The Metrorail system in Miami was met with the same white elephant boondoggle criticisms when it was built in the mid 1980s

It connects 4 or 5 separate communities in the Miami-Dade area over about 30 miles of track with stations every 2 miles or so. It has become a critical element of transportation infrastructure

Most of it is on raised tracks and the open space beneath was mostly useless until recently. Now there are playgrounds, walking/jogging/biking paths and small sitting gardens. People use them

It’s recently been connected to MIA and talk about connecting to Miami Beach spark up from time to time

Perhaps the HSR proposals should seek similarly less ambitious goals. Instead of connecting 1000s of miles across many states the proposals should seek to connect several metro areas across a few 100 miles (like Miami/Tampa/Orlando) a time

These lower scale projects are easier to judge costs/benefits and lower risk interns of capital investment. Such projects still offer a transportation option that is unique & of a different quality, kind & scale than the current alternatives

Expand full comment

A good point, which notes that a key is the pragmatic interconnectivity to "make it work" as a viable solution.

Just look at how the NYC Metro Airports (EWR, JFK, LGA) are connected to the City ... there's no "Gatwick Express" paradigm going on here.

For when one can't even make a 5 mile nonstop connection, there's no hope for a 50 or 500 mile one.

Expand full comment
Apr 11, 2021Liked by Noah Smith

PS: Did you paint that bunny with vegetable dyes to get that color, or do you have an orange tabby rabbit? I've never seen one like that before...

Expand full comment

Some jerks dyed that bunny orange and then surrendered her to a shelter. She was the sweetest rabbit. Eventually we found a good home for her!

Expand full comment

Oh, that's unfortunate. :-(

It is possible to _safely_ apply a vegetable-based dye to a white pet, depending how cooperative they are, although it doesn't last very long. We have some pictures of our white cat with a purple mohawk... Required gently painting it on when she was in a cooperative mood, and providing the occasional kibble reward.


Expand full comment

When I take the train, I'm so confused why it's more expensive than air travel. On a per kg basis, it seems like air travel should be way more expensive than train travel since propelling something along a track is way less energy-intensive than holding the same thing floating in the air. So what it is that makes trains so expensive to operate?

Expand full comment

What? Are you kidding? A train ride takes 3-4 days to cross the USA, with massive, massive infrastructure required. Megatons of steel and concrete. Bridges, food, cooking, etc. A plane zips along that same route in 6 hours.

Airplanes are much more efficient for long distance travel. So funny to me that so few HSR nitwits don't understand that.

Expand full comment

Except that you're looking at HSR all wrong. If you built HSR from NY to SF you wouldn't have many people riding FROM NY to SF but you would have a LOT of people riding from NY to Philadelphia and Pittsburg to Cleveland and Chicago to St. Louis and all of the cities in between. You look at HSR like it was a freeway built between NY and SF with no exits at any of the cities in between the two. Not many people would use that. It's usefulness is all the places in between that would now be connected to each other.

Expand full comment

Airplanes still involve megatons of steel and concrete. Have you never seen an airport?

Expand full comment

Sure - but airports are the equivalent of train stations - not track.

Expand full comment

Not quite. DFW Airport is larger than Manhattan. So a LOT more steel and concrete than a train station.

Expand full comment

For cargo, rail *is* considerably cheaper than air travel on a cost per weight basis. But passengers aren't shipped by weight - they tend to be less dense than cargo, since they need room to stretch and move around (plus dining and sleeping areas if it's a long trip). So you don't really get the benefits of that massive weight capacity.

Passenger trains are also much shorter than cargo trains - even if you *can* pull 100 passenger cars with a single locomotive, you don't have enough demand to fill 100 cars. So again, you don't get the benefit from not needing to lift that weight off the ground - there just isn't that much weight to be moved.

Expand full comment

That seems to imply that if you had way more demand for train travel, it might get cheaper than air travel.

Expand full comment

That's correct.

The constraint on the length of a passenger train is the length of the platforms at the stations, but they can be lengthened (this is costly, but not infinitely so).

There is a limiting constraint even then, though, which is the passenger circulation space. If a train disgorges 1500 or 2000 people all at once, there has to be somewhere for them to go, without crowding so hard that people get crushed, and they have to get out of the way quickly so another 2000 can get on and the train moves away before the next train comes in. At a busy city station, that can be only two or three minutes. You can clear 2000 people in two minutes if you build enough wide corridors and escalators and elevators and have vast circulation areas. But that's about the physical limit.

If you could operate a high-speed rail line that was moving 2000 people per train, 20 trains per hour, 20 hours per day, then the tickets would be much cheaper than flying. The infrastructure costs of that are the same as a more typical 500 people per train, 3 trains an hour, 16 hours per day, but the per-ticket costs are clearly vastly lower.

The problem is that there are no routes on Earth where 800,000 people will buy a ticket every day.

Now, that doesn't mean that trains aren't cheaper than flying. The Shinkansen is cheaper in operating costs on Tokyo-Osaka than flying (but they charge higher prices for tickets anyway, which means that JR Central makes enormous profits). LGV Sud-Est (Paris-Lyon) so completely outcompeted flying that there weren't any flights on that route even before the French government banned them.

But the capacity of high-speed rail is absolutely gigantic compared to demand.

London-Paris (two huge cities, only just over two hours apart) just about fills a 902-seat train once an hour each way (pre-covid). They could physically run 12 of those trains per hour. The tickets on the one they do run were servicing all of the debt incurred building the line and the trains right up until covid (the British and French governments have banned most travel between those two countries for over a year).

Expand full comment

Building HSR tracks & associated infrastructure costs a lot of money. It is built on land that costs a lot of money.

Expand full comment

HSR trains are extremely expensive & shouldn’t be built as a means to provide an alternative to air travellers especially when travel times are longer & fares won’t be cheaper.

Where HSR trains work well are between large cities in densely populated regions with congestion.

In the US the most obvious place for HSR is in the corridor from Boston to Washington DC serving New York, Philadelphia, Baltimore & other cities. That should be the main focus of HSR advocates.

Expand full comment

Even along that corridor, NIMBY is going to be a major problem. The greatest lengths of it will be through CT and NJ, both notoriously populated by uppity White suburbanites who will bring considerable resources into a tooth-and-nail fight against eminent domain.

Expand full comment

From my view in Portland... I think my desire for HSR in the US stems from:

- monument building

- a belief that the experience of a HSR ride is much, much better than a plane and given the choice, people would choose the latter

- a sense that local trains and HSR each go up in value as the other comes to exist. I hear you saying Local should be the chicken and HSR should be the egg and I'm okay with that, but would love to see them both come in, in concert.

- even though I moved from the mega-polis of LA to Portland over a decade ago, I think if Portland wants to thrive in the global economy, that it needs to link up with Big Bro Seattle, to more effectively compete with Miami, Austin/San Antonio/SF etc. As cities de-centralize and Zoom rises, I could see people in Portland wanting to live farther and farther into Washington state, which gives much deeper tax breaks to home-owners. If there was a HSR (or even pretty damn fast rail) linking downtown Seattle and downtown PDX, and each had a decent local network, in the long run, I could see that playing to the region's advantage, as it competes for talent and investment. Or maybe I have no clue and should stick to my background in marketing?

Expand full comment

It would be helpful if folks doing the maps would ballpark the ridership numbers and the costs and publish realistic numbers. Lots of city pairs are not likely to meet reasonable criteria on this basis, I would guess. Upkeep costs would be high. Money is scarce and we're going to be in a more difficult financial position nationally as more boomers retire while immigration is held down. The progressive agenda maybe needs some prioritization...

Also, I have ridden the same trains in Japan, China, and France and I also thought the experience was great. When possible I took these trains rather than planes. But the top trains are expensive in Europe...RyanAir etc move people for a lot less money, which has to be factored in. And Noah you are right about the network of local and regional passenger trains in Europe and Japan and even in China. Those were built first and make the HSR system a lot more usable. Seems like a good point that we start with those.

Also, the moon program generated a lot of basic research and built up our scientific and engineering infrastructure. Hard to see that argument carrying much weight in a nationwide US system, for reasons you point out.

Expand full comment

For me, at least, it's not monument building. I find train travel to be extremely pleasant, and air travel to be extremely unpleasant. I can do work or read easily on the train. I'm not cramped or uncomfortable. Getting to and from the station is easy (Penn Station is badly designed, but come on -- you're out in ten minutes or less), and I don't have to do it 1-2 hours early or wait in gratuitous lines. I'm not dreaming of how nice the tracks would look on a map: this is very practical for me. If there were a hi-speed train connecting Montreal and New York, or New York and Toronto, I would use them often. I would visit my friends for a weekend, often. I already use the Amtrak from New York to Boston often, but I would use it even more if an hour or more were cut off the commute.

Expand full comment

This post reminds me of the first controversy set off by Robert Fogel, one of the chief architects of the New Economic History and a winner of the Nobel Prize in Economics. In a highly influential book he argued that the economic benefits to building a transcontinental railway system in the United States were exaggerated. In making this argument he crossed swords with W. Rostow and J. Schumpeter who argued the transcontinental project gave a huge boost to economic growth and capitalism in the United States. Fogel compared railways to canals, many of which had been planned by state governments before Congress got behind the transcontinental project. His point was that transport systems compete with one another. What you need to look at are the marginal gains due to developing a new project. In the case of the United States we currently have airline transport for passengers, pipelines for liquids like petroleum, and trains most of which are devoted to cargo, largely shipped in containers. Airlines are fast and relatively cheap to fly, particularly on long-distance hauls (the main costs being getting up from the airport and landing at the other end). I fail to see how high speed transcontinental trains can compete with that mode of transport. Where competition is possible is on short-hauls. You can make a case for a high speed train going between San Francisco and Los Angeles for instance, largely on the grounds that the cost in time and resources of getting to airports are pretty steep and a high speed train with large central stations coupled to massive parking structures at both ends might be cost-effective relative to the airline solution. Beyond that the only other thing you can say is that light rail - not high speed but normal speed - is a good thing for within metropolis commuting. Portland is one example; BART in the Bay Area is another.

Expand full comment

I understand and mostly share your opinions. But yet, I remember the building of Dulles airport. Moved to DC in 1968 and IIRC it was called a "white elephant" then. Over the years it's had its problems but it seems a solid piece of our infrastructure today.

I remember driving the Beltway to get to NY from Fort Belvoir in 1966--it was a 4-lane highway then, and often empty.

So it's possible our timeframe is too small. HSR if built may be underused and a waste of money for its first 80 years or so, but something your grandchildren may accept as a vital element in a transportation grid serving a country of 1 billion population.


Expand full comment

To be fair the new Amtrak train hall at Penn Station is open and quite nice.

Expand full comment

Tend to agree (not American). I haven’t heard any talk about a bipartisan bill to reduce NIMBYism and improve construction tendering. However, I am pretty optimistic about the availability of short range eVTOL from about 2025, and hydrogen/electric narrow body jet by about 2030. The biggest constraint rn is for the FAA to figure out how to certify this stuff. At scale, eVTOL can be very cheap, so think of a network of sky taxis with a range of 250km in any direction and a speed of 250km/hr. These may be heavily regulated in major metros due to the nuisance effect, but a great way to connect small cities and towns.

Expand full comment

I'm glad to see eVTOLs mentioned and I'm also very optimistic of their potential and I think that they could radically transform society and have an even greater impact than the introduction of the car. It's not just that they could connect even small cities and towns, which they could, but that they can also connect every single point in-between - fundamentally reducing the need to congregate into cities to begin with. There is just something very revolutionary about a transportation system that hardly requires any physical infrastructure.

Expand full comment

Also, I feel like asteroid defense is something worth its own discussion -- it's a thing we definitely should do, but it's also extremely hard to determine exactly how. Any system capable of defending against asteroids can almost certainly also be used as a weapon, either directly, or, if it's the type that re-directs rocks rather than trying to blast them, then it can also probably grab a rock that's big enough to devastate a city, but not the whole world, and aim it.

It would almost certainly need to be an internationally funded project, with a control center that's internationally staffed, with 100% of control code open sourced, subject to incredibly strenuous scrutiny to make sure the control and update mechanisms can't be hacked by some doomsday cult.

Expand full comment

It will only take 17 short years to build a fast train from Madera to Bakersfield (assuming the CURRENT schedule holds). It’s going to be a tremendous monument to our American Greatness and the people of Bakersfield will rejoice because they are mostly HSR nerds.

Expand full comment

Highspeed trains never publish BTU per served real passenger mile. They never talk about real door to door times and how much Uber is involved to complete those trips. Advanced dual mode electric vehicles powered by PV solar is cleaner, faster and door to actual door.

Expand full comment

CA HSR has a bit problem that federal government would be good at solving. They're routed to cross three mountain ranges with their alignment. Pacheco in Northern california, and then Bakersfield-to-Palmdale, followed by Palmdale to Los Angeles, both in southern california.

Bridging these gaps with Passenger rail is an enormous undertaking, and mountain crossings take incredibly long amounts of construction time. One of the original sins of CA HSR was not in identifying one of those crossings, and prioritize it to begin right away, (Pacheco would be the obvious one, to start). Which would have been simultaneous with the federally required start that HSR chose in the central valley (to receive fed funds they had to have a MOS that could achieve top speeds on the first revenue day (and also was appropriate for top speed testing of train consists).

But that can got kicked down the road, which is unfortunate and ennables the potential for politicians and bad actors to scuttle the whole project even half built because of those gaps.

Expand full comment

To try to take your questions in turn:

Why HSR before local trains? I think Alon Levy has done much of the answering here, but I'll add that in a lot of cases the track is shared, so building HSR also improves or creates local train routes, and that you'd have to design local train upgrades such that they can be used by HSR.

To pick an obvious example, any improvements on paths into Penn Station in New York City from the New Jersey side will not just benefit Acela passengers to/from Washington, but also NJT commuters from all of North Jersey. If you can save five minutes by better track layouts into Penn Station, then everyone gains those five minutes.

Will HSR make America more connected? Yes, America is pretty easy to get around by car and plane (though, as someone with a physical disability that means I can't sit in most airline seats and can't afford the ones I can sit in, HSR would make it much easier for me personally to get around - but I'm in a small minority here). The more sensible designs include out-of-town stations with giant parking lots that people can drive to, meaning that it's no worse than airports and mostly better. Indeed, these stations would ideally be at the airport (though the exact location of the airport and the rail routing means this isn't always possible, e.g. Dulles is west of Washington and the logical rail route is North-South).

To the extent it moves travellers from the road to rail, then yes, they are less likely to interact with the areas they move through - though one wonders how much benefit there really is from drivers on the interstate who likely only stop at a roadside diner occasionally. In design terms, when going through a very rural area, you probably want a station about every hundred miles or so - people would be no more than 50 miles or so by car to the nearest station.

Ideally, you'd be closing lots of smaller airports and using the HSR as the local connection, both for people travelling to the nearest big city (from a rural station they drive to) and to the nearest hub airport for longer distance connections.

Who will use it, and how will this benefit the economy?

I'm much less optimistic than you are about people choosing not to travel for business. We'll see what impact the pandemic and Zoom actually have in a couple of years, though.

But the argument that people travelling for tourism and spending money doesn't benefit the economy seems to me to be just wrong. Hotels are still a big lump of economic activity; so are restaurants and entertainment venues. Hotels are almost exclusively used by travellers, restaurants and entertainment not exclusively but still extensively. And this activity isn't displaced from one place to another; people travel to go to a concert or to see a particular play, to visit a museum. Take Cleveland - it has a few tourist attractions that are national rather than regional in nature (e.g. the Rock and Roll Hall of Fame) but not enough for many people to make a specific journey there. If you have high-speed rail from NYC to Chicago, it will inevitably pass through Cleveland. That means that a lot of people from a wide area of the Northeast and Midwest will be tempted to take one or two days there, perhaps on the way to somewhere else. If you're flying from JFK to ORD, that's not really an option.

Why haven’t we been able to do this so far?

Yeah, your organisation and costing issues are a big deal. It's worth pointing out that the Bakersfield-Merced line will be (if built) Shinkansen-speed, so tails into LA and SF could eventually turn it into a true high-speed line; building part of a high-speed line can help if you can connect a regular speed line to it (this isn't an option in Japan for boring technical reasons, but they do it in France all the time - the line from Paris to Bordeaux was only LGV as far as Tours for 27 years, but there were TGVs in Bordeaux station throughout). If you can get a train from LA to Bakersfield and Merced to SF, then you can run a train all the way through and then upgrade the LA-Bakersfield and SF-Merced legs a bit at a time, shaving another 30 or 60 minutes off for another $20-50bn every decade or so. Not ideal, but you can build the line a bit at a time (and I note that there are trains from LA to SF already, so this is technically possible).

Is HSR monument-building?

Perhaps, and that is rather wasteful. I'd argue that the best non-economic case for it is that it makes it politically possible to ban some short-distance flights (if you can get from city A to city B by train in less then, say, four hours, then you're only allowed to fly if you're getting a connection) and that has a big impact on carbon emissions. If you could get NYC-Chicago times down enough that you could ban flights on that corridor, then you really would hit emissions in a big way.

Expand full comment