In that sense, you might think the most room is out West. In an America where people often insist their cities are too “full” to accommodate growth, this is where you would want to build more.
But that’s not quite the right way to look at the question. Most of us aren’t aiming to maximize the amount of land we own. We want some space, but space within commuting distance of jobs, stores and schools. Looking only among metropolitan areas, it turns out that the roomiest ones are not in the West. They’re in the South.
To see how much room there is, look at the density -- measured in households per square mile, of the typical household’s neighborhood in each major metro area in the United States. In other words, in each metro, how crowded is the neighborhood where the typical household lives?* By this measure, New York is the most crowded. It has, by far, the highest density of all metro areas in the country, followed by three in the West: Honolulu, San Francisco and Los Angeles.
|Metro||Households per square mile|
|1||New York-Newark-Jersey City, NY-NJ-PA||13,370|
|2||Urban Honolulu, HI||6,021|
|3||San Francisco-Oakland-Hayward, CA||5,630|
|4||Los Angeles-Long Beach-Anaheim, CA||4,260|
|9||Miami-Fort Lauderdale-West Palm Beach, FL||3,169|
|10||San Jose-Sunnyvale-Santa Clara, CA||2,857|
|11||San Diego-Carlsbad, CA||2,677|
|12||Las Vegas-Henderson-Paradise, NV||2,431|
In general, metro areas with more people tend to have higher densities. But there are striking exceptions.
Atlanta, the ninth largest metro area by population, ranks only 132nd for density. The typical household in Atlanta lives in a neighborhood with just 931 households per square mile -- or 1.5 households per acre. This is less than one-fourth the density of the typical household in Los Angeles or Chicago.
Other Southern cities have plenty of room, too. Among metro areas with populations of at least two million people, Charlotte, N.C., has the lowest density, followed by Atlanta. Among those with one to two million people, Birmingham, Ala., is the least dense, followed by Nashville and Raleigh, N.C.
Where to find a spread-out neighborhood
These figures are averages, so they hide what’s really going on across different neighborhoods within metro areas. To see this, let’s compare New York and Los Angeles. New York is three times as dense as Los Angeles. That’s because New Yorkers are far more likely to live in high-rise neighborhoods than people in Los Angeles are.
Yet this is not a complete picture. While many New Yorkers may be tightly packed in big apartment buildings, millions also live in single-family homes in the outer boroughs and surrounding region. By contrast, far fewer people in L.A. live in crowded high-rises, but Los Angeles residents are less likely than New Yorkers to live in spread-out neighborhoods. In metro Los Angeles, just five percent of households live in neighborhoods with less than one household per acre (that’s 640 households per square mile).
Thirteen percent of people in the New York metro area live in neighborhoods with less than one household per acre — a higher share than in Miami, San Jose, the San Francisco Bay Area, and even Las Vegas. Eight of the 10 large metro areas where few residents have lots of space are in the western half of the country.
|Metro||Share of households living in low-density neighborhoods (<1 household per acre)|
|1||Los Angeles-Long Beach-Anaheim, CA||5.1 percent|
|2||Miami-Fort Lauderdale-West Palm Beach, FL||7.7 percent|
|3||San Jose-Sunnyvale-Santa Clara, CA||9.6 percent|
|4||San Francisco-Oakland-Hayward, CA||11.5 percent|
|5||Las Vegas-Henderson-Paradise, NV||11.7 percent|
|6||New York-Newark-Jersey City, NY-NJ-PA||13.2 percent|
|7||Salt Lake City, UT||13.4 percent|
|8||San Diego-Carlsbad, CA||14.2 percent|
|9||Denver-Aurora-Lakewood, CO||17.2 percent|
|10||Urban Honolulu, HI||17.3 percent|
Density isn't always the opposite of sprawl
A lot of these figures run counter to conventional wisdom, where we assume that places with a lot of sprawl — development pushing out from city centers, forcing long commutes — would also be the least dense. And this intuition is usually right: Less dense areas tend to be more car-dependent, and denser areas rely more on public transit.
But, again, there are exceptions. Portland residents are 1.7 times as likely to take transit to work as Las Vegas residents (and almost three times as likely to walk or bike). But Las Vegas actually has denser neighborhoods than Portland.
Likewise, transit usage is higher in Boston, Philadelphia, Washington, D.C., and Chicago than in Los Angeles. But L.A. is more densely built than all of those metro areas.
This unexpected finding tells us something important about density. People don’t give up their cars just because they live close together.
Rather, a metro area’s level of transit usage is more closely tied to where people work than where people live. Metros with the highest job density – especially if those jobs are clustered in a single downtown – have the highest transit usage. People who live near transit are more likely to commute by transit, as you’d expect – but working right near a transit station is even more important to get people out of their cars. If your home is two miles from a subway or rail stop, you might drive to the station. But if your job is two miles from the station, you probably can’t get there easily.
As you’ve probably guessed, those metros with surprisingly high housing density but low transit usage don’t have particularly high job density. Job density, for instance, is three times as high in metro Washington, D.C., as in Los Angeles.
The politics of density reflect how closely together we live and where there’s room to build -- and the places with lots of room to grow aren’t always where you’d expect.
But while the politics of density are about housing, the benefits of density -- at least in terms of transportation and the environment -- have as much to do with where people work as with where people live.
*This is known as weighted density and equals the average of Census tract densities within a metropolitan area, weighted by the number of households in the tract. It reflects the density that residents actually experience and avoids the problem that conventional density measures face of being skewed by empty land within jurisdictional boundaries.