Sunday, May 6, 2012

The problem with South Coast Rail

Readers of this blog will know that I am a very big supporter of public transportation and the general goal of urban accessibility without a car. With that in mind, I still believe that it is proper to devote resources to their best use, and the long planned South Coast Rail project is a very bad idea in its current form. The plan is to extend the Stoughton line south and establish separate branches to Fall River and New Bedford as commuter rail service. Unfortunately, the project is plagued by elementary planning errors and outlandish costs. It represents a waste of resources as well as a threat to other far more deserving projects.

Currently, there exists commuter bus service from New Bedford to Boston South Station. Scheduled bus times range from 75 to 110 minutes depending on time of day. Plain old driving directions put it at 70 min without traffic. By comparison, the South Coast Rail project has published a table of vital statistics about train service, and also a bus alternative, which I am going to analyze in this section:

The numbers I am drawing from the table are based on the fact that the Stoughton line has been selected for the route, and it is highly likely that the MBTA will continue to employ diesel locomotives, so I think it is fair to say that the projected travel time from New Bedford to South Station is 85 minutes. The Rapid Bus alternative considered is within the range of existing service, at 103 minutes. However, the planned headways for peak service are 40 minutes for rail and 15 minutes for bus. Waiting time is an important part of overall transit time, and it makes a big difference here. With these numbers, your average trip is going to take 105 minutes by commuter rail, and 110.5 minutes by rapid bus. The worst-case trip time is 125 minutes by rail, and 118 minutes by bus. For New Bedford commuters, South Coast Rail is a terrible value proposition. It may save them 5 minutes on average but will cost the state in excess of $1.4 billion.

Off-peak the situation is even worse for rail. The table shows 38 daily trips to/from New Bedford and Fall River combined, meaning 19 for each separately. Suppose you wanted to ride the train inbound from New Bedford off-peak. Of the possibly 10 trips, 3 are peak, so that leaves you with 7 departures to choose from. Those 7 trips must be distributed throughout the off-peak period, so let's generously presume that means 9am - midnight, or 15 hours. That means there is an average headway of nearly 130 minutes in between off-peak trains inbound from New Bedford.

On the other hand, there would be 218 Rapid Bus trips per day, so 109 for New Bedford, and let's say 55 inbound. Of those, 8 are peak, therefore 47 are off-peak. Having 47 trips in the 15 hour off-peak span means an average headway of 20 minutes. Also, judging by the existing bus schedule, the trip should be faster off-peak, maybe taking as little as 75 minutes. The buses would probably not be able to serve points in between as effectively, as they would have to exit and re-enter the highway for each stop. But the average trip time could be as little as 85 minutes, and the worst case off-peak would likely be 120 minutes. Compare that to the average off-peak rail trip of 150 minutes, and the worst case of 215 minutes, if you just miss the train. It simply can't compete against the bus, even when using the numbers provided by South Coast Rail planners. This table summarizes the analysis:

ModeTimeHeadwayDuration of travelAverage overall trip timeWorst-case overall trip time
RailPeak40 min85 min105 min125 min
BusPeak15 min103 min110.5 min118 min
RailOff-Peak130 min85 min150 min215 min
BusOff-Peak20 min75-103 min85-113 min120 min

The only category in which rail comes out ahead is average trip time during peak travel, and then, only by 5 minutes. The bus alternative is superior otherwise.

Here is the cost breakdown from the Plan For Action:

Action ItemCost
Economic Development and Land Use Planning $2,000,000
Public Outreach $1,000,000
Environmental Permitting $8,000,000
Project Design $75,500,000
Station Siting $27,600,000
Project Construction $1,074,500,000
Vehicle Procurement $163,100,000
South Station Capacity $31,400,000
Midday Layover Facility $52,100,000
Total Project Costs $1,435,200,000

The MBTA estimates that the annual operating cost for the South Coast Rail will be approximately $26 million (in 2017 dollars) and that the new annual revenue associated with the new rail line will be approximately $5 million. This will result in an annual shortfall of approximately $21 million that will need to be funded from other sources. 
These costs have sky-rocketed from previous estimates by the MPO. Those numbers look positively quaint, in fact, by claiming a construction cost of $670 million, even assuming that was given in 2004 dollars. And they expected 8,700 riders to use the thing? That's $77,000 in capital costs per customer, a pretty high amount, even by U.S. standards. New York's Second Avenue Subway is clocking in at about $25,000 per expected rider. And now with South Coast Rail expected to cost over $1.4 billion, that puts it at a ridiculous $160,000/rider. And we can't really trust the ridership projections of the MPO. Since that document was written, one of their projects was actually built: Greenbush Commuter Rail. How did their numbers pan out? Well, they claimed that it would cost $470 million to build Greenbush, and it would attract 11,400 riders, of which 4,600 were completely new to transit. In fact, it cost $534 million to build Greenbush, and ridership is approximately 3,000 per weekday, largely siphoned away from the ferries. It seems that the MPO was overly optimistic about Greenbush, and it is likely that they are overly optimistic about South Coast Rail, especially given the fundamental failures in its planning. A more recent projection of South Coast Rail ridership puts it at 4,070 riders for the diesel alternative, which equals an astounding $340,000 / rider in capital costs.

I have been assuming diesel locomotive service so far, and the $1.4 billion number cited above is for diesel trains. If they actually choose to electrify, as the EPA says they should, service will be improved somewhat, but then the cost will go up to over $1.8 billion. And that probably doesn't include all the changes the MBTA will have to make to accommodate electric locomotives. Although, if that is the spur which pushes them to electrify their other lines (besides Providence), that could be a good thing in the long run. But it would not be cheap up-front.

I've said a lot of negative things. Is there any positive side? Taunton, New Bedford and Fall River do seem to have sufficient density to support stations. It is possible that rail could stimulate development. But, the plans call for
[Estimating] future station parking demand for a period of 25 years following South Coast Rail’s completion date, incorporating a reserve of 15% to 20% additional capacity. Conduct a legal assessment on whether the Commonwealth can bank land for future parking demand and, if so, identify the mechanisms and parameters for banking. If land banking is implemented the South Coast Rail project team should design and landscape such banked properties, providing opportunities for passive open space.
It looks like the MBTA is just going to implement its usual one-size-fits-none plan of plopping down humongous parking lots at taxpayer expense and then subsidizing their operation. Serving such peak-direction-only riders and using parking lots to block off pedestrian access to the station will never produce much revenue; the system will be on life-support from day one and forever onward.

The plans for South Coast Rail are extremely dependent upon spurring future development in faraway places such as Freetown, and industrial highway wastelands such as Whale's Tooth. Although it is conceivable that such development would be a good thing, it is just words on a piece of paper right now. In the meantime, we have actual people and existing neighborhoods in places like Somerville that have been promised better service for decades. Finishing the Green Line extension may not sound as flashy as a new commuter rail extension, but it is a much more sound proposition from the standpoint of transportation economics. There is also a multi-billion dollar backlog in repairs and renovation which desperately needs to be tackled.

Is there a way to do SCR right? Maybe if they cut the costs severely. No lavish stations, just basic service. Build in stages, first to Taunton, with connecting buses, and then consider matters from there. The trains also have to be significantly faster in order to attract enough ridership. That means electrification; but the MBTA isn't about to electrify a single line and create unique maintenance headaches for itself. After all, they still run diesel trains on the Providence line. In the future, if the MBTA ever gets serious about rail service, they will need to electrify all their lines to improve performance and reduce operating costs. At that point, it may be worthwhile to look at providing no-nonsense service to Fall River and New Bedford with travel times of 60-70 minutes and reasonable headways throughout the day. Until then, it's just a losing proposition.


  1. I think part of the issue is that service from Boston to Fall River and New Bedford is not commuter service; it's intercity. Fall River is a suburb of Providence, and has a minimal number of Boston-bound commuters. New Bedford has a minimal number of commuters bound for both Boston and Providence. It would in principle make the most sense to connect Fall River with Providence, if only the legacy line weren't cut at multiple difficult locations. The upshot is that the service should be uniform throughout the day (and at worst hourly) without a strong peak, and also that at first hourly frequency is enough, since trips are less spontaneous at this distance.

    A plan that wanted to spend less money would just drop Fall River and New Bedford. There are more important places to build MBTA stations at - Union Square, Tufts, Allston, Brighton, Revere.

    A plan that wanted to spend more money to do it right would not have a line that splits at all. It would do what it takes to restore the connection to Newport and the connection across Providence's East Side, and have Fall River as a junction connecting on one side to New Bedford and Newport and on the other side to Boston and Providence.

    1. That's a good point, which I pretty much glossed over with my MA-centric focus. Not only that, but it could actually be quicker to get from Fall River to Boston via Providence. Amtrak only takes 30-40 minutes, so a bus connection could be faster than SCR. Schedules aren't currently conducive to a commute, but, for example, there's a Peter Pan bus from Fall River to PVD at 9:45 that arrives at 10:15, enough time to catch the 10:22 regional to BOS arriving at 11 or so. That's a 75 minute trip with a connection.

  2. The issue of rail vs bus is a tricky one. While you rightfully point out that headways for bus service would be much less than for rail, there is also the issue of quality of service. We've seen over and over again that rail attracts more "riders of choice" (i.e. people with easy access to cars) than buses do. Rail cars are roomier and more comfortable than buses. Rail service is also smoother and less jarring. Dealing with stop and go traffic as well as the twists and turns of getting off and on a highway to make intermediate stops on a bus is a far less comfortable experience than the smooth ride of a train, which stops only at intermediate stations directly along its path of travel. Furthermore, rail is far more likely to spur transit oriented development around the stations (as long as the MBTA and local municipalities take advantage of that opportunity). While I do think South Coast Rail is expensive, I think the rail option is better than the bus option overall. Perhaps in the future, headways could be improved over what is proposed. I do think it's very important that all TOD opportunities are taken advantage of to increase ridership as much as possible. And I also agree that transit within and between South Coast cities and Providence remains an important issue as well.

    1. I agree that rail rides more comfortably than buses. But I don't think that ride quality is worth $1.4 billion. Especially when it might produce worse average trip times. They should instead consider investing in some decent long distance commuter or intercity buses, which are considerably nicer to ride than city buses.

      As for TOD, I have almost zero faith in the MBTA and local municipalities in MA to get it right. The MBTA has a thing for subsidizing massive parking lots around every station, which discourages development around the station, and also discourages people from living near or walking to stations. And if people won't walk to the stations, they won't walk *from* the stations either, so there's little possibility of establishing two-way all-day travel, which is essential for real ridership gains.

    2. Matthew is understating just how far the MBTA commuter rail is from TOD. Plymouth, an old town, is not even served downtown: they cut the line just short of the town center, so the commuter rail station is in a peripheral part of town. Then they barely even serve that station - most trains divert to Kingston-Route 3, a middle-of-nowhere park-and-ride just outside Plymouth.

      The other question is why even spend scarce money on Fall River and New Bedford, money that the MBTA will then have to pay subsidies for and cut service to more productive parts of the system. If there's money for a service expansion, they should expand service within Boston and other places where the TOD already exists or is close, which means bus-competitive fares (and not the stupid, stupid Zone 1 fare) and schedules (and not 40-minute headways in the middle of Dorchester).

  3. The information is probably in one of the links you have posted, but do you happen to know what the projected cost would be for a round trip ticket from Boston to New Bedford on this planned line? Would it compare favorably with the cost of the same ticket for the existing bus line?

    1. That information is surprisingly well hidden, and I think it is subject to change, but the older plan was to have Zones 9-11 cover Taunton, Freetown and Fall River/New Bedford.

      Currently, you can purchase a one-way ticket to New Bedford for $13. A 10-ride pack ($80) for commuters comes out to $8 per-ride, and a 40-ride pack ($310) comes out to $7.75 per-ride.

      MBTA doesn't have a Zone 11 currently, but in July, Zone 9 will be $10.50/ride and Zone 10 will be $11/ride. So you could extrapolate that Zone 11 would be $11.50 until the next fare increase. Monthly passes will be $329, $345 and $360.

      Assuming 22 working days per month, that comes out to about $8.20 per-ride which is slightly higher than the existing bus, but also includes a subway pass, which makes it a better deal if you need to use the T.

      I have heard that the private bus companies which operate this route are profitable on it, but I don't have any solid evidence in hand. I don't know what kind of subsidies they may be receiving, other than the usual freeloading on highways.

  4. The issue with bus service is the same as the problem with driving from Fall River to Boston: it is hamstrung by traffic in the northernmost 10 miles of the route. If there was a coherent HOV system in Boston, it would be better, but there's only the few-miles-long zipper lane along the Southeast Expressway, and still ten miles of daily gridlock from the north end of Route 24 in to the city. So what should be an hour-long drive becomes, at rush hour, closer to two, and rush hour on the Expressway anymore stretches from about 5 a.m. to 11 a.m. and 2 p.m. to 7 p.m. In both directions. At ±70 minutes (for whatever reason, a top speed of 100 mph was assumed, when 110 is the class break), train service would be very time-competitive with driving. South of Easton, the corridors—especially to New Bedford, are arrow-straight, meaning that top-speed operation should be feasible; if built correctly faster travel times should be feasible. Buses max out at 65 mph, but have to deal with traffic, and there's not room to build a bus transitway easily in to Boston.

    Waiting time is important, but much less so on commuter rail than other modes. Not many of us plan subway rides based on the schedule, we just build in the headways in to the trip. If you have a 20 minute subway ride and a 12 minute headway, the average trip is 26 minutes, and the worst-case 32. But for commuter rail, it's different. Few people just show up at the station and hope the next train is leaving—there's a scheduled to check, and the trains tend to come on time (or at least they should on a newly-built system). So the average trip time is likely only 5 or 10 minutes longer than the actual travel time; it just requires more planning. And the T should have at 60 minute headways with clockface scheduling on this line: running a train of 2 EMUs is a lot cheaper than running a six car diesel set midday.

    Finally, as Alon mentioned, this should be seen as more of an intercity rail project than commuter rail, with more even headways and less focus on straight commuter service. Fall River is closer to Providence, but the more dynamic job market in Boston should be more of a draw, except that it's hard to access with the current congestion. As with a lot of transit projects, it comes down to jobs access and housing: if New Bedford and Fall River were reliably 70 minutes from South Station, they would be more desirable for a lot of people to live.

    And additional perk of intercity rail: the New Bedford service runs directly in to the ferry terminal there. From there, it's a 1 hour ride to Martha's Vineyard. The state stupidly build a bike path in the center of the rail line to Woods Hole, precluding its use to access the ferries there (that's a long story), but the Cape Flyer new connects people going to Nantucket by rail to Boston. Getting to the Vineyard can be a three-hour ordeal from Boston, with traffic getting on-Cape, parking, and then the ferry ride. A timed connection in New Bedford would provide much better service to the Vineyard from the Boston area, potentially adding quite a bit of off-peak travel. (cont.)

    1. So, about that 70ish minute trip time... I have some evidence from the project documentation that the actual trip times will be in the 80ish minute range from terminal to terminal. In addition, the preferred Stoughton alternative as planned could lead to a degradation in service for all North East Corridor users. I prepared a post on it but never got around to finishing it. I will look at it again though, since you brought it up.

      Just because people check schedule before departure doesn't mean that they wouldn't be helped by more frequent service. Unless you are fortunate enough that the train schedule exactly matches your desires, you will find yourself waiting for the train one way or another. Perhaps you can do that in the comfort of your own home or office because you are confident in the schedule. Real-time data has extended that sort of advantage to rapid transit and bus as well. But it's still a wait.

      I agree with Alon that treating it as an intercity connection makes more sense. If the commuter market was so strong then why can they only come up with 4,500 projected riders (almost all of them snatched from private bus services)? And that's with the 70ish minute trip time assumption -- which may not be true.

      And all of this really depends on having a reasonable price tag. Reactivation of a former passenger line for a few hundred million makes some sense. The same for over two billion dollars is outright insane -- especially at the same time that desperately needed maintenance goes underfunded in the existing system.

  5. (cont) Is this the top priority for the Boston area? It shouldn't be. But it is a good start for what should be a wider look at intercity rail in Massachusetts (which will see more service on the Greenfield-Springfield-Hartford-New Haven Line and potential service from Pittsfield to New York in the western part of the state. Boston to Springfield and Amherst (or beyond) should also be considered, especially since at peak periods the Turnpike has been gridlocked on weekends this summer, with hours-long delays between Boston and Springfield. If you could reliably get to Pittsfield in 2:30, you could probably entice trainloads of people to take a day trip to Tanglewood (especially if you had a bar car onboard); it's a 10 minute bus ride from the nearest rail station to the music shed (and there are plenty of BRTA buses which could be put to use shuttling people on weekends). Obviously, these are long-term ideas, but we—and the state—shouldn't view rail transit as solely a way to move people in and out of Boston at peak times.

    1. The former Boston & Albany has the advantage of being continually in use, unlike SCR. And yes, it's true that there's many problems and lots of work to do to make B&A a decent passenger line again. Much of that comes along with the work to make Worcester commuter rail suck less, to fix Beacon Park and to modernize the entire inside-Framingham line from its current decrepit status.

      Incremental upgrades for reasonable prices is the name of the game.

  6. I am not disagreeing with you conclusion that this may not be the highest priority project. But I think your travel time comparison is flawed. When headways are more than 20 minutes, people do not go and wait at the station without knowing the schedule. People arriving ~15 min before the scheduled departure. So your 215 min travel time is misleading. That said, a higher frequency is surely a benefit of bus service, that may well compensate for the other drawbacks of riding a bus.

    1. You are right that people are unlikely to wait at the station for low frequency service. However, the travel time comparison is still valid. If you are not waiting at the station, you might be waiting at home or waiting in Boston. You are rearranging your schedule around an infrequent commuter rail train. You are losing your freedom to set your own schedule. If you are fortunate enough that it all works out for you -- great. But if not, then it's likely that you will seek out an alternative; often by driving instead of riding.

      I know that the terrible off-peak & weekend commuter rail service frequencies are one of the reasons that many people choose to not bother with it. Knowing that you could be stuck at your destination for 2 extra hours, or worse, really does drive people to... well... drive.


Note: Only a member of this blog may post a comment.