3 - Invitations to Intervene after the Cold War

Summary

Aiming to move beyond the limited primary sources on which polarised debate is usually based, this chapter reviews new data on UN Security Council practice in response to consensual interventions. From 1990 to 2013, the Council passed resolutions on 76 per cent of all internal conflicts. This chapter evaluates that response in light of four leading theories: of the Court in Nicaragua, that governmental invitations are always valid; of the Institut de droit international (IDI), that pro-government interventions are ‘allowable’ until a conflict becomes a civil war; that intervention is allowable at the invitation of an elected ‘democratic’ government to secure or restore its power; and that it is allowable in response to an invitation to counter ‘terrorist’ threats. The data shows that the Council does not unequivocally support the Nicaragua or IDI views but has approved regularly the anti-terrorist, and occasionally the pro-democracy, views. Its active voice is more marked than its alignment with any one theory. Among other implications, the IDI view – a Cold War response to abuses of supposed invitations – may be less salient when a multilateral check on such abuses is available.

I. Introduction

The argument supporting a right to intervene at a government’s request would seem straightforward. The UN Charter endows a state with a virtually absolute right to bar outside forces from entering upon its territory. But a government, as the state’s agent, has broad discretion to exercise this right or not to do so – and, if not, to invite foreigners to assist in any actions the government could lawfully undertake itself. That choice is simply an example of states’ general ability to consent to actions otherwise considered unlawful. Consent – so the argument goes – precludes the wrongfulness of the foreigners’ presence and vitiates any violation of the state’s territorial integrity.

Despite this appealing logic, few scholars believe the claim accurately describes contemporary international law.¹ The debates are legion and wide-ranging. Some address doctrinal issues, such as the disagreement about how a regime’s consent to intervention operates: as conduct simply not prohibited by the primary norm against aggressive force, or as a circumstance precluding wrongfulness under the secondary norm of state responsibility.² Others concern how the consent theory relates to cognate principles of international law. For example, the prohibition on the use of military force across borders is widely recognised as a ius cogens norm, prohibiting contrary agreements;³ if that is correct, then – by definition – a state cannot consent to the violation of its own territorial integrity by inviting outside forces.⁴ But a broad ius cogens prohibition on consensual intervention is not widely accepted,⁵ and indeed the problem may be avoided altogether if one views lack of consent as an element of the primary norm against aggressive force, since then consent would not operate to set aside a purportedly absolute rule.

Another criticism drawing on a cognate doctrine concerns the interests to be furthered by an invitation. The consent theory is agnostic on the reasons why an invitation is issued and the goals to be sought by the intervening state. An internal conflict may involve a pure contest of power and not implicate any significant interests of international law apart from ending the human tragedy that attends any war. But other conflicts may implicate those objectives directly. A government may be challenged by rebels tied to transnational terrorist networks.⁶ Or a legitimately elected leader may be ousted from power and take up arms to regain her office.⁷ Or an insurgency may challenge a regime engaged in mass human rights violations.⁸ All these scenarios raise the question of whether international law should be concerned not with the simple fact of intervention by invitation but with the reasons for intervention.⁹ The consent theory set out above does not take such substantive objectives into account.

A third set of critiques focus on the policy consequences of consent theory. One argues that a party to an internal conflict is most likely to seek assistance at precisely the time outside intervention in national politics is least desirable. Governments and opposition parties to an internal conflict need help when they are unable to prevail on their own. But such weakness, especially on the part of incumbent governments, evidences a deep division among citizens on fundamental questions of national policy and leadership.¹⁰ Should the balance in such circumstances be tipped by outside forces? Or should the principle of internal self-determination protect the citizenry as a whole from intrusions on their political autonomy?¹¹ Relatedly, if external assistance enhances the strength of the weaker party, then the addition of foreign forces is likely to prolong the conflict.¹² Some argue that civil wars fought to conclusion solely among national parties are both shorter and less likely to reoccur.¹³

A final critique, cutting across all categories, is that a rule permitting consensual intervention is ripe for abuse.¹⁴ Requests for assistance may come after the fact or, as happened during the Cold War, from groups created specifically for the purposes of issuing invitations.¹⁵ In some conflicts, it may be difficult to tell which faction is in effective control of the state and thus, under traditional doctrine, empowered to issue an invitation.¹⁶ A solution that relies on the extent to which one faction or another is recognised by other states only replicates the problem in another legal domain, since an intervening state will almost certainly recognise the faction issuing the invitation as the target state’s legitimate government.

If there is unity in the criticism of consensual intervention – or at least the version described above – there is little in describing the actual content of contemporary international law.¹⁷ Primary sources, state practice, and scholarship on this topic have been described as ‘scattered and incomplete,’¹⁸ ‘profoundly divided,’¹⁹ ‘all over the map,’²⁰ ‘a tangle of opinions’,²¹ and providing ‘no conclusive guidance’.²² How is one to find coherence, for example, among:

• statements in three UN General Assembly resolutions prohibiting ‘interference in civil strife in another State’ and ‘assisting or participating in acts of civil strife’;²³

• the UN Security Council’s proclamation of ‘the inherent and lawful right of every State, in the exercise of its sovereignty, to request assistance from any other State or group of States’;²⁴ and

• the unqualified statement of the International Court of Justice (ICJ) that intervention is ‘allowable at the request of the government of a State’?²⁵

The lessons of recent state practice are particularly contested, and one finds authors citing the same practice to support completely opposite views of the law.²⁶ In 2011, the Institut de droit international (IDI) decided to revisit its 1975 Wiesbaden Resolution III, which famously forbade assistance to governments engaged in civil wars.²⁷ But divisions among IDI members effectively led to its paralysis and the new 2011 Rhodes Resolution II did not even address the permissibility of intervention in civil wars.²⁸

One steps gingerly into such swirling waters. There seems little point in another effort to harmonise the small set of canonical sources that have so far defied synthesis. And an effort to find coherence among scholars – among whom disagreement abounds – seems even more futile. Instead, this chapter will focus on two aspects of consensual interventions that recent scholarship has not addressed at length. First, it will ask whether any of the various theories of consent find support in a comprehensive assessment of post-Cold War practice. To my knowledge, no effort has been made to compile all examples of consensual intervention after the end of the Cold War and examine systematically how the United Nations, regional organisations, and leading states have reacted. The discussion of this practice will rely on a new dataset compiled for this purpose.

Second, the chapter will focus particular attention on the practice of the UN Security Council. The data reveal that the Council has issued resolutions or presidential statements on an overwhelming proportion (82 per cent) of consensual interventions since 1990. Many scholars have focused on the international community’s inability to agree on factual aspects of contested interventions. These include whether an invitation was in fact issued, whether the inviting party exercised effective control over a state, and whether a conflict had reached the level of a ‘civil war’. But controversies over these factual predicates for a valid invitation are rendered largely irrelevant through collective determinations by the Council, which enjoys an authority to characterise legally significant facts and to distinguish between lawful and unlawful uses of force.

The Council’s heretofore unexamined record on consensual interventions also provides important support for this chapter’s particular focus on the post-Cold War era. The view that governments cannot invite outside forces to assist in a civil war came to prominence in the mid-to-late Cold War, building on older belligerency doctrine,²⁹ but drawing new support from a series of General Assembly resolutions on the right to internal self-determination and two widely cited secondary sources: the IDI’s 1975 Wiesbaden Resolution III and Louise Doswald-Beck’s extraordinary 1986 article in the British Yearbook of International Law.³⁰ Both the IDI and, in particular, Doswald-Beck proceeded from the then widely shared (and objectively correct) observation that – as a matter of fact, not law – no international organisation – especially the Security Council – could effectively review the factual basis for claimed consensual interventions. The very reason why consensual intervention was highly problematic in practice – the polarised camps in the Cold War made a series of dubious claims of invitation to secure and further their spheres of influence – was also the reason why international organisations were paralysed and unable to react.³¹ The IDI and Doswald-Beck advocated a strong prophylactic rule intended to reach the most consequential interventions (when the government feared it might be losing a civil war) and render them all unlawful, regardless of the purported justification. The need for ‘objective’ review by an international organisation was thereby diminished.

But we are no longer in an era in which the Security Council and other international organisations are mere peripheral players in legal determinations on the use of force.³² To the contrary, the Security Council now regularly addresses almost all non-international armed conflicts (NIACs), long the most prevalent form of conflict.³³ From 1990 to 2013, the Council passed resolutions on 76 per cent of all NIACs, increasing to 80 per cent for conflicts that began after 1990.³⁴ Further, the Council aggressively shaped the legal framework for ending and remediating NIACs by imposing a series of binding legal obligations on the conflict parties that, in some cases, deviated from existing international law.³⁵ Of course, the Council has remained deadlocked on Crimea and Syria, two widely discussed cases that share many attributes of superpower interventions during the Cold War. But the data show these to be a distinct minority. While the Council’s reactions in the majority of cases have varied, the critical point is that, in most cases, the Council faces few political obstacles to engagement. As a result, not only has Council paralysis ended in responding to NIACs, but also the Council has become aggressive and omnipresent.

The new Council activism vastly complicates the assumptions of Cold War-era doctrine on consensual interventions. If the Council is, in theory, available to pass judgment on the legality of interventions, is the prophylactic rule advocated by the IDI and Doswald-Beck still necessary? One may well answer ‘yes’ if a conflict involves one or more of the five permanent members of the Council or their close allies. But the data will show that these are in a distinct minority, meaning that whatever rule exists in customary law, the Council can override it in the majority of conflicts. The utility of a rule that actually ends up governing few conflicts is certainly open to question.

Further, if the Council regularly approves certain types of request for assistance but not others, should we not only reassess a general prohibition on interventions in civil wars but also ask whether customary international law should now be understood as allowing, or even favouring, interventions to accomplish particular goals? Post-Cold War international law has coalesced around a series of policy objectives unknown in the prior era. Collective action against terrorist groups and the promotion of democratic elections are two examples, both of which find support in Council practice. Has customary law, so understood, moved towards supporting interventions in furtherance of those policies?

In sum, Security Council activism in the post-Cold War era makes examining its record essential to understanding the content of international law. While this chapter will review older practice and scholarship for the purposes of understanding the various theories contending for pre-eminence, it will focus most of its attention on new Council practice and the conclusions to be drawn from aggregated data about that practice.

After first discussing in more detail the reasons for focusing exclusively on the post-Cold War era, this chapter will review the major theories on intervention by invitation (section II). After explaining the methodology used in collecting and sorting the data (section III), it will then test each of these theories by asking whether the record of Security Council reactions supports or negates each theory or does not point in either direction (sections IV–V). Finally, it will ask how the new Council practice should be seen as affecting international law (sections VI and VII): as a self-enclosed lex specialis, or as evidence of customary international law of a particularly useful kind?

II. The State of Debate

One can discern three critical periods in the Charter era during which doctrine on intervention by invitation evolved in tandem with the legal and political landscape of the times: (i) the mid to late-Cold War era; (ii) the post-Cold War era up to the 9/11 terror attacks; and (iii) the post-9/11 period. While this typology may be inexact at the margins, each period is readily identified with political developments that spurred legal innovations. Understanding how each doctrinal shift corresponded to these specific challenges is important, since there is a tendency in the literature to present the current lack of legal clarity as an inexplicable set of contradictions – a rift between groups of states with competing interests, or simply an unfortunate failure of political will. Each of the four major theories identified in the literature – discussed in this section and then tested against international reaction – bears a clear mark of the era in which it emerged.

Two preliminary observations are necessary. First, the four major theories described are not equal in terms of the scope of actions they encompass; rather, they are partially overlapping. The final three involve circumstances that represent subsets of the conditions encompassed by the first. Their relationship is illustrated in Figure 3.1.

• The first, the ‘Nicaragua view’,³⁶ which permits invitations by governments in all cases, is the broadest theory.

• The second, the ‘IDI view’, addresses a subset of cases permitted by Nicaragua in which a government invites outside forces into a conflict that has not yet reached the level of a ‘civil war’. Invitations to intervene in civil wars are thus encompassed (and permitted) by the Nicaragua view but not the IDI view.

• The third, the ‘democratic legitimacy view’, permits interventions only when requested by governments or rebel groups that have won an election verified as free and fair by credible international actors. This is best seen as another subset of the Nicaragua view since, as I will argue, the civil war threshold imposed by the IDI view is not consistent with the fully legitimising nature of an invitation from an elected government. However, the democratic legitimacy view covers fewer cases than the IDI view.

• Finally, the ‘anti-terrorism view’ also represents a small subset of the Nicaragua view, but that subset is different from those captured by the IDI view. The IDI view permits interventions in domestic conflicts not rising to the level of a civil war, such as riots or widespread criminal activity. While the anti-terrorism view also does not legitimise intervention in a civil war either, the reason is different: the terrorist groups involved are almost always transnational in nature. Their challenge to governments does not involve the fracturing of domestic political opinion that is the hallmark of a civil war.

The second observation is that the descriptions of the various theories here provided are not intended to be comprehensive. Such efforts have been ably undertaken by other scholars. Instead, the descriptions are designed only to highlight the central question raised by the data on international reaction to modern invitations: do the theories’ underlying assumptions and circumstances of origin remain relevant to contemporary state and international organisation practice?

Figure 3.1. Interrelation between Major Theories on Consensual Intervention

A. The Cold War Setting

The fracturing of international law on invitations to intervene into various doctrinal and policy-driven schools did not emerge in a vacuum; rather, it followed a period during the mid-to-late Cold War in which scholars struggled to find coherence in shifting normative currents.³⁷ Some writers invoked pre-Charter doctrine as effectively unchanged – in particular, the rules of belligerency.³⁸ According to Wolfgang Friedman, writing in 1965, ‘[w]hat is probably still the prevailing view is that the incumbent government, but not insurgents, has the right to ask for assistance from foreign governments, at least as long as insurgents are not recognised as “belligerents” or “insurgents”’.³⁹ Others recognised that belligerency doctrine had fallen into desuetude and that the UN Charter pointed towards a collective response to internal conflicts through decisions of the Security Council, but they bemoaned its dysfunction.⁴⁰ Still others focused primarily on the stark disconnect between prohibitions of intervention on either the government or rebel side and the avalanche of interventions practised by the competing East–West camps.⁴¹ Finally, many found, in the recent rise of decolonisation and norms of self-determination, fresh justifications for the prohibitory approach of belligerency doctrine, but they expressed caution about their salience in the face of so much contrary state practice.⁴²

The scholarship of this era was hesitant, uncertain about how to reconcile the cacophony of new and old norms and state practice that could arguably support multiple positions.⁴³ Many concluded their review of this unhappy situation with suggestions that law should follow wise policy and minimise the spread of violence by replicating the old belligerency rules, albeit without the futile requirement of recognition.⁴⁴

B. The Nicaragua View

The first view I will test using Security Council practice arose during the late Cold War era. It categorically favours governmental requests for assistance over those of opposition groups.⁴⁵ This view is most closely associated with a brief passage in the ICJ’s 1986 Nicaragua decision:

[T]he principle of non-intervention derives from customary international law. It would certainly lose its effectiveness as a principle of law if intervention were to be justified by a mere request for assistance made by an opposition group in another State – supposing such a request to have actually been made by an opposition to the regime in Nicaragua in this instance. Indeed. It is difficult to see what would remain of the principle of non-intervention in international law if intervention which is already allowable at the request of the government of a State, were also to be allowed at the request of the opposition. This would permit any State to intervene at any moment in the interna1 affairs of another State, whether at the request of the government or at the request of its opposition. Such a situation does not in the Court’s view correspond to the present state of international law.⁴⁶

The passage is famously cryptic and there are good reasons why its normative value might be seen as limited.⁴⁷ Yet when the Court had an opportunity to clarify the Nicaragua language almost twenty years later in its Congo/Uganda case, it failed to do so.⁴⁸ The question in that case was whether, at certain points in time, Ugandan troops were lawfully present on Congolese territory. The Court found that, in an early period, the Democratic Republic of the Congo had failed to object to the troops’ presence and thus effectively gave its consent, which it could have withdrawn at any time.⁴⁹ The Court thoroughly reviewed the record of negotiation and the agreement between the two parties, and it determined that a later modus operandi for the withdrawal of Ugandan troops did not embody consent to their continued presence. It concluded that Congo had in fact withdrawn its consent.⁵⁰ The Congo/Uganda opinion thus assumed the validity of a governmental invitation in the same underanalysed manner as Nicaragua.

The Nicaragua view is grounded in the idea that the ius ad bellum of UN Charter Article 2(4) and its progeny exist to secure states against coercive intrusions upon their political independence and territorial integrity. Neither, this view argues, is infringed by a consensual intervention.⁵¹ This view is echoed in language in the General Assembly’s 1970 Friendly Relations Declaration⁵² and its 1974 Definition of Aggression,⁵³ both of which juxtapose the consensual and non-consensual presence of foreign troops, singling out the latter as impermissible.⁵⁴ The legitimacy of consent would also seem to underlie every bilateral and multilateral status-of-forces agreement, which remain in force during civil wars, as well as UN and regional peacekeeping missions that are frequently based (at least in part) on governmental consent and which are either initially sent to states in which civil wars are ongoing or remain in place after civil wars break out.⁵⁵

Finally, the Nicaragua view is arguably consistent with the general role of consent in state responsibility law as a circumstance precluding wrongfulness of state action.⁵⁶ The requirements that consent be given authoritatively by a state, that it be given freely, and that the acts in question stay within the bounds of the consent are reflected in Article 20 of the International Law Commission (ILC) Articles on State Responsibility.⁵⁷ A host of questions specific to consent to the use of force are widely debated and will be addressed in this chapter – but few argue that these principles of consent do not apply as a general matter to the use of force.⁵⁸

Given the categorical phrasing of the General Assembly resolutions and the two ICJ decisions, the Nicaragua view has come to connote a blanket approval of governmental invitations and a blanket disapproval of invitations from opposition groups.⁵⁹ Three doctrinal consequences would seem to follow. First, the Nicaragua view is incompatible with both pre-Charter belligerency doctrine and the IDI view, both of which hold that when a conflict reaches a certain intensity threshold, intervention at the request of the government is prohibited.⁶⁰ Some argue that because the Court in Nicaragua was concerned, first and foremost, with non-intervention doctrine, its language referred to invitations in general and not specifically to civil wars, which would have required it to address the existence, or not, of an intensity threshold. But this view is hard to square with the Court’s sweeping language, not to mention its failure to limit that language in the Congo/Uganda case.⁶¹ Moreover, just twelve pages earlier, the Court had found that hostilities between the contras and the Nicaraguan government amounted to a NIAC.⁶² It seems unlikely that the Court would have crafted the rest of its opinion without regard to its prior legal characterisation of the conflict.

Second, the Nicaragua view is incompatible with limitations on consent based on the argument that an inviting or invited state is pursuing an unlawful objective. Some have suggested that a request to assist in committing widespread human rights abuses would contravene the legal obligations of both states.⁶³ The request would be incompatible with the requesting state’s obligations to its own citizens under customary and treaty-based human rights norms. For the invited state, two arguments are possible: that it would incur responsibility for assisting in the inviting state’s violation of the latter’s human rights obligations within its own territory; or that the invited state’s own human rights obligations would apply extraterritorially.⁶⁴ Such qualitative limitations on the reach of state consent are compelling, but they would need to be grounded in sources other than the Nicaragua opinion itself, which contains no limitation on invitations based on the objective the invited state would seek to achieve.⁶⁵

Third, Nicaragua has an uncertain relationship with norms more directly concerned with the legitimacy of inviting governments. The categorical nature of the Nicaragua view does not necessarily validate all invitations issued by ‘the government’. The not-uncommon scenario in which competing factions claim to speak for a state during a civil war was one of the critical factors underlying the belligerency doctrine, as well as the IDI view (discussed next). The Nicaragua view does not itself resolve competition among different would-be governing factions, since the United States simply did not raise the question of governmental legitimacy before the ICJ nor did the Court address the issue in its brief aside. To adopt the Nicaragua view, then, is simply to default to an entirely separate set of norms on questions of governmental legitimacy. Much scholarship on intervention by invitation takes a significant detour to discuss this question. Whether a specific theory of governmental legitimacy might apply to certain invitations is the subject of section D. But Nicaragua itself should be understood as agnostic on the subject.

Can one ascribe a particular worldview or historical provenance to the Nicaragua view? Given the opinion’s extraordinary brevity, one can only speculate. But in light of (i) the ICJ’s traditional aversion to advancing the law in bold leaps, (ii) the relatively new and unsettled nature in 1986 of the self-determination limitation on intervention in civil wars, and (iii) the Court’s determination that the Nicaragua conflict constituted a NIAC, one could argue the following. The Court was addressing the permissibility of invitations by rebel groups, which it wanted to reject in no uncertain terms. How better to make clear the destabilising and unacceptably intrusive nature of invitations to rebels than to place them in stark contrast to the right of the government to invite assistance? If one assumes the Court was well aware that many NIACs during the Cold War were internationalised by assistance to rebels, one might understand its language as an effort to address rampant interventionism without departing in any significant way from existing law. The Court could well have understood the law of the time not to have absorbed the self-determination view but to have left belligerency doctrine behind. That combination would result in no extant limit on interventions based on a civil war threshold. In addition, even Louise Doswald-Beck (whose article was published in the same year as the Nicaragua judgment) believed that international law accorded a presumption of continuity to governments that had lost territory to rebels.⁶⁶ So a rule (such as Nicaragua) containing no civil war threshold might, in practice, produce results only marginally different from a situation in which a government did not lose its authority to act for the state even when rebels controlled significant portions of territory. In other words, the categorical Nicaragua approach might not, in fact, permit more interventions supporting governments than did existing law on recognition. Both Nicaragua and international law on recognition in the 1980s would continue to view a government as legitimate (and thus empowered to invite foreign forces) even when it was in conflict with a well-organised rebel force in control of substantial territory.

C. The IDI View

The second view involves a full prohibition of intervention in civil wars. Once a conflict becomes a ‘civil war’, the government joins opposition groups in being unable to invite external assistance. This view encompasses a subset of cases captured by the Nicaragua view by excluding those invitations by governments that Nicaragua would permit. This view was most famously articulated in the IDI’s 1975 Wiesbaden Resolution III: ‘Third States shall refrain from giving assistance to parties to a civil war which is being fought in the territory of another State.’⁶⁷ I shall refer to this as the ‘IDI view’.

While the IDI view shared much in common with traditional belligerency doctrine, which required third-party neutrality when rebels reached a certain level of organisation and territorial control, it emerged from a new set of concerns. Coming to fruition in the mid-Cold War period, the IDI view arose in reaction to the superpowers’ use of invitations as a pretext under which to maintain or expand their spheres of influence in the developing world. The conflicts in which ‘invitations’ were cited as justification were seen by the superpowers as zero-sum contests: either an allied government was threatened by opponents seen as sympathetic to the other superpower, in which case the country might be ‘lost’, or a government allied with the other side was threatened by ideologically sympathetic opponents, in which case the country might be ‘gained’. In either case, the niceties of an invitation to defend the government were decidedly secondary to the perceived need to intervene and ‘save’ the state for one’s own side.⁶⁸ Invitations not infrequently came from groups not actually exercising any governmental authority.⁶⁹ Some such groups were created for the specific purpose of issuing an invitation.⁷⁰ Some issued invitations after the fact.⁷¹ Some invitations were coerced.⁷² The point of creating the appearance of consent was to align the purpose of the intervention with ‘the will of the people’. External support thus became not a break with legitimate government but support for the true ‘legitimate authority’, whose claim to a popular mandate was somehow superior to that of the regime being ousted. Of course, such claims of legitimacy were purely instrumental.

Given the zero-sum terms in which the Cold War protagonists viewed internal conflicts, as well as the lengths they were willing to go to ensure favourable outcomes, a rule that permitted interventions at the invitation of the government was doomed to ineffectiveness. Requirements of effective control and issuance by appropriate authorities simply led to the elaborate fictions noted above.⁷³ Continuing to permit consensual interventions in those circumstances would end up undermining a value it purported to protect: a ‘legitimate’ government’s freedom to control the presence of foreign troops on its territory. What was needed was a rule that prohibited interventions by invitation once it was clear that civil authority in a state had broken down or was imminently threatened – that is, when the Cold War camp aligned with that authority was most likely to intervene.

The rise of the right to self-determination in the late 1950s – culminating in its codification in common Art. 1 of the United Nations’ International Covenants on Civil and Political Rights (ICCPR), and on Economic, Social, and Cultural Rights (ICESCR) in 1966 – provided an important doctrinal foundation for this view.⁷⁴ First and foremost, self-determination was the legal vehicle facilitating decolonisation. But, in the view of newly independent states (and many others), external independence was hardly adequate to protect autonomous political decision-making. Continued interference by former colonial powers and Cold War antagonists deprived citizens of the ability to choose their own political direction. That choice might be manifest in the relatively orderly conduct of elections or in the violent outcome of a civil war. However citizens manifested their choice, it was theirs to make. What, after all, is the point of joining the ranks of autonomous states if the most fundamental act of sovereignty – choosing the national leadership – is influenced or indeed fully determined by outsiders?

It is important to emphasise that the IDI view protected the opportunity for national choice of regimes rather than any actual choice. The idea that non-democratic means of choosing a government might nonetheless represent a legitimate choice by citizens sits uneasily with international law’s contemporary emphasis on free and fair elections. But the argument by proponents of the IDI view stressed ‘the absence of outside interference rather than the quality of internal government’.⁷⁵

The self-determination rationale for prohibiting assistance to governments in civil wars also created a useful symmetry with the wholly non-controversial prohibition of assisting rebel groups: ‘Once a considerable [number] of people starts a civil rebellion in an attempt to change its political status, intervening from the outside on the government’s side would mean meddling in that State’s internal affairs as much as helping the rebels.’⁷⁶

But not all instances of unrest are manifestations of discontent with an incumbent government. Riots, other kinds of low-level disturbances, or widespread criminal activity do not necessarily indicate a fundamental rift in the body politic. Proponents of the IDI view thus mirrored belligerency doctrine – which they acknowledged had fallen into desuetude⁷⁷ – by imposing a threshold of a civil war.⁷⁸ The Wiesbaden Resolution III itself provided that its prohibition did not apply to ‘local disorders or riots’ but rather required armed conflict not of an international character for control of the state.⁷⁹

Once that level of organised and sustained violence was reached, the government lost its capacity to invite outside forces.⁸⁰ Stated another way, up to the point of civil war, a government could validly represent the state externally for purposes of issuing an invitation; past that point, the interests of the government and the interests of the people were deemed to be presumptively at odds. While a government generally enjoys a presumption of continued legitimacy even when its effective control is diminished, once a civil war commences, the government loses its ability to subordinate the interest of (again potentially) a majority of its people to its own interest in survival.⁸¹

The genesis for the self-determination rationale lay in a series of General Assembly resolutions passed between 1965 and 1981.⁸² Each of these resolutions articulated a prohibition of intervention in states’ internal affairs, including interference in ‘civil strife’.⁸³ Each also grounded this prohibition in every state’s ‘inalienable right to choose its political, economic, social and cultural systems, without interference in any form by another State’. The Friendly Relations Declaration of 1970 most famously linked these non-interference principles to the right of self-determination: ‘By virtue of the principle of equal rights and self-determination of peoples enshrined in the Chapter of the UN, all peoples have the right to determine, without external interference, their political status.’⁸⁴

The Wiesbaden Resolution III – in opaque, but hardly obscure, language – also cited pervasive Cold War realities as justification: the zero-sum way in which the superpowers viewed civil wars and the reality that intervention on one side inevitably led to counter-intervention on the other.⁸⁵ While the IDI’s members were far from unanimous and Special Rapporteur Dietrich Schindler expressed doubt that the Resolution accurately reflected settled law,⁸⁶ it has acquired a semi-authoritative status. The IDI view also found resonance in a widely cited report, commissioned by the European Union, of the Independent International Fact-Finding Mission on the Conflict in Georgia (IFFMCG). After reviewing the canonical sources and acknowledging their uncertainty, the report articulated a ‘negative equality’ principle that parallels the IDI view: ‘[A] military intervention by a third state in a state torn by civil war will always remain an illegal use of force, which cannot be justified by an invitation.’⁸⁷ Like Cold War-era sources, the IFFMCG justifies its position as a response to self-interested interventions. Negative equality, it asserts, ‘removes the pretext of “invitation” relied on by third states in order to camouflage interventions motivated by their own policy objectives’ and ‘relieves lawyers of the difficult task of identifying and proving a valid invitation’.⁸⁸

D. The Democratic Legitimacy View

The third theory posits that principles of democratic legitimacy should play a limited, but significant, role in evaluating the lawfulness of invitations.⁸⁹ The theory is limited because it is restricted to cases in which one party (usually the opposition) claims an electoral mandate to govern but is prevented from taking office or is ousted from office. The theory is nonetheless significant because traditional international law emphatically rejected democratic legitimacy criteria in favour of the effective control doctrine in evaluating a regime’s capacity to issue invitations.⁹⁰ The new theory does precisely the opposite.

Advocates of the democratic legitimacy approach – and they are few – ground their claim in the pervasiveness of democratic principles in international law after the end of the Cold War. Three trends are particularly relevant. The first is the pervasiveness of election monitoring. As Susan Hyde reports, as of 2011, ‘80% of all national elections are now monitored’ by international observers.⁹¹ The chances that any given regime’s claim to democratic legitimacy can be empirically validated are thus substantially higher than they were in the Cold War period. The second is the modest, but real, impact that principles of democratic legitimacy have had on state practice in the recognition of states and of governments.⁹² Regional organisations such as the Organization of American States (OAS), the European Union, the African Union, the Southern Common Market (Mercado Común del Sur, or Mercosur), and the Economic Community of West African States (ECOWAS) have established ‘democracy protection’ mechanisms that permit the collective non-recognition of regimes that depose or otherwise interrupt elected governments.⁹³

These regional mechanisms complement election monitoring in two ways. First, they seek prospectively to ensure the stability of elected governments after they take office by threatening to sanction anti-democratic actors who undermine or overthrow those governments.⁹⁴ In other words, they address anti-democratic events well beyond the election itself. Second, they provide collective judgments on when democratic governance has been interrupted. The question of when a regime, once elected, loses its ‘democratic’ character has remained highly controversial.⁹⁵ Military coups are obvious cases, but what of suspending civil liberties, removing judges, dissolving the legislature or closing opposition media outlets? If international law is now expressing a preference for ‘democratic’ governments, then whether one characterises any or all of these actions as ‘non-democratic’ takes on great significance. The regional regimes avoid the cacophony of individual states answering these questions by providing for collective determinations, undertaken by bodies such as the African Union Peace and Security Council or the OAS Permanent Council.⁹⁶ Just as election monitoring is intended to move the question of a new government’s entitlement to hold power from the domestic to the international realm, the democracy protection regimes similarly internationalise the question of an elected regime’s ongoing democratic bona fides.

The third trend underlying the democratic legitimacy view is the practice of UN-sponsored post-conflict missions to states emerging from NIACs, which have consistently emphasised the importance of elections, human rights, and other democratic principles in the new institutions they help to establish.⁹⁷ The Security Council has unanimously approved most of these missions.

This ascension of democratic legitimacy criteria inevitably leads to the following question: why should the legitimacy of an elected regime not include a capacity to invite foreign forces to uphold an electoral outcome?⁹⁸ If the effective control principle was challenged in other contexts, why not here? Sceptics of the democratic legitimacy approach are probably correct that these trends do not supersede some, or even all, competing values in evaluating invitations to intervene.⁹⁹ Glaring failures to challenge coups in cases such as Egypt (2013) or Thailand (2014), as well as the recent ‘democratic recession’, are common critiques. But democracy is a sufficiently significant presence in Security Council practices, in particular, to justify their empirical study.

The specific role accorded to democratic legitimacy depends on the nature of each case and how one understands the reach of the ‘democratic entitlement’ in international law more broadly.¹⁰⁰ There may be interventions in which democratic principles play no role whatever, such as where both an incumbent regime and opposition groups lack an electoral mandate, and the opposition group makes no promise of democracy once in power. From that ‘democratic vacuum’, one can imagine a continuum of increasingly well-grounded claims of democratic legitimacy on which an invitation to intervene might be based:

1. (i) one side in a conflict was elected at some point in the past without international monitors;

2. (ii) one side was elected recently, with international monitors certifying the process as free and fair; and

3. (iii) not only did one side win an internationally monitored election, but also international organisations – perhaps including the Security Council – affirmed their support for that side as the legitimate government.

Which of these scenarios should qualify as validating an invitation because it furthers the democratic legitimacy view?

In addition, the question of when a government may issue an invitation is separate from the question of who qualifies as the government of a state for the purposes of issuing an invitation. Democratic legitimacy criteria clearly answer the latter question but not necessarily the former. The questions of ‘who’ and ‘when’ could interact in several different ways. In each instance, let us assume the paradigmatic case of an elected regime ousted in a military coup. That regime then takes up arms to regain power. The conflict then becomes a civil war. Therefore, the regime invites in foreign forces.

• First, the law could contain a civil war intensity threshold but (unlike under the IDI view) find it satisfied by the electoral mandate. In other words, the electoral mandate resolves the question of whether the state is actually divided over its future leadership, which was the basis for the IDI/self-determination rule of abstention during civil wars. This approach would thus conclude that electoral mandate both qualifies the regime as the legitimate government and endows it with authority to invite foreign support, despite the conflict having crossed the civil war threshold.

• Second, the democratic legitimacy view, paired with the Nicaragua view, would also permit assistance to the ousted elected regime. This is because no intensity threshold would be imposed, thus making the existence of a civil war irrelevant to the validity of the invitation. The democratic legitimacy view would designate the ousted group as ‘the government’ for the purposes of issuing an invitation, despite it not exercising actual power.

• Third, the elected, but ousted, group would qualify as the ‘government’ but, pursuant to the IDI view, still not be permitted to issue an invitation. This view would emphasise that internal self-determination is not equivalent to democratic choice and instead functions as a shield for states to resolve their internal disputes by any and all means, free from outside influence. That this particular dispute was resolved by the forceful removal of an elected regime is of no consequence.

Any one of these modes of integrating democracy criteria is plausible. One might look for guidance in the few cases in which they have been invoked – namely, the ECOWAS interventions in Liberia in 1992, the ECOWAS intervention in Sierra Leone in 1998, and the threatened ECOWAS intervention in The Gambia in 2017.¹⁰¹ But these cases are so factually distinct from one another that one cannot imply a common legal template for the use of a democracy justification. In Liberia in 1992, besieged President Samuel Doe consented to the ECOWAS intervention in the midst of a NIAC, but his democratic bona fides were highly questionable.¹⁰² In Sierra Leone in 1998, the Security Council praised the ECOWAS intervention that restored elected President Ahmed Tejan Kabbah to power.¹⁰³ A NIAC was in progress at the time of the intervention, and the Council had previously denounced a military coup that deposed the elected Kabbah government and called for that government to be restored.¹⁰⁴ President Kabbah appealed from exile to the chair of ECOWAS for assistance.¹⁰⁵ After the ECOWAS action, the Council issued a presidential statement welcoming ‘the fact that the rule of the military junta has been brought to an end’, and commended ‘the important role’ of ECOWAS.¹⁰⁶ Some have questioned whether the Council statement amounted to an ex post ratification of the ECOWAS intervention.¹⁰⁷

Finally, in The Gambia in 2017, Adama Barrow defeated long-time President Yahya Jammeh in an election that Jammeh initially conceded but later denounced, refusing to leave office.¹⁰⁸ Barrow, after somehow being sworn into office in the Gambian Embassy in Senegal on 17 January 2017, asked the United Nations, the African Union, and ECOWAS for assistance in taking office.¹⁰⁹ On 19 January, the Security Council adopted a resolution condemning Jammeh’s refusal to leave office and urging respect for the electoral results, urging ‘all Gambian parties and stakeholders to respect the will of the people and the outcome of the election which recognised Adama Barrow as the President-elect of the Gambia’.¹¹⁰ The Council tied its conclusions to similar determinations by regional international organisations, endorsing ‘the decisions of ECOWAS and the African Union to recognise Mr. Adama Barrow as president of the Gambia’.¹¹¹ The Council did not, however, authorise the use of force. With ECOWAS troops massing in Senegal on the Gambian border, Jammeh left the country on 21 January. The significant factors in the Gambian case that differ from those of Liberia and Sierra Leone are that no foreign troops actually entered the territory and the person issuing the invitation had never actually held power.

Despite this ambiguity in state practice – essentially a problem of too many variables to support a one-size-fits-all rule – one can make a strong argument for a version of the first option above. This is the claim that a democratically legitimate regime ousted from power can still issue a valid invitation despite the existence of a NIAC. The central objections to this view fall into two categories. The first is doctrinal – that the international law of recognition of governments still favours the effective control test, even if that support has weakened in the post-Cold War era; the second is normative – during a NIAC, no outsider can presume to judge which competing faction should be permitted to entrench itself in power by inviting outside forces, and external efforts to designate one faction as legitimate are likely to be more subjectively political than empirically objective. Both of these critiques originated prior to the end of the Cold War and the rise of democratic legitimacy criteria.¹¹² More importantly, neither of the critiques appears valid where an international organisation – often the Security Council, but also a regional organisation – determines that an electoral result entitles one faction to hold power and the other not. The first objection, relying on the effective control test, would have to argue that such a collective determination of legitimacy would be valid for all purposes except inviting outside forces. Prior such determinations by international organisations contain no such distinction.¹¹³ On what basis could the winning faction, for example, be entitled to appoint ambassadors, enter into treaties or exercise diplomatic protection, but not consent to the use of force on its territory? The entire purpose of multilateral validation of one faction’s entitlement to rule is to grant it exclusive access to all aspects of the state’s sovereign prerogatives.

The second critique – that outsiders simply cannot presume to judge the legitimacy or illegitimacy of competing national factions – is simply of no consequence if an international organisation has already observed an election and determined the winner. The intervening state does not make its own subjective determination that the inviting regime is democratically legitimate; it simply acts on a prior determination by an international organisation to that effect.¹¹⁴

This multilateral component is, of course, critical. Such validations would (hopefully) remove the feared politicisation of a recognition decision. Cases with no involvement by an international organisation would be more susceptible to Cold War critiques. But in cases addressed by the Security Council and/or regional ‘democracy protection regimes’ – under which member states agree in advance to non-recognition and sanctions where democratic government is interrupted – recognition would be less political and a matter of legal obligation.¹¹⁵ That the losing faction resists such a determination and begins an armed resistance is not a reason for the international organisation to retract its legitimacy determination.¹¹⁶

If the democratic legitimacy view is thus seen as not subject to the IDI view’s exclusion of invitations issued in civil wars, it represents a subset of cases coming within the Nicaragua view. That subset would likely include some cases of civil war and thus, as shown in Figure 3.1, the democratic legitimacy view overlaps with in both the IDI and Nicaragua categories.

E. Anti-Terrorist Operations

The final view holds that invitations to assist governments in conflict with transnational terrorist groups are legitimate in all cases.¹¹⁷ This claim is perhaps the least controversial of the four presented.¹¹⁸ While there is some question as to whether counter-terrorism was the sole reason for some interventions in the dataset, there is little, if any, evidence of state reaction against the legitimacy of counter-terrorist intervention.

The terrorism view is primarily asserted as an exception to the IDI view. Unlike rebel groups representing some portion of a state’s citizens dissatisfied with their government, terrorist groups frequently count foreign fighters among their ranks and operate across different states simultaneously.¹¹⁹ As a result, ‘terrorist groups cannot be regarded as a “People”, denying any claim to the right of self-determination’.¹²⁰ The IDI view, seeking to secure the integrity of autonomous political decision-making within states, is unaffected by assistance to governments in conflict with groups not part of the national body politic.¹²¹

A rule permitting counter-terrorist interventions confronts the common problem of the term’s lack of a clear definition.¹²² The malleability and highly political nature of a ‘terrorist’ designation presents the obvious danger of incentivising governments to label their civil war opponents ‘terrorists’ to legitimise external assistance. Definitional ambiguity also creates problems for coding: if one were to count as relevant state practice all cases in which the inviting state designated its opponents ‘terrorists’, one could not ensure uniformity across cases.

Fortunately, the cases in the dataset coded as anti-terrorist interventions do not suffer from definitional ambiguity. Since 1999, the Security Council’s 1267 Committee has maintained a list of individuals and organisations associated with the Taliban and Al-Qaeda.¹²³ Those on the list are subject to a comprehensive set of sanctions, overseen by the Committee.¹²⁴ The list began as an effort to combat the harbouring of terrorist groups in Afghanistan and was later expanded to encompass many ‘associated’ groups elsewhere, including so-called Islamic State in Iraq and the Levant (ISIL) in 2015.¹²⁵ The Council has articulated criteria for listing individuals and groups, and it created an ombudsperson to review requests for delisting from those who claim to have been listed erroneously.¹²⁶ The 1267 sanctions list, in other words, reflects a collective effort to identify and sanction specific ‘terrorists’ in the name of the international community as a whole. The Council repeatedly underlines this latter point by employing the terminology of international criminal law and the ius ad bellum to describe the acts of listed terrorists.¹²⁷

As detailed in Table 3.1, all but three of the non-state-conflict parties in interventions coded as ‘anti-terrorist’ have appeared on the 1267 list. The three exceptions do not involve disagreements over whether the groups involved were ‘terrorists’.¹²⁸ Of course, future conflicts may involve non-listed groups, or alleged terrorist groups may participate in conflicts the Security Council has not yet addressed. But, for the purposes of assessing Council practice to date, the definitional debates plaguing other areas of international law are not a complicating factor here.

Table 3.1 Terrorist Groups Participating in Coded Conflicts

Conflict Name	Terrorist Groups	1267 List [Y if on the list; N if not on the list]	Criteria for Inclusion*
Afghanistan	Taleban	Y	1, 2, 3
	Hizb-I Islami-yi Afghanistan	N
Afghanistan	IS	Y	1, 2, 3
Algeria	AQIM	Y	1
Cameroon	Jama’atu	Y	1
Cameroon	IS	Y	1, 2, 3
Libya	IS	Y	1, 2, 3
Mali	Ansar Dine	Y	1, 2, 3
	AQIM	Y	1
	MUJAO	Y	1
	Signed-in-Blood Battalion	Y	1, 3
	al-Murabitum	Y	1, 3
Mauritania	AQIM	Y	1
Niger	IS	Y	1, 2, 3
Nigeria	Jama’atu	Y	1
Nigeria	IS	Y	1, 2, 3
Syria v. IS	IS	Y	1, 2, 3
Uganda	LRA	*N [Listed instead pursuant to Res. 2262 (2016)]
	ADF	*N [Listed instead pursuant to Res. 2078 (2012)]
Uzbekistan	IMU	Y	1, 2, 3
Yemen (North Yemen)	AQAP	Y	1, 2, 3

Note

Listing criteria taken from UN Security Council, ‘Sanctions: Security Council Committee pursuant to resolutions 1267 (1999) 1989 (2011) and 2253 (2015) concerning Islamic State in Iraq and the Levant (Da’esh), Al-Qaida and associated individuals, groups, undertakings and entities – Summary of Listing Criteria’, available at www.un.org/securitycouncil/sanctions/1267#listing_criteria:

1. 1. Participating in the financing, planning, facilitating, preparing, or perpetrating of acts or activities by, in conjunction with, under the name of, on behalf of, or in support of –

2. 2. Supplying, selling or transferring arms and related material to –

3. 3. Recruiting for –

or otherwise supporting acts or activities of, ISIL (Da’esh), Al-Qaida or any cell, affiliate, splinter group or derivative thereof.

III. Methodology for Assessing Recent State Practice

We have now reviewed four theories that can plausibly claim grounding in contemporary international law: the IDI view, the Nicaragua view, the democratic legitimacy view, and the anti-terrorism view. How have these theories been received by the international community in practice since the end of the Cold War? To answer this question, we coded interventions in armed conflicts from 1990 to 2017 and the reaction of critical international actors to those interventions.¹²⁹ Among international organisations, we coded the UN Security Council and General Assembly, the European Union, the OAS and the African Union. Among states, we coded the United States, the United Kingdom, France, South Africa, Argentina, Australia, and Japan. For each intervention, we asked whether each actor approved, disapproved or issued a statement evidencing neither approval nor disapproval.¹³⁰

Before discussing the data, it is first important to describe how the cases of consensual intervention were selected, with attention to two aspects in particular.

A. Selecting Conflicts

All cases included in the dataset are taken from the Uppsala Conflict Data Program (UCDP), which creates widely used compilations of historical and contemporary data about armed conflict.¹³¹ We began with the UCDP External Support Dataset, which provides ‘information on the existence, type, and provider of external support for all warring parties (actors) coded as active in UCDP data, on an annual basis, between 1975 and 2009’.¹³² The dataset for this chapter is modified in three ways.

1. 1. We eliminated all cases prior to 1990, using that year as a proxy for the end of the Cold War and the beginning of an era in which the UN Security Council was capable (with obvious and notable exceptions) of addressing most armed conflicts around the world.

2. 2. Although the UCDP codes a wide variety of forms of external support for warring parties, our dataset includes only cases in which troops were supplied to a primary warring party.¹³³

3. 3. Because the Uppsala External Support dataset ends in 2009, our dataset adds post-2009 NIACs from a second Uppsala dataset. We used the UCDP’s main dataset – UCDP/PRIO Armed Conflict Dataset, version 17.2 – to supply cases of external support from 2009 through its end date of 31 December 2016 (as of 22 January 2018).¹³⁴ Because this later data includes international armed conflicts and NIACs, we eliminated the former from our data.¹³⁵

4. 4. Finally, the UCDP does not code for whether the interventions in either dataset were invited or not. However, a member of the Uppsala project clarified that an intervention in an NIAC would not have been coded as such unless the party receiving assistance consented to that assistance.¹³⁶ That assurance meant that all the interventions we coded using these criteria were consensual interventions.

The final set of cases resulting from these modifications is set out as Appendix II. That table includes information on the invited and inviting states, the party being supported, the nature of the intervention, and the reaction, if any, by the UN Security Council.

B. Defining Civil Wars

The second methodological question involves how to define a ‘civil war’. To investigate the IDI view prohibiting interventions in civil wars, it is necessary to define which conflicts in our data set qualify as such. The IDI’s 1975 Wiesbaden Resolution III itself employs a definition with three elements: rebels must have a minimum level of organisation; the conflict must pass an ‘intensity threshold’; and the rebel groups must have certain specific goals.¹³⁷ But whether the IDI view aligns with how the international community generally – and the Security Council in particular – defines ‘civil wars’ is an exceedingly complex question.¹³⁸

‘Civil war’ is not a term of art in international law.¹³⁹ International humanitarian law (IHL) refers instead to NIACs, with competing definitions grounded in two different international instruments.¹⁴⁰ The first is common Article 3 to the four Geneva Conventions of 1949, which refers to ‘armed conflict not of an international character occurring in the territory of one of the High Contracting Parties’.¹⁴¹ The most widely accepted definition of a NIAC as the term is used in common Article 3 appears in the 1995 Tadić decision of the International Criminal Tribunal for the former Yugoslavia (ICTY) – that is, ‘protracted armed violence between governmental authorities and organised armed groups or between such groups within a State’.¹⁴²

The second is found in the Additional Protocol II (AP II) on Non-International Armed Conflicts, which sets out narrower criteria for application than those found in Tadić. The Protocol applies to armed conflicts:

… which take place in the territory of a High Contracting Party between its armed forces and dissident armed forces or other organized armed groups which, under responsible command, exercise such control over a part of its territory as to enable them to carry out sustained and concerted military operations and to implement this Protocol.¹⁴³

The Tadić definition and AP II criteria share two common elements: the opposition groups must have some minimum level of organisation or structure;¹⁴⁴ and the conflicts must have reached a certain level of intensity.¹⁴⁵ The Protocol adds a third: the rebel groups must exert control over a part of the state’s territory. Each definition seeks to distinguish internal armed conflicts from lower-level disturbances, which few doubt a government can quell with external assistance.¹⁴⁶

While the Tadić test is widely understood as reflecting customary international law, its use as a metric to identify the NIACs in our dataset presents a number of difficulties.¹⁴⁷ First, the specific factors relevant to the ‘intensity threshold’ are quite unclear and thus that aspect of the definition is not easily quantified.¹⁴⁸ One metric that would seem especially well-suited to clear line-drawing – the number of fatalities at the time of an intervention – is not uniformly employed by tribunals applying the Tadić test.¹⁴⁹

Second, it is not at all clear that a test developed for IHL purposes is appropriate for defining a ‘civil war’ for ius ad bellum purposes. The argument for applying IHL to an internal conflict is that the individuals affected – civilians in particular – deserve protection from violence in which they play no role. This concern for individual dignity, it is argued, remains compelling whether a conflict is inter-state or intra-state.¹⁵⁰ By contrast, the ius ad bellum argument for prohibiting external assistance during ‘civil wars’ rests on the collective entitlement of a citizenry to determine its political future during periods of extreme polarisation. One could well imagine the threshold for recognising such polarisation being much higher than the threshold for applying individual IHL protections. The point at which the level of individual suffering becomes intolerable, such that IHL protections are necessary, could be much lower than the level at which it is clear that a substantial portion of the population finds the government so unacceptable that its violent removal becomes justified.¹⁵¹

Third, it seems unlikely that states and international organisations regularly employ a legal test for civil wars when issuing political reactions to interventions. Even if they did, it would be unclear which of the two tests (Tadić or AP II) they would use.

Given this lack of clarity on legal thresholds, this chapter will employ the rather straightforward UCDP definition of an ‘internal armed conflict’¹⁵² – namely, that it is one that ‘occurs between the government of a state and one or more internal opposition group(s)’,¹⁵³ with two additional characteristics: it must involve at least twenty-five battle-related deaths in a calendar year;¹⁵⁴ and the dispute must concern government (the ‘type of political system, the replacement of the central government or the change of its composition’) or territory (‘the change of the state in control of a certain territory (interstate conflict), secession or autonomy (intrastate conflict)’).

The UCDP definition does not include the first element of the Tadić test concerning the opposition group’s level of organisation. But the requirement that the conflict be about either government or territory can be seen as making up for this omission in performing the similar function of distinguishing politically oriented violence from mere criminal activity or low-level unrest. Uppsala requires that conflicts concern either government or territory because those are the root causes of most significant armed conflicts. To the extent that the international community seeks to resolve those conflicts, it must also engage with issues of territory or governance. The types of internal conflict that Uppsala codes, in other words, are those most likely to engage the international community.¹⁵⁵

In sum, the problem of defining civil wars arises because of the need to test the IDI view, which relies on a particular definition of ‘civil war’. But unless the Uppsala definition employed here is coextensive with the IDI definition, showing international approval of interventions in ‘civil wars’ would not necessarily demonstrate disapproval of the IDI view. Is that the case?

Table 3.2 shows how each of the definitions of civil war discussed above – Uppsala, IDI, common Article 3/Tadić, APII – employs the four elements common to some, but not all, of them: the rebels’ level of organisation, the conflict’s intensity threshold, whether rebels hold significant territory, and the rebels’ goals. The critical comparison of the Uppsala and IDI definitions, located in the third column in the table, shows that the two largely overlap. Neither requires rebel control of territory. The nature of the rebels’ goals is virtually identical. The differences in the rebels’ level of organisation is marginal: IDI has no such requirement and Uppsala requires only that the rebels be an ‘internal opposition group’.

Table 3.2 Civil War Definitions

Criteria	Uppsala	IDI	Uppsala or IDI Broader?	CA 3/Tadić	AP II
Rebels have minimum level of organisation	Rebels must be ‘internal opposition group’	None	IDI	Rebels must be ‘organised armed groups’	Rebels must be ‘under responsible command’
Intensity threshold	At least 25 battle-related deaths per conflict year	Does not cover ‘local disorders or riots’	Unclear	Must be ‘protracted armed violence’	Instrument does ‘not apply to situations of internal disturbances and tensions, such as riots, isolated and sporadic acts of violence and other acts of a similar nature’
Rebel control of territory	None	None	Same	Rebels must ‘exercise such control over a part of its territory as to enable them to carry out sustained and concerted military operations and to implement this Protocol’	None
Rebel goals	Conflict must concern either the ‘type of political system, the replacement of the central government or the change of its composition’ or ‘the change of the state in control of a certain territory (interstate conflict), secession or autonomy (intrastate conflict)’	‘[O]ne or more insurgent movements whose aim is to overthrow the government or the political, economic or social order of the State, or to achieve secession or self-government for any part of that State’	Same	None	None

There is arguably a divergence in the final factor, the intensity threshold. For IDI, conflicts must rise above ‘local disorders or riots’; Uppsala requires at least twenty-five battle-related deaths per conflict year. One could argue that conflicts which are somewhat more intense than local riots or disorders would not produce twenty-five deaths per year, meaning that using the IDI definition would produce more conflicts than Uppsala’s – but this difference is again marginal.

The comparison between the Uppsala and IDI definitions is the only one that matters. Their virtual identity ensures that data employing Uppsala can properly be used to assess international approval or disapproval of interventions in civil wars.

IV. Post-Cold War Practice: An Overview

What does the data show about post-Cold War practice?¹⁵⁶ Using the criteria described above, we coded a total of forty-four interventions by invitation in conflicts that were ongoing between 1990 and 2016. The most important conclusion to emerge is that the UN Security Council and General Assembly made statements on an overwhelming number of these interventions. As shown on Chart 3.1, the Council reacted to 82 per cent (36/44) of the interventions, condemning 22 per cent (10/44), approving 41 per cent (18/44), and issuing statements that neither approved nor condemned in 18 per cent (8/44).¹⁵⁷ The Council had no reaction to 18 per cent (8/44) of the interventions, none of which began after 2010.¹⁵⁸

Chart 3.1 Overview of Security Council Reaction to Interventions

The second conclusion is that patterns in Council actions are difficult to discern. As shown on Chart 3.2, the Council most frequently condemned interventions that began prior to 2000 (50 per cent). It most frequently approved interventions from 2000 to 2010 (53 per cent), although this was quite close to its approval of 50 per cent of the interventions it addressed from 2011 to 2020. The Council’s decision to approve or condemn does not appear connected to the severity of the conflicts, measured by the number of fatalities at the time of the intervention.¹⁵⁹ As shown in Chart 3.3, the Council approved of 50 per cent of the interventions it addressed in conflicts with 500–1,000 fatalities and 35 per cent in conflicts with 1,000–5,000 fatalities. The Council reviewed only four interventions in conflicts with more than 5,000 fatalities and one cannot say the greater severity was correlated with a specific Council reaction.¹⁶⁰

Chart 3.2 Security Council Reaction to Interventions by Decade

Chart 3.3 Security Council Reaction to Interventions by Severity of Conflict

Third, patterns are also difficult to discern based on conflict length. The vast majority of conflicts were either of short or long duration at the time of the intervention: twenty-one had been active less than one month and eighteen had been active for more than twelve months.¹⁶¹ As shown in Chart 3.4, for conflicts spanning up to one month, the Council condemned 37 per cent of those interventions and approved of 42 per cent.¹⁶² For conflicts lasting twelve months or longer, the Council condemned 15 per cent, approved 35 per cent, and either made no statement or a non-committal statement in 25 per cent.

Chart 3.4 Security Council Reaction to Interventions by Duration of Conflict

The General Assembly reacted to fewer of the interventions, passing resolutions in 34 per cent (15/44) of the cases. None of these reactions came in the eleven cases in which the Council did not issue a statement. Indeed, the six cases met with silence by the two UN bodies were also ignored in all but a few instances by major regional organisations and states whose reactions we also coded.¹⁶³

Of the forty-four cases of intervention, the largest groupings consisted of assistance to governments in conflict with rebels seeking to overthrow that government – 36 per cent (16/44) – and assistance to governments in a conflict with one or more terrorist organisations – 32 per cent (14/44) (see Chart 3.5). Next were cases of assistance to rebels seeking to overthrow a government, which occurred in 22 per cent (10/44) of the cases. Finally, assistance to an individual or group not in effective control of the government but which claimed an electoral mandate to hold office, or assistance to a regime that is in effective control and claims a democratic mandate occurred in 9 per cent (4/44) of the cases.

Chart 3.5 Purpose of Intervention

What general conclusions can we draw from these data? First, the Security Council has been a central player in reacting to post-Cold War consensual interventions. It has issued statements in the overwhelming majority of conflicts coded (82 per cent), condemning 22 per cent of the interventions it addressed (10/44) and approving 41 per cent (18/44) of those interventions (see Chart 3.1). The post-Cold War era has thus been dominated by a collective approach to interventions, in stark contrast to the atomised reactions of earlier eras, when either no mechanism for collective reaction existed (pre-1945) or such mechanisms were effectively paralysed (1945–90).

Second, because the Council either condemned or approved of 63 per cent (28/44) of the interventions it addressed, a strong argument can be made that Council practice ought to inform our understanding of contemporary norms. If the Council had been non-committal in reacting to most interventions – which one could well understand, given the delicate diplomacy necessary to resolve NIACs – then its reactions could be seen as simply an example of largely extralegal diplomatic manoeuvring. Instead, the Council took clear positions on most interventions.

Third, as we might have predicted, Council reactions appear to be case-specific. One might have predicted that the length and severity of conflicts in which interventions occurred would have been important factors in determining (i) whether the Council reacted and (ii) the nature of its reaction. But those factors are more or less evenly distributed across the conflicts we coded.

V. UN Security Council Views on the Prevalent Legal Theories

None of this general analysis tells us whether the Council has affirmed or rejected the major legal theories on consensual intervention. In this section, we will assess the relevance of Council practice to each theory.

A. The Nicaragua View

The first is the Nicaragua view, which would permit a government to invite foreign forces in all circumstances and never permit rebel groups to do so. The government portion of this view has not been borne out by Council practice. As shown on Chart 3.6, in the sixteen cases of assistance to a government in conflict with rebel groups, the Council condemned 19 per cent (3/16), approved of 37 per cent (6/16), and neither approved nor condemned in 25 per cent (4/16).¹⁶⁴ Clearly, there are circumstances in which the Council believes governmental invitations are permissible and others in which they are not.

Chart 3.6 Security Council Reaction to Different Types of Intervention

The United States’ support for the government of Iraq from 2004 to 2008 provides a good example of the Council’s approving aid to a government fighting rebels in the midst of a NIAC. The Iraqi government was in conflict with the Al-Mahdi Army, a group formed in 2003 by Shi’a cleric Moqtada Al-Sadr.¹⁶⁵ Critical indicators of a NIAC were present: the International Committee of the Red Cross (ICRC) concluded that IHL applied to the conflict and the UCDP estimates the conflict resulted in 1,258 fatalities that year, undoubtedly meeting the intensity threshold.¹⁶⁶

Following the official end of the US/UK occupation of Iraq on 30 June 2004 and after the Coalition Provisional Authority handed governmental control over to an elected Iraqi regime, the Security Council approved a continued US presence under the umbrella of a ‘multinational force’.¹⁶⁷ The resolution was accompanied by a letter from the US secretary of state offering military assistance and a letter from the Iraqi prime minister accepting the offer. The United States’ letter described the troop’s mission as involving, among other tasks, ‘combat operations against members of these groups, internment where this is necessary for imperative reasons of security, and the continued search for and securing of weapons that threaten Iraq’s security’.¹⁶⁸ This authorisation was renewed several times until 31 December 2008.¹⁶⁹

A second example of Council support for the Nicaragua view is the 2006 Ethiopian intervention in Somalia, which involved tacit, rather than explicit, Council approval.¹⁷⁰ After the anarchy of the 1990s, a regional initiative established a Transitional Federal Government (TFG) for Somalia. But it failed to exercise any substantial control over Somali territory and the TFG fell into conflict with the Islamic Courts Union (ICU), an extremist Islamist group.¹⁷¹ In 2004, the leader of the TFG requested the deployment of regional forces to assist his regime – a request that was soon endorsed by most member states of the Intergovernmental Authority on Development (IGAD), a Horn of Africa regional organisation, as well as a body of the African Union.¹⁷² The UN Security Council had previously imposed an arms embargo on Somalia and such an intervention would require that an exception be made to the embargo. This exception came in the form of Resolution 1725, in which the Council permitted the deployment of an IGAD peacekeeping mission to Somalia.¹⁷³

While this process played out, Ethiopian troops entered the country to support the TFG.¹⁷⁴ Their presence allowed the TFG to survive.¹⁷⁵ It seems virtually inconceivable that Ethiopian forces were in Somalia without the consent of the TFG. In several reports to the Council, the UN Secretary-General noted that TFG forces were frequently supported by Ethiopian troops in key battles.¹⁷⁶ Ethiopian forces provided crucial support for the TFG while it waited for, first, the IGAD and, then, the African Union to deploy forces. For its part, the UN Security Council had numerous opportunities to condemn the Ethiopian presence, which the UN Secretary-General specifically noted in his reports – yet it issued no such condemnation. As Eliav Lieblich concludes, ‘[t]he international response to the intervention by Ethiopia was largely one of acquiescence’.¹⁷⁷

An example of the Council acting inconsistently with the Nicaragua view is its reaction to Senegal’s assistance to the government of Guinea-Bissau. While the Council did not reiterate the self-determination rationale for disapproving of the intervention, its actions tracked one of the classic arguments against the Nicaragua view: that foreign assistance to a government will simply prolong a conflict. The 1998 conflict in Guinea-Bissau involved a military junta seeking to dislodge increasingly unpopular President João Bernardo Vieira. Vieira had earlier halted his country’s support for rebels in neighbouring Senegal and, in recognition of this action, the Senegalese government provided 2,000 troops to support Vieira within 48 hours of the junta’s rebellion.¹⁷⁸ At one point, government forces managed to hold the presidential palace only with the assistance of Senegalese soldiers.¹⁷⁹ After the parties signed a series of peace documents in mid-to-late 1998, the Security Council commended the end of violence and called for ‘the withdrawal of all foreign troops in Guinea-Bissau’.¹⁸⁰

In contrast to its failure to follow the Nicaragua view on assistance to governments, the Council does appear to agree with Nicaragua’s blanket disapproval of assistance to rebel groups. The Council did not approve any of the nine interventions assisting rebel groups and specifically disapproved of seven.¹⁸¹ The Council used quite general language in many of these cases, condemning all outside intervention in the states.¹⁸²

B. The IDI View

The second view is the IDI claim that responding to invitations from governments is permissible up until a conflict becomes a civil war. The IDI view would permit intervention in only a subset of conflicts encompassed by the Nicaragua view. On first blush, the Council seems to have almost wholly ignored the IDI ‘civil war’ limitation. As noted earlier, all cases in the dataset qualify as internal conflicts, according to Uppsala criteria, which I argue largely correlate to the contemporary understanding of a NIAC in international law. The Council approved 41 per cent (18/44) of all interventions in the dataset (see Chart 3.1). Even limiting our examination to the cases in which interventions are designed to assist governments fighting rebels – the same cases considered for the Nicaragua view – the Council approved 37 per cent.¹⁸³ As a result, one could well conclude that the IDI view finds no support in Council practice.¹⁸⁴

But this obscures a difficulty in testing the IDI view. Although all forty-four cases qualify as internal conflicts according to UCDP, Uppsala classified 84 per cent (37/44) of those conflicts as ‘internationalised internal conflicts’ – that is, as internal conflicts ‘with intervention from other states … on one or both sides’.¹⁸⁵ Of course, one would assume that these conflicts were ‘internationalised’ because of the consensual intervention – but that is not the case: Uppsala codes six conflicts in the dataset as pure ‘internal conflicts’ despite the presence of a consensual intervention.¹⁸⁶ And the Council approved only one of those interventions (in the Central African Republic¹⁸⁷), suggesting that an IDI civil war limitation might be operating in cases of ‘pure’ internal conflict.

An alternative view is that these conflicts were internationalised because of an intervention to assist rebel groups that preceded the consensual intervention on the government side. Under the Uppsala coding scheme, this seems unlikely.¹⁸⁸ But if there were a prior intervention supporting rebels, the government would not need to invoke the intervention by invitation doctrine to support a counter-intervention in its favour; it could rely instead on collective self-defence in response to the armed attack represented by the initial assistance to the rebels.¹⁸⁹

These considerations do not, however, alter the conclusion that the Council has not disapproved of intervention in civil war – that is, that it has not adopted the IDI limitation. First, the Council’s views on an invitation justification are not made irrelevant by the existence of an alternative theory of justification. Indeed, if one were to insist that the only legally useful cases were those in which an invitation was the sole justification advanced for the use of force, one would be left with very little state practice at all.¹⁹⁰ Second, it is not clear in these cases that the Council was aware of prior interventions when it gave its approval, or that any prior interventions rose to the level of an ‘armed attack’ triggering a right of collective self-defence.¹⁹¹ Finally, if the IDI view truly guided Council actions, then the Council would have disapproved of all six interventions in ‘pure internal conflicts’. In fact, the Council disapproved only in the case of Angola.¹⁹²

Another way of approaching how the IDI view fared in Council practice is to ask whether, in any of the three cases in which the Council disapproved of pro-government interventions, it did so because the conflict had reached the threshold of a ‘civil war’. The Council did not do so explicitly. Its resolutions on South Sudan, Guinea-Bissau, and DR Congo (Kabila) condemned external intervention as part of an overall condemnation of ongoing or renewed conflict.¹⁹³ One might interpret the Council’s call for ending hostilities as consistent with favouring an indigenous resolution to the conflicts, free from the skewing effect of foreign support for the government. But bringing an end to fighting and facilitating peace negotiations are goals the Council pursues in every NIAC, whether or not foreign forces are involved.¹⁹⁴ The involvement of foreign forces in these civil wars, in other words, does not appear to be the reason the Council condemned the interventions.

The case of France’s 2013 intervention in Mali has been cited both as an example of the Security Council rejecting the IDI view and as an example of its endorsement of the anti-terrorist view.¹⁹⁵ A review of the Council’s reaction and that of its members suggests that the case may plausibly support both theories. The Mali case begins with discontent on the part of the Tuareg people, a nomadic group with origins in northern Mali, near the borders of Algeria, Niger, and Libya.¹⁹⁶ After the overthrow and death of Libyan leader Muammar Gaddafi in 2011, many Tuareg who had been living in Libya returned to northern Mali and founded a Tuareg separatist group, the National Movement for the Liberation of Azawad (MNLA). Those forming the MNLA rejected a potential leader, Iyad Ag Ghali, alienating him from the group.¹⁹⁷ Shortly thereafter, Ghali formed Ansar Dine, a group that was also predominantly Tuareg but which sought to bring a fundamentalist form of Islam to Mali. Despite their separate origins, MNLA and Ansar Dine both fought against Malian troops in the north. By mid-March 2012, the Malian Army had lost a third of the country’s territory to the two rebel groups.¹⁹⁸

The growing lack of confidence in the Malian government led to protests in Bamako, culminating in a coup d’état on 22 March 2012.¹⁹⁹ Malian Army Captain Amadou Sanogo and his followers seized power and suspended Mali’s constitution.²⁰⁰ The coup was widely condemned and, after negotiations led by ECOWAS, Sanogo agreed to a transitional political process under the leadership of interim President Dioncounda Traoré.²⁰¹

On 6 April 2012, having occupied a series of towns in the north, the MNLA declared independence for the state of Azawad.²⁰² From this point, the dynamics of the conflict became fluid. In May, Ansar Dine and several other Islamist groups fighting in the north formed an alliance with the MNLA.²⁰³ Shortly thereafter, the relationship soured. Ansar Dine had secured support from Al-Qaeda in the Islamic Maghreb (AQIM) and its splinter group MUJAO. In June, these groups forced MNLA out of many of the occupied towns in the north of the country and began advancing south.²⁰⁴

The UN Security Council began to react in the spring of 2012. Its resolutions and presidential statements initially addressed only the rebel groups, but then later expanded to address both rebel and ‘terrorist’ groups.²⁰⁵ Critically, in other words, the Council did not refer to ‘rebel’ and ‘terrorist’ groups as one and the same, but as distinct. In Resolution 2056 on 5 July, for example, the Council expressed its ‘categorical rejection of statements made by the National Movement for the Liberation of Azawad (MNLA) regarding the so-called “independence” of Northern Mali, and further reiterating that it considers such announcements as null and void’.²⁰⁶ In the same Resolution, it called on all groups in northern Mali, including the MNLA, Ansar Dine, and foreign combatants on Malian soil, ‘to renounce all affiliations incompatible with peace, security, the rule of law and the territorial integrity of Mali’.²⁰⁷ As the terrorist groups advanced south, they became the focus of Council attention, although it still occasionally mentioned the ‘rebels’.²⁰⁸

With the security situation in the north deteriorating, the transitional authorities requested military assistance from ECOWAS on 1 September 2012.²⁰⁹ This was followed by requests from both the transitional authorities and ECOWAS that the UN Security Council authorise the deployment of an international military force.²¹⁰ On 12 October, the Security Council did so, in Resolution 2071.²¹¹ On 20 December, the Council created the African-led International Support Mission in Mali (AFISMA), ‘[t]o support the Malian authorities in recovering the areas in the north of its territory under the control of terrorist, extremist and armed groups’.²¹² The disjunctive listing of the three types of group suggests that the Council did not consider them one and the same.

But the international force came together slowly and, by early January 2013, the armed groups were only 700 kilometres from Bamako. On 11 January, French President François Hollande announced that he had received a request for assistance from the transitional government and that France had agreed to help.²¹³ The intervention effectively stopped the groups’ advances.²¹⁴ The Council discussed the French intervention on 22 January. Many speakers praised the French action only for halting ‘terrorist’ advances.²¹⁵ Others, following the Council’s lead, referred to both terrorist and rebel groups.²¹⁶

In Resolution 2100, the UN Security Council welcomed ‘the swift action by the French forces, at the request of the transitional authorities of Mali, to stop the offensive of terrorist, extremist and armed groups towards the south of Mali’.²¹⁷ The Permanent Representative from Mali repeated the distinction between terrorist-affiliated groups, on the one hand, and the Tuareg separatists, on the other, describing the French intervention as supporting the government in opposing both.²¹⁸ The Council did not authorise the French intervention: although Resolution 2100 invoked Chapter VII of the UN Charter, the French action is mentioned only in the preambular paragraphs in the language quoted above. Legal authority for the request could thus emanate only from the invitation by the Malian transitional authorities. Despite the transitional regime being unelected and the Council urging it ‘to hold free, fair, transparent and inclusive presidential and legislative elections as soon as technically possible,’²¹⁹ Resolution 2100 does not treat them as incapable of issuing the invitation.

Thus the Council and its members clearly distinguished between the two sets of antagonists in Mali. There was good reason for them to do so: as noted, the Tuareg-focused MNLA had broken with the jihadist-focused groups (Ansar Dine and MUJAO) prior the French intervention, and the two factions began fighting with each other even as they were also fighting with the Malian Army and its French allies.²²⁰ It is thus difficult to conclude that the Council viewed the French intervention as assisting only in repelling the Islamist groups. Moreover, Mali is not a case in which ordinary rebels were rebranded as ‘terrorists’, so that the government could gain international support. The Council had already distinguished the groups by listing both AQIM and MUAO, but not MNLA, as terrorist groups subject to sanctions under Resolution 1267.²²¹

Finally, the ongoing efforts by various international actors to facilitate a political solution to the conflict appears incompatible with viewing Mali as solely an intervention concerned with terrorism. As the Under-Secretary-General for Political Affairs told the Council on 5 December 2012, the Secretary-General’s Special Representative:

… has significantly increased his political engagement with the authorities in Mali and key regional stakeholders to provide momentum to a Malian-owned political process focused on three main objectives: first, broad-based and inclusive national dialogue aimed at formulating a road map for the transition; secondly, negotiations with armed groups in the north that renounce violence and terrorism; and thirdly, preparations for the holding of elections.²²²

Critically, that peace process is described as one involving reconciliation among competing Malian opposition groups:

Despite concerted international efforts, the political landscape in Mali remains complex and fragmented. It is critical that the key political actors arrive at a unified vision as soon as possible if they are to effectively focus efforts on the main transition challenges, in particular national dialogue and negotiations with the armed groups. The support of the international community will continue to be critical in helping the Malians to bridge differences and arrive at a national consensus.²²³

A conflict involving a ‘complex and fragmented’ political landscape that requires citizens to ‘bridge differences and arrive at a national consensus’ sounds very much like a civil war.²²⁴ Terrorist groups, as noted, are generally seen as operating outside such a process of national self-determination. Anti-terrorism was certainly an objective articulated by virtually all international actors who characterised the intervention – but it was decidedly not the only objective. Mali thus stands as a substantial obstacle to grounding the IDI view in Council practice.

C. The Democratic Legitimacy View

The third theory is the democratic legitimacy view, finding invitations to be valid when they come from individuals or parties with a clear electoral mandate, who have been denied their office. One drawback of the data on this question is the small number of coded cases involving claims of democratic legitimacy: the Central African Republic (2002), Lesotho (1998), Sierra Leone (1997), and Yemen (2015) (see Appendix II).²²⁵

The record from this small sample size is mixed. The Council approved two pro-democratic interventions (Sierra Leone and Yemen) and either issued no statement or an equivocal statement for the remaining two.

1. Sierra Leone (2000)

The United Kingdom intervened in Sierra Leone in May 2000, after the failure of a peace agreement between the elected government of Ahmad Tejan Kabbah and the brutal Revolutionary United Front (RUF).²²⁶ With RUF forces threatening both Kabbah’s hold on the presidency and a newly deployed UN peacekeeping mission, the United Kingdom made a series of troop deployments with the consent of the Kabbah regime.²²⁷ The deployments are credited with halting a RUF advance that would almost certainly have toppled the regime.²²⁸ It also brought the RUF back to the negotiating table and eventually led to new elections in 2002.²²⁹ Protection of the elected regime was one of several justifications given by the UK government.²³⁰ Importantly, when Kabbah consented to the UK intervention, the RUF controlled at least 40 per cent of the country – one factor in the NIAC threshold.²³¹ Although the UN Security Council did not refer to the UK intervention in a resolution or presidential statement, the overwhelming number of states present expressed their approval at a Council meeting on 11 May 2000.²³² Many also described the RUF as threatening, in the words of the United States, ‘yet again to undermine the democratically elected government of President Kabbah’.²³³ The lack of a collective endorsement weakens Sierra Leone as support for the democratic legitimacy theory, but it does not undermine it altogether.

2. Yemen (2015)

The 2015 intervention in Yemen followed a three-year long deterioration in the country’s internal security. In November 2011, in the midst of the Arab Spring uprisings, President Ali Abdullah Saleh resigned pursuant to a Gulf Cooperation Council (GCC) initiative that included a long-term political transition process.²³⁴ Then Vice-President Hadi stood for election on 21 February 2012 and won, with 99.8 per cent of the vote.²³⁵ Hadi then formed a government of national unity.²³⁶

The UN Security Council gave its full support to the GCC-led transition process in its Resolution 2014.²³⁷ However, the Houthis (a Zaydist group based in the north of Yemen) rejected the GCC process, claiming it did not include the entire Yemeni people, and boycotted the election.²³⁸ The Houthis aligned themselves with the still-influential former President Saleh and his remaining supporters.²³⁹ They soon moved from the north to expand their territorial control.²⁴⁰ By September, the Houthis had taken control of the capital, Sana’a.²⁴¹ The Security Council condemned the Houthis’ action, imposed sanctions, and maintained its support for Hadi.²⁴²

Against this background, on 21 September 2014, Hadi’s government and the Houthis signed the Peace and National Partnership Agreement (PNPA), which was intended to create a unity government with Houthi representation in the cabinet.²⁴³ The Security Council welcomed the agreement, and it once again stressed that ‘Hadi is the legitimate authority based on election results and the terms of the GCC Initiative and Implementation Mechanism’.²⁴⁴

But the Houthis failed to realign their forces, as required in the agreement,²⁴⁵ and rejected a draft constitution submitted to Hadi on 7 January 2015.²⁴⁶ In early 2015, President Hadi and his cabinet were put under house arrest, and they collectively resigned on 22 January. Houthi forces once again took control of Sana’a.²⁴⁷ In February, an expert panel created by the Security Council concluded that the Yemeni conflict had risen to ‘the threshold of internal armed conflict in accordance with the international definition’.²⁴⁸ On 6 February, the Houthis terminated the then-ongoing UN-led negotiations, and announced the dissolution of Parliament and the establishment of a ‘presidential council’ to run the country temporarily.²⁴⁹ On 24 March, President Hadi requested military assistance from the GCC.²⁵⁰ Two days later, Saudi Arabia and other GCC states launched Operation ‘Decisive Storm’.²⁵¹ The intervention tilted the balance of the civil war in favour of the exiled government forces.²⁵²

In his letter requesting intervention, Hadi referred, first and foremost, to the acts of ‘Houthi coup orchestrators’.²⁵³ Hadi stated that while he had sought a peaceful solution to the conflict, ‘our peaceful and constant efforts have been categorically rejected by the Houthi coup orchestrators, who are continuing their campaign of aggression aimed at subjugating the rest of the country’s regions, particularly the south’.²⁵⁴ The Houthi actions are described as ‘acts of aggression’ – a phrase from the UN Charter normally applied to inter-state actions. While the letter then continues to focus on Houthi actions, it also states that the Houthis were supported ‘by internal forces that have sold their souls and are concerned only with their own interests’. This appears to be a reference to Al-Qaeda in the Arabian Peninsula (AQAP). Finally, Hadi states that the Houthis were ‘being supported by regional Powers that are seeking to impose their control over the country and turn it into a tool by which they can extend their influence in the region’.²⁵⁵ This appears to be a reference to Iran.

At the end of the letter, Hadi summarised his request to the GCC thus:

I urge you, in accordance with the right of self-defence set forth in Article 51 of the Charter of the United Nations, and with the Charter of the League of Arab States and the Treaty on Joint Defence, to provide immediate support in every form and take the necessary measures, including military intervention, to protect Yemen and its people from the ongoing Houthi aggression, repel the attack that is expected at any moment on Aden and the other cities of the South, and help Yemen to confront Al-Qaida and Islamic State in Iraq and the Levant.²⁵⁶

Saudi Arabia and the other GCC states described their acceptance of Hadi’s invitation in the same document submitted to the Security Council. The relevant passages are worth quoting in full:

We note the contents of President Hadi’s letter, which asks for immediate support in every form and for the necessary action to be taken in order to protect Yemen and its people from the aggression of the Houthi militias. The latter are supported by regional forces, which are seeking to extend their hegemony over Yemen and use the country as a base from which to influence the region. The threat is therefore not only to the security, stability and sovereignty of Yemen, but also to the security of the region as a whole and to international peace and security. President Hadi has also appealed for help in confronting terrorist organizations.

The Houthi militias have failed to respond to repeated warnings from the States members of the Gulf Cooperation Council and the Security Council. They have continued to violate international law and norms, and to build up a military presence, including heavy weapons and missiles, on the border of Saudi Arabia. They recently carried out large-scale military exercises using medium and heavy weapons, with live ammunition, near the Saudi Arabian border. The Houthi militias have already carried out a bare-faced and unjustified attack on the territory of Saudi Arabia, in November 2009, and their current actions make it clear that they intend to do so again. Our countries have therefore decided to respond to President Hadi’s appeal to protect Yemen and its great people from the aggression of the Houthi militias, which have always been a tool of outside forces that have constantly sought to undermine the safety and stability of Yemen.²⁵⁷

Resolution 2216 – the first adopted after the Saudi-led intervention – did not explicitly support the military action, although it noted Hadi’s request and the Saudi response.²⁵⁸ The Council did reiterate ‘its support for the legitimacy of the President of Yemen, Abdo Rabbo Mansour Hadi’, and called for the end to any actions that undermine ‘the legitimacy of the President of Yemen’.²⁵⁹ The Council also declared its ‘support for the efforts of the Gulf Cooperation Council in assisting the political transition in Yemen and commend[] its engagement in this regard’.²⁶⁰ In a debate over the Resolution, ‘no single Council member (not even Russia) explicitly questioned the legality of Operation Decisive Storm’.²⁶¹ The United Kingdom was the only state to address the Saudi intervention directly: it expressed support and tied the intervention to Houthi aggression.²⁶²

How should the Yemen intervention be classified? Hadi’s request and the Saudi response articulated three grounds for the intervention: defending the legitimate government against Houthi advances, countering terrorist forces, and responding to a prior intervention by ‘regional powers’ (i.e., Iran). Of these three, the claim of support for Hadi’s legitimate governmental authority best accords with the facts described in the letters, the reactions of other states, and the facts on the ground. For that reason, Yemen is coded as a pro-democracy intervention. But because both Hadi and the Saudis also mention ‘external’ intervention, Yemen may also be seen as a counter-intervention. This is a tenuous claim at best, however, as discussed below.

A) Hadi’s Invitation and the Saudi Response

The two letters from Hadi and Saudi Arabia (on behalf of the GCC) are overwhelmingly devoted to buttressing the legitimacy of Hadi’s presidency and countering the threat of the Houthi offensive. Hadi describes the threat as coming from ‘Houthi coup orchestrators’. The references to terrorist groups and external support for the Houthis are almost afterthoughts, asserted without supporting facts. Indeed, the closing paragraph of Hadi’s letter, while citing Article 51 of the UN Charter, makes no reference to an attack by Iran or any other state. Similarly, the Saudi letter summarises Hadi’s request as seeking ‘immediate support in every form and for the necessary action to be taken in order to protect Yemen and its people from the aggression of the Houthi militias’. The Saudis speak of their decision ‘to respond to President Hadi’s appeal to protect Yemen and its great people from the aggression of the Houthi militias’. While the Saudis also refer vaguely to ‘support’ from ‘regional forces’, they do not describe this support as involving troops or as a military intervention. The letter also notes that ‘President Hadi has also appealed for help in confronting terrorist organizations’, but says no more about the threat posed.²⁶³

B) State Reaction to the Intervention

An assessment of state reaction to the intervention begins with that of the UN Security Council itself. As Ruys and Ferro note, Resolution 2216 supported the two essential predicates for the democratic legitimacy rationale: the legitimacy of President Hadi;²⁶⁴ and the illegitimacy of Houthi actions (including a demand that they cease ‘all actions that are exclusively within the authority of the legitimate Government of Yemen’).²⁶⁵ The Resolution’s first operative paragraph focused on the Houthi threat to the democratic transition process.²⁶⁶ The Council imposed sanctions only on Houthi leaders, not on terrorist groups or Iran.²⁶⁷ Indeed, the Resolution makes no mention whatsoever of Iran, external support for the Houthis or Yemen’s right to self-defence.²⁶⁸ The Resolution condemns acts by AQAP but takes no action in response.

Outside the Council setting, the states supporting the intervention based their position largely on the Houthi threat to Hadi’s legitimate government. Most did not refer to terrorism or a prior intervention. This was true of the Arab League,²⁶⁹ the United States,²⁷⁰ the United Kingdom,²⁷¹ France,²⁷² and Canada,²⁷³ among others. Russia and the European Union criticised the intervention but focused on the potential for escalation rather than the invalidity of Hadi’s invitation.²⁷⁴ Importantly, none of the critics argued that the intervention was unlawful because the Yemen conflict had passed the NIAC threshold. A UN expert panel had determined one month prior to the intervention that the Yemen conflict constituted a NIAC.²⁷⁵ Yemen may thus contribute to the view that the democratic legitimacy theory – to the extent that it is accepted – is not constrained by the IDI view.

Since the intervention, the Council has continued to demand that the Houthis abide by the GCC transition process and has reaffirmed the centrality of that process to political reconciliation in Yemen.²⁷⁶ In particular, at the time of writing, the Council has not deviated from its support in Resolution 2216 for ‘the legitimacy of the President of Yemen, Abdo Rabbo Mansour Hadi’.²⁷⁷

C) Facts on the Ground

Finally, do the facts described by independent observers support the offhand, yet undeniably articulated, counter-intervention and anti-terrorist justifications for the Saudi intervention? The UCDP does not list an intervention by Iran (or any other state) prior to the Saudi intervention. The UN Panel of Experts on Yemen also does not mention an Iranian intervention in its 2015 and 2016 reports.²⁷⁸ The Panel did find that anti-tank missiles supplied to the Houthis were ‘likely to have been maintained or overhauled in the Islamic Republic of Iran’, but it did not mention Iranian personnel in Yemen.²⁷⁹ External observers of the war’s origins describe Iranian assistance to the Houthis as ‘minor and irrelevant to the balance of power in the ongoing war’.²⁸⁰ Some suggest that Saudi Arabia in fact exaggerated Iranian assistance to the Houthis to further justify its intervention.²⁸¹

As for AQAP, the UCDP does not list it as a party to the Yemeni conflict. The Panel of Experts describes AQAP as primarily engaged in conflict with the Houthis, complicating an anti-terrorism rationale on the part of the Saudis, who were also fighting the Houthis.²⁸² It appears that AQAP was not so much a presence in the main conflict between supporters of Hadi and the Houthis as it was a beneficiary of the power vacuum left by the breakdown of state authority brought on by the war.²⁸³

D) Assessing the Yemeni Case

The rationale dominating Hadi’s request for assistance was support for his government against a Houthi rebellion that undermined the GCC- and UN-backed transitional process. That rationale also dominates the Saudi response, the response of the UN Security Council, and the reaction of other states to the Saudi intervention. It is also the view most consistent with the facts on the ground. The UCDP does not identify either Iran or AQAP as party to the Yemeni conflict. Their absence is consistent with the marginal status of the counter-intervention and anti-terrorism justifications in the Hadi and Saudi letters.

Yet if Yemen is not a weak case for the democratic legitimacy view, neither is it an unambiguously strong one. First, Yemen presents the difficult question of how international law should process rationales for intervention that are either pretextual or only minimally grounded in fact. The Saudi claims to have responded to terrorists or a prior Iranian intervention simply are not supported by the facts. During the Cold War, such abuses of intervention by invitation were the primary justification for the restrictive IDI approach. I have argued that such factual conflicts are precisely the ones the Council can now resolve, meaning that the IDI limitations have lost much of their rationale.

Second, while the Council did not explicitly approve the GCC action, it did affirm the essential elements of the democratic invitation theory: the democratic legitimacy of the Hadi regime, the unacceptability (owing to the lack of democratic bona fides) of Houthi control, the continuing validity of Hadi’s claim to power despite his lack of effective control, and the validity (i.e., non-fictitious nature of) his invitation to the GCC states. That all of these factors were affirmed by the Council, as opposed to only the intervening state, adds credibility to the claim.

Third, much international support for Hadi and, by extension, his GCC benefactors was phrased not as favouring democratic legitimacy as such but as supporting the transitional process that the Houthi offensive had interrupted. That process was intended to culminate in a ‘democratic’ constitution and elections, so this may be a distinction without a difference. But it does somewhat attenuate the invention from a specific democratic outcome.

In sum, at a minimum, both Sierra Leone and Yemen presented the Council with opportunities to reject the democratic legitimacy theory in favour of the traditional effective control test. The Council did not do so in either case.

D. Anti-Terrorism

The fourth theory supports invitations by governments for assistance in conflicts with transnational terrorist groups. In the fourteen such cases in the dataset, the Council approved intervention in 71 per cent (10/14) and disapproved none (see Chart 3.6).²⁸⁴ In the case of the United States aiding the government of Yemen in its conflict with AQAP, while the Council issued no statement, the European Union approved of the action. If one were to take the EU approval as indicative of larger international opinion, the percentage of anti-terrorist interventions receiving international approval would rise to 78 per cent (11/14).

Since the Council appears to have accepted anti-terrorism interventions, there is little need to review individual cases. As noted above and shown in Table 3.1, all but three of the groups involved in anti-terrorist interventions had appeared on the Council’s 1267 list of terrorist organisations.²⁸⁵ The interventions have ranged from troops provided by one state (Mali), to troops provided by a small group of states (the Multinational Joint Task force that sent Chadian forces into Cameroon in 2015 to counter Boko Haram²⁸⁶), to troops provided by a larger group of states (the United States and its allies in Afghanistan, to counter the Taliban and other groups²⁸⁷).

E. Conclusions

Security Council practice does not reveal a preference for one dominant theory. The Council’s consistent approval of counter-terrorism interventions is obviously relevant only to a limited number of cases, as is its approval of pro-democracy interventions. Both the Nicaragua and IDI views are applicable to all possible conflicts, but the Council has not unequivocally endorsed either one.

VI. A New Paradigm? The Multilateralisation of Consensual Interventions

Both general theories of consensual intervention – IDI and Nicaragua – emerged during the Cold War. I argue that while the rise of UN Security Council practice does not resolve the competition between the two on their merits, it does reveal that their historically bound assumptions have been substantially eroded. I will argue further that, as a result, the international community should be open to treating Security Council practice as important evidence of customary international law in evaluating the lawfulness of consensual interventions.

A. The Demise of Rules for a Polarised World

1. The IDI and Nicaragua Views in Contemporary Context

Section II described how both the IDI and Nicaragua views were deeply embedded in the realities of Cold War politics.

Those realities have changed and the theories built upon their assumed continuation face two important challenges. First, both theories were premised on the absence of collective mechanisms to sort legitimate from illegitimate invitations. Despite the obvious desirability of centralised decision-making by the United Nations, the organisation was all but irrelevant to most NIACs.²⁸⁸ Individual states, largely divided into Cold War camps, were left to judge the legality of interventions for themselves. The solution devised by the IDI was to impose a broad prophylactic rule to minimise the number of divisive factual questions surrounding interventions. Highly politicised issues, such as whether an invitation was real or fictitious, whether a government exercised effective control, and whether a regime had ‘popular mandate’, were not to be entrusted to the self-judgment of states invested in the conflict; instead, they were made irrelevant in the most consequential cases – when a civil war had broken out. Taking the opposite approach, the Nicaragua view dealt with these divisive questions by ignoring them.

With the end of Cold War polarisation, multilateral engagement with NIACs increased as superpower investment in their outcomes receded. As the data has shown, the Council and some regional organisations regularly take positions on NIACs, including on questions of regime legitimacy. They have done so in a variety of ways: by condemning foreign intervention, by supporting particular sides in NIACs, by supervising elections in post-conflict states, and by designating the winners of those elections as the legitimate leaders of the state. In each case, the unilateral and self-interested views of Cold War antagonists have been replaced by a collective judgment. As a result, categorical prophylactic rules such as the IDI and Nicaragua views seem unnecessary to check the good faith of the antagonist states.

The diminished importance of the two theories may be seen as an illustration of Thomas Franck’s distinction between the legitimacy of ‘categorical rules’ and ‘complex elastic rules’.²⁸⁹ Franck argued that categorical rules, addressing problems with a simple and definite clarity, are most useful when no ‘authoritative interpreter’ of a norm is available – that is, when no entity is empowered to apply a complex scheme of rules to unclear facts and reach a determinate conclusion. Simple rules can effectively apply themselves and thus have less need of adjudicatory or evaluative institutions to achieve compliance. When such credible institutions do exist, however, more nuanced rules can be substituted if the institution is perceived as legitimate. Such ‘process legitimacy’ may ‘credibly mitigate the elastic quality’ of more complex rules.²⁹⁰

The IDI and Nicaragua views have the virtue of simplicity and clarity. Their lack of complexity minimised states’ ability to evade compliance. During the Cold War, those attributes were clearly essential. But since the Security Council has emerged as the authoritative interpreter of ius ad bellum norms (obviously not all cases), more nuanced rules may be appropriate.

Second, the more widely subscribed IDI view rests on a needlessly shallow conception of state autonomy. The IDI view prevents an inquiry into the question of how states actually choose their leaders, substituting the legal fiction that prohibiting foreign intervention allows a national ‘choice’ to be made. As Brad Roth has observed, conventional wisdom during the Cold War ‘held that empirical investigations to ascertain public opinion in a foreign state was most often impracticable; that “popular will” itself was a complex and normatively loaded concept; and that any imposition from abroad of procedures calculated to measure “popular will” was presumptuous at best, and a usurpation at worst’.²⁹¹ Preserving ‘autonomy’ through non-intervention, in other words, involved not creating opportunities for actual popular choice but indulging a presumption that any leadership in effective control had, for the rest of the international community, been acceptably ‘chosen’.

In an era when disagreements on theories of political legitimacy lay at the very heart of the superpower divide, this disinterest made sense. International law of the period did not ‘generally address domestic constitutional issues, such as how a national government is formed’.²⁹² Moreover, most, if not all, superpower interventions at the behest of ‘legitimate’ governments were assumed to support unpopular regimes that might otherwise fall. In such circumstances, a rule permitting intervention would become ‘an instrument to prevent social change, which is a vital aspect of national self-determination’.²⁹³

But international law is no longer deliberately indifferent to questions of regime legitimacy and how governments treat (or mistreat) their citizens.²⁹⁴ In an era of normative commitment to democratic elections and human rights, as well as the omnipresence of election monitors and human rights reporting, deliberately avoiding the question of whether a given regime is actually supported by its citizens seems anachronistic.²⁹⁵ Certainly, the Council has not practised democratic avoidance: ‘Since 1993 approximately a hundred resolutions referred to “democracy” as a form of governance that needs to be enhanced, strengthened or supported.’²⁹⁶ Consider the remarkable growth in election monitoring since the end of the Cold War, which has made information on most elections available for external scrutiny: ‘During the Cold War, only one in five elections outside of the consolidated democracies was monitored by international observers, whereas by 2010 the share of monitored elections increased to four in five.’²⁹⁷ Even when elections have been successfully stolen or an elected regime ousted from power, states, international organisations, and non-government organisations frequently (although not always) issue critiques that make it clear that international standards have been violated.²⁹⁸ In some cases, their reactions result in the restoration of elected regimes; in other cases, not. But the success of these critiques is not the relevant measure of how legitimacy pronouncements affect the non-intervention doctrine; rather, it is that international standards of regime legitimacy, are consistently reaffirmed and demands made that they be respected. The IDI policy of deliberate indifference is difficult to reconcile with this body of state and international organisation practice.

It is true that, in the past decade, there has been a widely noted decline in both electoral democracy and human rights observance.²⁹⁹ But the effect of these developments should not be overstated. In particular, they do not support a claim that international law has returned to the era of IDI’s agnosticism on regime legitimacy. First, these developments have not involved states repudiating the regional ‘democracy protection’ regimes that most clearly codify principles of regime legitimacy. The OAS regime, for example, was employed in 2019 to deny a seat to the ambassador appointed by Venezuelan President Nicolas Maduro on the grounds that Maduro’s election ‘lacked legitimacy’.³⁰⁰ Also in 2019, the African Union regime was used to condemn ‘the overthrow of the democratically elected President’ of Sudan.³⁰¹

Second, recent anti-democratic practices may be seen as adaptive strategies that reflect the success of the first generation of pro-democratic norms and institutions. The two most important events those norms sought to confront – military coups and blatant election-day fraud – have dramatically declined in recent years.³⁰² Military coups have been replaced in many instances by what Nancy Bermo has called ‘promissory coups’, in which regimes ‘frame the ouster of an elected government as a defence of democratic legality and make a public promise to hold elections and restore democracy as soon as possible’.³⁰³ Election-day fraud has diminished in the face of extensive international election monitoring. It has been replaced as a tool of democratic usurpation by ‘a range of actions aimed at tilting the electoral playing field in favour of incumbents’.³⁰⁴ If the decline of coups and blatant election fraud is understood as a ‘rational response[] to local and international incentives’,³⁰⁵ then international law confronts not a wholesale challenge to democratic legitimacy principles but a problem of normative and institutional design. Instruments such as the Inter-American Democratic Charter, invoked in the Venezuela case, may need to be reworked to confront the rise of smarter and more adaptive anti-democratic actors.

Third, the relevant baseline for purposes of reassessing the IDI view is not the mid-2000s, when the ‘democratic recession’ arguably began; rather, it is the Cold War era in which democratic legitimacy and liberal principles of governance were almost wholly absent from international legal discourse.³⁰⁶ This was the legal milieu in which IDI’s mandatory agnosticism arose. None of the legal infrastructure now supporting democratic legitimacy existed (or could have existed) at that time.

Finally, the argument for a departure from IDI and Nicaragua is not that every NIAC presents a clear binary choice between democratically legitimate and illegitimate factions; rather, it is that there is enough international consensus on democratic and human rights norms, as well as enough information from reliable sources on their implementation, that international law need no longer avoid, in every case, asking whether a regime is democratically legitimate. That the Security Council cannot make a binary choice in all cases is not a reason to return to the Cold War approach of not giving an answer in any case.

2. The Defence of the IDI View

Olivier Corten argues that the IDI’s 1975 Wiesbaden Resolution III ‘reflects established practice’ because ‘states never avowedly support a government acting against its own population’.³⁰⁷ But this claim begs two questions. First, which practice is being consulted? During the Cold War, the major powers intervened regularly to support favoured governments and insurgents. Commentators, as noted earlier, despaired that this practice rendered the law incoherent.³⁰⁸ The IDI Resolution was, in effect, a remedial response to these interventions.

Whatever one thinks of the wisdom of this approach, the 1975 Resolution cannot be cited as a reflection of custom; it was instead a reaction to uncertain and unhelpful state practice. Division among members of the IDI reflected the law’s uncertainty. The vote was sixteen in favour, six opposed, and sixteen abstaining.³⁰⁹ Then Special Rapporteur Dietrich Schindler believed that the prohibition of assistance to governments during civil wars under Article 2 of the Resolution ‘deviates from the classical rule, according to which assistance to the established government is lawful, at least until when the third state recognizes the insurgents as belligerent’.³¹⁰ Gerhard Hafner, the next IDI Special Rapporteur on the subject, believed that ‘there was no certainty on whether the [1975] resolution reflected lex lata or proposed articles de lege ferenda’.³¹¹ Reviewing the 1975 Resolution and its 2011 successor, the Rhodes Resolution II, Georg Nolte similarly concluded that, because of divisions among IDI members, ‘the 1975 resolution of the Institut did not lead to a clarification of existing law’.³¹²

What of practice since 1975? Corten reviews ‘a few emblematic cases’.³¹³ In each – Yemen, Iraq and Syria, Mali, and The Gambia – he analyses the reaction of states and international organisations separately, with no explanation of how the two relate to each other. The practice of international organisations seems not to enter into the legal conclusions to be drawn from each case. This is a highly incomplete picture. These four cases in fact demonstrate the importance of international organisation practice.

The UN Security Council was deeply involved in each case and, contrary to the IDI view, did not condemn any of the invited interventions. Indeed, as Corten notes, recent Council practice evidences ‘a new arrangement consisting of the informal validation of interventions by consent’.³¹⁴

Take the 2017 ECOWAS intervention in The Gambia. ECOWAS, the African Union, and the Security Council all condemned President Yahya Jammeh’s refusal to leave office after losing an election to Adama Barrow. Each international organisation also declared Barrow the legitimate president of the country.³¹⁵ ECOWAS troops responded to Barrow’s request for assistance, precipitating Jammeh’s departure. Shortly thereafter, the Security Council expressed support for the ECOWAS process and for the African Union Peace and Security Council’s declaration that ‘outgoing President, Yahya Jammeh, will cease to be recognized by the AU as legitimate President of the Republic of the Gambia’.³¹⁶ Surely the most distinctive feature of the Gambian episode is the absence of individual states as the dominant actors. Two regional and one global international organisation spoke essentially in unison, from the initial declaration of Jammeh’s illegitimacy to their support for an intervention. Yet Corten excludes these collective actions from relevant practice.

In sum, an assessment of practice that (i) assumes the 1975 IDI Resolution reflected customary law of the time, (ii) focuses on a few high-profile recent cases, and (iii) wholly excludes the reaction of international organisations is simply incomplete. Why not assess all practice, both of individual states and of international organisations? The need for such a holistic, empirical assessment, done with methodological rigour, is the starting premise of this chapter.

The second question raised by reliance on the IDI Resolution is why the mere fact of external support for ‘a government against its own population’ should violate citizens’ right to self-determination. Corten endorses the IDI view of self-determination as a legal fiction, which supports the opportunity for citizens to choose their own government but ignores any actual choice they may have made. According to this view, while elections may be acts of internal self-determination, they do not alter the barrier erected by external self-determination to exclude invitations by those who win elections.³¹⁷

This distinction, rooted in the Cold War, was attractive when claims of democratic legitimacy were often little more than subjective assertions by the intervening states. But the post-Cold War era has mostly (although not completely) erased the line between internal and external notions of democratic legitimacy. Most elections in emerging or nascent democracies are monitored by outside groups. Regional organisations in the Americas and Africa have well-established legal regimes to respond to interruptions of democratic government. The Security Council regularly congratulates electoral winners and emphasises that their victories bestow an entitlement to govern.

In such cases, there is no need to invoke self-determination as a legal fiction to protect a hypothetical popular ‘choice’. Citizens voting in an election will have made an actual choice. The legitimacy of that choice will have been verified by multilateral actors. The international community is thus fully aware of citizens’ preferences in a conflict pitting ‘a government against its own population’. To pretend that choice is unknowable to outsiders, and therefore in need of protection against their subjective judgments, is to ignore the immense body of international practice directed precisely at that resolving question. To put it another way, there is no need, in such cases, to invoke external self-determination to protect the integrity of internal self-determination.

The Gambia is again illustrative. ECOWAS and the African Union invoked their democracy protection regimes to declare Barrow the winner of the election. Those regimes, when invoked, are premised on the organisations’ ability to distinguish between democratically legitimate and illegitimate regimes. Following on from those determinations, the UN Security Council affirmed, in Resolution 2337, the primacy of actual electoral choice, urging ‘all Gambian parties and stakeholders to respect the will of the people and the outcome of the election which recognized Adama Barrow as President-elect of the Gambia and representative of the freely expressed voice of the Gambian people as proclaimed by the Independent Electoral Commission’.³¹⁸ And the Council extended this internal act of self-determination externally, urging ‘countries in the region’ to ‘cooperate with President Barrow in his efforts to realize the transition of power’.³¹⁹

Of course, these are the easy cases. Others exist on a spectrum, ranging from cases of undoubtedly free and fair elections monitored by objective observers, the results of which are affirmed by international organisations, to those in which election outcomes are disputed and no multilateral institutions identify victory by one party or another. Then there are breakdowns in democratic institutions short of defying electoral outcomes. In such cases, the nature of the popular choice is much less clear. As a result, the lawfulness of an intervention by one of the disputing parties will have only a tenuous connection to principles of democratic legitimacy.

But this distinction between easy and harder cases is one of fact. All internal conflicts are not the same. An inquiry, where possible, into the fairness of elections and the position of international organisations should allow principled distinctions to be made. Where international opinion is united, no further resort to the legal fiction of protecting ‘choice’ is necessary.

B. The Contribution of UN Security Council Practice

The data discussed earlier makes clear that the Security Council is now a consistent presence in evaluating the lawfulness of invitations. As Olivier Corten highlights in his chapter title, the recent era has been marked by ‘the expanding role of the UN Security Council’.³²⁰ This is a marked change from the pre-1990 period. But the data do not reveal any consistent patterns in Council views on IDI and Nicaragua, the two theories that might cover all cases. This is in contrast to evident Council support for anti-terrorist interventions and, in an admittedly few cases, pro-democratic interventions.

How should international law react to, and perhaps assimilate, this body of Council practice? Because Council practice on consensual intervention is not yet uniform, this question is one of legal process and not substantive doctrine. In this section, I discuss two possible responses: viewing the practice as a lex specialis, with no relevance to the ius ad bellum; or – quite differently – viewing the practice as evidence of customary law directly relevant to the ius ad bellum.

1. Council Practice as Lex Specialis

The first position sees Council practice as a lex specialis, deriving from the Council’s unique power to bind conflict parties and to legitimise or delegitimise particular uses of force. The practice can be seen as a lex specialis in that its effects are limited to the conflicts and actors the Council addresses in specific resolutions. Council practice, according to this view, would have no effect on the direction or substance of customary international law.

The UN Charter describes the Security Council has having competence to deal with particular incidents threatening or breaching international peace and security, not authority to alter the law applicable to state behaviour more generally.³²¹ To be sure, the Council has extraordinarily broad political authority to resolve particular conflicts. But, the lex specialis view would assert, one should not mistake a broad authority to resolve particular conflicts for an authority to reach beyond their resolution and shape the direction of the customary ius ad bellum. If that were the case, one would expect some evidence of opinio iuris. But neither the Charter nor the resolutions underlying the data contain any evidence of an intent to affect the content of custom.

The core of the lex specialis argument is a distinction between the powers of the Council and those of states acting individually. The Council enjoys the unique authority to deem an intervention lawful or unlawful. In Thomas Franck’s description, when the Council addresses an armed conflict, it acts as a kind of jury, hearing evidence both for and against the legality of state action and coming to an authoritative conclusion.³²² This unique power is by design. The Council’s expansive powers derive precisely from it not being a self-interested state with a national policy agenda and territory to protect.³²³ The Council, by definition, cannot materially benefit from its decisions, either in specific cases or more generally through the interpretations of international law underlying its decisions. Unlike states, whose authority to use force is extraordinarily limited precisely because their self-interest poses a danger of abuse, the Council’s authority is broad because it is understood to lack narrow self-interest.³²⁴

The separate domains of state and Council authority to judge and utilise military force, the argument goes, should extend to the normative consequences of their respective actions. If the Council has vastly more discretion to authorise or employ force, how can its actions inform the narrower legal grounds governing state behaviour? A Council decision to permit an elected, but ousted, regime to invite foreign assistance, for example, does not support that regime’s ability to issue an identical invitation absent Council authority. The collective judgment of the Council in such a case cannot be delegated to states, which would be the consequence of interpolating Council practice into the ius ad bellum. Indeed, the Council’s ability to authorise force in circumstances in which a state could not act is an important argument against expanding the realm of unilateral action. Why expand that realm, with all its attendant dangers of self-judgement and motivated reasoning, when a much safer multilateral option exists? This point is often made in reference to humanitarian intervention: it is precisely the Council’s willingness, in some cases, to authorise force in response to mass human rights violations that negates the existence of states’ unilateral right to engage in the same action.³²⁵

Beyond the argument that the Council and states inhabit separate normative domains, there are also process-based critiques of treating Council actions as evidence of customary international law. As the United States argued in a comment to the ILC:

It is axiomatic that customary international law results from the general and consistent practice of States followed by them out of a sense of legal obligation. This basic requirement has long been reflected in the jurisprudence of the International Court of Justice. It is also reflected in the practice of States in their own statements about the elements required to establish the existence of a rule of customary international law.³²⁶

More specifically, one can argue that Security Council resolutions lack critical attributes that have led the ICJ and others to treat certain General Assembly resolutions as evidence of customary international law.³²⁷ Those General Assembly resolutions are structured much like treaties, setting out broad and prospective rules of general application. Articulating broad standards is precisely the purpose of General Assembly resolutions such as the Friendly Relations Declaration.³²⁸ Council resolutions, however, are almost always conflict-specific. Moreover, while every state in the world may vote on a General Assembly resolution, the Council is a small, elite body.

Finally, knowing that Council resolutions may affect custom might cause some Council members to vote against resolutions they might otherwise support. Many conflicts on the Council’s agenda are of marginal strategic significance to at least some Council members. Those members may nonetheless support Council initiatives for the simple reason that there are no compelling reasons to withhold their support. But knowing that provisions of such resolutions may become building blocks for new or enhanced customary norms could change that calculus and lead to negative votes.

Although the lex specialis view is clear about the role Security Council practice on consensual interventions should not play, it is less clear about how it should be relevant, if at all, to international law. Perhaps it could serve a quasi-precedential function – not in the sense of formal stare decisis, but as a repertoire of successful best practices. If the Council is, as is often remarked, an essentially political body, perhaps this kind of political consensus on acceptable grounds for invitations is the most the body can offer.

2. Council Practice as Evidence of Customary International Law

The second view takes the opposite perspective: Council practice can serve as evidence of customary international law for purposes of understanding norms on consensual intervention.³²⁹ This view, which colleagues and I have discussed at length elsewhere, relies on three propositions.³³⁰ First, when the Council imposes obligations on conflict parties, it acts as an agent for all UN member states. Article 24(1) of the UN Charter provides that member states ‘confer on the Security Council primary responsibility for the maintenance of international peace and security, and agree that in carrying out its duties under this responsibility the Security Council acts on their behalf’.³³¹ Viewing the Council as an agent for member states embodies the logic of the Charter’s collective security regime. The Charter famously discarded the traditional view of armed conflict as primarily (and often solely) the concern of the warring parties, providing instead that all member states share an interest in maintaining the peace.³³² The agency theory ensures that the official positions of member states on conflicts do not diverge from executive decisions of the Council on the same conflicts by making the two legally indistinguishable. The Special Court for Sierra Leone relied on an Article 24(1) agency theory to hold that an agreement between Sierra Leone and the United Nations was, as a result of Council approval, ‘an agreement between all members of the United Nations and Sierra Leone’.³³³

Does the agency theory mean that member states have delegated to the Council a capacity to contribute to customary international law? In its ILC submission, the United States argued emphatically not: the mandates of international organisations are ‘carefully negotiated treaties’ that ‘rarely, if ever’, provide an express authorisation ‘that the organization exercise the powers of member states to generate practice for purposes of customary international law’.³³⁴ This was obviously true when the UN Charter was negotiated, but the US position seems anachronistic today.

First, consider the consequences, since 1990, of states having delegated to the Council the authority to address an extraordinary range of legal questions arising from NIACs but withholding any customary international law consequences of that delegation. Those consequences (i.e., evidence of custom) would not be attributable to the Council. But neither would they be attributable to member states who delegated authority to the Council to act on their behalf. For customary international law purposes, they would be neither acts of the Council nor acts of individual member states. An entire realm of rich international practice on NIACs would be lost to customary international law. That idea would lead to an unacceptable result, as my co-authors and I have written elsewhere:

For member states to authorize the Council to act on their behalf but withhold normative consequences of that action would consign the ‘acts concerned’ to a legal black hole: U.N. member states would not be acting in their own capacities, and thus no ‘state practice’ would be created, but, with normative consequences withheld, the Council’s corporate acts would make no contribution to customary law. As a result, no actor could claim as its own the potentially significant contributions to custom.³³⁵

Second, new data show the Council has been involved in almost all contemporary NIACs.³³⁶ The Council has addressed NIACs in every year, in every region, of varying duration, of varying numbers of actors, of varying battle deaths and civilian casualties, at various points in the conflicts, and both inside and outside the spheres of influence of every hegemonic state. No state or group of states comes close to matching this breadth of practice. The Council’s involvement in NIACs has also been remarkably deep, ranging from simply imposing obligations of conduct, to dispatching peacekeeping missions, to imposing sanctions. To take sanctions as an example, only four of the sixteen Council sanctions regimes in place in 2017 targeted state actors exclusively; the rest targeted non-state actors or both state and non-state actors. Obviously, no state or group of states has addressed NIACS more broadly or more comprehensively.

Viewing Council practice as evidence of custom may appear controversial when applied to customary norms not directly linked to the Council’s core set of competences. For example, the Council now regularly takes positions on issues of human rights, IHL, and treaty law. The Charter prescribes no special role for the Council in these areas. But finding evidence of custom should be much less controversial when the Council addresses the ius ad bellum. The UN Charter both revolutionised the substance of that law and empowered the Council to respond to virtually every significant use of force. Of course, the Charter’s scheme for authorised force was never implemented, but the UN system’s evolutionary adaptations – primarily peacekeeping and ‘authorised operations’ – have retained the Security Council as the central decision-maker.³³⁷ The idea of Security Council primacy in evaluating uses of force is even accepted by those states that have, on occasion, supported unilateral force when the Council is unable or unwilling to act.³³⁸

Thus Council actions and inaction have become central to debates over the non-annexation norm,³³⁹ self-defence against non-state actors,³⁴⁰ humanitarian intervention,³⁴¹ and other ius ad bellum questions. Participants in these debates do not argue that while the Council is the central actor in contemporary peace and security law, its views are irrelevant to the substance of that law when applied purely between states.

Of course, not all Council resolutions can be understood as interpretations or applications of Article 2(4) and its doctrinal progeny. Many scholars argue that, when the Council describes state acts as ‘threats to the peace’ under Charter Article 39, it may venture beyond the ius ad bellum and exercise a general discretion that is unmoored from specific norms restricting state behaviour.³⁴² But it is difficult to argue that the legality of an invitation to intervene falls into a zone of Council discretion beyond the ius ad bellum. Among other problems with such a claim is that none of the condemnatory resolutions in the dataset describes the interventions as a ‘threat to the peace’.

3. Critiques of Council Practice as Evidence of Custom

Olivier Corten’s position on the legal significance of Security Council practice is puzzling. On the one hand, he argues in his contribution to this volume that the Council has become a central actor in addressing invitations to intervene. In the case of Iraq, for example, he describes the Council as playing a ‘decisive’ role.³⁴³ On the other hand, he consigns that practice to a lex specialis, walled off from the development of customary international law. This view – that the Council is politically paramount but legally irrelevant – presents several problems.

First, it stifles the development of customary law, disconnecting its evolution from the reality of international practice. The Security Council has passed resolutions on 80 per cent of NIACs started after 1990.³⁴⁴ More specifically, it has been omnipresent in three of the four conflicts Corten analyses: Yemen (twelve resolutions since 2015³⁴⁵), Mali (thirteen resolutions since 2013³⁴⁶), and The Gambia (one critical resolution in 2017 endorsing the ECOWAS intervention³⁴⁷).³⁴⁸ As Corten himself observes, ‘the Security Council intervened in all the recent case studies on which this chapter [focuses]. By adopting resolutions, it pronounced on the authority that was entitled to give its consent, and in parallel on the legitimacy of the object and effects of the intervention.’³⁴⁹

States, in other words, have chosen the Council as their vehicle for articulating and executing policies towards these conflicts. Did they also choose to deprive Council actions of any relevance to custom? No evidence exists to support this claim. If it did, the results would be unfortunate. The customary law on invitations would either stagnate (because it would not take into account the most consequential actor involved in NIACs) or develop in directions reflecting only minority views (because it would credit only acts of the few states directly engaged with NIACs independently of the Council). In either case, custom would become marginalised as it diverges from the reality of international practice. Perhaps for this reason, international courts and other bodies have regularly cited Security Council practice as evidence of customary law.³⁵⁰

Second, and relatedly, analysing interventions without reference to legal frameworks established by the Council allows arguments to be raised that the Council had already foreclosed. In Mali, for example, the Council determined that the transitional government was competent to invite both a regional force and French troops. This was despite the government not controlling substantial portions of the national territory and lacking democratic legitimacy. In Yemen, the Council repeatedly supported the legitimacy of President Hadi, who issued the invitation to the GCC. Similarly, in The Gambia, the Council sided with two regional organisations in affirming the democratic bona fides of elected President Barrow. And in Iraq, the Council directly approved the post-occupation government’s invitation to a multinational force to assist in its internal conflict with the Al-Mahdi Army and other forces.³⁵¹ With the exception of Resolution 2337 on The Gambia, all of these determinations came in the form of resolutions passed under Chapter VII of the UN Charter.³⁵²

What legal issues are left for individual states to resolve after the Council took these steps? Very few – and that was the Council’s intent. In each case, the Council sought to unify the international community around a single legal conclusion. Yet, without using Council practice as evidence of custom, the issues decided by the Council can be treated (as does Corten) as open questions. In the case of The Gambia, for example, Corten asks whether the election won by Adama Barrow was really free and fair, and whether Barrow’s failure to control Gambian territory at the time he invited in ECOWAS troops presented ‘a problem with respect to the condition of effective control being exercised by the inviting authority’.³⁵³ But the Council had addressed both these issues and hence the result is confusion. How can Council decisions reflect consensus views of the international community at the moment they are issued but remain open to criticism for the purposes of assessing their customary law consequences?

Third, ignoring Council practice seeks to prioritise state practice that may not, in fact, exist, or may exist only at the margins. When the Security Council addresses NIACs in a comprehensive fashion, states have less of a need to take their own unilateral actions, or even to comment on actions by other states. It is quite telling that Corten’s comprehensive review of the four cases contains no mention of unilateral actions or statements by Germany, Italy, Spain, Turkey, Japan, Nigeria, South Africa, Canada, Australia, Brazil, Mexico, Pakistan, Indonesia, or North or South Korea (to mention only a few major military powers). The single cited statements by China, Egypt, and India relate only to Mali, and they do not address the French intervention. No statements of Saudi Arabia appear beyond those related to its own intervention in Yemen.

This should not be surprising. With these and other states ceding leadership to the Council, any need for unilateral action on the four conflicts diminished substantially. So customary law will not be shaped by these states’ actions or statements. But if Council actions are excluded, the authority they ceded to the Council will also not produce relevant practice. As a result, there will be little or no international practice relevant to custom emerging from these cases. This is a highly troubling outcome. It is the legal black hole to which I referred earlier.

One might respond that statements in debate over Council resolutions represent individual state practice. But there are two problems with this response: first, only fifteen states sit on the Council, meaning its debate cannot contain a representative sampling of state opinions on a given conflict; and, second, these are statements divorced from state action. Any acts resulting from Council debate would be corporate acts of the United Nations, not of the individual states making the statements.

VII. Conclusions

The four prevalent theories on the legitimacy of intervention by invitation emerged from specific historical circumstances. Those settings carried with them a set of assumptions about the relations between inviting and intervening states, the capacity of the international community to respond collectively to such interventions, and the propriety of pursuing certain substantive goals by military force. We have not yet left the historical moments in which the anti-terrorism and democratic legitimacy theories were incubated. Perhaps as a result, as the data shows, the Security Council has been favourably disposed towards both, although it has not accepted the democratic legitimacy view in all cases.

The anti-terrorism and democratic legitimacy views, however, apply only in a narrow set of circumstances. The IDI and Nicaragua views are the main competitors for a general framework regulating consensual interventions, potentially applicable to all interventions. The IDI view appears inconsistent not only with the Security Council’s approval of several interventions in NIACs but also its own record of intervening in NIACs in a variety of ways. This practice simply cannot be reconciled with the idea that civil wars are purely internal affairs. After almost thirty years of the Council finding NIACs to be a ‘threat to the peace’ and recommending liberal democratic institutions for post-conflict societies, can it really be said that all locally chosen options for governance are due equal respect?

But neither is the Nicaragua view wholly supported by Council practice. In several cases, the Council has disapproved of interventions requested by governments in effective control of their territories. While the Nicaragua view rejects the IDI view that governments in civil wars lack the capacity to invite outside forces, it defers questions of legitimate governance to other bodies of international law. As a result, Nicaragua may not, in practice, result in approving all interventions requested by a regime in effective control. International law on recognition of governments could well deem some of those requests illegitimate.

But the international reaction to post-Cold War interventions is not significant primarily for its acceptance or rejection of either theory; rather, it demonstrates that the Security Council has assumed a central role in passing on the legality of particular interventions. This collectivisation of global reaction requires us to be sceptical of theories premised precisely on the unavailability of such mechanisms. While one might dismiss Council practice as a lex specialis of limited relevance to norms resulting solely from state practice, this is not, in my view, the most compelling approach. Instead, Council practice should occupy a position in customary international law commensurate with the primacy UN member states have accorded the Council in responding to NIACs.