Summary of findings
- The April–June 2026 official bulletin reported 486,000 EPCs lodged across domestic and non-domestic buildings in England and Wales; domestic lodgements were 440,000 in England and 24,000 in Wales. These are different subsets and must not be misrepresented as an all-housing census. [1]
- In England, 89% of new dwellings lodged in the quarter had A/B energy-efficiency ratings, whereas 88% of existing-dwelling lodgements were in bands C or D. These describe certificates registered in different groups, not a longitudinal retrofit outcome. [1]
- Official statistics estimate EPC coverage at roughly 60% of English housing stock and 55% of Welsh stock. Missing certificates, repeated lodgements, assessment methods and building characteristics matter when mapping geographical inequality. [1,2]
Abstract: what can EPC records tell us about unequal housing efficiency?
Context. The energy efficiency of existing housing influences estimated heating demand, affordability, building decarbonisation and exposure to future retrofit costs. Energy Performance Certificates (EPCs) offer standardised assessments across many properties but do not constitute an unbiased measurement of every home or a direct record of household energy bills.
Objective. Evaluate the interpretation of recent EPC statistics, distinguish certificate lodgements from housing stock, and establish a defensible approach to comparing the spatial distribution of energy efficiency in England and Wales. The analysis also distinguishes new-build and existing-property certificates and the difference between modelled energy performance and realised household outcomes.
Methods. Critical synthesis of the Ministry of Housing, Communities and Local Government (MHCLG) official April–June 2026 statistical release, accompanying technical notes and the national certificate-data service. Reported counts are directly attributed to the official release; the paper does not claim to have recomputed a 2026 national address-level EPC map or drawn unverified local authority rankings.
Key interpretation. New dwellings registered in the period are far more likely to achieve high modelled efficiency bands than existing properties, but this descriptive contrast is affected by different building standards, stock age and transaction selection. Because certificate coverage is incomplete, the distribution among recorded EPCs is not automatically the distribution among all occupied homes. [1–3]
What an EPC band measures—and what it does not
Domestic Energy Performance Certificates summarise modelled building energy efficiency using property attributes and standardised assessment assumptions. An Energy Efficiency Rating (EER) and Environmental Impact Rating (EIR) may be reported; the letter bands range from A (higher rated performance) through G (lower rated performance). The model is built to compare buildings under standard assumptions, not to explain the complete variation in every household’s metered consumption. [1,2]
Actual bills and indoor conditions also depend on occupant behaviour, energy tariffs, heating schedules, weather, appliance use and maintenance. Two households in otherwise similar homes may experience different costs and levels of comfort. EPC records cannot directly quantify fuel poverty, because that requires additional information about households, income and actual or required costs.
A certificate relates to a particular dwelling assessment at a particular time. A record may have been superseded following improvements or another transaction; newly registered certificates are not equivalent to newly constructed homes unless the new-dwelling subset is specifically selected.
| Measure | Meaning | Primary caution |
|---|---|---|
| Energy Efficiency Rating / band | Modelled relative energy-cost efficiency, A–G | Not a household metered-bill observation |
| Environmental Impact Rating | Modelled environmental impact under an assessment method | Distinct from household carbon footprint |
| Current vs potential rating | Present assessment and modelled recommended potential | Potential is not an implemented improvement |
| Property type / age | Context on construction and dwelling characteristics | Can confound raw geographic comparisons |
| Lodgement date | Date certificate entered the register | Does not by itself establish building completion |
| Address / UPRN / local authority | Geographic join and grouping fields where available | Coverage, disclosure, duplicate and location issues |
The 2026 certificate evidence: strong differences, incomplete coverage
MHCLG reports 486,000 certificates lodged for domestic and non-domestic properties in England and Wales during April–June 2026. In England, 440,000 domestic EPCs were registered, comprising around 51,000 for new dwellings and 389,000 for existing dwellings. Wales recorded around 24,000 domestic lodgements, with approximately 2,000 new and 22,000 existing. The national total and the two domestic-only totals therefore do not share the same population definition. [1]
For English new dwellings registered in the quarter, 89% received A or B efficiency ratings. In the existing-home lodgements, most certificates were in the middle bands: 88% of existing dwelling lodgements received C or D. These figures should not be subtracted to produce a “retrofit gap”; they refer to different populations and different sets of bands. [1]
The statistical bulletin estimates certificate coverage equivalent to roughly 60% of dwellings in England and 55% in Wales. It cautions that the register is not a complete census of housing stock and that quarterly lodgement patterns are seasonal. Comparisons between the same quarter in different years are usually more appropriate than raw comparisons of consecutive quarters. [1]
Figure 1. Published percentage measures in different EPC subpopulations
Source: MHCLG Energy Performance of Buildings Certificates, April–June 2026 [1] · Bars refer to different rating groups and property cohorts; they must not be read as competing shares of one common denominator.
| Reporting group | Lodgements | Population represented |
|---|---|---|
| England and Wales, all domestic and non-domestic | 486,000 | All recorded EPCs in the quarter |
| England, domestic | 440,000 | New and existing English dwellings with EPC lodgements |
| England, new domestic | 51,000 | New-dwelling subset of English domestic records |
| England, existing domestic | 389,000 | Existing-dwelling subset of English domestic records |
| Wales, domestic | 24,000 | New and existing Welsh dwellings with EPC lodgements |
| Wales, new domestic | 2,000 | New-dwelling subset of Welsh domestic records |
The selection and duplication problems in certificate-based mapping
EPCs are commonly produced following specified transactions or construction events; a property can be assessed more than once, and some homes lack a recent certificate altogether. The observed share of a rating band among current registrations will differ from the true stock-wide share if certification is associated with building age, tenure, refurbishment, local sales activity or the type of dwelling.
This is not merely a small technical adjustment. Where certain neighbourhoods have fewer market transactions or a different property mix, raw certificate counts may primarily map administrative activity. A “low-energy-efficiency borough” ranking that ignores stock coverage and repeated certificates may be heavily biased.
The available bulk EPC service covers England and Wales, with Scottish EPC records managed separately; combining national sources without harmonising methods, geographies and dates can create spurious UK-wide patterns. The published certificate service includes older records since 2008, some replaced or expired; spatial analysts must define how the current record is identified. [2,3]
Which denominator should a map use?
The number of EPC records is not the denominator for a proportion of all housing. To estimate stock-wide rating shares, the researcher should use a housing-stock denominator (such as the appropriate official dwelling stock estimate) and explicitly report the observed EPC coverage fraction. A certificate-only rating distribution can still be useful, provided the map title makes that narrower meaning clear.
Where an address has multiple records, a declared “latest valid certificate per dwelling” rule is preferable to treating all lodgements as independent properties. De-duplication should be checked against the data provider’s documentation and building identifiers.
Interpretative rule: never describe a certificate-based local-authority map as “the share of all homes in EPC band E–G” without evidence of appropriate dwelling-level coverage and missing-data adjustment.
A practical research protocol for mapping inefficient housing
A regional or local-authority map should begin with a fully specified estimand: the distribution of latest EPCs among certified homes, or an adjusted estimate for all homes. These are not interchangeable. The intended unit may be dwelling, household, building or address; an apartment block can contain many dwellings and must not be treated as a single household without justification.
The protocol below identifies the transformations needed for a more rigorous national analysis. It is a specification for reproducibility, not a claim that the current page has processed the entire address-level certificate database.
- Select a fixed extraction date and official EPC data release, and record jurisdiction, certificate method, version and licence restrictions.
- Normalise available UPRN and address identifiers; identify the latest appropriate certificate by dwelling, excluding or separately marking superseded records.
- Separate new and existing dwellings. Do not mix domestic EPCs with non-domestic asset ratings or Display Energy Certificates.
- Join each selected certificate to a contemporaneous authority or small-area geography; preserve records with missing or uncertain geographic identifiers.
- Report the distribution across A–G bands among certified dwellings, and give the number of unique certified dwellings behind every geography.
- Add an independent housing stock estimate if making stock-wide claims; publish coverage share and test sensitivity to different assumptions about unobserved homes.
- Stratify by property type, age, tenure proxies or construction cohort where valid, and avoid interpreting compositional differences as causal effects of place.
- Document duplicate-resolution rules, small-cell disclosure controls, model-version changes, regional sensitivity and all source URLs.
What the energy-efficiency divide means for households and policy
Differences between new and existing EPC lodgements reflect a range of factors including building age, design, thermal fabric, heating systems and regulatory standards. The reported divergence is important for understanding the challenges facing existing housing, but it is descriptive. It should not be interpreted as a causal estimate of the effectiveness of any single regulation or retrofit grant.
A local concentration of lower EPC bands may be relevant for planning retrofit investment, but it does not prove that residents are fuel poor. Building condition, actual incomes, household composition and energy costs need separate evidence. Similarly, a high EPC rating does not guarantee healthy indoor temperatures or affordable bills.
A stronger empirical programme could track dwellings with repeat assessments, compare changes within the same home and examine which improvement recommendations were actually implemented. Even that design would need caution: a changed EPC rating may reflect assessment methodology as well as physical improvements.
The regulatory and policy interpretation should also be versioned. EPC assessment frameworks and minimum energy efficiency requirements can change; the research date and the policy version should be given rather than treating a future proposal as a current statutory requirement.
Combining EPC geography with other forms of land and housing intelligence
A geospatial enquiry can use energy performance as one analytical layer alongside building form, conservation constraints, flood exposure and local housing conditions. At individual property level, the official certificate register is the primary source for reported ratings; Maplify can support broader spatial context such as property location and relevant planning designations.
For retrofit planning, the ability to work on a building can be affected by listed status, conservation policy, housing type, landlord consent and the location of heat or electricity infrastructure. These constraints are research prompts, not a claim that upgrades are prohibited in every designated area.
No map layer can determine a household’s actual energy bills or the effectiveness of a proposed retrofit without appropriate property-specific information.
| Question | Primary evidence | Caution |
|---|---|---|
| What does the certificate say? | Government EPC Register or downloadable certificate data | Check date, address match and whether superseded |
| Where are lower bands clustered? | Deduplicated domestic EPCs joined to geographies | Describe certificate coverage and uncertainty |
| Is external alteration constrained? | Listed-building and conservation-area sources | Seek current authority and professional guidance |
| What other risks affect the property? | Flood mapping and local planning evidence | Do not infer energy ratings from flood classification |
| What policy is currently applicable? | Government and local planning or housing publications | Confirm current jurisdiction-specific rules |
How to make an EPC inequality argument that can withstand scrutiny
The central limitation is missingness: EPC certification is neither universal nor uniformly updated. A well-designed geographic study should publish the extent of coverage, define its observation unit and avoid interpreting the observed certificate distribution as a random sample of all homes.
The April–June 2026 statistical figures are release-level observations, not a spatial microdata extract processed by Maplify. Accordingly this report offers verified national and country-level published comparisons and a full analytical protocol, without fabricating ward-level counts, local authority rankings or a national geospatial dataset.
The next research step is to combine deduplicated records with an independent housing denominator and appropriate dwelling characteristics, then test whether spatial contrasts persist after accounting for stock composition. Strong conclusions require more than a coloured postcode map.
A reliable report should allow readers to trace each percentage from a source table through a clearly documented denominator to a reproducible calculation—and to distinguish that calculation from the wider social and policy interpretation.
References, data sources and citation
References are numbered in citation order. Where sources refer to different nations, reference dates or units, the report states the distinction instead of treating them as interchangeable. The links lead to originating official publishers and their methodology or downloads.
- MHCLG — Energy Performance of Building Certificates, April–June 2026 official statisticswww.gov.uk
- Government service — Get energy performance of buildings data (England and Wales)get-energy-performance-data.epb-integration.digital.communities.gov.uk
- MHCLG — Energy performance of buildings certificates technical noteswww.gov.uk
- ONS — Energy Performance Building Certificates official weekly serieswww.ons.gov.uk
- GOV.UK — Find an energy certificatewww.gov.uk
Suggested citation
Maplify Research Desk (2026). The energy efficiency divide: housing-stock coverage, certification patterns and the geography of energy performance. Maplify Research, 27 October 2026. Available at: https://maplify.co.uk/research/energy-efficiency-divide (accessed on date of use).
Research and explanatory analysis; not a peer-reviewed journal article, statutory planning determination, professional site survey, legal opinion or valuation. Data and policy can change. Consult the original publisher and competent authorities before consequential decisions.