CORE PATH Stop 96 / 106

Polycentricity and subsidiarity October 4, 2021 Updated: September 11, 2026 17 min read THY-REALITY

When is larger scale genuinely necessary?

Larger scale is not automatically better — but sometimes it is genuinely necessary. This article shows how to identify economies of scale, rare expertise, network functions and risk pooling, while balancing them against access, responsiveness and local autonomy.

“Subsidiarity in Practice: Which Level Should Decide?” turned subsidiarity into a method for choosing scale, while the articles from “How Can Communities Cooperate Without a Single Center?” through “Mobility and migration between communities” showed how autonomous communities can cooperate, use shared standards, trade and enable mobility. This article now has to state something just as important to the project as the critique of centralisation: sometimes larger scale is genuinely better — and for some functions safer, cheaper or more accessible as well.

The project backbone, as already formalised in “Local Sovereignty: How to Build Resilience Without Isolation”, treats local sovereignty as capability and resilience rather than autarky. “Health Care Close to People: What Can Be Local and What Requires a Larger Scale?” provided the first concrete health-care example: primary care needs proximity, while some rare and demanding functions need a sufficiently large pool of patients, specialists and equipment. the articles from “Who Maintains Shared Infrastructure?” through “Mobility and migration between communities” then reopened the same question across infrastructure, cities, preparedness, ecology, subsidiarity, standards, supply chains and mobility.

This article therefore does not search for an optimal size of a state or municipality. It searches for evidence that a specific function at a wider scale can achieve something a smaller level cannot achieve equally well — and that this benefit exceeds the costs of distance, coordination and concentration of power.

Scale is a question of function, not ideological identity

The biggest mistake in debates about decentralisation is to turn scale into an identity. Small becomes automatically humane and free while large becomes automatically bureaucratic; or, in the opposite ideology, large becomes automatically professional while small becomes parochial and inefficient. The evidence is far less romantic.

The OECD's 2026 review Federalism of scale finds that scale effects vary strongly across services depending on fixed costs, population density, technology and service characteristics. The literature often shows a U-shaped cost relationship: very small units have high per-user costs because fixed costs are spread across few people; at very large scale, congestion, coordination costs and loss of responsiveness can appear.

This article therefore uses the term larger scale rather than 'centralisation'. A function can operate at a larger scale through a federation, consortium or shared specialised service without one institution taking over every other competence. Scale describes the scope of a function; centralisation describes concentration of decision-making. They are not the same phenomenon.

The clearest economies of scale arise when the initial cost is high while additional use is relatively cheap. A laboratory, complex imaging equipment, a highly specialised surgical team, a treatment plant, data centre or heavy maintenance machinery all have costs that a very small user base may struggle to carry.

If ten communities each buy rarely used equipment, much capacity may sit idle while every unit must maintain the equipment, knowledge and spare parts. If the same function is shared, fixed costs can be spread across more users and specialised knowledge can be used frequently enough to remain sharp.

This is not an automatic argument for one large organisation. It is an argument for pooling rare capacities. Members can retain local decision-making elsewhere and jointly finance only what is indivisible or infrequently needed.

Second reason: quality sometimes needs sufficient volume

For some services, higher volume is not only a financial advantage but can improve quality. Highly specialised health care is the clearest example. A team that performs a difficult procedure only rarely has more difficulty maintaining routine, coordination and support systems than a centre that treats the same type of case regularly.

In its 2026 Economic Survey of Croatia, the OECD recommended concentrating complex care in high-performing specialised facilities and using minimum-volume requirements for certain procedures, while also protecting access for remote areas. This is almost a textbook “Health Care Close to People: What Can Be Local and What Requires a Larger Scale?” example: proximity for frequent needs, wider professional scale for rare and complex needs.

The right question is therefore not whether a hospital is local or national. It is how many cases, specialists, on-call teams, diagnostics and supporting services a function needs to remain safe and high quality.

Electricity, rail, water systems, telecommunications and parts of logistics have a network property: the value of each point depends on connection to others. Local capability remains valuable, but part of the function necessarily operates across a wider area.

The European Commission explicitly states that cross-border electricity interconnections strengthen resilience, enable more efficient use of generation and support renewable integration. The new Spain–Portugal interconnector inaugurated in July 2026 increased transfer capacity between the countries and was presented as reinforcement of both systems, not as a reduction of their energy capability.

“Energy: From Consumer to Energy Cooperation” therefore correctly rejected the idea that local solar generation removes the need for a grid. This article adds the opposite safeguard: a larger grid does not remove the value of local generation, storage and microgrids. Resilience often comes from combining local reserves with wider exchange.

Water, wastewater and infrastructure make the same scale test even more concrete.

In water and wastewater services, economies of scale often arise in treatment, laboratories, professional staffing and major investment. But the OECD also points to the opposite effect: a larger facility may be cheaper per unit while connecting users to it requires longer pipes, pumping, maintenance and energy.

The size of a treatment plant therefore cannot be judged without geography. In a dense area, regional provision may substantially lower costs and improve professional capacity; in a dispersed valley, long centralised infrastructure can cost more and create additional failure points. OECD work on water services therefore treats regionalisation as valuable where it genuinely pools capacity and expertise, not as a universal rule.

“Who Maintains Shared Infrastructure?” already required life-cycle stewardship. This article adds that we must compare the full system cost, not merely the price of the largest facility.

Education: a larger school may lower cost but increase distance

Education shows why scale is always a trade-off between efficiency and access. The OECD finds higher per-student costs in sparsely populated rural areas because schools and classes are smaller. Larger scale can make better use of specialised teachers, laboratories and other facilities.

Yet closing every small school and transporting pupils to a distant centre is not the only way to capture scale benefits. OECD work highlights school clusters, shared specialists, common facilities, distance learning and service co-location. This matters for THY-REALITY: economies of scale can often be captured through a network rather than by eliminating the local institution.

Larger scale is therefore justified for the sub-function where it truly improves provision or reduces unreasonable cost, while local presence can remain where proximity itself creates value.

Fourth reason: a wider pool can carry rare risks and reserves

Some functions are expensive not because they are used frequently, but because they must be ready for an uncommon bad event. Specialised rescue equipment, reserve transformers, advanced technical teams or specific emergency stocks have high value precisely because they are not needed most of the time.

If every small unit must maintain the full spectrum of rare reserves, the system may spend heavily on idle capacity. “When the System Fails: Local Preparedness for Disasters and Major Disruptions” showed a more sensible nesting: households and local communities carry basic reserves and first response, while wider levels maintain rarer specialist capacity that can be moved to where it is needed.

This is risk pooling in physical form: many units contribute to reserve capability because they are unlikely to need the same rare resource simultaneously. Where risks are correlated — for example a major regional flood — the reserve must itself exist at an even wider or geographically diversified level.

Larger scale can improve equal access — or worsen it

Wider financing and a larger user base can help provide services in places where a local tax or user base could not carry the cost alone. This is an important argument for regional or national solidarity in health care, infrastructure, specialised education and other functions with high costs per case.

But wider scale can also increase physical and administrative distance. OECD work on health and education stresses precisely the trade-off between cost and access in rural areas. A centralised specialist centre may be higher quality, but it helps little if reaching it is practically impossible.

The scale test must therefore include travel time, transport cost, digital accessibility, support for vulnerable users and continuity of a local first point of contact. Efficiency on a spreadsheet is not enough if the service becomes inaccessible.

Diseconomies of scale: when larger starts working worse

The same OECD review that confirms economies of scale also highlights diseconomies of scale. Large systems can develop more layers of coordination, longer feedback loops, standards that fit diverse places poorly and weaker accountability. At very large and dense scale, congestion and complexity can also rise.

The political cost is real as well. A pooled service can become more efficient while decision-making becomes less transparent and less responsive to local differences. A larger system therefore cannot measure success only by unit cost; it must also track quality, response time, resilience, innovation and the user's ability to influence the service.

“Subsidiarity in Practice: Which Level Should Decide?” therefore required periodic review. A scale that once made sense is not sacred. Technology, population and transport can change; functions can later be divided differently or decentralised again.

Check cooperation before merger

If three municipalities cannot each maintain a specialist, it does not follow that they must politically merge. “How Can Communities Cooperate Without a Single Center?” showed a full spectrum of intermediate options: a shared service, consortium, contract, federative body or polycentric network.

The OECD's 2026 report on inter-municipal cooperation explicitly notes that such cooperation can capture economies of scale and improve technical capacity in water, waste, infrastructure and spatial planning without municipalities losing their entire autonomy.

This is one of this article's most important lessons: larger operational scale does not necessarily require a larger political ruler. The shared function can be large while the mandate of the shared body remains narrow.

“Subsidiarity in Practice: Which Level Should Decide?” already noted that the same public function contains standards, financing, delivery, supervision, data, specialised support and appeals. This article uses that insight as its main tool for capturing scale economies without unnecessary centralisation.

In health care, the first point of contact can be local, specialised care regional, safety standards national and very rare reference expertise international. In energy, part of generation can be local, distribution regional, transmission wider and technical standards supranational. In education, the everyday relationship can remain local while specialist teachers or laboratories are shared across schools.

Instead of asking 'how large should the institution be?', it is often better to ask 'which sub-function needs which scale?'

Practical audit: is larger scale genuinely justified?

Before moving a function to a wider level, use the following audit:

1. Which fixed cost or rare capacity is actually used better at larger scale? 2. Does quality improve with higher case volume or deeper specialisation? 3. Is this a physical network whose parts must be coordinated across territories? 4. Does a wider user or financing pool make possible a service the smaller unit cannot carry alone? 5. What is the saving after transport, connection infrastructure, administration and maintenance are included? 6. What is lost through distance — time, local knowledge, responsiveness, access or democratic control? 7. Can the same scale benefit be achieved through cooperation or a shared service rather than political merger? 8. Can only the specialised sub-function be centralised while the first point of contact stays local? 9. What backup path exists if the large shared node fails? 10. When will the scale be reviewed again, and against which indicators?

If the main advantage of the wider level remains merely 'it is easier to administer this way', the evidence is not yet strong enough. But if larger scale demonstrably lowers total cost, improves quality, enables rare expertise, increases access or pools risk — and those benefits exceed the new costs — the wider level can be the correct subsidiarity decision.

This article is a deliberate counterweight to any romanticisation of localism. A small community is not freer if its size prevents it from providing a safe operation, demanding diagnostics, reliable water, a reserve transformer or access to specialist knowledge. Dependence on a wider network is not in itself a defeat of sovereignty; it can be rational cooperation.

But the reverse safeguard matters equally. An economy of scale in one sub-function does not automatically grant a mandate to centralise the entire sector. Wider scale is justified where it is demonstrably necessary, and only for as much of the function as that reason covers.

The the articles from “Subsidiarity in Practice: Which Level Should Decide?” through “When is larger scale genuinely necessary?” network arc therefore ends with a balanced rule: start with the lowest capable level; then cooperation among units; shared standards and open connections; and larger scale where fixed costs, rare expertise, physical networks, risk pooling or equal access demonstrate a genuine advantage. “After Harm: Restitution, Protection and the Limits of Punishment” will leave the question of scale and open a different problem: what should happen after serious harm has already occurred?