It is a pleasure to have been invited to contribute to the discussion of the article by Juliane Kokott and Christoph Sobotta on balancing constitutional core values and international law against the background of the Kadi case. At the outset I must state that I find myself in broad agreement with what I understand to be the authors’ central argument: ie that the CJEU (or ECJ) employed a variant of the Solange argument, if implicitly, in its Kadi judgment of 2008. I have in fact also argued this in a paper I presented in Oslo in 2009, which appeared in print in January 2012 (here) and indeed on this blog (here and here). I would kindly ask readers (if any) to also read the latter blog posts, as the present comment builds on the premises there laid out.
I will proceed with the discussion of three major issues that arise from the authors’ discussion of Kadi as a balancing exercise between constitutional core values and international law. The first issue refers to the perceived ‘dualism’ of Kadi and consequently of any attempt to employ a Solange argument. The second issue deals with the content of the Solange argument, in particular with the rules it seeks to establish and / or safeguard. And the final issue deals with the justification of the Solange argument in (international) law and the ‘battle for the analogy’.
I. Dualism, Monism, Primacy, and the Premises of a Solange Argument
Kokott and Sobotta note that Kadi has been criticised by some for taking what might be seen as an (even extreme) dualist position. Indeed the authors seem to concede the point, noting that the approach of the CJEU was merely ‘somewhat’ dualist, and indeed justifiably so under the circumstances, namely the exercise of ‘governmental’ authority in multilevel systems of governance. The dualist approach was, according to the authors, ‘required’ due to insufficient protection at the UN level: by strictly separating the two legal orders (UN and EU) the Court of Justice managed to subject sanctions—internationally imposed and locally implemented—to judicial control for the protection of fundamental rights (as locally guaranteed).
But such a dualist approach was not required to employ a Solange argument. Even if the EU legal order is an ordre juridique propre, separate from the international legal order, and thus a dualist approach is conceivable, the European Court of Human Rights has also employed Solange arguments, as the authors themselves mention. Indeed, in a number of cases, whether along the lines of Waite and Kennedy, or in Bosphorus, the ECtHR resorted to Solange II deferral to the control exercised at the level of an international organisation. Solange II-type arguments are implicitly based, by necessity, on an underlying Solange I argument: should the conditions for deferral no longer be fulfilled, the Court will resort to ‘local’ control.
Could the ECtHR be seen as subscribing to some sort of ‘dualist’ understanding of the relationship between the partial legal order established by the ECHR and that established by international organisations, whether the European Space Agency, or the EU, or the UN? I think probably not—but in order to understand this we must discuss the premises for the adoption of a Solange argument. Simply put, these are: (a) there must be some power that is being exercised, and whose exercise the ‘local’ court would normally control (eg exercise of ‘sovereign’ power controlled for compliance with locally guaranteed fundamental rights); (b) this power is being exercised by an actor who cannot be controlled in the instance by the ‘local’ court due to ‘local’ jurisdictional restrictions; (c) the actor who cannot be controlled by the ‘local’ court is either not controlled at all at the (non-local) level of exercise of its ‘sovereign’ power, or is subject to control that is significantly inferior to that to which it would normally be subject at the ‘local’ level.
This distinction between actors who can and cannot be directly controlled by the ‘local’ court can of course be achieved by reference to distinct legal orders (hence the ‘dualism’ ‘charge’). But it can also be achieved through the use of rules of attribution in international law: when certain conduct is imposed by one actor and implemented by another actor, said conduct may be attributable to either one of these actors, or to both; but if it is attributable to at least one actor that the ‘local’ court has jurisdiction over, this is the only thing that the ‘local’ court needs. The ECtHR will customarily attribute acts undertaken by State organs in implementation of international obligations to the State, which—being a party to the ECHR—can be controlled for compliance with the latter. If the State argues that it was merely implementing obligations it owes to an international organisation, the ECtHR will either hold the State responsible for having conferred powers on the organisation without imposing adequate controls on the international organisation level, or it will presume that the State has complied with its obligations under the ECHR, because adequate control exists at the international organisation level. This attribution point was what made Bosphorus possible, and this attribution point is what made Behrami clearly problematic.
So a Solange approach does not presuppose the existence of separate, distinct legal orders. It merely presupposes courts that are jurisdictionally limited as to the actors that they can control. The CJEU was not ‘required’ to adopt a ‘dualist’ stance in Kadi in order to introduce a Solange argument. But it was required—in its own eyes—to do so in order to avoid deciding on the question of primacy within a unified legal order, the question of hierarchical ordering. Let me note here that monism does not necessarily require the primacy of international law: monism merely postulates the existence of a unified legal order; within that unified order it may be domestic law that is supreme over international law, as Kelsen himself had conceded (see briefly the last chapter of his Introduction to the Problems of Legal Theory, a translation of the 1st edn of the Pure Theory of Law by the Paulsons). As such, the CJEU could have adopted a monist approach postulating the primacy of EU law over everything else. This the Court did not want to do; but it could have done it (in the final analysis, it does decide on the operation of international law by relying on the relevant rules of EU law as to how international law operates within the level of the EU partial order: in this sense, the EU legal order is monist).
International law, itself ‘dualist’ to some extent, introduces two basic rules which may be seen as establishing its primacy and a hierarchy of sorts: a general rule providing for the irrelevance of domestic law as a justification for the breach of an international obligation (Art 27 VCLT; Art 3 Articles on State Responsibility); and a special rule providing for the primacy of obligations under the UN Charter over other international obligations (Art 103 UN Charter). The first excludes the possibility of invoking local constitutional rules as against international obligations; the second excludes, according at least to part of the literature, the possibility of invoking any international obligations, and thus even obligations established within partial orders other than those of States, as against obligations under the UN Charter, including the obligation to comply with binding decisions of the Security Council (on ‘hierarchy’ and Art 103 see further here, with references to the abundant literature). A monist approach accentuates the potentially conflicting content of rules established within different partial orders, and requires them to be hierarchically structured in order to resolve conflicts. A dualist approach allows for the externalisation of the conflict: everything is fine within the partial order (which may now be seen as separate and independent); the conflict is exported as a conflict between legal orders, and is now someone else’s problem (ie at least not the ‘local’ court’s problem; perhaps the Comptroller-General’s problem).
The Solange argument elegantly avoids the confusion of the monist-dualist discussion, and blunts the edges of the question of primacy, which poses itself with some force within the framework of such a discussion. In the next section, I discuss the content of the Solange argument, before going on to the various justifications that can be seen to support it.
II. The Content of the Solange Argument: the Great Equaliser
The Solange argument seeks to establish the following point: the exercise of ‘sovereign’ powers cannot be freed from judicial control simply by being conferred on an actor that escapes the controlling court’s jurisdiction. It could be seen as an implementation of the maxim nemo plus juris transferre potest quam ipse habet. One may not circumvent limitations on the exercise of power by conferring the power to someone else. As employed by the German Federal Constitutional Court, the CJEU, the ECtHR, and the UK Supreme Court in Ahmed, the Solange argument requires that the exercise of ‘sovereign’ power be subject (a) to control offering guarantees of judicial protection (b) for compliance with certain fundamental rights. So in effect, reference is made to a catalogue of substantive rights that must be protected, as well as to a right to meaningfully challenge exercise of power for compliance with this catalogue of rights. The latter, a right in and of itself, is established as the pinnacle of all other rights, an enabling right allowing all other rights to be effective because effectively protected.
The Solange argument requires equivalence on both these points: (a) equivalence as to the rights to be protected, and (b) equivalence as to the mechanics of their protection. This requirement of mere equivalence (or adequacy) allows for flexibility on both fronts: as to the catalogue and substantive content of the rights to be protected, these need not be identical at the different levels of exercise of the ‘sovereign’ power. The rights sought to be protected may be drawn from the ‘local’ level and projected onto the current (non-local) level of the exercise of the ‘sovereign’ power, but they are in reality only core rights—or core aspects of rights—that are also protected on the current (non-local) level of exercise of power. To put this simply, the rights that the German Federal Constitutional Court sought to ‘impose’ on the EEC (as it then was) were not rights peculiar to the German Basic Law; they were drawn from the Universal Declaration of Human Rights, and indeed the Court was satisfied with EU protection of rights ‘drawn from the common constitutional traditions of the Member States and inspired by the ECHR’. There is a fundamental feedback loop between rights protected at various international and domestic levels: domestic constitutional protection across various States influenced the catalogue of rights to be internationally protected through multifarious regional and global instruments. These instruments have in turn shaped, through their domestic implementation, the subsequent evolution of the content of constitutionally protected rights. It is thus meaningless to try and establish primacy between the local/domestic and non-local/international in these circumstances: the rights are ‘consubstantial’, ie in their core aspects they have the same ‘substantive’ content (for the term see here at pp 163–4 and regarding the terms ‘fundamental rights’ and ‘human rights’ see further here at section VI).
Equivalence also refers to the method, or mechanics, of protection: this protection need not be, strictly speaking, judicial protection. Rather, it must offer guarantees of judicial protection. What is established here is the right to challenge the exercise of ‘sovereign’ powers before an instance that offers such guarantees, ie an independent and impartial instance with the power to make binding determinations at the very least. This the problem of the UN Ombudsperson: while independent and impartial, it still cannot make binding determinations, even after recent improvements which establish the requirement of an inverse consensus within the Sanctions Committee for the Ombudsperson proposal being disregarded: the question can be removed to the Security Council, where normal voting rules, including the veto, operate. But the challenge also needs to be ‘meaningful’, ie allow access to evidence and knowledge of the specific allegations that are to be disputed, etc. It makes sense that the precise level of equivalence required, ie the specific content of the guarantees of judicial protection may also vary from context to context: eg the guarantees need not be as exacting in the context of staff disputes in international organisations (eg in Waite and Kennedy-type cases) or in the context of private law disputes with international organisations protected by immunity, but they will be significantly more exacting in the context of crippling sanctions against individuals which, especially in view of their duration, may be deemed more akin to criminal sanctions than to mere preventive measures.
The Solange argument thus seeks to achieve, through its pattern of defiance (Solange I) and deferral (Solange II), equivalent protection of certain fundamental rights at various levels of exercise of sovereign powers. It acts as an equaliser, subjecting all exercises of sovereign power to similar, ‘adequate’, limitations, with a view to avoiding circumvention. The final section discusses how the Solange argument may be justified legally.
III. The Justification of the Solange Argument: the Battle for the Analogy
I think that it was Vaughan Lowe who said something along the lines that (international) law is a constant battle for the analogy. Let us assume he did. The justification of the Solange argument is such a battle. It may be seen as a method of imposing the primacy of the ‘local’ core constitutional value on a separate, foreign system (and thus is open to accusations ranging from local constitutional resistance to constitutional imperialism): you must do as I do, and as long as (‘solange’ in German) you don’t, I will meddle. It may be seen as a method of regulating conflicting jurisdictions through comity: as long as you offer some protection that is roughly equivalent to mine, I will not meddle. It may be seen as a method of structuring control in multilevel systems by operation of law: what is local becomes relative, and once protections at the level of exercise of the relevant power (the local level in the instance) are exhausted, other protections at non-local levels may spring into action, as long as local protection was inadequate—controlling however not against their own, peculiar, and potentially parochial, standards, but against standards that are universally grounded and universally protected. This latter reading is what I take Kokott’s and Sobotta’s argument on exhaustion of ‘local’ remedies to mean.
All these ways of conceptualising the Solange argument, whether in Kadi or in any other relevant setting, may have varying degrees of merit, and in any event may be (and have been) justified or rejected ‘legally’ on a number of different bases: the various levels are separate and distinct, and we merely focus on controlling local action at our level, even if that local action is imposed by action at another level; or: we need to harmonise protection at the various levels in order to avoid a breakdown of the whole system, as strict separation is dangerous for effective functioning of multilevel governance; or even: you are merely breaking the law when seeking to control action at one level for compliance with rules at another level.
My preferred analogy is to classical international law, and if it is as open to question as any other, in my reading it justifies the adoption of a Solange argument under international law. Non-protection of certain fundamental human rights, universally guaranteed, through a process that offers guarantees of judicial protection, constitutes a breach of an international obligation, namely the obligation to protect the particular right in the particular manner, imposed on an international organisation by virtue of customary international law. This customary law has developed through the practice of States requiring such protection at various levels of exercise of ‘sovereign’ authority. For as long as this breach is continuing, the local court will force a breach of the obligation of the relevant actor (State or international organisation) to comply with the normative acts of another actor (another international organisation) by controlling the acts of latter at the local level and potentially disregarding them. This breach is justifiable as a countermeasure taken by the injured actor as against the responsible actor. Once the wrongful act (the breach by the responsible actor) has ceased and some (at least juridical) restitution has been made, the obligation to comply with the responsible actor’s normative acts will be resumed, by eschewing local control and deferring to equivalent control (now) established at the level of exercise of the ‘sovereign’ power.