Melbourne Journal of International Law
ARE CONTROL TESTS FIT FOR THE FUTURE?
THE SLIPPAGE PROBLEM IN ATTRIBUTION DOCTRINES
The Slippage Problem in Attribution Doctrines
KRISTEN E BOON[*]
When do subjects of international law bear responsibility for the acts of others? It is often a question of control. Control is an essential element of the doctrine of attribution, defining the legal relationship between states, international organisations (‘IOs’), and individuals. Control is also a factor in determining what is properly within a state or IO’s purview, legally demarcating the public and private spheres. Yet while control tests are intended to operate according to objective standards, they have important normative dimensions because they can determine the outer bounds of state action, define the allocation of responsibility between states and IOs and have feedback effects for state sovereignty. This article argues that control tests under prevailing doctrines of attribution present a ‘slippage’ problem. Slippage is occurring because the essence of the state, as a primary subject of international law, is changing. In response, various techniques have emerged to adapt control thresholds, locating responsibility within omissions, the duty to prevent, or under the due diligence rule and articulating principles of shared responsibility. This development demonstrates great movement within attribution doctrines and the potential scope of state and IO responsibility. One consequence of this movement is that it may foretell the eclipse of general, secondary rules of attribution. Another consequence of this dynamism is the perception that the effective control test as an objective, portable, general concept of law will become increasingly suspect.
When do subjects of international law bear responsibility for the acts of others? It is often a question of control. Control is an essential element of the doctrine of attribution, defining the legal relationship between states, international organisations (‘IOs’) and individuals. Control is also a factor in determining what is properly within a state or IO’s purview, legally demarcating the public and private spheres. Yet while control tests are intended to operate according to objective standards, they have important normative implications. In particular, control tests can determine the outer bounds of state action and define the allocation of responsibility between states and IOs. They consequently have feedback effects for state sovereignty.
Because the regime of international responsibility remains seriously underdeveloped — in particular, because responsibility may be attributed to states, individuals under international criminal law and, somewhat controversially, to IOs, but not to entities like multinational corporations,
non-governmental organisations (‘NGOs’) and individuals outside of the criminal context — purportedly objective control tests have been harnessed in a bigger contest about the appropriate reach of international law and the definition of its primary subject: the state. The orthodox view remains that the stringent ‘effective control’ test is appropriate for attributing private conduct to a state and for allocating responsibility between states and IOs, unless primary norms or
lex specialis dictate otherwise. Nonetheless, efforts to indirectly expand or adapt control tests to multi-level governance situations, joint management arrangements between states and IOs and corporations or partnerships that perform public functions are symptoms of the lag between the changing nature of statehood and the limited category of subjects of international law.
Alternative techniques for redressing the limited reach of state responsibility have surfaced in response, such as lowering thresholds of control, attributing responsibility for omissions, establishing/developing a duty to prevent certain acts subject to a due diligence obligation and where circumstances and doctrine warrant, recognising shared responsibility between actors. All of these techniques are being used to navigate the new forms of regulatory power, and bridge the so-called ‘accountability gap’ in international law.
In combination, these techniques are indicative of great movement within attribution doctrines and in the potential scope of state and IO responsibility. This movement may foretell the eclipse of general, secondary rules of attribution and cast further doubt on the notion that the effective control test is an objective, portable, general concept of law. Most fundamentally, the shifting landscapes around existing doctrines of attribution provide an opportunity to revisit fundamental normative choices in international law.
This article argues that control tests under prevailing doctrines of attribution compound the problem of ‘slippage’. By slippage, I mean the decline of government control of functions traditionally associated with the state, resulting in the actual or perceived failure to meet standards under international law. Control tests both fail to reign in an existing tendency and are causing the problem to become manifest. Slippage is occurring because the essence of the state, as a primary subject of international law, is changing. state sovereignty has evolved with successive waves of globalisation, liberalisation and privatisation, resulting in a shift away from the state as the primary source of regulation. A restrictive or outdated view of the state, or reliance on secondary rules that do not grasp the complexities about how responsibilities are allocated between states and IOs, will only hasten slippage.
Part II of this article explores why control tests are so common in international law. Part III evaluates control tests under the law of state responsibility and explores why high control thresholds are used in some contexts of international law, whereas lower thresholds are being advocated in other contexts. It concludes that control, whilst appearing to have objective standards is actually very dependent on the facts. This has meant that despite the ICJ’s affirmation of the effective control test in the Application of the Convention on the Prevention and Punishment of the Crime of Genocide case (‘Bosnian Genocide’), the test is not as portable as is often assumed. Part IV examines the International Law Commission’s (‘ILC’) proposed effective control test for the responsibility of IOs and the metamorphosis and limits of this distinct joint management approach in light of IO practice. Part VI is devoted to duties and techniques that have surfaced to overcome the limitations of control tests, namely the duties of states and IOs to prevent and act with due diligence. I argue that the limitations of control tests have indirectly propelled these duties to prevent and due diligence requirements onto centre stage which, in turn, are affecting the scope and content of state sovereignty. I conclude with some observations about whether these alternative routes to state and IO responsibility address the problem of slippage.
The concept of control plays a significant role in at least ten different
sub-fields of international law:
• questions regarding whether an act or omission of an individual, organ or agent is rightfully attributed to a state, under art 8 of the Draft Articles on Responsibility of States for Internationally Wrongful Acts (‘Articles on State Responsibility’);
• questions regarding whether an act or omission of an individual, organ, agent or state is rightfully attributed to an IO under art 7 of the Draft Articles on the Responsibility of International Organizations (‘Articles on the Responsibility of IOs’);
• questions as to whether a military or civilian superior is in effective command or control, or indirect control, of a subordinate under international criminal law;
• questions as to whether a territory is under the effective control of a hostile state, triggering the application of the Hague Regulations and Geneva Convention IV;
• questions as to whether a state exercises control over a territory, which determines whether an armed conflict is considered international or
• questions as to whether a new state or government should be recognised, where effective control is an element in international recognition;
• questions as to whether a state has effective control over a space or territory, and hence has a general responsibility for upholding human rights conventions extraterritorially;
• questions as to whether a state incurs responsibility for acts on its territory, even if that territory is no longer under its control;
• questions as to whether a state is in effective control of a vessel flying its flag; and
• questions as to whether a state can revoke or withhold a transit permit when the air transit enterprise is not under the effective control of a contracting state.
As this list makes clear, control tests have become a default mechanism in multiple issue areas of international law. Their popularity can be explained by their flexibility. Designed to permit case-by-case assessments, control tests appear to offer objective standards to decision-makers across sub-fields of international law. Moreover, they permit differentiated burdens on subjects of international law in that they may confer greater obligations on entities with a strong nexus to the act or omission in question. Because there are different ideological views about the inherent functions of the state, the control test is also attractive because it is a way of sidestepping controversy about public and private functions. As Nigel White explains, control tests focus on the nature of the relationship between a state and private entity, rather than the function being performed, which might be considered public or private, depending on the particular state and context.
Despite their prevalence, however, there are limitations to control tests. One drawback is that control tests with a high threshold, such as effective control, are premised on the concept of limited state responsibility and, as a consequence, they do not always adapt well to modern manifestations of states that outsource functions of a traditionally public nature (eg law enforcement, immigration, prisons) or states that are themselves embedded in other
supra-structures, such as the European Union. A second limitation is that the effective control test misses the nuances of many relevant practices of IO. As such, transposition of effective control to the IO responsibility context has not been met with great enthusiasm. James Crawford described the test as one of ‘essential ambiguity’ which ILC members hoped would be fleshed out in practice. Finally, control tests may not in fact be objective enough to provide adequate notice to subjects of international law or be sufficiently adapted to the modern state. Because the appropriate standard of control is inherently connected to primary rules, the adoption of effective control as a general standard says more about presumptions about the state than it does about the currency of the standard in law.
Broadly speaking, there are two distinct categories of control tests: control over territory and control over persons. The first type of control test focuses on spatial control, where a state or IO’s territorial presence may trigger positive obligations to act, such as to prevent certain harms from occurring, to ensure respect for human rights or to protect populations in territories under a subject’s control. In contrast, the second kind of control test focuses on the attribution of acts where one entity exercises power over another. The focus of this article is on the latter test — attribution with regard to acts of persons, organs, agents and other entities. In other words, it addresses how control mediates power relationships, as opposed to how control affects obligations flowing from control over territory. The next Part will argue that despite the ICJ’s position that the effective control test is now de rigueur, there has been a longstanding contest over the appropriateness of the threshold that gives every appearance of continuing.
The core jurisprudence on the control threshold in the doctrine of attribution comes from three ICJ judgments involving the attribution of acts of non-state entities to states: Military and Paramilitary Activities in and against Nicaragua (‘Nicaragua’), Armed Activities on the Territory of the Congo (‘Armed Activities’) and Bosnian Genocide. In these cases, discussed in more detail below, the ICJ applied an effective control test and ultimately determined in each case that the state in question did not specifically and factually control the acts of the relevant non-state actors, despite sometimes extensive state support.
If one were to end the inquiry here, it would be easy to assume that the standard of effective control is settled under the core doctrine of attribution. The ICJ’s interpretation of control is, however, distinguishable in several important ways from the ILC’s approach to control under the Articles on State Responsibility. Moreover, the ILC’s interpretation of control under the Articles on State Responsibility and the Articles on the Responsibility of IOs is itself increasingly distinguishable from control-based attribution tests in the terrorism, trade, investor–state and international criminal law contexts. While the ICJ’s rulings are certainly most weighty they are not formally binding on parties outside the dispute in question, which affects their generality. This movement indicates the concept of control is a contentious one, which, despite its prevalence, has been harnessed into a larger debate about the potential reach of international law.
A quick tour d’horizon reveals that the calculus of control within doctrines of attribution can vary greatly. The restrictive effective control approach, which classically requires evidence of factual control over specific conduct, is favoured in the ICJ jurisprudence on international humanitarian law (‘IHL’) and in claims against IOs where there is joint management between an IO and state(s) or between two IOs. By contrast, in the contexts of terrorism, the World Trade Organization (‘WTO’), investor–state arbitration and in a determination of whether an international conflict exists, there have been movements towards lower thresholds because the primary rules in these contexts suggests the requisite level of control should be lesser. To that end, ‘meaningful’ or ‘overall control’ tests are often advocated on the basis that they are truer to the nature of the problem. The next Parts detail how these schisms have been apparent in fundamental texts and decisions in the area.
Although attribution-based control tests have shared criteria, standards and principles, there are important differences in application that quickly become apparent when one scratches the surface. The historic case on control that defined the current and restrictive paradigm of state responsibility is the ICJ’s 1986 decision in Nicaragua. In assessing whether violations of IHL were committed during the civil war in Nicaragua, the Court considered three different categories: acts of members of the United States Government, certain acts of the Latino Assets (Latin American operatives known as the UCLAs) and acts of the Contras. The Court determined that while acts of the first two categories were attributable to the US, acts of the third were not. On this, the key background finding, as Crawford explains, was that although the US ‘did not “create” the contra force, it was responsible for financing it and for providing logistical support to the movement’. Moreover, it had trained ‘the contras and provided them with intelligence as to Sandinista troop movements’, and some ‘contra operations had been planned in conjunction with US military advisers and ... the US had identified suitable targets for contra attacks’. Thus, despite extensive US involvement with, and influence over, the Contras, the ICJ held that the Contras were not essentially organs of the US government and that, while the US supported the Contras, it did not control them.
In making this determination, the ICJ identified two relevant levels of control: strict control and effective control. Strict control is based on complete dependence, which involves an assessment of whether or not the acts of an entity are essentially those of a de facto state organ. In essence the de facto organ must be shown to have no real autonomy or independence and to have acted as a mere instrument of the outside power. Effective control, in contrast, is based on partial dependence, where specific acts of private individuals or groups are controlled by the state. In order to meet the effective control test in this context, the applicant would have had to demonstrate the existence of (i) a de facto link by virtue of factors such as financing, organising, training, selecting targets and planning, and (ii) control such that it is clear that the acts had been ordered or imposed on the relevant individuals and entities by the state. The Court consequently adopted a high control threshold, reflecting a restrictive approach to the state, and thus state responsibility.
The most famous schism between international courts over control arose when the International Criminal Tribunal for the Former Yugoslavia (‘ICTY’) challenged the Nicaragua standard of effective control in Prosecutor v Tadić (‘Tadić’). The ICTY Appeals Chamber proposed a lower threshold, an overall control test, in recognition of the influence of organisation and hierarchy in groups. Under the overall control test, specific instructions are not necessary whereas, under the effective control test, they would be. The ICTY’s less stringent standard was based on the argument that ‘a member of the group does not act on his own but conforms to the standards prevailing in the group and is subject to the authority of the head of the group’.
The ICJ did not support the overall control test. In two subsequent decisions, the Armed Activities case of 2005 and the Bosnian Genocide case of 2007, the ICJ reaffirmed the effective control test. In the Armed Activities case, the ICJ concluded that there was no probative evidence that Uganda controlled, or could control, the manner in which the rebel group, Movement for the Liberation of the
Congo, was provided assistance. Nonetheless, the ICJ noted that, through its use of force and intervention, Uganda had violated the Declaration on Friendly Relations, which constitutes customary international law. Interestingly, it also noted that even in the absence of attribution, Uganda, the occupying power, had control over the territory and this created an obligation to prevent looting, which ‘extends ... to cover private persons in this district and not only members of Ugandan military forces’. As such, the Court partially addressed the slippage problem by upholding a duty to prevent under art 43 of the Hague Regulations, which extends to private persons.
In the Bosnian Genocide case, the court went further than it had in Armed Activities, noting that the competing and broader overall control test would stretch ‘almost to breaking point, the connection which must exist between the conduct of a State’s organs and its international responsibility’. The Court reiterated that a state is ‘responsible only for its own conduct, that is to say the conduct of persons acting, on whatever basis, on its behalf’. Based on this test, it then found that the massacres at Srebrenica were not committed by organs of the Federal Republic of Yugoslavia (‘FRY’), on the directions or instructions of organs of the FRY or in operations where the respondent exercised effective control. It also clarified ‘that the State’s instructions were given, in respect of each operation in which the alleged violations occurred, not generally in respect of the overall actions taken by the persons or groups of persons having committed the violations’. Finally, it rejected the argument that a lex specialis on attribution applied to the crime of genocide. As will be discussed below, however, the ICJ did determine that the FRY was under a duty to prevent acts of genocide, as a technique to address slippage.
As has been extensively analysed elsewhere, the apparent difference between the control standards advocated by the ICTY on the one hand, and the ICJ on the other, can be explained with reference to primary rules. The Nicaragua decision addressed state responsibility for violations of IHL, whereas Tadić involved a different question: the existence of international versus
non-international armed conflict. There was no reason why the same test needed to apply to both situations. Nonetheless, the ICJ’s adherence to the effective control standard in the Bosnian Genocide case indicates its belief that there is a portable, universal standard of effective control that applies to questions of attribution in all contexts, unless primary norms or lex specialis dictate otherwise. Moreover, it reaffirms a very high threshold for the attribution of responsibility to the state, a threshold that may not correspond, as an empirical matter, to changes in public authority in many states.
Slippage refers to the decline of government control or oversight, whether through regulation, ownership or delegation of functions traditionally associated with the state. The result of slippage may be a change in enforcement of particular standards or laws, whether declining or increasing, or it might be the actual or perceived failure to meet standards of law. I make no normative statement about the optimal level of regulation. Rather, my point of departure is that an important debate about control-based attribution tests is whether one high standard, premised on a limited conception of the state, is the appropriate default.
Regulatory slippage may occur for numerous reasons. It might be a product of declining government resources or a rational choice on the part of a local government to prioritise certain activities. It might be the result of the creation of new forms of partnerships, such as public–private partnerships. It may be that public and private actors are nestled together in multi-level governance situations and engage in ongoing negotiation and deliberation. Any of these situations may prompt a move along the regulatory continuum between, at one end, full government ownership and, on the other end, the free market. Between these two poles are various levels of delegation of powers and self-regulation. The next Part will examine how the ILC has approached the calculus of control within attribution doctrines, given the changing phenomenon of the modern state.
Using the ILC’s Articles on State Responsibility as a baseline, most of which are considered as constituting customary international law, this Part analyses the way control, and specifically effective control, is defined and applied by the ILC on one hand and by courts and tribunals in a variety of contexts on the other. Although the ILC’s goal was to supply generally applicable secondary rules, the failure to provide an overarching standard of control either encouraged a proliferation of different tests in practice, failed to reign in an existing tendency or it may have been right on the money: control is necessarily dependent on primary norms. The ILC recognised that sub-regimes might develop, and provided states the opportunity to contract around the secondary rules under art 55 on lex specialis. Nonetheless, an analysis of the ILC’s approach to control as an element of attribution has important differences from the ICJ jurisprudence, the jurisprudence of other tribunals and analyses of which control tests should be applied to contemporary problems in international law, including terrorism and accountability for the acts of private military contractors (‘PMCs’). The next Part will illustrate considerable movement in control-based theories of attribution, which has implications for the alleged unity of the law of responsibility.
As a preliminary matter, fundamental to the ILC’s approach to the law of state responsibility is the distinction between attribution of conduct and attribution of responsibility. Under the threshold question of attribution, an evaluation is made of whether acts can be attributed to a state directly or indirectly. Although courts will look at formal de jure relationships in making this determination, it is well-established that control is assessed through de facto arrangements as well. The inquiry into attribution is distinct from the secondary determination of whether a given act is contrary to international law or, in the case of positive obligations, is demanded by international law. Because responsibility only follows a wrongful act or omission, it is only upon a positive response to the second inquiry that attribution of conduct will lead to attribution of responsibility.
The attribution tests in the ILC’s Articles on State Responsibility revolve around three different categories of links: institutional (structural and
agency-based), functional and control-based. Institutional links are based on the status of an entity within a state or IO. Functional links are based on the exercise of governmental authority. Control links involve the conduct of private persons who are acting under governmental instructions or control. Control is a common element in all three categories, although it is most prominent in the third category of control-based inquiries.
From some perspectives, institutional links constitute the clearest form of control in the sense that an organ of a state is viewed as acting as the state itself. In other words, attribution is automatic because the link between the physical actor and the state is organic and absolute: acts by de jure state organs are attributable to the state prima facie, due to the principle of the unity of the state. As the ICJ wrote: ‘According to a well-established rule of international law, the conduct of any organ of a State must be regarded as an act of that State. This rule ... is of a customary character’. However, the organic connection means attribution goes farther than control. Even ultra vires acts of an organ will be attributable to the state, where the organ is formally independent and the state does not control it.
A second theory for attribution of conduct to the state on the basis of institutional links is a de facto test, where entities are completely dependent on a state. In the Nicaragua case, for example, the ICJ posited that if the relationship between the Contras and the US Government was one of dependence and control, it would be right to equate the Contras with an organ of the government or as acting on its behalf. Although the ICJ concluded that, despite the heavy subsidies and support provided to the Contras by the US, there was ‘no clear evidence of the United States having actually exercised such a degree of control in all fields as to justify treating the contras as acting on its behalf’, the vitality of the ‘strict control’ test remains.
The first clue of discord arose when the ILC was silent on the proliferation of different control tests in international jurisprudence in the Articles on State Responsibility, mentioning neither the strict control test nor the status of de facto organs. It is true that the ILC shied away from specifying any thresholds for control in the Articles on State Responsibility on the basis that control depends on the context but the explanation for the gap is not clear. What is apparent is that art 4 refers to the ‘internal law of the state’ and prioritises municipal legal systems in determining what constitutes the apparatus of the state, such as organs and agencies. In addition, art 4(2) of the Articles on State Responsibility states that ‘an organ includes any person or entity which has that status in accordance with the internal law of the State’. This formulation leaves open the possibility that this category is not restricted to designations under internal law, but instead depends on who exercises governmental functions on a permanent basis.
Nonetheless, states may also create organs on a de facto basis, where both domestic and international law will be relevant. Crawford, in fact, might be seen as taking a more liberal view of this category than the Articles on State Responsibility, explaining that ‘in some legal systems the status of state organ may be bestowed not only by internal law but also by internal practice, creating a category of de facto organ’, although he considers this status ‘exceptional’. A first difference is therefore apparent: unlike the ILC, the ICJ explicitly acknowledges the category of ‘de facto organ’ under a state’s strict control. This approach opens the door to an expansionist approach to responsibility in the case of institutional or legal (as opposed to factual) links.
The second category envisaged by the ILC involves functional links based on governmental authority under art 5 of the Articles on State Responsibility. Here, the focus is on bodies that are authorised to exercise governmental authority such as parastatal elements that exercise or retain certain public, governmental or regulatory functions. Although such authority might be manifested by control, it does not need to be. In fact, in its commentary on
art 5, the ILC was exceedingly clear that ‘there is no need to show that the conduct was in fact carried out under the control of the State’.
The resort to control in certain sub-fields like WTO law is still, however, common. In the 2003 case of Canada — Measures Affecting the Importation of Milk and the Exportation of Dairy Products, for example, the Appellate Body stated that the essence of government is that it enjoys the effective power to regulate, control or supervise individuals, or otherwise restrain their conduct, through the exercise of lawful authority. Provincial dairy boards, in other words, fulfil a governmental function through industry supervision and control and, in this regard, states can incur liability for their acts and omissions. This same approach was adopted in a 2012 case addressing the interpretation of ‘public body’ under the Agreement on Subsidies and Countervailing Measures. In United States — Countervailing Duty Measures on Certain Products from China, the WTO Appellate Body decided that ‘what matters is whether an entity is vested with authority to exercise governmental functions, rather than how that is achieved’. One way of demonstrating government authority under art 5, according to the Panel, is evidence that an entity exercises meaningful control.
A difference in approach is thus apparent. The Articles on State Responsibility indicate that control is irrelevant to art 5, while the WTO’s emphasis on meaningful control indicates the contrary: control matters when evaluating corporate identity. In this regard, WTO panels appear circumspect in recognising corporate separateness from a parent state where those corporate bodies exercise regulatory functions. When corporate bodies exercise public functions like regulation, therefore, their connection to a state may be assessed on the basis of control. Thus, while art 5 was meant to address the end result of a process of outsourcing for the ILC, including public corporations, semi-public entities, agencies and even private companies on occasion, the control-based approach opens the door for tribunals to address a range of issues, including cultural attitudes on public and private functions that go far beyond orthodox attribution doctrines. This stands in sharp contrast to a fundamental precept in international law, that attribution should not follow state ownership because of separate legal personality.
The concept of control is most important to the third category of the ILC’s schema: the indirect attribution of acts of irregular groups, individuals or entities that act under a state’s direction, control or instructions. This category deals with individuals or groups of individuals who are not formally part of the state but who still act under its influence (the principal inquiry of the Nicaragua decision). This is consequently a second de facto category: in addition to the de facto organ test under art 4, art 8 creates a subsidiary de facto category that addresses acts carried out by irregular groups or entities. It is a subsidiary test, as Stefan Talmon notes, in that ‘[t]he ICJ only resorts to it when it has found that the requirements of the “strict control” test for the determination of an agency relationship cannot be proved’.
In the Articles on State Responsibility, art 8 reads as follows:
The conduct of a person or group of persons shall be considered an act of a State under international law if the person or group of persons is in fact acting on the instructions of, or under the direction or control of, that State in carrying out the conduct.
Instructions refer to situations where state organs supplement their own action by recruiting or instigating private persons or groups who act as auxiliaries outside of the official state structure. Directions typically require the showing of a situation where corporations attempt to achieve a particular result.
While the heart of art 8 involves control, the ILC controversially did not specify the level of control required for attribution, indicating instead, in reference to the Tadić decision, that ‘it is a matter for appreciation in each case whether particular conduct was or was not carried out under the control of a State, to such an extent that conduct controlled should be attributed to it’ and that the ‘full factual circumstances and particular context’ need to be assessed. Despite the absence of a threshold, the principle is that private entities can act on behalf of a state as an ‘extended arm’. As such, a flexible and fact-based view of what falls within a state’s control is required.
As discussed above, the ICJ has identified the appropriate standard of control in Nicaragua, Armed Activities and the Bosnian Genocide case as effective control, which requires evidence of specific conduct over operations. Although the issues before the court were framed in the context of IHL, the ICJ’s wording suggested a general application. To prove effective control, the claimant must show direct interference, such as financial assistance, military assistance, intelligence sharing, selection, support and supervision of the leadership. Moreover, the control must be over the activities or operations giving rise to the internationally wrongful act, not over the entities as such. This level of control is to be contrasted with overall control, which is generic, flows from a general mandate and is based on a legal relationship. As Talmon writes, ‘unspecified claims of “involvement” or “direct participation” ... will not be enough to establish effective control over a particular activity or operation’.
This view of control is closely linked to the issues before the court in the Nicaragua and Bosnian Genocide cases. Indeed, as Kimberley Trapp writes, the test of ‘effectiveness’ was designed to address non-inherent acts contrary to international law such as violations of IHL:
In Nicaragua, the ICJ was dealing with two separate levels of activity: the first was the paramilitary operations of the contras ... the second level of activity involved violations of humanitarian law perpetrated in the course of those paramilitary operations.
As such, the elements developed by the ICJ were intended to apply to the determination of effective control during a military operation subject to the laws of war. This approach meant that the portability of the effective control test has been problematic from the start because it is tied to violations of IHL.
Since Nicaragua, questions about the portability of the effective control test have percolated in important sub-fields of international law because the primary rules of law and the relationship between states and private entities is changing. In the investor–state context, for example, a number of prominent and recent awards have interpreted the effective control test as requiring ‘both a general control of the State over the person or entity and a specific control of the State over the act the attribution of which is at stake’. Despite superficial similarity to the ICJ’s Nicaragua test, and liberal citing of ICJ jurisprudence by arbitral tribunals, there are two important differences. First, in the context of
state-owned companies or companies that are de facto controlled by states, the element of hierarchy is de-emphasised. Instead, tribunals have assessed whether or not a state exercises general control over an entity and specific control over the particular acts in question. Secondly, arbitral tribunals do not always inquire into what effective control would require under the applicable primary rules of law. In the context of the Nicaragua decision, for example, the test of ‘effectiveness’ was designed to address non-inherent acts contrary to international law such as violations of IHL.
Decision-makers in other contexts have picked up on these tremors. In Bayindir Insaat Turizm Ticaret Ve Sanayi A S v Islamic Republic of Pakistan, a case involving a contract dispute between an investor and a public authority appointed to run the national highways in Pakistan, the tribunal stated:
the Tribunal is aware that the levels of control required for a finding of attribution under Article 8 in other factual contexts, such as foreign armed intervention or international criminal responsibility, may be different. It believes, however, that the approach developed in such areas of international law is not always adapted to the realities of international economic law and that they should not prevent a finding of attribution if the specific facts of an investment dispute so warrant.
Similarly, the application of the effective control standard to the topical questions of terrorism, PMCs and the right to use force in self-defence against non-state actors has been problematic.
On terrorism, for example, the Articles on State Responsibility predate the terrorist attacks of 9/11 and there was little deep analysis about the applicability of those articles to what is now a pressing international problem. The obligations to comply with anti-terrorism measures, including by treaty and through UN Security Council resolutions, led to spirited debates about whether the existing responsibility framework was sufficient. Because state entanglement with terrorist groups can range from the rare situation in which a state directs and controls a terrorist group (which would meet the art 8 test) to the much more common scenario in which a state provides some logistical support, financing or safe harbor, there are only limited situations in which the stringent effective control test would be met. The result is that the universe of other common but potentially dangerous activities would fall through the cracks of the existing state responsibility framework.
The legality of the use of state force against terrorists in self-defence is also problematic under the effective control standard. Article 51 of the Charter of the United Nations gives states a right to respond to armed attacks, but does it permit the use of force (in self-defence) against non-state actors? When
Al-Qaida attacked US targets in 2001, for example, the United Kingdom and US responded with an invasion of Afghanistan under Operation Enduring Freedom. The legality of these actions rested on whether Afghanistan controlled Al Qaida, because attribution of Al Qaida’s acts to Afghanistan would elevate it into a traditional state-to-state conflict. As Jutta Brunée explained, however, it was apparent that the links between terrorists and states that harbour them would not typically meet the effective control standard, necessitating a new standard:
Since inter state force may be used only in self-defence, military action against another state will require that the state is implicated in the relevant attack. And it is precisely with respect to the nature of the link between the target state of a self-defence action and perpetrators of attacks, such as terrorists, that certain adjustments to the self-defence regime are required. The Nicaragua decision of the International Court of Justice required agency, assessing the use of force within the framework of state responsibility. State practice and opinio juris since 11 September suggest that a shift away from this approach is underway.
Along these lines, scholars like Christian Tams argued for a much lower threshold of control in this context, such as complicity.
The same criticism of the accountability gaps perpetrated by the effective control standard under the state responsibility framework has arisen with regard to PMCs. PMCs are a common element of the conflict landscape today and, like national soldiers, they may engage in acts that violate international law. Because PMCs are rarely integrated into national militaries, their conduct would only be attributable on the basis of art 4 of the Articles on State Responsibility in narrow circumstances. When they exercise governmental authority, they may fall within the scope of art 5, although as White notes, this too is unlikely. If neither art 4 nor art 5 applies, the effective control test under art 8 will become relevant in determining whether a PMC is acting under the direction or control of the state. It has been argued that traditional principles of state responsibility are sufficiently flexible to accommodate PMCs and that the contract itself will be particularly relevant to determining whether or not a state exercises control. Most PMCs are independent entities with different clients, including states that enter into contractual relationships. Sometimes states will have a great deal of control over these situations, however, if a PMC is hired to escort an aid convoy and the state agrees that force can be used to protect the convoys, and lethal force is in fact used, the question may arise whether the PMC is acting under the instructions of the state in using force. White argues that it would seem that the effective control test would apply. This slippage becomes more apparent with the reality that ultra vires acts of PMCs will not be captured by art 8.
It would be a mistake to assume that the challenges posed to the effective control standard by terrorism, self-defence and PMCs are aberrations or isolated examples. In fact, in light of the profound changes occurring within and around states, they represent what is becoming a common phenomenon: greater assumption of functions and standard-setting by non-state actors, which leads to a real or perceived accountability gap. I say ‘real or perceived’ because it is clear that some non-state actors are active and often effective standard-setters, but there may be no central locus of authority to discuss the proper application, interpretation or enforcement of such standards, which could be provided in the context of state regulation.
Despite the ICJ’s efforts in Bosnia to settle the matter, the ongoing dialogue about effective control indicates that tribunals and commentators concerned with diverse questions and different primary rules of law are reassessing the relationship between control, attribution and ultimately responsibility. On the one hand, the elaboration of elements in recent ICJ cases has not quieted the debate about what level of control is truly ‘effective’. Talmon takes the position that ‘[c]ontrol must not be confused with “support”’ and argues that, ‘unlike complete dependence, partial dependence does not allow the Court to treat the authorities of the secessionist entity as a de facto organ of the outside power whose conduct as a whole can be considered acts of the outside power’. In counter-position, Francesco Messineo argues that ‘planning, direction, and support’ could trigger the test under art 8, which would support the less onerous approach advocated in the context of terrorism. Moreover Antonio Cassese notes that
[i]t seems clear ... that by ‘effective control’ the [ICJ] intended either (1) the issuance of directions to the contras by the US concerning specific operations (indiscriminate killing of civilians, etc), that is to say, the ordering of those operations by the US, or (2) the enforcement by the US of each specific operation of the contras, namely forcefully making the rebels carry out those specific operations.
On the other hand, some courts have addressed the slippage problem by using alternative judicial techniques, including finding states internationally responsible for breaching primary norms, such as a duty to prevent, in the absence of meeting the effective control test.
In sum, what has become apparent is that the definition of effective control is heavily dependent on primary norms, the portability of effective control outside the context of IHL has not been seamless and, to avoid a perception of an accountability gap, commentators and some tribunals have advocated lower control thresholds. In the next Part, I analyse the migration of the effective control standard into one final context, the Articles on the Responsibility of IOs, and discuss slippage and the accountability gap in this context. I will then turn to a discussion of omissions, the duty to prevent and act with due diligence and conclude with a discussion about how attribution should be conceptualised in light of the changing nature of the state in international law.
In the Articles on the Responsibility of IOs, the ILC proposed a theory of attribution for IOs. This was the first time that an attempt had been made to articulate general principles of attribution applicable to IOs, which are both subjects of international law and possess separate legal personality.
In developing the Articles on the Responsibility of IOs, the ILC adopted the principle of effective control as a ‘base unit of analysis’ with art 7, the key provision on attribution of conduct between IOs and states, providing:
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.
The ILC’s point of departure is that, prima facie, the conduct at issue is attributable to the lending entity since it would be an organ of a state or an organ or agent of an IO. Nonetheless, in some circumstances, an organ or agent will fall under the effective control of an IO. The question regulated by
art 7 is therefore whether the circumstances are such that an act should be attributable solely to the borrowing entity, or potentially to the borrowing and lending entity, under the principle of shared responsibility.
The ILC does not define effective control in the Articles on the Responsibility of IOs or indicate the extent to which the Nicaragua and Bosnian Genocide decisions influence its application in this context, although the commentary emphasises that the determination is heavily dependent on the facts. Tellingly, Crawford describes ‘effective control’ as a principle of essential ambiguity that ILC members hoped would be fleshed out in practice.
While the common reference to control in the state and IO responsibility context might suggest a shared standard, there are in fact significant differences. The starting premise under the Articles on State Responsibility is that the state is only responsible for the acts of its organs or agents. If the state exercises a high degree of control over irregular groups, that conduct may also be attributable to the state as discussed with reference to art 8 above, but the structure of the Articles on State Responsibility leaves open the possibility that certain conduct will not be attributable to a state at all. In other words, irregular groups that do not operate under a state’s control fall outside the purview of the Articles on State Responsibility, leading to slippage and the so-called accountability gap. Moreover, the Articles on State Responsibility are silent on the level of control required for the attribution of conduct. Although control comes up in the various attribution tests, ‘effective control’ is a judicial standard which was supplied by the ICJ in Nicaragua, confirmed in Bosnian Genocide and applied by different courts and tribunals in subsequent decisions.
The Articles on the Responsibility of IOs use control in a different way. First, they are more specific: the effective control standard is supplied in art 7 of Articles on the Responsibility of IOs, unlike art 8 of the Articles on State Responsibility, which refers generically to control. Secondly, the question under the Articles on the Responsibility of IOs is not if conduct is attributable but rather to which entity it should be attributed: the IO or the state. In this sense, there is no accountability gap under the Articles on the Responsibility of IOs because where there is joint management between states or IOs either or both entities will be responsible. There is, in other words, no minimum quantitative threshold required to satisfy the effective control test: acts could be attributed to either or both entities where an organ is placed at the disposal of another IO. The question of distributing responsibility might, for example, arise where two or more states or international organisations conduct joint military operations in which some soldiers violate IHL.
Control consequently has a different function under the Articles on State Responsibility and Articles on the Responsibility of IOs. In the former, the control test is used to determine whether acts of individuals or groups of individuals that do not have status under international law are attributable to the state. In the latter, effective control is part of a comparative inquiry: assuming states and IOs are both involved, which entity (or both) exercises effective control?
Some of these differences in the content of the effective control standard applicable to IOs can be explained by the nature of the relationship. In most joint ventures, for example, the states and IOs involved would have a formal relationship. Additionally, because peacekeepers are, by definition, part of a national military and act as organs of their national governments, they will never operate like irregular groups.
The effective control standard for IOs has been largely conceived in the peacekeeping context. This context is important for two reasons. First, in its comments to the ILC, the UN questioned the effective control standard, stating that in principle it has exclusive control over national contingents in a peacekeeping force:
As a subsidiary organ of the United Nations, an act of a peacekeeping force is, in principle, imputable to the Organization, and if committed in violation of an international obligation entails the international responsibility of the Organization and its liability in compensation.
As such even the UN, the IO with the deepest connection to the context for which the ILC hoped to develop default rules, expressed concern about the test. However, it should be noted that, in practice, courts, commentators (and even the UN in other contexts) have recognised that the statement above is not as clear-cut as it might appear because who makes the decisions and is in operational command is important in any analysis of attribution. Indeed, one report of the UN Secretary-General notes:
The principle that in coordinated operations liability for combat-related damage in violation of international humanitarian law is vested in the entity in effective command and control of the operation or the specific action reflects a
well-established principle of international responsibility.
The UN has thus recognised the importance of factual control, noting that
states retain disciplinary control over their troops and also citing
instances in which the UN has apportioned damages to a
troop-contributing country on the basis of fault or sought refunds from
On the other hand, IOs — including the EU, International Labour Organization and International Monetary Fund — questioned the applicability of the effective control test outside the peacekeeping context. The EU, for example, wrote that the commentaries to the draft articles were largely devoted to UN practice and to a discussion of the case law of the European Court of Human Rights (‘ECtHR’). It also asked whether international practice is presently clear enough and whether there is identifiable opinio juris that would allow for the proposed standard of effective control. The application of this effective control standard has thus been challenged by organisations outside of the peacekeeping context on the basis it does not take an inside-out approach to IOs, but instead anticipates that IOs will be able to conform to a set of general rules, regardless of structure.
There are thus three important limitations to the conceptualisation and lack of practice associated with the effective control test. First, how should effective control be translated into multi-level governance situations, given the transforming
relationship that actors like the EU have with member states? The same concern would appear to apply to the increasingly common practice by IOs of operating through partnerships, including public–private partnerships, whereby the traditional distinction between public and private functions change. Secondly, how does the concept of effective control, an inherently military concept, translate into the civilian context or into mandates of IOs engaged in development assistance, lending or other activities that are consent-based? Although the ILC suggested that the same principles could apply regardless of context, citing the rather unique situation in which acts of the Pan American Health Organization might be attributed to the World Health Organization on the basis of a long-term contract, it appears that IOs with advisory and consultative mandates will only rarely fall within the purview of art 7. Finally, assuming effective control can be shared by two or more actors, what factors apply to determining how responsibility should be allocated thereafter?
The slippage problem that arises with regard to IOs is therefore somewhat different from the one that arises with regard to states. For IOs, the real issue is the dearth of primary norms defining IO behaviour, and the subsequent development of an effective control test that is almost exclusively derived from the peacekeeping context. The effective control test as a standard for attribution may simply not apply to the majority of IOs in their usual work. Compensating somewhat for this limited scope, however, is the ILC’s recognition of shared responsibility, which opens up the opportunities to reach partners in joint enterprises.
Despite these concerns, the effective control standard in the Articles on the Responsibility of IOs quickly migrated into the case law. There is a growing corpus of decisions addressing effective control in the peacekeeping context, including a 2007 admissibility decision by the ECtHR, Behrami v France, which interpreted and applied the effective control standard while the Articles on the Responsibility of IOs were still in draft form. Subsequently, the effective control standard was applied and interpreted before the UK House of Lords in R (on the Application of Al-Jedda) v Secretary of State for Defence (‘Al-Jedda (HL)’) and later by the ECtHR, and appeared again in domestic courts in the Netherlands in Nuhanović v Netherlands (‘Nuhanović’), which reached the Dutch Supreme Court. These decisions reveal differing views about the content, threshold and application of the principle of control to complex peacekeeping operations, where two core areas of contestation emerge: legal versus factual control, and positive versus negative control. I discuss
these decisions and views in reverse order with the most recent
decision — Nuhanović — first.
On 6 September 2013 the Dutch Supreme Court handed down the decisions of the hotly anticipated cases of State of the Netherlands v Mustafic-Mujic and Nuhanović. Both cases raised the issue of whether the Dutch state was responsible for the deaths of individuals who attempted to take refuge in the Dutch compound during the Srebrenica massacre. The people in issue had sought refuge in the compound of the Dutch battalion (‘Dutchbat’), but commanders decided not to evacuate them along with the battalion and instead sent them away from the compound on 13 July 1995. Outside the compound they were murdered by the Bosnian-Serb army or related paramilitary groups. Citing both the Articles on State Responsibility and the Articles on the Responsibility of IOs, the decision found that the Dutch state exercised effective control over Dutchbat pursuant to art 8 of the Articles on State Responsibility, defined as ‘factual control over specific conduct’. The Court also confirmed that the issue fell under art 7 of the Articles on the Responsibility of IOs because the Netherlands had placed troops at the disposal of a UN peacekeeping mission.
The definition of effective control in Nuhanović emphasises factual control over specific conduct (as opposed to legal control), noting that ‘all factual circumstances and the special context of the case must be taken into account’. The decision picks up the threads of prior case law including the Nicaragua decision by emphasising specific conduct, although the decision does not cite ICJ jurisprudence. Moreover, it leaves the door open to a positive conception of control, through responsibility for the failure to prevent, as discussed in more detail in Part VI.
One unusual aspect of this case is that the Dutchbat mission had conclusively failed at the time the deaths occurred. Nonetheless, the Court invoked the concept of control over territory to support its conclusion that while Nuhanović ‘could therefore no longer exert any influence outside the compound, this does not detract from the fact that the State had effective control over [Nuhanović’s] conduct in the compound’. In so doing, the Court took an expansive view of control in that it found that it was still being exercised even though the mission had failed. The Court was similarly expansive in determining that international law permits dual attribution. Citing arts 7 and 48 of the Articles on the Responsibility of IOs, the Supreme Court held that international law does not exclude the possibility of dual attribution of conduct. This finding enabled the Supreme Court to determine that the Netherlands was responsible, while leaving open the question of whether the UN had effective control in the early evening of 13 July 1995. Unlike the Behrami v France applications before the ECtHR, discussed below, which were struck down at the admissibility stage due to the UN’s involvement (and which led the Court to conclude it could not exercise jurisdiction), the shared responsibility approach leaves the door open to attribute conduct to more actors, and hence expands the range of potential responsibility amongst states and IOs. The possibility of joint control has already been recognised in situations where states share control over individuals
in indefinite detention.
On the relationship between the UN and the Netherlands, the Court noted that ‘it is not necessary for the State to have countermanded the command structure of the United Nations by giving instructions to [Nuhanović] or to have exercised operational command independently’. According to the Dutch Supreme Court, there is no requirement that the state go against the UN’s formal command to determine attribution of conduct and responsibility, which eases the evidentiary burden for demonstrating effective control. Moreover, this decision stands in contrast to the UN’s position in its comments submitted to the ILC during the Articles on the Responsibility of IOs drafting process that only direct contradictory orders to the UN can give a state effective control.
Effective control was the centrepiece of another prominent 2011 decision, the ECtHR’s Grand Chamber decision in Al-Jedda (ECtHR). The ECtHR was asked to determine whether a security detention in Basra constituted a violation of art 5 of the European Convention on Human Rights (‘ECHR’) and if wrongful conduct should be attributed to the UK. The UK had taken the position that responsibility should be attributed to the UN, because the situation was subject to several Security Council resolutions. Like the House of Lords, the ECtHR found that the UK was responsible because
the United Nations Security Council had neither effective control nor ultimate authority and control over the acts and omissions of troops within the
Multi-National Force ... the applicant’s detention was not, therefore, attributable to the United Nations. [However] the internment took place within a detention facility in Basra City, controlled exclusively by British forces, and the applicant was therefore within the authority and control of the United Kingdom throughout.
There are several noteworthy aspects of this decision. First, the ECtHR indicated that it was common ground between the parties and the UK House of Lords that art 5 of the Articles on the Responsibility of IOs applied, which sets the standard as effective control. The ECtHR then proceeded to identify the effective control test and the ultimate authority and control test, stating that neither was met in this case. While it did not define effective control, the ECtHR noted that UK responsibility flowed in part from the UK’s presence on the ground. Moreover the decision to hold and continue holding the applicant in internment was exclusively made by the UK. Indeed, the UN had protested the UK’s practices of indefinite confinement. Importantly, the decision also indicates the potential for dual attribution between states and IOs, in dismissing the argument that an overarching Security Council resolution transfers responsibility to the UN prima facie.
These decisions correct, to a certain extent, the earlier and now infamous ECtHR admissibility decision on effective control in Behrami v France, which involved alleged human rights abuses by states that were both UN members and troop-contributing countries for the UN and North Atlantic Treaty Organization operations following the 1999 Kosovo conflict. In determining whether the alleged violations of the ECHR were attributable to the UN or to the states involved, the ECtHR purported to apply the effective control standard promulgated in the ILC’s Articles on State Responsibility. It has been extensively chronicled that the ECtHR in fact applied the lower test of ‘ultimate
authority and control’. It was on this basis that the Court determined that the Kosovo Force (‘KFOR’) and the United Nations Interim Administration in Kosovo (‘UNMIK’) exercised powers delegated to them by the UN Security Council, and further, that the UN Security Council ‘retained ultimate authority and control and that effective command of the relevant operational matters was retained by NATO’ and had ‘the power to establish, as well as the operational command of, the international presence, KFOR’. As a result, the proceedings were dismissed because the ECtHR determined it did not have jurisdiction. Despite widespread criticism, the ECtHR did not shy away from the ‘ultimate authority and control’ test in subsequent decisions. Some commentators have gone so far as to say that politics and uncomfortable confrontations with the collective security system were behind the decision to focus on the issue of attribution. Indeed, there may be truth in this as the Al-Jedda (ECtHR) decision deliberately affirmed both tests.
The concept of effective control played an under-examined role in another aspect of these decisions as well: the extraterritorial reach of the ECHR. In Behrami v France, the Court determined that Kosovo was no longer controlled by the FRY, but instead that Kosovo was under the effective control of the international presences exercising public powers normally exercised by the Government of FRY. As such, the Court decided that the acts should be attributed to the UN because it exercised control over the territory through its international presence. This reference to effective control over territory is common to other recent decisions including Nuhanović and the Al-Jedda, where a jurisdictional finding of territorial control was used to bolster the stricter effective control standard for attribution of conduct.
A backwards glance is now in order. As a general rule, the deeds of non-state actors are not attributable to states on the basis that states are only responsible for the acts of their agents or organs. There are, however, a number of exceptions, two of which have been examined so far. First, where an agency relationship exists between a state and a private actor, the latter’s conduct may be attributed if it acts under the state’s control. Secondly, where an IO is involved, such as in a UN peacekeeping mission, states and IOs may share effective control. The third exception is that if a state fails to fulfil a primary obligation such as the duty to prevent, it may incur responsibility for any harms that arise as a result. For example, a state may incur responsibility if it did not adequately prevent certain private misconduct subject to a due diligence standard. These exceptions represent three distinct paths to state responsibility for non-state conduct. This article has addressed the first two routes to this point. I will now examine the duty to prevent, which has become an increasingly prevalent way of working around the limits of attribution-based doctrines, signalling a shift in emphasis from secondary to primary norms.
It is uncontroversial that responsibility may arise either by act or omission. Thus it is not only the affirmative act that constitutes a breach of an international obligation, but in some cases a breach can be established by the failure to act. As Crawford reminds us however, ‘an omission is more than simply “not-doing” or inaction: it is legally significant only when there is a legal duty to act which is not fulfilled, and its significance can only be assessed by reference to the content of that duty’. Primary norms are consequently essential to the analysis of whether or not responsibility for an omission arises.
In the Bosnian Genocide case, the ICJ famously illustrated how the secondary rules of attribution and the primary rules containing the duty to prevent interact: while the conduct of secessionist entities and paramilitary groups could not be attributed to the Republika Srpska for lack of control, the Republic was nonetheless responsible for failure to prevent acts of genocide under art 1 of the Articles on State Responsibility. Thus the failure to fulfil a duty to prevent may lead to international responsibility even in the absence of attribution of conduct, if a state’s organs or entities exercising government authority have not fulfilled their duties. While states are not responsible for the acts of private individuals absent a showing of control, they will be responsible for their own failure to protect.
Where the application of the stringent effective control test results in slippage, the duty to prevent has become a favoured strategy. In the case of terrorism, for example, disconnects between the effective control test and state complicity in modern manifestations of terrorism are being addressed in part by focusing on primary obligations to combat and prevent terrorism. Some existing sectoral treaties make state participation in terrorist acts an international wrong under art 4 of the Articles on State Responsibility. Shortfalls have been addressed by Security Council Resolution 1373, which is binding on all UN member states and declares that states shall ‘[p]revent and suppress the financing of terrorist acts’ and ‘prevent those who finance, plan, facilitate or commit terrorist acts from using their respective territories for those purposes against other States or their citizens’. Crawford argues that this is the right approach: better to develop primary norms than artificially extend the effective control test under the doctrine of attribution.
Slippage is a problem for PMCs as well, and two current regulatory efforts are underway to develop primary norms that incorporate a duty to prevent: a convention on an International Code of Conduct, currently in draft form, and the Montreux Document, a non-binding ‘soft law’ document of principles. Both efforts introduce due diligence obligations and the subsidiary obligation to prevent. While these mechanisms would not normally result in the direct attribution of PMC conduct to the state, they would place clearer obligations on the state to regulate and prevent unlawful activities.
Although there is no general duty to prevent, there is a growing list of treaties, Security Council resolutions and judicial decisions that articulate a duty to act. For example:
(i) The Convention on the Prevention and Punishment of the Crime of Genocide, the Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment and the International Convention for the Suppression of Terrorist Bombings are among contemporary treaties that contain a duty to prevent;
(ii) human rights law places positive obligations on states to protect and fulfil; and
(iii) the concept has taken root in international criminal law where a superior ‘should have known’ that crimes were being perpetrated under their de facto control, but did not intervene to prevent them.
To this we can also add the most developed field, environmental law, where states must ensure their territory is not used to cause environmental harm. In this regard, it is important to recognise that a general regime of fault-based responsibility has been proposed in environment law, which contemplates liability for lawful acts.
These duties to act have arisen through custom in certain areas of law, through Security Council resolutions and in new regulatory instruments. Given their contextual specificity, the scope of any such duty is much narrower than that of a secondary standard like effective control under the attribution doctrine promulgated by the ICJ. Indeed, in the Articles on State Responsibility, for example, due diligence is only mentioned in passing, because it addressed a standard of behaviour, which is applicable to primary but not secondary rules. Nonetheless, I now turn to a discussion of the areas in which primary norms such as a duty to prevent relate to slippage and state sovereignty.
A duty to prevent creates an obligation on all states within the jurisdictional reach of a primary rule, and potentially IOs, to curb the effects of conduct of private parties that may breach an international obligation. The obligation to prevent is subject to the due diligence rule. A second, but distinguishable obligation is the duty to punish, which I will not focus on here.
With regard to the first element, the Bosnian Genocide clarified that the content of the duty to prevent genocide is dependent on the capacity of states ‘to influence effectively the action of persons likely to commit, or already committing, genocide’. Capacity is determined by context: the geographic distance between the state and the events, the strengths of the political links between the state and the main actors, and the legal restrictions of action imposed on the state based on the situation. Giorgio Gaja notes that the same capacity-based assessment would presumably apply to an IO’s duty to prevent.
While there is no precise definition of the duty of due diligence, given its very close connection to primary rules, three factors have been suggested as to its content: (i) the degree of effectiveness of the state’s control over territory, (ii) the degree of predictability of harm, and (iii) the importance of the interest to be protected. In addition, it is generally acknowledged that due diligence is to be measured against an international, rather than domestic, standard, and that it has objective not subjective content. A subsidiary consideration, of general importance to duties to prevent and act with due diligence is the expectation that all states possess and use a legal and administrative apparatus able to guarantee respect for prevention.
Due diligence is an elastic and relative concept and in this sense it shares commonalities, some of them negative, with the context specificity of the effective control standard. Although it has been described as a ‘basic principle’ of international law that requires states and IOs to prevent or react to acts or omissions of their organs or agents, capacity, in particular, is an inherently variable concept. It could conceivably lead to less diligence for developing countries or to a variable timescale for completion. On the other hand, it might lead to more diligence if there is a greater degree of risk, such as for hazardous activities.
Another limitation of the diligence standard is that it plays a role only in some areas of international law: there is no general duty of diligence. Where there is no duty to prevent or act with due diligence, it will not be an available tool to address the problem of slippage. Nonetheless, because it undermines the public versus private divide that had hitherto strictly separated activities for which the state could and could not be held responsible, due diligence is relevant to the accountability gap.
One issue that the due diligence standard has not resolved is when the obligation to prevent is triggered. The ICJ held that, in Bosnian Genocide, the obligation arises when there is evidence of a serious risk that genocide will be committed. Because it is construed as one of conduct not result (in the sense that there is no obligation to succeed), the ICJ set a high standard and stated that the state must have manifestly failed to take measures to prevent acts of genocide to be found responsible. Those who seek a robust response to perceived accountability gaps, therefore, have not been assuaged by the articulation of a state’s duty in this regard. As Marko Milanović notes, there is no lex lata to suggest that all states have a duty to prevent or intervene in genocide, although it might be possible to distinguish between degrees of state complicity with regard to potential state responsibility.
Typically, the duty to prevent is derived from primary rules, but in at least one instance, it was read into secondary rules. In Nuhanović, the Netherlands Court of Appeal noted that effective control might be demonstrated in one of two ways: by the implementation of specific instructions from the UN; or, if there was no specific instruction, by the capacity to prevent the wrongdoing. The Court of Appeal consequently used the attribution doctrine to promulgate a positive conception of control, by determining that the Netherlands was responsible if it had the capacity to prevent removal of the plaintiffs from the compound. This interpretation both acknowledges the limitations of relying on direct orders and indicates that states and IOs have a range of powers available to them to prevent against wrongful conduct. Under this conception, responsibility is linked to causation.
On appeal, the Supreme Court did not squarely address the duty to prevent, although it would be wrong to suggest it was completely silent on the matter. At various places in the judgment it hinted that Dutchbat could have prevented the conduct in question. Moreover, the Supreme Court specifically approved of the Court of Appeal’s interpretation and application of the effective control test, writing:
The Court of Appeal’s ruling that the State had effective control over the conduct of which Dutchbat and hence the State as well are accused by Nuhanović does not reveal an incorrect interpretation or application of the law on the concept of effective control.
Thus, while the Supreme Court left for another day the robust approach to control that motivated the Court of Appeals’ decision, it is a direction that remains open to courts in new cases. Such a direction would have pros and cons. The pros might include attributing conduct to the entity best placed to prevent wrongdoing, which might then close the accountability gap in international law. On the other hand, such a concept of control turns on its head the lawmaking process and suggests a very different function for secondary rules than those envisioned by the ILC.
The duty to prevent, as a model of a positive duty of control, has parallels to the doctrine of superior responsibility in international criminal law. Although state and IO responsibility regimes are sui generis, and important differences between these different regimes of responsibility make comparisons difficult, some policy questions are relevant by analogy.
Under the doctrine of superior responsibility, superiors, whether military or civilian, who are not directly involved in the commission of a crime may still be criminally liable if they assert control over others. Because the accused must be shown to exercise control over others and demonstrate the requisite intent
(ie ‘knew or should have known’), the inquiry centres on dominance and influence of one individual over others. Here, the military or civilian superior incurs responsibility for their failure to prevent the acts of subordinates under their control, even if that individual is not at the scene of the crime.
Superior responsibility appears in the statutes of all contemporary international criminal tribunals, with three common elements: (i) a
superior/ subordinate relationship, (ii) where the superior knew or should have known the forces were committing such crimes; and (iii) the superior failed to take all necessary and reasonable measures to prevent the commission of the crime. Control is relevant to the first and third elements, in that it defines the nature of the relationship and the ability of the superior to act and influence subordinates. Although most statutes do not specify the standard, ‘effective command and control’ has consistently been incorporated into the jurisprudence, and ‘effective control’ is defined as the material ability of the military or civilian leader to prevent, repress or submit the matter to the competent authorities.
Consistent with decisions in other fields, criminal tribunals have generally adopted a fact-based approach to control, determining that the actual position and power of a superior is more relevant than any formal rank or position in an organisation. Although the core jurisprudence to date addresses military leaders, there are a growing number of cases that involve civilians who are not in a position to demand unquestioned obedience. In the Prosecutor v Brima (‘Armed Forces Revolutionary Council’) case, for example, the Special Court for Sierra Leone (‘SCSL’) took the position that the factors which may be useful in assessing effective control outside the military context include
that the superior had first entitlement to the profits of war, such as looted property and natural resources; exercised control over the fate of vulnerable persons such as women and children; the superior had independent access to and/or control of the means to wage war, including arms and ammunition and communications equipment; the superior rewarded himself or herself with positions of power and influence; the superior had the capacity to intimidate subordinates into compliance and was willing to do so; the superior was protected by personal security guards ... the superior fuels or represents ideology to which the subordinates adhere; and the superior interacts with external bodies or individuals on behalf of the group.
While some of these grounds seem more relevant to status than control, they nonetheless suggest that: (i) the power to issue superior orders and capacity of taking disciplinary action, and (ii) command structure, are both very important. For example, to establish command responsibility, the SCSL inquired into whether a commander’s orders were obeyed (he could reap the benefits); whether his power remained constant (was not sporadic); whether he could still exercise control over his troops even after they were forced to retreat (despite breakdown of the battalion, the Armed Forces Revolutionary Council fighting force was still cohesive); and whether the accused fitted into the command structure.
These decisions indicate that what are ‘reasonable and necessary measures’ will depend on the circumstances, although they must be lawful and feasible. As such, a heavy emphasis on the facts persists. The primary duty is to prevent future crimes and stop subordinates who are about to commit them, and the secondary duty (specific to the criminal law context) is to punish past crimes. Finally, effective control means that the accused must not only be able to issue orders, but also that the orders are in fact followed. As a result, resistance or disobedience to the orders will create a presumption against effective control.
The very high level of control required under contemporary command responsibility tests, however, has meant that few prosecutions are successful. The 2012 acquittal of Ante Gotovina by the ICTY Appeals Chamber, for example, was justified on the grounds that the Trial Chamber failed to provide specific information to support the conviction, such as whom the accused should have contacted and what additional steps he should have taken to prevent the acts. Similarly, the Chamber indicated that the failure to prevent would need to make a ‘substantial contribution’ to the crimes. These exacting requirements are further illustrated by other cases before ad hoc tribunals that have clarified that, in order to prove effective control, the power to issue orders is not enough, particularly if confirmation of orders is required by others. The difficulties of obtaining convictions under the command responsibility doctrine mean that another doctrine has come to the fore: indirect co-perpetration, codified by art 25(3) of the Rome Statute of the International Criminal Court. In contrast to superior responsibility, this mode of liability attaches to positive acts, not to the failure to prevent and punish. As such, the same shift noted in the field of state responsibility, greater reliance on direct responsibility through primary norms, appears to be taking place simultaneously in international criminal law (‘ICL’).
Stepping back, the role of effective control in ICL offers some insights into the control, attribution and responsibility matrix. First, it demonstrates a common emphasis on a bottom-up approach, which is heavily dependent on the facts. Second, there are commonalities in reasoning in regards to control and the requirement to prevent and punish. For example, Tom Dannenbaum’s argument that effective control, for the purposes of apportioning liability in situations of the kind addressed by art 6 of the Articles on the Responsibility of IOs, should be held by ‘the entity that is best positioned to act effectively and within the law to prevent the abuse in question’, has many parallels with superior responsibility. Nonetheless, while there has been controversy over lower control thresholds in case they expand the potential for attribution of acts to states and IOs, the same cannot be said for ICL. International criminal tribunals have been consistent in insisting on more exacting standards, requiring clear evidence that the superior could have produced end results.
I have argued that the relevance of effective control as a de facto standard for the secondary rules of attribution is waning, despite the ICJ’s affirmation of effective control in the Bosnian Genocide decision. Four concurrent patterns bear this out: (i) there is a palpable movement towards lower control thresholds in certain fields including WTO law, anti-terrorism, PMCs and self-defence against non-state actors, (ii) there is a growing interest in basing responsibility on omissions and primary norms that both emphasise the duty to prevent and create due diligence obligations for states and IOs, (iii) there are efforts to develop criteria for effective control shared by two or more actors, which broaden the potential scope of responsibility, and (iv) IO comments suggest that the ILC’s proposed effective control standard is not relevant to the practice of most organisations, particularly those that engage in consultative functions.
While at first blush it might appear that lowering control thresholds are the best response to the slippage problem, there are several potentially negative consequences of broader state and IO responsibility. First, because the rules of attribution are relevant to defining the nature of the entity suing or being sued, broader rules of attribution can affect the potential liability of a subject under primary rules of international law. In the investment context, for example, the International Centre for Settlement of Investment Dispute’s (‘ICSID’) jurisdiction is limited to investor–state disputes. Typically, the state is a respondent, although sometimes subdivisions, agencies or corporations controlled by a state are sued in an effort to reach the state, which may have deeper pockets. Some decisions on attribution of conduct will affect the scope of potential state liability, and some have cautioned that low control thresholds adopted for the purposes of determining standing ICSID cases could eventually reach state-to-state disputes.
Similarly, questions of attribution have affected jurisdiction over the UN as well. Because IOs are protected by privileges and immunities, the nature of the control test adopted can determine whether or not a court has jurisdiction ratione personae. The Behrami v France decision, discussed above, is a concrete example of this possibility, where the overall control test resulted in an admissibility decision that all acts were attributable to the UN, which was then determined to be outside of the Court’s jurisdiction. A higher control test or the incorporation of a shared responsibility paradigm might have resulted in acts being attributed to member states instead, which have proceeded to be decided on the merits. Recognition of shared responsibility, as noted above with regard to the Nuhanović decision, may also affect the availability of a remedy to plaintiffs and affect the potential scope of state and IO responsibility.
Another consequence of a low attribution threshold is the burden it can place on weak and under-resourced states. As Hessbreugge writes:
today, states more and more share their power with international organizations but even more importantly with non-state or transnational, sub-state actors. Multinational corporations, strengthened by free trade and privatization achieve annual turnovers that dwarf the gross domestic product of developing countries and can wield enormous economic power. Transnational networks of NGOs force countries to adopt new rules of international law such as those embodied in the Ottawa Treaty against landmines ... Armed non-state groups also take advantage of the opportunity to control territory that is left unprotected by weak states and then seek to expand their power even further.
The ILC recognised this trend in art 5 of the Articles on State Responsibility and IO comments to the ILC on the Articles on the Responsibility of IOs emphasised this reality as well. Where the real power of decision resides in non-state actors in any given area, it places greater burdens on states to police them. As Trapp notes:
While a wrongdoing state’s responsibility is not invoked as often as it might be in the terrorism context ... Holding states responsible as a matter of law for more than they are responsible for as a matter of fact will certainly not encourage more reliance on the regime of State responsibility.
Developed states typically have better capacity to do so through legislation and controls, but developing states struggle to implement responsibilities of due diligence. Broad responsibility may also increase the implementation gap in these circumstances and overshadow the alternative approach, which is to develop more comprehensive responsibility regimes for non-state actors.
Far from simply providing technical standards, rules on attribution embody judgments about the scope of state and IO obligations, the range of persons bound by a given set of norms and the potential spread of losses that give rise to remedial rights. Because the content of control tests is heavily dependent on primary norms, superficial similarities in the generic effective control standard tends to mask the very different values at play in sub-fields of international law. These differences are magnified when one considers how effective control can, depending on the circumstances, implicate not only power relationships (control over persons or entities), but also concurrent spatial control (over territory).
While states remain the fundamental building blocks of the international legal system, and in some instances, principal organs and authors of international law, their power is increasingly diffuse. High control thresholds have usually been justified on the basis that they safeguard states from being held responsible for too broad of a range of acts; acts which they might not instigate, direct or be able to prevent. However this article has argued that control is inherently context specific, and as such, abstract secondary rules are ceding to factual reality. This shift towards specific contexts is also borne out by greater emphasis on primary rules and direct responsibility from omissions.
The International Tribunal for the Law of the Sea noted that
while it is not considered reasonable to make a State liable for each and every violation committed by persons under its jurisdiction, it is equally not considered satisfactory to rely on mere application of the principle that the conduct of private persons or entities is not attributable to the State under international law.
The changing nature of the state and movement towards new divisions of labour between public and private, state and non-state, and public and commercial actors in international law reinforce the importance of this statement. Innovative power-sharing arrangements and new theories about the responsibilities associated with state sovereignty mean that control tests may no longer be fit for the future.
Going forward, some pressing issues remain to be resolved. It is apparent that there are gaps in the architecture of legal responsibility, particularly with regard to non-state actors, which are increasingly implicated in many of the harms we encounter as a society. In particular, an international responsibility framework applicable to corporations, joint partnerships, public private partnerships and non-government organisations is in need of development. The development of unified set of principles that addresses states, IOs, non-state actors and individuals, will resolve some of the struggles and inconsistencies apparent in contemporary control tests. Moreover, given the jurisdictional limitations of international tribunals, it will be important to focus on developing primary rules in particular subject areas, rather than relying on particular subjects of international law, in order to better define the responsibilities of different actors in fulfilling positive obligations.
Better alignment of control with primary norms will also be relevant to principles beyond the law of responsibility. For example, the definition of the ‘state’ is relevant to identifying unilateral acts, recognition and the identification of customary international law and norm generation generally. Similarly, where the ‘state’ or ‘IO’ is defined differently for purposes of attribution and immunity, there may be inconsistencies that lead to irrational results. Analogous issues have arisen in international criminal law, with regard to whether a state official who commits an international crime does so in a public or private capacity, which would affect the immunities available.
[*] MA, BA (McGill), JD (New York), JSD (Columbia), Professor, Seton Hall Law School. Special thanks to André Nollkaemper for his comments on this paper, and the opportunity to work on this project as a Visiting Researcher in the SHARES program at the Faculty of Law, Amsterdam. Thanks also to Francesco Messineo, Greg Fox, Jonathan Hafetz, Alice Ristroph and Nicholas Tsagourias for very helpful comments and suggestions.
 James Crawford and Jeremy Watkins, ‘International Responsibility’ in Samantha Besson and John Tasioulas (eds), The Philosophy of International Law (Oxford University Press, 2010) 283, 288 (‘States, lacking bodies of their own, must act through the agency of others’).
 Gordon A Christenson, ‘The Doctrine of Attribution in State Responsibility’ in Richard B Lillich (ed), International Law of State Responsibility for Injuries to Aliens (University Press of Virginia, 1983) 321, 321:
Properly understood, the doctrine of attribution in international law serves the purpose of allocating responsibility to the State for the consequences of certain wrongful acts or omissions of its organs and officials. It also defines the sphere of private or non-State conduct for which the State bears no responsibility.
 As Alain Pellet writes, ‘[r]esponsibility interacts with the notion of sovereignty, and affects its definition’: Alain Pellet, ‘The Definition of Responsibility in International Law’ in James Crawford, Alain Pellet and Simon Olleson (eds), The Law of International Responsibility (Oxford University Press, 2010) 3, 3.
 See Kristen E Boon, ‘New Directions in Responsibility: Assessing the International Law Commission’s Draft Articles on the Responsibility of International Organizations’ (2011) 37 Yale Journal of International Law Online <http://www.yjil.org/docs/pub/o-37-bo
on-new-directions-in-responsibility.pdf> (describing some of the controversies associated with the Articles on the Responsibilities of IOs, see below n 13).
 States, and under certain circumstances, organisations, are subjects of international law: James Crawford, Brownlie’s Principles of Public International Law (Oxford University Press, 8th ed, 2012) 115 (‘[A] subject of international law is an entity possessing international rights and obligations and having the capacity (a) to maintain its rights by bringing international claims; and (b) to be responsible for its breaches of obligation by being subjected to such claims’ (citations omitted)).
 Indeed, James Crawford has recently written that the standard of control is now a settled question: James Crawford, State Responsibility: The General Part (Cambridge University Press, 2013) 156:
So far as the law of state responsibility is concerned, this determination [the ICJ’s Bosnian Genocide decision] effectively ends the debate as to the correct standard of control to be applied under Article 8. Moreover it does so in a manner that reflects the ILC’s thinking on the subject from the time the term ‘control’ was introduced into then-Draft Article 8.
 Nigel D White, ‘Due Diligence Obligations of Conduct: Developing a Responsibility Regime for PMSCs’ (2012) 31 Criminal Justice Ethics 233, 239 (‘Disputes in international legal doctrine about the nature of the control test for the attribution of acts of private actors are set to continue and reflect the failure of international law to keep pace with changes in the structure of states and organizations’).
 See generally Carsten Hoppe, ‘Passing the Buck: State Responsibility for Private Military Companies’ (2008) 19 European Journal of International Law 989; John Cerone, ‘Minding the Gap: Outlining KFOR Accountability in Post-Conflict Kosovo’ (2001) 12 European Journal of International Law 469; Liesbeth Zegveld, Accountability of Armed Opposition Groups in International Law (Cambridge University Press, 2002).
 See Jan Arno Hessbreugge, ‘The Historical Development of the Doctrines of Attribution and Due Diligence in International Law’ (2004) 36 New York University Journal of International Law and Politics 265 (arguing that movement in the law of state responsibility is nothing new).
 See Sheila R Foster, ‘Collective Action and the Urban Commons’ (2011) 87 Notre Dame Law Review 57, 59 (discussing regulatory slippage in the context of common resources, where private actors are given scope to manage collective resources).
 Application of the Convention on the Prevention and Punishment of the Crime of Genocide (Bosnia and Herzegovina v Serbia and Montenegro) (Judgment)  ICJ Rep 43, 214  (‘Bosnian Genocide’).
 International Law Commission, Report of the International Law Commission on the Work of Its Fifty-Third Session, UN GAOR, 56th sess, Supp No 10, UN Doc A/56/10 (2001)
ch IV(E) (‘Draft Articles on Responsibility of States for Internationally Wrongful Acts’) (‘Articles on State Responsibility’).
 International Law Commission, Report of the International Law Commission on Its
Sixty-Third Session, UN GAOR, 66th sess, Supp No 10, UN Doc A/66/10 (2011) ch V(E) (‘Text of the Draft Articles on the Responsibility of International Organizations’) (‘Articles on the Responsibility of IOs’).
 Hague Convention (IV) respecting the Laws and Customs of War on Land, opened for signature 18 October 1907 (entered into force 26 January 1910) annex (‘Regulations respecting the Laws and Customs of War on Land’) s 3 .
 Geneva Convention relative to the Protection of Civilian Persons in Time of War, opened for signature 12 August 1949, 973 UNTS 287 (entered into force 21 October 1950) (‘Geneva Convention IV’).
 See generally David D Caron, ‘The Basis of Responsibility: Attribution and Other
Trans-Substantive Rules’ in Richard B Lillich and Daniel Barstow Magraw (eds), The Iran–United States Claims Tribunal: Its Contributions to the Law of State Responsibility (Transnational, 1998) 109.
 White, ‘Due Diligence Obligations’, above n 7, 236–7.
 See discussion below in Part III.
 Crawford, The General Part, above n 6, 205.
 On the definition of primary rules, see Antonio Cassese, International Law (Oxford University Press, 2nd ed, 2005) 244 (emphasis in original):
It is now generally acknowledged that a distinction can be made between ‘primary rules’ of international law, that is, those customary or treaty rules laying down substantive obligations for States (on State immunities, treatment of foreigners, diplomatic and consular immunities, respect for territorial sovereignty, etc), and ‘secondary rules’, that is, rules establishing (i) on what conditions a breach of a ‘primary rule’ may be held to have occurred and (ii) the legal consequences of this breach.
 See, eg, Human Rights Committee, General Comment No 31: The Nature of the General Legal Obligation Imposed on States Parties to the Covenant, 80th sess, UN Doc CCPR/C/21/Rev.1/Add.13 (26 May 2004)  (providing that states have the duty to guarantee and respect the International Covenant on Civil and Political Rights (‘ICCPR’) at home and abroad for individuals within their ‘power or effective control’). Although practice under the ICCPR and the European Convention on Human Rights is not clear cut, some important cases that address effective control over territory include: Banković v Belgium (European Court of Human Rights, Grand Chamber, Application No 52207/99,
21 December 2001); Issa v Turkey (European Court of Human Rights, Second Section, Application No 31821/96, 16 November 2004). See also R (Al-Jedda) v Secretary of State for Defence  1 AC 332 (‘Al-Jedda’).
 This article also does not address attribution in related fields of domestic law, such as attribution under the Foreign Sovereign Immunities Act 28 USC § 1330 (1976) or under the ‘act of state’ doctrine, although there are parallels and shared insights.
 See Crawford, Brownlie’s Principles of Public International Law, above n 5.
 Military and Paramilitary Activities in and against Nicaragua (Nicaragua v United States of America) (Judgment)  ICJ Rep 14 (‘Nicaragua’).
 Armed Activities on the Territory of the Congo (Democratic Republic of the Congo v Uganda) (Judgment)  ICJ Rep 168 (‘Armed Activities’).
 Bosnian Genocide  ICJ Rep 43, 214 .
 See, however, the discussion of the International Court of Justice’s (‘ICJ’) decision in Nicaragua where the ICJ found that certain acts were attributable to the United States, although acts of the Sandinistas were not: Nicaragua  ICJ Rep 14, 62–3 –.
 See especially Prosecutor v Tadić (Judgement) (International Criminal Tribunal for the Former Yugoslavia, Appeals Chamber, Case No IT-94-1-A, 15 July 1999) (‘Tadić’) (finding that the appropriate test to be overall control for the determination of the existence of an international conflict). See discussion below at Part III(b).
 The variety of control thresholds, such as ‘effective’, ‘strict’, ‘overall’, ‘ultimate’ and ‘meaningful’ control are discussed in the pages that follow.
 See discussion below on terrorism, private military contractors and self-defence against non-state actors.
 Nicaragua  ICJ Rep 14.
 Acts of US agents were attributable because they were organs of the US. Acts of the ‘Latino Assets’ (‘UCLAs’) were attributable either because the UCLAs had been given specific instructions by US agents or officials, and had acted under their supervision, or because those agents had planned, directed, or supported specific operations: Nicaragua  ICJ Rep 14, 45–8 –. As the ICJ wrote: ‘The execution was the task rather of the “UCLAs”, while United States nationals participated in the planning, direction and support’: at 50–1 . For a helpful discussion of the three categories, see Hannah Tonkin, State Control over Private Military and Security Companies in Armed Conflict (Cambridge University Press, 2011) 117.
 Crawford, The General Part, above n 6, 147 (emphasis in original).
 Ibid (emphasis in original) (citations omitted).
 Nicaragua  ICJ Rep 14, 62 . As the International Law Commission (‘ILC’) notes in the commentary to art 8 in the Articles on State Responsibility, ‘conduct will be attributable to the State only if it directed or controlled the specific operation and the conduct complained of was an integral part of that operation’: Articles on State Responsibility, UN Doc A/56/10, ch IV(E) 104  (commentary to art 8) (also noting that in Nicaragua, the ICJ rejected Nicaragua’s claim that all the conduct of the Contras was attributable to the US by reason of its control over them: at 105–6).
 Nicaragua  ICJ Rep 14, 62 . See also Bosnian Genocide  ICJ Rep 43, 202–7 –.
 Stefan Talmon, ‘The Responsibility of Outside Powers for Acts of Secessionist Entities’ (2009) 58 International and Comparative Law Quarterly 493, 498–9 (arguing it is not enough that an outside power take advantage of an independent group, support it or share common objectives. Instead, it is necessary to demonstrate that the assistance is so crucial to the entity’s operations that dependence and control are mirror images of each other); Crawford, The General Part, above n 6, 125 (citing Nicaragua, Crawford identifies relevant factors as to whether a state created the non-state entity, whether the state selected the leaders of the group and whether state involvement exceeded the provision of training and financial assistance: Nicaragua  ICJ Rep 14, 62–3).
 Nicaragua  ICJ Rep 14, 64 .
 Nicaragua (Separate Opinion of Judge Ago)  ICJ Rep 14, 188  (‘Only in cases where certain members of those forces happened to have been specifically charged by United States authorities to commit a particular act, or carry out a particular task of some kind on behalf of the United States, would it be possible so to regard them’).
 Tadić (International Criminal Tribunal for the Former Yugoslavia, Appeals Chamber, Case No IT-94-1-A, 15 July 1999). See, eg, Antonio Cassese, ‘The Nicaragua and Tadic Tests Revisited in Light of the ICJ Judgment on Genocide in Bosnia’ (2007) 18 European Journal of International Law 649.
 Tadić (International Criminal Tribunal for the Former Yugoslavia, Appeals Chamber, Case No IT-94-1-A, 15 July 1999) 49 .
 The Court continued:
In the view of the Court, the conduct of the [Movement for the Liberation of the Congo (‘MLC’)] was not that of ‘an organ’ of Uganda (Article 4, International Law Commission Draft Articles on Responsibility of States for internationally wrongful acts, 2001), nor that of an entity exercising elements of governmental authority on its behalf (Art 5). The Court has considered whether the MLC’s conduct was ‘on the instructions of, or under the direction or control of’ Uganda (Art 8) and finds that there is no probative evidence by reference to which it has been persuaded that this was the case. Accordingly, no issue arises in the present case as to whether the requisite tests are met for sufficiency of control of paramilitaries.
Armed Activities  ICJ Rep 168, 226  (citations omitted).
 The ICJ cited the Declaration on Principles of International Law concerning Friendly Relations, which provides that
[e]very State has the duty to refrain from organizing, instigating, assisting or participating in acts of civil strife or terrorist acts in another State or acquiescing in organized activities within its territory directed towards the commission of such acts, when the acts referred to in the present paragraph involve a threat or use of force. ... [N]o State shall organize, assist, foment, finance, incite or tolerate subversive, terrorist or armed activities directed towards the violent overthrow of the régime of another State, or interfere in civil strife in another State.
Declaration on Principles of International Law concerning Friendly Relations and
Co-Operation among States in accordance with the Charter of the United Nations, GA Res 2625 (XXV), UN GAOR, 6th Comm, 25th sess, 1883rd plen mtg, Agenda Item 85, UN Doc A/RES/2625(XXV) (24 October 1970) annex 123–4 (‘Declaration on Principles of International Law concerning Friendly Relations and Co-Operation among States in Accordance with the Charter of the United Nations’); Armed Activities  ICJ Rep 168, 226 –.
 Armed Activities  ICJ Rep 168, 253 .
 Cf Jay Butler, ‘Responsibility for Regime Change’ (2014) 114 Columbia Law Review 503, 551 (arguing that the Armed Activities case reveals a ‘lex specialis as to attribution’ (emphasis added)).
 See Bosnian Genocide  ICJ Rep 43, 208 –, 210 , citing Nicaragua  ICJ Rep 14, 64 ; and Articles on State Responsibility, UN Doc A/56/10,
ch IV(E) art 8 (which, the ICJ noted is customary international law: at 209 ).
 Bosnian Genocide  ICJ Rep 43, 210 .
 Ibid 214 .
 Ibid 208 .
 Ibid 208–9 .
 Ibid 224 . See also Mukeshimana-Ngulinzira v Belgium, RG No 04/4807/A and 07/15547/A, 8 December 2010, reported in  Oxford Reports on International Law in Domestic Courts 1604.
 See Tonkin, above n 32, 118–19:
The former is determined by the primary rules of international law, which govern the substantive obligations on states, whereas the latter is determined by the secondary rules of international law, which govern the circumstances in which states will be considered responsible for wrongful conduct and the legal consequences flowing from that responsibility.
 Daniel A Farber, ‘Taking Slippage Seriously: Noncompliance and Creative Compliance in Environmental Law’ (1999) 23 Harvard Environmental Law Review 297, 299–300 (defining ‘affirmative’ and ‘negative’ regulatory slippage: affirmative slippage refers to action which gets ahead of regulatory baselines or standards, whereas negative slippage refers to actions that fall behind those baselines or standards).
 A public–private partnership (‘PPP’) is an arrangement whereby the private sector is involved in providing public infrastructure. PPPs come in many shapes and sizes, the
high-water mark being the private sector (usually a consortium of various private companies) designing, financing, building, operating and maintaining public infrastructure under a long term (30-year-plus) contract. PPPs have been used for transportation (roads, bridges, tunnels, airports and ports), social infrastructure (hospitals, schools, university, courthouses and civic buildings) and water/wastewater facilities projects.
 For a definition of multi-level governance, see Philippe C Schmitter,
‘Neo-Neofunctionalism’ in Antje Wiener and Thomas Diez (eds), European Integration Theory (Oxford University Press, 2004) 45, 49 (emphasis in original):
[Multi-level governance] can be defined as an arrangement for making binding decisions
that engages a multiplicity of politically independent but otherwise interdependent
actors — private and public — at different levels of territorial aggregation in more-or-less continuous negotiation/deliberation/implementation, and that does not assign exclusive policy compétence or assert a stable hierarchy of political authority to any of these levels.
 Dan Assaf, ‘Conceptualising the Use of Public–Private Partnerships as a Regulatory Arrangement in Critical Information Infrastructure Protection’ in Anne Peters et al (eds), Non-State Actors as Standard Setters (Cambridge University Press, 2009) 61, 65 (showing a continuum from more interventionist to less interventionist government regulation).
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) art 55.
 This separation between attribution and responsibility was considered a great intellectual contribution of Roberto Ago, ILC Special Rapporteur, who produced a series of important and detailed reports on the topic, including the first draft of the Articles on State Responsibility. Along with the subsequent five reports, see ‘First Report on State Responsibility’ (1969) II Yearbook of the International Law Commission 125, UN Doc A/CN.4/217; ‘Second Report on State Responsibility’ (1970) II Yearbook of the International Law Commission 177, UN Doc A/CN.4/223.
 See discussion below of de facto arrangements.
 See discussion below of omissions, the duty to prevent and due diligence.
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) art 2 (‘There is an internationally wrongful act of a State when conduct consisting of an action or omission: (a) Is attributable to the State under international law; and (b) Constitutes a breach of an international obligation of the State’).
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) 71–2 – (commentary to art 2).
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) arts 4, 5, 8. See also Francesco Messineo, ‘Multiple Attribution of Conduct’ (SHARES Research Paper No 11, Amsterdam
Center for International Law, 2012) 7–8 (Francesco Messineo’s approach places functional and institutional in the same category, on the basis that de facto and de jure organs exercise functions of the state or IO and factual in a separate category, where factual is understood as constituting instructions, direction or control).
 De jure organs might be anywhere in the hierarchy and, in practice, acts by municipal authorities, courts, government agencies, cities and ministries are all attributable to central governments: see Tokios Tokelés v Ukraine (Jurisdiction) (ICSID Arbitral Tribunal, Case No ARB/02/18, 29 April 2004) – (actions of municipal authorities are attributable to the central government); Panel Report, United States — Measures Affecting the Cross-Border Supply of Gambling and Betting Services, WTO Doc WT/DS285/R
(10 November 2004) [6.126]–[6.130] (actions by the International Trade Commission (an agency of the US Government) attributable to the US); Panel Report, Korea — Measures Affecting Government Procurement, WTO Doc WT/DS163/R (1 May 2000) [6.4]–[6.5] (state responsible for answers given by ministry of commerce); Tecnicas Medioambientales Tecmed SA v United Mexican States (Award) (ICSID Arbitral Tribunal, Case No ARB (AF)/00/2, 29 May 2003)  (actions by National Ecology Institute of Mexico attributable to Mexico); Compañiá de Aguas del Aconquija SA v Argentine Republic (Award) (ICSID Arbitral Tribunal, Case No ARB/97/3, 20 August 2007) [7.4.18] (acts of Tucumán Province are attributable to Argentina); LaGrand Case (Germany v United States of America) (Provisional Measures)  ICJ Rep 9, 16  (‘international responsibility of a State is engaged by the action of the competent organs and authorities acting in that State’).
 See Articles on State Responsibility, UN Doc A/56/10, ch IV(E) art 4:
1. The conduct of any State organ shall be considered an act of that State under international law, whether the organ exercises legislative, executive, judicial or any other functions, whatever position it holds in the organization of the State, and whatever its character as an organ of the central government or of a territorial unit of the State. 2. An organ includes any person or entity which has that status in accordance with the internal law of the State.
 Difference relating to Immunity from Legal Process of a Special Rapporteur of the Commission of Human Rights (Advisory Opinion)  ICJ Rep 62, 87 .
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) art 7.
 Talmon, above n 37, 499–501.
 Nicaragua  ICJ Rep 14, 62 :
What the Court has to determine at this point is whether or not the relationship of the contras to the United States Government was so much one of dependence on the one side and control on the other that it would be right to equate the contras, for legal purposes, with an organ of the United States Government, or as acting on behalf of that Government (emphasis in original).
See also Bosnian Genocide  ICJ Rep 43, 207 :
[if] at the time in question, the persons or entities that committed the acts of genocide at Srebrenica had such ties with the [Federal Republic of Yugoslavia] that they can be deemed to have been completely dependent on it; it is only if this condition is met that they can be equated with organs of the Respondent for the purposes of its international responsibility.
 Nicaragua  ICJ Rep 14, 62  (emphasis in original); Bosnian Genocide  ICJ Rep 43, 207 .
 As Olleson notes, ‘the question of the possibility of attribution to the State of conduct of a persons [sic], gropus [sic] or entities on this basis is not dealt with expressly by the ILC’s Articles, nor is it discussed in the ILC’s Commentaries to Chapter II’: Simon Olleson, ‘The Impact of the ILC’s Articles on Responsibility of States for Internationally Wrongful Acts’ (Preliminary Draft, British Institute of International and Comparative Law, 10 October 2007) 28 <http://www.biicl.org/files/3107_impactofthearticlesonstate_responsibility
 In its commentary to art 8, as discussed below, the ILC acknowledged such variations were appropriate when it stated: ‘it is a matter for appreciation in each case whether particular conduct was or was not carried out under the control of a State, to such an extent that the conduct controlled should be attributed to it’: Articles on State Responsibility, UN Doc A/56/10, ch IV(E) 107  (commentary to art 8).
 For a general discussion of how to interpret the de facto category, see Jörn Griebel and Milan Plücken, ‘New Developments regarding the Rules of Attribution? The International Court of Justice’s Decision in Bosnia v Serbia’ (2008) 21 Leiden Journal of International Law 601. But see Marko Milanović, ‘State Responsibility for Acts of Non-State Actors: A Comment on Griebel and Plücken’ (2009) 22 Leiden Journal of International Law 307.
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) art 4.
 Ibid art 4(2) (emphasis added). See also Aurel Sari and Ramses A Wessel, ‘International Responsibility for EU Military Operations: Finding the EU’s Place in the Global Accountability Regime’ in Bart Van Vooren, Steven Blockmans and Jan Wouters (eds), The EU’s Role in Global Governance: The Legal Dimension (Oxford University Press, 2013) 126.
 Crawford, The General Part, above n 6, 124.
 Ibid 125.
 See Claus Kress, ‘L’organe de facto en droit international public. Reflexions sur l’imputation a l’Etat de l’acte d’un particulier a la lumiere des developments recents [The De Facto Organ of Public International Law. Reflections on the Attribution to the State of the Act of an Individual in Light of Recent Developments]’ (2001) 105 Revue Générale de Droit International Public 93 (describing the doctrine of attribution as expansionist).
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) art 5:
The conduct of a person or entity which is not an organ of the State under article 4 but which is empowered by the law of that State to exercise elements of the governmental authority shall be considered an act of the State under international law, provided the person or entity is acting in that capacity in the particular instance.
 Ibid. Common examples are prisons and immigration functions.
 See, eg, ibid 93  (noting that executive control is not a decisive criteria).
 Ibid 94 .
 Appellate Body Report, Canada — Measures Affecting the Importation of Milk and the Exportation of Dairy Products, WTO Doc WT/DS103/AB/R, AB-1999-4 (13 October 1999) – (finding that provincial milk marketing boards are agencies of Canada’s governments, and noting that ‘[a]lthough the provincial boards enjoy a high degree of discretion ... governments retain “ultimate control” over them’).
 Appellate Body Report, Canada — Measures Affecting the Importation of Milk and the Exportation of Dairy Products — Second Recourse to Article 21.5 of the DSU by New Zealand to the United States, WTO Doc WT/DS103/AB/RW2, AB-2002-6 (20 December 2002) – (finding liability for the failure to properly supervise the acts of private parties).
 Marrakesh Agreement Establishing the World Trade Organization, opened for signature 15 April 1994, 1867 UNTS 3 (entered into force 1 January 1995) annex 1A (‘Agreement on Subsidies and Countervailing Measures’).
 Appellate Body Report, United States — Definitive Anti-Dumping and Countervailing Duties on Certain Products from China, WTO Doc WT/DS379/AB/R, AB-2010-3
(2 March 2011) 123  (where the US Department of Commerce had determined that various Chinese State-Owned Entities and State-Owned Commercial Banks be characterised as ‘public bodies’. It further found that a government must exercise ‘meaningful control’ over an entity for conduct to be attributable).
 Ibid 123 .
 Cf EDF (Services) Limited v Romania (Award) (ICSID Arbitral Tribunal, Case No ARB/05/13, 8 October 2009)  (‘EDF v Romania’) (discussing the importance of corporate separateness).
 Crawford, The General Part, above n 6, 127.
 See Christine Chinkin, ‘A Critique of the Public/Private Dimension’ (1999) 10 European Journal of International Law 387, 390 (arguing that any universal claims to a strict
public–private dichotomy are not sustainable, and that a state responsibility regime that ‘assumes a commonly accepted rationale for distinguishing between the conduct of state organs and that of other entities ... in fact depends upon philosophical convictions about the proper role of government and government intervention’).
 See Barcelona Traction, Light and Power Company, Limited (Belgium v Spain) (Second Phase)  ICJ Rep 3, 39.
 Tonkin, above n 32, 93–4.
 Talmon, above n 37, 502.
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) art 8.
 Crawford, The General Part, above n 6, 145; Tonkin, above n 32, 114–15. A contemporary example might involve a private military corporation who is hired by a state to take on certain activities, with instructions being incorporated into their contract or issued by the state while they are in the field. Here, the state would not need to exercise any particular level of control, because the instructions would fall within the scope of art 8.
 See, eg, EDF v Romania (ICSID Arbitral Tribunal, Case No ARB/05/13, 8 October 2007)  (where an ICSID Tribunal found that Romania was using its ownership interest in and control of the corporations AIBO and TAROM to ‘achieve a particular result’, namely bringing the contractual arrangements to an end with two entities, which met the art 8 test).
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) 107 
(commentary to art 8), quoted in Appellate Body Report, United
States — Countervailing Duty Investigation on Dynamic Random Access Memory Semiconductors (DRAMS) from Korea, WTO Doc WT/DS296/AB/R, AB-2005-4 (27 June 2005)  n 188.
 Articles on the Responsibility of IOs, UN Doc A/66/10, ch V(E) 88  (citations omitted).
 Crawford, The General Part, above n 6, 141.
 Cf Bosnian Genocide  ICJ Rep 43, 208  (argument that genocide was of ‘a particular nature, in that it may be composed of a considerable number of specific acts separate, to a greater or lesser extent, in time and space’ with the consequence that the Court would be justified in assessing the effective control over the entirety of the operations carried out by the individuals committing genocide, rather than in relation to each specific act making up the genocide). See also Olleson, ‘Impact’, above n 71, 84–5.
 This is reinforced by the ICJ’s later decision in Bosnian Genocide in which it rejected the ICTY’s approach to control in Tadić: see Tadić (International Criminal Tribunal for the Former Yugoslavia, Appeals Chamber, Case No IT-94-1-A, 15 July 1999) 55–60
 Bosnian Genocide  ICJ Rep 43, 142 , 203 , 206 ; Nicaragua  ICJ Rep 14, 63 . While Nicaragua’s pleadings identified US involvement in the contra movement in general, including funding, organising, training, supplying and equipping of a group, they did not identify support for individual missions: at 86 .
 Talmon, above n 37, 502.
 Tadić (International Criminal Tribunal for the Former Yugoslavia, Appeals Chamber, Case No IT-94-1-A, 15 July 1999) 62 . This standard goes beyond equipping and financing a group, and also involves ‘participation in the planning and supervision of military operations’. It does not require issuance of specific orders.
 Talmon, above n 37, 503.
 Kimberley N Trapp, State Responsibility for International Terrorism (Oxford University Press, 2011) 43.
 White Industries Australia Ltd v Republic of India (Final Award) (UNCITRAL Arbitral Tribunal, 30 November 2011) [5.1.27] (‘White Industries (Final Award)’), citing Jan de Nul NV v Arab Republic of Egypt (Award) (ICSID Arbitral Tribunal, Case No ARB/04/13, 6 November 2008) 55 . See also Gustav F W Hamester GmbH and Co KG v Republic of Ghana (Award) (ICSID Arbitral Tribunal, Case No ARB/07/24, 18 June 2010) 52 .
 See, eg, the UNCITRAL Arbitral Tribunal’s dismissal of arguments relating to organisational structure: White Industries (Final Award) (UNCITRAL Arbitral Tribunal, 30 November 2011) 83 [8.1.5]–[8.1.6].
 Ibid 86–7 [8.1.18].
 See, eg, ibid 87 [8.1.19]–[8.1.21] (discussing absence of Indian control over particular acts, but not referring to primary rules of international law).
 Trapp, above n 106, 43 (‘In Nicaragua, the ICJ was dealing with two separate levels of activity: the first was the paramilitary operations of the contras ... the second level of activity involved violations of humanitarian law perpetrated in the course of those paramilitary operations’ (emphasis in original)).
 Bayindir Insaat Turizm Ticaret Ve Sanayi AŞ v Islamic Republic of Pakistan (Award) (ICSID Arbitral Tribunal, Case No ARB/03/29, 27 August 2009) 35  (contract dispute involving Bayindir and the National Highway Authority, a public corporation established by the National Highway Authority Act 1991 (Pakistan) to assume responsibility for the planning, development, operation and maintenance of Pakistan’s national highways and roads).
 But see Luigi Condorelli, ‘The Imputability to States of Acts of International Terrorism’ (1989) 19 Israel Yearbook on Human Rights 233.
 See, eg, SC Res 1540, UN SCOR, 59th sess, 4956th mtg, UN Doc S/RES/1540 (28 April 2004); SC Res 1989, UN SCOR, 66th sess, 6557th mtg, UN Doc S/RES/1989 (17 June 2011).
 See, eg, SC Res 1373, UN SCOR, 56th sess, 4385th mtg, Agenda Item 166, UN Doc S/RES/1373 (28 September 2001) (‘Resolution 1373’); Tal Becker, Terrorism and the State: Rethinking the Rules of State Responsibility (Hart, 2006) 239 (‘prevailing conceptions of State responsibility ... are inappropriate, from a theoretical perspective, for the particular problems posed by terrorism’). See also at 258: ‘it is difficult to understand why a State should be spared direct responsibility for the consequences of wrongful private conduct that it has knowingly helped bring about, simply because its conduct does not fit tidily into an agency construct’; Anne-Marie Slaughter and William Burke-White, ‘An International Constitutional Moment’ (2002) 43 Harvard International Law Journal 1,
19–20; Rudiger Wolfrüm, ‘State Responsibility for Private Actors: An Old Problem of Renewed Relevance’ in Maurizio Ragazzi (ed), International Responsibility Today: Essays in Memory of Oscar Schachter (Martinus Nijhoff, 2005) 423.
 See Steven R Ratner, ‘Jus ad Bellum and Jus in Bello after September 11’ (2002) 96 American Journal of International Law 905 (arguing that harbouring terrorists should give rise to state responsibility).
 Trapp, above n 106, ch 2.
 Jutta Brunnée, ‘The Security Council and Self-Defence: Which Way to Global Security?’ in Niels Blokker and Nico Schrijver (eds), The Security Council and the Use of Force: Theory and Reality — A Need for Change? (Martinus Nijhoff, 2005) 107, 122–3. See also Gilbert Guillaume, ‘Terrorism and International Law’ (2004) 53 International and Comparative Law Quarterly 537, 546–7. But see Sean D Murphy, ‘Terrorism and the Concept of “Armed Attack” in Article 51 of the UN Charter’ (2002) 43 Harvard International Law Journal 41, 50–1.
 See, eg, Christian J Tams, ‘The Use of Force against Terrorists’ (2009) 20 European Journal of International Law 359, 385:
contemporary practice suggests that a territorial state has to accept anti-terrorist measures of self-defence directed against its territory where it is responsible for complicity in the activities of terrorists based on its territory — either because of its support below the level of direction and control or because it has provided a safe haven for terrorists.
 White, ‘Due Diligence Obligations’, above n 7, 237. But see Tonkin, above n 32, 111.
 Tonkin, above n 32, 80. If courts were to adopt a context specific approach and examine a particular contract, Tonkin argues the effective control test would be satisfied:
Yet the elements of control identified by the Court would be highly significant if exercised over a single PMSC operation, rather than over the company itself. The hiring state will generally have a preponderant or decisive role in selecting, financing, organising and planning the particular PMSC operation to be performed under the contract, and in some cases the state will also supply and equip the contractors for the operation. The contract will ordinarily set out the specific goals of the operation, and in some cases it may also detail how the contractors must be trained, as well as identifying any specific weapons or equipment that must be supplied by the company itself. Any failure on the part of the company to comply with these terms may result in contractual penalties and even termination: at 120 (emphasis in original).
 White, ‘Due Diligence Obligations’, above n 7, 238. See also Oliver R Jones, ‘Implausible Deniability: State Responsibility for the Actions of Private Military Firms’ (2009) 24 Connecticut Journal of International Law 239.
 Hoppe, above n 8, 992.
 Julia Black, ‘Legitimacy, Accountability and Polycentric Regulation: Dilemmas, Trilemmas and Organisational Response’ in Anne Peters et al (eds), Non-State Actors as Standard Setters (Cambridge University Press, 2009) 241, 247.
 Talmon, above n 37, 502–3 (emphasis in original).
 Messineo, above n 63, 9 (discussing para 86 of the Nicaragua judgment).
 Cassese, ‘The Nicaragua and Tadić Tests Revisited’, above n 40, 653 (emphasis in original).
 Prior to the Articles on the Responsibility of IOs, the ILC had not addressed the relationship of international responsibility between states and IOs, as indicated by art 57 which provides that the Articles are without prejudice to any question of responsibility under international law of an international organisation: Articles on the Responsibility of IOs, UN Doc A/66/10, ch V(E) art 57.
 Reparation for Injuries Suffered in the Service of the United Nations (Advisory Opinion)  ICJ Rep 174, 179.
 Crawford, The General Part, above n 6, 195–6 (‘The emphasis on effective control in determining the division of international responsibility between the UN and contributing states, was adopted as the base unit of analysis by the ILC in developing the [Articles on Responsibility of IOs]’).
 Articles on the Responsibility of IOs, UN Doc A/66/10, ch V(E) art 7.
 In this regard, see Borhan Amrallah, ‘The International Responsibility of the United Nations for Activities Carried out by UN Peace-Keeping Forces’ (1976) Revue Egyptienne de Driot International 32, 73–4, cited in Crawford, The General Part, above n 6, 190:
To determine whether an unlawful act is imputable to the UN, the fundamental rule of [the] international law of responsibility ... should be applied, ie the international responsibility should be borne by the state whose organ or agent had committed the wrongful act. The UN may be held responsible for the unlawful act committed by a member of its force so long as this member could be considered [as] acting as an organ or agent of the UN ... The UN should not be held responsible for activities carried out by a member state using its own organs and under its full organic jurisdiction, even if those activities were in application of a decision [taken] by the UN.
 In a 2004 report, the ILC described ‘effective control’ as ‘the factual control that is exercised over the specific conduct taken by the organ or agent placed at the receiving organization’s disposal’: International Law Commission, Report of the International Law Commission, UN GAOR, 56th sess, Supp No 10, UN Doc A/59/10 (6 August 2004) 111 para 3.
 Crawford, The General Part, above n 6, 205 n 216, citing Kjetil Mujezinović Larsen, ‘Attribution of Conduct in Peace Operations: The “Ultimate Authority and Control” Test’ (2008) 19 European Journal of International Law 509, 518.
 Tom Dannenbaum, ‘Public Power and Preventive Responsibility: Attributing the Wrongs of International Joint Ventures’ in André Nollkaemper and Dov Jacobs (eds), Distribution of Responsibilities in International Law (Cambridge University Press, 2015) (forthcoming).
 Articles on the Responsibility of IOs, UN Doc A/66/10, ch V(E) 88  (commentary to art 7).
 André Nollkaemper and Dov Jacobs, ‘Shared Responsibility in International Law: A Conceptual Framework’ (2013) 34 Michigan Journal of International Law 359, 361.
 Ibid 403. Tom Dannenbaum helpfully observes that
whereas the forces under consideration in Nicaragua and Bosnian Genocide had no de jure relationship to any official government — and in particular no official relationship to the United States or the former Federal Republic of Yugoslavia, respectively — peacekeepers have genuine de jure connections to both their home states and the United Nations.
Tom Dannenbaum, ‘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’ (2010) 51 Harvard International Law Journal 113, 155.
 International Law Commission, Responsibility of International Organizations: Comments and Observations Received from International Organizations, UN GAOR, 56th sess, UN Doc A/CN.4/545 (25 June 2004) 17. See also International Law Commission, Responsibility of International Organizations: Comments and Observations Received from International Organizations, UN GAOR, 63rd sess, UN Doc A/CN.4/637/Add.1
(17 February 2011) 7–8.
 Larsen compares the position of Peck (‘The question of who makes the political, strategic, and operational decisions that together comprise the right to command and control United Nations forces is central to determining who is responsible for actions taken by UN Soldiers’) with Shraga (‘[i]n enforcement actions carried out by States under the authorization of the Security Council ... operational command and control is vested in the States conducting the operation, and so is international responsibility for the conduct of their troops’): Larsen, above n 134, 513, quoting Julianne Peck, ‘The UN and the Laws of War: How Can the World’s Peacekeepers be Held Accountable?’ (1995) 21 Syracuse Journal of International Law and Commerce 283, 293; Daphna Shraga, ‘The United Nations as an Actor Bound by International Humanitarian Law’ in Les Nations Unies et le Droit International Humanitaire: Actes du Colloque International à l’Occasion du Cinquantième Anniversaire de l’ONU [The United Nations and International Humanitarian Law: Proceedings of the International Symposium on the Occasion of the Fiftieth Anniversary of the United Nations] (Pedoné, 1996) 317, 330.
 Financing of the United Nations Protection Force, the United Nations Confidence Restoration Operation in Croatia, the United Nations Preventive Deployment Force and the United Nations Peace Forces Headquarters: Administrative and Budgetary Aspects of the Financing of the United Nations Peacekeeping Operation — Financing of the United Nations Peacekeeping Operations — Report of the Secretary-General, UN GAOR, 51st sess, Agenda Items 129 and 140(a), UN Doc A/51/389 (20 September 1996) 6 .
 Responsibility of International Organizations: Comments and Observations Received from International Organizations, UN GAOR, 58th sess, UN Doc A/CN.4/568/Add.1 (12 May 2006) 15 (commenting on secondment arrangements and lending officers between organisations).
 The International Monetary Fund noted, in 2004, that
[t]he issue of attribution of the conduct of peacekeeping forces to the United Nations is specific to that organization and, in the absence of particular proposals on this issue, we do not have any comments. We have reservations, however, about including such a specific issue, which applies to a limited number of organizations, in draft articles that aim at setting forth the principles of responsibility of all international organizations. If any principles or rules applicable to peacekeeping operations are included, the scope of such principles and rules should be explicitly limited to peacekeeping activities and organizations that conduct such activities.
Responsibility of International Organizations: Comments and Observations Received from International Organizations, UN GAOR, 56th sess, UN Doc A/CN.4/545 (25 June 2004) 17 (comments of the International Monetary Fund).
 See, eg, International Law Commission, Responsibility of International Organizations: Comments and Observations Received from International Organizations, UN GAOR, 63rd sess, UN Doc A/CN.4/637 (14 February 2011) 22.
 See ibid 10.
 Pieter Jan Kuijper and Esa Paasivirta, ‘EU International Responsibility and Its Attribution: From the Inside Looking Out’ in Malcolm Evans and Panos Koutrakos (eds), The International Responsibility of the European Union: European and International Perspectives (Hart, 2013) 35, 54:
the ‘organic model’ of attribution is unsatisfactory insofar as the EU’s acts in a ‘vertical mode’, ie the case where the EU acts are carried out via the authorities of its Member States, instead of the EU itself having its own administrative presence in its Member States. This is in fact the EU’s normal practice, which falls basically in the [Treaty on the Functioning of the European Union] context rather than in the [Common Foreign and Security Policy] context. In that situation the ‘organic model’ does not capture the core features of the EU action, since the Member States are seen as remaining sovereign and not constituting organs of the organisation in a formal sense. Consequently, the Member States would always be responsible, as the immediate actors, should acts be attributed on purely organic lines. The fact that normative decisions are taken at the EU level (Council, Parliament) and on the basis of a proposal by an independent institution (Commission) is basically ignored. This is not a satisfactory outcome. It is also out of tune with the fact that the EU is a recognized global actor, alongside states, all across the ‘civil’ areas falling under the TFEU.
 Responsibility of International Organizations: Comments and Observations Received from International Organizations, UN GAOR, 56th sess, UN Doc A/CN.4/545 (25 June 2004) 28.
 Articles on the Responsibility of IOs, UN Doc A/66/10, ch V(E) 87–93 – (commentary to art 7).
 Behrami v France (2007) 45 EHRR SE 10.
 Al-Jedda  1 AC 332.
 Al-Jedda v United Kingdom (2011) 53 EHRR 23 (‘Al-Jedda (ECtHR)’).
 State of the Netherlands v Nuhanović, Hoge Raad [Supreme Court of the Netherlands], 12/03324, 6 September 2013 <http://www.rechtspraak.nl/Organisatie/Hoge-Raad/Over DeHogeRaad/publicaties/Document s/12%2003324.pdf> (‘Nuhanović’).
 State of the Netherlands v Mustafić-Mujic, Hoge Raad [Supreme Court of the Netherlands], 12/03329, 6 September 2013 <http://www.rechtspraak.nl/Organisatie/Hoge-Raad/
 Nuhanović, Hoge Raad [Supreme Court of the Netherlands], 12/03324, 6 September 2013, [3.11.3].
 Ibid [3.10.2] (explaining that the Netherlands is a troop-contributing state that retained control over the personnel affairs of the military personnel, who remained in the service of the Netherlands, and retained the power to punish the military personnel under disciplinary and criminal law). Here, the Court rejected the arguments of the UN and the Netherlands that art 6 was applicable. This was also the position taken by the Procurator-General, Mr Paul Vlas, in his Advisory Opinion: State of the Netherlands v Nuhanović (Advisory Opinion of Procurator-General), Hoge Raad [Supreme Court of the Netherlands], 12/03324, 3 May 2013 <http://www.rechtspraak.nl/Organisatie/Hoge-Raad/OverDe HogeRaad/publicaties/Documents/advisory%20opinion%2012%2003324.pdf> (‘Nuhanović Advisory Opinion’). See André Nollkaemper, ‘Procurator General of the Dutch Supreme Court Concludes to Reject Appeal against Srebrenica Judgment’ on SHARES
(3 May 2013) <http://www.sharesproject.nl/procurator-general-of-the-dutch-supreme-cour
 Nuhanović, Hoge Raad [Supreme Court of the Netherlands], 12/03324, 6 September 2013, [3.11.3].
 Ibid [3.12.3].
 See Milanović, ‘State Responsibility for Acts of Non-State Actors’, above n 73, 315.
 Nuhanović, Hoge Raad [Supreme Court of the Netherlands], 12/03324, 6 September 2013, [3.11.2]. This aspect of the decision differs from the Nuhanović Advisory Opinion, in which the Procurator-General argued that art 7 of Articles on the Responsibility of IOs does not permit dual attribution: Nuhanović Advisory Opinion (Hoge Raad [Supreme Court of the Netherlands], 12/03324, 3 May 2013) [4.13]–[4.14]. As Nollkaemper explains:
[the Procurator-General] cites Crawford and Olleson for the proposition that the purpose of Article 7 ARIO ‘is not to determine whether particular conduct is attributable as such, but rather it addresses the question of to which of two entities (the “borrowing” international organization or the “lending” State (or international organization), the conduct is to be attributed’.
Nollkaemper, ‘Procurator General of the Dutch Supreme Court Concludes to Reject Appeal against Srebrenica Judgment’, above n 156, citing Nuhanović Advisory Opinion, Hoge Raad [Supreme Court of the Netherlands], 12/03324, 3 May 2013 [4.12] (emphasis in original). The Prosecutor-General thus construes art 7 as a ground for exclusive responsibility, not allowing for ‘independent responsibility’ of the troop-contributing state: at [4.13].
 See, eg, Rahmatullah v Secretary of State for Defence  UKSC 48 (UK forces had apprehended Mr Ramatullah in 2004 in Iraq and handed him over to the US who held him in indefinite detention. Pursuant to a Memorandum of Understanding, his lawyers argued that the UK had the power to request his release from the US under the writ of Habeas Corpus. Finding that the UK did not need to have actual custody to exercise control, the transfer of Mr Rahmatullah to Pakistan was deemed a violation of the Fourth Geneva Convention). The UK Supreme Court noted that
[t]here can be no plausible argument, therefore, against the proposition that there is clear prima facie evidence that Mr Rahmatullah is unlawfully detained and that the UK government was under an obligation to seek his return unless it could bring about effective measures to correct the breaches of [Fourth Geneva Convention] that his continued detention constituted. It is for that reason that I am of the view that the real issue in this case is that of control: at 635 .
 Nuhanović, Hoge Raad [Supreme Court of the Netherlands], 12/03324, 6 September 2013, [3.11.3].
 Bérénice Boutin, ‘Responsibility of the Netherlands for the Acts of Dutchbat in Nuhanović and Mustafić: The Continuous Quest for a Tangible Meaning for “Effective Control” in the Context of Peacekeeping’ (2012) 25 Leiden Journal of International Law 521, 528 (explaining that the UN is reluctant to admit the possibility that it does not control troops).
 See Al-Jedda (ECtHR) (2011) 53 EHRR 23.
 Ibid 42 .
 Ibid 49 –.
 Al-Jedda  1 AC 332, 338 . See also Al-Jedda (ECtHR) (2011) 53 EHRR 23, 29 , quoting Al-Jedda  1 AC 332, 348  (Lord Bingham):
a number of questions must be asked in the present case. Were UK forces placed at the disposal of the UN? Did the UN exercise effective control over the conduct of UK forces? Is the specific conduct of the UK forces in detaining the appellant to be attributed to the UN rather than the UK? Did the UN have effective command and control over the conduct of UK forces when they detained the appellant? Were the UK forces part of a UN peacekeeping force in Iraq? In my opinion the answer to all these questions is in the negative.
 Al-Jedda (ECtHR) (2011) 53 EHRR 23, 61 .
 Ibid 48 :
It is difficult to conceive that the applicant’s detention was attributable to the United Nations and not to the United Kingdom when UN organs, operating under the mandate of Resolution 1546, did not appear to approve of the practice of indefinite internment without trial and, in the case of UNAMI, entered into correspondence with the US Embassy in an attempt to persuade the Multi-National Force under American command to modify the internment procedure.
 Ibid 47 :
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.
 Behrami v France (2007) 45 EHRR SE 10. Note, however, that the ECtHR in Al-Jedda (ECtHR) avoided overruling this decision, and instead distinguished the UN presence in Kosovo from Iraq on the facts: Al-Jedda (ECtHR) (2011) 53 EHRR 23, 60 .
 Behrami v France (2007) 45 EHRR SE 10, 118 . See Aurel Sari, ‘Autonomy, Attribution and Accountability: Reflections on the Behrami Case’ in Richard Collins and Nigel D White (eds), International Organizations and the Idea of Autonomy: Institutional Independence in the International Legal Order (Routledge, 2011) 257, 264:
While the Security Council might have retained such ‘ultimate authority and control’ over the international security presence as was necessary to render the delegation of its powers lawful under the UN Charter, the question the Court should have asked itself is whether or not the Security Council exercised such control over KFOR as was sufficient to render the delegation of its powers lawful under the UN Charter, the question the Court should have asked itself is whether or not the Security Council exercised such control over KFOR as was sufficient to render the conduct of KFOR attributable to the UN in accordance with the law of international responsibility. The widely held view is that the necessary level of control required in this context is that of ‘effective control’ ... [and] the Security Council lacked
both the practical means and the legal authority to exercise such a degree of control
over ... KFOR.
 Behrami v France (2007) 45 EHRR SE 10, 119 , 120 .
 Pasquale De Sena and Maria Chiara Vitucci, ‘The European Courts and the Security Council: Between Dedoublement Fonctionnel and Balancing of Values’ (2009) 20 European Journal of International Law 193, 2029; Larsen, above n 134.
 Al-Jedda (ECtHR) (2011) 53 EHRR 23, 60 :
For the reasons set out above, the Court considers that the UN Security Council had neither effective control nor ultimate authority and control over the acts and omissions of troops within the Multi-National Force and that the applicant’s detention was not, therefore, attributable to the United Nations.
 See Sari, ‘Autonomy, Attribution and Accountability’, above n 173, 267.
 Behrami v France (2007) 45 EHRR SE 10, 104 .
 Marko Milanović, however, suggests control over territory seems less important with regard to surveillance and drones: Marko Milanovic, ‘Foreign Surveillance and Human Rights, Part 3: Models of Extraterritorial Application’ on European Journal of International Law, EJIL: Talk! (27 November 2013) <http://www.ejiltalk.org/foreign-surveillance-an
 See Articles on State Responsibility, UN Doc A/56/10, ch IV(E) art 2 (providing that ‘there is an internationally wrongful act of a State when conduct consisting of an action or omission: (a) Is attributable to the State under international law; and (b) Constitutes a breach of an international obligation of the State’). The identical provision is contained in art 4 of the Articles on the Responsibility of IOs, UN Doc A/66/10, ch V(E). See also Corfu Channel Case (United Kingdom v Albania) (Merits)  ICJ Rep 4 (‘Corfu Channel Case’) (nothing was attempted by the Albanian authorities to prevent the disaster and thus those grave omissions involved the international responsibility of Albania).
 The paradigmatic example of liability resulting from the failure to act is the Corfu Channel Case, where the ICJ determined that the failure of the Albanian authorities to take measures to prevent British vessels from sailing into mines in the sea was a grave omission that created international responsibility for Albania: Corfu Channel Case  ICJ Rep 4. See also Franck Latty, ‘Actions and Omissions’ in James Crawford, Alain Pellet and Simon Olleson (eds), The Law of International Responsibility (Oxford University Press, 2010) 355, 358 (defining omission as ‘an abstention consisting of the fact of not doing that which ought to be done’).
 Crawford, The General Part, above n 6, 218.
 Daniel Bodansky and John R Crook, ‘Symposium: The ILC’s State Responsibility
Articles — Introduction and Overview’ (2002) 96 American Journal of International Law 773, 783.
 See Bosnian Genocide  ICJ Rep 43, 232 .
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) 81  (commentary to art 8):
A receiving State is not responsible, as such, for the acts of private individuals in seizing an embassy, but it will be responsible if it fails to take all necessary steps to protect the embassy from seizure, or to regain control over it. In this respect there is often a close link between the basis of attribution and the particular obligation said to have been breached, even though the two elements are analytically distinct (citations omitted)
(discussing the Tehran Hostages case, where the ICJ found that the initial phase of the attacks upon the embassy were not attributable to Iran since the attackers were private individuals acting on their own, but nonetheless noted that the state will be responsible if it fails to take all necessary steps to protect the embassy from seizure or to regain control over it: Case concerning United States Diplomatic and Consular Staff in Tehran (United States of America v Iran) (Judgment)  ICJ Rep 3).
 See discussion in Crawford, The General Part, above n 6, 158.
 Resolution 1373, UN Doc S/RES/1373, .
 Crawford, The General Part, above n 6, 161.
 White, ‘Due Diligence Obligations’, above n 7, 234 (discussing in detail the Montreux Document and the draft convention on an International Code of Conduct).
 Convention on the Prevention and Punishment of the Crime of Genocide, opened for signature 9 December 1948, 78 UNTS 277 (entered into force 12 January 1951) art 1 (‘Genocide Convention’): ‘The Contracting Parties confirm that genocide, whether committed in time of peace or in time of war, is a crime under international law which they undertake to prevent and to punish’.
 Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment, opened for signature 10 December 1984, 1465 UNTS 85 (entered into force 26 June 1987) art 2(1): ‘Each State Party shall take effective legislative, administrative, judicial or other measures to prevent acts of torture in any territory under its jurisdiction’.
 International Convention for the Suppression of Terrorist Bombings, opened for signature 15 December 1997, 2149 UNTS 256 (entered into force 23 May 2001) art 15: ‘States Parties shall cooperate in the prevention of the offences set forth in article 2’.
 See, eg, the duty of non-refoulement, whereby states must not return persons to their home country if that will expose them to a risk of human rights violations: Cordula Droege, ‘Transfers of Detainees: Legal Framework, Non-Refoulement and Contemporary Challenges’ (2008) 90 International Review of the Red Cross 669, 670; See also Human Rights Committee, General Comment No 31 : The Nature of the General Legal Obligation Imposed on States Parties to the Covenant, UN GAOR, 80th sess, UN Doc CCPR/C/21/Rev.1/Add.13 (26 May 2004) . For an extensive discussion of due diligence obligations in the human rights context, see Duncan French and Tim Stephens, ‘ILA Study Group on Due Diligence in International Law’ (First Report, International Law Association, 7 March 2014) 16.
 Michael Duttwiler, ‘Liability for Omission in International Criminal Law’ (2006) 6 International Criminal Law Review 1, 2. See also discussion below in Part VI(B).
 See generally Pulp Mills on the River Uruguay (Argentina v Uruguay) (Judgment)  ICJ Rep 14. See also Riccardo Pisillo-Mazzeschi, ‘The Due Diligence Rule and the Nature of the International Responsibility of States’ (1992) 35 German Yearbook of International Law 9, 38.
 See International Law Commission, International Liability for Injurious Consequences Arising Out of Acts not Prohibited by International Law (9 July 2014) <http://legal.un.org/ilc/texts/9.htm> .
 Pisillo-Mazzeschi, above n 195 : arguing that
[a]n analysis of international practice shows that the due diligence rule has been applied in the areas of customary international law concerning: a) the security of aliens and representatives of foreign States; b) the security of foreign States; and c) the conservation of the environment.
 See, eg, Resolution 1373, UN Doc S/RES/1373.
 See, eg, International Code of Conduct Association, ‘International Code of Conduct for Private Security Service Providers’ (Code of Conduct, 9 November 2010).
 James Crawford, The International Law Commission’s Articles on State Responsibility: Introduction, Text and Commentaries (Cambridge University Press, 2002) 82.
 This can be determined by a jurisdictional clause in a treaty, such as the Convention on the Prevention of the Crime of Genocide, which provides that ‘[t]he Contracting Parties confirm that genocide, whether committed in time of peace or in time of war, is a crime under international law which they undertake to prevent and to punish’: Prevention of the Crime of Genocide, opened for signature 9 December 1948, 78 UNTS 277 (entered into force 12 January 1951) art 1. This arguably creates a general duty to prevent genocide. Absent a jurisdictional clause, the spatial effective control test is usually employed to determine the extraterritorial effect of treaties.
 See, eg, Giorgio Gaja, ‘The Relations between the European Union and Its Member States from the Perspective of the ILC Articles on Responsibility of International Organizations’ (SHARES Research Paper No 25, Amsterdam Centre for International Law, 2013) 2  <http://www.sharesproject.nl/publication/the-relations-between-the-european-union-an d-its-member-states-from-the-perspective-of-the-ilc-articles-on-responsibility-of-internat ional-organizations/> (discussing the EU’s obligation to prevent in: Conservation and Sustainable Exploitation of Swordfish Stocks in the South-Eastern Pacific Ocean (Chile v European Union) (Order) (International Tribunal for the Law of the Sea, Case No 7,
16 December 2009)).
 Pisillo-Mazzeschi, above n 195, 26.
 Marko Milanović, ‘State Responsibility for Genocide’ (2006) 17 European Journal of International Law 553, 568 (discussing the duty to punish in the context of the Genocide Convention, and noting it includes a duty to criminalise the act in national law, and prosecute or extradite).
 Bosnian Genocide  ICJ Rep 43, 221 .
 Gaja, above n 202, 2 .
 French and Stephens, above n 193.
 Pisillo-Mazzeschi, above n 195, 44.
 Hessbreugge, above n 9, 268 (noting a true obligation of due diligence would be breached by a failure to engage in the specific conduct even if the result did not occur, whereas a duty to prevent requires the event to occur).
 Pisillo-Mazzeschi, above n 195, 26, citing Walter A Noyes (United States of America v Panama) (Decision) (1933) 6 RIAA 308, 311.
 See, eg, Responsibilities and Obligations of States Sponsoring Persons and Entities with respect to Activities in the Area (Advisory Opinion)  ITLOS Reports 10 (‘Seabed Advisory Opinion’), citing United Nations Convention on the Law of the Sea, opened for signature 10 December 1982, 1833 UNTS 3 (entered into force 16 November 1994) art 153(4). However, the Tribunal notes (at 43 ) that
‘due diligence’ is a variable concept. It may change over time as measures considered sufficiently diligent at a certain moment may become not diligent enough in light, for instance, of new scientific or technological knowledge. It may also change in relation to the risks involved in the activity.
 Condorelli, ‘The Imputability to States’, above n 113, 240.
 French and Stephens, above n 193.
 Pisillo-Mazzeschi above n 195, 46 (noting diligence does not constitute an element present in all international obligations of the state and is not a general criterion of international responsibility).
 Chinkin, above n 90, 393–4.
 See Eyal Mayroz, ‘The Legal Duty to “Prevent”: After the Onset of “Genocide”’ (2012) 14 Journal of Genocide Research 79, 83–4, 86 (arguing that a state that has means which are likely to have a deterrent effect on would-be perpetrators is under the duty to make use of them ‘as the circumstances permit”’: at 84 (emphasis in original). Nonetheless, the duty to prevent genocide does not include the duty to intervene).
 In Bosnian Genocide, the ICJ wrote that ‘the obligation of States parties is rather to employ all means reasonably available to them, so as to prevent genocide so far as possible’. The ICJ also noted that responsibility is incurred only if the state manifestly fails to take measures to prevent genocide: Bosnian Genocide  ICJ Rep 43, 221 .
 Milanović, ‘State Responsibility for Genocide’, above n 204, 553.
 As Nollkaemper explains, ‘the removal of Nuhanovic and Mustafic from the compound could be attributed to the Netherlands, if the Netherlands was able to prevent that removal’: André Nollkaemper, ‘Dual Attribution: Liability of the Netherlands for Removal of Individuals from the Compound of Dutchbat’ on SHARES (8 July 2011) <http://www.sharesproject.nl/dual-attribution-liability-of-the-netherlands-for-removal-o f-individuals-from-the-compound-of-dutchbat/> (emphasis in original).
 See Boutin, above n 163, 529.
 Ibid 531 (arguing that the duty to prevent includes an element of cause-and-effect in that the evacuation of the victims was related to Dutchbat’s control over the evacuation).
 See Nuhanović, Hoge Raad [Supreme Court of the Netherlands], 12/03324, 6 September 2013, [3.12.2].
 Ibid [3.12.3].
 Prosecutor v Halilović (Judgement) (International Criminal Tribunal for the Former Yugoslavia, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) 23  (‘Halilović’).
 Beatrice I Bonafé, ‘Finding a Proper Role for Command Responsibility’ (2007) 5 Journal of International Criminal Justice 599, 600 (noting that large-scale atrocities are often carried out by a variety of perpetrators, control theories are one of the primary ways in which ICL can reach the leaders or the masterminds behind the scenes).
 Harmen van der Wilt, Command Responsibility (30 September 2013) Oxford Bibliographies <http://www.oxfordbibliographies.com/view/document/obo-978019979 6953/obo-9780199796953-0088.xml> .
 See, eg, Agreement between the United Nations and the Government of Sierra Leone on the Establishment of a Special Court for Sierra Leone, signed 16 January 2002, 2178 UNTS 137 (entered into force 12 April 2002) art 6(3) (‘Statute of the Special Court for Sierra Leone’): providing that
[t]he fact that any of the acts referred to in articles 2 to 4 of the present Statute was committed by a subordinate does not relieve his or her superior of criminal responsibility if he or she knew or had reason to know that the subordinate was about to commit such acts or had done so and the superior had failed to take the necessary and reasonable measures to prevent such acts or to punish the perpetrators thereof.
See also Law on the Establishment of Extraordinary Chambers in the Courts of Cambodia for the Prosecution of Crimes Committed During the Period of Democratic Kampuchea 2004 (Cambodia) art 29 (‘Statute of the ECCC’) (providing that ‘the fact that the
[acts] ... were committed by a subordinate does not relieve the superior of personal criminal responsibility if the superior had effective command and control or authority and control over the subordinate’). Statutes of other international criminal law tribunals include similarly-worded articles: see SC Res 827, UN SCOR, 48th sess, 3217th mtg, UN Doc S/RES/827 (25 May 1993) art 7(3) (‘Statute of the International Criminal Tribunal for the Former Yugoslavia’); SC Res 955, UN SCOR, 49th sess, 3453rd mtg, UN Doc S/RES/955 (8 November 1994) annex art 6(3) (‘Statute of the International Criminal Tribunal for Rwanda’); SC Res 1757, UN SCOR, 62nd sess, 5685th mtg, UN Doc S/RES/1757 (30 May 2007) annex art 3(3) (‘Agreement between the United Nations and the Lebanese Republic on the Establishment of a Special Tribunal for Lebanon’). See also Rome Statute of the International Criminal Court, opened for signature 17 July 1998, 2187 UNTS 90 (entered into force 1 July 2002) art 28 (‘Rome Statute’):
A military commander or person effectively acting as a military commander shall be criminally responsible for crimes within the jurisdiction of the Court committed by forces under his or her effective command and control, or effective authority and control as the case may be, as a result of his failure to exercise control properly over such forces, where: ... (ii) That military commander or person failed to take all necessary and reasonable measures within his or her power to prevent or repress their commission or to submit the matter to the competent authorities for investigation and prosecution.
A difference between the Rome Statute and the statutes for the ad hoc tribunals is that the former requires causality between the superior’s dereliction of duties and the commission of crimes: Prosecutor v Bemba (Decision pursuant to Article 61(7)(a) and (b) of the Rome Statute on the Charges of the Prosecutor) (International Criminal Court,
Pre-Trial Chamber II, Case No ICC-01/05-01/08, 15 June 2009)  (‘Bemba’); cf Prosecutor v Delalić (Judgement) (International Criminal Tribunal for the Former Yugoslavia, Trial Chamber, Case No IT-96-21-T, 16 November 1998)  (‘Delalić’).
 Prosecutor v Gotovina (Judgement) (International Criminal Tribunal for the Former Yugoslavia, Appeals Chamber, Case No IT-06-90-A, 16 November 2012) – (‘Gotovina’), citing Prosecutor v Orić (Judgement) (International Criminal Tribunal for the Former Yugoslavia, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) (‘Orić’).
 Cf Statute of the ECCC.
 Bemba (International Criminal Court, Pre-Trial Chamber II, Case No ICC-01/05-01/08,
15 June 2009) .
 See, eg, Delalić (International Criminal Tribunal for the Former Yugoslavia, Trial Chamber, Case No IT-96-21-T, 16 November 1998) :
whereas formal appointment is an important aspect of the exercise of command authority or superior authority, the actual exercise of authority in the absence of a formal appointment is sufficient for the purpose of incurring criminal responsibility. Accordingly, the factor critical to the exercise of command responsibility is the actual possession, or non-possession, of powers of control over the actions of the subordinates.
 The superior–subordinate relationship encompasses both military and civilian spheres, and the statutes of the ad hoc tribunals do not distinguish between the two. The Rome Statute does make a distinction, however, by introducing a higher threshold for civilian superiors, and integrating a nexus requirement. Article 28(b) provides for criminal responsibility under the following circumstances:
(i) The superior either knew, or consciously disregarded information which clearly indicated, that the subordinates were committing or about to commit such crimes; (ii) The crimes concerned activities that were within the effective responsibility and control of the superior; and (iii) The superior failed to take all necessary and reasonable measures within his or her power to prevent or repress their commission or to submit the matter to the competent authorities for investigation and prosecution.
 See Harmen van der Wilt, ‘Command Responsibility in the Jungle: Some Reflections on the Elements of Effective Command and Control’ in Charles Chernor Jalloh (ed), The Sierra Leone Special Court and Its Legacy: The Impact for Africa and International Criminal Law (Cambridge University Press, 2014) 144, 145 (emphasis in original):
the doctrine of superior responsibility ... [is] strongly wedded to the concept of military hierarchy. Its crucial element, effective control, presupposes a streamlined organization, with adequate channels of information, clear-cut patterns of hierarchy, and unquestioned obedience, which functions even (or should we rather say: in particular?) in the heat of fighting.
 Prosecutor v Brima (Judgement) (Special Court for Sierra Leone, Trial Chamber II, Case No SCSL-04-16-T, 20 June 2007) .
 van der Wilt, ‘Command Responsibility’, above n 235, 155.
 Ibid 149–50.
 Ibid 150–1.
 Ibid 153 (finding that Sesay exercised effective control over subordinates in the Revolutionary United Front ).
 Orić (International Criminal Tribunal for the Former Yugoslavia, Trial Chamber II, Case No IT-03-68-T, 30 June 2006) .
 Halilović (International Criminal Tribunal for the Former Yugoslavia, Trial Chamber I, Case No IT-01-48-T, 16 November 2005) –.
 In Bemba, the International Criminal Court stated that effective control is the ‘material ability’ to prevent, repress or submit the matter to the competent authorities: Bemba (International Criminal Court, Pre-Trial Chamber II, Case No ICC-01/05–01/08, 15 June 2009) . Note that unlike ad hoc tribunals, art 28 of the Rome Statute requires causality between dereliction of duties and the underlying crimes.
 Gotovina (International Criminal Tribunal for the Former Yugoslavia, Appeals Chamber, Case No IT-06-90-A, 16 November 2012) .
 The court also noted that there was ample evidence on the record that he promoted discipline against troops under his command: ibid .
 See, eg, Halilović (International Criminal Tribunal for the Former Yugoslavia, Trial Chamber, Case No IT-01-48-T, 16 November 2005).
 Dannenbaum’s illustrations of levers of controls include the powers to discipline and punish, hire and promote and train: Dannenbaum, ‘Translating the Standard’, above n 138, 164.
 See, eg, Abby Cohen Smutny, ‘State Responsibility and Attribution: When is a State Responsible for the Acts of State Enterprises? Emilio Agustín Maffezini v The Kingdom of Spain’ in Todd Weiler (ed), International Investment Law and Arbitration: Leading Cases from the ICSID, NAFTA, Bilateral Treaties and Customary International Law (Cameron May, 2005) 17, 17 (explaining that in regards to claims against states under investment treaties that ‘rules of attribution ... often constitute a critical aspect of the dispute. When a foreign investor has suffered losses, the question of whether the acts or omissions alleged to have caused the losses may be attributable to the State is assessed at the threshold’).
 Convention on the Settlement of Investment Disputes between States and Nationals of Other States, opened for signature 18 March 1965, 575 UNTS 159 (entered into force 14 October 1966) art 25, which provides:
(1) The jurisdiction of the Centre shall extend to any legal dispute arising directly out of an investment, between a Contracting State (or any constituent subdivision or agency of a Contracting State designated to the Centre by that State) and a national of another Contracting State, which the parties to the dispute consent in writing to submit to the Centre. When the parties have given their consent, no party may withdraw its consent unilaterally.
(2) ‘National of another Contracting State’ means:
(a) any natural person who had the nationality of a Contracting State other than the State party to the dispute on the date on which the parties consented to submit such dispute to conciliation or arbitration as well as on the date on which the request was registered pursuant to paragraph (3) of Article 28 or paragraph (3) of Article 36, but does not include any person who on either date also had the nationality of the Contracting State party to the dispute; and
(b) any juridical person which had the nationality of a Contracting State other than the State party to the dispute on the date on which the parties consented to submit such dispute to conciliation or arbitration and any juridical person which had the nationality of the Contracting State party to the dispute on that date and which, because of foreign control, the parties have agreed should be treated as a national of another Contracting State for the purposes of this Convention.
 Mark Feldman, ‘The Standing of State-Owned Entities under Investment Treaties’ in Karl P Sauvant (ed), Yearbook on International Investment Law & Policy 2010–2011 (Oxford University Press, 2012) 615, 624 (advocating a nature and purpose test to curb this eventuality).
 See, eg, Behrami v France (2007) 45 EHRR SE 10 (determining that wrongful acts in Kosovo were attributable to the UN on the basis of an ultimate control test, and that the Court had no jurisdiction ratione personae over the UN).
 Nuhanović, Hoge Raad [Supreme Court of the Netherlands], 12/03324, 6 September 2013, 22–3 [3.11.2].
 Hessbreugge, above n 9, 303.
 See Articles on State Responsibility, UN Doc A/56/10, ch IV(E) 92–5 – (commentary to art 5).
 See, eg, International Law Commission, Responsibility of International Organizations: Comments and Observations Received from International Organizations, UN GAOR, 63rd sess, UN Doc A/CN.4/637/Add.1 (17 February 2011) 9–14.
 Trapp, above n 106, 61.
 Hessbreugge, above n 9, 306.
 See, eg, Steven R Ratner, ‘Corporations and Human Rights: A Theory of Legal Responsibility’  YaleLawJl 32; (2001) 111 Yale Law Journal 443, 449–52.
 See Crawford and Watkins, ‘International Responsibility’, above n 1, 283, 288.
 H Lauterpacht, ‘Recognition of States in International Law’ (1944) 53 Yale Law Journal 385 (discussing how states act as organs of international law).
 Articles on State Responsibility, UN Doc A/56/10, ch IV(E) 80 :
In theory, the conduct of all human beings, corporations or collectivities linked to the State by nationality, habitual residence or incorporation might be attributed to the State, whether or not they have any connection to the government. In international law, such an approach is avoided, both with a view to limiting responsibility to conduct which engages the State as an organization, and also so as to recognize the autonomy of persons acting on their own account and not at the instigation of a public authority. Thus the general rule is that the only conduct attributed to the State at the international level is that of its organs of government, or of others who have acted under the direction, instigation or control of those organs, ie, as agents of the State.
 See Articles on State Responsibility, UN Doc A/56/10, ch IV(E) 104  (‘Bearing in mind the important role played by the principle of effectiveness in international law, it is necessary to take into account in both cases the existence of a real link between the person or group performing the act and the State machinery’). See generally Hiroshi Taki, ‘Effectiveness’ (February 2013) Max Planck Encyclopedia of Public International Law.
 Seabed Advisory Opinion  ITLOS Reports 10, , citing Articles on State Responsibility, UN Doc A/56/10, ch IV(E) 103  (commentary to art 8).
 See generally Amitai Etzioni, ‘Sovereignty as Responsibility’ (2006) 50 Orbis 71.
 Hessbreugge, above n 9, 306.
 See Luigi Condorelli and Claus Kress, ‘The Rules of Attribution: General Considerations’ in James Crawford, Alain Pellet and Simon Olleson (eds), The Law of International Responsibility (Oxford University Press, 2010) 221–3.
 Smutny, above n 248, 33.
 On this debate, see Steffen Wirth, ‘Immunity for Core Crimes? The ICJ’s Judgment in the Congo v Belgium Case’ (2002) 13 European Journal of International Law 877; Marina Spinedi, ‘State Responsibility v Individual Responsibility for International Crimes: Tertium Non Datur?’ (2002) 13 European Journal of International Law 895.