Wednesday, December 4, 2013

MBTA still won't provide real all-night transit

The MBTA recently announced a pilot program to extend service hours on Friday and Saturday nights. This is a great step for the T, which hasn't had subway service after midnight in over 50 years. While it does reduce the amount of time the T will have to inspect and fix its aging infrastructure, the other five nights will still have time for track work, while serving the busiest nights for late-night ridership. This is all good.

What Boston will still lack is an actual late-night transit system.

In every other major transit city in the country (More than 25% mode share: NY, Chicago, SF, DC, Philly or more than 1m daily ridership: LA), there is an option for getting around between midnight and 5 a.m. Not just on weekends. Not just until 3. (DC is a slight exception, some overnight bus routes shut down for 30 to 120 minutes around 3:00; but some routes, like the 70, have no more than an hour gap in service.) Service may only come every hour, and it may be a bus instead of a train, but if you need to get somewhere at 2 a.m., you might be able to take a bus there. This is not the case in Boston, and even with added Friday and Saturday night service, it still won't be. Several smaller cities operate all-night transit systems as well.

Late night transit service is going to cost money, but depending on how it is structured it can serve two economic purposes:

  1. "Cosmetic" late-night service (such as what is being proposed by the T) makes Boston more competitive compared to other cities. Apparently, workers are more likely to want to live in Boston because if they work late on a Friday night and have a couple of drinks, they don't have to worry about the train shutting down an hour before last call. (This Globe article makes that point.) By calling it "cosmetic" I don't mean to make light of this, as moving Boston towards being more of a 24-hour city is a laudable goal, and pushing train service later on weekends is certainly a good move. DC operates similarly; the Metro there stays open until 3 on weekends while shutting down around midnight most nights, but they manage to run a few popular bus lines all night, or close to it. And, yes, the workers behind the nightlife will be able to finish their shifts and catch the last train instead of ponying up for a taxicab or driving.
  2. Full late-night service, however, serves a much wider economic purpose: it provides access to employment centers which would otherwise be inaccessible during certain hours. This page has argued this point in the past, making the case that the MBTA or Massport should fund and provide at least a low level of 24-hour service to the airport, where many shifts begin or end between midnight and 5 a.m. But there are other overnight workers too. Overnight service on major transit routes, even if it were only every hour, would provide job access, especially for low-income service workers who spend an inordinate portion of their income for driving and (at the airport especially) parking costs.
How does the rest of the country do it? Because I love charts, here are some charts. I'll explain more below:

Obviously, New York skews this whole chart. It has more than four times the transit ridership of any other city, and 138 all-night routes (18 subway lines, the PATH, the Staten Island RR, 16 bus routes in Manhattan, 29 in Brooklyn, 47 in Queens, 17 in the Bronx and 8 on Staten Island). Let's remove New York. Here's the same chart, sorted by transit ridership, with the number of overnight routes highlighted:

Notice something conspicuously absent in the chart for Boston? It's the only city with high transit ridership without overnight service. The next largest? Atlanta, which is not known for it's transit friendliness (i.e. we're not moving the Red Sox to Danvers) and has a third the daily ridership of the MBTA. And several smaller cities have pretty comprehensive late-night transit systems. Las Vegas makes sense; the city basically operates 24/7. And last call in Miami is 5 a.m. But Cleveland and Baltimore? They're not what we think of as 24-hour cities. Yet they provide overnight transportation.

So the question is: is Boston going to put a cosmetic "hey look the trains run late filled with drunk people" band-aid on the situation? Or are we going to actually have a discussion of how to provide 24-hour transportation for citizens, and to jobs?

(There's a longer history of cuts to overnight service. Many cities, including Boston, had owl service in the 40s and 50s. In quite a few cases it survived longer. For example, Portland Oregon cut overnight service in 1986, and Minneapolis had a more extensive network until 1998.)

(By the way, if anyone knows of any other cities with overnight service—other than Newark which runs the 62 all night to EWR, I know about it—let me know. I was surprised that I could find neither a list of transit agencies by overall ridership nor a list of cities with overnight bus service. If anyone wants to help fill out these lists, I'm all ears.)


  1. Don't forget that Boston had the "Night Owl" service in the early 2000s where they ran buses along the subway lines after the T stopped. It was poorly implemented because it stopped at 2AM, precisely when people would have used it. I think they shut it down due to low ridership, which every college student saw coming from a mile away.

    The thing is that Boston has no businesses open late. As long as that's true, there's no reason to want to be out late. It's impossible to find a slice of pizza, cup of coffee, or a diner meal late at night, with just a few exceptions. As long as that's true, just having 24 hour transportation won't change anything. The city needs more businesses open later for this to be a success. We'll see if they allow it.

  2. In the Bay Area, in addition to Muni's rather comprehensive if infrequent all-night network, AC Transit runs a few routes in the East Bay and one across the bay, mostly paralleling BART lines, which have no overnight service (and in fact stop running earlier than even the T's subways). On the Peninsula, SamTrans and the VTA each run one route all night, roughly along the El Camino corridor, with a somewhat coordinated transfer in Palo Alto. Also, for what it's worth, the Bay Area is not really big on late night businesses either. Hard to say if it's better or worse than Boston, but it's certainly nowhere near NYC in that regard.

  3. @Wass, I've argued before that the Night Owl was a "failure" because there is no reason to subsidize the target demographic—drunk college students. When there's a funding shortfall, do we want to cut service on bus routes, or cut service for drunk college kids? Easy choice. They are framing it much better this time, by making the case that it makes the city more economically competitive. And running trains late, when they'll be better patronized since they're much faster than buses trying to follow train routes and people know where they go, will increase ridership further. Since the main cost of operating service is paying for staffing, it won't take appreciably more staff to run trains than to run buses. So it won't cost too much more.

    It will also allow people working at businesses open until 2 a.m. to finish their shift and grab the last train home. For a lot of lower income folks working in bars and restaurants, a $2 train ride rather than a $30 cab ride (if you can even find a cab) will be a major economic help.

    But 24 hour transportation is not really used to create nightlife; it's used to serve the people who need transportation then anyway. I've linked above to the post I wrote a while back regarding service to the airport. But it's astounding that it's nearly impossible to get to the airport by 6 a.m., when most airport shifts start at 4 or 5. The airport employs 20,000 people, and the fact that there's not at least 24 hour service from downtown to the airport means that anyone who works an early shift there needs a car. Even running the silver line 24 hours—on the surface to South Station—would allow people to commute without a car, or a wetsuit.

    Finally, a next step might be to run the service all night long on Friday and Saturday nights. With service ending around 3, the latest trips will probably reach their termini after 3; say, 3:30. The first trains leave outlying terminals like Riverside and Braintree right around 5:00 (4:55 from Riverside). Running half hour service from 3 to 5 would fill that gap, and they're not going to do any trackwork in 90 minutes.

    @Anon, I actually accounted for this in a couple of ways. I counted all of those lines, the Samtrans, VTA and AC Transit lines as well as the MUNI ones. They have the best resource for finding information of any agency as well, here (I have no idea why no one else has a good map or description of overnight service). I also used the CMSA population for the Bay Area, which includes San Jose, instead of the MSA, since service extended in to the San Jose area. Thanks for the tip, though.

  4. Note: there is early morning airport transit aimed at workers. Several bus routes have an early morning version which extends to Haymarket, and runs 1 or 2 times. There is also a route from Haymarket to the airport, and one from Dudley to the airport.

    Extensions of this system may prove to be the easiest way to create all-night service.

    Look up bus route "193" on for instance, that's the early morning version of the 57 bus. Or read the fine print on schedules like the 117 to see the extra trips. There's also routes like the 171 which combine a couple normal routes to provide early morning service. The buses that are out on the street are signed with the number of whatever route they are currently running along, so you won't see "171" on a rollsign but it's named that in the system.

  5. @Matthew: There are these routes, and they serve a very small number of people and are not publicized. I mentioned them in a post a couple years back. They also don't necessarily get people to the airport in time for those early shifts. If the T had its druthers, it would make sense for them to publicize these schedules and run an early-AM transit service from Haymarket with that early service to the airport. It's much cheaper than taxicabs, even for (cheap) travelers (like me). I've still never actually taken one of these early buses, but I do know of them.

  6. Greetings,

    As both a transit operations planner in Las Vegas and a Master's student researching the ridership effects of span of service changes, I find this to be fascinating subject, and I tend to agree with your categorization of the two types of overnight services; operating late night or overnight on Friday and Saturday night cannot possibly be used for dependable employment purposes.

    Anyway, the point of my post is to point out what I believe is the smallest city in the United States (dare I say, perhaps even the world?) with 24-hour bus service: Laughlin, NV, population 7,300. I have actually analyzed the ridership by hour of Laughlin's 777 route, and I was pleasantly surprised to find decent ridership -I believe the minimum of around 9 boardings per hour at 3 AM, but even 1 AM was in the mid teens. Overall, the two fixed routes in Laughlin experience relatively very strong ridership: around 1,000 boardings per day. The very high mode share is likely due to the service-sector, 24 hour nature of the economy (casinos) and the fact that the residential area of Laughlin is about three or four miles away from the employment area (the casinos), meaning walking is not a viable option (especially considering that it is one of the hottest cities in North America -hotter than Phoenix).

    Speaking of Phoenix, did you know all bus service there between 10 PM and 5 AM was eliminated in late 2008 amid draconian budget cuts? Phoenix, just slightly smaller than Boston, ends all transit service (other than the one light rail line) at 10 PM!

  7. Roughly speaking, late-night transit usage can be divided into three trip purposes; employment, recreation, or people in seek of shelter.

    In general, not many shift changes occur between Midnight and 6 am, with the notable exceptions of shift ends in the hospitality and restaurant industries until approximately 2-3 am, and shift starts in transportation and related industries beginning at approximately 4 am. Absent any other noticeable industry influences (i.e. one-industry towns), any shift patterns not listed above are not common enough to be the basis of economical transit planning.

    Recreational customers are essentially identical to hospitality and service employees in their trip patterns, but leave slightly earlier (the employees have to remain to clean up after their customers go home). This market segment is more likely travelling on Friday and Saturday nights (early morning Saturday and Sunday), as well as on the eves of certain holidays, and conversely are less likely to travel on nights preceding a work or school day. The above two groups share the same common travel characteristic that they are moving away from destinations in the very early hours, with that flow decreasing as a flow of customers towards destinations increases closer to daytime.

    In many cases, all night service attracts people in seek of shelter due to homelessness, domestic abuse, etc. These customers tend to favor long, uninterrupted round-trip runs where they are able to sleep most effectively, essentially turning the bus into a roving shelter. Route 22 in San Jose has been referred to as "Hotel 22" due to the clientele; customers are riding because the bus is a shelter, not because the bus is transportation (for that specific trip). As heartbreaking it is to see the pictures of the 10-year-old girl slouched over bus seats, as well as read the stories of the various people in seek of shelter, and as important as addressing homelessness is as a public policy initiative, the practical reality is that transit firms are not, nor should not, be in the business of combating social issues. Transit firms are in no position to offer the services homeless or temporarily dislocated people need, and from a public policy perspective should make no effort to present any image in the collective public mindset that they are a shelter resource for homelessness.

    Technically not customers as of now, the fourth market segment of note are people who would use transit service for other purposes. Extrapolating my personal habits; most weeknights I'm in bed inclusive of at least 1 to 5 am; the bus line by my house could operate every 5 minutes, once an hour, or not at all and it would not affect my travel habits; if a tree falls in the woods and nobody is around to hear it.... For people with no car availability, inelastic but rare demand for immediate night travel is best met with taxicabs and similar services.

    When the above is taken into account the "symbolic" service offered by the MBTA on a 15-20 minute headway, but only until 3 am is likely superior to service every 30-60 minutes but at all times. More potential customers are in a position to use frequent service to return home than infrequent service offered at times they have no need to travel. When these trips reach the outbound end of the line, if they are not needed for a later outbound trip they can be sent to the garage.

    At the same time, the early morning routes tailored to Logan Airport are able to provide superior service to customers and travelers who need to arrive very early. Instead of local-stop service with a connection, customers are able to enjoy express service from a central neighborhood station or park and ride directly to their destination.


  8. This is an awesome post. Really very informative and creative contents.

  9. Thanks for sharing valuable information.