Parents Involved in Community Schools v. Seattle School District, No. 1

Opinion by Judge O’SCANNLAIN; Dissent by Judge GRABER

O’SCANNLAIN, Circuit Judge.

Following the Washington Supreme Court’s resolution of certified state-law questions, we must decide whether the use of race in determining which students will be admitted to oversubscribed high schools in Seattle, Washington, violates the federal Constitution’s Equal Protection Clause.

I

This opinion marks the fourth time a federal court has addressed the Seattle Public Schools’ use of an explicit “racial tiebreaker” in choosing which student applicants it will admit to the City’s most popular public high schools. See Parents Involved in Cmty. Schs. v. Seattle Sch. Dist. No. 1, 137 F.Supp.2d 1224 (2001) [Parents Involved I], rev’d, 285 F.3d 1236 (9th Cir.2002) [Parents Involved II], withdrawn, 294 F.3d 1084 (9th Cir.2002), certifying questions, 294 F.3d 1085 (9th Cir. 2002) [Parents Involved III]. We draw the following restatement of facts largely from Parents Involved II.

A

Seattle School District Number 1 (the “School District”) operates ten public high schools: Ballard, Chief Sealth, Cleveland, Franklin, Garfield, Ingraham, Nathan Hale, Rainier Beach, Roosevelt, and West Seattle. Four of these (Ballard, Ingraham, Nathan Hale, and Roosevelt) are located north of downtown Seattle; of the remaining six, five (Chief Sealth, Cleveland, Franklin, Garfield, and Rainier Beach) are located south of downtown, and one (West Seattle) is located directly west of downtown.

These schools vary widely in quality, as measured by such factors as standardized test scores,¹ numbers of college preparatory and Advanced Placement (AP) courses offered and the availability of an Internal Baccalaureate (IB) program, percentages of students taking AP courses and SATs, percentages of graduates who attend college, Seattle Times college-preparedness rankings, University of Washington rankings, and disciplinary statistics. Moreover, some of the schools offer unique educational programs or opportunities not offered in other schools.²

The School District has never been segregated by law. However, due to Seattle’s racially imbalanced housing patterns,³ if Seattle’s children were simply assigned to the high schools nearest their homes, those schools would tend to reflect such imbalance. That is, the demographic profile of the individual high schools would not mirror the demographic makeup of the city’s student population as a whole.⁴ As part of its continuing efforts to prevent such imbalance and to promote racial diversity in its high schools, the School District has adopted an open choice plan instead of simply assigning students to the high schools nearest their homes. Pursuant to this system, each student may choose to attend any of the ten high schools in the city, so long as there is room available in that school.

The District’s open choice plan provides for a multi-step application process. Each student is first asked to rank the high schools he or she would like to attend. If a student is not admitted to his or her first-choice school because that school is full, the School District attempts to assign him or her to his or her second-choice school, and so on. If a student is not admitted to any of his or her chosen schools, he or she receives a mandatory assignment to a school with available space.

Not surprisingly, a significant problem arises when a school becomes “oversubscribed” — that is, when more students want to attend that school than there are spaces available. For the academic year 2000-01, five of the School District’s high schools were oversubscribed and five were undersubscribed.⁵ The magnitude of over-subscription during the 2000-01 school year underscores its problematic nature: Approximately 82 percent of students selected one of the oversubscribed high schools as their first choice, while only about 18 percent picked one of the under-subscribed high schools as their first choice.

To resolve the dilemma of oversubscription, the School District’s high school assignment plan uses a series of four “tiebreakers” to determine which students will be admitted to each oversubscribed school. The first tiebreaker gives a preference to students with siblings already attending the requested school. This tiebreaker accounts for somewhere between 15 percent and 20 percent of high school assignments. If a school is still oversubscribed after applying this first tiebreaker, the School District proceeds to a second tiebreaker, which is based entirely on race. For purposes of the racial tiebreaker, students are deemed to be of the race specified in their registration forms, which ask parents to identify their child’s race. Because registration must be completed in person by a parent, if a parent declines to specify a racial category, the School District assigns the student a category based on a visual inspection of the parent (and, if present, the student) at the time of registration. It is this second — racial—tiebreaker that spawned the present suit.

Use of the racial tiebreaker is designed to balance the racial makeup of the city’s public high schools. Accordingly, if an oversubscribed school’s demographic profile deviates from the overall demography of Seattle’s student population (approximately 40 percent white and 60 percent non-white) by more than a set number of percentage points, the School District designates that school “integration positive.” The racial tiebreaker is then applied in the course of determining admissions to such schools, so that students whose race (coded by the School District simply as white or non-white) will push an integration positive school closer to the desired racial ratio are automatically admitted.⁶ Thus, at Franklin (for instance), whites are admitted preferentially because they are white; and at Ballard, non-whites are admitted preferentially because they are not white.⁷ Ultimately, the School District’s use of this racial tiebreaker determines where about 10 percent of applicants will be admitted.

Once all students of the preferred racial category are admitted to an oversubscribed high school, any remaining “ties” are broken by resort to a third variable: distance. Quite simply, applicants are admitted on the basis of the mileage between their homes and the school to which they seek admission, with those who live closest admitted first. Although a fourth tiebreaker exists — a random lottery — it rarely is invoked because distances are calculated to one hundredth of a mile for purposes of the preceding tiebreaker.

B

Parents Involved in Community Schools (“Parents”) is “a nonprofit corporation formed by parents whose children have been or may be denied admission to the high schools of their choosing solely because of race.” It commenced this legal action in July of 2000, contending that the School District’s use of the racial tiebreaker for high school admissions is illegal under both state and federal law. Specifically, Parents alleged that by using race to decide who will be admitted to the oversubscribed high schools, the School District engages in illegal racial discrimination prohibited by the Washington Civil Rights Act (“Initiative 200”),⁸ the Equal Protection Clause of the Fourteenth Amendment,⁹ and Title VI of the Civil Rights Act of 1964.¹⁰

Both Parents and the School District moved for summary judgment on all claims; neither contended that genuine issues of material fact precluded summary judgment. In a published opinion dated April 6, 2001, the district court upheld the use of the racial tiebreaker under both state and federal law, granting the School District’s motion and denying the Parents’s. See Parents Involved I, 137 F.Supp.2d at 1240. Parents timely filed an appeal in this court and, on April 16, 2002, we issued an opinion reversing the district court’s decision. Acting “in our constitutionally ordained role as oracles of Washington law,” Parents Involved II, 285 F.3d at 1243, we prophesied that the School District’s use of the racial tiebreaker violated Initiative 200. Id. at 1244.¹¹ Simultaneously, we enjoined the School District from using the racial tiebreaker in its system of high school admissions pending further order from this court. Id. at 1257.

While the School District’s petitions for rehearing and rehearing en banc were pending before us, it “bec[a]me clear that [we could not] provide a definitive [legal] answer before assignments [were to] be made for the 2002-03 year, and therefore, ... that our sole reason for not certifying this question to the Washington Supreme Court ha[d] dissolved.” Parents Involved III, 294 F.3d at 1086. Consequently, we granted the petition for rehearing, withdrew our opinion, and vacated our injunction. See id. Simultaneously, we entered an order certifying to the Supreme Court of Washington the question whether

[b]y using a racial tiebreaker to determine high school assignments, [the] Seattle School District Number 1 “discriminate[s] against, or grantfs] preferential treatment to, any individual or group on the basis of race, ... color, ethnicity, or national origin in the operation of public education” in violation of Initiative 200.... ?

Id. at 1087.

The Supreme Court of Washington accepted certification, heard oral argument in the matter, and on June 26, 2003 issued an opinion concluding that 1-200 “does not prohibit the Seattle School District’s open choice plan tie breaker based upon race so long as it remains neutral on race and ethnicity and does not promote a less qualified minority applicant over a more qualified applicant.” Parents Involved in Cmty. Schs. v. Seattle Sch. Dist. No. 1, 149 Wash.2d 660, 72 P.3d 151, 166 (2003). It therefore “returned] the case to the federal court for further proceedings consistent with [its] resolution of the questions of Washington law,” id. at 167, and formally notified this court of its actions by delivery of a Certificate of Finality on September 8, 2003.

All state law issues having been definitively decided, the parties prepared supplemental briefing on the remaining federal constitutional question in light of the Supreme Court’s intervening decisions in the University of Michigan affirmative actions cases—Gratz v. Bollinger, 539 U.S. 244, 123 S.Ct. 2411, 156 L.Ed.2d 257 (2003), and Grutter v. Bollinger, 539 U.S. 306, 123 S.Ct. 2325, 156 L.Ed.2d 304 (2003)—followed by reargument.

II

As a preliminary matter, we must address whether the passage of time has mooted Parents’s action. Article Ill’s case-or-controversy requirement mandates that the parties to a federal court action must “continue to have a personal stake in the outcome of the lawsuit” at all stages of the proceedings. United States v. Verdin, 243 F.3d 1174, 1177 (9th Cir.2001) (internal quotation marks and citation omitted). “This means that, throughout the litigation, the plaintiff must have suffered, or be threatened with, an actual injury traceable to the defendant and likely to be redressed by a favorable judicial decision.” Spencer v. Kemna, 523 U.S. 1, 7, 118 S.Ct. 978, 140 L.Ed.2d 43 (1998) (quoting Lewis v. Cont’l Bank Corp., 494 U.S. 472, 477-478, 110 S.Ct. 1249, 108 L.Ed.2d 400 (1990)). As with any jurisdictional inquiry — and notwithstanding the parties’ unhesitating agreement that Parents’s action remains a live controversy appropriately subject to federal adjudication on the merits — we are charged with an independent constitutional responsibility to verify our authority to resolve their litigation. See Dittman v. California, 191 F.3d 1020, 1025 (9th Cir.1999).

Initially, we have little doubt that the associational aspect of Parents’s standing has not been mooted. At reargument, counsel for Parents informed us that several of the association’s members have children who, over the course of the next several years, will be applying for admission to the School District’s public high schools and who thus will be subject to the admissions policies established by the School Board. Because “some members of the [association] [c]ould [continue to] have ... standing to bring this suit in their own right,” Int'l Union, United Auto., Aerospace & Agric. Implement Workers of Am. v. Brock, 477 U.S. 274, 286, 106 S.Ct. 2523, 91 L.Ed.2d 228 (1986), and because the passage of time has not called into question Parents’s satisfaction of the other requirements for associational standing, see Friends of the Earth, Inc. v. Laidlaw Envtl. Servs. (TOC), Inc., 528 U.S. 167, 181, 120 S.Ct. 693, 145 L.Ed.2d 610 (2000), we are satisfied that the associational aspect of Parents’s standing has continued vitality.

Perhaps more troubling, however, is the disclosure that the School District is not currently employing — and has not since the 2001-02 school year employed— the racial tiebreaker that Parents challenge in this litigation. As noted earlier, we enjoined the School District’s use of the racial tiebreaker with our initial disposition of this case. See Parents Involved II, 285 F.3d at 1257. And although we vacated that injunction with the withdrawal of our initial opinion, see Parents Involved III, 294 F.3d at 1086, the School District has voluntarily declined to reinstate its racial tiebreaker during the pendency of this litigation. With the passage of time, the voters of Seattle have elected a new School Board,¹² and there is at least a remote possibility that the new Board will opt not to resume its use of the racial tiebreaker that prompted this lawsuit.

Nonetheless, it is beyond cavil that “ ‘a defendant’s voluntary cessation of a challenged practice does not deprive a federal court of its power to determine the legality of the practice’ unless it is ‘absolutely clear that the allegedly wrongful behavior could not reasonably be expected to recur.’ ” Buckhannon Bd. & Care Home v. W. Va. Dep’t of Health & Human Res., 532 U.S. 598, 609, 121 S.Ct. 1835, 149 L.Ed.2d 855 (2001) (emphases added) (quoting Friends of the Earth, 528 U.S. at 189, 120 S.Ct. 693, with internal quotation marks and citations omitted). Indeed, in these circumstances, a “heavy burden of persuading the court that the challenged conduct cannot reasonably be expected to start up again lies tvith the party asserting mootness.” Adarand Const., Inc. v. Slater, 528 U.S. 216, 222, 120 S.Ct. 722, 145 L.Ed.2d 650 (2000) (emphasis in original) (internal quotation marks and citations omitted). The problem here, of course, is that neither party has asserted that this case is moot: When asked at oral argument about the possibility that we lack jurisdiction over Parents’s action, counsel for the School District not only maintained that this controversy remains live, but questioned whether the Board would have him defend the racial tiebreaker if it did not intend to reinstate the challenged policy in the future.¹³

Indeed, where a court must address sua sponte the possibility that the passage of time has mooted litigation on alternative grounds (for instance, that the associational aspect of a plaintiffs standing no longer satisfies Article III jurisdictional requirements), we find it hard to imagine that a defendant’s voluntary cessation could ever operate itself to moot the underlying litigation. By virtue of the fact that neither party will have alleged mootness in the first instance, there is no one to “satisfy the heavy burden of persuasion” that well-established doctrinal precepts require a party to demonstrate before voluntary cessation can be held to moot a once live case or controversy. See United States v. Concentrated Phosphate Export Ass’n, 393 U.S. 199, 203, 89 S.Ct. 361, 21 L.Ed.2d 344 (1968). We therefore conclude that the present case remains live.

Ill

We now turn to the heart of Parents’s claim: that the School District’s use of race to determine who will be admitted to its oversubscribed public high schools constitutes illegal racial discrimination in violation of both the Fourteenth Amendment and Title VI.¹⁴

A

Forged in the crucible of Reconstruction and “[purchased at the price of immeasurable human suffering,” Adarand Const., Inc. v. Peña, 515 U.S. 200, 240, 115 S.Ct. 2097, 132 L.Ed.2d 158 (1995) (Thomas, J., concurring), the Equal Protection Clause of the Fourteenth Amendment mandates that “No state shall ... deny to any person within its jurisdiction the equal protection of the laws.” U.S. Const, amend. XIV, § 1. “Because the Fourteenth Amendment ‘protects persons, not groups,’ all governmental action based on race' — a group classification long recognized as in most circumstances irrelevant and therefore prohibited — should be subjected to detailed judicial inquiry to ensure that the personal right to equal protection of the laws has not been infringed.” Grutter, 539 U.S. at 326, 123 S.Ct. 2325 (quoting Adarand, 515 U.S. at 227, 115 S.Ct. 2097 (1995) (quoting Hirabayashi v. United States, 320 U.S. 81, 100, 63 S.Ct. 1375, 87 L.Ed. 1774 (1943))) (emphasis in original); see also Ho v. S.F. Unified Sch. Dist., 147 F.3d 854, 865 (9th Cir.1998) (“It is as a person that each of us has these rights that are so majestically secured.”). Therefore, “any person, of whatever race, has the right to demand that any governmental actor subject to the Constitution justify any racial classification subjecting that person to unequal treatment under the strictest of judicial scrutiny.” Adarand, 515 U.S. at 224, 115 S.Ct. 2097; see also Coalition for Econ. Equity v. Wilson, 122 F.3d 692, 702 (“The standard of review under the Equal Protection Clause does not depend on the race or gender of those burdened or benefited by a particular classification .... ‘[A]ny individual suffers an injury when he or she is disadvantaged by the government because of his or her race.’ ”) (quoting Adarand, 515 U.S. at 230, 115 S.Ct. 2097).

For race-based educational policies “[t]o withstand strict scrutiny analysis, respondents must demonstrate that the[ir] use of race in [their] current admission program employs ‘narrowly tailored measures that further compelling governmental interests.’ ” Gratz, 539 U.S. at 270, 123 S.Ct. 2411 (quoting Adarand, 515 U.S. at 227, 115 S.Ct. 2097); see also Hunter v. Regents of Univ. of Calif., 190 F.3d 1061, 1063 (9th Cir.1999) (“To meet the strict scrutiny test, the Regents must demonstrate that ... consideration of race/ethnicity is narrowly tailored to serve a compelling governmental interest.”) (first emphasis added); Ho, 147 F.3d at 865 (“Once the plaintiffs established the School District’s use of racial classifications ... the School District has the duty to justify them.... At trial, the School District will bear the burden of proving that [its use of race] is a ‘narrowly tailored measure that furthers compelling government interests.’ ”) (quoting Adarand, 515 U.S. at 227, 115 S.Ct. 2097) (emphasis added); Monterey Mech. Co. v. Wilson, 125 F.3d 702, 713 (9th Cir.1997) (“The burden of justifying different treatment by ethnicity or sex is always on the government.”). Notwithstanding its remarkable assertions to the contrary, it is thus quite plainly the School District which bears the weighty burden of demonstrating that its use of the racial tiebreaker in its open choice admissions program satisfies the “most searching examination” demanded by strict scrutiny, Gratz, 539 U.S. at 270, 123 S.Ct. 2411 (quotations and citations omitted): that is, that the racial tiebreaker is designed to further a compelling governmental interest, and that the manner in which it does so is narrowly tailored to achieve that interest.

B

In papers prepared for purposes of this litigation, the School District has proffered an array of interrelated and putatively compelling interests in pursuit of which it seeks to employ the racial tiebreaker in its open choice high school admissions program. These myriad interests include the School District’s desires to achieve: “the educational benefits of attending a racially and ethnically diverse school”; “integration of schools which, as a result of housing patterns and the tendency of many parents to choose schools close to home, would otherwise tend to become racially isolated”; “ensuring that public institutions are open and available to all segments of American society”; “alleviating de facto segregation”; “increasing racial and cultural understanding”; “avoiding racial isolation”; fostering “cross-racial friendships”; and “reducing] prejudice and increasing] understanding of cultural differences.” Perhaps its most articulate statement supporting use of the'racial tiebreaker is the School Board’s policy “Statement Reaffirming [the] Diversity Rationale.” It explains:

Diversity in the classroom increases the likelihood that students will discuss racial or ethnic issues and be more likely to socialize with people of different races. Diversity is thus a valuable resource for teaching students to become citizens in a multi-raeial/multi-ethnie world.

Providing students the opportunity to attend schools with diverse student enrollment also has inherent educational value from the standpoint of education’s role in a democratic society.... Diversity brings different viewpoints and experiences to classroom discussions and thereby enhances the educational process. It also fosters racial and cultural understanding, which is particularly important in a racially and culturally diverse society such as ours.

Based on the foregoing rationale, the Seattle School District’s commitment is that no student should be required to attend a racially concentrated school. The District is also committed to providing students with the opportunity to voluntarily choose to attend a school to promote integration. The District provides these opportunities for students to attend a racially and ethnically diverse school, and to assist in the voluntary integration of a school, because it believes that providing a diverse learning environment is educationally beneficial for all students.

To the extent Parents once may have been able to make out a colorable claim that the only interest sufficiently compelling to justify the use of racial classifications is the remediation of past official discrimination,¹⁵ such an argument no longer obtains. In Smith v. University of Washington, this court followed Justice Powell’s solo concurrence in Regents of the University of California v. Bakke, 438 U.S. 265, 98 S.Ct. 2733, 57 L.Ed.2d 750 (1978), observing that “educational diversity is a compelling governmental interest that meets the demands of strict scrutiny of race-conscious measures.” 233 F.3d 1188, 1201 (9th Cir.2000). And in its landmark 2003 opinion in Grutter, the Supreme Court settled any debate over the validity of employing racial preferences for non-remedial purposes by asserting that it had “never held that the only governmental use of race that can survive strict scrutiny is remedying past discrimination.” Grutter, 539 U.S. at 328, 123 S.Ct. 2325 (O’Connor, J.).¹⁶ Indeed, it expressly sanctioned the so-called “diversity rationale” articulated by the University of Michigan in support of employing such preferences in determining which applicants would be offered admission to its selective law school. See id. 328-33, 123 S.Ct. 2325. It is upon Justice O’Connor’s elaboration of the diversity rationale that we now focus our attention.

1

In part due to a recognition that the diversity rationale had often been criticized as “amorphous,” “abstract,” “malleable,” and “ill-defined,” see, e.g., Metro Broad., 497 U.S. at 612, 110 S.Ct. 2997 (O’Connor, J., dissenting); Wessmann v. Gittens, 160 F.3d 790, 796 (1st Cir.1998); Lutheran Church-Missouri Synod v. FCC, 141 F.3d 344, 354 (D.C.Cir.1998); Johnson v. Bd. of Regents, 106 F.Supp.2d 1362, 1371 (S.D.Ga.2000); Tracy v. Bd. of Regents, 59 F.Supp.2d 1314, 1321 (S.D.Ga.1999); cf. Grutter, 539 U.S. at 350 & 354 n. 3, 123 S.Ct. 2325 (Thomas, J., dissenting) (deriding the interest in “diversity” as “a faddish slogan of the cognoscenti” and describing the concept as being “more a fashionable phrase than it is a useful term”), the University of Michigan and its aligned amici mounted a concerted effort to bring much-needed clarity. In a remarkable series of briefs, these groups assembled both social scientific evidence and observational reports from business, industry, and military leaders regarding the “substantial” educational and societal benefits that flow from an educational institution’s “enrollment of] a critical mass of minority students.” Grutter, 539 U.S. at 330, 123 S.Ct. 2325 (citation and quotation omitted).

Among the benefits attributed by the University and its amici to the enrollment of a minimal core of minority students, and embraced by the Court under the broad rubric of the diversity rationale, are the promotion of “cross-racial understanding,” the “breaking] down of racial stereotypes,” and the fact that “classroom discussion is livelier, more spirited, and simply more enlightening and interesting when the students have the greatest possible variety of backgrounds.” Id. at 330, 123 S.Ct. 2325 (citations and quotations omitted). Justice O’Connor’s opinion for the Court also explained that “student body diversity promotes better learning outcomes, and better prepares students for an increasingly diverse workforce and society,” and noted “that the skills needed in today’s increasingly global marketplace can only be developed through exposure to widely diverse people, cultures, ideas, and viewpoints.” Id.; see also Brief for Respondents at 11, Gratz v. Bollinger, 539 U.S. 244, 123 S.Ct. 2411, 156 L.Ed.2d 257 (2003) (No. 02-516) (“Racial and ethnic diversity is educationally important because, notwithstanding decades of progress, there remain significant differences in our lives and perceptions that are undeniably linked to the realities of race. Continuing patterns of residential segregation, for example, mean that the daily events and experiences that make up most Americans’ lives take place in strikingly homogenous settings. As a result, most students entering college have had few opportunities for meaningful interactions across lines of race and ethnicity. This separation ... provides little opportunity to disrupt racial stereotypes .... ”).

Finally, the majority emphasized testimony that, in the absence of race-conscious admissions, “underrepresented minority students would have comprised 4 percent of the [school’s] entering class in 2000, instead of the actual figure of 14.5 percent,” id. at 320, 123 S.Ct. 2325, and that a principal aim of the program was to prevent racial isolation. Id. at 318 & 319, 123 S.Ct. 2325; see also id. at 380-81, 123 S.Ct. 2325 (Rehnquist, C.J., dissenting) (noting that the university’s focus on achieving a “critical mass” of minority students was premised on “enrolling] enough minority students to provide meaningful integration of its classrooms and residence halls” and reducing the effects of “isolation] by racial barriers”) (quoting Brief for Respondents at 5).

2

Recognizing that each of the School District’s proffered interests in using its racial tiebreaker falls comfortably within the diversity rationale as that justification’s aims and benefits were articulated to (and embraced by) the Court, see supra at 960-61, Parents and their amici seek to cabin Grutter’s reach by contending that the Court’s compelling interest analysis was expressly limited to the use of race in admissions in the context of “the expansive freedoms of speech and thought associated with the university environment.” Grutter, 539 U.S. at 330. We of course acknowledge that Grutter addressed the use of racial classifications in higher education, and that language in the Court’s opinion reflects that factual underpinning.¹⁷ But we cannot identify a principled basis for concluding that the benefits the Court attributed to the existence of educational diversity in universities cannot similarly attach in high schools. We simply do not see how the government’s interest in providing for diverse interactions among 18 year-old high school seniors is substantially less compelling than ensuring such interactions among 18 year-old college freshmen. Cf. Grutter, 539 U.S. at 347, 123 S.Ct. 2325 (Scalia, J., dissenting) (“The ‘educational benefit’ that the University ... seeks to achieve by racial discrimination consists, according to the Court, of ‘cross-racial understanding,’ and ‘better preparation of students for an increasingly diverse workforce and society,’ all of which is necessary not only for work, but also for good ‘citizenship.’ This is not, of course, an ‘educational benefit’ [but] the same lesson taught to ... people three feet shorter and twenty years younger ... in institutions ranging from Boy Scout troops to public-school kindergartens.”) (quoting Grutter, 539 U.S. at 331, 123 S.Ct. 2325) (citations and alterations omitted).

3

At bottom, Grutter plainly accepts that constitutionally compelling internal educational and external societal benefits flow from the presence of racial and ethnic diversity in educational institutions. In support of its racial tiebreaker, the School District invokes precisely the interest sanctioned by the Supreme Court: securing those benefits. Those benefits are as compelling in the high school context as they are in higher education. We therefore conclude that the District has satisfied its first burden under strict scrutiny: It has articulated a compelling interest in pursuit of which it seeks to use a racial classification.¹⁸

C

Of course, to hold that the School District has invoked a compelling interest in pursuit of which it seeks to employ the racial tiebreaker is merely to begin our inquiry. Because “ ‘racial classifications are simply too pernicious to permit any but the most exact connection between justification and classification,’ ” Gratz, 539 U.S. at 270, 123 S.Ct. 2411 (quoting Fullilove, 448 U.S. at 537, 100 S.Ct. 2758 (Stevens, J., dissenting)), the School District also bears the burden of demonstrating that its use of the racial tiebreaker is narrowly-tailored to further that interest. See Hunter, 190 F.3d at 1063; Ho, 147 F.3d at 865. As with respect to compelling interest analysis, Grut-ter and Grata shed much-needed light on the once crepuscular contours of the narrow tailoring test applicable to the non-remedial use of racial preferences in educational admissions. Careful attention to these decisions — and the ways they addressed the divergent undergraduate and law school admissions schemes at the University of Michigan — is especially warranted.

1

As Grutter outlined, the admissions process at the University of Michigan Law School functions roughly as follows. Every completed application received by the Law School is both read and considered holistically by admissions officials. A significant focus of the decisionmakers is on an applicant’s academic ability, as measured by his or her undergraduate grade point average and score on the Law School Admissions Test (LSAT). Even so, these “hard” measures are insufficient to resolve admissions decisions: Just as “the highest score does not guarantee admission!, neither] does a low score disqualify an applicant.” Grutter, 539 U.S. at 315, 123 S.Ct. 2325. Instead, admissions officials must look beyond those measures to “soft” variables, including “the enthusiasm of the [applicant’s] recommenders, the quality of the undergraduate institution, the quality of the applicant’s essay, and the areas and difficulty of undergraduate course selection,” which in turn are used to help measure “an applicant’s likely contributions to the intellectual and social life of the institution.” Id. (internal quotations omitted).

This focus on “soft” variables aims to ensure that the Law School can “ ‘achieve that diversity which has the potential to enrich everyone’s education and thus make a law school class stronger than the sum of its parts.’ ” Id. (quoting the Law School’s written admissions policy). While recognizing that there are “ ‘many possible bases for diversity admissions,’ ” the admissions policy

“reaffirm[s] the Law School’s longstanding commitment to ‘one particular type of diversity,’ that is, ‘racial and ethnic diversity with special reference to the inclusion of students from groups which have been historically discriminated against ..., who without this commitment might not be represented in our student body in meaningful numbers.’ ”

Id. at 316, 123 S.Ct. 2325 (quoting policy). Of note, the policy seeks to ensure the enrollment of a “critical mass” of underrepresented minority students through the use of a calibrated racial preference — in the absence of which such students would comprise just 4 perc^it of the law school’s entering class, and thereby be less likely as a group fully to “ ‘make unique contributions to the character of the Law School.’ ” Id. (quoting policy). Crucially, “[t]he policy does not define diversity ‘solely in terms of racial and ethnic status,’ ” id. (quoting policy), but rather gives “serious consideration to all the ways an applicant might contribute to a diverse educational environment.” Id. at 337, 123 S.Ct. 2325.

Turning to the constitutionality of the program, Justice O’Connor immediately honed in on its core features: its flexibility and breadth. In concert with the baseline constitutional prohibition against quotas, id. at 334, 123 S.Ct. 2325 (“[A] race-conscious admissions program cannot use a quota system — it cannot ‘insulate each category of applicants with certain desired qualifications from competition with all other applicants.’ ”) (quoting Bakke, 438 U.S. at 315, 98 S.Ct. 2733 (Powell, J., concurring)), the law school “consider[s] race [and] ethnicity more flexibly as a ‘plus’ factor in the context of individualized consideration of each and every applicant.” Id. Moreover, she explained, the Law School’s policy strenuously avoids evaluating individual applicants “in a way that makes [their] race or ethnicity the defining feature of [their] application.” Id. at 337, 123 S.Ct. 2325. Conforming to the “paramount” constitutional requirement of truly “individualized consideration in the context of a race-conscious admissions program,” id. at 337, 123 S.Ct. 2325, the Law School’s admissions officers shun a “policy, either de jure or de facto, of automatic acceptance or rejection based on any ‘soft’ variable,” and awards “no mechanical, predetermined diversity ‘bonuses’ based on race or ethnicity.” Id.

Quite in contrast, the school refuses to “limit in any way the broad range of qualities and experiences that may be considered valuable conti«ibutions to student body diversity,” id. at 338, 123 S.Ct. 2325, and illustrates the strength of its commitment to that principle by highlighting a variety of non-racial criteria the institution considers valuable — for instance, that an applicant has lived or traveled abroad, speaks more than one language, has overcome personal adversity, or has a strong record of community service or even a prior career in a non-legal profession. Id. (discussing policy). And the Law School pays more than mere lip service to this commitment: It “actually gives substantial weight to diversity factors besides race,” allowing those factors to “make a real and dispositive difference for nonminority applicants as well.” Id. In short, the law school’s admissions program “considers race as one factor among many, in an effort to assemble a student body that is diverse in ways broader than race,” id. at 340, 123 S.Ct. 2325, and as a consequence, “does not unduly harm nonminority applicants.” Id. at 341, 123 S.Ct. 2325.

Finally, the Court observed that the Law School has “sufficiently considered workable race-neutral alternatives,” id. at 340, 123 S.Ct. 2325, and that periodic reviews (along with an eventual legal cutoff) ensure that the Law School’s use of race will be time-limited, in concert with the Constitution’s demand that any “ ‘deviation from the norm of equal treatment of all racial groups [be] a temporary matter, a measure taken in the service of the goal of equality itself.’ ” Id. at 342, 123 S.Ct. 2325 (quoting Croson, 488 U.S. at 510, 109 S.Ct. 706 (plurality opinion)). In light of the policy’s careful design and its adherence to the strict limits placed on the non-remedial use of race, the Court upheld the Law School’s program as narrowly tailored.

2

At issue in Grate, however, the undergraduate admissions program at the University’s College of Literature, Sciences, and the Arts (LSA) — though purportedly pursuing the same benefits from diversity as the Law School — had structured its admissions program around such a crude racial classification that the school had not even come close to satisfying the narrow tailoring requirement.¹⁹ Prior to its 1998 admissions cycle, LSA developed an admissions “selection index” that assigned each applicant a score of up to 150 points based on his or her grades, test scores, high school quality and curricular rigor, instate residency, legacy status, personal essay, and personal achievement or leadership. The index was then divided into strict dispositional bands: students scoring 100-150 points were admitted; students scoring 95-99 points were either admitted or had consideration of their application postponed; students scoring 90-94 points either had consideration of their application postponed or were admitted; students scoring 75-89 points were delayed or postponed; and students scoring fewer than 75 points were delayed or rejected. Gratz, 539 U.S. at 255, 123 S.Ct. 2411.

Although the LSA system superficially appeared to provide for individualized consideration of each applicanU-at least to the extent that each application was reviewed to ascertain the presence of various “soft” variables for mechanical scoring— such a rosy portrait was belied by the underlying reality of the policy. Tucked into a “ ‘miscellaneous’ category, an applicant was entitled to 20 points based upon his or her membership in an under-represented racial or ethnic minority group.” Id. Indeed, during 1999 and 2000, “every applicant from an underrepresented racial or ethnic group was awarded 20 points,” id. at 256, 123 S.Ct. 2411, at least one-fifth of the points necessary to secure admission to LSA.²⁰

It therefore is not surprising that the Court rejected LSA’s program on narrow tailoring grounds. Put simply, the program provided no serious individualized consideration of the applicant’s potential contributions to educational diversity.

The LSA’s policy automatically distributes 20 points to every single applicant from an ‘underrepresented minority’ group, as defined by the University. The only consideration that accompanies this distribution of points is a factual review of an application to determine whether an individual is a member of one of these minority groups. Moreover, unlike Justice Powell’s example, where the race of a ‘particular black applicant’ could be considered without being decisive, the LSA’s automatic distribution of 20 poiiits has the effect of making ‘the factor of race ... decisive’ for virtually every' minimally qualified underrepresented minority applicant.

Id. at 272, 123 S.Ct. 2411 (quoting Bakke, 438 U.S. at 317, 98 S.Ct. 2733 (Powell, J., concurring)); see also id. at 273, 98 S.Ct. 2733 (“[A]s the University has conceded, the effect of automatically awarding 20 points is that virtually every qualified underrepresented minority applicant is admitted.”). Having so easily concluded that the LSA policy was not narrowly tailored, the Court concluded by swiftly rejecting the University’s lame suggestion that “ ‘the volume of applications and the presentation of applicant information make it impractical’ ” to predicate admissions on individualized consideration. Id. at 275, 98 S.Ct. 2733 (quoting LSA’s Brief). “[T]he fact that the implementation of a program capable of providing individualized consideration might present administrative challenges does not render constitutional an otherwise problematic system.” Id. (citing Croson, 488 U.S. at 508, 109 S.Ct. 706).²¹

3

From the Court’s decisions in Grutter and Gratz—and drawing upon well-established narrow tailoring principles—we derive the following governing constraints. First, where an institution pursues non-remedial objectives, racial quotas are strictly prohibited. Gratz, 539 U.S. at 293, 123 S.Ct. 2411 (Souter & Ginsburg, JJ., dissenting) (“Justice Powell’s opinion in [Bakke ] rules out a racial quota or set-aside, in which race is the sole fact of eligibility for certain places in a class.”); Grutter, 539 U.S. at 334, 123 S.Ct. 2325.

Second, any consideration of race for non-remedial purposes must be flexible; an educational institution may not treat an applicant’s race or ethnicity as the touchstone of his or her individual identity, but instead must meaningfully evaluate each applicant’s potential diversity contributions in light of all pertinent factors. Gratz, 539 U.S. at 271-74, 123 S.Ct. 2411; id. at 279, 123 S.Ct. 2411 (O’Connor, J., concurring); Grutter, 539 U.S. at 337-39, 123 S.Ct. 2325; Bakke, 438 U.S. at 315 & 317-18, 98 S.Ct. 2733; Belk v. Charlotte-Mecklenburg Bd. of Educ., 269 F.3d 305, 345 (4th Cir.2001) (Trader, J., concurring); Wessmann, 160 F.3d at 798 & 800; Eisenberg v. Montgomery Cty. Pub. Schs., 197 F.3d 123, 132-33 (4th Cir.1999); Tuttle v. Arlington Cty. Sch. Bd., 195 F.3d 698, 707 (4th Cir.1999).

Third, it follows that an institution’s use of race must be neither mechanical nor conclusive. Gratz, 539 U.S. at 271-72, 123 S.Ct. 2411; id. at 278-79, 123 S.Ct. 2411 (O’Connor, J., concurring); Grutter, 539 U.S. at 336-37, 123 S.Ct. 2325. After all, automatically awarding a fixed racial preference to every single racially-preferred applicant signals an institutional disregard for the far broader array of diversity characteristics that produce the educational and social benefits deemed compelling by the Court. Bakke, 438 U.S. at 317, 98 S.Ct. 2733 (Powell, J., concurring). And any racial preference that necessarily results in the admission of an applicant demonstrates the pursuit of prohibited racial balancing simpliciter. See, e.g., Grutter, 539 U.S. at 337, 123 S.Ct. 2325 (“There is no policy, either de jure or de facto, of automatic acceptance.... ”); Wessmann, 160 F.3d at 799; cf. Eisenberg, 197 F.3d at 131; Tuttle, 195 F.3d at 707.

Fourth, narrow tailoring demands that the institution seeking to employ racial preferences at the very least demonstrate an earnest consideration of race-neutral alternatives. Grutter, 539 U.S. at 339-40, 123 S.Ct. 2325; Wygant, 476 U.S. at 280 n. 6, 106 S.Ct. 1842; Tuttle, 195 F.3d at 706; Podberesky, 38 F.3d at 160-61.

Fifth, serious efforts must be made to minimize the adverse impact of racial preferences on non-preferred group members; a programmatic use of race should be no more potent than necessary to achieve the compelling interest being pursued. Grutter, 539 U.S. at 341, 123 S.Ct. 2325; Wygant, 476 U.S. at 287, 106 S.Ct. 1842 (O’Connor, J., concurring in part and dissenting in part); Bakke, 438 U.S. at 308, 311, 314-15, 98 S.Ct. 2733 (Powell, J., concurring); Wessmann, 160 F.3d at 798.

Sixth, and finally, any program of racial preferences, regardless of its ultimate aspirations, must be time-limited. Grutter, 539 U.S. at 342, 123 S.Ct. 2325; Croson, 488 U.S. at 510, 109 S.Ct. 706 (plurality opinion by O’Connor, J.); Hayes v. N. State Law Enforcement Ass’n, 10 F.3d 207, 216 (4th Cir.1993).

4

The School District’s racial tiebreaker fails virtually every one of the narrow tailoring requirements.

a

First (and second), in contrast to the “flexible, nonmechanical,” evaluation of race employed by the University of Michigan Law School in the course of a “highly individualized, holistic review” which scrupulously “ensure[s] that each applicant is evaluated as an individual and not in a way that makes an applicant’s race or ethnicity the defining feature of his or her application,” Grutter, 539 U.S. at 337, 123 S.Ct. 2325, the School District’s racial tiebreaker is virtually indistinguishable from a pure racial quota. As Grutter defined that forbidden fruit of non-remedial racial preferences, “Quotas impose a fixed number or percentage which must be attained, or which cannot be exceeded, and insulate the individual from comparison with all other candidates for the available seats.” Grutter, 539 U.S. at 335, 123 S.Ct. 2325 (citations and quotations omitted). Yet this is almost precisely how the District itself has described the operation of its program— with a single variance: rather than impose a racial floor or ceiling, the School District’s racial tiebreaker establishes both a floor and a ceiling.

[I]f an oversubscribed school has fewer than 45 [percent] students of color/more than 55 [percent] whites, students of color will be assigned ahead of white students who live closer to the school. Conversely, if an oversubscribed school has fewer than 25 [percent] white students/more than 75 [percent] students of color, white students will be assigned ahead of students of color who live closer.

b

Indeed, to an even greater degree than the University of Michigan’s undergraduate admissions program, the School District — third—automatically and mechanically admits, using a computer algorithm designed to implement the ceilings and floors framing its racial tiebreaker, hundreds of white and non-white applicants solely because of their race. Cf. Gratz, 539 U.S. at 270, 123 S.Ct. 2411 (“[T]he University’s policy, which automatically distributes ... one-fifth of the points needed to guarantee admission ... to every single ‘underrepresented minority’ applicant solely because of race, is not narrowly tailored ....”) (emphasis added); id. at 271-72, 123 S.Ct. 2411 (“The only consideration that accompanies this distribution of points is a factual review of an application to determine whether an individual is a member of one of these minority groups.”). Thus, in stark contrast to the program sanctioned by Grutter, the racial tiebreaker not only fails to “serious[ly] eonsider[ ] all the ways an applicant might contribute to a diverse educational environment,” but is in fact a “de jure [policy] of automatic acceptance or rejection based on a[ ] single ‘soft’ variable.” Grutter, 539 U.S. at 337, 123 S.Ct. 2325. This the Constitution categorically forbids; such an impliably reflexive use of race “cannot be said to be narrowly tailored to any goal, except perhaps outright racial balancing.” Croson, 488 U.S. at 507, 109 S.Ct. 706 (plurality opinion).²²

c

Fourth, although numerous alternative admissions structures have been proposed to solve the School District’s oversubscription dilemma without so prominently featuring race in the equation, not all have been (or ever were) seriously considered by the Board.²³ Three such alternatives stand out.

i

First, the School Board has never seriously considered the use of a citywide high school admissions lottery. Though perhaps not palatable to the electorate — a consideration that cannot justify the use of race in its stead — a randomized lottery would necessarily produce levels of school diversity statistically comparable to (and perhaps even more proportional than) the District’s racial tiebreaker. Yet, when asked about using a lottery to meet the District’s diversity targets, Board member Barbara Schaad-Lamphere actually argued that a lottery would not result in racially proportional representation “because of probabilities, the law of probabilities.” Let us be clear: We are not forcing the School District to adopt a random assignment lottery. But given its evident commitment to achieving diversity, there is no question but that the Board should have earnestly appraised such a program’s costs and benefits. Whatever reasons there may be to reject a lottery, the demonstrably false pretext that “the law of probabilities” would render it ineffectual is not one.

The dissent takes us to task for suggesting that the School District must seriously consider using a lottery to achieve its diversity goals when Grutter itself explicitly rejected the plaintiffs’ claim that Michigan’s Law School should have considered a lottery. See post at 1009-1010. We find such criticism unavailing. Grutter rejected the plaintiffs’ demand that the Law School consider a lottery because such a program would necessarily diminish the quality of its admitted students and might not produce adequate educational diversity due to potential under-representation of various (not necessarily racial) kinds of diversity in its limited applicant pool.²⁴ Yet as the dissent itself notes, the School District’s adoption of a lottery is subject to neither of these potential pitfalls. Post at 1000 (noting that in this case “there is absolutely no competition or consideration of merit.... All high school students must and will be placed in a Seattle public school. The students’ relative merit is irrelevant.”) (emphasis in original). As a result, the District’s unconstrained applicant pool is not subject to a possible demographic skew, and there is absolutely no possibility that a lottery would diminish the quality of admitted students. Thus, neither of Grutter’s grounds for rejecting consideration of a lottery is present here.

ii

Second, the School District could have considered adopting a diversity-oriented policy that does not rely exclusively on race, but which instead accounts for the wider array of characteristics that comprise the kind of true diversity lauded by Justice Powell in Bakke and by the Court in Grutter and Gratz. In this respect, we observe that — for purposes of its internal school funding formula — the School District already collects a much wider array of data on students and families than merely their racial and ethnic identities. Among that data is information on whether a child lives at home or in “an agency”; if she lives at home, with whom; whether the child’s home and most proficient languages are English or some other language; and the child’s eligibility for free or reduced price lunch. Yet, the District considers none of those factors in admissions, and although individual Board members have occasionally suggested using one or more of those factors as an alternative tiebreaker, the Board has declined even to study how such an alternative would impact school diversity.²⁵

The dissent once again strays in its criticism of this suggestion. It first suggests that a programmatic focus on “true diversity” is no alternative at all “because it is not directed toward achieving the District’s interest in a racially integrated learning environment.” Post at 1008. But this claim is fundamentally mistaken. A programmatic focus on true diversity— within which race is one of many factors considered by the School District — subsumes the District’s interest in achieving racial diversity: It does not supplant it. Indeed, such a focus may not only help the District achieve the racial diversity it desires, see infra at 1001 n. 26, but might actually serve the District’s socialization interests to a far greater degree than its presently narrow focus on race alone. For, if the District’s fundamental interest is, as the dissent characterizes it, “to prepare children to be good citizens — to socialize children and to inculcate civic values,” see post at 992, and to achieve “a more democratic and inclusive experience for all citizens,” id. at 991-92 n. 9 (quoting one of the District’s expert witnesses), then accounting for factors other than race — like socioeconomic status — would bolster the District’s “democratic” mission by fostering, for instance, cross-class (and not just cross-racial) interaction. See Richard D. Kahlenberg, All Together Now: Creating Middle-Class Schools Through Public School Choice (2001) (documenting the civic and educational benefits of socioeconomically integrated schooling).

The dissent also errs in suggesting that “even though there has been no formal study of [this] proposal,” the deposition testimony of a few Board members is sufficient to demonstrate “legitimate reasons why the majority of the Board rejected the use of poverty measures.... ” Post at 1008. We disagree. The Board members’ blithe dismissal of a sincerely presented proposal simply cannot satisfy the constitutional requirement that the government earnestly appraise race-minimal alternatives prior to adopting race-conscious policies. Matters not formally evaluated cannot be “rejected” in a constitutionally-relevant sense: Such appraisal— whether with regard to the need for race-based action, or to the shape such action is to take — must be conducted “on the record.” See, e.g., Croson, 488 U.S. at 498-511, 109 S.Ct. 706 (chronicling Richmond’s failure adequately to document the basis for its use and design of a racial quota and stressing the constitutional demand that it do so); see also Shaw v. Hunt, 517 U.S. 899, 909-10, 116 S.Ct. 1894, 135 L.Ed.2d 207 (1996); Fullilove, 448 U.S. at 533-35, 100 S.Ct. 2758 (Stevens, J., dissenting); Rothe Dev. Corp. v. United States Dep’t of Def., 262 F.3d 1306, 1322-28 (Fed.Cir.2001); Associated Gen. Contractors of Ohio, Inc. v. Drabik, 214 F.3d 730, 735-37 (6th Cir.2000) (Boggs, J.); WH Scott Constr. Co. v. City of Jackson, 199 F.3d 206, 217-19 (5th Cir.1999) (King, C.J.); H.K. Porter Co. v. Metro. Dade County, 975 F.2d 762 (11th Cir.1992).²⁶

iii

Third (and at least for the moment, finally), acting in response to Parents’s filing of this lawsuit, the Seattle Urban League convened a working growp which included, among others, a representative from the NAACP, one of the Parents, a former member of the School Board, a retired high school principal, the then-current President of the Seattle Council Parent Teacher Student Association (PTSA), and a former PTSA President. In September 2000, they developed — and then formally proposed to the School Board — a comprehensive plan that would seek to enhance the quality of education in the City’s schools by focusing on educational organization, teacher quality, parent-teacher interaction, raising curricular standards, substantially broadening the availability of specialized and magnet programs (which could attract a broader cross-section of students to undersubscribed schools), and supporting extra-curricular development.²⁷

At the same time, their plan proposed decreasing the School District’s reliance on race in the admissions process by adding a new, primary tiebreaker based on pairing neighborhoods with particular schools and structuring the scope and size of the component residential areas such that no single region would contain enough students to fill its linked high school to capacity. Under the plan, preferences initially would be given to students choosing a school in their paired region, and the existing racial tiebreaker would be dropped from second to third in the process of resolving any remaining oversubscription (and any residual racial concentration not already solved by improving the attractiveness of previously racially concentrated schools). Finally, the Urban League working group proposed that the School District add an eleventh public high school, and that it strenuously market to the public the existing (and proposed additional) specialty programs throughout the City’s other high schools.²⁸

We cannot know whether its proposed reforms would have been successful in achieving the working group’s ambitious goals, but there is no doubt that the Urban League presented the Board with an especially thoughtful proposal for addressing the dilemmas plaguing contemporary urban education²⁹ while simultaneously striving to attain educational diversity without unduly relying on the use of crude racial preferences. Yet, this proposal was never formally discussed at a Board meeting. Indeed, some members of the Board even refused to read it. Consider the following remarkable deposition testimony by Board member Michael Preston:

Q: Are you familiar with the plan proposed by the ad hoc work group of the Urban League?

A: Somewhat.

Q: Did you ever consider that plan as a viable alternative to the current assignment plan?

A: No.

Q: Why not?

A: I thought they hadn’t done their homework. And yeah, they seemed too liberal and unbusinesslike. But it didn’t recognize the legitimate concerns that the people of Ballard and Magnolia have about the school.

Q: What in particular do you believe are the shortcomings of that Urban League plan that caused you not to consider it to be a viable alternative?

A: That it came from the Urban League. Even though [Urban League President and CEO] James Kelly is a good friend of mine, the Urban League has not been a bastion of enlightened thought, in my view, historically.

Q: Did you read the proposal?

A: No. I heard it characterized and summarized.

Q: By whom?

A: By the superintendent. I have a copy of it. I chose not to read it. I’d rather play with my bass lunker fishing game.

Q: Than consider the Urban League’s proposal?

A: Well.

Q: That might give some offense to the people who spent a good deal of time working that proposal up. .

A: Okay.

Q: We don’t need to show them a copy of the deposition transcript.

A: I’m sure it will eventually fall into their hands.

Q: Are you familiar with the broad outlines of how that proposal was structured?

A: Yeah.

Q: What’s your understanding of that? Not necessarily the minutia, but what’s your understanding of the general way that the Urban League’s proposal would have worked?

A: I don’t understand the relevance of the Urban League’s proposal, because it wasn’t considered, it wasn’t used. I don’t understand what difference it makes.

Q: Well, I’m just — in all honesty, one of the issues in the case, as I see it, is what alternatives were available to the school board.

A: Well, that wasn’t an alternative.

Q: Why is that?

A: Well, the Urban League is not the school board, it’s not the administration, it’s not the superintendent....

Without belaboring the point, this is not exactly the stuff from which narrow tailoring is made. While it may be the case that educational institutions need not exhaust every conceivable alternative to the use of racial classifications to satisfy strict scrutiny, narrow tailoring at least demands that schools earnestly consider using race-neutral and race-limited alternatives in order to provide for the kind of diversity that, properly constituted, can further compelling educational and social interests. Grutter, 539 U.S. at 339, 123 S.Ct. 2325. Given the tragic history of race in our country, the Constitution demands no less — our education policymakers’ enthusiasm for handheld electronically simulated “bass lunker fishing game[s]” notwithstanding.

d

Fifth, the School District’s racial tiebreaker is not designed to minimize its adverse impact on third parties; the extent to which it uses race is not calibrated to the benefits sought. Over time, “the band,” see supra n. 6, has ranged from as much as +/25 percent to as little as +/10 percent, and it currently sits at +/15 percent. And while such variances conceivably could be interpreted to suggest that the School District is carefully trying to optimize its realization of diversity’s benefits, the record belies such an interpretation. In October 2000, School Superintendent Joseph Olchefske formally recommended that the band be expanded from +/10 percent to +/20 percent because “[ajfter review and 2 years experience, there was not strong evidence that utilizing a +/10 [percent] band provided a materially better educational experience than would a band of +/20 [percent]. Accordingly, in order to fulfill our narrow tailoring obligation, staff is recommending a +/20 [percent] band.”³⁰ Yet, even after its chief educator and his staff twice had reported that twice as much of an adjustment would have no adverse impact on the diversity payoff, the Board adjusted the band by just 5 percentage points. This is not the measure of tailored proportionality. Instead, it represents a stubborn adherence to the use of race for race’s sake, with the effect that some non-preferred student applicants will be displaced solely because of their racial and ethnic identities — to no benefit at all.³¹

Taken alone, any of these shortcomings would doom the School District’s program. Together, they reveal an unadulterated pursuit of racial proportionality that cannot possibly be squared with the demands of the Equal Protection Clause.³²

5

In an effort to evade the clear import of the narrow tailoring analysis applied by Grutter, Gratz, and their progenitors, the School District argues that such analysis simply does not apply in the K-12 context: “[T]he Michigan decisions have meaning only in the context of selective admissions and other ‘zero sum’ programs.... [The] argument that race may be considered [only] in a holistic individualized review as one factor among many contributing to diversity is not applicable to non-selective school assignments.”³³ Indeed, they argue, the use of racial classifications in their allegedly “non-zero sum” open choice high school admissions process “is narrowly tailored, as a matter of law.”

For support, they unearth language from Associated General Contractors of California v. San Francisco Unified School District, a 1980 case in which we addressed a challenge to a School Board policy requiring that “bidders for construction contracts ... must be minority general contractors or must utilize minority subcontractors for 25[percent] in dollar volume of the contract work.” 616 F.2d 1381, 1383 (9th Cir.1980). At the heart of their challenge, the contractors asserted that the quota violated California law requiring that all school contracts (except those involving de minimis expenditures) be awarded solely to “the lowest responsible bidder.” Id. at 1385 (discussing and quoting Cal. Educ.Code § 39640). In defense, the School Board asserted not only that its policy was permissible, but that the policy was required in order to remedy past discrimination — and therefore that insofar as it would prevent maintenance of such a policy, California’s low-bid law violated the Supremacy Clause, U.S. Const, art. VI, § 2. Id. at 1386.

We flatly rejected that argument, drawing a distinction between “reshuffle programs” and “stacked deck programs,” id. at 1386, and placing the Board’s quota in the latter category of race-based policies. We explained the central difference between these two categories as follows: While reshuffle programs “neither give[] to nor withhold[ ] from anyone any benefits because of that person’s group status, but rather ensure[ ] that everyone in every group enjoys the same rights in the same place,” so-called stacked deck programs “specifically favor[ ] members of minorities in the competition with members of the majority for benefits that the state can give to some citizens but not to all.” Id. In passing, we twice suggested that programs designed to desegregate schools are reshuffle programs, see id. at 1386 & 1387, and elsewhere stated “that ‘stacked deck’ programs trench on Fourteenth Amendment values in ways that ‘reshuffle’ programs do not.” Id. at 1387.

Ultimately, we held simply that although “the state has an affirmative constitutional duty to use ‘reshuffle’ programs to cure the effects of past or present de jure segregation ..., there is no constitutional duty to engage in ‘stacked deck’ affirmative action.” Id. Of particular import, we never reached the Fourteenth Amendment question squarely implicated by this case: whether notwithstanding the absence of an obligation to adopt such a policy, its implementation of such a program was permissible. Instead, given our resolution of an underlying state law question, our assessment of that law’s constitutionality, and the fact that the Court recently had granted certiorari in Fullilove, we explicitly declined to “test the policy itself against the standard of the United States Constitution.” Id. at 1391.

Although it might seem obvious that a case cannot control the resolution of an issue it expressly declined to confront, the School District nonetheless grasps at Associated General’s furtive references to school desegregation.³⁴ It argues that its racial classification is equivalent to the programs arguably favored by Associated General, and thus (reading that opinion at the highest level of generality) not viola-tive of the Fourteenth Amendment.³⁵ We disagree.

Let us begin with two interrelated observations. First, Associated General addressed neither school desegregation nor the use of race in educational admissions; it assessed a mechanical set-aside governing distribution of construction projects. Second, Associated General did not even purport to apply strict scrutiny to a particular racial classification (or, more specifically, an actual school desegregation program). Its Fourteenth Amendment analysis assessed only whether the use of certain types of classifications is required to remediate prior official race discrimination — not whether constitutional limits circumscribe the use of such classifications for non-remedial purposes, what those limits might be, and how they might affect the constitutionality of a particular policy.

Two conclusions follow. First, Associated General sheds no light on how we are to apply the narrow tailoring analysis rendered applicable to any racial classification by Croson, 488 U.S. at 493-95, 109 S.Ct. 706, Adarand, 515 U.S. at 224, 115 S.Ct. 2097, Grutter, 539 U.S. at 326, 123 S.Ct. 2325, and Gratz, 539 U.S. at 270, 123 S.Ct. 2411,³⁶ to the tiebreaker at issue in this case. And second, Associated General’s assertions about the nature of school desegregation programs are — however one defines the term — obiter dicta.³⁷

Insofar as the School District argues that because reshuffle programs may be less invidious than stacked deck programs and thus, as a constitutional matter, somehow not subject to the stringencies of the narrow tailoring requirement — a proposition never embraced by this court³⁸ and certainly not one reconcilable with binding Supreme Court holdings³⁹ — it misapprehends the concept of narrow tailoring. The contours of narrow tailoring do not turn on the importance of the interest supporting the government’s use of a racial classification (though such analysis, whatever its shape, surely is triggered by use of a racial classification in pursuit of a compelling interest), nor on the asserted degree of the intrusion that a particular use of race might render. The personal right to equal treatment is implicated any time the government employs race for any reason. See Coalition, 122 F.3d at 702 (quoting Adarand, 515 U.S. at 230, 115 S.Ct. 2097).

Instead, narrow tailoring analysis focuses only upon the fit between the ends in pursuit of which a racial classification is used, the particular way in which the policy at issue uses that racial classification, and the baseline legal limits placed upon the use of racial classifications — and it does so because any policy using race, for any reason and in any way, is “inherently suspect,” “by [its] very nature odious to a free people,” and simply “too pernicious to permit any but the most exact connection between justification and classification.” Adarand, 515 U.S. at 223, 115 S.Ct. 2097 (quoting Fullilove, 448 U.S. at 523, 100 S.Ct. 2758 (Stewart, J., dissenting)); id. at 224, 115 S.Ct. 2097 (quoting Hirabayashi, 320 U.S. at 100, 63 S.Ct. 1375); id. 229, 115 S.Ct. 2097 (quoting Fullilove, 448 U.S. at 537, 100 S.Ct. 2758 (Stevens, J., dissenting)). It is for those reasons that, as Judge Selya — in typically perspicuous fashion — has observed, a court applying strict scrutiny must focus on “whether the concrete workings of the Policy merit constitutional sanction. Only by such particularized attention can we ascertain whether the Policy bears any necessary relation to the noble ends it espouses. In short, the devil is in the details.” Wessmann, 160 F.3d at 798. Under strict scrutiny, no racial classification, no matter its context and no matter the nature or strength of the interest it serves, is exempt from the strictures of narrow tailoring, and this program plainly fails to satisfy them.

Finally, even if were we to accept that Associated General bears on this case, we observe that the School District’s racial tiebreaker operates as a “stacked deck” program. Quite simply, it “specifically favors members of [some races] in the competition with members of [other races] for benefits that the state can give to some citizens but not to all,” Associated General, 616 F.2d at 1386—here, of course, admission to particular high schools.⁴⁰ As noted earlier, operation of the tiebreaker at Franklin causes whites to be admitted preferentially because they are white, and its operation at Ballard causes non-whites to be admitted preferentially simply because they are not white. To the argument that this program is “non-selective,” we can only express bewilderment: The racial tiebreaker is used to determine student admissions solely to oversubscribed' — ■ and thus necessarily selective■ — schools.⁴¹ Whatever Associated General says about school desegregation programs, this is not the programmatic design that we had in mind.

Having accepted the School District’s invocation of Grutter and Gratz in support of the proposition that racial and ethnic diversity can generate constitutionally compelling benefits within the educational setting itself and our society at large, we ultimately are compelled to reject the School District’s strained efforts both to eat its cake and have it too. Its racial tiebreaker — though enlisted in the service of admittedly worthy ends — plainly fails the narrow tailoring component of the Constitution’s strict scrutiny test.⁴²

IV

Approaching this case from a fundamentally divergent perspective, the dissent offers a spirited and thoughtful defense of the School District’s racial tiebreaker. Although we have responded to a few of its more specific criticisms, we see little to gain from the kind of note-by-note combat so pervasive in modern judicial opinions. Rather, given the thoroughness with which we have articulated our respective analy-ses of the racial tiebreaker, we think there is more to gain by elucidating the core foundations of our dispute. We perceive three significant points of departure, and address each in turn.

A

The dissent repeatedly suggests that we should simply defer to the School District’s decision to employ its tiebreaker in pursuing racial proportionality. Post at 993-94 n. 13, 995-96, 996-98 & 1007-1008. Indeed, the dissent gradually shifts from noting how “compelling” the District’s policy is to focusing on its “reasonablefness],” “legitima[cy],” and “permissibility].” Id. at 993-94 n. 13, 1005, 1008, 1008, 1009, 1013. We believe such unfettered deference is inconsistent with our obligations under strict scrutiny — and, contrary to the dissent, that it is especially inconsistent with constitutional demands in this context. Because Grutter’s decision to accord universities a measure of deference occasioned a sharp debate over this issue, careful attention to its rationale for doing so is in order.

1

After identifying the Law School’s “compelling interest in attaining a diverse student body,” Grutter, 539 U.S. at 328, 123 S.Ct. 2325, the Court promptly stated:

The Law School’s educational judgment that such diversity is essential to its educational mission is one to which we defer.... Our scrutiny of the interest asserted by the Law School is no less strict for taking into account complex educational judgments in an area that lies primarily within the expertise of the university. Our holding today is in keeping with our tradition of giving a degree of deference to a university’s academic decisions ....

Id. at 328-30, 123 S.Ct. 2325 (citations omitted) (emphases added).

In developing its rejection of the dissenting Justices’ charge that such deference was indeed at odds with strict scrutiny, see id. at 379-86, 123 S.Ct. 2325 (Rehnquist, J., dissenting); id. at 387-94, 123 S.Ct. 2325 (Kennedy, J., dissenting); id. at 362-67, 123 S.Ct. 2325 (Thomas, J., dissenting), the Court began by reiterating its “long recognition] that, given the important purpose of public education and the expansive freedoms of speech and thought associated with the university environment, universities occupy a special niche in our constitutional tradition.” Id. at 330, 123 S.Ct. 2325. It then expressly sanctioned Justice Powell’s statement in Bakke that, given these unique First Amendment interests, universities have a “right to select those students who will contribute the most to the ‘robust exchange of ideas.’ ” Id. (quoting Bakke, 438 U.S. at 313, 98 S.Ct. 2733 (Powell, J., concurring)).

Deference thus was due to the University not because its interest was “simply to assure within its student body some specified percentage of a particular group merely because of its race or ethnic origin ... but [because it] was defined by reference to the educational benefits that [such] diversity is designed to produce,” id. (quotations and citations omitted) — and of crucial importance, because those “educational benefits” were not merely socially compelling, id. at 330-33, 123 S.Ct. 2325, but fundamentally internal to the university’s “academic” mission. Id. at 328, 123 S.Ct. 2325. The Court emphasized: “The benefits are important and laudable, because classroom discussion is livelier, more spirited, and simply more enlightening and interesting when the students have the greatest possible variety of backgrounds.” Id. at 330, 123 S.Ct. 2325 (quotations and citations omitted).

2

Beyond the fact that Grutter deferred only to the Law School’s identification of its interests as compelling — and, contrary to the dissent here, not to its tailoring analysis — neither of Grutter’s grounds for affording deference is present here. First, secondary schools do not occupy the same “special niche in our constitutional tradition” as higher education, and the Court has never held they possess a similar First Amendment right of academic freedom. Indeed, while the Court has been willing to afford secondary schools some limited leeway to enable them to meet their most basic need (preserving the orderly school environment necessary to enable academic learning⁴³), the Court has never suggested that public secondary schools have a constitutional right to select their student body, much less that such a right entails selecting students based solely on their race — a power that not even universities enjoy, despite their uniquely privileged status. Quite in contrast, the Court has repeatedly condemned racial balancing, held that a State’s creation of a system of compulsory public education endows students (not schools) with a constitutionally-protected interest, and has pointedly reminded school authorities that “ ‘[t]he Fourteenth Amendment ... protects the citizen against the State itself and all of its creatures — Boards of Education not excepted.’ ” Goss, 419 U.S. at 574, 95 S.Ct. 729 (quoting W. Va. Bd. of Educ. v. Barnette, 319 U.S. 624, 637, 63 S.Ct. 1178, 87 L.Ed. 1628 (1943)); see also Grutter, 539 U.S. at 330 & 336, 123 S.Ct. 2325.

Second, the principal benefits the School District seeks through its pursuit of racial proportionality are neither internal to the school environment nor within the special expertise of school administrators. For although the District’s asserted interests in the educational and societal benefits derivable from diverse schools mirror those embraced by Grutter, see supra at 960-65, the dissent quite properly notes that the School District has an appreciably different focus. Rather than primarily seeking the internal academic benefits of diversity, the District’s chief focus is on the broader social benefits diversity can stimulate. See post at 991-92.

We do not believe this divergent emphasis changes the nature of the District’s interests; they remain compelling, and thus can justify an appropriately limited use of race. See supra Sections III.B.3 & III.C.3. But it does affect the degree to which the District can claim deference. With regard to its secondary pursuit of diversity’s internal academic benefits, we believe that while limited deference to educational institutions arguably could be due when they pursue core goals, such deference is entirely unwarranted when they court tangential ones. Having proceeded from the premise that the internal academic benefits of diversity are secondary to the District’s socialization mission, it is at best curious that the dissent nonetheless maintains we should defer to the District’s particularly crude use of a most disfavored classification in their pursuit.⁴⁴

Far more important, we see a crucial difference between a school’s pursuit of the internal academic benefits of diversity and its pursuit of diversity’s external social benefits. For although the former manifest within the District’s schoolhouses, and thus are susceptible to ready appraisal exclusively by education policymakers, the “democratic” benefits attributable to classroom diversity are diffuse, manifest long after students leave the classroom, do so in contexts not subject to the exclusive oversight of teachers, and cannot be measured with skills possessed uniquely by educators. That is to say: While it is clear that educators are uniquely positioned to gauge how classroom discussions respond to shifts in classroom racial composition, they are not similarly well-positioned to assess how marginal changes in schoolhouse racial demographies affect how students interact with each other years after they leave school for the “real world.” And to the limited extent they may have such insight, the record does not suggest the District ever sought to appraise the long-term social effects of engineering proportional demographic changes in connection with its design of the tiebreaker.

B

Beyond the dissent’s unfettered deference to the School District — a constitutional departure that we think undermines its supposed “strict scrutiny” of the tiebreaker — one of its most striking features is its lack of focus on the data. For if it is true that a picture is worth a thousand words, there are times when statistics give Proust a run for his money.⁴⁵

1

The dissent simply seems to assume that absent the racial tiebreaker, Seattle’s public high schools would be “racially concentrated, or racially isolated,” and thus unable to attain diversity’s benefits. Post at 993; id. at 1001-1002, 1009-1010, 1011 & n. 39. The record belies its assumption. Uncontroverted data produced by the School District during this litigation and contained in the record indicate that, in the absence of the racial tiebreaker, 2000-2001 school year enrollments at Seattle’s public schools would have been as follows:

[[Image here]]

Far from revealing “racially concentrated” or “racially isolated” student bodies, these figures demonstrate that each of the District’s schools would enroll a vibrant array of students without considering race at all.⁴⁶ (Only by indulging the fallacious proposition that inter-ethnic diversity is irrelevant to the District’s putative democratic mission could one conclude otherwise.)⁴⁷ It is thus particularly hard to credit the dissent’s assumption that the tiebreaker is necessary to the District’s fulfillment of its diversity oriented goals.⁴⁸ & ⁴⁹

Such an assumption becomes even less tenable in light of the tiebreaker’s limited effect where it does operate. Indeed, the data demonstrate that the tiebreaker produces only the most trivial annual changes in school demography.

[[Image here]]

Given enrollment totals at these schools, the tiebreaker’s annual effect is thus merely to shuffle a few handfuls of different minority students between a few schools— about a dozen additional Latinos into Ballard, a dozen black students into Nathan Hale, perhaps two dozen Asians into Roosevelt, and so on. The District has not met its burden of proving these marginal changes substantially further its interests,⁵⁰ much less that they outweigh the cost of subjecting hundreds of students to disparate treatment based solely upon the color of their skin.⁵¹

2

Beyond casting serious doubt on the School District’s need to use its racial tiebreaker, the numbers undermine the dissent’s assessment of its efficacy — and yet again, on the dissent’s own terms. For present purposes, we take at face value the dissent’s identification of the District’s primary interest as ensuring “that children learn to interact with peers of different races,” post at 1001; see also post at 992-93, and accept arguendo its claim that “when a racially integrated school system is the goal ..., there is no more effective means than a consideration' of race to achieve a solution.” Post at 1001, 1010. We credit for the moment its assessment that it is appropriate to “link[] the integration tiebreaker to the racial demographic of the District’s population [because] the District is trying to teach its students to be effective participants in the racially diverse environment in which they exist,” post at 1010-1011 (emphasis added), and we ignore both that the tiebreaker does not operate where it would be most needed to meet the dissent’s articulation of the District’s goals and that its demographic impact is trivial where it does operate. Yet even then, it is clear to us that the School District’s racial tiebreaker falls well short.

Because the tiebreaker relies on a crude white/non-white metric of racial identity, representation of particular minorities varies widely within the schools where it operates (indeed, throughout the school system). Once again, consider data produced by the School District and contained in the record:

[[Image here]]

Given that the demographic impact of the tiebreaker during the 2000-2001 school year was merely to shift overall white/ non-white student ratios by at most 6.1 percent at the schools where it operated, supra at 984, Table 2, these are truly enormous variations. The range in representation of Asians at the tiebreaker schools (19.3 percent, as Asian representation ranged from just 17.5 percent of students at Ballard to a whopping 36.8 percent of students at Franklin) was more than 3 times the size of the tiebreaker’s maximum annual impact on overall white/ non-white ratios; and the range in black representation (24.6 percent, ranging from 32.3 percent of students at Franklin to just 7.7 percent of students at Roosevelt) was more than 4 times the maximum annual white/non-white impact.⁵² & ⁵³

These representational variances cut to the core of the dissent’s defense of the tiebreaker’s design. For if the tiebreaker’s goal is (as the dissent characterizes it) to assure every student the opportunity to interact with the right ratio of different-looking peers in order to prepare them for life in our diverse society, the program plainly fails huge numbers of students. Consider: Roughly 600 black students at Garfield during the 2000-2001 school year were deprived of the chance to interact with a sufficient number of Asian students (12.5 percent of Garfield students versus 27.5 percent of students in the school population at large); more than 100 Latino students at Roosevelt were denied a chance to interact with an adequate number of African-American peers (7.1 percent of Roosevelt students versus 22.8 percent of the student population at large); 600 Asian students at Franklin were unable adequately to interact with Native American students (0.8 percent of Franklin students versus 2.5 percent of the student population at large). And thousands of other students left their schools not having been inadequately exposed to interaction with certain racial minorities, but rather having been overexposed to them — and thus likewise not having been readied “to be effective participants in the racially diverse environment' in which they exist.” Cf. post at 1010-1011. Even taking at face value the dissent’s embrace of racial balancing, these data indicate the tiebreaker does not even rationally further the impermissible ends the dissent trumpets.

C

The final major point of departure between our opinion and the dissent is perhaps the most significant: We believe the dissent errs in failing to recognize the injury rendered by the tiebreaker. We have already explained that the individual right to equal treatment is implicated any time government uses race to apportion benefits or burdens. Here, in determining where some 300 students will attend high school, the School District crudely approximates the shades of their skin and assigns them to schools accordingly. Nonetheless, the dissent surprisingly suggests that not one of those students suffers a legally cognizable injury (whether such an injury can be justified or not) because the constitutional prohibitions against determinative racial preferences and quotas do not apply to secondary education, and because each of those students will get a basic education anyway. We address these claims in turn.

1

The dissent suggests that the constitutional prohibitions against determinative racial preferences and quotas do not apply within the secondary educational context because secondary schools do not employ race as a proxy for merit, and thus pose little risk of stigmatizing or stereotyping those they putatively benefit. Post at 999-1002. Its analysis cannot be squared with precedent.

The Court has long prohibited the use of outright quotas in contexts where merit and qualification are — as the dissent asserts them to be here — completely irrelevant. Thus, Croson rejected Richmond’s racial quota for construction contracts on grounds that the “quota cannot be said to be narrowly tailored to any goal, except perhaps outright racial balancing.” Croson, 488 U.S. at 507, 109 S.Ct. 706. Of course, when dealing with contracting, the only “qualification” or indicum of “merit” is submission of a low bid. Croson’s constitutional objection to the City’s use of a quota was thus not that it threatened to stigmatize minorities: It was that quota-driven racial balancing is flatly at odds with the Fourteenth Amendment’s “ ‘ultimate goal’ of ‘eliminat[ing] entirely from governmental decisionmaking such irrelevant factors as a human being’s race.” Id. at 500, 109 S.Ct. 706 (quoting Wygant, 476 U.S. at 320, 106 S.Ct. 1842 (Stevens, J., dissenting)) (alteration in original).

Likewise, courts have rejected the use of race-based quotas in legislative redistricting. As the Fifth Circuit has put the point, “[Djistricters are not bound — or allowed — to sacrifice traditional districting concerns to meeting quotas of diversity, just as they are not allowed to do so in order to meet quotas of racial concentration.” Chen v. City of Houston, 206 F.3d 502, 511 (5th Cir.2000); see also Easley v. Cromartie, 532 U.S. 234, 241 & 257, 121 S.Ct. 1452, 149 L.Ed.2d 430 (2001) (recognizing that race may not be the predominant factor in legislative districting, and noting that an email referencing “racial balance” constituted evidence of such an “improper[ ]” use of race). But surely redistricting involves no “competition” or “consideration of merit,” and of course “no stigma results from any particular [dis-tricting] assignment.” Cf post at 1000. At bottom, the problem with the imposition of a racial quota and the use of inflexibly mechanical racial preferences is not simply that they impermissibly conflate skin color with merit or qualification; it is that, more than any euphemistic “thumb on the scale,” they breed deep-seated cross-racial resentment and do violence to the constitutional principle that “we are just one race here. It is American.” Adarand, 515 U.S. at 239, 115 S.Ct. 2097 (Scalia, J., concurring).⁵⁴

2

Even more baffling is the dissent’s claim that because the District ultimately assigns each student to a “quality” high school offering a “baseline ... education,” no applicant rejected on the basis of his or her race “suffers a constitutionally significant burden” from the tiebreaker’s operation. Post at 1000-1001,1012-1013.

We certainly agree with the dissent’s observation that no student has a right to attend the school of his or her choice or to attend a school offering anything more than the standard education mandated by state law. Post at 1012-1013. What the dissent seems to overlook is that individuals likewise have no right to welfare benefits, Shapiro v. Thompson, 394 U.S. 618, 627 n. 6, 89 S.Ct. 1322, 22 L.Ed.2d 600 (1969), unemployment compensation, Sherbert v. Verner, 374 U.S. 398, 404, 83 S.Ct. 1790, 10 L.Ed.2d 965 (1963), overruled on other grounds by Employment Div. v. Smith, 494 U.S. 872, 110 S.Ct. 1595, 108 L.Ed.2d 876 (1990), a tax exemption, Sherbert, 374 U.S. at 406, 83 S.Ct. 1790 (discussing Speiser v. Randall, 357 U.S. 513, 78 S.Ct. 1332, 2 L.Ed.2d 1460 (1958)), or a public job, Perry v. Sindermann, 408 U.S. 593, 597-98, 92 S.Ct. 2694, 33 L.Ed.2d 570 (1972). Yet no one would seriously maintain that, as a result, states or localities could condition the distribution and extent of such benefits on the basis of race.⁵⁵

The dissent’s appeal to the rights/privileges distinction in this context is particularly ironic because it readily parallels arguments long ago repudiated in this context. Indeed, it quite unintentionally evokes the State of Missouri’s argument that because it had no duty to supply legal education, and there was therefore no personal “right” to a legal education, no one could suffer a legally cognizable injury from the University of Missouri Law School’s use of a racial classification in admissions. Yet as the Court explained in rejecting that claim, “The question here is not of a duty of the State to supply legal training, or of the quality of the training which it does supply, but of its duty when it provides such training to furnish it to the residents of the State upon the basis of an equality of right.” Missouri ex rel. Gaines v. Canada, 305 U.S. 337, 349-50, 59 S.Ct. 232, 83 L.Ed. 208 (1938).

Likewise, we think the issue here is not whether students have a right to attend the school of their choice, or one of significantly above-average quality. It is whether having made available a choice-based system of public education the District may use race to circumscribe parental choices in the way it has. The dissent’s suggestion that it may do so because all are “equally subject” to such discrimination, post at 1012 (emphasis in original), cannot be correct: Across-the-board wrongs do not, as the dissent reasons, make a right.

D

Unfortunately, the dissent’s thin scrutiny of the District’s racial tiebreaker strays far from constitutional norms — and it inadvertently threatens to entrench a permanent regime of racial discrimination. For if public education’s most compelling mission really is to prepare children to interact with those who look different by balancing each school’s racial profile, then the Constitution’s promise of equal justice will remain unfulfilled for the inestimable number of future generations during which race inevitably will be perceptible.

V

Because the School District’s use of the racial tiebreaker violates the equal protection mandate of the Fourteenth Amendment, we REVERSE the decision of the District Court and REMAND with instructions to ENJOIN the School District from using the racial tiebreaker, as most recently constituted, in making future high school assignments.

1 . For instance, year 2000 data indicates that the average combined score on the Scholastic Achievement Tests (SATs) at Garfield was 1208 — some 154 points above the state average — while at Cleveland it was 838, some 216 points below the state average. Similarly, while just one-quarter of students at Roosevelt scored below the 25th percentile on the Iowa Tests of Educational Development, more than three-quarters of Rainier Beach students scored below the 25th percentile.

2 . For instance, Ballard High School offers a unique "Biotech Academy.” Ballard describes its Biotech program as "[a] specialized learning program that brings together science, mathematics and language arts to prepare students for advanced study and a career in science.” Ballard Biotech Academy Website <http://ballard.seattleschools.org/aca-demics/academies/biotech.html> (visited Mar. 24, 2004). The program has its own separate admissions procedure with required prerequisite classes. Admission to the program does not, however, guarantee admission to Ballard — which is governed by the School District’s open enrollment plan.

3 . For graphic representations of the racial and ethnic dispersion within Seattle's population, interested readers may wish to consult the various thematic maps derived from year 2000 U.S. Census data and made available by the City of Seattle at: <http://www.cityofseat-tle.net/DCLU/demographics/data_census.asp> (visited March 29, 2004). They indicate that census tracts north of Seattle’s downtown and those along the City's waterfronts tend to be predominantly white, while those south of downtown (and particularly in the City's southeast quadrant) tend to reflect more substantial non-white populations.

4 .Seattle's student population is approximately 40 percent white and 60 percent non-white. Splitting Seattle along a north-south axis, data introduced by the School District indicates that 74.2 percent of the District's Asian students, 83.6 percent of its black students, 65.0 percent of its Hispanic students, and 51.1 percent of its Native American students live in the southern half of the city. By contrast, 66.8 percent of the District’s white student population lives in the northern half of the city. Overall, approximately 77.2 percent of students in the southern half of the city, and just 35.7 percent of students in the northern half of the city, are non-white.

5 . Oversubscription was apparently not tied to geographic location. The oversubscribed schools included three high schools north of downtown (Ballard, Nathan Hale, and Roosevelt) and two high schools south of downtown (Garfield and Franklin). The un-dersubscribed schools included one north of downtown (Ingraham), three south of downtown (Chief Sealth, Cleveland, and Rainier Beach), and one west of downtown (West Seattle).

6 .During the 2000-01 school year, the acceptable deviation (called "the band”) was fixed at 4-/10 percent. Thus, if an oversubscribed school had fewer than 31 percent or more than 51 percent white students, the tiebreaker would operate. Between the 2000-01 and 2001-02 school years — while this litigation was pending — the school board expanded the band to 4-/15 percent, meaning that the tiebreaker would operate only if an oversubscribed school had fewer than 26 percent or more than 56 percent white students. During the 2000-01 school year, four of the city’s five oversubscribed schools were considered integration positive and therefore employed the tiebreaker: Ballard, Franklin, Nathan Hale, and Roosevelt. With the expansion of the band to 4-/15 percent prior to the 2001-02 school year, the tiebreaker ceased to operate at Roosevelt.

Two further changes were made to the program prior to the 2001-02 school year. First, a so-called "thermostat” was added to the plan: The School District would cease to use the racial tiebreaker for the year at any school once its use had brought the school into racial balance. Second, the integration tiebreaker would be used only in determining the makeup of entering ninth grade classes, but would not be applied to assignments involving the limited number of students seeking to transfer high schools before the tenth, eleventh, or twelfth grades.

7 . Of course, this also means that at Franklin, non-whites are denied admission because they are not white; and at Ballard, whites are denied admission because they are white.

8 . Wash. Rev.Code § 49.60.400 ("The state shall not discriminate against, or grant preferential treatment to, any individual or group on the basis of race, sex, color, ethnicity, or national origin in the operation of public employment, public education, or public contracting.”).

9 . U.S. Const, amend. XIV, § 1 ("No state shall ... deny to any person within its jurisdiction the equal protection of the laws.”).

10 . 42 U.S.C. § 2000d ("No person in the United States shall, on the ground of race, color, or national origin, be excluded from participation in, be denied the benefits of, or be subjected to discrimination under any program or activity receiving Federal financial assistance.”). Title VI has long been held to be essentially co-extensive with the guarantees of the Equal Protection Clause of the Fourteenth Amendment and the equal protection component of the Fifth Amendment. See Regents of Univ. of Cal. v. Bakke, 438 U.S. 265, 287, 98 S.Ct. 2733, 57 L.Ed.2d 750 (1978) (Powell, J., concurring) ("Title VI must be held to proscribe only those racial classifications that would violate the Equal Protection Clause or the Fifth Amendment.”); id. at 329, 98 S.Ct. 2733 ("Title VI prohibits only those uses of racial criteria that would violate the Fourteenth Amendment if employed by a State or its agencies....”) (Brennan, White, Marshall, and Blackmun, JJ., concurring in part); see also Alexander v. Sandoval, 532 U.S. 275, 280-81, 121 S.Ct. 1511, 149 L.Ed.2d 517 (2001).

11 . Specially concurring in the court’s decision, Judge Graber likewise divined that "the racial tiebreaker that Seattle School District No. 1 uses to assign some public high school students to desirable schools plainly 'grants preferential treatment’ to those students on the basis of their race, in violation of Initiative 200.” Parents II, 285 F.3d at 1253 (Graber, J., specially concurring).

12 . See Deborah Bach, New School Board Must Work Together, Observers Say, Seattle Post-Intelligencer, Nov. 6, 2003, at B1; David Postman, Incumbents Hammered: 3 on Seattle School Board Out; 3 on Council Headed There, Seattle Times, Nov. 5, 2003, at A1.

13 . We note further that the School District’s 2004-05 secondary education "Enrollment Guide” continues to describe the operation of the racial tiebreaker as part of its open choice assignment program, explaining that "The integration positive tiebreaker has been suspended for the 2004-2005 assignment period due to the pendency of a lawsuit challenging its use.” See Seattle Public Schools, Middle and High School Choices 2004-2005: Enrollment Guide for Parents 44, available at http://www.seattleschools.org/area/eso/secondary_guide_04_05.pdf (last visited June 8, 2004) (emphasis added).

14 . Because "discrimination that violates the Equal Protection Clause of the Fourteenth Amendment committed by an institution that accepts federal funds also constitutes a violation of Title VI,” we address Parents’s twin challenges to the racial tiebreaker simultaneously. See Gratz, 539 U.S. at 276 n. 23, 123 S.Ct. 2411.

As with any grant or denial of summary judgment, the district court's resolution of cross-motions for summary judgment is reviewed de novo. United States v. City of Tacoma, 332 F.3d 574, 578 (9th Cir.2003). Because "[njeither side contends that there are any genuine issues of material fact ..., our task is to determine whether the district court correctly applied the relevant substantive law.” Arakaki v. Hawaii, 314 F.3d 1091, 1094 (9th Cir.2002). Such purely legal determinations are, of course, subject to de novo review on appeal.

15 . See, e.g., Freeman v. Pitts, 503 U.S. 467, 494, 112 S.Ct. 1430, 118 L.Ed.2d 108 (1992) ("Racial balance is not to be achieved for its own sake. It is to be pursued when racial imbalance has been caused by a constitutional violation.”); Metro Broad., Inc. v. FCC, 497 U.S. 547, 612, 110 S.Ct. 2997, 111 L.Ed.2d 445 (1990) ("Under the appropriate standard, strict scrutiny, only a compelling interest may support the Government’s use of racial- classifications. Modem equal protection doctrine has recognized only one such interest: remedying the effects of racial discrimination.”) (O'Connor, J., joined by Rehnquist, C.J., Scalia and Kennedy, JJ., dissenting); Richmond v. J.A. Croson Co., 488 U.S. 469, 493, 109 S.Ct. 706, 102 L.Ed.2d 854 (1989) (O'Connor, J., joined by Rehnquist, C.J., and White and Kennedy, JJ.) ("Classifications based on race carry a danger of stigmatic harm. Unless they are strictly reserved for remedial settings, they may in fact promote notions of racial inferiority and lead to a politics of racial hostility.”); id. at 524, 109 S.Ct. 706 (Scalia, J., concurring) ("[TJhere is only one circumstance in which the States may act by race to 'undo the effects of past discrimination'; where that is necessary to eliminate their own maintenance of a system of unlawful racial classification.”) (emphasis in original); Fullilove v. Klutznick, 448 U.S. 448, 489, 100 S.Ct. 2758, 65 L.Ed.2d 902 (1980) (Burger, C J., joined by White and Powell, IT.) ("That the use of racial and ethnic criteria is premised on assumptions rebuttable in the administrative process gives reasonable assurance that application of the [contracting set-aside] program will be limited to accomplishing the remedial objectives contemplated by Congress....”); id. at 530, 100 S.Ct. 2758 (Stewart, J., joined by Rehnquist, J., dissenting) ("Since the [set-aside] provision was in whole or in part designed to effectuate objectives other than the elimination of the effects of racial discrimination, it cannot stand as a remedy that comports with the strictures of equal protection, even if it otherwise could.”); Ho, 147 F.3d at 864 ("[T]he Supreme Court has not banished race altogether from our governmental systems. The concept, so long the instrument of governmental evil, so fraudulently promoted by pseudo-science, so corrosive of the rights of the person, may still be employed if its use is found to be necessaiy as the way of repairing injuries inflicted on persons because of race. Deployed for that limited purpose....”); Monterey Mech., 125 F.3d at 713 ("For a racial classification to survive strict scrutiny in the context before us, it must be a narrowly tailored remedy for past discrimination, active or passive, by the governmental entity making the classification.”); Coral Constr. Co. v. King County, 941 F.2d 910, 920 (9th Cir.1991) ("Race-based classifications must be reserved strictly for remedial settings."); see also Hopwood v. State of Texas, 78 F.3d 932, 944 (5th Cir.1996); Contractors Ass’n v. City of Phila., 91 F.3d 586, 596 (3d Cir.1996); Aiken v. City of Memphis, 37 F.3d 1155, 1162-63 (6th Cir.1994); In re Birmingham Reverse Discrimination Employment Litig., 20 F.3d 1525, 1544 (11th Cir.1994); O’Donnell Constr. Co. v. District of Columbia, 963 F.2d 420, 424 (D.C.Cir.1992); Podberesky v. Kirwan, 956 F.2d 52, 56 (4th Cir.1992) (en banc); Cunico v. Pueblo Sch. Dist. No. 60, 917 F.2d 431, 437 (10th Cir.1990).

16 . But cf. Metro Broad., 497 U.S. at 612-13, 110 S.Ct. 2997 (O'Connor, J., dissenting) (condemning the majority’s decision to apply intermediate scrutiny to racial classifications on grounds that doing so "too casually extends the justifications that might support racial classifications, beyond that of remedying past discrimination”); cf. also Croson, 488 U.S. at 493, 109 S.Ct. 706 (plurality op. by O'Connor, J.).

17 . See, e.g., Grutter, 539 U.S. at 322, 123 S.Ct. 2325 ("We granted certiorari to resolve ... [w]hether diversity is a compelling interest that can justify the narrowly tailored use of race in selecting applicants for admission to public universities.") (citation omitted); id. at 325, 123 S.Ct. 2325 ("[Tjoday we endorse Justice Powell's view that student body diversify is a compelling state interest that can justify the use of race in university admissions.”); id. at 331, 123 S.Ct. 2325 ("[T]he diffusion of knowledge and opportunity through public institutions of higher education must be accessible to all ....”); id. at 331-32, 123 S.Ct. 2325 ("[Ejnsuring that public institutions are open and' available to all segments of American society, including people of all races and ethnicities, represents a paramount government objective. [Njowhere is the importance of such openness more acute than in the context of higher education.”) (quoting Brief of the United States); id. at 332, 123 S.Ct. 2325 ("[Ujniversities, and in particular, law schools, represent the training ground for a large number of our Nation’s leaders.”).

18 . We express no opinion on the extent to which the diversity rationale extends beyond the secondary educational context. Cf. Petit v. City of Chicago, 352 F.3d 1111 (7th Cir.2004) (addressing the diversity rationale in the context of race-conscious promotions within a police department).”

19 . For present purposes, we focus on the University's post-1998 admissions program rather than its more troubling predecessors. See Gratz v. Bollinger, 122 F.Supp.2d 811, 831-33 (E.D.Mich.2000) (describing the myriad constitutional defects of the 1995-98 program).

20 . In actuality, LSA’s mechanical award of 20 points may have been worth more than one-fifth of the points necessary to secure admission. In 1999, LSA also established a discretionary review process in which individual applications would be given special additional consideration from the Admissions Review Committée (ARC) if the applicant met a threshold selection index score (80 points for Michigan residents, 75 points for out-of-state residents) and was selected by an admissions officer. Id. at 256-57 & n. 8, 123 S.Ct. 2411. If chosen, ARC would consider whether the applicant "possesses a quality or characteristic important to the University's composition of its freshman ( class” — including race and ethnicity — and could then choose to admit such students outside the strictures of the normal selection index grid. Id. Thus, under LSA’s plan, membership in an underrepresented racial group would secure an applicant at least one-fifth of the points necessary to ensure admission, as well as one-quarter of the points necessary for an individualized review that, by facilitating the double-counting of race, could lead to admission.

21 . Of some note, Justice O'Connor's concurrence (joined by Justice Breyer) highlighted key distinctions between the Law School and under-graduate programs:

The law school considers the various diversity qualifications of each applicant, including race, on a case-by-case basis. By contrast, the Office of Undergraduate Admissions relies on the selection index to assign every underrepresented minority applicant the same, automatic 20-point bonus without consideration of the particular background, experiences, or qualities of each individual applicant. And this mechanized selection index score, by and large, automatically determines the admissions decision for each applicant. The selection index thus precludes admissions counselors from conducting the type of individualized consideration the Court’s opinion in Grutter requires: consideration of each applicant’s individualized qualifications, including the contribution each individual's race or ethnic identity will make to the diversity of the student body, taking into account diversity within and among all racial and ethnic groups.

Gratz, 539 U.S. at 276-77, 123 S.Ct. 2411 (O’Connor, J., concurring, joined by Breyer, J.) (emphases in original).

22 . Although we admire the School District's use of a "thermostat” designed to curtail the deleterious impact of its racial tiebreaker once the requisite racial proportion is attained in a given school, the fact that it turns the tiebreaker on and off with such numerical precision ultimately helps confirm that the School District's aim is simple proportional representation by race. Cf. Eisenberg, 197 F.3d at 132 (“The fact that the ... diversity profile for each school is reviewed and adjusted each year to avoid the facilitation and the creation of a racially isolated environment does not make the policy narrowly tailored. Instead, it manifests Montgomery County's attempt to regulate transfer spots to achieve the racial balance or makeup that most closely reflects the percentage of the various races in the county’s public school population.”).

23 . Indeed, when asked whether the District gave "any serious consideration to the adoption of a plan for the assignment of high school students that did not use racial balancing as a factor or goal,” Superintendent 01-chefske responded: "I think the general answer to your question is no. ... I mean it’s possible informal ideas were floated here or there, but I don’t remember any significant staff work being done.” When asked whether he could "ever recall the board considering any race neutral plans,” John Vacchiery (who heads the District’s Facilities, Planning, and Enrollment Department) answered simply "No.” And responding to inquiries regarding the possibility of using a “system kind of similar to the one you have now but without race as a tiebreaker,” Board member Don Nielsen replied that "It’s never been considered.”

24 . As the Court explained:

The Law School[ ] ... considers race as one factor among many, in an effort to assemble a student body that is diverse in ways broader than race. Because a lottery would make that kind of nuanced judgment impossible, it would effectively sacrifice all other educational values, not to mention every other kind of diversity.

Grutter, 539 U.S. at 340, 123 S.Ct. 2325.

25 . In particular, the School District’s data expert, Morgan Lewis, testified that although Board member Don Nielsen once suggested instituting an income-based tiebreaker, "[t]hat particular proposal never went anywhere, in terms of assessing how you would implement it” and that School District staff was never asked to (and so never did) examine the effect that such a tiebreaker would have on the demographic composition of the District's high schools.

26 . Even beyond this evidentiary inadequacy, we observe that the individual Board members’ post-hoc litigation rationales for the Board's allegedly having "rejected” the use of broader diversity considerations do not withstand even limited scrutiny. Whether it is "often inaccurate” to assume that poverty and race are "coextensive” and "insulting to minorities” to suggest they may be, see post at 1008, it is beyond dispute that despite remarkable racial progress during the past century, see generally Stephan and Abigail Thernstrom, America in Black and White (1997), race and socioeconomic status remain correlated, so that the latter might serve as a workable race-neutral alternative satisfying the District's interest in obtaining the benefits of enrolling an otherwise diverse student population. (Of course, we can only determine whether it would do so if the School District would evaluate it.) Likewise, it is beyond dispute that any insult generated by recognizing the opportunity presented by this sad reality pales in comparison to the insult of rejecting an applicant solely because of the color of her skin (whether they are white or minority — as in this case, see supra at 955-56 & n. 7).

Finally, we find unsupportable the individual Board members’ suggestion that socioeconomic integration would not result in appreciable socialization advantages because "implementation would be thwarted by high school students’ reluctance to reveal their socioeconomic status to their peers.” Post at 1008. For if the theory underpinning integrated schooling is (quite accurately) that it will foster cross-bounded friendships, it is inevitable that students of varying backgrounds will not only interact in the classroom, but outside it — where exposure to their peers' varying communities, families, and lifestyles will almost certainly reveal their divergent socioeconomic experiences.

27 . As their report put the point, "With quality high schools throughout the city, assignment issues will disappear.”

28 . Of note, the report also concluded by discussing four additional alternatives that it ultimately declined to propose to the Board— but nonetheless deemed worthy of consideration.

29 . Our nation’s public schools, especially those in central cities, currently face a crisis of epic proportions. Although black and Latino students made substantial educational gains following the demise of legalized governmental segregation, the best data on educational attainment demonstrates that no further progress has been made in the past fifteen years — and indeed that the gap between white and Asian students’ educational achievement and that of blacks and Latinos has actually expanded on some measures during that period. See Stephan Thernstrom & Abigail Thernstrom, No Excuses: Closing the Racial Gap in Learning 18-21 (2003); see also John E. Chubb and Tom Loveless, eds., Bridging the Achievement Gap 1-2 (2002); David Grissmer, Ann Flanagan, and Stephanie Williamson, Why Did the Black-White Test Score Gap Narrow in the 1970s and 1980s?, in Christopher Jencks & Meredith Phillips, eds., The Black-White Test Score Gap 185-91 (1998). As the Thernstroms poignantly observe:

Racial progress on many fronts has been enormously heartening. But in a society committed to equal opportunity, we still have a racially identifiable group of educational have-nots — young African-Americans and Latinos whose opportunities in life will almost inevitably be limited by their inadequate education. .. . [Such] [o]ngoing racial inequality is not only morally unacceptable; it corrupts the fabric of American society and endangers our future.

Id. at 271-72.

30 . Indeed, Olchefske also recommended expanding the band in 1999.

31 . The dissent suggests that even though schools "may well have been sufficiently diverse to promote interaction across the white/nonwhite axis and to prevent the token-ization of nonwhite students” had the District used a 20 percent band (or, in something of a surprising concession, had it used no integration tiebreaker at all), see post at 1006-1007 & 1011, the District's selection of a 15 percent band was appropriate because it would facilitate greater movement of students between the south end of the District and the north. Id. Given Superintendent Olchefske’s statement that the District's interests in the tiebreaker would not be served by setting the band at 15 percent instead of 20 percent, we find it hard to credit the assertion that those interests must have included such movement of students. In any event, we hardly think that non diversity-enhancing north-south integration is sufficiently compelling to justify the use of a racial classification.

32 . The only narrow tailoring requirement the School District’s racial tiebreaker even arguably satisfies is the final one — that the institutional use of race be time-limited. Addressing temporal limitations in Grutter, the Court explained that "[i]n the context of higher education, the durational requirement can be met by sunset provisions in race-conscious admissions policies and periodic reviews to determine whether racial preferences are still necessary to achieve student body diversity.” 539 U.S. at 342, 123 S.Ct. 2325 (emphases added). Yet, notwithstanding the conjunctive nature of its stated requirements and in spite of the fact that Michigan’s policy did not itself contain a sunset provision, the Court held the policy to be adequately limited after unhesitatingly crediting the University’s pledge "that it would like nothing better than to find a race-neutral admissions formula,” id. (quotation omitted), and stating that "We expect that 25 years from now, the use of racial preferences will no longer be necessary to further the interest approved today.” Id. at 343, 123 S.Ct. 2325. Whether that vague assertion represented a "holding that racial discrimination in higher education admissions will be illegal in 25 years,” id. at 351, 123 S.Ct. 2325 (Thomas, J., dissenting), or merely a "hope, but not [a] firm[] forecast, that over the next generation’s span, progress ... will make it safe to sunset affirmative action” in accordance with the ever-powerful "international understanding of the office of affirmative action,” id. at 344 & 346, 123 S.Ct. 2325 (Ginsburg, J., concurring), we cannot say. But the School District's annual review of the racial tiebreaker and the at least theoretical availability of a judicially-enforceable end-point to the School District’s racially discriminatory admissions program may well bring the policy into line with the durational limits required by strict scrutiny.

33 . Of course, the School District makes this bold assertion only when pressed by the full gravity of the Court’s pronouncements. Where otherwise convenient, the School District argues that "the Michigan decisions do provide crucial guidance to this Court,” and claims that beyond illuminating the contours of the compelling interest test, those decisions “also provide considerable assistance in applying the narrow tailoring element of strict scrutiny.”

34 . To its credit, the dissent declines to follow the School District's lead.

35 . We emphasize the District's over-reading of Associated General at the outset because— even beyond our express failure to address the issue joined here — it is not clear that Associated General implicitly suggested that reshuffle programs do not violate the Constitution. There is a world of difference between stating that " 'stacked deck' programs trench on Fourteenth Amendment values in ways that 'reshuffle' programs do not,” Associated General, 616 F.2d at 1387 (emphasis added), and holding that reshuffle programs designed to integrate schools cannot violate the Fourteenth Amendment at all.

36 . See also Shaw v. Hunt, 517 U.S. 899, 904-05, 116 S.Ct. 1894, 135 L.Ed.2d 207 (1996); Miller, 515 U.S. at 904, 115 S.Ct. 2475; Shaw v. Reno, 509 U.S. 630, 643, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993).

37 . Compare Miller v. Gammie, 335 F.3d 889, 902 (9th Cir.2003) (en banc) (Tashima, J., concurring) (citing Best Life Assurance Co. v. Comm'r, 281 F.3d 828, 834 (9th Cir.2002), for the proposition that "dictum [i]s a statement made during the course of delivering a judicial opinion, but one that is unnecessary to the decision in the case and therefore not precedential.”) (citations and quotations omitted) with Miranda B. v. Kitzhaber, 328 F.3d 1181, 1186 (9th Cir.2003) (per curiam) (" ‘[Wjhere a panel confronts an issue germane to the eventual resolution of the case, and resolves it after reasoned consideration in a published opinion, that ruling becomes the law of the circuit, regardless of whether doing so is necessary in some strict logical sense.’ ”) (quoting United States v. Johnson, 256 F.3d 895, 914 (9th Cir.2001) (Kozinski, J., concurring)) with United States v. Crawley, 837 F.2d 291, 292-93 (7th Cir.1988) (Posner, J.) (suggesting that rather than define dicta, courts should look to the "reasons there are against ... giving weight to a passage found in a previous opinion[:] One is that the passage was unnecessary to the outcome of the earlier case and therefore perhaps not as fully considered as it would have been if it were essential .... A closely related reason is that the passage was not an integral part of the earlier opinion — it can be sloughed off without damaging the analytical structure of the opinion, and so it was a redundant part of that opinion. ... Still another reason is that the passage was not grounded in the facts of the case and the judges may therefore have lacked an adequate experiential basis for it; another, that the issue addressed in the passage was not presented as an issue, hence was not refined by the fires of adversary presentation. All these are reasons for thinking that a particular passage was not a fully measured judicial pronouncement....”).

38 .In our intervening 24 years of jurisprudence addressing the use of race, we have referenced Associated General just once for the proposition that reshuffles are less suspect than stacked deck programs, and we did so only in discussing the interaction between a state ballot initiative and the Court's political structure doctrine, see Washington v. Seattle Sch. Dist. No. 1, 458 U.S. 457, 102 S.Ct. 3187, 73 L.Ed.2d 896 (1982); Hunter v. Erickson, 393 U.S. 385, 89 S.Ct. 557, 21 L.Ed.2d 616 (1969), rather than in applying the narrow tailoring requirement of the Fourteenth Amendment to a particular race-based program. Coalition, 122 F.3d at 707 n. 16. To the limited extent that Coalition’s passing reference to school desegregation programs might bear on the Equal Protection inquiry at issue here, we note simply that — among other crucial differences — the desegregation program referred to in the Coalition footnote "bas[ed] student assignments on attendance zones rather than on race,” Seattle, 458 U.S. at 461, 102 S.Ct. 3187, and thus would not even be necessarily subject to the tailoring analysis we must apply here (and to the use of all racial classifications). Adarand, 515 U.S. at 224, 115 S.Ct. 2097; Croson, 488 U.S. at 493-95, 109 S.Ct. 706.

39 . We recognize that more than 30 years ago the Court — in what is well recognized as dicta, see, e.g., Brewer v. W. Irondequoit Cent. Sch. Dist., 212 F.3d 738, 750 (2d Cir.2000)—suggested that school districts may pursue a measure of racial balance independent of constitutional requirements to remedy past discrimination. Swann v. Charlotte-Mecklenburg Bd. of Educ., 402 U.S. 1, 16, 91 S.Ct. 1267, 28 L.Ed.2d 554 (1971). In contrast to the dissent, which recognizes that "these statements must be considered in the light of the Court's later decisions ... establishing] that the government may not act in furtherance of racial balance without a compelling nonracial reason” but then declines to so read those statements, we do not feel free to privilege a gratuitous statement over the Court’s clear holdings.

40 . Although Associated General describes stacked deck programs as "specifically favoring] members of minorities in the competition with members of the majority," 616 F.2d at 1386, it is beyond any serious dispute that such a distinction is utterly bereft of force in the post-Croson, post-Adarand world — where our focus is on the use of racial classifications per se, rather than upon the race of those who are benefitted or burdened by the operation of the classification. Grutter, 539 U.S. at 326, 123 S.Ct. 2325; Gratz, 539 U.S. at 270, 123 S.Ct. 2411; Adarand, 515 U.S. at 224, 115 S.Ct. 2097; Croson, 488 U.S. at 493-95, 109 S.Ct. 706; see also Shaw II, 517 U.S. at 904-05, 116 S.Ct. 1894; Miller, 515 U.S. at 904, 115 S.Ct. 2475; Shawl, 509 U.S. at 643, 113 S.Ct. 2816; Jed Rubenfeld, Affirmative Action, 107 Yale L.J. 427, 434 ("The critical holding of Adarand was that all laws employing a racial classification must undergo strict scrutiny, with no exception made on the basis of allegedly benign intentions. The classification itself is the constitutionally suspect feature of the law, the feature that triggers heightened scrutiny, regardless of which race happens to be burdened, and regardless of the particular burdens imposed.”).

41 . Indeed, the oversubscribed Seattle high schools to which children of the Parents members seek admittance are "selective” in precisely the same way as the University of Michigan: Due to the quality of the education they provide, the availability of special academic programs, and their location, more students than can be accommodated seek admission — and the District must therefore determine which applicants will be offered a coveted seat in a more desirable school. We simply disagree with the dissent's assertion, post at 1012-1013, that school quality exists in some objective vacuum distinct from market assessment.

42 . We further hold that the School District’s racial tiebreaker violates Title VI. See Gratz, 539 U.S. at 275 & n. 23, 123 S.Ct. 2411.

43 . See generally Abigail Themstrom, Where Did All the Order Go? School Discipline and the Law, in Diane Ravitch, ed., Brookings Papers on Education Policy: 1999, 299-314 (1999).

44 . We offer another basis for declining to defer to the District's pursuit of racial proportionality in seeking diversity's benefits, infra at 984 n. 49.

45 . See Marcel Proust, Á la recherche du temps perdu [Remembrance of Things Past, or In Search of Lost Time] (1913-27).

46 . The same picture emerges from 2002-03 school year data contained in the School District’s "Individual School Profiles” report, submitted as an appendix to the supplemental amicus brief of Pacific Legal Foundation et al.

47 . Indeed, as one dissenting Board member noted in deposition testimony, "What we’ve done is we’ve defined ourselffrzc ] a problem by lumping all minorities together.... [O]ne of the things that the [District still hasn’t come to grips with is that minorities are quite different among and between themselves and sometimes have vast differences." It is in part for this reason that the history of racial conflict in this country is not one limited to tensions among whites and non-whites. See, e.g., Bill Ong Hing, In the Interest of Racial Harmony, 47 Stan. L.Rev. 901 (1995) (discussing tensions between the African-American and Asian communities). Inter-ethnic tensions persist even within diverse high schools. See, e.g., Carl Campanile, Now It's a Federal Case, N.Y. Post, Aug. 23, 2001, at 2 (detailing a series of incidents at Brooklyn's Lafayette High School between African-American students and Asians); Elissa Goot-man, City to Help Curb Harassment Of Asian Students at High School, N.Y. Times, June 2, 2004, at B9 (describing a recent consent decree designed to address these incidents, and noting the diversity of Lafayette's student population). Surely helping resolve these tensions is just as crucial to the District’s mission as addressing tensions between whites and non-whites.

48 .While we do not mean to suggest that better-calibrated pure racial balancing would be constitutionally permissible, it bears note that the tiebreaker does not even operate to correct "racial imbalance” at the two schools where the white/non-white student ratio is most out of line with the District demographics (Rainier Beach and Cleveland). Supra at 955-56 n. 6; post at 993-94 n. 13. Taking the dissent on its own terms, it is thus that much harder to credit its claim that the tiebreaker is adequately designed to address the "flip side” of the District's interest — ensuring that " '[n]o student ... attend a racially concentrated school.’ ” Post at 993-94 (alteration in original); see also id. at 1011.

49 . We further note that these numbers show the District's high schools would enroll a proportion of underrepresented minority students already recognized by the Court as adequate to spur the internal educational benefits of diversity. This fact further counsels against deferring to the District's decision to use a racial classification here. Cf. Grutter, 539 U.S. at 320, 123 S.Ct. 2325 ("[A] race-blind admissions system would have a very dramatic, negative effect on underrepresented minority admissions.... [Lfjnderrepresented minority students would have comprised 4 percent of the entering class in 2000 instead of the actual figure of 14.5 percent.”) (quotation omitted).

50 . Asked whether she was aware of any studies suggesting that the District’s goals were better achieved by shifting enrollment proportions by "three, four, or five percentage points,” Board Member Schaad-Lamphere said she was not aware of any such evidence and that she "ha[d] not used research to base flier] decisions on.”

51 . It is not even clear that the tiebreaker itself is fully responsible for this trivial shuffling of students. The dissent, post at 10011 n. 40, observes that among students who were denied admission to any school of their choice by operation of the tiebreaker, some 35 percent were assigned to the same school they would have been assigned to had the tiebreaker not operated. As a consequence, the numbers in Table 3 likely overestimate substantially the demographic impact of the tiebreaker— and thus cast further doubt over its usefulness.

In turn, the fact that these changes are so marginal dovetails our holding that the District. was obligated to seriously consider whether alternative arrangements could have met its goals. See supra at 970-75. Race-neutral alternatives not only may have produced equivalent levels of diversity in the District's schools, but far greater diversity than the tiebreaker.

52 . Likewise, while Latinos accounted for less than one-tenth of the students at these schools, and Native Americans less than one-twentieth of students, representational ranges for those student populations were, respectively, almost as large as the maximum effect and more than half the size of the maximum effect.

53 . The point holds even using the dissent's preferred metric (ninth grade — rather than overall — enrollment changes): Individual inter-ethnic disparities either trump or rival overall white/non-white enrollment variances attributable to the tiebreaker's operation, and the schools continue to enroll widely diverse ninth grade classes — a point essentially conceded by the dissent. Post at 1006-1007 & 1011.

54 . The dissent's assertion that the constitutional requirement of holistic review does not apply in this context fails for the same reason. Gnitter and its progenitors require that one's race be considered only as one among many factors not because "holistic review ... provides a closer fit with a university's interest in viewpoint diversity,” post at 1000-01, but because any interest in race alone is unconstitutional: "Outright racial balancing ... is patently unconstitutional.... Racial balance is not to be achieved for its own sake." Grutter, 539 U.S. at 330, 123 S.Ct. 2325 (quoting Freeman, 503 U.S. at 494, 112 S.Ct. 1430, and citing Bakke, 438 U.S. at 307, 98 S.Ct. 2733). Given such a clear — and oft-repeated — condemnation of the practice by the Supreme Court, it is hard seriously to credit the dissent's peculiar assertion that we misread the case law's per se ban on racial balancing. See post at 999-1000 n. 22.

55 . Recall the undisputed fact that the gap in average SAT scores between students at Garfield and Cleveland is nearly 400 points. See supra at 953-54 n. 1.