Professor Alexander Orakhelashvili
One year has passed since the International Court of Justice delivered its judgment in Azerbaijan v Armenia, in which it ruled on its jurisdiction to assess alleged violation of the 1965 UN Convention on Racial Discrimination in the armed conflict between those two States. Azerbaijan and Armenia became parties to CERD on 15 September 1996 and 23 July 1993, respectively, and the contended question was whether the Court had jurisdiction over acts and conduct that took place prior to the entry into force of CERD as between the Parties on 15 September 1996.
The Court has declined jurisdiction over acts prior to Azerbaijan’s ratification, the way that hardly squares with its 1996 decision in in Bosnia v FRY and offers, moreover its rather distorted representation. The Court suggests that the earlier decision “did not address … the broader issue of the application ratione temporis of compromissory clauses in human rights conventions” and that it was instead concerned with State succession in the process of the dissolution of the former Yugoslavia (para. 49). On closer inspection it will appear, however, that even if the 1996 Judgment has touched upon the succession issue, it did not conclusively clarify it or use it as the central basis of its decision; instead, the 1996 Judgment was based on a far wider range of legal reasons, and that for this and other reasons specified below, several outcomes upheld in the 2024 Judgment are flawed and not supported by evidence.
The Court had noted in Bosnia v FRY that, following the thesis of State continuity (which is not the same as State succession, see here, Chapters 5 and 14), on 27 April 1992 FRY declared that, as a State “continuing the State, international legal and political personality” of FRY, it would respect obligations the latter had assumed. The succession approach was adopted by Bosnia with effect from 6 March 1992 (the date of its independence), regarding the Genocide Convention. The Court (p.611) noted that on 22 May 1992 Bosnia became a member of the UN and hence eligible to become a party the Genocide Convention under Article XI, though at that time it did not do anything of the kind, and instead acted through its succession notice of 29 December that year. Voluntary succession did not have retrospective effect as far-reaching as Bosnia’s possible automatic succession would have had. Even if automatic succession to human rights treaties was discussed by individual judges, the whole Court did not base its decision on it. It may be that as the debate has developed since then, today’s Court would be more willing to endorse the thesis of automatic succession, but that can do little to change the fact and the record indicating that the 1996 Court did not do so. As for Bosnia’s consensual succession as of the later date of 29 December 1992, it would not have accounted for the period between 6 March and 29 December that year, and hence, contrary to the Court’s assertion at paragraph 49 in Azerbaijan v Armenia, the Convention would not have “remained applicable in the relevant territory at all times of the conflict”. Something more was clearly needed to account for that period.
The Court (p.612) was clear that State succession issues were not the basis of its decision and that it “not consider it necessary, in order to decide on its jurisdiction in this case, to make a determination on the legal issues concerning State succession in respect to treaties which have been raised by the Parties”. What mattered more was that whichever version of State succession could have accounted for Bosnia’s position, or whether its succession statement was to be treated as accession, Bosnia was “a party to [the Genocide Convention] on the date of the filing of its Application on 20 March 1993”, i.e. the date when it seized the Court. And (p.614), even if the Convention applied to relations between two States as of 1995 when the Dayton Agreement was signed and FRY and Bosnia have recognised each other, “the Court could not set aside its jurisdiction on this basis, inasmuch as Bosnia and Herzegovina might at any time file a new application, identical to the present one, which would be unassailable in this respect.” Needless to emphasise, this would have been a later and hypothetical new application, based on which it would have been possible to subsume earlier periods within the Court’s jurisdiction. This looks even more unlike any version of State succession to treaties.
Once all the aforesaid was established, then the matter was no longer, principally or exclusively at least, one of State succession, but one of application, ratione temporis, of the Genocide Convention and its jurisdictional clause under Article IX. Even if the applicant State would be deemed to become party to the Genocide Convention at the latest of all relevant dates – whether one of independence, or succession (claimed or actual), or accession to the UN, or the parties’ mutual recognition manifested, as late as 1995, through their conclusion of the Dayton Peace Agreement – the Convention would apply in the case to events before that particular date, in effect to as far backwards as the factual background of the case would go. That was the case simply because “the Genocide Convention – and in particular Article IX – does not contain any clause the object or effect of which is to limit in such manner the scope of its jurisdiction ratione temporis, and nor did the Parties themselves make any reservation to that end” (p.617).
It is therefore obvious that Bosnia v FRY was anything but about a binary difference between succession to treaties and their temporal application, or about treating those two things strictly as alternatives, as paragraph 49 in Azerbaijan v Armenia seems to pretend it was. The assertion that the Court “did not address” the ratione temporis issue back in 1996 is even more counterfactual.
The Court suggests in Azerbaijan v Armenia that “The present case is different [from Bosnia v FRY]. There is no question that the Parties to this case are both parties to CERD.” In Bosnia v FRY the Court had said pretty much the same thing, namely that “at all events [Bosnia] was a party to [the Genocide Convention] on the date of the filing of its Application on 20 March 1993” (p.612), as was FRY (albeit on different grounds). There is no significant difference between the two cases in that respect. Moreover, in 1996 the Court was clearly able to distinguish between the two issues: the party status and the ratione temporis applicability of substantive and jurisdictional provisions of the Genocide Convention. The latter was separate from the former, or consequential upon it, and the former was not alluded to the way that would prejudice or curtail the latter.
By contrast, in Azerbaijan v Armenia the two issues were addressed disjunctively; and the party status issue now meant more than it meant back in 1996 and covered the position at the litigation time as well as at the time of alleged violations, while in Bosnia v FRY it did not relate to that latter issue. The Court said that “Azerbaijan was not a party to CERD at the relevant time when the alleged acts took place” and derived from that fact the conclusion opposite to the one reached in Bosnia v FRY which held that, even if Bosnia was not a party to the Genocide Convention at the time when certain alleged violations took place (or even if treaty relations between the two parties did not come into effect till 1995), it could still invoke FRY’s responsibility for them simply because it was a party to the Genocide Convention at the time of seizing the Court. In addition, Bosnia’s application was treated as flexibly as possible (among others by applying to it the date of application approach pursuant to PCIJ’s jurisprudence) (p.613); while Azerbaijan’s application did not receive any of that flexible treatment.
The Bosnia v FRY Court has also made it clear that the outcome upheld in that case was in accordance with the continuous operation of the Genocide Convention in accordance with the erga omnes nature of Convention obligations following from their objective nature (p.617). It is important to bear in mind that a clear connection was find between these two issues, as opposed to the link between erga omnes obligations and automatic succession on which the Court did not pronounce (p.612). Since the well-known pronouncement in the Court’s 1951 Advisory Opinion, the doctrine of objective treaty obligations has become generally accepted by international courts and tribunals with regard to pretty much all humanitarian treaties (for detail see here, Chapter 4). This doctrine is too entrenched and generally accepted to be abolished or modified through the Court’s single decision.
It appears more likely that the Court is merely unwilling to give effect to that doctrine in this case. This is clear from the Court’s allusion to reciprocity. It stated that, while judicial jurisdiction is based on State consent, “Procedurally, if Azerbaijan were permitted to make claims against Armenia for the latter’s alleged acts that occurred during that period while Armenia could not exercise that right against Azerbaijan for Azerbaijan’s conduct during the same period because of its non-party status, there would be no reciprocity and equality between the Parties” (para. 51). In Bosnia v Serbia, the Court did what it did, even if FRY could not have been able to sue Bosnia before certain of the dates with which Court identified Bosnia’s party status.
Moreover, there is no general principle of consent that is relevant beyond enabling a State to undertake or not to undertake international obligations. Effect of State consent in depends in specific cases on the legal framework to or within which consent is given. Hence, not all instances of consenting to jurisdiction are the same (see also here); some are given in a free-standing manner, while others are given as part of a wider or pre-existing legal framework. Therefore, what matters is not only whether consent is given but also what (namely what type of legal framework) is being consented to and if that framework is not based on reciprocity, consent to it should not be interpreted to pretend as though it did. In the context of CERD, reciprocity could not be soundly insisted upon, because CERD, just like any other human rights or humanitarian law treaty, is a body of objectively operating obligations aimed not at protection of the rights of States-parties but at the protection of vulnerable persons or groups from discrimination; treating these as reciprocal obligations involves a category error.
The CERD Committee’s approach in Palestine v Israel has been different from that of the ICJ. The Court adopts a rather obscure comparison between the jurisdiction of CERD committee under Article 11 and its own jurisdiction under Article 22 CERD, suggesting that “The first aims to monitor compliance by States parties with their obligations under the Convention and can be used “[i]f a State Party considers that another State Party is not giving effect to the provisions of [the] Convention” (Article 11). The latter aims to settle disputes relating to obligations which States, by becoming parties to the Convention, have accepted to undertake vis-à-vis each other, and the judicial settlement may result in the engagement of the respondent’s responsibility towards the applicant” (para. 54). This is a distinction without a difference. It evades one’s mind how “not giving effect to the provisions” is different from the issues arising as a matter of State responsibility, or how CERD Committee does not aim “to settle disputes relating to obligations” of States-parties. The fallacy of the Court’s conclusion is furthermore evident from the fact that, if the CERD Committee has dealt with, in the relevant case, with acts occurring prior to one State-party’s ratification of CERD, those matters would ipso facto as well as ipso jure become, under of Article 22 CERD, matters that have not been settled by other means provided for in the Convention, which include the CERD Committee. This is also not the first time when the Court has given a distorted representation of dispute settlement framework under CERD.
While it is difficult to identify precise intention driving the Court in adopting a particular decision, one commentary suggests that the Court might have been motivated by gatekeeping concerns. If that were the case, then the Court must have adopted the approach that the ends justify the means, and a case of opportunism would be rather obvious. For, even if keen on a particular outcome, there is no warrant for an international tribunal to take liberty with the use of treaty-based compromissory clauses the same way as a court in common law jurisdiction could do with rules governing its jurisdiction.
Clarifying the real aim of propagating an allegedly insurmountable difference between erga omnes obligations and consensual jurisdiction is another pressing issue. The Court propagates that thesis in paragraph 48 the Armenia-Azerbaijan judgment, referring to two previous cases of East Timor and DRC v Rwanda, in effect presenting these two cases that had different context and differentials as though they manifested some single or integrated doctrine of non-justiciability of public interest obligations that is allegedly suitable to determine issues of a yet different type arising in Azerbaijan v Armenia. Although some people might take what they see in paragraph 48 as granted rather than go on and verify the underlying rationale, one still might wonder whether the Monetary Gold principle (which was, by the way, disapplied by all other major international courts) is so wide and all-encompassing, all-singing and all-dancing as to cover most or potentially all cases involving any relation between the case and third party interest, and thus even to require the invocation of erga omnes obligation to offset its effect; or whether it was adequate that the Court in DRC v Rwanda and Sudan v UAE persistently chose not to address or discuss the complex issue of compatibility of reservations to Article IX of the Genocide Convention except through one sentence that contained no analysis and stated a preconceived outcome, most likely to avoid discomforting the US government that also maintains Article IX reservation; and hence whether what the Court says at paragraph 48 could feasibly provide a generalised doctrine applicable to all or most cases where public interest norms are invoked before the Court.
If one wonders what is behind such putative general doctrine of non-justiciability, one is bound to discover rather little apart from the ideology that that anything preventing accountability and cementing impunity is to be welcomed as sound and well-founded, and anything that promotes accountability and enforcement of international law is to be treated as unsound and unrealistic. Well, if you believe in any of that, the only question you have to ask is whether you are at the directing or at the receiving end of the fabrication process.