ACCOUNTABILITY MECHANISMS OF MULTILATERAL DEVELOPMENT BANKS AND THE LAW OF INTERNATIONAL RESPONSIBILITY

Abstract

Multilateral development banks (MDBs) are international organizations subject to the law of international responsibility. Yet, the relationship between their accountability mechanisms and the International Law Commission (ILC) Articles on the Responsibility of International Organizations (ARIO) remains unclear. Understanding this relationship is essential in fully realizing the right to remedy in the development finance context. A comparative analysis of these legal frameworks clarifies that notwithstanding their different rationale, scope and functions, the two are not normatively conflicting and both serve to control public power. While the accountability mechanisms correct the ARIO's State-centric orientation by granting legal standing to project-affected people, they have their own deficiency concerning the actions they can prescribe to MDBs upon a finding of noncompliance. Highlighting that the MDBs’ mandate to ‘do no harm’ and pursue sustainable development is left unfulfilled by the accountability mechanisms’ deficient remedial function, this article identifies specific ARIO provisions to complement rather than undermine the MDBs’ accountability system. The ARIO's residual character, combined with the proposition that remedies arise not only from wrongful conduct but also from harm suffered by one party due to another's risky activities, justify this complementarity.

I. INTRODUCTION

Accountability as a concept in international law was born in 1993, to paraphrase David Hunter, since that year marked the creation of the World Bank Inspection Panel, the pioneer independent accountability mechanism (IAM), which ‘disrupted the normal channels of accountability in international law’.¹ It remarkably took another decade for international lawyers to study the concept of accountability systematically, as applied to international organizations (IOs).² They focused instead on responsibility: the International Law Commission (ILC) finalized the Articles on the Responsibility of States for Internationally Wrongful Acts (ARSIWA) and the Articles on the Responsibility of International Organizations (ARIO) in 2001 and 2011, respectively.³

Since their establishment after World War II, international financial institutions (IFIs)—or multilateral development banks (MDBs)⁴—like the World Bank have significantly expanded their mandates. In nearly 80 years of existence, however, they have been largely excluded from legal critique and the harms resulting from their conduct left unremedied. Similar observations have been made regarding other IOs, including the United Nations (UN).⁵ The advent of accountability in international law, particularly in the law of IFIs,⁶ is thus widely celebrated. The IFIs themselves have readily embraced the notion of accountability, in stark contrast to their chronic and obstinate rejection of international responsibility for social and environmental harms connected to their activities.

Still, accountability is not yet fully understood as an international legal concept, and its relationship with responsibility needs further clarification. Indeed, despite the growth of noncompliance mechanisms, ‘the nature and content of the accountability principles and their relationship with the law of responsibility remains ill developed, in particular where it concerns principles relevant to situations of shared responsibility’.⁷ More concretely, the characterization of IAMs and their interaction with the international responsibility regime remain unsettled.

These uncertainties have partly allowed IAMs to become a pretext or smokescreen for not scrutinizing IFIs’ responsibility under international law. Yet, because IFIs are IOs and subjects of international law, they bear international legal rights and obligations and their conduct merits closer legal scrutiny. Within the justifiably criticized legal framework of the IFIs, the notion of human rights is absent and deliberately evaded. Rather than rehashing the credible and strong arguments against such evasion, this article examines the concepts and principles that IFIs seem to be utilizing in lieu of responsibility, rights and remedies. It inquires how accountability in the development finance context relates to the law of international responsibility, and examines the implications of this relationship for the right to remedy of persons and communities harmed by IFI-supported projects, focusing on the rules and policies of MDBs concerning IAMs, and the ARIO.

The literature on this topic is limited and authors vary in their portrayal of IFIs’ accountability vis-à-vis responsibility under international law. Orakhelashvili distinguishes the concepts, but concludes that although the Inspection Panel can only determine the World Bank's accountability, it ‘can [also] deal with the Bank's [international responsibility] by flagging violations of the applicable legal standards that may be inherent in certain violations of the Bank's policies and procedures’.⁸ Similarly, Ong finds that ‘the development of accountability mechanisms for public and private IFIs is a viable alternative to the attribution of responsibility to these non-State actors under international law’,⁹ while conceding that such alternatives ‘still fall short of ensuring that justice prevails in every instance of a breach of international environmental obligations and/or standards’.¹⁰ Arguing that ‘[t]he draft articles are inadequate for an institution like the World Bank’, which plays a ‘specific supervisory role at various project cycle stages’, Baimu and Panou stress an irreconcilable difference between the Bank's accountability system and the law of international responsibility, while also claiming that ‘the remedial regime under the panel's legal framework is consistent with the main principles of the draft articles’.¹¹ Since it relies on a questionable and unsubstantiated difference between the characterization of the Inspection Panel as a compliance mechanism and the ‘uncertainty about the Bank's international obligations’,¹² their argument is unconvincing. It can be gleaned from these works that the main obstacle to applying the responsibility regime to MDBs is the purported imprecision, if not lack, of the international legal obligations borne by IFIs qua IOs. This point is addressed below in discussing the MDBs’ sustainability mandate.

Accountability and responsibility are not mutually exclusive alternatives, but Ong correctly highlights that IFIs’ accountability mechanisms have the potential to overcome some conceptual and practical difficulties—eg identifying an IO's international obligation—involved in establishing institutional responsibility. While disagreeing with the conclusion that the ARIO and Inspection Panel are separate paths that are unlikely to converge, this article adopts Baimu and Panou's approach of ascertaining ‘the added value [per Simma and Pulkowski] that general international law could bring to the panel’,¹³ meaning the circumstances in which reliance on general international law ‘is expedient to serve the purposes of the special regime’.¹⁴ One area where this approach can be useful—and thus employed here—concerns the implementation of project-affected people's right to remedy in the development finance context. Results of the comparative analysis in Section V corroborate the persisting criticism that IAMs ‘have not extended ‘the notion of (institutional) accountability into the requirement under international (State) responsibility … for full reparation for any injury … suffered by the victims of breaches of international norms’.¹⁵

Remedy in the development finance context finds bases in the banks’ sustainable development and ‘do no harm’ mandates. As the Office of the High Commissioner for Human Rights (OHCHR) posits, doing no harm does not only mean prevention, rather it also logically requires remedying any harms done.¹⁶ To introduce remedy in the development finance context, it is necessary to ‘transcend the punitive assumptions and associations with remedy and approach the issue from the standpoint of contingency planning’.¹⁷ Otherwise stated, it needs to be clarified that wrongful conduct is not always and necessarily a prerequisite for remedies, and that the latter may instead be based on or justified by the occurrence of harm, which can result from lawful but risky or hazardous activities.¹⁸

Two main claims are advanced here. First, the relationship between the IFIs’ accountability system and the ARIO is one of complementarity rather than of inconsistency and preclusion, since both serve to control public power. Second, neither acting alone can fully implement the human right to remedy. The law of international responsibility is deficient in this regard, because its definition of injured parties is restricted to States. The IAMs resolve this limitation but have their own deficiency concerning the actions they can prescribe to IFIs upon a finding of noncompliance.

The article is structured as follows. Section II following this Introduction outlines the conceptual foundations and analytical framework adopted to compare accountability and responsibility in the development finance context. The next section describes what IAMs do and why—to explain how the IFIs’ accountability system fits the responsibility as answerability model. It also discusses the importance of the ostensibly increasing emphasis on IFIs’ duty of due diligence in fulfilling their legal mandates to ‘do no harm’ and pursue sustainable development, and its implications for IAMs’ remedial function. Section IV outlines the ILC's responsibility regime and explains how it embodies the responsibility as liability model. It additionally examines how IAMs could be understood vis-à-vis certain ARIO provisions. Section V delves into the similarities and differences between the law of international responsibility and the rules and principles governing MDBs’ accountability. It elaborates some of IAMs’ advantages, including the much-lauded feature of providing access to justice to project-affected individuals and communities. It also highlights a gap in IAMs’ functions—concerning remedies—that could be filled by the ILC's responsibility regime acting residually. The Conclusion restates the core thesis: the IFIs’ accountability framework and the law of international responsibility are complements, and this relationship needs to be studied and developed further to realize the right to remedy in the development finance context.

II. ACCOUNTABILITY, RESPONSIBILITY AND COMPLEMENTARITY

Control of public power constitutes the essence and purpose of accountability.¹⁹ Attention to power imbues the concept and practice of accountability with an extra- or non-legal dimension, according to the International Law Association (ILA) and other authors.²⁰ As the ILA aptly stresses, ‘[p]ower entails accountability, that is the duty to account for its exercise’ and accountability, ‘as a matter of principle … is linked to the authority and power of an IO’.²¹ In the ILA's study, IO accountability takes different, non-mutually exclusive forms—legal, political, administrative, financial—and consists of three interrelated and mutually supportive levels: ‘internal and external scrutiny and monitoring’; ‘tortious liability for injurious consequences arising out of acts or omissions not involving a breach of any rule of international and/or institutional law’; and ‘responsibility arising out of acts or omissions which do constitute a breach’.²²

According to Brunnée, the concept of international legal accountability involves ‘the legal justification of an international actor's performance vis-à-vis others, the assessment or judgment of that performance against international legal standards, and the possible imposition of consequences if that actor fails to live up to applicable legal standards’.²³ This conception frames international legal responsibility as ‘a particular form of legal accountability, focused upon the legal consequences of breaches of international law that are attributable to an international actor’.²⁴ Simply put, international legal accountability is broader than international legal responsibility. Noting Brunnée's approach and concerned about accountability gaps resulting from ‘shifts in governance and public authority … towards different forms and levels of governance’ beyond the State,²⁵ Curtin and Nollkaemper portray ‘accountability as a set of concentric circles where, from a legal perspective, responsibility and liability may form the core, legal accountability (other than responsibility and liability) provides a second circle and non-legal (yet legally relevant) forms of accountability in more distant circles’.²⁶

The responsibility regime embodied in the ILC's works is analysed here as ‘responsibility as liability’, which is compared and contrasted with ‘responsibility as answerability’ that frames the accountability model used in, and arguably more appropriate to, the IFIs’ legal order. According to James Crawford and Jeremy Watkins, a person deemed responsible in the answerability sense is ‘called to account for their conduct and made to respond to any moral or legal charges that are put’, without the need to adjudicate whether the relevant conduct constitutes a wrong.²⁷ The liability model, on the other hand, conveys ‘the idea that a person has violated their obligations and become liable to some negative response such as punishment, censure, or enforced compensation’.²⁸ For analytical clarity, this model should be differentiated from the legal conception of sine delicto liability (as used in treaties and the ILC's other works²⁹) that resembles more closely the ILA's second level of accountability above. In its current conception, responsibility does not include liability for significant transboundary harm resulting from hazardous but lawful activities, and hence is narrower than what was originally understood by jurists before the ILC's codification efforts began.³⁰

Crawford and Watkins’ framework valuably illustrates that one model does not necessarily prevail over the other, because answerability and liability serve different purposes and operate at different stages of a legal controversy.³¹ It places the concepts along a processual continuum or spectrum. In the development finance context, IAMs act at the initial phase and enforce IFIs’ responsibility, in the answerability sense, by providing legal standing to non-State actors, who generally lack capacity to bring claims before international tribunals. Unlike the ARIO, reflecting the responsibility as liability model, IAMs do not go further to ascertain the wrongfulness of IFI conduct. However, the ARIO are unable to redress harms suffered by project-affected people, because they only consider States as injured parties.

These distinctions underscore the ‘liberating effect’ of the concept of accountability that could help international lawyers become more attuned to contemporary realities and enable them to think more broadly about, among others, the entities (beyond States) to whom account should be rendered.³² It is important to look beyond the traditional notion of responsibility and see ‘complementary relationships between … various modalities’ of accountability, according to Curtin and Nollkaemper, given the accountability gaps ‘result[ing] from shifts in power and that create the risk of abuses of power’.³³ This expansion is indeed imperative, because there are activities and relationships at the global level that create damages and losses, which a State-centric international law has yet to address meaningfully. The case of harms arising from MDB-supported projects and the concomitant demands for redress by affected non-State actors demonstrate the urgent need for the international legal system to expand its concepts and mechanisms for controlling power.

III. ACCOUNTABILITY OF MULTILATERAL DEVELOPMENT BANKS: ANSWERABILITY

IAMs exist in almost all MDBs and several national development finance institutions (DFIs).³⁴ As their names signify, the functions they perform—compliance review and dispute resolution—involve an actor being called upon to answer to another actor for an alleged harm occurring in the course of a development project. In most IAM proceedings, the bank management or staff, as ‘respondent’, have to explain their acts or omissions that are claimed to have caused harm to individuals or communities, ie the complainant(s).

Consistent with the ‘responsibility as answerability’ model, IFIs’ accountability framework focuses on justifying conduct towards an injured party, not on determining the wrongfulness of such conduct and imposing sanctions or penalties. Framed through Brunnée's concept of international legal accountability, IAMs satisfy two of its three elements: (i) an IFI's (international actor) legal justification of its performance vis-à-vis project-affected people (others); and (ii) an evaluation of such performance against the IFI's environmental and social policies (international legal standards). Imposition of consequences is the missing element.

A. Rationale for Establishing IAMs

Accountability mechanisms are bodies whose creation is not expressly mentioned in IFIs’ Articles of Agreement—suggesting that they were not originally considered essential or necessary to these IOs’ operations. Yet once the realities of IFI-supported projects causing the destruction of natural habitats and community displacement emerged, the MDBs’ governing bodies had to react. The origin story of the World Bank Inspection Panel, the pioneer IAM, reveals that the rationale for establishing these mechanisms is two-fold: ensuring the cost-effectiveness and economic efficiency of IFI-funded projects; and responding to growing demands from project-affected people and civil society organizations during the 1980s and 1990s for MDBs to take action and assume responsibility for harms to people and environment occurring in connection with development (typically infrastructure) projects.³⁵ In performing their functions, IAMs contribute to preventing or mitigating development projects’ adverse effects on people and environment.

The first reason is primarily a matter of internal accountability: member States, as creators and shareholders of an IFI, call upon the latter to account for the use of (their) public money in relation to disastrous projects. Parenthetically, the World Bank's majority shareholders such as the United States³⁶ also supported the Panel's creation due to domestic public pressure. With similar motivations, other IFIs followed suit. For example, the Inter-American Development Bank (IDB) established its Independent Investigation Mechanism in 1994 to increase transparency, accountability and effectiveness.³⁷

Of greater interest here is the second reason, which concerns external accountability: the demand for answers from project-affected people, who do not have any formal legal tie with the MDB. This change is crucial, as prior to IAMs being established, project-affected people were unable to even ask the IFI qua IO to provide project-related information. There existed no forum where individuals’ and local communities’ voices could be heard and where the IFI could be held to account for its conduct. It is the departure from this status quo—giving project-affected people the legal standing to initiate transnational, quasi-judicial processes—that makes these mechanisms innovative and valuable in remedying the disregard³⁸ of non-State actors in the international legal system. Indeed, the configuration of IAMs is unorthodox and state-of-the-art for institutionally recognizing a relationship—material and legally relevant, albeit non-contractual—between an IO and the individuals affected by the former's actions and decisions.³⁹

These changes are important, because accountability entails creating or recognizing a relationship. Accountability is ‘a moral or institutional relation in which one agent … is accorded entitlements to question, direct, sanction or constrain the actions of another – particularly where these actions involve the exercise of public power or authority within a governance system’.⁴⁰ It is relational: an actor (‘accounter’) being called to account for their conduct is doing so by responding to another actor (‘account-holder’).⁴¹ More aligned with the premise of the ILA's study is the terminology used by Ruth Grant and Robert Keohane, who refer to the accounter as the ‘power-wielder’.⁴² Here, the IFI is the power-wielder. The establishment of IAMs increased the number of account-holders. In addition to their member States, IFIs are now also accountable to select non-State actors.

B. Safeguard Systems

To appreciate better how IAMs promote accountability and sustainability,⁴³ it is important to situate them within IFIs’ ‘safeguard systems’. Safeguard systems comprise two components: (i) the IFI's policies on social and environmental matters that are addressed to management and staff; and (ii) the accountability mechanisms that interpret and enforce the safeguards.

1. Environmental and social policies

The standards against which the Inspection Panel evaluates the conduct of World Bank management and staff are the safeguard policies, previously known as operational policies and procedures (OPPs). One curious feature of the OPPs is that, while they are considered internal law or organizational rules that legally bind only the Bank and its agents, they also prescribe certain actions to the borrowing State. These requirements become international legal obligations through the loan agreement (between the borrower and the IFI),⁴⁴ which incorporates the OPPs by reference.

In 2018, the World Bank transformed the OPPs into the Environmental and Social Framework (ESF) with three components: (i) the World Bank's Vision for Sustainable Development; (ii) the World Bank Environmental and Social Policy for Investment Project Financing; and (iii) ten Borrower Requirements or Environmental and Social Standards (ESSs). The ESF's overall purpose is to translate the Bank's aspirations regarding environmental and social sustainability ‘into practical, project-level applications within the context of the Bank's mandate’.⁴⁵ With this goal, the Bank envisions moving beyond ‘do no harm’ towards maximizing development gains.

In other MDBs, OPPs are alternatively called ‘environmental and social policies’ or ‘safeguards’. As explained below, these policies are vital to determining what IFIs are being held accountable for when the IAM's functions are triggered. This section briefly describes the three policies that IAMs often address.

a) Environmental assessment

Key to operationalizing sustainable development is an environmental impact assessment (EIA). Most IFIs now refer to this requirement—more accurately—as an environmental and social assessment (ESA).⁴⁶ Although the Asian Development Bank (ADB) has yet to update its terminology, it formulates its EIA requirement to include the identification of:

potential direct, indirect, cumulative, and induced impacts and risks to … socioeconomic (including impacts on livelihood through environmental media, health and safety, vulnerable groups, and gender issues), and physical cultural resources in the context of the project's area of influence.⁴⁷

To emphasize, it is the ‘client’, typically the borrowing State, who is legally tasked with conducting an ESA,⁴⁸ but it is the IFI that classifies projects according to the risks—ie high risk, substantial risk, moderate risk, or low risk—involved. Classification accounts for several factors, including:

the type, location, sensitivity, and scale of the project; the nature and magnitude of the potential environmental and social risks and impacts; and the capacity and commitment of the Borrower (including any other entity responsible for the implementation of the project) to manage the environmental and social risks and impacts in a manner consistent with the ESSs.⁴⁹

Additionally, clients of the European Bank for Reconstruction and Development (EBRD) are required to establish and maintain an environmental and social management system, even if third (private) parties carry out the project.⁵⁰

An ESA importantly requires consultation with project-affected people and/or stakeholder engagement to gather information and scope issues.⁵¹ This requirement often includes or closely interrelates with transparency and information disclosure that are essential for meaningful participation in a process, which likewise entails ‘examin[ing] project alternatives; identify[ing] ways of improving project selection, siting, planning, design and implementation … and seek[ing] opportunities to enhance the positive impacts of the project’.⁵² More specific consultation, participation and consent requirements often apply where a project involves indigenous peoples.

b) Involuntary resettlement

As development projects traditionally involved (and still involve) large-scale infrastructure and construction activities, IFIs have specific safeguards for adverse impacts of displacement. Improperly managed, inadequately implemented, or unmitigated physical and economic displacement could, and often does, give rise to severe and long-term economic, social and environmental consequences, including the dismantling of production systems, weakening or breaking down of community institutions and social networks, loss of cultural identity and potential for mutual help, and intensified competition for natural resources and other basic services that could possibly create civil unrest.⁵³ MDBs’ common approach to involuntary resettlement issues thus consists of ‘avoid[ing] and minimiz[ing] physical and/or economic displacement, while balancing environmental, social, and financial costs and benefits, paying particular attention to impacts on the poor and vulnerable’ by requiring the borrower to consider feasible alternative project designs.⁵⁴

Involuntary resettlement policies almost uniformly provide three vital requirements for the borrowing country: (i) resettlement⁵⁵ and/or livelihood restoration⁵⁶ plan; (ii) compensation or other forms of resettlement assistance; and (iii) meaningful consultation with project-affected people. Although not couched as such across all IFIs, these requirements correspond to human rights, especially of the involuntarily resettled individuals. Exceptionally, the EBRD Performance Requirement provides that its application ‘is consistent with the universal respect for, and observance of, human rights and freedoms, specifically the right to private property, the right to adequate housing and to the continuous improvement of living conditions’, citing the Universal Declaration of Human Rights and the International Covenant on Economic, Social and Cultural Rights (ICESCR).⁵⁷

The consultation requirement links all three policies discussed here. Indeed, for the World Bank, EBRD and IDB, the required consultation and participation under the involuntary resettlement policy are conducted through the process described in a separate policy for stakeholder engagement.⁵⁸

c) Indigenous peoples

Central to the indigenous peoples policies are consultation and participation—as with other project-affected people—but also and more specifically, free, prior and informed consent (FPIC). The rationale for requiring consent is the indigenous peoples’ particular vulnerability when projects involve:

(i) commercial development of the cultural resources and knowledge of Indigenous Peoples; (ii) physical displacement from traditional or customary lands; and (iii) commercial development of natural resources within customary lands under use that would impact the livelihoods or the cultural, ceremonial, or spiritual uses that define the identity and community of Indigenous Peoples.⁵⁹

As most MDBs claim, however, FPIC has no universally accepted meaning. They thus adopt slightly varying definitions, generally agreeing that FPIC does not require unanimity but is an enhanced version of meaningful consultation. For the EBRD, ‘consent refers to the collective support of affected indigenous peoples for the proposed project activities’ that is ‘established through good-faith negotiations between the client and affected indigenous peoples, at the conclusion of which the latter arrive at a decision, in accordance with their cultural traditions, customs and practices’.⁶⁰

The ADB importantly begins its indigenous peoples policy by mentioning that many States in Asia and the Pacific voted in favour of the UN Declaration on the Rights of Indigenous Peoples (UN DRIP).⁶¹ It then focuses on the problem of development projects increasingly intruding into areas that indigenous peoples ‘traditionally own, occupy, use, or view as ancestral domain’, and emphasizes the need for special efforts to engage them in the development decision-making process, so that their specific needs and aspirations are taken into account.⁶² The UN DRIP is likewise among the international conventions and instruments guiding the IDB's policy requirements.⁶³ The World Bank recognizes indigenous peoples as having a ‘vital role in sustainable development’.⁶⁴ Its objective is ‘ensur[ing] that the development process fosters full respect for the human rights, dignity, aspirations, identity, culture and natural resource-based livelihoods’⁶⁵ of this distinct, particularly vulnerable social and cultural group.

Similar to involuntary resettlement policies, borrowing States are required to (i) assess the nature and degree of the expected direct and indirect economic, social, cultural (including cultural heritage) and environmental impacts on indigenous peoples; (ii) engage the latter in meaningful consultation and ensure their informed participation in project design and implementation; (iii) obtain FPIC in certain circumstances; and prepare, in consultation with the affected indigenous peoples, a time-bound indigenous peoples (development) plan which specifies protective or mitigating measures whose scope and scale are proportionate to the project's potential risks and impacts.⁶⁶ IFIs’ clients are additionally expected to ‘explore feasible alternative project designs to avoid the relocation of indigenous peoples from their communally held traditional or customary lands’, especially because they often have close ties to these lands and ‘their forests, water, wildlife, and other natural resources’, and such ties ‘can relate to livelihoods, cultural, ceremonial, or spiritual dimensions and can define indigenous peoples’ identities and communities’.⁶⁷

2. Functions of IAMs

All IFIs have IAMs serving dual functions: compliance review (investigation) and dispute resolution (problem-solving). Together these functions help build ‘a credible and responsive structure to ensure that projects are environmentally and socially sound and enhance [the IFIs’] contribution to sustainable development’.⁶⁸

Some IAMs likewise have an advisory function. Inspired by, and improving upon, the World Bank Inspection Panel,⁶⁹ the Compliance Advisor/Ombudsman (CAO) was established in 1999 to serve as the IAM for the private sector-oriented members of the World Bank Group, namely, the International Finance Corporation (IFC) and the Multilateral Investment Guarantee Agency (MIGA). The CAO's three discrete functions are indicated by its very name: compliance review or audit, dispute resolution or problem solving (ombudsman), and advice-giving.⁷⁰

a) Compliance review

An IAM's principal function is to receive and investigate complaints from individuals or communities claiming that they were harmed, or threatened to be harmed, by an IFI's violation of its safeguard policies. This mandate and the attendant requirement for IFI management to respond to the compliance review report fittingly demonstrate the axiom of speaking truth to power and highlight how power-wielders can be compelled to justify their conduct.⁷¹

As a fact-finding exercise,⁷² compliance review proceedings follow a general pattern: IAM reports enumerate the applicable and/or invoked safeguard policies; present the respective views of complainants and IFI management; set out the relevant facts, including those gathered from site visits; ascertain compliance; and make recommendations, as necessary and if so authorized, to rectify any violation of the environmental and social policies.

A typical allegation in a request for investigation or inspection involves a failure to conduct an EIA that resulted in the project area covering a protected natural habitat that should have been excluded. Complainants can also claim that an EIA is insufficient, for ignoring certain areas that are also likely to be impacted by the project. In a case brought before the Inspection Panel, for instance, Chadian residents alleged, among others, that the EIA (for a World Bank-funded transboundary pipeline project) defined the project area too restrictively to the oil production and transmission zones, thereby excluding testing and exploration sites in other regions.⁷³ The Panel affirmed this allegation: ‘The Project's spatial dimensions have to be explicitly defined to embrace all areas that will experience significant impacts from the Project’.⁷⁴ In its Action Plan, Management disagreed that it was noncompliant with the EIA policy, because the latter ‘does not “require” Regional EA’.⁷⁵ It nevertheless proposed to ‘intensify its efforts to work with the relevant Chadian agencies to prepare the [Regional Development Plan] … to address spatial issues and satisfy the objectives of Regional EA’.⁷⁶

It is noteworthy that the Management Response, which is submitted following the request for investigation and serves as the power-wielder's rendering of account, essentially reads like a defence. This document typically includes a rebuttal of the IAM's factual findings, especially of causality. Management's justification often states that it adequately supervised the stakeholder consultation conducted by the borrowing State; or that it exercised due diligence in ensuring that the borrower undertook the proper EIA; or that ‘the serious failures that may exist are exclusively attributable to the borrower or to other factors external to the Bank’.⁷⁷ Thus, while IAMs stop short of determining whether an IFI engaged in wrongful conduct, the management and staff are still required under the IO's rules to answer the project-affected people's complaint and explain their conduct based on what the safeguard policies provide. Significantly, there may be instances where the causality analysis is not disputed, but rather the Management Action Plan indicates concrete steps to correct noncompliant conduct. These corrective measures could include, for instance, telling the borrower to conduct another or a supplemental EIA or to hold another consultation session that includes groups who were previously not given an opportunity to participate.

Despite its avowed objective of giving a voice to project-affected people, however, compliance review tends to be inward-looking in practice. For instance, the CAO's objective in investigating alleged noncompliance with IFC and MIGA policies is to ‘improv[e] IFC/MIGA environmental and social performance’⁷⁸ by holding ‘discussions with IFC/MIGA team working with the specific project and other stakeholders to understand which criteria IFC/MIGA used to assure itself/themselves of the performance of the project’.⁷⁹ Other IAMs similarly perceive compliance review as primarily serving supervisory purposes. Moreover, although it is the project-affected people who trigger compliance review proceedings—ie they are given the opportunity to call out the possible errors of IFI management and staff—a closer examination of the ensuing procedures depicts a different set of account-holders. For example, IAM reports are addressed to the IFI's governing body, the Board of Executive Directors, or the Bank President. Likewise, the Management's Response containing its explanation and/or proposed action plan is directed to the Board, not to the complainants.

Yet, the subject matter of IAM proceedings involves, if not revolves around, the harm or threat thereof suffered by the project-affected people due to management or staff failure to comply with the IFI's policies and procedures. It thus seems logical and fair to expect that the explanation is owed to them as much as—if not more than—to the Board. That this legitimate expectation is unfulfilled makes the accountability obtained through the Inspection Panel and similar mechanisms indirect and incomplete. In this regard, IAMs hardly deviate from the traditional framework, as portrayed by Brown Weiss, wherein IOs remain directly ‘accountable to the Governments that established and that fund them’ and are merely ‘accountable at least indirectly to those that are affected by what they do, be they affected peoples or communities, the private sector, or non-governmental bodies’.⁸⁰

b) Dispute resolution

Dispute resolution involves various procedures, such as mediation, conciliation, consultative dialogue, information sharing and fact-finding. Also called problem-solving, this function complements the investigation by pragmatically addressing project-affected people's grievances and ‘improving environmental and social outcomes on the ground’,⁸¹ without focusing on the issue of IFI compliance with its policies and procedures.⁸² Simply put, the problem-solving function is flexible, outcome-driven, and does not entail identifying and allocating blame.⁸³

To ‘empower project-affected people … rather than just letting them be recipients of the [investigation] results’,⁸⁴ the ADB pioneered the problem-solving function—performed by the Special Project Facilitator (SPF)—to complement the inspection function of its accountability mechanism. The SPF is mandated to ‘engage with all relevant parties, including the complainants, the borrower, the ADB Board member representing the country concerned, Management, and staff to gain a thorough understanding of the issues to be examined’.⁸⁵ Concretely, it has handled a grievance about inadequate compensation to owners whose shops were removed due to an ADB-supported road improvement project. Its first step was to organize several meetings with complainants, borrowing country officials, and the ADB operations department about possible solutions. The SPF also helped complainants put together the requisite documentation to have their properties evaluated. The borrower's transport ministry then paid the compensation, and the SPF even ‘visited to confirm the payment and the satisfaction of the complainants’.⁸⁶

Significantly, the ADB's ‘dual approach [or bifurcated model] had a ripple effect with its adoption by EBRD, JBIC [Japan Bank for International Cooperation], AfDB, followed by EIB [European Investment Bank], and (in 2010) by IADB [Inter-American Development Bank] when it overhauled its previous mechanism focused on investigation with this combined approach’.⁸⁷ The CAO's Ombudsman or Dispute Resolution role also performs this function.⁸⁸

The flexibility and consensus-oriented character of the problem-solving function could be as vicious as it is virtuous, because it usually involves ‘difficult compromises about what can be achieved and may result in significant harms being left unaddressed’—a situation that the OHCHR finds objectionable, because it undermines the inalienable nature of human rights.⁸⁹ These features obscure the unequal bargaining strengths—a function of ‘technical, legal, financial, or fiscal resources and capacities’⁹⁰—of the parties to the dispute resolution process. In this respect, the following provision in the African Development Bank (AfDB) Independent Recourse Mechanism (IRM) Operating Rules and Procedures is noteworthy: ‘The Problem-Solving exercise shall be particularly sensitive to the existence of power asymmetries between the negotiating parties, particularly concerning the availability of information needed and the capacity of the parties to participate effectively in these processes.’⁹¹

Using recent World Bank reforms, the next sub-section further explains both compliance review and dispute resolution functions.

c) New World Bank Accountability Mechanism

Before the World Bank Accountability Mechanism's creation in 2020, the Inspection Panel served as the sole IAM for the IBRD and the International Development Association (IDA). While its ‘progeny’ in other IFIs have assumed both investigation and problem-solving functions, the Panel has only performed the former since 1993. With the new Accountability Mechanism, the Bank decided to create a separate unit to perform the latter function, instead of adding it to the Inspection Panel's authority. As expressly provided in the relevant Board resolutions, ‘The Inspection Panel will have no role in dispute resolution and will not opine on policy compliance in dispute resolution or the outcome of the dispute resolution process.’⁹²

There have thus been minimal changes to the Panel's mandate and operating procedures.⁹³ It continues to have jurisdiction to receive and investigate complaints wherein individuals and communities allege that their ‘rights and interests have been or are likely to be directly affected by an action or omission of the Bank as a result of a failure of the Bank to follow its operational policies and procedures’ and ‘that such failure has had, or threatens to have, a material adverse effect’.⁹⁴ Additionally, its investigation remains limited to Bank management and staff conduct. Hence, the Panel report can only discuss ‘those material adverse effects, alleged in the request, that have totally or partially resulted from serious Bank failure of compliance with its policies or procedures’.⁹⁵ The Panel's authority to comment on the Management Report and Recommendation likewise remains restricted to submitting its view, for the Executive Directors’ consideration, on ‘the adequacy of consultations with affected parties in the preparation of the management action plan’.⁹⁶

The expressly stated purpose of the World Bank Dispute Resolution Service is ‘to facilitate a voluntary and independent dispute resolution option for Requesters [ie project-affected people] and borrowers (the “Parties”) in the context of Inspection Panel Requests for Inspection’⁹⁷ and assist the Parties in reaching a mutually agreed solution.⁹⁸ The jurisdiction ratione materiae of the mediator jointly selected by the Parties to deliver dispute resolution services is the same as that of the Inspection Panel, namely, ‘project-related issues raised in the Request for Inspection and identified as the issues to be investigated in the Inspection Panel's report to the Executive Directors recommending investigation’.⁹⁹ Consistent with the optional and consensual nature of this function, any of the Parties can withdraw at any time to conclude the dispute resolution process.¹⁰⁰

Thus the only major change effected by the new Accountability Mechanism is that Inspection Panel proceedings could be temporarily suspended, should the project-affected people and the borrowing State mutually agree to pursue dispute resolution.¹⁰¹ Moreover, a possible outcome of the dispute resolution is preventing the Panel from taking further action on the request for investigation if ‘the Parties have reached agreement and signed a Dispute Resolution Agreement’, since the case shall then already be considered closed.¹⁰² Recalling the asymmetries among the Bank, the borrower and the project-affected people, this ‘solution’ is problematic, as it creates the unjust and objectionable possibility that Parties could ‘lawfully waive – by agreement – violations of Bank policies and procedures, regardless of the legal consequences of the material adverse effects of these violations on the Requesters’.¹⁰³

C. Sustainable Development and the Right to Remedy

The emergence of accountability as an institutional priority is contemporaneous and intimately connected to IFIs’ incremental recognition of sustainable development as part of their operations, if not the core of their legal mandates as IOs. Several authors have analysed the evolution of the constituent instruments of the World Bank and other MDBs that resulted in their formulation of the safeguard policies.¹⁰⁴ Such accounts will not be repeated here, but it is important to contextualize the creation of IAMs in these substantive changes to MDBs’ internal law. The AfDB fittingly describes the IRM's purpose by referring to its role in ensuring compliance with the bank's ‘policies and procedures related to sustainability’.¹⁰⁵ Moreover, through their interpretive functions, IAMs clarify, elaborate and operationalize the concept of sustainable development, which lies at the core of IFIs’ legal or constitutional mandates.¹⁰⁶

1. Do no harm

Most MDBs’ financing activities and sustainability efforts implement a ‘do no harm’ principle. The IDB, for instance, relates its implementation to establishing ‘clear provisions for effectively managing project-related environmental and social risks and impacts, and whenever feasible, facilitating the enhancement of social and environmental sustainability beyond the mitigation of adverse risks and impacts’.¹⁰⁷

IFIs construe and implement ‘do no harm’ as a due diligence obligation—consistent with how some tribunals¹⁰⁸ and scholars¹⁰⁹ interpret the harm prevention rule. Their duty of prevention covers harms to both people and environment. This coverage contrasts with the harm prevention rule in international environmental law and is closer to the older Trail Smelter ‘no harm’ rule that focused on State responsibility for transboundary harm.¹¹⁰ It aligns with the ILC's definition of harm as one ‘caused to persons, property or the environment’ in relation to hazardous activities.¹¹¹

While IFIs still vehemently deny bearing human rights obligations, their interpretation of the harm prevention rule leaves space to argue that they have at least an obligation to respect human rights. Significantly, IFIs are enjoined under the ICESCR to pay greater attention to protecting human rights in their financing activities, including the policies they prescribe or require as condition for funding.¹¹² One study succinctly explains that due diligence operationalizes ‘do no harm’, and as applied to potential human rights infringements, IOs ‘have the ex ante duty to identify and mitigate potential human rights risks exacerbated by their operations, as well as the ongoing duty to monitor and mitigate newly emerging human rights violations’.¹¹³ Additionally, as the World Bank is considered a UN specialized agency (pursuant to a Relationship Agreement between these IOs), it is bound by the UN Charter, including its human rights-related provisions. In this regard, the Bank's obligation to respect the Charter involves undertaking due diligence tasks.

2. Due diligence

The ILC explains that the harm prevention duty is one of due diligence, rather than a guarantee of total prevention, such that a State must exert ‘best possible efforts to minimize the risk’ or undertake ‘all appropriate measures to prevent significant transboundary harm’.¹¹⁴ Rather than a free-standing obligation, due diligence qualifies behaviour and describes the duty of care ‘needed when a risk has to be controlled … to prevent harm and damage done to another actor or to a public interest’.¹¹⁵ From an international project finance perspective, due diligence is a set of practices and methods aimed at identifying and assessing risks. As Diane Desierto expounds, ‘[t]he scope of the risk assessment is necessarily extensive as to subject-matter, multidimensional as to tools and methods used for risk analysis, interrelated in the approach to identifying individual and collective environmental and social risks, and continuing as part of the Bank Management's project oversight’.¹¹⁶

Based on the MDBs’ formulation, ‘do no harm’ is the underlying substantive obligation that the procedural duty of due diligence operationalizes. The World Bank now explicitly acknowledges its own duty of due diligence, ‘proportionate to the nature and potential significance of the environmental and social risks and impacts related to the project’, in order to implement the Environmental and Social Policy (ESP).¹¹⁷ An entire sub-section in the ESP details mandatory requirements for Bank management and staff in carrying out environmental and social due diligence, which is intended to ‘assist the Bank in deciding whether to provide support for the proposed project and, if so, the way in which environmental and social risks and impacts will be addressed in the assessment, development and implementation of the project’.¹¹⁸ Most relevant here is this specification:

32. The Bank's due diligence responsibilities will include …: (a) reviewing the information provided by the Borrower relating to the environmental and social risks and impacts of the project, and requesting additional and relevant information where there are gaps that prevent the Bank from completing its due diligence; and (b) providing guidance to assist the Borrower in developing appropriate measures consistent with the mitigation hierarchy to address environmental and social risks and impacts in accordance with the ESSs. The Borrower is responsible for ensuring that all relevant information is provided to the Bank so that the Bank can fulfil its responsibility to undertake environmental and social due diligence in accordance with this Policy.¹¹⁹

As early as 2009, the ADB similarly recognized its duty under the safeguard policies to ‘help borrowers/clients meet [policy] requirements during project processing and implementation through capacity-building programs, ensure due diligence and review, and provide monitoring and supervision’.¹²⁰

Apart from IFIs’ own declaration of their ‘do no harm’ and due diligence commitments, the harm prevention rule is custom that binds IOs as subjects of international law. Although the State arguably bears the primary international legal obligations towards environmental and human rights protection during development projects, IFIs qua IOs cannot absolve themselves of the customary duty under international law to take diligent steps to prevent harm. Moreover, given the material and power inequalities between MDBs and many borrowing States, it is reasonable to expect the former to be more capable of ensuring that development projects do not cause environmental and social harms and that remedies are provided should such harms nevertheless occur.

3. Right to remedy

Remedy is highly important in the development finance context for several reasons, foremost of which are IFIs’ legal mandates to pursue sustainable development and ‘do no harm’ and the necessity to meet ‘evolving social expectations and relevant normative developments’.¹²¹ Among such developments are the elaboration of the right to remedy into procedural (effective access) and substantive components, its crystallization as a customary rule, and the implication that IOs are legally bound to go beyond the first level of accountability.¹²² As the OHCHR explains, ‘remedy is the functional corollary of the “do no harm” mandates of DFIs, going to the heart of their missions’,¹²³ and IAMs aid DFIs in performing such missions by ‘support[ing] the voice, empowerment and participation rights of people directly affected by projects [and] bringing inputs, knowledge and feedback loops that may not otherwise be available, to the benefit of equity and sustainability’.¹²⁴ However, as one commentator critically notes, ‘[t]he IAMs’ general lack of consequential authority is a major shortcoming of the IFI accountability systems’.¹²⁵ A notable exception is the AfDB IRM, which is expressly authorized to recommend, upon finding noncompliance, ‘[t]hat redress be provided to those harmed, which may include financial and/or non-financial considerations’.¹²⁶ Whether other IFIs would follow the AfDB's example remains subject to conjecture.

It is interesting to see how the World Bank views its sustainable development mandate and relationship with international law:

[T]he World Bank's activities support the realization of human rights expressed in the Universal Declaration of Human Rights. Through the projects it finances, and in a manner consistent with its Articles of Agreement, the World Bank seeks to avoid adverse impacts and will continue to support its member countries as they strive to progressively achieve their human rights commitments.¹²⁷

This statement represents the long-standing official position of the World Bank, as well as other IFIs, that it does not bear international human rights obligations but simply acts in an assistive or supportive capacity vis-à-vis member States, who are the ones legally bound to respect, protect and fulfil the human rights implicated in development projects.¹²⁸ A few observations and recommendations by treaty bodies, as well as academic commentaries, convincingly refute this claim.¹²⁹

IAMs’ contributions to the overall improvement of project performance and, possibly, to decreasing the likelihood of future environmental and social harms in development projects, should not be ignored. It nonetheless bears emphasis that they still fail to adequately and appropriately satisfy the project-affected people's right to remedy, especially concrete reparation.¹³⁰ This deficiency could be attributed to the fact that, except in a few existing accountability mechanisms, the involvement of project-affected people does not go further than initiating a complaint. Complainants are excluded from subsequent processes and are not even given an opportunity to provide inputs to the management action plan, which is meant to correct the project failures or violations that have caused social or environmental harm. Additionally, the considerable discretion of the IFI—mainly the Management and, to a lesser extent, the Board of Executive Directors—regarding the appropriate actions upon a finding of noncompliance is problematic. According to Hunter, enabling the Bank to exercise greater discretion and professional judgment exacerbates the glaring gaps in the standards, ‘most notably the lack of a clear commitment to ensure Bank projects do not contribute to violations of human rights’.¹³¹

An examination of the management action plans formulated by different MDBs after a compliance review shows that the actions are corrective but not reparative. Put differently, the outcome of establishing harm-causing noncompliant behaviour is not to wipe out the adverse consequences to project-affected people. Instead, the actions usually taken by IFI management are forward-looking and geared towards improving project implementation in subsequent phases. Notably, this procedure is consistent with the ILA's recommended rules and practices about reporting and evaluation, specifically that reports ‘should not only contain a genuine account of action or inaction but also explanations for the course of conduct’ and past operations should not perfunctorily be used ‘as a model for future operational activities’.¹³²

IV. RESPONSIBILITY OF INTERNATIONAL ORGANIZATIONS: LIABILITY

States are interested in controlling and ensuring IOs’ performance of their functions in accordance with international law.¹³³ A step in this direction, the ARIO attempt to codify and progressively develop rules that would prevent IOs from becoming Frankenstein's monster,¹³⁴ inflicting harm on people and the environment. Its text and principles derive from the ARSIWA. The ILC acknowledges, however, that IOs, as international actors, have idiosyncrasies that distinguish them from States. The difficulties in drafting the secondary rules stem in large part from ‘the limited availability of pertinent practice’, the unavailability or inadequate use of third-party dispute settlement procedures that cover IOs, and the ‘great diversity among [them]’.¹³⁵

In contrast to IFIs’ accountability system, the ILC responsibility regime is objective and non-relational. This characterization means that determining the existence of international responsibility requires looking only at an international actor and their conduct, without regard to the effect of the latter or to any relationship of said actor to another entity, international or otherwise. What distinguishes the ILC framework, conceptualized here in the responsibility as liability sense, from the IFI accountability regime (responsibility as answerability) is its focus on identifying the violation of a legal obligation and specifying the consequences of such violation for the responsible actor. According to Crawford and Watkins, the principles behind this conceptualization ‘shape the judicial response to international lawbreaking and the legal obligations which responsible parties thereby acquire’.¹³⁶ This emphasis can be observed in the overall structure of the ARIO, wherein the part dealing with the Content of the International Responsibility immediately follows the part that tackles The Internationally Wrongful Act itself.

A. Definition and Consequences of International Responsibility

At the crux of the law of international responsibility is an internationally wrongful act. International responsibility requires two elements: (i) an act or omission attributable to an actor, either a State or an IO; that (ii) constitutes a breach of an international obligation of that actor.¹³⁷ The notions of ‘injury’ and an ‘injured State’, become relevant only at a later phase, namely, in invoking responsibility and implementing the legal consequences of an internationally wrongful act.

The ARIO share the premises underlying the ARSIWA's provisions on reparation. Legal consequences of an internationally wrongful act serve at least two purposes: (i) maintaining or restoring the rule of law in the interest of the international community as a whole (‘legality function’); and (ii) making whole the parties injured by the breach of an obligation (‘remedial function’).¹³⁸ Although not entirely clashing, these purposes and the provisions that operationalize them sometimes stand in tension, to the detriment of affording full reparation to injured parties.¹³⁹ For instance, the ILC's choice to restrict compensation to ‘financially assessable damage’ is intended to ensure that the ARIO do not punish the responsible State, instead only compelling it to restore or uphold the international rule of law. This solution, however, could be inadequate for an injured party, who might additionally be interested in imposing punitive damages to deter future violations. The standard of ‘financially assessable damage’ also poses difficulties in establishing non-pecuniary environmental harms, such as air or marine pollution and loss of biodiversity.¹⁴⁰

The ARIO are limited to enforcing and protecting rights and obligations of States and IOs. Despite the ‘communitarian turn’ in the international responsibility regime,¹⁴¹ it appears that the international community, as understood in the ARIO, excludes individuals and other non-State actors. The latter's conduct only becomes legally relevant if it is attributable to a State or an IO.¹⁴² Conversely, responsibility for human rights violations and breach of other treaties, wherein the ultimate rights-bearers are individuals, can only be implemented by States and IOs.¹⁴³

Whether IOs have international obligations owed to non-State actors—as could reasonably be argued in the case of MDBs vis-à-vis project-affected people—is a question that the ILC barely tackled. However, it noted, albeit still quite vaguely, that the wording of Article 10(2) of the ARIO ‘is not intended to exclude the possibility that other rules of the organization may form part of international law’,¹⁴⁴ meaning, some parts of IFIs’ internal law could give rise to international legal obligations. The ‘institutional theory of law’ more readily admits such possibility and the plurality of legal phenomena, given the shift of this school of thought's inquiry from how the legal system legitimizes legal norms, to how certain results of human activity and social reality ‘can obtain legal validity as elements of the legal system’.¹⁴⁵ As Dekker aptly concludes, although ‘principles, inducements and purposes’ may not amount to legal obligations, they are ‘institutional legal realities’ that create ‘different expectations and different patterns of legitimization of conduct’, and therefore IOs can be held accountable for noncompliance with ‘a legally valid – but not necessarily also legally binding – rule of international law’.¹⁴⁶ Relatedly, Brunnée explains that the rise of alternative modes of accountability reflects the conceptual inadequacy of international responsibility and ‘the fact that the traditional conception of international law as a set of inter-state rules flowing from certain formal sources is no longer a sufficient account of international legal relations’.¹⁴⁷

B. Provisions Most Salient to IFIs

As with most works entailing generalizations to cover a broad range of situations and an equally broad and diverse group of actors, the ARIO do not consider the specific circumstances of any IO. Judging, however, from the comments submitted to the Commission by the World Bank and other IFIs, certain provisions are particularly salient to this group of international economic organizations. The relationship between the ARIO and IAMs can be analysed along three issue-areas: (i) lex specialis character of IO rules; (ii) IO rules’ legal nature (international versus internal); and (iii) interaction between State and IO responsibility.

1. Lex specialis

For the World Bank, like many IOs, Article 64 on Lex specialis is the ARIO's key provision.¹⁴⁸ Apart from strongly encouraging the ILC to emphasize more clearly the ‘centrality of lex specialis and the residual character’ of the ARIO, it insists that the internal law of the organization should prevail over all international obligations, save for jus cogens norms.¹⁴⁹ A related provision is Article 32 of the ARIO, which precludes an IO's invocation of its rules to justify noncompliance with obligations in Part Three, ie the legal consequences of international responsibility including, among others, the reparation provisions. Notably, while an IO and its member States may agree on special rules concerning remedies between them, such agreement cannot affect the rights and interests of third parties, including individuals injured by the IO's activities.¹⁵⁰

Application of the lex specialis principle contemplates two possible relationships: one wherein the general standard and the specific rule point in the same direction, but with the latter elaborating the former; and another wherein the two have ‘no express hierarchical relationship, and provide incompatible direction’ about tackling the same set of facts.¹⁵¹ As applied here, the issue rests on two questions. Does the legal framework governing IAMs cover all topics that the ARIO address, such that the former can displace the latter? Is there a normative conflict between the regimes?

Focusing on the first question, Baimu and Panou answer in the affirmative, highlighting the peculiarity of a given IO and the greater specificity of its rules.¹⁵² They fail, however, to satisfy ‘thresholds of specificity and genuine inconsistency’¹⁵³ by, for instance, explaining how IAMs’ applicable rules or the safeguard policies elaborate ARIO provisions. In contrast, Palchetti disagrees that IFIs’ accountability system constitutes lex specialis, arguing that the two regimes are complements rather than competitors, with the ARIO providing external control and IAMs operating ‘“from within” by imposing procedural limitations on the way in which the organization exercises its functions’.¹⁵⁴

The comparative analysis below yields a similar conclusion, showing that the law of international responsibility and the IFI accountability framework serve the same broad objective but address different account-holders and situations. While IAMs address harms caused by IFIs’ noncompliance with safeguard policies, they do not adjudicate on the legality of such noncompliance. The ARIO tackle IFIs’ breach of international obligations and the legal consequences thereof, but only consider, as a secondary matter, injury to States. Although not covering the same subject-matter, they are not conflicting in their rules and operation. Therefore, IAMs cannot entirely replace the ARIO (and vice versa) and the two can apply simultaneously and complementarily to control IFIs’ power better.

2. Characterization of ‘rules of the organization’

Whether IO rules¹⁵⁵ are internal or international law is addressed, albeit inconclusively, in the second paragraph of Article 10 of the ARIO, stating that the breach of an international obligation by an IO ‘may arise … under the rules of the organization’. The World Bank's view on this matter is worth quoting in full:

Precisely because the World Bank agrees with the Commission in its cautious approach to [the controversial question about the legal nature of IO rules], it thinks that the clarification provided in the commentary would be better reflected by the deletion of paragraph 2 from draft article 9. In fact, paragraph 1 already states that a breach is a breach regardless of the origin and character of an obligation binding an international organization, thus clearly implying that this origin may also be in the rules of the organization. On the contrary, retaining paragraph 2 may wrongly lead to the unsubstantiated conclusion (expressly denied in the Commission's commentary) that the breach of any rule of the organization is necessarily a breach of an international obligation.¹⁵⁶

This statement reflects the oft-criticized position of almost all IFIs that they are not bound by obligations under international human rights law and international environmental law.¹⁵⁷ It contradicts their acts of formulating environmental and social policies and creating IAMs. Whether IFIs’ safeguard policies create international obligations links to whether noncompliance with them, as determined by IAMs, amounts to international responsibility. The ILC itself does not claim to have a definitive answer.¹⁵⁸ It nonetheless points out that not all breaches of obligations under the rules of the organization are necessarily breaches of obligations under international law.¹⁵⁹ It is possible for an IAM's findings of noncompliance to implicate also an IFI's breach of international law, since some safeguard policies overlap with treaty or customary rules.¹⁶⁰ Moreover, if international lawmaking is viewed as a communicative process among relevant actors, IFIs qua IOs are considered participants in such process,¹⁶¹ and their environmental and social policies could contribute to international law.¹⁶²

This issue emphasizes the outstanding need to develop the primary rules further, specifically concerning harm prevention and due diligence, applicable to IFIs when they assist States and fund development projects. Clarifying and elaborating the primary rules that bind IOs are critical not only in ‘enhanc[ing] the predictability of international organisations’ action towards the different entities of their constituencies’, but also ‘provid[ing] the necessary preconditions for any serious attempt at remedial action against them’.¹⁶³

3. Responsibility of multiple actors

Other provisions likely applicable to IFIs are those on aid or assistance (Article 14 of the ARIO) and direction and control (Article 15 of the ARIO). Regarding the former, the World Bank enjoins the ILC to reiterate, for clarity and emphasis, its commentary to Article 16 of the ARSIWA in the commentary to the ARIO that ‘organizations providing financial assistance do not, as a rule, assume the risk that assistance will be used to carry out an international wrong’.¹⁶⁴ Such an interpretation is critical, the Bank claims, to alleviate the ‘dangerous chilling effect [that Article 14 of the ARIO may create] for any [IFI] providing economic assistance to eligible borrowers and recipients’.¹⁶⁵ Relatedly, the Bank seeks clarification about the knowledge threshold for these two provisions and pushes for setting it to actual, rather than presumed, knowledge. It details the distinction between ‘oversight’ and ‘direction and control’, arguing that in an agreement between the IFI and the borrowing State, ‘direction and control for the implementation of a project or programme activities are never really ceded, because the responsibility for implementation remains with the borrower or recipient, while the [IFI] engages at most in the exercise of oversight’.¹⁶⁶

The ILC seems to encourage States and IOs to delineate clearly and allocate expressly their respective competences and responsibilities when they undertake joint projects or operations. Some authors suggest that ex ante clarification and apportionment of responsibilities could be relevant to inter-State relations and particularly useful in managing litigation among them.¹⁶⁷ Discussing the model contribution agreement between the UN and the States contributing military contingents for peacekeeping operations, the Commission notes, however, that such agreement only binds the international actors parties thereto and ‘could … not [deprive] a third party of any right that this party may have towards the State or organization that is responsible under the general rules’.¹⁶⁸

Viewed in this light, one can construe IFIs’ safeguard policies as heeding the ILC's advice, since they distinguish between the policy applicable to Bank management and staff and the standards for the borrowing State to meet. The apportionment of tasks between the Bank and the borrowing State is illustrated by the provisions on the formulation of an Environmental and Social Commitment Plan (ESCP), which forms part of the legal agreement between the Bank and the Borrower and lays down ‘the material measures and actions required for the project to meet the [environmental and social standards] over a specified timeframe’.¹⁶⁹ Remarkably, the implementation of the measures and actions specified in the ESCP is the Borrower's obligation, and the Bank only undertakes to monitor the environmental and social performance of the project in accordance with the ESCP. These decisions, however, prompt inquiry whether such distribution of duties aligns with the power and capabilities of the actors concerned to prevent and to remedy any resulting harms.

V. INDEPENDENT ACCOUNTABILITY MECHANISMS AND THE ARIO: CONVERGENCE AND DIVERGENCE

International legal accountability, in principle, encompasses both the responsibility as answerability and the responsibility as liability models. It can thus be hypothesized that the rules governing IAMs and the law of international responsibility bear similarities but retain certain distinctions. This part specifies ways that these two regimes converge and diverge and examines their implications, particularly for remedying harms to people and the environment that result from IFI-supported development projects.

The comparative analysis uses as reference points some questions that Edith Brown Weiss¹⁷⁰ associates with accountability: (i) for what types of conduct is an account due; (ii) to whom the accounter must give account; (iii) what consequences follow from being held accountable or responsible; and (iv) whether and how institutional learning takes place.

A. Noncompliance or Breach

Both the ARIO and IAMs are concerned with conduct that deviates from a particular norm. Under the former, one of two elements of an internationally wrongful act is the breach of an international obligation. Under the latter, among the requisites for triggering an IAM's jurisdiction is the MDB's noncompliance with its safeguard policies. Both thus have a similar object of inquiry, ie deviant behaviour, but arguably different standards to evaluate such conduct. In an IAM's case, the obligation stems from the IFI's safeguard policies, whereas the ARIO's concern an obligation based on treaty, custom or general principles of law.

B. Harm or Injury

A crucial distinction between the two frameworks concerns the role of harm or injury. Given the ILC's delineation between primary and secondary rules, harm is not a constitutive element of international responsibility. Instead, the content of the primary rule determines whether this element is required.¹⁷¹ In both the ARSIWA and ARIO, injury is discussed in the part entitled ‘Content of the International Responsibility’ that is separate from the part concerning ‘The Internationally Wrongful Act’. More specifically, injury is described in the provision defining the consequence or obligation of reparation: ‘Injury includes any damage, whether material or moral, caused by the internationally wrongful act of an international organization.’¹⁷² Likewise, the notion of an ‘injured State’ exists in the ILC regime for the purposes of implementing international responsibility and identifying which actor can do what (eg invocation; non-recognition; countermeasures) in response to the internationally wrongful act.

In contrast, harm is an eligibility requirement for persons bringing a case before IAMs. It gives individuals the legal standing to trigger the IAM's exercise of its jurisdiction, which entails the scrutiny of the IFI's conduct vis-à-vis its policies and procedures. The IAM's fact-finding exercise largely involves determining the cause(s) of harms that project-affected people allegedly suffered or will probably suffer. More particularly, it ascertains whether such harms result from a failure by management and staff to comply with the IFI's safeguard policies. The causation analysis, however, does not include a determination or judgment that the noncompliance constitutes a legal wrong—a feature that distinguishes the IAM's tasks from that of a judicial body applying the ARIO.

C. Standing or Invocation

The ARIO do not allow non-State actors like individuals to invoke and protect their rights. They do, however, clarify that the chapter pertaining to invocation of responsibility ‘is without prejudice to the entitlement that a person or entity other than a State or an international organization may have to invoke the international responsibility of an international organization’.¹⁷³ The Commission, again reverting to the primary–secondary rule distinction, explains that the relevant primary rule will have to ‘determine whether and to what extent persons or entities other than States are entitled to invoke responsibility on their own account’.¹⁷⁴ Hence, in a human rights treaty breach, for example, while it is undisputed that the rights-bearers are the natural persons themselves, under the ARIO, the injured party who can invoke responsibility would presumably be limited to the State of nationality¹⁷⁵ of those natural persons whose rights were violated, or any other State or international organization who is a party to such treaty and the norm involved is of an erga omnes character.¹⁷⁶ The ILC nonetheless acknowledges that a primary obligation, such as a human rights treaty or a bilateral investment treaty, that accords rights to non-State actors may itself provide for a procedure that entitles the latter to invoke the responsibility on their own account.¹⁷⁷

It is this traditional State-centric view of international law that IFIs attempt, rather successfully, to overcome through the IAMs. Indeed, these mechanisms are one of the first international fora where individuals can directly submit complaints. The novelty, which the Inspection Panel and its counterparts in other MDBs has been widely praised for, is the fact that through these mechanisms, IFIs qua IOs become answerable to project-affected people, ie non-State actors who have neither international legal personality nor any formal relationship with the IFIs.

D. Remedies and Consequences

In the ILA study, the types of remedies differ according to level of accountability.¹⁷⁸ Remedies may assume a ‘positive, injunctive, corrective or compensatory nature’, and ultimately, the potential outcome of remedial action against IOs is contingent on ‘the identity of the party seeking redress, the kind of accountability involved, and the forum before which the remedial action has been brought’.¹⁷⁹ IAMs operate at the first level of accountability, which entails forms of internal and external scrutiny that could result in modifications of an IO's conduct, but without regard to potential and subsequent liability and/or responsibility, which are the second and third levels, respectively.

The consequences of an IAM's finding of noncompliant behaviour differ considerably from the consequences of an internationally wrongful act. The first difference concerns cessation and non-repetition. In the responsibility as answerability model, because no legal wrong is determined to have been committed, the accounter—ie the IFI—is not obliged to cease its conduct, even when the IAM finds that a noncompliant behaviour caused harm to project-affected people. Neither is there an obligation to offer assurances and guarantees of non-repetition. The AfDB resolution creating its IRM, however, presents a noteworthy innovation, as it attempts to be more specific about what Management is required to do after a finding of noncompliance: ‘Management shall ‘[i]dentify any changes to AfDB practices, policies, guidance or systems to bring the Bank into compliance and/or avoid recurrence of similar situations’.¹⁸⁰ Generally, however, it is discretionary on the IFI Management's part to stop the harmful activity, to perform an omitted action that would have prevented the harm, or to commit to ensuring that noncompliance will not recur in the future.

In contrast, due to the rule-of-law-restoration philosophy of the responsibility as liability model, the ARIO provide that the obligations to cease the wrongful conduct and to offer guarantees of non-repetition accrue to the responsible actor almost automatically, ie without need of invocation.¹⁸¹ Arguably, when a responsible State or IO stops its internationally wrongful conduct, the non-State actors injured by such conduct still indirectly benefit from the cessation. In the LaGrand Case,¹⁸² for example, the guarantees of non-repetition were given by the responsible State to the victims’ State of nationality. One can nonetheless claim that the ultimate beneficiaries of such assurances are the victims and the injured State's other citizens.

It bears stressing, again, that due to the international responsibility regime's traditional international law orientation, individuals and other non-State actors are not entitled to the remedies provided in the ARIO. Rather, the entities who can claim reparation are limited to States and IOs. The reference to other ‘beneficiaries of the obligation breached’ in Article 49(4)(b) of the ARIO and in Article 48(2)(b) of the ARSIWA ‘provides a means of protecting the community or collective interest at stake’,¹⁸³ as in the case of human rights treaties. The practice of States and IOs relative to this provision remains limited, however. The ARIO thus appear ill-equipped to redress the harms experienced by project-affected people. It is highly unlikely that the borrowing State would espouse their claims, considering that those harms could or would typically also be caused by the State's own acts or omissions. On an even more pragmatic level, a borrowing State might not be willing to risk the cancellation of the loan or grant and the cessation of the development project, especially if it believes that the latter would bear economic gains for the country.

E. Opportunity to Learn

Remedy, as ‘an acceptable outcome arrived at through a procedure instigated by an aggrieved party’, encompasses other kinds of redress such as ‘prospective changes of policy or practice by the IO’.¹⁸⁴ While the IFI accountability framework does not require offering assurances of non-repetition like the ARIO do, the expectation is that being held to account would become a learning opportunity. Brown Weiss, who endorses this view, proposes focusing less on sanctions and instead reconceiving accountability ‘as mutual accountability and as a dynamic process in which learning takes place in response to holding actors accountable’.¹⁸⁵ The principle of mutual accountability ‘anticipates a potentially broader set of consequences, including incentives or capacity-building and transparency, as well as penalties, sanctions, or restitution … [and] looks to a broader array of mechanisms for holding parties accountable, including non-judicial bodies’.¹⁸⁶

Several IAMs seem to be moving in this direction. According to the ADB, its ‘Staff, Management, and the Board increasingly see the Accountability Mechanism as a tool for ADB to respond positively to public scrutiny … [and] helps ADB to learn lessons and improve its project quality’.¹⁸⁷ The AfDB IRM is expressly authorized to perform an advisory function, which is triggered when the President and/or the Boards ‘feel that projects, programs, policies and procedures of the Bank Group can benefit from the accumulated experience of the IRM and support efforts of staff and Management to strengthen the social and environmental impact of the projects funded by the Bank Group’.¹⁸⁸ The advisory function is meant ‘to provide independent opinions on systemic issues, and technical advice on projects and programs of the Bank Group’.¹⁸⁹

Likewise, part of the CAO's advisory work is to ‘advanc[e] the boundaries of environmentally and/or socially responsible behaviour on the part of IFC/MIGA by advising on emerging, strategic, or systemic issues or trends or processes’.¹⁹⁰ It has wide discretion to advise based on diverse sources, including international norms concerning human rights and environmental protection.¹⁹¹ The CAO is also envisioned to advance the same behaviour ‘in the private sector, civil society, and academia through lessons derived from CAO cases’.¹⁹² In recent reports, the World Bank Inspection Panel appears also to be performing an advisory function (despite the absence of a clear and express mandate from the Bank) by adding ‘a final chapter in which it discusses “systemic issues” related to the Bank's compliance with [safeguard policies] in light of its conclusions in the case at hand’.¹⁹³

VI. CONCLUSION

The distinct components and emphases of the IFI accountability framework and the ILC responsibility regime have posed challenges about whether and how they could fit together. In some aspects, eg the role of harm and the account-holders’ identity, the former has a larger, more inclusive and progressive scope. Hence, IAMs may seem ‘better’—outsizing the law of international responsibility. However, aligning them can simply involve a straightforward adjustment in perspective. Viewing the frameworks along a spectrum representing different phases of the legal process showed that, while IFIs have focused on establishing non-State actors’ legal standing—essential to answerability—the ILC has concentrated on determining wrongfulness, which is important for liability. Using the open-ended conception of international legal accountability also helps in understanding that the identified divergence points are only differences in the regimes’ methods and standards of regulating IFIs’ power. Indeed, despite having different rationales, scope and functions, the two regimes are not normatively conflicting and they share the objective of controlling IFIs.

Two key points emerged from this study. First, IAMs improve upon the State-centric international responsibility regime, by enabling individuals to demand an explanation from IFIs, but fall short of providing remedies. The IFI accountability framework thus leaves a gap that ARIO provisions on reparation could fill given their residual character. Second, neither the international responsibility regime nor the IFI accountability framework can exclusively protect project-affected people's rights and remedy violations thereof. It is therefore suggested that IAMs should be authorized to direct IFIs to make reparation in any of its forms (restitution, compensation, satisfaction) when IFIs’ violation of their own policies has caused harm. This reform can be done without need to change the IFI accountability system's underlying logic that IAMs are not intended to establish internationally wrongful conduct. Remedies under international law do not solely arise from unlawful conduct but can also be justified by harm suffered by one party due to another's risky activities.