This week Dapo Akande was kind enough to allow Max du Plessis and I to post the following on the EJILTalk blog, as part of an ongoing discussion on the topic:
Earlier this week, Dapo had a post dealing with the obligations of contracting parties to the Genocide Convention to implement ICC arrest warrants and pointing to the UCLA Online Forum debate on this topic. In what follows we hope to contribute to the discussion around point (ii) of the topics raised by Dapo’s post and the UCLA debate: the obligations of African Union States Parties to implement ICC arrest warrants. Our contribution is drawn from our upcoming position paper for the Institute for Security Studies’ International Crime in Africa Programme. The paper provides an analysis of the various obligations Kenya and other States Parties must meet to the ICC in respect of both al-Bashir and the Court’s ongoing investigation into the post-election violence in Kenya, and which considers the nature of the obligations on African States Parties such as Kenya in respect of the AU decision, and in particular the demand for non-cooperation in respect of al-Bashir. This paper builds on the ISS’s previous position paper – An African expert study on the African Union concerns about article 16 of the Rome Statute of the ICC, by Dapo Akande, Max du Plessis and Charles Jalloh, previously discussed on this blog (here and here) – and similarly the Institute plans to launch the paper internationally in due course.
Firstly, decisions of the AU Assembly are potentially binding on member states. Although there is no express provision in the AU’s Constitutive Act conferring this power, it is clear from article 23 – which sets out the consequences for failing to abide by such decisions – as well as a thorough contextual reading of the Constitutive Act that the Assembly is empowered to do so. Even if the text of the AU Constitutive Act is considered insufficient or equivocal in this regard, given the considerable mandate the body has been given by its member states, those advocating for the binding nature of Assembly decisions could rely on the doctrine of implied powers to support their position. Further, as Bill Schabas notes in his piece on the UCLA Law Forum, the AU Commission clearly views the AU’s Bashir Decision as binding on its members.
As far as the July 2010 AU Decision (Assembly/AU/Dec.296(XV) is concerned, using the Security Council as an analogue, various interpretive techniques can be employed with the view to ascertaining the “intention” of the Assembly. Firstly, the plain language (the primary indicator) of operative paragraph 5 of that decision – whereby the AU “Reiterates its Decision that AU Member States shall not cooperate with the ICC in the arrest and surrender of President El-Bashir of The Sudan” – clearly suggests it was intended to be binding and not merely exhortatory. Moreover, a consideration of the practice of the AU Assembly reveals that the word “decision” and its variations are used sparingly and deliberately. Unfortunately, due to the controversy that surrounded its adoption, and the complex three-tiered decision-making structure of the AU Assembly, the discussions leading up to the Bashir Decision are of little assistance in this regard. Nevertheless, the decision that AU states shall not cooperate clearly creates a prima facie obligation on such states not to do so.
As Bill Schabas rightly notes, the upshot of this is “with respect to Member States of the African Union that are also States Parties to the Rome Statute, there would appear to be a conflict between the binding obligations imposed by the Rome Statute and the binding obligations imposed by the Decisions of the African Union”. This ‘norm conflict’ can be resolved in one of two ways: with reference to hierarchy (such as a jus cogens norm or per article 103 of the UN Charter) or through “techniques of interpretation”.
As far as the use of hierarchy is concerned, in our view there is little to be gained from article 103 of the UN Charter in this regard. Although it might have done so, by its terms UNSCR 1593 does not bind states (other than Sudan) under Chapter VII to cooperate with the Court in respect of the Sudan situation. As for jus cogens obligations in respect of genocide, leaving aside debates over when these are engaged, their scope cannot (perhaps yet) be said to extend to the execution of an arrest warrant.
So far so good.
Where we differ respectfully with Bill is in respect of the second conflict resolution tool: techniques of interpretation. Bill dismisses Amnesty International’s attempt to argue that “the African Union intended that its Decision ordering States not to enforce the arrest warrants be consistent with the Rome Statute, and at the same time to declare the interpretation of a provision of the Rome Statute upon which the Decision of the African Union is based to be erroneous”. At this point, having dismissed AI’s attempts to read down the AU Decision, Bill concludes that the two obligations “cannot be reconciled” through interpretive means and, in the absence of “any apparent rule or formula establishing a hierarchy by which one prevails over the other”, the “conflict of legal norms requires a political solution”.
However, we believe that there is an interpretive means to avoid this “conflict of legal norms”, and that solution lies within the AU Decision itself. While the controversial paragraph in the AU Decision demanding non-cooperation has attracted much attention, one must also consider the paragraph that follows immediately after, which “[r]equests Member States to balance, where applicable, their obligations to the AU with their obligations to the ICC”. This ’balancing paragraph’ – included at the insistence of states such as South Africa whose implementing legislation obliges them, under domestic law, to cooperate with the ICC – makes a straightforward textual interpretation of the Decision as a whole difficult. It suggests a decision that drives at a categorical imperative yet at the same time provides allowance for a measure of discretion. On their face these two commands appear irreconcilable by simultaneously commanding member states not to cooperate in the arrest of al-Bashir and requesting them to balance this edict with their obligations under the Rome Statute. In such circumstances it becomes necessary to look beyond the text in order to give meaning to these two paragraphs.
Here there is unfortunately little guidance provided by the discussion in the AU Assembly. Indeed, at the time the Decision was adopted there was concern raised by some states that the AU Commission – responsible for drafting the voluminous decisions of the Assembly as they pass through the various stages of the decision-making process – had on more than one occasion distorted the discussions and decision-taking during the preparatory stages in order to push its own hawkish line on the question of the Bashir Arrest Warrant. The difficulty then is that reference to the discussions leading up to the Decision is not a reliable means of infusing plainly contradictory language with any singularity or clarity of purpose.
Arguably there is only one tool remaining by which to avoid the norm conflict. In this regard we submit that the internal contradiction in the AU Decision occasioned by the inclusion of this ‘balancing paragraph’ should be resolved by employing the doctrine of effective construction. This doctrine takes on different forms but has been held to require that one “avoid interpretations which would leave any part of the provision to be interpreted without effect”, and that “an interpretation which would make the text ineffective to achieve the object in view is prima facie suspect”. In the past it has been used by both the ICJ and the ICTY Appeals Chamber, and Bill references it in his discussion on immunity. Although those occasions involved the interpretation of treaties, there appears to be no reason why the doctrine is not applicable in the interpretation of organisational acts.
Applying the doctrine of effective construction to the AU Decision means that to read it as peremptory would render permissive opt-out meaningless. Therefore, in terms of this doctrine the text of the Decision is best rendered exhortatory: that is, it is not an unequivocal command but rather is diluted when states (“where applicable”) are under another obligation stemming from the Rome Statute. States parties may thus avoid the apparent conflict of norms represented in the AU Decision by an interpretive turn which takes the text of the decision seriously and attempts to render it meaning-ful.
Incidentally, the AI report considers this ‘balancing paragraph’ and concludes that it “provides AU member states with a tool to solve any possible conflict between the obligations arising from the 2009 and 2010 Decisions and the obligations towards the ICC”. The problem is that it does not suggest how this tool might work, although it’s clear that on AI’s reading this balancing exercise would inevitably “result in the obligations towards the ICC prevailing over the obligations towards the AU” in respect of all AU member states. We would avoid a reading of this paragraph and the use of balancing as a substitute for subjugating the AU Decision to the ICC. The AU’s Decision is (textually and contextually) a reflection of internal inconsistencies around deeply-felt concerns about the Court’s work in Africa. We suggest that, although the result may be the same, the doctrine of effective construction is a better means of simultaneously recognising that these concerns animate the AU’s Decision, whilst addressing the contradictory aspects of a decision by a regional organisation comprising 53 sovereign states.