1200
627
Published 6 October 2026
Onward, Onward is a UK centre-right think tank developing bold and practical ideas to boost economic opportunity, build national resilience, and strengthen communities.
310
310
Category:
For too long we have failed to ask hard questions about the state of Britain’s housing benefit system. Growth in the cost of housing benefits has been relentless – with a nearly 20% increase in spending projected over just 5 years – and costs set to reach £44 billion by the end of the decade.
We have not asked if this is fair. Fair to taxpayers, paying their own rents and mortgages. Fair to housing benefit claimants in the private rental sector, where the generosity of housing benefits seesaws unpredictably through freezes and uplifts, with no ability to plan. Or fair to families who really and often urgently need a stable roof over their heads, stuck on council waiting lists whilst high earners – and even some foreign nationals – benefit from lifetime tenancies.
This excellent paper asks these fundamental questions. It makes a compelling case that so much of our housing benefit system is failing. It sets out a range of thoughtful proposals for how government can control costs and build a better, fairer system. As we do the hard work in opposition to plan our reforms to housing benefits, I welcome contributions like this that add to this vital debate.
Fixing housing benefits and social housing is a difficult but essential challenge. We must ensure our housing stock and the benefits that fund it support those most in need and enable opportunity, rather than incentivising people to remain stuck in homes that hard-up taxpayers can ill-afford to subsidise. Yet the most urgent pressure is to grip the system and end the inexorable rises in spending that form a growing part of our ballooning welfare bill. This package is a compelling solution to that pressing problem.
Helen Whately
Shadow Secretary of State for Work & Pensions
Housing is the largest single living expense that most families face. Many people work and save for years to own their home outright to be freed from high housing costs. But another group is also living rent-free, with their costs met by the state. One in five households in Britain have their housing costs covered at least in part by the benefits system, through Housing Benefit or the Universal Credit housing element.[1] This is equivalent to all the households in England’s nine largest cities combined.[2]
Paying the rent of over 5.75 million households a year comes with a heavy price – one that is rising fast. Official forecasts suggest that by the end of the decade, the cost of housing benefits will be more than double the cost in 2018.[3] Even this alarming trajectory is severely under-pricing recent and upcoming policy decisions which will accelerate costs further. By the end of the next parliament, core housing benefits spending could exceed £52 billion, up from £30 billion today.[4]
This paper reveals that the current estimates for future housing welfare spending are implausibly low and builds a more realistic baseline for spending in the private and social rented sectors for households without additional housing support needs. It then looks at how this higher projected spending can be brought under greater control quickly and without the need for new primary legislation, saving £4.5 billion a year by the end of the next parliament in 2033/34.
Without the usual price signals that drive other renters to choose carefully a home in the right location they can afford, the benefits system creates an array of perverse incentives for claimants that are neither in their best interests, nor in taxpayers’. As well as addressing the unsustainability of housing benefits spending, this paper also considers ways to correct some of the underlying unfairness to those paying for both their own housing and their neighbours’.
Housing benefits are paid by the Universal Credit (UC) housing element to most working age claimants or by Housing Benefit (HB) to pension age claimants, those with additional housing support needs and those still on legacy working age benefits.[5] These benefits are paid on different terms to those in the private rented sector and those in the social rented sector, creating a series of perverse incentives:
The dynamics governing how claimants respond to housing benefits policy sit both within the housing benefits themselves, and in the wider structure of the benefit system and local government. This paper focuses on measures possible within allowances and rent-setting that could be implemented immediately at the start of a new parliament in 2029/30, without changing allocation policies, work conditionality, tenancy terms or other benefit eligibility.
With so many reasons to stay dependent on the welfare state or choose more expensive areas from which to draw housing benefits, the impact on the public finances should be no surprise. The latest 2026 Spring Forecast indicates that housing benefits spending will be the same proportion of welfare spending this financial year as it was in 2010/11 at 10.9%.[7] Against a much larger overall welfare bill, this stubborn constancy in share is concerning, especially as housing benefits were an area of spending restraint through much of the intervening period.
After a period of cutting, capping and freezing affecting both tenures in the 2010s, housing benefits spending dipped to £20.7 billion by 2018/19. This was followed by two sharp increases when allowances were restored to the 30th percentile of local market rents in 2020 and again in 2024. In the current financial year, £38.8 billion is forecast to be spent on housing benefits.[8] By 2030/31, this is expected to exceed £43.5 billion – more than double the cost in 2010/11 or 2018/19.
Figure 1 – Housing benefits spending, 2010/11-2030/31
Source: DWP, Benefit expenditure and caseload tables 2026[9]
The increase across the forecast period shows that, even with an assumed continuation of the private rented sector freeze, there is no sign that spending will start to come down as it did during the freeze in the latter half of the 2010s. Likewise, in the social rented sector forecast, the increase in permitted social rent rises in the latest Rent Settlement will add an average of £300 million a year to housing benefits spending.[10] Even with the private rented sector freeze in place, total spending continues to rise sharply.
The deeper problem with this already bleak forecast for the public finances is that it does not even begin to capture the true extent of likely housing benefits growth within the forecast period and beyond. The section that follows explains why the forecast is implausible and seeks to build a more likely baseline from which savings from policy changes can then be measured.
The forecast assumes that the current freeze in Local Housing Allowance (LHA) rates that keeps rates at 2024/25 levels will continue at least until 2030/31. This is based on the 2026 Spring Forecast stating that LHA rates will remain frozen until at least March 2027.[11] The OBR simply takes the current policy and has no choice but to assume it continues, absent an affirmative policy decision beyond next March. To understand why the freeze being renewed in the next financial year is highly improbable, it is first necessary to understand how LHA rates are set in policy and legal terms.
LHA rates are the maximum weekly rate of housing benefits claimants in the private rented sector can receive. Claimant households receive benefits for the full amount of their rent if it is lower than this rate, or part of their rent up to this maximum. Households may receive less after their rate has been applied, such as if their total benefit award exceeds the benefit cap or if they have savings above certain thresholds. There are also national caps that apply to LHA rates in the most expensive areas.
The rate applied depends on which area a claimant lives in, known as the Broad Rental Market Area (BRMA), and the bedroom eligibility of the household. Size criteria for LHA rates usually allow one bedroom for each of the following: a couple; any other person aged 16 or over; two children of the same sex; two children under 10 regardless of sex; and any other child. Eligibility is capped at four bedrooms however large the household. A single claimant under 35 with no dependants is limited to the shared accommodation rate (the rate for a room in a shared house) rather than the one-bedroom rate, subject to exemptions for disability or vulnerability.
There are 192 BRMAs in Great Britain, varying in size and population. In England, rates are calculated by Valuation Office Rent Officers collecting a sample of rental information from letting agents, landlords and tenants about private market rents each year. In Wales, they are set by Rent Officers Wales and in Scotland by the Rent Service Scotland via a similar process. From these lists of rents, Rent Officers are bound by legislation to calculate the 30th percentile of local rents for each BRMA.[12] Secondary legislation is then laid each year for which a freeze is applied to override this policy.[13] These negative instruments can be “prayed against” in the House of Commons or the House of Lords, which can lead to debate on the issue. The statutory default remains to uplift rates to the 30th percentile in each area each year, not a continuous freeze.
Second, the current freeze has already eroded the value of LHA rates significantly, making even one further year of freeze difficult for the government to sustain. Shortly after the last uplift to the 30th percentile in 2024, 45.6% of private resented sector households on UC with a housing element were subject to an LHA rate that did not fully cover their rent.[14] In the latest month for which information is available, shortfalls stand at 58.2%. By April 2027, the number of claimant households with shortfalls is on course to have well exceeded the number with shortfalls directly before the last LHA rate uplift to the 30th percentile, and has likely surpassed one million already.[15]
Already, the size of the gap between LHA rates and the affordability of tenancies is significant. In 2026/27, out of the 960 different BRMAs and bedroom categories, 34% already have frozen LHA rates that are lower than the 10th percentile of local market rents in that year.[16] In the next financial year when the freeze or uplift decision comes, this is likely to exceed 55% of areas and bedroom categories, based on the ten-year average of ONS price index of private rents (PIPR).[17] By 2030/31, 95% are likely to be below the 10th percentile.
The extent of the shortfall between the default 30th percentile and a continuation of the LHA freeze at 2024/25 rates into next year is acute. The gap between the likely 30th percentile of market rents for 2027/28 and a continued freeze of LHA rates will leave an average shortfall of around £156 per month. For larger properties, the gap is even more cavernous, with four-bedroom frozen LHA rates an average of £241 per month below the 30th percentile. Even in the BRMAs with the median shortfall across all bedroom categories – Wigan and Barrow-in-Furness – the gap is around £135 per month. For claimant households, these mounting gaps represent a significant share of their total awards, making it difficult to bridge the shortfall from other parts of their benefits awards.
Claimants who entered private rented sector tenancies that were at or above the LHA rate at the 30th percentile in 2024 will likely have experienced a profound erosion in what their housing benefits cover. They would not have known when taking on the tenancy that their housing award would remain frozen for years afterwards as their rent increased. Tenants cannot make long-term decisions about which properties and areas are affordable for their family circumstances, including for sustaining employment, if housing benefits rates see-saw between generous leaps and punitive freezes at unpredictable intervals.
Figure 2 – Average monthly difference between 30th percentile and frozen LHA rates in 2027/28
Source: Onward analysis, MHCLG, Welsh Government, Scottish Government and ONS PIPR data[18]
Third and finally, there is the political likelihood of the freeze continuing into next year or longer. The current freeze has held LHA rates at the same cash level for three years already. If it continues into next year, it will be on par with the 2020/21 to 2023/24 freeze and one year off the 2016/17 to 2019/20 freeze.[19] Continuing to 2030/31, as per the OBR forecast assumption, would be the longest LHA freeze ever. Given how fiercely these long freezes were opposed by those now on the government side, it will be increasingly challenging to maintain discipline in support of this position.
The likelihood of the Government supporting another year of a freeze is reduced by the fact that Labour’s original LHA rates were far more generous than the 30th percentile. When LHA rates were introduced during the last Labour Government, they were set at the 50th percentile (median rent). In addition, if tenants chose tenancies with rents below the median, they were allowed to keep up to £15 per week of the cash difference to encourage them to choose cheaper properties than the maximum and avoid rent inflation up to the median. Claimants could also be eligible to claim for up to five bedrooms, rather than four. The case for reverting to these policy positions was made vociferously by opposition parties during the coalition and Conservative governments.
This historic position has not been forgotten. Ahead of the decision to freeze LHA rates last year, the then Labour Chair of the Housing, Communities and Local Government Select Committee, Florence Eshalomi MP, lent her support to a coalition of forty organisations calling on the government to restore LHA rates to at least the 30th percentile and to assess the impact of returning to the median.[20] She is now a Minister of State in MHCLG. Early Day Motions were signed by Labour, Liberal Democrat, Green and Plaid Cymru MPs calling for LHA rates to be restored to the median.[21] The chorus of stakeholders calling for an end to the freeze has only become louder this year.[22]
The Prime Minister himself has repeatedly referred to the damaging impact of LHA freezes and the need for regular uplifts. In a major speech as the candidate to become the Labour leader in June 2026, three weeks before becoming Prime Minister, Andy Burnham stated:
“We are forced to chase rents in the private-rented sector through the benefits system. When governments try to control these costs by freezing Local Housing Allowance, it makes families homeless and places unfunded pressures on councils when they have to pay for temporary accommodation.”[23]
This landmark speech is consistent with his earlier calls for ending the freeze to reflect rental market conditions. A fortnight before the November 2025 Budget, while Mayor of Greater Manchester, he called for “restoring LHA to reflect real rental costs,” although he fell short of specifying whether this means median rents or another market link.[24] When the previous uplift to the 30th percentile was announced in the 2023 Autumn Statement, he issued a press release saying, “one of our biggest calls has been the urgent need to unfreeze Local Housing Allowance and I am pleased that the Chancellor has listened.”[25]
Recent Resolution Foundation analysis casts doubt on the Prime Minister’s claim about benefits chasing, or driving up, rents in the private rented sector. Findings indicate that only around 10p in every pound of uplift was fed through to higher rents at the bottom of the market after the April 2024 uplift.[26] This means the Prime Minister has less need to avoid the uplift out of a principled aversion to passing on the increase to private landlords.
Against this legislative, affordability and political backdrop, it is highly improbable that a freeze will continue into the next financial year, let alone through to the end of 2030/31 as the Spring Forecast assumes. This context is consequential for any incoming government after the next election. Housing benefits spending will almost inevitably be materially higher than is expected currently – and the imperative to exert spending restraint will be overwhelming.
This paper models a plausible baseline for private rented sector housing benefits spending: it assumes that LHA rates are uplifted to the 30th percentile each year from 2027/28, in line with the legislative default. A list of the key assumptions is given below.
Figure 3 – Private rented sector baselines
Source: Onward analysis, DWP outturn and forecast, Spring 2026[27]
To create this baseline, the model analyses the latest Valuation Office lists of local market rents, known as the Shadow List of Rents, and the devolved equivalents, to commence with the latest snapshot of the distribution of market rents by BRMA on the same basis that LHA rates are set.[28]
The model then applies the ten-year average growth rate from the ONS’s PIPR to estimate how rents rise until the end of the next parliament in 2033/34.[29] At a national level, the compound annual growth rate is 3.81%. This aligns with the latest PIPR, which shows average UK monthly private rent increased by 3.8% in the 12 months to August 2026.[30] If more recent multi-year compound annual average growth rates were used, such as three or five years, the rate would be materially higher, at around 6% nationally. A ten-year average is used to avoid embedding distortion from the period following the Covid-19 pandemic when rents surged and to avoid possible local anomalies from taking single year values alone.
In order to model how many claimants are eligible for different LHA rates, it is necessary to estimate the proportions of claimants in each BRMA and each bedroom rate. For example, some areas will attract a higher proportion of younger claimants only eligible for the shared accommodation rate, whereas others will have more family homes at the lower end of the rental market and therefore attract a higher proportion eligible for three- and four-bedroom rates.
As UC household caseload data does not split by bedroom category, nor gives sufficient detail on household characteristics from which to estimate this, bedroom level caseload proportions applied for each BRMA are based on bedroom category proportions from 2018 Housing Benefit caseload data. While this is lagged, it is the last year before managed migration to UC began, giving the most up to date and reasonably complete picture of the general shape of the entire caseload by eligible bedrooms in each area.[31] At a national level, just over two thirds of the caseload were eligible for one or two bedroom rates, with the rest split between three bedrooms, the shared accommodation rate and four-bedroom rates in descending order of proportion.
Figure 4 – National share of private rented sector caseload by bedroom eligibility, April 2018
Source: Onward analysis, DWP Stat-Xplore[32]
Next, the model assesses how much the caseload is likely to grow in each BRMA across the seven year time horizon, against which the bedroom shares for that area are applied. Caseload growth rates are calculated from the compound annual growth rate in caseload across those receiving housing benefits through UC and HB in each BRMA since April 2018.[33] Extreme outliers in caseload growth are capped to ensure that the total growth remains in line with the overall national rate.
There is no published breakdown of the amount of UC award claimants receive from just the housing element, let alone by each BRMA and for each bedroom category. Again, the model relies on HB, taking the latest month for which data is available for each LA (then converted to BRMA) and each bedroom category.[34] This is then calibrated against the 2026 Spring Forecast, which gives the only breakdown of how much is spent in total on UC housing element in the private rented sector as a separate component within awards to combine with private rented sector HB awards.[35]
To reflect that housing benefits in the private rented sector pay the lower of LHA rates and rent, an adjustment factor is modelled to give an implied pass through of any change in rate. Claimants who already have a shortfall between their LHA rate and rent stand to benefit from a rate rise to the 30th percentile, whereas those for whom the frozen LHA rates still cover their rent will see no change from an uplift.
Those with rents that sit somewhere between the old LHA rate and the new one will gain part of the uplift amount. The model considers the number of claimants in each BRMA that have rents above the previous year’s LHA rate, the gap between old and new LHA rate and where rents are likely to sit between the old and new rates according to the local rent growth rate. At a national level, the model estimates a pass through to benefits of 56.3%.[36]
The UK still has one of the largest social rented sectors in Europe as a proportion of total dwellings, despite the stock declining steadily since the 1980s.[37] There are around 3.4 million homes for general needs social rent, and a further 400,000 for general needs affordable rent.[38] The almost three quarters of the sector that comprises general needs social rent is governed by a jumble of archaic rent policies, tied to the shape of the labour and property market in 1999.
Before the millennium, there was no single basis for setting social rents, with a series of policy changes through the 1970s and 80s having led to wide rent variation between both council housing and housing association tenants, as well as between newer tenants and longstanding ones. From 2002, a new rent-setting policy was introduced, called formula rent, which was meant to provide a single basis for the vast majority of social rents after a ten-year period of gradual adjustment.
The basic input values and caps for formula rent are based on 1999 figures, including individual property values, manual worker earnings and rent distributions from properties of different sizes. This snapshot means formula rent is comparatively lower than market rents in London, while it over-values rents in other areas that have experienced decline.
After a long period of social rent cuts and caps from 2015, only around a quarter of eligible social tenancies have converged up to formula rent. Just 0.1% of local authority social rents and 43% of private registered provider social rents have arrived at formula rent.[39] Lack of progress towards a single social rent-setting basis means that social rents and their associated housing benefits costs are still determined by a mismatch of legacy inputs. There are also discrepancies between housing benefits in the social rented sector and the private rented sector in how bedrooms are weighted in rent calculations, with more generous allowances for very large homes in the social rented sector compared to four in the private rented sector.
Soon after the 2024 general election, the new government consulted on changes to social rent policies. However, it declined to re-open the underlying basis of formula rent, citing a need to move quickly and avoid instability or uncertainty in the sector.[40] Instead, the scope was limited to the annual uprating basis (and some minor technical amendments).
In January 2026, the Government confirmed that convergence would resume over a ten-year period, stepping up to £1/week from April 2027, rising to £2/week from April 2028 onwards, on top of CPI+1%.[41] Social rents above formula rent and affordable rents can increase by CPI+1% only. Accelerating towards a rent goal linked indelibly to a formula based in conditions over a quarter of a century old will further embed distortion in the sector.
The array of different stock types, rent-setting bases and benefit eligibility means the trajectory of housing benefits in the social rented sector is far more complex to project into the future than the private rented sector. The Spring 2026 forecast for housing benefits in the social rented sector captures a reasonable baseline for this tenure type to 2030/31 in Great Britain. Crucially, it includes a significant policy change announced in January 2026, which will accelerate English social rent increases across the period, beyond the usual CPI+1%.[42]
Figure 5 – Social rented sector tenures and benefit eligibility
| Paying benefit | Eligible costs | ||
| General needs | Working age: UC housing element (+HB legacy) Pension age: HB | Social rent: Actual rent + eligible service charges | Affordable rent: <80% market rent (including service charges) |
| Supported housing | HB | Social rent: Actual rent + eligible service charges +IHM if applicable | Affordable rent: <80% market rent (including service charges) + IHM if applicable |
Housing benefits awards are a reserved matter, but rent setting policies are devolved. To assess the impact of the social rented sector rent policies discussed in this paper, it is therefore necessary to model a more limited baseline than the Spring 2026 forecast to exclude Scotland and Wales, service charges, supported housing (and the associated intensive housing management costs, or IHM) and temporary accommodation costs paid by DWP.
This paper’s baseline therefore only includes general needs social rented sector claimants in England whose rent type is social rent (excluding eligible service charges) or affordable rents. The starting difference in baselines between the 2026 Spring Forecast for the relevant combined categories and the modelled baseline is therefore largely down to service charges being excluded in the model. By 2030/31 the modelled baseline moves above the Spring 2026 forecast baseline due to more recent higher-than-expected inflation forecasts used for uprating purposes.[43]
Figure 6 – England general needs social rented sector baselines
Source: Onward analysis, DWP outturn and forecast, Spring 2026[44]
Stock, tenure and average rents by local authority and bedroom size come from the Regulator of Social Housing’s 2025 statistical data returns for private registered providers (PRPs) and local authority registered providers (LARPs).[45] The model calculates formula rent for each authority and bedroom size according to the methodology in the 2026 policy statement on rents for social housing and January 1999 property values.[46] It calibrates this to the Regulator’s published formula rents and compares it with actual rents. Homes below formula rent are treated as still converging.
Each April, social rents can rise by the previous September’s CPI+1%. For social rent, the different rent policies are applied for those above formula rent or still converging up to it across the period according to the 2026 Rent Settlement terms. Estimated turnover is calculated from official social housing lettings data, showing a national churn rate of 4.97% per year.[47] In line with modelling from the Chartered Institute of Housing, the National Housing Federation and others, it is assumed that providers re-let vacant properties at formula rent.[48] Affordable rent is modelled using the Shadow List of Rents for local market rent data with CPI+1% applied.[49] Both paths use the latest CPI forecasts compiled by HMT or 2% for beyond that horizon (for 2029 uprating onwards).[50]
A new Social and Affordable Homes Programme was introduced in 2026, shifting the priority towards social rent homes.[51] Previous local stock growth rates are measured from the Regulator of Social Housing stock data between 2020 and 2025.[52] These rates are then scaled down for affordable rent and up for social rent to reflect the new planned tenure mix. While the new Prime Minister has indicated that he will expand social rent building further, there are not yet details sufficient to quantify this.
As rents change, not all of this feeds through into housing benefits awards, because some tenants are not claimants and some claimants have deductions from their awards (such as the under-occupation deduction or the benefit cap). The draft Impact Assessment for the 2026 rent convergence policy notes that social security impacts are calculated as two thirds of the change in rental income.[53] This factor is applied to the model to scale the pass through to benefits of rent changes.
By combining the private and social rented sector baselines, a substantial underestimation of the likely cost of housing benefits is revealed. By the end of the forecast period in 2030/31, spending on housing benefits in the private rented sector and the general needs social rented sector is set to exceed £40 billion, around £7.6 billion higher than the 2026 Spring Forecast indicates. By the end of the next parliament in 2033/34, it is on course to have more than doubled since 2023/24 if the likely private rented sector uplift occurs.
Figure 7 – Combined baseline
Source: Onward analysis
The policy options in this package are costed against the combined baseline above, using the same underlying basis from which to build the new policy scenarios. It is assumed that they are introduced at the start of a new parliament in the financial year 2029/30 and implemented over the five-year period to the end of 2033/34.
Rates move from the 30th to the 10th percentile of local market rents and are uplifted to the 10th percentile every April thereafter. Existing claimants receive transitional protection that keeps their rate frozen in cash terms from 2029 until the yearly uplift to the 10th percentile overtakes it, or a new tenancy is signed. Claimants above state pension age may claim up to the 30th percentile.
The constant freezing and sporadic uplifting of rates to the 30th percentile is bad for tenants and taxpayers alike. The arbitrary and compounding impact of freezes on claimants, as demonstrated above, shows why this is a particularly regrettable way to achieve spending restraint. If LHA rates remain frozen in cash terms in line with the OBR forecast, they will erode to implausibly low rates, leaving many claimants exposed to growing shortfalls if they remain in their current tenancies or unable to find a property in a given area at all.
As the modelled baseline highlights, the 30th percentile is clearly an unaffordable long-term basis for uplifting housing benefits, and could see the total housing benefits bill quadruple from current levels by 2040 as the caseload continues to grow.[54] It is equally unfair to those who are not reliant on benefits to be priced out of some parts of the country while those on benefits can choose to live in more expensive areas and be subsidised to access effectively a third of the housing market there.
Shifting permanently down to subsidise only up to the bottom tenth of the market in each area is both fairer and more affordable for the government. To prevent further volatility for claimants who are already assumed to have seen LHA rates uplifted to the 30th percentile from 2027/28, a grandfathering arrangement is proposed. Existing claimants in the private rented sector would have their housing benefits awards frozen going into 2029/30 and held in cash terms until the yearly uplift to the 10th percentile overtakes their award. In areas with high local market rent growth rates, this protection will erode faster than in areas with slower growth.
Figure 8 – Mean year for 10th percentile overtake of frozen 2028/29 30th percentile LHA rates
Source: Onward analysis[55]
New claimants or those taking on a new tenancy are only eligible for the 10th percentile LHA rate, calculated on the same basis as the baseline. Claimants above State Pension age, who are less able to adjust their income and outgoings, remain able to claim up to the 30th percentile throughout the period. Over a five year period from 2029/30 to 2033/34, this recommendation is estimated to save a cumulative £6.9 billion. In 2033/34, the in-year saving reaches £1.9 billion.
Figure 9 – 10th percentile LHA rate policy savings
Source: Onward analysis
Eligible rent is capped at 10th percentile LHA rates for general needs social and affordable rent tenancies. Existing claimants have a maximum cash loss of £10 a week for the duration of that tenancy.[56] Pensioners are exempted from the cap.
Claimants in the social rented sector receive high effective subsidies from the state when compared to the full rental market value of the properties they are occupying. On top of this, there is no LHA-style maximum on eligible rent, so housing benefits can exceed private rented sector rates for that area. There is no equivalent to the shared accommodation rate for under 35s in the social rented sector. Extending 10th percentile LHA rates into the social rented sector as a cap on eligible rent creates a single, consistent maximum.
Eligible rent is currently set on a property basis rather than by household characteristics. This means that this policy would act as an additional underoccupancy penalty for those with rents hitting the LHA maximum. Almost a fifth (18.1%) of general needs social rented sector claimant households have one or more spare bedrooms.[57] As with the private rented sector, pensioners are exempted as they will be less able to adjust to this change in circumstances by increasing their income to cover any shortfall. To protect any claimants at the LHA maximum from sudden losses of housing benefits, a further cap is proposed to limit the weekly deduction to £10 per week.
Bringing LHA rates into the social rented sector was proposed previously at the Spending Review in 2015, but never took effect. According to analysis from the National Housing Federation, 18.2% of eligible households were expected to be affected by the cap.[58] A principal reason for this policy’s downfall last time was the inclusion of supported housing, which this recommendation excludes. As social rents have since been held low for so long, this policy only affects a much smaller proportion of total social rented sector properties, predominantly affordable rent tenants.
In 2029/30 when this policy is proposed to be introduced, just over 207,000 social and affordable rent homes are estimated to exceed a 10th percentile cap.[59] This is much lower than estimates for the 2015 cap, which was expected to affect around 800,000 homes in the long run.[60] Just over a third of those estimated to be impacted if this were introduced in 2029/30 are social rent properties, with the majority being affordable rent where the rents can be set up to 80% of market rents. In practice, many providers already use LHA rates as a guide for the maximum affordable rent to charge by number of bedrooms in the house, which is why the cap affects so few claimant households.[61]
As LHA rates only allow claims for up to four bedrooms, those with five or more bedrooms in the social and affordable rent sector would not be eligible to claim higher rates that account for additional bedrooms. There are around 12,500 social and affordable rent homes with five or more bedrooms, or around 0.34% of the total general needs stock.[62] There are only expected to be around 700 homes facing a further reduction at the cap due to having more bedrooms than they are eligible for under LHA rates.
The shared accommodation rate would impact those in the general needs social rented sector who are living alone and under 35 in the small number of areas hitting the LHA cap. As with the private rented sector, exemptions from the shared accommodation rate for disability and vulnerability would apply here too. The 2021 Census indicates that there are just under 89,500 households in England, or 2.1% of social rented sector households, with one resident under 35 with no disability or long-term health condition.[63] 44% of these households are also recorded as not in employment.[64]
Some of these single under 35s will be in supported housing or would be exempt from the shared accommodation rate by vulnerability criteria, such as being a care leaver under 25, a victim of domestic abuse or a former homeless hostel resident. As there is no equivalent to shared accommodation in the general needs social rented sector, the small number affected would need to move into the private rented sector if they cannot afford the shortfall, freeing up a one bedroom property for a couple or tenants over 35.
This policy saves between £42 million and £45 million a year between 2029/30 and 2033/34. As this recommendation interacts with savings from Recommendation 3, these savings should not be summed and the total savings are given below.
Social and affordable rents are held at their 2028/29 cash level until 2033/34 while a new rent-setting basis is developed.
The basis for setting and uprating social rent is highly distorted. It is providing unreasonable effective subsidies to millions of tenants while private renters not on housing benefits are squeezed out of many desirable areas. Racing stock that sits below formula rent towards a calculation tied to 1999 does not to fix the underlying system for determining the value of social homes. As demonstrated by Recommendation 2, affordable rent is also distortionary, providing relatively generous rent levels in some areas compared to the local market.
As the consultation for the 2026 rent settlement acknowledged, preparing an entirely new system for rent-setting requires time for providers to adapt. It remains to be seen what the detail of the proposed devolution settlements will entail too, which could radically alter the structure of council responsibilities by 2029/30. Amidst this uncertainty, there remains a central fiscal imperative: to bring down the cost of welfare. As bringing LHA caps into the social rented sector is not enough to make significant savings on its own, a further step is proposed. A rent freeze should be brought in during the next parliament, while a new long-term rent-setting system is devised.
In addition to welfare savings, it is relevant to account for the loss of revenue to local authorities in the form of rent. The final saving is therefore given in public sector net borrowing terms, in line with the 2015 HMT assessment of the rent cut which includes this as a behavioural effect.[65] Housing associations were reclassified into the public sector in 2015 and then back into private sector classification from 2017, so they are not included in offset accounting.
Providers previously absorbed a 1% rent cut for four years from 2016/17 to 2019/20. The financial position of providers going into a freeze from 2029/30 is uncertain. The latest published forecasts from the Regulator of Social Housing do not include the 2026 Rent Settlement changes for convergence. There are also a number of new legal requirements on property standards and safety that increase costs to providers.
For local authority providers, previous analysis from Savills for the Local Government Association projects that a ten-year rent convergence policy would put Housing Revenue Accounts into a cumulative revenue surplus after around nine years and overall surplus (including borrowing for capital needs) after 30 years, while CPI alone deepens deficits.[66] A freeze from 2029/30 would arrive two years into convergence, before many HRAs are likely to have reached cumulative revenue surplus. The Regulator of Social Housing judged the majority of private registered providers to be financially robust in 2025, prior to convergence increasing their revenue.[67] Pre-convergence forecasts for private registered providers suggest that interest cover was due to go from 89.9% in 2026 to 118.8% in 2030.[68]
In recognition that the financial position of providers would deteriorate during a freeze, it is limited to five years while a new rent-setting basis can replace it that does not chase the outdated formula. There are existing hardship measures that allow an exemption where complying would jeopardise a provider’s viability.[69] Providers are also allowed to set rents for households with incomes above £60,000 at full market rent, so at least some of the windfall for non-claimant social tenants can be recouped.
Recommendations 2 and 3 interact with each other, as the rent freeze reduces the number of households subject to LHA caps as they are uplifted each year. Together, they are expected to save £7.5 billion by 2033/34, net of LA revenue loss. In 2033/34, the in-year saving reaches £2.5 billion. In addition to Recommendation 1, the three policies save a total of £14.4 billion over the five years, and an in-year saving of £4.5 billion in 2033/34.
Figure 10 – Total savings from recommendations 1, 2 and 3
Source: Onward analysis
Update the national LHA rate caps to uplift with the 10th percentile of the highest outer London BRMA, plus 20%.
There are national caps on LHA rates which limit rates in more expensive areas where market rents are highest. When these were introduced in the 2010 Budget, they were based on the principle that those on benefits should not be able to live in areas that were out of reach to many people in work. Recent implementation of the caps has followed the 30th percentile rent of the highest outer London BRMA, plus 20%.[70] Current caps have been in place since 2024/25 and have likewise been frozen since. The shared accommodation rate national cap is the same as the one bedroom rate cap.
As LHA rates are moved to the 10th percentile, the national caps should also reflect the 10th percentile to remain in step. This prevents inner London LHA rates from becoming more generous in relative terms over the period. The current national cap policy is assumed to continue uprating with the 30th percentile plus 20%, along with LHA rates uprating to the 30th percentile. Changing to a 10th percentile-linked cap creates an additional saving of £120 million over the five years to 2033/34.[71] In 2033/34, it saves an extra £38 million in-year.
Figure 11 – Proposed weekly national LHA caps
| Shared | 1 bed | 2 bed | 3 bed | 4+ bed | |
| Current frozen caps | £331.39 | £331.39 | £412.86 | £497.10 | £704.22 |
| 2029/30 caps | £366.38 | £366.38 | £429.31 | £529.89 | £716.22 |
| 2033/34 caps | £430.63 | £430.63 | £508.81 | £628.01 | £848.84 |
Raise same-sex sharing age threshold to 18 and apply the housing cost contribution or non-dependant deduction from 18.
The housing benefits system currently treats young people inconsistently. According to the official statistical measure of overcrowding, known as the bedroom standard, a household is considered overcrowded if any single person (not in a couple) aged 21 or over does not have their own room.[72] This minimum standard is referenced in social rented sector allocations guidance to local authorities for the purposes of determining household needs, but local allocations policies can vary.[73] At the same time, social rented sector households are not considered underoccupied for the purposes of a housing benefits under-occupation deduction if someone between 16 and 21 has their own room.
When a child in a private rented sector claimant household turns 16, their family becomes eligible to claim for a separate room if they are sharing with a same-sex sibling. If the family does not move house to take up this extra room, they simply receive this as a larger benefit award if their rent is above their previous LHA rate. In UC, a housing cost contribution only applies from the age of 21. HB applies a non-dependant deduction from age 18 for some households.
There is a clear mismatch between the statistical measure of occupancy which applies to all homes, including those not claiming benefits, and the benefit treatment of occupancy. Uneven thresholds make housing benefits more generous for families from 16, right at the point where childcare costs are diminishing and that young person could start contributing to household income. It incentivises moving or being allocated to larger homes that may then become unaffordable when that 16 year old leaves in just a few years when the housing cost contribution, non-dependant deduction or under-occupation deduction applies.
Adults can live independently making their own UC claim from 18, with individuals under 35 usually limited to the shared accommodation rate. Moving bedroom sharing rules all the way to the bedroom standard at 21 could therefore be more expensive if more young people moved out to start their own claim. Instead, this recommendation proposes aligning same-sex sharing rules up to 18 while bringing the housing cost contribution or non-dependant deduction to 18 too, aligning with the threshold for an independent claim.
These measures set a clear expectation for young people that having their own room (if previously same-sex sharing) requires contribution from them or their family. Making their own claim at 18 likewise comes with usual conditionality requirements. Claimant families will continue to be eligible for the UC child element for young people in the household until the August after they turn 19 if they are in non-advanced education or training.
Together with the harmonisation of housing-related age boundaries, these steps create a stronger incentive for young people to take on employment, education or training – or for their parents to work to meet the shortfall. Benefit savings are expected, but likely to be small. There are insufficient breakdowns of household composition available to quantify the effect.
This package presents a series of options to control a housing benefits bill that is likely to grow rapidly without reform. There are many other ways in which benefit eligibility, conditionality and tenancies can – and should – be changed to improve the housing benefits system, but most require significant time for primary legislation and implementation.
Despite many attempts to restrain spending through freezes, caps and cuts, the housing benefits bill risks spiralling to even more unsustainable levels if likely policy decisions are taken over the next few years. Making serious savings within a single parliament requires decisive action through existing powers and regulation changes.
Permanently uplifting LHA rates to the 10th percentile offers a more sustainable path that avoids the cycle of arbitrary freezes and sudden cost jumps in the future. Extending the same ceiling to general needs social housing creates a single, consistent maximum across tenures. An immediate social and affordable rent freeze can keep spending under control while longer term reform is considered that leaves behind a rent formula crafted in the last century.
The final two measures on national caps and bedroom eligibility begin to address some of the geographic and contribution-related unfairness that offers choices to claimants that are not available to many working families. This package provides the starting point for rationalising housing benefits spending that is more predictable for claimants and more affordable for the state.
[1] Onward analysis of: DWP, Annual DWP Benefits Statistics Compendium, September 2026, link; ONS, Households by household size, regions of England and Great Britain constituent countries, April 2026 link.
[2] Onward analysis: ONS, Census 2021, Number of households, link.
[3] DWP, Benefit expenditure and caseload tables 2026, link.
[4] Onward analysis; Local Housing Allowance-bound private rented sector tenants and general needs social rented sector tenants only.
[5] The Universal Credit housing element and Housing Benefit are referred to collectively as “housing benefits.”
[6] MHCLG, Social housing lettings in England, tenants: April 2024 to March 2025, November 2025, link.
[7] DWP, Benefit expenditure and caseload tables 2026, link.
[8] DWP, Benefit expenditure and caseload tables 2026, link.
[9] DWP, Benefit expenditure and caseload tables 2026, link.
[10] OBR, Economic and fiscal outlook – March 2026, Annex B, link.
[11] HM Treasury, Spring Forecast 2026, link.
[12] As per paragraph 2 of Schedule 3B to The Rent Officers (Housing Benefit Functions) Order 1997 (link) made under s.122 of the Housing Act 1996. Paragraph 2(2) sets the LHA as the lower of the 30th percentile, determined under sub-paragraphs (4) to (8), and the national cap. The same structure appears in Schedule 1 para 2 of the Rent Officers (Universal Credit Functions) Order 2013 (link) and in Schedule 3B of the equivalent Scotland Order (link).
[13] For example, the latest freeze year is set out under The Rent Officers (Housing Benefit and Universal Credit Functions) (Modification) Order 2026, link.
[14] Onward analysis, DWP Stat-Xplore, Local Housing Allowance indicator by Month, counting Households on Universal Credit, split by proportion with LHA that covers rent and LHA that does not cover rent.
[15] Onward analysis, DWP Stat-Xplore, Local Housing Allowance indicator by Month, counting Households on Universal Credit. By March 2024, claimants with LHA that did not cover rent was 1,009,604. In December 2023, the number of households with LHA that did not cover rent was 985,697, roughly the same as the volume in May 2026 of 985,372, meaning the climb took around three months.
[16] Onward analysis, 329 out of 960 BRMAs and bedroom categories fall below the 10th percentile, calculated from the Valuation Office Rent Officers’ Shadow Lists of Rents and their Scotland and Wales equivalents, link.
[17] Onward analysis, as above with expected growth rate calculated from the ONS Price Index of Private Rents 10 year growth rate by local authority, applied to BRMAs by share of each local authority in each BRMA, link.
[18] Central London BRMA is omitted from the scale due to an extremely high gap of £1,538 per month. 30th percentile comparison does not include national caps.
[19] The 2015/16 freeze started from below the 30th percentile when rates had been uprated by CPI and then 1% in the years previously. It also later included recycling some of the savings through Targeted Affordability Funding to raise LHA rates in areas that had fallen furthest below the 30th percentile.
[20] Chartered Institute of Housing, CIH among 40-strong coalition urging government to unfreeze local housing allowance, October 2025, link.
[21] EDM 2202: Freezing of Local Housing Allowance, link; also this year, EDM 616: Local Housing Allowance rates, link.
[22] See, for example, Crisis, No Home to Go To, July 2026, link; Resolution Foundation, Saving Private Renters, September 2026, link.
[23] Andy Burnham, Andy Burnham Speech at the People’s History Museum, Manchester, June 2026, link.
[24] PoliticsHome, Boost Local Housing Allowance To Tackle Rough Sleeping, Says Andy Burnham, November 2025, link.
[25] Greater Manchester Combined Authority, Statement from The Mayor on today’s Autumn Statement, November 2023 link.
[26] Resolution Foundation, Saving Private Renters, September 2026, link.
[27] Onward analysis, DWP, Benefit expenditure and caseload tables 2026, link.
[28] Onward analysis, DWP, Shadow List of Rents April 2026: collated 1st October 2024 – 30th September 2025, link; Scottish Government, Private Sector Rent Statistics, Scotland, 2010 to 2025, link; Wales does not publish a consistent shadow list of rents, but publishes this data by local authority under ad-hoc statistical requests, the latest of which contains information from 23 October to 3 November 2023 and is adjusted to align with the English and Scottish start years, link.
[29] Onward analysis of ONS, Price Index of Private Rents, September 2026, link.
[30] ONS, Price Index of Private Rents, September 2026, link; note that the PIPR figure is UK wide, while the model excludes Northern Ireland where this policy area is devolved. The national figure given is the average while the model itself uses local growth rates.
[31] Only small volumes of claimants migrated onto UC via natural migration (triggered by a change of circumstances), voluntary migration or new claims before this point.
[32] Onward analysis, DWP Stat-Xplore, Local authority, entitled Bedrooms (LHA only), April 2018. LA to BRMA proportions applied.
[33] HB data on Stat-Xplore is split by pre-April 2018 and post-April 2018, so this date was selected to give the longest possible view from single consistent extracts.
[34] Regional and national scaling factors are used for missing values where volumes are too low to show values.
[35] DWP, Benefit expenditure and caseload tables 2026, link.
[36] Onward analysis, DWP Stat-Xplore LHA covers rent and LHA does not cover rents and Shadow list of rents calculations.
[37] OECD, Social rental housing stock, 2022, link; MHCLG, English Housing Survey, Chapter 1: Profile of households and dwellings, July 2026, link.
[38] Regulator of Social Housing, Registered provider social housing in England – stock and rents, October 2025, link.
[39] MHCLG, How to implement Social Rent convergence, March 2026, link.
[40] MHCLG, Future social housing rent policy, Consultation outcome, January 2026, link.
[41] MHCLG, How to implement Social Rent convergence, Consultation outcome, March 2026, link.
[42] MHCLG, How to implement Social Rent convergence, Consultation outcome, March 2026, link; OBR, Economic and fiscal outlook – March 2026, Annex B, link.
[43] HMT, Forecasts for the UK economy, September 2026, link.
[44] Onward analysis, DWP, Benefit expenditure and caseload tables 2026, link.
[45] Regulator of Social Housing, Registered provider social housing stock and rents in England, link. Regulator of Social Housing, Local authority registered provider social housing stock in England, link; Regulator of Social Housing, Private registered provider social housing stock in England, link.
[46] MHCLG, Policy statement on rents for social housing, April 2026, link.
[47] MHCLG, Social housing lettings in England, tenants: April 2024 to March 2025, November 2025, link.
[48] Chartered Institute of Housing, Rent consultation analysis, December 2024, link; National Housing Federation, How to implement rent convergence, link.
[49] DWP, Shadow List of Rents April 2026: collated 1st October 2024 – 30th September 2025, link.
[50] HMT, Forecasts for the UK economy, September 2026, link.
[51] Homes England, Social and Affordable Homes Programme (SAHP) 2026 to 2036, November 2025, link.
[52] Regulator of Social Housing, Registered provider social housing stock and rents in England, link. Regulator of Social Housing, Local authority registered provider social housing stock in England, link; Regulator of Social Housing, Private registered provider social housing stock in England, link.
[53] MHCLG, Future social housing rent policy from 2026-36: CPI+1% plus convergence options, July 2025, link.
[54] Onward analysis.
[55] Aberdeen and Shire’s frozen 2028/29 30th percentile LHA rate does not cross the 10th percentile due to negative rental market growth.
[56] Onward analysis; this policy is costed for England only in line with the England only baseline.
[57] Onward analysis, DWP Stat-Xplore, February 2026 figures.
[58] National Housing Federation, Impact of applying Local Housing Allowance rates to general needs social tenants, October 2016, link. Note the analysis uses CORE data from 2012/13 due to lack of more recent breakdowns.
[59] Onward analysis, DWP Stat-Xplore, February 2026 figures.
[60] Institute for Fiscal Studies, The Budget and Spending Review: what do they mean for social housing?, December 2015, link.
[61] National Housing Federation, Considerations on rent, 2022, link.
[62] Regulator of Social Housing, Registered provider social housing in England – stock and rents, October 2025, link.
[63] Onward analysis, Census 2021, Lifestage of household reference person, number of people in household with no long-term health condition and tenure of household, link.
[64] Onward analysis, Census 2021, Lifestage of household reference person, number of adults in employment in household, number of people in household with no long-term health condition and tenure of household, link.
[65] HM Treasury, Summer Budget 2015: policy costings, July 2015, link.
[66] Savills for the Local Government Association, Housing Revenue Account Research Update, October 2024, link.
[67] Regulator of Social Housing, 2025 financial forecasts of private registered providers, January 2026, link.
[68] Regulator of Social Housing, 2025 financial forecasts of private registered providers, January 2026, link.
[69] MHCLG, The Direction on the Rent Standard, January 2026, link; MHCLG, Policy statement on rents for social housing, April 2026, link.
[70] Explanatory Memorandum to The Rent Officers (Housing Benefit and Universal Credit Functions) (Amendment) Order 2024, link.
[71] Onward analysis of impact in addition to Recommendation 1.
[72] MHCLG, English Housing Survey 2024 to 2025, Chapter 1: Housing quality, January 2026, link.
[73] MHCLG, Allocation of accommodation: guidance for local authorities, updated April 2026, link.
If you value the work we do support us through a donation. Your contribution will help fund cutting edge research to make the country a better place.