A simultaneous model of residential location, mobility tool ownership and mode choice using latent variables

Abstract

This paper investigates the interplay between three key choice dimensions in transport modeling: (i) Residential location, (ii) mobility tool (e.g., car and/or public transport season ticket) ownership and (iii) mode choice. We use a unique panel dataset of recently moved Swiss respondents to better assess the impact of changes in the respective dimensions. An expected trend is that the more accessible the residential location, the more likely the respondent owns a public transport season ticket, and the more likely public transport rather than a car is chosen. However, our model allows to disentangle the decision process and provides a more comprehensive picture of respondents’ behavior. For example, our simple mode choice model shows a strong and positive effect of income on the choice of car. In the simultaneous approach, given respondents’ increase in car ownership for increasing income, this effect becomes even more substantial if the residential location is assumed to be constant. With the latter assumption lifted, however, the effect decreases again, since respondents with higher income tend to live in areas with higher accessibility, lowering their car usage. An important general finding is that if the longer-term decisions (i.e., the choice of residential location and mobility tools) are considered, the short-term effects of level-of-service attributes (such as travel time) on mode choice is substantially weakened by more than 20%.