# Belk v. Charlotte-Mecklenburg > Court of Appeals for the Fourth Circuit · September 19, 2001 URL: https://www.frixlaw.com/law-library/cases/2967477 ## Case - **Court:** Court of Appeals for the Fourth Circuit - **Decided:** September 19, 2001 - **Precedential status:** Published - **Opinion:** Opinion - **Cited by:** 0 later opinions in the Frix Law Library ## Citator (automated) - No negative treatment found by the automated citator. That is not the same as a confirmation that the case is good law; read the citing cases. - Full citator and citing cases: https://www.frixlaw.com/law-library/cases/2967477 ## How later opinions describe it (automated extraction) - observing "that 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 judicial scrutiny" ## Opinion text Filed: September 19, 2001 UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT Nos. 99-2389(L) (CA-97-482-3-P, CA-65-1974-3-P) Terry Belk, et al., Plaintiffs - Appellants, versus William Capacchione, etc., et al., Plaintiffs - Appellees. O R D E R The court amends its opinion filed November 30, 2000, as follows: On page 3 -- the list of amici curiae is corrected to read “United States of America; North Carolina School Boards Association; National School Boards Association.” On page 4, section 2, line 4 -- the district court numbers are corrected to read “CA-97-482-3-P, CA-65-1974-3-P.” On page 23, first full paragraph, line 12 -- a comma is added after “(1979)” to complete the cite. - 2 - On page 60, indented quotation, line 8 -- the single quotation marks around “know-how” are corrected to double quotation marks. On page 108, first paragraph, last line -- “See Ante at 34" is corrected to read “See ante at 34.” For the Court - By Direction /s/ Patricia S. Connor Clerk Rehearing en banc granted by order filed 1/17/01; published opinion filed 11/30/00 is vacated Volume 1 of 3 PUBLISHED UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT TERRY BELK; DWAYNE COLLINS, on behalf of themselves and the class they represent, Plaintiffs-Appellants, v. WILLIAM CAPACCHIONE, Individually and on behalf of Christina Capacchione, a minor; MICHAEL P. GRANT; RICHARD EASTERLING; LAWRENCE GAUVREAU; KAREN BENTLEY; CHARLES THOMPSON; SCOTT C. WILLARD, Plaintiffs-Appellees, and No. 99-2389 THE CHARLOTTE-MECKLENBURG BOARD OF EDUCATION; ERIC SMITH, Superintendent, in his official capacity; ARTHUR GRIFFIN, Chairman of the Charlotte-Mecklenburg School Board, in his official capacity, Defendants. UNITED STATES OF AMERICA; NORTH CAROLINA SCHOOL BOARDS ASSOCIATION; NATIONAL SCHOOL BOARDS ASSOCIATION, Amici Curiae. WILLIAM CAPACCHIONE, Individually and on behalf of Christina Capacchione, a minor; MICHAEL P. GRANT; RICHARD EASTERLING; LAWRENCE GAUVREAU; KAREN BENTLEY; CHARLES THOMPSON; SCOTT C. WILLARD, Plaintiffs-Appellees, and TERRY BELK; DWAYNE COLLINS, on behalf of themselves and the class they represent, Plaintiffs, No. 99-2391 v. THE CHARLOTTE-MECKLENBURG BOARD OF EDUCATION; ERIC SMITH, Superintendent, in his official capacity; ARTHUR GRIFFIN, Chairman of the Charlotte-Mecklenburg School Board, in his official capacity, Defendants-Appellants. UNITED STATES OF AMERICA; NORTH CAROLINA SCHOOL BOARDS ASSOCIATION; NATIONAL SCHOOL BOARDS ASSOCIATION, Amici Curiae. 2 WILLIAM CAPACCHIONE, Individually and on behalf of Christina Capacchione, a minor; MICHAEL P. GRANT; RICHARD EASTERLING; LAWRENCE GAUVREAU; KAREN BENTLEY; CHARLES THOMPSON; SCOTT C. WILLARD, Plaintiffs-Appellees, and TERRY BELK; DWAYNE COLLINS, on behalf of themselves and the class they represent, Plaintiffs, No. 00-1098 v. THE CHARLOTTE-MECKLENBURG BOARD OF EDUCATION; ERIC SMITH, Superintendent, in his official capacity; ARTHUR GRIFFIN, Chairman of the Charlotte-Mecklenburg School Board, in his official capacity, Defendants-Appellants. v. UNITED STATES OF AMERICA; NORTH CAROLINA SCHOOL BOARDS ASSOCIATION; NATIONAL SCHOOL BOARDS ASSOCIATION, Amici Curiae. 3 WILLIAM CAPACCHIONE, Individually and on behalf of Christina Capacchione, a minor; MICHAEL P. GRANT; RICHARD EASTERLING; LAWRENCE GAUVREAU; KAREN BENTLEY; CHARLES THOMPSON; SCOTT C. WILLARD, Plaintiffs-Appellees, and TERRY BELK; DWAYNE COLLINS, on behalf of themselves and the class they represent, Plaintiffs, No. 00-1432 v. THE CHARLOTTE-MECKLENBURG BOARD OF EDUCATION; ERIC SMITH, Superintendent, in his official capacity; ARTHUR GRIFFIN, Chairman of the Charlotte-Mecklenburg School Board, in his official capacity, Defendants-Appellants. UNITED STATES OF AMERICA; NORTH CAROLINA SCHOOL BOARDS ASSOCIATION; NATIONAL SCHOOL BOARDS ASSOCIATION, Amici Curiae. Appeals from the United States District Court for the Western District of North Carolina, at Charlotte. Robert D. Potter, Senior District Judge. (CA-97-482-3-P, CA-65-1974-3-P) Argued: June 7, 2000 Decided: November 30, 2000 Before MOTZ, TRAXLER, and KING, Circuit Judges. 4 Affirmed in part, reversed in part, vacated in part, and remanded by published opinion. Judge Motz and Judge King wrote the opinion. Judge Traxler wrote a concurring and dissenting opinion. _________________________________________________________________ COUNSEL ARGUED: Stephen Luke Largess, FERGUSON, STEIN, WALLAS, ADKINS, GRESHAM & SUMTER, P.A., Charlotte, North Carolina; John W. Borkowski, HOGAN & HARTSON, L.L.P., Washington, D.C., for Appellants. Allan Lee Parks, PARKS, CHESIN & MILLER, P.C., Atlanta, Georgia, for Appellees. ON BRIEF: James E. Fergu- son, II, John W. Gresham, C. Margaret Errington, FERGUSON, STEIN, WALLAS, ADKINS, GRESHAM & SUMTER, P.A., Char- lotte, North Carolina; Elaine R. Jones, Director-Counsel, Norman J. Chachkin, Gloria J. Browne, NAACP LEGAL DEFENSE & EDU- CATIONAL FUND, INC., New York, New York; Allen R. Snyder, Maree Sneed, HOGAN & HARTSON, L.L.P., Washington, D.C.; James G. Middlebrooks, Irving M. Brenner, Amy Rickner Langdon, SMITH, HELMS, MULLISS & MOORE, L.L.P., Charlotte, North Carolina; Leslie J. Winner, General Counsel, CHARLOTTE- MECKLENBURG BOARD OF EDUCATION, Charlotte, North Car- olina, for Appellants. Kevin V. Parsons, PARKS, CHESIN & MIL- LER, P.C., Atlanta, Georgia; John O. Pollard, MCGUIRE, WOODS, BATTLE & BOOTHE, Charlotte, North Carolina; William S. Hel- fand, MAGENHEIM, BATEMAN, ROBINSON, WROTENBERY & HELFAND, Houston, Texas; Thomas J. Ashcraft, Charlotte, North Carolina, for Appellees. Bill Lann Lee, Acting Assistant Attorney General, Mark L. Gross, Rebecca K. Troth, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Amicus Curiae United States. Michael Crowell, Lisa Lukasik, THARRINGTON SMITH, L.L.P., Raleigh, North Carolina; Allison B. Schafer, General Counsel, NORTH CAROLINA SCHOOL BOARDS ASSOCIA- TION, Raleigh, North Carolina; Julie K. Underwood, General Coun- sel, NATIONAL SCHOOL BOARDS ASSOCIATION, Alexandria, Virginia, for Amici Curiae Associations. _________________________________________________________________ 5 OPINION DIANA GRIBBON MOTZ & ROBERT BRUCE KING, Circuit Judges: Since 1954, the school boards throughout this country, including the Charlotte-Mecklenburg Board of Education, have been operating under a standing Supreme Court mandate to integrate their school sys- tems and eliminate all vestiges of de jure segregation. Brown v. Board of Educ., 347 U.S. 483, 490 (1954) ("Brown I"). During the twenty years following the Supreme Court's mandate, the Charlotte- Mecklenburg Board of Education ("CMS" or the"Board") resisted all efforts to expedite desegregation, essentially arguing that, in light of the centuries over which the dual system of education had come to fruition, the Board would need a proportional period of time to develop remedies aimed at correcting past wrongs. Faced with this intransigence, the Supreme Court unanimously decided in 1971 that the Constitution required the Board to take affirmative measures, including the use of race-based ratios in student assignment, to eradi- cate vestiges of its invidious discrimination. See Swann v. Charlotte- Mecklenburg Bd. of Educ., 402 U.S. 1 (1971). Finally, in 1975, the Board began seeking to fulfill the Supreme Court's mandate that public schools be desegregated with "all deliber- ate speed." Brown v. Board of Educ., 349 U.S. 294, 299 (1955) ("Brown II"). Today, with the Board having had less than twenty-six years to implement appropriate remedies, we must decide whether the task of desegregating the Charlotte-Mecklenburg schools has reached its end. We hold that it has not. Over the Board's own admission to the contrary, the district court concluded that the school system had achieved unitary status across the board. While the district court made findings sufficient to hold that CMS had achieved unitary status in some respects, the court failed to adequately explore the return of pre- dominantly one-race schools as a vestige of segregation, rendering its findings insufficient to conclude that CMS has achieved unitary status in every respect. In an equally unprecedented ruling, the district court held that the school system, although operating under court orders to desegregate 6 its schools, violated the Constitution by employing a magnet school program that considered race in student assignment. On the contrary, because the Board's expanded magnet schools program -- and the race-sensitive method of student selection it employed -- was undertaken both to remedy the effects of past segregation and to com- ply with governing court orders, they did not and could not violate the Constitution. In this appeal, we consider the above rulings along with a number of related issues. As explained below, we affirm in part, reverse in part, vacate in part, and remand this case for further proceedings. I. A. In order to better understand the issues presented in this case, we must briefly review our country's history of school desegregation liti- gation, in which CMS has played a prominent role. Even after slavery had been abolished for almost a full century, African-American children were, for the most part, either excluded from the public schools or educated separately from white children. "In fact, any education of Negroes was forbidden by law in some states." Brown I, 347 U.S. at 490; see also Martin v. Charlotte- Mecklenburg Bd. of Educ., 475 F. Supp. 1319, 1324 (W.D.N.C. 1979) ("For three centuries racial segregation was the law of the land."). Indeed, throughout the early part of the 1900s, CMS operated a segre- gated school system within the safe harbor created by the Supreme Court's doctrine of "separate but equal" articulated in Plessy v. Fergu- son, 163 U.S. 537 (1896). In the middle of the 1900s, the Supreme Court began dismantling the great wall of segregation constructed under the imprimatur of Plessy. The Court initially sought to determine whether various "sepa- rate" African-American schools were genuinely "equal" to white schools by evaluating the quality of physical facilities, curricula, fac- ulty, and certain "intangible" considerations. See, e.g., Sweatt v. Painter, 339 U.S. 629 (1950); Sipuel v. Board of Regents of Univ. of 7 Okla., 332 U.S. 631 (1948). In each instance, the Court concluded that they were not. Id. In 1954, the Supreme Court at last overruled Plessy, declaring that "in the field of public education the doctrine of `separate but equal' has no place. Separate educational facilities are inherently unequal." Brown I, 347 U.S. at 495. Just one year later, the Court mandated that federal courts and school authorities take affirmative steps to achieve desegregation. Brown II, 349 U.S. at 299. Specifically, federal courts were to retain jurisdiction over desegregation cases during the period of transition, wielding their equitable powers to supervise school boards' efforts to effectuate integration. Id. at 300-01. One of the most important obligations of the federal courts was to ensure that school boards were proceeding in good faith to desegregate the public schools "with all deliberate speed." Id. at 301. With these seminal decisions -- Brown I and Brown II-- the Supreme Court promised the citizens of this country, and particularly African-American chil- dren, school systems "in which all vestiges of enforced racial segrega- tion have been eliminated." Wright v. Council of the City of Emporia, 407 U.S. 451, 463 (1972). Notwithstanding the Court's repeated admonition that segregation and its vestiges be eliminated "root and branch," Green v. County Sch. Bd. of New Kent County, 391 U.S. 430, 437-48 (1968), many school boards -- CMS included -- adopted "an all too familiar" response to the mandate of Brown II, interpreting "all deliberate speed" "as giving latitude to delay steps to desegregate." Freeman v. Pitts, 503 U.S. 467, 472 (1992). And so, lower federal courts, with the guidance and oversight of the Supreme Court, began fashioning equitable remedies to contend with school board recalcitrance. For example, in Green, the Supreme Court held that a "freedom of choice" plan, which per- mitted students -- regardless of race -- to choose the school they would attend, was by itself insufficient to meet the mandate of Brown. 391 U.S. at 430. In so holding, the Court recognized that more inten- sive efforts would be necessary in order to make "meaningful and immediate progress toward disestablishing state-imposed segrega- tion." Id. at 439. Subsequently, in this very case, the Court approved significant federal court intervention into a school system in order to eliminate segregation "root and branch," including the busing of stu- dents from schools close to their homes to schools farther away, the 8 use of race-based "mathematical ratios," and the alteration of student attendance zones. Swann, 402 U.S. at 15, 25, 28, 30-31. The Supreme Court has made clear, however, that a federal court's "end purpose must be to remedy the violation and, in addition, to restore state and local authorities to the control of a school system that is operating in compliance with the Constitution." Freeman, 503 U.S. at 489. Hence, as a school system eliminates the vestiges of past offi- cial segregation from certain facets of its operations, courts possess the authority to relinquish supervision in a commensurate fashion. Id. at 489-91. In this context, we examine the steps taken by CMS to eliminate the vestiges of segregation. B. 1. North Carolina's most significant initial response to the mandate of Brown II was the "Pupil Assignment Act of 1955-56, under which [the Board had] the sole power to assign pupils to schools, and chil- dren [were] required to attend the schools to which they [were] assigned." Swann v. Charlotte-Mecklenburg Bd. of Educ., 300 F. Supp. 1358, 1361 (W.D.N.C. 1969). This was an ineffectual measure -- perhaps intentionally so -- and by 1964, no more than a few dozen (out of more than 20,000) African-American children in CMS were attending schools with white children. Id. at 1362. 2. In 1965, the parents of African-American children attending CMS (hereinafter the "Swann plaintiffs")11 filed a class action seeking injunctive relief, claiming that the Board's policies and practices were perpetuating a segregated school system. Swann v. Charlotte- Mecklenburg Bd. of Educ., 243 F. Supp. 667, 668 (W.D.N.C. 1965). _________________________________________________________________ 1 Since this case was first filed in 1965, the various successor plaintiffs have been referred to as the Swann plaintiffs, a practice we continue to observe here. 9 On July 14, 1965, the district court approved a Board-proposed plan that closed certain black schools, built new schools, and established school attendance zones based on neighborhoods. But the linchpin of this plan was its grant of permission to each student -- regardless of race -- to freely transfer to a different school (often described as a "freedom of choice" plan). Id. In approving this plan, the district court held that CMS had no affirmative duty to "increase the mixing of the races"; instead, the Board's obligation under Brown II, according to the court, was to act without the intent to perpetuate segregation. Id. at 670. The following year, this Court affirmed the district court's interpretation of Brown II. See Swann v. Charlotte-Mecklenburg Bd. of Educ., 369 F.2d 29, 32 (4th Cir. 1966) ("Whatever the Board may do in response to its own initiative or that of the community, we have held that there is no constitutional requirement that it act with the con- scious purpose of achieving the maximum mixture of the races in the school population."). However, in the wake of the Supreme Court's 1968 decision in Green, which struck down a desegregation plan founded predomi- nantly on "freedom of choice," it became clear that school boards did possess an affirmative obligation to desegregate, not merely an obli- gation to implement race-neutral policies. Green, 391 U.S. at 437-38. Invigorated by the developing law, the Swann plaintiffs promptly filed a motion for further relief with the district court, seeking to expedite the desegregation process. 3. In 1969, Judge James B. McMillan, newly assigned to the Swann case,2 2 reexamined the Board's actions in light of Green and deter- mined that its "freedom of choice" plan, when coupled with geo- graphic zoning, were "not furthering desegregation." 300 F. Supp. at 1372. On the fundamental matters of assigning students and faculty, and the siting of new schools, the court made the following findings: - Student assignment: The court noted that a ratio of seventy percent white students to thirty percent black students, which approxi- _________________________________________________________________ 2 For clarity's sake, we will often refer within to the presiding district judge by name. 10 mated the ratio of white to black students in the county, tended to aid "better students [in holding] their pace, with substantial improvement for the poorer students." Id. at 1369. - Faculty assignment: Although faculty members were not being assigned with a discriminatory purpose, there was also "no sus- tained effort to desegregate faculties." Id. at 1370. The court ordered CMS to work actively to integrate the faculties, so that "a child attending any school in the system will face about the same chances of having a black or a white teacher as he would in any other school." Id. - School siting: The court underscored that the desirability of imple- menting a "neighborhood school" policy, under which efforts were made to locate schools in neighborhoods and within walking dis- tance for children, could not override the constitutional duty to desegregate. Id. at 1369. At the same time, CMS was not to avoid locating new facilities in black neighborhoods. Id. at 1371. In light of Green, Judge McMillan also ordered CMS to submit a new, amended desegregation plan, and he outlined certain possible remedies, including busing and re-zoning. Swann, 300 F. Supp. at 1360; Swann v. Charlotte-Mecklenburg Bd. of Educ., 306 F. Supp. 1299, 1302 (W.D.N.C. 1969). Once again, however, CMS was slow to respond, prompting Judge McMillan to impose a deadline of August 4, 1969, by which the Board was to submit a detailed desegregation plan to the court. See Swann v. Charlotte-Mecklenburg Bd. of Educ., 300 F. Supp. 1381, 1382, 1386 (W.D.N.C. 1969). CMS complied, and its proposed desegregation plan appeared to accept, for the first time, the constitu- tional duty to desegregate students, teachers, principals, and staffs "`at the earliest possible date.'" Swann v. Charlotte-Mecklenburg Bd. of Educ., 306 F. Supp. 1291, 1293 (W.D.N.C. 1969). The Board's pro- posed desegregation plan, approved by the district court on an interim basis ("interim desegregation plan"), included programs for faculty desegregation, the closing of seven all-black schools, and the reas- signment of pupils from the closed schools to outlying, predominantly white schools. Id. at 1298-99. In approving the plan on an interim basis, the district court noted that black children were bearing a dis- 11 proportionate burden of the desegregation efforts, but the court none- theless concluded that some action -- even if interim -- was preferable to none at all. Id. at 1298. Judge McMillan also ordered the Board to submit another desegregation plan within three months. In November and December 1969, the district court determined that the school system's compliance with the interim desegregation plan was unsatisfactory, finding that the Board was continuing to per- petuate segregation: The School Board is sharply divided in the expressed views of its members. From the testimony of its members, and from the latest report, it cannot be concluded that a majority of its members have accepted the court's orders as repre- senting the law which applies to the local schools. By the responses to the October 10 questions, the Board has indi- cated that its members do not accept the duty to desegregate the schools at any ascertainable time; and they have clearly indicated that they intend not to do it effective in the fall of 1970. They have also demonstrated a yawning gap between predictions and performance. Swann, 306 F. Supp. at 1306. At that time, the district court also reviewed and rejected the Board's newly submitted amended desegre- gation plan. Id. at 1313-14. Then, the court appointed Dr. John A. Finger, Jr. as an expert consultant to prepare a more acceptable plan. This appointment came nearly two years after the Supreme Court's Green decision and more than fifteen years after Brown I. The district court ultimately adopted Dr. Finger's proposed plan for elementary schools and the Board's plan, as modified by Dr. Finger, for secondary schools (collectively the "Finger Plan"). Swann v. Charlotte-Mecklenburg Bd. of Educ., 311 F. Supp. 265, 268-70 (W.D.N.C. 1970). In doing so, the court again observed the Board's failure to make an effective beginning to desegregation: "The School Board, after four opportunities and nearly ten months of time, have failed to submit a lawful plan (one which desegregates all the schools). This default on their part leaves the court in the position of being forced to prepare or choose a lawful plan." Id. at 267. 12 The Finger Plan included several components. First, students were to be assigned "in such a way that as nearly as practicable the various schools at various grade levels have about the same proportion of black and white students." Id. at 268. Second, "no school [could] be operated with an all-black or predominantly black student body." Id. Third, in redrawing the school system's attendance zones, the Board was authorized to use bus transportation and noncontiguous "satellite zones"33 to accomplish its goals. Id. Fourth, the district court restricted the student transfer policy in order to safeguard against resegregation. Id. at 268-69. Fifth, the race of faculty members at each school had to approximate the ratio of black and white faculty members through- out the system. Id. at 268. Sixth, the overall competence of teachers at formerly black schools could not be inferior to those at formerly white schools. Id. Finally, the district court mandated that the Board monitor and report on its progress in implementing the plan. Id. at 269. The Finger Plan was challenged on several occasions and, in 1971, the Supreme Court upheld it as a valid exercise of the district court's equitable powers. Swann, 402 U.S. at 31-32. Indeed, the Court specif- ically found that the district court's adoption of a student assignment plan that used race-based "mathematical ratios" as a starting point was well within the court's "equitable remedial discretion." Id. at 25. Even after the Supreme Court's decision in Swann, the district court found that the Board's desegregation efforts failed to meet con- stitutional requirements. For example, Judge McMillan ordered stu- dent assignment proposals revised in June 1971, finding that the proposals "were discriminatory in detail and in overall result; they _________________________________________________________________ 3 CMS used "satellite zones" in connection with elementary schools. Under this method, students from a small geographic area located outside an elementary school's primary attendance area were assigned to that school. J.A. 15571, 16052; see also Swann, 402 U.S. at 9 & n.3. The use of satellite zones was implemented by "pairing" elementary schools -- students from a predominantly black neighborhood were bused to a school in a predominantly white neighborhood for grades K-3, and stu- dents from a predominantly white neighborhood were bused to a school in a predominantly black neighborhood for grades 4-6. J.A. 15571, 16052; see also Swann, 402 U.S. at 9-10. 13 placed increasing burdens upon black patrons while partially relieving white patrons of similar burdens." Swann v. Charlotte-Mecklenburg Bd. of Educ., 328 F. Supp. 1346, 1347 (W.D.N.C. 1971). During the 1971-72 and 1972-73 school years, the district court attempted a "hands-off" approach, leaving the Board to remedy problems as they arose, but the court twice found that the Board still had not adopted sufficient measures to guard against resegregation and ensure that whites were bearing an appropriate share of the desegregation burden. See Swann v. Charlotte-Mecklenburg Bd. of Educ., 362 F. Supp. 1223, 1230 (W.D.N.C. 1973); Swann v. Charlotte-Mecklenburg Bd. of Educ., 379 F. Supp. 1102 (W.D.N.C. 1974); see also discussion of specific findings infra. The 1974 order expressed somewhat more optimism about the Board's desegregation efforts. In that order, Judge McMillan approved a student assignment proposal that, if implemented prop- erly, would result in "a fair and stable school operation" and would permit the court to close the case as an active matter. See 379 F. Supp. at 1103. The proposal made provisions for several "optional schools" -- schools that would offer some specialized program or curriculum and thereby attract students of all races from across Charlotte and Mecklenburg County. Although Judge McMillan approved the incor- poration of these schools into the plan, he cautioned that the optional schools would be inconsistent with the school board's constitutional obligations if they merely served to re-institute "freedom of choice." Id. at 1104 ("`Freedom of choice' was a synonym for segregation for many years, and . . . it should not be resurrected at this late date sub nom. `optional schools' without adequate safeguards against discrimi- natory results."). To ensure that the optional schools served their stated purpose of furthering the process of desegregation, Judge McMillan decreed that "optional school enrollments will be con- trolled starting with 1974 so that they . . . have about or above 20% black students." Id. Finally, in July 1975, over twenty years after the mandate of Brown II, Judge McMillan for the first time observed, albeit with reserva- tions, that the Board was actually working toward desegregation: "The new Board has taken a more positive attitude toward desegrega- tion and has at last openly supported affirmative action to cope with recurrent racial problems in pupil assignment." Swann v. Charlotte- 14 Mecklenburg Bd. of Educ., 67 F.R.D. 648, 649 (W.D.N.C. 1975). Although the district court cautioned that problems remained, the new vigor with which the Board was pursuing desegregation persuaded Judge McMillan to close Swann as an active matter of litigation and to remove it from the court's docket. Id. at 649-50. In so acting, the court reaffirmed that its orders still stood: "[t]his case contains many orders of continuing effect, and could be re-opened upon proper showing that those orders are not being observed." Id. at 649. 4. Between 1975 and 1992, two significant actions were taken in con- nection with the CMS desegregation litigation. a. First, in 1978, a group of white parents and children brought suit against CMS, seeking an order prohibiting the Board from assigning children pursuant to the Board's latest student-assignment plan. See Martin, 475 F. Supp. at 1320. The Martin plaintiffs claimed that the Supreme Court's then-recent decisions in Pasadena City Bd. of Educ. v. Spangler, 427 U.S. 424, 436 (1976), and University of Cal. Regents v. Bakke, 438 U.S. 265, 305 (1978), prohibited any consideration of race in student assignment. 475 F. Supp. at 1321. The Swann plain- tiffs intervened in Martin, joining the Board's opposition to the con- tentions of the Martin plaintiffs. Id. A brief review of Spangler and Bakke is necessary to an under- standing of Martin. In Spangler, the Supreme Court held that because the Pasadena Unified School District ("PUSD") had achieved racial neutrality in its school attendance pattern, "the District Court was not entitled to require the PUSD to rearrange its attendance zones each year so as to ensure that the racial mix desired by the court was main- tained in perpetuity." 427 U.S. at 436. All parties in Spangler agreed that the plan initially achieved racial neutrality in student attendance; nonetheless, the district court had believed it was empowered to annu- ally readjust school boundaries to ensure in perpetuity that there would be no majority of any minority race at any Pasadena school. Id. at 433, 436. In Bakke, the Supreme Court determined that a public university with no history of discrimination could not constitutionally 15 reserve sixteen out of one hundred admission slots for racial minori- ties. 438 U.S. at 319-20. In striking down this admissions plan, the Court had made clear that "[w]hen a classification denies an individ- ual opportunities or benefits enjoyed by others solely because of his race or ethnic background, [it must] be regarded as [constitutionally] suspect." Id. at 305. Judge McMillan, who retained jurisdiction over Swann and pre- sided over Martin, first held that because CMS had not achieved racial neutrality in student attendance, consideration of race in student assignment policies was appropriate under Swann . See Martin v. Charlotte-Mecklenburg Bd. of Educ., 626 F.2d 1165 (4th Cir. 1980). He explained that because the student assignment policy in the CMS school system had been independently adopted by the Board, it was not established, as the Spangler policy had been, via judicial coercion or order. 475 F. Supp. at 1340-43. Second, Judge McMillan ruled that Bakke was inapposite to the claims of the Martin plaintiffs. Specifi- cally, the court reasoned that no child was being denied access to equal educational opportunity because of race, see id. at 1321, and the actions of the Board were therefore not constitutionally suspect under Bakke. In upholding the independent actions of the Board, Judge McMil- lan made several important findings. For example, he found that dis- crimination had not ended; indeed, it was this very finding that led the court to uphold the 1978 race-conscious student assignment policy. Id. at 1346-47. Also, although for the first time the district court praised the efforts of the Board without reservation, it underscored yet again the need for patience and continued efforts: It took three centuries to develop a slave culture, to fight a bloody civil war, and to live through the century of racial turmoil after that war. *** The culture and attitudes and results of three centuries of segregation cannot be eliminated nor corrected in ten years. Human nature and practices don't change that fast, even in the hands of people of good will like the members of the 16 present School Board. They need time to work their own experiments, and to find their own ways of producing the sustained operation of a system of schools in which racial discrimination will play no part. I vote to uphold their efforts to date, and to give them that time. Id. at 1347. In 1980, we affirmed the district court's decision in Mar- tin. See 626 F.2d at 1165. b. The second significant phase of litigation between 1975 and 1992 was initiated in 1980. At that time, CMS and the Swann plaintiffs notified the district court that the black student population in CMS elementary schools had grown from twenty-nine percent to forty per- cent, making it increasingly difficult to comply with the desegregation order's mandate to avoid majority-black elementary schools. In response to this change, Judge McMillan approved a modification to the desegregation plan. Instead of prohibiting a "predominantly black student body," the court permitted CMS to operate elementary schools with a black student population of "plus 15 percent" above the district-wide average. Thus, if the school district averaged forty percent black students, any individual school could have fifty-five percent black students. 5. From 1981 to 1992, the Board continued to operate its desegrega- tion plan as approved by the district court, focusing, inter alia, on sat- ellite attendance zones, a feeder plan (assigning middle-school students from a certain neighborhood to identified high schools), school closings, and construction of new schools. Then, in 1992, CMS substantially increased its reliance on "optional" or magnet schools (the "expanded magnet schools program"). The Board placed new emphasis on magnet schools in order to phase out"pairing" and heavy reliance on busing, and to give parents more choice in school selection. It was the expanded magnet schools program that ultimately led to the present phase of this litigation. 17 6. In September 1997, William Capacchione, individually and on behalf of his daughter Christina, sued CMS claiming that Christina was unconstitutionally denied admission to a magnet school. Chris- tina is Hispanic and Caucasian, and her suit under 42 U.S.C. § 1983 sought declaratory, injunctive, and compensatory relief. In response, CMS moved to dismiss Capacchione's suit and, almost simulta- neously, the Swann plaintiffs moved to reactivate Swann, claiming that CMS was not yet in compliance with past desegregation orders and had not yet achieved unitary status. Because Judge McMillan had died, the cases were assigned to Senior Judge Robert D. Potter, who restored Swann to the district court's docket, consolidated the cases, denied CMS's motion to dismiss, and granted Capacchione's motion to intervene.44 The Capacchione plaintiffs claimed that CMS had long since elim- inated the vestiges of segregation in its schools, and that its formerly dual system of white and black schools had, for some time, been uni- tary. They also contended that CMS, while still operating under the court's desegregation orders, had violated those orders and the consti- tutional rights of white students in its efforts to desegregate the school system by employing a race-conscious assignment lottery in its expanded magnet schools program. The Swann plaintiffs countered that the school system had not yet achieved unitary status. CMS acknowledged that it was not yet in compliance with past desegrega- tion orders and agreed that it should not be declared to have achieved unitary status. CMS also contended that, in any event, the expanded magnet schools program constituted an entirely constitutional and _________________________________________________________________ 4 Since filing suit, the Capacchiones have moved to California. Based on that fact and other findings, the district court determined that William Capacchione no longer possessed standing to seek injunctive or declara- tory relief, but that he did have standing to pursue compensatory relief. Capacchione v. Charlotte-Mecklenburg Sch., 57 F. Supp. 2d 228, 240 (W.D.N.C. 1999). Another group of white parents intervened in the con- solidated action and that group, represented by plaintiff Michael Grant, claimed that CMS has achieved unitary status. The various groups of plaintiffs that have joined in Capacchione's claims are hereinafter referred to as "the Capacchione plaintiffs." 18 appropriate integration tool authorized under the desegregation orders in this case. The Swann plaintiffs, while endorsing the concept of magnet schools, argued that the expanded magnet schools program, as implemented, was contributing to the resegregation of the school system. Following a bench trial conducted from April 19 to June 22, 1999, the court, on September 9, 1999, filed its Memorandum of Decision and Order, from which this appeal is taken. See Capacchione v. Charlotte-Mecklenburg Sch., 57 F. Supp. 2d 228 (W.D.N.C. 1999). Although the Board claimed that unitary status had not been achieved, the district court found that it had. In its ruling, the district court then found that the Board's expanded magnet schools program, even though instituted to effect court-ordered desegregation, was unconsti- tutional. Furthermore, the court enjoined the Board from "assigning children to schools or allocating educational opportunities and bene- fits through race-based lotteries, preferences, set-asides, or other means that deny students an equal footing based on race." Id. at 294. Finally, the court awarded the Capacchione plaintiffs nominal mone- tary damages and substantial attorney's fees. Following the filing of timely notices of appeal, the Swann plain- tiffs and CMS sought a stay of Judge Potter's September 9, 1999 injunction. On December 30, 1999, we granted the requested stay pending further order of this court. Thereafter, the Capacchione plain- tiffs petitioned for an initial hearing en banc, which was denied by an eight-to-three vote of the Court. Belk v. Charlotte-Mecklenburg Bd. of Educ., 211 F.3d 853 (4th Cir. 2000). The panel heard argument in these appeals on June 7, 2000. C. We are now called upon to review the district court's various deter- minations relating to these consolidated appeals. Having reviewed and carefully considered each of the important questions determined by the district court, we affirm in part; however, we must reverse in part, and we must vacate and remand on certain issues. The district court's findings suffice to uphold its determination that the Board achieved progress toward desegregation in the 1970s and 19 1980s, and the court was therefore justified in concluding that CMS had achieved unitary status in some respects. However, the district court's findings do not support its conclusion that CMS has attained unitary status in every respect. Moreover, even if CMS could now properly be found to have achieved unitary status in toto, it would be inappropriate to assess the expanded magnet schools program as if there had been no court order in place. The expanded magnet schools program and the race-conscious lottery it employed were undertaken to remedy the effects of past segregation and were in compliance with court orders governing this case; they do not violate the Constitution. Consequently, the award of nominal damages, substantial attorney's fees, and an overbroad injunction barring any future consideration of race in school assignments and other fundamental aspects of school operations cannot stand. II. We first address the district court's unitary status decision. The determination of whether any part of a school system has achieved unitary status is a factual one; therefore, we review the district court's findings as to unitary status for clear error. See Riddick v. School Bd. of the City of Norfolk, 784 F.2d 521, 533 (4th Cir. 1986); see also Jacksonville Branch, NAACP v. Duval County Sch. Bd., 883 F.2d 945, 952 n.3 (11th Cir. 1989) (citing United States v. Texas Educ. Agency, 647 F.2d 504, 506 (5th Cir. Unit A 1981)). We do not, how- ever, defer to the district court on conclusions of law, including the district court's understanding of controlling law or the various bur- dens of proof and presumptions; consequently, we review any such conclusions of law de novo. See, e.g., In re Brice, 188 F.3d 576, 577 (4th Cir. 1999). A. 1. Indisputably, the school system of Charlotte-Mecklenburg County subjected African-Americans to nearly a century of segregation and discrimination. Indeed, the Supreme Court recognized as much in Swann, noting that North Carolina was one of the states with "a long history of maintaining two sets of schools in a single school system 20 deliberately operated to carry out a governmental policy to separate pupils in schools solely on the basis of race." 402 U.S. at 5-6. In this context the remedies forcefully endorsed in Brown II, including the use of race conscious measures, are necessary to eradicate the invidi- ous segregation at which they are aimed. Moreover, court supervision over local school boards, also embraced in Brown and its progeny, is entirely appropriate whenever "school authorities fail in their affirmative obligations" "to take what- ever steps might be necessary to convert to a unitary system in which racial discrimination would be eliminated." Swann, 402 U.S. at 15. Not only are the federal courts entitled to supervise and direct the actions of local school boards under those circumstances, but the scope of federal authority is almost plenary: "Once a right and a vio- lation have been shown, the scope of a district court's equitable pow- ers to remedy past wrongs is broad, for breadth and flexibility are inherent in equitable remedies." Id. There is no doubt that CMS was justifiably subjected to federal court supervision; in fact, even after the Board had been subjected to court supervision, it had to be repeat- edly ordered to begin the process of desegregation. Ultimately, however, the goal in a desegregation case such as this is to reach the point at which federal supervision is no longer war- ranted and the use of race-conscious measures is no longer necessary. See Freeman, 503 U.S. at 489. The Supreme Court has identified six factors (collectively the "original Green factors") that must be free from racial discrimination before the mandate of Brown is met: (1) student assignment, (2) physical facilities, (3) transportation, (4) fac- ulty, (5) staff, and (6) extracurricular activities. Green, 391 U.S. at 435. Not only are reviewing courts to ascertain whether these original Green factors are free from racial discrimination, but courts also are entitled, in their discretion, to identify other factors ("ancillary factors")5 5 and "determine whether minority students were being disadvantaged in ways that required the formulation of new and further remedies to ensure full compliance with the court's decree." 503 U.S. at 492. _________________________________________________________________ 5 For convenience, we refer to the original Green factors and any ancil- lary factors identified by the district court as "Green factors." 21 2. For school systems proceeding through the difficult process of desegregation, the Supreme Court has adopted the goal of achieving unitary status. Freeman, 503 U.S. at 486-87; Board of Educ. of Okla. City Pub. Sch. v. Dowell, 498 U.S. 237, 245-46 (1991). Although prior to the Court's Dowell and Freeman decisions federal courts used the term "unitary status" somewhat inconsistently, see Freeman, 503 U.S. at 486-87; Green, 391 U.S. at 437-38, the term has now come to mean that the school system has been unified such that the vestiges of segregation have been eliminated to the extent practicable. Free- man, 503 U.S. at 487; Green, 391 U.S. at 437-38. When a school sys- tem achieves unitary status, federal courts must withdraw supervision over the local school board. In this case, Judge Potter declared that CMS had achieved unitary status in every respect. The Supreme Court has directed that an appel- late court review a district court's unitary status determination by applying a two-part inquiry (the "Freeman inquiries"). An appellate court must determine if (1) a school Board has, in good faith, com- plied with the desegregation decree since it was entered; and (2) the vestiges of de jure segregation in the school system have been elimi- nated to the extent practicable. See Freeman, 503 U.S. at 492 (citing Dowell, 498 U.S. at 249-50). If the party seeking a declaration of unitary status cannot demon- strate that the school system has achieved unitary status in its entirety, we then undertake to determine whether the school system has achieved unitary status with respect to one or some of the Green fac- tors ("partial unitary status"). At that point, we apply, with respect to each Green factor, the two Freeman inquiries along with one addi- tional Freeman-mandated inquiry: "whether retention of judicial con- trol [over one aspect of the school system] is necessary or practicable to achieve compliance with the decree in other facets of the school system." Freeman, 503 U.S. at 491. This third Freeman inquiry rec- ognizes that the Green factors are -- to a great extent -- interrelated, and when determining whether judicial supervision over a school board may be withdrawn, the overlap between the Green factors is a crucial consideration. 22 The Freeman analysis brings us to the most difficult questions presented in any desegregation case: whether present racial isolation is a vestige of past segregation and, if so, whether a school board can practicably reduce that racial isolation. It is even difficult to define "vestige" in this context. See id. at 502 (Scalia, J., concurring) ("We have never sought to describe how one identifies . . . a `vestige' or a `remnant' of past . . . ."). The vestiges "that are the concern of the law may be subtle and intangible but nonetheless they must be so real that they have a causal link to the de jure violation being remedied." Id. at 496 (Kennedy, J.); see also id. at 512 (Souter, J., concurring) (citing Columbus Bd. of Educ. v. Penick, 443 U.S. 449, 465 & n.13 (1979), and Keyes v. School Dist. No. 1, Denver, 413 U.S. 189, 211 & n.17 (1973)) (court must order affirmative remedy where school board's conduct "create[d] or contribute[d] to" racial identifiability of schools). We adhere to the most common-sense meaning of "vestige": it is a condition or occurrence causally related to the former de jure system of segregation. Because a school system's duty to eliminate such vestiges is restricted by the availability of practicable measures for doing so, see Freeman, 503 U.S. at 492, it is also incumbent on us to consider prac- ticability. In determining the practicability of further measures, the district court must look to numerous indicia of the system's operation. Practicability depends on the feasibility of the proposed method, from both a financial and an administrative perspective. Cf. id. at 481-83, 493-97. Whether a measure is practicable also depends on whether it is "directed to curing the effects of the specific violation," and whether it is likely to do so. Id. at 497. Our duty, in reviewing Judge Potter's decision, see Capacchione, 57 F. Supp. 2d at 228, is clear. We must examine each Green factor and ascertain whether unitary status has been achieved with respect to any or all of them. Because the district court declared the entire CMS school system to have achieved unitary status, we must assess, with respect to each Green factor, whether the Board has complied, in good faith, with the desegregation decree and whether the vestiges of segregation have been eliminated to the extent practicable. See Freeman, 503 U.S. at 492 (citing Dowell, 498 U.S. at 249-50). If the school system has not achieved unitary status in its entirety, then, 23 consistent with Freeman, we also must weigh the degree of interre- latedness existing between the various Green factors. B. By way of introduction to our analysis of this case, we first address a fundamental flaw in the district court's proceedings -- a flaw aris- ing from the district court's failure to give any consideration to a remedial plan sought to be admitted as evidence by CMS. Following the filing of the Capacchione plaintiffs' Complaint in Intervention, the Board undertook to produce a comprehensive analysis of whether vestiges of de jure segregation existed in CMS and whether any such vestiges could be practicably remedied. The Board analyzed available data and identified several vestiges remaining; then, in line with the mandate of Freeman, the Superintendent of CMS developed a plan containing practicable remedial steps. The Board independently reviewed this plan and, on March 30, 1999, adopted the "Charlotte- Mecklenburg Schools' Remedial Plan to Address the Remaining Ves- tiges of Segregation" (the "Plan" or "Remedial Plan"). J.A. 11029. Consistent with pretrial deadlines, CMS filed the Remedial Plan with the district court as a potential exhibit at trial. J.A. 11028. At the pretrial conference conducted on April 13, 1999, the Capacchione plaintiffs moved in limine to exclude the Remedial Plan. In essence, the Capacchione plaintiffs maintained that the trial had been bifur- cated into two phases and that only unitary status was at issue in the first phase. They further maintained that the Remedial Plan contained proposed remedies that could only be implemented if CMS was deter- mined not to have achieved unitary status. Because the unitary status question had not yet been resolved, they claimed that the Remedial Plan (which the Capacchione plaintiffs characterized as a damages report) was irrelevant. In opposing exclusion of the Remedial Plan, CMS and the Swann plaintiffs relied on the Supreme Court's Freeman analysis. J.A. 1421. Specifically, they asserted that each unitary status determination encompassed in the first phase of the trial turned on "whether the ves- tiges have been remedied to the extent practicable." Id. (emphasis added). The Remedial Plan, they claimed, was not merely relevant, 24 but crucial, to establishing both the existence of vestiges of segrega- tion and the practicability of remedial measures. Judge Potter responded with two rulings. First, Judge Potter explained in assessing whether CMS had achieved unitary status that he believed Freeman required him to consider just one thing: "only . . . what CMS has done, not what it may do in the future." See Order of April 14, 1999 at 4. Second, based on this understanding of Free- man and the unitary status test, Judge Potter concluded that the Reme- dial Plan was irrelevant: "If the Court later determines that additional remedial measures are needed, it may consider the plan. Until that time comes, however, the Court will not get mired in the complex details and mechanics of a proposed plan." Id. at 5.6 6 We believe Judge Potter erred in both of these rulings. First, he misapprehended Freeman and its test for unitary status. At the outset, Freeman explicitly rejects, as a matter of law, the very analysis adopted by the district court. That is, under Freeman, a district court must consider (1) compliance with prior orders (i.e., "what CMS has done"), and (2) whether vestiges have been eliminated to the extent practicable (i.e., "what [CMS] may do in the future"). See Freeman, 503 U.S. at 491; see also Order of April 14, 1999 at 4. By construing Freeman's unitary status test to include the former ("what CMS has done") but not the latter ("what [CMS] may do in the future"), Judge Potter erred as a matter of law. _________________________________________________________________ 6 Judge Potter also chided CMS for proffering the Remedial Plan "after the deadline for fact discovery and expert witness discovery had expired." On the contrary, CMS fulfilled all of its duties under the fed- eral rules, appropriately supplementing its responses to discovery requests as soon as the Remedial Plan had been adopted. Furthermore, a more precipitant proposal could not have incorporated the various expert perspectives developed during discovery. A similar plan submit- ted earlier in the course of the litigation necessarily would have been based largely on speculation and supposition, and therefore would have been far less useful and pertinent. As it was, the Plan was tendered in advance of the non-jury trial, and, of great significance, almost five months before the district court issued its decision. Neither the court nor the parties could have been inconvenienced by the necessary timing of the Remedial Plan's submission. 25 The Remedial Plan directly addresses the latter inquiry, and it does so in an apt, informed manner, relying on the considered opinions of highly capable professionals retained to analyze the latest available data. In other words, the district court's second reason for excluding the Plan -- relevancy -- also fails to withstand scrutiny.77 There is no doubt that Judge Potter had wide discretion on this issue, but rele- vancy is a fluid concept under the Federal Rules of Evidence. See Fed. R. Evid. 401 (defining relevant evidence as "having any ten- dency to make the existence" of any material fact"more probable or less probable than it would be without the evidence"). Consequently, relevancy typically presents a rather low barrier to admissibility. See, e.g., United States v. Van Metre, 150 F.3d 339, 349 (4th Cir. 1998) (citing United States v. Powers, 59 F.3d 1460, 1465 (4th Cir. 1995)). However, we need not rely on the minimal threshold encompassed in the test for relevancy because this Remedial Plan would be relevant under any reasonable test. The Remedial Plan identified record evi- dence (including the deposition testimony of several experts) support- ing the Board's belief that vestiges of de jure segregation in CMS remain apparent in (1) faculty assignment and quality, (2) physical facilities and the allocation of instructional resources, (3) student achievement, and (4) student assignment. More importantly, the Remedial Plan detailed specific steps that the Board proposed to undertake over the course of the ensuing five years "with a goal of achieving unitary status at that time." J.A. 11029. _________________________________________________________________ 7 Taking the district court at its word that the only question before it initially was the extent of the Board's compliance with the prior desegre- gation orders, the Remedial Plan was nonetheless highly relevant for even that purpose. The ease with which some of the proposed Plan reme- dies could be realized, e.g., merely distributing available funding to address the stark disparity in basic resources such as instructional materi- als and media centers, see J.A. 11040, strongly suggests that the Board had not fully implemented the long-standing dictates of the prior orders. The court nonetheless observed that "while the goal of perfect compli- ance with court orders has remained elusive, no evidence has been presented that school authorities were guilty of easily correctable errors." Capacchione, 57 F. Supp. 2d at 283. To the contrary, the Plan thoroughly documented the Board's failings and the facility with which they could be rectified. The district court simply chose to ignore this highly relevant evidence. 26 Without a doubt, federal courts possess the final word in deciding whether a particular school system is operating within the parameters of the Constitution. Appreciable weight must be given, however, to the views of those selected by the community to administer the sys- tem. See Dowell, 498 U.S. at 248 (noting specialized knowledge pos- sessed by local school officials).8 8 In refusing to consider the Plan, the district court erroneously failed to accord the Board's official position any weight, much less the respect that it was due. That the district court so completely disregarded this crucial evi- dence is telling. Nonetheless, we have carefully examined each con- clusion below, ever mindful of the deference accorded the factfinder. The manifest importance of this case (quite apart from the substantial time and energy invested by the parties and the court below) demands that we carefully explain the myriad aspects of our decision today. We now embark upon that task. 1. Student Assignment Of all the Green factors, the most fundamental is the degree of racial imbalance in student assignment. Freeman, 503 U.S. at 474. Uniformity in the racial composition of a given school was the hall- mark of official discrimination, "for under the former de jure regimes racial exclusion was both the means and the end of a policy motivated by disparagement of, or hostility towards, the disfavored race." Id. Court-ordered desegregation was designed to meet the enemy head- on; the long-term stability of attempts at racial balancing in student assignment is often seen as the most conspicuous indication of the courts' success (or lack thereof) in combating the underlying societal evil. _________________________________________________________________ 8 Though we need not grant CMS the same deference afforded the pro- mulgations and adjudications of a federal administrative agency, the for- mal declarations of its governing Board "do constitute a body of experience and informed judgment to which courts and litigants may properly resort for guidance." Ritter v. Cecil County Office of Housing & Community Dev., 33 F.3d 323, 328 (4th Cir. 1994) (quoting Skidmore v. Swift & Co., 323 U.S. 134, 140 (1944)). 27 We must now determine whether present racial isolation in CMS may be a vestige of the former dual system, and, if so, whether there are practicable measures CMS could take to reduce or eliminate that isolation. In doing so, we are bound to focus particularly on the Board's record of compliance with the district court's desegregation orders. See id. at 492 (citing Dowell). Because significant and grow- ing racial imbalances in student assignment do exist in CMS, because the Board for decades has failed to comply with certain specific decrees of the district court (particularly regarding the siting of new schools), because these failures may have contributed to current racial isolation, and because future compliance might practicably reduce this racial isolation, we must vacate the district court's finding that CMS has achieved unitary status with respect to student assignment. a. In the wake of the 1970 desegregation order, virtually all of the schools in CMS operated in racial balance for a considerable time. By 1998-99 however, nearly thirty percent of the schools in the system had become racially identifiable.9 9 Of the 126 schools included in the CMS desegregation plan, twenty-three are identifiably black and thir- teen more are identifiably white. J.A. 11587. Further, virtually all of the identifiably black schools are located in either the inner city or in the immediate northwest-to-northeast suburbs, the areas of Mecklen- burg County with the highest concentration of African-Americans. In stark contrast, all thirteen of the identifiably white schools are found in the extreme northern and southern areas of the county, both of which (and particularly the latter) have seen dramatic increases in white population during the past thirty years. The trend in CMS toward resegregation of its schools has accelerated markedly since the move to de-emphasize satellite zones and mandatory busing in 1992. In the last seven years, the number of CMS African-American stu- _________________________________________________________________ 9 Judge McMillan's final desegregation order mandated, inter alia, that no school should become "racially identifiable." Swann, 311 F. Supp. at 268. Judge Potter interpreted the phrase synonymously with "racially imbalanced," which, as noted within, describes a school with an African- American student population deviating more than fifteen points in either direction from the county-wide norm. See Cappachione, 57 F. Supp. 2d at 246. 28 dents who attend racially identifiable schools (now almost three in ten) has risen fifty percent. J.A. 9589. Indisputably, from 1981 until 1997, the CMS school system went through significant demographic changes. For example, the total pop- ulation of Mecklenburg County has grown from 354,656 in 1970 to 613,310 in 1997. J.A. 16247. Almost 100,000 children attend CMS, making it the twenty-third largest school system in the country. J.A. 7107. During the period from 1970 to 1997, the black school-age pop- ulation (ages 5 through 17) in the county has increased by approxi- mately 10,000. J.A. 16247. Over the same period, the corresponding white school-age population has decreased by approximately 3,000, id., and by 1997, African-Americans comprised 34 percent of the county's school-age population, the total of which numbered approxi- mately 108,600. Evidence before the district court revealed that, since 1970, the growing African-American population has migrated out- ward from the inner city into formerly white suburbs. In turn, many white citizens who formerly populated the city's periphery have moved even farther into the county's outlying reaches. Though parts of the county have become more integrated as the result of these shifts, a disproportionately large number of African-Americans still reside in contiguous clusters generally north and west of the down- town area. The primary issue we must address is whether the thirty-six racially identifiable schools in CMS represent a vestige of segregation -- that is, whether the present racial isolation is causally related to the prior system of de jure segregation. The Swann plaintiffs argue, and CMS agrees, that current racial isolation, like the racial isolation of the 1960s and 1970s, results both from past inequities that, to some extent, have persisted to this day, and from the Board's failure to comply with certain specific directives in the remedial decrees in this case. Because CMS has not previously been adjudged to have achieved unitary status in student assignment, we are bound under Swann to presume that the current racial imbalance in the school population constitutes a continuing vestige of segregation. 402 U.S. at 26. The Capacchione plaintiffs have the burden of showing that the present existence of predominantly one-race schools in CMS "is not the result 29 of present or past discriminatory action." Id.; see also Riddick, 784 F.2d at 535. Our unwillingness to affirm the conclusion that CMS is unitary with respect to student assignment centers on the Board's failure to comply with court orders regarding selection of sites for the construc- tion of new schools. The role of school siting in achieving sustainable desegregation should not be underestimated. In fact, the importance of site selection has been apparent since the early stages of this case. As the Supreme Court explained in 1971: In the past [site selection] choices . . . have been used as a potent weapon for creating or maintaining a state-segregated school system. . . . [S]chool authorities have sometimes, since Brown, closed schools which appeared likely to become racially mixed through changes in neighborhood residential patterns. This was sometimes accompanied by building new schools in the areas of white suburban expan- sion farthest from Negro population centers in order to maintain the separation of the races with a minimum depar- ture from the formal principles of "neighborhood zoning." Such a policy does more than simply influence the short-run composition of the student body of a new school. It may well promote segregated residential patterns which, when combined with "neighborhood zoning," further lock the school system into the mold of separation of the races. . . . In ascertaining the existence of legally imposed school seg- regation, the existence of a pattern of school construction and abandonment is thus a factor of great weight. Swann, 402 U.S. at 21. Subsequent to the Supreme Court's decision in Swann, Judge McMillan specifically ordered that site selection for new schools could not "be predicated on population trends alone." 379 F. Supp. at 1107. New schools were "to be built where they can readily serve both races." Id. In the 1979 Martin decision,10 10 Judge McMillan _________________________________________________________________ 10 Judge Potter incorrectly declared that "Martin was not a unitary sta- tus hearing[.]" Capacchione, 57 F. Supp. 2d at 250. In fact, as the accom- 30 devoted an entire section of his opinion to demonstrating that "con- struction, location and closing of school buildings continue to pro- mote segregation." 475 F. Supp. at 1329. Judge McMillan explained that "[t]he location of schools plays a large if not determinative role in . . . insuring that any given assignment and feeder plan will provide meaningful desegregation, rather than just the predictably short lived appearance of desegregation." Id. at 1332. In the years since this decree was issued, CMS has built twenty- five of twenty-seven new schools in predominantly white suburban communities. In the mid-1980s, CMS adopted a formal policy of building "midpoint" schools -- schools located midway between black and white population centers. There is little evidence, however, to suggest that CMS faithfully adhered to this policy. Rather, record evidence strongly indicates that the policy influenced the site selec- tion for, at most, four of the twenty-seven new schools. See J.A. 15404-06. Meanwhile, as we discuss infra, there is substantial evi- dence that CMS has allowed many of its older school facilities in the city -- schools attended in disproportionate numbers by African- American students -- to fall into a state of disrepair. The Board's record of building the great majority of its new schools on the predominantly white suburban fringe of the county supports two possible conclusions. On one hand, CMS could have been responding to demographic reality -- a demand for new class- rooms in areas of high population growth (although we note that the number of white students in CMS has decreased since 1970, while the black student population has greatly increased). On the other hand, the Board's pattern of school construction could have facilitated or even hastened white flight to the suburbs. As the Supreme Court explained in Swann, "[p]eople gravitate toward school facilities, just as schools are located in response to the needs of people. The location of schools may thus influence the patterns of residential development _________________________________________________________________ panying text indicates, the white parents in Martin contended, as the Capacchione plaintiffs do today, that CMS had achieved unitary status. Intervening African-American parents, like those herein, maintained to the contrary. In actuality, there is little difference between today's case and Martin, and Judge McMillan's findings in the latter are as binding on the parties as any others made in the course of this litigation. 31 of a metropolitan area and have important impact on composition of inner-city neighborhoods." 402 U.S. at 20-21. The Board's school sit- ing policies could well evidence its lack of political will in the face of pressure to abandon desegregative policies -- pressure from fami- lies who "are concerned about the racial composition of a prospective school and [who] will make residential decisions accordingly." Free- man, 503 U.S. at 513 (Blackmun, J., concurring). There is certainly no evidence that CMS has intentionally sought, through its school siting policies, to "lock the school system into the mold of separation of the races" in the way that the Supreme Court described in Swann. But the actual choices the Board has made with regard to school siting may in fact be quite similar to the "pattern of school construction and abandonment" described by the Court, with the actual effect that the Court feared of "lock[ing] the school system" into a condition of racial isolation. 402 U.S. at 21. We cannot con- clude, at least in the absence of further fact-finding, that CMS, in choosing sites for new schools, has pursued "meaningful desegrega- tion, rather than just the predictably short lived appearance of deseg- regation." 475 F. Supp. at 1332. Rather, the Board's practice of siting new schools such that they could not reasonably be expected to serve a racially balanced student population and Judge McMillan's determination that this practice, in the past, represented the school system's failure to eliminate the ves- tiges of segregation, together raise a strong inference that those ves- tiges remain today. When this inference is viewed in combination with the burden borne by the Capacchione plaintiffs to show that cur- rent racial imbalances have no causal link to past discrimination, we are compelled to conclude that a remand to the district court is required. Although we defer to a district court's findings of fact unless clearly erroneous, Judge Potter's error here came in his application of the legal standard to the evidence regarding the Board's school siting policies. Judge Potter found that (1) CMS had not discriminated on the basis of race in choosing sites for new schools and that (2) CMS had incorporated racial diversity as one of its factors in site selection. Even assuming arguendo that both findings are not clearly erroneous, neither is sufficient to support the legal conclusion that in siting new 32 schools CMS acted in compliance with the governing court orders and Constitution to eliminate the vestiges of segregation to the extent practicable. "To fulfill this duty, school officials are obligated not only to avoid any official action that has the effect of perpetuating or reestablishing a dual school system, but also to render decisions that further desegre- gation and help to eliminate the effects of the previous dual school system." Harris v. Crenshaw County Bd. of Educ., 968 F.2d 1090, 1095 (11th Cir. 1992) (citing Pitts v. Freeman , 755 F.2d 1423, 1427 (11th Cir. 1985)). Therefore, CMS had to do more than merely select sites for new schools on a nondiscriminatory basis. It had to do more, too, than simply give some consideration to "diversity" in its selection of sites. To the extent practicable, CMS had to site new schools "where they can readily serve both races." 379 F. Supp. at 1107; see also Swann, 402 U.S. at 21; Martin, 475 F. Supp. at 1329-32. Judge Potter never found that CMS had met this standard, and as outlined within, there is substantial record evidence that CMS did not do so. In accordance with Swann, the burden is on the Capacchione plain- tiffs to prove that vestiges of past discrimination do not remain, or that nothing can practicably be done to remedy them. We note that Judge McMillan, in his last published decision in this case, clearly evidenced his understanding both that CMS had not done all that it could do in the area of school siting and that future school siting deci- sions could practicably advance the process of desegregation. On remand, it is thus incumbent on the Capacchione plaintiffs to demon- strate that conditions in Charlotte and Mecklenburg County have changed sufficiently such that school siting no longer represents a practicable opportunity to eliminate the vestiges of segregation. The Swann plaintiffs have identified additional areas in which CMS has fallen short of its obligations under the court orders. For the life of the desegregation orders, CMS has consistently placed the heaviest burden of mandatory busing on African-American students. Currently, 80% of those students who ride the bus as a result of a mandatory assignment are African-American. J.A. 11515. Judge McMillan repeatedly ordered CMS to distribute this burden more fairly. See 475 F. Supp. at 1339-40; 379 F. Supp. at 1103-04; 362 F. Supp. at 1232-33. Yet, CMS has utterly failed to do so. In addition, 33 CMS has never developed an effective system for monitoring student transfers to ensure that the overall effect of such transfers is not to increase the racial imbalance in the system as a whole. Again, this represents a failure to comply with the explicit instructions of the dis- trict court. See 475 F. Supp. at 1337-38; 379 F. Supp. at 1103-04; 362 F. Supp. at 1229-30. We are troubled by these failings on the part of CMS. They provide additional support for a conclusion that, in the face of political pressure, CMS has not done all that it could do to eliminate the vestiges of segregation. Finally, the Board has itself taken the remarkable step of admitting its noncompliance with prior orders in this case. A school board's frank acquiescence in a position inuring to its detriment (in this case, the potential of ongoing judicial intervention), if not treated as con- clusive, should at least be considered with the utmost gravity. Under these circumstances, we have no difficulty in determining that the dis- trict court's conclusion that the Board's level of compliance was "full and satisfactory" must be vacated. b. If the vestiges of official discrimination have indeed been elimi- nated to the extent practicable with respect to student assignment, then there is little reason to prolong court supervision. In light of the district court's failure, however, to recognize the Board's continuing noncompliance with respect to student assignment -- administered as recently as twenty years ago in a manner reinforcing the once-official notion that African-Americans are inferior -- we have no confidence in the court's ultimate finding that these vestiges have now disap- peared. We are therefore obliged to vacate the portion of the judgment below relating to student assignment and remand for further proceed- ings. On remand, the district court must first determine whether, since Judge McMillan's decision in Martin, CMS has fulfilled its constitu- tional and court-imposed obligations with regard to site selection for new schools. If CMS has failed to fulfill its obligations, the district court must then determine whether this failure contributed to the pres- ent condition of racial isolation in the school system. Finally, if CMS did fail to live up to its constitutional and judicially decreed obliga- 34 tions, and if that failure did contribute to the present racial imbal- ances, then the court must determine if proper site selection is a practicable remedy for the lingering effects of the Board's past dis- criminatory practices. If not, then the district court should relinquish control over student assignment: there is nothing further that CMS can practicably do to eliminate the vestiges of the prior de jure sys- tem. If, however, proper sites can be found, then the district court should retain control over student assignment. The court might decide, for example, that most or all new schools constructed over the next sev- eral years be located proximate to the inner city or in midpoint areas already integrated residentially. Conversely, the district court may conclude that more flexibility is required because of real estate costs, crushing demand in the suburban fringes, or for some other sufficient reason. The court should also consider the efficacy of the Board's Remedial Plan as a limited term remedy for the racial isolation that would otherwise continue to exist until the Board's newly redirected school siting policies can begin to take effect.11 11 Of course, some reasons will not be sufficient to deny African- American students a remedy, should corrective action be deemed jus- tified. For example, political pressure and perceived resistance to change by certain groups in the community will not suffice. Addition- ally, logistical barriers merely making "difficult" the transport inward of outlying white students will likewise, if reasonably surmountable, not be enough. Cf. Capacchione, 57 F. Supp. 2d at 253 (district _________________________________________________________________ 11 The strategies described in the Remedial Plan may be of particular help to the court in deciding whether practicable measures are available. The Plan proposes, among other things, to divide Mecklenburg County into three to five demographically similar "clusters," within which stu- dents may choose to attend any school, magnet or otherwise. Where the demand for a given school exceeds the available room, spots would be assigned by lottery based on factors such as proximity, sibling atten- dance, and racial, ethnic, and economic diversity. The Plan also outlines a formal mechanism to disseminate information regarding the enrollment process, and it provides that the Board will work with the business com- munity and local government to secure subsidies for disadvantaged fami- lies wishing to relocate to areas in which low-cost housing is scarce. See J.A. 11053-59. 35 court's observation that "transport[ing] white students in from satel- lite zones . . . is difficult given the rush hour traffic patterns"). Although what is "practicable" need not extend to all that is "possi- ble," rectifying the grievous constitutional wrongs of the past surely justifies reaching beyond the "difficult" or purely "problematic." 2. Physical Facilities After describing how CMS has allocated its physical facilities and resources among its students, Judge Potter concluded that "the Swann plaintiffs have failed to overcome the Court's previous findings on facilities by establishing the requisite discriminatory intent and causa- tion." Id. at 267. Judge Potter's mention of "previous findings" refers to excerpts from various opinions and orders authored by Judge McMillan: April 1969 -- "No racial discrimination or inequality is found in the . . . . quality of the school buildings and equip- ment. . . . Schools described by witnesses as `white' ranged well up and down on both sides of [the average per-pupil expenditure], and schools described by witnesses as `black' showed a similar variation." 300 F. Supp. at 1366. August 1969 -- "The defendants contended and the court found in its April 23, 1969 order that facilities and teachers in the various black schools were not measurably inferior to those in the various white schools. It is too late now to expect the court to proceed upon an opposite assumption." 306 F. Supp. at 1298. October 1971 -- "[T]he formerly black schools are not shown nor suggested to be inferior in faculty, plant, equip- ment or program." 334 F. Supp. at 625. Toward the close of the prior proceedings in 1975 (and consistent with the above), Judge McMillan awarded attorney's fees to the Swann plaintiffs as prevailing parties, "[e]xcept for the refusal of the court to find in the plaintiffs' favor . . . regarding adequacy of physi- cal plants and equipment and teacher quality." Swann, 66 F.R.D. at 484. 36 Judge Potter acknowledged that no court "ha[d] [ ]ever granted uni- tary status to CMS, nor . . . partially withdrawn supervision as to facilities or any other Green factor." Capacchione, 57 F. Supp. 2d at 262. The court nevertheless relied on the above 1969 and 1971 find- ings to release the Capacchione plaintiffs from their burden of prov- ing CMS unitary with respect to facilities, stating that to proceed otherwise would "defy logic." Id. at 263. Judge Potter thus accepted the premise that Judge McMillan's 1969 and 1971 findings "consti- tute collateral estoppel and law of the case" regarding facilities, "thereby shifting the burden to CMS and the Swann plaintiffs to show discriminatory intent." Id. at 262. We hold the district court's burden-shifting analysis to constitute an error of law. Once the existence of an unlawful dual school system has been established and court supervision begun, it is presumed that racial disparities arising during the period of intervention "are caus- ally related to prior segregation." School Bd. of the City of Richmond v. Baliles, 829 F.2d 1308, 1311 (4th Cir. 1987). Following the imposi- tion of judicial control, a party seeking to end the status quo bears the burden of overcoming the presumption of causation. If this burden is met and the school system is declared to have achieved unitary status as to the particular factor at issue, the presumption ends. Id. Gener- ally, in any subsequent proceeding involving new allegations of dis- parate treatment, the complaining party must show purposeful discrimination. Riddick, 784 F.2d at 537 (concluding that Swann and its progeny require proof of "discriminatory intent on the part of the school board of a unitary school system" in order to resume court super- vision).12 12 To be sure, the absence heretofore of any finding to the contrary would have been an important consideration in determining whether the Capacchione plaintiffs had proved CMS to have achieved unitary status with respect to facilities. However, that Judge McMillan did not intend his initial observations regarding facilities to be construed as a finding of unitary status is obvious from his subsequent actions. In _________________________________________________________________ 12 However, if a district court retains jurisdiction over one or more Green factors, it may, upon a proper showing, reassert control over a fac- tor previously adjudged to have attained unitary status. Freeman, 503 U.S. at 508-09 (Souter, J., concurring). 37 1973, Judge McMillan assumed control over facilities and resources, found inequities, and ordered CMS to remedy those disparities. See Swann, 362 F. Supp. at 1235 (finding Double Oaks Elementary access road still undeveloped two years after court's identification of the problem -- "No $80,000,000 budget is so powerless."); id. (finding Double Oaks library not restored to standards several years after fire); id. at 1238 (ordering athletic facilities at West Charlotte High School immediately upgraded to level comparable with other schools in the county). We must conclude that the Board has been subject to the court's jurisdiction as to its facilities since at least 1973. See Dowell, 498 U.S. at 246 (school boards entitled to a "rather precise statement" terminating a desegregation order). The asserted lack of a prior adverse finding should not have been determinative of the issue, especially as the district court in 1969 was not focusing on a school system suddenly thrust into the judicial arena, but was instead examining one that had been subject to court supervision for nearly four years. Between the commencement of the initial Swann lawsuit in 1965 and the district court's first mention of the facilities issue in April 1969, CMS closed sixteen black schools. The Board's en masse action gives rise to an almost undeniable infer- ence that these schools were shut down because they were inferior, and the timing also suggests strongly that the closures were prompted by the judicial proceedings then underway. Viewed in context, the most plausible conclusion is that the puta- tive equality mentioned by the district court in 1969 and 1971 was actually an endorsement of the steps that had been taken by the Board to remedy the inequities in facilities. In any event, CMS could not be said to have achieved unitary status absent a finding by the lower court that the Board had "eliminated the vestiges of its prior discrimi- nation," embodied in an "adjudicat[ion] . . . through the proper judi- cial procedures." Georgia State Conference of Branches of NAACP v. Georgia, 775 F.2d 1403, 1413 n.12 (11th Cir. 1985), quoted in Dowell, 498 U.S. at 245 (noting distinction between school systems operating in an unitary fashion and those that have achieved unitary status, and observing that the former "could be called unitary and nev- ertheless still contain vestiges of past discrimination"). 38 Thirty-five years have passed since the Board first acted to equalize its facilities, yet serious questions remain as to whether it has finally realized that goal. Dr. Dwayne E. Gardner, an impressively qualified educational planner and consultant, compiled an exhaustive report for the Board in which he evaluated the suitability of its school facilities.13 13 Dr. Gardner examined and personally visited more than half of the schools in CMS (including all of the high schools), analyzing a host of factors affecting educational quality. For the purposes of his study, Dr. Gardner divided the subject schools into three groups: (1) all imbalanced-black schools; (2) all racially balanced schools in imbalanced-black census tracts; and (3) each remaining high school, along with a set of elementary and middle schools randomly selected from the remaining schools and approximately equal in number to those already included within the first two groups. Each school in the study was assigned a composite score from 0- 100, indicating its worthiness. Schools scoring 44 or lower were, in Dr. Gardner's opinion, so deficient as to merit replacement, while those with scores between 45-59 were classified as needing "major improvements." Any school that scored 60 or above was "considered to have the ability to serve the educational program adequately." J.A. 12174. The results of Dr. Gardner's study are troubling. The average score for the forty Group 3 schools (racially balanced or imbalanced-white in predominantly white or balanced areas) was 61.7. Although the Group 3 data indicate a situation that is far from ideal, the ten Group 2 schools (racially balanced in predominantly black areas) fared much worse, with an average score of 56.3. The scores of the twenty-three Group 1 schools (imbalanced-black) were worse still, averaging just 53.3.14 14 At trial, Dr. Gardner confirmed that the disparities apparent _________________________________________________________________ 13 The district court found that the expert called on behalf of the Capac- chione plaintiffs, Dr. David J. Armor, could offer no reliable testimony on the subject. See Capacchione, 57 F. Supp. 2d at 264. 14 It has been famously said (by either Mark Twain or, earlier, Benja- min Disraeli, depending on one's source), "There are three kinds of lies -- lies, damned lies and statistics." A common difficulty in dealing with statistics is illustrated by the district court's analysis of Dr. Gardner's 39 from the above numbers were "substantial" with respect to the facili- ties generally available to white and African-American children attending CMS. J.A. 6196-99. The anecdotal accounts of a number of witnesses effectively cor- roborated Dr. Gardner's conclusions. See, e.g., J.A. 4992 (testimony of Board member Pamela R. Mange) (schools with "more severe" problems tended to be predominantly black); J.A. 4769 (testimony of Annelle Houk) ("[T]he schools that were in the worst repair and had the poorest supply of resources . . . were on the west side and they were predominantly populated by black students."). John A. Kramer, co-chair of an advisory task force created by the Board, made formal visits to several CMS schools in 1997. Among the locales on Mr. _________________________________________________________________ study. The court first noted that, of the four schools scoring in the lowest category, two were in Group 1 and two were in Group 3. Capacchione, 57 F. Supp. 2d at 264-65. Next, the court observed that the two highest ratings accorded elementary schools were again split between Groups 1 and 3. Id. at 265. Based on this selective culling of the data, the lower court concluded that "the results of Dr. Gardner's analysis do not show disparities along racial lines." Id. at 264. The forest that is CMS is not sufficiently mapped by the documenta- tion of a few trees. We could accurately say, for example, that omission from Group 1 of the brand-new elementary school referred to by the dis- trict court as having one of the highest ratings would lower the Group 1 average by more than a full point. Or we could state without error that seven of the twenty-three Group 1 schools (more than 30 percent) scored below 50, while only five of the forty Group 3 schools (12.5 percent) scored similarly. Indeed, we note that none of the Group 1 high schools scored higher than 46, yet all those in Groups 2 and 3 scored at 50 or above. Of course, one would rightly view this latter declaration with some skepticism once it became known that there are but fourteen high schools in CMS, only two of which were included by Dr. Gardner in Group 1. The pick-and-choose method gets us nowhere. The value of Dr. Gard- ner's research lies in the general conclusions that can be drawn from the entirety of the data. The most obvious conclusion is that, as a general matter, imbalanced-black schools in CMS are in worse shape than those attended by larger proportions of white students. Once we accept that premise, the lone remaining question of any significance is "Why?" 40 Kramer's itinerary were Elizabeth Lane Elementary, a predominantly white school located in a prosperous suburban area of the county, and Shamrock Gardens Elementary, a downtown school with an African- American student population exceeding sixty percent. Mr. Kramer's descriptions of his visits contrasted sharply: [T]o compare Elizabeth Lane Elementary as an example, which is a relatively new school located in Matthews, I walked into that school, I was overwhelmed because I had never set foot in a school that was like that before. It was clean, it was light and airy, it was a beautiful facility. . . . My overwhelming feeling was, wow, I wish my kids could go to this school. And another observation that was very clear was that when I looked at the student body, it was vir- tually all white students, obviously, affluent, happy kids having a great time. On the other hand, my experience, for example, at Sham- rock Gardens was shocking by comparison. I had never vis- ited either one of these schools before, but to visit that school which is in the inner city, the students are predomi- nantly black students, it reminded me of a rundown 1950s motel. There was literally no access to the rooms except by outer walkways that were covered by rusted, dilapidated overhead fixtures. . . . They were using closets and things to teach children in. The carpets were stained and thread- bare. . . . It just didn't feel clean, it didn't feel good. And I can honestly say that as a parent, my heartfelt reaction was relief that my children didn't have to go to school there. J.A. 6098-99. Even those Board members who voted to pursue a determination of unitary status before the district court admitted that disparity in facilities was a problem within CMS. J.A. 1817, 1820 (testimony of James H. Puckett); J.A. 1918-19 (testimony of John W. Lassiter); J.A. 2095-96 (testimony of Lindalyn Kakadelis). Although it seems reasonably clear that a racial disparity in facili- ties exists in CMS, its cause is somewhat less apparent. The Capacch- ione plaintiffs maintain that no discrepancies exist in CMS facilities, and even if they do, such discrepancies are totally benign in origin. 41 Had the Capacchione plaintiffs proved their theory, we would be con- strained to affirm the district court's conclusion that unitary status has been achieved with respect to the facilities factor. The district court, however, required the Capacchione plaintiffs to prove nothing; it instead erroneously placed the burden on CMS and the Swann plain- tiffs to affirmatively show that the present inequities in facilities are a vestige of official discrimination, i.e., causally related to the prior de jure system of segregation. Capacchione, 57 F. Supp. 2d at 267. The district court erred as a matter of law in foreclosing the devel- opment of evidence relevant to a proper vestige analysis. We must therefore remand this portion of the case to permit the parties and the district court to elicit the additional facts necessary to fully consider the question of causation with respect to the current racial inequities in facilities. Because CMS has not been previously adjudged to have attained unitary status, the Capacchione plaintiffs are charged on remand with the burden of demonstrating that the vestiges of past de jure racial discrimination in the context of the school system's facili- ties have been eliminated "root and branch" to the extent practicable.15 15 _________________________________________________________________ 15 The district court made no findings as to whether practicable reme- dies exist with respect to facilities. In light of the court's refusal to con- sider the Board's proposed five-year Remedial Plan, we cannot determine in the first instance whether practicable remedies to the current disparities exist. We therefore remand to the district court for develop- ment on this point. We note, however, that the Remedial Plan specifi- cally identifies disparities associated with race in baseline needs for schools' instructional materials and media centers, and the lack of any standardized criteria to evaluate the adequacy of these resources. J.A. 11037-38. The Plan proposes to achieve uniformity in resources across schools by imbalanced allocations that reflect the schools' current resource gaps and imbalances. J.A. 11038-40. Likewise, the Remedial Plan identifies disparities associated with race in the instructional facili- ties, and proposes building replacements or renovating existing facilities for sixteen schools that are either racially identifiable as black or are located in a predominantly black census tract. J.A. 11041-42. Uniform building maintenance standards and procedures are proposed. J.A. 11043. Monitoring, evaluation, and development of appropriate criteria for evaluation are also proposed to maintain equity across the school sys- tem's resources and facilities. J.A. 11038-40, 11042-43. 42 3. Transportation School bus transportation was at the epicenter of the original Swann litigation, specifically the degree to which involuntary busing could be used to implement a remedial desegregation decree. The Supreme Court in Swann, of course, approved busing as a "normal and accepted tool of educational policy," 402 U.S. at 29, at least to the extent that the rigors of time and distance would pose little risk to the affected students' health or to the educational process as a whole. See id. at 30-31. In the intervening twenty-nine years, CMS has taken the Court's license to heart; during the 1998-99 school year, five of every six students in the school system rode a school bus. Upon review of the Green factor of transportation, Judge Potter concluded that "a court may grant unitary status when transportation is provided on a non-discriminatory basis." 57 F. Supp. 2d at 267. In other words, according to the district court, a school system achieves unitary status with respect to transportation once it provides access to transportation non-discriminatorily to black and white children. Because CMS provides all children, regardless of race, access to transportation, Judge Potter concluded that CMS had achieved unitary status with respect to this Green factor. We must be mindful of the Supreme Court's command to consider the interrelatedness of the various Green factors. See Freeman, 503 U.S. at 491 (court must consider "whether retention of judicial control is necessary or practicable to achieve compliance with the decree in other facets of the school system"). In this context, we can only con- clude that the Green factor of transportation is so inextricably inter- twined with the Green factors of student assignment and facilities that our vacatur on those issues also mandates vacatur on the factor of trans- portation.16 16 _________________________________________________________________ 16 Pursuant to Freeman, the district court accepted the invitation of the Board and the Swann plaintiffs to consider whether vestiges of official discrimination remain concerning the ancillary factors of student achievement and student discipline. The court found in the negative, con- cluding that CMS had attained unitary status in both areas. With respect to the ancillary factor of student achievement, we must vacate Judge Potter's holding that unitary status had been achieved, and 43 The Swann plaintiffs maintain and offer substantial record evidence that the burdens of busing for desegregation purposes are being borne disproportionately and unfairly by African-American children. Brief of Appellants at 31-32, 33-35; see Swann, 306 F. Supp. at 1298 (dis- trict court commenting in initial stages of remediation that it did not intend "to endorse or approve any future plan which puts the burden of desegregation primarily upon one race"). Eighty percent of stu- dents who currently ride the bus as a result of a mandatory assignment are African-American. Judge Potter rejected any consideration of this evidence, holding that a school district has achieved unitary status with respect to transportation as soon as it is provided on a race- neutral basis. The evidence, however, demonstrates the close interre- lationship of transportation with student assignment. In view of our conclusion that CMS is not yet unitary with regard to student assign- ment, we think it is premature to relinquish control over transporta- tion at this stage. On remand, if the district court determines that CMS must remain under court order to correct the current imbalances in student assignment, it should also retain control over transportation to ensure that those imbalances are corrected in a way that is fair to all students.17 17 _________________________________________________________________ we do so on a basis similar to our analysis of the Green factor of trans- portation. Judge Potter found that disparities in student achievement existed but that the disparities (1) were not vestiges of de jure segrega- tion and (2) could not be remedied by any practicable measure. Capacch- ione, 57 F. Supp. 2d at 280-81. An analysis of disparities in student achievement may only be appropriate once the school system has achieved unitary status in other respects. See Swann, 306 F. Supp. at 1309 ("Until unlawful segregation is eliminated, it is idle to speculate whether some of this [achievement] gap can be charged to racial differ- ences or to `socio-economic-cultural' lag."). At the very least, as with transportation, student achievement in this case is inextricably inter- twined with the other Green factors, particularly student assignment. Therefore, having vacated certain of the district court's rulings on unitary status, including its ruling with respect to student assignment, we must also vacate the district court's conclusion on student achievement. We have reviewed and considered the district court's consideration of student discipline, and we affirm the district court's resolution as to this ancillary factor. 17 While the Remedial Plan does not specifically address transportation as a Green factor, it does propose siting new schools in a manner calcu- 44 4. Faculty Our analysis of this factor must take two concerns into account. We must determine both whether CMS has generally eliminated the ves- tiges of discrimination in faculty assignment, and whether the teach- ers assigned to predominantly black schools are of comparable quality to those teaching in schools with large numbers of white students.1818 See Swann, 311 F. Supp. at 268 (final desegregation order directing that the racial composition of faculty assigned to each school reflect that of the system at large, with the proviso that "the competence and experience of teachers in formerly or recently black schools will not be inferior to those in the formerly or recently white schools in the system"). The evidence at trial demonstrated that CMS assigned its faculty in substantial compliance with the desegregation order at least until 1992, when school principals were granted the leeway to actively recruit new teachers without the strictures of maintaining a specific racial proportion. As a result of this gravitation from centralized to site-based control of faculty assignments, a trend away from propor- tionality has emerged. In 1998-99, one-third of the 126 schools cov- ered by the remedial decree had a proportion of black faculty deviating more than ten percent from the system-wide norm (about _________________________________________________________________ lated to promote racial balance in CMS. J.A. 11042. If CMS chooses sites for new schools that are more accessible to the majority of the black population, we presume that fewer black students would have to be bused to the suburbs for purposes of desegretation. A new approach to school siting would address the vestiges of past discrimination, if such vestiges remain, in those areas in which CMS has not yet achieved uni- tary status. 18 The district court considered a particular school to be racially imbal- anced if its proportion of African-American students varied more than fifteen percent from the district-wide average. In 1998-99, African- Americans represented 42.7% of the elementary students in CMS, 41.7% of the middle school students, and 39.6% of the high school students. J.A. 11574. An elementary school would therefore be designated imbalanced-black if more than 57.7% of its students are African- American; conversely, if African-Americans constituted less than 27.7% of the student body, the school would be designated imbalanced-white. 45 twenty-one percent). Prior to the 1992 change in policy, no more than one-sixth of the schools had ever been so situated. We are satisfied that the current trend toward faculty imbalance is neither a vestige of the dual system nor the product of subsequent dis- crimination. There is no evidence that this trend results from legal or administrative compulsion within CMS or from perceptions about the desirability or undesirability of teaching positions in schools that serve students of predominantly one race. In short, we do not perceive a causal relationship between past de jure segregation and the present assignment of faculty members to schools within CMS.19 19 Nor do we think that this trend toward more racially imbalanced faculties has resulted in disparities in the quality of teaching, as mea- sured by the instructors' years of experience and post-graduate work. Indeed, there is no significant difference in experience between facul- ties at imbalanced-black schools as compared to those that are imbalanced-white. Faculties at black schools are about one year less experienced than the district-wide average, while faculties at white schools are correspondingly more seasoned. This disparity may arouse some initial concerns, until one is informed that the typical CMS teacher has spent more than ten years in the classroom. The upshot is that black and white students alike are, with no meaningful distinction, enjoying the benefits of their teachers' substantial experi- ence. The difference in post-graduate education between black-school and white-school faculties is more pronounced. For every three teach- _________________________________________________________________ 19 Even if the pattern of faculty assignments were somehow shown to be a vestige of past official discrimination, the evidence before the dis- trict court casts substantial doubt upon the Board's ability to effect a practicable remedy. See Capacchione, 57 F. Supp. 2d at 258-59: CMS runs the risk of losing significant numbers of teachers if its faculty assignment policies become too restrictive. . . . Another practical problem faced by the district is the fact that it must con- stantly hire thousands of new teachers in the midst of a national teacher shortage . . . [which] is especially pronounced with regard to black teachers, particularly in this region of the coun- try. 46 ers holding advanced degrees who ply their craft at imbalanced-white schools, there are only two similarly qualified teachers assigned to schools that are imbalanced-black. Compared to the district average, white schools have a somewhat larger proportion of these highly trained instructors, while the allotment granted to black schools is slightly less than the norm. Although these facts give us reason for concern, we think it impru- dent to disturb the district court's conclusion that the trial evidence affirmatively disclosed no link between past discrimination and the current asymmetry. Most revealing on this point is that, until now, the issue of teacher quality within CMS has not been contested. The 1970 desegregation order mandating equal competence and experience in faculty assignments was not meant to remedy disparities then exist- ing, but was instead intended to caution against future imbalances. In the intervening thirty years, there is little indication that CMS has neglected to heed the warning inherent in that order. The district court did not clearly err in concluding that the develop- ing disparities in teacher assignments and any (perhaps superficial) deficiency in the quality of instruction currently afforded African- American children are unrelated to the de jure segregation once prev- alent in the school system. We therefore affirm the lower court's find- ing that CMS has attained unitary status with respect to faculty.20 20 5. Staff In substantially the same manner as it spoke to the allocation of teachers, the final desegregation order provided that "the internal operation of each school, and the assignment and management of school employees, of course be conducted on a non-racial, non- discriminatory basis." Swann, 311 F. Supp. at 269. Inasmuch as the Swann plaintiffs raised no challenge to the school system's compli- ance with the desegregation order in this regard, the court below _________________________________________________________________ 20 Although the Board's official position, as outlined in its Remedial Plan, is that remediable vestiges of de jure segregation do remain as to faculty assignments and quality, the clear weight of the evidence is to the contrary. The district court's failure to consider the Plan was therefore harmless in this narrow respect. 47 found CMS to have achieved unitary status with regard to its support staff. We agree, and we affirm that aspect of the district court's judg- ment. 6. Extracurricular activities According to the evidence at trial, African-American students in CMS participate in athletics and hold class office at a rate proportion- ate to their numbers. These same students lag far behind, however, when it comes to participating in co-curricular clubs and honors pro- grams. J.A. 11634. However, the scope of our inquiry concerning extracurricular activities is limited. We need only determine whether the school system permits its students equal access to extracurricular activities, without regard to race. Coalition to Save Our Children v. State Bd. of Educ. of Delaware, 90 F.3d 752, 768-69 (3d Cir. 1996) (citation omitted); see also Swann, 402 U.S. at 18-19 ("[T]he first remedial responsibility of school authorities is to eliminate invidious racial distinctions. With respect to such matters as transportation, sup- porting personnel, and extracurricular activities, no more than this may be necessary. . . . In these areas, normal administrative practice should produce schools of like quality, facilities, and staffs."). The criterion of equal access is surely satisfied in this regard. Par- ticipation in honors programs and co-curricular clubs is strictly volun- tary, and there is no evidence that the lack of participation by African- American students in certain activities reflects the efforts of CMS to exclude them. We discern no error in the district court's conclusions regarding this Green factor, and we therefore affirm its finding that CMS has achieved unitary status with respect to extracurricular activ- ities. C. Pursuant to the foregoing, we conclude that the district court should be affirmed in its determination of unitary status with respect to faculty, staff, extracurricular activities, and student discipline. However, we also conclude that we must vacate and remand for fur- ther consideration in the areas of student assignment, facilities, trans- portation, and student achievement. 48 Volume 2 of 3 49 III. We now turn to the question of whether the Board's expanded magnet schools program violates the Constitution. Even if we had concluded that the district court was correct in decreeing that CMS has now achieved unitary status in all respects, any preceding remedial measures could not be analyzed as if they had been taken by a "de facto" unitary school district. Prior to the decision below, no court had ever determined that CMS had attained unitary status as to any Green factor. As the Capacchione plaintiffs concede, Judge Potter's decision -- not some earlier event-- "terminated [the] injunction" issued by Judge McMillan and affirmed by the Supreme Court. Brief of Appellees at 3. The Supreme Court has made it clear that unitary status is a legal concept that has no "fixed meaning or content" independent of the judicial proceeding in which "unitariness" is conferred. Freeman, 503 U.S. at 486-87. The Court has consistently held that desegregation decrees remain in effect until they are terminated by a court in unam- biguous terms. See Dowell, 498 U.S. at 244-46 (finding district court's prior determination of "unitariness" too ambiguous to dissolve desegregation decree and emphasizing that litigants are entitled to a "rather precise statement" if a decree is to be terminated); Spangler, 427 U.S. at 438-40 (holding that even an unconstitutional court order remains in effect until a "definitive disposition" is reached on the con- 50 stitutional objection). Thus, Judge Potter properly acknowledged that the law would not support a judicial finding that a school district had attained unitary status at a point in time prior to the court's own uni- tariness determination. See Capacchione, 57 F. Supp. 2d at 285 ("The Court finds no legal basis for a finding of de facto unitary status that would abrogate CMS's immunity retroactively. In other words, the termination of court supervision cannot `relate back' to an earlier time."). Judge Potter nevertheless held that the expanded magnet schools program was "ultra vires," beyond the scope of action authorized under the series of injunctions and orders governing desegregation of CMS. Id. After holding that the magnet schools program exceeded the scope of these injunctions and orders, the district court proceeded to analyze whether the program violated the Capacchione plaintiffs' constitutional rights. Applying strict scrutiny, the court concluded that the expanded magnet schools program could not be legally justified. Id. at 287-90. We review the district court's factual findings for clear error and its legal conclusions de novo. See Freeman, 503 U.S. at 474; United States v. Texas, 158 F.3d 299, 306 n.8 (5th Cir. 1998); Little Rock Sch. Dist. v. North Little Rock Sch. Dist., 109 F.3d 514, 516 (8th Cir. 1997). For the reasons that follow, we must reverse. In fact, the injunc- tions and orders governing this case specifically authorize every sig- nificant aspect of the expanded magnet schools program, including the use of racial proportions in assigning students to magnet schools. Furthermore, the Board's obligation to obey these court orders insu- lates it from constitutional attack for actions taken in compliance with them. A. Magnet schools are designed to achieve desegregation by offering some kind of special program or curriculum that will attract students, regardless of race, from throughout a school district. See Missouri v. Jenkins, 495 U.S. 33, 40 n.6 (1990); see also 20 U.S.C. § 7204 (1999) (a magnet school is "a public elementary or secondary school . . . that 51 offers a special curriculum capable of attracting substantial numbers of students of different racial backgrounds"). Magnets "were first con- ceived and developed in large, urban school districts seeking a volun- tary alternative to busing as a means of decreasing racial segregation." J.A. 10654. Critical to the magnet school concept is voluntary choice -- students choose to attend magnet schools because of their desire for the special programs such schools offer. Thus, magnet schools, when not permitted to become dominated by one race, act as "incen- tives for parents to keep their children in the public school system and to send their children to integrated schools." J.A. 15509. Since the 1970s, school boards throughout the country have uti- lized magnet schools as part of desegregation plans that have been routinely approved by the courts. See, e.g., Milliken v. Bradley, 433 U.S. 267, 272 (1977) (Milliken II); Stell v. Savannah-Chatham County Bd. of Educ., 888 F.2d 82, 85-86 (11th Cir. 1989); Little Rock Sch. Dist. v. Pulaski County Special Sch. Dist. No. 1, 839 F.2d 1296, 1309-12 (8th Cir. 1988); United States v. Yonkers Bd. of Educ., 837 F.2d 1181, 1237-39 (2d Cir. 1987); Liddell v. Missouri, 731 F.2d 1294, 1310-11 (8th Cir. 1984). Almost invariably, magnet school programs include an assignment policy that takes race into account "to assure to the greatest extent possible that these voluntary attendance schools not work to under- mine the progress of desegregation." Davis v. East Baton Rouge Par- ish Sch. Bd., 721 F.2d 1425, 1440 (5th Cir. 1983). Such a policy is necessary to prevent magnet schools from "serv[ing] as a haven for those seeking to attend a school predominantly composed of their own race". Morgan v. Kerrigan, 530 F.2d 401, 423 (1st Cir. 1976); see also Jenkins v. Missouri, 942 F.2d 487, 488-89 (8th Cir. 1991); Stell, 888 F.2d at 83; Little Rock Sch. Dist., 839 F.2d at 1311; Yonkers Bd. of Educ., 837 F.2d at 1215; Liddell, 731 F.2d at 1310. The various court decisions reflect a cautious enthusiasm for the utilization of magnet schools, both because such schools allow for more flexibility in student assignment and because they rely more heavily on voluntary choice than mandatory busing. See, e.g., Jenkins, 515 U.S. at 92 ("Magnet schools have the advantage of encouraging voluntary movement of students within a school district in a pattern that aids desegregation on a voluntary basis, without requiring exten- 52 sive busing and redrawing of district boundary lines.").21 21 Further, no authority suggests that magnet programs with race-conscious assign- ment policies constitute an inappropriate vehicle for achieving deseg- regation under a court-sponsored plan. Indeed, it is difficult to see how a magnet program devised, with court approval, to eliminate ves- tiges of segregated schools could do so if school authorities were not permitted to control the racial composition of magnet school enroll- ment. By the 1970s, CMS had established some magnet schools, which it called "optional schools." These early magnet schools had race- conscious assignment policies. See 379 F. Supp. at 1106, 1108; J.A. 2489. Moreover, they offered two special curricula -- "open" and "traditional" -- both of which constituted "very rigorous academic program[s]" not offered in "conventional schools." J.A. 2489, 15683. Judge McMillan approved these magnet schools, including their race- conscious assignment policies. See 379 F. Supp. at 1105-106. In 1992, the Board expanded its magnet schools program into a district-wide system with a wider range of curricular choices. In the expanded magnet schools program, the Board retained the curricula first available in the early magnet or "optional" schools -- the "open" curriculum, emphasizing "interdisciplinary approaches," and the "tra- ditional" curriculum, featuring a "highly structured program." J.A. 16722-23. Furthermore, six of the early magnets that offered such curricula prior to 1992 -- Myers Park, Elizabeth, Hawthorne, Irwin Avenue, Piedmont, and West Charlotte -- continue to do so today under the expanded magnet schools program. Compare J.A. 13448, 13529-40, 15683 (pre-1992 "open" and "traditional" magnets) with J.A. 16722-23 (1998-99 "open" and "traditional" magnets); see also J.A. 10061 (report indicating that pre-1992 magnet schools were incorporated into the 1992 expanded magnet schools program).22 22 _________________________________________________________________ 21 The courts' caution essentially anticipates the position of the Swann plaintiffs in this case: that magnet schools are insufficiently desegrega- tive at best, and that at worst they simply provide an "escape hatch" for white students who would otherwise attend majority black schools, leav- ing those majority black schools even more segregated than they had been before. 22 Judge Potter recognized that the optional schools "were similar to today's magnet schools," both having countywide enrollment and a racial 53 Race is considered in assigning students to the magnet schools instituted under the 1992 expanded program, just as it was in assign- ing students to the original magnet or optional schools. See 379 F. Supp. at 1108. Specifically, under the expanded program, CMS allo- cates 40% of the seats in its magnet schools for black students and 60% for students of other races. This ratio reflects the student popula- tion of the school system, which is approximately 41.0% black, 52.2% white, 3.7% Asian, 2.5% Hispanic, and .5% American Indian.23 23 CMS generally assigns students to its magnet schools using two paral- lel lotteries, one for black students and one for students of other races. When there has been insufficient interest from black students to fill the seats allocated to them in a particular school, CMS has sometimes _________________________________________________________________ balancing target." Capacchione, 57 F. Supp. 2d. at 286. He nonetheless concluded that the schools established after 1992 under the expanded magnet schools program "differ from optional schools in that [the new] magnets offer specialized curricula and thereby confer a benefit above and beyond the regular academic program." Id. at 286-87 n. 49. But nothing in the record supports this view. To the contrary, assuming arguendo that "specialized curricula" constitute a "benefit," the magnet schools instituted after 1992 provide precisely the same "benefit" as the pre-1992 "optional schools." See J.A. 10552 (proposed 1992 pupil assignment plan recommending continuation of six magnet schools already in place); J.A. 15504 (1993 letter noting that magnet schools were called "optional schools" prior to 1992); J.A. 10651 (Summary of Findings From Research on Magnet Schools explaining that "[o]ur optional schools function as magnet schools"); J.A. 13606, 15581 (Stolee Plan explaining that "[the traditional schools presently existing in Charlotte-Mecklenburg are good examples" of curriculum specialty schools, "sometimes called magnet schools"). After all, it was only because the optional schools did offer certain "specialized curricula" that parents (including Michael Grant, one of the Capacchione plaintiffs, J.A. 2489) were willing, well prior to the 1992 expanded magnet schools pro- gram, to enroll their children in desegregated optional schools. See J.A. 13641, 15616. In fact, the original six "open" and "traditional" schools remain among CMS's more heavily subscribed magnets. See J.A. 10292- 340. Myers Park Traditional, for example, had 245 students on its wait- ing list for the 1998-99 school year. See J.A. 2159. 23 For simplicity, we often refer within to non-African-American stu- dents in the magnet schools as "white." 54 refused to allow students of other races to fill those slots. Thus, race may affect a student's chances of being assigned to a magnet school. B. The Capacchione plaintiffs contend that the expanded magnet schools program violates the Equal Protection Clause. Recognizing, if only implicitly, the difficulty in maintaining that actions taken pur- suant to court orders violate the Constitution, they principally argue that the expanded magnet program was not implemented under the court orders governing this case. Specifically, they first contend that the Board's increased reliance on magnet schools constituted a "vol- untary desegregation plan implemented to counteract demographic change," rather than a good faith effort to eliminate the vestiges of discrimination as required by the existing desegregation orders. Sec- ond, they argue that the expanded program's race-conscious assign- ment lottery violated the desegregation orders. Finally, they maintain that even if CMS expanded its magnet schools program pursuant to and in compliance with governing court orders, strict scrutiny none- theless applies and requires that the program be held unconstitutional. We consider each contention in turn. 1. The Capa [Text truncated at 120,000 characters. The full text is on the page linked above.] --- Source: Frix Law Library, https://www.frixlaw.com/law-library/cases/2967477. Public record. Not legal advice.