This post signposts readers to the official sources of housing statistics relating to new supply and dwelling stock for England collated and reported by the Department of Levelling Up, Housing and Communities (DLUHC), with some wider reference to UK-wide statistics that the Office of National Statistics (ONS) reports.
Section 1 provides links to the four main relevant source time series, the housing supply, net additional dwellings (net new supply), and the affordable housing, the dwelling stock, and the indicators of new supply (termed new housebuilding), with relevant data summarised in the commentary and in the seven annex tables.
All data and commentary refer to England unless otherwise stated.
Links are also provided to dashboards recently published by the DLUHC within their respective series and by the House of Commons Library that will allow them to access some data (sometimes incomplete) at an individual local authority level at a touch of a mouse.
The aim is to provide policymakers and other interested users with a user-friendly guide to the availability of such statistics and their most appropriate application and use depending upon the purpose of the enquiry, along with their associated limitations and uncertainties, hopefully reducing reader use of midnight oil and wet towels in the process.
Section 2 concludes by summarising the primary issues and problems connected with the series data, before making recommendations concerning their future presentment to users.
Comments and observations are welcomed to asocialdemocraticfuture@outlook.com. If you would like a pdf please request through this address.
1 The main housing supply statistical series
Housing supply: net additional dwellings (net new supply) series
The net new supply series is presented by the DLUHC as the primary and most comprehensive available new housing supply metric for England. This is for two primary reasons.
First, it includes additions and losses to the stock as well as new build completions, which can be higher or lower than net new supply, taking account of net additions or losses.
For example, between 1961 and 1980, as large-scale post war slum and redevelopment demolitions and clearances proceeded apace, net new supply lagged new build completion totals.
Conversely, across recent decades, as gains to the stock from net conversions and from net changes in use have progressively tended to exceed losses from demolitions, annual net new supply has exceeded new build completion totals.
Second, as the DLUHC technical notes on the series sets out, its method of compilation is more comprehensive and accurate.
LAs are required to take account of all the changes to the housing stock within their areas over the previous financial year before inputting them into – what is officially termed – a standardised housing flows reconciliation (HFR) form, and to submit that no later than the subsequent September. The Greater London Authority (GLA) collates similar information from the London boroughs.
Each LA, therefore, has up to five months from the end of the financial year to submit their annual HFR return, compared to the much shorter six-week submission period connected with the building control based quarterly new housebuilding series, considered later.
By having more time to reconcile diverse potential sources of completed new build activity, including from council tax, building control and other records, as well as site visits, LAs can consequently compile a more complete and comprehensive count of additions to their local stocks.
Each November successive to the preceding financial year, the DLUHC publishes a release and a set of accompanying live tables that together describe changes to England’s net supply position.
For example, the 2021 to 2022 annual statistical release published in November 2022 reported the net change that took place between 1 April 2021 to 31 March 2022.
Live Table 118 reports total annual net additions for England, by region, since 2000-1.
Table 120, since 2006-2007, has broken down the different components of net supply to include:
- new build completions;
- net changes to the dwelling stock resulting from gains or losses, whether attributable to approved changes of planning use, to conversions, or from demolitions; and,
- from 2015-16 onwards, changes in use resulting from permitted development rights (PER’s).
Table 1 below provides a summary of the data provided in Table 120.
It confirms that net supply exceeded new build completions by c20,000 dwellings per annum between April 2006 and end March 2022.
More significantly, it shows the proneness of net new supply to fluctuate in a lagged response to the wider macro-economic and housing market conditions, with 2012-13 net supply (reflecting collapse of private speculative activity during and in the wake of the 2008-10 global financial recession), barely half of the level subsequently achieved during 2018-20.
It can be expected that annual net new supply could dip below 230,000 dwellings during the next few years in response to recent housing market uncertainty induced by rising post-covid inflation and interest rates.
Certainly, there is no immediate prospect of the 300,000-dwelling target – as the data shows is clearly aspirational and rhetorical rather than policy-led in nature – being even remotely approached.
Table 1: Net new supply summary, April 2006 to end March 2022
| Year (Apr-Mar) | DLUHC Table 120: New build completions | DLUHC Table 120: Non-new build total net additions | DLUHC Table 120: Total new net additions (new supply) | Non-new build net additions as % of total new supply | Total new supply as % of peak (2019-20=100) |
| 2006-07 | 193,080 | 21,860 | 214,940 | 10.2 | 88.6 |
| 2007-08 | 200,300 | 23,230 | 223,530 | 10.4 | 92.1 |
| 2008-09 | 157,630 | 25,140 | 182,770 | 13.8 | 75.3 |
| 2009-10 | 124,200 | 20,670 | 144,870 | 14.3 | 59.7 |
| 2010-11 | 117,700 | 19,690 | 137,390 | 14.3 | 56.6 |
| 2011-12 | 128,160 | 6,740 | 134,900 | 5.0 | 55.6 |
| 2012-13 | 118,540 | 6,180 | 124,720 | 5.0 | 51.4 |
| 2013-14 | 130,340 | 6,270 | 136,610 | 4.6 | 56.3 |
| 2014-15 | 155,080 | 15,610 | 170,690 | 9.1 | 70.3 |
| 2015-16 | 163,940 | 25,710 | 189,650 | 13.6 | 78.1 |
| 2016-17 | 183,570 | 33,780 | 217,350 | 15.5 | 89.6 |
| 2017-18 | 195,390 | 26,890 | 222,280 | 12.1 | 91.6 |
| 2018-19 | 214,410 | 27,470 | 241,880 | 11.4 | 99.7 |
| 2019-20 | 219,120 | 23,580 | 242,700 | 9.7 | 100.0 |
| 2020-21 | 191,820 | 20,050 | 211,870 | 9.5 | 87.3 |
| 2021-22 | 210,070 | 22,750 | 232,820 | 9.8 | 95.9 |
| Average 2006-22 | 168,959 | 20,351 | 189,311 | 10.5 | 78.0 |
| Average 2017-22 | 199,962 | 23,515 | 223,477 | 10.5 | 92.1 |
Tables 122 and 123 reports total net additional supply by LA district and their component flows, respectively, since 2001-2.
The House of Commons library has produced a very useful online housing supply dashboard in which users at a touch of their mouse can access and print net new supply data covering the most recent ten year reported for individual LAs, as is required.
Alternatively, the DLUHC’s own interactive dashboard can be consulted, although it is less print-friendly.
It does, however, offer a limited facility to access reported net new supply at a local LA level per 1,000 dwellings basis over a range of years.
Whilst the incompleteness of data coverage reported at that local level should be borne in mind, the range of performance displayed between even neighbouring authorities, suggests an under-researched area.
Table 124 reports annual changes in the flows of communal non-self-contained accommodation by LA district. These changes are separate to and are not included in the other tables or headline net supply figures.
Column C of Annex Table One, also since April 2006, charts significant variation between the between the total new build completions reported by the net new supply series and the DLUHC quarterly new housebuilding series.
That series undercounts new build completions by an average annual c17%, increasing to c24% during the recent 2017-22 period.
That tendency underscores the need to use this net new supply series measure not only new net supply, but new housebuilding completions since April 2006.
The 2012-21 dwelling estimates – calibrated to the 2021 census, as the next sub-section explains – suggest that the net new supply series has even itself under recorded net new supply over the past ten years.
The net new supply statistics are also compared against the latest census on its release every ten years to ensure that the sum of net additions over that period are calibrated to the latest census total dwelling count.
Any difference between the that and previous census total dwelling figure is spread evenly across the previous decade consistent with the adjustments made to the dwelling stock estimates.
Such an adjustment to the 2012-21 net supply series will be reported in the November 2023 net new additions (supply) release.
The 2021 census dwelling count will then provide the baseline for future reported annual net supply series changes that, in turn, will provide the data source for successive reported total dwelling stock estimates in England, until the next 2031 census adjustment takes place.
Other revisions to the series live tables are generally limited to those provided by LAs relating to the previous two financial years.
In summary, the net supply series identifies and measures the different components of net new supply, including new build completions, since April 2006.
It provides consequently the primary and most comprehensive and accurate available statistical record of the different components of new supply at both national and LA levels.
The series, however, is only published annually in November for the preceding financial year; given that and because it does not cover starts, it is not forward-looking.
Accordingly, to obtain a more timely but imperfect indicator of new build completions for the year preceding November publication of this series and of future activity, the demonstratable inaccurate new housebuilding series must be consulted.
The department also in 2011-12 removed the information reporting requirement on LAs to categorise new supply by tenure and the net supply series does not break down the reported net supply total by tenure.
This is significant insofar that gross reported new affordable supply greatly exceeds net new affordable supply – as the affordable housing series sub-section will show and discuss.
The dwelling stock series
The dwelling stock series reports total stock estimates back to 1801, when the first census took place, and from 1961 onwards its tenure breakdown.
Until April 2000, the annual dwelling stock estimates provided the source of the annual net additions figure. Starting with 2000-01, the annual dwelling stock estimate has been equal to the previous year’s dwelling stock estimate plus or minus the latest annual net supply of housing figure, subject to the census adjustment process outlined below.
Both the dwelling and net supply series are subject to a decadal retrospective and new baseline census adjustment, usually applied in the year following each successive census.
The May 2023 statistical release for the year ending 31 March 2022 calibrated dwelling stock figures for 2012 to 2021 to the Census 2021 dwelling count figure.
This involved an annual c6,000 dwelling upward adjustment to the previous annual dwelling estimates (the annual net supply change reported in that series from 2012 onwards).
These previous annual net supply figures, as the previous sub-section noted, likewise should be adjusted in the November 2023 net new supply release and its associated tables.
The dwelling count estimates reported for year ending end March 2022 are provisional and subject to revision.
The statistical release and the live tables break down the total dwelling estimate into owner-occupied, private rented sector, local authority housing and housing association tenures, save that it is not possible to split the private housing stock into owner-occupied and private rented sector at the local authority district level.
The tenure split is based on periodic survey data mainly for the owner occupied and privately rented sector, including the English Housing Survey, Labour Force Survey, and on local authority and regulator of social housing returns.
The technical notes accompanying the series provides more background and information.
They confirm that a dwelling in this series is defined in line with the census definition, which defines a dwelling as a self-contained unit of accommodation where all the rooms (including kitchen, bathroom, and toilet) in a household’s accommodation are behind a single door, which only that household can use.
A dwelling can therefore consist of one self-contained household space or two or more non-self-contained household spaces at the same address.
Ancillary dwellings (for example, ‘granny annexes’) are included provided they are self-contained, pay separate council tax from the main residence, do not share access with the main residence (for example, a shared hallway) and that there are no conditional restrictions on occupancy.
Communal establishments, that is establishments providing managed residential accommodation, are not counted in overall housing supply. These include university and college student, hospital staff accommodation, hostels/homes, hotels/holiday complexes, defence establishments (not married quarters) and prisons.
But purpose-built (separate) homes (for example, self-contained flats clustered into units with four to six bedrooms for students) are included in the dwelling estimates, with each such self-contained unit counted as a dwelling.
Non-permanent (or ‘temporary’) dwellings are included if they are the occupant’s main residence and council tax is payable on them as a main residence. These include caravans, mobile homes, converted railway carriages and houseboats.
Permanent Gypsy and Traveller pitches should also be counted as dwellings, if they are, or likely to become, the occupants’ main residence.
Shared ownership dwellings are currently counted as owner occupied within the dwelling stock series.
Dwellings classified as Other Public include dwellings owned by government departments, such as Ministries of Justice and of Defence. They can be vacant awaiting sale or redevelopment.
Table 2 below provides summary data for the decadal years 2001, 2011, 2021, and for 2022.
Table 2: Total dwelling stock estimates and percentage share by tenure, England.
| Year Ending 31 March | Owner-Occupied (1,000s) | Private Rented Sector (1,000s) | Social and affordable rented (1,000s) | Other Public (1,000s) | TOTAL |
| 2001 | 14,735 | 2,133 | 4,236 | 103 | 21,207 |
| 2011 | 14,828 | 4,105 | 3,981 | 62 | 22,976 |
| 2021 | 15,914 | 4,875 | 4,105 | 33 | 24,927 |
| 2022 | 16,124 | 4,885 | 4,117 | 34 | 25,160 |
| % of total stock | |||||
| 2001 | 69.5% | 10.1% | 20% | 0.5% | 100% |
| 2011 | 64.5% | 17.9% | 17.3% | 0.3% | 100% |
| 2021 | 63.8% | 19.6% | 16.5% | 0.1% | 100% |
| 2022 | 64.1% | 19.4% | 16.4% | 0.1% | 100% |
| Source: | Tables 2 and 4, DLUHC Dwelling stock estimates 2022 in release | ||||
DLUHC live Table 100 reports from April 2009 onwards, the total dwelling estimates and its tenure breakdown to down district local authority (LA), save that owner occupied and private rented sectors are subsumed into one private sector category at that LA level.
Table 104 provides a historical annual dwelling stock series, dating back to 1801, with a tenure breakdown first provided in the census year 1961 and then annually from 1969 onwards.
The short story it tells is that in 1961 the LA and private rented (PR) stock combined exceeded the c6.1m dwelling owner-occupied stock.
By 1980 the LA stock continued to climb to c5.2m, before plummeting to its current level below 1.6m.
The PR stock declined from c4.4m in 1961 to 1.6m in 1986, before recovering to its current level of c4.9m.
The owner-occupied dwelling stock progressively increased in total numbers throughout the period to its current c.16.1m dwellings, notwithstanding that its percentage tenure share fell from its recorded high of c70% in 2001 to c64.5% in 2011 and then flatlined for much of the last decade.
Annex Table Two derived from DLUHC live Table 104 reports the total stock dwelling estimates for each year since April 1991 up to 2022, by tenure.
Annex Table Three reports the annual change in the estimated dwelling stock for the same period relative to the previous year and the total period change, also by tenure. It offers the best available backward-looking guide to longer term interlinked stock and net supply trends calibrated to the decadal census.
The ONS publishes a dwelling stock by tenure, UK series recording the estimated number of dwellings in the UK by tenure for each of the UK’s constituent countries, where available, as well as at an aggregate UK and Great Britain level, formerly produced by DLUHC predecessors.
Dwelling stock data in total and by tenure for each English LA can be obtained by a flick of mouse using the House of Commons Library dwelling stock interactive dashboard.
Affordable housing series
The affordable housing series reports the gross annual flow of additional dwellings for rent or sale provided for specified eligible households whose needs are not met by the market.
Such housing can be newly built, acquired, or result from a net gain secured through conversion or from a change in use.
The gross totals reported in the series take no account of demolitions or sales of existing affordable dwellings.
Until recently, these greatly exceeded gross gains, resulting in a substantial net annual reduction in the affordable stock, and then for the last decade or so only a small net increase (see Table 4): a bit akin to trying to run up an accelerating downward escalator.
Headline affordable series figures tend therefore to mask that key outcome.
Affordable rented housing as a sub-tenure (termed government affordable rent in this post), was introduced in 2011 by the incoming Coalition government, to be let by providers of social housing to eligible households at a rent of no more than 80% of the local market rent (including service charges, where applicable).
In contrast, social rented housing is let at c50%-60% of local market rents according to nationally set guidelines.
Intermediate affordable housing is homes for sale and rent provided at a user cost above social rent but below market levels, subject to it remaining at an affordable price for future eligible households and/or to the recycling of expended subsidy into the support of alternative affordable housing provision.
It can include shared ownership and tenure forms where some equity is retained by the provider, as well as other low-cost homes provided either for sale and intermediate rent and, from 2021-22, through First Homes, a tenure form that the current government apparently intends to eventually account for at least 25% of all affordable housing units delivered through planning obligations.
Dwellings purchased under the post-2013 Help-to-Buy programmes were not, however, considered or recorded as affordable housing as purchaser access was not made subject to an income qualification.
Affordable starts and completions for the preceding financial year are published in the succeeding late November/early December alongside updated live tables, which are also subject to scheduled revisions of earlier key figures each June.
Table 1000 reports the most up-to-date summary estimates of affordable housing provided in England since 1991-92.
Table 1000C additionally breaks down that summary data to type of scheme, while Table 1000S reports affordable starts but – as across the series – only from 2015-16 onwards.
Tables 1008C and 1008S break down the total affordable completions and start information to region and LA district, while Tables 1006 and 1007 do likewise for defined tenure types.
Table 1009 breaks the total affordable completion data down according to whether the provided dwellings were newly built or acquired, by tenure.
Tables 1011C and 1011S report from 1991-92 affordable total completions provided and starts (from 2015-16), according to type and source of funding (including nil grant S106), broken down to LA district level.
Annex Table Four collates affordable completion data taken from Tables 1009 and 1011C, by sub-tenure, for the entire 1991-92 to 2021-22 period covered by the affordable housing series.
It reports that gross affordable provision for the April 2018-22 period averaged annually c57,000 – the highest level since April 2007-11.
This was considerably above that above the April 1999-2003 low point average of 33,500, but lower than the c68,500 annual average recorded during the April 1992-96 high point period.
However, Table Annex Four also reports that most affordable dwellings provided between 1999-2003 were let at social rents; during that latest 2018-22 period most were rather let at higher government affordable rent or on intermediate tenure terms.
This compositional variance (amongst other potential ones, such as changes in the bedroom composition of the provided dwellings and/or their location) is a salutary example of the imperative to drill down into the detail and institutional/policy context of subject data before jumping to judgement about headline data and to be wary of soundbite political uses of it. Context is all-important.
Sight should also not be lost that provision trends are a function of previous start levels, which provide a forward but imperfect indicator of future completions; nor that the affordable housing series is particularly prone to cyclical fluctuation related to funding programme changes and their associated delays.
Table 3 simply reports from Table 1011S starts for the April 2018-22 period, by tenure (when known), to provide an indication of future short-term gross completion levels.
New build completions, however, tend to lag for longer than might be expected where a one-to-three-year construction period is assumed. Large scale regeneration projects involving multi-phases invariably straddle many years but could possibly relate to incentives to record starts for programme monitoring purposes or to other unquantified reasons.
Table 3: Affordable starts, by tenure, April to end March, 2018-22
| Social Rent | Government Affordable Rent | Intermediate Tenure | Unknown | Total Affordable Starts | |
| 2018-19 | 8,143 | 25,913 | 25,667 | 1,107 | 60,830 |
| 2019-20 | 12,108 | 24,523 | 25,642 | 5,972 | 68,245 |
| 2020-21 | 11,968 | 19,609 | 17,155 | 8,972 | 57,704 |
| 2021-22 | 14,877 | 13,976 | 17,672 | 16,666 | 63,191 |
Note 1: Social rent includes London Affordable Rent (LAR)
Note 2: Intermediate Tenure total for 2021-22 includes 119 First Homes starts.
Annex Table Five also uses Table 1011C to estimate the number and proportion of affordable dwellings provided using S106 without the use of public grant – in effect through developer cross subsidy realised from sales of dwellings sold at market values then used to help finance affordable social housing dwellings – secured through the planning system and the associated use of S106 agreements.
This table identifies a general trend since the nineties for the proportion of affordable housing across tenures provided through S106 to progressively increase to become its primary funding mechanism, save for some trend interruption in the wake of the Global Financial Crisis (GFC) and for some recent indications of its stalling, probably related to wider conditions in the housebuilding market associated with input and interest rate costs rising in the wake of covid.
Dwellings reported as being provided through nil grant S106 exceeded 50% of total affordable completions in 2019-20.
More detail on how S106 became the primary funding mechanism of affordable (social) housing is provided in Section 1 of The New Infrastructure Levy (IL): Going Round the Mulberry Bush.
Table 1012 reports affordable housing starts and completions, funded by Homes England, combined with the GLA from April 2012, both for starts and completions, between 2009-10 up to 2022-23, reproduced in Annex Table Five.
Further useful background is provided by the latest June 2023 Homes England Housing Statistics publication, while GLA dedicated data can be assessed here.
The affordable housing series is wider in scope and coverage. It aims to provide a complete picture on affordable housing delivered, irrespective of funding mechanism or its source, using more disparate sources, most notably LA annual housing statistical (LAHS) returns.
It is to be expected, therefore, that the affordable completions and starts (where applicable) reported by the DLUHC affordable housing series in its Tables 1000 to 1011 are higher than the combined Homes England and GLA figures that Table 1012 reports.
c44% of 2021-22 total completions of c60,000 dwellings were provided through nil grant S106 – c27,000 dwellings, as Annex Table Five reports.
Added to the c37,000 dwellings reported that year as funded from Homes England GLA combined programmes produces c63,000 dwellings: c5per cent greater than the Table 1011C total.
LAs are asked to only record affordable housing that has not been reported by Homes England or the GLA including affordable housing that did not receive grant funding or developer contributions under planning agreements.
Some nil grant completions could be included in the Homes England/GLA data. Homes England has confirmed to this website that some nil grant S106s are indeed included in its data returns, and it is possible that some LAs double count by including them also in their LAHS returns.
New DHLUHC Table 1013 breaks down affordable provided completions to type of provider, whether LA, registered and non-registered registered providers (RPs), or unknown.
Historically, only non-profit-making organisations, generally and previously known as housing associations, could be registered as providers of social housing or social housing landlords.
Since April 2010 profit-making organisations have been able to register but their output is not meant to be recorded by the affordable housing series, even though it is possible that some such units are included in LA statistical returns.
Registered private providers of social housing (termed RPs in this post) can include organisations providing supported housing and care, local authority subsidiary companies, community groups seeking to develop new housing, commercial developers setting up small subsidiaries to receive Section 106 affordable housing, subsidiaries of investment companies and funds, entities established by registered provider groups, either new parents for group structures or new subsidiaries, as well as small charities, such as alms houses.
According to Table 1013, RPs provided 47,885 of the total 59,356 estimated or recorded new affordable 2021-22 gross affordable (social housing) new supply.
As was highlighted at the beginning of this sub-section, the main drawback of the affordable housing series is that its affordable provision totals are reported gross and do not take account of dwelling losses resulting from sales, demolitions or from other causes. It thereby does not report actual net affordable supply.
The difference between gross and net new affordable supply – as for gross and net total supply the preceding sub-section noted – can be wide and significant.
Most pertinently, the introduction of the Right-to-Buy in 1980-81 led to the losses in the social housing stock resulting from sales progressively outpacing the gross supply of new supply of affordable housing for rent.
Table 678 shows that since 1981 more than two million social housing dwellings were sold. Although sales of RP owned social housing have taken a growing share in recent years, 94% were to local authority tenants across the entire period.
This propelled a reoccurring and generally widening annual net loss of social rented stock, peaking at c98,000 dwellings in 2002-2003.
A January 2023 release on social housing and demolitions 2021-22 updated on nascent but continuing DLUHC efforts to address the absence of a net affordable supply series.
One key output is a seminal time series tracking sales and demolitions of low-cost rental social housing from 1997-98 against the gross new supply of low-cost rental dwellings provided by LAs and RPs.
Figure One (figure 1.7 of the release) provides a graphical representation of that series.
The release explained or claimed that this exhibited secular trend or pattern change was driven mainly by a sharp decrease in sales since 2003-04, as Table 678 reports; by a general decline in the number of demolitions since 2001-02; and, since 2008, by increased gross affordable homes investment.
It did caution, however, that the incompleteness of the sources of the new time series meant that it cannot be taken “as a comprehensive and fully accurate measure of net new supply”, before introducing a more “experimental” statistical dataset aiming to offer a net flow-based measure estimate of annual changes in the net supply of affordable housing for rent, summarising the in- and out-flows to this sector of the housing stock in England.
38,668 new affordable housing units for rent, according to this flow-based dataset, were delivered during 2021-22, while c19,000 low-cost rental dwellings were sold and c7,500 demolished.
Although the LA rental stock decreased by c7,200, the RP stock increased by c18,900, resulting in an estimated annual net total stock increase of 11,700 affordable homes for rent, according to the release, slightly more than the previous year’s 11,200 figure.
The results summarised in Table 4 below – derived from the dwelling statistics reported in DLUHC Table 104 – are supportive of the release’s tentative conclusions, suggesting that they were, perhaps, couched in an over-cautious manner.
Moreover, Table 3 of DLUHC’s Dwelling stock estimates 2022 in release reports a 12,000 increase in the stock of social and affordable rented sub-tenure stock in 2021-22 and over 13,000 on average annually since April 2009, when the loss of stock in the other public sector category is not considered.
Table 4: Average annual change in stock during stated periods, 1992-2022, by tenure
| 31-Mar | Owner Occupied | Rented Privately or with a job or business | Rented from Private Registered Providers | Rented from Local Authorities | LA+RP+other public sector | All Dwellings |
| 1992-2009 | 97 | 108 | 84 | -116 | -36 | 168 |
| 2009-2022 | 89 | 91 | 32 | -19 | 10 | 190 |
| 1992-2022 | 93 | 101 | 62 | -75 | -17 | 177 |
The existing official published data therefore indicates that during the 2009-2022 continuing annual net gains meant that the total affordable stock (RP+LA+other public sector) grew between April 2009 and 31 March 2022 by an annual net average of above c10,000 dwellings (see Annex Tables Two and Three for more detail).
That net increase was, however, crucially far lower – because of the continuing impact of sales especially and demolitions – than the headline average annual gross increase of c57,000 new affordable housing dwellings completed or provided, reported in the affordable housing series, (see Annex Table Four), during the same period.
As was noted earlier, a key and under highlighted outcome and metric.
New Housebuilding: new build
The DLUHC changed the title of this release from “House building: new build dwellings’ to “Housing Supply: indicators of new supply” in September 2020.
A confusing change insofar that the live tables associated with it are only concerned with new build. It also risks conflating the series with the “Housing supply: net additional dwellings” series described by the department as the “primary and most comprehensive measure of housing supply” and “a leading indicator of overall housing supply”.
Indeed, it was for that reason that the Housing supply: net additional dwellings” series was the first to be described and discussed by this post.
To avoid such possible confusion, this post terms the series in line with what its live tables really report.
The Office of National Statistics (ONS), which continues to publish new housebuilding data according to country and at a pan Great Britian and UK level in its House building, UK: permanent dwellings started and completed by country statistical series.
ONS tables relating to England replicate the live tables below published by the DLUHC, updated to incorporate revisions, unlike the annual statistical release commentaries.
Table 213 covers both new build starts and completions, by sector tenure, both on an annual financial year (from 1969-70) and on a quarterly non- seasonally adjusted basis, (from quarter, Q1, 1978 to Q1 2023 at the time of writing, but as is updated by the quarterly reporting cycle).
Table 222 provides similar but quarterly seasonally adjusted figures.
Table 244 catalogues the longer-term new housebuilding record, broken down by tenure also when provided, but on a calendar year basis, from 1946 onwards.
It can and should be used to identify and review longer term trends and features concerning new build completions and its tenure split, as to when provided, prior to April 2006 in conjunction, as appropriate, with Table 213.
Tables 217, 253 and 253a report dwellings completed and started by tenure, broken down to region (quarterly) since 1990-91 (217), LA district (253, by financial year, 252a, quarterly since 1980-81).
Table 254 reports completions and starts, according to whether it was a house of flat, by bedroom size, and by tenure.
All these tables, as noted above, breakdown according to tenure, as to when started or completed according to provider, in contrast to their final tenure status when occupied.
That status can diverge for reasons that include that a progressively larger proportion of affordable dwellings since the early 1990s have been built by private developers in accordance with section 106 planning agreements.
Although ownership of many of these dwellings may later transfer to housing associations, they are often reported and recorded as private enterprise completions within new housebuilding tables.
The most serious and fundamental drawback by far, however, of the series live tables is that they under-enumerate recent new build completions, in total and across tenures, often by a large magnitude, as the net new supply sub-section earlier highlighted.
Column C of Annex Table One reported then that the annual discrepancy between new completions reported in Table 213 of this series and net supply Table 120 reached an annual average of 24% across the most recent 2017-22 period.
This discrepancy also appears sensitive to the wider economic and construction cycle, and/or other determining factors: new build completions reported for 2014-15 in net new supply Table 120 were 25% higher than such completions reported in the new housebuilding series in Table 244; but in 2009-10, such reported completions were only 3.6% higher (Column C of Annex Table One).
The ONS series likewise consistently undercounts recent new build activity and is incomplete insofar that data from Wales is missing.
The new housebuilding series began in 1946 solely based on local authority building control. National House Building Council (NHBC) data was added from 1985 and then 2007 legislation then allowed data from independent approved inspectors.
The sources of the series data accordingly remain building-control-based, considered the best source to identify the start of new build dwellings (specifically the commencement of construction in laying of foundations) and the timeliest measure of new build completion (as measured by the completion certificate), but in recent years building control-based sources have become increasingly fragmented, contributing to the partial coverage problem identified above.
DLUHC statisticians make allowances for non-response (reported as 78% for independent approved inspectors in Q1 2023) through a process of imputation from past returns, revising the series when late returns are submitted.
The technical notes accompanying the series recognises, however, that data coverage in terms of late and non- reporting is a problem – one that demonstrably has and not resolved by the imputation process.
Although the DLUHC has produced a one page guide to housing statistics which shows users how the new housebuilding series relate to other releases, the precise reasons for the under-reporting of new completions by this series remains to be accurately determined, reported, and rectified, while its seriousness and impact appear to be downplayed by both the DLUHC and ONS.
2 Guidelines and recommendations for the future
That the net new supply series provides the best measure of not only new net supply but also for new housebuilding completions since April 2006 has become progressively better recognised over recent years. It remains, however, a lagged backward-looking metric that does not cover starts.
Although flagged as such by the DLUHC in the relevant official releases, this website considers that the department should go further and advise users that the new housebuilding series should only really be used for two reasons.
First, to consider longer-term trends prior to 2006-2007.
The DLUHC interactive dashboard published within its indicators of new supply (new housebuilding) series, reproduced in , reports new build completions according to the tenure of provider from building control sources until 2005-2006, but subsequently by final tenure using the more accurate net supply and affordable housing series. It should be updated and highlighted in future new net supply and new housebuilding releases and perhaps be made a live table.
Second, as a leading but inadequate indicator of new starts.
Given that recent trends in housebuilding activity are an important determinant of wider macro-economic as well as of future housing market outcomes, greater accuracy new build start data is certainly required if it is to provide a fit for purpose leading snapshot indicator of the direction and scale of current and near-future new build activity.
This will probably require a shift to a mandatory site-based time-sensitive reporting requirement placed on the housebuilder backed up by the necessary enforcement and associated political commitment.
On an interim basis DLUHC/ONS should actively consider on how best to mitigate the demonstrable undercounting of new build starts by the series and upgrade its risk rating.
The net new supply series does not breakdown according to tenure, a drawback related to the problem that the provision totals reported in the affordable series are reported gross, not taking account of dwelling losses resulting from sales, demolitions or from other causes.
That series thereby does not report actual net affordable supply – a fraction of the gross completion totals subject to official fanfare and political soundbite.
This should be kept well in sight, which future affordable housing series could more clearly and transparently highlight.
The DLUHC meanwhile is undertaking work to address a request from the July 2020 recommendation by the Housing Communities and Local Government select committee “on publishing net supply of affordable housing by tenure”.
The ancillary information it has produced – as reported earlier – on net affordable additions by tenure should be both further developed and reported and/or cross-referenced should be moved to the annual net supply and affordable housing release commentaries , as appropriate, as should the accompanying commentary on sales and demolitions, rather than be buried in the technical notes and dashboard of the indicators of new supply series (really the new housebuilding series: the one, as above, that needs a stronger health warning!)
In that light, Annex Table Seven could be updated and customised in parallel with the progressive and timely development of the new experimental flow-based series as an addition to the annual net supply and affordable housing series, insofar that stock-based figures are likely to become less accurate as the decade progresses towards the next 2031 decadal census adjustment.