The sending side

Every European integration platform I have looked at was built by someone who first encountered migration at the point of arrival. That is not a criticism. It is where they live and it is where the problem became visible to them.

But a labour migration corridor has two ends, and almost everything that determines what happens at the receiving end was decided at the sending end, often years earlier. I spent twenty-one years administering that end.

What the departure end actually is

For most of my career it was a set of institutions with unglamorous names. The Department of Foreign Employment. The Department of Labour. Later the Ministry of Labour, Employment and Social Security. What they do is licence recruiters, negotiate destination country agreements, run the pre-departure system, and maintain the data infrastructure underneath all of it.

Several hundred thousand workers leave Nepal each year. Every one of them passes through a process that records who they are, what they can do, where they are going, and under what terms. The state that issues that record knows a great deal about the person before they board the plane.

By the time they arrive, almost none of it travels with them.

The thing this makes obvious

Here is what twenty-one years on that side of the desk teaches you, and it is not what people expect.

The skills information exists. It is just discarded at the border. A worker who was assessed, documented, and cleared to work in a specific occupation arrives in a country that has no way of reading any of that. Their competence becomes invisible at the moment they cross into the system that most needs to know it. Not lost, not falsified. Discarded, because there is no receiving structure that reads it.

Institutions on both sides are working blind about each other. A sending state does not know how its workers are actually faring in the destination labour market, beyond remittance flows and consular incidents. A receiving municipality does not know anything about the person's documented history before arrival. Two administrations, each holding half of a record, neither able to see the other half, both making decisions as though their half were the whole.

The failure is administrative, not cultural. This is the part I want to be most precise about. The public conversation about integration in Europe is largely a conversation about culture and belonging. The failures I watched for two decades were overwhelmingly failures of data, timing, and institutional plumbing. A qualification that could not be read. A pathway that stalled because two offices were not joined. A pressure that was visible in one department's caseload and invisible in the department that could have responded.

Those are engineering problems wearing political clothes.

Why this changes what you build

When your first encounter with migration is arrival, the natural product is a service for managing arrivals. When you have spent your career on the departure side, the natural product is different: it is the layer that carries what is already known across the boundary where it currently stops.

That is why Humaniwork is built around competence rather than status, and around institutional decisions rather than individual case management. The platform's job is to make the invisible half of a record legible to the institution that needs it, and to do it soon enough that the institution can act.

It is also why I am firm about what the Observatory is not. It monitors service demand and cohesion indicators at neighbourhood level, for the institution accountable for provision. It is not an enforcement tool, not a policing input, and not a targeting instrument, and a contract that asks it to become one is declined. I have watched what happens to migration data infrastructure in states where that line was not held. It stops being infrastructure and becomes something else, and the workers stop telling the truth to the system, and then the data is worthless as well as harmful.

The other reason I am here

I moved to Barcelona and found a city receiving extraordinary talent through systems that were, by the standards of the institutions I had run, analogue. Not underfunded. Not badly intentioned. Working on annual national statistics to make weekly local decisions, which is the same problem I had spent a career watching from the other end.

There is a particular frustration in seeing an institution make a decision about people's lives on information that is eighteen months old when better information exists two departments away. I felt it for twenty-one years. I did not expect to feel it again in a wealthy European city.

What I would say to a municipal director

You are not failing. Your instrumentation is failing, and you have been asked to be accountable for something nobody has given you the means to see.

That is a much more tractable problem than the one you are probably being blamed for.