Linear Compression: How HSR is Like Rapid Transit

A post from last month on Keep Houston Houston notes how high-speed rail transformed Japanese geography to the point that it’s faster to get from Osaka to Nagano via Tokyo than direct despite a doubling of travel distance. The same comment could equally be made about rapid transit within a city: for example, for some origin-destination pairs in Vancouver, it’s faster to go the long way around the Millennium Line than to take a direct bus, and the same principle works in every other city. For both modes of transportation, this comes from high capital costs and high capacity, which make them useful primarily on the thickest travel markets, which tend to be radial around the largest center.

The next step is to look at the effect this change in transportation on economic geography. As I’ve argued before, in both cases the result reinforces preexisting centralization. This is both feedforward and feedback: a dominant city creates enough travel demand to support an HSR network and a dominant CBD creates enough demand to justify digging subways, while at the same time the quickness of travel along the rapid lines makes people emphasize connections along them and deemphasize others.

Concretely, this means that in Manhattan, with its wealth of north-south subway lines and paucity of east-west lines north of Midtown, people identify with the East Side or the West Side. Although the Upper East Side and Upper West Side are socially and demographically similar and are geographically close to each other, the social connections I’ve seen are primarily north-south. A gaming group I participate in many of whose members have recently moved to New York concentrates on the West Side since the earliest members moved to the Upper West Side, and so more people who were living or looking to live in Brooklyn or Queens are moving to Uptown Manhattan in general and the West Side specifically. The subway helps the Greater Upper West Side project influence as far north as Inwood. In contrast, the east-west connection is deemphasized to the point that people I know talk up the cultural differences between the Upper East Side and the Upper West Side, even ones who are not from either neighborhood and are not from the usual high-income demographic (though, of course, the two neighborhoods are culturally dominant and can discuss their own issues via mass media).

I do not know if the above trend is also the case for countries with developed HSR networks. However, another corollary trend is. The importance of the CBD and areas easily accessible from it is that the CBD becomes the more or less neutral choice for where people from different sectors can meet. Midtown can be easily accessed from the Greater Upper West Side, Greater Williamsburg, Greater Bed-Stuy, and so on. This effect then not only reinforces the rapid transit lines but also their nodes, to the point of creating possible centers around accidental transfer stations. In Vancouver, the Commercial Drive area functions as a major meeting location for social groups that are too widely distributed around the metro area for a place in Burnaby or along the Canada Line to be as acceptable. Although the Commercial Drive area hasn’t turned into a CBD and most likely never will, Midtown Manhattan became a CBD largely because of subway lines leading to Uptown Manhattan and Queens. Social meetings and job centers obey similar geographic rules.

In a fractal manner, in each sector there can also be a relatively neutral meeting location when the primary CBD is too expensive or too far, based on either a highway network (for example, White Plains for Westchester) or a rapid transit network (for example, Downtown Brooklyn for all of Brooklyn except Eastern Brooklyn), or even an arbitrary choice of zoning that then becomes self-reinforcing (for example, Metrotown in Burnaby). It promotes a perverse kind of equality, one in which no sector is favored over others, and the social hierarchy is based on the ease of getting to the center, in a similar manner to how in former British colonies with few whites, English sometimes arises as the politically neutral choice of language (or French in former French colonies, etc.), replacing a hierarchy between speakers of different local languages with a hierarchy between people with varying degrees of English fluency.

The exact same node effect can be observed in HSR. Japan’s become more centralized around Tokyo since the Shinkansen was built. In France and Britain there’s heavy centralization, going back many decades; from the start, the lines connecting the capital to the major secondary cities were treated as fast main lines while the others were slower branches. In South Korea, there’s mixed evidence about the role of the KTX in promoting development in secondary cities, but there has been growth in outer exurbs of Seoul that the KTX put within reasonable commute distance, such as Cheonan and Asan, even beyond the general growth of Seoul’s suburbs in the last 30 or so years. It is likely that of the secondary cities, the one emerging the best from this development is Daejeon, both the closest to Seoul and the junction of the lines to Busan/Daegu and Gwangju; for what it’s worth, even before the KTX opened, its metro area had faster population growth than the other major metro areas, excluding satellite metro areas that should really be thought of as suburbs of larger cities.

The meaning of this analogy is that an urban rapid transit network and a national HSR network will look similar. We can now extend the analogy and think in terms of connecting transportation. S-Bahn/RER-style regional rail generally involves routing preexisting commuter lines through new tunnels to provide rapid transit-style urban service; this is analogous to making HSR use legacy lines at lower speed in parts of the system that don’t justify the construction costs of a new line. Branch regional lines and buses feed people into rapid transit stations, in the same manner that legacy rail lines feed people into HSR stations. Some of the alignment questions, such as whether to tunnel or build complex viaducts to reach secondary city centers or to go around them on easier rights of way to save money, are similar, though the answers are often different (i.e. the benefits of the higher-cost alternative are much higher for rapid transit than for HSR since more people ride local transit than intercity transit, while the extra costs are comparable).

It can even explain some of the political coalitions. Rapid transit and HSR are both high-construction cost, high-capacity, long-term investments. They scale up but not down, and therefore cannot be undertaken by a cheeky entrepreneur with a moderate amount of venture capital; they are instead built by governments or very large conglomerates or sometimes both combined, and require careful planning (for example, upzoning) to ensure economic development patterns can reorient along the new infrastructure. They are also signature investments generating a lot of press, to the point that in some cases they can pursued purely for the ribbon cutting, while other forms of rail usually aren’t unless a politician is trying to oversell them as equivalent to rapid transit or HSR but cheaper.


  1. Joseph Eisenberg

    “Midtown Manhattan became a CBD largely because of subway lines leading to Uptown Manhattan and Queens. ”

    Is this correct? I thought that the commuter railroad stations at Grand Central and Penn Station for the reason that Midtown overtook Downtown for New York development. Certainly those railroads were built before the subways, if not before all of the street railways or elevated lines. Grand Central Depot opened in 1871 at the current site.

    • Alon Levy

      It has to be correct; how else do you explain the fact that peak employment density is so far from Penn Station but quite close to the Queens Boulevard Line? Of course Grand Central is older, but in 1910 the main CBD was still Lower Manhattan.

      • Adirondacker12800

        Things moved uptown because they were running out of space downtown.
        The busiest subway line is the Lexington Ave which doesn’t go to Queens.

        Things had been moving uptown for decades by the time Penn Station opened and Grand Central reopened. The epitome of Grand Central office building is the Chrysler Building, The epitome of Penn Station office building is the Empire State Building. Both opened before the IND did. Union Square was the happenin’ place in the 1890s. By the time the Flatiron building opens in 1902 it was Madison Square. The New York Times moved to Longacre Square in 1903. The Metropolitian Life Building, on Madison Square, was the world’s tallest until the Woolworth Tower displaced it.
        There’s a few reasons why Midtown East has more office space. The El getting to the Bronx and the Brooklyn Bridge opening gave it better access to what were then the suburbs. Keep in mind that the Flushing Line and the Astoria line served Second Ave until 1940. It was much easier to assemble land especially in the immediate vicinity of Grand Central because the New York Central owned it all and was willing to make deals. And probably one of the major ones. The movers and shakers in corporate America come from all over. They don’t have fantasies of living in Kings Point or Far Hills, they have fantasies of living in Greenwich. … the Man in the Gray Flannel Suit doesn’t live in Forest HIlls Gardens, he lives in Westport. Even if they don’t have fantasies of living in Greenwich the perception is that 1 East 48th Street is a more prestigious address than 1 West 48th. Park Ave versus Sixth Ave. too. Though they will probably insist on using Avenue of the Americas on the stationary.
        On the other hand Midtown West has lots of office space, lots of retail and lots entertainment. Tear down Radio City Music Hall, build a 50 story cube farm and there would be more employment. Less subway riders though…. Or Macy’s. … Or Madison Square Garden…

        • Alon Levy

          The Corona Line and the Astoria Line opened in the 1910s, though. Worth asking why Madison Square never became the center of the CBD. Or even Union Square, which has the Canarsie Line…

          • Adirondacker12800

            It’s cheaper to build skyscrapers north of 34th Street than it is to build skyscrapers south of 34th. Unless you are south of Chambers. So when the NY Times is considering a skyscraper their engineers prefer the site at Longarce Square to the one at Madison Square. Anyway the tenements in the west 40s are cheaper than the upscale stuff around Madison Square. And by that time the theaters are talking about a move uptown to Longacre Square. Met Life builds at Madison Square because they had foreclosed on the property. NY Life builds across the street to rub it in their face. The original Waldorf Hotel is built on the site of the family’s mansion and to annoy the neighboring relatives. The relatives take revenge by building their own hotel and call it the Astoria. It gets torn down for the Empire State. They move to Park Ave in the 20s because the New York Central has conveniently assembled some land for them. The schlub looking for new office space takes into consideration that the the big kahunas live in Connecticut. It’s more complex than the IND goes to Kew Gardens

    • Nathanael

      Joseph is correct; the intercity train stations are what dragged employment from Downtown to Midtown. The failure to get train stations into Downtown made Midtown naturally more attractive for businesses, which are far more focused on long-distance travel than residents. Especially in a town founded around import-export and developing around finance!

      So Manhattan actually shows a development based on intercity travel locations. Something similar happened in London, where development was dragged away from the City in the directions of the intercity train terminals.

        • Nathanael

          It does sound like the 1910s zoning laws had an effect. The truncation of the steam lines from the north to Grand Central Depot (with lines south being carried strictly by horse) circa 1870 was definitely part of the situation too, though. It’s more attractive to build next to the intercity train station, rather than a 30-minute horse-drawn carriage ride away.

          • Alon Levy

            Maybe… but then why is there not that much commercial development near Penn Station? The Empire State Building is nearby, but peak employment density is far north of Penn.

          • Adirondacker12800

            Herald Square can hardly be mistaken for a residential neighborhood.
            In 1880 if you didn’t want to take the streetcar up to 42nd Street from your Wall Street office you couldtake the El. To the burgeoning suburbs in Harlem if you were middle management. Or go to the St John’s terminal of the Hudson line. Or take the ferry to the LIRR terminal in Brooklyn or the multiple terminals in New Jersey. Whisked to your estate in Llwellyn Park in under an hour. Where the station was right across the street from the gates to the community.

  2. Ryan

    I can’t see the national HSR network doing anything other than further fragmenting the country’s metro regions, to be honest. Even the incompetently-designed and glaringly incomplete map of all the HSR corridors the FRA wants leaves huge holes in the network, and I somehow doubt that 100% of the corridors on that map will end up getting built, at least in the short-term.

    So, while the transformative effect of an HSR line on its associated region is undeniable, I somehow doubt that having (as an example) two HSR lines from Houston to Dallas via San Antonio and Austin and from Oklahoma City to Kansas City is going to do anything for anyone between Dallas and Oklahoma City except perhaps to make those commute/travel patterns look less attractive, enforcing fragmentation.

    This probably would have also happened had Florida’s HSR and the Southeast HSR from Washington DC to Charlotte (or even Atlanta) managed to both get built – rather than cultivating demand for an eventual connection between Atlanta and Orlando via Jacksonville, the demand for such a corridor that’s there today would likely be diminished as more and people moved or adjusted their travel patterns to reflect the better options inherent in either corridor rather than remain wedged between them, gaining the benefits of neither.

    Of course, it isn’t like there’s a good way to avoid this potential outcome.

    • ant6n

      I think you ‘zoomed out’ too much. You have to consider the development patterns within a large region, not the country as a whole.

    • Eric

      Better to have fragmented regions than the current situation – fragmented individual cities.

  3. Pingback: Coordinated Planning and High-Speed Rail | Pedestrian Observations

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.