US Commuter Rail: What it Is and What It Could Be

Jake Blumgart has a superb piece in Governing (free but click “Continue to site.”) about the distinctly North American artifact called commuter rail, and why it’s so different from the way heavy rail infrastructure is used for transit in most other developed countries.

The key difference is that most other countries want their heavy rail services to be useful all day, while the default in most of the US has been to run only at rush hour with at best minimal service the rest of the time, as though the briefcase commuter is the only conceivable customer.

There are not such sharp contrasts between regional rail and the rest of transit systems in most wealthy European or East Asian nations. But in North America, the divide was sacrosanct. As recently as 2016, then-MBTA General Manager Frank DePaola drew a bright line between this service and the rest of the agency’s subway, bus, and light rail services: “Commuter rail is commuter rail. It’s not transit. It’s designed to bring people into the city in the morning and take them home at night.”

The COVID-19 pandemic, however, calls that whole concept into doubt.

“It’s not transit.” “The divide was sacrosanct.” Of course this division of the market by trip purpose also implied a distinction of social class, manifested especially in very high fares.  But high ridership is diverse ridership.  If the goal is to help everyone go places, it’s always best to design services that are useful to everyone and make those services connect, rather than run two systems side by side (commuter rail and slow buses, for example).

Commuter rail services also tend to run long distances across core cities without providing much relevant service to them.  They either run nonstop or offer a frequency that’s too low to be relevant for the shorter in-city trips happening along the line.

I have my own trauma about this, because I spent too many years in the 1980s trying to advocate for an all-day frequent service on Caltrain, the commuter rail line between San Francisco and San Jose in California.  Caltrain has the geography of a frequent all-day rapid transit service: it runs through the historic downtown of almost every city it serves, because the downtowns grew around the rail line.  In terms of the useful transit provided, it could have functioned as another BART line.

Yet decades later, Caltrain service levels are still terrible in the context of the corridor it serves.  It was big event when the midday frequency was improved from two hours, as it was in my day, to one hour, but that still makes it irrelevant to most trips along the corridor, especially those of less than 10 miles or so where waiting time becomes more onerous.  So a bus system has to run alongside it, inefficiently serving some of the same trips that rail could be serving, and unable to efficiently feed the rail line very well because of the rail’s low frequency and erratic schedule.

Since then, as a bus network planner, I’ve encountered the same problem in many other cities.  There’s just no way to integrate commuter rail with a local bus network, because good bus networks involve regular patterns of frequency that are not what US commuter rail does.  At best you have to provide dedicated shuttles that meet the trains, and that’s a form of duplication that leads to worse access for everyone.

Blumgart’s piece touches on the effort that’s now being made to rethink commuter rail to make it more like what it’s always been in Europe, East Asia, and even low-density Australia and New Zealand.  In all these countries, rail is how you travel longer distances across the region all the time, not just at rush hour.

That difference arises in part from a different geography of social class.  When I lived in Sydney, I wasn’t happy with midday frequencies on the all-day rail network — mostly every 30 minutes then, mostly 15 now — but that was far superior to what most big US cities are used to.  The difference is the suburbanization of poverty, which has been happening for decades in Europe and Australia but is only now accelerating in the US.  The most remote parts of Sydney are some of the poorest.  In this urban structure, lower income people have to travel longer distances, but mostly to non-peak jobs.  Making heavy rail services part of the total transit system isn’t just a better use of infrastructure, it’s increasingly going to be an equity issue.

If Covid-19 causes a permanent drop in rush-hour commuting, we could see a golden opportunity to make better use of all the existing infrastructure of commuter rail.  There are plenty of obstacles, but they aren’t physical. They’re mostly cultural issues embedded in regulation and labor practices.  When it becomes important enough, those can be solved.

32 Responses to US Commuter Rail: What it Is and What It Could Be

  1. Alon Levy April 26, 2021 at 10:31 am #

    Canada has US-style commuter rail. Montreal specifically severed the Mont-Royal Tunnel from the mainline network to build REM. Toronto is trying to increase frequency on GO Transit but the project is plagued with trad railroader resistance and very high construction costs. Envying Canada here is actually bad.

    • Jack O April 26, 2021 at 11:31 am #

      And the REM pretty much eliminates any hope of through-running for Canada’s VIA HFR project. Unless they build a new tunnel for such. I feel like that’s a worse crime than railroader resistance for Toronto’s GO RER project.

    • Peter L April 26, 2021 at 1:35 pm #

      REM is a very bad joke told poorly. There is huge potential for it to damage Montreal’s ability to build transit for a generation or more.

      And it’s ugly

      • Uche April 26, 2021 at 7:48 pm #

        What will damage Montreal’s ability to build transit is those high construction costs. The Blue Line extension of the Montreal Metro costs $6B for a 5.5 km extension. $1B/km. Not even Toronto has reached $1B/km in their new subway lines/extensions to my knowledge. As a result, the Blue Line extension will probably be the last extension of the Montreal Metro for decades to come.

        As for the REM, it’s a noble attempt to build extensive rapid transit service at lower costs, via the use of shorter but more frequent trains. If anything, Montreal will have the second most extensive rapid-transit network in North America when both REM projects are complete which is truly impressive. But that comes at the cost severing the Mont-Royal tunnel and losing your only through-running mainline. VIA HFR cannot run to Quebec City because of that and any sort of EXO commuter rail improvements die along with it,

        Montreal should’ve done what Toronto did and bought up their mainline tracks so that they can run more trains and move towards and S-Bahn/RER system like in Toronto. Not saying we should completly envy Toronto as they’ve made questionable decisions, insanely high costs and had railroader resistance for such a conversion as Alon mentioned.

      • Dan April 28, 2021 at 9:38 am #

        Surely REM is a step in the direction Jarrett’s is suggesting? The project converts 2 commuter rail lines (Deux Montagnes and Mascouche) with an all day bi-directional service with headways similar to a metro.

        Yes, the Mont Royal tunnel conversion means that the Mascouche line gets ‘chopped’ and will affect potential HFR services. However in terms of equity I would suggest that dramatically expanded urban transit provision across the entire Montreal region (REM) represents a ‘better’ outcome than potential impact on long distance rail services in terms of equity.

        • Peter L April 28, 2021 at 1:33 pm #

          Except that it’s a modern version of the Expo ’67 Minirail — OK, a slight exaggeration — and not what Mr Walker is talking about. My numbers are from a few years back when the first horrifying numbers came out of CDPQ-Infra’s team of cilvil engineering school seniors (who developed this as their “capstone” project – I made that up … or did I?) so may be out of date.

          4-car trains with 30 seats per car, so 120 *seats* per train. Claimed crushload of 780/train (5.5 standees/seat).

          Former system — completely rebuilt ballast-to-power-distribution in 1995, mind you — usually used 9-car trains of MR-90 EMUs each of which sat about 90 (depends on presence of cab, washroom, etc) giving 830 seats per train and a claimed crush load of 1900 per train (1.3 standees/seat).

          The problem is that the STCUM didn’t buy any more MR-90s. Nor did successor AMT. Even after they renamed to exo (lower case). But there were only ever enough sets to send 9 trains in to Central Station at morning peak (0600-0900). Even with the Mascouche trains with their Unicorn dual-modes and bought-for-the-tunnel double deck cars, the 100+ year old tunnel was nowhere near capacity.

          The only way that REM gets better numbers is that they claim (again, numbers may be stale) is by running 54 of their minirails in the same time period. Each carrying a crush load. And if you think that “well, they can just run double sets of cars” … apparently the platforms are all exactly one train-length long.

          I *don’t* think my numbers include trains to/from Baie d’Urfe or YUL, only from the “main line” to Sainte Eustache.

          Had AMT ever invested in more train EMU sets, REM likely never would have happened.

          They are trying to build an airport train and build a train over the (new) Champlain bridge to the South Shore and, well, let’s also throw a bone to the West Island and pretend that it will be paid for by investors (it will not lose any money any more than a man can have a baby). And, as a bonus, they managed to get a “shot” at the unions by making it driverless!

          Have you seen the western stations? Imagine waiting on the platform in January. Think Red Line stations along the Dan Ryan or Blue Line stations along the Kennedy. Yeh. Brilliant putting them 10-20 m above ground.

  2. James S April 26, 2021 at 10:33 am #

    Not just social class, but also race. By essentially allowing one-way service (inbound morning, outbound peak) they prevent the “urban” people from taking the train to the suburbs where naturally, they “will commit crimes”.

    Some regions are more explicit about this kind of racism than others.

    Also why commuter rail trains even within the city will charge a $8 fare to go one stop when a bus line serving the same corridor charges $2. It’s to keep “those people” away. NYC is big on this policy. Of course the official justification is that the trains are very expensive to run, but having 8 staff on one train is a choice.

    NJTransit has 4 or 5 rail stations within Newark that were closed in the 80s. Now, suburban trains go by without having to stop in the city. Newark residents are directed to the bus. Im sure the legacy of closed inner stations is true in most big cities.

    • Peter L April 26, 2021 at 1:45 pm #

      > they prevent the “urban” people from taking the train to the suburbs where naturally, they “will commit crimes”.

      Did you know that the children of the instigators of the LOOT RAIL thread on USENET have taken up where their parents left off when they died? True story that I just made up.

      I am fine with a unionized workforce (and kinda wish I was in one) and don’t want anyone to lose their job, but for god’s sake, why are so many CR systems still using on-board fare collection?

      I moved away from Metra 15 years ago now, so may have out-of-date info, but back then not only did Metra have a hierarchy of people (Conductor, Assistant Conductor, Collector) on the trains collecting fares but they had human-staffed stations to sell you a ticket! And this in a system where most (peak hour) riders had a monthly and many of the rest of us had 10-rides.

      • I B April 30, 2021 at 4:48 pm #

        OMG, I remember “Big Don” of Washington state and his rants on misc.transport.urban-transit!

        • John Charles Wilson May 10, 2021 at 1:56 pm #

          OMG, I remember him too. Used to get a charge out of pissing him off. Unfortunately, I probably pissed off many innocent people too. I’m the Communist who believes Laura Ingalls Wilder is God. Pretty sure if you were on m.t.u-t, I don’t need to say more.

  3. Steven April 27, 2021 at 10:55 am #

    Given the stunted and distorted nature of rail transit in North America, is high-speed rail an expensive waste of money and focus in the US?

    • AJ April 29, 2021 at 3:06 pm #

      If it is just fancy branding to create dedicated ROW for passenger rail and run frequent, all day trains, no.

      If it is an attempt to run trains faster without improving horrible frequency and span of service, yes.

  4. Jarrett April 28, 2021 at 10:13 am #

    NOTE: Not wishing to launch any more bouts of Canadian self-criticism, I’ve deleted the reference to Canada to which some of you are responding. I was thinking mostly of Toronto, but Vancouver’s commuter rail is definitely US-quality and I guess I was wrong about Montreal.

    • Uche May 8, 2021 at 6:54 pm #

      You weren’t wrong really. Toronto’s RER and Montreal’s REM are the most ambitious attempts for Regional Rail in North America and should be commended. Toronto’s and Montreal’s transit networks would be among the best in NA. But the way they were done is nothing to envy and deserves scrutiny. What it should do is light a spark in every other NA commuter rail agency to transform their Commuter Rail systems to RER/S-Bahn. Looking at you METRA, MBTA, LIRR, Metro-North.

  5. AJ April 29, 2021 at 3:11 pm #

    Do you think the mooted merger of Caltrain with BART will drive a major change in how Caltrain sees itself?

    For all of BART’s criticisms as an urban subway system, it provides frequency and span of service that would be consider good for regional rail in most of the world. If BART acquires Caltrain and starts with the mindset of “let’s run this line like we run our BART lines,” that seems to be a good thing?

    • PM May 16, 2021 at 9:31 am #

      No way. BART is an essential tool in the Daly City-Berkeley -Oakland triangle, but outside that triangle it has insufficient ridership to pay for its exorbitantly high fixed capital and operating costs. South of Daly City, BART is expensive, infrequent, badly designed, and poorly run.

      On weekends sometimes a minor delay on BART resulted in 90-minute waits to connect to CalTrain. It took BART and CalTrain 19 years to float the idea of coordinating transfers at Millbrae.

      BART’s Milpitas station opened 3 years late, and each passenger embarked or disembarked at Milpitas cost hundreds of tax Dollars )according to the Mercury News). Neither BART nor CalTrain has managerial competence; for now it is safer to contain their individual damage.

  6. asdf2 April 29, 2021 at 6:01 pm #

    A lot of the reason you off-peak commuter rail service in the US is so abysmal is that it’s operated on a cost structure with a very high marginal cost per trip.

    In Seattle, for example, the agency didn’t have the money to actually build their own track, so commuter rail operates on rented track from a private freight rail company. Each additional round trip the agency wants to operate requires negotiating an agreement with the rail company, and the railroad company always has the upper hand because the transit agency needs them (they have no alternatives except to give up and run buses) more than the railroad needs the agency’s money (they make plenty of money as it is, just running freight trains). So, each additional weekday round trip costs tens of millions dollars annually, just in trackage rights, before even getting into the cost of actually running the train. On top of this, antiquated rules require the line to be operated with gigantic trains and multi-person crews. This results in an insanely high cost per passenger, unless enough passengers exist to actually fill the gigantic trains, which, of course, happens only during rush hour.

    It’s the classic tradeoff between capital cost and operating cost. When the intent is to only run the train a couple times per day, paying more in operating cost to save on capital cost seems to make sense. But, the tradeoff is that if you later decide you want an all-day, frequent train, the cost quickly becomes prohibitive.

    • Peter L April 29, 2021 at 6:38 pm #

      Which “antiquated rules require the line to be operated with gigantic trains and multi-person crews”? Please be specific.

    • AJ April 30, 2021 at 3:31 pm #

      I think the line of criticism in this article is generally directed at ROW fully owned by the relevant public agency, like Caltrain, rather than a service where specific time slots are purchased for passenger rail service, like Sounder or many Amtrak routes.

      I believe in Chicago’s Metra it is a mixed bag, where some lines’ ROW is fully owned by Metra but others are shared with freight.

      • Peter L April 30, 2021 at 5:30 pm #

        Metra does it all ways – UP owns/operates UP-N, UP-NW, UP-W and BNSF owns/operates the line to Aurora.

        Metra operates trains over CN for North Central Service and Heritage Corridor and over NS for the SW Service.

        The other lines are “Districts” — Metra Electric, Milwaukee North, Milwaukee West, and Rock Island — are all Metra, ballast to dispatching.

    • Mike May 18, 2021 at 11:27 am #

      The difference in the rest of the world is the government owns the tracks and prioritizes passenger rail. Even when private companies own the track (as I think is the case in Japan), they’re fulfilling the government’s passenger rail plan. The US has always tried to cheap out, initially by giving the railroads free land to avoid spending taxpayer money, so the railroads run what benefits themselves, which is lucrative freight. They have a minimal requirement to run Amtrak, and some states and metros have negotiated leases to run regional or commuter rail. But the railroads hold all the cards so they demand a robber-baron rate, if they allow any passenger rail at all beyond the Amtrak mandate.

      The “antiquated rules require the line to be operated with gigantic trains and multi-person crews” asdf2 mentioned refers to the difference in regulations between the US and Europe. Europe uses lighter passenger trains to reduce cost and energy use and increase speed. To minimize collisions between heavy freight trains and light passenger trains, they put their highest priority on crash avoidance, using technologies like positive train control (PTC).

      The US takes a different approach (although this is changing). It requires passenger trains to be as heavy as freight trains so they’ll withstand a collision. This makes passenger trains expensive, slow, and gas-guzzling. It also means we can’t use off-the-shelf state-of-the-art train designs, they have to be customized to US regulations. The US passenger-train market isn’t large enough to attract many US companies or US factories of foreign companies, so the “Buy American” provision means that international-standard passenger-train models either aren’t available in the US or cost more than they otherwise would. Lighter trains have to run on separate tracks from freight or be time-isolated. (E.g., light trains daytime, freight trains nighttime.) This affects both “light rail” and lighter forms of heavy rail. It costs billions of dollars to buy a new right of way and build a new track, so the regulations effectively make lighter trains too expensive for metros and states to afford. There has been some loosening of the regulations over the past few years, so that lighter European trains can be allowed in more circumstances. But then you run into the inertia of states and transit agencies, which are sometimes slower to change than the federal government is.

      I don’t know about the “multi-person crews” part, but this might be another area where US regulations are anachronistic.

    • Mike May 18, 2021 at 11:44 am #

      There is one good thing about the US approach. More of US freight goes by rail than in other countries, so that makes it more energy-efficient than countries that rely more on trucks.

      But this is where a conflict arises with high-speed rail. The ubiquidous American freight rail trains are *slow*. Customers have found that they don’t need commodities shipped *quickly* as long as there’s one arriving *every day*. It doesn’t matter *which* batch of iPhones or oranges you get today as long as you get something. So freight can waltz for three days across the country like they’re dancing the Capulets and Montagues, and nobody cares. But this causes complications if freight trains are running at 50 mph and passenger trains are running at 79, 90, 125, or 150 mph on the same track. Slower trains mean less fuel and maintenance costs for the railroad, but passengers want to arrive quickly.

      • Henry May 20, 2021 at 12:45 pm #

        Normally this is resolved by running freight and passenger trains on different tracks, if their needs conflict too much.

        As an example, CAHSR is not using the single track line through the Tejon Pass, nor should it; a new alignment is being built instead. (The placement of that new alignment is a whole different questions.)

        To give another example, a big impetus behind the Chinese HSR network was to free up slots for slower freight trains on the legacy network.

        Toronto is doing this by building new pairs of tracks alongside the old freight tracks. Of course, once you get to that conclusion, you may also reach the conclusion that if you’re going to build a new set of tracks that the current right-of-way is not the best place to put them anyways for passenger traffic.

      • Dondegroovily May 22, 2021 at 1:15 pm #

        Freight trains at 50 MPH? Lol, try 25, that’s more typical.

  7. Glandu May 14, 2021 at 6:21 am #

    As an European who never lived in the USA, I must say what I thought was obvious, is not. Thanks for hte pointer.

    The most impressive memory I have on the topic is the 3 cities network of Gdansk, Sopot, Gdynia, in Poland. One train every hour during the night. We took that kind of suburban train from Sopot around 2h30 to the central railway in Gdansk, for our main travel back in city express. We did see a few people back to their homes after a party – and it seemed like usual for them.

    That would be unthinkable in Paris, where closing the traffic between 01h00 & 005h00 is a religion. But that’s still 20 hours of trains per day. At least one per 30 minutes. I didn’t even realize it was thinkable to close outside peak hours.

    You’re spot on also on diversity. In the route I was taking still a few years ago, wealthy workers (there was even a colonel of the French Air Force) were crowded on peak hours, while people in the poorest jobs were scattered at impossible times (mostly blacks, Paris is segregated that way, by time of the day). And of course, some tourists, at any possible time of the day. I can’t even imagine the usefulness of a network that would be so limited in hours of usage.

  8. david vartanoff May 17, 2021 at 12:03 pm #

    Metra’s Electric District in Chicago is the saddest case. The Illinois Central built a multi track electrified dedicated ROW for commuter services in 1926. Trains ran very frequently all day–locals,expresses,’Specials’, making them useful not only for the 9-5 crowd, but also for shorter trips from neighborhood to neighborhood.
    Today, Metra operates skeletal service much as Jarrett describes on Caltrain years ago. Ridership is minimal, with many on one branch park and ride passengers from Indiana. There is advocacy in Chicago calling for fare integration of MED and Rock Island into CTA. Unfortunately CTA is more protective of its turf than rider needs.
    As to BART and Caltrain, this is a real problem. BART for its first 40+ years dissed traditional subway riders, branding itself regional rail. (BART also pioneered the US exorbitant cost/mediocre results pattern so well documented by Alon Levy and others) BART’s fare structure and service patterns discourage short trip rides. While a few of the elected BART directors are more transit friendly, staff is not.
    Caltrain OTOH seems to be moving to all day short headway scheduling.

  9. RossB May 19, 2021 at 9:46 am #

    I think there are political reasons and practical reasons for the existing approach. From a political standpoint, a lot of transit in America is designed for two types of people. Those who can’t afford a car, and those who commute to downtown 9 to 5. Taking that mindset becomes self-fulfilling. With barely adequate transit (long waits, indirect routing) those who can afford to, use a car.

    From a practical standpoint, it is worth noting that often these commuter lines tend to have low ridership outside of rush hour. There are exceptions (California being the one you mentioned). But you could run the trains all day in most of the United States, and it wouldn’t get many riders. For example, consider “Sounder South”, a fairly successful commuter rail line serving greater Seattle. It is peak-direction oriented, but they do run a few trains in the middle of the day and reverse peak. They perform very poorly: https://www.soundtransit.org/sites/default/files/documents/2020-service-implementation-plan.pdf#page=79. Instead of around 700 riders per train, they get less than 100.

    That could still be a decent value, depending on how the trains are managed. Usually, costs per trip go down the more you add trips, for the same reason as buses (https://humantransit.org/2017/08/basics-the-high-cost-of-peak-only-transit.html). But if the tracks are leased, it is bound to be the opposite. It is one thing to give up a couple hours in the morning and evening to passenger rail. But running trains 16 hours a day could put a real cramp in the system. Each additional run becomes more costly to the freight operators.

    There is no question that some cities could do a much better job in running their heavy rail lines. Even if they can’t be frequent in the middle of the day, they could be consistent (e. g. every hour). But just as with high speed rail, we shouldn’t delude ourselves into thinking that this is a solution that works for every American city, or even most of them. Nor should it be our first priority (this should: https://humantransit.org/2021/01/fixing-us-transit-requires-service-not-just-infrastructure.html).

  10. Donde Groovily May 22, 2021 at 1:13 pm #

    Except that there are physical problems with high frequency, and a big financial problem. A lot of these commuter lines run on active freight rail lines, so they need to be scheduled to work around all the freight trains. And the financial problem is that these are often privately owned freight lines, so more trains means more money to lease the tracks.

    In Seattle, the only commuter lines where Sound Transit owns the tracks is the line from Tacoma to Lakewood (which is also used for freight, but not a lot of it). The rest of the system runs on tracks owned by BNSF that have more than 50 freight trains every day.

  11. Carl May 22, 2021 at 5:15 pm #

    If you look at what Germany has done with its S-Bahn that should be a model for North America.

    Many of the routes were former freight and regional rail routes. Some of them have mixed service even today. In others they have added tracks adjacent to the right of way to expand capacity.

    They are almost all electrified, generally on 15-20 minute headways (outer branches/tails and Sundays sometimes less) and all EMUs or DMUs.

    Caltrain would be such an easy candidate for S-Bahn operation. Virtually no freight. Should have been electrified long ago. Who cares about Baby bullets – run the thing every 15 minutes all day long, every 7 minutes at peak, and lower the fares. Start with DMUs if you must, but electrify it asap.

    Crazy that the MBTA runs diesel trains under the wire between Boston and Providence – on what planet does that make any sense? Add stations and frequency and EMUs.

    Sadly Boston has a history of unforced errors. Not building a rail connection between North and South stations as part of the Big Dig has to be the top one. But they also had a legacy of trolley tunnels that they could have used. Still many corridors in Boston that can be used.

    Chicago electric should be an S-Bahn. Seattle-Tacoma should also be made into a S-Bahn, even if they need to add 1-2 tracks to do it and disentangle the freight. In most places the right of way has room

  12. david vartanoff May 24, 2021 at 8:40 pm #

    Further thoughts on BART and Caltrain. Currently Caltrain is running 30′ headways all day, BART is doing the same with a few extra trains in rush. Caltrain, as is traditional with commuter systems, sells passes, BART refuses to. Caltrain adopted the clipper card several years ago, and now honors clipper on a cell phone.

    • Carl May 25, 2021 at 4:41 pm #

      30 minute headways on BART are sad. Are they due to over generous union contracts? Given the capital investment made in that system, they ought to maximize operations to generate a return on that capital investment. Life is returning to normal.

    • Christopher Parker June 1, 2021 at 9:59 am #

      When I last paid attention a few years ago, BART was running only every 20 minutes during the day. Not frequent enough.

      A system like BAR would be revolutionized by automated running without drivers (it’s already most of the way there). Imagine breaking up a 10 car consist now running every 20 minutes so a single car now runs every 2 minutes. Since several lines run together, frequency would need to be a little less with 2 or 3 car trains.