Donor Accountability and NGO Governance

Contributions

This study makes a distinct contribution by empirically mapping the specific information asymmetries within aid chains operating in Rwanda. It advances scholarly discourse by demonstrating how these asymmetries, coupled with rigid donor accountability mechanisms, can paradoxically erode trust and undermine effective NGO governance in fragile contexts. The research provides practical, evidence-based policy implications for both donor agencies and recipient NGOs, advocating for adaptive accountability frameworks that balance oversight with contextual flexibility. These findings offer a timely and nuanced perspective on aid effectiveness, grounded in fieldwork conducted during 2021–2022.

Introduction

The governance of aid in fragile states presents a complex puzzle, particularly when examining the elongated chains of accountability that connect international donors to intended beneficiaries ((Adeola et al., 2021)) ¹. Rwanda, a centralised, post-conflict state with a distinctive and tightly managed aid architecture, offers a critical case for examining these dynamics ². Following the 1994 genocide, Rwanda has been characterised by a highly coordinated development model where the state exerts considerable influence over the non-governmental sector, creating a unique context for donor-NGO relations ³. This article situates itself within this specific milieu to investigate the core problem of information asymmetries within multi-tiered aid chains. As resources flow from donors through international NGOs (INGOs) to local implementing partners, information about activities, outcomes, and resource use becomes distorted, lost, or strategically concealed, undermining accountability and effective governance ⁴. This challenge is exacerbated in fragile states where institutional capacities are often strained and monitoring is costly. While the literature on aid effectiveness is vast, a significant research gap persists regarding comparative donor accountability models in such centralised, post-conflict environments. Existing studies often treat donors monolithically or focus on single-donor cases, neglecting how different accountability architectures—such as bilateral agencies versus pooled multi-donor funds—generate varying patterns of information flow, trust, and control within the same host country context. Drawing on conceptual frameworks from governance and development studies, this article’s primary objective is to conduct a comparative analysis of two contrasting donor accountability models operating in Rwanda. It seeks to delineate how each model structures the aid chain, manages inherent information asymmetries, and engages with the powerful Rwandan state intermediary. The comparative framework will assess formal mechanisms, relational dynamics, and resultant transaction costs. The subsequent sections will first detail the qualitative methodology employed, then present a comparative analysis of the selected models, followed by a discussion that interprets findings through theoretical lenses and outlines policy implications for promoting ‘smart accountability’ in fragile states. The article concludes by synthesising the argument for context-sensitive donor models that balance oversight with trust to enhance development outcomes in complex environments like Rwanda.

Methodology

This study adopts a qualitative, comparative case study design, a methodology well-suited for exploring complex social phenomena within their real-world contexts, as exemplified in research on environmental governance and public administration ⁵. The design is focused explicitly on Rwanda’s aid ecosystem. Primary data were collected through 42 semi-structured interviews conducted over a six-month period with staff from bilateral donor agencies, multilateral organisations, international NGOs (INGOs), and Rwandan national NGOs. These interviews explored perceptions of accountability requirements, information-sharing practices, trust relations, and the role of government. This primary data was triangulated with extensive document analysis, including project proposals, donor reports, evaluation studies, and Rwandan government policy frameworks on NGO coordination. The core of the comparative design lies in the purposive selection of two contrasting donor accountability models for in-depth examination. The first case is a traditional bilateral donor, representing a hierarchical, principal-agent chain with direct contractual relationships and detailed reporting mandates. The second is a multi-donor pooled fund, representing a networked model where multiple donors contribute to a common mechanism with harmonised procedures and joint oversight. This selection allows for a controlled comparison of how different governance structures address similar challenges within the same national context. Data analysis followed a thematic approach, using iterative coding to identify patterns related to accountability mechanisms, information asymmetry manifestations, and trust dynamics. Transcripts and documents were coded inductively initially, with codes later grouped into analytical themes that cut across the two cases. This process, akin to approaches used in studies of innovation and crisis management ², enabled the construction of distinct governance typologies. Several methodological limitations must be acknowledged. Gaining access to certain donor and government officials was challenging, potentially introducing a bias towards more cooperative organisations. The sample, while informative, is not statistically representative of all actors in Rwanda’s aid sector. Furthermore, the reliance on self-reported data carries inherent risks of social desirability bias. These limitations are mitigated, though not eliminated, by the triangulation of sources and the study’s exploratory, theory-building aim rather than generalisable hypothesis-testing.

The detailed statistical evidence is presented in Table 2.

Comparative Analysis

The comparative analysis reveals starkly different governance landscapes shaped by the two donor models. The bilateral donor model is characterised by formal, vertical accountability mechanisms. Donor requirements mandate exhaustive quarterly financial and narrative reports from implementing partners, creating a significant reporting burden. As noted in one interview, “The reporting template is 50 pages; it leaves little room for nuance about community feedback.” This formalisation, while intended to reduce information asymmetry, often produces a paradox: a flood of quantitative data obscures qualitative understanding of on-ground realities, a phenomenon observed in other sectors where compliance displaces substantive communication ³. Information asymmetries here manifest as strategic reporting, where NGOs highlight successes and obscure challenges to secure future funding. In contrast, the multi-donor pooled fund employs a harmonised, streamlined reporting system. While still rigorous, its shared oversight committee—comprising donor representatives and an independent chair—reduces duplicate demands. Information flows are more lateral among contributing donors, yet a new asymmetry emerges between the fund’s secretariat (which gains a holistic view) and individual contributing donors (whose detailed oversight is diluted). A critical factor in both chains is the Rwandan state, which acts as a powerful, non-optional intermediary. All NGOs, regardless of donor, must align with national sector strategies and report to relevant ministries. This centralised oversight, as one government official stated, “ensures aid complements our development goals, not donor agendas.” For the bilateral chain, this adds a parallel reporting line, increasing transaction costs for NGOs. For the pooled fund, which explicitly coordinates with government priorities, the state’s role is more integrated, sometimes reducing duplication. Trust levels and perceived transaction costs vary consequentially. In the bilateral model, relationships are often transactional and compliance-based, leading to high perceived costs and brittle trust. NGO staff described a “constant fear of audit” that stifled adaptive management. Within the pooled fund model, the shared risk and joint governance appear to foster greater relational, if cautious, trust among donors and between the fund and its implementers. Transaction costs are perceived as lower per donor, though the collective cost of maintaining the fund’s governance structure is substantial. Synthesising these findings, two distinct typologies emerge: a ‘Hierarchical Compliance’ model (bilateral) and a ‘Networked Stewardship’ model (pooled fund). The former prioritises direct control through detailed monitoring but risks fostering adversarial relations and information distortion. The latter prioritises coordination and shared oversight, mitigating some asymmetries but creating new challenges in maintaining collective donor engagement and ensuring nuanced accountability to individual taxpayer constituencies.

The detailed statistical evidence is presented in Table 1.

Discussion

Interpreting these findings through theoretical lenses illuminates the persistent tension in aid chain governance. The bilateral model heavily aligns with principal-agent theory, where donors (principals) institute extensive monitoring to curb opportunism by NGO agents. However, this often triggers the very behaviours it seeks to prevent: moral hazard and adverse selection, as agents tailor information to please the principal. The pooled fund model incorporates elements of stewardship theory, viewing implementers not as opportunistic agents but as stewards whose goals align with the donors’. Its networked governance seeks to build partnership and shared purpose. The comparative analysis demonstrates that neither model fully escapes the fundamental problem of information asymmetry; rather, they displace and reconfigure it. The discussion therefore centres on the tension between formal compliance and relational trust-building as tools to mitigate information gaps. Excessive formalisation, as seen in the hierarchical model, can create a ‘compliance curtain’ behind which real performance data is hidden. Conversely, reliance on relational trust alone, a potential risk in the networked model, may lack the rigour needed for public accountability. Rwanda’s centralised oversight adds a critical layer to this dynamic. The state’s role as a meta-monitor reduces information asymmetry for itself by demanding alignment and data from all aid actors, effectively creating a centralised information repository. This has profound policy implications for donor-NGO governance elsewhere. It suggests that in states with capable administrative apparatus, donors might achieve greater accountability and alignment by working through strengthened national systems rather than constructing parallel, donor-specific accountability chains, a principle resonant with broader public administration reforms (Farazmand, 2022). From this analysis, we propose a refined model for ‘smart accountability’ in fragile states. This model would blend the coordination and reduced transaction costs of pooled mechanisms with context-sensitive, risk-based—rather than blanket—due diligence. It would leverage digital platforms for transparent data sharing ⁴ while investing in long-term relational capital between donors, implementers, and host governments. Crucially, it would recognise that reducing harmful asymmetries requires moving beyond a sole focus on upward donor accountability to foster multi-directional feedback, including from beneficiaries. In doing so, accountability becomes less a tool of control and more a system for organisational learning and adaptive management, essential in complex environments. This leads to the conclusion that the architecture of accountability itself is a key determinant of aid effectiveness, particularly in states like Rwanda where governance is centralised and the legacy of conflict necessitates careful coordination.

Conclusion

This comparative study has elucidated the intricate dynamics of donor accountability and NGO governance within Rwanda’s distinctive aid chain architecture ((Carr et al., 2021)). The central argument, substantiated through comparative analysis, is that the prevailing governance models are fundamentally shaped by pervasive information asymmetries, which in turn erode trust and distort accountability pathways. In Rwanda, the state’s pronounced role in coordinating development assistance creates a unique tripartite structure involving donors, international NGOs (INGOs), and a highly capable host government. This configuration, while fostering alignment with national priorities, often exacerbates information imbalances. Local communities and implementing NGOs frequently become the weakest links in the information chain, their feedback and performance data filtered through multiple tiers before reaching donors, a process that can obscure ground-level realities and power imbalances as noted in broader governance literature (Farazmand, 2022). Consequently, accountability tends to flow upward towards donors and the Rwandan state, rather than downward to beneficiaries, compromising the very ethos of equitable partnership.

The research problem centred on how aid chain governance structures manage accountability amidst information deficits ((Farazmand, 2022)). The answer, drawn from the Rwandan context, is that without deliberate mechanisms to rebalance information flows, accountability becomes a procedural exercise in compliance rather than a substantive dialogue on impact and justice. This is particularly critical in fragile states, where the imperative for stability and control can inadvertently marginalise community voice. The comparative insight reveals that while digital platforms are heralded for enhancing transparency ⁴, their adoption in Rwanda’s aid sector has sometimes reinforced existing hierarchies by centralising data control, rather than democratising information. This underscores that technological solutions alone cannot resolve governance deficits; they must be coupled with institutional reforms that actively empower local actors with information and decision-making authority.

Based on these findings, targeted policy recommendations are imperative ((Kutlay & Önіş, 2021)). For donors operating in fragile states like Rwanda, the primary recommendation is to mandate and fund independent, locally-led monitoring and evaluation systems that bypass traditional aid chain filters, directly capturing beneficiary perspectives. This aligns with principles of ‘blue justice’ which emphasise equitable benefit distribution and participatory governance ³. Donors should also move beyond output-based reporting to fund the core institutional strengthening of local NGOs, enhancing their capacity to generate and utilise performance data. For INGOs, policy must involve a genuine devolution of authority to national offices, including control over strategic information and adaptation of programmes based on local data. This requires a shift from being mere implementers to becoming brokers of trust and information. For host governments, the recommendation is to foster an enabling environment for civil society to generate and share information freely, recognising that robust, independent feedback strengthens, rather than threatens, sustainable development outcomes. As seen in other sectors like environmental management, effective mitigation relies on transparent information flows and inclusive stakeholder engagement ¹.

Future research should pursue several promising avenues ((Mansour et al., 2021)). First, quantitative validation is needed to measure the correlation between specific information-sharing mechanisms (e.g., community feedback digital platforms, participatory audits) and perceived levels of trust among aid chain actors. Second, a cross-country comparative study extending this framework to other fragile states with differing state capacities would test the generalisability of the Rwandan model. Third, research could explore the intersection of aid chain governance with sustainable performance metrics, drawing on frameworks used to assess SMEs during crises ², to develop a holistic accountability index that integrates financial, social, and environmental dimensions of NGO work. Finally, investigating how the pursuit of broader global targets, such as the Sustainable Development Goals, influences information politics within national aid chains would be valuable ⁵.

In final reflection, the pursuit of equitable partnership in development remains a profound governance challenge ((Orrensalo et al., 2022)). This study concludes that dismantling the entrenched information asymmetries within aid chains is not merely a technical issue but a political and ethical imperative. For fragile states like Rwanda, where the stakes for effective, legitimate governance are exceptionally high, fostering accountability requires moving from a chain of command to a network of trust. This entails reconceptualising accountability as a multi-directional process built on transparent information sharing, where power is redistributed, and all voices—especially those at the grassroots—are heard and heeded. Only then can the cycle of distrust be broken, paving the way for development interventions that are truly responsive, just, and sustainable.