English Housing Survey: local authority housing stock condition modelling, 2023 - main report
Published 28 May 2025
Applies to England
Introduction
This Official Statistics in Development series provides sub-regional estimates of housing stock condition. Specifically, modelled estimates of the number and proportion of occupied homes that are deemed non-decent according to the Decent Homes Standard, have Category 1 hazards, or have a problem with damp in each local authority, by tenure and dwelling type.
The models used to produce these estimates were developed using the 2022-23 English Housing Survey (EHS) housing stock data. The methodology uses EHS findings on local household and dwelling characteristics, known to be associated with housing quality at a national level, to predict housing quality for each local authority in England. More detail on the modelling methodology is provided later in this release.
EHS data used in this publication comes from full inspections of dwellings between April 2022 and March 2024. The previous publication in this series represented data from 2020, where EHS housing quality data was partially modelled due to limitations in physical observations of dwellings during the COVID-19 pandemic. The pandemic also affected data collection in 2021 and 2022. Due to these limitations, we have opted to skip intervening years and use observed EHS data from the 2022 and 2023 datasets to produce a more robust analysis.
The data used in this report was produced by Tad Nowak and Tom Gregory at BRE in collaboration with NatCen Social Research and MHCLG. The responsible analyst and author of this report is Alistair Rice, Data, Analysis, Statistics and Surveys Division, MHCLG.
These statistics have been developed in response to a growing interest in housing quality. We will continue to develop these metrics to support the needs of users and welcome any feedback on this release to ehs@communities.gov.uk.
Advice for users
Totals and percentages reported here are household weighted, i.e. they refer to the number of households who live in poor quality housing, and not the total number of occupied dwellings in poor quality. In this way, occupied dwellings with multiple households resident will be counted multiple times. We refer to these as ‘non-decent homes’, ‘homes with Category 1 hazards’ or ‘homes with a problem with damp’.
We expect differences in the modelled data in this publication and the self-reported data from social sector landlords in related publications. This is because the data are collected for different reasons, and are subject to uncertainty, often leading to differences in reported levels of poor quality housing. Users should consider differences in methodological approaches when making comparisons, e.g., the modelled data here only refers to occupied dwellings and is household weighted.
We do not recommend that users directly compare estimates for individual local authorities from this release with published data from previous years due to data and methodological considerations, which prevent the use of statistical testing to determine whether changes are significant. For more detail, see How the data should be used and Modelling methodology.
Further details on the various published statistics covering housing quality are outlined in a statistical coherence article, published alongside this note.
Main findings
- Across both rented tenures, homes in local authorities in the South West, the Midlands and Yorkshire and the Humber were generally more likely to be non-decent than homes in local authorities in the South East, London and the East of England.
- London, and areas in the Midlands and the South which are geographically closest to London, had the lowest rates of homes with Category 1 hazards in 2023. Additionally, there was a low prevalence of homes with Category 1 hazards in the North East.
- Local authorities in the South West, East of England, and the North East had the lowest levels of damp in homes, in contrast to local authorities across the South East, the Midlands, Yorkshire and the Humber and the North West. This pattern appears to be driven by a higher prevalence of damp within terraced homes in those areas.
- In the private rented sector, damp was most prevalent in Yorkshire and the Humber and the West Midlands, as well as throughout the East Midlands and parts of the South East. In contrast, the highest levels of damp in the social rented sector were present in London boroughs.
- While levels of non-decency and the prevalence of Category 1 hazards fell in many regions from 2019 to 2023, most regions saw an increase in the proportion of damp homes across tenures over that time.
1. Variation in housing quality across England
Introduction
Approximately 15% of properties in England were non-decent in 2023, 8% had a HHSRS Category 1 hazard, and 5% had a problem with damp [footnote 1].
At the local authority level, the modelled data shows variation in housing conditions across England, with 89% of local authorities having fewer than 20% of homes predicted to be non-decent. Similarly, in 81% of local authority areas, less than 10% of homes are estimated to have a HHSRS Category 1 hazard. In almost all (98%) local authorities, fewer than 10% of homes are estimated to have damp.
Table 1. Number of local authorities with levels of non-decency, Category 1 Hazards and damp
Thresholds | Non-Decent | HHSRS Category 1 Hazard | Damp |
---|---|---|---|
Less than 5% | 0 | 55 | 139 |
5 to 9.9% | 45 | 185 | 151 |
10 to 14.9% | 132 | 51 | 6 |
15 to 19.9% | 86 | 4 | 0 |
20 to 24.9% | 27 | 0 | 0 |
25% or more | 6 | 1 | 0 |
Decent Homes Standard
For a dwelling to be considered ‘decent’ under the Decent Homes Standard, it must:
- meet the statutory minimum standard for housing (the Housing Health and Safety Rating System, since April 2006) - homes which contain a Category 1 hazard under the HHSRS are considered non-decent
- be in a reasonable state of repair
- have reasonably modern facilities and services
- provide a reasonable degree of thermal comfort
In 2023, 15% of all dwellings in England were non-decent [footnote 2]. Non-decent homes were spread across England. Local authority areas containing a consistently high proportion of non-decent homes were more likely to be located in the South West peninsula, as well in the East Midlands, the North West and Yorkshire and the Humber.
The West Midlands, and coastal local authorities in the South and East of England also contained some local authority areas with predicted high levels of non-decency, though levels were not as consistent.
Figure 1.1. Proportion of Non-Decent Homes by Local Authority in England, 2023
The proportion of non-decent dwellings varies by tenure. In 2023, 21% of private rented sector dwellings were non-decent. This was higher than in the owner-occupied sector (14%) and the social rented sector (10%).
Across both rented tenures, local authorities in the South West, the Midlands and Yorkshire and the Humber were more likely to have high levels of non-decency compared to local authorities in the South East, London and the East of England.
Figure 1.2. Proportion of Non-Decent Rented Homes by Local Authority, 2023
In 2023, across all regions except London, private rented sector dwellings were more likely to be non-decent than dwellings in the social rented sector.
In London, predicted levels of non-decent homes were similar for the private and social rented sectors (Figure 1.3). This was not the case across the rest of England, where the modelled rate of non-decency in the social rented sector was always less than the modelled rate of non-decency in the private rented sector (below the 1:1 dotted line).
Figure 1.3. Relationship between rates of non-decency in the social rented and private rented sector by local authority, 2023[footnote 3]
Across England in 2023, a quarter of dwellings (25%) were semi-detached. Semi-detached homes were more likely to be non-decent if they were in the North of England. This differs from the distribution of all homes observed in Figure 1.1, where higher levels of non-decency overall were also present in parts of the South.
Non-decency in bungalows, which made up 8% of dwellings in England in 2023, was more concentrated in the Midlands, the South and London.
Figure 1.4. Regional distribution of non-decent semi-detached homes and bungalows, 2023
HHSRS Category 1 hazards
The HHSRS is a risk-based assessment that identifies hazards in dwellings and evaluates their potential effects on the health and safety of occupants and their visitors, particularly vulnerable people. The most serious hazards are called Category 1 hazards and, where these exist in a home, it fails to meet the statutory minimum standard for housing in England. A dwelling will fail the Decent Homes Standard if it has one or more HHSRS Category 1 hazard.
In 2023, 8% of all dwellings in England had a Category 1 hazard.
London, and those areas of the Midlands and the South which are geographically closest to London, had the lowest rates of homes with Category 1 hazards in 2023. Additionally, there was a low prevalence of homes with Category 1 hazards in the North East. Other regions across the country showed higher rates of homes with Category 1 hazards.
Figure 1.5. Proportion of Dwellings with Category 1 Hazards by Local Authority in England, 2023
In 2023, 4% of social rented dwellings had HHSRS Category 1 hazards, compared to 10% of private rented dwellings and 8% of owner occupied dwellings.
In the private rented sector, the presence of one or more Category 1 hazards in homes was highest in local authority areas in the North West and Yorkshire and the Humber, as well as those in the South West and on the Welsh border. This mirrored the regional trend for non-decency in the private rented sector.
Within the social rented sector, only a few local authorities in Yorkshire and the Humber had Category 1 hazards in more than 8% of their homes.
Figure 1.6. Proportion of Rented Homes with Category 1 Hazards by Local Authority, 2023
Across England in 2023, 21% of dwellings were flats. Converted flats in particular had the highest rates of Category 1 hazards (18% in 2022 [footnote 4]). Flats (of all types) with Category 1 hazards were more prevalent in the Yorkshire and the Humber, and in local authorities on the Welsh border.
In contrast, detached homes, which made up 18% of all dwellings in 2023, were more likely to have Category 1 hazards if they were in the East of England.
Figure 1.7. Regional distribution of flats and detached homes with Category 1 hazards, 2023
Damp
In the English Housing Survey, a home is considered to have damp, or a problem with damp, if the surveyor records damp which is significant enough to be taken into consideration when making a HHSRS assessment. Therefore, minor issues of damp are not recorded.
In 2023, 5% of all dwellings in England had a problem with damp. Damp is known to be associated with age of dwelling, e.g. in 2022, 10% of dwellings built pre-1919 had a problem with damp compared to only 1-4% of those built in later periods [footnote 5]. In addition, terraced houses (29% of all dwellings) and converted flats (4% of all dwellings) had higher rates of damp (6% and 9% respectively) compared to all other dwelling types (2-4%).
Local authorities in the South West, East of England, and the North East had the lowest levels of damp, in contrast to local authorities across the rest of the country. This pattern appears to be driven by a higher prevalence of damp in terraced homes in those areas.
Figure 1.8. Proportion of All Homes and Terraced Homes with Damp by Local Authority in England, 2023
In 2023, 9% of private rented sector dwellings had a problem with damp, higher than 7% of social rented sector dwellings and 4% of owner occupied dwellings.
In the private rented sector, damp was most prevalent in homes in Yorkshire and the Humber and the West Midlands, as well as throughout the East Midlands and parts of the South East. In the social rented sector, the highest levels of damp were present in homes in London boroughs.
Figure 1.9. Proportion of Rented Homes with Damp by Local Authority, 2023
Historical regional trends in housing quality
The methodology used to produce this data is unable to take advantage of statistical testing to identify whether apparent differences within local authorities between years are representative of real changes.
The English Housing Survey, from which this modelled data is partially derived, can statistically test whether the prevalence of housing quality issues has changed over time within the nine Government Office regions.
Changes to the prevalence of poor quality housing over the last four years (2019 to 2023) show regional variation.
- Levels of non-decency have generally remained similar across most regions between 2019 and 2023. However, there were decreases in non-decency in the social rented sector in the South West and in the private rented sector in the East of England, as well as in London when considering all tenures together.
- Decreases in the prevalence of Category 1 hazards across the West Midlands, North East and North West were driven by improvements in the owner occupied sector. While there has been a significant drop in the proportion of private rented sector homes with Category 1 hazards since 2019, the sample size of the English Housing Survey does not have the power to detect whether these differences have occurred in specific regions.
- The prevalence of damp increased across both regions and tenures. In the South East and West Midlands, for example, this effect appears to be driven by increases in levels of damp for both the owner occupied and private rented sectors.
Figure 1.10. Statistically significant changes in the proportion of households who live in poor quality housing, 2019 to 2023
2. Regional profiles of housing quality
Introduction
This chapter provides a profile of local authority areas within each government office region in England in relation to housing quality. It includes maps for each of the nine government office regions and compares how the prevalence of non-decency, Category 1 hazards and damp differ between tenures. By focussing on each region in turn, we can observe more detailed sub-regional patterns.
This modelled data is subject to uncertainty, and trends within regions are described below in general terms. To understand further detail about individual local authorities in their regional context, users are recommended to consider this data alongside self-reported data published in the Statistical Data Return and Local Authority Housing Statistics.
North West
In the north of the North West, rural local authorities in Cumbria, as well as some in Lancashire, showed the highest likelihood of having homes that were non-decent or had Category 1 hazards, compared to a lower likelihood in the more urban areas of Cheshire, Greater Manchester and Merseyside in the south of the region. Across the entire region, Category 1 hazards were very low in prevalence in the social rented sector.
The regional pattern of the prevalence of damp across the North West varied by tenure, being more concentrated towards the south of the region in the private rented sector, but more evenly distributed across the region in the social rented and owner occupied sectors.
Figure 2.1. Regional distribution of housing quality by tenure in the North West, 2023
North East
Urban local authorities in Tyne and Wear and the Tees Valley had a lower proportion of homes that were non-decent, compared to the larger rural areas of Northumberland and County Durham. This difference was most present in the private rented sector.
There was a low prevalence of Category 1 hazards and damp across the North East, with no clear sub-regional pattern emerging.
Figure 2.2. Regional distribution of housing quality by tenure in the North East, 2023
South East
Areas on the South coast of the South East region, particularly in Kent in the east of the region, had a higher likelihood of having homes which were non-decent or had Category 1 hazards.
Damp was most prevalent in the private rented sector, particularly in Oxfordshire, Buckinghamshire, Berkshire, Hampshire in the west of the region, as well as in Surrey. There was no distinctive subregional pattern in the prevalence of damp in either the social rented or owner occupied sector.
Figure 2.3. Regional distribution of housing quality by tenure in the South East, 2023
South West
Areas with higher levels of non-decency or Category 1 hazards across all tenures were more likely to be on the western end of the South West Peninsula in rural local authorities in Cornwall and Devon. Gloucestershire in the north of the region had a comparatively higher likelihood of private rented sector homes being non-decent or having Category 1 hazards.
In contrast to non-decency, there was no clear geographical trend in the presence of damp in homes in the South West.
Figure 2.4. Regional distribution of housing quality by tenure in the South West, 2023
East of England
Local authorities in Norfolk and Suffolk on the east coast of the region showed a comparatively higher likelihood of being non-decent or having Category 1 hazards. Social rented sector homes were unlikely to be in poor quality in all parts of the East of England.
In contrast, private rented sector homes were more likely to be damp in Cambridgeshire in the west compared to the rest of the region.
Figure 2.5. Regional distribution of housing quality by tenure in the East of England, 2023
Yorkshire and the Humber
Homes in Yorkshire and the Humber had a high likelihood of being non-decent across all tenures with this likelihood slightly lower in local authorities in South Yorkshire. This was also true for the presence of Category 1 hazards and damp in the private rented sector.
Figure 2.6. Regional distribution of housing quality by tenure in Yorkshire and the Humber, 2023
West Midlands
Across all tenures, local authorities with a higher likelihood of homes being non-decent or having Category 1 hazards in the West Midlands were generally found on the Welsh border.
Damp homes in the private rented sector were present in high proportions across almost all local authorities across the West Midlands, though there was no clear subregional trend in either the social rented or owner occupied sector.
Figure 2.7. Regional distribution of housing quality by tenure in the West Midlands, 2023
East Midlands
Local authorities with high levels of non-decency in all tenures were present across the East Midlands.
In both the private rented and owner occupied sectors, local authorities in the south of the region showed a lower prevalence of Category 1 hazards compared to local authorities in the north, whereas all local authorities had a low rate of Category 1 hazards in the social rented sector.
Damp in the social rented sector was lower in the south of the region than the north. In contrast, levels of damp in the private rented sector were comparatively higher in the south, as well as in parts of Lincolnshire on the East coast. The owner occupied sector showed low levels of damp.
Figure 2.8. Regional distribution of housing quality by tenure in the East Midlands, 2023
London
All London boroughs, apart from a few central boroughs, had similarly low levels of non-decency and Category 1 hazards, a pattern which was independent of tenure. However, levels of damp were comparatively higher in both rented sectors in outer London boroughs, particularly in social rented homes in the south of London.
Figure 2.9. Regional distribution of housing quality by tenure in London, 2023
Technical notes
Summary
Official Statistics in Development are defined in the Code of Practice for Statistics as “newly developed or innovative official statistics undergoing evaluation” and were previously called ‘experimental statistics’. They are published to involve users and stakeholders in their development and to build in quality at an early stage.
These statistics are modelled from the English Housing Survey, which provides robust regional and national data (based on physical inspections of properties, undertaken by qualified surveyors). Application of the model requires specific local area data on a variety of demographic and socio-economic factors. Ideally, EHS data would be used directly to model at the local level. This is not possible, however, because of the relatively small sample used for the survey, which does not give sufficient coverage for each of the 296 Local Authorities, and certainly not for the individual Census Output Areas (COAs) within them. Instead, EHS data has been used to create a model which can then be applied to a national (dwelling level) dataset.
It should be noted, however, that small areas which are atypical in condition, may not be identified by the model. It is therefore essential, wherever possible, to compare the modelled results to local data.
See details on Official Statistics sources that publish statistics on housing quality metrics.
The Department will continue to consider how these statistics should be developed and updated going forward, including in light of feedback received on this statistical release. Our assumption at this stage is that they will be published annually.
We welcome feedback on the data, especially where the modelled estimates do not align with other local authority level data and/or intelligence on housing quality. Please send any comments to EHS@communities.gov.uk.
Definitions
For a dwelling to be considered “decent” under the Decent Homes Standard, it must:
- Meet the statutory minimum standard for housing (the Housing Health and Safety Rating System (HHSRS). Homes which contain a Category 1 hazard under the HHSRS are considered non-decent
- Provide a reasonable degree of thermal comfort
- Be in a reasonable state of repair
- Have reasonably modern facilities and services
The Housing Health and Safety Rating System (HHSRS) is a risk-based assessment that identifies hazards in dwellings and evaluates their potential effects on the health and safety of occupants and their visitors, particularly vulnerable people. The most serious hazards are called Category 1 hazards and where these exist in a home, it fails to meet the statutory minimum standard for housing in England.
A home is considered to have damp, or a problem with damp, if damp is significant enough to be taken into consideration when making a HHSRS assessment. Therefore, minor issues of damp are not considered. For further detail, see English Housing Survey 2023 to 2024: Headline findings on housing quality and energy efficiency.
How the data should be used
How the data should be used
- To show local authorities where housing quality might be poorer/worse than average.
- To examine how housing quality might vary by tenure and dwelling type.
- To indicate the likely scale of the “problem”, i.e., by estimating the number of non-decent homes in each local authority.
- To use alongside other local level data and/or intelligence to develop an understanding of housing condition.
How the data should not be used
- To identify vacant homes in poor condition (only occupied homes are included).
- As a replacement for survey data at the local authority level. The Local Authority Housing Statistics data collection contains self-reported stock condition data regarding local authority stock. The Private Registered Provider Social Housing Stock and Rents in England publication contains self-reported stock condition data regarding private registered provider stock.
- To estimate all dwelling stock by tenure at a local authority level.
- To estimate the number of dwellings where HHSRS Category 1 Hazards, non-decent conditions or damp have emerged or been remedied.
- To compare effectiveness of local authority enforcement performance, including comparisons between years.
Modelling methodology
Data available
For each local authority, estimates are available for:
- Total number of non-decent dwellings, the number of dwellings with Category 1 hazards, and the number of dwellings with a problem with damp
- Non-Decent dwellings, dwellings with a Category 1 hazard, and dwellings with a problem with damp by tenure
- Non-Decent dwellings, dwellings with a Category 1 hazard, and dwellings with a problem with damp by dwelling type
The key datasets contributing to the modelling process were:
- Combined year 2022 and 2023 EHS dwelling and household data
- 2023 Experian dwelling-level data
-
Lodged EPC data to March 2024
-
2021 Census demographic data
- April 2023 Local Authority / Unitary Authority geography
Decent Homes and Category 1 Hazards (HHSRS)
The 2023 sub-regional Decent Homes model and HHSRS model are based on national data drawn from the EHS combined 2022 and 2023 dataset. Application of the models requires specific local area data on a variety of demographic and socio-economic factors, which are derived from the 2021 Census small area datasets and 2023 Experian dwelling-level data.
A large number of independent variables are individually tested against the binary Decent Homes indicator to determine which have the highest coefficient of determination, i.e. which are most sensitive in predicting the Decent Homes or Category 1 hazard outcome. The variables are then grouped according to this coefficient and an iterative selection is made by running logistic regression on a developing group of these variables, with those least likely to contribute to final model predictions being dropped at each stage. The final set of variables contributing to the model is dependent on the results of this regression testing.
The 2023 Decent Homes model is based on the following variables, derived from the public and commercial sources listed above:
- Dwelling age – pre-1919, 1919 to 1944, 1945 to 1980, post-1980 (Experian)
- Predominant age of dwelling per postcode area – 8 bands (Experian)
- Mosaic household classifications by postcode – 15 socio-economic groups as detailed in the Mosaic guide (Experian)
- Proportion of centrally-heated dwellings per COA (2021 Census)
- Tenure – owner occupied, private rented, social rented (Experian)
- Government Office Region (2023 EHS and Experian)
- Dwelling type – terraced, semi-detached, detached, bungalow, flat (2023 BRE dwelling model)
-
Proportion of households containing a full-time employee per COA (2021 Census)
- Household affluence measure per postcode area – 20 bands (Experian)
- Household composition – 6 bands (Experian)
In addition to the regression modelling, the national set of lodged EPC data covering around 65% of the dwelling-level dataset is used to model the thermal comfort criterion as closely as possible.
The 2023 HHSRS final predictive model is based on the following variables, derived from the public and commercial sources listed above:
- Dwelling age – pre-1919, 1919 to 1944, 1945 to 1980, post-1980 (Experian)
- Predominant age of dwelling per postcode area – 8 bands (Experian)
- Mosaic household classifications by postcode – 15 socio-economic groups as detailed in the Mosaic guide (Experian)
- Tenure – owner occupied, private rented, social rented (Experian)
- Proportion of non-centrally-heated dwellings per COA (2021 Census)
- Government Office Region (2023 EHS and Experian)
- Proportion of households renting from Housing Association/RSLs per COA (2021 Census)
- Proportion of households containing a full-time employee per COA (2021 Census)
In addition to the regression modelling, the national set of lodged EPC data covering around 65% of the dwelling-level dataset is used to model the excess cold hazard as closely as possible, identifying additional cases that may have a Category 1 hazard.
The coefficients for each category of the final models’ independent variables are run to give the predicted likelihood for each dwelling in each area to be non-decent or have a Category 1 hazard. A further coefficient is included where the EPC data predicts that a dwelling falls below the thermal comfort measure.
These likelihoods are then combined to produce the total number and proportion of non-decent dwellings and dwellings with a Category 1 hazard at COA level. This is run separately for each region using a multiplying factor calculated from the modelled and published regional Decent Homes proportions applied to each COA non-decent total, thus ensuring that the regional proportions from the model are consistent with the EHS analysis (household weighted).
Due to geographical differences between the 2021 Census and other national data used in the modelling, a small number of COAs do not have Decent Homes or HHSRS data. These gaps are filled by identifying COAs with complete data, which shared identical or very similar characteristics to the missing cases. Census data is used for this comparison as it records proportions of particular characteristics rather than discrete categories, and so the process can more accurately match pairs of COAs.
To provide consistency between the modelled dwelling and Decent Homes numbers and the EHS regional figures, an adjustment is made in which small incremental changes are made randomly at COA level, until the totals of all dwellings, non-Decent dwellings and dwellings with Category 1 hazards are consistent with regional totals (household weighted).
The adjusted figures are re-checked against the published totals before the data is aggregated to Local Authority and regional level.
Further validation is applied to ensure the proportion of non-decent homes and homes with a Category 1 hazard in each COA is within an expected range, and that regional totals are consistent at each level of aggregation. Finally, a level of consolidation is applied between the dwelling type and tenure totals stated by the model and those from grossed EHS data. Although these will follow a similar pattern, they will not precisely match since the key objective of the model is to estimate the different proportions of non-decent homes and homes with Category 1 hazards between local areas in the first instance. Any attempt to then standardise the precise numbers of homes by dwelling type and tenure as stated by EHS data (rather than sub-totals from the dwelling level dataset) would lead to a recalculation of these proportions, thereby losing the geographic pattern established by the regression model.
Damp
This is the first EHS Local Authority Data report to estimate sub-regional levels of homes with damp problems, alongside established reports focussing on Decent Homes and HHSRS hazards. A home is considered to exhibit damp problems if one or more rooms are found with either penetrating damp, rising damp, or serious condensation.
The methodology used is broadly consistent with similar sub-regional reports on Decent Homes and HHSRS and is provided at English Local Authority (LA) level. These data report on the estimated number and proportion of homes which exhibit damp problems, by LA. They also provide this data disaggregated by tenure and by dwelling type.
The methodology for the damp model includes some developments beyond previous models to help deal with the smaller level of unweighted EHS surveys contributing to damp data, than those contributing to the Decent Homes or HHSRS models. These developments are summarised below.
The sub-regional damp modelling is based on national data drawn from the English Housing Survey (EHS) to produce 2023 sub-regional statistics. These are based on the 2022 and 2023 combined years EHS derived damp indicators.
The SPSS statistical regression method used in previous sub-regional modelling was again used for the 2023 HHSRS and Decent Homes predictive element. However, an alternative method using the CatBoost machine learning algorithm was used for the damp model. This is due to the relatively small sample of EHS surveys identified as having damp problems (5.7% of the combined EHS sample), which are used to train the model.
In addition, the damp model used key indicators developed within the BRE dwelling level stock model as potential independent variables in the regression stage, alongside established Experian and Census variables.
These independent variables are run through numerous feature engineering stages, to remove highly correlated variables and eliminating variables with 50% or more missing values. In addition, any features with unique values or very low variance are also eliminated. The remaining features are checked for overall correlation.
The 2023 damp model is based on the following variables, derived from the public and commercial sources listed above:
- Dwelling age – pre-1919, 1919 to 1944, 1945 to 1980, post-1980 (Experian)
- HHSRS dwelling level indicator (BRE stock model)
- Disrepair dwelling level indicator (BRE stock model)
- Excess cold (BRE stock model and EPC surveys)
- EPC SAP rating band (BRE stock model and EPC surveys)
- Tenure – owner occupied, private rented, social rented (Experian)
- Households with children indicator (Experian)
- HHSRS Category 2 level hazard indicator (BRE stock model)
Training of the model involved two stages. The first included all the collected variables across several models, using different configurations and parameters, with the best performing configuration selected. The second stage included only the most relevant variables and re-running the model with the best performing parameters and configuration.
Following this selection process the most relevant variables are run through a CatBoost machine learning classifier which outputs propensity results, classifying how likely a dwelling is to exhibit damp problems. This is converted into a binary indicator determining whether the property fails the damp criteria.
The binary indicator is aggregated to Local Authority level and the proportions of damp homes are standardised to ensure they match reported EHS regional totals when aggregated per region. To provide consistency between the modelled dwelling numbers and the EHS regional figures, further small adjustments are made in which small incremental changes are made randomly at LA level, until the totals of all dwellings and those with damp problems are consistent with regional totals (household weighted).
Further validation is applied to ensure the proportion of dwellings with damp problems in each Local Authority is within an expected range, and that regional totals are consistent at each level of aggregation. Finally, a level of consolidation is applied between the dwelling type and tenure totals stated by the model and those from grossed EHS data. Although these will follow a similar pattern, they will not precisely match since the key objective of the model is to estimate the different proportions of damp homes between local areas in the first instance. Any attempt to then standardise the precise numbers of damp homes by dwelling type and tenure as stated by EHS data (rather than sub-totals from the dwelling level dataset) would lead to a recalculation of these proportions, thereby losing the geographic pattern established by the regression model.
Related data
English Housing Survey
The English Housing Survey is a continuous national survey commissioned by the Ministry of Housing, Communities and Local Government (MHCLG). It collects information about people’s housing circumstances and the condition and energy efficiency of housing in England. It provides robust regional and national data on housing quality, based on physical inspections of properties, undertaken by a qualified surveyor. National Statistics publications from the English Housing Survey, on housing quality can be found here:
- Data on dwelling condition and safety
- Housing quality and condition report 2022-23
- Housing and health fact sheet 2023-24
Energy Performance Certificate (EPC) data
Based on certificates held on the Energy Performance of Buildings Register, these Energy Performance of Buildings Certificates statistics present information about the energy efficiency of domestic and non-domestic buildings in England and Wales that have been constructed, sold, or let since 2008, and of larger public authority buildings recorded since 2008. For the modelled estimates of housing quality, this data was used to model the thermal comfort criteria of the Decent Homes Standard, the Excess Cold hazard in the HHSRS Category 1 hazard model, and the prevalence of damp.
There is a strong relationship between energy efficiency and housing quality, and most homes with poor energy efficiency do not meet the Decent Homes Standard.
Experian data
The modelling is supported by Experian household level data which provides an up-to-date portfolio of address, geographic and demographic data, which is used to inform BRE’s stock modelling estimates. Information on dwelling age, household composition, tenure and dwelling type have informed the likelihoods used in the Decent Homes and HHSRS Category 1 hazard models and helped to estimate the number of dwellings failing to meet each indicator.
2021 Census demographic data
During the regression analysis stage of this modelling, the 2021 Census data at COA level, was used to help identify demographic categories that were most likely to contribute to estimates of these housing quality indicators. The data were further used to provide a baseline position for the estimated number of dwellings in each local area, which were then refined through consolidation with recent data from EHS and other central government sources.
Welsh Housing Conditions Survey
The Welsh Housing Conditions Survey collects information about the condition and energy efficiency of all types of housing in Wales.
Scottish House Condition Survey
The Scottish House Condition Survey contains statistics on house condition, based on an annual survey sample of around 3,000 dwellings.
Northern Ireland Housing Statistics
The Northern Ireland Housing Statistics report contains information and statistics relating to the condition of homes in Northern Ireland.
Additional notes
Actual numbers of homes by Local Authority, tenure and dwelling type are indicative and based on the outputs from the modelling methodology described above.
As such, totals and the numbers of homes which are non-decent, have Category 1 hazards, or with damp problems, whether explicitly stated or inferred, should be viewed as an approximate measure of relative size against other Local Authorities, but should not be treated as surveyed totals.
The proportions of homes have been consistently modelled across the local areas as described above, and are indicative of their relative levels, but are not based on a surveyed total of such homes.
-
Estimates in this release are modelled using a household weighting, whereas data published in English Housing Survey reports uses a dwelling weighting. This can lead to differences in reported proportions, especially where sample sizes are small. ↩
-
English Housing Survey 2023 to 2024: headline findings on housing quality and energy efficiency. ↩
-
Scatter plot excludes the Isles of Scilly. ↩