Structural equation models of VMT growth in US urbanised areas

Authors

Reid Ewing, University of UtahFollow
Shima Hamidi, University of Utah
Frank Gallivan, ICF International
Arthur C. Nelson, University of Utah
James B. Grace, US Geological SurveyFollow

Date of this Version

2014

Citation

Urban Studies 2014, Vol. 51(14), pp. 3079–3096.

Comments

U. S. government work.

Abstract

Vehicle miles travelled (VMT) is a primary performance indicator for land use and transportation, bringing with it both positive and negative externalities. This study updates and refines previous work on VMT in urbanised areas, using recent data, additional metrics and structural equation modelling (SEM). In a cross-sectional model for 2010, population, income and freeway capacity are positively related to VMT, while gasoline prices, development density and transit service levels are negatively related. Findings of the cross-sectional model are generally confirmed in a more tightly controlled longitudinal study of changes in VMT between 2000 and 2010, the first model of its kind. The cross-sectional and longitudinal models together, plus the transportation literature generally, give us a basis for generalising across studies to arrive at elasticity values of VMT with respect to different urban variables.