The Pluralism of Global Administrative Law

The European Journal of International Law Vol. 17 no.
1 © EJIL 2006; all rights reserved
...........................................................................................
The Pluralism of Global

Administrative Law
Nico Krisch*
Abstract
As public power is increasingly exercised in structures of global governance, principles of
domestic law and politics are extended to the global level, with serious repercussions for the
structure of international law. Yet, as this article seeks to show for the emerging global
administrative law, this extension is often problematic. Using administrative law
mechanisms to enhance the accountability of global regulation faces the problem of
fundamental contestation over the question of to whom global governance should be
accountable. National, international and cosmopolitan constituencies are competing for
primacy, and this results in an often disorderly interplay of accountability mechanisms at
different levels and in different regimes. This pluralist structure, based on pragmatic
accommodation rather than clear decisions, strongly contrasts with the ideals of coherence
and unity in modern constitutionalism and domestic administrative law. However, given the
structure of global society, it is likely to endure and it is also normatively preferable to
alternative, constitutionalist approaches. It helps avoid the friction that may result from a
federal-type distribution of powers and the practical problems of a consociational order, and
by denying all constituencies primacy it reflects the legitimacy deficits of each of them.
Mirroring divergent views on the right scope of the political order, it also respects
everybody’s equal right to political participation. A pluralist global administrative law thus
presents an alternative to problematic domestic models for ensuring accountability in the
circumstances of global governance.
1 Introduction
As more and more state functions are performed on the global level, the concepts that
have guided politics and law are increasingly under pressure. International law’s self-
understanding as an order that organizes commonality among grossly different value
* Merton College, Oxford University. I am grateful to Gráinne de Búrca, Benedict Kingsbury, Terry Macdonald,
Neus Torbisco Casals, and participants in the International Justice Seminar at Merton College and in
Oxford’s Public International Law Research Seminar for helpful comments and criticism. Email:
[email protected].
........................................................................................................................................................
EJIL (2006), Vol. 17 No. 1, 247–278 doi: 10.1093/ejil/chi163
248 EJIL 17 (2006), 247–278
systems is questioned by those who refuse to give up democracy and constitutionalism

merely because the exercise of public power has now moved to the international level.
Likewise, the extension of the guiding ideas of domestic law and politics to global
governance faces serious problems; these ideas often lose their normative force under
the different political and social circumstances that prevail in the global realm.
Global administrative law seeks to explore these questions from the angle of
accountability; it asks whether and to what extent ideas from domestic administrative
law can help us solve accountability problems in global governance.1 In this article,
I seek to take this enterprise further by focusing on one central challenge to the use of
administrative law concepts for global regulation, namely the fundamental contesta-
tion over the question of to whom global governance should be accountable. As I will
argue, the strength of this contestation contrasts with the unitary, hierarchical basis
on which domestic administrative law rests, and it leads to a peculiarly rugged land-
scape of accountability mechanisms on the global level. In many instances, these
mechanisms are disorderly and highly unstructured; they operate in different regimes
and are driven by oppositional dynamics. Global regulation faces challenges by
domestic courts, by international civil society, and by competing international
regimes, yet all of them function relatively independently – through mutual chal-
lenge rather than through fulfilling assigned roles in a coherent overall system.
I call this often unstructured interaction a ‘pluralist administrative law’, and I will
argue that it is unlikely to be a passing phenomenon that will wither away in the pro-
cess of building more robust mechanisms of accountability. It results from fundamen-
tal and durable contestation over the right constituency of global governance, and
there are many reasons to regard it also as a normatively adequate response to this
contestation. In the circumstances of global governance, attempts at ‘constitutionalizing’
the political order by forcing it into a coherent, unified framework are problematic as
they tend to downplay the extent of legitimate diversity in the global polity; and this
has important consequences for the structure of accountability mechanisms. If I am
right, the shape of global administrative law is likely to be, and should be, essentially
different from that of domestic administrative law.
I try to make my point in five steps. In Section 2, I set the stage for the argument by
clarifying the importance of constituency questions for accountability in general and
by discerning the central, though often neglected role of such questions in domestic
administrative law settings. I also try to sketch the three main archetypes in the con-
test around the right constituency of global governance: the national, international
and cosmopolitan constituencies. In Section 3, I use this framework to analyse in
greater depth one area of global regulation in which this contest, and the resulting
pluralist order, is particularly apparent: the regulation of transboundary trade with
genetically modified organisms (GMOs). Here, WTO regulation faces challenges from
domestic courts, especially in the European Union (EU), but also from the competing
regime of the Biosafety Protocol, and we can understand the resulting fragmented
1
See Kingsbury, Krisch, and Stewart, ‘The Emergence of Global Administrative Law’, 68 Law and Contem-
porary Problems (2005) 15.
The Pluralism of Global Administrative Law 249
picture as a response to the open question of which constituency should control this
type of regulation. In Section 4, I then ask whether such pluralist accountability
mechanisms are merely an unwelcome step on the way to a global administrative law
along well-ordered domestic lines, or whether they are themselves normatively defen-
sible. I begin with a prudential argument, based on the limits of competing
approaches of a more constitutionalist character in dealing with the constituency
contest. A federal model of assigning defined powers to different constituencies would
have to take sides rather than remain neutral in this contest; and a consociational
model, relying on power-sharing and veto rights of the various constituencies, would
likely prove impractical given the number of actors and the resulting risk of blockade.
A pluralist approach, granting the different constituencies the possibility to challenge
but not formally veto a decision, would reflect the essential contestation and also
achieve a sufficient degree of practicality. It would respond to the claims of the con-
stituencies not through their full co-decision, but through their distance from the
decision: none of them would be entitled to fully control the outcome, but each of
them could equally contest it. In Section 5, I seek to develop elements of a principled
defence of such a pluralist order. One pillar of this defence is based on the normative
deficits of all constituencies: since none of them can claim full legitimacy for global
regulatory decisions, they need to complement one another; and since co-decision is
largely barred for practical reasons, such complementarity should result in a pluralist
order. Another, normatively more demanding pillar of a principled defence rests on
the individual right to participate in the determination of the scope of the polity. If this
right is exercised in diverging ways, all choices can (in certain boundaries) claim
respect and need to be reflected in the institutional order. Granting one of them
primacy over the others would then be normatively problematic. In Section 6, I
briefly address some of the central problems of a pluralist order. Many challenges
remain, especially as regards the role of power in a pluralist approach, but in many
respects, alternative conceptions do not seem to provide better solutions. In a global
setting devoid of the certainties of the nation state, a pluralist global administrative
law might be the best option we have.
2 Accountability, Administrative Law, and Its Many

Constituencies
A Accountability: To Whom?
If ‘[t]o be accountable means to have to answer for one’s action or inaction, and
depending on the answer, to be exposed to potential sanctions, both positive and
negative’2, most institutions are accountable. This is most obviously true for a gov-
ernment that has to leave office when it loses elections, but it extends to such bodies
as courts that face reputational sanctions if their judgments do not find favour among
2
Oakerson, ‘Governance Structures for Enhancing Accountability and Responsiveness’, in J.L. Perry
(ed.), Handbook of Public Administration (1989), 114, at 114.
250 EJIL 17 (2006), 247–278
their peers and the public, and also to private entities. The board of a corporation is
accountable to the shareholders, and any company to the market, in the most exact-
ing case to the stock market. Likewise, most institutions of global governance are
accountable in this sense. The World Bank has to respond to the will of its member
states, and in particular to those with strong voting rights. The Financial Action Task
Force (FATF) is closely tied to the limited circle of its member entities, just as the
United Nations Security Council is highly accountable to its members, and in a more
diffuse way to the broader membership of the UN. And also a private body like the
International Organization for Standardization (ISO) is accountable to all its member
organizations as well as to the market that has to accept its standards.3
To speak of an ‘accountability deficit’ in global governance is therefore somewhat
misleading. Many regulatory institutions on the global level are in fact highly
accountable – up to the point that they often enjoy little freedom of independent
action and are closely tied to the wishes of their constituents. This is particularly true
for intergovernmental organizations.4 The problem with these institutions is, then,
not an absolute accountability ‘deficit’, to be overcome by improving or strengthen-
ing accountability mechanisms in a technical exercise. Rather, the problem is that
these institutions are often accountable in the wrong way: in part, they are accountable
to the wrong constituencies.5 The World Bank, it is often claimed, should respond to
the people affected by its decisions, rather than primarily to the (mostly developed)
countries that fund it. The FATF should be accountable to those states subject to its
measures, not just to its members. Or the Security Council should have to answer to
the individuals it targets directly with its sanctions, not only to its member govern-
ments or the broader membership of the UN.
Claims for stronger accountability of global governance thus involve not only the
fairly uncontroversial argument that any public power should be accountable as
such, but also much more contested normative positions on who should control public
power – since, after all, accountability is about control.6 Central to arguments about
accountability is then the question: ‘to whom’ should a given institution be account-
able?7 In the growing field of global administrative law, it is on this issue that the
stakes are highest for the distribution of power: if an institution becomes accountable
to a wider group, those that previously controlled it become less influential. If, for
example, the World Bank allowed for stronger representation and participation of
3
On the different forms of accountability mechanisms, see Fisher, ‘The European Union in the Age of
Accountability’, 24 Oxford J Legal Studies (2004) 495, at 501–508; Grant and Keohane, ‘Accountability
and Abuses of Power in World Politics’, 99 American Political Science Review (2005) 29, at 35–37.
4
See also ibid., at 37.
5
See Keohane, ‘Global Governance and Democratic Accountability’, in D. Held and M. Koenig-Archibugi
(eds), Taming Globalization: Frontiers of Governance (2003), 130, at 145.
6
See Scott, ‘Accountability in the Regulatory State’, 27 J Law and Society (2000) 38, at 39. It should be
noted, though, that accountability (which operates mostly ex post) is not the only form of control but
operates alongside various forms of potential ex ante control. See ibid.; also H.F. Pitkin, The Concept of
Representation (1967), at ch. 3; T. Macdonald, ‘We the Peoples’: NGOs and Democratic Representation in
Global Politics, D.Phil. Dissertation, Oxford (2005), at ch. 8.
7
See also Scott, supra note 6, at 41; Grant and Keohane, supra note 3, at 42.
developing countries and their citizens, the funders that so far dominate the Bank
would lose influence. Accountability is not necessarily zero-sum, but in many cases,
more accountability to one actor means less to another.
This article seeks to explore the question ‘accountable to whom?’ – the question of
the constituencies of global governance. Its claim is that the structure of global
accountability mechanisms and of global administrative law, especially its current
irregular shape, is best understood as a product of deep-seated disputes over the right
constituencies of global regulatory action. Yet the question of the right constituency
is not limited to the global level: it is also a central (though often less visible) element
of domestic administrative law systems.
B The Constituencies of Domestic Administrative Law

In domestic administrative law, the constituency question is usually not addressed as
such, and this is probably because the answer appears to be obvious, and constitu-
tionally mandated. The two primary constituencies seem indisputable: on the one
hand, the individual (or individual entities like companies) insofar as their rights or
interests are affected; and on the other hand, ‘the people’ insofar as administrative
action is supposed to serve democratic purposes, or more broadly the common good.
Yet especially with respect to the latter category, this clarity is only superficial, and
constituency questions resurface in a number of areas of administrative law, of which
I will only mention two central ones here.
The first, and structurally most important, area is the balance between the
national community (‘the people’) and smaller parts of it in the organization of
accountability mechanisms. This balance is crucial in all orders in which sub-entities
possess their own normative weight, as is often the case for local self-government; but
it finds its clearest expression in federal systems. In the case of Germany, where the
constitution prescribes that ‘all state authority is derived from the people’,8 this
includes both the people of the whole of Germany and the several peoples of the
Länder. The duality of constituencies leads to a complex system of interlinkages in
administrative organization, procedure and judicial review.9 In general, administra-
tion is a domain of the Länder, with the effect that organizationally, administrative
officials answer to the elected government on this level and thus eventually to the
people of the Land.10 Insofar as they execute federal law, however, they are subject to
supervision by the federal government11 as well as to judicial review not only by
Länder courts (which mixes accountability to the federal level through the content of
8
Art. 20(2) of the Grundgesetz für die Bundesrepublik Deutschland; for an English version, see
www.oefre.unibe.ch/law/lit/the_basic_law.pdf.
9
I can sketch only some principal traits here. For more detail, see Rudolf, ‘Kooperation im Bundesstaat’, in
J. Isensee and P. Kirchhof (eds), Handbuch des Staatsrechts der Bundesrepublik Deutschland (1990), iv, at
1091; Blümel, ‘Verwaltungszuständigkeit’, in ibid., at 857. On the tension of these interlinkages with
the general constitutional prohibition of a common administration (Mischverwaltung), see Hermes, ‘Art.
83’, in H. Dreier (ed.), Grundgesetz: Kommentar (2000), iii, at 1, 24–27.
10
Art. 84(1) Grundgesetz.
11
Art. 84(3) of the Grundgesetz. This supervision is limited to the conformity with federal law.
252 EJIL 17 (2006), 247–278
the legislation with that to the Länder level through the composition of the courts) but
ultimately also by federal courts. In contrast, regulation by the federal government is
often tied back to the Länder: in all areas in which federal legislation requires the con-
sent of the Bundesrat (the Upper House representing the Länder) or is to be executed by
the Länder, administrative rule-making also requires Bundesrat consent unless federal
law provides otherwise.12 Interlinkages between the different constituencies are even
stronger in the case of the European Union, where regulation in the Comitology
framework accounts, by virtue of its participants, to both the Union and the Member
State levels.13 But this approach contrasts starkly with that of the US, in which the
two levels of administration are in principle separate (yet state administration is
subject to judicial review by federal courts when compliance with federal law is at
issue).14 In all multi-level systems, though, the multiplicity of constituencies is a cent-
ral factor for the shape of accountability mechanisms.
The constituency question has been the focus of more explicit attention in the liter-
ature as regards the participation of private actors in governmental agencies’ admin-
istrative decision-making. In the US, this has been theorized as pluralist interest
representation, seeking to strengthen regulatory accountability directly to the people.
It represents a response to the increasingly limited capability of legislation to steer
regulatory outcomes and the resulting limits of substantive judicial review in tying
regulation back to electoral processes. Yet such interest representation has provoked
serious concerns that accountability is not to ‘the people’ but to particularly well-
organized and powerful interest groups.15 In Germany, the idea that accountability to
the people should be organized through responsibility to an elected government as
well as the ‘transmission belt’ of legislation is still dominant and enjoys even constitu-
tional status. Although the inclusion of affected interests in decision-making bodies
was initially seen with great scepticism, recent constitutional jurisprudence envis-
ages circumstances in which accountability to affected actors can complement that
to the people as a whole, e.g. in cases of functional self-government or representation
of particularly affected groups in collegial administrative bodies. Yet here, too, there
are concerns about potentially excessive influence of these new constituencies, and
the organizational structure must in any event be decided by parliament, so as to
ensure the ultimate dominance of the constituency of the people as a whole.16
In domestic administrative law, mechanisms to involve federal, state and local con-
stituencies as well as general and particular publics operate within a framework in
12
Art. 80(2) of the Grundgesetz. In practice, 40% of these regulations require Bundesrat consent: see
Bauer, ‘Art. 80’, in Dreier, supra note 9, (1998), ii, at 1526, 1534.
13
See C. Joerges and E. Vos (eds), EU Committees: Social Regulation, Law and Politics (1999). On the coopera-
tive structure of European administration more generally, see E. Schmidt-Aßmann and B. Schöndorf-
Haubold (eds), Der Europäische Verwaltungsverbund (2005).
14
But see Stewart, ‘Federalism and Rights’, 19 Georgia L Rev (1985) 917, on important deviations from
this principle in practice.
15
See Stewart, ‘The Reformation of American Administrative Law’, 88 Harvard L Rev (1975) 1667.
16
See Bundesverfassungsgericht, 2 BvL 5/98 and 2 BvL 6/98, 5 Dec. 2002, 107 BVerfGE 59; E. Schmidt-
Aßmann, Das Allgemeine Verwaltungsrecht als Ordnungsidee (2nd edn., 2004), at 87–101, 262–266.
which ‘the people’ (however understood) enjoys a paramount place; it remains the
ultimately decisive constituency. This structure reflects the aspiration to coherence,
unity and normative hierarchy characteristic of modern constitutionalism.17 On the
global level, not only does the number of important constituencies rise; but also none
of them enjoys clear primacy. Thus, if the question of the right constituencies is
important for the shape of domestic administrative law, it is likely to be central to the
structure of global administrative law.
C Three Competing Constituencies in Global Administrative Law

The question to whom global administration is, and should be, accountable is heavily
contested.18 Several constituencies compete for primacy, with different patterns
emerging in different institutional settings, and although the resulting picture is
highly varied and inconsistent, we can identify three dominant approaches.
The nationalist approach is the classical and probably still the dominant among the
three; it claims that control over regulatory decisions should lie with the national
constituency.19 This approach has for a long time been central to the international
legal system: since the rise of voluntarist positivism in the 19th century, international
law derives its legitimacy and binding force mainly from the consent (or at least the
acquiescence) of states, and each state is in theory free to decide whether to be bound
or not. In the realm of administration, the dominant role of the national constituency
is reinforced by the need for domestic implementation: it is through domestic admin-
istrative actors, accountable to a national public, that international norms are con-
cretized and applied. The requirement of domestic ratification and implementation
shields and separates the national level from the international one, allowing the
former to retain full control. The resulting dominance of the national constituency
has a strong normative basis: it respects political and cultural pluralism by allowing
each society to bring its own values to bear in regulatory decision-making.20 And in
the views of many, it is the only approach capable of realizing democratic aspirations:
cultural homogeneity, communicative coherence, or a strong sense of solidarity,
which are often seen as preconditions for strong democracy, are usually present only
in national communities. It is therefore the national constituency that needs to have
17
On the ideal of unity in modern constitutionalism, see J. Tully, Strange Multiplicity: Constitutionalism in an
Age of Diversity (1995). For attempts at reconceptualizing this model within states (though from a different
perspective from the one I focus on here), see also Ladeur, ‘Towards a Legal Theory of Supranationality:
The Viability of the Network Concept’, 3 European LJ (1997) 33, at 43–46; Dorf and Sabel, ‘A Constitution
of Democratic Experimentalism’, 98 Columbia L Rev (1998) 267. For a related attempt with a primary
focus on the European Union, see Walker, ‘The Idea of Constitutional Pluralism’, 65 MLR (2002) 317,
and the references infra in note 84.
18
I do not deal here with accountability to individuals with respect to their affected rights; on this, see
Kingsbury, Krisch, and Stewart, supra note 1, especially at 45–48.
19
I simplify here by treating nation and state as coextensive; especially in recent years, the issue of sub-
state nations has (rightly) attracted much attention; see W. Kymlicka, Contemporary Political Philosophy
(2nd edn., 2002), at ch. 8. I do not take this up here for reasons of clarity alone.
20
See Kingsbury, ‘Sovereignty and Inequality’, 9 EJIL (1998) 599, at 623–625.
254 EJIL 17 (2006), 247–278
the right to take all important decisions, even in international contexts.21 This consti-
tutes, in essence, the approach of the German Constitutional Court towards the Euro-
pean Union,22 and it seems to be increasingly influential in attitudes towards
international law in the US.23 It is also reflected in arguments for a strong role of regu-
latory networks in global administration. In those networks, it is claimed, domestic
accountability mechanisms with a high democratic pedigree are more effective than
in classical international organizations: they can influence the national regulators
that participate in decision-making in those networks, and they can later review the
decisions, as these are usually non-binding.24
The internationalist approach, in contrast, regards the international community of
states as the main constituency. In this view, on such issues as human rights or the
environment, international law has already moved beyond the narrow confines of
the state-voluntarist perspective and is increasingly shaped by concerns common to
all states.25 In global administration, this is reflected in the rise of international bodies
for rule-making, decision-making and review: international organizations, as far as
possible independent from member states, are seen as guardians of a broader, impar-
tial international interest, and they keep national administrators in check. The UN or
the World Health Organization are examples, just as human rights institutions or the
WTO Dispute Settlement Body. The internationalist approach is also supported by
strong normative reasons: on many issues, regulation produces effects well beyond
national boundaries and thus cannot be left to national constituencies if all affected
interests are to be dealt with in a fair way.26 In democratic terms, too, this is a power-
ful argument: insofar as democracy is regarded as self-government, the scope of the
polity has to reflect the range of affected individuals.27 For internationalists, though,
this does not imply a move towards a single global community: for them, the diversity
of political orders deserves recognition and is best protected through states in which
different political visions – liberal, collectivist, etc. – can find expression. Global gov-
ernance should thus be accountable to the national constituencies together, not to
each of them individually or to a single global body of citizens.
21
See E.-W. Böckenförde, Staat, Nation, Europa (1999), at 103–126; Grimm, ‘Does Europe Need a Constitu-
tion?’, 1 European LJ (1995) 282, at 292–297; also J. Habermas, Die postnationale Konstellation (1998), at
91–169. For a broader defence of nationalism, see D. Miller, On Nationality (1995), and for an overview
of nationalist arguments, see Kymlicka and Straehle, ‘Cosmopolitanism, Nation-States, and Minority
Nationalism’, 7 European Journal of Philosophy (1999) 65, at 68–72.
22
See Bundesverfassungsgericht (1993) 89 BVerfGE 155.
23
See Goldsmith, ‘Liberal Democracy and Cosmopolitan Duty’, 55 Stanford L Rev (2003) 1667; Rubenfeld,
‘The Two World Orders’, 27 Wilson Quarterly (Autumn 2003) 22. See also Dahl, ‘Can International
Organizations be Democratic? A Skeptic’s View’, in I. Shapiro and C. Hacker-Cordón (eds), Democracy’s
Edges (1999), at 19.
24
See A.-M. Slaughter, A New World Order (2004), at 230–244.
25
See, e.g., Simma, ‘From Bilateralism to Community Interest in International Law’, 250 Recueil des Cours
(1994-VI) 217.
26
See, e.g., D. Held, Democracy and the Global Order (1995), at 236; I.M. Young, Inclusion and Democracy
(2000), at 246–251; Kumm, ‘The Legitimacy of International Law: A Constitutionalist Framework of
Analysis’, 15 EJIL (2004) 907, at 922–924.
27
See Pogge, ‘Cosmopolitanism and Sovereignty’, 103 Ethics (1992) 48, at 63–69.
The cosmopolitan approach goes even further in that it insists on a genuinely global
constituency for issues of global governance. It shares with the international
approach the view that accountability to national constituencies is insufficient, but it
is based firmly on a liberal individualism according to which states are not valuable as
expressions of fundamental diversity but only as (albeit important) organizational
tools to ensure a division of labour and check the dangers of a world state. For cosmo-
politans, the right constituency would not consist in the community of states, but in
the global community of individuals, in a truly global public.28 They disagree on how
to institutionalize accountability to that constituency, with proposals ranging from
representative options such as a world parliament to more deliberative proposals.29
Most common, though, is an emphasis on NGOs as representatives of ‘global civil
society’ alongside state representation. Efforts to involve NGOs in the global administra-
tive space have indeed succeeded in a number of areas, from active engagement in
environmental contexts, for example in proceedings of the World Bank Inspection
Panel, to observer participation in standard-setting bodies such as the Codex Alimen-
tarius Commission or representation in forms of private regulation, as for example in
the Forest Stewardship Council. Cosmopolitans are certainly aware of the problems of
relying on NGOs, and they would not see them as sufficient to satisfy their vision of
global democracy but as a first step in the organization of a transnational global public.
These three constituencies – the several states, the community of states, the global
public – certainly represent only rough approximations of the main strands in the
debate on accountability in global governance, and many more approaches could be
identified: for example, the accountability to funders as in the international financial
institutions, or that to particular stakeholders, as in some forms of standard-setting.30
Moreover, in the way I have sketched the three constituencies here, they represent
archetypes or structures rather than ‘real’ constituencies:31 by speaking of an ‘inter-
national constituency’, for instance, I merely point to the level of community to
which accountability is due, not the precise shape of the accountability holder. This
abstracts from the particular reasons for regarding this constituency as relevant:
international control over environmental issues, for instance, might be preferred to
national control because it helps limit negative externalities, because it protects state
sovereignty against encroachments, or because in democratic terms it ensures that
all affected individuals can participate. Whichever approach one chooses will determine
the precise shape of the accountability holder: from a democratic perspective, this
might be the peoples of the several states affected, while from a sovereignty perspective,
28
At least alongside constituencies on other levels; see Held, supra note 26, at 226–238; Pogge, supra note
27, at 57–69; Young, supra note 26, at 265–271.
29
See Falk and Strauss, ‘Toward Global Parliament’, 80 Foreign Affairs (Jan./Feb. 2001) 212. See also the
useful survey in Howse, ‘Transatlantic Regulatory Cooperation and the Problem of Democracy’, in
G.A. Bermann, M. Herdegen, and P.L. Lindseth (eds), Transatlantic Regulatory Cooperation (2000), at 469.
30
See Woods, ‘Making the IMF and the World Bank More Accountable’, 77 Int’l Affairs (2001) 83, at 85–87;
Mattli and Büthe, ‘Global Private Governance: Lessons from a National Model of Setting Standards in
Accounting’, 68 Law and Contemporary Problems (2005) 225, at 250–259.
31
I am grateful to Gráinne de Búrca and Benedict Kingsbury for insisting that I clarify this point.
256 EJIL 17 (2006), 247–278
it might be the sum of national governments. My account brackets this problem and
seeks to highlight only the three abstract archetypical groups within which, despite
all differences in detail, various normative approaches as well as strands in interna-
tional society converge as primary candidates for holding global governance
accountable.
The main point of this section was to show the fundamentally contested nature
of the constituency of global administrative law: the competition of normatively
attractive and practically influential positions for primacy in the accountability
mechanisms of global governance. While in domestic administrative law multiple
constituencies coexist in an order that, as a matter of constitutional principle, has
at its top ‘the people’, in global administrative law no such order prevails.32
Instead, it is the contest between different constituencies that shapes the ‘constitu-
tional’ framework of the global polity as well as global administrative law and its
institutions.
3 Contested Constituencies in Global Risk Regulation: The

Case of GMOs
Contestation over the right constituency of global governance is written into the
institutional structures of global governance and global administrative law, and it is
the shape of these structures that this section seeks to analyse further. For this pur-
pose, I will focus on one particular case – the regulation of genetically modified organ-
isms (GMOs) – which provides an example of how accountability mechanisms in
global regulatory structures differ from those in typical domestic settings.33
As the previous section has indicated, in domestic settings we typically find uni-
tary, hierarchical structures of regulation and administrative law in which multiple
constituencies are accommodated through participation and review mechanisms as
well as checks and balances and a distribution of competences among different levels
of decision-making. In global regulation, such clearly ordered structures are unusual;
here, we find a much more disorderly institutional landscape, with decision-making,
participatory procedures, and review often taking place in different sites at the same
time and with no formal connection among them. This institutional ‘disorder’ makes
final decisions often difficult to reach, yet it also allows for an institutional expression
of the views of different constituencies, for mutual influences and gradual approxima-
tion, while permitting the operation of buffers that prevent full control of one over the
others.
32
But see on similar conflicts in early federal states and in the EU, Oeter, ‘Souveränität und Demokratie als
Problem in der “Verfassungsentwicklung” der Europäischen Union’, 55 ZaöRV (1995) 659, at 664–687;
Goldsworthy, ‘The Debate About Sovereignty in the United States: A Historical and Comparative Per-
spective’, in N. Walker (ed.), Sovereignty in Transition (2003), at 423; MacCormick, ‘The Maastricht
Urteil: Sovereignty Now’, 1 European LJ (1995) 259.
33
I am grateful to Richard Stewart for many insights into GMO regulation and for allowing me to use his
unpublished draft on ‘The GMO Challenge to Global Governance of Risk’ (on file with the author).
A The Substantive Dispute over GMOs

On the issue of GMOs, two fundamentally opposed approaches confront each other,
and both respond to deeply held convictions about risk, nature and scientific
progress. I can only sketch them here and have to leave out many nuances.
On the one hand, the ‘permissive’ approach that is today dominant in the
United States sees restrictions on the production, sale and use of foodstuffs as jus-
tified only when there are scientifically proven risks for human health, the envir-
onment, or other important goods. Absent such proof, the production, sale and
use of food and feed is free, and since for many products that contain GMOs or
have been produced on the basis of GMOs risk assessments have not revealed
ascertainably higher risks than for other products, restrictions are not warranted
under this approach.
On the other hand, the ‘precautionary’ approach that is favoured in Europe
emphasizes the scientific uncertainty that even thorough risk assessments leave and
insists that in situations of uncertainty and potentially serious risks, one should err
on the side of caution. Since the consequences for public health and the environment
for products containing or based on GMOs cannot be fully determined – in part
because of the short time that has so far been available for testing and in part because
testing is usually limited to small contexts rather than extending to entire ecosystems –
the precautionary approach tends to restrict the production, sale and use of such
products significantly.
On a more general level, these approaches reflect attitudes to risk in scientific
progress and in particular to alterations of nature and its potential consequences.
But in the case of the EU, the more cautious approach also stems from recent experi-
ences in the area of food safety, in particular the BSE scandal, as well as concerns
about the effect of a shift towards GMO food and feed for the agricultural landscape.
In this way, it may also be connected with a desire to shield the relatively small
European agricultural businesses from the pressures towards stronger industriali-
zation that GMO agriculture and competition with large-scale American farms
would bring.34
These two approaches, usually coexisting peacefully, clash over questions of global
trade. US exports of agricultural products containing or based on GMOs to Europe are
severely limited by stringent EU rules, and exports to developing countries are often
hampered because of the wish of these countries to export agricultural products to
Europe, which is more difficult with GMOs in the food chain. Accordingly, the contest
between the two approaches takes place mostly in trade-related institutions.
34
On the two general approaches, see ibid., at 11–16, 22–32; Vogel, ‘The Politics of Risk Regulation in
Europe and the United States’, 3 Yearbook of European Environmental Law (2003) 1; for a cautionary note
on the differences between the US and Europe, see Wiener and Rogers, ‘Comparing Precaution in the
United States and Europe’, 5 J Risk Research (2002) 317. On the development of European GMO regula-
tion, including important differences between EU Member States, see Shaffer and Pollack, ‘Regulating
Between National Fears and Global Disciplines: Agricultural Biotechnology in the EU’, Jean Monnet
Working Paper 10/04, available at www.jeanmonnetprogram.org/papers/04/041001.html.
258 EJIL 17 (2006), 247–278
B The Institutional Contest

Institutionally, the conflict is situated mainly between the EU and the WTO, with two
other institutions – the Codex Alimentarius Commission and the Cartagena Protocol
on Biosafety – also playing a significant role.
Europeans understand the WTO as a threat to their approach to GMOs mainly because
the WTO’s Agreement on the Application of Sanitary and Phytosanitary Measures (SPS
Agreement) emphasizes the role of a scientific risk assessment when trade restrictions for
human or animal health are at issue.35 The SPS Agreement is widely regarded as seri-
ously limiting precautionary considerations; and the WTO Appellate Body has in a
number of cases found it to be violated by measures based on such considerations.36 One
of these cases concerned politically highly sensitive restrictions by the EU on hormones in
beef37 and has certainly contributed to European concern that WTO rules challenge its
own food safety regulation. Yet, on the other hand, the Appellate Body has often left con-
siderable scope for national regulation in public health and environmental matters, and it
has indicated that a precautionary approach may find wider application than just for the
provisional measures for which it is explicitly admitted in the SPS Agreement.38
As a result, the legal issues under the SPS Agreement are far from clear-cut and a
good case can be made for the conformity of European measures with it;39 thus, the
central issue is about the decision-making body and in particular about the level on
which decision-making should take place. While from an American perspective this
decision is best taken on an international level and with regard mainly to an interna-
tional (trade) constituency, Europeans insist that the dispute is about values and
policies and that it is primarily for the European polity to determine: to simplify and
put it in the terms of the previous section, Europeans insist on the superiority of the
national over the international constituency in this regard.40 Accordingly, they have
35
On the dominant role of science in the SPS Agreement and WTO decisions, see Peel, ‘Risk Regulation
under the WTO SPS Agreement: Science as an International Normative Yardstick?’, Jean Monnet Work-
ing Paper 02/04, available at www.jeanmonnetprogram.org/papers/04/040201.html.
36
Precautionary considerations are explicitly allowed only with respect to provisional measures in situations
of insufficient scientific evidence: see Art. 5(7) of the SPS Agreement. For the cases, see EC Measures Con-
cerning Meat and Meat Products, Report of the Appellate Body, WTO Docs. WT/DS26/AB/R & WT/DS48/
AB/R (16 Jan. 1998) (hereinafter: EC Beef Hormones); Australia – Measures Affecting Importation of Salmon,
Report of the Appellate Body, WTO Docs. WT/DS18/AB/R (20 Oct. 1998); Japan – Measures Affecting Agri-
cultural Products, Report of the Appellate Body, WTO Docs. WT/DS76/AB/R (22 Feb. 1999); Japan – Measures
Affecting the Importation of Apples, WTO Docs. WT/DS245/AB/R (26 Nov. 2003). For an overview of the first
three cases, see Victor, ‘The Sanitary and Phytosanitary Agreement of the World Trade Organization: An
Assessment After Five Years’, 32 NYU J Int’l L and Politics (2000) 865, at 895–913.
37
EC Beef Hormones, supra note 36.
38
See Howse, ‘The WHO/WTO Study on Trade and Public Health: A Critical Assessment’, 24 Risk Analysis
(2004) 501.
39
See Howse and Mavroidis, ‘Europe’s Evolving Regulatory Strategy for GMOs – The Issue of Consistency
with WTO Law: Of Kine and Brine’, 24 Fordham Int’l LJ (2000) 317.
40
I simplify here also by treating the European constituency as ‘national’ for the purpose of this article. Given
the tensions between the European and Member State levels within the EU, interesting constituency con-
tests can also be observed in that framework; see, e.g., Shaffer and Pollack, supra note 34. For reasons of
clarity, though, I leave them out here and focus on the central commonalities of the approaches in Europe.
challenged the complaint brought in 2003 by the US, Canada and Argentina against
EU restrictions on GMOs as denying states their right to a sovereign decision and plac-
ing the issue in the wrong forum:
[I]t is not the function of the WTO Agreement to allow one group of countries to impose its
values on another group. ... There is a serious question as to whether the WTO is the appropri-
ate international forum for resolving all the GMO issues that the Complainants have raised in
these cases.41
The dispute before the WTO can thus be interpreted as a contest of constituencies,
with both international and European constituencies claiming a right to determine
the limits of trade with GMOs. From this perspective, we can also understand the dif-
ferent, and competing, accountability mechanisms that are at play here as expres-
sions of this same contest. The WTO dispute settlement regime, in reviewing
European regulatory decisions and thus providing a mechanism of judicial review of
administrative action,42 seeks to tie domestic administration back to an international
constituency: it serves, in Robert Keohane’s words, to mitigate the ‘external accounta-
bility gaps’ of national regulation that produces external effects.43 From an opposite per-
spective, the European mechanism that determines the status of WTO law in European
Union law ties global regulatory action back to the European, national constituency.
This mechanism is very political in nature: although in principle, international trea-
ties enjoy direct effect in EU law, European courts have denied this effect to WTO law,
arguing that otherwise the political organs of the EU would be severely limited in
their negotiations over trade disputes.44 Therefore, WTO law is only considered by the
courts if EU legislation intends to implement it domestically or if it expressly refers to
it. The EU judiciary has created a buffer: if a WTO decision does not find favour with
the European constituency as expressed through political channels, the courts will
disregard it. In many ways, this buffer re-enacts a dualist system, but because of its
specificity in the context of WTO regulation, it might well be seen as a mechanism for
ensuring that a particularly intrusive form of international regulatory action is
accountable to a European, a ‘national’, constituency.45
41
European Communities, First Written Submission in EC – Measures Affecting the Approval and Marketing
of Biotech Products, 17 May 2004, para. 10, available at http://trade-info.cec.eu.int/doclib/docs/2004/
june/tradoc_117687.pdf. See also the Press Release of 17 June 2003, available at http://trade-
info.cec.eu.int/doclib/docs/2003/november/tradoc_114665.pdf.
42
For such an interpretation see Kingsbury, Krisch, and Stewart, supra note 1, at 36–37; Cassese, ‘Global
Standards for National Administrative Procedure’, 68 Law and Contemporary Problems (2005) 109.
43
See Keohane, supra note 5, at 146–152. For a similar account, see Scott, ‘European Regulation of GMOs:
Thinking about “Judicial Review” in the WTO’, Jean Monnet Working Paper 04/04, available at
www.jeanmonnetprogram.org/papers/04/040401.html, at 12–13.
44
European Court of Justice, Judgment of 23 Nov. 1999 in Case C–149/96, Portugal v. Council [1999] ECR
I–08395, at paras 34–49; Judgment of 1 Mar. 2005 in Case C–377/02, Van Parys v. Belgische Interventie-
en Restitutiebureau, available at http://curia.eu.int/en/content/juris/index_form.htm at paras 38–54.
45
This approach is not confined to Europe: see Cottier and Nadakavukaren Schefer, ‘The Relationship
between World Trade Organization Law, National and Regional Law’, 1 J Int’l Economic L (1998) 83, at
106–110.
260 EJIL 17 (2006), 247–278
These mechanisms balance each other to some extent. They reject claims to hierar-
chical superiority by the other order, but do not exclude mutual influences and
approximation over time. Tracing the precise operation of such influences is difficult,
but one might interpret WTO Appellate Body attempts to – at least verbally46 – leave
the door open to precautionary approaches as an attempt to gain broader acceptance
and also to retain a marge de manoeuvre for future conflicts. Likewise, European courts
have given the ‘precautionary principle’ in EU law a restrictive interpretation that is
close to WTO rules, and WTO disciplines have had a major impact on recent EU legisla-
tion and regulatory reform in the area.47 The competing accountability mechanisms
and superiority claims of the different constituencies may thus not result in antago-
nism, but rather in mutual observation and gradual and pragmatic approximation.48
Likewise, interactions with other institutions can also be understood as driven by
claims to make the global regulation of GMOs accountable to further constituencies.
As mentioned above, the two additional institutions with a significant role in the
dispute are the Codex Alimentarius Commission (CAC) and the Biosafety Protocol.
The CAC49 is closely linked with the WTO as it is presumed that trade measures in
conformity with Codex Alimentarius standards are in compliance with the SPS
Agreement (even though here, too, there may exist a degree of flexibility for WTO
bodies as to the status they accord a given standard50). This brings a broader interna-
tional constituency into WTO regulation: not only does the CAC strengthen the influ-
ence of public health rather than trade expert communities, but it also represents a
significantly broader, almost universal membership and includes NGO observers
(even if the influence of developing countries and NGOs on CAC decision-making is
weak).51 The broader membership, however, has so far prevented a clear stance on
GMOs: since its membership is divided, it has only been able to adopt procedural rules
on the performance of risk assessment, while leaving the substantive questions
open.52 Yet the open expression of this division, coupled with some pragmatic steps
for narrowing the dispute, might already provide an important signal for regulation
in other fora, especially in the WTO.53
46
The outcomes of actual cases indicate a more restrictive approach: see Peel, supra note 35, at 53–86.
47
See Scott, ‘European Regulation of GMOs and the WTO’, 9 Columbia J European L (2003) 213, at
228–229, 232–233; Skogstad, ‘The WTO and Food Safety Regulatory Policy Innovation in the Euro-
pean Union’, 39 J Common Market Studies (2001) 485, at 497–500.
48
For a similar assessment, see Shaffer and Pollack, supra note 34, at 41–45.
49
For a brief overview of the work of the CAC, see Victor, supra note 36, at 885–892.
50
See Scott, ‘International Trade and Environmental Governance: Relating Rules (and Standards) in the
EU and the WTO’, 15 EJIL (2004) 307, at 331–333.
51
In Sept. 2005, the CAC had 171 member states: see http://www.codexalimentarius.net. On the limited
role of developing countries, see Chimni, ‘Cooption and Resistance: Two Faces of Global Administrative
Law’, IILJ Working Paper 2005/16, available at www.iilj.org/papers/2005.16Chimni.htm, at 12–17.
52
See Ostrovsky, ‘The New Codex Alimentarius Commission Standards for Food Created with Modern
Biotechnology: Implications for the EC GMO Framework’s Compliance with the SPS Agreement’, 25
Michigan J Int’l L (2004) 813, at 818–821; Poli, ‘The European Union and the Adoption of International
Food Standards within the Codex Alimentarius Commission’, 10 European LJ (2004) 613, at 626–629.
53
See also the positive evaluation by ibid., at 629–630.
The Biosafety Protocol,54 in contrast, has been established in part as a counter-

weight to WTO rules. Negotiated in the framework of the Biodiversity Convention
and thus a part of the Rio Process on environmental protection, it goes back to an
initiative to provide a stronger regulatory regime for trade with GMO products and
was driven by constituencies with little influence on WTO law-making. Especially
developing countries and environmental NGOs played a stronger role in the pro-
cess leading up to the Protocol, and the resulting rules reflect their views (and
those of Europeans) more strongly than the other international instruments.55 In
this sense, the Protocol can be seen as reflecting the views of a broader, more ‘cos-
mopolitan’ constituency, to stay within the terminology introduced above.56 It
places emphasis on the ‘Advance Informed Agreement’ of importing countries of
GMO products that are to be released into the environment57 and, besides its
requirement of a scientific risk assessment as a basis for the importing decision, the
Protocol makes several references to precautionary measures.58 It states in par-
ticular that a lack of scientific certainty shall not prevent a party from taking mea-
sures to avoid or minimize potential adverse effects of GMOs to be imported as food
or feed.59 It is thus relatively close to European approaches to GMOs and, as an
‘international standard’, it may have a similar impact on decision-making under
the SPS Agreement as Codex Alimentarius standards would have.60 Yet the ulti-
mate relationship with WTO law is unclear: after long negotiations, the Biosafety
Protocol contains two preambular clauses that emphasize that the Protocol is not
intended to change rights and obligations under other agreements, while also
insisting that this proviso is not meant to subordinate the Protocol to other agree-
ments.61 In the WTO dispute, the EU exploits the resulting indeterminacy and
insists that the Biosafety Protocol should heavily influence the interpretation of
WTO law on GMO issues.62
54
Cartagena Protocol on Biosafety, available at http://www.biodiv.org/biosafety/. On the protocol and its
negotiation, see M. Böckenförde, Grüne Gentechnik und Welthandel: Das Biosafety-Protokoll und seine
Auswirkungen auf das Regime der WTO (2004), at 118–240; C. Bail, R. Falkner, and H. Marquard (eds),
The Cartagena Protocol on Biosafety: Reconciling Trade and Biotechnology with Environment and Development
(2002).
55
See Stewart, supra note 33, at 56–57; Falkner, ‘Regulating Biotech Trade: the Cartagena Protocol on
Biosafety’, 76 Int’l Affairs (2000) 299, at 301–302, 313; Böckenförde, supra note 54, at 140–144.
56
The importance of the forum and its composition was highlighted, e.g., by the vigorous opposition of
developing countries and the EU to the creation of a WTO Working Group on biotechnology parallel to
the negotiations on the Biosafety Protocol; see Falkner, supra note 55, at 305.
57
Art. 7 of the Biosafety Protocol.
58
See especially the Preamble to and Art. 1 of the Biosafety Protocol.
59
Art. 11(8) of the Biosafety Protocol.
60
See Howse and Mavroidis, supra note 39, at 354–370; see also Boisson de Chazournes and Mbengue,
‘GMOs and Trade: Issues at Stake in the EC Biotech Dispute’, 13 Review of EC and Int’l Environmental Law
(2004) 289, at 297–303; Böckenförde, supra note 54, at 333–336.
61
See Safrin, ‘Treaties in Collision? The Biosafety Protocol and the World Trade Organization Agreements’,
96 AJIL (2002) 606, at 614–628; also Falkner, supra note 55, at 309–310.
62
See EC, First Written Submission, supra note 41, at paras 453–459.
262 EJIL 17 (2006), 247–278
C Why Competing Accountability Mechanisms?

The CAC and the Biosafety Protocol as counters to WTO dynamics, and the EU limita-
tions on direct applicability of WTO norms and decisions in EU law, all have account-
ability functions vis-à-vis the WTO. They are not, of course, classical accountability
mechanisms as we know them from domestic administrative law; they do not operate
inside a given regime, as participation or review mechanisms, but outside, trying to
influence the regime through different, partly political means. Yet they fulfil functions
very similar to those classical mechanisms, as they make regulatory processes
respond to particular concerns and constituencies and provide checks against regula-
tory excesses; we might thus classify them as ‘non-institutionalized’ accountability
mechanisms.63
To some extent, these mechanisms help to give a voice to constituencies that are
under-represented in formal procedures, as for example developing countries and
NGOs in the WTO. By creating an alternative regulatory site, the Biosafety Protocol
provides an external challenge to such exclusion, which may over time lead to
changes in WTO mechanisms and a reinterpretation of WTO norms, but which even
so may persist as the institutionalized expression of a particular, more environmentally-
minded approach to GMOs that would be diluted in fora that give trade-oriented
perspectives a greater role. In a ‘regime complex’,64 rather than in a single regime, a
multiplicity of voices can find institutional homes.65
Yet in other respects, these irregular accountability mechanisms give voice to con-
stituencies that are well represented in formal mechanisms. The EU was a central
player in negotiating the WTO agreements,66 continues to play a central role in WTO
committees and reform negotiations,67 is increasingly active in the CAC,68 and does,
of course, also enjoy participation rights in disputes before WTO bodies. Here, it can
express its views through its submissions as well as review a panel decision by taking
the case to the Appellate Body. Yet these formal, classical accountability mechanisms
must appear insufficient if, from an EU perspective, it is the European polity, not an
international constituency, that ought to decide on GMO regulation in Europe. In
fact, every participation right short of a right to veto would be insufficient, and veto
rights are increasingly impractical in global regulatory structures that are intended
to work quickly and effectively. In this case, ‘irregular’ mechanisms serve to express
the views of the European polity and limit the effect of decisions that are mainly
63
See Keohane, supra note 5, 139–140.
64
See Raustiala and Victor, ‘The Regime Complex for Plant Genetic Resources’, 58 Int’l Org (2004) 277, at
279, on the concept of ‘regime complex’.
65
For a similar argument, see Koskenniemi and Leino, ‘Fragmentation of International Law? Postmodern
Anxieties’, 15 Leiden J Int’l L (2002) 553; and, in the trade/environment context, O. Perez, Ecological
Sensitivity and Global Legal Pluralism: Rethinking the Trade and Environment Complex (2004), at 259–262.
66
On EU involvement in the negotiation of the SPS Agreement, see Skogstad, supra note 47, at 492–494.
67
See Vogel, ‘Trade and the Environment in the Global Economy: Contrasting European and American
Perspectives’, in N. Vig and M. Faure (eds), Green Giants? Environmental Policies of the United States and the
European Union (2004), at 231.
68
See Poli, supra note 52.
accountable to an international constituency. The buffers the EU has erected against

the direct impact of WTO rules and decisions in its own legal order are one such
mechanism.
Thus, from one perspective, the EU stance may look very much like standard non-
compliance with international law, but from another, it appears as part of a contest of
different constituencies in global regulatory governance. Through the different, not
formally connected and thus ‘irregular’ mechanisms sketched here, national and cos-
mopolitan views check a regulatory process through which an international com-
munity seeks to hold national regulatory action accountable. All of them enter this
process as irritation: from each regime’s perspective, the others are formally at best of
limited relevance.69 This process has the advantage of giving voice to different constit-
uencies, none of which accepts the others as fully determining the issue; and as we
have seen, it leads to mutual influences and pragmatic approximation and might
thus help keep regulatory outcomes in a balance between several constituencies. But
it has, of course, the disadvantage that it may often not reach stable results; that reg-
ulatory outcomes are fluid and provisional rather than embodied in one final and
clear decision. I will return to these and other problems later. We should note, how-
ever, that such instability will not reflect the normal, day-to-day operation of a
regime complex such as this one: in most situations, coordinated and smooth action
will be much more prominent.70 Decisions will become unclear and unstable once
they concern issues of high salience. Yet in a highly contested environment, this
might be better than pressing for a final decision for which the decision-making body
lacks full legitimacy in the eyes of central addressees.71
4 Accommodating Multiple Constituencies: A Pluralist

Approach
In the eyes of many, a pluralist order like the one just sketched with respect to the
GMO dispute will appear highly unsatisfactory. The unclear distribution of compe-
tences, the instability of decisions and the highly political nature of accountability
mechanisms will be seen as lacking in many respects, and not only because they
promise serious difficulties for quick and reliable decision-making procedures. Also
from a normative standpoint, such an order will seem problematic since it is unable to
reflect a normatively justified distribution of powers among the different levels. If one
believes, for example, that the size of the polity and the shape of public institutions
should be determined by the extent of obligations of justice,72 then one might want a
69
For a fruitful (yet exaggerated) description of such mechanisms in terms of Luhmannian systems theory,
see Fischer-Lescano and Teubner, ‘Regime Collisions: The Vain Search for Legal Unity in the Fragmenta-
tion of Global Law’, 25 Michigan J Int’l L (2004) 999.
70
On the relatively standard implementation of CAC standards by the EU, e.g., see Poli, supra note 52, at
616–617. On the oscillation between smooth and fracturous operation of the regime complex on plant
genetic resources, see Raustiala and Victor, supra note 64.
71
See also Stewart, supra note 33, at 66–67, 108–110.
72
See, e.g., Young, supra note 26, at 250.
264 EJIL 17 (2006), 247–278
cosmopolitan constituency to prevail on issues of global environmental impact; a plu-

ralist order that keeps hierarchies in the balance will then seem hardly attractive.
From such a perspective, it would thus appear preferable to organize accountability
in global governance according to clearer and normatively defensible principles and
thus, following constitutionalist lines, with an aspiration to coherence and unity. The
most obvious institutional model would then be a federal order in which particular
powers would be assigned to the different levels. Each constituency would then con-
trol decisions on the matters assigned to it, subject to some hearing and participation
rights of other constituencies.73
The viability of such a solution, however, would be highly doubtful. In a situation
of fundamental contestation over who should have the final say, any institutional
order that responds primarily to one of the different constituencies is likely to lack
legitimacy and will thus be unable to produce lasting and stable decisions. As long as
the contestation persists, alternative institutional mechanisms will therefore be
necessary.74 This situation is similar to that of plural, divided societies in which deci-
sions by a majority group are likely to destabilize the polity and undermine the
authority of the institutions. For such societies, ‘consociational’ models have often
been found to yield superior results; and they might also help devise strategies to
accommodate multiple constituencies in global regulatory governance.75
A The Consociationalist Road and Its Problems

Consociationalist approaches, unlike majoritarian democracy, are characterized by
an insistence on accommodation: prominent in a number of smaller European
countries, especially in the post-war period, they seek to manage deep disagreement
by grand coalitions and the creation of veto positions that force all actors to reach
common solutions rather than imposing their views on the others; none of the con-
stituencies enjoys primacy. According to its defenders, in particular Arend Lijphart,
consociationalism is much more likely than majoritarian systems to create a stable
political order under conditions of diversity.76 Whether or not this view is based on
sound empirical evidence is contested,77 but many indicators do indeed point to stabi-
lizing results, and the main thrust of the argument bears sufficient plausibility. And in
73
See Held, supra note 26, at 235–237; Young, supra note 26, at 267; Kumm, supra note 26, at 920–924;
O. Höffe, Demokratie im Zeitalter der Globalisierung (1999), ch. 10; see similarly J. Habermas, Der
gespaltene Westen (2004), at 131–145.
74
Historically, such contestation (of national-level authority) has often been overcome through ‘nation-
building’ and nationalist ideologies, but at a high cost. See Connor, ‘Nation-Building or Nation-Destroying?’,
24 World Politics (1972) 319.
75
Federal and consociational models are not necessarily opposed; in fact, they are often linked: see
A. Lijphart, Democracy in Plural Societies (1977), at 42. As will become clearer below, I distinguish here
between federal orders in which powers are assigned to specific levels, and consociational orders in
which they are shared.
76
See ibid., passim.
77
See Barry, ‘Political Accommodation and Consociational Democracy’, 5 British J Political Science (1975)
477; Lusztick, ‘Lijphart, Lakatos, and Consociationalism’, 50 World Politics (1997) 88.
fact, the international order in its classical mould deals with diversity in a very similar
way78: decisions are usually taken by unanimity, and every state enjoys not only
strong autonomy rights but also a veto position with respect to international obliga-
tions.79
For our purposes, however, consociationalism is inapplicable in this classical mode:
dealing mostly with differences on a horizontal level – between different groups in
society – it cannot be applied immediately to the vertical frictions between different
levels that are the focus of our attention. Classical international law is based on the
primacy of the national constituency, which, as we have seen, is increasingly the
object of dispute. Yet the central consociationalist idea, namely the importance of a
co-decision of the different constituencies involved, is useful for our vertical constella-
tion, too; and it is also reflected in those strands of the theory and practice of federal-
ism that stress vertical interlinkages between the federal and state levels, such as in
Germany.80 Thus, adapted to the circumstances of global governance, a consocia-
tionalist approach would insist on the co-decision of the different constituencies
involved: all of them – the national, international and cosmopolitan constituencies –
would have to have a voice, and a right to veto, in global regulation.
Such an approach, however, would have obvious limits. First, like any consocia-
tional setting, it would face serious problems of practicability in quick and effective
decision-making. Veto positions for the different constituencies may work reasonably
well for small numbers; with higher numbers, the problems of negotiation and the
risk of blockade rise rapidly, and dangers of a ‘joint-decision trap’ emerge.81 The effi-
ciency problems that plague consociational approaches on the domestic level would
be exacerbated in a global context. This, in turn, would likely cause a serious norma-
tive bias: either in favour of non-regulation (if this were the default rule governing
cases in which a consensual decision cannot be reached) or in favour of national reg-
ulation (if this remained the default rule of global affairs, as it is mostly now, with
some – increasingly important – exceptions as can be found in WTO law82). Either
way, the result would probably be undesirable.
Secondly, co-decision in a consociational vein would only be of limited use for
ensuring stronger accountability not only to governments but also to publics. One of
78
Lijphart, supra note 75, at 45. See also the important critique by Donald Horowitz, e.g. in ‘Constitutional
Design: Proposals Versus Processes’, in A. Reynolds (ed.), The Architecture of Democracy: Constitutional
Design, Conflict Management, and Democracy (2002), at 15, 19–25, as well as Lijphart’s rejoinder in ‘The
Wave of Power-Sharing Democracy’, in ibid., at 37, 43–44, 47–49.
79
This holds true even for most newer forms of secondary law-making, at least from a formal perspective.
See Brunnée, ‘COPing with Consent: Law-Making Under Multilateral Environmental Agreements’, 15
Leiden J Int’l L (2002) 1, at 15–32.
80
See Scharpf, ‘Die Politikverflechtungsfalle: Europäische Integration und deutscher Föderalismus im
Vergleich’, 26 Politische Vierteljahresschrift (1985) 323, at 324–331; S. Oeter, Integration und Subsidi-
arität im deutschen Bundesstaatsrecht (1998).
81
Scharpf, supra note 80, at 346–350. Accordingly, also for Lijphart consociational orders ideally operate
with no more than four main groups: see Lijphart, supra note 75, at 56.
82
As we have seen in the GMO example, in some areas the WTO severely limits national regulation in the
absence of a corresponding international standard.
266 EJIL 17 (2006), 247–278
the central elements of consociationalism is its reliance on the cooperation of elites:

because on many issues genuine consensus among the different groups will be elu-
sive, problem-solving requires bargaining, package-deals, logrolling. This, however,
can only be achieved by elites that stand in constant contact with each other and are
socialized into cooperation. Stronger participation of the general public in the various
groups renders this cooperation difficult because it is usually focused only on a par-
ticular decision, not the whole of the deal struck. Accordingly, as Lijphart stresses
himself, ‘it is ... helpful if [leaders] possess considerable independent power and a
secure position of leadership’.83 Even though this is not incompatible with public par-
ticipation in general, it considerably limits its scope. Moreover, the introduction of
further accountability mechanisms into the already difficult framework of negotia-
tions on the global level would only aggravate the risk of a complete blockade of
decision-making.
B From Co-Decision to Mutual Challenge

Co-decision will thus often not be a preferable option, and it also seems to fall short of
accommodating the different views on the right constituency. Yet, given the inherent
tensions between these views, it is probably impossible to achieve their full integra-
tion in any institutional structure. In order to nonetheless respect them all, we might
thus choose a path that aims not so much at integration but at dissociation: one that
keeps a distance from the ideals of all of them.
Any such path will, in order to avoid the problems outlined above, refrain from
according full control over decisions – through veto rights or otherwise – to either of
the competing constituencies, and it will recognize that each of the constituencies will
have to be involved in decision-making processes through multiple accountability
mechanisms rather than merely through one precisely defined procedure. At the
same time, it will not be sufficient to allow for their participation just through hear-
ings and submissions. Such procedural participation is important in order to canalize
and publicize the standpoints of the different constituencies, but it is too weak to bal-
ance the influence of the constituency to which the decision-maker is most directly
accountable. Merely allowing statements by NGOs, for instance, will hardly recali-
brate a decision-making process centred on state representatives.
Steering a path between co-decision and the deficiencies of mere hearings thus
seems to call for participation of each constituency in the form of some, but not full,
decision-making powers: to a pluralization rather than concentration of authority. If
all constituencies are to have decision-making powers beyond merely being listened
to, but shall not be able to dictate or veto a particular decision, then no decision can
fully bind them all, and each constituency has to retain the right to challenge
(though not to veto) it. The resulting picture of global governance would then be one
of a constant potential for mutual challenge: of decisions with limited authority that
may be contested through diverse channels until some (perhaps provisional) closure
83
Lijphart, supra note 75, at 50.
might be achieved. The GMO case described above provides an example of how such a
regime complex can work in practice. But it also shows that in such a pluralist order
contestation does not necessarily form the normal mode of operation: as indicated
above, challenges are likely to be the exception in particularly salient cases, whereas
the day-to-day functioning will be dominated by smooth cooperation, compromise
and mutual accommodation.
Such a pluralist picture would be quite far from a classical, constitutionalist
approach to legal and political ordering, but it is in fact relatively close to current reali-
ties of global governance well beyond the food safety regime complex exemplified by
the GMO case. The model example is certainly the interplay between national consti-
tutional courts and the European Court of Justice (ECJ).84 Beginning in the 1970s, the
highest courts of several Member States of the European Communities questioned the
competence of the ECJ to ultimately decide on contested issues. Their focus first con-
cerned human rights, and in response to their assertion of final authority on human
rights protection, the ECJ readjusted its jurisprudence on that matter, without, how-
ever, ceding ground on the more fundamental question of final authority. Likewise, in
the 1990s, Member State courts reclaimed the authority to decide on issues of the
delimitation of powers between the national and the European levels, thereby again
forcing the ECJ to take this issue more seriously.85 As mentioned above, the German
Constitutional Court based its challenge explicitly on considerations of democracy,
insisting on the superiority of the national over the European constituency.86 None of
these cases, though, eventually produced serious friction: national courts mostly sig-
nalled potential non-compliance in the future but did not follow through with it, due
largely to the pragmatic steps of European authorities to accommodate their concerns.
Thus, the fundamental contestation in this case helped recalibrate the overall system,
leaving open the question of ultimate authority and thus balancing accountability to
the different constituencies87 while allowing for smooth functioning most of the time.88
On the global level, we can see parallels, for instance, in the use of national courts to
hold Security Council sanctions decisions accountable to individuals.89 Since the late
84
For an analysis from the perspective of a changed understanding of constitutionalism (or an alternative
to it), see Walker, supra note 17; Weiler, ‘In Defence of the Status Quo: Europe’s Constitutional Sonder-
weg’, in J.H.H. Weiler and M. Wind (eds), European Constitutionalism Beyond the State (2003), at 7;
Maduro, ‘Europe and the Constitution: What If This is as Good as it Gets?’, in ibid., at 74; Kumm, ‘The
Jurisprudence of Constitutional Conflict: Constitutional Supremacy in Europe before and after the Con-
stitutional Treaty’, 11 European LJ (2005) 262; see also Krisch, ‘Europe’s Constitutional Monstrosity’, 25
Oxford J Legal Studies (2005) 321.
85
See A.-M. Slaughter, A. Stone Sweet, and J.H.H. Weiler (eds), The European Court and National Courts –
Doctrine and Jurisprudence (1998); F. Mayer, Kompetenzüberschreitung und Letztentscheidung (2000).
86
See supra note 22.
87
See also Maduro, supra note 84, at 81–86.
88
See also Oeter, ‘Föderalismus’, in A. von Bogdandy (ed.), Europäisches Verfassungsrecht (2003) at 59,
110–119.
89
On the general potential of extending the argument of domestic review of international organizations
beyond the European context, see Walter, ‘Grundrechtsschutz gegen Hoheitsakte internationaler
Organisationen’, 129 Archiv des öffentlichen Rechts (2004) 39.
268 EJIL 17 (2006), 247–278
1990s, the Council has increasingly targeted individuals directly, imposing travel and
financial restrictions on them, without, however, providing them with procedural
rights or possibilities of judicial review. As a result, in a number of cases national courts
have been seized by those affected by sanctions.90 In one of them, several Swedish
nationals of Somali descent challenged measures implementing restrictions imposed by
the Council before the European Court of First Instance (ECFI). Soon after the first steps
in the proceedings (and after significant diplomatic pressure), the Security Council
Sanctions Committee reconsidered the issue, striking two of the applicants from its list
of targeted individuals and, more importantly, establishing a general procedure for
revising the list upon request by a government (with only indirect participation by the
individuals concerned, however).91 These procedural improvements were central for
the ECFI’s eventual rejection of the applicant’s case on the merits. However, the Court
did not abdicate review of Security Council action, as had been sought by the EU institu-
tions and some Member States. Instead, it emphasized its own power to assess Security
Council action in light of peremptory norms of international human rights law, thus
signalling that it could in the future police the limits of international sanctions.92
Thereby, it would insist on enhanced accountability, not only to the affected individu-
als, but also to the European constituency more broadly, as Security Council decisions
would be judged on the basis of a European interpretation of those international norms.
In a similar vein, though potentially with even broader impact, the European Court of
Human Rights has reviewed the implementation of Security Council sanctions by EU
and national measures on the basis of the European Convention of Human Rights.93
The court has so far not assessed Security Council procedures as such, but the logic of its
reasoning could be extended to include such an assessment in the future. Thereby, glo-
bal security regulation could be tied back not only to international human rights stan-
dards, but even to standards established by a European constituency. Yet, despite these
reservations in principle, serious friction is unlikely in practice; in most cases, the pos-
sibility of challenge will suffice to produce pragmatic accommodation.94
90
See Gutherie, ‘Security Council Sanctions and the Protection of Individual Rights’, 60 Annual Survey of
American Law (2004) 491; Dyzenhaus, ‘The Rule of (Administrative) Law in International Law’, 68 Law
and Contemporary Problems (2005) 127, at 140–150.
91
See Cramér, ‘Recent Swedish Experiences with Targeted UN Sanctions: The Erosion of Trust in the Secu-
rity Council’, in E. de Wet and A. Nollkaemper (eds), Review of the Security Council by Member States
(2003), at 85, 90–97. Three months before the Committee decision, the CFI had rejected the request for
provisional measures but reserved a decision on the merits: see the order of the President of the CFI of
7 May 2002 in Case T–306/01R, Aden et al. v. Council and Commission [2002] ECR II–02387.
92
See the judgment of the CFI of 21 Sept. 2005 in Case T–306/01, Yusuf and Al-Barakaat v. Council and Com-
mission, available at http://curia.eu.int/en/content/juris/index_form.htm. It is under appeal to the ECJ.
93
See the judgment of the ECtHR of 30 June 2005 in App. No. 45036/98, Bosphorus Hava Yollari Turizm v.
Ireland, available at http://cmiskp.echr.coe.int.
94
On the general reserved deferral of domestic courts to UN sanctions and their implementation, see
Gowlland-Debbas, ‘Implementing Sanctions Resolutions in Domestic Law’, in V. Gowlland-Debbas (ed.),
National Implementation of United Nations Sanctions: A Comparative Study (2004), at 33, 55–65. On paral-
lel cases, see also Lammers, ‘Challenging the Establishment of the ICTY before the Dutch Courts’, in de
Wet and Nollkaemper, supra note 91, at 107; Marko, ‘Challenging the Authority of the UN High Repre-
sentative before the Constitutional Court of Bosnia and Herzegovina’, in ibid., at 113.
Similar pluralist orders, in which different constituencies balance each other

through mutual challenges, can be observed in many other areas of global regulation
too, ranging from national checks on global anti-doping decisions to European
attempts to shift the accountability structure of the international accounting stan-
dards regime.95 Likewise, the turn towards soft law and government networks in glo-
bal regulation96 can be understood as reflecting a need to balance different
constituencies: by limiting claims to international authority, it allows for an interplay
between the different levels which both cooperation and contestation are already
part of.97
As is evident from these few examples, pluralist accountability mechanisms are a
pervasive feature of global regulatory governance. We can explain and to some
extent justify them as a response to the contested constituencies of global regulation
that seeks to avoid the downsides of federal and consociational models of a more uni-
tary, constitutionalist character. Pluralist mechanisms certainly have downsides too
(I will consider some of them below), but on closer inspection, they are neither as
anomalous nor as impractical as they may seem at first sight.
5 A Principled Case for a Pluralist Order

So far, I have defended a pluralist global administrative law mostly on prudential
grounds: it seems superior to more unitary – especially federal and consociational –
models of order in terms of its capacity to accommodate different views about the right
constituency of global governance, and thus in its ability to create a relatively stable,
and reasonably effective, institutional order. Yet, compared with more coherent, con-
stitutionalist models, this approach has seemed to lack appeal because of its inability to
reflect a distribution of powers according to fully justified principles.98 As I will try to
show in this section, this contrast between prudence and principle is overdrawn. There
is actually a strong case to be made for a pluralist order on principled grounds.
A The Constituencies’ Equal Deficits

This case can rest, in the first place, on the weakness of the supremacy claims of all
constituencies, grounded in particular in the normative deficits of each of them.99
95
See Van Vaerenbergh, ‘Regulatory Features and Administrative Law Dimensions of the Olympic Move-
ment’s Anti-doping Regime’, IILJ Working Paper 2005/11, available at www.iilj.org/papers/
IILJ2005.11Vaerenbergh.htm, at 17–21; Mattli and Büthe, supra note 30, at 257–258. Other examples
include conflicts over the regulation of water services: see Morgan, ‘Turning Off the Tap: Urban Water
Service Delivery and the Social Construction of Global Administrative Law’, in this issue; or, in a different
mode, mutual challenges between certification regimes in sustainable forestry: see Meidinger, ‘The Admin-
istrative Law of Global Private-Public Regulation: The Case of Forestry’, in this issue.
96
See D. Shelton (ed.), Commitment and Compliance: The Role of Non-Binding Norms in the International Legal
System (2000); Slaughter, supra note 24.
97
See Barr and Miller, ‘Global Administrative Law: The View from Basel’, in this issue.
98
See supra, text at notes 72, 73.
99
For a similar argument with respect to the EU, see Maduro, supra note 84, at 82–86.
270 EJIL 17 (2006), 247–278
The national constituency is limited in that it cannot fully respond to the needs and
interests of those outsiders that are affected by its decisions or that have a claim to be
considered, for example for reasons of transboundary justice. The international con-
stituency is not capable of instituting structures of democratic participation that are
nearly as thick and effective as those possible on the national level; it is too far
removed from individuals, and intergovernmental negotiations will never come with
the deliberative structures necessary for real public involvement. In a cosmopolitan
constituency, too, we face serious limits of communication across cultural, linguistic
and political boundaries. Making the international or cosmopolitan constituency
supreme would also neglect the strong attachments of individuals to national (or sub-
national) groups and the importance this has for their identity.100 Consequently,
since none of the constituencies can make a convincing claim for primacy, we should
regard them as complementary and recognize that they stand in a non-hierarchical
relationship. Given the impracticability of consociational arrangements, such a sta-
tus will best be reflected in a pluralist order as I have sketched it above.
Such an argument is, however, not without problems. One could allege, for
example, that it ignores the differences in weight of these deficits in different issue
areas. In some areas, externalities of national action may be high, and thus one may
want to give priority to the international constituency. In others, they will be low,
and the national constituency will have greater weight. Instead of a pluralist order
we would then arrive, again, at a federal-style order with a distribution of powers
among the various levels; which level to grant priority in a given area would be
merely a question of balancing the different concerns.
This objection is of course to some extent valid: nobody thinks of establishing inter-
national or cosmopolitan control over issues such as regulating homes for the elderly
or curricula for elementary schools.101 Yet there is a wide range of issues on which
solutions are far less evident; and in many cases, issue areas will be meshed, as we can
see in the GMO dispute where trade, environmental and broader cultural concerns
are closely linked. Here, the federal-type vision reaches its limits, and the right con-
stituency is, for good reason, heavily contested.
B Equal Respect for Constituency Choices

A second, and probably stronger, defence of a pluralist order could start from this fact of
reasonable contestation and base itself on an argument from liberty and autonomy.102
100
See also the arguments presented supra, in sect. 2 C. Here, as there, I do not present a fully theorized
argument, based on a particular normative vision, but rather an approach that might resonate with a
number of different broader normative perspectives.
101
Although, in the latter case, the OECD’s efforts in the area of education may reflect inroads into the
national domain.
102
Basing the international order on such normative grounds is, of course, in itself problematic as it stands
in tension with the strong value pluralism of international society. I cannot deal with this problem here,
but the emphasis placed here on actual preferences may alleviate the general concern. On these prob-
lems with respect to global administrative law in general, see Kingsbury, Krisch, and Stewart, supra note
1, at 42–52.
Insofar as autonomy is linked to political self-government, it includes a right to parti-

cipate in decisions on the constitutional framework of the political order; and this also
has to encompass a right to participate in decisions on the shape and extent of the
political group in which self-government is exercised and to whose decisions individ-
uals are subject.103 If individuals choose different and conflicting political groups, an
order that seeks to equally respect their choices cannot grant any group primacy.
And in our case, again, since non-hierarchical relationships among constituencies
cannot be organized through consociational means, they will best find expression in a
pluralist order.
The basic idea behind this argument is close to political theories based on the
freedom of association. An early influential strand of this kind was English political
pluralism, associated especially with Frederick William Maitland, G.D.H. Cole,
John Neville Figgis, and Harold Laski.104 For them, a political order based on volun-
tary associations appeared superior to a state-centred one because it promised indi-
viduals greater influence in their affairs. Because they originated in individual choice,
such associations were, in their view, also independent from the state in their basis of
legitimacy and possessed non-derived powers. Laski, in some of his works, took this so
far as to assert that the state was in effect just another association, with no a priori
claim to supremacy and dependent on acceptance by other associations and individu-
als whenever it sought to act on them.105 Yet despite their general emphasis on the
importance of associations, most English pluralists, including Laski in his most influ-
ential writings, accepted a superior role of the state as a guardian of the system: as a
guarantor of the freedom of association, as an enforcer of common norms, and as an
arbiter between associations.106 This general line is also followed by contemporary
theorists close to this tradition, in particular Paul Hirst and William Galston. Both
emphasize the centrality of societal associations but regard them as embedded in an
overarching order that is upheld by state institutions, however minimal their func-
tions.107 In effect, this recreates a federal order with a strong subsidiarity norm, albeit
one that is organized along personal rather than territorial lines.
103
E.g., David Held defines autonomy as follows: ‘persons should enjoy equal rights and, accordingly, equal
obligations in the specification of the political framework which generates and limits the opportunities
available to them; ...’: D. Held, Models of Democracy (2nd edn., 1996), at 301. Much of Held’s further
specification of the principle, also with respect to global contexts, assumes particular political frame-
works (along the lines of existing power-wielding institutions). But if participation is understood as an
individual right (rather than the collective right of a given community), the general principle is easily
applied also to the definition of the framework itself. In a related vein, Thomas Franck sees the increasing
recognition of individuals’ nationality choices as a reflection of their basis in autonomy: see T.M. Franck,
The Empowered Self: Law and Society in the Age of Individualism (1999), at ch. 4.
104
See P.Q. Hirst (ed.), The Pluralist Theory of the State (1989), at 1–47; D. Nicholls, The Pluralist State: The
Political Ideas of J.N. Figgis and his Contemporaries (2nd edn., 1994); also D. Runciman, Pluralism and the
Personality of the State (1997).
105
See Laski, ‘Law and the State’, in Hirst, supra note 104, at 197, 214; also Hirst, in ibid., at 28.
106
See Nicholls, supra note 104, at ch. 5; Hirst, supra note 104, at 28–30; Laski, ‘The Problem of Adminis-
trative Areas’, in ibid., at 131, 155.
107
See P.Q. Hirst, Associative Democracy (1994), at 47, 56–61; W.A. Galston, Liberal Pluralism (2002), at
ch. 9; in a similar vein, see Young, supra note 26, at 180–195.
272 EJIL 17 (2006), 247–278
A similar emphasis on freedom of association underlies Chandran Kukathas’

recent work, but here we can find hints at more radical conclusions.108 In his vision,
society is an ‘archipelago’ of (partly overlapping) associations that coexist both next
to each other and on different levels, but not in hierarchical relationships: all depend
on negotiations and compromises with the others; none can command; and the basic
operational principle is toleration. In this order, the state occupies an elevated place
but is confined to an even more minimal role than in the approaches mentioned
above. It is supposed to ensure order as an ‘umpire’ between associations, but ques-
tions of justice are out of its reach since they are contested among different associa-
tions and no neutral ground can be found to adjudicate them. What is just and right
may therefore remain undecided; competing views will seek to broaden their support
but cannot be enforced against associations that are unwilling to share them.109
Kukathas’ approach assumes a society made up of relatively unconnected groups –
‘islands’ separated by the sea – which can govern many of their affairs without much
effect on others; in that, it differs significantly from the situation we are facing in glo-
bal regulatory governance.110 Yet the basic normative idea holds much force also in
these circumstances. It reclaims respect for the individual’s exercise of her freedom of
association through determining the group that forms the framework for political
action. And it follows through with this respect by according all associations equal
status in matters of justice; given their equal basis in individual freedom, none of
them can claim superiority in this regard. For Kukathas, this results in toleration
between groups that stand (mostly) in a horizontal relationship and are often rela-
tively unconnected. In our circumstances, it would result in heterarchy and mutual
challenge between the three vertically related, and closely linked, constituencies: the
national, international, and cosmopolitan ones.
C The Range of Relevant Constituencies

If we start from this basis, though, it seems difficult to sustain limits on the range of
eligible constituencies. We might thus have to extend their circle beyond the
national, international and cosmopolitan groups to other constituencies that reflect
individual preferences, such as, for example, the funders of a given institution. These
groups would compete with the national, international and cosmopolitan constituen-
cies; none of them would a priori emerge as supreme; and control over outcomes
might eventually depend mostly on the power play between them.
From a perspective such as Kukathas’, this result would be inevitable; he does not
impose limits on the creation of associations or their equal recognition. Yet this posi-
tion is founded on a particular conception of freedom; and if we adopt an alternative,
108
C. Kukathas, The Liberal Archipelago (2003). Kukathas bases freedom of association not on autonomy but
on freedom of conscience (ibid., at 36–37); but this difference is of little importance in the present context.
109
See ibid., at ch. 6, and especially at 252 (‘[t]he state should not be concerned about anything except
order or peace’).
110
The difference escapes Kukathas who sees international society as remaining in a mostly coordinatory
mode: see ibid., at 28.
more substantive conception, limits on the range of relevant constituencies will

follow much more naturally. Kukathas’ approach is based on a minimalist, libertarian
idea of freedom that deliberately excludes considerations of actual effectiveness. For
him, freedom is essentially a formal notion, and this is reflected in many elements of
his political vision, most visibly perhaps in the weak requirements to make the right
to exit from a group effective.111 Yet this starting-point is not the only conceivable
one. Critics like Will Kymlicka would emphasize effective autonomy rather than
formal freedom of conscience as a basis for ascribing rights to groups; and as a result,
the circle of groups that are granted such rights is limited.112 Likewise, in our context,
it might seem more attractive to ground freedom of association in a more substantive
idea of freedom, for example in one that focuses on political autonomy or non-
domination.113 In this case, the range of relevant choices will be more limited – such
choices that, in effect, serve to impose rules and decisions on others, will be excluded
from equal respect; the eligible associations will have to be sufficiently inclusive. Just
as domestic institutions are no longer formally accountable primarily to the rich, the
constituencies of global regulatory institutions will have to reflect some degree of
effective freedom for all.114
This normative ‘pre-screening’ of eligible constituencies may seem like a negation
of the central role of individual choice in determining the scope of the polity, but it
really is an attempt to strike a balance between elements of justice and choice, as we
find it in political orders in general. Even if we generally respect political choices, as
we do in domestic democratic orders, we do not do so without limits. We shape them
through democratic procedures and limit them through substantive constitutional
rules, such as individual rights. Even more so, we do not accept all choices about the
political process itself; instead we recognize only such choices that fall within a range
of normatively defensible options and we reject in particular those that undermine
the basic principle of equal participation of all. Participation in shaping the frame-
work of politics is important, but it is not the only relevant factor.115
The same holds true for the determination of the size of the polity – of the constituency
of the political order. During the era of the nation-state, this determination seemed
self-evident, but once the question is posed explicitly, the democratic procedures as
we know them domestically are of no help in answering it – they already presuppose
111
See ibid., at 93–114.
112
W. Kymlicka, Multicultural Citizenship (1995), at ch. 8; see also Kymlicka, ‘The Rights of Minority
Cultures: Reply to Kukathas’, 20 Political Theory (1992) 140, at 142–143.
113
See Held, supra note 103, at 299–305; P. Pettit, Republicanism (1997).
114
This does not imply that national constituencies that do not follow the model of liberal democracy would
be excluded as relevant constituencies. Compared to actual alternatives, especially foreign domination,
even authoritarian regimes will often have a better claim to ensure non-domination of their citizens.
115
Another way of expressing this idea is to emphasize the simultaneous origin of private and public auton-
omy and the mutual constitution of rights and democracy: see J. Habermas, Faktizität und Geltung (1992),
at ch. III. See also J. Waldron, Law and Disagreement (1999), at chs 11 and 13, for whom participation is
central, yet only in so far as it concerns matters of ‘good-faith disagreement’ and thus presumably not
about the question whether everybody’s participation should matter equally in principle: ibid., at 303–304.
The question who decides what is ‘good-faith disagreement’ or ‘justice’ remains, of course, open here.
274 EJIL 17 (2006), 247–278
the primacy of a particular polity: the national one.116 Therefore, we should answer it
in the same way as we answer other questions about the shape of the political pro-
cess: in both cases, actual political choice and conceptions of a just order have to be
brought into balance. By introducing normative constraints on the range of eligible
groups, we seek to achieve that balance; we seek, in particular, to ensure that the
exercise of choice does not violate the very basis on which it rests: the idea of effective
freedom through self-government.
The three main contenders for primary constituencies here – the national, the
international, and the cosmopolitan ones – all come with a strong and defensible nor-
mative basis.117 However, admitting others, such as the constituency of funders,
might put equal self-government at risk if it granted non-inclusive groups a central
role. This does not mean that none other than the three constituencies mentioned
would be acceptable. Yet all of them would have to meet prior normative standards;
their recognition does not follow merely from the fact that they are chosen by some
person.118 Unless one sees freedom best realized in a libertarian world, certain con-
straints on the eligible constituencies will be necessary and justified.
In our case, if we accept the basic normative strength of the claims of the national,
international and cosmopolitan groups to be constituencies of global regulatory gov-
ernance, we should deny any of them formal primacy; they can all make a valid claim
to hold global regulatory governance accountable. Their relative strength will then
result not from a predetermined hierarchy, but from their influence, their allegiances
and support, which will be determined in the political processes of a pluralist order.
This has the added advantage of allowing for shifting weights of the different constitu-
encies in the fluid process of constructing global governance; it opens up space for
politics and for attempts at radical transformation that might be barred by rigid insti-
tutionalization and constitutionalization. Such an order might appear unsatisfactory
to those who have clear views on the single right constituency for particular issues.
But it respects the fact that people have different views and make different choices;
respect for this disagreement is, after all, respect for everybody’s right to self-government,
to equal participation in the design of the political order.119
6 Problems of Pluralism
Even if there are good prudential and moral reasons to favour a pluralist global
administrative law in principle, many problems with this approach remain. In the
space of this article, I cannot explore them all, but I want to highlight two central dif-
ficulties and point to possible ways of countering them.
116
See Shapiro and Hacker-Cordón, ‘Outer Edges and Inner Edges’, in Shapiro and Hacker-Cordón, supra
note 23, at 1.
117
See supra, sect. 2 C. As pointed out there, all of them can be defended on the basis of different broader nor-
mative approaches.
118
One might also think, e.g., of subnational groups as additional constituencies.
119
See also Waldron, supra note 115, at ch. 11, especially on Wollheim’s paradox, at 246.
One problem concerns lack of certainty: with the disappearance of a clearly compe-
tent authority and the resulting fluidity of decisions, the clarity and stabilization of
expectations that we usually expect from the law would be severely compromised.
This may be of particular concern in economic contexts where strategic decisions
depend on a predictable legal environment; and given that global administrative law
is in many instances concerned with economic regulation, legal certainty may gain
particular importance here. To take the example of GMOs, an enduring lack of clarity
about the legal framework will make decisions for all economic actors involved more
difficult and might cause them to adopt a cautionary attitude; a result that may not
be ideal in economic terms. However, similar charges could be brought against other
mechanisms of accountability too: most of them are inherently destabilizing.120 They
challenge decisions that have already been taken, and especially when they involve
court proceedings, they might leave the law unclear for many years. Even in the case
of judicial review, we thus face a trade-off between efficiency and legal certainty on
the one hand and fairness, individual rights and democratic accountability on the
other. A pluralist model of challenge and contestation may indeed aggravate the
problem, but this would only reflect the fact that the political and social structure at
the global level is much more complex than in smaller and more homogeneous
national communities. In order to achieve similar goals, we might have to sacrifice
more in terms of certainty and predictability than on the domestic level. And eco-
nomic actors are, in fact, well equipped to deal with this problem: it merely raises the
costs of engagement, but as long as the benefits are high enough, it will not deter
them.121 We should also keep in mind that usually decisions will remain unclear over
a longer period only when they concern issues of high salience, as in the GMO case.
On most issues, though, a pluralist order will operate much more smoothly. And if it
nonetheless appears necessary to provide for greater clarity and stability in some con-
texts, one might decide to establish institutions along consociational lines, or even on
a federal model by sacrificing some procedural fairness for the sake of substantive
goals. Yet these are significant trade-offs, and it is not clear that the added uncer-
tainty of a pluralist order is as problematic as the risk of blockade and the lack of
inclusiveness in these alternative models.
A yet more serious problem concerns power disparities. Pluralist approaches (even
in a less radical shape) have long had to face the objection that some groups in society
possess superior organizational capabilities and, in general, more power than others,
so that relying on their free interplay will merely favour the powerful at the expense
of the weak.122 To some extent, this risk has been contained here by limiting the range
of eligible constituencies; only sufficiently inclusive groups have been found to pass
the kind of normative ‘pre-screening’ required. Still, the general objection retains
120
Fisher, supra note 3, at 513–514.
121
E.g., multinational water companies have changed their strategies but have not ceased to invest in
developing countries in the face of unruly regulatory environments: see Morgan, supra note 95.
122
See M.G. Schmidt, Demokratietheorien (2nd edn., 1997), at 159. See also the parallel critique of
Kukathas’ approach in the book review by Spinner-Halev, 67 Journal of Politics (2005) 595, at 596–597.
276 EJIL 17 (2006), 247–278
force in the context of global administrative law. In our GMO example, the EU chal-
lenge to international regulation bears much weight, whereas a similar challenge by
a developing country would likely remain ineffectual. Likewise, for an international
body to object to regulation in, say, Zambia, will usually be far more successful than
to object to US regulation.
Yet it is open to question whether a pluralist landscape would be much different in
this respect from a classical, hierarchically ordered, unitary structure. This will ulti-
mately have to be assessed in specific contexts of global regulation, but as we know
from domestic contexts, differences in organizational capacities are extremely import-
ant also in procedural models with clearly defined participation rights for affected
interest groups;123 the power relations outside an institution are always to some
extent reflected inside it, despite provisions for formal equality. And power disparities
may be more effectual when the regulatory process leads to one final decision; incentives
for capture are biggest in a concentrated decision-making process.124 The dominant
influence of powerful actors on the final decision will then reinforce these actors’
power and legitimize its exercise.125 In contrast, recognizing the provisional and con-
testable nature of regulatory decisions limits their claim to legitimacy and may open
up space for less powerful actors to articulate their position. Their contestation can
remain visible and is not submerged in a process that claims to produce a decision
reflecting the ‘common good’.126
A second way to counter the argument from power is to devise substantive and
procedural norms that would limit the influence of power on the pluralist process.
Some prominent candidates are principles of mutual recognition, toleration, coher-
ence, or inclusiveness of political communication.127 The latter probably goes furthest
in demanding not only respect for difference, but also the transformation of purely
self-regarding argument into one that takes into account and responds to the
perspective of the other.128 If a principle like this could be established as the benchmark
123
See Stewart, supra note 15, at 1760–1805; Schmidt-Aßmann, supra note 16, at 265–266.
124
On the related tendency towards a ‘politicization’ of the judicial process when politics becomes judicial-
ized, see Ferejohn, ‘Judicializing Politics, Politicizing Law’, 65 Law and Contemporary Problems (2002) 41,
at 63–65.
125
On dominant influence in the shaping of global regulatory institutions see Steinberg, ‘In the Shadow of
Law or Power? Consensus-Based Bargaining and Outcomes in the GATT/WTO’, 56 Int’l Org (2002) 399;
Shaffer, ‘Power, Governance, and the WTO: A Comparative Institutional Approach’, in M. Barnett and
R. Duvall, Power in Global Governance (2005), at 130 (on the WTO); Mattli and Büthe, supra note 30 (on
the setting of international accounting standards). On the generally precarious position of international
legal norms between reinforcing and limiting power, see Krisch, ‘International Law in Times of Hegem-
ony: Unequal Power and the Shaping of the International Legal Order’, 16 EJIL (2005) 369.
126
For arguments along these lines, see Young, supra note 26, at 40–50, 108–120, 167–173; Kukathas,
supra note 108, at 164, 191–192, 268. See also the discussion of counter-hegemonic, subaltern plural-
ity in B. de Sousa Santos, Toward a New Legal Common Sense: Law, Globalization, and Emancipation (2nd
edn., 2002), at 471–493. Laski’s pluralism seems to have been motivated by similar concerns: see
Nicholls, supra note 104, at 86.
127
See Tully, supra note 17, at 117–124; Kukathas, supra note 108, at ch. 4; Maduro, supra note 84, at
99–100, Young, supra note 26, at ch. 2.
128
See ibid., at 35–36.
for decision-making by all the constituencies relevant for global administration, the
problem of power could be limited, even if – as in all other forms of global administra-
tive law – not eliminated. This may sound utopian, but not necessarily so. In the
European Union, the fact that the question of ultimate authority has remained open
and contested, may actually have helped to anchor a principle of ‘constitutional toler-
ance’ in political practice.129 And in the discussion of the GMO dispute, we have seen
that the different constituencies seemed often not only to observe the demands of
their counterparts, but to take steps to accommodate them pragmatically.130
7 Conclusion
Organizing accountability in the global realm faces a dilemma. The classical mecha-
nisms of legitimizing international law and international organizations have lost
strength with the rise of global governance functions and their increasingly inde-
pendent exercise, often at a far remove from conventional treaty frameworks. Conse-
quently, the role of treaties as ‘transmission belts’, ensuring accountability to states
and through the ratification process also to the public within states, has become
weaker and weaker, just as demands for stronger accountability have risen in the face
of ever more intrusive global regulation of formerly domestic affairs.131 On the other
hand, filling the resulting gap by using domestic models of accountability mecha-
nisms also meets with significant obstacles. In domestic settings, standard elements of
administrative law, such as procedural participation and judicial review, perform
particular, limited functions, and most of the burden of ensuring accountability lies
with the electoral processes that check parliamentary law-making as well as govern-
mental action. Once this electoral anchor is lost, as it is on the global level, adminis-
trative law-type mechanisms risk being overburdened by the demands made on
them.132 As I have tried to show in this article, differences in the constitutional foun-
dations present a serious challenge also in another respect: in global regulatory
governance, unlike in domestic administrative contexts, the question of to whom
accountability is primarily due – the question of the right constituency – is highly
disputed. National, international and cosmopolitan constituencies are all vying for
ultimate control, and this undermines the hierarchies that domestic administrative
law is built upon. The clearly ordered accountability mechanisms that we know from
national contexts, based as they are on the idea of a coherent and unitary system so
characteristic of modern constitutionalism, are thus of limited use for the contested
and decentred global administrative space.
129
See Weiler, supra note 84, at 18–23.
130
See supra, sect. 3 B.
131
See Weiler, ‘The Geology of International Law – Governance, Democracy and Legitimacy’, 64 ZaöRV
(2004) 547.
132
See Kingsbury, Krisch, and Stewart, supra note 1, at 48–49. For an attempt at conceptualizing non-
electoral accountability mechanisms, see Macdonald and Macdonald, ‘Non-Electoral Accountability in
Global Politics: Strengthening Democratic Control in the Global Garment Industry’, in this issue.
278 EJIL 17 (2006), 247–278
The ‘pluralist global administrative law’ that I have sketched in this article
suggests a possible direction out of the dilemma this creates. It refrains from establish-
ing a clearly structured institutional order and accepts that constituency contests will
be played out in mutual challenges between different regimes and different levels in
global regulatory governance. The connections between the parts of the overall order
are heterarchical, not hierarchical, and often of a political rather than legal nature.
Stability is thus created not by final decisions based on ultimate authority, but
through processes of negotiation and compromise as well as challenge and conces-
sion between the different constituencies involved. The resulting picture is disorderly
and fluid but, as we have seen in the examples of GMO regulation, Security Council
sanctions and the European Union, not unworkable. Pluralist accountability mecha-
nisms have emerged in many areas of global governance, and while they lead to open
institutional confrontation on issues of high salience and contestation, they operate
in most instances relatively smoothly through pragmatic mutual accommodation.
Such a pluralist order is not perfect; in fact, it poses many important problems, par-
ticularly with respect to legal certainty and power disparities. However, as I have
tried to argue, there are good reasons for regarding it as superior to more coherent
and unitary competitors on both pragmatic and moral grounds. A federal-style model
that would grant each of the different constituencies primacy for distinct issue areas
would likely create serious friction in a world in which this primacy is heavily con-
tested. A consociational approach, while better at dealing with this contestation,
faces significant problems of practicability, due to its cumbersome consensual proc-
esses. A pluralist order, in contrast, may be able to combine a reasonable degree of
practicability with a respect for the different views and allegiances of people when it
comes to determining the right constituency. These views differ for good reason: there
is a strong normative case to be made for each of them, but also important arguments
for why none of them should enjoy primacy. Rejecting formal supremacy claims in a
pluralist order is a way of respecting the diverging individual choices while leaving
space for shifts in their respective weights.
A pluralist global administrative law may not correspond to anybody’s ideal; its
design is far too unstructured and leaves too much room for political struggle. Yet this
may be precisely its virtue. Being nobody’s ideal, it refrains from taking sides in the fun-
damental contests that characterize the global order. Bracketing the issue of principle,
and working around it pragmatically, may after all not only be prudent but also mor-
ally preferable. And it might be politically advantageous: rather than stabilizing a par-
ticular institutional setting, a pluralist order might open up space for the political
transformation of a structure of global governance whose legitimacy is far from settled.

The Pluralism of Global Administrative Law

Uploaded by

Copyright:

Available Formats

The Pluralism of Global Administrative Law

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Pluralism of Global Administrative Law

Uploaded by

Copyright:

Available Formats

The European Journal of International Law Vol. 17 no.

1 © EJIL 2006; all rights reserved

The Pluralism of Global

systems is questioned by those who refuse to give up democracy and constitutionalism

2 Accountability, Administrative Law, and Its Many

B The Constituencies of Domestic Administrative Law

C Three Competing Constituencies in Global Administrative Law

3 Contested Constituencies in Global Risk Regulation: The

A The Substantive Dispute over GMOs

B The Institutional Contest

The Biosafety Protocol,54 in contrast, has been established in part as a counter-

C Why Competing Accountability Mechanisms?

accountable to an international constituency. The buffers the EU has erected against

4 Accommodating Multiple Constituencies: A Pluralist

cosmopolitan constituency to prevail on issues of global environmental impact; a plu-

A The Consociationalist Road and Its Problems

the central elements of consociationalism is its reliance on the cooperation of elites:

B From Co-Decision to Mutual Challenge

Similar pluralist orders, in which different constituencies balance each other

5 A Principled Case for a Pluralist Order

A The Constituencies’ Equal Deficits

B Equal Respect for Constituency Choices

Insofar as autonomy is linked to political self-government, it includes a right to parti-

A similar emphasis on freedom of association underlies Chandran Kukathas’

C The Range of Relevant Constituencies

more substantive conception, limits on the range of relevant constituencies will

You might also like