Tuesday, April 28, 2009

The feedback loop: segregation and education in the last 30 years

A study shows that clustering of income classes indicates that the U.S. is increasingly segregated now compared to previous decades. This is problematic, as a positive feedback loop develops, and the segregation spreads. One example is present in the Census data released on education: educational attainment by race, has improved only slightly since 1969; a wedge is still quite evident at the bottom and top tails of the distributions.

A visible pattern of income-class clustering (living in groups) emerged. From the WSJ's Real Time Economics blog (Mark Thom provides an excerpt of the paper here):
America is a more segregated society today than it was in 1970 — when one looks at the tendency of rich, poor and middle class to cluster together, says economist Tara Williams, who studied 216 cities and finds that the bulk of this new segregation occurred during the 1980s at the same time as the gap between the incomes of rich and poor widened substantially.


“As inequality increases, it becomes less likely that rich and poor households are willing” — or able — “to pay similar amounts for a given set of neighborhood amenities,” she says. “As income inequality rises, the rich will be more likely to outbid the poor for high-quality neighborhoods and the rich and the poor will be less likely to live in close proximity.”

RW: A tendency for income classes to cluster could magnify the segregation across the economy, affecting, earnings, job availability, mobility, and overall welfare. And a report by the Census Bureau shows just that: educational attainment by race, has improved slightly; however, a wedge is still quite evident at the bottom and top end of the distributions.

by race in
As always, click to enlarge

The chart illustrates the percentage of the White and Non-white non-institutional population that achieved each level of education in 1969. The White population achieved a higher level of education across all categories of higher education.

The disparity across race is especially evident at the bottom and top end of the distributions. In the Non-white population, 39% achieved just an elementary school level of education, and 12% attained at least some college education, with 5% receiving a college degree. On the other hand, a lower 25% of the White population achieved just an elementary school level of education, while 20% attained at least some level of college education, with 9% (almost double the Non-white) receiving a college degree.

Education by race in
2008. Note: since the 1969 report breaks down race into 2 categories, White and Non-white, I present the 2008 data in the same manner. However, the Census now follows four categories of race.

The chart above shows that in 2008 a similar pattern to that in 1969. There is a larger share of the Non-white population that achieved less than a high-school education (1st-11th grade), 25% Non-White versus 9% Non-Hispanic White, and a smaller share that attained a Bachelor, Master's or Professional, or Doctorate level of education, 20% total Non-white versus >30% total Non-Hispanic White.

According to the Census report,
"workers with a high school degree earned an average of $31,286 in 2007, while those with a bachelor’s degree earned an average of $57,181."
That's a $26k premium on earnings if a worker has a B.A. Hmm....

To be sure, the size of the 2008 non-white population has increased substantially since 1969; and furthermore, the 1969 data includes data on the population aged 14 yrs and older, while the 2008 statistics consist of the population aged 18 yrs and older. However, the trend is clear: there are potentially harmful feedback loops between income inequality and education by race.

The negativepositive feedback loop: income inequality - disparate education opportunities - diverging relative earnings - income inequality.

Rebecca Wilder


  1. Is this limited to only those born in the US? Because, if it includes the entire population, immigrants (1st generation) will show the education level, in most cases, of their country of origin. That would explain the larger uneducated group. Some will take classes but most don't have the time/money and rather encourage their children to go to school.

  2. Hi Janie,

    The data span the non-institutional population, which includes foreign-born residents and citizens.

    As you point out, a more quantitative analysis must account for measurement issues, especially since the Non-white population is so much larger in 2008 than it was in 1969.

    Thank you for a great comment!


  3. Greater than in 69?

    I don't believe the figures. Maybe this is true in other cities. You need a pie chart that distinguishes between all ethnic groups. Something else is skewed? I.e, this is not a race distribution but an economic distribution.

    Also, isn't there a trend to move back to the city (as opposed to stay the burbs)?

    Clustering? It's logical that as the economic disparity between classes grows so will segregation. But that may be for another reason. This is also the way homes and neighbor hoods are designed or built (in mass).

    Also it is logical that education separates the income classes. But who is going to school. And why single out ba's ma's ph.d's?

    Aren't the community colleges booming? And who is attending these? There are a lot more foreign students. Maybe also it now costs relatively more money to go to school.

    What good are stats that separate the classes between whites & non-whites? How are the boundaries defined?

  4. Also, was the population growth between economic groups considered?

  5. Take yourself back to the race riots. You have no idea about segregation by economic class.

  6. Flow5,

    Come on, it's just a qualitative assessment of the data - I am not "modeling" an empirical test. Clearly, as I state, there are a lot of economic issues that need to be filtered out before assessing the statistical significance of education and clustering. However, the qualitative differences are certainly suggestive.

    Also, the clustering argument is not mine - that is a paper at the NBER.


  7. It is an important topic and one that Bernanke has spoken about (i.e., the rest of the world has progressed beyond us).

    One reason for the divergence in the 80's wast that the inner cities were decimated by drugs. In fact, the churches fed more indigents in the 80's than in this recession/depression.

    This divergence was more of a social phenomenon than an economic one (stagflation).

    The drug war has been won and now there are a lot of ambitious kids.

    But I still find it telling that the porches were built in the front in the old days, and now the decks are made in the back of the house.

  8. Here at the U.S. Chamber Education and Workforce Summit, I don’t really need to convince you that, as an investment, education provides excellent returns, both for individuals and for society. As executives accustomed to making hard cost-benefit decisions, you doubtless assign a high priority to the quality of your business’s workforce because you know that a key--perhaps the key--to your success is the capabilities of the people you employ. To a significant extent, those capabilities are the product of education. Here I am speaking not just of education acquired formally in classrooms before entering the workforce but also of lifelong learning that, yes, includes the formal classroom training that might first come to mind but that also includes early childhood programs, informal mentoring on the job, and mid-career retraining, to name a few examples. And when I speak of capabilities, I mean not only the knowledge derived from education but also the values, skills, and personal traits acquired through education, which are as important as, and sometimes even more important than, the specific knowledge obtained. These include such qualities as the ability to think critically, to communicate clearly and logically, and to see a project through from start to finish.

  9. "The mind has an enormous capacity for error, self-deception, illogic, sloppiness, confusion and silliness. All of these tendencies may be diminished by training, and that, of course, is the function of education"

    John Gardner - former Secretary of Health Education and Welfare, May 12, 1963