Principal investigator: Bruno Benevit
Author Peter Hinrichs
Original title: Affirmative Action and Racial Segregation
Location of the Intervention: United States
Sample Size: 2.068 state-years
Sector: Economics of Education
Primary Variable of Interest: Racial segregation in colleges
Type of Intervention: Removal of affirmative action policies
Methodology: OLS, Event Study
Summary
Given the racial inequalities present in the United States, disparities in access to higher education between whites and minorities are part of this context. Affirmative action policies were conceived with the intention of reducing these historically established differences within the college environment in the country. This study analyzed how state bans on affirmative action in colleges at different times affected racial segregation in these educational institutions. The results indicated that affirmative action bans increased segregation among colleges in some cases, but reduced it in others. Specifically, early bans on affirmative action in states with highly selective public universities were associated with less segregation, while more recent bans were associated with greater segregation.
- Policy Problem
Affirmative action policies were designed to increase minority enrollment in higher education, and their implementation spread throughout the United States during the 1960s and 1970s. In the US context, these policies were designed to reduce historically established differences within the country's college environment (HINRICHS, 2020). Recently, affirmative action has been defended based on the educational gains resulting from increased diversity in academic environments, an argument supported by the Supreme Court when it ruled it constitutional.
However, in recent years, several states in the country have ceased employing affirmative action in their public universities, resulting from direct voting decisions, legislative actions, and executive orders from different states. As a result, several studies have identified that the discontinuation of affirmative action has led to the entry of minorities into selective colleges, but has not reduced the participation of minorities in general.
The removal of affirmative action policies can have diverse effects on racial interaction in higher education. If the end of these policies reduces minority enrollment in more selective institutions, displaced students may be redistributed to universities with high or low concentrations of minorities (ARCIDIACONO; AUCEJO; HOTZ, 2016; REARDON; BAKER; KLASIK, 2012). Thus, it is relevant to understand how this redistribution affects the overall landscape of universities, as the impact is not only on the diversity of an institution, but on racial interaction throughout the higher education system.
- Policy Implementation Context
Considering the application or discontinuation of affirmative action policies, the distribution of minorities in colleges varies according to the level of selectivity of the institutions. The relationship between the proportion of minority students and the selectivity of colleges does not follow a fixed pattern, being influenced by several factors, including the presence or absence of affirmative action policies. More selective colleges tend to have lower representation of minority students, especially in states where affirmative action has been prohibited. Without these policies, it is likely that minorities will be displaced from highly selective institutions to less selective colleges. However, this displacement may occur differently depending on the level of the college, with minorities being relocated to medium-sized or less selective institutions that previously had little minority representation.
The effects of affirmative action on racial segregation among colleges are ambiguous. Banning these policies could increase segregation by displacing minorities to already segregated institutions, or reduce segregation if students are redistributed to colleges that previously had little racial diversity. Furthermore, there are scenarios where these changes do not significantly affect segregation, since displacements that increase segregation can be offset by those that decrease it.
In this sense, analyzing the impact of affirmative action on segregation requires detailed data on the racial composition of each faculty, since changes in the average racial composition of institutions do not necessarily capture the actual segregation between them. While the impacts of banning affirmative action may not have significant effects on the representation of minorities in universities in general, the compositions among universities can change drastically. Especially when considering the quality of the institutions, the authors argue for the existence of a U-shaped relationship between the quality of the faculties and the representation of minorities: affirmative action promotes a distribution with a greater concentration of minorities in more and less selective institutions. Thus, the results vary according to the dynamics of student redistribution and the returns from the faculties.
- Evaluation Details
The study used data from the Integrated Postsecondary Education Data System (IPEDS), a dataset on higher education institutions, collected annually by the U.S. National Center for Education Statistics. Institutions participating in federal financial aid programs are required to complete IPEDS surveys, providing information on enrollment, costs, institutional finances, and other characteristics.
For this study's analysis, data from four-year colleges on the number of first-time undergraduate students studying full-time, categorized by race, from 1995 to 2016, were considered. This data allowed for the construction of racial segregation indices at the state level.
The segregation metrics used by the authors included three standard indices: (i) the white-to-black exposure index, (ii) the black-to-white exposure index, and (iii) the white-to-black dissimilarity index. The exposure indices were constructed separately for whites and blacks, and measure the potential for interaction between racial groups. The white-to-black exposure index calculated the percentage of black students at the institution attended by the average white student, while the black-to-white exposure index measures the percentage of white students at the institution attended by the average black student. Finally, the dissimilarity index assessed how unevenly distributed students of different races are among institutions, indicating the percentage of students of a given race that would need to be reallocated so that all colleges have the same racial composition.
- Method
The study used the Ordinary Least Squares (OLS) regression method to analyze how affirmative action bans affect the demographic composition of universities at different levels of selectivity between 1995 and 2016. The model estimated the participation of students from different racial groups in an institution, considering whether or not a state implemented an affirmative action ban. The regressions included control variables such as fixed effects of institution, year, and state-specific temporal trends. The OLS method was weighted by the total number of enrollments at each institution and with robust standard errors, with clustering at the state level. In addition to the full four-year sample of institutions, the study also considered analyses for subsamples of public universities and institutions ranked among the best in college rankings.
The study also estimated an Event Study model to assess the impact of affirmative action bans on racial segregation in colleges, taking into account segregation rates calculated for each state over the years. Regressions were weighted according to the racial composition of the institutions and include fixed state effects and specific temporal trends to reduce bias arising from different segregation trajectories among states. This method aimed to analyze whether segregation trajectories before the ban are comparable between states that adopted the ban and those that did not.
Finally, the study presented a California case study to analyze the relationship between college quality and minority representation after the state's ban on affirmative action. Thus, the research examined how the ban affected the redistribution of Black students among universities with different levels of selectivity. The aim was to explore the hypothesis that racial segregation could decrease due to the redistribution of Black students to less selective colleges, leading to greater homogeneity in the racial composition of institutions and reducing racial segregation among them.
- Main results
Analyses using the OLS method revealed that bans on affirmative action were associated with a decrease in the representation of Black, Hispanic, and Native American students, and an increase in the representation of Asian and White students, especially in more selective institutions. These effects were more pronounced in public universities ranked among the top 50 in the rankings. US News and World ReportBetween 1995 and 2016, the ban led to statistically significant reductions in minority enrollment, with varying impacts over time. In more recent periods, the declines in Black and Hispanic enrollment were smaller, suggesting heterogeneity in the treatment's effects over time and across different levels of selectivity.
Regarding analyses that used the OLS method to estimate the impact of the ban on affirmative action on segregation between whites and blacks, the results showed no significant effects. For the period from 1995 to 2016, the exposure of whites to blacks decreased slightly, while the exposure of blacks to whites and the dissimilarity index between blacks and whites showed statistically insignificant variations. However, when considering specific periods, such as 1995–2003 and 2004–16, there were indications that the bans affected segregation differently over time. While older bans were associated with a reduction in segregation, more recent bans pointed to an increase in segregation.
Regarding the Event Study, the results suggested that, prior to a ban, there were no significant differences in pre-existing trends between states with and without the ban. For the period 1995–2003, there was a significant increase in exposure of Black people to White people and a decrease in racial dissimilarity in the year of the ban. Conversely, the period 2004–16 showed a slight reduction in exposure of Black people to White people and an increase in dissimilarity.
The results identified in the analysis of the relationship between college quality and minority representation in the state of California showed significant changes after the ban on affirmative action in 1998. An increase in the exposure of Black students to white students was observed that year, despite a downward trend over time. Furthermore, the dissimilarity index between Black and white students showed a considerable reduction immediately after the ban. Additionally, a reduction in the heterogeneity in the distribution of minorities according to college selectivity was observed, indicating a flattening of the U-shaped curve that describes the relationship between university quality and minority presence, suggesting a decrease in segregation among universities. The likely explanation is that, without affirmative action, Black students were redistributed, especially among moderately selective institutions.
- Lessons in Public Policy
In this article, the authors conducted several empirical approaches to analyze how state bans on affirmative action in the United States affected racial segregation in colleges. The results indicated that the bans had heterogeneous effects over time, with a reduction in racial segregation in the most selective institutions soon after the first bans, while in more recent periods, a trend of increased segregation was observed in some regions. Additionally, the authors found evidence that the bans resulted in a redistribution of Black students, especially among institutions with different levels of selectivity.
The evidence presented in this article provides relevant information on the consequences of affirmative action bans for racial segregation in higher education, offering insights for the formulation of educational public policies. The authors highlight that, considering the observed dynamics, policies that promote greater inclusion and reduce institutional barriers in universities with different levels of selectivity can contribute to mitigating the effects of increased racial segregation in some regions, especially in states that have adopted the bans more recently.
References
ARCIDIACONO, P.; AUCEJO, EM; HOTZ, VJ University Differences in the Graduation of Minorities in STEM Fields: Evidence from California. American Economic Review, v. 106, n. 3, p. 525–562, 1 Mar. 2016.
HINRICHS, P. Affirmative Action and Racial Segregation. The Journal of Law and Economics, v. 63, n. 2, p. 239–267, 1 May 2020.
REARDON, SF; BAKER, R.; KLASIK, D. Race, Income, and Enrollment Patterns in Highly Selective Colleges, 1982–2004. StanfordStanford University, Center for Education Policy Analysis, 2012.
