The End of a Form of Administration

1Opening

German public administration is not failing in individual cases but as a form. It is failing in construction, in migration, in education, in the health sector, in research funding, in digitalisation. It is not failing because its staff are incompetent — most are trained, conscientious, often committed. It is failing because a particular historical form of administration — the administration of the 20th century — has lost the conditions under which it could function. The crisis becomes visible as a symptom: a building permit that takes five years; an asylum procedure that never ends; a funding programme whose effect is never measured; an electronic file that is merely a scanned piece of paper. These are not mishaps. They are signs that an entire construction has lost its preconditions.

This working paper attempts to describe the phenomenon in such a way that it is not mistaken for a programme proposal. It is not about *how* the administration ought to be reformed. There are proposals enough. It is about what is happening structurally, why the reform routines the system has provided for itself do not take hold, and which observer position is appropriate for the diagnosis.

The finding is old. Max Weber theorised bureaucracy as a historical form around 1920 and named its limits even then. Renate Mayntz described the problem of the self-observation of programme policy in the 1980s. Lewis Mumford, in The Pentagon of Power in 1970, developed the thesis of a Megamachine that cannot reform itself. The theoretical references are at hand. What is missing is a sober inventory of what we are currently observing.

2Bureaucracy as a historical form

Bureaucracy in today’s sense is an achievement of the late nineteenth and early twentieth centuries. It was built in a particular social constellation and optimised for that constellation’s tasks. Its design conditions can be summed up in four preconditions:

First — stability of rules. Bureaucracy works when rules hold over a longer period. It can apply a new rule; it cannot constantly draft new rules while implementing the old ones at the same time. Its files, forms, training courses and procedures need time to become established. A rule that is changed again after two years has scarcely any half-life in the administration.

Second — territorial closure. Bureaucracy is made for a domain of validity that coincides with the borders of the state. Its competences end where the state ends. Tasks that run across several states or cut across state borders can be represented in its logic only to a very limited extent.

Third point — separability of subject matter. Bureaucracy is divided into ministries. That presupposes that substantive questions can be broken down separably into specialised competences. Economy, transport, environment, research, health, education — each ministry has its files, its officials, its rules. Whatever lies between the ministries is, in organisational terms, a blind spot.

Fourth — reliance on writing. Bureaucracy is record-keeping. A case exists when it has been written down; it is closed when it has been recorded. The file is not a side issue; it is what constitutes administration. It makes the case traceable, verifiable, legally secure.

Today these four preconditions no longer hold, in part or in full. Change happens in weeks, not decades. Problems are transnationally intertwined. Substantive questions cut across ministries. Writing is now only one of several media forms — data, images and algorithms stand alongside it or take its place. Bureaucracy tries to carry on working as if the preconditions were intact. The result is a form that formally continues to function but substantively no longer fulfils its tasks.

3Four fault lines

The erosion of the four preconditions has not progressed equally far everywhere. But it can be described precisely as a diagnosis in each of the four areas.

3.1 Industry

Industrial mass production was the basis of a form of administration that in turn processed mass cases. Tax collection, social insurance, trade supervision, market regulation — bureaucracy and industrial society developed in parallel. Today industry is deterritorialised; value chains run across continents. The administration still tries to capture it by location, ownership and legal form. Concepts such as company or location are in many areas no more than accounting fictions.

3.2 The nation state

Bureaucracy was the bureaucracy of a nation state. Today the level of action above the national — the EU with its regulations and funding programmes — and the level below it — the region, the federal state, the municipality — have both become more important than the national level itself. German administration sits between two directions of pressure that it cannot serve at the same time without being ground down. The EU demands standardisation; the region demands distinctiveness. Berlin tries to mediate and in doing so produces a language no one understands.

3.3 Writing

The file is the past. What counts today are data flows — structured, searchable, combinable information. A file describes an individual case; a database links patterns. The administration has remained trapped in the logic of writing. It produces files, scans them and calls that digitalisation. But it has no idea what it means to work in a *data-based* way — that would be not only a technical question but an epistemic one. Data connect according to different rules from file references.

3.4 Hierarchy

Bureaucracy is built hierarchically because it assumes that those at the top know more than those below. In the past this assumption was mostly correct. Today it is mostly wrong. The case officer in the building authority knows more about the building problem than her head of department; the programmer of a funding database knows more about its weaknesses than the head of unit responsible for it. Hierarchy works when knowledge runs along the hierarchy. When that is no longer the case, hierarchy becomes an obstacle rather than a source of clarity.

4Digitalisation as a symptom

Anyone wanting to observe German administration can begin with digitalisation. It makes visible what remains invisible in the classical file structure. When you apply for an online public service, you typically go through a process that leaves the old structure unchanged: you enter data into a form, the form is sent off, a case officer receives it, checks it according to the old rules and writes a notice that is delivered by post. Digitalisation is confined to the input channel. Processing remains analogue.

That is not *missing technology*. The technology is available. What is missing is the structural, organisational, legal and mental readiness to leave the old logic behind. Estonia shows how it could be done differently. There, the administration was built after 1991 without any historical baggage. Digitalisation was not an appendage but the basic assumption. Today most Estonian administrative procedures can be completed in seconds, because the data flow between authorities instead of being managed in files. A marriage is registered electronically with one office; all other authorities — tax, social insurance, residents’ registration — are informed automatically. That is not *advanced administration* but a different administration.

Germany will not go down this road, because it would destroy the existing structure. Whoever builds an electronic connection between authorities takes away the authorities’ autonomy. Whoever lets data circulate freely between the federal government and the Länder grinds down federalism at a decisive point. Digitalisation does not succeed because it cannot succeed without shaking the political fabric that would have to carry it.

The lack of digitalisation is therefore no proof of incompetence. It is the logical consequence of a form of administration that cannot give itself up.

5Symptoms in the policy areas

The same diagnosis can be traced in almost all major areas of administration. Here are six examples, kept brief.

5.1 Research funding

More than twenty project management agencies administer more than four hundred funding programmes on behalf of some twelve federal ministries, plus the EU, plus the Länder. In 2026 the Bavarian Supreme Audit Office put the number of Bavarian programmes alone at more than four hundred, with administrative costs of up to sixty per cent in individual cases. The programmes’ success is monitored — if at all — through commissioned reports by the funders’ own research organisations. In July 2025 scientists at Fraunhofer ISI said in a blog post that the learning ambition of previous strategies had not been fulfilled.1

5.2 Construction

Planning applications in Germany today take four times as long as twenty years ago. Building regulations have not become simpler but more complicated, because every crisis adds a new layer of rules: energy, fire protection, accessibility, structural engineering, noise, heritage protection. There is no body whose task would be simplification. Complexity grows because it accumulates. A thoroughgoing reform would affect lobby groups, associations, professional bodies, the insurance industry and the courts all at once — and is therefore politically impossible to push through.

5.3 Migration

Asylum procedures take two to three years on average. The federal offices are chronically overloaded. Family reunification works formally, but not in practice. Integration is invoked, but the administrative structure meant to make it possible is fragmented between the Federal Office, the foreigners’ authority, the job centre, the school authority, the municipality, the welfare association, the language school and the recognition office. Each authority demands its own documents, its own forms, its own office hours.

5.4 Schools

A shortage of teachers in almost all federal states. Dilapidated buildings. Outdated curricula. The answer is always the same: one more reform, one more programme, one more strategy. The structural question — how do you school children in a digital, multilingual, climate-critical world — is not asked, because it is too big for the forms of administration that are available. The federal structure guarantees sixteen parallel answers to every question and no common solution.

5.5 Health care

Hospital reforms for thirty years, without thoroughgoing success. Long-term care insurance was introduced in 1995 on the assumption of a demographic development that is today taking a completely different course. But the structure does not adapt; it is fed with additional money. The digital patient record has been announced for more than ten years and is still not functional.

5.6 The tax system

More than two hundred special provisions, exceptions, privileges. No one fully understands the system any more. A fundamental simplification has been promised for thirty years and never implemented. Complexity grows because every new provision is formulated as the answer to a specific lobby demand, without the system as a whole being taken into account.

The list could be extended. Pensions, social insurance, the judiciary, internal security, environmental law — everywhere the same structure is found: an old form of administration, overloaded with tasks it was not made for, helpless at adapting, because its construction does not provide for adaptation.

6Inability to reform itself

Why does reform not succeed? This question has three answers that work together.

First — the structural answer. By its construction, bureaucracy is a rule-executing institution. It knows how to follow rules. It does not know how to turn itself into an institution that develops, tests, discards and replaces rules. That would be a learning institution — something classical bureaucracy structurally excludes, because learning presupposes discretion and discretion runs counter to the bureaucratic principle. If administration is allowed to learn, it could become arbitrary. Weber explicitly wanted to prevent that.

Second — the personnel answer. Anyone who works in an authority as a civil servant for thirty years becomes a bureaucrat. That is not a moral failing but conditioning. The structure shapes behaviour. It rewards risk avoidance, record-keeping, attendance at meetings. It punishes initiative, taking responsibility for failure, criticism directed upwards. Over four decades a population has accumulated in the German ministerial bureaucracy that is predisposed to administering, not to shaping. Even committed officials who know that the funding logic is wrongly conceived cannot change it — they have to check proof-of-use statements because the Federal Budget Code prescribes it.

Third point — the incentive answer. Even someone who wanted to has no reason to commit to reform. Careers are made by superiors, not by impact outside. Whoever stands out is not promoted. Whoever keeps quiet inherits the post. The promotion structure trains people to conform; conformity reinforces the structure.

The three answers interlock: the structure produces the people. The people reproduce the structure. The incentives stabilise both. The result is a self-preserving system in the strict cybernetic sense — a system that produces precisely the conditions of its own continuation. Not because anyone wants it that way, but because the feedback loops are closed accordingly.

The reform routines the system has provided for itself are all ineffective:

• A new minister changes the top, but not the apparatus.
• An administrative reform changes forms, not incentives.
• A new strategy replaces the old strategy without anything having been learned from the old one.
• An evaluation produces a report that is filed away.
• An audit office report generates a day of press coverage that passes.

The system cannot reform itself. It would have to abolish the structures on which its actors live.

7The question of transition

If the diagnosis is right — if bureaucracy is a historical form that has lost its preconditions, and if it cannot reform itself — then the question is no longer *how it can be reformed*. It is *what comes after it*.

In his late writings Lewis Mumford formulated a similar diagnosis for large-scale industrial society. He called the construction in which we live a Megamachine — a large structure held together by bureaucracy, hierarchy and standardisation that cannot observe its own functional logic. It is, he wrote, incompetent in a higher sense. Not incompetent at carrying out its routines — but incapable of self-observation, of learning, of correction. Mumford saw no possibility of reform from within. He expected either a deliberate replacement by other structures — he called them polytechnics: decentralised, rooted in craft, capable of learning — or disintegration.

Mumford’s answer was unsatisfying because it was too closely tied to wishful thinking. But his diagnosis has stood the test. What we are observing today is not a crisis *within* a form of administration. It is the crisis *of* a form of administration — and the question is what will take its place.

There are tentative indications of transitional movements:

The generational question. The civil servants who began their careers in the 1970s and 1980s will retire over the next ten years. Younger people are coming up who were socialised in a different world — with a different relationship to hierarchy, to transparency, to responsibility. That is no substitute for structural reform, but it creates possibilities.

The fiscal crisis. Over the next ten years demographic change and the fiscal burden will narrow the financial room for manoeuvre so severely that the existing system can no longer be continued for lack of money. That will force reforms — probably hard, painful, unplanned ones.

Alternative structures. SPRIND (the Federal Agency for Disruptive Innovation), the Cyber Agency, individual pilot projects — recent years have produced attempts to work within the state but outside the ministerial bureaucracy. These structures are small and are watched with suspicion by the classical administration; some are ground down. But they show that alternatives are conceivable.

Subnational movements. In Estonia, Denmark, parts of Switzerland and individual German municipalities, administrative structures are emerging that are built differently from the national apparatus. They are not directly transferable, but they demonstrate that the 20th-century form is not the only conceivable one.

It is possible that a different form of administration will gradually emerge from these movements. It is just as possible that the old form will run on until it collapses at some point — through a fiscal crisis, a political upheaval, a demographic dislocation. What does not seem possible is a *reform under its own steam*. The Megamachine does not reform itself. It is replaced — or it disintegrates.

8Open questions

This working paper is not a closed analysis but an attempt to put a field of observation in order. Several questions remain open.

First — how long is the period of transition? In the 1970s Mumford reckoned with time horizons of decades. But crises can accelerate. We do not know whether the dissolution of today’s form of administration will stretch over twenty or over fifty years.

Second — what breaks first? If the administration collapses, it will not do so equally everywhere. Some areas are more robust (social insurance, the land registry), others more fragile (asylum, construction, schools). Which areas collapse first — and what consequences that has — is an empirical question that is not yet being asked.

Third point — what role does the public play? Structures are changed by pressure; pressure arises through attention. Germany’s administrative misery produces no continuous pressure, because it has no scandal format. It is slow, structural, without personalisation. As long as that remains so, the pressure for change will not arise. That may change when the consequences become directly tangible for the electorate.

Fourth — what role do technical developments play? Artificial intelligence will arrive in public administration in the coming years. It is open whether it will stabilise the old structure (by making it more efficient) or undermine it (by showing what administration could be instead). Probably both — differently in different areas.

Fifth — who observes? If the administration cannot observe itself, then the observer position outside becomes important. Who takes it up? The audit offices try, but without resonance. Academia produces commissioned studies for the funding bodies. The media focus on scandal formats. What is missing is a form of external observation that argues structurally, stays with it over the long term and generates resonance. Whether it will emerge is open.

The diagnosis of this working paper is sober, not resigned. It does not claim that nothing can be done. It claims that what needs to be done cannot be found within the logic of the administration itself. Whoever expects reform from the system is mistaken. Whoever considers the system unchangeable is mistaken too — it will change because it can no longer hold. The only question is whether the change will be deliberately shaped or suffered as disintegration.

That is a different question from the reform question of the twentieth century. It demands a different observer position, a different language, a different patience. This working paper is an attempt to take up that position — provisionally, with open questions, in the awareness that the real work is only beginning.