Upzoning London: the solution to Britain's housing crisis
Upzoning London is how Labour can deliver prosperity and housing abundance.
Here is what I am going to try doing in this post:
Praise the new Labour government for its housing policies.
Highlight that they are stop-gap measures, at best.
Warn about some ways that they might go wrong.
Suggest a few ways to fix housing for good in Britain.
First: I’m impressed with what Labour has said on housing so far. Rachel Reeves used her debut speech as Chancellor to say that Britain’s ‘national mission’ is getting economic growth, that building more housing and infrastructure is the best way to do that, and that the planning system is the biggest barrier to that happening. Having this come from the Chancellor, and framing it as a growth issue rather than a cost of living or “young people” issue, is a sign that Labour grasps how important it is to fix housing in Britain. It’s remarkable to see how far YIMBYism has come in such a short space of time.
The actual measures the Chancellor has announced are decent. Hiring more planning officers will rebuild state capacity that the last government left to decay. Approving more important infrastructure projects at ministerial level could get us a lot of high-value stuff built, including projects that have been in limbo for years. Reclassifying damaged green belt land as developable ‘grey belt’ will probably free up land in useful places, and sets a precedent for future releases of green belt land. And bringing back housing targets, and setting them at a high level, signals the government’s commitment to addressing the housing shortage problem. But these are stop-gaps, at best.
The wrong targets
Housing targets in particular are a bad mechanism for getting houses built. One problem is that local authorities often allocate the land for them far from existing amenities and infrastructure, to reduce the imposition on existing residents.
More importantly, they put houses in the wrong places at the national level. Each area’s target is set to follow the area’s so-called “housing need”, determined by local trend population growth. That means that they do not allow places to grow above their past trend growth. Places that have built too little in the past, and thus had slow population growth, get lower targets in future, which causes them to have lower population growth than they would otherwise, and so on.
If we’re serious about fixing Britain’s growth problem by building more homes, we need some places to be able to grow much more rapidly than they have been growing. During the Industrial Revolution, some cities grew by orders of magnitude as people moved to them from the countryside to find better-paying jobs. Cardiff grew forty times in size between 1821 and 1891. Glasgow grew from 70,000 people in 1800 to over 700,000 in 1900. Most of Britain’s other major cities grew by at least five to ten times during the 19th Century.
High internal migration is normal in rapidly-growing countries. America’s Sun Belt has absorbed tens of millions of people from what we now call the Rust Belt over the past half-century, driven by better jobs and cheaper housing. China’s urban population has risen from 20% in 1980 to over 60% today, driven by and driving the country’s industrialisation over that time.
Something similar should be happening in Britain today, with people moving to prosperous cities like London, Oxford and Cambridge for better jobs and lives. But they can’t, because we don’t build enough in those places.
Since 2016, London has actually underbuilt for its size. London’s is 15.7% of England’s population, but only 10.8% of housebuilding starts and 12.6% of completions in England since 2016 have been in London. Really, it should be building much more than its share of population, so it can grow.
[EDIT, 21 July]: Jim Gleeson, of the Greater London Authority, tells me that this data is wrong. For some reason, and nobody seems to really be sure why, the government’s official data seems to undercount housing supply in London more than the rest of the country. Using other data that Jim says is more reliable, London builds about the same share of housing nationwide as its population share. That doesn’t change my general point that it should be building much more than its share, but I apologise for the error and I have removed a chart based on the government’s bad data.]
This isn’t just stopping us taking advantage of today’s growth, it also draws out the political pain of building houses. Housing targets work through local plans, which have to be written up by local authorities and then approved centrally by the Planning Inspectorate, and updated every five years. This means that they involve a never-ending political battle between central government and local authorities.
Targets are persistently unpopular, and create easy wedges for opposition politicians to campaign against them – something the Lib Dems and Conservatives have both excelled at in the past. Under current plans, Labour will face the same ongoing fights over new housing that have been a feature of the last sixty years of British politics.
Just say yes to tunnels and interconnectors
Similarly, the plan to call in new infrastructure projects is fine in principle, but in practice will draw out political pain for as long as the government actually uses it to approve things.
It is tempting to think that local objections will have little resonance nationally, but that often isn’t how things go. The Aquind Interconnector in Portsmouth – which would add as much clean electricity to Britain’s grid as a large new nuclear power plant – was already decided centrally in the way the government intends to decide on other projects, and it was blocked by the last government because of local objections. (Amazingly, it doesn’t even seem as if a majority of locals opposed the project, just a vocal minority.)
Even after that was overturned, that government delayed the decision indefinitely, for the next one – ie, the Labour government that has just been elected – to handle. What it decides will be an early test of the new government’s willingness to actually approve these infrastructure projects. To be clear, I expect this government to be much better at this than the last one – last week it approved 1.5 GW of solar projects that the last government had left in limbo, including one that was delayed by more a year.
But the Aquind project’s troubles were not just an exception owing to the unfortunate influence of Portsmouth’s former MP, Penny Mordaunt. Locally controversial projects can quickly catch national attention. The campaign against the project to bury the road near Stonehenge ended up being led by nature campaigners and one of Britain’s leading historians, and it remains in legal limbo.
The decision on the Lower Thames Crossing, which is already the largest planning application in the UK running at 359,000 pages, has been delayed until the 4th October. Unlike the quick action on the solar projects, the Government has issued a consultation letter (the fifth one since the end of its 6 month inquiry process). That project would unlock development potential for local ports, and relieve the Dartford Crossing, providing an avenue for growth. It is Britain’s only remaining transport megaproject after the cancellation of HS2. If, after £800m spent before a shovel has hit the ground, that project doesn’t get consent quickly, Labour’s pro-planning mantra may ring hollow.
Delaying the pain
Both of the measures above – housing targets and approving projects at the national level – have low up-front political costs and high ongoing political costs. It is costless to announce that you’ll decide on major infrastructure projects centrally: the costs come each time you actually have to do it. It is fairly low cost (though not costless) to announce the introduction of ambitious housing targets. Most of the costs come when people find out that that means new housing estates near them.
We don’t yet know enough about the ‘grey belt’ plans to know how frontloaded the political pain will be, but if the government drags out the process instead of doing it quickly, it will be another weight around its neck. Knight Frank has estimated that only about 1% of green belt land might be eligible for release, much of which will not be in very useful places. Tim Leunig has written today on this problem.
Despite its landslide in terms of MPs, Labour only won 35% of the vote. Maybe some of the factors that caused that (like Muslim voters voting for independents focused on the Gaza war) will fade away by the next election. Hopefully, for everyone’s sake, the economy grows quickly, and hopefully for Labour’s sake voters reward them for that. But there is every possibility that things look quite bleak politically for Labour in five year’s time. Will it still feel as bullish on forcing unpopular housing and infrastructure on marginal constituencies then as it does now? What if immigration doesn’t come down as quickly as Labour hopes – could Reform or a resurgent Tory Party sharpen people’s anxieties about that by tying that to the new housing being built near people?
How many houses is a lot?
Even if these plans do succeed, 1.5 million homes is really not that many compared to what we need. On average over the next five years it is 300,000/year – only 20% more than we managed in 2019. I don’t say that to knock it: 2019 was the most we’ve built for many years, and if we really do manage 1.5 million over this Parliament, it will mean we are building more annually towards the end of the Parliament than that.
But we are way behind where any reasonable estimate says we ought to be. In 2004, the Barker Review said the UK needed to build 297,000 homes a year. In practice, we’ve built about 150,000 a year since then – so that’s twenty years of a roughly 150,000 annual shortfall, or 3 million homes, just to catch up with where we “should” be. If you go by the Centre for Cities’s standard of reaching housing parity with Western Europe, we need to be building 442–654,000 homes per year for the next two decades.
And my view is that the UK needs to build vastly more homes than that.
Partially that’s because some parts of the country need to become much bigger than they are: Cambridge and Oxford, as 21st Century Cardiffs and Glasgows, could probably grow by 10-20 times. London might double in population size if we let it.
Partially, it’s because a lot of the housing we have built, especially since the Second World War, has been small, ugly, poor quality, and sometimes unsafe. That includes both private housing built during the suppressed-demand years of the 1960s and 70s, and the council housing built rapidly and cheaply after the Second World War that ought to be replaced with larger, better-built homes for council tenants.
Housing policy should not just be about keeping our heads above water: it should aim to create the conditions where we can overhaul huge amounts of Britain’s existing housing stock. Everyone – social tenants, private renters, owner-occupiers – should be able to expect a home that is bigger, warmer, safer, more beautiful, and more liveable than most of Britain’s current housing stock.
These are not criticisms of Labour’s plans. If it sees them through, Labour will have done more on housing in one week than the Tories did in fourteen years. But they will not even begin to fix the problem.
A framework for actually fixing British housing
I think the three most important features of permanent fixes to Britain’s housing problems are:
One, they should be demand-led. Unlike housing targets, they should go where prices for new homes are highest, where the best jobs are, where there is the least vacant stock, and where council waiting lists are longest. That means lots more in places like Islington, Westminster, Lambeth, Cambridge, and Oxford – not just in line with their share of Britain’s population (which would still be an increase in many cases), but significantly higher.
Two, they should be scalable. This means that its political costs are frontloaded, with minimal further capital necessary to get stuff built after enacted. This follows the same model that allows software companies like Google and Facebook to build products that the whole world uses: high fixed costs (to develop and implement the products) and low or zero marginal costs from extra people using them.
Three, they should be durable. This means that once in place, it is difficult to reverse them. This normally means either by rewarding certain interest groups (like local authorities) for taking housing and infrastructure near them, or using opt-outs to allow the most passionate opponents to avoid things they don’t want, so they don’t bother campaigning against it nationally and don’t block the whole policy.
Here are three examples of policies that are demand-led, scalable, and durable.
1. Estate regeneration
This would allow the residents of housing estates (mostly social tenants) to vote on developer proposals to redevelop their estate, adding new private market units and using the proceeds from those to pay for the redevelopment of their own homes.
There have already been thirty ballots of projects like these around London, of which all have passed, some by incredible margins:
Estates can increase enormously in density: in Tower Hamlets, affordable housing provider One Housing is redeveloping 24 homes into 202, of which 50 percent will be rented at below-market rates. All 24 existing households will get new, higher-quality homes, and 84 percent of residents approved the proposals with 100 percent turnout. In Lambeth, 135 homes are being redeveloped by Riverside Group into 441 larger and higher-quality homes, of which almost half are submarket. This was approved by existing residents at 67 percent with an 87 percent turnout. The development comes with not only new homes, but new amenities like a gym, a community center, cycle parking spaces, and a new communal outdoor space.
Another proposal to replace 330 homes with 1,582 new ones, including 116 new socially rented units (a 45% increase), in an estate 30 minutes’ walk from Canary Wharf passed with 93% of the residents in favour with 91% turnout.
But at the moment the policy is not scalable because it requires the approval of the local council for a project to be voted on, and then the council’s approval again even after residents endorse it. Change the balance of presumptions so that residents can trigger ballots themselves, and the local council cannot block it without clearly defined reasons for doing so, and we could get many projects like this approved rapidly across London and the rest of the country.
Because the value of the new homes determines how good the offer can be to existing residents, it is demand-led, and because stronger Estate Regeneration powers would create the prospect of better housing for estate residents across the country, it may be durable since these people would lose directly if the powers were scaled back later.
2. Permitted development rights for mansard roofs and upward extensions
A mansard roof is an extra storey added to a terrace house that looks to the outside like a sloped roof. The National Planning Policy Framework already includes recently-added language to favour these. At a minimum, Labour can just preserve this wording, it could go a lot further by clarifying further that any terraced home that wants to a mansard roof is allowed to do so, and the local authority cannot object except on the grounds of safety. (Houses in conservation areas and listed buildings should have to do mansards that are in keeping with the existing neighbourhood and/or property’s design.)
Doing this would be demand-led, because these terraced homes are some of the most valuable properties in Britain; it would be scalable, because it would apply to a huge amount of the existing housing stock; and it would be durable, because every terraced homeowner in the country – 6.93 million homes, 26% of the total housing stock – would gain a valuable right that they would fight to keep.
3. 100% business rates retention for strategic infrastructure
This one isn’t about housing, but it highlights how useful this framework can be. I wrote about this proposal recently with Ben Southwood, but the basic idea is to let local councils keep the full business rates bill from certain kinds of infrastructure projects in their area.
This is demand-led because the payments to councils are larger the more valuable the project is; it’s scalable because it applies to every project of this kind, rather than being done on a case-by-case basis; and it’s durable because local councils that could benefit from it will fight to keep it in place if and when it is introduced, because it benefits them so much.
Maximum London
None of those three reforms are going to “solve” housing either. At scale, they might add a few hundred thousand extra homes over a single term, and a couple of dozen of billions of pounds of extra infrastructure. Importantly, they would do so in high-value cases.
The only way to truly fix Britain’s housing market is to repeal or completely overhaul the 1947 Town and Country Planning Act (and its 1990 successor). I’m not convinced that that’s politically possible, and I don’t expect Labour to risk its majority for the sake of doing it. I also doubt it would be durable: I think the temptation for a Tory opposition to pledge to bring it back would be irresistible.
So here’s a second-best suggestion: massively upzone London. The Housing Secretary could do this without primary legislation by making a Special Development Order granting the right to build up to eight stories on residential land within the Greater London Built-Up area and within half a mile of any tube station, London commuter railway station, or Elizabeth Line station outside the Built-Up area.1 Places like Southall, Ilford and Gidea Park along the Elizabeth Line could get huge amounts of extra density. Anya Martin has described a version of this as the 1894 Plan.
They could couple this with new mechanisms for local approval of design codes, decided on by ordinary voters, not planners or architects, so areas have the option to get the sort of dense, beautiful buildings that exist in London’s most popular areas. Let Stockwell, Peckham and Hackney grow to be as beautiful as Bloomsbury, Kensington, and Marylebone.
This approach would meet two of my conditions. It would be demand-led, because London is where there is most demand to live, and within London the most housing would be built in the places where demand is highest. It would be scalable, because it would be a “one and done” rule.
But would it be durable? NZ Labour lost the election after its most ambitious nationwide upzonings. Nobody can ask Britain’s Labour to risk the same thing – and it already has experience of an electoral backlash after attempting an upzoning in Croydon in South London. But in New Zealand the upzoning that backfired applied across the whole country; doing it in Greater London only would limit the risk to a smaller number of seats, and could help Labour outside London if it meant fewer houses being needed elsewhere.
But that might not be enough. In Houston, a Democrat-led city-wide upzoning was made durable by allowing parts of the city to opt out. Contiguous areas of London (ie, neighbouring blocks and streets) could, similarly, be allowed to opt out if 51 percent of the homeowners signed a petition to do so. This meant that areas where a majority of people wanted to keep things the way they were would be able to do so without obstructing upzoning in the rest of the city. The map above shows how little of Houston actually opted out this way.
As well as being good for growth and housing affordability, this sort of upzoning would complement Labour’s Net Zero plans, since walkable density reduces people’s reliance on cars to drive, and greater in-city density makes more public transport options viable.
Obviously, this would require a lot more transit infrastructure. Here is one area where “land value capture” could be powerful. Landowners near a new Tube station will enjoy a large windfall in the value of their properties as a result. Taxing that windfall would allow us to capture a lot of the benefit of new Underground lines, and make it possible to pay for new ones altogether.
The economic effects of this could be huge. I won’t re-do the Housing Theory of Everything any more here, but study after study estimates large wage, economic growth and innovation gains from making it easier for people to move to rich, dense cities. The effect is probably even larger in the UK than the US.
Incidentally, I would let other cities apart from London do these things as well, but leave it up to Metro Mayors to decide how and where to upzone in their areas. London is different: it’s Britain’s preeminent city, and one of the world’s most important cities. It is large enough that densifying it alone would solve a lot of Britain’s housing shortage, but small enough that this could be done without a fatal political backlash.
Don’t settle for half measures
There are two big reasons Labour might not do this. The first is just political: it would be a big move. But I think a significant amount of political pain is unavoidable in building new homes, and this approach has benefits over the slow grind that Labour is currently planning for itself. They might as well do the houses where they’ll produce the most new growth.
The second reason is intellectual – Labour is very interested in capturing some of the land value that arises when land gets planning permission. But unlike, say, a new Crossrail line, planning permission is artificially scarce. In a world where there was no planning barrier to building homes and infrastructure, there would be no land value uplift to capture from grants planning permission. So the better your reform is in terms of ending artificial scarcity, the less land value uplift there is to capture.
On the other hand, land value uplift from genuine improvements would be higher under this plan, because things like new parks and infrastructure, beautiful design codes, safer and cleaner streets, etc, would be enjoyed by more people in any given area. Developing a way to capture that uplift would be a sensible way to pay for more of them.
It is to the new government’s immense credit that it understands the fundamental problem with Britain’s economy. The measures it has announced are decent, but can only be the first step to fixing the housing shortage. The framework I’ve set out might help the government to think about effective smaller improvements to the system, like the 100% business rates retention policy that could make councils happier to take new prisons and data centres. Piecemeal improvements like this are the most likely path to fixing the problem.
But remember how far we are from where we ought to be: we could be as rich and dynamic as America, with a welfare state surpassing that of Denmark or Sweden. That’s what growth gets you, and we all agree that more housing is how to get that growth. So let’s really go for it. Take away the limits on London’s growth, let it become the beautiful, dense, booming global capital that it wants to be, and create a Britain with better, cheaper, and more abundant housing than we’ve had for almost a century.
A new National Development Management Policy that supports building would help stop this decisions from being held up in the courts.