A joint model of residential mobility and location choice decisions

ABSTRACT

Residential relocation is a process involving the interdependent decisions of mobility (i.e. the decision to move) and location choice. Location choice is a discrete choice decision, whereas mobility has a continuous time dimension. Since households relocate along their life course for different reasons such as the birth of a child, these reasons could influence the duration of stay as competing events. This paper develops a joint model for mobility and location choice decisions. Data come from a retrospective survey conducted in the Central Okanagan of British Columbia, Canada. A reason-based competing hazard model is developed for the mobility component of the joint model. For the location choice component, a latent segmentation-based logit model is developed to capture unobserved heterogeneity. The influence of location utility on mobility is addressed through log sum parameters. The model confirms the effects of life-cycle events, parcel, socio-demographic, land use, transportation, neighbourhood and accessibility characteristics. The findings reveal that households relocating for the occurrence of life-cycle events are active in the housing market following the birth of a child, marriage and vehicle purchase. Relocation to live in a desirable neighbourhood is found to delay the move for the reason of living closer to activity points. Results also indicate that duration is likely to be shorter with the increase in expected utility from the following location. Regarding the location choice, preferences for dwellings closer to schools are evident. In the case of heterogeneity, it is observed that urban dwellers are more likely to prefer locations closer to the workplace with the addition of jobs in the household. In contrast, suburban dwellers are more likely to prefer residing far from the workplace. The developed joint model has been included in an integrated urban model, currently under development at the University of British Columbia.

KEYWORDS:

• Residential mobility
• Residential location choice
• Reasons for moving out
• Competing hazard model
• Latent segmentation-based logit model

Acknowledgements

The authors would like to thank BC Assessment for providing the parcel-level dwelling unit data for analysis.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada (NSERC) (discovery grant).