The mathematics of school segregation
--
America’s schools are deeply divided by race. That situation perpetuates income inequality between races — white people live in more expensive neighborhoods, white children go to better schools, they end up with better jobs, and they are able to stay in good neighborhoods.
It’s difficult to get to the heart of why school segregation persists. Let’s start the discussion with this Vox article, which tries to pin the problem on gerrymandering of school districts.
The article suggests that school zoning can be used to increase or decrease racial segregation in schools, and provides a fun, interactive data tool to map racial divisions in your own area. After looking at the data for a bit, though, I quickly got the sense that the article was missing the point entirely. When you look at the map of any big city, you see that minorities are clustered in one area of the city while white people are clustered in another. Most big cities have the same pattern, as a few examples, let’s look at Atlanta, Denver, or New York City. Purple shading indicates a mostly black or hispanic neighborhood, white indicates a mostly white neighborhood:
School zoning can only change the size and shape of the small zone around each school, so it can only make a difference in a border region where one racial neighborhood meets another. You can use zoning policy to integrate these border regions, or to make them more segregated. The majority of the city’s schools will still be segregated, though, because the majority of neighborhoods are polarized between whites and minorities, in patterns larger than the size of each school’s population. Here’s a view of Chicago’s current zones vs. zones with children sent to the closest school:
There’s not much improvement in racial integration. So, the article misses the scope of the existing segregation, and it also doesn’t get into some bigger questions — why are neighborhoods so segregated to begin with? Is there anything we can do to fix that?
Here’s another Vox article, from the same author, which discusses that problem directly.
Let’s highlight some of the charts. Here is what whites, blacks, and latinos in Chicago say they want in a neighborhood, vs. what they get in a neighborhood:
And here is a comparison of the racial mix in a person’s ideal neighborhoods vs. their worst choice neighborhood:
The overall racial mix in Chicago is 32% black, 32% white, 28% hispanic, 5% Asian, and 3% other.
The article takes all this evidence and goes with the explanation that white people are racist, they flee from minorities, and that leaves neighborhoods segregated. It would be more accurate to say that each group participates in self segregation, to some degree. White people’s preferences are most out of line with reality, they want a neighborhood that’s at least 50% white, in a 32% white city. Black people want a 40% black neighborhood, in a 32% black city. Latinos’ preferences are almost compatible with reality, but they are also too prejudiced against black people to settle for the average racial mix.
Because each group is unwilling to settle for the average racial mix, they choose to live in completely polarized neighborhoods instead. Here’s a color coded map of Chicago’s racial groups:
Using three colors clearly shows the segregation by neighborhood between blacks, whites, and hispanics. It should be evident by now that this is not simply a case where whites are racist but other groups happily coexist. If you look closely, or zoom in on the source data, you’ll also see small majority Asian neighborhoods.
So far, it sounds like segregation in Chicago arises because no one wants a neighborhood with the same racial mix as the city, so they settle for a heavily segregated racial mix instead. Maybe if we could change people’s preferences a little bit, things would work out and neighborhoods would integrate?
Unfortunately, that might fail too. As we’ll see, moderate prejudices and preferences can reinforce themselves to create big divisions. Consider the simplest case, of a town that’s half black and half white. Suppose everyone in this town is happy if half or more of their closest neighbors are the same race as they are. People in this imaginary case start out randomly scattered around the city, but are free to move if their preferences are not met. You can easily simulate that case (this is called Schelling’s model of segregation). At first, the distribution looks uniform and random:
Some people are happy, but others are unhappy because they don’t like their neighbors, the randomness makes some surrounded by too many dissimilar people. The unhappy people randomly move somewhere else. Within a few moves, everyone ends up clustered by race:
No one individual was particularly racist, in that simulation. Everyone was willing to have up to half of their neighbors be dissimilar, which is more generous than the average person in America is. All it took was a few people unhappy with their neighbors, and big divisions still arose.
So, it seems that both Vox articles are going the wrong direction with their conclusions. The first article blamed government policy for segregation. The second article blamed racist white people running away. A simpler explanation is that it’s an emergent phenomenon that comes from the choices of many individual people, making decisions about the mix of their own neighbors. You don’t need a racist government or a population of incredibly racist people to cause racial segregation, it can still happen with well-meaning, fairly tolerant people who just prefer a majority of similar neighbors.
Imagine another city where the black population is only 10% and the white population is 90%. If blacks in that city also prefer to have at least 40% similar neighbors, then you’ll see black people cluster in one neighborhood. This kind of dynamic plays out often, it’s described at the beginning of that same Vox article, when the author talks about hispanics moving to Worthington, Minnesota. In general, you could say that smaller minority groups tend to be responsible for their own self segregation, while larger minority groups eventually trigger shifts among majority groups by exceeding some tolerance threshold in those groups.
The original question was what we can do about school segregation. It seems that most adults are individually more comfortable with living in racially homogeneous neighborhoods, but some of us are collectively uncomfortable with the segregated schools that this creates.
In the past, some cities have bused students across town to achieve racial integration. This is a blunt but effective way to reintegrate schools. Here’s an article describing the way Seattle used to integrate its public schools by busing, but no longer does so.
There’s only one small problem — when you racially integrate a school, the children tend to self segregate by race, inside that school.
If Vox wrote about this, I’m sure they’d blame racist white kids, but I suspect it’s a tendency of all races. Black kids as young as 3rd graders are encouraged to form all black peer groups, because it makes them more popular than if they have integrated groups of friends.
It’s well known that some white people move their children to avoid integrated schools in mixed districts. However, minorities sometimes do, as well. Black parents in Oakland choose to send their children to lower performing schools with a higher black population, because they feel more welcome than at the majority white schools. Various minorities send their children to charter schools in Minneapolis.
I grew up in a racially mixed school district in California, so I’m personally familiar with these kinds of divisions. I’d guess that my school was 50% white, 35% hispanic, and 15% asian. Like any high school, kids formed cliques, they self segregated by interests and traits. Among the white kids, we had plenty of divisions — jocks, kids who were into country music, stoners, skaters, goths, and nerds (you can probably guess which of those groups I fit into). Every racial group also had a tendency to self segregate, but it was more pronounced between whites and hispanics than it was between whites and asians.
There was a general sense of animosity between whites and hispanics, but it went both ways. You could easily argue that historical white prejudice against hispanics created this animosity, but it certainly seems to be a self perpetuating thing, where neither group trusted the other or was willing to befriend the other.
The same animosity wasn’t as strong, between white and asian kids, though the level of mixing still wasn’t what you’d want. At the very least, if an asian kid was a computer geek, they would hang out with the other computer geeks. I think that’s a reasonable starting goal for racial integration in America — when kids are more free to self segregate by interests than they are by race, then we’re on the path towards racial integration.
As far as academic achievement went, my school had separate tracks (College Placement and Advanced Placement) that high schoolers could go down, which was meant to separate kids by academic ability. The AP classes were generally a mix of white and asian kids. The CP classes were a mix of white and hispanic kids. If you attempt to integrate schools by race, to make children’s academic achievement more uniform, this is another trap that you can fall into — the integrated schools may offer dual tiered educational tracks. Children will likely get sorted by race into higher achieving and lower achieving tracks, perpetuating the same unequal outcomes within the same schools, instead of in separate schools.
Let’s go back to public policy and think again about what we can do about all this. Zoning laws aren’t a strong tool to fix society, they can only create an integration experiment at the boundary between racial regions. Busing students around town can effectively integrate schools, but they’ll still self segregate within schools, and tracking may prevent those schools from educating everyone equivalently.
A few other ideas might work. We could redistribute school funding around districts, so that lower performing schools could get the same funding or more as higher performing ones. We could incentivize better teachers to work at lower performing schools. That might reduce inequality in academic performance. The best you could achieve, with that approach, is a separate but equal world.
To get an integrated and equal world, we would need to reach a point where children and adults of all races are so blind to race that they don’t care who their friends are or who they live next to. To achieve this, we might need active anti-racist education, from an early age, younger than 3rd grade. Another newsweek article describes some of the difficulties in this.
Some parents are overtly racist, and their children learn to mimic these behaviors. Others hope to raise children without racial preferences, but they often fall into a few common problems. In one case, they assume that they can just not talk about race, and children will grow up blind to it. In reality, children do notice race, and come up with their own theories about it if they aren’t explicitly told what to think. In another case, parents try to teach their children about race, but do so in very vague terms, because the specifics are uncomfortable to discuss. One parent in that article continually answered racial questions by telling her child that “all races are equal”. After hearing this for months, her young child eventually asked the simple question, “what does equal mean?”
At this point, I might say that this situation sounds hopelessly complicated. One positive point is that some children and adults wield more influence than others. Maybe you only have to convince the most popular kids that integration is cool, and have some popular role models in entertainment and politics, and everyone else should eventually follow.
If that doesn’t work, I guess we could still try the solution proposed in the movie Bulworth (video is NSFW):
