Does TOD Need the T? Auto Use, Residential Sorting, and Access to Rail

Smart growth planning often focuses on building new housing near rail stations, on the assumption that transit-oriented development can reduce driving and therefore mitigate climate change, pollution and congestion. But research has rarely investigated how transit-oriented development affects auto use—or asked whether new development should be oriented away from autos, rather than toward rail. This study addresses two questions. First, do households choosing new housing near rail stations have different patterns of auto ownership and use? Second, are those differences a result of rail access, or other factors associated with rail access? Previous studies have not used comparable data to compare nearby and farther away housing units; have lacked data on parking supply; have not controlled for characteristics of the housing itself, particularly the age of the housing; and have generally failed to account for how residential choices may affect TOD housing outcomes. I surveyed over 1,100 households within two miles of ten rail stations, and conducted a field count of over 6,200 on-street parking spaces on 818 block faces near the stations. The household survey collected information on housing characteristics, parking, travel, and household demographics as well as stated residential choice criteria to control for heterogeneity in preferences. The survey data were geocoded and joined to on-street parking supply data and other spatial data from secondary sources. I used regression analysis to examine how housing, parking, neighborhood and subregional spatial characteristics are correlated with automobile ownership, commuting, and grocery trips. Auto ownership and use is much lower among households living in new housing near rail stations, but these differences are not explained by rail access. Off- and on-street parking availability, housing type and tenure, local and subregional density, and bus service are much more highly correlated with residential choices and subsequent auto ownership and travel. Rail access is associated with lower auto use when combined with factors like parking supply and housing type, but rail by itself has little explanatory power. I conclude that the details of TOD matter, and some of the things that “work” about TOD have little or nothing to do with rail access by itself. Planners should broaden efforts to develop dense, mixed-use housing beyond rail station areas. Denser housing development coupled with good management of automobile parking and improved bus service could be more effective, and less expensive, than a development policy oriented around rail stations.