What American Commuter Rail Can Learn From Paris


For years, transit advocates in New York and other U.S. cities with legacy rail systems have advocated for modernizing commuter rail to run more like a subway, often using Paris as an example. For good reason: Data suggests Parisian regional rail lines are well-used throughout the day, meaning the system helps people forgo driving for all types of trips, not just the trip to work.

Starting in the 1970s, Paris connected its suburban rail lines to form the RER network. Instead of ending in stubs in the central city, the RER runs as an express subway through the city proper. Using time-of-day trip data*, we can compare the RER to its American counterparts.

While the largest share of boardings in the Paris suburbs is in the morning peak, there is also substantial ridership in the afternoon and evening. In contrast, commuter rail in the U.S. is tilted much more heavily to peak hour ridership.

In Paris, I attempted to exclude all stations that do not serve originating commuters: the airport, the city center stations, stations dominated by transfers, and the big suburban job centers. Across the remaining stations, representing the region’s bedroom communities, 46 percent of weekday boardings are in the peak hours between 6 and 10 a.m.

On the LIRR, the corresponding figure, excluding city center stations, is 67 percent. On Metro-North, excluding job centers White Plains and Stamford, it is 69 percent (including them, it is 63 percent). On the MBTA, it is 81 percent.

U.S. commuter railroads are often full to seated capacity during peak hours but only carry a few people per car at other times. An entire off-peak MBTA train might have, at its fullest, 60 passengers. This is not the case in Paris, where midday RER trains routinely leave the city with every seat full.

There are two main reasons for this: service frequency and land use. This article focuses on frequency, with land use covered in a follow-up post.

Commuter rail service in the U.S. is simply not frequent enough throughout the day. Outside of rush hour, it’s common for trains to run once per hour in New York and Boston. Some MBTA lines have midday service gaps of two hours or longer. On LIRR branch lines, stations only get hourly service. Even on some LIRR and Metro-North trunk routes, service runs every 30 minutes.

This service pattern is not frequent or reliable enough for people to orient their lives around the trains, so instead, anyone who can afford a car in the suburbs served by these trains gets one, and drives it to all destinations except the urban core. To American regional rail operators to think less in terms of peak-hour commuters and more in terms of good overall service, especially for people who do not own cars (even if they travel to the rail station from another community).

Parisian commuter trains are oriented toward all-day ridership. The RER network is one of the most important models worldwide for running commuter rail like a subway, complete with an urban service pattern not too different from that of New York’s express subway lines. On the busiest RER trunk lines, the RER A and B, off-peak trains still come every four and five minutes, respectively.

Even RER branch lines, and some suburban rail branches not connected to the RER, have high off-peak frequency. Typical off-peak trains on these lines come every 15 minutes, rising to 10 minutes on two out of three RER A branches. A few come every 20 minutes, and fewer still come every 30. The trains are usually either on time or only two to three minutes late. (Only distant exurbs have hourly service, and there, ridership is indeed concentrated in peak hours.)

Moreover, off-peak schedules are predictable — trains always come the same number of minutes after the hour (known as clockface scheduling). For example, at Bures-sur-Yvette, a southern Paris suburb, inbound RER B trains leave at :03, :18, :33, and :48 after the hour between 10:33 and 3:48 every weekday. It’s easy for people to memorize the schedule, and if they live within walking distance of this or any other RER station, they can time their arrival at the platform to coincide with the train. In the U.S., Metro-North and some LIRR lines have off-peak clockface schedules (every hour, rarely every half hour), but on the MBTA only the Lowell Line does.

There’s growing emphasis on frequency among transit advocates in the U.S. This is because even rush hour ridership benefits from higher off-peak frequency: Office workers are more likely to ride the train when the trains will still be frequent at 7 or 8 p.m. if work runs late.

But off-peak frequency is the most useful for off-peak ridership, of course. This is why it is so significant that Paris gets especially high off-peak ridership. It suggests that if frequency on U.S. commuter rail improves, more people will choose to orient their lives around the better transit service. Some will choose not to own a car, while others will still own one but use it less than they do today.

Commuter rail operators may think that adding frequency would just result in running more empty trains, but the fact that off-peak RER trains are full suggests otherwise. If transit agencies choose to add more service, off-peak ridership is likely to rise disproportionately to the increase in frequency.

*A note of caution on the data: the Parisian numbers do not cover the entire commuter rail network. Paris has two different operators: RATP, which runs the Metro, most of RER A, and half of RER B; and SNCF, which runs the RER lines that RATP does not, as well as the more traditional Transilien lines. Only SNCF makes time-of-day data available to the public, based on one-day counts; RATP only has annual numbers. In the U.S., three big commuter railroads provide time-of-day data: the LIRR and Metro-North in New York, and the MBTA Commuter Rail in Boston. Of those, Metro-North and the MBTA only provide data on inbound boardings, though the MBTA also says it has very few outbound boardings in the suburbs.

36 thoughts on What American Commuter Rail Can Learn From Paris

  1. RATP lines are even busier. They have been by far way more successful. Based on what I have read, the RER A and B would be incredibly busy even during off peak hours (perhaps less with RER A now that metro line 14 has been built to alleviate its infamous overcrowding).

  2. The RATP lines are definitely busier, I just don’t have time-of-day data for them. I suspect they’d have an even lower percentage of traffic at rush hour, since they’re much more urban.

    I don’t know how it was before M14 opened, but the RER A isn’t that crowded. The three-door bilevels help a lot, and the trains are very long, about 225 meters (equivalent to 9 cars on the LIRR or Metro-North). The RER B is actually more crowded – it has single-level trains with narrow corridors and is limited to 20 tph because it shares a tunnel with the RER D.

  3. Meanwhile we get stuff like the current ACE train in the bay area: Rerouting freight trains due to the Oroville disaster means no ACE service for 2 days, because the freight outweighs the people. 6000 more cars per day.

  4. We should also remember that gasoline (and probably diesel fuel) cost about twice as much in France and other European countries as in the US. There are probably other factors making automotive travel less attractive over there. I once did an essay on why the automobile came to dominate local and intermediate distance travel in the US.


    One way to start a discussion about local passenger transportation is to consider why the personal automobile has become the dominant mode in the US. In a few places, such as Boston, New York City, Chicago and San Francisco, public transit plays an important role. In most other metro areas, getting people out of their cars and into a bus or train is an uphill battle.

    One cynic observed: Of course the automobile took over–look at all the wonderful human character traits it plays to: Impatience, selfishness and laziness. Impatience is a part of American culture; I once saw a personalized license plate that read H8 2 W8 (“hate to wait”). We tend to be in a hurry, packing our days with activity. “Road rage” resulting from this frantic lifestyle makes the headlines. One of the annoyances of using public transit systems is waiting for a bus or train. The observation that railroad stations have “waiting rooms” points up this fact. Compare this with the automobile, ready to go at a moment’s notice, any time, day or night. Then there’s selfishness. From time to time various state and local agencies tout the virtues of ride-sharing and/or van-pooling. They’re wonderful ways to cut fuel consumption, air pollution and road congestion, but many drivers are reluctant to give up the control represented by the steering wheel. When you’re in your own car, you choose the route. If you want to stop at your favorite hobby store on the way home, nobody will object. If you want to listen to music, it’s your choice from the radio or the stereo player. Is your car showroom fresh or a wastebasket with wheels? It’s your choice. The laziness aspect can be seen in several ways: ever watch someone drive around the mall parking lot looking for a parking space 25 feet closer to the door? Unless one lives next door to a bus stop or railway station, there’s a walk involved in getting to public transit, while the car is right out in the driveway or garage. And once you get to the destination stop, there may be more of a hike to the actual building. And there’s more to it than an aversion to walking: my wife worked for many years in Times Mirror Square at 1st & Broadway in LA. She usually drove and parked in the parking structure on 2nd St., which had direct access to the building. When her car was in the shop, she’d take the bus from San Gabriel (it stopped near the service station) and go downtown. The problem was getting from the bus stop to the TM building–dodging all the winos and weirdoes that infest that part of town. Yeeccchh!

    When I wrote this essay some years ago, I forgot another aspect of “automobility”: Vanity. One luxury import “marque” (to use the car snobs’ term) even uses “You are what you drive” as a slogan. If it wasn’t for the concept of “making a statement”, most of us would be driving Honda Accords, Toyota Camrys or similar cars that get the job done without attempting to impress relatives, friends and total strangers with your “coolness”. For family excursions and errands which require more cubic feet than a sedan, there’s the SUV vs. minivan question. Consumer Reports reviewed the Honda Odyssey recently, and started off saying: Sorry, image conscious parents. Despite the popularity of SUVs, none of them can match a minivan for overall versatility and practical family transportation. I suppose to some people, “image” is important. Motor vehicle ads on TV certainly play up to this part of the human psyche, telling all the guys in the audience that the sponsor’s product will attract more “hot chicks” than you’d know what to do with, or turning your shy, nerdy son into a “big man on campus” because you drive him to school in your new luxury car. Then there’s the athlete who finally makes the “big leagues” and signs a multi-million-dollar contract to hit, throw, carry or catch some kind of round or spheroidal object. It’s not true in all cases, but one of the first things many of these players do (especially if they came from a “financially challenged” background) is buy the biggest SUV or the sportiest coupe they can find. Alas, a sad number of them use more of the big bucks to go out and party, knowing they can cover any bar tabs, and proceed to run their fancy ride into something large and unyielding.

    An illustration of the “control” aspect of driving may be taken from a song that Roy Clark recorded back in the 60’s: “Right or Left at Oak Street”. One might call it a “Suburban and Western” song, it’s the sad tale of a man whose life isn’t truly wretched, but it’s not very happy. His wife is a nag and his children are brats, and everyone he knows seems to be doing much better financially and emotionally than he is. To borrow a song title from the Punk Rock era, “Life Sucks, Then You Die”. Every morning Monday through Friday he climbs into his beat-up old car and drives to work. Every morning he comes to Oak Street. Every morning he has to make a decision: one way will take him to his workplace for another eight hours of soul-deadening toil; the other leads to an Interstate on-ramp and ????? We don’t know. Johnny Cash recorded one possibility in “Understand Your Man”, in which the narrator hits the road. My guess is that Roy Clark’s subject will just keep plugging away in his rut until he finally keels over. Guys like him may gripe a lot, or stoically carry on, but they’re not likely to do anything daring. He probably knows that trying a road trip in his old car would be asking for trouble. Unless he had an easily marketable skill, he would think long and hard about leaving his job, no matter how dull it is. Unless he had friends or relatives with whom he could bunk down while sorting things out, he probably wouldn’t want to be paying for daily lodging, even at Motel 6 rates. Now, what does all this have to do with transit? Consider if our man was in a car pool–would he ask the driver of the day to “let me off at the Greyhound station” on the way to work? If he rode a bus or train, it would look suspicious if he walked to the stop or station with a suitcase. With a car it’s easier to hide–he could smuggle clothes and other travel items out to the garage over a period of weeks and stash them in an old box (his garage is probably such a mess that it’s easy to hide things). When he’s finally ready to make a break for it, he just has to move the box into the trunk and try not to act nervous (most wives have a “sixth sense” that can detect when husbands are “up to something”.) It may have an element of masochism, the decision at Oak Street every day, the thought that escape from his weary existence is there with the turn of a steering wheel, or maybe it’s all that keeps him sane.

    My first wife rode the local transit bus as a teenager in the 50’s (rail transit in our area being only a memory), but as soon as she had enough money saved up, she bought a car. She later told me that there was a man who used to ride the bus and harass girls and young women, and she’d had enough of that, plus the bus service was fairly infrequent and didn’t always go where she wanted to go. As far as I know, she never rode a bus again. This was back when one could buy a used car for about $100, fill it with gas at 25 cents a gallon, run it until something major failed and find another clunker. I later learned that this was part of the GM strategy: sell new cars on the GMAC installment plan, encourage the more prosperous motorists to trade in fairly often by making incremental improvements in performance and styling, and have an ever-widening stream of used cars traveling down the economic food-chain. Many years ago I had folks who had been to Europe saying, “You’d love it in Germany (or Switzerland, or Holland)”—trains go everywhere and most of the cities have streetcars and/or metros. This led me to visit the German Consulate General in San Francisco (on one of the cable car lines, for bonus points) and interview one of the consular officials, who explained some of the differences between Germany and the US regarding cars. Driver’s licensing is much more rigorous in Germany—none of the “Where did you get your drivers license? In a box of corn flakes?” that we have here. Instructors must be licensed, and their “alumni” are tracked—if too many of an instructor’s graduates are cited or get into accidents, his or her certificate can be revoked. Cars, too, are kept under much tighter scrutiny. They have to be inspected at specified intervals, a process that takes at least a day and maybe two. Even body damage has to be repaired before approval is given—no banged-in doors or primer splotches. Add-on equipment is also regulated—car owners can’t just go down to the custom shop and have chrome rims installed; they have to be authorized for that particular make and model. Imagine trying to register a “Low Rider” in Germany.

    Some transit advocates assert that the typical American got sold a big bill of goods by the auto makers, the oil companies and the housing developers. According to this school of thought, most of us have been “brainwashed” into thinking that “Real Americans” drive cars and live in single-family houses. Apartment dwellers and transit riders are “second-class citizens”. Cities are thought by many to be terrible places to raise children. Once again you have the issue of “control”. Part of our legal tradition is a man’s home is his castle. Or to quote a folk song parody: This land is my land, it isn’t your land, and you better get off, ‘fore I blow your head off. One “conspiracy theory” holds that after World War II, the federal government passed the GI Home Loan program to encourage veterans to buy single-family homes in the burgeoning suburban tracts. The idea was to disperse the “working class” into suburbs where they would be so busy maintaining houses, mowing lawns and taking care of cars that they wouldn’t have time to listen to radical “agitators”. (I don’t have a source for this “theory”, but over the years I’ve worked with and had classes with people with a wide variety of viewpoints, so I picked it up somewhere)

    One development that hurt both bus and trolley business was the advent of broadcast television. Instead of taking transit to the movie theater, folks would stay home and watch Uncle Miltie, Ed Sullivan or a whole host of other entertainment choices (even with only a half dozen channels to choose from).

    We can also look at how automobiles kept improving, while many transit operators kept pre-World War I streetcars in service, usually because they didn’t have the money or borrowing power to buy new ones. As late as 1949, Pacific Electric was still running cars that dated back to before the first Model “T” rolled out of a Ford plant. Autos, on the other hand went from open-air roadsters and touring cars, to all-weather sedans and coupes. Engines became more powerful, riding quality became smoother, tires were less likely to blow out, and features such as radios and heaters became popular. World War II put private cars on “hiatus”, but after a few years of pre-war carryover models, things started changing again. The 50’s saw automatic transmissions and V-8 engines in the “low priced” three brands. Air conditioning showed up in luxury makes, becoming common as the century wore on. Of course, by now the battle was over, and the “gas buggy” had won. Anyone who used public transit, outside of a few major cities, was considered a loser, and an object of pity. Then came the 70s, and a pair of “gas crunches”. Suddenly there wasn’t a filling station on every major corner (and lots of minor ones) dispensing gas at 25 cents a gallon. “Detroit iron gas guzzlers” didn’t look so impressive, but to this day, even when gasoline went over $4.00 a gallon for a while the vast majority of local transport is provided by individually owned motor vehicles.

    I’ve often thought that one of the selling points for individually owned and operated cars can be expressed in Huey Long’s slogan, “Every Man a King”. And the king doesn’t ride in a bus or streetcar with the commoners*. According to one author, when the economy improved to the point where even the “colored people” in the South could afford, if nothing else, a fourth-hand Flivver, some of the “white folks” were upset because their legislators couldn’t figure out how to apply “Jim Crow” segregation laws to the streets and highways. Another story from the same era concerns the 1940 movie version of John Steinbeck’s Grapes of Wrath. The report was that as soon as this tale of the downtrodden “Okies” leaving their ruined and foreclosed farms in Oklahoma and migrating to the supposed “promised land” of California in their caravans of wheezy old jalopies hit the screen, the Soviet film bureau got a print and ran off bootleg copies, presumably with Cyrillic subtitles. These prints were distributed to cinemas throughout The Motherland, but were soon withdrawn when the “buzz” reached the film authorities’ ears. The audiences were more interested in the fact that in “Amerika” even the “peasants” had motorcars, than in the miseries of the migrants. Owning an automobile was far beyond the wildest dreams of the average Ivan and Olga. The only kind of motor vehicle anyone outside the higher echelons of the Communist Party could expect to drive would be a tractor, or maybe a truck in a Red Army convoy.

    *As I recall the King in one of Europe’s constitutional monarchies was photographed riding a bicycle some years ago.

  5. It does! But I’m on purpose looking at ridership in the suburbs rather than in the city proper. Since I only have data for the SNCF side, the only city stations I have are either central stations with more afternoon than morning boardings, which are excluded from the analysis, or some RER C branch stations with not much ridership. The major city neighborhood stations are all on the RATP side: Nation, Etoile, and the Left Bank RER B stations.

  6. “If transit agencies choose to add more service, off-peak ridership is
    likely to rise disproportionately to the increase in frequency.”

    If and only if suburban commuter rail stations are redeveloped. Many (most?) have nothing within walking distance. I look forward to your discussion on surrounding land use, but I imagine that until the land use side is dealt with, adding frequency *will* just result in running more empty trains (at least for those stations located outside the city limits).

  7. Well, yes! But, at the same time, #notallstations. On Metro-North, there’s a fair amount of development close to New Rochelle, Yonkers, Mount Vernon, and Fleetwood; they still don’t get much off-peak usage. On the LIRR, it’s better in the sense that stations with decent development around them, like Rockville Centre and Port Washington, are sub-60%, and Mineola and Long Beach are (barely) sub-50%. It helps that some of these lines have half-hourly off-peak service, and Mineola especially has decent off-peak frequency. It’s still nothing close to the 46% average over all bedroom communities in Ile-de-France, with many town centers at 30-35%.

  8. The RER trunk line portion of the system are also very useful for quick cross-city or center city to “ring city” trips, which is why the trunk lines are crowded. For NYC, this is the equivalent of being able to:
    -take the LIRR to Flushing instead of the 7 train
    -take MNR from grand central to fordham instead of the 4 to the D
    -take non-stop regional rail from Penn Station to Queens Plaza instead of the E or to Port Morris instead of the E to the 6.
    Because dense jobs centers are built on the outer ring RER hub stations, these kinds of trips are very popular mid-day.

  9. Speed also helps. Many commuter rail services in the US are incredibly slow. Even in Southern California’s famously hellish traffic that grinds highways to slogging stop and go, driving often still beats the commuter Metrolink services by 2:1 and that assumes you are going station to station.

  10. “Because dense jobs centers are built on the outer ring RER hub stations”

    …The author hints that more concentrated TOD is likely also a major contributing factor in attracting heavy ridership on the RER (and will be a topic on his next post). But that begs the question, does higher frequency service only happen once more TOD has happened or will TOD only happen once higher frequency service is established? Chicken or egg?

    I’ve long thought the MBTA in Boston (and state govt in general) has done a terrible job of promoting TOD opportunities around its stations.

  11. It’s by far the former. Poorly situated commuter rail stations with thousands of park-and-ride spots gobbling up valuable land obscures the land value and potential to build densely around the station, because even though commuter rail may reach fairly dense towns with diverse trip profiles, they are still designed to serve single-family commuter subdivisions surrounding the town, miles out.

  12. And white plains and stamford and greenwich and new haven…. okay probably a couple more but those are the other biggies.

    One other issue I see is that even outbound AM rush hour trips are charged as Peak fares – though the trains are 1/2 empty. Maybe 1/3rd empty.

  13. one odd thing about Paris is its metro covers an unusually confined area.
    Yes, stations are very very dense in Paris and sometimes , just across the ring highway – but they do not (yet) tend to extend very far outside of Paris (with a couple exceptions). So many areas that are just as far from Chatlet Les Halles as Jamaica Queens or Southern Brooklyn is from Midtown are not served by the Metro at all (yet!), but are only served by regional rail. Hence a lower Metro ridership and a higher RER / Transilien ridership with greater, metro-like frequencies.
    Just another contributing factor in Paris’ high frequency regional rail system.

  14. Sometimes, they’re built simultaneously. The busiest branch of the RER A, going east to Marne-la-Vallee, is entirely new: the first segment opened simultaneously with the central RER A tunnel, and it’s since been extended farther east. The newer parts of the branch were built simultaneously with the plan designating it as a suburban job growth region, with office parks close to some of the train stations and Eurodisney at the eastern end. One of the western branches of the RER A, to Cergy, is the same: the edge city was built simultaneously with the rail link serving it.

  15. Glad to see Alon writing for a much bigger audience. Another good example would be Taipei’s commuter lines, although Asian systems are often less explored due to the language barrier (Japan, South Korea, Taiwan).

  16. This is great, and I can certainly agree. It would be great if this article (and others similar to this topic), could talk more about SEPTA Regional Rail, and the challenges it has to do something quite like this. Clockface headways exist, though only running every 30 to 60 minutes during off-peak headways. Thanks to the Center City Commuter Rail Tunnel, built in 1984, trains can now pass through the Urban Center of Center City, Philadelphia, rather than just terminate there. Most off-peak trains travel through to another branch.

    Local Transportation Expert Vulcan Vuchic studied the potential of making the Regional Rail network more like a “Metrorail”: https://trid.trb.org/view.aspx?id=572810

    For some time, SEPTA Regional Rail lines were numbered, with an “R”, but that was eliminated due to confusion and the fact that lines ended up “transferring” to other lines as a rule, rather than the exception (an R6 train from Norristown would never continue to the R6 Cynwyd Line, but would continue as an R2 to Newark instead).

    Honestly, out of all of the Commuter Rail systems in the country that exist today, SEPTA Regional Rail is the closest to operating in the manner recommended in this article. It’s just too bad there were unfunded mandates (Positive Train Control), crumbling infrastructure, a long period of time where there was a complete lack of dedicated funding, and other operational issues that have prevented SEPTA from moving forward with the “Metrorail” idea.

  17. Do they have conductors on commuter trains in Paris? In the Baltimore and DC area, while heavy rail and light rail do not have conductors (having fare gates and proof of payment), the commuter rail (MARC and VRE) all has one or more conductors per train walking up and down. That is a significant marginal cost for running an extra off peak train that has to dissuade running them. The MARC could check tickets at major stations like Union Station and ignore fare jumpers riding, e.g., Odenton to Bowie State and in exchange run more trains.

  18. While I don’t expect SEPTA’s experience to be quite as applicable on the national level, there is far too few critical analysis of its service and potential among local urbanist groups and governmental legislatures. If someone could tell me the last Delaware Valley politician to actively campaign for a service improvement outside SEPTA’s own volition, I would gladly contact them. But there’s not a single one in recent memory that really cares like politicians in other cities do. There’s a certain disconnect in how much riders and politicians are personally invested into the system because nearly 100% of SEPTA’s budget comes from the state and feds, and they’ve gone for decades without a local contribution to balance things out. Moreover, SEPTA’s regressive actions in areas like fare implementation go unchecked, and they’re planning to entrench conductors in their roles as cash dispensaries and trapdoor openers forever because they’re too lazy to maintain 50 more TVMs at outer stations. Like most agencies, SEPTA displays a certain arrogance whenever someone like Vuchic proposes service improvements, and gives a laundry list of excuses not to go for it. So whenever they do try and propose something from the SEPTA-DVRPC echo chamber, like the overkill Schuylkill Valley Metro or trivial Roosevelt Blvd pseudo-BRT, it’s always horribly misinformed and misprioritized.

    For reference, the full Metrorail paper is here. It’s a very important yet ignored prototype of how frequent service could be implemented on the cheap for legacy systems.

  19. BART and the Washington Metro are examples of heavy rail systems that act like commuter rail systems. Of course, they have very high capital costs and are facing mid-life crises in terms of infrastructural renewal. However, they brought several innovations being “newer” systems: relatively frequent (15-20 minute) off-peak service and one-person train operation. While legacy commuter rail systems like LIRR and Metro-North generally don’t have equivalent “station agents”, they do have unnecessarily high numbers of assistant conductors, which means the marginal cost of providing additional trains is extremely expensive.

    The other problem with existing commuter rail in the United States and Canada is that there are only limited numbers of central stations (e.g., Penn and Grand Central, Washington Union Station). In some cases, those stations aren’t quite in the right locations (San Francisco 4th & King, San Jose Diridon and Oakland Jack London Square). So then there’s the issue of capacity mismatch, uncoordinated fares and outright inconvenience to connecting services. Short of billions and billions of dollars in investments, a key short term fix would be to coordinate fares so people don’t have to pay extra for the privilege of transferring.

    Pricing is a huge issue. In the reverse peak direction, for example, Metro-North charges peak fares during the AM rush hour. So the 18-mile trip from Harlem-125th St to White Plains is $11.75, or $259 monthly (+$116.50 for New York City Transit = $375.50) – and this is for a transportation network under one umbrella. Even off peak fares at $8.75 are hefty. This can be cost-prohibitive even for someone in the middle class.
    Finally, suburban bus connections are much poorer than in European countries. Much of suburban Chicago, for example, is a transit desert during off-peak times. So you have a Metra train coming every 1 to 2 hours, connecting to either a non-existing Pace bus or one that might run every 40-60 minutes. NiceBus in Nassau County (LIRR) is not much better, having cut back service due to insufficient county funding. Contrast that with Paris, where suburban bus and tram services are much more robust.

  20. Thanks for this, Alon!

    Just for giggles, I computed the morning-rush proportion of ridership for the PATH system. It’s around 56% or so. That means that a *rapid transit* system in NYC has lower off-peak ridership than a commuter rail system in Paris! I guess it doesn’t help that the PA runs trains only once every 20 minutes on some routes midday, and every 35 minutes (with no clockface scheduling) in the overnight hours. It’s pathetic.

    Do you have similar numbers for the MTA for comparison?

  21. In White Plains nearly all of the residential development is more than a 10-minute walk from the train station. And that walk is a miserable one, across 100-foot-wide one-way streets with no street-level retail and windswept plazas.

    Stamford looks quite similar development-wise. Greenwich is very small. I don’t know about New Haven.

    You did miss Port Chester, though, which has a walkable core around the station (though it’s almost as small as Greenwich).

  22. I know its mostly office and retail development in DT WP.
    But there are several residential towers there. But yes, the arterials are overly wide there, the bus station is a (needless) block away from the train station, and there are all those parking lots RIGHT IN FRONT of the train station that could be redeveloped into something mixed use…. I mean, in a sane country that is.

  23. I don’t! I wish I did. Where are you PATH numbers from? I’ve seen PATH entries per station but not by time of day.

    My guess is that PATH is just horrifically commute-to-Manhattan oriented. It doesn’t serve enough of Jersey City to really get local workers to jobs at Exchange Place, Journal Square, or Newport. And evidently, there aren’t a lot of reverse-commuters into Newark, or else Uptown Tube trains would run to Newark rather than short-turning at Journal Square. It probably also has about zero reverse-commuting.

    Most Transilien lines do a lot more than just getting people to Paris. Part of it is La Defense, or other suburban job centers, which are excluded from the analysis. But another part is people working in suburbs that are not really job centers, but still get 20-25% of their boardings in the afternoon rush, like Aulnay-sous-Bois, Sartrouville, Melun, or Argenteuil.

    By the way, the nighttime service frequency here is zero. Trains shut down overnight.

  24. So much this. Connecting bus service is not planned in coordination with trains in the US; I’ve heard some horror stories specifically out of Boston.

    I don’t know to what extent Transilien relies on connecting buses with timed transfers. I know that Swiss planning relies on them extensively, but I’m less familiar with suburban bus planning here. The suburban buses here get very high ridership, so clearly they’re not like Pace or Bee-Line or Nice, but I don’t know if they’re RER feeders or just strong corridors in their own right. The trams are the latter, with tons of ridership that’s clearly not just connecting to the RER. By analogy, crosstown bus routes like Western in Chicago, Fordham in the Bronx, and the 66 in Boston intersect a lot of subway/L lines, but it’s not really appropriate to call them subway/L feeders, since they do much more.

  25. There are no conductors here. The city stations have faregates like the Metro (and like American subways), the suburban stations are barrier-free but have proof-of-payment checks (like American light rail lines). The result is that the RER can run frequently, and charge the same fare as the Metro regionwide, and keep the farebox recovery ratio to the same levels as the workhorse American systems.

  26. …do you have time-of-day data for Taipei? (Also, in Taipei, is this commuter rail, or the MRT?)

    In Tokyo I know how to get ridership per station and train schedules, but I have no clue what the ridership is by time of day.

  27. Excellent work, as always.

    At the other extreme, cities with a lifeline bus system used predominantly by the under- or unemployed and those unable to drive also have very flat ridership/time-of-day profiles. Given that Paris has higher unemployment rates than comparable US cities and a greater prevalence of part-time work, I would expect that some portion of the difference can be explained by working differences outside the control of the transit company (without evidence, I’d guess about 10-25% of the difference).

    This suggests that cities could be grouped into three rough stages of transit ridership.
    1. Lifeline – Flat time-of-day profile because the few who ride use the system for most of their trips.
    2. Commuter – Peaky time-of-day profile, with lifeline riders from stage 1, but attractive to CBD-bound workers who ride peak services but generally do not ride for other trips.
    3. Full-service – Flat time-of-day profile, because the system is competitive to a broad range of customers for most of their trips.

  28. The prime age employment to population ratio in France is higher than in the US. Lower overall employment/population comes from earlier retirement (retirement age here is 60) and students not working (tuition is 500 euros a semester). There are more job seekers here than in the US, but fewer people who dropped out of the labor force entirely.

    I bring this up because Paris really needs to be compared with New York in wages and economic output, and not with places in the US that have 8% unemployment or 50% employment/population.

    I also don’t know to what extent there is really more part-time work here than in the US. Germany and the Netherlands are full of part-timers, but my understanding is that people work fewer days per week for family reasons, rather than fewer hours per day.

  29. In Taipei I’m referring to the legacy TRA service, specifically the through running from Banqiao to Nangang along the east west main line. There is no “commuter rail” in Taiwan but regular rail lines used by commuters (in addition to HSR).

  30. Fist of all I really love to read your blogs. They are both simple, but challenging at the same time.I want slightly deepen the time table of the commuter rail roads. It is not that simple as it looks. Prior to make or even to commit to any schedule the transit agency has to make a deed research on the travel pattern. Recently the MTA has conducted an interesting study similar to what I am talking about https://youtu.be/HJ0seVq85Lc (fast forward to 43:50). This report says almost nothing about per-station demand and this information is extremely important to run trains effectively, frequently with overall good service as you said.

  31. I watched 4 minutes of this, and maybe it’s addressed elsewhere, but do they at all consider how biased the surveys are? People who travel off-peak and drive aren’t included, people who work outside Manhattan and drive aren’t included, people who travel off-peak but chose to live in the city and not in the suburbs because they didn’t want to rely on off-peak commuter rail aren’t included. Best industry practice is to integrate planning of rolling stock, infrastructure, and schedules, rather than to plan each in isolation.

  32. I agree that they should have an integrated planning, and no off peak details, I guess because they were surprised themselves. The research does not provide a good reference for future planning for LIRR itself, however should ring some bell at Suffolk and Nassau transit professionals.

  33. It was much more crowded than now because it was all single-level trains, and the MI84 did’nt even extended the whole lenght of the platforms. Yet it does carry more people today, about 1,2 millions daily, about 300 millions a year, and in peak, the busiest direction (east-west in the morning, opposite in the evening) see more than 60 000 passengers an hour!

    RER B carries 900 000 passengers daily and it’s southern portion date back to 1937. And the RER D about 600 000 even if it’s limited to 12 trains/hour in the shared tunnel between Chatelet and Gare du Nord.

Leave a Reply

Your email address will not be published. Required fields are marked *


At Bourg-la-Reine, outside Paris, the rail station is surrounded by dense, mixed-use development and walkable streets. Image: Google Maps

What American Commuter Rail Can Learn From Paris, Part 2

In Europe it's common for regional rail systems to get ridership comparable to that of the subway in the central city. But in America, this is unheard of. One reason for the discrepancy is land use: American commuter rail stations are typically surrounded by parking, while in the Paris region you see a different pattern with ample development next to suburban train stations.

What If “Commuter Rail” Was for Everyone, Not Just 9-to-5 Commuters?

Rhode Island has been investing in commuter rail — long distance service connecting Providence to Boston and towns in between. But lackluster ridership at a new park-and-ride rail station at the end of the line (by a Walmart!) is sapping support for much more useful investments, reports Sandy Johnston at Itinerant Urbanist. Anti-rail critics are piling on. The libertarian Rhode Island Center […]