The National Academy of Sciences Transportation Research Board released "Driving and the Built Environment," a study of the potential impacts of increasing density on carbon dioxide levels. (A very useful summary can be found here.) The results, as I told USA Today in an article reporting on the study's release, are underwhelming, to say the least.
According to the analysis in the report, if we funnel 75 percent of future land use into higher density settings--effectively doubling current regional population densities--we might reduce carbon emissions by 11 percent. But, the chances of that happening are so remote they almost aren't even worth considering. Anthony Downs, one of the study's coauthors, said as much and is quoted in a report on the study's release by Technology Review.
One of the study's authors doubts whether major increases in housing density are even possible. "I think the 75 percent figures are completely unrealistic," says Anthony Downs of the Brookings Institution. "Twenty-five percent is much closer to realistic and that may even be high. Nationally we've had no increase in housing density in the last 30 years; I don't see that reversing."
A more moderate scenario that assumes about 25 percent of replacement housing is built in high density settings suggest CO2 emissions will fall by about 1 percent. That's probably more plausible. But even this estimate may be high short of planning policies that dramatically restrict suburban development using acreage of one quarter of an acre or more for new housing.
But this assumes these studies are reliable. The TRB report points out that most of the studies examining the relationship between land use and transportation are weak and lack methodological consistency.
While many proponents of Smart Growth, including draconian policies such as urban growth boundaries that explicitly limit land available for development to force higher densities, believe the study vindicates their position, it doesn't. While the research shows that higher densities in the right place and with the right uses can reduce vehicle miles traveled (VMT), this is a very blunt industry if the goal is to influence carbon emissions (and climate change).
It's not land use per se that produces carbon emissions; it's the fuel for the technology we use to get from place to place, heat our homes, and run the conveniences of life. The most direct and efficient way to reduce carbon emissions is to replace the fuel technology, not limit housing or transportation choices, or force people into "second best" alternatives.