Research Highlights Podcast
May 15, 2026
The wrong side of the tracks
Eric Chyn discusses the effects of racial segregation on long-run economic outcomes.
Source: Tracks_Rails from Chicago, United States, Public domain
The place where a child grows up in America shapes their economic future to a significant degree. One long-suspected explanation is racial segregation, but proving whether segregation actually causes worse outcomes—rather than just correlating with them—has been challenging for economists.
In a paper in the American Economic Journal: Applied Economics, authors Eric Chyn, Kareem Haggag, and Bryan A. Stuart provide evidence that racial segregation shapes the long-run economic prospects of American children.
Using the placement of railroad tracks in the 19th century, they found that a one standard deviation increase in segregation—roughly the gap between Minneapolis and Philadelphia—cost a Black child from a poor family about $4,200 a year in income as an adult. While lower-income Black children were hit the hardest, segregation also hurt higher-income Black children and lower-income White children.
Chyn recently spoke with Tyler Smith about why segregation hurts low-income kids in particular and what his findings imply for policymakers.
The edited highlights of that conversation are below, and the full interview can be heard using the podcast player.
Tyler Smith: There's been a lot of work among economists, yourself included, to try to figure out why kids born in one American neighborhood end up with very different life prospects than kids born maybe just a few blocks away. This work is in that vein, but you look at segregation in particular. Why did you want to focus on segregation as a source of the disparities we see in the United States?
Eric Chyn: A lot of my prior work, just to give you a little bit of the origin story, has been on why kids in certain neighborhoods within a city have such different outcomes than other kids, even in the same city. Specifically, I've done work on public housing, looking at programs that help move children to more advantaged areas. I've learned a lot and been fascinated by that topic, but a natural related question when you're studying neighborhoods and these moving programs is how segregation operates across a whole city, which is itself defined by many neighborhoods. In lots of cities, we see this tremendous sorting by race and income. Do the more segregated cities really have different outcomes for the children living there? It's a natural extension of the other work I've done.
Smith: You use the placement of railroad tracks in the nineteenth century as an instrument for later segregation. Can you walk me through that and help me understand why where a track was laid in the nineteenth century tells us something causal about a child's outcomes today?
Chyn: The core problem with just looking at correlations between segregation and children's outcomes is confounding. Cities differ from each other on so many dimensions—industrial history, political culture, labor markets, housing policy—and many of those things can both make a city more segregated and independently affect how kids grow up there. So if you just compare more segregated cities to less segregated ones, you can't tell what's actually driving the difference in outcomes. What the railroad geography gives us is a piece of segregation that's plausibly disconnected from those other city characteristics. The reason is essentially history. The railroad tracks were laid in the nineteenth century for reasons having to do with terrain and trade routes between cities, and this was decades before the Great Migration, decades before any of the modern segregation we observe today. But when Black migrants started arriving in these northern cities decades after the railroads were put in, those tracks ended up serving as coordinating devices for residential settlement patterns and segregation. This is the literal origin of the term "living on the wrong side of the tracks." So when we use the nineteenth-century railroads and how they chopped up a city as a source of variation, we're isolating a piece of segregation that was determined long before any of these cities had the racial demographics or the political dynamics that we're studying today.
Smith: You look at several different outcomes, but the main outcome of interest in this paper is intergenerational mobility. What do you mean by intergenerational mobility and how did you measure it?
Chyn: What we use is data from the Opportunity Atlas. This gives us measures of children's income rank within the distribution of national income. You take a child, look at their earnings in adulthood, and then you can rank all children on a measure of how well they're doing later on in life. We're using the standard income rank measures that the Opportunity Atlas has provided to lots of researchers, and that's really one of the key reasons we’ve innovated in this paper, because this data didn't exist decades ago. Now we can take advantage of it to revisit some of these old questions, like what are the effects of segregation, but through the new lens of trying to understand long-run consequences for kids.
Smith: Turning to your results, what did you find? What was the impact of segregation?
Chyn: Using this research design that revolves around comparing cities that are more or less segregated because of railroads, we look at a lot of different long-run outcomes for kids. We start by looking at measures of labor market outcomes and earnings, and we do this separately for African American kids and white kids. The units we use in the paper are about a one standard deviation increase in segregation. You might ask, what is a one standard deviation increase in segregation? Two real-world examples: if you compare Minneapolis and Philadelphia, Minneapolis is much less segregated than Philadelphia. So this is roughly comparing kids who grew up in Philadelphia to kids growing up in Minneapolis. What we see in our analysis is that for low-income Black children at the first percentile, a one standard deviation increase in segregation reduces their annual income by about $4,000—about 16 percent of income for that particular group. But even when you look at more advantaged Black children, for example Black children at the 75th percentile of family income background, you still see that they are losing ground, and a one standard deviation increase in segregation reduces their income by about 7 percent. That's really the story—broadly across the income distribution for Black kids, you see these reductions in income. But when we turn to White children, we see some heterogeneity. The most disadvantaged White children, those coming from the lowest income households, also take a hit from living in a more segregated city—a reduction of about 9 percent in annual earnings. Through the rest of the distribution of White children, we see less tangible negative impacts.
We're seeing a real big difference in public good availability in cities that are more segregated versus less.
Eric Chyn
Smith: Why is segregation having this big effect on lower-income kids in particular?
Chyn: There are a number of different things we do in the paper to try to unpack that, but one of the most compelling parts of our analysis looks at a very natural set of outcomes for a place, which is its public goods. We look at spending on government programs, schools, things like this. One of the big headline findings is that more segregated cities have lower per capita government spending. That is being driven primarily by the schooling channel—they're spending less on schooling in these areas. So there's this critical aspect of public good provision that's different across places that are more or less segregated, and that's one of the leading reasons why, in particular, we're thinking about lower-income children. They might be the ones who critically rely on the provision of public goods in the cities they live in. We're seeing a real big difference in public good availability in cities that are more segregated versus less.
Smith: Given what you've learned from this work, and some of your previous work, what can we say about trying to improve the outcomes of kids growing up in these disadvantaged neighborhoods?
Chyn: Policy implications are super important whenever you're studying a topic like segregation. The first thing I want to say is that although segregation has been declining since 1970, it's still a defining feature of most US cities. With that point in mind, our results suggest that reducing segregation, or, alternatively, policies that may offset its effects on public spending and public goods, could meaningfully improve mobility. Our results suggest that policies that reduce residential segregation or policies that offset the impacts of segregation on public spending and public goods—generally supporting the provision of public goods in cities that are very segregated—could meaningfully improve economic mobility for children. That's how I interpret a lot of this—through the lens of segregation and its policy consequences.
♦
“The Effects of Racial Segregation on Intergenerational Mobility: Evidence from Historical Railroad Placement” appears in the April 2026 issue of the American Economic Journal: Applied Economics. Music in the audio is by Sound of Picture.