Affirmative Action Bans

What the evidence says

The strongest evidence is on enrollment and admission. In the manual-drop study of California's ban, many underrepresented minority (URM) applicants became 50 percentage points less likely to be admitted to the University of California's (UC) most selective campuses, relative to academically comparable non-URM applicants. A separate review reports that bans decrease enrollment at the most selective public institutions. It cites a 1.74 percentage-point drop for African American students (about 30%) and a 2.03 percentage-point drop for Latino students (27.5%). In medicine, a matched difference-in-differences study finds the share of underrepresented students in ban schools fell 4.8 percentage points five years after the ban. Control schools rose 0.7 points, so the net difference is −5.5 percentage points, or about a 37% relative reduction from a 14.8% baseline. Another study reports a 17% decline in first-time medical matriculation (from 18.5% to 15.3%). Graduate enrollment of students of color also falls: 12.2% overall, 26% in engineering, 19% in the natural sciences and 15.7% in the social sciences.

The California work identifies two channels. The ban lowers admission odds at selective campuses, and it deters applications: about 250 fewer Black and 900 fewer Hispanic applications a year, almost 80 percent of whom would likely have been admitted to at least one UC campus. A qualitative study reports the same deterrence in Washington State. The California study finds that URM applicants then cascade into lower-quality public and private universities.

The long-run outcomes in the California study run in the same direction. Bachelor's degree attainment falls 4.3 percentage points among URM applicants in the bottom academic-index quartile. STEM degree attainment falls 1.0 percentage point. The likelihood of a graduate degree falls 1.3 percentage points. Average URM applicant wages fall 5% a year between ages 24 and 34, almost wholly driven by Hispanic applicants. Each applicant becomes 1.4 percentage points less likely to earn $100,000 a year in their early 30s. The number of early-career URM Californians earning over $100,000 falls by at least 3%.

The graduation-rate evidence is mixed. Across all institutions, overall graduation rates do not change much. At selective schools, the share of Black college entrants who graduate from top-two-tier universities falls 0.93 percentage points, and the share graduating from top-50 universities falls 1.54 points. Net of these effects, bans lead to fewer URM graduates of selective colleges. The one rise is among Hispanics at top-two-tier public universities, where six-year graduation increases 2.36 percentage points. The author notes this may reflect a change in who enrolls. The claims show no measured effect of the ban on STEM course grades, which bears on the hypothesis that admitted students were under-prepared.

Evidence on non-economic outcomes is thin. One study finds URM adults exposed to bans in late high school were 1.8 percentage points more likely to smoke, with no measured effect among non-Hispanic White adolescents. A qualitative law-school study (tier E) links bans to racial isolation, stigma and overt racism. This page has no quantified evidence on these outcomes beyond those two studies.

Where the evidence splits

Application behavior splits. One study finds no change in the SAT-sending behavior of highly qualified Black and Hispanic students after California and Texas ended race-conscious admissions. It also finds no change in the number of schools they applied to. The California study finds about 250 fewer Black and 900 fewer Hispanic applications a year, and the Washington account finds minority students chose not to apply. The claims do not say why these differ, so the split is unexplained by the sources on this page. The measured groups (highly qualified SAT-senders versus all URM UC applicants) are not the same, but no source states that this reconciles them.

Graduation effects split by group and selectivity. Overall graduation rates barely move, Black graduation at top-tier universities falls, and Hispanic graduation at top-two-tier public universities rises. The author attributes the Hispanic rise possibly to changing student composition. The claims give no test of that explanation.

The Black–Hispanic comparison depends on the measure. The medical-school study finds a larger decrease for Hispanic than for Black students. The selective-institution review reports a larger percentage-point drop for Latino students (2.03 versus 1.74) but a larger relative drop for African American students (30% versus 27.5%). The claims do not state the metric in the medical-school comparison.

The net effect on practicing Black lawyers is theoretical. One argument holds that, under certain assumptions, eliminating affirmative action could increase the number of practicing Black lawyers even as elite law-school enrollment falls. No measured result on this page supports or refutes it.

Thresholds and dose-response

No threshold established in the current sources.

Open empirical questions

• Why do SAT-sending and application counts diverge between the California and Texas study and the California application-deterrence study?
• Do the California earnings and degree effects appear in other states with bans?
• Does the Hispanic graduation increase at top-tier public universities persist after accounting for changes in who enrolls?
• How large are the health-behavior and stigma effects beyond the single smoking estimate and the qualitative law-school account?