Build it – but will they come?

When are transport investments justified?

I have an article out in Works In Progress about through running, the most cost-effective way of upgrading (sub)urban railway lines. It means taking two or more lines on opposite sides of a city and running them through the city centre, typically in a tunnel, in a network that (copying German practice) I call an S-Bahn. An S-Bahn enables more frequent suburban services, which have better connections to the city centre – and it increases the capacity of city-centre termini into the bargain.

You should go and read the full piece, which goes into detail about how this was done in Munich and London, and the opportunities for British cities that through running presents. It also includes some very cool graphics:

This post is about something that I allude to in the article: when are transport investments justified?

‘No brainer’ schemes

Some investments can be justified purely because a city’s transport network is crying out for more capacity. This was the case with the Elizabeth Line. These could be called ‘no-brainer’ schemes. Crossrail 2 falls into this category, and so would a Manchester S-Bahn. Looking at mainline railways, although I have criticisms of the route of HS2, a high-speed line from London to Birmingham and the North West is also a ‘no-brainer’.

Britain has an unfortunate tendency to build only ‘no-brainer’ rail schemes, often many decades late, and not to build very much else. We know that they are built late because of how quickly they fill up. If a new railway line is full within a few years of opening, as has essentially happened with the Elizabeth line, then that is a pretty clear signal that it should have been built a long time beforehand.

It always looks like this at rush hour. Image from IanVisits.

By contrast, the Munich S-Bahn took many decades to fill up: to cope with demand, the number of trains using it in each direction in the peak has risen from 18 per hour in 1972, to 24 in 1997, to 30 today. This means it can cater for 840,000 journeys each workday, even though originally it was only designed for 250,000 passengers per day. The gradually increasing patronage is a sign that it was built at roughly the right time.

That is why (in descending order of priority) schemes like Crossrail 2, the Bakerloo Line Extension all the way to Hayes, and a Chelsea–Hackney Line are so vital. We should have built them, or something like them, decades ago. London is crying out for new capacity. There is no doubt that these lines will be well-used.

Alon Levy suggests that, as a rough-and-ready rule of thumb, a transit project is viable if its cost per weekday journey is less than the country’s GDP per capita. There’s no special reason to choose GDP per capita specifically, but as a proxy for national wealth it encodes the fact that when a country is richer, the marginal journey will produce greater economic benefits.

The Elizabeth Line has 700,000 journeys per weekday, and cost £21.3 billion in today’s money. That works out to £30,428 per daily journey, which is well below Britain’s GDP/capita. Had the Elizabeth Line been built at Swedish costs, it would be even better: £8,700 per daily journey.

Build it and they will come?

This crude-but-useful heuristic is a form of cost:benefit analysis, which uses the number of daily journeys as a proxy for benefits.

Cost:benefit analysis (in a more sophisticated form) is used for all railway projects in Britain. It is not the decisive factor in choosing whether a project gets built: the Thameslink Programme, which was built, had a benefit:cost ratio of only 1.4:1, while the Leeds Tram, which was not built, was 2.2:1. It does, however, inform decision-making, especially at the earlier stages of a project’s development.

But cost:benefit analysis has a flaw: it ignores the fact that transport projects can be agents of change. They can help catalyse growth in ways that do not easily show up in estimates of their benefits.

It is, for instance, highly unlikely that Canary Wharf would have been built without the Docklands Light Railway. The first section of the DLR was built on a shoestring budget: at only £90 million (£260 million in today’s money), it re-used disused railway lines and avoided the expense of running into central London by terminating on the edge of the City of London at Tower Gateway. By the time that the first section opened in 1987, however, events had overtaken the DLR, as the plans to redevelop the docks into Canary Wharf were under development. Before the scheme was even finished, in 1986 the extension to the heart of the City at Bank was approved.

Recognising the importance of rapid transit access to the success of its scheme, the developer of Canary Wharf was prepared to fund a third of the cost of the 1.2 kilometer tunnel. (The tunnel, incidentally, cost more than three times the original scheme.) The fact that the developer was willing to pay for it is a pretty clear signal that access by rail was necessary to make the scheme viable. Subsequently, the owners of Canary Wharf promoted the charmingly-named ‘Waterloo and Greenwich Railway’, which morphed into the Jubilee Line Extension, to provide more capacity to the development.

The site of Canary Wharf in 1981. Image by David Dixon from National Maritime Museum, Greenwich Park on Wikimedia Commons.

Canary Wharf in 2022, from a similar angle. Image from Wikimedia Commons.

The fact that transport projects can be agents of change is a reason to be somewhat sceptical of cost:benefit analysis. Any cost:benefit analysis done on the DLR in the early 1980s would have ignored Canary Wharf, because nobody thought that out of the wastelands of the docklands would rise a little corner of urban America.

One could go so far as to argue, in fact, that the tool of cost:benefit analysis systematically underestimates the benefits of new infrastructure, which means that less gets built. As a methodology, it prioritises the visible, costs and easily-quantifiable benefits, and ignores the invisible – the ability of transport to change a place.

This agent of change argument, however, needs to be used with extreme caution. Transport is a necessary condition for growth, but it is not sufficient: as well as the DLR, Canary Wharf was enabled by the deregulation of Britain’s financial industry in the ‘Big Bang’ of the late 1980s, an economic boom, foreign capital investment, and a special regime for urban planning.

For every Canary Wharf, there is an Ebbsfleet, a project to build a new town on former industrial land in the Thames Estuary. The construction of HS1 was meant to spur the development of a new town; instead, 18 years after the line opened and 11 years after the formation of a special government body to built it, Ebbsfleet only has 4,500 homes out of the planned 15,000, down from the original ambition of 43,000, partly because the discovery of rare spiders caused half the site to be designated for wildlife protection.

Ebbsfleet is supposedly a ‘garden city’. Image from New Civil Engineer.

Inducing demand?

Justifying transport because it can act as an agent of change is an argument from induced demand. This phrase is well-known because it is a common canard among urbanists, who sometimes argue that we should not build roads, because experience shows us that new roads are full pretty much as soon as they open: new roads do nothing to alleviate congestion.

‘Induced demand’ makes little sense as an argument against roads. The phenomenon being described is not induced demand (new demand created because of the road’s opening) but latent demand. If a road is widened from two lanes to four lanes, and is full as soon as it opens, then that means there was two lanes’ worth of demand for travel which could not be satisfied, because there was not enough road space. The demand was not ‘induced’ by the new road: it was there all along, but suppressed by the size of the road. I wish urbanists would stop using the argument: there are much, much better arguments against building new roads in cities.

That said, new transport infrastructure certainly can, in the medium- to long-term, induce demand. (The false urbanist claim is that it induces demand in the short term.) When a new road opens, it will over time cause people’s behaviour to shift. People will get jobs that they otherwise wouldn’t, or make leisure journeys that they wouldn’t have considered before the new infrastructure opened; and as a result homes, workplaces and amenities will gradually cluster in the better-connected places.

Putting the argument in these terms, new underground railway lines in London can be justified purely on the basis of the vast amount of latent demand in the city. This applies to many schemes outside of London as well, such as an S-Bahn scheme in Manchester, or some kind of light rail or tram-train in Cambridge.

But there are also many possible schemes that are reliant, to some extent, on the ability of the new railway line to induce demand. Again, this is a question of degree. A Leeds tram(-metro) probably relies on induced demand much less than a second Glasgow Subway line, because the places that people currently want to go to in Glasgow are already reasonably well-served by the city’s rail network, whereas Leeds is famously the biggest city in western Europe without a metro or light rail system. A second Glasgow Subway line would be marginal unless it were able to catalyse the city’s East End into as much a magnet for travel as the West End.

In any city, in any country, inducing demand is a risky business. The development might not materialise, for a host of good reasons: there might be an economic downturn, or it might never have been viable in the first place. The phrase ‘white elephant’ exists for a reason.

But in Britain, there is an added complication: the planning system. The fate of transport schemes that rely on induced demand is in the lap of the planning system.

Marching hand-in-hand

Britain’s planning system (construed widely, to include things like environmental regulations) makes it very difficult to build things, and this is the biggest reason Britain’s economy has stagnated.

This obviously holds back any scheme that relies on inducing demand. To induce demand, you usually need to build things: houses, workplaces and amenities that are well-connected to the new transport infrastructure. Ebbsfleet is an example of this: the fact that it is still a derelict wasteland is downstream of the planning system. What should have been a great magnet for development was transmuted into a white elephant by the planning system. ‘Build it and they will come’ does not work if they are prevented from coming.

We can draw two conclusions from this. The first is that the infrastructure and development need to march hand-in-hand.

There is substantial precedent for this. In the 1920s the Metropolitan Railway developed estates of middle-class suburbia around its stations. Although other railway companies had to sell off surplus land, thanks to the initiative of its chairman in the nineteenth century, Sir Edward Watkin, the Met was allowed to retain land that it believed was necessary for future railway use. Fifty years later, the railway ultimately used this land for housing, expanding London’s urban footprint with mock-Tudor suburban houses with gardens. The railway marketed these developments as ‘Metro-land’, offering the promise of living a peaceful, semi-rural life, with speedy journeys into the centre of London on the Metropolitan Line. This, in turn, increased the Met’s income from fares. Suburban development and railway infrastructure were in symbiosis with one another. Similarly, in Japan, the President of Tokyu, one of the big private railways in the Tokyo area, has described his company as a ‘city-shaping company’, rather than a railway company.

Obviously, Britain’s planning system does not ignore transport links. The Greater London Plan, for instance, uses a measure called PTAL, public transport accessibility level, to determine how much housing should be built in a particular place given its transport links. We often speak of a transport investment ‘unlocking’ housing, which is an ugly phrase, but an accurate one: the better-connected a place is, the more densely it can be built. Likewise, infrastructure is designed to enable housing: a big part of the reason Crossrail 2 should go to Tottenham Hale and then northwards is the opportunity to build more housing there.

But while planning and transport influence one another, the relationship goes no further than that. There needs to be a tighter link between the two. England’s planning system is often criticised (especially by those within the planning profession) for being reactive to forces happening around it, such as a new railway line being built, rather than proactively trying to plan good places. This is unfortunate, and undermines the legitimacy of the entire planning system, whose core function ought to be planning rather than responding.

We need a mechanism to enable delivery of transport and housing that are joined at the hip: Metroland, not Ebbsfleet. The Holy Grail would be to enable transport schemes to be funded by new development. This would kill many birds with one stone. More transport schemes would become viable. It would help solve the problem of paying for infrastructure. It would reduce the basic unfairness that landowners near to new infrastructure get an unearned windfall from development. We would build more housing, workplaces and amenities. Transport schemes would be able to create better places. This wouldn’t mean that every project would magically become possible – the new development would still need to be economically viable – but it would mean that a transport project’s ability to act as an agent of change would not be stymied by planning. Land readjustment is one way that we could do this.

The second conclusion relates to the types of schemes that we should try to build. Planning reform is very, very hard. As long as development is stymied by planning, we need to focus on schemes that do not depend for their viability on inducing demand.

The DLR is one such scheme. It was extremely cheap, which means that using the rough-and-ready comparison with GDP/capita, which in 1992 was £27,498 in today’s money, it needed fewer than 10,000 daily journeys to be viable after five years. It is true that the DLR was initially grossly underbuilt, but this was solved with upgrades: longer trains, the Bank extension, bigger stations and ultimately the Jubilee Line Extension and Elizabeth Line to improve access to Canary Wharf.

More generally, this militates in favour of through running. To repeat the table in the full piece, many of our cities have a vast amount of suburban railway, more than many other comparable cities in the developed world:

I think that schemes to introduce (or increase) the amount of through running in Manchester, Birmingham and Glasgow can all be justified purely on the basis of the benefits they would bring to the existing network. They will also catalyse development. Probably not as much as we would like. But they give us optionality. Even if they are unable to be agents of change, through running projects will be worthwhile in and of themselves.