What I Mean When I Say Cities Have no Transit

On social media and various forums, I have an expression for a variety of cities: “it has no public transportation.” This concerns just about the entire United States excluding a handful of cities like New York, San Francisco, and Chicago; Los Angeles notably is not among these handful, but has no public transportation, and neither do any cities in the South or the Midwest except Chicago. I want to talk a bit more about what I mean by this. I obviously don’t mean that literally there is no scheduled public transportation in these cities; I’ve taken these non-existent transit systems, in Los Angeles on a visit and in Providence when I lived there. But I mean that there’s something about such places distinguishing them from the bad-but-existing transit category of Boston, Chicago, Nice, etc.

Whatever you’re doing isn’t working

Let’s use an 8% cutoff for trips to work. This number is fully motivated reasoning: the metro area (MSA, not CSA) of Philadelphia is just above this cutoff, and I would not say it has no public transportation, at least not in the current state of the system. Bad, yes, but it exists. I may be missing some areas, but I don’t think I am: the list of American metro areas that meet that cutoff is New York, San Francisco, Boston, Washington, Chicago, Fairfield County, Seattle, Kitsap County, Philadelphia, Honolulu. 70% of American transit commuters live in one of these MSAs. Go down to 6% and you also get Portland and Baltimore, adding about 2.5% of US transit commuters.

Nor are things improving. Some parts of the US are seeing rising mode shares. The most notable is Seattle, which is serious about permitting urban housing, and has tunneling construction costs that would only get Europeans fired rather than simply not existing in democratic Continental Europe. But other cities that occasionally win accolades from American urbanists for investing in public transportation just aren’t cutting it. In the 2006-17 period, chosen because that’s what the ACS makes available, Denver went from 4.6% to 4.4%, Los Angeles from 6.1% to 4.8%, and Portland from 6.4% to 6.3%; in the praised-by-urbanists set, only Minneapolis went up, from 4% to 4.8%.

Let’s unpack what this means: whatever Los Angeles has been doing in the last 10+ years has gotten its mode share down – and that’s without counting the fact that the Inland Empire, officially a separate metro area, is growing much faster and has an even lower mode share, as people drive further and further from jobs to qualify for a mortgage. Portland and Denver have done a lot of supposedly good things with their light rail networks, but are standing still. Portland’s stagnation goes back at least to 1980, while Vancouver has built SkyTrain, a high-rise downtown, and Metrotown, and at 20% has a higher (and rising) mode share than any American metro area save New York.

Tabula rasa

When a metro area has 2-3% mode share, it’s best to treat it as tabula rasa. Yes, there are people who ride the buses and trains today, but so few that the advantages of from-scratch design are usually greater than the disadvantage coming from the risk to current ridership. The 2-3% figure really depends on the situation – I don’t want to give it as an ironclad figure.

Suburbs of very large cities (read: New York) approaching 10% may still be best treated the same way: commuter rail systems like the LIRR are really shuttles that extend auto-oriented suburbia into the city rather than the reverse. Sadly, where I say such suburbs have no transit as a positive statement, an MBTA general manager said “commuter rail is not public transit” as a normative statement.

The situations of extremely low-mode share metro areas and low-mode share suburbs are not exactly the same. For one, existing ridership is higher on Long Island than in Cleveland or St. Louis so there’s more risk if (for example) supernumerary workers go on strike to fight efficiency improvements, but the reward is much greater. We know how to squeeze high ridership out of regional rail in the suburbs, even low-density ones, since the city has so many jobs in the center. Moreover, we know which ready sources of ridership are suppressed by current operating patterns: working-class reverse-commuters, people who work non-traditional hours regardless of class, and peak-direction commuters getting off short of city center.

The tabula rasa concept notably does not mean the infrastructure doesn’t exist. Los Angeles has the physical infrastructure of a rail network. Long Island and Westchester have many rail lines pointing toward Manhattan. However, the operating patterns and development are deficient and little to no accommodation should be made for them. In the suburbs of New York and a handful of other American cities this concerns premium fares, low off-peak frequency, and lack of integration with local buses. In American metro areas with low overall ridership this concerns weak city centers, lack of TOD even when it could succeed (for example in Los Angeles and San Diego), local political systems that view transit as an excuse to get federal funds for other things such as road repaving, and, as in suburbia, low off-peak frequency. The problems vary, but the fact that there are severe problems remains.

The other element of tabula rasa is social. There is almost never any knowledge base in those areas about how good transit works, because people who’ve only lived there have by definition not regularly used even bad-but-existing public transport. Whatever local activists of all stripes have been doing in Los Angeles is not working. Understanding why from them can be valuable, for the same reason I talk to planners at poorly-run agencies like the MTA and the MBTA to understand what’s wrong, but all local practices should be considered suspect unless corroborated in an area with at least decent public transportation.

On giving offense

The people who complain about my use of “no transit” to refer to the vast majority of the United States are not making a semantic nitpick or asking for clarification. They specifically complain I give offense by erasing 2-3% of the population of Cleveland and St. Louis, or 1% of the population of Kansas City. (I name these cities and not 6% Portland because that’s what people have complained about to me.)

So let’s unpack what this means. I point out that in the vast majority of the United States, excepting a handful of regions all of which are politically stereotyped as Not Real America partly because they have public transit, has buses and trains that are so useless they might as well not exist. I point out that this remains the case despite extensive construction in many cities – Dallas has 150 km of light rail, which is respectable for a city of its size, Denver keeps expanding its network and has something resembling frequent regional rail, and so on. The problem is that I do not conveniently blame this on a political faction of others, be it Republicans, unions, moderates, drivers, or whoever. I genuinely think it’s the fault of everyone who’s had any amount of power, and this includes community organizations that keep identifying as always losing even when they repeatedly succeed in blocking changes they dislike.

This is American culture. Even the denigration of New York and other cities where there is public transportation is part of that culture; there are certain aspects of San Francisco, Boston, and Philadelphia that are useful for other parts of the US to emulate. But accepting that requires understanding that there is to a good approximation no contribution coming from no-transit cities (and this again includes Portland and Los Angeles, it’s not just Cleveland or Dallas).

Part of the problem is that the US defines itself so much around cars and car culture that the presence of public transportation is enough to make something feel not really American. The result is that any exhortation to learn from places with trains with decent ridership is bound to offend; I might as well tell Americans to move to Tokyo and learn Japanese and never come back to the West. But sadly for Americans, reality can be offensive. The culture of Real America has to change, at least when it comes to how to treat transportation and cities.


  1. Hugh B

    Alon, how many of the “no transit” cities do you think could go above 8% without any concrete i.e. Fleet expansion, better operating practices, transit lanes, and up zoning? Also, how high do you think Philadelphia could go without concrete (given its extensive infrastructure)?

    • Alon Levy

      Philadelphia is currently going down… I would prescribe better bus-subway-RR integration mainly, including mode-neutral fares, much higher RR frequency as Vukan Vuchic recommended already in the 1980s, and bus redesign that acknowledges the RR network and works with it.

      Portland and LA can probably both go into the low teens just through better land use, and maybe San Diego could too, I’m not sure. Portland is doing too little TOD, creating localized housing shortages. LA is a lot worse – it permits so little housing, and could generate a lot of ridership out of existing and under-construction lines by KWCIMBYing the parts of the Westside near the stations and commercializing near-downtown areas. Atlanta is nowhere near as NIMBY as LA, but it still has serious land use problems, like single-family housing 2 MARTA stops east and west of Five Points.

      • Jim Shilliday

        “I would prescribe better bus-subway-RR integration mainly”

        Have you ever done this analysis in detail? Hugh B rightly notes that Philadelphia has extensive infrastructure – two great decisions stand out: the early 20th century copying of the four-track New York model for our Broad Street and Market-Frankford lines, and later, building the four-track heavy rail connection between the Reading Terminal and (ex-PRR) Suburban Station (in the face of opponents’ claims that it would be cheaper to invest the money and provide free taxi shuttles in perpetuity!).

        On the other hand, we also have an enormous transit fail: An inexplicable design decision meant that the major commuter rail connection to New Jersey (PATCO) crosses the SEPTA system at 8th and Market Streets and then wanders off to terminate uselessly several blocks southwest of City Hall. New Jersey commuters must make an expensive and time-consuming transfer to get to locations directly west of City Hall (which is now Philly’s business hub and the only area east of the Schuylkill River zoned for high-rise office buildings), or to AMTRAK’s 30th St. Station (the gateway to the emerging high-rise area west of the river). I know people who have quit their jobs when faced with that commute. Sure, no one knew Philadelphia’s future when the PATCO system was designed, but they did know that they weren’t interchanging with the Pennsylvania Railroad or the subway. Unfortunately, we have no truly regional governmental body to operate our transit system. Nothing SEPTA or PATCO does, short of a merger, will solve that.

        But (as an interested layperson), I share your sense that much could be done in Philadelphia without major infrastructure changes, and I wish that SEPTA would hire you to plan here as you did in Brooklyn.

        • Jacob Manaker

          Unfortunately, we have no truly regional governmental body to operate our transit system. Nothing SEPTA or PATCO does, short of a merger, will solve that.

          Really? Fare integration between SEPTA, PATCO, the ACL, and NJ transit buses would go a long way without requiring a regional body to operate all four.

          If I had to summarize the problem with the NJ-side transit services, I would complain about the ACL, rather than PATCO’s downtown terminals. It’s the same story as regional rail throughout the US: it serves an express shuttle for auto-oriented suburbs to reach downtown, not local transit for urbanites. Rather than serve Merchantville and North Broad, the only stop between Cherry Hill and 30th St. gets no ridership that couldn’t get the same effects from PATCO. Conversely, there’s a missed opportunity to double as expresses Haddonfield-Lindenwold for PATCO commuters. The latter in particular addresses PATCO’s forced transfers via a reverse branch: commuters headed to east and south Center City would take PATCO, whilst commuters to west Center City & 30th Street would change to the ACL at Haddonfield.

          • adirondacker12800

            The Atlantic City line goes through North Philadelphia not North Broad. The only station between the bridge and 30th Street.

        • Sick Transit

          Jim: To answer a couple of your questions –

          First, only the Broad Street line is 4-tracked. The Market-Frankford El has only two tracks, a limitation that prevents any true form of express service. The closest it had was the old “skip stop” practice where trains alternately served or skipped various stations. However increased ridership and customer pressure has pretty much ended skip-stop service except at a limited number of stations.

          Second, the PATCO “rump” stub line in Philadelphia wasn’t a specific decision but rather the result of multiple failures of funding and municipal will during the early and mid-20th century. The stub was initially supposed to be part of a major center city loop that would have run down Locust Street and then turned north to Arch Street. (in fact there’s actually a partial tunnel sitting unused underneath Arch Street). The loop was intended to connect to the Broad Street Line via what’s now the similarly vestigial Broad-Ridge spur. However both money and political will ran out and the loop, as well as many other proposed lines, were never constructed. In the 1970s there were plans to extend the PATCO stub westward to connect with the operationally-compatible NHSL (ex-P&W) in Upper Darby, but those plans too died on paper.

          As an aside, it’s often asked why the Broad Street and Market-Frankford lines were never connected. That WAS a blindered decision, indirectly forced on the MFL’s owners by steam railroad companies. The history’s a bit convoluted: In the late 19th century traditional railroads became concerned that interurban trolley companies were siphoning off more and more freight business. They lobbied the state to require that new trolley lines use special broad-gauge tracks to prevent any form of interoperation with standard-gauge railroads. The four trolley lines that were built from the western suburbs to Upper Darby thus had to use the new gauge. When the Market Street Elevated was proposed, its owners hoped to connect directly with those trolley lines and also adopted the special broad gauge for compatibility. OTOH the Broad Street Line had no such limitations and was built to normal railroad specifications.

          Unfortunately corporate disagreements prevented the trolley / El connection from being implemented, leaving the Market Street Line as a broad-gauge “orphan”. The idea of converting it to standard gauge looks attractive on paper (e.g. it could allow through-running via the NHSL as well) but the cost would be astronomical: its broad-gauge tracks would have to be ripped up, its third-rail power system would have to be converted, and most importantly tunnel clearances would have to be rebuilt. So we live with it.

          • Nilo

            The MFL is the express service the trolleys are the local, at least for that short stretch East of the River. Why would you want to connect the broad street line to the MFL? They have a transfer, there’s no reason to engage in some ridiculous interlining practice to prevent that. Anyways the Philadelphia subway loop was a dumb idea, there’s a reason nobody builds distributor loops, they’re just not that useful per dollar spent.

            Why would you want a trolley-el connection?

      • Hugh B

        I’m not sure if this is actually the case, but I would think that Philadelphia regional rail suffers from a combination of generous padding and poor track condition, and I know that there are many low platforms in the system.

        • Nilo

          Yeah but raising platforms costs peanuts next to electrification and a regional rail tunnel, both of which it has.

          • Eric

            I suppose nowadays in the US, it probably costs $100 million per station to raise platforms…

          • jd

            Seems like it would make a lot of sense, if an agency could standardize and stick to a cheap no-frills design, to directly employ a small engineering/PM team and a dedicated construction crew to churn out platform after platform, year after year.

          • Alon Levy

            Yeah, they’re working on it now, it’s buried inside the South Coast Rail budget but they are contracting out (because The Private Sector Is Always Better) standardized designs.

        • danieltrubman@gmail.com

          Actually raising the platform only costs SEPTA $2-3 million. But they typically only do it as part of larger station projects that costs several times more (including ADA enhancements, landscaping, increasing/maintaining parking infrastructure).

          SEPTA also claims that some stations can’t have higher platforms (presumably because they don’t actually have full control over all the rail they operate on). I’ve asked for a list of stations where high platforms wouldn’t work, but haven’t gotten a response. Need to finish my Right-To-Know request form on the matter now that I think of it.

          • Nilo

            Where is their freight on the SEPTA network? Considering it owns all the track not owned by Amtrak the above claim seems straight up implausible.

          • Nathanael

            There is only one phsyical characteristic which actually prevents high platforms — platforms on a sharp curve. This is the excuse made for rebuilding Amtrak’s Schenectady, NY station with low platforms. (I don’t even think that excuse is valid, personally, but it’s currently considered valid because building high platforms means much larger gaps between the train and the platform. Use gap fillers, guys.)

            At Exton, they appear to be planning mini-highs as a result. (*Part* of the track is straight…) At other stations, they’ve relocated the station away from the curve.

            With the disentanglement of the West Trenton line complete, freight is not an issue for high platforms anywhere on SEPTA.

      • Mike

        San Diego is strange because there are three or four dense job districts: downtown, Sorrento Valley, UTC/Golden Triangle and Kearny Mesa. Downtown is the nexus for transit, but the other three locations are pretty terrible to get to via transit.

        Sorrento Valley has a busy commuter rail stop but it’s only really good for weekday commutes from North County; I think only one peak rush hour bus goes into that area and it’s from the direct east. UTC has three or four bus lines go to the mall, but that’s it until the Mid-Coast trolley (light rail) line comes online in 2021-22. Kearny Mesa has basically no transit to speak of. TOD would help but we just need more frequent/reliable transit.

        • Eric

          Looks to me like it would be easy to add a branch of the Mid-Coast line from UCSD to Sorrento Valley…

          • Mike

            It looks easy but I don’t believe it is. UTC to Sorrento Valley has an 8% grade over a mile on the flat part next to Interstate 5, plus the train station at Sorrento Valley is in a canyon with no really good way of getting out of there or expanding it besides the current line. I think it’s protected wetland north of Sorrento Valley along the rail line as well.

            SANDAG is looking into another trolley line that will connect Sorrento Valley but I think it’ll arrive from the east along Rose Canyon. It’s in blank-paper-pen-and-a-gallon-of-coffee stage.

        • Nathanael

          The San Diego system is only just getting to UCSD (Golden Triangle, UTC) just now with the “Mid-Coast Trolley”.

          Getting to UCSD is pretty critical as it is a bigger concentrated job *and* residential center than almost anything in the county — 35,000 students and 28,000 employees. Only the Navy has higher employment, and *nowhere* has more residential density than the campus proper. (The students will definitely be taking the Trolley a lot.)

          Frankly, I think the line from UCSD to downtown is a “missing trunk route” in San Diego. The transit-inaccessibility of UCSD was the reason my parents needed a car every time my father went to San Diego (he’s a professor). Locating UCSD in the middle of nowhere was a bad move in the first place, but now we’re stuck with it, and it’s more of an urban center than the actual downtown of San Diego is.

          The Kearny Mesa area is extraordinarily car-oriented, being full of auto dealerships, and is also not on the way from anywhere to anywhere else. The job development — even the part which isn’t auto dealerships — is exceptionally auto oriented. I honestly don’t know how you’d serve it.

          The Sorrento Valley area (Qualcomm) could be served by one line straight down Mira Mesa Blvd, though connecting it to UCSD (as is necessary) would be difficult.

          San Diego would definitely benefit from some more high-density-residential land use near its stations. They mostly run through commerical districts, which is great, but leaves people in postage-stamp single-family residences dependent on cars. The first several lines were revivals of old railroad routes, which weren’t *bad* but weren’t *optimal*.

          A lot of the existing slightly-more-dense areas, areas dense enough to be urban (rather than suburban) are along University Avenue from Hillcrest to La Mesa — which doesn’t have rail, and the buses get caught in traffic.

          (Citation: https://www.voiceofsandiego.org/topics/land-use/san-diego-the-eighth-largest-city-but-still-mostly-a-suburb/ )

          Other higher-density areas are also unserved: Otay Mesa and Bay Terraces. Pacific Beach and Clairemont are just barely going to be served, tangentially, by the Mid-Coast Trolley line which *hasn’t opened yet*. So prior to the Mid-Coast Trolley, nearly all the denser residential areas were completely bypassed by the rail system. Combine that with no rail to the single largest employer and single densest set of residences (UCSD), and what chance did it have?

          • Nathanael

            …. and I suppose this is a way of saying that before the opening of the Mid-Coast Trolley, if you attended or worked at UCSD, it felt like San Diego had “no public transit”. To use your words.

      • Nathanael

        I do wonder what best practices would be in Philly. I suspect that allowing denser zoning would be important and valuable. The disconnection of PATCO, as mentioned by Jim, is glaring, as is the historic-museum-style ticketing on SEPTA Regional Rail. (Not to mention that they only just stopped using *tokens* on the subway and trolleys.)

        SEPTA has one real oddity of history: they have really extensive, arguably excessive, supply of transit along *one corridor* through downtown (the Trolley/Market-Frankford/Regional route down Market Street), and it’s much more difficult to get to a lot of other areas. This transit access is NOT matched by the zoning. The extreme transit access along Market Street really should lead to *even denser* development there than anywhere without a four-track subway, but as far as I can tell it hasn’t been allowed to; there are still surface parking lots within a block of Market Street. Oy.

        I have to agree that the problems in Portland and LA are almost entirely due to zoning, and probably some of the problems in San Diego too; I can analyze San Diego further, though.

    • Robert Jackel

      SEPTA is frustrating because it seems they’re spending lots of money on things that make transit worse. SEPTA key is very difficult for casual users, and is tough for people (like me) who use SEPTA to take their kids to school.

      The SEPTA regional rail implementation seems to be an expensive way of making plausible deniability to keep homeless people out of Suburban Station.

      • Sick Transit

        SEPTA’s much-ballyhooed Key Card is a textbook case of how NOT to implement a fare system. Without going into the gory details, they opted not to start with a clean slate and look at what worked (or didn’t) in other cities. Instead, they decided to create an electronic version of the legacy fare system that was a patchwork inherited from SEPTA’s multiple predecessor agencies. The Key preserves many of those predecessors’ worst practices, such as high transfer costs and separate fare levels for transit and rail lines in the CBD. Unfortunately despite rider complaints, major cost overruns, and professional evaluations detailing the Key’s numerous deficiencies, SEPTA’s management continues to plow ahead.

        • Robert Jackel

          It’s truly awful, which becomes apparent when you want to use Regional Rail. Once the keycards are activated, there will be obvious throughput problems.

          Imagine trying to catch your train at Narberth and there’s a line to use the janky tag-in kiosk.

        • Nathanael

          Yeah, they needed to establish a unified zone-based fare system like London before trying to change fare media. But they didn’t.

    • Henry Miller

      8-10% – exactly 0. There are too many cars in the way to run buses often enough. Even if you paint bus lanes, cars won’t respect them, the police won’t enforce them. For that matter politicians will remove the paint on any useful lane because cars are to too important to everybody and the resulting traffic mess from taking away busy lanes will get political attention fast – anyone who tries to stand by the bus lanes will be voted out of office. Same applies to wide spread signal priority – more drivers will complain than bus riders will enjoy it even if we (wrongly!) assume all bus riders vote for the politician who stands by, the drivers will win the next election and that will all be gone.

      > 50% – all of them. There is more than enough concrete on the ground now, and more than enough density in most suburbs to support “a bus every 10 minutes all day every day”. With that many riders the bus (and sometimes train) is arranged so that if you can’t get there in a short rider you get to an express bus (real express – for cross city trips the bus get up to 90 mph on the now-empty freeways thus beating cars at the speed game for longer trips). With that many trips there will be useful routes (with easy transfers) to anywhere you want to go in a reasonable amount of time for less the cost of owning a car and so many people won’t even have a car. Thus political forces ensure that transit gets the priority needed to work.

      You can’t get to 50% with going through 8% though: we can’t get there from here without concrete. The concrete isn’t needed once we are there, but we are not there.

  2. threestationsquare

    It does seem like some metro areas in the 5% modeshare range still sometimes have enclaves with decent transit usage? And pushing the boundaries of such an enclave outward can sometimes be a viable strategy. Notably it seems like Seattle did something like this, and Denver’s approach was very much not this. (LA does seem to have attempted something like this and I’m a bit confused why they’re failing so badly.)

    • michaelrjames

      LA does seem to have attempted something like this and I’m a bit confused why they’re failing so badly.

      Agreed. I imagine most transit and city planners expect LA’s investment in public transit to begin paying off … eventually. Is the statistic just a reflection of the greater growth of its distant exurban periphery relative to its more central core? That is, is transit use dropping in a absolute numbers or just not growing as fast as these others?

      • Henry Fung

        Transit use is clearly dropping in absolute numbers. It turns out, when immigrants don’t feel they will lose their livelihoods by driving (because they can’t legally get a license due to their unauthorized status), they start driving. Also, with so many cars available, and California’s tendency to buy Japanese imports, there are plenty of good quality 10, 15, and 20 year old cars out there for a song.

        • Herbert

          Rail usage has risen while bus usage has shrunk. Los Angeles has cut bus service quite a bit…

      • Jacob Manaker

        LA might be investing in transit, but it isn’t also committing to constricting car use in the necessary fashion. Just a week-and-a-half ago, Metro’s blog wrote a brief summary of all the topics that would appear on an upcoming board meeting. That the 405 ExpressLanes project doesn’t require widening was so unusual, they made a point of explicitly mentioning it.

        Freeway widening. In this day and age!

    • danieltrubman@gmail.com

      MSAs might have some uses, but these types of considerations, I think “Urban Area” might be better. UAs still includes areas that I believe go beyond suburban to exurban, but they aren’t as likely to have include truly empty space.

      Matters more for considering density and land-use, because the relative emptiness of the fringes mean there aren’t many people to pull down mode shares, but it doesn’t take too many people living in an areas with a mode share at or approaching literally 0.0% to start making the numbers look funky.

      Of course UAs don’t always have a neat international comparison, but it’s not like every country does MSAs exactly the same either.

      • Alon Levy

        Yeah, but when we talk about commutes, commute-defined MSAs do work better. (And American, French, and Canadian metro area definitions are pretty similar. Japanese ones are a bit different but and are more analogous to American CSAs in practice.)

        • adirondacker12800

          If people weren’t commuting between the MSAs they wouldn’t be in the CSA.

        • danieltrubman@gmail.com

          I just don’t see how including statistics from Cecil County Maryland tells you anything about the quality, usefulness, or ridership of transit in Philadelphia?

          • danieltrubman@gmail.com

            I’m going to guess you’ve never been to Cecil County Maryland? Though maybe you’ve driven through it, as I-95 goes through it. Anyway, the county is almost entirely rural, plus a state forest. I mean it’s part of the Eastern Shore. About 10k commute in to New Castle County (according to 2011-2015 5-Year ACS Commuting Flows) but only 391 commute to Philadelphia. That pales in comparison to the 1,177 who commute to Baltimore, but makes sense when you consider MARC’s Penn Line terminates in Cecil County. Identity isn’t everything, but I honestly doubt 1% of county residents would identify as being part of the Philadelphia region.

          • danieltrubman@gmail.com

            Hey if you want to consider Cecil County when evaluating transit usage in metro Wilmington, that makes sense. But Wilmington itself is at the fringe of the greater Philadelphia labor market. Less than 9k commute from New Castle to Philadelphia, which isn’t that impressive considering County has population over half a million.

          • adirondacker12800

            Wlimington isn’t it’s own MSA, it’s part of Philadelphia’s. It’s not just people who commute from Delaware to Philadelphia it’s people who commute to Delaware County, Pennsylvania too. And any other counties in the MSA or CSA. And people from the rest of the MSA or CSA who commute to New Castle county.

          • john

            As a native of Ceciltucky (now in Baltimore), the main economic pull comes from New Castle Co/Wilmington, DE (Philly MSA), with Harford County, MD on the other side (Baltimore MSA) a close-ish second, thanks to Aberdeen Proving Ground (big fat military base) and the job-sprawl-blob that is Bel Air.
            Culturally, most people just think of themselves as Cecil Co/rural MD/MD’s Eastern Shore, tho’ I remember an old quote from a source long forgotten that “The Eastern Shore don’t want us, the Western Shore won’t have us, we might as well be part of Delaware.” And there are definitely lots of fans of both Philly & Baltimore sports teams, but more Philly. And lots of people shopping in sales-tax-free Delaware (border arbitrage!), blissfully unaware that DE has a gross receipts tax half the rate of MD’s sales tax but applied to a much broader base (more products), hidden inside the sticker-price and giving them rather less savings than they think.
            Demographically, the land may still be mostly rural, but most of the people are divided between: A) A handfull of old railroad towns along the Amtrak/I-95 corridor that have spilled over into mini-sprawl blobs well past their municipal boundaries; and B) exurban sprawl spattered across the landscape of the county’s northern half (mostly north of Amtrak).
            Which circles us back to the matter at hand: Little things add up, and if a sprawling metropolis’ people are bleeding off at the edge into a bunch of exurbs, and enough jobs follow them, that drains resources (people and jobs) away from the center that the center could have relied on (politically & financially, as well as a daily user base) to support things like more and better transit, even if each exurb is, by itself, proportionately small. Cecil Co isn’t alone in tugging metro Philadelphia outward, it is joined by Chester County, PA; Salem Co, NJ.; the outer communities of Bucks & New Castle & Burlington & Gloucester, etc, etc,. And all that distant sprawl tugs the region’s culture and politics in an anti-urban direction, too.

    • Ross Bleakney

      It does seem like some metro areas in the 5% modeshare range still sometimes have enclaves with decent transit usage? And pushing the boundaries of such an enclave outward can sometimes be a viable strategy. Notably it seems like Seattle did something like this

      Actually it was the opposite. They simply improved the enclave (Seattle proper). Almost all of the transit increase from 2010 occurred in the city. There are several reasons for this. First, they finally built the section of subway they should have started with (downtown to the University of Washington). This allowed a restructure, which increased frequency on the bus (still by far the largest share of transit). Seattle itself passed a levy to increase bus service, even though the county runs the buses. Traffic has been getting worse, while they city chips away at the problem (building bus lanes, etc.). Finally, the city proper has grown faster than the suburbs, and much of that growth has occurred in urban areas (where people are likely to use transit).

      The only other area that has seen an increase in transit (and a much smaller one) is Bellevue. It is a suburban city, but with a fairly large downtown, that has seen improved bus service to it. It has also seen some growth (a lot more than a lot of the suburbs) with that growth being centered around a few, relatively convenient areas.

      The train does go out to the suburbs, but ridership there is relatively small.

    • Nathanael

      Denver didn’t really start with a “transit enclave” and expand it, sadly.

      Denver did the rather traditional thing of building passenger rail on all the old steam railroad lines. This has its good sides and its bad sides. Due to the way land use developed, it usually works better to build the passenger rail on the old *interurban electric* lines, but they didn’t do that.

      The big problem is that Denver has not really invested in transit *in* the downtown, dense enclave. People have been campaigning for a Colfax streetcar or light rail (with its own exclusive lanes) for a long time, and it’s the obvious corridor for it, and it used to have it, but nope, they haven’t built it, and they haven’t even made bus lanes. Instead Denver’s built suburban rail dependent on park-and-rides. The extreme example is “Lone Tree City Center”, which is an empty prairie where some developer might eventually build a suburb, mayble.

  3. Chad Newton

    Your Tabula Rosa point about “no transit” cities is why I think fare-free transit in the likes of Kansas City is a good idea. To most everyone there, a standard transit fare is a poor value compared to toll-free highways and nearly always free parking. Free transit levels the playing field, and approximates the perceived value. Fares were providing such as low proportion of operating costs that a minor and affordable funding increase can cover for the lost fares.

    With a new free transit system, the city has a clean slate, where curious people who have never ridden transit can try it out without having to worry about how to pay. Ridership can rapidly scale if people find a use for it. And once the budget accommodates free transit, all future funding increases can go right into increased service. Freed of the friction of fares, additional subsidies could become more popular if useful services are created.

    • Alon Levy

      Yeah, I know the theory. The practice is the opposite – budget-cutters have a much easier time arguing for frequency cuts whenever the economy has >4% unemployment. You do not need to be profitable for cuts in off-peak frequency to be financially wrong on the margins, especially if you’re running trains rather than buses; New York barely saves any money from subway frequency cuts *at the peak*, and off-peak it almost certainly would gain from matching or near-matching peak frequency.

      For what it’s worth, where I am right now, Taipei has a profitable metro system, and I vaguely think that so do the other Taiwanese cities. This coexists with great political will to expand public transportation, which is useful since Taiwan is building subways pretty late in its development, so it needs to revitalize public transit rather than transition straight from jitneys to trains the way Singapore and Hong Kong did. Its recent infrastructure package spends, rescaled to US population and PPP, $50 billion a year on rail expansion; the US spends that in about 4 years’ worth of federal transportation bill – and the rest of the package isn’t roads.

      • Tonami Playman

        I thought Kaohsiung metro is constantly running at a loss and far from hitting it’s breakeven ridership. It’s also unique for an Asian city to have such a low ridership especially in per kilometer terms. 184k average daily riders for a system length of 47km.

        • michaelrjames

          Is it that it is only a decade old and, like LA, has trouble displacing old habits. Although it is a city approaching 3m perhaps it hasn’t reached a sufficient pain threshold to get people off the roads. Fares seem to be cheap enough. Do they have good deals on Monthly cards?

          • Henry Fung

            Everyone’s on a four cycle engine motor scooter, which works better for Kaoshiung than Taipei because of weather (further south, more humid) and parking availability.

      • Tonami Playman

        Alon, regarding Taipei, how does the RFID token tickets compared to card based systems in Singapore, Hong Kong or other cities. Their metro is constantly ranked highly for efficiency, but most metro are moving to card based systems.

        I was always of the opinion that card based fare systems are more efficient, but looks like Taipei must be doing something right.

        • jonahbliss

          The vast majority of people use a card on the Taipei Metro. The tokens are essentially single use rfid cards. That setup is not too uncommon in Asian metro systems.

        • Alon Levy

          Taipei is a card-based system, and getting a card is easy – there’s a prominent booth to buy one at the airport, which only takes cash, but is located about 20 meters from a moneychanger with better rates than my banks give me on international transactions.

    • RVA_Exile

      Isn’t the point that “all future funding increases” = LOL in an improving economy, and the free service will be the first thing to be cut at the next downturn?

  4. jonahbliss

    It seems like your message would be a lot clearer if you said “no transit ridership” instead of “no transit.” It’s not like the actual service provided in Honolulu is good (as an example), but the geography and factors relating to expenses and incomes there create more ridership than would be found otherwise…

      • Alon Levy

        I actually don’t know what’s Kitsap County’s deal; the others’ deals I do know (a bunch of old big cities, a suburban county of New York, and a dense place that’s so mountainous that people and jobs cluster along a coastal ribbon).

        • jd

          I figured it was the worker driver buses, but it looks like that is only 8% of ridership. I imagine the island geography and ferries play a part.

        • SB

          In guessing but Kitsap Country is across Seattle and fastest link is though ferry because there is no direct bridges/tunnel.

        • Brendan Dawe

          It’s probably going to be similar to some of the insular communities near Vancouver, such as Bowen Island (21%) & Gibsons (9%) or Gabriola Island (14%) where car ferry users presumably get counted as transit commuters.

          Bowen Island isn’t getting big-city modeshare from the one or two short-buses that Translink operates over there


          • Rico

            I could easily be wrong but if I lived on Bowen and commuted to Vancouver I would find a way to get to the ferry terminal (car, bike, bus, walk) then take a bus from Horseshoe Bay.

          • Brendan Dawe

            People do, but supposing you live in Snug Cove, by the Bowen terminal, it’s still a 75 minute commute to downtown. There is a cottage industry of Horseshoe Bay landowners renting parking to Bowen & Sunshine Coast commuters’ mainland cars

    • Alon Levy

      I say no transit because urban geography adapts to how most people travel. The reason there isn’t any transit ridership in (say) Dallas isn’t that the light rail is good but the roads are so much better. Rather, the region’s distribution of jobs has oriented itself around 100% auto usage, so city center is weak (job density comparable to Lyon and Vancouver, roughly), and non-central jobs don’t have any structure that transit could plausibly serve (e.g. linear along a railway).

      In contrast, Honolulu has compact job distribution – on the metric I use, jobs in the central 100 km^2, it’s not far behind Lyon or Vancouver, in a metro area one third the size, and moreover the jobs are along a linear corridor. It’s so perfect that we should absolutely chew the state and the city for not having higher transit usage, e.g. because the HART project is extremely overpriced and delayed. So the service provided in Honolulu is clearly terrible, as in Boston and Chicago, but it’s firmly in the bad-but-it-exists category.

      • jonahbliss

        I don’t think there’s anywhere in the US where transit “works” that doesn’t have a severe geographic constraint essentially forcing it to work. So in most cases that’s some sort of body of water that both constrains where the city developed density, and also essentially limits the road network access to the city core. DC is probably the place that least fits that mold (although the Potomac is pretty wide and unbridged in many areas) but that’s obviously helped by the strong downtown job presence created by the federal government, and then that same federal system helping to create a WMATA that had a stronger downtown presence than the other two transit systems created in the same era…

        • Lee Ratner

          The places in the United States with the best transit are the ones where the downtown area of the core city remained the economic focus of the metropolitan region. Jobs in New York City, Chicago, and to a lesser extent San Francisco, Portland, and Seattle suffered less from suburbinzation than other downtowns. You also had a lot of entertainment and shopping remaining in the downtown areas. Since driving into Manhttan to go to the MET or see a Broadway show or the Knicks play at Madison Square Gardens is a pain, you have people willing to take transit into the city from the suburbs for fun.

      • Lee Ratner

        Dallas built their light rail system after the jobs disbursed because of decades of car oriented development. They should have possibly built a light rail system that was less focused on getting people downtown but more focused on the employment centers.

      • Nathanael

        HART was always going to be overpriced and delayed because it’s the first major construction the state has done in, oh, my lifetime (so — no expertise), and they have to import everything (so — elevated costs). The NIMBYs threw everything they could think of at it to delay it and make it more expensive. It will, of course, be a smash hit when it opens. Hopefully later extensions will have a better political dynamic.

          • Nilo

            Thankfully there’s exactly one short extension east to the line that makes any sense, so even as they hose up this project cost wise its not like there’s much more rail Honolulu could possibly need for at least the next like 50 years.

      • Nathanael

        San Diego is a curiosity because it actually has some highly compact job distribution *but the trunk transit routes don’t go to the job centers*. Nor do they go to most of the concentrated residential areas! The UCSD extension will start to change that.

        San Diego has the system it has out of historical accidents and political “we can do this on the cheap” decisions, but if we were starting with tabula rasa, we’d start by connecting UCSD to downtown. So at least that is getting built finally.

  5. SB

    LIRR has weekday average ridership of 360,000 and ridership per mile of 1,120 despite the fact it is expensive and peak orientated.
    This doesn’t feel like tabula rasa to me
    How much will improving just organization will increase a LIRR ridership?

    Also using urban areas rather than MSA seems better comparison to me.

    • Alon Levy

      There’s an RER line with more ridership than every New York-area commuter train combined, and another with about as much as the sum total of any two out of the three.

      And LIRR ridership could, if the trains were useful for non-Don Drapers, include commuters to Jamaica, Downtown Brooklyn, Flushing, and LIC; reverse-commuters who are currently stuck in traffic on NICE buses; people working in Mineola (with about 25,000 jobs and 250 morning peak alightings); people going to Mineola for government business like jury duty; people in Eastern Queens who are currently stuck in traffic on buses to the E and F trains; and people in inner suburbs like Great Neck and Hempstead making non-commuter trips. And that’s even without TOD.

      • adirondacker12800

        I’ll bite. I live in Massapequa and I have jury duty in Mineola, how does the LIRR get me there? How does reverse commute on the Mainline get me from Flushing to a job in Hicksville? The high frequency NICE buses don’t run near the LIRR which is why they are high frequency. You are also assuming people are schlepping all the way into Manhattan on the E or the F and not to a good job somewhere in Queens that doesn’t have an LIRR station.

        • Alon Levy

          It doesn’t. But a commenter on (I think) SAS talked about how they got out of jury duty in Mineola when living farther west, I forget where (maybe Queens, maybe far west LI?), because they had to report in the morning and with no reverse-peak LIRR service they’d have had a 2.5-hour trip, which is legally long enough to be an excuse not to show up.

        • Alon Levy

          Plenty of middle-class riders, yeah, but not many working-class ones. Nassau County’s transit commuters have 50% higher median salaries than the county average, Suffolk County’s have 65% higher median salaries.

          • adirondacker12800

            Because it doesn’t make sense to commute long distances for a low wage job when there are low wage jobs closer to home.

          • adirondacker12800

            What about Mineola? If I live in Massapequa and have jury duty in Mineola I could change trains in Jamaica to get there. But since I live in Massapequa it’s very likely I own an automobile and even with the execrable traffic in Nassau County, driving to Mineola would be faster. Without looking at a map I remember three ways to do that. And which ones I couldn’t use if the vehicle has commercial plates. I’m a teller for Chase Bank I’m not going to schlep to Mineola when there is a branch in Massapequa. Or a cashier’s job at Walgreens. Walgreens closes I’m not going to schlep to Rego Park, I’m going to apply for a cashier’s job at CVS in Massapequa.

        • Alon Levy

          Wait, the LIRR link is telling me that 80 out of 90 million riders are traveling to or from a western terminal. The Metro-North link tells me that east of the Hudson, 68% of weekday riders travel at the peak in the peak direction (Transilien: 46%, even excluding suburban job centers like La Defense and the airport).

          • Lee Ratner

            Transfer between different lines is vaguely plausible on the LIRR but implausible on Metro North because all the lines fan out from Grand Central in one direction. There aren’t transfer stations with the system. This isn’t even a post-car phenomenon. The systems were built to go to and from Manhattan before car use was a thing.

  6. Damien Smith (@indysmith)

    What are your specific concerns with Chicago–you’ve mentioned it’s bad transit across a number of posts. Aaron Renn (I think) once said on a podcast that the L system is designed to get people from outer suburbs to the Loop for work but is pretty bad for moving people within the city. Some key high-traffic destinations, such as Union Station and Soldier, Field are not served by the L. You can easily not notice that Metra exists, and there’s no integration with the L. I’ve visited a few times (two of my sisters used to live there) but your impressions would be helpful. I also don’t know now it could be easily fixed; your post on regional rail to O’Hare pointed out that it could not be done easily.

    • Alon Levy

      That’s basically it, really. Chicago has a wonderful transit network if you work in the Loop, and unlike commuter lines like the LIRR, the L doesn’t require you to be working 9-to-5. But connections between different lines are inconvenient, and outside the Loop job density drops like a rock.

      • Eric

        I was recently visiting an inner suburb of Chicago. Bus lines had a reasonable geographic density and ran every 15 minutes (though they were slow with frequent stops). I could get to the Loop with a 30 minute bus ride and 30 minute L ride. Certainly most people chose to drive, but I wouldn’t say this was a case of “no transit”.

      • Nathanael

        The ludicrous way lines just fail to connect to each other (by one or two blocks) is almost unique to Chicago; I’ve never seen it anywhere else. Very hard to fix though.

        • Fernando

          I think the main missed connection is Illinois Medical District (blue) and Polk (pink). Most of the other missed connections are excusable since the lines connect a stop away: the orange line breezes past Chinatown without a stop but connects with the red (and green) lines at Roosevelt a stop away, the blue line stop at Grand is right next to the green/pink line stop at Clinton but all three lines connect at Clark/Lake a stop away, the red line and brown line overlap but don’t connect near the North/Clybourn red line stop but connect a stop away at Fullerton. Even the IMD/Polk missed connection is due to the fact that the pink line used to be a branch of the blue line that diverged a stop away at Racine.

          • adirondacker12800

            The IRT and the BRT/BMT used to be much more extensive.

    • Nilo

      Chicago problems in rough order of fixability

      1. Buses are slow and not frequent enough (all door boarding and more bus lanes solve this.)

      2. L capacity on the Blue and Red line is a major problem right now. (They’re doing electrical work on the blue and concrete work on the red to try to alleviate this.)

      3. Alon’s job density thing, I’m not sure exactly what definition of the Loop Alon is using, but there has been quite a bit of office development in what’s known as the “West Loop.” There’s also a rather ambitious proposal to extend the Loop rather dramatically southward onto an empty lot. Thats apparently where Amazon was considering, but of course you need a tenant to promise to come to get construction there.

      4. Metra is run by morons. I’ve pointed this out on other mediums, but at least in ’10 straight up more Loop commuters arrived via Metra than the L. Metra has a litany of problems, but two of the most obvious, terrible rolling stock and poor frequency would do a lot to help the situation. Metra of course owns a rapid transit line, the electric line, but runs it like an infrequent commuter line on much of its length. There are hard Metra problems (how to connect the lines across the Loop being the obvious one), but so much could be done by modernizing the four lines Metra owns with electrification, running them frequently, charging reasonable in city fares, and entering into talks with BNSF about four-tracking that line to provide more service on it. Metra on the southwest, south, and northwest sides runs through a bunch of neighborhoods without the L. These neighborhoods also provide the vast majority of drive alone into the city center commuters, since suburbanites have a commute that’s too brutal to do by car so they take the train.

        • Nathanael

          The Near North Side, despite being booming, has poor service east of the Red Line, which is actually another point which isn’t usually noted.

          The biggest problem is really Metra’s unbelievably backwards attitudes. But it’s worse than that! Apparently Metra is now on-board with the rapid transit renovation of the Electric line, but the Chicago mayor is opposing it for idiotic turf war reasons. Yeeeaaaarrrrrgh….

        • Nilo

          Fair though Metra does a pretty poor job of servicing the edges of those regions. Through running and better L/Metra transfers are the obvious (is expensive) fix for this.

      • Nathanael

        Oh, P.S., the BNSF line is already three-tracked, which is enough for all-day two-way service even with freight and Amtrak in the mix. The stations within the City of Chicago, however, need to be totally rebuilt (they’re not ADA-accessible and are quite miserable, and they don’t have platforms on the right tracks.) Metra attitude problems (“we’re here for the suburbs, not for Chicago”) have prevented this.

      • Gok (@Gok)

        You can really generalize point #4 into the fact that the CTA/Metra/Pace/NICTD split makes everything worse. There’s a sick incentive to prevent transit users from using more than one system. The city actively sabotages any effort to use Metra within city limits. It’s telling that the only L expansion actively being considered is not a circle line, but an extension of the Red Line south, almost entirely paralleling the Metra line that could already run like rapid transit.

        There have been repeated analyses showing that the current structure of RTA is terrible and should be replaced, but of course there is no political will to fire a bunch of incompetent political appointees.

  7. Untangled

    At the other end, at what mode share to work would you put the cut-off between “bad-but-existing” transit and decent transit?

  8. adirondacker12800

    commuter rail systems like the LIRR are really shuttles that extend auto-oriented suburbia into the city rather than the reverse.

    Because people in the city don’t have a burning desire to go to a Walgreens in the suburbs when there is a CVS a block away. Or McDonalds when there is a Burger King. There aren’t many destinations out there.

  9. Eric

    I think it would have been simpler to say “transit bad enough that there are no choice riders”.

    • Benjamin Turon

      I understand the point being made her in this blog post; but without the Metro Rail and NFTA buses when I went to SUNY Buffalo I would have been trapped on the campus. For a person from a place where “there was no transit — period” that was quite a change, and a whole lot of freedom. Sure, Buffalo has no transit — unless you come from someplace that actually has no trains or buses, and no much within walking distance.

      • Luke

        That surely must be the summation of why transit in the U.S. is so bad. Also living in a place where the nearest bus stop is 2mi/4km away, and the nearest railway station requires that I travel over a state line 30mi/50km away–the outskirts of Manchester, NH–it becomes easy to see why people in this country have such low expectations for transit system functionality where they live. Ignoring equality and environmental concerns (which, of course, oughtn’t be ignored), that can be fine in the relatively low-density environments that many Americans live in; you just drive if you can afford to. However, that creates a culture (much-referred to in this blog and elsewhere) of low expectations for transit suis generis, and autocentric transport systems simply do not work without the space for them, i.e., cities. So, operators in dense areas like the MBTA can get away with being garbage, even when the Boston metro area NEEDS better transit because it’s not a sprawling suburb, because it’s about as bad as anywhere else in the country. Initiate positive feedback loop, yada yada.

        Having traveled and resided overseas in the UK, France, China, and South Korea, it drives me batty that the whole thing–top to bottom, inside and out–is so bad here.

        • Alon Levy

          Yeah, and the frustrating thing is, the UK and France aren’t even that amazing relative to city size. Stockholm, Munich, Vienna, Prague, and Budapest all manage comparable transit usage to London and Paris per capita, on a metro population in the range of Manchester, Lyon, and Birmingham.

        • Lee Ratner

          Years ago on an alternative history group on Usenet, somebody referred to cars as America’s ethnic form of transportation. I think there really was something about cars that fit into American self-images of itself as a nation of free wheeler-dealers. When transit use became the provenance of the people to poor to afford cars in American society, which was a very small group in the United States because policy made driving cheap, or people living in a select few cities that kept their pre-WWII rail transit and had commuter railways, transit became even more disfavored. Transit use is literally associated with poverty, and this means to a large extent African-Americans and Hispanic-Americans, in the American consciousness. It’s seen as a social service rather than a general transportation service.

      • Nilo

        Buffalo Light Rail is actually a relatively high performer given the length of the system. Buffalo has just been hollowing out for decades. Even good service can’t counter heavy suburbanization and massive metro area decline.

        • Benjamin Turon

          That could be changing a bit with renewed investment Canalside, Downtown, and the Medical Campus. I worked as a rail advocate to keep the new rail station on its current site downtown at Exchange Street — opposed to Buffalo Central Terminal. The state has adopting my idea of connecting the Amtrak station with the Metro Rail by building a new broad walkway under the elevated I-190 expressway to the nearest Metro Rail Station, about two blocks/5-min walk away. I have made this train-subway transfer several times. I would agree that US public transit isn’t that great, but beyond how the the service itself is run; what is need is more walkable environment and increase density along transit corridors.

        • Nathanael

          Buffalo has been revitalized by immigrants recently — many refugees, and especially from Somalia. Entrepeneurs.

          Ben, while I’m glad they’ve connected the Exchange Street rail station to the MetroRail by a good covered foot walkway, it doesn’t solve the long-term problems for Amtrak in Buffalo. Trains to Chicago can’t go to the Exchange Street station, so a second Amtrak station is needed. The Buffalo-Depew station will have to be replaced sooner (not later), and *it* should be replaced with Central Terminal. Oh, and MetroRail should be extended on the preserved, NFTA-owned corridor from downtown through Central Terminal and the Galleria to the Airport… but nobody’s talking about it right now.

          • Luke

            I didn’t know that Chicago-bound trains can’t go to the Exchange St. station; is it that they can’t, or that they’d have to back in and out?

            I certainly love the idea of Central Terminal coming back into service, but was against it for local reasons. Yes, Buffalo is starting to come back, but my concern was that spreading new activity between downtown and the area around Central Terminal risked short-circuiting the whole renaissance; I just don’t know that there’s enough energy in things–at this point–to bring the whole triangle (downtown, Exchange St., Central Terminal) back to life. Certainly, it’s plausible, and desuburbanizing the city by rebuilding its entire downtown to bring in tens of thousands of new residents is conceivable (cities in Korea build residential complexes for thousands on a regular basis), but I don’t know how likely it is. The bigger worry is taking the wind out of sails by making the sails too big, to turn a phrase.

  10. Transportation Justice CNY

    But you’re still erasing those small populations who ride and depend on their cities’ bad transit systems. They look more numerous when they’re crowded onto a bus and you can’t sit down. I probably wouldn’t board and say “Everybody off, Alon says you don’t exist.”

    • Henry Miller

      If there is a bus that is nearly full it wouldn’t be cut in the redesign. It might be rerouted a bit (which might make one or two get out, but more will get on to replace them), but it would exist. It would be faster to where they are going, it would come more often, and run latter at night in case they needed to work late. Probably some weekend service too. Making the existing good lines better is the key to making a transit system work.

      What would be cut is those lines to nowhere with only a couple people on them, and no potential to grow that ridership.

    • Luke

      I think the idea is that while there are populations in these cities who depend on transit for their very livelihoods–that’s true even in my suburbanizedhometown–that’s not by choice. Transit in these places with “no” transit exists–as others have mentioned–as a social service to those in need, not as a fundamental amenity to 21st century urban life. Transit in these places exists only for those who wouldn’t use it if they didn’t have to, and is run to accordant levels of service (bad). That feeds into the whole idea that transit is crappy by virtue of it being transit and not private automobiles. The idea is to reframe the conception of transit by contrasting it with an honest assessment of how most people in this country think about those who use it–transit is crappy because of how it’s run in America, not who uses it.

      • Joseph Brant

        It’s run poorly partially because of who uses it, especially once racial segregation comes into play. It’s assumed that the transit needs of the very poor don’t matter because they have no other choice. So you can treat them like a tabula rasa and not worry about what they think. I think this is why people get offended by the “no transit” characterization.

  11. Ross Bleakney

    When you include MSA and CSA, you are bound to get very low numbers. For a lot of suburban areas, you are completely correct — there is no transit there. What little transit they have is going to have very low ridership compared to driving.

    Part of the problem with this sort of statistical analysis is that rarely is a suburb purely a bedroom community. There are bound to be jobs in that area, along with people commuting to the big city. In most U. S. cities, the suburbs are extremely low density, compared to similar cities around the world. If these suburbs were cut off from the big city, then you wouldn’t expect them to have good transit share, any more than a small town doesn’t have good transit share.

    For example, look at Snohomish County, north of Seattle. Population density is very low, overall, and there is nothing the least bit spiky (more like rolling hills in terms of density) https://arcg.is/D09Tf. Employment density is a bit more interesting. The biggest hotspot for jobs is south Everett (where there is a major Boeing plant) but there are jobs in downtown Everett as well as Lynnwood and Bothell. Most of these are spread out (Boeing is the only real spike). This would not be the end of the world if not for the fact that the county is so sprawling and so small. You just can’t expect an area this size to invest a lot in transit. There are too many places to serve, over too wide an area.

    Yet this counts as Seattle in the MSA, all the way up to Everett. Even though it is part of the same statistical area, there just aren’t that many people commuting from Everett to Seattle. Most just work in Everett (or the surrounding area). As you get closer to Seattle, the number of people who work in Seattle increases. But you still have plenty of people who don’t work in Seattle. As it turns out, the commuter bus service into from Snohomish County is fairly good. It wouldn’t surprise me if a substantial portion of those who commute from Snohomish County to Seattle do so via transit. Thus the whole reason that the area is included in the MSA (commuting to Seattle) actually has decent transit ridership, while other trips (purely within the county) water down the numbers substantially.

    Ideally this would be part of the analysis, but I don’t think it is. They just look at commuting as a hole, not commuting into the urban center.

    You could also look at numbers simply within each “urban core”. It is obviously challenging to come up with such a definition. I don’t know if the census department has done that. They have done the opposite, drawn larger and larger circles around the city, and considered them part of the same thing. There is certainly value in that, but the farther out you go, the less these places really are part of the nearest big city. We are left with definitions based on municipalities. These are challenging (San Fransisco and Anchorage are extremes in terms of size) but likely more revealing. For example, Seattle has a transit commuter share of 23% up from 18% just a few years ago. Driving alone has dropped to 43%. In contrast, Snohomish County has gone from 5.7% to 5.9% in the same period, and 72% drive to work (their only significant change is a 2% increase in the number of people working at home).

    I’m guessing that there are plenty of other cities with similar numbers. It is quite possible that many U. S. cities don’t have public transportation. But I bet there are plenty that have adequate public transportation, except that those who live outside it (and aren’t even trying to get in it) don’t.

    Regardless, it means that most of America lacks good public transportation, since so much of America lives in low density suburbs.

    • Alon Levy

      I mean, these dropoffs happen everywhere. Paris has a 64% mode share, Essonne (where IHES is, where I reverse-commuted for 3 months) has a 28% mode share, and everything together averages out to 43%. Likewise, Portland proper is at 12.6% (which is 3/4 bus, 1/4 rail), and the rest of the metro area is even lower, and everything together averages out to 6.3%.

      What’s more, the extensive suburban job sprawl is precisely why such regions have no public transit. You will never serve haphazardly strewn office parks with public transit well. They’re built at automobile scale, so a bus can only serve one place at a time, and unless it’s one giant corporate HQ with a large number of workers in one origin, like the Googleplex and San Francisco, nobody is going to bother.

      • Ross Bleakney

        Yeah, of course these dropoffs happen everywhere. The point I’m making is that focusing on wide regions instead of cities is interesting, but not exactly revealing. No one in Puget Sound would consider Everett a suburb of Seattle. Lynnwood, definitely; Everett, of course not. The MSA regions are drawn so broadly as to include anyone who ever once commuted from (or to) there.

        In general, things average out (which I guess is your point). If the city proper has lots of transit riders, then it can make up for bad transit in the suburbs. But my guess is the numbers are largely a reflection of how suburban these areas are, and how loosely the MSA are drawn. Everett does not have great transit — neither does Spokane. But neither are part of Seattle.

        I’m just saying that the title, and everything else that followed is misleading. It is really the suburbs and the small cities that lack transit ridership. Suburban job sprawl hurts urban ridership a little, but it more or less destroys suburban (and small city) transit ridership. The midsize cities are nowhere near their European (or Asian) counterparts, but they aren’t horrible.

        The other drawback with this approach is the numbers have to be really low to pass the sniff test. This means that you include Kitsap County, which sounds absurd. I doubt there is any place in Kitsap County that comes close to the transit ridership of Seattle proper (let alone the core of the city). Fairfield County also made the cut, but are its numbers actually much better than Portland proper?

    • adirondacker12800

      Yet this counts as Seattle in the MSA

      That the way it works. If there weren’t a lot of people commuting from Snohomish County to King County or vice versa they wouldn’t be in the MSA. They don’t have to commute to Seattle, they just have to commute to King County. Or vice versa. Even if it’s a quarter of mile walk to a minimum wage job. That’s the way it works.

      • Ross Bleakney

        If there weren’t a lot of people commuting from Snohomish County to King County or vice versa they wouldn’t be in the MSA.

        MSA boundaries aren’t drawn with that much detail. For Seattle, the MSA includes entire counties. This means you include areas over hundreds of kilometers away, separated from Seattle (and Puget Sound) by major mountain passes.

        They don’t have to commute to Seattle, they just have to commute to King County.

        Which is why the definition is absurd. Of course there is a relationship between these areas, but no one would call these places greater Seattle. The definition includes small, independent towns, as well as suburbs of other cities (most people would call Marysville a suburb of Everett). Drawing a map of greater Seattle would involve something tighter, based on current data (such as commute share to Seattle). It wouldn’t automatically include the entire county, nor would it include people who commute across the county border, a long way from the city.

        The MSA (and CSA) were meant to lump together various areas without overlap. That is why they are drawn so broadly. But in the real world, many of the areas aren’t that large, and you are bound to have overlap. But naming all of those overlapping areas and calling it the biggest city seems silly.

  12. Lee Ratner

    The problem is that Americans were early adapters of mass automobile ownership for a variety of reasons, that during the Cold War mass automobile ownership became ideological enough that there was no need to develop transit after World War II and that combined with suburbinzation led to a transit death spiral. We have hundreds of millions of Americans who have no experience with transit despite living in some of the densest and built up environements ever, and American transit planners and builders are messing up revitalizing transit.

  13. Martin

    While focus of much of this is how cheap driving is in US, the other aspect is time which is broken down into cost of car ownership, operation and parking. Registration fees are low, gas prices are low, parking free, and anytime transit shares road space with cars – always slower and unable to recover any time from driving and parking.

    There are tons of dials that need to move together which is why Seattle ridership is going up while in other towns, it’s going down.

    Might be worth an article whether it’s worth having a single agency control all the dials or not.

    • Lee Ratner

      Increasing transit use in the United States is going to require big natural or artificial increases in the cost of driving. I don’t think any politician is going to have the political bravery for the later because the pain inflicted would be political suicide. You are also going to need to re-concentrate jobs and entertainment/ and increase housing density. The later are more plausible but still politically hard because of the great baptist and bootlegger NIMBY coalition. Increasing transit use will require going what against many voters want.

  14. James S

    If your metric results in a statement like “LA has no transit” than your metric is wrong. it reminds me of those ridiculous lists with headlines like “Las Vegas found to be most sustainable city in the US!” Clearly the methodology is wrong.

    Even with the decline, Metro has 1,155,119 daily riders (as of November 2019). That;s a whole lot of “nobodies.” Stand on any corner in LA and you will see a bus go by within 2 minutes. That’s not something you get in Kansas, where you might need to go 2 miles to find an arterial with a bus that comes every hour from 7am to 5pm.

    What exactly do you gain by making such inflammatory statements? If your goal is twitter engagement, great, but as an advocate or policy person, I don’t see how this argument is productive.

    • Alon Levy

      …2 minutes? Wilshire 12-minute frequency on the rapids, and that’s the busiest corridor.

      And 1,000,000 unlinked trips in a region of 12 million (really 18 million) is pitiful.

        • jonahbliss

          And while I won’t disagree that overall ridership is extremely disappointment, if you’re going to talk about the entire region you have to include the ridership of all the other operators. Adding in OCTA, Omnitrans, RTA, BBB, Foothill, Long Beach, etc gets you another ~500k trips)

        • Alon Levy

          Off-peak service matters and one of the ways US transit fails is in providing too little of it, often because the people who manage the agencies work 9-to-5 and think everyone else does too.

          • Lee Ratner

            The everybody works 9 to 5 stuff was true for quite awhile.

          • adirondacker12800

            People who 10 to 6 or noon to midnight don’t have to deal with rush hour and drive?

          • adirondacker12800

            I’ve been in Los Angeles traffic. They exaggerate. Rush hour isn’t awful either.

    • Henry Fung

      I’m from LA and get the statement. By global standards, Los Angeles has no transit, especially compared to the other GAWC cities in its classification, save Miami, Bangkok, Kuala Lumpur, and maybe Istanbul and Jakarta.

      And rush hour traffic is bad in a central core area but outside of that, it is fine. Worse than Fresno, Kansas City, Charlotte? Yes, but I’ve been in mid size cities (Portland, Seattle) with worse traffic.

      • Eric

        “less transit than similar cities” is not the same as “no transit”. I have less hair than most humans (I get short haircuts, even for a man), but it would be very misleading to say I have no hair.

      • Alon Levy

        Hey, in KL the transit mode share is high by US standards, it just sucked by Singaporean ones. Per an ASEAN-wide overview, PDF-p. 135, KL’s mode share bottomed at 16% in 2010 and has begun climbing (and this paper says 17% of all trips) and sources here and here seem to be saying it’s around 20-25% now (it’s weird, they say this is for Malaysia or urban Malaysia, but the numbers are clearly inconsistent with that, but could imply 20-ish% in KL with low workforce participation).

        Bangkok and Jakarta both have a lot of bus ridership out of poverty. Per the ASEAN overview, 5% of trips in Bangkok are by rail and another 37% are by bus; this source says 6% of trips are by rail as of 2015 (and please ignore the future projections, MRT/BTS expansion has been delayed).

    • Ross Bleakney

      I agree. It is one thing to say that L. A. didn’t make the cut, but then you go on to write that Kitsap County (and Portland) did. That just doesn’t pass the sniff test.

      I think the problem is two fold. One is the way that MSA borders are drawn. L. A. is bunch of overlapping cities and the MSA borders are huge. If you consider it one big city then of course the city has terrible (to no) transit. Many of the surrounding cities are really big and really sprawling. As a result they have enough people to skew the results, while hardly any transit commuters. But transit within the core of the city is much like the core of San Fransisco — slow, not very good, but still used by a substantial portion of the populace. The difference isn’t that the surrounding cities (or suburbs if you prefer) has worse transit than greater L. A., it is that that they count for a much bigger portion of the population.

      The other problem is that you are focused on commuting. Want to get around central L. A. (a very large area, with millions of people)? Yes, you can do that on transit, even in the middle of the day. Want to get around just about anywhere in Kitsap county in the middle of day? Don’t be ridiculous. You just can’t do it.

  15. Pingback: Quick Note: What is Culture, Anyway? | Pedestrian Observations
  16. anonymouse observer

    Haven’t Americans been in love with cars before the industrialization: far before engines replaced horses and other work animals?

    I believe American cities missed out the opportunity to invest heavily in public transit. They should have invested heavily to public transit probably during the heyday of interurban. There is a paper written in similar subject, which include interesting figures and arguments even though it mainly talks about East Asian cities:


    It says that timing of large-scale investment to public transportation really matters in terms of slowing down the motorization while maintaining or even recovering the transit ridership. If the timing is right, the transit ridership could potentially bounce back to pre-motorized era level. The paper focuses on cities in East Asia, but this framework could be applied to cities outside of East Asia.

    Majority of these cities you characterize as “city with no transit” started putting major efforts and investment into public transit too late, and thus, the it does not generate ridership big enough to increase the transit mode share (probably except in Vancouver, BC). Based on this, I think it is highly unlikely that L.A. can reverse the century long downward trend in transit ridership even with that level of investment to transit.

    • Eric

      The issue is not so much investment as city form. US cities invest huge amounts in transit – not only the expenses (astronomical per mile of course) but the number of miles built. However due to the low density of US cities, this infrastructure is less useful. For example Dallas light rail (built from scratch in recent years) is about the same length as the Budapest tram, but the Budapest tram has about 12 times as much ridership. Vancouver has increased its transit share precisely because along with SkyTrain it has built forests of skyscrapers. Asian cities are already dense and probably getting denser, so transit will have no problem succeeding there. American cities, which are unlikely to densify the way Vancouver did, are a more difficult question though.

      • Ross Bleakney

        Yeah, I think it is a combination of low density and transit systems that sprawl. These go together, for political reasons. Rather than focus on building up the core of Dallas, it is easier, politically, to build a subway from Dallas to Fort Worth. BART and Denver’s light rail are similar. These don’t work very well, in the long or short term. Employment sprawls, and so does residential development. Despite having some transit to those areas, it isn’t enough to get good transit share. It is just harder to serve a long skinny line with good transit along the way (other than the train itself). The long lines are often close to (or even adjacent to) freeways. A long distance subway has trouble competing with a car, and yet there are very few trips between suburban stations. A long line is also more expensive to serve; without good ridership, frequency is reduced.

        Yet these sorts of proposals are all popular. Partly it is because they can please suburban interests. Like all politicians, suburban leaders love a big ribbon cutting ceremony, and think the work is done as long as there is a line into the city. But if the city itself has bad transit, then they are out of luck. They are way better off with a good suburban bus network (which has a good connection(s) to a strong urban subway system) and maybe some affordable commuter rail. The other problem is that the populace as a whole makes the same basic mistake — probably because they think in terms of driving. Long distance subway lines sound great — they seem like a big freeway. Distance becomes synonymous with value, which is ridiculous.

        For example, consider greater San Fransisco. More people ride the Muni buses (just the buses, mind you) than BART. Let that sink in. BART makes the key connection across the bay. It is 175 kilometers long, one of the longest in the world. There are massive systems that carry millions of riders each day that are smaller. It is also very fast, with large stop spacing and trains capable of 110 KM (faster than the driving speed limit). It covers the entire region. Muni buses are notoriously slow — the slowest in North America. They only serve San Fransisco proper (a city that is relatively small). On paper, BART sounds great, and the buses look like they would be irrelevant. Yet the buses carry more people than BART. The main reason BART does so poorly is because they didn’t build the core. There are only a handful of stations — all on one shared line — in San Fransisco. East Bay is no different, with only a handful of stops, and one line. Instead of having multiple lines criss-crossing the urban core (San Fransisco, Oakland and Berkeley) with lots of stops and minimal extensions into the suburbs (focused on bus to rail intercept) it is the opposite.

        In contrast, the systems that are built out from the center, like a traditional subway have great success. Washington D. C. did not try to run a subway line to Baltimore — they covered the city instead. Just about all of the city itself has a stop within walking distance (Georgetown being the notable exception). It is not like the city is any less suburban than other cities. But people in various suburbs often have to take a bus to the nearest train station — or a bus right into town. But once they get there, they can get around just fine using public transit.

        The sad part is that cities seem to be following the failed BART model. Seattle is embarking on a huge mass transit operation, that will rival BART in terms of size and cost. Yet most of the inner city — where you have the highest concentrations of people — will get nothing. Trips within the city that are notoriously difficult (because of the challenging geography of Seattle) will continue to be very slow. Huge amounts will be spent, and yet relatively few will benefit.

        • Lee Ratner

          BART was built to be a commuter rail system but operating with much greater frequency. The entire idea was people live in the suburbs but come into the city for work and fun. That’s still true to a bigger extent in SF than other metropolitan areas but not really. There was also supposed to be a line down Geary, which never got built.

    • Lee Ratner

      I think for the heyday of the interurban, the problem was that most Americans cities were of the size where streetcars were all you needed in the form of urban transportation. Most of them didn’t reach subway size until the peak urbanization of the mid-20th century and then started to decline because of suburbs and cars. Detroit apparently planned to build a subway system in 1918 but that got defeated because the Mayor wanted it to be entirely public rather than privately owned, so something transformative didn’t get built for political reasons. The Cincinnati subway was abandoned before it got completed. The Los Angeles 1925 plan was not implemented because LA residents favored their personal interests over the culture, although the plan also called for elevated rail more than subways so that didn’t help. There were a bunch of ambitious plans in the 1960s and 1970s that went no where or only got partially built. So its not that transit planners didn’t know what to do, its just that there was a lack of political will to do it.

  17. anonymouse observer

    So as in urban areas in other countries, I think. What actually happened in other countries with functioning high performance transit system/network (or what did not happened in the U.S.) is that streetcars, interurban, and railroads all evolved to the higher performance passenger rail by innovating themselves.

    For instance, majority of non-JR urban rail lines in large metropolitan areas were originally streetcar line (e.g. Hanshin Main Line, Meitetsu Line, etc.) or interurban (e.g. Kintetsu Lines, Keihan Lines, etc.). These streetcar and interurban lines gradually became modern rapid transit-like service over years by addressing crowding, dealing with competition, and/or building new or additional tracks in dedicated right of way (including grade separations using viaducts and tunnels/subway), which requires continuous investments and improvement efforts. At one point, the national railway (pre-JNR era Ministry of Railroad) built their own version of interurban-like service on Tokaido Main Line (the primary intercity rail corridor in Japan) in Osaka region (a part of JR West Tokaido-Sanyo Local Line between Kyoto and Kobe).

    This evolution allows one-size-fits-most rail system you see in Japan today. Each passenger rail service or infrastructure (except Shinkansen) offers multiple tiers of services serving both local/intra-city demands and intercity demands (as well as freight service on JR lines). Because they mix both local and intercity service on the same track, you get better service in terms of frequency and speed. I believe this increase return of investment since they do not have to build a new infrastructure for each type of service while the better frequency and speed attracts more passengers.

    Such evolution somehow did not happen in the U.S.. Railroads and interurban did not learn from each other or improved themselves like it happened in other countries, and rail transit eventually died; what’s left was dysfunctioning legacy systems or new underperforming systems which was built too little and too late.

    American railroads appeared to be the worst in this according to a book “The Passenger Train in the Motor Age” by Gregory Lee Thompson:

    Railroads literally let the passenger service lose against buses in relatively early days (between 1910s and 1940s) just because they didn’t know the cost structure and their strength and made the trains longer (more infrastructure needed to operate frequently) and heavier (cost more to operate and maintain) instead of lighter, shorter, and more frequent.

    • Lee Ratner

      To be fair to American railroads and interurbans, they faced a lot of political restraints that prevented them from updating themselves even if they wanted to. Detroit’s street car system wanted to turn itself into a private subway in 1918. The Detroit City Council approved but the Mayor vetoed the project because he wanted the Detroit subway to be public owned. A similar mid-1920s plan for a private-public partnership to grade separate the Pacific Electric was rejected by voters for short term reasons. Other plans faced similar defeats or couldn’t be built because of not enough money due to the Great Depression, etc. There were lots of private companies and governments that knew what needed to be done. The public will to do it, necessary for getting the budget to build, wasn’t there.

  18. Pingback: How Minneapolis Can Become A Transit City In One Decade | streets.mn
  19. Pingback: Metcalfe’s Law for High-Speed Rail | Pedestrian Observations
  20. Pingback: America’s Most Liveable City? A 1-Year Retrospective on St. Paul | streets.mn
  21. Pingback: ¿La ciudad más habitable de Estados Unidos? Una retrospectiva de un año sobre St. Paul - Baristas y Café - Noticias del mundo del Café para el Barista y el Entusiasta
  22. Pingback: Quick Note on Los Angeles and Chicago Density and Modal Split | Pedestrian Observations

Leave a Reply to Herbert Cancel reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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.