That last part is pretty important. The authors think you could do it for under $20 billion dollars. That’s a lot of money, but it’s about five times less than the budget Amtrak says it would require. What’s the difference? How is it that when Amtrak gets asked to price out high-speed rail, it gives a quote that much higher?
We brought in Alon Levy, transit guru and the lead author of the report, to answer the question, and to explain a bunch of transit facts to a layman like me. Is this project actually technically feasible? And, if it is, could it actually work politically? (...)
I’m excited for this conversation, largely because although I'm not really a transit nerd, I enjoyed this report from you and your colleagues at the Transit Costs Project. But it's not really written for people like me. I'm hoping we can translate it for a more general audience.
The report was pretty technical. We wrote the original Transit Costs Project report about the construction cost of various urban rail megaprojects. So we were comparing New York and Boston projects with a selection of projects elsewhere: Italian projects, some Istanbul subway and commuter rail tunnels, the Stockholm subway extension, and so on.
Essentially the next step for me was to look at how you would actually do it correctly in the US, instead of talking about other people's failures. That means that the report on the one hand has to go into broad things, like coordination between different agencies and best practices. But also it needs to get into technical things: what speed a train can go on a specific curve of a specific radius at a specific location. That’s the mood whiplash in the report, between very high-level and very low-level.
The report was pretty technical. We wrote the original Transit Costs Project report about the construction cost of various urban rail megaprojects. So we were comparing New York and Boston projects with a selection of projects elsewhere: Italian projects, some Istanbul subway and commuter rail tunnels, the Stockholm subway extension, and so on.
Essentially the next step for me was to look at how you would actually do it correctly in the US, instead of talking about other people's failures. That means that the report on the one hand has to go into broad things, like coordination between different agencies and best practices. But also it needs to get into technical things: what speed a train can go on a specific curve of a specific radius at a specific location. That’s the mood whiplash in the report, between very high-level and very low-level.
I think you guys pulled it off very well. Let's get into it — I'll read a passage from the intro:
“Our proposal's goal is to establish a high-speed rail system on the Northeast Corridor between Boston and Washington. As the Corridor is also used by commuter trains most of the way… the proposal also includes commuter rail modernization [speeding up trains], regularizing service frequency, and… the aim is to use already committed large spending programs to redesign service.”
As a result, you think we could get high-speed rail that brings both the Boston–New York City trip and the New York City–Washington trip under two hours. You'd cut more than a third of the time off both those trips.
And here’s the kicker: you argue that the infrastructure program would total about $12.5 billion, and the new train sets would be under $5 billion. You're looking at a $17–18 billion project. I know that's a big sticker price in the abstract, but it's six to eight times cheaper than the proposals from Amtrak for this same idea. That’s my first question: Why so cheap?
First of all, that $18 billion is on top of money that has already been committed. There are some big-ticket tunnels that are already being built. One of the things that people were watching with the election was if the new administration was going to try to cancel the Gateway Tunnel, but they seem to have no interest in doing so. Transportation Secretary Sean Duffy talks about how there’s a lot of crime on the New York City subway, and how liberals want people to ride public transportation more and to drive less, but I have not seen any attacks on these pre-existing projects. So, as far as I’m concerned, they’re done deals.
The second thing is that along the length of the Northeast Corridor, this investment is not all that small. It’s still less than building a completely new greenfield line. With the Northeast Corridor, most of the line pre-exists; you would not need to build anything de novo. The total investment that we’re prescribing in Massachusetts, Rhode Island, New Jersey, Pennsylvania, Delaware, and most of Maryland is essentially something called a track-laying machine.
The Northeast Corridor has this problem: Let’s say that you have a line with a top speed of 125 mph, and the line has six very sharp curves that limit the trains to 80 mph. If those six curves are all within a mile of each other, there’s one point in the middle of the line where you have six 80 mph curves. That couple-mile stretch is 80 mph, while the rest of the line is 125. Now, what happens if these curves are evenly spaced along the line?
And here’s the kicker: you argue that the infrastructure program would total about $12.5 billion, and the new train sets would be under $5 billion. You're looking at a $17–18 billion project. I know that's a big sticker price in the abstract, but it's six to eight times cheaper than the proposals from Amtrak for this same idea. That’s my first question: Why so cheap?
First of all, that $18 billion is on top of money that has already been committed. There are some big-ticket tunnels that are already being built. One of the things that people were watching with the election was if the new administration was going to try to cancel the Gateway Tunnel, but they seem to have no interest in doing so. Transportation Secretary Sean Duffy talks about how there’s a lot of crime on the New York City subway, and how liberals want people to ride public transportation more and to drive less, but I have not seen any attacks on these pre-existing projects. So, as far as I’m concerned, they’re done deals.
The second thing is that along the length of the Northeast Corridor, this investment is not all that small. It’s still less than building a completely new greenfield line. With the Northeast Corridor, most of the line pre-exists; you would not need to build anything de novo. The total investment that we’re prescribing in Massachusetts, Rhode Island, New Jersey, Pennsylvania, Delaware, and most of Maryland is essentially something called a track-laying machine.
The Northeast Corridor has this problem: Let’s say that you have a line with a top speed of 125 mph, and the line has six very sharp curves that limit the trains to 80 mph. If those six curves are all within a mile of each other, there’s one point in the middle of the line where you have six 80 mph curves. That couple-mile stretch is 80 mph, while the rest of the line is 125. Now, what happens if these curves are evenly spaced along the line?
You have a way longer commute, right?
Yes. If you have to decelerate to 80 mph and back five times, that’s a lot slower. That’s the problem in the Northeast Corridor: there are faster and slower segments. Massachusetts is faster. Rhode Island is mostly fast. Connecticut is slow. If you have a line that’s slow because you have these restrictions in otherwise fast territory, then you fix them, and you’ve fixed the entire line. The line looks slow, but the amount of work you need to fix it is not that much.
The Northeast Corridor (red is stretches with commuter rail)
In southeast Connecticut, the curves are sharp, and there’s no way to fix any of them. This is also the lowest-density part of the entire Northeast: I-95, for example, only has four lanes there, while the rest of the way, it has at least six. I-95 there happens to be rather straight, so you can build a bypass there. The cost of that bypass is pretty substantial, but that’s still only about one-sixth of the corridor. You fix that, and I’m not saying you’ve fixed everything, but you’ve saved half an hour.
Your proposal is not the cheapest possible high-speed rail line, but I want to put it in context here. In 2021, there was a big proposal rolled out by the Northeast Corridor Commission, which was a consortium of states, transit providers, New Jersey Transit, Amtrak, and federal transportation agencies. Everybody got in on this big Connect Northeast Corridor (Connect NEC) plan, and the top line number was $117 billion, seven times your proposal. And this is in 2021 dollars.
They didn’t think that they could do Boston to New York and New York to DC in two hours each, either. There are two different reasons for their high price tags. The first reason is that they included a lot of things that are just plain stupid.
For example, theirs involved a lot of work on Penn Station in New York. Some of it is the Gateway Project, so that money is committed already, but they think that they need a lot beyond the tunnel. They have turned Gateway into a $40 or $50 billion project. I’m not going to nitpick the Gateway spending, although I’m pretty sure it could be done for much cheaper, but they think they need another $7 billion to rebuild Penn Station, and another $16 billion to add more tracks.
And you don’t think that’s necessary.
No. We ran some simulations on the tracks, and it turns out that the Penn Station that currently exists, is good enough — with one asterisk — even if you ran twice as much service. You can’t do that right now because, between New Jersey and New York Station, there is one tunnel. It has two tracks, one in each direction. They run 24–25 trains per hour at the peak. This is more or less the best that can be done on this kind of infrastructure. (...)
Unfortunately, they think Penn Station itself can’t handle the doubled frequency and would need a lot of additional work. Amtrak thinks that it needs to add more tracks by condemning an entire Midtown Manhattan block south of Penn Station called Block 780. They’re not sure how many tracks: I’ve seen between 7 and 12.
To be clear, the number of additional tracks they need is 0, essentially because they’re very bad at operations.
Well, let’s talk about operations. You say one way to drive down the cost of high-speed rail is just better-coordinated operations for all the trains in the Corridor. The idea is that often fast trains are waiting for slow trains, and in other places, for procedural reasons, every train has to move at the speed of the slowest train that moves on that segment.
What’s the philosophical difference between how you and the rail managers currently approach the Corridor?
The philosophical difference is coordinating infrastructure and operations. Often you also coordinate which trainsets you’re going to buy. This is why the proposal combines policy recommendations with extremely low-level work, including timetables to a precision of less than a minute. The point of infrastructure is to enable a service. Unless you are a very specific kind of infrastructure nerd, when you ride a train, you don’t care about the top speed, you don’t care about the infrastructure. You care about the timetable. The total trip time matters. Nobody rides a TGV to admire all the bridges they built on the Rhone.
I think some people do!
I doubt it. I suspect that the train goes too fast to be a good vantage point.
But as I said, you need 48 trains per hour worth of capacity between New Jersey or Manhattan. You need to start with things like the throughput you need, how much you need to run on each branch, when each branch runs, how they fit together. This constrains so much of your planning, because you need the rail junctions to be set up so that the trains don’t run into each other. You need to set up the interlockings at the major train stations in the same way. When you have fast and slow trains in the same corridor, you need to write timetables so that the fast trains will not be unduly delayed.
This all needs to happen before you commit to any infrastructure. The problem is that Connect NEC plans (Connect 2035, 2037) are not following that philosophy. They are following another philosophy: Each agency hates the other agencies. Amtrak and the commuter rail agencies have a mutually abusive relationship. There’s a lot of abuse from Amtrak to various commuter rail operators, and a lot of abuse by certain commuter rail operators, especially Metro North and Connecticut DOT against Amtrak. If you ask each agency what they want, they’ll say, “To get the others out of our hair.” They often want additional tracks that are not necessary if you just write a timetable.
To be clear, they want extra tracks so that they don’t have to interact with each other?
Exactly. And this is why Amtrak, the commuter railways, and the Regional Plan Association keep saying that the only way to have high-speed rail in the Northeast Corridor is to have an entirely separate right of way for Amtrak, concluding with its own dedicated pair of tunnels to Penn Station in addition to Gateway.
They’re talking about six tracks, plus two tracks from Penn Station to Queens and the Bronx, with even more urban tunneling. The point is that you don’t need any of that. Compromising a little on speed, the trip times I’m promising are a bit less than four hours from Boston to Washington. That’s roughly 180 kilometers an hour [~110 mph]. To be clear, this would be the slowest high-speed line in France, Spain, or Japan, let alone China. It would probably be even with the fastest in Germany and South Korea. It’s not Chinese speed. For example, Rep Moulton was talking about high-speed rail a couple of months ago, and said, “This is America. We need to be faster. Why not go 200, 250 mph?” He was talking about cranking up the top speed. When we were coming up with this report, we were constantly trying to identify how much time a project would save, and often we’d say, “This curve fix will speed up the trains by 20 seconds, but for way too much hassle and money.” The additional minutes might be too expensive. Twenty seconds don’t have an infinite worth. (...)
I want to go back to something you said earlier. You were contrasting the aesthetic of this proposal with Representative Moulton’s proposal, who wants our top speeds to be faster than Chinese top speeds. How do you get voters to care about — and I mean this descriptively — kinda boring stuff about cant angles?
Voters are not going to care about the cant angle efficiency on a curve. They’re not going to care about approach speed. However, I do think that they will if you tell voters, “Here's the new timetable for you as commuters. It looks weird, but your commute from Westchester or Fairfield County to Manhattan will be 20 minutes faster.”
With a lot of these reports, the issue is often that there are political trade-offs. The idea of what you should be running rail service for, who you should be running it for, that ended up drifting in the middle of the 20th century.
But also, the United States is so far from the technological frontier that even the very basics of German or Swiss rail planning, like triangle planning of rolling stock/infrastructure/operations, that's not done. Just doing that would be a massive increase in everything: reliability, frequency, speed, even in passenger comfort.
The main rail technology conference in the world, it's called InnoTrans, it's in Berlin every two years. I hear things in on-the-floor interviews with vendors that people in the United States are just completely unaware of.
by Santi Ruiz and Alon Levy, Statecraft | Read more:
Image: uncredited
[ed. Fascinating stuff! (I think, anyway). And, for something completely different, see: How to Be a Good Intelligence Analyst (Statecraft):]
***
I think the biggest misconception about the community and the CIA in particular is that it's a big organization. It really isn't. When you think about overstuffed bureaucracies with layers and layers, you're describing other organizations, not the CIA. It is a very small outfit relative to everybody else in the community. (...)What kinds of lessons were consistently learned in the Lessons Learned program?
There's an argument that the lessons learned are more accurately described as lessons collected or lessons archived, rather than learned.
Because learning institutionally is hard?
Learning institutionally is hard. Not only is it hard to do, but it's also hard to measure and to affect. But, if nothing else, practitioners became more thoughtful about the profession of intelligence. To me, that was really important. The CIA is well represented by lots of fiction, from Archer to Jason Bourne. It's always good for the brand. Even if we look nefarious, it scares our adversaries. But it's super far removed from reality. Reality in intelligence looks about as dull as reality in general. Being a really good financial or business analyst, any of those kinds of tasks, they're all working a certain part of your brain that you can either train and improve, or ignore and just hope for the best.
Learning institutionally is hard. Not only is it hard to do, but it's also hard to measure and to affect. But, if nothing else, practitioners became more thoughtful about the profession of intelligence. To me, that was really important. The CIA is well represented by lots of fiction, from Archer to Jason Bourne. It's always good for the brand. Even if we look nefarious, it scares our adversaries. But it's super far removed from reality. Reality in intelligence looks about as dull as reality in general. Being a really good financial or business analyst, any of those kinds of tasks, they're all working a certain part of your brain that you can either train and improve, or ignore and just hope for the best.
I don't think any of those are dull, but I take your point about perception vs. reality.
I don't mean to suggest those are dull, but generally speaking, they don't run around killing assassins. It's a lot less of that.
