Localization is a legitimacy problem, not a capacity problem.

The sector's answer for why localization keeps stalling is that local organizations are not ready yet. Not enough financial controls, not enough compliance capacity, not enough of whatever the next audit will ask about. I have heard it in every forum I have sat in. It is not a lie. It is an explanation that happens to require nobody to give anything up, which is a useful property in an explanation and worth noticing when one lasts this long.

For my MBA thesis I interviewed sixteen people about it, across international NGOs, UN agencies, donor bodies, local organizations, federated organizations and academia. I went in expecting capability to show up as the binding constraint somewhere in the data. It did not. Decision authority stays concentrated in global roles. Local actors carry the operational responsibility without holding the authorship or the decision rights. Of 264 coded references, the largest single dimension, with 101, was the conditions layer. Not capability. Conditions. The funding instruments, the compliance regimes and the governance structures that determine where authority is permitted to sit.

Compliance is the interesting one, because nobody built it in bad faith. It exists to manage fiduciary risk and it does that well. It also hardens the arrangement, because every requirement written for a large international body is a requirement that a smaller local one has to grow into, and the growing takes years during which the decision stays where it was. The system is not defending its power. It is defending against loss, and holding power is what that produces.

So the word I use is legitimacy, and I do not mean the formal kind that arrives with a title. I mean the working kind. Whether people act on a decision or wait quietly to see whether it holds. You can hand somebody a job title, a budget line and a place on the organigram and still not have given them that. It is granted by the people around the decision, and it does not transfer on paper.

This is what my framework asks about. Three questions, and you can put them to any relationship where one party carries the work and another holds the power. Who decides, meaning who can say yes or no without asking. Whose name is on it. And what holds that in place. Organizations answer the first two at headquarters, assign the responsibility somewhere else, and never touch the third. What that produces is responsibility without authority, which is the most common structural failure I encounter and it is not confined to aid.

Here is the version of this I carry with me.

In 2008 I was posted to a base at Rutshuru in North Kivu, in eastern DR Congo, with Merlin. Operating cash came up by road from the finance office in Goma. Word got out when money was moving, reliably, and organizations were being stopped on that road at gunpoint and relieved of it. The pattern that made an organization a target was built out of local knowledge, who talks to whom, how word travels between agencies and offices, and none of us who had flown in could see it.

A driver I had hired designed the answer. He argued we should never be seen leaving the finance office carrying anything. He came to the office early and took the cash in advance, serviced the vehicle, and concealed the money in a strapped storage space next to the engine under the bonnet. I left the finance office with no luggage at all, with everyone already sitting in the car, which told anybody watching that nothing was being carried. On arrival I walked straight into the house. He followed later with his own bag and his paperwork, and handed the money over.

My decision was to accept a security design authored by my driver, and to trust him to run it. We were never stopped and we never lost money, on a road where our peers were being robbed. Those organizations were, as far as I know, following correct procedure.

He was never a capacity problem. He held knowledge that could not be acquired from outside, and he was on the payroll as a driver.

The second decision I made about that protocol was never to write it down, and that one has stayed with me longer, because my profession runs on capturing what works and circulating it. Some knowledge dies on contact with a document. The thing that kept us safe was one man's read of a specific town in a specific year, and a best practice circulated between offices would have destroyed the only property that made it work, which was that nobody else knew it. So it went nowhere. It protected one team on one road and then it ended, and the cost of that is real. The cost of the alternative would have been that it stopped working.

That is the choice the sector keeps declining to make. Not whether to extract local knowledge and standardize it, which is the reflex, but whether to move the authority to the people who already hold the knowledge and let them decide what to do with it.

The full argument runs through Pathways and Power: who decides, whose name is on it, and what holds that in place.