The ADB's Accountability Mechanism: An uphill battle for communities

When the Asian Development Bank (ADB) emerged in the 1960s, the premise was that developing countries’ priorities, especially in the Asia-Pacific region which was then amongst the world’s poorest, required moving away from Western-dominated financial institutions and establishment of a regional bank. Proponents of this new bank believed that it could supplement the Bretton Woods Institutions, which were mired in controversies around governance and voting rights. By 2024, ADB has grown to become the largest Multilateral Development Bank (MDB) in the region in terms of project numbers, overtaking its counterparts, the International Finance Corporation and World Bank (which are placed at a distant second and third respectively) and finally, the Netherlands Development Finance Company (FMO). 

Since its inception, ADB has admittedly done some things differently - including consolidating its capital base to ensure larger financial capacities,¹ promising to address not simply economic growth but also inequalities, centering gender in its workstream and as an operational priority, and more recently, even taking on the mantle of being “Asia-Pacific’s climate bank”, committing to increase its financial capacities and disbursements specifically for climate finance purposes. From January 2016 to October 2024 (our period of analysis), ADB investments in the South Asian region² stood at a ball-park figure of USD 63 billion. ADB had invested in or approved investments for close to 834 projects during this time (projects approved, active, closed, and completed), of which some were transboundary infrastructure projects. 

Background

Despite its vision however, ADB has consistently shown that it is not very different from its contemporaries, both, in terms of its governance and decision-making processes, and also in its operations and project implementation. In South Asia specifically, ADB invested in 834 projects (excluding proposed and cancelled projects). Between the same time (2016-2024), ADB received over 180 complaints to its Accountability Mechanism (AM) from the region, also the highest amongst the MDBs.³ The Accountability Console, a database that tracks complaints to MDB accountability mechanisms from across the globe, found that most of the ADB complaints were “closed without outputs”, meaning that most of the complaints failed to produce any investigation report, or an agreement between the parties (complainants, borrower-clients, and the MDBs). 

But why are these couple hundred complaints, against the thousands of ADB-financed projects, that significant? For one, community or CSO complainants may not be fully aware of the presence of the AM itself, since it is unclear how many borrower clients consistently disclose the existence of the AM, and how extensively the ADB monitors this. The number of complaints that evoke lack of access to project information, and the AM’s own acknowledgement in its Impact Assessment that the lack of information is a possible reason for the low number of complaints processed, could indicate not very. But let us move to the complaints that do find their way into the system. 

Once they are into the system, they face another hurdle their way - the eligibility requirement. 

Crossing the great barrier: Eligibility Assessment

Usually on receipt of complaints, the AM screens it for admissibility where it adjudges if the complaint falls within its mandate. This includes assessing that it does not relate to procurement or corruption issues, is not frivolous or anonymous, and is made by at least two complainants. Next, the complaint is forwarded to the Office of the Special Project Facilitator (OSPF) or the Compliance Review Panel (CRP), which are the two separate offices within the AM. Depending on whether the complainants opt to solve their problems, or focus on an inquiry into whether the ADB has complied with its own policies, the complaint moves within these two offices respectively. 

The complaint is screened for eligibility, where the AM adjudges if the substance of the complaint is related to the actions or omissions of the Bank in terms of formulating, processing, or implementing the project, and is serious enough to warrant further processing of the complaint (later, for compliance review). While these make up the criteria for finding eligibility, a complaint can also be excluded from eligibility on the ground that the complainants did not make a “prior good faith efforts” with the ADB. Finally, the OSPF and the CRP produce their opinions in the form of Eligibility Assessments. Where complaints are found ineligible, the OSPF/CRP forwards them to the concerned Operations Department for resolution at their level, or at the project-level.

At least 52 out of the over 180 complaints have been found ineligible on the lack of “prior good faith efforts” grounds.⁴ Other reasons include, (1) the Operations Department is working with the concerned authorities in order to resolve the complaint (with or without involving the complainants also), (2) complainants are not directly affected peoples, or not foreseen to be project affected, (3) the complaint is not within their mandate (without further explaining why), (4) the AM’s intervention is “unhelpful” (again, not further explained), (5) the complaint was filed beyond limitation periods, and (6) the existence of a parallel complaint with another mechanism.

Good faith efforts: But only for the Complainants?

At one of the AM’s Outreach meetings recently, participants - all project affected peoples or those who stood to be potentially affected, and wanted to bring complaints to the AM - were curious about what this “prior good faith efforts" implied. The response was, “reaching out to any person within the ADB was enough”. But in practice, demonstrating “prior good faith efforts” has not been as easy as the AM claimed in its Outreach exercise. In practice, the AM has often interpreted “prior good faith efforts” in an overly technical way, even where complainants’ have demonstrated genuine efforts. 

For instance, the AM excluded the complaint, simply because complainants did not specifically reach out to the Operations Department (also, this case complaint), or, the Department did not respond to their complaint. This puts an additional onus of engagement on the complainant, of either reaching out to additional actors, or of waiting for a response even while they continue to suffer project related violations and harm (whereas interpreting good faith engagement would have required the concerned staff to respond).

In the Rajasthan Urban Sector Development Program case, the complainants reached out to the Operations Department, admittedly after submitting the complaint (so, not “prior”), but as soon as they were apprised of the good faith policy, in their first meeting with the OSPF. They received no further response from the Department during the length of the complaint. The OSPF possessed special information that the Operations Department was already made aware of the issues, and was making “internal efforts” that were never even communicated to the complainants. OSPF closed this complaint despite the acknowledgement that there was nothing further that the complainants could do in this case, and needless to say, they received no further response even as of the date of complaint closure.

In the Shapoorji Affordable Housing Project case, a complaint where amongst others, complainants claimed involuntary resettlement and forcible land acquisition, the CRP dismissed the complaint on grounds of a single technicality - the complainants had not elaborated all of their “distinct and shared concerns”. This raises several concerns: (1) the complainants had already raised their issues multiple times with the project authorities over a two-year period - first in 2020, with the formal complaint submitted in 2022;  (2) the Operations Department itself acknowledged that the issues raised in the complaint before the CRP matched those they had previously received; and, (3) it prompts the question - should the CRP have required such legal exactedness of the complainants? If the purpose of demonstrating “prior good faith efforts” is to alert the Operations Department and provide an opportunity to address the issues, then how should it be interpreted? Is the AM’s insistence on being the “last resort” ultimately counterproductive? And what becomes of all these internal lessons that the AM highlights after every complaint, admissible or not. Is the Operations Department failing to learn from these complaints, especially when multiple complaints reveal recurring issues and patterns for concern with respect to the same project?

To its credit, in some cases, the AM did forward these complaints with the added observation on the Operations Department’s crucial role in enablement of the good faith standard, and how this was pending in the immediate case, but that does not do much for a complaint that is still dismissed. 

In some cases, however, the AM actively jeopardised the complainant’s claims. For example, in the Nepal Decentralised Rural Infrastructure and Livelihood Project complaint, while still finding it ineligible, the OSPF in its assessment went on to echo the Management, reiterating the benefits that the industrial road project posed for the nation as a whole. This implication that the project’s potential benefits offset the negative impacts relied solely on the project completion reports and mission reports. For that matter, these reports also mentioned the issue of compensation being delayed, which was also the core of the SPF complaint, but the AM conveniently ignored them. Complainants had already raised their concerns with both state authorities as well as the ADB staff. 

Which raises the question - should eligibility - the first stage of complaint - be such a pain-staking process?

When “good faith” becomes too late

There is also the very real possibility of complaints turning futile by this technical exactedness of the AM over its “prior good faith efforts”.

In some cases, the complainants engaged over project related issues with the Operations Department and the ADB team repeatedly, without response, until the matter was brought to the court and was no longer eligible, or the company went insolvent and the loan closed.  

In the former scenario, the OSPF denied eligibility on the grounds that the matter was not taken up with the Operations Department specifically (even though complainants had informed the executing agency, ADB project officers). Eventually, when the period for rationalising its ineligibility decision came, the SPF also claimed that the project team sought to close it on their part, now that the matter was before the courts (“sub-judice”). But there is no provision under the AM’s Policy currently that excludes sub-judice matters as ineligible. Moreover, the complaint brought before the SPF went beyond the specific issue subjudice (ownership of land). On the other hand, the project team seemed willing to re-consider the executing agency’s request to continue with the project nevertheless. 

In the latter scenario, for over 3 years, between 2014 and 2018, complainants who were aggrieved by serious labour rights violations, continued approaching the project level authorities, state authorities and eventually, the ADB project team. They even appeared at the ADB’s Annual General Meetings in an attempt to convince the Board of Directors to take action. By the time the complaint finally came to the OSPF, the project proponent had turned insolvent and the project had passed onto the government. Since the period of 2 years had passed since the loan closing, there was nothing that the OSPF could do - except, make a comment “strongly encouraging” discussions between ADB Management and the complainants. 

For some cases, requiring good faith also defies logic. For example, in several cases, complainants might have resubmitted complaints to the AM over the same project, only to be dismissed on account of some technicality. Or, they might have submitted complaints to the responsible executing authorities, filed court complaints, and subsequently filed a complaint with the AM desiring the exercise of compliance review function, since the project issues were never rectified. In two such cases, complainants had even received court orders in their favour, orders which clearly vindicated the illegalities that the project suffered from. Despite that, the executing agency with support from the ADB Management, continued with the project, in what could be termed as a clear contempt of court process (see cases Power Distribution Enhancement Investment and Rajasthan Renewable Energy Transmission Investment Program). These blatant violations of even the national laws as the minimum standard highlight serious gaps in the exercise of monitoring and supervision obligations of the ADB team. Or in other words, they raise non-compliance concerns. 

The AM in its adjudication should consider the immediacy that the issues demand, and the fact that often project complainants try different avenues that are more immediately available / tangible before filing complaints. For example: in several cases, complainants reached out multiple times to the executing agencies, district officials, or even lodged cases before judicial systems, without receiving response or redress. At least from a compliance review perspective, the refusal from the project authorities to rectify Safeguard violations that pose negative impacts for communities, should indicate a preliminary lack of supervision by the ADB team over the project’s “processing and implementation” - an observation which as it appears, is possible to make, if willing. The AM should realise the enormity of efforts that it takes on part of the communities to come together to file a complaint to such a mechanism in the first place, as communities are often told that their efforts are in vain, or would result in them losing the little benefits or compensation they were offered initially, or simply, to avoid confrontation with officials that might exacerbate pre-existing risks. 

What precludes further scrutiny of the AM’s complaints processing is a lack of adequate rationale - at least one that is publicly available, and the lack of all complaint-related information (minus the identification of complainants that would be validly retracted in case of request, or threat of reprisal). Although this lack of information goes beyond complaint and rationale for ineligibility, this is a matter for another article. In general though, this lack of information makes it difficult to assess whether the ineligibility was justified, or whether the complainant truly achieved redress for their issues. Additionally, it smacks of a lack of predictability around the AM process, which contravenes the effectiveness criteria for non-judicial grievance mechanisms, per the UN Guiding principles. 

It is unfortunate that so far within the region, the few instances in which the AM has provided some meaningful result for the community, or overseen the provision of such result through its tracking of the complaint, once forwarded to other departments, is limited to a mere 13 out of over 180 complaints. These have taken the form of realignment of parts of the project, information clarification/disclosure, agreements towards renewed consultations, re-valuation of land rates or revised studies and assessments, increased compensation rates, and plenty of internal lessons.⁵ This does not however imply that all agreements have been realised, or even further that communities have achieved remedy. Imagine if for a Bank-financed project that expropriates your lands to construct a power plant, that already displaces, uproots livelihoods, and the social fabric of the community - no amount of renewed consultations, or internal lessons on instituting more grievance mechanisms, can effectively remedy that which the community has already lost. The AM as it itself admits, does not have the power to suspend project works in the interim, nor are its recommendations enforceable. For the limited results that it can achieve though, it should not be gatekeeping the AM, if its objective is as it claims, to provide a “forum for those affected to voice their concerns”. 

Instead, the AM should ensure that it does not interpret the already high barrier that is set for “prior good faith efforts”, in a manner that further precludes accessibility for complainants. Instead, as a mechanism dedicated to both, internal as well as external accountability for ADB projects, it should ensure that its eligibility determinations are context-dependent. This could include, assessing the viability of raising issues with project authorities or the ADB Operations Department, the possibility of retaliations, the fact that the project loan or project construction is at an advanced stage, strong and independent evidence supporting the complainants’ claims. Instead of dismissing complaints outright, it should see how it can support complainants in overcoming an otherwise technical exclusion that has no bearing on the substance of their claims - for instance, by actively working with the complainants to remedy this defect, or even when it forwards the complaint to the Operations Department, by taking regular updates on the resolution of cases. 

Note: The AM is currently undergoing review, and even the external reviewer’s report for Phase 1 of the AM Review acknowledged the “prior good faith efforts” requirement as an issue to accessing the AM, and the recommendations of the Good Policy Paper, published by 11 international organizations on IAM policies, that this should not be a pre-condition to filing complaints.

ENDNOTES

[1] This includes for example, combining its ordinary and special fund capital bases for increased lending, and risk-bearing, promoting financial instruments like blended or syndicated finance, exploring co-financing, reinvesting earnings to not rely heavily on additional external capital infusions.

[2] According to ADB, Pakistan is a part of the Central-West Asia Regional Department, but for purposes of this Research, I have referred to the South Asia region geographically* and not in terms of the ADB’s Operation Department divisions. 

[3] The number of ADB projects and complaints filed to the mechanism cannot be exact equivalents for several reasons since complaints against newer projects might only emerge later in the project cycle, meaning that there might be projects approved before the 2016 cut-off period that are currently seeing complaints, or the number of complaints might also be relatively lower if newer projects have since come up for which complaints might only be lodged later than our period of analysis. Additionally, the 181 complaints is also an estimated figure as some of the complaints are only recorded on the AM Registry and not on the SPF-CRP complaint sites.

[4] Note that out of the 180 complaints, there are also complaints where full complaint related information as well as rationale for ineligibility might not have been available. 

[5] This is with some exceptions, like the Malir expressway sub-project under the Supporting Public-Private Partnership Investments in Sindh Province, where the complaint directly or indirectly led to discussions between the Operations Department and the Pakistan government, clarification of the projects’ non-alignment with ADB’s priorities as a climate bank, and the government agreeing to take the project off their proposed projects for ADB funding. See the assessment.