1. Introduction

Annual estimates of the tenure of dwellings for subnational geographies will provide evidence to help planning authorities set housing policy, and allows them to monitor the distribution of tenure over time within an area and between areas.

Published data on dwelling stock by tenure are available at the national level, which are produced by the Ministry of Housing, Communities and Local Government (MHCLG). They also produce dwelling stock estimates for local authority districts in England, however, this only includes data for the private sector overall. Therefore, in England, annual statistics on the breakdown of the private sector into owner-occupied and privately rented dwelling stock at the subnational level are not available. There is a need for regular annual statistics that provide a full dwelling stock by tenure breakdown at the subnational level.

This Research Output provides the latest information about research into new methods and data sources used to produce annual subnational estimates of the dwelling stock by tenure in England. This is an update to our previous Research Output, following the same method but with two years’ extra data.

There are three main types of tenure for which dwellings can be categorised: owner-occupied, privately rented and socially rented. As National Statistics are produced by MHCLG on the number of socially-rented dwellings for local authority districts in England, we are only producing statistics that break down tenure within the private sector.

Nôl i'r tabl cynnwys

2. Main points

  • In 2017, the percentage of dwellings in England that were owner-occupied ranged from 24.5% in Tower Hamlets in London to 84.5% in Ribble Valley in the North West.

  • Within the English regions, the percentage of owner-occupied dwellings in most local authorities followed a similar trend to the region overall in 2017; however, Newcastle upon Tyne, Manchester and Kingston upon Hull were the only local authorities that had a significantly lower percentage of owner-occupied dwellings than any other in their regions.

  • Purbeck in the South West had the largest decrease in the percentage of owner-occupied dwellings over the last five years, while Hertsmere in the East of England had the largest proportionate increase in privately rented dwellings.

  • Local authorities that had the highest percentage of privately rented dwellings in 2017 were typically in London and surrounding areas, whereas the local authorities that had the highest percentages of owner-occupied dwellings were distributed across English regions, excluding London and the North East.

  • The percentage of privately rented dwellings in 2017 ranged from 6.5% in East Hampshire to 44.1% in Westminster.

  • The higher rate of private renting has become increasingly more common between 2012 and 2017, with increases in the proportion of local authorities that had between 20% and 39% of privately rented dwellings.

Nôl i'r tabl cynnwys

3. Disclaimer

We have published these Research Outputs to provide an indication of the tenure breakdown of dwellings within the private sector at the subnational level. Research Outputs are produced to provide information about new methods and data sources being investigated.

Official statistics on private dwellings by tenure are currently only available at the country level. Statistics on dwelling stock by tenure are available for local authorities but do not provide a breakdown of owner-occupied and privately rented dwellings.

These statistics are subject to marginal error as they are estimates based on a survey, therefore users should refer to the coefficient of variation (CV) and confidence intervals when making interpretations.

Nôl i'r tabl cynnwys

4. Things you need to know about this release

These subnational estimates of dwelling stock by tenure were created using data from the Annual Population Survey (APS) to produce statistics on the number and percentage of owner-occupied and privately rented dwellings out of the total dwelling stock, for subnational geographies in England.

The APS is a household survey and so only dwellings that are occupied are included in the data, because it would not be possible to obtain a survey response from a vacant dwelling. For these estimates, we are interested in the total number of dwellings that are taken up by each tenure type, and so we make an adjustment to the survey-based estimate to account for the dwelling stock that is vacant.

To make this adjustment, occupancy rates by year, region and tenure from the English Housing Survey (EHS) are applied to the owner-occupied and privately rented stock estimates derived from the APS. This is useful as the likelihood of a dwelling being vacant varies by tenure, with privately rented dwellings more likely to be vacant than owner-occupied dwellings.

Occupancy rates are not the same in all regions of England. For example, 83.4% of privately rented dwellings in the South West were occupied in 2016, whereas 94% were occupied in the North West. Occupancy rates were used for each year apart from 2017, as this year’s data are not currently available. We have applied the occupancy rates for 2016 to the 2017 APS data in the calculation of these estimates.

These statistics provide an estimate of the total number of dwellings within each tenure type, including dwellings that are vacant, which have been constrained to the published National Statistics on private sector stock, produced by the Ministry of Housing, Communities and Local Government (MHCLG). The private sector stock (owner-occupied plus privately rented dwellings) for regions sum to England overall, to the nearest thousand dwellings. However, aggregating the private sector stock of local authorities does not sum to the regional or England-level private sector stock. Data for local authorities are taken from MHCLG’s Table 100 (XLS, 597KB), whilst regional data and country level data are taken from Table 109 (XLS, 40KB) and Table 104 (XLS, 37KB), respectively.

We use the number of dwellings in each of these tenures to provide percentages of owner-occupied and privately rented dwellings out of the total dwelling stock. We do not include statistics on the number of social-rented dwellings, which are published by MHCLG in Tables 100, 115 and 116. MHCLG also publishes a breakdown of tenure within the private sector for England overall, and we include this figure in our datasets.

Significant differences throughout this research output refer to 95% confidence intervals around the estimates not overlapping.

Nôl i'r tabl cynnwys

5. Research findings

Highest rates of private renting in London and surrounding areas

According to the Ministry of Housing, Communities and Local Government (MHCLG), nearly two-thirds (62.9%) of the total dwelling stock in England was owner-occupied in 2017. This percentage has decreased since 2012, when 63.8% of dwellings were owner-occupied. However, between 2016 and 2017 the percentage of owner-occupied dwellings increased 0.5%. This was the first increase since 2002.

While owner-occupation remains the most prevalent tenure in England, there has been an increase in privately rented dwellings of 2.9% between 2010 and 2017. In 2017, of the total dwelling stock, 20% was privately rented. This percentage increased year-on-year between 2012 and 2016, but had a decrease between 2016 and 2017 of 0.4%.

Figure 1 shows that the South East had a larger percentage of owner-occupied dwellings in 2017 (69.1%), while London had a lower percentage (51.0%) than any other region.

Figure 2 shows that London had a larger percentage of privately rented dwellings (26.5%) than any other English region. The other regions were more similar to each other for privately rented dwellings than for owner-occupied dwellings.

Local authority-level estimates allow the exploration of differences in tenure within regions, to better understand tenure at a more local level, whilst taking account of the uncertainty around the estimates described by the 95% confidence intervals. As the sample sizes in the Annual Population Survey (APS) are smaller for lower level geographies, the 95% confidence intervals are generally wider.

Figure 3 shows that in 2017, the local authorities with the highest rates of owner-occupied dwellings out of the total stock were distributed across English regions, except for local authorities in the North East and London. Out of the 50 local authorities with the highest rates of owner-occupied dwellings, 18 were in the South East and 12 were in the East Midlands, whereas there were no local authorities in the top 50 in the North East or London.

The local authorities with the highest percentages of privately rented dwellings were in the greater south east. In 2017, out of the 50 local authorities with the highest rates of private renting, 19 were in London and 14 were in the neighbouring South East and East of England regions.

Rate of owner-occupied dwellings varies within regions

The percentage of owner-occupied dwellings can vary significantly across local authorities within the same region, with 95% confidence intervals that do not overlap. For example, Rochford in the East of England was in the top 1% of all local English authorities in terms of the percentage of owner-occupied dwellings (84.1%), whereas Norwich, in the same region, was in the lowest 3% of local authorities (38.9%). Figure 4 shows that the five local authorities with the lowest percentage of owner-occupied dwellings in the East of England were significantly lower than the five local authorities with the highest percentages in 2017.

Figure 5 shows that in Newcastle upon Tyne in the North East, 49.9% of the total dwelling stock was owner-occupied in 2017. This was the lowest percentage of owner-occupied dwellings in the North East, with the next lowest being Middlesbrough (56.7%). Manchester and Kingston upon Hull were the only other local authorities to be significantly lower than any other local authority within their regions, North West, and Yorkshire and The Humber, respectively. This shows how these estimates are useful in identifying areas that have significantly different tenure distributions, despite being relatively close geographically.

In 2017, the percentage of dwellings that were owner-occupied ranged from 24.5% in Tower Hamlets in London to 84.5% in Ribble Valley in the North West. Figure 6 shows that the 60 to 69.99 percentage band was the band that had the largest number of local authority districts in both 2012 and 2017, with over 40% of local authorities in this band. When comparing the rate of owner-occupied dwellings in 2012 and 2017, there was a slightly lower percentage of local authority districts in the higher bands (70 to 79.99% and 80 to 89.99%) in 2017 than in 2012. A higher proportion of local authorities were in the middle percentage bands (40% to 69.99%) for owner-occupied dwellings in 2017.

Higher rates of private renting becoming increasingly common

In 2017, the percentage of privately rented dwellings ranged from 6.5% in East Hampshire to 44.1% in Westminster. Figure 7 shows that the rate of private renting was between 10% and 19% in the majority of local authorities, in both 2012 and 2017. However, between 2012 and 2017, there were increases in the proportion of local authorities that had between 20% and 39.99% of privately rented dwellings. For both years, Westminster was the only local authority in which between 40% and 49.99% of dwellings were privately rented.

Purbeck had largest decrease in owner-occupied dwellings

Figure 8 shows the local authorities that had significant changes to the percentage of owner-occupied dwellings over the last five years. Six local authorities had significant changes in the percentage of owner-occupied dwellings between 2012 and 2017, three of which were in the South West. Three out of the six had a decrease in owner-occupied dwellings and three had an increase. Purbeck, in the South West, had the largest fall in the percentage of owner-occupied dwellings, decreasing from 76.6% in 2012 to 59.6% in 2017.

Hertsmere had the largest proportionate increase in privately rented dwellings

Figure 9 shows the four local authorities that had a significant increase in the percentage of privately rented dwellings between 2012 and 2017. Hertsmere, in the East of England, had the largest proportionate increase; 22.7% of dwellings were privately rented in 2017, which is more than three times the rate in 2012 (7.3%).

Figure 10 shows the changes in tenure over time for the three local authorities with the highest percentage of owner-occupied dwellings in 2017 (Ribble Valley, Rochford and Fareham) and the three local authorities with the highest percentage of privately rented dwellings in 2017 (Westminster, Tower Hamlets and Newham).

Ribble Valley, in the North West, had the highest percentage of owner-occupied dwellings in 2017 (84.5% of all dwellings). However, in 2012, Ribble Valley had the 41st highest percentage of owner-occupied dwellings (76.0%), an increase of 8.5 percentage points over the five-year period. Westminster had the highest percentage of privately rented dwellings in 2017 (44.1%), having remained relatively stable since 2012, with the lowest rate of private renting being 43.2% in 2015.

Figure 10 shows that despite the order of local authorities changing over time, the local authorities with the highest percentage of owner-occupied dwellings did not change significantly between 2012 and 2017.

Nôl i'r tabl cynnwys

6. Feedback

We are keen to receive feedback on this Research Output and the methodology used to produce these estimates, including how they might be improved and potential uses of the data. Please email your feedback to better.info@ons.gov.uk.

Nôl i'r tabl cynnwys

7. Next steps

One of our next steps is to further improve these estimates by incorporating data from multiple sources, including administrative data, on the size of the private rented sector. We will also explore the possibility of applying local authority level occupancy rate adjustments to further increase the accuracy of the estimates.

Nôl i'r tabl cynnwys

8. Quality and methodology

Data sources

Annual Population Survey

Household level data from the Annual Population Survey (APS) were used to calculate the breakdown of owner-occupied and privately rented dwellings. Data for this Research Output were obtained from the APS for the survey months January to December for each year (2012 to 2017).

The variables that were used from the APS are as follows:

  • PHHWTA17 – Household weight

  • LLORD – Landlord of accommodation

  • PCODE – Postcode

  • RELHRP6 – Relation to household reference person

  • TEN1 – Tenure

Further information can be obtained by emailing socialsurveys@ons.gov.uk.

MHCLG live tables on dwelling stock

The Ministry of Housing, Communities and Local Government (MHCLG) produces live tables on dwelling stock (including vacants), which provide the tenure breakdown for England and for local authority districts within England. At the national level, the tenure breakdown is available within the private sector, separating the “owner-occupied” dwellings from the “rented privately or with a job or business” dwellings. However, the tenure data for local authorities in England only provide information on the private sector stock as a whole and so does not provide a breakdown of owner-occupied and privately rented dwellings.

MHCLG has discontinued tables on regional dwelling stock by tenure in Table 109 (XLS, 40KB), with data only published up to 2011, however, this data can still be requested. For geographies that are not available in MHCLG data, we have aggregated the stock data from local authorities where possible.

National Statistics are available on the size of the private sector by country, region and local authority in England (Tables 100 and 109), and so we apply our percentage breakdown of owner-occupied and privately rented stock from the APS to the private sector stock for a given area to derive the owner-occupied and privately rented stock breakdown. Once we have the stock breakdown within the private sector, we can use the total stock published by MHCLG in Tables 100 and 109 and to calculate the percentage of owner-occupied and privately rented dwellings out of the total dwelling stock.

English Housing Survey

The English Housing Survey (EHS) is a national survey conducted by MHCLG, based on households’ housing circumstances in England. When the interviewers visit a dwelling to survey, the interviewer assesses whether it is vacant, and clarification is sought from neighbours. A dwelling is classified as vacant if the dwelling is in-between lets or vacant for a longer period. Surveyors are then required to gain access to any dwelling believed to be vacant, to undertake full inspections. As a result, vacancy rates by tenure, region and year are available from the EHS.

These vacancy rates are used to adjust the owner-occupied and privately rented stock estimates derived from the APS household data, to estimate the total number of dwellings within each tenure type, rather than just the total number of occupied dwellings. It is important to make a vacancy rate adjustment that adjusts for tenure because the likelihood of a privately rented dwelling being vacant is usually higher than for owner-occupied dwellings. Vacancy rates also change over time and across geographical areas.

The latest year that the vacancy rates are available for is 2016, so we applied the 2016 vacancy rates to the 2017 data. This will be revised in future publications to include the relevant years’ vacancy rate.

Other sources of data

Information on tenure can be found for other countries.

Wales

Dwelling stock estimates by local authority and tenure

The Welsh Government publishes annual statistics on dwelling stock by tenure, which provide a breakdown of owner-occupied, privately rented and social housing stock for local authorities. These estimates are produced using a similar method, but the method uses the Labour Force Survey (LFS), which the APS has been boosted from. In addition, the Welsh dwelling stock by tenure estimates are not adjusted for vacancy rates by tenure and so are not directly comparable with the estimates for England in this Research Output.

Scotland

Estimated stock of dwellings by tenure

The Scottish Government produces estimates of the stock of dwellings by tenure and local authority on an annual basis, which also breaks down the private sector into owner-occupied and privately rented dwellings. Since 2001, data have been included on vacant dwellings, but this is only available for the private sector as a whole. Therefore the percentages given for owner-occupied and privately rented dwellings do not account for vacant dwellings.

Northern Ireland

Household tenure (XLS, 235KB)

The Department for Communities – Northern Ireland produces statistics on household tenure at the national level, where a breakdown of owner-occupied and rented households is available, but no distinction can be made between private and social rent.

Tenure

Tenure provides information about whether a dwelling is owned or rented by those who occupy it. If rented, information about the type of landlord who owns or manages the accommodation helps determine the type of renting at the dwelling. There are typically three main tenure categories: owner-occupied, privately rented and social housing.

To derive tenure from the APS, the tenure and landlord of accommodation variables are used to categorise households into each of the three categories in line with census definitions:

Tenure:


Owned outright -> Owner-occupied

Being bought with a mortgage or loan -> Owner-occupied

Part rent, part mortgage -> Owner-occupied

Rented -> Rented

Rent-free -> Rented

Squatting -> Rented

For the households that have Tenure equals Rented, look at Landlord of accommodation variable.

Landlord of accommodation:


LA/Council/Scottish Homes -> Social rent

Housing association, charitable trust or local housing company -> Social rent

Employing organisation -> Private rent

Another organisation -> Private rent

Relative of household member -> Private rent

Individual employer -> Private rent

Other individual private landlord -> Private rent

If there is no response to the tenure question from the APS, or no landlord of accommodation response if the household was rented, the household was excluded from our analysis.

Households versus dwellings

According to census definitions a household is an individual or a group of people living at the same address with common housekeeping such as sharing a living room or at least one meal a day. A dwelling is the physical unit of accommodation in which a household resides (household space).

The APS data used to produce this Research Output are from a household survey and so only cover dwellings that are occupied and do not provide data on any vacant dwellings. We make an adjustment so that we are not only estimating the occupied dwelling stock (households), but are estimating the number of dwellings available to be taken up by each of the tenure types (total dwelling stock). To do this, we apply vacancy rate adjustments from the EHS to the survey-based household estimates so that our estimates refer to “dwellings” rather than “households”.

Data from the APS were filtered to only include the household reference person, to ensure that only one individual per household was included. This method requires an assumption that households sampled in the APS are equivalent to dwellings, with only one household per dwelling, but in reality that may not always be the case. It is possible for a dwelling to contain multiple households and the households within it to be in different tenures. Statistics presented in this Research Output reflect one tenure per dwelling, regardless of the number of households within.

Geography

To identify the geographical location of a household in the APS, the postcode variable was linked to the National Statistics Postcode Lookup (NSPL), which allowed geography codes to be linked to the data, which enables the production of subnational statistics. Estimates are available for England, regions, local authorities (LAs), county and unitary authorities (county and UAs), Nomenclature of Territorial Unit 3 areas (NUTS 3) and combined authorities.

Statistics are not reported for a local authority in which the APS sample size was smaller than 50 for the private sector (owner-occupied and privately rented combined), after the data was filtered. Given that local authorities fit into the larger geographies (county and UAs, NUTS 3 areas and combined authorities), these areas were suppressed if they contained at least one local authority that was suppressed. We suppressed these larger geographies even when the private sector sample size for that larger area was 50 or more so that the suppressed local authority estimates cannot be derived. An exception for this is regions, because the regional APS data are constrained to regional MHCLG data, from which you cannot derive an LA’s stock. County and UA, NUTS 3 and combined authority data have been constrained to aggregated LA data from MHCLG.

Weighting to national population totals

Survey data are based on a sample of the population, which is not always representative of the true population. Even with a randomly selected sample, there could be biases in sample data, but weights are created to deal with these biases and adjust a sample to make it more representative of the population it is designed to reflect. In the APS, there are person and household weights available, and we have used the household weights for this Research Output.

During the process of providing the tenure breakdowns by geography from the APS data, through cross-tabulations, the household weighting variables are applied. Weights assign a value to each survey record to indicate how much “weight” the specific record should have during data analysis. The outcome is a representative weighted breakdown of tenure for each of the given geographies.

Applying vacancy rates and smoothing

The weighted stock for each of the tenure types from the APS was then divided by the occupancy rate (1 minus the vacancy rate) from the EHS. The occupancy rate applied was based on the relevant year, tenure and region. We currently do not have occupancy or vacancy rates for smaller geographies than regions. The weighted stock for a given tenure and area was divided by the occupancy rate, giving a new stock figure for the number of total dwellings for the given tenure, including vacant dwellings.

We smoothed the estimates to take into consideration the previous and following years’ estimates. This helps minimise any effects of year-on-year variation from the survey data. We then derived the percentage breakdown of owner-occupied and privately rented dwellings within the private sector. These percentages were applied to MHCLG private sector stock to derive the final stock figures, which were then used to work out their percentage from MHCLG total stock for a given area.

Measures of uncertainty

There is an element of uncertainty about estimates that are produced from a sample. It is possible to consider how precise an estimate is by constructing a confidence interval around the estimate, to show the range of values in which the true population is likely to lie, in the absence of bias. We have produced 95% confidence intervals, which means that 19 times out of 20, the true value of the population should fall within the upper and lower boundaries. The 95% confidence intervals are calculated using the following formula:

P± 1.96 x standard error, where P is the proportion of dwellings in the private sector.

As the standard error considers the sample size from the APS, areas that have a smaller sample have generally wider confidence intervals.

Throughout this article, we refer to changes being significantly different if the 95% confidence intervals do not overlap. This means that we are confident that 19 out of 20 times, changes will be down to real changes in our population, rather than a result of sampling variability or randomness variation.

We also produced a coefficient of variation (CV) for each estimate, by considering the proportion of dwellings in the private sector (p), and the sample size from the APS for the number of dwellings in the private sector (n), using the following formula:

The lower the CV, the lower amount of uncertainty there is around an estimate. Generally, a CV of lower than 10% is considered good and lower than 20% is considered acceptable. If a CV is greater than 20%, this indicates a relatively large amount of uncertainty around the estimate, which would need to be considered if making conclusions or decisions based on these estimates. As the proportion of owner-occupied dwellings is usually higher than the proportion of privately rented dwellings, the CV tends to be higher for the estimates of privately rented dwellings.

The full methodology is available in the previous Research Output.

Nôl i'r tabl cynnwys

Manylion cyswllt ar gyfer y Erthygl

Nigel Henretty
better.info@ons.gov.uk
Ffôn: +44 (0)1329 447934