Loading summary
Transcript71 lines
- [00:00]
Alon Levy
Foreign.
- [00:04]
Santi Ruiz
Hi, I'm Santi Ruiz and you're listening to Statecraft. At the very end of April of this year, a report from the Transit Costs Project came out. It's called how to Build High Speed Rail on the Northeast Corridor. As the name suggests, the authors of the report had a simple goal. The stretch of the US from Washington D.C. and Baltimore through Philadelphia to New York and then up to Boston is the densest stretch of the country. It's an ideal location for high speed rail. How could you actually build trains to get you from D.C. to New York in two hours, or New York to Boston in two hours without breaking the bank? That last part is pretty important. The Transit Costs Project team, the authors of this report, think you could do it for well under $20 billion. That's a lot of money, obviously, but it's about five times less than the budget Amtrak says it would need to do it. Well, 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 on Alon Levy, the lead author of the report, to explain a bunch of transit facts to a layman like me. Is this project actually technically feasible? Why are these numbers so far apart? And if this project could actually work, what would it take to actually get it done politically? As a reminder, this is the fourth installment in a Statecraft mini series we've been doing on transit. I'm really happy with how these episodes have turned out, and if you like this one, check out the others. I'll link them in the show notes. One is with the head of Massachusetts Transportation on saving a doomed project, another with Biden's head of the Federal Transit Administration, and another with the head of DC Metro. You can find all these episodes at www.statecraft.pub. if you like Statecraft, give us a rating on Spotify or on Apple Music or wherever it is you're actually listening. Without further ado, Alon Levy Alon, it's good to have you on.
- [01:47]
Alon Levy
Thank you, Santi, for having me on.
- [01:49]
Santi Ruiz
I'm really excited for this conversation, largely because I'm not really a transit nerd. But I would say, and I think you would probably self identify as one, yes I am. And I really enjoyed this report of yours. It's called how to Build High Speed Rail on the Northeast Corridor. It's from you and a couple other folks at the Transit Cost Project and it's a fantastic report. I really enjoyed reading it as a lay reader, but it's not really written for people like me. And I'm hoping we can do a little translation of it for an audience that's interested in policy and I think largely in transit, but maybe not have the familiarity with rail junctions that a lot of your readers do.
- [02:23]
Alon Levy
Yeah, the, the report was, I mean, the report was technical, for there's kind of a historic reason, which is we wrote the transit cost project report about the construction costs of various urban rail mega projects. So we were comparing the New York and Boston project, the Second Avenue subway in New York and the Greenland extension in Boston with a selection of projects in places that we were looking at a bunch of Italian Metro projects and at some Istanbul subway and commute rail tunnels and at the extension of the subway in Stockholm and the recently opened extension of their commuter rails. Essentially, the next step for me was I was thinking, okay, instead of talking about other people's failures, I wanted to look at how would you actually do it correctly? And the problem is that means that the report on the one hand has to go into broad things about what needs to be done, things about coordination between different agencies and the best practices on that. But also it needs to get into technical things about what speed a train can go on on a specific curve of a specific radius at a specific location. That's the mood whiplash between things that are very high level, but also things that are extremely low level.
- [03:33]
Santi Ruiz
Well, I think you guys pull it off very well. And I'll just say diagnosis is always easier than prescription in policy. So I think it's always valuable to have actual blow by blow prescription, especially if it's technical and thorny. Let's get into it. I'll read a passage from the intro of this report, which again is about building high speed rail in the Northeast Corridor, and then I want you to flesh it out for me. So here's, here's, you quote, 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. You say the proposal also includes commuter to rail modernization, so speeding up the trains, regularizing service frequency. And you say for both intercity and commuter trains, 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 lets you do the Boston to New York City trip and the New York City to Washington trip each in under two hours. So you'd cut more than a third of the time off both those trips. I think the kicker here is the report argues that you could Put this together for an infrastructure program totaling about 12 and a half billion dollars and new high speed train sets that are under $5 billion, you're looking at a 17, $18 billion project all told. I know that's a big sticker price in the abstract, but it's six to eight times less than the proposals from Amtrak, for instance, for this same kind of idea. Why? Why so cheap?
- [04:55]
Alon Levy
So first of all, it's on top of things that have already been committed. So there are some big ticket tunnels that are being built, but the money is committed. There's already construction at this point. Even things that are very signature initiatives that a politician who wants to wreck things would wreck are not being wrecked. So one of the things that people were watching with the election was is the new administration going to try to cancel the Gateway Tunnel. But they seem to have no interest in doing so. Sean Duffy is talking 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 attack on these pre existing projects. So as far as I'm concerned, there are non deals. So part of it is some of the money has been committed and the 17 or 18 is on top. And the second thing is for a line the length of the Northeast Corridor, this investment is not all that small. It's still less than building a completely new greenfield line. Let's call this line Chicago to Cleveland. So you can do a system where you build a high speed line departing one of the Chicago urban stations like Millennium Station or Union Station, going east in northern Indiana and Ohio to Toledo and then you split to Cleveland and Detroit. That would be a rather cheap line to build. It's extremely easy terrain to build in. It's very flat. So you do any extensive earthworks. But it's also not a floodplain. So you don't need to worry about earthworks like the North China Plain floods too much. So their costs are rather high. But with the Northeast Corridor, most of the line pre exists. You would not need to build anything like that de novo. The total investment that we're prescribing in let's say Massachusetts, Rhode Island, Jersey, Pennsylvania, Delaware, most of Maryland is essentially run something called a track laying machine. The Northeast Corridor has this problem where probably the best way to describe it is let's say that you have a line with a top speed of let's say 125 miles an hour. Okay. Then the line has six curves that are very sharp and Limit the trains to 80 miles an hour? Sure. 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 curves, each of which is 80 miles an hour. What happens? That stretch of a couple miles is 80 miles an hour. The rest of the line is 125. Now, what happens if, instead, these curves are evenly spaced along the line?
- [07:31]
Santi Ruiz
You have a way longer commute, right?
- [07:33]
Alon Levy
Yeah, if they're evenly spaced along the line, essentially the amount of time that it takes you to go down from 125 to 80 and back, you need to do it five times. And if these are curves that are right next to each other, you need to only do it once. That's the problem. And the Northeast garter, there are faster and slower segments. As I said, Massachusetts is fast. Rhode island is mostly fast. Connecticut is slow. On the other hand, if you have a line that's slow because you have these short, slow restrictions in otherwise fast territory, then you fix them and you fix the entire line. So this is a line that looks slow. I mean, it is slow, though. Trip time is longer than you would like it to be. But the amount of work you need to fix it is not that much. And that is a lot of the situation with the Northeast corridor, the slowest part in terms of track geometry. Let's forget everything else right now and focus on sharp curves, because most of the reason that people will tell you that the line is slow is sharp curves.
- [08:26]
Santi Ruiz
Okay.
- [08:26]
Alon Levy
There are other things that can be done. I'm going to get to them in a bit. But the difficult stuff to do is fixing the sharp curves. So the area with the sharpest curves is between New Haven and southern Rhode Island. The curves essentially start widening around the point where you cross between Connecticut and Rhode island, and shortly thereafter into Rhode island transitions into the fastest part of the corridor in southeast Connecticut. The curves are sharp. There's no way to fix any of them. However, this is also the lowest density part of the entire Northeast. This is where i95, for example, only has four lanes. The rest of the way, it has at least six. So this is the least urbanized part of the corridor. I95 there happens to be rather straight. You just can build a bypass there, and the cost of that bypass is pretty substantial. But that's still only about one sixth of the corridor. So you fix that. And I'm not saying you fix everything, but you've saved half an hour.
- [09:23]
Santi Ruiz
Sure. Let me back up a little bit here, because I take Your point that it's still not the cheapest possible high speed rail line, but I just want to put it in in contrast here. So in 2021, as you know, there was a big proposal rolled out by the Northeast Corridor Commission, which was a consortium of all these states, the transit providers, NJ Transit, Amtrak, Federal transportation agencies, everybody got in on this big connect Northeast Corridor plan. And this is 2021. The top line number was it would cost $117 billion. So seven times your proposal. And this is in $2021.
- [09:57]
Alon Levy
And they didn't think that they New York and New York to D.C. and two hours each. It was actually slower than that.
- [10:03]
Santi Ruiz
Right.
- [10:04]
Alon Levy
The reason for their high price tags is there are two different reasons. The first reason is they included a lot of things that are just plain stupid. I'm sorry. So this involved, for example, a lot of work on Penn Station in New York. Some of it is the Gateway project. So that's part of the 1.17 and Gandalf committed already. But they think that they need a lot of stuff beyond the tunnel. They have turned Gateway into some, I think a 40 or $50 billion project. So the new tunnels, the money is committed. I'm not going to nitpick the cost, even though I'm pretty sure they can be done for much cheaper. It's about 16, but on top of that 16, they think they need seven to rebuild Penn Station and another 16 or 17 to add more tracks.
- [10:47]
Santi Ruiz
And you don't think that's necessary in.
- [10:49]
Alon Levy
Your proposal at all? We actually ran some simulations on the tracks and it turns out that the Penn Station that exists with one asterisk is good enough, even if you run twice as much service, which you can't right now because there's only one tunnel. So between New Jersey and New York Penn station there is one tunnel. It has two tracks, one in each direction. They run 24 to 25 trains per hour at the peak. This is more or less the best that can be done on this kind of infrastructure. If you've heard of places that run more trains per hour, these are not mainline trains. These are captive metros where it's much easier for each train to substitute for another train if they all make the same stops. Or maybe they have a two way branching and otherwise they make the same stops. So that's where you can crank it up. But on a highly branched mainline system, this is more or less the best that you can do. So they, for about 30 years have wanted to build a second pair of tracks this Used to be called arc, or access to the region's core. It was funded, but then Governor Chris Christie canceled it on grounds that there could be cost overruns that the state would be on the hook for. And then Amtrak took over the project, renamed it Gateway. The cost increased subsequently. So instead of two tracks, they're going to have four tracks. 24 times 2 is 48. They think that they need 48 trains per hour, rush hour into Penn Station, which actually makes a lot of sense if you are mildly optimistic about returning to the office in New York. You don't need to be aggressively optimistic about this at this point. But New Jersey is building a lot of housing by Northeastern American standards. Is it building like it's Texas? No. Is it building like it's Istanbul or Tokyo or the other other cities that advocates point to as places that build aggressively? No, it doesn't. But it still builds more than the rest of the Northeast. And the demand there is obviously mostly commuters to New York City. So in the medium term, ridership will increase as much as it can. Given the infrastructure. It's a sensible project to build this. Unfortunately, they think that Penn Station itself can't handle the doubled frequency and would need a lot of additional work. Amtrak thinks that they need to add more tracks through condemning an entire block south of Penn Station. It's called Block 780. Penn Station is between 7th and 8th Avenues, for the most part, between 31st and 33rd Streets. Block 780 is between 7th and 8th, between 30th and 31st. So they want to condemn an entire Midtown Manhattan block and build additional tracks, and they're not sure how many tracks. I've seen numbers between 7 and 12 additional tracks. To be clear, the number they need is zero additional tracks, essentially because they're very bad at operations.
- [13:48]
Santi Ruiz
So let's get into that. The operations side. There's a lot going on in the proposal. I think it's quite dense, and I'm going to oversimplify it here, but a big part of what you propose is you can get the cost of high speed rail in the Corridor way down, in part by just coordinating and better running operations for all the trains in the Corridor. And correct me if I'm wrong here, but the idea is that often fast trains are waiting for slow trains in other places. Every train has to move at the speed of the slowest train that moves along that segment for reasons that can be solved. Will you just, at the 30,000 foot level, articulate what's the difference between a proposal like yours and how Amtrak and the other rail managers currently approach this corridor. What's the philosophical difference here?
- [14:33]
Alon Levy
So the philosophical difference is coordinating infrastructure and operations. Often you also coordinate the rolling stock decisions, as we say, which train sets you're going to buy. It's a triangle. So this is why the proposal combines these policy recommendations with extremely low level work, including timetables to precision of less than a minute. The point of infrastructure is to enable a service. Unless you're 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. Yes, places that are trying to brag about their technology will have a speedometer that you can view from inside the train. You can see that it's going 300 kilometers an hour or more in China. But fundamentally, when you ride, you don't care. You care about the timetable. Now, the timetable also has an arrival time minus departure time. So the total trip time matters. So the average speed matters, but the top speed, the infrastructure, these don't matter. Nobody rides a TGV to admire all of the bridges that it built on the Rhone.
- [15:41]
Santi Ruiz
I think some people do.
- [15:43]
Alon Levy
I doubt it. I suspect that if you're trying to look at the bridges on the lawn, you will not lose it from the train because the train goes too fast. You will find a good vantage point where you can see the trains from the lawn. So you're by definition not riding these trains at the same time. So, as I said, you need 48 trains per hour worth of capacity between New Jersey or Manhattan. You can maybe get a little bit less if you need to make a compromise. You don't, but if you do, you can do 42, 44 or something. But you need to start with things like, this is the throughput that you need, this is how much you need to run on each branch so that you know how to run each branch separately. And of course, they share tracks. So you need to figure out when each branch runs and 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 are not going to run into each other. You need to set up the interlockings at the major train stations, let's say Penn Station, Grand Central, South Station, Washington, Union Station, in the same way. And when you have fast and slow trains sharing the same corridor, then you need to write timetables so that the fast trains will not be unduly delayed. And all of these need to happen before you commit to any Infrastructure. The problem is that the Connect nec plans, Connect 2035, Connect 2037, they're not following that philosophy. They're following a philosophy of each agency hates the other agencies. Ambrec and the commuter rail agencies have a mutually abusive relationship. There's a lot of abuse by Amtrak against various commuter operators. There's a lot of abuse by certain commuter operators, especially Metro north and especially Connecticut, DOD against Amtrak. So you ask each agency, what do you want? And they say, get the others out of our hair. They will want additional tracks that often are not necessary. If you just write a combined timetable.
- [17:41]
Santi Ruiz
Just to be clear, the point of the additional tracks would be, now we get to ride separately. We don't ever have to think about them again.
- [17:47]
Alon Levy
Exactly.
- [17:47]
Santi Ruiz
We can separate out our domains.
- [17:49]
Alon Levy
Exactly. And this is where you're getting Amtrak. The commuter railways, the Regional Plan association, they keep saying that the only way to really have high speed rail in the Northeast Corridor is to have an entirely separated right of way for Amtrak, including its own dedicated pair of tunnels to Penn Station in addition to Gateway. So now they're talking sort of four tracks between New Jersey and Penn Station. They're talking about six tracks and also two more tracks from Penn Station to Queens and the Bronx with even more urban tunneling. And the point is that you don't need any of that compromising a little bit on speed. The trip times that I'm promising are a bit less than four hours. Boston to Washington, that is, what is it, 180 something kilometers an hour? Miles? 115 maybe. To be clear, this would be the slowest high speed line in France or Spain or Japan, let alone China. It would be probably about even with the fastest in Germany and South Korea, which are not as fast. It's not Chinese speed or anything like that. For example, Congressman Malden, he was talking about high speed rail a couple months ago and said, this is America, we need to be faster. Why go 200, not 250 miles an hour? And he was talking about trying to crank up the top speed and the average speed. But this is a point where you can keep cutting minutes from what I'm proposing, but the additional minutes might be too expensive, especially as a startup. So constantly what we were thinking when we were coming up with these lists of infrastructure projects was, how much time does this save? And often we how to say, okay, this is a curve fix. It's going to speed up the trains by 20 seconds for way too much hassle and way too much money. 20 seconds don't have an infinite Worst. I'm sorry. So there's coordination between infrastructure and operations. There's coordination between intercity and commuterial planning. There's some kind of coordination between how you're going to maintain the system and additional construction. These are two separate things that unfortunately are being bundled together in America, but not coordinated. Does that make sense to you?
- [19:44]
Santi Ruiz
It doesn't. You're going to have to explain that.
- [19:46]
Alon Levy
Okay, so maintenance is a recurring expense. Now it's treated as a capital expense because let's say that I'm running the trains in a city. If I don't spend any money on operations, if I'm deciding not to pay the train drivers for the next three months, then I'm not going to have trains running. The train drivers don't work for free. If I decide not to do maintenance for the next three months, then I could do maintenance the subsequent three months and I can defer it to later. It will cost me more in the long run, but sometimes it's worthwhile to take this cost.
- [20:16]
Santi Ruiz
As a side note, we say this a lot in parenting. When you're trying to decide do you use the screen at this particular moment? You just have to understand it as a loan. You know you will pay for it. But sometimes it's okay.
- [20:26]
Alon Levy
Yeah, exactly, exactly. Everything is trade offs. And here the trade off with very routine maintenance can be in a matter of months, but something that can be a matter of years or even decades. So there is a temptation to defer maintenance, especially the long term maintenance, especially renewal, which involves more long term things. Often they might even require a shutdown of the line for a few weeks or even months while you rebuild the infrastructure. These are decisions that can be very easily deferred over many years or even more. And okay, you defer maintenance in the short run and then in the long run your system collapses. There was a solution to this in the 80s, which was called State of Good Repair on the New York City subway, where they came with these five year capital plans that characterize how subway investment is done today. So up until then, operations and maintenance were funded as one thing. And then outside capital intuition could go to expansion. For example, they constantly wanted to build Second Avenue subway in the 50s, 60s, 70s and in the 80s, because the system was collapsing, one third of it was under slow restrictions. And when I say slower sections, I don't mean you're supposed to go 50 miles an hour and you go 40, I mean you go 10, 15 miles an hour. The trains broke down in 1982. Every 6.2 thousand miles that seems like.
- [21:43]
Santi Ruiz
A lot of breaking down.
- [21:44]
Alon Levy
Yes. The breakdown rate after they fixed the system by the 2000 2010s was breakdown every somewhat more than 100,000. So maybe 110,000 miles between breakdowns. You could do better. There are places that scratch the seven digits. Yeah, but at the end of the day, six digits means you see a breakdown on your own train once in a generation, which means you're stuck behind a train that's broken down maybe once every year with 6,000 miles. You see a breakdown probably once every year, and you get stuck behind a train that's broken down every week or two. So they had this program where they decided to do long term capital spending on renewal and maintenance. And it worked. Unfortunately, after it worked, everyone else copied that and they copied it in very different circumstances in which they essentially reduced maintenance to a capital megaproject. Amtrak did that in the 2000s. So the chair of Amtrak, the president, David Cohn, refused to defer maintenance. He said, no, maintenance is the priority. And at the time it was the Bush administration, they were trying to privatize Amtrak. And when you intend to privatize, you will defer maintenance. It's not even unique to Amtrak or to the United States. American railways in the 1960s often deferred maintenance in order to look financially good for a merger. The Milwaukee Railroad was doing that in the 60s. It kept looking for a merger and deferred maintenance through that. And because the mergers kept getting rejected and kept looking for bigger, bigger mergers, the result was that their maintenance collapsed after 10 years and the railway just went bankrupt in most of its fixed plant is at this point bike trails. So the private sector does that as well. The public sector outside the US does that as well. So anyway, David Gahn refused to defer maintenance, so the political fired him. The administration has changed. The Obama administration was not interested in privatizing Amtrak, but was interested in investing in high speed rail, which in the 2000s was viewed as kind of a flashy thing. America is the greatest country in the world, but could learn from French and Japanese rail speeds. So Amtrakride poverty and said, oh look, we have this big maintenance backlog. Please give us money for State of Good Repair. Essentially it became a way to spend megaproject money with nothing to show for it.
- [23:55]
Santi Ruiz
Let me cut in here. What should we do with this existing system of State of Good Repair funding where I apply for megaproject money to do this kind of maintenance? What's the TLDR for this proposal?
- [24:07]
Alon Levy
Outside funding. That would be federal funding. If It's a state or local agency should not go to state of good repair. It should go to capital spending that you can say no to. So expansion is such an example. It can be something that you can maybe even bond around for better operations, but not ongoing maintenance. Instead, the federal government should demand that maintenance be funded on an ongoing basis locally or regionally or statewide. Again, this doesn't cover amdrac, which is federal. I can tell you what can be done with mainline rail, let's say commuter rail. These state regional rail programs like the Hartford line in Connecticut, which is you want maintenance to be done in a very mechanized fashion with track laying machines which go on the tracks. They travel to the work site at the speed of a slow crane. Then they do the work at the speed of a very slow walker. They do about a mile an hour at best, one third of a mile an hour because they're literally changing everything. They maybe unscrew the two rails from the cross ties and then they replace the cross ties. Obviously you can't go at any real speed on that because then you will wreck the rails. They will just shift, they will go away, they will move outside. So you need to go at an extremely slow and methodical speed. And then you replace the rails, you replace the ties. You can even change. So curves are always banked or canted. The outside rail is a little higher than the inside rail. The reason is when you're going on the curve, you have centrifugal force. So you're going to use gravity to partly countermand that force. You can change the cant angle. The cant in the United States is not as aggressive as it can be, I'm guessing.
- [25:50]
Santi Ruiz
Sorry. Just maybe 3% of our listeners will know what a cant angle is real quick.
- [25:54]
Alon Levy
So you see this a lot with racetracks as well, or on interstates. Honestly, on curves, curves are banked, which means that the outside of the curve is higher than the inside of the curve. So if you're standing on this curve, you're not moving, you're standing. Then gravity is going to pull you a little bit to the inside. But if you're moving at a high speed, then centrifugal force is going to pull you to the outside. So you do this. It's called banking on a road or in physics classes and it's called canting on a rail.
- [26:22]
Santi Ruiz
And we're just not as aggressive about the angle of that cant as many, many other countries.
- [26:28]
Alon Levy
Exactly. Now, American regulations are not different from the normal European regulations on this But American practice is constantly too conservative. They don't go as far as the regulations will permit. And it's another way that you can speed things up cheaply by running the trains as fast as the regulations permit and not as fast as accumulated practices over the last couple decades of American railroading say that you can. There are all these 10 or 15 mile an hour restrictions and station floats that don't need to be there. And yeah, it doesn't sound flashy if you're going from 10 miles an hour to 30, but you can save four minutes on just the last mile into Grand Central. On every kilometer train on Metronox, you can save two minutes between Back Bay and South Station, both intercity trains into Boston.
- [27:13]
Santi Ruiz
Well, let me zoom out and I want to go back to something you said earlier when you were contrasting the way this proposal looks and the kind of the aesthetic of this proposal with people like Representative Moulton who say things like, why can't we get our top speeds faster than the Chinese top speeds? Because it really is striking. When you read your proposal, which is this 150 page think tank report, the top line takeaways to me are very attractive, very striking. You know, you save an hour on each of these trips. You do it for way cheaper than the standard proposals. Then you get into it. And a lot of the actual meat and bones of the proposal is, and I say this neutrally, but it's much more boring, right? It's not building new infrastructure largely. There's a couple places where you do, but you pick those really specifically. We'll come back to the political coordination because I have a lot of questions about how you could actually implement that, whether you could. But a bunch of the rest of it is this stuff that's incredibly arcane, that's not interesting at all to voters. When you think about voters are not.
- [28:10]
Alon Levy
Going to care about cat efficiency on a curve. They're not going to care about approach speed. However, I do think that if you actually tell voters, okay, here's a new timetable for you as commuters. Yeah, it looks weird and the trains are going to be less express. But your commute from your suburb in let's say Westchester or Fairfield county or in or near New Jersey to Manhattan is going to be 50 minutes faster, 20 minutes faster in Boston to Providence, the current trip time is an hour 10, more or less. I think that it can be done in around 50. And this is with the trains holding for overtakes by high speed trains twice. Wow.
- [28:46]
Santi Ruiz
Matty Glaciers had a good write up of this report, and it's a pretty short one, but he just flags that in his view, this kind of report is relatively undersupplied. Nonprofits don't fund it to the extent they should. The kind of analytical what would you do if you were actually interested in increasing speed and reliability and those were your two focuses? I'm curious, do you think that's true? Is this kind of work undersupplied? And if so, why?
- [29:10]
Alon Levy
I don't know enough about the American NGO funding world to be able to give you a very strong answer. I think I may typically technical in how I think about these kinds of policy things. You are with a lot of these reports the issue is often 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 kind of drifting in the middle of the 20th century onward. So at this point, quite a lot of the technology involved is one where the United states is literally 70 years behind best practices. In this case, the US is just so far from the technological frontier that even things like the very basics of German or Swiss rail planning, like this triangle planning of rolling stock infrastructure and operations, that's not done. And just doing that would be a massive increase in everything, honestly, and reliability and frequency and speed, even in passenger comfort. Usually you slow down the trains on curves because of passenger comfort issues, right? You don't need to make these trade offs in the OS relative to where you guys are right now. Because the OS is so far from the technological frontier, it gets to the point where the main rail technology platform in the world is called Enotrans. It is in Berlin every two years. And there are things there that I hear on the floor, interviews with people who work for these vendors as things that people in the United States are just completely unaware of. And that's going to matter for things that are really boring, like how you rebuild the overhead catenary that provides the electricity that the trains are using.
- [30:46]
Santi Ruiz
I'll confess I had to read that section twice because the first time I could not pay attention because this is.
- [30:51]
Alon Levy
To be clear, boring. I am a nerd for public transportation, but not for that. I had to outsource some of the nerdery to people who are that kind of overhead wire nerds. But that is actually important just because the OS is so far behind this cutting edge of technology. The basic kind of catenary that exists on not just high speed, but practically every low speed regional line in Europe, that would be a major innovation at rather low cost. I mean, I keep talking about construction costs and how for our previous report, Second Avenue subway in New York was least per mile was in today's money, I'm going to say 3.2, maybe $3.3 billion a mile. And the world median, is it maybe 500, 400, 500? I think it's 500. And if you exclude American projects or projects in countries where the predominant language is English, then you're getting less of that. The high construction cost problem is specifically New York, but the rest of the English speaking world is not too far behind, unfortunately. Los Angeles, San Francisco, Boston, London, Toronto, Hong Kong, Singapore, Melbourne, Sydney. So this is a factor of 7, 8. We found a factor of about 9, premium unconstruction costs for the subway in New York. You don't have these factor of nine differences in other things.
- [32:16]
Santi Ruiz
Before I move on to the political constraints of this proposal, which I think are real, and I want to get into with you what that would actually look like to solve the politics of this. Will you just give me just really briefly, what are some of the other kinds of wonky or quote unquote boring things that you come up with in the proposal as a result of focusing really exclusively on how can you actually get reliability and frequency? I mean, we've talked about, sure, it sounds obvious, but you focus on increasing speed in slow areas.
- [32:41]
Alon Levy
Things about things about regulations on speed on a given curve. Just because you can turn a 70 mile an hour curve into a normal 100 mile an hour curve. And yeah, a hundred miles an hour is not an especially flashy speed on a train.
- [32:55]
Santi Ruiz
Well, you, as you pointed out, if you're going 70, you get to 100. That's huge.
- [32:59]
Alon Levy
Exactly. It's huge. So that's the bigger thing, the other boring things. So as I mentioned, there's the catenary issue. There's something about rail junctions and timetabling around bottlenecks. Sometimes you do need to build additional flyovers.
- [33:14]
Santi Ruiz
And just for listeners, a flyover is where instead of the train tracks being crossed physically, the trains go at different levels and they never touch each other.
- [33:23]
Alon Levy
Exactly. Same with, same with highways. I mean, the point of the Interstate system was to build flyovers because they recognize that first of all, when two highways cross at grade with, let's say, stoplights, that first of all is very slow. And second, that causes a lot of accidents. The Interstate system is much safer than the surface roads. So these are probably the biggest things.
- [33:44]
Santi Ruiz
A couple other ones that you mentioned, and I'll just throw these out for listeners. There's stuff like, and you've talked about this quite a bit, but elevating platforms so that there's no step up to get onto trains, and that saves you time on how long the train has to be in the station for everybody.
- [33:57]
Alon Levy
It's also an important accessibility feature that unfortunately, the accessibility is sometimes ghettoized as just a civil rights ADA issue. Unfortunately, the agencies have gotten used to not being proactive about it. They just do the minimum. They won't get them sued. So the they don't think proactively how to make it easier to get on the trains. The Paris Metro is the least accessible in the world and I'm not disabled. But if I'm flying, I have luggage. I don't want to drag this luggage down multiple flights of stairs. If you're pushing a stroller, same thing.
- [34:34]
Santi Ruiz
Yeah. As someone who roughly once a month boards the Northeast Regional train in Washington, D.C. with a stroller, this is a personal pet peeve of mine. I would be very happy to see that platform elevated. One of the big challenges, I think, is you propose a full integration of planning across seven or eight states. Amtrak, the fta, the Feds, and that if you integrated all those organizations in a certain way and you actually synthesize that work, you could get a lot of gains just from timetabling. For instance, in practice these days, when those organizations coordinate, what you see is something like the Connect 2035 project, like, that's a. That's a coordination, an existing coordination of those states and all the other relevant players. And the project they give you is. I don't know if you'd call it a boondoggle, but it's the whole wish list of all these agencies. Everybody gets their own priorities. You don't get this synthesized, efficient model. So break that down for me. Why be optimistic that you could do this?
- [35:31]
Alon Levy
The problem is this is a stable job. It's a staple job of New Jersey Transit's list of priorities, some of which are very good, and made our proposal. You staple that with Metro north, both in New York City and the Connecticut side's list of priorities, a few of which are good, most of which are terrible. You staple that with mbta, Wishlist, with septa, marc, Amtrak, some state dots that are maybe even working independently of their commuter rail agencies, because they're not used to coordinating, because coordination is not something natural coordination kind of needed to be invented. And in the United States, it was never invented because the railroads were used to competing against each other, and then they were used to competing against, for Example Greyhound. So they did not coordinate intercity bus and intercity rail service. To this day, these are competitive rather than in any kind of combined system. The commuter railways in the United States did not think about coordination very much because of the demographics of the riders. The commuter rail in the United States is generally designed around Don Draper and Pete Campbell from Advent. So high income, nine to five workers who live in the suburbs because that's where you're expected to live as a middle class American way. More people commuted at rush hour to these city centers in the 1960s than do today. Today more people commute to the Manhattan core, which is Manhattan south of, I never remember 59th or 60th street south of Central Park. Basically, yeah. Manhattan south of Central park gets more commuters today, and by today I mean 2019 than they did in the 60s, but fewer of them, not a lower percentage, fewer in absolute numbers arrive in the peak hour than did in the 1960s. But the mentality is still very 9 to 5. And they didn't think about how to coordinate because they didn't think they needed to. And this created a two tier planning system where the buses were viewed as for the poor. And to this day, I mean, for example, in, not in Boston, the city in the MPTA zone, apparently they're not allowed to coordinate their commuter rail schedules with the buses. It's considered something different. So a lot of it's just needing to invent this kind of coordination, but you don't need to invent it. It was done already. It was done mostly in Germany in the 1970s to create the modern intercity rail network in Germany, which is not very fast, it's just much better coordinated than anywhere else. So the ridership is the same as in France per capita, and France has way faster trains.
- [38:06]
Santi Ruiz
So I take your point and I think there's a lot to like about that diagnosis of why we're incredibly uncoordinated in places where it would make sense to be coordinated. But I want to read you this is a quote that's been quoted on the on statecraft in the past. Jen, Paul and I have talked about it from a book about implementation and the challenges of it. And the upshot is, quote, no suggestion for reform is more common than what we need is more coordination. I also would love coordination, but I think by definition when we look at things like Connect 2035, that is the current state of coordination among a bunch of rail agencies that have, as you say, an abusive relationship with each other. Explain to me why this is not pie in the sky.
- [38:43]
Alon Levy
The reason it's not pie in the sky is they all get their money from the same source. The source being FRA funding for the Northeast Garder. They're not actually competing with each other. I mean sure they're competitive grants, but the FRA can actually indicate its priorities. And when it's important enough to the Feds, they educate these priorities. The example I'm thinking of is of the Northeast garter in the Biden administration. Secretary Buttigieg wanted to fund one big high speed rail line so that people would see that this is possible in America. And there was a question of which one would get funded. And the answer was bright analyst. It didn't get 100% funding, I think like 25%. And they think that they are going to be able to get the rest private at that level. It is actually possible for the top level officials who can be politicals or they can be the top level servants. I'm literally seeing these decisions made when they need to be because all of these agencies, they're getting their funds from the same pot of money. This is something where the FRA could be prioritizing. Not all, I think maybe about half of what we're proposing, not of the 17, but of the 12 and a half. So just the infrastructure costs are projects that have already been on the list. Just the highest priority projects on the list is junction fixes. Some of that is a novelty by which I mean they haven't applied for funding. But it's being planned. It's being stuck behind certain things like that bypass that I mentioned along i95. That is not something that I made it on the spot. This is something that was planned in the Northeast Carter future plans in the 2010s they ran into some NIMBYism in southeastern Connecticut. The alignment that I drew for this was specifically trying to mollify as far as possible. So the people in Old Sabre had three complaints and I think that what we drew addresses two of them. Okay. The third you can't. But there's a limit to how much you can modify. So most of it is something that is already there. It's just they don't know how to prioritize. Partly because they think that they can't prioritize. But again they think they can't coordinate. But again I'm saying they've coordinate. A lot of it is just knowing that it's possible. By it, I don't mean coordination, I mean it being this kind of high speed rail plan because as I said, it's something that's not the flashy thing that you see when you go on a three day junket to France to ride a TGV or on a three day junket to Japan to ride the Shinkansen. You need to actually drill down and talk to intermediate level people and see how the planning was done in France and what they did and didn't do. Or in Germany what they did and didn't do here. A lot of it is just not invented here syndrome. And the thing is, you can get around not invented here syndrome. You can say, oh my God, our trains are much slower and less reliable than they are elsewhere. We need to learn even from places that we are not used to learning from. This does happen. It requires a very large gap in capabilities to actually take this kind of cultural embarrassment of having to learn from people that you're not used to learning from. But that gap exists. The trains over the Northeast Corridor are genuinely much worse than the top line corridors are in every developed country of the same scale as the northeastern United States. I mean, sure, the trains are faster than in Australia. So the entire population of Australia is half that of the Northeast Corridor. And Australia is very spread out. The northeastern US is rather dense.
- [42:14]
Santi Ruiz
Yeah, it's a good place for high speed rail. Let me ask a slightly different version of this question, which is how do you get buy in from all the states that would be required for a proposal like this? And I'll flag a couple of the examples of benefits and costs are distributed pretty unequally, just by definition, right? So there'd be a lot of construction in New Jersey, but the benefits to the state are pretty modest. Whereas you have the reverse dynamic for Pennsylvania that Philadelphia gains a ton from these increases in speed, but there's relatively little construction you have to deal with. Or for Virginians, one of the biggest downsides or locally concentrated negatives of this proposal is from now on, if you want to go from Richmond directly to New York, you have to switch a train in dc. How do you sell that to Virginia politicians and to voters? How do you even think about the political economy of getting all the governors on board for their model? Because I think the classical political economy idea about this stuff is you propose this and then every state says, okay, sure, but to get us on board, we need our pound of flesh. And then you end up with something close to what you get in practice, which is the Connect 2035 or 2037 models that are 100 billion-plus.
- [43:19]
Alon Levy
So the answer is that every state actually gets a pound of flesh on this. For example, you mentioned New Jersey. In New Jersey, yeah, there's extensive construction, but all of it is driven by the state. Everything that I'm proposing to be included in Jersey, it's three projects. It's Portal Bridge south, it is Hunter Flyover, which is a flyover between the Raritan Valley Line and the Northeast Corridor, and something called the Mid Lane Loop, which is kind of an elevated project that lets local trains in central Jersey turn without mucking the track for the express trains. All of these are on the New Jersey Transit wish list. I believe it applied for bipartisan infrastructure law funding for all three of them and it got funding for other things that it applied for. This is driven by the state. And the reason it's driven by the state is that all of these are of great utility to the commuter trains in the state. The utility to intercity trains is there, but it's actually kind of limited. It's just that the commuter trains are part of the system, so they get this funding. So all of this funding in New Jersey has extensive benefits for enabling this higher throughput for commuter trains, making them faster, making them more reliable, having high platforms. So there will be also accessibility. And that repeats itself in every state between Massachusetts and the District of Columbia. To the extent that D.C. has a state, they all get things for intercity and commuter trains in Virginia. You're right that there is a more serious trade off. The trade off being that unless they wire the line down to Richmond, they're going to change trains. Now, can they wire the line to Richmond? Absolutely. I don't want to tell you how much it's going to cost. They can give you an order of magnitude estimate, which is D.C. to Richmond is 125 miles, right?
- [45:00]
Santi Ruiz
Seems right, yeah.
- [45:01]
Alon Levy
Yeah, I think it's about that. And that means that the line should be 600, 700, 800 million to electrify. And if Virginia says sure, but we want money for electrifying and some upgrades, sure. As far as I'm concerned, that needs to happen anyway for better service to Virginia. They actually are spending the big ticket money on that already because the rail bridge and the long bridge between the District and Virginia is currently owned by csx and it's freight primary. It's an important freight carver. So there's no freight rail going into Union Station. There's a bypass just south of it, but the bridge itself is shared. They're building an additional bridge for passenger trains. So the high spending often is already happening and it's just doing a little bit extra, in this case electrification. So In Virginia, it is an imposition if people have to change trains at Washington Union Station. On the other hand, the trains they're going to change to are significantly faster to New York. So you have a change and that is a real trade off. So obviously different writers have different opinions on this. Probably if you're a very regular, if you're a very seasoned rider, and if you're by yourself and without heavy luggage, you probably don't care. If you're traveling as a family, you care more. If you're an irregular rider, it's confusing for you. Maybe you care more. But on average, what I read is that the impact is on the order of 15 to 20 minutes. So people will take a transfer to speed up the trip by 15 to 20 minutes. Again, it depends on the riders or ones who will take it to speed up by two minutes or ones who will want hours. But if it's 20 minutes on average, I'm promising an increase in speed that turns the Northeast corridor into an hour. 56 New York D.C. and today it's just under three hours if you're paying the seller premium and it is three and a half hours if you're not. So that is where you have a trade off. But even there, the speed benefit is such that it doesn't quite bind. Again, it comes close to binding. And the closer you get to Frontier, the more you're getting to binding trade offs where you do need to buy people off.
- [47:08]
Santi Ruiz
Let me change gears here. No pun intended. I want to get your thoughts on a previous conversation we've had in this series with former head of Massachusetts Transit, Stephanie Pollock. She came in on the Green Line extension project in Boston, which you've written quite a lot about. You've been pretty critical of her in the past. I have a quote from you here from 2020 describing some problems with Massachusetts transit as quote nothing that firing Stephanie Pollack and hiring an actually competent person wouldn't fix. I take it you had some criticisms of her comments on Statecraft.
- [47:37]
Alon Levy
Yes, I have. So what she said was that you can't really compare different projects because the costs in the United States are inclusive of things that are not included elsewhere. And the problem is that she's wrong. So for example, our database for construction costs, we very dutifully included only infrastructure, not train sets. We have a different database for train sets. There is an American premium in the acquisition cost of these train sets. It's much smaller than the premium for infrastructure. And so everything is without rolling stock. There are a few lines where we couldn't net out rolling stock, but these are usually in the cheapest countries in the world. When I look at some Spanish construction calls or some Italian ones or some Portuguese ones, sometimes you just can't get the numbers without rolling stock. So, yes, these lines look a little more expensive. These are also the cheapest lines in the world that we're telling Americans to be more like that. Likewise with other kinds of infrastructure, things like heavy maintenance, facilities, planning, all of that is included elsewhere.
- [48:40]
Santi Ruiz
Sorry, why does it matter?
- [48:42]
Alon Levy
It matters because two things. The first is that the cost in the United actually are comparable.
- [48:48]
Santi Ruiz
You can measure apples to apples exactly.
- [48:50]
Alon Levy
The delivery is practically the same. It's just that the soft costs in the United States are excessive. And I don't mean that everything is excessive, although it is. What I mean is that you take the hard costs and in Italy or in France, the planning costs on top are between 5 and 10%. And in the US that might be 20% on top. And then you look at all the other soft costs and there are going to be maybe an additional 50% in the US so 20 plus other things, it's 50%. And then maybe in Italy it's going to be 20%, 25%. There's just much less white collar job productivity in the United States for public sector, partially the private sector is the American white collar office sector. What people deride as email jobs. They're very efficient. So on top of being kind of inefficient in the public sector in the United States, you also need to pay salaries to be competitive with an extremely efficient private sector. But if you look at, for example, how many white collar supervisors are required per blue collar worker for construction, it's consistently much more than that for, let's say, the Green Line extension, than it is for private sector construction projects in the Northeast, for example. So there's this myth of apples to apples that the United States can't be this bad. But it is. It's just coming from people who don't know how things work in any country that doesn't speak English natively. So they come up with these really weird ideas about what is and is not included in a Swedish construction project or a German one or a Turkish one or an Italian one.
- [50:26]
Santi Ruiz
Two questions for you. One is we talked to Stephanie Pollock about her work basically coming in on the Green Line extension for listeners, what goes northwest out of Boston into the surrounding suburbs, some pretty dense suburbs. And Paula came in when the project was well over budget and tried to bring the budget back down. We talked about that on stake, I found it a really enjoyable episode. Elan, how do you rate her performance on the Green Line extension?
- [50:51]
Alon Levy
First of all, her performance started with the fact that she sued people into building it, which means that the previous cost overruns are to an extent her fault as well, because she created a situation in which the project was considered a must build. When you make something a must build, you're giving everyone an excuse to lade it with a bunch of extras with their pound of flesh. A pound of flesh is almost never actually necessary. It's just people want to say yes to things and they don't watch the budget. Then not watching the budget is partly her legacy as well. I don't want to say that she's the main villain of the story. She isn't. She is one of many. But the reduction in costs under Governor Baker, it was real, but it was to levels even after the reduction in cost. That would have been rather expensive. Not extremely expensive, but rather expensive subway project. It was not a subway project. It was a light rail project that was in an existing trench that needed to widen a little bit. I can't tell you how much it should have been because usually people don't build exactly that. Because usually where I live, if you had that kind of infrastructure, you would build a commuter rail line. But there's been so much siloing of commuter rail, as I mentioned, away from every other mode of public transportation. This is why I'm quoting the former MPTA general manager, Paola, who said, commuter rail is not mass transit. The problem is that all of these decisions were bad. And again, I'm not going to make Pollock the main villain. What annoys me about her, I guess, is that she didn't learn. She didn't say, yeah, we did a bad job in here, or the lessons learned not to do this again. Instead, she's defending that and defending worse decisions. Things like trying to privatize decision making further, doing these special purpose delivery vehicles that they did under Governor Baker, where they hired experienced project managers and got things under control. And then instead of keeping them in house for further projects that they knew they said was going to build, they just let it all dissipate. In Paris, by the way, for the 200 kilometer, 125 mile expansion of the metro into the suburbs as a bunch of driverless lines in the inner suburbs of Paris. Some extensions of existing lines, some new orbital lines, nearly all of it is underground. It's rather expensive because 125 miles of subway are going to be expensive. Anywhere. And their costs are running over because they're used to building a few miles at a time, not hundreds of miles at a time in Paris. So they established this special purpose delivery vehicle, but instead of letting it go to waste, as they did in Boston, as they did in London for Crossrail, they're making it a permanent institution. Because once you're hiring all these experienced project managers and engineers, planners and architects and transport economists, what are you going to do? Let them all go so that each of them is going to pick the few pieces that they learned and take them to a consultancy? No, you're going to keep them because you're going to build more. In the case of France, they're going to use it as a public sector consultant to help build similar things in all of the other cities of France. And in Boston, they could have done this. They are building more things and they're planning to build more things. They just didn't make this long term commitment. And this is one of the things I'm complaining about, is to connect with high speed rail a little bit. It's something that requires this top down coordination where someone needs to say no. And again, the FRA is capable of saying no. It's not captive to any of the states that want something else. Because they can say, yeah, we can't give you this because we can give you that, and it's going to give you these benefits. But they're just not used to doing it. They're used to the idea that the federal government just staples other people's proposals together. Sure.
- [54:32]
Santi Ruiz
So I guess just to wrap that point, your contention is the power to force the coordination that you want to see already exists.
- [54:40]
Alon Levy
Exactly. Just as the boring things about the exact speed of a train on occurs. The regulations already permit that in the United States due to these reforms that were done in the 2010s. It's just the power is not being used because of this not invented here syndrome, where the railroaders are used to having been taught certain standards that they're not going to question. Because what they're going to do, Go look at the German engineering standards. First of all, they're written in German. And second, there's this oh, America is different, we can't mentality. But this also includes the understanding of how to do this coordination, which means you start with a timetable. I've been told by people who are more competent than Stephanie Pollack that a plan like what I'm doing never goes down to the level of detail of timetables to the mint. And they had to send links to the long term plans of Deutsche Bahn it's all called the Deutschlandtagt where they're going to invest in a medium speed Germany wide system where all the regional trains are going to connect better. And the thing is these plans that were supposed to open in 2013 are going to open much later at this point and in Switzerland when they go to on these projects they include the timetable to the minute because that's actually important. It's an integration of top level planning to say are we spending 500 million Swiss francs and in Switzerland and in small cities that's a lot of money and the exact timetable to the minute these actually have to connect with each other whereas in the US you come up with the big plan first and you expect the assistants to the assistants to fill out the details.
- [56:14]
Santi Ruiz
Great. Well Alon, this has been a real pleasure. Thank you for making so much time to get in the weeds with me.
- [56:20]
Alon Levy
Yeah, thank you, thank you for inviting me.