Stop Trying to Get Workers Out of Their Cars
"Smart growth" is dumb about commuting.
If you hate urban sprawl, you're probably familiar with the complaints of the "smart growth" movement: Roadways blight cities. Traffic congestion is the worst. Suburbanization harms the environment. Fortunately, say these smart growthers, there is an alternative: By piling on regulations and reallocating transportation-related tax money, we can "densify" our urban communities, allowing virtually everyone to live in a downtown area and forego driving in favor of walking or biking.
Smart growth proponents have been gaining influence for decades. They've implemented urban growth boundaries (which greatly restrict the development of land outside a defined area), up-zoning (which tries to increase densities in existing neighborhoods by replacing single-family homes with apartments), and "road diets" (which take away traffic lanes to make room for wider sidewalks and bike lanes).
Alas, there are inherent flaws in the "smart growth" approach—beginning with the idea that it makes sense for everyone to live and work in the same small area. In fact, that idea flies in the face of what economists call urban agglomeration.
Urban agglomeration is why there are more jobs in and around big cities. Job seekers have access to a large number of potential employers, which increases each person's likelihood of finding one that can make the best use of her unique talents and skills. The same is true for business owners, who have a much better chance of finding people in a large populous urban area who match their needs.
Transportation turns out to be a key factor in enabling these wealth-increasing transactions. Imagine drawing a circle around the location of your residence, defined by how far you are willing to commute to get to a satisfying job. The larger the radius of that circle, the more potential work opportunities you have. Likewise, a company's prospective-employee pool is defined by the number of people whose circles contain that company's location.
Most people measure that radius in time rather than distance; studies show they are generally unwilling to spend much more than 30 minutes commuting each way on a long-term basis. That means the size of their opportunity circle is critically dependent on how quickly they can get around.
Despite urban sprawl and ever-increasing congestion levels, economists Peter Gordon and Harry Richardson of the University of Southern California have documented, using census data, that average commute times in various metro areas have hardly changed at all over several decades. More recently, Alex Anas of the University of Buffalo modeled what would happen as a result of a projected 24 percent increase in Chicago's metro area population over three decades. He estimated that auto commute times would increase only 3 percent and transit trip times hardly at all. The reason is that people tend to change where they live or work in order to keep their travel times about the same. But this happy result comes about only if the transportation system expands accordingly.
A recent empirical study from the Marron Institute of Urban Management at New York University likewise found that, on average, the labor market of an urban area (defined as the number of jobs reachable within a one-hour commute) nearly doubles when the workforce of the metro area doubles. The commute time increases by an average of only about 7 percent, however—assuming an efficient region-wide transportation network. To achieve higher economic productivity, they recommend fostering speedier rather than slower commuting; more rather than less commuting; and longer rather than shorter commutes.
These policies would expand the opportunity circles of employers and employees, enabling a more productive urban economy. But these are exactly the opposite of the policy prescriptions of smart growth, which generally seek to confine people's economic activity to a small portion of a larger metro area.
One early manifestation of this was the attempt by urban and transportation planners in the '80s and '90s to promote "jobs-housing balance," where each county of a large metro area has comparable percentages of the region's jobs and of its housing. The rationale was that this would reduce "excessive" commuting by enabling people to find work close to their homes. But urban agglomeration theory makes it clear that that is a recipe for a low-productivity urban economy. Census data show that many suburban areas are now approaching jobs-housing balance on their own, but this does not necessarily reduce commute distances—to get to the jobs they want, many people still travel across boundaries.
A fascinating example is Arlington County, Virginia. Since 2000, the number of jobs and the number of working residents in the county have been approximately equal. But it turns out that only 52 percent of those working residents have jobs in the county. Out of 582,000 resident workers, 280,000 commute to adjacent counties or the District of Columbia. And out of 574,000 jobs in the county, 272,000 are filled by workers from other places.
A less extreme version of smart growth says that we should discourage car travel and shift resources heavily toward transit. People should be encouraged to live in high-density "villages" where they can easily obtain transit service to jobs elsewhere in the metro area. The problem with this vision is the inability of transit to effectively compete with the auto highway system.
Simply put, cars work better for workers. A 2012 Brookings study analyzing data from 371 transit providers in America's largest 100 metro areas found that over three-fourths of all jobs are in neighborhoods with transit service—but only about a quarter of those jobs can be reached by transit within 90 minutes. That's more than three times the national average commute time.
Another study, by Andrew Owen and David Levinson of the University of Minnesota, looked at job access via transit in 46 of the 50 largest metro areas. Their data combined actual in-vehicle time with estimated walking time at either end of the transit trip, to approximate total door-to-door travel time. Only five of the 46 metro areas have even a few percent of their jobs accessible by transit within half an hour. All the others have 1 percent or less. Within 60 minutes door-to-door, the best cities have 15–22 percent of jobs reachable by transit.
Meanwhile, Owen and Levinson found that in 31 of the 51 largest metro areas in 2010, 100 percent of jobs could be reached by car in 30 minutes or less. Within 40 minutes, all the jobs could be reached by car in 39 of the cities. Within an hour, essentially every job in all 51 places could be reached by car. The roadway network is ubiquitous, connecting every possible origin to every destination. The contrast with access via transit—let alone walking or biking—is profound.