The effects of labor market spatial structure and the built environment on commuting behavior: Considering spatial effects and self-selection

In response to the problems caused by urban sprawl, smart growth strategies have gained increasing popularity among researchers and policy makers. Although numerous studies have identified the importance of land use policies on travel behavior, the effects are still inconclusive mostly due to two issues: residential self-selection and spatial spillover effect. Using instrumental variables with a generalized spatial two-stage least square estimator (GS2SLS), I investigate the effect of land use on travel behavior after controlling for the self-selection problems and spatial autocorrelation together. Particularly, I attempt to explain the effects of not only land use variables, but also neighborhood labor market spatial structures on commuting patterns. The findings clearly reveal that labor market spatial structure and land use play a significant role in determining commuting behavior at an aggregate level. Specifically, higher jobs-housing balance, more retail and service jobs in a neighborhood, higher density, and walker-friendly environment reduce commuting time and car use and increase non-motorized travel.