Statistical Modeling of Spatial and Temporal Variation in Vehicle Kilometers Traveled

VKT (Vehicle Kilometers Traveled) is one of the most commonly-used measure of transportation system performance, and often a key variable in transportation operations, planning, and sustainability research. VKT models are frequently employed by decision-makers to understand the impact of various policy initiatives (e.g. fuel price/tax increase, vehicle registration/use regulation, emission control) on vehicle use, congestion, fuel consumption, and pollution/green house gas emissions. Based on state- level data in the U.S., we built two statistical models to estimate short- and long-run VKT dynamics: fixed-effect time-series model and structural equation model. Both models capture the spatial and temporal variance in VKT very well, and the results obtained from the two models are largely consistent.