Data Story № 3

Two Ladders
on the Same Street

Take two children, both raised in low-income families, both in the same census tract. If one is White and one is Black, the Opportunity Atlas can tell you — decades later, on average — who ended up higher. In 93 percent of American neighborhoods, it's the White child.

Opportunity Atlas, 1978–83 birth cohorts, 40,452 tracts
p44.8
White kids' mean adult rank
p34.3
Black kids' mean adult rank
92.7%
of shared tracts favor White kids

The Opportunity Atlas is an outstanding resource for anyone interested in studying inter-generational economic mobility across the United States. It is a collection of academic research papers investigating geographic and racial disparities in economic mobility, and a repository of the data used in the studies. Using anonymized tax and census records, the data follows more than 20 million children born between 1978 and 1983 from the census tract where they grew up to their adult incomes in their mid-thirties.

01

The Geography of Opportunity

For each tract, the data reports the average adult income percentile reached by children whose parents stood at the 25th percentile (poor families, but not the poorest).

So, does location matter? For children from similar income families, where they grow up matters a lot for how much they earn in adulthood. Below I replicate a familiar map that demonstrates the extent of geographic disparity in economic mobility in detail. Counties where poor children rose furthest are blue; counties where they stayed poorest are red. The rural Great Plains and Mountain West lift low-income children better than almost anywhere; the Southeast and the old industrial cores hold them down.

p30p55+ mean adult income rank of low-income children
Where poor children rise. Mean adult household-income percentile of children raised at the 25th parental-income percentile, by the county where they grew up. Gray: insufficient data. Source: Opportunity Atlas county outcomes.
02

The Racial Divide

Because the Atlas reports outcomes separately by race, it can answer a deeper question: are racial disparities in income mostly due to where people live, or does the same address provide different opportunities to different children?

Each dot below is one census tract where the Atlas measured both Black and White children from 25th-percentile families. If the same neighborhood provided the same opportunities, the dots would crowd the diagonal. They do not. Nearly the entire cloud hangs below it — these are tracts where White children grew up to out-earn their Black neighbors from families of similar income.

Adult income rank of Black vs. White children from the same tract, same parental income. 5,000-tract random sample of the 40,452 tracts with both estimates. Dots below the diagonal are tracts where White children ended up higher. Source: Opportunity Atlas, kfr_p25 by race.

The vertical distance from the diagonal averages about ten percentiles — and it persists in rich tracts and poor ones, in the high-mobility Plains and the low-mobility South. The researchers report that the divide is driven almost entirely by boys: “In 99% of neighborhoods in the United States, black boys earn less in adulthood than white boys who grow up in families with comparable income.“ The gap is smallest in a few neighborhoods that combine low poverty, high father presence, and low racial bias.

Girls
Boys
The same gap, split by sex. Black vs. White children from the same tract and parental income (25th percentile); 5,000-tract random samples. Among girls the cloud sits closer to the line; among boys it falls well below it — the divide is driven mostly by sons. Source: Opportunity Atlas, kfr_white/black_female_p25 & _male_p25.

Note that same tract does not mean 'same everything'. Within a tract, Black and White families at the same income percentile may still differ in factors such as wealth, schooling, and treatment by institutions. The same-tract comparison removes neighborhood differences, not all differences — that is precisely what makes the residual gap informative.

Below, pooling every measurable tract and comparing across all racial groups yields four distribution curves. Note that the Black and Asian-American distributions barely overlap. Children of low-income Asian-American families reach the highest average ranks; Black children the lowest, with Hispanic and White children in between.

Distribution of tract-level outcomes for low-income children, by race. Each curve is the distribution across tracts of the mean adult income rank; ticks mark group means. Source: Opportunity Atlas.

Splitting White and Black by sex confirms the pattern in the scatter above: the Black–White distance between the mean ticks is wider for sons than for daughters.

Daughters
Sons
Distribution of tract-level outcomes for low-income White and Black children, by sex. Full-sample distributions across U.S. tracts; ticks mark group means. Source: Opportunity Atlas, kfr_white/black_female_p25 & _male_p25.

Place matters a lot, and race is a big part of the story. The same neighborhoods that send poor White children toward the middle class lead their Black neighbors toward a median ten percentiles lower. A ZIP code is a powerful summary of opportunity — but it is not the same opportunity for everyone who lives in it.

Notes & data