2.1. The existence of DNSA obligations

Traditionally, international law was concerned with the behaviour of, and relationship between States. The regulation of natural and legal persons has typically been considered an issue for domestic law. As one leading textbook puts it:

There is now, however, a multiplicity of instruments on the international plane which purport to set standards and create obligations for persons traditionally regulated by domestic law. Further, recent years have seen an increase in the popularity of such instruments; where individual States are unwilling or unable to effectively regulate DNSAs at the domestic level, the international sphere is increasingly considered a viable alternative.

As a preliminary matter, it is important to recognise that the efforts to regulate DNSAs come in several forms. Existing literature which focuses on broader notions of accountability rather than the particular conceptual problem being discussed here often does not distinguish between different types of instruments or obligations. The focus of this article, however, makes drawing such distinctions necessary. Particularly helpful in this respect is the definition adopted by Markos Karavias in his work on corporate obligations in international law. Karavias’ account explains that it is only possible to speak of a binding international obligation where three conditions have been met.Footnote ⁸

First, the relevant obligation must exist in international law and apply directly to the conduct of corporations (or other DNSAs).Footnote ⁹ Obligations must therefore arise from treaties, rules of custom or general principles of law. So-called ‘indirect’ obligations which impose obligations on States to regulate the activities of DNSAs in certain ways are therefore excluded from this definition. Whilst indirect obligations are perhaps one of the most effective ways for international law to regulate natural and legal persons, they do not hold any conceptual mysteries: the obligation of the DNSA is mediated through the State and exists at the level of domestic rather than international law. Put differently, a failure to comply with the relevant obligation will give rise to liability at the domestic but not the international level. Should the State fall short of fully implementing or achieving what international law requires, it is the State rather than the relevant natural or legal person that is internationally liable. There are several instruments which operate in this manner, such as the draft treaty to ‘regulate, in international human rights law, the activities of transnational corporations and other business enterprises’ (being negotiated at the time of writing under the auspices of the United Nations (UN) Human Rights Council) which is addressed exclusively to States.Footnote ¹⁰

Second, a breach of the primary obligation must give rise to responsibility, and hence secondary obligations, of the DNSA in international law.Footnote ¹¹ As Anne Peters explains, in the absence of responsibility and secondary obligations, fulfilment of primary obligations is merely a matter of goodwill rather than a legal duty.Footnote ¹² Whilst clearly related to the first condition, this ‘responsibility’ condition provides a way to exclude certain international instruments which are addressed to DNSAs but nevertheless do not impose international obligations upon those actors. The biggest category of such instruments is non-binding instruments, such as Corporate Social Responsibility (CSR) codes or standards which purport to set out the best practice for responsible business conduct in a range of areas. Particularly prominent examples include the Organisation for Economic Co-operation and Development Guidelines for Multinational Enterprises and the UN Global Compact.Footnote ¹³ It would be wrong to casually dismiss CSR standards as merely wishful thinking; such standards have an important role to play in managing the potentially harmful effects which may result from the activities of multinationals. The impact of CSR standards is increasingly clear in international investment law, as some investment treaties make investment protection conditional on compliance with CSR standards.Footnote ¹⁴ More generally, ‘soft law’ instruments can, over time, lead to the emergence of hard-edged rules of international law.Footnote ¹⁵ Yet, as the language of ‘non-binding’ suggests, without more, businesses are not legally required to comply with such standards, and a failure to comply does not lead to any international liability. The problem of obligation without consent therefore does not arise.

The third and final condition, which flows from the previous two, is that the implementation of a DNSA’s international obligations must take place on the basis of international rather than domestic law.Footnote ¹⁶ In practical terms, this means that any court or tribunal dealing with a case based on the alleged violation of international obligations by a DNSA must be able to apply international law. It is important to clarify that the absence of a mechanism to adjudicate upon and enforce a particular obligation does not mean that the relevant obligation does not exist or that it is not ‘legal’ in nature. What is, however, important is the potential to establish mechanisms for determining the compliance of DNSAs with their international obligations.

International courts and tribunals are the most obvious institutions to consider here. It is worth noting that not all international courts and tribunals are competent to consider the liability of DNSAs. Most obviously, the jurisdiction of the International Court of Justice (ICJ) is limited to inter-State disputes so it cannot adjudicate on any internationally wrongful acts committed by DNSAs.Footnote ¹⁷ Nevertheless, it is now well-established that it is possible for international courts and tribunals to consider the liability of DNSAs. The International Criminal Court is perhaps the best-known example of such an institution, regularly trying individuals for international crimes. Other notable international criminal tribunals include the International Criminal Tribunal for the Former Yugoslavia and the International Criminal Tribunal for Rwanda. Outside of the context of international criminal law, there have also been recent suggestions that under certain treaties international investment tribunals have jurisdiction to adjudicate on the international liability of DNSAs.Footnote ¹⁸

Domestic courts may also have the capacity to apply international law to determine the international liability of DNSAs.Footnote ¹⁹ Whilst care must be taken to properly identify the source of the relevant obligation, the fact that an obligation is being applied by a domestic rather than international court does not affect its status as an international legal obligation. Indeed, where domestic courts are empowered to apply international law, bringing a case before the domestic courts can often be a more effective means of enforcing international obligations against DNSAs than attempting to bring a claim before an international court. This is commonly the case where domestic courts exercise universal jurisdiction to determine liability under international law or adopt specific legislation allowing defendants to be sued for breaches of certain international obligations in domestic courts, such as the Alien Tort Statute (ATS) in the United States (US).Footnote ²⁰

On Karavias’ definition, there are several examples of international obligations which bind DNSAs. Perhaps the best-known example is international criminal law. Whilst the language of international crimes extends as far back to the law dealing with international piracy in the time of Hugo Grotius, the nature of international criminal responsibility was only put on a firm footing following the trials of the International Military Tribunal (IMT) at Nuremberg. On the basis that ‘[c]rimes against international law are committed by men, not by abstract entities’ the IMT held that international criminal law imposes international duties and liabilities directly on individuals without the interposition of domestic law.Footnote ²¹ This position was subsequently supported by both the International Law Commission (ILC) and the Sixth Committee of the UN General Assembly (UNGA).Footnote ²² Although the source of international criminal law is generally taken to be customary international law rather than treaty law,Footnote ²³ that such obligations directly bind individuals is also reflected in various treaties.Footnote ²⁴ The result is that individuals can be, and have been, held liable for a range of international crimes, including genocide, crimes against humanity and war crimes.

Similarly uncontroversial is the applicability of certain rules of international humanitarian law to non-State armed groups. Such groups typically have no formal legal status under domestic law as they are considered or characterised by the State as being illegitimate and amounting to nothing more than criminal organisations. Nevertheless, international humanitarian law is clear in recognising the existence of such groups and imposing certain obligations upon them. The best-known examples are Common Article 3 and Additional Protocol II of the Geneva Conventions. Common Article 3 provides that ‘[i]n the case of armed conflict not of an international character occurring in the territory of the High Contracting Parties, each Party the conflict shall be bound to apply as a minimum the following provisions’, before going on to list several obligations.Footnote ²⁵ A number of international courts, including the ICJ in Military and Paramilitary Activities in and against Nicaragua, have confirmed that the reference to ‘each Party to the conflict’ encompasses non-State armed groups.Footnote ²⁶ Whilst such language cannot be found in Additional Protocol II, the obligations contained therein have nevertheless been considered to apply to non-State armed groups as well as the States Parties to the Conventions.Footnote ²⁷ The International Committee of the Red Cross has suggested that other principles of the Geneva Conventions and also of Additional Protocol I have acquired the status of customary international law which also bind ‘parties to the conflict’.Footnote ²⁸ Other notable examples include Amended Protocol II to the Conventional Weapons Convention, and Article 19(1) and the Second Protocol to the Cultural Property Convention.Footnote ²⁹

The UN Security Council (UNSC) has also issued several resolutions addressing DNSAs. In a recent study on the UNSC, Michael Wood and Eran Sthoeger note that the UNSC regularly addresses non-State actors in terms which purport to be legally binding, and that it has long done so.Footnote ³⁰ Recent examples include Resolution 2178 which demanded ‘that all foreign terrorist fighters disarm and cease all terrorist acts and participation in armed conflict’.Footnote ³¹ In the Kosovo Advisory Opinion, the ICJ appeared to accept, or at least did not dissent from, the view that UNSC resolutions can be legally binding on DNSAs when it considered whether a declaration of independence was a breach of UNSC Resolution 1244.Footnote ³²

Outside of these clear-cut examples, there are also areas where the existence of DNSA obligations is less certain. One such area is international human rights law. The orthodox position is that only States can have international human rights obligations. In recent years, however, there has been a notable movement towards the view that DNSAs should be regarded as having at least some international human rights obligations.Footnote ³³ The argument for DNSA obligations has been made in relation to a range of human rights treaties, but the African Charter on Human and Peoples’ Rights is perhaps the strongest candidate.Footnote ³⁴ Whilst the majority of the treaty is phrased in general terms, Articles 27–29 contain ‘duties’ addressed to individuals concerning, inter alia, non-discrimination, national security and cultural values.Footnote ³⁵ These provisions have not, however, received much attention from either the African Commission or the African Court on Human and Peoples’ Rights, leading to continuing controversy about the content and nature of these obligations.Footnote ³⁶ Others argue that even if DNSAs do not currently have international human rights obligations, there should be a move towards creating such obligations.Footnote ³⁷

At the domestic level, the jurisprudence of the Canadian and US Supreme Courts offers particularly useful insights. In Nevsun Resources Ltd v Araya, the Canadian Supreme Court was dealing with the question of whether customary international law obligations were capable of binding corporations (in that case, a mining company). In her leading judgment, Justice Abella relied on a range of academic commentary to find that rules of customary international law are capable of binding corporations.Footnote ³⁸ However, neither the court nor the academic commentary cited engaged with the conceptual question of how DNSA obligations come into existence.

In the US, the issue of the international human rights obligations of DNSAs has arisen in the context of the ATS.Footnote ³⁹ In Sosa v Alvarez Machain, the US Supreme Court held that it is possible to sue governmental officials for breaches of international human rights obligations under the ATS, so long as the relevant international rule was ‘specific, universal, and obligatory’ and it was a proper exercise of judicial discretion.Footnote ⁴⁰ The position in relation to corporations is, however, less clear. A split emerged between the different circuits of the Federal Court of Appeals, with divisions even between judges in particular cases.Footnote ⁴¹ The US Supreme Court has also been unable to reach a clear position, despite dealing with the issue on more than one occasion.Footnote ⁴²

Similar debates can also be seen outside the realm of international human rights law. The question of the existence and scope of the international obligations of DNSAs has been debated in international criminal law and international environmental law.Footnote ⁴³ Understanding how and why DNSAs can be bound by international legal obligations is therefore not merely a matter of academic interest. A convincing explanation as to how and under what conditions international obligations can bind DNSAs is needed if courts and tribunals are to be able to determine whether a DNSA is subject to an international obligation in a particular case. Even where the international obligations of DNSAs are well-established, a lack of clarity as to their foundations may negatively impact their perceived legitimacy and lead to a reluctance to enforce them.

2.2. The uncertain foundations of DNSA obligations

The discussion in Section 2.1 shows that, as a matter of positive law, there is no question that DNSAs can and do have international obligations. The more difficult issue is explaining how and why such obligations exist. It is trite that, generally speaking, the binding force of international obligations comes from consent. This is made apparent in Articles 34 and 35 of the Vienna Convention on the Law of Treaties (VCLT), which provide that States are only bound by treaty obligations to which they have consented.Footnote ⁴⁴ Indeed, Gerald Fitzmaurice went as far as arguing that the importance of consent was such that it was more accurate to describe treaties as a source of obligation between the consenting parties, akin to a private contract, than as a source of law per se.Footnote ⁴⁵ The importance of consent is also reflected in the formation of customary international law, particularly the opinio juris requirement and the persistent objector doctrine.

The concept of consent has been under increased scrutiny in recent years, with writers identifying areas where States seem to become subject to international obligations without their consent and pushing back on a purely ‘voluntarist’ understanding of international law more broadly.Footnote ⁴⁶ Despite such debates, consent retains a central role in explaining the existence of international legal obligations. The statement from the ICJ in Barcelona Traction that ‘a body of rules could only have developed with the consent of those concerned’ remains as true today as it was when it was made.Footnote ⁴⁷

As a consequence of the central role of consent, international obligations generally have a self-imposed character. Such obligations exist because of the choice of the State to be bound by, or at least not to object to, a particular obligation. There is no formal hierarchy or relationship of authority between States, unless certain powers have been specifically delegated to a group of States such as the UNSC. Instead, in formal terms at least, the relationship between States is a horizontal one pursuant to the principle of the sovereign equality of States.

Despite the central role of consent in international law in general, the consent of DNSAs plays no role in the creation or validity of international obligations binding DNSAs. It is generally accepted that the validity of such obligations does not depend on the consent of such actors; indeed, the consent of DNSAs is typically neither sought nor obtained. Whilst there are a few exceptional examples where the consent of DNSAs is a condition of the existence of the relevant international legal obligation, this is due to the decision of States to make these obligations contingent on such consent.Footnote ⁴⁸ In the context of UNSC resolutions, only States, and not DNSAs, have consented to Articles 25 and 48 of the UN Charter, which give UNSC resolutions their legally binding character.Footnote ⁴⁹ In the absence of a consent requirement, the international obligations of DNSAs are not self-imposed in the same way as those of States. This general lack of a consent requirement therefore poses a conceptual challenge when seeking to understand such obligations; put simply, without the consent of the actors bound, the foundations of DNSAs’ international obligations are difficult to identify.

The existence of international obligations which bind DNSAs also blurs the boundaries between the international and domestic spheres. Consistent with the traditional focus of international law on the obligations of States, the obligations of and relations between non-State entities are normally considered to be a matter for domestic law. In the words of James Crawford, the jurisdiction to regulate DNSAs ‘is an aspect of sovereignty’.Footnote ⁵⁰ This is given effect by Article 2(7) UN Charter, which provides that, subject to enforcement measures in Chapter VII (threats to or breaches of the peace), ‘[n]othing contained in the present Charter shall authorize the United Nations to intervene in matters which are essentially within the domestic jurisdiction of any state …’.Footnote ⁵¹ The exclusive authority to regulate within territorial bounds is central to the modern understanding of statehood and delimits the roles of domestic and international law. States therefore have a wide discretion as to the domestic obligations of DNSAs in pursuit of the cultural and political values and policies considered domestically important. The existence of international obligations which bind DNSAs represents a potential rival to the authority of domestic law, and the trend towards imposing obligations on DNSAs raises the difficult questions of the relationship between and the proper roles of domestic and international law.

Several writers have recognised this problem and attempted to provide an explanation as to how such obligations come into existence. Section 3 discusses the leading accounts and explains why none of them can provide a satisfactory general explanation of DNSA obligations. The discussion will, however, argue that each account contains several valuable insights which provide the basis for the State imposition account developed in this article.

3.1. Legislative jurisdiction

Perhaps the leading account in the existing literature is that DNSAs become bound by international obligations as a result of the doctrine of ‘legislative jurisdiction’.Footnote ⁵⁴ This account builds on the views of several States attending the Diplomatic Conference of 1949 and the Diplomatic Conference of 1974–77 that an international treaty signed by a State also binds those within the territory of the State.Footnote ⁵⁵ Put differently, on this view a State imposes international obligations on all individuals within its territory whenever it becomes a party to an international treaty.

The legislative jurisdiction account has several advantages, including the ability to accommodate the existence of several examples of international obligations with which other potential justifications struggle.Footnote ⁵⁶ Particularly important is the account’s ability to explain how non-State armed groups can be bound by international obligations regardless of whether they ever actually become the government of the State.Footnote ⁵⁷ However, the main insight offered by the legislative jurisdiction account is that it relies on States as the primary actors and sole creators of international law.Footnote ⁵⁸ Whilst some other potential explanations depend on giving DNSAs certain attributes of States,Footnote ⁵⁹ such as the ability to contribute to the formation of international law, the ‘legislative jurisdiction’ account avoids the need to engage in the controversial reimagining of such foundational propositions. The account developed in Section 4 recognises this principle and places the sole law making authority of States at its centre.

The account does, however, suffer from several drawbacks which mean that it is not a suitable candidate for developing into a general account of DNSA international obligations.Footnote ⁶⁰ The main shortcoming is an uncertainty as to the source of DNSA obligations. The leading exposition of the account is provided by Sandesh Sivakumaran who describes it in the following way:

This short explanation falls between two distinct possibilities for the source of DNSA obligations. The reference in the first sentence to ‘the competence of the government to legislate’ is consistent with a view that DNSA obligations are imposed on DNSAs. This view recognises a distinction between a State and DNSAs but suggests that former has the ability to regulate the latter in the form of international obligations. On this understanding, there is little conceptual difference between the domestic and international obligations binding DNSAs. The idea of international obligations being imposed on DNSAs is developed further in Section 4.

However, the second sentence of the passage is consistent with a different understanding of DNSA obligations. The second sentence appears to assimilate a State and ‘all individuals within its territory’ so that the consent of the State to an international obligation is deemed to also be that of the persons within its territory. This suggests that deemed consent, rather than the will and regulatory authority of the State, is the basis for DNSA obligations.

This uncertainty as to the basis of DNSA obligations may give rise to practical issues. For example, the significance of an express declaration that a DNSA does not wish to be bound by a particular obligation may be approached differently depending on which of the two understandings above is adopted. Whereas the first sentence of the passage would suggest that such a declaration is irrelevant, the second sentence may suggest that the effect of such a declaration is that the DNSA cannot be deemed to have consented to the obligation and is therefore not bound. The passage also creates some uncertainty as to when exactly DNSA obligations arise. The first sentence suggests that such obligations are dependent on a conscious decision of the State to impose such obligations, so that it is possible for a State to become party to a treaty without the obligations in that treaty also being automatically imposed on DNSAs. The second sentence, however, suggests that DNSAs are automatically bound whenever a State enters into a treaty, even if the State does not consciously or actively intend to impose such obligations on DNSAs.

The above explanation of the source of DNSA obligations also raises several other conceptual questions. Perhaps most importantly, the identity of the persons which may be the subject of DNSA obligations is unclear. The above passage refers both to a State’s ‘nationals’ and ‘individuals within its territory’. These are clearly overlapping but not identical categories. Further issues arise once corporations are considered. It is, for example, unclear how the legislative jurisdiction account manages corporations which have a presence in the territory of another State or corporations which have a presence in several States.

Insofar as the legislative jurisdiction account does rest on some notion of deemed consent on the part of the DNSAs (as per the second sentence of the passage from Sivakumaran), this is inherently problematic and should be rejected. Where individuals are acting on behalf of a State, it is undoubtedly right to assimilate the State and the individuals through which the State acts. However, in order to account for the international obligations of DNSAs, this aspect of the legislative jurisdiction account has to go a step further and assimilate a State with all natural and legal persons with the State’s nationality or in the State’s territory. Only by doing so is it possible to claim that the consent of the State can be attributed to DNSAs in any meaningful sense. Yet such assimilation is difficult to accept.Footnote ⁶² Since Emer de Vattel’s work, international law has recognised that States possess some existence which is distinct from their populations or rulers, at least at a formal level.Footnote ⁶³ The legislative jurisdiction account is inconsistent with this understanding of international law as it elides the identity of the State and its nationals and others within its territory, without any explanation for doing so. It is therefore hard to understand why the consent of the State should be treated as also amounting to the consent of DNSAs. In the absence of a convincing justification for this position, the legislative jurisdiction account must be considered flawed.

Following this logic to its natural conclusions demonstrates its inherent difficulties. Taken at face value, the argument that ‘when a state ratifies a treaty, it does so not just on behalf of itself but also on behalf of all individuals within its territory’ suggests that all of a State’s international obligations are also prima facie binding on all those within its territory.Footnote ⁶⁴ Such a position may also lead to natural and legal persons facing conflicting obligations under national and international law. States do of course have the power to adopt laws and regulations which impose obligations on DNSAs, but such obligations operate on the domestic rather than international level. This is arguably true even in so-called ‘monist’ States, as the status of international law within the domestic legal system depends on a rule of domestic law rather than being an inherent feature of international law.

Some proponents of the legislative jurisdiction account attempt to cater for this problem by arguing that the applicability of international obligations to DNSAs depends on the existence of an intent to bind such actors.Footnote ⁶⁵ But this brings the reader back to the discussion regarding the uncertainty as to the source of DNSA obligations under the legislative jurisdiction account. If DNSAs have international obligations simply because ‘the relevant [obligation] binds all individuals within [a State’s] territory’,Footnote ⁶⁶ it is not clear what role intent can have in narrowing the number of those obligations.

The legislative account also struggles to deal with several key issues in international humanitarian law, the area on which the account focuses. One such difficulty is how to accommodate non-State armed groups.Footnote ⁶⁷ International humanitarian law recognises that such groups, distinct from the individuals of which they comprise, are important actors as they facilitate the organisation and coordination of activities which would not otherwise be possible. Non-State armed groups also have greater political significance than individuals as they may represent an alternative to existing governmental structures and policies. Yet these groups generally receive no legal recognition from States, making it difficult to explain their liability within the legislative jurisdiction model. Put simply, if a State does not recognise the existence of a group, it is difficult to understand how that State can impose international obligations upon it.

The reliance on nationality in the first sentence of the passage from Sivakumaran creates further problems. Reliance on nationality fails to explain how international obligations can bind DNSAs in circumstances where the State of nationality has not itself consented to such obligations.Footnote ⁶⁸ Given that not all States have consented to all rules of international humanitarian law, reliance on nationality leaves the status of such rules unclear. More generally, Jann Kleffner points out that nationality is an increasingly unstable concept in international law, with international courts increasingly looking beyond formal nationality to more substantive considerations.Footnote ⁶⁹

Invoking a State’s territorial regulatory jurisdiction instead does avoid some of these problems, but it does not provide an answer to the criticisms regarding the recognition of groups. Further, relying on territorial jurisdiction also introduces other problems given the cross-boundary nature of many of the relevant activities. Non-State armed groups, for instance, frequently operate across borders but there is no support for the proposition that the content of their legal obligations changes depending on the territory they happen to be in at a particular time.

3.2. International legal personality

A second account of why DNSAs can have obligations under international law rests on the notion of international legal personality.Footnote ⁷⁰ This account suggests that international legal personality extends beyond States to encompass a range of other actors, most relevant of which for the current discussion are non-State armed groups.Footnote ⁷¹ If such actors do have international legal personality, the argument goes, they are bound by rules of custom and general principles in the same way as States.Footnote ⁷²

The main advantage offered by the ‘personality’ account is that it does not require the interposition of the State between international law and DNSAs in order to explain the international obligations of such actors.Footnote ⁷³ Instead, DNSAs are bound directly. On this account, DNSAs’ international obligations are not merely an extension or instantiation of a State’s right to regulate; instead, DNSAs are recognised as important international actors in their own right which international law has a distinct interest in regulating. This is particularly important in relation to non-State armed groups. Dispensing with the reliance on concepts of nationality and territory also helps the personality account to avoid many of the problems of the legislative jurisdiction approach discussed in Section 3.1.

Nevertheless, an attempt to generalise this account runs into problems. At a basic level, this account can only accommodate a limited range of the international obligations of DNSAs. In particular, the international legal personality account seeks to explain only those rules which exist in customary international law or as general principles of law. The account cannot explain the existence of international obligations binding DNSAs which exist only in treaties which DNSAs have not signed or consented to.Footnote ⁷⁴ DNSAs, on this view, are not bound by any international obligations which only exist by virtue of a treaty at the time of the relevant act. Yet there are treaty obligations which are generally understood to bind DNSAs, despite their status in customary international law or as general principles being unclear.Footnote ⁷⁵

The main problem, however, is that the concept of international personality offers little clarity as to how the international obligations emerge, both generally and in connection with DNSAs in particular. As a general matter, the definition of international legal personality which is based on whether an actor has rights or obligations under international law cannot, without circularity, itself provide a basis or explanation for recognising the existence of such international obligations.Footnote ⁷⁶ It is instead necessary to turn to another concept. In the context of State obligations, this further concept is consent. Further, the recent expansion of actors which are recognised as having international legal personality has given rise to the question of how to distinguish between different types of legal personality.Footnote ⁷⁷ Some go as far as to argue that the concept of international legal personality and the subject/object distinction should be abandoned in favour of the more neutral concept of ‘participants’.Footnote ⁷⁸

These problems are amplified when considering DNSA obligations. As explained by Alvarez, relying on the concept of international legal personality is a fundamentally ‘top-down’ approach to international law, which does not provide any explanation as to when DNSAs are subject to international obligations and why, and risks obscuring differences between different actors.Footnote ⁷⁹ In particular, reliance on the concept of international legal personality fails to explain why DNSAs are treated differently to States when it comes to consent to be bound. Whilst there are calls from some quarters to recognise a role for DNSAs in the creation of customary international law,Footnote ⁸⁰ this view is yet to become mainstream and the orthodox position remains that only States have lawmaking powers.Footnote ⁸¹ This is further reflected in the fact that DNSAs have no opportunity to avoid the application of such rules by persistently objecting.

There is also no obvious way in which DNSAs can influence the recognition of general principles of law. In some ways, DNSAs are not entirely unique in this position; newly formed and newly independent States are taken to be bound without their consent by international law as it exists at the time the State came into being.Footnote ⁸² Yet, going forward, such States are able to contribute to customary international rules and general principles in a way that DNSAs cannot. In this way, DNSAs are treated differently to States and are rule-takers without also being rule-makers. Hence, even if it is correct to describe the relevant actors as having international legal personality, it does not take the search for the source of DNSA obligations any further.

3.3. DNSAs as third parties

A third account advanced by Antonio Cassese focuses on the effect that treaties can have on third parties.Footnote ⁸³ Writing specifically on Additional Protocol II, Cassese argues that it is possible to apply Article 35 VCLT, which provides that non-parties to a treaty may be bound by the treaty when ‘the parties to the treaty intend the provision to be the means of establishing the obligation and the third State expressly accepts that obligation in writing’.Footnote ⁸⁴ As non-State armed groups do often express their willingness to comply with international humanitarian law,Footnote ⁸⁵ this account argues that the VCLT rules on third parties are an adequate way of explaining the foundations of such obligations.

There are two particularly helpful elements of Cassese’s account which are worth drawing out. The first is that this account explicitly starts from the premise that DNSAs are not parties to the relevant treaty and therefore would not ordinarily be considered to be bound by the treaty. Whereas the other accounts discussed in this section treat the international obligations of DNSAs as routine or, rather, consider the liability of DNSAs as the normal position, Cassese’s account treats such obligations as exceptional. This coheres better with the orthodox position that international legal obligations normally and primarily bind the States which have consented to those obligations. The second helpful element, related to the first, is that this account recognises the significance of obligations being specifically addressed to DNSAs.Footnote ⁸⁶ In contrast to the other accounts discussed in this section, which are potentially too broad in supposing that all international obligations potentially bind DNSAs, Cassese’s focus on whether DNSAs are addressees of the relevant obligation gives effect to the exceptional nature of such obligations. Although this account considers this requirement through the lens of Article 35 VCLT, the importance of identifying the addressee is a valuable insight which will be developed further in Section 4.

Cassese’s third party account takes a more orthodox approach than the other existing accounts to the source of DNSA obligations. Whereas the other accounts seek to find a basis for such obligations which does not rely on the consent of DNSAs, Cassese’s reliance on Article 35 VCLT assumes that consent is a necessary condition for the existence of obligations. This reliance on consent does, however, lead to several problems, both on Cassese’s own terms and when determining the wider relevance of his account.

The main problem is that Cassese cannot explain DNSA obligations which exist despite the lack of the consent required by Article 35 VCLT. As explained in Section 2.2, whilst some international humanitarian obligations are dependent on the consent of the relevant non-State armed group, many are not. Non-State armed groups may express their willingness to comply with such obligations but this does not make any difference to the existence of a legal obligation which is binding upon them.Footnote ⁸⁷ Cassese’s account is even more difficult to square with the reality of international criminal law, where there is no suggestion that the existence of international obligations which bind individuals is dependent on the consent of those individuals. Indeed, it is difficult to conceive of how such consent could realistically be obtained. The account also does not provide a basis for obligations existing in customary international law or arising as a result of general principles of law, due to its focus on treaty obligations. There is also the matter of whether verbal consent would satisfy the requirements in Article 35, given that Article 35 requires the third party to accept the obligation in writing.Footnote ⁸⁸

Adopting this account therefore requires taking one of two paths. The first is to argue that DNSAs are not bound by international obligations where the consent requirement of Article 35 is not met. This would, however, be a challenging argument to defend given the significant number of international obligations which purport to be binding on DNSAs without their consent, discussed in Section 2. The result would be a patchwork of ad hoc liability.Footnote ⁸⁹ Looking forward, Cassese’s account would also create significant obstacles for any efforts to impose international human rights obligations on individuals or corporations as it would be necessary to seek the consent of all relevant actors. The second path is to suggest that the non-consensual obligations of DNSAs rest on another basis. Yet there is no indication in Cassese’s account as to what this alternative basis might be.

Relatedly, it is unclear what justification there is for invoking Article 35 VCLT in this context.Footnote ⁹⁰ Cassese does not provide a clear and principled justification for invoking the provision and it is doubtful whether one can be identified. No such justification can be found in the text of the VCLT itself. Article 1 VCLT provides that ‘[t]he present Convention applies to treaties between states’ and Article 35 refers only to third States and not DNSAs.Footnote ⁹¹ That the VCLT reflects customary international law does not take the matter any further, as those customary rules may also apply only to States.

It is also difficult to identify any justification for relying on the provision as a matter of principle or analogy. In the words of the ILC, the consent requirement in Article 35 is ‘one of the bulwarks of the independence and equality of States’.Footnote ⁹² This concern for independence and sovereign equality plainly does not apply to DNSAs. It is therefore unclear why it is appropriate to invoke this provision and the consent requirement when explaining the source of the international obligations of DNSAs.

3.4. Governmental functions

Two final accounts which relate to the exercise of DNSAs’ governmental functions can be dealt with together. The first, which could be termed the ‘de facto’ account, argues that non-State armed groups become bound by international humanitarian law by virtue of their exercise of governmental functions over part of a State’s territory.Footnote ⁹³ If, the argument goes, the State has consented to certain international obligations, groups which are de facto exercising governmental functions should also be considered to be bound by such obligations. The second account, the ‘succession’ account, adopts similar reasoning but takes a slightly different focus. According to the succession account non-State armed groups which claim to be the legitimate government of a particular State are bound by the international humanitarian obligations of that State.Footnote ⁹⁴

A central aspect of both accounts is that the binding force of international obligations is understood to derive from the consent of the relevant State rather than imposed directly on the DNSA. Both accounts therefore resemble the legislative jurisdiction account discussed in Section 3.1, insofar as the legislative jurisdiction account depends on an assimilation of the State and the relevant DNSA. There are, however, important differences. Instead of taking a ‘top-down’ approach which focuses on the capacity of the State against which the armed group is fighting, the two governmental control accounts considered here adopt a ‘bottom-up’ approach which focuses on either the power which the armed groups have on the ground or their aspirations to become the government.Footnote ⁹⁵ By focusing on governmental functions, both accounts also recognise that non-State armed groups are often political actors and not merely common criminals. Hence, whilst State-centric, both accounts avoid the problems associated with the legislative jurisdiction approach discussed in Section 3.1 and recognise the existence and significance of groups as actors. The principle of recognising the distinct identity of DNSAs is valuable and one which will be taken forward in the next section.

The limited scope of these accounts makes them particularly unsuited to generalisation. By focusing on non-State armed groups which assume some governmental functions, neither account can provide an explanation as to how individuals or corporate actors may become bound by international legal obligations. It is therefore difficult to conceive of how either account could be used to explain the existence of international criminal obligations or provide the foundation for the potential future imposition of obligations on corporate actors.

Even within the context of non-State armed groups, however, both accounts are problematic. A closer analysis of the State-centric nature of both accounts reveals a number of assumptions which may not be borne out. For instance, the de facto control account assumes both that the non-State armed group has control over a stable area and that the group at least attempts to exercise its authority in a manner similar to how a conventional government would. Yet, in many cases, neither of these two assumptions are true.Footnote ⁹⁶ Slightly different—but no less problematic—assumptions underpin the succession account. The succession account assumes that the non-State armed group aspires to govern the relevant State, but this is not necessarily the case; such groups have a range of different goals.Footnote ⁹⁷ Whilst some groups may aim to remove the existing government and govern in their place, some may hope only to cause disruption in the hope of achieving limited political or economic objectives, whilst secession may be the goal of others. Even where a group does aspire to become the government of the relevant State, the succession account remains conceptually deficient. By focusing on succession, the account has no way of explaining why non-State armed groups have international humanitarian law obligations before they become the government of a State, or why groups which hope but fail to become the government of a State are considered to have such obligations. Both accounts therefore struggle to fit with the existing law on non-State armed groups, which imposes obligations on such groups without any regard as to whether the group exercises governmental functions or succeeds in becoming the government of the State (or aspires to do so).

In sum, none of the existing accounts are capable of resolving the conceptual puzzle of how DNSAs can be bound by international obligations in the absence of their consent. Each account does, however, contain useful insights into the problem and the issues at stake. The next section will seek to draw upon these insights to formulate an account of DNSAs’ international obligations which is both rooted in existing legal principles and capable of accounting for the full range of actors and obligations discussed in Section 2. These insights are: first, States are the only actors with the capacity to create international law; second, there is a need to recognise DNSAs as conceptually distinct from the States with which they may be related; third, DNSAs’ international obligations are unusual and there are reasons to recognise such obligations only when there is a clear intention to bind DNSAs; and, fourth, the importance of recognising that DNSAs are significant actors in their own right with their own political and economic interests.

4.1. The capacity to create DNSA international obligations

In light of the limitations of the accounts explored in Section 3, it is necessary to formulate a more general account which is capable of accommodating the existence of the full range of DNSA international obligations. After explaining the key features and advantages of the suggested account, this section will discuss the implications for understanding international obligations and identify the potential risks of creating such obligations.

The new account of DNSA obligations put forward in this article is that States have a capacity to impose international obligations on DNSAs.Footnote ⁹⁸ The capacity recognised by the State imposition account is, prima facie, a broad one. However, the scope of that capacity is limited by other rules and principles of international law which constrain its exercise in important ways.

There are three key features of this ‘State imposition’ account which serve to explain the existence of the current international obligations of DNSAs, guide States which seek to impose additional DNSA international obligations in the future and distinguish this account from those discussed in Section 3. First, under the State imposition account it is States and States alone which have the ability to create international legal obligations. Put differently, the existence of international obligations binding DNSAs is dependent on the will of States. The international obligations of DNSAs are therefore not self-imposed in the same way as State obligations are. Although DNSAs may have an informal role in the lawmaking process, it is only States which can formally contribute to the formation of international law. To this extent, the State imposition account resembles the legislative jurisdiction and governmental functions accounts and rejects the aspects of the international legal personality and third party accounts which suggest that DNSAs have some part to play in the formation of international legal obligations.

This account has the benefit of recognising and reinforcing the regulatory authority and legitimacy of States. This is not simply a matter of defending the status quo. Rather, excluding non-State actors from the law making process is an issue of principle, which can be put in both negative and positive terms. In negative terms, there is no reason why DNSAs should be endowed with lawmaking capacities. As discussed in Section 2.2, there is no independent principle of international law equivalent to the equality of States which requires DNSAs to have some lawmaking role or to consent to the obligations which bind them. Even Cassese’s third party account, which has DNSA consent at its core, does not provide a clear, principled explanation of why such consent is necessary.

In positive terms, there are good reasons of principle why DNSAs should not have any formal lawmaking role in the international legal system. International obligations inevitably have significant impact on third parties, including natural and legal persons. This is true even in traditional issues of international law such as boundary disputes, which have implications for where the nationals of each State can reside, the resources they have access to and the domestic laws they are subjected to. The impact on natural and legal persons of more recently developed areas of international law, such as human rights and trade and investment regimes, is more obvious but the basic idea is the same. This is particularly true of international obligations imposed on DNSAs. The obligations discussed in Section 2 have clear consequences for other natural and legal persons, given that they often regulate how DNSAs interact with each other. International obligations are therefore not, in substance at least, a matter of private relations between a limited number of parties. Instead, they have an inherently regulatory character which is to protect public interest values, however those values are defined.

DNSAs are fundamentally ill-equipped to have proper regard for, or to give effect to such values, even if they are well-intentioned. The proper pursuit of such values requires balancing competing interests. States have various mechanisms for the weighing of such interests and for preventing the pursuit of any particular interest at the expense of all others. However, by their nature, DNSAs do not have any such mechanisms, and will naturally be motivated by their own interests, whether personal, political or financial, even when those interests are tension with others or the broader public interest. Indeed, the very existence of some actors, particularly companies, depends on the pursuit of a particular interest and they may be under domestic legal duties to pursue that interest. Any international lawmaking capacity which such actors are endowed with would inevitably be exercised with individual rather than public interests in mind.

The second feature of the State imposition account is that States’ capacity to create international obligations on DNSAs is ‘free-standing’. It therefore does not depend on any pre-existing link between the relevant State and DNSA, such as through nationality or territory. Instead, States are able to impose international obligations on DNSAs with which they have had no prior relationship. In this way, the State imposition account is a rejection of the legislative jurisdiction account and instead bears greater resemblance to the international legal personality and the third party accounts.

Whilst this results in a very broad prima facie capacity to create DNSA international obligations, there are good reasons for recognising such a broad capacity. Perhaps most importantly, it is only by adopting such a broad view that it is possible to explain the range of DNSA international obligations discussed in Section 2. Departing from the view that it is necessary to mediate the existence of DNSA international obligations through States also better reflects the role which DNSAs have on the international plane. Instead of treating the role of such actors as parasitic on a particular State, this approach takes seriously the independent role of DNSAs as ‘participants’ in the international legal system and recognises the distinct interest which international law may have in regulating such actors. This gives effect to many of the considerations underlying the international legal personality account, without becoming trapped in the conceptual difficulties of that approach. Importantly, this broad prima facie capacity is subject to the limitations imposed by other rules and principles of international law, explained further below.

The third feature of the State imposition account is that international obligations only bind DNSAs if they are addressed to such actors. This will naturally be a matter of interpretation which will need to be assessed on a case-by-case basis. However, the mere fact that an obligation binds a State, or even all States, is not itself sufficient to support the view that the obligation also applies to DNSAs.Footnote ⁹⁹ This concern played a role in the decision of the tribunal in Urbaser v Argentina to reject the respondent’s counterclaim, which was based on the notion that the claimant investor had a positive obligation under international law to provide the population with drinking water and sanitation services.Footnote ¹⁰⁰ It may also be necessary to recognise the variety of DNSAs and distinguish between them, so that an obligation may apply to some DNSAs, or categories thereof, but not others. This feature bears some resemblance to the third party account modelled on Article 35 VCLT favoured by Cassese in that both require evidence of an intention on the part of States to impose an obligation. But, not being rooted in that provision, it is not accompanied by the second requirement of consent on the part of the DNSA.

This approach is consistent with a prominent interpretation of the decision of the PCIJ in the Danzig case. Relying on the passage which this article opened with, Kate Parlett has argued that the findings of the PCIJ in Danzig should be interpretated as stating that in order to create rights for third parties ‘states must clearly intend to adopt rules’ to that effect.Footnote ¹⁰¹ The same conclusion can be reached with regard to third party obligations. In practice, this ‘addressee’ requirement can be best given effect by recognising a rebuttable presumption that States do not intend to impose obligations on DNSAs.

Accepting the existence of this requirement and the associated presumption throws doubt on the approach of some writers who favour the view that broad and generally worded treaties impose direct international obligations on DNSAs.Footnote ¹⁰² Nevertheless, there are several advantages to recognising the existence of such a requirement. The main advantage is that this requirement serves to ensure that international regulation of DNSAs remains exceptional, and emphasises that domestic law has the primary role in regulating such actors.Footnote ¹⁰³ Only where States have identified acts of DNSAs which they consider to impact a distinct international interest in need of protection would a DNSA be subject to international obligations. This also reflects the reality that the majority of international legal obligations are relevant only to States and not to DNSAs. A further advantage is that this requirement serves to reinforce the notion that DNSAs are significant actors in their own right. In practical terms, the requirement avoids the problem of how to distinguish between obligations which do bind DNSAs and those which do not, discussed in relation to the legislative jurisdiction and international legal personality accounts in Sections 3.1 and 3.2. This would provide greater legal certainty to both States and DNSAs.

This final requirement that obligations are directly addressed to DNSAs raises the issue of how UNSC resolutions can create legal obligations for DNSAs. Even if UNSC resolutions are specifically addressed to DNSAs, their binding legal force is derived from Articles 25 and 48 UN Charter, which do not apply to DNSAs. Other provisions do, however, provide an answer consistent with the general approach developed in this article. According to Article 24 UN Charter, Member States of the UN agree that ‘the Security Council acts on their behalf’ when carrying out its various functions. The plain meaning of this provision supports the view that a resolution of the UNSC represents the will of States as a whole. The ability of the UNSC to bind DNSAs can then be explained on the basis that States have effectively authorised the UNSC to impose such obligations within the scope of its powers. Even where a resolution is opposed by a non-permanent member of the UNSC or a member of the UNGA, the effect of Article 24 is that such opposition is not legally relevant to the resolution.

This capacity does not, however, exist in a vacuum. Rather, whilst States have a broad prima facie capacity to impose international obligations on DNSAs, this capacity remains subject to existing rules and principles of international law. There are two limitations of particular importance which should be noted here. The first is that States cannot impose international obligations on DNSAs which conflict with peremptory norms of international law.Footnote ¹⁰⁴ This limitation naturally follows from the hierarchically superior position which peremptory norms enjoy over other rules of international law. States can therefore not impose international obligations on DNSAs where doing so would either require the DNSA to act contrary to a peremptory norm, or would in some way facilitate a State to act contrary to such a norm.

The second such limitation follows from the obligation of States to respect the sovereignty and regulatory authority of other States.Footnote ¹⁰⁵ As explained by the PCIJ in its judgment in the Lotus case, ‘the first and foremost restriction imposed by international law upon a State is that – failing the existence of a permissive rule to the contrary – it may not exercise its power in any form in the territory of another State.’Footnote ¹⁰⁶ This second limitation has important consequences when considering the ability of small numbers of States to impose international obligations on DNSAs. There is no conceptual requirement that international obligations binding DNSAs be conditional on the unanimity of States. Unanimity is highly unlikely to be achieved and even the well-established examples of DNSA obligations do not have the unanimous support of States. In principle, therefore, a small group of States could decide to impose international obligations on DNSAs. A problem may, however, arise if such an obligation purports to bind a DNSA which is present in the territory of a State which does not support the creation of an international obligation. In such a situation, the States seeking to impose the obligation would be vulnerable to claims that they have infringed the sovereignty and regulatory authority of the States where the DNSAs are present, unless the consent of the latter has been obtained. This would be especially so if the international obligation conflicts in some way with a domestic obligation.

The State imposition account of DNSAs’ international obligations requires a shift in how international obligations are understood. As discussed at the outset of this article, international obligations are generally rationalised as being self-imposed and thus dependent on the consent of the actors which are bound. If, however, the State imposition account is accepted, this must be reconsidered. On the account developed in this article, international obligations may arise from two bases: first, an expression of consent to be bound; or second, the right of States to create international obligations. In the case of State obligations, the two bases clearly overlap. However, only the latter is relevant to an international obligation of a DNSA, which is not self-imposed but rather is the result of the will of States. Further, DNSA obligations do not exist in the context of a horizontal relationship between equals but in a hierarchical relationship, in which States have regulatory authority over DNSAs.

Rationalising DNSA obligations in this way has several notable consequences. The first is that such obligations cannot be understood in terms of a traditional bilateral or multilateral structure in which obligations are owed to a limited number of actors. Rather, the international obligations of DNSAs are more akin to restrictions on liberty similar to those in domestic criminal law. Moreover, such obligations typically purport to pursue or protect an interest of the international community as a whole rather than an interest of a particular actor. Such obligations are therefore not ‘owed’ to anyone and do not have the same reciprocal structure as traditional inter-State obligations. The rules or jurisdictional competence of a court or tribunal may identify particular actors who have standing to bring a claim, but that does not change the fundamental nature of the underlying obligation.

Second, it is necessary to dispense with the concept of international legal personality as an analytical tool. Not only does the concept fail to provide any analytical guidance, as discussed in Section 3.2, it also has the potential to lead to conceptual confusion. Although the notion of international legal personality may imply that the existence of international rights and obligations is somehow dependent on the characteristics of the beneficiary or duty bearer, this is incorrect. There is no clear division between actors which may benefit from international rights or be subject to international obligations, and those which may not. Nor is it correct to understand the international legal system as a matter of ‘all or nothing’, in which a particular actor is subject to the full range of rights and obligations or none at all. Instead, the existence of international rights and obligations is dependent on the policy choices of States when exercising their lawmaking and regulatory authority. The consequence of this is an international legal landscape characterised by variable geometry, in which the rights and obligations of different actors can vary depending on the subject matter and the relevant actor. The concept of international legal personality obscures these complexities and reduces the status of actors in international law to a binary position.

Third, it is necessary to reconsider the relationship between consent and the existence of an international obligation. Contrary to what is commonly assumed, the discussion in Section 2.1 shows that there is no inherent connection between consent and an international obligation, and that international obligations are not necessarily self-imposed. International obligations can and do exist despite a lack of consent on the part of the actors bound. Instead, the right to consent, or not, to international obligations is best understood as a characteristic or privilege of States that is a corollary of the principle of the equality of States. However, as a matter of practical reality, DNSAs may have the opportunity to influence the content of their international obligations through lobbying efforts, and States may seek to obtain the agreement of particularly important or influential DNSAs. As a matter of formal legal principle, however, the privilege of the right to consent to international obligations is limited to States.

4.2. Policy considerations

The State imposition account developed in this article may give rise to various issues of legal policy. A natural objection is that it envisages a broad power on States to create international obligations for DNSAs. There is a risk that some obligations may suffer from problems of legitimacy or practicality, and that the effectiveness of such obligations may be undermined by a lack of compliance. These concerns may have influenced features of the existing accounts discussed in Section 3. These are legitimate concerns which should be given proper weight, but the various attempts to address them through the creation of legal rules or requirements have created the conceptual problems discussed in Section 3. Many of these concerns are dealt with by recognising the limits imposed by the rules relating to peremptory norms and the obligation on States to respect the sovereignty and regulatory authority of other States. Alongside these legal limitations, there are various policy concerns which are relevant to whether it is desirable for States to exercise their capacity to impose international obligations on DNSAs in particular circumstances and, if so, what form such obligations should take.

The types of concerns that arise differ depending on the nature of the particular obligation and the identity of the DNSA. Nevertheless, it is possible to outline some broad themes which are relevant across international law. One group of concerns relates to the imposition of obligations on domestic private actors outside of the domestic legal system. Inherent in the notion of an international obligation is that States agree that it is appropriate to have a common set of standards or rules and that divergence is disadvantageous or undesirable in some way. Where DNSAs are concerned, however, agreement at the international level will often not be appropriate because of differing cultural and political values.Footnote ¹⁰⁷ Imposing direct international obligations in such circumstances therefore risks disempowering States and government institutions at a domestic level in a context when there is a range of legitimate approaches to balancing competing interests and values.Footnote ¹⁰⁸ The concerns of several writers in the context of attempts by the European Union to introduce a harmonised code of contract law may be instructive in this regard.Footnote ¹⁰⁹ As a result, before imposing international obligations on DNSAs, States should be sure that a single common rule is in fact desirable. A failure to ask this question risks DNSAs becoming bound by an international obligation from which it is difficult or impossible to depart, or States entering into lengthy and expensive negotiations which may prove to be fruitless.

To avoid such a risk, States would be advised to only impose obligations on DNSAs where there is a clear and distinct international interest which is in need of international legal protection such that there is a risk of a ‘regulatory deficit’.Footnote ¹¹⁰ As argued by Miriam Gur-Arye and Alon Harel in the context of international criminal law, the international obligations of DNSAs should not be seen as second best to, or a replacement for, domestic legal obligations.Footnote ¹¹¹ Rather, such obligations protect a distinct set of specifically international interests which do not exist, or cannot be adequately protected, at the domestic level. Acting in this way also reduces the likelihood that the creation of a DNSA obligation will violate a State’s sovereignty and regulatory authority.

Related to this is the concern that a growth in the creation of international obligations for DNSAs may result in international law rivalling domestic law for the authority to regulate such actors. The potential problems arising from such a ‘regulatory turn’Footnote ¹¹² can be seen from two different perspectives. From the perspective of domestic lawmakers and institutions, there is a risk that the international legal system undermines or circumvents the domestic lawmaking processes, challenging the regulatory authority of those domestic actors.Footnote ¹¹³ Such concerns would be particularly acute where international law imposes obligations on DNSAs which had been rejected by the domestic legislature in a particular State. But such concerns may also exist where a government favours regulation on the international rather than the domestic plane for reasons of political expediency.

From the perspective of the relevant DNSA, there is a possibility that they will be subject to conflicting obligations. Whilst these obligations take effect on different planes, this does not help the DNSA resolve the conundrum of how it should act in a given situation. The possibility that corporations may be bound by international human rights obligations is especially relevant here. If a domestic legal system imposes no similar human rights obligations but does impose an obligation to act for the best interests of the shareholders, the legal obligations of the corporation may be irreconcilable, leading to liability in one system or the other. Such a situation would evidently diminish the ability of either system to guide the behaviour of the actors it purports to regulate. The State imposition account manages this risk of inconsistency by requiring that international obligations are specifically addressed to DNSAs if they are to be binding upon them. This eliminates the risk of inconsistency between a domestic obligation and a generally addressed international obligation and forces States to consider any inconsistency between a domestic obligation and a specifically addressed international obligation. The possibility for inconsistency is further reduced by recognising the effect of the obligation on States to respect the sovereignty and regulatory authority of other States.

Finally, when deciding whether to impose obligations on DNSAs, States would be wise to consider whether any obligation is likely to be effective in practice. The existence of direct DNSA international obligations plainly does not depend on the degree to which they are followed or enforced, but lawmakers generally hope that the law they create will have some practical effect. Voluntary compliance is the most important way in which legal obligations have effect. There is no one single factor which determines whether or not a DNSA will comply with an international obligation, but the perceived legitimacy of any obligation is a concern which permeates much of the literature explored in Section 3. If DNSAs do not consider the relevant obligation to be ‘legitimate’, so the argument goes, they will be less likely to voluntarily comply and so the practical effect of the obligation will be diminished. Instead of relying on the ability of the legal obligation to guide the behaviour of DNSAs, it would be necessary to resort to dispute resolution and judicial remedies to encourage compliance and attempt to reverse any harm caused by the DNSA’s failure to comply. This concern with compliance appears to be particularly influential in the work of those such as Cassese who suggest that consent does have a role in explaining the international obligations of DNSAs.Footnote ¹¹⁴

Some of these legitimacy and compliance concerns may at least be partially addressed by recognising the legal limitations to States’ capacity to impose obligations on DNSAs discussed in Section 4.1, and following the other suggestions in this section. Obligations which protect a distinct international interest are more likely to be considered legitimate by the relevant DNSAs than those which do not. The same goes for international obligations which seek to complement rather than undermine or circumvent domestic legal obligations. Although a rejection of a DNSA consent requirement is a key feature of the State imposition account suggested in this article, this position does not mean that DNSAs cannot or should not have any input into the lawmaking process. By consulting with the DNSAs which will be bound by an international obligation, States will both be able to create a more effective obligation which takes account of the practical realities of the actors bound and foster a relationship of trust which is more likely to induce voluntary compliance. Consultations are already common in negotiations creating obligations upon States which affect DNSAs in some way, and there is no reason why this approach would not be equally appropriate in this context. This is not a silver bullet; there may be circumstances where it is impractical or inappropriate to consult DNSAs (such as where an obligation will bind a large number of individuals), or where DNSAs refuse to engage with the process or are obstructive. In such circumstances, States will have to weigh the importance of creating a DNSA international obligation against the potential of routine non-compliance. In some cases, DNSAs may refuse to voluntarily comply with their obligations despite having been a part of the lawmaking process. Yet such possibilities do not detract from the general point that, where it is possible, consulting DNSAs may help to bolster the legitimacy of the resulting international obligation and, consequently, the practical effectiveness of the obligation through voluntary compliance.

Even if DNSAs do voluntarily comply with their obligations, disagreements are likely to nevertheless arise and compliance is unlikely to be perfect. The practical effect of obligations will therefore often depend on the existence of a means of adjudication and enforcement. States considering imposing international obligations on DNSAs should therefore also consider what mechanisms are available to hold DNSAs to account for their breaches. Many international courts and tribunals only hear inter-State disputes, and those which do hear disputes involving DNSAs usually only consider claims against States and not DNSAs. Even where there is the possibility of enforcing DNSA obligations through a counterclaim, such as in investor-State proceedings,Footnote ¹¹⁵ the initiative remains with the DNSA to bring the initial claim and the decision of whether to bring the counterclaim rests with the State rather than the other actors directly affected by the alleged breach. Enforcement through domestic courts remains a possibility; however, courts willing to exercise universal jurisdiction for DNSA breaches of international law outside of international criminal law remain exceptional.

States wishing to impose international obligations on DNSAs will therefore need to reflect on whether the creation of a new adjudication and enforcement mechanism is possible to ensure the practical effectiveness of such obligations. The creation of a new court or tribunal is not impossible, as shown by the establishment of standing courts and tribunals such as the International Criminal Court. However, creating a new body is likely to be difficult. Establishing dispute settlement institutions and funding their operations on a long-term basis is expensive and is likely to be relatively low in the funding priorities of States and their populations.Footnote ¹¹⁶

The creation of any new international dispute settlement institution would be even more difficult if the relevant obligations are not considered legitimate by the broader community of States. The establishment and maintenance of any international institution is invariably a joint effort; in addition to the obvious question of funding, the compliance of other States is often needed to ensure the effectiveness of dispute resolution proceedings. This is relevant where the court issues interim or procedural orders, awards remedies or relies on States to ensure the appearance of a party in proceedings. If the relevant international obligation does not have widespread support or is considered to be illegitimate for some reason, other States are unlikely to provide the cooperation or assistance required for the effective practical implementation of the obligation. Some States may even take steps to shield DNSAs from the effects of any decision. States seeking to impose international obligations on DNSAs may therefore also need to consider whether the obligation will be effective in practice in light of the attitudes of other States.

It is important to note that just because it may be undesirable to impose legal obligations on DNSAs at the international level in certain circumstances, it does not necessarily mean that it is undesirable to impose any legal obligations on such actors. As highlighted in Sections 1 and 2.1, domestic law is available for imposing obligations on DNSAs where the creation of international obligations for such actors is considered inappropriate. Domestic legal systems generally do not run into the same problems of legitimacy when imposing obligations on DNSAs within their territory in pursuit of the public interest. Such legal systems are also generally better placed than international law to manage tensions between potentially conflicting legal obligations, such as the example of human rights interests and shareholder interests noted earlier in this section, particularly in light of the availability of both public and private law techniques.Footnote ¹¹⁷ Further, domestic systems generally have more effective means of ensuring compliance with legal obligations. This discussion is not intended to suggest that is never appropriate to impose international legal obligations on DNSAs; rather, it is intended to serve as a reminder that international law is not the only, or even the default, option for regulating DNSAs and that international and domestic law can and should act as complementary rather than competing systems of regulation.