The National Planning Policy Framework is a central document, introduced in 2012, with which all local plans in England are required to be consistent (planning is devolved to Northern Ireland, Wales, and Scotland). Usually, tweaks to it take several years to be reflected in plans, and while plans are being updated, existing policies remain partly in force.

But housing minister Matthew Pennycook, in his speech, said that local plans should be given ‘very little weight’ where they contradict the new framework, effectively invalidating them. The new framework could come into force in just a few months. This means that, in principle, developers could be submitting successful planning applications in line with the new NPPF in the summer of 2026.

Share

The changes would, among many other things:

1. Remove the ability of councils to block housing and mixed use development on the grounds that it is in one of England’s green belts within 800 metres of a train station (GB7.1.h), so long as any proposed developed is denser than 50 dwellings per hectare, as well as creating a presumption that such development should be approved unless its benefits would be ‘substantially outweighed’ by adverse effects (S5.1.h).

2. Introduce a presumption that anyone, anywhere in the country, be allowed to build additional homes in their back garden (L2.1). Strikingly, this need not be a single granny flat, but could be multiple units, or even a small block of flats, so long as its overall footprint (the amount of floor area it uses) is no greater than double the existing property, and so long as it does not use up more than half the undeveloped area of the plot.

3. Remove the ability of councils to ‘gold plate’ building regulations (PM13), the rules that govern the basic standard of homes that developers are allowed to build. No more could individual councils or mayoralties require stricter rights to light, or balconies, or dual aspect, and so on.

4. Give very small sites (under a quarter of a hectare) an exemption from the requirement that otherwise exists for all development to prove it generates a net improvement for biodiversity.

There are some serious hurdles that the policy will need to clear during the consultation period. It may leave big loopholes. For example, GB7.h.iv may allow a clever council to prevent any of the new presumptions applying by generating a comprehensive redevelopment plan for the area around a station, even if it rejects any planning applications made to fill out parts of that plan.

It may be held back by other parts of the system. Viability tests appear to have been removed as a constraint on social and affordable housing demands, meaning that all developments pursuant to these changes may all be required to be 35 percent social housing, as well as 15 percent below-market rent (‘affordable’). In many areas, this will make development unviable. Even where it does not push development out of viability, it will use up an enormous fraction of any value generated by granting the permissions.

Even if the Government removes glitches like these after the consultation, there will be political blowback to handle. Britain has seen many rounds of abortive planning reform: in recent decades, there was New Labour’s attempt to allow more suburban densification (subsequently revoked), the Conservatives’ attempt to reform planning rules in the Planning for the Future white paper (largely abandoned), the current government’s attempt to reform environmental restrictions in the Planning Bill (hollowed out), and many others. The Government will need courage, vigour and resourcefulness to win the political battle.

The infrastructure deficit

These are all issues. One other issue which could be huge, though, is infrastructure. Development unlocked by the new rules is likely to be concentrated in the South East, where housing scarcity is greatest and prices highest. YIMBY Alliance estimates that the new rules on development near stations could create space for an estimated one million homes in this region. The rules around garden flats could create many millions more – there are seven million houses with gardens in London and the South East.

With 2.35 people in the average household, this could mean enormous growth in the population of the London metropolitan area – perhaps by millions – over ten years. This vast new population will need a huge amount of new transport infrastructure, a fundamental upgrading of the transport capacity of southeast England.

The NPPF draft makes no real provision for this. Development will be subject to the usual charges.1 These systems are often adequate for funding routine local transport infrastructure like access roads or bus stops. But nobody expects these to fund massive region-wide projects of the kind that will be needed.

This is concerning. There are now more rail passengers into English cities than before Covid. At peak times, most lines entering London are already at capacity: over 100,000 people are forced to stand on the trains into London every morning. In 2019, London was rated by the OECD’s researchers as the major city with the worst roads in the entire developed world, the only place in their entire sample where driving was generally slower than taking public transport. If millions of people are added to the South East they will either be unable to access road and rail, or they will cannibalise the road and rail use of existing residents.

Why we can’t just build first and ask questions later

One might suppose this problem will be easy enough to solve further down the line. If we really did build a million homes in the South East, it would release enormous economic growth, perhaps four percent per year, as we converged towards our pre-2005 growth trend. The Treasury would be replete with cash. One might assume that it could just fund the infrastructure later, as this happens.

Don’t depend on it. The UK government is incredibly bad at responding elastically to shifts in demand around the country. The population of Britain has grown by ten million, and shifted significantly around since 1992, but we have gained no reservoirs in that 33 years. Heathrow actually has fewer runways now than when it was first built, and the number of aircraft movements (takeoffs and landings) has been flat for 24 years. Our motorway programme stalled after 1997, since when France has added as many miles of motorway to their network as Britain has in total.

Since the A13 and the M25, London and the wider South East have seen barely any major roads built, even though there are 50 percent more cars on the roads than 30 years ago, and even though between 1991 and 2021 its population increased by six million, more people than live in Denmark. No wonder that the region has the most congested roads in the country. Nationwide, the number of railway services increased by 32 percent between 1998 and 2019, while passenger numbers doubled.

Successful infrastructure projects can take an incredibly long time to get off the ground, even when the government isn’t hard up. The Elizabeth Line was first proposed in a report in 1974. It received government backing in the late 1980s but was cancelled in 1994. Planning restarted in 2001, it was approved in 2008, and it finally opened late in 2022. Outside London, cities are left waiting even longer.

Many British projects are also very expensive: HS2 is the most expensive new railway in the world, costing 5.5 times as much per kilometre as the next-most-expensive high-speed line, which is being built between Lyon and Turin and includes a tunnel under Mont Blanc. This makes the taxpayer, and its representative, HM Treasury, reluctant to fund more infrastructure, which is part of why it takes so long to build.

If this infrastructure doesn’t arrive, it will create two big problems. First, the public will be more likely to revolt against the increase in building, because it is making their lives worse as they sit in more traffic and can’t get a seat on the train. Some NIMBY grievances may be pretextual or frivolous, but overloaded transport infrastructure is a completely real way in which development makes the lives of existing residents worse.

Second, it will undermine the benefits of new housing itself. The main purpose of adding more homes to successful urban agglomerations is to allow people to move towards better paid jobs from lower-productivity areas. But if congestion on trains and cars worsens sufficiently – and 2.3 million people would do that – then the incomers will be improving their own lot at the expense of the existing residents. Slower transport means smaller job markets and lower productivity.

Too many people and not enough roads is a big part of why developing countries are poor. A World Bank report found that traffic congestion in Cairo makes Egypt about four percent poorer, roughly equivalent to the contraction in the UK economy after the Financial Crisis. We do not want the London metropolitan area to become a ‘Cairo of the North’.

Share

Crossrails

Luckily, there is a surprisingly easy solution to rail overcrowding in London: crossrails. Most of Britain’s railways were built in the 1840s and 50s, at which time it was almost impossible to tunnel a railway under a city centre. As a result, London’s railways generally terminate at a big station on the edge of the centre, like Waterloo or Victoria.

This was the best the Victorians could do with the technology available to them, but it imposes a low ceiling on capacity. Terminating a train takes far longer than stopping and carrying on, especially in the centre, where most passengers are going. Many more platforms are thus required at a terminus station than there are pairs of lines going in if platform capacity is not to become the binding constraint on the capacity of the entire network. For example, Waterloo has four pairs of lines going in, which could carry perhaps 96 trains per hour, but only 24 platforms, which turn around no more than 42 trains per hour. Terminus capacity at Waterloo thus halves the total capacity of the enormous rail network that fans out from it, compromising the transport system of the whole of southwest London and much of southern England.

The solution is joining up two terminating networks, each on opposite sides of the city centre with a tunnel (or occasionally a viaduct). The old suburban railway systems of Paris, Madrid, Berlin, Munich, Frankfurt, Hamburg, Zurich, Leipzig, Stuttgart, Vienna, Copenhagen, Oslo and Stockholm have all been joined up, creating immensely powerful networks like the Paris RER, the Madrid Cercanías, and the German S-Bahns. Crossrailing has also been practised on a huge scale in Japan, with Tokyo joining up no fewer than thirty lines between the 1960s and the 2010s.

Urban expansion in postwar Europe has been fundamentally shaped by it too, with much development concentrated around railway stations on the joined-up nineteenth-century lines. In effect, crossrailing has systematically enabled the kind of rail-orientated development the Government is proposing for decades in countries like Spain, Germany and Denmark.

Share

Crossrailing is not unknown in British history either. London completed one of the world’s first crossrail projects with the Snow Hill Tunnel, creating what became Thameslink in the 1980s. The Central Line extension was another example in the 1930s. Liverpool did this twice in the 1970s, and Newcastle once. Much of the Manchester Metrolink operates on a similar principle, weaving together the alignments of old suburban railways with on-street trams through the city centre.

The Elizabeth Line is the most recent example: on its own it added 1.5 million, or 30 percent, to London’s labour market catchment – the people who can reach one of London’s three central business districts within 45 minutes by public transport. It has proved to be immensely popular: six of the ten busiest stations in the UK appear in a row in the Elizabeth Line’s core.

The total length of the Elizabeth Line is around 118 kilometres, but 100 kilometres of that existed already. Yet ridership still rose by about 300 per cent when it opened. In other words, London got three quarters of the ridership of a whole new branched metro line while having to build only one sixth of its length. This pattern is why crossrails have so persistently been a good investment around the world.

Compared to the major cities of the continent or Japan, however, Britain’s crossrailing efforts have been modest. London has crossrailed perhaps a fifth of its Victorian railway system, whereas Paris, Madrid and Tokyo have joined up nearly all of their terminating legacy lines. Huge swathes of South London, along with most of the great cities of the Midlands and the North, rely exclusively on the railways of the 1800s, low-frequency services terminating at overcrowded stations like Waterloo and Liverpool Street. Manchester has dozens of stations, but no capacity in the city centre to run frequent trains from these stations. The scope for further crossrails remains immense.

Funding through value capture

There are three main ways to fund rail infrastructure: taxpayer subsidy; borrowing against future fares; and value capture (both private and public). Wednesday’s de facto cancellation of the Leeds tram, to which the government had committed £2.5 billion, is a reminder of how risky it is depending on government subsidy in today’s fiscal constraints. Fare income funded the construction of most of Britain’s railways in the nineteenth century, but modern urban rail is both heavily price controlled and faces tougher competition than Victorian railways did, from aeroplanes and cars.

The remaining opportunity is value capture. When a railway service is improved, properties around its stations become more valuable because of their greater access to jobs and amenities. ‘Value capture’ means using some of that uplift to cover the cost of the improvements.

Value capture was widely used as a means of funding railways in the USA and Canada in the nineteenth century, and it was used in England to help fund the famous Metropolitan Line Extension into ‘Metro-Land’ in the early twentieth century. It has remained routine in Japan down to the present day. The recent Northern Line Extension was funded largely through value capture. Value capture has been used to fund crossrails in Paris, Hong Kong, and Tokyo, as well as about one third of the Elizabeth Line.

There are three main ways to do it. The first is speculation. A private or public corporation buys land near where a station is going to go, at near-farmland prices. When the railway is built, they sell or rent that land out. This is what Japan’s private railways and Hong Kong’s municipal railway operator do. It is how the Kansas Pacific Railroad was funded in the US, along with many streetcar suburbs.

The second is tax. With the tax model, a public authority, usually a local government, taxes some fraction of the value that the new infrastructure generates for property and landowners. Sometimes this is passive: countries like Switzerland, France, Denmark, the US and Japan with high taxes on property automatically get back some of the money they spend on infrastructure when it drives up home values. In other cases it is more active. Business improvement districts often approve new taxes on commercial landlords to fund investments that boost all of their revenues. American local voters often vote by referendum for their local authority to issue billions of dollars of bonds to build infrastructure like roads, repaid by increased property taxes.

The third is permissions. Many countries keep the permission to build homes scarce. In these countries, being granted permission to develop farm or industrial land into residential properties immediately raises the value of the land dramatically. The Resolution Foundation estimates that giving a landowner permission to build homes on London greenfield adds about £35 million of value per hectare. Charging for permission, or extra permission, to build around infrastructure is another standard way to fund it. Brazil’s CEPACs are a kind of auctioned development certificate that have raised billions for infrastructure in this way.

Since the government already wishes to give out extra permission, this third option is perfect.

Share

A worked example

Here is one concrete example. The Great Northern line runs through the Hertfordshire green belt into Moorgate Station. Moorgate Station is small and currently achieves only eight trains per hour in the peak, although the tracks could take twenty-four. Meanwhile in South London the Sutton Loop line is limited to a train every half an hour in each direction, which contributes to the overloading of the Northern Line.

We propose constructing a new tunnel between the lines at Old Street and Elephant and Castle to join these two lines up. With some targeted infrastructure improvements including grade separation at Herne Hill and additional platforms for terminating, this route could be segregated from other services, allowing up to 24 trains per hour through the new core under central London, the same peak-time service as the Elizabeth line. We might call this the ‘Oldcastle Line’ as a portmanteau of ‘Old Street’ and ‘Elephant and Castle’, taking our lead from the Bakerloo Line.

The Oldcastle Line would roughly triple frequency on the Great Northern line north of Moorgate as well as sextuple frequencies at stations like St Helier in south London. It would also provide passengers from these stops with direct services to the City and the Southbank, and to direct interconnections with nearly every Underground line. Like the Elizabeth Line, the Oldcastle Line would be transferred to Transport for London.

In the north, the Oldcastle Line would take over a string of green belt stations, which would suddenly receive frequent and highly interconnected services into the capital. Many have abundant developable land under the Government’s new planning policies.

Consider Bayford, a small village with its own railway station on the GNR. Bayford Station has nearly 200 hectares of farmland within one kilometre’s walk. Granting it planning permission for 50 dwellings per hectare, as the government is intending, will give each landowner an immediate uplift of £35 million for each hectare they have, for a gross development value of £7 billion. We have identified ten stations like this on the Oldcastle Line alone. Many other lines are similar.

Working with a team of specialists, we are working on a plan for connecting suburban railways across Britain in this way, funded by the proceeds of a hypothecated charge on new planning permissions granted near stations. There are many engineering challenges, but they are surmountable.

The government need not accept our exact proposals. The important thing is that they accept the principle. Mass housing development without infrastructure will cause gridlock and backlash. Building infrastructure in arrears, when the taxpayer can be convinced to fund it, is foolhardy. Instead, the enormous uplift that will be generated by handing out millions of homes’ worth of planning permissions should be turned to providing the infrastructure they will require. In practice, doing that means development charges, hypothecated to new Crossrails.

The future of British urban rail

The 1930s remains the most prolific decade for housing delivery in British history in percentage terms, peaking in 1936 at nearly three percent of stock. Today that would be equivalent to 760,000 new homes built in a year, enough to achieve the Government’s target in just two years.

Doing this required a truly gigantic amount of new infrastructure. Under Sir Herbert Walker, the Southern Railway electrified hundreds of miles of track into Surrey and Sussex, creating the modern commuter belt. The Central, Northern, and Piccadilly, District were extended, connecting vast new suburbs to London and giving millions of Outer London swift connections to jobs and amenities.

The Ministry of Transport pushed through a massive program of arterial roads and bypasses, funded by the new tax on petrol, to unclog Victorian streets. This included the A40 Western Avenue, the A12 Eastern Avenue, and the North Circular (A406). A vast fleet of motorbuses and coaches created new transport connections across England.

Homes and infrastructure go together naturally. The Government’s schemes could deliver enormous numbers of homes. Land value capture would give us the infrastructure those homes need.

Subscribe now

We should implement the sweeping and radical reforms to planning, environment and habitats protections, and radiological rules that the taskforce proposes. But if I had the power to push through one, and only one, recommendation from the list it would be Recommendation 36: Proposals in relation to community benefits.

The UK planning system is built around the principle that planners around the country are deciding on the socially optimal outcome, without reference to anyone’s individual interests. Everyone knows that planning is driven by democratic preferences in practice, but there is a widespread social myth to go with the law. A Victorian would have objected to new houses because they affected his view, or brought in social undesirables to the area. A modern Briton is obliged to pretend that he has some broader social purpose, like protecting newts, or maintaining the character of the area, or preventing excess nutrient runoff in the water.

This has been backed up by case law. In 1981, a local area was not allowed to impose the condition that a hangar was packed up after ten years on the grounds that if it was objectively good for the human race that the hangar be built, it couldn’t also be good to remove it after ten years. In 2019, a prospective wind farm developer offered four percent of the turbines’ receipts to local residents in exchange for allowing them to be built. The local community and its representatives said yes. Lord Sales ruled that planning permission ‘cannot be bought or sold’.

Development is allowed to mitigate harms, whatever the cost. So, for example, you can spend £700 million on preventing the deaths of 10 or 20 protected fish every year or two, or £120 million saving 300 somewhat rare bats. You can prevent every new building in London over building safety, or an appreciable fraction every new greenfield house in the country over alleged nutrient runoff.

Decisions like this create obvious perverse incentives. If councils are allowed to block development, so long as they can come up with some nice fluffy-sounding objection, like heritage, nature, or the environment, and their voters have every incentive to block it, because all development is obliged to benefit society as a whole but never the local community, then they will always find a way. Which is why I have said before that infrastructure costs are high in the UK because people want them to be high.

This change is the most radical in the report, and would fundamentally upend the whole purpose of the Town and Country Planning Act 1947, and its associated reforms, which was to replace local government incentives for development with rational scientific planning. The TCPA’s problem was not that it gave local governments extra planning powers, though it did do that, but that it deliberately removed all the incentives local governments had to approve housing, and other development. This would reverse that, albeit only for nuclear power projects.

Once locals have an incentive to see development go ahead, whether that is housing or nuclear power, their position on bat tunnels, fish discos, pointless radiological interventions and so on changes dramatically. Suddenly they prefer development over delay and proportionality over uncontrolled cost. Evidence around the world shows this.

Share

Hundreds of thousands of homes in Seoul were built through joint redevelopment projects, a kind of land readjustment where Korean shantytown dwellers could vote by supermajority to upzone their neighbourhood to 55 storeys. Senakw, is a project in Vancouver where the local Squamish nation voted to upzone their own land outside the city centre to similarly huge towers. When asked whether they would impose a swathe of environmental, nature, and other cost-increasing bylaws that they happened, as a First Nation, to be exempted from, a spokesperson said ‘we have decided to maximise economic benefits for the community’, or in other words, ‘no’.

Israeli apartment dwellers, who now have the power to agree by supermajority to demolish their building and put a bigger one on its site, now often sue their local government to be more permissive. Houston, USA’s permissiveness is backed up by a community right to opt out of reforms.

Most importantly, the gigantic French buildout of nuclear power was enabled by extreme generosity to locals. ‘Avoine, the commune that hosted the first EDF reactors, for example, saw annual revenue jump from tens of thousands of francs in the mid 1950s up to nine million a year by the time EDF2 and EDF3 were being brought online in 1964–1966. Locals dubbed their town “the Kuwait of Indre-et-Loire”…The northeastern town of Chooz has offered its residents free high-speed internet and a 120-channel TV subscription since 1999. In 2012, local councils in France paid an average of €35 in subsidies or benefits per local resident. In the 19 areas that hosted nuclear power, the average was €450.’

No wonder France had such a weak anti-nuclear movement. Most of the people who might have campaigned against it – angry locals – were neutralised. And no wonder that France is having trouble with its latest wave of nuclear power, as it is shifting substantially away from local taxes towards nationalisation.

It would take a while for this change to bed in. Therefore, I think the ideal option for the government is to simply take the flak and implement the report in full. Eventually the local benefits may make the equation work. But if I could get only one reform from the report it would be this one, which could fundamentally change the public’s view. The public is softly pro nuclear power, but with little sense of the upside of fixing regulation. By giving them a direct stake, this could change that, and while it’s doing so, create a case for changing the planning system across the board.

Subscribe now

Ever since we started Works in Progress, people have asked us, again and again, if we would produce a print edition. We never wanted to. We thought it was a bit obvious of us to produce a print edition, and we thought it would be a radical and alternative design choice to be an online-first, online-only and online-forever magazine.

We held onto our cherished little belief for years. Nick Whitaker, our only Generation Z colleague, wore us down steadily over time, arguing that the ephemeral digital world was really a cringe millennial, gen X, and boomer thing, and that youngsters these days had a strong affinity for the tangible and physical. Evidence kept accumulating that he was right, but we kept resisting his message, though he told the truth.

Eventually we gave in. I prefer to read absolutely everything on my phone. I gave away most of my physical books over the last ten years, having held onto only the ones that work as wallpaper. Even if I own a book, I will buy the epub to read it, as I prefer the experience. I acknowledge that this is a strange preference, and that most of our readers, especially the younger generation, strongly prefer to read things in physical print. So I am excited about better reaching that audience.

There is another benefit. In once realised, during the time I worked for UK think tank Policy Exchange, that I had a chance to put a research paper in the hand of then-Prince now-King Charles. I had reason to believe that that paper, Strong Suburbs, would be something he would like, given its tilt towards popular design, traditional urbanism, and the Great Estates. But I could hardly hand him a QR code, or get the Prince of Wales to Google my paper! So we rushed out a whole print run of 1,000 in order to have them in time to get three handed over to this important individual. I have no reason to believe that it made any difference, but it was a worthwhile gamble. I think the same will be true of the new print Works in Progress: it will be something I can hand over to Prime Ministers and Presidents, should I meet them.

I think there is a third reason why print is good, as well as pleasing zoomers and VIPs. Banks are said to have grand headquarters because it was a credible, unfakeable signal that they had a lot to lose by taking risks and going bankrupt, or by running off with depositors’ savings. Beacon, a gorgeous town about an hour north of New York City, through which I travelled to escape New York (via Stewart International Airport and Iceland) during the Heathrow fire earlier this year, has an old bank HQ that is literally just a giant triumphal arch. It is a common form for American small towns. I think our physical magazine will perform a similar function for some people, proving that we are a real organisation, rather than ‘just an online blog’.

If you are a fan of what we do, I can promise you that you will want to read the physical magazine. All of our main articles will still go onto our website and Substack, and if that’s your preferred way to dip in and out of what we do, then nothing will change. But I am confident that the extra content and the presentation will make our magazine the ultimate version of Works in Progress. If you are interested, then subscribe now, and you will receive the first issue in November, as soon as it is printed, and the following five issues as well. I am very excited for our next chapter!

You can find it here, in full: https://ukfoundations.co/

We show how Britain led the world, and then Europe, and then declined to be a European also-ran, in productivity terms, far behind France and Germany, let alone the USA. We show how this mirrored our decline from being the world’s greatest builder of homes, infrastructure, and energy, between around 1650 and 1940, to becoming the world’s slowest builder, and most energy-starved rich nation.

Investment has been correctly identified as a core reason why Britain is behind. But so many think this is just a case of convincing the Treasury to be ever more profligate, and invest £150 billion in HS2, for example, five or ten times the price it should cost. After all, it all counts as investment.

We explain that it is blocking investments the private sector would like to do – in buildings, railways, roads, and more – that explains decline. Thankfully, that means fixing it, and returning to French and German levels of productivity, or, as in history, higher ones, is quite straightforward. We just need to build.

I know that was then, but it could be again!

Since the 1800s, birth rates have fallen across the developed world. They temporarily rose in many developed countries before, during, and after the second world war, but have since declined further, and this time they have stayed low. The rest of the world has followed.

Childcare costs are very high in some countries. Housing costs, making extra bedrooms for children expensive, are high in many. And, increasingly, the Western countries with the highest fertility rates are those with high female labour force participation rates, like Sweden: countries that tend to have egalitarian gender norms and generous maternity leave policies. More socially conservative developed countries, like Italy, South Korea, Japan, Greece, and Taiwan tend to have the lowest birth rates. Richer people now have more children. 

All of these facts together have led some to argue an ‘economic’ story behind fertility rates.

Total fertility rates predict the number of children the average woman currently of childbearing age will have by the time she reaches the end of her fertile years. The gaps between national fertility rates in 1980 and 2000 are pretty small, and have narrowed since. Spain’s 2023 total fertility rate was 1.19. Sweden’s was 1.49, as was the UK’s. France managed 1.68. These differences matter quite a bit – France is having 80,000 more births than the UK each year – but they are not dramatic.

The gap between Swedish and Spanish fertility, 0.3 births per women, is dwarfed by the gap between native and non-EU-born residents of Belgium (0.8), or Germans and non-EU-born citizens of Germany (just under 0.6). And these are dwarfed in turn by the fertility rate gaps between either of these groups and ultra-high-fertility groups like Hasidic Jews, who often have six or seven surviving children, or the Old Order Amish, who have about five. 

These larger gaps have led people to argue that it clearly must be ‘culture’ that is most important.

Subscribe now

Culture responds to economic incentives.

Culture can also be called ‘folkways’: the way that your group does things. The idea is that you follow a script which lays out the established fertility patterns of what your group does, rather than doing a complicated economic calculation and working out whether you value the time playing with your children, their visits in old age, and so on, as balanced against the costs of extra space, the reduced freedom, the time spent on drudgery, and the inconvenience. 

Clearly some version of this story is true. The most important human trait is copying and mimicking. Doing complex economic calculations around every one of your decisions is very difficult. 

But culture changes, implying that it isn’t just about mindlessly following what the last generation did. My view is that family-level culture is mostly about copying the perceived behaviors of those who are successful, because working out all the exact answers is too difficult for every individual. This is a bit like how names tend to start off in more economically successful ‘classes’ or families, before making their way down the social hierarchy.

My view helps explain how fertility transitions work. French fertility declined much earlier than the rest of the world. It appears to have been linked to a ‘culture’ of rising secularism: areas seeing the decline also saw wills use less religious language, and fewer testators (those giving wills) donating money to Catholic churches so that priests would remember them in masses for as long as the church existed.

But this cultural shift could also be seen as copying a cultural invention that ‘worked’.

Aristocratic fertility had declined even earlier in France. More generally, fertility fell first among the richer classes in all European fertility transitions.

We could see this as the most successful people first in society discovering a norm that worked (investing more in fewer children) and later the others copying them. So we might well say this transition was mostly about ‘economics’ – which patterns are most successful within society – and culture is just a description of different folkways people adopted as they proved to generate successful children.

This happens to immigrants to modern Western countries too. They come from vastly different cultures and usually follow a different ‘script’ in the first generation – with different languages, accents, clothes, sexual and relationship standards – and of course fertility rates, as I pointed out a few paragraphs up. But they usually converge to native fertility within one generation, presumably because they adopt the native culture (aka ‘learn what the accepted script of respected behavior is’).

Share

Cultural and economic interventions are hard to disentangle. Realistic ‘cultural’ interventions look economic.

Take drink driving. The first step to tackling it was implementing weak, slap-on-the-wrist penalties for those caught driving under the influence to indicate that the behaviour was abnormal, and that society disapproved of it.

Paragons of the community – hardcore norm followers – rapidly began following the new norm, now common knowledge had been established. This removed the stigma against punishing drink driving, which had been ‘normal’ behaviour. This meant penalties could be steadily ratcheted upwards, until only a tiny minority of people who don’t care what society thinks of them were drinking and driving. Fines and incentives changed culture.

If we wanted to encourage fertility in the same way it would be totally inappropriate, probably impossible, and borderline fascist and evil to start fining or punishing people for not having children. But there are some antisocial behaviours that we might want to softly discourage, though not with criminal penalties.

Lots of expectations, perpetuated through behaviours, raise the cost of having children by socially pressuring people to spend more money or drudgery on them. We could discourage these.

We might wish to tax mostly positional goods, like private tutors, expensive weddings, or luxury children’s shoes. We could discourage the social norm of putting in enormous effort to make children costumes for special days at school, of giving children separate bedrooms before they are teenaged, and of buying them new clothes and toys rather than giving them hand-me-downs.

On the positive side we could encourage (and subsidise) things that lower the cost of kids, like alloparenting – parenting by those other than the parents themselves. We could make degrees and career paths more flexible. The more high status people who access these schemes the more such schemes are likely to be viewed in a positive light. Society tends to respect and admire those things which go along with economic success.

Shifting the behaviours of those who are most responsive to social cues could cause rapid cascading effects and make certain costly showing-off behaviours feel ‘cringe’. Paradigm economic interventions have cultural effects.

Share

These kinds of interventions have worked before.

Economic factors – albeit undirected ones – are also the main reason for past increases in fertility rates, as documented in work by the amazing new group Boom. The Baby Boom came about when it became more affordable to get married and start a family, plus giving birth became less risky to the mother’s health, and new technologies made it easier to raise the children you did have.

But deliberate economic changes have worked too. South Tyrol, a region in the German-culture part of Italy, had a birth rate that was about average for Italy, and well below German-speaking countries before 1990. From that point they began making a few small, but consistently administered, changes that mixed helping shoulder the cost of children with showing that the state (and by extension, society as a whole) respected parents and mothers. By 2020 it was seeing vastly more births than anywhere else in Italy, and even beating the German-speaking world. 

Boom

The province defying Italy’s birth dearth

In March, the Italian equivalent of the ONS, ISTAT, reported that Italy’s total fertility rate (TFR) – the average number of children per woman – had declined to 1.2. Nobody was surprised – Italy’s birth rate has declined almost every year since 2010. In fact in 2020, the Italian population declined by 384,000. That’s like if all the residents of a town…

Read more

2 years ago · 4 likes · Boom

France has bucked decades of fertility declines and also has a raft of policies that help people have children: lots of housebuilding, tax rates that are much more generous for parents than for childless couples and singletons, and much more.

According to the French national demographic institute, their healthy birth rate largely comes from high native fertility, rather than being driven up by immigrants. Also, French fertility doesn’t appear to correlate especially well geographically with current migration flows. Lyman Stone has also shown that those with three or four French grandparents are having practically the same number of children as those with zero.

But France doesn’t collect statistics on French people by ethnicity or religion, so it’s not possible to completely rule out the possibility that the children of migrants who follow different religions are partly responsible for French fertility.

Subscribe now

Explaining the ‘cause’ of the long-term and recent downward trend in fertility  is instinctively held by many to be important. First we should understand why we’re here, the idea goes, and then we can reverse whatever brought us here. 

I think this perspective is wrong. Every social trend is massively ‘overdetermined’, meaning that there are hundreds of (theoretically adjustable) factors pushing it both up and down.

‘Why has British growth been slow since 2008’ is a complex, abstract, philosophical question. Whereas ‘what could we do to raise growth’ is a simple question, which doesn’t require taking a stance on why growth has been slow. You could raise British growth by relaxing constraints on the supply of housing. But that doesn’t imply that Britain’s growth has been slow over recent decades because of a lack of housing (though I think housing is a part of the story, of course).

The very high fertility populations do achieve their rates through retaining ‘cultural’ elements that we lost. They jettison nearly everything that typical Westerners hold dear: personal space, possessions, gender egalitarianism, rationalist schooling, expressive clothing. In some sense it’s true that our low fertility comes from not living in this way. Most would not consider it desirable to go back.

But there are cultural changes we could make to raise our fertility rate – it’s just that they are also economic interventions too.

Consumer Surplus

One weird trick to get data centres built

Labour has pledged to make decisions on important infrastructure projects, like data centres and prisons, at the ministerial level, instead of letting the local council decide. The hope is that this will allow them to approve these projects even when there is a lot of local opposition…

Read more

2 years ago · Sam Bowman and Ben Southwood

I also explained how fragmented ownership leads to ugly and run down high streets, because no self-interested landlord has an interest in picking up litter or beautifying facades, as those benefits spill over, and it’s easier to free ride. Consequently, the solution is either single ownership (like a shopping mall, which are all car- and crime-free) or some sort of property taxation. Both are types of land value capture. I also did a popular thread on this.

The Works in Progress Newsletter

Fixing retail with land value capture

Don’t miss Issue 15, which we released just over two weeks ago…

Read more

2 years ago · 47 likes · Ben Southwood and Phil Levin

I didn’t write this one, but I did edit it, and I think I contributed to how well it came off. It’s by Nan Ransohoff, the legendary, swashbuckling leader of Stripe Climate’s Frontier. And it explains how you can start your own Advance Market Commitment. See my thread on it (also popular!)

The Works in Progress Newsletter

How to start an advance market commitment

Don’t miss Issue 15, which we released two weeks ago. We have had hundreds of applications for Invisible College, our week-long residential seminar in Cambridge in August. Applications close at the end of today if you want to add yours, or know someone who does…

Read more

2 years ago · 23 likes · Nan Ransohoff

And you mustn’t miss Works in Progress’s Issue 15, which was out in mid-May. I find it impossible to pick favourites from any of our issues after Issue 05, when we got a bit more time to devote to working on it, but I think this one is easily one of the best. Check out the official WiP Twitter account thread introducing all the articles.

There is a second common narrative, which holds that these so-called NIMBYs oppose new development near them because it causes concentrated harms, such as congestion, competition over state-provided resources, noise, and disruption. They are able to organise, such that even though its dispersed benefits – infinitesimally better housing for everyone in the country – outweigh these concentrated costs, opponents of development are able to win and block building and keep zoning or planning rules strict.

This story is also false, because these benefits can be concentrated as well, in the form of a single developer or landowner who ‘captures’ them all, to pay off the local objectors. If the benefits of development outweigh the costs, this developer should be able to pay them enough to get nearly all of them to say ‘yes’.

Share

Let’s illustrate this with an example.

Lower Manhattan (the bit South of Central Park, including Midtown, the Financial District, and everything inbetween), is about the same size as Palo Alto in San Francisco excluding its parks and nature reserves: around ten square miles.

Although they are about the same size, as you must surely know, Lower Manhattan is incredibly densely built up, with an enormous supply of offices, hotels, restaurants, bars, artist studios, shops, and more. The daytime population of the whole of Manhattan is about double its nighttime population: the 24,000 people per square kilometre who live there. By contrast, Palo Alto sees very few visitors, and has ten times fewer people per square kilometre, because it is largely one-storey and two-storey suburban homes.

Property prices average $1,800 per square foot in Palo Alto. Building costs in San Francisco are about $400 per square foot (the most expensive in the world by some metrics). But to get to (East) Midtown Manhattan floor area ratios of 14.3 (14.3 means that the building has 14.3 times as much floorspace as the plot of ground it uses), we’re going to have to build taller than most Bay Area construction, so let’s call it $700 per square foot. Midtown East is about 60 percent covered in buildings if you include roads (each 18 hectare block is 11 hectares of buildings, three hectares of gardens and courtyards, and four hectares of road and sidewalk space), meaning that the net floor area ratio could be about eight.

If there was no price response, there would be $1,100 profit per square foot. Duranton and Puga’s estimates imply that several million (probably tens of millions) of people would migrate to the Bay Area from other parts of the US if housing was more affordable there. I think adding around 500,000 people to get to Manhattan’s population density, while adding a huge amount of office, lab, and retail space as well, would by no means be enough to drive prices to the cost of construction, but let’s assume that they just do, so the last block profits by $0.01. That (a 45 degree demand curve) means that the profit per square foot will be $550.

Multiply that up:

• $550 of profit per square foot, times

• A floor area ratio of eight, times

• 10 square miles in square feet – 278,800,000

You get $1.2 trillion in profit after building costs, which includes the sorts of basic infrastructure developers usually provide. I think we need to put in some expensive transport infrastructure: let’s say we build 30 miles of subway, about what Manhattan has in that area. We can build it cut and cover on freshly demolished land, so it would likely be quite cheap, but let’s assume that we have to do it at BART prices of $2 billion per mile: total of $60 billion. Let’s say we also need to tunnel 30 miles of road at the exorbitant costs of the Fort McHenry Tunnel (about $1.5 billion per mile). That’s another $45 billion.

So we have about $1 trillion left over. There are about 26,000 households in Palo Alto. That’s enough for $30 million per household who lives there plus a $220 billion contingency fund.

So if Palo Alto decided to sell off its entire land to a developer, then upzone it to Midtown East levels, they could give each household $30m (or a share in the development worth that much), build America’s second-biggest subway network and longest road tunnels, and have $220 million left over for contingencies. The average property in that city is worth about $3 million. Who would prop up their $3 million house price at the cost of losing $30 million?

Subscribe now

In short: tunnel-boring machines got a lot better, and as people got wealthier they got a lot more sensitive about the huge disruption caused by building cut and cover.

In typical Potter style you learn along the way how disc cutters were invented and work, and how fast Isambard Kingdom Brunel’s father Marc built his tunnel under the Thames at Rotherhithe (about eight feet per week).

I think Brian is right: building cut and cover has serious political issues that have not been addressed. However, I am a little more optimistic than he is that they could, theoretically, be addressed, partly by parallel to housing.

Share

Europe and North America have both seen very little urban intensification since the second world war. Where cities have grown, it has been almost entirely through outward expansion – for example two thirds of Paris inside the peripherique is surviving fabric from before the second world war, and the population there has fallen from nearly three million to a little over two. But the broader built-up area has almost tripled in extent, and the metropolitan area has doubled in population. City population densities have fallen around the world. The heights of the tallest skyscrapers in the world have barely grown in the entire postwar period.

One conclusion to draw would be that building upwards is impossible due to our wealth and greater intolerance of the disruption such development causes. Sprawl is inevitable. But through changing institutions, people around the world have managed to make intensification acceptable to the public, or even popular.

Houston has added 80,000 townhomes on existing low density neighbourhoods. Washington, DC, has densified its metro stations. Israeli apartment owners have opted into demolishing their own properties and rebuilding them. New Zealand managed to upzone a significant fraction of the entire country by exempting the most precious historic neighbourhoods. Fifty-thousand accessory dwelling units have been added to California backyards, and ADU builders expect half a million more over the next ten years.

Vancouver’s Squamish people, a First Nation there, have been completely explicit about how they have been able to bring cost down substantially by junking burdens like affordable housing (they are able to avoid normal Vancouver bylaws if they wish), when 87 percent of them voted to add 6,000 apartments and 200,000 square feet of commercial space on 12 acres they own next to the center of Vancouver. They said the project was designed to ‘maximize economic benefits for the Squamish Community’.

Lots of the costs and complaints associated with modern development are cynical exploitations of institutions by groups that have been given vetos. If whole communities could choose for themselves (and keep the money saved) they would often choose a more cost-efficient and quicker option, even given its attendant costs. After all, HS2’s tunnels cost £700m per mile.

What would this institution look like? Choosing the exact thresholds is a bit of a lumping/splitting question – there is no perfect answer. For cutting and covering an urban road, I propose that the developers get majority (or perhaps supermajority) approval from residents living on that road, or living up to 50 metres away on any road feeding onto that road.

For cutting and covering the A2 between Vauxhall and New Cross, for example, that would mean asking about 9,000 people in the polygon above (created with Tom Forth’s wonderful tool).

Of course, these would not be the only people affected by closing such an important spine of South London. With current technology, it might be possible to identify and ask the other affected people – though perhaps automatic numberplate recognition systems could also tell us who drove the road regularly over the past three years, and we could get their approval too.

Such a system could have other benefits: it could be the starting point of an extra way to approve infrastructure, going directly to the voters. This is done widely for funding infrastructure. For example, American counties and cities often ask their voters to approve higher property taxes to guarantee bonds they issue – for billions of dollars – to build new roads, pipes or hospitals. This mechanism was exploited to its greatest extent in New Zealand, where such bonds built the Auckland Harbour Bridge.

Directly approving funding with the voters helps with approvals, because it’s inarguable that the public wants it to go ahead – it has a powerful mandate. It might be possible to take this even further and carve out much of the approvals process with direct public votes. If this did happen, then the mechanism could be a way around our infrastructure morass in general, even for projects that do not involve cut and cover or extreme local spillover costs.

Subscribe now

Of course, I could be completely wrong, and tunnel-boring may have improved so much that cut and cover plus compensation is always more expensive. But the sample of projects is so small, messy, and distorted that I think we should find out first.

As I said before, roads satisfy a latent demand for journeys, make housing cheaper by opening up much more land for productive development, support economic agglomeration, and helped start the Industrial Revolution. It’s a mistake to try and come up with motivated reasoning to argue that roads have no upsides at all. 

The lesson we should take from this is not, however, that roads are always the best transport mode, that historical schemes to bulldoze our cities for them were good ideas, and that we should push against every scheme for pedestrianisation and reducing traffic in city centres.

The first issue is that cars take up a lot of space relative to the person they are transporting. Online urbanists have pointed this out so many times that you might think it's a canard, but it is important. Including stopping distances, cars take up the space of 20 people.

This means that if lots of people want to do exactly the same trip, then despite the fact cars have engines and can go pretty fast, they're often slower at moving lots of people than just walking, as you have to squeeze so much extra material through the available space.

For example, all of New York City’s bridges carried more people per day when they were devoted to transit and pedestrians, usually about 100 years ago, than they do now when all or nearly all the spaces has been given over to cars. The Brooklyn Bridge had 426,000 crossings at its peak time vs 178,000 now.

Share

Or for another example, London has over 2m railway journeys per day and 5m tube journeys on the 1% of its land that is railways and stations. And there are still some lines with spare capacity if the trains could be found to run. By contrast there are about 6m trips per day (including both car passenger and driver) on the 12% of its space that is road – slightly fewer journeys on twelve times more land (though buses and bicycles use the space too). 

Setting aside roadspace in the centre of cities at ground level does not just spend it on that use, nor just make the area polluted and noisy, it also prevents other modes of transport. Walking around Manhattan feels endless, because of all the waiting to cross roads.

And this is all much worse if we expect to store cars while parked as well – as opposed to say, having driverless cars or taxicabs that are always carrying passengers around rather than sitting idle, because we use a whole bunch of extra land up for that storage.

Cars have also got much bigger over time as we have got richer (on the demand side) and more efficient/productive at making them (on the supply side). This makes for a much nicer drive, but it worsens the space costs of driving, and these bigger and heavier cars are somewhat more dangerous for everyone who’s not in them.

For roads like motorways, which are mostly in the countryside, this doesn't really matter. Road links between cities are very important, and the alternative use of the land is usually just farming, worth about £10,000 per acre (£25,000 per hectare). We have 16.843 million hectares of farmland in the UK, and it covers 69% of the UK, and while it is very important in total, we can generally spare a few hundred acres of it for a motorway. For much less densely populated countries like France, Spain, or the USA, this goes even more strongly.

Land in cities, by contrast, is very scarce, especially in the centre. Land in Westminster, London's most expensive borough, averages £93 million per hectare – 9,300 times more than farmland – and that's with today's planning constraints. A plot of land in Westminster without any planning constraints would go for quite a lot more. 49 Park Lane went for £3,300 per square foot. At 55 storeys (like the Hilton down the road), 1/2 hectare would have about 1 million square feet of usable space, £3.3 billion worth, and about 300,000 times more than farmland.

So land in the most prime areas can be nearly a million times more valuable than land outside cities. This is of course, why we build upwards when land is scarce, and we build at low density in low value areas, since we have ample cheap land. (The early sections of this recent spatial economics paper from New Zealand explains this economics of land very carefully.)

This is also part of the reason why urban railways have often been shifted underground. (The other part of the reason is the second half of this article: the noise and pollution externalities railways generate, especially steam railways.)

This all suggests we want to direct driving around the centre of cities with bypasses and ring roads. Roads in the centre should mostly be for access. Where possible, arterials should go underground as they approach the centre, just like metros/tubes and utility cables.

Take Rennes (France) and Bristol (UK), which have core cities and metropolitan areas of roughly the same population. Rennes is neatly ringed, Bristol has major roads driving right through the centre. My story implies the Rennes approach is superior. Rennes actually has more lane-miles of road than Bristol in total, but this is mostly segregated from where people live, work, recreate, and so on.

So the first problem with roads and private car travel is that they are extremely space inefficient, meaning that they are inappropriate in the centre of cities.

Notice anything about this fine-grained map of Nitrogen Dioxide pollution? Things get worse as you get to the centre, but above all, pollution is driven by roads. The same is true for PM2.5s and other air pollutants. Air pollution is really bad for brain development, lung health, and more. And road use is the number one cause of air pollution in the UK (as in most other developed countries) – which everyone knows intuitively from tasting and smelling the air near roads.

Hopefully, electric vehicles make a big difference to solving it; tyres and brakes mean they will not completely eliminate pollution but I gather they will roughly half it. Sadly, however, they will make very little difference to solving noise pollution, as EVs make a very similar amount of noise above about 20mph.

People hate noise, and this is reflected in house prices. One 1976 paper found that house prices got higher the further you were away from a major new road – the I-495 in North Springfield, Virginia – for the first 1,150 feet, as the cost of the noise and air pollution dissipated, but then began to decline as the benefits of access to the network declined.

A more recent Swedish study found that each decibel of extra road noise cut 0.6% from a property’s value – being in the loudest places cut a full 30% off its price. A comparable study for Glasgow, UK found similar, but smaller, results. An ingenious study pairing two million 100 square metre units, and sound measurements there, with 100,000 property sales, in the Netherlands, also found there was a very large impact from noise.

We can back this up with evidence from aeroplane noise too. According to a 2009 study of 66,000 house sales in the Netherlands, each decibel of aeroplane noise a property is regularly exposed to above 45 decibels (roughly the volume of light rain), wipes about €1,900 off the property’s value. Different studies find different exact values, in part due to differences in methodology, data availability, local context, and so on, but results are fairly consistent: some 0.5% to 3% is taken off house prices by each extra decibel. 

No one wants to live on a through road – in addition to the noise and the pollution, it’s generally not safe for your children to play around in the street. Apparently, residents on cul-de-sacs tend to lend their neighbours tools more, and are more likely to go over to their neighbours’ for tea.

All this together is why homes on cul-de-sacs tend to sell for a lot more. One study found they were worth 30% more than other similar properties. Thus, if allowed to do so, developers will invariably develop all new neighbourhoods as cul-de-sacs, to prevent people from living with traffic going by their house. Great Estates always blocked through traffic if they could.  

All this means that we should consider the massive externalities of roads, as well as the benefits they give us. We have two ways of doing this: first, we can try and make sure most of our important roadspace is not near where most people spend their time; and second we can try and reduce the harms that road use has on those who live by it.

How can we keep traffic away from people? Two major options we’ve gone for in the past are bypasses, which have been built to allow traffic to go past a town or a village so its ancient high street need not be full of whizzing vehicles, and ring roads, like the M25, which are effectively the same thing, but from all directions. (Though not all ring roads are created equal.)

When you have a credible bypass, then even without significant driving restrictions in the centre, you will see most of the traffic just go the quickest, most convenient route on the infrastructure meant for it. This is absolutely clear to me from driving around Spain.

If paired with sensible restrictions on through traffic in the centre of these towns, villages, and cities, ring roads and bypasses can give us a win-win: good driving infrastructure, but with most of the traffic happening relatively far away from people’s houses. In fact, this was the principle behind the 1963 Buchanan Report – Traffic in Towns.

Another option is low traffic neighbourhoods, which restrict through traffic in residential areas to main arterials, to prevent ‘rat running’ through minor streets. Effectively, they turn minor streets into cul-de-sacs for cars, but leave them as through streets for cyclists, pedestrians, and emergency vehicles.

I think LTNs are generally a good thing, because the benefit to the people on the closed streets outweighs the cost to the drivers who have worse commutes from the LTNs on their routes.

But right now, these drivers have a very fair case that they are being disadvantaged by the extra traffic they face. It is always distressing to have the rules of the game changed in a way that disadvantages you. I think the fair way to solve this problem would be to build more and better major roads so that they can get around as well as before – but in a way that allows us to massively improve residential streets. I honestly think that such a win-win approach could make a lot of the LTN controversy go away.

A third way to move traffic away from where people live or hang out  is tunnels. We should be tunneling many more roads. The tunneled road could be wider than its surface-level predecessor  – happy drivers – and we could even fund the tunneling by building homes on top of the land. And the pollution and noise has been solved. But this is a whole topic of its own, and I intend to write much more about it.

In addition to moving traffic off the streets on which people live, how can we reduce the harm of the roads we do have? The first thing is to try and make traffic less polluting. The best way to do this is just to wait, as electric vehicles continue to improve massively every year, and pollute less than half as much as regular cars. They will hopefully take over. And at 20mph they are nearly silent except for their beautiful metallic safety drone, meaning they will also vastly reduce the nasty noise impact of traffic.

But it will take decades for our ‘fleet’ to completely shift to electric, and my lungs are suffering right now. So one thing we can do that has a more rapid impact is impose restrictions on which cars can come into the centre of cities. This is the principle behind the original ULEZ, and it’s a good one (I say this as someone who lives right on the boundary of the zone before the recent expansion, and who paid a great many fines before I eventually had to buy a new car to meet the standards). I’m not so sure about the massive and controversial recent expansion of the zone (pictured below), as it reaches areas that are not nearly dense inner city, and so the negative impact of emissions and noise there is correspondingly much smaller.

Another thing I recently learned that could make a huge difference, is to put noise barriers on major roads. Apparently this is common outside the UK, especially on viaducts. They are widely used in central Europe, particularly in Poland. California has 760 miles of them and Florida 250. 

In the UK, it has been implemented on and off around the strategic road network, e.g. around Luton on the M1 and Cambridge on the A14. Grow some vines to cover them up and they even improve the view. Or put solar panels on top. We could start with the Westway and other major roads that go into cities. 

Many sound barriers claim reductions of 15 decibels or more which is approximately the difference between nearby heavy traffic, and the sound of rain, or a normal conversation in the background – which would be night and day near the UK’s major roads.

We should also be enforcing our existing laws against excessive road noise, especially motorbikes with illegal exhausts, ideally with automatic radar cameras as in Paris.

These two facts: that roads are essential infrastructure, but also cause harm to those nearby, suggests there is space for a win-win political compromise. More roads in the right places, fewer cars in the wrong places. More agglomeration, less noise where it matters.

It’s clearly not impossible: the Netherlands uses about 3x as much land as we do, per capita, for roads, but has legendarily quiet and safe neighbourhoods. These two things are just not at odds. I think there could be massive political rewards to politicians who could come up with the deft policies that square this circle.

Subscribe now

Subscribe now

1. Induced demand is not a useful concept – the demand is always there, latent

There is a popular idea that if you build extra road infrastructure, it will just be filled up with new traffic, and thus, this extra roadspace hasn’t done anything to relieve traffic congestion. This theory is known as ‘induced demand’. Strictly, it is false, and we shouldn’t let it stop us building more roads, as doing so would make us richer and happier. However, this doesn’t mean more roadspace is good in every location. In fact, there are many places, like residential streets and city centres, which may have too much driving.

In a sense, the claims behind induced demand are true: in many cities there is such latent demand for driving that building extra road infrastructure – and charging nothing to those who wish to use it – will lead to no reduction in congestion. 

Famously, RJ Smeed predicted in 1949 that London’s average road speeds would always be roughly 9mph, which they remained at for seventy years after this prediction. Any slower, and people took the tube, cycled, walked, or stayed at home; any faster, and more drivers flooded onto the road until it came back down.

But in another more important sense it is not true: more roadspace means more journeys. Even if the road looks just as congested, four congested lanes are carrying more people to their desired destinations than three or two congested lanes. One study of widening the M25 found that journey times worsened over the same period. But even in this study, the road let over 10% more people through the relevant junctions. (In fact, roads in general were seeing worsening times, and M25 speeds would likely have worsened more if not for the expansion.)

The demand is not really ‘induced’: it’s more true to say that it was always there – when the roads don’t exist to accommodate it you could call it ‘suppressed demand’. If we offer more subsidised social housing, more people tend to live in it – would we then say that demand for social housing has been ‘induced’?

Imagine a city with acute food shortages and controlled prices (e.g. Caracas, Pyongyang). People join long queues before bakeries daily, and still only average half their calorie needs. The supply of bread then improves so that they average 60% of their calorie needs instead. Would we really expect the queues to wither away? Of course not. They might not even get discernibly shorter. Does this mean that demand for bread is in some peculiar and distinctive sense ‘induced’? Of course not. Does it mean that going from 50% to 60% of calorie needs is not an improvement? Again, of course not.

2. More and better infrastructure lets us spread out and enjoy more space

A sophisticated version of the induced demand story says that as well as driving more on a better road network, people faced with a better road network, over time, shift towards living more car-dependent lifestyles. This is true. When infrastructure is reliable, people begin to depend on it. As a train commuter, I live a train-dependent lifestyle and when Southeastern trains (currently run by the Department for Transport) recently reduced the frequency at my station by 2tph I was incensed. If the trains had never existed, I would never have moved to Zone 3. 

When our best mode of transport was walking, cities had a maximum size. Once things got more than a few miles – an hour’s walk – wide, it was beyond the distance where agglomeration was possible. So if there were sufficiently big advantages to the city – as in Ancient Rome, medieval Constantinople and Paris, or Georgian London – people built apartment buildings up and squeezed more and more people into the existing space. 

Life was horribly overcrowded. In 1910, the densest tract of New York had around 150,000 people per square kilometer. Overall Victorian Manhattan had about 60,000 people per square kilometer across the city – as densely packed as its densest district today, and higher than famously crammed Dhaka today. 

This was all lifted with infrastructure like the horse-drawn omnibus, and commuter train lines (Greenwich to London Bridge was London’s first – the world’s Dublin to Kingstown). And, yes, the car. Because cars can take you point to point and go anywhere you like, they are much more land-hungry than transit, but they also open up much more land for useful development. If you look at Atlanta on a satellite map it barely looks like developed land at all, but instead like a gigantic partly-inhabited forest – it has essentially achieved the goals of the Garden Cities Movement of Ebenezer Howard.

If house prices are above construction costs, that means that developers could profitably build, but they’re not being allowed, and it is a strong indicator that there is an undersupply of housing. In eighty percent of American cities, house prices are roughly equal to construction costs, suggesting that housing is relatively abundant. This in turn means it is generally affordable in these places to get married and have children. They also tend to have low homelessness. And in all of these places, it is because the city makes it easy to build outwards in all four directions.

People like space, and although things like crime push people to spread out more than is efficient, even in near-zero-crime societies with tolled highways and expensive parking like Japan, nearly everyone wants to own a car, and likes to spread out into quiet suburbs if they can. 

So, yes, if roads exist, people plan their lives around them and use them more. But they do this with every technology. This is just how demand works – it’s not a special type of ‘induced’ demand.

So induced demand is not a helpful concept. We use roads because we want to get around, which we want whether or not roads exist. The upshot is that road infrastructure is good. Overall, for example, the US interstate highways are responsible for about four percent of US GDP. 

The UK’s decision in the 1700s to allow private organisations to take over 11,000 miles of roads, charge tolls, and use the money for improvements, which led to ‘turnpike mania’, increased average road speeds from 2.6mph to 6.2mph. It, along with the canal boom at the same time, was surely an important ingredient into the industrial revolution.

3. Transport is very important, and roads can provide a lot of it

When we get in our cars to go somewhere, it’s because we value doing so – because we can go to work, to pick up something we need, meet a friend, or visit a family member. Roads are also an essential way we transport goods around and get emergency services where they are needed. We have an enormous latent demand for trips, and the more infrastructure we build, the more trips we do. 

On top of the value you get from the roads you actually drive on, there is a bonus value to having infrastructure out there that you could use. You don’t know whether you will want to visit an Ethiopian restaurant, or get job on a specific side of town, regularly, or even ever, but you will likely still value this option.

Infrastructure owners, such as airports, can profit from this. But even if we do have tolls or road pricing it’s hard for those running the roads to charge you prices that fully reflect this benefit you get, since they can only charge you for the trips you do take, which means that the value these options create is usually ‘captured’ not by infrastructure operators, but by house prices, because people pay more to live near infrastructure which offers lots of potential.

Airports show this very clearly. You will take only a tiny infinitesimal fraction of the potential flights you could take from your nearest airport. But each year you enjoy having the option of taking them, since you don’t know where you might want to travel in future. Thus, airlines and airports capture only a tiny fraction of the value they create, and instead of paying airlines people pay to live nearby airports. 

Living near Heathrow gives you many more options than living near Stansted, and living near Stansted is worlds apart from living near a small regional airport. If you are not in the flight path, every 1% further you get away from a major airport reduces your house price 0.15%. 

The same is true for the road network. Roads increase the pool of jobs workers can get to in a given time – or on the flipside, the pool of employees a given employer can hire. A larger labour market means higher productivity. Homes with access to these higher productivity jobs are worth more. One Norwegian highway linking up villages in the South raised prices 5% across the region. Every mile closer homes were to a pair of toll road networks in Orange County in the 1980s was worth thousands of dollars in home value. And so on.

Sufficient road infrastructure can even allow what we might cheekily call a ‘15 minute city’ – pick anywhere in Lubbock, Texas and any other place and see how you can drive between those two places in 15 minutes.

Although American cities – with more roads than nearly anywhere else – should have the worst congestion, through induced demand, they actually enjoy large agglomeration benefits, and allow people to get around pretty well. It would almost definitely be more efficient to provide a little less road infrastructure in some cities, and a little more transit, and to use congestion charges as well, to relieve peak time traffic, but the system’s success proves it basically works.

Share

All of this doesn’t mean that we should bulldoze the best parts of our cities to build roads, as London attempted in the 1970s. Car traffic is noisy, it pollutes, it’s space-hungry, and it can be dangerous. This means that we may end up favouring things like low-traffic neighbourhoods and pedestrianisation – in the right places. We may also favour tolls or congestion charges. But equally, we should not use flawed arguments to try and pretend that roads are not an important form of transport that support our prosperity.

In the UK, our roads are horribly congested, and compared to competitors – such as for example France – we have many fewer and worse quality ones. Building them is insanely expensive, in part due to mistakes we have made around judicial review. We should do something about this.

One: we don’t have enough infrastructure. Commuters across the country face congestion and overcrowding; we haven’t built a new reservoir since 1992 and it’s blocking the delivery of new homes; new wind and solar farms can’t get plugged into the grid. 

Two: the infrastructure we do build is far too expensive. Just the planning application for the Lower Thames Crossing cost more than Norway’s longest road tunnel; Britain’s cheapest tram project was double the price of a typical tram project in France, and only just cheaper than France’s most expensive; HS2 is over £200m more expensive per kilometer than France and Italy’s latest high speed rail projects.

These problems are related. One major reason we don’t have enough infrastructure is that it costs a lot to build it, meaning that the Treasury – speaking on behalf of the UK’s taxpayers – is unwilling to fund it. One major reason it costs so much to build infrastructure is that we do it so infrequently, meaning that local governments have little experience of procuring it and we often lack the workers with specialised skills to deliver the projects.

This relationship hints at the dark secret of infrastructure costs: it is a feature, not a bug of the system. Infrastructure costs have got high because lots of people want them to be high. And fixing those incentives is the only robust way towards solving the problem.

Share

One story of NIMBYism, which I have used myself in the past, says that it arises when the benefits of something are dispersed but the costs are concentrated. Then, the argument goes, locals will inevitably oppose development nearby, be it housing, or railways, or roads, or reservoirs, due to the disruption, shade, noise, or traffic. Beneficiaries, by contrast, are spread all over the place.

More flights from Heathrow benefits literally millions of people around the country who get to enjoy a better labour market in London, or work for airlines, or in tourism, or just get to enjoy lots of convenient flights to a wide range of destinations. If you are not in the flight path, every 1% further you get away from a major airport reduces your house price 0.15%.

But if you live in the flight path, then flights from Heathrow cause horrible noise pollution. It’s much worse for any individual person in potential flight paths than it is better for any of the beneficiaries of a bigger airport and stronger South East economy, so it’s easier for the objectors to organise and block the third runway the country has been planning for seven decades.

The problem with this theory is that if doing something is overall socially beneficial, by definition the benefits of doing it should be larger than the costs to the objectors. In principle you should be able to share some of those benefits with the objectors until the benefits outweigh the costs for almost everyone, except perhaps some tiny group of diehards. That may not literally mean handing them cash, but could mean building them new roads, buying them new houses, laying them out a new park, or adding buildings to their local school. There should be a package that locals would approve.

Lest this seem unlikely, consider the Squamish Nation, 87% of whom recently voted to approve 6,000 flats and 200,000 square feet of commercial space in towers on 11.7 of the 350 acres they own abutting central Vancouver. Vancouver bylaws do not apply to the Squamish reserve, despite its location, and this was confirmed by the Supreme Court. Deciding for themselves, the Squamish did not impose affordable housing/inclusionary zoning, saying ‘the project was designed to maximize economic benefits for the Squamish Community’. They don’t seem to have replicated any of the other requirements Vancouver requires either.

Consider also estate regeneration. In 2018 the authorities in London changed the system for redeveloping social housing. Those intending to redevelop the estate, usually at a much higher density, would have to get approval from the residents themselves. If residents approved, it would generally be waved through by the planning authority. Since then, there have been 31 ballots on 30 schemes: 29 out of 30 were won the first time, many by huge majorities. The 30th was lost narrowly then the offer was improved and the ballot was won.

How did the providers win the ballots? They massively improved the design quality of the estates, turning concrete into bricks and restoring streets. They also gave every existing resident much nicer and larger properties. And they paid for it all by selling off the extra flats and houses they produced by increasing the density of the site which the locals were happy to approve. For example, Kidbrooke Village, with 5,300 homes, replaced the Ferrier Estate, which had 1,900.

Compare with HS2. With HS2 the locals had no power to block or approve any particular proposal, and the developers had no ability to go directly to the locals and make a deal to get approval. What do rational locals, especially the diehard opponents, do at this point? They push on every lever in the system to try and add as much cost as possible to the projects, aiming to ideally stop the project, but if not then at least massively delay it. In the UK they usually succeed with both.

In the UK planning system there are many ‘veto players’ who either legally or practically must give their say-so for a project to go ahead. If you get to your approval decision with the Planning Inspectorate or the relevant minister, and some element of process has not been followed, the diehard opponents of the scheme will make sure this is raised. So each project has an incentive to massively increase paperwork, assessments, newt surveys (not a joke), and consultations – the planning equivalent of defensive medicine, using every test or remedy in the book just in case.

How did these veto players get there in the first place? The same mechanisms. There is no support around the country for a systematic plan of cracking down on locals everywhere. Everyone is a local somewhere. When Robert Jenrick proposed to ‘tear up the planning system’ and it became clear that this meant not just a concentrated dose of housing in a few important places but an increased supply everywhere, his scheme was toast. As was every previous postwar attempt to increase housing supply by forcing locals to accept it more.1

Top-down force actually makes things worse, ratcheting the effect upwards each time. Every time the government attempts to steamroller infrastructure or housing through, there is support for added process, especially slow and expensive discretionary process, to make sure everything is being done ‘properly’. The justifications will always be made in socially acceptable terms, like the importance of biodiversity, bats, nitrogen levels in rivers, affordable housing, and so on. But the support for these rules generally has a deeper, underlying, and personal motivation.

So delays and process dramatically increase the cost of projects, meaning there is nothing left over to compensate the other locals – those who are not diehard opponents – for the real, but negotiable downsides they experience from the development. In terms of the Squamish Nation, they have had to do so many environmental assessments that they can’t make a substantial profit from developing precious ancestral land, so all the potential moderates tip towards opposing them as well. And as they tip towards opposing projects, they also support the cynical accretion of costs that we see across the system.

This is what David Foster and Joseph Warren call ‘the NIMBY problem’ in a brilliant paper from 2022. If we don’t cut out veto players, by offering locals a direct way to approve proposals from infrastructure (and housing) developers, then we will never be able to tackle all of the crazy cost additions that are preventing us from developing world-class infrastructure.2

Subscribe now

The simple supply and demand story about immigration and the housing market is appealing, intuitive, and broadly right. More immigrants in the UK means more demand for housing in the cities where those immigrants intend to live. Because the supply of housing in the UK is extremely ‘inelastic’ – it doesn’t respond to increased demand due to restrictions on development – this drives up house prices and rents. One government paper put this impact overall at 20% from 1991 to 2016. Since prices rose about 300% above inflation during that period, that would be a relatively small fraction of the overall rise.

Subscribe now

Scarce housing has all sorts of problems. But most of these happen through deterring migrants (internal and international) from coming to the city, so too many people coming can’t be a cause of these knock-on problems like innovation (of course it can still be bad for the natives who face more expensive housing!)

Recently, the UK has had a wave of over one million migrants coming in during one year (about five times the rate seen in the USA right now, close to the rates seen in Australia and Canada). Seemingly this rise above previous levels has been driven in part by pent-up student demand from the Covid years, in part by refugees from the Ukraine war and China’s takeover of Hong Kong, and in part by generous rules on NHS and social care workers.

This rapid spike in immigration may be the key reason why not only has there been a rapid worsening in the UK’s urban rental markets with many lurid stories of huge queues at viewings and massive over-asking-price offers, but that this has spread beyond the usual hotspots of Cambridge, Oxford, and London. Cities like Glasgow with important universities, and which have typically faced very low property prices by UK standards, have seen rents spike – one wealthy American student was apparently forced to buy a flat as rental property was so scarce.

How does higher demand for housing affect Britons? Around two-thirds of the country owns their own home. There are also 4.4m landlords, and 95% of these are British. For two-thirds of the country, plus the landlording six percent, this higher demand is mostly good, unless they are eager to upsize. For the other third, it is probably bad. It is definitely bad for private tenants, who face higher rents (they lose to the same extent landlords win). Social tenants generally have very secure tenancy, meaning existing tenants may not be heavily affected. But for those prospective social tenants who miss out on a flat or house due to an incoming immigrant taking their place it is pretty bad: as well as being secure, social housing is very heavily subsidised.

So two thirds of the population win, and one third lose. But how exactly does this happen?

Increased demand for housing leads to appreciating land values. But it is not really the land which increases in value, or even the structures on top of the land, but the permissions that allow farmland, worth a mere few thousand pounds per acre, to become developed land, worth millions per acre. Or, to say the same thing, allow low density developed land to become moderate or high density developed land.

So we drive up the value of these permissions by allowing many more people into our country without releasing more land for development. But the great thing is that these permissions don’t cost us anything to provide. We are effectively storing up value that careful housing policy could release. 

The more high-skilled, positive-externality-rich immigration the more, for example, the planning permission associated with a community land auction is worth. Giving planning permission to a single 10 or 20 acre field near a suburban station might be worth tens of millions of pounds for an affected local area, which runs to a lot of potholes, council tax reductions, bus services, and new trees. This would be, in effect, a transfer from the migrants (and other renters) to the local community (likely shared with the landowner, developer, and national government, depending on how we release the money).

Of course, right now the planning system isn’t well set up for releasing this permit value. In fact, it never has been suited to doing this, as it is built around the idea that central planners should be able to tot up costs and benefits, make the socially optimal decision, and ride over political objections without needing to incentivise acceptance or compensate those who lose out. It flat out rejects win-win trades.

Once we release this permit value, all of the country should benefit, not least because we would also make housing and buildings – one of the most important factors of production – much more freely available.

Of course, you may well say that Britain has had 75 years of housing failure without changing anything or at any point building enough homes, the last time we did that was in a fundamentally different world, the 1930s. You would be right. But I have hope. We have only recently fully understood why the planning system fails to deliver enough homes, and come up with schemes, like community land auctions mentioned above, to capture the stored up permission value, and give it to the affected communities to unblock the system. These schemes are based on numerous success stories around the world, and are already influencing UK policy. It’s a matter of time.

Of course, if I am wrong, then the one third of society that is private renters, or those who would like to be social renters, but are blocked by a foreign tenant, can reasonably complain about losing out, though the other two-thirds who own land should be pleased. But as a whole, it’s good for UK plc if there’s more demand for the biggest asset we own: our land.

Share

1. Cul-de-sacs are streets with only one entrance to cars. They are popular with homebuyers and residents, because unless you live there or are visiting, there is no point driving in. Ths means they are quiet, the air is cleaner, and they are safe. This, in turn, means neighbours know one another better, and there is generally a stronger sense of community. Cul-de-sacs are popular with residents.

2. However, cul-de-sacs also add ‘circuity’ to a system. If they lack cut-through alleyways, then they make walking journeys extremely long (see diagram below), encouraging extra driving. By definition, cul-de-sacs generate roadspace that doesn’t accommodate traffic in the network. Cul-de-sacs, done badly, can generate traffic.

3. If a cul-de-sac does have alleyways, then it’s like a grid for pedestrians, but a cul-de-sac for drivers (would you prefer the 7 min drive below or the 3 min walk). This is the principle behind ‘Low Traffic Neighbourhoods’ in the UK, and ‘superblocks’ in cities like Barcelona. By closing streets to certain traffic, they make many people’s streets into what are effectively cul-de-sacs, with less pollution, more safety (e.g. for children to play in the street), and less noise. By adding circuity for cars but not other transport modes, these types of cul-de-sacs may not encourage driving. Why drive five minutes when you can walk it in three? Cul-de-sacs, done well, need not generate traffic.

4. Each individual LTN or superblock, when it’s introduced, closes various roads. Since people still want to get around, one intuitive prediction is that this just means more traffic on the remaining open streets. Evidence had appeared to reject this, appearing to show not just driving redistributed around the neighbourhood from the closed interior streets to the open boundary streets, but also reduced driving overall. However, close inspection of the evidence suggests this probably isn’t true. Because an individual LTN only affects a small fraction of most drivers’ journeys, it doesn’t affect their decision to drive much, and people tend to just shift their routes to the now-more-congested boundary streets. Each individual LTN largely shifts driving around.

Share

4. However, each LTN introduced likely reduces the total amount of driving in the city a small amount. When roads do not charge users for driving, the main regulation on how much people drive is the speed they expect to go at – this has tended to keep London road speeds around 8-9 mph for about 100 years. If LTNs reduce driving speeds on arterial boundary streets, LTNs would be expected to lead to reduced driving overall, even if each individual one barely affects local driving. This is the flipside of the standard result that in high income countries with no road pricing and moderate or high population density, more roads don’t reduce experienced traffic or increase road speeds, but they do allow more people to get through the equally congested roads (so-called ‘induced demand’). Thus, although each LTN worsens traffic on the boundary roads nearby, if every residential neighbourhood was an LTN, traffic on all boundary roads would be no worse than beforehand. A city with more LTNs is one with less driving.

5. The costs of traffic are likely nonlinear. Going up from a cul-de-sac to a through road appears to have larger effects on house prices and community than going from a congested main road to an especially congested main road. Plus, more people in total live on interior streets than on main streets. Finally, those who chose to live on a boundary street generally already opted for higher traffic, suggesting they are more tolerant to it. This suggests that just because traffic is shifted to other streets where people live, it could still be a net improvement. Traffic shifting is not necessarily bad, and can be good.

6. In theory, we could solve the problem of excessive traffic on boundary streets with pricing systems (in fact – this could improve traffic to be much better than before LTNs). This would be much, much easier (and cheaper) than charging road pricing on every road in the country, plus more defensible politically. Indeed, if all residential neighbourhoods had the option of being LTNs, then achieving fast flowing roads across the country with tolls would be much more straightforward. Rural roads could be exempted; leaving only motorways, ring roads, and arterials, like the boundary streets relevant to LTNs. In practice, it’s reasonable for those on boundary streets to complain that this may not happen. LTNs make road pricing easier. Road pricing could solve boundary road congestion.

7. Another way we could compensate those on boundary roads for the short term cost to them is a policy like halving their council tax. Since those on interior streets are enjoying gigantic benefits, this should be considered fair.  And it should make LTNs more resilient and lasting where they do indeed have net benefits. Since very few people live on boundary streets, this would be cheap. Residents of boundary streets of LTNs should be given discounts on their council tax.

   • As an example, Ashburnham Triangle in Greenwich, London is a historic low traffic neighbourhood, with through access blocked for decades. It includes at least 150 terraced or semi-detached houses, each worth over £1m, plus about 50 social and private rented flats worth about £300,000. Each has perhaps a fifth of its value tied up in it being low traffic – in total over £30m of proprty value (or £750,000 of rental value each year). Far fewer people live on the much more commercial boundary streets (though one of them is Liz Truss) – completely exempting the perhaps 100 properties from about £1,500 council tax per year would come out at far lower than the uplift on all the beneficiary properties. Of course, actually capturing this uplift to use it would be difficult with today’s institutions.

Subscribe now

My first love was macroeconomics. My first job was economics reporter at City A.M. during the Eurozone crisis in 2012. My second job was head of macroeconomic policy at the Adam Smith Institute, working on trying to update the Bank of England’s remit. Back then, during the double-dip recession – which was later revised away – we were concerned that without a change in the Bank of England’s remit, it would mis-perceive supply-side inflation as indicating an overheating economy and deliberately cause a recession by crashing total spending in the economy (which is equivalent to nominal GDP or aggregate demand).

We ultimately failed to switch the Bank of England’s remit, from targeting inflation to targeting nominal GDP. According to friends who are much closer to the internal workings, there was a brief period during Mark Carney’s early months as Bank of England governor when it was on the table as an option, and it might have come through. But alas it did not.

Subscribe now

Still, as much as I thought and think that nominal GDP targeting is the most robust policy out there when it comes to avoiding recessions, I wasn’t too worried. Firstly, that’s because it’s always hard to focus on macroeconomics when you’re outside of a recession. A slump focuses the mind and it’s hard to think of anything but macro. But outside a downturn the eyes glaze over. How can this really matter, you think to yourself. 

But secondly and more importantly, I thought that macroeconomic policymakers had learned the lessons of the Great Recession. I thought they agreed that flexible inflation targeting was flexible because it told you to ‘look through’ inflation that is clearly coming from the supply side – like a pandemic, major war, and gigantic energy crisis – and contract policy only in response to inflation when it’s driven by the demand side. It’s not always possible to know where inflation is coming from, of course, and it’s even harder to judge when core inflation – typically used as an indicator of the stance of monetary policy because it’s less affected by input prices than broader inflation – is running high too. So I wasn’t too worried. Properly operated flexible inflation targeting will have similar outcomes to nominal GDP targeting.

The Bank of England’s monetary policy report yesterday, however, has put the fear of God into me. Why have I been wasting my time thinking about irrelevant microeconomic supply side issues all these years since 2014, I think. I’ve forgotten all the finer points of macro. I partied during summer, and now winter is here. This week the Bank of England’s monetary policy committee raised interest rates 0.75pp in order to tame inflation. I think this is probably the wrong decision. But what worries me much, much, much more is that they are planning for dramatically below-trend total spending growth in the coming years (that is, aggregate spending across the whole economy, including the private sector and government spending).

Not only is the Bank expecting a long recession. But they’re expecting total nominal spending, which they control, to rise 3.3%, 1.3%, and 0.7% over the next three years – compared to trend growth of around 4-5%. This will make the unavoidable downturn from higher energy costs and other supply-side problems into a much deeper and longer recession where far more people lose their jobs than really have to.

Mike Bird cautions that this forecast is based on the market being right about the path of interest rates – which the Bank itself expressly disagrees with in its report. It seems to me that if the market disagrees this strongly with the Bank, its position is not likely to be very credible, and if they do have a credible position somewhere in there, they should be communicating it much more loudly and strongly.

In a basic macro model of the economy, total demand – the total amount of spending people are doing in the economy – interacts with total supply – the total amount of goods and services on offer. Total supply is determined by our technology, institutions, capital, skills, and resource inputs. Generally it increases slowly over time as we invent new technology – but it sometimes drops rapidly if there are natural disasters, wars, or big institutional dislocations like Brexit. 

Total demand is the amount of money we can all spend in nominal terms – either total spending, or total earnings, or total output, which are three sides of the same three-sided coin. This is affected by monetary policy: which simplistically is how much money there is in the economy to spend, which the central bank controls either directly by creating new digital money and using it to buy government bonds (gilts) or indirectly by changing its policy interest rate.1

Now, clearly, printing more money does not make us richer. It doesn’t make us more or less able to produce iPhones, or drycleaning services, or to redecorate homes, or to produce energy. It can even make us poorer. If we print too much money, especially in an unexpected way, then it makes it hard to decide between different product offers and different jobs, causing dislocations in the economy. People don’t take the best jobs, stay unemployed for too long, some businesses can’t afford to buy inputs or pay workers, and so on.

But printing too little money can make us poorer too, for symmetrical reasons. The economy is like a game of musical chairs. When it’s going well, there are enough chairs for all of us. Playing with too little money is like playing with too few chairs. In economics we say that wages are sticky downwards, i.e. that people don’t like to see their numerical wage fall. The labour market is the most important market in the economy. If we expect there to be a certain amount of money in the economy to settle our contracts with, and then there is less, money will stop doing its job as a common mediator of value as various people will value it differently. Markets will stop clearing. We will get unemployment. Firms will hold off on investment.

The market will fix this eventually, but only after a completely unnecessary period of adjustment where we redenominate all our prices and contracts under the terms implied by the new growth path of spending, rather than the old one, i.e. a recession or even a depression.  This is the recession that the Bank of England yesterday has said it wants, in order to get prices down (for its own sake). This will mean hundreds of thousands or millions of people out of work, and therefore a loss in all the stuff they could be producing: reduced consumption and investment for everyone.

Money’s value is inherently meaningless – inherently we only care about the stuff we get with it. Yen having many more zeroes on prices than the pound has not made Japanese poorer than Brits. Why constantly change the expectations for where the macroeconomy is going and make firms uncertain about whether they can afford certain investments and certain employees?

This is controlled flight into terrain. If the Bank of England goes ahead with its plan to cut inflation by tightening policy so much that we have a recession, we will regret it big time.

Share

P.S. I also had a long post on the Works in Progress substack, talking about agglomeration benefits, and why a major recent academic paper does not prove they don’t exist (or that they aren’t important).

Over the past few years compelling evidence has come out suggesting that housing policy errors are behind slow productivity growth, wealth inequality, and the cost of living crisis. But a raft of new work suggests housing policy errors might also be a key explanation for rising regional inequality in the West, and more specifically the rise of England’s South East and the relative decline of the Midlands and North.

Subscribe now

A speculative working paper from Greg Clark and his collaborator Neil Cummins looks far back to determine why the North, the site of the UK's industrial revolution, has declined relative to much of the country, especially London. He looks at a genealogy of 78,000 people and finds that people with distinctively Northern surnames actually have not lost out relative to those with distinctively Southern surnames.

But many of the best and brightest have moved to London, and this has accelerated in recent years, even into the late 2000s, just as the productivity gap has opened up. He links these two phenomena together: being unable to keep their smartest has hampered the success of commerce and industry there. Large fiscal transfers make relatively little long term difference if they don't attract people. And one-size-fits-all policies (like national pay bargaining) that might be thought to help poorer areas actually worsen the problem, incentivising people to queue for the far more generous pay that the public sector offers in poorer areas.

What Clark's work lacks is an explanation. Why has the migration South been skill-biased? Why has it sped up so much recently? A recent literature suggests that restrictive housing supply, which raises rents (and house prices), deters low-skilled movers much more than high-skilled movers, raising the average talent in the host city and cutting it in the sending city. This compounds itself, because the move itself affects the desirability of the places to live in, especially for the high-skilled.

The data comes from the USA, but the situation is the same. Between 1880 and 1980 per capita income between US states converged steadily and significantly. If you heard that things were better somewhere else, you upped sticks and moved. It's the point of America! But convergence fell from nearly two percent per year to under half of that in the last two decades. And a recent NBER working paper points the finger directly at falling migration.

The mechanism we propose for explaining the decline in income convergence can be understood through an example. Through most of the twentieth century, both janitors and lawyers earned considerably more in the tri-state New York area (NY, NJ, CT) than their colleagues in the Deep South (AL, AR, GA, MS, SC). This was true in both nominal terms, and after adjusting for differences in housing prices. Migration responded to these differences, and this labor reallocation reduced income gaps over time.

Today, though nominal premiums to being in the New York area are large for these two occupations, the high costs of housing in the New York area have changed this calculus. Lawyers continute to earn much more in the New York area in both nominal terms and net of housing costs, but janitors now earn less in the New York area after subtracting housing costs than they do in the Deep South.

This sharp difference arises in part because for lawyers in the New York area, housing costs are equal to 21% of their income, while housing costs are equal to 52% of income for New York area janitors. While it may still be worth it for lawyers to move to New York, high housing prices offset the nominal wage gains for janitors.

This effect only appears in places that regulate land use and housing supply tightly. Another paper, by Andrii Parkhomeno, puts 23% of the rise in wage dispersion down to housing issues. In the forty years to 1940-1980 inequality both regional and overall income inequality fell, with migration explaining a huge chunk of this. The opposite has happened more recently both in the USA and the UK, as the incentive and ability to move has crashed.

The world of 2017 has a lot of big problems. Clearly some of them (dealing with new technology, North Korea, and terrorism) are things that reforming housing has little to do with. But many of the most important issues today are almost entirely driven by housing problems. At Berlin or Tokyo rents per square metre my life would be completely transformed. I could save go on an extra holiday a year, save up for a house, put more into my pension, and go out for dinner more—all at the same time.

If this flow were driven by inevitable socio-economic factors then maybe the infamous paper would be right: we should just abandon declining places and move their residents. But it looks like regional decline is a policy choice just as the productivity puzzle, wealth inequality, and high rents are a policy choice. We absolutely can do something about them.

P.S. There is a brilliant new special issue of Works in Progress, which you must check out.

In economics there is a relationship called Goodhart’s Law, after Charles Goodhart, LSE professor and onetime member of the Bank of England’s Monetary Policy Committee. Goodhart’s Law states that the relationship between any indicator of some phenomenon that is not the phenomenon itself and the phenomenon will break down when you try and exploit it.

The most famous example is the Phillips Curve. The Phillips Curve established a remarkably strong negative relationship between unemployment and inflation between 1861 and 1961. Based on this relationship, Keynesian postwar economic planners tended to try and fine tune aggregate demand to keep the economy at full employment. But as soon as they started doing this, the relationship started to break down. In the 1970s higher inflation tended to coincide with higher unemployment.

Recall the concept of ‘surrogate endpoints’ from my last post on drugs. Usually, the FDA approves your drug based on whether it affects the lifespan of patients. However, in some cases, it approves it based on surrogate endpoints – things that correlate with lifespan that we can track much earlier. This makes drug development, especially for long term conditions, dramatically cheaper to do.

Well-defined surrogate endpoints do achieve their goals. Low density lipoprotein (LDL) cholesterol seemed in the 1960s and 1970s to be a strong indicator of cardiovascular problems - heart disease, heart attacks and CVD deaths. Statins lower LDL cholesterol, and were approved by the FDA to treat those at risk of heart disease for that reason. Now, Goodhart’s Law did still operate: statins do not lower CVD as much as we would expect given how much they lower LDL. But it didn’t fully wipe out the link—there remained a correlation between LDL and mortality. With hindsight, now that we have large meta-analyses covering tens of thousands of those who have used statins we find they did indeed lower overall mortality.

Much more aggressive versions of Goodhart’s Law have operated when it comes to other surrogate endpoints. Oncology drugs can be approved both by the FDA and the EMA (its European equivalent) on the basis of tumour response rate (the percentage of patients whose cancer tumour shrinks after treatment) or progression-free survival (how many patients survive without their cancer growing).

Intuitively, and based on established correlations, these endpoints both correlate with overall all-cause mortality (and cancer mortality). However, in practice drugs approved on the basis of improving these metrics have not improved survival much—or in some cases at all. One obvious reason is that drugs can have extremely bad side effects, potentially in some cases outweighing or balancing out the benefits seen via the endpoints. In 2011, the FDA removed the breast cancer indication1 it had given to Avastin (bevacizumab), citing safety concerns—at this point it had also shown little impact on overall survival.

Subscribe now

Chul Kim and Vinay Prasad review all 36 cancer drugs approved in the USA between 2008 and 2012 on the basis of surrogate endpoints. Of these 36, only five were shown to improve survival at all. Eighteen did not improve survival in the follow-up trials; and at the time of their paper (2015) a further 13 had unclear effects on overall mortality. The same authors, with Mauricio Burotto, followed up this work by systematically reviewing and meta-analysing the literature on these oncology surrogate endpoints. They found in 2015 that most studies of current oncology surrogate endpoints have a low correlation with overall survival. A 2019 follow-up review found similar results.

This has combined with perverse funding mechanisms to create some severely distorted incentives for research. Medicare, which funds healthcare for around 60 million Americans, and nearly all Americans over 65, is obliged to fund any cancer treatment that gets FDA approval. Combined with the fact that cancer drugs can be approved on the basis of weakly validated surrogate endpoints, and you have a perfect storm where relatively unhelpful drugs are paid for - whatever the price. Critics of the system like Vinay Prasad allege that only fear of public backlash restrains firms from setting essentially infinite yearly prices on their drugs. Worse still, firms know this, and are incentivised to spend heavily on researching new marginally effective oncology drugs. 

A final problem is that this effective loosening of standards, while not necessarily a problem if combined with rational funding mechanisms, is distorting our overall data on the number of new molecular entities approved in given years.

For decades, the numbers of drugs approved per billion dollars spent has been falling—this is called ‘Eroom’s Law’. Eroom’s Law is the inverse of Moore’s Law, which describes the consistent and steady progress we have made on fitting more transistors on a computer chip. It describes the consistent and steady decline in the number of new molecular entities approved by the FDA per billion dollars spent by Big Pharma.

Since the late 00s there has been a huge increase in the number of drugs designated orphans (for rare diseases), as well as oncology drugs approved based on surrogates. This recent uptick has driven a flattening of the otherwise endless Eroom’s Law seen for previous decades. An Eroom’s Law with QALYs saved per $1bn drug R&D on the y-axis would not show this flattening of the trend. In a sense, Scannell’s metric itself suffered a Goodhart’s Law style breakdown as soon as it began to be measured.

So what can be done? I don’t personally think the issue is so simple as surrogates bad, survival good. I think the issue is that surrogates are approved and validated in a much more arbitrary way than drugs themselves. Drugs go through the most rigorous scientific process of pretty much anything in our whole society. They apply basically every rule and procedure suggested by advocates of good science like Stuart Ritchie: pre-registration, randomisation, multiple centres, open results, and so on. But the FDA approves surrogates in a more qualitative way.

Why not require surrogates to go through a similar procedure to drugs – proof of their meaningful link to survival through rigorous tests – but then allow drug companies to patent and license the surrogates for a time. Surrogates can be tested in randomised controlled trials with another main outcome, but as a pre-registered secondary outcome. And surrogates can be tested through other non-RCT but quasi-random methods, for example Mendelian Randomisation.

In this way, they would have an incentive to discover and prove the most valuable links, and we could dramatically cut costs and speed up drug development, without generating a whole range of drugs that don’t work, like the current system.

P.S. Works in Progress has just released its first ever ‘special issue’, which tackles the question ‘Dude, where is my flying car?’

One of my favourite articles of all time is ‘Catching Crumbs from the Table’ by Ted Chiang. Ted Chiang is a brilliant writer, whose work is perhaps categorised as a type of science fiction. In writing, he tends to change one small thing about our world and then work out all of the consequences. Inevitably, his world building is careful and immersive, and his stories are compelling. His most famous contribution is penning the novella or short story that inspired the film Arrival (2016).

‘Catching Crumbs from the Table’ is, I think, the only work of fiction ever published in Nature, one of science’s ‘big three’ journals. I recommend it to people all the time, and given my interests sometimes also have it recommended to me. Despite its short length, it gets across a realistic world in a very evocative way, expressing speculations that I think are almost entirely true, or at least as good predictions as you could plausibly make.

The story imagines a world where (friendly, well-aligned) artificial general intelligences have vastly surpassed human capacity for science. These AIs continually improve the technology we can use, in tangible ways that we experience when we use the advances. But we can’t understand how the advances have been made. 

In a world where the frontier of science is so vastly far away from where humans had taken it, the most efficient way to do science is mostly not to theorise, experiment in labs, and so on, but just to take the data outputs the AIs make and try and decode them. How great is an AI really if it can’t explain what it’s doing to us? A reasonable question, but could we explain our reasoning to an ant?

Subscribe now

I think this is an intriguing look at what might happen in a world with aligned AI. But this week, when markets in the UK have been going wild, it brought an idea to mind. Aren’t we already catching crumbs from the table?

According to the fundamental theorems of welfare economics, or Friedrich Hayek’s theories around information, market participants don’t need to know much to together interact and come to an efficient market price. Markets work because they take the tiny snippets of information the participants have and aggregate them together into something useful. 

If the price is wrong, the market will pay you to correct it. The more it’s wrong, the more it’ll pay you. In a sense, entrepreneurs are people who notice some price has become wrong, perhaps due to a technological or social shift, and who set out to prove it. If they are right, they get rich.

This is all to say that markets are collective intelligences that may know more than you or I, or indeed any individual in the world. 

In some extremely simple markets, it’s easy to tell what the market is telling you, and why. I was watching England play Pakistan on Sunday, in one of the best T20 games I’ve ever seen. Every time Liam Dawson hit a boundary in the England chase, the odds of an England win rose several percentage points, and the odds of a Pakistan win fell concomitantly. It’s obvious what was going on here. Each boundary was better than what was expected from the number eight, shifting the expectations bettors had about the future, and the prices they were willing to pay for contracts that pay out on different outcomes.

Some markets are a little harder to peer into. When house prices rise, the economist Scott Sumner might tell you not to reason from a price change, but if this happens just as we see a fall in interest rates, a rise in immigration, a fall in housing taxes, or promises to restrict the supply of housing, then we might feel reasonably confident in attributing the change.

Financial markets, however, can defy simple explanation. How about last Friday’s crash in the pound? It slid a little just before Kwasi Kwarteng went out to give his mini budget last week, then slipped another percentage point in the ten or so hours afterwards. Over the weekend the pound was obliterated in Asian markets, dropping nearly five percent on Sunday evening UK time.

Since then we have been catching crumbs from the table, attempting to explain this fall. 

A handful of commentators, like Rishi Sunak, predicted ahead of time that there would be pressure on the pound under Liz Truss. But since she was, at the time, his rival in a Tory leadership election, and since, at that time, he was throwing everything he could at her, it seemed hard to put much faith in. Dominic Cummings tweeted in July that Liz Truss becoming leader meant ‘guaranteed fiasco’. But almost every recent UK prime minister has presided over quite a lot of fiasco. In both cases, standard forecasting rigour was not in place: it is extremely easy to predict downturns or generic bad happenings if you’re nonspecific about when they are going to happen and to what extent.

No commentators I have seen predicted at the time the budget happened, or immediately afterwards, that the pound would come under the sort of pressure it did. (Please correct me if I’m wrong.)

Since, we have been catching crumbs from the table. An intuitive argument might be that the fiscal event’s very significant spending commitments (from energy price caps) and tax cuts must have spooked the markets – perhaps Britain’s good credit is at issue. In fact, this argument seems so intuitive to people, that a great many have reached for it.

I believe, however, that our methods of catching crumbs from the table have moved on a little from this sort of reasoning. You don’t need to believe the most extreme and stringent version of rational expectations to think that the bulk of those commitments – which have been promised consistently for weeks or months – were already expected and priced in by the financial markets. Similarly, the sacking of Tom Scholar, which happened on 8th Sept, shortly before the pound peaked at $1.17, seems like an implausible reason that the pound fell on 23rd and 25th, given what we know about the forward-looking nature of markets. And, remember, that although the markets did react substantially to the announcements, the biggest drops came several days later.

A more plausible attempt at crumb-catching says that although only a few billion of the fiscal commitments were new information, these commitments were strong signals of the new Government lacking the commitment to fiscal rectitude that had until now been a hallmark of post-2010 big C Conservatism. This meant, in the medium to long run, UK financial assets had become riskier, perhaps because inflation has a chance of becoming unanchored. Underlining this might be the fact that the bigger collapse in the pound happened just after Kwasi Kwarteng went on Sunday television to promise even more tax cuts.

An alternative way of interpreting these crumbs – doing crumblinology – is to imagine you were a creditor to a struggling and mediocre firm which nonetheless sweated its assets as hard as possible to pay your interest and pay you back – but at the cost of attempting more high-risk strategies that might ultimately grow the company. If this firm attempted a radical new growth path, in part by focusing less on diverting all resources to paying your loan, you might be slightly worried – even if the high-risk strategies had a chance of working.

Some amusing crumb catching I have also seen takes a totally different line. The fiscal event will be good for the UK, this argument goes, except that the public (or press, and via them, the public) mistakenly hate it, and it will shift them towards voting for Labour. A rising chance of a Labour victory is bad for the pound for standard rational expectations reasons, this story goes. Though this would be a fun one to believe, and though there were indeed several polls showing gigantic poll leads for Labour, Betfair’s odds of Labour getting a majority at the next election, or the most seats, have barely shifted at all since Friday, so I suspect this one is not quite true.

Another option is to try and interrogate the price data we get with much greater rigour. Since gilt yields rose just as the pound fell, you might say that the pound might have fallen much further if the market didn’t expect the Bank of England to step in with higher policy interest rates. Things are actually worse than they seem! But then US and European government bond yields have risen at the same time (though less than those in the UK) so perhaps these are in substantial part international issues. Things are less bad than they seem!

Still, it is difficult to know. I certainly don’t pretend to have any idea which narrative is true, if any of these, I am far too ignorant about financial markets. The market’s view is not an average of the opinions of the inputs into it, but something entirely separate from any of them, a completely new view entirely. It is a black box, a superior collective intelligence whose motives and reasoning we can only reconstruct hypothetically.

We are already catching crumbs from the table, as indeed we were when Chiang wrote his story.

Market power is when a firm can charge more than the cost of making a new unit, because of barriers to entry. In a competitive market where firms rent their capital, they can only charge marginal cost: if not, other firms would pop up and steal the market by undercutting them. Some of the most competitive real live markets are close to this—supermarkets make tiny margins and offer near-identical prices. But most markets are to some degree away from this.

This means there are rents. In an extreme case these are monopoly rents. Imagine you own and run a railway between Manchester and Leeds. Building and operating a new one costs a lot, and this new railway would have to charge lower prices to attract customers. Assuming there are no other modes of transport then this gap—expected potential profits—determines what prices the monopolist can set before they'll face competition. The gap between a competitive price and the monopoly price is a rent. This is a pure redistribution from users of the network from whoever happens to own it.

It helps to distinguish between barriers to entry and costs of entry. If building a new railway was trivial technically, and the only cost was regulatory—e.g. you had to pay off a corrupt bureaucrat but rails appeared magically—then potential competitors face barriers to entry. Real costs that society has to bear—like using workers, capital, and management or entrepreneurship to organise building a railway line instead of doing something else—are costs of entry.

Barriers to entry reduce competition with no corresponding benefit, but when there are large costs of entry, the loss of competition is balanced by keeping resources spare. Competition requires "unnnecessary duplication", something that greatly troubled early socialists and led them to believe that the socialist economy would not only be more moral than the capitalist one, but more efficient.

In any case, a canny monopolist will set prices such that no competitor enters, enjoying their rents for as long as possible. Modern regulators aim to set prices so that natural monopolies do not earn rents, and consumers get higher consumer surplus instead. So far, so sensible. But this approach may have a key defect: it requires Herculean feats of innovative winner-picking elsewhere when you think the economy might be dynamic, not static.

This is because rents that are caused by costs of entry (but not those caused by barriers to entry) are automatic prizes that reward entrepreneurs in direct proportion to how much they can alleviate inefficiency in the marketplace. Higher rents not only motivate investment, as in the example above, where someone might build a second railway (or indeed a road, canal, bus service, or air link) between Manchester and Leeds, but also innovation.

This is quite general. Some innovation does not take away profits, but instead reduces the need for labour, land, or capital. But the cost of these is a rent too—a factor rent. Just as we don't see these as fixed, nor should we see market rents as fixed. As long as we think innovation is reasonably possible, we should be open to allowing market rents to exist to direct innovation. Efficiency-enhancing innovations do not fall from heaven—they come from where we focus our research activities as a society.

It's obvious how this applies to intellectual property. When I have a patent to produce a drug, if this drug is useful I will earn large rents. But it is precisely those rents that indicate that finding a substitute is so valuable. If a firm can find a close substitute, they stand to get some of those rents for themselves—driving innovation into the highest value areas. So we should be more sanguine about less competitive markets and natural monopolies—where competition is restrained by facts of the world, not regulation—they help tell us where improvement is most valuable. And they pay people for doing that improvement!

Cooperation is why we are successful as a species. Notwithstanding the emergence of new technologies like Zoom, cooperation still happens best when we are physically near one another. To work physically near someone you have to be within commuting distance. So cities are important.1

Many of the most successful cities around the world have shortages of housing, with houses that are much more expensive to buy than build. In ‘The Housing Theory of Everything’ John Myers, Sam Bowman and I looked at all the terrible knock-on effects these shortages have on fertility, growth, inequality, and the climate – and possibly more.

In general, these cities have housing shortages because development control systems deliberately prevent either extra density or extra outward expansion or both. They do this where there is popular demand: development tends to impose costs on locals that they are not compensated for, like noise and congestion. When clever policy means that they are compensated for noise, congestion, and so on, locals do often opt for development, as in clever development policies seen in Tel Aviv, Seoul, Houston, and Vancouver.

One further potential cost of development is crime. Crime is hyperlocal. People pay substantial amounts to buy housing in neighbourhoods with less crime. Since crime is extremely bad, people are risk averse about getting more of it near them, and when crime is high they tend to be more opposed to development nearby.

Crime in America is extremely high, relative to developed countries. For example there were 24,576 homicides in the USA in 2020, compared to 318 in Japan, a country with about 40% of its population, or about 30x higher per capita. There were just over 4,000 across the EU as a whole, which has more than 100 million extra citizens compared to America – or a rate about 8x lower.

The USA actually has somewhat less of a housing shortage than somewhere like Britain. Whereas British cities have been contained almost completely by green belts, with practically no outward expansion at all since the Second World War, American cities have sprawled legendarily.

In cities with four directions to sprawl into – like Atlanta, Phoenix, Houston, and Dallas – this has led to house prices that are not far off the cost of construction.

In cities facing physical limits, usually lakes, oceans, and mountains, housing shortages appear, with high rents and prices. San Francisco and Silicon Valley are hemmed in by the bay, the Pacific, and the mountains; New York City has the Hudson River and the Atlantic. Both cities have significant housing shortages.

There are upzoning efforts going on across the USA – attempts to increase the amount of building that is permitted on existing suburban plots. YIMBYs have had some impressive success. To the extent that these will succeed, I think it is because American cities had been getting much safer between the early 1990s and about 2014, driving property values in the centre up dramatically, but zoning had not always caught up to reflect this.

A New York City with 2,200 murders in a year is one where you will be minded to block every bit of housing people want to build nearby. Residents of a 300-murder NYC will balance up the probabilities differently. This is true across all categories of crime – for example California has more than 3x fewer murders per capita today than in 1993, 2.5x fewer car thefts, 8x fewer burglaries, and 3x fewer robberies. But I tend to use murder as the best indicator, since it is measured most consistently and accurately.

Sadly, these attempts to allow more development in suburban areas of cities will for the most part have trouble succeeding, in part due to worries about crime. Overall, the gigantic overwhelming majority of new housing in American cities is not being built in the places where housing is most expensive, near the best jobs in San Francisco and New York City, but instead forced to the places where it is easiest to sprawl, with all the concomitant costs.

How much is this costing America?

There are two prominent answers to this question in the academic literature, which I will call Hsieh & Moretti and Duranton & Puga. Both are helpfully open access.

Hsieh & Moretti look at 220 American metropolitan areas between 1964 and 2009. They find that these strict zoning laws that locals have imposed are keeping people from moving to superstar cities. They find that this lack of migration is driving the widening productivity gaps that opened up between cities during this period. And summed together they find that growth would have been 36% higher over the period, which implies that overall output would have been about 14% higher in 2009 with rules that allowed more upzoning.

Duranton & Puga model the productivity upsides and congestion downsides of city size and density, and look at a counterfactual USA where planning restrictions were stripped away. In this scenario many large productive US cities like Boston, San Jose, New York City, and San Francisco grow by millions in population and see rapidly rising productivity through extra agglomeration and division of labour. Overall, average incomes rise by over a quarter.

So the prize overall is somewhere between 14% and 25% of US national income. But crime isn’t responsible for all of this. We need to work out some sort of way of judging how much extra restrictionism extra crime causes.

I have puzzled over several methods. Firstly, I can’t seem to find any literature at all that directly addresses this question by creating some sort of measure of the demand for zoning, and then seeing how it varies over time, over space, or before and after some sort of treatment condition, event, or natural experiment.

I wondered, therefore, if there might be some sort of measure of zoning stringency around the world which I could draw on and compare to crime rates. Japanese zoning is legendarily liberal; American zoning is strict. Some of the gap between these is down to the varying crime rates. But measuring the stringency of zoning systems is a detailed, careful job, and working out how much of that gap is due to crime is a further delicate statistical job. And sadly I can’t find such a measure or metric – and having asked field experts like Nolan Gray, they don’t know of such a thing either.

So I’m coming up empty. How can I work out empirically how much fears about crime drive some of the demands for restrictive development control?

Post scripts

• I appeared on the Aesthetic City podcast with Ruben Hanssen, which I enjoyed very much.

• The UK has a new housing secretary, Simon Clarke. Read more about him here.

• My work on crime (so far) has helped influence an excellent post on the Effective Altruism forums.

• Ed West wrote about how certain crimes are de facto legal in the UK.

This article was inspired by ‘Age of the Bacteriophage’, an excellent article by Léa Zinsli in the latest issue of Works in Progress.

We have become increasingly familiar with the story of antibiotic resistance.

Antibiotics work, but bacteria can evolve resistance to them. By ‘evolve’, we mean within extremely large groups of bacteria that we try and wipe out with our drugs, there are sometimes bacteria who, due to random mutation, have adaptations that make them immune to the antibiotic’s mechanism of action.

After their bacterial siblings and competitors have been wiped out, the immune bacteria reproduce and build the colony/infection back up to its previous size. Like chopping off a hydra’s head – another one regrows – but even worse as your sword now doesn’t cut through it.

Since we are not developing new antibiotics as fast as bacteria are developing immunity to our old ones, we potentially have a growing problem, hence the revival of interest in alternative therapies like bacteriophages – see more in Léa’s article.

My interest here is using this as a makeshift analogy that might help us think about another societal problem we have. Before modernity, people by and large aimed to have large families. If they could afford to get married, and feed, clothe and house them, and they were lucky with infant mortality and the harsh vagaries of premodern life, they ended up having a large number of children. Before the demographic transition, completed fertility rates were usually around 7.5 per woman, as they are in countries like Niger today.

Modernity, and especially the industrial revolution, changed all this. Starting with France, followed by Britain, Germany, the USA, and eventually the rest of the developed world, parents began to have something more like 2.5 children. Most of this happened before the pill or the cheap, widely-available condom, and before legal and available abortion, not to mention before the major waves of feminism culturally, legally, and economically. In some countries it happened before the decline of child mortality, in others it happened after.

There is still a debate about what the deeper causes are – some ascribe it to wealth itself; others to a modern system of production where ‘quality’ of children is worth more than quantity; others still to cultural changes and secularisation; and others to exogenous changes in the cost of having children like increased pressure to school children and norms and laws proscribing child labour.

Whatever the causes, post-demographic transition countries have generally produced only just-above-replacement fertility, with one exception, in their post-Second World War baby booms. In recent years, this just-above-replacement fertility has steadily slid to somewhat below replacement, sparking widespread concern and the beginnings of some policy responses.

Subscribe now

But what I want to argue here is, whatever the ultimate cause, humans are like bacteria. Within human groups, and within humanity between groups, the future belongs to those who show up. Those who reproduce steadily increase their fraction in the group and those who don’t steadily decline. Quakers were a vital part of the American founding population, but since they had very few children, and Puritans had as many as ten per family, Puritans ended up as a much more important part of what made later America.

The demographic transition is not quite like an antibiotic: though it does cause childlessness and the complete and permanent cessation of some lineages, it mostly works through reducing the number of children per family, while most families still have at least one. Eighty percent of British women born in 1975 (the youngest cohort to have finished having children) had at least one. But like an antibiotic, some groups seem more resistant than others.

First let’s consider groups who have never gone through the demographic transitions, like Nigeriens, Hasidic Jews, and the Old Order Amish. It’s entirely possible that they will come up against whatever it is that has caused fertility to decline amongst everyone else. But if they don’t, then naive projections of population suggest they will become strikingly large fractions of the overall human population in coming decades and centuries. Humans as a whole will have evolved around the problem, like, say, homo sapiens hypothetically beat out Neanderthals due to their greater cooperativeness, even if most of us leave very few genetic descendants.

But I think that even if these groups do go through a demographic transition, we are already evolving resistance. Recent data suggests that wealth is now correlated with fertility – the rich are having more babies than the middle class. Women in countries with higher female labour force participation now have more children than those who don’t.

If everyone has lots of kids, then wanting to have kids is not selected for. But if nearly everyone has very few children, then intrinsic desire to procreate is one of the most important things we select for.

To put this another way, before the demographic transition, yours and your parents’, and your children’s fertilities were all only correlated because of your wealth and other similar things that drove fertility. Since wealth, even in the past, varied quite a bit between generations, this is fairly low correlation overall.

However, since the demographic transition, the fertility of different generations of particular families has begun to correlate substantially. Children from families with four children tend to be more likely to go on to have four themselves.

One downside here is shared from the bacterial example. What if having resistance to the demographic transition antibiotic correlates – as it sometimes does in bacteria – with something that makes us less effective. Some people believe that being religious is evidence of serious flaws in a person’s reasoning. But in some countries it is precisely the religious that are having the most offspring. If we select only for fertility, then we relax selection on other attributes of people.

Another downside is speed. Antibiotics kill billions of bacteria really quickly, leaving a tiny fraction to build back better, which they do incredibly quickly. Here, the selection is (thankfully) much less extreme. This means that populations could get pretty low before they recovered.

A simplistic little model I built suggests that if one percent of the fertile UK population had a fertility rate 1.2 above the 1.8 norm, it would take over 200 years for the population to beat its current level, assuming zero immigration and everything staying static.

But I still think the story is true and important: humans are like bacteria, and we are evolving bacterial resistance to the antibiotic of low fertility.