Abstract
The determinants of long-distance commuting (i.e., commuting trips longer than 50 miles one way) in the U.S. appear to be poorly understood even though long-distance commuting may have substantial environmental, social, and economic impacts. A review of the commuting literature also shows that commuting and residential self-selection have rarely been studied together, and that few studies have considered how residential land values influence long-distance commuting. In this paper, the authors start addressing these gaps by analyzing the long-distance travel component of the 2012 California Household Travel Survey (CHTS). The authors estimate a generalized structural equation model that explains the impacts of different socio-economic, land use, and land value variables on the likelihood that households commute long-distance. In particular, the authors’ model captures residential self-selection resulting from household socio-economic characteristics. We find that households who reside in areas with a higher mixed density index are less likely to commute long-distance (OR=0.91*), whereas households with a higher car to drivers ratio are more likely to commute long-distance (OR=1.20**). Moreover, although median residential home value is significant, it has no practical importance (OR=0.99***â??1). The authors’ results also illustrate the presence of residential self-selection as land use and land value are influenced by household socio-economic and demographic characteristics. More generally, the empirical evidence presented in this study shows that land use policies can influence long distance commuting and highlights the effectiveness of mixed development for curbing vehicle-miles traveled, which is one way of reducing the environmental impacts of transportation.