How much should we blame planning for the degree to which cities sprawl? As much time as we (justifiably) spend here on this blog explaining how conventional U.S. planning drives excessive sprawl, it’s worth periodically remembering that, at the end of the day, the actual extent of the horizontal expansion of cities is largely outside the control of urban planning.
Consider Houston. Whenever I say anything nice about Houston’s relatively liberal approach to land-use regulation, someone invariably comments some variation of the following: “Yes, that’s all well and good in theory. But in practice, heavily regulated cities like Boston are far more urban and walkable, so maybe relaxed land-use regulations aren’t so great.”
Indeed, most of Houston is classic sprawl. But this begs the question: to what extent can urban planning policy be blamed for sprawl?
The urban economist Jan Brueckner, drawing on an extensive literature, distinguishes between the “fundamental forces” that naturally drive urban growth outward and the market failures that push this growth beyond what might occur in an appropriately regulated market. (For the purposes of this post, I’ll be using “sprawl” and “horizontal urban expansion” interchangeably. In the same paper, Brueckner thoughtfully distinguishes the two.) The latter, urban planners can address. The former, not so much.
Let’s look first at the “fundamental” variables that planners have little to no control over. Brueckner identifies three: population growth, rising income, and falling commuting costs. The first variable is obvious: as cities grow, demand for all housing goes up, and some of that housing goes out on the periphery regardless of planning policy. Metropolises like Houston, Dallas, and Atlanta are currently experiencing 2% population growth every year, meaning they are on track to double in population in the next 35 years. You would expect a lot of horizontal expansion, all else being equal!
The second variable is income. Houses are what an economist might call a normal necessity good. It’s a normal good in that, as we get richer, we spend more on it. But it’s also a necessity good, meaning that as we get richer, we spend proportionally less of our income on it. This first stage is essentially what has been happening in Sun Belt cities like Charlotte, Houston, and Orlando over the past few decades, where median incomes have in many cases doubled. With all that new demand for housing, you would expect a lot of it to go out on the periphery on greenfields, regardless of planning policy.
The third variable is commuting costs. If it’s cheap to commute long distances, people will do it. Nested in this variable is the fact of technological change and progress. The cities that are booming today are taking form in an age when most people can afford to buy a one-ton metal machine and commute alone to work at roughly 60 miles per hour. It’s easy to get far out of town, where land is cheap and even a middle class resident can afford a decent sized home on a quarter acre lot. When Boston and Philadelphia were building up and out, the big new thing in transportation was the electric streetcar, and many people still had to walk. Land values within walking distance of these transit options were unsurprisingly quite high, meaning that housing had to go up, not out. We shouldn’t expect these classes of cities to come out looking the same.
On top of Brueckner’s three variables, I’ll add two more that come up a lot in these discussions. The fourth variable is agricultural productivity. Under the standard urban model, the outer edge of a metropolis is the point at which residential, commercial, and industrial developers are unable to profitably outbid agriculture for land. That means that in regions with high agricultural productivity, you would expect to see more compact cities, and in regions with low agricultural productivity, you would expect to see more sprawling cities.
This theory has been repeatedly tested and validated. And short of running a few hundred regressions, you can intuitively work it out: in a place like Las Vegas and Phoenix, land holds no agricultural value whatsoever, so once some form of development pencils out, it happens. On the flip side, in regions of outstanding agricultural productivity like Western Oregon and Central Kentucky, rents for urban uses must be quite high before developers can outbid agricultural uses. Why do you think Lexington, Kentucky and Portland, Oregon were the first and second cities, respectively, to adopt urban growth boundaries? Because they both are home to large, power agricultural interest groups (the horse industry is quite powerful in my home state of Kentucky) and land rents were so high that the pressure to expand outward was already weak relative to other cities. Of course, as rising residential rents inch further above agricultural rents in these two cities, their growth boundaries may gradually weaken.
A fifth and final variable is landscape. No matter what all the other variables are doing, if a city is bounded by the ocean—Seattle, San Francisco, Miami—or mountains—Los Angeles, Pittsburgh, Asheville—you could expect it to quickly start building up, since it’s either very expensive or physically impossible to build out. Take a look at a random sampling of of metropolises regularly derided as sprawling—Oklahoma City, Raleigh, Columbus—and you will find that many sit on a flat, featureless plain with few barriers to horizontal growth. The effect of geographic features like this is so robust that it’s standard practice for researchers to control for them when studying topics like urban form and housing affordability.
So let’s bring it all together. Houston is a rapidly growing city, where incomes have approximately doubled since 1990 and nearly everyone can afford to commute alone by car. Agricultural productivity is not unusually high and the metropolis sits on a flat, featureless plain. It would be weird if Houston didn’t sprawl under these conditions. Indeed, it would almost be a miracle if planners and policymakers could have forced the city to do anything else.
None of this is to say that planners and policymakers have no control over sprawl. As Brueckner and others have pointed out, virtually every U.S. city under-prices the negative externalities associated with long commutes—namely, congestion and air pollution—and fails to internalize the cost of new infrastructure involved in suburban development. At the same time, most cities—even Houston and many of its suburbs—make it tough to build dense new housing in existing urban areas and require new housing on the periphery to sit on large lots along wide roads. All of this is bad and it needs to be reevaluated. To be clear, I don’t mean to diminish the harmful impact these policies! But these market and policy failures are only part of the set of variables driving urban form.
Cities are vast, complex systems, beyond the comprehension or control of any single individual or group. As the urbanist Alain Bertaud puts it, different cities should often be treated as different species entirely; Atlanta and Barcelona have about as much in common as an elephant and a mouse. They’re both mammals, sure, but try to treat them the same and the results could be messy. All of this should leave us humbled—but not incapacitated!—about the power of planning and policy to reshape cities. And it should leave urbanists less confident in their harsh moral judgement of today’s sprawling cities. You’re not going to turn Houston into Boston. But that doesn’t mean that tinkering on the edges won’t help.
For future content and discussion, follow me on Twitter at @mnolangray.