I’ve previously discussed the myth that density relieves congestion, yet it persists. Most recently, planners in Fairfax County, Virginia say they want to put thousands of high-rise apartments in Tysons Corner in an effort to increase the density and relieve congestion around proposed rail stations.
Planners claim that Ballston, a rail station on the DC Orange line, proves that this strategy is successful. The opening of the Ballston station in 1979 led to a lot of transit-oriented development, and today many people in the area walk or take transit to work.
However, planners fail to mention that a major freeway, I-66, opened at about the same time, and it probably did more to stimulate development than the rail line. At least, other stations that were not close to new freeway interchanges failed to develop as planners hoped.
The myth that density can reduce congestion goes at least as far back as 1973 with publication of a book titled Compact City: A Plan for a Livable Urban Environment. The book called for high-rise housing projects, but suggested that Jane Jacobs’ “lively streets” model could work as well.
The year before, Edwin Mills’ models of urban economies had already proven than density only increased congestion because large increases in density would only produce small reductions in per-capita driving. But Mills was not a popular writer (except among economists), and his work was ignored by planners.
The modern smart-growth concept of density came from the LUTRAQ research program sponsored by 1000 Friends of Oregon. Based largely on hypothetical data, LUTRAQ found that density, pedestrian-friendly design, and improved transit service would lead to small reductions in per-capita driving. 1000 Friends equated such reductions with reductions in congestion — the two are not at all the same — and convinced the state of Oregon to require all major cities in the state to adopt such policies with the aim of reducing per-capita driving by 10 to 20 percent.
USC Professor Genevieve Giuliano pointed out that most of the driving reductions projected by LUTRAQ came not from density or land-use patterns but from an assumption that all employers and retailers would charge for parking and employers would give all their employees free transit passes. As far as I know, no city has dared require retailers or employers to charge for parking; instead, they focus on less effective (and counterproductive) practices like density and transit-oriented developments.
The Ballston story, and many similar stories, suffer from the problem of self selection. Even if it is true that people who live near Ballston drive less, it is largely because people who want to drive less choose to live near transit lines. This does not mean that putting people who drive a lot in high density transit-oriented developments will reduce driving. By extension, it does not mean that increasing the density of an urban area will reduce overall driving in that area.
One of the most striking examples of this can be seen in a study by junk scientists in the Sierra Club, Surface Transportation Policy Project, and other advocacy groups. The paper (abstracted here) claimed to prove that higher densities, pedestrian-friendly design, and intensive transit service led to less driving.
To reach this conclusion, the junk scientists looked at individual neighborhoods in the San Francisco, Chicago, and Los Angeles urban areas. But they also ranked each of these three urban areas, and San Francisco had the highest density, most intensive transit service, and (by their measures) twice the pedestrian friendliness of the other urban areas.*
Guess which urban area had the most per-capita driving? Yep, San Francisco. It also had the most vehicles per capita. (This information is from table one in their paper.) And not just slightly more driving either: San Francisco was supposed to have nearly 10 percent more per-capita driving and 20 percent more cars per capita than L.A. and close to 20 percent/40 percent more than Chicago.
On a neighborhood scale, urban design may correlate with driving, but this is due to people choosing their neighborhoods according to their transport preferences. On an urban scale, if there is any correlation at all, it is that more density means more driving, or at least no less. And even if per-capita driving were the same or slightly less in a denser area, it would mean more congestion and more air pollution.
As long as planners say “less congestion” when they mean “less per-capita driving,” planning will be a junk science.
* According to Census Data, Los Angeles has a higher density than the San Francisco-Oakland urban area. The Sierra Club-STPP study used the “San Francisco urban area,” which presumably includes only the west Bay Area (San Francisco and San Mateo).