Tuesday, July 23, 2013

Geography is destiny?

The New York Times yesterday reported at length on a study of economic mobility by metropolitan area. The results reported in the Times focused on the chances of a child whose family was in the bottom 20 percent of incomes moving into the top 20 percent by age 30. Not surprisingly the chances aren't very great overall (8 percent), but the variation across areas was striking. They ranged from 11 percent in Salt Lake City, San Jose and San Francisco, to 4 percent in Atlanta, Charlotte and Indianapolis. That may not seem an awfully wide range, but note that if all outcomes were totally random 20 percent of people in each quintile would wind up in each quintile. Getting 11 percent of a poor population into the top 20 percent is fairly impressive, while 4 percent is pretty darned close to zero.

[For the record, Chicago's score is a weak 6.1 percent. Cedar Rapids scores 12.6 percent, higher than any of the larger metropolitan areas. I'd call that unbelievably high. I'm not sure 12.6 percent of all Cedar Rapids children are going to wind up in the richest quintile, much less 12.6 percent of poor Cedar Rapids children. There are areas of west Texas and North Dakota that score over 30 percent, which is ridiculous. I'm not sure why these anomalies exist... maybe too small a sample size? Anyway, it suggests treating data from smaller population areas with caution, which I do not do below, but even if I did the points I make still stand.]

The results are so intriguing that it's easy to overlook some details and some important qualifications. So before I comment on the findings, let's address these.

First of all, the "quintiles" are groups of 20 percent of the U.S. population, ranked by family annual income. For the starting point:
  • top quintile (80-99th percentile) = $107,000+
  • second quintile (60-79th percentile) = $73,000-107,000
  • middle quintile (40-59th percentile) = $47,000-73,000
  • fourth quintile (20-39th percentile) = $25,000-47,000
  • bottom quintile (0-19th percentile) = less than $25,000
For the end point, the study took age into account. A 30-year-old reaches the top quintile at $70,000 annual income. For 45-year-olds it's $100,000.

Secondly, movement from the bottom to the top is used as a measure of mobility, but it certainly is not the whole story. One can live very well in the second and middle quintiles, for example, and they can be worthy aspirations; but being truly stuck at the bottom can be cause for despair. Data showing movement from the bottom quintile to the top 60 percent might be more illuminating. Age 30 (chosen because they analyzed a group born in 1980 and 1981) is not necessarily the finish line. It is the age at which many people have begun having families, but if mobility were occurring more gradually--say, if poor children were regularly making it above the median by age 40 or 50--that would still indicate a strong degree of economic opportunity.

The authors of the study were specifically investigating the impact of tax policies on mobility, but found (to their surprise, they said) "only slightly" positive impact of larger tax credits for the poor and higher taxes on the wealthy. What did have an impact:
  • size and dispersion of social classes: are poor families isolated in ghettos, or more dispersed across neighborhoods with a mix of incomes?
  • marriage, or at least the proportion of households with two parents
  • education, quality schools at the elementary and high school levels [The news story quoted Lawrence Katz, an economist, who noted that the areas of highest mobility correspond roughly with areas that established high schools earliest and have a long tradition of strong education]
  • civic engagement, greater membership in religious and community groups
What had little or no impact:
  • number of local colleges
  • cost of college (whew!)
  • extreme wealth
  • overall wealth (e.g. Seattle and Atlanta are similar in average income but Seattle has way more mobility)
  • race (greater mobility for whites is explainable by location, but see below)
The study is a brief for the role of place, particularly city design when we're talking about how the poor are or aren't connected to the parts of the city where the jobs are. (Do they live there already? Do they have access to public transportation? Is traffic prohibitively congested?) As noted in earlier posts, city design can also affect levels of civic engagement... that's the premise behind walkable cities, and the movement to establish third places. For Cedar Rapids, it shows why it's critically important to get the MedQuarter district and the West side right. If they develop in a way that creates jobs and builds connections from the downtown out to the poorer neighborhoods, that will do much to create economic opportunity in this city. If they develop in ways that create even more physical barriers, that will have the opposite effect, and institutionalize those barriers for decades.

Eyeballing the map suggests some relationship to politics, though I confess that relationship is unclear. The Times map has the areas under 5 percent in black, so they're easy to spot. They occur in parts of Alaska, Arizona, South Dakota, Indiana, Arkansas, Mississippi, Alabama, Georgia, South and North Carolina. The 2012 electoral vote total from those ten states was Romney 87, Obama 0. All but Arkansas have Republican governors. [NOTE: Looping in Louisiana and Virginia, which have specks of low mobility connected to commuting zones in Mississippi and North Carolina, respectively, makes it Romney 97, Obama 13, and 11 of 12 Republican governors.] Those states also have higher rates of inequality and poverty in spite of economic growth 2001-07, more child poverty, and higher percentages of women addicted to painkillers. [10/7/13 update: add declining female life expectancy.] Something's going on there, either in policy or culture, that's producing suboptimal outcomes.

That should give us pause before we embrace the Mississippi economic model: low taxes, low public services, little labor or environmental regulation, weak-to-no unions.

Finally, I wonder if the authors were too quick to dismiss the corrosive effects of racial prejudice. They note that whites have higher rates of mobility than blacks, but point out that that's due to where they live. Whites and blacks in the same area (e.g. Atlanta) have the same levels of economic mobility, but more whites tend to live in higher-mobility areas (e.g. Seattle). Race, then, is spuriously correlated to economic mobility. 

But I wonder if, instead, race is an antecedent variable? In plain English, do metropolitan areas have lower social mobility because there are high proportions of blacks living there? It's been pretty well documented that racial diversity affects how whites think about politics. Particularly lower-income whites are less likely to support redistributive policies, even though they'd benefit from them, if they live near blacks. If my mental image of poverty has a black face, is it easier not to pursue policies that create economic opportunity (and the resulting mobility)?


Emily Badger, "A Child in Seattle Has a Much Better Chance of Escaping Poverty Than a Child in Atlanta," The Atlantic Cities, 22 July 2013 [http://www.theatlanticcities.com/jobs-and-economy/2013/07/child-seattle-has-much-better-chance-escaping-poverty-child-atlanta/6275/]

Raj Chetty, Nathaniel Hendren, Patrick Kline and Emmanuel Saez, "The Economic Impacts of Tax Expenditures: Evidence from Spatial Variation Across the U.S.," http://www.equality-of-opportunity.org/.

David Leonhart, "Geography Seen as Barrier To Climbing Class Ladder," New York Times, 22 July 2013, A1, 12.


  1. Very interesting! Thanks for posting. I thought of you and Lynda yesterday when I was listening to a story on NPR about a study suggesting states with more isolated state capitols have a higher rate of political corruption. I haven't read the actual study - or really looked into its credibility (I have a habit (good, bad or otherwise) of believing anything on NPR is credible, although I'm fairly certain they consistently misspelled capitol in the transcript) - but I thought it was interesting nonetheless.


  2. Hi Kyle! I'm glad you found my blog. To defend the honor of NPR, "capital" is the town where the government is headquartered, "capitol" is the building. So their spelling was correct.

    Bruce N.


Bike to Work Day 2018

This year's Bike to Work observation finds me in Washington, D.C., where it's mostly confined to one day, Friday, which I guess i...