Analysis of household car and two-wheeler ownership considering spatial dependency

This paper investigates the spatial spillover effects of built environment features on household car and motorcycle ownership through aggregate ward-level analysis and household-level disaggregated analysis. Using a primary data collected in Hubli-Dharwad city in Southern India, the paper estimates Spatial Durbin models to assess the direct and spillover effects of socio-economic characteristics (household income, the number of children and household size), built environment characteristics (population density, diversity, distance, and subjective built environment measures), travel attitudes (pro-car, pro-NMT and pro-public transport), and commute behavior (distance and time) on household car and motorcycle ownership. The paper finds that spatial dependencies exist in both aggregate and disaggregate levels of analysis. There is a significant spatial dependency in car ownership due to household income, household size, the number of children, and pro-car attitude, and the correlations are significant in both aggregate and disaggregate levels of analysis. The positive lag effects of these variables suggests that socio-economic settings and attitudes of neighboring wards affect car and motorcycle ownership in a ward. The effects of diversity on car and motorcycle ownership maintain similar correlations in both levels of analysis. The likelihood of motorcycle ownership increases with improvements in diversity, while car ownership likelihood decreases. The study also shows that the spatial effects of built environment features could vary between the aggregate and disaggregate levels of analysis.