Tiling the Nation Differently: Five Reasons Britain's Statistical Geography Has Passed Its Expiry Date
Tiling the Nation Differently: Five Reasons Britain's Statistical Geography Has Passed Its Expiry Date
Every ten years, Britain counts itself. The census is a remarkable logistical achievement — a near-complete enumeration of a population of more than 67 million people, conducted simultaneously across thousands of communities and subsequently disaggregated into a hierarchy of geographic units that underpin public policy for the decade that follows.
The 2021 Census, conducted amid the extraordinary disruptions of the pandemic period, was no exception. Its data will inform school funding formulae, NHS commissioning decisions, local authority grant settlements, and transport investment priorities well into the 2030s. The geographic units through which that data is organised — output areas, lower super output areas, middle super output areas, and ward boundaries — will shape the spatial logic of British public administration for years to come.
The problem is that those units are already obsolete. Not in the sense that they contain errors — the Office for National Statistics applies considerable rigour to their construction — but in the deeper sense that they encode assumptions about how people relate to place that are increasingly difficult to defend. Below, CodexGeo examines five specific dimensions of that obsolescence, and the alternative approaches that deserve serious consideration.
1. Ward Boundaries Reflect Political History, Not Lived Geography
Electoral ward boundaries are perhaps the most deeply embedded layer of Britain's statistical geography. They are used not only to organise voting but to aggregate a vast range of administrative and analytical data, from deprivation indices to public health surveillance. Yet ward boundaries are drawn primarily to achieve electoral equality — roughly equal numbers of registered voters per ward — rather than to reflect any coherent geographic or social logic.
The result is that statistically adjacent wards frequently contain communities with radically different characteristics, artificially separated by boundaries that correspond to roads, rivers, or administrative conveniences rather than to any meaningful distinction in how people live. Conversely, communities that function as a single social and economic unit are regularly split across ward boundaries, making it impossible to analyse them coherently using standard census geographies.
This is not a marginal distortion. Deprivation funding allocated on the basis of ward-level indices can systematically underserve communities that happen to straddle boundaries, or overserve areas where the aggregation masks significant internal variation. The ward, as a statistical unit, is a political artefact masquerading as a geographic one.
2. Output Areas Cannot Capture Population Fluidity
The output area — the smallest unit in the census geographic hierarchy — is designed to contain between 100 and 625 households, grouped to maximise internal social homogeneity. This is a reasonable approach for capturing the residential population at a single point in time. It is an inadequate approach for understanding how that population actually uses space across the course of a day, a week, or a year.
Britain's urban population is increasingly fluid. Commuting patterns, hybrid working arrangements, and the fragmentation of retail and leisure activity across multiple centres mean that the daytime population of many output areas differs dramatically from the residential population captured by the census. A city-centre output area may house relatively few residents but serve as the daily destination of tens of thousands of workers and visitors. A suburban output area may appear well-resourced on residential metrics while its population disperses entirely during working hours.
Public services dimensioned on the basis of residential output area data — from GP practice capacity to library provision — are therefore routinely miscalibrated. The geography of need, as the census records it, and the geography of demand, as services actually experience it, are increasingly divergent.
3. School Catchment Calculations Are Built on Shifting Sand
Few areas of public policy are more geographically sensitive than school admissions. Catchment areas, oversubscription criteria, and the fair access protocols that govern admissions to popular schools are all deeply entangled with the statistical geographies through which population data is reported. When those geographies misrepresent residential patterns — as they increasingly do in areas experiencing rapid new development, demographic turnover, or boundary changes — the downstream effects on individual families can be severe.
Rapidly growing communities on the periphery of established urban areas present a particular challenge. New housing developments frequently straddle output area and ward boundaries, or fall within statistical units whose demographic profile was established before the development existed. School place planning based on census-derived population projections for these areas routinely underestimates demand, producing the capacity crises that have become a familiar feature of British suburban growth.
Several local authorities are now supplementing census-derived data with administrative records — GP registration data, school roll information, and council tax records — to build more dynamic pictures of local population change. These approaches are promising, but they operate outside the official statistical geography, creating a parallel data landscape that complicates cross-authority comparison and national-level analysis.
4. NHS Catchment Planning Is Anchored to the Wrong Scale
The relationship between NHS service planning and census geography is long-standing and deeply institutionalised. Clinical commissioning geographies, health needs assessments, and the allocation of primary care resources all draw extensively on output area and middle super output area data. The assumption embedded in this approach is that the populations served by health services are meaningfully represented by the residential geographies through which census data is organised.
For acute and emergency services, this assumption is particularly questionable. Hospital catchments are defined not by residential proximity but by the complex interplay of transport accessibility, referral patterns, clinical specialisation, and patient choice. A hospital serving a regional catchment of several hundred thousand people will draw patients from dozens of output areas and multiple ward boundaries, none of which corresponds to the functional geography of its patient population.
Integrated Care Boards, introduced under the Health and Care Act 2022, represent an attempt to rationalise NHS planning geographies around more coherent functional areas. But their boundaries were drawn through a process of negotiation and compromise that did not systematically engage with census geography, creating a new layer of spatial misalignment rather than resolving the underlying problem.
5. The Alternatives Are Ready — But Lack Institutional Champions
The limitations of Britain's inherited statistical geography are well understood within the geospatial and statistical communities. What is less well understood — at least outside specialist circles — is the degree to which credible alternatives are already being developed and, in some cases, deployed.
Hexagonal grid systems, of the kind popularised by platforms such as Uber's H3 framework, offer a mathematically consistent tiling of geographic space that is indifferent to administrative boundaries and scalable across multiple resolutions. Several local authorities and research institutions are exploring hexagonal grids as a basis for spatial analysis, attracted by their geometric consistency and their compatibility with modern data science workflows.
Dynamic activity zones — geographic units defined by patterns of movement and interaction rather than residential location — represent a more radical departure. Derived from mobile network data, transport modelling, and aggregated location intelligence, these units can represent how populations actually use space rather than where they happen to sleep. Pilot applications in urban transport planning and retail catchment analysis have demonstrated their analytical value, though significant questions about data governance and privacy remain to be resolved.
The Office for National Statistics has itself acknowledged the limitations of the current geographic framework and is exploring options for more flexible statistical geographies in future. But institutional momentum is slow, and the constituencies invested in the current framework — from local government to NHS planning — are numerous and entrenched.
A Foundation Worth Rebuilding
The argument for reforming Britain's statistical geography is not that the current system is useless. It is that the system's limitations are now sufficiently consequential — in terms of misallocated resources, distorted analyses, and policy decisions built on spatial fictions — that the cost of inaction is measurable and growing.
Location intelligence is only as reliable as the geographic framework through which it is interpreted. Britain has invested significantly in improving the quality of its geospatial data. It has invested far less in questioning whether the tiles through which that data is organised still reflect the country they are meant to describe.
