Modeling Car Ownership in Urban Areas of Developing Countries

Determinants of household car ownership in Bogotá, Colombia, based on disaggregate data from mobility surveys in 1995 and 2005, are analyzed. The development of discrete choice models (multinomial logit and ordered logit) allows the individual effects of income, household size, and location variables to be disentangled. Results showed that income had the greatest impact on car ownership, with an aggregate-level elasticity between 0.908 and 1.110 in 1995. At the same time, the impact of other variables was relatively low, because income was a major limitation to car ownership. Analysis of behavioral changes in both years showed that distance to work had a negative effect on car ownership in 1995 but a positive effect in 2005; this result suggests an interesting area for further research. Additionally, population density was found to have a negative effect in 1995 but no effect in 2005; this change indicates an uncertain role for densification strategies to reduce car ownership. Finally, a higher quality of public transportation was found to have a generally negative effect on car ownership.