Case Study 02 — SFFA v. Harvard / UNC (2023)

The plaintiff and the cases

Students for Fair Admissions, Inc. (SFFA) is a membership nonprofit founded in 2014 by Edward Blum, a longtime opponent of race-conscious admissions and the strategist behind Fisher v. University of Texas (2013, 2016). SFFA filed two lawsuits in 2014: one against Harvard University, alleging that Harvard's race-conscious admissions discriminated against Asian-American applicants in violation of Title VI of the Civil Rights Act of 1964; the other against the University of North Carolina at Chapel Hill, alleging that UNC's race-conscious admissions violated both Title VI and the Equal Protection Clause of the Fourteenth Amendment (UNC, as a state institution, is subject to both).

The cases proceeded through district court trials at Harvard's, with extensive expert testimony on Harvard's admissions data. Plaintiffs' expert, Peter Arcidiacono of Duke, presented regression analyses suggesting that Asian-American applicants were systematically rated lower than otherwise comparable applicants on Harvard's "personal rating" — a subjective score covering "likeability," "courage," and similar traits. Harvard's expert, David Card of Berkeley, contested the regression specifications. The district court found in favor of Harvard; the First Circuit affirmed; the Supreme Court granted certiorari.

The cases were consolidated for argument and decided together in Students for Fair Admissions v. President and Fellows of Harvard College and Students for Fair Admissions v. University of North Carolina, 600 U.S. 181 (2023). Justice Jackson recused from the Harvard case because she had served on Harvard's Board of Overseers; she participated in the UNC case.

The holdings

Chief Justice Roberts wrote for a 6-3 majority (6-2 in the Harvard case), joined by Justices Thomas, Alito, Gorsuch, Kavanaugh, and Barrett. Justice Sotomayor wrote a dissent joined by Justices Kagan and (in UNC) Jackson. Justice Jackson wrote a separate dissent in UNC. Justices Thomas, Gorsuch, and Kavanaugh wrote concurrences.

The core holding: Harvard's and UNC's race-conscious admissions programs violate Title VI of the Civil Rights Act and the Equal Protection Clause. The reasoning proceeds in several steps.

Step 1: Strict scrutiny applies. The Court reaffirmed that any racial classification by government (or by entities accepting federal funds, under Title VI) is subject to strict scrutiny. Universities, including Harvard as a private federal-fund recipient, are subject to this standard.

Step 2: The diversity rationale fails strict scrutiny as the universities have framed it. The Court's prior decisions (Bakke, Grutter, Fisher) had accepted "the educational benefits flowing from a diverse student body" as a compelling state interest. The SFFA majority does not formally overrule those holdings, but holds that the diversity goals as articulated by Harvard and UNC ("training future leaders," "preparing engaged and productive citizens," "enhancing appreciation, respect, and empathy") are not sufficiently coherent or measurable to qualify as compelling government interests. Roberts: "Although these are commendable goals, they are not sufficiently coherent for purposes of strict scrutiny. ... How is a court to know whether leaders have been adequately 'trained'; whether the exchange of ideas is 'robust'; or whether 'new knowledge' is being developed?"

Step 3: The programs use race as a stereotype, not as an individualized factor. The majority emphasizes that the universities' programs treat race as a fixed characteristic associated with a fixed perspective — a stereotype that the Equal Protection Clause forbids. Roberts: "If a university is to consider race only to permit individuals to enrich the prospective student's contribution to a diverse class, it follows that race-based classifications cannot be used merely because of stereotypes about the perspectives that a person of a particular race holds."

Step 4: The programs lack a logical endpoint. Grutter (2003) had said that race-conscious admissions "must have a logical end point." Twenty years later, the universities had no end point. The majority reads this as an additional failure of strict scrutiny.

Step 5: The programs result in racial balancing. Even though the universities deny using racial quotas, the empirical pattern of admissions from year to year (the proportion of Black, Hispanic, white, and Asian American admits remaining within narrow bands) suggests, the majority argues, that the universities are practicing racial balancing in fact if not in form.

The narrow caveats. The opinion preserves two openings. First, applicants may discuss in personal essays how their race has affected their lives — Roberts: "Nothing in this opinion should be construed as prohibiting universities from considering an applicant's discussion of how race affected his or her life, be it through discrimination, inspiration, or otherwise." Second, the opinion does not extend to U.S. military academies, where "potentially distinct interests" may apply (a holding subsequently litigated in Students for Fair Admissions v. United States Military Academy at West Point, ongoing as of 2026).

The dissent

Justice Sotomayor's principal dissent (54 pages, joined by Justices Kagan and Jackson) offers a comprehensive defense of race-conscious admissions. The major moves:

Historical interpretation. The Reconstruction Congress that drafted the Fourteenth Amendment also enacted race-conscious legislation — the Freedmen's Bureau Acts, the Reconstruction-era veterans' benefits — without considering them constitutionally suspect. The original public meaning of the Fourteenth Amendment, Sotomayor argues, included the permissibility of race-conscious remediation.

Doctrinal continuity. Brown v. Board did not require race-blindness; it required desegregation, which is itself a race-conscious remedy. Bolling v. Sharpe, Green v. County School Board, and Swann v. Charlotte-Mecklenburg all required race-conscious remedies. The majority's reading of "color-blind" as a constitutional principle, Sotomayor argues, is not faithful to the civil-rights tradition or to the original meaning of the Fourteenth Amendment.

Empirical claims. The dissent argues that race-blind admissions in unequal societies produce racial disparities — that K-12 educational opportunities, family wealth, residential segregation, and access to test preparation produce systematically different metrics by race that are not the result of meritocratic differences. Race-conscious admissions, on this view, are how universities identify excellence under conditions of structural inequality.

Predictive judgment. The dissent predicts that the elimination of race-conscious admissions will produce significant declines in Black and Latino enrollment at selective institutions, citing the post-Proposition 209 California experience and modeling studies on Harvard's likely future composition.

Justice Jackson's separate dissent in UNC focuses on the historical wealth gap and on the empirical inadequacy of race-blind alternatives.

The empirical question: what happens after SFFA?

The first admissions cycle subject to SFFA's holding was Fall 2024. The data, as released through 2025, are partial and contested:

• MIT — Black enrollment dropped from 15 percent of the entering class (2023) to 5 percent (2024). Hispanic enrollment dropped from 16 percent to 11 percent. Asian-American enrollment rose from 40 percent to 47 percent.
• Amherst College — Black enrollment dropped from 11 percent to 3 percent.
• Harvard — Black enrollment dropped slightly (from 18 to 14 percent of the new class), Hispanic enrollment held roughly steady, Asian-American enrollment increased.
• Princeton — racial composition reported broadly stable.
• Yale — racial composition reported broadly stable.
• Duke, UNC — modest declines in Black enrollment, modest increases in Asian-American enrollment.
• Texas A&M, University of Texas at Austin — already operating under the state's percent plan and largely unaffected.

Two cautions on these data: first, schools differ in admissions methodology, so single-cycle comparisons across schools are noisy; second, schools differ in their specific responses to SFFA (some emphasized class-based factors, some intensified geographic recruitment, some maintained essay prompts that effectively elicited race-relevant content), so the variation in outcomes reflects both SFFA's effect and the schools' adaptive choices.

A serious empirical assessment will take five to ten admissions cycles. As of 2026, the preliminary picture is mixed: significant declines in Black enrollment at some elite institutions, stability or modest declines at others, and a modest reshuffling among Asian-American subpopulations.

Steel-manning the two readings

The chapter's posture, faithfully applied, requires presenting both readings of SFFA in their strongest forms.

The conservative reading. SFFA is the proper application of a principle that has been latent in American constitutional law since the Fourteenth Amendment and explicit since Justice Harlan's Plessy dissent: the Constitution is color-blind. Government — and federally funded institutions standing in the same constitutional position — may not classify citizens by race in order to advantage some at the expense of others, even with benevolent intentions. The civil-rights tradition that ran through Brown and the 1964 and 1965 Acts was a tradition of dismantling racial classifications, not of building benevolent ones. The diversity rationale of Bakke and Grutter was a doctrinal compromise that allowed race-conscious admissions to continue under the guise of an interest (diversity) different from the interest (remediation) the universities really cared about, and the Court has correctly recognized that the compromise was unstable. The mismatch literature (Sander, Ho), the stigma literature, the resentment literature, and the principle of treating individuals as individuals all support SFFA's direction. The universities will adapt — with class-based, geography-based, and essay-based methods — and the resulting student bodies will be more genuinely meritocratic without being less diverse on the dimensions that matter.

The progressive reading. SFFA is a constitutional and historical mistake. The Reconstruction Amendments were enacted to enable race-conscious remediation, not to forbid it; the Reconstruction Congress passed multiple race-conscious statutes, and the original public meaning of the Fourteenth Amendment encompassed such measures. Brown and the 1964 Act dismantled formal racial caste, but they did not eliminate the lasting effects of two and a half centuries of caste — effects measurable in wealth, education, neighborhood quality, and life outcomes. Race-blind admissions, applied to credentials produced by unequal opportunity, will not produce equality; they will reproduce inequality. The post-SFFA enrollment data from MIT and Amherst — sharp declines in Black and Latino enrollment within a single cycle — illustrate the point. The First Amendment's protection of expressive activity, combined with Title VI's reach, will leave universities navigating a complex web in which they cannot consider race directly but must somehow continue to assemble diverse student bodies for the legitimacy of their educational mission. The claim that the Constitution is "color-blind" reads Harlan's Plessy dissent against its context — Harlan was attacking a caste system, not articulating a principle of race-neutrality applicable to remedies for caste.

Both readings have substantial textual, historical, and prudential support. Both are advanced by serious legal scholars across the Court's ideological spectrum. The point of the chapter is not to choose between them; the point is to ensure that the reader can articulate each in its strongest form.

What the case shows

Three lessons:

1. The civil-rights tradition is internally contested. The principle of "equal treatment under law" admits multiple readings. The disagreement between SFFA's majority and dissent is not a disagreement between civil-rights advocates and civil-rights opponents. It is a disagreement within the civil-rights tradition over what equal treatment requires.

2. Doctrinal change is real and consequential. Bakke (1978) governed for forty-five years. Grutter (2003) governed for twenty years. SFFA (2023) substantially narrows both. American constitutional doctrine moves; what was settled in 2003 is unsettled in 2023; what is settled in 2023 may be unsettled in 2043.

3. Justice Harlan's "color-blind Constitution" is now invoked across the political spectrum. That is not a contradiction; it is a sign of a phrase that has acquired meanings beyond what its author intended. Both readings — the demand for race-neutral law, and the demand for the abolition of caste — are part of the civil-rights tradition. American law is in the process of working out which reading governs which contexts. The work is not finished.

Sources

• Students for Fair Admissions, Inc. v. President and Fellows of Harvard College, 600 U.S. 181 (2023). The full opinion, including concurrences and dissents, is approximately 230 pages and rewards reading.
• Justice Sotomayor's dissent (Section IV especially) is the most comprehensive recent statement of the historical case for race-conscious admissions.
• Justice Thomas's concurrence is the most detailed recent statement of the originalist case for color-blindness.
• Peter Arcidiacono's expert reports in the Harvard trial (publicly available through court filings) for the regression-analysis evidence.
• David Card's expert reports (also public) for the alternative analysis.
• The Atlantic, The New York Times, The Wall Street Journal, and The Chronicle of Higher Education have continuing coverage of post-SFFA admissions cycles.
• Randall Kennedy, For Discrimination: Race, Affirmative Action, and the Law (2013) — pre-SFFA defense of affirmative action by a Black Harvard Law professor.
• Thomas Sowell, Affirmative Action Around the World: An Empirical Study (2004) — comparative critique.
• Sander and Taylor, Mismatch (2012) — the empirical mismatch hypothesis.