The government’s decision to drop cash-in-lieu flexibility for medium sites landed on the same day as the revised NPPF. It was framed as a victory for affordable housing delivery. The National Housing Federation called it “extremely positive.” Kate Henderson, the NHF’s chief executive, said government had “listened to the social housing sector.”
But read the data that underpins both sides — the NHF’s own detailed consultation response, the HBF’s uncontracted units surveys, and the evidence from practitioners operating at the sharp end of medium-site delivery — and a more contextualised picture emerges. The government has rejected the market mechanism without putting the public subsidy in place to replace it. And the NHF’s own submission contains evidence that makes the case for a better-designed cash-in-lieu route rather more clearly than their public position acknowledges.
What the government decided — and why
The original proposal was straightforward: developers building between 10 and 49 homes — now formally defined in Annex B as ‘medium development’ — would have had discretion to meet affordable housing requirements through cash payments rather than on-site delivery. The consultation triggered an overwhelmingly negative response: 41% of the 910 respondents were strongly opposed, against just 18% who strongly supported the change.
The SMEs lost on the polling figures but it’s worth remembering that SMEs are not a sustainable presence in the market. They ride cycles. Affordable housing businesses are operational businesses with sustained income. They have a more stable presence. A stable presence is a more consistent voice. They of course have huge pressures too, but they are better at articulating some of their needs and concerns then SME developers who are living pay cheque to pay cheque.
Back to the poll result. The government’s stated reason for dropping it was that it “could have a detrimental impact on the manifesto commitment to deliver the biggest increase in social and affordable housebuilding in a generation.” The Treasury, which had been keen on the measure as a supply unlock, conceded the field and the SMEs take it on the chin. A government that is pro-development, as the NPPF makes clear throughout, is not necessarily pro-developer.
The consequence is that medium sites — a 12-home scheme on a tight urban brownfield plot, a 35-home rural development, a 49-unit conversion — now face the full HO5 and PM12 affordable housing regime that applies to a 1,500-home strategic allocation. The NPPF has applied a single answer to a highly differentiated problem.
The NHF’s case — and what it actually says
The NHF’s public position rests on three data points. First, 12% of all S106 acquisitions by housing associations over the past three years were on medium sites — rising to 26% for small and medium housing associations. Second, almost 10,000 affordable homes delivered nationally over that period would not exist without the on-site requirement. Third, cash contributions rarely result in new affordable homes because councils lack the capacity, land, and expertise to deploy them.
These are serious points and deserve serious engagement. Their rural argument in particular is strong and I find it compelling: in England’s most rural areas, more than half of all affordable homes are built on medium sites, and the absence of alternative development opportunities means that cash contributions collected in one community genuinely risk being spent somewhere else.
But the NHF’s own detailed consultation response — submitted in March 2026 and running to 38 pages — is considerably more nuanced than the public-facing position. On the cash-in-lieu question directly, the NHF states: ‘As is the case currently, we recognise that limited, exceptional circumstances … may justify some flexibility regarding on-site delivery.’ It then sets out detailed conditions under which cash-in-lieu should be permitted, including that contributions must be tightly ring-fenced, directed to registered providers rather than LPA general funds, and set at a level that includes a 20% uplift to reflect the true cost of delivery.
That is not a defence of the status quo. That is a design specification for a well-functioning cash-in-lieu mechanism. The NHF’s objection is not to cash-in-lieu per se — it is to the poorly governed version that has historically produced the outcomes they describe. The distinction matters enormously for what good policy would actually look like and gives me hope that something can be done still.
The denominator the NHF doesn’t use
The 10,000 homes figure — approximately 3,300 per year — is a count of what was successfully delivered on medium sites under the current regime. It is not a measure of what the regime attempted to deliver. The relevant question is: how many affordable units were triggered on medium sites over the same period, and what proportion successfully found a registered provider to contract with?
HBF data answers part of this question. An HBF survey of 31 developers in October 2024 found 17,432 S106 affordable housing units with detailed planning permission across all site types that remained uncontracted. A subsequent FOI exercise in June 2025 found approximately 8,500 S106 affordable homes either under construction or due to start within a year that had no RP contract. Around 900 completed affordable homes were standing empty. More than 700 development sites had been delayed or stalled in the previous three years due to the inability to secure an RP.
Meanwhile, NHF’s own consultation submission cites HBF data showing an estimated £9 billion of developer contributions sitting unspent across England and Wales — a rise of £800 million despite declining housing delivery. Almost £3 billion has been held for over five years. £700 million earmarked specifically for affordable housing is waiting to be spent. Tower Hamlets alone holds over £260 million which I found slightly alarming. The NHF uses this data to argue for better governance of contributions – fair point. But it also demonstrates comprehensively that the problem with cash-in-lieu is not the mechanism — it is what happens to the money once collected. The system actually appears to work well at getting money in.
What is the probability of finding an RP on a medium site?
The data is rich enough to attempt a rough estimate, though the methodology must be stated clearly: the HBF and NHF datasets do not share a common denominator, and any probability figure derived from them carries material uncertainty.
If approximately 3,300 affordable homes per year are successfully contracted and delivered on medium sites (NHF data), and medium sites represent roughly 12% of the S106 market by volume of acquisitions, then approximately it could be summised that 1,020 of the 8,500 near-term uncontracted units at any given time are likely on medium sites. That implies a contracted-to-total ratio of approximately 3,300 / (3,300 + 1,020) — roughly 76%.
In other words, on the available data, the probability of finding a registered provider on a medium site is somewhere in the region of three in four. But that figure carries two significant downward pressures that push the real-world probability materially lower.
First, the NHF and HBF data are both backward-looking to a period when RP balance sheets were less constrained than they are today. RPs are currently investing heavily in building safety remediation, damp and mould, and decarbonisation — expenditure the HBF estimates at over £50 billion over five years. The HBF’s own assessment is that the uncontracted pipeline ‘shows little sign of abating.’ I would contend a forward-looking probability is materially lower than a backward-looking one.
Second, the 26% figure for small and medium housing associations is the more relevant population for medium sites. Larger RPs increasingly focus on land-led development and are less likely to bid for six affordable units on a 20-unit scheme. The effective RP pool for the smallest medium sites is shallower than the aggregate figure suggests.
My take is the current system implies a failure rate on medium sites of somewhere between 25% and 40% under current conditions, rising toward the higher end on sites below approximately 30 units where the RP minimum threshold problem bites hardest.
Most RPs will not engage with fewer than 10 affordable units. On a 30% affordable housing requirement, that means sites of fewer than 33 dwellings may have no realistic prospect of RP delivery regardless of what the planning obligation requires. The obligation does not conjure a buyer into existence.
The cascade the HBF asked for and the NHF almost agreed to
What is striking about the debate is how close the two sides actually are when you read the primary sources rather than the press releases. The HBF’s January 2025 parliamentary briefing called explicitly for ‘a greater acceptance of cascade agreements by local planning authorities’ — a mechanism that moves from on-site delivery to alternative tenure to cash payment in lieu as a last resort. The NHF’s consultation response endorses cash-in-lieu in exceptional circumstances with proper controls. Kate Henderson, in direct exchange on LinkedIn, acknowledged that S106 works well “when there is a strong relationship between the developer and housing association” — a formulation that concedes it often doesn’t.
A practitioner in that same exchange put the structural problem precisely: most RPs operate to a minimum requirement of around 10 dwellings per site. On a 30% affordable housing requirement, a 49-dwelling site may be manageable. A 30-dwelling site starts to struggle. Below that, the realistic prospect of RP delivery on a standalone basis becomes thin. Again it would be good to have more granular data on this but it doesnt really exist.
Of course, the absence of a national default cash-in-lieu route does not remove the need for pragmatic LPAs to incorporate sensible cascade provisions in S106 agreements from the outset — it simply means that whether this happens depends entirely on the sophistication of the LPA involved. That is not a national housing policy. It is a postcode lottery based on politics and capability.
The Homes England rescue that may not come
The implicit logic of the government’s position is that Homes England grant funding will plug the viability gap that cash-in-lieu flexibility would have addressed. The Social and Affordable Homes Programme is cited as the mechanism. But Homes England’s existing Affordable Homes Programme is already heavily committed to large strategic sites in towns and cities. The clearing service had listed fewer than 600 homes against a backdrop of 17,400 uncontracted units when last reported, despite over 200 registered providers having signed up.
The rescue funding may not be available at the scale and speed that medium-site delivery requires. And unlike a large strategic site, a 20-unit scheme in a rural village cannot wait three years for a grant funding decision. The capital doesn’t work that way. The SME developer carrying the scheme on a development loan cannot wait that way either. I come back to the point at the start, SMEs are precarious, they are not rental businesses and dont have operational income streams. They cannot afford to take elongation on their programmes without an end in sight.
Fund it or free it
The debate has now produced a clear analytical conclusion, even if the policy hasn’t caught up with it. There are two coherent approaches to affordable housing delivery on medium sites. The first is to fund it properly — to give Homes England the resource and the mandate to act as a genuine backstop buyer on uncontracted S106 units, to capitalise the clearing service properly, and to ensure that grant funding reaches small sites and rural communities rather than concentrating on large urban schemes. The second is to free it — to design a cash-in-lieu mechanism that is tightly ring-fenced, directed to registered providers rather than LPA general accounts, set at a level that reflects the true cost of delivery including a land value component, and subject to time limits that prevent funds sitting unspent for five years.
What is not coherent is the current position: maintaining the on-site requirement as the default, while the clearing service is underfunded, while £700 million of affordable housing contributions sits unspent, while 8,500 affordable homes in the near-term pipeline have no RP, and while the probability of securing an RP on a site of fewer than 33 units is materially below 50%.
The NHF’s own consultation response makes the case for the second option more clearly than their public position suggests. The HBF’s data makes the case for the first option more urgently than the government has acknowledged. The NPPF’s decision to sidestep both is not a protection of affordable housing delivery. It is a deferral of a problem that is already costing the country homes it urgently needs.
The pie can only be split so many ways. Either fund the affordable housing properly, or provide the flexibility for the market to find a route to deliver it. Doing neither is not a housing policy. It is a holding position dressed up as one.
