Mapping Vehicle Diffusion Dynamics in the United States

Abstract

Household vehicle fleet composition is an important dimension of transportation policy and planning. Vehicle choice and diffusion may be considered as a function of two network dimensions. First, spatial proximity may be associated with diffusion of a new vehicle powertrain, category, or model. A household purchases a vehicle based on local incentives or seeing the vehicle on the road. Second, social proximity as measured by similarity in demographics. We use a historical county-level vehicle registrations database for the United States to examine these factors. Results of the Moran I coefficient of correlation confirm the spatial diffusion phenomena for plugin electric vehicles (PEV)—including battery and plug-in hybrid electric, hybrid electric vehicles (HEV), hydrogen fuel cell electric vehicles (FCEVs), and US brand vehicles. The spatial autoregressive (SAR) model is estimated to measure the spatial diffusion effect after accounting for demographic variables associated with social diffusion. Larger county populations are linked to higher PEV and HEV adoption rates but lower US brand vehicle ownership. Elevated state grid eqCO₂ intensity is associated with higher adoption rates for both PEV and HEV. Centrally located states maintain high US brand vehicle ownership rates, while West and East Coast states show consistently low rates, with declining diffusing into neighboring regions.