Making Sense of the Affordable Housing and Lettings Statistics

This is an update, revision, and extension of part of Making Sense of the Housing Supply Statistics published last year in www.asocialdemocraticfuture.org

?It focuses on the most recently published Department of Levelling Up, Housing and Communities (DLUHC) affordable housing (December 2023) and social letting statistical series (January 2024), providing links to customised tables that reproduce series live tables with accompanying commentary, as well as to the source material.

?I have been a user of official housing statistics for many decades, spending many a midnight hour, trying to make sense of them, and in All in One Place: The British Housing Story 1971-90, compiled and wrote on behalf of the Catholic Housing Aid Society (CHAS), one if not the first guide to myriad sources of housing statistics, which was well received and used at the time.

?That I wrote, as I do this, to make it easier for people and organisations (but first and foremost, myself) seeking hard facts to put into objective rather than political spin context to base assessments on the real-world housing situation and policy responses to it.

?As someone who does not work in or conduct housing policy and research continuously, I don’t want to reinvent the wheel every time I revisit an issue, identifying and trawling different sources from scratch each time, but, on the other hand, I don’t want to rely on others to tell me what the facts are, sometimes in a cherry-picked and selective way. I want to put my own interpretation on information.

?Hopefully, interested readers of similar bent find it useful and helpful on that score.

?It includes some new initiatives. New affordable supply is weighted to adjust for the shift in programme composition since 2011 away from Social Rent (SR) towards Affordable Rent in Table 5. The post also links together the affordable housing and social letting series to highlight the neglected importance of letting trends on housing opportunity.

Comments and observations on the above and the material generally are welcome, which will be considered prior to the incorporation of this post in a reworked and wider Making Sense of the Housing Statistics, to be published on www.asocialdemocraticfuture.org, no later than this March.

All data unless otherwise stated refers to England. Table numbers are not in sequence as they are derived from the previous publication.

?Introduction and definitions

The affordable housing series reports the gross annual new supply of additional affordable dwellings for rent or sale provided for specified eligible households whose needs are not met by the market, according to the official definition, including properties not officially defined as ‘social housing’.

Dwellings classified and recorded as such, accordingly, encompass a range of sub-tenures, defined briefly below.

?Housing let at Social Rent levels is set according to national guidelines involving a complicated formula.

A ‘formula rent’ for each property is calculated based on the relative value of the property, relative local income levels, and its size, with reference to a stated aim to ensure that “similar rents are charged for similar properties”, not always realised in practice. Dwellings let in accordance with such arrangements are termed Social Rent (with block capitals) in this post.

?Social Rent (SR) is often stated as let at c50%-60% of local market rents, although the proportion can vary within and between regions and between areas across a wider c30% and c70% range, at the extremes.

?The DLUHC in its January 2024 DLUHC Social housing-lettings April-2022 to March-2023 (Tenancies) statistical release (January 2024 DLUHC social lettings release) advised that the national median SR in 2023 was 47% of market rental levels. ?Figure 26 of that release graphically represented median SR, GAR, and IR (see below) median rents, against weekly amount, according to region, which showed considerable variation.

Affordable Rented housing (termed Government Affordable Rent (GAR)) 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).

According to the same DLUHC release, it was let at a median 71% of market rent levels in 2023-23.

?Intermediate Rent (IR) should not exceed 80% of the current market rate, with the reduced rent officially defined “as an opportunity for the tenant to save towards a house purchasing deposit” and possibly purchase all or a share of the property they currently rent, although the release suggested that IR rents could be higher than that.

?SR + GAR + IR is officially defined in the Social Letting series (as per above link) as social rented housing (no capitals), which can, nevertheless be confusing as it is wider in sub-tenure scope than the Social Rent (with capitals) sub-tenure.

?Intermediate affordable housing is homes for sale and rent provided at a user cost above Social Rent levels 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.

This can include shared ownership and other tenure forms where some equity is retained by the provider, as well as other low-cost homes provided either for sale or for intermediate rent and, from 2021-22, through First Homes - a sub-tenure form that was intended 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 have not, however, been recorded as affordable housing on the ground that purchaser access was not made subject to an income qualification.

Social landlords of officially defined affordable accommodation can be a local authority (LA), often known as council housing, or be a private registered provider (RP), now registered with the Regulator of Social Housing (RSH).

Historically, only non-profit-making organisations, generally and previously known as housing associations, could be registered as providers of social housing or as social housing landlords.

Since April 2010 profit-making organisations, however, have been able to register with the RSH.

Registered private providers of social housing can now 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 their group structures or new subsidiaries;

·???????? small charities, such as alms houses.

Dwellings provided by non-registered providers should not be counted a s affordable housing, but, in practice, sometimes are, but the potential numbers involved are small.

Affordable housing can be newly built, acquired, or result from a net gain secured through conversion or from a change in use.

?Gross totals reported by the affordable housing series take no account of demolitions (representing a direct dwelling loss) or of sales of existing affordable dwellings (representing a transfer of ownership into a different tenure no longer officially defined as affordable).

The series, therefore, does not report changes in annual net affordable supply: the combined flow of stock gains minus losses across each sub-tenure - which can be positive or negative - or its availability, as indicated by new lets.

Rather, it measures new affordable provision activity (predominantly new build, but also including acquisitions and net gains or losses from conversions and from changes in use).

Such activity impacts on but does not determine either net supply or letting availability.

In that light, DLUHC Table 678 reports that during the 40-year period from the introduction of the Right-to-Buy (RTB), starting in 1981 and ending 2022, more than two million dwellings were sold to LA and RP tenants exercising their RTB, c93% of which were to sitting LA tenants.

Such annual combined RTB sales peaked at c84,000 in 2003-04, before collapsing to c.3,100 (mainly because of changes in discount and eligibility arrangements) in 2009-10, before rising to a more stable 2013-22 average of 13,200 dwellings.

These fluctuating RTB stock losses contributed to the results reported in Annex Table Three and Table 4 (both derived from the dwelling stock series).

These tables, when taken together, show that since 1991 the LA net stock lost a net c2.3m dwellings, continuing to annually contract but much more slowly from 2011 onwards.

The combined LA and RP tenure stock (defined in Table 4 as ‘social and affordable rented’ - pertaining to stock let as both Social Rent and Government Affordable Rent) increased by c12,000 dwellings in 2021-22, recording an average annual net stock increase of c13,000 dwellings (disregarding dwellings categorised as ‘other public sector’) during the 2011-21 decadal period.

Changes in net social rented supply appears poorly correlated with new letting supply, at least over the short- to medium-term, although, any increase in gross new affordable supply of social rented dwelling should have a corresponding impact on potential short-term letting supply of new or first lets.

Instructively, the January 2024 DLUHC social lettings release recorded a decrease in social housing lettings made since 2013/14 across all English regions, most marked in London. That is a far more immediate outcome impinging on housing opportunity than marginal changes in net supply.

Before turning to the fuller story told by that and the relevant data tables across both series, it is worth noting that neglected outcome with connected drivers and remedies requires the attention that it deserves.

Affordable housing series data

Affordable starts and completions for the preceding financial year (2022-23) are published in the succeeding late November/early December (2023) alongside updated live tables, which are also subject to scheduled revisions each June (2024).

Table 1000 reports the most up-to-date summary estimates of affordable housing provided (primarily new build completions plus acquisitions and net additions provided resulting from conversion or from change in use: ‘completions’ and ‘provided’ are used interchangeably in this post, unless otherwise stated) since 1991-92.

Table 1000C additionally breaks down that summary data to sub-tenure by type of scheme and funding. Table 1000S reports also affordable starts on a similar basis but – as across the series - from 2015-16 onwards only.

Tables 1008C and 1008S break down the total affordable completion and start information to region and LA district level, while Tables 1006 and 1007 do likewise for different defined sub-tenure types.

?Table 1009 breaks the total affordable completion data down according to whether the provided dwellings were newly built or acquired, by sub-tenure.

Tables 1011C and 1011S report from 1991-92 onwards, affordable total completions and starts (from 2015-16), according to type and source of funding (including nil grant S106), as does Table 1000C but further broken down to LA district level.

?New DHLUHC Table 1013 breaks down affordable provided completions to type of provider, whether LA, or registered, or non-registered registered providers, or unknown.

?It advised that RPs provided 49,844 of all the 63,605 affordable dwellings recorded for 2022-23, with non-registered providers accounting for 2,273, while 2,582 were of unknown sub-tenure. LAs accounted for 8,906: the most since the series started in 1991-92, but still only c18% of the RP delivery total.

?Annex Table Four summarises affordable completion data taken from Tables 1009 and 1011C, broken down to sub-tenure, across the entire 1991-92 to 2022-23 series period.

Since 2001, annual affordable new supply has accounted for c27% of total annual new net additions (supply) on a decadal average basis.

?Affordable new build completions taken as the proportion of the total new build completions recorded in DLUHC Table 120, displays a similar average since 2006.

?Such long-term averages, however, mask considerable annual variations, primarily caused by the proportion of total new build and supply activity taken by affordable new build and net supply increasing sharply in the wake of economic downturns (as private speculative activity contracts) and decreasing likewise during upswing periods marked by private sector recovery.?

63,600 new affordable homes in total were provided in England in 2022-23: seven per cent greater than the previous year: the highest number of affordable completions reported since 2014-15.

This latest high point coincided with the end of the 2016-23 Affordable Homes Programme, as did the earlier 2014-25 uptick peak with the end of the 2011-15 Affordable Homes Programme.

Taking long term decadal averages from April 1991, new affordable supply has been seemingly stable, averaging close to 51,000 dwellings throughout, save for a c10 per cent dip occurring during 2001-11.

?That said, 1991-2001 average completions would have been c10% higher if the first reported year of the entire series, 1991-92, was disregarded - indeed, the peak period of the entire series came early during the 1992-97 period, when 66,200 dwellings were provided on average annually, culminating in a 1996-97 peak of 74,500. Most of the provided dwellings, 56,900, were for Social Rent – more than two and half times the 21,700 Social Rent completions reported in 2004-2005.

Gross annual affordable provision for the most recent April 2018-23 period averaged c57,000 dwellings. This represents the highest level recorded since April 2007-11, considerably above the April 1999-2003 low point average of 33,500, yet lower than the 68,200 annual average dwellings provided during the April 1992-96 series high point period.

However, Table 5 documents that most affordable dwellings provided between 1999-2003 were let at Social Rents or its broad equivalent, when, in contrast, during the latest 2018-23 period, nearly three quarters (74%) of the average annual 35,800 combined social and affordable rent completions total were let at higher Government Affordable Rental levels.

Such compositional variance (amongst other potential ones, including changes in the bedroom composition of the provided dwellings and/or their location) provides a salutary example of the care needed to interpret time-series data on a consistent like-for-like basis.

?Accordingly, Table Five adjusts for that shift towards government affordable housing provision by applying the following weightings: Social Rent = 1; Government Affordable Rent = 0.625; Other Intermediate = 0.5.

?When these are applied, the 2011-21 decadal average annual completion figure compared to that of previous 2001-11 period transposes from around nine per cent higher to around nine per cent lower.

?The most recent 2018-23 annual average is also greatly reduced from c27% higher to around just one per cent higher (when also compared to the decadal 2001-11 average).

Tweaks to the weightings applied would vary that outcome and readers are invited to suggest evidenced improved alternatives.

?In short, the shift to government affordable rent from 2011 onwards has allowed more output to be squeezed from existing resources but at the cost of higher rents.

?Overall, since 1991, disregarding programme and other cyclical variations – and even allowing for the compositional effects identified above - gross affordable supply has on a cross-decadal basis proved broadly stable in total volume terms, despite a generally hostile and contracting wider public expenditure environment for housing investment.

?That, of course, is a different issue as to whether gross affordable supply in volume and compositional terms was – and is – sufficient and/or well-targeted relative to prevailing and changing socio-economic circumstances.

These can include demographics, wider housing supply and its composition, its affordability, as well on other factors impacting on the overall need for housing that is affordable for different households given their own individual income and other circumstances rather than by official definition.

?And, as was earlier identified and will be shown in the next sub-section, new gross affordable supply does not equate with new housing opportunities, especially with respect to access to Social Rented homes.

Affordable starts for the April 2018-23 period, by sub-tenure (when known) provide an indication of future short-term gross completion levels. c71,000 affordable starts were recorded in 2022-23, and an average of c61,000 over the past five years.

Sight should not be lost that they provide a forward but imperfect indicator of future completions; or that the affordable housing series is particularly prone to cyclical fluctuation related to funding programme profiles and to changes to them and associated delays.

New build completions, in practice, tend to lag for longer than might be expected if a one-to-three-year construction period is assumed. Large scale regeneration projects involving multi-phases invariably straddle many years. Lags could also possibly relate to incentives to record starts for programme monitoring purposes or result from other unquantified reasons.

Annex Table Five uses DLUHC Table 1011C to estimate the number and proportion of affordable dwellings provided using S106 without the use of public grant.

Such affordable dwellings are generally secured through the planning system through developer cross subsidy realised from sales of dwellings sold at market values.

?This table identifies a general trend since the nineties save for some trend interruption in the wake of the Global Financial Crisis (GFC) for the proportion of affordable housing across tenures provided through S106 to progressively increase.

This is to a point where it has become the primary funding mechanism of affordable housing, with dwellings reported as being provided through nil grant S106 exceeding 50% of total affordable completions in 2019-20, and 47% in the latest reported 2022-23 year.

The story of how S106 became the primary funding mechanism of affordable housing is recounted in Section 1 of The New Infrastructure Levy (IL): Going Round the Mulberry Bush.

As ever, average figures can mask variation. Generally nil grant S106 provided c53% of the intermediate sub-tenure affordable dwelling total; in London, however, only four of the 1,567 dwellings provided through conventional Social Rent (excluding London Affordable Rent) was provided though nil grant S106, although it did for 39% of the London Affordable Rent dwellings provided. ?

?DLUHC 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 Six.

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 than the Homes England and GLA series, aiming 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.

Affordable completions and starts (where applicable) reported by the DLUHC affordable housing series in its Tables 1000 to 1011 exceed the combined Homes England and GLA figures that Table 1012 reports.

Although 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 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.

?Social Housing Lettings

?Data reported below is taken from the previously cited DLUHC January 2024 Social housing-lettings April-2022 to March-2023 (Tenancies) statistical release, unless otherwise stated. The source of data reported from tables is stated at the end of each linked table.

Notwithstanding the shift to Government Affordable Rent (GAR) since 2011, dwellings let at Social Rent levels still comprised 83% of total new social housing lettings (let at sub-market rents) in 2022-23, with GAR accounting for the remaining c16% of that total (it provided seven per cent in 2012-13, and 13% in 2015-16, and Intermediate Rent (IR) for only one per cen. 2022-23 and 2021/22 outturns were similar.

RPs in 2022/23 provided 71% of these total social housing lettings; LAs the remaining 29%.

75% were for General Needs (GN) and 25% were Supported Housing (SH), defined as housing provided with special design facilities or features targeted at a specific client group requiring support, for example housing designed for older people or those with disabilities.

38% GN lettings in 2022/23 were of properties that had one bedroom (this includes bedsits), 41% had two bedrooms, 19% had three bedrooms and 2% had four or more bedrooms.

87% of total social housing lets were relets of existing social housing (LA+RP) stock.

The remaining 13% were first (new) lets of newly built (and some converted and acquired) dwellings, relating to the annual flow supply of new gross affordable dwellings.

DLUHC Live Table 602 ?advises that total LA lets, including relets, averaged over 400,000 between 1981-2000. These overwhelmingly would have been at Social Rent equivalent levels. They thus held up well during that post 1980 RTB period, despite the massive net loss of stock that marked it; indeed, the proportion of such Social Rent lets as a proportion of stock (the churn) increased to a record 12.3% in 1996-97.

However, as Table 6 (derived from DLUHC live table 602) shows, total LA lets subsequently have fallen significantly and continuously, collapsing in volume terms (on a long-term basis) to a reported c87,000 total lets (including mutual exchanges and short-term lets) in 2022/23.

An expected outcome, given that the LA net stock continued progressively to deplete for much of the period: the product of stock transfers to RPs; of demolitions and redevelopment; and of the continuing impact of the RTB combined with the cessation of new council housing building - only recently resumed at a modest level.

?It should be noted, however, that social lettings data has been subject to classificatory changes during that period.

Table 602 is also derived from the Local Authority Housing Statistics (LAHS) source. It does not cover RP lettings.

The Continuous REcording of Lettings in Social Housing in England (CORE) series involves a different collection methodology than LAHS across different timescales, and covers both LA and RP lets. Tables 6A and 6B report CORE data.

CORE data excludes mutual exchange, short-term and some other lets, which at least partly explain why LAHS lettings figures are higher. New first lettings to RPs are not available in a consistent DLUHC time series table, however.

Figure 13 of the January 2024 DLUHC release did, however, report that over half (51%) of total Affordable Rent properties (including new and relets) were first lets, compared to 38% of new Intermediate Rent lets, whereas only 5% of new Social Rent lets were first lets, even though they provided the great majority of total new lets, as stated above.

New or first lets of social housing represent new housing opportunities to households not previously social tenants, formerly residing in the PRS or in other tenures, or were sharing with others, or were in temporary accommodation or were otherwise homeless.

Table 6 above shows that first lettings to LA new tenants, recorded at c55,000 in 2022-23, were less than a quarter of their 2000-01 levels.

Such plummeting levels of LA new lets constitute a significant outcome, indeed. They mean that LAs operating in the most housing stressed areas, especially in London, facing a growing homelessness crisis, have had to put ballooning numbers in temporary accommodation that is often unsuitable for sustainable family living and invariably highly costly in public expenditure terms – both direct and indirect.

Unsurprisingly, given that wide flow deficiency, 60,040 households at the end of March 2023 were in temporary accommodation according to the GLA Housing in London 2023 publication: 57% of the England total in temporary accommodation, a record 104,510 households.

As Table 7 documents, first lets to new social tenants (including those accepted as statutorily homeless) in London fell even more calamitously from c.40,500 in 1996-97 to c8,500 dwellings in 2021-22.

That year, 12,040 households across the capital were accepted as owed a ‘main homelessness’ duty, where the applicant was adjudged as unintentionally homeless and statutorily eligible for assistance, usually because they are in priority need because of dependent children responsibility or for other defined reasons, such as vulnerability (see Table MD1, homelessness statistics).

What is puzzling, however, is why since 1996-97, the LA churn rate more than halved between 1997-2023, according to Table 6, and from 7.5% to 4.6% as a percentage of the total stock between 2012 and 2023, according to Table 6A.

?Not only has the local authority stock contracted sharply in size, therefore, so has its productivity over time - when defined as the number of new housing opportunities that it generates relative to its size, across any one year.

This suggests that factors other than a continuing net reduction in LA stock – whose decline, as Annex Table Three conveyed earlier, has slowed since 2009 – underlie that declining churn rate. These in isolation, or in combination, could include:

?·???????? a widening affordability gap between social and market rents;

·???????? an intensifying inability or unwillingness of social tenants to exit the tenure due to their persistent and unchanging disadvantaged labour market and their other socio-economic characteristics;

·???????? changing tenancy age structure profiles;

·???????? longer void turnaround times;

·???????? allocation policies; or

·???????? other unidentified factors.

The data is insufficient on which to frame firm conclusions.

That said, the DLUHC in its January 2024 social lettings statistical release, pointed out that a higher proportion of stock was relet (churned) in northern England than in the south: 7.1% of the combined social stock (LA + RP) in the North East region was relet during 2022/23, compared to 2.3% for London.

It suggested a trend connection with regionally varying social rent levels relative to local market affordability. This accords with common sense observation supported by any passing acquaintance with the London’s prevailing house prices and market rental levels (recognising their variation within the capital) – requiring for access, dual incomes of at least sub-professional levels and above.

Yet, RP lettings have fallen despite the RP stock increasing in net numbers (size) since 2011/12, as Figure 7 of the January 2024 DLUHC social lettings statistical release shows across the country and Table 7 did for London, where lettings to new social housing tenants fell 45% across the RP sub-tenure - more than the 33% fall in new council lettings recorded during the same 2012-22 decadal period.

Across London, at least, the productivity of the RP stock consequentially has fallen even more sharply in the most housing stressed region, most in need of additional social lets – a matter of considerable public policy concern given RP’s dominant sub-tenure and repository of public investment status and position.

Information on the volume trend of LA nominations to RPs could be helpful and relevant in that regard.