A geographic and statistical analysis of the relationship between neighborhood and regional urban form and household car-dependency in Montreal (Quebec)

The negative consequences (especially urban smog and global warming) of the heavy dependence on motorized vehicles of residents in urban areas across Canada and the United States have rightly motivated a great deal of research into potential policies aimed at reducing car ownership and use. This research focuses on the potential role of neighborhood and regional urban form and land use pattern in reducing car-dependency in cities. The study relies primarily on an origin-destination travel survey collected by the Agence Metropolitaine de Transport (AMT) in 1998, consisting of a 5% representative sample of residents living in the Montreal CMA. The first statistical model of household car-dependency is a multinomial logit model (MNL) of household automobile ownership levels on the Island of Montreal. The results suggest that neighborhood form has only a modest relationship automobile ownership levels. However, the location of a household's neighborhood relative to Montreal's CBD and whether an adult in it is employed there are both strongly associated with the number of vehicles a household owns, especially of multiple vehicles. The results indicate the importance of the spatial distribution and density of employment opportunities to vehicle ownership The second statistical model, an OLS regression of vehicle kilometers driven (VKD) by households for shopping purposes, found that neighborhood population density and proximity to a large retail cluster are both associated with meaningful reductions in motorized travel. Having said that, the results from both models suggest the most important determinants of household car-dependency are its socioeconomic and demographic composition. These findings suggest new suburban developments based on the principles of New Urbanism are unlikely to have any meaningful effect on car-dependency in urban areas.