Full opinion text
OPINION OF THE COURT EN BANC, Lay, Chief Judge, Heaney, Bright, Ross, McMillian, Arnold and Fagg, Circuit Judges, with John R. Gibson, Circuit Judge, concurring in part and dissenting in part, Bowman, Circuit Judge, dissenting. The Caldwell and Liddell plaintiffs, representing black students and parents of the St. Louis City School District, the City School District, and several suburban school districts have entered into a unique and comprehensive settlement agreement designed to further desegregation in the city schools. The United States District Court has approved the agreement and has entered orders to fund the plan. With the exceptions and limitations noted in the opinion, we approve the agreement and the order entered by the district court with respect to: The voluntary transfers of students between the city and suburban schools and the establishment of additional magnet schools and integrative programs in the City School District as necessary to the successful desegregation of the city schools; The quality education programs for the nonintegrated schools in the City School District; The quality education programs for all schools in the City School District, but only insofar as these programs have been shown to be necessary for the city to retain its Class AAA rating or to be essential to the successful desegregation of the city schools as hereinafter set forth; The provisions of the district court’s order requiring the State of Missouri, as the primary constitutional violator, to pay the full cost of city to suburb and suburb to city transfers, magnet schools and integrative programs in the city schools, and one-half of the cost of the quality education programs in the city schools. We decline to approve the district court order insofar as it requires the State to fund student transfers between suburban school districts and to fund magnet schools or integrative programs in those suburban districts; Improved facilities for the city schools. We require further planning, however, before construction begins, to identify with particularity the projects that will be undertaken, and to take account of a probable decline in the city school population in the next few years. We outline the steps that the district court must take before it can require an increase in real estate taxes to fund the City Board’s share of the quality education component of the plan without a vote of the people, and the steps that the court must take before it can require that bonds be issued to fund the City Board’s share of capital improvements without a similar vote. We make it clear, however, that no party found to have violated the Constitution will be permitted to escape its obligation to provide equal educational opportunity to the black children of St. Louis. We make it clear that the suburban schools meeting the goals set forth in the plan will receive a final judgment declaring that they have satisfied their desegregation obligations. Finally, we recognize that the settlement agreement and the district court’s order will have to be modified to conform to this opinion, and we are aware that the cost of the plan, particularly to the State, will be significantly reduced. In our view, however, the changes do not alter the essential character of the plan, and they preserve its constitutionality. The parties to the settlement agreement are required to decide promptly whether they will accept the changes set forth in this opinion. If they refuse to do so, the interdistrict trial will proceed. I. PROCEDURAL HISTORY. In February, 1972, a group of black parents (the Liddell plaintiffs) filed a class action against the City Board, the board members, and school administrators, alleging racial segregation in the city’s schools in violation of the fourteenth amendment. The defendants’ motion to join the State of Missouri and St. Louis County (containing the suburban school districts) as codefendants was denied on December 1, 1973. A year later, the parties entered into a consent agreement which provided for an increase in the number of minority teachers and included a pledge by the City Board to attempt to “relieve the residence-based racial imbalance in the City schools.” Liddell v. Bd. of Educ., 469 F.Supp. 1304, 1310 (E.D.Mo.1979). The case first came before this Court im 1976, when the Caldwell plaintiffs appealed the district court’s denial of their right to intervene. We granted intervention, but declined to pass on the constitutionality of the consent decree. Liddell v. Caldwell, 546 F.2d 768 (8th Cir.) (Liddell I), cert. denied, 433 U.S. 914, 97 S.Ct. 2987, 53 L.Ed.2d 1100 (1976). We encouraged the United States and State of Missouri to intervene, recommended the creation of a biracial citizens committee to assist in formulating a desegregation plan, and suggested voluntary interdistrict student transfers as one remedial tool. Id. at 774. Desegregation plans were developed and submitted to the district court by the City Board, the Liddell plaintiffs, the Caldwell plaintiffs, and the United States as amicus curiae. Before approving any plan, the district court ordered a trial to determine whether there had been a constitutional violation and to frame a remedy if a violation was found. The United States, the City of St. Louis, and two white citizens’ groups were allowed to intervene as plaintiffs. The State of Missouri, the State Board of Education, and the Commissioner of Education were added as defendants. The district court found no constitutional violation, and held that the City Board had achieved a unitary school system in 1954-56 through its “neighborhood school policy.” Liddell v. Bd. of Educ., supra, 469 F.Supp. at 1360-1361. We reversed the district court in Adams v. United States, 620 F.2d 1277 (8th Cir.) (en banc), cert. denied, 449 U.S. 826, 101 S.Ct. 88, 66 L.Ed.2d 29 (1980), holding that the City Board and the State were jointly responsible for maintaining a segregated school system. In reaching this decision, we noted that the Missouri State Constitution had mandated separate schools for “white and colored children” through 1976, that the State had not taken prompt and effective steps to desegregate the city schools after Brown v. Bd. of Educ., 347 U.S. 483, 74 S.Ct. 686, 98 L.Ed. 873 (1954) {Brown I), and that the City Board’s policies and practices since 1956 had contributed to the existing segregation. We remanded to the district court and directed that the schools be promptly desegregated. We suggested the following techniques: (1) Developing and implementing compensatory and remedial educational programs. * * * (2) Developing and implementing programs providing less than full-time integrated learning experiences. (3) Developing and implementing a comprehensive program of exchanging and transferring students with the suburban school districts of St. Louis County. * * * (4) Maintaining existing magnet and specialty schools, and establishing such additional schools as needed to expand opportunities for an integrated education. (5) Establishing an Educational Park. (6) Continuing and expanding a policy of permissive transfers in the district. Adams v. United States, supra, 620 F.2d at 1296-1297 (citations omitted). After holding extensive evidentiary hearings, the district court approved a system-wide desegregation plan for the city schools beginning with the 1980-81 school year. Liddell v. Bd. of Educ., 491 F.Supp. 351 (E.D.Mo.1980). This plan included a comprehensive program of exchanging and transferring students between the city and suburban schools, the establishment of magnet schools and integrative programs, and a quality education component. In approving the plan, the district court concluded: In sum, the State defendants stand before the Court as primary constitutional wrongdoers who have abdicated their affirmative remedial duty. Their efforts to pass the buck among themselves and to other state instrumentalities must be rejected[.] Id. at 359. We affirmed the district court’s plan on appeal. Liddell v. Bd. of Educ., 667 F.2d 643 (8th Cir. 1981) (Liddell III), cert. denied, 454 U.S. 1081, 1091, 102 S.Ct. 634, 656, 70 L.Ed.2d 614, 629 (1982). In so doing, we decided that it was constitutionally permissible to allow a number of all-black schools to remain in the city. We noted that no all-white schools would remain, that a plan of voluntary interdistrict transfers would be initiated, that magnet schools and integrative programs would be established, and that a substantial part of the desegregation budget would be spent to improve the quality of education in the all-black schools. We affirmed the State’s liability for desegregation costs and remanded for continued implementation of the plan. Questions about this plan’s implementation came before us in early 1982, when the State again protested its liability for certain desegregation costs. Liddell v. Bd. of Educ., 677 F.2d 626 (8th Cir.) (Liddell V), cert. denied, 459 U.S. 877, 103 S.Ct. 172, 74 L.Ed.2d 142 (1982). We affirmed the district court’s allocation of costs, placing one-half of the actual desegregation costs on the State. We also required the State to pay the costs of voluntary interdistrict transfers and the costs of merging city and county vocational educational programs. Meanwhile, the City Board and the Liddell and Caldwell plaintiffs continued to seek the consolidation of the city and county schools into a single integrated school district on the theory that the suburban schools had also violated the Constitution. They successfully moved to add the county school districts and St. Louis County officials as defendants to this litigation. We noted that the suburban schools could not be held as constitutional violators without further evidentiary hearings and findings by the district court. We again noted that the State and City Board — already adjudged violators of the Constitution — could be required to fund measures designed to eradicate the remaining vestiges of segregation in the city schools, including measures which involved the voluntary participation of the suburban schools. Liddell V, supra, 677 F.2d at 641. The district court entered an order on August 6, 1982, which disclosed the mandatory interdistrict plan it would impose in the event the suburban school districts were found liable for constitutional violations. This plan would create one unified metropolitan school district with a uniform tax rate. The court then scheduled inter-district liability hearings. Before these hearings were held, however, the City Board, the Liddell plaintiffs, the Caldwell plaintiffs, and all twenty-three county school districts developed a settlement agreement with the assistance of a court-appointed expert and filed a proposed consent decree on March 30, 1983. This agreement settled the plaintiffs’ interdis-trict claims against the county school districts, and also enabled the State and City Board to take important steps to desegregate the city schools through the voluntary participation of the county schools, as we outlined in Liddell V The settlement plan has several components. It provides for voluntary interdis-trict transfers between city and suburban schools and includes fiscal incentives to encourage these transfers. Each county school district which receives enough transfers within five years to satisfy its desegregation obligations under the plan will receive a final judgment. Affirmative hiring requirements and voluntary teacher transfers are included in the plan to assure it will have a substantial impact in the county schools. To attract white student transfers to the city, and also to provide remedial programs for city students, the plan creates additional magnet schools in the city and the county, and has several compensatory and remedial education components. These latter components are designed to improve the quality of education in the city schools, and to make special improvements in the all-black schools. After the parties filed the settlement agreement, the district court conducted hearings in April and May of 1983 to determine whether the settlement plan is fair, reasonable, and adequate. In its July 5, 1983, order, the court concluded the plan met these standards and allocated the costs of the plan between the State and City Board. Liddell v. Bd. of Educ., 567 F.Supp. 1037 (E.D.Mo.1983). The State is totally responsible for the costs of the voluntary interdistrict transfers, the magnet schools, and various part-time and alternative integrative programs. Further, the State will pay one-half of the cost of the quality improvements in the city schools and one-half of the capital improvements required by the plan. The City Board is required to pay the remaining costs. The district court ordered the City Board to submit a bond issue to its voters before February 1, 1984, to fund its share of the capital improvements required under the plan. In the event this bond issue failed to obtain the necessary two-thirds vote the court reserved authority to consider an appropriate order to fund these capital improvements. The district court also deferred a scheduled reduction in the City Board’s operating levy otherwise required by Mo.Rev.Stat. § 164.013 (Proposition C) insofar as this revenue is necessary to fund the City Board’s share of desegregation costs. It further reserved authority to order an increase in the City Board’s property tax rate, following notice and a hearing on the amount, if the revenue necessary to fund the City Board’s constitutional obligation to desegregate the city schools is not otherwise available. Several weeks after the district court entered its order approving the settlement, the State filed a motion to stay the implementation of the plan. The City of St. Louis filed a petition for a writ of prohibition seeking the same result. The district court denied both of these motions, and the State and City of St. Louis appealed to our Court. In an en banc order, Liddell v. Missouri, 717 F.2d 1180 (8th Cir.1983) (Lid-dell VI), we denied the stay with certain exceptions. We froze the number of inter-district transfers and deferred any further district court action concerning the City Board’s property tax rate. We also deferred action on the writ of prohibition until we considered the case on its merits. Appeals were filed from the district court’s July 5, 1983, order by the State of Missouri, the City of St. Louis, the North St. Louis Parents and Citizens for Quality Education, and the St. Louis Teachers Union. The State contends on appeal that the district court erred: (1) in approving additional interdistrict transfers of students, and requiring the State to pay the full cost of the additional transfers; (2) in approving additional magnet schools and part-time integrative programs, and requiring the State to pay their full cost; (3) in approving certain programs to improve the quality of education in the city schools, and requiring the State to pay one-half the cost of these programs; and (4) in ordering a deferral of scheduled property tax reduction for the city schools, and in stating that it would order a further increase in property taxes to fund the City Board’s share of the cost of the quality education programs in the city schools. The City of St. Louis joins in questioning the authority of the district court to enter the taxing order referred to in (4) above. The St. Louis Teachers Union contends that the district court erred in denying its motion to intervene. The Northside Parents Organization contends that the district court erred in failing to provide more extensive relief to the black students who would remain in the nonintegrated schools. The United States did not file a notice of appeal or cross-appeal. It did file a brief and it was permitted to argue its position before the Court en banc. It appears to argue that many of the programs authorized by the district court may be necessary to desegregate the city schools, but questions whether the district court’s factual findings are sufficient to support all aspects of the district court’s remedial order. It asks this Court to remand to the district court to correct the alleged deficiencies. II. INTERDISTRICT TRANSFERS. On July 2,1981, the district court entered an order authorizing voluntary interdistrict transfers and requiring the State to pay the cost of the transfers. The program was initiated at the beginning of the 1981-82 school year, and by the end of the 1982-83 school year, it had grown so that 873 city students were attending county schools and 318 county students were attending city schools. All but seven of the 318 were enrolled in city magnet schools. The State of Missouri paid the cost of these transfers, including transportation costs and fiscal incentives, to the sending and receiving schools. The settlement agreement calls for an expanded program of interdistrict transfers. City-to-county transfers of black students will be permitted to grow incrementally until they reach 15,000. No limit is placed on the county-to-city transfers, but the number is not expected to exceed 3,000. These transfers are expected to be primarily to city magnet schools and programs. Transfers between county districts are also permitted. All student transfers are voluntary. The State’s funding obligations remain as they were under the July 2, 1981, order: It must pay transportation costs and must pay to the receiving district for each transferring student an amount equal to the receiving district’s cost per pupil, less State aid and trust fund allocation. It is further required to provide fiscal incentives to sending districts which may elect payment under one of two formulas: either one-half of the State aid the district would have received had the student not transferred; or, beginning in 1984-85, if a district sends more students than it receives, State aid based on the district’s enrollment for the second prior year. To be eligible for transfer, students of good standing must be in the racial majority in their home districts and must transfer to districts where they would be in the racial minority. After approval of the settlement agreement, transfers rose dramatically. During the current school year, 2,294 city students have transferred to suburban districts and three hundred and eighty-nine suburban students have transferred to. city schools. Thirty-four suburban students have transferred to other suburban districts. One thousand nine-hundred and sixty-five additional city-to-county transfer applications are on file. The settlement agreement provides that participating districts will receive a final judgment releasing them from further liability if they achieve the plan ratio within five years. Litigation is stayed during this period. If the school district does not reach the plan ratio, litigation can be renewed after first pursuing various negotiating procedures. If the liability of any individual school district is litigated, the plaintiffs must prove liability and may not seek reorganization or consolidation of school districts, nor may they seek a minority enrollment exceeding twenty-five percent of the school district. The State argues that the district court order approving the settlement agreement and requiring the State to pay the full cost of interdistrict transfers cannot be sustained because it imposes an interdistrict remedy based on an intradistrict violation. We disagree for two reasons: First, the issue has previously been decided adversely to the State; second, the interdistrict transfers are intrinsic to an effective remedy for the intradistrict violation and are justified by precedent. A. THE PROPRIETY OF THE DISTRICT COURT’S ORDER WITH RESPECT TO INTERDISTRICT TRANSFERS HAS BEEN PREVIOUSLY DECIDED. This Court has repeatedly authorized the interdistrict transfer of students as a fundamental element of an effective remedy for the unconstitutional segregation of the city schools. In Adams v. United States, supra, 620 F.2d at 1296, we specifically approved the development and implementation of “a comprehensive program of exchanging and transferring students with the suburban school districts of St. Louis County.” In Liddell III, supra, 667 F.2d at 650, we rejected the State’s argument that the district court was without authority to formulate an interdistrict plan without finding an interdistrict violation. We also noted that voluntary interdistrict pupil exchanges “must be viewed as a valid part of the attempt to fashion a workable remedy within the City.” Id. at 651. In an order appended to that opinion, we noted that the State had been “judicially determined to be a primary constitutional violator,” and we held that an interdistrict transfer plan would be salutary and would be entirely enforceable against the State. Id. at 659. Finally, in Liddell V, supra, 677 F.2d at 630, we reiterated our conclusion that, because the State had been found a primary constitutional wrongdoer, it can “be required to take those actions which will further the desegregation of the city schools even if the actions required will occur outside the boundaries of the city school district.” After discussing broad-based inter-district proposals and dismissing them as unsuitable, we addressed the proper limits of the district court’s equitable remedial authority: [T]he district court can require the existing defendants — the state and city school board — to take the actions which will help eradicate the remaining vestiges of the government-imposed school segregation in the city schools, including actions which may involve the voluntary participation of the suburban schools. For example, the district court could * * * (4) require the state to provide additional incentives for voluntary interdistrict transfer. Id. at 641-642 (footnote omitted). We did not act hastily or arbitrarily in approving voluntary interdistrict transfers. We outlined the reasons for our decision in Adams v. United States, supra, 620 F.2d at 1291-1297. We reviewed the parties’ proposed remedial alternatives, several of which involved extensive cross-busing between city schools. The Caldwell plaintiffs proposed a seventy-five percent black/twenty-five percent white racial mix within the district. The Liddell plaintiffs, through their expert witness, Dr. David Colton, proposed a four-tier division of the schools by age groups, which would integrate schools above fourth grade to achieve a sixty percent/forty percent or fifty-five percent/forty-five percent ratio of black to white students. All whites above third grade would attend integrated schools and all blacks would receive at least one-third of their education above third grade in integrated schools. The Department of Justice, through its expert witness, Dr. Gary Orfield, proposed maintenance and expansion of integration in all grades, voluntary interdistrict and intradis-trict transfers, magnet schools, integration of personnel, and community involvement. The Board of Education proposed the creation of integrated junior high schools which would funnel students to high schools in a balanced fashion. Magnet schools would supplement these junior high schools. The white parents proposed that the schools be left as they were or, alternatively, that the city and county schools be merged and a comprehensive plan for inter-district student transfers be developed. Of the four plans submitted by the parties, we found that only the Colton and Orfield plans were constitutionally permissible. We rejected the City Board’s plan as too little too late: elementary schools would remain entirely segregated and desegregation of the upper tiers would be delayed four to seven years. We rejected the Caldwell plan because the record supported the district court’s finding that implementation of the plan would probably result in an all-black school system within a few years. We found that the Colton plan was permissible with some substantial changes, but that plan was discarded by the district court after it found that the plan was “educationally unsound” and that it would “fail to achieve effective desegregation.” Liddell v. Bd. of Educ., supra, 491 F.Supp. at 356. The approach suggested by the United States’ expert, Dr. Orfield, was ultimately adopted by the district court as the plan that held “the promise of providing ‘the greatest possible degree of actual desegregation, taking into account the practicalities of the situation.’ ” Id. at 359, citing Davis v. Bd. of School Comm’rs, 402 U.S. 33, 37, 91 S.Ct. 1289, 1292, 28 L.Ed.2d 577 (1971). We reaffirmed our support of the Orfield plan in Liddell III, supra, 667 F.2d at 649-653. We noted that it was the only constitutionally permissible plan submitted that could achieve stable, effective integration while minimizing transportation of students and maintaining integrated schools in integrated neighborhoods. Id. at 650. The State defendants have raised the question of remedial scope twice before the Supreme Court. On June 17, 1981, the State filed a petition for certiorari from our panel opinion in Liddell III. In that petition, the State argued that there was no basis for State liability: The evidence in this case indicates that the State of Missouri took the necessary and appropriate steps to remove the legal underpinnings of segregated schooling as well as affirmatively prohibiting such discrimination. State’s Petition for Certiorari, No. 80-2152, June 17, 1981, at 17. It further argued: The District Court exceeded its authority in ordering the preparation of a plan of voluntary pupil exchanges between the St. Louis School District and nonparty school districts because (1) an interdis-trict violation has neither been pleaded nor proven, and (2) the District Court cannot, consistent with Milliken v. Bradley, [418 U.S. 717, 94 S.Ct. 3112, 41 L.Ed.2d 1069] order the State of Missouri to fund such a voluntary plan simply on the basis of an intradistrict violation. Id. at 20. The Supreme Court denied certiorari. Missouri v. Liddell, 454 U.S. 1091, 102 S.Ct. 656, 70 L.Ed.2d 629 (1981). Not satisfied with this answer, the State raised the same arguments again before our Court in Liddell IV and Liddell V. Unsuccessful in our Court, the State filed a second petition for certiorari with the Supreme Court on April 30, 1982. The State again argued that ordering an inter-district remedy [the 12(a) voluntary transfers, funded by the State] without first finding an inter-district violation and inter-district effect is in conflict with this court’s decision in Milliken v. Bradley I [418 U.S. 717, 94 S.Ct. 3112, 41 L.Ed.2d 1069] [and Hills v. Gautreaux]. State’s Petition for Certiorari, No. 81-2022, April 30, 1982, at 7; see also id. at 10. Again, the Supreme Court denied certiora-ri. Missouri v. Liddell, 459 U.S. 877, 103 S.Ct. 172, 74 L.Ed.2d 142 (1982). Both of the State’s petitions for certiorari came after the Supreme Court’s decision in Hills v. Gautreaux, 425 U.S. 284, 96 S.Ct. 1538, 47 L.Ed.2d 792 (1976). As a result of our previous holdings and of the Supreme Court’s inaction, the use of interdistrict transfers is settled as law of the case. While this doctrine does not foreclose this Court from correcting its errors, it prevents repeated litigation of the same issue and promotes uniformity of decision. In Re Exterior Siding and Aluminum Coil Antitrust Litigation, 696 F.2d 613, 616 (8th Cir.1982), vacated en banc; 705 F.2d 980 (8th Cir. 1983), cert. denied, — U.S.-, 104 S.Ct. 204, 78 L.Ed.2d 178 (1983). We will reconsider a previously decided issue only on a showing of clear error and manifest injustice. United States v. Unger, 700 F.2d 445, 450 n. 10 (8th Cir.), cert. denied, — U.S. -, 104 S.Ct. 339, 78 L.Ed.2d 308 (1983); Wrist-Rocket Mfg. v. Saunders Archery Co., 578 F.2d 727, 730-731 (8th Cir.1978). We are loath to retract our previous declarations on settled issues when a ease returns on appeal; to do so ignores important considerations of judicial economy and ignores our interest in protecting the settled expectations of parties who have conformed their conduct to our guidelines. In this case, our conclusion that State-funded interdistrict transfers are an appropriate remedy is strengthened by our previous invocation of the law of the case doctrine. Liddell V, supra, 677 F.2d at 629-630. The State argues that we should not be bound by our earlier decisions because the magnitude of the proposed plan, with respect both to cost and numbers of students, distinguishes it from existing plans. Neither this Court nor the district court placed any limitation on the number of students that could transfer under the plan in existence during the last two school years, nor were we requested to do so. Moreover, it was clear that the number of transfers would have to be large if the opportunity for an integrated education was to be provided to a significant number of the 30,000 black students that remained in the all-black schools in the city. Notwithstanding our view that the issues regarding interdistrict transfers have been heretofore decided, we again reach the merits of the matter and, alternatively, hold that the plan and the funding order, as they relate to interdistrict transfers, meet constitutional standards. B. THE DISTRICT COURT’S ORDER WITH RESPECT TO INTERDIS-TRICT TRANSFERS MEETS CONSTITUTIONAL STANDARDS. Since Brown v. Bd. of Educ., 349 U.S. 294, 300, 75 S.Ct. 753, 756, 99 L.Ed. 1083 (1955) (Brown II), principles of equity have guided courts in devising remedies to eradicate segregation and its effects. Yet for equitable remedies to pass eonstitutional muster, they must conform to three overlapping criteria. [First], the nature of the desegregation remedy is to be determined by the nature and scope of the constitutional violation. * * * The remedy must therefore be related to “the condition alleged to offend the Constitution.” 'f * * Second, the decree must indeed be remedial in nature, that is, it must be designed as nearly as possible “to restore the victims of discriminatory conduct to the position they would have occupied in the absence of such conduct.” * 4 * Third, the federal courts * * * must take into account the interests of state and local authorities in managing their own affairs, consistent with the Constitution. Milliken v. Bradley, 433 U.S. 267, 280-281, 97 S.Ct. 2749, 2757, 53 L.Ed.2d 745 (1977) (Milliken II) (citations and footnotes omitted). Examination of voluntary interdistrict transfers confirms that, as a remedy for an intradistrict violation, such transfers comply with constitutional standards. 1. The remedy was closely tailored to the nature and scope of the violation. The Missouri Constitution requires the State to provide a free public education. Mo. Const, art. 9, § 1(a). The State supervises instruction, distributes funds for public education to local school districts, approves school bus routes, provides free textbooks, and passes on applications by school districts for federal aid. See Mo. Rev.Stat. §§ 161.092, 163.021, 163.031, 163.-161, 170.051, 170.055; and Liddell v. Bd. of Educ., supra, 469 F.Supp. at 1313-1314. Before the Civil War, Missouri prohibited the creation of schools to teach reading and writing to blacks. Act of February 16, 1847, § 1, 1847 Mo.Laws 103. State-mandated segregation was first imposed in the 1865 Constitution, Article IX § 2. It was reincorporated in the Missouri Constitution of 1945: Article IX specifically provided that separate schools were to be maintained for “white and colored children.” In 1952, the Missouri Supreme Court upheld the constitutionality of Article IX under the United States Constitution. See State ex rel. Hobby v. Disman, 250 S.W.2d 137, 141 (Mo.1952). Article IX was not repealed until 1976. Adams v. United States, supra, 620 F.2d at 1280. Under the segregated system, the State bused suburban black students from St. Louis County into the city’s black schools to maintain the dual system. Id., at 1281. The city schools remained largely segregated until this Court’s decision in Adams. It is clear from the foregoing that the State’s presence in public education is immense and that the State’s Constitution and statutes mandated discrimination against black St. Louis students on the broadest possible basis. It is equally clear that the discriminatory policies continued after the Supreme Court decided Brown I, supra, in 1954. Given the breadth of the State’s violation, it was appropriate for the district court to mandate an equally comprehensive remedy. The potential for integration within the district, however, was limited by the fact that almost eighty percent of the students were black, and by the district court’s finding that if it integrated the city schools by imposing an eighty/twenty ratio in each school, an all-black school system would probably result. With that in mind, the district court properly considered the alternative of voluntary transfers to county districts. The opportunity for effective integration became a reality when the county schools agreed to accept the voluntary transfer of several thousand black students. 2. The remedy restores the victims of discrimination as nearly as possible to the position they would have occupied absent that discrimination. We have heretofore enumerated the alternative remedies suggested by the parties, and we have explained why the district court selected a remedy which included voluntary interdistrict transfers and why this Court approved that remedy. See supra pp. 1303-1304. We are met for the first time on this appeal with a new, or at least a more precisely framed, argument against inter-district transfers. The State asserts that the district court cannot require the State to fund extensive interdistrict transfers unless the record supports and the district court finds that the black children of St. Louis would have attended schools in the county had it not been for the State’s constitutional prohibition against black and white students attending schools together. Nothing in the cases cited by the State suggests or requires us to hold that the district court abused its discretion when it required the State to fund interdis-trict transfers of students to consenting districts. Indeed Milliken II states that the remedy should correct conditions that “flow from such a violation” and should return victims “to the position they would have enjoyed in terms of education,” but for the violation. Milliken II, supra, 433 U.S. at 282, 97 S.Ct. at 2758. This remedy does precisely that: It returns the largest number of victims to integrated schools and provides integrative opportunities and compensatory and remedial programs for those who cannot participate in the transfer plan. As the primary constitutional violator, the State is in no position to complain that some of the victims may elect to transfer to integrated schools in another school district that is willing to accept them. In our view, Hills v. Gautreaux provides precedent for the remedy mandated by the district court. In that case, the Supreme Court considered a remedy against the United States Department of Housing and Urban Development (HUD) for discrimination in public housing in the City of Chicago. The United States Court of Appeals for the Seventh Circuit had reversed the district court’s dismissal and ordered the district court on remand to enter summary judgment against HUD for violations of the Fifth Amendment and the Civil Rights Act of 1964 by knowingly sanctioning and assisting the Chicago Housing Authority’s (CHA) racially discriminatory public housing program. Hills v. Gautreaux, supra, 425 U.S. at 291-292, 96 S.Ct. at 1543-1544. Thereafter, the plaintiffs requested that the district court require HUD to provide public housing outside Chicago’s city limits. The district court refused, holding that the wrongs were committed solely against city residents and within the city’s boundaries. On appeal, the Court of Appeals for the Seventh Circuit reversed and the Supreme Court affirmed. The Supreme Court stated: We reject the contention that, since HUD’s constitutional and statutory violations were committed in Chicago, Milliken precludes an order against HUD that will affect its conduct in the greater metropolitan area. The critical distinction between HUD and the suburban school districts in Milliken is that HUD has been found to have violated the Constitution. That violation provided the necessary predicate for the entry of a remedial order against HUD and, indeed, imposed a duty on the District Court to grant appropriate relief. * * * Our prior decisions counsel that in the event of a constitutional violation “all reasonable methods be available to formulate an effective remedy,” North Carolina State Board of Education v. Swann, 402 U.S. 43, 46 [91 S.Ct. 1284, 1286, 28 L.Ed.2d 586], and that every effort should be made by a federal court to employ those methods “to achieve the greatest possible degree of [relief], taking into account the practicalities of the situation.” Davis v. School Comm’rs of Mobile County, 402 U.S. 33, 37 [91 S.Ct. 1289, 1292, 28 L.Ed.2d 577]. As the Court observed in Swann v. Charlotte-Mecklenburg Board of Education: “Once a right and a violation have been shown, the scope of a district court’s equitable powers to remedy past wrongs is broad, for breadth and flexibility are inherent in equitable remedies.” Hills v. Gautreaux, supra, 425 U.S. at 297, 96 S.Ct. at 1547 (emphasis added; citations omitted). The Supreme Court then discussed Milliken v. Bradley, 418 U.S. 717, 94 S.Ct. 3112, 41 L.Ed.2d 1069 (1974) (Milliken I), and the limitation it imposed on the scope of the federal courts’ equity powers. In Milliken I, the respondents alleged that the Detroit school system was racially segregated and they sought the creation of a unified school district as a remedy. Without finding constitutional violations by the suburban districts and without finding significant segregative effects in those districts, the district court ordered the consolidation of the Detroit school system with fifty-three independent suburban school districts. After the Court of Appeals for the Sixth Circuit affirmed this desegregation order, the Supreme Court reversed, holding that the order exceeded the district court’s equitable powers: the courts must tailor “the scope of the remedy” to fit “the nature and extent of the constitutional violation.” Id. at 744, 94 S.Ct. at 3127. In evaluating the remedy in Hills according to Milliken I’s standards, the Supreme Court noted that nothing in Milliken I “suggests a per se rule that the federal courts lack authority to order parties found to have violated the Constitution to undertake remedial efforts beyond the municipal boundaries of the city where the violation occurred.” Hills v. Gautreaux, supra, 425 U.S. at 298, 96 S.Ct. at 1546 (footnote omitted). In Hills, the Supreme Court approved the remedy because it did not coerce uninvolved governmental units and because CHA and HUD had the authority to operate outside Chicago’s city limits. Id. Justification for requiring the State to fund transfers between city and county schools is stronger than the justification for the remedy in Hills. Its role in education is much broader than HUD’s role in housing. See supra p. 1305. In addition, the breadth, gravity and duration of the State’s violation here was much greater. The violation scarred every student in St. Louis for over five generations and it gained legitimacy through the State Constitution and through the State’s preeminent role in education. In following the Supreme Court’s guidelines in Hills, we echo its conclusion concerning Milliken I. If we barred the use of interdistriet transfers solely because the State’s constitutional violation took place within the city limits of St. Louis, we would transform Milliken [I]’s principled limitation on the exercise of federal judicial authority into an arbitrary and mechanical shield for those found to have engaged in unconstitutional conduct. Hills v. Gautreaux, supra, 425 U.S. at 300, 96 S.Ct. at 1547. 3. The district court’s order with respect to interdistrict transfers does not infringe on State or local government autonomy. The Supreme Court in Hills v. Gau-treaux, supra, 425 U.S. at 298, 96 S.Ct. at 1546, has interpreted Milliken I to mean that district courts may not restructure or coerce local governments or their subdivisions. This remedy does not threaten the autonomy of local school districts; no district will be coerced or reorganized and all districts retain the rights and powers accorded them by state and federal laws. See Hills v. Gautreaux, supra, 425 U.S. at 305-306, 96 S.Ct. at 1550. We also find unpersuasive the State’s argument that funding this remedy will compel other budget cuts, which would interfere with the autonomy of state and local governments. If we accepted this argument, violators of the Constitution could avoid their remedial responsibility through manipulation of their budgets, leaving victims without redress. Simply put, parsimony is no barrier to a constitutional remedy; “it is obvious that vindication of conceded constitutional rights cannot be made dependent upon any theory that it is less expensive to deny than to afford them.” Watson v. Memphis, 373 U.S. 526, 537, 83 S.Ct. 1314, 1321, 10 L.Ed.2d 529 (1963). Interdistrict transfers between the city and the county schools may proceed pursuant to the settlement agreement, subject to the following exceptions: (1) No additional transfers will be permitted for the balance of the current school year. Such transfers would disrupt the education of students in both sending and receiving schools. Planning and recruitment may continue so that enrollment may reach the levels contemplated in the settlement agreement. (2) City-to-county transfers will be limited to a total of 6,000 students in the 1984-85 school year and to not more than 3.000 additional total transfers in each succeeding school year until the limit of 15.000 is reached. A shortfall of enrollment in one year may be made up in succeeding years. (3) In the event the number of applicants for transfer exceeds the spaces available, priority shall be given to applicants who would otherwise attend an all-black school. (4) In Liddell V, supra, 677 F.2d at 631-632, we warned of the need for vigilance to control the costs of desegregation. Budgetary constraints persist, and so does the need for frugality. We are unwilling, however, to accept the State’s suggestion that “complementary zones” be established, which would effectively limit schools that transferees could attend. This would destroy the voluntary nature of the plan. Nevertheless, constant effort and careful planning must be made by all concerned to limit the costs of transportation, insofar as is consistent with the Constitution and the voluntary nature of the plan. C. COUNTY TO COUNTY TRANSFERS. Although we approve State funding of transfers of students between the city and county, we are unable to give similar approval to the funding of transfers of students between county districts. We emphasize again that the objective of transfers between the city and county is the eradication of segregation within the city. Such transfers are closely tailored to the violation and are clearly remedial with respect to that violation, according to the standards announced in Milliken II which were discussed above. Transfers between county districts, however, are not geared to remedy the violation found within the city. Nor does the record establish that inter-county transfers will materially assist in desegregating the city schools. We recognize that some suburban school districts have majority black enrollments and others have nearly all-white enrollments. We acknowledge that the suburban districts would achieve a further degree of desegregation by such transfers. We neither prohibit nor discourage such voluntary transfers between county schools but we cannot compel the State to pay for them absent a finding of an inter-district violation. III. MAGNET SCHOOLS AND INTEGRATIVE PROGRAMS. A. MAGNET SCHOOLS. The district court and this Court previously authorized the creation of magnet schools and integrative programs. About 8,000 students (one-half of whom were blacks) participated in these schools and programs in the 1982-83 school year. Three hundred participants resided in the county. No one suggests that the magnet schools or integrative programs be discontinued. The settlement agreement approved by the district court provides for the expansion or replication of existing magnet schools and programs and the development of new magnet schools and programs — in both the city and the county — with total enrollment to reach 20,000 students, twelve to fourteen thousand to be enrolled in city magnets and the balance in county magnets. The new schools would be phased in over the 1983-87 period. To be eligible for transfer to the magnet schools, students in good standing must be in the racial majority in their home districts and must meet the qualifications for the magnets. Special eligibility requirements allow white students from the city to attend city magnets if the students now attend schools that are less than ten percent or over fifty percent white. Black students in majority black districts are eligible to attend magnet schools and programs in other black majority districts if seats remain open after all of the host district’s black students have been accommodated. The State argues that insufficient attention has been devoted to developing a curriculum designed to attract county students. It also objects to being required to pay the full cost of building and operating the new magnets. Before reviewing the State’s specific arguments, we observe that the utility and propriety of magnets as a desegregation remedy is beyond dispute. In Adams v. United States, supra, 620 F.2d at 1296-1297, we evaluated the remedies we had previously found to be constitutionally permissible. We recommended “[mjaintaining existing magnet and specialty schools, and establishing such additional schools as needed to expand opportunities for an integrated education.” Id. at 1297. We reiterated our approval of magnet schools in Liddell III, supra, 667 F.2d at 658 (emphasis omitted), where, in considering an in-tradistrict remedy, we directed the city and suburban school districts to undertake a “study of the feasibility of establishing magnet schools located in suburban districts with attendance open to students of both the suburbs and the city. * * * The location of these magnet schools should be determined by agreement between the St. Louis Board of Education and the suburban school districts involved.” Finally, in Liddell V, supra, 677 F.2d at 642, we reaffirmed our conclusion that the district court could “require that additional magnet schools be established at state expense within the city or in suburban school districts with the consent of the suburban districts where the schools would be located.” As with interdistrict transfers, our previous determinations in this case concerning magnet schools are law of the case. Had we not in our previous decisions explicitly examined and approved the use of magnet schools and programs, the weight of precedent would nevertheless oblige us now to approve their use. In Milliken II, supra, 433 U.S. at 272, 97 S.Ct. at 2753, the Supreme Court mentioned magnet schools as a supplement to the compensatory and remedial programs which it approved in that case. Dissenting in another case, Justice Powell observed that the Supreme Court in Swann v. Charlotte-Mecklenburg Bd. of Educ., supra, 402 U.S. at 26-27, 91 S.Ct. at 1281, implicitly encouraged the use of magnet schools: Incentives can be employed to encourage [majority-minority] transfers, such as creation of magnet schools providing special educational benefits and state subsidization of those schools that expand their minority enrollments. * * * These and like plans, if adopted voluntarily by States, also could help counter the effects of racial imbalances between school districts that are beyond the reach of judicial correction. Columbus Bd. of Educ. v. Penick, 443 U.S. 449, 488, 99 S.Ct. 2941, 2992, 61 L.Ed.2d 666 (1979). This Court also approved magnets as a means of desegregating the Little Rock schools in Clark v. Bd. of Educ. of Little Rock, 705 F.2d 265, 269, 272 (8th Cir.1983). Courts of Appeals in several other circuits have also approved desegregation plans which include magnets. Arthur v. Nyquist, 712 F.2d 809, 811-813 (2d Cir. 1983); Berry v. School District of Benton Harbor, 698 F.2d 813, 819 (6th Cir.), cert. denied, — U.S. -, 104 S.Ct. 236, 78 L.Ed.2d 227 (1983); United States v. Texas Education Agency, 679 F.2d 1104, 1110 (5th Cir.1982); Hart v. Community School Bd. of Educ., 512 F.2d 37, 54-55 (2d Cir. 1975) (citing successful magnet programs in Boston, Massachusetts; Providence, Rhode Island; and Coney Island, New York); Stout v. Jefferson County Bd. of Educ., 483 F.2d 84, 85 (5th Cir.1973). District courts have also approved plans that include magnets. Tasby v. Wright, 520 F.Supp. 683, 741 (N.D.Tex.1981), affd in part, rev’d in part, on other grounds, 713 F.2d 90 (5th Cir.1983); Smiley v. Blevins, 514 F.Supp. 1248, 1260 (S.D.Tex.1981). A survey of the literature reveals that magnets are being used in at least eighteen cities. Rossell, Magnet Schools as a Desegregation Tool, 14 Urban Education 303, 320 (1979). Despite the widespread approval of magnet schools by the federal courts, critics maintain that magnet schools cannot correct the deep-seated evils of school desegregation. See, e.g., Morgan v. Kerrigan, 530 F.2d 401, 410 & n. 10 (1st Cir.), cert. denied, 426 U.S. 935, 96 S.Ct. 2648, 49 L.Ed.2d 386 (1976); Bradley v. Milliken, 484 F.2d 215, 243 (6th Cir.), rev’d on other grounds, 418 U.S. 717, 94 S.Ct. 3112, 41 L.Ed.2d 1069 (1974); Kelley v. Guinn, 456 F.2d 100, 108-109 (9th Cir. 1972), cert. denied, 413 U.S. 919, 93 S.Ct. 3048, 37 L.Ed.2d 1041 (1973). Yet the criticisms in these cases generally apply to desegregation plans in which magnets are the principal tool in a “freedom of choice” plan. They function differently in the settlement agreement approved here by the district court. Magnet schools are a single element of the panoply of remedies approved by this Court and the district court. Like the magnet schools in Stout v. Jefferson County Bd. of Educ., supra, 483 F.2d at 86, they are “part of a complex and many-faceted” plan. Magnets perform the salutary function of allowing “non-white as well as the white students so enrolled a chance to widen their horizons through the interplay of ideas and the absorption of diverse sub-cultural attitudes.” Hart v. Community School Bd. of Educ., supra, 512 F.2d at 54. Magnet schools under this plan will be distinguished by the features that have made them successful in other cities: individualized teaching, a low pupil-teacher ratio, specialized programs tailored to students’ interests, enriched resources and active recruitment. See Rosenbaum and Presser, Voluntary Racial Integration in a Magnet School, 86 U.Chi.School Rev. 156, 156 (1978); Levine and Eubanks, Attracting Nonminority Students to Magnet Schools in Minority Neighborhoods, 19 Integrateducation 52, 57 (1981). Because they are supplemented by the extensive program of interdistrict transfers and compensatory education, these magnets will not resegregate, nor will they create a dualistic system with elitist schools. We do not believe that the district court erred in ordering the State to pay the full capital and operating cost of magnet schools. As we noted earlier, the State’s status as a violator of the Constitution compels the district court to remedy the deprivations the State has caused. In Lid-dell V, supra, 677 F.2d at 642, we held that the State could be ordered to undertake as a part of its remedial responsibility the development of magnets. Now we reaffirm that conclusion. While we approve magnet schools and affirm the district court’s decision concerning their funding, we see merit in the State’s argument that careful study and planning must precede replication or expansion of magnets. New magnet schools must be approved by the Magnet Review Committee and the district court. The planning process should focus on those schools and programs that present a reasonable probability of attracting suburban white students; only those schools which demonstrate such a probability should be approved. The new schools should be phased in over a period of four years as provided for by the settlement agreement. The total number of students enrolled in city magnet schools shall not exceed 14,-000. We impose an additional limitation on the development of suburban magnets. Although a panel of this Court approved the use of suburban magnet schools in Liddell III, supra, 667 F.2d at 658-659; and Liddell V, supra, 677 F.2d at 641-642, the Court en banc does not believe that the record sufficiently supports this development. The county districts may proceed on their own, of course, without state funding. Any black city students who transfer into county-funded magnet schools would count toward achieving the district’s plan goal and would contribute to the district’s final judgment. State fiscal incentives would include payments to districts sending transferees to county-funded magnets, but the State will not be required to pay the capital or operating costs of county magnet schools as such. B. PART-TIME INTEGRATIVE PROGRAMS. Part-time integrative programs are primarily intended to provide integrative learning experiences for students attending all-black schools. Adams v. United States, supra, 620 F.2d at 1296; Liddell IV, supra, 693 F.2d at 727; Liddell V, supra, 677 F.2d at 642. These programs have been, and should continue to be, an important element of the overall plan to integrate the city schools. In determining the need for continuing the existing programs, or developing new ones, the City Board and the Budget Review Committee must keep the above standard in mind. They must also recognize that the number of black students in nonintegrated schools will decline dramatically over the next four years. We thus approve the district court’s decision insofar as it permits the continuance of part-time integrative programs and requires the State to pay full cost of the approved programs. We do not, however, specifically approve the new or expanded programs or the dollar amounts for these programs listed in the proposed budget (items A.4.10, A.4.11, A.5.01, A.5.02, A.5.04, A.5.05, A.6.01, A.6.03, and A.6.04). We rather require the City Board to resubmit to the Budget Review Committee, discussed infra Section VI, a list of the new or expanded programs that they would propose to implement. The total cost of these programs should not exceed $1 million. Further, these programs must not duplicate any programs approved in the quality education section of this opinion. Any dispute that emerges between the City Board and the State concerning these programs should be submitted for resolution by the Budget Review Committee and the district court in light of this discussion. IV. QUALITY EDUCATION IMPROVEMENTS. The settlement plan approved by the district court includes compensatory and remedial programs to improve the quality of education throughout the St. Louis public schools and additional programs for the same purpose in the nonintegrated schools. The district-wide improvements include a reduction in class size; restoration of art, music, physical education, and extracurricular programs; creation of pre-school centers and all-day kindergarten programs; additional staff to address the needs of handicapped students; additional nursing and counseling staff; and expansion of library and other media resources and services. Administrative improvements include curriculum and staff development, evaluation and performance assessment, and enhanced long-range planning. The additional improvements for the non-integrated schools include a further class-size reduction in grades K through 8, to twenty pupils per teacher; additional remedial instruction time through after-school, Saturday, and summer school programs; parental involvement programs; and alternative education options for black students unable to attend magnet schools. Other programs address motivational needs of students in the all-black schools by stimulating opportunities for student success and recognition, by introducing role models for academic achievement, and by establishing student concerns committees to address the morale, attendance, and behavior issues which emerge during the implementation of the plan. A. LEGAL PRECEDENT FOR INCLUDING COMPENSATORY AND REMEDIAL PROGRAMS IN DESEGREGATION REMEDIES. This Court suggested the necessity for remedial and compensatory programs in Adams v. United States, supra, 620 F.2d at 1296, and reiterated that need in Liddell V, supra, 677 F.2d at 641-642. We thus approve them in principle as law of the case. See supra p. 1304. Moreover, such programs have solid support in the case law as proper components of a desegregation remedy so long as they relate to the constitutional violation, are remedial in nature, and account for state and local autonomy. Milliken II, supra, 433 U.S. at 280-281, 97 S.Ct. at 2757. In Brown I, the Supreme Court recognized that segregation harms black children by generating “a feeling of inferiority as to their status in the community that may affect their hearts and minds in a way unlikely ever to be undone.” Brown I, supra, 347 U.S. at 494, 74 S.Ct. at 691. In light of this harm, federal courts have often required the inclusion of remedial programs in desegregation plans to overcome the inequalities inherent in dual school systems. Milliken II, supra, 433 U.S. at 283, 97 S.Ct. at 2758. See, e.g., Arthur v. Nyquist, supra, 712 F.2d at 811; Oliver v. Kalamazoo Bd. of Educ., 640 F.2d 782, 789-790 (6th Cir.1980); Evans v. Buchanan, 582 F.2d 750, 767-769 (3d Cir.1978) (en banc), cert. denied, 446 U.S. 923, 100 S.Ct. 1862, 64 L.Ed.2d 278 (1980); United States v. Texas, 447 F.2d 441, 448 (5th Cir.1971); United States v. Jefferson County Bd. of Educ., 380 F.2d 385, 394-395 (5th Cir.), cert. denied, 389 U.S. 840, 88 S.Ct. 77, 19 L.Ed.2d 104 (1967); Berry v. School Dist. of Benton Harbor, 515 F.Supp. 344, 369-373 (W.D.Mich.1981), affd and remanded, 698 F.2d 813 (6th Cir.1983); United States v. Bd. of School Comm’rs of Indianapolis, 506 F.Supp. 657, 671-673 (S.D.Ind.1979), vacated in part on other grounds, 637 F.2d 1101 (7th Cir.), cert. denied, 449 U.S. 838, 101 S.Ct. 115, 66 L.Ed.2d 45 (1980). Such programs “assist students who previously attended all-Negro schools when those students transfer to formerly all-white schools____ The remedial programs ... are an integral part of a program for compensatory education to be provided Negro students who have long been disadvantaged by the inequities and discrimination inherent in the dual school system.” Milliken II, supra, 433 U.S. at 284, 97 S.Ct. at 2759 (emphasis in original), quoting Plaquemines Parish School Bd. v. United States, 415 F.2d 817, 831 (5th Cir. 1969). Crucial to the Supreme Court’s analysis in Milliken II is the concept that segregation not only inflicts harm on individual black students, but also builds “inadequacies [into the] * * * educational system.” Milliken II, supra, 433 U.S. at 284, 97 S.Ct. at 2759 (emphasis added). Thus, to remedy the effects of a dual system which operated for decades with the sanction of law, remedial efforts must also concentrate on systemic educational improvements. A secondary remedial objective of the quality education improvements is to enha