The Paradox of Evaluation in International Development

In 2019, a major international NGO completed a rigorous evaluation of its flagship education program in South Asia. The findings were sobering: test scores had barely improved, teacher training was ineffective, and community engagement was far lower than reported. The evaluation team prepared a comprehensive report with actionable recommendations. Six months later, the organisation's annual report highlighted the program as a "great success," citing carefully selected metrics and anecdotal stories.

While this is a hypothetical scenario, it illustrates a real contradiction that often plays out across the development sector.

Evaluation is supposed to help development organisations understand what is working, what is not, and how to do better. Yet, more often than not, evaluation remains trapped in a paradox, shaped by institutional incentives, donor relationships, and the broader political economy of aid.

In today's changing landscape, this paradox matters more than ever.

The Basic Model and Its Inherent Tension

Most international development organisations follow a familiar cycle: they identify a critical development problem, build donor confidence that the organisation is well-positioned to address it, mobilise funding, and implement solutions. This is how development work is sustained — and here lies the paradox.

The very act of sustaining this cycle depends on maintaining donor confidence. Yet the role of evaluation is to surface gaps — to show where programmes fall short, what is not working as intended, and what could be improved. In an environment of tightening funding, openly acknowledging such gaps can feel risky, potentially undermining the very confidence that keeps the work going.

As a result, evaluation can become trapped in a defensive posture — used to manage reputational risk, rather than to drive genuine learning and accountability.

The Observation–Reality Gap

One factor that reinforces the evaluation paradox is the wide gap between programme realities on the ground and what donors can observe — the observation–reality gap.

In the private sector, the relationship between producer and consumer is direct, and feedback loops are fast. If a product fails to meet expectations, paying customers (the equivalent of donors) provide immediate signals, and companies (like development organisations) can quickly adapt.

In development, the dynamic is very different. Donors, as the primary paying customers for development work, fund programmes in distant contexts with limited direct visibility into actual performance. They often do not have the means to easily verify whether the “product” provided by the development organisation — improved outcomes on the ground — has truly met expectations.

Because donors cannot experience outcomes first-hand, they tend to place disproportionate weight on the few formal signals they do receive, including evaluation reports. When funding is uncertain and trust must be actively maintained, this heightens the perceived risks around evaluation. Candid findings that expose major problems may shake donor confidence, not because of malice or mistrust, but because donors might lack the contextual understanding to interpret these findings in a balanced way.

As a result, organisations face strong incentives to manage the message and strike a perceived “balance” — but often at the expense of candour and of the core intention of optimising programme delivery. In practice, this means evaluation reports may lean toward a positive or cautiously balanced tone, even when programmes underperform, to preserve donor confidence and protect funding relationships. When visibility is low and funding is fragile, evaluations risk becoming defensive tools — serving institutional risk management, rather than engines of genuine learning and adaptation.

Moreover, the actors intended to provide an independent perspective — such as think tanks, research institutions, and consultancies — are also competing for visibility, funding, and influence within the same political economy of aid. Their incentives are often aligned with those of the organisations commissioning the work, which can reinforce risk aversion and further narrow the space for unvarnished reflection.

Beyond Comfortable Narratives

There is an uncomfortable truth we rarely acknowledge: Over time, the development sector has become skilled at telling a convenient story.

In this story, development challenges are always serious, but never so difficult that more funding, time, or coordination cannot solve them. This narrative keeps everyone comfortable.

But this narrative holds us back.

Development is not just a delivery problem; it is a learning problem. Our institutions, like the work itself, are imperfect, adaptive, and continuously evolving.

There will always be missteps. Programmes will underperform, assumptions will prove wrong, and contexts will shift in ways we did not anticipate. These are not failures of intent — they are part of working in complex environments.

The challenge is not whether to accept imperfection, but how to build systems that learn from it.

Moving forward will require fundamental shifts in culture, incentives, and practice that go beyond the scope of this piece, but are truly essential if development institutions are to become fit for purpose.

Building a genuine learning culture — with evaluation at its heart — must be part of that transformation.

These reflections draw from ongoing conversations with colleagues across the development community. They represent one perspective in an ongoing dialogue about how to strengthen development institutions for the challenges ahead.

#InternationalDevelopment #Evaluation #OrganizationalLearning #AidEffectiveness #DevelopmentPractice