The trick would be to build lots of itty bitty extensions continuously and have a small constant staff rather than big bang projects where the knowledge poofs into thin air, although now the cost has gotten so out of hand that we can’t afford to do that either.

This is the answer but small improvements aren’t “flashy” enough for the politicians (and voters) involved.

> You won't get competence in managing the budgeting, planning, and execution of a subway line without doing it a lot.

> So how do you break the cycle of inefficiencies leading to people scared to start enough projects necessary to build expertise?

I wonder if this could be something that's only possible when you're riding on fast exponential growth - fast enough that you can afford starting more and larger projects faster than population growth (or at least faster than competent people entering management work), with confidence that this "spending spree" is self-sustaining. Once you lose that confidence, you start worrying, lose velocity, and then lose capability.

This would explain why it seems that nations tend to rapidly industrialize and improve across the board, in a kind of push that lasts for couple decades, and then runs out of fuel, leaving the nation barely able to do any large project anymore.

An analogy in my mind is to a scramjet: an engine that allows you to achieve extremely high velocities, but which itself needs you moving very fast to even start it - and conversely, if you let your airspeed drop past the threshold, it will stop working.

Japanese housing policy has what seems like an inefficient system of knocking down any buildings older than some age when sold, with wood being 22 years, reinforced concrete being 47 years.

But it has three really great side effects

1. Because they always need to be building, they never run into this problem of losing institutional knowledge to keep building. 2. Buildings are periodically refreshed up to code for things like Earthquakes. 3. Because housing is never seen as an investment they don't get the NIMBY problem.

Maybe NYC doesn't need to build more. However there are a lot of cities that don't have nearly the metro system that NYC does. Minneapolis could use one downtown (they have light rail, but downtown it is just buses), and this could go to St Paul (ie make the green line a subway). Kansas City, Denver, Dallas are just a few of the large cities that could use a subway system but don't have it. Des Moines has more population than Stockholm when Stockholm started building their subway - that enabled a transformation of Stockholm's build form to become much denser (as opposed to become a car first city as the people got rich enough to afford more)

NYC needs to build more.

Transport networks really benefit from network effects; each additional line benefits from connecting to the existing network. The IBX rail project was recently proposed in New York and found to have 100k+ daily riders, which blows pretty much everything else in the US out of the water.

The other issue is that land use planning in the US except in a select few places like New York is really bad for transit. Transit can certainly be "build it and they will come", but also the land around transit needs to legally permit that as well. Rail is only really efficient at high utilization, and most US cities refuse to permit the development that makes rail useful.

This comparison picture really brings home the point that in most US cities the land use is causing rail value per dollar spent to be really low: Atlanta and Barcelona had similar population numbers and metro lengths in 1990. https://ti.org/images/AtlantavBarcelonametro.jpg

You need to set up management at a level where there is sufficient demand for it. Right now most projects are managed at the local level where yeah a rail extension is probably a once in a generation project, but at the state, regional, or federal level there's generally always something that needs doing.