Land use statistics revamp breaks historic comparisons

Land use statistics in England face a major methodological reset that could change the way planners, councils and developers present evidence about brownfield housing, land consumption and development density. The Ministry of Housing, Communities and Local Government is consulting on a replacement system, with responses invited until 28th October 2026.

The proposals, originally published on 16th September, would restart official land use releases using Ordnance Survey’s National Geographic Database. They would provide more detail about solar installations, residential gardens and the way land is classified, but introduce an important break with figures published under the old method.

That break is the central practical issue. Historic statistics are frequently used to describe whether development is shifting towards or away from previously developed land. A change in the data and classifications could alter the apparent trend even where the underlying land has not changed in the same way.

Land use statistics: what is changing?

The MHCLG consultation covers three related statistical products: the annual snapshot of land use in England, the annual measurement of new residential addresses by previous land use, and the three-yearly measurement of how land changes from one use to another. These are evidence products, not amendments to planning legislation or a new definition of development.

The older land use series was drawn from Ordnance Survey MasterMap and AddressBase data and assigned land to 28 categories originally developed in 2004. The underlying source products are being retired. MHCLG says the replacement National Geographic Database offers richer information, but its classifications and mapped features cannot simply be fitted into the previous method.

Instead, the Department proposes to adopt Ordnance Survey’s two-level land use and land cover classifications. Land use describes an activity, such as housing, education or retail. Land cover describes a physical surface, such as trees, grassland or water. The consultation favours a combined list that uses the most informative available label for each piece of land.

That approach would also recognise a wider site, such as a hospital or school, rather than automatically giving each building, garden or separate feature its own apparent use. Where a feature forms part of nested sites, the proposal is to apply the classification of the largest containing site. This may produce more realistic figures for complex landholdings, but users are being asked whether that hierarchy is appropriate.

Brownfield measurements and the developed-land divide

One of the most consequential questions concerns which new categories should count as developed or non-developed land. The previous statistical classifications contained this attribution; the replacement Ordnance Survey land use categories do not. MHCLG proposes to assign a provisional status to each new category using the National Planning Policy Framework definition of previously developed land as a reference.

The consultation acknowledges that allocating a whole statistical category to one side of the divide involves compromises. Sports facilities, for example, may include developed stadia and relatively undeveloped playing fields. Residential gardens are another critical distinction: gardens in built-up areas are generally excluded from the NPPF definition of previously developed land, even though they form part of residential property.

The proposal also adds specific categories for residential gardens and solar installations, which are not discrete categories in the standard Ordnance Survey Tier A classification. For solar development, this could help users understand an increasingly important competing demand on land. But a statistical category for solar panels should not be mistaken for a planning conclusion about the lawful use of a particular parcel, its Green Belt status or the merits of a scheme.

A useful consultation response would test the provisional classification against real examples. A council or developer relying on a district-wide brownfield percentage should ask whether the method accurately represents mixed-use sites, former commercial yards, outdoor recreation, vacant sites and land requiring remediation. Errors or category differences at that level could affect how persuasive a headline figure appears.

Vacant sites and housing density would be counted differently

Another proposed change removes vacant land as a separate reporting category. The former series distinguished vacant previously developed land from vacant land that was not previously developed. Under the new approach, land that is vacant would instead generally be classified by its last known use. MHCLG notes that the two old vacant-land categories historically accounted for large numbers of new residential addresses.

Removing those distinct headings would redistribute addresses across other categories. It does not mean those sites have disappeared, nor that the Government is proposing to remove vacant land from planning consideration. It changes the way one statistical series describes their origin, so published comparisons would need careful explanatory notes.

MHCLG also wants a more intuitive housing-density measure. The earlier release assessed the level of development within one hectare of each new address, describing how built-up its surroundings were. The replacement would count new addresses per unit of area to better identify where new homes are being concentrated. The final spatial scale remains undecided.

For practitioners, these two measurements answer different questions. A new dwelling surrounded by established urban development can appear in a high-density setting under the old approach without necessarily being part of a dense new scheme. The proposed address-count measure instead focuses on the concentration of new development. It should not be quoted as if it were an automatic substitute for plot density, dwellings per hectare in an application, or a design assessment.

Why past land use statistics will not match the new series

The Department is explicit that a new data source, different mapping granularity and different land-use labels will cause a break with the figures collected between 2013 and 2022. In its consultation, MHCLG warns that figures from the new series should therefore not be directly compared with those published up to 2022. That is an official warning about comparability, rather than evidence that land use changed sharply at the point the methodology changed.

Officials considered producing an approximate match between old and new categories but are concerned that it might create a false impression of precision. They are seeking views on whether an indicative correspondence table would still be useful, provided that its limitations were made clear.

For Local Plan examinations, the important distinction is between a change on the ground and a change in the way that land is recorded. Any evidence base spanning the transition will need to identify the data vintage, the definition of each measure and whether a valid comparison exists. Trend graphics that join the two datasets without qualification could mislead readers.

The proposed new series would begin from an October 2024 baseline, with 2024, 2025 and 2026 land-use snapshots targeted for publication by the end of 2026. MHCLG also proposes to publish the 2024 to 2025 and 2025 to 2026 new residential address releases by then, with the first new hectarage-change release covering 2025 to 2027 expected by the end of 2027. These dates remain proposals and publication targets, not completed releases.

What should planning professionals do now?

Local planning authorities, consultants and developers who rely on land-use figures should consider responding to the consultation. The proposed information continues to include district-level breakdowns, the Green Belt and areas of flood risk. The Department is also seeking views about retaining neighbourhood-scale data and whether additional geographies, including National Landscapes or Sites of Special Scientific Interest, would be valuable.

Useful responses can be specific. Explain which historical measures are used in development plan evidence, whether users need a transparent bridge between the old and new series, and which categories matter to current land allocations. Consultees can also comment on the developed/non-developed split, the loss of a discrete vacant-land class and the proposed new density measure.

The consultation runs until 28th October 2026 through the official response portal. MHCLG is not proposing through this exercise to change permitted development rights, planning application procedures, statutory planning tests or the definition of previously developed land in a site-specific planning decision. It is consulting on the statistical evidence used to observe and debate the consequences of those rules.

The territorial scope is England only. The National Geographic Database covers Great Britain, but the statistical consultation concerns MHCLG’s England series, not equivalent Welsh, Scottish or Northern Irish planning guidance. Separate planning systems and datasets must be assessed on their own terms.