Can Transit Tame Sprawl?

Anthony Downs
Anthony Downs Former Brookings Expert

January 1, 2002

Housing and transportation costs are inversely related. In outlying settlements, households with more than one worker usually need two or more vehicles to get everyone to their jobs. That increases their transportation costs. Because their homes tend to be built on lower-cost rural land, however, their housing costs are relatively low. Conversely, in denser urban areas that are well served by public transit, households with two or more workers can get by with one vehicle—or even none. For these households, transportation costs are likely to be a lower percentage of their budget than their rural counterparts. But urban land is usually valuable real estate, so their housing costs are likely to be steeper.

Data on consumer spending in specific regions backs up these observations. The highest fraction of consumer spending on transportation was 22.8 percent in the Houston area, which is
notorious for traffic congestion. The lowest was 15.3 percent in the New York City area, which has an extensive public transit system. But Houston households spent only 30.6 percent on housing, whereas New York households spent 36.9 percent. Combined, both types of spending were about the same in the two cities. Similar inverse relationships between housing and transportation spending prevailed in all of the 28 metropolitan regions for which data were available.

Another piece of evidence that partially supports this relationship comes from “The Costs of Sprawl 2000,” a recent study conducted by Rutgers University, the Brookings Institution and several other organizations. Part of the research examined how housing prices vary with distance from the regional downtown of each metropolitan area. Although only a few areas were analyzed, the study showed consistently that prices of similar homes tended to decline about 1.2 to 1.5 percent per additional mile from the regional downtown, except where proximity to the ocean had more influence on prices—as in Southern California. Meanwhile, longer-distance commutes added to fuel and travel-time costs by about the same amount per mile in every region. The study also found that per-mile housing-cost savings from added commuting distance were much larger in regions with absolutely very high housing costs than in those with absolutely low housing costs. Therefore, it was more likely to be economically worthwhile for households to move farther out to gain cheaper housing in high-housing-cost regions such as the San Francisco Bay and Boston areas than in low-housing-cost areas.

Different advocacy groups derive opposite policy conclusions from these findings. Supporters of more public transit claim that greatly improving a region’s transit services and encouraging more compact future growth can reduce the total of both housing and movement costs, even if higher densities raise land prices and thus cause higher housing costs: Higher housing costs would be offset by lower transportation costs. Therefore, public policy should encourage more transit usage and limit outward growth, which would also preserve more open space.

But advocates of more road building claim that enabling households to move farther out onto cheaper land would also reduce their combined housing and movement costs, at least in regions where housing costs are high. Moreover, they argue that in regions where public transit services are limited—which is most of America—it would be impossible to get many residents to stop using their cars for most of their transportation, even if enormous sums were spent improving transit. Americans’ preference for moving in private vehicles is too strong. Therefore, building more peripheral roads even though that might lead to both further sprawl and more traffic congestion would encourage construction of relatively low-cost housing.

Which view will prevail? The biggest factor influencing a policy decision is the nation’s need to accommodate projected population growth of about 64 million residents by 2020—a 23 percent gain in population. It is hard to believe that this population bulge can be fitted into existing metropolitan settlements mainly through increasing density in already-built-up areas. Moreover, conversion of any significant fraction of American households from their almost total dependence on private vehicles to major use of public transit is highly unlikely. Therefore, at least some continued outward sprawl seems probable in most metropolitan areas experiencing growth, even though peripheral development will generate more traffic congestion. This will occur even if such areas improve their transit services considerably, as many should.

The reality is that sprawl will prevail not because it will provide lower-cost housing—although it may do that. Rather, the alternative of substantially raising densities in existing neighborhoods will be decisively rejected by NIMBY-oriented residents there. As long as power over land-use decisions and housing location remains totally in the hands of local governments, their continued support of exclusionary local zoning rules will dominate future urban policy.