IN THE UNITED STATES COURT OF APPEALS
FOR THE FIRST CIRCUIT
STUDENTS FOR FAIR ADMISSIONS, INC.,
PRESIDENT AND FELLOWS OF HARVARD COLLEGE,
ON APPEAL FROM THE UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF MASSACHUSETTS
BRIEF FOR THE UNITED STATES AS AMICUS CURIAE
SUPPORTING APPELLANT AND URGING REVERSAL
ERIC S. DREIBAND
Assistant Attorney General
ELLIOTT M. DAVIS
Acting Principal Deputy Assistant Attorney General
THOMAS E. CHANDLER
MATTHEW J. DONNELLY
U.S. Department of Justice
Civil Rights Division
Benjamin Franklin Station
P.O. Box 14403
Washington, D.C. 20044-4403
TABLE OF CONTENTS
INTRODUCTION ..................................................................................................... 1
INTEREST OF THE UNITED STATES .................................................................. 4
STATEMENT OF THE ISSUE ................................................................................. 5
STATEMENT OF THE CASE .................................................................................. 5
SUMMARY OF ARGUMENT ................................................................................. 8
ARGUMENT ........................................................................................................... 10
I HARVARD FAILED TO PROVE THAT ITS USE OF RACE IS NARROWLY TAILORED ................................................ 11
A. Harvard Engages In Racial Balancing ..................................... 11
1. The Remarkably Stable Racial Composition Of Harvard’s Incoming Classes Reflects Racial Balancing ............................................................ 12
2. Harvard Achieves Racial Balancing By Considering Race Throughout The Admissions Process And Continually Monitoring And Reshaping The Class’s Racial Composition................... 15
B. Harvard’s Use Of Race Penalizes Asian Americans ................ 23
1. Harvard’s Own Statistics Reflect A Penalty Against Asian Americans ................................................ 23
2. The District Court Relieved Harvard Of Its Obligation To Justify The Lower Personal Ratings It Assigns To Asian-American Applicants ....................................................................... 26
II HARVARD FAILED TO ARTICULATE A SUFFICIENTLY CONCRETE AND MEASURABLE INTEREST .............................. 29
TABLE OF CONTENTS (continued): PAGE
CONCLUSION ........................................................................................................ 32
CERTIFICATE OF COMPLIANCE
- ii -
TABLE OF AUTHORITIES
City of Richmond v. J.A. Croson Co., 488 U.S. 469 (1989) (plurality opinion) .............................. 4
Fisher v. University of Tex. at Austin, 570 U.S. 297 (2013) ............................passim
Fisher v. University of Tex. at Austin, 136 S. Ct. 2198 (2016) ............. 16-17, 29, 30
Gratz v. Bollinger, 539 U.S. 244 (2003) ................................................................. 10
Grutter v. Bollinger, 137 F. Supp. 2d 821 (E.D. Mich. 2001) ................................ 21
Grutter v. Bollinger, 539 U.S. 306 (2003) ........................................................passim
Heller v. Doe, 509 U.S. 312 (1993) ......................................................................... 28
Parents Involved in Community Schools v. Seattle School District No. 1, 551 U.S. 701 (2007) ............ 10, 12
Republican Party of Minnesota v. White, 536 U.S. 765 (2002) .............................. 28
42 U.S.C. 2000d .................................................................................................. 1, 10
Federal Rule of Appellate Procedure 29(a)(2) ........................................................... 5
- iii -
IN THE UNITED STATES COURT OF APPEALS FOR THE FIRST CIRCUIT
STUDENTS FOR FAIR ADMISSIONS, INC.,
PRESIDENT AND FELLOWS OF HARVARD COLLEGE,
ON APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF MASSACHUSETTS
BRIEF FOR THE UNITED STATES AS AMICUS CURIAE
SUPPORTING APPELLANT AND URGING REVERSAL
Title VI of the Civil Rights Act of 1964 commands that “[n]o person in the United States shall * * * be subjected to discrimination under any program or activity receiving Federal financial assistance” based on her “race, color, or national origin.” 42 U.S.C. 2000d. Although the Supreme Court has held that colleges receiving federal funds may nevertheless consider applicants’ race in certain limited circumstances, it has made clear that colleges may do so only if
- 2 -
they satisfy strict scrutiny—that is, only if the college demonstrates that its use of race is narrowly tailored to a compelling interest.
Harvard College receives millions of dollars in federal taxpayer-financed assistance every year. By accepting federal funding, Harvard subjected itself to Title VI’s stringent restrictions on the use of race. As Harvard undisputedly considers applicants’ race when selecting its incoming freshman class, it bore the burden at trial of proving that its consideration of race in the admissions process is narrowly tailored to a cognizable compelling interest. Yet the unvarnished record shows that Harvard’s use of race is hardly tailored at all.
The trial record established that Harvard actively engages in racial balancing that Supreme Court precedent flatly forbids. Indeed, the racial composition of Harvard’s admitted class is strikingly stable from year to year. That result is no accident. The school considers applicants’ race at virtually every step, from rating applicants to winnowing the field of applicants when attempting to avoid an over-subscribed class. And its inclusion of race in the analysis frequently makes a dispositive difference. The district court found that Harvard’s use of race was “determinative” for “approximately 45% of all admitted African American and Hispanic applicants.” ADD84. Moreover, Harvard meticulously tracks and shapes the racial makeup of its emerging incoming class throughout the process, continuously comparing the new class’s racial composition with that of the
- 3 -
previous year. This overt engineering of racial stasis bears no resemblance to the flexible, nonmechanical “plus” factor that the Supreme Court’s cases to date have permitted.
The evidence also showed that Harvard’s process has repeatedly penalized one particular racial group: Asian Americans. Indeed, Harvard concedes that eliminating consideration of race would increase Asian-American admissions while decreasing those of Harvard’s favored racial groups. The resulting racial penalty stems in part from one component of Harvard’s admissions rubric—a nebulous and entirely subjective “personal rating”—that consistently and inexplicably produces poorer scores for Asian Americans than for other applicants. That disparity is undisputed, and unexplained. Harvard bore, but did not carry, the burden of proving that this disparity is not the product of racial discrimination. As the district court observed, based on the trial record, one could not rule out racial discrimination as the source.
Despite highlighting various holes in Harvard’s evidence, the district court credited the school’s assurances that it does not employ racial biases in the personal rating, and gave the school the benefit of the doubt when the experts’ analyses conflicted. But strict scrutiny demanded much more. Faced with the results of Harvard’s process—a barely fluctuating racial composition and less favorable evaluations of Asian Americans—and the mechanisms that produced
- 4 -
them, the court should have put Harvard to its proof. And even if the parties’ evidence were in equipoise, a tie should have gone to the plaintiff, not to Harvard. On this record, the court should have found that Harvard’s failure to demonstrate affirmatively that its admissions process meets the Supreme Court’s stringent criteria was fatal.
The wisdom of race-based admissions policies like Harvard’s is subject to vigorous debate. But Title VI and Supreme Court precedent impose limitations on Harvard’s ability to consider race in its admissions process—limitations that Harvard has not respected. It now falls to this Court to enforce these limitations and reverse the district court’s judgment.
INTEREST OF THE UNITED STATES
The United States has a substantial interest in protecting its citizens from racial discrimination. The United States enforces multiple statutes that prohibit racial discrimination in public accommodations, housing, voting, education, and employment, among other contexts. It also has a fundamental interest in ensuring “that public dollars, drawn from the tax contributions of all citizens, do not serve to finance the evil of private prejudice.” City of Richmond v. J.A. Croson Co., 488 U.S. 469, 492 (1989) (plurality opinion). As particularly relevant here, the United States distributes billions of dollars in federal financial assistance every year, and it
- 5 -
has a significant interest in ensuring that recipients of such assistance comply with Title VI’s antidiscrimination mandate.
The United States respectfully submits this brief pursuant to Federal Rule of Appellate Procedure 29(a)(2).
STATEMENT OF THE ISSUE
Whether Harvard carried its burden at trial of proving that its overt consideration of race in its admissions process is narrowly tailored to advance a compelling interest, as required by Title VI and Supreme Court precedent.
STATEMENT OF THE CASE
1. Harvard considers race throughout its admission process. See ADD17-18; ADD21-24; ADD26-31; Pl.’s Br. 6-9. Admissions officers assign four numerical ratings referred to as the academic, extracurricular, athletic, and personal ratings. ADD18. Ratings generally range from 1 to 4, with 1 being the strongest. ADD18-19. Admissions officers also assign an overall rating. ADD18. Harvard uses race in the overall rating and in admission determinations made by its subcommittees and its full committee. ADD21-26. Throughout the admissions cycle, admissions leaders closely monitor the racial composition of the applicant pool and tentatively admitted students and compare them to the previous year’s racial composition. ADD27 28. Harvard’s close attention to its racial composition
- 6 -
has produced a remarkably consistent racial balance in the admitted class from year to year. JA4434-4435; JA4446-4447; see Part I.A.1., infra.
Harvard’s use of race benefits African-American and Hispanic applicants. For example, as the district court found, “race is a determinative tip for approximately 45% of all admitted African American and Hispanic applicants.” ADD84; see also JA6121; JA3063-3065 (Card); JA5749.
2. Plaintiff Students for Fair Admissions (SFFA) commenced this action alleging, inter alia, that Harvard violates Title VI through its use of race in its admissions process. JA108. After denying the parties’ cross-motions for summary judgment, ADD132-133, the district court conducted a three-week bench trial.
Much of the trial centered on disparities in Asian Americans’ admission rates and personal ratings. ADD53-56. The personal rating purportedly captures intangible qualities such as “integrity, helpfulness, courage, kindness, fortitude, empathy, self-confidence, leadership ability, maturity, or grit.” ADD20. And it carries significant weight in the admissions process: although fewer than 20% of applicants receive a personal rating of 1 or 2, they represent almost 80% of the admitted class. JA2215-2216 (Arcidiacono); JA5997. Admissions officers testified that they did not directly consider race in the four subsidiary profile ratings, including the personal rating, ADD45, but Harvard never gave them written guidance on how to use race in assigning ratings until 2018, when it revised
- 7 -
its reading procedures on the eve of trial, ADD17 (citing JA4586-4590; JA647-649 (Fitzsimmons)). Moreover, the racial distribution of the personal rating nearly mirrors that of the overall rating, for which Harvard admits it uses race. JA2228-2229, JA2255-2258 (Fitzsimmons); JA6005-6006.
Both sides relied heavily on expert statisticians, who modeled Harvard’s admissions process based on data from Harvard’s applicant classes of 2014 through 2019. ADD50-51; ADD62 63. The experts offered competing analyses of whether Harvard’s process discriminated against Asian Americans, and whether race infected the personal rating. ADD53-80. The district court found both experts’ models “worthy of consideration.” ADD75. With the personal rating excluded, both experts’ models show Harvard’s program inflicts a statistically significant penalty against Asian-American applicants. ADD66; ADD74-79; JA2317-2318 (Arcidiacono); JA3149-3152 (Card).1
After trial, the district court concluded that Harvard has a compelling interest in student-body diversity that justified its use of race in its admissions program and that its program is narrowly tailored to that interest. ADD122-127. The court held that Harvard uses race as a “plus” factor “in a flexible, nonmechanical way.” ADD108 (quoting Grutter v. Bollinger, 539 U.S. 306, 334
1 The district court discusses a model that blends aspects of both experts’ models. ADD79. But it did not publish that model’s numerical results, making those results impossible to scrutinize.
- 8 -
(2003)). The court also concluded that Harvard had “not imposed racial quotas or otherwise engaged in impermissible racial balancing.” ADD83. The court acknowledged that narrow tailoring requires “that a race-conscious admissions program not unduly harm members of any racial group.” ADD109 (citations omitted). After speculating about possible causes of disparities in the personal rating, the court concluded that those disparities “did not burden Asian American applicants significantly more than Harvard’s race-conscious policies burdened white applicants.” ADD109-111. Finally, the court concluded that “no workable and available race-neutral alternatives would allow” Harvard “to achieve a diverse student body while still maintaining its standards for academic excellence.” ADD122.
SUMMARY OF ARGUMENT
Harvard intentionally uses race in its admissions process. It thus bore the burden to prove that its process satisfies strict scrutiny by showing that its use of race is narrowly tailored to a compelling interest. Harvard did not carry its burden.
I. Even assuming that Harvard has articulated a sufficiently concrete interest in diversity to warrant any consideration of race, Harvard failed to show that its consideration of race is narrowly tailored, for at least two reasons.
A. First, Harvard’s process entails prohibited racial balancing. The manifest steadiness of the racial composition of successive admitted classes speaks
- 9 -
for itself. Harvard’s consideration of race throughout the admissions process and its constant monitoring of the class that is taking shape are the key mechanisms by which Harvard has achieved such remarkable stability from one year to the next. And Harvard’s own internal study of race-neutral alternatives confirms that the school’s goal is maintaining consistency from year to year. The district court approved Harvard’s process as the mere application of a “plus” factor permitted by precedent. But in reality, Harvard’s process employs a system of de facto quotas.
B. Second, Harvard’s process imposes a racial penalty by systematically disfavoring Asian-American applicants. It does so in part through the subjective personal rating that admissions officers apply with minimal guidance or supervision. That rating produces consistently poorer scores for Asian Americans. Harvard did not prove that the personal rating is race-neutral. The district court nevertheless ruled for Harvard by positing alternative explanations that might account for the disparity. In doing so, it relieved Harvard of its well-established burden of proof.
II. Because Harvard failed to show that its process is narrowly tailored, the Court need not decide whether Harvard has asserted a sufficient diversity interest. But to rule for Harvard, the Court would need to confront this issue head-on. Articulating a concrete objective is critical to ensure that considering race is not an end in itself, and to enable judicial review of a school’s claim that its
- 10 -
asserted ends cannot effectively be achieved by alternative, race-neutral means. Harvard, however, has not articulated any such objective. Although its process demonstrably results in (and is engineered to produce) incoming classes with consistent racial compositions, it does not claim a compelling interest in maintaining such stability. Nor does Harvard identify any other concrete, cognizable educational goal by which its use of race can be tested. That failure frustrates judicial review. And Harvard’s failure to provide an objective measure should not redound to its benefit. Unless a court can conclude with confidence that a school’s use of race is lawful, it is not.
Title VI forbids schools that receive federal assistance from discriminating based on race. 42 U.S.C. 2000d. A recipient violates that prohibition by engaging in discrimination barred by the Equal Protection Clause. See Gratz v. Bollinger, 539 U.S. 244, 276 n.23 (2003). Because Harvard receives Title VI funds and considers race in its admissions process, it must satisfy strict scrutiny, which requires showing that its use of race is narrowly tailored to serve a compelling interest. E.g., Fisher v. University of Tex. at Austin, 570 U.S. 297, 309-311 (2013) (Fisher I); Parents Involved in Cmty. Schs. v. Seattle Sch. Dist. No. 1, 551 U.S. 701, 720 (2007). The district court erred in concluding that Harvard carried its burden.
- 11 -
I HARVARD FAILED TO PROVE THAT ITS USE OF RACE IS NARROWLY TAILORED
Strict scrutiny required Harvard to “prove that the means chosen by” the school to achieve its stated interest in diversity “are narrowly tailored to that goal.” Fisher I, 570 U.S. at 311. “On this point,” Harvard “receives no deference.” Ibid. Whether its chosen means for attaining racial diversity are “‘specifically and narrowly framed to accomplish that purpose’” is a question “for the courts, not for university administrators.” Ibid. (quoting Grutter, 539 U.S. at 333). Harvard did not carry that burden. The evidence shows that it engages in racial balancing and that its process imposes an unlawful racial penalty on Asian-American applicants.
A. Harvard Engages In Racial Balancing
“Outright racial balancing” is “patently unconstitutional” under the Equal Protection Clause, Grutter, 539 U.S. at 330, 343, and thus also violates Title VI. Imposing flat racial quotas clearly violates that prohibition. See id. at 334 (“To be narrowly tailored, a race-conscious admissions program cannot use a quota system.”). But as the district court recognized, forbidden racial balancing can exist even if a school does not seek precise, perfectly stable numbers. “[A] university could run afoul of Title VI’s prohibition on quotas even where it stopped short of defining a specific percentage and instead allowed some fluctuation around a
- 12 -
particular number.” ADD113 (citing Parents Involved, 551 U.S. at 712). The record shows that Harvard has engaged in that kind of racial balancing.
1. The Remarkably Stable Racial Composition Of Harvard’s Incoming Classes Reflects Racial Balancing
a. The racial breakdown of Harvard’s admitted classes over time reflects that they are the product of deliberate racial balancing. The following table shows the minority racial composition of Harvard’s admitted classes over an eight-year span, calculated using what Harvard now calls its “Old Methodology”:
|Percentage of Admitted Class by Race – Harvard’s Old Methodology|
|Class of 2010||Class of 2011||Class of 2012||Class of 2013||Class of 2014||Class of 2015||Class of 2016||Class of 2017||Range|
|Combined African American and Hispanic American||20.1||20.4||18.9||21.0||19.9||22.7||19.3||21.8||3.8|
JA4434-4435. The stability of each minority race’s representation in the class over time is striking. The percentage of each class consisting of African Americans stayed within a 1.6-percentage-point range (10.0% to 11.6%). Likewise, the percentage of Hispanic Americans stayed within a 2.3-percentage-point band
- 13 -
(8.8% to 11.1%). The combined percentage of those two groups varied no more than 3.8 percentage points (18.9% to 22.7%).
Harvard now prefers a different methodology (the “New Methodology”). See ADD28 & n.22 (explaining methodology preference). But that new rubric paints the same jarring picture for the years analyzed (2014-2017):
|Percentage of Admitted Class by Race – Harvard’s New Methodology|
|Class of 2014||Class of 2015||Class of 2016||Class of 2017||Range|
|Combined African American and Hispanic American||21.0||23.7||21.1||22.9||2.7|
JA4446-4447. Once again, the fraction of each year’s admitted class composed of members of particular races remained remarkably stable. The percentage of African Americans stayed within the same 1.6-percentage-point range (10.0% to 11.6%), and that of Hispanic Americans stayed within a 2.1-percentage-point band (10.0% to 12.1%). And the combined percentage of those two groups under Harvard’s preferred methodology varied only within 2.7 percentage points (21.0% to 23.7%).
- 14 -
These numbers speak for themselves. The minimal variation, including in the percentages of underrepresented minorities that Harvard seeks to benefit, over a multi-year period is much narrower than the 6.6-percentage-point range in underrepresented minorities the Supreme Court sustained in Grutter. 539 U.S. at 336.
b. The district court discounted those stable percentages as evidence of racial balancing. ADD80-82. The court relied on Harvard’s repackaging of the same data using a different, irrelevant metric that compared variation within each racial group’s representation year by year. ADD81 (citing JA6114-6115; JA5735-5742). For example, that metric portrays the 1.6 percentage-point decline in the percentage of African Americans from 2015 (11.6%) to 2016 (10.0%) as a drop of more than 13%. But that comparison elides the central question: how stable is the racial composition of the class over time.
A simple example illustrates the court’s error. Suppose racial group A fluctuates back and forth between 1% and 2% of the class from year to year, and racial group B similarly fluctuates between 10% and 11%. In any year, each group increases or decreases by only one percentage point. If that pattern persisted for years, it would be evidence of racial balancing. Yet Harvard’s intra-group-comparison methodology that the district court adopted tells an entirely different story. For group A, it would portray the one-percentage-point fluctuation between
- 15 -
1% and 2% as a series of 100% increases followed by 50% decreases. And for group B, Harvard’s method would instead depict the same pattern of one-percentage-point increases and decreases as a series of 10% increases followed by 9.1% decreases. That analysis distorts the changes in the class’s composition, arbitrarily magnifying the size of changes in a group that is smaller to start. It accordingly obfuscates the parallel stability in the two groups’ representation in the class over time. The district court should have rejected Harvard’s effort to obscure the results its process has produced.
2. Harvard Achieves Racial Balancing By Considering Race Throughout The Admissions Process And Continually Monitoring And Reshaping The Class’s Racial Composition
The decidedly stable racial composition of Harvard’s incoming classes should come as no surprise. The school considers race at virtually every step of its admission process. And its officials constantly monitor and continually reshape the racial makeup of each admitted class as it emerges. Those mechanisms confirm that Harvard’s racial balancing is no accident; it is engineered.
a. Harvard considers race at nearly every step of its admissions process. Admissions readers “generally begin with the application summary sheet”—“a two or three page document that is prepopulated with much of the key information about an applicant,” including the applicant’s race. ADD22. The admissions office’s regional subcommittees consider race in making initial admissions
- 16 -
recommendations to the full committee. ADD24. During some subcommittee meetings, the applicant’s summary sheet, including his race, is projected on a large screen for all decisionmakers to see. ADD23-24; see also, e.g., JA2108-2109 (Ray). Admissions officers may give an applicant a racial bonus even when an application does not discuss the applicant’s racial identity. ADD13; ADD104; see also, e.g., JA654-657 (Fitzsimmons); JA1117-1118 (Looby). Finally, admissions officers again consider race during the so-called “lop process,” when the committees cut down their list of prospective admitted students at the end of the cycle. ADD26; see also JA2046-2048 (Kim); JA2113-2116 (Ray). Indeed, race represents one of the five important factors on the “lop” sheet. JA4156 (lop sheet example); JA2113-2116 (Ray).
Harvard’s multi-stage consideration of race far exceeds the “one stage only” use of race approved by the Supreme Court in Fisher v. University of Texas at Austin, 136 S. Ct. 2198, 2207 (2016) (Fisher II). In upholding the university’s process in that case, the Court stressed that the University of Texas considered race at only one step: in calculating a “Personal Achievement Score.” Id. at 2206-2207. That score was one component of an applicant’s “Personal Achievement Index.” Id. at 2205-2206. Admissions officers then made admissions decisions based on the Personal Achievement Index and a race-neutral “Academic Index.” Id. at 2205, 2207. Crucially, they “ma[d]e that decision without knowing the
- 17 -
applicant’s race.” Id. at 2207. As the Supreme Court observed, race was thus “but a ‘factor of a factor of a factor’ in the holistic-review calculus.” Ibid. (citation omitted). “Race enters the admissions process, then, at one stage and one stage only—the calculation of the [Personal Achievement Score].” Ibid.
In contrast, race infiltrates Harvard’s admissions process at essentially every stage. Harvard admissions officers consider race when they score applicants, when they make initial admissions decisions, and when they winnow the list of possible admittees. That continual consideration of race for each applicant provides a ready way for the school to maintain the stable composition reflected in the data.
Moreover, Harvard’s process accords significant weight to race—and that weight makes a material difference. For example, Harvard’s expert presented an analysis of “average marginal effects,” i.e., the change in the likelihood of admission associated with changing one particular variable. ADD65-66; ADD79-80. For example, if left-handed applicants’ average chance of being admitted is 10%, and a school establishes a preference for left-handed applicants that increases their average chance of being admitted to 15%, the average marginal effect is five percentage points—the difference between the actual rate (15% with the preference) and the base rate (10% without it). Here, Harvard’s expert testified that the base rate for African American applicants is approximately 3%, and the average marginal effect of being African American in Harvard’s process is six
- 18 -
percentage points. JA3111-3112, JA3247-3248 (Card). Harvard’s process thus triples the likelihood of being admitted for African-American applicants (from 3% to 9%). Similarly, Harvard’s expert testified that the average marginal effect for being Hispanic was 3.73 percentage points, which roughly doubles the chances of admission for Hispanics. JA3111-3112. The effect of race is even more pronounced for the strongest applicants. Harvard’s own expert calculated that, for the strongest applicants, the average marginal effect of Hispanic identity was approximately 30 percentage points, and the average marginal effect of African-American identity was more than 40 percentage points. See JA6111-6112; JA3042-3046 (Card).
The significant boost that Harvard’s process accords minorities that it favors affects admissions outcomes. The district court found that the bump was “determinative” for “approximately 45% of all admitted African American and Hispanic applicants.” ADD84. And “[a]t least 10% of Harvard’s admitted class, including more than one third of the admitted Hispanics and more than half of the admitted African Americans, would most likely not be admitted in the absence of Harvard’s race-conscious admissions process.” Ibid. Harvard’s process thus not only has the potential to alter the racial composition of its incoming classes, but in fact does so.
- 19 -
b. When Harvard admissions officials consider each applicant’s race, they do not consider the race of that applicant in isolation as part of an individualized inquiry. Instead, throughout the admissions cycle, admissions leaders track the racial composition of the entire class. ADD27. They do so to see how the prospective class is “shaping up,” JA1392-1393 (McGrath), and then convey that information to admissions officials making decisions.
Throughout the process, Harvard’s admissions leaders secure as many as twenty “one-pagers”—documents that compare the racial demographics of the projected class to the demographics of the previous year’s class. JA1392-1397 (McGrath); JA5982 (documenting receipt of one-pagers and similar documents); JA1869-1883 (Yong). The district court found that the dean would “share the breakdown of the admitted class as reflected on the one-pagers with the full committee from time to time.” ADD28. “For example, at the start of the full Admissions Committee meetings, he usually states how many students are being recommended for admission by the subcommittees and how the breakdown of the class compares to the prior year in terms of racial identities and other demographics.” Ibid. The dean also requested “a one pager and his ethnic stats” before the lopping process. JA4112 (e-mail reflecting Dean Fitzsimmons’s request).
- 20 -
The trial evidence also leaves little doubt what Harvard’s admissions leaders are looking for: maintaining Harvard’s current racial composition. The “Smith Committee”—a group of three deans—was tasked in 2017 with examining race-neutral alternatives to Harvard’s current race-based process. ADD40. That committee concluded “that no workable race-neutral alternatives will currently permit Harvard to achieve the level of racial diversity it has credibly found necessary for its educational mission.” ADD83. But the committee’s own description of the school’s diversity objective speaks volumes. The committee focused on achieving “a student body comparable in diversity to current classes” as well as “levels of racial diversity commensurate with those at the College today.” JA4421 (emphases added). It analyzed the magnitude of a race-neutral alternative necessary “to reach the current level of” underrepresented minorities “admitted to Harvard.” JA4426 (emphasis added). And it weighed the impact of another race-neutral alternative by determining what would be necessary “to produce a combined proportion of” underrepresented minorities “comparable to that of Harvard’s current classes.” JA4430 (emphasis added). The committee’s chair conceded that “[they] weren’t looking to go backwards for any particular racial group.” JA1841 (Smith); accord JA1830 (Smith).
It is little wonder Harvard’s leaders failed to identify an adequate race-neutral alternative that would meet the school’s goal, given that their goal was to
- 21 -
maintain Harvard’s existing racial makeup. For present purposes, the key point is that Harvard’s leaders are not only acutely aware of each admitted class’s racial composition and how it compares to previous classes, but they actively seek to maintain that makeup. Seeking to “assure within its student body some specified percentage of a particular group merely because of its race or ethnic origin” “amount[s] to outright racial balancing.” Grutter, 539 U.S. at 329-330 (citations omitted).
The district court downplayed this evidence of racial balancing. It observed that the law school in Grutter also issued reports that tracked racial composition. See ADD115 (citing Grutter, 539 U.S. at 336). But Harvard’s ongoing reliance on such reports is materially different. The Court in Grutter did not indicate, for example, that the reports the law school considered compared the racial composition to the prior year’s composition. See also Grutter v. Bollinger, 137 F. Supp. 2d 821, 832 (E.D. Mich. 2001) (daily reports tracked racial composition of current applicant pool), rev’d, 288 F.3d 732 (6th Cir. 2002) (en banc), aff’d, 539 U.S. 306 (2003). Nor was there any indication that the law school’s dean read the comparative statistics aloud at the start of full committee meetings convened to discuss admitting applicants.
The district court also noted that Harvard considers the racial distribution of its admitted students to assist in predicting its yield rate, so as to avoid over-
- 22 -
enrolling the class. ADD29; ADD116. But Harvard’s one-pagers most often do not include yield rates. See JA4130-4146. Moreover, admissions officers testified that the Dean discussed racial breakdowns of the prospective class, not that he told them to think about their yield rate. See, e.g., JA2109-2113 (Ray); JA2516-2518 (Howrigan). In any event, the previous year’s racial composition itself has no apparent bearing on the yield rate. Harvard thus has not explained why—apart from the school’s objective of racial balancing—it provided that particular datapoint to decisionmakers. At any rate, contrary to the district court’s suggestion, ADD116, nothing in the Supreme Court’s decisions permits a school to favor or disfavor applicants of a particular race based on generalizations about their behavior, such as the likelihood that they will accept an offer of admission.
* * *
Viewed in light of that and other trial evidence, the district court’s conclusion that Harvard merely “considered” race “as a ‘plus’ factor in the context of individualized consideration of” individual applicants, ADD108 (citation omitted), does not hold up. Harvard takes account of race throughout its process. And that process consistently produces incoming classes with strikingly similar racial compositions. Wherever the outer limits of permissible consideration of race lie under the Supreme Court’s precedents, Harvard’s practice exceeds them.
- 23 -
B. Harvard’s Use Of Race Penalizes Asian Americans
Harvard’s failure to refute the powerful evidence that it engages in racial balancing should be fatal. But the school’s admissions process also is invalid for an additional reason. A race based admissions program that penalizes “individuals who are not members of the favored racial and ethnic groups” cannot satisfy narrow tailoring. Grutter, 539 U.S. at 341 (citation omitted). The predictable and now proven result of Harvard’s efforts to benefit African Americans and Hispanics has been to impose a racial penalty on Asian Americans. Even if Harvard were not actively seeking to maintain the overall racial composition of its incoming classes, it may not ignore the harmful effects of its efforts to assist certain racial groups on the members of another.
1. Harvard’s Own Statistics Reflect A Penalty Against Asian Americans
The parties disputed the statistical evidence extensively, but one point should be common ground: Asian-American applicants have become a casualty of Harvard’s efforts to boost members of other races. The plaintiff points (at 29-43) to statistical models showing that Asian Americans fare less well than white applicants when one variable—the personal rating each applicant was assigned—is excluded from the model. The district court downplayed that evidence based on its view that the personal rating should not be excluded from the models. ADD68-73; see also ADD109-112. The resolution of that statistical-methodology dispute
- 24 -
depends, in turn, on whether the personal rating itself is influenced by race. ADD68-73; ADD109; Pl.’s Br. 30. As discussed below, strong evidence exists to support the plaintiff’s position that the personal rating is influenced by race, and the district court improperly failed to require Harvard to carry its burden of proving otherwise. See Part I.B.2., infra. But even setting aside the dispute about the competing statistical models for assessing whether Harvard penalizes Asian Americans as compared to white applicants, the evidence clearly shows that Harvard imposes a racial penalty on Asian Americans as compared to members of other minority races that Harvard favors.
Grutter’s test requires examining burdens imposed on “individuals who are not members of the favored racial and ethnic groups.” Grutter, 539 U.S. at 341 (citation omitted). That test necessarily contemplates a comparison between members of racial groups that are favored and those that are not. The data relied upon by Harvard and the district court demonstrate that Harvard imposes a penalty on Asian Americans as compared to African Americans and Hispanics.
Based on Harvard’s calculations, the district court found that “[i]n the absence of any other adjustments to Harvard’s admissions policy, eliminating consideration of race would cause the African-American representation at Harvard to decline from approximately 14% to 6% of the student population and Hispanic representation to decline from 14% to 9%.” ADD84; JA4420. Conversely, using
- 25 -
the framework accepted by the district court, Asian-American representation at Harvard would, under a race-neutral admissions program, increase from 24% to 27%. See JA6121. Put differently, Harvard’s use of race inflicts an 11.1% penalty (i.e., the decrease from 27% representation to 24% representation) on Asian Americans while simultaneously providing a 133% bonus (i.e., the increase from 6% representation to 14% representation) to African Americans.
In addition, data Harvard submitted show that the percentage of Asian Americans admitted in each class was, without exception, smaller in a 40-year span (from 1980 to 2019) than the percentage of Asian Americans in the applicant pool. JA5743-5744. In contrast, the percentage of African Americans admitted in each incoming class was almost invariably higher than the percentage of African- American students who applied. See ibid. The same is true of Hispanics. Ibid.
The district court never grappled with this issue in concluding that Harvard does not impose a racial penalty. It focused on comparing Asian-American applicants with white applicants. See ADD109-112. But that comparison is incomplete. Nothing in Grutter or the Supreme Court’s other decisions suggests that a court should compare the members of a college’s least favored race only against members of its second-least-favored race, or that a school may freely discriminate in favor of one minority racial group and against another. Instead, Grutter’s framework contemplates a spectrum-wide comparison between
- 26 -
Harvard’s disfavored racial groups and its “favored racial and ethnic groups,” Grutter, 539 U.S. at 341 (citation omitted)—in this case, African-American and Hispanic applicants.
2. The District Court Relieved Harvard Of Its Obligation To Justify The Lower Personal Ratings It Assigns To Asian-American Applicants
Although the starkly disparate outcomes for Asian Americans as compared to other minority groups may stem from multiple aspects of Harvard’s race-based admissions process, it is clear that Harvard’s subjective personal rating of applicants is a contributor. The personal rating plays a significant role in admissions decisions. And the district court found that, “holding constant any reasonable set of observable characteristics,” “the data demonstrates a statistically significant and negative relationship between Asian American identity and the personal rating assigned by Harvard admissions officers.” ADD69. Although the court credited Harvard’s assertions that its “admissions officers do not believe that Asian American applicants, as a group, have worse personal qualities than other applicants,” it found that they “assign Asian American applicants personal ratings that are, on average, slightly weaker than those assigned to applicants from other racial groups.” ADD55 (citing JA4530). In other words, Harvard’s admissions officers tended to evaluate Asian Americans, as compared to members of other racial groups, as having less integrity, being less confident, constituting less-qualified leaders, and so on.
- 27 -
The district court acknowledged that why “Asian American applicants score lower on the personal rating” was an important question. ADD109. If that personal rating itself embodies racial bias, then it is one mechanism by which Harvard penalizes Asian Americans. Moreover, if the personal rating is influenced by race, then it should be excluded from statistical models designed to measure the effect of race on admissions decisions. See ADD63 (citing JA2253-2254, JA2377 (Arcidiacono) & JA3219-3220 (Card)); ADD73. That is significant because, when the personal rating is excluded, all of the models presented show a statistically significant penalty against Asian Americans as compared to white applicants. ADD66; ADD74-79; JA2317 2318 (Arcidiacono); JA3149-3152 (Card); see also Pl.’s Br. 30.
The district court posited various potential explanations for Asian Americans’ undisputedly lower personal-rating scores. ADD109-110. It acknowledged that the disparity might reflect “overt discrimination or implicit bias at work to the disadvantage of Asian American applicants.” ADD110. Alternatively, the court speculated “that the self-selected group of Asian Americans that applied to Harvard during the years” at issue might “not possess the personal qualities that Harvard is looking for at the same rate as white applicants.” ADD109. Ultimately the court concluded that it “c[ould not] definitively explain the difference in the personal ratings.” ADD110. That
- 28 -
uncertainty should have ended the analysis and resulted in a ruling against Harvard, which “must prove” that its “means chosen” to attain diversity “are narrowly tailored to that goal.” Fisher I, 570 U.S. at 311.
Instead, the district court excused Harvard from carrying its burden by treating the lack of conclusive evidence as to why Harvard assigns lower personal ratings to Asian Americans as a reason to sustain Harvard’s process. See ADD109-110. Finding “no evidence of any discriminatory animus or conscious prejudice,” the court concluded that the disparity “reflects neither intentional discrimination against Asian American applicants nor a process that was insufficiently tailored to avoid the potential for unintended discrimination.” ADD111. That analysis has things backward. Unlike rational-basis review, under which a policy must be sustained so long as “there is any reasonably conceivable state of facts” that could support it, Heller v. Doe, 509 U.S. 312, 320 (1993) (citation omitted), strict scrutiny requires more than “assertion and conjecture,” Republican Party of Minn. v. White, 536 U.S. 765, 781 (2002) (citation omitted). The district court also dismissed the statistically significant disparity as “small,” and because it “did not burden Asian American applicants significantly more than Harvard’s race-conscious policies burdened white applicants.” ADD110-111. But the court cited no legal authority for disregarding the penalty imposed on Asian Americans on either ground.
- 29 -
II HARVARD FAILED TO ARTICULATE A SUFFICIENTLY CONCRETE AND MEASURABLE INTEREST
Given Harvard’s failure to prove that its use of race in admissions is narrowly tailored to the school’s asserted interest, this Court need not address whether Harvard’s asserted interest in diversity is sufficiently concrete to satisfy strict scrutiny. To sustain Harvard’s policy, however, the Court would also have to confront whether Harvard has articulated an interest that is “sufficiently measurable to permit judicial scrutiny.” Fisher II, 136 S. Ct. at 2211. Apart from an apparent and impermissible objective of maintaining the existing racial composition of its class, see Part I.A., supra, Harvard has offered only an “elusory” and “amorphous” diversity goal that is incapable of objective analysis. Fisher II, 136 S. Ct. at 2211.
Harvard’s own officials testified repeatedly that its diversity objective could not be measured. For example, when asked by the district court how Harvard “measure[d] what’s enough,” the college’s dean responded:
I don’t think there’s like a quantitative number. What it is, is that when you have a lot of people who share different backgrounds and experiences and no one is more salient than others; that is, you have just a lot of ways that you have to understand each person as an individual, that you’ve had enough experiences where your notions of stereotypes are broken, I think that educational experience is what gets us [t]o that point.
- 30 -
JA1750 (Khurana); see also, e.g., JA660 (Fitzsimmons, Dean of Admissions) (“There are no formulas. There are no sort of numerical guidelines that would do justice to anything like that. It just isn’t that simple.”). And Harvard’s outside diversity expert candidly explained that the school lacks the “capacity to know when enough diversity is enough diversity.” JA2784 (Simmons); accord JA2801.
Harvard’s admitted inability to measure progress toward its stated diversity goal independently renders its race-based process infirm. See Fisher II, 136 S. Ct. at 2210-2211. And deeming such an interest sufficient would frustrate, if not foreclose, meaningful judicial review. Courts cannot perform their function of carefully scrutinizing whether a school’s policy is narrowly tailored to advance a compelling interest, see Fisher I, 570 U.S. at 311, if progress toward that interest using a particular policy cannot be assessed.
The lack of a measurable interest in diversity also impedes judicial scrutiny of whether adequate race-neutral alternatives are available. Plaintiffs here contend (at 56-62) that viable alternatives to achieve racial and socioeconomic diversity exist but that Harvard considered such alternatives only belatedly and superficially after this litigation was underway. The district court found that none of the proposed alternatives would enable Harvard to “reach the level of racial diversity that it believes necessary.” ADD85 (emphasis added); accord ADD83. But because Harvard never articulated any measurable diversity objective, the court
- 31 -
had no way of verifying Harvard’s bare assertion that no race-neutral alternative could adequately advance the same objective. The court in effect was forced to take the school’s word for it—precisely the type of “deference” the Supreme Court has prohibited. Fisher I, 570 U.S. at 311.
In addition, although the Supreme Court has emphasized that “race-conscious admissions policies must be limited in time,” Grutter, 539 U.S. at 342, a court’s inability to compare a school’s policy against any objective measure risks “[e]nshrining a permanent justification for racial preferences,” which “would offend * * * fundamental equal protection principle[s].” Ibid. Harvard’s inability to identify objectively measurable goals that its race-based admissions process advances provides further reason to hold that it has failed to carry its burden.
- 32 -
The judgment below should be reversed.
ERIC S. DREIBAND
Assistant Attorney General
ELLIOTT M. DAVIS
Acting Principal Deputy Assistant Attorney General
s/ Matthew J. Donnelly
THOMAS E. CHANDLER
MATTHEW J. DONNELLY
U.S. Department of Justice
Civil Rights Division
Benjamin Franklin Station
P.O. Box 14403
Washington, D.C. 20044-4403
CERTIFICATE OF COMPLIANCE
I certify that the attached BRIEF FOR THE UNITED STATES AS AMICUS CURIAE SUPPORTING APPELLANT AND URGING REVERSAL:
(1) complies with the type-volume limitation of Federal Rules of Appellate Procedure 29(a)(5) and 32(a)(7)(B) because, excluding the parts of the document exempted by Federal Rule of Appellate Procedure 32(f), it contains 6488 words; and
(2) complies with the typeface requirements of Federal Rule of Appellate Procedure 32(a)(5) and the type-style requirements of Federal Rule of Appellate Procedure 32(a)(6) because it has been prepared in a proportionally spaced typeface using Microsoft Word 2019, in 14-point Times New Roman font.
s/ Matthew J. Donnelly
MATTHEW J. DONNELLY
Dated: February 25, 2020