Abstract

This three-essay dissertation focuses on understanding linkages between urban form, travel behavior, ownership of alternative fuel vehicles, active commuting, congestion, fuel consumption, and air pollution (including greenhouse gas emissions). These essays estimated different specifications of Generalized Structural Equation Models (GSEM) to explicitly account for residential self-selection and vehicle choice endogeneities. The first essay analyzes the influence of land use policies and gasoline prices on driving patterns. I estimated a Generalized Structural Equation Model (GSEM) with a Tobit-link specification on a Southern California subsample of the 2009 National Household Travel Survey (NHTS). These data haves a quasi-experimental nature thanks to large exogenous variation in gasoline price during the survey period. I analyzed separately home-based work trips and non-work trips under the hypothesis that households have more flexibility to adjust their non-work trips when gasoline prices change, whereas most of the literature does not take trip purpose into account. To measure urban form, which is treated as a latent construct, I used fine-grained geospatial information including population density, land use mix, employment density, distance to employment centers and transit availability. I found that, in the short run, households drive 0.171% less for non-work trips when gasoline prices increase by 1%, while work trips are not responsive to gasoline price changes. This suggests that, in the short term, higher fuel prices reduce discretionary driving such as shopping and recreational trips, but they do not affect non-discretionary driving such as commuting trips. My results also suggest that policies that seek to increase transit service and housing opportunities near employment centers will reduce driving. The second essay investigates the impact of government incentives such as access exemption to High Occupancy Vehicle (HOV) lanes and parking privileges on household ownership of Alternative Fuel Vehicles (AFVs) using Generalized Structural Equation Models (GSEM), and accounts for residential self-selection, household demographics and ambient political-environmentalism. I analyzed geocoded travel diary data from the 2012 California Household Travel Survey (CHTS), linked with fueling station data from the US Department of Energy Alternative Fuels Data Center and precinct level election data from the UC Berkeley Statewide Database. My findings suggest that, on average, households with alternative fuel vehicles drive approximately 10 miles more on weekdays and about 0.5 miles more on non-discretionary trips than otherwise similar households. In addition, households who live closer to a freeway with HOV lanes, work closer to an AFV charging facility (that provides free parking), and are likely supportive of pro-environmental measures are more likely to own alternative fuel vehicles. The third essay examines the influence of urban form on transit use and non-motorized travel (NMT, including biking and walking) for households (with at least one employed adult) in Los Angeles and Orange Counties in California based on 2009 National Household Travel Survey (NHTS) data. The objectives of the research are (1) to assess several methods for measuring urban form features in the near-residence and near-workplace environments and (2) to assess the importance of these urban form features on transit use and NMT after accounting for the influence of these features on household vehicle ownership and residential selection. Results provide insights into the relative influence of several specifications of population density, transit access and walkability measures on transit use and NMT for commute and non-work trips. Reduced form models suggest that the dominant determinant of discretionary travel is household socio-demographic status. In terms of residential selection, lower income, younger, and smaller households are more likely to choose a dense, pedestrian friendly, and transit rich neighborhood. In terms of vehicle ownership, households living in high density, pedestrian friendly, and transit rich neighborhoods are less likely to own vehicles. After accounting for the influence of urban form on vehicle ownership and residential selection, workplace transit accessibility has greater influence on transit commuting than transit access near a household’s residence. Results vary by how urban form is specified and by the source of travel data. Finally, there is some evidence that population density affects active travel for discretionary purposes.