I first want to thank Jordan McGillis for his careful engagement with my piece on the environmental benefits of sprawl. I couldn’t ask for a more thoughtful response, and I appreciate his pointing out the many areas where we agree. But I wouldn’t be doing my job as respondent if I didn’t take a moment to point out the issues on which we remain apart.
McGillis points out that the United States continues to emit more greenhouse gases per capita than countries like Japan and Germany and that part of this difference comes from high transportation emissions. Yet most of our differences in those emissions cannot be attributed to “sprawl” per se. Only about 16 percent of all US emissions come from cars. And much of our excess there has to do with the United States’ larger and less fuel-efficient automobile fleet (not a result of sprawl) and its more distant cities (no amount of density is going to make Chicago as close to New York City as Osaka is to Kyoto). Most importantly, transportation overall generates the same proportion of overall greenhouse gas emissions in America as in Europe, again making clear that sprawl is not at the root of the emissions divergences.
Yet the most important point that I would emphasize about sprawl and driving uses the good ole’ economic term “sunk costs.” The argument in my original article was not that US sprawl hasn’t increased driving to some extent. It likely has, but that is irrelevant for the policy debate since those costs are sunk, having already been incurred. The real question is the costs and benefits of densification going forward.
In 2022, the United States has over 140 million existing homes. We build about 1.5 million new homes a year on average. Since most new homes won’t be demolishing old ones (and demolition would have environmental costs of its own), it will take almost 50 years for the United States to change even one-third of its total housing stock. And even if those new homes were built at three times the previous density—an impossible task considering current trends—it would only double US density in that period. McGillis has not cited any studies countering a metasurvey from the National Research Council that found a doubling of residential density would only reduce Vehicle Miles Traveled (VMT) by 5 to 12 percent. Given that, in my imaginary densification campaign, we would be engaging in an unprecedented act of social engineering to only slightly reduce VMT five decades hence!
As I also noted, not just the increasing use of electric cars (which, McGillis and I agree, given our current electricity mix are not an unalloyed good) but the increasing efficiency of gasoline cars and the increasing adoption of hybrid and autonomous cars will sever the connection between VMT and carbon emissions in the future. Even if we don’t assume substantial increases in fuel efficiency, which seem likely, but instead extrapolate from trends since 1975, my imaginary densification campaign would reduce the 16 percent of US carbon emissions coming from cars by less than one percentage point over the course of a few decades. Again, that would be one of the slowest and most costly imaginable means to reduce greenhouse gas emissions.
In regard to buildings, McGillis compares energy use between large apartments and single-family homes. But the crucial difference is that even large apartments are usually smaller than detached homes and contain fewer wealthy people in them, both of which lead to lower energy use overall. That’s why, I think, the best comparison measure for efficiency is energy use per square foot, and I did not see any disagreement from him with the studies I cited on that score—all of which show more energy use for the construction of taller buildings and more energy spent on electricity and energy in them.
Finally, in terms of the economic costs of the US suburbs, McGillis brings up familiar arguments, emerging from the 1974 and 2000 “Costs of Sprawl” reports, that elongated roads, pipes, and electric lines add lots of public costs that are not paid by suburban homeowners. This ignores the fact that in most suburbs, especially newly incorporated suburbs on the fringes of cities, the property tax on homes pays for all of these things. In much of Texas, for example, municipal utility districts are formed along with new developments to pay for new infrastructure costs, for which they only charge the homeowners. This does not keep overall housing costs in, for example, Houston, Texas, from being less than a third of those in San Jose, California. In reality, except in extremely sparsely populated areas of less than 200 people per square mile, the costs of public services have been found to be much higher in dense areas and in large municipalities (anyone who’s witnessed a rehabilitation or construction project in New York City can vouch for this).
In some sense, of course, all of these debates are speculative. While McGillis and I both admit there are costs and benefits to sprawl as well as density, we disagree on where to draw the line. If a correct price were put on carbon, we could see the effects of that price on automobile efficiency, building height, and density, so that we would not have to imagine them through dueling studies. We both want Americans to live in the environment they prefer, as long as they pay for any “externalities.” We can’t know what that will really cost, though, without an accurate carbon price.
McGillis notes that my “Sprawl Is Good” title may have been tongue in cheek. Though it was intended to provoke, I meant it. Just as we should celebrate our amazing increases in wealth and health in the past 100 years, we should celebrate the fact that technology has allowed us to escape the dense, dirty, noisy, and cramped five-story walk-ups of our urban ancestors, and allowed more of us to live in clean, expansive homes surrounded by trees and grass. It’s what most people, here and abroad, want, and we should all celebrate that economic and environmental success.