30 Comments

You are right that the greatest efficiency gains could be reaped by those cities which are best able to accommodate auto-autos. Or maybe people will move toward [smaller] metropolitan areas which are already incidentally well-suited for such cars.

Saying that cities would be more efficient neglects the coordinative and capital costs of adjustment. Maybe instead, changing smaller cities and seeing migration is the most efficient outcome. (Some large and rigid cities would become like Detroit. Dynamic ones would be revived after property values have fallen such that adjustment is cheaper.)

Expand full comment

Probably huge potential under South Korean guidance, though the initial costs would be enormous (South Korea has hundreds of billions of headroom though because of their current low debt/GDP ratio).

Expand full comment

I was wondering about the potential North Korea might have if reconstruction ever begins there.

Expand full comment

"Self-driving cars could drive fast close together to increase road throughput" This is usually a formula for less density, not more. Post-war metropolitan areas are more spread out than older cities for a reason - more efficient road structures.

Expand full comment

They're not actually doing this, unless you're talking about the biggest cities in other countries.

Expand full comment

One thing that may help is that in the USA the biggest cities are gaining population at the expense of the other cities.

Expand full comment

But does anyone think it would be particularly difficult for them to meticulously map essentially all the roads in the U.S. in a year or two, if they chose to put their minds to it?

Expand full comment

Well, wouldn't there be a long wait during rush hour?

I find waiting for a taxi one of the more unpleasant ways to travel.

[When I lived in New York, I felt in transportation heaven. But while cars were generally unnecessary and I rarely drove mine, many still insist on driving their own car.]

Perhaps a reason the infrastructural investments needed for auto-driving cars won't be forthcoming is that they would be so speculative, so dependent on assumptions about public tastes.

Expand full comment

It's not completely unjustified. Large cities offer a much higher degree of labor market flexibility. People are much more likely to stay in their jobs or accept a sub-optimal one if it means not having to move. Especially if they have a family.

Expand full comment

Sorry about the spelling :/

Expand full comment

I aggree that think self driving cars won't get people to share rides but rather to share cars. It will be like a giant Taxi service. When you don´t have to pay Taxi drivers the ride should be a lot cheaper. That combined with better efficiency will make using these services cheaper that owning a car. Most people use their cars maybe 1-5% of the time, while a service with thousands of cars in constant circulation should be able to have a much higher efficiency. Instead of the car standing all day in a parking lot it could be driving other people. You just order a car with an app and it drops you off without parking. Then you get a new car for the ride home. The result: No need for as many parking spaces.

Expand full comment

Hanson: "four times bigger cities could plausibly be twenty-five percent more productive"

Researchers observe that efficiency for cities seems to increase more than proportionately to size, say 2.2 times the output when the city doubles in size. That is close to the plausible 25% greater than proportional output for a city 4 times as large. This seems to recommend collecting people together into larger cities to gain that efficiency.

I want to put a fly in that ointment. This could be another case of the Planner's Fallacy (my term). Planners observe a relationship which has evolved naturally and think that they can change one side of the relationship to affect the other side. That confidence is not at all valid because they have no idea about cause and effect and unaccounted third factors.

Something has caused the bigger cities to grow bigger. Maybe there are natural efficiencies for each of them which has caused its growth. Maybe they grow until that natural advantage is REDUCED by the additional population to the point where they still have an advantage, but not as great a one. I don't know and I don't think that the planners know.

Detroit has lost about half its population in 60 years. Whatever advantage it once had seems to have disappeared.

The current planning craze to collect everyone into megacities for a utopian future is completely unjustified. People consume both GDP and unpriced goods (like nice views and ample parking). It is a big mistake to promote plans merely because of dreams of greater GDP, and an even worse mistake to not know what causes what.

EasyOpinions.blogspot.com

Expand full comment

The reason Robin gave for why less parking space would be required is people would share rides, and I don't see why they would do so that much more than with driven vehicles.

But it now occurs to me what might have been obvious to Robin: the self-driving cars can park themselves somewhere in the periphery. (One hopes gasoline will be cheap.)

Expand full comment

My money would be on malls and other areas with huge parking lots in the more successful suburbs. People already want to be there, and the redevelopment costs are far, far lower than in the old traditional neighborhoods. Pre-auto areas often have the density to support walking neighborhoods but they were designed for a very different technology and community than we have today.

Expand full comment

The subway doesn't go to your door. In many cases, it doesn't even go to your state. Auto autos will go to everybody's door.

The reasons for higher density are very simple. Developers want to make money, and they get paid basically by the square footage of the buildings they put up. Payment for parking space is trivial. Now go to any suburban commercial or industrial area and look at the lots. The overwhelming majority is parking spaces and driveways. If you no longer need much parking, the developer can build twice the building on the same lot - and they will, because they'll get twice as much money.

Longer term, there's a potential for a big shift in city orientations. There are low-density car-oriented areas and high-density pedestrian-oriented areas. There are economic forces which encourage denser areas over time, but it's hard to shift because there's a range of densities that are too dense for cars but not dense enough for pedestrians. Once neighborhoods hit this range, they tend to stop getting denser and so never transition to pedestrian. With auto autos, car-oriented areas will be able to "densify" up to the point where people are willing to just walk instead and so you'll start seeing the transition.

I do think the issue of stabilizing population is less of an issue that Robin thinks because a lot of recent construction is relatively low-quality and not designed to last, and the pre-auto construction is now getting really old. So there will be a lot of replacement of buildings and opportunities to build for a more compact community even without population growth.

Also, since denser neighborhoods are more productive and valuable, there will be a race because property owners in areas becoming denser will benefit at the expense of those in areas that don't. Huge sections of perfectly viable cities in the Rust Belt become partially abandoned in the late 20th century, even as construction continued in their suburbs. We'll now see a similar phenomenon, but sort of in reverse, where remoter suburban areas get partially abandoned as people concentrate in livelier suburbs where their friends hang out and their jobs are.

Expand full comment

My tastes run for greater agglomeration. But I too don't see the reasoning that more efficient automobile transportation makes for greater population densities. I would think the way to get greater density is efficient public transportation in the urban center. (Think New York versus L.A.)

[Maybe I'm a Luddite, but why do we need self-driving cars, given the huge environmental effects of autos? If you want to be driven, take the subway. Moreover, self-driving cars running bumper to bumper will force driven cars off the roads, and taking a real drive is fun.]

Expand full comment