Reflections on a Comparative Perspective Within the U.S.
Roger Andersson’s reflection on segregation in the U.S. and Sweden reminds us of the importance of examining segregation using a comparative perspective, which can help us understand the relative importance of different factors that contribute to segregation, as well as the types of policies that seem to reduce it or mitigate its effects. Andersson’s call to undertake comparative studies can be applied to examining differences between countries but also differences between racial/ethnic groups within countries. He suggests that while in Sweden racial residential segregation is relatively recent and due to increasing immigration, in the U.S. segregation has been shaped by the African American experience. While this general point is valid, today’s segregation in the U.S. also reflects the experience of the Latino population—the largest racial/ethnic minority group—and its rapid growth fueled by immigration in the last five decades. I am aware of and have done research on the nuanced differences between black and Latino segregation. Yet, what I find most striking is the “efficiency” of the U.S. spatial stratification system in building such segregated patterns for Latinos over a few decades. Andersson has rightly identified some of the key elements of this stratification system, for example, high municipal governance fragmentation and very limited cross-jurisdictional housing and school systems.
Research on segregation in the U.S. continues to focus largely on the African American experience, while research on Latinos more commonly uses the immigrant assimilation paradigm. This may be a helpful distinction, and one would hope eventually Latinos will emulate previous immigrant groups and integrate spatially and otherwise. However, today more than two-thirds of the Latino population is second or higher generation, i.e., born in the U.S., and multiple indicators suggest limited residential integration.
Because of current demographic trends in the U.S., we should examine and try to undo patterns of segregation not only among blacks, but also among other groups, especially Latinos. There is, of course, a wealth of research on U.S. immigrants, and some researchers routinely calculate indices for all the main U.S. racial/ethnic groups.1 However, when it comes to housing policy research or just policy debate, Latinos are often missing. When I attend housing policy (research) conferences, I am often struck by the limited discussion of Latino housing issues. Housing problems –from affordability to crowding to segregation—disproportionately affect Latinos and blacks. However, possibly due to eligibility restrictions for immigrants and/or limited outreach and cultural competency, Latinos are underrepresented in housing programs. For example, the Housing Choice Voucher program serves approximately 34% of income-eligible renter black households with children, but only 6% of similar Latino households and 10% of similar white households. Latinos are also under represented in public housing and project-based housing vouchers. Thus, paradoxically, and in sharp contrast with our demographic reality—close to 25% of U.S. children are Latino—Latinos may not be seen as a major housing policy constituency. This may help explain why housing policy (research) discussions often center on black/white issues.
Another reason why Latino segregation may not be a main focus of policy discussions is that Latinos are not as highly residentially segregated as blacks. However, they are highly segregated and the patterns we observe are not promising. For example, school segregation is now higher for Latino children than for black children, and the trend is that their segregation will continue to increase.2 Residential segregation of black and Latino children declined modestly over 2000-2010. However, there was increasing segregation of Latino children in many of small- to medium-sized metro areas in the South and Midwest, which are experiencing some of the fastest Latino growth.3
Residential and school segregation are associated with vast differences in neighborhood and school environments. Here the trend too is that Latino children’s neighborhood and school environments are not as bad but fairly similar to those of black children. According to the Child Opportunity Index—an index that our project diversitydatakids.org has developed in collaboration with the Kirwan Institute for the Study of Race and Ethnicity—across the 100 largest metropolitan areas, 32% of Latino and 40% of black children are concentrated in the lowest opportunity neighborhoods in their metro area, compared to only 12% of Asian children and 9% of white children. Similarly, today the majority of Latino (and black) children in large metro areas attend a school in which the majority of the students are low-income.4
While the pattern seems to be one of convergence, the effects of segregation may be different for Latinos than for blacks. In public health, several studies suggest that first generation Latino immigrants seem to be resilient to concentrated neighborhood disadvantage, possibly due to beneficial effects of ethnic enclaves or to health selectivity. However, segregation has negative effects on second generation Latino immigrants, which are now the majority of the Latino population.5 Andersson makes the point that it would be helpful to know the benefits of income mixing in housing, and that in Sweden income mixing, but not racial mixing (e.g., avoiding ethnic clustering), has been found to have positive effects. I worry that his point ignores that while for first generation immigrants, ethnic enclaves may be beneficial, or at least not detrimental, this may not be true for the second generation. In the U.S. the evidence indicates that segregation is not beneficial for second or higher generation immigrants for which segregation is a sign of downward assimilation.
Also it is important to make a clear distinction between our knowledge of the effects of mixed-income housing and the stark reality of extremely unequal distributions of neighborhood opportunity by race/ethnicity. It is true that at least from a health perspective, the research evidence that mixed-income housing improves health outcomes is mixed.6 Yet, as discussed earlier, we have extreme racial/ethnic inequities in children’s neighborhood environments. Can we really wait to have strong research evidence on mixed-income housing to state as a normative planning ambition (borrowing Andersson’s terms) that children’s neighborhood and school environment should not be so strongly predicted by their race/ethnicity?
 I thank Nancy McArdle for calculating these figures based on tabulations of the U.S. Department of Housing and Urban Development Public Use Microdata Sample (2012) and Joint Center for Housing Studies of Harvard University tabulations of the U.S. Census Bureau American Housing Survey (2011).
1. Cutler DM, Glaeser EL, Vigdor JL. Is the melting pot still hot? Explaining the resurgence of immigrant segregation. US Census Bureau, Center for Economic Studies, 2004 August 2004. Report No.: Contract No.: CES 04-10.
2. Orfield G, Frankenberg E, Ee J, Kuscera J. Brown at 60: Great Progress, a Long Retreat and an Uncertain Future. Los Angeles,: UCLA, The Civil Rights Project, 2014.
3. McArdle N, Osypuk TL, Hardy E, Acevedo-Garcia D. Segregation Falls for Black Children in Most Metro Areas But Remains High; Fewer Metros Experience Declines for Latinos. Issue Brief. Boston, MA: diversitydata.org, 2011 July. Report No.
4. McArdle N, Osypuk T, Acevedo-Garcia D. Segregation and Exposure to High-Poverty Schools in Large Metropolitan Areas: 2008-09. Boston: DiversityData.org, 2010.
5. Osypuk TL, Bates LM, Acevedo-Garcia D. Another Mexican birthweight paradox? The role of residential enclaves and neighborhood poverty in the birthweight of Mexican-origin infants. Social Science & Medicine. 2010;70(4):550-60. Epub Epub 2009 Nov 18.
6. Briss PA, Zaxa S, Pappaioaunou M. Developing an evidence-based guide to community preventative services – methods. American Journal of Preventative Medicine. 2000;18:35-43.