I much enjoyed this article by Dr. Cameron Murray, but there are a few things I'm having a hard time with. So this post is me wrestling with the implications.
Explainer: Markets efficiently delay building feasible new homes
Specifically, while I understand that for housing, while the stock dominates, there is a unit of analysis problem. A comment was made here (Comments - Inescapable Equilibrium? - by Jeff Fong) that that Houston halving lot sizes didn't do anything for housing affordability.
Houston? Yeah, probably mostly a reallocation with the city, and nationally a reallocation for sure.
This essentially hinges on the contention that making housing cheaper in Houston just induced someone from not-Houston to move to Houston, doing nothing at all to make housing prices more affordable. Which seems counter-intuitive, but I can see it in action in Salt Lake City, which has made strenuous efforts to improve housing affordability through new construction. But locals aren't really seeing the full benefit of that, as the affordable housing (housing cost compared to income) is attracting people from all over to move there. And there is not really anything anyone can do about that.
Not to say the supply isn't helping--this article is nicely explicit on the relationship between additional supply and lower rents: 1% more supply means 1-2% lower rents, and that is really hard to lower rents, because adding 1% more supply to a metro with a million housing units requires ten thousand housing units. So, in a large metro, it requires a lot to move the needle. (Salt Lake has been doing it anyway, by building thousands of apartments). But the awkward reality is that in order to keep rents from increasing, Salt Lake is going got need to keep doing it. Which means at some point, it will run into the 'Vancouver Problem', where all the depreciated low-rent strip-malls and auto dealerships have been developed, and the only low-density development that remains are the single-family homes that make up 80% of the urbanized area. And eventually, (theoretically), Salt Lake would have the Barcelona problem, where developing a seven-story apartment requires demolishing a three-story apartment. (An issue, ironically, arising not from the geographic constraint of an island or peninsula, but from regulatory constraint of military planners).
And for whatever else, Dr. Patrick Condon has done a nice job of showing that no matter how much new density you add, single family detached homes are still nice, and people will still pay quite a bit for them, and the smaller the share of the stock they represent, the more valuable they are. I still remember a story of the last private house facing Central Park, owned by an heiress who'd lived there, life-long, and didn't care to sell at any price. Her heirs sold and immediately realized some millions in windfall.
Nothing short of a Rustbelt-style dying city is actually going to lower land values, and no one wants to achieve housing affordability that way anymore than they want to achieve weight loss through amputation.
Dr. Cameron Murray did also comment that:
Planning DOES change the spatial distribution of homes, as intended. But make the spatial distribution more efficient will make average dwellings better and hence higher value.
Which is kind of a bid deal, from a transportation perspective. The spatial distribution of buildings is extremely important from a transportation perspective. And that uniform density generates pretty much the worse possible outcomes in terms of maximizing required travel distances, limiting labor market access by non-car owners, making effective transit infeasible, losing time in traffic, and suffering air pollution.
From an economic perspective, I suppose it seems like it doesn't matter--the cost of all those externalities will be priced into location values, and people will make location choices based on how they allocate their consumption between their pleasures - clean air or a daily pack of cigarettes.
But from a planning perspective, it matters quite a lot. There is plenty of evidence of low-income folks clinging to incredibly substandard housing because of locational amenities. Exiling someone without a car to a car-dominated district effectively stands them, something most clearly seen in that people prefer being homeless* in an urban setting to marginal housing in a rural setting. It's not well known, but the majority of the homeless population experience it only temporarily, as gaps in being housed (however inadequately) and only a hard core are long-term homeless. Consequently, there are people who are willing to pay incredibly high proportions of their income toward housing, simply because they cannot meet the threshold requirements of car ownership. And the core of the many anti-blight measures that characterize our modern planning regime are simply anti-density measures, hostile design to the poorest and least well-off, and that the slum-reforming efforts describing the ills of density readily translated into a slum destroying and slum clearance mentality.
It's easy to forget that while we now associate streetcar and subway use as being working class, it was (at first) very middle class--a way to escape the central city to streetcar suburbs without suffering the transportation cost of needing to walk for hours. Which in turn enabled both lower density and higher quality housing (as landlords and developers competed on quality rather than location). Indeed, if we could costlessly teleport, and go anywhere, instantly, and location ceased to matter, competition would be purely on quality. However, the internet supplies an immediate counter-factual: in a space where we can be anywhere in an infinite space, instantaneously, we've still clustered in a small number of high-density websites, via the network effect.
People will spend their money for A or B, but how much A or B people get matters, and especially so for the marginal. Tokyo is a megalopolis, the largest city in the world, and while rents haven't decreased, the amount of space per person has, which represents a very real improvement over previous conditions.
Even if halving Houston's lost sizes didn't do anything for average rents in Houston, it still enabled a lot more people to move to Houston, and to enjoy the amenities of density. People talk endlessly of the ills of density--the smells and sounds of neighbors, of traffic congestion. But too little is said of its benefits--access to employment, services, and amenities that enrich life. There are, of course, cities that are Doing It Wrong and have managed to obtain all the ills of density with realizing many of the benefits. Partially, this is a matter of the non-linearity of costs and benefits. Ten times the density doesn't make traffic congestion 10x as bad, while ten-times the density is 100x the job opportunities. But partially it's a matter of failing to make a regime shift, not away from auto-mobility, but away from auto-dependence. The automobile works at basically any scale--if you've the time, you can drive across Eurasia. It doesn't get less efficient at smaller distances (you can still drive across the parking lot), but other (slower) modes become more competitive. After all, humans don't measure travel in distance, but in time. A 30m trip is a 30m trip, regardless of if that trip covers 5 miles or 10, all else equal.
Growing up in suburbia, walking wasn't really feasible--a 10-minute walk didn't get you anywhere you wanted to be. 20 minutes would do the job, but a 20-minute trip is really a pair of 20-minute trips--one outbound and one inbound. Regardless, the empirical evidence is clear--the relationship between trip distance and frequency is inversely proportional--a trip that is twice as long is made half as often. Longer trips mean fewer trips, which means less access, which means less interaction, which means less employment, entertainment, and interaction.
As traffic congestion slows the car, alternative modes become more attractive. Hmm. I think it is important to think of parking as the 'access mode' to the car, in much the same way walking is the access mode to a bus stop. Where there isn't a bus stop, you can't embark on a bus. But likewise, where there isn't parking, you can't disembark from your car. And that is a significant limitation on car use. Of course, where parking is sparse, people can (and do) disembark from a travel time, further slowing other cars, and again eroding the value of automobility.
That enormous effort is spent to maintain auto accessibility should come as no surprise. The highway network is a network, and reduced access to even a single node degrades the whole network. And if the highway network were to lose access to a high-density location, such as a CBD, the impact to network value would be enormous. Which will make the effects of congestion pricing in New York very interesting. And interestingly implies that part of the reason that auto congestion has fallen is that auto access to the congestion zone is suddenly a lot less valuable. Which will have knock-on effects to the value of parking in the zone, which will in turn affect parking supply, in turn affecting auto accessibility.
Amsterdam an interesting example--a rational understanding that land is limited and that creating more of it (by reclaiming it from the sea) is expensive, and so a space expensive project like a highway system has opportunity costs that other places don't face. A national macrocosm of the problem faced by every city in the world.