As this article from New Geography reports
“Around the world planners are seeking to increase urban densities, at least in part because of the belief that this will materially reduce automobile use and encourage people to give up their cars and switch to transit, or walk or cycle (Note 1). Yet research indicates only a marginal connection between higher densities and reduced car use. Never mind that the imperative for trying to force people out of their cars has rendered largely unnecessary by fuel economy improvements projected to radically reduce greenhouse gas emissions from cars (see Obama Fuel Economy Rules Trump Smart Growth).
“Transit Use and Density: A Tenuous Connection at Best
“In a widely cited study, Reid Ewing of the University of Utah, and UC Berkeley’s Robert Cervero reported only a minimal relationship between higher density and less driving per capita. In a meta-analysis of nine studies that examined the relationship between higher density and per household or per capita car travel, they found that for each 1 percent higher density, there is only 0.04 percent less vehicle travel per household (or per capita). This would mean that a 10 percent higher density should be associated with a reduction of 0.4 percent in per capita or household driving.
“More people in the same area driving a little less means overall driving is greater, as Peter Gordon reminds us. This is illustrated by the Ewing-Cervero finding — a 10 percent increase in population density is associated with 9.6 percent increase in overall driving, as is indicated in Figure 1 (the calculation is shown in the table). Ewing and Cervero placed this appropriate caution in their research: “we find population and job densities to be only weakly associated with travel behavior once these other variables are controlled.”
“There is another limitation to the density-transit research. The comparison of travel behaviors between areas of differing density provides no evidence that conversion of an area from lower to higher density would replicate the travel behavior of already existing (historic) areas of higher density.”