In North America, commuter trains run with conductors, often several per train. On most systems they walk the entire length of the train to check every passenger’s ticket, whereas on a few, namely in California, they do not do that anymore, but there are nonetheless multiple conductors per train. In addition, the scheduling is quite inefficient, in that train drivers do not work many revenue hours. I investigated what effect this has on operating costs, and it turns out that the effect on the marginal operating costs, which are important for off-peak service, is large: on the LIRR and Metro-North, nearly fivefold improvements in revenue train-hours per on-board employee (driver or conductor) are possible, which would halve the marginal operating cost per train-km. The bulk of this post is dedicated to explaining the following breakdown of variable operating costs:
The National Transit Database has figures for service in car-km and car-hours for a variety of US transit agencies. In New York State, the Empire Center has lists of every public employee’s position and pay, which we can use to figure out the average pay of a train driver and conductor and the productivity of their labor. The NTD numbers are as of 2011, so I will use the number of employees of 2011, but the pay per employee of 2014 (at any rate, there have been no major service changes since 2011, so numbers are similar). In 2011, the LIRR averaged 5,000 car-hours per driver-year, and Metro-North averaged 4,000; the LIRR runs longer trains than Metro-North, so the figure for both railroads appear to be about 500 train-hours per driver-year. Both railroads had a little bit more than 2 conductors per driver on average (2.14 Metro-North, 2.47 LIRR). The average pay of a driver, as of 2014, is $109,000 on the LIRR and $120,000 on Metro-North, whereas the average pay of a conductor is $112,000 on the LIRR and $96,000 on Metro-North.
From this, we can piece together the average operating cost of commuter rail derived from on-board labor, per train-hour: $771 on the LIRR, $714 on Metro-North. Assuming 8 cars per train (and again, the LIRR tends to run longer trains), this is around $90-95 per car-hour. According to the NTD, the average operating cost of both was about $550 per car-hour in 2011, but this includes fixed costs, such as management and rolling stock. As we will see, variable operating costs are much lower.
As a digression, I’d like to point out that the peaky schedule of commuter rail contributes to the low productivity of the drivers. Crew schedules include substantial gap time between trips, and occasionally, especially on low-frequency diesel branches, they deadhead. That said, the subway’s number of revenue train-hours per driver is not materially different. For higher figures, one must leave New York. Toei got about 700 revenue hours per driver when I last checked, but I can no longer find the reference. On the London Underground, I do have fresh references, pointing in the same range: 76.2 million train-km per year at 33 km/h average speed (from TfL’s facts and figures), and a bit more than 3,000 train operators. In 2012, the last year for which there’s actual rather than predicted data (see also PDF-p. 7 of the TfL Annual Report), there were 720 revenue hours per train driver. This is in tandem with a less peaky schedule than in New York: although the average speed is barely higher than that of the New York subway, as reported in the NTD, the trains travel about 180,000 km per year (see fact 149 here), twice as long as in New York. In Helsinki, metro trains run every 10 minutes all day on each branch, every day, without any extra peak service, contributing to even higher utilization: the schedules show 65,000 revenue-hours per year, whereas a factsheet from 2010 shows 75 metro drivers, for a total of 867 revenue hours per driver. In both the UK and Finland, average hours per employee are marginally shorter than in the US; London Underground drivers have 36-hour workweeks.
The importance of this computation is not just to highlight that 44-73% improvement in revenue-hours per employee is possible, but to point out that, on the margins, adding off-peak service would make crew schedules more efficient, since higher frequency would reduce the need to deadhead and to wait between trains. This means that, although the average operating cost may be about $750 per train-hour, the marginal cost is lower, even without changes to work rules.
Suppose now that trains run without conductors, using proof-of-payment as on light rail lines, even ones in North America, and on countless commuter rail systems in Continental Europe. Suppose also that there are 720 revenue-hours per driver, and that a driver is paid $115,000 per year. This means that running extra trains would not cost $90-95 in on-board labor per car-hour, but only $20, a nearly fivefold improvement. At Helsinki’s level of productivity, a nearly sixfold improvement to $16.60 is possible. At the LIRR’s present average speed of 50 km/h (compared with 53 on New Jersey Transit and 59 on Metro-North), the fivefold improvement based on London Underground productivity would cut the average cost per car-km from $1.80-1.90 to $0.40; at a higher but still realistic 67 km/h, it’s a cut from $1.35 to $0.30. A large majority of this cut comes from eliminating conductors, which, by itself, would cut costs threefold, but raising driver productivity would allow an additional cut of 30-40%. I again stress that the marginal cost is lower than the average cost computed here, since less peaky schedules come with simpler crew scheduling; more off-peak service would by itself cut the average cost, which means its marginal cost would be quite low.
Let us now look at other variable costs than on-board labor. Two years ago, I did this computation for high-speed rail, and found that, provided the schedules did not have extra rush hour service, operating expenses would be very low. We can do the same computation for commuter rail, and note that the lower speeds imply that operating and maintenance costs are spread across less passenger-km, raising costs. Let us consider train maintenance, cleaning, and energy.
I do not have information about train maintenance costs on commuter rail. Instead, I will use those of high-speed rail, for which standards are higher. As I noted in my computation from two years ago, the reference here is California HSR’s 2012 Business Plan, which aggregates these figures from around the world on PDF-p. 136. Maintenance costs per train-km are $4.47 for the Tokaido Shinkansen (with 16-car trains) and $2.58 per the UIC (with what I assume are 8-car trains), both in 2009 dollars. These figures cluster around $0.30 per car-km in 2009 dollars, or $0.30-35 per car-km in 2014 dollars.
With cleaning, there is some information about commuter rail: the Empire Center has lists of coach cleaners on Metro-North (there are 314) and their pay (on average, a little less than $50,000 a year). This seems high given the amount of service Metro-North runs – about $0.15 per car-km. Shinkansen trains are cleaned on a seven-minute turnaround in Tokyo, using one cleaner per standard-class car; this includes tasks that are not required on commuter rail, such as flipping seats to face forward. A cleaner making $30 per hour cleaning a single car per 15 minutes, with each train cleaned once per 150 km roundtrip, would cost $0.05 per car-km. I suspect that part of the low productivity of Metro-North cleaners is again a matter of low off-peak frequency – Shinkansen cleaners work almost continuously – but I don’t have comparative data to back this up; New York City Transit pays even more per cleaner per car- or bus-km, but this is on much lower average speed, and per car- or bus-hour, it pays about $6.40, vs. about $8.90 for Metro-North. I’m going to pencil in $0.10 per car-km as the cost of cleaning.
Energy costs we can compute from first principles. This is easier than for HSR, since commuter trains travel at such speed that a large majority of their energy consumption is in acceleration, rather than cruising. The explicit assumptions I am making is that the top speed is 130 km/h (the two main LIRR lines are mostly 80 mph territory), each car weighs 54 metric tons (the LIRR M7s weigh 57.5 and the Metro-North M8s even more, but this is very high by international EMU standards, thanks to FRA regulations), the average distance between stations is 4 km (the LIRR’s average is less than that if all trains make all stops and more if there are some express trains), and the track resistance per unit of train mass is the same as for the X 2000, for which data exists on PDF-p. 64 of a thesis on tilting trains. Regenerative braking is assumed to exactly cancel out with losses in transmission. Train acceleration performance is assumed to be like that of the FLIRT, which would take about a kilometer to accelerate to line speed and have about 2 km of cruising before slowing down for the stop; the M7 has inferior performance, but this would reduce energy consumption since trains would spend more time at lower speed.
With the above assumptions, each acceleration, cruise, and deceleration cycle between stations consumes about 13 kWh, of which 10 kWh is required to accelerate the train to top speed, and the other 3 are for overcoming track resistance. See rough computations in a subthread on California HSR Blog starting with this comment, and bear in mind the initial comment made a large computational error. As for April of this year, transportation electricity costs in the state are $0.1245 per kWh, giving us about $1.60 per 4-km interstation, or $0.40 per car-km.
Overall, those three items are $0.80 per car-km. This means that going from paying train crew $1.35 per car-km to paying them $0.30 per car-km represents halving of direct marginal operating expenses: it means going from $2.15 to $1.10 per car-km. Finally, let us add management costs, which are not exactly marginal costs, but do grow as the workforce grows, since more employees require supervisors. At RENFE, we can extract 0.27 support and management employees per operations employee from the data on PDF-p. 46 of its 2010 executive summary. On the Helsinki urban rail network, the corresponding figure is 0.34 as per the factsheet referenced above. This affects train crew, cleaning, and maintenance staff, but not energy. If this means 30% extra costs, this means going from $2.675 to $1.31 per car-km – again, we see costs are halved.
The off-peak LIRR fare is 15 cents per kilometer at long distances (14 to Ronkonkoma, but much more at shorter distances, for example 21 to Hicksville). If the marginal cost of running off-peak service is $1.31 per car-km, it means a car needs to have 9 passengers without season passes on it paying 15 cents per km for the trip to break even. If it’s $2.675, it needs 18. Passengers who commute off-peak and get season passes for those commutes also contribute, but less – a monthly pass for Ronkonkoma is $377, which at 46 trips a month is 10 cents per kilometer. It is not hard to have 9 passengers even on a long train, or even 13 (at the lower rate of season passes); Ronkonkoma itself is a park-and-ride, where this is less likely, but high enough passenger volumes as far as Mineola and Hicksville and all over the Babylon Branch are quite likely. If the required minimum is 18, let alone 26, this is substantially harder.
I harp on North American mainline rail operations for a variety of antiquated practices, but the on-board overstaffing is by far the worst. While improvement in train driver productivity can occur as a natural byproduct of improvement in off-peak frequency, getting rid of conductors is not so easy. It means a fight with the unions over job losses. Some of the required layoffs can be mitigated by retraining conductors as train drivers and running more service, but this would not boost service hours by a factor of 5; on the Ronkonkoma Branch, the peakiest of the three long LIRR lines, boosting off- and reverse-peak frequency to half the peak frequency would only increase train service by a factor of about 1.8.
I am not an expert on labor relations, so I do not know if any solution barring a prolonged SEPTA-style strike could work, alone or in combination. One possibility would be to commit to reducing working hours in the next five or ten years instead of hiking pay; working hours would be gradually reduced to core Western European levels, with 35-hour workweeks and 6 weeks of paid vacation, and hourly pay would rise as scheduled while annual pay would be frozen. Another possibility is that the MTA would help laid off employees find private-sector work, as happened in the 1980s with Japan National Railways (see PDF-pp. 103-4 of a handbook on rail privatization). This possibility requires implementing the reform at a time of wage growth and low unemployment, when private-sector work is easier to find, but the US is posting strong job growth numbers nowadays and is projected to keep doing so for at least another year.
But whatever happens, the most important reform from the point of view of reducing marginal off-peak service provision costs is letting go of redundant train crew. Halving the variable operating costs is exactly what is required to convert the nearly empty off-peak trains from financial drains to an extra source of revenues, balancing low ridership with even lower expenses. This would of course compound with other operating efficiencies, limiting the losses of branch lines and turning the busier main line trains into profit centers. But nowhere else is there the possibility of cutting costs so much with one single policy change as with removing conductors and changing the fare enforcement system to proof-of-payment.
Update 7/31: first, check comments below about maintenance costs: as far as I can tell from poorly presented Empire Center data, they are about 2.5 times higher, for both trains and the infrastructure, than the maintenance costs of high-speed rail. Although the effect of reducing those costs to conventional HSR level is larger than the effect of eliminating conductors, the details of reducing maintenance costs are far more delicate than those of eliminating conductors and running trains more often so that train drivers have less downtime.
Second, there is a small error in the above calculations: the figure of $90-95 in crew salary per car-hour is based on two conflicting assumptions. To get to $771 per train-hour on the LIRR, I assumed the LIRR ran 10-car trains. To get down to the $90-95 range, I assumed 8-car trains; 10-car trains would make this $77/hour. If we redo the entire calculation with 10-car trains, still with HSR maintenance costs, then instead of a cut from $2.675/car-km to $1.31/car-km, improved labor efficiency would cut costs from $2.415/car-km to $1.21/car-km. This is based on exact LIRR salaries, whereas the original calculation assumes hybrid LIRR/Metro-North salaries, and Metro-North pays drivers better than the LIRR.
Now, trains are somewhat longer at the peak than off-peak. If off-peak service is already with 8-car trains, and the average number of conductors is constant, then the original calculation (a cut from $2.675 to $1.31) still holds. After all, the salaries of train drivers and conductors are the same no matter how long the train is. But the number of conductors is not constant – let’s say it is proportional to train length, so 8-car LIRR trains have 2 conductors instead of 2.47, just as Metro-North’s average number of conductors per train is shorter than the LIRR’s, in tandem with its shorter consists. This changes the calculation to a cut from $2.535 (reflecting fewer conductors than in the original calculation) to $1.31. Observe that no matter what assumption we use, the operating cost cut coming from removing conductors and using drivers more efficiently is about 50%, give or take 1-2%.
There’s an article in the New York Times by its architecture critic Michael Kimmelman, making a forceful case for the Gateway Project’s necessity. Like nearly all transit activists in New York, I think new Hudson tunnels are the top infrastructure priority for regional rail; like nearly all transit activists, I groan at Amtrak’s proposed budget, now up to $16 billion (but unlike most, I think that it should not be built unless costs can be brought down – I’d peg their worth at $5 billion normally, or somewhat more in a crunch). I would like to explain one specific piece of scope in Amtrak’s plan that can be eliminated, and that in fact provides very little transportation value: Penn Station South.
Like all proposals for new Hudson tunnels, Gateway is not just a simple two-track tunnel between New Jersey and Penn Station. No: the feuding users of Penn Station all think it needs more tracks. The rejected ARC proposal had a six-track multilevel cavern, and Gateway has Penn Station South, a proposal to demolish an entire block south of Penn Station and build seven additional platform tracks. The cost of just the real estate acquisition for Penn South: $769 million to $1.3 billion, at today’s prices. Trains using the preexisting tunnels would have to go to the preexisting Penn Station tracks, which I will call Penn Classic; trains using the new tunnels could go to either Penn Classic or Penn South, but the junction is planned to be flat. For illustration, see PDF-p. 12 of a press release of the late Senator Lautenberg, and a clearer unofficial picture on Railroad.net.
As a result of this proposed track arrangement, train services would initially suffer from the capacity limitations of flat junctions. Like Penn Station’s tracks 1-4, Penn South would be terminal tracks. This means that the only service possibilities are as follows:
1. Schedule all through-trains, such as Amtrak trains, through the preexisting tunnels.
2. Do not schedule any westbound trains from Penn South or any eastbound trains entering the preexisting Penn Station tracks: for example, no westbound trains from Penn South in the morning peak, and no eastbound trains entering Penn Classic in the afternoon peak.
3. Schedule around at-grade conflicts between opposing traffic.
Option #2 is impossible: Penn South has 7 tracks. If trains can enter but not leave in the morning, there will be room for 7 trains entering in the morning, a far cry from the several dozens expected. Option #1 is the better remaining option, but is ruled out, since Amtrak wants to use the new tunnels for its own trains. This leaves option #3, which restricts capacity, and complicates operations. Thanks to Amtrak’s imperialism, taking over regional rail projects for its own ends, Penn South has negative transportation value relative to just building new tunnels to Penn Classic’s tracks 1-4 (the transportation value relative to doing nothing is of course positive).
I emphasize that the negative transportation value of Penn South comes entirely from Amtrak’s involvement. The same infrastructure, used by passenger rail agencies that were more interested in providing high-quality public transportation than in turf wars, would have positive transportation value. However, as I explained to Kimmelman, this positive transportation value is low, and does not justify even the cost of real estate acquisition, let alone that of digging the station.
Briefly, as can be seen in the diagrams, the interlocking between the two new tunnel tracks and Penn’s eleven terminal tracks – tracks 1-4 of Penn Classic, and all of Penn South – is exceedingly complicated, which would limit approach speed, and not provide much flexibility relative to the number of tracks provided. This is to a large extent unavoidable when two approach tracks become eleven station tracks, but it does lead to diminishing returns from extra tracks. This is one of the reasons it’s easier if trains branch: it’s easier to turn 12 trains per hour on two tracks than to turn 24 on four (although both are done in Tokyo – indeed, the Chuo Line still turns 27 tph on two tracks).
Avoiding large crunches like this at urban terminals a benefit of through-running. This is hard to realize initially unless the new tunnel is what I call ARC-North. It’s still possible to through-run trains, pairing the new tunnels with the southern pair of East River Tunnels and the old tunnels with the northern pair, but it requires a lot of diverging moves at interlockings, limiting speed. Penn Station plans should be built with a long-term goal of simple moves at interlockings, to (slightly) increase speed and capacity and reduce maintenance needs.
However, it’s still possible to square the circle by requiring trains to turn fast on tracks 1-5 of Penn Station (track 5 splits to a terminating end and an end that runs through east of New York). Tokyo would be able to turn a full complement of 24 trains per hour on these tracks. Most other cities would not. However, as somewhat of a limiting European case, the RER A turns a peak train every 10 minutes on single track at Le Vésinet-Le Pecq, the next-to-last station on the Saint-Germain-en-Laye branch; Le Pecq has two through-tracks (also hosting a train every 10 minutes) and one terminal track. See map and schedule. This does not scale to 24 tph on four tracks; somewhat tellingly, those trains do not continue to the terminus, which is a three-track station, implying turning 12 tph on three tracks is problematic. The RER E turns 16 tph at the peak at Haussmann-Saint Lazare, a four-track city terminus, pending under-construction extension of the line to the west, which would make it a through-station.
If we accept 16 tph as the capacity of new Hudson tunnels without new Penn Station tracks, then the question should be what the most cost-effective way to raise future capacity is. An extra 9 tph, the equivalent of the difference between 16 tph and the 25 tph that the current tunnel runs and that Amtrak projects for Gateway, is within the capabilities of signaling improvements and better schedule discipline. Again looking to Paris for limiting cases, the combined RER B+D tunnel between Gare du Nord and Châtelet-Les Halles runs 32 tph, without any stations in the tunnel (the RER B and D use separate platforms), while the moving block signaling-equipped RER A runs 30 tph on its central segment, with stations (as do the S-Bahn systems of Berlin and Munich). The RER E was planned around a capacity of 18 tph, but only 16 tph are run today. 18+32 = 50 = 25+25. France is not Japan, with its famous punctuality: French trains are routinely late, and as far as I remember, the RER B has on-time performance of about 90% based on a 5-minute standard, worse than that of Metro-North in its better months.
More importantly, dropping Penn South from the Gateway plan saves so much money that it could all go to through-running, via a new tunnel from tracks 1-5 to Grand Central. This is about 2 km of tunnel, without any stations; in a normal city this would cost $500 million, the difficulty of building in Midtown canceling out with the lack of stations, and even at New York construction costs, keeping the tab to $2 billion should be doable. The 7 extension is $2.1 billion, but includes a station; an additional proposed infill station at 10th Avenue, dropped from the plan, would’ve $450 million, giving us $1.6 billion for about 1.6 revenue route-km, rising to 2.3 km including tail tracks – less than a billion dollars per kilometer.
At $2 billion, the premium over $1 billion of impossible-to-cut real estate acquisition costs is down to $1 billion. It’s unlikely the construction cost of Penn South could be just $1 billion, without general reductions in city construction costs, which would enable the Penn-Grand Central link to be cheaper as well. Each Second Avenue Subway station is about a billion dollars, and those stations, while somewhat deeper than Penn Station, are both much shorter and narrower than a full city block. The result is that building a Penn-Grand Central link is bound to be cheaper than building Penn South, while also providing equivalent capacity and service to a wider variety of destinations via through-running.
One difficulty is staging the tunnel-boring machines for such a connection: building a launch box involves large fixed costs, especially in such a crowded place as Midtown. One of the reasons Second Avenue Subway and the 7 extension are the world’s most expensive subway project per kilometer is that they’re so short, so those fixed costs are spread across less route length. The best way to mitigate this problem is to build the link simultaneously with the new Hudson tunnels. The staging would be done on Penn’s tracks 1-4, whose platforms would be temporarily stripped; the construction disruption involved in the tunnels is likely to require shutting those tracks down anyway. Depending on the geology, it may even be possible to use the same tunnel-boring machine from New Jersey all the way to Grand Central.
This doesn’t save as much money – the Penn-Grand Central link is extra scope, with its own costs and risks. However, unlike Penn South, it is useful to train riders. Penn South allows terminating trains at Penn Station more comfortably, without having to hit the limit of large-city terminal capacity; the Penn-Grand Central link creates this capacity, but also lets riders from New Jersey go to Grand Central and points north (such as Harlem, but also such more distant commercial centers as Stamford), and riders from Metro-North territory go to Penn Station and points west (such as Downtown Newark).
Normally, I advocate unbundling infrastructure projects, because of the tendency to lump too many things together into a single signature plan, which then turns into political football, a sure recipe for cost overruns. However, when projects logically lead to one another, then bundling is the correct choice. For example, building an entire subway line, with a single tunnel-boring machine and a single launchbox, usually costs less than building it in small stages, as is the case with Second Avenue Subway. New Hudson tunnels naturally lead into a new tunnel east of Penn Station, regardless of where this tunnel goes; and once a tunnel is built, its natural terminus is Grand Central.
The night before last, a Northeast Corridor Amtrak train derailed in Philadelphia, killing seven people. For some overviews of what happened, see Vox and Huffington Post. I am not going to talk directly about the accident here; it appears to be the same kind of derailment as on Metro-North a year and a half ago. Instead, I’m going to talk about the general issue of redundancy, which I saw people bring up in response to the train shutdowns that followed the crash. This is not the first time I hear about this; redundancy figures prominently into the list of benefits touted for new rail tunnels across the Hudson, allowing Amtrak to shut down the existing tunnels for repairs. Even before Amtrak proposed the Gateway project, transit activists talked about redundancy as a positive feature, for example Cap’n Transit. In this post, I am going to explain why, in public transportation and intercity rail, redundancy is in fact far less useful than other investments for the same amount of money.
First, let us list the various high-caliber rail networks of the world. In high-speed rail, the biggest networks are those of China, Japan, and France. None of them has redundancy, in the sense that there is more than one way to get between two cities on high-speed track. JR Central is building a second line from Tokyo to Osaka, but this is because the existing line is at capacity, running about 14 trains per hour into Tokyo at the peak; redundancy is a minor consideration. In regional rail, the busiest networks do have some redundancy, in the sense that if one line is shut down then people can take a parallel line, but this is because these networks are so busy that in most directions there’s enough demand to fill multiple lines. In Tokyo, which has the largest regional rail network, the parallel line is usually run by a competing company, so within each company’s network there’s little redundancy.
The reason for this non-redundant operation is simple: building new rail lines is expensive, while maintaining them adequately so that they don’t break down is cheap. Amtrak thinks that the Gateway tunnel will cost $16 billion. The program to repair the damage the preexisting tunnels suffered in Hurricane Sandy is $700 million, which assumes an accelerated construction schedule in which the tunnels will be shut down one track at a time, but conversely also includes work in the worse-damaged East River tunnels and not just the tunnels across the Hudson. This is a one-time repair after salt water intrusion, not annual ongoing maintenance. New Hudson tunnels are a necessary project for capacity reasons, but whatever benefit they have for redundancy is a fraction of their cost.
For high-speed rail, too, the cost of maintenance are far smaller than those of construction. The average maintenance costs of a single route-km of HSR are about €100,000 per year, versus €20 million for construction (see PDF-p. 9 of a study by Ginés de Rus about HSR between Stockholm and Gothenburg). With this amount of maintenance, there need not be any closures or disruptions in service.
Consider the Northeast Corridor, more concretely. To guarantee redundancy everywhere, so that train accidents do not disrupt the line, is to restore some passenger service along the former Baltimore and Ohio and tie-ins. Between Philadelphia and New York this means the West Trenton Line; between Philadelphia and Washington this means the CSX freight line. This also requires new Hudson tunnels. The cost of each of these elements is in the billions, and for the most part, with the exception of the new Hudson tunnels the transportation benefit is very low, especially south of Philadelphia, where there aren’t enough people to justify a second commuter line. Between New York and New Haven, there are no good alignments for a second route except for short bypasses; that’s what makes constructing HSR there so difficult.
Redundancy is a good feature of networks where failures are frequent and unavoidable; for such systems, redundancy is useful, as is the concept of failing gracefully. Rail transit is not such a network. It is both possible and desirable to reduce accident rates to levels approaching zero. Natural disasters remain hazardous, but are extremely infrequent, and at any rate when a deadly earthquake strikes, there are higher priorities than providing alternative passenger rail routes.
This is not to say that redundancy has no uses. Dense subway systems are redundant in the sense of providing multiple routes through the city – although, at the peak, they’re usually all very crowded. This makes it possible to shut down lines off-peak for maintenance; New York and London are both notorious for weekend service changes, and Paris shuts down short segments of lines for maintenance for a few weeks at a time (see for example here). But small subway systems manage to make do with just ordinary overnight shutdowns, and Copenhagen even runs trains 24/7, shutting down one track at a time at night and using the driverless operation to run trains on single track. It’s just more convenient to have more options, but not necessary.
The upshot is that when a subway or mainline rail network chooses where to lay additional lines, it should ignore all needs of redundancy, except possibly as tie-breakers. The benefits are there, but do not outweigh the cost of building less optimal lines. The operator should instead invest in systems, worker training, and maintenance regimes that ensure high reliability, and expand the network based on ordinary criteria of expected ridership and capacity needs. There’s no need to worry about failure, and it’s much better to design the network not to fail in the first place.
Last week, Bill de Blasio proposed a citywide ferry system in his otherwise perfectly boilerplate State of the City speech. Ferries, as Ben Kabak notes, are a tried and failed solution in New York, with a $30 per passenger subsidy on the ferry to the Rockaways, one of the neighborhoods mentioned in de Blasio’s speech. At the same time, some ferry routes do attract large numbers of passengers, including the Staten Island Ferry and SeaBus; in addition, MBTA Boat attracts fewer passengers than SeaBus, but achieves better cost recovery than the MBTA’s land transportation services. The purpose of this post is to explain which urban geographies could be well-served by ferries, and why New York could not.
Until the invention of the railroad, the fastest, cheapest, and most reliable form of transportation was the boat. Inland transportation of goods was by canal whenever possible. Overland transportation was so expensive that, as noted by Andrew Odlyzko, the cost of coal would double twelve miles away from the mine (see p. 14). As a result, cities were founded on shorelines and in river estuaries, and shrank if their rivers silted.
Railroads inverted this equation. Even in the 1830s, trains achieved higher speeds than ferries do today: the London and Birmingham averaged 31 km/h at opening, whereas SeaBus, which uses fast catamarans, averages at most 20 km/h. They could climb grades without resorting to locks and derailed much less often than boats sank; and, with the world still in the tail end of the Little Ice Age, the railroads did not freeze in winter. In this situation, a seaside location is no longer an advantage. At coastal locations, railroads have to cross more rivers, as did roads before; the current route of the Northeast Corridor in Connecticut was not the first but the third rail connection to be built between New York and Boston, after the Long Island Railroad (with ferry connections at both ends) and the inland Hartford and New Haven Railroad route.
The 19th century was a period of fast population growth in the industrialized world, especially the US, and fast urbanization. The industrial cities were then sited based on the optimal locations of a railroad network and not that of a shipping network. Birmingham and Manchester were already the largest cities in the UK outside of London, but the first railroad was, not coincidentally, built precisely to give Manchester port access without relying on the Manchester Ship Canal. In the US, we can see this in action, especially in New England: Boston has always been New England’s largest city, but many other early-settled cities – Salem, Newport, Plymouth, Provincetown, Portsmouth – declined, and now New England’s second cities include not just coastal New Haven and Providence but also inland Hartford, Worcester, and Nashua-Manchester.
In some areas of Long Island and New England, we can see towns with dual centers: an older coastal center, and a newer inland center, near the train station or a highway interchange. As Long Island had extensive suburban growth in the postwar era, the inland centers there are usually the larger ones, whereas in Massachusetts and Rhode Island, the coastal centers are usually larger.
Boston’s ferries serve these coastal centers. The Greenbush Line is locally infamous for its low ridership, about 3,000 per weekday in each direction. And yet, the ferries serving Hingham are fairly well-patronized: about 3,500 weekday passengers in both directions. (Both figures are from the 2014 Blue Book.) Now, the trains still carry nearly twice as many passengers as the ferries, but, relatively speaking, the ferries are doing quite well, since that part of the South Shore was settled before the railroad came, so the ferry serves passengers better than the trains do.
The other issue is which mode of transportation offers the most direct route. On the South Shore, the ferries go in a straighter line than the trains, which have to detour to remain on land. The Staten Island Ferry goes in a straight line, whereas roads and trains take big detours, especially for passengers leaving from St. George and not from near the bridges to Brooklyn and New Jersey. SeaBus, likewise, takes a direct route.
The significant fact for the Staten Island Ferry and SeaBus is that there economic centers of Staten Island and North Vancouver are right next to the ferry docks, coming from the fact that those areas were settled as suburban regions connected to the center by ferry. Because constructing a road or rail link across the New York Harbor or Burrard Inlet is difficult, those ferries were never replaced by fixed links; this is in contrast with Jersey City, which was also connected to New York by multiple ferry lines, but had enough demand a hundred years ago to fill the Hudson Tubes and later the Holland Tunnel with commuters.
None of these histories and geographies applies to the routes proposed by de Blasio and other ferry supporters. A Rockaway ferry has to detour around all of Brooklyn to reach Manhattan. The various waterfront ferries between Manhattan and Queens don’t really serve neighborhood centers, which are located around subway stations. Subway stations, like railroads, dislike coastal locations, not because of construction difficulties but because half their walk sheds would be underwater. Even Red Hook, which is cut off from the rest of the city by the Brooklyn-Queens Expressway and has no subway service, is not centered around the waterfront: the projects are several blocks inland, and Ikea Dock is facing the wrong way, south instead of west.
New York’s commercial centers, likewise, are inland. Why would a Midtown office developer waste any time building a skyscraper on the East River when the easternmost subway stations in Midtown are at Lexington Avenue? Thus the high-rise towers that line First Avenue are more residential than commercial, making them poor candidates for ferry connections. Lower Manhattan is better-connected to the water, but it is served by a large number of subway lines in all directions, none of which is at capacity since Midtown is the bigger office cluster. It’s also far from the waterfront condo clusters de Blasio wants to serve with ferries.
Even service between Staten Island and Manhattan shouldn’t be a ferry. A rail tunnel would offer a large improvement in trip times: about 8 minutes or even less, compared with 25 by ferry, and one to two transfers less than today. The question is entirely whether the costs could be contained enough to be in line with a realistic demand projection. Of course this is best realized as part of a regionwide commuter rail modernization plan, but even without such a plan, a connection to the 1 train would substantially reduce Staten Island’s commute time, which, at least last decade, was the longest of all US counties.
And this is an origin-destination pair that, given current infrastructure, is actually well-served by ferry, unlike the routes that de Blasio proposed. Ben tried to propose a better way of running ferries in New York, but with no real anchors to connect to, Ben’s proposal is a polite way of what I would phrase as “just don’t.”
Unlike Cuomo, de Blasio is not inherently hostile to public transit. However, he does not particularly care about transit, either. In this view, what he says about ferries is of limited consequence; the amounts of money in question are trivial. He’s not like Bloomberg, who directed $2 billion of city money to the 7 extension ahead of more deserving subway investments. Perhaps it’s wiser to focus on his plan to deck over Sunnyside Yards, or, more specifically, his invocation of massive projects including Stuyvesant Town, Coop City, and Starrett City – precisely the models that a Sunnyside decking should avoid.
However, there’s a good reason to focus on this, unimportant as it is. Cuomo’s failings are characteristic of an autocrat who is hostile to transit. De Blasio’s are characteristic of an autocrat who is indifferent. Although there is a long-term transit plan in New York, centered around completing Second Avenue Subway, this is not what de Blasio talked about, at all. Instead, he went for projects that can be done during his first term: off-board fare collection on a few more bus routes (“Select Bus Service,” complete with the pretense that they are bus rapid transit), and ferries. He won’t just follow an agenda set by others a long time ago: he has to remind people he exists on this issue as on his signature issues, but, as he doesn’t actually care about it, he will propose distractions that would at best do little (Select Bus Service) and at worst would be complete wastes of money (the ferries).
In a democracy, good transit advocates can push themselves into key positions at the ministry of transport, or its equivalent, such as a parliamentary committee on transportation (including the Congressional one, even). The same is true for people who care about other aspects of government spending and policy: housing, health care, education, defense, social welfare, policing. In an autocracy, such as the strong mayor system, it boils down to asking the autocrat to care and to take the right position. But the autocrat is just one person, and cannot pay equal attention to everything. Hence, ferries and Select Bus Service, in lieu of real transit investment.
Transit lines branch. Core routes have more demand than outlying ones, so naturally trains and buses run on trunk lines in the core and then branch farther out, to match frequency to demand. I gave an overview of this years ago. This is both normal across nearly all significant transit systems, and good practice. In this post, I’d like to focus on the opposite kind of branching, which I am going to call reverse branching, when one outlying line splits into two core routes. This is much less common, but exists in multiple cities, and leads to problems including restrictions on capacity and disappointing ridership. Cities should avoid building new lines that reverse branch, and in one famous existing case, London’s Northern line, the city is working on changing the situation by building a new outlying branch.
London’s Northern line, as can be seen on the Underground map, has three branches to the north and two in the center, but just one to the south. The highest ridership demand is in the center, but because both branches feed into just one southern branch, there is less than full capacity on the central branches, about 20 trains per hour each, compared with 30 tph on the southern branch and 33 tph on the Victoria and Jubilee lines. As a result, Transport for London has made recurrent plans to split the line for good: one central branch (through the City of London) using the existing southern branch and two of the northern ones, and one (through Charing Cross) using one northern branch and terminating at Kennington, the junction with the southern branch. An under-construction extension of the line from Kennington to Battersea can then be tied to the Charing Cross branch. There is some NIMBY opposition from a member of Parliament representing a constituency on one of the northern branches, who would like her constituents to have one-seat rides to both branches, but most likely, Transport for London’s need for capacity will make the split inevitable once the Battersea extension opens, ending the reverse branching practice.
In New York, routes branch and recombine, and thus it is common to have trains of different colors (which only denote Manhattan trunks) running together on a branch in Brooklyn, Queens, or the Bronx. The single busiest entry point into the Manhattan core is via 53rd Street Tunnel (connecting to Queens Boulevard), technically a branch since it runs trains connecting to both the Eighth and Sixth Avenue Lines. This, again, causes capacity problems. It’s not so bad on the numbered lines, where four trunk tracks (the Manhattan express trunks, carrying the 2/3 and 4/5) recombine in a different way to four tracks in Brooklyn (pairing the 2/5), but the lettered lines’ reverse branching in Uptown Manhattan and Queens initially forced eight trunk tracks (the Sixth and Eighth Avenue services, the B/D/F and A/C/E) to converge to six branch tracks (the two Queens Boulevard express tracks via 53rd, and the four Central Park West tracks). New subway connections have replaced this situation with twelve trunk tracks (including the Broadway Line’s N/Q/R) splitting to ten, spreading the problem around but not dealing with the fundamental restriction on capacity. The under-construction Second Avenue Subway will connect to the Broadway Line and run Q trains, raising the number of lettered tracks Uptown and in Queens to twelve, but this will not be enough to disentangle the tracks and provide full capacity on each core track; see below for proposed examples.
In Delhi, the Green Line splits into short branches, to provide transfers to two different Metro trunk lines. As seen on the system map, the Green Line does not enter central Delhi, and the current setup allows passengers to travel to central Delhi via two different routes. However, the Phase 4 extension plan extends the one branch to go out of the city in a V-shaped direction (the light green Kirti Nagar-Dwarka Section 28 line on this map), and has an extension that may connect to the other branch (Inderlok-Indraprastha, colored ocher on the map) to connect it to central Delhi, which may cause a serious mismatch in demand on the outlying common segment.
Finally, in Tokyo, subway lines reverse branch in two locations. The Namboku and Mita Lines share their southernmost three stations and the tracks in between. Although most Tokyo subway lines, including Namboku and Mita, run through to commuter lines, which provide the normal kind of branching, the Mita and Namboku Lines only do so either to the north or via the shared segment, as seen on this map, constraining capacity. They run only 12 peak tph each, and have low ridership by Tokyo subway standards. The Fukutoshin and Yurakucho Lines are in a similar situation, but the Fukutoshin Line does run through to a commuter line, the Tobu Tojo Line, without going through the shared segment (it is not depicted on the map, which is a few years out of date). The Fukutoshin Line has low ridership (see last page here), but the Yurakucho Line does not.
In all examples I’ve listed so far, the two core branches serve very central areas (as in London, New York, and Tokyo), or neither of them does (as in Delhi). Tokyo is somewhat of an exception, since the Yurakucho and Mita Lines serve Central Tokyo and the Fukutoshin and Namboku Lines serve secondary centers, but those secondary centers are very dense themselves; the Mita and Namboku Lines in particular are quite close in ridership. I am more wary of proposals to split an outlying line in the core that have one branch serving the CBD and one branch avoiding it, as in Delhi, assuming I understand the proposal correctly.
Also of note, all the examples I’ve listed involve subways. This is because conventional branching, with a core trunk splitting into multiple outlying branches, is more limited on urban rail than on both buses and regional rail. Most subway lines do not have more than two branches feeding into a trunk. In New York, not counting the split in the A, which is inherited from the LIRR, there is exactly one place where three subway routes share tracks: the N, Q, and R from Manhattan to Queens. In Stockholm, with its highly branched subway network, only one line, in one direction, splits into three. This is because even a split into three branches requires limiting off-peak frequency on the branches to less than a train every ten minutes, which is undesirable in large subway systems. The result is that reverse branching can easily create a situation in which there are more tracks in the core than in the outlying areas, as it does in all four cities surveyed above, restricting capacity on each core track.
In contrast, regional rail tends to operate at lower frequency on the branches, and this permits conventional branching with more than two branches per trunk. In addition, there are often turnback facilities at through-stations, and substantial four-track segments on otherwise two-track lines. The result is that reverse branching is possible without any constraint on core track capacity. The Berlin S-Bahn is highly branched in both the conventional and reverse senses. The RER E is being extended to the west, including a takeover of an RER A branch. And the Tokyo commuter rail network has extensive reverse branching, coming from through-service between commuter lines and subway lines but also from the Shonan-Shinjuku Line’s split from the Tokaido and Tohoku commuter lines. In none of these cases is there a significant restriction on core capacity, simply because there’s enough slack in the branches that they can’t fill to track capacity unless the core has filled as well.
In the US, I am familiar with three proposals for new subway lines that involve reverse splits, in Boston, Washington, and New York.
In Boston, the proposal actually involves commuter rail rather than the subway: the Worcester Line would use the Grand Junction Railroad to go through Cambridge to reach North Station, bypassing South Station. See map on page 38 of the statewide transportation capital budget proposal. This would not reduce capacity, since the Worcester Line is nowhere near exhausting the capacity of a two-track railroad, and moreover, the Grand Junction line would terminate at West Station within Boston proper, where there’s a railyard. However, this is still bad transit, for other reasons. West Station serves a residential neighborhood, without enough density to justify a fork toward both North Station and South Station. On top of that, since North Station lies outside the Boston CBD, the proposal is essentially a mixture of a radial and a circumferential line, with all the problems that would bring – and despite running as a circumferential line through Cambridge, there is no transfer planned with the Red Line, although the Grand Junction passes close to the Kendall/MIT station.
It would be better to bag all plans to use the Grand Junction until such time that the state builds the North-South Rail Link, connecting North Station with South Station. Then, the Grand Junction would make an almost perfect alignment for a circular line, with its eastern leg connecting North and South Stations and its western end going through Cambridge, making several stops, including a transfer to Kendall/MIT. This would require high investment – besides being a single-track at-grade line, the Grand Junction would require a new junction to connect to the Worcester Line to go east toward South Station, whereas today it only connects to the west, toward Allston and Brighton – but still a fraction of the cost of the North-South Rail Link, which is getting some serious political support, including from former governors Michael Dukakis and William Weld.
In Washington, there already is some reverse branching: the Yellow and Blue Lines share tracks in Virginia, but run on two different trunk lines in Washington proper, each shared with other lines, so four central tracks become four tracks in Virginia. But now with the opening of the Silver Line, raising the number of Virginia tracks to six, WMATA would like to separate the Blue Line from the Orange Line, which it shares tracks with in Washington, in order to provide six tracks across the District as well. This can only lead to awkward service patterns and wasted core capacity, as Matt Johnson demonstrates on Greater Greater Washington: because the Orange and Silver Line will keep interlining under any plan, reckoned from their split east there are only four tracks in Virginia and not six. Moreover, the Yellow Line interlines with the Green Line in the District, which means that even if it’s separated from the Blue Line, it could not run at full capacity.
Washington built itself into a corner with its Metro route decisions. There’s no corridor in the city that really needs a subway line; unlike New York, Los Angeles, and San Francisco, Washington has no corridor with so much bus ridership that it should be a subway line. A fourth subway line would be useful for service to Georgetown, but that’s about it. So decisions about a fourth line in the District should be based on the capacity needs of the branches, not those of the core. On a list of possible changes that WMATA looked at, Greater Greater Washington included a separated Silver Line, including separation up to the junction with the Orange Line so that they share no tracks. I’ll add that if WMATA wants to go down that route, then it should give the Orange Line its own route through the District and keep the Silver and Blue Lines together; this is because the Orange Line is the busiest of the three, so that it should be the least branched, in this case not branched at all whereas the other two do branch.
Finally, New York. Second Avenue Subway is going to change the nature of the reverse branching used by the lettered lines, for the better. Because the plan for Phases 1 and 2 is to run only the Q train, the city will finally have matching numbers of lettered tracks in and north and east of the Manhattan core: twelve tracks in the core, and twelve in Uptown Manhattan and Queens. Unfortunately, it is impossible to match service, because that would sever too many connections. Second Avenue Subway only connects to the Broadway express line, so to match service there couldn’t be any other service using the Broadway express.
Recall the London NIMBYism mentioned at the beginning of this post: that was about a service change that would give commuters a
cross-platform (see comment with diagram) transfer between their branch of the Northern Line and the central segment of the other branch. In New York, the transfers in western Queens involve a lot of walking between platforms, if they even exist. Then all the Broadway locals (the N/R) would go to Queens through 60th Street Tunnel, and thence to the Astoria Line, severing the connection to the Queens Boulevard Line. The Queens Boulevard Line has two ways into Manhattan: 53rd Street, which connects to both Eighth and Sixth Avenues, and 63rd Street, which connects only to Sixth Avenue. Moreover, either all F trains (through 63rd) have to run express in Queens and all E trains local, or the reverse; mixing and matching would produce at-grade conflict at the junction, as seen on the Queens 1 track map on nycsubway.org. The transfer between the E and F would be located at 74th Street in Queens, several kilometers east of the split, which is located just to the east of the westernmost express/local station, Queens Plaza. Neither the E nor the F would have a transfer to the N/R near their respective intersection points. The Q would not have a transfer to the E (it would have one to the F, though). This puts many commuters in an impossible situation and the capacity gains from it are frankly not enough to be worth it.
Instead, the capacity gains would be limited to running some more express trains on the Broadway Line. Before the service cuts in 2010, the N ran express on the Broadway Line, the Q terminated at 57th Street at the north end of Midtown, and a fourth Broadway route, the W, ran local and served the Astoria Line. Once the Q is extended up Second Avenue, the restored W could beef up Broadway Line service. Second Avenue would only get a branch despite its high ridership, but it’s still only a segment of a line.
Then there are Phases 3 and 4 of Second Avenue Subway, serving Midtown and the Lower East Side, where the Q runs on Broadway. The official plan introduces another reverse branching: a new route, the T, is planned to run the entire length of Second Avenue: see map here. When both phases are complete, there will be fourteen lettered tracks in Midtown but only twelve Uptown and in Queens.
To resolve this, the MTA should activate a connection that is included in the Second Avenue Subway plan as a non-revenue connection: a connection from Second Avenue south of the Q/T split at 63rd Street to 63rd Street Tunnel; currently, 63rd Street is the least used connection from Manhattan to Queens, since the reverse branching limits capacity and 63rd Street is the least useful connection since it enters Manhattan north of Midtown. This implies there should be a Queens Boulevard-Second Avenue service, which I will call the U, one letter next to the T. The Queens Boulevard express tracks are filled to capacity and the local ones are not, so the T should run local, cutting the frequencies on the existing local R and M trains a bit to make room. It would still leave New York with twelve Uptown and Queens tracks diverging to fourteen Midtown tracks, but it would distribute the load better, in the same way the present system distributes the load better than the 1930s-era reverse branching from six to eight tracks did.
New York is in a somewhat special case, in that its subway system is based on heavy branching and reverse branching, and moreover it’s historically based on three different systems, with poor transfers between them. Fully untangling the lines after Second Avenue Subway’s Phases 1 and 2 are built is not possible because there are no transfers between the lines that would result, and the station placement is such that any new transfers would involve long walks between platforms.
Other cities, especially cities planning new systems from scratch, should not emulate this feature, and should instead design all lines to either not branch at all or only branch conventionally. A system designed from the ground up could have cross-platform transfers between lines, and even make sure they’re timed, reducing the cost to passengers of having to transfer in lieu of using a reverse branch. It could be coherent, in the sense of making it easy for an unfamiliar passenger to understand how to get from each station to each other station. And it could be built for maximum capacity in the most crowded segments, where it matters the most.
On Tuesday, Andrew Cuomo proposed a new signature initiative: a $450 million AirTrain to LaGuardia, connecting to the Mets’ stadium on both the 7 train and the LIRR. The proposal has practically no merit even as an airport connector: Ben Kabak and Yonah Freemark both note, with helpful graphics, that the connection is so circuitous it’d be slower than the existing bus-subway options to nearly every destination, including everywhere in Manhattan. Capital New York notes that in general, transit activist reactions to the plan were cold, precisely because it’s such bad transit.
The interesting aspect of this is about the counter-criticism, and the discussion it led to. (In contrast, Cuomo’s general hostility to transit and intercity rail is not news, and it’s unlikely someone with such a history could come up with cost-effective transit plans.) The main reaction to the criticism is not “where would you spend $450 million instead?”. That question has a few answers, all of which are boring: the general MTA capital plan, or, if the money is to go to expansion, Second Avenue Subway Phase 2, the next item on the city’s transit agenda now that Phase 1 is nearing completion.
Instead, the main reaction is “how would you connect to LaGuardia instead?”. That question, too, has a definite answer, which Ben talked about in his post, and which I pointed out in my post about airport connectors last year: an extension of the N to the east, with several stops (for example, at Steinway and Hazen) to serve more of Astoria and not just airport riders. The N takes a direct route to Manhattan, passing through or next to the top areas for LaGuardia passengers, as seen in the second map here. But even that is the wrong question. There are probably more cost-effective subway extensions in New York, having nothing to do with LaGuardia; I have to say probably, since at no point has the MTA proposed large enough a slate of possible extensions that we can compare projected costs per rider and say “this is the best.” There might even be better ways to extend the N eastward than to LaGuardia: an elevated line over Ditmars, a short segment of the Grand Central Parkway, and Astoria Boulevard would serve East Elmhurst, a dense, transit-deprived section of Queens, and would probably produce higher ridership than a swerve from the GCP to the airport.
Such is the power of a governor who’s accountable to nobody: he proposes a scheme, and even the criticism is on the governor’s own terms of providing service to LaGuardia. Yonah compares travel times to various destinations on various alignments for connecting LaGuardia to the subway. Nate Silver’s response has an infographic with travel times from the airport to city hall in various American cities – an infographic that is of little use to New York, where the main destination is far north of city hall, but is well within the general topic of LaGuardia’s airport connections. Even I, cognizant of this agenda-setting power, have to at least mention an alternative LaGuardia connector, knowing readers will want a plan.
The cheeky response to this is that in a democracy, this wouldn’t happen. Now, the US is a democracy. Cuomo has to stand for election every four years. The worst infrastructure disasters tend to be in countries that are authoritarian through and through: Russia’s elevated winter Olympics costs in Sochi and Qatar’s human rights abuses in the World Cup preparations are the two biggest recent examples. But democracies with insufficient checks on political power are susceptible to this as well. This is common in the third world, where corruption is more common – hence the abuses of the World Cup last summer, in a solidly democratic country – but can also happen in developed countries with democratic deficits.
Usually, the phrase democratic deficit refers to the EU, and by analogy other supranational organizations. But in the US, it’s a useful framework for thinking of local and state governments. Rick Scott, Scott Walker, and John Kasich needed nobody’s approval to reject federal funding for intercity rail. Chris Christie did not need anyone’s approval to cancel ARC, or to cause traffic jams in retribution against a mayor who refused to endorse him; in a recent article in New York YIMBY, defending the cancellation of ARC as originally proposed, I made sure to take multiple barbs at Christie, just to avoid playing into the agenda of canceling ARC to posture about government waste while diverting rail money to the New Jersey Turnpike.
Cuomo’s power is if anything even greater: the New York state government works by a three men in a room model, in which the governor, the speaker of the State Assembly (just indicted for corruption), and the majority leader in the State Senate (currently relatively powerless and dependent on Cuomo) wield all practical power. In such a system, Cuomo does not have the power to shoot protesters, thankfully, but does have the power to propose megaprojects that glorify him, without a broad discussion with stakeholders, in which the MTA’s long-term expansion plans and cost-benefit ratios would come into play.
Last year, in writing about elite infrastructure projects that are not about meeting a service need, I noted that talking about such projects in terms of cost-effectiveness is moot, because they were never intended to be about benefiting the wider public. We could discuss where to spend money on transit in New York in the way that would benefit the largest number of riders. We could even discuss what the optimal way of connecting to LaGuardia is, before comparing the best connection with non-airport projects to see where it should lie on the list of future expansions. But it would be pointless, because Cuomo is not interested in spending money on benefiting the largest number of riders; he frankly does not care about transit riders. When the time came to support transit riders, for example in signing a lockbox bill guaranteeing that money the state government had promised the MTA would indeed go to the MTA, he vetoed the bill instead.
In such a climate, as soon as we talk about tweaks to Cuomo’s plan, Cuomo’s already won; whatever happens, he will reap the credit, and use it to buy political capital to keep building unnecessary megaprojects. Even trying to make the best of a bad situation by making the airport connector better is of little use, since Cuomo will support the plan that maximizes his political capital and not the one that maximizes transit usage even within such constraints as “must serve LaGuardia.”
This is evident in his response to criticism among transit activists. After listing the many pundits and activists who oppose the plan, Capital New York included a response from the governor’s office, which said, in so many words, “our plan is better because it doesn’t go through populated neighborhoods, where there would be NIMBYs.” What those of us who want good transit view as a feature – connecting to underserved neighborhoods and not just to the airport – Cuomo regards as a bug. A plan that included additional stops in Astoria might well attract community support, while still offering much faster trip times to Manhattan because of the direct route, but would rely on non-airport ridership, which Cuomo doesn’t care about, to keep the cost per rider reasonable.
Because of this disconnect between what would work for transit users and what would work for Cuomo, the only reasonable answer to the plan is a simple no, which should be said as sharply as possible. No working with the proposal: it’s terrible, a true stone soup. No tweaks: Cuomo wouldn’t want any ingredients that would improve the soup, and would insist on keeping the stone in anyway. (He doesn’t have to eat it, he doesn’t use transit either way.) And, within the parameters of a transit conversation in which people are desperate to see expansions, no discussion that validates Cuomo’s original plan.
Update 7/28: in a joint announcement with Joe Biden, Cuomo has just announced $4 billion in airport improvements at LaGuardia, bundling the rail connector into the larger projects. I have nothing to add that I didn’t already cover in this post and in my older post about elite infrastructure investments.
I’ve had an argument in comments with the author of Purple City about who commuter rail should serve. He’s argued before that cities should make sure outer suburbanites can get to the center via express commuter rail, and I will add that American cities do do that, and orient commuter rail too much around the needs of peak-hour outer suburbanites. Insofar as I think cities should have commuter rail there’s no disagreement, but what I think they do wrong is focusing too much on the peak. The two practices in contention are the low off-peak frequency (for example, Metra’s Union Pacific-North Line, which has no freight to speak of, has worse than hourly off-peak service), and the stop distribution, which has trains making few or no stops in the city proper.
The common thread of these two practices is that they optimize one variable: peak travel time for a suburban commuter to the CBD. This neglects other sources of ridership on commuter rail, which are suppressed in the US but significant in countries with more modern operating practices. I will contrast the peak-focused approach with a rapid transit approach, using examples that I believe will show that the latter is bound to get far more ridership, even in the suburbs.
First, let us imagine a contrasting system, one in which North American commuter rail looks more like an RER, an S-Bahn, or a Japanese commuter rail network. Such a system will have the following features:
1. Relatively consistent stopping pattern. The busier lines may have local and express trains, but the express trains will stop at the same major stops. Local trains will make all local stops over a fairly wide stretch.
2. Low ratio of peak to off-peak frequency, in the vicinity of 2:1 or even less. In a major city like Chicago or New York, a line that can’t support half-hourly service all day, at a bare minimum, will likely have no service at all; the only exceptions I can think of are services at range so long they’re practically intercity, like New York-Hamptons or New York-Allentown.
3. An urban stopping pattern that’s not too express. If there’s a parallel subway then it’s okay to have a somewhat wider stop spacing than in the inner suburbs beyond the subway’s range, but still closer to the 2-3 km range than the 4-5 km range of Metra.
If it’s possible to do so technologically, then the commuter line may be interlined with a subway line, even. This is usually hypothetical, since subways and commuter trains, where both exist, are almost always technologically incompatible; Tokyo and Seoul are the two major exceptions, with London a borderline case. However, it’s useful to consider such hypothetical cases, to examine what would happen to train service. I will consider two such cases: having Vancouver’s Evergreen Line take over West Coast Express (the original argument), and having Boston’s Red Line take over Old Colony Lines. Neither situation is technologically possible, even ignoring FRA and Transport Canada regulations, as both Boston and Vancouver build subway tunnels for much smaller trains than run on the mainline, but this discussion may be useful in cases where a takeover is feasible, such as when the commuter line is an isolated branch. I prefer to discuss the hypotheticals since the two examples in question are purer examples of priorities: outer-suburban peak service, or rapid transit-style service.
Vancouver’s rail service consists of the SkyTrain network, which gets about 400,000 weekday riders, and the West Coast Express, a peak-only commuter rail network running 5 trains per day per direction, with 11,000 weekday riders. SkyTrain’s under-construction Evergreen Line will intersect the West Coast Express at Port Moody and Coquitlam, and then serve more stations in Coquitlam off the mainline, while the WCE continues much farther to the east, into the Vancouver exurbs. The WCE connects Port Moody to Waterfront in 25 minutes and Coquitlam in 30 minutes; the Evergreen Line is projected to take 33 and 38 minutes respectively, with a transfer at Broadway/Commercial. Despite the slower service, the much higher frequency, all-day service, and connections to more of the Vancouver metro area win: the projected ridership for the Evergreen Line is about 23 million a year (see Table 2 on PDF-p. 4 here), which corresponds to about 75,000 per weekday.
Now, what’s in contention is whether it would be wise to have the same treatment at WCE stations farther east. The potential ridership at those stations is lower since they’re in less built-up areas, so it is likely cost-ineffective to build an Evergreen Line branch along the Canadian Pacific mainline and have it replace the WCE, but if such a line were built, it would most likely have the same effect on travel times: people would have to transfer at Broadway/Commercial, and not including the transfer time take 8 minutes more to get to Waterfront. The eastern end of the line, Mission, has 75-minute service now, and this would change to 83-minute service plus a transfer.
I claim that Mission residents would still take the train more often if it were 8 minutes lower. The reason is simple: as a proportion of overall travel time, the 8 minutes are more important to a 25-minute Port Moody commuter than to a 75-minute Mission commuter. Mission commuters live farther out, so they’re somewhat less likely to care about service to various neighborhoods along the way, but they’re even less likely to care about 8 minutes. They also are less likely to care about very high frequency, since their trips are longer, but they do care about service availability all day, even if they’d be okay with half-hourly service. Moreover, the Evergreen Line will connect to secondary nodes like Metrotown better than the WCE does, and eventually have direct service to Central Broadway and UBC, both of which draw commuters from the entire region.
In the present, the WCE works as a placeholder – it’s possible to reduce staffing and improve turnaround times to allow off-peak service, but there’s too little population east of Coquitlam to justify a SkyTrain extension, and so far population growth is fastest in inner-suburban Port Moody and Surrey (see here and here) and not east of Coquitlam. In the future, if those areas grow then it will make sense to replace the WCE with SkyTrain. WCE upgrades are unlikely – adding infill stations is practically impossible, as the line hugs an active port, with no good station sites. While SkyTrain’s driverless configuration keeps operating expenses down, it makes it impossible to extend branches to the suburbs cheaply by running them at-grade and in mixed traffic with freight.
Several of Boston’s subway branches are parallel to extant or closed commuter lines. The Orange Line runs alongside the Northeast Corridor to Forest Hills, the Blue Line took over parts of the narrow-gauge Boston, Revere Beach and Lynn Railroad, the Green Line D Branch took over a commuter rail loop used by the Boston and Albany, and the Red Line took over a New Haven Railroad branch line to Ashmont and runs alongside the Old Colony Lines to Braintree. At the time the Braintree extension opened the Old Colony Lines were closed for passenger service, but they have been since reopened, running from Braintree to South Station with just one stop in between, either JFK-UMass or Quincy Center (never both, except on trains that skip Braintree); off-peak frequency is about every two hours on each of two lines, and with some off-peak trains skipping Braintree, service to Braintree is worse than hourly. The Red Line takes 26-27 minutes to go from Braintree to South Station, the Old Colony Lines take 19-21 minutes.
As is projected in Vancouver, ridership on the Red Line is much higher: according to the 2014 Blue Book, on PDF-pp. 14 and 74, the busiest MBTA commuter rail station, Providence, gets 2,325 riders per weekday and the busiest Old Colony station, Bridgewater, gets only 1,036, while the Braintree extension’s five stops get 6,975, 4,624, 8,655 (Quincy Center), 4,785, and 5,122 (Braintree). Those five stops get 30,000 riders between them, meaning 60,000 since it’s unlikely people ride internally on the extension; this is nearly half the entire MBTA commuter rail ridership, and three times the ridership on the Old Colony Lines (counting Greenbush, which diverges at Quincy, as a third line).
As in Vancouver, I claim that a Red Line extension taking over the Old Colony Lines would have much higher ridership. Of course the frequency per line, already middling since the Braintree extension is a branch, would not be very good; but at the range of the suburbs served by these lines, half the current frequency of the Red Line, giving about 20 minutes at the peak and 30 off-peak, is enough, and is a massive improvement over multi-hour headways. The extra 5-8 minutes of travel times matter less as one moves farther out, again; travel time to South Station from the first Old Colony stations past Braintree, South Weymouth and Holbrook/Randolph, is 28 minutes, about the same as from Braintree on the Red Line, and those two stations have a bit more than 500 weekday riders each.
Moreover, the Red Line has something the commuter trains don’t: service to multiple centers within the inner Boston region. Downtown Crossing is closer to most jobs than South Station, saving people the walk. Cambridge is a major job center in its own right (it has more jobs than any New England city except Boston, ahead of Providence, Worcester, and Hartford). Back Bay is a bit more accessible via the Orange Line at Downtown Crossing or the Green Line at Park Street than via commuter rail at South Station.
Like SkyTrain, the Red Line can’t run on mainline rail tracks, and there is not enough population to justify an extension, nor enough population growth in New England for such an extension to ever pencil out. However, it’s possible to modernize commuter rail, as I have written before. This would not provide direct service to Downtown Crossing or Cambridge, but could provide cross-platform transfers to Back Bay, decent frequency all day, and, since regional EMUs can have very good performance characteristics, much higher average speeds than with today’s slow diesel locomotives even if trains make more stops.
The examples of Boston and Vancouver’s ridership patterns suggest that it’s okay to sacrifice speed to provide coherent service. It’s worth noting here that the bulk of present-day ridership on North American commuter rail would not benefit too much from such sacrifice. North American commuter rail provides awful service in the off-peak or to non-CBD destinations: even the Newark CBD, relatively well-served by New Jersey Transit, has a 26% mode share as a job center as of 2000, as per an Alan Voorhees Transportation Center report called Informed Intuition (PDF-p. 13). There’s a huge amount of latent ridership on North American commuter rail, which is why rapid transit gets so much more ridership than peak-focused commuter rail.
This doesn’t change much at different ranges of distance from the center. The few minutes saved by expressing through the city to the CBD matter a great deal to the suburbs right beyond city limits, but those innermost suburbs are precisely the ones that could make the most use of service to multiple city nodes. Farther out, where commuters to the city tend to be more likely to be working at the CBD, since it is more specialized than most secondary nodes, frequency and service to everywhere matter less, but the extra few minutes matter even less.
However, since present-day riders are precisely the narrow slice of potential users who are okay with the current setup, they have the potential to engage in NIMBY protests against any attempt at modernization. Why change what works for them? This is why Long Island representatives oppose such modernization attempts as letting Metro-North access Penn Station; it’s entirely a turf war. Even reforms that do not degrade trip times to the CBD are unlikely in this political situation, for example mode-neutral fares: the people paying premium fare to ride the LIRR or (to some extent) Metra are the ones who are okay with paying this fare, and who may object to increased train crowding coming from lower fares.
Judging by the ridership multiple between the Evergreen Line and WCE, there are likely to be a few million weekday rides coming out of Eastern Queens and Long Island if the LIRR is modernized, but those are not the Manhattan-bound commuters who dominate the discussion today. Instead, they are people who have gotten used to unusable commuter rail, and drive to work, or take long bus-subway commutes to avoid paying higher fares. They do not seem like a significant source of regional rail ridership because they are not current riders (or they ride local transit instead), but they are precisely what makes the difference between the low ridership of every North American commuter rail system and the higher ridership of many European systems.
Nearly all rapid transit lines belong to one of two categories: radial lines (a large majority), which connect city center with outlying neighborhoods or suburbs; and circumferential lines (a minority), which go around city center and often serve secondary centers and usually intersect all or nearly all radial lines perpendicularly, such as Paris’s Lines 2 and 6, Moscow’s Circle Line, Seoul’s Line 2, New York’s G train, and Shanghai’s Line 4. In this post, I’m going to discuss an uncommon third category, that of lines that combine circumferential and radial functions: they go toward city center, like a radial line, but then change direction and become circumferential. The G train in New York was like this until 2001, and Line 3 in Shanghai is like this today. This is apropos a proposal by a team Penn Design graduate students to build a variant of Triboro RX in New York that combines Triboro’s circumferential orientation with a radial commuter line. I believe such mixed lines are a recipe for low ridership and strained transfer points, and the Penn Design proposal is inferior to the original Triboro proposal.
First, some details about the mixed lines in question. The example most accessible to most readers is the historical G train in New York. When it first opened in the 1930s as part of the IND, it was designed to both connect Brooklyn and Queens without going through Manhattan and provide local service along radial lines, running alongside express trains that would serve Manhattan. Thus the northern half of the G ran under Queens Boulevard as a local, while the E and F trains provided express service and went to Manhattan. From the start, this arrangement was unstable. Demand for service to Manhattan was much greater than to Brooklyn, so people riding the G inbound changed to the E or F at the first express station after the one they boarded. With overcrowded express trains and undercrowded local ones, the Transit Authority was compelled to build a track connection in 1955 to add a Manhattan-bound local service, and to build a second track connection in 2001 to add another Manhattan-bound local train and remove the G from Queens Boulevard entirely.
In Shanghai, Line 3 was built as an (almost) entirely above-ground line, interlined for part of the way with the circular Line 4. The northern half of Line 3 is radial, running parallel to the overcrowded Line 1. However, where Line 1 enters the traditional center and serves People’s Square, Line 3 swerves west to go around it (missing Lujiazui, the new high-rise CBD to the east of People’s Square), interlining with Line 4, and leaving the loop southward to intersect Line 1 again at Shanghai South Railway Station. Its ridership disappoints not only by the standards of Line 1, but also by those of Line 4: 642,000 on 2014/4/30, the system’s busiest day, compared with 1,384,000 on Line 1 and 907,000 on Line 4. Line 6, which likewise combines a radial function at its northern end with a circumferential one at its center, serving Century Avenue but not Lujiazui or People’s Square, has even lower ridership, 376,000, although this is several times the original projection.
There’s a discussion on Human Transit, in which consensus is that the best circumferential lines connect secondary activity nodes that generate trips in their own right. Now, the G train connects Downtown Brooklyn (the largest business district in New York outside Manhattan) with Long Island City (one of the business districts of Queens), but it lacks the other positive feature of circumferential lines: transfers to the radial lines, to allow one-transfer trips from anywhere to those secondary nodes. The G has good transfers only to other IND lines, and at the Queens end, its transfer to the Queens Boulevard trains was cut in 2001 since, for operational reasons, it was cut not to its old junction with the E and F (Queens Plaza) but one station short (Court Square). Other G transfers are very recent and require a considerable amount of walking.
In contrast to the underperforming G, circumferential lines that both connect important activity nodes and have plenty of radial transfers are backbones of their cities’ transit systems. Shanghai’s Line 4 is fairly busy as noted above. Seoul’s Line 2 is according to a forum post the busiest in the system. Paris’s Lines 2 and 6 are only about average in ridership but combined would be the second busiest after Line 1 (and per route-km are third and fourth, only behind Lines 1 and 4, but are only narrowly ahead of many other lines). The juxtaposition of Shanghai’s Lines 3 and 4 in particular suggests that subway lines shouldn’t try to mix radial and circumferential functions.
Let us go back to the impetus for the post, Triboro RX. The proposal is to largely use existing freight rail lines, all of which are lightly used and could be turned over to the subway, to provide a semicircular line connecting nodes of activity in the Bronx, Queens, and Brooklyn. Because of the focus on using an existing right-of-way to reduce costs, the line misses the most important nodes in Brooklyn and Queens, which are served by the G in any case. However, it passes within half a kilometer of the Hub in the Bronx and, via a short greenfield tunnel, connects to Yankee Stadium, the Bronx’s busiest subway station; it also connects to Brooklyn College and 74th Street/Broadway in Queens, both busy stations if not as central as Downtown Brooklyn or Long Island City. Moreover, it provides direct Bronx-Queens service, which in the existing system requires circuitous routes through Manhattan with difficult transfers, and has reasonable transfers to nearly all subway lines. At the end, the lack of service to Downtown Brooklyn ensures it cannot be a very well-patronized line, but as the right-of-way is almost entirely in place, its cost per rider could be quite low.
In contrast, the Penn Design proposal, called Crossboro, severs the connection to the Hub and Yankee Stadium, and replaces it with service along the Northeast Corridor to Coop City, making sparse stops, at the same locations Metro-North plans to for Penn Station Access; this is 4 stops in 10 km in the Bronx, compared with stops spaced roughly every 800 meters in Queens and Brooklyn, as in the original proposal. The trains would be certified for mainline operation, on the model of the London Overground, rather than segregated from mainline traffic.
The problem with the Crossboro idea is essentially the same as that of the G train until 2001. Most riders at the four Bronx stops are interested in getting to Manhattan, and not to the neighborhoods served by Crossboro, so they’d look for transfer opportunities. The only such opportunity in the Bronx is at Hunts Point, to the 6, which unlike the E and F on Queens Boulevard runs local and provides slow service to Manhattan. With the extra transfer, there is no advantage to Bronx riders over continuing to walk or take the bus to the 2, 5, and 6 trains to Manhattan. Moreover, because of the poor transfers within the Bronx, it’s impossible to use the line to connect from Queens to anywhere in the western half of the Bronx, including Yankee Stadium and the Hub: there’s a connection only to the 6, and none to the 2, 4, 5, or B/D, or to Metro-North.
The principle in action here is that, especially when there are no compelling destinations, it’s critical to make sure the line provides connectivity between large regions. This means connecting to all or almost all radial lines in a region well-served by radial subways but poorly served by preexisting circumferential ones. Not including the 1, which no proposal connects to, the Bronx has five subway lines, all providing radial service to Manhattan; it is a feature of Triboro that it connects to all five (though the connection to the 2/5 requires a long walk), ensuring that people from nearly everywhere in the Bronx can use the line to get to its destinations in Queens and Brooklyn with one transfer. To get from anywhere to anywhere would require two transfers, but to get from a random station in the Bronx to one in Queens or Brooklyn often already requires two transfers, usually at busy Manhattan stations that are out of the way for the crosstown traveler.
Mixing radial and circumferential service interferes with this principle, since the radial component has to come at the expense of completing the circle (or semicircle in a geographically-constrained city like New York). Thus, it’s harder to use the line to get to a large enough variety of points of interest to make up for the fact that it misses the city’s most important destinations. Of course, such a line is also wanting as a radial line, since it misses the center. Thus, ridership underperforms, and the line usually fails to achieve its stated purpose.
In the last few years New York’s MTA has gone through multiple cycles in which a new head talks of far-reaching reform, while only small incremental steps are taken. The latest is the MTA Transportation Reinvention Commission, which has just released a report detailing all the way the MTA could move forward. Capital New York has covered it and hosts the report in three parts. Despite the florid rhetoric of reinvention, the proposals contained in the report are small-scale, such as reducing waste heat in the tunnels and at the stations on PDF-pp. 43-44 of the first part. At first glance they seem interesting; they are also very far from the reinvention the MTA both needs and claims to be engaging in.
Construction costs are not addressed in the report. On PDF-p. 53 of the first part, it talks about the far-reaching suburban Grand Paris Express project for providing suburb-to-suburb rapid transit. It says nothing of the fact that this 200-km project is scheduled to cost about 27 billion euros in what appears to be today’s money, which is not much more than $150 million per km, about a tenth as much as New York’s subway construction. (Grand Paris Express is either mostly or fully underground, I am not sure.) The worst problem for transit in the New York area is that its construction costs are an order of magnitude too high, but this is not addressed in the report.
Instead of tackling this question, the report prefers to dwell on how to raise money. As is increasingly common in American cities, it proposes creative funding streams, on the last page of the first part and the first six pages of the second part: congestion pricing, cap-and-trade, parking fees, a development fund, value capture. With the exception of congestion pricing, an externality tax for which it makes sense for revenues to go to mitigation of congestion via alternative transportation, all of these suffer from the same problem: they are opaque and narrowly targeted, which turns them into slush funds for power brokers. It’s the same problem as the use of cap-and-trade in California.
One of the most fundamental inventions of modern government is the broad-based tax, on income or consumption. Premodern governments funded themselves out of tariffs and dedicated taxes on specific activities (as do third-world governments today), and this created a lot of economic distortion, since not all activities were equally taxed, and politically powerful actors could influence the system to not tax them. The transparent broad-based tax, deeded to general revenue through a democratic process, has to be spent efficiently, because there are many government departments that are looking for more money and have to argue why they should get it. Moreover, the tax affects nearly all voters, so that cutting the tax is another option the spending programs must compete with. The dedicated fund does neither. If the broad-based tax is the equivalent of market competition, a system of dedicated funds for various government programs is the equivalent of a cartel that divides the market into zones, with each cartel member enjoying a local monopoly. In this way there’s a difference between the hodgepodge of taxes the MTA levies and wants to levy and Ile-de-France’s dedicated 1.4-2.6% payroll tax: the payroll tax directly affects all Francilien workers and employers, and were it wasted, a right-wing liberal politician could win accolades by proposing to cut it, the way New York Republicans are attacking the smaller payroll tax used to fund the MTA.
The proposals of where to spend the money to be raised so opaquely are problematic as well. There is a set of reforms, based on best practices in Continental Europe and Japan, that every urban transit system in the first world should pursue, including in their original countries, where often only some of those aspects happen. These include proof-of-payment fare collection on buses, commuter trains, and all but the busiest subway systems; all-door boarding on buses; mode-neutral fares with free transfers; signal priority and bus lanes on all major bus routes, with physically separated lanes in the most congested parts; a coherent frequent bus network, and high off-peak frequency on all trains; and through-service on commuter rail lines that can be joined to create a coherent S-Bahn or RER system. As far as I can tell, the report ignores all of these, with the exception of the vague sentence, “outfitting local bus routes with SBS features,” which features are unspecified. Instead, new buzzwords like resiliency and redundancy appear throughout the report. Redundancy in particular is a substitute for reliability: the world’s busiest train lines are generally not redundant: if they have parallel alternatives those are relief lines or slower options, and a shutdown would result in a major disruption. Amtrak, too, looks for redundancy, even as the busiest intercity rail line in the world, the Tokaido Shinkansen, has no redundancy, and is only about to get some in the next few decades as JR Central builds the Chuo Shinkansen for relief and for higher speeds.
The only foreigners on the Commission are British, Canadian, and Colombian, which may have something to do with the indifference to best industry practices. Bogota is famous for its BRT system, leveraging its wide roads and low labor costs, and Canada and to a lesser extent the UK have the same problems as the US in terms of best industry practices. Swiss, French, German, Japanese, Spanish, and Korean members might have known better, and might also have been useful in understanding where exactly the cost problems of the US in general and New York in particular come from.
The final major problem with the report, in addition to the indifference to cost, the proposal for reactionary funding sources, and the ignorance of best industry practices, is the continued emphasis on a state of good repair. While a logical goal in the 1980s and 90s, when the MTA was coming off of decades of deferred maintenance, the continued pursuit of the maintenance backlog today raises questions of whether maintenance has been deferred more recently, and whether it is still deferred. More oversight of the MTA is needed, for which the best idea I can think of is changing the cycles of maintenance capital funding from five years, like the rest of the capital plan, to one year. Long-term investment should still be funded over the long term, but maintenance should be funded more regularly, and the backlog should be clarified each year, so that the public can see how each year the backlog is steadily filled while normal replacement continues. This makes it more difficult for MTA chiefs to propose a bold program, fund it by skimping on maintenance, and leave for their next job before the ruse is discovered.
I tag this post under both good categories (“good transit” and “good/interesting studies”) and bad ones (“incompetence” and “shoddy studies”) because there are a lot of good ideas in the report. But none of them rises to the level of reinvention, and even collectively, they represent incremental improvement, of the sort I’d expect of a city with a vigorous capital investment program and industry practices near the world’s cutting edge. New York has neither, and right now it needs to imitate the best performers first.
North Americans are in love with trains that go in highway medians. A large fraction of urban rail construction since World War Two, both light rail and full metro, has used highway medians as cheap at-grade rights-of-way to extend train service, often deep into the suburbs. Some proposed longer-range lines are supposed to go in medians as well: Florida had reserved space in the I-4 median for Orlando-Tampa high-speed rail, and Xpress West planned to go from Las Vegas to the outskirts of the Los Angeles area in the I-15 median. The Texas Central Railway, a private group backed by JR Central planning high-speed rail between Dallas and Houston, is considering several alignments, but markets the route as following I-45 (no mention of median) in some public discussions. In nearly all cases, both urban and intercity, it borders on incompetent to design rail lines in highway medians; intercity lines frequently follow highways on one side, but even that tends to be overrated in American discussions in my experience.
For urban rail, the reason to use highways is that, in most of North America, they’re everywhere, and they’re usually equipped with generous medians and shoulders, allowing relatively cheap placement of rail tracks. Of note, this is generally not the cheapest option: construction on extant (often disused) rail rights-of-way tends to be cheaper. However, in many cases, a rail right-of-way is unavailable, hosts heavy freight traffic, has been permanently turned into a trail, or has commuter trains without integration into the rest of the urban transit network. Examples include the Dan Ryan half of the Red Line and both halves of the Blue Line in Chicago, the Orange and Silver Lines in Washington, the outer ends of BART, the Spadina line in Toronto, and several light rail lines. Often they run on one side of the road, but more frequently they’re in the median, which was often reserved for it when the road was built (as in Chicago and Calgary).
The problem is that nobody wants to live, work, or hang out next to a busy grade-separated road. Living or working a kilometer or two away, with easy access by car, is great for the driver, but within close walking distance, there is just too much noise, pollution, and blight, and the pedestrian environment is unwelcoming. The transit-oriented development in Metrotown and Arlington could not have happened next to a freeway. Christof Spieler frames this as a decision of spending more money on routing trains near where people live versus staying on the easy rights-of-way. But this isn’t quite right: the Expo Line in Vancouver was assembled out of an interurban right-of-way and a city center tunnel, both out of service; the line’s high ridership comes from subsequent development next to Metrotown and other stations.
Other times, the routing comes from a deliberate decision to integrate the trains with cars, with large park-and-rides at the ends. This is common on newer light rail systems in the US (though not Canada, as Calgary prefers integration with connecting buses) and in the Washington and San Francisco suburbs. This makes things even worse, by extending the radius within which the environment is built for cars rather than for people, and by encouraging the same park-and-ride construction elsewhere, along abandoned railroads and greenfield routes, where the preexisting environment is not car-oriented.
I do not want to categorically say that cities should never build urban rail alongside highways. But I cannot think of a single example in which this was done right. Calgary is a marginal case: it did build light rail along highways, and had some success with transit-oriented development, but those highways are arterials rather than freeways, and this makes the pedestrian environment somewhat better.
The situation is somewhat different for suburban rail, but usually the scale of suburban rail is such that there’s not much new construction, only reappropriation of old lines. These lines are long and the environments low-density, making it hard justify the costs of new lines in most cases. Where new suburban rail is built, for examples the Grand Paris Express, and various airport connectors, it is typically in environments with such expected traffic density that the rules for urban rail apply, and we tend to see more underground construction or usage of extant rights-of-way.
The reasons favoring highway alignments intercity rail in the US are somewhat different. Tellingly, HSR in Europe is frequently twinned with motorways. It is not about integration with cars, since those alignments are rarely if ever meant to have major stops in their middle. Instead, it’s about picking a pre-impacted alignment, where there are fewer property takings and fewer NIMBYs. This logic is sound, but I often see Americans take it to extremes when discussing HSR.
The first problem is that roads are almost never as straight as HSR needs to be. The design standards I have seen after briefly Googling give the radius of a motorway capable of about 120 km/h as, at a minimum, 500-700 meters. With these curves, trains, too, are capable of achieving about 120 km/h – less at 500 meters without tilting, more at 700 meters with tilting. The most recent high-speed lines are built with a minimum curve radius of 7 km; about the absolute minimum that can be done, with design compromises and tilting trains, is 4 km. This implies that the trains have to deviate from the motorway alignment whenever it curves. In flat regions the road curves are much gentler than the minimum, but still too sharp for full-speed running. Both Florida HSR and Xpress West noted that the trains would have to slow down whenever the Interstate curved, because the need to run in the median would prevent them from curving gently enough to maintain full speed.
Of note, the European examples of HSR running in motorway alignments have it running alongside the roads, not in the medians. I invite the reader to spend a few minutes following French LGVs on Google Maps and seeing this. This is because there invariably have to be small deviations from the road, which in a rural area are trivial when one runs next to the road but require viaducts when one runs between the road’s two carriages.
There may also be an issue regarding reusing the Interstates. To transit supporters who view HSR as a replacement for freeways, this has an element of poetic justice, or just plain practical reuse of infrastructure they think is obsolete. I chanced upon this while looking up Interstate design standards, but I’ve seen similar proposals elsewhere, as well as dissimilar proposals making use of interstate terminology, as a reminder of past national greatness. It comes from the same place as proposals to reuse auto factories to produce rolling stock: there’s a romantic aspect in addition to or instead of an economic one.
But the most fundamental problem is that the contentious experiences of the freeway revolts and modern-day NIMBYism have soured Americans on any process that involves brazen takings. What I mean by brazen is that carving a new right-of-way, especially through a populated area, looks obvious on a map. In contrast, sticking to a preexisting right-of-way and incrementally widening it or straightening curves is less controversial, even when it involves eminent domain as well, and opposition remains much more local, based on the specific properties being taken, rather than stated in general principles. I am not completely sure why this is so; my suspicion is that widening and straightening are more easily justified as things that must be done, whereas a new right-of-way looks gratuitous.
In either way, Americans have convinced themselves that NIMBYs are a major obstacle to infrastructure construction. While zoning is a notoriously NIMBY-prone process, infrastructure often isn’t. In the English common law world, expropriations are if anything easier than in France, where farmers are especially powerful, or Japan, where rioters threatened to block the construction of Narita Airport. NIMBYs are good at getting their names out in the media, but when it comes to blocking construction, they are relatively powerless; California HSR is facing NIMBYs in the Central Valley, many of whom are conservative and politically opposed to the project regardless of local impact, but so far they have not managed to delay construction.
However, NIMBYs are a convenient bogeyman for public projects, as their motives are openly selfish. They give charismatic, authoritarian leaders the opportunity to portray their infrastructure projects as battles between the common good and backward-looking parochial interests. As I’ve noted multiple times before, New York’s livable streets community (which is similar politically to the set of HSR supporters in the US) tends to overblow the importance of NIMBYs to the point of seeing NIMBYs even when the concerns have nothing to do with NIMBYism: see, for example, the reaction to the opposition of two Harlem politicians to a plan to speed up only the whitest bus route through the neighborhood.