Dual attribution: liability of the Netherlands for removal of individuals from the compound of Dutchbat

 Posted by: Nollkaemper André

On 5 July 2011, the Court of Appeal of the Hague decided that the State of the Netherlands had acted unlawfully and is liable for evicting Bosnian nationals from the compound of Dutchbat in Srebrenica on 12 July 1995. Ibro Nuhanovic, Muhamed  Nuhanovic, Nasiha Nuhanovic and Rizo Mustafic were subsequently killed by Bosnian Serbs, as part of what the ICTY and the ICJ later found to be acts of genocide.

The decision adds another chapter to the tortuous attempt of the Netherlands to cope with its multiple failures, with dramatic consequences, in its policies and decisions regarding the conduct of Dutch peacekeeping troops in Srebrenica in 1995, In 2002, the Government of then prime minister Wim Kok resigned after a report held it partly to blame for the failure to offer protection in Srebenica. At the time, Kok said that he accepted political, but no legal responsibility. Almost ten years later, the Court of Appeal has made clear that the responsibility is not only political, but that the Dutch policy in regard of Srebrenica also has engaged its legal liability.

When not overturned by the Supreme Court, the decision of the Court of Appeal will stand as a groundbreaking ruling on the possibility of dual attribution of an act (or omission) to the United Nations (UN) and a troop contributing state, in this case the Netherlands.

Ibro Nuhanović, Nasiha Nuhanović and Muhamed Nuhanović were respectively the father, mother and brother of Hasan Nuhanović, an interpreter who worked for Dutchbat, a Dutch battalion under command of the United Nations peacekeeping force UNPROFOR. On 12 July 1995, he was with other refugees inside the compound protected by Dutchbat. After Srebrenica had fallen, the refugees were taken away by the Bosnian Serbs. Local staff of Dutchbat who were employed by the UN and had a UN identity card were allowed to stay in the compound. Hasan Nuhanović had such a card, but Ibro, Nasiha and Muhamed Nuhanović were compelled to leave the compound the following day. All three were killed.

Rizo Mustafic had since 1994 been working as electrician for Dutchbat. After the fall of Srebrenica, Mustafic had sought refuge on the compound with his family. Though he wanted to stay, he was removed from the compound. He then was separated from his family and killed by Bosnian Serbs.

Relatives of Nuhanovic and Mustafic brought a tort claim against the State of the Netherlands. Their main claim was that the State had committed a wrongful act (tort) and that it should compensate any damage incurred as a consequence of that wrongful act.

In September 2008, the District Court of the Hague rejected the claims of the families of Nuhanovic and Mustafic against the Netherlands holding that the acts of Dutchbat could only be attributed to the UN. Nuhanovic and Mustafic then appealed.

The Court of Appeal in the Hague quashed the first instance judgment of the District Court and found that the disputed conduct was attributable to the Netherlands, and that the Netherlands had acted wrongfully. Three aspects of the decision require brief comments: 1) the disputed conduct 2) attribution of the conduct 3) wrongfulness of the conduct.

1. THE DISPUTED CONDUCT

An important preliminary point is that the Court only rules on the actual removal of Nuhanovic and Mustafic from the compound, not on any failure of the Netherlands to subsequently protect them.

Plaintiffs based their claims on both the removal of Mustafic and Nuhanovic from the compound and on the failure to intervene when Mustafic and Nuhanovic were (outside the compound) separated from their relatives.

The Court only rules on the first of these claims. The plaintiffs had asked that the State had committed a wrongful act (tort) and that it should compensate any damage incurred as a consequence of that wrongful act. Since that claim could be sustained on the basis of the findings of the Court in regard to the first claim, considering the second claim (however interesting that would have been for other cases) would not have led to a different outcome and the Court thus does not consider that claim (par. 6.22). This means that the judgment in respect of the wrongfulness of the conduct of the Netherlands in principle is strictly limited to the active removal from the compound.

This limits the possible relevance of the decision for other claims in connection to the conduct of Dutchbat in Srebrenica or comparable peacekeeping operations. However, the Court did also rule on the legal implications of the instructions of general Gobillard (UNPROFOR) ‘to take all reasonable measures to protect refugees and civilians in your care’, and thus may have broader implications for obligations to protect of peacekeeping forces and troop contributing states.

2. FINDINGS ON ATTRIBUTION OF CONDUCT

The Court of Appeal holds that the question of attribution should be decided solely on the basis of international law. Plaintiffs had argued that the question of attribution should have been decided on the basis of national (ie Bosnian) law, probably in view of the case-law on attribution of the European Court on Human Rights (ECtHR), which present substantial hurdles to a finding of attribution of conduct to a troop contributing state.

The Court rejects the argument that it should rule on attribution on the basis of Bosnian law and states that the question whether troops have been put at the disposal of the UN, and what the contents and consequences of an agreement to that effect are (including the consequences for a civil liability claim) should be assessed on the basis of international law (par. 5.3.2).

The Court adds however that even if it would have answered the question under Bosnian law, the outcome would have been the same since Bosnian law does not contain a relevant rule on attribution and that the Court therefore should have relied on international law to fill the gap (par. 5.4).

Against this background, the decision of the Court of Appeal essentially rests on a three-fold construction of the principles of attribution.

◆    First, it determines that the standard for attribution is ‘effective control’. In doing so, it rejects the standard for attribution of conduct that was used by the District Court (‘operational overall control’), the standard used by the ECtHR in Behrami (‘ultimate authority and control’) as well as the basis of attribution that the UN itself believes to be correct when it claims that peacekeeping troops are to be considered as subsidiary organs of the UN.  However, the Court does align itself with the criterion formulated in article 7 of the Draft Articles of the International Law Commission on the Responsibility of International Organizations, as adopted by the International Law Commission on second reading on 3 June 2011. This article has had broad support from states, and this aspect of the decision of the Court thus rests on solid ground.

◆    Second, the Court takes the position that for determining whether the state had effective control it is not only relevant whether the act was an implementation of a specific instruction by either the United Nations or the State,

‘but also whether, if there is no such specific instruction, if the United Nations or the State was able to prevent the conduct in question.’ (par. 5.9, author’s translation).

That is: the removal of Nuhanovic and Mustafic from the compound could be attributed to the Netherlands, if the Netherlands was able to prevent that removal. The exact language used by the Court is a matter of some importance. The Court says, in Dutch, that for determining whether there is effective control, it is relevant whether the UN or the Netherlands ‘het in zijn macht had het het desbetreffende optreden te voorkomen’. This could be translated by ‘had the power to prevent the conduct in question’, but because of the various connotations of ‘power’, the translation in terms of ‘being able’ is to be preferred. In any case both the Dutch words and the translation can either be interpreted in legal or factual terms, and their construction follows from the remainder of the judgment, as analysed below.

◆   Third, directly implied by the preceding point, the Court states that

‘it is possible that more than one party has ‘effective control’, and that it therefore is not excluded that application of this criterion leads to attribution of conduct to more than one party’. (par. 5.9)

In view of its finding on the possibility of dual attribution, the Court leaves aside the question whether the United Nations also possessed effective control, and it proceeded to examine whether the State exercised effective control over the disputed action. That it could do so follows from the individual nature of attribution. In such a case of possible dual attribution, the question whether an act can indeed be attributed to the UN would not affect its attribution to the Netherlands. Under the principles adopted by the ILC, it would also not affect the liability of the Netherlands and its possible obligation to provide reparation.

The Court does not elaborate on the first of these three steps (it limits itself to referring to the ILC Commentary), nor on the third (other than stating that the possibility of dual attribution is ‘generally accepted’ – somewhat of an overstatement, but the third step as such rests on solid ground). In contrast, much of the decision explains, in not always clear reasoning, what it means to exercise effective control, that is, in the construction of the Court, not only whether the state gave a specific instruction or whether the state was ‘able to prevent the conduct in question’.

The construction of effective control in terms of ability to prevent an act seems to differ from the ILC’s construction, which emphasizes ‘the factual control that is exercised over the specific conduct taken by the organ or agent placed at the receiving organization’s disposal’ (par 3 of the 2009 Commentary to what now is art. 7 of the DARIO). But this certainly does not mean that is therefore wrong.

Interestingly, this approach is quite close to a position defended by Tom Dannenbaum in a recent seminal piece on attribution on peacekeeping operations (‘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)). This piece was relied on by counsel for the plaintiffs and is cited by the Court (though on another point). Dannenbaum writes that ‘effective control (…) is held by the entity that is best positioned to act effectively and within the law to prevent the abuse in question’ (p. 158): His interpretation ‘aims at ensuring that the actor held responsible is the actor most capable of preventing the human rights abuse.’  The Court’s approach bears a close similarity to this reasoning.

Saying that a state exercises effective control in regard to a particular act if it is able to prevent that act may be open a wide door. Does a troop contributing state not always have the factual possibility to send orders or instructions to its nationals who serve in a UN operation if that is necessary to make them act in a certain way or to prevent them from acting in a certain way? If one accepts this position, conduct of peace keeping forces almost by definition can be attributed to the state (whether or not in parallel to the UN), since there was always the possibility for that the state to exercise control in a way that prevents the impugned conduct from occurring. In effect, this comes close to the position that some scholars have taken that conduct of contributing troops can always be attributed to both the sending state and the UN.

While there is much to say for this position from a theoretical and a normative perspective, there is little practice that supports this broad construction. Likewise, the comments by states and international organizations to the DARIO do not offer much support for this construction. The Court therefore rightly backs away from this broad construction of dual attribution.

The Court emphasis that effective control should be assessed in the concrete circumstances of the case, not in terms of an abstract possibility to exercise control. In its reasoning, whether control is ‘ effective control’ depends both on the factual and the formal aspects of control in a specific context.

As to the factual nature, the Court appreciates effective control in concreto and states that the disputed conduct in question (the removal of Nuhanovic and Mustafic from the compound) was directly connected to decisions and instructions of the government of the Netherlands. After 11 July, the UN and the Dutch government had decided to evacuate Dutchbat with the refugees. The Court found that during this process of evacuation, the Netherlands was actively involved in the process of evacuation. For instance, the Court notes that together with UN force commander Janvier, two Dutch military officials took, on behalf of the Dutch Government, the decision to evacuate Dutchbat and the refugees. The Court concludes that the government in the Hague actually instructed the Dutch military officials. It also notes that the Dutch General Nicolai fulfilled a double role because he acted both for the UN and for the government of the Netherlands.

It is in this factual context that the removal of Nuhanovic and Mustafic from the compound has to be assessed. The Court find this to be a consequence of the way in which the evacuation from the compound was organized and the way the instructions from the government were implemented (par 5.19-5.20).  The Court thus infers from the fact that the Netherlands Government in fact was closely involved in the evacuation, that the removal of Mustafic and Nuhanovic from the compound had to be attributed to the Netherlands. It tellingly notes that if the Government would have given Dutchbat the order to keep Mustafic and Nuhanovic on the compound, that order would have been implemented.

In other words: it is not the abstract possibility that the state could intervene and order its nationals who act as part of a peacekeeping mission to act in a particular way that triggers effective control. Rather, it was the specific factual situation in which the government in fact was so involved with the evacuation that it has be assumed that its orders would have been effective, that triggers effective control.

It might well be argued that even if the Netherlands had until that moment not been involved at all, an order from the Hague to keep Mustafic and Nuhanovic on the compound, would have been implemented just as well. In such a case, the Netherlands would not yet have exercised any form of control, but it always could have done so. However, that mere possibility would not have been sufficient as a basis of attribution in terms of the ILC’s construction, and it appears that likewise for the Court it was not the abstract possibility of control that mattered, but the actual exercise of control, that made the possibility of prevention more than a theoretical one.

However, the Court does not limit itself to a factual account of the involvement of the Netherlands government as a basis for its finding on effective control. The Court connects the ability to prevent an act (and thus the existence of effective control) with the legal power (or the normative control) to do so. On this point two aspects should be distinguished.

First, the Court emphasizes that a troop contributing state retains formal power in regard to personal and disciplinary matters, as well as to withdraw the troops (par 5.10). For instance, it states that the removal of Nuhanovic and Mustafic from the compound was contrary to the instructions of general Gobillard to protect refugees, and that the State had the authority (or: legal power) to take disciplinary measures against these acts (par .5.18).

It would seem that what the Court means is that when a battalion fails to act to protect civilians in a situation where it should offer such protection (certainly if that required by the mandate or by specific instructions given by the UN), the state not only can as a factual matter act, but also has the legal authority to do so. This points directly to possible attribution to the state. This finds some support in the ILC’s commentary, which argued that attribution based on effective control is linked with the retention of these powers by the state (par 6 of the 2009 Commentary to what now is art. 7 of the DARIO). This construction is also in line with the abovementioned construction of attribution by Dannenbaum, who argues that  effective control is held by the entity that is best positioned to act effectively ‘and within the law to prevent the abuse in question’ (p. 158, emphasis added).

While in this particular case this reasoning has a rather narrow scope (it is limited to conduct by which specific individuals were evicted from the compound against orders of the UN to protect them and indeed against both Bosnian and international law), there may a broader implication of this reasoning. If we accept an obligation of peacekeeping forces to protect (for instance on the basis of the Genocide Convention), it may be argued that the state should on that basis intervene or at least take disciplinary measures against those who act contrary to that obligation. Such an obligation thus could not only provide a ground for wrongfulness but, in the reasoning of the Court of Appeals, also a ground for attribution. Though plaintiffs had relied on the Genocide Convention and in particular the ICJ’s interpretation of prevention of genocide, the Court of Appeals does not pursue this – neither in relation to attribution nor in relation to  wrongfulness (as it finds wrongfulness on other grounds).

Second, the Court appears to attribute legal relevance to the fact that the peacekeeping operation had entered a new phase on 11 July. The Court found that during this process of evacuation, not only the United Nations but also the government of the Netherlands formally had authority (‘zeggenschap’) over Dutchbat, because this concerned the preparation of the withdrawal of Dutchbat from Bosnia-Herzegovina.

While this new situation (transition of a functioning peace keeping mission to the evacuation of Dutchbat) primarily is relevant to explain the factual involvement of the Netherlands, it also has direct legal relevance. It allows the Court to distinguish this case from the facts in the Behrami case, that the Court considers as an example of the ‘normal situation’ in which troops that have been put at this disposal of the UN function. The State of the Netherlands had strongly relied on Behrami, and as indicated earlier this even caused plaintiffs to opt for Bosnian law. This fear of Behrami proved unjustified. The distinguishing factor was that after July 11 the mission to protect Srebenica had failed. As Srebrenica had fallen in the hands of the Bosnian-Serbs, Dutchbat, or UNPROFOR would no longer continue the mission (par. 5.11). The decision to evacuate Dutchbat and the refugees was taken in mutual consultation between Janvier (on behalf of the UN) and high representatives of the Netherlands.

It is noted that it is by no means to be accepted as a given that in the ‘normal’, Behrami type of situation the acts would exclusively be attributed to the UN – even in such cases a strong argument can be made that there can be double attribution (see eg A. Sari, Jurisdiction and International Responsibility in Peace Support Operations: The Behrami and Saramati Cases, Human Rights Law Review 2008, 151). But the Court does not enter that path since it considered the facts to be different.

In this respect the judgment bears some similarity to the decision of the Belgium Court of First Instance of Brussels of 8 December 2010, where the Court found that the failure by the UN peacekeeping contingent to prevent the killing of Tutsis’ in the 1994 Rwanda genocide could be attributed to Belgium in a situation where the Belgian government had decided to withdraw itself from the peacekeeping operation (Mukeshimana-Ngulinzira and others v Belgium and others, First instance judgment, R.G. n° 04/4807/A et 07/15547/A; ILDC 1604 (BE 2010), 8 December 2010). Counsel for the plaintiffs has relied on this case, but the Court does not refer to it.

It is to be added that while this legal context gave the Netherland legal authority to act, and allowed the Court to find that the Netherlands had effective control, it does not follow that in the absence of legal authority a troop contributing state cannot exercise effective control. This was accepted by the District Court (which on this point was relied upon by the ILC), that held that

‘Acts of Dutchbat might have been attributable to the state had Dutchbat been 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. The same would have been true if Dutchbat to a greater or lesser extent had backed out of the structure of UN command, with the agreement of those in charge in the Netherlands, and considered or shown itself as exclusively under the command of the competent authorities of the Netherlands’

Thus: an (effective) instruction to Dutchbat will in all  cases lead to a finding of effective control and thus to attribution to the Netherlands – with or without legal authority. What is specific for the case at hand is that there was no instruction, but an involvement combined with legal authority that enabled the Netherlands to act – it was that ability that was the basis of attribution.

3. FINDINGS ON WRONGFULNESS

Based on the finding that the Netherlands had effective control and that the disputed conduct had to be attributed to the Netherlands, the Court then finds that the acts were wrongful. It rejected the argument of the Netherlands that the acts of Dutchbat in Bosnia-Herzegovina should exclusively be assessed on the basis of international (and not national law). Apart from questions of immunity, the acts of peace keeping forces remain subject to the national legal order, and in principle thus can lead to liability in a tort action under Bosnian law (that was the applicable law on the basis of Dutch rules of private international law). While the Court bases attribution on international law, it thus based its finding on wrongfulness primarily on national law. This can be explained in part by the fact that the Court found that Bosnian law did not contain a relevant rule on attribution. But as indicated above, for the Court this was only a subsidiary argument; its prime argument being that attribution between two subjects of international law should be decided on the basis of international law. While that may be true, the question should be raised whether the legal status of individuals in international law should not imply that also the question of wrongfulness, and responsibility, should be considered under international law, which would allow the Court to maintain a connection between primary and secondary rules. For the outcome that probably would not have a difference, but it would from the perspective of international law have been a more consistent approach.

The possibility of basing wrongfulness on international law is recognized by the Court itself. Even though the Court opts for determination of wrongfulness under Bosnian law, it adds that the acts were also wrongful based on a breach of the principles contained in articles 2 and 3 ECHR and 6 and 7 ICCPR (right to life and right to freedom from inhumane treatment), arguing that these principles have to be considered as part of customary international law that bind the state. The Court adds that it assumes that the State, which had argued that the treaties were not applicable to acts of Dutchbat in Bosnia, did not aim to argue that it should not comply with the principles contained in articles 2 and 3 ECHR and 6 and 7 ICCPR during peacekeeping missions such as that of Dutchbat.

The Court did not deal with the argument advanced by counsel that the human rights treaties were applicable based on effective control on the compound – an argument that was supported by the judgment of the ECtHR two days later in Al-Skeini and Others v. the United Kingdom (application no. 55721/07).

Perhaps to protect itself against claims that the application of human rights standards to extraterritorial military actions is controversial, the Court adds that on the basis of art. 3 of the Constitution of Bosnia-Herzegovina, the rights in question have direct effect and that since in any case the ICCPR was in force for Bosnia in 1995, articles 6 en 7 ICCPR are part of the Bosnian law and have supremacy over any conflicting rules of Bosnian law.

The Court then determines on the basis of the facts that the commanders in question (Karremans and Franken) should have known that the Bosnian men who were to be ‘evacuated’ from the compound faced a real risk of being killed or at least being subjected to inhumane treatment. It follows that Dutchbat on the basis of Bosnian law and the directly applicable rights of the ICCPR was not allowed to send Mustafic and Nuhavonic from the compound and that the state thus acted wrongfully. It adds that Mustafic and Nuhanovic would have been alive if they would not have been removed, and that there thus was a causal connection between the removal and their death (without considering any possible intervening factors, that broke the causal chain, as the ICJ did in the Genocide Case in looking at the relationship between Serbia’s failure to prevent and the eventual genocide that occurred).

The Court concludes that, on the basis of article 155 of the Law of Contracts of Bosnia and Herzegovina, the state is liable for the immaterial damage of the plaintiffs in this case.

The Court did not yet make a final ruling on reparation due to the fact that it still wishes to examine a possible breach of the right to a fair trial in connection to a replacement of the judges in the District Court – a matter that is left aside here.

4. CONCLUSION

In many respects, this was a ‘hard case’. The voluntary eviction of own employees, in the face of clear evidence of threatened death and genocide, while there were ample possibilities to offer protection, constitutes a rather unique and extreme set of facts. It seems inevitable that these facts influenced the interpretation and construction of the relevant legal principles.

Whether the decision will survive a possible cassation with the Supreme Court remains to be seen, but it should be observed that the Supreme Court cannot revisit the facts, and the facts in this case strongly support the claim of the plaintiffs. Likewise, the Supreme Court cannot revisit questions of foreign (Bosnian) law. If the Supreme Court would annul the decision, it would have to be based on a different interpretation of the principles of international law – something that has never been done in the history of the Supreme Court. While there is a first time for everything, this is neither to expected nor to be desired. The decision of the Court of Appeal in large part is solidly based in positive international law. In its innovative parts, it is a major contribution to the clarification and development of the principles pertaining to the (dual) attribution of acts of peacekeeping troops.

The departure from the dominant black and white approach that was adopted in Behrami, is to be welcomed. As such it is also supported by the ECtHR itself, that on 7 July 2011 in Al-Jedda v. the UK (application no. 27021/08) held that  ‘The Court does not consider that, as a result of the authorisation contained in Resolution 1511, the acts of soldiers within the Multi-National Force became attributable to the United Nations or – more importantly, for the purposes of this case – ceased to be attributable to the troop-contributing nations.’ (para. 80)

Given these unique facts, and given the fact that attribution was based on the active involvement of the Netherlands in the evacuation process (a mere possibility to intervene would not have been enough) and that the case rests largely on the fact that the mission de facto had been completed, one should be very cautious in using the judgment as a possible basis for other claims in regard to liability of troop contributing countries.

However, the main message of the Court of Appeal (effective control, and thus attribution, depends on a test of ‘appreciation in concreto of the ability to prevent’) may apply to broader situations. Depending on the facts, attribution can be approached as a sliding scale, that includes on the far sides exclusive attribution to the UN and exclusive attribution to the troop contributing state. In the middle of this continuum, in factual situations where both the state and the UN have normative control and are factually are involved, dual attribution is the proper approach.

Expert Seminar on Shared Responsibility in International Refugee Law – In Search for a Legal Basis of Burden-Sharing

 Posted by: Swerissen Isabelle

Do States – and other subjects of international law – have a collective obligation to protect refugees? And if this is the case, does a breach of this obligation lead to shared international responsibility? At a time when the burdens and responsibilities that flow from massive displacement of people have been distributed so unevenly among the world’s regions and countries (see UNHCR Global Trends 2010), these two questions have attracted growing interest and were discussed at the Expert Seminar on Shared Responsibility in International Refugee Law that the SHARES Project organized on 30 May 2011 (see Programme). While a more detailed account will be provided in the forthcoming seminar report, we will use this blog post to give an overview of the discussions held at the Expert Seminar. Despite the variety of views presented on the above-mentioned questions, we intend to outline some of the general points and arguments made on the theory and practice of shared responsibility in international refugee protection at the international and regional levels.

INTERNATIONAL REFUGEE PROTECTION AND BURDEN-SHARING

At first sight, international refugee protection seems to be an area of international law that involves questions of sharing and distributing obligations and/or responsibilities. After all, “international protection” – while lacking a generally agreed definition – is understood as the protection provided by the international community because the refugee, due to a well-founded fear of persecution, is unwilling or unable to avail himself of the protection of his country of nationality (compare Article 1 of the 1951 Convention Relating to the Status of Refugees). UNHCR itself often refers to the “collective responsibility” to share the burden of the global refugee problem on the part of the State parties of the 1951 Convention and its 1967 Protocol (on burden-sharing see J.P. Fonteyne, ‘Burden-Sharing: An Analysis of the Nature and Function of International Solidarity in Cases of Mass Influx of Refugees’, (1978-1980) 8 Australian Year Book of International Law 162-188).

But is this burden-sharing really based on a legal obligation, or does it merely represent a moral principle? Some participants at the SHARES Expert Seminar observed that, in light of State practice, a collective obligation or ensuing shared responsibility, as manifested in equitable burden-sharing, does not exist. Whereas refugee protection was indeed considered a responsibility incumbent upon the international community during the Second War, UNHCR was not established with an operational mandate, which distinguished it from its different predecessor organizations, in particular the International Refugee Organization (IRO). Instead the responsibility for refugee protection was assigned to individual States or, more specifically, to the State on whose territory a refugee sought refuge. This responsibility to protect refugees on the part of individual States has been most prominently enshrined in Article 33 of the 1951 Convention, stipulating the prohibition to return a refugee to a place of persecution (non-refoulement).

The obligation of non-refoulement subsequently developed into the centerpiece of the international refugee protection regime, and has arguably even achieved the status of ius cogens. It is therefore all the more paradoxical that the obligation of non-refoulement counteracts an equitable sharing of the burden of refugee protection, for it imposes a larger burden on those States that are geographically closer to re-producing countries. These receiving countries are often less capable of accommodating large numbers of refugees for various reasons – one only has to think of the conflict-loaded and difficult socio-economic situation of most countries in sub-Saharan Africa. Consequently, it is not surprising that UNHCR’s mandate, enshrined in General Assembly 428 (V) (1950), has become quite operational over the years (as a Sudanese refugee in Egypt observed: “We live in a country of UNHCR”), and has been stripped of its temporary nature. The principle of non-refoulement does thus not present a valid legal basis for the collective obligation to protect refugees. It could even be argued that the principle has been (ab)used to prevent a more equitable of refugees among developed and developing nations due to its inherent link to with State territory, which prevents overburdened countries of asylum to send refugees to other potential host countries with better reception capacities.

Despite the absence of a legal basis for the collective obligation to protect refugees, shared responsibility may still be the result of a breach of obligations that concern the protection of individuals in other areas of law such as international human rights law and international humanitarian law. Although notions such as the “roots causes” of displacement or “preventive protection” have been tainted by the negative experiences during the war in the Former Yugoslavia, the current situation of mass flights from Syria to Turkey due to the repressive measures of the regime of Bashir Al-Assad proves the continuing relevance of the debate on the reasons of forced displacement (see SHARES news item on the situation in Syria), which are all too often found in massive human rights and humanitarian law violations. In this context, it could certainly be contended that the international community has a “responsibility to protect” civilian populations, to use the well-known catchphrase (see the forthcoming book by Julia Hoffmann and André Nollkaemper on the Responsibility to Protect: From Principle to Practice, based on the similarly titled conference). Strictly speaking, this “responsibility to protect” is not a responsibility but rather an umbrella concept for different collective obligations that are incumbent on the international community as a whole (erga omnes status). Failure to observe these collective obligations to protect may inter alia lead to forced internal and external displacement. Following this line of argument, the international community would incur shared responsibility for not having complied with its primary obligations.

Against this background, the participants discussed whether UNHCR could be seen as burden-sharing mechanism – that is a mechanism to implement the shared responsibility of the international community – as proposed, for instance, by former High Commissioner Lubbers. However, as mentioned above, UNHCR was de jure not designed to be a burden-sharing mechanism, although the continuously pressing needs of the global refugee problem have de facto led to the operationalization of the organization.  As a result of this discrepancy between normative and factual reality, the efforts of UNHCR are generally considered to be insufficient in reducing the disparities in international refugee burden-sharing. The activities of UNHCR are not guided by principles of equity but rather humanitarian necessity, as conditioned by budgetary and political constraints since UNHCR is largely dependent on the voluntary contributions of UN member States. UNHCR has made an attempt to remedy this situation by means of special agreements in the framework of the so-called “Convention Plus” Initiative, which was intended to lead to a more equitable burden-sharing. Yet, the Convention Plus Initiative ultimately failed to achieve its main objective for different reasons, mainly because it addressed burden-sharing eclectically rather than globally (see generally Marjoleine Zieck, ‘Doomed to from the Outset? UNHCR’s Convention Plus Initiative Revisited’, (2009) 21 International Journal of Refugee Law 387-420). Forced displacement is a global problem, which calls for global solutions and for a global approach in terms of burden-sharing, for instance, in the form of an amendment to the UNHCR Statute and/or the 1951 Convention.

REGIONAL REFUGEE PROTECTION AND BUCK-PASSING

The finding that burden-sharing at the global level does not really work begs the question whether burden-sharing functions better at the regional level, in particular the context of the European Union (EU). However, the practice of the European Union and its member States seems to somewhat mirror the problems that the refugee protection regime faces at the global level. Instead of sharing the burden of refugee protection by means of the more sophisticated legal framework of the EU treaties, the European Union and its member States have rather engaged in a practice of buck passing in their internal and external relations.

In its external or international relations, a policy of buck-passing may be observed in situations of where refugees are prevented from reaching the territory of the EU in the first place. This usually occurs as a result of a deliberate acts to circumvent refugee arrivals, referred to as ‘externalization of European border policies’, by taking action outside the territory, for example, by interdicting refugees in international waters or by exercising control over refugees in the territory of another (non-European) State through the operation of some kind of extraterritorial processing center (on extraterritorial immigration policies see M. den Heijer, ‘Europe Beyond its Borders: Refugee and Human Rights Protection in Extraterritorial Immigration Control’ in B. Ryan and V. Mitsilegas (eds.), Extraterritorial Immigration Control: Legal Challenges (Martinus Nijhoff Publishers: Leiden/Boston, 2010) 169-198). Frontex, the EU border agency, plays an important role in the context of buck-passing outside the EU. Equipped with high-tech detection equipment, helicopters and boats, it strengthens border security by ensuring the coordination of member States’ actions in the implementation of EU measures relating to the management of the external borders.

During the Expert Seminar it was argued that any time a State seizes the power or authority over a refugee to determine his or her destination, whether this occurs on the high seas or in the territory of another State, Article 33 of the 1951 Convention is implicated.  Moreover, the conduct of the EU and its member States is not only regulated by the 1951 Convention, and the prohibition of non-refoulement at the frontier, but also by other international obligations, which can in particular be found in the law of the sea (see the recent discussions on EJIL: Talk! here and here). A breach of these obligations will most likely result in a form of shared responsibility, depending on the degree of cooperation between the multiple actors involved. Various scenarios were discussed by participants in which both the State exercising extra-territorial powers, with assistance of Frontex, and the territorial State may have obligations in regard to refugees and may share responsibility in case of breach of such obligations. Although the competences of Frontex itself are limited in the execution of its border control operations, it was also suggested the EU – to which Frontex’s conduct is attributable – may be responsible to the extent that it shares competences with its EU Member States (see the recently adopted ILC Draft Articles on the Responsibility of International Organizations).

While the practice of extra-territorial refugee policies illustrates the shifting of obligations in relation to refugees who have yet to reached EU territory from EU member States to third States, the participants of the Expert Seminar also extensively discussed buck-passing in cases where refugees have been successful in reaching territory of the EU. This has been the subject of much debate, in large part due to the controversy surrounding the Dublin II regulation. This regulation, which replaced the 1990 Dublin Convention, is based on the presumption that protection standards in all EU member States are the same or at least equivalent. Mutual trust in comparable standards of protection, rather than actual harmonization of refugee rights, has thus formed the basis for assigning protective responsibility over refugees to member States. By virtue of the Dublin regulation, only one State, usually the country where the refugee first entered the Union, is responsible for the examination of an asylum application. As such, the system imposes an increased burden on countries at Europe’s external borders, who, due to the happenstance of geography, find refugees crossing their borders. Instead of distributing protection obligations more fairly, the regulation has resulted in an institutionalized form of buck-passing within the EU.

Greece’s broken asylum system shows this all too clearly. With a vast Mediterranean coastline and a porous land border with Turkey, Greece has become the gateway to Europe. The difficulties in coping with the increased influx of refugees have been recognized by the European Court of Human Rights in M.S.S. v. Belgium and Greece (also see SHARES blogpost), a case regularly referred to during the Expert Seminar. The case concerned the return of an Afghan asylum seeker to Greece by the Belgian authorities in application of the Dublin regulation. By knowingly exposing him to structural shortcomings in the asylum procedure and systematic problems in the detention and reception of asylum seekers in Greece – for which Greece itself was found responsible – the ECtHR ruled that Belgium had violated Article 3 of the European Convention on Human Rights. Rather than as a collectivity, both countries have individually breached their own human rights obligations, not making this case any less relevant from the perspective of shared responsibility, which can also arise from acts that cannot be attributed to the same actor. This judgment has dealt a major blow to the Dublin system, and underlined the continued international obligation of EU Member States to actively verify how the receiving States apply internal EU asylum legislation and procedures in practice, effectively debunking the myth of equality in protection.

Buck-passing or the assignment of ‘protection elsewhere’, whether extraterritorially or within the territory of the EU, is now surprisingly well entrenched. One striking feature, however, is that the idea of requiring a refugee to seek protection elsewhere is not explicitly anchored in the text of the 1951 Convention. In light of the absence of such authority, a key but less frequently asked question is whether protection elsewhere policies and practices are permitted under international law. A second related question is whether, assuming that some kind of transfer to another country of asylum is permitted, any legal impediments operate on States in choosing to send a refugee to a particular country. With regard to the latter question, States have mainly focused on the prohibition of refoulement as enshrined in Article 33 of the 1951 Convention as the only protection obligation that must be respected. States reason that, as long as they do not breach this obligation, they are not necessarily required to provide protection to refugees who reach their territory, but may send them to other States.

However, a perusal of the 1951 Convention shows that, Article 33 apart, there is a range of requirements imposed upon contracting States with respect to refugees, some of which can likewise be characterized as protection obligations. These obligations apply as soon as the refugee is within the territory of a State. This is so even if State agencies have not yet formally determined that the applicant satisfies the refugee definition, since recognition of refugee status is declaratory. In the words of the UNHCR Handbook: “recognition of refugee status does not make a person a refugee, but only declares him or her to be one”. Having said that, requiring a refugee to go somewhere else may therefore not be a simple ‘dumping’ of refugees abroad on the amorphous ground that they will be admitted and protected from expulsion, and expulsion only, in that third country. European policies have been controversial precisely for that reason. Rather, the sending State is under an obligation to ensure that refugees will be guaranteed certain fundamental rights or, as one of the speakers eloquently observed, “the onus is on the sending State”. Failure to observe this obligation and any ensuing violations of those rights in the State to which they are transferred, committed individually and separately by the sending and receiving States, can result in shared responsibility as illustrated by the M.S.S. case. In order to be lawful and meaningful, a burden-sharing scheme must therefore be implemented in a manner that takes into account all obligations incumbent on States under the 1951 Convention.

SOME PRELIMINARY CONCLUSIONS: BUCK-PASSING INSTEAD OF BURDEN-SHARING?

The discussions at the SHARES Expert Seminar revealed wide divergences between the theory and practice of shared international refugee protection at the global and regional levels. At the global level, only few geographically disadvantaged countries share the largest part of the burden of international refugee protection with the support of UNHCR, which could be characterized as a – albeit ill-equipped – de facto burden-sharing mechanism that implements the shared responsibility of the international community. Geography also appears to play a crucial role at the level of the European where EU Member States have used the framework of the Dublin II regulation to pass the buck to European border States such as Greece, or prevented refugees from entering EU territory in the first place.

It is widely accepted that States cannot “contract out” of their protection obligations or evade their international responsibility by passing it on to other States or by establishing an international organization, be it UNHCR or the European Union (including its border control agency Frontex), without a certain degree of effectiveness. However, an effective implementation of the obligations in the area of international refugee protection presupposes that the actors involved, be it States or international organizations, are aware of and recognize the nature and scope of their respective international obligations. While the discussions at the SHARES Expert Seminar have shown that States may have various collective obligations with regard to the prevention of forced displacement and individual obligations with regard to the protection of refugees, States often simply deny the existence of these obligations and ensuing responsibilities for political reasons. This observation is quite evident in the negatively connoted term “burden-sharing” that was criticized by several participants of the Expert Seminar as inappropriate in light of the compelling human tragedy that characterizes international refugee protection, which may actually not only come with ‘burdens’ but also benefits and opportunities for host countries.

It seems that the international community needs to be faced with yet another cataclysm – such as the two World Wars in the first half of the 20th century – in order to recognize the global dimension of international refugee protection as a shared responsibility. Interestingly, such a historical momentum may be created in another area of international law that could give rise to collective obligations and/or shared responsibility, namely international environmental law (see our next Expert Seminar on Shared Responsibility in International Environmental Law). For more recently, forced displacement induced by climate change has received increasing attention by scholars and practitioners alike and may lead to new and unforeseen challenges in the area of international refugee protection that call for a further clarification of collective rights and obligations.