Number of vacant and second homes, England and Wales: Census 2021

• Of all dwellings, it was estimated that there was a slightly higher vacancy rate in Wales (7.0%) compared with England (5.4%); Wales also had a higher percentage of second homes with no usual residents (1.2%) compared with England (0.6%).

• The South West had the highest percentage of dwellings that were second homes with no usual residents (1.3% of all dwellings); this was closely followed by Wales with 1.2%.

• In 2021, we estimate that 89.7% of unoccupied dwellings in England on Census Day were truly vacant, while 10.3% were second homes with no usual residents.

• In 2021, we estimate that 85.4% of unoccupied dwellings in Wales on Census Day were truly vacant, while 14.6% were second homes with no usual residents.

• Second homes with no usual residents were most likely to be flats, maisonettes or apartments in England (36.7%) and detached houses or bungalows in Wales (32.4% of second homes).

• Flats, maisonettes or apartments were the most common accommodation type of truly vacant dwellings in both England and Wales (45.3% and 26.8% of truly vacant dwellings, respectively).

On Census Day, 21 March 2021, there were 1.5 million unoccupied dwellings in England and 120,450 in Wales. This is 6.1% of all dwellings in England and 8.2% in Wales.

Unoccupied dwellings are units of accommodation that have no usual residents. Some may be used by short-term residents or visitors as second homes, while some are truly vacant, that is, have no indication of being used as a second home and are not inhabited by short-term residents.

Estimates in this release combine intelligence from census data and administrative data sources to provide a breakdown of unoccupied dwellings into truly vacant dwellings and second homes (with no usual residents), which is difficult to determine from census data alone.

Figure 1: Process of determining truly vacant dwellings and second homes with no usual residents

Download this image Figure 1: Process of determining truly vacant dwellings and second homes with no usual residents

.png (205.5 kB)

Of the 1,507,100 unoccupied dwellings in England in 2021, it is estimated that 89.7% were truly vacant and 10.3% were second homes. There were 120,450 unoccupied dwellings in Wales in 2021, of which 85.4% were truly vacant and 14.6% were second homes.

Out of all English regions and Wales, the West Midlands had the highest percentage of unoccupied dwellings that were truly vacant (95.4%), although this region had the lowest percentage of unoccupied dwellings overall (4.8% of all dwellings). The South West had the highest percentage of unoccupied dwellings that were used as second homes (18.1%). London had the highest percentage of unoccupied dwellings (8.0% of all dwellings) of English regions; 91.6% of these were vacant and 8.4% were second homes with no usual residents.

Figure 2: Percentage of unoccupied dwellings, which are truly vacant, and second homes with no usual residents, English regions and Wales, 21 March 2021

When looking at the percentage of truly vacant dwellings out of all dwellings, it was estimated that Wales had a slightly higher percentage (7.0%) compared with England (5.4%). Wales also had a higher estimated percentage of all dwellings that were second homes with no usual residents (1.2% compared with 0.6% in England).

Second addresses

We previously published Second addresses by location and type: Census 2021, which provides census data on second addresses for respondents who said they stayed at another address for 30 days a year or more. The data included second addresses that had usual residents living there, for example, another parent or guardian’s address, as well as those with no usual residents. There were 1,627,450 dwellings used as second addresses in England and 99,610 dwellings used as second addresses in Wales.

Second homes

This release focuses on second homes located in England and Wales that had no usual residents. This is derived from a combination of census data and administrative data (Figure 1). These estimates capture all second homes with no usual residents regardless of length of stay per year and location of the usual residence of individuals staying there (that is, covers non-UK residents). See Section 8: Strengths and limitations for information on how these data on second homes compare with Second addresses by location and type: Census 2021.

In 2021, it was estimated that there were 172,545 dwellings in England and Wales that were second homes without usual residents on Census Day, 21 March 2021, (154,970 in England, 17,575 in Wales). This is the number of second homes that do not have any usual residents, for example, holiday homes, short-term lets, or another address working away from home.

In England, the South West had the highest percentage of second homes with no usual residents (1.3%), out of all English regions. This was closely followed by Wales (1.2%). The lowest percentage was found in the West Midlands (0.2%). In London, 0.7% of all dwellings were second homes with no usual residents. Census data on the second addresses of usual residents found a higher percentage of second addresses used as “another address when working away from home” in London than any other English region or Wales (10% of all second address types).

Figure 3: Percentage of all dwellings that are second homes with no usual residents, by English regions and Wales, 21 March 2021

We estimate the percentage of all dwellings that were second homes with no usual residents ranged from 7.7% in the Isles of Scilly in the South West, to 0.04% in Dudley in the West Midlands. The majority of local authorities had less than 5% of dwellings that were second homes with no usual residents. The exceptions were located:

• in the South West, the Isles of Scilly and South Hams, with 7.7% and 6.0% of all dwellings, respectively

• in the East Midlands, East Lindsey, with 7.4% of all dwellings being second homes with no usual residents

• In London, City of London, and Kensington and Chelsea, with 6.8% and 6.0% of all dwellings, respectively

• in Yorkshire and The Humber, East Riding of Yorkshire, with 6.3% of all dwellings being second homes with no usual residents

• in Wales, Gwynedd, with 6.1% of all dwellings being second homes with no usual residents

• in the East of England, North Norfolk, with 5.8% of all dwellings being second homes with no usual residents

Kensington and Chelsea was the least affordable local authority in England and Wales in 2021, but had the sixth-highest percentage of second homes with no usual residents.

Figure 4: Percentage of all dwellings that are second homes with no usual residents, and vacant dwellings, local authorities, England and Wales, 21 March 2021

Second homes by accommodation type

Accommodation type data are taken from census data, to provide an estimation of second homes and truly vacant dwellings by accommodation type.

Flats, maisonettes or apartments were estimated as the most common accommodation type of second homes in England, making up 36.7% of second homes with no usual residents. In Wales, second homes were most likely to be detached houses or bungalows (32.4% of second homes with no usual residents).

Semi-detached houses or bungalows were the most common accommodation type in both England and Wales overall (30.6% and 30.9% of all dwellings, respectively). However, this was the least common accommodation type when looking at the distribution of second homes that had no usual residents (10.7% in England and 12.6% in Wales).

Figure 5: Distribution of all dwellings and second homes with no usual residents, by accommodation type, country, England and Wales, 21 March 2021

When looking at the percentage of all dwellings within an accommodation type that were used as second homes, caravans or other mobile or temporary structures had the highest percentage in both England (16.5%) and in Wales (27.6%).

Yorkshire and The Humber had the highest percentage of caravans or other mobile or temporary structures that were unoccupied (71.7%). Of all caravans, 50.9% were second homes with no usual residents, more than double the percentage of any other English region. This leaves 20.8% of this accommodation type being truly vacant in Yorkshire and The Humber.

Figure 6: Percentage of all caravans or other mobile or temporary structures being used as second homes with no usual residents, English regions and Wales, 21 March 2021

For other accommodation types:

• of all detached houses or bungalows, in England 0.5% were second homes and in Wales 1.4% were second homes

• of all semi-detached houses or bungalows, in England 0.2% were second homes and in Wales 0.5% were second homes

• of all terraced houses or bungalows, in England 0.5% were second homes and in Wales 0.8% were second homes

• of all flats, maisonettes or apartments, in England 1.0% were second homes and in Wales 1.5% were second homes

Second homes by number of bedrooms

Data on number of bedrooms are taken from Valuation Office Agency (VOA) Property Attributes data, to provide an estimation of second homes and truly vacant dwellings by number of bedrooms. However, we could not allocate a number of bedrooms to 23.5% of second homes in England, and 25.1% in Wales, because of data availability, so a “Missing” category has been included in this analysis.

The most common number of bedrooms within second homes with no usual residents is two, for both England and Wales (31.5% of second homes in England and 31.4% in Wales).

Three-bedroom dwellings were the most common in both England and Wales (42.0% and 49.7% of all dwellings, respectively). However, the percentage of three-bedroom dwellings within second homes with no usual residents was only around half of this (20.9% in England and 26.0% in Wales).

Figure 7: Distribution of all dwellings and second homes with no usual residents, by number of bedrooms, country, England and Wales, 21 March 2021

Two bedrooms were most common within second homes with no usual residents for all English regions and Wales, ranging from 39.5% of second homes in the North East to 18.2% in the East Midlands. London had a higher percentage of second homes that had one bedroom (32.1%) than other English regions and Wales

Figure 8: Percentage of second homes and vacant dwellings, by number of bedrooms, English regions and Wales, 21 March 2021

It was estimated that there were 1,455,010 dwellings in England and Wales that were truly vacant on Census Day 2021 (1,352,130 in England and 102,875 in Wales), using census and administrative data sources. This means there were no usual residents living in these dwellings and there was no indication of these being used by short-term residents or visitors.

Although Wales had the highest percentage of unoccupied dwellings overall (8.2% of all dwellings), it was estimated that London had the highest percentage of truly vacant dwellings (7.3%), out of all English regions and Wales. The East of England had the lowest percentage of truly vacant dwellings (4.3%) and the second-lowest percentage of unoccupied dwellings overall (4.9% of all dwellings).

As Census 2021 was carried out during the coronavirus (COVID-19) pandemic, some people, for example, overseas students or those privately renting, may have moved back in with family members, leaving more unoccupied dwellings.

Figure 9: Percentage of all dwellings that are truly vacant, by English regions and Wales, 21 March 2021

The majority of local authorities had fewer than 1 in 10 dwellings that were truly vacant (317 out of 331 local authorities). Of the 14 local authorities that had a percentage higher than 10%, five were in London, five in Wales, two in the North West, and one in both the East Midlands and the South West.

The percentage of truly vacant dwellings ranged from 26.1% in City of London with the highest percentage, to Dudley in the West Midlands with the lowest percentage (2.3%). A local authority map with the percentage of all dwellings that were truly vacant can be seen in Figure 4.

Vacant dwellings by accommodation type

Flats, maisonettes or apartments were estimated as the most common accommodation type among truly vacant dwellings in England (45.3%). This is nearly double the percentage of flats, maisonettes or apartments out of all dwellings in England (23.5%). Although London had the highest number of truly vacant flats, maisonettes or apartments, this was also the most common accommodation type for truly vacant dwellings in the rest of England, excluding London (36.9%).

In Wales, truly vacant dwellings were also most likely to be flats, maisonettes or apartments (26.8%), closely followed by terraced houses or bungalows (26.4%). Although these percentages are similar, when looking at the percentage of each accommodation type overall that were truly vacant in Wales, this was nearly double for flats, maisonettes or apartments (13.9%), compared with terraced houses or bungalows (7.0%).

Figure 10: Distribution of all dwellings and truly vacant dwellings, by accommodation type, country, England and Wales, 21 March 2021

When looking at the percentage of dwellings within each accommodation type that were truly vacant, caravans or other mobile or temporary structures had the highest percentage that were truly vacant in both England and Wales. In Wales, 38.3% of all caravans or other mobile or temporary structures were truly vacant, and in England, 24.1% were truly vacant. On Census Day, many caravan sites were shut because of the coronavirus (COVID-19) pandemic, so respondents who usually lived on such sites may have lived elsewhere. Respondents could still respond to say the caravan park was where they usually lived.

For other accommodation types:

• of all flats, maisonettes or apartments, in England 10.4% were truly vacant and in Wales 13.9% were truly vacant

• of all detached houses or bungalows, in England 3.5% were truly vacant and in Wales 5.8% were truly vacant

• of all semi-detached houses or bungalows, in England 2.9% were truly vacant and in Wales 4.2% were truly vacant

• of all terraced houses or bungalows, in England 4.9% were truly vacant and in Wales 7.0% were truly vacant

Vacant dwellings by number of bedrooms

In England, 30.3% of truly vacant dwellings had two bedrooms, making this the most common number of bedrooms for truly vacant dwellings. When looking at the number of bedrooms distribution in all dwellings, three bedrooms were most common (42.0% of all dwellings).

In Wales, there was a higher percentage of truly vacant dwellings that had three bedrooms (33.5%). Three-bedroom dwellings were also most common for all dwellings in Wales (49.7%), but the percentage was lower for truly vacant dwellings.

We could not allocate the number of bedrooms to 14.9% of truly vacant dwellings in England, and 18.3% in Wales because of data availability.

Figure 11: Distribution of all dwellings and truly vacant dwellings, by number of bedrooms, country, England and Wales, 21 March 2021

Out of English regions and Wales:

• London was the only area to have the highest percentage of truly vacant dwellings having one bedroom (33.5% of truly vacant dwellings); this may be because of the higher proportion of one-bedroom households in London

• two bedrooms were the most common number of bedrooms in truly vacant dwellings for the North East, the North West, Yorkshire and The Humber, the East of England, the South East and the South West

• three bedrooms were the most common number of bedrooms in truly vacant dwellings for the East Midlands, the West Midlands and Wales

Figure 8 shows the percentage of truly vacant dwellings by number of bedrooms for all English regions and Wales, alongside second homes data.

Accommodation type

The type of building or structure used or available for use by one or more household spaces.

This could be:

• the whole house or bungalow

• a flat, maisonette or apartment

• a temporary or mobile structure, such as a caravan

Whole house or bungalow

This property type is not divided into flats or other living accommodation. There are three types of whole houses or bungalows:

• detached – none of the living accommodation is attached to another property but can be attached to a garage

• semi-detached – the living accommodation is joined to another house or bungalow by a common wall that they share

• terraced – a mid-terraced house is located between two other houses and shares two common walls; an end-of-terrace house is part of a terraced development but only shares one common wall

Flats (apartments) and maisonettes

An apartment is another word for a flat. A maisonette is a two-storey flat.

Dwelling

A dwelling is a self-contained unit of accommodation that may be empty or being lived in, for example, houses or flats. They are usually made up of one household, but those with more than one household are shared and called a “shared dwelling”.

If a dwelling has no usual residents living in them, for example, they are empty after being sold, these are called “unoccupied dwellings”, but may be used by short-term residents or visitors on Census Day, 21 March 2021, for example, holiday homes.

Number of bedrooms

The number of bedrooms available for use in a household’s accommodation.

Second address

An address (in or out of the UK) a person stays at for more than 30 days per year that is not their place of usual residence.

Second addresses typically include:

• armed forces bases

• addresses used by people working away from home

• a student’s home address

• the address of another parent or guardian

• a partner’s address

• a holiday home

If a person with a second address was staying there on census night, they were classed as a visitor to the second address but counted as a usual resident at their home address.

Second home

A second home in this context is defined as a unit of accommodation with no usual residents, which has either:

• been confirmed as a census respondent’s second address where the person stays at for more than 30 days per year that is not their usual place of residence

• been identified as being a “second home” by its owner through census responses

• has information from administrative data sources, which indicates the likelihood of being a second home, such as Council Tax premiums.

Short-term residents

A census short-term UK resident is anyone born outside of the UK who, on 21 March 2021, had stayed or intended to stay in the UK for a period of three months or more but less than 12 months.

Unoccupied dwelling

An unoccupied dwelling refers to a unit of accommodation with no usual residents, although they may be used by short-term residents or visitors on Census Day. They can be recorded as either of the following on Census Day, 21 March 2021:

• second residence, holiday home or vacant home only

• occupied by short-term residents or visitors only

Usual resident

For Census 2021, a usual resident of the UK is anyone who, on Census Day, was in the UK and had stayed or intended to stay in the UK for a period of 12 months or more, or had a permanent UK address and was outside the UK and intended to be outside the UK for less than 12 months.

Vacant dwelling

A truly vacant dwelling is unoccupied on Census Day with no usual residents, has no indication of being used as a second home and is not inhabited by short-term residents.

Producing estimates of vacant and second homes

Census data provide the number of unoccupied dwellings in England and Wales on Census Day, 21 March 2021. Producing a breakdown of unoccupied dwellings that are truly vacant or used as second homes is difficult to calculate from census data alone. Statistics in this release combine intelligence from census data and administrative data sources to provide this breakdown, through the following hierarchical steps.

1. Dwellings that were unoccupied on Census Day provide a list of Unique Property Reference Numbers (UPRN) to base our data from.

2. If a census respondent (usual resident in England and Wales) indicated that an unoccupied address we have listed is where they stay at for more than 30 days a year and is not their usual place of residence, it is flagged as a second home. The census collects data on the following second address types:

• holiday home

• another address when working away from home

• armed forces base address

• student’s term-time address

• student’s home address

• another parent or guardian’s address

• partner’s address

Of these, most will have usual residents staying at them so would not be counted as second homes in this release, but we are likely to capture holiday homes and another addresses when working away from home at these second addresses.

3. A census response was received indicating a truly vacant dwelling.

4. During census collection, some housing associations provided a list of all of their vacant properties, which allowed us to allocate these dwellings as truly vacant. Data do not cover all housing associations, but others would have filled out responses for each individual vacant dwelling.

5. For non-responding households, there were addresses that received a field visit. If they were indicated as a “holiday accommodation” or “second residence” by a field officer, they were flagged as a “second home”.

6. Where an address had exemption, discount or premium code from the Council Tax data, these were used to classify dwellings as either a “second home” or “truly vacant”.

7. If a dwelling had an indication from the utilities (ElectraLink) data of being unoccupied or de-energised (that is, the supplier does not believe it should be consuming electricity), it was given a classification of “truly vacant”.

8. Remaining dwellings that received a field visit indicating them as being a “vacant dwelling” were given a value of “truly vacant”. This rule was placed last in the hierarchy because of the difficulty for field officers to distinguish between households that were temporarily absent and those that were truly vacant.

9. Remaining unoccupied dwellings that did not meet the conditions of any of the rules were assigned as “truly vacant” or a “second home” through a binomial logistic regression model, including the following variables:

• from census: dwelling output area, sub-group (for example, town living or university town), accommodation type

• from Council Tax: exempt type, discount type, premium type, recently updated flag

• from Personal Demographic Service data (NHS Digital) – recent updated flag, number of people present

• from Driver and Vehicle Licensing Agency Driver Data – latest application type, recently updated flag

• from English School Census (Department for Education) – number of residents. These data are only available for England, an equivalent source was not used for Wales.

The source of information used to allocate an unoccupied dwelling into either vacant or second home were:

• census response – 43.1% in England, 47.8% in Wales

• housing associations’ vacant lists – 1.3% in England, 0.6% in Wales

• non-responding households field visit – 30.0% in England, 26.4% in Wales

• Council Tax data – 9.6% in England, 12.3% in Wales

• utilities data – 3.9% in England, 2.5% in Wales

• binomial logistic regression model – 12.1% in England, 10.4% in Wales

Accommodation type

Census data provide the accommodation type split of vacant dwellings and second homes, at the address level. Accommodation type is available for all unoccupied dwellings.

Number of bedrooms

Census collects data on the number of bedrooms, but this is only available for households with usual residents. To provide the distribution of vacant dwellings and second homes by number of bedrooms, we linked April 2021 Valuation Office Agency (VOA) Property Attributes data to 2021 Census data at the address level (UPRN). The number of bedrooms from VOA data was available for 96.5% of all dwellings recorded in the census for England and 95.7% for Wales. Flats, maisonettes or apartments were least likely to have number of bedrooms data (44.2% of dwellings with missing data).

We were able to add the VOA number of bedrooms for 84.2% of unoccupied dwellings in England and 80.7% of unoccupied dwellings in Wales. This suggests that unoccupied dwellings have a higher likelihood of missing number of bedrooms data in VOA.

Data sources

Census data provide estimates of the characteristics of all people and households in England and Wales on Census Day, 21 March 2021. It is carried out every 10 years and gives us the most accurate estimate of all the people and households in England and Wales.

At the Office for National Statistics (ONS), we made every effort to ensure Census 2021 counted everyone. We used AddressBase Premium to ensure every household in England and Wales received invitations to complete their census questionnaire. We supplemented this with other data sources identifying communal establishments such as halls of residence and care homes. We ensured address data were as accurate as possible by working with GeoPlace to update our address frame prior to Census Day.

Council Tax data are collected for domestic properties in England and Wales, and the ONS receive this data from every individual local authority. The data indicate whether there are exemptions, discounts or premium codes for a property, which allow us to identify whether a property is vacant or used as a second home. Each local authority has their own criteria for exemptions and discounts. Data used were for March and April 2021 combined, but only included active records on Census Day.

ElectraLink data for England and Wales from May 2021, which cover households where there is a live electricity contract, was used to determine electricity usage at a dwelling. Census supplied ElectraLink with a list of problematic postcodes where it was difficult to identify vacant dwellings, and data were returned for addresses that had a smart meter, a recent reading taken since 1 March 2021 or an average daily consumption of greater than or equal to two kilowatt hours. A report was also received covering all domestic addresses, and provides information pertaining to likely occupation status of them based on customer activity or consumption levels. A dwelling would be flagged as vacant if a dwelling is de-energised for more than three months, the last prepayment read or top-up more than six months ago or the latest advance (covering census) below the 95th percentile.

Comparing with other releases

The Office for National Statistics (ONS) previously published census data that looks at the location of second addresses that are used as holiday homes across England and Wales, as well as how many people use them and where they travel from.

Data were published on second address types, providing a breakdown of the type of second address for those who stay at another address for 30 days a year or more.

These both rely solely on census responses of usual residents in England and Wales confirming they stay at another address for 30 days a year or more.

This current release differs to what is previously published in that:

• it only focuses on second homes with no usual residents

• it captures second homes that do not have anyone staying at the address for 30 days a year or more, for example, short-term lets

• it captures second homes in England and Wales that are only stayed at by residents who live outside of England and Wales

Estimates in this release may differ from other published sources (including annual figures on vacant dwellings published by DLUHC covering the period 2004 to 2022 (Table 615)) because of different data collection periods and possible implications of the Census 2021 occurring during the pandemic. 

We can compare the percentage of unoccupied dwellings between 2011 and Census 2021, see Housing in England and Wales: 2021 compared with 2011, but this is the first census we have been able to provide the split of unoccupied dwellings into truly vacant and second homes.

Quality

Quality considerations along with the strengths and limitations of Census 2021, more generally, can be found in our Quality and Methodology Information (QMI) for Census 2021.

Further information on our quality assurance processes is provided in our Maximising the quality of Census 2021 population estimates methodology.