« Assorted links | Main | What do kids find worth fighting over? »
A cost-benefit analysis of high-speed rail
Matt Yglesias points us to this survey of costs and benefits from a Dallas-Houston high-speed rail link. I'm not convinced by many of the particulars of the argument, which claims to show that the link is a good idea. For instance will the train line really be built with green energy? Will 80 percent of flyers take the train? Is Madrid-Barcelona a good analogy?
More generally, my jaw dropped when I read the denouement:
In this more comprehensive model that takes into account trivialities like regional population growth and a reality-based route, the annual benefits total $840 million compared with construction and maintenance costs of $810 million.
I'm not sure what discount rates he is using but even if we put that problem aside this screams out: don't do it. Given irreversible investment, lock-in effects, and required hurdle rates of return, this still falls into the "no" category. And that's an estimate from an advocate writing a polemic on behalf of the idea. I'm not even considering the likelihood of inflation on the cost side or the public choice problems with getting a good rather than a bad version of the project. How well has the Northeast corridor been run?
So, on high-speed rail, count me as still unconvinced. Nonetheless if you know of a good cost-benefit study, of a single rail link, not in the Northeast corridor, favoring HSR, let me know in the comments. I'll try to read and report on it.
General remark: It's not about population density per se. It's about how many independent, hard-to-connect nodes the system has and that is why high-speed rail on the whole works better in Europe or Japan than in many other locales. To give an example from a slightly different realm, I live right near the Metro in a high-density suburban area. Yet I don't take the Metro to my Arlington office, which is about two minutes from a Metro stop. I'd rather do the 37-minute drive. Why? Because I stop at the supermarket and the public library on my way home at least half of the time or maybe I stop to eat at Thai Thai. If those conveniences were right next to my house I'd consider the Metro but they're not. The fact that my neighborhood has lots of people doesn't help me any. In Tokyo you could live for years within the confines of many (most?) individual city blocks.
Posted by Tyler Cowen on August 27, 2009 at 01:31 PM in Economics | Permalink
Comments
The proposed Southeast High Speed Rail Corridor is an attempt to extend the NEC from DC to Richmond-Raleigh-Charlotte (and then eventually Atlanta). Their feasibility study is here, with claimed cost-benefit analyses. This study was done in 1999, and the basis for them proceding.
Of some note, another part of the analysis mentions that the DOT has performed cost-benefit analyses of various other proposed HSR links and found them generally to have total benefits below total costs. The benefits of the SEHSR only exceed costs because it links up to the already existing NEC.
Posted by: John Thacker at Aug 27, 2009 1:42:13 PM
It's about how many independent, hard-to-connect nodes the system has and that is why high-speed rail on the whole works better in Europe or Japan than in many other locales.
Cato argues is that high-speed rail doesn't 'work' in Europe or Japan either -- in the sense that France and Japan have spent about as much on their high-speed rail networks as on their freeway systems, yet the freeway systems carry 10-20 times more traffic.
Posted by: Slocum at Aug 27, 2009 1:43:08 PM
Yes, but regular low-speed rail and subways work in Japan fine and is even profitable, which makes sense from Tyler's point about Tokyo. HSR is a little different, especially the other shinkansen lines besides the Tokaido (Tokyo-Kyoto-Osaka) one, built only because everywhere else in the country had to have one too.
The point about the Metro and Tokyo is really about subways and regular commuter and heavy rail, not about high speed rail.
Posted by: John Thacker at Aug 27, 2009 1:51:18 PM
One assumption in the article Tyler links to is that cities will successfully encourage thousands people to move to inner city station areas so that high-speed rail can be viable. The author describes this as a "conservative" assumption. I would choose another word, like "dubious".
Posted by: Millian at Aug 27, 2009 1:56:26 PM
The HSR in Texas is about to get really interesting, as both of the main contenders for governor are record in favor, but there is tremendous public doubt. (Note: we stopped electing Ds to state-wide office in 1998. This is a R-primary fight.)
Posted by: Right Wing-nut at Aug 27, 2009 1:56:51 PM
Tyler,
If you, and those who shop like you do, used more rail, it is likely that shops like you prefer would open nearer to the rail stations. Certainly, if you lived in Dupont Circle, you would use rail more.
If gas cost $8/gallon or you had to pay $20/day to park, would you use rail more?
Posted by: Allan at Aug 27, 2009 1:57:29 PM
The basic thing that Yglesias misses in his obsession is that rail travel is an inferior good.
It's not that people outside a select few cities drive because they don't have the opportunity to take the train, it's that people in these cities take the train because they can't drive.
Posted by: Chris at Aug 27, 2009 1:57:36 PM
"Construction and maintenance costs"? What about other operating expenses, including the costs of actually running the trains? Or is Y. just being careless with terms?
Posted by: George Selgin at Aug 27, 2009 1:58:16 PM
Tyler wrote:
To give an example from a slightly different realm, I live right near the Metro in a high-density suburban area. Yet I don't take the Metro to my Arlington office, which is about two minutes from a Metro stop. I'd rather do the 37-minute drive. Why? Because I stop at the supermarket and the public library on my way home at least half of the time or maybe I stop to eat at Thai Thai. If those conveniences were right next to my house I'd consider the Metro but they're not. The fact that my neighborhood has lots of people doesn't help me any.
That's because you don't live in a properly designed linear city. Then again, nobody really wants to live in a properly designed linear city.
Posted by: Ironman at Aug 27, 2009 2:02:37 PM
Anyone that is in favor of investing a huge amount of money in high-speed rail on the basis of a serious c/b analysis should first explain how he/she assess the ongoing reconstruction of the Bay Bridge. Although the benefits of this reconstruction are huge, the actual costs are going to be much larger than estimated initially and may even be "too close" to any updated estimation of the benefits (of course, now that much of the investment is a sunk cost, there is no reason to stop the reconstruction). The story of the Bay Bridge cannot be ignored by anyone proposing large infrastructure projects.
Posted by: E. Barandiaran at Aug 27, 2009 2:03:01 PM
Sorry--I inadvertently referred to Matt as if he'd authored the survey! My bad.
Posted by: George Selgin at Aug 27, 2009 2:03:19 PM
Michael Moore had a line in his Huffington Post essay about this that still makes me laugh:
http://www.chelseagreen.com/content/michael-moore-goodbye-gmhello-high-speed-rail/
The money quote:
"The fact that the technology already exists for us to go from New York to L.A. in 17 hours by train, and that we haven’t used it, is criminal."
The best feature of high speed or light rail is that with our federal system, other people will pay for it. Even considering that, it's probably not a great deal.
Here is a good skeptic site:
http://ti.org/antiplanner/
Posted by: kebko at Aug 27, 2009 2:11:05 PM
When Mr. Freemark says things like:
while transit-friendly conditions are desirable – and it bears mention that both Dallas and Houston are expanding their transit systems significantly – there is little evidence those networks are vital in attracting customers to high-speed rail.
it makes me wonder if he has ever set foot in either city.
Posted by: Blackadder at Aug 27, 2009 2:14:20 PM
Has the author read "Bent Flyvbjerg, Mette K. Skamris Holm, and Soren L. Buhl, “Inaccuracy in Traffic Forecasts,” Transport Reviews, vol. 26, no. 1 (2006)"
The GAO had this to say about it: "Another study that included 27 rail projects, 1 of which was a high speed rail project, from around the world found that rider forecasts for over 90 percent of the rail projects studied were overestimated, and 67 percent were overestimated by more than two-thirds."
Posted by: Eric at Aug 27, 2009 2:16:16 PM
"Author" in my comment is meant to refer to the Infrastructurist, not Tyler.
Posted by: Eric at Aug 27, 2009 2:17:44 PM
Tyler, Why do you exclude consideration of the Northeast corridor? It would make more sense to exclude all other locations and focus exclusively on this region.
The Northeast corridor is the most obvious place for HSR in the U.S. Most of the right-of-way rights are already in place but just need upgrades. Boston-NY-Philly-Washington would get a ton of monthly, weekly and even daily travelers. The agglomeration benefits and spillovers from knowledge sharing from this region would likely surpass any other megalopolis in the world.
Posted by: a student of economics at Aug 27, 2009 2:18:12 PM
And what is a "reality-based" route?
Everytime I read 'reality-based ' it seems to mean that the author pulled it out of his butt because its the only thing that'll work, even though it has never, or will never, exist.
Posted by: Tom at Aug 27, 2009 2:19:32 PM
There are 5 shinkansen lines in Japan:
Tokaido-sen - Tokyo to Osaka
Sanyo-sen - Osaka to Fukuoka
Joshinetsu - Tokyo to Nigata (to the Japan Sea side)
Tohoku-sen - Tokyo to up north towards Hokkaido
Hokuritsu-sen - Along the Japan Sea side from Nigata down through Komatsu (Ishikawa-ken) to Nagoya
Of these shinkansen lines, only the Tokaido-sen is profitable. The others are subsidized by the national government. Having ridden it many times, I can tell you why the Tokaido-sen is profitable. A train runs every 3-5 minutes from 8AM until 5PM every work day and everyone of these trains are full up. Each train has 15 carriages, with each carriage with the seating capacity of a 737. That is, each train carries the equivalent of 3 747's worth of passengers, every 3 to 5 minutes.
Outside the NYC to Washington DC corridor, there is no traffic corridor anywhere in the U.S. that comes close to the kind of traffic density on the Tokaido-sen.
This is the reason why high speed rail cannot be profitable in the U.S.
Posted by: kurt9 at Aug 27, 2009 2:26:29 PM
A ~4% Roi, and only if things keep on getting better. On a project that will take MANY years to complete? This screams boondoggle to me.
Posted by: Doc Merlin at Aug 27, 2009 2:29:10 PM
Lots of data on energy consumption per passenger mile. The relevant stuff starts on page 285.
http://www.bts.gov/publications/national_transportation_statistics/2008/pdf/entire.pdf
Not as great as one would hope.
Posted by: Floccina at Aug 27, 2009 2:29:55 PM
I often fly from Dallas to Houston and back for a day trip, back home by dinner. I don't think I would be able to do that on a train.
Posted by: rob at Aug 27, 2009 2:35:24 PM
I'm continually amazed to what degree the supposed "reality-based" community is drunk on the HSR kool-aid.
Rail advocates just refuse to see the HUGE advantages of an individual vehicle that can travel an individual route on an individual schedule.
I just wish there was some way to make money betting against the HSR advocates and their outlandishly optimistic predictions!
Posted by: ed at Aug 27, 2009 2:37:39 PM
I would certainly have doubts about such a line. Most flyers are probably taking connecting flights, destination traffic is more likely to need cars once they get there since both are sprawling, etc. But payback in less than a year does not seem to be one of them (though what is designated maintenance is open to interpretation).
Posted by: Lord at Aug 27, 2009 2:39:48 PM
I have no idea of the specifics of the Texas line, so I won't comment on it. But what I can say is that, in large chunks of Europe, we'd rather take a trip by train instead of driving, even though we can drive.
For medium size routes, like Madrid-Barcelona, even in perfect traffic, the train crushes a car in convenience: The slowdown by not being able to leave until the train leaves is dwarfed by the difference in speed over driving and the convenience of being able to do something constructive on the train instead of paying attention to the road.
So it's not that people in those cities can't drive: It's just that the train really is better in quite a few cases. I know, I've been in that situation, and kept my car at the garage.
Posted by: hibikir at Aug 27, 2009 2:41:01 PM
Another issue I never see addressed: HSR advocates assume thousands of people per hour are taking the train. How are you going to get all these people to the train station?
Yet another issue: one purported benefit of HSR is that there are more security lines at the airport. But then why not just lower security at the airports, or put more resources into speeding it up? I think rail makes an easier terrorist target than planes anyway.
Posted by: ed at Aug 27, 2009 2:42:52 PM
