The United States Learned Little from Obama-Era Rail Investment
A few days ago, the US Department of Transportation (USDOT) announced Bipartisan Infrastructure Law grants for intercity rail that are not part of the Northeast Corridor program. The total amount disbursed so far is $8.2 billion; more will come, but the slate of projects funded fills me with pessimism about the future of American intercity rail. The total amount of money at stake is a multiple of what the Obama-era stimulus offered, which included $8 billion for intercity rail. The current program has money to move things, but is repeating the mistakes of the Obama era, even as Secretary of Transportation Pete Buttigieg clearly wants to make a difference. I expect the money to, in 10 years, be barely visible as intercity rail improvement – just enough that aggrieved defenders will point to some half-built line or to a line where the program reduced trip times by 15 minutes for billions of dollars, but not enough to make a difference to intercity rail demand.
What happened in the Obama era
The American Recovery and Reinvestment Act (ARRA), better known as the Obama stimulus, included $8 billion for what was branded as high-speed rail. Obama and Secretary of Transportation Ray LaHood spoke favorably of European and East Asian high-speed rail at the time. And yet, the impetus to spread the money across multiple states’ programs meant that the sum was, by spending, around half for legacy rail projects euphemistically branded as higher-speed rail, a term that denotes “faster than the Amtrak average.” Ohio, Wisconsin, and Illinois happily applied that term to slow lines. The other half went to the Florida and California programs, which were genuinely high-speed rail. In Florida, the money was enough to build the first phase from Orlando to Tampa, together with a small state contribution.
Infamously, Governor Rick Scott rejected the money after he was elected in the 2010 midterms, and so did Governors Scott Walker (R-WI) and John Kasich (R-OH). The money was redistributed to states that wanted it, of which the largest sums went to California and Illinois. And yet, what California got was a fraction of the $10 billion that the High-Speed Rail Authority had been hoping for when it went to ballot in 2008; in turn, the cost overruns that were announced in 2011 meant that even the original hoped-for sum could not build a usable segment. The line has languished since, to the point that Governor Gavin Newsom said “let’s be real” regarding the prospects of finishing the line. Planning is continuing, and the mostly funded, under-construction segment connecting Bakersfield, Fresno, and Merced is slated to open 2030-33 (in 2008 the promise was Los Angeles-San Francisco by 2020), but this is a fraction of what was promised by cost or utility; Newsom even defended the Bakersfield-Merced segment on the merits, saying that connecting three small, decentralized metro areas to one another with no onward service to Los Angeles or San Francisco would provide good value and taking umbrage at the notion that it was a “train to nowhere.”
In Illinois, the money went toward improving the Chicago-St. Louis line. However, Union Pacific owns the tracks and demanded, as a precondition of allowing faster trains, that the money be spent on increasing its own capacity, leading to double-tracking on a line that only run five trains a day in each direction; service opened earlier this year, cutting trip times from 5:20-5:35 in 2010 to 4:46-5:03 now, at a cost of $2 billion. This is a 457 km line; the cost per kilometer was not much less than that of the greenfield commuter line to Lahti, which has an hourly commuter train averaging 96 km/h from Helsinki and a sometimes hourly, sometimes bihourly intercity train averaging 120. In effect, UP extracted so much surplus that a small improvement to an existing line cost almost as much as a greenfield medium-speed line.
Lessons not learned
The failure of the ARRA to lead to any noticeable improvement in rail service can be attributed to a number of factors:
- The money was spread thinly to avoid favoring just one state, which was perceived as politically unacceptable (somehow, spending money on a flashy project with no results to show for it was perceived as politically acceptable).
- The federal government could only spend the money on projects that the states planned and asked for – there was no independent federal planning.
- There was inattention to best practices in legacy rail planning, such as clockface timetabling, higher cant deficiency (allowed by FRA regulation since 2010), etc.; while the high-speed rail program aimed to imitate European and East Asian examples, the legacy program had little interest in doing so, even though successful legacy rail improvements in such countries as the UK, Germany, Sweden, Switzerland, Austria, and the Netherlands were available already.
- A dual mandate of both jobs and infrastructure, so that high costs were a positive to an interest group that the federal government and the states announced they wanted to support.
- California specifically was a series of unforced errors, including a politicized High-Speed Rail Authority board representing parochial rather than statewide interests, disinterest in developing any state capacity to plan things (the expression “state capacity” wouldn’t even enter common American political discourse until the late 2010s), early commitment, and, once the combo of cost overruns and insufficient money on hand meant the project had no hope of finishing in the political lifetime of anyone important, disinterest in expediting things.
I have not seen any indication that Buttigieg and his staff learned any of these lessons, and I have seen some indication that they have not.
For one, the dual mandate problem is if anything getting worse, with constant invocations of job creation even as unemployment is below 4% where in 2010 it was 10%, and with growing protectionism; the US has practically no internal market for modern rolling stock and the recent spate of protectionism is leading to surging costs, where until recently there was no American rolling stock cost premium. This is not an intercity rail problem but an infrastructure problem in general in the US, and every time a politician says “this will create jobs,” a surplus-extracting actor gets their wings.
Then, even though the NGO space has increasingly been figuring out some best practices for regional trains, there is still no integration of these practices into infrastructure planning. The allergy to electrification remains, and mainline rail agency officials keep making things up about rest-of-world practices and getting rewarded for it with funds. Despite wide recognition of the extent of surplus extraction by the Class I freight carriers, there is no attempt to steer funding toward lines that are already owned by passenger rail-focused public-sector carriers, like the Los Angeles-San Diego line, much of the Chicago-Detroit line, and the New York-Albany line.
The lack of independent federal planning is if anything getting worse, relative to circumstances. In the Obama era, the Northeast Corridor was put aside. Today, it is the centerpiece of the investment program; I’ve been told that Biden asks about it at briefings about transportation investment and views it as his personal legacy. Well, it could be, but that would require toothy federal planning, and this doesn’t really exist – instead, the investment program is a staple job of parochial interests. Based on this, I doubt that there’s been any progress in federal planning for intercity rail outside the Northeast.
And finally, the money is still being spread too thinly. California is getting $3.1 billion, which is close to but not quite enough to complete Bakersfield-Merced, whose cost is in year-of-expenditure dollars at this point $34 billion for a 275 km system in the easiest geography it could possibly have. Another $3 billion is slated to go to Brightline West, a private scheme to run high-speed trains from Rancho Cucamonga in exurban Los Angeles, about 65 km from city center, to a greenfield site 4 km south of the Las Vegas Strip; the overall cost of the line is projected at $12 billion over a distance of 350 km. It’s likely that this split is worse than either giving all $6 billion to California or giving all of it to Brightline West. But as I am going to point out in the following section, it’s worse than giving the money to places that are not the Western United States.
The frustrating thing is that, just as I am told that Biden deeply cares about the Northeast Corridor, Buttigieg has been quoted as saying that he cares about developing at least one high-speed rail line, as a legacy that he can point to and say “I did that.” Buttigieg is a papabile for the 2028 presidential primary, and is young enough he can delay running for many cycles if he feels 2028 is not the right time, to the point that “I built that” will strengthen his political prospects even if he has to wait until opening in the 2030s. And yet, the money committed will not build high-speed rail. It might build a demonstration segment in California, but a Bakersfield-Fresno line and even a Bakersfield-Merced one with additional funds would scream “white elephant” to the general public.
Is it salvagable?
Yes.
There are, as I understand it, $21.8 billion in uncommitted funds.
What the $21.8 billion is required to achieve is a) a complete high-speed line, b) not touching the Northeast Corridor (which is funded separately and also poorly), c) connecting cities of sufficient size that passenger ridership would make people say “this is a worthy government investment” rather than “this is a bridge to nowhere on steroids.” Even a complete Los Angeles-Las Vegas line is not guaranteed to be it, and Brightline West is saving money by dumping passengers tens of kilometers along congested roads from Downtown Los Angeles.
Given adequate cost control, Chicago-Detroit/Cleveland is viable. It’s around 370 km Chicago-Toledo, 100 km Toledo-Detroit, 180 km Toledo-Cleveland, depending on alignments chosen; $21.8 billion can build it at the same cost projected for Brightline West, in easier topography. If money is almost but not quite enough, then either Cleveland or Detroit can be dropped, which would make the system substantially less valuable but still create some demand for completing the system (Michigan could fund Toledo-Detroit with state money, for example).
But this means that all or nearly all of the remaining funds need to go into that one basket, and Buttigieg needs to gamble that it works. This requires federal coordination – none of the four states on the line has the ability to plan it by itself, and two of them, Indiana and Ohio, are actively hostile. It’s politically fine as a geographic split as it is – that part of the Midwest is sacralized in American political discourse due to its industrial history, which history has also supplied it with large cities that could fill trains to Chicago and even to one another; politicians can more safely call Los Angeles “not real America” than they can Detroit and Cleveland.
But so far, the way the Northeast Corridor money and the recently-announced $8.2 billion for non-Northeast Corridor service have been spent fills me with confidence that this will not be done. The program is salvageable, but I don’t think it will be salvaged. There’s just no interest in having the federal government do this by itself as far as I can see, and the state programs are either horrifically expensive (California) or too compromised (Midwest, Southeast, Pacific Northwest).
So what I expect will happen is more spreading of the money to lines averaging 100 km/h or less, plus maybe some incomplete grants to marginal high-speed lines (Atlanta-Charlotte is a contender, but would get little traffic until it connects to the Northeast Corridor and would cost nearly the entire remaining pot). Every government source will insist that this is high-speed rail. Some parts will be built and end up failing to achieve much, like Chicago-St. Louis. Every person who is not already bought in will learn that the government is inefficient and it’s better to cut taxes instead, as is already done in Massachusetts. Americans will keep making excuses for why it’s just not possible to have what European and a growing list of Asian countries have, or perhaps why there’s no point in it since if it were good it would have been invented by the American private sector.
Pedestrian Observations 
Why is Chicago to St. Louis still so slow after all that investment? The average speed is under 100km/hr, despite top speeds of 180km/hr.
It is alternating double-single track, plus work is need on the northern end into Chicago I believe, to add capacity and cut travel times.
You can check out the work on Google Maps and Streetview on the Chicago-St Louis line, look around at the upgrade level crossings and stations.
Because it’s owned by (thus the physical improvements to it are controlled by) and operated daily by UP rather than Amtrak. Naturally, UP is interested in making UP’s trains flow more smoothly and efficiently. To them, the passenger trains are somewhere between “afterthought” and “annoyance”.
Plausibly, the most valuable part of the investment would have been simply buying the line from UP, just so that the dispatchers could be given a new uniform and the instruction to prioritize passenger trains. And then improve the physical plant from there, when you have it under your control, rather than wrestling UP into doing on your behalf the improvements you want (as opposed to using your money to do the improvements they want).
To be fair the vast majority of “high speed lines” don’t get an average speed anywhere close to 186mph. The only people who do are the Chinese who get about 180mph by my calculations. Then the French and Spanish manage around 150mph, Tokyo-Osaka does 130mph, and everyone else is slower than that. Frankfurt-Cologne is only approximately 115mph as is the fastest Milan-Rome service as far as I can tell.
The fastest Diesel service is probably the 2307 Chiltern Mainline service from Marylebone to Birmingham Moor Street which has a top speed of 100mph and averages 62mph over the whole trip including stations and with little freight contention. The Americans averaging a similar speed with freight and a top speed of only 110mph is really not bad.
Actually forgot the Reading-Exeter line is still mostly diesel and has a top speed of 110mph. Then the trains with two stops manage that 137 mile trip at an average of 79mph. Of course this is with full priority over freight and Brunel quality engineering meaning it is probably 110mph pretty much throughout.
It is alternating double-single track, plus work is need on the northern end into Chicago I believe, to add capacity and cut travel times.
The line opened already but they’re not running more trains, and the overall travel time cut is small relative to the cost.
Worth mentioning that Michigan is a key swing state, Ohio is a past and maybe future swing state, and Pennsylvania (the next stop after Cleveland) a key swing state. So the line you suggest is pretty ideal in terms of electoral geography, especially if announced in the next year.
A downside is that all these states are depopulating.
An alternative line which serves a growing swing state might by LA-Phoenix. This would also get Brightline West all the way to LA, and allow for LV-Phoenix trips. Ridership might be higher than Chicago-Detroit-Cleveland. But it might be seen as favoring one part of the country at the expense of the rest.
Phoenix-Tucson would be good, it seems that the combination of interstate and railroad right-of-way could allow for a Brightline West type of infrastucture from city center-to-city center, plus the Phoenix airport.
Phoenix Tucson got half a million dollars for a study: https://www.axios.com/local/phoenix/2023/12/11/passenger-rail-restoration-federal-grant
In round numbers Tuscon’s Combined Statistical Area is the same as Albany’s. There’s a whole lot of nothing between the western suburbs of Phoenix and the eastern suburbs of Palm Springs. 230 miles/370 km of it between Goodyear AZ and Indio CA. For a metro area that is a bit bigger than Utica’s. It makes a nice crayon line but there are places in the Northeast and Midwest that make a lot more sense.
Tuscon is the same size as Albany. They have an airport just like Albany does. Why would they need to go to Sky Harbor/Phoenix?
Yeah, Michigan and Ohio are depopulating, but at a very slow rate and from a high basis. The current population of Detroit, Cleveland, and Chicago is sufficiently high for this line right now – no need for any assumptions on future growth, unlike with Vegas.
Thr population of the city of Cleveland itself is down, but the overall population of the metro region hasn’t changed. The population of downtown Cleveland has grown and is limited by the available buildings and space for new housing.
You don’t build intercity passenger rail only in areas where the population is growing. You build it as an econimic development tool. There is a lot of travel density in the region too. Cities like Detroit and Cleveland are good locations for investing in passenger rail.
Alon: Sorry, I hit send too soon and can’t edit my comment. I meant to start by saying I agree with your statement on population.
“The current population of Detroit, Cleveland, and Chicago is sufficiently high for this line right now – no need for any assumptions on future growth, unlike with Vegas.”
2020 Census MSA populations:
Chicago 9.45M
Detroit 4.39M
Cleveland 2.19M
LA 13.2M
Riverside/S Bernadino 4.6M
Las Vegas 2.27M
I’m failing to see how LA-Vegas is lacking population compared to Chicago-Detroit/Cleveland.
Plus, Las Vegas obviously draws more travelers than its size alone would dictate. Air traffic from LA-LV is ~7,100 daily, with San Diego-LV adding ~1,800 more. Chicago to Cleveland and Detroit combine for just ~2,000 daily air passengers.
“Despite wide recognition of the extent of surplus extraction by the Class I freight carriers, there is no attempt to steer funding toward lines that are already owned by passenger rail-focused public-sector carriers, like the Los Angeles-San Diego line, much of the Chicago-Detroit line, and the New York-Albany line.”
Actually the “S-Line” south of Richmond and a lot of the share freight and passenger within North Carolina is state owned, and the Richmond-DC project is going ahead with both federal and state funding, Virginia as both state capacity and dedicated funding for passenger rail.
Yeah, there’s that, but what I’m not seeing is any concerted focus on these lines to the exclusion of lines where the Class Is extract surplus to the point of demanding a second track for the not extensive traffic of Worcester-Springfield.
I think you have called out the corridor that will be the biggest home run for true HSR: DC-Richmond.
Getting real high speed through suburban DC is probably a fool’s errand, but electrifying the RF&P for the first 50km through Quantico, then running 90km of greenfield HSR from Quantico through 5km short of Richmond Main Street Station, and then a slower ride from there – could probably yield a trip time under 1 hour from DC to Richmond (down from roughly 2h 30m today).
An average speed of about 130mph is not exactly world-leading but it’s good enough, and given the run-though potential onto the NEC, would be a runaway success.
130mph average speed is top 5 worldwide as far as I am aware. It is very very decent.
I think my measurements and calculations were off, it would be:
50km of conventional speeds DC to Quantico
120km of high speed Quantico to Richmond
5km of conventional speed on approach to Richmond Main Street
A WAS-RVM trip time of 1 hour requires an average speed of 175km/h or 110mph. Which is more like what I was thinking. Not world-leading but good enough.
Improving WAS-anything south is kind of out of the question until the Long Bridge project is complete, since there’s too much conflict between the freight and passenger traffic until that finishes. At least the Long Bridge project seems to finally be nearing construction, even if they’re still refusing to contemplate adding electrification to the bridge when it’s built (although supposedly everything is being designed so as not to preclude electrification).
For something that could be developed rather quickly, Richmond-Raleigh is a better segment, even if hooking up DC-Richmond is necessary to make the route viable.
I’d be interested in seeing some serious crayoning on DC/VA regional rail infrastructure a la Transit Matters is doing for Boston–is there anyone who’s up to that task?
The thing is, most of this money actually did go to only a few projects. There was a small pool of money that got distributed in many $500,000 chunks to do “studies”, presumably so that they could make a nice map with lots of lines. But the vast majority of the $8.2b are going to eight or so projects.
Do we need more studies? Amtrak has done studies in the NEC since the 1970s, they should be sitting on a shelf in some backroom. Dust one off and start building it.
I’m starting to look a not executing on a study is a sign of corruption: since the study never gets turned into concrete things the consultant’s incompetence is hidden and you can keep sending them money. (unless the study actually says all alternatives are bad – do those ever happen?)
I don’t see the studies as all that serious right now. I see it as a political game. Which I also see as fair play with a relatively small amount of money. Whatever I can say about Biden’s administration, it’s quite clear that if he loses next year things would be worse for intercity rail. If this chunk of money prevents that and does nothing else, then it has provided its value.
75% of it goes to two projects, but even that is insufficient for either, to the point that it would have been better to spend the entire $6.1 billion on one or the other.
But neither project *asked* for the whole $6.1 billion.
People ask for what they think they’ll get.
Did both projects have information that they wouldn’t receive more than $3b ahead of time? If not, I’d expect them to have asked for what they needed to complete their project.
California HSR doesn’t have the money to complete Bakersfield-Merced by 2033 even with the $3 billion; for 15 years the HSR Authority has been using magic asterisks hoping for private funding and what not.
While there is some truth in what you say, you miss the bigger picture here. The reason the U.S. hasn’t advanced passenger rail to the standard of other developed nations (and an increasing number of developing ones) because we’ve never committed the level of funding necessary to make it happen. When we decided to build our two-lane highways, we spent what was necessary and established gasoline taxes to pay for them. Ditto our interstates. When we committed to building airports and an air traffic control system in the 1950s, we spent what was necessary and established an aviation trust fund. We’ve never done that for passenger rail, and until we do, we’ll continue to limp along like this.
As for California, they are the first real effort at true high-speed rail. Cost over-runs aren’t the issue. Even Japan’s Shinkansen cost double the original estimates. I-69 is decades behind schedule and way over budget. As California is the first in the nation, their is no knowledge pool in the US for building HSR. They are building it from scratch. This is not an excuse, it’s an explanation. Also as far as costs go, it’s more expensive to build infrastructure in the U.S. than it is in Europe and other countries. Their are various solvable reasons for this, but until we address those reasons, this stuff is always going to be expensive.
What do you mean, you’ve never committed to the level of funding necessary to make it happen? The LGV Sud-Est was built for $3.5 billion in 2023 prices, and was on-budget.
If a line was being built across the Great Plains that would be an achievable cost as the population density along the route is similar.
In the North East United States probably hitting double HS1 is probably all you can aim for – I.e. $180m/km. I say double to account for wages in America being higher and it being unrealistic to assume additional automation that doesn’t make sense in developed Europe and Asia is unrealistic – plus if you try it you risk it blowing up costs like the Great Western Mainline electrification did.
There’s no point in implying you can hit price points that the legacy unhelpful people within these organisations can just blow up by pointing out that the population density of rural France near the LGV Sud Est is functionally zero so you don’t need any mitigations or stations.
The point about population density is pretty crappy, sorry. Germany has higher population density than France and also higher costs, but they’re mostly reducible to more tunneling, sometimes because of topography, sometimes because of freight-friendly grades that are then not used for actual freight. There is a rough equivalent of that in the UK with the extensive tunneling in the Chilterns, but even relative to the extent of tunneling, HS2 costs are very high – and California HSR costs manage to be comparably high with no such tunneling (the Central Valley is not dense).
Re wages: American wages are high for white-collar workers, but in the working class, even among well-paid skilled craft workers, there’s no big American premium. New York sandhogs, who are practically labor aristocracy, earn about as much as Swedish miners, who are an extensively immigrant workforce; very well-paid Norwegian electricians manage to electrify trains cheaply just fine.
Re mitigations: in France it’s more or less legal for white people to riot; farmers demand and get mitigations, like overcompensation for land, and state-supervised land swap deals.
Population density absolutely matters.
For a start if Labour had promised a HS2 station for Stoke, Leamington and Wycombe/Amersham they’d probably have won the general election in 2010/2015/2017. And each of those stations would have got at least 2 million rides a year so would justify a decent service.
Then there’s the much greater need for mitigations. Much harder to justify mitigations for villages of less than a thousand where the electoral impact is zero compared to denser places like England and the North East USA.
If blue collar wages are the similar to Europe then perhaps $100m/km is a reasonable target.
Why? When it isn’t tunneling, Germany builds for a third of that cost.
Germany uses legacy approaches into its cities. I am assuming America won’t be able to. If it can it will obviously be cheaper.
Plus I am pretty sure Germany is less dense along its high speed lines than the North Eastern United States.
The US absolutely can use legacy approaches.
Population density should not matter if you have compulsory orders etc. England has high population density but highly nucleated despite its geography. Certainly easier than building through Gunma, Ishikawa or Shizuoka with their inanely fine land subdivisions with sprawl on top and weak coercive land assembly powers.
If Korea, Japan and China can manage to build at less than 1/3 of HS2’s cost per/km despite vastly more difficult geography.
(Yes Alon I know JR West screwed up on Hokuriku Shinkansen extension to Tsuruga, have you figured out why?).
Also the idea that extra stations would have changed the outcome of those elections is insane. Transport is not even Tier 2 issue outside central London, and a Tier 1 issue nowhere. Transport is not even a major dividing line between the major parties (it may become so). Those elections were fought on national identity, immigration, Brexit, “economic credibility” and judgements on the parties leadership.
This always drives me mad about UK transport people, they think everybody cares as much as they do. They don’t even understand the relationship to housing crisis/planning system. Reactionary whiners.
Recent Shinkansen lines aren’t too visible to me beyond topline numbers. I think the media is just checked out, since at the end of the day, a Shinkansen extension to Kanazawa isn’t the sort of project that gets followed very closely, especially not in English-speaking media sources.
I could dig up the Prefectural and Ministry of Land documentation. I noticed the lack of Kyushu stuff in the data set too.
Do you know the Kyushu topline costs?
@Borners, I think good public transport is important if you want to deliver other government priorities. So yeah it’s not a tier 1 issue, but most obviously to support new housing developments and to spread opportunity it’s important behind the scenes.
And HS2 stations might well have helped win swing votes, opposition to HS2 did come up as an issue on the doorstep along the route. Stations counters that. And a station for Stoke on Trent in particular would be levelling up.
Personally not quite. Here:s some basic sleuthing.
This from Kagoshima Prefecture details the original budget of the Hakata-Kagoshima Kyushu Shinkansen, then divides it into the two sections opened in 2004 and 2011, with budget increases included.
https://www.pref.kagoshima.jp/ac08/infra/kotu/shinkansen/gaiyou/gaiyo_routeplan.html
]Tunnelling share figures
https://www.pref.kagoshima.jp/ac08/infra/kotu/shinkansen/gaiyou/gaiyo_routeplan.html
More detail on those tunnelling etc on the first page of this pdf.
Click to access shinkansen.pdf
Nishi-Kyushu Shinkansen
Here is the original budget plus the additions to cost overruns.
https://www.mlit.go.jp/report/press/tetsudo09_hh_000064.html
The basic details of construction are here with tunnelling/viaduct numbers.
https://www.jrtt.go.jp/project/kyushu.html
The chief industry of Stoke is resentful pensioners. I don’t think they would care, especially if they thought (as much of the public thought) that HS2 came at the expense of the NHS etc. If you actually read what people say they want in the Levelling up places they want pensions, NHS, fewer foreigners and somebody to spruce their local government buildings. Interest in economic transformation is very much limited to the major city centres where only a minority of Outer England’s voters live. See also the Celtic fringe despite have 20% more infrastructure money to spend have managed….some electrification and a single tram line. Scotland has had Danish budgets and Northern English service.
If you’re looking for transit oriented swing voters that’s in Outer London and the Home Counties. That’s also where the New Towns are likely to be because why would you build another Warrington/Peterlee/Skemersdale/Telford when existing Northern settlements are pathetically low density and frankly empty.
That being said just getting railways outside London back to 2019 services would take a lot of effort. HS2 extensions will have to wait until the still intact section opens and everybody then wants one. Till then electrification, passing loops and operations reform are way more important. Rail advocates/industry need to take on that the failure of British railway is their fault almost as much as the Treasury/Tories. Until that happens they don’t deserve new toys.
Speaking of Japanese construction costs. Alon you might be interested in an old book chapter which has some detail on Japanese subway construction issues in English. Its in Japanese Urban Enviroment editor Gideon Golany which a selection of Japanese urban infrastructure engineers from 1995.
On of the thing it highlights for high costs is that Central Tokyo is soft alluvial soils. That doesn’t excuse other failures on cost control of Japanese subways I know but Tokyo mimicry is an issue in smaller cities (Fukuoka, Sendai).
I stated it pretty clearly. What don’t you understand? Instead of a permanent, sizeable and dedicated trust fund to fund the mode year after year like highways and aviation have had since the mid-20th century, passenger rail gets money when a friendly president makes it a priority. Then years go by before anything significant happens again. Add in the lack of a healthy planning apparatus, and you have a recipe for a lack of progress.
So France did a HSR line for $3.5 billion. So what? That doesn’t change the fact that the U.S. has a flawed system for planning and building transportation projects that ends up spending more and getting less in return than other modern nations. Check out this article from Vox:
Why does it cost so much to build things in America?
https://www.vox.com/22534714/rail-roads-infrastructure-costs-america
Did you just cite an article at Alon, that quotes Alon?
Would LGV Sud-Est cost $3.5B today, at recent French construction prices? I know one has to use inflation adjustment to compare projects over time, but I’m not sure it is always accurate.
Also, although I wish to make now excuses for any of CAHSR’s mistakes, to be fair LGV Sud-Est is only about 65% as long as SF-LA (before any further route to Anaheim, SD, San Bernadino, etc.) and didn’t have to cross anything like the Coast Range or the Tejon/Tehachapi passes. There is not a single meter of tunnel along the entire LGV Sud-Est route, that simply cannot happen for HSR in California, terrain won’t allow it.
San Jose to LA is 340 miles or 550km. They have $23.4 billion to spend or $42m/km. Well assuming you can use legacy approaches into LA as well as Caltrain into San Francisco it’s probably doable. Electrification of Caltrain probably should cost $250m or something.
LGV Sud-Est used legacy approaches into Paris and Lyon as well. In Lyon they also took the step of using legacy track to a freight station (Part-Dieu) and then developing a new CBD around it instead of getting the TGV into the old city center at Lyon-Perrache. The cost of building a new city center (not charged as a rail cost) versus bringing the line to the old city center (which would have been a rail cost) is of course another thing that makes LGV Sud-Est cheaper.
SJ-LA via where the people are (Fresno, Bakersfield) is 590km.
You can’t just use Caltrain to reach SF, because travelling 80km at Caltrain speeds would take so long it would affect overall SF-LA travel time. Unlike the NEC, where New York is in the middle, in California the important cities are at the ends of the line so it is a place where maximizing absolute speed becomes important to keep travel times competitive. Unlike the “only as fast as necessary” mantra where going a little slower to make takt at intermediate locations will provide better ridership, too many compromises in CA could leave the most important city pair (SF-LA) too slow to draw from air. So you need to spend the money to both upgrade track speeds along Caltrain and also provide additional track so HSR can bypass regional services instead of slowing down.
Based off the baby bullet service a non stop service from San Jose to San Francisco assuming 3 minutes a stop (which is what the Chiltern line manages with albeit old DMUs) would take 48 minutes. If the stop penalty with those locomotives is actually 3.5 minutes which doesn’t seem silly then a non stop run would be 45 minutes.
A non-stop run for 50 miles out on London on the West Coast Mainline with a top speed of 125mph takes 30 minutes and it would be about 39 minutes on the Chiltern Mainline with DMUs and a 100mph top speed for that distance.
So certainly to start with taking 45 minutes to do San Francisco-San Jose wouldn’t be bad at all. And that assumes no real speed up at all from electrification for any slow points en-route.
“So certainly to start with taking 45 minutes to do San Francisco-San Jose wouldn’t be bad at all.”
Given that it’s 15 min slower than what the WCML does at sub-HSR speed, it is bad. SF-LA is like Bos-Baltimore, it’s far, and it needs every minute it can spare. Plus, you can’t run SF-SJ non stop with current infrastructure because you would run into a Caltrain doing slower than that. So there has to be infrastructure cost on the 80km used by Caltrain.
The problem is that dedicated lines in urban areas are super expensive compared to in rural areas.
I think you’d struggle to get a dedicated new route from San Jose to San Francisco for under $20 billion or so.
So the question is whether the time saving is worth it.
You don’t need a dedicated new route, you just need to make the old route dedicated. The Caltrain ROW is wide enough for four tracks most of the way, where it isn’t it either doesn’t matter (the approach to SF where even HSR isn’t going fast) or can be worked around by scheduling overtakes in the wider stretches. Even better is sending HSR via Altamont Pass instead of SJ, which among its many benefits gets HSR off the Caltrain corridor sooner to limit conflicts.
So you don’t need $20B, but you can’t waive away the cost of improving the corridor to make CAHSR cheaper.
Clem Tiller’s blog, https://caltrain-hsr.blogspot.com has a wealth of information on this subject.
Caltrain track geometry is pretty straight to the point where their diesel trains outperform most LIRR EMUs travel time.
That said, the EMUs are capable of 110 mph, so given signaling updates, HSR could do the 50 miles from SJ to SF in close to 30 mins.
Martin, you just keep writing things that are incorrect.
Stop it.
Caltrain track geometry is not “pretty straight”. In fact, the assholes who have “run” Caltrain for the past three decades have self-inflicted even worse curves with even worse speed limits.
And “signaling updates” have less than nothing do with the matter.
Try reading and understanding Caltrain HSR Compatibility Blog “The Top 10 Worst Curves” from 2009 (2009!) Not that this is the first time you’ve ignored it.
Actual SF Transbay to SJ Cahill non-stop trip times, along Caltrain’s actual line, with aggressive but not crazy 7% schedule padding, range from 47 minutes (today’s diesel locomotive hauled trains and speed limited to 79mph) down to at best 37 minutes (modern EMU, assuming many tens of billions of dollars for straighten curves and for fully grade separating to allow 200kmh operation and for extensive quadruple tracking to allow any such fast non-stop trains to run at all while sharing a 80km long corridor a score of intermediate stations.) See actual run times based on numerical integration of real train performance.
30 minutes SF–SJ is never happening, under any combination of circumstances. Of course real data have never stopped anybody from typing unrealistic nonsense.
I do appreciate the explanation-not-excuse elucidation; people often get confused by that. Just the same, one of the benefits of being a latecomer to something is that you can learn from whoever came before you. Repeating mistakes that lead to the same cost overruns as were made by the first system—60 years ago—is pretty hard to excuse.
Not even meaning to imply a turnkey Shinkansen would be the answer; that’s not working out too well in India or Texas. But if we’re throwing around stupid amounts of money, we may as well offer an absurd salary to some JR/Korail/SNCF/Trenitalia people and have them tell us what to do.
“When we decided to build our two-lane highways, we spent what was necessary and established gasoline taxes to pay for them. Ditto our interstates. When we committed to building airports and an air traffic control system in the 1950s, we spent what was necessary and established an aviation trust fund.”
This is all true, the problem is that these both made financial sense because the revenue from gas taxes and airline fees exceeded the capital outlays to build the roads and runways. The question becomes is this true for HSR. I know Alon has reviewed lines that turn an operational profit, but is it enough to also cover capital costs? Even if there were locations where it did, in the US the places where this would be so are limited, US size and geography are such that in many places air and car travel are superior to train/HSR travel time. The NEC would obviously work, as would the Toronto Montreal corridor in Canada. SF-LA as well, just based on California’s size. Texas Triangle or T-Bone again based on sheer size. Miami-Orlando-Tampa because of linear geography. Maybe Chi-Det/Cleve based on size, right-size distance, and historical connections. Everywhere else you end up subsidizing the rail transport, which means you won’t see a major commitment to a national network like you had with the interstates or air travel (or with rail transport in the 1880s, through the different mechanism of the land grant process).
Speaking of updating the NE Corridor,
in todays Wall Street Journal:
New Jersey’s New $2.1 Billion Bridge Leaves a Sludge Barge Stranded
A tugboat captain who transports sewage on the Hackensack River says a bridge will be too low for his vessel, which could force the waste onto trucks.
https://www.wsj.com/articles/new-jerseys-new-2-1-billion-bridge-leaves-a-sludge-barge-stranded-1653ca19?mod=hp_listc_pos1
Just buy the guy a new tug boat that is 5′ lower, which is a rounding error for this type of project.
Lots of these movable bridge projects would be cheaper by buying out the marinas upstream of the crossing.
Ugh read the article fully and it’s worse than dumb. The guy currently can’t use the bridge for 10 hours per day because it won’t open during rush hour.
The new bridge will be fixed, and he won’t be able to get under it at high tide. “My tugboat is 49 feet, so at high tide I won’t get under it.”
So basically, he won’t be able to get under 10 hours per day, same as it is right now.
https://marineweather.net/tide/amtrak-rr-swing-bridge-hackensack-river-nj-tides
This guy is just trying to squeeze money out of a project because tides. Don’t give him anything. If he wants to keep the contract, he can move shit (literally) during medium and low tides.
The problem is that one’s right to take a larger boat up the river is probably grandfathered in.
Probably a common law thing from the island nation that ruled America before its independence if nothing else.
From the article: “The U.S. Coast Guard concluded in 2013 that the new bridge wouldn’t disturb river traffic, and work on the span began last year.”
Ah fair.
Some refreshingly honest journalism:
“He has considered changing his tugboat by installing hinges on the mast and antennas[…] “We are most probably going to have to buy a smaller tugboat,” he said.”
In other words: there isn’t any story, publishing and reading this was a waste of everyone’s time.
The Wall Street Journal is owned by News Corp. The “soft” news can have lots of F.U.D. in it. The editorials can directly contradict the “hard” news. …. the editorials could directly contradict the hard news before News Corp bought it too.
He also mentioned that he might need to spend $250K on a hinged antenna mast. Amtrak might as well just drop that check in the mail before spending 10x on lawyers debating with him.
It’s not Amtrak’s problem or the Coast Guard’s problem or the sewage authority’s problem that he has a boat that doesn’t fit in the Coast Guard’s limits. And there are other barge haulers in the world too.
Lots of references to Obama, LaHood, and Buttiegeg; but little about Congress or the actual legal requirements and constraints of the law and the legal structure for American planning.
Congress is our legislative body. If the current setup can’t get us a functional transportation planning apparatus, they can/should/are obliged to change the laws that will.
Yes, because this is an entirely executive issue. There already is a federal planning framework in the NEC Commission; it’s just run with the same competence level as the California HSR Authority of the 2000s, as a collection of parochial interests rather than any kind of coordinated planning apparatus.
I read somewhere that federal funds might help restart the apparently dormant Texas Central project between Dallas and Houston. I don’t know how much commitment of private funding is still on the table, but completing an HSR line as a showpiece public-private partnership in a red state would make a big impact. I would just hinge the funding on several conditions to correct for the major weaknesses which would have likely resulted in a very costly and nearly useless white elephant if the initial plan had ever been built:
1. Downtown termini which either use a good existing Amtrak station (Union Station in Dallas) or a newly built station that integrates existing Amtrak service and adjacency to public transportation (Houston).
2. A route that includes intermediate stops in mid-sized cities to increase overall ridership and anticipate a future Dallas-San Antonio line along the I-35 corridor. Instead of the ridiculous “Brazos Valley” stop in the middle of nowhere, College Station and Waco provide sizable populations, the former with the largest student population of any US university, and the latter which would result in the construction of about one third of a future line from Dallas to San Antonio. If the Hardy Road alignment out of north Houston were to be used, I could even imagine a useful stop serving the Woodlands-Conroe area (+400k population) on the way to College Station.
The love for PPPs is another unforced error, one that unlike the others I’m detailing in this post is new rather than an intensification of errors from 12-14 years ago.
Yes, certainly the tracks and stations themselves I think should ideally be 100% public infrastructure to prevent the compromises that private investment entails. With my proposal I was just assuming participation of the existing private investors behind Texas Central to form a PPP with the federal government as at least an improvement over the purely privately financed model it initially operated under.
The pronounced boosterism of the project by Texas authorities that was exhibited in the initial years after the project announcement always gave me the impression that the ambitions behind the rail proposal were more than merely a means to provide a good and effective transportation alternative between Texas’ two biggest cities, but it was fundamentally an ideological competition with California’s publicly funded HSR project — to prove the superiority of the free market over public spending. It didn’t quite work out the way they had hoped.
In terms of the remaining pot, they should create a “Federal Rail Tunneling Agency” partner with the army corp of engineers and have them use five billion of the money to design build and maintain the HSR tunnels to Gilroy and the HSR tunnels from Bakersfield to Palmdale. If you bridge those gaps and take all the private sector out you could have a hodgepodge SF to LA around the 2033ish deadline the private sector will deliver the Central Valley spine of the HSR.
If the tunnel funding stays in the private sector, the tunnels of each of those reaches will each cost 100 billion and won’t be usable until 2150 at the earliest.
Just a nitpick: Kerava-Lahti is a commuter line, yes, but it was created also to lower travel times to the east of Finland and St Petersburg, being the first Finnish line to be built with a geometry allowing for 300 km/h running if needed (although as no trains are capable of it and freight trains are also running on the line, the track max speed is currently 220 km/h for Pendolino trains and 200 km/h for Intercity trains).
Wait, how did Finland build a 300 km/h line for this cheap? This is LGV Sud-Est cost; I thought it was just a medium-speed line with geometry designed for 220 km/h.
The line is not canted enough for more than 220 km/h but the turn radii etc. base geometry were future-proofed with 300 km/h running in mind. Sadly Finnish wikipedia does not cite the source, but it is cited in various other places as well, including this website from the time: https://web.archive.org/web/20050215005210/http://www.oiko.phnet.fi/artikkelit/Oikoradan-urakat-2003-2006.html
An inflation calculator tells me that 330 million in 2006 is 475 million in 2023, so a half a billion project. Unsure how this was achieved, as I did not actively monitor the subject at that time.
Ha, found a better source, a blog post from the traffic infrastructure agency: https://helsinkipietari150.vayla.fi/pietarin-rata-nyt-ja-tulevaisuudessa/
“Kerava–Lahti -oikoradan geometria tehtiin kaukoviisaasti jo alun perin 300 km/h nopeudelle, sähköistys jo nyt 250 km/h vauhdille ja raideväli 4,7 metriä 300 km/h nopeudelle. Koeajoissa on jo ajettu noin 250 km/h ongelmitta.”
-> Base geometry is 300 km/h, electrification allows 250 km/h and the track separation 4,7 metres is enough for 300 km/h, test runs have been done at around 250 km/h.
I’m curious what your thoughts are on why Chicago/Cleveland/Detroit and not the Texas triangle? Perhaps we just have different takes on American politics, but I feel like getting one conservative state to buy-in is easier than two conservative and two liberal to work together, and the landscape should be similarly easy. Especially if a foreign corporate operator that the Texas state government would like would agree to run the trains, and not Amtrak.
The cities of Austin/Dallas/Houston/San Antonio (whichever parts of that are fundable) would *probably* represent higher ridership if neither proposed plan is connected to anything else (obviously tying Cleveland to NEC would change the calculus). And Texas has shown a willingness to ignore NIMBY complaints in ways that much of blue America can only dream of.
Obviously very unlikely, but part of me thinks that JR Central, Trenitalia, or DB executives would actually get along pretty well with Texas state officials if funding was no object.
I see now this is kind of addressed in a comment above… so perhaps I should go look up how the current Texas HSR project is going instead!
1. The Midwest is flatter than Texas.
2. Texas Central made a likely irrevocable mistake of locking itself into an elevated alignment to avoid splitting farms (cf. French land swap deals); high-cost elevated construction is the norm in Japan, and at this point Texas won’t tell the farmers to get stuffed while working out land swap deals, Texas YIMBYism not extending to tolerance of takings. The Midwest can more easily avoid this error, partly because JR isn’t involved, and partly because in the absolute worst-case scenario in which land swaps are not possible, the line can run due east along the boundary between square farm plots.
3. Chicago is bigger than Dallas and Houston – on current population the expected traffic density is considerably higher on Chicago-Toledo than on Dallas-Houston and almost as high to Detroit and Cleveland.
4. Chicago is much more centralized than Dallas and Houston. Trusting ridership estimates trained on European or rich-Asian data in Texas is risky, since Dallas and Houston have very weak city centers for their size by non-US standards. Detroit and Cleveland are very decentralized too, but because they connect to Chicago, it’s a smaller risk than decentralized-to-decentralized connections.
Makes sense to me, thanks for the answer!
(I am from Chicago for what its worth)
On your notes about Indiana/Ohio hostility to high-speed rail, and on “committing too early,” perhaps the move would be to announce that the DOT will build one complete high-speed corridor, federally planned, and have states lobby for it.
If all the states on a route sign up to a program of prerequisites – matching construction and/or operating funds commitment, deferring to DOT planning, environmental clearance, a laundry list of limitations on “other people’s money” demands – they can be in the running.
DOT hires some international experts, puts together alignments, timetables, and budgets, and makes the big announcement about who’s getting the complete HSR line in Summer 2024.
As you say, they won’t do the big commitment anyway, but I think this method of project selection would help with publicity and knocking down hurdles.
Yeah, they could, but DOT is so worried about shovel-readiness that it wants to only fund things that are in more advanced design (like California HSR) or pretend to be (like Brightline West) or occasionally things that have rushed designs and apply for these funds to design the project that they’re simultaneously asking for funding for (this happened with GLX).
If they don’t pick something already in planning they won’t be spending construction money until 2040.
…. sumptin is seriously wrong with how we do things. The Lackawanna Cutoff railfans are all a twitter about service to Scranton being on the list. And has a hope of making it to the next phase because the right of way is in place and there are completed studies. According to Wikipedia the Delaware Lackawanna and Western decided it needed to upgrade, in 1905. And the whole double tracked Cutoff opened in 1911. NJTransit decided to extend service to Andover in 2008 and they are projecting that it might be in service by 2026. Along a route they already owned. Andover to the Delaware moves along at the same pace Scranton might have trains in 2045.
Rancho Cucamonga is a suburb, Victorville is an exurb 😉
Wait, you state that “committing too early” to a project is an ARRA problem that’s being repeated, but then your proposed solution is an HSR project along a route that’s had zero planning done so far? That doesn’t make sense to me.
Everything Alon describes here is the result of a lack of political coalition in favor of rail, at least outside the Northeast. Roads and car use are massively subsidized in the US. Rail can’t compete until vast sums are shifted from roads to rail. The sunk costs of the road system create far too large a set of vested interests in roads to allow for the creation of a political coalition for rail, except in the Northeast. This is about the gargantuan advantages of the US road system, not about the inability of Americans to understand rail.
A large percentage of the population things trains are a Communist plot to take away their cars. And sap and impurify their precious bodily fluids. There is another group who think traveling is sinful especially if you are going to those cesspits of immorality, cities. Both groups overlap and are joined by people who don’t want to be around anyone who isn’t exactly like them.
Yes, but then you get to cities where people don’t believe that, and those cities can’t build either. Fucking Tel Aviv is building more than New York, and this is a city that up until the moment the Red Line opened had similar levels of faith that a subway would ever open that New Yorkers have about future phases of Second Avenue Subway.
No city is an island. NYC is part of the US. It’s only able to pursue it’s own agenda in the context of the vast interstate/automotive industrial complex’s hold on power in the US as a whole. The important point isn’t that NYC has vastly overpriced subway projects. The important point is that NYC even exists and has a subway at all.
US states especially the urban blue ones have the most fiscal autonomy of any democratic Federal system. They also control their land governance and labour laws. They can totally do it. They have done it before.
New York is really an outlier in terms of the small amount of transit construction it has done since the 1950s. Even cities like Manchester, Liverpool and Newcastle have all built significantly more stuff for Christ sakes.
US States get a large share of their transportation spending through federal funding systems that heavily favor roads. That doesn’t contradict my point at all. NYC HAS done a lot in the face of a vast federal bias towards roads.