Dr. Rachel Weinberger Authored a Paper in the journal of Environment and Planning titled: “Separating Contextual from Endogenous Effects in Automobile Ownership Models”.

Paper Citation: Goetzke F, Weinberger R, 2012, “Separating contextual from endogenous effects in automobile ownership models” Environment and Planning A 44(5) 1032 – 1046

Abstract:

Using the 1997/98 New York Metropolitan Transportation Council household survey and United States Census, we estimate an instrumental variable probit model to test the impact of contextual and endogenous social interaction effects on auto ownership and determine that the probability of car ownership is affected by both types of social interaction effects. Previous research focused only on either contextual effects, or, increasingly, on endogenous effects using contextual effects variables as instruments. Therefore we were unable to find studies looking at both social interaction effects simultaneously. Consistent with earlier results, we find that households have a higher probability to
possess a vehicle if they are surrounded by other automobile-owning households (endogenous effect). However, we find that contextual effects are only be correctly measured when the endogenous effect is included. In our case,  everything else being equal, that households in poorer neighborhoods are more likely to own vehicles and households in neighborhoods with higher proportions of people with graduate degrees are less likely to own vehicles. This suggests that car ownership in New York City is a status symbol for poorer households and that noncar ownership is a status symbol for people with post baccalaureate education. The results are important in two policy contexts: as auto ownership is a precursor to trip generation and mode choice, auto ownership estimation is important to effective travel forecasting; as vehicle miles traveled (VMT) is tied to auto ownership, VMT reduction strategies, as a way to improve air quality, reduce congestion and reduce greenhouse gas emissions may depend on auto ownership reduction strategies. In either case, correct modeling of auto ownership will lead to more effective policy outcomes.