Home PANYNJ Let’s talk about that PATH Train station again

Let’s talk about that PATH Train station again

by Benjamin Kabak
Santiago Calatrava's PATH Hub soars through the Lower Manhattan sky. (Photo by Benjamin Kabak)

Santiago Calatrava’s PATH Hub soars through the Lower Manhattan sky. (Photo by Benjamin Kabak)

While in Lower Manhattan for the opening of the Fulton St. Transit Center in early November, I had a few minutes to wander around the much-transformed area. As I strolled over to the World Trade Center site, I couldn’t help but notice Santiago Calatrava’s PATH Hub. It looms above the area, piercing the sky in a rather impressive way. If you don’t know anything about the price tag or tortured history of the project, you would be right to marvel at this structure. But there’s something odd about it: Not even open to the public yet, its visible joints are already rusting.

In the various renderings of the $4 billion structure, the joints were neither visible nor rusting, and I wondered if this were part of the plan or not. And then, out come David W. Dunlap’s in-depth look at the PATH Hub with this gem at the end:

What did nearly $4 billion buy? Certainly an arresting structure, but one whose details do not match the shimmering images that Mr. Calatrava used to seduce officials a decade ago.

For instance, the ribs of the mezzanine looked sleek as silk in the renderings but in reality have the texture of stucco because of a fire-protective coating. Asked in March why no one had smoothed the surfaces, Mr. Calatrava’s office answered, “The client was not prepared to spend the additional money.”

That’s right: After falling to meet his already-lofty budget by nearly 100 percent, Calatrava tried to milk more money out of the Port Authority. If that’s not symbolic of the entire project, I don’t know what is.

This anecdote aside, Dunlap’s profile of this project is well worth the read. He delves into the spurious numbers that supported a big expense on a subway station and tracks the lack of leadership at the Port Authority as no one was in a position to stop project costs from spiraling out of control. Somehow, the PA expects 160,000 PATH riders per day, a jump of four times the current daily ridership, and it’s not clear how or where this number originates as the $3.7 billion station included no money for additional service. Here’s a key excerpt:

The price tag is approaching $4 billion, almost twice the estimate when plans were unveiled in 2004. Administrative costs alone — construction management, supervision, inspection, monitoring and documentation, among other items — exceed $655 million. Even the Port Authority of New York and New Jersey, which is developing and building the hub, conceded that it would have made other choices had it known 10 years ago what it knows now. “We would not today prioritize spending $3.7 billion on the transit hub over other significant infrastructure needs,” Patrick J. Foye, the authority’s executive director, said in October.

The current, temporary trade center station serves an average of 46,000 commuters riding PATH trains to and from New Jersey every weekday, only 10,000 more than use the unassuming 33rd Street PATH terminal in Midtown Manhattan. By contrast, 208,000 Metro-North Railroad commuters stream through Grand Central Terminal daily. In fact, the hub, or at least its winged “Oculus” pavilion, could turn out to be more of a high-priced mall than a transportation nexus, attracting more shoppers than commuters…

But whatever its ultimate renown, the hub has been a money-chewing project plagued by problems far beyond an exotic and expensive design by its exacting architect, Santiago Calatrava, according to an examination based on two dozen interviews and a review of hundreds of pages of documents. The soaring price tag has also been fueled by the demands of powerful politicians whose priorities outweighed worries about the bottom line, as well as the Port Authority’s questionable management and oversight of private contractors.

Read through the whole piece as Dunlap finds fault with then-Gov. George Pataki’s plans, then-Mayor Michael Bloomberg’s meddling and the Port Authority’s inability to lead. It’s a sobering look at a flawed project.

At this point, I’ve written extensively about the wasteful spending we’ve seen in the PATH Hub, and I’m almost tilting at windmills. It’s likely that the mall will offset some of the costs, but it’s clear maintenance expenditures will be far higher than they should be. As the Port Authority gears up to invest a lot of money into our region’s airports, we can wonder how we could have better used these dollars, but the key is to learn from this mistake. If we can’t, we’ll be doomed to repeat it — at Moynihan Station perhaps or elsewhere — and that’s something the region, with its myriad transportation needs, simply cannot afford.

You may also like


John-2 December 8, 2014 - 12:25 am

The only problem here is some of these problems were fairly common knowledge 5-10 years ago, when something might have been done about the mounting costs. As it is, the story’s more of an autopsy about how the Port Authority killed off $4 billion in federal funds by building the Tweed Courthouse of subway stations (mainly because the subway part was only the excuse for what was to go above it).

Stephen Smith December 8, 2014 - 1:03 am

The sad part about all this is that we don’t seem to have learned anything from it. In any of the recent discussions about the next major NY/NJ rail infrastructure project – a new Hudson River crossing and Penn Station (whether it’s the ARC cavern, or Penn Station South, or something more major) – do you ever hear any talk of what happened last time and how to avoid it?

Nyland8 December 8, 2014 - 5:09 am

When one considers how long we’ve all been witnessing this travesty unfold, one can’t resist conjuring the image of a “slow moving” train wreck, where the impending crash seems inevitable for minutes, but no amount of breaking can prevent that impact. Time after time we’ve seen how the Port Authority mishandles their little train set. It should be taken from them, subsumed into the MTA, and the bloated and redundant middle and upper management sent packing.

And if we learn anything from these events, it should be that the last thing we need to do is to build a horizontal Trump Tower under midtown to do the job that sending something like the L Train over to Lautenberg would do much cheaper, much faster and with much better results for the region.

Ryan December 8, 2014 - 7:06 am


The auxiliary station is a horrible idea and should be killed, but the most important rail tunnel project in the entire country is another two (ideally another four) intercity rail tracks underneath 34 St.

The L and the 7 should both be extended into New Jersey for different reasons, but neither can replace additional capacity for NJT and Amtrak between Secaucus (west of which the line is overbuilt) and New York City.

Alon Levy December 8, 2014 - 7:17 am

*Capacity for NJT, not Amtrak. Amtrak is running significantly below capacity because it runs short trains (at most 8 cars, not 16), runs even shorter trains on the Keystones, and wastes 37.5% of the Acela’s length on locomotives and a cafe. NJT is the one that’s having a capacity crunch even with long-ish trains.

Also, if I had to choose between a second Hudson tunnel and the completion of SAS, I’d choose SAS and so should you.

Ryan December 8, 2014 - 7:48 am

Train length isn’t relevant to the number of slots available for train movement versus the number of trains actually moving. By that metric (the only capacity metric which actually matters because you can run mile-long trains with 50+ passenger cars and stop as many times as you need to unload in 12 or 16 car chunks if there’s sufficient demand, there is no actual upper bound on capacity as a function of train length), Amtrak is at capacity, NJT is at capacity, the tunnel is at capacity and running 12-car or 16-car trains exclusively doesn’t do a damn thing to solve the very real problem that you’re about to have when one of the two tubes has to close for emergency repairs and then close again for long-term repairs.

I wouldn’t choose anything over solving the Hudson tunnel capacity crisis. Not even 2 Av rates higher on the priority scale. (2 Av, in fairness, comes close.) There’s always the option to continue busing people up and down 1/2/3 Avs and to go to CBTC on the Lexington Line to squeeze out another 16 trains there while we wait to build 2 Av out, but there are no parallel corridors over which you can get the sheer volume of traffic between NJ and NY Penn into NYC when (not if) the tunnels need to go down.

Ryan December 8, 2014 - 8:12 am

I’ll revise that statement slightly.

Train length is a variable in the calculations to determine what the minimum tip-to-tail headway is, which in turn sets the absolute maximum capacity of a rail line in terms of pure space available. So, yes, eventually at some point if you have high-speed switches and a bleeding edge signaling system and you’re already dispatching on the order of 45 trains per track per hour then train length is going to play a role in determining what capacity is.

But you need to max out very other variable first, which isn’t practical to do. (The mile-long train example, while technically feasible, is implausible except in the dystopian future where we’re down to one track into Penn Station and lashing up 50 car trains twice per hour because we can no longer run 10 car trains every six minutes.) There simply will never be the kind of demand where we’re running 16-car Acela IV sets every two minutes and seriously having to figure out how much more room between the tail end of the :00 run and the tip of the :02 run there is to add another X cars.

BenW December 8, 2014 - 11:28 am

I think you’re sort of missing Alon’s point: when you say “Amtrak is at capacity”, you mean Amtrak is at the limit of the number of slots they can use, no? But if Amtrak could (hypothetically, since they clearly don’t have the hardware available for it today) be running trains that carry 25%-100% more passengers, using (as you point out) the same number of slots… then why do they need more slots? If moving passengers is the goal, not moving power cars, then Amtrak has a lot of capacity to expand within their current slot allotment.

None of which changes the long-run problem of losing one of the North River tubes for extended rehab, of course—but the point about Amtrak vs. NJT and who needs more capacity is very much about the number of linear feet of passenger car that goes through those tubes per slot those carriers are using right now

Ryan December 8, 2014 - 12:26 pm

I consider the true goal of mass transit to be moving people to where they want to go when they want to go there. Frequency trumps speed trumps passenger comfort, but you need two out of three (and you want ideally to have all three) for a successful mass transit service.

I would prefer 8 trains an hour at 75% loads each to 6/hour with 100% load, 4/hour at 150% load (the situation we have today in some cases) or 4/hour at 100% load with double the available space relative to the trains in my preferred 8 per hour alternative. The dismissive way to restate that is indeed that I would rather move as many power cars as possible.

And by the way, I do believe that the overwhelming majority of the cross-Hudson traffic should be local/regional runs. Specifically, I’d say that there should be 16 slots per hour allocated to Amtrak traffic through the tubes and 80 slots per hour allocated to everyone else (NJT, the MTA, and the MTA can all fight it out over who deserves to get what out of that chunk). That’s 24 trains per hour per track in a four track system, working out to regional/commuter services roughly every 90 seconds in each direction, and Amtrak each way roughly every 7.5 minutes. (Given their propensity for “let’s all sit and wait at this station” you could literally have the next Amtrak run arriving as the previous one leaves. Show up and go.) That’s enough capacity for pretty damn close to forever. If we do the right thing and overbuild to six tracks now, we’ll have enough capacity literally forever and never have to worry about taking track pairs offline again.

Alon Levy December 8, 2014 - 6:11 pm

I think you misunderstand the point of frequency. Nobody really needs an intercity train every 7.5 minutes, not when these trains largely go to the same destinations. Show-up-and-go is nice, but when you’re traveling for three hours, it’s not really necessary; even at hourly trains you start hitting diminishing returns, and below half-hourly frequency there’s not too much extra benefit. The TGVs only run half-hourly trains on the thickest markets, like Paris-Lyon, with a few more at the peak, and they still get very high modal share. Oh, and they require reservation ahead of time because of the yield management system. It drives me up the wall that SNCF works that way, but passengers ride the TGVs just as much as they ride the show-up-and-go Shinkansen trains.

The TGVs also run in double traction, so they are as long as 16 American cars. They have power cars, but the coaches are bilevels, so the total number of seats per train is 1,090, more than 3 times as many as on the Acelas. They also have high frequency on the busiest trunk line, Paris-Lyon, it’s just that instead of running 10 trains between Paris and Lyon, they run 2 between Paris and Lyon, 2 between Paris and Marseille (not stopping in Lyon), 1 between Lille and Lyon (bypassing Paris), etc., so to the passenger the effective frequency is usually half-hourly or hourly.

Given that there’s a capacity constraint, there’s no excuse for sticking with 8-car trains. It’s not like mile-long trains, because, as I noted in my reply to eo, most Amtrak platforms are already long enough for 12 cars and many are long enough for 16, and lengthening the ones that aren’t costs a pittance.

The mockery of Amtrak’s “let’s sit and wait” mentality is another example of why Amtrak shouldn’t be trusted with anything. If it let people board from any access point, and not check tickets at the platform when it’s already checking them on board, it could let people get on the train much closer to departure than it currently recommends. Instead of waiting in line 10 minutes, people could get to the station 3 minutes ahead of time and board. That’s 7 minutes of effective trip time that at current rates Amtrak would have to spend billions on in order to shave by raising top speed or removing slow zones.

adirondacker12800 December 8, 2014 - 7:05 pm

Replacing 20 trains with a capacity of 300 with 28 trains with a capacity of 425 almost doubles the number of seats. How many should they provide? Three times as many and run them half empty for a decade or two?

Ryan December 8, 2014 - 8:57 pm

Okay, so intercity means “three hours” right now, but it shouldn’t have to. New York to Harrisburg is an intercity route and will always be an intercity route. New York to Harrisburg is also doable in 106 minutes at an average speed of just under 115 mph, which is well below international averages (between 130~150 mph). At an average of 140 mph you get on your train in Harrisburg and get off in Penn Station just about 84 minutes later. (And by the way, 140 mph averages drop travel time to Philadelphia to 39 minutes.)

In the future where we’re running trains at those speeds, you’re going to see the mileage range of commuters expand as well. New York’s daily commuter market already includes places as far as way as Hartford and Philadelphia, and it would expand to include places like Baltimore and Schenectady and Providence. People are already grinding out the 2 hour trips to places like Harrisburg on a regular basis. Many, many more people are going to want to make that trip when it’s 45 minutes instead.

It’s those people I want to give high intercity frequency to, and once-per-hour or twice-per-hour trains that ultimately terminate (or “terminate” and simply become some other train, possibly two or three other trains) in far-flung cities several states away add up to trains every 7.5 minutes into New Jersey, and every 7.5 minutes in the other direction.

adirondacker12800 December 8, 2014 - 11:05 pm

They have 4 an hour now, mid afternoon….

Alon Levy December 9, 2014 - 5:37 am

Amtrak’s current intercity service is about three hours on the top city pair.

For what it’s worth, the French city pairs are usually shorter. Paris-Lyon is two hours.

Commute trips happen even on the infrequent TGV. They’re actually not a great thing to have on your system; they’re peaky, so you need to buy trainsets that will only run a few hours a day. About the best sort of travel pattern is what goes on between New York and New Haven today: a lot of intercity trips and some day trips, but not many commutes, so that passenger traffic at New Haven is the same on weekdays and weekend days, 4,000 Metro-North boardings.

Talking about 8 tph peak is nice, and Amtrak should be planning on possibly 12, in the future. But it has so many other things to worry about first, and insisting on Gateway as phase 1, especially given its extreme cost, is a nonstarter. For starter HSR service, 4 long tph on NY-DC (Keystones should be cut to Philly if there’s no demand for a long NY-Harrisburg train) are perfectly fine.

Ryan December 9, 2014 - 8:17 am

12 is excessive, considering that we’re probably not going to go to six tracks. 12 (which is really 24, because 12 each way) drops the available slots for all other users of the tunnel down to 72 which then gets cut in half for two direction operation which in turn means you’re down to 36 TPH for NJT-LIRR or NJT-MNCR joint runs. Is it a lot? It’s the difference between one more set of midtown direct trains, which are more important to regional operations than taking ~90 seconds off the top of the time between Amtrak trains in the direction of (Newark, Trenton, and) Philadelphia on the aggregate.

I disagree that commutes are a bad thing to have in your intercity system – leisure travel’s peak hours aren’t the same as business travel’s peak hours and there’s very little overlap. Most business travel is going to peak between 4-7 am or 7-10 pm, leisure travel (barring some exceptions) between 9 am and 2 pm, and the conventional commuter peak blocks are 6-9 am and 3-6 pm. The increasing prevalence of flex time and staggered schedules will smooth all the business peaks out to some extent, but ultimately, there shouldn’t need to be extra train deployment in a relatively narrow window like you’re suggesting would be needed. There’s also the counter point to consider: the commuters/short-haul intercity uses are the group that most needs this frequent service. If you’re gearing HSR only to people committing to 2~3 hour trips and largely ignoring the 40~80 minute trip set, then we’re right back at your original premise of not needing to bother with any more than hourly or half-hourly service for all of time, because what’s another half hour to wait when you’re already going to spend 3 on the train?

I think we agree that there’s a point where increasing frequency hits diminishing returns, I just don’t think we agree on where that point is. 8 is enough to satisfy Amtrak’s needs for a very long time considering that Philadelphia is really the only key place that Amtrak needs to go because all of its services will either emanate out from Philadelphia, won’t be using the tunnel at all (New York-originating services pointed east towards Boston or north towards Canada), or will operate as continuations off of services coming up from Philadelphia.

And just because we’re planning to have 8 slots for Amtrak in the future doesn’t mean it needs all 8 of them right now. In the short term, NJT is the biggest beneficiary of the tunnel project because it needs the extra 44 slots that would be allocated to local/regional service far more than Amtrak needs another 4 slots to grow into. We’re not doing Gateway or son of ARC or whatever you want to call the new tunnel for the impact it’s going to have on Amtrak’s operations. We must build another pair of tracks under 34 St because the existing tracks will need to be closed for maintenance sooner rather than later, because the existing tracks are already largely at capacity, because the existing tracks are very much a critical point of failure, and because even after we build the NJT-LIRR tunnel under Liberty St and connect Atlantic to Secaucus there is going to be a huge mismatch between available capacity in NY and NJ. Amtrak is a throw-in – a very important throw-in, but a throw-in nonetheless.

Alon Levy December 9, 2014 - 8:44 am

12 is excessive for now. It’s not excessive for the timeframe Amtrak thinks it will take to build things. Worth thinking about, just not orienting a tunnel that should take 5 years to build around.

Although current peak traffic through the tunnel is 24-25 tph, it’s possible to go up to 30 tph with better signaling. Mainline signaling systems are capable of 30 tph today, although not reliably. RER signaling systems, which involve somewhat shorter captive lines, go up to 32, provided each tunnel track splits up into two station tracks; this is achieved on the combined RER B and D tunnel between Gare du Nord and Chatelet-Les Halles, where each station has dedicated tracks for each line. We’re also talking about future capacity, and signaling systems are improving; expecting 30-32 tph per track in the future is reasonable.

Of course, the positive train control mandate provided a perfect opportunity for American railroads to install good signals, i.e. European Train Control System Level 2… and instead they squandered it on a low-grade overlay. European stuff is foreign and icky, can’t have that on the LIRR.

Finally, don’t think in terms of total trains per hour, not when it comes to highly branched commuter rail system. Think in terms of a service plan. Does it make sense to provide a service plan with a certain tph count between Jersey and Midtown? Personally I think that, on top of the two tunnels that should be built as soon as it’s possible to build them for reasonable cost (which is closer to $3 billion than to Amtrak’s stated cost) – ARC Alt G and Jersey City-Lower Manhattan-Flatbush – the region should be thinking of a fourth tunnel pair, from Hoboken to the ESA part of Grand Central, giving the Lackawanna lines a dedicated path to Manhattan. But that’s 30 years from now.

Ryan December 9, 2014 - 9:27 am

The Lackawanna Lines can be given a direct path into Manhattan for far cheaper if you just build a flyover from the upper level tracks east of Secaucus to the lower level tracks either south of Secaucus (so they can stop there) or north of it (so they can bypass the stop.) It’s almost absurd how much space there is to work with out that way, and no amount of environmental mitigation could ever make connecting Hoboken to GCT the cheaper option.

If you want to talk about this in terms of service patterns, then the three places that regional rail should enter Manhattan are downtown (Liberty St), midtown (34 St), and somewhere uptown (presumably 86 St). Any other crosstown connections have no business being anything other than part of MTA expansion.

I think of what happens with 34 St’s service as a function of tph or train slots specifically because it’s the only one of those crosstown rail lines that should ever carry intercity traffic.

Alon Levy December 9, 2014 - 10:16 am

The Lackawanna lines already have a direct path to Manhattan – that’s the Kearny Connection. What I mean is giving them a dedicated path, without track-sharing with other lines, assuming traffic rises to the level that justifies it.

(More details at my place, soon.)

adirondacker12800 December 9, 2014 - 12:07 pm

There’s almost as many people in Brooklyn as there are in metro Baltimore. If Brooklyn was it’s own MSA it would be the 22th largest in the country – between metro Denver and metro Pittsburgh. The country’s biggest business district is Midtown Manhattan, the second biggest is the Loop in Chicago and the third biggest is Wall Street. Which implies Wall Street is the second biggest business destination along the NEC. There’s some demand for intercity travel. It would be almost “free” if we build a tunnel from Atlantic Ave to Jersey City to get New Jerseyans who work on Wall Street off the PATH trains and Long Islanders who go to Penn Station out of Penn Station.

Ryan December 9, 2014 - 12:10 pm

Sorry, I think I misunderstood.

Is the current NJT Main Line not the former Erie Lackawanna Main Line? The Main Line, and the other lines that branch off of it north of Secaucus, were the Lackawanna Lines that I was thinking of and they are not served by the Kearny Connection.

My proposal would be a flyover that joined the Main Line to the approach tracks for the 34 St tunnel. Send half of all the trains from Suffern or Port Jervis or Rockland County or wherever else you’d eventually branch out to, and send the other half down through the Liberty St tunnel. Maybe keep a couple trains per hour around for Hoboken, but the vast majority of traffic would go to one of those two tunnels.

adirondacker12800 December 9, 2014 - 1:05 pm

Since the 80s, Amtrak, NJTransit and the LIRR have been talking about what happens when the lines in New Jersey get direct access to Manhattan. The plan has always been to someday build a loop near the new station in Secaucus so that all the trains in New Jersey could go to Manhattan.
Connect the LIRR to NJTransit someplace in downtown Manhattan, Hoboken becomes obsolete. People who want to go to Hoboken can change to PATH or the HBLR at the station in Jersey City.

Alon Levy December 9, 2014 - 1:36 pm

What NJT calls the Main Line descends from the Erie Main Line; the Lackawanna owned the Morris and Essex Lines and the Montclair Branch.

The problem with the loop is that it’s super-awkward. Don’t even bother building it temporarily if the Midtown tunnels open before the Lower Manhattan ones (which they should); just rip the faregates at Secaucus, add access points directly between the two platform levels, and tell people to transfer.

adirondacker12800 December 9, 2014 - 2:07 pm

Onto trains that are already full…

Nathanael December 10, 2014 - 3:19 am

Amtrak isn’t running hourly service yet, though.

I totally agree that Amtrak needs to abolish its crazy boarding “wait in line” procedures at Penn Station. However, frankly, there are deeper problems with the way Penn Station is operated.

There’s no signage on the platforms to indicate what train is what.

When a train fails and we are sent to a different track (as happened on the last NJT train I was on)… notifications are mostly missing.

There’s a bunch of standard-practice stuff which is done by nearly every commuter rail system, and also at most minor Amtrak stations, which simply isn’t done properly at NY Penn.

adirondacker12800 December 8, 2014 - 1:30 pm

The weekend schedules are arranged for single track service. That’s what you get when one track is out.

Eric December 8, 2014 - 9:21 am

“Also, if I had to choose between a second Hudson tunnel and the completion of SAS, I’d choose SAS and so should you.”

SAS is more of a need now than a Hudson tunnel. But were the current Hudson tunnels to fail, which according to the following article will happen within 20 years, it would immediately become a much bigger problem than the lack of an SAS.

adirondacker12800 December 8, 2014 - 10:18 am

Wall Street is the country’s third largest business district and the only access is by subway. Instead of sending people to 34th Street so they can get to Wall Street just send them to Wall Street.

Michael K December 8, 2014 - 3:12 pm

The tunnel from Hoboken to Atlantic Terminal via Wall Street will do just that, as well as make our two most underutilized rail assets (Brooklyn Division and Erie Lines) useful.

Eric December 16, 2014 - 4:13 am

That’s a pretty obvious need, but are there any actual plans for it to be built?

eo December 8, 2014 - 2:02 pm

I disagree. The tunnels are not your only constraint. A physical constraint is the length of the platforms. No I am not speaking of the NYP platforms, the platforms at other stations because your long trains need to stop elsewhere and double stopping or making everyone walk to the back of the train to get off does no wonders for throughtput — just check some of the NJ Transit line: on current length trains the last three cars cannot platform at Bay Street so the train typically dwells 3-4 times the normal time for high level platform just because people need to walk the length of several cars. Furthermore with longer trains you lose the important ability to stop two trains on the longest platform in NYP — the second train will have to wait in the tunnel for another platform, so here went your increase in passengers carried — the train throughtput is down.

Then come other unfortunate constraints. For example, there are no more Acela cars and there won’t be any any time soon. Despite all the noise Amtrak makes about Acela2, I will believe it only when the President signs the bill authorizing such a purchase. Also, the Keystones do not need to be longer than current length because the demand is not there — there just are not enough passengers to fill more cars on these trains between NYP and 30th Street and west of there the demand also is not there for more cars (and you have the platform length problem). Politically you cannot take the slots from the Keystones, so sorry they are staying and you cannot use these slots on the higher demand between NYP and Washington. There are also other reasons, but lets just stop here.

The reality is that because of a few physical constraints and a few self imposed political constraints that I see as not going away you are only left with new tunnels and station as a way to increase passengers.

Alon Levy December 8, 2014 - 5:59 pm

1. Indeed, platform length is a constraint. But nearly all Acela stations are already 12 cars long or longer. The shortest, New Haven, is 8 cars long but has space for lengthening platforms, which can be done for about three orders of magnitude less money than new tunnels.

2. Penn Station has 4 access tracks at most and 21 station tracks; in no circumstance does it need to double-berth.

3. The constraint on the number of cars is also a constraint on how frequently trains can come in! Amtrak needs to use its equipment more efficiently; the Acelas have low equipment availability (17 out of 20 at the peak if I remember correctly) and long turnaround times, and if they’d run based on best practices they could run in double south of New York and decouple at Penn Station continuing in single set to Boston.

4. Politically Amtrak wants escalating amounts of money for a new tunnel. It’s now up to $16 billion. “We can’t cut the Keystones to Philly and run longer Regionals to meet them because politics” tells me Amtrak is irresponsible with money (and the implied “we can’t split consists in Philly in a reasonable amount of time” is also kind of stupid).

5. New tunnels and a new station are two very different things. New tunnels are great, if they come with a decent service plan and cost a reasonable amount, which for the purposes of this discussion is well below $5 billion. A new station is just another excuse for a PATH-style disaster, only bigger. Penn Station is not more crowded than Stockholm Central. It can keep its crappy structures as they are and just punch two extra tunnels to tracks 1-5 and two extra tracks to the east to connect to Grand Central.

Spendmor Wastemor December 8, 2014 - 6:54 pm

Even I cannot understand how having a locomotive on a high speed train is wasteful. Two seems reasonable, as turnaround tracks are missing in many places where they are needed, and today’s hypercomplex kit can quit more often then is acceptable for reliable service.

Alon Levy December 9, 2014 - 5:42 am

Ew, no. First, trains are capable of running push-pull; the MBTA has just one loco, without any turn tracks at either South or North Station. The FRA claims it’s unsafe for intercity service on the NEC for stupid reasons and that’s why Amtrak uses two on the Acela, but one is sufficient.

And second, most high-speed trains in the world today are EMUs. Most of the countries that use locomotives are transitioning to EMUs. EMUs far outperform locomotives, can run double-ended without any trouble, and allow seating in nearly 100% of the train rather than 75% of it. The problem with immediately transitioning to them is that locomotives and coaches often need to be replaced at different times; that’s the excuse Amtrak uses for never EMUifying.

Nyland8 December 8, 2014 - 7:30 pm

” … but neither can replace additional capacity for NJT … ”

No – that is exactly what it would do. Running the L Line over to Lautenberg would substantially reduce the number of NJTransit trains into NYPenn. Having commuted into that station for years, I know that the overwhelming preponderance of people who take NJTransit into NYPenn do not walk to work from there. They head straight for one of the Manhattan trunk lines. Getting on a crosstown subway in Secaucus obviates the need to go into NYPenn. It also means that dozens of NJT busses could stop at Lautenberg and don’t have to make the trips through the Lincoln Tunnel.

In terms of bang for the buck, and in terms of the quickest solution to several pressing problems, running the L Line over to Secaucus is by far the best deal.

And while I agree that the 2nd Ave contracts should be a higher priority for the MTA, it shouldn’t be an either/or situation. I don’t think the MTA should be responsible for building the tunnels or the station. The design, the engineering, the construction … everything … should be borne by the State of New Jersey. It solves three of their biggest problems – a) sharing Amtrak’s tunnels and station; b) east side access; c) Port Authority Bus Station overcrowding. It is in the best interests of their commuters to get them where they need to go as quickly, and as cheaply as they can. We don’t need to pump more New Jersey commuters into NY Penn Station just to watch them frantically scramble out of there every morning looking for the subway that takes them to their final destination.

It even gives them east side access without spending the billion$, and waiting the years it will take to tunnel from Grand Central to NYPenn.

It’s as elegant a solution as anyone has come up with, it can be completed in the shortest timeframe, and by not involving any underground New York City real-estate hewn from Manhattan schist, it costs a fraction of any other option.

It would be a win-win-win for the region.

adirondacker12800 December 8, 2014 - 7:51 pm

They aren’t pouring into the subway to get to 14th Street. Why would someone who works north of 28th or 26th opt for a longer commute?

Nyland8 December 9, 2014 - 1:10 am

They certainly ARE pouring into the subways to get down to the southern business district. Just stand on the uptown 34th St. platforms of the A,C,E,1,2,3 during the evening rush and try not to get knocked over by the hurried bodies desperate to get off those trains, looking to make their connections back to Jersey. 14th Street is 20 blocks closer to Wall Street than 34th St is. And likewise witness the tsunami that crosses 7th Ave with every changing traffic light in the morning as they head for the B,D,F,M,N,R,Q at Herald Square. The people that come into Penn Station every day work throughout the city – even out in the boroughs.

The fact is that thousands of both New Jersey and Long Island commuters go North, South and East from NYPenn every morning, and those connections are simply better made from a crosstown subway. Which is not to say that every one of them would prefer to bypass NYPenn, but I’d venture to say more than 75% would prefer to do without piling in and out of that building, if there was an attractive alternative. Now imagine the possibilities if 6 or 7 out of every 10 NJTransit trains didn’t have to go all the way into NYPenn. It’s a totally different landscape. More slots for Amtrak, more slots for MetroNorth, etc.

For at least half of the Pascack Valley, Bergen, and Main Line commuters, making their first subway connection in Secaucus is far better, and far quicker than waiting for the next Northeast Corridor line to bring them into NYPenn from Lautenberg . . . especially if their destination is the east side. Because going into Hoboken and taking the PATH won’t get you there at all. Getting into an empty, waiting subway car that brings you directly to the trunk line of your choice is far more inviting than cramming into a standing-room-only commuter train that’s coming up from Trenton and only brings you one stop before you have to hoof it over to the same train.

And because the subway will run more frequently than the incoming commuter rails, rerouting busses through Lautenberg makes a great deal of sense. In fact, they already do that on occasion when the tunnels get jammed up. The L or the 7 to Lautenberg Station would be instant relief to the PABT.

Too many people seem convinced that some form of Taj Mahal under Macy’s basement is the answer to the problem, but I’m saying that the new station that relieves all that current and future pressure has already been built. It’s out in Secaucus. All we have to do is find the most elegant way to tap into it.

Alon Levy December 9, 2014 - 5:43 am

14th Street is a secondary business district. If the point is that it’s closer to Wall Street than 34th, then build a dedicated commuter rail tunnel to Lower Manhattan in addition to the existing tunnel to Midtown.

Nyland8 December 9, 2014 - 9:45 am

No. The point is that it is BETWEEN the upper and lower skylines, thus providing equal access to BOTH! Other points include the fact that it offers a wider train with wider side platforms that will not be overwhelmed when eastbound and westbound trains enter the station at the same time – like the center-platform 7 WILL; it offers BETTER east side access because it will connect to the 2nd Ave subway at 3rd Ave – which the 7 Line will NEVER do; it solves the problem of the terrible terminal at 14th & 8th, where during daylight hours the westbound train NEVER enters the station without cuing up in the tunnel for a minute or two first, and then ALWAYS crawls into the station at 5mph because that is the law. Conversely, extending the L Line to Lautenberg means entering 14th & 8th at full speed; It is a relatively straight shot – i.e. “the shortest distance between two points” – to Lautenberg, meaning a shorter, cheaper tunnel; it will connect easily and logically to the HBLR at 9th & Congress; etc, etc …

From the standpoint of distributing New Jersey commuters throughout the subway system, from the standpoint of costs, and from the standpoint of engineering, it is simply a better choice than the 7. Any advantage the 7 Line might have originally had over it when first proposed by Bloomberg was wasted once the TBMs were demobilized. That train has left the station.

Don’t get me wrong. I’m an advocate of any and all of these projects. I vilified Christie for unilaterally squashing the ARC project, even though I knew damned well it was less than an ideal solution. And I have no objection to through-running a couple of tracks from Secaucus to the Atlantic yards via Lower Manhattan either. But now you’re talking about hewing a very deep new cavernous station, threaded between skyscraper foundations in the financial district, including dodging all the existing subway lines. What do you realistically think that timetable will be? 2114? And what on earth do you imagine the costs?

I’m just pointing out that by far, the biggest bang for the buck, the most utile improvement, the easiest engineering, and the project we can realistically see accomplished in my lifetime, would be running a cross town subway to Secaucus. It pales every other option, in every conceivable way, no matter how attractive we might imagine the final result would be.

It’s cheap, it’s fast, it’s elegant, and by relieving the cross-Hudson crush, it buys time for all the other long-range projects that Amtrak might yet devise. And because it should be undertaken by New Jersey, it should NOT compete with the MTA’s timetable or commitment to the T Line.

Alon Levy December 9, 2014 - 9:57 am

7 trains are longer than L trains. They also get people closer to the largest business district. Either build one tunnel for each business district, or build one just for the larger one; there is no good reason to build something in the middle. Transportation doesn’t work that way.

As an exercise, imagine that there are 1000 people at point A and 2000 at point B, located 10 km away from A. What’s the ideal location for a single service station for something (car repair, maybe), assuming you want to minimize the total distance people have to go to the station?

L tail tracks would be nice for other reasons, but I don’t think that they have to be bundled into a tunnel across the river.

adirondacker12800 December 9, 2014 - 11:49 am

Theres enough demand for Midtown and enough demand for Wall Street for each of them to have their own tunnels.

Ryan December 9, 2014 - 8:50 am

The most elegant way to tap into it is with an extension of the tail tracks at Atlantic Avenue into a new tunnel under State St and the BQE in Brooklyn, Liberty St in Manhattan, and Grand St in Jersey City, surfacing at whatever point is most convenient for a straight shot up into Secaucus – possibly with a detour to hit Journal Square.

Do I think the L and the 7 should be extended into Jersey too? Yes, but certainly not at the expense of the 34 St tunnels or the downtown Manhattan tunnels or the 2 Av Subway. Those are priorities #1, #2, and #3 and in that order. The L and the 7 are fighting it out for something on the order of priority #8.

Ryan December 9, 2014 - 9:06 am

No, that is not at all what it would do.

You’re arguing to replace NJT capacity with something else. The argument you are making is that NJT is over-represented in the existing tunnels into Penn, that sending the L or the 7 or both out to the Taj Malautenberg is a better way to serve existing NJ-originating traffic, and that you want to effectively zero out (or leave less than 4 TPH and say “look, see, we’re not zeroing you guys out”) NJT traffic into Penn Station instead.

That’s an argument I can’t agree with and frankly I think it’s a crappy argument to make from a regional connectivity standpoint – but you’re welcome to that argument, if you want to make it.

But let’s be absolutely clear in that the capacity slack to permit the North River tunnels to close one at a time for repairs is going to come from somewhere, and you, Nyland8, are suggesting that it be taken away from NJT instead of given to the entire region in the form of a new pair of tracks. You, and nobody else, are suggesting that Amtrak deserves the lion’s share of the tunnel traffic underneath the Hudson River going forward.

If you want to make that argument, that’s fine. It’s not nearly as much of a disaster when one of the 34 St tracks goes offline in the world where we can just nullify all the NJT runs and go to 6 slots to spread out across 6 Amtrak trains operating in both directions over the single track. Just remember that when that happens, it will be exactly what you asked for when you said “don’t expand under 34 St, redirect all that money to 14 or 42 instead.”

Larry Greenfield December 8, 2014 - 7:27 am

What struck me the most in Dunlap’s article was the ‘Leadership Churn” in which “Consistent direction was rendered almost impossible by constantly changing leadership: four New York governors who appointed five executive directors of the authority, and five New Jersey governors who appointed four chairmen.”

If this doesn’t lead to either radically changing or eliminating PATH, nothing will.

Larry Littlefield December 8, 2014 - 7:43 am

After a process dominated by Municipal Art Society types you had two form over function proposals: the Liberskind tower for One WTC whose primary goal was to stand up while looking like it was about to fall down, and the Calatrava station.

The private sector refused to build the Liberskind tower.

Using our money, politicians decided to build the Calatrava station.

Folks who ride around in taxis aren’t going to be worried about the opportunity cost to the serfs.

Kenn December 8, 2014 - 11:51 pm

http://ny.curbed.com/archives/.....illion.php — Comment #4

Administrative costs? Translation: corruption. A giant taxpayer funded cookie jar available to the usual sleaze bags. That they could in any way justify 4 billion dollars for this ridiculous gateway just shows how stupid they think the public is. And they may be right.

is he right?

Nathanael December 10, 2014 - 3:22 am

Excessive “administrative costs” isn’t always corruption. Sometimes it’s merely gross incompetence.

AlexB December 8, 2014 - 8:47 am

The point in the nyt article about the smooth fireproofing is that it would be a much more reasonable and minor expense compared to all the other things the authority wasted money on for political reasons. 100 grand or whatever it is for a better finish makes much more sense than spending hundreds of millions extra to build the above ground plaza before the underground subway. The better finish would have nothing to do with calatrava’s fee. In fact, they may have had to pay the architect extra to change the documents the remove the better finish shown in the renderings.

Eric December 8, 2014 - 9:33 am

Every expense seems justified and minor at the time. You have to start by putting your foot down somewhere.

Nathanael December 10, 2014 - 3:23 am

The porcupine was the place to put one’s foot down. There is no reason for it. Build any one of a dozen traditional headhouse, dome, or skylight designs instead — a lot cheaper and, frankly, prettier.

gimbels lover December 8, 2014 - 10:11 am

As Alex says, the $1M for sanding the intumescent paint or whatever is peanuts compared to the billion dollars added to the project by political expediency.

The massive, custom-section exposed vierendeel trusses underground are a huge cost that are worth pinning on the architect. The alterations made to appease security fetishists, a group that also gave us an uncreative, dull skyscraper that is also the most expensive skyscraper in the world, are not.

The picking on the artistic side of this plays into comfortable stereotypes about frivolity and ego, when the Times article pretty evenly distributes blame across the board. I realize you’re trying to advocate for capacity increases but I’d hope you could do it more precisely.

I was down there yesterday, and the truth is that it’s attractive enough that most people will forget the cost, which is a real problem. Unfortunately, the lightly-informed nitpicking of architects on this blog isn’t giving anyone insight into how to actually get great architecture on a budget.

I look forward to Gene Kaufman designing SAS Phase II.

Benjamin Kabak December 8, 2014 - 10:18 am

Your criticisms of my piece are valid. I don’t dispute that. We’ve read enough about Calatrava over the years to know that he’s a party responsible for the costs, stemming from the impracticalities of his original design, but we also know that political meddling and the demands of the security state are driving the price tag as well. I’ll be mindful of the coverage in the future.

gimbels lover December 8, 2014 - 10:50 am

I very much appreciate that. I just thought that it was an odd way to read the Times article. As a CM trained as an architect, when I hit the real world, I was pretty surprised to see where costs come from in projects, and also how hard it can be to parse cost saving approaches.

I certainly don’t want to defend Calatrava, who does seem to have a long, expensive record of white elephants.

al December 8, 2014 - 1:16 pm

There’s also the “toss it into the station budget” factor. Much of the sub surface components got tossed in with the station regardless of whether it was directly related to the station or not. This is on top of the costs from added complexities and difficulties that political meddling caused. How much was to cover overruns or change orders?

From the NY Times piece it stated $400 million for these adjacent items.
From WSJ the cost for added cost for rushing the 9/11 Memorial for 9/11/2011 is $100 million.
From The NY Times the cost to micropile the IRT subway tunnel box while trains were running through it instead of shutting it down periodically to underpin it cost $350+ million. WSJ puts it in the $300-$500 million range.
That is $1.3+ billion out of $4 billion.

How much of the $2.7 billion is the a) underground mall, b) the subsurface station, and c) the above ground Stegosaurus?

I recall the Stegosaurus was $300-700 million from various estimates (that was from 2 years ago, so update is needed). That leaves $2-$2.4 billion for the sub-surface station/mall. That is still one very expensive station/mall.

SEAN December 8, 2014 - 11:24 am

I’ve made this point before Ben, but it bears repeating here…everyone needs to remember the mindset in 2001 vs 2014. In 2001, after 9/11 the mindset was “we must rebuild otherwise the terrorists have one.” Although there was a minority group that believed the site was hallow ground & shouldn’t be rebuilt beyond the mamorial & PATH station.

Enter the polititions & federal funds for rebuilding in 2003 along with some ridiculous schemes to bring either AirTrain downtown from JFK or the LIRR from Jamaica. Since that time, it’s been a battle of epic proportions for the future of the WTC & lower Manhattan. The PATH station is just one part of the mess – all be it an expensive one.

Now in 2014 we’re looking back over the past dozen years with regrets saying we shouldn’t have… but it’s too late – the money was spent & we need to move forward & learn from this. The problem is can we? The political system may make it next to impossible to do so.

Eric F December 8, 2014 - 11:26 am

“although there was a minority group that believed the site was hallow ground & shouldn’t be rebuilt beyond the memorial & PATH station”

Rudy Giuliani was in the minority!

SEAN December 8, 2014 - 1:35 pm

Rudy Giuliani was in the minority!

Ironic isn’t it? Rudy would have given in to the terrorists based on logic at that time – if you could even call it logic. And yet George wanted to extend AirTrain & or the LIRR to lower Manhattan despite federal funding appearing shaky for such projects.

As I remember, the feds were uncertain of the number of passengers on the LIRR segment despite the actual number of LIRR riders carried. As for AirTrain, the way it was built & it’s ridership potential made it a ridiculous project in the first place.

normative December 8, 2014 - 4:10 pm

“demands of the security state are driving the price tag as well.”

We don’t talk about that enough. Didn’t the article say the NYPD made him change the original design because the spines weren’t blast proof, or something to that effect. I actually liked the original design a lot better.

lawhawk December 8, 2014 - 9:20 am


I don’t think that you’re seeing rusting; rather it’s likely the anti-corrosive steel treatment that they’ve used (similar to cor-ten steel like what’s used at Barclay’s).

Eric F December 8, 2014 - 9:21 am

I’m not sure they really have a rusting problem. The rafters are to be painted after installation. A couple of iterations back, the plan was to have them all installed in August. Of course, delay No. 1,000 set in, and they were finally installed in roughly the beginning of December, and they are still welding. I think by now it’s too cold to paint these things, so they’ll be dowdy until the Spring, is my guess.

I think your point on the texture is just wrong. You wrote:

“That’s right: After falling to meet his already-lofty budget by nearly 100 percent, Calatrava tried to milk more money out of the Port Authority.”

I don’t think that money to smooth the rafters’ texture would have went to Calatrava. That is a materials and finishings point, not a design point. I’m not sure why you think the architect was referring to himself as a recipient of money in that quote.

All in all, amazing that the PA is spending that much money as the terminus of an agonizingly slow system, stuck using a cira 1900 1×1 track alignment with perpetual operational delays. It really is astounding.

Larry Littlefield December 8, 2014 - 11:51 am

“All in all, amazing that the PA is spending that much money as the terminus of an agonizingly slow system, stuck using a cira 1900 1×1 track alignment.”

I would have been content to have Hudson Terminal rebuilt. And the existing #1 platform at South Ferry merely extended northward to accommodate a full trail, and the station rehabbed, despite the ongoing presence of the curve.

Jeff December 8, 2014 - 4:00 pm

Technically the joints are “rusting” because they are still welding the steel. You can’t cover up those weld joints until the welding is done, so no primers or finish paint have been applied yet.

This type of stuff is a normal part of the construction process. Nothing special whatsoever.

Bgriff December 8, 2014 - 10:36 am

What do you bet the Port Authority and MTA engage in a bidding war for their respective transit stop malls and one gives Apple (and Coach, Brooks Brothers, etc.) a cheap lease at the expense of both public authorities.

Eric F December 8, 2014 - 11:25 am

The scale of the WTC station utterly dwarfs Fulton Center.

BoerumHillScott December 8, 2014 - 11:58 am

Westfield controls the retail at both stations

The was potential for a bidding war with Brookfield at Brookfield Place (former WFC), but there seems to be an informal agreement where Brookfied will get the higher end stores while WTC/Fulton get more mass market.

Alon Levy December 9, 2014 - 12:22 pm

Is that collusion legal? It seems obvious to me that in a major business district with several malls, the malls will have to compete for tenants, and their anchor tenants will have to compete for customers. In Singapore, Orchard Road is lined with malls on both sides, one after the other, and I don’t think the major malls have any informal agreements about who gets what.

If malls specialize, it’s a matter of reputation, and not collusion. For example, Lucky Plaza is a low-end mall, and caters to immigrant maids on their days off. A high-end store can open a branch at Lucky Plaza if it wants, but the high-end customers aren’t going to go there spontaneously, because most of what they’re interested in is in nearby high-end malls like Paragon and Wisma Atria. It works the same way with shopping streets in cities without much of an enclosed mall culture, like New York: the high-end clothing stores stay on Fifth Avenue and don’t go to streets even though street rents are massively cheaper, because the high-end customers walk on Fifth Avenue to look for clothes and aren’t going to walk on a street to window-shop.

D in Bushwick December 8, 2014 - 1:38 pm

Is it time for PATH Hub Sagas?

Frank December 9, 2014 - 7:06 am

I have to agree since he is always bringing this up.

j.b. December 11, 2014 - 4:58 pm

When $4 billion transit dollars are spent so uselessly in a transit-dependent city like ours, the scale of the lost opportunity resists immediate comprehension.

BruceNY December 8, 2014 - 1:50 pm

While in the meantime, the PA has just raised the tolls on the Hudson River crossings to $14!!!

SEAN December 8, 2014 - 7:59 pm

And soon enough it will be $15!

Nathanael December 10, 2014 - 3:28 am

Still free to cross the Verazanno eastbound and the Holland Tunnel westbound, though. So the Truck Loophole persists.

paulb December 10, 2014 - 8:35 am

I have nothing to contribute except, Ben, super photo!


Leave a Comment

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More

Privacy & Cookies Policy