Date of Graduation
Chambers College of Business and Economics
David R. Martinelli.
Network analysis in transportation economics has traditionally focused on congestion as a negative externality stemming from supply-side capacity constraints. In my first paper paper, an analytical mode choice model is developed to examine the demand-side network effects. The assumption behind the approach is that, because of social network effects, the utility of people taking the mode increases with its mode share. It is found that social network effects change the modal aggregate demand curve for the mode to an inverted u-shape. This result has far-reaching policy consequences, since multiple equilibria become a possibility, causing positive externalities and path-dependency.;Transportation planners have always been aware of positive network effects in public transit use, which can be attributed to the fact that people choose transit, because other people already take it. In my second essay, I employ a spatially autoregressive mode choice mode to econometrically test for the existence of social network effects. It is found that the coefficient estimate for transit use network effects is positive and significantly different from zero. Furthermore, if social network effects are not included, it can be shown that an omitted variable bias is introduced into the model, which can lead to a systematic error in travel forecasts.;The third essay explains municipal differences in bicycle mode share with social network effects. Using data from the nation-wide travel behaviour survey, Mobility in Germany 2002, a binary logistic regression model was developed to identify in how much a city-specific ''biking culture'' has an impact on the city's bike modal split. To avoid endogeneity of the biking culture variable, a social network effects instrument was developed. It was found that not only bicycle infrastructure, but also social network effects change municipal bike mode share. Further results were that work/educational and leisure trips depend less on social network effects than other trip purposes. The outcome of this research has significant policy implications, such as, that transportation planners can target biking culture in a city as a mean to improve bike mode share.
Goetzke, Frank, "Network effects and spatial autoregression in mode choice models: Three essays in urban transportation economics" (2006). Graduate Theses, Dissertations, and Problem Reports. 2497.