What the Government’s planning reform has - and hasn’t - achieved

“The difficulty lies, not in the new ideas, but in escaping from the old ones,” wrote John Maynard Keynes, one of history’s greatest economists but, of course, not a town planner.

His observation nevertheless captures the central challenge of planning reform. National policy can change overnight, but the assumptions, processes and practical constraints that govern development may persist for years.

Reform may therefore open new routes to housing without clearing the obstacles that prevent sites from progressing along them. For a policy change to produce a new home, its effects must travel through every stage of the development process.

A ministerial announcement must first become operative policy. That policy must then influence land decisions, shape a viable application and result in an implementable planning permission. Even then, the scheme must overcome the demands of technical design, infrastructure, finance and construction. Progress at one stage does not guarantee progress at the next.

Since July 2024, the Government has introduced substantial changes, from restoring mandatory housing targets and creating the Grey Belt route to reforming local plans and publishing the NPPF 2026.

Together, these measures have expanded the range of sites on which a credible case for housing may be made. They have created new planning opportunities, but they have not removed many of the physical and financial barriers that determine whether an allocation or planning permission ultimately becomes a completed home.

In this article, I provide a detailed analysis of the planning reforms across that entire sequence, examining what they mean for planning decisions, site capacity, land value and housing delivery. I distinguish the routes applicants can use now from the measures whose full consequences will become clear only as local plans are adopted, permissions are implemented, and homes are built.

The changes arrived through different instruments with different effects. A consultation signals direction; operative policy can influence an application; an Act may create powers for later regulations; and a programme may remain a proposal until the government confirms it.

Labour took office in July 2024, succeeding the Conservative government, with an ambition to deliver 1.5 million homes in England during the Parliament. 

In her 8 July 2024 statement, Chancellor Rachel Reeves set out the first planning measures and promised consultation on restoring mandatory housing targets. That consultation opened on 30 July. It established the Government’s direction, but did not itself change the rules for deciding applications.

The Government’s first major policy changes took effect through the NPPF 2024, published on 12 December 2024. Mandatory housing targets returned under a new method that identified a national need of about 370,000 homes annually. The framework also established Grey Belt as a formal category and required relevant major housing developments in the Green Belt to meet the Golden Rules.

The consequences extended beyond the higher figures councils had to plan for. Some landowners gained a route to apply for housing on Grey Belt land without waiting for a local plan review to change the Green Belt boundary. That route still required evidence of unmet need, a suitable location and an assessment of the remaining Green Belt. 

For major schemes, affordable housing, infrastructure and accessible green space also had to be accounted for. Those requirements could strengthen the public benefits of development while changing its capacity, costs and land value.

The Planning and Infrastructure Act 2025, which received Royal Assent on 18 December 2025, operates at a different level. It provides statutory powers for spatial development strategies, a national scheme for delegating planning functions and the Nature Restoration Fund, among other measures.

Its practical effect depends on later regulations and guidance. The national scheme of delegation, for example, was enabled by the Act but does not take effect until 31 October 2026.

The new local plan-making system came into force on 25 March 2026. Its structured gateways and expected 30-month programme aim to accelerate plan production. The clock starts with the Gateway 1 self-assessment, so earlier evidence gathering sits outside it.

The two systems are running in parallel. Legacy plans can generally be submitted until 31 December 2026, while other authorities use the new route. Its effect on allocations will emerge as new plans reach examination and adoption.

The NPPF 2026, published on 17 August 2026, did more than replace the NPPF 2024. It recast national planning policy into a more structured, rules-based format, replacing continuous paragraph numbering with coded policies and clearly separating plan-making from national decision-making. 

It also made the annexes part of policy and introduced clearer development routes outside settlements and around well-connected stations.

More importantly for live pending planning applications, its decision-making policies became material considerations immediately upon publication. Annex A directs decision-makers to give very limited weight to materially inconsistent local policies, subject to specified exceptions. Age alone does not make a local policy inconsistent, and the development plan remains the statutory starting point.

As of October 2026, the position can be reduced to four categories:

• operative policy, including the NPPF 2026 decision-making policies that can affect an application now;

• a process already in force, including the new local plan system whose effect will appear as plans pass through the gateways;

• enacted measures still being implemented, including the delegation scheme, spatial development strategies and the Nature Restoration Fund; and

• a proposal that is not yet final policy, including the draft New Towns Programme.

That distinction is not merely procedural. For a landowner or developer, a route available to a current application has a very different value from one that depends on a future plan or regulation. Treating them alike risks converting policy expectation into land price before the route actually exists. 

The first signals are mixed. More homes are appearing in outline applications, but fewer are reaching detailed permission, and completions have barely moved. That does not yet prove success or failure, because each dataset captures a different stage of schemes often begun under earlier rules.

The latest housing supply indicators estimate that England added 437,900 net homes between 9 July 2024 and 20 September 2026. That is about 29% of the Government’s 1.5 million ambition.

For 2025-26, the provisional estimate is 196,900, around 6% below the 208,600 recorded in 2024-25. It uses Energy Performance Certificate data adjusted for demolitions and will be replaced by the annual net-additions series.

Most homes completed in 2025-26 were allocated, permitted or started before December 2024. The 2026 framework was six weeks old at the date of this article. Current completions cannot fairly be credited to, or blamed on, either framework.

The planning applications statistics for the year to June 2026 show outline applications proposing 246,000 homes, up 41%. Detailed applications proposed 274,000 homes, down 8%, while the provisional number of homes granted detailed permission fell 12% to about 212,000.

Ten outline applications of at least 2,000 homes accounted for 47,700 units, almost one fifth of the total. A few strategic submissions can therefore move the series without indicating a broad recovery.

The 87% overall grant rate covers all planning decisions, many unrelated to housing. It is not the proportion of proposed homes approved, and an outline application is not an outline permission.

My reading is that the outline increase is consistent with promoters seeking optionality under the new routes, while the weaker detailed figures describe the nearer-term pipeline to build. This is an interpretation, not proof: the series do not track the same schemes.

Building control data recorded 35,910 seasonally adjusted starts between April and June 2026, 20% higher than a year earlier. The release cautions that the increase was affected by much higher reporting from the Building Safety Regulator, which recorded 5,758 starts in the quarter. Completions, at 35,800, were only 1% higher.

Across the year to June 2026, starts rose 15% and completions were unchanged. The 2024 reforms may already influence promotion and applications; the newer 2026 framework should appear first in advice and decisions, then later in starts and completions. A fair judgement must follow the same sites.

They can evidence unmet need for particular routes, but measure different weaknesses. Five-year housing land supply asks whether enough deliverable sites exist for the next five years. The Housing Delivery Test compares delivery with the previous three years’ requirement. One looks forward; the other looks back.

Our September 2026 review identifies 153 councils with a reported supply below five years, represented through 156 authority and plan-area entries. The 2025 Housing Delivery Test records 93 authorities or plan-area entries below 75%. Sixty-one entries appear in both datasets, while 32 of the entries below 75% fall outside our supply-shortfall list.

These are different planning conditions, not competing lists. The 61 entries in both datasets show weak recent delivery and a short forward supply. The 32 below 75% but outside the supply list show where historic under-delivery may evidence unmet need without a recorded supply shortfall.

The reverse is equally revealing. Sixty-three entries in our supply review scored at least 95% in the 2025 test. They kept pace with the past requirement yet report an inadequate future pipeline, often against a newer or higher figure. Historic delivery does not prove future supply.

A joint plan area is an entry, not another council. Absence from our supply list also does not prove five years of supply. For a particular site, the latest published position, base date, buffer and plan geography still need to be checked.

Annex D of the NPPF 2026 sets out a stepped response. Delivery below 95% requires an action plan. Below 85%, a 20% buffer is added to the supply requirement. Below 75%, evidenced unmet housing need is deemed to exist for the outside-settlement housing route in Policy S5(1)(j).

Within the Green Belt, S5 is not the decision route. Policy GB7(1)(g) asks for evidenced unmet need for the development proposed, alongside the Grey Belt, strategic Green Belt, sustainable-location and Golden Rules tests. A five-year-supply shortfall or a Housing Delivery Test result below 75% supplies the housing evidence for that limb.

The evidence also has different durability. Supply may change with a new statement or plan. A Housing Delivery Test result normally remains until the next measurement, although it may be recalculated after a new housing requirement. The route can change while the land does not.

For more than a decade, our work on Green Belt projects has led us to argue that not all Green Belt land is green, environmentally valuable or equally important to the designation’s purposes. Some land maintains a critical gap or a clear boundary against sprawl. Other areas contain hardstanding, commercial yards, infrastructure, golf courses or fragmented land that performs a weaker role.

That weaker role does not justify automatic development, but it does justify honest assessment. We developed this argument in Green Light to Green Belt Developments, which I co-authored with Ufuk Bahar, managing director of Urbanist Architecture. The book examined the gap between stated purposes and sites such as distribution yards or inaccessible scrubland.

Before December 2024, national policy contained no formal Grey Belt classification or distinct route for weaker-performing Green Belt land. Applicants relied on existing exceptions, very special circumstances or plan-led release. What had required a more exceptional case gained a defined policy route in 2024.

The NPPF 2026 has made the assessment more structured through Annex E. Assessment areas must be small enough to reveal meaningful differences in their contribution to Green Belt purposes (a), (b) and (d), while the effect on the remaining Green Belt across the plan area must also be considered. Appearance, dereliction and previous development alone do not decide the answer.

Grey Belt is not a licence to build anywhere. It directs growth towards less harmful land and recognises where the designation performs a weaker role. That can strengthen Green Belt policy when classification is followed by serious assessment of location, infrastructure, design and delivery.

This is where the difference between planning opportunity and development reality becomes clear. Grey Belt classification can apply to land that remains a poor housing prospect because it tests Green Belt function, not whether a workable scheme can be delivered. The first question is which parts qualify; the more important one is what can realistically be built on them.

Consider an illustrative 2.5-hectare site beside a town: a commercial yard and an adjoining field, a station apparently 700 metres away, mature trees along one edge and one possible vehicle access through a narrow lane. A promoter may see a ready-made Grey Belt opportunity.

Treating yard and field as one area could conceal the field’s role in checking sprawl or preventing settlements from merging. A boundary drawn tightly around the yard solely to obtain a favourable answer is equally weak. Annex E requires a defensible scale, while GB7 also tests the remaining Green Belt across the plan area.

Suppose the yard and a contained part of the field pass that assessment, and the relevant Housing Delivery Test result is 68%. The unmet-need element may be satisfied. The scheme still needs a sustainable location, with particular reference to Policy TR3, and major housing must comply with the Golden Rules.

The national Connectivity Tool can inform the accessibility evidence. It does not prove that the lane, crossing, footway or local services can support the scheme.

An early sketch may show 60 homes. Access widening takes frontage; trees, drainage and biodiversity reduce the developable area; green space takes land; and utility or junction works add cost. Once designed rather than listed, the workable capacity might be 38 homes.

The figures are illustrative, but the lesson is real. Without safe access or a funded connection, favourable classification may add little value. An unconditional purchase based on the first image can create a viability problem.

Grey Belt status, a qualifying station or unmet need can increase competition before capacity is known. Policy opportunity becomes land value only after its obligations, costs, and risks are accounted for. If price assumes the highest unit count, later design is pressured to defend a number the site may not sustain.

The Golden Rules direct value towards affordable housing, necessary infrastructure and accessible green space. They should shape land price before agreement, not appear as deductions after acquisition.

Until an applicable development-plan policy is in place, the affordable housing requirement is generally 15 percentage points above the highest existing requirement that would otherwise apply, capped at 50%. Where no requirement exists, 50% applies by default. A local requirement of 25% would ordinarily become 40%; one of 35% would become 50%.

Policy GB8 permits a viability case only in defined circumstances, including qualifying legacy costs, strategic sites and a wholly different development model. The maximum possible contribution is still expected. National viability guidance is clear that a high land price does not justify a policy shortfall.

Return to the illustrative site. At 60 homes and 25% affordable housing, a buyer might initially assume 45 open-market homes. At 38 homes and 40% affordable housing, there would be about 23. The total unit count has fallen by just over a third, but the open-market homes that largely support land value and infrastructure have fallen by roughly half.

That difference can matter more to price than the label attached to the planning route. A layout that ignores access overstates capacity; an appraisal that ignores the Golden Rules overstates value. Planning, architecture and the transaction must therefore be tested together.

A conditional option, promotion agreement or price linked to capacity and infrastructure milestones may handle uncertainty better than an unconditional purchase. The most valuable consent is the one most likely to remain viable and buildable.

The new ‘default yes’ for development near railway stations includes a separate Green Belt route under Policy GB7(1)(h). Qualifying residential or mixed-use development is treated as not inappropriate in the Green Belt without having to establish Grey Belt status or unmet need. Major development must still meet the Golden Rules in Policy GB8.

The station must lie within one of the top 80 Travel to Work Areas by economic output and normally provide four trains or trams an hour during the weekday daytime, or two an hour in one direction. The site must be well related to the station or settlement, supportable by infrastructure and compatible with comprehensive development.

Reasonable walking distance is around 800 metres, or about ten minutes where topography, route quality or barriers discourage the longer walk. A radius is insufficient: tracks, major roads, steps or an unusable path may change the answer.

Our research assessed 83 station-and-place pairs and shortlisted 20 after considering access, land, constraints and delivery. The service threshold identifies an opportunity, not a good site. 

The NPPF 2026 supports greater housing density where suitable land, existing buildings and infrastructure can be used more effectively. 

Around a well-connected station, Policy L3 normally requires at least 35 dwellings per hectare across the net developable area, rising to 45 where service frequency is at least twice the qualifying minimum. Those figures are a starting point for design, not a substitute for it. 

The homes must still work in terms of daylight, living conditions, access, servicing and landscape, and the density calculation must reflect the land the scheme can actually develop.

This changes the commercial assessment. A Grey Belt case that depends on unmet housing need may weaken if supply is restored. The station route has no equivalent unmet need test, so its policy basis may prove more durable, provided the station service, walking route, infrastructure and design support the proposal.

There is also a distinct opportunity through plan making. Under Policy GB3(4), where a development plan’s spatial strategy identifies suitable land around a well-connected station, exceptional circumstances need not be demonstrated to justify altering the Green Belt boundary. The boundary still has to be changed through the plan process; the policy does not itself grant planning permission for an application.

They should where the evidence identifies a better-connected extension and enduring boundary. Grey Belt is a classification, not a spatial strategy. The easiest land to classify is not necessarily best able to support homes and infrastructure.

A previously developed yard reached mainly by car may be straightforward to classify. Land that makes a stronger contribution to a Green Belt purpose could still form a more coherent extension around a station, with connected streets and a defensible landscape edge. If that alternative is to be considered, the proper forum is plan-making, where locations and cumulative infrastructure can be compared.

Policy GB4 prioritises previously developed Green Belt land, then other Grey Belt, then other Green Belt locations, provided the pattern is sustainable, and the remaining Green Belt is not undermined. Plans must still compare transport, landscape, ecology, infrastructure and assembly.

This strategic choice is separate from the tests for an individual application. A promoter cannot use a theoretically better alternative to redraw a boundary or rely on a plan-making provision as though it were consent. The plan chooses between locations; the application tests one site and one scheme.

Several reforms act after policy support is established. Faster plans, strategic planning and delegation may shorten the process; the Accelerator, new towns and Nature Restoration Fund intervene more directly. All still depend on infrastructure, land control, technical approval and finance.

A 30-month plan process should reduce the wait for an up-to-date plan. It cannot secure a grid connection, enlarge a sewer, assemble ownerships or fund a school. An allocation dependent on an unnamed third party may remain years from construction.

Spatial development strategies can coordinate growth above local plans. Policy PM1 identifies broad growth locations and infrastructure, but cannot allocate individual sites. Local plans must convert that strategy into allocations and a credible sequence.

The delegation scheme may reduce inconsistent call-ins from 31 October 2026. Yet 97% of district-level decisions in April to June 2026 were already delegated. Its greatest effect may be on the smaller group exposed to local call-in rules.

Delegation can shorten the decision stage. It does not complete a section 106 agreement, discharge a condition or obtain a technical approval.

The New Homes Accelerator uses brokerage and technical support on stalled sites. Government reported that it had helped progress more than 48,500 homes and addressed wider barriers affecting another 76,500. Those are homes affected, not completions caused by the programme.

The draft New Towns Programme proposed seven locations in March 2026, but final locations were unconfirmed in October 2026. New towns may coordinate long-term supply and infrastructure, although delivery extends well beyond one Parliament.

The Nature Restoration Fund would replace some project-by-project mitigation with Environmental Delivery Plans and developer payments. The first draft plan, covering nutrient pollution in Norfolk, entered consultation on 17 September. Existing duties remain until an approved plan covers the impact and area.

Access rights, utilities, drainage, highway agreements and land assembly sit outside planning permission. For new higher-risk buildings, the Building Safety Regulator reported a median Gateway 2 approval time of 22 weeks between June and August 2026.

Finance adds another filter. A viable appraisal requires a realistic sales rate, cost allowance, funding structure and risk margin. Government coordination may unblock a stalled scheme, but it cannot make an unviable or physically impossible scheme buildable.

A planning permission with an unfunded junction, unobtainable access or unaffordable utility connection may have no credible start date. A refusal on a better located site may identify a correctable design or evidence problem. This is a judgement about particular schemes, not a preference for refusal.

The commercial question is whether realistic capacity can reach construction sooner and with less risk than a larger consent. Lenders, purchasers and contractors value an implementable permission, not a theoretical maximum.

Outline permission may establish the principle of development without resolving reserved matters, conditions, section 106, highways or utilities. Its value depends on whether those steps have owners, costs and realistic programmes.

At a station site, a new public route can shorten the real walk, organise a legible street and support the required density. The drawings and transport evidence must also establish its width, gradient, crossings, land control, funding and opening phase. A line across land outside the applicant’s control is not an access strategy.

At a Green Belt edge, a landscape corridor may retain trees, manage drainage, support habitats and create an enduring boundary. Those functions must work in the engineering and management evidence. A route without rights or a swale without maintenance will not deliver the promised benefit.

The true test is not how many sites enter the planning system, but how many emerge from it as completed homes. Success should therefore be judged by following the same sites from allocation or outline stage through to construction and completion. 

As our analysis of why planning permissions alone will not solve the housing crisis explains, applications indicate intent, permissions establish an acceptable proposition, starts show construction, and completions provide homes.

A more useful national dataset would track each cohort through:

• allocation or outline application and decision, including the capacity first proposed;

• detailed consent or reserved matters, showing how much capacity survives design development;

• section 106 obligations, infrastructure agreements and pre-commencement conditions;

• starts on site and the time from implementable permission to construction; and

• completions measured against the number, tenure and phasing originally approved.

That would reveal where attrition occurs. If outline schemes lose homes at reserved matters because access or design parameters were unrealistic, the problem began early. If complete schemes wait for utilities or funding, the blockage lies elsewhere. Slow completions may point to phasing or market absorption.

The same evidence would evaluate reforms more clearly. The Accelerator should be judged by time saved and homes started, faster plans by funded allocations reaching applications, and Grey Belt and station policies by the places they deliver.

Before buying a development site or promoting land, separate the route available today from one dependent on a future plan. Before fixing price, test a credible range of homes after access, drainage, landscape, infrastructure and affordable housing.

Before treating permission as realised value, give every reserved matter, condition, obligation, right, approval and utility connection an owner, cost and date. That exercise may show that a different access, phase boundary or housing mix is worth more than a few additional units. It also produces the delivery information that national monitoring currently lacks.

The Government has changed the planning choices available in England. Housing need now carries greater weight, Grey Belt and station-led development provide clearer routes to approval, and strategic planning is returning. As a result, suitable sites that might previously have been dismissed too early can now be assessed on their evidence and merits.

The deeper change concerns how land is evaluated: whether it performs an essential planning function, which locations can support sustainable growth, and what public benefits development should provide in return. This is closer to the evidence-led approach we argued for before Grey Belt entered national policy.

There is good reason for cautious optimism. The NPPF 2026 provides a clearer and more positive framework for addressing housing need, directing growth towards sustainable locations and making better use of land. If applied consistently and supported by sufficient resources, infrastructure and investment, it may unlock sites that the previous system struggled to bring forward.

A wider route to permission is not, however, the same as a housing recovery. England still faces an acute shortage, and the provisional estimate of 196,900 net additional homes in 2025–26 amounts to little more than half the approximately 370,000 homes a year generated by the standard method.

Most of those completions came from sites planned or permitted under earlier rules, so the figure cannot fairly be treated as a verdict on the reforms. It does, however, reveal the scale of the delivery gap that the new system must eventually close.

Reform has not yet made routine decision-making consistently fast either. Major applications may still extend well beyond their statutory determination periods while section 106 terms are negotiated. Once permission is granted, development can then be delayed by reserved matters, conditions, highways requirements, utility connections or building control approval. 

Greater delegation and faster plan-making may remove some delays, but they cannot by themselves provide missing officers, funded infrastructure, land control or development finance.

There is also a risk that the promise of reform becomes reflected in land values before a site’s true capacity is understood. 

A site may now appear more plausible for housing, but affordable housing, infrastructure, landscape, access and design requirements can all reduce its developable area and residual value. A smaller scheme that is viable and capable of being built may ultimately be worth more than a larger permission that cannot be implemented.

For any particular site, success still depends on bringing together the right location, a realistic capacity, funded infrastructure, an acceptable design and an implementable permission. Planning reform has widened the entrance to that sequence, and the NPPF 2026 gives more suitable sites a genuine opportunity to progress. Its lasting success, however, will be measured not only by the permissions it enables, but by how many of those permissions become homes, and how quickly.

New policy can strengthen a site’s planning case, but it may not resolve the reason an earlier scheme stalled or was refused. At Urbanist Architecture, we examine both: what the current framework now supports and which design, access, infrastructure, or evidence problems still need an answer.

As a multidisciplinary team of architects and town planners, we test alternative layouts, scales and housing mixes against those constraints. Where the first proposal assumed too much capacity, a revised scheme may offer a stronger route to permission and construction. Where a policy change creates genuine additional potential, we can establish what evidence is needed to make use of it.

This helps you decide whether to revise an application, promote the land through a plan or redirect investment elsewhere.