Abstract
Research on changes in car ownership decision has often considered purchase interval times using hazard models. A handful of studies competing risks models allowing not only to study duration, but also the kind of car ownership change. In these studies car purchasing and disposal behavior are usually assumed mutually independent. Ignoring possible correlation between these behaviors may lead to bias. In addition, to the extent that life events have been considered, usually identical effects have been assumed on the probability across the population, perhaps because traditional competing risk models cannot properly capture heterogeneity. This paper, therefore, proposes a latent class competing risk model, which relaxes the independence assumption between competing risks and captures heterogeneity among individuals in analyzing car ownership decisions. Particularly, the model incorporates the effects of socio-demographics and life events on car purchasing and disposal decisions. The model was estimated using the Bayesian Markov Chain Monte Carlo method. Results show significant heterogeneity in households’ car ownership decisions. The covariates are found to have significantly different effects for different segments of households. Baby birth and partner moving in increase the probability of car purchasing for nearly three quarters of the households, but have no significant effect for the remaining households. The partner leaving the household has a significant negative effect on the probability of car disposal for most households.
Original language | English |
---|---|
Title of host publication | Proceedings 97rd Annual Meeting of the Transportation Research Board, Washington DC, USA |
Publisher | Omnipress |
Number of pages | 14 |
Publication status | Published - 2018 |
Event | 97th Transportation Research Board Annual Meeting - Washington, United States Duration: 7 Jan 2018 → 11 Jan 2018 |
Conference
Conference | 97th Transportation Research Board Annual Meeting |
---|---|
Country/Territory | United States |
City | Washington |
Period | 7/01/18 → 11/01/18 |