Planned but Not Built: Why the Netherlands Cannot Deliver Enough Homes
The Netherlands has ambitious housing targets and a vast plan inventory. This explainer follows the chain to a completed, usable home and shows how fragmented responsibility can amplify real constraints.
On 1 July 2026, the Netherlands largely brought into force a law intended to restore public control over housing delivery. The Wet versterking regie volkshuisvesting, roughly the Act to Strengthen Direction over Public Housing, strengthens national and provincial powers to steer housing programmes, locations, and the groups those programmes are intended to serve. It also promises shorter procedures and a clearer division of responsibility. For years, the state announced targets that municipalities, housing associations, developers, infrastructure operators, and courts were somehow expected to assemble into homes. It has now written “direction” into the law’s name.
The timing invites an obvious comparison. In 2025, the Netherlands registered 69,200 newly built homes, plus a net 10,700 homes from conversions and other changes. After 9,500 demolitions, the registered housing stock grew by about 70,000. By contrast, the government’s spring 2026 inventory contained 823,400 potential homes for 2026 through 2030, of which 386,200 were described as approved by a municipal council.
That looks like an enormous conversion failure. It is not a conversion rate at all. The figures cover different periods and measure different stages. A proposed home in a rolling plan inventory is not the same thing as an approved environmental plan, a permit, a financed project, a construction start, a completed dwelling, or an occupied home. Dividing one number by another would produce a satisfyingly precise answer to a question the data do not ask.
The new law cannot yet be judged by completions. Most of it had been in force for barely a month when this article was written, while important programme and urgency duties remain phased. For now, its significance is diagnostic. The law recognizes that the Netherlands does not merely need a larger national number; it needs institutions capable of turning that number into serviced land, lawful decisions, viable tenure programmes, finance, construction, allocation, and use.
The provisional answer is that fragmented authority and accountability amplify genuine constraints. Housing projects can fail because the project appraisal no longer balances, a grid connection is unavailable, nature law has not been satisfied, a site is unsuitable, staff are missing, or local costs have not been resolved. Those constraints are real. The deeper problem is that the power, money, information, risk, and duty needed to address them are distributed among institutions that control different stages, operate on different timetables, and do not necessarily bear the full social cost when the final home does not appear.
The new law meets the old pipeline
The Wet versterking regie volkshuisvesting tries to repair the front of the pipeline. Staatsblad 2026, 156 amended the Omgevingswet, Huisvestingswet 2014, Woningwet, and other legislation so that national, provincial, and municipal governments can coordinate housing programmes and, where necessary, issue instructions. The accompanying commencement decree brought the law largely into force on 1 July, but expressly excluded several provisions from immediate commencement. National and municipal housing programmes were expected by 1 July 2027, provincial programmes six months later, and mandatory municipal urgency arrangements by 1 January 2028. The detailed implementing decree was still in the Council of State advice phase and was expected, if adopted and commenced on schedule, to take effect on 1 January 2027.
That sequence matters. The statute has already altered powers and legal terminology, but the full programming system is not yet operating as a settled administrative routine. Existing municipal housing visions can remain in place during the transition. Provinces and municipalities can prepare programmes before the formal deadlines. Some procedural provisions are operative while related delegated rules are still to come. “The law took effect” is true, but too crude. It does not explain which obligations can already change a project and which still depend on later implementation.
The affordability rules require a sharper distinction. The statute creates the delegation and regional-programming architecture, but the draft implementing decree proposed the detailed formula: two-thirds of programmed new homes would be affordable and 30 percent would be social rental housing at regional level. It also proposed differentiated municipal programming, regional compensation, and limited departures. Those rules had not yet been adopted, published, or commenced when this article was revised. Official implementation guidance describes the intended system in the present tense, but guidance cannot turn draft delegated legislation into operative law. If the decree is adopted substantially as drafted, it would not impose an identical 30 percent social-housing formula on every municipality or project. That would be a regional allocation system, not a national cookie cutter.
Stronger direction can improve the plan inventory because the enacted statute gives higher government stronger programme and instruction powers. Once the remaining programme provisions and delegated rules operate, a national target can be translated into regional programmes, provinces can press municipalities to identify sufficient locations, and affordability requirements can be distributed rather than merely announced. These can be meaningful changes because ambiguity at the planning stage can migrate downstream as redesign, renegotiation, or litigation. They remain changes to the front of the pipeline, not proof that the rest of it will move.
But a programme is still a programme. It does not purchase the land, clean contaminated soil, expand a substation, finance a school, lower a legacy land price, complete an ecological assessment, recruit municipal planners, secure a contractor, or persuade a lender that the rent and sales assumptions remain credible. The state can require a municipality to reserve capacity for homes. It cannot make every underlying input appear through statutory syntax.
The latest national figures make the distinction concrete. The spring inventory’s 823,400 potential homes is large relative to recent annual output, but only 386,200 had reached the monitor’s council-approved category. Even that category does not mean “ready to build.” A council-approved plan may still lack land control, a final environmental decision, infrastructure, a permit, a balanced project account, a construction contract, or committed finance. Meanwhile, the same official forecast expects 99,700 homes in 2027. That is evidence that a production recovery is considered possible. It is not an observed result, and it certainly is not 99,700 occupied homes waiting politely in the future.
The relevant question is not whether planners have typed a sufficiently large number into a national spreadsheet. It is where executability can be lost after a home enters that spreadsheet, and why existing arrangements do not consistently carry every necessary input across the boundary to the next stage.
What counts as a missing home
The headline shortage first needs a definition. The official 2026 method estimates a shortage of about 384,000 homes, or 4.6 percent of the dwelling stock. It starts with roughly 427,000 defined housing seekers and subtracts about 43,000 dwellings treated as available. The seeker side includes potential new households and people in specified living arrangements who are assumed to want an independent home. The available side is not simply every address that happens to be empty.
This is a model of market imbalance. It is not a headcount of people sleeping outside, a tally of everyone on a social-housing waiting list, or a count of every household paying too much rent. It does not measure every unsuitable home, every couple postponing separation, every adult child who would prefer to leave the parental home, or every worker unable to find housing near a job. Some of those experiences influence the model; others require separate evidence. Folding them into one emotionally comprehensive number would make the figure more dramatic and less meaningful.
The method also uses 2 percent of the housing stock as the point at which the market is treated as broadly balanced. On that convention, the policy task is not necessarily to erase the entire 4.6 percent estimate; the government calculates the gap to the 2 percent benchmark separately. The benchmark is administratively useful, but it is not a law of nature. Reaching it would not abolish homelessness, waiting lists, affordability problems, poor-quality housing, or local scarcity in Amsterdam, Leiden, Utrecht, or Eindhoven.
These distinctions affect what counts as a solution. A high-priced apartment can add to the physical stock and reduce aggregate pressure while remaining inaccessible to the household with the most urgent need. A social rental home can improve affordability and allocation for a selected household even if it replaces another dwelling and adds little net stock. Converting an office can create a registered home; splitting a large unit can create two; demolishing obsolete homes can reduce the stock before replacement arrives. A completed building can remain temporarily unoccupied because a utility connection, legal-use decision, sale, or allocation is pending.
The 2025 production figures therefore need to remain separate. CBS provisionally recorded 69,200 new-build completions. Other additions and withdrawals, including transformations and changes of use, made a net contribution of 10,700. Together they produced about 79,900 additions before demolition. After 9,500 demolitions, the registered stock grew by about 70,000. None of these figures directly reports how many households moved into a suitable and affordable home during the year.
Demand also complicates any simple population-versus-construction story. CBS projects about 567,000 additional households between 2025 and 2035, with one-person households accounting for roughly 72 percent of that increase. Population growth, migration, separation, aging, and young people forming households all matter, but the composition of growth matters too. A country adding many one-person households needs a different mix of units than one adding the same number of people in families of four.
The demand argument gets something important right. Even a well-run delivery system would face a substantial task because household growth continues and the inherited shortage is already large. It does not follow that the present supply chain is performing as well as reasonably possible, or that migration alone explains why approved projects lack infrastructure, why municipal teams cannot complete agreements, or why viable tenure mixes are renegotiated late. Demand determines the scale of the challenge. It does not excuse losses inside the delivery process.
The outcome that matters for this article is narrower than the official shortage and more demanding than the completion count: a legally usable home, in a location and tenure that someone can actually access, has been occupied. In the sources reviewed for this article, no representative national dataset connected that outcome back through every prior project stage. The state can report plans, permits, programmes, grants, and stock changes, but the published national record does not yet show, with stable project identifiers, where the pipeline lost each unit or month.
The long chain from target to key
A national housing target begins as a political and demographic judgment. It becomes a regional allocation, then a location, then an entry in a plan inventory. After that, a project still needs control of the land, a legally usable spatial decision, a workable tenure and unit mix, infrastructure, permits, finance, a contractor, a start, completion, legal use, allocation, and occupation. These stages do not form a neat government conveyor belt. They are a chain of conditional commitments owned by different actors.
Consider a municipality that identifies a station-area site for 2,000 homes. The municipal council can amend the omgevingsplan, the environmental plan that contains the local rules for the physical environment. Under the Omgevingswet system, that decision must integrate housing with transport, noise, soil, water, nature, safety, public space, and other legally relevant interests. The developer may control only part of the land. The grid operator may have no capacity for a large connection. A school or tram may require national or provincial money. The affordable share may reduce expected revenue. The landowner may still expect a price based on an earlier, more profitable programme. Each problem can be manageable in isolation and fatal in combination.
The sequence is unforgiving. A lender may not commit until the permit is sufficiently secure and enough sales or rental income is contracted. A contractor may not hold a price while an environmental decision remains exposed. A municipality may hesitate to fund infrastructure until the project is certain, while the project cannot become certain without the infrastructure. The grid operator cannot reserve capacity indefinitely for an immature plan, but a mature project may be unable to start after joining a queue too late. Every actor can have a defensible reason for waiting. Together, those incentives can keep a project waiting for years.
Land control introduces another boundary. A planning decision can increase the potential value of land enormously, but it does not transfer ownership. A municipality may buy, enter a development agreement, use pre-emption rights, or ultimately pursue expropriation under legal conditions. Each route requires money, expertise, time, and political tolerance for risk. A private owner can refuse a residual price that has fallen below earlier expectations. A developer that bought at the top of the market may be unable to accept the write-down needed to support a new affordable programme. The plan exists, but the price embedded in the land keeps the old plan alive.
Once a municipality allows development, it must also secure the public costs that the project creates. Under the Omgevingswet, cost recovery for area development can be arranged through an agreement or public rules. Roads, public space, planning work, and other attributable costs do not disappear because housing is socially valuable. Yet the amount, timing, and allocation of those costs can make a marginal project unfinanceable. This does not prove that municipalities systematically extract the maximum possible contribution. It shows that public amenities and project viability meet in the same account, often before anyone has decided which taxpayer, landowner, developer, buyer, or tenant should carry the difference.
Permits provide a useful warning against reading stages as outcomes. CBS recorded nearly 86,000 new-build permits in 2025, materially above completions. CBS also changed the permit series to use the national address and building register, the BAG, rather than the former municipal survey. The new method produced substantially higher counts in the overlap period, so a naïve historical comparison would mistake a measurement change for a building boom. Some permits are later withdrawn, and many authorized projects still need finance, presales or leasing commitments, infrastructure, and a final construction decision.
Even the measured interval from permit to completion describes only part of the process. Among homes completed in the first three quarters of 2025, the median permit-to-completion time was almost 23 months. The slowest quartile took more than 35 months. Those figures exclude the years spent identifying the site, assembling land, designing the programme, conducting studies, changing the plan, and negotiating infrastructure. They also exclude projects that never completed. The government’s statement that a full development process takes roughly ten years is therefore a programme diagnosis, not the same statistic measured over a larger window.
Objection and appeal add another dependency, but not one that can be reduced to a generic delay penalty. Residents, competitors, and environmental organizations can challenge decisions within the law. Courts determine whether the government has followed the correct procedure, respected rights, applied the right legal test, and supported its conclusions with evidence. A judgment that annuls a permit may reveal an unavoidable conflict, a curable evidence defect, a poor initial location, or an administrative shortcut that saved no time at all. There is no representative national basis for claiming that appeals routinely add a fixed number of years to every housing project.
The chain continues after completion. The building must be safely and legally usable. Addresses and status must be registered. Utilities and public space must function. A housing association or municipality may allocate the unit under eligibility and urgency rules; a private owner must sell or let it; a household must be able to pay. Completion is a strong production measure, but occupation is the final public outcome.
No single institution controls this sequence. National government can legislate, fund, and direct, but it usually does not own the site or issue the municipal permit. A municipality controls crucial land-use decisions and local land-policy instruments, but cannot order a bank to lend or a grid operator to create capacity. A developer can redesign or finance a project, but cannot legalize an unsuitable location. A court can invalidate an unlawful decision, but cannot repair the project account. The pipeline is not failing because one actor forgot to pull a master lever. There is no master lever.
Nor does the evidence permit a national ranking of the stages. Planning, land, finance, infrastructure, nature law, staffing, local politics, and market conditions all have documented mechanisms, and several have well-documented project effects. In the sources reviewed, no representative national project series and credible counterfactuals showed that one stage accounts for a specified share of missing homes. Any reform claiming to solve “the bottleneck” should first establish which bottleneck it has actually found.
Where the business case breaks
The strongest alternative to the institutional diagnosis begins with arithmetic. A project will not start merely because government has assigned it a target and a postcode. Expected sales or rental income must cover construction, land, infrastructure, planning, finance, taxes, regulation, and the return required for taking development risk. When the total no longer fits under expected revenue, the project is not being ideologically stubborn. The project appraisal does not balance.
Work backward from the finished homes. Estimate what buyers, tenants, an investor, or a housing association can pay. Subtract construction and professional costs, financing during a long and uncertain development period, required public works, taxes, and a risk allowance. What remains is the residual land value. For that project, the land is worth what the permitted programme can support after the other claims have been paid, not whatever price the owner once hoped to receive.
In theory, a policy change that reduces revenue should therefore lower the land price rather than eliminate the project. In practice, adjustment is neither automatic nor painless. The land may have been bought years earlier with debt. Several owners may have different reservation prices. A developer may already have committed substantial predevelopment capital to a design that no longer works. The parties may disagree over who absorbs the loss, or an owner may simply prefer to wait. Economics can identify where the loss should land; it cannot make the affected party volunteer.
This is why “regulation stopped the project” and “the landowner captured the gain” can both be incomplete accounts. A rent cap, affordable-housing requirement, or higher infrastructure contribution can reduce the residual. That effect is real. Whether it reduces the number of homes depends on who can absorb the adjustment, whether the rule was anticipated when land was purchased, whether public support is available, and whether the site can be redesigned. Lower land value, higher density, a different unit mix, public or housing-association ownership, cheaper finance, subsidy, phasing, or a lower required return can restore feasibility. Sometimes none is available at the required scale.
Middle-rental housing shows the trade-off. Since July 2024, the Wet betaalbare huur, the Affordable Rent Act, has extended the points-based rent system into the middle segment. For 2026 contracts, homes with 144 through 186 points generally fall between the social boundary and a maximum middle-rent boundary of €1,228.07 per month. A qualifying new middle-rental home can receive a 10 percent new-build premium for 20 years if construction began before 1 January 2028 and the home entered use after the law took effect.
The government announced in June 2026 that it intended to extend the construction-start window through 2031 and adjust parts of the points system. That proposal was not yet operative law. For projects being financed now, the distinction is material. An investor can model a proposed future premium, but cannot prudently treat it as guaranteed current law while consultation, legal drafting, official advice, and implementation remain pending.
It is plausible that regulated revenues, higher interest rates, taxes, and construction costs have made some private rental projects marginal or unviable. A July 2026 DNB analysis by staff authors estimates that the 100,000-home ambition would require about €40 billion in annual financing, including roughly €6.4 billion for private rental construction. The analysis reports that Dutch institutional investors financed a little more than half of the estimated private-rental requirement in 2025, while international participation in new rental development had fallen sharply. These are model estimates and market observations, not a count of homes that rent regulation alone prevented, and the authors note that their views do not necessarily represent DNB’s official position.
Other data prevent the opposite overstatement. Capital Value, a housing-market adviser with a commercial interest in investment activity, reported €4.8 billion in Dutch residential investment transactions in the first half of 2026, including €2.2 billion classified as new-build investment. Most international capital went to existing portfolios, and international investors accounted for only 1 percent of the provider’s new-build volume. The categories are not CBS construction statistics, and a transaction involving future homes is not a start or completion. Still, the figures show why “investors have returned” and “capital avoids construction” can both mislead. Capital returned unevenly, and meaningful new-build finance continued, largely from domestic institutions and other Dutch actors.
The relevant distinction is between stabilized assets and development risk. Buying an occupied rental portfolio purchases current cash flow, an established building, more observable operating costs, and an earlier or more established exit route. Forward-funding a project, buying it on completion, supplying development equity, or lending against land and construction exposes capital to planning changes, cost overruns, interest during delay, contractor and lease-up risk, and future regulation. The same investor can rationally accept one exposure and reject the other. “Available capital” is therefore not a common pool waiting to be pointed at whichever project government prefers.
Housing associations occupy a different position. They can accept long holding periods, operate under a social mandate, and finance through a sector with public guarantees and regulatory oversight. They are not profit-maximizing developers in disguise. Nor are they exempt from the project account. The National Performance Agreements call for 30,000 new social rental homes per year from 2029, with an ambition to reach that level earlier. A January 2026 assessment found a €19.4 billion gap against a €115 billion investment programme. The June assessment projected that announced measures would increase aggregate investment capacity, enough in theory to address the earlier sector-wide gap, but resources and obligations remained unevenly distributed among associations and regions. The official assessment still described execution as a concern.
This alternative thesis is therefore substantially correct: a more directive state can reproduce underproduction if it assigns projects whose allowed revenue cannot cover their lawful and physical costs. Public agencies and housing associations cannot build below cost indefinitely merely because their required return is lower. The boundary is that project economics are not an external verdict delivered to policy. Government influences land use, tenure, infrastructure, subsidy, tax, finance, timing, and risk allocation. A project residual is the place where those choices meet.
A capable delivery system does not deny the residual. It decides, transparently, whether to reduce land value, increase density, change the programme, invest public money, accept a lower return through a patient owner, move the site, or abandon it. Existing arrangements can postpone that decision until several institutions have built incompatible expectations into the same project.
Constraints that cannot be planned away
Environmental and infrastructure constraints are often treated either as excuses for policy failure or as immovable facts that end the discussion. Both positions avoid the useful question: which limits are legally or physically hard in the short run, which can be changed through investment and sequencing, and which reflect a poor choice of location or design?
Nitrogen provides the clearest legal example. Under EU and Dutch nature law, a project that may significantly affect a Natura 2000 site requires an assessment capable of excluding harmful effects with the legally required certainty. On 18 December 2024, the Administrative Jurisdiction Division of the Council of State changed its approach to intern salderen, internal offsetting. A developer may no longer subtract the site’s existing emissions in the preliminary screening simply to conclude that no nature permit is needed. The offset may still be considered, under conditions, as mitigation in an appropriate assessment and permit decision.
The Council of State applied that doctrine to a 162-home project on previously fertilized land in May 2025. In January 2026, the Council of State confirmed that the same sequencing applies when a municipal environmental plan authorizes development. That is a significant evidentiary and permitting burden. It is not a categorical prohibition on internal offsetting, a universal requirement to purchase nitrogen rights, or a national ban on housing. The practical result depends on the site, the affected habitat, the reference situation, the evidence, and whether the mitigation satisfies the additionality test.
Courts do not balance an unlawful permit against the national housing target and decide that 100,000 is a sufficiently important number. Their job is to apply the legal standard. Government can reduce delay by choosing less exposed sites, collecting ecological evidence earlier, coordinating decisions, improving the quality of assessments, and funding lawful mitigation. It cannot accelerate construction by pretending the standard is optional. A decision that arrives quickly and is then annulled is not fast delivery. It is an expensive rehearsal.
Electricity capacity is a different kind of constraint. In congested regions, a substation or network may simply lack room for a large connection. The Authority for Consumers and Markets’ current prioritization framework governs how network operators allocate scarce transport capacity and permits priority for specified projects serving important social needs. Priority changes the queue. It does not thicken the cable.
The implementation rules also vary by connection. Under the small-user process introduced from 1 July 2026, applications in congested areas can enter a queue rather than receiving automatic capacity, with transition arrangements continuing until 1 January 2027. The effect on a housing site depends on the region, connection type, phasing, heating design, shared facilities, and available flexibility. No defensible universal Dutch grid delay was located that could be attached to every planned home.
Water and transport demonstrate the difference between a hard short-run limit and an investable long-run one. The government has an action programme intended to avert drinking-water shortages by 2030, which means new extraction, treatment, permits, and regional capacity cannot simply be assumed. Wastewater treatment and sewerage can create separate local dependencies. A location on weak soil or in a flood-exposed area may be buildable only at higher cost, lower density, or with substantial adaptation; another location may be a bad bet regardless of the housing target.
These trade-offs are not an argument for never building in the Randstad or always moving growth elsewhere. In modeled 2050 scenarios, Deltares found that restricting construction in flood- and water-sensitive locations could reduce risk while producing large side effects, including displacement of housing to less preferred regions and pressure on other land uses. In those scenarios, allowing higher density reduced some of the side effects. Whether that result carries to a particular location depends on transport, public space, services, infrastructure, and the model assumptions. “Water and soil should guide development” is a planning principle, not a free map of painless sites.
Transport and social infrastructure run on still other timetables. In November 2025, the government announced €2.5 billion for transport measures associated with 273,000 planned homes, plus €877 million for area measures associated with 128,000 homes. These figures show the scale of the dependency. They do not mean the funding created those homes, that all projects were additional, or that the associated dwellings have started. A tram line, road, soil clean-up, or grid measure may be necessary without being sufficient.
Construction capacity interacts with the rest of the pipeline. Industrialized construction can shorten factory and on-site work, but it cannot manufacture a location or a grid connection. The Netherlands Court of Audit found that many centrally purchased flex homes remained unplaced after manufacture because sites, permissions, finance, and local implementation were unresolved. The factory had solved its stage. The system had not solved the rest.
The institutional thesis must remain modest. Fragmentation cannot repeal ecological limits, create water, or make a risky site sensible. It can affect whether evidence is collected early, infrastructure is financed before the project reaches crisis, locations reflect real capacity, and risks are assigned to an actor able to manage them. Some constraints must be respected. Others must be bought, built, redesigned, or avoided. Calling all of them “planning delay” conceals the decision that each one requires.
Why nobody owns the final result
The present division of labor was not designed in one sitting. Before the financial crisis, many Dutch municipalities combined land-use control with active land development: they acquired land, prepared sites, and expected to recover public investment through later sales. Housing associations also had broader development and cross-subsidy capacity. That arrangement could align land, infrastructure, and construction, but it placed property-market risk on public and quasi-public institutions. When the market collapsed after 2008, a CPB review of the subsequent fall in municipal investment identified the property crisis and reduced national contributions as the main causes, alongside substantial setbacks on municipal land positions. Municipalities did not all abandon active land policy, but many had good reason to carry less land, investment, and risk. A national target can be restored much faster than a municipal land department.
Housing associations changed as well. The 2015 Woningwet imposed clearer rules intended to protect social housing and limit financial risk, while narrowing associations’ commercial activities and concentrating their mandate on affordable rental housing. The OECD’s 2025 review treats those reforms, together with the landlord levy, as part of the explanation for reduced activity outside the sector’s core role. The levy was abolished from 2023. This was not the destruction of the association sector; associations continued to own, manage, and build affordable rental housing. It did narrow commercial activity and cross-subsidization and placed a substantial charge on the sector until the levy ended.
Since 2022, national policy has moved back toward direction. Abolishing the landlord levy, negotiating National Performance Agreements with housing associations and municipalities, funding project gaps and infrastructure, and enacting the Wvrv all shift responsibility upward or make national expectations more explicit. They do not recreate the old model. The current system still leaves municipal land-use decisions and many permits with municipalities, construction and much finance with developers and lenders, long-term affordable rental ownership with associations, sector-specific network responsibilities with regulated operators, and review of legality with courts. The Netherlands has rebuilt ambition and legal steering faster than it has rebuilt arrangements capable of carrying delivery risk across the whole chain.
Power is widely distributed because the interests are real and the decisions consequential. National government sets legislation, taxes, subsidies, housing-association rules, infrastructure budgets, and national spatial priorities. Provinces coordinate regional programmes and balance housing against nature, water, landscape, agriculture, energy, and economic development. Municipalities choose locations, adopt environmental plans, issue many permits, negotiate project agreements, recover costs, provide local infrastructure, and face the voters who experience construction first.
Housing associations build and operate affordable rental homes, but need land, planning permission, finance, contractors, and a programme that their regulated revenues can support. Developers assemble land, design projects, arrange finance, and bear market and construction risk, but cannot compel public approval or a grid connection. Investors and lenders can supply capital, but their fiduciary and prudential duties do not include meeting the national housing target. Landowners can wait. Courts review legality. Grid, water, transport, and wastewater bodies have sector-specific duties and plan networks for several users at once. Contractor availability and pricing remain project risks.
Each institution can explain why the missing input belongs somewhere else. National government can point to a municipality that has not approved enough hard plans. The municipality can point to a landowner, an unaffordable infrastructure package, national rent rules, or a grid queue. The developer can point to late public requirements and rising costs. The province can point to an ecologically unsuitable site. The operator can point to a network investment programme that takes years. These explanations may all be accurate. Accuracy at each boundary can still produce collective failure.
The fiscal and political timing reinforces the pattern. National politicians receive visible credit for a target or grant. Municipalities may incur or help finance early costs for staff, consultation, roads, schools, parks, drainage, and disruption. Some revenues may arrive later, in another budget, or not at all. A developer bears predevelopment risk before a permit is secure. A housing association commits to long-term management and affordability. Existing residents experience construction traffic and service pressure now; many prospective residents receive a home later and are not yet identifiable participants in the local process.
Local opposition is therefore not always adequately explained by homeowner selfishness. A neighborhood asked to accept thousands of homes without credible transport, schools, green space, flood protection, or parking management may have identified a delivery problem, not merely expressed a dislike of newcomers. A proposed tower can also impose real local losses, even when the broader project is justified. Democratic participation and legal challenge are not design defects simply because the benefits are dispersed and delayed.
The asymmetry still matters. Many prospective residents are absent from local participation because they do not yet live in, or cannot identify themselves with, the project area. Some objections may remain after reasonable mitigation because local costs are concentrated while regional benefits are broad. A fair system must distinguish impacts that should change the project, impacts that can be mitigated or compensated, and claims that amount to a permanent local veto over regional need. The answer should follow from the measured cost, not precede it.
Public programmes expose the same problem of stage ownership. The Realisatiestimulans gives municipalities €7,000 for each qualifying affordable home whose construction began in the previous year. That is more outcome-oriented than paying for a plan, but the outcome is a start, not a completion or occupied home. It may help cover local costs. Whether it causes additional starts, rather than rewarding projects that would have begun anyway, still requires evaluation.
The Woningbouwimpuls offers a sharper warning. The Netherlands Court of Audit compared funded projects with rejected applicants and found that the programme did not produce faster starts in the studied rounds. It remained uncertain whether the programme produced more homes overall. It did, however, help preserve a higher affordable share in funded projects. A subsidy can therefore succeed at distribution while failing to demonstrate acceleration. “Homes associated with a grant” is not the same category as “homes caused by the grant.”
Responsibility also cannot be inferred from who benefits from scarcity. Existing owners may gain nominal equity, but, as an illustration, a leveraged family moving to a larger home also faces a higher replacement price. A mortgage-free downsizer typically needs a suitable cheaper home to realize much of the gain while continuing to be housed. Landowners may receive planning uplift; developers may earn a margin; lenders receive interest; municipalities may obtain revenue or avoid losses. These are different positions, not a conspiracy with a membership list.
The supported judgment is narrower. Fragmented authority can amplify real constraints because no institution consistently owns the final result and the resources needed to secure it. Several actors can lawfully withhold, or simply be unable to supply, a necessary input. The accumulated social costs of waiting, overcrowding, high rents, delayed household formation, and longer commutes are diffuse rather than assigned to one project actor. Existing arrangements can document why each actor could not proceed. They do not reliably identify who must make the next decision, with whose money, by what date, and what happens if that decision is not made.
What a delivery state would actually do
A delivery state is not a ministry that announces a more ambitious target or a public developer instructed to ignore arithmetic. It is a system that identifies the binding stage of a project, assigns an institution with authority to move it, gives that institution the necessary money, staff, information, and risk capacity, and makes failure visible early enough to change course. Stronger direction is necessary. It is not sufficient.
The first reform should be informational but not merely statistical: a national stage-gate register with stable project identifiers from initial plan entry through occupation. For each material site, the register should distinguish land control, environmental-plan status, nature evidence, infrastructure capacity, permit, finance, start, completion, legal use, allocation, and occupation. National government could establish common definitions and reporting duties, using existing powers where sufficient and new statutory or contractual authority where necessary. Provinces and municipalities could maintain project records, while operators and programme administrators could contribute stage data under clear data-sharing, confidentiality, and privacy rules. Commercial details can be protected without allowing every delayed project to become a fresh anecdote.
Such a register would not build a home. It would make it harder for plans to appear mature because a missing field is hidden, reduce the temptation to treat rolling inventories as delivery, and allow government to test whether a reform moved the intended stage. It could also help expose optimistic dates and double counting. The risk is that reporting becomes another administrative product polished for the next housing summit. That risk can be reduced by using a small set of auditable gates and linking public money to verified transitions rather than self-reported confidence.
Second, selected strategic sites need an accountable lead institution. Depending on the project, that could be the municipality, province, or national government. The lead should have an agreed programme, budget, milestones, access to specialist staff, and a defined route for escalating unresolved decisions. The Wvrv can strengthen programme direction and instruction, but it does not by itself provide site-acquisition budgets, infrastructure command, data-sharing authority, or general cross-sector escalation powers. Those would require specific legal authority, agreements, safeguards, and funding. Centralizing the meeting calendar without centralizing any capacity would merely produce better minutes.
The lead institution should not override every local or sectoral judgment. It should identify unresolved decisions and force them into the open: the project either receives grid and transport investment, changes its energy or mobility design, reduces scale, moves phase, or leaves the programme. An affordable programme is supported by a lower land value, association ownership, subsidy, cross-subsidy, or a revised design, or its cost is acknowledged explicitly. A nature conflict is lawfully mitigated, redesigned, or treated as a reason not to build there. Escalation should end ambiguity, not suspend law.
Third, national funding for municipal and provincial implementation capacity must become sustained rather than episodic. Planners, land economists, environmental specialists, project directors, lawyers, and permit staff are production inputs. The government’s 2026 acceleration package recognized this by allocating money for implementation and digitalization. Temporary grants do not guarantee durable teams. Shared regional expert units and multi-year funding would be more credible where individual municipalities cannot maintain specialist capacity.
Capacity should support parallel planning where the risk is manageable. Land negotiation, infrastructure design, environmental evidence, procurement preparation, and finance can proceed concurrently instead of waiting for every preceding document to become final. Standard methods and reusable evidence can reduce reinvention, especially for recurring building types and well-understood impacts. Parallel work shifts some spending earlier and creates a risk of sunk costs. That trade-off can be acceptable when the alternative is discovering a fatal dependency after years of sequential optimism.
Fourth, infrastructure must be treated as part of the housing programme rather than an external condition checked near the end. National government, provinces, municipalities, and regulated operators should jointly identify the grid, water, wastewater, transport, school, and public-space investments needed for strategic locations, with budgets and delivery dates tied to housing phases. Public money should go first to sites that are otherwise lawful, viable, and likely to proceed, not to the projects most skilled at attaching a large housing number to an infrastructure request.
Infrastructure-first funding is expensive, and some investment will precede certainty. The alternative is to ask a developer to finance a network-scale asset through the first phase of homes or to discover that a public operator’s capital plan arrives years after the permit. Costs should be allocated transparently among general taxpayers, network users, land value, and project contributions. Where new planning permission creates a substantial land uplift, active land policy or a carefully designed planning-gain charge can help fund public works. DNB’s indicative analysis of a possible plan-gain tax shows that the potential base is material, but it also warns that owners still need a return to bring land forward. The estimated base is not a pot of collectible cash waiting at the municipal treasury.
Fifth, project-gap finance should be milestone-based and expose the residual. A municipality or national programme should identify the verified gap after realistic revenue, land, construction, infrastructure, finance, and risk assumptions. Support can then be paid at meaningful stages, with clawbacks if costs fall, the affordable programme shrinks, land is sold at a windfall, or the project fails to proceed. This will not make additionality perfectly observable, but it is better than counting every home attached to an award as a public-policy creation.
Sixth, housing associations should receive better access to suitable land and finance where long-term affordable rental ownership is the objective. That may require municipal land allocation, lower or deferred land payments, guarantees, project subsidy, or redistribution within the association sector. The implementing institutions are national government, municipalities, associations, and the guarantee and supervisory system, not “the public sector” in the abstract. Expansion must also recognize regional financial differences and the need to maintain existing homes. An association cannot spend the same euro on a new unit, a leaking roof, and energy renovation.
Private middle-rental construction needs a stable rule set rather than a promise of permanent deregulation. Tenant protection has an explicit social purpose; investors need a credible account of future revenue. Government should settle the new-build premium and points-system changes through law, preserve clear protections for existing contracts, and avoid repeated late changes that shift the residual after land and design are fixed. Where regulated rents cannot support the desired project, the choice should be visible: lower land value, public support, association delivery, a different programme, or no project. Hiding the gap inside optimistic assumptions serves neither tenant nor builder.
Finally, strategic growth needs an explicit spatial and distributional bargain. Selective density around transport can economize on land and support public transport, but may require expensive public space and local services. Edge or greenfield development can add scale, but may create transport, water, landscape, flood, and ecological burdens. The balance is location-specific, and the Deltares scenarios do not establish a universal ranking. National and provincial government should compare locations using those full costs, then fund credible local benefits and mitigation. Existing communities should not be handed the bill for a national target; neither should they receive an uncompensated permanent veto over future residents.
This package is less dramatic than announcing that government will build everything or that deregulation will unleash the market. It is also more demanding. It requires the state to distinguish stages, reject non-executable plans, carry financial risk where public goals justify it, accept that some land values must fall, invest before scarcity becomes a crisis, and tell voters who pays. Private developers, lenders, investors, associations, and contractors remain essential because public direction does not replace design, finance, construction, or long-term operation. Public authority remains essential because markets cannot create legal permission, decide how public affordability goals should be distributed, or finance every network and social cost through individual projects.
The Netherlands does not lack institutions that touch housing. The Wvrv can improve direction at the top of the chain. Its success will depend on whether the country also builds the capacity, infrastructure, viable projects, land arrangements, and accountability needed below it.
Continued delay risks perpetuating a familiar pattern: large plan inventories, recurring grants, rising expectations, and scarcity reflected in existing-property and land values, while prospective residents wait for the system’s many individually reasonable decisions to add up to a key. They will not do so by themselves.