Category: Articles

Come Back to Croydon

In 2022, one of the first acts of incoming Mayor of Croydon, Jason Perry, was to revoke the London borough’s Suburban Design Guide, an award-winning bit of planning policy which is widely held to have resulted in an impressive increase in the number of completed homes but which, at the same time, caused consternation within the leafier parts of the borough. This wasn’t a surprise: scrapping the SPD had been a key tenet of his election manifesto.

“In order to address the imbalance in planning I will drop the dreaded design guide that is driving so much destruction of homes in our borough and revisit the local plan, so that I can remove the intensification zones designed to further destroy the character of Croydon. For too long planning has been density-led in Croydon and we need to get back to a design-led approach that respects the wonderful character found across the borough.”
Jason Perry, Election Manifesto, (no longer accessible online)

The success of Croydon’s SPD has been the subject of much subsequent analysis, which suggested that the document had a profound impact, resulting in nearly 2,000 homes being built on small sites during a three-year period from 2018 to 2021. Research outfit Centre for Cities covered this extensively, finding a huge uplift—and subsequent collapse—in the number of homes consented within minor planning applications (those delivering fewer than 10 homes) either side of the policy being in place. This conclusion was based on earlier research from the Greater London Authority which noted a leap in small site housing delivery in the period immediately after the document was in place.

Extract from the GLA’s Planning in London 2023 report, October 2023

“There was a particularly large increase in building on small developments in Croydon after it introduced supportive planning guidance, with net completions rising from 770 in 2012/13 to 2016/17 to 1,965 in 2017/18 to 2021/22 – almost three times as much as the next placed borough (Barnet, with 710 net completions).”
Housing in London, James Gleeson / GLA, October 2023

This figures are striking, but they’re not the whole story. The research considered only minor planning applications (comprising fewer than 10 homes) rather than the small site definition established in the London Plan that was published in late 2017 and adopted in 2021. For example, the Croydon SPD promoted the idea of aggregated intensification, focused in particular around three suburban neighbourhood, where rows of detached houses could be demolished and replaced with blocks of flats. This extract from the document shows one example of how this might be achieved:

A real-life illustration of this can be found in Purley, where three large houses were demolished to make way for 30 flats, resulting in a net gain of 27 homes. This application, any many like it, were not captured by the GLA’s research, even though it falls within the London Plan’s definition of a small site.

Google Streetview image from 2015, showing 4, 6 and 8 Russell Hill prior to demolition.

Google Streetview image from 2024, from approximate the same viewpoint, showing the completed development of 30 flats.

The application on Russell Hill was approved under delegated powers by Croydon’s planners in December 2017; nearly a year before the SPD was published in draft (and the same month that Sadiq Khan’s draft replacement London Plan was released).

The officer’s report accompanying the approval noted that “Representations have raised concern over the intensification of the site and overdevelopment. The site is in a suburban setting with a PTAL rating of 4 and as such the London Plan indicates that the density levels ranges of 200-350 habitable rooms per hectare (hr/ha) and the proposal would be in excess of this range at 364hr/ha. However, the London Plan further indicates that it is not appropriate to apply these ranges mechanistically, as the density ranges are broad, to enable account to be taken of other factors relevant to optimising potential, such as local context, design and transport capacity.“

The report makes interesting reading. Throughout, the case officer makes it clear that this type of site should be developed for high-density housing, and subjective reasons for refusal are quickly batted away. There is a strong presumption in favour of intensification, taking precedence over concerns about scale, overdevelopment and impact on neighbours.

When Khan’s first London Plan was published in draft at the end of 2017, policy H2 defined a small site has having both an area no greater than 0.25ha (2,500sqm) and yielding no more than 25 new homes. During the lengthy adoption process, this latter threshold was dropped, with only the plot size remaining. As part of this dedicated policy, H2 also established targets for each of the boroughs, setting out how many dwellings must be found on small sites.

At 15,110 homes, Croydon was given the highest absolute target of any of the London’s planning authorities, although this was slashed to 6,410 homes when the plan was adopted in 2021. Boroughs were encouraged to introduce dedicated policies to respond to this requirement: the Suburban Design Guide was the council’s way of rising to meet this challenge (our own version, for Lewisham, was adopted in 2021).

But if we look more closely at the numbers, we can see that Croydon’s small site approvals had begun climbing well before this, jumping from 307 homes approved in 2015 to 555 the following year, before rocketing to over 1,200 the year after. For comparison, Barnet’s figure for 2017—the next highest in London—was a mere 494 homes.

In fact, by the time that the SPD was published in draft, Croydon’s small site approvals had already peaked, dropping to 1,005 homes in 2018, and 898 in 2019, just as the policies of the suburban design guide kicked in. In line with every other borough, 2020 saw a catastrophic collapse in small site housing approvals as the pandemic took hold.

Given the length of time it takes to prepare, submit and determine a planning application (sixth months at best; more likely between nine months and a year) applicants must have been confident to make decisions to proceed several years ahead of the SPD being in place. What can have given them the confidence that taking this kind of planning risk, on such ambitious proposals, was worthwhile?

Using data extracted from the London Planning Datahub, the animation below shows the chronology of small site approvals across Croydon during the decade from 2015. Each dot on the map represents a planning application that was approved during each month, the size of the dot proportional to the net number of homes it contained. Orange ones represent those approved prior to the SPD being consulted, purple during the consultation period; and blue after its adoption.

With the exception of the Brighton Road cluster, the four areas identified for intensification within the SPD saw very little development during this period. “War on the suburbs” this was not.

If we look at the quarterly breakdown for 2015-2020, we can see that nearly 400 homes were approved (and subsequently built) in the final three months of 2017, almost a year before the SPD was published in draft for the first time. Approvals dropped by half in the new year, before jumping back up once the design guide broke cover.

Given the time required to assemble a compliant planning application, it’s likely that these schemes were being prepared at least six months to a year ahead of approval, meaning that applicants clearly had confidence that their development would be granted consent.

At the same time, Croydon Council was embarking on an ambitious, but ultimately ill-fated, programme of direct delivery through its housing vehicle Brick by Brick. In the second quarter the council approved 291 of its homes. But the contribution to the overall number of dwellings delivered after 2015 was modest.

By far the greatest criticism of the policy was from the southern wards; those largely Conservative-voting neighbourhoods on the edges of the green belt. Yet the distribution of new homes under this regime was broadly evenly split across Tory and Labour-voting areas, although it is true to say that there was a slight shift in the years immediately following the formal adoption of the SPD. Claims that this was the imposition of mass development on Conservative-voting wards by a Labour administration were unfounded – particularly when one considers that the denser northern (Labour-voting) wards cover around a third of the southern ones, whilst accommodating the same number of new homes; as well as the huge towers rising around East Croydon station, which in themselves added many thousands of new flats.

To be fair, what was happening at the time in some of Croydon’s suburbs was pretty wild. I understand why residents of Banstead Hill might have felt upset by the scale of redevelopment happening around them…at least until the developers started knocking on their doors, chequebooks in hand.

This aerial view of Banstead Road in Purley illustrates the scale of change taking place in suburban Croydon, where suburban houses are replaced by large apartment buildings.

What seems apparent from the data, however, is that the SPD was probably not the catalyst for the huge number of homes delivered across the borough in the decade following Labour taking control. Given that the peak year for approvals happened before the document was even published in draft, it’s impossible to ascribe the high volume of permissions to the implementation of the SPD itself. In fact, there was a distinct reduction in small site approvals post-2019 period; just as the document can into force.

I’m convinced that the reason so many homes were approved on small sites in Croydon was not the introduction of the Suburban Design Guide SPD, but rather a culture of permissiveness that stemmed from the Labour administration’s re-election in 2014—and that the SPD was actually a product of this permissiveness, rather than the catalyst for it.

So, what lessons can we draw from this conclusion? The introduction of dedicated policy guidance is appropriate if boroughs want to see an uptick in the number of homes delivered on small sites. But this should be seen as a means by which to improve the quality of schemes that come forward.

To enact real change—to make a meaningful dent in the shortage of homes that the capital, and the country, desperately needs—planning departments have to fundamentally rethink the way in which they consider intensification. Change is good; preservation is not always the correct option.

As the draft London Plan from December 2017 clearly stated, “the emphasis of decision-making should change from preserving what is there at the moment towards encouraging and facilitating the delivery of well-designed additional housing to meet London’s needs.”

This paragraph was removed from the document on its adoption in 2021. Perhaps it’s time to put it back?

Timeline of the Croydon Suburban Design Guide SPD

Nov 2017: Strategic Housing Land Availability Assessment indicates need for high small-site targets.
Feb – Mar 2018: Informal workshops, engagement with residents’ associations and the Planning Committee regarding suburban intensification.
Sept – Oct 2018: Public consultation held on the draft Suburban Design Guide
April 2019: Croydon’s Suburban Design Guide (SPD2) formally adopted
March 2021: London Plan formally adopted, reducing small site targets for Croydon
June 2022: Executive Mayor proposes revocation of the SDG to move away from density-driven targets
July 2022: Suburban Design Guide formally revoked following local opposition to densification

London Election Map

Frustrated by the slow rate at which the results from last week’s local elections were reported, and the poor quality of the way in which these were displayed online, I set about creating an online, interactive map showing every borough and every ward in London, with the results for each candidate harvested from the borough websites.

Quite a lot of this was done using Claude Code and Cowork, which sped up the extraction of the candidate vote counts from each of the boroughs, each of which decided to publish their data in a slightly different way (or not at all, in the case of Newham and Brent, who waited until Monday before they released the data).

Utilising a set of ward boundaries from Ordnance Survey, the map includes each result from every ward, the results for the council in which they’re located, the strength of the vote for the winning party and ranking of all candidates according to the number of votes and share of the total.

You can view the map online here.

Going Solo

In May this year, the government launched a consultation on the potential sub-division of planning application categories. Up to now, full planning applications fell into one of two classifications: minor, and major, with the threshold for those schemes falling into one or the other being the net addition of 10 homes. Any less than this, you’d be a minor application; any more, a major—regardless of whether this involved the creation of 10 homes or 1,000. This has consequences in the level of information required to validate the application and the time taken to determine it: minor applications should receive an outcome in no more than eight weeks from the point of validation, but for major schemes this rises to 13 (although very few applications are ever actually determined within this period).

The proposed bands retain the minor category at between 1 and 9 inclusive, but introduces a new medium definition of between 10 and 49 homes, with the major classification now including those schemes yielding more than 50 dwellings.

This makes a lot of sense, although perhaps it doesn’t go far enough. Applications for 50 homes should not be treated in the same way as those for 500, so there should be a further division of the major application type, probably around the 250-home mark.

But what about at the smaller scale? SME developers are likely to be the ones building the majority of homes within the minor category, but community builders, councils and housing associations also have a part to play. But what about those new homes for private clients; the boutique developers who have managed to acquire a tiny scrap of land in a fashionable area, or the families who want to carve out a sliver of their back garden in order to build a home for their kids? Single family homes are an opportunity for creativity and innovation, as well as being a vital tool in our attempt to intensify our towns and cities. Many well-known architects have cut their teeth on crafty infill projects that squeeze delightful and spacious homes on the most constrained urban sites.

He Lives in a House, a Very Big House

The National Planning Policy Framework (NPPF) has, for some time, included a clause which allows new homes to be built in the countryside provided that certain criteria are met. The most recent iteration of the NPPF encapsulates this objective in paragraph 84, which states that “…decisions should avoid the development of isolated homes in the countryside unless one or more of the following circumstances apply:”

e) the design is of exceptional quality, in that it:
i. is truly outstanding, reflecting the highest standards in architecture, and would help to raise standards of design more generally in rural areas; and
ii. would significantly enhance its immediate setting, and be sensitive to the defining characteristics of the local area.
Extract from paragraph 84, National Planning Policy Framework, December 2025

This is great for wealthy actors, hedge fund managers and oligarchs, but why should they have all the fun?

I think it’s time to introduce a similar policy to allow the development of single homes in urban areas, free from the many restrictive planning constraints which consistently make small-scale development costly and unpredictable. In keeping with the alliterative minor/medium/major naming convention, we need a new category: micro.

Micro Scope

Here’s my proposal:

Within urban and suburban areas, but outside conservation areas and not within the curtilage of a statutorily listed building, it should be possible to build a new house with automatic planning permission, provided that the external volume of the house fits within a set of maximum parameters determined by the outlook of neighbouring windows and private amenity space. Applications must demonstrate only how they respond to a) flood risk, b) highways and transport requirements, c) waste and recycling provision, and d) risk of contamination.

The dimensional parameters might include the following limitations:

The development must comprise a single dwellinghouse on a plot no larger than 200sqm;
The development volume must be no taller than the highest part of an immediate neighbour;
The development volume must be no closer to the highway than the principal elevation of an immediate neighbour;
The development volume must not extend beyond a 45% line drawn in a horizontal plane from the nearest jamb of any primary window serving any habitable room;
The new dwelling must be no closer to the site boundary than the neighbouring dwelling is;
The distance between the principal window serving a habitable room in the new development must be no closer than 16m from that of a principal window serving a habitable room within a neighbouring dwellinghouse.

This is how these policies might appear in three dimensions, using a street infill site type:

External appearance should be entirely down to the applicant: local character does not have to be “adhered to” or “respected”. I have no say in the colour or model of the car my neighbour chooses to park outside the front of their house; nor should they have a say in what I want my home to look like.

Building Regulations compliance is distinct from planning, so each home delivered this way will need to confirm to minimum levels of structural integrity and feature basic amenities, such as a toilet. It’s unlikely a mortgage lender is going to let self-builders create anything too outlandish.

Beyond these constraints, there should be no requirement to adhere to local planning policies, including space standards. You want to create a voluminous home consisting of a single room? No problem.

Not every site will allow a house to be built using these parameters, but then there’s always the normal planning route to fall back on.

Be Careful What You Wish For

Of course, there’s a risk that such a permissive policy could result in some dreadful interventions, but on the other hand, it could also unleash a wave of creative and extraordinary homes to meet a range of different needs. And can we really say that the planning policies of the last century has allowed only buildings of exceptional quality to be built? I don’t think so.

Existing planning policies are necessarily prescriptive to ensure that rogue developers don’t throw together exploitative shoeboxes that can’t meet the basic standards necessary for a comfortable existence, but these requirements also assume that everyone neatly slots into a conventional family unit. What about the extended multi-generational families; co-habiting friends happy to share a living room, but not a bathroom; or the retired couple who don’t need the space afforded by their large family home, but who have a lifetime of possessions that can’t be squeezed into a one-bed flat?

Giving almost complete flexibility to design an entirely bespoke houses—free from planning conventions and context—would allow families to create homes to suit their specific needs, and encourage architects to come up with innovative ways to craft extraordinary homes from the most unlikely of plots: witness the brilliance of the Japanese “jutaku” houses which exploit tiny urban sites to deliver some of the world’s most striking architecture. This would have the added benefit of adding new layers of interest and surprise into otherwise homogenous residential neighbourhoods.

There are risks, of course: many planning applications are already of extremely low quality, with scant attention paid to even the most basic tenets of good design. But this, I think, is a risk worth taking. Any damage would be limited, and local.

What policy levers might be pulled to achieve it? There are several options. The first might be the use of a National Development Management Policy (NDMP). An NDMP allows the government to establish a nationwide planning policy that overrides local plans, although planning permission would need to be sought in the usual way. A second route is through a new class of permitted development within the GDPO (General Permitted Development Order). This would allow homes to be built without the need to secure planning permission, instead seeking “prior approval” to ensure that (for example) the new home considers flood risk, contamination, impact on transport and highways, and the means through which to dispose of waste and recycling. Beyond that, anything goes. The latter option would be the most credible, but the most controversial too.

The dream of designing and building one’s own home should not be out of the reach of ordinary people, and at the same time we desperately need to fill the forgotten gaps and vacant plots of our cities with as many homes as we can. Single family houses provide an opportunity for innovation and the possibility of people owning a home that is designed specifically for their needs. We should let them get on with it.

Growing Pains

Inexplicably, allotments have been in the news this week. Responding to a parliamentary question raised by Kevin Hollingrake, the government confirmed that Angela Rayner had approved the disposal of eight allotments by the local authorities that own them.

“Is this government going to put the nail in the coffin of the joy of digging ground for potatoes on a cold, wet February Sunday afternoon?” pondered Jeremy Corbyn in a letter to the famously left-leaning Telegraph. Of course, Angela Rayner didn’t instruct these sales, but under the Allotments Act of 1925, has the final say on whether their disposal can take place.

The Ordnance Survey’s green space dataset counts around 150,000 allotments in the UK, although reports elsewhere suggest there’s more than double this figure. Clearly the loss of eight of these is a tiny fraction of the total amount, but could the release of a few more redundant growing spaces really be a practical means to build new homes – particularly in urban areas where the need is greatest?

London alone has around 770 allotments across 31 boroughs. Only the City of London and Kensington & Chelsea are without at least one allotment within their boundary.

London has around 770 allotments within 31 of the 33 local authorities. Only the City of London and Kensington & Chelsea don’t have a single one.

By and large, allotments don’t take up much space: just shy of 1,000 hectares in total. That’s about one third of one percent of London’s total area.

The largest single allotment in the city is the Fuel Land Allotments in Finchley, Barnet, which is around 11.5 hectares. It’s owned by The Finchley Charities, a small housing association that provides housing for elderly people. Its foreboding steel fence that faces the Great North Road conceals a vast landscape of neatly parallel plots extending into the distance.

Coming a close second is the Addiscombe, Woodside & Shirley Leisure Gardens, a privately-owned allotment society in Croydon with a total area of around 10.5 hectares.

The Fuel Land Allotments in Barnet are owned by The Finchley Charities.

Addiscombe, Woodside & Shirley Leisure Gardens in Croydon; the country’s oldest privately-owned allotment society.

These are far from the norm. The median size of London’s allotments is a modest three-quarters of a hectare. Barnet, home to London’s largest allotment, also has the most space dedicated to them, with just over 100 hectares of growing space.

Haringey has the largest proportion of its area occupied by allotments, but even then it’s only a modest 1.4%.

Their size pales into insignificance when compared to the city’s golf courses, however. The amount of space dedicated to hitting a little white ball around is the same as a mid-sized borough; together allotments occupy about the same area as Richmond Park.

Even if we were to build homes across every plot in the capital, how many homes might result? At a reasonably dense 50 dwellings per hectare, replacing peas with people would yield about 48,000 homes—just six months’ supply. This perhaps is a battle that even the developer-friendly Labour government would be unwise to fight.

Losing local authority land to housing isn’t the whole story, however. While allotments provide vital outside space for many people who have none at home, waiting lists in many parts of London are decades long, and because of this, favour long-term residents. There are many examples of owners objecting to development adjacent to their plots, in some of the best located and most appropriate locations for new homes. Anecdotally, many of those continuing to enjoy the benefit of allotments for many years no longer live in the same borough – or even the same city – as their discounted growing plot. One can imagine how a young person who is struggling to find a home might consider it unfair that publicly-owned but privately-managed space is rented at low cost to those who are no longer residents of the borough.

This week’s hoo-ha around the disposal of a few surplus allotments is likely to quickly fade. There are far easier—are considerably larger—locations for housing than allotments. The capital’s comedic carrots and amusing marrows are safe, for now.

London, Open

Five years ago, during the early months of the pandemic lockdown, I taught myself how to use the open-source mapping software QGIS, applying my new-found skills to locating and measuring the golf courses which pepper London’s suburbs. I’d suspected that the area of London occupied by golf was large, but I hadn’t anticipated it being as much as a mid-sized borough.

What I wanted to know was: given the huge area of land dedicated to this single use, how many homes might be built if we were to allocate some of this space for development.

In May this year, the Mayor of London published his long-awaited outline proposals for how the city might evolve over the coming years. “Towards a New London Plan” is his consultation on the the future of the London Plan, the spatial policy plan for London.

In my article I’d argued that, on the whole, golf courses fail to meet the criteria for Metropolitan Open Land; a specific planning designation within the London Plan, designed to protect open space from development, effectively benefiting from the same protections as green belt. A key criterium for MOL is that the open space in question must “serve either the whole or significant parts of London”. While it’s the case that some courses have public paths within them, woe betide anyone who fancies a Saturday afternoon picnic on the fairway of the ninth. As for biodiversity: well, golf courses are largely monocultural with sterile soil. Research from abroad has even suggested that living close to a golf course can give you Parkinson’s Disease.

It was a welcome surprise to see the Towards a New London Plan agreeing with my analysis. The chapter covering MOL included the following text (my emphasis):

2.11 Metropolitan Open Land

The Mayor will continue to give protection to MOL given its vital role for Londoners and providing a liveable city as London grows. However, some areas of MOL, such as certain golf courses are not accessible to the wider public and have limited biodiversity value. This undermines the purpose of the designation. These areas could be assessed to understand whether they should be released from MOL. They may be able to help to meet London’s housing and accessible open space provision (for example opening up strategic new open spaces accessible to Londoners alongside new homes). At the same time, they could improve biodiversity through landscape-led redevelopment. Clearly there are key issues to explore. For example, could golf courses with Sites of Importance for Nature Conservation (SINC) designations be released (with compensatory biodiversity uplift), and if so, in what circumstances.
Towards a New London Plan, Mayor of London, May 2025

Later in the document, the following paragraph expands on this point, suggesting that some courses could be released for genuine public benefit.

Given the challenging housing target, there may be some very specific circumstances where certain MOL, such as golf courses, could be considered for release for housing. These are often not publicly accessible and offer limited biodiversity value. They could also provide new accessible open spaces and parks alongside housing and other development.
Towards a New London Plan, Mayor of London, May 2025

In fact, since I first conducted my research, Ealing Council did exactly this, having converted Perivale Golf Course into Pear Tree Park.

40 of 93 of London’s golf courses fall within the designation of Metropolitan Open Land, totalling 1,578 hectares. In simple terms, that’s a little larger than the borough of Islington, which has a population of nearly 250,000 people.

Building at an approximate gross density of 50dph (an average for London, equivalent to compact Victorian terraced streets), building on all of the city’s MOL golf courses would yield some 79,000 homes. I don’t think London’s golfers would thank us for that. Besides which, many of the city’s courses are not in easy reach of public transport or high streets – two things new housing will likely need access to. So if we exclude those bits of the courses that are less convenient to get to, we arrive at around 625ha of land. That’s still 31,150 homes – not to be sniffed at, but it’s still only a third of London’s annual target. The graph below shows the accessible areas of those golf courses protected by Metropolitan Open Land, and within 800m of a station or high street (a useful policy definition set out in H2 of the London Plan).

Nevertheless, this is a number of homes worth having. Given that the total area of golf courses in London is well over four thousand hectares, losing just 17% of this for housing doesn’t seem too much of an ask.

The proposed changes to the London Plan are heading in the right direction.

Balancing the Books

Planning is all about compromises. It nigh on impossible for any scheme to be fully compliant with every policy: a presumption against height might limit the amount of affordable homes that can be provided; accepting a taller building might mean that the garden of a neighbouring house is overshadowed at certain times of the day.

The purpose of planning is to determine which of these criteria has the most importance – one person’s right to enjoy their garden might be diminished, but the lives of future occupants of the flats next door could be greatly enhanced. It’s the job of the planning system to distil these competing objectives into a binary outcome.

Often the responsibility for making such a call falls on the planning case officer, whose job it is to compare the application against the local, regional and national planning policies and decide whether it should be approved or refused. In particularly contentious cases—where there are a large number of objectors, for example, or for a large development—the decision will often be made by a planning committee consisting of elected politicians, many of whom discharge their responsibilities with diligence and care. Often, though, smaller applications are left to be determined by officers without the appropriate experience and confidence to reach a considered and reasonable conclusion.

The issue of balancing competing demands can be particularly challenging when considering applications for housing on smaller developments. Because of their size, such schemes are often handed over to more junior officers. They also tend to be more complex, often close to existing homes and squeezing the most out of fiddly sites. An aggrieved neighbour might be particularly vocal about losing an extra hour of sunlight on their prized carrots, or simply dislike the idea of an open plot of land they’ve looked across for years being occupied by a new home. How is a local authority officer supposed to determine which of the multitude of competing policies takes precedence? And how can they turn this information into a single binary decision?

I was intrigued by an article in The Economist newspaper (and an accompanying methodology) which described a system for weighing up competing priorities in a fair and transparent manner. A set of questions is presented, with a single pool of points that can be assigned to a positive or negative response to each. As each preference is made, the pool depletes; but expressing a stronger opinion for each challenge reduces the remaining points by a square factor. So, if you have a mild concern about the potential impact on neighbouring amenity of a new development, you can use a single point to express this. If you have a significant concern a second vote will cost you not two points, but four—two squared. Expressing a major concern would then diminish the available points by a further seven (16 less the nine you’ve already spent), and so on. You continue until there are no points remaining or it’s not possible to assign them to any further questions. Allocating no points to one question simply means that the assessor believes there the policy implication is neutral.

Using a standard set of topics which planning officers usually assess an application against it’s possible to construct a series of positive and negative questions against which an application is assessed. This allows the different aspects of a scheme to be weighed up against one another: a less-than-substantial harm to a conservation area might be mitigated by an outstanding design, for example.

The following application is a crude example of how such a system might work.

Available Points

Housing Delivery Test 2023

The 2023 Housing Delivery Test score for every English planning authority was published last week, and I’ve mapped the results to show those areas with the highest and lowest performance against their targets.

The southeast has, as expected, performed particularly poorly with several of the London boroughs falling below 50%. Of these, Lewisham is the worst, with a score of just 32% and a placement sixth from bottom in the national ranking. The highest score of all of the boroughs (with the exception of the City of London) is Croydon, with a score of 160%—although this is likely to diminish in the future now that its new local plan has been adopted, scrapping many of the progressive planning policies which enabled it to build more homes than any other borough bar Brent and Tower Hamlets.

A map showing the 2023 Housing Delivery Test score of every planning authority in England.

The highest score in the country was Richmondshire, which managed a stonking score of 6,121% which seems extraordinary until one realises that significant overperformance was likely when its absolute housing target was a pitiful 24 homes: just one for every 5,500 hectares of land. It actually built nearly 1,500 homes in the three-year period from 2020-2023; not to be sniffed at, but hardly an exceptional amount considering its size. Richmondshire is no more: in 2023 it was absorbed into a new Unitary Authority covering North Yorkshire.

Update 11 February 2025

Here’s a scatter graph showing the Housing Delivery Test scores for all of London’s planning authorities, including the 33 boroughs, the City of London, and the two Mayoral Development Corporations.

And here’s the same graph but including all of England’s planning authorities.

Get on board

Around England’s rural railway stations there is the capacity to build between 1m and 2m homes depending on where one sets an appropriate development density.

My own research has established this figure at around 1.2m homes, assuming development is constrained to just 10 minutes’ walk (approximately 800m) from a station, assuming a gross residential density of between 40 and 75 dwellings per hectare.

The eastern branch of the Thameslink railway franchise links London St Pancras (and beyond it, Brighton) to Cambridge, passing through Essex and the south Cambridgeshire countryside. The total journey time from Brighton to Cambridge takes just under two-and-a-half hours, with St Pancras the halfway point. Three quarters of an hour later the train arrives at the first of four rural stops: Ashwell & Morden, Meldreth, Shepreth and Foxton. In the middle of these sits the town of Royston (see map, below).

Foxton is just 12 minutes from Cambridge station; Ashwell & Morden 47 minutes from St Pancras and an hour from Kings Cross. These stations are extremely well connected, but serve a very low number of homes within the surrounding area.

Map showing potential locations for new towns along the London to Cambridge Thameslink route. Contains OS data © Crown copyright 2024.

In the year ending March 2023, Ashwell & Morden station had 140,000 passenger movements (entrances and exits) with just under a third of these in the direction of Cambridge.

Connections to the road network are also strong, with Meldreth, Shepreth and Foxton all located on, or close to, the A10, which links London to Cambridge. Ashwell & Morden is immediately adjacent to the A505, linking the M1, A1(M) and the M11.

Currently, Thameslink trains call at Ashwell & Morden and Royston, but Meldreth, Shepreth and Foxton are served by the Great Northern franchise, linking Cambridge to Kings Cross. The track is the same, and there is no reason why these two could not be consolidated into a single service. The proposed route for East West Rail also joins this line north of Foxton, with the potential to link these towns with Bedford and Oxford. It is also close to the new Biomedical Campus at Cambridge South, and the line passes through the new station being built there.

Extract from East West Rail preferred route map showing the potential connection to Foxton on the southwest branch.

According to my research there is the potential for approximately 10,000 homes in the undeveloped land around Ashwell & Morden station alone. It is unencumbered by green belt designation nor is it covered by ecological land protections. The developable area around the station totals approximately 170ha within an 800m radius, and with an average density of just 60dph this would deliver more than 10,000 homes.

Google Satellite image showing potential development area around Ashwell & Morden station (white) and freeholds (yellow).

There is further potential for urban extensions to other villages along the line. Shepreth could accommodate 7,200 homes through a northern expansion of the settlement (the station itself separates the town from the countryside). 2,800 homes could be built through infill and enlargement of Meldreth.

Foxton straddles the boundary of Cambridge’s southern green belt yet has the capacity for 6,800 homes around a station that is largely surrounded by open fields. Together, the areas around these stations could accommodate at least 27,000 homes—considerably more, if the radius was expanded slightly further into the countryside.

The proposition becomes more attractive when one starts to think of these stations not as individual settlements but as a single entity. The potential journey time from Ashwell & Morden to Foxton is no longer than 16 minutes, meaning that a series of smaller towns, each with a population of some 40 to 50 thousand, could support significant new social infrastructure: several new primary schools in each location, with a new secondary school at Ashwell & Morden, and perhaps a second a Shepreth.

A new Development Corporation, structured along the railway line rather than around a single settlement, could coordinate physical and social infrastructure in the appropriate locations, and produce a masterplan and design codes to lock in placemaking and quality objectives. The Corporation would adopt planning powers to ensure that delivery and quality objectives are met. With the exception of a small part around Ashwell & Morden, all of these proposed settlement expansions fall within South Cambridgeshire District, simplifying planning powers.

The land either side of the track is mostly owned by Network Rail. With some creative thinking this could be exploited to provide cycleways linking the expanded settlements, supporting a transport modal shift away from car dependency.

Based on the assumptions set out above, the approximate capacity of each new or expanded settlement would be as follows:

Station	Developable Area	Density at 60dph
Ashwell & Morden	170ha	10,200 homes
Meldreth	47ha	2,800 homes
Shepreth	120ha	7,200 homes
Foxton	113ha	6,800 homes
Total	450ha	27,000 homes

There are many other examples of such stations along the railway lines that extend out from our cities where this approach to Development Corporations could be adopted too.

Update March 2025

On 11 March 2025, the government published its long-awaited Planning and Infrastructure Bill. Included within the bill is an important amendment to the scope of Development Corporations, under paragraph 79, which states:

“A single development corporation may be established for the purposes of the development of more than one new town in England if the Secretary of State considers that having a single development corporation would facilitate efficient development.”
Paragraph 79(3), Planning and Infrastructure Bill, March 2025

Previously, the area of individual Development Corporations could only be contiguous. This amendment enables exactly the configuration that I am advocating for above, clearing the way for a single administrative body to cover several new settlements even where these do not share a common boundary.

Grey belt identification using Bluesky

For the last few months I’ve been attempting to map every bit of “grey belt” land within London’s green belt. As there’s been some uncertainty over exactly what qualifies as grey belt, I’ve adopted a cautious definition, including areas which appear to have been previously developed, including car parks and yards, earthworks and hard standing.

This builds on work I’ve previously undertaken which attempts to use AI to locate small site development opportunities in London’ssuburban areas. You can read more about this here.

I used our existing learning model (based on the InceptionV3 convoluted neural network) to assess an area of green belt using higher-resolution aerial photography provided to me by aerial mapping specialist Bluesky with the purpose of categorising the types of land cover. Our learning model was trained on aerial photography / satellite imagery from Google.

The following describes the methodology followed and some of the initial findings.

The area of sample data provided by Bluesky is contained within the white rectangle shown within the following image. The pink masked area is not within metropolitan green belt; the remainder is. The total area of the sample data provided by Bluesky covers an area of approximately 390ha, with just over three quarters of this within the metropolitan green belt.

Within this area I created a grid of points at 25m centres (i.e. a 625sqm tile). The image to the right shows the distribution of these within the examination area.

Using our existing learning model (based on Google satellite data), I assessed the ground conditions at each data point. The image on the right shows how the AI learning model has categorised the landscape features at each point. The pattern of ground cover can be seen here, with dark green representing woodland and tree cover, blue corresponding with areas of water (although there are some limitations to this, described below).

One of the limitations of this is that using the higher-resolution Bluesky photography some tiles contain multiple “types” and the AI struggles to distinguish between, say, a domestic garden and a residential house. Because of the limited resolution of the Google imagery this was not a significant issue, but at higher resolutions the data points can miss key features, such as entire houses, classifying the gardens to the front and rear, but not the structures themselves.

To address this, I introduced a data point grid with double the resolution, with points at 12.5m intervals. The image here shows the distribution of these on the same map. Given that the 25m grid across the entirety of London’s metropolitan green belt results in more than 8m data points, adopting a 12.5m grid will result in four times this amount.

Running the AI again using data with the Bluesky imagery results in this image, which includes the non-green belt area in the south west of the map, (which was included in error, but demonstrates the efficacy of the process). The patterns of land cover are clearly visible at this scale.

The image below shows a small area of the sample data. Moving the slider across the page reveals the AI predictions for each point at 12.5m intervals.

Golf courses are generally identified correctly, likely due to the presence of sand bunkers and the curved shapes of the fairways visible in the image, although there are some odd predictions for “water” which need further investigation.

Water is being identified correctly, as this lake within Aldwickbury Park Golf Club shows. There are some anomalous categorisations occurring around the periphery of the lake, likely due to surface markings fooling the AI into thinking these are sports pitches, as the pale line around the edge resembles the white lines of a football pitch.

As with the golf course, crops and woodland which appear as dark areas of consistent colour are incorrectly being identified as water. An improved learning model will be needed to account for this, as with this resolution these types should be distinguishable from bodies of water.

Car parks appear to be identified correctly, although due to the resolution of the image some areas within the car parking areas are being incorrectly identified. In some cases the markings on roads, or areas where no cars are parked, are fooling the AI into thinking that these are hard courts. The presence of parked cars is generally necessary for the AI to correctly identify land used for this purpose.

Domestic gardens need to be included within the “allotments and garden centres” type. Previously, most gardens fell within the “buildings – domestic” type as the 25m tile would generally include both. At 12.5m, the two are usually found in separate image tiles and therefore correctly identified.

For an unknown reason, some crops appear to be categorised as golf courses. Unclear as to why. Further investigation required. It is likely that adding these areas to the test and training data will teach the AI to distinguish between these with a greater degree of accuracy.

Fallow fields appear to be correctly identified, although the sample data from Bluesky does not include and “earthwork” sites so the accuracy of this prediction cannot be assessed. Note some anomalous identifications as “water” to the right of this image. Futher training of the learning model using a larger sample of the Bluesky data should assist in correcting these.

This exploratory exercise has demonstrated that higher-resolution mapping data, with imagery taken within a narrower seasonal timeframe, results in more accurate predictions, although to take advantage of this, a significant increase in data points will be required. The length of time required to generate the imagery and process the data will be significant and demand an increase in processing power.

Cornering the Market

At the weekend MHCLG published a draft working paper setting out plans for a “brownfield passport”, with the intention of finding ways to make the intensification of urban areas easier.

This is long overdue. The planning system is disproportionately complex for small-scale development, and this is one of the primary reasons why the country has witnessed a collapse in the SME developer market in recent decades. In London, Barratt Homes currently builds one in ten of the city’s new homes—this is not a healthy state of affairs. I have written elsewhere about the need to provide greater certainty for small developers who are less financially resilient than their corporate counterparts.

Importantly, the document recognises the woefully low density of many of our towns and cities:

Given our relatively low densities, there is scope in many areas for increases. While such increases should take account of local character, existing character should not be used to block sensible changes which make the most of an area’s potential, and which can create sustainable, well-designed and productive places to live and work.

This is an encouraging acknowledgement, as for too long, “character” has been used as a means to refuse new homes. The draft NPPF which was published in June helpfully removes paragraph 130, which requires LPAs to refuse applications which are “wholly out of character with the existing area”. Combined with a national policy which sets out parameters for intensification, this could be powerful indeed.

Policy could, for example, say that development should be of at least four storeys fronting principal streets in settlements which have a high level of accessibility, and/or set acceptable density ranges that allow for suitable forms of intensification.

We know from previous experience that liberal design codes can be difficult for existing communities to accept. An example of this is in Croydon, where a suburban intensification policy was so hated that a new mayor was voted in just to scrap it, despite delivering nearly 2,000 new homes in a three-year period. The reason for the disgust at the Croydon policy was because it concentrated new homes in a relatively small area, and the pace of change was rapid.

How might a similarly ambitious policy, set at national rather than local level, provide certainty to small developers whilst delivering high-density development in sustainable locations? The following paragraph from the document is intriguing:

[W]e are keen to explore how this might be done – for example, whether densification in some areas should focus on corner plots and those adjoining them rather than whole streets, or linking densification opportunities to accessibility.

This makes a lot of sense. Corner developments tend to be better suited to intensification as they are less likely to overshadow or overlook adjoining gardens. They can act as wayfinding devices at key road intersections, helping bring clarity and legibility to suburban areas. And they can help provide non-residential uses such as small shops at ground floor level.

We included a specific design code for corner plots in our Small Sites SPD for Lewisham Council.

In this example a single corner dwelling is replaced with a small block of flats—four storeys in height, as it happens, as the document suggests—and provides six new homes. As Lewisham was keen not to see a net loss of family-size dwellings, this includes a pair of duplex flats with rear gardens.

Extract from the Lewisham Small Sites Supplementary Planning Document, adopted in 2021, by Ash Sakula Architects and RCKa

In outlying areas there are even more opportunities for intensification. This lovely scheme by architects Harp & Harp replaces a sprawling family home on a corner plot with no less than seven family homes. Repeating this on each of the remaining three corners of the block would result in an increase of some 24 homes; that’s almost doubling the density, with no discernible detriment to character (I’d argue a significant improvement).

Design for a new development of seven family homes in Croydon by architects Harp & Harp

Here’s another striking example by OB Architects that we included as a case study in the Lewisham SPD. It replaced a single family house on a corner plot close to South Croydon station with an attractive four-storey apartment block containing eight new homes. We should be doing this on every suburban corner.

A land-hungry detached house occupies a corner plot close to South Croydon Station.

OB Architect’s Greyfort House replaces this property with eight new homes.

So, how many of these opportunities exist? With support from MHCLG’s PropTech Round 4 funding, I’ve been developing an AI tool that can identify and characterise different types of small development plot. I’ll be writing about this more in due course, but one of the powerful features of this is that it can find combinations of different site types. For each site it assesses, the AI determines a “ranking” according to the confidence in its prediction. So, it may think that a site has a 60% chance of being an “infill” type, but also a 30% chance of being a “semi-detached” type (it could be both, of course), reducing the percentage confidence each time until it has accumulated five predictions.

Searching for sites which fit both the “corner” (ranking first or second) and “detached” categories (again, ranked first or second), for instance, reveals approximately 320 properties in Croydon which could be intensified in this way—that’s a potential net increase of around 2,000 homes. Croydon’s small sites target in the current London Plan is 6,410 homes, so this would deliver nearly a third of this total alone—as well as concentrating low-rise intensification in those areas best suited to accommodate it.

There are fewer of these sites—around 150—in Lewisham, partly because it’s a smaller borough, but also lacks the large areas of suburban development. Even so, there’s plenty of potential: take the example below, a stone’s throw from Grove Park station. My suburban intensification study suggests that this neighbourhood has a prevailing density of some 16 dwellings per hectare (dph). This is a dreadful figure for somewhere with such good access to public transport and a Public Transport Accessibility Level of 4.

This corner plot in Lewisham is ripe for intensification, with a large single-family home and deep garden, close to a suburban station, this is exactly where we should be enabling a significant uplift in density.

A development here could increase the number of homes on the site from one to fifteen or so, but as we know, local politics means that “character” inevitably takes precedence over additional homes. We cannot, therefore, rely on local policy to take the bold steps needed to intensify our suburbs: the only solution is to introduce a strategic policy that removes the potential for political interference and starts to deliver the homes we desperately need.

A national brownfield passport would be a great place to start.

Right on Target

When the new Labour government’s proposed housing targets were published in July there was some surprise that many planning authorities, particularly urban ones, had seen a significant reduction in the number of homes they were being expected to deliver, and that there was a discernible shift away from the south-east of England to the north-west.

The map below shows the percentage change for each planning authority in England. The authority with the largest reduction is Tower Hamlets, going from 5,190 homes to 2,177 using the proposed Standard Method; a reduction of 58%.

At the other end of the scale, Redcar & Cleveland’s increase of over 1,300% seems large until you realise this is a jump from just 45 homes per annum to 642. That’s just 30% of Tower Hamlet’s new total, despite Redcar & Cleveland having an area more than 11 times greater than the east London borough.

Maybe, then, the changes in numbers expressed as a percentage of previous calculations are not that useful? A different method might be to consider the total number of homes in terms of density. The map below shows what this looks like.

As expected, those areas that you’d assume would have the highest housing demands (in and around England’s major cities) see the greatest number of homes her hectare. Top of the list is London Borough of Kensington & Chelsea, with 3.45 homes per hectare required to meet the new targets. This is an outlier: next on the list is neighbouring City of Westminster, with a target of 1.72 homes per hectare. The top 25 authorities by density are all within London, and the first ranked authority outside the south-east is the City of Bristol, with a target of just 0.28 homes per hectare.

These housing targets have been greeted with considerable uproar, particularly from the anti-development brigade which spuriously claims that this amounts to nothing less than “concreting over the countryside”.

As an example, in an open letter to Angela Rayner published shortly after the targets were revealed, East Hampshire District Council raised the following concerns:

“East Hampshire’s housing targets, as determined by the standard methodology, show we must identify sites for almost 11,000 homes by 2040.

“But those calculations take no account of the fact that 57 per cent of the district is inside the South Downs National Park, an area where development is restricted.

“That leaves the remaining 43 per cent of the district to take the lion’s share of development.

“Inevitably that will put pressure on our highly-prized countryside and our rural towns and villages, which have already seen so much change over the past few years.

“This is not sustainable development. It is the damage done by a blunt instrument – a planning policy that takes no view of unique local factors.”

Whilst it’s true that a large proportion of East Hampshire is located within a national park, to suggest that these figures are unachievable is ludicrous. Even building at a relatively low density of 40 homes per hectare, its five-year target of 5,370 would require just a quarter of one percent of its total area to deliver—even if it were to be accommodated in a single location. The reality is, in fact, that most of this will be achieved through the intensification and expansion of existing settlements.

Here’s a map of the district showing the land required to deliver 5,370 homes at 40 dwellings per hectare.

Similar complaints have been heard in Cherwell, which sits to the north of Oxford. Cherwell is similar in size to East Hampshire, albeit without a chunk of national park, so it would take rather a lot of effort to “carpet over” it, as Conservative Councillor Eddie Reeves suggests will happen under the new targets.

To be fair, Cherwell does accommodate some of Oxford’s green belt in the southern region, but other than that is pretty much free from planning constraints. Again, even with a 55% uplift in housing numbers it could find all 5,475 homes using just 0.23% of its land. Suggestions that significant parts of the countryside will be lost to development are very wide of the mark.

Down in Kent, Conservative councillors of Tonbridge & Malling got particularly excited about their new targets, being required to find space for 5,285 homes over the next five years—an increase of just under a third. A large part of the district is within London’s green belt, with some of this coinciding with the Kent Downs AONB. But, as the green belt is no longer a constraint on development, there’s more than enough space for these homes. Hardly the “naked opportunism” that the local Tory Councillor Matt Boughton seems to think it is.

West Berkshire is on a war footing after council leaders described the new targets as a “bombshell“. The CPRE, which can always be relied on to keep a level head in the time of crisis, backed these protestations, claiming that the authority’s new targets of 5,285 homes over five years are “excessive and unsustainable”.

A large chunk of West Berkshire is, to be fair, covered by the North Wessex Downs AONB, and this will certainly prevent housebuilding at scale. But the remaining areas to the southeast, around Newbury and Thatcham, have no such constraints, so there’s no reason why these homes cannot easily be accommodated, particularly as most will be located within the boundaries of existing towns.

It’s tempting to snigger at the hysterical language employed by the many politicians who oppose the building of homes in their local areas, but I genuinely believe that they have a very limited understanding of the scales of land involved. It’s easy to look out of your window at miles of fields and worry that a housebuilder is going to come and plonk a bunch of executive homes in the way of your view, and in some cases that is bound to happen.

But as for “concreting over the countryside”, well, nothing could be further from the truth. The countryside is huge.

We know from previous research that the public has little understanding of how much of the country is built on. A survey in 2018 showed that people tend to think that nearly half of it is. But England’s urban areas take up about 14% of its total area—and that includes urban parks and gardens. The land covered by buildings themselves is even smaller than this. That leaves around 11.5 million hectares of open countryside, most of which is for agriculture.

Now imagine that the government’s 1.5m home target was being located in an entirely new settlement somewhere in the countryside. At a modest density of 40 homes per hectare it would take up an area of 37,500 hectares. That’s just over 2% of the size of England’s green belt. This is what a circle with an area of 37,500 hectares overlaid on a map of England looks like:

Whichever way you look at it, the countryside is safe.

Maps of Every Planning Authority in England

You can view a map of every planning authority in England by choosing from the dropdown list below.

Comment or Object? Bias Against Development in London’s Planning Portals

A few years ago I took it upon myself to submit a supporting comment on Transport for London’s plans to build new homes on the car park at Cockfosters station. Although it’s just across the borough boundary in Enfield, it’s only a few hundred metres from where I live in New Barnet, and the application was worthy of support. Then Chipping Barnet MP, Theresa Villiers, was running an active campaign against the plans, which was as a good a reason as any to add my name to the list of those in favour.

I’d not thought much about it before then, but having found the application reference (handily provided on one of the many anti-development leaflets that dropped through my door), and registered on Enfield’s website, I found myself presented with a long list of checkboxes setting out the reasons why I might want to object to the proposals. This included a veritable NIMBingo card of objections; yet missing was anything that could be considered a positive justification for the development.

Many weren’t even valid planning objections at all: included on the list was “general dislike of proposal”, but nothing about the desperate need for new homes. “Noise nuisance” made an appearance, but missing was a checkbox acknowledging a positive contribution to the local area.

Typical page from a London borough planning portal, showing an entire list of reasons to object to a planning application – but none to support.

If you found yourself wanting to object to a planning application but you weren’t sure why, this provided you with a perfect “to do” list of reasons. It’s difficult to imagine anyone sufficiently enraged by the idea of new housing that they’d be selective about what they were ticking and why. “I just don’t like the look of it, but yeah – the land is probably contaminated too!”

To Enfield’s credit, when I raised this with the head of planning, he acknowledged that this list was unnecessarily biased, and it was duly revised to include an equal balance of positive and negative sentiments. Alongside “development too high” appeared “improves the quality of the area”; “loss of parking” was countered by “makes sustainable use of land”, and so on. Simply ticking every box from top to bottom was now counter-productive: those wishing to object, or support, an application were now forced to carefully consider each option.

This got me thinking about the inherent bias in the planning system in favour of those railing against development. I spend a lot of my life raking through planning applications and am familiar with many of the terms used, but for someone who’s engaging with the system for the first time this can be an intimidating and confusing experience.

You just need to look at the front page of many council homepages to see this bias at first hand. The screen grab below is from the front page of Barking & Dagenham’s website (I’m not picking on LBBD here: it just happens to come first in the alphabet), and demonstrates a tacit assumption that nobody would, in their right mind, actually write in favour of development:

Clicking through to the planning portal itself, it gets worse.

The impenetrable nature of Idox is for another day, but even once you’ve managed to locate the reference of the application you want to comment on, you are required to complete a long and complicated registration form to do so.

I suppose that, unlike me, most people will be commenting mostly on applications which immediately affect them and will only every need to go through this procedure once, but you can see how someone without much time on their hands, who moves home regularly as a consequence of precarious circumstances, or is unfamiliar with engaging with the authorities, might be put off by this step.

I’m in two minds about the need to register to make a comment: on one hand I can see how it might limit spurious or trivial comments that take up officer time; on the other, I can also see that it might be off-putting for those more likely to support new development (older, established residents are, I suspect, more likely to have the capacity to spend time grappling with user-unfriendly web portals). But, on balance, I’m not sure that requiring commenters to register is a good thing.

Some planning authorities—my home borough of Barnet included—have removed this list entirely, so it’s up to respondents to decide for themselves whether they believe a particular scheme meets local planning policy. The bias against supporters is apparent elsewhere on the site, however.

It’s well known that public consultation is a bit of a nonsense anyway as every application needs to be assessed on its own merits, and officers are perfectly capable of determining whether a proposal for development is broadly compliant with policy (and if the planning committee disagrees, the Planning Inspectorate certainly is). The only purpose served by the commenting process is to apply pressure on elected officials to resist development.

Islington Council’s website is entirely neutral in its language, offering visitors the opportunity simply to “view or comment” planning applications; although the ability to find anything without the specific planning reference is impossible. Not to have an interactive web map displaying all current planning applications cannot surely be acceptable in a planning website that was only updated this year.

Richmond-upon-Thames’ website is marginally better: searching for applications is painful (another borough without an interactive map), nowhere does the site refer to “objecting”, and there’s no need to register in advance. Sensibly, personal details are limited to a name and email address.

All in all, the general state of London’s planning portal is woeful. A lack of online mapping, anachronistic interfaces and dysfunctional search facilities abound. Given the importance of housing in London, surely we deserve better than this? It’s no wonder people, including those for and against new development, feel disengaged from the planning process when it’s so hard to register an opinion. Perhaps the new government might want to invest in a unified platform, provided freely to local planning authorities, to speed up the planning process.

As it happens, despite Theresa Villiers’ intervention and a vigorous local campaign, the Cockfosters application was narrowly approved by Enfield’s planning committee—albeit in a meeting that went on into the early hours. The application was later stalled by an intervention from then Minister of Transport Grant Shapps, and when this was overturned, by issues around financial viability. In total, Enfield received 2,852 formal objections, with just 15 in support. That the committee voted in favour of the scheme is a credit to elected members. But it does suggest, at least to me, that the system is rigged and it’s time to do something about it.

If you’re interested in knowing more about the planning system works, and how you can help support planning applications for new homes in your area (or anywhere else, for that matter), I wrote a handy guide which you can download here.

Towards a Suburban Renaissance

“Boroughs should…recognise in their Development Plans and planning decisions that local character evolves over time and will need to change in appropriate locations to accommodate additional housing provision and increases in residential density through small housing developments.”
Draft London Plan, December 2017

I live in suburban north London, in a neighbourhood which sprang from almost nothing in the late nineteenth century with the arrival of the railway. New Barnet station is at the end of my road and my house was built on land previously bought by the railway company: we still have the original deed of transfer from 1899 which details the sale of our plot from the railway company to the developer who bought it and built the home in which I now live.

New Barnet in 1897. The railway arrived in 1850 when farmland was acquired by the Great Northern Company to enable the construction of the route, and then the land around it was sold to the British Land Company (image from National Library of Scotland).

The streets here are largely lined with Victorian terraces, with some grander villas dotted around on larger plots. Sprinkled among these are houses and flats built in the later years as a result of incremental intensification, some on the former gardens of the bigger homes, others on the site of houses destroyed by bombs in the Second World War. : There’s a handful of more recent interventions: nearby, a backland site turned into eight contemporary houses which, of course, Theresa Villiers—our local MP (for the time being)—objected to, and so on.

A typical suburban neighbourhood with deep rear gardens and lots of parking. Most of this area is less than 10 minutes’ walk from New Barnet Station (Google Maps).

Let’s Take a Ride…

Although we’re in Zone 5, trains run into town in less than half an hour; to Moorgate via Finsbury Park and Old Street, and recently our line has been connected to Thameslink, with peak-hour trains connecting commuters with the Elizabeth Line at Farringdon and on to southeast London. It’s a convenient place to live.

But, having said all of that, New Barnet—and other parts of suburban London just like it—simply aren’t that dense. Pockets of new development have been snuck into redundant land, former garages and car parks, derelict pubs and disused warehouses, and even so, the most recent Census data from 2021 tells us that the density of this part of north London is just 18 dwellings per hectare (dph). Maida Vale, as a comparison, has a density of more than five times that. And Maida Vale is hardly somewhere you could describe as an unpleasant place to live.

Whichever way you look at it, suburban London can clearly do more to meet the city’s housing needs.

The shortage of housing in London is at crisis levels and manifests itself in many ways. Young people have been particularly badly hit, and the consequences for our economy and society are dire. In Hackney, schools are closing because young couples are unable to afford to start families. Homelessness is at record levels, with one in ten children in parts of London classed as effectively homeless. Median house prices in the capital are now 14 times average incomes while wages have stagnated. While this cannot be entirely blamed on our inability to build enough homes, it certainly plays a very large part.

Land in London is precious, yet the suburbs have a hegemony over it. Those lucky enough to own a house in the suburbs and, in particular, those living close to public transport, surely have a moral duty to allow more housing to be built around them so that others can benefit from convenient access to all of the amenities that the city has to offer.

So where might these houses go? Do we have the space? And how can we encourage intensification to happen?

Diminishing Ambitions

The current Mayor of London’s strategic plan for the city, the “London Plan”, was finally adopted in 2021 it set ambitious targets for new homes across the city, compelling each of the planning authorities to meet specific annual housing targets both, with a proportion of these to be delivered on small sites, that is any plot with an area of less than 0.25ha (about a third of a standard football pitch).

An early consultation version of the Plan was accompanied by a series of policies which provided a framework for intensification, clearly stating that boroughs needed to accept that “local character evolves over time” and that it would “need to change in appropriate locations to accommodate additional housing provision”.

I’ve written elsewhere about the push-back from many of the outer-London boroughs to this policy which resulted in the final version eviscerating the small sites targets, and Croydon’s progressive attempts to densify suburban areas that were unceremoniously chucked out by an incoming NIMBY mayor. But even in the short period of time that Croydon’s policy was in place, it resulted in a remarkable outcome, delivering around 2,000 homes within developments of fewer than 10 homes, with house prices and rents levelling off as a result.

So, what if the lessons from Croydon could be repeated across the rest of suburban London? That’s what I’ve set out to establish.

It’s about to get a bit geeky from here on in.

Cum On Feel The (Voro)nois

Using a combination of data from Ordnance Survey and Greater London Authority, I assembled a map of London and marked on it every station in the city – both Underground and mainline stations. Many stations are within 800m of each other, so I created Voronoi polygons to establish the closest station to every area in London.

Around each station I created an 800m diameter circle, which equates roughly to a ten-minute walking distance. By combining the two geometries I established the closest station to every area in London that’s no more than ten minutes’ walk away.

Clearly this approach doesn’t take into account the various constraints on potential development, including lots of areas which would, of course, be impossible to build on. The Thames, for example, but also areas of protected land such as Strategic Industrial Land (or “SIL”) Locally-Significant Industrial Sites (“LSIS”), green belt and Metropolitan Open Land. There’s a discussion to be had about whether protecting any space close to stations is sensible, and whether golf courses and industrial land might be put to better use. But for the purposes of this exercise, I’ve excluded them; together with parks, gardens, sports pitches and any other type of open space. Given my focus on suburban areas, I’ve also excluded the “Central Activities Zone”, which covers central London. Further refinements exclude a buffer either side of national and regional roads, and existing railways.

The resulting map of London looks something like this:

So now that we have a map showing all of the potential areas that might be intensified around London’s stations, we need to introduce some data which tells us more about the neighbourhoods around them.

Census Sensibility

The 2021 Census provides a huge set of data broken down into geographic zones that enables us, with a bit of mapping jiggery-pokery, to intersect them with our areas of interest.

Using Census data broken down by Medium Super Output Area (MSOA) I divided the mapping areas by the equivalent polygon areas. First, though, I ran a series of mapping exercises to establish some additional figures for each of these regions: for example, using Ordnance Survey Zoomstack data to measure the approximate coverage of buildings for each MSOA polygon. Bringing the two together enabled me to examine each of these areas in more detail. Here’s an example: MSOA ref. E02000028 which is located immediately to the west of New Barnet station.

Measurements taken from GIS tell us that MSOA E02000028 has a total area of 105.61 hectares, and the census data tells us that this contains 2,783 homes (45% of which are detached or semi-detached houses) – an equivalent density of about 26 dwellings per hectare. The footprint of all the buildings is about 18.7% of the MSOA, which makes sense given the large rear gardens, even though there are no large areas of open space within it. The census also tells us that, with a total population density of 62 people per hectare, the occupancy level is only 2.35 people per dwelling…which is surprising given the number of very large houses found here (the highest dwelling occupancies in London tend to be in the East End, with parts of Bethnal Green exceeding eight people per home).

You can see the hatched areas overlaid on the image above, which represent the different Voronoi polygons described early. To the bottom right of the image you’ll find Oakleigh Park station, and this MSOA is divided into three sub-areas, each part closest to a different station: in addition to New Barnet and Oakleigh Park, the north-west corner is within 800m of High Barnet Underground Station.

26 dwellings per hectare is pretty low, although not untypical of suburban London. A modest increase over this area could result in a significant number of new homes – let’s imagine for a moment that this is increased by just 25% (hardly a transformative figure). Yet, even at these modest numbers this results in 686 additional homes – an uplift in density from 26 to 32 dwellings per hectare.

Even 32 dwellings per hectare is pretty modest when compared to other parts of London. MSOA E02000589 covers the area around High Street Kensington, topping out at 137 dwellings per hectare (dph). This is probably a bit much for Zone 5, but Herne Hill (MSOA E02000642) achieves a density of 40 dph and can hardly be considered overcrowded.

With all of this in mind, I’ve established a few rules to apply to my data to try and estimate what a modest uplift in density might achieve. Arguably, nowhere in London that’s within 800m of a station should have a density of less than 40dph, so I’ve set that as a minimum. And, although some parts of the capital exceed this, I suggest that the increase in density should not push an area beyond 100dph. Within these thresholds, I’ve set a few additional rules: where detached and semi-detached houses form more than 40% of the total dwellings, I’ve set the potential density increase at 50%; where they’re less than 10% of the total housing stock, it’s 10%. For everything else I’ve assumed a 25% increase.

I’ve made a further adjustment where buildings cover less than 25% of the available land, adding a compound increase of 40% to this figure. The resultant algorithm is something like this (where “familyHouses” means a semi-detached or detached dwelling):

# First, calculate the initial uplift in density based on the proportion of "family homes"
IF familyHouses > 40% THEN newDensity = existingDensity x 1.5
ELSE IF familyHouses < 10% THEN newDensity = existingDensity x 1.1
ELSE newDensity = existingDensity x 1.25
# Then add a compound density based on the total percentage coverage (footprint) of buildings over the MSOA area
	
IF coverage < 20% THEN newDensity = newDensity x 1.4
# Finally, if the new density exceeds 100 dph, cap the increase to this level (this means that any areas that already exceed 100dph see no increase)
IF existingDensity > 100 THEN newDensity = existingDensity
ELSE IF newDensity > 100 THEN newDensity = 100
ELSE IF newDensity < 40 THEN newDensity = 40

Applied across the entire city, this results in a net increase of some 900,000 homes, with each of the boroughs seeing the following uplift:

Borough	Net New Homes
Barking & Dagenham	15,070
Barnet	54,129
Bexley	37,985
Brent	32,719
Bromley	68,426
Camden	8,625
Croydon	67,165
Ealing	34,616
Enfield	57,520
Greenwich	28,739
Hackney	10,886
Hammersmith & Fulham	9,755
Haringey	18,646
Harrow	39,122
Havering	33,385
Hillingdon	47,922
Hounslow	27,313
Islington	7,336
Kensington & Chelsea	5,537
Kingston upon Thames	27,862
Lambeth	16,601
Lewisham	31,801
Merton	23,681
Newham	26,094
Redbridge	31,856
Richmond upon Thames	31,589
Southwark	17,817
Sutton	32,490
Tower Hamlets	11,633
Waltham Forest	21,822
Wandsworth	15,593
Westminster	5,040
Total	898,776

Unsurprisingly, those boroughs with the largest area see the greatest net increase in new homes, with Bromley at the top with 68,426 new dwellings, and Croydon slightly behind with 67,165. The inner London boroughs such as Camden, Kensington & Chelsea and Westminster see the least. The City of London is at zero and doesn’t appear in this table because it’s entirely within the Central Activities Zone and excluded as a result.

It’s important to remember that the figures I’ve listed above are limited to those areas within 800m of a station. That means there’s a lot of outer London excluded from my estimates, but imagine that we increase the density here as well, perhaps by a more modest amount…there would surely be many more thousands of homes that could be built in addition to the 900,000 I’ve suggested above.

Due to limitations in the mapping data, there are some anomalies which skew the figures in a few areas. For example, the Ordnance Survey mapping data doesn’t identify football stadia within its “sites” geometry, and while I’m ambivalent about the so-called beautiful game, and would be quite happy for every football stadium in London transformed into housing, I’m not sure Arsenal fans are quite ready for the Emirates Stadium to go the same way as their former ground just yet.

Using this methodology, the Emirates Stadium is identified as a location for intensification; Arsenal’s previous ground can be seen in the top right of the image, which was converted into homes after the club moved to its new location in 2006.

There’s also no adjustment made for those areas subject to wider regeneration schemes or empty sites. The large car park to the east of Stratford Westfield, which was going to be the home of London’s version of the Madison Square Gardens’ Sphere has an area of around 2.5 hectares and could feasibly provide 200-300 homes, but my methodology only shows an uplift of eight, as the density calculation is based on the entire MSOA area rather than this small section of it.

There are some other issues which could do with refinement. The MSOA boundaries do not take into account the type of space within them so, for example, with two polygons of equal size might have varying levels of undevelopable space. The total number of existing dwellings might be the same in both cases, and therefore the overall density would be shown as equal, however in reality the same number of homes could be crammed into a smaller area. This would mean that the impact of intensification would be more profound in the latter.

In reality, though, I’m not sure these anomalies make much of a difference overall as they seem to balance out across the wider picture.

So, these oddities aside, what does suburban intensification look like when applied to largely residential neighbourhoods?

Learning from south London

The troubled history of Croydon’s Suburban In t ensification SPD is beyond the scope of this article (I wrote about its demise for OnLondon), but it really was the gold standard for how outer London boroughs might encourage development on small sites in residential areas.

The guide provided a series of simple diagrams which mapped out the evolution of suburban blocks to show how, over an 18-year period, infilling gap sites and the replacement of some large houses with a combination of flats and houses. Let’s take a look at these to see what this means in numerical terms.

The first extract, below, shows a typical suburban block of detached houses. In the top example (2016) there are 37 detached houses. Although the plan is supposed to be a generic example, it’s almost certainly based on a real part of Croydon. There are large rear gardens and gaps of varying widths between the houses themselves.

In the “evolved” condition of 2036, several of the houses have been replaced with new buildings, and some have had new homes erected in rear gardens. From this plan it’s impossible to count the new number of dwellings that might be delivered in this way (it’s not really the point of the drawing), but the drawing does attempt to show the subdivision of the new buildings into the individual demises. Assuming nothing is taller than three storeys, I count at least 50 new homes, including a mix of houses and flats—and 29 of the original houses remain. In total, that’s a doubling of density – and it can hardly be said that the character has changed beyond all recognition: the large rear gardens largely remain and the coverage of buildings relative to undeveloped space is minimal.

This demonstrates that the kind of intensification we’re talking about is entirely achievable, and any objection on the basis of unacceptable change in character is for the birds.

Such an approach is entirely possible if we’re prepared to implement to bold policy reforms needed to enable this kind of development to come forward. In the brief period between the Croydon SPD being adopted in 2018, and it’s unceremonious scrapping in 2022, there was a remarkable uptick in small site development across the borough. The GLA’s annual Housing in London report shows that during this time Croydon delivered (delivered, not just approved) nearly 2,000 homes within developments consisting of fewer than 10 homes: more than three times the next highest, Barnet.

It’s time to adopt a London-wide policy which encourages similar levels of development across all of London’s suburbs. We know we need the homes. We now know we have the capacity. Let’s get on and do it!

You can have a play with my online map showing all areas of suburban intensification by clicking the image below.