Skip to main content
    • Aa
    • Aa


  • Access
  • Open access


      • Send article to Kindle

        To send this article to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle.

        Note you can select to send to either the or variations. ‘’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        Authorising humanitarian intervention: a five-point defence of existing multilateral procedures
        Available formats
        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your Dropbox account. Find out more about sending content to Dropbox.

        Authorising humanitarian intervention: a five-point defence of existing multilateral procedures
        Available formats
        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your Google Drive account. Find out more about sending content to Google Drive.

        Authorising humanitarian intervention: a five-point defence of existing multilateral procedures
        Available formats
Export citation



Even scholars who support multilateralism in principle frequently question the value of securing approval from existing multilateral bodies for humanitarian intervention. The United Nations (UN) and regional organisations such as NATO, the argument goes, are far from democratic; furthermore, multilateralism is often a recipe for doing nothing; therefore, unauthorised intervention should be permissible in circumstances of ‘humanitarian necessity’. This article maintains that although today’s multilateral organisations and related procedures for authorising armed intervention may be suboptimal, they have significant output legitimacy. First, existing authorisation procedures reduce the risk of destabilising conflict spirals among powerful states. Second, they diminish the likelihood that humanitarianism will be used as a pretext. Third, they reduce epistemic problems concerning the identification of a just cause for intervention and thus the risk of accidental abuse. Fourth, they minimise the ‘moral hazard’ of humanitarian intervention. Finally, compliance with multilateral procedures is increasingly required for successful peacebuilding. This leads me to conclude that humanitarian warfare should always be authorised by the UN or regional multilateral organisations.


Over the last two and a half decades, powerful states – primarily the United States and its allies – have intervened for what are widely regarded as humanitarian purposes in northern Iraq (1991), Somalia (1992), Haiti (1994), Bosnia (1995), Kosovo (1999), East Timor (1999), Sierra Leone (2000), Cȏte d’Ivoire (2002), Liberia (2003), Libya (2011), and the Central African Republic (2013). Most of these interventions were carried out with the approval of the United Nations and/or regional multilateral organisations such as NATO and the African Union. By why should states seek multilateral approval for their humanitarian interventions? Why can’t the ostensible humanitarian purpose be a sufficient source of normative legitimacy?

Even scholars who support multilateralism in principle frequently question the value of securing approval from existing multilateral bodies for humanitarian intervention. The United Nations and regional organisations such as NATO, the argument goes, are far from democratic; furthermore, multilateralism introduces severe coordination costs and is often a recipe for doing ‘too little, too late’. Many analysts conclude that until existing multilateral organisations become more accountable, more democratic, and less prone to deadlock, unauthorised (that is, unilateral) intervention should be permissible in circumstances of ‘humanitarian necessity’. 1

This article maintains that although today’s multilateral organisations and related procedures for authorising military intervention may be suboptimal from the point of view of input legitimacy (which refers to the ways in which decisions are reached), they have significant output legitimacy (which refers to the achievement of beneficial outcomes).

First, I argue, current authorisation procedures reduce the risk of destabilising conflict spirals among powerful states resulting from misperception of one another’s intentions. Second, by exposing the fraudulent use of human rights rhetoric, these procedures diminish the likelihood that humanitarianism will be used as a pretext for self-serving interventions. Third, they reduce epistemic problems related to the identification of a just cause for intervention and thus the risk of accidental abuse. Fourth, they minimise the ‘moral hazard’ of humanitarian intervention, which occurs when a strategic and normative context that favours international intervention motivates rebel groups to deliberately provoke government atrocities. Finally, compliance with multilateral procedures is increasingly required for international burden sharing during the post-bellum phase and thus for successful peacebuilding.

The first three benefits have been identified to some extent in previous research, but this article offers a more systematic discussion of related causal mechanisms, specifies scope conditions, and provides empirical examples to illustrate the argument. The last two benefits are examined for the first time in this article. I focus primarily on the United Nations Security Council (UNSC); however, I also suggest that seeking and securing approval from regional organisations such as NATO or the African Union (AU) may offer similar functional benefits.

As James Pattison notes, ‘to have the right to intervene – for its intervention to be permissible – an intervener needs to be likely to be effective’, and in the context of humanitarian intervention, effectiveness should be understood primarily in terms of promoting ‘the enjoyment of human rights’. 2 I am sceptical that unauthorised intervention can be effective at promoting human rights. Even well-intentioned interveners may end up making the human rights situation worse, and the risk is heightened considerably in the absence of multilateral approval and support. This output-oriented perspective leads me to conclude that humanitarian warfare should always be authorised by the UNSC or appropriate regional multilateral organisations.

The article is structured as follows: Part One reviews the classic argument in favour of multilateral approval linked to global democracy and prominent criticisms of this line of reasoning. Part Two discusses some of the criteria for just humanitarian intervention and clarifies key features of appropriate multilateralism. Part Three presents a detailed five-point defence of existing multilateral authorisation procedures from the perspective of output legitimacy. Part Four, finally, explicitly considers whether there should be a ‘unilateralist exception’ in cases of putative humanitarian emergency.

The traditional argument and its critics

The classic argument in favour of multilateral approval holds that it is desirable because it signals compliance with widely-accepted international norms and comes close to expressing the ‘general will’ of international society. As Inis Claude observed, under circumstances in which democracy has become a central measure of legitimate decision-making, ‘the United Nations [is viewed] as the most impressive and authoritative instrument for the expression of a global version of the general will’. 3 Among normative political theorists, Michael Walzer has argued perhaps most explicitly that faith in multilateral decision-making results from the association of multilateralism with ideals of global democracy:

Behind this preference [for multilateralism] is an argument something like Rousseau’s argument for the general will: in the course of a democratic decision procedure, … the particular interests of the different parties will cancel each other out, leaving a general interest untainted by particularity. As with individuals in domestic society, so with states in international society. 4

A variant of this argument is that multilateral rules and procedures ought to be followed because they reflect collective agreement on the governance of international society. 5 Others emphasise that multilateral approval is desirable because, apart from reflecting agreement on procedures (either globally or within a particular region), it demonstrates that ‘the values at stake are commonly held’ 6 – or, put differently, it signals an action’s ‘soundness or conformity with other [communal] norms’. 7

The view of multilateral bodies, and the UNSC in particular, as representatives of community interests and embodiments of a global ‘general will’ has been subjected to significant criticism. Robert Keohane, for instance, argues that the UNSC’s composition and modus operandi, which allows any of its five permanent members to veto decisions at will, ‘does not reflect any principled set of criteria for representation … [and] cannot be justified on the basis of principles of either democracy or elementary fairness and reciprocity’. 8

Furthermore, the veto rights of permanent members of the UNSC and the demanding supermajority requirements of regional multilateral bodies (NATO, for example, requires that decisions are adopted by consensus, and the AU requires a two-thirds majority) often result in institutional deadlock. Even if military intervention is approved at the end of a cumbersome negotiating process, it may be too late to have much of a positive impact, as thousands of civilians may already have died and the worst of the humanitarian crisis may be over. In short, critics contend, existing multilateral institutions, and specifically their procedures for authorising the use of force, are ‘morally problematic’. 9

Therefore, it is frequently argued, bypassing existing multilateral bodies should be viewed as permissible in circumstances of humanitarian necessity. 10 In the longer term, multilateral bodies ought to be fundamentally reformed. Proposals range from establishing a ‘concert of democracies’ that would complement or replace the UNSC in authorising armed intervention, 11 to creating a new global court that ‘would grant or refuse “warrants” to intervene’, 12 to setting up a democratic world parliament that would authorise interventions by a ‘cosmopolitan UN force’. 13 But until such more legitimate institutions emerge, the argument goes, we should be willing to countenance intervention without formal multilateral approval. During this interim period of indefinite duration, the legitimacy of intervention should be assessed through a loose jurying process involving state governments and non-state actors as well as multilateral bodies. 14

An output-oriented perspective

Reforming multilateral bodies in the direction of greater democracy, accountability, and effectiveness is certainly a worthwhile objective. A strong case can be made that, at a minimum, the UNSC should be made more representative and its decision-making processes more transparent, as this would increase input legitimacy. However, the successful implementation of such reforms is likely to take a long time. This article is mainly concerned with the interim period from now until then.

Furthermore, theorists who critique existing multilateral institutions on grounds of their unrepresentative character may be applying an exceedingly one-sided normative standard. I argue that instead of becoming fixated on the deficiencies of these institutions and related authorisation procedures from the point of view of democratic theory, we should focus on their ability to limit the negative consequences of intervention and help achieve desirable outcomes (or outputs).

Just humanitarian intervention

I follow the conventional understanding of humanitarian intervention as the deployment of ‘military force across borders for the purpose of protecting foreign nationals from man-made violence’. 15 A just humanitarian intervention needs to comply with the jus ad bellum proportionality requirement – that is, the good that will follow must outweigh the inevitable pain and destruction likely to result from the international use of force. Concretely, this means one has to be confident that the human rights balance will be positive: significantly more civilians need to be saved from imminent death or physical injury than are killed or injured as a direct consequence of the intervention. 16

Traditional humanitarian interventions that merely used to provide vital aid to at-risk civilians, while avoiding any action related to the political causes of civilian suffering, were unlikely to leave the target society noticeably better off. Indeed, the impartial provision of humanitarian aid without further political measures often fed into and prolonged local conflicts. 17 Contemporary jus post bellum requires that interveners look beyond saving threatened civilians in the immediate present and aim to achieve sustainable improvements in physical integrity rights. 18 Closely related to this, the Responsibility to Protect doctrine is widely understood to involve a responsibility to rebuild war-torn societies, with the ultimate goal of ‘build[ing] a durable peace’. 19

Research indicates that durable peace can be achieved by facilitating the emergence of more inclusive domestic political institutions. 20 Concretely, therefore, humanitarian interveners should aim to stabilise war-torn societies by promoting domestic power-sharing arrangements or, perhaps more controversially, enforcing deeper changes in the domestic political regime.

As Walzer argues, ‘in the case of humanitarian intervention, jus post bellum involves the creation of a new regime, which is, minimally, nonmurderous. And it is more than likely that the creation of a new regime will require some period, perhaps an extended period, of military occupation’. 21 Regardless of whether one embraces the goal of post-bellum regime change, ‘durable peace’ sets the bar for successful intervention rather high: interventions that cannot prevent a relapse into large-scale civil violence, at least over the medium term, should be viewed as unsuccessful. It follows that unless outsiders also have a workable strategy for longer-term peacebuilding, they should refrain from military intervention altogether. 22

Appropriate multilateralism

In terms of multilateral approval, I focus chiefly on the UN Security Council. The UNSC has primary responsibility for international peace and security under the UN Charter and remains the sole body that can authorise humanitarian intervention under international law. From an international law perspective, regional organisations such as NATO or the AU cannot authorise the use of force. Armed intervention without UNSC authorisation is compatible with international law only when it is carried out in self defence; even then, as Article 51 of the Charter notes, such action is admissible only for an interim period, until the Security Council has reviewed the situation and has ‘taken measures necessary to maintain international peace and security’. 23

I acknowledge the importance of compliance with international law and the centrality of the UNSC. Nevertheless, I also consider, as a second best, the possible functional benefits of seeking and securing approval from standing regional organisations. To qualify as an appropriate multilateral body for the purpose of authorising military intervention, a regional organisation needs to have a mandate in the field of international security (hence purely economic organisations, for instance, do not qualify). In addition, it needs to function on the basis of generalised principles of conduct – that is, ‘principles which specify appropriate conduct for a class of actions, without regard to the particularistic interests of the parties’. 24 The approval has to be granted by the organ designated for that purpose in the organisation’s charter: for NATO, it is the North Atlantic Council; for the AU, the Assembly of Heads of State and Government; for the Organisation of American States (OAS), the organisation’s Permanent Council.

A further requirement of appropriate multilateralism is that there be no excessive concentration of decision-making power: the organisation’s most powerful member should always have to take into account the interests and opinions of other members rather than simply being able to ram through its preferred policies. 25 This is essential if multilateral procedures are to restrain powerful states and limit the risk of abuse. For instance, in the 1960s, the OAS was too clearly dominated by the United States. When the OAS approved Washington’s intervention in the Dominican Republic in 1965, that helped US policymakers build support at home, but they did not have to adapt their policy in any way. 26 The power distribution within the OAS has become less blatantly skewed in favour of the United States in recent decades, as demonstrated by the fact that the organisation explicitly condemned the unilateral US intervention in Panama in 1989. 27

Several other regional organisations – including the AU, the Arab League, and NATO – are far from being merely instruments of hegemonic power projection for their most powerful member(s). Within NATO, the United States is certainly the most influential state, but the power imbalance has never been large enough for Washington to be able to use the organisation to simply rubber-stamp its preferred policies. 28 To secure NATO’s approval for military interventions in the Balkans, the United States had to engage in sustained and often difficult negotiations with its European partners, taking their views into account and making significant concessions. In the 2003 Iraq case, reflecting widespread European scepticism about American goals, NATO always remained very far from endorsing the US-led invasion. 29

Finally, appropriate multilateralism excludes the possibility of endorsement by improvised ‘coalitions of the willing’. Powerful states with the capabilities to project military force abroad often find it easy to cobble together a nominal multinational coalition to manage public perceptions. As Marc Grossman, who served as a senior US State Department official in several administrations, explains, ‘Some coalition is always available’. 30 Efforts to secure the backing of improvised coalitions do not meaningfully constrain the intervener’s freedom of manoeuvre and do not help overcome information deficits; thus, they are unlikely to promote any of the benefits discussed here.

The benefits of compliance with existing multilateral procedures

In what ways, then, do existing multilateral procedures limit the negative consequences of intervention and help achieve desirable outcomes? Specifically, I ask, what impact can compliance with these procedures be expected to have on the protection of physical integrity rights? I also address two related questions: how might physical integrity rights be affected by greater consensus among policymakers, scholars, and the general public that: (a) multilateral approval is always required for humanitarian intervention or, conversely; (b) states may bypass multilateral bodies whenever they identify a situation of ‘humanitarian necessity’? Lest our answers be entirely speculative, they should rely, fundamentally, on what we know about state behaviour and foreign policy decision-making; in other words, they should be informed by the best empirical research available.

I. Reduced risk of conflict spirals among powerful states

All-out war among major powers, such as the Thirty Years’ War, the Napoleonic Wars, and the two World Wars, has dramatic and probably unparalleled implications in terms of physical integrity rights violations. Today, if a military dispute between nuclear great powers such as the United States and China, or China and Russia, spiralled out of control, it might end human civilisation. All-out war between regional rivals, such as Iran and Saudi Arabia, or Malaysia and Indonesia, could also cause unfathomable suffering and destruction.

Therefore, to the extent that multilateral rules and procedures regulating the use of force help manage international rivalries and reduce the risk of accidental war among powerful states, they can be said to have a beneficial impact on physical integrity rights. Similar arguments have been made by pluralist theorists of international society, who emphasise the importance of a strong international rule of non-intervention and of multilateral authorisation for the use of force as foundations for international order. 31 They point out that over time, a procedural consensus has emerged among states about the rules that ought to govern their external behaviour, including in matters of military intervention; these rules create stable expectations among states and should not be forsaken lightly. ‘Unilateral intervention’, as Hedley Bull wrote, whether by individual states or ad hoc coalitions, ‘threatens the harmony and concord of the society of states’. 32

From a rational-choice perspective, as well, multilateral rules and decision-making procedures can function as ‘focal points’ that reduce coordination problems among powerful states. The formal institutions that embody these rules also offer a platform for states to communicate their concerns, thus increasing transparency and reducing the risk of accidental escalation. 33 Allowing states to intervene without multilateral approval, when such approval is not readily forthcoming and they determine on their own that a foreign crisis has reached the threshold of ‘humanitarian necessity’, could have destabilising consequences: while state leaders tend to believe that others will share their own benign view of their state’s policy, foreigners (especially foreign adversaries) usually interpret such policies based on their pre-existing (and often hostile) images and assumptions. 34 Consequently, absent multilateral validation, what appears as a humanitarian mission to some may be interpreted as an imperial war of aggression by others, fuelling security dilemmas and spirals of misperception that could heighten regional tensions and, eventually, result in major-power war. 35

Compliance with the rules of positive international law, and thus approval of humanitarian interventions by the UNSC, is most likely to prevent conflict spirals and accidental escalation among powerful states. But approval by regional organisations, too, may signal benign intensions and thus help manage international tensions, especially in the following circumstances: first, when the regional organisation includes among its members dyads of potentially antagonistic states that might otherwise see their relationship deteriorate due to one state’s participation in the intervention, such as for Greece and Turkey in the case of NATO’s interventions in the Balkans (Greece had strong reservations about the US-led Bosnia and Kosovo interventions, while Turkey contributed militarily); and second, when the organisation approves a military intervention by outside powers against one of its own members, thus providing a potentially powerful counter to accusations of neo-imperialism, as in the case of Arab League-endorsed interventions by Western powers in the Arab world.

One important caveat is that to effectively signal benign intentions, intervening states not only have to secure multilateral approval; they also have to abide by the terms of the authorising resolution. If interveners blatantly exceed the available mandate, as they arguably did during the 2011 Libya intervention, this may exacerbate broader international tensions. Security Council Resolution (SCR) 1973 on Libya, adopted on 17 March 2011, provided a narrowly tailored authorisation of the use of force to ‘protect civilians and civilian populated areas under threat of attack’ and to enforce a no-fly zone over Libya. However, Washington and its allies soon shifted to a more ambitious policy of regime change, arguing that this was essential to achieve peace and secure longer-term improvements in human rights. 36 Russian authorities, who had abstained on SCR 1973, felt strongly that as a result, the intervention exceeded the available UN mandate, which led to significant tensions between Moscow and the West. 37

II. Reduced likelihood of self-serving interventions

Scholars and political authorities from the developing world, in particular, worry that humanitarian intervention has the ‘potential of becoming a tool for the interference by the strong in the affairs of the weak, with humanitarian considerations providing a veneer to justify such intervention’. 38 A strong case can be made that existing multilateral rules and procedures make it more difficult for powerful states to use humanitarianism as a pretext when the real purpose is something else. 39

As noted, a genuine humanitarian intervention has as its primary purpose the protection of foreign nationals from manmade violence. This purpose plays a key role in making the intervention legitimate. 40 The intervener’s purpose needs to be distinguished from his or her underlying motives. The purpose is what one aims at, whereas the motive is why one aims at it. Traditional Catholic just war theory was fundamentally concerned with an intervener’s inner motives, as the ultimate goal was the salvation of the soul; however, from a secular perspective, ‘pure’ or altruistic motives are not essential for legitimate humanitarian intervention. What matters is that the motives are compatible with and support the pursuit of a humanitarian objective. 41

Thus, the United States might intervene in Haiti with the aim (purpose) of ending a humanitarian crisis there, and France and the United Kingdom might intervene in Libya with the same aim, but their motives for doing so could be self-interested – for example, stopping refugee flows or a desire for international prestige. Indeed, ‘mixed motives’ involving a degree of self-interest may be desirable, as they make it more likely that an intervener will marshal the resources and have the commitment necessary to bring about a positive humanitarian outcome. 42

When the purpose of an intervention is entirely self-serving, the intervention cannot be called ‘humanitarian’. States might nevertheless use humanitarian rhetoric fraudulently: for example, a state might declare that its purpose is to save an ethnic or religious minority abroad, but its actual goal is to overthrow a troublesome foreign government regardless of human rights considerations, annex a particular territory, or exploit that territory’s natural resources. For instance, Russia claimed in 2008 that its military intervention in Georgia, motivated largely by neo-imperial ambitions, was justified under the Responsibility to Protect doctrine in response to an imminent threat of mass atrocities against South Ossetians. 43 Similarly, in 1983, President Ronald Reagan sought to justify the US invasion of Grenada, whose main purpose was to overthrow that country’s communist government, as a new type of pro-democracy humanitarian intervention. 44

To argue persuasively that contemporary multilateral rules and procedures make it more difficult for powerful states to use humanitarianism as a pretext, one would need to be able to show the following: first, these procedures are reasonably successful at exposing the fraudulent use of humanitarianism; and second, prospective interveners generally seek to comply with multilateral procedures because they expect that non-compliance can be costly.

Existing multilateral institutions and procedures can help bring to light a state’s true purpose – or, at a minimum, expose the state’s justification as unsound – in various ways. Multilateral bodies can dispatch impartial fact-finding missions and observers to the territory in question to establish whether there is evidence of massive human rights violations that might warrant a humanitarian intervention. When a prospective intervener seeks multilateral approval, it may also be required to disclose its own intelligence assessments to other member states, which will then be able to closely scrutinise the methodologies employed and related findings. Should the available evidence indicate that conditions on the ground might warrant a humanitarian intervention in principle, the prospective intervener will then need to persuade other member states that a good faith effort has been made to resolve the crisis peacefully and that it has both the capabilities and the political will to achieve a positive humanitarian outcome. When a state simply refuses to subject its policy to multilateral vetting and authorisation, this inherently raises questions about its purpose.

Political realists have traditionally doubted the ability of multilateral institutions to meaningfully influence and constrain the behaviour of powerful states. Recent research, however, indicates that even powerful states value multilateral approval for their interventions. Democracies – including most Western countries – crave multilateral approval for domestic political reasons, to satisfy groups of their own citizens that mobilise around international norms 45 and to reassure sceptical legislators that the operational burden will be shared internationally. 46 Even non-democracies for the most part desire multilateral approval as a way of signalling benign intentions to other states and averting potentially costly retaliation across issue areas. 47

Most states most of the time cherish the international status quo and are not threatened in their survival; consequently, they are likely to intervene militarily only if they anticipate that the costs to them will be relatively low. Multilateral institutions have the ability to constrain powerful states by threatening to delegitimise, and thus increase the cost, of interventions that are not clearly in self defence and are carried out without those institutions’ approval. Because multilateral bodies are unlikely to authorise self-serving interventions and even powerful states desire multilateral approval, it seems fair to conclude that existing multilateral procedures make self-serving interventions less likely. The suggestion from some quarters that the presumption against unauthorised humanitarian warfare should be relaxed, if widely accepted, probably would make it easier for powerful states to engage in self-serving interventions masked as humanitarianism.

III. Reduced epistemic uncertainty and lower risk of accidental abuse

A related but different problem arises when a prospective intervener genuinely has a humanitarian purpose, aiming to save foreign nationals from manmade violence, but is mistaken in its judgement. For instance, the state wishing to intervene might be mistaken in its assessment of the situation on the ground, its estimate of the costs of intervention, and/or its evaluation of the likelihood of success. Such mistaken judgements can be expected to result in accidental abuse, whereby the human rights situation of the target population is unintentionally made worse. 48 International deliberations aimed at securing multilateral approval, as Jürgen Habermas affirms, can limit the risk of such accidental abuse, for instance, by helping well-meaning interveners distinguish their own partial viewpoint from the ‘universalisable interests that all the other nations could share’. 49 But why should we expect that deliberation within inter-national, as opposed to merely national, fora will result in less biased assessments?

Decision-makers operating within the bureaucracies and public sphere of a particular country are especially prone to motivated biases – that is, they are likely to misjudge a situation because of powerful psychological and political forces. Research in cognitive psychology has identified a tendency for individuals and tightly integrated organisations such as national governments to reach ‘premature cognitive closure’, coming to images, beliefs, and conclusions on the basis of limited and fragmentary information, and then to interpret ambiguous or even discrepant information as confirming these beliefs. 50 National bureaucracies are also prone to ‘groupthink’, which occurs when decision-making in highly cohesive groups produces a psychological drive for consensus. This inhibits the expression of internal dissent and the consideration of alternative viewpoints – reinforcing the propensity to reach decisions prematurely, based on limited evidence and preconceived notions. 51

Political pressures often reinforce these psychological tendencies. 52 Thus, a national leadership subject to intense domestic political pressure to ‘do something’ about a foreign humanitarian crisis may conclude after a quick review of the evidence and brief internal debate that military intervention not only would be politically advantageous but also is warranted on human rights grounds and that a positive outcome can be achieved at little cost. Once a decision has been taken to follow such a course of action, discrepant opinions within the national bureaucracy, emphasising that the distinction between local victims and aggressors may not be as clear-cut as initially thought and that the costs of intervention could be much higher, are likely to be silenced – because politically, it is disadvantageous to express public doubts about one’s favoured policy.

The requirement to seek external validation, or approval, of one’s national perspective through multilateral procedures can break through such motivated biases. First, it helps reduce epistemic uncertainty related to the existence of a just cause for intervention; second, it may result in a more accurate assessment of the likely costs of humanitarian warfare and its chances of success. Multilateral organisations provide an institutionalised platform for sharing intelligence and military assessments: when a crisis situation arises, established routines make it easier for member governments to compare and contrast the information generated by their respective national bureaucracies. In addition, multilateral bodies can gather their own information about alleged human rights violations by dispatching fact-finding missions, deploying international monitors, or setting up commissions of inquiry. 53

In the case of Bosnia, for instance, the presence of UN personnel on the ground was essential in 1995 to ascertain Serb responsibility for mass atrocities in the town of Srebrenica and in the capital, Sarajevo, triggering US-led airstrikes that facilitated an end to the Bosnian war. 54 In the Kosovo case in late 1998 and early 1999, the presence of international monitors was again central to determining that Serb forces had employed indiscriminate violence, this time against ethnic Albanians, thereby strengthening the hand of those who were calling for military intervention. 55

For both Bosnia and Kosovo, if President Bill Clinton had not been committed to securing NATO’s approval for air strikes, the United States might have intervened much sooner, with only an ad hoc coalition of allies and under much greater uncertainty as to whether one of the parties on the ground reasonably could be termed the aggressor. On Bosnia, from the spring of 1993 onward, US ambassador to the United Nations Madeleine Albright, as well as Vice President Al Gore, National Security Adviser Tony Lake, and other senior US officials, relied on fragmentary information to advocate coercive air strikes against the Bosnian Serbs. 56 Albright, in particular, dismissed evidence that the Bosnian Muslims were responsible for their own share of war crimes and insisted from early on that the United States had a moral obligation ‘to resist evil’. 57 In a secret April 1993 memo, Albright argued that given the hesitations of America’s Western European allies, the United States might have to use air power ‘unilateral[ly] … to demonstrate the commitment and will of the U.S. to prevent “ethnic cleansing”’. 58 Likewise on Kosovo, from the spring of 1998 onward, Albright (then secretary of state) and other senior US officials pushed for unilateral air strikes against Belgrade – at a time when many international observers still had doubts as to whether Serb security forces had exceeded the bounds of legitimate counterinsurgency. 59

In both the Bosnia and Kosovo cases, President Clinton long held out against the entreaties of his hawkish advisers. Mindful of keeping the Atlantic alliance together and securing international legitimacy for the use of force, Clinton worked hard to forge a NATO consensus behind coercive air strikes. 60 Washington’s continental European allies were reluctant to go to war against the Serbs; the Europeans were also sceptical about the ability of air power alone to achieve sustainable improvements in human rights; therefore, it seems likely that the need to bring NATO on board raised: (a) the evidentiary threshold for establishing the Serbs’ primary responsibility for violations of international humanitarian law; and (b) ensured that when NATO intervened, its members were committed to contributing the necessary resources for long-term success – including armed peacekeepers on the ground.

There are nevertheless circumstances in which multilateral procedures may fail to reduce epistemic uncertainty and thus the risk of accidental abuse. The 2011 Libya intervention, for instance, was justified (and swiftly authorised by multilateral bodies) on the basis of claims of an impending humanitarian disaster, but it is unclear whether the intervention met the jus ad bellum proportionality requirement (saving more civilians than are killed or injured as a consequence of the intervention).

After a rebel uprising in Libya in early 2011, by mid-March, forces loyal to the country’s president, Muammar Qaddafi, seemed poised to crush the anti-government forces. Qaddafi’s threat to show the rebels holding out in Benghazi, their last stronghold, ‘no mercy or compassion’, was seized upon by advocates of humanitarian intervention as evidence of the regime’s intent to commit mass atrocities – resulting in the adoption of SCR 1973. 61 However, as Alan Kuperman has shown, ‘although the [Libyan] government did respond forcefully to the rebels, it never targeted civilians or resorted to “indiscriminate” force’. 62 In his warning to Benghazi’s residents, Qaddafi promised amnesty for those ‘who throw their weapons away’. 63

US defence officials have since conceded that the decision to intervene in Libya was ‘an intelligence light decision’ based on speculative claims about what might happen to civilians rather than on facts reported from the ground. 64 The main advocates of military intervention in the Barack Obama administration – US ambassador to the United Nations Susan Rice, National Security Council staffer Samantha Power, and Secretary of State Hillary Clinton – may have fallen prey to motivated bias: they made up their minds quickly, relying on fragmentary evidence and extrapolating from previous experiences in Bosnia and Rwanda. 65

Why did the multilateral authorisation process not help break through such bias in the Libya case? The most likely reason is that by early March, there was considerable international support for military intervention; hence the United States did not have to meet a high evidentiary threshold for proving large-scale violations of international humanitarian law. Washington’s principal European allies – notably, France and Britain – had been calling for military action for weeks; Qaddafi was unloved in the Arab world, which encouraged two regional organisations, the Gulf Cooperation Council and the Arab League, to endorse the imposition of a no-fly zone over Libya; and Russia, though sceptical, did not consider its national interests sufficiently threatened to veto the UNSC resolution. 66 This suggests that, perhaps unsurprisingly, multilateral authorisation procedures are less likely to help overcome motivated bias (and thus reduce the risk of accidental abuse) when securing multilateral approval is relatively easy.

IV. Reduced moral hazard

Advocates of humanitarian intervention believe that a credible threat of international military action will often deter states from committing atrocities. If deterrence fails, actual intervention can then compel an abusive government to stop its atrocity crimes and facilitate a political solution. 67 Others, however, contend that a low threshold for intervention ‘creates moral hazard that unintentionally fosters rebellion’ by encouraging armed opposition groups to deliberately provoke state atrocities in the expectation that this will trigger intervention on their behalf. 68 According to a weaker form of this argument, expectations of humanitarian intervention may prolong civil wars, making rebels less prone to accept negotiated solutions because they have incentives to hold out for a better deal in the future. 69 In short, the prospect of military intervention in support of anti-government opposition groups may operate much like an insurance policy, which creates moral hazard by encouraging greater risk taking on the part of its intended beneficiaries. 70

The ‘moral hazard of humanitarian intervention’ argument remains contested. Two studies, in particular, find little evidence that expectations of humanitarian intervention cause substate groups to initiate violence against government forces. 71 However, critics have not been able to disprove the moral hazard argument in its weaker form – namely, that the prospect of humanitarian intervention fosters escalation (rather than initiation) of rebel violence and prolongs violent rebellions by making negotiated settlements more difficult. 72

In the Kosovo case, for example, in October 1998, US envoy Richard Holbrooke persuaded the Serb-dominated government of Yugoslavia to end its crackdown on Kosovar Albanian civilians. Soon thereafter, as revealed by declassified documents, senior US officials determined that ‘there exists very substantial FRY [Federal Republic of Yugoslavia] compliance’ with international demands. 73 But the Kosovo Liberation Army (KLA), the region’s main rebel group, had by then little interest in a negotiated compromise solution, given that earlier in the fall the Americans had explicitly threatened military action against Serb targets. Indeed, there is considerable evidence that the KLA took advantage of Serb restraint to reorganise itself, and by mid-November, it had dramatically stepped up its attacks against Serb authorities in Kosovo. According to Wolfgang Petritsch, the European Union’s envoy to Kosovo at the time, these increased attacks were clearly intended to bring about Serb retaliation, with the goal of triggering a US-led military intervention. 74

Others have found that in the case of Darfur, rebels similarly prolonged the fighting in 2003–4, ‘either to take advantage of the possibility that the international community would intervene in their defence, or to obtain better peace terms under international sponsorship’. 75 Likewise, there is some indication that in the case of Libya in the spring of 2011, the prospect of international intervention emboldened anti-government rebels to keep on fighting despite of their inferior firepower rather than seeking a negotiated settlement. In early March of that year, Qaddafi reportedly accepted international offers of mediation, yet the rebel leadership flatly rejected any possibility of negotiations. 76 Kuperman concludes that the early and significant signals of support for the Libyan rebels from Britain, France, and the United States, and those countries’ growing talk of military intervention, are important factors that ‘help explain why the otherwise feeble rebels continued fighting’ at great cost to the civilian population. 77

In short, one ought to take seriously the possibility that a high ex ante likelihood of military intervention in support of anti-government rebels may embolden the rebels to step up their violent activities and adopt a more intransigent stance, at the cost of increased human rights violations. How can this moral hazard be reduced without leaving civilian populations exposed to the threat of government-sponsored mass atrocities? Kuperman identifies several strategies for mitigating moral hazard: first, randomisation, which consists in ‘add[ing] uncertainty about insurance payouts, thereby lowering the insured’s incentive to take excessive risks’; 78 second, insulation from lobbying by adopting various measures to reduce the insurance provider’s exposure and vulnerability to political pressures aimed at generating a payout; and third, better regulation aimed at increasing the likelihood that payouts will be made only to those who behave responsibly. 79

Current norms placing value on multilateral approval for humanitarian intervention are likely to have already reduced the problem of moral hazard. A strong argument can be made that greater consensus among scholars, pundits and policymakers that multilateral approval is always required for humanitarian intervention would help further mitigate moral hazard through all of the aforementioned mechanisms.

To begin with, the presumption in favour of requiring multilateral approval, reaffirmed by UN member states in the 2005 World Summit outcome document, 80 introduces an element of ambiguity into the rules for insurance protection. It thereby contributes to randomisation, as even the most optimistic among rebels cannot expect a payout in the form of intervention to occur automatically once a certain threshold of violence has been crossed. In addition, and related to this, delegating the authorisation to multilateral bodies insulates intervention from lobbying, as persuading a single powerful government that may be easily swayed by advocacy is no longer sufficient. Finally, the presumption in favour of requiring multilateral approval results in better regulation, reducing the likelihood that payouts will be made to opposition groups that behave irresponsibly. For instance, the fact-finding missions and observers dispatched by multilateral organisations can, in theory, delegitimise a rebel group if they find that it has used anti-government violence as a deliberate instrument of provocation and/or has engaged in large-scale human rights violations of its own.

Current norms that place value on multilateral approval cannot entirely overcome the moral hazard problem. Indeed, when several powerful states are pushing for intervention and the international community is inclined to support them from an early stage (as in the 2011 Libya case), multilateralism may not mitigate moral hazard at all. But the Libya case was probably sui generis. In most cases in which humanitarian intervention is contemplated in support of anti-government rebels, securing multilateral approval is likely to be anything but easy, given that such interventions are prima facie incompatible with the principle of non-interference in states’ domestic affairs enshrined in Article 2 of the UN Charter. Consequently, compliance with multilateral procedures can generally be expected to have a restraining effect on would-be interveners, reducing moral hazard through all the aforementioned mechanisms.

Completely eliminating moral hazard would necessitate much deeper restrictions on humanitarian intervention, such as altogether excluding the possibility of intervention in support of groups that have engaged in anti-government violence. That would make it more difficult for oppressed groups to defend themselves; it would also limit the ability of outside actors to impose costs on indiscriminately oppressive governments, which would be morally problematic. 81 But some selectivity in humanitarian intervention, based on multilateral vetting, should be welcomed to minimise moral hazard.

V. Increased likelihood of successful peacebuilding

Finally, approval of the initial intervention from the UNSC or other appropriate multilateral bodies is increasingly required for successful post-bellum peacebuilding. Humanitarian interventions are typically carried out by states that perceive no imminent threat to their own national security, making it difficult for these states to generate the necessary political support at home for open-ended peacebuilding commitments – unless there is evidence of significant international burden sharing. Such burden sharing is significantly more likely to materialise if the original intervention was approved by appropriate multilateral bodies; indeed, multilateral approval may be essential for longer-term burden sharing (and thus ultimately for successful peacebuilding).

In the United States, congressional opposition to foreign deployments of US troops usually intensifies as commitments become protracted. 82 At the same time, there is strong bipartisan demand in the US Congress for burden sharing with foreign allies and partners, which suggests that evidence of burden sharing will help keep congressional opposition in check. 83 Other democracies – whether in Europe, Asia, or Africa – that might play important roles in humanitarian intervention are similarly constrained by their domestic politics from committing significant troops and resources to open-ended stabilisation missions, making international burden sharing highly desirable. 84

Multilateral approval of the initial intervention facilitates longer-term burden sharing in several ways. First and most explicitly, the resolution of approval can contain a formal commitment on the part of the multilateral body and its principal members to set up a follow-on peacekeeping force with contributions from several countries. For instance, SCR 940, which authorised a US intervention in Haiti in 1994, explicitly mandated the establishment of a follow-on UN force as soon as basic security in the country had been restored. More recently, UN resolutions have become increasingly specific in this regard: SCR 1497, which authorised a US-led humanitarian intervention in Liberia in 2003, committed the Security Council ‘to establish … a follow-on United Nations stabilisation force’ led by a coalition of countries from the region within a maximum of two months (para. 2).

Even in the absence of such explicit pledges of longer-term burden sharing, multilateral approval involves a public, and therefore potentially costly, commitment to support the intervener’s policy on the part of all the organisation’s members that have offered their affirmative vote. Once member states are thus committed, subsequent resistance by them to establishing a multilateral stabilisation mission strongly wanted by the lead intervener becomes less likely. Member states may also independently value organisations such as the UN, NATO, and the AU for the security benefits they provide; consequently, once a multilateral stabilisation mission has been approved and the organisation’s reputation becomes linked to mission success, members may be willing to maintain significant troop contributions even in the face of mounting costs. 85

Finally, and crucially, advance multilateral approval facilitates sustained burden sharing by legitimising the intervention and thus reducing domestic political obstacles for foreign leaders to cooperate with the intervener. 86 What matters most in this context is not whether institutions such as the UN, the AU, and NATO are themselves legitimate according to abstract democratic standards but rather whether international audiences are likely to perceive interventions authorised by such bodies as legitimate in accordance with socially constructed norms. 87 Interventions carried out without multilateral approval can be expected to raise suspicions among foreign publics about the intervener’s intentions, making it difficult even for the leaders of sympathetic states to contribute troops and resources, let alone maintain their contributions in the face of mounting costs.

In 2003, for instance, after the United States and Britain invaded Iraq without UN approval, they struggled to persuade other countries to contribute stabilisation troops. Once the United States agreed to offset the financial cost of these contributions, in the short run, it was able to recruit about 16,000 troops from other countries besides Britain. 88 However, the improvised coalition showed little staying power. By May 2007, when America ‘surged’ its own troops to more than 150,000 to control the Iraqi Civil War, the non-US/UK component had shrunk to only about 7,000 troops. 89 This pointed lack of burden sharing resulted in growing domestic political pressure on Washington and London to withdraw all their troops, well before they had achieved sustainable peace. 90

By contrast, in the Balkans, the United States intervened militarily in 1995 and again in 1999 only after securing multilateral approval from the United Nations and/or NATO. That greatly facilitated sustained international burden sharing, which, in turn, helped forestall serious US congressional opposition to America’s peacebuilding commitments in the Balkans. 91 The United States similarly secured UN approval for the 1994 Haiti intervention, which then made it possible in a matter of months to establish a follow-on UN force with contributions from several countries. Former deputy US Secretary of State Strobe Talbott recalls that members of Congress who ‘were eager either to cut or restrict appropriations for Haiti’ were mollified by this evidence of burden sharing. 92

Advance multilateral approval may, of course, not be sufficient in itself to generate sustained international contributions to stabilisation and peacebuilding, as indicated by the recent Libya experience. For third-party countries to contribute, the main intervening states usually must be willing to deploy sizeable numbers of their own troops in a stabilisation role, at least for several months. Absent such political will on the part of the lead interveners, no amount of multilateral legitimation may be able to persuade other countries to take on the task on their own. In Libya in 2011, once Qaddafi’s regime fell, the principal interveners – Britain, France, and the United States – had no appetite for deploying their own troops in a stabilisation role; consequently, no international stabilisation force was established, and the result has been an ‘enduring state of lawlessness in Libya’ that may well have made the overall human rights situation worse relative to the status quo ante. 93

Discussing the problematic outcome of the Libya intervention, President Obama acknowledged that ‘we [and] our European partners underestimated the need to come in full force if you’re going to do this’. Specifically, Obama conceded, after the cessation of major hostilities, ‘at that moment, there has to be a much more aggressive effort to rebuild societies that didn’t have any civic traditions’. 94 Securing multilateral approval for the initial intervention may not be sufficient for the success of such longer-term efforts – but it is increasingly likely to be essential.

Unilateralism in circumstances of ‘humanitarian necessity’?

Before concluding, it may be worth discussing more explicitly whether the rule that requires states to obtain multilateral approval for humanitarian intervention should be regarded as peremptory (demanding absolute compliance), or instead as something flexible. Those who believe that the rule should be flexible generally ask some variant of the following question: would it be better not to intervene in a genocidal, Rwanda-like situation ‘than to intervene without meeting the conditions … [of] the authorisation rule?’ 95

Precedents from the 1970s: Vietnam in Cambodia and Tanzania in Uganda

Supporters of a flexible interpretation of the multilateralism rule sometimes point to Vietnam’s intervention in Cambodia in 1978 and Tanzania’s intervention in Uganda one year later. Although blatantly unilateral, it is argued, these interventions achieved positive humanitarian outcomes by stopping large-scale atrocities and toppling two barbaric regimes – Pol Pot’s Khmer Rouge regime in Cambodia and Idi Amin’s military dictatorship in Uganda. 96

These examples, however, are problematic. First, it is debatable whether the Cambodia and Uganda interventions should count as ‘humanitarian’, given that the purpose of the interveners was not to alleviate human suffering but rather to tackle more traditional threats to national security. According to most accounts, humanitarianism was not even a second-order consideration. The Vietnamese government, as Nick Wheeler notes, simply concluded after repeated cross-border incursions by Cambodian forces onto its territory between 1977 and 1978 ‘that its long-term security required the removal of Pol Pot’. 97 Similarly, the Tanzanian government decided after Ugandan forces invaded a Tanzanian border region in October 1978 that its national interest required it ‘to remove any future military threat to Tanzania by toppling Amin from power’. 98

The positive humanitarian outcome of these interventions – narrowly conceived as saving civilians who were imminently threatened by violence – was thus largely accidental, a fortuitous by-product of self-serving military operations. Even today, self-serving interventions (say, a Saudi intervention in Syria aimed at toppling the regime of Bashar al-Assad and installing in its stead a Sunni regime) could more or less by chance achieve a positive humanitarian outcome in the short term. But it does not follow that one should condone or even welcome such unauthorised interventions as a matter of principle, given their potential downsides as discussed in previous sections of this article.

It should also be noted that the Cambodia and Uganda interventions stopped short of putting the target country on a path toward durable peace and thus would be viewed as failures if measured against modern principles of jus post bellum. In Cambodia, following Pol Pot’s removal, a fragile cease-fire among the country’s principal factions was not reached until 1991, and in 1992 the UNSC concluded that a renewed international intervention was required to stabilise the country. 99 In Uganda, after the fall of Idi Amin, a brutal five-year civil war tore the country apart, claiming an estimated 500,000 lives, most of them civilian (thus exceeding the death toll from Idi Amin’s previous rule, estimated at 300,000). 100

Understanding non-intervention in Rwanda and Syria

Furthermore, today (as opposed to the 1970s), in the presence of a humanitarian emergency widely recognised as such, states that are determined to intervene to alleviate the suffering and can be trusted to comply with international humanitarian law have a good chance of securing multilateral approval. One should not overstate the ease of obtaining such approval, as related negotiations can often be tortuous and time-consuming. Nevertheless, blaming veto players at the UNSC or other multilateral bodies for the absence of intervention may be unfair.

In the case of Rwanda in 1994, which is often mentioned as indicative of the inadequacy of existing multilateral procedures, what prevented timely intervention was primarily a lack of political will on the part of the United States and its most militarily capable allies. Wheeler persuasively argues that ‘had any state or group of states come forward in April or May 1994 requesting a UN mandate to end the atrocities …, this would have been readily agreed to’. 101

More recently, confronted with complex humanitarian emergencies in Darfur and Syria, the United States and its Western allies have similarly been unwilling to intervene decisively to relieve the suffering. 102 In the Syria case, in particular, President Obama never had the political will for a major humanitarian intervention that would probably have resulted in the collapse of the al-Assad regime, because he: (a) does not believe that the humanitarian crisis there poses a direct security threat to the United States; and (b) views Syria as ‘a slope potentially as slippery as Iraq’. 103 The ostensible opposition of Russia and China at the UNSC may thus have offered Western countries a convenient smokescreen to hide their own unwillingness to act.

No determined effort has been made to secure multilateral approval for a humanitarian intervention in Syria – for instance, by seeking a ‘bargain’ with Russia, whereby the West would offer Russia economic inducements and/or full reintegration into the G8 group of major powers, in exchange for Moscow’s support at the Security Council. To put things in perspective: during the run-up to the 1991 Persian Gulf War, to obtain Moscow’s and Beijing’s cooperation at the UNSC, Washington offered hundreds of millions of dollars in economic inducements to Russia and agreed to lift economic sanctions imposed against China after the Tiananmen massacre. 104

Even superpowers cannot secure multilateral approval under all circumstances, as demonstrated by Washington’s failure to secure a UN mandate for the 2003 Iraq War. 105 However, if there is a widely recognised just cause for intervention, chances are good that prospective interveners will be able to obtain such approval – provided they make a major effort. Scholars who claim that current multilateral procedures bear heavy responsibility for morally problematic ‘humanitarian under-intervention’ 106 may thus misidentify the problem. When no intervention occurs, in spite of a genuine humanitarian emergency, it is often because no major power is making a determined push for intervention (and for multilateral approval) to begin with.

Conclusion: Ideal theory and a non-ideal world

This article has argued that existing multilateral procedures for authorising armed intervention should be evaluated first and foremost against their ability to limit the negative consequences of intervention and promote physical integrity rights. By this consequentialist standard, I have shown, they have performed reasonably well, and we can expect them to continue to do so.

From the vantage point of ideal theory, it may not be possible to rule out the permissibility of unauthorised intervention: if an unbiased, well-informed, militarily capable, and strongly motivated state were willing to intervene to stop large-scale human suffering in a situation clearly involving genocide, war crimes, or crimes against humanity and, ex hypothesi, no appropriate multilateral body were willing to authorise the intervention, then we should most likely welcome a unilateral intervention by that state. But under the non-ideal circumstances of the world as it currently exists, the odds of such a combination of factors occurring are rather low.

In practice, for reasons explained in this article, there is a high probability that unauthorised interventions will fail to promote human rights. If powerful states were able to intervene without multilateral approval, whenever they concluded or simply claimed that a situation of humanitarian necessity exists, we could expect: (a) increased international tensions between the intervener(s) and other powerful states; (b) more deliberate and accidental abuse, as countries sought to manipulate humanitarian rhetoric for their own ends or might be mistaken in their assessments; (c) increased moral hazard, with violent opposition groups engaging in high-risk behaviour to persuade powerful states to intervene on their behalf; and (d) a significantly lower likelihood of successful peacebuilding.

To conclude, then, how should countries concerned about large-scale human rights violations in foreign lands behave, when multilateral approval for military intervention is genuinely unavailable – for example, because other member states strongly disagree that there is a just cause for intervention, or worry that intervention would threaten their core national interests and/or make the situation on the ground worse? In such circumstances, concerned countries may provide vital medical assistance and food aid, offer hospitality to refugees, and attempt to mediate a political solution to the crisis, but they should refrain from intervening militarily. The outcome-oriented approach embraced in this article suggests that in our current non-ideal world, humanitarian warfare unauthorised by mulitlateral bodies should indeed be viewed as impermissible.


The author would like to thank James Pattison, Jennifer Welsh, Lea Ypi, and the three anonymous reviewers for helpful comments on previous versions of this article. Open-access publication was made possible by a Marie Curie Career Integration Grant from the European Commission.

1 Tom J. Farer , ‘A paradigm of legitimate intervention’, in Lori Fisler-Damrosch (ed.), Enforcing Restraint: Collective Intervention in Internal Conflicts (New York: Council on Foreign Relations, 1993), p. 327 ; Nicholas Wheeler , Saving Strangers: Humanitarian Intervention in International Society (Oxford: Oxford University Press, 2000), pp. 298299 ; Thomas Franck , ‘Legality and legitimacy in humanitarian intervention’, in Terry Nardin and Melissa Williams (eds), Humanitarian Intervention (New York: New York University Press, 2006); Fernando R. Tesón , ‘The vexing problem of authority in humanitarian intervention: a proposal’, Wisconsin International Law Journal, 24:3 (2006), pp. 761772 ; James Pattison , Humanitarian Intervention and the Responsibility to Protect: Who Should Intervene? (Oxford: Oxford University Press, 2010), pp. 188189 , 196; Ian Hurd , ‘Bomb Syria, even if it is illegal’, New York Times (27 August 2013).

2 Pattison, Humanitarian Intervention, p. 74.

3 Inis Claude , ‘Collective legitimization as a political function of the United Nations’, International Organization, 20:3 (1966), p. 372 , emphasis added.

4 Michael Walzer , ‘The politics of rescue’, Social Research, 62:1 (1995), pp. 5366 . Nevertheless, Walzer offers at best qualified support for multilateralism, which, he believes, ‘is no guarantee of anything’ (p. 63).

5 Thomas Franck , The Power of Legitimacy among Nations (Oxford: Oxford University Press, 1990); Jennifer Welsh , ‘Authorizing humanitarian intervention’, in Richard M. Price and Mark W. Zacher (eds), The United Nations and Global Security (London: Palgrave Macmillan, 2004), pp. 183186 .

6 Ruth Wedgwood , ‘Unilateral action in a multilateral world’, in Stewart Patrick and Shepard Forman (eds), Multilateralism and U.S. Foreign Policy (Boulder, CO: Lynne Rienner, 2002), p. 168 .

7 Ian Clark , Legitimacy in International Society (Oxford: Oxford University Press, 2005), p. 193 .

8 Robert O. Keohane , ‘The contingent legitimacy of multilateralism’, in Edward Newman, Ramesh Thakur, and John Tirman (eds), Multilateralism under Challenge (Tokyo: United Nations University Press, 2006), pp. 6061 .

9 Pattison, Humanitarian Intervention, p. 5. See also Keohane, ‘The contingent legitimacy of multilateralism’; Tesón, ‘The vexing problem of authority’.

10 See fn. 1.

11 Allen Buchanan , Justice, Legitimacy, and Self-Determination: Moral Foundations for International Law (New York: Oxford University Press, 2003), pp. 450468 .

12 Tesón, ‘The vexing problem of authority’, pp. 771–2.

13 Daniele Archibugi , ‘Cosmopolitan guidelines for humanitarian intervention’, Alternatives, 29:1 (2004), pp. 121 ; Pattison, Humanitarian Intervention, pp. 233–4.

14 Franck, ‘Legality and legitimacy’; Catherine Lu , ‘Whose principles? Whose institutions? Legitimacy challenges for “humanitarian intervention”’, in Nardin and Williams (eds), Humanitarian Intervention ; Pattison, Humanitarian Intervention, pp. 188–9.

15 Martha Finnemore , The Purpose of Intervention: Changing Beliefs about the Use of Force (Ithaca, NY: Cornell University Press, 2003), p. 53 .

16 Pattison, Humanitarian Intervention, pp. 101–10.

17 Mary B. Anderson , Do No Harm: How Aid Can Support Peace – or War (Boulder, CO: Lynne Rienner, 1999), esp. ch. 4. For a useful discussion of the tensions inherent in traditional humanitarian intervention, see also Alan J. Kuperman , ‘Humanitarian intervention’, in Michael Goodhart (ed.), Human Rights: Politics and Practice (2nd edn, Oxford: Oxford University Press, 2013), pp. 291293 .

18 Brian Orend , ‘Justice after war’, Ethics & International Affairs, 16:1 (2002), pp. 4356 ; Robert E. Williams , Jr, and Dan Caldwell , ‘Jus post bellum: Just war theory and the principles of just peace’, International Studies Perspectives, 7:4 (2006), pp. 309320 .

19 See International Commission on Intervention and State Sovereignty, The Responsibility to Protect (Ottawa: International Development Research Centre, 2001), p. 39 .

20 Michael W. Doyle and Nicholas Sambanis , Making War and Building Peace: United Nations Peace Operations (Princeton, NJ: Princeton University Press, 2006), pp. 2030 ; Madhav Joshi and T. David Mason , ‘Civil war settlements, size of governing coalition, and durability of peace in post-civil war states’, International Interactions, 37:4 (2011), pp. 388413 . See also Stefano Recchia , ‘Just and unjust postwar reconstruction: How much external interference can be justified?’, Ethics & International Affairs, 23:2 (2009), pp. 165187 .

21 Michael Walzer , ‘The aftermath of war: Reflections on jus post bellum ’, in Eric Patterson (ed.), Ethics Beyond War’s End (Washington, DC: Georgetown University Press, 2012), p. 38 .

22 As Michael Blake puts it, ‘we should not think ourselves licensed to intervene, unless we have both the means and the will to rebuild’. See Blake , ‘The costs of war: justice, liability, and the Pottery Barn rule’, in Don E. Scheid (ed.), The Ethics of Armed Humanitarian Intervention (Cambridge: Cambridge University Press, 2014), p. 134 . See also Pattison, Humanitarian Intervention, p. 74.

23 The Security Council may give its retrospective approval to the exercise of self defence, or it may refrain from doing so and may insist on the cessation of the unilateral action. See Yoram Dinstein , War, Aggression, and Self-Defence (5th edn, Cambridge: Cambridge University Press, 2012), pp. 234237 .

24 John Ruggie , ‘Multilateralism: the anatomy of an institution’, International Organization, 46:3 (1992), p. 571 .

25 Sarah Kreps , ‘Multilateral military interventions: Theory and practice’, Political Science Quarterly, 123:4 (2008), pp. 573603 .

26 Jerome Slater , ‘The limits of legitimization in international organizations: the organization of American states and the Dominican crisis’, International Organization, 23:1 (1969), pp. 4872 .

27 The OAS condemned the Panama intervention in a 20:1 vote. See John M. Goshko and Michael Isikoff , ‘OAS votes to censure U.S. for intervention’, Washington Post (23 December 1989).

28 See, for example, Thomas Risse-Kappen , Cooperation Among Democracies: The European Influence on U.S. Foreign Policy (Princeton, NJ: Princeton University Press, 1997).

29 On the Balkans, see Steven L. Burg , ‘Coercive diplomacy in the Balkans: the U.S. use of force in Bosnia and Kosovo’, in Robert J. Art and Patrick M. Cronin (eds), The United States and Coercive Diplomacy (Washington, DC: United States Institute of Peace, 2003). On the Iraq War, see Philip H. Gordon and Jeremy Shapiro , Allies at War: America, Europe, and the Crisis over Iraq (New York: McGraw-Hill, 2004), pp. 136141 .

30 Grossman, author interview, Washington, DC, 13 January 2011.

31 See, for example, R. J. Vincent , Nonintervention and International Order (Princeton, NJ: Princeton University Press, 1974); and Hedley Bull (ed.), Intervention in World Politics (Oxford: Clarendon Press, 1984).

32 Bull, Intervention in World Politics, p. 195. See also Robert Jackson , The Global Covenant: Human Conduct in a World of States (Oxford: Oxford University Press, 2000), pp. 278293 .

33 Robert O. Keohane and Lisa L. Martin , ‘The promise of institutionalist theory’, International Security, 20:1 (1995), pp. 3951 .

34 Robert Jervis , Perception and Misperception in International Politics (Princeton, NJ: Princeton University Press, 1976), pp. 6277 .

35 On the importance of multilateral approval for signalling benign intentions, see Alexander Thompson , ‘Coercion through IOs: the Security Council and the logic of information transmission’, International Organization, 60:1 (2006), pp. 134 .

36 David Cameron , Barack Obama , and Nicolas Sarkozy , ‘The bombing continues until Gaddafi goes’, The Times (London) (15 April 2011).

37 Catrina Stewart , ‘Russia accuses NATO of “expanding” UN Libya resolution’, The Independent (4 July 2011). For a useful discussion, see also Dag Henriksen and Ann Katrin Larssen (eds), Political Rationale and International Consequences of the War in Libya (Oxford: Oxford University Press, 2016), ch. 4.

38 Mohammed Ayoob , ‘Humanitarian intervention and state sovereignty’, International Journal of Human Rights, 6:1 (2002), p. 92 .

39 For similar arguments, which, however, do not explore the causal mechanism in detail, see Farer, ‘Legitimate intervention’, pp. 324–6; and Michael W. Doyle , ‘The ethics of multilateral intervention’, Theoria, 109 (April 2006), pp. 4142 .

40 Terry Nardin explains that ‘the “cause” that makes a given action “just” is its end or purpose – defending the innocent from violence, for example’. Cf. Nardin, ‘Introduction’, in Nardin and Williams (eds), Humanitarian Intervention, p. 10.

41 Ibid., pp. 9–11.

42 See, for example, Walzer, Politics of Rescue, pp. 59–60; Mark S. Stein , ‘Unauthorized humanitarian intervention’, Social Philosophy and Policy, 21:1 (2004), p. 31 ; and Pattison, Humanitarian Intervention, pp. 159–60.

43 Roy Allison , Russia, the West, and Military Intervention (Oxford: Oxford University Press, 2013), pp. 155158 .

44 Sean Murphy , Humanitarian Intervention: the United Nations in an Evolving World Order (Philadelphia: University of Pennsylvania Press, 1996), pp. 108111 .

45 Andrew P. Cortell and James W. Davis , ‘How do international institutions matter? The domestic impact of international norms and rules’, International Studies Quarterly, 40:4 (1996), pp. 451487 ; Joachim Krause , ‘Multilateralism: Behind European views’, Washington Quarterly, 27:2 (2004), pp. 4359 .

46 Atsushi Tago , ‘Determinants of multilateralism in U.S. use of force’, Journal of Peace Research, 42:5 (2005), pp. 585604 ; Stefano Recchia , ‘Why seek international organisation approval under unipolarity? Averting issue linkage vs. appeasing Congress’, International Relations, 30:1 (2016), pp. 78101 .

47 Thompson, ‘Coercion through IOs’. See also Erik Voeten , ‘The political origins of the UN Security Council’s ability to legitimize the use of force’, International Organization, 59:3 (2005), pp. 527557 .

48 On accidental abuse in the context of humanitarian intervention, see Pattison, Humanitarian Intervention, p. 58.

49 Jürgen Habermas , The Divided West (London: Polity, 2006), p. 184 . See also Ned Dobos , ‘Is U.N. Security Council authorisation for armed humanitarian intervention morally necessary?’, Philosophia, 38:3 (2010), pp. 499515 .

50 Robert Jervis , ‘Bridges, barriers, and gaps: Research and policy’, Political Psychology, 29:4 (2008), pp. 578579 .

51 Irving L. Janis , Groupthink (rev. edn, Boston: Houghton Mifflin, 1982). See also Dina Badie , ‘Groupthink, Iraq, and the War on Terror: Explaining US policy shift toward Iraq’, Foreign Policy Analysis, 6:4 (2010), pp. 277296 .

52 Jervis, ‘Bridges, barriers, and gaps’, pp. 580–1.

53 See Jennifer Welsh and Dominik Zaum , ‘Legitimation and the UN Security Council’, in D. Zaum (ed.), Legitimating International Organizations (Oxford: Oxford University Press, 2013), pp. 7680 .

54 Burg and Shoup, The War in Bosnia–Herzegovina, pp. 307–59. The Sarajevo market bombing of 28 August 1995 remained shrouded in controversy: several authors surmise that the Bosnian Muslims may have engineered the shelling of their own ethnic kin, precisely to trigger an international intervention against Serb forces. See, for example, Nora Beloff , Yugoslavia: An Avoidable War (London: New European Publications, 1997), p. 112 ; and Michael Parenti , To Kill a Nation: the Attack on Yugoslavia (London: Verso, 2002), pp. 7576 . In 2007, however, the International Criminal Tribunal for the Former Yugoslavia (ICTY) concluded after examining all the evidence that Bosnian Serb units were almost certainly responsible for the market shelling, thus validating the initial assessment of UN ballistic experts. See ICTY, ‘PROSECUTOR v. DRAGOMIR MILOSEVIC’, Case No. IT-98–29/1-T, 12 December 2007, p. 241, para. 724, and pp. 220–41 more generally, available at: {} accessed 29 July 2016.

55 Alex J. Bellamy , Kosovo and International Society (London: Palgrave, 2002), pp. 114116 . Some authors have questioned whether the killing of a few dozen ethnic Albanians in the village of Racak in early 1999, which provided a major impetus for intervention, should count as a ‘civilian massacre’, as it seems likely that several of the individuals killed on that occasion were rebel fighters. See, for example, Diana Johnstone , Fool’s Crusade: Yugoslavia, NATO and Western Delusions (London: Pluto, 2002), pp. 241243 ; Mark A. Wolfgram , ‘Democracy and propaganda: NATO’s War in Kosovo’, European Journal of Communication, 23:2 (2008), pp. 153171 . However, the weight of the evidence subsequently assembled by independent international authorities suggests that at Racak, as well as more generally during the run-up to the Kosovo intervention, Serb security forces failed to adequately discriminate between rebel forces and innocent civilians, which, while often difficult, remains a key marker of legitimate counterinsurgency campaigns. ICTY, ‘PROSECUTOR v. MILAN MILUTINOVIĆ et al.’, Case No. IT-05–87-T, 26 February 2009, vol. 1, pp. 324–48, available at: {} accessed 29 July 2016.

56 David Halberstam , War in a Time of Peace: Bush, Clinton, and the Generals (New York: Touchstone, 2001), pp. 196197 ; Stefano Recchia , Reassuring the Reluctant Warriors: U.S. Civil-Military Relations and Multilateral Intervention (Ithaca, NY: Cornell University Press, 2015), pp. 116118 .

57 Madeleine Albright , Madam Secretary (New York: Miramax, 2003), p. 177 . As early as April 1993, the NGO Helsinki Watch (precursor to Human Rights Watch) reported that ‘Bosnian Muslim and Croatian troops have forced the displacement of Serbs in southwestern and central Bosnia’ and ‘Muslim forces have summarily executed civilians and disarmed combatants in Eastern Bosnia’. See Helsinki Watch , War Crimes in Bosnia-Hercegovina, Volume II (New York, 1993), pp. 12 , 14.

58 Madeleine Albright , ‘Options for Bosnia’, memorandum for the National Security Adviser (14 April 1993), available at: {} accessed 29 July 2016.

59 Recchia, Reassuring the Reluctant Warriors, pp. 159–61. See also Ivo Daalder and Michael O’Hanlon , Winning Ugly: NATO’s War to Save Kosovo (Washington, DC: Brookings Institution Press, 2000), p. 30 .

60 Clinton writes that he ‘didn’t want to divide the NATO alliance by unilaterally bombing Serb military positions’. See Clinton , My Life (New York: Knopf, 2004), p. 513 . See also Recchia, Reluctant Warriors, pp. 120–38 (on Bosnia) and pp. 162–77 (on Kosovo).

61 David Kirkpatrick and Kareem Fahim , ‘Qaddafi warns of assault on Benghazi as U.N. vote nears’, New York Times (17 March 2011).

62 Alan J. Kuperman , ‘A model humanitarian intervention? Reassessing NATO’s Libya campaign’, International Security, 38:1 (2011), p. 110 .

63 Kirkpatrick and Fahim, ‘Qaddafi warns of assault on Benghazi’.

64 Kelly Riddell and Jeffrey Scott Shapiro , ‘Hillary Clinton’s “WMD” moment: U.S. intelligence saw false narrative in Libya’, Washington Times (29 January 2015).

65 David Fitzgerald and David Ryan , Obama, U.S. Foreign Policy and the Dilemmas of Intervention (London: Palgrave Macmillan, 2014), pp. 99104 .

66 Henriksen and Larssen, Rationale and Consequences of War in Libya, chs 2–6.

67 See, for example, Matthew C. Waxman , Intervention to Stop Genocide and Mass Atrocities: International Norms and U.S. Policy (New York: Council on Foreign Relations, 2009).

68 Alan J. Kuperman , ‘The moral hazard of humanitarian intervention: Lessons from the Balkans’, International Studies Quarterly, 52:1 (2008), p. 51 . See also Dane Rowlands and David Carment , ‘Moral hazard and conflict intervention’, in Murray Wolfson (ed.), The Political Economy of War and Peace (The Hague: Kluwer, 1998); and Timothy W. Crawford , ‘Moral hazard, intervention and internal war: a conceptual analysis’, Ethnopolitics, 4:2 (2005), pp. 175193 .

69 Roberto Belloni , ‘The tragedy of Darfur and the limits of the “Responsibility to Protect”,’ Ethnopolitics, 5:4 (2006), pp. 327346 ; Roberto Belloni , ‘The trouble with humanitarianism’, Review of International Studies, 33:2 (2007), pp. 459461 .

70 See also Andrew H. Kydd and Scott Straus , ‘The road to hell? Third-party intervention to prevent atrocities’, American Journal of Political Science, 57:3 (2013), pp. 673684 .

71 Jon Western , ‘Illusions of moral hazard: a conceptual and empirical critique’, Ethnopolitics, 4:2 (2005), pp. 225236 ; Alex J. Bellamy and Paul D. Williams , ‘On the limits of moral hazard: the “Responsibility to Protect”, armed conflict and mass atrocities’, European Journal of International Relations, 18:3 (2011), pp. 539571 .

72 Bellamy and Williams (‘On the limits of moral hazard’, pp. 549–50) find that rebellions have generally been shorter since the rise of humanitarian intervention norms after 1990, which they claim disproves the argument that a higher likelihood of intervention prolongs rebel violence. Observed changes in the duration of rebellions, however, may be due simply to the end of Cold War proxy wars and related superpower funding. The authors’ inference would be warranted only if their findings held up after limiting the sample to post-Cold War cases in which humanitarian intervention was in fact seriously considered.

73 US National Security Council, ‘Deputies Committee Meeting on Kosovo’, Summary of Conclusions (26 October 1998), available at: {} accessed 29 July 2016.

74 Wolfgang Petritsch and Robert Pichler , Kosovo-Kosova: Der Lange Weg zum Frieden (Klagenfurt: Wieser, 2004), p. 148 . See also David Gibbs , First Do No Harm: Humanitarian Intervention and the Destruction of Yugoslavia (Nashville, TN: Vanderbilt University Press, 2009), ch. 7.

75 Belloni, ‘Tragedy of Darfur’, p. 226.

76 Kuperman, ‘Model humanitarian intervention?’, pp. 110–13.

77 Ibid., p. 124. See also Daniel Larison , ‘Libyan ceasefire and the moral hazard of intervention’, The American Conservative (18 March 2011).

78 Alan J. Kuperman , ‘Mitigating the moral hazard of humanitarian intervention: Lessons from economics’, Global Governance, 14:2 (2008), p. 228 .

79 Ibid., pp. 229–31.

80 See UN General Assembly, ‘World Summit Outcome’, Sixtieth Session, A/RES/60/1 (24 October 2005), § 139.

81 For a similar argument, see Michael W. Doyle , The Question of Intervention (New Haven, CT: Yale University Press, 2015), p. 23 .

82 David P. Auerswald and Peter F. Cowhey , ‘Ballotbox diplomacy: the war powers resolution and the use of force’, International Studies Quarterly, 41:3 (1997), pp. 505528 .

83 Joshua Busby , Jonathan Monten , Jordan Tama , and William Inboden , ‘Congress is already post-partisan’, Foreign Affairs (28 January 2013), available at: {} accessed 27 July 2016; Recchia, ‘Why seek international organisation approval?’

84 See, for example, UK House of Commons, ‘Working with International Organisations’, Select Committee on Defence, Seventh Report, Part 3 (18 March 2010), available at: {} accessed 30 July 2016; and David S. Sorenson and Pia Christina Wood (eds), The Politics of Peacekeeping in the Post-cold War Era (London: Frank Cass, 2005).

85 Sarah Kreps , ‘Elite consensus as a determinant of alliance cohesion: Why public opinion hardly matters for NATO-led operations in Afghanistan’, Foreign Policy Analysis, 6:3 (2010), pp. 191215 ; Recchia, ‘Why seek international organisation approval?’, p. 83.

86 Wedgwood, ‘Unilateral action in a multilateral world’, p. 173; Thompson, ‘Coercion through IOs’.

87 Claude, ‘Collective legitimization’; Finnemore, The Purpose of Intervention, pp. 80–2.

88 Paul Richter , ‘U.S. enlists more countries in Iraq, at taxpayers’ expense’, Los Angeles Times (22 June 2007); Christopher Blanchard and Catherine Marie Dale , ‘Iraq: Foreign contributions to stabilization and reconstruction’, CRS Report for Congress (Washington, DC: 26 December 2007).

89 Joseph Christoff , ‘Stabilizing and rebuilding Iraq: Coalition support and international donor commitments’, Testimony before the House Subcommittee on International Organizations, Human Rights, and Oversight (Washington, DC: 9 May 2007); Blanchard and Dale, ‘Iraq: Foreign contributions’, pp. 11–18.

90 Jeff Zeleny and Carl Hulse , ‘Senate supports a pullout date in Iraq War bill’, New York Times (28 March 2007); Richard Norton-Taylor , ‘Out by June: UK plans Iraq withdrawal’, The Guardian (10 December 2008).

91 Steven Woehrel , ‘Future of the Balkans and U.S. policy concerns’, CRS Report for Congress (Washington, DC: 13 May 2009); Stephen J. Cimbala and Peter K. Forster , Multinational Military Intervention: NATO Policy, Strategy and Burden Sharing (Burlington, VT: Ashgate, 2010).

92 Strobe Talbott , The Great Experiment (New York: Simon & Schuster, 2008), p. 302 .

93 Roland Paris , ‘The “Responsibility to Protect” and the structural problems of preventive humanitarian intervention’, International Peacekeeping, 21:5 (2014), p. 585 . See also Kuperman, ‘Model humanitarian intervention?’, pp. 125–8.

94 Barack Obama, interview with Thomas L. Friedman, New York Times (8 August 2014).

95 Stein, ‘Unauthorized humanitarian intervention’, p. 33. See also Franck, ‘Legality and legitimacy’, pp. 143–4; Tesón, ‘The vexing problem of authority’, p. 766.

96 Wheeler, Saving Strangers, pp. 100–5 and 13–36. Finnemore (Purpose of Intervention, p. 73) argues that ‘strong humanitarian claims were certainly credible’ in these two cases. On the Uganda intervention, see also Fernando R. Tesón , Humanitarian Intervention: an Inquiry into Law and Morality (2nd edn, Irvington-on-Hudson, NY: Transnational Publishers, 1997), pp. 179195 .

97 Wheeler, Saving Strangers, p. 81.

98 Ibid., p. 114.

99 David Chandler , ‘Foreign interventions in Cambodia, 1806–2003’, in William J. Lahneman (ed.), Military Intervention: Cases in Context for the Twenty-First Century (Lanham, MD: Rowman & Littlefield, 2004).

100 On casualty figures, see, respectively, Global Security, ‘The War in the Bush’, available at: {} accessed 27 July 2016; and Amnesty International, ‘Human Rights in Uganda’ (June 1978), p. 13, available at: {} accessed 28 July 2016.

101 Wheeler, Saving Strangers, p. 299.

102 On Darfur, see Rebecca Hamilton , Fighting for Darfur: Public Action and the Struggle to Stop Genocide (New York: Palgrave Macmillan, 2011), ch. 6. On Syria, see Fitzgerald and Ryan, Obama and the Dilemmas of Intervention, ch. 6.

103 Jeffrey Goldberg , ‘The Obama doctrine’, The Atlantic, 317:3 (April 2016), p. 73 .

104 On incentives to Russia, see James A. Baker , The Politics of Diplomacy: Revolution, War & Peace, 1989–1992 (New York: Putnam’s Sons, 1995), pp. 294295 and 287–313 more generally. On China, see Noële Crossley , Multilateralism versus Unilateralism: the Relevance of the United Nations in a Unipolar World (New York: Peter Lang, 2008), p. 73 .

105 For a discussion, see Stefano Recchia , ‘Did Chirac say “non”? Revisiting UN diplomacy on Iraq, 2002–03’, Political Science Quarterly, 130:4 (2015), pp. 625654 .

106 See, for example, Tesón, ‘The vexing problem of authority’, p. 766.