Lasting Effects of Segregation on Political Behavior and Economic Opportunity

Featured in print Reporter
By Eric Chyn and Kareem Haggag

Segregation based on race and income is a defining feature of cities and schools across the United States. While Black Americans make up less than 14 percent of the overall population, the typical Black child lives in a neighborhood where Black families make up the majority of residents and attends a school where at least half their peers are also Black.1 These neighborhoods and schools also tend to have relatively high rates of poverty.

Theory posits that segregation in terms of neighborhoods and schools plays important roles in understanding poverty for disadvantaged Black communities.2 This segregation may also shape the development of White residents’ attitudes, behaviors, and outcomes.

Although individual outcomes are notably correlated with exposure to segregation in neighborhoods and schools, it has proven difficult to assess whether these descriptive relationships reflect independent causal impacts. For example, descriptive work robustly documents that minorities who live in or attend schools in areas of concentrated poverty have worse economic outcomes than their counterparts who live outside these areas. It is well recognized that these patterns are difficult to interpret since differences in outcomes may reflect unmeasured differences in family background that may drive residential selection or school choice.3

During the past decade, a large body of research has exploited natural experiments to provide new causal evidence on the effects of segregation in schooling and neighborhood contexts. In a recent set of papers, we have added to this literature by studying how segregation shapes civic participation and political identity, as well as broader social policy and economic opportunity. Our work relies on newly available data sources and demonstrates that segregation in cities and schools has wide-ranging impacts on both political and economic behavior. This piece summarizes our key findings.

Can Moving to Higher-Opportunity Neighborhoods in Childhood Increase Civic Engagement?

An important body of research shows that childhood neighborhoods exert powerful influences on later-life economic outcomes,4 but there has been relatively less evidence on the links between childhood residence and social outcomes. Political engagement is an important outcome for which existing theory suggests neighborhoods play a critical role. For example, growing up in a disadvantaged community may reduce an individual’s access to the types of institutions, civic norms, or social networks that foster engagement with the political process and voting.

We provide new quasi-experimental evidence on the link between childhood neighborhoods and later-life political behavior.5 Our analysis is based on a natural experiment created by public housing demolitions.6 During the 1990s, the Chicago Housing Authority (CHA) began a process of destroying high-rise public housing, prioritizing closing buildings that had histories of poor maintenance. Households living in buildings selected for demolition received housing vouchers and relocated to lower crime, higher income neighborhoods. We compare the long-run voting behavior of people who as children were displaced by public housing demolition to that of their peers who lived in nearby public housing buildings that were not destroyed.

We find that relocating to lower-poverty areas due to public housing demolition has large impacts on measures of political participation. Our analysis is based on linking administrative records of displaced and nondisplaced children to statewide voter registration records from Illinois. As shown in Figure 1, displaced children were 2.3 to 2.8 percentage points — 13 to 16 percent — more likely to vote in the presidential elections of 2008, 2012, and 2016. Overall, we find that displaced children were 3.3 percentage points — 12 percent — more likely to vote in any general election up to 2018. We also find that registration increased by 2 percentage points — 5 percent — for displaced children, demonstrating that part of the overall effects were due to new voters. These results suggest that childhood neighborhood inequality casts a long shadow over access to the political process.7

This figure is a vertical bar graph titled, Voting Participation of Displaced and Nondisplaced Children. The y-axis is labeled, mean voting rate. It ranges from 0 percent to 25 percent, increasing in increments of 5.  The x-axis represents year, ranging from 2008 to 2018, increasing in increments of 2 years.  There are two categories of bars: nondisplaced children and displaced children. The bars for displaced children are consistently higher than those for non-displaced children in both presidential and midterm election years. During presidential election years, the percentage for displaced children is around 20 percent, compared to approximately 17 percent for non-displaced children. In midterm election years, there's a steady increase for both groups, from about 5 percent in 2008 to nearly 10 percent in 2018. The gap between the two groups varies between 1 to 3 percent across the years, with the differences being more pronounced during presidential election years. The note on the figure reads, bars represent 95% confidence intervals. The source line reads: Source: Eric Chyn and Kareem Haggag. NBER Working Paper 26515.
Figure 1

What are the implications of these findings? Recently, housing authorities and policymakers have introduced new housing counseling programs, such as the Creating Moves to Opportunity program in Seattle and King County, and have reformed housing voucher payment caps to encourage moves to more-advantaged neighborhoods. While such policies are primarily motivated by a desire to improve the economic circumstances of poor families, our results suggest these policies may have the added benefit of increasing long-run involvement in the political process, which may further feed back into political and economic outcomes.

Does Intergroup Contact in Schools Matter?

Prominent theories across the social sciences posit that racial attitudes are importantly shaped by exposure to racial out-groups. According to contact theory, meaningful interactions with racial minorities can reduce bias among majority group members.8 On the other hand, the “racial threat” hypothesis posits that Whites’ proximity to minorities can trigger hostile attitudes as Whites perceive their status as threatened.9 While there has been a flurry of tests of real-world intergroup contact in recent years, understanding the long-run effects of sustained intergroup contact in early life — the period before preferences and beliefs can be calcified — has been a challenge.10

In work with Stephen Billings, we provide novel evidence on the long-run effects of changes in intergroup contact due to reforms that reshaped schooling environments in a large urban school district.11 In 2002, Charlotte-Mecklenburg Schools (CMS) ended race-based busing and redrew existing school district boundaries. As a result of these reforms, some White students attended schools with varying percentages of Black students. In our work we ask whether the racial composition of schools attended during childhood affects one’s political behavior more than a decade later. To answer this question, we assembled data linking administrative schooling records from CMS and voting registration records from North Carolina.

Comparing party affiliation in adulthood of students who lived in the same neighborhood but were assigned to schools with a different racial composition after the reforms, we find that a 10 percentage point increase in the share of minorities in a White student’s assigned school decreased their likelihood of registering as a Republican by 2 percentage points (12 percent). Our results also provide suggestive evidence that White students are more likely to be registered as Democrats or unaffiliated voters. What is the link between exposure to minorities and party affiliation? Prior research has robustly documented a tight link between racial attitudes and Republican Party affiliation in the US.12

These results may have broader implications. Hundreds of school districts have been released from court-ordered desegregation during the past 30 years — a policy shift that has led to the gradual resegregation of schools within these districts.13 In addition to any negative effects of school segregation on economic outcomes, our estimates suggest that these policy changes could have led to important shifts in the partisan identities of Americans.14

Citywide Consequences of Residential Segregation

What impact does racial segregation have on the local political and policy environment? Residential segregation in cities implies segregation in schools, workplaces, and a range of avenues through which groups may have otherwise interacted. As we’ve noted above, a lack of intergroup contact can foster stereotyped and prejudicial views among majority group members. Given the documented link between racial attitudes and support for inequality-reducing programs, it is possible that segregation translates into impacts on redistributive social policies.15

In recent work with Bryan Stuart, we analyze the causal effects of citywide racial segregation with a focus on both political and economic outcomes,16 building on pioneering work by Elizabeth Ananat that introduced a novel instrument for racial segregation based on historical railroad placement.17 The strategy relies on the fact that cities that were subdivided to a larger extent by railroads during the nineteenth century became more segregated after the arrival of Black residents in the North and Midwest during the Great Migration. We apply this quasi-experimental research design to newly available large-scale surveys on political and racial attitudes as well as data on local government expenditures.

Figure 2 summarizes our key results on the effects of racial segregation across 121 cities. A 1 standard deviation increase in a city’s segregation rate worsens an index measure of White residents’ racial resentment by 0.69 standard deviation, echoing prior work that relies on related measures.18 Moreover, we find that these changes in attitudes are accompanied by lowered self-reported measures of support for redistributive policies, including minimum wage increases and social program spending. These survey-based findings appear to be matched by realized measures of government expenditures. We find that a 1 standard deviation increase in racial segregation results in a 39 percent decrease in total per capita government expenditures. This reduction is widespread, affecting vital areas such as education (38 percent of the total decrease), welfare and health, infrastructure, and public safety. Across all these measures of attitudes, we find that the results for Black residents move in the opposite direction.

This figure is a three panel vertical bar graph titled, Segregation, Racial Attitudes, Government Spending, and Income Mobility. It is subtitled, Bars show the effect of a one standard deviation increase in city’s segregation rate. The following pertains to the left-side panel. The y-axis is labeled, standard deviation change in attitudes. It ranges from negative 1 to 0, increasing in increments of 0.2. The x-axis has two categories, Racial and Redistribution. The Racial bar has a value of negative 0.6 and redistribution bar has a value of near negative 0.4.  The following pertains to the middle panel. The y-axis is labeled, standard deviation change, 1990 per capita dollars, $1000. It ranges from negative 1 to 0, increasing in increments of 0.2.  The x-axis has one category, government expenditures. The value of its bar is nearly negative 0.6. The following pertains to the right-side panel. The y-axis is labeled, change in upward mobility, income percentiles. It ranges from negative 0.06 to positive 0.02.  The x-axis has 4 categories, listed in order: White 25th, Black 25th, White 75th, and Black 75th.  The White 25th bar has a value of over about negative 0.02 and the Black 25th bar is nearly double that. The White 75th bar is nearly 0 while the Black 75th bar is about negative 0.03.  The following pertains to all panels. The note on the figure reads: Bars represent 95% confidence intervals. The source line reads: Source: Eric Chyn, Kareem Haggag, and Bryan A Stuart. NBER Working Paper 30563.
Figure 2

Finally, our work also traces the effects of residential segregation on economic opportunity based on newly available data on upward mobility by race and parental income from the Opportunity Atlas.19 For a child whose parents are at the 1st percentile of the nationwide income distribution, we find that a 1 standard deviation increase in racial segregation leads to a 4.5 percentile decline in the child’s long-run income rank, which amounts to 17 percent of the average mobility for this group. Since Black children born to parents in the 1st percentile end up in the 27th percentile — $17,500 in annual household income — on average, a drop to the 22nd percentile amounts to $4,834 in lost income each year. The negative effects of segregation on mobility are also sizable and statistically significant for Black children whose parents have income at the 25th, 50th, and 75th percentiles of the distribution. For White children, we find evidence of heterogeneous impacts with segregation worsening outcomes for those from lower-income households and benefiting children from the very top of the distribution.



 “Black-White Segregation Edges Downward since 2000, Census Shows,” Frey WH. Brookings Institution, December 17, 2018. “US Public School Students Often Go to Schools Where at Least Half of Their Peers Are the Same Race or Ethnicity,” Schaeffer K. Pew Research Center, December 15, 2021.


The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy, 2nd ed., Wilson WJ, editor. Chicago: University of Chicago Press, 2012.


The Social Consequences of Growing Up in a Poor Neighborhood,” Jencks C, Mayer SE. In Inner-City Poverty in the United States, Lynn Jr. LE, McGeary MGH, editors, pp. 111–86. Washington, DC: National Academy Press, 1990.


 “Neighborhoods Matter: Assessing the Evidence for Place Effects,” Chyn E, Katz LF. NBER Working Paper 28953, June 2021, and Journal of Economic Perspectives 35(4), Fall 2021, pp. 197–222.


Moved to Vote: The Long-Run Effects of Neighborhoods on Political Participation,” Chyn E, Haggag K. NBER Working Paper 26515, November 2019.


 Moved to opportunity: The long-run effects of public housing demolition on children,” Chyn E. American Economic Review 108(10), October 2018, pp. 3028–3056.


These effects are in addition to the various other structural barriers faced by Black Americans in the political process, e.g., the much longer average voting wait times documented in “Racial Disparities in Voting Wait Times: Evidence from Smartphone Data,” Chen MK, Haggag K, Pope DG, Rohla R. NBER Working Paper 26487, November 2020, and The Review of Economics and Statistics 104(6), November 2022, pp. 1341–1350.


The Nature of Prejudice, Allport GW, editor. Addison-Wesley, 1954.


 Southern Politics in State and Nation, Key VO, editor. University of Tennessee Press, 1984.


 “Building Social Cohesion between Christians and Muslims through Soccer in post-ISIS Iraq,” Mousa S. Science 369(6505), August 2020, pp. 866–870.


 “The Long-Run Effects of School Racial Diversity on Political Identity,” Billings SB, Chyn E, Haggag K. NBER Working Paper 27302, June 2020, and American Economic Review: Insights 3(3), September 2021, pp. 267–284.


 “Why Did the Democrats Lose the South? Bringing New Data to an Old Debate,” Kuziemko I, Washington E. NBER Working Paper 21703, November 2015, and American Economic Review 108(10), October 2018, pp. 2830–2867.


Brown Fades: The End of Court-Ordered School Desegregation and the Resegregation of American Public Schools,” Reardon SF, Grewal ET, Kalogrides D, Greenberg E. Journal of Policy Analysis and Management 31(4), Fall 2012, pp. 876–904.


Long-run Impacts of School Desegregation & School Quality on Adult Attainments,” Johnson RC. NBER Working Paper 16664, September 2015. “School Segregation, Educational Attainment, and Crime: Evidence from the End of Busing in Charlotte-Mecklenburg,” Billings SB, Deming DJ, Rockoff JE. NBER Working Paper 18487, October 2012, and The Quarterly Journal of Economics 129(1), February 2014, pp. 435–476.


 Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy, Gilens M. Chicago: University of Chicago Press, 2000.


The Effects of Racial Segregation on Intergenerational Mobility: Evidence from Historical Railroad Placement,” Chyn E, Haggag K, Stuart BA. NBER Working Paper 30563, October 2022.


The Wrong Side(s) of the Tracks: Estimating the Causal Effects of Racial Segregation on City Outcomes,” Ananat EO. NBER Working Paper 13606, November 2007, and as “The Wrong Side(s) of the Tracks: The Causal Effects of Racial Segregation on Urban Poverty and Inequality” in American Economic Journal: Applied Economics 3(2), April 2011, pp. 34–66.


"Segregation and Black Political Efficacy,” Ananat EO, Washington EL. NBER Working Paper 13606, February 2008, and Journal of Public Economics 93(5–6), June 2009, pp. 807–822.


"The Opportunity Atlas: Mapping the Childhood Roots of Social Mobility,” Chetty R, Friedman JN, Hendren N, Jones MR, Porter SR. NBER Working Paper 25147, February 2020.