The principled case for humanitarian accountability is relatively straightforward. Proponents point to the power of aid agencies intervening in complex humanitarian emergencies, to their moral obligations as beneficent actors, and to their fiduciary responsibilities as bearers of public funds (e.g. HAP 2008b: 3–4; VanRooyen 2013). Especially since the Rwandan genocide, humanitarianism has experienced what practitioners have deemed an “accountability revolution,” with the sector embracing self-regulatory initiatives, reporting mechanisms, and learning networks (Knox-Clarke and Mitchell 2011); accountability has thus moved from “emergent language” to increasingly embedded practice (Stein 2008: 126).

However, the practice of accountability has proven decidedly more complicated than the principle. As Stein (2008) asks, to whom are humanitarians accountable? For what? And how? Each of these questions implicates the others: different accountability relationships—with governments, donors, beneficiaries, fellow international nongovernmental organizations (INGOs), and the public—yield different accountability demands and are ultimately informed by different accountability logics (Berghmans et al. 2017). In practice, then, humanitarian accountability is negotiated; coherence must be crafted and consensus sustained. It is, or should be, a process that gives rise to contestation.

To date, despite the wave of academic research into INGO accountability, few empirical studies have seriously grappled with the phenomenon’s inherently contested nature. In the dominant principal-agent framework, contestation appears only in a limited sense, as competition among instrumentally rational agencies pursuing their individual interests—namely survival and expansion—rather than as more fundamental contestation over the ideas and practices of accountability itself. INGOs design and implement standards to credibly signal proper practice to their principals, generally donors, and to distinguish themselves from their competitors (e.g. Gugerty and Prakash 2010). As for existing constructivist approaches, which foreground shared principles, while ideas (of proper practice) are taken seriously, competition gives way to cooperation; here, accountability is seen as arising from joint endeavor and mutual accord among aid practitioners (e.g. Deloffre 2016). The challenge, Berghmans et al. (2017: 1532) have explained, is that the harmonious reading risks ignoring “the political and conflictual nature of these processes of negotiation.”

This paper explores the tensions that arise in the process whereby humanitarians translate abstract accountability principles into concrete accountability practices. Through a study of the Humanitarian Accountability Partnership (HAP) International, I find that initial agreement on the desirability of beneficiary (or “downward”) accountability quickly gave way to a series of principled disagreements and power struggles. I identify four central tensions. The initiation stage of HAP was marked by two conflicts—a debate about enforcement and a turf war over location and control—and culminated in the rebranding of the Humanitarian Ombudsman into HAP and its relocation from the UK to Geneva. The implementation stage of HAP was characterized by ideological tensions over certification and intra-organizational struggles over leadership of the initiative. The contemporary practice of humanitarian accountability is thus shown to be a contingent and contested social process, with humanitarian identity and practice ultimately at stake.

These ideas are developed as follows. I begin by surveying the literature on INGO accountability, classified according to principal-agent and stakeholder approaches, and outline my analytic framework. I take seriously the constructivist premise that ideas (which are intersubjectively held) and identities (which are socially constituted) are critical variables in political life, particularly for public benefit organizations like INGOs whose legitimacy (and hence authority) hinges on public acceptance. I argue, however, that constructivist insights have only partially been applied to INGO accountability, resulting in the very situation that Berghmans et al. (2017) lament: the inattention to struggle and conflict that characterizes existing research. Drawing on international relations (IR) theory, I differentiate between “liberal” constructivism, the variant most often applied to INGO accountability, and more “critical” strands, which foreground the power that inheres in social practices, the exclusions upon which group identity is founded, and the inherent contestability of concepts like accountability. After outlining my methods, I present the results of my research into HAP International; I focus on the contest to define and control the institutions of humanitarian accountability. In the conclusion, I discuss several implications of the findings.

Accountability, Contested

Though accountability is widely accepted as positive and necessary, the concept is peculiarly complex by virtue of the multivariate environments within which INGOs operate and the diverse constituencies to which they account (Stein 2008). Consequently, accountability is “chameleon-like” for its different meanings in different contexts (Everett and Friesen 2010: 469). This variability is reflected in the multitude of associated terms, the chief approaches to which are discussed below. The common denominator is that accountability is a relational concept. As Bendell (2006) explains, accountability concerns a relationship between A and B, where A is accountable to B if they must explain their actions to B, and could be adversely affected by B if B does not like the account.

Traditionally, accountability has been studied from a principal-agent perspective founded on contracting relationships between principals, often donors or states, and agents, the implementing INGOs (Edwards and Hulme 1995; Ebrahim 2003; Gugerty and Prakash 2010). Here, accountability is an externally motivated construct (Dhanani and Connolly 2015), the logic of which is shaped by the power exercised by principals to control agents (Pallas and Guidero 2016). Coule (2015) labels this the “systems-control” approach, and the result is frequently a narrow, short-term, compliance-based accountability directed upward (Ebrahim 2003). The risk is that accountability, thus conceived, can squeeze out broader impacts and affected groups. Indeed, this concern is borne out by research conducted on a wide range of INGOs, from advocacy (O’Dwyer and Unerman 2008) to environmental (Coule 2015) to humanitarian (Stein 2008) agencies. Thus, despite aspirations for broader accountability, Schmitz et al. (2012) find that INGO leaders overwhelmingly highlight donor expectations and financial accountability, evidence of which is also seen in INGO annual reports, which tend to emphasize statutory reporting requirements and the interests of funders to the detriment of other considerations (Dhanani and Connolly 2015).

This article’s focus case, HAP International, is representative of an alternate trend in accountability, which seeks to incorporate a wider array of stakeholders and thereby to more holistically address the ethical and societal dimensions of INGO practice. The stakeholder approach, as it is often called, endeavors to transfer the “right to accountability from exclusively those that have authority over an organization to anyone that has been affected by the organization’s policies” (Lloyd 2005: 3). Accountability, in this view, includes both being held responsible and taking responsibility, a stance that scholars variously label “expressive” (Coule 2015), “intrinsic” (Dhanani and Connolly 2015), or “moral” (Everett and Friesen 2010).

Two points bear on the discussion of stakeholder accountability. First, INGOs are potentially accountable to multiple stakeholders at every stage of activity. The cardinal division is between “upward” accountability to states or donors and “downward” accountability to beneficiaries or clients, with the literature frequently emphasizing the relative powerlessness of beneficiaries, who are often excluded from decision making (Ebrahim 2003; Schmitz et al. 2012; Coule 2015). The “horizontal” level also merits attention; van Zyl and Claeyé (2018) highlight accountability to the INGOs themselves, to their boards, to members and staff, and to self-regulatory organizations, while Lloyd (2005) addresses inward and peer dimensions. Indeed, the horizontal level, specifically board, member, and staff relations, plays a central role in the HAP International case, given the power of the secretariat. Second, because these individuals and groups have divergent interests and priorities, not to mention different levels of efficacy, negotiations (formal or informal) are at the heart of the stakeholder approach (Coule 2015; Berghmans et al. 2017; van Zyl and Claeyé 2018).

This paper investigates stakeholder approaches to INGO accountability, specifically initiatives to strengthen beneficiary accountability in humanitarianism. It focuses primarily on the interplay among constituent groups and the intensity of deliberations in the unfolding of accountability. Following recent scholarship (e.g. Kennedy 2019; Crack 2018; Berghmans et al. 2017; Deloffre 2016), it adopts a constructivist framework of analysis. Constructivism is well-suited to the analysis of stakeholder accountability, given that it takes seriously the motivating and structuring capacities of shared ideas, not only strategic interests (Deloffre 2016), while encompassing the multiplicity of relationships into which INGOs enter (Berghmans et al. 2017). But the application of constructivist theory to INGO accountability has also been criticized. Berghmans et al. (2017) probe constructivism’s seeming inability to account for the complexities and tensions inherent in the crafting of accountability initiatives. I share these concerns. The literature has tended to emphasize “deliberative dialogue” (Williams and Taylor 2013) or “social learning” (Deloffre 2016) as vehicles for reaching consensus; power and contention largely drop out, despite the centrality of authority and control to the practice of accountability. Moreover, because existing research has largely focused on the creation rather than the actual operation of accountability initiatives (Crack 2018), we know relatively little of the disputes that continue into or emerge during the implementation of accountability standards.

The issue, I suggest, has less to do with constructivism, per se, than with its selective application to INGO accountability. In IR, as Havercroft and Duvall (2017: 157) wryly observe, the approach is sufficiently popular that “anyone studying international norms must necessarily be a social constructivist of some kind.” But this labeling conceals considerable variation: thin versus thick (Wendt 1999), mainstream (or conventional) versus critical (Hopf 1998), liberal versus realist (Mattern 2004); and even agonistic (Havercroft and Duvall 2017) constructivism. While the precise demarcations are open to debate, existing constructivist research into INGO accountability fits comfortably in the liberal (or mainstream) category, which emphasizes variables like persuasion, dialogue, and shared norms.

By drawing on critical constructivist traditions in IR, we have space to consider factors like contestation and exclusion as equally important as (and coexisting with) collaboration and mutuality (Hopf 1998: 185). Specifically, this broader framing lends insight into three elements of INGO accountability. First, critical scholarship underscores the ubiquity of power, which is seen as inhering in all social relationships, even those involving principled actors. Indeed, as Carpenter (2014: 19–20) has found of transnational activist networks, they are “sociopolitical structures with their own kinds of hierarchies, power relations, and governing mechanisms.” Even as transnational civil society seeks to elevate the voices of the marginalized, not all have equal access or control. In this reading, power extends beyond material capabilities and conscious coercion to encompass ideas and institutions (Barnett and Duvall 2005: 3). Applied to INGO accountability, critical constructivism is attuned to the structured (and often unequal) relationships among stakeholders, to the power of dominant discourses (including “accountability” itself), and to the importance of institutions (like HAP), which stabilize and direct sanctioned practice.

Second, stemming from the constructivist analysis of identity (as antecedent to interests), critical variants also focus attention on exclusions. Campbell (1998: 8), following Judith Butler, puts it as follows: “Inescapable as it is, identity—whether personal or collective—is not fixed by nature, given by God, or planned by intentional behavior. Rather, identity is constituted in relation to difference.” In the context of INGO accountability, the focus is on constitutive rules as well as on the implications of these acts of demarcation. The cumulative effect of these practices, Kennedy (2019) explains, “is to establish (often implicitly) ‘this is what we are,’ and, hence, also ‘this is what we are not.’”

Finally, inasmuch as concepts are by definition social facts—their meaning hinges on human agreement—they are inherently contestable. As Niemann and Schillinger (2017: 47) discuss, critical constructivism poses the challenge of “theorizing with the unfixity” of concepts. That accountability is a social fact is exemplified by the aforementioned ambiguities in the academic literature. Efforts to regulate a practice are simultaneously attempts to standardize its meaning (Barnett and Duvall 2005: 21), but these meanings are always contingent and subject to debate and (over time) modification. Stein (2008: 125) argues that the persistent accountability debates within humanitarianism reveal “deep divisions about principles and practices that are entangled with shifting currents of ethics, power, and politics.”

Collectively, these observations direct our attention to various ways in which contestation and cooperation comingle in the unfolding of INGO accountability.

Methods and Case Selection

Humanitarians have long grappled with questions of quality and accountability, but discussions intensified in the 1990s, particularly following the Rwandan genocide, resulting in a proliferation of intra-sectoral initiatives (Hilhorst 2002: 194). While it is often said that humanitarianism experienced an accountability revolution, Lawday (2006: 40) argues that what actually transpired “was a revolution in accountability to donors, not to beneficiaries.” Ten years after the Rwandan genocide, an analysis of 35 nongovernmental codes reported one dominant way of looking at accountability: “to ensure compliance with reporting requirements, laws, and regulations” (Lloyd and de las Casas 2006).

The Humanitarian Accountability Partnership, or HAP International, was an exception to this rule. From 2003 until 2014, HAP was the best-known initiative advocating for beneficiary (or downward) accountability, and its values and models live on in the Core Humanitarian Standard on Quality and Accountability (CHS). According to a HAP staffer, “the aim of HAP was to say, ‘Well, what about the beneficiary? Who is listening to the beneficiary?’” (H1) HAP has been celebrated by academics and practitioners alike for its principled and collaborative model of social transformation (e.g. Ramalingam et al. 2009: 33; VanRooyen 2013). As Deloffre (2016: 744–745) argues in an extended treatment of the topic, HAP “reflects a logic of participation that emphasizes mutual accountability and learning.” Research on HAP thus follows the accountability literature in highlighting cooperation; conflict and contestation are largely absent from the narrative.

This paper process traces the development of HAP International, beginning with the Humanitarian Ombudsman (HO) Project in 1997, continuing through the initiative’s rebranding as HAP in 2003, and culminating with the creation of the CHS in 2014. Process tracing seeks to understand development, change, and evolution through the investigation of causal mechanisms; the focus is on turning points and critical junctures (George and Bennett 2005: 6). Data is drawn from two sources of evidence: semi-structured interviews with key HAP and HO figures and archival materials. In total, 29 interviews were conducted between 2010–2014 and included several generations of HAP and HO staff, 4 HAP/HO executive directors, board members, and implementing organizations (Table 1). Depending on their position, participants were asked to comment on the design, implementation, and evolution of HAP. What prompted creation of the initiative? Which issues were debated? How was the initiative run? How did it develop? Participants spoke candidly and care has been taken to preserve anonymity, including labeling easily identifiable executive directors as “staff.” To quote a board member, HAP was located in “the cauldron of Geneva. It’s an insular place where everyone knows everything that’s going on” (B4). Archival materials supplement the interviews. These documents were available online, accessed in person at the HAP office in Geneva, and provided directly by participants.

Table 1 Interviews
Full size table

Findings

HAP International’s formal existence was relatively brief, spanning an 11-year period from 2003 to 2014, but its roots go deeper, while today its ideas on beneficiary accountability and management processes form a constituent part of the Core Humanitarian Standard. This section outlines the HAP model and the shared sentiments informing its creation.

Though HAP officially dates to 2003, the initiative is fundamentally rooted in the aftermath of the Rwandan genocide. It was at this point, according to HAP (2006: 8), that “many agencies learned that good intentions were no guarantee of quality.” In interviews, leading figures reflected on the emotional state of practitioners at the time. To quote an Ombudsman Project staff member: “People in humanitarianism had a deep sense of almost shame and guilt that they didn’t do the right thing. The core of the feeling was a belief that they wanted to do better” (O3). Similar sentiments were expressed in interviews with HAP staff (H1), board members (B4), and member organizations (M1, M4). HAP was one expression of a wider reform agenda sweeping humanitarianism, with the Sphere Project, People in Aid (PIA), and the Active Learning Network for Accountability and Performance in Humanitarian Action (ALNAP) among other self-regulatory and quality initiatives created in this period (Krause 2014: 128–130).

Krause (2014: 140) labels HAP “the most far-reaching and radical [reform project] in terms of impact on the organizations that embrace it.” This is because HAP was more than a standard-setting initiative; its staffers were adamant in interviews that they would not provide “how to” accountability guides to member agencies (H4). Rather, HAP sought to address beneficiary accountability through intra-organizational change, namely through its quality management framework and certification in the HAP 2010 Standard in Accountability and Quality Management. The Standard’s six benchmarks were intended to promote beneficiary inclusion at all stages of the product cycle, from commitments to complaints to continual improvement (HAP 2010c). HAP was also innovative in that, together with the personnel-focused PIA, it was one of only two sector-wide initiatives promoting third-party certification. Compare this approach to the Sphere Project, as HAP staff frequently did. Sphere set minimum standards for emergency response (a “how to” guide) and left implementation to the discretion of agencies (Kennedy 2019).

HAP promoted beneficiary accountability for reasons both principled and practical. From a principled standpoint, HAP was preoccupied by power relations in aid work and the potential for harm and abuse (Everett and Friesen 2010: 475). The Guide to the HAP Standard explains that “inequality between provider and receiver means that the act of giving is often exercised without the consent of the person in need” and relief occurs in a “state of virtual judicial impunity” (HAP 2008b: 3–4). Consequently, accountability for HAP (2010c: 1) is defined as “the means through which power is used responsibly. It is a process of taking into account the views of, and being held accountable by, different stakeholders, and primarily the people affected by authority or power.” HAP thus embodied the stakeholder approach to accountability. Practically, HAP (2010d: 7) argued that “humanitarian programmes will be delivered more effectively; saving more lives; and improving the quality of more people’s lives; if their intended beneficiaries participate in all stages of the programme cycle.”

The principles motivating HAP were widely shared among practitioners. In a retrospective on humanitarian accountability, ALNAP’s Paul Knox-Clarke and the Ombudsman’s John Mitchell explain that the movement was informed by moral claims rooted in humanitarian principles and the rights-based approach and a belief that accountability would yield better performance (Knox-Clarke and Mitchell 2011: 3). MSF’s Austen Davis characterizes the prevailing view as follows: “accountability is one of those unquestionably ‘good things’ that it is taboo even to question” (Davis 2003: 16). However, agreement on principles did not translate into consensus on mechanisms (Kennedy 2019). One of the Ombudsman figures put it as follows: “Everyone’s heart was damaged by Rwanda. But what you do about it is a different question. There’s a lot of horrible stuff that goes on in humanitarian responses. It puts you to tears. But where do you draw the regulatory line? It’s a constant debate and there are never clear answers” (O3). HAP (2011: 54) itself acknowledged the “lack of a widely shared vision of what an ‘accountable’ humanitarian system would look like” and bemoaned the “challenging years convincing the sector of its value” (HAP 2010b: 36).

The following sections investigate these debates and challenges, tracing the launch of the Humanitarian Ombudsman Project in 1997 and its transformation into the Humanitarian Accountability Partnership in 2003. The rise and fall of the HO provide a window into the contentious nature of humanitarian accountability, particularly into debates and leadership struggles at the “horizontal” sectoral and intra-institutional level, while the different models proposed by the HO and HAP point to variability and contingency in the enforcement of accepted principles.

Initiation

The Rwandan genocide precipitated a period of reflection and reassessment of humanitarian assistance. The key component was the Joint Evaluation of Emergency Assistance to Rwanda (JEEAR), a multi-actor process proposed by Danida (Denmark) in September 1994 and tasked with deriving lessons from the Rwanda experience (Borton 2004). Over a 15-month period, concurrent with the post-genocide Goma refugee crisis, experts and representatives from developed country bilateral donors, multilateral organizations, and INGOs researched and deliberated, producing four lengthy reports. Most of the reports and their lessons actually implicated states and multilateral organizations, not INGOs. JEEAR Study III was the exception; this report advised—among multiple recommendations—that the international community: “Identify a respected, independent organization or network of organizations to act on behalf of beneficiaries of humanitarian assistance and member states to perform the functions described in option (ii) above” (Eriksson 1996: 61). The word “ombudsman” was mentioned, once; the recommendation was that an ombudsman be established in the UN Department of Humanitarian Affairs. Proponents of beneficiary accountability picked up on this suggestion, launching the Humanitarian Ombudsman Project at the World Disasters Forum in London in June 1997.

What was an “ombudsman”? Remarkably, its proponents were initially unsure. “People thought it sounded good but no one knew what it was” (O2), one of the HO’s employees recalled. An early project proposal draft explained: “Instinctually, it was thought that an Ombudsman was the best way forward to increase accountability to beneficiaries, while facilitating improved performance by agencies” (Ombudsman Project 1998a). Research conducted by the HO would subsequently determine that “ombudsman” is an old Swedish word used to describe someone who “represents or protects the interests of another” (Mitchell and Doane 1999: 115). Following a feasibility study, ground consultations were launched at the 1998 World Disaster’s Forum. The Project was coordinated by the British Red Cross, based in London, and guided by 11 leading British agencies and institutions.Footnote 1 Beneficiary accountability now had institutional life. From May to August 1999, the Ombudsman held consultations in Costa Rica, Rwanda, Sri Lanka, and Sierra Leone. During August 1999, research was conducted in Kosovo with the intent of developing a possible model. This research confirmed that there was a clear need for and interest in a mechanism for holding humanitarians to account, but yielded different perspectives on its shape and scope.Footnote 2

From the beginning, despite considerable enthusiasm, the Ombudsman wrestled with competing visions of accountability and, as its meeting minutes attest, even struggled for coherence in its own Steering Committee (e.g. Ombudsman Project 1998b). Future HAP executive director Nicholas Stockton (2000: 19) observed:

It is worth noting that our efforts to promote greater accountability to legitimate humanitarian claimants through the promotion of the Humanitarian Ombudsman Project has met with outright opposition from some of the Code [of Conduct]’s signatories. While some are fearful of unrealistic expectations and unreasonable litigation, others are opposed in principle to subjecting the humanitarian act of compassion to technical, legal or contractual norms.

This opposition was recalled in multiple interviews (B2, B3, O1, M10). An Ombudsman figure observed that outside of the UK, INGOs were “very hostile to this” (O2). There were two overriding reasons. First, there was unease about the Ombudsman and its objectives, especially with respect to enforcement. Second, the HO’s proponents had difficulty demonstrating the wider applicability of the project outside of Great Britain.

Standards and Enforcement

First, concerns about enforcement tended to dominate the debate. Ian Christoplos, a Project Researcher, noted that agencies had different perspectives on the proper role of standards, particularly on the balance between punishment and incentives. One camp felt that humanitarian standards should be used to constrain and sanction deviant agencies; others viewed standards as flexible guidelines to facilitate efforts at learning and practice (Christoplos 1999: 127). Sphere fell into the latter camp, prioritizing learning over sanctioning, and its staff recounted the oft-heard fear that the HO would be used to enforce its technical standards (S1, S2). This claim bemused HO staff. Said one: “there was a lot of talk about policing. But how on earth would we police?” (O2) Another was “mystified and fascinated” by the talk of enforcing Sphere: “there’s no way in a million years that this had anything to do with Sphere” (O1).

For its critics, French agencies most vocally, HAP embodied a worrisome humanitarian tendency to shift political responsibility from states onto INGOs and, subsequently, onto disaster-stricken populations (Terry 2002: 53). In interviews, prominent critics questioned whether accountability was even meaningful or achievable in the field (F1). As Austen Davis (2003), then head of MSF-Holland, framed it: “The degree to which a humanitarian worker can be accountable to people in societies that have been destroyed from within is questionable… A sectoral or systemic attempt to generate humanitarian accountability has little currency. It blurs responsibilities and differences of capacity.” Concerns such as those expressed by Davis and Terry may stem from different perspectives on the nature and practice of humanitarianism (Kennedy 2019; Krause 2014). This view was confirmed in interviews with HAP and HO staff, who attributed MSF’s opposition to their “internal organizational philosophy” (H1, H2). One put it thus: “In Kosovo, for instance, the English sent the technocrat; the French sent the philosopher” (O3).

Organizational philosophies aside, there was an element of truth to concerns that the HO would enforce boundaries. As Christoplos (1999: 136) himself wrote in 1999:

The HAO should not work from a basic assumption that all actors have a role to play in the humanitarian arena. The Codes provide a basis for identifying those actors that share a basic set of common humanitarian values. Those that do not, including those international agencies that are too incompetent to provide a significant contribution, should be publicly labelled as such. Donors, host authorities and the humanitarian community should naturally be encouraged to treat these actors as being outside the humanitarian consensus.

Christoplos (1999: 132–133) indicated that the HO’s founders would not shy away from a “public critique of some international agencies, local institutions and individuals, if they demonstrate that they do not share our basic moral values, i.e. that they do not operate within the broad moral frameworks of the Codes.” This could require a “joint decision to isolate certain actors who are making the situation worse”. Nicholas Stockton (2000: 21), too, supported “driv[ing] a wedge between those that do comply and those that only want to sign for purposes of window-dressing. The membership rules of the humanitarian club, as defined by the Code, need to be much tougher.” This is not surprising; Gugerty and Prakash (2010) note that codes by design distinguish good from bad practice. But by which measures is good practice defined? Who decides?

Turf and Territory

To an extent, these accountability debates were ultimately about turf and territory, which is to say, whose prerogatives were safeguarded or threatened and where would the project be based? Ideas—in this case about appropriate practice—are difficult to disentangle from their social and institutional contexts, including and especially national INGO environments, with prior research underscoring the impact of national identity on transnational action (Stroup 2012). Ombudsman documents suggest that the perception that the project was British—in location and cultural context—was pervasive and damaging. There was a call for further internationalization of the project at a Steering Committee meeting in December 1998, where it was also acknowledged that “the choice of location is symbolically important—London may be a poor choice to ‘headquarter’ the project” (Ombudsman Project 1998b). Similar findings were reported from consultations conducted in the United States in 1999 (Ombudsman Project 1999).

Consequently, the Ombudsman was cautiously received outside of England. For instance, aside from the British Red Cross, the HO reported “strong opposition” from the Red Cross family, with the International Committee of the Red Cross (ICRC) particularly concerned that the Ombudsman’s remit might include protection (Doane 1999). HO staff attributed this stance to territory—protection, or the preservation of civilian safety, integrity, and dignity, is central to the ICRC’s mission—and philosophy—French skepticism of accountability was perceived to permeate the ICRC system (O1, O3). UN agencies were equally wary, with the Office for the Coordination of Humanitarian Affairs labeling the Project “threatening” for fear that INGOs “would shame the UN into becoming involved at a later date” (Ombudsman Project 1999). The American InterAction federation offered mixed support; its members were somewhat supportive, but InterAction’s head, Jim Bishop, personally opposed the HO (O1, O3). For their part, UK agencies feared a loss of ownership were the Project to be moved, especially as they felt that accountability debates were much further advanced in the UK relative to other national contexts (O3).

Despite these misgivings, the Project was, in fact, internationalized. At the conclusion of Stage III, in June 2000, the Humanitarian Ombudsman Project moved from London to Geneva, hired an executive director, and was rechristened the Humanitarian Accountability Project. Under new leadership, HAP was transformed. It undertook five research projects in Sierra Leone, Afghanistan, and Cambodia to test different mechanisms, concluding that accountability would best be served by the creation of an international self-regulatory body. This was a dramatic shift from the light, field-based Ombudsman. Fourteen humanitarian agencies endorsed this recommendation in January 2003, and the Humanitarian Accountability Partnership International was officially registered in Geneva in March 2003 (Callamard 2003).

For many of the former Ombudsman staff, it was anything but clear that other accountability mechanisms needed to be tested. The general view in interviews with HAP and especially HO staff was that the change in direction was less a product of the research conducted than it was a case of new people “wanting to shape the thing in their own way” (O1). Power manifests in this instance through new executive officers taking HAP in a different direction. Two veterans spoke of the “suddenness” of the change from the context-specific Ombudsman (O3, B4). The new location also played a role, given that HAP was now a part of the Geneva “cauldron” (B4).Footnote 3 Several highlighted a March 16, 2000 meeting in Geneva, which brought together major networks and principals. Borton (2012) recalls this as the point at which “the consensus that had been established among the members over the previous two and a half years fell apart as the concept of the Ombudsman came in for questioning and criticism from agency chief executives,” some of whom had been represented by mid-level staffers and were only now realizing that a new accountability initiative was about to be launched.

Implementation

From HO to HAP, from London to Geneva, invested with new leadership—much had changed in the world of beneficiary accountability, but the contested nature remained constant. Interviews from across the movement speak to this point. While Ombudsman staff were, perhaps unsurprisingly, critical in their appraisals of HAP—one labeled its history “terribly fraught” (O3) and another, colorfully, observed “there’s a lot of blood on the walls” (O2)—HAP figures actually shared this somber assessment. HAP’s institutional existence was “turbulent,” to quote a board member, while the early years especially were “a tough period” (B4). Overall, in the words of a staffer, “HAP has been cautiously received by the sector” (H2). Why was this? What were the key sticking points? This section draws out two tensions that characterized HAP’s institutional existence: disagreement over certification and a struggle over leadership.

Certification

As aforementioned, HAP was one of the few humanitarian self-regulatory initiatives to certify the compliance of its members. This was a point of distinction but also a bone of contention between HAP’s pro-certification staff and its often ambivalent membership.

By way of background, in the HAP 2010 Standard, certification is defined as: “Through a thorough and independent audit, the HAP certification scheme verifies and recognises that an organization meets the HAP Standard. HAP, or a body accredited by HAP for this purpose, carries out the audit” (HAP 2010c: 4). Certification was optional but strongly encouraged; it resulted from an 18-month process that included the development of an accountability framework and a quality management system, document review, and headquarters and field visits. INGOs could then display the “HAP 2010″ certification mark.

For HAP staff and leadership—the Geneva-based secretariat—certification was a constitutive element of the initiative’s identity and a marker of its distinctiveness. “It’s the fact that we verify, that’s what makes it different” (H4) and “fresh” (H2), I was told. A staffer explained that they consciously differentiated themselves from the Sphere Standards’ technical guidelines: “We are a bit pooh–pooh about Sphere, too, as you can tell. It’s not fair, but we are, because their concern was to get this broad buy-in as the way to get change, as opposed to actually pushing beyond that” (H4). This impression was corroborated by an external evaluation of the project, commissioned by HAP in 2009 and authored by Geoffrey Salkeld. In the report, Salkeld wrote of HAP’s “single-minded emphasis on its version of compliance verification,” suggesting that in HAP’s strategic plan: “The phrase ‘and to accredit its members accordingly’ in objective (4) has emerged as a leading edge objective, morphing into ‘compliance verification through the HAP Certification process’” (Salkeld 2009: 6).

HAP’s membership did not universally share the secretariat’s enthusiasm for certification. Indeed, according to Salkeld (2009: 36, 48), certification was the most controversial issue in his study and the source of “considerable tensions in the relationship between the HAP Secretariat and some Members.” He found that compliance verification through certification had “been done without the whole-hearted support and engagement of some Member agencies and Board members” (Salkeld 2009: 6). In an interview, a board member put it as follows: “There’s this tension about whether certification is the only route to accountability and whether HAP should really be encouraging all forms of accountability, not just certification, but keeping certification as a long-term objective, perhaps” (B2). Archival materials paint a similar picture. For example, at the very first Board and Advisory Committee meeting in 2003, it was noted that HAP had had difficulties because “the regulation component of HAP International creates this fear” (HAP 2003b). Fears were again expressed at the second board meeting, in 2004 (HAP 2004a: 4), and general assembly meetings were much the same. One notable exchange, from 2008, pitted members—“Certification itself is not the ultimate goal”—against HAP’s secretariat, namely Executive Director Stockton: “To drop compliance verification would remove HAP’s reason for existence” (HAP 2008a).

While some members, particularly mid-sized organizations, embraced certification, HAP’s membership as a whole was thus divided (B4). This was directly reflected in the slow pace of certification. Though HAP’s 2007–2009 Strategic Plan envisioned 36 agencies certified by the end of 2009, by December 2009, only 7 had completed certification and, as of January 2015, when the CHS was launched, there were still only 17 certified members (HAP 2010b: 35).

Why promote certification? Interestingly, according to Salkeld’s evaluation, the evidence did not yet support this objective; he found “little field-level evaluation or assessment of the impact of the HAP certification on the quality of life, autonomy or dignity of beneficiaries” (Salkeld 2009: 7). The HAP secretariat disputed these findings—to its credit, publishing them anyway—but Salkeld’s observations are amply supported by archival data. From the earliest meetings, there were consistent calls by signatories as well as board members for evidence that HAP membership and certification made a difference on operations—for a “business case.”Footnote 4 Impact assessment, and the difficulties thereof, was also a consistent theme of HAP’s annual Humanitarian Accountability Report. For instance, in the 2008 edition, material reviewed indicated that “more evidence is required from agencies… to convincingly demonstrate a clear ‘business case’ for improved accountability to intended beneficiaries and local communities” (HAP 2009: 62). Acknowledging gaps in the evidence became something of an annual ritual (e.g. HAP 2007: 30; 2008d: 52; 2013: 59).

If not impact, what was driving certification? Based on interviews and document review, the answer appears to be ideology: key figures in the secretariat drew inspiration from the quality management movement, also known as “quality assurance” and “new public management.” Quality management approaches conceive of the public as users and consumers who hold service providers to account through accountability mechanisms and by seeking other providers (Power 1999; Krause 2014: 142–144; Hilhorst 2002). HAP owes its embrace of this philosophy to two individuals in particular. The first, Agnès Callamard, was HAP’s first executive director, hired during the transition from the Ombudsman. Callamard came from the human rights sector, where public sector reforms had already exerted considerable influence, and was central to HAP’s evolution from a floating field mechanism to an initiative pushing intra-organizational change (O1). The second, Nicholas Stockton, was the former Emergencies Director for Oxfam GB, a key figure in Sphere, and a member of the Ombudsman’s Steering Group. As executive director at HAP from 2003 to 2010, he had an enduring impact on the initiative. While Callamard was more invested in justice and Stockton enforcement, participants cited their shared embrace of the market approach (O1, O3). As Stockton (2005) wrote, “we believe that meaningful consultation with ‘customers’ is both an ethical objective and a strategy for achieving quality management of humanitarian action.”

Quality management was integral to the HAP vision. Both the HAP 2007 and 2010 Standard followed ISO quality management guidelines, namely the 9000 series. ISO 19011 was also used to review performance of HAP’s independent auditors on the Certification and Accreditation Review Board (HAP 2010b). HAP followed ISO closely for guidance on accreditation and certification, “not necessarily because it’s the best, but because it’s the one that’s around, and it is the international standard” (H2). As Everett and Friesen (2010: 476) observe, HAP’s core documentation was infused with business terminology, including “product,” “quality,” “customers,” “costs,” and “value.” HAP (2004c: 2) acknowledged that “while it may be argued that humanitarian action is not simply a ‘product’ or ‘service’, and that the intended beneficiaries are not just ‘customers’, the core ISO objective of promoting quality management is exactly analogous with HAP’s transformative agenda for the humanitarian system.”

Leadership

A second divide, linked to the first, was over leadership of HAP and, ultimately, implicated the horizontal accountability of the secretariat to its board and members. It bears emphasizing that institutional structures are not unitary, nor is power evenly distributed. In HAP’s case, functions were divided across three bodies, plus committees and working groups. The general assembly was the largest body, including full and associate members (100 total, at the end), and structurally weakest, vested with few powers beyond oversight and board elections. The board was 12-strong and elected from full members (with one-third independent). It met twice a year to set HAP’s strategic direction, including reviewing the budget and member applications. Finally, the executive, or secretariat, led by a board-appointed executive director, was tasked with research, collaboration, training, assisting, monitoring, and advocating for implementation of HAP’s mission (HAP 2008c).

Though HAP was a membership-based initiative, its members held few levers of power. As was noted at the first board meeting: “It is difficult at this stage to identify a meaningful role for the General Assembly. It elects the Board members, but somehow it is cut from the center of discussion” (HAP 2003a: 3). Salkeld (2009: 6, 24, 50) found that HAP had “become increasingly Secretariat-driven,” yielding tensions between the head office and many board members, who felt that the secretariat was “exceeding its mandate.” Oxfam GB and World Vision International even resigned from the board. Similar views were expressed in interviews with HAP board members, with one lamenting the secretariat’s “go it alone” mentality (B4) and another referencing “years of counterproductive infighting” (B3).

Leadership style, personality, and ultimately institutional culture also emerged as major obstacles. “HAP,” a board member explained, “has always been fueled by a sense that something has to change; it was founded with an element of anger, which inflects its culture and energy” (B4). HAP was, to quote a staffer, “trying to change the whole global system” (H2) and change it in ways consistent with HAP’s “transformative agenda” (Davidson 2002: 42), as a HAP publication framed it. One staffer even envisioned a future in which 75% of humanitarian organizations were certified (H1). Certification was the lynchpin, at least for the secretariat, because it provided the only assurance that agencies would live up to their commitments. A staffer explained:

I think that is why HAP irritates people, aside from other reasons that we’ve been irritating in the past, that we don’t let them get away with just saying, “We’ve done training,” or “We’ve signed up to HAP like we’ve signed up to the Code of Conduct,” that we actually keep pushing it… So HAP will come along and say, “So you keep saying you’re signed up to Sphere… How do you make sure it’s used and it doesn’t just sit around?” (H5)

A colleague accounted for opposition to certification as follows: “Everybody goes back to the same donors… and why would you have someone come into the organization that might show that you’re not as good as your competitor?” (H2) Another suggested that large members did not want certification because they were “beyond the sovereign” (H6)—untouchable. This mistrust was at the root of what multiple respondents referred to as Nicholas Stockton’s “big stick” approach to accountability (O3, B4).

Personality and leadership style emerged repeatedly in interviews, unprompted. A senior staffer at a large American INGO expressed in highly personal terms her antipathy toward HAP’s secretariat, in contrast to Sphere staff, who “simply were not arrogant” (M5). An Ombudsman staffer explained that HAP’s initial leadership team was composed of “extremely headstrong people who didn’t get along with themselves or with others, much less with their own board.” “In general,” this respondent concluded, “the psychology and influence of individuals is a very big issue in HAP” (O1; also O2). HAP’s external evaluation, too, indicated that the personalities of the secretariat’s staff had been internalized into its organizational identity, manifested in a level of “defensiveness” and what some members called an “aggressive style” (Salkeld 2009: 7, 27). A HAP board member labeled it “forthright—probably over so” (B4). HAP staff conceded these points: “When we’re looking inward, we feel, oh gosh, we’re not as loved as Sphere, but partially that is because it’s much harder to be part of our ‘club’ than it is the Sphere club.” “Shock tactics” was the term used (H5). Another added, “we did have this idea that we had to go into organizations and the way to do it is like this” (H4).

Discussion and Conclusions

Accountability is inherently contested. A social fact, its meaning hinges on human agreement. Following the Rwandan genocide, humanitarians largely did agree on a set of basic propositions: that outcomes did not always match intentions, that aid could paradoxically harm local populations, and that all stakeholders—not just donors—should have a part to play in relief and reconstruction. From a normative standpoint, HAP reinforced and elevated ongoing efforts by aid professionals to incorporate beneficiaries into previously donor-centric accountability frameworks (Deloffre 2016). Indeed, as Krause (2014: 145) observes, it is precisely because HAP oriented itself around beneficiaries, and because “the beneficiary has such symbolic importance in relief, [that] those who did not join the initiative did not feel it appropriate to contest it very loudly.” In the words of one respondent, “they’ve staked out high moral ground and put a HAP flag on it” (O1). HAP was thus shielded from some critiques—but not from all. From enforcement to certification, from turf to leadership, the key debates from HO to HAP implicated the processes by which humanitarians put accepted principles into practice. Ostensibly methodological, as McGee and Gaventa (2011) suggest, the debates actually expressed fundamental—and not easily overcome—ideological and epistemological stances. What is the essence of humanitarian practice? Which (and whose) ideas hold sway?

Interviews with proponents of beneficiary accountability thus offer insights resonating with ongoing scholarly debates about the purposes and practical implications of INGO self-regulation. In particular, as anticipated by critical constructivists, the findings point to the persistence of power in shaping accountability initiatives and to the intrinsic connection between shared practices and collective identity.

First, we see in HAP’s history the omnipresence of power, a power simultaneously material and ideational. From the outset, HAP officials called attention to the structural privilege of aid agencies intervening in emergency situations and to the abuses that so often resulted. Here, we see power as direct control or coercion (Barnett and Duvall 2005: 13) wielded by the materially advantaged INGO over the disempowered “victim” and by the affluent institutional donor to compel a response from the INGO. Through beneficiary accountability, HAP envisioned a means by which to flip the power in favor of affected populations. How powerful was that mechanism, though? Could local groups seek redress or change service providers? Would their complaints be heard? (Davis 2003) Perhaps not, at least, not automatically, hence the HAP secretariat’s focus on third-party certification and the “big stick” approach.

Power also manifested in subtler institutional and discursive forms. Various forces competed for control of the HO/HAP institution. An Ombudsman staffer, recalling the move from London to Geneva, explained that: “The debates were very much power plays about who should rule and how humanitarianism should be done” (O3). A colleague admitted that, in the end, “it was definitely a bit of a battle of positions. Who would lead the big quality initiative?” (O2) These battles did not transpire because HAP was materially strong; indeed, its funding situation was rarely secure and it did not even exercise its sanctioning capacity (H1, H5). Rather, HAP’s power, and especially that of its secretariat, flowed from its daily existence—from the banal tasks by which it promoted and embedded reformed humanitarian practices, thereby reinforcing its own position as the focal point for accountability. Rebranded and rooted in Geneva, invested with new leadership, HAP’s secretariat expressed institutional power, “guid[ing], steer[ing], and constrain[ing] the actions (or non-actions) and conditions of existence of others” through its rules and procedures (Barnett and Duvall 2005: 15). Take the debate over certification. Faced with the fears of some members, HAP officials bolstered their position through reference to founding documents. At the second board meeting, for instance: “Nick [Stockton] reaffirmed that the HAP-I Statutes (Article 5 point 4) place an obligation upon us to establish an accreditation system” (HAP 2004a: 4).

Second, as was emphasized in interviews with members, HAP’s standards were about more than “ticking the box” (M1) or “cheap talk” (M8). They were verifiable and certifiable, targeting the inner workings—the management framework, and hence the standard operating procedures—of members so as to produce, as one HAP publication labeled it, “The Accountable Organisation.” This publication explains that “changes in practices at field level… require broader transformations, in the first place at the level of the humanitarian organizations themselves and the value they may place (or not) on accountability in general and accountability to the beneficiaries in particular as a key determinant and characteristic of interventions and interactions” (Davidson 2002: 42). In this way, Kennedy (2019) argues, self-regulatory initiatives serve as vehicles for reforming humanitarian subjectivities or, in constructivist parlance, the rules are simultaneously regulatory and constitutive. But what kind of identity was inculcated? Everett and Friesen (2010) point to HAP’s embrace of quality assurance, suggesting that “by relying on a new, commercial script the organization is helping re-form its members’ identities as commercial identities” (Everett and Friesen 2010: 476). Though HAP subsequently dropped the term “beneficiary,” it reasserted its commitment to quality assurance—to enabling organizations to meet the “needs and expectations of its customers” (HAP 2010c: 6–7).

I raise two additional issues in concluding. First, the impact of personality and leadership style emerged as a surprise finding. Over one-third of interviewees spoke, unprompted, of the temperament of key figures and of HAP’s institutional culture, narrating them as obstacles or barriers to the broader acceptance of HAP’s model. As much as the message, the messenger shaped perceptions of beneficiary accountability. Second, from its inception, the INGO accountability project has been marked by uncertainty and even serendipity. The Rwandan genocide and JEEAR process provided critical junctures, bringing conversations about humanitarian reform from the margins to the center. An isolated mention of an “ombudsman” was taken on by British INGOs, shaping the early direction of the initiative. The initial field research was conducted under conditions of uncertainty in Kosovo, while the final direction—toward quality assurance—owed greatly to a change in leadership and location. HAP owed its existence to the concerted advocacy efforts of key figures, certainly, but it is equally clear that they did not act on a terrain entirely of their own making. Powerful societal and institutional forces have enabled the spread of management and audit practices (Power 1999). A respondent put it as follows: “Study 3… put the idea on the table and to some extent did put NGOs on notice that if they didn’t come up with something themselves that sooner or later the donor community would come up with ideas of their own. Of course if it hadn’t been for John Mitchell and his colleagues in the British Red Cross…, the JEEAR recommendation could have fallen on infertile ground and the idea disappeared” (O4). HAP was the contingent outcome of uncertain circumstances.

Uncertainty is profoundly uncomfortable. As Katzenstein and Seybert (2018: 29) note in a recent volume, “it cuts against the grain of institutional and organized life in the twenty-first century… Our risk-based thinking expresses a deep desire for and faith in control.” But uncertainty is also generative. It “can create conditions ripe for improvisation. It can also incite unexpected innovations as political actors try to make the future meaningful by linking the self to something bigger than its singular, present existence” (Katzenstein and Seybert 2018: 40). So it was with HAP International. The trauma of Rwanda created conditions of possibility for beneficiary accountability. The ensuing inter- and intra-agency tensions, while “grossly irritating” (B4) from an organizational perspective, were also generative of innovative ideas and practices. Today, through the CHS, the HAP approach is a core element of contemporary humanitarian practice. But, as this research has demonstrated, the CHS is unlikely to mark the final resolution in the debates over humanitarian accountability.