“There is always a well-known solution to every human problem: neat, plausible, and wrong.”
This 1917 quote from satirist H.L. Mencken came to mind when I read the Times 21
st October piece summarising the new Centre for Policy Studies (CPS) and Create Streets report.
[1] The Times reported the advocated approach would enable 4.6 million homes to be built:
“without harming the green belt [and]
avoid damage to farming or the environment.”
It would be unfair to suggest the CPS/Create Streets report is “
wrong”. There is in fact much to agree about it: Green Belt designation is not a proxy for landscape quality, biodiversity, agricultural value or public access. There are plainly locations where legacy Green Belt boundaries are no longer sensible in view of growth needs.
Development around rail stations is generally preferable to pushing housing beyond the Green Belt into more car-dependent locations. The report also recognises that what it describes as “
intuitive grey belt” (the archetypal
derelict petrol stations, abandoned depots and other obviously brownfield sites) is only a small proportion of the Green Belt release necessary.
But the report (or more how it was headlined in the Times) points indirectly at a wider problem. Analysis of solutions to the housing crisis often comes with very large numbers associated with individual policy changes (in this case, 4.6 million homes) attracting headlines, as intended.
It is not that the individual ideas or studies are necessarily wrong: most capture a tangible opportunity. The problem is:
-
that the apparent precision and certainty of the answer (such as 4.6 million homes) – at least in the public debate - can conceal how dependent it is on assumptions; and
-
the size of the prize it dangles in front of policy makers can lead people to treat measures that are complements as substitutes for other things that are also needed.
The CPS/Create Streets report is a good example of the first. Some other recent contributions to the housing debate provide examples of the second.
The challenge of assumptions
The new CPS/Create Streets report raises some good points: its spatial work is a useful reminder that not every Green Belt field is environmentally off-limits or that fragments of brownfield land alone can meet the need; it recognises that the Government's adopted 800-metre station policy will capture only a subset of the wider opportunity; and it explores what might happen if the radius were expanded to 1,200 metres and attention shifted from existing stations to railway lines where new stations might be created.
[3]
These are legitimate avenues to explore, but the journey from mapped land to the headline 4.6 million homes relies on the upper end of its assumptions. The report identifies 152,347 hectares of low- or medium-quality agricultural land within 1,200 metres of railway lines. The report says its calculations assume 35 dwellings per hectare (dph) and development on 60 per cent of the land (a typical development platform, accounting for land needed for infrastructure and open space). Those inputs produce about 3.2 million homes. The 4.6 million figure is generated only by applying 50 homes per hectare to 60 per cent of the land.
[4] That matters because the Government expressly reduced its minimum density around rail stations in the NPPF from 50 dph to 35 dph because:
“Some respondents ... considered they could create viability and market demand challenges in some locations, particularly where density requirements rely on partially flatted development to come forward. Some respondents thought that more rural locations, and areas with lower land values, could be most affected by these challenges.”
In other words, in many of the locations suggested in the report, these densities are going to be unviable.
Of course, three million homes of theoretical capacity is still a very large number. Nor is 50 dph too high in every rail-connected location. The point is that the result is assumption-led. Changes to density, developable area, or the distinction between proximity to a railway line and genuine accessibility can add or subtract hundreds of thousands of homes from the estimate before any developer has tested a site’s feasibility. Treating all Grade 3 agricultural land as 'medium quality' obscures the distinction between Grade 3a, which forms part of the best and most versatile agricultural resource, and Grade 3b, which does not. Once one reflects site-specific constraints, actual walking routes, land assembly, infrastructure, railway capacity and the requirement for viable places, and the number will move again.
The report's move from the Government’s NPPF focus on well-connected stations to railway lines further illustrates the point. A 1,200-metre geometric buffer around a line identifies a useful search area, but a railway line is not a station where people can access the train, a new station must be funded and services timetabled without compromising the wider network.
Where new stations are realistic, the land value created by development may help fund them as the report advocates. Metro-land is an appealing precedent.
But once the proposition relies on land assembly, a new station, enhanced railway capacity, schools, utilities and roads, it has become a major infrastructure and delivery programme.
Capacity vs delivery
This brings us to perhaps the fundamental issue: the gap between physical capacity, planning capacity and delivery capacity.
Spatial analysis can identify land where homes might fit, but planning must determine whether those locations can make coherent new neighbourhoods and how impacts will be addressed. Delivery then depends on landowner participation, infrastructure, utilities, finance, public sector capacity, demand and the rate at which different tenures can be absorbed.
The delivery part is only as strong as the weakest link in the chain. Brownfield land exists without being viable. Rail lines can exist without technical ability to introduce a new station. Local-plan allocations can exist without infrastructure funding. Planning permissions can exist for unviable developments in a failing land market. Planned s.106 affordable homes can exist without Registered Provider demand to buy them. Removing one constraint may simply reveal the next.
This is not a reason to do nothing, but it is a reason to ask demanding questions about the practicalities and timing of posited ideas. Without those answers, capacity is useful evidence, but it is not the same as an effective policy.
Policy measures as additives not substitutes
Now to the second problem: there can be a familiar pattern in debates over the housing crisis: an important piece of evidence about the benefits or impacts of one lever acquires a headline number, and that is then argued to be as a reason to not also pull the other. A few examples:
-
Brownfield land is the archetype: CPRE's ‘State of Brownfield 2025’ identifies minimum capacity for 1.41 million homes on previously developed land in England, including 770,000 on sites with some form of planning permission.
[8] This is routinely taken to support a further proposition: because 1.4 million homes can be accommodated on brownfield land, substantial greenfield or Green Belt release is unnecessary or can be reduced.
Yet 1.4 million is less than five years of what is needed. Our
progress report on building out land on the 2021 Brownfield Registers suggests some of the 1.4m potential homes already existed at the base date and at most half of the capacity is deliverable within 15 years. Further, the late 1990s policy fixation on brownfield was associated with
a precipitous collapse in housing delivery.
-
The same danger can arise in tenure debates and the widely recognised need to increase the supply of affordable homes, including for social rent. Whilst few reputable bodies would suggest there is no need to also increase the supply of market homes,
the active role of market housing in meeting need is often overlooked – or even denigrated - by some in the debate.
And yet the strategic economic and fiscal case for more Social Rent housing recently cited by the Prime Minister is fundamentally
dependent on delivering significantly more private homes than currently.
-
New Towns or Garden Cities have been discussed in similar terms. They are important components for a serious long-term supply strategy, but a programme of large new settlements will mature over decades and cannot remove the need for urban extensions, brownfield regeneration, intensification or smaller sites, at least in the meantime. Yet the CPRE in 2025 said areas with New Towns should have their housing targets
“radically reduced” to reflect the huge changes to their communities that a new town will bring.
-
The supply-side planning reforms launched by the Government in 2024-2026 – much needed and positive as they were for development - were sometimes cited as having fulfilled everything the industry said was required, and the-then Prime Minister said what was now needed was for developers to just “
get on with it”.
Yet, from the beginning, it was always clear the situation was more complex.
The Government has subsequently had to play ‘wack-a-mole’ on a host of other barriers to housing delivery,
Most recently, the current market downturn has led - after sustained pressure - to a return to
demand-side support via its Your First Home scheme.
-
Public sector land assembly is another: some propose that if the state taxed land differently, or bought it closer to existing use value, it could capture the uplift itself and boost supply.
Some have argued more public sector intervention in the land market would
“weaken the power of the speculative housing delivery model.” Of course, where scale or fragmented ownership defeat the market, as with New Towns and comprehensive regeneration, more public sector intervention may be right. But CPO is a scarce, slow and contested instrument. Only 61 were confirmed by authorities in England in 2024.
Public land assembly can do more, but most homes – built on literally hundreds of sites every year - will still need to come from land that promoters and housebuilders option, promote and build out. And of course, it still needs an effective planning system.
The popular attraction of the single answer – especially that with a big number - is understandable: concentrate the growth or responsibility somewhere else and reduce the pressure on other parts of the system. Unfortunately, real world housing delivery is not quite so obliging.
The missing theory of change
Housing policy does not suffer from a shortage of individual ideas. It suffers from an insufficiently integrated account of how they might fit together, how different factors combine to raise output, and the critical limitations and interdependencies. A ‘theory of change’ would provide that account.
It works backwards from the required output, through the delivery chain from planned idea to actual homes built, and makes explicit the causal links, assumptions and binding constraints involved at each stage. It would identify the potential contribution of private sale, Social Rent, other affordable tenures, build-to-rent, specialist housing and self-build. It would identify the land required from brownfield regeneration, urban intensification, suburban renewal, Green Belt review, urban extensions, station growth and New Towns. It would then test the enabling conditions for each: plan coverage, land assembly, infrastructure, public investment, development finance, utilities, skills and market demand.
It would also recognise geography and time. A home in a low-demand market is not a substitute for a home in an unaffordable city region that needs housing to support agglomeration benefits. A New Town completing in the 2040s does not meet a shortfall in the next five years. A brownfield permission stalled by remediation costs is not equivalent to a serviced strategic site that can build now. Social housing can help absorption and meet acute need, but its delivery still requires land, developers and, under the funding model advocated by Shelter/NHF, a functioning private market to generate Section 106 contributions. It would address that one particular measure might work well in ten years, but would not be an effective solution in the current market.
Creating a theory of change is not easy and it is much less satisfying than a silver bullet because it reinforces the hard choices. More brownfield development does not eliminate the case for Green Belt release. Equally, the latter does not remove the need to secure urban intensification. New Towns do not mean existing towns and villages do not need to grow. Social housing does not make market delivery irrelevant; any more than market delivery removes the need for affordable homes. An effective regime of Strategic and Local Plans can support longer term growth, but it will do relatively little or nothing for what gets built before the middle of the next decade and thus is no substitute for measures that support planning applications for sustainable development under the decision-making policies of the NPPF.
The practical conclusion is that tackling the housing crisis in England needs the whole lot. The policy aim should not be to choose the most attractive measures. It should be to understand the contribution, dependencies and failure triggers of each and then have in place a portfolio of approaches, some of which will deliver small marginal gains rather than what is counted in millions. These measures will then need continuous and rapid recalibration to rapid changes in different market cycles.
This blog started with the CPS/Create Streets report. It makes a useful contribution to the policy portfolio, strengthening the case for a more discriminating Green Belt policy and for serious consideration of rail-led growth. CPRE is on to something when it says brownfield capacity is an opportunity not be squandered. Shelter and the NHF are right that England needs far more Social Rent. And so on; there are lots of good ideas.
As the political parties inevitably begin contemplating their policy propositions for the next election (whenever that might be called),
the mistake would be for them not to think properly about how the different measures must work together. The housing crisis is the product of many factors, many of them embedded over decades. The tests for any manifesto housing target is simple: show us the theory of change, or it is just a number.
[3]Centre for Policy Studies, The Future of the Green Belt, executive summary and pp. 54-62
[4] See pp. 5 and 57-61 of the CPS Report. The calculations in the text use the report's published 152,347 hectares and its stated assumptions of 35 dwellings per hectare and 60% site utilisation
[7] Of course, Lichfields produced its own capacity-based estimates of the Government’s station-based policy in the draft and final NPPFs, albeit observing the challenges around land availability and density assumptions.”
[8]CPRE, State of Brownfield 2025, pp. 2 and 15
