Graduation Year


Document Type




Degree Granting Department


Major Professor

Joseph S. DeSalvo, Ph.D.

Committee Member

Brad Kamp, Ph.D.

Committee Member

Ram M. Pendyala, Ph.D.

Committee Member

Gabriel A. Picone, Ph.D.


urban economics, residential location, travel behavior, activity-based modeling


This study presents an analytical model of the interaction between urban form and the demand for transit travel, in which residential location, transit demand, and the spatial dispersion of non-work activities are endogenously determined. In this model, travel demand is considered a derived demand brought about by the necessity to engage in out-ofhome activities whose geographical extent is affected by urban form. In a departure from the urban monocentric model, residential location is defined as a job-residence pair in an urban area in which jobs, residences, and non-work activities are dispersed. Transit demand is then determined by residential location, work trips, non-work trip chains, and goods consumption.

Theoretically derived hypotheses are empirically tested using a dataset that integrates travel and land-use data. There is evidence of a significant influence of land-use patterns on transit patronage. In turn, transit demand affects consumption and non-work travel. Although much reliance has been placed on population density as a determinant of transit demand, it is found here that population density does not have a large impact on transit demand and, moreover, that the effect decreases when residential location is endogenous. To increase transit use, urban planners have advocated a mix of residential and commercial uses in proximity to transit stations. In this study, it is found that the importance of transit-station proximity is weakened by idiosyncratic preferences for residential location. In addition, when population density and residential location are jointly endogenous, the elasticity of transit demand with respect to walking distance to a transit station decreases by about 33 percent over the case in which these variables are treated an exogenous.

The research reported here is the first empirical work that explicitly relates residential location to trip chaining in a context in which individuals jointly decide residential location and the trip chain. If is found that households living farther from work use less transit and that trip-chaining behavior explains this finding. Households living far from work engage in complex trip chains and have, on average, a more dispersed activity space, which requires reliance on more flexible modes of transportation. Therefore, reducing the spatial allocation of non-work activities and improving transit accessibility at and around subcenters would increase transit demand. Similar effects can be obtained by increasing the presence of retail locations in proximity to transit-oriented households. Although focused on transit demand, the framework can be easily generalized to study other forms of travel.