University of California Transportation Center

This dissertation is an exploration of the relationship between the transportation-land use system in New York City and the transportation and residential location choices made by New Yorkers. The focus is on understanding these location and travel choices made by urbanites. Specifically, this research uses discrete choice models to identify and quantify the effects of the variables that factor into New Yorkers' decisions about where to live, whether to own a car, and how to get around in their daily lives. These models, along with GIS technology, are used to answer the following questions:

1. How far off are the results of models that do not take all three of these decisions as endogenous?

2. In a densely populated urban environment, what are the policy-sensitive factors that determine whether households own cars and how often walking is the mode of choice?

3. How does the relative importance of these factors change across different neighborhoods within the city?

4. How much of the relationship between land use patterns and travel behavior is due to the indirect effects of neighborhood and car ownership choice, and how large is the direct effect of land use patterns on travel behavior?

Results indicate that the choices of residential location and commute mode are closely related; models of only commute mode choice produce biased results. Models that do not take the choice of car ownership as endogenous in New York do not appear to be severely biased.

Full models of the three choices indicate that the most important policy- sensitive factors influencing car ownership and mode choice are commute cost, commute time, and population density. A set of spatial scenario analyses illustrate that the importance of these factors does indeed vary across neighborhoods within the city.

Finally, a methodology is developed to separate the direct effect of land use patterns on travel behavior from the indirect effects. The example used here identifies the direct and indirect effects of population density on the propensity to walk, finding that approximately half of the total effect is direct.