Thank you. I am mystified by the apparently widespread mental framework that roads don't actually do anything other than attract masses of people in cars, inexplicably like insects toward bright light.

It's true that roads convey people & permit economic activity.

But that road down the way that is so congested right now? If we expanded it, and it could convey more cars at the same level of congestion- where did they come from?

The latent demand represents potential traffic that either:

1) is doing just fine taking a different route right now

2) isn't particularly economically valuable, i.e. not worth waiting through traffic for

3) things like future sprawl enabled by wider roads today

In order, #1 is not particularly valuable to reroute, it was doing ok. #2 is not very valuable either, though perhaps a little. #3 creates new demand, but the only thing we get in return is lower density.

So I guess in sum the point is that all the most economically valuable travel is already happening, right now, and by widening the road we add only the travel that has such marginal economic value it's not worth doing today. So the more we widen the road, the less & less economic value we get in return.

>2) isn't particularly economically valuable, i.e. not worth waiting through traffic for

With schedules and business operating hours, many economic activities are impossible due to hard time constraints.

I don't see how your argument would be less applicable to a city with two dirt roads.

I know you didn't mean it, but honestly your analogy is kind of brilliant. It does feel that way as somebody who enjoys walking.

Versus what?

Presumably if someone wants to go from A to B, or take goods from A to B, they derive some social or economic benefit from that (as do the people at B).

Just aiming to maximize the throughput of private cars is not sustainable. The primary way to alleviate congestion problems should be making people less dependent on personal cars, both by urban planning that facilitates shorter commutes and by making mass transit more attractive and available.

That may be true, but not because of latent demand. Latent demand happens with mass transit and bike lanes too. It doesn’t mean we shouldn’t build those.

The point is to shift latent (and, indeed, actual) demand from personal vehicles to alternative means of transport. This is a net win when it comes to congestion because private cars have by far the greatest surface area per person transported (especially given that their passenger seats tend to be woefully underutilized).

There is a classic coordination problem regarding private/mass transport: as long as mass transit lags behind personal vehicles in convenience (and status), it is in the interests of car users that everybody except themselves switch to transit.