It's the land price. Rising land prices are the problem, and land is in completely fixed supply. You cannot induce more supply with rising prices. Land is the dominating factor, ergo local variations in housing elasticity don't really matter. Clearly there is some inelastic thing that is eating all the increased local income without inducing more supply: it's the land.
Edit: To the people saying "you can build more on each piece of land!" sure you can. And that should definitely mitigate price increases, but that's exactly what this paper says actually isn't predictive of prices, clearly because cost is dominantly driven by something that isn't affected by this elasticity. The value of higher density development gets immediately baked into the land prices, which then does not induce more supply of land.
Lots of people can live in a single unit of land, unless local zoning rules actively forbid that.
Tokyo occupies about twice as much land as Los Angeles, but has four times as many people.
Too many American cities forbid even relatively modest density such as duplexes or small apartment buildings (with, say, 6 apartments per building).
Demanding that people in metropolitan regions build exclusively single family homes on large lots is insanely uneconomic, and, arguably, a failure of democracy at a local level.
You also don't even have to urbanize all that much to get higher density.
Nassau County in New York is one of the birthplaces of postwar suburbia and has a population density higher than Los Angeles County, ranking at 22nd in the country. (4,705 people per square mile vs 2,420.)
To be fair, LA County includes mountains (Angeles National Forest among others) and a desert and a whole lot of nothing on the other side of those mountains. Nassau County is almost all developable land.
I bet if you compared like to like, the density won't be too far off.
There's externalities in having a home that aren't just the land itself, but also the infrastructure per person. The infrastructure costs are why very few people are buying up land in e.g. California City[0] and popping a cheap concrete[1] or steel[2] box (depending on your preference of 3D printing or prefab/shipping container houses) in the desert for less than many people here earn per month.
Did you want a road with that house? Running water? Electricity? Internet? A police force?
Yes, these things are all doable. But they also add to the value proposition of a property ("it's in a good area"), driving up demand, meaning people can charge more for the land, and if you're going to spend a lot more on the land anyway then you might as well make the structure of the home itself a bit nicer, and it all blows up rapidly.
I'm not sure how costs of sewage etc. change with increased population density. Pipes have to get wider per person, but low-density also means they're longer.
> I'm not sure how costs of sewage etc. change with increased population density. Pipes have to get wider per person, but low-density also means they're longer.
Probably accurate to say that in general, the great majority of the cost of infrastructure is in labor, not materials. It doesn't matter if pipes are a little wider or longer; the labor of digging up dirt and installing pipes, and doing that over time as things need maintenance, would far exceed.
IIRC where suburbs get the short end of the stick is lifecycle replacement. The primary costs for replacing pipes is in labor hours, not really pipe size, and it takes longer to replace more miles of pipe.
But trying to suppress land prices by limiting its uses just pushes those price increases nearby.
The most disastrous idea in urban planning is that prices can be kept low by limiting use, and trying to preserve "affordability" for a narrow slice of the middle class. It has failed entirely.
if you permit more density on a small amount of land, the value of that land will go up. there's only a few places to put the new units, so the good land is scarce, and the price goes up
if you do it over a whole city, i'm not sure it still holds. not as much, anyway, since the developers have more choices on where to build
I'm not sure that's an experiment we'd see unfold due to NIMBYism. (Not commenting on whether that's an appropriate reaction, just that it's inevitable.) So I wonder if within realistic constraints it's still a fairly safe assumption?
Land prices have been rising mostly because of productivity gains. Land is completely free. There is no inherent cost to it. All cost is derived from its productive power, and productive power rises as technology improves and as people congregate in greater density.
You could outlaw loans altogether and land would still not be anywhere close to free.
“Supply” in economics is the function mapping market clearing price to quantity supplied. Even if all supply curves slope upward, that's not increasing supply with price, it is increasing quantity supplied with price.
Edit: To the people saying "you can build more on each piece of land!" sure you can. And that should definitely mitigate price increases, but that's exactly what this paper says actually isn't predictive of prices, clearly because cost is dominantly driven by something that isn't affected by this elasticity. The value of higher density development gets immediately baked into the land prices, which then does not induce more supply of land.