Tram-Trains

I recently covered the Stadtbahn, a mode of rail transportation running as rapid transit (almost always a subway) in city center and as a tramway farther out. The tram-train is the opposite kind of system: it runs as a tramway within the city, but as rapid transit farther out. There’s a Human Transit blog post about this from 2009, describing how it works in Karlsruhe, which invented this kind of service pattern. Jarrett is bearish on the tram-train in most contexts, giving a list of required patterns that he says is uncommon elsewhere. It’s worth revising this question, because while the tram-train is not very useful in an American context, it is in countries with discontinuous built-up areas, including Germany and the Netherlands but also Israel. Israeli readers may be especially interested in how this technology fits the rail network away from the Tel Aviv region.

What is a tram-train?

Let’s dredge the 2*2 table from the Stadtbahn post:

Slow in centerFast in center
Slow in outlying areasTramwayStadtbahn
Fast in outlying areasTram-trainRapid transit

The terms fast and slow are again relative to general traffic. The Paris Métro averages 25 km/h, less than some mixed-traffic buses in small cities, but it still counts as fast because the speed in destinations accessed per hour is very high.

Be aware that I am using the terms Stadtbahn and tram-train to denote two different things, but in Karlsruhe the system is locally called Stadtbahn. German cities use the term Stadtbahn to mean “a tramway that doesn’t suck,” much as American cities call a dazzling variety of distinct things light rail, including lines in all four cells of the above table. Nonetheless, in this post I am keeping my terminology distinct, using the advantage of switching between different languages and dialects.

Tram-trains and regional rail

The Karlsruhe model involves trains running on mainline track alongside mainline trains, diverging to dedicated tramway tracks in the city, to connect Karlsruhe Hauptbahnhof with city center around Marktplatz. This also includes lines that do not touch the mainline, like S2, but still run with higher-quality right-of-way separation outside city center; but most lines run on mainline rail part of the way.

North American light rail lines, with the exception of the Boston, Philadelphia, and San Francisco Stadtbahn systems, tend to run as tram-trains, but never have this regional rail tie-in. They run on entirely dedicated tracks, which has two important effects, both negative. First, it increases construction costs. And second, it means that the shape of the network is much more a skeletal tramway map than the more complicated combination of an S-Bahn and a tramway that one sees in Karlsruhe. San Diego has a short segment sharing tracks with freight with time separation, but the shape of the network isn’t any different from that of other American post-1970s light rail systems, and there’s an ongoing extension parallel to a mainline railroad that nonetheless constructs a new right-of-way.

In this sense, the Karlsruhe model can be likened to a cheaper S-Bahn. S-Bahn systems carve new right-of-way under city center to provide through-service whenever the historic city station is a terminus, such as in Frankfurt, Stuttgart, Munich, or German-inspired Philadelphia. They can also build new lines for more expansive service, higher capacity, or a better connection to city center, like the second S-Bahn trunk in Hamburg; Karlsruhe itself is building a combined road and rail tunnel, the Kombilösung, after a generation of at-grade operation. The tram-train is then a way to achieve some of the same desirable attributes but without spending money on a tunnel.

It follows that the tram-train is best when it can run on actual regional rail tracks, with good integration with the mainline system. It is a lower-speed, lower-cost version of a regional rail tunnel, whereas the North American version running on dedicated tracks is a lower-cost version of a subway. Note also that regional rail can be run at different scales, the shorter and higher-frequency end deserving the moniker S-Bahn; the Karlsruhe version is long-range, with S1 and S11 reaching 30 km south of city center and S5 reaching 70 km east.

Where is a tram-train appropriate?

Jarrett’s 2009 post lays down three criteria for when tram-trains work:

  • The travel market must be small enough that an S-Bahn tunnel is not justified.
  • The destination to be served isn’t right next to the rail mainline.
  • The destination to be served away from the mainline is so dominant that it’s worthwhile running at tramway speeds just to get there and there aren’t too many people riding the line beyond it.

The center of Karlsruhe satisfies the second and third criteria. It is borderline for the first – the region has maybe a million people, depending on definitions, and the city proper has 312,000 people; the Kombilösung is only under-construction now and was not built generations ago, unlike S-Bahn tunnels in larger cities like Munich.

Jarrett points out that in the urban world he’s most familiar with, consisting of the United States, Canada, Australia, and New Zealand, it is not common for cities to satisfy these criteria. He does list exceptions, for example Long Beach, where the Blue Line runs in tramway mode before heading into Los Angeles on a mostly grade-separated right-of-way, whereupon it goes back into the surface in Downtown LA before heading into an under-construction tunnel. But overall, this is not common. City centers tend to be near the train station, and in the United States there’s such job sprawl that just serving one downtown destination is not good enough.

That said, the Long Beach example is instructive, because it is not the primary city in its region – Los Angeles is. I went over the issue of outlying S-Bahn tunnels a year ago, specifying some places where they are appropriate in Israel. The tram-train must be a key tool in the planner’s box as a cheaper, lower-capacity, lower-speed version of the same concept, diverging from the mainline in tramway mode in order to serve a secondary center. Karlsruhe itself is a primary urban center – the only time it’s the secondary node is when it connects to Mannheim, and that train doesn’t use the tramway tracks – but a secondary tram-train connection is being built in outlying areas there, namely Heilbronn.

Different models of urban geography

In the American model of urban geography, cities are contiguous blobs. Stare at, for example, Chicago – you’ll see an enormous blob of gray stretching in all directions. Parkland is mostly patches of green in between the gray, or sometimes wedges of green alternating with wedges of gray, the gray following commuter railroads and the green lying in between. Boundaries between municipalities look completely arbitrary on a satellite map.

In the German model of urban geography, it’s different. Look at Cologne, Frankfurt, Mannheim, or Stuttgart – the built-up area is surrounded by green, and then there are various suburban towns with parkland or farmland in between. This goes even beyond the greenbelt around London – there’s real effort at keeping all these municipalities distinct.

I don’t want to give the impression that the United States is the weird one. The contiguous model in the United States is also common in France – Ile-de-France is one contiguous built-up area. That’s how despite being clearly a smaller metropolitan region than London, Paris has the larger contiguous population – see here, WUP 2007, and see also how small the German and Dutch urban areas look on that table. Urban agglomeration in democratic East Asia is contiguous as in the US and France. Canada looks rather American to me too, especially Vancouver, the city both Jarrett and I are the most familiar with, while Toronto has a greenbelt.

This distinction moreover has to be viewed as a spectrum rather than as absolutes. Boston, for example, has some of the German model in it – there’s continuous urbanization with inner suburbs like Cambridge and Newton, but beyond Route 128, there are many small secondary cities with low density between them and the primary center. Conversely, Berlin is mostly American or French; the few suburbs it has outside city limits are mostly contiguous with the city’s built-up area, with the major exception of Potsdam.

The relevance of this distinction is that in the German or Dutch model of urban geography, it’s likely that a railway will pass through a small city rather far from its center, fulfilling the second criterion in Jarrett’s post. Moreover, this model of independent podlike cities means that there is likely to be a significant core, which fulfills the third criterion. The first criterion is fulfilled whenever this is not the center of a large metropolitan area.

It’s not surprising, then, that the Karlsruhe model has spread to the Netherlands. This is not a matter of similarity in transport models: the Netherlands differs from the German-speaking world, for examples it does not have monocentric S-Bahns or S-Bahn tunnels and it builds train stations with bike parking where Germany lets people bring bikes on trains. Nonetheless, the shared model of distinct municipalities makes tram-train technology attractive in South Holland.

Israel and tram-trains

In Israel, there are very few historic railways. A large share of construction is new, and therefore has to either swerve around cities or tunnel to enter them, or in a handful of cases run on elevated alignments. Israel Railways and local NIMBYs have generally preferred swerving.

Moreover, the urban layout in Israel is very podlike. There do exist contiguous areas of adjacent cities; Tel Aviv in particular forms a single blob of gray with Ramat Gan, Givatayim, Bni Brak, Petah Tikva, Bat Yam, and Holon, with a total population of 1.5 million. But for the most part, adjacent cities are buffered with undeveloped areas, and the cities jealously fight to stay this way despite extensive developer pressure.

The final important piece in Israel’s situation is that despite considerable population growth, there is very little rail-adjacent transit-oriented development. The railway was an afterthought until the Ayalon Railway opened in 1993, and even then it took until last decade for mainline rail to be a significant regional mode of transport. The state aggressively builds new pod-towns without any attempt to expand existing towns toward the railway.

The upshot is that all three of Jarrett’s criteria for tram-trains are satisfied in Israel, everywhere except in and around Tel Aviv. Tel Aviv is large enough for a fully grade-separated route, i.e. the already-existing Ayalon Railway. Moreover, because Tel Aviv needs full-size trains, anything that is planned to run through to Tel Aviv, even as far as Netanya and Ashdod, has to be rapid transit, using short tunnels and els to reach city centers where needed. A tram-train through Ashdod may look like a prudent investment, but if the result is that it feeds a 45 meter long light rail vehicle through the Ayalon Railway then it’s a waste of precious capacity.

But Outside Tel Aviv, the case for tram-trains is strong. One of my mutuals on Twitter brings up the Beer Sheva region as an example. The mainline going north has a station called Lehavim-Rahat, vaguely tangent to Lehavim, a ways away from Rahat. It could get two tramway branches, one diverging to the built-up area of Lehavim, a small suburb that is one of Israel’s richest municipalities, and the other to Rahat, one of Israel’s poorest. There are also interesting options of divergence going south and east, but they suffer from being so far from the mainline the network would look scarcely different from an ordinary tramway.

Beer Sheva itself would benefit from tramways with train through-service as well. The commercial center of the city is close to the train station, but the university and the hospital aren’t, and are not even that close to the subsidiary Beer Sheva North station. The station is also awkwardly off-center, lying southeast of the city’s geographic center, which means that feeding buses into it with timed transfers screws internal connections. So tramway tracks on Rager Boulevard, cutting off Beer Sheva North for regional trains, would do a lot to improve regional connectivity in Beer Sheva; intercity trains should naturally keep using the existing line.

In the North, there are similar examples. Haifa is not going to need the capacity of full-size trains anytime soon, which makes the case for various branches diverging into smaller cities to provide closer service in tramway mode strong. Unlike in Beer Sheva, the case for doing so in the primary center is weak. Haifa’s topography is the stuff of nightmares, up a steep hill with switchback streets. The mainline already serves the Lower City well, and climbing the hill is not possible.

This creates an interesting situation, in which the technology of the tram-train in the North can be used to serve secondary cities like Kiryat Ata and Tirat Carmel and maybe enter the Old City of Acre, but the operational pattern is really that of a Stadtbahn – fast through Haifa and up most of the Krayot, slow through smaller suburbs.

What is Neoliberalism, Anyway?

It increasingly looks like the cause of high construction costs in the English-speaking world is the trend of the privatization of the state since the 1980s. Instead of public planning departments, there is growing use of consultants. This trend is intensifying, for example with increasing use of design-build contracts, introduced into Canada just before costs exploded.

Q. Does this mean neoliberalism is to blame?

A. Not really.

Nearly every political and economic trend in the last 40 years in a developed country can be connected with neoliberalism. The transition in South Korea from military government to something like social democracy has been reasonably compatible with the Washington Consensus principles.

In fact, two opposite trends have both been criticized as neoliberal: the move from income support to workfare in the Clinton administration in the 1990s, but also calls from some liberals, greens, and social democrats today for basic income.

Even things that are mostly about things that are what people on the left criticize as neoliberalism are not necessarily about the privatization of the state. Any of the following agenda items can be plausibly called neoliberal:

  • Privatization of state-owned enterprises like the mail, the national airline, the national railway, road maintenance, and health care.
  • Reduction in top income tax rates from historic levels that were sometimes higher than 90% to something closer to today’s 50% in various rich Western countries.
  • Liberalization of foreign exchange and foreign investment.
  • Voucher systems for public services like schools.
  • Fiscal and monetary austerity.

The key here is that none of these items is exactly privatization of planning. Germany had welfare reform in the Schröder era, Hartz IV, that SPD and the Greens don’t even like anymore. It’s had austerity budgets under Angela Merkel, and inflation has been below the 2% target. The Netherlands privatized health care. Sweden has contracted out operations of rail and many other infrastructure services.

However, the privatization of the state itself is mostly a British and American program, which has spread to other English-speaking countries through their cultural cringe. Under Macron, the most neoliberal of French leaders, Grand Paris Express staffed up as a public planning agency, rather than contracting everything out to consultants.

Even when France engages in design-build, it’s not the same as in the Anglosphere. Design-build in France means that the three teams that are typically kept separate – public planning, private design, private construction – talk to one another more regularly, still with public oversight. There is still strong civil service, and no impetus so far to privatize it or discount its advice on the American and increasingly British model.

There is neoliberalism in Japan, and in Germany, and in France, and in Scandinavia. And in none of these do we see Anglo construction costs. This matters.

Stadtbahn Systems

I made an off-hand remark about subway-surface systems, called Stadtbahn in German (as is, confusingly, the fully grade-separated east-west Berlin S-Bahn line), regarding a small three-line single-tail network that Brooklyn could build. I also talked about it in a little more detail last year. I want to go more deeply into this now. It’s a public transportation typology that’s almost nonexistent outside Germany and Belgium; Tel Aviv is building one line, and the US has three but two of those are from more than 100 years ago. But there are interesting examples of good places to use this technology elsewhere, especially elsewhere in Europe.

What is the Stadtbahn?

The Stadtbahn (“city rail”), or the subway-surface line in US usage, is an urban line running light rail vehicles, with grade separation in city center and street running outside city center. All examples I know of are in fact underground in city center, but elevated lines or lines running in private rights-of-way could qualify too, and in Cologne, there’s a semi-example over a bridge dropping to the surface at both ends.

It’s best illustrated as a 2*2 grid:

 Slow in centerFast in center
Slow in outlying areasTramwayStadtbahn
Fast in outlying areasTram-trainRapid transit

The terms fast and slow are relative to general traffic, so a mixed-traffic bus in a low-density city that averages 30 km/h is slow whereas the Paris Métro, which averages 25 km/h, is fast; the speed in km/h may be higher on the bus, but the speed in destinations accessed per hour is incomparably higher on the Métro.

The tram-train is confusingly also called Stadtbahn in Germany, for example in Karlsruhe; this is nearly every light rail built in North America. It is not the topic of this post.

What is the purpose of the Stadtbahn?

Historically, Stadtbahn systems evolved out of pure surface tramways. City center congestion made the streetcars too slow, so transit agencies put the most congested segments underground. This goes back to Boston in 1897 with the Tremont Street Subway and Philadelphia in 1906 with the Subway-Surface Lines. The contrast both in that era and in the era of Stadtbahn construction in Germany from the 1960s to the 80s is with pure subways, which are faster but cost more because the entire route must be underground.

Stadtbahns always employ surface branching. This is for two reasons. First, there’s more capacity underground than on the surface, so the higher-capacity rapid transit segment branches to multiple lower-capacity tramways to permit high throughput. And second, there’s generally less demand on the outer segments than in the center – lines with very strong demand all the way tend to turn into full subways.

This is therefore especially useful for cities that are not huge. In a city the size of Cologne or Stuttgart or Hanover, there isn’t and will never be demand for a rapid transit system with good citywide coverage. Instead, there is something like a sector principle. For example, in Cologne, the Deutz side of the city, on the right bank of the Rhine, has service to city center on the S-Bahn, on tramway lines over the Deutzer Bridge branching on the surface, and on tramway lines over the Mülheimer and Severin Bridges feeding into the north-south ring Stadtbahn. Smaller cities have simpler systems – Hanover for has three underground trunk lines meeting at one central station, and Dortmund has three meeting in a Soviet triangle. This maintains good coverage even without the budget for many rapid transit lines.

Where are Stadtbahns appropriate?

Cities should consider this technology in the following cases:

  • The city should not be too big. Tel Aviv is too big for this, and people in Israel are starting to recognize this fact and, in addition to the under-construction three-line Stadtbahn system are proposing a larger-scale three-line fully grade-separated metro system. If the city is big enough, then a full metro system is justified.
  • There should be a definitive city center for substantial traffic to funnel to. The purpose of the Stadtbahn is to have comparable throughput to that of a metro, albeit with shorter trains.
  • There should be wide swaths of sectors of the city where having multiple parallel lines is valuable. This, for example, is the case in cities that are not exceptionally dense and cannot expect transit-oriented development to completely saturate one metro corridor.
  • The street network should not be too gridded, because then the sector-based branching is more awkward, and the combination of rapid transit to city center and a surface transit grid can be powerful, as in Toronto.
  • There should be too much city center congestion for a pure surface system to work, for example if most streets are very narrow and traffic funnels to the few streets that can use

These circumstances are all common to German urbanism: city centers here are strong, but residential density peaks at 15,000/km^2 or thereabouts in near-center neighborhoods and drops to 3,000-6,000/km^2 farther out. Moreover, Germany lacks huge cities, and of the largest four urban cores – Berlin, Hamburg, Munich, Frankfurt – three have full rapid transit systems. Finally, grids are absent here except at very small scale, as in Mannheim.

However, these are not unique features to Germany. They’re common around Europe. European cities are not very big, and the only ones that can genuinely fill any subway line with transit-oriented development are a handful of very big, very rich ones like London and Paris. Even Stockholm and Munich have to be parsimonious; they have have full metro systems with branching.

The French way of building rapid transit does not employ the Stadtbahn, and perhaps it should. In a city the size of Bordeaux or Nice, putting a tramway underground in city center and then constructing new branches to expand access would improve coverage a lot.

This is likely also the case in Italian cities below the size class of Milan or Rome. Many of these cities are centered around Renaissance cores with very narrow streets, which are nonetheless auto-centric with impossibly narrow sidewalks, Italy having nearly the highest car ownership in Europe. Finding one to three good corridors for a subway and then constructing tramways funneling into them would do a lot to speed up public transit in those cities. Bologna, for example, is planning a pure surface tramway, but grade-separated construction in the historic center would permit trams to have decent coverage there without having to slow down to walking speed.

Are there good examples outside Europe?

Yes! From my original post from 2016, here is one proposal for New York:

The B41 could be a tramway going between City Hall and Kings Plaza, using two dedicated lanes of the Brooklyn Bridge. In that case, the line would effectively act as subway-surface, or more accurately elevated-surface: a surface segment in Brooklyn, a grade-separated segment between Manhattan and Brooklyn. Subway-surface lines should branch, as all current examples do (e.g. Boston Green Line, Muni Metro, Frankfurt U-Bahn), because the subway component has much higher capacity than the surface components. This suggests one or two additional routes in Brooklyn, which do not have strong buses, but may turn into strong tramways because of the fast connection across the river to Manhattan. The first is toward Red Hook, which is not served by the subway and cut off from the rest of the city by the Gowanus Expressway. Unfortunately, there is no really strong corridor for it – the streets are not very wide, and the best for intermediate ridership in Cobble Hill and Carroll Gardens require additional twists to get into the core of Red Hook. Court Street might be the best compromise, but is annoyingly a block away from the F/G trains, almost but not quite meeting for a transfer. The second possible route is along Flushing Avenue toward the Navy Yard; it’s not a strong bus by itself, but the possibility of direct service to Manhattan, if a Flatbush tramway preexists, may justify it.

Note that this proposal is opportunistic: Brooklyn Bridge just happens to be there and point in the right direction for at least one strong surface route in Brooklyn, and conversely would connect too awkwardly to the subway. In a city the size of New York, Stadtbahn lines must be opportunistic – if the region intentionally builds new river crossings then they must carry the highest-capacity mode of transportation, which is rapid transit, not a light rail variant.

American cities smaller than New York are often very big by European standards, but also very decentralized. This hurts the Stadtbahn as a mode – it really only works for a monocentric city, because if there are multiple centers, then all but the primary one get slow transit. The Rhine-Ruhr notably uses the S-Bahn, which is rapid transit, to connect its various cities, and only run Stadtbahn service internally to each center, like Cologne or Dortmund.

There are a number of places in the United States where burying a light rail line in city center is advisable, but this is for the most part conversion of a tram-train to rapid transit, for examples in Portland and Dallas. The only example that come to mind of a decent Stadtbahn in the US that doesn’t already exist is Pittsburgh, converting the BRT system to rail.

Outside the United States, I get less certain. Canada is bad geography for a Stadtbahn because of its use of grid networks; Ottawa may be good for a Stadtbahn using the Confederation Line tunnel, but that’s probably it. Australia may be better, combining decently strong city centers with very low residential density; transit-oriented development potential there is very high, but it could plausibly come around multiple distinct corridors as well as regional rail stations. Melbourne’s tramways thus may be a candidate for Stadtbahn conversion.

In both East Asia and in the developing world, it’s likely best to just build full metros. East Asian cities are big and have high rates of housing construction (except Hong Kong). I can see a Stadtbahn succeed in Taichung, extending the under-construction Green Line on the surface and building intersecting lines, but that’s probably it. Kaohsiung already has a (very underused) subway, what I think is Daejeon’s best next corridor on top of Line 1 and the planned Line 2 is unusually bad for a Stadtbahn because the streets are too gridded west of the center, Daegu is too gridded as well.

A similar set of drawbacks is also true for the developing world. The urban population of the developing world tends to cluster in huge cities. Moreover, these cities tend to have high residential density but low city center job concentration; the Addis Ababa light rail is bad at serving people’s work trips because so few people work in the center. Finally, the developing world has high rates of increase in urbanization, which make future-proofing systems with higher capacity more valuable.

Some Examples of Falling Costs

Question. Are there any historical examples of construction costs actually falling in a city, rather than just rising slower than elsewhere?

Answer. Yes! Not many, though.

I know of three examples, but the first is fairly irrelevant and is included here for completeness.

Example 1. London’s District line, going by Wikipedia data, cost 3 million pounds, which in today’s money translates to $90 million per kilometer. This was astonishingly expensive, and even today London Underground extensions, as opposed to Crossrail, cost less than that relative to British GDP per capita. The reason for the high cost: the line was built cut-and-cover without any street to go under, so it needed to carve a new right-of-way through Kensington, demolishing houses in an expensive area. No further cut-and-cover lines were built. Costs fell to about $30 million per kilometer with the invention of deep-bore tunneling a generation later; today, bored tunnel costs more than cut-and-cover, but with the technology of the late 19th and early 20th century, this was not the case.

Example 2. Milan built its first two lines very cheaply; in today’s money, M1 cost around $50 million per kilometer. It was built using a method invented specifically for the city’s narrow Renaissance streets called the Milan method or cover-and-cut, allowing vertical construction with retaining walls rather than sloped ones that require more street width. M2 was very cheap as well, but M3’s costs were much higher, I believe around $250 million per km in today’s money, built in the 1980s at the peak of Milanese corruption. Costs fell dramatically after a series of anti-corruption prosecutions that put much of the Italian political elite in prison. The Passante Railway was in today’s money around $140 million per km, not all underground, but it’s regional rail with difficult city center construction under three older lines. The more recent lines, M5 and M4 (in this order), run up to $120-160 million per km.

Example 3. Istanbul began building its subway system with M2, M3, and M4; the first Istanbul Metro line, M1, is light rail and its original section had very little tunneling. It used Italian designs and costs were low, not much more than $100 million per kilometer, but subsequently value engineering has led to slightly lower costs. The city had a learning process in which it reduced station footprints to save money, engaged in more extensive prior engineering before putting out new lines to bid, and generally gained experience in managing a project. Newer lines have cost slightly less, for example around $80 million/km for M5, all underground.

The angle of cleaning up corruption and building up state capacity is probably relevant – probably. Italy and Turkey remain very corrupt and clientelist states. In Turkey, the former mayor of Istanbul openly said he was going to prioritize metro construction in neighborhoods that voted for AKP, and then when the opposition won the city election the state stopped giving it money for new lines; construction goes on because the new mayor went to the European Investment Bank for financing. In Italy, for all the clientelism elsewhere, public-sector engineering is fiercely depoliticized and professionalized nowadays.

I might even speculate, without much knowledge yet (we’re still early in the work in Istanbul and even earlier in Milan), that Southern Europe may have such reputation for corruption that it has mundane mechanisms that professionalize public works. The clientelism in Turkey as far as we can tell extends to macro-level decisions of where to build lines, and evidently Istanbul managed to identify alternative sources of financing to the Erdoganist state.

If I’m right, then these same mechanisms of anti-corruption and public-sector professionalization can also be replicated in other parts of the world with state capacity problems. This cannot possibly be everything – Milan reduced its costs from levels that were not extremely high, and Istanbul was cheap from the start – but it does point in a more optimistic direction.

Job Sprawl as Deurbanization

A few years ago, Aaron Renn was writing, I think about the General Electric headquarters’ move from suburban New York to Downtown Boston in 2016, that in the future, city center jobs would go to high-value industries like corporate HQs and professional services, and then lower-end stuff like call centers would go in suburban office parks. At the time I didn’t understand the full meaning of this – I was still thinking of employment in a narrow city center of a few blocks rather than a broader region, like the 100 km^2 zone I use to compare the US with Canada and France because that’s the most granular data I have in the latter two countries. But in retrospect, Aaron was getting at a dangerous trend in which job markets deurbanize. This is not a new trend – office park sprawl goes back to the 1970s, and industrial sprawl even earlier – and to some extent it’s less about deurbanization and more about the urban job market reaching maximum size. But whatever the history of it, it’s a serious threat to economic performance – and the solution to it requires better public transportation.

Cities as job markets

I’ve written before about production theory. The only thing I have to add on the theory side is that since I wrote that post, I was at a talk that Alain Bertaud gave at Marron, about urbanization. The main topic of the talk was about urban growth and sprawl in the developing world, but at the beginning of the presentation, he gave some remarks about cities and corona. Zoom meetings like the one we had, he warned, were fine, but cities are fundamentally job markets that succeed through spontaneous interaction, and this spontaneity does not exist with remote work. This is to a large extent the new urban geography thesis of Paul Krugman or the work of Ed Glaeser – cities exist as places of production first, and this production requires close proximity.

Now, close proximity depends on technology. In a city with the transport technology of London circa 1800, close proximity means the scope of the City of London, and even 5 km is uncomfortably far. In a city with cars and highways, the distance is much greater – but it is not the same as commute distance. A half-hour drive is not spontaneous. When I asked American friends and coworkers about their productivity through the spring corona lockdowns, a Boston lawyer told me that lawyers wouldn’t even travel midday for clients for 20-30 minutes, since their time was too valuable – they’d schedule conference calls.

This does not mean that the entire work market has to be within such a short distance. It certainly helps, but different industries can cluster in different parts of the city. But there is a maximum distance within which the city is recognizably a single job market.

Aaron Renn’s bifurcation

Aaron talks about bifurcation a lot, between winners and losers. He relates the move of large corporate HQs to city centers to this bifurcation: city centers win by having higher-value added, higher-paying jobs, everyone else gets saddled with lower-end jobs. Moreover, these lower-end jobs are commodities – a call center can be anywhere – and therefore they compete on price and not quality, frustrating the attempt of any region on the margins of the US to climb up the value chain.

That said, even the sort of job sprawl of the 1970s, spearheaded by big companies’ move out of city centers to rich suburbs like GE to Fairfield and IBM to Armonk, represents the same threat to urban productivity. That was driven by snobbishness – the elite suburbanized, and then dragged jobs outside the city with it, for example GE did partly on spurious grounds of resilience in face of nuclear war destroying city centers. Today, the city gains higher-end jobs at the expense of the suburbs, the opposite of the situation in the 1970s. But the same situation of jobs outside one major core persists.

Is this polycentricity?

No. It’s become fashionable to speak of polycentric cities as the next evolution, to decongest old cores. But doing so requires the urban geography to have centers. I pointed out previously that Los Angeles may claim to be polycentric but is just weak-centered – the secondary centers have a few tens of thousands of jobs each at most. This is not like the big city centers one finds in Kyoto, Osaka, and Kobe, or even in the Rhine-Ruhr or Randstad.

Keihanshin, the Rhine-Ruhr, and Randstad are all agglomerations of historic cities. It is possible to also form polycentric regions out of new development – for example, Yokohama was founded as a 19th-century treaty port and then grew as a Tokyo suburb. Both New York and Paris have moved their central business districts by a few kilometers gradually, New York from Lower Manhattan to Midtown and Paris from Les Halles to around the Opera; both also have near-center business centers, like Long Island City or La Défense. Even then there’s likely to be some efficiency loss in decentralizing city center jobs this way, but it’s still easier to shuttle between Times Square and World Trade Center than between either and New Brunswick.

The public transit solution

In the 1970s, the abandonment of city centers was motivated by a desire to escape their poverty and a belief that the suburbs were the future. Urban poverty still exists but inner-urban wealth is considerable and increasing, and the belief that the suburbs were the future turned out to be incorrect – one cannot be a suburb of nowhere.

The model of suburbanization that can be sustained is one built from the late 19th century to about the 1950s and early 60s: jobs stay in the city, people go wherever.

Doing so requires three things: offices, dwellings, and a way of getting between them.

Offices mean commercial upzoning – some American cities are good about it, but the ones with the most demand, like New York, aren’t. In general there’s little appetite for commercializing near-center neighborhoods in the US, whereas Europe is looser about it and therefore new firms can sprout a few subway stops outside the primary center, for example Spotify two stops outside T-Centralen. Residences likewise require upzoning, especially for mid- and high-rise apartment buildings near subway stations where they exist and have capacity.

But in many cases, it’s required to also build up public transportation. Big central business districts feature hundreds of thousands of people converging on a small area at the peak, and the biggest go up into the millions. The highest-capacity form of transportation is required, which is rapid transit, never cars or surface transit.

Rapid transit and city centers are symbiotic, now as in 1910. An expansive rapid transit system, with high service quality, is required to serve city centers from multiple directions; and city centers are required to give people something to take the trains to, or else they’ll just drive everywhere and only take the train to the sports stadium or the airport.

And ultimately, city centers are required for economic efficiency, because of the importance of proximity for spontaneous economic and social interactions. Rapid transit also benefits from high efficiency – it’s very cheap to operate compared with the cost of car ownership. The alternative is a kind of deurbanization, in which people may live at high density relative to travel speeds but don’t form large clusters enabling the highest productivity.

Urbanization in Europe and East Asia

After writing this post about the urban layout of Germany and high-speed rail, I got interested in city size in Europe and Asia more generally. East Asians live in much larger cities than Europeans, and I’m not just talking about Tokyo and Seoul. I was vaguely aware this was the case, but underestimated the size of this effect. By modern definitions of metropolitan areas as those with at least 1 million people, Europe is not majority-urban.

Definitions

The definitions are hard to harmonize – there are international lists but I don’t fully trust them, many just collating national statistics (like the UN’s) and others making some judgments I am uncertain about, like CityPopulation. I believe the effect I’m discussing is robust to changes in definitions, but its magnitude may be different. I use the following definitions:

  • In Japan, I’m using the major metropolitan area. This is the part I’m least certain about, hence a subsidiary definition based on smaller metropolitan employment areas (MEAs).
  • In South Korea, there are no definitions of a metro area. Seoul is defined as Seoul + Gyeonggi + Incheon, the other cities use city limits.
  • In Taiwan, as in Korea, cities are defined by city limits, except that the capital region consists of Taipei + New Taipei + Taoyuan.
  • In France, a metropolitan area is an aire urbaine.
  • In Germany, a metropolitan area is either the Rhine-Ruhr region, or a Verkehrsverbund.
  • In Italy and Spain, a metropolitan area is cobbled from NUTS-3 regions in 2021 definitions; this can also be used in Germany as a secondary check, but not in France, as French NUTS-3 regions are departments, which aren’t granular enough. Population figures can be found here.

Results

Japan

Using the MMA definition, slightly more than half the population lives in the metro areas of Tokyo, Osaka, and Nagoya. Nagoya goes barely over the 50% mark, so the median resident lives either in Nagoya (9.4 million people) or Fukuoka (5.5 million). 56% of the population lives in these four metropolitan areas, 64% lives in metro areas of at least 2 million, 70% lives in metro areas of at least 1 million.

The MEA definition splits polycentric regions like Osaka-Kyoto-Kobe and Fukuoka-Kitakyushu, and also thinks Nagoya is much smaller. On that definition, the median Japanese resident lives in Sapporo (2.4 million), and overall 51% live in metro areas of at least 2 million and 61% in metro area of at least 1 million.

South Korea

The metropolitan area of Seoul has 26.2 million people, which is just a hair over half the population of the country. But it’s so close it’s obligatory to also include the next largest region, which is Busan, population 3.4 million. Together, they have 57% of total population. Daegu adds 2.5 million, and with it 62% of the country’s population lives in cities of at least 2 million. Down to 1 million, the share rises to 70%.

Taiwan

Taipei, New Taipei, and Taoyuan collectively have 8.9 million people, or 38% of total population; Taoyuan is somewhat independent, but Taipei and New Taipei are very closely integrated. The median Taiwanese lives in one of the two second cities, Kaohsiung and Taichung, each with around 2.8 million people; together, 61% live in one of these five municipalities. The next city, Tainan, has 1.8 million people, and including it raises the proportion to 69%.

France

The Paris region has 20% of the country’s population, and the urbanity of the population craters afterward. Only one more metro area has more than 2 million, Lyon; together, the two regions have 23% of national population. Five more metro areas have between 1 and 2 million people; together, these seven regions have 33% of total population. Even allowing metro areas down to 500,000 only increases the metropolitan share to 44%. The median is either Valenciennes or Le Mans, populations 369,000 and 347,000 respectively.

Germany

By any definition, we lump the Rhine-Ruhr into one region of 11,430,361 people, based on the Metropolregion definition (see data here) – usually the German Metropolregionen are way too loose, but in this polycentric region the definition agrees with regional rail extent and with a number of other metro area population lists. In theory this is Germany’s only megacity, but because there are at least 7 independent centers, it doesn’t feel like it’s another London or Paris. Still, it’s 14% of total population.

Using the rather loose Verkehrsverbund limits (still tighter than the Metropolregionen!), we add Frankfurt (6.7 million), Berlin-Brandenburg (6.2 million), Hamburg (3.5 million), Rhine-Neckar (3 million), Munich (2.9 million), Nuremberg (2.8 million), Stuttgart (2.4 million), Bremen (2.2 million), Leipzig-Halle (1.7 million), Karlsruhe (1.4 million), Hanover (1.2 million), Dresden (0.8 million). As a note of caution, some of these regions are drawn very loosely, like Nuremberg, Bremen, and Frankfurt; Hanover and Munich are a lot tighter, Hanover corresponding to a county and Munich to the S-Bahn extent.

On these very loose definitions, the 2+ million regions still only sum up to 49% of the population, and the 1+ million regions to 55%, the median German living in Leipzig-Halle.

Italy

Italy has many distinct definitions of metro area, some disagreeing on Milan by factors approaching 2. Eurostat, using provinces as its level of analysis, has metropolitan area for Milan (4.4 million), Rome (4.3 million), Naples (3.1 million), Turin (2.3 million), Brescia (1.3 million), Palermo (1.3 million), Bari (1.3 million), Bergamo (1.1 million), Catania (1.1 million), Bologna (1 million), Florence (1 million), Padua (0.9 million), Verona (0.9 million), Venice (0.9 million), Genoa (0.8 million), Perugia (0.7 million), Messina (0.6 million), Taranto (0.6 million), Reggio Emilia (0.5 million), Parma (0.5 million), Cagliari (0.4 million), Prato (0.3 million).

The four main metro areas comprise 23% of Italy’s population. Going down to 1 million raises the proportion to 37%. These regions combined have only 48% of Italy’s population. This does not mean Italy is majority-rural – the remaining provinces are full of cities in their own right, some larger than Prato or Cagliari – but it’s hard to detect that level of granularity.

Spain

In Spain, as in Italy, the division isn’t too granular, at the level of a province. We have metro areas for Madrid (6.7 million), Barcelona (5.6 million), Valencia (2.5 million), Seville (2 million), Alicante (1.8 million), Málaga (1.7 million), Murcia (1.4 million), Cádiz (1.2 million), Biscay (1.1 million), A Coruña (1.1 million), Asturias (1 million), Zaragoza (1 million), Vigo (0.9 million), Tenerife (0.9 million), Granada (0.9 million), Mallorca (0.9 million), Gran Canaria (0.9 million), Córdoba (0.8 million), Gipuzkoa (0.7 million), Navarre (0.6 million), Cantabria (0.6 million), Valladolid (0.5 million), Álava (0.3 million).

The two main metro areas are together 26% of Spain’s population; the next two, down to 2 million people, add up to 35%; the next ones going down to 1 million go up to 55%, the median Spaniard living in a metropolitan province of 1.1 million.

Quick Note: What is This for?

I’d like to propose a test whenever someone proposes a new conversation into public transit: what tangible changes come from this that would not have come about otherwise?

I put this out because I’ve seen a lot of people discuss the impact of corona on transportation and bring up solutions that they believed before the crisis began and that are at best loosely related to the pandemic. In the US these are BRT, higher off-peak bus and train frequency, bus network redesigns, and off-board fare collection, among others. All of these have been popular among American transit advocates for years, but now there’s spin talking about how they’re useful for corona.

Another buzzword in the US now is equity, in which every person is expected to figure out how to be less racist, which in practice means justifying what they already believed as a solution to racism. It’s weird – I asked on Twitter what the most useful transportation investments are from an equity perspective, and I got a lot of really good ideas in comments, but almost invariable they are things that are good even without the equity perspective. There are some differences in priorities and focus, but for example the value of (say) Second Avenue Subway Phase 2 is high regardless of any equity concerns.

Note that this does not mean all topical or newsworthy discussions regarding public transportation are useless. Most of them are, but there are some interesting ones. The most notable, I think, is the issue of equity for women as opposed to the more standard measures of looking at equal access for the working class or racial minorities. Nicole Badstuber, for example, wrote about it last year, and specifically mentioned an example: women walk at higher rates than men and drive at lower rates, and snow clearing priorities that had roads ahead of sidewalks were therefore sexist. Crush dummies for cars are man-sized and therefore result in cars that are less safe for smaller-size people, such as the average woman. Nicole also brings up the issue of trip chaining, which a number of commentators brought up in 2012 as well.

All of these have concrete implications that one would not have otherwise thought of: dummies should be sized for the average person and not the average man (and really have a variety of sizes to test car safety for), public transportation should be designed to facilitate trip chaining, etc.

However, this is not the typical case of trying to connect public transportation with another political idea or current event. To distinguish real additions from cases of “I am anti-racist, I like this, therefore this is anti-racist” that just create more red tape, it’s always important to ask what new concrete actions this recommends that would not be otherwise undertaken.

Corona and Europe’s Idiocy

550 new coronavirus cases in Berlin yesterday. 7,000 in Germany. 110,000 in the European Union, which at 240 per million people is even higher than the US, which is at 200/million. French hospitals are flooded with corona patients, and the state expresses its grave concerns but will still not set up quarantine hotels or universalize the use of surgical masks or do anything else that Taiwan did in less time. This is the second wave, and seven months after Taiwan showed the way how to deal with this and ended up being the only country this year to have positive economic growth to boot, Europe (and the US) still stays in its comfort zone of mass death.

It’s worth discussing the excuses, because so many of them port well to the realm of public transportation, where Europe is not so bad (there are even things East Asia can learn from us); Europe’s real disaster compared with rich Asia is in urbanism and its resistance to tall buildings. But the United States is horrific on all matters of transportation and urban redevelopment and the excusemaking there is ensuring no infrastructure can be built.

Excuse #1: the restricted comparison

The Max Planck Society (MPG) put out a statement three weeks ago, with some interesting insights about the need for a multi-pronged strategy, including contact tracing, hygiene, and social distancing. But it kept engaging in these silly intra-European comparisons, praising Germany in contrast with Britain. At no point was there any engagement with East Asia, even though we know that Taiwan has not had community spread since April, and that in Korea and Japan the current rates are about 2 and 4 daily cases per million people respectively.

Excuse #2: bullshit about culture

I’m told that there is general understanding within Germany that Taiwan and South Korea are doing far better. However, people keep making up cultural reasons why Europe can’t have what East and Southeast Asia have. This excuse unfortunately is not restricted to people who are totally unaware: a few months ago, Michel Bauwens, a Belgian degrowth advocate who lives in Thailand, talked up Thailand’s corona suppression, but attributed it to a communitarian, collectivist culture. The Thais are mass-protesting their autocratic government’s state of emergency (while wearing masks, unlike Western anti-regime protesters); what collectivism? The actual policy differences – mandatory centralized quarantine for people who test positive, mask wearing mandates – were not mentioned.

When I bring up the necessity of centralized quarantine, and even the fact that Israel used corona hotels to nearly eradicate the virus in the first wave (the second wave came from mass abandonment of social distancing – MPG is right about multi-pronged strategies), Europeans and Americans keep making a “but freedom” line. It’s strange. Yes, Thailand is autocratic. But Taiwan and South Korea are not – and they had authoritarian governments within living memory, and both are currently run by political parties that emerged out of democratic opposition to autocracy in the 1980s and 90s, and that far from becoming autocrats themselves, ceded power peacefully when they lost reelection in the past.

Excuse #3: the fake tradeoff

Many aspects of policy involve genuine tradeoffs. But many others don’t, and corona protection is one. Taiwan is the only developed country that is projecting positive economic growth in 2020. South Korea is projecting 1% contraction, the smallest contraction in the OECD, of which Taiwan is not a member. There is no economy-death tradeoff. Plowing through with reopening before the virus has been suppressed just means mass closures later and a weaker economy. The only major suppression country that is seeing economic contraction is Thailand, whose economy is based heavily on tourism and therefore more sensitive to crises outside its borders than are the industrial export-based economies of Taiwan and Korea.

Excuse #4: learned helplessness

I write occasionally about the importance of state capacity, but centralized quarantine is not some specialized technique only available to the most advanced states. It was routine in developed countries until the 1960s, when the incidence of infectious disease had fallen to a point that it was no longer necessary. The same is true of social distancing – Nigeria for example has used it and appears to be successful, with semi-decent test positivity rates and lower per capita confirmed infections than Korea.

However, various leaders keep saying “we can’t.” This is not about technical matters. Rather, it’s about political ones: we can’t established corona hotels, we can’t ban indoor dining and drinking, we can’t scale up surgical mask production like Taiwan did 8 months ago and require people to wear surgical masks in public. The only thing Europe seems capable of doing is prohibiting travel from countries that at this point often have less corona than we do.

This is learned helplessness. Risk-averse politicians who know on some level what needs to be done are still too spineless to do it, even knowing very well that successfully suppressing corona is an amazing crowd booster.

The connection with infrastructure

All of the above problems also lead to disastrous infrastructure projects. This is to some extent a problem in Germany, where “we can’t” is a common excuse for surrender to NIMBY opposition; this is why certain key high-speed rail segments have yet to be built. But it’s a truly massive drag on the English-speaking world, especially the United States. I have seen advocates engage in internal-only comparisons within the last 24 hours; the other excuses, I saw earlier this week, and many times in the last few months, with so many different American transit managers incanting “it’s not apples-to-apples” whenever Eric and I ask them about costs. One literally said “we can’t” and “it’s not possible” and is regionally viewed as progressive and forward-thinking.

In the same manner Europeans discount any knowledge produced outside of Europe and the United States, Americans discount everything produced outside their country. Occasionally they’ll glance at Canada and Britain to affirm prior prejudices. They treat foreign language fluency as either dilettantism or immigrant poverty and not as a critical skill in the modern world right next to literacy and numeracy. They’ll flail about as they die of corona and blame one another when, just as the EU flag today is twelve yellow coronaviruses on a field of blue, the US flag is fifty white viruses on a field of blue with red and white stripes.

High-Speed Rail and Cities

When preparing various maps proposing high-speed rail in Germany, I was told that it looks nice but it overfocuses on the largest cities and not about connecting the entirety of the country. I’ve seen such criticism elsewhere, asserting that high-speed rail is a tradeoff in which the thickest connections get fast trains but the long tail suffers, whereas the medium-speed system of Germany or Switzerland or Austria serves everyone. So with that in mind, let’s look at the actual population served by a Germany-wide high-speed rail program.

I made a proposal last year, but then made some additional tweaks, posted as part of a Europe-wide map. The most important tweak: the main east-west trunk line was extended to Dortmund, which trades off some additional tunneling in the Ruhr for both higher frequency on Berlin-Dortmund and fast, frequent Dortmund-Cologne and Dortmund-Düsseldorf service. To my later map I’ll add one proposal: moving the Hanover-Dortmund tracks so that trains can stop in Bielefeld. Otherwise, take the maps as given.

The question is, what population is served by those maps? The answer of course depends on what this exactly means. The sum total of the populations of the cities served – Berlin, Hamburg, Hanover, Dortmund, Würzburg, Erfurt, Mannheim, etc. – is 18.2 million, or 21% of Germany’s population. But is that the full story? This just includes central cities, where people in many nearby suburbs and satellite cities would travel to the rest of the country via the primary city center anyway.

For example, let’s go down this list of German cities. The largest city without a stop on my proposed network is Bonn. But Bonn is very close to Cologne and there are Stadtbahn subways connecting the two cities, in addition to regional rail lines; Bonn also has a shorter-range Stadtbahn to a suburban station at which some high-speed trains on the Cologne-Frankfurt line call. Is it really correct to say Bonn is unserved? Not really. So its population should be added to the 18.3 million.

Going down the list, the same can be said of Wiesbaden (near Frankfurt), Mönchengladbach (Düsseldorf) and likewise many Rhine-Ruhr cities, Halle (near Leipzig and potentially on some slower Berlin-Erfurt trains like today), Potsdam (near Berlin), Ludwigshafen (near Mannheim), and many others. Some cities remain unserved – the largest is Münster, like a few other northwestern cities not really near anything bigger or on any line – but overall this adds another 5.3 million. So we get 23.6 million, around 30% of Germany’s population.

But that list is just cities of 100,000 people or more, and there are smaller suburbs than that. These form counties (“Kreise”), which should be added as well when feasible, e.g. when they lie on S-Bahn systems of large cities or when they are right across from cities with stations, such as Neu-Ulm to Ulm. For example, the Kreise served by the Munich S-Bahn, excluding the city proper, have a total of 1.3 million people, and people in those suburbs would be connecting to the rest of Germany by train at Munich Hauptbahnhof anyway – a lower-intensity, higher-coverage network would do nothing for them.

Overall, these suburbs add another 18.7 million people. In Berlin I used this list of suburbs; elsewhere, I went by S-Bahn reach, or in a few cases used an entire region where available (Hanover, Göttingen, Fulda). There are a few quibbles on the margin in the gaps between the Frankfurt and Rhine-Neckar and in places that probably should count but aren’t on any big city S-Bahn like Frankfurt an der Oder, but it doesn’t change the big picture: a dedicated high-speed rail network would serve around half of Germany’s population pretty directly.

Very little of the remaining half would be genuinely bypassed the way Magdeburg and Brunswick were when Germany built the Berlin-Hanover line. Regensburg, for example, is and will remain peripheral under any rail plan, with regional connections to Ingolstadt and Nuremberg; high-speed rail serving those cities is the best way to connect it to destinations beyond Bavaria. Kiel, at the other end of the country, is and will always remain connected to Hamburg by regional rail. Münster, genuinely unserved, is not really bypassed, not with how close it is to Dortmund. And so on.

Such a plan cannot serve the entire country, but it can definitely then serve a majority of it. It mostly serves the largest metropolitan areas, but that’s fine – Germany is an urban country, around 40% of the country lives in metro areas of at least 1 million people (defined again mostly by S-Bahn reach, which is a conservative definition by American MSA or French aire urbaine or Japanese MMA standards) and much of the rest is either in metro areas somewhat below the cutoff or in exurbs served by regional trains but not the S-Bahn.

Public Transportation in Megacities

I’ve been talking so much lately about integrated timed transfer in the context of Boston that people started asking me if it’s also applicable to New York. The answer is that the basic principles are not scale-dependent, but the implementation is, so in very large cities, public transport planning should not look like in Switzerland, a country whose largest metro area is staring at 2 million people from the bottom.

The one caveat here is that most cities are not huge. The developed world has seven megacities: Tokyo, Seoul, New York, Los Angeles, Osaka, London, Paris. And Los Angeles doesn’t really have public transportation, so we’re down to six. The middle-income world has a bunch more for sanity checking – Mexico City, São Paulo, Rio de Janeiro, Buenos Aires, Johannesburg, Moscow, Istanbul, Tehran, Beijing, Shanghai, Guangzhou, Shenzhen, Bangkok – but all are either still in convergence mode building up their networks or (mostly in Latin America) have given up. So much of this comes down to the idiosyncrasies of six cities, of which the largest three networks are substantially in the same planning tradition.

Demand is huge

Big cities have big centers, which can’t really be served by any mode except rapid transit. Even in Los Angeles, what passes for a central business district has around a 50% public transport modal split. This means that the transport network has to deliver high throughput to a relatively small city center. Even in a low-kurtosis city like Paris, most Métro lines converge on a narrow area ranging from Les Halles to Saint-Lazare; in a high-kurtosis one like New York or Tokyo, there are a few square kilometers with 200,000 jobs per km^2, which require an exceptionally dense network of rapid transit lines.

Some other network design principles follow from the need to amply serve city center. Specifically, high frequency is rarely a worry, because there’s so much demand even off-peak that usually megacity subway systems do not venture into the frequency range where long waits deter traffic; New York’s 10-minute midday gaps are bad, but that’s unusual and it comes from a combination of the legacy of postwar fear of subway crime suppressing demand and excessive branching.

But other principles require careful planning still.

Electronics before concrete, megacity version

The driverless lines in Paris support peak throughput of 42 trains per hour – a train every 85 seconds. CBTC on Line 13 without driverless operation supports 38 tph, and London’s CBTC-equipped lines support 36 tph when the branching isn’t too complex. It is imperative for other cities to learn from this and do whatever they can to reach similar headways. The difference between 21 tph, as in Shanghai, and Paris’s 42, is equivalent to building a brand new subway line. And what’s more, in a city in the size class we’re talking about, the primary concern is capacity – coverage is already good, so there really is no reason to build two 21 tph lines instead of one 42 tph one.

The situation in Paris is in a context with self-contained lines. That said, extremely busy self-contained lines do exist in other megacities – London has a bunch with near-Parisian levels of throughput, New York has some, Tokyo has a few, Seoul and Osaka are both more self-contained than Tokyo is.

Throughput and organization

The primacy of throughput means that it’s worthwhile to build small infrastructure upgrades, even with concrete, if they help with capacity. Right now the Northern line reverse-branches with the branches to the north recombining with those in the center, and Transport for London would like to split the line in two, reducing branching complexity, which would increase capacity. But doing so requires improving pedestrian circulation in the corridors of the branch point, Camden Town, where TfL expects very large transfer volumes if there’s a split and already there are circulation problems today without a split. Hence the plan in the medium term is to upgrade Camden Town and then split.

If there are bumper tracks at the end of a line, as at 8th Avenue on the L or Flushing-Main Street on the 7, then it’s useful to dig up the street for another block just to add some tail tracks. That way, trains could enter the station at full speed. This increases throughput, because the terminal interlocking has trains heading in opposite directions crossing each other at-grade, which imposes schedule constraints; it’s best if trains can go through the interlocking as fast as possible to reduce the time they’re in a constrained environment, but that in turn requires short tail tracks so that an overrun of a few meters is not catastrophic. Ideally the tail tracks should even extend a full train length past the platform to place the interlocking on the other side of it, as is done in Paris and Moscow; in that case, trains cross the interlocking out of service, when it’s easier to control their exact timings.

Such projects are disruptive, but the disruption is very localized, to just one transfer station for a deinterlining project as in London or one terminal as in New York, and the impact on capacity is very large, if not quite as large as the full suite of signaling and track upgrades that make the difference between a train every 3 minutes and a train every 1.5 minutes.

Network design

The ideal metro network is radial. Megacities already support that just because so many lines have to serve city center. However, it’s important to make sure every pair of lines intersects, with a transfer. No large metro network in the world achieves this ideal – Mexico City’s network is the largest without missed connections, but it is not radial and its only three radial lines are overburdened while the other lines have light ridership. Paris has just a single missed connection on the Métro proper, not counting the RER, but it has many pairs of lines that do not intersect at all, such as M1 and M3. London is more or less a pure radial, but there are a handful of misses, including one without any transfer between the two lines anywhere, namely the Metropolian line (including Hammersmith and City) and the Charing Cross branch of the Northern line.

Big cities that plan out a metro network have to make sure they do better. Missed connections reduce passenger ridership and lead riders to overload the lines that do get connections; for example, in Tokyo one reason cited for the high ridership of the Tozai Line is that until Fukutoshin opened it was the only one with a transfer to every other subway line, and in Shanghai, Line 1 was extremely congested as long as the alternatives going north either had critical missed connections (like Line 8) or avoided city center (like Line 3).

The role of regional rail

Regional rail as a basic concept is mostly scale-invariant. However, the design principles for trains that come every half hour are not the same as those for trains that come every 5 minutes. If trains come every half hour, they had better connect cities in a roundtrip time equal to an integer number of half hours minus turnaround times, so that they don’t have to loiter 25 minutes at a terminal collecting dust and depreciating. If they come every 5 minutes, they’re not going to loiter 25 minutes anyway, and the difference between a 5-minute turnaround and a 7-minute turnaround is not really relevant.

The design principles are then mostly about throughput, again. The most important thing is to build independent trunk lines for trains to serve city center. Even in a huge city, the finances of building a purely greenfield subway deep into suburbia are poor; Tokyo has done it with the Tsukuba Express but it’s mostly above-ground, and for the most part regional lines there and elsewhere come from taking existing suburban lines and linking them with city center tunnels.

Tokyo’s insistence on making these city center tunnels also form a coherent metro network is important. Only one non-Tokyo example is worth mentioning to add to all of this: this is Berlin, which is not a megacity but has three independent S-Bahn trunk lines. Berlin, unlike London and Paris, painstakingly made sure the S-Bahn lines would have transfers with the U-Bahn; its network has only one U-Bahn/S-Bahn missed connection, which is better than the situation in Tokyo, Paris, or (with Thameslink and Crossrail) London.

The role of development

All first-world megacities, and I believe also all megacities elsewhere, have high housing demand by domestic standards. All are very wealthy by domestic standards except Los Angeles, and Los Angeles is still incredibly expensive, it just doesn’t have the high wages to compensate that London and New York and Paris have. In such an environment, there’s no need to try to be clever with steering development to transit-oriented sites. Anywhere development is legal, developers will build, and the public transport system has a role to play in opening more land for more intense development through fast trips to the center.

A laissez-faire approach to zoning is useful in such an environment. This contrasts with smaller cities’ reliance on finger plans, like the original one in Copenhagen or the growing one in and around Berlin. No limits on development anywhere are required. The state’s planning role remains strong through transportation planning, and the suburbs may well form natural finger plans if developers are permitted to replace single-family houses with apartment buildings anywhere, since the highest-value land is near train stations. But state planning of where housing goes is counterproductive – high transit ridership comes from the impossibility of serving a large central business district by cars, and the risk of politicization and policy capture by homeowners is too great.

The advantage of this approach is also that because in a high-demand city public transport can to some extent shape and not just serve development, it’s okay to build lines that are good from the perspective of network coherence, even if the areas they serve are a bit light. This principle does not extend indefinitely – subway and regional rail lines should still go where people are – but for example building key transfer points in near-center neighborhoods that are not in high demand is fine, because demand will follow, as is building lines whose main purpose is to close some gap in the network.

Construction costs

The larger the city, the more important cost control is. This may sound counterintuitive, since larger cities have more demand – only in Manhattan could a $1.7 billion/km extension like Second Avenue Subway pencil out – but larger cities also have a bigger risk of cost blowouts. Already Tokyo has stopped building new rapid transit in the core despite very high crowding levels on the existing network, and London builds next to nothing as well. New York’s poor cost control led Philip Plotch to entitle his book about Second Avenue Subway The Last Subway. Even Paris builds mostly in the suburbs. Extensive city center and near-center construction continues in Seoul, in the context of very low construction costs.

The flip side is that a New York (or even London) that can build subways at the cost of Paris, let alone Seoul, is one that can rapidly solve all of its transport problems. My Assume Nordic Costs map fixates on a region of the world with small cities, but the construction costs in South Korea are if anything lower than in the Nordic countries. And even that map, given free reins for developers, is underbuilt – some lines would look ridiculous at current costs and zoning but reasonable given low costs and liberal zoning, for example something meandering through currently industrial parts of New Jersey.

Small cities designed their public transportation philosophy around scarcity: Switzerland really can’t just draw crayon and build it, because housing and transport demand there are finite and limited. Cities like New York and London, in contrast, should think in terms of abundance of infrastructure and housing, provided their regulations are set up in a way that permits the state to build infrastructure at low costs and private homebuilders to redevelop large swaths as they become easily accessible to city center.