Introduction
In the early 2000s, terror suspects included in the blacklist of the United Nations (UN) Security Council sanctions regime against Al-Qaeda and the Taliban had no right to know if they were blacklisted, let alone what allegations had been made against them. Nor did they have the right to challenge their listing before a political or administrative body, let alone a court. Suggesting that this represented a flagrant violation of the right to due process, a Canadian judge famously likened the fate of a blacklisted terror suspect he represented to that of Kafka’s Josef K. in The Trial.Footnote 1 Since then, protection of the due process rights of blacklisted terror suspects has been continuously improved. Today, listed individuals not only have the right to know that a travel ban and assets freeze have been imposed upon them, but also why. Moreover, the Security Council has appointed an ombudsperson who accepts complaints by listed parties and publicly advises the Council on whether the individuals who make the complaints should be delisted. So far, the Security Council has in most cases followed this advice, thus de facto accepting a significant restriction of its scope of action. Within a decade the UN Security Council’s sanctions regime has thus transitioned from one that blatantly violated human rights to one that even formerly fierce critics now attest shows ‘essential elements of due process’.Footnote 2
This is just one among many examples illustrating how international organisations (IOs) have established provisions for the protection of the human rights of individuals affected by their policies. But how did this specific type of institutional change within IOs come about? Standard International Relations (IR) theories do not answer this question satisfactorily. According to rational institutionalism, state interests are what drive IOs to legalise their activities. States favour legalisation if this reduces transaction costs and helps them to make credible commitments and solve the problem of incomplete contracting.Footnote 3 None of these interests are served by IO procedures that ensure due process for individuals. Rational institutionalists ought, therefore, to be sceptical about IOs committing to human rights protection. Liberal theories that stress the importance of domestic parliaments, courts and civil society in bringing about international legalisation have so far been used to explain states’ commitment to, and compliance with, international human rights law.Footnote 4 What has yet to be examined is whether domestic actors in IO member states also drive the evolution of human rights safeguards. From a two-level perspective, however, one would expect governments to have an incentive to keep IOs free from obligations that might reduce their opportunities to delegate ‘difficult’ decisions to the IO level.Footnote 5 Constructivist scholars have shown that persuasion and shaming, especially by transnational civil society organisations (CSOs), can lead states to establish strong mechanisms for the protection of human rights at the international level.Footnote 6 But again, no one has yet explored which CSO strategies bring IOs to accept their responsibility for the protection of human rights.Footnote 7
The purpose of this article is to fill this gap. We develop the authority-legitimation mechanism (ALM) and argue on the basis of a test covering ten case studies that it can explain why IOs develop provisions for the protection of human rights.Footnote 8 In short, the ALM posits that expanding the authority of IOs increases their propensity to violate human rights. This in turn provokes efforts at delegitimation and demands for the creation of human rights safeguards. If the authority-challengers find coalition partners, then IOs accept their responsibility for the protection of human rights and introduce relevant provisions. This explanation hints at a causal mechanism derived from the Weberian sociology of domination, according to which actors who are perceived to exercise authority need to nurture beliefs in their legitimacy among constituencies and audiences.Footnote 9
The ALM, we argue, can develop through different pathways and is therefore characterised by equifinality. Equifinality means that an outcome – in our case human rights safeguards in IOs – can be reached in a variety of ways.Footnote 10 We therefore combine the deductive element of testing a mechanism (the ALM) with an inductive element uncovering the more specific pathways through which the mechanism materialises. In one such pathway, parliaments in IO member states put pressure on domestic governments to campaign for human rights safeguards in IOs (legislative institution-building). In another pathway, domestic or international courts demand human rights safeguards in IOs (judicial institution-building). A decisive input may also come from so-called like-minded actors such as CSOs, middle powers and IO bodies with little formal power who expose rights violations by IOs and develop proposals for reform (like-minded institution-building). Finally, IOs may develop safeguards following the example of a role model (anticipatory institution-building). All in all, we show that IOs introduce human rights safeguards as a legitimation device to counter real or anticipated delegitmation challenges, which can build up through the interventions of domestic parliaments in IO member states, courts, or coalitions of like-minded actors. In this inductive part of our research, we also develop hypotheses on the conditions that trigger the individual pathways and determine whether the pathways lead to comprehensive or limited provisions of protection.
Finally, we suggest that the rise of human rights protection in IOs indicates a step towards the international rule of law and the emergence of elements of a liberal, human-rights-based pattern of IO legitimation.
The article is structured as follows. Sections II and III introduce the ALM and our conceptual framework. Section IV summarises the results of the empirical analysis. Section V presents the theoretical implications and sets out our conclusions.
The authority-legitimation mechanism (ALM)
IOs have, in operational terms, institutionalised authority when they are able to take decisions or develop interpretations that bind states or their societies, and these states accept this, even if compliance is not in their short-term interest.Footnote 11 This presupposes that states accept that IOs no longer take decisions or develop interpretations on the basis of unanimity. The institutionalised authority of IOs has grown significantly over time. If we compare today’s IOs with those of three decades ago, it becomes evident that the former clearly display more delegated and pooled authority.Footnote 12 IOs sometimes directly regulate the behaviour of individuals and in that way exercise authority not only over states, but also over persons.Footnote 13 In this case, states then cease to be mediators between IOs and their citizens. UN sanctions, for example, were originally devised to alter the behaviour of states; today, the addressees of sanctions are often terror suspects, warlords, or dictators and their entourages, hence individuals.Footnote 14 Likewise, when the UN, the European Union (EU), or the North Atlantic Treaty Organization (NATO) assume government functions in transitional administrations, they enter into direct authority relationships with individuals. Furthermore, when the World Bank and the International Monetary Fund (IMF) sponsor projects and programmes whose substance they heavily influence and whose implementation they closely supervise, they come close to exercising authority over individuals.
These new facets of IO authority imply that IOs have, under certain circumstances, the capacity to violate the human rights of individuals.Footnote 15 UN sanctions led to a humanitarian disaster in Iraq when child mortality rose dramatically as a result of the shortage of food and medical supplies. UN and NATO peacekeepers sexually exploited women and children and facilitated human trafficking.Footnote 16 Projects funded by the World Bank have expelled indigenous people from their ancestral homelands without compensation and have resulted in massive environmental destruction. IMF-sponsored programmes have aggravated poverty in recipient countries that were compelled to cut social spending. The EU’s Frontex agency has conducted pushbacks resulting in the violation of the right to asylum of refugees. The World Trade Organization’s (WTO) TRIPSFootnote 17 agreement initially prohibited the selling of generic medicaments in developing countries, thus putting a heavy strain on their public health services.
The ALM posits that authoritative IOs lose legitimacy if they violate human rights. Originally, responsibility for the protection of human rights was ascribed to states. Today, IOs conceived as public authorities are also expected to ensure that their policies do not violate human rights. IOs can no longer disclaim responsibility and shift the blame to their member states if rights violations occur in connection with their policies.Footnote 18 The permissive consensus vis-à-vis IOs is over, IOs are politicised and can no longer take it for granted that they will be perceived as legitimate actors. Actors whose rights are violated by IOs, or by other actors on behalf of IOs, demand that IOs acknowledge their ‘right to justification’.Footnote 19 In fact, human rights violations by IOs are challenged by different actors in different fora. The World Bank and the IMF experienced mass demonstrations in front of their headquarters. EU refugee policy is widely condemned by the media and CSOs. And NATO was sued by Serbia before the International Court of Justice for violating international humanitarian law in relation to the airstrikes during the Kosovo war.
According to the ALM, IOs increase their self-legitimation efforts when their legitimacy is challenged; the legitimacy of an organization is not stable but needs to be constantly cultivated through processes of legitimation.Footnote 20 If IOs confront a crisis of legitimacy, they take steps to (re)legitimate their claim to authority to make sure that they are once again recognised as legitimate wielders of authority. When IOs face allegations of human rights violations, they are expected to draw on a liberal narrative of legitimation. This narrative is based on legal accountability, the protection of individual rights, and legal equality. Accordingly, introducing provisions for the protection of the human rights of individuals affected by IO policies increases the perceived legitimacy of IOs. Focusing on procedures, the liberal pattern of legitimation is more demanding than the technocratic pattern, in which the emphasis is on output or the fulfilment of a mandate.Footnote 21 Yet, it is easier to achieve than the participatory pattern of legitimation, with its focus on the principle that all those affected by a decision should have a say in the decision and on democratic accountability.Footnote 22
We can thus summarise the logic of the ALM and conceptualise the mechanism as a succession of interlinked steps: (1) IOs have experienced a rise in authority and increasingly take decisions that directly affect individuals, that way circumscribing states’ role as mediators between IOs and their citizens or other individuals; (2) The increase in the authority of IOs has increased the likelihood that IOs commit human rights violations; (3) If IOs violate human rights, legitimacy is withheld from them, provided that concerned actors find opportunities to express disapproval and form coalitions; (4) To (re)legitimate their claim to authority, IOs draw on the liberal pattern of legitimation and introduce provisions for the protection of human rights (Figure 1).
Figure 1 The authority-legitimation mechanism (ALM).
Conceptual framework
We first apply the method of comparative process tracing in order to test the ALM. This theory-testing use of process tracing implies that we are examining whether we can detect the mechanism’s specified individual steps in one or several cases in which the mechanism’s starting point and endpoint are given – in our study ten cases in which IOs exercise authority and introduce provisions for the protection of human rights. In a second step, now following an inductive logic, we reconstruct different analytical pathways through which the ALM materialises.Footnote 23 In this step, we also investigate the variation in the quality of the protection provisions that emerge in different cases, employing the method of structured-focused comparison.Footnote 24 Each of the inductively identified pathways exemplifies the ALM and together they suggest equifinality. We thus combine different logics: we test the ALM, but inductively construct the different pathways.
Mechanisms and process tracing
Mechanisms are ‘recurrent processes linking specified initial conditions and a specific outcome’.Footnote 25 They are thick chains of events, or reactive sequences, that map the distance between an independent and a dependent variable.Footnote 26 Unlike others, we do not assume that a mechanism follows only one logic of action, such as coercion, competition, or emulation.Footnote 27 Rather, we conceive of a mechanism as a pattern of sequences of action and reaction involving both strategic and deliberative practices.
To test the ALM, we use the method of process tracing. Process tracing denotes ‘the analysis of evidence on processes, sequences, and conjunctures of events within a case for the purposes of either developing or testing hypotheses about causal mechanisms that might causally explain the case’.Footnote 28 The method is particularly suitable for discovering interacting causal processes in the context of complex causality.Footnote 29
Process tracing allows causal inferences to be drawn only if the researcher succeeds in reconstructing an uninterrupted causal story. We therefore draw on a variety of different empirical sources: primary documents produced by IOs and actors who interact with them, 70 semi-structured interviews conducted with IO officials, representatives of member states and CSOs, and secondary literature.
Pathways
We draw on the method of theory-building process tracing to expose a number of different causal pathways through which the ALM materialises.Footnote 30 These pathways share the general characteristics of mechanisms, but are more specific with regard to the actors that drive them, the strategies these actors apply, and the conditions on which these actors become engaged.
For the purpose of identifying these causal pathways, we assume that IO decision-makers face different modes of input from within and without the IO, namely reform proposals, arguing, shaming, defiance, coercion, and litigation. We also assume that there are different modes of reaction to that input on the part of IOs: IOs first start a reflection process before they opt for either defensive behaviour (ignoring, denial, immunisation, diversionary behaviour) or institutional change. To get from the identified modes of input and reaction to a specific pathway, we examine what sequence of input and IO reaction connects the human rights violation with the creation of human rights safeguards in IOs. We also assess what input by which actor has been the most important.Footnote 31
Human rights protection provisions in IOs
According to the ALM, all pathways are expected to lead to the development of some human rights protection provisions in IOs. Yet, the provisions vary across IOs. Human rights protection provisions in IOs may include prevention and complaints provisions, that is, provisions that prevent human rights violations in the first place and provisions that enable complaints to be made by aggrieved individuals.Footnote 32 To assess the quality of human rights safeguards in IOs we developed six criteria for each type of provision. Each criterion is assigned the value 0, 1, or 2, depending on the extent to which it is fulfilled.
To assess the quality of IOs’ prevention provisions, we rely on the following criteria: obligation, precision, scope, addressee, compliance management, and mainstreaming. Obligation and precision refer to the degree of bindingness and clarity, respectively, of the provisions. Scope relates to their substantive breadth. Addressee refers to the actor to whom the provisions apply. Compliance management corresponds to the managerial arrangements to guarantee implementation of the provisions. Mainstreaming refers to the question of whether the provisions apply to all or only some applications of the policy.
To assess the quality of IOs’ complaints provisions, the criteria we use are: delegation, obligation, complainant, accessibility, remedy, and mainstreaming. Delegation refers to the extent to which an independent body exists that has the competence to adjudicate on complaints. Obligations relates to the degree of bindingness of the decisions on complaints. Complainant refers to the question of whether aggrieved individuals themselves or only third parties can file an action. Accessibility corresponds to the ‘user-friendliness’ of the complaints mechanism. As above, mainstreaming refers to the question of whether the provisions apply to all or only some policy applications.
We trace the evolution of human rights safeguards in each IO starting from the year in which the safeguards were first established and ending in 2012. For each year, we create an aggregate value between 0 and 2. We label provisions with an aggregate value between 0.5 and 1.0 ‘limited protection provisions’, and provisions with an aggregate value above 1.0 ‘comprehensive protection provisions’. To create the aggregate values, we first assess separately the annual values of the prevention and complaints provisions. We then create the overall annual values by aggregating the values for prevention and complaints.Footnote 33
Trigger and success conditions
The initiation of the pathways to human rights protection in IOs and their quality depend on specific conditions relating to the features of the IOs, their environments and the human rights violations they have perpetrated. To generate hypotheses on the trigger and success conditions of different pathways related to the ALM, in each case study we ask why a specific pathway has been activated and why it led to limited or comprehensive protection provisions.
To guide this inductive part of our empirical analysis, we assume that the trigger and success conditions of the pathways are likely to be related to the following four aspects: first, the fact that the IO disposes of an organisational culture that facilitates learning. This type of organisational culture is typically associated with a flat decision-making processFootnote 34 and boundary-spanning units that transfer expertise from outside the IO into the IO.Footnote 35 Second, the vulnerability of an IO to pressure. An IO may be vulnerable if the human rights violations are likely to generate a powerful campaignFootnote 36 and if its identity makes it an easy target for such a campaign.Footnote 37 IOs may also be vulnerable if the rights violation is considered justiciable and if human rights safeguards entail no costs for the actors that propagate them. Third, the power of the actors who demand rights protection. Domestic parliaments are powerful, for instance, if they are independent from the executive and if they are positioned in a strong IO member state. Courts can be powerful actors if they have jurisdiction over the IO or important member states. Finally, IOs may be more likely to establish human rights protection provisions if there are role models they can follow. IOs can learn from each otherFootnote 38 or from domestic models among their member states.
Case selection
To test the ALM, we selected cases that exhibit the mechanism’s first (IO with authority) and last (human rights protection provisions in IO) component. We test whether the cases display the expected sequence in between, and we explore via which pathways the mechanism unfolds. We selected ten cases, establishing something that can be labelled a comparative process-tracing approach.
This approach comes with costs for the second step in our research design – inductively identifying the trigger and success conditions for the different pathways. From this perspective, we have made our selection on the basis of the dependent variable. To arrive at generalisable findings and increase the probability of discovering different pathways, the selected IOs that cover different issue areas, execute different policies, and commit different types of rights violations. To discover the success conditions of the pathways, the sample includes cases in which IOs have created limited protection provisions only and cases in which IOs have proceeded to comprehensive protection provisions. Finally, our research design is based on five pairwise comparisons to lessen the drawbacks of selecting on the dependent variable. In three comparisons we juxtapose pairs of cases in which IOs have committed the same rights violations, but established different safeguards, namely either limited or comprehensive ones. This allows us to examine the conditions that determine the different quality of the safeguards, while keeping the effect of confounding variables to a minimum. In the two remaining comparisons, we match pairs of cases in which IOs have developed provisions of the same quality in relation to the same rights violations. These comparisons are intended to examine the conditions that activate different pathways, again minimising the effect of confounding variables.
On the basis of these considerations we selected the following cases for analysis: four case studies of UN and EU sanctions policy – two that trace the evolution of provisions in these two IOs designed to eliminate the violation of subsistence rights connected with comprehensive trade embargoes, and two that reconstruct the development of provisions for the protection of due process rights for blacklisted individuals; four case studies dealing with UN and NATO peacekeeping – two that analyse the evolution of provisions to prevent sexual exploitation and/or human trafficking by IO personnel and provide avenues for complaints and two that trace the emergence of provisions in the two organisations to protect the due process rights of detainees; and, finally, two case studies on World Bank and IMF lending that analyse the development of provisions for the protection of subsistence and, in the case of the World Bank, cultural rights (Table 1).
Table 1 Cases.
Results
The growth of human rights protection in IOs
In each of the ten cases we analysed, IOs arranged for provisions to eliminate, or at least minimise, the violation of human rights related to their policies (prevention provisions). In some cases, IOs introduced additional provisions that enable aggrieved individuals to hold them accountable (complaints provisions). Overall, the protection provisions that emerged in the different cases differ in their quality. In five cases IOs established comprehensive protection provisions, namely with values between 1.21 and 1.84 points. In the other five cases, only limited protection provisions emerged that reached values between 0.63 and 0.92 points (Figure 2).Footnote 39
Figure 2 Human rights safeguards in IOs (2012).40
Comprehensive protection provisions were institutionalised in the two UN peacekeeping cases. In the case of sexual exploitation and abuse by UN peacekeepers, the UN Secretariat introduced regulations governing how mission personnel were to behave, and developed mandatory training modules.Footnote 41 Complaints and investigation provisions in the form of Conduct and Discipline Units were established in the Secretariat and in the field (1.73 points).Footnote 42 In the case of the violation of detainees’ due process rights in peace operations, the UN Secretary-General issued a regulation that bound the treatment of detainees to the standards established by the Geneva Conventions and customary international law.Footnote 43 In the UN operation in Kosovo, a detention review panel and an ombudsperson institution were created to which detainees could address their complaints (1.21).Footnote 44
Comprehensive provisions were also established in the cases involving due process violations in UN and EU sanctions policy. The UN Security Council has granted blacklisted individuals the right to know the reasons for their listing.Footnote 45 It has also appointed an ombudsperson to take complaints by listed terror suspects and advise the Security Council on individuals to be delisted from the sanctions regime (1.42).Footnote 46 The EU has also granted individuals the right to know the reasons for being listed in a sanctions regime. Moreover, individuals listed in an EU sanctions regime can appeal to the European Court of Justice (ECJ) against their listing (1.84).Footnote 47 Finally, comprehensive protection provisions emerged in the World Bank, which introduced social and environmental standards to guide the design and implementation of the projects for which it provided funding.Footnote 48 It also created the Inspection Panel to which individuals can direct their complaints if the Bank fails to respect its safeguards (1.73).Footnote 49
Limited provisions for the protection of human rights were introduced by the UN and the EU with regard to protecting the subsistence rights of individuals affected by their sanctions regimes. Both IOs have committed themselves to adopting targeted sanctions instead of sweeping trade embargoes.Footnote 50 Moreover, both organisations grant humanitarian exemptions and require an assessment of the humanitarian impact of their sanctions (0.84 and 0.75 for the UN and the EU, respectively).Footnote 51 Furthermore, limited provisions were established by NATO to prevent human rights violations in peacekeeping operations. To prevent NATO peacekeepers from being involved in human trafficking and sexual exploitation, NATO introduced a code of conduct and training modules (0.92).Footnote 52 As for the treatment of detainees, NATO introduced guidelines and a review panel for its Kosovo operation (0.63).Footnote 53 Lastly, the IMF created limited provisions to mitigate the negative impact on human rights of its lending policy. With its Poverty and Social Impact Analysis (PSIA), the organisation established at least some standards for evaluating the social impact of IMF-funded projects (0.67).Footnote 54
The ALM, its pathways, and related conditions
The ALM holds that IOs, which, during the course of exercising authority over individuals, violate human rights, must introduce human rights safeguards when the legitimacy of their claim to authority is challenged. The ALM can be observed in nine of our ten cases: the UN and the EU created safeguards for the protection of due process rights in their sanctions policies when they faced a legitimacy problem; similarly, the UN’s establishment of provisions to protect the subsistence rights of individuals affected by its sanctions was also a response to efforts at delegitimation. The UN and NATO introduced human rights safeguards for their peacekeeping operations after having been confronted with widespread allegations of human rights violations. And the World Bank and the IMF agreed to introduce provisions for the protection of human rights in response to accusations that the projects and programmes that received their funding gave rise to human rights violations, thus challenging their legitimacy. In one case, the ALM could not be confirmed: the EU established provisions for the protection of subsistence rights in its sanctions policy without being under pressure. This response followed allegations made against the UN, however; the EU wanted to avoid the same fate.
In line with the notion of equifinality, the ALM materialises in the form of different pathways. Based on the input–reaction scheme introduced in Section III we have inductively identified three pathways that differ in terms of who is the most important actor pushing the reform process and what strategy that actor applies. We have named these pathways legislative, judicial, and like-minded institution-building, indicating the pathways’ key actors and strategies. In the EU case study just mentioned we discovered a pathway that deviates from the ALM inasmuch as the IO introduced human rights safeguards without having either violated human rights or faced a challenge to its legitimacy; as we show below, the case nevertheless provides indirect support to the ALM. In this section, we outline the four pathways identified in our ten cases and formulate insights on the trigger and success conditions of each pathway.
Legislative institution-building. In this pathway, parliaments in democratic member states significantly influence the development of human rights protection provisions in IOs. The pathway is rooted in liberal foreign policy theory, which attributes a key influence on foreign policy in democratic states to domestic actors.Footnote 55 The parliament is the main organ through which individuals, interest groups and CSOs can channel their demands. While parliaments are often portrayed as obstacles to the implementation of international agreements, the promotion of human rights in foreign policy can actually profit from parliaments’ influence on the executive.Footnote 56 With regard to the establishment of human rights protection provisions in IOs, the pathway legislative institution-building comes in two variants. In the direct variant, a parliament tries to promote human rights safeguards in IOs by passing laws that bind the executive, or by refusing to ratify a treaty or allocate funding unless its demands are met. In the more indirect variant, a parliament impacts on the executive’s foreign policy by conducting investigations, agenda setting, and shaping the domestic discourse.Footnote 57
The direct form of this pathway occurred in the World Bank case. The Bank’s introduction of safeguards and of the Inspection Panel was heavily influenced by the United States (US) Congress making the allocation of government funds to the Bank dependent upon such reforms. Congress had been mobilised before by CSO campaigns that exposed the negative social and ecological impact of World Bank projects. In response to Congress’s actions, the US Executive Director at the World Bank convinced other Executive Directors and the management to agree to reforms.Footnote 58 The indirect form of this pathway is evident in the case study on the emergence of protection provisions in NATO following peacekeepers’ involvement in human trafficking and sexual exploitation. The US and Norwegian ambassadors to NATO introduced these provisions against the background of investigations by the US Congress into reports of misconduct by US troops in Korea and the Balkans. In reaction to these allegations, President Bush had introduced a ‘zero-tolerance policy’ regarding human trafficking in peace operations.Footnote 59 Following this model, the two ambassadors managed to persuade the other NATO members to adopt the ‘NATO Policy on Combating Human Trafficking’.Footnote 60
The cases also provide insights into the conditions that trigger the pathway legislative institution-building. First, the pathway seems to require a substantial degree of autonomy for the parliament in question vis-à-vis the executive, as is the case in presidential systems in particular. In both instances, the US Congress was able to further the adoption of human rights protection provisions in IOs thanks to its considerable independence from the executive, whether in terms of its formal competences or its informal powers – which parliaments in parliamentary systems normally lack.Footnote 61 Second, the pathway was in both cases facilitated by the existence of comparable human rights protection provisions at the domestic level in the state whose parliament became engaged. In fact, the reforms adopted by the World Bank and NATO were inspired by domestic scripts; namely, US legislation on social and environmental standards for unilateral US development lending and legislation on the prohibition of sexual exploitation by US and Norwegian military staff in the two countries.Footnote 62
Finally, the two cases indicate the success conditions for the pathway legislative institution-building, that is, the conditions under which the pathway leads to comprehensive rather than limited protection provisions. One condition of success seems to be that the parliament is located in a powerful IO member state. This is in line with rational institutionalism, but also with hegemonic law-making theories, which assume that IOs are shaped by the interests of their dominant member states.Footnote 63 The cases also suggest that comprehensive provisions only emerge if they do not entail sovereignty costs for the state the parliament represents. The US pushed for ambitious regulations in the World Bank as these applied only to Bank staff and borrowing statesFootnote 64 while, with regard to NATO, the US promoted only limited provisions as these also bound the actions of US soldiers in NATO operations.Footnote 65 This ambivalence corresponds to the assumptions of delegation theories regarding the interaction of dominant member states with IOs in generalFootnote 66 and US human rights policy more specifically.Footnote 67
Judicial institution-building. This pathway implies that the decisive input leading to the adoption of human rights protection provisions in IOs is provided by higher courts. These courts can engage in law-making at the expense of the autonomy of the legislature and the executive,Footnote 68 either for the purpose of filling legal gaps through specification or by judicial activism. The latter is regarded as more controversial since it involves judges deliberately using their role as interpreters of the law to transform existing legislation.Footnote 69 The very nature of law empowers judges as it needs to be interpreted, applied to specific contexts and evaluated with regard to potential conflicts of norms.Footnote 70 At the same time, however, judges depend on plaintiffs putting forward cases and on member states accepting their authority. Thus, while states increasingly delegate far-reaching competences to autonomous courts, thereby increasing the autonomy of judges,Footnote 71 judges at the same time need to be careful not to permanently disregard the interests of powerful member states.Footnote 72 If judges go too far, IO staff or member states that do not accept the courts’ judgements can protract or circumvent their implementation.Footnote 73
This pathway occurred in the case of the protection of due process rights in EU sanctions policy. Terror suspects blacklisted by the EU filed complaints to national courts and to the ECJ, claiming that their due process rights were being violated by the EU. In a landmark decision in 2006, the court annulled the sanctions against a complaining partyFootnote 74 and, as a consequence, the Council of the EU reformed its listing and delisting procedures.Footnote 75 Further protection provisions were institutionalised as more and more lawsuits brought before the ECJ were decided against the EU.Footnote 76 The due process protection provisions that evolved in UN sanctions policy also resulted from the pathway. In this case, affected individuals also filed complaints before national courts and the ECJ, challenging the implementation of UN sanctions in EU member states. In response to the ECJ’s famous Kadi decisionFootnote 77 and related judgements, in which the court ruled that UN sanctions violated fundamental rights as guaranteed by EU law, the UN Security Council significantly improved its listing and delisting procedures.Footnote 78
The two cases indicate that the pathway of judicial institution-building is only triggered if a right that is considered to be justiciable is violated – in both cases a civil right, namely the right to due process, which is regarded as enforceable by a court and which has been regularly enforced by courts in the past. Unlike civil and political rights, social and economic rights are generally not regarded to be justiciable, nor is there a tradition of legal enforcement.Footnote 79
Regarding the achievement of comprehensive protection provisions, the cases suggest that it makes a difference if the activated court is a strong one. This is the case whether the state, or the community of states, over which the court has jurisdiction is a key player within the IO, as in the UN case, or if the court has jurisdiction over the IO as a whole, as in the EU case. This resonates with the observation that the ECJ has become a powerful court when it comes to protecting individual rights vis-à-vis the EU. The high reputation of the court increases compliance with its decisions, which has made strategic litigation before the ECJ a promising instrument for enforcing comprehensive human rights protection provisions not only in the EU but also, indirectly, in other IOs.Footnote 80
Like-minded institution-building. This pathway is characterised by decisive input from a coalition of norm entrepreneurs within and outside the IO. These actors shame the IO for its human rights violations and put forward reform proposals. While the term ‘like-minded’ is usually employed to describe a coalition of CSOs and states that pursue a common agenda,Footnote 81 we use this concept to describe actors who possess neither core institutional power resources within the IO, nor are executives from powerful member states. These actors can form part of a coherent and sustained social movement, or their cooperation can be looser and less durable.Footnote 82 Like-minded actors organise powerful transnational campaigns to increase moral pressure on the IO in order to make it introduce protection provisions.Footnote 83 Through framingFootnote 84 and persuasion,Footnote 85 they expose the discrepancy between the IO’s actions and existing norms to national and transnational publics and provide arguments for reform.
This pathway can be identified in five of the ten cases in our study, three of them in the area of peacekeeping. In the case of sexual abuse by UN peacekeepers, numerous reports by non-governmental organisations (NGOs) and international media, as well as reform proposals put forward by the UN Secretariat, led to the adoption of protection provisions by UN intergovernmental bodies.Footnote 86 Similarly, the protection provisions regarding the violation of due process rights in UN detention evolved in response to criticism and reform proposals by NGOs and regional organisations.Footnote 87 In the case of the violation of due process rights in NATO detentions, criticism and reform proposals came from local and transnational NGOs as well as regional intergovernmental organisations, which prompted the organisation to adopt at least limited protection provisions.Footnote 88
Two further cases in which the ALM unfolded via the pathway are that of subsistence rights violations in UN sanctions policy and of IMF lending. In the UN sanctions case, NGOs and humanitarian UN agencies organised a powerful campaign targeting the disastrous UN sanctions regime against Iraq, which had provoked massive suffering among the Iraqi population.Footnote 89 At the same time, a coalition of middle powers and academics joined forces and developed recommendations, which eventually led the Security Council to introduce at least limited protection provisions.Footnote 90 Likewise, in the case of the IMF, public campaigns as well as recommendations by the IMF’s own Independent Evaluation Office disposed the organisation to adopt at least limited provisions to minimise the human suffering that its structural adjustment programmes in particular caused in recipient countries.Footnote 91
The cases suggest that the pathway like-minded institution-building can be activated if there is a structure of political opportunity that facilitates a powerful campaign. This implies that the human rights violations can be made visible, either because innocent victims’ physical harm can be portrayed visually in campaigns, or because a short causal chain between the violating IO and the victims can be constructed.Footnote 92 In all five cases, at least one criterion was given. In the case of sexual exploitation in UN peacekeeping, it was possible to generate both empathy with the victims and a short causal chain designating blue-helmets as the perpetrators.Footnote 93 In the cases of violation of due process rights in UN and NATO detention, empathy with the victims, who were suspected terrorists, was more difficult to establish. Instead, campaigners successfully focused on attributing the responsibility for the violations to the international authorities in charge of the detainees.Footnote 94 In the cases of violation of subsistence rights in UN sanctions policy and IMF lending, the causal chain was rather diffuse. Instead, campaigners resorted to portraying the physical harm innocent victims incurred.Footnote 95
The cases also provide hints on the conditions for success under which the pathway leads to comprehensive protection provisions. Mobilisation seems to be successful only if the IO is vulnerable to human rights campaignsFootnote 96 and displays an organisational culture that is conducive to learning.Footnote 97 In the two UN peacekeeping cases, the Secretariat was vulnerable to human rights campaigns because the organisation’s identity was strongly imbued with the protection of human rights.Footnote 98 Moreover, the UN Secretariat’s organisational culture was open and receptive to external knowledge. As a consequence, the campaigns and recommendations fell on good soil and led to the adoption of comprehensive provisions.Footnote 99 In the other cases, none or only one of these criteria was given. In the case of UN sanctions, the Security Council was vulnerable to human rights campaigns, yet its organisational culture was not particularly open to learning.Footnote 100 In the cases of NATO and the IMF, the organisations were less vulnerable as their mandates focused on military and economic or financial aspects. Moreover, their organisational culture was rather closed and hierarchical, therefore not conducive to learning.Footnote 101 As a consequence, in the latter four cases only limited protection provisions evolved.
Anticipatory institution-building. Finally, we discovered a pathway through which an IO establishes human rights safeguards, even if there is no direct pressure for such reforms. We call this pathway anticipatory institution-building. This pathway may be based on internal lesson-drawing or on the evaluation of the experiences of peer IOs in the same policy field.Footnote 102 In the latter case, IO decision-makers can learn from the policy failures of other IOs and draw lessons for their own policies.
The pathway of anticipatory institution-building renders comprehensible why the EU introduced provisions for the protection of subsistence rights of individuals affected by its own sanctions. The EU itself was not accused of subsistence rights violations, but could observe that the UN was facing harsh criticism for them.Footnote 103 Moreover, EU representatives participated in meetings that served to develop proposals for how the civilian population could be spared the negative effects of UN sanctions.Footnote 104 As these proposals proved to be relevant for the EU as well, the organisation preventively adopted comparable protection provisions to avoid becoming the target of a harmful campaign similar to that which faced the UN.Footnote 105
The EU case suggests that the activation of this pathway depends on the presence of a reference organisation, which executes the same policy, but commits human rights violations in the process, and therefore comes under heavy criticism to which it responds with the introduction of human rights protection provisions. Such a reference organisation can serve as a role model not only in terms of the safeguards it introduces but also in terms of showing that human rights violations do not go unnoticed.Footnote 106 The case also suggests that, for the pathway to be triggered, the IO needs to possess an organisational culture that is conducive to learning and information processing, that is, an organisational culture marked by a flat decision-making process and an open, boundary-spanning institutional structure. Given that, in our case, this pathway only led to limited provisions, no success conditions can be derived from it. It might well be, however, that lesson-drawing not only takes place regarding whether or not to adopt protection provisions but also regarding their quality. Learning from the UN’s experience, the EU could conclude that limited provisions were sufficient to reduce external pressure.
Strictly speaking, the case does not provide direct support for the ALM. The EU introduced provisions that were designed to make sure that subsistence rights are not violated when it imposes sanctions, without it having violated such rights or having been criticised for violating them. We argue, however, that the case does not disprove the ALM either, because the EU introduced human rights safeguards in order to avert the dynamics associated with the ALM. The EU observed that the UN, which did violate subsistence rights, was delegitimated as a result, and pre-emptively introduced human rights safeguards to escape the fate of its sister organisation.Footnote 107 In this sense, the case does provide indirect support for the ALM.
Table 2 summarises our findings:
Table 2 Pathways, cases, and hypotheses on conditions.
Conclusion and theoretical implications
In this article we have argued that the ALM accounts for the establishment of provisions to ensure that IO policies do not violate human rights. We were able to observe the ALM in nine of our ten case studies on UN and EU sanctions policy, UN and NATO peacekeeping, and World Bank and IMF lending. In all these cases IOs introduced human rights safeguards to regain a legitimacy that had been challenged when human rights violations were exposed. The remaining case study on the EU introducing provisions for the protection of subsistence rights in its sanctions policy is the anomaly, in that the EU introduced provisions without having been criticised for human rights violations. Yet, it does provide indirect support for the ALM, given that the EU learned from the UN’s experience and pre-emptively introduced human rights safeguards to prevent challenges to its legitimacy.
We have, moreover, inductively shown that the ALM can materialise in the form of different pathways, confirming the assumption of equifinality according to which one outcome can be the reached via different routes. Which of the pathways is taken depends on the presence or absence of specific conditions that trigger each pathway. The main input may come from domestic parliaments in IO member states, if these parliaments enjoy sufficient autonomy from the executive (legislative institution-building). National and international courts can drive the introduction of human rights protection provisions, provided the violated right is considered justiciable (judicial institution-building). If the rights violation can be made visible, formally weak actors such as CSOs and middle powers can construct a damaging campaign (like-minded institution-building). If IOs learn from a role model, the fourth pathway (anticipatory institution-building) comes into play, which is at least indirectly linked to the ALM.
Finally, we have inductively identified the conditions under which the pathways are particularly effective. It seems that whether or not the pathway we call legislative institution-building leads to comprehensive protection provisions depends on the position of the parliament and on the sovereignty costs the provisions entail. The pathway we call judicial institution-building may result in comprehensive provisions if the court involved has jurisdiction over the IO or powerful member states. And the pathway we label like-minded institution-building seems to be comparably effective if the vulnerability of the IO can be exploited and if the IO has the capacity to learn.
Our findings suggest that the concept of the rule of law when applied to IOs has undergone a significant transformation. Traditionally, scholars, if they thought of the rule of law in relation to IOs at all, applied a thin or formal conception of it. Accordingly, legitimate IOs were expected to act in conformity with their legal mandate, generate clear and certain law, and, more recently, introduce legalised dispute settlement bodies to adjudicate conflicts arising in the application of the law.Footnote 108 Today, IOs are increasingly expected to abide by a thick, or substantive, conception of the rule of law. IOs are no longer expected only to respect procedural requirements, but also to protect the fundamental rights of individuals affected by their policies, thereby committing to what has been termed humanity’s law.Footnote 109 This also suggests that what we called a liberal pattern of IO legitimation can complement and partially substitute other legitimation patterns.Footnote 110 While much of the debate on the legitimation of IOs centres on IOs’ democratic deficit, legitimation through participatory democratic procedures is difficult, because important preconditions, such as a transnational demosFootnote 111 or equal access to global democratic institutions,Footnote 112 are seldom given. Legitimation through human rights protection should be less demanding, as it essentially requires IO decision-makers and member state representatives to commit to basic rule of law standards.
Can we extrapolate from our findings the future of the current trend of IOs accepting their responsibility for the protection of human rights? We believe that there are several indications that the trend will solidify. If we look beyond our case studies, there indeed seems to be a general trend of authority-wielding IOs creating provisions for the protection of human rights. Fifteen of the twenty IOs with the highest name recognitionFootnote 113 have meanwhile introduced rules for the protection of human rights: the African Union (AU), the Council of Europe, the EU, the Food and Agriculture Organization of the UN (FAO), the International Criminal Court (ICC), the International Labour Organization (ILO), the IMF, the Organisation for Economic Co-operation and Development (OECD), the Organization for Security and Co-operation in Europe (OSCE), the Southern African Development Community (SADC), the UN Development Programme (UNDP), the UN Educational, Scientific and Cultural Organization (UNESCO), the UN Refugee Agency (UNHCR), the World Health Organisation (WHO), and the World Bank.Footnote 114 Hence, the emergence of human rights safeguards in IOs is not confined to those that are composed of, or dominated by, Western states like those in our sample. Nor is the trend restricted to IOs operating in the areas of security and development or finance, like those in our sample.
Moreover, the conditions that have given rise to the pathways associated with the ALM are likely to hold. States continue to delegate competences to IOs; this tendency may have slowed down to some extent in recent years, but it has not been reversed.Footnote 115 We might therefore expect more human rights violations by IOs in the future, and thus more demand for respective protection provisions. Furthermore, the pathway that describes the impact of courts, and which in our case studies had the best success rate of all the pathways we discovered, may become more important in the years to come. Indeed, states not only delegate competences to IOs but also create more and more international courts with private access and compulsory jurisdiction.Footnote 116 Besides, domestic courts not only increasingly issue judgements on the domestic implementation of IO policies, but have also begun to concern themselves with disputes between IOs and private litigants, at least with regard to private law issues.Footnote 117
Lastly, there seems to be an increasing number of cases in which IOs introduce human rights safeguards without having violated human rights in the first place. Examples are regional development banks introducing accountability mechanismsFootnote 118 and the EU introducing provisions to prevent sexual abuse by its peacekeepers.Footnote 119 In these cases IO decision-makers may well have observed that peer IOs (the World Bank, the UN) have come under pressure for violating human rights when executing similar policies. These cases also show that direct pressure on IOs is not necessarily always needed to make IOs introduce human rights safeguards.
At the same time, we cannot fail to mention that the global power transition, and especially the rise of China, may slow down the spread of human rights safeguards among IOs. This has already become evident in the case of the World Bank. The organisation, under pressure from China and other rising powers, is currently watering down its social and environmental safeguards, jeopardising what has been achieved in terms of human rights protection in the past decades.Footnote 120 More generally, two of the ALM’s pathways, namely the pathways that describe the impact of domestic parliaments and of like-minded actors, may lose some of their leverage if states which oppose strong human rights safeguards gain influence in IOs: Western parliaments, especially the US Congress, will lose much of their influence when the US and other Western powers lose their preeminent position in IOs. Likewise, IOs will be less vulnerable to pressure from like-minded campaigns if they face strong counter-pressure from China and other powers with less regard for human rights. Nonetheless, our cases do indicate that challenges by states with differing views are unlikely to succeed in fully dismantling those safeguards in IOs or to halt their evolution completely. In the case of the World Bank, for instance, China is not fighting for the abolition of the Bank’s safeguards but rather for their dilution. Likewise, in the case of the UN Security Council’s sanctions policy, China eventually gave up its opposition to the creation and subsequent strengthening of an ombudsperson. How the trend of IOs introducing human rights provisions will evolve in the years to come is, therefore, an open question. It is certain, however, that IOs cannot go back to the era in which human rights protection was not expected of them.
Acknowledgements
This article is based on the results of a project titled ‘International Organizations and the Protection of Fundamental Rights’. We gratefully acknowledge generous financial support by the German Research Foundation for this project. We are also very grateful for all the helpful comments on the project we received from the many colleagues with whom we had the opportunity to discuss the project, especially Thorsten Benner, Philipp Dann, Klaus Dingwerth, Andrea Liese, Peter Mayer, Georg Nolte, and Bernhard Zangl. We thank Theresa Reinold for conducting the case study on the IMF and Rebecca Majewski, Markus Patberg, Friederike Reinhold, and Lina Staubach for research support. Finally, we cordially thank our interview partners from various institutions for sharing their insights with us.
Supplementary material
To view supplementary material for this article (Appendices I and II), please visit: https://doi:10.1017/S0260210517000420
