A quick look at sky-high housing prices in San Francisco and Manhattan makes it clear that land-use restrictions make places more expensive—a pattern that has been documented in a large body of research. Affluent residents of these places put pressure on local politicians to limit development, in order to protect the character of their neighborhoods and property values. As I highlighted in The New Urban Crisis, the most expensive—and the most unequal and segregated—places in America today are also the most liberal and progressive. They are the places run by progressive mayors like Bill de Blasio, London Breed, and Eric Garcetti. They are also the places that gave the largest shares of votes to Hillary Clinton in the 2016 election.
A new study published in the journal Political Geography shows that places with strict land-use regulations become more Democratic over time. The study, by political scientist Jason Sorens of Dartmouth College, finds that land-use restrictions gradually tilt places leftward—not just by attracting more highly-educated Democrats, but even more so by repelling non-college-educated workers who have become more Republican over time. Sorens defines land-use regulations as the stringency of residential zoning based on the Wharton Residential Land Use Regulatory Index (WRLURI), a measure that is commonly used in such studies.
Sorens examines the connection between land-use restrictions and voting patterns across states, counties, and subdivisions of counties. He looks specifically at the effects of land-use restrictions, as well as housing prices, housing cost-to-income ratios, and cost of living, on partisan voting.
Looking at presidential voting patterns by state level for the period 1984 to 2016, Sorens finds clear evidence that more restrictive places tilt Democratic over time, controlling for other factors. An increase in housing restrictions and the related increase in housing cost cause a rise in Democratic vote share in the subsequent four-year period.
This Democratic tilt is not a result of higher incomes—the effect of income in Sorens’ model is either negative or insignificant. Rather, it is an effect of housing restrictions. As he puts it, “[T]here is little evidence that richer states per se tended to move toward the Democrats over these periods. States with more costly housing, due largely to restrictions on supply, have become more Democratic.”
He finds the same basic pattern across counties. At the county level, two things stand out. One, more restrictive counties tend to shed Republicans and become more Democratic over time—a pattern that seems to have accelerated since 2012. And two, more restrictive counties tend to gain more college-educated people and shed fewer educated ones, which contributes to their partisan shift. As Sorens notes, college education may causally mediate a substantial part of the relationship between zoning and partisanship at the county level. In other words, a county’s becoming more educated as a result of building restrictions makes it more Democratic.
He also finds clear evidence of this pattern across county subdivisions. In the study, he considers county subdivisions in New England. In the early 2000s, there was only a weak connection between land-use restrictions and partisanship. Democratic places were only slightly more regulated than Republican ones. Interestingly enough, many towns that started out more Republican actually became more restrictive over time. This, for Sorens, is a key piece of evidence: “These data support the central claim of this paper: Democrats do not cause stricter zoning, but stricter zoning causes more Democrats (relative to Republicans),” he writes. In other words, when Republican towns increase land-use restrictions, they tend to drive away more Republicans.
Ultimately, then, it is not that Democratic places simply impose more land-use restrictions. It is that places with more land-use restrictions tend to attract more Democrats and shed more Republicans. “Do places with more Democrats enact stricter housing regulations?” Sorens wrote on Twitter, summarizing the key results of his study. “No.” Land-use restrictions tilt places to the Democratic side of the aisle. Our tendency to sort by education is the key mediating variable. It’s not just that expensive places attract more and more highly educated Democrats, it’s more the case that they repel less educated (non-college-educated) workers who have become more Republican over time.
Sorens adds an important new wrinkle to our understanding of America’s deepening political divide. We know all too well that Americans sort by education, by density, by size of city, and across urban, rural, and suburban areas. And we increasingly understand that Americans sort by homeownership and by use of the car. This study unveils an additional dimension to our sorting: the way that land-use restrictions help turn places more liberal, by making them more expensive and shutting out less advantaged working-class voters who increasingly vote Republican. It further fuels the politics of resentment that increasingly define, and perhaps imperil, our nation.