Vehicle miles traveled and the built environment: evidence from vehicle safety inspection data

This study examines the linkage between household vehicle usage and their residential locations within a metropolitan area using a newly available administrative dataset of annual private passenger vehicle safety inspection records (with odometer readings) and spatially detailed data on the built environment. Vehicle miles travelled (VMT) and a set of comprehensive built-environment measures are computed for a statewide 250 x 250 m grid cell layer using advanced geographic information systems and database management tools. We apply factor analysis to construct five factors that differentiate the built-environment characteristics of the grid cells and then integrate the built-environment factors into spatial regression models of household vehicle usage that account for built environment, demographics, and spatial interactions. The empirical results suggest that built-environment factors not only play an important role in explaining the intraurban variation of household vehicle usage, but may also be underestimated by previous studies that use more aggregate built-environment measures. One-standard-deviation variations in the built-environment factors are associated with as much as 5000-mile differences in annual VMT per household. This study also demonstrates the potential value of new georeferenced administrative datasets in developing indicators that can assist urban planning and urban management.