Residential choice, the built environment, and nonwork travel: evidence using new data and methods

Residents of dense, mixed-use, transit-accessible neighborhoods use autos less. Recent studies have suggested that this relationship is partly because transit-preferring and walk-preferring households seek and find such neighborhoods. If this is so, and if the number of such households is small, policies to alter the built environment may not influence auto use very much. I argue that many of these studies are inconclusive on methodological grounds, and that more research is needed. A purpose-designed survey of households in two urban regions in California is investigated, with the aid of a new methodological approach. I find that most surveyed households explicitly consider travel access of some kind when choosing a neighborhood, but that this process of residential self-selection does not bias estimates of the effects of the built environment very much. To the extent that it does exert an influence, the bias results both in underestimates and overestimates of the effects of the built environment, contrary to previous research. The analysis not only implies a need for deregulatory approaches to land-use and transportation planning, but also suggests that there may be value in market interventions such as subsidies and new prescriptive regulations.