Tom Dannenbaum is a Graduate Associate in the Law and Public Affairs Program at the Woodrow Wilson School of Public and International Affairs at Princeton University. He is author of Translating the Standard of Effective Control into a System of Effective Accountability: How Liability Should be Apportioned for Violations of Human Rights by Member State Troop Contingents Serving as United Nations Peacekeepers, 51 Harv. Int’l L.J. 113 (2010)
In an earlier post, I reported on the Hague Court of Appeal’s decisions in Nuhanović v. The Netherlands and Mustafić-Mujić et al v. The Netherlands regarding the wrongdoing of Dutchbat at Srebrenica. Here, I examine the Court’s holding on the attribution of that wrongdoing to the Dutch state.
The decisions provide stronger and clearer jurisprudential affirmation of the principles of “effective control” and dual attribution than does the Grand Chamber’s judgment in Al-Jedda v. United Kingdom (handed down just two days later). Moreover, the Court of Appeal’s elaboration of “effective control” establishes several key features of the concept as applied in the peacekeeping context. First, the “effective control” analysis should be applied equally to the contributing state and the receiving international organization. Second, “effective control” includes not just giving orders, but also the capacity to prevent the wrongdoing. Third, though the Court’s position on this is slightly more ambiguous, troop-contributing states may sometimes hold that “power to prevent” in virtue of their authority to discipline and criminally punish their troops for contravening U.N. orders. I would go beyond the Court’s reasoning on this third feature to add that the state’s authority with respect to selecting and training troops and contingent commanders is also relevant in this regard.
Since the decisions do not differ on any significant matters of substance, the citations below are to Nuhanović, but apply equally to Mustafić-Mujić.
The District Court had attributed the impugned conduct of Dutchbat exclusively to the U.N. on the grounds that it had “operational command and control” over the peacekeeping force and the state had not acted to negate that control by “cutting across” the U.N. command structure or withdrawing Dutchbat from the structure. (paras. 4.8-4.14.5, Nuhanović District Court [the District Court holdings are available here and here]). Elaborating this standard of attribution, the District Court explained:
If Dutchbat was instructed by the Dutch authorities to ignore UN orders or to go against them, and Dutchbat behaved in accordance with this instruction from the Netherlands, this constitutes a violation of the factual basis on which the attribution to the UN rests. This then creates scope for attribution to the State. The same is true if Dutchbat to a greater or lesser extent backed out of the structure of UN command, with the agreement of those in charge in the Netherlands, and considered or shown themselves as exclusively under the command of the competent authorities of the Netherlands for that part. If, however, Dutchbat received parallel instructions from both the Dutch and UN authorities, there are insufficient grounds to deviate from the usual rule of attribution. (Para. 4.14.1, Nuhanović District Court) [emphasis added].
Presumably in an effort to avoid this framework and/or the “ultimate authority and control” standard articulated by the Grand Chamber in Behrami and Saramati (para. 133, Behrami and Saramati), the plaintiffs argued before the Court of Appeal that attribution should be determined as a matter of Bosnian law (para. 5.2). The Court rejected this argument, finding attribution in such circumstances to be solely a matter of international law (para. 5.3) and reasoning in the alternative that in the absence of Bosnian legal provisions on the issue, the gap would anyway be filled by international law (para. 5.4).
However, departing from both the District Court and Behrami and Saramati, the Court of Appeal identified the international law standard of attribution to be “effective control.” Two days later, the Grand Chamber granted the “effective control” rule further jurisprudential authority in Al-Jedda. In several respects, the Court of Appeal’s analysis is more satisfying in this regard. The first relevant paragraph reads as follows:
In international law literature, as also in the work of the ILC, the generally accepted opinion is that if a State places troops at the disposal of the UN for the execution of a peacekeeping mission, the question as to whom a specific conduct of such troops should be attributed, depends on the question which of both parties has ‘effective control’ over the relevant conduct. Cf. M. Hirsch, The Responsibility of International Organizations Towards Third Parties: Some Basic Principles (1995) p. 64; F. Messineo, NILR 2009 p. 41-42; A. Sari, Human Rights Law Review 2008 p. 164; T. Dannenbaum, Harvard International Law Journal 2010 p. 140-141. This opinion has also found expression in the Draft Articles on the Responsibility of International Organizations of the ILC, of which Article 6 [now Article 7] reads as follows:
The conduct of an organ of a State or an organ or agent of an international organization that is placed at the disposal of another international organization shall be considered under international law an act of the latter organization if the organization exercises effective control over that conduct.
Although strictly speaking this provision only mentions ‘effective control’ in relation to attribution to the ‘hiring’ international organization, it is assumed that the same criterion applies to the question whether the conduct of troops should be attributed to the State who places these troops at the disposal of that other international organization. (Para. 5.8) [emphasis added].
Two things are worthy of note here. First, unlike the Grand Chamber in Al-Jedda, the Court of Appeal is categorical in holding that “effective control” is the applicable rule of attribution. Compare Marko Milanovic’s analysis of the “vastly entertaining” equivocations in para. 84 of Al-Jedda as to whether “effective control” or “ultimate authority and control” applies. Second, rather than adopting a presumption that conduct is attributable to one of the two entities (the U.N. or the state) and then examining whether the other entity might usurp responsibility in virtue its effective control, the Court instead holds that the effective control analysis applies equally to both entities. As articulated in the article cited by the Court, I think this is the correct interpretation (51 Harv. Int’l L.J. 113, 141).
The necessary implication of this interpretation is that “effective control” over conduct cannot refer only to direct orders resulting in that conduct. Much unlawful action occurs directly contrary to superior orders or authorizations. Rather than applying a default attribution to one of the two entities when orders are contravened, this interpretation of the standard requires a deeper analysis as to the forms of control exercised by the relevant entities. In the absence of effective orders, this means considering the capacity to prevent the wrongdoing – i.e. that control that could have an effect on the conduct. The Court recognized this implication,
The question whether the State had ‘effective control’ . . . must be answered in view of the circumstances of the case. This does not only imply that significance should be given to the question whether that conduct constituted the execution of a specific instruction, issued by the UN or the State, but also to the question whether, if there was no such specific instruction, the UN or the State had the power to prevent the conduct concerned. (Para. 5.9) [emphasis added]
As noted by the ILC in its Commentary to the Draft Articles on the Responsibility of International Organizations, “When giving a wide meaning to the concept of ‘effective control’ so as to include also the ability to prevent, the Court followed the approach taken by T. Dannenbaum, ‘Translating the Standard of Effective Control into a System of Effective Accountability.’” [U.N. Doc. A/66/10 p.91, n.129; see also André Nollkaemper on this point]. In the cited article, I argue for an interpretation whereby effective control “is held by the entity that is best positioned to act effectively and within the law to prevent the abuse in question.” (51 Harv. Int’l L.J. 113, 157). I return to the Court’s application of this “power to prevent” standard below.
First, it is worth noting that the Court of Appeals also prefigured Al-Jedda in accepting the possibility of dual attribution. Specifically:
[T]he Court adopts as a starting point that the possibility that more than one party has ‘effective control’ is generally accepted, which means that it cannot be ruled out that the application of this criterion results in the possibility of attribution to more than one party. For this reason the Court will only examine if the State exercised ‘effective control’ over the alleged conduct and will not answer the question whether the UN also had ‘effective control’. (Para. 5.9).
The Court of Appeal’s position is again more satisfying than that of the Grand Chamber. While the latter implied the possibility of dual attribution in such contexts (para. 80, Al-Jedda), the Court of Appeal stated that possibility explicitly. Moreover, unlike the Grand Chamber, which indicated a lingering preference for an “either/or” standard of attribution (considering and rejecting attribution to the U.N. [not a party to the litigation] before attributing the relevant acts to the U.K (paras. 80-86, Al-Jedda)), the Court of Appeal cut straight to an assessment of whether the Dutch state had effective control over the conduct in question. This suggests that the fact that the conduct may be attributable to one entity is irrelevant to its potential attribution to the other.
In applying “effective control”, the Court of Appeal first emphasized the U.N. command structure under which Dutchbat operated. (para. 5.7). The battalion was subject to a U.N. chain of command running from UNPF Force Commander Lt. Gen. Bernard Janvier, through UNPROFOR HQ Commander Lt. Gen. Rupert Smith and his Deputy Commander Gen. Hervé Gobillard. In principle, Dutchbat was supposed to take orders exclusively from these commanders, not from the Dutch government.
However, the Court of Appeal found that following the fall of Srebrenica an extraordinary situation arose in which the Dutch government became more actively involved in decision-making than would ordinarily be the case in a U.N. force. The Court explicitly distinguished this context from that which pertained in Behrami and Saramati (para. 5.11). Once Srebrenica fell, the Netherlands decided together with the U.N. that Dutchbat would be withdrawn from the peacekeeping mission. (paras. 5.10-5.12). This initiated a “transition period” prior to the actual withdrawal “in which matters in Potocari were being completed [including] the aid to and the evacuation of the refugees.” (para. 5.17).
During this transition period, besides the UN, the Dutch Government in The Hague had control over Dutchbat as well, because this concerned the preparations for a total withdrawal of Dutchbat from Bosnia and Herzegovina. In this respect [Dutch General Cornelis Nicolai – Smith’s Chief of Staff at UNPROFOR HQ] fulfilled a double role because he acted on behalf of the UN and also on behalf of the Dutch Government. The fact that The Netherlands had control over Dutchbat was not only theoretical, this control was also exercised in practice: the Government in The Hague, represented by two of its highest military officers, Van den Breemen and Van Baal, together with Janvier took the decision for the evacuation of Dutchbat and of the refugees, Minister Voorhoeve gave the instruction that Dutchbat was not allowed to cooperate in a separate treatment of the men, and he told Karremans that he had to save as much as possible. Through the intermediary of Nicolai in his double role, the Dutch Government also gave orders to Karremans regarding the evacuation. According to the judgment of the Court, in all these cases it was a matter of orders being given and not just transmitting the wishes or expressing the concerns which Nicolai understood very well (‘if the Dutch Government says something like that, as a military officer you just carry it out’). . . . . (para. 5.18) [emphasis added].
The Court concluded that the Dutch government
would have had the power to prevent the alleged conduct if it had been aware of this conduct at the time. The facts do not leave room for any other conclusion than that, in case the Dutch Government would have given the instruction to Dutchbat not to allow [Hasan Nuhanović’s brother] (as well as his father Ibro Nuhanovic) to leave the compound or to take him along respectively, such an instruction would have been executed. (para. 5.18) [emphasis added]
Therefore, the Dutch Government possessed “effective control” over the conduct of Dutchbat regarding the relevant evictions (para. 5.20).
Notably, the Court held that no order from the Dutch government or from the U.N. required Dutchbat to evict the Mustafić or Nuhanović families or to cooperate with the BSA “evacuation” once the risks to able-bodied male refugees became apparent (para. 6.8). The Dutch government’s “effective control” therefore took the form of the “power to prevent,” rather than “direct control” over the conduct in question (para. 6.20).
Overall, despite the mysteriousness of the “transition period” on which it relies, the Court’s application of the “power to prevent” interpretation to the unusual facts before it is plausible. However, the Court could have reached the same conclusion via a simpler application of that interpretation of “effective control.” Doing so would have provided a more broadly applicable precedent, albeit one that would have required a more explicit departure from Behrami and Saramati – something the Court seemed reluctant to do (raising the question of whether the reasoning would have differed had the decisions been authored after Al-Jedda).
Interestingly, the Court of Appeal does consider and grant some support to this simpler approach. Crucial here is Gen. Gobillard’s 11 July 1995 order requiring that Dutchbat “[t]ake all reasonable measures to protect refugees and civilians in your care.” (Para. 2.16). The Court found that Dutchbat “acted contrary to [Gobillard’s] instruction … [because] it had become clear at the latest at the end of the afternoon on l3 July 1995 that the evacuation of the men was (had become) life-threatening.” In this context, Gobillard’s order to protect became the overriding command to which Dutchbat ought to have been responsive. (Para. 6.8).
In the above-cited article, I contend that when peacekeepers engage in wrongful conduct contrary to U.N. orders, those acts must be attributed to the troop contributing state. Specifically, I argue, “by disobeying [the U.N. Commander’s] direct orders, peacekeepers undermine the bases for effective U.N. control. The acts fall beyond the reach of that control precisely because by their very nature they deny the efficacy of that control.” (51 Harv. Int’l L.J. 113, 160). U.N. control is undermined in such contexts because the U.N. Commander’s authority typically includes only the authority to issue orders. Troop-contributing states retain exclusive authority to select the members of the troop contingent, to appoint contingent leaders and define the internal hierarchy, to train the troops (including on their legal responsibilities), to discipline and/or dismiss from the contingent those that breach orders, and to punish those that commit crimes. “Through acting with purpose and vigor along these dimensions, states can exert considerable effective control” to ensure conduct in compliance with lawful orders. (51 Harv. Int’l L.J. 113, 164). Indeed, these are the only mechanisms by which conduct in breach of direct orders can be prevented. Therefore, effective control with respect to that class of conduct rests exclusively with the state. (51 Harv. Int’l L.J. 113, 164).
Under this interpretation, the finding that the relevant Dutchbat conduct contravened Gobillard’s order would be sufficient to attribute that wrongful action to the state, even if the Dutch government had not been involved so intimately in decision-making regarding Dutchbat at that point. Although the Court did not rely exclusively on this approach, it seems to have accepted the basic theory. Abstractly, it found,
When applying the ‘effective control’ criterion it is important to establish that it is not disputed that the state that provides the troops keeps control over the personnel matters of the assigned soldiers, who are and will remain employed by the state, as well as the power to take disciplinary action and start criminal proceedings against these soldiers …. (para. 5.10).
Concretely, the Court held,
[I]t is important [to the issue of attribution] that … the alleged conduct was contrary to the instruction given by General Gobillard to protect the refugees as much as possible, and that the State held it in its power to take disciplinary actions against that conduct. (para. 5 .18) [emphasis added].
The italicized sections emphasize the Court’s correct understanding that when orders are contravened, the relevant levers of control over the conduct are those that impact discipline and obedience to lawful commands. Because in peacekeeping forces those levers are held exclusively by the state, it is the state that has the control most likely to be effective in preventing wrongdoing that contravenes U.N. orders.
Due to the broader applicability of this analysis, the Court’s endorsement in this regard may prove to be of greater doctrinal consequence in elaborating the concept of “effective control” than its more detailed examination of the Dutch government’s involvement in decision-making during the “transition period” prior to full withdrawal. In either case, however, the adoption of the “power to prevent” approach to “effective control” analysis represents an important advance on what has come before.
The Court of Appeal’s decisions provide robust jurisprudential support for the principles of “effective control” and dual attribution with respect to the conduct of state organs or agents placed at the temporary disposal of an international organization. Elaborating the “effective control” standard, the Court established: (i) the “effective control” rule of attribution applies equally to both entities under examination; (ii) “effective control” includes not just giving orders, but also the capacity to prevent the wrongdoing; (iii) troop-contributing states may hold the “power to prevent” in certain contexts in virtue of their authority to discipline and criminally punish their troops for contravening U.N. orders. I would suggest that the state’s authority with respect to selecting and training troops and contingent commanders is also relevant in this regard.
The prospect of reversal by the Dutch Supreme Court is limited. The Supreme Court cannot review findings of fact or interpretations of foreign law. Moreover, the Al-Jedda decision makes reversal on the international law issue of attribution significantly less likely than might have been the case when Behrami and Saramati was the principal international precedent. Nuhanović and Mustafić-Mujić are therefore likely to remain landmark decisions on this topic.