotiation of agricultural production quotas and prices or the renegotiation of co-
hesion funding every five to seven years, the Commission has significant agenda-
setting power (on the latter, see chapters 6 and 7). Finally, Commission proposals re-
flect treaty commitments, but even here the Commission is by no means passive, for
reasons noted above. To the extent that treaty commitments are vague, the Commis-
sion has leeway in pressing them into institutional form. The great reform of EU co-
hesion policy was, for example, mandated by treaty in 1986 but was hammered into
innovative institutional form by the Commission (Marks 1992; Hooghe I996a). Ac-
cording to the Commission, which has no reason to belittle national governments,
treaties generate only one-tenth of its legislative proposals.

15
14 Chapter 1
The picture that emerges is one where the Commission holds the pen but is sub-
ject to pressures from many actors. Policy initiation in the European Union is a
multi-actor activity.
It
includes, in addition to the Commission, the European
Council, the European Parliament, the Council of Ministers, and interest groups
alongside individual member states.
European
Council
A potentially powerful principal with respect to the Commission is the Euro-
pean Council, the summit of the political leaders of the memher states (plus the
president of the Commission), which is held three or four times a year. The Eu-
ropean Council has immense prestige and legitimacy and a quasi-legal status as
the body that defines "general political guidelines" (Title 1, Article 4 TEU, ex-D).
However, its control of the European agenda is limited because it meets rarely
and it provides the Commission with general policy mandates rather than specific
policy proposals. European Council mandates have proven to be a flexible basis
for the Commission to build legislative programs.
A striking example of this is the European Energy Charter, a formal agreement
between Russia and Western European states guaranteeing Russian energy sup-
ply after the collapse of the Soviet Union (Matlary 1993, 1997). Energy policy
became an EU policy because the Commission preempted an alternative inter-
governmental approach preferred by the Dutch, German, and British govern-
ments. Acting on a vague mandate of the European Council in June 1990, the
Commission negotiated a preliminary agreement with the Russian government in
1991. National governments, presented with a fait accompli, accepted the Euro-
pean Union as the appropriate forum for the Charter and gave the Commission a
toehold in international energy policy (Matlary 1993), a noteworthy incursion in
a policy area that had formerly been determined by national governments.
European Parliament
and
Council
of
Ministers
More direct constraints on the Commission originate from the European Par-
liament and the Council of Ministers. Indeed, the power of initiative has increas-
ingly become a shared competence. permanently subject to contestation, among
the three institutions. The European Parliament (Article 192 TEC, ex-138b) and
the Council (Article 208 TEC, ex-152) can request the Commission to produce
proposals, although they cannot draft proposals themselves. So far, the European
Parliament has made relatively little use of its recently gained competence in
Ar-
ticle 192, which enables it, by an absolute majority of its members, to request the
Commission to act. By 1999 only a handful of such requests had been made (Nu-
gent 1999).
The Council
of
Ministers, and particularly the presidency of the Council, began
to exploit this window in the legal texts from the mid-I 980s (Nugent 1999). Gov-
Multi-Level Governance
ernments often bring detailed proposals with them to Brussels when they take
over the Council presidency. The Council can also circumvent the Commission's
formal monopoly of legislative proposal by making soft law, i.e., by ratifying
common opinions, resolutions, agreements. and recommendations (Nugent 1999;
Peterson and Bomberg 1999; Snyder 1994). More often, though, national repre-
sentatives and Commission officials work hand in hand to push a new issue up
(or down) the agenda. Most initiatives germinate in the machinery of advisory
committees and working groups that the Commission has set up for consultation
and pre-negotiation. Many committees are made up of national government nom-
inees (usually civil servants), but others consist of interest group representatives
or experts (Page 1997; Wessels 1997). As it is the Commission that organizes and
pays for these committees, it is well placed to shape their agenda. National rep-
resentatives wishing to raise an issue need to cultivate the Commission officials
in charge, for they must be persuaded that an initiative is important enough to go
on the agenda.
Interest groups
Diffusion
of
control over the
EU's
agenda does not stop here. Interest groups
have mobilized intensively in the European arena and, while their power is diffi-
cult to pinpoint, it is clear that the Commission takes their input seriously. The
passage of the Single European Act precipitated a sharp increase in interest group
representation in Europe (Marks and McAdam 1996; Fligstein and McNichol
1998). National and regional organizations of every kind have mobilized in Brus-
sels, and these are flanked by a large and growing number of European peak or-
ganizations and individual companies from across Europe (Greenwood 1997;
Greenwood and Aspinwall 1998). According to a Commission report, some 3,000
interest groups and lobbies employing about 10,000 people were based in Brus-
sels in 1992. Among these are 500 "Euro-groups," which aggregate interests at
the European level (McLaughlin and Greenwood 1995) and some 150 offices in
Brussels representing regional and local authorities (Bomberg and Peterson 1998;
Hooghe and Marks 1996b; Hooghe 1995b; Marks, Salk, Ray, and Nielsen 1996;
chapter 5). Most groups target their lobbying activity at the European Commis-
sion and the European Parliament, for these are perceived to be more accessible
than the secretive Council (Mazey and Richardson 1993b, 1999).
The Commission's ability to create new advisory committees has helped it
reach out to new constituencies, and these include many subnational groups. An
example of this strategy was the creation of the Advisory Council for Local and
Regional Authorities in 1988 to advise the Commission on initiatives in cohesion
policy. The Commission hoped to mobilize support from below for a "partner-
ship" approach to structural programming in which the Commission, national,
and subnational governments would jointly design, finance, and implement eco-
nomic development programs. One of the Commission's longer-term goals was
L

17
16 Chapter 1
to institutionalize regional participation, and a step was taken in this direction
with the establishment of a Committee of the Regions in 1993. While the Com-
mission alone was not responsible for this
outcome-pressure
by the German
Lander and the Belgian regions on their respective governments was
pivotal-the
Advisory Council laid the groundwork. The purview of the Committee of the Re-
gions was extended in the Amsterdam Treaty of 1999 (see chapter 5).
European institutions compete for control over agenda setting. Interest groups
and subnational governments vie to influence the process. One consequence is
that it is difficult to assign responsibility for particular initiatives. This is true for
the most intensively studied initiative of
all-the
internal market
program-
which was pressed forward by business interests, the Commission, and the Euro-
pean Parliament, alongside national governments (Cameron 1992; Cowles 1995;
Dehousse 1992; Garrett and Weingast 1993; Majone 1994; Moravcsik 1998). Be-
cause the Commission plays a subtle initiating role, one cannot capture its influ-
ence by examining which institution formally announces a new policy. For ex-
ample, the European Council mandated the white paper Growth, Competitiveness,
and Employment in June 1993, but it did so in response to detailed guidelines for
economic renewal tabled by the Commission president. Another example is
Agenda 2000: For a Stronger and Wider Union, which the Commission produced
in July 1997 in response to a formal request by the Madrid European Council of
December 1995. The product was a 1,300-page white paper outlining Commis-
sion opinions on the ten applicants for European Union membership, reform pro-
posals for common agricultural policy and the structural funds, and proposals
about how to finance enlargement. The Commission used the opportunity to set
the agenda for the 1999 negotiations on the multi-annual budget, including re-
forms in common agricultural policy (CAP) and cohesion policy. As one observer
noted, "much of the policy drive on the enlargement of the EU to the Central and
Eastern European countries has been provided by the Commission working
within a framework authorized by the European Council" (Nugent 1999, 119).
Policy initiation engages a wide range of participants, but the Commission is the
critical actor in this phase, whether one looks at formal rules or political practice.
The Commission's leverage on setting the agenda depends on its ability to antici-
pate and mediate demands, and its capacity to employ expertise derived from its
role as the think tank of the European Union. While the Commission uses its for-
mal powers of initiative from time to time to shape the agenda autonomously, it is
usually responsive to the wishes of the European Council, the Council of Minis-
ters, the European Parliament, or interest groups.
It
is inaccurate to claim that the
Commission's role is merely an agent of national governments. A balanced read-
ing of the evidence suggests instead that the Commission operates in a system of
multi-level governance involving competition and interdependence among it and
the European Council, Council of Ministers, and European Parliament. These in-
stitutions share authority in the intricate game of policy initiation.
Multi-Level Governance
Decision Making: State Sovereignty in Retreat
The thrust of the state-centric argument is to give great weight to the legislative
powers of national governments in the decision-making stage. In this view, na-
tional governments adjust policies to their collective preferences, define the lim-
its of European collaboration, and determine the role of the European Commis-
sion and the ECJ and, if need be, curtail their activities.
If
previous decisions have
unintended consequences, the Council can correct them. National governments
may then be said to be in complete control.
According to the EU treaties, the main legislative body is the Council of Min-
isters, composed of national governments. Before the Single European Act of
1986, the chief legislative weapon of the Parliament was to slow down legislation
by withholding its opinion. Such actions were rare, and the Council was effec-
tively the sole legislative authority. This is no longer the case, however. In the
first place, individual governments have operated under serious constraints since
the Single European Act. Second, even collectively, national governments exert,
at most, conditional control. National government control has been eroded by the
legislative power of the European Parliament, the role of the European Commis-
sion in overcoming transaction problems, and the efforts of interest groups to in-
fluence outcomes in the European arena.
Limits on individual control
The most transparent blow to individual sovereignty has come from the pro-
gressive extension of qualified majority voting in the Council. Table 1.1 dis-
aggregates Council voting rules as described in the treaties into two cate-
gories: provisions that lay down unanimity, and those that prescribe some form
of majority voting (simple majority, qualified majority, or other special major-
ity). Two broad observations can be made. First, treaty negotiators never
intended unanimity to be the default rule for the Council of Ministers. From
the start, there were at least as many treaty provisions with majority rule as
with unanimity. And second, the proportion
of
rules stipulating unanimity
in the Council has steadily declined from 49 percent under the Treaty of
Rome (1958-1987), to less than 45 percent under the Single European Act
(1987-1993), to 35 percent under the Maastricht Treaty (1993-1999), with a
slight increase to almost 37 percent under the Amsterdam Treaty (from 1999).
Qualified majority voting is now the rule for decisions under the first pillar for
Community policies, such as the single market, competition policy, economic
and monetary union, regional policy, trade, environment, research and devel-
opment, transport, employment, immigration and visa policy, social policy,
and education. Qualified majority voting also applies to some provisions under
the second pillar for Common Foreign and Security Policy (CFSP pillar),
namely to "proposed actions related to agreed strategies," and to a handful
of

19
18 Chapter 1
Table 1.1 Decision Rule in the Council of Ministers According to the Treaties
Percentage
of
treaty provisions where
the Couneil decision rule is:
Simple, Qualified, or
Unanimity Special Majority
51.0
49.0
Treaty of Rome (1951\-1987) 55.5
44.5
Single European Act (1987-1993) 64.9
35.1
Maastricht Treaty (1993-1999) 63.3
36.7
Amsterdam Treaty (1999+)
Sources: Wessels, Mauer, and Mittag 1999, 10. The figures for the Amsterdam Treaty are our own calcula-
tions based on a categorization of treaty provisions by Simon
Hix
(1999, appendix).
decisions under the third pillar for Justice and Home Affairs (JHA pillar). The
decision rules are complex,
but
the bottom line is clear: over broad areas of EU
competencies, individual governments may be outvoted.
One might point out that these formal rules do not necessarily determine be-
havior. Under the informal Luxembourg compromise of 1966, a national govern-
ment can veto a decision subject to majority voting if it claims that its "vital na-
tional interests" are at stake. This gentlemen's agreement, one may argue, is proof
of the ultimate sovereignty of member states. However, the Luxembourg com-
promise has always featured more strongly in academic debates than in the prac-
tice of European politics.
It
was invoked less than a dozen times between 1966
and 1981, and it has been used only a handful of times since.
The real impact of the compromise was to reinforce a veto culture, which in-
hibited majority voting if a national government expressed serious objections.
During the 1970s, this paralyzed hundreds of Commission proposals, but the very
existence of this veto culture was its undoing. It eroded during the I980s as the
European Parliament and many national leaders became intolerant of deadlock
(Teasdale 1993). A turning point was the inability of the British government in
1982 to veto a decision on agricultural prices to extract a larger British budgetary
rebate. Prime Minister Margaret Thatcher's veto was ruled inadmissible, and a
qualified majority vote was taken in the Council of Ministers over British objec-
tions. Thereafter, national governments became more reluctant to invoke the
compromise or tolerate its use by others. The last successful use of the Luxem-
bourg veto was in June 1985 (Teasdale 1993). As Neill Nugent has observed, the
Luxembourg compromise "is in the deepest of sleeps and is subject only to occa-
sional and largely ineffective awakenings" (1999, 169).
It
has become much more difficult for national governments to justify the veto
so that others will accept its use on a particular issue normally determined by
qualified majority. Every government will find itself in a minority from time to
time. Why should a particular government be able to escape the consequences of
Multi-Level Governance
this? Unless that government forces its will on other others by threatening to shut
down decision
making-a
strategy as rare as it is
risky-it
must gain acquies-
cence within the Council for its veto. In practice, the conditions are restrictive:
the issue must be clearly linked to vital national interests, and the government
concerned must convincingly claim that it risks severe domestic political damage
(e.g., mass demonstrations or a cabinet crisis). An individual government can sus-
tain a veto only if other governments approve. This is a far cry from the original
intention of the Luxembourg compromise, which legitimized unconditional de-
fense of national
sovereignty-French
President Charles de Gaulle vetoed the EU
budgetary reform of 1965 on the grounds that it was too supranational. The no-
tion of vital national interest has evolved to justify defense of substantive inter-
ests, not defense of national sovereignty itself.
The Luxembourg veto is, in any case, a dull weapon. It cannot block alterna-
tive courses of action, as the German federal government learned in 1985 after it
had vetoed a Council regulation on lower prices for cereal and colza. The Com-
mission responded by invoking its emergency powers and achieving virtually the
same result unilaterally (Swinbank 1989; Teasdale 1993). Six months later, the
Council, with German acquiescence, abandoned the status quo on cereal and
colza prices (Teasdale 1993).
There are other ways for national governments to defend individual interests,
but they depend on the consent of the other governments. For example, special
safeguards can be built into the treaties, a practice that has proliferated since the
Maastricht Treaty. Particular states have been granted derogations, that is, special
exemptions. The United Kingdom and Denmark each have derogation from Eu-
ropean monetary union. Several derogations have been granted in the areas of
state aid, environmental policy, and energy policy. Sometimes they are written
into special protocols, such as those attached to the Amsterdam Treaty that meet
concerns of Denmark, Ireland, and the U.K. on border controls and EU immigra-
tion and visa policy. The Amsterdam Treaty also inserted a new decision
rule-
constructive
abstention-that
allows a member state to abstain from voting on an
issue and to formally declare that it will not implement a decision that nonethe-
less commits other EU member states. Constructive abstention is, however, re-
stricted to certain issues under the CFSP pillar (Stubbs 1999). In addition, the
treaties preserve unanimity for the most sensitive and contested policy areas, par-
ticularly major foreign policy decisions, nearly all decisions on justice and home
affairs, and much of fiscal policy. And finally, there is the norm within the Coun-
cil of Ministers that it should operate consensually on delicate political issues.
The norm appears to be weakening, however. A recent count found that the Coun-
cil overrules one or more national governments on around one-quarter of all de-
cisions (Hayes-Renshaw and Wallace 1997). The weekly European Voice esti-
mated that between January 1995 and January 1998, Germany was most often
outvoted in the Council, followed by Britain and, at some distance, Italy (15-21
October 1998,4).

21
20 Chapter 1
Derogations, constructive abstention, Council norms, and the continued vital-
ity of unanimity soften the blow of European integration for national sovereignty.
But qualified majority voting has been extended, the Luxembourg veto has re-
treated, and decision making in the Council has become more contentious. There
is evidence that larger, as well as smaller, countries find themselves outvoted. Au-
thority over broad areas of policy has shifted from individual national states to
collective decision making in the Council of Ministers (Wessels 1992; Scharpf
1994; Meny, Muller, and Quermonne 1994).
Collective national control
The Council of Ministers shares decision-making authority with other Euro-
pean institutions. Over the last three decades, the European Parliament has been
transformed from a decorative institution to a directly elected co-legislator (Ja-
cobs and Corbett 1990; Westlake 1994). Direct election of representatives to the
Parliament was mandated in 1976 (followed by the first EU-wide elections in
1979); parliamentary consultation was strengthened in the early 1980s; a cooper-
ation procedure, giving the Parliament significant agenda-setting powers, and an
assent procedure were part of the Single European Act of 1986; and codecision,
giving the Parliament a veto over many areas of legislation, was introduced under
the Maastricht Treaty of 1993 and extended under the Amsterdam Treaty of 1999.
Through the assent procedure, the EP has a veto over enlargement of the EU and
over most association agreements and treaties between the European Union and
third parties. According to some observers, these changes constitute "a major step
towards a bicameral legislative model (with two strong players) at the EC level"
(Falkner and Nentwich 1999, 26).
Institutional analysis demonstrates that the Parliament must be taken seriously
in explaining legislative outcomes (Crombez 1996; Steunenberg 1994; Scully
1997; Tsebelis 1994, 1995; Tsebelis and Garrett 1999). Table 1.2 provides an
overview
of
the European Parliament's formal role in decision making. The Par-
liament was almost powerless under the Treaty of Rome. It played no role in 73
percent of the provisions and was consulted on the remaining 27 percent. By 1999
and the Amsterdam Treaty, the EP had emerged as a force to be reckoned with.
Roughly equal proportions of provisions mandated codecision, consultation, and
exclusion. The Council of Ministers is stronger, of course. It votes under alI treaty
provisions, but the trend seems perfectly clear, as are the implications for the col-
lective capacity of national governments to determine policy making. In place of
the original Council-dominated process, Council, Parliament, and Commission
now interact in making policy.
The cooperation procedure allowed the Commission to set the agenda (Tsebelis
1994, 1995; Garrett and Weingast 1993; Weiler 1991; compare with skeptical
early prognoses: Bieber, Pantalis, and Schoo 1986). It could decide to take up or
drop amendments from either the
Councilor
Parliament, a power that made it a
Multi-Level Governance
Table 1.2 Role of
the
European Parliament in EU Legislation According to
the
Treaties
Percentage
of
treaty provisions where the European
Parliament's role in EU legislation is:
Decisional
(cooperation,
codecision,
assent) Consultative
None
Treaty of Rome (1958-1987) 0.0 27.1 72.9
Single European Act (1987-1993) 18.2 27.3 54.5
Maastricht Treaty
(1993-1999)
22.9
37.0 40.1
Amsterdam Treaty (1999+) 33.1 34.5 32.4
Sources: Wessels, Mauer, and Mittag 1999, 10. The figures for the Amsterdam Treaty are our own calcula-
tions based on a categorization of treaty provisions by Simon Hix 11999, appendix),
broker-a
consensus
crafter-between
the two institutions. The Council could
not decide legislation without the support of either the Commission or the Euro-
pean Parliament unless it was unanimous.
Under the codecision procedure the European Parliament can veto Council leg-
islative proposals. A conciliation committee, consisting of representatives from
both institutions with a representative of the Commission as broker, tries to ham-
mer out a compromise if Parliament and Council are deadlocked. To become law.
a compromise proposal needs to be approved by a majority in the Parliament and
a qualified majority in the Council. The codecision procedure comes close to put-
ting the European Parliament "on an essentially equal footing with the Council"
(Falkner and Nentwich 1999, 26). The Commission retains important agenda-
setting powers, though its broker role is weaker than under the cooperation pro-
cedure (Tsebelis and Garrett 2000). Under both procedures the Council is locked
in a complex relationship of cooperation and contestation with the two other in-
stitutions. Tftis is multi-level governance in action and is distinctly different from
what one would expect in a state-centric system.
One reason why the collective control of national governments has eroded is
because the Council often lacks information, expertise, and the coordination to
act quickly and effectively. The Council is an egalitarian institution, and this can
complicate coordination, particularly now that there are fifteen member states
(Garrett and Weingast 1993; Pollack 1997; Scharpf 1988; Majone 1994). The
Commission is more coherent because it is more hierarchical. It is far from a uni-
tary organization (see chapter 9), and there are often rivalries between its depart-
ments, but it is usually more able than the Council to present a united front on a
particular proposal. Formal decision rules in the Council of Ministers often help
the Commission focus discussion or broker compromise. While national repre-
sentatives preside at Council meetings, the Commission sits in to clarify, redraft,
and finalize the proposal; in short, it holds the pen.

23
22 Chapter 1
European defense policy illustrates the Commission's capacity to frame deci-
sion making. The regulation of defense industry in Europe is debated between
those who regard it as a normal industry that does not require special treatment
and those who emphasize its special security role. The former favor deregulation
in order to exploit economies of scale; the latter wish to preserve national firms
or at least organize them on a European-only basis. In the European Commission,
the directorate-generals for industry and the internal market favor deregulation
and liberalization, while the directorate-general for external relations conceives
the issue as one of security in the context of Common Foreign and Security Pol-
icy. This tug-of-war made it impossible for the Commission to weigh in on ED
decision making until November 1997, when the two protagonists, external rela-
tions and industry, found a way to combine their views in a report emphasizing
the dual character of the industry. The report, entitled Implementing European
Union Strategy on Defense-Related Industries, laid out two complementary
plans-a
European defense equipment policy in the framework of the CFSP and
an industrial action plan for defense-related industries ill the framework of the EC
pillar (Morth 2000). By linking Community policies to CFSP measures, the Com-
mission catapulted itself into the heart of decision making on European defense
industry, and it provided the Council of Ministers with a focal point in its search
for a feasible policy.
The frustrations of intergovernmental cooperation may lead national govern-
ments to voluntarily cede authority to supranational agents, as immigration pol-
icy illustrates (Stetter 2000; Falkner and Nentwich 1999; den Boer and Wallace
2000). When in the late 1980s the Commission made a case for an El.I-wide im-
migration policy to give backbone to the free movement of people in the single
market, national governments rejected a supranational solution. Cooperation on
immigration and border control under Schengen (1985) and on asylum (the 1990
Dublin Convention) were attempts by national governments to overcome regula-
tory problems among themselves. Intergovernmental problem solving reached its
zenith in the Maastricht Treaty, where national governments institutionalized co-
operation in immigration policy, visa policy, border control, and police coopera-
tion in the third pillar of Justice and Home Affairs. This arrangement excluded the
Commission, the ECJ, and the European Parliament from decision making, but
frustrations quickly piled up. The Schengen agreement did not enter into force
until 1995, and even then Britain, Ireland, and Denmark decided to stay outside.
The Dublin Convention was ratified only in 1997.
In the run-up to the Amsterdam conference, national governments openly con-
ceded that the third pillar did not function properly. The intergovernmental re-
flection group that prepared the Amsterdam Treaty observed that the voluntary
legal framework of Justice and Home Affairs created "uncertainty in legal pro-
tection" for citizens because it was not based on transparent principles, and "en-
forcement problems" because it was difficult for national governments to make
binding commitments. The reflection group noted further that unanimity voting
Multi-Level Governance
exacerbated these problems and recommended that immigration policy be incor-
porated into the first (EC) pillar of the European Union.'? In the Amsterdam
Treaty, national governments transferred the bulk of immigration policy to the EC
pillar, which means
that-after
a transition period
of
five
years-the
Commission
will have the sole right of initiative, the European Parliament will codecide, the
Council of Ministers will vote on most issues by qualified majority, and the ECJ
will adjudicate.
I)
"What used to be defined as areas of 'common interest' between
member states has now become an objective of the ED" (Stetter 2000,94).
The Commission's unique resources sometimes enable it to step beyond its role
of umpire to become a negotiator. Cohesion policy offers an example. In estab-
lishing the framework for structural funds for 1994 to 1999 during the summer of
1993, Commission officials negotiated bilaterally with officials from the relevant
states. The Belgian presidency served as umpire. In essence, the Commission be-
came a thirteenth participant around the bargaining table (Hooghe 1996a). Some-
thing similar has taken place from time to time even in the most intergovernmen-
tal
forum-treaty
bargaining-as
revealed in the Maastricht negotiations. When
the British government refused the watered-down social provisions at Maastricht,
Jacques Delors put his original, more radical, social policy program of 1989 on the
table and proposed to attach it as a special protocol to the Treaty, leaving Britain
out. Faced with the prospect that the whole negotiation might break down, the
other eleven national governments hastily signed up to a more substantial docu-
ment than they had originally anticipated (Pierson 1996; Lange 1993).
In sum, the Council is the senior actor in the decision-making stage, but the Eu-
ropean Parliament and the Commission are indispensable partners. The Commis-
sion's power is predominantly soft in that it is exercised by influence rather than
sanction. Except for agriculture, external trade, and competition policy, where it
has substantial executive autonomy, it can gain little by confrontation. The Com-
mission's influence depends on its
ability-and
indispensability-in crafting
consensus among institutions and among national governments. Extensive re-
liance on qualified majority voting should in principle enable the Commission to
be bolder, as it does not have to court every national government. Nevertheless,
ideological convergence in the Council since 1997 and ineffective leadership of
the Commission have weakened the Commission's pivotal role at the start
of
the
twenty-first century (Peterson 1999).
It
remains to be seen whether the Commis-
sion's relative decline will outlive these conjunctional factors.
The European Parliament's position is based more on formal rules. Its track
record under cooperation and codecision shows that it does not shy away from
confrontation. In return for parliamentary assent to enlargement and the GATT,
the Council agreed to allow parliamentary observers in the preparatory negotia-
tions for the intergovernmental conference
of
1996-1997, and the Parliament
emerged as the main institutional winner in the Amsterdam Treaty (Falkner and
Nentwich 1999). It is intent on making the most
of
its power, even if it treads on
the toes of its long-standing ally, the European Commission. Since the Amsterdam

25
24 Chapter 1
Treaty, the European Parliament's assent is needed for the appointment
of
the
Commission president, as well as for the whole team. "The result of these reforms
is a quasi-parliamentary system (i.e., the classic 'indirect' mechanism) for select-
ing this branch of the EU executive, whereby the Commission president is nom-
inated by the member states and ratified by the European Parliament immediately
following the EP elections" (Hix 1999c, 97). In spring 1999, a Parliamentary in-
quiry into fraud, mismanagement, and nepotism in the Commission forced the
Santer Commission to resign en masse under a cloud of accusations.
Authoritative competencies in Europe are exercised across multiple levels of
government. At the European level, national governments and supranational ac-
tors share authority, and the institutions in which they operate have intermeshing
competencies.
Implementation: Breaking the State
Mold
Multi-level governance is prominent in the implementation stage. The formal di-
vision of authority between the Commission, which had sole executive power,
and member states, which monopolized policy implementation, no longer holds.
National governments have come to monitor the executive powers of the Com-
mission, and the Commission has become involved in day-to-day implementation
in a number of policy areas, and this brings it into close contact with subnational
governments and interest groups. As for agenda setting and decision making, the
mutual intrusion
of
institutions into the other's terrain is contested.
The Commission's formal mandate gives it discretion to interpret legislation
and issue administrative regulations or decisions for specific cases. It used to an-
nounce at least 4,000 administrative regulations annually, and an equal number of
decisions, but in the late 1990s this number more than halved (Nugent 1999; Lud-
low 1991). Still, the Commission remains formally responsible for the bulk
ofEU
rules (Dogan 1997).
From the 1980s the Council of Ministers and individual governments became
intimately involved in the executive powers of the Commission. The term for this
is comitology, which refers to the practice of having a committee of national rep-
resentatives assist the Commission in its executive work. Many regulations have
their own committee attached to them. Rules of operation vary from committee
to committee, and they are a source of friction among the Commission. the Par-
liament, and the Council (Neyer I
()()();
Christiansen and Kirchner forthcoming).
Some committees are only advisory; others can prevent the Commission from
carrying out a certain action by qualified majority vote; a third category must ap-
prove Commission actions by qualified majority. In each case the Commission
presides.
At first sight, comitology may seem to give national governments control over
the Commission's actions in principal-agent fashion, but this impression is mis-
leading. Comitology is weakest in precisely those areas where the Commission
Multi-Level Governance
has extensive executive powers, e.g., in competition policy, state aids, agricul-
ture, commercial policy, and the internal market. Here, the Commission has sig-
nificant space for autonomous action (McGowan and Wilks 1995; Nugent 1999;
see also Tsebelis and Garrett 2000). Comitology does not alter the basic fact that
national governments have lost their monopoly
of
authority in many policy
areas: for example, they no longer control competition within their borders, they
cannot aid national firms as they deem fit, and they cannot autonomously con-
duct trade negotiations.
National governments often select people outside the central executive to rep-
resent them in comitology. Most participants in comitology are not national civil
servants but are subnational officials, interest group representatives (particularly
from farming, union, and employer organizations), technical experts, scientists,
or academics (Buitendijk and van Schendelen 1995; Page 1997; van Schendelen
1996). Subnational participation in comitology is most common for federal or
semi-federal states, though in recent years, more centralized states have followed
suit (Goetz 1995; Hooghe 1995b; McLeod 1999; see also chapter 5).
Though subnational officials, technical experts, interest group representatives,
and private actors are selected by their national government to participate in
comito logy, they have particular territorial or group interests, as well as the na-
tional interest, to defend. Comitology was designed to allow national govern-
ments to monitor the Commission, but it has had the additional, and unintended,
consequence
of
deepening subnational and group participation in the European
political process.
Commission officials now
playa
role in day-to-day policy implementation.
The Commission was never expected to perform ground-level implementation,
except in unusual circumstances (such as competition policy, fraud, etc.). Yet in
some areas this has changed. The most prominent example is cohesion policy,
which absorbs about one-third of the EU budget. The bulk of the money goes to
multi-annual regional development programs in the less developed regions of
the EU. As

detail in chapters 5, 6, and 7, the structural funds reform of 1988,
followed by revisions in 1993 and 1999, involves regional and local govern-
ments as well as social actors in all stages
of
the policy
process-the
selection
of priorities, choice of programs, allocation of funding, monitoring of opera-
tions, and evaluation and adjustment of programs. Each region or country re-
ceiving funding is required to set up monitoring committees with a general com-
mittee on top and a cascade of subcommittees focused on particular programs.
Commission officials can and do participate at each level of this tree-like struc-
ture. Partnership is implemented unevenly across the EU (Bache 1998; Heinelt
and Smith 1996; Hooghe and Keating 1994; Hooghe 1996c; Marks 1996b), but
just about everywhere it institutionalizes some form of direct contact between
the Commission and subnational governments. Such links break open the mold
of the state, so that multi-level governance encompasses actors beneath, as well
as above, central states.

27
26 Chapter 1
Adjudication: An Activist Court in a Supranational
Legal
Order
State-centrists have argued that a European legal order and effective European
Court of Justice are essential to state cooperation (Garrett and Weingast 1993;
Garrett 1995; Moravcsik 1993). Unilateral defection is difficult to detect, and
thus it is in the interest of states to delegate authority to a European court to mon-
itor compliance. The
ECl
also mitigates problems of incomplete contracting by
applying current agreements to future contingencies. From this point of view, the
ECl
is an agent of the member states. However, a number of scholars have con-
vincingly argued that the Court has become more than an agent of member states
(Alter 1998; Burley-Slaughter and Mattli 1993; Dehousse 1998; Mattli and
Slaughter 1995, 1998; Stone Sweet and Brunell 1998; Weiler 1991). With the
help of the Commission, and in collaboration with national courts, the
ECl
has
transformed the European legal order in a supranational direction.
The
ECl
has laid the foundation for an integrated European polity. The Court
has built an impressive body of case law establishing the Treaty of Rome as a
document creating legal obligations directly binding on national governments
and individual citizens alike. These obligations have legal priority over laws
made by member states. Directly binding legal authority and supremacy are core
attributes of sovereignty, and their application by the
ECl
suggests that the EU is
becoming a constitutional regime.
The Court was originally expected to act as an impartial monitor "to ensure
that in the interpretation and application of the treaties the law is observed" (Ar-
ticle 220 TEC, ex-I64; Article 136 Euratom; Article 31 ECSC), but from the be-
ginning the Court viewed these interstate treaties as more than narrow interna-
tional agreements (Alter 1998). The Court's expansive role is founded on the
failure of the treaties to specify the competencies of major EU institutions. In-
stead, the treaties set out "tasks" or "purposes" for European cooperation, such as
the custom union (Treaty of Rome), the completion of the internal market (Sin-
gle European Act), or economic and monetary union (Maastricht Treaty). The
Court has constitutionalized European law and European authority in other pol-
icy areas by stating that these were necessary to achieve these functional goals
(Weiler 1991).
Court rulings have been pivotal in shaping European integration. However, the
ECl
depends on other actors to force issues on the European political agenda and
condone its interpretations. Legislators (the European Council, Council of Min-
isters, Commission, and Parliament) may reverse the course set by the Court by
changing the law or altering the EU treaties. The
ECl
is no different from the
Council, Commission, or European Parliament in that it is locked in mutual de-
pendency with other actors.
One outcome of this interlocking is the principle of "mutual recognition,"
which became the core principle of the internal market program. In the landmark
case of Cassis de Dijon (1979), the Court stated that a product lawfully produced
Multi-Level Governance
in one member state must be accepted in another. But it was the Commission that
projected the principle of mutual recognition onto a wider agenda, the single mar-
ket initiative, and it did this as early as July 1980 when it announced to the Eu-
ropean Parliament and the Council that the Cassis case was the foundation for a
new approach to market harmonization (Alter and Meunier-Aitsahalia 1994).
National courts have proved willing to apply the doctrine of direct effect by in-
voking Article 234
(ex-In)
of the Treaty of Rome, which stipulates that national
courts may seek "authoritative guidance" from the
ECl
in cases involving Com-
munity law. In such instances, the
ECl
provides a preliminary ruling, specifying
the proper application of Community law to the issue at hand. While this prelim-
inary ruling does not formally decide the case, in practice the court renders a
judgment of the "constitutionality" of a particular statute or administrative action
in light of its interpretation of Community law. The court that made the referral
cannot be forced to accept the
ECl's
interpretation, but if it does, other national
courts will usually accept the decision as a precedent. Preliminary rulings expand
ECl
influence, and judges at the lowest level gain a de facto power of judicial re-
view, which had been reserved for the highest court in the state (Burley-Slaugh-
ter and Mattli 1993). Article 234 gives lower national courts strong incentives to
circumvent their own national judicial hierarchy. With their support, much of the
business of interpreting Community law has been transferred from national high
courts to thc
ECl
and thc lower courts.
ECl
decisions have become an accepted part of the legal order in the member
states, shifting expectations about decision-making authority from a purely na-
tional-based system to one that is multi-level. The doctrines of direct effect and su-
premacy were constructed over the strong objections of several national govern-
ments. Yet, the influence of the
ECl
has as much to do with creating opportunities
for other actors, including the Commission and lower national courts, to influence
European rule making as it does with its enlarged scope for unilateral action.
CONCLUSION
Multi-level governance does not confront the sovereignty of states directly. In-
stead of being explicitly challenged, states in the European Union are being
melded into a multi-level polity by their leaders and the actions of numerous sub-
.--
national and supranational actors. State-centric theorists are right when they argue
t-
that national states are extremely powerful institutions that are capable of crush-
ing direct threats to their existence. The institutional form of the state emerged be-
cause it proved a particularly effective means of systematically wielding violence,
and it is difficult to imagine any generalized challenge along these lines. But this
is not the only, or even the most important, issue facing the state. One does not
have to argue that states are on the verge of political extinction to believe that their
control of those living in their territories has significantly weakened.

29
28 Chapter J
It is not necessary to look far beyond the state itself to find reasons that might
explain such an outcome. When we disaggregate the state into the people and or-
ganizations that shape its diverse institutions, it is clear that key decision makers,
above all those directing the national government, may have goals that do not co-
incide with projecting national sovereignty into the future. The state is a means
to a variety of ends, which are structured by party competition and interest group
politics in a liberal democratic setting.
Even if national governments want to maintain national sovereignty, they are
often not able to do so. A government can be outvoted because most decisions in
the Council are now taken by qualified majority. Moreover, the national veto, the
ultimate instrument of sovereignty, is constrained by the willingness of other na-
tional governments to tolerate its use. But the limits on sovereignty run deeper.
Even collectively, national governments do not determine the European agenda
because they are unable to control the supranational institutions they have cre-
ated. The growing diversity of issues on the Council's agenda, the sheer number
of national principals, the mistrust that exists among them, and the increased spe-
cialization of policy making have made the Council of Ministers reliant upon the
Commission to set the agenda, forge compromises, and supervise compliance.
The most obvious blow to Council predominance has been dealt by the Euro-
pean Parliament, which has gained significant legislative power since the Single
European Act. Indeed, the Parliament has become a principal in its own right. The
Council, Commission, and Parliament interact within a legal order, which has
been transformed into a supranational one through the innovative jurisprudence
of the European Court of Justice.
Since the 1980s, these changes in EU decision making have crystallized into a
multi-level polity. With its dispersed competencies, contending but interlocked
institutions, and shifting agendas, multi-level governance opens multiple points
of access for interests. In this process of mobilization and counter-mobilization,
national governments no longer serve as the exclusive nexus between domestic
politics and international relations. Direct connections are being forged among
political actors in diverse political arenas.
Multi-level governance may not be a stable equilibrium. There is no explicit
constitutional framework. There is little consensus on the goals of integration. As
a result, the allocation of competencies between national and supranational actors
is contested.
It
is worth noting that the European polity has made two U-turns in
its
Sh0l1
history. Overt supranationalist features
of
the original structure were
overshadowed by the imposition of intergovernmental institutions in the 1960s
and 1970s (Weiler 1991). From the 1980s, a system of multi-level governance
arose, in which national governmental control became diluted by the activities of
supranational and subnational actors. The surreptitious development of a multi-
level polity has engendered strong reactions. The EU-wide debates unleashed by
the Maastricht Accord have forced the issue of national sovereignty onto the pub-
lic agenda. Where governing parties themselves have shied away from the issue,
Multi-Level Governance
opposition parties, particularly those of the extreme right, have raised it. States
and state sovereignty have become objects of popular
contention-the
outcome
of which is as yet uncertain.
NOTES
1. This chapter is based on an earlier version coauthored with Kermit Blank (Marks,
Hooghe, and Blank 1996). We would like to thank Simon Bulmer, Jim Caporaso, Stephen
George, John Keeler, Peter Lange, Andrea Lenschow, Christian Lequesne. Mark Pollack,
Michael Shackleton, and Helen Wallace for their useful comments on earlier drafts of this
chapter. We are indebted to Ivan Llamazares and Leonard Ray for research assistance.
2. While the roots of the state-centric model lie in neorealism, there are a variety of
state-centric approaches to European integration that take issue with certain neorealist as-
sumptions and attempt to encompass domestic politics as an int1uence on the formation of
state preferences. The most interesting of these is "liberal institutionalism," which, despite
its nuanced view of interstate cooperation and state preference formation, is firmly in the
state-centric mold.
Liberal institutionalism focuses on how international institutions foster gains from co-
operation where they otherwise might not arise. International institutions diminish anar-
chy, but the state-centric perspective remains intact: states are unitary and state preferences
are determined exogenously or by domestic politics (Caporaso 1996a). "The basic claim
...
is that the EC can be analysed as a successful intergovernmental regime designed to
manage economic interdependence through negotiated policy coordination
....
An under-
standing of the preferences and power of its member states is a logical starting point for
analysis" (Moravcsik 1993,474).
This approach allows that European institutions are strong: "Strong supranational insti-
tutions are often seen as the antithesis of intergovernmentalism. Wrongly so" (Moravcsik
1993,507). But they are at the service of member states, not independent:
'The
unique in-
stitutional structure of the EC is acceptable to national governments only insofar as it
strengthens, rather than weakens, their control over domestic affairs, permitting them to at-
tain goals

unachievable" (Moravcsik 1993, 507). Alan Milward claims that "the
political machinery of the Community resembles the court of a minor eighteenth-century
German state. There is a numerous and deferential attendance around the president of the
Commission. A hierarchical bureaucracy attends to the myriad facets of relationships with
the surrounding greater powers, for every decision has to be finely attuned to the wishes
of the real powers to which the Community's continued existence is useful. The struggles
to appoint to its offices are like those within the Imperial Diet" (Milward 1992,446).
European institutions are not essentially different from other international institutions.
All serve a precise function: "Like other international regimes, EC institutions increase the
efficiency of bargaining by providing a set of passive, transaction-cost reducing rules"
(Moravcsik 1993, 518). Consequently, supranational actors cannot achieve political au-
tonomy. In this respect, the ED looks strikingly similar to a consociational regime:
"Consociational theory sees the state apparatus as being an umpire rather than a promoter
of any specific ideology
....
[P]ressures to enlarge the role of the Commission as umpire
are increased rather than diminished as integration proceeds" (Taylor 1991, 118-119).

30 31
Chapter I
The state-centric model claims that member states have EU institutions firmly under
control. "The EC regime
...
fixes interstate bargains until the major European powers
choose to negotiate changes" (Moravcsik 1993, 31). In effect, "the most fundamental task
facing a theoretical account of European integration is to explain these bargains" (Morav-
csik 1993,473). To do so, one should refer back to the preferences of participating states:
"EC
institutions appear to be explicable as the result of conscious calculations by member
states" (Moravcsik 1993, 507). And when states choose to transfer sovereignty to supra-
national institutions "their principal national interest will be not only to define and limit
that transfer of sovereignty very carefully hut also meticulously to structure the central in-
stitutions so as to preserve a balance of power within the integrationist framework in favor
of the nation-states themselves" (Milward and Sorensen 1993, 19).
In the most general sense, European integration has served to rescue the nation state.
"The European Community has been its buttress, an indispensable part of the nation state's
post-war construction. Without it, the nation state could not have offered to its citizens the
same measure of security and prosperity which it has provided and which has justified its
survival" (Milward 1992, 3). "[Sltates will make further surrenders of prosperity if, but
only if. they have to in the attempt to survive" (Milward 1992,446). Stanley Hoffmann ar-
rived at the same conclusion along somewhat different lines: "In areas of key importance
to the national interest, nations prefer the certainty, or the self-controlled uncertainty,
of
national self-reliance, to the uncontrolled uncertainty of the untested blender
....
The logic
of diversity implies that, on a vital issue, losses are not compensated by gains on other (and
especially not on other less vital) issues: nobody wants to be fooled
....
The logic
of
in-
tegration deems the uncertainties
of
the supranational function process creati ve: the logic
of diversity sees them as destructive past a certain threshold: Russian roulette is fine only
as long as the gun is filled with blanks" (Hoffmann 1966, 882).
Despite these gloomy predictions, by the early I990s, the annual regulatory output of
the European Community was greater than that of most individual states and 75 to 80 per-
cent of national legislation was subject to prior consultation with the European Commis-
sion (Majone 1994). How do state-centrists account for this expansion? Some argue that
state competencies have merely shifted: "The European nation state has lost some eco-
nomic functions to the EC and some defense functions altogether, while gaining functions
in what had previously been more private and local spheres. Overall, the bars of the [na-
tional] cage may not have changed very much. Citizens still need to deploy much of their
vigilance at the national level" (Mann 1993, 130). For others, state sovereignty is still in-
tact: "policymaking in the Community has not in itself detracted from national sover-
eignty: what is changed is the wish of national legislatures and governments to do certain
things rather than their legal or constitutional right or capacity to do them" (Taylor 199 I,
123). Still others worry less ahout the scope as long as member states control the depth of
European intrusion. And here voluntarism and the individual
veto-"fundamental
deci-
sions in the EC can be viewed as taking place in a non-coercive unanimity voting system"
(Moravcsik 1993,
498)--combine
to make outcomes converge at the lowest common de-
nominator. "The need to compromise with the least forthcoming government imposes a
binding constraint on the possibilities for greater cooperation, driving EC agreements to-
ward the lowest common denominator. A lowest common denominator outcome does not
mean that final agreements perfectly reflect the preferences of the least forthcoming gov-
ernment-since
it is generally in its interest to compromise somewhat rather than veto an
agreement-but
only that the range of possible agreements is decisively constrained by its
Multi-Level Governance
preferences" (Moravcsik 1993, 50 I). However, many outcomes cannot be characterized as
lowest common denominator (see our argument below), a point that some state-centrists
are now conceding (Moravcsik 1995, n. 3).
Community institutions that try to challenge member states do not get very far: "As for
the common organs set up by the national governments, when they try to act like a Euro-
pean executive and parliament, they are both condemned to operate in the fog maintained
around them by the governments and slapped down if they try to dispel the fog and reach
the people themselves" (Hoffmann 1966, 910).
One contribution of liberal institutionalism, and of Andrew Moravcsik's work in partic-
ular, lies in the attempt to specify the conditions under which "international cooperation
.
..
tends on balance to strengthen the domestic power of executives vis-a-vis opposition
groups" (Moravcsik 1994, 7, his emphasis). However, even though the billiard ball model
of the nation state is cracked open to understand state preferences, state-centrists resort to
unitary actor assumptions to analyze interstate bargaining: "Groups articulate preferences;
governments aggregate them" (Moravcsik 1993, 483).
3. States or state leaders are conceived as monopolizing the interface between the
neatly separated arenas of European and domestic politics. European decision making is
seen as "a process that takes place in two successive stages: governments first define a
set of interests, then bargain among themselves in an effort to realize those interests"
(Moravcsik 1993, 481). State-centrists make short shrift of interest group representation
in Brussels:
"Even
when societal interests are transnational, the principal form
of
their
political expression remains national" (Moravcsik 199I, 26). European and national pol-
itics belong to two different worlds because there is no need for direct interplay:
"If
par-
ties have organized themselves only in a superficial way in the European Parliament, that
is because no more has been needed
....
[It] is within the nation that political parties have
to fulfill their task of organizing a democratic consensus" (Milward
1992,446).
Other
state-centrists argue that domestic and EU arenas are nested rather than interconnected
because it is in the interest of national governments to keep
them
that way: "The EC does
not diffuse the domestic influence of the executive; it centralizes it. Rather than domesti-
cating the international system, the EC internationalizes domestic politics. While coop-
eration may limit the external flexibility of executive, it simultaneously confers great do-
mestic influence
....
In this sense, the EC strengthens the state" (Moravcsik 1994, 3, his
.,
emphasis).
4. In his book The Choice for Europe, Andrew Moravcsik has dropped this concept of
power. Instead, he relies on the notion of "credible commitments" to explain under what
conditions national governments agree to "pool" or "delegate" sovereignty (Moravcsik
1998, 73-77).
5. The Amsterdam Treaty (1999) strengthens majority voting in the European Union's
common foreign and security policy (CFSP), creates a High Representative for the CFSP,
and sets up a supranational unit for foreign policy
analysis-the
Policy Planning and Early
Warning Unit. The latter unit pools expertise from the Commission, the Council secre-
tariat, the member states, and the Western European Union (Peterson and Bomberg 1999).
However, these steps fall well short of a profound transfer of authority in foreign and de-
fense policy to the European Union.
6. This has strong implications for how one can explain preference formation in na-
tional governments as well as in the European Parliament, and we examine some of these
in chapter 4.

32
2
Chapter 1
7. As the authority of the Parliament has grown, so its internal operation has become
more important. The norms that govern the EP's parliamentary procedures, its committee
structure, the selection of candidates on party lists, and the development of transnational
European party federations all lie outside the treaties.
8. These policy areas are summarized in table A 1.1 in appendix I.
9. The only exception was the European Defense Community, which was voted down
in the French Assemblee in 1954. After that debacle, plans for the European Political Com-
munity were quietly dropped.
10. For Britain, the latter are a constitutional innovation with immense knock-on effects.
11. This refers to "EC pillar" issues, which encompass the bulk of EU initiatives. EC
pillar, or pillar I, issues refer to economic integration, including economic and monetary
union, and all policies areas; pillar II refers to common foreign and defense policy (CFSP);
pillar III to cooperation on justice and home affairs (JHA).
12. The strongest proponents for a transfer of immigration and border control to pillar I
were the Dutch, Belgian, Luxembourgian, German, Italian, Portuguese, and Austrian gov-
ernments, while the French and Spanish governments were in favor of a partial transfer
(den Boer and Wallace 2000).
13. During the transition period, the Commission shares its right of initiative with the
member states, the EP is only consulted, and the Council of Ministers votes by unanimity.
A Historical Perspective
The
creation
of
a European polity over the past half century has been an ex-
periment in interstate coordination and supranational institution building.
While the European Union is a new kind of polity, scholars have sought to
gauge its particularities and understand its dynamics by comparison. This
chapter lies squarely in that tradition, one that goes back to the earliest at-
tempts to analyze European integration.
Comparison, but with what? Given the exceptional character of European in-
tegration, the question has no single answer. European integration does not fit
neatly into any class of political phenomena, though it shares interesting com-
monalities with several.
Two lenses have been used to gain comparative insight. The first treats the Eu-
ropean Union as an international regime. Like the United Nations, the General
Agreement on Tariffs and Trade, or the North American Free Trade Association,
the EU can be conceived as an organization created, sustained, and dominated by
national governments. Conceptualizing the EU as an international regime focuses
attention on intergovernmental bargaining and allows scholars to inquire into the
factors that lead to coordination among national governments (Moravcsik 1991,
1994; for a critique, see Sandholtz 1996). Why do national governments create
international regimes, and what functions does the European Union fulfill?
A second lens treats European integration as the development of a federal con-
stitutional
order-a
domestic regime. From this standpoint, the European Union
has been compared to a variety of existing federal regimes, including those in
Switzerland, Canada, Germany, and the United States (Sbragia 1992; Cappel-
letti, Seccombe, and Weiler 1986; Scharpf 1992).1 Here the focus has been on
institutional arrangements that link constituent governments to the center. What
is the role of constituent territorial units in central decision making, and how are
they constrained by the center? How are constituent territorial units represented
in EU institutions?
33