Jacob Huibers — Statecraft, April 2026

Sometime this spring machinery-of-government.vercel.app turned up in my timeline. A British developer has placed the entire Whitehall machine on a single screen. Departments, ministers, agencies, arm’s-length bodies, with budget and personnel figures included. Clickable, searchable, walkable in half an hour. Anyone who clicks through afterwards has a reasonably complete picture of who in the United Kingdom is responsible for what.

The question pressed itself: can this also be done for the Netherlands? The short answer: the data are there. The Register of Government Organisations at organisaties.overheid.nl maintains the inventory; Statistics Netherlands provides the financial figures; rijksfinancien.nl exposes budgets and personnel figures per agency and quasi-autonomous body. A developer with an evening to spare can technically replicate the Whitehall approach.

The longer answer is more interesting. Whoever does the exercise does not see a Dutch version of the British picture appear. Whoever does the exercise sees something illegible. And that illegibility is no technical problem to be solved with better visualisation. It is the story itself.

What the British project makes visible

The British model is in essence hierarchical: ministerial departments at the top, beneath them executive agencies and non-ministerial departments, and beneath those a layer of arm’s-length public bodies. Three layers, all centrally directed, all in principle under ministerial responsibility. The visualisation can show that because the structure lends itself to being shown. A layperson understands within a few minutes how Defra relates to the Environment Agency, or what the Driver and Vehicle Standards Agency does within the Department for Transport.

Local government remains out of view in this picture. That is in the British case defensible, because in budget and tasks local government is a fraction of what in the Netherlands has been decentralised. The municipalities of England, Scotland, Wales and Northern Ireland together spend less than the Department for Work and Pensions alone. Whoever pictures Whitehall pictures almost the entire state.

In the Netherlands that ratio is inverted. Dutch municipalities spent over 80 billion euros in 2024, inter-municipal partnerships 13.2 billion, provinces and water boards together about 10 billion. Added together that is a substantial part of what the Dutch state actually does. Whoever replicates the British approach here and leaves local out, does not tell the story. Whoever takes local in gets no visualisation any more but a tangle.

The figures that make the picture illegible

The Netherlands has 13 ministries, around 38 agencies, just under 120 quasi-autonomous bodies, 12 provinces, 21 water boards, 25 safety regions, 29 environmental services, and 342 municipalities. So far disordered but reviewable. Beneath that lies the actual finding: 438 inter-municipal partnerships (gemeenschappelijke regelingen), in four statutory forms, which together form just over three thousand hub-and-spoke connections between those 342 municipalities and the supra-local tasks they collectively try to deliver.

That figure of 438 calls for elaboration, because counting methods diverge. The register of inter-municipal partnerships at overheid.nl contains public bodies, business-operations bodies, joint organs and so-called arrangements without further structure. Whoever counts only the heaviest variant arrives a few hundred lower; whoever also counts the centre arrangements and the cooperation agreements arrives over seven hundred. The spread is itself telling: there exists no authoritative answer to the question of how many components the Dutch executive state consists of. A state that cannot count itself, also cannot draw itself.

Whoever does try arrives at a picture in which the average Dutch municipality sits in between eight and nine inter-municipal partnerships. An alderman in a municipality of 50,000 inhabitants is therefore a board member of a safety region, a public health service, an environmental service, a tax cooperation, a sheltered employment provision, an IT cooperative, an employment service and a youth-care region. Those are only the structural cooperations. Add to that temporary covenants, regional deals, programme organisations, and the dedicated account managers from central executive agencies who coordinate with the municipality.

In an interim assignment in a centre municipality, I worked with an alderman who sat on fourteen general boards. He described his diary as a matrix in which six to eight inter-municipal-partnership meetings fell each month, each with its own budget cycle, its own annual account, its own auditor’s report and its own consultation procedure. He led none of those organisations. He represented his municipality in collective boards where his vote was statistically never decisive, and where the decisions then returned via an apportionment system to the municipal budget. His council rejected the formal opinions he later in the general board could not block anyway. The system produced the paper reality of democratic control, while everyone in the chain knew it was ritual.

The UWV paradox

One observation is administratively instructive: large does not by itself mean unmanageable, small does.

The Employee Insurance Agency (UWV) covers a budget of around 44 billion euros. That budget is annually scrutinised by the House of Representatives via the standing committee for Social Affairs and Employment, tested by the Court of Audit, inspected by a national inspectorate, commented on by a trade-union delegation, and followed by media that have any incidental benefits error in the paper within a day. Ministerial responsibility is regulated. The democratic infrastructure per euro of public money is, for all the problems UWV has, impressive.

A random inter-municipal partnership of 200 million euros has none of that. No standing committee. No Court of Audit with dossier knowledge. No national inspectorate. Democratic control runs via representatives in a general board, who themselves sit on seven or eight other general boards and have no parliamentary support staff on the dossier. The municipal council receives a formal opinion for approval which everyone knows does not bind the outcome. By the time the annual account of an inter-municipal partnership reaches the council agenda, the decisions are already two years old and the money already spent.

Whoever works inside the system recognises this as the structural paradox: the smaller and more fragmented the administrative arrangement, the less control is exercised. Not because councillors are stupid or lax, but because it is mathematically impossible, with the time of a part-time controller, to follow eight to ten organisations, each with their own planning-and-control cycle and their own technical dossiers. The legal fiction of “extended local government” (verlengd lokaal bestuur), under which the Joint Provisions Act has worked since 1985, is in execution an empty shell.

The mathematical optimum as mirror

Whoever approaches the design problem rationally arrives at an unexpectedly clean answer. The classical organisation-design rule of thumb is that an effective span of control lies between five and nine entities. With seven as guide and 342 municipalities at the base, that yields exactly three layers: 7 × 7 × 7 = 343 ≈ 342. One national level, beneath it seven country regions, beneath those forty-nine regions, beneath those the municipalities. The mathematics that the British model implicitly uses turns out to be just as clean for the Netherlands.

This is no blueprint for a new state organisation. It is a mirror. It shows what the Dutch executive state in fact already tries to organise — namely a scalable layered structure in which local tasks are executed locally and supra-local tasks are bundled at a fitting level. It also shows how far the actual architecture has drifted from a minimal logic. Instead of seven country regions, the Netherlands has four informal country regions and twelve provinces whose execution capacity largely overlaps with regional divisions imposed from central government. Instead of fifty structural regions, the Netherlands has 25 safety regions, 29 environmental services, 25 public-health regions, 35 labour-market regions, 42 youth-care regions, and 30 regional energy regions, all using their own division and only partially overlapping each other.

Congruence of regional divisions is an open door, and it is tempting to end there. But congruence does not touch the actual question. Even with fully congruent regional borders, those would still be 25 or 50 or 80 separate legal persons with separate boards, separate budgets, separate decision structures, and the problem that none of those legal persons is directly elected by the residents. The structural shortage does not sit in the geographical lines, it sits in the absence of an administrative layer with mandate over what plays out at that scale.

What visibility would change

The British project is appealing because it makes the state manageable. A visualisation of the Dutch equivalent would do something else. It would show why the state no longer has itself in hand.

A large part of the Dutch executive state consists of organisations that are not directly mandated by any voter, that are not fully steered by any minister or alderman, and that cannot be effectively controlled by any representative body. It is no shadow government in the conspiracy sense. It is a real government, with real budgets and real decisions, that for lack of democratic infrastructure operates in a zone of structural invisibility. The effort a councillor must make to gain insight into one inter-municipal-partnership budget is out of proportion to what can subsequently be achieved with that formal opinion. The rational response is then that most inter-municipal-partnership budgets are approved as routine items, and so it goes.

The Act Strengthening Democratic Legitimation of Inter-Municipal Partnerships, in force since 2022, adds to this picture the characteristic pattern: rather than ask the structural question of why eight to ten legal persons per municipality is sensible at all, the legislature adds extra safeguards within the existing system. A citizen-participation paragraph in every arrangement, a more extensive consultation procedure, an evaluation obligation. Paper control is increased; the actual problem remains. The system defends itself by complexifying itself further legally, and thereby reduces the chance that a councillor still understands any of it.

A Wgr+ youth-care region, as the Act on improving the availability of youth care provides, would be the next example. A forty-third cooperation arrangement in the social domain, in a new legal jacket, with again its own board and its own budget. The explanatory memorandum to the act describes it as a tightening of current practice. Whoever knows the system reads an expansion of the problem the act claims to solve.

Closer to what a visualisation could expose is the question of what in the Netherlands is the minimal working unit of public decision-making. A minister decides with mandate, an alderman decides with mandate, a general board decides collectively and cannot be steered by any of its members alone. Whoever wants to get something done at the supra-local level must organise a conference. Whoever wants to prevent something from being done need only wait until coordination costs exhaust the others. The system is therefore a conserving mechanism of the first order. It can produce almost nothing new, and it can wind down almost nothing.

An intervention that genuinely wants to change this state of affairs touches not an individual arrangement but the design principle. That principle has held since 1985 that supra-local tasks are placed via “extended local government”, even when the scale of the task can no longer be carried locally and even when the councillors who should control it in fact no longer get round to it. A state that can no longer draw its executive layer is a state that can no longer defend its own design.

The British project shows what a state is when it still allows itself to be shown. The Dutch version would show what a state becomes when it no longer does.

---

Sources: Statistics Netherlands, Municipal Division 2026 and Government Finance 2024; CBS Iv3 model for inter-municipal partnerships; Register of Government Organisations (organisaties.overheid.nl); Joint Provisions Act; Act Strengthening Democratic Legitimation of Inter-Municipal Partnerships (2022); machinery-of-government.vercel.app.