Full opinion text
TABLE OF CONTENTS I. BACKGROUND..........................................................1190 II. STANDARD OF REVIEW................................................1192 III. DISCUSSION............................................................1193 A. Admissions Policies and Practices.......................................1193 1. Background Facts........................................... 1193 2. Undergraduate Admissions Standards................................1194 a. District court ruling............................................1194 b. Arguments on appeal...........................................1196 c. Analysis......................................................1198 i. Rejection of plaintiffs’ proposals.............................1199 ii. Reliance on spring screening and summer remedial program____1200 iii. Elimination of existing remedial courses......................1201 iv. Timing...................................................1202 d. Conclusions regarding undergraduate admissions standards.........1203 3. Scholarship Policies................................................1203 a. District court ruling............................................1203 b. Arguments on appeal...........................................1204 c. Analysis......................................................1204 d. Conclusions regarding scholarship policies ........................1209 B. Enhancement of Historically Black Institutions...........................1209 1. Background Facts.................................................1209 2. New Academic Programs...........................................1210 a. District court ruling............................................1210 b. Arguments on appeal...........................................1212 c. Analysis......................................................1213 d. Conclusions regarding new academic programs....................1215 3. Land Grant Programs..............................................1215 a. District court ruling............................................1215 b. Arguments on appeal...........................................1216 c. Analysis......................................................1216 d. Conclusions regarding land grant programs.......................1217 4. Duplication of Programs............................................1217 a. Fordice.......................................................1217 b. District court ruling............................................1218 c. Arguments on appeal...........................................1219 d. Analysis......................................................1220 e. Conclusions regarding program duplication........................1221 5. Funding..........................................................1221 a. District court ruling............................................1221 b. Arguments on appeal...........................................1223 c. Analysis......................................................1223 d. Conclusions regarding funding...................................1225 C. Employment of Black Faculty and Administrators.........................1225 D. System Governance...................................................1227 IV. CONCLUSION...........................................................1228 Before KING, JOLLY, and DENNIS, Circuit Judges. KING, Circuit Judge: This case concerns the obligation of the State of Mississippi and the other defendants to dismantle the system of de jure segregation that was maintained in public universities in Mississippi. After we heard the initial appeal of this case in 1990, the Supreme Court established, for the first time, the standards for determining in the university context whether a state has met its affirmative obligation to dismantle its prior de jure system. We now review the district court’s ruling following trial on remand to determine whether it erred in its application of these standards. For the reasons set forth below, we affirm in part, reverse in part, and remand the case to the district court for further proceedings consistent with this opinion. I. BACKGROUND Mississippi’s system of public four-year universities was formally segregated by race from its inception in 1848 through 1962, when the first black student was admitted to the University of Mississippi by order of this court. See Meredith v. Fair, 306 F.2d 374 (5th Cir.), cert. denied, 371 U.S. 828, 83 S.Ct. 49, 9 L.Ed.2d 66 (1962). The racial identifia-bility of Mississippi’s eight public universities changed little during the decade following the landmark admission of James Meredith. The student composition of the University of Mississippi, Mississippi State University, Mississippi University for Women, University of Southern Mississippi, and Delta State University (collectively, “historically white institutions” or “HWIs”) remained almost entirely white, while that of Jackson State University, Mississippi Valley State University, and Alcorn State University (collectively, “historically black institutions” or “HBIs”) remained almost entirely black. See United States v. Fordice, 505 U.S. 717, 722, 112 S.Ct. 2727, 2732-33, 120 L.Ed.2d 575 (1992). The racial identifiability of these institutions persists to the present. Private plaintiffs initiated this class action in 1975, complaining that Mississippi was maintaining a racially dual system of higher education in violation of the Fifth, Ninth, Thirteenth, and Fourteenth Amendments to the United States Constitution, 42 U.S.C. §§ 1981 and 1983, and Title VI of the Civil Rights Act of 1964, 42 U.S.C- §§ 2000d to 2000d-4a. The United States intervened as plaintiff and alleged violations of the Equal Protection Clause of the Fourteenth Amendment and Title VI. For twelve years the parties attempted to resolve their differences through voluntary dismantlement of the prior segregated system. Unable to achieve ultimate agreement, the parties proceeded to trial in 1987. The district court ruled that Mississippi had discharged its affirmative duty to dismantle the former de jure segregated system of higher education through its adoption and implementation of good-faith, race-neutral policies and procedures in student admissions and other areas. Ayers v. Allain, 674 F.Supp. 1523, 1564 (N.D.Miss.1987) (Ayers I). Sitting en banc, this court affirmed. Ayers v. Allain, 914 F.2d 676 (5th Cir.1990). The United States Supreme Court granted certiorari. Ayers v. Mabus, 499 U.S. 958, 111 S.Ct. 1579, 113 L.Ed.2d 644 (1991). The Supreme Court vacated the judgment and remanded for further proceedings, holding that the mere adoption and implementation of race-neutral policies was insufficient to demonstrate complete abandonment of the racially dual system. Fordice, 505 U.S. at 731, 743, 112 S.Ct. at 2737, 2743. The Court stated that even after a State dismantles its segre-gative admissions policy, there may still be state action that is traceable to the State’s prior de jure segregation and that continues to foster segregation.... If policies traceable to the de jure system are still in force and have discriminatory effects, those policies too must be reformed to the extent practicable and consistent with sound educational practices. Id. at 729, 112 S.Ct. at 2736. Applying this standard, the Court identified admissions standards, program duplication, institutional mission assignments, and continued operation of all eight public universities as a nonexclusive list of “constitutionally suspect” remnants of the prior de jure system, “for even though such policies may be race neutral on their face, they substantially restrict a person’s choice of which institution to enter, and they contribute to the racial identifiability of the eight public universities. Mississippi must justify these policies or eliminate them.” Id. at 733, 112 S.Ct. at 2738. The Court directed that these and “each of the other policies now governing the State’s university system that have been challenged or that are challenged on remand” be examined “in light of the standard that we articulate today.” Id. On remand, the district court ordered each party to submit proposed remedies “to resolve the areas of the State’s liability pursuant to the Supreme Court mandate.” Without conceding liability, defendant Board of Trustees of State Institutions of Higher Learning (the “Board”) responded by presenting a detailed proposal for modification of the higher education system. This proposal contained, among other provisions, uniform standards of admission for all universities, as well as a plan to merge Delta State University and Mississippi Valley State University into one institution to serve students in the Mississippi Delta. The private plaintiffs and the United States (collectively, “plaintiffs”) responded by insisting that the range of constitutionally suspect policies and practices to be examined on remand had yet to be determined. Pursuant to a subsequent court order, plaintiffs identified the following policies and practices for examination: admissions standards that allegedly deny black students equal access to higher education and tend to channel black students to the HBIs; the use of ACT scores as a basis for awarding undergraduate scholarships at the HWIs; maintenance of institutional mission assignments that largely follow historical racial designations; funding policies that disproportionately benefit the HWIs; allocation of academic programs that is unfavorable to the HBIs; allocation of land grant programs between Alcorn State and Mississippi State that is unfavorable to Al-corn; duplication of the HBIs’ programs and course offerings at the HWIs; maintenance of facilities at the HBIs that are inferior to those at the HWIs; employment practices that perpetuate the racial identifiability of the universities and compensate faculty at the HBIs at a lower rate than faculty at the HWIs; maintenance of all eight institutions; and practices that limit the participation of black persons in system governance. Trial commenced on May 9, 1994, following lengthy attempts at settlement. After ten weeks of testimony, the district court made additional findings of fact and conclusions of law. The district court found vestiges of de jure segregation in the areas of undergraduate admissions, institutional mission assignments, funding, equipment availability and library allocations, program duplication, land grant programs, and number of universities. Ayers v. Fordice, 879 F.Supp. 1419, 1477 (N.D.Miss.1995) (Ayers II). The district court entered a remedial decree on March 7,1995. The remedial decree enjoins defendants from maintaining remnants of the prior system and engaging in practices impeding desegregation. Specific relief includes adoption of the uniform admissions standards proposed by the Board and allocation of additional resources to Jackson State University and Alcorn State University. The district court did not order implementation of the Board’s proposal to consolidate Delta State University and Mississippi Valley State University. The decree establishes a Monitoring Committee to monitor implementation of the terms and obligations imposed by the decree. The Monitoring Committee is to consist of three disinterested persons with experience in the field of higher education, agreed upon by the parties and appointed by the court. The Monitoring Committee is to receive and evaluate reports required of defendants and make recommendations to the district court, which has retained jurisdiction over the action. Plaintiffs now contend that the district court left in place practices that are traceable to the prior dual system and that have discriminatory effects and adopted reforms proposed by the Board without examining the soundness or practicability of alternative, less discriminatory proposals. Issues on appeal encompass undergraduate admissions standards, scholarship criteria, enhancement of historically black universities, system governance, and employment. No party appeals the district court’s rejection of the Board’s consolidation proposal. II. STANDARD OF REVIEW The standard set forth by the Supreme Court in Fordice guides our review of the district court’s judgment. Fordice established that “a State does not discharge its constitutional obligations until it eradicates policies and practices traceable to its prior de jure dual system that continue to foster segregation.” 505 U.S. at 728, 112 S.Ct. at 2735. More specifically, [i]f the State perpetuates policies and practices traceable to its prior system that continue to have segregative effects— whether by influencing student enrollment decisions or by fostering segregation in other facets of the university system — and such policies are without sound educational justification and can be practicably eliminated, the State has not satisfied its burden of proving that it has dismantled its prior system. Id. at 731, 112 S.Ct. at 2737. . We have read Fordice to require that “each suspect state policy or practice be analyzed to determine whether it is traceable to the prior de jure system, whether it continues to foster segregation, whether it lacks sound educational justification, and whether its. elimination is practicable.” United States v. Louisiana, 9 F.3d 1159, 1164 (5th Cir.1993). The State’s liability depends upon these factors. Id. Once liability is found, the offending policies and practices “must be reformed to the extent practicable and consistent with sound educational practices.” Fordice, 505 U.S. at 729, 112 S.Ct. at 2736. “[Sjurely the State may not leave in place policies rooted in its prior officially segregated system that serve to maintain the racial identifiability of its universities if those policies can practicably be eliminated without eroding sound educational policies.” Id. at 743, 112 S.Ct. at 2743. Accordingly, we have interpreted the directives of Fordice “as recognizing the need to consider the practicability and soundness of educational practices in determining remedies as well as in making an initial determination of liability.” Louisiana, 9 F.3d at 1164. We apply the directives of Fordice in conjunction with general standards of appellate review. This appeal challenges elements of the district court’s remedial decree and implicates several of its findings and conclusions. We do not disturb the district court’s findings of fact unless they are clearly erroneous, although we freely reassess its conclusions of law under the de novo standard of review. Ross v. Houston Indep. Sch. Dist., 699 F.2d 218, 226 (5th Cir.1983). A third standard applies to our review of the remedial decree itself. A desegregation remedy is an exercise of a trial court’s equitable power and as such is reviewable, within the context of Fordice, for abuse of discretion. Cf. Valley v. Rapides Parish Sch. Bd., 702 F.2d 1221, 1225 (5th Cir.), cert. denied, 464 U.S. 914, 104 S.Ct. 276, 78 L.Ed.2d 256 (1983). m. DISCUSSION A. Admissions Policies and Practices 1. Background Facts In 1961, less than one week after James Meredith applied to the University of Mississippi, the Board adopted a policy requiring all applicants for undergraduate admission to any state institution of higher education to take the American College Test (“ACT”). Ayers I, 674 F.Supp. at 1530-31. Several months later, the Board authorized each university to set a minimum ACT score for eligibility for admission. Id. at 1531. By 1963, the University of Mississippi, Mississippi State University, and the University of Southern Mississippi required an ACT composite score of at least 15 for all freshmen applicants. Id. At the time, the average ACT score among white students was 18, while that for black students was 7. Fordice, 505 U.S. at 734, 112 S.Ct. at 2738-39. When this case was tried initially in 1987, admissions standards for first-time freshman varied along with the historical racial identifi-ability of each institution. Four HWIs continued to require a composite score of at least 15 on the ACT for automatic admission; the other HWI, Mississippi University for Women, required a score of 15-17 together with a high school grade point average of at least 3.0 on a 4.0 scale, or a score of at least 18. Ayers I, 674 F.Supp. at 1533-34. The HBIs required a minimum ACT composite score of 13. Id. at 1534. Based on the undisturbed factual findings of the district court — and unmoved by lower court determinations that the admissions standards derived from policies enacted in the 1970s to redress the problem of student unpreparedness — the Supreme Court concluded in Fordice that the policies were traceable to the de jure system, were originally adopted for a discriminatory purpose, and continued to have discriminatory effects. 505 U.S. at 734, 112 S.Ct. at 2738-39. The Court found that the minimum ACT requirements “restriet[ed] the range of choices of entering students as to which institution they may attend in a way that perpetuate[d] segregation.” Id. Those students who received ACT scores too low to meet the admissions requirements at the HWIs were restricted to the HBIs or community colleges if they wanted a higher education. Id. at 734-35, 112 S.Ct. at 2738-39. As the Court stated, “[plroportionately more blacks than whites face[d] this choice: In 1985, 72 percent of Mississippi’s white high school seniors achieved an ACT composite score of 15 or better, while less than 30 percent of black high school seniors earned that score.” Id. at 735, 112 S.Ct. at 2739. The Court also deemed “constitutionally problematic” the fact that the State denied automatic admission if an applicant did not achieve the minimum ACT score specified for a particular institution, without also considering high school grades as an additional factor in predicting college performance. Id. at 736, 112 S.Ct. at 2739—40. Plaintiffs’ challenges on remand included the use of differential ACT-based admissions policies at the HWIs and HBIs, as well as the use of ACT cutoff scores and alumni connection in the award of undergraduate scholarships at the HWIs. The district court’s ruling on each of these issues is now before us on appeal. 2. Undergraduate Admissions Standards a. District court ruling The district court concluded that “Undergraduate admissions policies and practices are vestiges of de jure segregation that continue to have segregative effects.” Ayers II, 879 F.Supp. at 1477. More specifically, the court found that the admissions standards in place at the time of the 1987 trial were traceable to the prior de jure system and continued to have segregative effects in a system where racially identifiable institutions offer numerous duplicative academic programs. Id. at 1434. The court held that defendants had a duty to eradicate use of the ACT cutoff score “as a sole criterion for admission to the system when the ACT is used in conjunction with differing admissions standards between the HBIs and HWIs.” Id. Although admissions standards had been modified somewhat by the time of the trial on remand, the district court found that they “basically utilized a version of the 1987 standards with various exceptions.” Id. at 1431. In 1989, the ACT was replaced by the Enhanced ACT. Id. at 1430. Scores on the two tests are not equivalent; the American College Testing Program accordingly publishes concordance tables that correlate scores on the old ACT and Enhanced ACT according to percentile rank. The introduction of the Enhanced ACT prompted the Board to solicit recommendations from the eight universities for revised admissions standards based on the new test. Each HWI recommended use of an Enhanced ACT score of 18 for regular admission, which approximated the previous standard of an ACT score of 15. Each HBI recommended use of an Enhanced ACT score of 15 for regular admission, the concordant value of which was 11 on the old ACT. Because the HBIs had previously required an ACT score of at least 13 for regular admission, this recommendation represented an effective lowering of admissions standards at these institutions. Throughout the system, students not qualifying for regular admission could be admitted as “high risk” exceptions. The recommended Enhanced ACT scores for high risk applicants ranged from 14 to 17 at the HWIs, and from 12 to 14 at the HBIs. The Board approved all recommendations. Differential admissions standards thus persisted in the system through the 1994 trial and, as found by the district court, “resulted in the ‘channeling effect’ described in Fordice.” Id. at 1434. The district court’s remedial order responded to the standards in place in 1994. Defendants proposed, and the district court ordered implementation of, new admissions criteria that standardize requirements at all eight universities beginning with applications for admission in the fall of 1996. The new criteria grant “regular admission” to applicants who have (1) a GPA of at least 3.20 in a designated core curriculum, (2) a GPA of at least 2.50 in the core curriculum or class rank in the top 50% and an Enhanced ACT score of at least 16, or (3) a GPA of at least 2.0 in the core curriculum and an Enhanced ACT score of at least 18. Id. at 1477-78. The admissions policy ordered by the district court provides an important alternative to regular admission through a spring screening and summer remedial program for applicants who do not meet the requirements for regular admission. Students participating in the spring screening process will take the Mississippi College Placement Examination (the “accuplacer”) during the spring of their senior year in high school. Based upon these scores, Enhanced ACT subtest scores, and counselor interviews, students will either be admitted for the fall semester or invited to participate in the summer remedial program. The summer program is designed to provide ten to eleven weeks of remedial instruction in reading, writing, and mathematics, taught both in traditional classroom settings and through computer-assisted individual components. Id. at 1478. In addition, the program plan incorporates cultural and recreational activities to “climatize” students to the college campus. Id. Those students who successfully complete the summer program, by passing at minimum the remedial English and mathematics courses, will be admitted in the fall. The district court found that “the new admissions standards through their uniformity will eliminate the prior segregative effects of the previous differential admissions standards between the HBIs and HWIs, noted by the Supreme Court in Fordice.” Id. at 1481. The district court found that as compared with the standards litigated in the 1987 trial, the new standards would result in an overall increase in the number of black students eligible for regular admission to the university system. As compared with the standards in place at the time of the 1994 trial, which were less stringent than in 1987 as a result of the 1989 changes in requirements at the HBIs, the new standards would result in an overall decline in the percentage of black students eligible for regular admission to the system. The district court noted, however, that the summer program offers a distinct opportunity for applicants to gain admission. Id. at 1479. The court found the summer program to be “credible and educationally advanced. In its proposed form, it is considered by its developers as an educationally sound developmental system.” Id. at 1481. The district court concluded that [wjhile the new admissions standards may reduce the number of black students eligible to be admitted to the system without remedial courses required, it is not evident that the new standards will actually reduce the number of black students ultimately admitted to the system as either regular or remediated admittees. Id. Finally, although the State’s community college system is the subject of a separate lawsuit, the district court made findings and ordered relief in this regard because the community college system is relevant to the issue of access to higher education. The court found evidence that the community college system “can have an impact on the admissions policies of the universities and their ability to further diversify institutions of higher learning.” Id. at 1475. The court also found, however, that the community college system in Mississippi is not providing remediation for students unprepared for four-year institutions “to any great degree.” Id. The district court apparently linked this to at least two factors. First, in contrast to the open admissions policy that prevailed at all community colleges when this case was tried in 1987, some community colleges now require minimum ACT scores for admission to certain programs. Id. at 1474-75. Second, the “overwhelming majority” of students who start at the community college level do not transfer to four-year universities. Id. at 1475. The University of Southern Mississippi has the highest proportion of transfer students in its student body, largely attributable to its recruiting efforts and articulation agreements with several community colleges in surrounding regions. Id. Black students transfer at a significantly lower rate than whites, possibly because a high percentage of black students in community colleges are enrolled in two-year vocational programs. The district court concluded that the State “is losing a valuable resource in not' coordinating the admissions requirements and remedial programs between the community colleges and the universities.” Id. The remedial decree contains a provision ordering the Board “to study the feasibility of establishing system-wide coordination of the community colleges in the State in the areas of admissions standards and articulation procedures,” and to report its findings to the Monitoring Committee. Id. at 1496. b. Arguments on appeal The district court’s finding that undergraduate admissions policies and practices are vestiges of de jure segregation that continue to have segregative effects is not contested on appeal. Plaintiffs do contest the remedy thereupon ordered. Plaintiffs’ challenge to the admissions remedy has two parts. First, plaintiffs argue that the district court’s adoption of the Board’s proposed standards was improper because these standards will significantly reduce the number of black students eligible for regular admission to the university system, and thereby disproportionately burden black students with a loss of educational opportunity. Plaintiffs assert that the district court was obligated by Fordice to consider the educational soundness of alternative proposals that would have excluded fewer black students, but failed to do so. Second, plaintiffs argue that the district court’s reliance on the spring screening and summer remedial program to compensate for the projected decline in regular admission of black students was inappropriate because the program was untested and incompletely defined at the time of trial. Plaintiffs contend that although the district court found the summer program to be “credible and educationally advanced,” it did not specifically find that the program would be an effective means of identifying students capable of succeeding in college or that it could achieve the same results as “existing remedial programs.” In addition, plaintiffs argue that the summer program is not a viable option for the many black students who must work during the summer in order to afford to go to college in the fall, and that the community college system currently does not provide an adequate alternative. Plaintiffs therefore argue that the Board should be required to maintain existing remedial courses and to adopt standards that minimize any reduction in the number of black students eligible for admission, at least during the period that the summer program is being tested and the community college system undergoing change. Although their criticisms of the new admissions standards coincide, private plaintiffs and the United States advocate different admissions policies as alternatives. Private plaintiffs proposed below and re-urge here adoption of a tiered admissions policy, in which admissions requirements vary along with the mission of each university, with the most accessible tier having “open admissions.” By “open admissions,” private plaintiffs mean a policy of granting admission to students with a high school diploma and ACT score of 10. Id. at 1480. Under private plaintiffs’ proposal, the three comprehensive universities would use the admissions standards proposed by the Board, and Jackson State University would have open admissions for eight years with the option thereafter of gradually raising admissions standards to the level prevailing at the comprehensive universities. Id. Existing remedial programs would be strengthened in this scheme. The United States proposed below and re-urges here an admissions policy, which was presented to the Board in 1992 but never adopted, in which regular admission would be granted to students achieving (1) a 2.0 GPA in the core curriculum and a minimum of 16 on the Enhanced ACT or (2) a 2.50 GPA in the core, a ranking in the top 50% of the class, and a minimum of 13 on the Enhanced ACT. The United States contends that under this standard, an estimated 73.6% of black students who took the ACT would qualify for admission, as compared to 52.5% or 50.7% under the proposal adopted by the district court. The United States states that “ACT predictive data indicate that, at the [HBIs], where remedial instruction was given, freshmen with these qualifications could be expected to achieve at least a C average.” U.S. Br. at 12. Defendants argue that the new admissions criteria wholly eliminate prior policies traceable to de jure segregation. Defendants contend that the new admissions standards sufficiently address the concerns articulated in Fordice because they do not differentiate between universities according to historical racial designation and do not rely on the ACT as the sole criterion for admission. Defendants argue that under Fordice, the traceable admissions policy was the Board’s particular use of differential ACT cutoff scores, which effectively channeled black students to the HBIs, and not use of the ACT per se. Accordingly, defendants contend that the new policy is not traceable to the prior de jure system and may be implemented because the record discloses that it is educationally sound and was not adopted for a discriminatory purpose. While defendants maintain that Fordice does not require the district court to select the educationally sound alternative with the least discriminatory effect, they argue that even if the district court did have such an obligation, its findings regarding the segregative effect and educational soundness of the new admissions standards effectively discharged it. c. Analysis The district court’s findings that the new criteria for admission are educationally sound and will not perpetuate segregation within the system are not challenged on appeal. Plaintiffs contend, rather, that the district court erred by failing to consider the educational soundness of proposals that would have resulted in a smaller reduction in the number of black students excluded from regular admission. We agree with plaintiffs that it would be inappropriate to remedy the traceable, segre-gative effects of an admissions policy in a system originally designed to limit educational opportunity for black citizens by adopting a policy that itself caused a reduction in meaningful educational opportunity for black citizens. We do not, however, understand the district court to have done so. The district court considered and rejected alternative proposals as educationally unsound, and expressly contemplated that the remedial route to admission could alleviate any potential disproportionate impact on those black students who are capable, with reasonable remediation, of doing college level work. We understand the district court to have determined, in the specific context of formulating an appropriate remedial decree in this case under Fordice, that access to higher education must be provided only to those applicants who can demonstrate, based on educationally sound and constitutionally permissible indicators, an ability (with reasonable remediation) to do college level work and who therefore have a real prospect of earning a degree. The court found that admission of students unprepared to do college level work may result in significant attrition accompanied by unprofitable debt accumulation. Ayers II, 879 F.Supp. at 1435. Fordice does not require that all students who would have been admitted under the prior, unconstitutional admissions standards be admitted under the reformed admissions standards without regard to the educational soundness of the reformed standards. Instead, the district court’s mandate under Fordice was limited to reforming traceable, segregative policies “to the. extent practicable and consistent with sound educational practices.” 505 U.S. at 729, 112 S.Ct. at 2736. Having found admissions policies and practices to be traceable to the de jure system and to have present segregative effects, the district court properly focused its consideration of alternative admissions policies on their educational soundness and potential to eliminate existing segregative effects; its focus, in turn, on ability to do college level work is consistent with both the evidence as presented by plaintiffs and Fordice. i. Rejection of plaintiffs’ proposals The district court set forth in detail the respective admissions standards proposed by private plaintiffs and the United States. See Ayers II, 879 F.Supp. at 1479-80. Although the district court credited expert testimony indicating that differential or tiered admissions standards are both sound and routinely used, id. at 1482, it did not adopt private plaintiffs’ proposal in light of its finding that the open admissions component of this proposal was educationally unsound. Id. at 1481-82. The district court found that universities across the nation generally are moving toward higher admissions requirements, not lower ones. According to the testimony; students in working toward goals will usually do that which is expected of them. If they believe they need not prepare themselves for college by taking the core curriculum in high school, they will not do so. Such unpreparedness may bring them to college campuses unable to execute the rigors of college work and result in low retention rates, college debt accumulations and years expended with no degrees____ It has also been shown that institutions of higher learning which open their doors to unprepared students via open admissions not only do a disservice to many of the admittees, but can lower the quality and, concurrently, the prestige of the institutions generally. Id. at 1482-83. These findings are not clearly erroneous, and the district court did not abuse its discretion in rejecting private plaintiffs’ proposal. Even assuming that tiered admissions could be implemented without open admissions as a component thereof, it was not an abuse of discretion in this context for the district court to opt instead for a policy based on uniform standards. In the Mississippi system of higher education, differential admissions criteria were rooted in the de jure past and fostered both segregation of the races and the public perception that the institutions with lower standards — the HBIs— were of inferior quality. Id. at 1477, 1486. A tiered system would continue to differentiate among institutions based on, their respective missions. See id. at 1482. In light of the history of differential admissions in Mississippi higher education, and in light of its finding that policies and practices governing the missions of the universities are traceable to de jure segregation and continue to have segregative effects, the district court was within its discretion to unify standards across institutions. The standards proposed by the United States met this interest in uniformity, but were fixed at a level that the district court found to be educationally unsound. Under the United States’s proposal, students with a 2.5 GPA and a class rank in the top 50% would qualify for regular admission with an Enhanced ACT score of 13. While this formula adds high school grades and class rank into the eligibility determination, it nevertheless represents a lowering of the ACT score requirement from even post-1989 levels at the HBIs. In contrast, students with identical qualifications would need an Enhanced ACT score of 16 to qualify for regular admission under the Board’s proposal. The district court concluded that the requirements for regular admission under the Board’s proposal were “quite moderate,” and stated that it “does not find persuasive or educationally sound the adoption of open admissions or continually lowering admissions standards, as was done at the HBIs after the 1987 trial.” Id. We understand this finding to encompass the standards endorsed by the United States. Both plaintiffs and defendants cite ACT predictive data in support of their respective proposals. The United States points out that such data indicates that students with the minimum qualifications they propose would be expected to achieve at least a C average by the end of their freshman year at each of the HBIs. We note that such students are predicted to complete their freshman year with grades significantly below a C average, the minimum required for graduation, at any of the HWIs. See PP 39-R. Defendants highlight a different aspect of the same predictive data, which the district court apparently found persuasive: students with the minimum qualifications proposed by the Board would be expected to complete their freshman year with a C average or slightly below at each of the HWIs. The district court’s finding that the Board’s proposed standards are “quite moderate” is indeed supported by the evidence. On this record, the district court could fairly conclude that it would be educationally unsound to adopt an admissions policy under which students could do college level work at only three institutions in the system. We realize that no set of standards is without its flaws. Significantly, as we discuss below, the standards that the district court did adopt provide an alternative route to admission that does not rely on ACT scores whatsoever. The district court’s decision to order implementation of this system, rather than dilute standards for regular admission, was a proper exercise of its discretion. ii. Reliance on spring screening and summer remedial program The district court recognized the likelihood that the Board’s standards would reduce the number of black students eligible for regular admission as compared to then-prevailing standards, and chose to adopt them only in conjunction with the additional opportunity to gain admission through the spring screening and summer remedial program. The district court was unable to conclude that the new standards, which provide an alternative route to admission that does not rely on ACT scores whatsoever, would actually reduce the total number of black students eligible for admission either as regular or remediated admittees. In light of the district court finding that lowering admissions standards “as was done at the HBIs after the 1987 trial” is educationally unsound, the court apparently determined that to the extent any reduction in the number of black students eligible for admission relative to post-1989 standards does take place, it may reflect the educational unsoundness of prior policies. As contemplated, the new standards should result in the identification and admission of those applicants who, with reasonable remediation, can do college level work. This is consistent with For dice’s mandate of a reformed admissions policy that is practicable and educationally sound. The district court also recognized that the spring screening and summer remedial program was untested and its standards not fully established at the time of trial. See id. at 1478-79, 1481. We think that the program was sufficiently defined that the district court did not abuse its discretion in ordering its implementation. If, however, as plaintiffs suggest may be the case, the spring and summer program is unable to any significant degree to achieve its intended objectives of identifying and admitting otherwise eligible applicants — i.e., applicants who could, with reasonable remediation, successfully complete a regular academic program— for whatever reason, then the program must be reevaluated. The district court’s proper retention of jurisdiction over this action indicates its intent to examine this important component of the admissions system once the relevant data becomes available. If the district court ultimately concludes that the spring screening and summer remedial program (as it may be modified) is unable to any significant degree to achieve its objectives, then the court should, if possible, identify and implement another practicable and edu-eationally sound method for achieving those objectives. iii. Elimination of existing remedial courses We have thus far addressed the spring and summer program as a component of the reformed admissions policy. We turn now to the argument made by the plaintiffs that the district court erred in relying upon the summer remedial program to replace the existing remedial courses in the absence of a finding that the summer program could achieve the same results as the universities’ existing remedial courses in enabling students to succeed in and graduate from college. We note in this connection that the plan proposed by the Board provides that “[developmental studies are only offered during the summer session.” In ordering implementation of this plan, the district court tacitly approved the elimination of most, perhaps evén all, of the remedial courses that had been offered by all the universities at issue here, most notably by the HBIs. This is a troubling decision, implicating the reformed policies for regular admission as well as the spring screening and summer remedial program. On the one hand, there was evidence to indicate that an intensive, structured program of remedial instruction during the summer months prior to a student’s immersion in the college experience may actually be more effective at preparing students for college than a more diffused program of remedial instruction throughout the academic year. On the other hand, the district court appeared to base its decision not to consolidate Mississippi Valley State University with Delta State University, at least in part, on the significant percentage of students enrolled in remedial, or developmental, education at Mississippi Valley and on Mississippi Valley’s role as “a significant nurturer of underpre-pared blacks,” id. at 1492, a role that the district court apparently did not want to see eliminated. Further, it is not clear to what extent the operative predictive data assumes the existence of remedial programs insofar as it is based on historical achievement. It is clear that the predictive data relied upon by the State in support of its argument that its proposed admissions standards were “quite moderate” indicate that students who are admitted with the minimum qualifications required under the new standards are not predicted to achieve a C average during their first year at at least three of the HWIs. This suggests, as defendants note in their brief and indicated at oral argument before this court, that many students who are admitted under the reformed standards will need “substantial educational assistance,” possibly including remedial courses. Remedial courses may be an important part of the admissions policy at any school in which a significant number of students are not predicted to achieve a C average during their first year. Plaintiffs did not challenge the State’s existing remediation policies as traceable to the de jure era. There was therefore no requirement, under Fordice, for reformation of those policies as such. However, the Board’s proposed admissions standards (Bd.R-202) treated the adoption of the summer program and the elimination of the existing remedial courses as components of its admissions standards, and the district court, in ordering the implementation of the Board’s proposal, effectively did the same. The principle that apparently underlies the Board’s admissions policy (and, therefore, the district court’s decision) is that, in the ease of any applicant, what can and cannot be accomplished with reasonable remediation is a key element of the admissions decision. Clearly, this principle is educationally sound. But the court’s action in eliminating the existing remedial courses can legitimately be challenged by plaintiffs as an inappropriate feature of the court’s admissions remedy. We have recognized that there are some tensions in the district court’s findings in this regard. In the light of these tensions and the absence of specific consideration of the justification for, or reasonableness of, eliminating these unchallenged courses, we are sufficiently concerned about the district court’s exercise of its discretion in this regard to direct the court on remand to reconsider its decision to eliminate these courses. On remand, the district court should determine if remedial courses are needed to help ensure that students admitted under the new admissions criteria have a realistic chance of achieving academic success. iv. Timing The United States argues that it may take several years for the summer program to be thoroughly implemented, tested, and evaluated and argues that during the interim, an admissions policy that minimizes any reduction in the number of black students eligible for regular admission should be installed. We reject this argument. The summer program has sufficient promise, on the present state of the record, to allow it “to prove itself in operation,” Green v. County Sch. Bd., 391 U.S. 430, 441, 88 S.Ct. 1689, 1696, 20 L.Ed.2d 716 (1968), should the district court decide to continue on that path. There is no reason why, however, reconsideration of the district court’s decision to eliminate the existing remedial courses cannot be done promptly. We intimate no view on the outcome of that reconsideration. d. Conclusions regarding undergraduate admissions standards Except as set forth below, we affirm paragraph 2 of the remedial decree, which reads in relevant part as follows: “The 1995 admissions standards as proposed by the Board for first-time freshmen, effective for the academic year [1996-97], shall be implemented at all universities.” Ayers II, 879 F.Supp. at 1494. We do not affirm paragraph 2 insofar as it eliminates the remedial courses previously offered at each of the eight universities. We remand this latter issue for reconsideration in the light of this opinion. We understand ’ the district court’s continuing jurisdiction to encompass the evaluation of the effectiveness of the spring screening and summer remedial program, as a component of the admissions system, in achieving its intended objectives of identifying and admitting those students who are capable, with reasonable remediation, of doing college level work but who fail to qualify for regular admission. Should the district court ultimately conclude that this program (as it may be modified) is unable to any significant degree to achieve its objectives, then the court will need to identify and implement another method for achieving those objectives. 3. Scholarship Policies a. District court ruling While the district court found that undergraduate admissions policies in general are vestiges of de jure segregation that continue to have segregative effects, it found that scholarship policies in particular are not. On remand, plaintiffs challenged the use of ACT cutoff scores for the award of undergraduate academic scholarships at the HWIs, as well as the use of ACT cutoff scores and alumni connection in the award of nonresident fee waivers for out-of-state admittees. Unlike most other forms of financial aid, the scholarships challenged by plaintiffs are generally awarded on the basis of academic achievement, not financial need, and do not require repayment by the recipient; The district court found a significant disparity in the percentage of nonresident fee waivers awarded by race in any given year. Id. at 1433. The evidence indicated similar disparities in the award of academic scholarships. The district court concluded, however, that [t]he Board’s policy of allowing [nonresident fee waivers] to be based on ACT cutoffs and the use of ACT cutoff scores as the sole criterion for the receipt of academic scholarship monies has not been proven to have linkage with the de juré system, and there is no evidence that these practices currently foster separation of the races such as influencing student- choice. Therefore, reformation of these policies cannot be ordered consistent with the law of the case, absent evidence of discriminatory purpose of which the court finds none. The use of ACT scores in awarding sehol-arships is widespread throughout the United States and generally viewed as educationally sound. Id. at 1434-35 (footnote omitted). The district court did not make a specific finding with regard to the traceability of the alumni connection requirement for nonresident fee waivers. The remedial decree does not order alteration of any of the challenged scholarship policies. b. Arguments on appeal Plaintiffs argue that the district court clearly erred in finding that the use of ACT cutoffs in the award of academic scholarships and nonresident fee waivers at the HWIs is not traceable to the dual system and does not have segregative effects. Although the district court’s findings and conclusions with respect to academic scholarships focus specifically on policies that establish an ACT cutoff score as the sole criterion for award, plaintiffs’ challenge encompasses all instances in which the HWIs require a minimum ACT score for scholarship eligibility. Accordingly, plaintiffs have identified on appeal numerous scholarships at various HWIs that are available only to students with certain minimum ACT scores. Plaintiffs contend that the use of ACT cutoff scores for scholarship eligibility is traceable to the de jure system because under that system ACT cutoff scores were implemented for the purpose of excluding black students from the HWIs. The seg-regative effects of this practice, plaintiffs argue, are evident in the racial disparity in scholarship awards. Because black students receive only a very small proportion of such scholarships, yet are more likely than white students to be in need of financial aid, the policy effectively reduces the number of black students able to attend the HWIs. Moreover, plaintiffs argue that the record does not support the district court’s finding that the use of ACT cutoff scores in the award of scholarships is widespread. Plaintiffs also ■ contend that the district court erred in upholding the practice of limiting nonresident fee waivers to children of an institution’s alumni. Plaintiffs maintain that the alumni connection requirement is traceable to the de jure system in that parents of today’s students were systematically excluded from the HWIs under the de jure system. c. Analysis Although it is clear from the record that undergraduate scholarship policies were litigated on remand, the district court made virtually no fact findings with regard to specific policy criteria or operation. The parties’ original briefing of this issue on appeal was also scant. In response to our request for supplemental briefing, plaintiffs provided a summary of the challenged policies along with the racial breakdown of their distribution for the 1992-93 year (and in one instance, for the 1991-92 year). Defendants have not contested the accuracy of this summary, which is drawn from defendants’ answers to interrogatories and from other evidence introduced by defendants. We therefore accept plaintiffs’ factual summary. According to that summary, the scholarships alleged to be traceable to de jure segregation- and to have present discriminatory effects are as follows: DELTA STATE UNIVERSITY First-time freshman enrollment 1992-93: 21% black Scholarship Name Number of Recipients Dollars Received Minimum ACT Score Black White Total Black White Total Dean’s and Presidential 26 2 160 162 $ 1,375 $ 131,175 132,550 . 1% black 1% black MISSISSIPPI STATE UNIVERSITY First-time freshman enrollment 1992-93: 16% black Scholarship Name Minimum ACT Number of Recipients Dollars Received Score Black White Total Black White Total Entering Freshman ACT 8,000 31 294 299 $ 546,000 $ 555,000 $ 2,000 Sharp Forestry 31 0 7,500 7,500 Entering Freshman ACT 5,000 and Schilhg 29 454 468 16,250 596,836 626,836 Ramsey & Elaine O’Neal and Hearin-Hess 28 41 41 115,500 115,500 Entering Freshmen ACT 4,000, South Central Bell, and Jesse & Lillian Tims 28 248 267 7,944 239,444 261,388 Leadership 20 71 80 3,600 34,450 38,550 John C. Stennis 24 1 6 8 1,000 6,000 8,000 Alumni 21 N/A N/A N/A N/A N/A N/A TOTAL 20 1117 1166 $30,794 $1,545,730 $1,612,774 2% black 2% black MISSISSIPPI UNIVERSITY FOR WOMEN First-time freshman enrollment 1992-93: 21% black Scholarship Name Minimum ACT Number of Recipients Dollars Received Score Black White Total Black White Total Centennial and Eudora Welty 28 26 0 26 $ 0 $ 142,464 $ 142,464 Regional 21 2 68 70 1,200 74,400 75,600 Alumni 21 2 50 52 600 32,540 33,140 Academic 21 10 208 218 3,402 111,500 114,902 Valedictorian 21 0 6 6 0 7,075 7,075 Salutatorian 21 0 6 6 0 4,125 4,125 TOTAL 14 364 378 5,202 $ 372,104 377,306 Jo black ; black First-time freshman enrollment 1991-92: N/A Scholarship Name Minimum ACT Number of Recipients Dollars Received Score Black White Total Black White Total Special Conditions 21 34 154 188 $40,820 $ 139,163 $ 179,983 Academic 25 0 79 79 0 130,425 130,425 TOTAL 34 233 267 $40,820 $ 269,588 $ 310,408 13% black 13% black UNIVERSITY OF MISSISSIPPI First-time freshman enrollment 1992-93: 7% black Scholarship Name Minimum ACT Number of Recipients Dollars Received Score Black White Total Black White Total Children of Nonresident Alumni 21 1 305 307 529,512 $ 1,960 $ 533,432 Children of Faculty & Staff Post-1977 18 10 106 118 14,092 88,540 104,196 Children of Faculty & Staff Pre-1977-78 19 10 104 116 19,780 195,263 215,783 Academic 28 6 683 701 14,130 1,608,555 1,641,805 Academic 30 2 27 29 9,500 105,000 114,500 Academic 22 9 240 253 11,350 244,467 258,642 Special Conditions 22 6 130 140 6,810 211,550 224,240 TOTAL 44 1595 1664 $77,622 $2,982,887 $3,092,598 3% black 3% black UNIVERSITY OF SOUTHERN MISSISSIPPI First-time freshman enrollment 1992-93: 27% black Scholarship Name Minimum ACT Number of Recipients Dollars Received Score Black White Total Black White Total Presidential, Schillig-Baird, Pulley, Pulley, and Gough 29 0 36 36 $ 0 $ 194,043 $ 194,043 Academic Excellence 28 352 371 8,375 773,490 816,860 Regional 25 43 47 0 72,914 79,774 Alumni 21 143 146 1,960 230,333 236,213 TOTAL 574 600 $10,335 1,270,780 $1,326,890 1% black 1% black The district court found that basing scholarship eligibility on ACT cutoff scores is not traceable to the dual system and does not have current segregative effects. We agree with the principle articulated by the district court that use of an ACT cutoff is not unlawful in all circumstances. “Rather, its particular use in any circumstance must be examined to consider whether as a component of the policy challenged, the same is traceable to prior de jure segregation.” Ayers II, 879 F.Supp. at 1434. In light of the facts set out above, however, we conclude that the district court erred in arriving at its findings regarding traceability and segregative effects. The district court may have applied an erroneous view of traceability. As defendants point out in their supplemental letter brief, a traceable policy is one “rooted in” the prior dual system. See Fordice, 505 U.S. at 730 n. 4, 732 n. 6, 743, 112 S.Ct. at 2737 n. 4, 2737 n. 6, 2743. It is only “surviving aspects” of de jure segregation that a state need remedy. See id. at 733, 112 S.Ct. at 2738. That is not to say, however, that a challenged policy as it exists today must have been in effect during the de jure period in order to be constitutionally problematic. The undergraduate admissions criteria that the district court found to be traceable, for instance, had been modified several times since the de jure era but nonetheless were found to be rooted in the prior system. Similarly, the Supreme Court found Mississippi’s scheme of institutional mission classifications to be traceable to de jure segregation even though it was not put in place until several years after termination of official segregation. See id. at 732-33, 739-41, 112 S.Ct. at 2737-38, 2741 — 12. The Court noted that “[t]he institutional mission designations adopted in 1981 have as their antecedents the policies enacted to perpetuate racial separation during the de jure segregated regime.” Id. at 740, 112 S.Ct. at 2742. In United States v. Louisiana, this court implicitly recognized that Louisiana’s open admissions policy could be traceable to that state’s prior de jure system despite its adoption only after de jure segregation had ended. See 9 F.3d at 1167. Because the district court had not addressed the policy’s traceability, we left the issue open for resolution on remand. Id. In this case, plaintiffs concede that the record does not contain evidence directly linking the use of ACT cutoffs for scholarship purposes with any time prior to 1980. Such evidence apparently was not developed because plaintiffs concluded, in our view correctly, that the discriminatory use of ACT cutoffs to exclude black students from the HWIs during the de jure period establishes traceability with respect to all current practices that limit black student access to the HWIs by setting ACT cutoff scores at a level that disproportionately favors white students. Defendants contend that plaintiffs have failed to prove traceability because they have not produced evidence establishing that the practice of using ACT cutoffs in the award of scholarships was initiated either “(i) during de jure segregation, (ii) as an integral component of de jure segregation, (iii) to continue, perpetuate, or further segregation, or (iv) because of some intentionally segregative policy which formerly existed.” This argument misses the mark. First, to the extent defendants suggest it is lacking, evidence of discriminatory purpose is required to establish a constitutional violation only for present policies that are not traceable to the prior system; discriminatory purpose is not an element of traceability itself. Fordice, 505 U.S. at 733 n. 8, 112 S.Ct. at 2738 n. 8. Second, this argument ignores the relationship between scholarship awards and grants of admission, an element missing from the district court’s analysis as well. Scholarship decisions are not wholly independent of admissions in the way that most financial aid determinations are. Indeed, the record indicates that at University of Mississippi, Delta State University, and Mississippi University for Women, the application for admission also constitutes the application for scholarships. It is because scholarships are intended to reward exemplary academic achievement, as defendants point out, that scholarship decision criteria overlap more with those for admission than for financial aid. By their nature, scholarships are designed to attract outstanding students to the awarding institution; that scholarships need not be repaid is a powerful incentive for students to both pursu