Urban Sprawl

An easy culprit to pick on for some of our urban ills is sprawl. This is the beginning of something I wrote that was printed in the University of Illinois Real Estate Newsletter and in the Wharton Real Estate Review (I will post the remainder later):

NINE CAUSES OF SPRAWL
Richard K. Green


Since the late 1990s, sprawl has become a leading public policy issue. But in the debate over sprawl, few policy makers seem interested in investigating the phenomenon’s root causes. This failure is disturbing, for without understanding sprawl’s causes, politicians and planners may make policy choices that exacerbate its effects, while voters are unable to make choices among a set of alternative outcomes. The paper that follows presents nine important causes of sprawl. The list is not exhaustive, but it explains a substantial share of the suburbanization that has characterized post-World War II United States.

I. THE RENT GRADIENT

Modern urban economics has its roots in models developed by William Alonso, Edwin Mills, and Richard Muth. These models show that two key determinants of urban land values are: the value of undeveloped land at the metropolitan area edge; and the cost of transportation. Put simply, it is desirable to be near the hub of commercial activity, so people who live near job centers pay more for land and those living near the periphery pay less.
Where land is relatively expensive, it makes sense to economize its use. Therefore, building densities will be high in places with high land values. For that reason, the densest developments tend to be near city centers. Conversely, the land on cities’ peripheries is relatively inexpensive, so building densities are low. A result is that at the relatively less expensive urban periphery, each home takes up more land (houses sit on larger lots), and therefore these areas generally exhibit lower density (that is, more sprawl) than do city centers. This tendency is not necessarily a problem, because relative prices reflect the relative scarcity of resources. After all, construction components such as timber and labor are scarce resources, too. The question is whether the relative prices of land and improvements appropriately take into account all resource costs.

II. DEMOGRAPHIC CHANGES

Land use is driven partly by the composition of households. Since the end of World War II, Americans are: waiting longer to marry; more likely to divorce; having fewer children; and living much longer after their children have grown. These demographic changes have caused the average U.S. household size to fall from 3.5 persons in 1940, to 2.5 persons today. Thus, even if population had remained constant and housing density had remained unchanged, the amount of land required to house the population would have risen by 40 percent between 1940 and today.
It might be argued that smaller households could occupy smaller housing units, resulting in higher unit density. However, accomplishing this increase in density would not be easy. Consider an older house in the middle of a city. The house is in good condition, but not easy to subdivide into multiple units, so it is occupied by (on average) 2.5 people, rather than 3.5. Multifamily units would likewise be difficult, if not impossible, to resize for smaller households. Until consumer tastes change and the existing housing stock is replaced, smaller household size will not, by itself, reduce the amount of land per capita needed for housing our population. This outcome is, arguably, benign.

III. GROWING AFFLUENCE

The U.S. has become a more affluent nation. Households at all economic strata are materially better off now than they were in 1945 (or in 1965, for that matter). At the top end of the spectrum, the number of households earning in excess of $100,000 in real income has increased sixfold over the last ten years alone.
The effect that affluence has had on land use is profound from two perspectives: an income effect and a substitution effect. The income effect could be called the “George Carlin effect,” that is, as people have more money, they want more “stuff,” including larger houses and larger lots. The substitution effect chiefly concerns transportation. For example, if a large share of household income is spent on transportation, a family might want to economize, or better still avoid its cost altogether, by living in the center of a city. Conversely, as transportation becomes relatively less costly, people might seek to economize on land cost, and will therefore move to places (the suburbs or exurbs) where the per unit price of land is low.
In the context of contemporary society, transportation costs are largely fixed; while acquiring and maintaining a car are costly, the cost per mile for driving is small (though this may ultimately change as gas prices keep rising). Therefore, while transportation concerns dictate location decisions for low-earners, transportation costs are relatively unimportant for high-earners. Thus, for high-earners, who are the chief market for new suburban housing, rising affluence leads to greater land consumption.

IV. TRANSPORTATION

According to Joel Garreau’s Edge City, “the maximum desirable commute, throughout human history, regardless of transportation technology is forty-five minutes.” When people in ancient times walked to work, cities could extend only a mile or two. Omnibuses and streetcars increased this distance. With the use of private automobiles, and average driving speeds of 30 miles per hour, people today can live ten miles or more from their workplace. Garreau demonstrates that development has responded to the desired 45-minute commute by placing office space on the periphery of cities, so that commuters can avoid the density and traffic of downtowns. As a result, the amount of office space in suburbs is now generally greater than the amount in traditional downtowns, and the most common commuting pattern in America today is from suburban home to suburban office. Households’ desire to reap the private benefits of low density while avoiding lengthy commutes has pushed cities to spread out.
This result would all be ideal if households bore the full costs of their commuting. But they do not. For example, the City of Milwaukee’s Department of Administration calculates that automobiles in that city cost society about $400 apiece each year beyond what their owners pay in licensing fees and gasoline taxes. These added costs relate primarily to congestion and pollution. What is remarkable about the $400 figure is that Milwaukee has relatively little congestion, as large American cities go. In many cities, the costs are certainly much higher.
One method for making commuters pay these costs is to increase fuel taxes. For instance, someone who drives 10,000 miles per year in a car that gets 25 miles per gallon uses 400 gallons of gasoline per year. Increasing the gasoline tax by $1 per gallon would internalize the costs of congestion. But according to this argument, in places such as rural areas where congestion costs are much lower, gasoline taxes should also be lower. This would give metropolitan area residents an incentive to drive outside their cities to buy their gasoline. An important question remains as to whether people’s time is sufficiently inexpensive that it is worth their while to go out of their way to buy gas.
Other potential methods for internalizing commuting costs include highway tolls and commuter taxes. It could be argued that all such taxes have strengths and weaknesses, and that all are politically difficult to implement. But the consequences of underpricing the social cost of automobile use have also led to politically unacceptable outcomes.