When the Trump administration announced aid cuts, the development sector erupted in alarm. Billions of dollars at stake. Communities at risk. Progress threatened. But what progress, exactly?
I have watched hundreds of projects over the past two decades, and I can count on one hand those that left anything behind once the funding stopped. The rest vanished. Not immediately, but in a slow dissolve that followed a predictable pattern. For two or three years, the project would hum along, producing activity at an impressive clip. Workshops conducted, trainings delivered, stakeholder consultations held, glossy assessments printed, conferences convened. The donors were reassured, the reports stacked up, and the photos of smiling women filled the brochures.
Then the money ran out. Within months, the trained staff had moved on, the equipment sat unused, and the new systems quietly collapsed. Farmers returned to the methods they had used all along. Women went back to the same small enterprises they had run before the grant. Health workers kept showing up, but the monitoring systems, the referral networks, the capacity building protocols were already gone. Two years later, a new baseline survey would discover the same problems in the same communities. A new project was designed, a new set of consultants arrived, a new cycle began, slightly rebranded but fundamentally unchanged.
Here is what the vanishing looks like up close. A 2.8 million dollar agricultural livelihoods project trains 650 smallholder farmers in climate-smart techniques. For three years, field officers have visited regularly. Demonstration plots flourish. Farmers attend sessions, receive inputs, and sign attendance sheets. The project's mid-term evaluation glows with achievement. Then the funding cycle ends. I return eighteen months later. Not a single farmer is using the techniques. The demonstration plots have reverted to traditional crops. The equipment, a soil testing kit, a small processing machine, sits rusting in a storage room of the district agriculture office. When I ask what happened, a farmer shrugs. "They taught us things for their crops, not ours," he says. "The seeds cost too much. The market they promised never came."
The project's final report, printed on heavy stock paper with a glossy cover, sits in a donor's library somewhere. It documents 650 farmers trained, 14 demonstrations conducted, 4 manuals produced, and 22 workshops held. All true. All irrelevant. This is not failure. This is the system working exactly as designed.
The development sector has perfected something remarkable: the production of activity that resembles progress. Consultants arrive and conduct needs assessments, never mind that identical assessments were done three years prior for the previous project. Workshops are held where participants dutifully fill feedback forms rating the facilitation. Training modules are developed, printed, and distributed, then locked away because of "data privacy" or "intellectual property concerns." Knowledge products accumulate in hard drives and archives, never to be referenced again. Meanwhile, the same problems persist. The same photos appear in new proposals: women carrying water from distant wells, children in crowded classrooms, farmers working exhausted soil. Sometimes, literally the same photos, recycled across projects and years. The women's faces change, the framing adjusts, but the underlying reality, the poverty, the unmet needs, the laborious survival, remains frozen in place.
What looks like failure is actually the system succeeding at its true purpose: not solving problems, but producing the appearance of solving problems. Not creating development, but creating evidence of development work. The difference is everything.
Consider what disappears when a project ends. Not just the funding or the staff, though those vanish reliably. What disappears is knowledge. The field officer who spent three years learning which farmers were actual innovators and which just performed for visiting officials, gone to another job. The community health worker who figured out which households would never use the new clinic because of an unspoken conflict with the nurse moved to an NGO two districts away. The project manager who finally understood why women would not attend evening meetings retired. This knowledge never gets written down because it contradicts the project's official theory of change. It never makes it into the "lessons learned" annexe because it suggests the project was designed by people who did not understand what they were designing. So it simply evaporates. The next project starts from zero, destined to learn the same lessons at the same cost, only to lose them again when its funding ends.
But here is the deeper problem: the knowledge that could have prevented the failure in the first place, the knowledge possessed by the farmers, the women, the health workers themselves, never entered the system at all. Every vanishing project shares one feature: it was designed by people who will never live with its consequences and implemented on people whose knowledge never shaped what was built.
The project design happens in capital cities or foreign offices. Consultants, often quite intelligent, sometimes well-intentioned, sit with laptops and theories and donor priorities and devise interventions. They hold consultations, yes. But these happen after the framework is set, after the budget is allocated, after the theory of change is locked. The farmers are invited to "validate" a plan already drawn. The women are asked to comment on a training schedule already printed. This is not a consultation. This is theatre. The purpose is not to learn but to legitimate.
The farmers in the agricultural project knew from the beginning that the introduced seeds were wrong for their soil and that the promised market linkages would never materialise in their remote location. The women knew the evening meetings would fail because that is when they must cook for their families. The health workers knew the new protocols would conflict with how the community actually seeks care. But none of this knowledge ever travelled upstream to the designers because the structures of decision-making had no channels for it to flow through. The project proceeded on the planners' theories. It failed on the communities' reality.
Even when the right people are present, power tilts the conversation. Imagine a bureaucrat from the capital holding a meeting under a village tree. The farmers know this man controls permits, subsidies, and inspections. They weigh every word against the risk of losing access or drawing unwanted scrutiny. Crucial knowledge, the kind that could explain why the last irrigation scheme failed, remains unspoken. Reverse the scene and fly the farmer to a hotel workshop in the city. Around him, consultants flip through PowerPoints in fluent aid jargon. He knows his soil, his microclimate, the rhythms of rain. But he does not know how to translate decades of lived experience into the categories of a logframe or the language of indicators. His silence is taken for ignorance, when in fact the failure belongs to the room itself, which never created the conditions for his knowledge to count.
Why does this persist? Because everyone involved has incentives that point away from acknowledging it. The donor needs to show Parliament or Congress that funds are being deployed. Workshops conducted and farmers trained are easy to count and photograph. Actual transformation of agricultural practice takes a decade to assess and is devilishly hard to attribute. The implementing organisation needs to show the donor that funds were spent as promised and targets were met. The targets are about activities, not outcomes, so the organisation optimises for activities. Hold the workshops. Print the manuals. Take the photos. Submit the reports. Secure the next contract.
The consultants need to produce designs that donors will fund. Donors fund things that look rigorous and replicable, logic models, standardised approaches, and best practices imported from elsewhere. Donors do not fund proposals that say "we will spend six months listening to farmers to discover what the actual constraints are, and then design something custom to this context that may not work anywhere else." The government officials need projects that produce visible activity in their districts before the next election cycle. Workshops and training are visible. Slow, patient work building relationships and adjusting approaches based on what is learned is invisible until it is too late to matter politically.
Everyone is rational. Everyone is optimising. The system produces exactly what it is structured to produce: activity without permanence, visibility without transformation, expensive forgetting. What emerges is a theatre of progress where participation is staged, impact is performed, and reports are written to please donors rather than communities. Everyone is busy, everyone is professional, and everyone is accountable to someone above them in the hierarchy. The only people no one truly answers to are those who were meant to benefit.
Countries receive billions in aid year after year, decade after decade. The same countries. The same problems. Tanzania has received over 50 billion dollars in aid since its independence. Afghanistan received over 100 billion dollars in two decades. Haiti has been a focus of international development for generations. If this were any other industry, if pharmaceutical companies produced medicines that never cured anyone, if construction firms built roads that vanished after three years, we would call it malpractice. In development, we call it "complex" and "challenging contexts" and design the next project.
I do not say this as an outsider levelling accusations. I say it as someone who has participated in this system, who has watched it from inside, who has been complicit in it. I have written the reports that emphasise outputs while staying silent on absent outcomes. I have attended the workshops where we performed consultation while never genuinely holding the plan open to what we might learn. I have watched good people, smart, dedicated professionals, trapped in a structure that requires them to ignore what they see in order to produce what is required. The tragedy is not that people are incompetent or uncaring. The tragedy is that the system is designed such that competence and care cannot save it.
This brings us to the question that will drive this series: who are the stakeholders, and why does everything depend on whether their knowledge actually shapes what gets built? The conventional answer is that stakeholders are "involved" in projects; they attend consultations, they participate in workshops, they provide feedback. But this answer obscures the critical question: when do they enter the process, which stakeholders are identified, and do the power dynamics permit them to actually influence decisions?
A farmer summoned to approve a road route already surveyed possesses knowledge that might have changed the entire premise, but now can only adjust minor details. The woman who processes the catch but does not own the boat never makes it into the stakeholder meeting, so her knowledge about where boats actually land never enters the plan. The community health worker who knows the real barriers to clinic use stays silent in a workshop dominated by officials from the capital, so the project addresses imagined barriers while real ones remain. The issue is not whether stakeholders are present. The issue is whether the institutional arrangements permit the knowledge they possess to surface and reshape what gets done. And the answer, in most development practice, is no.
Projects vanish because they are built on theories held by planners rather than knowledge held by communities. When the planners leave, the theories depart with them. What remains are people going about their lives using the knowledge they always possessed, knowledge that could have made the project work, had anyone created the conditions to actually hear it.
Until that changes, development will remain a theatre of progress rather than progress itself. The test should be simple: can a community point to something real years after the project money is gone? If not, then no matter how many workshops were held or reports were produced, it was never development at all.