Piecing the Withdrawal Puzzle: May the ICC still open an investigation in Burundi?

[Sergey Vasiliev is an Assistant Professor of International Law, Grotius Centre for International Legal Studies, Leiden University. This is the first part of a two-part contribution.]

Questions raised by the ICC’s reaction to Burundi’s withdrawal

On 27 October 2017, one year after Burundi notified the UN Secretary-General of its intention to withdraw from the Rome Statute, the withdrawal became effective in accordance with Article 127(1). The preliminary examination of the situation in Burundi has been ongoing since 25 April 2016, as announced by the Prosecutor. The expiry of the ‘cooling off period’—and the first effective withdrawal in the ICC’s 15-year history—was an event of some significance; more so perhaps than those covered in its press releases on or around the same date. The observers expected an official acknowledgement from the Court that the Article 127(1) term had ended. The Court could have also taken the opportunity to provide clarity on the status of the situation to the public and, above all, to the withdrawing State whose representatives (like some Bijumbura residents), yearned for a sense of closure and had already bade the Court a festive farewell.

Disappointingly, no planned ICC statement followed. The ICC spokesperson’s curt and enigmatic response to the journalists’ queries (see BBC, AP, Al Jazeera and communications to Benjamin Dürr and Anna Holligan) only thickened the plot. The spokesperson intimated that an announcement regarding the results of the Burundi preliminary examination would be made in due course in accordance with the OTP’s practice. More controversially, he asserted that ‘the Burundi withdrawal does not affect the jurisdiction of the Court with respect to the crimes alleged to have been committed during the time it was a State Party, namely up until 27 October 2017’.

The latter point raises questions about the legal consequences of withdrawal: in particular, whether the Court retains jurisdiction over the crimes committed in Burundi while it was a State Party, and whether an investigation could still be opened in the aftermath of the withdrawal. The OTP’s 2016 Report on Preliminary Examination Activities neither unequivocally asserted nor ruled out the possibility of initiating an investigation after the withdrawal becomes effective: ‘According to its legal assessment, the Office could also initiate investigations at least during this one-year period.’ (para. 60). Notably, the Report said nothing about the impact of the withdrawal on jurisdiction. The position that jurisdiction remains unaffected is far-reaching and not as straightforward as presented in the statement; the issue of jurisdiction is not squarely addressed in Article 127. This led some observers (this writer included) to wonder whose legal opinion the spokesman expressed and what the legal basis for it was.

The—so far the only available—ICC’s official reaction to the Burundi’s effective withdrawal is in line with the view expressed by the authors of the Amnesty HRIJ blog post and other commentators (see Alex Whiting’s 2016 post), who consider that the consummation of withdrawal does not (necessarily) extinguish the ICC’s jurisdiction. The authors of the Amnesty HRIJ post further argue that, therefore, an investigation into the Situation in Burundi could still be opened, even after 27 October 2017. Somewhat differently, Alex Whiting found Article 127(2) of the Statute to be unclear on this point—which it certainly is—and, therefore, considered it safer for the Prosecutor to request judicial authorization under Article 15(3) before the withdrawal became effective. By contrast, Dov Jacobs and Kevin Jon Heller took the position in their recent posts that the ICC had missed the train and that no formal investigation could be launched as of 27 October 2017. I agree with this conclusion but take a different route in arriving at it. 

Confidential request scenario

Before explaining why opening an investigation is in my view no longer an option, absent any material change in circumstances (such as Burundi re-acceding to the Statute or filing an Article 12(3) declaration), I should clarify that this argument, which I set out in Part II of this post, is limited to the scenario under which the Prosecutor had not filed a confidential request for an authorization to launch an investigation before the withdrawal became effective.

I will briefly consider here the alternative scenario, namely that the Article 15(3) was filed confidentially. There is nothing in Article 15(3), Rule 50, or Regulations 45 and 49 RoC to preclude a confidential request, and I do not rule out the possibility that the Prosecutor did file one prior to 27 October. This is still a missing piece of the factual puzzle at present and it will be dispositive of the possibility to open an investigation in the aftermath of the withdrawal.

On the one hand, one would think of a confidential request as an unusual and unlikely move for the Prosecutor. Firstly, as Dov has noted, this would depart from previous practice: past requests under Article 15(3) were filed as public documents (with confidential annexes). Second, going confidential seems peculiar in the situation complicated by the impending withdrawal—which is a factor arguably calling for transparency rather than confidentiality—absent any (self-evident) situation-specific reasons for keeping both the Regulation 45 notice and the Article 15(3) application under seal. On the other hand, I admit that the OTP possibly had good grounds for preferring to proceed confidentially, which are not (yet) in the public domain relating, for example, to the serious security risks for information providers.

If the Prosecutor did make a confidential Article 15(3) request, I believe it could be decided upon by the PTC even after 27 October 2017, with the possibility of giving the green light to the investigation. The second part of the second sentence of Article 127(2) allows for a ‘continued consideration of [the] matter which was already under consideration by the Court prior to the date on which the withdrawal became effective’. The ‘matter’ here would be the OTP request under Article 15(3) (as opposed to the preliminary examination as such, which, as I explain below, is a different ‘matter’); the question before the judges being whether there is ‘a reasonable basis to proceed with an investigation, and [whether] the case appears to fall within the jurisdiction of the Court’ (Article 15(4)). The notion of ‘matter’ is sufficiently broad (and vague) to cover an OTP request to open investigation. Since it would be the PTC judges who would be seized of it at the time when the withdrawal became effective, it is not an issue whether ‘the Court’ in the second sentence of Article 127(2) refers to the judiciary alone or also includes other organs (such as the OTP).

If the (hypothetical) confidential request to open an investigation was indeed filed before the expiry of the Article 127(1) term and is going to be authorized by the PTC after that deadline, the investigation cannot be considered as having been ‘commenced prior to the date on which the withdrawal became effective’. In that case, Article 127(2) exempts Burundi from an obligation to cooperate with the Court in connection with the investigation. It is questionable whether opening the investigation within one year since the withdrawal notice rather than later would have made any difference in terms of the availability of cooperation. As others note, Burundi would likely be unwilling to cooperate with the Court even if it were under an obligation to do so. But it does matter, both legally and symbolically. If the investigation had been commenced before 27 October 2017, the Court would have been legally entitled to demand cooperation from the former State Party in connection with the investigation, possibly also giving it a stronger leverage with actors capable of inducing cooperation politically. Burundi would have also had a more difficult time justifying any failures to perform its cooperation duties.

In Part II of this post, I will look at the scenario under which no confidential Article 15(3) request had been filed before the withdrawal became effective (while accepting the possibility that the future will prove me wrong). It is important to consider, as a matter of law, whether a proprio motu investigation may be initiated in respect of a situation in a State that has effectively withdrawn from the Statute.

Initiation of an investigation by the OTP post-withdrawal

As I argued previously, no investigation in the Situation in Burundi can be initiated after 27 October 2017 unless the OTP made a confidential request under Article 15(3) prior to the date on which the withdrawal became effective (the scenario I considered in Part I of this post). The ICC retains (temporal, material, and personal/territorial) jurisdiction over the crimes committed while Burundi was a party. But as that jurisdiction remains dormant until it is activated under Article 13(c), and the expiry of the one-year period precludes such activation, the Court cannot exercise it by way of letting the situation progress to the formal investigation. Thus, while I agree with the authors of the Amnesty HRIJ blog that there is jurisdiction (in principle), I reject their claim that opening an investigation remains a possibility.

First, the distinction between mere existence of jurisdiction and the Court’s ability to exercise it, is important to maintain. It is the cornerstone of the Court’s jurisdictional mechanism which consists of two elements – [1] the preconditions to the exercise of jurisdiction (or dormant jurisdiction, as aptly phrased by Héctor Olásolo in his excellent 2005 book, p. 39) (Article 12); and [2] triggering mechanisms set out in Article 13. For the Court to be able to exercise the (otherwise dormant or latent) jurisdiction which it possesses in respect of crimes under Articles 5, 11 and 12, the jurisdiction needs to be triggered either by a State Party referral (Article 13(a)) or by the Prosecutor through the initiation of an investigation proprio motu (Article 13(c)).

Article 12(2)(a), (c) and (3) establishes the parameters of the acceptance by relevant States of the Court’s jurisdiction as the precondition to its exercise by the Court whilst Article 13 (‘Exercise of Jurisdiction’) governs the activation of that jurisdiction. It provides that ‘[t]he Court may exercise its jurisdiction with respect to a crime referred to in article 5 in accordance with the provisions of this Statute if: … (c) The Prosecutor has initiated an investigation in respect of such a crime in accordance with article 15.’ This means that if the Prosecutor has not triggered the ICC’s jurisdiction by initiating an investigation, i.e. by filing with the PTC a request to authorize it to open investigation accompanied by supporting material, the jurisdiction remains dormant and cannot be exercised by the Court. In case of Burundi, the ICC’s jurisdiction was not activated by the time when the withdrawal became effective since under the scenario I am considering here the Prosecutor did not make an Article 15(3) request prior to 27 October 2017.

As the second step, I posit that Article 127(2) precludes the activation of the Court’s jurisdiction by the Prosecutor in accordance Articles 13(c), 15(3) and 53(1) of the Statute in the post-withdrawal stage. Rather than reproducing the text of Article 127(2) verbatim, it would be helpful to paraphrase it in positive terms: [1] withdrawal discharges a State from any obligations that would have arisen for it under the Statute had it remained a Party, including any financial obligations [1^st sentence]; [2] its withdrawal shall affect cooperation with the Court in connection with criminal investigations and proceedings in relation to which the State would have otherwise had a duty to cooperate, and which had not been commenced prior to the date on which the withdrawal became effective [2^nd sentence; 1^st part]; and [3] withdrawal shall prejudice the consideration of any matter that was not already under consideration by the Court prior to the date on which the withdrawal became effective [2^nd sentence; 2^nd part].

As noted by other commentators, Article 127(2) of the Statute reflects and details for the purpose of the Statute the standard of Article 70 of the VCLT which governs the consequences of the termination of a treaty. Although this is open to debate, it is arguable that Article 70 VCLT informs the interpretation and application of Article 127(2) and/or applies subsidiarily pursuant to Article 21(1)(b) of the Statute. According to Article 70 VCLT, a State’s withdrawal from a multilateral treaty releases it from ‘any obligation further to perform the treaty’ but ‘[d]oes not affect any right, obligation or legal situation of the parties created through the execution of the treaty prior to its termination’. This implies that withdrawal from a multilateral treaty does affect any right, obligation or legal situation of the parties that was not created through the execution of the treaty prior to withdrawal.

At this juncture, let us apply the elements of the provision of Article 127(2), as paraphrased for convenience and seen in light of Article 70 VCLT, to the situation of Burundi.

[1] The scope of obligations the withdrawing State is released from as per the 1^st sentence of Article 127(2) is unqualified and subject only to a temporal limitation as it covers all putative obligations under the Statute postdating the withdrawal. The former State Party is exempted from further performing not only the duties of cooperation vis-à-vis the Court (specifically, yet only partially, addressed in the 2^nd part of the 2^nd sentence). The exemption also covers any obligations constituting the ‘legal situation of the parties created through the execution’ of the Rome Statute. These include duties vis-à-vis States Parties which, if it had not withdrawn, would have otherwise arisen for the State in question on account of the Court’s exercise of its complementary jurisdiction in respect of the crimes over which that State would normally exercise jurisdiction.

The ‘legal situation’ in connection with the Court’s exercise of jurisdiction encompasses, among others, the State Party’s duties to submit oneself to the ICC’s scrutiny as part of the complementarity analysis and to its admissibility determinations, as well as the duty to accept any prospective investigation or prosecution stemming from the activation of the Court’s jurisdiction by the Prosecutor in accordance with Article 15(3). For the withdrawing State to continue to be bound by these obligations post-withdrawal under Article 127(2), the Court must have commenced to exercise its jurisdiction before withdrawal becomes effective. As noted, this has not happened in case of Burundi under the present scenario and the Court’s jurisdiction remained dormant. Therefore, the Prosecutor’s attempt to trigger it after 27 October 2017 would fall outside of the scope of the 1^st sentence of Article 127(2) and Article 70(1)(b) VCLT insofar as it would amount to nothing else than seeking to create a new legal situation through a continued execution of the treaty as if the withdrawal did not happen.

For this reason, I find the Amnesty HRIJ blog authors’ VCLT-based argument not sufficiently sensitive to the nuances of the ICC’s jurisdictional setup and ultimately not convincing in justifying the conclusion that the Prosecutor may initiate investigation in Burundi following the withdrawal. To quote a passage from Olásolo’s book (p. 135): “the lack of activation of the Court’s dormant jurisdiction over the situation in which such crimes were allegedly committed prevents the assertion of the existence of a “right, obligation or legal situation of the parties created through the execution of the treaty prior to its termination.”

[2] The 1^st part of the 2^nd sentence of Article 127(2) does not concern the implications of the effectuation of the withdrawal for the Court’s jurisdiction. Rather, it is limited to its effects on the withdrawing State’s cooperation duties. For these duties to continue to apply to a former State Party in connection with any specific investigations and proceedings, these activities must have commenced prior to the date on which the withdrawal became effective. It is clear that if no proprio motu investigation in Burundi was ‘initiated’ by the OTP before 27 October 2017, none has been ‘commenced’ prior to that date either. The moment of the commencement of a proprio motu investigation is marked by its authorization, which is within the PTC’s competence. In the scenario I am considering, the PTC was not seized of the matter, and much less decided on it, by the expiry of the Article 127(1) period.

Even if one admits arguendo that the investigation could be opened after 27 October (which I do not believe is legally possible), there is no way Burundi would have a duty to cooperate in connection with this investigation. This is a valid policy consideration in itself that militates against the OTP seeking an authorization to investigate in the post-withdrawal stage. This concern has been raised by commentators (notably, by Alex Whiting) who, without making an argument on jurisdiction similar to mine, emphasized the importance for the OTP, in case it intended to launch an investigation and/or prosecutions in a withdrawing State, to care to do so while that State was still a party to the Statute. This is a matter of due diligence that goes to securing a waterproof legal basis for demanding and obtaining cooperation that is essential for effective investigations. As noted by Roger Clark, ‘[a] vigilant Prosecutor who bears pending withdrawals in mind will need to be careful to commence proceedings formally within the relevant time-frame.’ (R. Clark, ‘Article 127’, Triffterer/Ambos (eds), The Rome Statute, 3^rd ed. (Beck–Hart–Nomos, 2016) 2324, emphasis added).

[3] Finally, the basic point of the 2^nd part of the 2^nd sentence of Article 127(2) is that withdrawal prevents the consideration of any matter not already under consideration by the Court prior to the effectuation of the withdrawal. I am not convinced by the arguments positing the possibility of a post-withdrawal Article 15(3) request with reference to this limb of Article 127(2). Kevin and Dov may have a point when they challenge the notion that the ‘Court’ in this article should be read to include the OTP, and that a preliminary examination can properly be deemed a ‘matter’, respectively, although I am not interested to consider whether the interpretations they take issue with are invalid and implausible. I generally find these terms too elastic, arbitrary, and elusive to serve as the basis for any conclusive argument. The somewhat haphazard choice of terminology and less than impeccable coordination in the drafting of the different parts of the Statute sets limits on the cognitive and practical utility of systemic or contextual interpretation.

Instead, I do not think one can rely on the last sentence of Article 127(2) for extending the OTP’s power to file an Article 15(3) request to the post-withdrawal stage because the ‘matter’ underlying the request for an authorization to investigate coming before the PTC is materially and procedurally distinct from the ‘matter’ under consideration by the ‘Court’ (in the broad sense) in the context of the preliminary examination. In other words, the matter of the initiation of investigation (Article 15(1)-(3)) is distinguishable from that of its commencement (Article 15(4)), among others, given the different actors, legal bases, process, and substance. These ‘matters’ are similar due to an overlap in subject-matter and temporal proximity in the procedural chronology. But they are definitely not identical since the ‘consideration’ of each of them is carried out by different organs of the Court and is subject to distinct provisions. Presenting them as one and the same ‘matter’ in the singular would be stretching this notion to the point where it loses any meaning. The consideration of the OTP’s request by the PTC judges in accordance with Article 15(4) and Rule 50, aims to establish whether there is a reasonable basis to proceed and whether an investigation should be authorized. By contrast, the OTP’s preliminary examination (which does have a formal status in the ICC regime since it is mentioned in the Statute, even if once, – in Article 15(6)), is governed by distinct provisions (Article 15(1)-(3) and Rules 46-50) and directed towards the former determination but not the latter, which is for the Chamber to make. Thus I do not see why, despite the continuity between the Article 15(3) and Article 15(4) determinations, they should be considered as the same ‘matter’. Therefore, even though the preliminary examination in Burundi can continue post-withdrawal as per the last sentence of Article 127(2), it cannot progress to a formal investigation under that provision.

Conclusion

The fact that the Court retains jurisdiction over the crimes committed in Burundi while it was a State Party to the ICC Statute, does not mean that it is still possible for the Prosecutor to seek authorization to investigate and for the Court to exercise its jurisdiction over that situation after the withdrawal became effective. Statements to the effect that the withdrawal ‘does not affect’ the Court’s jurisdiction are not exactly accurate. While the dormant jurisdiction (or the preconditions to its exercise) is not thereby removed, the exercise of jurisdiction by the Court is precluded.

The OTP may not and should not request authorization to open investigation in Burundi after 27 October. This is so not only because it would be problematic for policy reasons—which I will not go into here other than mentioning the thin prospects of obtaining cooperation from this former State Party—, but also because such a step would be inconsistent with the Statute or, at the very least, rest on a shaky legal foundation. The OTP’s own legal analysis from one year ago may be taken as indirectly acknowledging this. It would be ill-advised to try to compensate for its inaction under Article 15(3) in the course of one year since the withdrawal notice (if there indeed was such inaction).

Regrettable as it were, the Burundi ship has sailed and the ICC should let go – at least until such time as the exercise of its jurisdiction might be rendered possible, for example, by Burundi’s accession to the Statute, Article 12(3) declaration, or a UNSC referral. If the history of States’ political maneuvering with international criminal justice institutions can teach us anything, it is perhaps that everything is possible and nothing is guaranteed.