Cross-sectional and quasi-panel explorations of the connection between the built environment and auto ownership

Auto ownership is a critical mediating link in the connection between the built environment and travel behavior: the built environment presumably influences auto ownership, which in turn impacts travel behavior. However, the way in which individual elements of the built environment affect auto-ownership choices is far from understood. Further, residential self-selection may confound the interaction between the built environment and auto ownership. And the absence of panel data impedes our understanding of the causal relationships. Using a survey of 1682 respondents in Northern California, we applied ordered probit and static-score modeling techniques to investigate the causal link from the built environment to auto ownership in both cross-sectional and quasi-panel contexts. Through variable selection in cross-sectional analysis, we found that individuals' attitudes regarding residential neighborhood and travel are more strongly associated with their auto-ownership decision than is the built environment per se. Specifically, when general preferences for various neighborhood traits were allowed to enter the model, they drove out from the model the (perceived) measure of the same trait for the neighborhood of current residence, a pattern suggesting that the observed correlation between neighborhood characteristics and auto ownership is primarily a result of self-selection. On the other hand, the quasi-panel results indicate that some built-environment elements such as outdoor spaciousness and mixed land use are causes of auto ownership (remaining even after attitudes were allowed to enter the model), but their effects are marginal. In contrast, the strong influence of sociodemographics suggests that households' auto-ownership decisions are fundamentally based on their mobility needs and purchasing power. Given the mixed findings, we do not definitively confirm a causal relationship between the built environment and auto ownership. However, we provide encouraging evidence that land-use policies designed to reduce auto ownership and use will lead to a marginal reduction in auto ownership.