Trump’s stunning and unexpected victory in last week’s election highlighted the overlapping divides of class and location in American politics. Almost a decade ago now, Bill Bishop pointed to the ‘Big Sort’ that is dividing Americans by income, education, and where we live. As the election showed, that big sort has become an even bigger sort as America’s middle class has declined and the nation has split into areas of concentrated affluence and even larger spans of concentrated disadvantage.
The steady erosion of America’s once sizable middle class is staggering. The middle class share of the population shrunk in 203 of 229 U.S. metros between 2000 and 2014, according to a detailed study from the Pew Research Center. Nationwide, 172 of 229 metros saw growth in affluent, upper-income households in the past decade and a half; 160 saw an increase in the share of low-income households; and roughly half, 108, experienced both.
The decline of the middle class is the key factor in America’s deepening divide between rich and poor. The share of American families living in middle class neighborhoods fell from nearly two-thirds (65 percent) in 1970 to 40 percent in 2012, according to a recent study by Sean Reardon and Kendra Bischoff. At the same time, the share of American families living in either all-poor or all-affluent neighborhoods more than doubled, increasing from roughly 15 percent to nearly 34 percent.
To better understand the geography of middle class decline, my colleague Charlotta Mellander and I took a deep dive into the Pew data, which charts the transformation of the American class structure across 229 metros between 2000 and 2014.
For the purposes of this analysis, the “middle class” is defined by households with incomes two-thirds to twice the median household income, adjusted for household size. A single person needs to earn between $24,000 and $72,000 to be considered middle class; two people have to earn between $34,000 and $102,001; a three-person household must make between $41,641 and $124,925; a four-person household between $48,083 and $144,251; and a five-person household between $54,000 to $161,000. Lower-income households include those with incomes less than two-thirds the median; the upper-income or upper class includes households with incomes that are more than double the median.
All of these figures are adjusted for differences in metro area costs of living. Taylor Blake of the Martin Prosperity Institute (MPI) generated the maps: Lighter blue indicates losses, while purple indicates gains.
The first map shows the simple change in the middle class across U.S. metros. The map is a patchwork of sorts. The biggest losses can be seen across parts of the East Coast, the Midwest, and Rocky Mountain regions, while the biggest gains are in California on the West Coast, parts of Texas and Florida, and some parts of the East Coast and Midwest.
But, what really stands out is that every single large metro (over one million people) saw its middle class decline. In fact, less than 10 percent of all U.S. metros saw any increase in their middle class whatsoever, with most of these gains being in the range of one percent.
The ten large metros with the biggest middle class decline are a mix of knowledge hubs such as Boston and Washington, D.C.; Sunbelt metros such as Orlando and Charlotte; and Rustbelt metros such as Milwaukee. That said, smaller metros such as Niles-Benton Harbor, Maine; Goldsboro, North Carolina; Springfield, Illinois; and Midland, Texas; plus college towns like Burlington, Vermont; Champaign-Urbana, Illinois; and New Haven, Connecticut saw even larger losses in their middle class.
Declining Economic Status
The next map is based on a broader measure which looks at the overall change in the share of people moving into the upper and lower classes. Specifically, this net change measure gauges the percentage point change in the share of adults who were upper income minus the change in the share who were lower income between 2000 and 2014. Across the nation as a whole, the share of upper class adults increased from 17 percent in 2000 to 20 percent in 2014, while the lower class share increased from 28 percent to 29 percent. This measure varies substantially across metros.
At the metro level, the biggest net gains can be seen across the Acela Corridor on the East Coast and California and the Pacific Northwest on the West Coast, along with parts of Texas and other small areas of the country, while the biggest net losses are in the Midwest and South.
|Declining Economic Status|
|Atlanta-Sandy Springs-Roswell, GA||-8.3%|
|Las Vegas-Henderson-Paradise, NV||-5.2%|
|Milwaukee-Waukesha-West Allis, WI||-5.1%|
|Kansas City, MO-KS||-5.0%|
Detroit experienced the biggest loss of economic status of any large metro, followed by Atlanta, Charlotte, Raleigh, and Orlando. Las Vegas, Milwaukee, Kansas City, Indianapolis, and Cleveland round out the list of large metros with the biggest losses on this metric. In fact, all ten are located in the Rustbelt or Sunbelt metros. But smaller metros mainly in Michigan (Jackson, Monroe, Muskegon), Ohio (Springfield, Mansfield), Indiana (Michigan City, Fort Wayne), and North Carolina (Goldsboro, Hickory, Rocky Mount, Winston-Salem, Burlington, Greensboro), saw the biggest losses in economic status.
|Improving Economic Status|
|New Orleans-Metairie, LA||12.7%|
|Oklahoma City, OK||8.0%|
|Virginia Beach-Norfolk-Newport News, VA-NC||6.1%|
|San Antonio-New Braunfels, TX||5.7%|
|Houston-The Woodlands-Sugar Land, TX||4.6%|
The metro that saw the biggest gain in its economic status was New Orleans, a consequence of the loss of low-income residents in the wake of Hurricane Katrina. This group also includes resource-based metros such as Houston and Oklahoma City; knowledge hubs such as Boston and D.C., and reviving industrial metros such as Pittsburgh, Providence, and Baltimore. The metros that saw the biggest gains were largely smaller oil metros in Texas (Odessa, Midland, Amarillo, San Angelo and Corpus Christi) and Louisiana (Lafayette, Baton Rouges) as well as Barnstable, Massachusetts; Lewiston, Maine; Johnstown, Pennsylvania; and Albany, New York. As the Pew study points out, these are not necessarily metros with lots of upper income. They saw big gains mainly because they grew their upper class shares from relatively low bases.
Metros with the Largest and Smallest Middle Class
Which metros have the largest and smallest middle classes overall? The next map charts the pattern.
The large metros where the middle class is smallest are a combination of superstar cities, tech hubs, resource economies and poorer places. L.A. has the smallest middle class overall, followed by San Francisco, New York, and San Jose. Houston, Miami, Boston, Sacramento, New Orleans, and Hartford round out the top ten. That said, the places with the smallest middle classes are mainly smaller metros such as Monroe, Louisiana; Midland, Brownsville, McAllen, Laredo, and El Paso, Texas; Bakersfield, Fresno, Visalia, and El Centro, California; as well as college towns such as Auburn, Alabama; Champaign, Illinois; and Morgantown West Virginia.
|Smallest Middle Class|
|Los Angeles-Long Beach-Anaheim, CA||46.5%|
|San Francisco-Oakland-Hayward, CA||47.7%|
|New York-Newark-Jersey City, NY-NJ-PA||48.1%|
|San Jose-Sunnyvale-Santa Clara, CA||48.5%|
|Houston-The Woodlands-Sugar Land, TX||48.5%|
|Miami-Fort Lauderdale-West Palm Beach, FL||48.5%|
|New Orleans-Metairie, LA||49.2%|
|Hartford-West Hartford-East Hartford, CT||49.3%|
Conversely, the metros with the largest middle classes are a combination of Sunbelt and Rustbelt places. Salt Lake City has the largest middle class of any large metro, followed by Rochester, New York; Virginia Beach, Las Vegas, Kansas City, Indianapolis, Minneapolis-St. Paul, Pittsburgh, Richmond, and Louisville. The places with the largest middle classes overall are all smaller, industrial or working class metros, mainly in Wisconsin (Wausau, Janesville, Sheboygan, Eau Claire), Indiana (Elkhart-Goshen, Fort Wayne), Pennsylvania (Lebanon, East Stroudsburg), Ohio (Youngstown, Canton), and Utah (Ogden, Provo).
|Biggest Middle Class|
|Salt Lake City, UT||58.2%|
|Virginia Beach-Norfolk-Newport News, VA-NC||56.1%|
|Las Vegas-Henderson-Paradise, NV||55.9%|
|Kansas City, MO-KS||54.8%|
|Minneapolis-St. Paul-Bloomington, MN-WI||54.6%|
|Louisville/Jefferson County, KY-IN||54.1%|
What’s Behind Middle Class Decline?
Why do some metros have a bigger or smaller middle class? What are the key motivating factors at work here?
To get at this, Mellander ran a basic correlation analysis between the middle class share, as well as the upper and lower class shares, of metros and their key economic, social, and demographic characteristics. As usual, I note that correlations like these do not imply causality and simply point to associations between variables.
We can start with the factors that are associated with the geography of the middle class. The middle class share of the population is associated with the working class (.37), the share of the population that is white (.34) and political conservatism (measured as the share of voters who went for Romney in 2012, .19). These are also all features of economically declining places.
On the flip side, the middle class share of the population is negatively associated with density (-.22), college grads (-.14), the creative class (-.20), left-leaning politics (with a correlation of -.19 to the share of Obama voters in 2012, -.19), and diversity (with correlations of -.46 to immigrants, -.47 to Latinos, and -.37 to the gay and lesbian community)—all key features of economically more vibrant places.
Simply put, a big part of our national economic dilemma is this: the middle class is larger in declining places and smaller in growing ones.
This is especially troubling as a large middle class remains a bulwark against rising economic inequality. The middle class share of the population is negatively associated with income inequality (-.64), wage inequality (-.38), economic segregation (-.43), and having a large lower class share of population (-.62).
Worse, metros that had bigger middle classes in 2000 are the ones that have seen the largest middle class declines by 2014. We find a negative and statistically significant correlation (-.48) between the middle class share in 2000 and the change in the middle class share between 2000 and 2014.
A look at the factors that bear on the two other classes—the upper and lower classes—can help us better understand the geography of middle class decline and America’s deepening divide into rich and poor places.
Lower-class metros are smaller (with a correlation of -.19 to population size) and have larger Latino (.45) and immigrant populations (.33). They have lower levels of college grads (-.45), and the creative class (-.34), high-tech industry (-.35) and well-being (-.43). They suffer from higher levels of income inequality (.29) as well.
Upper class metros are larger (.37) and denser (.31); more knowledge based, with positive correlations to college grads (.70), high tech industry (.56) and the creative class (.60). They also suffer from higher housing costs (.49), greater wage inequality (.48), and higher levels of segregation (.36).
Our analysis also quantifies the stark divide between America’s upper and lower class places: Across metros, the correlation between upper and lower income shares of the population is a whopping -.72. In other words, increasingly, metros are either one or the other.
In the U.S. today, economic inequality is also spatial inequality. Rich and poor increasingly occupy entirely different spaces and worlds. Understanding this and developing solutions for rebuilding the nation’s fading middle class may well be the defining issue of our time.