YIMBYs are right that the US needs a major expansion of its housing supply. Unfortunately, eliminating restrictions on private housing development won’t do much to get us there.
The Currie Commons housing development under construction in the Harrison neighborhood of Minneapolis, Minnesota, US, July 20, 2023. (Ben Brewer / Bloomberg via Getty Images)
Twenty years ago, “the hot-button issue of land use” was a phrase that could only have been uttered as a joke. Today, it would barely raise an eyebrow. Jacobin contributor Ross Barkan, writing in New York magazine last year, described the “YIMBY war breaking out on the Left,” in which “progressives and outright socialists find themselves on all sides,” as one of the fiercest ongoing. Even for those most inured to online Sturm und Drang, the sheer vitriol that’s unleashed these days at the first breath of terms like “zoning,” “density,” or “YIMBY/NIMBY” can be jarring.
Lately the fight has become a war of movement. The center of gravity of left opinion has shifted rapidly and unambiguously: away from what used to be a widespread, rather unreflective antidevelopment position (driven partly, as Barkan put it, by a desire to “keep bucolic, low-lying neighborhoods as they are”) to a much more intellectually considered support for pro-density public policy.
In the process, the Left has become more receptive to certain ideas once primarily associated with economists and policy writers, like Ed Glaeser or Matt Yglesias, who could — accurately and nonpejoratively — be classed as “neoliberals.” According to their line of thinking, the main obstacle to plentiful and affordable big-city housing is the restrictiveness of metro-area land-use policies. These policies, which favor single-family, low-rise homes, severely curtail the construction of apartment buildings and other density-friendly types of development. Eliminate the restrictions, these voices insist, and more affordable housing will proliferate.
There are plenty of good things to say about this shift. It reflects a greater savviness on the Left about urban policy and a heightened attentiveness to evidence and debates from the social sciences. It’s led activists and writers to push for a range of worthy pro-density policies, like an end to the “apartment bans” that perpetuate the soft apartheid of the heavily zoned surburbs. And it should go without saying that just because an idea is espoused by some people with the wrong politics doesn’t make the idea itself wrong.
However, the shift has also led to unrealistic expectations about how far cities can be changed for the better by “liberating supply,” due to a too-ready acceptance of models that try to adapt Econ 101 thinking (of questionable relevance even in the best of circumstances) to the sui generis problem of urban housing.
What Upzoning Actually Accomplishes
In the Econ 101–inspired picture of housing markets, the problem of housing scarcity is almost trivially simple: local metro-area governments have made it illegal to build more than a certain number of housing units on each section of urban land; this cap on supply, combined with rising demand, results in a bidding up of the price of the “product,” just as you’d expect in any “normal” industry. Lift the cap, and market incentives will send new housing supply rushing in.
But there’s a problem with this logic: it glosses over the critical role of land.
Urban land, whose value accounts for about 80 percent of the geographic variation in residential property prices, is what makes housing fundamentally different from other sectors of the economy. It’s unique among production inputs, for at least two reasons. For one thing, unlike machine tools or office supplies, it’s a speculative asset; its value fluctuates according to investors’ shifting guesses about future developments. For another, it is inescapably monopolistic: the land in a given location of a city is the only land there can be in that location.
In “normal” industries, the cost of production is driven by productivity: the more output can be squeezed out of a given amount of labor and capital, the less the product costs. And that logic does, in fact, apply to the nonland portion of housing development costs (most obviously construction costs — but also, importantly, lawyers and red tape).
Unlike machine tools or office supplies, land is a speculative asset; its value fluctuates according to investors’ shifting guesses about future developments.
But the cost of land itself is determined in an entirely different way: not by productivity but by how much rent the land is expected to yield. Just as a company’s stock price reflects investors’ beliefs about its future profits, the price of a parcel of land reflects investor beliefs about the parcel’s future rent-generating capacity.
The first point to note, then, is that when a city “upzones” — that is, when it allows denser development by lifting the cap on the number and size of housing units that can be built on a given piece of land — the price of land actually goes up, which makes it more expensive, all else equal, to build housing there.
Some may find this paradoxical: How can eliminating a restriction on the supply of something make it more expensive? The logic should be clear enough, but anyone not satisfied by mere a priori reasoning can find a wealth of empirical evidence in the spring 2020 issue of Valuation, the journal of the professional society of real estate appraisers, which features an article — snazzily headlined “In the Zone” — on exactly this question: “Many cities see upzoning as an effective way to increase density and affordable housing,” it begins, “but how does it affect property values and land use?”
The piece is full of quotes from real estate professionals around the country, all of whom say basically the same thing: “If there is the prospect of upzoning, investors will start sniffing around and land values will increase as sellers start doing the math on a larger bulk allowance.” (That was “Theresa M. Nygard, senior vice president at KTR Real Estate Advisors in New York”; a “larger bulk allowance” means more permitted floor space per land plot.)
So if upzoning increases the price of land, and if land is the decisive determinant of housing costs, does that mean upzoning — touted as a way to make housing cheaper — actually makes it more expensive?
No, not necessarily. But in order to work as intended, upzoning needs to clear a high hurdle: it needs to result in the number of homes per plot of land expanding by more, proportionally, than land prices increase. If it doesn’t, then housing is unlikely to get any cheaper.
A YIMBY Blind Spot
Let’s take a concrete example. Right now, a downtown apartment in a superstar city is an insanely expensive commodity, something many people would love to have but can’t afford. If you view housing as a more or less normal “product” — and land as a more or less normal production input — the mystery is why this hasn’t prompted enterprising capitalists to swoop in and offer a competing product at a more attractive price.
Why not find a dilapidated two-story building somewhere in town, buy it from the current owner, tear it down, and build a four-story building in its place? Then you can spread the cost of the land over twice as many apartments, offer each apartment for rent at somewhat less than the current going rates in town — say, 10 percent less — and be flooded with applications from eager would-be tenants. Everybody wins: apartment-seekers get lower rent, you make a huge profit, and the previous landowner gets to sell his dilapidated property at a price that surely exceeds what it would have been worth in the absence of redevelopment. The only losers, presumably, are the incumbent landlords in town, who will now be forced to offer their own tenants lower rents to stay competitive.
In order to work as intended, upzoning needs to clear a high hurdle: it needs to result in the number of homes per plot of land expanding by more, proportionally, than land prices increase.
Given how costly housing is these days, and given how everybody wins in this scenario, surely this sort of business should be going on all around us, at a massive scale. If it’s not, there must be something afoot that’s stopping it from happening: zoning regulations. Remove the regulations, and this kind of redevelopment will proliferate. Or so many people think.
But if you look at land as a speculative investment, rather than as a normal production input, the whole issue appears in a rather different light.
From this perspective, the mystery is why the owner of this land would want to sell it for a price that reflects rents that are lower, per apartment, than the current market level. The idea of “spreading the cost of the land over twice as many apartments” may sound clever to you. But from the owner’s point of view it amounts to a crude form of “shrinkflation.” Presumably, apartment rents in town have been rising — year after year, for many years; that’s exactly the problem upzoning is supposed to solve. Why should the owner believe that trend has suddenly gone into reverse?
At some point the owner will presumably engage the services of a professional real estate appraiser to advise them about how much the site might be worth, and why. The appraiser’s job will be to estimate its value at its “highest and best use,” in the jargon of the field, using data on recent rents for comparable apartments in the neighborhood, as well as information on current zoning limits. (Recall the quote from the real estate executive in Valuation magazine: in the wake of an upzoning “land values will increase as sellers start doing the math.”)
Imagine how foolish this owner would feel if they did sell the property to you now at the price you’re offering — a price that assumes lower rents — only to read in the newspaper a year from now that average rents in the neighborhood have, yet again, risen at a double-digit clip. You’re probably not the first developer who’s called to inquire about the property; you probably won’t be the last. Surely the best option for this owner is to politely decline your offer and wait for a better one.
Astute readers might recognize this scenario as conveying the insights formalized in “real options theory,” a mathematically abstruse branch of the finance literature that stresses the value of waiting, rather than going ahead with an investment, whenever the investment would be costly to reverse and the future is unknowable. Just a few months ago, a theoretical paper by a pair of Dutch economists applied real options theory to the problem of urban land use, and while I wouldn’t recommend the paper to anyone lacking an appetite for heavy jargon and glowering equations, its authors mercifully summarize their conclusion right in the paper’s title: “The option value of vacant land: Don’t build when demand for housing is booming” (my emphasis).
At a basic level, though, it’s a simple point: the advocates of upzoning say it will lead to a house-building boom and a consequent fall in rents. But in order for that to happen, investors must already believe it will lead to a housebuilding boom and a consequent fall in rents. Why should they? On what grounds? This is the blank spot at the heart of the YIMBY case.
When Land Management Is Private
YIMBY economics must, then, be based on a kind of circular reasoning: upzoning causes rents to fall because rents are expected to fall, due to the fall in rents.
But isn’t there something circular about my argument, too? After all, what I’m saying is that rents are determined by land values. But I’m also saying that land values are determined by expected rents. What, ultimately, determines both?
That question is easily answered. Rents and land values in a particular location are determined by the income and wealth of the people and businesses in and around that location. Put plainly, it’s the presence of rich people that makes rents expensive: they can pay more. This was the valuable kernel of insight embedded in the old “unreflective” left attitude toward urban development; its replacement by the new focus on supply — however valuable that focus might be in its own right — represents knowledge that’s been, if not lost, then at least injudiciously devalued.
If we want to know why housing has gotten so expensive, it helps to look at which locations have seen housing costs rise the fastest and which have seen costs rise more slowly, and to figure out what accounts for the difference. That’s the purpose of the chart below, which is based on data for about seventy-two thousand Census tracts: very small geographic units averaging about four thousand residents each, each of which can be thought of as a “neighborhood.”
The chart groups the seventy-two thousand tracts into eight “bins” along the horizontal axis, sorted according to the percentage change in median housing costs they experienced between 2012 and 2019; the bins range from -10 percent to +30 percent. (All data in the chart are from the Census.)
What we want to know is: What kinds of characteristics determine whether a neighborhood ended up in one of the far-right bins, with skyrocketing housing costs; and which neighborhoods ended up in the far-left bins, with only modest increases, or even outright declines, in housing costs?
Start with the yellow line on the chart. It shows how the median tract within each bin performed in terms of building more housing units: each dot displays the national percentile rank of the median tract in that bin in terms of housing-stock growth.
Did the neighborhoods experiencing less housing cost growth build lots of new homes? There’s a chart that’s been making the rounds on Twitter/X lately purporting to show (in a different way, using different data) that they did. But on my chart, you can barely see a difference: the median rankings of the eight housing-cost-growth bins range from the forty-sixth percentile to the forty-ninth percentile of housing-stock growth. So that doesn’t appear to be a major factor.
But looking at the rate of growth of housing units doesn’t necessarily settle the issue, because what we really want to know about, specifically, is the effect of zoning regulations. Luckily enough, there’s data on that. A team of economists at the Wharton School maintains an index of local zoning restrictiveness, called the Wharton Residential Land Use Regulation Index (WRLURI), based on a detailed survey of municipal zoning authorities.
That’s what the blue line on the chart shows. Like the yellow line, it displays the percentile rank of the median tract of each of the housing-cost-growth bins — but this time in terms of its score on the Wharton index. (Strictly speaking, it’s the score for the municipality to which the tract belongs.)
Did neighborhoods that experienced the fastest housing-cost growth have more restrictive zoning? Again, the difference is just barely perceptible: between the best and the worst housing-cost-growth performers, the range of median rankings on the Wharton index runs from the forty-sixth to the fifty-second percentile.
Now we come to local income growth, represented by the green line. Here there’s no need to squint at the chart. Every single bin in the housing-cost distribution experienced faster growth in median household income than the bin before it. Across the eight housing-cost-growth bins, the national income-growth rankings of the median tract range from the thirty-third to the sixty-fourth percentile, a remarkable spread. No other variable comes close in determining how much costlier housing has grown across different localities.
That housing is invariably expensive in places where rich people live and cheaper where they don’t is a familiar fact of life — so much so that it may be tempting to shrug at these data and say they’re just stating the obvious: plus ça change. But this, as they say on MSNBC, is not normal.
If it’s the speculative nature of urban land that stymies housing supply, it’s the monopolistic nature of it that gives its owners the power to extract so much wealth.
Other products and industries do not have prices that rise in lockstep with local incomes, hoovering up the fruits of all the other industries’ productivity gains. According to the McDonald’s app, a Big Mac costs $5.79 in the impoverished South Bronx and only sixty cents more ($6.39) on the gilded Upper East Side. By contrast, when it comes to housing, a recent study by the MIT labor economist David Autor finds that “local housing costs at least fully offset local pay premiums, implying that workers who move to [better-paying areas] have no higher net-of-housing consumption” (emphasis added).
If it’s the speculative nature of urban land that stymies housing supply, it’s the monopolistic nature of it that gives its owners the power to extract so much wealth. This is not a monopoly conferred by restrictive zoning; it’s inherent in the phenomenon of urbanization. The capitalization of mobile wealth into land values happens because location matters: urban land in one place is not a perfect substitute for land in another place (it’s barely a substitute at all). That’s what “density” is all about: agglomeration economies, network externalities, production synergies, and so on. These forces make economic activity more remunerative in a given place; but by the same token they raise the cost of switching to a different place. That’s what gives urban land its monopolistic power.
When the management and ownership of land is left to the free play of private speculation and investment, it creates a trilemma between density, affordability, and inequality. When inequality is high, you can have affordability but not with density (viz. the Sunbelt); or you can have density without affordability (Manhattan, Boston, etc). I’ll have more to say about what is to be done in a subsequent article.