There are two big theories about who controls the pace of development in American cities and suburbs.
One is the “growth machine.” In this telling, developed by academics like Harvey Molotch in the 1970s, urban elected officials and zoning boards are highly influenced by coalitions of business and civic leaders interested mainly in economic growth and maximizing the price of the land they own.
The other, developed later by the economist William Fischel, is the “homevoter hypothesis.” Fischel argues that real power—at least in the small to moderately-sized municipalities in which the majority of Americans live—is held by homeowners, who are also interested primarily in maximizing the value of their property: their homes.
These two theories closely track two of the major camps in the debate about what’s wrong with American housing policy. If you believe in the growth machine, either because you’re a reader of Molotch or it just happens to coincide with your general worldview, you’ll probably believe that US cities suffer from too much development, pushed on an unwilling populace by a profit-driven elite for whom zoning and planning is an inconvenience at most.
If you’re in the homevoter camp, conversely, you’re likely to think that the problem is too little development, as NIMBY homeowners scare local elected officials into blocking any housing development that might compromise their property values—either simply by increasing the housing stock, and thus the number of “competing” sellers, or by introducing “undesirable” kinds of people or buildings.
As you can see, there’s quite a bit riding on which of these theories is right—or, more realistically, in what proportions, where, and in which ways, each is right and wrong. One suggests that development laws ought to be made more restrictive; the other that they ought to be less. One suggests that there may already be a broad democratic coalition in favor of a rational housing market, but that coalition is frustrated by a smaller number of more powerful interests; the other suggests that the broad democratic coalition of homeowners, at least at a hyper-local level, is getting exactly what it wants when home values spiral upwards.
A study published last year from Vicki Been, Josiah Madar, and Simon McDonnell of NYU took a crack at pitting these theories against each other to see which did a better job of explaining zoning changes in New York City from 2003 to 2009, under former mayor Michael Bloomberg.
As the authors themselves admit, their choice of place and time give the growth machine position a leg up: even Fischel, who coined the “homevoter” name, suggests that his theory probably applies best to smaller municipalities, without the huge business interests of major cities. It’s also most likely to matter in places where most people own their own homes—and New York City has one of the lowest homeownership rates in the country, with just 36 percent of homes being owner-occupied. Finally, the billionaire businessman Michael Bloomberg is, at least superficially, as good an avatar of the “growth machine” coalition as one can imagine.
To distinguish between the two theories, the study compared “upzones” (a zoning change that increases allowed density) and “downzones” (a zoning change that reduces allowed density) with several factors:
- First, land close to infrastructure and services, like rail stations and high-performing schools. If the growth machine theory is correct, these places should see more upzones, as developers try to take advantage of these valuable locations. If the homevoter theory is correct, these places might see more downzones, as homeowners try to protect the value of their property against other, competing homes.
- Second, market growth and rising home prices. The growth machine theory would suggest places with rising prices would also see more upzones.
- Neighborhood demographics. Growth machine theory might suggest that demographics associated with high real estate values, like wealthier and whiter neighborhoods, would see more upzones. Homevoter theory would say those places would see more downzones, as more powerful homeowners are more able to enact their anti-development orientation.
- Homeownership rates and voter turnout. Both high homeownership rates and turnout would result in more downzones, according to homevoter theory.
So what did the results looks like?
- Interestingly, proximity to high-quality infrastructure and services made land more likely to be changed in both directions—that is, land far from high-quality infrastructure and services was more likely to remain in its original zoning category. But in almost every case, proximity was especially likely to lead to more downzones. For example, parcels in high-performing school districts were 43 percent more likely than the typical parcel to be upzoned—but 392 percent more likely to be downzoned.
- Correlations with market growth were weaker—but they suggested that growing markets were associated with downzoning. Parcels in neighborhoods seeing rapid population growth were 41 percent more likely to be downzoned, for example. Parcels in neighborhoods seeing rapid home value increases were about 20 percent less likely to be upzoned, although they were also 27 percent less likely to be downzoned.
- Downzoning was very strongly correlated with whiter neighborhoods: parcels in Census tracts that were over 80 percent white were more than seven times more likely to be downzoned than parcels in tracts that were less than 20 percent white.
- Parcels in tracts with high homeownership rates were 43 percent more likely to be downzoned, and 25 percent less likely to be upzoned. Parcels in districts with high voter turnout were 230 percent more likely to be downzoned, and 53 percent less likely to be upzoned.
So what does this tell us? Well, in every case where the evidence clearly points to one theory or the other, the winner is the homevoter hypothesis. Homes near high-performing schools—sources of great value to which Fischel says homeowners pay particular attention—were overwhelmingly more likely to have their maximum density reduced, rather than increased, as developers and business interests would surely prefer. Neighborhoods with rising demand, similarly, tended to see more anti-development zoning, to the extent that a growing population tended to lead the government of New York City to allow fewer homes to be built. Very white neighborhoods were particularly likely to see density caps lowered. And homeownership and voter turnout rates had exactly the results expected by the homevoter theory. (Although the voter turnout results could, arguably, also be seen as compatible with the growth machine.)
These outcomes are all the more stunning because of where and when they took place: New York City under Michael Bloomberg, which, again, is one of the very last settings you would expect to find “homevoters” in charge of development.
Of course, none of this means that big landowners, businesses, and developers aren’t also trying to influence the development process in their favor, and sometimes succeeding. Perhaps the most fascinating results of the study actually suggest a somewhat more complex dynamic: if land near parks and high-performing schools are more likely to be both upzoned and downzoned, then you can imagine a story in which these amenities lead to increased competition between the growth machine and homevoters. According to this report, homevoters appear to win the vast majority of the time—but developers also score a few victories, getting more upzones than are granted (or, perhaps, requested) in less valuable territory.
As we wrote in our review of William Fischel’s new book, Zoning Rules!, and contra Ilya Somin’s argument about a growing pro-zoning-reform consensus, these findings also suggest that those who would like to moderate home price increases through smarter development policy may have a daunting political hill to climb. More about that in an upcoming post.