Category: Amtrak
Timetable Padding Practices
Two weeks ago, the Wall Street Journal wrote this piece about our Northeast Corridor report. Much of it was based on a series of interviews William Boston did with me, explaining what the main needs on the corridor are. One element stands out since the MTA responded to what I was saying about schedule padding – I talk about how Amtrak and Metro-North both pad the timetables on the Northeast Corridor by about 25%, turning a technical travel time of an hour into 1:15 (best practices are 7%), and in response, the MTA said that they pad their schedules 10% and not 7%. This is an incorrect understanding of timetable padding, which speaks poorly to the competence of the schedule planners and managers at Metro-North.
The article says,
Aaron Donovan, a spokesman for the Metropolitan Transportation Authority, says the extra time built into Metro-North schedules generally averages 10%, depending on destination and length of trips, and takes into account routine track maintenance and capital work that can increase runtime. Metro-North continually reviews models, signal timing, equipment, and other elements of operation to improve travel times and reliability for customers, he says.
This is, to be clear, incorrect. Metro-North routinely recovers longer delays than 10%; delay recovery on the New Haven Line can reach well over 20 minutes out of a nominally two-hour trip, around 25% of the unpadded trip length. The reason this is incorrect isn’t that Donovan is dishonest or incompetent (he is neither of these two things), but almost certainly that the planners he spoke with genuinely believe they only pad 10%, because they, like all American railroaders, do not know how modern rail scheduling is done.
Modern rail scheduling practices in the higher-reliability parts of Europe and Japan start with the technical timetable, based on the actual speed zones and trains’ performance characteristics. This includes temporary speed restrictions. The ideal maintenance regime does not use them, instead relying on regular nighttime maintenance windows during which all tracks are out of service. However, temporary restrictions may exist if a line is taken out of service and trains are rerouted along a slower route, which is regrettably common in Germany. Modern signaling systems are capable of incorporating temporary speed restrictions – this is in fact a core requirement for American positive train control (PTC), since American maintenance practices rely on extensive temporary restrictions for work zones and one-off slowdowns. If the signal system knows the exact speed zones on each section of track, then so can the schedule planners.
The schedule contingency figure is computed relative to the best technical schedule. It is not computed relative to any assumption of additional delays due to dispatch holds or train congestion. The 7% figure used in Switzerland, Sweden, and the Netherlands takes care of the high levels of congestion on key urban segments.
The core urban networks in these countries stack favorably with Metro-North in track utilization. The Hirschengraben Tunnel in Zurich runs 18 S-Bahn trains per hour in each direction most of the day and 20 at rush hour with some extra S20 runs, and the Weinberg Tunnel runs 8 S-Bahn trains per hour and if I understand the network graphic right 7.5 additional intercities per hour. I urge people to go look at the graphic and try tracking down the lines just to see how extensively branched and reverse-branched they are; this is not a simple network, and delays would propagate. The reason the Swiss rail network is so punctual is that, unlike American rail planning, it integrates infrastructure and timetable development. This means many things, but what is relevant here is that it analyzes where delays originate and how they propagate, and focuses investments on these sections, grade-separating problematic flat junctions if possible and adding pocket tracks if not.
Were I to only take timetable padding into account relative to an already more tolerant schedule incorporating congestion and signaling limitations, I would cite much lower figures for timetable padding. Switzerland speaks of a uniform 7% pad, but in Sweden the figures include two components, a percentage (taking care of, among other things, suboptimal driver behavior) and a fixed number of minutes per 100 km, which at current intercity speeds resolve to 7% as in Switzerland. But relative to the technical trip time, the pad factors based on both observed timetable recovery and actual calculations on current speed zones are in the 20-30% range, and not 10%.
Of course, at no point do I suggest that Metro-North and Amtrak could achieve 7% right now, through just writing more aggressive timetables. To achieve Swiss, Dutch, and Swedish results, they would need Swiss, Dutch, or Swedish planning quality, which is sorely lacking at both railroads. They would need to write better timetables – not just more aggressive ones but also simpler ones: Metro-North’s 13 different stopping patterns on New Haven Line trains out of 16 main line peak trains per hour should be consolidated to 2. This is key to the plan – the only way Northern Europe makes anything work is with fairly rigid clockface timetables, so that one hour or half-hour is repeated all day, and conflicts can be localized to be at the same place every time.
Then they would need to invest based on reliability. Right now, the investment plans do not incorporate the timetable, and one generally forward-thinking planner found it odd that the NEC report included both high-level infrastructure proposals and proposed timetables to the minute. In the United States, that’s not the normal practice – high-level plans only discuss high-level issues, and scheduling is considered a low-level issue to be done only after the concrete is completed. In Northern European countries with competently-run railways and also in Germany, the integration of the timetable and infrastructure is so complete that draft network graphics indicating complete timetables of every train to the minute are included in the proposal phase, before funding is committed. In Switzerland, such a timetable is available before the associated infrastructure investments go to referendum.
Under current American planning, the priorities for Metro-North are in situ bridge replacements in Connecticut because their maintenance costs are high even by Metro-North’s already very expensive standards. But under good planning, the priority must be grade-separating Shell Interlocking (CP 216) just south of New Rochelle, currently a flat junction between trains bound for Grand Central and ones bound for Penn Station. The flat junctions to the branches in Connecticut need to be evaluated for grade-separation as well, and I believe the innermost, to the New Canaan Branch, needs to be grade-separated due to its high traffic while the ones to the two farther out branches can be kept flat.
None of this is free, but all of this is cheap by the standards of what the MTA is already spending on Penn Station Access for Metro-North. The rewards are substantial: 1:17 trip times from New Haven to Grand Central making off-peak express stops, down from 2 hours today. The big ask isn’t money – the entire point of the report is to figure out how to build high-speed rail on a tight budget. Rather, the big ask is changing the entire planning paradigm of intercity and commuter rail in the United States from reactive to proactive, from incremental to comfortable with groun-up redesigns, from stuck in the 1950s to ready for the transportation needs of the 21st century.
Against State of Good Repair
We’re releasing our high-speed rail report later this week. It’s a technical report rather than a historical or institutional one, so I’d like to talk about a point that is mentioned in the introduction explaining why we think it’s possible to build high-speed rail on the Northeast Corridor for $17 billion: the current investment program, Connect 2037, centers renewal and maintenance more than expansion, under the moniker State of Good Repair (SOGR). In essence, megaprojects have a set of well-understood problems of high costs and deficient outcomes, behind-the-scenes maintenance has a different set of problems, and SOGR combines the worst of both worlds and the benefits of neither. I’ve talked about this before in other contexts – about Connecticut rail renewal costs, or leakage in megaproject budgeting, or the history of SOGR on the New York City Subway, or Northeast Corridor catenary. Here I’d like to synthesize this into a single critique.
What is SOGR?
SOGR is a long-term capital investment to bring all capital assets into their expected lifespan and maintenance status. If a piece of equipment is supposed to be replaced every 40 years and is currently over 40, it’s not in good repair. If the mean distance between failures falls below a certain prescribed level, it’s not in good repair. If maintenance intervals grow beyond prescription, then the asset to be maintained is not in good repair. In practice, the lifespans are somewhat conservative so in practice a lot of things fall out of good repair and the system keeps running. The upshot is that because the maintenance standards are somewhat flexible, it’s easy to defer maintenance to make the system look financially healthier, or to deal with an unexpected budget shortfall.
Modern American SOGR goes back to the New York subway renewal programs of the 1980s and 90s, which worked well. The problem is that, just as the success of one infrastructure expansion tempts the construction of other, less socially profitable ones, the success of SOGR tempted agencies to justify large capital expenses on SOGR grounds. In effect, what should have been a one-time program to recover from the 1970s was generalized as a way of doing maintenance and renewal to react to the availability of money.
Megaprojects and non-megaprojects
In practice, what defines a megaproject is relative – a 6 km light rail extension is a megaproject in Boston but not in Paris – and this also means that they are not easy to locally benchmark, or else there would be many like them and they would be more routine. This means that megaprojects are, by definition, unusual. Their outcome is visible, and this attracts high-profile politicians and civil servants looking to make their mark. Conversely, their budgeting is less visible, because what must be included is not always clear. This leads to problems of bloat (this is the leakage problem), politicization, surplus extraction, and plain lying by proponents.
Non-megaprojects have, in effect, the opposite set of problems. Their individual components can be benchmarked easily, because they happen routinely. A short Paris Métro extension, a few new infill stations, and a weekend service change for track renewal in New York are all examples of non-megaprojects. These are done at the purely professional level, and if politicians or top managers intervene, it’s usually at the most general level, for example the institution of Fastrack as a general way of doing subway maintenance, and that too can be benchmarked internally. In this case, none of the usual problems of megaprojects is likely. Instead, problems occur because, while the budgeting can be visible to the agency, the project itself is not visible to the general public. If an entire new subway line’s construction fails and the line does not open, this is publicly visible, to the embarrassment of the politicians and agency heads who intended to take credit for it. In contrast, if a weekend service change has lower productivity than usual, the public won’t know until this problem has metastasized in general, by which point the agency has probably lost the ability to do this efficiently.
And to be clear, just as megaprojects like new subway lines vary widely in their ability to build efficiently, so do non-megaproject capital investments vary, if anything even more. The example I gave writing about Connecticut’s ill-conceived SOGR program, repeated in the high-speed rail report, is that per track- or route-km the state spends in one year about 60% as much as what Germany spends on a once per generation renewal program, to be undertaken about every 35 years. Annually, the difference is a factor of about 20. New York subway maintenance has degraded internally over time, due to ever tighter flagging rules, designed for worker protection, except that worker injuries rose from 1999 to the 2010s.
The Transit Costs Project
The goal of the Transit Costs Project is to use international benchmarking to allow cities to benefit from the best of both worlds. Megaprojects benefit from public visibility and from the inherent embarrassment to a politician or even a city or state that can’t build them: “New York can’t expand the subway” is a common mockery in American good-government spaces, and people in Germany mock both Bavaria for the high costs and long timeline of the second Munich S-Bahn tunnel and Berlin for, while its costs are rather normal, not building anything, not even the much-promised tram alternatives to the U-Bahn. Conversely, politicians do get political capital from the successful completion of a megaproject, encouraging their construction, even when not socially profitable.
Where we come in is using global benchmarking to remove the question marks from such projects. A subway extension may be a once in a generation effort in an American city, but globally it is not, and therefore, we look into how as much of the entire world as we can see into does this, to establish norms. This includes station designs to avoid overbuilding, project delivery and procurement strategies, system standards, and other aspects. Not even New York is as special as it thinks it is.
To some extent, this combination of the best features of both megaprojects and non-megaprojects exists in cities with low construction costs. This is not as tautological as it sounds. Rather, I claim that when construction costs are low, even visible extensions to the system fall below the threshold of a megaproject, and thus incremental metro extensions are built by professionals, with more public visibility providing a layer of transparency than for a renewal project. This way, growth can sustain itself until the city runs out of good places to build or until an economic crisis like the Great Recession in Spain makes nearly all capital work stop. In this environment, politicians grow to trust that if they want something big built, they can just give more money to more of the same, serving many neighborhoods at once.
In places with higher costs, or in places that are small enough that even with low costs it’s rare to build new metro lines, this is not available. This requires the global benchmarking that we use; occasionally, national benchmarking could work, in a country with medium costs and low willingness to build (for example, Germany), but this isn’t common.
The SOGR problem
If what we aim to do with the Transit Costs Project is to combine the positive features of megaprojects and non-megaprojects, SOGR does the exact opposite. It is conceived as a single large program, acting as the centerpiece of a capital plan that can go into the tens of billions of dollars, and is therefore a megaproject. But then there’s no visible, actionable, tangible promise there. There is no concrete promise of higher speed or capacity. To the extent some programs do have such a promise, they are subsumed into something much bigger, which means that failing to meet standards on (say) elevator reliability can be excused if other things are said to go into a state of good repair, whatever that means to the general public.
Thus, SOGR invites levels of bloat going well beyond those of normal expansion megaprojects. Any project can be added to the SOGR list, with little oversight – it isn’t and can’t be locally benchmarked so there is no mid-career professional who can push back, and conversely it isn’t so visible to the general public that a general manager or politician can push back demanding a fixed opening deadline. For the same reason, inefficiency can fester, because nobody at either the middle or upper level has the clear ability to demand better.
Worse, once the mentality of SOGR is accepted, more capital projects, on either the renewal side or the expansion side, are tied to it, reducing their efficiency. For example, the catenary on the Northeast Corridor south of New York requires an upgrade from fixed termination/variable tension to auto-tension/constant tension. But Amtrak has undermaintained the catenary expecting money for upgrades any decade now, and now Amtrak claims that the entire system must be replaced, not just the catenary but also the poles and substations. The language used, “the system is falling apart” and “the system is maintained with duct tape,” invites urgency, and not the question, “if you didn’t maintain this all this time, why should we trust you on anything?”. With the skepticism of the latter question, we can see that the substations are a separate issue from the catenary, and ask whether the poles can be rebuilt in place to reduce disruption, to which the vendors I’ve spoken with suggested the answer is yes using bracing.
The Connecticut track renewal program falls into the same trap. With no tangible promise of better service, the state’s rail lines are under constant closures for maintenance, which is done at exceptionally low productivity – manually usually, and when they finally obtained a track laying machine recently they’ve used it at one tenth its expected productivity. Once this is accepted as the normal way of doing things, when someone from the outside suggests they could do better, like Ned Lamont with his 30-30-30 proposal, the response is to make up excuses why it’s not possible. Why disturb the racket?
The way forward
The only way forward is to completely eliminate SOGR from one’s lexicon. Big capital programs must exclusively fund expansion, and project managers must learn to look with suspicion on any attempt to let maintenance projects piggyback on them.
Instead, maintenance and renewal should be budgeted separately from each other and separately from expansion. Maintenance should be budgeted on the same ongoing basis as operations. If it’s too expensive, this is evidence that it’s not efficient enough and should be mechanized better; on a modern railroad in a developed country, there is no need to have maintenance of way workers walk the tracks instead of riding a track inspection train or a track laying machine. With mechanized maintenance, inventory management is also simplified, in the sense that an entire section of track has consistent maintenance history, rather than each sleeper having been installed in a different year replacing a defective one.
Renewal can be funded on a one-time basis since the exact interval can be fudged somewhat and the works can be timed based on other work or even a recession requiring economic stimulus. But this must be held separate from expansion, again to avoid the Connecticut problem of putting the entire rail network under constant maintenance because slow zones are accepted as a fact of life.
The importance of splitting these off is that it makes it easier to say “no” to bad expansion projects masquerading as urgent maintenance. No, it’s not urgent to replace a bridge if the cost of doing so is $1 billion to cross a 100 meter wide river. No, the substations are a separate system from the overhead catenary and you shouldn’t bundle them into one project.
With SOGR stripped off, it’s possible to achieve the Transit Costs Project goal of combining the best rather than the worst features of megaprojects and non-megaprojects. High-speed rail is visible and has long been a common ask on the Northeast Corridor, and with the components split off, it’s possible to look into each and benchmark to what it should include and how it should be built. Just as New York is not special when it comes to subways, the United States is not special when it comes to intercity rail, it just lags in planning coordination and technology. With everything done transparently based on best practices, it is indeed possible to build this on an expansion budget of about $17 billion and a rounding-error track laying machine budget.
The Problems of not Killing Penn Expansion and of Tariffs
Penn Station Expansion is a useless project. This is not news; the idea was suspicious from the start, and since then we’ve done layers of simulation, most recently of train-platform-mezzanine passenger flow. However, what is news is that the Trump administration is aiming to take over Penn Reconstruction (a separate, also bad project) from the MTA, in what looks like the usual agency turf battles, except now given a partisan spin. I doubt there’s going to be any money for Reconstruction (budgeted at $7 billion), let alone expansion (budgeted at $17 billion), and overall this looks like the usual promises that nobody intends to act upon. The problem is that this project is still lurking in the background, waiting for someone insane enough to say what not a lot of people think but few are willing to openly disagree with and find some new source of money to redirect there. And oddly, this makes me think of tariffs.
The commonality is that free trade is not just good, but is more or less an unmixed blessing. In public transport rolling stock procurement, the costs of tariffs are so high that a single job created in the 2010s cost $1 million over 4-6 years, paying $20/hour. In infrastructure, in theory most costs are local and so it shouldn’t matter, but in practice some materials need to be imported, and when they run into trade barriers, they mess entire construction schedules. Boston’s ability to upgrade commuter rail stations with high platform was completely lost due to successive tightening of the Buy America waiver process under Trump and then Biden, to the point that even materials that were just not made in America (steel, FRP) could not be imported. The problem is that nobody was willing to say this out loud, and instead politicians chose to interfere with bids to get some photo-ops, getting trains that are overpriced and fail to meet schedule and quality standards.
Thus, the American turn away from free trade, starting with Trump’s 2016 campaign. During the Obama-Trump transition, the FTA stopped processing Buy America waivers, as a kind of preemptive obedience to something that was never written into the law, which includes several grounds for waivers. During the Trump-Biden transition, the standards were tightened, and waivers required the approval of a political office at the White House, which practiced a hostile environment, hence the above example of the MBTA’s platform problems. Now there are general tariffs, at a rate that changes frequently with little justification. The entire saga, especially in the transit industry, is a textbook example not just of comparative advantage, but of the point John Williamson made in the original Washington Consensus that trade barriers were a net negative to the country that imposes them even if there’s no retaliation, purely from the negative effects on transparency and government cleanliness. This occurred even though tariffs were not favored in the political elite of the United States, or even in the general public; but nobody would speak out except special interests and populists who favored trade barriers.
And Penn Expansion looks the same. It’s an Amtrak turf game, which NJ Transit and the MTA are indifferent to. NJ Transit’s investment plan is not bad and focuses on actual track-level improvements on the surface. The MTA has a lot of problems, including the desire for Penn Reconstruction, but Penn Expansion is not among them. The sentiments I’m getting when I talk to people in that milieu is that nobody really thinks it’s going to happen, and as a result most people don’t think it’s important to shoot down what is still a priority for Amtrak managers who don’t know any better.
The problem is that when the explicit argument isn’t made, the political system gets the message that Penn Expansion is not necessarily bad, but now is not the time for it. It will not invest in alternatives. (On tariffs, the alternative is to repeal Buy America.) It will not cancel the ongoing design work, but merely prolong it by demanding more studies, more possibilities for adding new tracks (seven? 12? Any number in between?). It will insist that any bounty of money it gets go toward more incremental work on this project, and not on actually useful alternatives for what to do with $17 billion.
This can go on for a while until some colossally incompetent populist of the type that can get elected mayor or governor in New York, or perhaps president, decides to make it a priority. Then it can happen, and $17 billion plus future escalation would be completely wasted, and further investment in the system would suffer because everyone would plainly see that $17 billion buys next to nothing in New York so what’s the point in spending a mere $300 million here and there on a surface junction? If it were important then Amtrak would have prioritized that, no? Even people who get on some level that the agencies are bad with money will believe them on technical matters like scheduling and cost estimation over outsiders, in the same manner that LIRR riders think the LIRR is incompetent and also has nothing to learn from outsiders.
The way forward is to be more formal about throwing away bad ideas. Does Penn Expansion have any transportation value? No. So cancel it. Drop it from the list of Northeast Corridor projects, cancel all further design work, and spend about 5 orders of magnitude less money on timetabling trains at Penn Station within its existing footprint. Don’t let it lurk in the background until someone stupid enough decides to fund it; New York is rather good lately at finding stupid people and elevating them to positions of power. And learn to make affirmative arguments for this rather than the usual “it will just never happen” handwringing.
Quick Note: Report on Electrification and Medium-Speed Rail Upgrades
Nolan Hicks has wrapped up nearly a year of work at Marron on a proposal called Momentum, to upgrade mainline rail in the United States with electrification, high platforms, and additional tracks where needed, short of high-speed rail. The aim is to build low- or perhaps medium-speed rail; the proposed trip times are New York-Albany in 2:05 (averaging 109 km/h) and New York-Buffalo in 5:38 to 5:46 (averaging 123 km/h). The concept is supposed to be used US-wide, but the greatest focus is on New York State, where the plan devotes a section to Network West, that is New York-Buffalo, and another to Network East, that is the LIRR, in anticipation of the upcoming state budget debate.
The costs of this plan are high. Nolan projects $33-35.6 billion for New York-Buffalo, entirely on existing track. The reasoning is that his cost estimation is based on looking at comparable American projects, and there aren’t a lot of such upgrades in the US, so he’s forced to use the few that do exist. A second track on single-track line is costed cheaply with references to various existing projects (in Michigan, Massachusetts, etc.), but third and fourth tracks on a double-track line like the Water Level Route are costed at $30 million/km, based on a proposal in the built-up area of Chicago to Michigan City.
In effect, the benefits are a good way of seeing what upgrades to best American industry practices would do. The idea, as with the costing, is to justify everything with current or past American plans, and the sections on the history of studies looking at electrification projects are indispensable. This covers both intercity and regional rail upgrades, and we’ve used some of the numbers in the drafts at ETA to argue, as Nolan does, against third rail extensions and in favor of catenary on the LIRR and Metro-North.
(Update 4-3: and now the full proposal is out, see here.)
Cos Cob Bridge Replacement
The Northeast Corridor has eight movable bridges in Connecticut; other than one that was replaced in the 1990s, all are considered by Amtrak and Connecticut DOT to be both critical priorities for replacement and also major undertakings. The Bipartisan Infrastructure Law funded two, on the Connecticut and the Norwalk Rivers. The costs are enormous, beyond any justification: the Walk Bridge replacement is funded at $1 billion for a four-track bridge of 200 meters, and the replacement will still be a movable bridge rather than a fixed span with enough clearance below for boat traffic. The cost can be compared with an order of magnitude of tens of millions of dollars for comparable or longer bridges, for examples $50 million for one of the Rhone bridges on the LGV Méditerranée and $32 million for an 800 m viaduct on the Erfurt-Nuremberg line.
The goal of this post is to focus on the Cos Cob Bridge on the Mianus River. Among the eight bridges, it is the one with the least advanced plans for rehabilitation, such that no cost figure is given, but rumors put it in the mid-single digit billions for a viaduct of about 1 km, crossing about 250 m of water. Among the bridges west of New Haven, it is also the one with the most constrained alignment making replacement more necessary to fix the right-of-way geometry: the bridge itself is straight but flanked by two short, sharp curves, and replacement should be bundled with a wider curve.
The NEC Webtool outlines one alignment, with a wide curve, 2,400 meters in radius. The snag is the vertical alignment. The bridge needs to be high enough to clear boat traffic below; I-95 slightly upriver has a clearance below of 14.9 meters (Wikipedia says 21 meters but that’s the top of the deck, not the bottom), and with a typical deck thickness of 1.5 meters it means top of rail needs to be about 16.5 meters above sea level – but the Riverside station 450 meters east of the midpoint of the river has top of rail 10 meters above sea level and the Cos Cob station under the I-95 overpass 450 meters west of the midpoint is 8 meters above sea level. To build it as a high span thus requires rising 8.5 meters over 450 meters.
The current Northeast Corridor plans hew to a much lower ruling grade. The Walk Bridge is being replaced with another movable bridge and not a high fixed span because the standards call for a 1% grade. This is, frankly, dumb. The passenger trains are electric, either commuter rail EMUs or powerful intercity trains capable of climbing 4% grades over a short section, even the medium-speed Northeast Regionals. The freight trains are long enough that what matters isn’t so much the maximum grade as the maximum grade averaged over the length of a train, in which case peaking at 4% over a length of 450 meters is not at all problematic.
With a 4% standard, the question is not about the grade, but about the vertical curve radius. Standards for those are tighter than for horizontal curves. Vertical and horizontal curve radii both follow the formula ar = v^2, but the acceleration limit a is much tighter since there is no tilting or superelevation, and on a crest a high speed also reduces the effective weight acceleration and thus reduces train stability. In Germany, a is limited to 0.482 on a crest and 0.594 on a hallow, both requiring special permission; in Sweden, the German crest value is the minimum limit, with no special dispensation on a hallow. The upshot is that at 250 km/h, the exceptional vertical curve radius is 10,000 m and thus it takes 400 meters just to get to 4%; over a length of 450 meters, the maximum average gradient is 1.125% if the higher acceleration rate on a hallow isn’t used or 1.25% if is and the tracks can only rise respectively 5 or 5.5 meters. To make it 8.5, the speed limit needs to be reduced: at 200 km/h, the vertical curve radius is 6,400 meters and then over 225 meters the trains can get up to 3.5% and, if it’s symmetric, over 450 they can climb 7.9 meters, and if it’s asymmetric then they can climb more than the required 8.5%. It’s dirty but it does work.
The issue is then how this affects construction. I don’t know why the Connecticut bridge replacements are so expensive, beyond the observation that everything in Connecticut is exceptionally expensive, usually even by the standards of other Northeastern American rail projects (for example, infill stations), let alone European ones. The local press articles talk about staging construction to avoid disturbing the running track, and if this is the main difficulty, then building a new bridge 50 meters upriver should be much easier, since then the only part of the project interfacing with the existing track is the track connections on firma.
Whatever it is, a multi-billion dollar pricetag is not believable given the required scope. More difficult construction has been done for two orders of magnitude less on this side of the Pond. On a different mode but in the same region, the 10-lane 1.4 km long Q Bridge cost $554 million, around $790 million today, which, relative to the size of the bridge, is still around an order of magnitude cheaper than Walk and more than an order of magnitude cheaper than what Cos Cob is rumored to be.
16-Car Trains on the Northeast Corridor
The dominant length of high-speed rail platforms in China, Japan, South Korea, and Europe is 400 meters, which usually corresponds to 16-car trains. The Northeast Corridor unfortunately does not run such long trains; intercity trains on it today are usually eight cars long, and the under construction Avelia Liberty sets are 8.5 cars long. Demand even today is high enough that trains fill even with very high fares, and so providing more service through both higher frequency and longer trains should be a priority. This post goes over what needs to happen to lengthen the trains to the global norm for high-speed rail. More trains need to be bought, but also the platforms need to be lengthened at many stations, with varying levels of difficulty.
The station list to consider is as follows:
- Boston South Station
- Providence
- New London-HSR
- New Haven
- Stamford
- New York Penn Station
- Newark Penn Station
- Trenton
- Philadelphia 30th Street
- Wilmington
- Baltimore Penn Station
- BWI
- Washington Union Station
Some of these are local-only stations – the fastest express trains should not be stopping at New London or BWI, and whether any train stops at Stamford or Trenton is a matter of timetabling (the headline timetable we use includes Stamford on all trains but I am not wedded to it). In order, allowing 16-car trains at these stations involves the following changes.
Boston
South Station’s longest platforms today are those between tracks 8 and 9 and between tracks 10 and 11, both 12 cars long. To their immediate south is the interlocking, so lengthening would be difficult.
Moreover, the best platforms for Northeast Corridor trains to use at South Station are to the west. The best way to organize South Station is as four parallel stations, from west to east (in increasing track number order) the Worcester Line, the Northeast Corridor and branches, the Fairmount Line, and the Old Colony Lines, with peak traffic of respectively 8, 12 or 16, 4 or 8, and 6 trains per hour. This gives the Northeast Corridor tracks 4-7 or possibly 4-9; 4-7 means the Franklin Line has to pair with the Fairmount Line to take advantage of having more tracks, and may be required anyway since pairing the Franklin Line with the Northeast Corridor (Southwest Corridor within the city of Boston) would constrain the triple-track corridor too much, with 12 peak commuter trains and 4 peak intercity trains an hour.
The platform between tracks 6 and 7 is 11 cars long, but to its south is a gap in the tracks as the interlocking leads tracks 6 and 7 in different directions, and thus it can be lengthened to 16 cars within its footprint. The platform between tracks 4 and 5 is harder to lengthen, but this is still doable if the track that tracks 5 and 6 merge into south of the station is moved in conjunction with a project to lengthen the other platform.
Of note, the other Boston station, Back Bay, is rather constrained, with nearly the entire platforms under an overbuild, complicating any rebuild.
Providence
Providence has 12-car platforms. The southern edge is under an overbuild with rapid convergence between the tracks and cannot reasonably be extended. But the northern edge is in the open air, and lengthening is possible. The northern edge would be on rather tight curves, which is not acceptable under most standards, but in such a constrained environment, waivers are unavoidable, as is the case throughout urban Germany.
New London
This is a new station and can be built to the required length from the start.
New Haven
The current station platforms are only 10 cars long, but there is space to expand them in both directions. The platform area is in effect a railyard, a good example of the American tradition in which the train station is not where the trains are (as in Europe) but rather next to where the trains are.
A rebuild is needed anyway, for two reasons. First, it is desirable to build a bypass roughly following I-95 to straighten the route beginning immediately north of the station, even cutting off State Street in order to go straight to East Haven rather than curve to the north as on the current route. And second, the current usage of the station is that Amtrak uses tracks 1-4 (numbered west to east as in Boston) and Metro-North uses tracks 8-14, which forces Amtrak and Metro-North trains to cross each other at grade from their slow-fast-fast-slow pattern on the running line to the fast-fast-slow-slow pattern at the station. In the future, the station should be used in such a way that intercity trains either divert north to Hartford or Springfield or go immediately east on a flying junction to the high-speed bypass toward Rhode Island, without opposed-direction flat junctions; the flying junction is folded into the cost of the bypass and dominates the cost of rebuilding the platforms, as the space immediately north and south of the platforms is largely empty.
Stamford
Stamford has 12-car platforms. Going beyond that is hard, to the point that a more detailed alternatives analysis must include the option of not having intercity trains stop there at all, and instead running 12-car express commuter trains, lengthening major intermediate stops like South Norwalk (currently 10 cars long) and Bridgeport (currently 8) instead.
To keep the mainline option of stopping at Stamford, a platform rebuild is needed, in two ways. First, the station today has five tracks, a both literally and figuratively odd number, not useful for any timetable, with the middle track, numbered 1 (from north to south the numbers are 5, 3, 1, 2, 4), not served by a platform. And second, the platform between tracks 3 and 5 can at best be lengthened to 14 cars, while that between tracks 2 and 4 cannot be lengthened without moving tracks on viaducts. This means that some mechanism to rebuild the station should be considered, to create four tracks with more space between them so that 16-car platforms are viable; this should be bundled with a flying junction farther east to grade-separate the New Canaan Branch from the mainline.
A quick-and-dirty option, potentially viable here but almost nowhere else, is selective door opening, at the cost of longer dwell times. Normally selective door opening should not be used – it confuses passengers, for one. However, here it may be an option, as intercity traffic here is unlikely to be high; traffic today is 323,791 in financial 2023, the lowest of any station under consideration in this post unless one counts New London. The only reason to stop here in the first place is commuter ridership, in which case mechanisms such as restricting unreserved seats to the central 12 cars can be used.
New York
Penn Station has multiple platforms already long enough for 16- and even 17-car trains, including the one we pencil for all high-speed intercity trains in the proposal, platform 6 between tracks 11 and 12, as well as the two adjacent platforms, 5 and 7. (Note that unlike at New Haven and Boston, platform numbers at Penn increase south to north, that is right to left from the perspective of a Boston-bound traveler.)
Thank the god of railways, since platform expansion requires a multi-billion dollar project to remove the Madison Square Garden overbuild in the most optimistic case; in a more pessimistic case, it would also require removing the Moynihan Station overbuild.
Newark
Newark Penn Station’s platforms are in a grand structure about 14.5 cars long. Thankfully, they extend a bit south of it, producing about 16 cars’ worth of platform on the west (southbound) side, between tracks 3 and 4; as in New York, track numbers increase east to west. On the east side, PATH interposes between the two tracks, which have a cross-platform transfer from northbound New Jersey Transit trains to PATH. The platform structures and their extensions do have enough length to allow 16-car trains – indeed they go as long as 18 – but the southern ends are currently disused and would require some rehabilitation.
Trenton
Trenton has a 12.5 car long southbound platform and an 11.5 car long northbound platform. There is practically no room for an expansion if no tracks are moved. If tracks are moved, then some space can be created, but only enough for about 14 cars, not 16.
However, traffic is low, the second lowest among stations under consideration next to Stamford. The suite of Stamford solutions is thus most appropriate here: selective door opening with only the middle 12 cars (naturally the same as at Stamford) open to commuters, or just not stopping at this station at all. The only reason we’re even considering stopping here is timetabling-related: trains should be running every 10 minutes around New York but every 15 between Baltimore and Washington, or else significant expansion of quad-tracking on the Penn Line is required, and so a local stop should be added as a buffer, which can be Trenton or BWI, and BWI has twice the current Amtrak traffic of Trenton.
Philadelphia
30th Street Station has 14-car platforms. Selective door opening is basically impossible given the high expected traffic at this station, and instead platform expansion is required. There is an overbuild, but the tracks stay straight and only begin curving after a few tens of meters, which gives room for extension; from the north end to the overbuild to where the tracks begin curving toward one another to the south is 15.5 cars, and there is room north of the overbuild between the tracks.
Whatever reconstruction project is needed is helped by the low traffic at these platforms. SEPTA uses the upper level of the station, with tracks oriented east-west. The north-south lower level is only used by Amtrak, which could be easily reduced to three platform tracks (two Northeast Corridor, one Keystone) if need be, out of 11 today. Thus, staging construction can be done easily and intrusively, with no care taken to preserve track access during the work, as half the station platforms can be closed off at once.
Wilmington
Wilmington is frustrating, in that there is platform space for 16 cars rather easily, but it’s on inconsistent sides of the tracks. Track numbers increase south to north; track 1 has a side platform, there’s an island platform between tracks 2 and 3, and then track 3 also has a side platform on the other side, extending well to the east of the island platform. The island platform and the track 1 platform are about 12.5 cars long, and the track 3 side platform is 13.5 cars long. Thus, an extension, selective door opening, or a station rebuild is required.
The island platform can be extended about one car in each direction, so it cannot be the solution without selective door opening. Both side platforms can be extended somewhat to the west: the track 1 platform can be extended to 16 cars, but it would need to be elevated in the narrow space between the track viaduct and the station parking garage; the track 3 platform can be extended in both directions, avoiding a new elevated extension over North King Street.
If for some reason an extension of the track 1 platform is not possible, then selective door opening can be used, but not as reliably as at lower-traffic Stamford or Trenton, and overall I would not recommend this solution. A station rebuild then becomes necessary: the station has three tracks but doesn’t need more than two if SEPTA and Amtrak can be timetabled right, and then the removal of either track 1 or track 2 would create space for a longer platform.
Baltimore
Baltimore Penn has seven tracks, numbered from south to north 1, 3, 4, 5, 6, 7, F. Their platforms are 10 to 13 cars long. Northbound trains are more or less forced to use the platform between tracks 1 and 3, since the way the route tapers to a three-, then four-track line to the east forces all eastbound trains to use mainline track 1; this platform is rather narrow at its east end but has space to the west for a 16-car extension. Westbound trains can use either the platform between tracks 4 and 5 or that between tracks 6 and 7, with tracks 4 and 6 preferred over 7 as they reach the express westbound track (track 5 stub-ends). Both platforms can be extended, with the platform between tracks 6 and 7 requiring a one-car extension to the east where a ramp down to track level for track workers exists whereas that between tracks 4 and 5 has ample unused space to its west.
BWI
The two side platforms at BWI are just under 13 cars long. However, nowhere else on the corridor is an extension easier: the station is located in an undeveloped wooded area, with space cleared on both sides of the track so that tree cutting is likely unnecessary west of the tracks and certainly unnecessary east of them.
The station itself needs a rebuild anyway, due to already existing plans to widen it from three to four tracks. This is required to enable intercity trains to overtake commuter trains anyway, unless delicate timetabling on triple track is used or another part of the Penn Line is set up as a four-track overtake. The plans are rather advanced, but platform extensions can be pursued as an add-on, without disturbing them due to the easy nature of the right-of-way.
Washington
Washington is set up as two separate stations, a high-platform terminal to the west and a low-platform through-station to the east on a lower level. Track numbers increase west to east, the western part taking 7-20 (though only 9-20 are high and wired) and the eastern part 23-30. None of the western platforms is long enough, but multiple options still exist:
- The platform between tracks 9 and 10 has room for an extension.
- The platforms between tracks 15 and 16 and between tracks 16 and 17 look like they already have extensions, if not open for passengers.
- The platforms between track 17 and track 18 and between tracks 19 and 20 are only 12 cars long, but tracks could be cannibalized in the open air to make a long enough platform, especially since the reason track numbers 21 and 22 are skipped is that there used to be tracks there and now there’s empty space.
- The platform between tracks 25 and 26 is long enough, and could be raised to have level boarding.
The existing platforms that can be extended easily are sufficient in number, but probably not in location – it’s ideal for the platforms to be close together, to simplify the interlocking as trains have to be scheduled to enter and leave the station without opposite-direction conflicts. If it’s doable even with a split between platforms separated by multiple tracks then it’s ideal, but otherwise, the extra work on tracks 17-20 may be necessary, converting a part of the station that presently has six tracks and four platforms into likely four tracks and two platforms.
Conclusion
All of this looks doable. The hardest station, Stamford, is skippable if selective door opening is unviable after all and a rebuild is too expensive. Among the other stations, light rebuilds are needed at Boston, Wilmington, and maybe Washington; New Haven needs a more serious rebuild as part of the bypass, but the station platforms are a routine extension where there is already room between the tracks. The most untouchable station, New York, already has multiple platforms of the required length at the required location within the station.
Northeast Corridor Profits and Amtrak Losses
In response to my previous post, it was pointed out to me that Amtrak finances can’t really be viewed in combination, but have to be split between the Northeast Corridor, the state-supported routes, and the long-distance trains. Long-distance is defined by a 750 mile (1,200 km) standard, comprising the night trains plus the Palmetto; these trains have especially poor financial performance. The question is what level of Northeast Corridor profitability is required to cover those losses.
In financial 2024 (ending 2024-09-30), Amtrak finances per route category were as follows, in millions of dollars or passenger-km or in dollars per p-km:
| Category | Ridership | P-km | Cost | Cost/p-km | Revenue | Revenue/p-km |
| NEC | 14 | 4,053.3 | 1,146.8 | 0.283 | 1,414.6 | 0.349 |
| State-supported | 14.5 | 2,972.6 | 1,110.7 | 0.374 | 859.2 | 0.289 |
| Long-distance | 4.3 | 3,505.8 | 1,261.2 | 0.36 | 626.1 | 0.179 |
The long-distance trains don’t actually have higher cost structure than the state-supported ones. Their greater losses are because fares are degressive in distance, and so the longer distances traveled translate to lower revenue per kilometer. This is also observable on some high-speed routes in Europe – the fares on TGVs using the LGV Sud-Est are very degressive, with little premium on Paris-Nice over Paris-Lyon despite the factor of 2.5 longer distance and factor of almost 3 longer time.
Revenue per passenger-km in France and Germany is around $0.15, as I explain in this post with links, and revenue per passenger-km in Japan is $0.25, both with average trip lengths similar to those of the Northeast Corridor and state-supported trains. Getting operating costs for just high-speed trains in France and Germany is surprisingly tricky; the Spinetta report says the TGV costs 0.06€/seat-km without capital, which at current seat occupancy is around 0.08€/p-km or around $0.11/p-km.
The upshot is that Northeast Corridor profits need to be $886.6 million a year to cover losses elsewhere, and if the operating costs on the corridor were the same as on the TGV, this could be achieved now with no further increases in service.
Now, in reality, high-speed rail would both massively increase ridership and also have to involve reducing fares to more normal levels than $0.35/p-km. If the revenue is $0.15/p-km and the cost is $0.11/p-km, then traffic in p-km has to rise to 22.165 billion/year, a fivefold increase, to cover. This is less implausible than it sounds – my gravity-based ridership model predicts about that ridership. Potentially, operating costs could be lower than on the TGV, if the entire corridor is (relatively) fast, with no long sections on slow lines as in France, and if traffic is less peaky than in France. But to first order, the answer to the profits question should be “probably but not certainly.”
Amtrak’s Failure
An article in Streetsblog by Jim Mathews of the Rail Passengers Association talking up Amtrak as a success has left a sour taste in my mouth as well as those of other good transit activists. The post says that Amtrak is losing money and it’s fine because it’s a successful service by other measures. I’ve talked before about why good intercity rail is profitable – high-speed trains are, for one, and has a cost structure that makes it hard to lose money. But even setting that aside, there are no measures by which Amtrak is a successful, if one is willing to look away from the United States for a few moments. What the post praises, Amtrak’s infrastructure construction, is especially bad by any global standard. It is unfortunate that American activists for mainline rail are especially unlikely to be interested in how things work in other parts of the world, and instead are likely to prefer looking back to American history. I want to like the RPA (distinct from the New York-area Regional Plan Association, which this post will not address), but its Americanism is on full display here and this blinds its members to the failures of Amtrak.
Amtrak ridership
The ridership on intercity rail in the United States is, by most first-world standards, pitiful. Amtrak reports, for financial 2023, 5.823 billion passenger-miles, or 9.371 billion p-km; Statista gives it at 9.746 billion p-km for 2023, which I presume is for calendar 2023, capturing more corona recovery. France had 65 billion p-km on TGVs and international trains in 2023.
More broadly than the TGV, Eurostat reports rail p-km without distinction between intercity and regional trains; the total for both modes in the US was 20.714 billion in 2023 and 30.89 billion in 2019, commuter rail having taken a permanent hit due to the decline of its core market of 9-to-5 suburb-to-city middle-class commuting. These figures are, per capita, 62 and 94 p-km/year. In the EU and environs, only one country is this low, Greece, which barely runs any intercity rail service and even suspended it for several months in 2023 after a fatal accident. The EU-wide average is 955 p-km/year. Dense countries like Germany do much better than the US, as do low-density countries like Sweden and Finland. Switzerland has about the same mainline rail p-km as the US as of 2023, 20.754 billion, on a population of 8.9 million (US: 335 million).
So purely on the question of whether people use Amtrak, the answer is, by European standards, a resounding no. And by Japanese standards, Europe isn’t doing that great – Japan is somewhat ahead of Switzerland per capita. Amtrak trains are slow: the Northeast Corridor is slower than the express trains that the TGV replaced, and the other lines are considerably slower, running at speeds that Europeans associate with unmodernized Eastern European lines. They are infrequent: service is measured in trains per day, usually just one, and even the Northeast Corridor has rather bad frequencies for the intensely used line it wants to be.
Is this because of public support?
No. American railroaders are convinced that all of this is about insufficient public funding, and public preference for highways. Mathews’ post repeats this line, about how Amtrak’s 120 km/h average speeds on a good day on its fastest corridor should be considered great given how much money has been spent on highways in America.
The issue is that other countries spend money on highways too. High American construction costs affect highway megaprojects as well, and thus the United States brings up the rear in road tunneling. The highway competition for Amtrak comprises fairly fast, almost entirely toll-free roads, but this is equally true of Deutsche Bahn; the competition for SNCF and Trenitalia is tollways, but then those tollways are less congested, and drivers in Italy routinely go 160 km/h on the higher-quality stretches of road.
Amtrak itself has convinced itself that everyone else takes subsidies. For example, here it says “No country in the world operates a passenger rail system without some form of public support for capital costs and/or operating expenses,” mirroring a fraudulent OIG report that compares the Northeast Corridor (alone) to European intercity rail networks. Technically it’s true that passenger rail in Europe receives public subsidies; but what receives subsidies is regional lines, which in the US would never be part of the Amtrak system, and some peripheral intercity lines run as passenger service obligation (PSO) with in theory competitive tendering, on lines that Amtrak wouldn’t touch. Core lines, equivalent to Chicago-Detroit, New York-Buffalo, Washington-Charlotte-Atlanta, Los Angeles-San Diego, etc., would be high-speed and profitable.
But what about construction?
What offends me the most about the post is that it talks up Amtrak’s role as a construction company. It says,
Today, our nationalized rail operator is also a construction company responsible for managing tens of billions of dollars for building bridges, tunnels, stations, and more – with all the overhead in project-management staff and capital delivery that this entails.
The problem is that Amtrak is managing those tens of billions of dollars extremely inefficiently. Tens of billions of dollars is the order of magnitude that it took to build the entire LGV network to day ($65.5 billion in 2023 prices), or the entire NBS network in Germany ($68.6 billion). Amtrak and the commuter rail operators think that if they are given the combined cost to date of both networks, they can upgrade the Northeast Corridor to be about as fast as a mixed high- and low-speed German line, or about the fastest legacy-line British trains (720 km in 5 hours).
The rail operations are where Amtrak is doing something that approximates good rail work – lots of extraneous spending, driving up Northeast Corridor operating costs to around twice the fares on German and French high-speed trains, probably around 3-4 times the operating costs on those trains. But capital construction is a bundle of bad standards for everything, order-of-magnitude cost premiums, poor prioritization, and agency imperialism leading Amtrak to want to spend $16 billion on a completely unnecessary expansion of Penn Station. The long-term desideratum of auto-tensioned (“constant-tension”) catenary south of New York, improving reliability and lifting the current 135 mph (217 km/h) speed limit, would be a routine project here, reusing the poles with their 75-80 meter spacing; an incompetent (since removed) Amtrak engineer insisted on tightening to 180′ (54 m) so the project is becoming impossibly expensive as the poles have to be replaced during service. “Amtrak is also doing construction” is a derogatory statement about Amtrak.
Why are they like this?
Americans generally resent having to learn about the rest of the world. This disproportionately affects industries where the United States is clearly ahead (for example, software), but also ones where internal American features incline Americans to overfocus on their own internal history. Railroad history is rich everywhere, and the relative decline of the railway in favor of the highway lends itself to wistful alternative history, with intense focus on specific lines or regions. New Yorkers are, in the same vein, atypically provincial when it comes to the subway’s history, and end up making arguments, such as about the difficulty of accessibility retrofits on an old system, that can be refuted by looking at peer American systems, not just foreign ones.
The upshot is that an industry and an advocacy ecosystem that both intensely believe that railroad decline was because government investment favored roads – something that’s only partly true, since the same favoring of roads happened more or less everywhere – will want to learn from their own local histories. Quite a lot of advocacy by the RPA falls into the realm of trying to revive the intercity rail system the US had in the 1960s, before the bankruptcies and near-bankruptcies that led to the creation of Amtrak – but this system was what lost out to highways and cars to begin with. The innovations that allowed East Asia to avoid the same fate, and the innovations that allowed Western Europe to partly reverse this fate, involve different ideas of how to build and operate intercity rail.
And all of this requires understanding that, on a basic level, Amtrak is best described as a mishmash of the worst features of every European and East Asian railway: speed, fares, frequency, reliability, coverage. Each country that I know of misses on at least one of these aspects – Swiss trains are slow, the Shinkansen is expensive, the TGV has multi-hour midday gaps, German trains barely run on a schedule, China puts its train stations at inconvenient locations. Amtrak misses on all of those, at once.
And while Amtrak misses on service quality in operations, it, alongside the rest of the American rail construction industry, practically defines bad capital planning. Cities can build the right project wrong, or build the wrong project right, or have poor judgment about standards but not project delivery or the reverse, and somehow, Amtrak’s current planning does all of these wrong all at once.
Amtrak Doubles Down on False Claims About Regional Rail History to Attack Through-Running
Amtrak just released its report a week and a half ago, saying that Penn Expansion, the project to condemn the Manhattan block south of Penn Station to add new tracks, is necessary for new capacity. I criticized the Regional Plan Association presentation made in August in advance of the report for its wanton ignorance of best practices, covering both the history of commuter rail through-running in Europe and the issue of dwell times at Penn Station. The report surprised me by making even more elementary mistakes on the reality of how through-running works here than the ones made in the RPA presentation. The question of dwell times is even more important, but the Effective Transit Alliance is about to release a report addressing it, with simulations made by other members; this post, in contrast, goes over what I saw in the report myself, which is large enough errors about how through-running works that of course the report sandbags that alternative, less out of malice and more out of not knowing how it works.
Note on Penn Expansion and through-running
In the regional discourse on Penn Station, it is usually held that the existing station definitely does not have the capacity to add 24 peak trains per hour from New Jersey once the Gateway tunnel opens, unless there is through-running; thus, at least one of through-running and Penn Expansion is required. This common belief is incorrect, and we will get into some dwell time simulations at ETA.
That said, the two options can still be held as alternatives to each other, even as what I think is likeliest given agency turf battles and the extreme cost of Penn Expansion (currently $16 billion) is that neither will happen. This is for the following reasons:
- Through-running is good in and of itself, and any positive proposal for commuter rail improvements in the region should incorporate it where possible, even if no dedicated capital investment such as a Penn Station-Grand Central connection occurs. This includes the Northeast Corridor high-speed rail project, which aims to optimize everything to speed up intercity and commuter trains at minimal capital cost.
- The institutional obstacles to through-running are mainly extreme incuriosity about rest-of-world practices, which are generations ahead of American ones in mainline rail; the same extreme incuriosity also leads to the belief that Penn Expansion is necessary.
- While it is possible to turn 48 New Jersey Transit trains per hour within the current footprint of Penn Station with no loss of LIRR capacity, there are real constraints on turnaround times, and it is easier to institute through-running.
The errors in the history
The errors in the history are not new to me. My August post criticizing the RPA still stands. I was hoping that Amtrak and the consultants that prepared the report (WSP, FX) would not stick to the false claim that it took 46 years to build the Munich S-Bahn rather than seven, but they did. The purpose of this falsehood in the report is to make through-running look like a multigenerational effort, compared with the supposedly easier effort of digging up an entire Manhattan block for a project that can’t be completed until the mid-2030s at the earliest.
In truth, as the August post explains, the real difficulties with through-running in the comparison cases offered in the report, Paris and Munich, were with digging the tunnels. This was done fairly quickly, taking seven years in Munich and 16 in Paris; in Paris, the alignment, comprising 17 km of tunnel for the RER A and 2 for the initial section of the RER B, was not even finalized when construction began. The equivalent of these projects in New York is the Gateway tunnel itself, at far higher cost. The surface improvements required to make this work were completed simultaneously and inexpensively; most of the ones required for New York are already on the drawing board of New Jersey Transit, budgeted in the hundreds of millions rather than billions, and will be completed before the tunnel opens unless the federal government decides to defund the agency over several successive administrations.
The errors in present operations
The report lists, on printed-pp. 40-41, some characteristics of the through-running systems used in Paris, Munich, and London. Based on those characteristics, it concludes it is not possible to set up an equivalent system at Penn Station without adding tracks or rebuilding the entire track level with more platforms. Unfortunately for the reputation of the writers of the report, and fortunately for the taxpayers of New York and New Jersey, those characteristics include major mistakes. There’s little chance anyone in the loop understands the RER, any S-Bahn worth the name, or even Crossrail and Thameslink; some of the errors are obviously false to anyone who regularly commuted on any of these systems. Thus, they are incapable of adjusting the operations to the specifics of Penn Station and Gateway.
Timetabling
A key feature of S-Bahn systems is that the trains run on a schedule. Passengers riding on the central trunk do not look at the timetable, but passengers riding to a branch do. I memorized the 15-minute off-peak Takt on the RER B when I took it to IHES in late 2016, and the train was generally on time or only slightly delayed, never so delayed that it was early. Munich-area suburbanites memorize the 20-minute Takt on their S-Bahn branch line. Some Thameslink branches drop to half-hourly frequency, and passengers time themselves to the schedule while operators and dispatchers aim to make the schedule.
And yet, the report repeatedly claims that these systems run on headway management. The first claim, on p. 40, is ambiguous, but the second, on the table on p. 41, explicitly contrasts “headway-based” with “timetable-based” service and says that Crossrail, the RER, and the Munich S-Bahn are headway-based. In fact, none of them is.
This error is significant in two ways. First, timetable-based operations explain why S-Bahn systems are capable of what they do but not of what some metros do. The Munich S-Bahn peaks at 30 trains per hour, with one-of-a-kind signaling; major metros peak at 42 trains per hour with driverless operations, and some small operations with short trains (like Brescia) achieve even more. The difference is that commuter rail systems are not captive metro trains on which every train makes the same stops, with no differentiation among successive trains on the same line; metro lines that do branch, such as M7 and M13 in Paris, are still far less complex than even relatively simple and metro-like lines like the RER A and B. The main exception among world metros is the New York City Subway, which, due to its extensive interlining, must run as a scheduled railroad, benchmarking its on-time performance (OTP) to the schedule rather than to intervals between trains. In the 2000s and 10s, New York City Transit tried to transition away from end-station OTP and toward a metric that tried to approximate even intervals, called Wait Assessment (WA); a document leaked to Dan Rivoli and me went over how this was a failure, leading to even worse delays and train slowdowns, as managers would make the dispatchers hold trains if the trains behind them were delayed.
The second consequence of the error is that the report does not get how crucial timetable-infrastructure planning integration is on mainline rail. The Munich S-Bahn has outer branches that are single-track and some that share tracks with freight, regional, and intercity trains. The 30 tph trunk does no such thing and could not do such thing, but the branches do, because the trains run on a fixed timetable, and thus it is possible to have a mix of single and double track on some sporadic sections. The Zurich S-Bahn even runs trains every 15 minutes at rush hour on a short single-track section of the Right Bank of Lake Zurich Line. Recognizing what well-scheduled commuter trains can and can’t do influences infrastructure planning on the entire surface section, including rail-on-rail grade separations, extra tracks, yard expansions, and other projects that collectively make the difference between a rail network and crayon.
Separation between through- and terminating lines
Through-running systems vary in how much track sharing there is with the rest of the mainline rail network. As far as I can tell, there is always some; near-complete separation is provided on the RER A, but its Cergy branch also hosts Transilien trains running to Gare Saint-Lazare at rush hour, and the Berlin and Hamburg S-Bahn systems have very little track-sharing as well. Other systems have more extensive track sharing, including Thameslink, the RER C and D, and the Zurich S-Bahn; the RER E and the Munich S-Bahn are intermediate in level of separation between those two poles.
It is remarkable that, while the RER A, B, and E all feature new underground terminals for dedicated lines, the situation of the RER C and D is different. The RER C uses the preexisting Gare d’Austerlitz, and has taken over every commuter line in its network; the through-connection between Gare d’Orsay and Gare d’Invalides involved reconstructing the stations, but then everything was connected to it. The RER D uses prebuilt underground stations at Gare du Nord, Les Halles, and Gare de Lyon, but then takes over nearly all lines in the Gare de Lyon network, with the outermost station, Malesherbes, not even located in Ile-de-France. Thameslink uses through-infrastructure built in the 1860s and runs as far as Petersborough, 123 km from King’s Cross on the East Coast Main Line, and Brighton, the terminus of its line, 81 km from London Bridge.
And yet, the report’s authors seem convinced the only way to do through-running is with a handful of branches providing only local service, running to new platforms built separately from the intercity terminal; they’re even under the impression the RER D is like this, which it is not. There’s even a map on p. 45, suggesting a regional metro system running as far as Hicksville, Long Beach, Far Rockaway, JFK via the Rockaway Cutoff and Queenslink, Port Washington, Port Chester, Hackensack, Paterson, Summit, Plainfield, New Brunswick, and the Amboys. This is a severe misunderstanding of how such systems work: they do not arbitrarily slice lines this way into inner and outer zones, unless there is a large mismatch in demand, and then they often just cut the outer end to a shuttle with a forced transfer, as is the case for some branches in suburban Berlin connecting to S-Bahn outer ends. Among the above-mentioned outer ends, the only one where this exception holds is Summit, where the Gladstone Branch could be cut to a shuttle or to trains only running to Hoboken – but then trains on the main line to Morristown and Dover have no reason to be treated differently from trains to Summit.
Were the report’s authors more informed about just the specific lines they look at on p. 41, let alone the broader systems, they’d know that separation between inner and outer services is contingent on specifics of track infrastructure, including whether there are four-track lines with neat separation into terminating express trains and through locals. But even if the answer is yes, as at Gare de Lyon and Gare d’Austerlitz, infrastructure planners will attempt to shoehorn whatever they can into the system, just starting from the more important inner lines, which generate more all-day demand. There don’t even need to be terminating regional trains; the Austerlitz system doesn’t, and the Gare de Lyon and Gare de l’Est systems only do due to trunk capacity limitations. In that case, they’d recognize that there is no need to have two commuter rail systems, one through-running and one not. Penn Station’s infrastructure already lends itself to allowing through-running on anything entering via the existing North River Tunnels.
Branching
S-Bahn systems usually try to keep the branch-to-trunk ratio to a manageable number. Usually, more metro-like systems have fewer branches: Crossrail has two on each side, the RER A has two to the east and three to the west, the Berlin Stadtbahn has two to the west plus short-turns and five to the east, the Berlin North-South Tunnel has three on each side. The Munich S-Bahn has five to the east and nine to the west, and the combined RER B and D system has three to the north and five to the south, but the latter has more service patterns, including local and express trains on the branches. Zurich has so much interlining that it’s not useful to count branches, and better to count services: there are 21 S-numbered routes serving Hauptbahnhof, of which 13 run through one of the two tunnels, as do some intercity trains.
If there are too many branches, then they’re usually organized as sub-branches – for example, Munich has seven numbered routes through the central tunnel, of which two have two sub-branches each splitting far out. Zurich has fewer than 13 branches on each side, but rather there are several services using each line, with inconsistent through-pairing – for example, the three services going to the airport, S2, S24, and S16, respectively run through to two separate branches of the Left Bank Line and to the Right Bank Line.
The table on p. 41 gets the branch count mildly wrong, but the significant is less in what it gets wrong about Europe and more in what it gets wrong about New York. A post-Gateway service plan is one in which New Jersey has 12 branches, but some can be viewed as sub-branches (like Gladstone and the Morristown Line), and more to the point, there are going to be two trunk lines. The current plan at New Jersey Transit is to assign the Northeast Corridor and North Jersey Coast Lines to the North River Tunnels alongside Amtrak, which is technically two branches but realistically four or even five service patterns, and the Morris and Essex, Montclair-Boonton, and Raritan Valley Lines to Gateway, which is four branches but could even be pruned to three with M&E divided into two sub-branches. The Erie lines have no way of getting to Penn Station today; to get them there requires the construction of the Bergen Loop at Secaucus, with an estimated budget of $1.3 billion in 2020, comparable to the total cost of all yet-unfunded required surface improvements in New Jersey for non-Erie service combined.
If the study authors were more comfortably knowledgeable of European S-Bahn systems, they’d know that multi-line systems, while uncommon, do exist, and divide branches in a similar way. The multiline systems (Paris, Madrid, Berlin, Zurich, and London) all have some reverse-branching, in a similar manner to how New York is soon going to have the New Haven Line reverse-branch to Penn Station and Grand Central. The NJT plan is solid and stands to lead to a manageable branch-to-trunk ratio, even with every single line going to Penn Station via the existing tunnel running through.
The consequence of the errors
The lack of familiarity with through-running commuter rail is evident in how the report talks about this technology. It is intimately related to the fact that the way investment should be done is different from what American railroaders are used to. For one, there needs to be much tighter integration between infrastructure and scheduling. For two, the scheduling needs to be massively simplified, with fewer operating patterns per line – usually one, occasionally two, never 13 as on the New Haven Line today. The same ignorance that leads Amtrak and its consultants to assert that the S-Bahn runs on headway management rather than a fixed timetable also leads them not to even know how through-running commuter rail networks plan out their routes and services.
From my position of greater familiarity as both a regular user and a researcher, I can point out that the required investments to make through-running happen in New York are entirely in line with the cheap surface projects done in the comparison cases. New rolling stock is required, with the ability to run on the different voltages of the three networks – but multi-voltage commuter rolling stock is the norm wherever multiple legacy electrification systems coexist, including Paris, London, and Hamburg. Some extensions of electrification and high platform conversions are required – but these are not expensive, and the latter is already partly funded at reasonable unit costs. Some rail-on-rail grade separations are required – but those are already costed and very likely to be funded, potentially out of the Bipartisan Infrastructure Law.
Penn Station would be used as the universal station in this schema, without the separation into a surface terminal and a through- underground station seen in Munich and Paris. But then, Paris and Munich don’t even universally have this separation themselves; Ostbahnhof was reconstructed for the S-Bahn but is still a single station, and the same is true of the RER C. In a way, Penn Station already is the underground through-station, built generations before the modern S-Bahn concept, complementing and largely replacing surface terminals like Hoboken and Long Island City because those are not in Manhattan.
None of this is hard; the hard part is the Gateway tunnel and that’s already fully funded and under construction. But it does require understanding that the United States is so many decades behind best practices that none of what American railroaders think they know is at all relevant. It’s obligatory to understand how the systems that work, in Europe and rich Asia, do, because otherwise, it’s like expecting someone who has never learned to count beyond 10 to prove mathematical theorems. The people who wrote this report clearly don’t have this understanding, and don’t care to get it, which is why what they write is not worth the electrons that make up the PDF.
Reports on High-Speed Rail and the Northeast Corridor
Two reports that I’ve collaborated on are out now, one about high-speed rail planning for Marron and one about Northeast Corridor maintenance for ETA. A third piece is out, not by me but by Nolan Hicks, about constant-tension catenary and its impact on speed and reliability. The context for the latter two pieces is that the Northeast Corridor has been in a recurrent state of failure in the last three weeks, featuring wire failures, circuit breaker failures, track fires, and transformer fires. The high-speed rail planning piece is of different origin – Eric interviewed officials involved in California High-Speed Rail and other American projects that may or may not happen and this led to synthesizing five planning recommendations, which aren’t really about the Northeast Corridor but should be kept in mind for any plan there as well.
The broader context is that we’re going to release another report specific to the Northeast Corridor, one that’s much more synthetic in the sense of proposing an integrated infrastructure and service planning program to cut trip times to about 1:53 New York-Washington and 2:00 New York-Boston, informed by all of these insights. Nolan’s piece already includes one key piece of information that’s come out of this work, about the benefits of constant-tension catenary upgrades: 1:53 requires constant-tension catenary, and if it is not installed, the trip time is 2:04 instead, making this the single biggest piece of physical infrastructure installation the Northeast Corridor needs.
The catenary issue
Trying to go to Philadelphia, I was treated to a train stuck at Penn Station without air conditioning, until finally, after maybe 45 minutes of announcements by the conductor that it would be a while and they’d make announcements if the train was about to move, I and the other passengers got out to the station, waiting for anything to change, eventually giving up as the train and several subsequent ones were canceled. My post from three days ago about Germany has to be read with this context – while publishing I was waiting for all three pieces above to appear.
I encourage people to read the ETA report for more detail about the catenary. In brief, overhead wires can be tensioned by connecting them to fixed places at intervals along the tracks, which leads to variable tension as the wires expand in the heat and contract in the cold; alternatively, they can be tensioned with spring wires or counterweights, which automatically provide constant tension. The ETA report explains more, with diagrams, some taken from Garry Keenor’s book on rail electrification, some made by Kara Fischer (the one who made the New Mexico public transit maps and others I’ll credit upon request, not the USDOT deputy chief of staff). The catenary on the Northeast Corridor has constant tension north of New York, and for a short stretch in New Jersey, but not on the vast majority of the New York-Washington half of the line.
Variable-tension catenary is generally unreliable in the heat, and is replaced with constant-tension catenary on main lines even in Europe, where the annual temperature range is narrower than in the United States. But it also sets a blanket speed limit; on the Northeast Corridor, it is 135 mph, or 217 km/h – the precision in metric units is because 217 km/h is the limiting speed of a non-tilting train on a curve of radius 1,746 meters, a common radius in the United States as it is a round number in American units (it’s 1°, the degree being the inverse of curve radius). This blanket speed limit slows trains by 11 minutes between New York and Washington, subject to the following assumptions:
- The tracks otherwise permit the maximum possible speed based on curvature, up to 320 km/h; in practice, there are few opportunities to go faster than 300 south of New York. There is an FRA rule with little justification limiting trains to 160 mph, or a little less than 260 km/h, on any shared track; the rule is assumed removed, and if it isn’t, the cost is about one minute.
- Trains have the performance of the Velaro Novo, which trainset is being introduced to the United States with Brightline West. Other trainsets may have slightly better or worse performance; the defective Avelia Liberty sets are capable of tilt and therefore the impact of maximum speed is larger.
- Intercity trains make one stop per state, counting the District of Columbia as a state.
- Intercity and regional trains are timetabled together, on a clockface schedule with few variations. If a train cannot meet these requirements, it stays off the corridor, with a forced transfer at Philadelphia or Washington. All train schedules are uniformly padded by 7%, regardless of the type of catenary. If variable-tension catenary requires more padding, then the impact of constant-tension catenary is increased.
The bulk of the difference between 1:53 and the current trip time of about 2:50 is about timetabling, not infrastructure – when the trains are running smoothly, there is extensive schedule padding, in one case rising to 35 minutes south of New York on a fast Regional. Rolling stock quality provides a boost as well, to both reliability and acceleration rates. Faster speeds on curves even without tilt matter too – American standards on this are too conservative, and on a built-out line like the Northeast Corridor, being able to run with 180 mm of cant and 130 mm of cant deficiency (see explanation here) is valuable. But once the regulatory and organizational issues are fixed, the biggest single piece of infrastructure investment required is constant-tension catenary, simultaneously reducing trip times and improving reliability.
Nolan’s piece goes more into costs for catenary repair, and those are brutal. The Northeast Corridor Project Inventory includes $611 million to just replace the catenary between Newark and New Brunswick, without constant-tension upgrades. This is 36.5 route-km, some four- and some six-track; the $16.7 million/cost electrifies a new line from scratch around six times over in non-English-speaking countries, and while the comparison is mostly to double-track lines, around half the cost of electrification is the substations and transformers, and those aren’t part of the project in New Jersey.
State of Good Repair projects always end up as black holes of money, because if half the money is spent and there’s no visible improvement, it’s easy for Amtrak to demand even more money, without having to show anything for it. An improvement project would be visible in higher speeds, better ride quality, higher reliability, and so on, but this is free money in which the cost is treated as a positive (jobs, the appearance of work, etc.) and not something to be minimized in pursuit of another goal. One conclusion of this is that no money should be given to catenary renewal. Money can be spent on upgrades with visible results, in this case constant-tension catenary. On all else, Amtrak cannot be trusted.
High-speed rail planning
The report we wrote on high-speed rail planning at Marron is longer than the ETA report, but I encourage people to read it as well, especially anyone who wishes to comment here. In brief, we give five broad recommendations, based on a combination of reviewing the literature on high-speed rail, cost overruns, and public infrastructure management, and interviewing American sources in the field.
- The federal government needs to nurture local experimentation and support it with in-house federal expertise, dependable funding, and long-term commitment.
- The FRA or another federal entity should have consistent technical standards to ensure scale and a clear operating environment for contractors.
- The federal government should work with universities to develop the technology further, which in this case means importing standards that work elsewhere – high-speed rail in 2024 is a mature technology, not requiring the inventions of new systems that underlay the Japanese, French, and German networks.
- Agencies building high-speed rail should have good project delivery, following the recommendations we gave in the subway construction costs report. Using consultants is unavoidable, but there needs to be in-house expertise, and agencies should avoid being too reliant on consultants or using consultants to manage other consultants.
- Agencies and states should engage in project planning before environmental reviews and before making the decision whether to build; the use of environmental reviews as a substitute for planning leads to rushed designs, which lead to mistakes that often prove fatal to the project.
Currently, all American high-speed rail plans should be treated as case studies of what to avoid. However, this does not mean that all of them fail on all five criteria. For one, California High-Speed Rail largely used pan-European technical standards in its planning; Caltrain did not in related planning including the electrification project and the associated resignaling (originally intended to be the bespoke CBOSS). The criterion on technical standards becomes more important as different projects interact – for example, Brightline West is inconsistent about what it’s using. Then there’s Texas Central, which uses turnkey Shinkansen standards, but as it’s turned over to Amtrak is bound to get modifications that conflict with what Japan Railways considers essential to the Shinkansen, such as total lack of any infrastructure mixing with legacy trains.
Notably, none of this is about the Northeast Corridor directly. My own interpretation of the report’s recommendations points out to other problems. For example, the Northeast Corridor’s technical standards are consistent but also bad, coming from an unbroken legacy of American railroader traditions whose succors can barely find Germany on a map, let alone bother to learn from it or any other foreign country. This way, the New Haven Line, which with modern trainsets and associated standards has few curves limiting trains to less than 150 km/h, is on a blanket speed limit of 75 mph, or 121 km/h, in Connecticut, with several further slowdowns for curves. There’s long-term planning for the corridor, and it’s bipartisan, but this long-term planning involves agencies that fight turf wars and mostly want to get the others out of what they perceive as their own turfs. There is lush funding, but it goes to the wrong things – Moynihan Train Hall but no improvements at the track level of Penn Station, extensive track renewal at 1.5 orders of magnitude higher cost than in Germany, in-place bridge replacements on curvy track instead of nearby bypasses.
The current planning does use too many consultants – in fact, Penn Reconstruction’s interagency agreement stipulates that they use consultant-centric project delivery methods, with one possibility, progressive design-build (what most of the world calls design-build; what New York calls design-build is different and better), not even legal in New York state law, but the local power brokers are trying to legalize it and break their own construction cost records. But it’s not quite the same as not bothering to develop in-house talent – there is some, and sometimes it isn’t bad, but poor project management and lack of interagency coordination has caused the budgets for the big-ticket items that Amtrak wants to explode beyond anyone’s ability to manage. The five recommendations, applied to the Northeast, mostly speak to the low quality of the existing agencies, rather than to a hodgepodge of standards as is happening at the interface between California High-Speed Rail and Caltrain or Brightline West.
The ultimate problem on the Northeast Corridor is that it is held together with duct tape, by people who do not know how to use more advanced tools than duct tape. They constantly fight fires, sometimes literally, and never ask why fires always erupt when they’re around; it’s not the heat, because the Northeast isn’t any warmer than Japan or South Korea or Italy, and it’s not underinvestment 30+ years ago, because Germany has that history too. Nolan points out the electric traction backlog on the Northeast Corridor grew from less than $100 million in 2018 to $829 million today; the people in charge are substantially the same ones who deferred this much maintenance over the six-year period that included the Bipartisan Infrastructure Law. I didn’t get into this project in order to study other people’s failures again, as we did with the construction costs report. But everything I’m seeing on the Northeast Corridor, even more than in California or Texas, points to what may be the worst intercity rail planning of any even vaguely modern country.
Pedestrian Observations 
