A zoning map can be extraordinarily precise and socially mute.
It can tell us that a parcel is residential, how intensely it may be developed, how tall buildings may become, where streets and public facilities belong, and which parts of a site must remain open. With enough regulatory detail, the physical envelope of a future neighbourhood can be established before a single foundation has been poured.
Yet residential capacity says surprisingly little about the housing that will ultimately occupy it. The same authorised floorspace could become large owner-occupied flats, small rental units, subsidised housing, public housing, investment property, or some mixture of them. The built volume could remain identical while the households able to access it, the tenure structure and the long-term affordability of the stock differ radically.
This is not evidence that spatial planning and housing policy occupy separate worlds. Planning already distributes several of the most consequential ingredients of housing: permission, location, development capacity and, often, land value. Where housing may be built, how readily capacity can adjust, and what obligations accompany newly created rights all shape the housing system before a rent subsidy, mortgage or eligibility rule reaches an individual household.
The comparison that follows therefore does not treat Tokyo, Vienna, Munich, São Paulo, Singapore and Yerevan as formally equivalent cities or their plans as equivalent legal documents. It asks a narrower question: at what points do planning and land policy intervene in the process that turns development capacity into housing, and what do they require beyond the simple permission to build?
What zoning makes visible: how selected planning systems encode land use, development intensity and housing obligations.
Where planning enters the housing-production system
The comparison becomes clearer if land value is not treated as a stage in a linear housing chain. Planning can intervene through at least five different functions: land or site control; spatial designation; development control; housing conditions or programmes; and delivery or production. Land value is an economic effect that can arise when those rules change. Affordability and household access are outcomes produced by the interaction of all of them.
The six cases occupy this framework differently. Tokyo is examined chiefly through use zoning and intensity controls that keep residential capacity comparatively responsive.
Vienna combines land assembly, a subsidised-housing designation and a wider subsidised-housing delivery system. Munich attaches SoBoN obligations when new planning creates significant land-value uplift. São Paulo uses ZEIS to establish a spatial claim for social-interest housing before an individual project is delivered. Singapore places statutory land-use planning alongside a state-land system and a public producer, HDB. Yerevan has a legally consequential spatial-planning chain, but the reviewed planning instruments do not systematically translate residential capacity into an affordable-housing requirement.
Where housing affordability enters the planning and development process
Capacity before affordability: Tokyo
Tokyo makes the capacity question unusually visible. Japan's use-zone system is nationally standardised, while metropolitan and municipal authorities allocate zones and layer additional controls. Residential buildings are possible across several use categories, and intensity is regulated through instruments such as floor-area ratio, building coverage and district-level controls. The system is highly regulated, but it does not make every increment of housing dependent on an exceptional rezoning event.
That matters because scarcity can be produced by regulation as well as by land itself. Where replacement, infill and intensification remain ordinary possibilities, the planning system can accommodate demand without forcing every increase in housing supply through a singular political bargain. This does not make the resulting dwellings affordable to every household. It addresses a different problem: whether legal development opportunity is unnecessarily scarce.
The distinction has become more explicit in Tokyo itself. In June 2026, the Tokyo Metropolitan Government revised several urban-development incentive systems so that affordable housing can be encouraged alongside private development. The change is important precisely because it does not replace capacity with affordability. It adds a targeted housing condition to selected bonus and redevelopment mechanisms within a planning system whose core comparison here remains use and development intensity. [2]
The national production series provides context rather than proof of causation. Japan continued to record substantial housing starts per 1,000 residents through a period of population decline, with a clear pandemic-era dip. That pattern is consistent with comparatively elastic housing production, but it should not be read as a Tokyo-specific estimate or as evidence that zoning alone caused the construction rate. [3]
Japan: housing production continues as population declines. Housing starts per 1,000 residents, 2015-2024. The national series is contextual evidence for the capacity argument, not a causal estimate of Tokyo zoning.
Source: MLIT housing-start statistics and Statistics Bureau of Japan population estimates/censuses; 2024 starts are the estimate used in the figure.
When a housing condition enters the plan: Vienna
Vienna intervenes at a different point. Its planning framework can attach a subsidised-housing condition - Gebiet für geförderten Wohnbau (GF) - to qualifying areas and plan changes. The important point is not that Vienna has invented another generic residential land-use colour. It is that part of the residential capacity created or reorganised through planning can arrive already subject to a housing condition.
For larger relevant developments, the planning basis generally requires about two-thirds of the residential usable floor area under the GF condition to be subsidised housing. [4] That is more precise than saying that Vienna simply reserves two-thirds of all new housing: the rule applies through a specific planning instrument, to a specific legal base, under specified conditions.
The timing changes the economics of development. If land is rezoned or substantially intensified first and affordability is addressed only afterwards, an affordable-housing provider must compete for land whose price already reflects the newly granted development potential.
A housing condition attached at the planning stage constrains what can be done with those rights from the outset and therefore becomes relevant to the land price that can be sustained.
GF also makes sense only as part of a wider housing system. Vienna does not rely on the designation alone: wohnfonds_wien acquires, banks and develops land for subsidised housing and coordinates development competitions and other delivery mechanisms. [5] The planning condition and the delivery apparatus are analytically distinct, but their connection is precisely what allows a planning rule to become housing rather than remain a notation on a map.
The land-use decision can therefore become an instrument of housing policy without becoming the whole of housing policy. That distinction matters throughout the comparison.
The invisible economic drawing: Munich
Munich introduces another dimension: the value created by planning. A plan normally depicts use, density, streets and built form. It does not depict the change in economic value that may occur when a low-intensity parcel receives additional development rights or a non-residential site becomes suitable for housing. Nothing has yet been built, but the legal possibilities of the land - and therefore its economic position - have changed.
Munich's Sozialgerechte Bodennutzung (SoBoN) makes that second effect part of the planning bargain. The city applies SoBoN in development-plan processes that generate costs and burdens and produce a significant increase in land value for planning beneficiaries. Under the current basic model on private land, 60 per cent of newly created residential development rights are in subsidised or price-controlled housing, 20 per cent in privately financed rental housing and 20 per cent in privately financed owner-occupation.
Relevant rental and affordability bindings extend for forty years, and planning beneficiaries also contribute to infrastructure and land requirements. [6]
The distinction between rights and uplift is essential. The 60/20/20 allocation is not a division of the land-value uplift into three percentages. It describes the housing mix attached to newly created residential development rights. The uplift matters because SoBoN applies where the planning process produces a significant increase in land value, making that economic effect part of the basis for the bargain. SoBoN is therefore neither simply a zoning category nor a mechanical division of uplift; it is a framework that connects a value-creating planning change to housing, infrastructure and land obligations.
In Munich, SoBoN applies where planning creates substantial land-value uplift and links the new rights to housing, infrastructure and land-related obligations.
Reserving a spatial claim: São Paulo
São Paulo moves the intervention towards the spatial designation of land itself. Its Zonas Especiais de Interesse Social (ZEIS) identify areas in which social-interest housing receives a protected planning claim. The significance of that move is temporal as much as spatial: lower-income housing is recognised before the land has been left to compete entirely with higher-value uses.
The designation is not merely symbolic. In qualifying ZEIS developments, minimum shares of computable built area must be devoted to social-interest housing. Under the current framework, ZEIS 1-4 require a minimum 60 per cent share for HIS-1, while ZEIS 5 requires a minimum 40 per cent share for HIS more broadly. [7] The precise categories and eligibility thresholds matter, but the larger planning logic is straightforward: a social purpose is attached to territory before a specific building proposal determines the final use of that land.
Cities routinely protect future land for roads, parks, schools, utilities and environmental systems. Affordable housing is often dealt with later, after the land market has already sorted locations by price. The result is familiar: public or non-profit providers search for the land they can still afford, and social need is displaced towards places where transport, employment and services may be weaker. ZEIS reverses part of that chronology by giving a future lower-income resident a spatial claim earlier.
Designation, however, is not delivery. A ZEIS boundary does not by itself supply finance, assemble land, appoint a provider or guarantee completion. Its value in this comparison is precisely that it solves a different problem from Vienna or Munich: it protects a social-interest use in space.
When planning has a producer: Singapore
Singapore brings planning and production unusually close together. The Master Plan 2025 is the statutory land-use plan that guides permissible land use and density over the medium term. It does not, by itself, constitute the public-housing programme or confer a universal public-housing quota on every residential site. [8]
What changes the housing outcome is the wider institutional system around that plan. State land provides a supporting land base, while the Housing & Development Board plans towns, develops public housing and delivers flats at scale. HDB describes its role not only as a housing authority but as a planner and developer of towns. [9] The public-delivery channel is therefore institutional rather than a percentage printed on the Master Plan.
This is why Singapore is useful but difficult to imitate. The lesson is not that another city should copy a particular HDB share or insert the letters HDB into a zoning legend. It is that planned residential capacity and delivery capacity are different things. Singapore has built institutions that connect them. A city can otherwise plan a site perfectly well while possessing no reliable mechanism by which the desired housing will actually appear there.
The same institutional connection also extends beyond the dwelling itself. HDB plans towns with amenities and community infrastructure as part of the public-housing environment, so delivery is conceived at neighbourhood scale rather than as the production of isolated units. [10]
The same capacity, different housing systems
To make these differences visible, imagine that each planning system accommodates the same 100,000 m² of residential floorspace. The number is deliberately arbitrary. It provides a constant quantity against which the institutional content of residential capacity can be compared.
Under Tokyo's ordinary zoning system, that capacity may be accommodated through a comparatively responsive regulatory framework without a general affordable-housing quota. In Vienna, a qualifying planning change can make roughly two-thirds of the relevant residential usable floor area subject to a subsidised-housing condition. In Munich, newly created residential rights can be divided 60/20/20 under the SoBoN basic model while also carrying infrastructure and land obligations. In São Paulo, land within ZEIS can carry minimum social-interest housing requirements before a particular project is delivered.
In Singapore, a public producer can translate planned land into public housing at scale without a universal site-level percentage. In Yerevan, the reviewed planning instruments can define use and development capacity without attaching an equivalent citywide affordable-housing condition.
These percentages should not be compared as performance scores. Vienna, Munich and São Paulo use different denominators, triggers and legal categories; Tokyo and Singapore illustrate mechanisms that do not depend on a universal site quota at all. The thought experiment does not standardise unlike systems. It makes visible how very different legal and institutional consequences can follow from the same nominal quantity of residential capacity.
The same nominal residential capacity can carry very different planning, housing and delivery conditions. Quota bases differ: Vienna = residential usable floor area under GF; Munich = newly created residential development rights under SoBoN; São Paulo = minimum built-area shares in ZEIS.
Yerevan: planning capacity, housing need - the missing link
Armenia's planning law gives spatial planning real regulatory force. The community General Plan and the urban-development zoning project are distinct instruments. The General Plan establishes the primary spatial framework, while zoning can regulate permitted use and development criteria including density, building height and the relationship between built and green surfaces. Urban-development activity that conflicts with zoning requirements is prohibited. [11]
Yerevan's current General Plan was adopted by the Council of Elders on 10 September 2024. It should not be described as a parcel-level zoning map: operative zoning is a separate planning layer. [12] That distinction is important because the argument is not that the General Plan should contain every housing-policy rule. It is that the mandatory planning chain can create residential development capacity with major economic consequences while the socioeconomic content of that capacity is determined elsewhere.
The reviewed planning instruments can answer where housing may be built and under what spatial parameters. They do not systematically connect that capacity to a citywide affordable-housing requirement comparable to Vienna GF, to a SoBoN-type obligation attached to planning-created value, or to a ZEIS-type social-interest reservation.
This is a statement about the planning instruments reviewed here, not a claim that Armenia lacks separate housing, social-protection or public-land policies.
The gap becomes clearer when the two sides are placed next to one another. The spatial-planning chain can move from General Plan to zoning project to permits and construction and end in residential development capacity. The housing-policy questions begin with need: which households require housing, under what tenure, at what affordable rent or price, with what land, finance and delivery mechanism, and with what protection of long-term access. Those questions do not form a single automatic causal sequence, but they all have to be resolved somewhere if a residential target is intended to answer a housing need.
A parcel of public land illustrates the stakes. It can be sold after development rights are secured, retained for a public or non-profit housing programme, leased with long-term affordability conditions, or used in an agreement that exchanges part of planning-created value for public obligations. The resulting built mass might be almost identical. The housing system would not be.
Yerevan’s planning system creates residential capacity without a systematic link to housing need.
From residential capacity to housing outcomes
The six cases are not a ladder with Yerevan at the bottom and a model city at the top. They address different failures. Restrictive planning can make development opportunity artificially scarce. A generous supply of market housing can still exclude households whose incomes cannot reach market rents. Inclusionary obligations that ignore development economics can suppress production. Social-housing designations without land, finance or a capable provider can remain unbuilt. Public housing delivered far from transport and services can reproduce spatial disadvantage. A value-capture system can work only where planning creates an uplift and the public authority can enforce the bargain.
The more useful distinction is therefore between residential capacity and the capacity of a housing system to turn that permission into accessible homes. The first can be expressed in hectares, floor area, units and density. The second depends on how land is controlled, what the plan designates, how development rights are regulated, whether housing conditions are attached, who can finance and deliver the dwellings, and whether affordability and access endure after completion.
This does not mean every zoning plan should contain an affordable-housing percentage. Tokyo demonstrates why the ability to add housing matters in its own right. Vienna shows what happens when a subsidised-housing condition enters planning. Munich makes planning-created value available to a negotiated system of obligations. São Paulo protects a social-interest claim in space. Singapore links planning to a public delivery institution. The relevant question is not which model wins, but which problem a city is trying to solve - and what its planning decisions are already doing before the housing policy arrives.
Yerevan already regulates the spatial side of this process in considerable detail. The unresolved question begins where the residential polygon ends: what legal and institutional consequences follow from the development capacity the plan creates, and which households can actually access what is produced?
A residential target is therefore not an answer to a housing question. It establishes the capacity to build. The political question is what conditions, institutions and claims a city attaches to that capacity — and what kind of housing it ultimately allows that capacity to become.
Sources and notes
[1] Ministry of Land, Infrastructure, Transport and Tourism (MLIT), Japan. Introduction to the urban land-use planning system; Shinjuku City, current use-zone planning map.
[2] Tokyo Metropolitan Government, Bureau of Urban Development. Revision of urban-development incentive policies to encourage affordable housing in conjunction with private urban development, 30 June 2026.
[3] Ministry of Land, Infrastructure, Transport and Tourism (MLIT), Japan, housing-start statistics; Statistics Bureau of Japan, population estimates and censuses. The 2024 starts value is the estimate used in the figure.
[4] City of Vienna. Planning basis for Gebiet für geförderten Wohnbau (GF) and the Wien-Plan settlement-development provisions; the usual benchmark is two-thirds subsidised housing in qualifying cases.
[5] wohnfonds_wien. Land acquisition, land banking and development of sites for subsidised housing; developer competitions and quality-assurance procedures.
[6] Landeshauptstadt München. Sozialgerechte Bodennutzung (SoBoN), current 100-point model and basic model.
[7] Municipality of São Paulo. Strategic Master Plan / ZEIS provisions and current implementing regulation, including minimum computable built-area shares for HIS in ZEIS.
[8] Urban Redevelopment Authority, Singapore. Master Plan 2025: statutory land-use plan showing permissible land use and density.
[9] Housing & Development Board, Singapore. Our Role; HDB plans towns and develops public housing.
[10] Housing & Development Board, Singapore. Planning and design of HDB towns, amenities and community facilities.
[11] Republic of Armenia. Law on Urban Development, current consolidated text, including zoning and development-control provisions.
[12] Yerevan Council of Elders. Decision approving the new General Plan, 10 September 2024.