Richard Florida is a co-founder and editor at large of CityLab and a senior editor at The Atlantic. He is a university professor in the University of Toronto’s School of Cities and Rotman School of Management, and a distinguished fellow at New York University’s Schack Institute of Real Estate.
The highly educated tend to live apart in college towns like State College, Pennsylvania, and big cities like Birmingham and Houston.
The postindustrial economy requires talented and educated workers—or, in economic parlance, those with high levels of human capital. And, as my previous research has demonstrated, nothing attracts smart people like other smart people, who concentrate in urban centers, amplifying innovation, entrepreneurialism, and economic growth.
Urbanists and economists agree that this clustering of human capital is a basic motor of economic growth. Its flip side, however, is geographic sorting—the tendency of talented and educated people to concentrate more in some places than others. Writing in The Atlantic, I dubbed this tendency “the means migration”; Bill Bishop calls it “the big sort.”
But how much of this kind of talent sorting goes on within metros? To what extent do the most educated citizens of any given metro area live separately from the rest of the population, in their own exclusive enclaves?
To get at this, my Martin Prosperity Institute (MPI) colleague Charlotta Mellander calculated the geographic segregation of highly educated people across the more than 70,000 Census tracts that make up America’s 350-plus U.S. metros. Using the conventional proxy for human capital, the share of adults with a Bachelor’s degrees or higher, she measured their segregation across tracts within metros. Mellander used an index of dissimilarity, developed by sociologists Douglas Massey and Nancy Denton, which compares the distribution of a selected group of people with all others in that location (the index ranges from 0 to 1, where 0 reflects no segregation and 1 reflects complete segregation). The MPI’s Zara Matheson mapped the data.
• • • • •
The map below charts the segregation of highly educated people across U.S. metros. The dark blue areas are metros where college grads are the most segregated; lighter blue metros are one where they are more moderately segregated; green areas show metros where they are somewhat less segregated; and the yellow reflect metros where college grads are more integrated with other segments of the population.
College graduates are most segregated in metros along the Eastern seaboard, throughout the South, in the Midwest and Texas, and up and down the Pacific Coast from California up through Washington. Large metros tend to fall into the top two quintiles, while the bottom quintile is mostly made up of smaller metros, especially those in sparsely populated states such as Montana and North Dakota.
The tables below show the ten large metros (those with one million of more people) where highly educated people are most and least segregated from other groups.
|Large Metros Where College Grads Are Most Segregated|
|Rank||Metro||Index||Rank of All Metros|
|2||Houston-Sugar Land-Baytown, TX||0.419||7|
|3||Los Angeles-Long Beach-Santa Ana, CA||0.406||8|
|6||San Antonio-New Braunfels, TX||0.395||12|
|7||Louisville/Jefferson County, KY-IN||0.388||16|
|8||Dallas-Fort Worth-Arlington, TX||0.386||17|
|9||Charlotte-Gastonia-Rock Hill, NC-SC||0.384||20|
The metros where college grads are most segregated are mainly in the Sunbelt and old South, with Birmingham, Alabama, and Houston, Texas, topping the list. The rest of the top ten include Los Angeles, Columbus, Memphis, San Antonio, Louisville, Dallas, Charlotte, and Chicago.
Leading knowledge-based metros rank further down the list of large metros. Washington, D.C. is 17th (45th overall), San Francisco 18th (49th overall), New York 19th (51st overall), San Jose 23rd (55th overall), Seattle 28th (64th overall), and Boston 33rd (86th overall).
When we look at the pattern across all 350-plus U.S. metros, a number of small and medium-sized metros, especially college towns, rise to the very top. State College, Pennsylvania (home of Penn State) has the highest level of human capital segregation of any metro in the country. Salinas, California, is second; Trenton-Ewing, New Jersey (which includes Princeton University) is third; Bloomington, Indiana (home of Indiana University Bloomington) is fourth; and Bryan-College Station, Texas (Texas A&M) is fifth. Birmingham, Alabama, falls to sixth; Houston is seventh; L.A. is eighth; and Columbus, Ohio (Ohio State University) drops to ninth. Blacksburg, Virginia (Virginia Tech) is now tenth overall.
The well-educated are also highly segregated in college towns like Durham-Chapel Hill (University of North Carolina and Duke); Tucson (University of Arizona); Tallahassee (Florida State); Gainesville (University of Florida); Morgantown (West Virginia University); Athens (University of Georgia) and Auburn, Alabama (Auburn University). In most of these places, the economy is sharply divided between professors, doctors, researchers and administrators, and the workers who provide the colleges with basic services.
|Large Metros Where College Grads Are Least Segregated|
|Rank||Metro||Index||Rank of All Metros|
|50||Virginia Beach-Norfolk-Newport News, VA-NC||0.284||190|
|49||Las Vegas-Paradise, NV||0.288||182|
|48||Providence-New Bedford-Fall River, RI-MA||0.290||176|
|47||Hartford-West Hartford-East Hartford, CT||0.294||165|
|46||Minneapolis-St. Paul-Bloomington, MN-WI||0.297||159|
|45||Tampa-St. Petersburg-Clearwater, FL||0.300||155|
|43||Miami-Fort Lauderdale-Pompano Beach, FL||0.316||124|
|42||Buffalo-Niagara Falls, NY||0.317||123|
Conversely, the large metros in which highly educated people are the least segregated include Orlando, Tampa, Miami, and Las Vegas in the Sunbelt, as well as Providence, Hartford, Minneapolis-St. Paul, Rochester, and Buffalo in the Frostbelt.
That said, most large metros register relatively high levels of human capital segregation. There are roughly 170 smaller and medium metros with lower levels of human capital segregation than the least segregated large metro.
The U.S. metros with the least human capital segregation are all smaller places, like St. George, Utah, which has the lowest level of human capital segregation of all. St. George is followed by Lewiston, Idaho; Sherman, Texas; Fond du Lac, Wisconsin; Elizabethtown, Kentucky; Mankato, Minnesota; Great Falls, Montana; Joplin, Missouri; and Barnstable, Massachusetts.
• • • • •
So what separates the places with low levels of segregation by education from those where college grads tend to live apart? What factors are associated with the geographic segregation of the highly educated?
To get at this, Mellander ran a basic correlation analysis between our measure of human capital segregation and a number of key economic, social, and demographic characteristics of metros. As usual, I note that correlation does not equal causation and points only to associations between variables.
For all of the disparities between town and gown in smaller, college towns like State College and College Station, the segregation of highly educated people is greater in larger, denser metros. The geographic segregation of college grads is closely associated with the size of population (.54) and more modestly with density (.39). These kinds of places tend to have greater levels of gentrification and more high-end neighborhoods that price out less skilled workers, thus concentrating the more skilled (and better paid) in their own enclaves.
The geographic segregation of college grads is greater in more affluent regions with significant high-tech industry. It is associated with two measures of regional development – average wages (.34) and economic output per person (.35) – though it is quite a bit less associated with per capita income (.15). The segregation of the highly educated is more closely associated with concentrations of high tech industry (.50), which clusters in locations with highly educated talent. Tech firms also tend to provide highly paying jobs for well-educated workers, so the association with further segregation is unsurprising.
The geographic segregation of highly educated households reflects long-standing racial cleavages. Highly educated people are more segregated in metros with larger shares of black (.34), Latino (.25), and Asian (.24) populations and less so in areas with greater shares of whites (an even larger negative correlation of -.45).
Metros with greater levels of segregation of college graduates also tend to have greater levels of economic inequality, much more so than for the other types of segregation we have examined in this series. Human capital segregation is closely associated with both wage inequality (.55) and income inequality (.55). More specifically, Mellander found that income inequality explains roughly a third of the variation in the segregation of highly educated people across metros in a simple regression analysis.
• • • • •
Education is the most important economic asset a person can have. Children have more opportunity for mobility when they grow up in an area with good schools, a low dropout rate, lots of books, and access to libraries and museums. In contrast, children are far more likely to be entrapped in a cycle of long-run concentrated disadvantage when they grow up in a neighborhood with overcrowded and underfunded schools, a higher dropout rate, and few libraries and other cultural institutions.
My next post, the final in this series, turns to the geographic segregation of knowledge, professional, and creative workers across the United States.
Top Image: AP Photo/Stew Milne