Fragmentation is a choice, not a condition.

I closed a speech in December with a sentence I had been carrying around for a while. Fragmentation is our risk; interoperability is our responsibility. I want to take the second half seriously, because it is the half that costs something.

We talk about fragmentation as though it were weather. It is described as a feature of the sector, an unfortunate inheritance, something everyone deplores and nobody owns. It is not weather. It is the sum of a great many decisions, each of which was defensible at the moment it was taken, taken by people who could each name a good reason. Twenty years of collaboration initiatives have produced real results and uneven progress, and most agencies still run their supply chains independently, which means the parallel version keeps being chosen.

The research I know best names the barriers plainly. Competition for funding and visibility. Weak incentives to collaborate. Organizational and operational differences. Resistance to change. I want to sit on the second word in that first barrier, because it is the one nobody says out loud in a coordination meeting.

Visibility.

Collaboration means somebody else's name on the result. If two agencies run one procurement instead of two, one logo goes on the outcome and one does not, and the one that does not still has to go back to its donors at the end of the year and account for what it did. In a sector where funding follows attribution, giving up visibility is not vanity. It is a budget risk taken by a person who will be held to a number. Every incentive that organization faces is telling it to keep its own warehouse.

That is the second question in my framework, arriving in an unexpected place. Whose name is on it. I built that question to describe what happens between an international organization and a local one, where the outsider keeps the authorship and the local actor carries the work. The same mechanism runs horizontally, between peers, and it produces duplication instead of dependency. Same question, different axis, and it is why the two problems will not be solved separately.

Which is where the word independence gets used, and where I want to be careful, because it is doing two jobs. Some independence is real. A mandate that requires an agency to act when others will not, an operational capability that has to exist inside the organization because it is needed at three in the morning, a principled refusal to depend on a partner whose access is negotiated differently. That independence is worth defending and I would defend it.

The rest is duplication with a good name. The study found the duplication sitting exactly where it costs most, in procurement and transport, and in the customs arrangements and supplier data underneath both. Agencies negotiate the same clearance in the same country separately. They keep parallel records of the same suppliers. Then the arrangement gets called resilience, and it is not. Resilience is having a second option when the first one fails. Running one fragile thing several times over is not a second option, it is the same exposure bought repeatedly.

Here is why the word responsibility matters more to me than the word risk. Risk is collective and therefore nobody's. Every organization can agree that fragmentation is bad for the system while acting rationally to preserve its own position within it, and no individual step in that sequence is wrong. Responsibility divides. It is mine, and it is separately yours, and it survives the fact that the other party has not moved yet. If interoperability is only a risk we share, it will keep being described in communiqués. If it is a responsibility each of us holds, somebody has to go first and absorb the cost of going first.

What that does not mean is another coordinating body. The sector's reflex when coordination fails is to build an institution to coordinate the coordination, which adds a seat to every future negotiation and gets counted as progress. The research I am drawing on refused that route deliberately, and looked instead at whether the mechanisms already running could be strengthened, extended or aligned. That is the harder discipline and the right one. The first question about any coordination proposal should be which existing thing it replaces.

So the practical test I would put to my own organization, and to anyone else's. Name one capability you maintain that somebody else already runs well. Not one you could theoretically share, one you actually could stop building tomorrow. If the answer is none, the answer is not that you are unusually efficient. It is that nobody has been asked to look, because looking is how you find the thing you would have to give up.

So we went first. I committed 2.5 million Swiss francs to interoperability, digitalisation and local voices, and the figure is the least interesting part of it. What makes a commitment real is what you are prepared to stop doing in order to fund it. If holding that priority means pulling other projects back, they get pulled back.

I hold that as a working rule rather than a virtue. Spreading resource thinly across everything is how you explore, and exploring is worth doing. It is a poor way to deliver any of it. Once you know what the priority is, funding it properly means letting go of things you would have enjoyed running, and an organization that funds every good idea is telling you it has not decided yet.

Which is where the word lands. Responsibility is not agreeing that fragmentation is bad. It is the thing you gave up.

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