The causation goes the other way. When an area has a lot of housing demand, that's where they put the tall buildings. Nobody is going to build a skyscraper in the middle of a corn field.
But in order for the density to keep up with demand, you can't suppress new construction.
> What are the two densest cities in the US? Which are the two most expensive cities in the US?
Which are the two cities in the US with the most restrictive zoning and building codes? Which are the two most expensive cities in the US?
Arlington VA added a ton of mid-rise and high-rise development in the 80s/90s (Clarendon, Courthouse, Ballston) with good results. Reston VA is now doing the same and home that were $400k a decade ago are selling for nearly double. In both cases, the scope of in-fill was tightly controlled - the local governments created dense zones around the Metro stations - which makes sense.
This is a common source of confusion. If you relax zoning, the instantaneous effect is that the value of land is increased in the areas where the zoning is made more permissive (because now you can do more with the land) and reduced where zoning was already permissive (because now land zoned for those uses is less scarce).
So the immediate effect is that the value of the land where single-family homes are will go up, and that effect is permanent. Then the value of indoor space goes down as people are allowed to and do build more of it. Which is what we're after, because the important thing is that people can get a couple thousand sq ft of living space for an affordable price, regardless of the number of floors the building they're in has. If the value of the land doubles but the amount of housing per acre quintuples, you've still made the land cost per housing unit decline by 250%.
What that does to the long-term value of existing single-family homes depends on how much demand there is. In areas where the zoning had been oppressively restrictive, the gain in land value from relaxing the zoning could exceed the amount that it's even possible to lose from the value of indoor space.
For example, you have a $400,000 house and the new zoning makes the land itself worth $500,000, but housing still becomes more affordable because someone can buy that $500,000 piece of land, build a dozen housing units there and sell the property for $2.4M, pricing each of the units at $200,000. The existing single family home still sells for $500,000+ though, not because it's a housing unit but because it's an acre of land where someone could build a dozen.
> In both cases, the scope of in-fill was tightly controlled - the local governments created dense zones around the Metro stations - which makes sense.
It doesn't. It makes a significant amount of sense to put a 3-5 story building in an area with mostly single-family homes, because that can often be supported with no additional infrastructure. If you want to put them all in the same place then you have to upgrade the utilities and transportation infrastructure there. Which is fine if that's what happens naturally because people choose to develop a particular area, but there is no reason to force it.
And trying to concentrate them all in one area also slows down the construction and makes it cost more, because then the land cost there remains high for lack of alternatives, forcing developers to buy out those particular people (and giving them undue gains that should have been evenly distributed to the community at large) instead of buying whatever properties were already on the market.
If builders were selling units at their cost plus a fixed profit margin, you're right.
But builders are in the business of maximizing profit. If in this area there was a backlog of people willing to pay 400K for a housing unit, the builder will sell each of these dozen units for 400K.
If you feel I'm wrong, please point me at a few cities in the US where housing became cheaper as density increased.
Or people wouldn't pay quite as much as the existing single-family homes, but they'd still pay more than the builder's costs to construct the building, so it's still profitable. It only becomes not profitable when the price of a housing unit drops below the builder's costs. Which is why they'd want to keep building until that happens.
> If you feel I'm wrong, please point me at a few cities in the US where housing became cheaper as density increased.
I keep saying this. The causation goes the other way. Density increases where demand is higher. Higher demand increases prices, then density blunts the increase so that it's smaller than it would be if no additional construction occurred. That doesn't mean it would become cheaper in the higher demand area than the lower demand area -- particularly if the area is still growing, in which case construction lags the increase in demand and has to catch up. And if construction is inhibited, it can't, so you get as much density as the zoning allows and then prices skyrocket because supply can no longer respond to increases in demand. Which can happen even if the area already has above-average density.
What you'd need to see a price decrease is an area where zoning was relaxed after being strict, allowing density to increase after it had previously been suppressed, so that density increases because of the legal change rather than a new increase in demand. But the lack of this happening in areas of high demand is exactly the problem. Maybe we should try it and see what happens?
I believe it's a feedback loop really, not exclusively in one direction or the other.
You're right that nobody wants to build a skyscraper out in the farmlands. They build it in the city, which makes that city a tiny bit more dense, pulls in more people and makes the city a small increment more attractive and pulls in even more people which in turn raises prices even more and the cycle continues. As cities get more dense they always get more expensive.
Look at Manhattan, full of highrises and skyscrapers as far as the eye can see, and is the #1 most expensive area in the US to live.
> Which are the two cities in the US with the most restrictive zoning and building codes?
I have no idea to be honest. Do you have info on that?
This is only a feedback loop if the city becomes more attractive by a larger amount than the increase in housing. For example, if you add new housing for 1000 people and this causes 1200 new people to move to the city. If adding new housing for 1000 people causes 800 new people to move to the city, you've still created more new housing for locals than you had before, and can just keep at it.
We can use basic math to see that even if the former ever happens, it can't sustainably happen. If e.g. the Chicago Metro area had the population density of Manhattan, it would have a larger population than the United States as a whole, 750M vs. 333M. A single metropolitan area could provide more housing than there are people to live in it. That many people would have to exist before they could move there. And obviously the entire US population is not going to move to a single metro area anyway, so you would reach saturation long before that.
Now, one thing that does happen is that tall buildings are more expensive to build, so when you get into the extremely high densities like in Manhattan, part of the high cost is the construction cost for a building that tall. But Manhattan is an outlier. If we're talking about converting single family homes into 12-unit condos, the cost per unit isn't much different and you get a 12X increase in density.
And here's another piece of the puzzle: Where are these people coming from? If you build new housing in New York, that might attract people who used to live in Boston. If you build new housing in Boston, that might attract people who used to live in Seattle. What happens if you build new housing in New York and Boston and Seattle and Los Angels etc.? If people move from everywhere to everywhere, there is no net migration and no increase in local demand. And even if you did only increase the amount of housing in New York, and that caused people to move to New York from Boston and Seattle and Los Angeles, then housing would become more affordable in Boston and Seattle and Los Angeles because the local demand there would go down. So you're talking about a hyper-local effect, whereas on a question of national policy, supply and demand is in full effect and "build more housing" makes the average unit cost less even if the ones that get more affordable aren't in the same place as the ones you build.
> I have no idea to be honest. Do you have info on that?
It's the same two cities. It's New York and San Francisco. They both make it extremely arduous to add new housing. Many of the existing buildings in Manhattan would be illegal to build today. SF is even worse.
The math checks out on paper, the arguments make sense.
And yet, whenever I see a very dense city, I also see an extremely expensive place to live.
When observations contradict theory, I have to think there's some flaw in the theory.
Suppose you have an area with 10,000 acres of land and local demand for 250 housing units, i.e. a rural location. Housing here is going to be inexpensive; land is not scarce so people will build a hundred single family homes on a hundred acres of land, there will still be 9750 acres available and the main cost will be the cost of constructing a house.
Now suppose you have the same 10,000 acres of land but local demand for 25,000 housing units. In the absence of zoning restrictions, this will still result in somewhat higher prices. Many of the lots will need more than one housing unit, these may cost a little more to build, and the cost of the land will be a little higher because there is more demand. In the presence of zoning restrictions requiring only single family homes on an acre of land, you're completely screwed, because the maximum number of units you can build is 10,000 and you have demand for 25,000. We see the latter in practice in many places, e.g. places where any single family home with an acre of land is selling for more than a million dollars.
Now suppose you have the same 10,000 acres of land but local demand for 250,000 housing units. It is not only not possible to satisfy this demand without multi-story buildings, it will take tall ones at that. Without any restrictions that is what would happen, and prices would be a little higher for the same reasons, but still be kept in check by new construction. However, if zoning only allows 125,000 units to be built, you would have a significant amount of density and still fall well short of the 250,000 you need, causing prices to be much higher.
The prices in the area where you need 25 units per acre are always going to be higher, but there is a difference between "higher because it costs somewhat more to construct a 10 story building" and "San Francisco prices".
Moreover, there is another reason why the highest density areas have such high prices. Common zoning practice is to assign some very small area where the tallest buildings can be constructed, e.g. here's the NYC zoning site:
Have a look at what proportion of the land is zoned R8-R10, allowing for the sort of densities seen in Manhattan. It's only allowed where it already is. It's not even allowed in all of Manhattan, which itself is less than 1% of the New York Metro area. So if you want to build a very tall building, it has to be in that tiny sliver in Manhattan, causing the price of land with that zoning to be astronomical because it's so artificially scarce, and adding that cost to the price of any housing there. You keep asking why housing costs so much in Manhattan -- that's why, because you're not allowed to build those buildings somewhere else, and the price of land operates on supply and demand. What's happening is exactly as the theory predicts: Artificial scarcity raises prices.
Or to ask your question in reverse, show me a place with a large amount of land area under permissive zoning that has higher housing costs than NYC or SF.
Ultimately, the question I find most interesting is: Which cities have achieved very high density while maintaining a low cost of living (relative to local salaries)?
We can debate "If only they did ..." forever, but if there are no actual cities where this has ever happened, maybe the answer is that it can't happen.
Density seems to always make an area more expensive. What are the two densest cities in the US? Which are the two most expensive cities in the US?
Increasing density does not ever seem to make an area cheaper to live. Maybe that is ok, but it does contradict the idea of building more to make it cheaper.