| The relationship between density and vehicle miles travelled is one of the most frequently studied topics in urban economics and planning. While, for decades, economists and planners had believed that density had a strong, negative relationship with how much people drive, modern studies have suggested otherwise. Two models, Brownstone & Golob (2009) and Bento et al (2005), achieved such results by accounting for self-selection bias in neighborhood choice. This paper builds off of the findings of these models and includes public transportation as an endogenous explanatory variable of VMT in an instrumental variables framework. The results are largely consistent with established theory, and support the notion that the absence of public transportation accessibility in many models may lead to an upward bias in the effect of density on VMT. These results should be viewed with some reservations, however, given identification issues regarding the public transportation variables. |
